Animal Communication Networks

P1: IYP P1: JZZ/JZZ 0521809592agg.xml

P2: JZZ CB777B/Carrington 0 521 80959 2 0521823617agg.xml CU1917B/McGregor 0 521 582361 7

April 9, 2005

15:59

April 12, 2005

This page intentionally left blank

ii

13:42

P1: JZZ/JZZ P2: JZZ 0521823617agg.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

13:42

Animal Communication Networks Most animal communication has evolved and now takes place in the context of a communication network: several signallers and receivers within communication range of each other. This idea follows naturally from the observation that many signals travel further than the average spacing between animals. This is self-evidently true for long-range signals, but at a high density the same is true for short-range signals (e.g. begging calls of nestling birds). This book provides a current summary of research on communication networks and appraises future prospects. It combines information from studies of several taxonomic groups (insects to people via fiddler crabs, fish, frogs, birds and mammals) and several signalling modalities (visual, acoustic and chemical signals). It also specifically addresses the many areas of interface between communication networks and other disciplines (from the evolution of human charitable behaviour to the psychophysics of signal perception, via social behaviour, physiology and mathematical models). P. K. McGregor was Head of the Department of Animal Behaviour at Copenhagen University; he is now Reader in Applied Zoology at Cornwall College, Newquay, UK. He is editor of the journal Bioacoustics and on the editorial board of several other academic journals.

i

P1: JZZ/JZZ P2: JZZ 0521823617agg.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

ii

13:42

P1: JZZ/JZZ P2: JZZ 0521823617agg.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

Animal Communication Networks Edited by

P. K. McGregor University of Copenhagen and Cornwall College, Newquay, UK

iii

13:42

   Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo Cambridge University Press The Edinburgh Building, Cambridge  , UK Published in the United States of America by Cambridge University Press, New York www.cambridge.org Information on this title: www.cambridge.org/9780521823616 © Cambridge University Press 2005 This book is in copyright. Subject to statutory exception and to the provision of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published in print format 2005 - -

---- eBook (EBL) --- eBook (EBL)

- -

---- hardback --- hardback

Cambridge University Press has no responsibility for the persistence or accuracy of s for external or third-party internet websites referred to in this book, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.

P1: JZZ/JZZ P2: JZZ 0521823617agg.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

13:42

Contents

List of contributors viii Preface xiii 1 Introduction 1 Peter K. McGregor

Part I

Behaviours specific to communication networks Introduction 9

2 Eavesdropping in communication networks 13 Tom M. Peake

3 Public, private or anonymous? Facilitating and countering eavesdropping 38 Torben Dabelsteen

4 Performing in front of an audience: signallers and the social environment 63 Ricardo J. Matos & Ingo Schlupp

5 Fighting, mating and networking: pillars of poeciliid sociality 84 Ryan L. Earley & Lee Alan Dugatkin

6 The occurrence and function of victory displays within communication networks 114 John L. Bower

Part II The effects of particular contexts Introduction 129

v

P1: JZZ/JZZ P2: JZZ 0521823617agg.xml

vi

CU1917B/McGregor

0 521 582361 7

April 12, 2005

Contents

7 Enlightened decisions: female assessment and communication networks 133 Ken A. Otter & Laurene Ratcliffe

8 Predation and noise in communication networks of neotropical katydids 152 Alexander B. Lang, Ingeborg Teppner, Manfred Hartbauer & Heiner Römer

9 Nestling begging as a communication network 170 Andrew G. Horn & Marty L. Leonard

10 Redirection of aggression: multiparty signalling within a network? 191 Anahita J. N. Kazem & Filippo Aureli

11 Scent marking and social communication 219 Jane L. Hurst

Part III Communication networks in different taxa Introduction 247 12 Waving in a crowd: fiddler crabs signal in networks 252 Denise S. Pope

13 Anuran choruses as communication networks 277 T. Ulmar Grafe

14 Singing interactions in songbirds: implications for social relations and territorial settlement 300 Marc Naguib

15 Dawn chorus as an interactive communication network 320 John M. Burt & Sandra L. Vehrencamp

16 Eavesdropping and scent over-marking 344 Robert E. Johnston

17 Vocal communication networks in large terrestrial mammals 372 Karen McComb & David Reby

18 Underwater acoustic communication networks in marine mammals 390 Vincent M. Janik

19 Looking for, looking at: social control, honest signals and intimate experience in human evolution and history 416 John L. Locke

13:42

P1: JZZ/JZZ P2: JZZ 0521823617agg.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

13:42

Contents

Part IV Interfaces with other disciplines Introduction 445 20 Perception and acoustic communication networks 451 Ulrike Langemann & Georg M. Klump

21 Hormones, social context and animal communication 481 Rui F. Oliveira

22 Cooperation in communication networks: indirect reciprocity in interactions between cleaner fish and client reef fish 521 Redouan Bshary & Arun D’Souza

23 Fish semiochemicals and the evolution of communication networks 540 Brian D. Wisenden & Norman E. Stacey

24 Cognitive aspects of networks and avian capacities 568 Irene M. Pepperberg

25 Social complexity and the information acquired during eavesdropping by primates and other animals 583 Dorothy L. Cheney & Robert M. Seyfarth

26 Communication networks in a virtual world 604 Andrew M. R. Terry & Robert Lachlan

Index 628

vii

P1: JZZ/JZZ P2: JZZ 0521823617agg.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

13:42

Contributors

Filippo Aureli School of Biological and Earth Sciences, Liverpool John Moores University, Byrom St, Liverpool L3 3AF, UK

John L. Bower Fairhaven Office 348, Fairhaven College, Western Washington University, Bellingham, Washington 98225-9118, USA

Redouan Bshary Department of Zoology, University of Cambridge, Downing St, Cambridge CB2 3EJ, UK. Present address: Evolutionary Psychology and Behavioural Ecology Research Group, School of Biological Sciences, Crown St, University of Liverpool, Liverpool L69 7ZB, UK

John M. Burt Cornell Laboratory of Ornithology, 159 Sapsucker Woods Rd, Ithaca, NY 14850, USA. Present address: Department of Psychology, Box 351525, University of Washington, Seattle, WA 98195-1525, USA

Dorothy L. Cheney Department of Biology, University of Pennsylvania, Philadelphia, PA 19104, USA

Torben Dabelsteen Department of Animal Behaviour, Copenhagen University Zoological Institute, Tagensvej 16, DK-2200 Copenhagen N, Denmark

Arun D’Souza Department of Animal Ecology and Tropical Biology, University of W¨ urzburg, 97074 W¨ urzburg, Germany

Lee Alan Dugatkin Department of Biology, University of Louisville, Louisville, KY 40292, USA

viii

P1: JZZ/JZZ P2: JZZ 0521823617agg.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

13:42

List of contributors

Ryan L. Earley Department of Biology, Georgia State University, 20 Peachtree Center Ave NE, 402 Kell Hall, Atlanta GA 30303, USA

T. Ulmar Grafe Department of Animal Ecology and Tropical Biology, University of W¨ urzburg, 97074 W¨ urzburg, Germany

Manfred Hartbauer Institute of Zoology, Karl-Franzens-University, Universit¨ atsplatz 2, A-8010 Graz, Austria

Andrew G. Horn Department of Biology, Life Science Centre, Dalhousie University, 1355 Oxford St, Halifax, Nova Scotia, Canada B3H 4J1

Jane L. Hurst Faculty of Veterinary Science, University of Liverpool, Leahurst Veterinary Field Station, Neston, South Wirral L64 7TE, UK

Vincent M. Janik Centre for Social Learning and Cognitive Evolution and the Sea Mammal Research Unit, Gatty Marine Laboratory, University of St Andrews, Fife KY16 8LB, UK

Robert E. Johnston Department of Psychology, Uris Hall, Cornell University, Ithaca, NY 14853, USA

Anahita J. N. Kazem School of Biological Sciences, University of Wales Bangor, Brambell Building, Deiniol Rd, Bangor LL57 2UW, UK

Georg M. Klump Carl von Ossietzky Universit¨ at Oldenburg, AG Zoophysiologie and Verhalten, FB 7, 26111 Oldenburg, Germany

Robert Lachlan Department of Biology, Coker Hall, University of North Carolina at Chapel Hill, North Carolina 27599, USA

Alexander B. Lang Institute of Zoology, Karl-Franzens-University, Universit¨ atsplatz 2, A-8010 Graz, Austria

Ulrike Langemann Carl von Ossietzky Universit¨ at Oldenburg, AG Zoophysiologie and Verhalten, FB 7, 26111 Oldenburg, Germany

Marty L. Leonard Department of Biology, Life Science Centre, Dalhousie University, 1355 Oxford St, Halifax, Nova Scotia, Canada B3H 4J1

ix

P1: JZZ/JZZ P2: JZZ 0521823617agg.xml

x

CU1917B/McGregor

0 521 582361 7

April 12, 2005

13:42

List of contributors

John L. Locke New York University, 719 Broadway (Suite 200), New York, NY 10003, USA. Present address: Department of Speech-Language-Hearing Sciences, Lehman College, City University of New York, 250 Bedford Park Boulevard West, Bronx, New York 10468, USA

Ricardo J. Matos Department of Animal Behaviour, Copenhagen University Zoological Institute, Tagensvej 16, DK-2200 Copenhagen N, Denmark

Karen E. McComb Experimental Psychology, School of Biological Sciences, University of Sussex, Falmer, Brighton BN1 9QG, UK

Peter K. McGregor Department of Animal Behaviour, Copenhagen University Zoological Institute, Tagensvej 16, DK-2200 Copenhagen N, Denmark. Present address: Centre for Applied Zoology, Cornwall College Newquay, Trenance Gardens, Newquay Cornwall TR7 2LZ, UK

Marc Naguib Department of Animal Behavior, University Bielefeld, PO Box 10 01 31, 33501 Bielefeld, Germany

Rui F. Oliveira ˜o em Eco-Etologia, Instituto Superior de Psicologia Aplicada, Unidade de Investigac¸a Rua Jardim do Tabaco 34, 1149–041 Lisbon, Portugal

Ken A. Otter Ecosystem Science and Management Program, University of Northern British Columbia, 3333 University Way, Prince George, British Columbia, Canada V2N 4Z9

Tom M. Peake Department of Animal Behaviour, Copenhagen University Zoological Institute, Tagensvej 16, DK-2200 Copenhagen N, Denmark

Irene M. Pepperberg MIT School of Architecture and Planning, Brandeis University and Department of Psychology, Waltham, MA 02454, USA

Denise S. Pope Department of Animal Behaviour, Copenhagen University Zoological Institute, Tagensvej 16, DK-2200 Copenhagen N, Denmark. Present address: Department of Biology, Trinity University, 1 Trinity Place, San Antonio, TX 78212-7200, USA

Laurene Ratcliffe Department of Biology, Queen’s University, Kingston, Ontario, Canada K7L 3N6

David Reby Experimental Psychology, School of Biological Sciences, University of Sussex, Falmer, Brighton BN1 9QG, UK

P1: JZZ/JZZ P2: JZZ 0521823617agg.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

13:42

List of contributors

Heiner R¨ omer Institute for Zoology, Karl-Franzens-University, Universit¨ atsplatz 2, A-8010 Graz, Austria

Ingo Schlupp Zoologisches Institut, Universit¨ at Z¨ urich, Winterthurerstrasse 190, CH-8057 Z¨ urich, Switzerland and Section of Integrative Biology C0930, University of Texas, Austin, TX 78712, USA. Present address: Biozentrum Grindel, Universit¨ at Hamburg, Martin-Luther-King Pl. 3, D-20146 Hamburg, Germany

Robert M. Seyfarth Department of Psychology, University of Pennsylvania, Philadelphia, PA 19104-6196, USA

Norman E. Stacey Department of Biological Sciences, University of Alberta, Edmonton, Alberta, Canada T6G 2E9

Ingeborg Teppner Institute of Zoology, Karl-Franzens-University, Universit¨ atsplatz 2, A-8010 Graz, Austria

Andrew M. R. Terry Department of Animal Behaviour, Copenhagen University Zoological Institute, Tagensvej 16, DK-2200 Copenhagen N, Denmark. Present address: IUCN – The World Conservation Union, Regional Office for Europe, Rue Vergote 15, 1030 Bruxelles, Belgium

Sandra L. Vehrencamp Cornell Laboratory of Ornithology, 159 Sapsucker Woods Rd, Ithaca, NY 14850, USA

Brian D. Wisenden Department of Biology, Minnesota State University Moorhead, 1104 7th St S, Moorhead, MN 56563, USA

xi

P1: JZZ/JZZ P2: JZZ 0521823617agg.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

xii

13:42

P1: JZZ/JZZ P2: JZZ 0521823617agg.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

13:42

Preface

This book attempts to reflect the state of current research on communication networks: groupings of several individuals that constitute the social context in which communication takes place. In my view, a structured collection of chapters by active researchers best conveys the excitement of the research findings as well as the underlying expertise of the authors, especially when a wide range of taxa and signalling modalities are addressed. The motivation to edit such a book came from the interest in the topic that was evident after seminars and conference presentations. However, it was the symposium on communication networks at the XXVIIth International Ethological Conference held in T¨ ubingen that converted motivation into action. The symposium showed (at least to my satisfaction) how well the topic integrated research on different taxa and signalling modalities. It was also the opportunity to meet Shana Coates of Cambridge University Press and to appreciate her enthusiasm for a ‘book of the symposium’. The book has turned out to be much more than a collection of symposium papers. First, it covers considerably more ground in its 26 chapters than was possible in a symposium of nine spoken papers. Second, some of the stimulating informal discussions that characterize a good conference have contributed to the section introductions. However, the main ‘added value’ comes from the willingness of the authors to comment on the chapters of others, to incorporate comments and cross-references into their own chapters and, above all, to look at communication from a network perspective. In many instances, this has led to insights that are likely to have a major effect on the direction of research on animal communication – real Eureka moments. It has been a privilege to share in these moments. Many people deserve my thanks for the role they have played in the creation of this book. Marc Naguib suggested that we submit the symposium topic

xiii

P1: JZZ/JZZ P2: JZZ 0521823617agg.xml

xiv

CU1917B/McGregor

0 521 582361 7

April 12, 2005

13:42

Preface ‘communication networks’ to the IEC scientific committee. This committee and the conference main organizer, Raimund Appfelbach, were kind enough to accept the topic and in doing so set the ball rolling. Shana Coates of Cambridge University Press gently, but firmly, ensured that my timescale for editing a book on communication networks was advanced from ‘some time in the future’to ‘in the next couple of months’. Considerable credit is also due to Tom Peake and Andrew Terry, who applied their own particular brand of pressure (accompanied by several cappuccinos) to stimulate me to draft the book proposal on the flight back from T¨ ubingen. Shana gave excellent advice in the early stages of the book, since when Tracey Sanderson and Martin Griffiths have overseen production. Of course, there would be no book without the authors. I am very grateful to all of them for finding time in overcrowded schedules to write their own chapters and to comment on those of others. Denmark’s Statens Naturvidenskabelige Forskningsr˚ ad has supported my research for the last 5 years. København Universitet supported me during the initial stages of the book, but the bulk of the work was done with support from the EU and Cornwall College via a Marie Curie Category 40 Fellowship. Last, but by no means least, I thank Leonie and Tom McGregor – for sustaining me throughout the project with their love, good humour and flexibility at all times, particularly when the going got tough.

P1: IKB/... P2: KOD 0521823617c01.xml

CU1917B/McGregor

0 521 582361 7

April 10, 2005

15:52

1

Introduction peter k. mcgregor University of Copenhagen, Denmark and Cornwall College, Newquay, UK

Some of the most conspicuous behaviours performed by an animal are related to communication – communication that mediates reproduction and survival. As explained below, a knowledge of animal communication is important in more respects than simply its role in understanding such fundamental aspects of animals’lives. This book is about a perspective that can increase our understanding of animal communication. One way in which animal communication is important is that it interfaces with and links several fields of study. In the field of behaviour, for example, communication is often used to illustrate Niko Tinbergen’s four types of question (function, mechanism, development and evolution) and how the answers complement each other (e.g. Krebs & Davies, 1993). Communication has interfaces with many other areas of biology including evolution, ecology, population genetics, neurobiology and physiology. For example, it can be a window into the cognitive worlds of animals (e.g. Ch. 24). Links with other sciences are shown by the use of ideas and techniques from psychology to understand how communication is perceived (Ch. 20), and using information from physics and chemistry to explain how communication is achieved (e.g. Bradbury & Vehrencamp, 1998). Communication cannot occur in isolation; it is an inherently social behaviour. This makes it even more surprising that the wider social context in which communication takes place is rarely considered explicitly. As explained in the next paragraph, it is likely that communication commonly occurs in the context of a network of several animals. This chapter is both a brief introduction to this context – animal communication networks – and an explanation of this book’s structure.

Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

1

P1: IKB/... P2: KOD 0521823617c01.xml

2

CU1917B/McGregor

0 521 582361 7

April 10, 2005

15:52

P. K. McGregor About communication networks A communication network is a group of several animals within signalling and receiving range of each other. If signals travel further than the average spacing between individuals, then there is potential for a communication network to exist. This is as true for the ocean-spanning songs of whales as it is for the begging calls of songbird nestlings crammed into a nest cavity, and it is why networks can be considered to be the commonest context in which communication occurs (e.g. McGregor & Peake, 2000). This would seem to be stating the obvious, especially to those new to the field of animal communication. Indeed, those studying chorusing animals, particularly insects and anuran amphibians, have long adopted a network perspective and recognized the importance of doing so (e.g. Otte, 1974). However, it is only relatively recently that other types of communication have been considered explicitly in a network context. Communication was, and still is in many instances, treated as occurring between two individuals – the signaller–receiver dyad – perhaps because this is the simplest relationship possible between the three basic components found in communication (the signaller, the signal and the receiver). In this sense, a dyadic view of communication follows from the stricture of Occam’s razor (also known as the law of parsimony) to employ the simplest explanation consistent with the facts. While agreeing wholeheartedly with this standard scientific practice, it is clear that a dyadic view of communication is often not consistent with the facts. One example is the high signal level used in close-range aggressive encounters – human antagonists nose to nose, yet shouting at each other – surely high signal levels are not needed to achieve signalling at such close range? In a network context, such high levels make more sense, because there may be more distant intended receivers (the gathering crowd in the human example) in addition to the opponent (Zahavi, 1979). Many further examples of communication that are best considered in the context of a communication network are found throughout this book. Another reason for explicitly considering communication in a network context is that it identifies communication behaviours that cannot occur in a dyad. A good example is eavesdropping, particularly social eavesdropping (Ch. 2) in which the eavesdropper extracts information from a signalling interaction between others. Social eavesdropping requires a minimum of three individuals (one eavesdropping, two more interacting) and, therefore, falls outside a dyadic view of communication. The evidence for eavesdropping and its wider implications (e.g. for comparative cognition) is presented in many of the chapters of this book. Eavesdropping and similar network behaviours discussed in this book are considered by many to be a compelling reason to adopt a network perspective.

P1: IKB/... P2: KOD 0521823617c01.xml

CU1917B/McGregor

0 521 582361 7

April 10, 2005

15:52

Introduction Communication networks and eavesdropping

It is perhaps worth emphasizing that, while eavesdropping is a good example of communication network behaviour, it is not the only one, and the value of the communication network perspective does not depend on a demonstration of eavesdropping. The reason for its current prominence is that it was considered first and, therefore, at the moment it is more prevalent in the literature. There is no merit in shoe-horning a natural example into a definition of eavesdropping, nor in judging the value of any natural communication behaviour by how well it fits this (or any other) definition. As several chapters demonstrate (e.g. Chs. 9 and 23), such examples from the real world can probe and challenge our definitions (e.g. of interactions and of communication more generally) and the thinking that follows from them. The result can be considerable insight and lead to progress for the whole field of communication.

A note on definitions Clear and workable definitions are the essential basis for meaningful discussion. I have tried to ensure that terms are used clearly and consistently within a chapter, but there may be good reasons why chapters differ in the detail of their definitions (e.g. for reasons discussed in the previous paragraph). There are no instances in this book where the same term is used in a markedly different way in different chapters, but readers should bear in mind that the detail of the definition may be important to the topics discussed by the chapter. There are two nice illustrations of the problems that definitions can create. The first concerns eavesdropping. Alan Grafen pointed out a problem with the term eavesdropping after I had used it when presenting ideas on communication networks at the Royal Society Meeting on Signalling in 1992 (McGregor, 1993). The problem he foresaw was that in everyday use the term means secret information gathering, and it was clear to him that there may be advantages to the signallers in providing information (i.e. promoting eavesdropping), especially if the signaller had won the agonistic contest (see also Zahavi, 1979). The everyday meaning of eavesdropping and its implicit association with acoustic signals have been at the root of several misunderstandings that could perhaps have been avoided if a more neutral term had been used (at the time Grafen suggested type II receivers). Tom Peake has sorted out this and other problems to do with definitions of eavesdropping with admirable clarity in Ch. 2. Nevertheless, information gathered without the source’s knowledge may have particular value, as John Locke discusses in Ch. 19. I think this demonstrates that identifying the secrecy or otherwise of information gathering is the route to progress, rather than rigidly applying a definition.

3

P1: IKB/... P2: KOD 0521823617c01.xml

4

CU1917B/McGregor

0 521 582361 7

April 10, 2005

15:52

P. K. McGregor The second example concerns the relationship between information and communication. In my view, the terms are clearly not synonymous; rather signals are a subset of information because they are specialized to transmit information (more details in McGregor & Peake (2000)). This could have created a problem with semiochemicals: if they are not signals (i.e. they contain information but are not specialized to transmit it) then the behaviour involving them is not communication and the concept of communication networks would not apply. Fortunately for the book, Brian Wisenden and Norm Stacey thought carefully about the issue and realised that there were many important similarities that gave them an opportunity to discuss the functional and evolutionary relationships between information, signals and networks (Ch. 23). So a problem arising from definitions has given real insight, rather than the acrimonious defence of definitions that is all too common in the literature.

About this book Coverage

There are several types of book on animal communication. Some are synoptic treatments of the whole topic (e.g. Hauser, 1996; Bradbury & Vehrencamp, 1998) whereas others concentrate on particular types of signal such as pheromones (Wyatt, 2003) or on a group of animals such as arthropods (Greenfield, 2002). Many books do both, for example dealing with acoustic communication in insects (Gerhardt & Huber, 2002) or birds (Kroodsma & Miller, 1996). This book is rather different in that it looks at a specific topic in communication and covers several modalities and taxonomic groups. Organization

Each chapter has been written so that it can be read alone, since this is a common way for edited volumes to be read. Inevitably, this has led to some similarity between chapters in their opening remarks, but I think this is more than offset by each chapter having its own reference section. The many cross-references to other chapters in the book also show the extent to which authors have taken account of material in other chapters and made links between them. A second way in which the book has been given overall coherence is to group the chapters into four parts that reflect major aspects of communication networks. Each of these parts is prefaced by a short overview that identifies chapter themes and highlights some of the issues that remain to be tackled. The fact that many chapters could have been put into any of the four parts further demonstrates the extent of overall coherence of the book and the wide-ranging nature of the

P1: IKB/... P2: KOD 0521823617c01.xml

CU1917B/McGregor

0 521 582361 7

April 10, 2005

15:52

Introduction chapters. Within each part, there is no particular order of chapters, although in Part III the order is loosely phylogenetic. The chapters grouped into Part I deal with communication behaviours, such as eavesdropping and audience effects, that involve three or more individuals (i.e. a communication network) and as such fall outside the ‘classical’ or traditional dyadic (one signaller and one receiver) approach to communication. Part II groups particular contexts that are fruitful to consider from a communication network perspective: mate choice, predation, begging, aggression and scent marking. The reason for grouping chapters in Part III is taxonomic: from fiddler crabs to humans via most groups of vertebrate. While communication networks may be more or less ubiquitous, features of different taxa (e.g. main senses, social organization) can have a major effect on the details of communication networks and provide insight into the topic as a whole. The final part contains chapters that, to a greater or lesser degree, link communication and other disciplines in biology and more widely in science. From the evidence of these chapters, a network perspective seems to be particularly valuable at such subject interfaces.

Summary There are several reasons for considering that the natural context in which communication occurs (and in which it has evolved) is a network of several animals in signalling and receiving range of each other. However, this context has not been considered explicitly in many studies of animal communication. The chapters in this book apply a communication network perspective to a variety of taxa using a number of signal modalities in several circumstances. The results are illuminating. To modify a marketing phrase used for mobile phones: the future is bright; the future is a network view of communication.

References Bradbury, J. W. & Vehrencamp, S. L. 1998. The Principles of Animal Communication. Sunderland, MA: Sinauer. Gerhardt, H. C. & Huber, F. 2002. Acoustic Communication in Insects and Anurans: Common Problems and Diverse Solutions. Chicago, IL: Chicago University Press. Greenfield, M. D. 2002. Signalers and Receivers: Mechanisms and Evolution of Arthropod Communication. Oxford: Oxford University Press. Hauser, M. D. 1996. The Evolution of Communication. Cambridge, MA: MIT Press. Krebs, J. R. & Davies, N. B. 1993. An Introduction to Behavioural Ecology, 3rd edn. Oxford: Blackwell Scientific.

5

P1: IKB/... P2: KOD 0521823617c01.xml

6

CU1917B/McGregor

0 521 582361 7

April 10, 2005

15:52

P. K. McGregor Kroodsma, D. E. & Miller, E. H. 1996. Ecology and Evolution of Acoustic Communication in Birds. Ithaca, NY: Cornell University Press. McGregor, P. K. 1993. Signalling in territorial systems: a context for individual identification, ranging and eavesdropping. Philosophical Transactions of the Royal Society of London, Series B, 340, 237–244. McGregor, P. K. & Peake, T. M. 2000. Communication networks: social environments for receiving and signalling behaviour. Acta Ethologica, 2, 71–81. Otte, D. 1974. Effects and functions in the evolution of signaling systems. Annual Review of Ecology and Systematics, 5, 385–417. Wyatt, T. D. 2003. Pheromones and Animal Behaviour: Communication by Smell and Taste. Cambridge, UK: Cambridge University Press. Zahavi, A. 1979. Why shouting? American Nauralist, 113, 155–156.

P1: JZZ/IKB P2: JZZ/... 0521840716agg.xml CU1917B/McGregor

0 521 582361 7

Part I B E H A V I O U R S S P E C I F I C T O COMMUNICATION NETWORKS

7

April 10, 2005

16:0

P1: JZZ/IKB P2: JZZ/... 0521840716agg.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

8

16:0

P1: JZZ/IKB P2: JZZ/... 0521840716agg.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

16:0

Introduction

The reason for grouping together the chapters that appear in this part of the book is that each of them concerns communication behaviours that are best viewed from a communication network perspective, rather than from the more common dyadic (one signaller to one receiver) standpoint. It is a fact that, with the exception of choruses, most studies to date have implicitly or explicitly considered communication between a dyad. Although the communication network perspective of several signalling and receiving individuals seems to follow logically from what we know of natural communication, the dyadic viewpoint has historical precedence and considerable inertia. A network perspective will become more commonly adopted only if it is clearly better able to explain communication behaviours than a dyadic approach. It is for this reason that a network perspective has long been adopted in studies of choruses; the effect on an individual’s signal timing of the signals of nearby conspecifics can be striking patterns, such as signal synchrony in the chorus (e.g. Greenfield, 2002; Ch. 13). Such patterns cannot be explained by considering communication as a dyad. All of the chapters in this book demonstrate the value of adopting a network perspective; however, it gives this demonstration more emphasis to begin with a section covering communication behaviours that are particularly suited to, or associated with, a network perspective. Eavesdropping In Ch. 1, eavesdropping is identified as a receiving behaviour that has been particularly identified with, and is only possible in, a communication network. The first two chapters of Part I look at eavesdropping in more detail. Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

9

P1: JZZ/IKB P2: JZZ/... 0521840716agg.xml CU1917B/McGregor

10

0 521 582361 7

April 10, 2005

16:0

Part I In Ch. 2, Tom Peake summarizes the evidence for eavesdropping in different contexts and also discusses the diverse use of the term in the literature. His division of eavesdropping into interceptive eavesdropping (e.g. predators locating prey from prey signals) and social eavesdropping (extracting information from a signalling interaction) is an important clarification. However, as Tom points out, clarifying definitions is more important as means of moving arguments on from the question of whether a given behaviour can be called eavesdropping or not and towards a more fruitful and general approach based on the nature of information transfer. Torben Dabelsteen deals mainly with social eavesdropping on the acoustic signals of birds in Ch. 3. He identifies the potential costs and benefits of eavesdropping and uses information from studies of how bird song transmits through the habitat to explore how eavesdropping is best achieved. The overall balance of costs and benefits of being eavesdropped upon will determine whether signallers promote eavesdropping on their signals or whether they try to avoid it. One intriguing possibility that Torben discusses is whether the costs of being eavesdropped upon could be avoided if signallers made their signals anonymous by removing information on signaller identity.

Audience effects In communication networks, several receivers are likely to be present during signalling interactions between others; these receivers do not take part in the interaction and have been referred to as an audience. The effects they can have on signalling behaviour are the subject of Ch. 4, in which Ricardo Matos and Ingo Schlupp draw the distinction between an apparent audience and an evolutionary audience. The distinction is important because selection pressures imposed by the presence of audiences in the evolutionary past of the animals may result in features of the signalling interactions despite the absence of an audience during any particular interaction. Also, whether an audience is apparent to the signallers involved in interactions may depend on signal modality: individuals have to be in the line of sight of visual signals to receive them and, therefore, an audience is likely to be apparent; however, the same is not true of widely broadcast acoustic signals.

Bystanders Being a bystander (i.e. present, but not directly involved) during an agonistic or mating interaction can affect subsequent aggressive and mating behaviour and is explored in Ch. 5. Ryan Earley and Lee Dugatkin focus on social eavesdropping (a subset of bystanding) by two species of poeciliid fishes that are likely

P1: JZZ/IKB P2: JZZ/... 0521840716agg.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

16:0

Behaviours specific to communication networks to be familiar to many – green swordtails Xiphophorus helleri and guppies Poecilia reticulata – and that communicate largely with visual signals. Their chapter shows how a network view can encompass and organize diverse aspects of fighting and mating behaviour (including mate copying); it also identifies the many conditions that favour eavesdroppers and how the effects of eavesdropping are manifested.

Victory displays In the final chapter of this section, John Bower examines victory displays: signals produced by the winner (but not the loser) after an aggressive interaction. There has been surprisingly little work specifically on this topic, despite the wealth of studies of signalling before and during aggressive displays, and such information is widely scattered. Chapter 6 collates the information on victory displays and then interprets its functional significance, first from a dyadic perspective and then from a network perspective. It may have been premature to include victory displays in this section, because on current evidence it is not clear that victory displays always function in a network context rather than in the winner–loser dyad. However, even if their main function is dyadic, their conspicuous nature makes it likely that other individuals could gain useful information by paying attention to victory displays.

Future directions The authors dealing with eavesdropping make several suggestions for the directions future research should take: incorporating eavesdropping into theoretical models to derive testable predictions that can contribute to understanding signal evolution (Ch. 2); finding evidence of eavesdropping in non-experimental natural contexts (perhaps by using a combination of tracking and acoustic location technologies to follow the individuals in a network), and continuing such studies long enough to identify differences in reproductive success (Ch. 3); unravelling the complex interrelationships between features of individuals, their social and wider environment and the role of bystanders in order to understand communication fully (Ch. 5). The authors dealing with eavesdropping clearly consider that the phenomenon is now well characterized. In contrast, victory displays clearly need more detailed study in order to establish the phenomenon and to elucidate its function and whether it is network phenomenon. It is likely that controlled laboratory experiments are the best way to investigate what effect, if any, victory displays have on other members of the communication network (Ch. 6). Progress in understanding audience effects seems likely to come from a different type of approach. In addition to modelling and controlled experiments, there is

11

P1: JZZ/IKB P2: JZZ/... 0521840716agg.xml CU1917B/McGregor

12

0 521 582361 7

April 10, 2005

16:0

Part I the potential to integrate information on audience effects with underlying mechanisms. Suitable candidate mechanisms exist in the literature (e.g. the hormonal basis of priming effects) and deserve to be investigated more fully (Ch. 4).

References Greenfield, M. D. 2002. Signalers and Receivers: Mechanisms and Evolution of Arthropod Communication. Oxford: Oxford University Press.

P1: JZZ/... P2: JZZ/... 0521823617c02.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

16:16

2

Eavesdropping in communication networks tom m. peake University of Copenhagen, Denmark

Introduction All communication occurs in a network environment with the exception of a subset of systems that unequivocally meet both of the following criteria: (a) a signal can never be received by more than one receiver; (b) a receiver can never receive more than one signal simultaneously. In other words, all communication networks have at least one of two defining properties: (a) signals can be, at least potentially, received by several receivers; and (b) receivers can, at least potentially, receive signals from several signallers at any one time. Consequently, in moving from a dyadic consideration of communication to a network view, signallers and receivers both take on a range of costs and benefits, which are the theme of this book. In this chapter, I will consider the implications of a particular type of receiving behaviour that becomes possible in a network, namely eavesdropping. I will begin by reviewing different definitions of eavesdropping that are found in the literature and the evidence for different types of eavesdropping, distinguishing between eavesdropping on signals and eavesdropping on signal interactions. I will then examine the costs, benefits and implications of eavesdropping on interactions, as recognition of this phenomenon emerged from considerations of qualitative differences between dyadic and network views of communication (McGregor, 1993; McGregor & Dabelsteen, 1996).

Defining eavesdropping The verb eavesdrop is defined by the New Oxford Shorter English Dictionary as ‘Listen secretly to (a person, private conversation), orig, by standing beneath the Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

13

P1: JZZ/... P2: JZZ/... 0521823617c02.xml CU1917B/McGregor

14

0 521 582361 7

April 10, 2005

16:16

T. M. Peake eaves of a house. Formerly also, stand beneath the eaves of (a building) in order to overhear conversation within.’ This word, at least to native English speakers, has an evocative quality that makes it appealing to authors in a variety of often quite different contexts. The use of the word in everyday language also has connotations that it may be useful to discard at this stage. First, in everyday use the term applies only to the acoustic modality; as a technical term in animal communication there is no good reason why this should be, although undoubtedly this has contributed to the term not being used by authors working in some modalities. Second, the idea of secrecy contained in the above definition need not necessarily be carried over to its use in animal communication. In the context of animal communication, the term has been used in a number of different ways that can be summed up by a general definition: the use of information in signals by individuals other than the primary target. This definition excludes the use of eavesdropping to describe behaviours such as detecting prey by cues that are not designed to enhance information transfer (e.g. extraneous noise caused by movement); in this sense the definition differs from that given by Bradbury & Vehrencamp (1998, p. 3). Eavesdroppers have been called ‘illegitimate’ (Otte, 1974), ‘unintended’ (Wiley, 1994) or ‘third party’ (Zahavi, 1979) receivers or ‘bystanders’ (e.g. Dugatkin, 2001) according to the context in which they were defined. In the general definition above, I use the phrase ‘individuals other than the primary target’ on the grounds that, as I shall outline below, eavesdropping individuals do not under all circumstances impose a cost on signallers as is implied by some of the alternative terms above. In situations where the presence of eavesdroppers benefits signallers, there may be selection pressure to allow information transmission to eavesdroppers, while the major selective force remains the more apparent (or primary) receiver. Within this general definition, there are two classes of use of the term eavesdropping that are sufficiently different, yet sufficiently commonly used, to warrant discussion and clarification. Here I call these classes interceptive and social eavesdropping. Interceptive eavesdroppers benefit by intercepting signals intended (in an evolutionary sense) for another individual, usually to the cost of the signaller. Social eavesdroppers gather information on other individuals by attending to their signalling interactions with conspecifics. At first glance, these two types of behaviour may seem very similar; however, as I shall argue in the remainder of the chapter, the nature of information transfer and resulting selection pressures differ markedly between them. These terms were chosen carefully to indicate the source from which eavesdroppers gather information (i.e. intercepting signals versus attending to social interactions), without undue prejudice towards certain aspects more commonly (but rarely exclusively) associated with either behaviour. The aim of this distinction is to move arguments or points of confusion away from the question of whether a given behaviour can be called

P1: JZZ/... P2: JZZ/... 0521823617c02.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

16:16

Eavesdropping in communication networks Table 2.1. Generalizations concerning the two types of eavesdropping behaviour defined in the text Type of eavesdropping

Interceptive

Social

Source of information

Signals

Signal interactions

Type of signal

Usually broadcast

Always directed

Eavesdropper–signaller

Usually heterospecific

Usually conspecific

relationship Payoff to signaller

Usually negative or zero

Positive, negative or zero

Information gathered

Absolute

Relative information also available

eavesdropping or not and towards a more fruitful approach based on the nature of information transfer. While these definitions of eavesdropping require that signals be transmitted to more than one receiver (network property (a) above), social eavesdropping further requires that receivers can detect more than one signal at the same time (network property (b)). One fact is unavoidable: eavesdropping is, by definition, a behaviour that can only occur in a network as it requires at least three individuals: a signaller, a target receiver and an eavesdropper. A number of generalizations may be made that show the distinctions between the two types of eavesdropping (summarized in Table 2.1). (a) Interceptive eavesdropping usually involves the reception of broadcast signals (i.e. those that have a class of targets such as females of the signaller’s species rather than a specific target); social eavesdropping, by definition, involves the exchange of signals directed towards specific receiving individuals. (b) Interceptive eavesdropping is most commonly identified in situations where eavesdroppers are a different species from the signaller; social eavesdropping is usually identified within a species. (c) Interceptive eavesdropping usually has a negative or zero effect on signallers; the payoff to a signaller resulting from social eavesdropping is much less clear, as will be discussed below. (d) Interceptive eavesdropping focuses on the absolute signalling behaviour of the signaller (in many cases this may be simple presence/absence of information); social eavesdropping additionally allows information to be gathered on the relative performance of interacting signallers (allowing both direct comparison of interactants and assessment of relationships between them).

Interceptive eavesdropping Wiley (1983) defined eavesdropping as the behaviour by which ‘signals intended for one receiver are intercepted by another’; this definition is explicitly given as an example of a receiver ‘obtaining information about the signaller

15

P1: JZZ/... P2: JZZ/... 0521823617c02.xml CU1917B/McGregor

16

0 521 582361 7

April 10, 2005

16:16

T. M. Peake against its own best interests’. Bradbury & Vehrencamp (1998) adapted this definition to include as eavesdropping situations where the signaller obtains a zero benefit, terming as ‘exploitation’ cases in which eavesdroppers are detrimental to signallers. In defining interceptive eavesdropping above, I make no assumptions about the nature of the signaller payoff. The definition is intended to be simply descriptive and one could imagine many subdivisions that could be made. One clear distinction is in the taxonomic relationship between signaller and eavesdropper. Where signaller and eavesdropper are of different species, as is the case in most examples of signal interception found in the literature, the payoff to the signaller is almost certainly negative. Information obtained by eavesdropping in this case may be something as simple as the location of a suitable prey item. The effect of eavesdropping within a species is likely to be more difficult to determine and the kinds of information gathered may well be more related to features of the signaller. As a final point, it is suggested by some authors (e.g. Bradbury & Vehrencamp, 1998; Greenfield, 2002) that signals intercepted by another species, particularly in such cases as predators locating prey, should be considered cues as they are not designed to enhance information transfer to those receivers. This is certainly true if one considers predator and prey in isolation; however, when considering the structure of the source of information and factors associated with production it is important that the wider context is included. Interspecific examples

The most commonly cited examples of eavesdropping, and those that have been best studied, are those that occur between trophic levels, i.e. predators and parasites detecting the signals of prey or hosts or prey detecting predator presence by their signals. Selection pressures imposed by these kinds of eavesdropper are understandably high and have been shown to lead to a range of counter-adaptations that aim to ameliorate or avoid such pressures (e.g. Greenfield, 1994, 2002; Heller, 1995; Stoddard, 1999; Gerhardt & Huber, 2002; Ch. 8). Examples are particularly prevalent in acoustic (e.g. Cade, 1975; Ryan et al., 1982; Sakaluk & Belwood, 1984; Belwood & Morris, 1987) and chemical (Aldrich, 1995; Roberts et al., 2001) signalling and there are good reviews available (e.g. Stowe et al., 1995; Zuk & Kolluru, 1998). While many visual signals are conspicuous and may be used by predators to find prey (e.g. Lloyd & Wing, 1983), it is rare to find such examples called eavesdropping (see Bruce et al. (2001) for such an example). Such interceptive eavesdropping reflects communication networks working on a community level. The selection pressures on communication between trophic levels are widely acknowledged in most considerations of the evolution of communication. Much less widely studied and appreciated is the importance of networks

P1: JZZ/... P2: JZZ/... 0521823617c02.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

16:16

Eavesdropping in communication networks operating within a species, and there is evidence that interceptive eavesdropping occurs at this level. Intraspecific examples

Within a species, individuals may eavesdrop on the signalling behaviour of others for a variety of reasons. In some cases, animals of one sex eavesdrop on signals intended for the opposite sex (Ch. 12). For example, Kiflawi & Gray (2000) looked at eavesdropping by male crickets Acheta domesticus on competing males’ mating calls. Smaller males showed a phonotactic response towards speakers broadcasting calls preferred by females in a two-speaker choice design, while larger males varied in their phonotaxis. The suggestion here is that males with unattractive calls can potentially intercept females as they move towards attractive males. A recent example concerns the use of female signals, apparently intended for mates, as a means of detecting fertile females for extra-pair copulations. Female robins Erithacus rubecula produce ‘seep’ calls to obtain provisioning from their mate (East, 1981). Mate removal experiments show that females may attract other males, which provide courtship feeding that may result in copulation (Tobias & Seddon, 2000). Tobias & Seddon (2002) found that neighbouring males approached ‘seep’ calls when they were played back at a high rate near the territory boundary, on occasion bearing provisions. They suggest that, if the female call is a hunger signal directed towards the mate, neighbouring males might be eavesdropping (in the interceptive sense). In this case, females may derive a benefit from the presence of eavesdroppers while the primary benefit comes from the response of the mate. Tobias & Seddon (2002) also acknowledged the possibility that the ‘seep’ call is directed towards extra-pair males as a means of ‘blackmailing’the mate into providing food. In either case, the results highlight the influence of operating in a social network. Individuals may eavesdrop on signals designed to warn others of the presence of predators. For example, Shennan et al. (1994) describe the behaviour of group members paying attention to the vigilance activities of others in order to avoid predators. Convict cichlids Cichlasoma nigrofasciatum fin-flick in order to warn young; fish not guarding young do not fin-flick. Parents were shown to fin-flick in response to fin-flicking models, suggesting that they are capable of monitoring the vigilance activities of others in order to warn their own young sooner. Here, the primary targets of the signal are likely to be relatives of the signaller, while non-relatives may also benefit by paying attention to the signals at no obvious cost to the signaller. This early warning feature of the signalling behaviour of others has been suggested as a possible function of territoriality (e.g. Eason & Stamps, 1993). Observations of red-capped cardinals Paroaria gularis showed that territorial males had

17

P1: JZZ/... P2: JZZ/... 0521823617c02.xml CU1917B/McGregor

18

0 521 582361 7

April 10, 2005

16:16

T. M. Peake a high chance of detecting intrusions where the intruder had been recently repelled from a neighbour’s territory as a consequence of the conspicuousness of behaviours involved in eviction. Intruders that had not been detected by neighbours were unlikely to be detected by territorial subjects. In all of these examples, individuals use signals produced by conspecifics for their own benefit, as fits the definition of interceptive eavesdropping. Less clear is the effect of such eavesdropping on the signaller. In some cases (e.g. Kiflawi & Gray, 2000), the signaller may suffer a cost because of the presence of eavesdroppers, while in some there is no obvious benefit or cost to the signaller (e.g. Shennan et al., 1994). In some cases (e.g. Tobias & Seddon, 2002), the presence of eavesdroppers may actually benefit the signaller. These examples show how definitions of eavesdropping and communication based on costs to the signaller (e.g. Wiley, 1983; Bradbury & Vehrencamp, 1998) may not apply to all circumstances. It is for this reason that I prefer the descriptive definition of eavesdropping in general as the use of signals by receivers other than the primary target (see above). Autocommunication and eavesdropping

Eavesdropping has also been used to describe the interception of information contained in sounds produced by animals in order to investigate their environment. Although not strictly within the general definition of eavesdropping given above, because these sounds are not designed to transmit information to others, the examples are interesting enough to be worth mentioning. Little brown bats Myotis lucifugus gather information by paying attention to the echolocation calls of foraging conspecifics (Barclay, 1982). Bats approached speakers broadcasting echolocation calls of conspecifics and a heterospecific Eptesicus fuscus, suggesting that eavesdropping could substantially increase potential prey detection distance. Balcombe & Fenton (1988) found similar results in M. lucifugus, a congener Myotis yumanensis and another species (Lasiurus borealis), in which individuals apparently used others’ echolocation calls to identify ‘vulnerable’ prey in order to steal them from the eavesdropped bat. Similarly, Xitco & Roitblat (1996) use the term eavesdropping to describe the ability of a bottlenose dolphin Tursiops truncatus to identify objects inspected by another dolphin, via the reception of the inspector’s echolocation clicks.

Social eavesdropping In one of the first explicit considerations of territorial systems as a communication network, McGregor (1993) suggested that the term eavesdropping could be applied within species, particularly in the context of paying attention to interactions between neighbours and rivals. While initially McGregor (1993) considered

P1: JZZ/... P2: JZZ/... 0521823617c02.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

16:16

Eavesdropping in communication networks this analogous to interceptive eavesdropping, McGregor & Dabelsteen (1996) later made the distinction, refining their definition of eavesdropping to ‘extracting information from an [signalling] interaction between other individuals’. They considered it ‘a prerequisite of eavesdropping that a third party (the eavesdropper) gains information from an interaction that could not be gained from a signal alone’. This, they suggested, was a different level of information transfer than simply locating a prey item by its signals. McGregor & Peake (2000) attempted to clarify the hierarchical relationship between information, signals and signalling interactions, concluding that interactions may be under additional selection pressures to those acting on the signals themselves. The most obvious source of these selection pressures is those with an interest in the outcome: rivals and mates. Experimental evidence for social eavesdropping has recently increased dramatically as clear experimental paradigms have emerged. Studies that explicitly address eavesdropping in this context have thus far been carried out exclusively on acoustic interactions in territorial songbirds and visual interactions in teleost fish. However, evidence from other experiments not designed to test eavesdropping per se are strongly supportive of the existence of eavesdropping as a means of gaining information on the qualities of and/or social relationships between conspecifics. Acoustic interactions in songbirds

McGregor et al. (1997) addressed the issue in songbirds using interactive playback (Dabelsteen et al., 1996) to simulate intrusion upon the neighbours of subject male great tits Parus major (Fig. 2.1ai). Neighbours were presented with one of two types of intruder. One type indicated its willingness to escalate by beginning each song immediately following the onset of neighbour song (overlapping: Hulsch & Todt, 1982; Dabelsteen et al., 1996, 1997) and increasing song length. The other type of intruder playback signalled a lower level of willingness to escalate by beginning songs only after the neighbour songs had been completed (alternating) and reducing song length. After a short amount of time, an intrusion by the same intruder was simulated in the subject male’s territory (Fig. 2.1aii) singing an alternating pattern with matched song length. Subjects responded to previously aggressive intruders by keeping their distance and overlapping song, while less-aggressive intruders were approached quickly. A similar experiment carried out on great tits looked at the behaviour of females in response to intruders interacting with their mates and neighbours (Otter et al., 1999). In this case, experiments were carried out on dyads of neighbouring territories each defended by a mated pair. Playback was used to intrude on each territory on successive days (Fig. 2.1bi, bii) such that the same intruder showed a high willingness to escalate to one male and a low willingness to escalate to the

19

P1: JZZ/... P2: JZZ/... 0521823617c02.xml CU1917B/McGregor

20

0 521 582361 7

April 10, 2005

16:16

T. M. Peake

(a)

(b) (i)

(i)

(ii)

(ii)

(c)

(d)

(e) (i)

(iii)

(ii)

(i)

(ii)

(i)

(ii)

(iii)

KEY Loudspeaker

Interaction between two loudspeakers

Loudspeaker broadcasting playback

Interaction between male and loudspeaker

Fig. 2.1. Schematic representations of experiments described in the text investigating social eavesdropping on acoustic interactions in songbirds. (a) Representation of design used by McGregor et al. (1997) showing (i) interaction between loudspeaker and neighbouring male and (ii) subsequent playback intrusion in subject’s territory. (b) Design used by Otter et al. (1999) and Mennill et al. (2002) showing (i) interaction between one male and a loudspeaker, (ii) subsequent interaction between neighbouring male and loudspeaker and (iii) observation of female behaviour. (c) Design used by Naguib & Todt (1997) and Naguib et al. (1999) showing (i) interaction between two loudspeakers inside the subject’s territory and (ii) subsequent playback from loudspeaker not approached initially by the subject. (d) Design used by

P1: JZZ/... P2: JZZ/... 0521823617c02.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

16:16

Eavesdropping in communication networks neighbouring male. Otter et al. (1999) then followed the females for some time subsequent to the treatments (Fig. 2.1biii) and showed that females whose mates had suffered from an intruder that a neighbour dealt with easily were more likely to trespass onto neighbouring territories, particularly that of the neighbour who had performed well during playback. These female forays were not converted into offspring, however (Otter et al., 2001), suggesting that the short-term nature of the information was not enough to convince females of the poor quality of their mates. Mennill et al. (2002), however, did find such an effect in female black-capped chickadees Poecile atricapillus using a similar experimental paradigm. In this case, Mennill et al. (2002) had information on males’ dominance ranks during winter feeding flocks so that dyads of neighbouring territories each consisted of one mated pair in which the male was high ranking and the other in which the male was low ranking. Playback was carried out to these dyads in a similar way to that used by Otter et al. (1999) and was followed by microsatellite paternity analysis in order to assess female reproductive decisions. The results showed that highranking males that had lost to playback showed a much greater incidence of lost paternity with extra-pair young in 12 of 23 nests, compared with 2 of 20 in control nests. Low-ranking males that did well against intruders lost paternity to the same extent as controls. In all of the above experiments, the conclusion is that the responses of subjects are a result of information gained by paying attention to the interactions between intruders and known males. This interpretation is, however, somewhat limited by the fact that, in each case, the response of the subject may be affected by the response of the known male: because the interaction of interest was between a male simulated by playback and a live male, the subsequent response of the subject may be affected by changes in behaviour of the live male. For example, in McGregor et al. (1997), the response of the subject may result from changed behaviour of the neighbour following different levels of intrusion; similarly, in the latter two studies, females may have changed their behaviour in response to changed behaviour of their mates. Mennill et al. (2002) did, in fact, examine the

Fig. 2.1 (cont.) Peake et al. (2004) showing (i) interaction between two loudspeakers outside the territory boundary and (ii) subsequent intrusion by one of the loudspeakers. (e) Design used by Peake et al. (2002) showing (i) intrusion by loudspeaker, (ii) subsequent interaction between that and another loudspeaker outside the territory boundary and (iii) intrusion by the second loudspeaker. Rounded rectangles represent territory boundaries; male and female symbols represent approximate positions of resident males and females; arrows represent movements by loudspeakers; arrows with curved lines represent monitoring of female movements. See text and cited references for more details of each experiment.

21

P1: JZZ/... P2: JZZ/... 0521823617c02.xml CU1917B/McGregor

22

0 521 582361 7

April 10, 2005

16:16

T. M. Peake behaviour of males following each playback treatment and could find no effect on subsequent behaviour. The problem of lack of control over the signalling behaviour of interactants and subsequent changes in behaviour can be avoided in songbirds by replacing males with playback; dyadic encounters can then be simulated using two loudspeakers. The first study to use this approach was carried out on nightingales Luscinia megarhynchos by Naguib & Todt (1997), who examined the effect of asymmetric interactions on the responses of territorial males. The asymmetry in this case was achieved by having one loudspeaker producing songs that overlapped the other. Loudspeakers were placed inside the territory boundary of the subject and interactions lasted for two minutes (Fig. 2.1ci). Males responded by spending more time near, spending more time singing near, and singing more songs near the overlapping speaker. Ten minutes after the ‘interaction’ had finished, playback for one minute was broadcast from the speaker that was not approached first during the interaction (Fig. 2.1cii). Males sang more at the location of the formerly overlapping speaker regardless of whether that speaker was producing song. Naguib et al. (1999) repeated this experiment with a different kind of interaction in which songs did not overlap but were still asymmetrical as one speaker (the follower) always directly followed the output of another (the leader). In this case, males showed a stronger response to the speaker that ‘led’; once again the subjects responded differently to the two types of apparent opponent (Naguib & Todt, 1997). In these two experiments, the design meant that subjects could associate roles during an interaction with the location of a singing intruder. Peake et al. (2001) looked at whether similar associations could be made between roles and song features using a similar experimental paradigm in great tits. In this case interactions were carried out between two loudspeakers situated outside the territory boundary and thus in an area that subjects would be less willing to approach directly (Fig. 2.1di). Information extracted by subjects was then assayed by means of a third speaker placed well inside the subject’s territory, which broadcast songs of one of the interactants 15 minutes after the interaction (Fig. 2.1dii). Three types of interaction were used, again based on song timing between the speakers: overlapping, alternating and random. Intruders were then one of four types: overlappers, alternators, random (i.e. no consistent role) or males that had been overlapped by the opponent. In response to the assay intrusion, subjects responded with equally high song output to overlappers, alternators and random interactants but showed a twofold reduction in song towards males that had been overlapped. In all three of these experiments (Naguib & Todt, 1997; Naguib et al., 1999; Peake et al., 2001), the only difference between intruders was in relative song timing during the interaction, i.e. there was no absolute information available

P1: JZZ/... P2: JZZ/... 0521823617c02.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

16:16

Eavesdropping in communication networks upon which subjects could base their responses. The results then clearly showed that subjects had eavesdropped on the interaction as a whole and associated the roles of interactants with either the location of the singer (Naguib & Todt 1997; Naguib et al., 1999) or features of his song (Peake et al., 2001). These three experiments also share the feature that, in each case, the subjects had no prior experience of either interactant; therefore, decisions must have been made purely on the basis of the interaction. In reality, territorial songbirds are likely to have knowledge of the relative strengths of neighbouring males as a result of direct interactions during territory establishment and maintenance and indirectly from hearing them interact with others. Therefore, individuals may be able to use these known individuals as ‘yardsticks’against which to measure previously unencountered individuals. The first three studies mentioned in this section attempted to address this issue by looking at eavesdropping on encounters between intruders and neighbouring males (McGregor et al., 1997) and/or mates (Otter et al., 1999; Mennill et al., 2002). In these studies, however, there was little control over eavesdroppers’ prior experience with these yardsticks and the possibility that they may have themselves contributed to the responses shown (see above). Peake et al. (2002) attempted to address these problems by carefully controlling prior experience with an individual. This experiment was similar to the twospeaker experiment mentioned above (Peake et al., 2001). The difference was that one of the interactants (A) was introduced to the subject prior to the interaction by means of a territorial intrusion simulated by interactive playback (Fig. 2.1ei). The initial intruder either played an aggressive role, overlapping the subject’s song, or a much less-aggressive role, beginning a song one second after the subject had finished each song, allowing the subject to overlap playback. Following this intrusion, an interaction was simulated outside the territory between the recent intruder and a male (B) unknown to the subject (Fig. 2.1eii); here either A or B played the aggressive role by overlapping the song of the other. By combining the outcomes of the two interactions, four treatment types were carried out that provided information on the status of B relative to the subject. In two cases the information available did not clearly show the status of B relative to the subject: either A was aggressive to both B and the subject or A received aggression from both. In the other two cases, the information available was clear: A showed low aggression to the subject but high aggression to B, indicating B to be of low status, or A was highly aggressive to the subject but received aggression from B, indicating that B was of high status relative to the subject. In response to subsequent intrusion by B (Fig. 2.1eiii), males showed a threefold reduction in song towards males that were of low relative status (as indicated by the information available from the treatment type), compared with the response to high-status males or males about which information did not reliably determine status. This result shows that males

23

P1: JZZ/... P2: JZZ/... 0521823617c02.xml CU1917B/McGregor

24

0 521 582361 7

April 10, 2005

16:16

T. M. Peake combined information from the two interactions, one they took part in and one they heard, in deciding how to respond to subsequent intrusion. In all of these experiments, great care was taken to ensure that the information available to eavesdroppers was purely relative. This is important in order to demonstrate social eavesdropping, i.e. that individuals pay attention to the interaction rather than simply the absolute outputs of either male. However, during real interactions it is likely that both absolute and relative information is available to, and indeed used by, social eavesdroppers. A recent experiment on great tits (Peake et al., 2004) attempted to address this issue. Male great tits have a repertoire of one to six song types, many of which are shared by neighbouring individuals and used during song interactions (matched counter-singing; Krebs et al., 1981; Falls et al., 1982). Peake et al. (2004) used the two-loudspeaker design of Peake et al. (2001; Fig. 2.1d) to simulate interactions in which interactants differed in their use of song types. In each interaction, one speaker (A) produced the same song type throughout the interaction. The other speaker (B) began producing a different song type from A and then switched song types halfway through. On half of the occasions, B switched to the same song type as A (matching); on the other half of occasions B switched to a song type that was different from A. Thus, there were four possible intruders during the assay intrusion: males that switched to match (matchers), males whose opponent switched to match (matched), males who switched but did not match their opponent (switchers) and males whose opponents switched but did not match (switched). In both types of interaction there is a clear absolute difference between males in signalling behaviour, i.e. singing one song type versus two song types. Between the two types of interaction there is also relative information, available only in the interaction as a whole, in whether the switching individual matched his opponent. Subjects responded to simulated intruders by singing much shorter songs to those individuals that used two song types compared with those singing one song type. In addition, subjects did not approach or spend time near switched intruders, compared with no difference in approach response to the other types of intruder. Therefore, it seems that in this case the response of males to simulated intruders used both relative information in the interaction (switching and matching) and absolute information in the signalling behaviour of the individual interactants (one or two song types). Visual interactions in fish

Eavesdropping on visual displays given by male Siamese fighting fish Betta splendens during male–male interactions has been shown by both males (Oliveira et al., 1998) and females (Doutrelant & McGregor, 2000). In these experiments, subjects (who could see other males without themselves being seen) were allowed to witness interactions between males displaying across a transparent barrier

P1: JZZ/... P2: JZZ/... 0521823617c02.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

16:16

Eavesdropping in communication networks (Fig. 2.2ai). At the same time, two other males were taking part in a similar interaction that could not be seen by the subject, allowing a control for changes in the male opponents providing information on the outcome rather than (or as well as) information from the dynamics of the interaction. Male subjects (Oliveira et al., 1998) were then introduced to each of the four interactants (two seen and two unseen) in turn (Fig. 2.2aii) and the response measured. Males responded to individuals that they had seen lose by approaching and displaying sooner than with males that they had seen win. No such differences were seen in response to the winners and losers of displays that had not been witnessed. Features of the subjects’ behaviour during the interaction strongly suggested that the information used by subjects in responding was gathered by eavesdropping (see discussion in McGregor & Peake, 2000). In experiments with females (Doutrelant & McGregor, 2000), the seen and unseen interactions were temporally separated rather than concurrent (Fig. 2.2bi, bii). Following interactions, female subjects were allowed to move freely so as to exhibit a proximity preference for either male (Fig. 2.2biii). Females visited seen winners first, more often and spent more time near and displaying to seen winners than seen losers. Unseen losers were visited first more often than unseen winners, with no difference in the time spent near or displaying towards either male. Similar results have been found in green swordtails Xiphophorus helleri by Earley & Dugatkin (2002). Males allowed to view contests between other males without themselves being seen (Fig. 2.2ci) responded more cautiously to perceived winners, being less willing to initiate contests. Males that had witnessed contests were much less willing to escalate contests than males that had not seen contests. Males allowed to interact with contesting males during the contest (Fig. 2.2cii), and hence assess contestants more directly, were as likely to win contests as those that had not seen contests, suggesting that individual differences between the fish settled those contests. Males that had not witnessed contests (Fig. 2.2ciii) tended to escalate, whereas males that had interacted with contestants were unlikely to escalate, suggesting that these individual differences were assessed previously. These results also suggested that, where information on an opponent’s fighting ability was available from direct interaction, a presumably more reliable source, males tended to ignore the less-direct information gathered by eavesdropping. The difficulties of providing visual stimuli in the absence of live animals makes it much less straightforward to achieve the level of control over interactions afforded to songbird studies by acoustic playback. The use of models (e.g. Shennan et al., 1994) or video playback (Oliveira et al., 2000) potentially allows control over stimuli at a comparable level to acoustic playback, but neither has yet been used to simulate interactions. In the absence of such an approach, one way to delve deeper into the relationship between interactions and eavesdroppers is to decouple the

25

P1: JZZ/... P2: JZZ/... 0521823617c02.xml CU1917B/McGregor

26

0 521 582361 7

April 10, 2005

16:16

T. M. Peake

(a)

(i)

(b)

(i)

(ii)

(iii)

(c)

(i)

(ii)

(iii)

(d)

(i)

(ii)

(ii)

KEY Opaque partition Transparent partition One-way glass (arrow shows visible direction) Fig. 2.2. Schematic representations of experiments described in the text investigating social eavesdropping on visual interactions in fish. All fish were physically isolated by partitions; opaque, solid line; transparent, dotted line; one-way glass, dashed line with arrow showing direction in which visual contact was possible. (a) Design used by Oliveira et al. (1998) showing (i) two visual signalling interactions across transparent partitions, one witnessed by the central male, the other not; and (ii) subsequent presentation (indicated by arrow) of each male to the subject. (b) Design used by Doutrelant & McGregor (2000) showing (i) interaction between two males witnessed by a female, (ii) not witnessed and (iii) subsequent observation of female movements (indicated by arrow). (c) Treatments used by Earley & Dugatkin (2002) in which (i) interacting males were witnessed by a male that could not be seen by the

P1: JZZ/... P2: JZZ/... 0521823617c02.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

16:16

Eavesdropping in communication networks experience of interactants and eavesdroppers, i.e. both parties view the interaction differently. McGregor et al. (2001) attempted to do this using male Siamese fighting fish. Subject males were allowed to view two conspecific males apparently interacting across a small gap into which the eavesdropper could not see. The gap was in reality filled by a small aquarium that either was empty (allowing males to interact across the gap: the ‘real’ interaction; Fig. 2.2di) or contained two fish separated by an opaque partition such that each of the males viewed by the eavesdropper was, in fact, interacting with a fish that could not be seen (the ‘apparent’ interaction; Fig. 2.2dii). In the apparent interaction, the eavesdropper’s interpretation of the aggressive signals given by each visible male was decoupled from that male’s actual experience in his own interaction with a hidden fish. The results showed that eavesdroppers responded more aggressively to individuals that had displayed more during apparent interactions, with no such differentiation between males involved in real interactions. McGregor et al. (2001) suggested that the proximity of interactants to (hidden) opponents during the apparent interaction (compared with the relatively large distance between males (about one fish length) in the real interaction) resulted in an increase in aggressive displays (tail beating) and behaviour (attempted biting) in these interactions. Therefore, males either paid more attention to the information in these particularly aggressive encounters or viewed the winner of a highly aggressive encounter as a different level of threat to the winner of an encounter of lower general aggression. The evidence presented above demonstrates that fish pay attention to information available in interactions between conspecifics; however, the source of that information has not been conclusively shown to be the signals exchanged during interactions. While the results are consistent with the idea that fish eavesdrop, the difficulties in presenting subjects with fully controlled visual signalling interactions (cf. acoustic playback) means that using the term ‘social eavesdropping’ for this behaviour may be premature. Other evidence for social eavesdropping

There is, in addition to the examples given above, a variety of evidence supporting the importance of social eavesdropping as a means of information gathering. In some cases, there may be clear physiological effects; adult male

Fig. 2.2 (cont.) interactants, (ii) were witnessed by a male that could be seen (and interacted with) or (iii) were not seen. (d) Treatments used by McGregor et al. (2001) in which subjects witnessed two males (i) interacting across an empty divide into which the subject could not see (the real interaction) or (ii) apparently interacting but in fact interacting with males hidden from the subject (the apparent interaction). See text and cited references for more details of each experiment.

27

P1: JZZ/... P2: JZZ/... 0521823617c02.xml CU1917B/McGregor

28

0 521 582361 7

April 10, 2005

16:16

T. M. Peake cichlid fish Oreochromis mossambicus that witnessed fights between conspecifics showed elevated androgen levels (testosterone and 11-ketotestosterone) compared with controls (Oliveira et al., 2001). Female domestic fowl Gallus gallus that witness a known dominant being defeated by a stranger readily submit to that stranger in subsequent interactions (Hogue et al., 1996). When the stranger lost to a known dominant, the witnessing individual was able to dominate this stranger subsequently on 50% of occasions. A similar situation was found in juvenile rainbow trout Onchorhynchus mykiss, when subjects were allowed to interact with fish that had been seen to be dominant in a previous encounter or who had been dominant but had not been seen (Johnsson & ˚ Akerman, 1998). Individuals that lost to either dominant (seen or unseen) reduced aggression more rapidly to seen dominants. Individuals that won over these dominants increased aggression more rapidly to seen dominants. These results suggest that, while the final outcome may be a result of individual differences, information obtained before direct interactions occurred enabled individuals to make decisions about how to respond to these individuals more quickly. In these studies, the extent to which social eavesdropping occurs is difficult to assess, as it is not clear whether information extracted by observers is contained in signal interactions between participants or in other aggressive behaviours. Knowledge of the social rank relationships of others through observation of social interactions is also an important part of forming strategic alliances in some primate species (Seyfarth & Cheney, 2002; Ch. 25). The required amount of knowledge of this kind quickly becomes enormous as group sizes increase and has been suggested as one selection pressure driving large brain sizes in primates (Seyfarth & Cheney, 2002).

Identifying types of eavesdropping So far I have considered examples that clearly fall into one of the two classes of eavesdropping; interception of signals or of signal interactions. Situations could be imagined that are not so clearly placed in one category or the other. Some of these problems may be caused by the difficulties of defining signal interactions (Chs. 9 and 14). All the examples considered so far have dealt with interactions in which the signallers use the same modality. Many behaviour patterns that can be viewed as signalling interactions may occur in different modalities or switch between modalities as the interaction proceeds. If an acoustic signal given by a male is responded to by a visual signal from the female, then we can clearly say there has been an interaction involving signals, but does this constitute a signalling interaction? There is no good reason why signalling interactions, at least in terms of information transfer, have to occur in the same modality for each party.

P1: JZZ/... P2: JZZ/... 0521823617c02.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

16:16

Eavesdropping in communication networks However, there may be logistic difficulties in demonstrating social eavesdropping in such cases; one would have to show that both signals are involved in producing a response by the eavesdropper, i.e. that neither signal alone produced the same response. An example in which the level of information transfer is currently unclear concerns the use of courtship signals by neighbouring males to detect reproductive attempts in the whitethroat Sylvia communis (Balsby & Dabelsteen, 2003). In this case, the response of neighbouring males was recorded during experiments that simulated a territorial male interacting with an intruding male (via playback) or a receptive female (via a remotely controlled dummy and playback of female calls). The simulation of a receptive female resulted in greater song flight activity by neighbours than simulation of an intruding male, and intrusions by the neighbour (and subsequent evictions) were only seen during simulated courtship events. Song output by the experimental male could not explain the incidence of intrusion during courtship interactions by neighbours. As Balsby & Dabelsteen (2003) acknowledged, their experiment does not rule out the possibility that either signal alone provides sufficient information to explain the pattern of intrusion, thus interceptive eavesdropping may be an appropriate description of this behaviour. However, the detection of a courtship attempt in which the female is receptive is much facilitated by the presence of both male and female signals; therefore, it may be the interaction that is important, making this an example of social eavesdropping.

Information gathering and implications of eavesdropping So far I have discussed the different ways in which animals may be considered to be eavesdropping and some of the contexts in which this behaviour has been shown to occur. As yet I have considered the kinds of information that eavesdroppers may gain in only the broadest terms. As discussed by McGregor & Peake (2000), the information available in signalling interactions that is not available in signals alone is the important distinction between social and interceptive eavesdropping. Just as signals are a subset of the information available in an animal’s environment, signal interactions are a further subset of signalled information. The way in which animals use signals during interactions may in many cases be more revealing than the underlying content of the signals themselves, and in any case the use and content of signals need not necessarily be directly related. A number of features of signalling interactions may be particularly important in this respect; many of these features are particularly obvious in agonistic signalling encounters. Signal exchanges during agonistic encounters are generally thought to function so as to reduce the likelihood of direct physical aggression

29

P1: JZZ/... P2: JZZ/... 0521823617c02.xml CU1917B/McGregor

30

0 521 582361 7

April 10, 2005

16:16

T. M. Peake (and its ensuing costs) by allowing assessment of the likely outcome of a direct fight (e.g. Enquist, 1985). Reliability of signals used in this context may be high because of the threat of having one’s bluff called by one’s opponent. To this end, and given the immediacy of the potential punishment for cheating, signals given during these kinds of exchange may be particularly reliable. It is likely that the most accurate picture of an opponent’s fighting ability (short of actually fighting them) is gained by becoming involved in an aggressive signal exchange. The reliability of signals given in this context may also make social eavesdropping a good alternative in terms of obtaining accurate information while avoiding the risk of escalation. Similarly, while information on the underlying quality of an individual may be available in signals, immediate quality (e.g. condition, motivation) may be assessed more accurately when an opponent calls those factors directly into question. Thus signal interactions may provide reliable and up-to-date information on the current quality of participants. Second, interactions enable a direct comparison to be made between participants on a relative scale. Simply knowing the outcome of an interaction provides the information that A is stronger than B. By paying attention to an interaction it may be possible to extract information on relative quality, e.g. A is much stronger than B. If selection favours individuals that ‘just do enough’ to win an interaction, the available information will underestimate the relative difference in quality between the opponents. However, the nature of the interaction may provide such information. For example, one might expect interactions involving highly asymmetrical opponents to be shorter and less intense than those involving closely matched opponents. Of course, this assumes that the relative quality of opponents is the only influence on the information contained in the interaction; the presence of eavesdroppers may well affect the dynamics of interactions (Ch. 4). In most situations, the presence of eavesdroppers imposes selection pressures on signallers. In the case of interceptive eavesdropping, the selection pressures may be severe, especially in the case of predatory eavesdropping. In this case, the signaller must accept or avoid the costs of signalling; such avoidance mechanisms are particularly well understood in insects and anurans (e.g. Gerhardt & Huber, 2002; Ch. 8). In the case of selection pressures imposed by social eavesdroppers, the situation may be much less clear. In some situations, it is apparent that eavesdroppers have imposed strong selection pressures on signalling interactions by the nature of those interactions; for example, intense song duels in birds often involve switching to ‘quiet’ song when they approach physical aggression (Dabelsteen et al., 1998). In some cases, however, the opposite seems to be true: signalling interactions seem to be much more conspicuous than necessary to transmit information between participants (Zahavi, 1979). In these cases, it may be reasonable to assume that signals used here are, at least partly, ‘intended’ for eavesdroppers,

P1: JZZ/... P2: JZZ/... 0521823617c02.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

16:16

Eavesdropping in communication networks i.e. that communication is no longer restricted to the interaction. This view is supported by work on the effects of audiences on the dynamics of interactions (Ch. 4) and recent work on altruism (see below). In cases where asymmetries become clear to interactants, there may be different pressures acting on each party: for ‘winning’ individuals to advertise the fact and ‘losing’ individuals to hide it. Here eavesdroppers, as part of the selection regime in which communication systems have evolved, provide individuals with different payoffs depending on the current social context. In this case, one would expect a variety of adaptations in signalling behaviour to allow individuals dynamically to advertise or privatize information in interactions; many such adaptations have been suggested (e.g. Chs. 3 and 10). While considerations of social eavesdropping have focused on aggressive interactions, there are a number of similarities between the considerations of information gathering in this context and in the context of acts of apparent altruism (Johnstone, 2001). Suggestions that altruists may benefit by being perceived as such assume that observers are able to associate those acts with the individual performing them and subsequently use that information (Nowak & Sigmund, 1998). Studies of eavesdropping provide clear evidence that this level of association occurs, albeit in a different context. Many other aspects of cooperative and noncooperative behaviour occurring between individuals may similarly be explained by the passage of information outside the apparent dyad (Ch. 22).

Costs of eavesdropping Social eavesdropping has been suggested as a relatively cost-free means of gathering reliable information on rivals or potential mates. However, the costs have not been explicitly discussed in the literature aside from a sentence by McGregor & Dabelsteen (1996) that ‘listening at a distance only involves forgoing other behaviours such as feeding’. While gathering information in this way is undoubtedly less costly than becoming involved in aggressive, possibly physical, interactions, the costs of listening may be greater than commonly assumed. Evidence from studies of prey detection strongly suggests that animals have limits to the attention they can give to competing tasks (Dukas, 1998; Dukas & Kamil, 2001). Dukas (1998) suggested that the brain has finite capacity to process information such that animals can only process a limited amount of information at any one time. While studies have so far concentrated on visual attention, presumably a more demanding process than listening, it may be that similar reasoning applies to the acoustic sense. In this case, eavesdropping may limit the attention available for other important tasks such as predator vigilance, particularly if social eavesdropping is a more cognitively demanding task than simply attending

31

P1: JZZ/... P2: JZZ/... 0521823617c02.xml CU1917B/McGregor

32

0 521 582361 7

April 10, 2005

16:16

T. M. Peake to signals. A study on humans provided support for this view (Pendry, 1998). Human subjects were asked to form an impression of a target person based upon information provided to them one item at a time on a computer screen. Without prior warning and with no specific instruction, participants were simultaneously played a tape recording of a conversation that was either relevant to them or not (the contents of the conversations were identical; relevance differences were achieved by changing the object of the conversation). Subsequent tests showed that participants extracted more information than expected by chance from the relevant conversation but not from the irrelevant one. Participants hearing the relevant conversation were much more likely to obtain stereotypical impressions of the target and recalled many fewer items of information related to the target than those hearing a conversation of little relevance. These sorts of cost may still be comparatively low, particularly if eavesdropping, or otherwise monitoring the social environment, is relatively rare. However, social monitoring may represent a large proportion of some animals’ lives. A study of brown capuchin monkeys Cebus apella suggested that monitoring the social environment was the main function of vigilance behaviour in this species, and that the amount of time spent in such activity was highly correlated with the number of neighbouring individuals (Hirsch, 2002). Individuals spent, on average, 12.7% of their time vigilant, of which nearly 30% could be directly attributed to social monitoring and less than 10% to predator vigilance (see also Chs. 19 and 25).

Summary and future possibilities A defining property of communication networks is that more than one receiver may detect signals. In many cases, at least some of the receivers are not the primary target of the signal and in some the majority of potential receivers may fall into this category. Cases where individuals other than the primary target use information obtained in signals have been termed eavesdropping by a number of authors. In this chapter, I have distinguished between examples of eavesdropping that involve the interception of signals (interceptive eavesdropping) and a more recently suggested phenomenon of gathering information from signalling interactions between conspecifics (social eavesdropping). Interceptive eavesdropping can have significant effects on signal design and signalling behaviour when eavesdroppers impose large costs on signallers (e.g. where such eavesdropping occurs between trophic levels). The effects of interceptive eavesdroppers within species and social eavesdroppers in general are less well understood. The focus of studies on social eavesdropping has so far been territorial species, reflecting the biases of researchers active in this area. However, many different social situations would seem ideally suited to promoting eavesdropping as a social

P1: JZZ/... P2: JZZ/... 0521823617c02.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

16:16

Eavesdropping in communication networks behaviour. Group and/or colonial living species presumably have an even greater opportunity to extract information from interactions because of the close proximity of individuals to one another. Similarly, the focus of most studies of eavesdropping so far has been on male–male aggressive interactions. Interactions between males and females (Ch. 7), between group members (Chs. 10 and 25) and between parents and offspring (Ch. 9) are just some of the areas that potentially offer an important source of information to eavesdroppers and would benefit from further research. Equally as interesting would be the possibility of taking studies of eavesdropping beyond the trio of two interacting signallers and an eavesdropper and in so doing place the effect of eavesdroppers in a more extensive and natural network environment. The study of social eavesdropping is still in its infancy, yet the number of studies showing that such eavesdropping occurs has increased dramatically since the idea was mooted by McGregor & Dabelsteen in 1996. As the prevalence of studies demonstrating eavesdropping in communication networks increases, the importance of eavesdropping as a selective force on signalling and social structure will be better understood. We currently lack a clear theoretical framework within which to place the importance of social eavesdroppers in the evolution of signalling systems. As empirical studies continue to provide evidence that animals clearly have these capabilities, we eagerly await the emergence of models including eavesdroppers, such as those of Johnstone (2001) and Terry & Lachlan (Ch. 26); models that make clear predictions about where, how and when eavesdropping should occur.

Acknowledgements I would like to thank the following people for providing discussion and comments that greatly improved both this chapter and my thoughts on communication networks: Pete McGregor, Thorsten Balsby, Torben Dabelsteen, Giuliano Matessi, Ricardo Matos, Denise Pope and Andy Terry. While writing this chapter, I was funded by the Zoological Institute, University of Copenhagen.

References Aldrich, J. R. 1995. Chemical communication in true bugs and exploitation by parasitoids and commensals. In Chemical Ecology of Insects II, ed. R. T. Card´e & W. J. Bell. New York: Chapman & Hall, pp. 318–363. Balcombe, J. P. & Fenton, M. B. 1988. Eavesdropping by bats: the influence of echolocation call design and foraging strategy. Ethology, 79, 158–166.

33

P1: JZZ/... P2: JZZ/... 0521823617c02.xml CU1917B/McGregor

34

0 521 582361 7

April 10, 2005

16:16

T. M. Peake Balsby, T. J. S. & Dabelsteen, T. 2003. Male calling between courtship sequences in whitethroats: a way to counter intrusions from neighbouring rivals. Behavioural Processes, 63, 149–157. Barclay, R. M. R. 1982. Interindividual use of echolocation calls: eavesdropping by bats. Behavioral Ecology and Sociobiology, 10, 271–275. Belwood, J. J. & Morris, G. K. 1987. Bat predation and its influence on calling behavior in neotropical katydids. Science, 238, 64–67. Bradbury, J. W. & Vehrencamp, S. L. 1998. Principles of Animal Communication. Sunderland MA: Sinauer. Bruce, M. J., Herberstein, M. E. & Elgar, M. A. 2001. Signalling conflict between prey and predator attraction. Journal of Evolutionary Biology, 14, 786–794. Cade, W. 1975. Acoustically orienting parasitoids: fly phonotaxis to cricket song. Science, 190, 1312–1313. Dabelsteen, T., McGregor, P. K., Shepherd, M., Whittaker, X. & Pedersen, S. B. 1996. Is the signal value of overlapping singing different from that of alternating singing during matching in great tits? Journal of Avian Biology, 27, 189–194. Dabelsteen, T., McGregor, P. K., Holland, J., Tobias, J. A. & Pedersen, S. B. 1997. The signal function of overlapping singing in male robins (Erithacus rubecula). Animal Behaviour, 53, 249–256. Dabelsteen, T., McGregor, P. K., Lampe, H. M., Langmore, N. E. & Holland, J. 1998. Quiet song in songbirds: an overlooked phenomenon. Bioacoustics, 9, 89–106. Doutrelant, C. & McGregor, P. K. 2000. Eavesdropping and mate choice in female fighting fish. Behaviour, 137, 1655–1659. Dugatkin, L. A. 2001. Bystander effects and the structure of dominance hierarchies. Behavioral Ecology, 12, 348–352. Dukas, R. 1998. Constraints on information processing and their effects on behavior. In Cognitive Ecology, ed. R. Dukas. Chicago, IL: University of Chicago Press, pp. 89–127. Dukas, R. and Kamil, A. 2001. Limited attention: the constraint underlining search image. Behavioral Ecology, 12, 192–199. Earley, R. L. & Dugatkin, L. A. 2002. Eavesdropping on visual cues in green swordtail (Xiphophorus helleri) fights: a case for networking. Proceedings of the Royal Society of London, Series B, 269, 943–952. Eason, P. K. & Stamps, J. A. 1993. An early warning system for detecting intruders in a territorial animal. Animal Behaviour, 46, 1105–1109. East, M. 1981. Aspects of courtship and parental care of the European robin Erithacus rubecula. Ornis Scandinavica, 12, 230–239. Enquist, M. 1985. Communication during aggressive interactions with particular reference to variation in choice of behaviour. Animal Behaviour, 33, 1152–1611. Falls, J. B., Krebs, J. R. & McGregor, P. K. 1982. Song matching in the great tit (Parus major): the effect of similarity and familiarity. Animal Behaviour, 30, 977–1009. Gerhardt, H. C. & Huber, F. 2002. Acoustic Communication in Insects and Anurans. Chicago, IL: University of Chicago Press. Greenfield, M. D. 1994. Cooperation and conflict in the evolution of signal interactions. Annual Review of Ecology and Systematics, 25, 97–126

P1: JZZ/... P2: JZZ/... 0521823617c02.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

16:16

Eavesdropping in communication networks 2002. Signalers and Receivers. New York: Oxford University Press. Heller, K. G. 1995. Acoustic signaling in palaeotropical bushcrickets (Orthoptera: Tettigonioidea: Pseudophyllidae): does predation pressure by eavesdropping enemies differ in the Palaeo- and Neotropics? Journal of Zoology, 237, 469–485. Hirsch, B. T. 2002. Social monitoring and vigilance behavior in brown capuchin monkeys (Cebus apella). Behavioral Ecology and Sociobiology, 52, 458–464. Hogue, M. E., Beaugrand, J. P. & Lague, P. C. 1996. Coherent use of information by hens observing their former dominant defeating or being defeated by a stranger. Behavioral Processes, 38, 241–252. Hulsch, H. & Todt, D. 1982. Temporal performance roles during vocal interactions in nightingales (Luscinia megarhynchos B.). Behavioral Ecology and Sociobiology, 11, 253–260. ˚ Johnsson, J. I. & Akerman, A. 1998. Watch and learn: preview of the fighting ability of opponents alters contest behaviour in the rainbow trout. Animal Behaviour, 56, 771–776. Johnstone, R. A. 2001. Eavesdropping and animal conflict. Proceedings of the National Academy of Sciences, USA, 98, 9177–9180. Kiflawi, M. & Gray, D. A. 2000. Size-dependent response to conspecifics mating calls by male crickets. Proceedings of the Royal Society of London, Series B, 267, 2157–2161. Krebs, J. R., Ashcroft, R. & van Orsdol, K. 1981. Song matching in the great tit, Parus major L. Animal Behaviour, 29, 918–923. Lloyd, J. E. & Wing, S. R. 1983. Nocturnal aerial predation of fireflies by light-seeking fireflies. Science, 222, 634–635. McGregor, P. K. 1993. Signalling in territorial systems: a context for individual identification, ranging and eavesdropping. Philosophical Transactions of the Royal Society of London, Series B, 340, 237–244. McGregor, P. K. & Dabelsteen, T. 1996. Communication networks. In Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca, NY: Cornell University Press, pp. 409–425. McGregor, P. K. & Peake, T. M. 2000. Communication networks: social environments for receiving and signalling behaviour. Acta Ethologica, 2, 71–81. McGregor, P. K., Dabelsteen, T. & Holland, J. 1997. Eavesdropping in a territorial songbird communication network: preliminary results. Bioacoustics, 8, 253–254. McGregor, P. K., Peake, T. M. & Lampe, H. M. 2001. Fighting fish Betta splendens extract relative information from apparent interactions: what happens when what you see is not what you get. Animal Behaviour, 62, 1059–1065. Mennill, D. J., Ratcliffe, L. M. & Boag, P. T. 2002. Female eavesdropping on male song contests in songbirds. Science, 296, 873. Naguib, M. & Todt, D. 1997. Effects of dyadic interactions on other conspecifics receivers in nightingales. Animal Behaviour, 54, 1535–1543. Naguib, M., Fichtel, C. & Todt, D. 1999. Nightingales respond more strongly to vocal leaders of simulated dyadic interactions. Proceedings of the Royal Society of London, Series B, 266, 537–542. Nowak, M. A. & Sigmund, K. 1998. Evolution of indirect reciprocity by image scoring. Nature, 393, 573–577.

35

P1: JZZ/... P2: JZZ/... 0521823617c02.xml CU1917B/McGregor

36

0 521 582361 7

April 10, 2005

16:16

T. M. Peake Oliveira, R. F., McGregor, P. K. & Latruffe, C. 1998. Know thine enemy: fighting fish gather information from observing conspecifics interactions. Proceedings of the Royal Society of London, Series B, 265, 1045–1049. Oliveira, R. F., Rosenthal, G. G., Schlupp, I. et al. 2000. Considerations on the use of video playbacks as visual stimuli: the Lisbon workshop consensus. Acta Ethologica, 3, 61–65. Oliveira, R. F., Lopes, M., Carneiro, L. A. & Canario, A. V. M. 2001. Watching fights raises fish hormone levels. Nature, 409, 475. Otte, D. 1974. Effects and functions in the evolution of signaling systems. Annual Review of Ecology and Systematics, 5, 385–417. Otter, K. A., McGregor, P. K., Terry, A. M. R. et al. 1999. Do female great tits Parus major assess extra-pair males by eavesdropping? A field study using interactive song playback. Proceedings of the Royal Society of London, Series B, 266, 1305–1309. Otter, K., Stewart, I. K., Terry, A. M., McGregor, P. K. & Burke, T. 2001. Extra-pair paternity in great tits in relation to manipulation of male signals. Journal of Avian Biology, 32, 338–344. Peake, T. M., Terry, A. M. R., McGregor, P. K. & Dabelsteen, T. 2001. Male great tits eavesdrop on simulated male-to-male vocal interactions. Proceedings of the Royal Society of London, Series B, 268, 1183–1187. Peake, T. M., Terry, A. M. R., Dabelsteen, T. & McGregor, P. K. 2002. Do great tits assess rivals by combining direct experience with information gathered by eavesdropping? Proceedings of the Royal Society of London, Series B, 269, 1925–1929. Peake, T. M., Matessi, G., Dabelsteen, T. & McGregor, P. K. 2004. Song type matching, song type switching and eavesdropping in male great tits. Animal Behaviour, in press. Pendry, L. 1998. When the mind is otherwise engaged: resource depletion and social stereotyping. European Journal of Social Psychology, 28, 293–299. Roberts, S. C., Gosling, L. M., Thornton, E. A. & McClung, J. 2001. Scent-marking by male mice under the risk of predation. Behavioral Ecology, 12, 698–705. Ryan, M. J., Tuttle, M. D. & Rand, A. S. 1982. Bat predation and sexual advertisment in a neotropical anuran. American Naturalist, 119, 136–139. Sakaluk, S. K. & Belwood, J. J. 1984. Gecko phonotaxis to cricket calling song: a case of satellite predation. Animal Behaviour, 32, 659–662. Seyfarth, R. M. & Cheney, D. L. 2002. What are big brains for? Proceedings of the National Academy of Sciences, USA, 99, 4141–4142. Shennan, M. G. C., Waas, J. R. & Lavery, R. J. 1994. The warning signals of parental convict cichlids are socially facilitated. Animal Behaviour, 47, 974–976. Stoddard, P. K. 1999. Predation enhances complexity in the evolution of electric fish signals. Nature, 400, 254–256. Stowe, M. K., Turlings, T. C. J., Loughrin, J. H., Lewis, W. J. & Tumlinson, J. H. 1995. The chemistry of eavesdropping, alarm, and deceit. Proceedings of the National Academy of Sciences, USA, 92, 23–28. Tobias, J. A. & Seddon, N. 2000. Territoriality as a paternity guard in the European robin Erithacus rubecula. Animal Behaviour, 60, 165–173.

P1: JZZ/... P2: JZZ/... 0521823617c02.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

16:16

Eavesdropping in communication networks 2002. Female begging in European robins: do neighbours eavesdrop for extra-pair copulations? Behavioral Ecology, 5, 637–642. Wiley, R. H. 1983. The evolution of communication. In Animal Behaviour, Vol. 2: Communication, ed. P. J. B. Slater & T. R. Halliday. Oxford: Blackwell, pp. 156–189. 1994. Errors, exaggeration and deception in animal communication. In Behavioral Mechanisms in Evolutionary Ecology, ed. L. Real. Chicago, IL: University of Chicago Press, pp. 157–189. Xitco, M. J., Jr & Roitblat, H. L. 1996. Object recognition through eavesdropping: passive echolocation in bottlenose dolphins. Animal Learning and Behavior, 24, 355–365. Zahavi, A. 1979. Why shouting? American Nauralist, 113, 155–156. Zuk, M. & Kolluru, G. R. 1998. Exploitation of sexual signals by predators and parasitoids. Quarterly Review of Biology, 73, 415–438.

37

P1: GFZ/... P2: JZZ/... 0521823617c03.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

17:1

3

Public, private or anonymous? Facilitating and countering eavesdropping t o r b e n da b e l s t e e n University of Copenhagen, Denmark

Introduction Animals often live in environments where several conspecifics are within signalling range of each other simultaneously. This is obvious for lekking species, but it also applies to territorial species. In theory, this allows complicated patterns of information flow between the individuals, which can be considered members of a communication network (Dabelsteen, 1992; McGregor, 1993). A special case of a network is when dyads of signalling individuals are temporarily or, in the case of a sparse population, permanently isolated from other signallers or receivers. Within a network, an individual may gather information about another individual from simple reception of its signals, but it may also be in a position that allows simultaneous reception of the signals from two individuals engaged in a dyadic signalling interaction. Such a position provides a special option for gathering relative information about the two interacting individuals, for example about their state, strength or quality (McGregor & Dabelsteen, 1996). The relative information results from how the two individuals use their signals in the interaction; the information is relative in the sense that it expresses the relative state or ‘value’ of the two individuals without necessarily giving information about their absolute states or values. The extraction of information from a signalling interaction is, therefore, fundamentally different from simple receiving and so deserves its own term, eavesdropping (McGregor & Dabelsteen, 1996) or, more specifically, social eavesdropping (Ch. 2). In theory, relative information about different individuals could also be obtained by receiving the signals from each individual in turn,

Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

38

P1: GFZ/... P2: JZZ/... 0521823617c03.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

17:1

Facilitating and countering eavesdropping followed by a process in which the absolute information about different individuals is compared. Do animals eavesdrop?

There are observations of how natural signalling interactions between two animals may cause conspecifics to approach the interactants and/or signal in immediate or delayed response to the interaction or directly to interfere with the interaction (e.g. McGregor & Dabelsteen, 1996). Such observations, whether anecdotal (e.g. Snow, 1958) or the result of planned studies (e.g. Bower, 2000), suggest that animals eavesdrop. There is also experimental evidence that a conspecific may interfere with a dyadic interaction by signalling predominantly towards the apparent superior individual of the interaction, suggesting that eavesdropping took place before interference (Naguib & Todt, 1997; Naguib et al., 1999; Ch. 14). The ability to gather relative information by eavesdropping and utilize this in later dyadic encounters with one of the previously interacting individuals has been demonstrated experimentally in field studies with male birds (e.g. McGregor et al., 1997), female birds (e.g. Otter et al., 1999; Mennill et al., 2002), captive male fish (e.g. Oliveira et al., 1998; McGregor et al., 2001; Ch. 5) and captive female fish (e.g. Doutrelant & McGregor, 2000). Some of the most convincing evidence comes from field experiments with territorial male great tits, Parus major, that were allowed to eavesdrop on male–male interactions simulated by means of playback from two different loudspeakers. The experiments with simulated interactions, which allowed the best possible control over the relative information made available to the test subjects, demonstrated that males have the ability to extract relative information about rivals engaged in a song dual and utilize this in later encounters with them (Peake et al., 2001). They also demonstrated the ability to combine such information with the eavesdropper’s own previous direct experience with one of interacting males (Peake et al., 2002). Given the potential advantages to eavesdroppers of gaining such relative information (see below), this ability is likely to be used also in non-experimental natural contexts, but we still lack firm evidence that this really happens.

Potential gains and costs of eavesdropping Gaining at least relative information about the state, quality or strength of rivals or potential mates when absolute information is not available, or when the capacity to compare such information is lacking, must constitute the aim of any assessment process preceding decision making (e.g. in mate choice). Eavesdropping was, therefore, predicted to be a widespread phenomenon in both sexes and because it may constitute a low-cost and low-risk alternative to gathering

39

P1: GFZ/... P2: JZZ/... 0521823617c03.xml CU1917B/McGregor

40

0 521 582361 7

April 10, 2005

17:1

T. Dabelsteen the same relative information through direct interactions with the individuals eavesdropped upon (McGregor & Dabelsteen, 1996). For instance, a male that is eavesdropping on male–male interactions probably uses less energy and runs a lower risk of injury than it would through a direct interaction, which might escalate to actual fighting. Escalated interactions, whether hostile or collaborative (as in courtship), may also increase predation risk because vigilance is reduced (e.g. Dabelsteen & Pedersen, 1990; Jakobsson et al., 1995). As eavesdroppers must divide their attention between interactants, they may have to reduce their vigilance more than simple receivers, which can focus their attention on a single signaller (e.g. Dukas & Kamil, 2001). It is difficult to identify other potential costs that are specific to eavesdropping rather than costs that are common to any sort of information gathering. Overall, the advantages of eavesdropping seem obvious and predict the evolution of eavesdropping strategies that increase the possibility of gaining relative information about the participants of an interaction. Whereas the advantages of eavesdropping are clear, it is not necessarily advantageous to give away information to eavesdroppers. For instance, an individual that ends up losing a hostile interaction is unlikely to benefit from having its loss advertised, whereas a winner would have a clear interest in advertising its superiority (see also discussions in Chs. 2, 4 and 10). An individual should only start an agonistic interaction if it has no prior knowledge that it will lose to its opponent (e.g. Dabelsteen, 1985). Therefore, participants in an agonistic interaction should not attempt to withhold information at the start of the interaction, but do so later should the outcome become uncertain and the interaction escalate. When the outcome is uncertain, both of the interactants would have an interest in keeping the interaction private until the interaction has been settled. I know of no experiments in which subjects were presented with an interaction that had no difference in relative information, but I predict that a possible response of subjects to such playback would be territorial intrusions and extra-pair behaviour in relation to the interactants. Courtship interactions may also be sensitive to eavesdropping. During early stages of courtship, the two sexes may have conflicting interests. Males have an interest in preventing rivals from discovering, and perhaps interfering with, their courtship (Balsby & Dabelsteen, 2003a,b). Females may wish to attract more males (e.g. Wiley & Poston, 1996), for instance to provoke an interaction between males upon which they could eavesdrop. When the female has chosen a mate, both the female and the chosen male have an interest in preventing rivals from eavesdropping on the intensive courtship interaction that often precedes copulation, because information that copulation is imminent could lead to attempts to prevent or even interrupt copulation by rivals. Intrusion by neighbours and subsequent

P1: GFZ/... P2: JZZ/... 0521823617c03.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

17:1

Facilitating and countering eavesdropping interference with courtship (including interruption of copulation) have been observed in a number of different species, for example the robin Erithacus rubecula (Lack, 1940), the dunnock Prunella modularis (Davies, 1992) and the blackbird Turdus merula (Snow, 1958). As Snow (1958, p. 86) wrote: The sight of a pair copulating or about to copulate has an immediate and powerful effect on neighbouring males. In nearly every case that I observed, copulations were interfered with by the sudden arrival of one or two males, who either knocked the copulating male off the female or prevented him from mounting. And these attacks have been directed against a territory-holder in the middle of his own territory, where the neighbours normally never go or, if they do, only with every sign of nervousness. Snow (1958) also noticed that courtship of an impassive blackbird female did not result in such interference. Only when the female responded with copulation solicitation behaviour (i.e. when there was a real courtship interaction) did the rival males intrude. In the whitethroat Sylvia communis, the presence of a female in a territory leads to significantly more intrusions from neighbouring males than when no female is present, and territory owners always respond by chasing intruders out of their territories even when they have to interrupt their courtship of the female (Balsby & Dabelsteen, 2003a,b). A recent experiment suggested that it is the courtship interactions that make neighbours intrude. The experiment compared the intrusion rate of male subjects that could eavesdrop on either their neighbours interacting with a loudspeaker playing normal full whitethroat song (song duel treatment) or courtship interactions between their neighbours and Jumping Sylvia (Balsby & Dabelsteen, 2002), a remotely controlled stuffed female that could jump and vocalize (courtship treatment). Whereas the song duel treatment never elicited intrusions into the neighbour’s territory by the male subjects, the courtship treatment did so in 56% of trials, and 44% of intrusions led to direct interference with the neighbour’s courtship of, or copulation with, Jumping Sylvia (Balsby & Dabelsteen, 2004). The different potential gains and costs of eavesdropping to interactants in agonistic and courtship contexts predict the evolution of strategies that counter, or reduce, the negative consequences and strategies that ignore, or even facilitate, it. In the rest of this chapter, I discuss how communication behaviour can be made public, private or anonymous and how eavesdroppers should behave. The focus is on vocal interactions of territorial songbirds.

41

P1: GFZ/... P2: JZZ/... 0521823617c03.xml CU1917B/McGregor

42

0 521 582361 7

April 10, 2005

17:1

T. Dabelsteen How best to eavesdrop An eavesdropper should attempt to achieve the best conditions for receiving the signals of both interactants simultaneously and at the same time reduce the potential costs of missing more important interactions or being detected by predators. Since eavesdropping always involves increased risk of predation because of divided attention, the duration of an eavesdropping session should be restricted to the time necessary for the ‘intended’ gathering of information. In addition, a male eavesdropper should attempt to stay undetected by the interactants because at least one (e.g. the loser) will always have an interest in concealing the interaction irrespective of its nature, i.e. whether it is agonistic or sexual. A detected female eavesdropper, by comparison, may benefit from the intensified interaction between male interactants. Such intensification of males’ displays and fights is often observed in lekking species when females pass or arrive on a lek (e.g. Lack, 1939; Hovi et al., 1995). Predictions on how best to eavesdrop

Simply approaching the interactants or moving to a position where the signals from both individuals can be received simultaneously and equally well will, of course, enhance eavesdropping. When the information gathered from an interaction depends on the timing of airborne sound signals from two interactants, eavesdropping may be complicated by the relatively slow speed of sound transmission in air. In such cases, eavesdroppers should approach to positions where the distances to the two interactants are equal (e.g. Dabelsteen, 1992). Like simple receivers, eavesdroppers may also improve the conditions for receiving sound signals by ascending to a high perch. Depending on the nature of the surrounding vegetation, this may improve the possibilities for observing visual displays. The evidence for the improvement of sound reception comes from sound transmission experiments that quantified the degradation of natural sound signals transmitted over natural communication distances using natural signaller and receiver positions in the appropriate habitat for the study species (Box 3.1). Sound transmission experiments indicate that simply leaving the ground and flying a few metres up to the undergrowth of a forest will not necessarily improve the receiving conditions if the signaller is already located higher up in the vegetation. This is perhaps not surprising given the ‘ground’effect (attenuation, especially of low-frequency sounds, when sounds are transmitted along the ground) when both the signaller and the receiver are close to the ground (e.g. Wempen, 1986; Embleton, 1996; Nemeth et al., 2001). However, ascent by the receiver to perches above the undergrowth may improve receiving conditions considerably depending on the type of habitat. For instance, a whitethroat that moves from

P1: GFZ/... P2: JZZ/... 0521823617c03.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

17:1

Facilitating and countering eavesdropping 4 to 9 m above ground level in an open whitethroat habitat will not improve sound receiving conditions further (Balsby & Dabelsteen, 2003c). In whitethroats, high perches mainly seem to help visual surveying of the surroundings. There is a very different result in a closed forest habitat before leaf burst. Experiments with songs of three different species, the blackbird (Dabelsteen et al., 1993), the wren Troglodytes troglodytes (Holland et al., 1998, 2001) and the blackcap Sylvia atricapilla (Schmitz et al., 2000; Mathevon et al., 2004), show that receivers may obtain a considerable improvement by ascending to high perches. For instance, a blackcap receiver that moves from 4 to 9 m above ground level obtains improvements that would correspond to a horizontal approach towards the singer of up to 23 m, i.e. almost half an average territory diameter (Fig. 3.1 and Box 3.1) (Schmitz et al., 2000; Mathevon et al., 2004).

Box 3.1 Transmission-caused sound degradation has at least four aspects. (a) Sound signals are attenuated because of spherical spreading (6 dB per doubling of distance) and excess attenuation (EA) caused by absorption and multiple scattering (e.g. Michelsen, 1978). (b) This attenuation will, together with the addition of background noise, reduce the signal-to-noise ratio (SNR). (c) Selective frequency filtering, atmospheric turbulence and reverberation will result in a distortion or blurring within the sounds of their frequency and amplitude patterns over time (e.g. Wiley & Richards, 1982), which can be quantified by a blur ratio (BR; Dabelsteen et al., 1993). (d) Reverberation will cause an elongation of the sounds with tails of echoes, which can be quantified by a tail-to-signal ratio (TSR; Holland et al., 2001). Sound transmission experiments indicate that all these aspects of sound degradation change with distance: EA, BR and TSR increase and SNR decreases with distance. The experiments also indicate that sound degradation decreases with increasing height above ground level of signaller (loudspeaker) and receiver (microphone). This means that both signaller and receiver may improve communication by sound signals by moving up to high perches. Exactly how large these improvements are is perhaps best understood when the regressions of the values of each of the four degradation measures against the logarithm of the distance are used to translate the improvements obtained by moving upwards into virtual horizontal distances that a signaller or receiver would have to approach a receiver or signaller, respectively, to obtain the same improvements. Figure 3.1 shows an example.

43

P1: GFZ/... P2: JZZ/... 0521823617c03.xml CU1917B/McGregor

44

0 521 582361 7

April 10, 2005

17:1

T. Dabelsteen

Sender SNR

BR EA TSR

Receiver TSR

SNR

BR

EA

Fig. 3.1. Virtual horizontal distances (see text) that a blackcap receiver or signaller would have to approach a signaller or a receiver, respectively, to obtain improvements similar to those obtained by moving in the vegetation from 4 to 9 m above ground level. In this example, the distance between signaller and receiver is 50 m and the calculations are based on average degradation values for 10 different blackcap song elements. For a signaller, the average improvements for each of four degradation measures (indicated by silhouettes of flying birds for SNR, BR, EA and TSR; defined in Box 3.1 text) correspond to virtual approaches only slightly longer than the 5 m ascent. For a receiver, the virtual approaches are considerably longer, corresponding to almost half the average territory diameter of 25 m, indicated by the horizontal line along which the bird silhouettes are flying. Movements and vegetation are to scale, whereas bird silhouettes are enlarged. (Modified from Schmitz et al., 2000.)

These sound transmission experiments predicted that eavesdroppers on vocal interactions should ascend to high perches in a forest habitat before leaf burst. By ascending to high perches instead of approaching interactants, an eavesdropper would save energy and also avoid moving too far away from locations where future interactions of potential interest might take place. By staying somewhere in the middle of its territory during periods with high-singing activity, for instance at dawn or during most of the morning, a perching bird would also potentially be able to monitor a number of simultaneously occurring interactions and switch its attention between different interactions depending on where the most interesting developments happen. Staying at some distance from the interactants would

P1: GFZ/... P2: JZZ/... 0521823617c03.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

17:1

Facilitating and countering eavesdropping also help to conceal eavesdropping activity. Exactly where and how high above ground level an eavesdropper should position itself will, of course, depend on the local vegetation and the availability of cover to avoid detection by predators or interactants. Female eavesdroppers have the same interest as males with respect to predators, but not necessarily with respect to the interactants (see above). If the predation risk constrains them to stay hidden in the vegetation, they could, in theory, announce their presence vocally. For how long should individuals eavesdrop? Eavesdropping should be as brief as possible to reduce the risk of predation caused by divided attention, but sufficiently long to allow the extraction of relative information about the interactants. What exactly ‘sufficient’ means seems to depend heavily on the context. For instance, intensive courtship interactions will almost immediately inform an eavesdropper that both members of a pair are ready to copulate, whereas agonistic interactions between males may sometimes progress very slowly and the signalling may only reveal the superior male after some time has passed. In such cases, we cannot predict the duration of eavesdropping because this will be controlled by the interactants. In some species, song repertoire size is correlated with morphometric measures of males, suggesting that repertoire size is capable of providing absolute information about male quality. In such species, relative information about the quality of different individuals could be deduced by receiving the songs from individuals in turn. However, in some of these species, studies have failed to show that repertoire size has signal value. In experimental studies where the design has only allowed simple receiving, the negative results could reflect the fact that the species do not have the capacity for determining absolute repertoire size. It would be interesting to investigate if such species eavesdrop on song duels to gather relative information about repertoire sizes. If so, the minimum time needed to obtain the relative information by eavesdropping could be predicted from cumulative plots of the number of new songs sung as a function of the total number of songs sung by each of the interactants. The time taken to get a reliable indication of the relative repertoire sizes of two interactants would, of course, depend on who is interacting with whom since the cumulative plots of some pairs of individuals become different sooner than other pairs. Male whitethroats, for instance, apparently do not vary their responses to playback of different repertoire sizes (Balsby & Dabelsteen, 2001): maybe because they do not consider repertoire size in agonistic contexts; maybe because they do not have the capacity for determination of absolute repertoire size. If they could perceive relative repertoire sizes, such males would have to eavesdrop for between two minutes (on an interaction between males A and B

45

46 0

100

200

300

400

500

600

0

4000

8000 Total number of elements

C

A

12 000

B

16 000

Balsby, 2000.)

10 minutes, approximately 1000 elements were sung. The letters refer to different males discussed in the text. (Modified from

0 521 582361 7

Fig. 3.2. Cumulative plots of new song elements found as a function of song elements sung by 18 different whitethroat males. In

Number of different elements

700

P1: GFZ/... P2: JZZ/... 0521823617c03.xml CU1917B/McGregor April 10, 2005 17:1

P1: GFZ/... P2: JZZ/... 0521823617c03.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

17:1

Facilitating and countering eavesdropping in Fig. 3.2) and more than 20 minutes (when males B and C are interacting) to detect which male has the largest repertoire (Balsby, 2000). Do eavesdroppers fit the predictions?

The experiments that sought to establish eavesdropping behaviour in songbirds did not monitor the locations of subjects in a way that allows the above predictions to be tested. However, there is evidence (personal observation) that male great tits may approach and/or be perched during experimentally induced opportunities for eavesdropping. There are also observations of neighbouring males approaching during interactive playback to territorial blackbirds, and male blackbirds may interrupt ongoing behaviour and fly to high perches and stay silent in response to other males starting to sing (personal observation). Although anecdotal, such observations support the idea that eavesdroppers may attempt to achieve the best conditions for receiving the signals of interactants. It is also possible that some of the variation in the results obtained in the field experiments with great tits (Otter et al., 1999; Peake et al., 2001, 2002) was due to the ‘uncontrollable’ eavesdroppers (i.e. subjects) having been in positions that varied in how well the interactants could be heard. Evidence suggesting that birds probably engaging in eavesdropping behave in a way that would best receive signals of interactants comes from a radio-tracking study of 11 unmated female great reed warblers Acrocephalus arundinaceus (Bensch & Hasselquist, 1992). Their routes were taken as evidence for female assessment preceding mate choice (Bensch & Hasselquist, 1992; but see Ch. 7). Figure 3.3 shows the route followed by one such female. During the eight hours the female was followed, it had four relatively long stays within a limited area, indicated by clusters of black dots in Fig. 3.3. The first cluster indicates a stay of 40 minutes at the location of the female early in the morning before it starts to move. The other three clusters, each corresponding to stays of about 60 minutes’ duration, are at locations with almost exactly equal, relatively short, distances to two males at a time: perfect positions for eavesdropping on vocal interactions between males. One cluster would allow eavesdropping on interactions between males VII and III, the next on males III and II, and finally on males V and VI. These results suggest that about 60 minutes was needed to gather information from two males at a time, including relative information from the occasional singing interactions between the males. It has been hypothesized that females of some bird species use loud fertility advertisement calls to incite male–male interactions (e.g. Montgomerie & Thornhill, 1989; Hoi, 1997). Although the actual fertility advertisement function of the loud female calls is doubtful, it is possible that such calls may initiate or even intensify ongoing male–male interactions and hence act as an aid in mate choice

47

P1: GFZ/... P2: JZZ/... 0521823617c03.xml CU1917B/McGregor

48

0 521 582361 7

April 10, 2005

17:1

T. Dabelsteen

Fig. 3.3. Map showing the movements of a female great reed warbler in an area with nine male territories, the positions of which are marked with bird silhouettes. The female, which was equipped with a radio transmitter and followed for eight hours, had its position monitored every 10 minutes, as indicated by black dots. Four clusters each consisting of four to six dots have been encircled. The one to the left marks the position of the female when she was first tracked at 04:50 hours. The three clusters to the right mark positions from which the female may have been eavesdropping on singing interactions between dyads of males.

(e.g. Sæther, 2002; Ch. 7). A female that utters such calls to intensify an ongoing interaction would be fitting one of the above predictions for how best to eavesdrop.

Public signalling: facilitating eavesdropping The term advertising is normally used for signalling that makes the signaller advertise itself to a wide audience with respect to some quality or capacity. Advertising signals transmit over relatively long distances and are used in solo

P1: GFZ/... P2: JZZ/... 0521823617c03.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

17:1

Facilitating and countering eavesdropping signalling as well as in signalling interactions. A good example of such an advertising signal is the full song of songbirds, which often has a dual function: to attract potential mates and repel rivals. Interactants using advertising signals will, of course, expose themselves to eavesdropping; however, if such eavesdropping has no immediate or subsequent adverse consequences, then it should not change the signalling behaviour of interactants. Yet there could also be situations where one of the interactants might benefit from the presence of eavesdroppers, for instance by making its superiority relative to the opponent widely known. Here the superior individual should continue signalling or even attempt to make it more effective, whereas the inferior individual should stop using advertising signals and ultimately stop interacting. It could be argued that an individual should always make advertising signalling as effective as possible. However, when effectiveness depends on energy used or some other cost, animals may wish to limit the costs during solo signalling more than during the usually shorter signalling interactions, perhaps to avoid exhaustion. Predictions for public signalling

Signallers could allow eavesdropping simply by choosing signals that transmit effectively in their physical environment and by signalling from positions and at times of the day that would maximize signal transmission. The evidence for such choices in use of sound signals comes from sound-transmission experiments. All of the experiments with bird song mentioned above show that some sound types transmit better than others, even among functionally equivalent types. For instance, low-frequency, narrowband and unmodulated sounds seem to transmit best in a forest habitat. In the full song of the blackbird (Fig. 3.4a), the introductory low-frequency whistle or motif sounds transmit much better than the terminating highly modulated and broadband twitter sounds; among the whistle sounds, the relatively unmodulated and constant frequency ‘CF-sounds’ transmit better than sounds that are frequency modulated (‘FM-sounds’)and/or have energy rich overtones (‘MIX-sounds’) (Dabelsteen et al., 1993). Also in great tit song, relatively ‘pure-toned’sound elements seem to transmit better than highly modulated ‘buzz’ elements (Blumenrath et al., 2004). In blackcap song the terminating relatively low frequency, pure tone and narrowband motif sounds are less attenuated than the introductory highly modulated broadband twitter sounds (Dabelsteen & Mathevon, 2002). In wren song, low frequency sounds are least attenuated (Holland et al., 1998), and in antbird (Thamnophilidae) song, the narrowband and low-frequency sounds transmit best (Nemeth et al., 2001). Low-frequency sounds also have another property that make them suited for advertising: they usually radiate from the vocalizing individual more or less equally well in all directions (i.e. they are omnidirectional). This has been shown

49

P1: GFZ/... P2: JZZ/... 0521823617c03.xml CU1917B/McGregor

50

0 521 582361 7

April 10, 2005

17:1

T. Dabelsteen

(a) kHz 10 8 6 4 2 0.5

1.0

1.5

2.0

2.5

3.0

0.5

1.0

1.5

2.0

2.5

3.0

s

(b) kHz 10 8 6 4 2 3.5

4.0

4.5

s

Fig. 3.4. Sound spectrograms of blackbird song. (a) A male full song recorded in the field; the introductory whistle part is underlined. (b) A male full song, which is framed, has the highest frequencies of the terminating twitter part overlapped by a female copulation trill, the duration of which is indicated by a horizontal line. This was recorded in an anechoic chamber and shows the part of a playback trial where playback of full song to an oestradiol-treated female elicits a copulation solicitation posture and a copulation trill (see Dabelsteen, 1988). The movements associated with the female posturing are indicated by broadband white noise, especially over the whistle part of the male’s full song but also over the last third of the copulation trill. Spectrograms were produced in Avisoft (FFT 512 points, flat top, overlap 75%, frequency resolution 43 Hz, time resolution 5.805 ms).

in laboratory studies with blackbirds (Larsen & Dabelsteen, 1990). Species using sounds with a more directional radiation pattern may facilitate advertising by moving their head from side to side during singing (Brumm & Todt, 2003). Sound transmission is normally believed to be most effective from high perches. The sound transmission experiments with bird song also indicate that a signaller should at least ascend to a few metres above ground level to make sound transmission effective (e.g. Nemeth et al., 2001); transmission may sometimes be improved further by ascending to high perches, but only slightly. In a forest, the improvement that a signaller obtains by moving to higher perches sometimes seems very small relative to that a receiver might obtain by the same movement (e.g. Fig. 3.1); consequently, high perches should perhaps be called listening posts rather than song posts in these cases (Schmitz et al., 2000; Mathevon et al., 2004). Sound transmission is also believed to be most effective at dawn (e.g. Henwood & Fabric, 1979).

P1: GFZ/... P2: JZZ/... 0521823617c03.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

17:1

Facilitating and countering eavesdropping However, very few studies have investigated this with sound transmission experiments using natural sound signals. A recent study with blackcap song, at the peak season of their singing immediately after their return to the breeding sites in the spring, failed to show that their songs propagate most effectively at dawn (Dabelsteen & Mathevon, 2002). Overall, the sound transmission experiments suggest that the most widely propagating sound signals would be relatively loud, low frequency, narrowband and unmodulated; they would be emitted from at least a few metres above the ground. The effect of higher song perches is small and the optimal time of day is uncertain and probably depends strongly on the weather. Windy and rainy conditions will constrain advertising for different reasons (e.g. Lengagne & Slater, 2002), although birds, in theory, might compensate for a high level of background noise by increasing the output level of songs (Lombard effect; Lombard, 1911). For instance, nightingales Luscinia megarhynchos may use a higher song output level in noisy environments than at less-noisy locations (Brumm, 2004). Do interactants use public signals to facilitate eavesdropping?

Observations suggest that birds sometimes increase the loudness of their singing when they shift from solo singing to counter-singing (i.e. interacting). Unfortunately, such observations have rarely been verified with sound pressure level (SPL) measurements. An exception is a study by Brumm & Todt (2002), which showed that male nightingales singing full songs increase the SPL by more than 5 dB when they shifted from solo singing to playback-induced counter-singing. However, it is a question of whether the increase in SPL has evolved to facilitate eavesdropping or whether it represents an increased arousal of the singers, which as a side effect inevitably facilitates eavesdropping on the vocal interaction. Playback experiments support the arousal hypothesis because a higher SPL elicits a stronger response (e.g. Dabelsteen & Pedersen, 1992). There is, as yet, no evidence that interactants facilitate eavesdropping during interactions by using signals from their repertoire of functionally equivalent advertising signals that transmit more effectively than those used during solo signalling. For song types, this would imply that blackbirds engaged in a song duel across the border between their territories should use more CF-sounds and fewer FM- and MIX-sounds than during solo singing. However, the opposite seems to happen, perhaps because the sound types are not functionally equivalent, with the FM-sounds expressing the highest degree of arousal (Dabelsteen & Pedersen, 1992). Neither is there any evidence that higher perches are used during interactive singing than during solo singing. In both contexts, the use of relatively high perches is probably mainly to improve sound reception, either of the songs of the opponent or of vocal responses to the solo songs.

51

P1: GFZ/... P2: JZZ/... 0521823617c03.xml CU1917B/McGregor

52

0 521 582361 7

April 10, 2005

17:1

T. Dabelsteen When singing interactions using advertising signals escalate, the interactants often shift to special types of close range, private singing (see below). However, after agonistic encounters, one or both of them may shift back to loud advertising songs. When both individuals do this and continue to interact, the shift back to advertising songs can be said to facilitate eavesdropping on their interaction. The advertising singing of the eventual winner of the escalated interaction is often different from that of the loser and, therefore, is referred to as an ‘acoustic victory display’ (e.g. Bradbury & Vehrencamp, 1998; Ch. 6). When only one of the birds shifts back to advertising songs, it is usually the winner of the interaction. In this case, the winner’s singing cannot be said to facilitate eavesdropping since there is only one signaller left.

Private signalling: countering eavesdropping An interactant should use private signals in an interaction whenever public signals would incur potential risks. This is true when the risks are immediate, for example from predators because of reduced vigilance during interactions compared with solo signalling, or from male eavesdroppers, which may take advantage of an interactant’s involvement with a rival to pay its mate a visit during the interaction. It is also true when the risks are less immediate, for example eavesdroppers that can extract information about an interactant and utilize this in future encounters with it. Whatever the risks, they seem likely to increase with the duration of the interaction. An important step to make an interaction private would, therefore, be to make it as brief as possible, because this would reduce the risk of detecting the signalling interaction and of obtaining useful information. However, as explained in the next section, there are other options. Predictions for private signalling

Anything that limits signal transmission and reception may, of course, help to make signal interactions private, in effect the opposite of advertising. Private signalling should, therefore, employ signals, signaller positions and signalling times that are the opposite of those used for advertising. Vocally interacting birds that wish to signal privately should use sound signals that are relatively high frequency, broadband and highly modulated, and emit them with a low SPL. This would reduce the number of potential receivers for two reasons: such sounds attenuate and degrade relatively fast with distance (e.g. Dabelsteen et al., 1993; Holland et al., 1998) and are directional in the sense of being beamed away from the sender in one direction (e.g. Larsen & Dabelsteen, 1990). Sound transmission experiments also show that privatizing would be most effective if the interactions take place close to the ground or in the undergrowth of a forest.

P1: GFZ/... P2: JZZ/... 0521823617c03.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

17:1

Facilitating and countering eavesdropping Background noise, constant as well as transient, may mask communication sounds and hence contribute to private signalling. This may be especially important when eavesdropping involves information gathered from the timing of songs. For instance, the degree to which songbirds delay their songs relative to each other or overlap each other’s songs seem to be important indicators of social dominance (e.g. Dabelsteen et al., 1997, 1998; Langemann et al., 2000) that are utilized by eavesdroppers (e.g. McGregor et al., 1997; Peake et al., 2001). Receivers in general may have internal representations of songs which would help them to reconstruct songs which are partially masked by transient background noise, for instance the vocalizations of other birds. However, such representations would probably not be of much help to eavesdroppers in reconstructing how interactants delay their songs relative to each other or overlap each other’s songs (e.g. Poesel et al., 2001). Therefore, private signalling should take place at times of the day when the level of background noise is high.

Do interactants use private signals to counter eavesdropping?

So-called quiet song in songbirds, sometimes referred to as soft or whisper song, seems to be a good candidate for a private signal (e.g. Dabelsteen et al., 1998). Quiet singing is still relatively unexplored, but wherever it has been discovered it has been accompanied by an active behaviour rather than perching, for instance posturing and/or escorting or chasing another individual, and it is often sung more or less continuously without the intersong pauses characteristic of full singing. Quiet singing usually occurs at close range during escalated interactions, collaborative as well as competitive, i.e. in contexts of eavesdropping (e.g. Titus, 1998; Balsby, 2000; Balsby & Dabelsteen, 2003a,b). Good examples of quiet singing occur in the blackbird, the redwing Turdus iliacus, the robin, the dunnock and the alpine accentor Prunella collaris (Dabelsteen et al., 1998). The blackbird has at least three types of quiet singing. Counter-singing males switch from loud full songs with whistle sounds to quiet twittering without whistles when the interaction escalates and the interactants approach each other to within about 10 m. A male also twitters quietly during intense courtship of a soliciting female immediately before copulation, and the soliciting female sings a very quiet copulation trill (Figs. 3.4b and 3.5). Until recently, the aggressive and sexual twitter were believed to be identical because that is how they sound to the human observer (e.g. Dabelsteen et al., 1998). However, recent spectrographic analyses have suggested that sexual twitter may also include specific sounds (Fl´eron, 2003) and hence have the potential to communicate the sexual arousal of the singer in the same way as the accompanying posturing of the male (e.g. Snow, 1958).

53

P1: GFZ/... P2: JZZ/... 0521823617c03.xml CU1917B/McGregor

54

0 521 582361 7

April 10, 2005

17:1

T. Dabelsteen kHz 10 8 6 4 2 0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.5

6.0

6.5

7.0

7.5

8.0

8.5

9.0

9.5

kHz 10 8 6 4 2

s

Fig. 3.5. Sound spectrogram of the quiet vocalizations uttered during 10 seconds of a natural blackbird courtship interaction early in the morning soon after the dawn chorus. The courtship vocalizations dominate among the sounds that are visible on the spectrogram above 4 kHz, most of them being male sexual twitter. A horizontal line indicates a trill that is likely to be the female copulation trill. Two full songs of a robin that is singing close by are indicated by dashed horizontal lines: the first has visible energy up to about 5 kHz; the second starts with a high-frequency part (approximately 8 kHz) and ends with a more low-frequency part (approximately 4 kHz). Most of the sounds between 1.8 and 3 kHz are whistle parts of more distant singing blackbird males. Spectrograms were produced as in Fig. 3.4.

Quiet singing in the blackbird seems to fit nearly all of the predictions for a private signal, with respect to sound type, variability and SPL, and with respect to where, when and for how long it is used. Relative to the whistle sounds of full song, all three types of quiet song consist of relatively high-frequency, broadband and modulated sounds (Figs. 3.4, and 3.5). Twitters are at least 10 dB(A) quieter than whistles (Dabelsteen, 1984) and sound transmission experiments have shown that the twitters degrade and attenuate much faster than whistles (Dabelsteen et al., 1993). Whereas whistles seem capable of transmitting over at least two to four territory diameters, twitters usually transmit less than one diameter (Dabelsteen et al., 1993). Twitters are also more directional than whistles (Larsen & Dabelsteen, 1990). Female copulation trills sound even quieter than male twitters (personal observations) and probably have a radiation pattern and transmission capability

P1: GFZ/... P2: JZZ/... 0521823617c03.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

17:1

Facilitating and countering eavesdropping resembling that of twitters. Both aggressive twitter interactions and courtship interactions are most frequently performed under cover close to the ground or in the undergrowth (personal observation) and early in the morning after the males’ dawn chorus at a time where the level of background noise from the vocalizations of other species is very high (Fig. 3.5) (e.g. Messmer & Messmer, 1956). Finally, the duration of aggressive twitter interactions can vary quite a lot, but courtship interactions are always very brief (e.g. Snow, 1958). The term quiet singing should not be taken too literally since some of the bird vocalizations referred to as calls seem to be used in the same contexts and to have the same acoustic structure as quiet singing. Good examples are the dscharpand ze-calls of the female whitethroat, which are used in courtship interactions where the male performs a special diving song display (Balsby & Dabelsteen, 2002, 2003a,b). The two calls fit the requirements for private signals: they have sound structures and low SPL (measured for dscharp-call, too low to be measured for ze-call), which make them short-range signals, and they are uttered from low positions in the vegetation (Balsby & Dabelsteen, 2003a,b). The eavesdropping contexts with quiet vocalizing are characterized by risks to the interactants from predation and eavesdropping. This probably applies to all eavesdropping contexts, and both these risks seem strong enough to cause the evolution of private signals. It will, therefore, be very difficulty to disentangle the relative influence of the two risks in the evolution of quiet singing and calling. Quiet vocalizing might also have evolved simply to save energy during close-range communication. However, this seems unlikely given that quiet vocalizations are always accompanied by movements and sometimes even by posturing. This, together with the continuous nature of quiet vocalizing, suggests that birds do not necessarily save energy by switching from loud advertising singing to quiet singing. If energy saving was the main purpose, the bird could simply lower the output level of singing as in the non-social subsong (e.g. Thorpe & Pilcher, 1959). They do not have to switch to another song type and otherwise behave in a way that would counter eavesdropping. However, there are contexts where predation risk may be the main factor responsible for private signals, but these contexts do not, as far as I know, involve a signalling interaction and, therefore, do not constitute eavesdropping in the sense used in this chapter. A good example is the so-called nest-relief song or calling-out song produced by males of many songbird species (e.g. Gompertz, 1961; Stork, 1971; Ficken et al., 1978; Lind et al., 1996). Structurally, these songs seem identical to the full songs used during advertising singing, but they are usually sung with a very low SPL (although higher than in subsong; personal observations) and they are also relatively short or few in number.

55

P1: GFZ/... P2: JZZ/... 0521823617c03.xml CU1917B/McGregor

56

0 521 582361 7

April 10, 2005

17:1

T. Dabelsteen Anonymity: another way to counter eavesdropping Private signals counter eavesdropping but do not necessarily prevent it. However, if interactants can remain anonymous, eavesdroppers cannot attribute information gained from the interaction to any particular individual and, therefore, this cost of eavesdropping to interactants is removed. For instance, eavesdroppers on an interaction between unknown individuals could not make use of information in subsequent encounters with them. Predictions for anonymity

Anything that makes individual identification of a signaller based on its signalling activity more difficult would help anonymity. One way to constrain individual recognition of vocalizing birds might be to increase the variability of their vocal output. For instance, experiments suggest that song repertoires of fewer than about 25 song types do not interfere with song-based neighbour recognition, whereas repertoires of more than about 100 do (e.g. Stoddard, 1996; Molles & Vehrencamp, 2001). Birds that want to make their singing anonymous should, therefore, change the way they use their repertoire or somehow increase it during interactions, for instance by switching to another type of singing. If a male can be identified by the repertoire being sung from particular song posts, switching to unusual posts combined with frequent post shifts might also help to achieve anonymity. Do animals make themselves anonymous in eavesdropping contexts?

Quiet singing in songbirds may aid anonymity because, unlike full song, it is used almost everywhere and not from preferred posts, and most importantly it seems a lot more variable than full song (e.g. Dabelsteen et al., 1998). For instance, when a rival intrudes, a male blackbird uses the aggressive twitter everywhere in its territory rather than from its usual song posts. Also its repertoire of different aggressive twitter motifs far exceeds the threshold value (> 100; Fl´eron, 2003) that constrains individual identification (e.g. Stoddard, 1996). The repertoire size of full-song whistles averages 44 (Rasmussen & Dabelsteen, 2001). The idea of anonymity in signals is novel and, therefore, not really studied yet, but it seems clear that, unlike private signals, anonymity is difficult to explain as a response to predation risk. Unlike the aspects of private signals mentioned above, which can reduce the risk of detection by predators, increased variation per se does not seem to have such an effect. Relative to full song, quiet song endowed with a variation similar to that of full song would still be very difficult for a predator to detect. The evolution of the relatively larger variation in quiet singing could have resulted from sexual selection, which is believed to play a role

P1: GFZ/... P2: JZZ/... 0521823617c03.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

17:1

Facilitating and countering eavesdropping in the evolution of avian repertoire sizes (e.g. Searcy & Yasukawa, 1996). However, this would require sexual selection to have acted more strongly on quiet song than on full song, even though both types of song seem to play important roles in mate choice and deterrence of rivals. Finally, the relatively large variation of quiet song could be a side effect of a sound production mechanism, coupling low SPL or high frequencies with large variation. However, this also seems unlikely given that full song is sometimes uttered in non-social contexts as very quiet so-called subsong (Thorpe & Pilcher, 1959) and that high-frequency quiet singing contains the same type of fixed combinations of sound elements as full song (e.g. Rasmussen & Dabelsteen, 2001; Fl´eron, 2003). At the moment, it seems likely that the large variation of quiet singing relative to full singing in some species reflects the need for singer anonymity to counter negative consequences of eavesdropping.

Summary There is now experimental evidence that animals have the ability to gather relative information about interactants by eavesdropping and utilize this information in subsequent decision making about how to behave towards the interactants. We still lack good observational evidence that this happens in non-experimental natural contexts, but this seems likely given the obvious advantages to the eavesdropper. It is also likely that eavesdroppers behave in ways that enhance their ability to eavesdrop. The potential gains of being eavesdropped upon are more difficult to identify but may exist in special situations and, therefore, have led to the facilitation of eavesdropping, including enhancing advertising signals. The potential costs of being eavesdropped upon are much more obvious and set the scene for an evolutionary arms race between eavesdroppers and interactants, with private signals and anonymity reducing the costs of being subjected to eavesdropping. In this chapter, I have concentrated on vocal interactions between songbirds and made predictions for what eavesdroppers and interactants should do in terms of their relative positioning and the type of sound signals that interactants should use. Some of the predictions may seem trivial or speculative, while others are more substantial because they are derived from the results of sound-transmission experiments. At the moment, there is anecdotal evidence for all of the predictions except aspects of public signals, where there is stronger evidence. There is also strong evidence for private signals. Quiet singing in songbirds fulfils most of the predictions for private signalling and seems to do so for anonymity as well. Other selection pressures are also important; for instance predation risk is likely to have played an important role in the evolution of private signals. An interesting challenge for future research on these matters will be to investigate eavesdropping in natural non-experimental contexts. In territorial songbirds,

57

P1: GFZ/... P2: JZZ/... 0521823617c03.xml CU1917B/McGregor

58

0 521 582361 7

April 10, 2005

17:1

T. Dabelsteen this could be done by quantifying how territorial subject males behave during vocal interactions between other males and in subsequent vocal interactions with them. Such a study would need recordings of the vocal activity in a local network of males using acoustic location systems, as done by, for instance, Bower (2000) and Burt & Vehrencamp (Ch. 15), combined with monitoring the movements of silent individuals by radio-tracking. If the study located birds in three dimensions, it might also test the predictions on positioning and signal use discussed above. If the study was long term and included collection of data on predation, survival and reproductive success, it might help to disentangle the influence of the different selection pressures on the evolution of communicating in a network.

Acknowledgements Thorsten Balsby kindly commented on the manuscript. Sandra Blumenrath drew Fig. 3.1 and Thorsten Balsby assisted with the production of the remaining figures. Henrik Brumm gave me access to unpublished manuscripts. The main part of my own research on communication network activities forming the basis for this chapter has been funded by the Danish National Science Foundation.

References Balsby, T. J. S. 2000. Song activity and variability in relation to male quality and female choice in whitethroats Sylvia communis. Journal of Avian Biology, 31, 56–62. Balsby, T. J. S. & Dabelsteen, T. 2001. The meaning of song repertoire size and song length to male whitethroats Sylvia communis. Behavioural Processes, 56, 75–84. 2002. Female behaviour affects male courtship in whitethroats, Sylvia communis. An interactive experiment using visual and acoustic cues. Animal Behaviour, 63, 251–257. 2003a. Male singing behaviour and female presence in the territory in whitethroats. Acta Ethologica, 5, 81–88. 2003b. Male calling between courtship sequences in whitethroats: a way to counter intrusions from neighbouring rivals. Behavioural Processes, 63, 149–157. 2003c. Degradation of whitethroat Sylvia communis song: implications for song flight and communication network activities. Behaviour, 140, 695–719. 2004. Simulated courtship interactions elicit neighbour intrusions in the whitethroat Sylvia communis. Animal Behaviour, in press. Bensch, S. & Hasselquist, D. 1992. Evidence for active female choice in a polygynous warbler. Animal Behaviour, 44, 301–311. Blumenrath, S. H., Dabelsteen, T. & Pedersen, S. B. 2004. Being inside nest boxes: does it complicate the receiving conditions for great tit Parus major females? Bioacoustics, in press.

P1: GFZ/... P2: JZZ/... 0521823617c03.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

17:1

Facilitating and countering eavesdropping Bower, J. L. 2000. Acoustic interactions during naturally occurring territorial conflict in a song sparrow neighborhood. Ph.D. Thesis, Cornell University, Ithaca, New York. Brumm, H. 2004. The impact of environmental noise on song amplitude in a territorial bird. Journal of Animal Ecology, 73, 434–440. Brumm, H. & Todt, D. 2002. Noise-dependent song amplitude regulation in a territorial songbird. Animal Behaviour, 63, 891–897. 2003. Facing the rival: directional singing behaviour in nightingales. Behaviour, 140, 43–53. Bradbury, J. W. & Vehrencamp, S. L. 1998. The Principles of Animal Communication. Sunderland, MA: Sinauer. Dabelsteen, T. 1984. An analysis of the full song of the blackbird Turdus merula with respect to message coding and adaptations for acoustic communication. Ornis Scandinavica, 15, 227–239. 1985. Messages and meanings of bird song with special reference to the blackbird (Turdus merula) and some methodology problems. Det Kongelige Danske Videnskabernes Selskab Biologiske Skrifter, 25, 173–208. 1988. The meaning of the full song of the blackbird, Turdus merula, to untreated and oestradiol-treated females. Ornis Scandinavica, 19, 7–16. 1992. Interactive playback: a finely tuned response. In: Playback and Studies of Animal Communication, ed. P. K. McGregor. New York: Plenum Press, pp. 92–109. Dabelsteen, T. & Mathevon, N. 2002. Why do songbirds sing intensively at dawn? A test of the acoustic transmission hypothesis. Acta Ethologica, 4, 65–72. Dabelsteen, T. & Pedersen, S. B. 1990. Song and information about aggressive responses of blackbirds Turdus merula: evidence from interactive playback experiments with territory owners. Animal Behaviour, 40, 1158–1168. 1992. Song features essential for species discrimination and behaviour assessment by male blackbirds (Turdus merula). Behaviour, 121, 259–287. Dabelsteen, T., Larsen, O. N. & Pedersen, S. B. 1993. Habitat-induced degradation of sound signals: quantifying the effects of communication sounds and bird location on blur ratio, excess attenuation, and signal-to-noise ratio in blackbird song. Journal of the Acoustical Society of America, 93, 2206–2220. Dabelsteen, T., McGregor, P. K., Holland, J., Tobias, J. & Pedersen, S. B. 1997. The signal value of overlapping singing in male robins (Erithacus rubecula). Animal Behaviour, 53, 249–256. Dabelsteen, T., McGregor, P. K., Lampe, H. M., Langmore, N. E. & Holland, J. 1998. Quiet singing in song birds: an overlooked phenomenon. Bioacoustics, 9, 89–105. Davies, N. B. 1992. Dunnock Behaviour and Social Evolution. Oxford: Oxford University Press. Doutrelant, C. & McGregor, P. K. 2000. Eavesdropping and mate choice in female fighting fish. Behaviour, 137, 1655–1659. Dukas, R. & Kamil, A. C. 2001. Limited attention: the constraint underlying search image. Behavioral Ecology, 12, 192–199.

59

P1: GFZ/... P2: JZZ/... 0521823617c03.xml CU1917B/McGregor

60

0 521 582361 7

April 10, 2005

17:1

T. Dabelsteen Embleton, T. F. W. 1996. Tutorial on sound propagation outdoors. Journal of the Acoustical Society of America, 100, 31–48. Ficken, M. S., Ficken, R. W. & Witkin, S. R. 1978. Vocal repertoire of the black-capped chickadee. The Auk, 95, 34–48. Fl´eron, G. M. W. 2003. Stille sang hos solsorten, Turdus merula. M.Sc. Thesis, University of Copenhagen. Gompertz, T. 1961. The vocabulary of the great tit. British Birds, 54, 369–394, 409–418. Henwood, K. & Fabric, A. 1979. A quantitative analysis of the dawn chorus: temporal selection for communicatory optimisation. American Naturalist, 114, 260–274. Hoi, H. 1997. Assessment of the quality of copulation partners in the monogamous bearded tit. Animal Behaviour, 53, 277–286. Holland, J., Dabelsteen, T., Pedersen, S. B. & Larsen, O. N. 1998. Degradation of wren Troglodytes troglodytes song: implications for information transfer and ranging. Journal of the Acoustical Society of America, 103, 2154–2166. Holland, J., Dabelsteen, T., Pedersen, S. B. & Paris, A. L. 2001. Potential ranging cues contained within the energetic pauses of transmitted wren song. Bioacoustics, 12, 3–20. Hovi, M., Alatalo, R. V. & Siikam¨ aki, P. 1995. Black grouse leks on ice: female mate sampling by incitation of male competition? Behavioral Ecology and Sociobiology, 37, 283–288. Jakobsson, S., Brick, O. & Kullberg, C. 1995. Escalated fighting behaviour incurs increased predation risk. Animal Behaviour, 49, 137–142. Lack, D. 1939. The display of the blackcock. British Birds, 32, 290–303. 1940. Observations on captive robins. British Birds, 33, 262–270. Langemann, U., Tavares, J. P., Peake, T. M. & McGregor, P. K. 2000. Responses of great tits to escalating patterns of playback. Behaviour, 137, 451–471. Larsen, O. N. & Dabelsteen, T. 1990. Directionality of blackbird vocalization. Implications for vocal communication and its further study. Ornis Scandinavica, 21, 37–45. Lengagne, T. & Slater, P. J. B. 2002. The effects of rain on acoustic communication: tawny owls have good reason for calling less in wet weather. Proceedings of the Royal Society of London, Series B, 269, 2121–2125. Lind, H., Dabelsteen, T. & McGregor, P. K. 1996. Female great tits can identify mates by song. Animal Behaviour, 25, 667–671. Lombard, E. 1911. Le signe de l’´el´evation de la voix. Annales des Maladies de l’Oreille et du Larynx, 37, 101–119. Mathevon, N., Dabelsteen, T. & Blumenrath, S. H. 2004. Are high perches in the blackcap Sylvia atricapilla song or listening posts? A sound transmission study. Journal of the Acoustical Society of America, in press. McGregor, P. K. 1993. Signalling in territorial systems: a context for individual identification, ranging and eavesdropping. Philosophical Transactions of the. Royal Society of London, Series B, 340, 237–244.

P1: GFZ/... P2: JZZ/... 0521823617c03.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

17:1

Facilitating and countering eavesdropping McGregor, P. K. & Dabelsteen, T. 1996. Communication networks. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca, NY: Cornell University Press, pp. 409–425. McGregor, P. K., Dabelsteen, T. & Holland, J. 1997. Eavesdropping in a territorial songbird communication network: preliminary results. Bioacoustics, 8, 253–254. McGregor, P. K., Peake, T. M. & Lampe, H. M. 2001. Fighting fish Betta splendens extract relative information from apparent interactions: what happens when what you see isn’t what you get? Animal Behaviour, 62, 1059–1065. Mennill, D. J., Ratcliffe, L. M. & Boag, P. T. 2002. Female eavesdropping on male song contests in songbirds. Science, 296, 873. Messmer, E. & Messmer, I. 1956. Die Entwicklung der Laut¨ ausserungen und einiger Verhaltensweisen der Amsel (Turdus m. merula L.) unter nat¨ urlichen Bedingungen und nach Einzelaufzucht in schalldichten R¨ aumen. Zeitschrift f¨ ur Tierpsychologie, 13, 341–441. Michelsen, A. 1978. Sound reception in different environments. In: Sensory Ecology, ed. M. A. Ali. New York: Plenum Press, pp. 345–373. Molles, L. E. & Vehrencamp, S. L. 2001. Neighbor recognition by resident males in the banded wren, Thryothorus pleurostictus, a tropical songbird with high song type sharing. Animal Behaviour, 61, 119–127. Montgomerie, R. & Thornhill, R. 1989. Fertility advertisement in birds: a means of inciting male–male competition? Ethology, 81, 209–220. Naguib, M. & Todt, D. 1997. Effects of dyadic vocal interactions on other conspecific receivers in nightingales. Animal Behaviour, 54, 1535–1543. Naguib, M., Fichtel, C. & Todt, D. 1999. Nightingales respond more strongly to vocal leaders of simulated dyadic interactions. Proceedings of the Royal Society of London, Series B, 266, 537–542. Nemeth, E., Winkler, H. & Dabelsteen, T. 2001. Differential degradation of antbird songs in a Neotropical rainforest: adaptation to perch height? Journal of the Acoustical Society of America, 110, 3263–3274. Oliveira, R. F., McGregor, P. K. & Latruffe, C. 1998. Know thine enemy: fighting fish gather information from observing conspecifics interactions. Proceedings of the Royal Society of London, Series B, 265, 1045–1049. Otter, K. A., McGregor, P. K., Terry, A. M. R. et al. 1999. Do female great tits Parus major assess males by eavesdropping? A field study using interactive song playback. Proceedings of the Royal Society of London, Series B, 266, 1305–1309 Peake, T. M., Terry, A. M. R., McGregor, P. K. & Dabelsteen, T. 2001. Male great tits eavesdrop on simulated male-to-male vocal interactions. Proceedings of the Royal Society of London, Series B, 268, 1183–1187. 2002. Do great tits assess rivals by combining direct experience with information gathered by eavesdropping? Proceedings of the Royal Society of London, Series B, 269, 1925–1929. Poesel, A., Dabelsteen, T. & Pedersen, S. B. 2001. Making yourself heard: a study of masking effects on blue tit Parus caeruleus singing interactions. In Proceedings of the XVIII International BioAcoustics Conference, p. 12.

61

P1: GFZ/... P2: JZZ/... 0521823617c03.xml CU1917B/McGregor

62

0 521 582361 7

April 10, 2005

17:1

T. Dabelsteen Rasmussen, R. & Dabelsteen, T. 2001. Song repertoires and repertoire sharing in a local group of blackbirds. Bioacoustics, 13, 63–76. Sæther, S. A. 2002. Female calls in lek-mating birds: indirect mate choice, female competition for mates, or direct mate choice? Behavioral Ecology, 13, 344–352. Schmitz, S., Dabelsteen, T. & Mathevon, N. 2000. High perches in the blackcap Sylvia atricapilla: song posts or listening posts? In Proceedings of the 8th International Behavioral Ecology Congress, p. 175. Searcy, W. A. & Yasukawa, K. 1996. Song and female choice. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma and E. H. Miller. Ithaca, NY: Cornell University Press, pp. 454–473. Snow, D. W. 1958. A Study of Blackbirds. London: Allen and Unwin. Stoddard, P. K. 1996. Vocal recognition of neighbours by territorial passerines. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma and E. H. Miller. Ithaca, NY: Cornell University Press, pp. 356–374. Stork, H.-J. 1971. Zur socialen Funktion des Gesanges der Amsel Turdus merula L. Zeitschrift f¨ ur Tierpsychologie, 28, 54–58. Thorpe, W. H. & Pilcher, P. M. 1959. The nature and characteristics of sub-song. British Birds, 51, 509–514. Titus, R. C. 1998. Short-range and long-range songs: use of two acoustically distinct song classes by dark-eyed juncos. The Auk, 115, 386–393. Wempen, J. 1986. Outdoor sound propagation close to the ground. In: Sound Propagation in Forested Areas and Shelterbeds, ed. M. J. M. Martens. Nijmegen: Faculty of Sciences, Catholic University, Nijmegen, pp. 83–106. Wiley, R. H. & Poston, J. 1996. Indirect mate choice, competition for mates, and coevolution of the sexes. Evolution, 50, 1371–1380. Wiley, R. H. & Richards, D. G. 1982. Adaptations for acoustic communication in birds: transmission and signal detection. In: Acoustic Communication in Birds, Vol. 1, ed. D. E. Kroodsma & E. H. Miller. New York: Academic Press, pp. 131–181.

P1: GFZ/... P2: JZZ/... 0521823617c04.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

17:49

4

Performing in front of an audience: signallers and the social environment ric ardo j. matos1 & ingo schlupp2 1 2

University of Copenhagen, Denmark University of Z¨ urich, Switzerland and University of Texas, Austin, USA

Introduction Several signallers and receivers sharing the same active signalling space constitute a communication network. This type of environment imposes additional selection pressures on both signallers and receivers other than those classically considered in signaller–receiver dyads. In this chapter, we shall discuss how communication networks influence the behaviour of a signaller and, more specifically, the effect of an audience (defined below) on signalling behaviour. An individual signaller has to cope with two main issues when signalling in a network: (a) it has to compete or cooperate with other signallers, and (b) it has to deal with the presence of several receivers. Signalling at the same time as other individuals poses a problem for the signaller: how does it ensure that its specific signal is detected by a receiver when other conspecifics are signalling? Signallers solve or minimize this problem by either cooperating or competing for the signal broadcast space. For example, in frog and insect choruses, individuals time their signals to avoid acoustic interference (e.g. alternating their calls) or compete for call order in the chorus (Gerhardt & Huber, 2002; Ch. 13). At the community level, different species with similar signals may broadcast their signals at different times of the day (Endler, 1992). The presence of several receivers presents two additional problems for the signaller. The first is how to direct the signal to a specific receiver. For example, bird song often has a range that encompasses several neighbouring territories. When a bird sings, the song could potentially reach all the neighbours in surrounding territories. During interactions with neighbours, individuals may need to direct the Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

63

P1: GFZ/... P2: JZZ/... 0521823617c04.xml CU1917B/McGregor

64

0 521 582361 7

April 10, 2005

17:49

R. J. Matos & I. Schlupp signal to a specific individual, for example because that neighbour starts to sing close to the territory boundary. McGregor & Peake (2000) discussed several ways in which songbirds can direct the signal to a specific rival neighbour or intruder. For example, matched counter-singing (Stoddard et al., 1992; Beecher et al., 1996) is a good candidate for directing the signal to a specific individual bird. The second issue that arises from the presence of several receivers, and one that this chapter covers in more detail, is how signallers communicate in the presence of additional receivers other than the primary target receiver. We will concentrate on conspecific receivers because heterospecific receivers, especially predators and parasites, have received considerable attention and are known to be important in shaping signals and signalling interactions (Bradbury & Vehrencamp, 1998; Chs. 2 and 8). The term audience has been used to describe conspecific receivers in the context of a communication network (McGregor & Peake, 2000; Doutrelant et al., 2001). In this chapter, we shall begin by discussing this term and its use in the context of communication networks. We shall then discuss how the presence of several receivers may affect signalling behaviour and the choice and evolution of signalling strategies.

Definitions of audience and audience effects Audiences

We define audiences as individuals that are present during, but do not take part in, signalling interactions between others. We distinguish two types of audience: evolutionary audiences and apparent audiences. Evolutionary audiences

By evolutionary audiences we mean individuals that were historically common in the environment of the signaller and that may have generated selection on the form and content of signalling behaviour. For example, it is widely accepted that bird song has a dual function, both as a signal to attract females and as a signal used in male–male competition (Berglund et al., 1996; Searcy & Nowicki, 2000). The evolution of this dual function has been widely discussed in the literature (e.g. Searcy & Nowicki, 2000). One hypothesis suggests that song first evolved as a male–female signal with males acting as eavesdroppers (see Ch. 12 for similar discussion on fiddler crabs). This eavesdropping pressure caused by male audiences may have induced new selective forces on the form and content of the signal, resulting in the appearance of a dual function signal. If this hypothesis is correct, then males have acted as an evolutionary audience in bird song evolution. An evolutionary audience does not need to be present or apparent to affect

P1: GFZ/... P2: JZZ/... 0521823617c04.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

17:49

Signallers and the social environment signalling behaviour at any instant in time, because selection has acted in the past (and presumably continues to act) on the signal (e.g. introducing or emphasizing features in the design of songs that males use in male–male competition). For more information on the effects and importance of evolutionary audiences, we refer the reader to Chs. 2 and 14. Apparent audiences

Apparent audiences are individuals that affect the behaviour of the signaller only when they are present and detected. For example, in the presence of females, interacting male Siamese fighting fishes Betta splendens decrease highly aggressive behaviours (attempted bites) and increase the intensity of conspicuous displays (tail beats and gill cover display) (Doutrelant et al., 2001). Unlike evolutionary audiences, the effects produced by this type of audience are triggered when the audience is present; males show no such effects on the different displays when the female is absent. In this chapter, we are mainly concerned with the study of apparent audiences, as their effects can be studied experimentally and, unlike studies of evolutionary audiences, they do not rely on historical inference. Audience effects

We define an audience effect as changes in the signalling behaviour during an interaction between individuals caused by the mere presence of an audience. Matos & McGregor (2002) found that male fighting fish engaged in visual signalling interactions changed their signalling (i.e. the visual displays directed towards the rival male) when a male audience was present. It is important to emphasize that the change in signalling behaviour occurred between the two individuals involved in the signalling interaction and not directly towards the audience. Whether the information content of signalling changes will depend on the balance of cost and benefit to the signallers (see below). This type of effect is specific to a communication network as it can only occur in situations where a minimum of three individuals is present: two individuals engaged in a signalling interaction and one individual making up the audience. Why audience?

Different authors have used different terms to designate extra potential receivers in a communication network, such as bystanders (Dugatkin, 2001), unintended receivers (Endler, 1993) or illegitimate receivers (Otte, 1974). Most of these terms, however, have been used in an interspecific context (with the exception of bystander) to describe predator detection of prey signalling behaviour (Otte, 1974; Endler, 1993; Bradbury & Vehrencamp, 1998; Ch. 2). Because we restrict our

65

P1: GFZ/... P2: JZZ/... 0521823617c04.xml CU1917B/McGregor

66

0 521 582361 7

April 10, 2005

17:49

R. J. Matos & I. Schlupp definition of an audience to conspecifics, we exclude predators or parasites responding to the signal (Ch. 2). We chose the term audience because it is more descriptive of the role of the individual during the signalling interaction in two ways. First, it implies that individuals are present but do not take part in the interaction, although they are clearly able to. Second, it implies that the individuals may pay attention to the signalling interaction and thus potentially extract information. We think that it is important to link the term audience to other network behaviours such eavesdropping in this way because the presence of eavesdroppers can impose costs and benefits on signallers and to link these costs–benefits to the information content of the interaction. For example, the finding that eavesdroppers behave more aggressively to individuals that behave as losers in an aggressive signalling interaction (e.g. Chs. 2 and 14) identifies an immediate cost of an audience on the losers. It is worth noting though that we do not have to show that individuals are able to extract information to cause an audience effect. For example, audiences may be costly just because there is a high risk of the audience disrupting the signalling interaction (e.g. intervention behaviour of semicaptive zebras Equus quagga; Schilder, 1990). By comparison, non-apparent eavesdroppers do not promote an audience effect because signallers are unaware of their presence. Other uses of audience and audience effect in the literature

The terms audience and audience effects have been used in the communication network literature to describe the effects on signalling interactions of the presence of additional potential receivers that do not take part in the interaction (Doutrelant et al., 2001; Matos & McGregor, 2002). However, these terms have also been used in other studies in animal communication. In the following paragraphs we shall talk about these studies and underline the differences between the two uses of the term audience. The first studies to use the terms audience and audience effects looked at the effect of the presence of a conspecific on the incidence of alarm and food calls in birds (Gyger et al., 1986; Marler et al., 1986; Gyger, 1990; Evans & Marler, 1994). These authors were interested in whether these calls were elicited by and directed to a specific class of individuals or audiences, namely conspecifics (e.g. conspecific versus predator; male versus female). In these studies, an audience is defined as any individual that is present in the same location as the subject (an apparent audience), and the audience effect is the change in signalling behaviour (e.g. increase in food call rate: Marler et al., 1986) caused by the presence of the audience. In both cases, the signal was assumed to be directed towards the audience; for example, Gyger et al. (1986) performed two experiments to investigate whether

P1: GFZ/... P2: JZZ/... 0521823617c04.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

17:49

Signallers and the social environment male cockerels Gallus domesticus modulated their alarm calls in the presence of an audience when a model of a predator was presented. The protocol of both experiments was the same; the birds were placed in a cage above which a model of a predator was ‘flown’. The audience was housed in another cage next to the male’s cage; both individuals could see the predator. In the first experiment, the audience was either their own mate or a female that was mated to another male, with an empty cage as a control. The second experiment was similar to the first one with the difference that instead of another male’s female the authors used an unfamiliar male. The authors found that in both experiments males increased the rate of alarm calls when a conspecific was present compared with when alone. No significant difference was found between the presence of the male’s mate compared with another male’s mate, or between the male’s mate compared with an unfamiliar male. The authors concluded from these results that the presence of a conspecific audience has an effect on alarm calling and that these calls may be primarily directed towards conspecifics and not towards the predator. Because there was no significant effect of the type of conspecific (own mate, other’s mate, unfamiliar male), one can rule out the hypothesis that the observed increase in call rate is a result of sharing the risk with the other prey (Gyger et al., 1986). There are two main differences between the use of the terms audience and audience effects in these studies and our own use. First, we restrict audience effects to the signalling interaction between the two individuals; the audience is not the primary receiver of the signals but acts as a potential non-targeted receiver. In the predator/food call studies, the distinction between the audience and a primary receiver of the signal is blurred as the target receiver is the audience (Fig. 4.1). The second difference is that our definition is specific to communication networks. In the predator/food call studies, this was not necessarily true; only two conspecifics were necessary to produce the audience effect: the signaller and the audience. For example, in a similar study to the one described above, Marler et al. (1986) showed that male cockerels increased their food calls in response to the presence of one hen; such a situation is a signaller–receiver dyad. We point out these differences in use of the terms to ensure that different phenomena are not confused by the use of a common term and suggest that the terms should be clearly defined when used.

Audience effects Relatively few studies have addressed directly the question of whether audience effects occur. In this section, we summarize these studies and discuss other systems in which audience effects appear to have an important influence on signalling behaviour.

67

P1: GFZ/... P2: JZZ/... 0521823617c04.xml CU1917B/McGregor

68

0 521 582361 7

April 10, 2005

17:49

R. J. Matos & I. Schlupp

Fig. 4.1. The audience effect in different types of study. (a) In predator/food call studies (Gyger, et al., 1986), the change in behaviour (dotted arrows) is triggered by the presence of the audience (the hen) and directed towards the audience. (b) In the audience effects described in this chapter, the change in behaviour (dotted arrows) is triggered by the presence of the audience (non-target receiver, the hen) and directed towards the target receiver (the other male).

Male–male aggressive signalling interactions

Individuals often use signals to compete for resources such as territories, food or mates. These displays are used to assess the opponents’ fighting ability and motivation (Huntingford & Turner, 1987; Bradbury & Vehrencamp, 1998). In a communication network, this information is available not only to the opponent but also to other individuals that are within signal range. This audience of nontargeted receivers may introduce extra costs or benefits to signallers; as explained above, some studies show that eavesdropping fish are more likely to initiate aggressive interactions with a loser than with a winner (Oliveira et al., 1998; Earley & Dugatkin, 2002; Chs. 2 and 5). If an audience has high costs or benefits to signallers, then signallers should adjust their behaviour towards the opponent in order to conceal or enhance information, respectively (McGregor & Peake, 2000). Siamese fighting fish

Siamese fighting fish often use signals to mediate competition over resources such as territories, food or mates, and such visual displays have been

P1: GFZ/... P2: JZZ/... 0521823617c04.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

17:49

Signallers and the social environment

(a)

(b)

Fig. 4.2. Representation of the experimental design used in Matos (2002) to study the effect of a male audience on male–male interactions in Betta splendens. (a) In the first 10 minutes, both males were allowed to interact in the absence of an audience. (b) In the second 10 minute period, either an audience or an empty tank was revealed (removal of the opaque partition) to the males. Ma and Mb are the interacting males; A is the audience tank; o.p. is an opaque partition; arrows represent the direction in which visual contact was possible.

used as a model system to address different questions related to communication networks (e.g. eavesdropping: Oliveira et al., 1998; McGregor et al., 2001). One of the first experiments to address specifically whether male Siamese fighting fish were affected by the presence of an audience during an aggressive interaction was performed by Matos (2002). Two males were allowed to interact through a clear partition (tank walls), and a third male (the audience) was placed at a small distance from these males (Fig. 4.2). This small distance prevented the audience from taking part in the interaction yet, at the same time, allowed both males to see the audience. Each trial of the experiment was divided into two 10 minute periods: in the first period the two individuals were allowed to interact without the audience being present; the second period started when an opaque partition that separated the audience from the two males was removed, allowing the males to see the audience while interacting. Previous studies have shown that one can predict the winner of a fight between two male fighting fish from display difference at the beginning of the interaction (Simpson, 1968). In this experiment, the winner of the signalling interaction was defined as the individual that displayed most during the first 10 minutes of the interaction (the other male was the loser). It is important to note that the barriers between males prevented actual fighting and none of the interactions reached an outcome (e.g. displaying submissive colouration). No displays directed towards the audience were observed. Matos (2002) found that ‘winners’ did not change their signalling behaviour in the presence of an audience. In

69

P1: GFZ/... P2: JZZ/... 0521823617c04.xml CU1917B/McGregor

70

0 521 582361 7

April 10, 2005

17:49

R. J. Matos & I. Schlupp contrast, when an audience was present ‘losers’ reduced the time they spent in gill cover display (a purely visual display) and the time spent near the opponent compared with when there was no audience. However, there was no significant change in the more aggressive displays that had both tactile and visual components (i.e. attempted bites and tail beats). This change in behaviour may be viewed as an attempt by the loser to restrict the information available to the eavesdropper while at the same time providing adequate information for assessment by the opponent. Another hypothesis is that by reducing the less-aggressive displays whilst maintaining the more aggressive forms, ‘losers’ may seem more aggressive to the audience. Thus even though the audience may have seen that individual lose, it would be more reluctant to interact with it because of its aggressiveness (‘good loser’ hypothesis: Peake & McGregor, 2004). This study (Matos, 2002) suggests that there is an audience effect when a male audience is present during male–male interactions and that the presence of the audience can be more costly for the individual that is losing the interaction than for the winner. The finding that the audience effects in this situation involved a change of signalling behaviour by the loser fits both observations that losers are more rapidly approached by males that saw them lose (Siamese fighting fish: Oliveira et al., 1998; McGregor et al., 2001; swordtail fish Xiphophorus helleri: Earley & Dugatkin, 2002) and that this effect disappears in combats where both individuals escalated (Earley & Dugatkin, 2002). In an earlier study, Doutrelant et al. (2001) also found that female audiences affected male–male B. splendens aggressive displays. In this experiment, a female audience was presented to a pair of males that interacted through a clear partition. The effect of the presence of an audience was then compared with a treatment where males were allowed to interact with no audience present. Males increased the amount of conspicuous displays (e.g. tail beats and time with gill cover erect) and decreased the more aggressive displays (e.g. attempted bites) towards opponents when a female was present. The authors interpreted this result as males trying to compromise between having to interact with an opponent and at the same time provide information to the audience by using more conspicuous displays, which are more often used in both aggressive and courtship contexts. Doutrelant et al. (2001) also performed a second experiment to examine whether male audiences affected signalling interactions but did not find an audience effect (except for a tendency for males to spend less time near the opponent). However, the result of these two experiments cannot be compared directly because of differences in the experimental design and procedure (i.e. the audience was closer to the males and the males were pre-exposed to the audience in the female experiment, while in the male experiment the audiences were further away and there was no

P1: GFZ/... P2: JZZ/... 0521823617c04.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

17:49

Signallers and the social environment

(a)

(b)

Fig. 4.3. Schematic representation of the experimental design used in both Matos & McGregor (2002) and Matos et al. (2003). (a) In a five minute pre-exposure period, both males could see the audience tank. (b) In the 10 minute interaction period, the opaque partition was removed and both males were allowed to interact with each other in front of or in the absence of the audience. Ma and Mb are the interacting males; A is the audience tank; o.p. is an opaque partition; arrows represent the direction in which visual contact was possible.

pre-exposure period). Both distance and pre-exposure to another individual have been shown to have a strong effect on male aggressive display (Bronstein, 1989; Halperin et al., 1998; also see below). In a more recent experiment, Matos & McGregor (2002) looked directly at the effect of the sex of the audience. Three different types of audience were used: male, female B. splendens and female Xiphophorus spp. (to control for responses not specific to conspecifics). A control with no audience present was also used. The design and procedure of the experiment was similar to that in Matos (2002), except that the males were first pre-exposed to the audience and then were allowed to see and interact with the opponent (Fig. 4.3). The audience was visible for the entire trial. No distinction was made between winners and losers as data were only collected from one of the individuals involved in the interaction. No differences were observed between the female Xiphophorus spp. treatment and no audience; therefore the Xiphophorus spp. treatment was used as the control. Males behaved more aggressively (i.e. attempted more bites and spent less time near the opponent) when a male audience was present than with a female audience (Matos & McGregor, 2002). To explain this difference, the authors suggested that the presence of a female might confront the males with a trade-off between expelling their male opponent and not driving away a potential mate. Males of this species often bite when courting a female and highly aggressive males may cause females to flee because of the high risk of injury (Bronstein, 1984). The results of these

71

P1: GFZ/... P2: JZZ/... 0521823617c04.xml CU1917B/McGregor

72

0 521 582361 7

April 10, 2005

17:49

R. J. Matos & I. Schlupp experiments suggest that the sex of the audience is important in determining how males should behave during aggressive signalling interactions. Field crickets

Tachon et al. (1999) studied male–male competition for resources in the field cricket Gryllus bimaculatus. They tested whether the presence of a female influences the aggressive behaviour between males. In each test, a group of five males in an arena under three different treatments was observed. Besides the two obvious treatments, presence and absence of females, they used a third condition where a paper impregnated with female scent was introduced into the arena. Previous studies had shown that this scent elicited behavioural responses from males of this species (Otte & Cade, 1976; Hardy & Shaw, 1983). Tachon et al. (1999) found that males increased their level of aggressive displays (e.g. aggressive stridulation and mandible flaring) towards other males in the treatment where the females were present. Interestingly, there was no evidence that the female scent produced the same effect as the actual presence of a female. Female scent alone in this system may be a poor predictor of female presence and the cost of escalating increases when there is a high probability that the female is not present. However, in this example, it is not clear what effect direct female– male interactions had on male–male competition, as opposed to the effect of the mere presence of the female. Further studies are needed to attempt to distinguish these effects and thus to confirm whether this is an example of an audience effect. Parental behaviour

Male parental care is common in many species. If there is a direct link between the care provided to the young and their survival until reproductive age, it might be of advantage for the females to choose a good father as a potential mate. One way of assessing paternal care is to observe male interactions with young (e.g. affiliate signalling behaviour). If females do choose a good father for their future mate, then it should be to the advantage of the male to try to perform as a better ‘parent’ when a female is present. Vervet monkeys

Vervet monkeys Cercopithecus aethiops have a complex social system where individuals influence their own or other group members’ dominance rank by socializing with individuals of different rank. In such a system, female mate choice or preference to associate with a male can influence the male’s future position in the hierarchy (Ch. 25). Interactions between males and infant are quite common and males often form strong protective relationships with the females and their young. These relationships may reduce the harassment that females and infants

P1: GFZ/... P2: JZZ/... 0521823617c04.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

17:49

Signallers and the social environment receive from other group members. Therefore, females may prefer to associate with males that perform more affiliatively towards their infant. Hector et al. (1989) investigated whether male vervet monkeys changed their interaction with an infant in the presence versus ‘absence’ of the mother. In this experiment, the females were placed (a) behind a one-way mirror, where they could see both male and infant but not vice versa; (b) behind a Plexiglas partition, where male, female and infant could see each other; and (c) behind a metal partition, where the female could not see the dyad and the male and infant could not see the female. The results of this experiment showed that males are sensitive to the presence of the mother and engaged in more affiliative and less-antagonistic behaviour toward the infant when the male was able to see the mother. However, it is not clear whether the effect is simply caused by the presence of the female or occurs because the females could still potentially signal to the dyad through the Plexiglas, affecting the behaviour of both infant and male. A further treatment would be needed to address this question, where the female is placed behind a one-way mirror and the dyad can see the female but not vice versa. The authors further studied if females varied their behaviour towards males that they saw performing more affiliative behaviours towards their infants and found that females tolerated the males more and also performed more affiliative behaviours towards them. In spite of the lack of an appropriate control, this study showed that potentially individuals may adjust their behaviour when an audience is present and that there are direct consequences to the individual. Budgerigars

Female birds may assess male parental care behaviour by the male’s extrapair behaviour during the period prior to egg laying. In species with obligate biparental care, males that provide more care to the young should be preferred as a mate, as less-committed males increase the female’scosts of feeding and spending more time with the young. Extra-pair activity by the male (e.g. displaying to another female) may provide information to the female on the male’s attentiveness towards the female and the nest. Budgerigars Melopsittacus undulatus are socially monogamous birds where both members of the pair provide parental care. The males of this species provide most of the food to the nest, both at the start of the nesting period and through brooding. As a consequence, male commitment to the female and brood is very important to the female and survival of the brood, and females may use cues of male commitment when they are choosing a potential mate. Baltz & Clark (1994) investigated whether male budgerigars were less likely to court another female when their own mate was present. In other words, they tested whether there is an effect of an audience (their mate) on the male’s extra-pair behaviour. The study was conducted on

73

P1: GFZ/... P2: JZZ/... 0521823617c04.xml CU1917B/McGregor

74

0 521 582361 7

April 10, 2005

17:49

R. J. Matos & I. Schlupp a captive population housed in an outdoor aviary. Nestboxes were provided, simulating the nests in natural cavities observed in the wild. The authors assumed that the females lost visual contact with the male when inside the nestbox. The behaviour of each male and its mate was recorded in the periods where the female was inside (no audience) and outside (audience) the nest. Males significantly increased extra-pair courtship behaviour when out of view of the female (i.e. when the female was inside the nestbox) relative to when the female was in view. However, the results of this experiment can also be explained by an alternative hypothesis. Males may reduce the time courting other females because with their mate outside the nest they are more vulnerable to extra-pair courtship and copulations by other males in the flock. Therefore, the reduction of courtship may be a result of mate guarding (Baltz & Clark, 1994). In another study, Baltz & Clark (1997) showed that the necessity for mate guarding did not change the males’ response to the extra-pair female. The authors used the same experimental design as before but this time the mate was separated from the rest of the flock in both treatments. This procedure prevented other males from interacting with the female (subject’s mate) and thus reduced the necessity for mate guarding. Once again, males reduced courtship behaviour towards extra-pair females when their mate was visible. Although this study suggests that there maybe an audience effect, we consider it poor evidence for audience effects as we define them in this chapter. The main problem with the experimental design of both studies is that the audience effect is not caused by the mere presence of the audience, the male–female pair are only separated visually by an opaque partition, and, as the authors state, both individuals could still contact each other through calls even when they could not see each other. We suggest that further studies would be required to confirm the presence of an audience effect in such system. Human behaviour

Social psychologists have long recognized that audiences have an important effect on human behaviour (e.g. Zajonc, 1965; Blumstein, 1973; Felson, 1982; Ch. 19). These effects extend from a change in the performance of simple motor tasks, when compared with apparently ‘non-social’ contexts (Zajonc, 1965), to changes in more complex forms of social behaviour such as interpersonal strategies used during social interactions (Blumstein, 1973). One interesting area of study with regard to audience effects in humans is impression management theory. This theory focuses on the principle that a person is aware of being characterized or typified by others when performing a behaviour and responds by trying to make these characterizations favourable. As a consequence, most human behaviour is designed to obtain ‘favourable’ reactions from an audience (Felson, 1978, 1982). For example, Felson (1982) found an effect of third-party

P1: GFZ/... P2: JZZ/... 0521823617c04.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

17:49

Signallers and the social environment presence on aggressive interactions between humans. The study was based on interviews with patients with previous mental health problems, with ex-criminals and with a sample of the general population. All groups answered a questionnaire asking them to describe in detail four aggressive incidents. The replies showed that the outcome of an interaction between individuals of the same sex was more severe when an audience was present (when allowing for third-party instigation or mediation of the fight). There was a higher probability that individuals would escalate from verbal insults to actual physical contact. However, the authors also found that the same was not true in conflicts between the sexes; the cause of such a difference may be that the audience is more likely to disapprove of severe aggression in between-sex conflicts (Felson, 1982). The general idea that individuals may try to manipulate their characterization by others has recently been used to explain altruistic behaviour in humans and non-human animals (Zahavi & Zahavi, 1997; Wedekind & Milinski, 2000; Milinski et al., 2001; Bshary, 2002). This idea is discussed by Bshary & D’Souza in Ch. 22.

Priming: a mechanism of audience effects or a functional alternative? In the experiments discussed above showing that male Siamese fighting fish behaved more aggressively towards an opponent when a male audience was present (Matos & McGregor, 2002), the trial procedure allowed males to see the audience before they started interacting. This procedure was used to ensure that the males were aware of the presence of the audience during the interaction. In a further series of experiments, Matos et al. (2003) found that the presence of an audience before an interaction affected how male B. splendens behaved during the interaction. Using a similar design to that described by Matos & McGregor (2002), the authors divided each trial into two continuous periods: a pre-exposure period (when males could either see an empty tank or a tank containing an audience) and an interaction period (when both males where allowed to interact with each other in the presence or absence of an audience). In the first experiment, four different treatments where used in which the audience was (1) present in the pre-exposure period, (2) present during the interaction period, (3) present in both periods or (4) absent in both periods. The authors then separated the behaviours overt aggression (i.e. attempted bites and latency to first bite) and a display score (combined measure of the other displays, i.e. time spent flaring the gill cover, number of tail beats and time spent near the opponent); for details on the method see Matos et al. (2003). Overall, males behaved more aggressively (i.e. shorter latency to attempt to bite the opponent) during the interaction in the treatments where the males were preexposed to the audience (treatments 1 and 3). This effect is similar to aggressive priming. The presence of the audience before the interaction may have increased

75

P1: GFZ/... P2: JZZ/... 0521823617c04.xml CU1917B/McGregor

76

0 521 582361 7

April 10, 2005

17:49

R. J. Matos & I. Schlupp

(a)

(b)

(c)

(d)

Fig. 4.4. Schematic representation of the second experimental design used in Matos et al. (2003). (a–c) The five minute pre-exposure period when both males were pre-exposed to an empty tank (no pre-exposure) (a); both males were pre-exposed to an audience (b); and only one of the males (Ma) was pre-exposed to the audience (c). (d) The 10 minute interaction period following all treatments, where both males were allowed to interact in front of an audience. Ma and Mb are the interacting males; A is the audience; thick lines between the tanks represent opaque partitions; arrows represent the direction in which visual contact was possible.

the motivation to behave aggressively. As a result, individuals escalated more rapidly into more aggressive forms of behaviour when they interacted with the opponent. The authors also found that priming effects overrode any effect of presenting the audience only during the interaction. The levels of aggression between the two treatments where males were pre-exposed (treatments 1 and 3) were similar, independent of audience presence during the interaction period, while much lower levels of aggression were seen in treatments 2 and 4. In fact, there was no significant difference between the treatments with the audience absent in both periods (treatment 4) and with the audience present during the interaction (treatment 2). These results may suggest that audiences do not affect male–male fighting fish interactions, as the audience affected only treatments with pre-exposure. In this respect the results matched those of Doutrelant et al. (2001), in which male audiences did not have an effect on male–male signalling interactions (see above). However, we should also note that in both studies the authors did not look at losers and winners separately. Matos et al. (2003) performed a second experiment to look at the interaction between audience effects and pre-exposure to the audiences; the design allowed independent pre-expose of the two opponent males (Fig. 4.4). As in the first

P1: GFZ/... P2: JZZ/... 0521823617c04.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

17:49

Signallers and the social environment experiment, the trials were divided in two periods: five minutes of pre-exposure and a 10 minute period in which the two opponents were allowed to interact. There were three treatments in the pre-exposure period: both males pre-exposed to an empty tank (no pre-exposure; Fig. 4.4a), both males pre-exposed to the audience (symmetric pre-exposure; Fig. 4.4b), and one of the opponents pre-exposed to the audience while the other male was pre-exposed to an empty tank (asymmetric pre-exposure; Fig. 4.4c). The audience was always present in the interaction period (Fig. 4.4d). The results confirmed that pre-exposed males tend to behave more aggressively (higher display scores and overt aggression); both the no pre-exposure and the symmetric treatments showed the same tendencies. In the asymmetric treatment, pre-exposed males also tended to display more than the ones not preexposed with one exception: non-pre-exposed males matched the number of attempted bites of the pre-exposed males. A possible explanation is that it may be costly for individuals not to retaliate when its opponent escalates, because of the high risk of injury, especially in a confined space such as the experimental tanks (Maan et al., 2001). However, this cost may be enhanced by the presence of the audience. By matching the opponent in more aggressive behaviour, males may be either decreasing the ability of an audience to discriminate loser from winner or manipulating the information to seem more aggressive. These results support the previously discussed idea that males, particularly losers, may gain by performing more aggressively during an interaction in the presence of an audience, as it may decrease the chances of future harassment by that individual (Earley & Dugatkin, 2002; Matos et al., 2003). These two studies taken together support the idea that previous studies on audience effects (i.e. Doutrelant et al., 2001; Matos & McGregor, 2002) have underestimated the effect of pre-exposure on male aggression. Nevertheless, they also suggest that the social environment (i.e. audiences) is important in determining the dynamics of signalling interaction. Previous studies have shown that priming is an important mechanism mediating aggressive interactions (e.g. Potegal & Popken, 1984; Bronstein, 1989; Halperin et al., 1992) as it affects the individual’s aggressive motivation. For example, priming may decrease the time to initiate aggression or increase the attack behaviour of individuals (e.g. Potegal & ten Brink, 1984; Halperin et al., 1998). However, the effect on the outcome of interactions is not always clear. It seems that priming may have a more pronounced effect during the initial stage of the fight, either causing the individual to display more actively at the beginning of the interaction or to escalate and initiate aggression more quickly (Potegal & Popken, 1984; Bronstein, 1989; Halperin et al., 1998). In several species, individuals that display more intensively and escalate earlier during an interaction usually gain a

77

P1: GFZ/... P2: JZZ/... 0521823617c04.xml CU1917B/McGregor

78

0 521 582361 7

April 10, 2005

17:49

R. J. Matos & I. Schlupp competitive advantage over their opponents (Huntingford & Turner, 1987). In such a case, priming may produce a positive effect as it increases the probability that the individual will win the fight. In some cases, however, priming can have a negative effect, male B. splendens that have been isolated and then primed with a conspecific image behaved more aggressively towards their opponents but lost most of the interactions (Halperin et al., 1998). These individuals could have been manipulated by priming into aggressive levels that they were not able to sustain during the entire fight, causing them to tire faster than the opponents and subsequently lose the interaction. We conclude that priming may have an important impact on the outcome of the interaction, but whether this impact is positive or negative may depend on whether the initial stages of the interaction determine the outcome and on the length of the interaction. One potential mechanism behind such aggressive priming is the production of hormones caused by the presentation of a social stimulus. In a recent study, Oliveira et al. (2000) showed that watching a fight raises the androgen levels of adult male cichlid fish Oreochromis mossambicus. Priming may involve a similar mechanism, and the facilitation of aggressive behaviour through pre-exposure may be caused by an increase in androgen levels initiated by the pre-exposure to the audience. Oliveira et al. (2000) suggested that these hormones mediate changes in the perceptual abilities and readiness to interact of males, which, in turn, would enhance their success in social interactions. Further studies are needed to comprehend fully the relation between the adaptive value of priming and the presence of an audience. Advances in the understanding of the effects of the social environment on the neuroendocrinological system may be an important contribution in this area (Ch. 21).

Summary and future directions One important question in the general context of communication networks is how narrowly or broadly we wish to define the social context of signalling. Recent studies have shown that mate preferences can be altered by viewing sexual interactions (Westneat et al., 2000). In several species, seeing a male mate enhances this male’s attractiveness to females (Dugatkin, 1992; Ch. 5). Such choosing females would be eavesdropping on the signalling–mating interaction of two other individuals and responding accordingly. The same reasoning might apply for other interactions as well. It appears that most dyadic interactions are actually embedded in a social context or network. This raises the question of how common the well-studied dyadic interactions actually are, as these studies have only considered them in a social void. This situation might be more of an exception than the rule. In this context, more knowledge on sensory ecology and

P1: GFZ/... P2: JZZ/... 0521823617c04.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

17:49

Signallers and the social environment especially the role of private channels would be very helpful. Communication via private channels uses sensory channels not available to the audience. This has been documented for swordtails (Xiphophorus spp.): males signal in the ultraviolet, a part of the spectrum that cannot be detected by a predator, the Mexican tetra Astyanax mexicanus (Cummings et al., 2003). True dyadic interactions may be brief and limited to signals transmitted in close contact. A potential example might be nipping in poeciliid fishes; here, males nibble a female’s genital region and chemical signals are transmitted (Parzefall, 1973). Such signals are not available to any other individual, although the male’s response to the signal might be (Parzefall, 1973). Another aspect to consider is that many social interactions relevant to several aspects of an individual’s life may happen simultaneously and influence each other. Any given individual will have to include this into its signalling decisions. For instance, a singing bird may simultaneously be faced with the problems of attracting a female, discouraging a neighbour from entering its territory and avoiding predators. This leads to a more complicated network of social interactions, the components of which may influence each other to shape a ‘social interaction network’. Our singing bird example also illustrates that each context alone would select for a different signal or signalling strategy. Signals have to be effective enough to transmit accurate information to target receivers but private enough to prevent this information from being detected by ‘unwanted’ untargeted receivers. Any signal that is under such conflicting demands will be a compromise, depending on the associated costs and benefits. Only recently has formal modelling been used to address this problem (Johnstone, 2001; Ch. 26). We have attempted to show that the presence and type of audience can have important effects on the signalling strategies of individuals. The nature of information and the extent to which it is broadcast may depend on the type of audience and on the role of each signaller during an interaction. Audiences may also influence the evolution of new types of signal. During signal evolution, different pressures may arise in signal design depending on whether it is specialized to advertize or privatize information (e.g. ‘normal’ song versus quiet song: Dabelsteen et al., 1998; Ch. 3). Audience effects may be closely linked with mechanisms such as priming effects, which may influence motivation of signallers and consequently their signalling strategy. In natural systems, the social environment affects how animals make behavioural decisions. Individuals can use signalling interactions between others as a source of information; this can in many ways have important consequences for the fitness of individuals. In order to improve our understanding of the evolution of signals and signalling strategies, we must take into account the individuals’ social environment and the costs and benefits associated with the presence of audiences and eavesdroppers.

79

P1: GFZ/... P2: JZZ/... 0521823617c04.xml CU1917B/McGregor

80

0 521 582361 7

April 10, 2005

17:49

R. J. Matos & I. Schlupp Acknowledgements We would like to express our sincere thanks to Peter McGregor for providing us with the opportunity to write this chapter and to Tom Peake, Ryan Earley, Denise Pope, Giuliano Matessi and Andrew Terry for their valuable comments and discussion of the manuscript. We thank the ˜o para a Ciˆencia e Tecnologia (Portugal) for funding R. M., whose Ph.D. provided data and Fundac¸a ideas included in this chapter. I. S. was supported by a Heisenberg Fellowship of the Deutsche Forschungogemeinschaft.

References Baltz, A. P. & Clark, A. B. 1994. Limited evidence for an audience effect in budgerigars, Melopsittacus undulatus. Animal Behaviour, 47, 460–462. Baltz, A. P. & Clark, A. B. 1997. Extra-pair courtship behaviour of male budgerigars and the effect of an audience. Animal Behaviour, 53, 1017–1024. Beecher, M. D., Stoddard, P. K., Campbell, S. E. & Horning, C. L. 1996. Repertoire matching between neighboring song sparrows. Animal Behaviour, 51, 917–923. Berglund, A., Bisazza, A. & Pilastro, A. 1996. Armaments and ornaments: an evolutionary explanation of traits of dual utility. Biological Journal of the Linnean Society, 58, 385–399. Blumstein, P. W. 1973. Audience, Machiavellianism, and tactics of identity bargaining. Sociometry, 36, 346–365. Bradbury, J. W. & Vehrencamp, S. L. 1998. Principles of Animal Communication. Sunderland, MA: Sinauer. Bronstein, P. M. 1984. Agonistic and reproductive interactions in Betta splendens. Journal of Comparative Psychology, 98, 421–431. Bronstein, P. M. 1989. The priming and retention of agonistic motivation in male Siamese fighting fish, Betta splendens. Animal Behaviour, 37, 165–166. Bshary, R. 2002. Biting cleaner fish use altruism to deceive image-scoring client reef fish. Proceedings of the Royal Society of London, Series B, 269, 2087–2093. Cummings, M. E, Rosenthal, G. G. & Ryan, M. J. 2003. A private ultraviolet channel in visual communication. Proceedings of the Royal Society of London, Series B, 270, 897–904. Dabelsteen, T., McGregor, P. K., Lampe, H. M., Langmore N. E. & Holland J. 1998. Quiet song in song birds: an overlooked phenomenon. Bioacoustics, 9, 89–105. Doutrelant, C., McGregor, P. K. & Oliveira, R. F. 2001. The effect of an audience on intra-sexual communication in male Siamese fighting fish, Betta splendens. Behavioral Ecology, 12, 283–286. Dugatkin, L. A. 1992. Sexual selection and imitation females copy the mate choice of others. American Naturalist, 139, 1384–1389. Dugatkin, L. A. 2001. Bystander effects and the structure of dominance hierarchies. Behavioral Ecology, 12, 348–352.

P1: GFZ/... P2: JZZ/... 0521823617c04.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

17:49

Signallers and the social environment Earley, R. L. & Dugatkin, L. A. 2002. Eavesdropping on visual cues in green swordtail (Xiphophorus helleri) fights: a case for networking. Proceedings of the Royal Society of London, Series B, 269, 943–952. Endler, J. A. 1992. Signals, signal conditions, and the direction of evolution. American Naturalist, 139, S125–S153. Endler, J. A. 1993. Some general comments on the evolution and design of animal communication systems. Philosophical Transactions of the Royal Society of London, Series B, 350, 215–225. Evans, C. S. & Marler, P. 1994. Food calling and audience effects in male chickens, Gallus gallus: their relationship to food availability, courtship and social facilitation. Animal Behaviour, 47, 1159–1170. Felson, R. B. 1978. Aggression as impression management. Social Psychology, 41, 94–105. Felson, R. B. 1982. Impression management and the escalation of aggression and violence. Social Psychology Quarterly, 45, 245–254. Gerhardt, H. C. & Huber, F. 2002. Acoustic Communication in Insects and Anurans: Common Problems and Diverse Solutions. Chicago, IL: University of Chicago Press. Gyger, M. 1990. Audience effects on alarm calling. Ethology Ecology and Evolution, 2, 227–232. Gyger, M., Karakashian, S. J. & Marler, P. 1986. Avian alarm calling: is there an audience effect? Animal Behaviour, 34, 1570–1572. Halperin, J. R. P., Dunham, D. W. & Ye, S. 1992. Social isolation increases social display after priming in Betta splendens but decreases aggressive readiness. Behavioural Processes, 28, 13–32. Halperin, J. R. P., Giri, T., Elliott, J. & Dunham, D. W. 1998. Consequences of hyper-aggressiveness in Siamese fighting fish: cheaters seldom prospered. Animal Behaviour, 55, 87–96. Hardy, T. N. & Shaw, K. C. 1983. The role of chemo-reception in sex recognition by male crickets: Acheta domesticus and Teleogryllus oceanicus. Physiological Entomology, 8, 151–166. Hector, A. C. K., Seyfarth, R. M. & Raleigh, M. J. 1989. Male parental care, female choice and the effect of an audience in vervet monkeys. Animal Behaviour, 38, 262–271. Huntingford, F. A. Turner, A. K. 1987. Animal Conflict. London: Chapman & Hall. Johnstone, A. R. 2001. Eavesdropping and animal conflict. Proceedings of the National Academy of Sciences, USA, 98, 9177–9180. Maan, M. E., Groothuis T. G. G. & Wittenberg, J. 2001. Escalated fighting despite predictors of conflict outcome: solving the paradox in a South American cichlid fish. Animal Behaviour, 62, 623–634. 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? Animal Behaviour, 34, 194–198. Matos, R. J. 2002. Social influences on signalling: audience effects and communication networks. Ph.D. Thesis, University of Copenhagen.

81

P1: GFZ/... P2: JZZ/... 0521823617c04.xml CU1917B/McGregor

82

0 521 582361 7

April 10, 2005

17:49

R. J. Matos & I. Schlupp Matos, R. J. & McGregor, P. K. 2002. The effect of the sex of an audience on male–male displays of Siamese fighting fish (Betta splendens). Behaviour, 139, 1211–1221. Matos, R. J., Peake, T. M. & McGregor, P. K. 2003. Timing of presentation of an audience: aggressive priming and audience effects in male Siamese fighting fish (Betta splendens). Behavioural Processes, 63, 53–61. McGregor, P. K. & Peake, T. M. 2000. Communication networks: social environments for receiving and signalling behaviour. Acta Ethologica, 2, 71–81. McGregor, P. K., Peake, T. M. & Lampe, H. M. 2001. Fighting fish Betta splendens extract relative information from apparent interactions: what happens when what you see isn’t what you get. Animal Behaviour, 62, 1059–1065. Milinski, M., Semmann, D., Bakker, T. C. M. & Krambeck, H.-J. 2001. Cooperation through indirect reciprocity: image scoring or standing strategy? Proceedings of the Royal Society of London, Series B, 268, 2495–2501. Oliveira, R. F., McGregor, P. K. & Latruffe, C. 1998. Know thine enemy: fighting fish gather information from observing conspecific interactions. Proceedings of the Royal Society of London, Series B, 265, 1045–1049. Oliveira, R. F., Lopes, M., Carneiro, L. A. & Can´ ario, A. V. M. 2000. Watching fights raises fish hormone levels. Nature, 409, 475–475. Otte, D. 1974. Effects and functions in the evolution of signalling systems. Annual Review of Ecology and Systematics, 5, 385–417. Otte, D. & Cade, W. 1976. On the role of olfaction in sexual and interspecies recognition in crickets (Acheta and Gryllus). Animal Behaviour, 24, 1–6. Parzefall, J. 1973. Attraction and sexual cycle of poeciliids. In: Genetics and Mutagenesis of Fish, ed. J. H. Schr¨ oder. Berlin: Springer, pp. 177–183. Peake, T. M. & McGregor, P. K. 2004. Information and aggression in fishes. Learning and Behavior, in press. Potegal, M. & ten Brink, L. 1984. The behavior of attack-primed and attack satiated female golden hamsters, Mesocricetus auratus. Journal of Comparative Psychology, 98, 66–75. Potegal, M. & Popken, J. 1984. The time course of attack priming effects in female hamsters. Behavioural Processes, 11, 199–208. Schilder, M. B. J. 1990. Interventions in a herd of semi-captive plains zebras. Behaviour, 112, 53–83. Searcy, W. S. & Nowicki, S. 2000. Male–male competition and female choice in the evolution of vocal signalling In: Animal Signals: Signalling and Signal Design in Animal Communication, ed. Y. Espmark, T. Amundsen & G. Roseqvist. Trondheim: Tapir Academic Press, pp. 301–315. Simpson, M. J. A. 1968. The display of the Siamese fighting fish Betta splendens. Animal Behaviour Monographs, 1, 1–73. Stoddard, P. K., Beecher, M. D., Campbell, S. E. & Horning, C. L. 1992. Song-type matching in the song sparrow. Canadian Journal of Zoology, 70, 1440–1444. Tachon, G., Murray, A.-M., Gray, D. A. & Cade, W. H. 1999. Agonistic displays and the benefits of fighting in the field cricket, Gryllus bimaculatus. Journal of Insect Behaviour, 12, 533–543.

P1: GFZ/... P2: JZZ/... 0521823617c04.xml CU1917B/McGregor

0 521 582361 7

April 10, 2005

17:49

Signallers and the social environment Wedekind, C. & Milinski, M. 2000. Cooperation through image scoring in humans. Science, 288, 850–852. Westneat, D. F., Walters, A., McCarthy, T. M., Hatch, M. I. & Hein, W. K. 2000. Alternative mechanisms of nonindependent mate choice. Animal Behaviour, 59, 467–476. Zahavi, A. & Zahavi, A. 1997. The Handicap Principle. Oxford: Oxford University Press. Zajonc, R. B. 1965. Social facilitation. Science, 149, 269–274.

83

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:24

5

Fighting, mating and networking: pillars of poeciliid sociality ryan l. earley1 & lee al an dugatkin2 1 2

Georgia State University, Atlanta, USA University of Louisville, USA

We are both spectators and actors in this great drama of existence Niels Bohr

Introduction Poeciliid fishes such as green swordtails Xiphophorus helleri and guppies Poecilia reticulata aggregate in social groups called shoals. In addition to reducing predation risk and increasing foraging efficiency (e.g. Magurran & Pitcher, 1987; Ranta & Juvonen, 1993), fish shoals promote the transfer of social information within the group. For instance, information about foraging routes is transmitted from trained individuals to naive fish in guppy shoals (Laland & Williams, 1997; Swaney et al., 2001; Brown & Laland, 2002). The type of information transfer demonstrated in the social learning and foraging literature involves the transmission of signals from one or more individuals to the remaining group members. Investigations of social foraging and anti-predator behaviour have demonstrated that poeciliids attend to a variety of cues emitted by both conspecifics and heterospecifics (e.g. predators: Brown & Godin, 1999; Mirza et al., 2001; Brosnan et al., 2003). Although social learning and anti-predator responses constitute important aspects of group living in poeciliids, this chapter focuses more on how individuals gain information from observing interactions that occur in their social environment. Indeed, the concept of communication networks was founded on the premise that the information exchanged during social interactions (e.g. agonistic or courtship displays) may be available not only to the participants but also to bystanders within signal detection range (McGregor, 1993; McGregor et al., 2000). Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

84

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:24

Poeciliid sociality: fighting, mating and networking Throughout this chapter, we make a clear distinction between bystanders and eavesdroppers, though the two have been used synonymously in the past (Ch. 2). Bystanders are any individuals within detection range of signalling interactions while eavesdroppers represent a subset of bystanders that extract information from these interactions. The primary aim of this chapter is to examine how social eavesdropping – extracting information from signalling interactions between others (Ch. 2) – influences aggressive contest behaviour and female mate choice in male X. helleri and female P. reticulata, respectively. Therefore, we focus on how eavesdropping affects the subsequent behaviour of poeciliid bystanders rather than on how the behaviour of participants in an interaction is modified in the presence of an audience (Ch. 4). Swordtails and guppies are well suited to investigations of networking phenomena because they are highly social, exhibit stereotypical agonistic and courtship displays and are especially responsive to a host of stimuli (e.g. visual, chemical) in their social surroundings.

Social eavesdropping and contest behaviour Although sociality confers fitness-related benefits to individuals within the group, competition for social status and limiting resources (e.g. food, mates) often increases with group size (Pulliam & Caraco, 1984). In many animals, overt aggressive interactions are most common during hierarchy or territory establishment. Among fishes, rank-order fights involve a series of gradually escalating displays that convey information about strength, size or willingness to persist in the encounter. The intensity and/or duration of aggressive contests depend largely on differences in fighting ability between adversaries. When substantial differences in fighting ability exist, the interaction may terminate following a bout of noncontact displays. Interactions between well-matched opponents, however, may intensify to physical combat, where behavioural tactics such as mouthwrestling are used to settle the dispute (Enquist et al., 1990). Escalated contests often yield unambiguous dominant–subordinate relationships among closely matched competitors but the costs associated with such interactions can also be quite high. For instance, Neat et al. (1998) revealed that prolonged fights result in the accumulation of anaerobic metabolites and depletion of sugar reserves in the muscle tissue of cichlid fish Tilapia zillii. Other potential contest costs include physical injury, increased susceptibility to predation, lost mating or foraging opportunities and increased stress hormone levels, which may impede future reproductive activity (Haller, 1995; Jakobsson et al., 1995; Schuett, 1997; Halperin et al., 1998; Neat et al., 1998). Most theoretical and empirical work on aggressive contest behaviour assumes that information about fighting ability is available only via direct interactions

85

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

86

0 521 582361 7

April 11, 2005

11:24

R. L. Earley & L. A. Dugatkin (Enquist & Leimar, 1983; Payne & Pagel, 1997; Mesterton-Gibbons & Adams, 1998; Payne, 1998). Although this may be the case for solitary species, the social environment of group-living animals is ripe with opportunities for indirect assessment of fighting ability via eavesdropping. Since observing fights may provide information about fighting ability without the associated costs of physical combat, eavesdropping should be an advantageous assessment strategy when both contest costs and the opportunities for watching interactions are high. In this section, we review a series of experiments that elucidate how eavesdropping modulates the agonistic behaviour of male X. helleri. Specifically, we focus on how a bystander’s behaviour changes after observing fights and the levels at which these behavioural modifications are manifest. Spectators in swordtail networks: empirical work Cast of characters

Green swordtail fish are an excellent system in which to examine visually based network effects. Although there are few studies on the costs of combat in X. helleri, corticosteroid hormone levels are elevated above control levels for at least six hours after contest settlement, particularly in subordinate fish (Hannes et al., 1984). This finding, together with the data from a number of other studies on fish (e.g. Neat et al., 1998), indicates that fighting is likely to be costly for male green swordtails. Moreover, in both laboratory and field settings, male X. helleri establish social hierarchies where rank-order fights and/or attack–retreat sequences are common (Beaugrand et al., 1984; Franck et al., 1998). Thus, bystanders probably have ample opportunity to observe aggressive interactions that occur within their social environment and, given the potential costs of fighting, may benefit from doing so. Figure 5.1 depicts a simplified version of the swordtail social network as being composed of fighters engaged in aggressive signalling interactions, solitary individuals not involved in dominance interactions, and bystanders. Furthermore, the solitary individuals and fighters can either be observed or not observed by a bystander within range to extract information relevant to fighting ability (e.g. signals exchanged by the fighters or subtle behavioural/morphological cues of solitary individuals). Of course, each individual within the swordtail groups can assume the solitary, fighter or bystander position depending on the social circumstances they are exposed to at any given moment. Basic paradigm

Within the swordtail network, bystanders may extract information from fighters or solitary fish and integrate this information for use in future encounters with the observed individual(s). Furthermore, eavesdropping may elicit changes

(a)

+

+

−

−

+

+

−

−

87

April 11, 2005

effect; −, avoidance response; +, initiation of aggression or escalation and less likely to concede.

respond in a variety of ways to others in its social network (b; solid lines). Effect on bystander behaviour: 0, no

watched or not watched. After observing fights or solitary individuals (a; dashed lines), the bystanders may

have on a bystander’s response toward observed winners, observed losers, or solitary conspecifics that had been

0 521 582361 7

Fig. 5.1. The individuals comprising swordtail networks and the hypothetical effects that watching fights could

(b)

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor 11:24

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

88

0 521 582361 7

April 11, 2005

11:24

R. L. Earley & L. A. Dugatkin in the motivational state of the bystander that could influence its interactions with individuals that had not been observed (Fig. 5.1). To address these issues in the laboratory, we established a protocol where a bystander was visually isolated from, or able to observe, fighters or solitary conspecifics. This was accomplished by placing either an opaque plastic partition (opaque treatment) or a one-way mirror (mirror treatment) between the bystander and either a pair of interacting individuals or a solitary individual in laboratory aquaria (Earley & Dugatkin, 2002; Earley et al., 2003, 2004). In all of the experiments, the bystander and the fighters (or solitary conspecific) were matched for lateral surface area, a composite measurement of body length, body depth and sword length that corresponds better with fighting ability than standard length alone (Beaugrand et al., 1996). Following opportunity or no opportunity to observe, the bystander was pitted against one of the following individuals: the observed solitary individual, the winner of the observed fight, the loser of the observed fight, or a solitary fish that had not been observed (solitary-naive). In the context of aggressive interactions, eavesdropping could influence a bystander’s behaviour toward the observed individual(s) in many hypothetical ways. First, observation could have no effect (0) on the behaviour of the bystander. In this case, the behaviour of bystanders that had observed should be similar to the behaviour of naive bystanders. Second, the bystander may exhibit an avoidance response (−) if, through eavesdropping, it assessed the fighting ability of the observed individual(s) to exceed its own. Avoidance responses include refraining from initiating aggression or escalation and withdrawing from the contest. Third, if the bystander assessed the fighting ability of the observed individual(s) to be less than its own, it may be more inclined to initiate aggression or escalate and less likely to concede to its opponent (+). It is also possible that eavesdropping could affect the aggressive behaviour of the bystander outside the context of interacting with the individuals that had been observed, for example when interacting with solitary conspecifics that were not observed (Fig. 5.1). Eavesdropping on fights: confronting winners and losers

Our work on communication networks in X. helleri began with a relatively simple question: does watching a fight influence the agonistic response of a bystander toward the observed contestants? To determine precisely how eavesdropping influences a bystander’s interactions with the observed contestants, one must also recognize that the fighters enter the interaction with previous dominance or subordination experience. Previous winning or losing experiences are known to influence a host of contest characteristics in many fish species (e.g. probability of initiating aggression or winning against future opponents: Bakker

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:24

Poeciliid sociality: fighting, mating and networking & Sevenster, 1983; Bakker et al., 1989; Chase et al., 1994; Hsu & Wolf, 1999, 2001). Prior winning experiences tend to increase the probability of future contest success, while prior losing experiences decrease the likelihood of winning in the future; these experiential effects have been dubbed the ‘winner effect’ and ‘loser effect’, respectively. The opaque treatment can be seen as a control for the effects of previous fighting experience. Because the bystander is not allowed to observe the interaction, the dynamics of bystander versus winner or bystander versus loser contests are influenced primarily by the fighters’ previous experience. In contrast, both eavesdropping and prior experience effects can mediate bystander contest dynamics in the mirror treatment. Thus, when the dynamics of contests involving the bystander are compared between the opaque and mirror treatments, the contribution of eavesdropping can be determined explicitly. When the effects of the fighters’ prior experience are controlled, we found that watching fights had a considerable influence on bystander behaviour, particularly when confronted with the winner of the observed interaction (Earley & Dugatkin, 2002). Bystanders that observed the contest exhibited a more pronounced avoidance response toward winners than bystanders that had not observed the interaction (Fig. 5.2; Winner). However, the intensity of the observed contest had no bearing on the eavesdroppers’response toward winners (see caption to Fig. 5.3). Thus, eavesdroppers avoided observed winners regardless of whether they defeated their opponent by escalated or non-escalated means. A radically different scenario emerged in the bystander versus loser contests. Here, bystanders responded in a similar way to all losers, regardless of whether their defeat was witnessed (Fig. 5.2; Loser). Nevertheless, our data revealed that eavesdroppers were less likely to initiate aggressive behaviour and win against losers that persisted versus losers that retreated immediately in the observed contest (Fig. 5.3). These results demonstrate that swordtails not only make the dichotomous assessment of ‘winner versus loser’ but also calculate the fighting ability of each contestant, particularly the losers, independent of final outcome. The capacity to tease apart fighting ability from final status (i.e. winner or loser) may be of particular benefit during the initial establishment of dominance hierarchies, where the rank of each group member relative to the others, including the bystander, remains unclear. For instance, if eavesdroppers base their future agonistic decisions on status assessment alone (winner versus loser), challenging a relatively strong loser may prove costly. However, eavesdroppers that assess the fighting ability of the observed contestants based on, for example, persistence or willingness to escalate may be better equipped to adjust their agonistic behaviour in a manner consistent with the actual fighting ability of others in the network.

89

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:24

Bystander initiates aggression (proportion of contests)

(a)

observed

naive

observed

naive

Proportion of escalated contests

(b)

Fig. 5.2. The effect of watching fights on a swordtail bystander’s propensity to (a) initiate aggression, (b) participate in escalated contests and (c) win contests against observed winners, observed losers, solitary conspecifics that had been observed, or previously unknown individuals (solitary-naive). Grey bars represent the mirror treatment, where a bystander could see either the fighters or the solitary individual without being seen; white bars represent the opaque treatment, where a bystander remained naive to the presence of the fighters or the solitary individual prior to confrontation. An asterisk indicates significant differences (p < 0.05) between the mirror and opaque treatments.

90

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:24

Poeciliid sociality: fighting, mating and networking

Bystander wins (proportion of contests)

(c)

observed

naive

Fig. 5.2. (cont.)

Eavesdropping on fights: observing solitary individuals

Given that male swordtails modify their agonistic behaviour based on what appears to be an independent assessment of each contestant, bystanders may be getting more information from the individual than from the interaction itself. To address this possibility, we allowed some bystanders to observe a solitary fish through a one-way mirror; the remaining bystanders were visually isolated from the solitary individual using an opaque plastic partition. After the observation period (or lack thereof), the bystander was confronted with its opponent. In this experiment, we allowed small variation in body length, body depth and sword length between the two contestants. Asymmetries in any of the three size measurements did not lead to substantial mismatches in lateral surface area. When asymmetries in body size between the bystander and the watched individual were not considered, observation did not appear to elicit modifications in the bystanders’ behaviour (Earley et al., 2003; Fig. 5.2, Solitary-observed). However, when small differences in body size were examined as potentially informative cues, an interesting result emerged. The observers’ propensity to initiate attack increased as a function of relative body length. Somewhat surprisingly, observers were more prone to initiate against larger opponents, a trend evident only in the mirror treatment. Rapid escalation was a key predictor of contest success, with initiators of attack winning 81% of the interactions (Earley et al., 2003). Therefore,

91

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

April 11, 2005

11:24

R. L. Earley & L. A. Dugatkin (b) Probability that bystander wins against loser

(a) Probability of initiating aggression against loser

92

0 521 582361 7

Fig. 5.3. The probability that bystanders will a) initiate aggression or b) win against losers from the observed contest as a function of the degree to which the observed contest in the mirror treatment escalated. The number of reciprocal acts refers to the frequency with which an aggressive act from the eventual winner of the initial contest was countered with an aggressive act from the eventual loser. Solid lines indicate the probability of the event; dashed lines indicate the upper and lower confidence limits. Results of the logistic analyses for initiation (Wald χ1 2 = 4.4; p = 0.036) and winning (Wald χ1 2 = 4.5; p = 0.034) against persistent losers were statistically significant. Similar analyses on bystander versus winner contests yielded insignificant results (initiation: χ1 2 = 0.4; p = 0.54; winning: χ1 2 = 0.003; p = 0.96) and are not shown here.

when a potential fighting disadvantage is perceived, observers adopt tactics that enhance the probability of contest success against slightly larger opponents. Because relative body size had no influence on the behaviour of individuals that did not preview their opponent, prior observation is the likely trigger for modifications in the bystanders’ attack behaviour. These findings demonstrate that swordtails are capable of detecting small disparities in body length and that they adjust their agonistic behaviour in response to perceived size asymmetries. The fact that behavioural modifications elicited by watching solitary individuals were distinct, even opposite, from changes generated by observing fights suggests that different information is being integrated in each case. Information that accurately reflects superior fighting ability, such as outcome or persistence, may be more effective at deterring eavesdroppers than information about relative body

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:24

Poeciliid sociality: fighting, mating and networking size alone. In systems where prior experience effects, or any other social factor, have considerable influences on contest behaviour, information about size alone may not be sufficient to deter eavesdroppers, particularly when asymmetries are small. An interesting point that was not addressed by the above experiment is whether observing individuals exhibiting contest-type behaviour, without actually witnessing the interaction itself, would modify a bystander’s agonistic response. The importance of this question lies in partitioning how the fighting tactics of each contestant versus the dynamics of the actual interaction influence bystander behaviour. McGregor et al. (2001) allowed a bystander to observe either ‘real’ or ‘apparent’interactions between two male Betta splendens. The ‘real’interaction provided the bystander with information about two contestants that were actually fighting with one another. In the ‘apparent’interaction, the bystander was exposed to two males that appeared to be interacting with each other but were actually fighting against different opponents (Fig. 5.3 and Fig. 2.2d, p. 26). In the ‘apparent’ interactions, bystanders responded to winners (i.e. the individual of a pair that displayed longest) more strongly than to losers. It is important to note that, in the ‘apparent’ interactions, an individual that was perceived to have won/lost by the bystander may have actually obtained a different experience. Therefore, these results demonstrate that bystanders utilize information about individual contest behaviour (e.g. display duration), in addition to interaction dynamics, to gauge their future agonistic decisions. Another way to test this idea would be to allow a bystander to observe an individual exhibiting aggressive behaviour toward a stimulus that is out of view of the bystander (e.g. mirror image; conspecific opponent). Following observation, the bystander could be exposed to the individual it had observed; as potential controls, the bystander could be pitted against individuals that were observed not interacting with the stimulus and/or individuals that were not seen interacting with the stimulus. Provided that the watched individuals show substantial variation in aggressive behaviour, this type of experiment could elucidate whether watching individuals exhibiting contest-typical behaviour is sufficient to elicit modifications in bystander behaviour. Eavesdropping on fights: confronting naive conspecifics

From the experiments described above, it is clear that observing aggressive interactions prompts an agonistic response in swordtail bystanders. Initially, we interpreted this as evidence that bystanders extract information about the fighting ability of each contestant and respond accordingly when confronted with the observed individuals. Although our results are consistent with the existence of eavesdropping in swordtails, it is possible that observing fights elicits general changes in a bystander’s aggressive motivation that could affect future

93

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

94

0 521 582361 7

April 11, 2005

11:24

R. L. Earley & L. A. Dugatkin contest behaviour. This alternative hypothesis does not require that bystanders extract information from signalling interactions between others. Studies on Mozambique tilapia Oreochromis mossambicus (Oliveira et al., 2001) and Siamese fighting fish B. splendens (Clotfelter & Paolino, 2003) have shown that observing fights increases urinary androgen levels and aggressive behaviour/contest success, respectively. This type of response to social stimuli is best labelled as ‘priming’ (Hollis et al., 1995). However, the motivational changes experienced by swordtail bystanders, if any, may be quite different from priming. Recall that swordtail spectators responded with increased avoidance behaviour toward winners and losers that had persisted in the observed contest. Based on these data, any changes in the motivational state of swordtail bystanders should be manifest as decreases, rather than increases, in aggressive behaviour, possibly as a consequence of elevated corticosteroid stress hormones. To address the ‘motivational’ hypothesis from a behavioural perspective, some bystanders were exposed to aggressive interactions while others were visually isolated from a pair of fighters. Following the observation period, or the lack thereof, the bystander was confronted with an inexperienced fish that was not seen. Bystanders that observed conflict were equally likely to initiate aggression, escalate and win against the inexperienced fish as bystanders that were not exposed to the fight (Earley et al., 2004; Fig. 5.2: Solitary-naive). Therefore, observing fights does not appear to precipitate general increases or decreases in the aggressive motivation of swordtail bystanders. Given the lack of support for the ‘motivational’ hypothesis, it is reasonably clear that swordtail behaviour is modulated by more sophisticated mechanisms than observation-induced priming or stress. Namely, the agonistic response of swordtail eavesdroppers is influenced by the acquisition, integration and retention of information that accurately reflects the fighting ability of others in the network.

Communication networks and fighting: future considerations Eavesdropping in aggressive contexts: influence of individual differences

Most investigations of networking phenomena in fishes employ a ‘symmetrical’ design; that is, all participants (e.g. fighters and bystanders) are matched for attributes such as size or previous social experience. This type of design minimizes the effect of extraneous variables on the dynamics of contests involving the bystander and has helped to pinpoint how eavesdropping modulates bystander behaviour in X. helleri (Earley & Dugatkin, 2002) and B. splendens (Oliveira et al., 1998). However, the symmetrical design is probably not representative of communication networks in nature where substantial individual variation in size, previous fighting experience, physiological condition and social status are likely

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:24

Poeciliid sociality: fighting, mating and networking to exist. Individual variation in characteristics related to physical prowess may have important consequences for how the benefits and costs of fighting and/or observing contests are perceived and, thus, how the effects of eavesdropping are manifest. For instance, dominant and subordinate members of a social hierarchy may respond differently to the winners and losers of observed interactions. Maintenance of dominance status is likely a priority for dominant individuals while increasing status may benefit subordinates. Therefore, dominants should respond more vigorously to individuals that pose the greatest threat of rank usurpation (i.e. winners), while subordinates should respond more aggressively to losers so as to take advantage of opportunities to increase rank. Similarly, territorial defence is essential for resident individuals while non-territorial individuals may be most interested in seizing a territory of their own. In this case, residents may respond most aggressively toward upstart winners (Naguib & Todt, 1997; Naguib et al., 1999; Peake et al., 2001) while intruders should exploit recent losers. In addition to mediating the ‘direction’ of a bystander’s agonistic response, differences in individual perceptions of the costs and benefits of fighting may influence the degree to which eavesdropping is utilized as an assessment strategy (Johnstone, 2001). The relationship between the benefits and costs of fighting may be perceived as high for consistent winners, intermediate for inexperienced animals and low for consistent losers. Given that eavesdropping is most advantageous under circumstances where combat bears a relatively high cost (low benefit to cost ratio), inexperienced animals and consistent losers may benefit most by observing fights. However, as an individual’s perception of combat costs increases past a certain threshold, it may refrain from aggressive interactions and eavesdropping altogether (e.g. playing ‘dove’; Johnstone, 2001). An individual’s perception of the costs and benefits of fighting can be influenced by a host of additional factors including size, status, ownership, physiological or immunological condition (e.g. hungry versus satiated; healthy versus weak), reproductive state and resource value. Whether animals eavesdrop or how they respond to observing fights may be integrally related to each of these factors. Though the state dependency of eavesdropping effects has not yet gained empirical attention, this is likely to be an important avenue of future research in the area of animal communication networks. Eavesdropping in aggressive contexts: environmental influences

Just as factors associated with physical prowess can affect eavesdropping, so too can environmental or population-based variables. The presence of predators, or cues indicative of predator presence, influences the frequency of aggressive interactions in fishes (e.g. Martel & Dill, 1993; Wisenden & Sargent, 1997; Ch. 23). Individuals engaging in aggressive contests become more conspicuous to

95

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

96

0 521 582361 7

April 11, 2005

11:24

R. L. Earley & L. A. Dugatkin predators; that is, the costs of fighting are increased considerably when predators are around. Predation risk may also have a negative impact on the frequency of eavesdropping via at least three potential mechanisms: (a) if fights are less common in the presence of a predator, (b) if observing fights makes the bystander more conspicuous to predators (e.g. by association with the fighters), or (c) if the capacity to dedicate simultaneous attention to fights and predators is limited (see Dukas (2002) for review on selective attention). Although the frequency of eavesdropping may decline with predation risk, the efficacy of eavesdropping as an assessment strategy would not necessarily be compromised. In fact, when aggressive contests are rare and information flow through the network is reduced, bystanders may take every opportunity to extract information about the fighting abilities of others. In this sense, predation risk decreases the frequency and intensity of aggressive encounters and, by necessity, the opportunities to eavesdrop and the absolute amount of information available. However, the net benefit of eavesdropping under these circumstances may remain unaffected (or may even be increased). To test whether the frequency of eavesdropping is modified under different predation pressures would entail an analysis of how the proportion of eavesdroppers in a population changes across predation regimes or how predation risk affects an individual’s propensity to eavesdrop. Mathematical models may be best suited to address population-level questions (e.g. how the proportion of eavesdroppers changes with predation risk). Questions more amenable to empirical study include whether predation risk affects the extent to which bystanders gather information or whether bystanders compromise information acquisition in order to remain hidden from predators. For instance, one could compare whether bystanders derived from high-predation and low-predation sites differ in their response to the watched contestants. In addition, one could vary the quality or availability of refuges, the degree of habitat heterogeneity or the distance of refuges from the focal fight and subsequently quantify how the bystander’sresponse to the observed contestants changes with environmental condition. Besides predation risk, the social and/or mating system of the species in question may have a significant impact on how, and to what extent, exposure to fights alters a bystander’s agonistic decisions. The challenge hypothesis postulated that individuals should respond to social instability with increased testosterone levels and, presumably aggression levels, to deter rival males and secure reproductive opportunity (Wingfield et al., 1990; Ch. 21). In the broad sense, ‘social instability’ could include instances where individuals are being challenged directly (e.g. in territorial disputes) or where individuals are exposed to but not directly engaging in aggressive interactions. Mozambique tilapia and Siamese fighting fish fit nicely within this broad scope, as males of these species respond to observing fights with increased 11-ketotestosterone levels or increased aggression levels (Oliveira

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:24

Poeciliid sociality: fighting, mating and networking et al., 2001; Clotfelter & Paolino, 2003). However, we failed to uncover evidence that general changes in agonistic motivation accompany eavesdropping in green swordtail fish (Earley et al., 2004). We argued that differences in the mating system (e.g. breeding seasonality; monogamous versus polygamous) and/or social system (e.g. territorial versus hierarchical; stable versus unstable) could explain apparent species-specific differences in the response to observing contests (Earley et al., 2004; Ch. 21 has a more in-depth treatment). A piece of the hierarchical puzzle?

Eavesdropping is inherently a social phenomenon because it requires at least two individuals actively engaging with one another and a third party that extracts information from the signalling interchange. Nevertheless, there have been no controlled studies in fishes that examine aspects of networking above and beyond its effects at the dyadic level (for birds see Peake et al., 2002; Ch. 15). For example, in our work with X. helleri, we exposed a previously inexperienced bystander to an aggressive interaction and then assessed its response toward one of the contestants by staging dyadic contests (bystander versus winner or bystander versus loser). However, is it possible that the effects of eavesdropping are manifest differently when bystanders interact with a previously observed individual in the presence of other network members (e.g. observed winners or losers, unknown conspecifics, previous opponents that defeated or were defeated by the bystander; see Ch. 4)? Also, since social groups provide the opportunity for a wide range of interactions, each individual may have a different blend of prior social experience (e.g. several winning, losing or eavesdropping experiences or any combination thereof). Could the ways in which several social experiences are integrated over time, and the mere presence or absence of winner, loser and eavesdropping effects, have implications for the structure of animal social systems? From an empirical standpoint, these questions remain unanswered. However, the role of bystanders in the establishment of linear dominance hierarchies has been a question of conceptual interest for quite some time and, more recently, has attracted theoretical attention. Chase (1980, 1982) developed a conceptual model of linear hierarchy formation, the ‘jigsaw model’,that involved two interacting individuals and a bystander. Once a dominance relationship was established between the initial contestants, four possible interaction sequences could follow: initial dominant defeats bystander (double dominance), initial subordinate submits to the bystander (double subordination), bystander defeats initial dominant, or bystander submits to the initial subordinate. Double dominance and double subordination most often led to the establishment of linear hierarchies while the remaining two interaction sequences generated intransitive dominance orders.

97

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

98

0 521 582361 7

April 11, 2005

11:24

R. L. Earley & L. A. Dugatkin Chase (1980, 1982) did not address explicitly the behavioural mechanisms responsible for the double dominance and double subordination sequences and, thus, the genesis of linear hierarchies. There are at least two mechanisms that could give rise to such sequences. On the one hand, the initial contestants may update their estimation of their own fighting ability after having won or lost, i.e. winners increase and losers decrease their perceived fighting ability (Hsu & Wolf, 2001). As a consequence, previous winners should be more likely to defeat a bystander with average fighting ability and losers should be more likely to defer to the same bystander. In this instance, winner and loser effects are the behavioural mechanisms responsible for the double dominance and double subordination sequences and, in turn, the formation of linear hierarchies. On the other hand, the bystander may update its perception of the fighting ability of the initial contestants after observing the fight. For instance, the bystander may increase its estimate of the winner’s fighting ability and decrease its perception of the loser’s prowess. As a result, the bystander may be liable to attack the loser and submit to the winner. In this case, eavesdropping is the behavioural mechanism that generates the double dominance and double subordination sequences and linear hierarchies. These two scenarios need not be mutually exclusive. That is, winner, loser and eavesdropping effects may act in concert to promote linear hierarchy formation in animal groups. Dugatkin (2001) developed a simulation model to illustrate the potential interactions between eavesdropping and prior-experience effects in shaping dominance hierarchies. In this model, individuals could increase or decrease their own fighting ability and that of others in the network. The model made two assumptions: the first was that winner, loser and eavesdropping effects change individual perceptions with equal magnitude; the second was that all bystanders were privy to every interaction that occurred in their social environment. When eavesdropping and prior experience were considered separately, only winner effects produced a linear hierarchy. Conversely, when eavesdropping and prior experience effects operated simultaneously, a linear hierarchy always emerged. Therefore, linear hierarchies are most likely to occur when some combination of winner, loser and eavesdropping effects operate. In order to improve understanding of how these effects operate in natural systems, some of the assumptions employed by Dugatkin (2001) need to be relaxed. For at least two reasons, all bystanders are probably not capable of observing every interaction that occurs in their social network: first, time spent observing one interaction interferes with a bystander’sability to observe additional contests and, second, bystanders are unlikely to be within signal detection range of all aggressive encounters. Moreover, eavesdroppers probably have less information about the fighting ability of each observed contestant than the contestants themselves.

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:24

Poeciliid sociality: fighting, mating and networking Therefore, updates of others’ fighting ability may be of lesser magnitude than updates of one’s own fighting ability. R. L. Earley, S. Brosnan & J. Bragg (unpublished data) developed a simulation model of hierarchy formation in animals, in part to examine the consequences of relaxing these assumptions (Ch. 26). Several signalling modalities: seeing is not everything

Our work on eavesdropping and aggression examined the influence of visual signals on a bystander’s future behaviour. Nevertheless, visual signals may convey only part of the story in piscine duels. Acoustic (Lugli, 1997; Ladich, 1998; Amorim & Hawkins, 2000; Thorson & Fine, 2002), chemical (Waas & Colgan, 1992; Giaquinto & Volpato, 1997) and electrical (McGregor & Westby, 1992) stimuli have all been implicated as potential modes of communication in aggressive contests in fishes. Whether these signals elicit similar changes in bystander behaviour as visual signals has yet to be tested. Thorson & Fine (2002) demonstrated that male gulf toadfish Opsanus beta emit acoustic signals during the calls of neighbouring males, a phenomenon they called ‘acoustic tagging’ and interpreted as an aggressive display. If overlapping versus non-overlapping acoustic signals in toadfish provide information about willingness to escalate (or putative status), then playback experiments such as those used in territorial bird systems (e.g. Peake et al., 2001; Ch. 2) may be worth conducting, provided an anechoic aquatic chamber can be developed. Insights into multimodal signalling, the transmission of these signals within the network and the availability of such signals to bystanders will surely weave a more comprehensive story of how communication networks operate in nature.

Social eavesdropping and female mate choice As a general rule, females invest more time and energy in the reproductive process than males (e.g. production of viable eggs, gestation, maternal care, etc.) and, therefore, should be the choosier of the two sexes. Over the past several decades, an abundance of conceptual, theoretical and empirical work has focused on the factors that mediate female mate choice or male success in attracting females (Ryan, 1997). Most of the female mate-choice models have investigated how exaggerated male secondary sex characters and female preferences for these characters evolve, through either direct or indirect selection. For instance, Fisher’s runaway selection hypothesis postulated that, over evolutionary time, the alleles responsible for the male trait and the female’spreference for the male trait become genetically correlated (Fisher, 1958). This genetic linkage initiates a positive feedback loop whereby male traits can become more exaggerated as the preference for

99

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

100

0 521 582361 7

April 11, 2005

11:24

R. L. Earley & L. A. Dugatkin such traits strengthens and, in turn, further exaggeration of male traits intensifies female preference. Fisher’s genetic model, together with alternative models that address direct (e.g. sensory bias) and indirect (e.g. good genes) selection on female preferences (Ryan, 1997), have helped to elucidate how exaggerated male traits exist when they appear to have a negative impact on survival and how strong female preferences for these traits arise and persist. However, almost all sexual selection models assume that females choose mates independent of the choices made by other females (an exception is Kirkpatrick & Dugatkin, 1994; also see below). Is it possible that female choice depends not only on intrinsic preferences but also on the preferences of other females in the network? Two lines of evidence suggest that the decisions a female makes with respect to choosing mates is influenced in large part by observing interactions. First, monitoring apparent male–male interactions alters female mating decisions, for example initial mate choice (Doutrelant & McGregor, 2000) or loyalty to partner (Otter et al., 1999; Mennill et al., 2002; Ch. 7). Second, observing male–female courtship and/or mating interactions influences the subsequent mate choice decisions of a female peripheral to the interaction (e.g. Dugatkin, 1992, 1996a; Dugatkin & Godin, 1992; Grant & Green, 1996; Witte & Ryan, 1998; Witte & Noltemeier, 2002). In this section, we focus on the latter aspect of networking in poeciliid fishes: namely, how intersexual courtship rituals mediate the mating decisions of female P. reticulata that are not directly involved in the interaction. The principal concept linking communication networks to courtship interactions and female mate choice is mate copying. Mate copying occurs when “the conditional probability of choice of a given male by a female is either greater or less than the absolute probability of choice depending on whether that male mated previously or was avoided, respectively” (Pruett-Jones, 1992, p. 1001). Furthermore, the female must obtain information about a male’s mating history (or some part of it) by observation (Dugatkin, 1996b). In other words, the information gained by eavesdropping on mating interactions may sway a female’s decision toward or away from mating with the observed male. We confine our discussion to unambiguous cases of mate copying, that is, where a shift in the mating decisions of females is based solely on observing interactions between males and females other than oneself.

Spectators in guppy networks: empirical work Cast of characters

Guppies are an ideal species for examining networking phenomena such as mate copying for a number of reasons. First and foremost, guppies live in mixedsex shoals, within which females likely have opportunities to view (and potentially

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:24

Poeciliid sociality: fighting, mating and networking copy) the mate choice of nearby conspecifics. In addition, ample evidence suggests that social information is utilized by guppies in the context of mate choice (see below), foraging (Laland & Williams, 1997; Laland & Reader, 1999) and in the dynamics of shoal motion (Lachlan et al., 1998). Mate choice has been studied extensively in this species (for reviews see Kodric-Brown, 1990; Endler & Houde, 1995; Houde, 1997) and guppies exhibit normal courtship behaviour when placed in small aquaria; therefore, they are ideal for manipulative laboratory experiments. Typical courtship interactions involve the male directing sigmoid displays at the female; receptive females respond to male displays with a ‘gliding’ motion (for a more complete description, see Liley (1966)). Lastly, female guppies from the Paria river (the population that was used in the experiments described below) exhibit heritable preferences for certain male traits (e.g. orange colour: Houde, 1987, 1988; Endler & Houde, 1995). Experiments on such populations provide a unique opportunity to examine the interaction between genetic-based preferences and those resulting from interactions within the guppy social network. The cast of characters in the guppy network resembles that described for the swordtails. The principal difference is that the network is partitioned into (a) males that are either quite similar with respect to exaggerated colour patterns or that differ by varying degrees; (b) at least one female being courted by, or exhibiting a preference for, a focal male; and (c) a female bystander within range to detect courtship and/or mating signals (Fig. 5.4). In theory, any of the females within the guppy network can assume a bystander role or a courtship role at any given time; males however, are restricted to a courtship role. Basic paradigm

In the guppy network, we are concerned primarily with how a female bystander responds to a male that was recently preferred by another female (here, the focal female). To address this question, one of the authors (Dugatkin) developed a protocol where a female was either exposed or not exposed to a courtship interaction between a male and a focal female. Following the observation period, the female bystander was given the opportunity to make a choice between the male that was preferred by the focal female and the male that was not preferred (Dugatkin, 1992). The preference of the focal female was ‘staged’ because she was restricted to the side of the aquarium occupied by one of the two males; therefore, the bystander female observed an apparent choice by the focal female. Eavesdropping on the mate choice of others could have three potential effects on the future behaviour of the female bystander (Fig. 5.4). If observing courtship interactions has no effect (0) on the female bystander then she should choose both males with equal frequency. If females increase their assessment of a male’s quality (+) after observing him successfully court, then the bystander should choose the male that

101

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

102

0 521 582361 7

April 11, 2005

11:24

R. L. Earley & L. A. Dugatkin Brightly coloured male

Relatively drab male

+ / 0 /−

+ / 0 /− Bystander

Brightly coloured male

Model female

Relatively drab male

Fig. 5.4. The individuals comprising guppy networks and the hypothetical effects that watching mating interactions could have on a female bystander’s response toward the observed, apparently successful male. After witnessing a courtship interaction (dashed lines), the female bystander can respond in a variety of ways toward males in her social network (solid lines). Effect on bystander behaviour: 0, no effect; −, decreased assessment of male quality; +, increased assessment of male quality.

was preferred by the focal female a significant majority of the time, i.e. mate copying. It is also possible that females decrease their assessment of a male’s quality (−) after observing a courtship interaction, for instance if a recent mating depletes the male’s sperm supply (Nakatsuru & Kramer, 1982). In this case, the bystander female should avoid the male that was recently preferred by the focal female. In the following sections, we provide a general overview of the empirical work on mate copying that has been conducted in P. reticulata, with special emphasis an how eavesdropping mediates a female’s subsequent mating decisions. Female mate copying: does it occur?

Dugatkin’s (1992) research on guppies provided the first controlled study of female mate copying. In this study, the bystander female chose the male that

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:24

Poeciliid sociality: fighting, mating and networking was preferred by the focal female in a significant proportion of the trials. In a series of five additional experiments, Dugatkin (1992) ruled out a host of alternative explanations, including female biases toward one side of the aquarium over the other, female preference for areas of the tank recently occupied by the largest group of fish (‘schooling preference’) and female choice for males that courted most recently and thus were more active. These findings provided substantial evidence that females do copy the mate choice of others and that female bystanders assess the quality of recently preferred males to be higher than those that were not preferred. Moreover, Dugatkin (1992) demonstrated that courtship interactions between the focal female and a male are crucial for eliciting mate copying behaviour. These results support the notion that mate choice decisions in P. reticulata are, in part, socially modulated. Two other studies (Brooks, 1996; Lafleur et al., 1997) found no evidence of mate copying in guppies. It is critical to note, however, that neither of these studies used guppies from natural streams in Trinidad. Nonetheless, support for mate copying has also been reported in other poeciliid species (Poecilia latipinna; Witte & Ryan, 1998) and the Japanese medaka Oryzias laticeps (Grant & Green, 1996). Given the pivotal role of male–female interactions in mediating a female bystander’s future mating decisions, the social system of guppies and other species in which there is unambiguous evidence for mate copying are clearly amenable to interpretation from the perspective of communication networks. Socially modulated versus intrinsic preferences

In the absence of mate copying opportunities, female guppies distinguish between males based on a number of phenotypic traits, for example tail size (Bischoff et al., 1985) or colouration patterns (Houde, 1988; Houde & Endler, 1990; Endler & Houde, 1995). Furthermore, female preferences for male traits such as orange colouration have a significant heritable component (Houde, 1988; Houde & Endler, 1990). Because female preferences in guppies are shaped by both genetic and social factors, an interesting question is whether and how these factors interact. Dugatkin & Godin (1992) conducted an experiment where a female was initially allowed to choose between two males that differed only in their colour pattern. Following this choice, the female was either exposed or not exposed to a focal female placed beside the male not chosen in the initial preference test. After the observation period, the female was again allowed to choose between the same two males. Interestingly, the female reversed her choice in 75% of the trials, suggesting that social factors (i.e. eavesdropping on the apparent preference of another female for the less preferred male) are capable of overriding intrinsic preferences. Dugatkin (1996a) employed a similar protocol but systematically altered the asymmetries in orange colouration between the two males provided to the female

103

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

April 11, 2005

11:24

R. L. Earley & L. A. Dugatkin

Proportion of times female preferred more-orange male

104

0 521 582361 7

Mean difference in total orange body colour between males Fig. 5.5. The proportion of times female bystanders chose the more orange of the two males in the presence (solid squares) or absence (open circles) of a focal female. (Adapted from Dugatkin, 1996a.)

bystander. When there was no opportunity for mate copying, females exhibited a more pronounced preference for males with more orange as the asymmetries in colouration increased (Fig. 5.5). However, when a focal female was placed beside the male with less orange, so as to simulate an apparent preference for drab males, the female bystander chose the less-orange male significantly more often in all cases except when the asymmetry was most drastic (40% difference in total orange body colouration). Even when substantial asymmetries in male colouration existed, a female bystander could be coerced into choosing the less-orange male by increasing his perceived attractiveness via social manipulation (Dugatkin, 1998). This was accomplished by increasing the number of focal females that exhibited an apparent preference for the drab male or by increasing the amount of time a single focal female spent near the less-orange male. Witte & Noltemeier (2002) obtained strikingly similar results in female sailfin mollies P. latipinna, a related poeciliid species where females presumably exhibit a genetically based preference for large males (Marler & Ryan, 1997). Witte & Noltemeier (2002) also demonstrated that reversals in mate choice incited by a copying bout persisted for long periods

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:24

Poeciliid sociality: fighting, mating and networking of time (five weeks), even when a female was allowed to choose between previously unknown large and small males. This marks one potential direction for the guppy work: to determine whether observing courtship bouts influences female choice even in the absence of the observed, previously chosen male. In all, these results provide compelling evidence that observing courtship interactions is sufficient to overturn female poeciliid’s genetically predisposed choice. Furthermore, as the amount of social information available to the female bystander about a male’s potential quality increases, the more apt she is to rely on social signals in lieu of the preference algorithm engrained in her genes. It is important to note that in all of the studies on mate choice copying the observed female was placed near a male of lesser quality (e.g. smaller or with less orange). The low-quality male was then considered to be a suitable mate by virtue of his being chosen earlier. This type of design is necessary to decouple mate choice copying from established, genetically based preferences. In nature, however, female bystanders likely observe other females choosing relatively high-quality mates. Therefore, mate choice copying is likely to reinforce rather than contradict preexisting preferences (Brooks, 1996). Mate copying: a theoretical perspective

Relative to eavesdropping in an aggressive context, mate choice copying has received an abundance of theoretical attention. Mathematical treatments of mate copying have addressed two principal evolutionary questions: first, how established mate copying strategies influence the evolution of female preferences and male secondary sex characteristics or the variance in male mating success (Wade & Pruett-Jones, 1990; Kirkpatrick & Dugatkin, 1994; Laland, 1994; Agrawal, 2001) and, second, the emergence and persistence of copying strategies (Losey et al., 1986; Pruett-Jones, 1992; Dugatkin & Hoglund, 1995; Servedio & Kirkpatrick, 1996; Stohr, 1998). An exhaustive comparison of these models is beyond the scope of this chapter. However, mate copying theory has revealed that network phenomena, in particular the behavioural modifications that result from observing social interactions, can have both short-term effects on individual decision making and substantial evolutionary consequences. For instance, established mate copying strategies can increase the variance in male mating success and, thus, the opportunity for selection on male phenotypic traits that correlate well with attractiveness (Wade & Pruett-Jones, 1990). When bystanders most often observe interactions between females and males that exhibit the most common phenotype and when the effects of observing courtship interactions are independent of male phenotype, mate copying impedes the spread of rare (or novel) male traits (Kirkpatrick & Dugatkin, 1994; Laland, 1994; Agrawal, 2001). Interestingly, when eavesdropping has a graded effect on bystander behaviour (e.g. depending

105

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

106

0 521 582361 7

April 11, 2005

11:24

R. L. Earley & L. A. Dugatkin on the phenotype of the observed male), a number of additional scenarios emerge, including the potential for the spread of rare male phenotypes (Agrawal, 2001). Given that mate copying can have dramatic effects on the distribution of male traits in a population, it is important to understand how mate copying strategies could emerge in the first place. A recurrent theme in the literature in this area, particularly in models that assume female choice is under direct selection (but see Servedio & Kirkpatrick, 1996), is that copying strategies will thrive when female choice is costly (e.g. sampling costs, search time, predation risk: Pruett-Jones 1992) or when substantial differences exist in females’ability to discriminate low- versus high-quality males (Stohr, 1998). In the next section, we elaborate on how the costs and benefits of female mate choice (or copying) potentially influence the degree to which copying is used as a mate-assessment strategy.

Communication networks and mate choice: future directions Eavesdropping and mate choice: influence of individual differences

In most studies on mate choice copying, the bystander and focal female are matched for characteristics such as size, age and previous mating experience (but see Dugatkin & Godin, 1993; Witte & Ryan, 1998). Under natural circumstances, females involved in the interaction are likely to differ in some respect. Individual differences may influence how observation of courtship interactions modulates bystander behaviour by adjusting the expected costs and benefits of mate copying or mate choice itself. Dugatkin & Godin (1993) demonstrated that small (young) females copy the mate choice of large (old) females, but not vice versa. If older females have more mate choice experience and if experience decreases the rate at which errors in mate choice are committed (e.g. choosing a poor-quality mate), then older females are liable to be better at discriminating low- from high-quality mates than younger females. Therefore, the fitness benefit of copying older females, namely having a higher probability of mating with superior males, may be quite high for young females. Conversely, the costs of copying a younger, error-prone female may be sufficiently high to discourage mate copying in older females. A host of other variables, including gravidity, previous mating experience independent of age and physiological condition may affect how the benefits and costs of mate choice and/or mate copying are perceived. For instance, if female mate choice entails substantial sampling or search costs, individuals who are under significant time constraints (e.g. hungry individuals whose time would better be spent foraging) should be more likely to rely on the choices of others. As with eavesdropping in an aggressive context, the state dependency of

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:24

Poeciliid sociality: fighting, mating and networking mate copying represents an understudied and potentially important aspect of this field. Eavesdropping and mate choice: environmental influences

Environmental factors may also modify an individual’spropensity to copy the choice of other females. Predation risk is known to influence female choosiness and this is likely because females encounter a trade-off between the benefits of remaining vigilant toward predators and the costs of spending time searching for, or assessing, potential mates (Magnhagen, 1991; Pocklington & Dill, 1995). As the costs of female choice intensify, mate copying should become increasingly beneficial; therefore, under high predation risk, we might expect females to increase their propensity to copy other females. However, this also assumes that bystanders are less conspicuous to predators than individuals involved in active mate choice. Briggs et al. (1996) found little support for this hypothesis in guppies derived from streams with relatively high predation risk; apparent predation risk did not influence the proportion of females reversing their choice in the presence of a focal female. However, as Briggs et al. (1996) acknowledged, females from populations derived from streams with a high predation risk may not exhibit differential mate copying responses under different predation regimes because, given the high cost to female mate choice under natural circumstances, they may already exhibit the maximal propensity to copy. Although Briggs et al. (1996) discarded this thesis, they did not test whether female guppies derived from low-predation streams exhibit divergent responses in the presence or absence of a predator. Nevertheless, their work marks one of the first attempts to broach a largely unexplored area of communication networks: namely how predation risk can modulate individual tendencies to eavesdrop on mating interactions. The probability or frequency for females to copy the choices of others may also be influenced by temporal factors (e.g. early vs. late in the breeding season: Dugatkin & Hoglund, 1995), the quality of one’s own mate, the availability of refuge or the opportunity to observe mating interactions. Identifying influential environmental factors may aid in formulating a comparative communication network concept and should certainly trigger empirical work dedicated to distinguishing the relative effects of individual- and environment-based factors on decision-making processes in animals.

Summary The aim of this chapter was to illustrate the importance of extending research on mating and aggressive behaviour beyond the dyad and into a broader social milieu. We have demonstrated that poeciliid fishes are capable of extracting

107

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

108

0 521 582361 7

April 11, 2005

11:24

R. L. Earley & L. A. Dugatkin information from various types of social interaction and integrating this information for use in future encounters with the observed individuals. A recurrent theme in this chapter is how the actual (or perceived) costs and benefits of fighting, mating or eavesdropping can influence the short-term effects of observing social interactions and the evolutionary viability of such strategies. This is where mate copying theory and current, albeit sparse, theory on eavesdropping in an aggressive context intersect. Irrespective of a bystander’s focus, when social interactions bear a high cost, it pays to be attentive to signalling exchanges between others. Whether the same selection pressures act on eavesdropping regardless of the context in which it is used remains to be explored theoretically. Nevertheless, we have attempted to highlight a wealth of factors that could affect the circumstances under which eavesdropping is favoured and how the effects of eavesdropping are manifest behaviourally. Daunting as the list of candidate influences may be, a comprehensive understanding of communication networks relies, in part, on our ability to partition the relative effects of each of these factors using comparative, theoretical or empirical approaches. Unveiling the complex interactions between individual (e.g. size, status, age), social (e.g. mating systems) and ecological (e.g. predation risk, seasonality) variation and bystander behaviour marks a compelling future direction for the field of communication networks.

Acknowledgements We express our sincere thanks to Matthew Druen, Trish Sevene-Adams, Meredith McGee, Michael Boles, Megan Tinsley and Blair Gilliland for their assistance in the laboratory. We are also grateful to Peter McGregor for extending the invitation to contribute to this book and to Matthew Grober, Cathleen Drilling and Ed Rodgers for insightful discussion on earlier versions of this chapter. The work described in this chapter was funded in part by the National Science Foundation, Sigma Xi, Kentucky Academy of Science, Animal Behavior Society and the American Livebearers Association.

References Agrawal, A. 2001. The evolutionary consequences of mate copying on male traits. Behavioral Ecology and Sociobiology, 51, 33–40. Amorim, M. & Hawkins, A. 2000. Growling for food: acoustic emissions during competitive feeding of the streaked gurnard. Journal of Fish Biology, 57, 895–907. Bakker, T. C. M. & Sevenster, P. 1983. Determinants of dominance in male sticklebacks (Gasterosteus aculeatus L.). Behaviour, 86, 55–71.

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:24

Poeciliid sociality: fighting, mating and networking Bakker, T. C. M., Bruijn, E. F. D. & Sevenster, P. 1989. Asymmetrical effects of prior winning and losing on dominance in sticklebacks (Gasterosteus aculeatus). Ethology, 82, 224–229. Beaugrand, J. P., Caron, J. & Comeau, L. 1984. Social organization of small heterosexual groups of green swordtails (Xiphophorus helleri, Pisces, Poeciliidae) under conditions of captivity. Behaviour, 91, 24–60. Beaugrand, J. P., Payette, D. & Goulet, C. 1996. Conflict outcome in male green swordtail fish dyads (Xiphophorus helleri): interaction of body size, prior dominance/subordination experience and prior residency. Behaviour, 133, 303–319. Bischoff, R., Gould, J. & Rubenstein, D. 1985. Tail size and female choice in the guppy (Poecilia reticulata). Behavioral Ecology and Sociobiology, 17, 252–255. Briggs, S., Godin, J. G. J., Dugatkin, L. A. 1996. Mate-choice copying under predation risk in the Trinidadian guppy (Poecilia reticulata). Behavioral Ecology, 7, 151–157. Brooks, R. 1996. Copying and the repeatability of mate choice. Behavioral Ecology and Sociobiology, 39, 323–329. Brosnan, S. F., Earley, R. L., Dugatkin, L. A. 2003. Observational learning and predator inspection in guppies (Poecilia reticulata). Ethology, 109, 823–833. Brown, C. & Laland, K. N. 2002. Social learning of a novel avoidance task in the guppy: conformity and social release. Animal Behaviour, 64, 41–47. Brown, G. & Godin, J. G. J. 1999. Chemical alarm signals in wild Trinidadian guppies (Poecilia reticulata). Canadian Journal of Zoology, 77, 562–570. Chase, I. D. 1980. Social process and hierarchy formation in small groups: a comparative perspective. American Sociological Review, 45, 905–924. 1982. Dynamics of hierarchy formation: the sequential development of dominance relationships. Behaviour, 80, 218–239. Chase, I. D., Bartolomeo, C. & Dugatkin, L. A. 1994. Aggressive interactions and inter-contest interval: how long do winners keep winning? Animal Behaviour, 48, 393–400. Clotfelter, E. & Paolino, A. 2003. Winner effect is induced in bystanders to aggressive contests between male fighting fish. Animal Behaviour, 66, 343–347. Doutrelant, C. & McGregor, P. K. 2000. Eavesdropping and mate choice in female fighting fish. Behaviour, 137, 1655–1669. Dukas, R. 2002. Behavioural and ecological consequences of limited attention. Philosophical Transactions of the Royal Society of London, Series B, 357, 1539–1547. Dugatkin, L. A. 1992. Sexual selection and imitation: females copy the mate choice of others. American Naturalist, 139, 1384–1389. 1996a. Interface between culturally based preferences and genetic preferences: female mate choice in Poecilia reticulata. Proceedings of the National Academy of Sciences, USA, 93, 2770–2773. 1996b. Copying and mate choice. In: Social Learning in Animals: the Roots of Culture ed. C. Heyes & G. Galef, New York: Academic Press, pp. 85–105. 1998. Genes, copying, and female mate choice: shifting thresholds. Behavioral Ecology, 9, 323–327.

109

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

110

0 521 582361 7

April 11, 2005

11:24

R. L. Earley & L. A. Dugatkin 2001. Bystander effects and the structure of dominance hierarchies. Behavioral Ecology, 12, 348–352. Dugatkin, L. A. & Godin, J. G. J. 1992. Reversal of female mate choice by copying in the guppy (Poecilia reticulata). Proceedings of the Royal Society of London, Series B, 249, 179–184. 1993. Female mate copying in the guppy (Poecilia reticulata): age-dependent effects. Behavioral Ecology, 4, 289–292. Dugatkin, L. A. & Hoglund, J. 1995. Delayed breeding and the evolution of mate copying in lekking species. Journal of Theoretical Biology, 174, 261–267. Earley, R. L. & Dugatkin, L. A. 2002. Eavesdropping on visual cues in green swordtail (Xiphophorus helleri) fights: a case for networking. Proceedings of the Royal Society of London, Series B, 269, 943–952. Earley, R. L., Tinsley, M. & Dugatkin, L. A. 2003. To see or not to see: does previewing a future opponent affect the contest behavior of green swordtail males (Xiphophorus helleri)? Naturwissenschaften, 90, 226–230. Earley, R. L., Druen, M. & Dugatkin, L. A. 2004. Watching fights does not alter a bystander’s response towards na¨ıve conspecifics in male green swordtail fish (Xiphophorus helleri). Animal Behaviour, in press. Endler, J. A. & Houde, A. E. 1995. Geographic variation in female preferences for male traits in Poecilia reticulata. Evolution, 49, 456–468. Enquist, M. & Leimar, O. 1983. Evolution of fighting behaviour: decision rules and assessment of relative strength. Journal of Theoretical Biology, 102, 387–410. Enquist, M., Leimar, O., Ljungberg, T., Mallner, Y. & Segerdahl, N. 1990. A test of the sequential assessment game: fighting in the cichlid fish, Nannacara anomala. Animal Behaviour, 40, 1–14. Fisher, R. A. 1958. The Genetical Theory of Natural Selection, 2nd edn. New York: Dover. Franck, D., Klamroth, B., Taebel-Hellwig, A. & Schartl, M. 1998. Home ranges and satellite tactics of male green swordtails (Xiphophorus helleri) in nature. Behavioural Processes, 43, 115–123. Giaquinto, P. C. & Volpato, G. L. 1997. Chemical communication, aggression, and conspecific recognition in the fish Nile tilapia. Physiology and Behavior, 62, 1333–1338. Grant, J. & Green, L. 1996. Mate copying versus preference for actively courting males by female Japanese medaka (Oryzias latipes). Behavioral Ecology, 7, 165–167. Haller, J. 1995. Biochemical background for an analysis of cost–benefit interrelations in aggression. Neuroscience and Biobehavioral Reviews, 19, 599–604. Halperin, J., Giri, T., Elliot, I. & Dunham, D. 1998. Consequences of hyperaggressiveness in Siamese fighting fish: cheaters seldom prosper. Animal Behaviour, 55, 87–96. Hannes, R. P., Franck, D. & Leimann, F. 1984. Effects of rank-order fights on whole-body and blood concentrations of androgens and corticosteroids in the male swordtail (Xiphophorus helleri). Ethology, 65, 53–65. Hollis, K. L., Dumas, M. J., Singh, P. & Fackelman, P. 1995. Pavlovian conditioning of aggressive behavior in blue gourami fish (Trichogaster trichopterus): winners become winners and losers stay losers. Journal of Comparative Psychology, 109, 125–133.

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:24

Poeciliid sociality: fighting, mating and networking Houde, A. E. 1987. Mate choice based on naturally occurring colour-pattern variation in a guppy population. Evolution, 41, 1–10. 1988. Genetic differences in female choice between two guppy populations. Animal Behaviour, 36, 510–516. 1997. Sex, Colour and Mate Choice in Guppies. Princeton, CT: Princeton University Press. Houde, A. E. & Endler, J. A. 1990. Correlated evolution of female mating preference and male colour pattern in the guppy, Poecilia reticulata. Science, 248, 1405–1408. Hsu, Y. & Wolf, L. L. 1999. The winner and loser effect: integrating multiple experiences. Animal Behaviour, 57, 903–910. 2001. The winner and loser effect: what fighting behaviours are influenced? Animal Behaviour, 61, 777–786. Jakobsson, S., Brick, O. & Kullberg, C. 1995. Escalated fighting behaviour incurs increased predation risk. Animal Behaviour, 49, 235–239. Johnstone, R. 2001. Eavesdropping and animal conflict. Proceedings of the National Academy of Sciences, USA, 98, 9177–9180. Kirkpatrick, M. & Dugatkin, L. A. 1994. Sexual selection and the evolutionary effects of copying mate choice. Behavioral Ecology and Sociobiology, 34, 443–449. Kodric-Brown, A. 1990. Mechanisms of sexual selection: insights from fishes. Annales Zoologici Fennici, 27, 87–100. Lachlan, R. F., Crooks, L. & Laland, K. N. 1998. Who follows whom? Shoaling preferences and social learning of foraging information in guppies. Animal Behaviour, 56, 181–190. Ladich, F. 1998. Sound characteristics and outcome of contests in male croaking gouramis (Teleostei). Ethology, 104, 517–529. Lafleur, D. L, Lozano, G. A. & Sclafani, M. 1997. Female mate-choice copying in guppies, Poecilia reticulata: a re-evaluation. Animal Behaviour, 54, 579–586. Laland, K. N. 1994. Sexual selection with a culturally transmitted mating preference. Theoretical Population Biology, 45, 1–15. Laland, K. N. & Reader, S. M. 1999. Foraging innovation in the guppy. Animal Behaviour, 57, 331–340. Laland, K. N. & Williams, K. 1997. Shoaling generates social learning of foraging information in guppies. Animal Behaviour, 53, 1161–1169. Liley, N. R. 1966. Ethological isolating mechanisms in four sympatric species of poeciliid fishes. Behaviour (Supplement), 31, 1–197. Losey, G., Jr Stanton, F., Telecky, T., Tyler III, W. and the Zoology 691 Graduate Seminar Class 1986. Copying others, an evolutionary stable strategy for mate choice: a model. American Naturalist, 128, 653–664. Lugli, M. 1997. Response of male goby, Padogobius martensii, to aggressive sound playback following pre-experimental visual stimulation. Behaviour, 134, 1175–1188. Magnhagen, C. 1991. Predation risk as a cost of reproduction. Trends in Ecology and Evolution, 6, 183–186. Magurran, A. E. & Pitcher, T. J. 1987. Provenance, shoal size, and the socio-biology of predator evasion behaviour in minnow shoals. Proceedings of the Royal Society of London, Series B, 229, 439–465.

111

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

112

0 521 582361 7

April 11, 2005

11:24

R. L. Earley & L. A. Dugatkin Marler, C. A. & Ryan, M. J. 1997. Origin and maintenance of a female mating preference. Evolution, 51, 1244–1248. Martel, G. & Dill, L. M. 1993. Feeding and aggressive behaviours in juvenile coho salmon (Oncorhynchus kisutch) under chemically mediated risk of predation. Behavioral Ecology and Sociobiology, 32, 365–370. McGregor, P. K. 1993. Signalling in territorial systems: a context for individual identification, ranging and eavesdropping. Philosophical Transactions of the Royal Society of London, Series B, 340, 237–244. McGregor, P. K. & Westby, G. W. M. 1992. Discrimination of individually characteristic electric organ discharges by a weakly electric fish. Animal Behaviour, 43, 977–986. McGregor, P. K., Otter, K. & Peake, T. M. 2000. Communication networks: receiver and signaller perspectives. In: Animal Signals: Signalling and Signal Design in Animal Communication ed. Y. Espmark, T. Amundsen & G. Rosenqvist. Trondheim: Tapir Academic Press, pp. 405–416. McGregor, P. K., Peake, T. M. & Lampe, H. M. 2001. Fighting fish Betta splendens extract relative information from apparent interactions: what happens when what you see isn’t what you get. Animal Behaviour, 62, 1059–1065. Mennill, D., Ratcliffe, L. & Boag, P. 2002. Female eavesdropping on male song contests in songbirds. Science, 296, 873. Mesterton-Gibbons, M. & Adams, E. 1998. Animal contests as evolutionary games. American Scientist, 86, 334–341. Mirza, R., Scott, J. & Chivers, D. 2001. Differential responses of male and female red swordtails to chemical alarm cues. Journal of Fish Biology, 59, 716–728. Naguib, M. & Todt, D. 1997. Effects of dyadic vocal interactions on other conspecific receivers in nightingales. Animal Behaviour, 54, 1535–1543. Naguib, M., Fichtel, C. & Todt, D. 1999. Nightingales respond more strongly to vocal leaders of simulated dyadic interactions. Proceedings of the Royal Society of London, Series B, 266, 537–542. Nakatsuru, K. & Kramer, D. L. 1982. Is sperm cheap? Limited male fertility and female choice in the lemon tetra (Pisces, Characidae). Science, 216, 753–755. Neat, F., Taylor, A. & Huntingford, F. 1998. Proximate costs of fighting in male cichlid fish: the role of injuries and energy metabolism. Animal Behaviour, 55, 875–882. Oliveira, R. F., McGregor, P. K. & Latruffe, C. 1998. Know thine enemy: fighting fish gather information from observing conspecific interactions. Proceedings of the Royal Society of London, Series B, 265, 1045–1049. Oliveira, R. F., Lopes, M., Carneiro, L. A. & Canario, A. V. M. 2001. Watching fights raises fish hormone levels. Nature, 409, 475. Otter, K., McGregor, P. K., Terry, A. M. R. et al. 1999. Do female great tits (Parus major) assess males by eavesdropping? A field study using interactive song playback. Proceedings of the Royal Society of London, Series B, 266, 1305–1309. Payne, R. 1998. Gradually escalating fights and displays: the cumulative assessment model. Animal Behaviour, 56, 651–662. Payne, R. & Pagel, M. 1997. Why do animals repeat displays? Animal Behaviour, 54, 109–119.

P1: JZZ/... P2: JZZ/... 0521823617c05.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:24

Poeciliid sociality: fighting, mating and networking Peake, T. M., Terry, A. M. R., McGregor, P. K. & Dabelsteen, T. 2001. Male great tits eavesdrop on simulated male-to-male vocal interactions. Proceedings of the Royal Society of London, Series B, 268, 1183–1187. 2002. Do great tits assess rivals by combining direct experience with information gathered by eavesdropping? Proceedings of the Royal Society of London, Series B, 269, 1925–1929. Pocklington, R. & Dill, L. M. 1995. Predation on females or males: who pays for bright male traits? Animal Behaviour, 49, 1122–1124. Pruett-Jones, S. 1992. Independent versus nonindependent mate choice: do females copy each other. American Naturalist, 140, 1000–1009. Pulliam, H. R. & Caraco, T. 1984. Living in groups: is there an optimal group size? In: Behavioural Ecology. An Evolutionary Approach, 2nd edn. ed. J. R. Krebs & N. B. Davies. pp. Oxford: Blackwell Scientific, pp. 122–147. Ranta, E. & Juvonen, S. K. 1993. Interference affects food-finding rate in schooling fish. Journal of Fish Biology, 43, 531–535. Ryan, M. J. 1997. Sexual selection and mate choice. In: Behavioural Ecology. An Evolutionary Approach, 4th edn. ed. J. R. Krebs & N. B. Davies. Oxford: Blackwell Scientific, pp. 179–202. Schuett, G. 1997. Body size and agonistic experience affect dominance and mating success in male copperheads. Animal Behaviour, 54, 213–224. Servedio, M. & Kirkpatrick, M. 1996. The evolution of mate choice copying by indirect selection. American Naturalist, 148, 848–867. Stohr, S. 1998. Evolution of mate-choice copying: a dynamic model. Animal Behaviour, 55, 893–903. Swaney, W., Kendal, J., Capon, H., Brown, C. & Laland, K. N. 2001. Familiarity facilitates social learning of foraging behaviour in the guppy. Animal Behaviour, 62, 591–598. Thorson, R. & Fine, M. 2002. Acoustic competition in the gulf toadfish Opsanus beta: acoustic tagging. Journal of the Acoustical Society of America, 111, 2302–2307. Waas, J. R. & Colgan, P. W. 1992. Chemical cues associated with the visually elaborate aggressive displays of threespine sticklebacks. Journal of Chemical Ecology, 18, 2277–2284. Wade, M. J. & Pruett-Jones, S. G. 1990. Female copying increases the variance in male mating success. Proceedings of the National Academy of Sciences, USA, 87, 5749–5753. Wingfield, J. C., Hegner, R. E., Dufty, A. M. & Ball, G. F. 1990. The challenge hypothesis: theoretical implications for patterns of testosterone secretion, mating systems, and breeding strategies. American Naturalist, 136, 829–846. Wisenden, B. D. & Sargent, R. C. 1997. Anti-predator behaviour and suppressed aggression by convict cichlids in response to injury-released chemical cues of conspecifics but not to those of an allopatric heterospecific. Ethology, 103, 283–291. Witte, K. & Noltemeier, B. 2002. The role of information in mate-choice copying in female sailfin mollies (Poecilia latipinna). Behavioral Ecology and Sociobiology, 52, 194–202. Witte, K. & Ryan, M. J. 1998. Male body length influences mate-choice copying in the sailfin molly Poecilia latipinna. Behavioral Ecology, 9, 534–539.

113

P1: JZZ/... P2: JZZ/... 0521823617c06.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:41

6

The occurrence and function of victory displays within communication networks john l. bower Fairhaven College, Western Washington University, Washington, USA

Introduction Much recent research has focused on communication that occurs prior to and during agonistic interactions in animals, leading to theoretical and empirical advances in our understanding of the evolution of signalling before and during agonistic contests (Maynard Smith, 1982; Huntingford & Turner, 1987; Bradbury & Vehrencamp, 1998; Johnstone, 2001). However, very little research has focused on the signalling that occurs at the conclusion of agonistic contests (but see Ch. 10) despite the fact that such signalling may have important consequences for animals involved in such contests and nearby conspecifics. Post-contest signals may be given by the winner or loser of a contest or may occur when there is no clear winner or loser. Such signalling by a winner or loser may be directed towards the other combatant or may be directed to others, such as potential rivals and mates. One type of post-contest signal has been called a ‘victory display’ (e.g. Bradbury & Vehrencamp, 1998). Here, I define a victory display as a display performed by the winner of a contest but not by the loser. In this chapter, I review the known occurrence of victory displays and then use those examples to explore the functional significance of victory displays. I first consider functions within the combatant dyad and then expand the view to consider functions within a communication network. Along the way, I illustrate some difficulties in studying victory displays and suggest areas for further research.

Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

114

P1: JZZ/... P2: JZZ/... 0521823617c06.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:41

The occurrence and function of victory displays Occurrence of victory displays It is not known how common victory displays are because few researchers have studied them or have looked for them, and I am unaware of any published review of these behaviours. Here, I present examples of displays that are plausibly victory displays. These examples come from a literature survey, enquiries to researchers and requests for information on victory displays posted to listservers that focus on behavioural ecology or taxonomic groups. Most potential victory displays discovered were either side notes within papers devoted to other aspects of behaviour or were anecdotes from researchers. I will then explore the displays described here to begin the discussion of how one identifies a victory display. Because of the inefficiency of the methods I was forced to use, I expect that my list of potential victory displays does not represent all the occurrences in the literature, and most certainly the list does not reflect the frequency of its occurrence in the natural world. In the following paragraphs, I begin with the earliest described victory display (in waterfowl), then consider other bird examples before travelling a conventional phylogenetic route from arthropods to humans. Waterfowl

The earliest described victory display is the ceremonie der triumphe in the greylag goose Anser anser (Heinroth, 1910). This display occurs in several contexts, but its use as a victory display occurs when males return to their mates and/or families after ritualized or actual contests. Returning males cackle loudly with their necks extended and wings half-raised. Their mates sometimes join the display, creating what Lorenz (1965a) considered to be ‘the most impressive vocal display of the greylag goose’. Similar displays appear to be common in swans and geese. Examples from the literature include Canada geese Branta canadensis (Radesater, 1974), barnacle geese Branta leucopsis (Bigot et al., 1995) and black swans Cygnus atratus (Kraaijeveld & Mulder, 2002). In each of these four examples, the displays are characterized by acoustic (raucous ‘cackling’ vocalizations) and visual (wing flapping and water splashing) signals that carry far beyond the area in which the contest took place. In particular, exaggerated rolling of the neck and cackling occur simultaneously after a male has won an agonistic contest in the greylag goose (Lorenz, 1965b). In Canada geese, male ‘high intensity cackling’, occurs after contests (Radesater, 1974). Raveling (1967) found that triumph ceremonies were performed by both males following some contests, but that only victors performed ceremonies following severe attacks. In addition, victors gave more prolonged displays after severe attacks than they did after less-severe attacks. Thus, while graded expressions of the triumph ceremony may occur in a variety of behavioural contexts, the exaggerated and simultaneous head rolling and cackling may be the

115

P1: JZZ/... P2: JZZ/... 0521823617c06.xml CU1917B/McGregor

116

0 521 582361 7

April 11, 2005

11:41

J. L. Bower identifying characteristics of a victory display. Furthermore, it may be only after particularly serious contests that the ceremony can be considered a victory display. Other bird groups

In other birds, there are several examples of victory displays. These include parrots of the genus Trichoglossus, which show post-contest behaviours analogous to the triumph display in geese (Serpell, 1981) and the ‘bow flipper spread’ given by winners of contests in the little blue penguin Eudyptula minor (Waas, 1990). In little blue penguins, aggression commonly occurs between unpaired males gathered on non-breeding sites within caves that contain a nesting colony. About 10% of agonistic interactions between males escalate to physical fights, which sometimes caused serious flesh wounds. At the conclusion of such a contest, the winner typically bows forward with his flippers spread and vocalizes while the loser remains stationary or retreats. Some examples involve mainly acoustic signals. For example, a victory display has recently been described in duetting tropical boubou Laniarius aethiopicus (Grafe & Bitz, 2004). In playback experiments involving 26 pairs of boubous, pairs sang one of their 12 shared song types much more often than any other song type following the cessation of playback. This song type was only rarely sung prior to playback or during playback, suggesting that the song type functioned as a postconflict display. Because only presumptive winners sang the duet and not losers, the song appears to be a victory display. This song type had unique signal design features (longer song, higher frequencies, more overlap of male and female notes: Grafe & Bitz, 2004) and was sung from higher perches and carried farther than other song types (T. U. Grafe, unpublished data). This is a particularly striking example of a bird species using a specific song type in a specific behavioural context. A second acoustic example of a victory display occurs in song sparrows Melospiza melodia, in which the winner of a naturally occurring territorial contest (defined as the bird who remains in the contest area after the contest) increases his song rate to match the highest song rates sung prior to territorial contests (Table 6.1; Bower, 2000). During the minute following the conclusion of a contest, the winning bird’s song rate almost always exceeds those of the other dozen or so males in the song sparrow neighbourhood (Table 6.2). Like the tropical boubous, winning song sparrow males typically sang from higher perches after contests ( J. Bower, unpublished data), making them highly conspicuous to their neighbours and suggesting that neighbours may be intended receivers of the victory display. A third example of a victory display with a striking acoustic component is the yodel call of black-throated divers Gavia immer. In a low-density Scottish population, yodels were produced by males that had just successfully defended their loch

P1: JZZ/... P2: JZZ/... 0521823617c06.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:41

The occurrence and function of victory displays Table 6.1. Song rates of winners, losers, neighbours and non-neighbours during the minute before and the minute after the end of territorial contests Songs/min (mean ± SE) Category

Before

After

No.a

p valueb

Winner

3.08 ± 0.99

6.08 ± 0.45

12

0.02∗

Loser

1.17 ± 0.66

2.25 ± 0.63

12

0.26

Unpaired neighbour

4.06 ± 0.59

3.25 ± 2.14

8

0.18

Paired neighbour

1.07 ± 0.41

1.86 ± 0.83

7

0.40

Unpaired non-neighbour

2.32 ± 0.44

2.36 ± 0.40

12

0.86

Paired non-neighbour

0.78 ± 0.25

1.05 ± 0.39

9

0.40

SE, standard error a Sample

sizes vary between tests because not all contests included birds of every category,

but males do not appear more than once in the data. b Differences

were compared with Wilcoxon matched-pairs signed-ranks test. Statistically

significant results are marked with an asterisk.

Table 6.2. Song rates of winners, losers, neighbours and non-neighbours during the minute following the end of territorial contests Category

Songs/min

Category

Songs/min

No.a

p valueb

Winner

6.1 ± 0.5

Loser

2.3 ± 0.6

12

0.002∗

Winner

6.3 ± 0.7

Unpaired neighbour

3.3 ± 0.8

8

0.03∗

Winner

6.0 ± 0.4

Paired neighbour

1.9 ± 0.8

7

0.03∗

Winner

6.1 ± 0.5

Unpaired non-neighbour

2.4 ± 0.4

12

0.02∗

Winner

5.7 ± 0.4

Paired non-neighbour

1.1 ± 0.4

9

0.01∗

a Sample

sizes vary between tests because not all contests included birds of every category.

b Differences

were compared with Wilcoxon matched-pairs signed-ranks test. Statistically

significant results are marked with an asterisk.

against another male, and playback of yodels elicits searching rather than yodels (Gilbert, 1993). Arthropods to humans

Victory displays have been described in several arthropod species. Reichert (1978) described one such display in female funnel-web spiders Agelenopsis aptera, in which the winners and losers of contests over web ownership showed striking differences in stereotyped behaviours after the contest. Females who take over a web, in particular, display behaviours such as biting the web manipulating prey, circling the web or laying new silk. The performance of these behaviours is

117

P1: JZZ/... P2: JZZ/... 0521823617c06.xml CU1917B/McGregor

118

0 521 582361 7

April 11, 2005

11:41

J. L. Bower exaggerated during post-contest periods beyond their non-contest performance. For instance, the spiders exaggerate their abdomen movements when laying silk and walk particularly slowly when circling the web. In crickets and their allies, post-contest stridulation by winners of male–male contest appears to be common in several species, including Gryllus bimaculatus (Alexander, 1961; Simmons, 1986; Adamo & Hoy, 1995), Teleogryllus oceanicus (Burk, 1983) and Acheta domesticus (Hack, 1997). In three species of Australian tree wetas (Hemideina crassidens, H. femorata and H. ricta), winners of male–male contests stridulate after winning contests while losers do not (Field & Rind, 1992; Field, 2001). One species of reptile and one amphibian show what appear to be victory displays. In some lizards, the end of a contest is marked by ritualized positions, which may communicate the status of winners and losers to others. For instance, in the pygmy Mulga monitor Varanus gilleni, individuals who win contests end up atop the loser and attempt to ride the loser until either a new contest occurs or the two separate and the loser leaves the contest area (Carpenter, 1976). In green frogs Rana clamitans, territorial males engage in splashing displays after expelling an intruding male from a territory (Wells, 1978). In mammals, potential victory displays have been recorded for a number of species. Natoli & de Vito (1991) and Natoli et al. (2000) reported that some feral domestic cats Felis catus roll on their back on the ground, exposing their undersides, in front of the contest loser. Males that engaged in this behaviour were highly ranked within the dominance hierarchy of the feral cat groups studied. In wolves Canis lupus and coyotes Canis latrans, winners of contests often run about with their tails held high in the air after winning contests ( J. Way, personal communication). In observations of eastern coyotes at the Stoneham Zoo (Stoneham, MA), Way reported that winners of pinning contests emerge from the contests with a high and bouncing gait while losers stay low to the ground in a submissive pose. Antelope territorial males sometimes engage in a scent-marking behaviour after expelling an intruder from their territory. For instance, hartebeest Alcelaphus buselaphus sometimes add to a boundary dung pile after contests (L. M. Gosling, personal communication). In marine mammals, the term ‘victory squeal’ is used for the vocalization given within a second of a fish being seized by a bottlenose dolphin Tursiops truncatus (S. H. Ridgeway, personal communication) or by a white whale Delphinapterus leucas (Ridgeway & Carder, 1998). Whether this is a victory display by my definition depends on whether the fish was caught in a competitive situation: analogous to goal scoring in human examples. Lastly, it is worth noting that various human behaviours may be considered victory displays. For instance, in ritualized sporting contests, winners often engage in conspicuous displays after goals are scored or victories occur: ice hockey players

P1: JZZ/... P2: JZZ/... 0521823617c06.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:41

The occurrence and function of victory displays raise their sticks high after a goal has been scored and football (soccer) players sometimes take off their shirts and run around the field after scoring a goal. Many American football players engage in stereotyped displays from ‘spiking’ the ball against the ground to displays involving dances that are unique to the individual performing it. At the conclusion of many important team sports events, winning teams climb on top of each other in ways that would seemingly leave them vulnerable to aggression from the other team. On a more serious note, victory parades following armed conflict can be considered a group victory display. It would be fascinating to know of other examples of victory displays following individual or group conflict in either adult or juvenile humans. Functional significance of victory displays Distinctive design features of victory displays

Before addressing the functional significance of victory displays, I must address the problem of determining if a victory display has a unique function and meaning that can be separated from that of signals occurring prior to and during contests. One difficulty in answering this question is that many post-contest signals have similar design features to signals used in other behavioural contexts. For example, wetas and crickets stridulate and birds sing prior to and during contests as well as after contests, leading one to question whether the meaning of a post-contest signal is to communicate the end of a contest or whether it is simply performed in anticipation of the continuation of aggression. One solution to this problem is to examine whether the context for the postcontest signalling differs markedly from contexts in which signals with similar design features are produced. If the sender and potential receivers can determine that the signal is given in the context of the conclusion of a contest, then one can surmise that the signal’s meaning in the post-contest context may be specific to that context. For instance, in song sparrows, the victory display of singing at a high song rate is similar to the high song rate typical of aggressors before they initiate contests (Bower, 2000). However, the context in which this post-contest signal occurs differs from the pre-contest singing in important ways. A major difference is that at the termination of song sparrow contests, the loser leaves the area, ending the almost continual chasing and physical contact that marks a song sparrow territorial contest. Following contests, losers typically remain hidden and quiet for 30 minutes or more after losing ( J. Bower, unpublished data). Thus, the contest winner and surrounding individuals are likely to be aware very soon after the contest ends that he has entered a post-contest period in which further aggression is unlikely to occur in the near future. During most pre-contest periods, both males sit on perches separated by only a few metres and sing. Consequently,

119

P1: JZZ/... P2: JZZ/... 0521823617c06.xml CU1917B/McGregor

120

0 521 582361 7

April 11, 2005

11:41

J. L. Bower the context for the victory display is very different from the context for singing in other stages of interactions. Further research on victory displays may further our understanding of how such signalling differs from signals with similar design features given in other contexts. A second solution to this problem is to test whether the design features of postsignal displays differ, even subtly, from similar signals given in other contexts. For instance, one difficulty in considering the triumph ceremony in geese as a victory display is that the triumph display is composed of a variable and complex mix of discrete displays, which occur in a variety of social contexts. For instance, the rolling of the neck and some degree of cackling occurs even when male greylag geese only feign attacks on other geese and do not actually engage in a contest. However, both the rolling and cackling are most exaggerated and occur simultaneously only after a male has won a real contest (Lorenz, 1965b). Similarly, in Canada geese, male ‘high intensity cackling’ is the part of the multifaceted triumph display that occurs most commonly after territorial aggression (Rades¨ ater, 1974). Therefore, while graded expressions of the triumph ceremony may occur in a variety of behavioural contexts, the exaggerated and simultaneous head rolling and cackling may be specific design features of a victory signal. In other species, behaviours or signal design features may also differ from signals used in different contexts. For instance, winners in song sparrow contests often rise quickly above the thick shrubby vegetation to sing from higher perches than they typically sing from, increasing the active space of their song and the number of potential receivers ( J. Bower, unpublished data). High perches may also facilitate listening (Ch. 3). In playback experiments with the duetting tropical boubou, pairs typically sang one of their 12 shared song-types following the cessation of playback (Grafe & Bitz, 2004). This song type is rarely sung outside of the post-contest context, suggesting that it functions as a victory display. Design features of this song type differ significantly from other song types in several ways. Male and female notes overlap more; male notes reach higher frequencies and the songs are longer in duration than other song types. So, while victory displays may at first appear to be very similar to other signals, with closer inspection one may find that they have specific and unique design features that identify them as victory displays. Functions of victory displays

Victory displays could function within the winner–loser dyad or more widely in the communication network. For example, victory displays could make the victory more memorable (in the sense of Guilford & Dawkins (1991)) to the loser of the dyad, to other receivers in a network or both. It is also possible that the displays have no function and are consequences of mechanisms driving

P1: JZZ/... P2: JZZ/... 0521823617c06.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:41

The occurrence and function of victory displays contests, such as the hormone changes underlying such interactions (e.g. Ch. 21). While recognizing that function and mechanism (ultimate and proximate) levels of explanation are complementary rather than alternatives, the detailed features of victory displays discussed below make it unlikely that they are functionless by-products of recent aggression. Functions within the dyad

If victory displays function within the winner–loser dyad, the winner would be directing the signal to the loser, most likely in an attempt to decrease the probability that the loser will initiate a new contest. By discouraging the loser from starting a new contest, the winner might achieve a lasting victory, thus avoiding the costs of further contests and creating the opportunity to resume other activities, such as searching for a mate, maintaining a pair bond, vigilance against other intruders and predators, and feeding and other maintenance tasks. The examples described above seem to fit into one of two categories. First, there are victory displays that seem to invite an extension of the contest. For instance, in the bow flipper spread display in little blue penguins the winner bows low the ground and spreads his flippers apart. This appears to put the winner in a position where he is vulnerable to attack from the loser. Likewise, feral cats that roll over on their back and expose their undersides would seem to be choosing a position that is vulnerable to further attacks. It is possible that by providing the loser with a stimulus for attacking the winner just after the loser has retreated from the contest, the winner helps to crystallize dominance over the loser. At a mechanistic level, it is possible that future displays by the winner similar to the victory display used after a contest victory may result in a change of mental state in the loser. Such associative learning (e.g. Staddon, 1983), in which the loser associates losing the contest with the signal used during the winner’s victory display, may function to discourage a contest loser from initiating future contests with the winner. Second, there are victory displays in which the winner may give a display that is energetically or otherwise costly. For instance, in my song sparrow study (Bower, 2000), winners almost always sang at very high rates in the minutes following a contest. Since singing is a moderately energetically costly behaviour (Oberweger & Goller, 2001), the ability to sing at high rates immediately following a contest may advertise the winner’s vigour or quality. This should be especially true since song sparrow contests are energetically costly, often characterized by almost continuous chasing and occasional physical fighting for an hour or more, with no breaks for feeding or resting (Bower, 2000). By singing at high rates just after a contest ends, a male may be sending an honest signal to the loser that the winner has sufficient endurance to defend the area he has just secured from subsequent

121

P1: JZZ/... P2: JZZ/... 0521823617c06.xml CU1917B/McGregor

122

0 521 582361 7

April 11, 2005

11:41

J. L. Bower challenges, despite the energy costs of the recent contest. Other examples of possible victory displays that may be metabolically or otherwise costly include postvictory stridulation in crickets and wetas, circling the web and laying new silk in funnel-web spiders, splashing displays in frogs and the bouncing gait of eastern coyotes. As with song sparrows, while victory displays may be relatively short in duration, and thus not likely to be more than moderately energetically costly, their occurrence at the end of often long and energetically costly contests may make them difficult to perform. Consequently, they may function as honest indicators of aspects of the winner’s quality (for further discussion of handicaps and honest signalling, see Dawkins (1995)). The ability to perform such a display following a contest may reduce the chances that the contest loser will re-engage the winner in a later contest. Functions within the network

Theory suggests that animals are likely to gain fitness benefits by assessing potential mates and rivals in contest situations and altering their behaviour according to their assessments (e.g. Cox & Le Boeuf, 1977; McGregor & Dabelsteen, 1996; McGregor & Peake, 2000). If this were so, then one would expect selection pressure for winners to advertise their victory to other members of a communication network. Victory displays may inform social eavesdroppers (Ch. 2) and other members of the network that did not pay attention to the interaction that the winner has just won a contest. As discussed above, the displays may also provide further information about the winner’s vigour and/or other measures of quality by displaying after energetically costly contests. This possible function of victory displays is distinct from audience effects (Ch. 4) and considerations of the nature of signals and signalling during interactions (e.g. private signals (Ch. 3) and the ‘good loser’ hypothesis (Peake & McGregor, 2004)). However, no studies have been attempted that test whether performing a victory display modifies the behaviour of network members. At present, investigating the network function of victory displays has to rely on indirect evidence. One obvious criterion, of course, is whether conspecifics other than (or as well as) the loser are able to receive the signal (for similar discussion on the intended receivers for post-copulatory displays in ducks, see Johnson et al. (2000)). This depends both on the signal features and the spatial arrangement of the conspecifics. For instance, of the examples given above, it is least likely that funnel-web spider post-victory behaviour has a communication network function, since the widely spaced webs (Reichert, 1978) and reliance on vibration for communication means that rivals are unlikely to be aware of contests on other webs. Likewise, since varanid lizard territories tend to be large (R. Earley, personal communication) and the possible victory display is a visual signal, it is unlikely that rivals in adjoining territories

P1: JZZ/... P2: JZZ/... 0521823617c06.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:41

The occurrence and function of victory displays will witness a contest between two lizards. Therefore, the pygmy Mulga monitor’s ‘back riding’ post-contest behaviour is more likely directed to the loser rather than for another potential rival. However, it is not known how often females might witness such contests, nor how often other males encroaching on the territory witness the contest. Therefore, even in territorial species in which territories are large and signals have relatively short ranges, there may be a network function if other animals are present on territories. In contrast to these two examples, the other victory displays described above occur in species in which potential mates and rivals are likely to witness the display. Territorial songbirds, for instance, often reside in neighbourhoods, with several territories constituting a neighbourhood. During the song sparrow territorial contests I studied, I noticed that neighbourhood males and females often flew up to perches from which they could see the contest ( J. Bower, unpublished data). Such behaviour seemed to indicate that members of the communication network were paying close attention to contests in their neighbourhood. Bird song is a moderately long-range signal, typically carrying into and beyond neighbouring territories (e.g. Brenowitz, 1982; Ch. 20). Therefore, the post-contest vocalizations of a winning song sparrow, tropical boubou pair or goose would likely be heard by other conspecifics in the bird’s neighbourhood. Similarly, stridulating crickets and wetas, as well as splashing green frogs, all produce acoustic signals that are likely to be accessible to rivals and potential mates. Visual signals made within open habitats or by species with close spacing may also be candidates for victory displays directed at the network. Thus, post-contest scent marking by hartebeest, the movements of geese and the bouncing gait of eastern coyotes all are accessible to conspecific receivers in the communication network.

Summary Victory displays are post-conflict signals given by the winner (but not the loser) of an agonistic contest. There has been little work specifically addressing such displays and most of the evidence for their existence comes from incidental descriptions or asides and footnotes. Much remains to be done to characterize victory displays and to identify their function, including whether they are network phenomena or are directed at the loser of the interaction. The difficult but exciting work that lies ahead is to demonstrate that such displays, in conjunction with or independent from the social eavesdropping that may occur during a contest, alter the behaviour of other members of the communication network. It is likely that such a test could be developed more easily in a laboratory setting than in situ; one approach would be to prevent observers from watching the contest but allow them to observe the following victory display.

123

P1: JZZ/... P2: JZZ/... 0521823617c06.xml CU1917B/McGregor

124

0 521 582361 7

April 11, 2005

11:41

J. L. Bower Acknowledgements I thank the many researchers who responded to my call for potential examples of victory displays, Ulmar Grafe for sharing unpublished data, and three anonymous referees and Peter McGregor for comments on previous drafts of this chapter.

References Adamo, S. & Hoy, R. 1995. Agonistic behaviour in male and female field crickets, Gryllus bimaculatus, and how behavioural context influences its expression. Animal Behaviour, 49, 1491–1501. Alexander, R. 1961. Aggressiveness, territoriality, and sexual behaviour in field crickets (Orthoptera gryllidae). Behaviour, 17, 130–223. Bigot, E., Hausberger, M. & Black, J. 1995. Exuberant youth: The example of triumph ceremonies in barnacle geese (Branta leucopsis). Ethology, Ecology and Evolution, 7, 79–85. Bower, J. L. 2000. Acoustic interactions during naturally occurring territorial conflict in a song sparrow (Melospiza melodia) neighborhood. Ph.D. Thesis, Cornell University, Ithaca, New York. Bradbury, J. W. & Vehrencamp, S. L. 1998. Principles of Animal Communication. Sunderland, MA: Sinauer. Brenowitz, E. A. 1982. Long-range communication of species identity by song in the red-winged blackbird. Behavioral Ecology and Sociobiology, 10, 29–38. Burk, T. 1983. Male aggression and female choice in a field cricket (Teleogryllus oceanicus): the importance of courtship song. In: Orthopteran Mating Systems: Sexual Competition in a Diverse Group of Insects, ed. D. T. Gwynne & G. K. Morris. Boulder, CO: Westview Press, pp. 97–119. Carpenter, C. C. 1976. A further analysis of the combat ritual of the pygmy mulga monitor, Varanus gilleni (Reptilia: Varanidae). Herpetologica, 32, 35–40. Cox, C. R. & Le Boeuf, B. J. 1977. Female incitation of male competition: a mechanism for mate selection. American Naturalist, 111, 317–335. Dawkins, M. S. 1995. Unravelling Animal Behaviour, 2nd edn. Harlow, UK: Longman. Field, L. H. 2001. Aggression behaviour in New Zealand tree wetas. In: The Biology of Wetas, King Crickets, and Their Allies, ed. L. H. Field & T. H. Jarman. Oxford: CAB International, pp. 333–350. Field, L. H. & Rind, F. 1992. Stridulatory behaviour in a New Zealand weta, Hemideina crassidens. Journal of Zoology, 228, 371–394. Gilbert, G. 1993. Vocal individuality as a census and monitoring tool: practical considerations illustrated by a study of the bittern Botaurus stellaris and the black-throated Diver Gavia arctica. Ph.D. Thesis, University of Nottingham, UK. Grafe, T. U. & Bitz, J. H. 2004. An acoustic victory display in the duetting tropical boubou (Laniarius aethiopicus): a signal of victory. BMC Ecology, 4, 1. http://www.biomedcentral.com/bmcecol/. Guilford, T. & Dawkins, M. S. 1991. Receiver psychology and the evolution of animal signals. Animal Behaviour, 42, 1–14.

P1: JZZ/... P2: JZZ/... 0521823617c06.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:41

The occurrence and function of victory displays Hack, M. A. 1997. The energetic costs of fighting in the house cricket, Acheta domesticus L. Behavioral Ecology, 8, 28–36. Heinroth, O. 1910. Beitrage zur Biologie, namentlich Ethologie und Psychologie der Anatiden. In: Verhandlungen der V Internationalis Ornithologen Kongressus in Berlin, pp. 589–702. Huntingford, F. & Turner, A. K. 1987. Animal Conflict. Cambridge, UK: Cambridge University Press. Johnson, K. P., McKinney, F., Wilson, R. & Sorenson, M. D. 2000. The evolution of postcopulatory displays in dabbling ducks (Anatini): a phylogenetic perspective. Animal Behaviour, 59, 953–963. Johnstone, R. A. 2001 Eavesdropping and animal conflict. Proceedings of the National Academy of Sciences, USA, 98, 9177–9180. Kraaijeveld, K. & Mulder, R. 2002. The function of triumph ceremonies in the black swan. Behaviour, 139, 45–54. Lorenz, K. 1965a. Here I Am – Where Are You?: The Behavior of the Greyleg Goose. New York: Harcourt Brace. 1965b. The triumph ceremony of the greylag goose. Philosophical Transactions of the Royal Society of London, Series B, 251, 477–478. Maynard Smith, J. 1982. Evolution and the Theory of Games. Cambridge, UK: Cambridge University Press. McGregor, P. K. & Dabelsteen, T. 1996. Communication networks. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca, NY: Cornell University Press, pp. 409–425. McGregor, P. K. & Peake, T. M. 2000. Communication networks: social environments for receiving and signalling behaviour. Acta Ethologica, 2, 71–81. Natoli, E. & de Vito, E. 1991. Agonistic behaviour, dominance rank and copulatory success in a large multi-male feral cat, Felis catus L., colony in central Rome. Animal Behaviour, 42, 227–241. Natoli, E., De Vito, E. & Pontier, D. 2000. Mate Choice in the domestic cat (Felis silvestris catus L.). Aggressive Behavior, 26, 455–465. Oberweger, K. & Goller, F. 2001. The metabolic cost of bird song production. Journal of Experimental Biology, 204, 3379–3388. Peake, T. M. & McGregor, P. K. 2004. Information and aggression in fishes. Learning and Behavior, 32, 114–121. Rades¨ ater, T. 1974. Form and sequential associations between the triumph ceremony and other behaviour patterns in the Canada Goose Branta canadensis L. Ornis Scandinavica, 5, 87–101. Raveling, D. G. 1967. Sociobiology and ecology of Canada Geese in winter. Ph.D. Thesis, Southern Illinois University, Carbondale, USA. Reichert, S. E. 1978. Games spiders play: behavioral variability in territorial disputes. Behavioral Ecology and Sociobiology, 3, 135–162. Ridgeway, S. H. & Carder, D. A. 1998. Net-aided foraging by two white whales. Marine Mammal Science, 14, 332–334. Serpell, J. 1981. Duets, greetings and triumph ceremonies: analogous displays in the parrot genus Trichoglossus. Zeitschrift f¨ ur Tierpsychologie, 55, 268–283.

125

P1: JZZ/... P2: JZZ/... 0521823617c06.xml CU1917B/McGregor

126

0 521 582361 7

April 11, 2005

11:41

J. L. Bower Simmons, L. 1986. Inter-male competition and mating success in the field cricket, Gryllus bimaculatus (De Geer). Animal Behaviour, 34, 567–579. Staddon, J. E. R. 1983. Adaptive Behavior and Learning. Cambridge, UK: Cambridge University Press. Waas, J. R. 1990. An analysis of communication during the aggressive interactions of little blue penguins (Eudyptula minor). In: Penguin Biology, ed. LS Davis & JT Darby. San Diego, CA: Academic Press, pp. 345–376. Wells, K. D. 1978. Territoriality in the green frog (Rana clamitans): vocalizations and agonistic behaviour. Animal Behaviour, 26, 1051–1063.

P1: JZZ/... P2: JZZ/... 0521823617agg.xml CU1917B/McGregor

0 521 582361 7

Part II T H E E F F E C T S O F P A R T I C U L A R CONTEXTS

127

April 11, 2005

10:51

P1: JZZ/... P2: JZZ/... 0521823617agg.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

128

10:51

P1: JZZ/... P2: JZZ/... 0521823617agg.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

10:51

Introduction

The rationale behind the grouping of chapters into this section is to facilitate comparisons between communication networks found in very different contexts: mate choice, predation, nestling begging, redirection and scent marking. One of the attractions of communication networks is that the idea applies to any context in which the signals used travel far enough to encompass several other individuals. However, each context will have distinctive features affecting the nature of the information transmitted, the signals used and their travelling power; therefore, the nature of the communication network may differ. Comparison of networks found in different contexts could, therefore, advance our understanding of the topic.

Mate choice It is probably a fair generalization to say that in recent years the most widely considered, modelled and experimented upon context for communication has been the simplest mate choice situation, i.e. that involving a male signaller and a female receiver. However, Ken Otter and Laurene Ratcliffe point out in Ch. 7 that a communication network is a more likely context because females have access to the widely broadcast mate attraction signals of several males. This chapter discusses which traits females in a communication network use when choosing between males: both as pair mates and as extra-pair partners. It also discusses the way choice is achieved (e.g. simultaneous versus sequential assessment) and how sampling by females can be inferred from the pattern of movement through a network of signalling males. Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

129

P1: JZZ/... P2: JZZ/... 0521823617agg.xml CU1917B/McGregor

130

0 521 582361 7

April 11, 2005

10:51

Part II Predation Communication networks are not restricted to individuals of the same species; predators have long been recognized to be the unwanted guests at communication feasts, with long-range signals advertising the location of potential prey. The commonest response of choruses of insects and anurans to the detection of a predator is to cease calling (Gerhardt & Huber, 2002). This may be an efficient way to avoid predators, but it is an inefficient way to communicate with the intended receivers of the signals, such as potential mates. In Ch. 8, Alexander Lang, Ingeborg Teppner, Manfred Hartbauer & Heiner R¨ omer explain how pseudophylline katydids can to some extent overcome this problem by signalling with vegetation-borne vibrations (tremulations) rather than airborne sounds, because tremulations cannot be detected by passive listening bat predators. While such bat predators are an important selection pressure on katydids, they are not the only ones. This chapter uses neurophysiological preparations in the field and decisiontree learning algorithms to investigate how katydids communicate in a noisy rainforest environment.

Nestling begging In birds, the begging of nestlings has come to rival mate choice as a model system for the study of the evolution of biological signalling (Wright & Leonard, 2002). However, in parallel with most of the work on mate choice, begging is considered as a dyad, with a nestling (or the brood collectively) as the signaller and one parent as the receiver. This seems odd, perverse even, given the close proximity of nestlings (both to nestmates and to their parents) and the conspicuous vocal and visual signals that nestlings produce. In Ch. 9, Andy Horn and Marty Leonard show how considering begging as a communication network can yield new insights into begging behaviour. Furthermore, issues for communication networks, in general, are raised by the close proximity of individuals and the possibility of direct physical action in a crowded nest. Such issues will be particularly relevant to many highly social species such as social hymenoptera.

Redirection of aggression The nature of communication in aggressive encounters has a long history of study, from Darwin’s antithesis principle (Darwin, 1872) and Lorenz’s classics King Solomon’s Ring and On Aggression (Lorenz, 1952, 1966) onwards. Most attention has focused on signal exchanges before and during aggressive encounters. However, redirection is a puzzling behaviour performed by losers after an aggressive contest. As the target of redirection is an individual other than the winner

P1: JZZ/... P2: JZZ/... 0521823617agg.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

10:51

The effects of particular contexts (commonly a lower-ranking individual in social groups), it extends even dyadic interactions into a network. Ani Kazem and Filippo Aureli discuss explanations for redirection in Ch. 10, focusing on primates where it has been best described. They conclude that redirection is best interpreted from a network viewpoint, in terms of how it can influence the behaviour of bystanders.

Scent marking Scent marking is a rather different context from the others considered in this section. Whereas the other contexts could be loosely considered to be aspects of social behaviour, scent marking is a particularly distinct aspect of chemical communication and can be involved in several social contexts. Scent marking is also rather different from most other signals because scent marks persist, often for considerable periods, in the absence of the signaller. Any conspecific visiting the scent mark can obtain information from it; in this respect, such marks could be considered analogous to public noticeboards. In Ch. 11, Jane Hurst considers the selection pressures that result from such undirected and long-lasting signals. Most of her examples come from studies of mice, where many studies have examined the behavioural and biochemical basis of scent communication.

Future directions A common theme of the chapters in this section (and indeed throughout this book) is that communication needs to be considered in a more complex way in order to make progress. Far from being a standard recourse to complexities of the real world when simple explanations fail (and being even further from a counsel of despair), these chapters demonstrate how adopting a network perspective can explain troublesome aspects of communication and also indicate directions for future research. One future challenge in mate choice is to characterize female assessment behaviour, because in many instances potential mates can be assessed at long range, with close approach possibly representing the outcome of choice rather than assessment in action. Scent marks seem to offer an opportunity for such characterization because close approach and perhaps contact are required to gather information from scent marks. Video tracking individuals in a naturalistic enclosure may relatively easily provide data on patterns of visits to scent marks and, by extension, on information gathering. Many communication behaviours including mate choice are likely to be constrained by the presence of predators, as the chapter on katydids demonstrates. In such circumstances, information gathering could be a costly exercise if proximity to signallers increases the risk of being preyed upon. Similarly, being a bystander

131

P1: JZZ/... P2: JZZ/... 0521823617agg.xml CU1917B/McGregor

132

0 521 582361 7

April 11, 2005

10:51

Part II at an aggressive interaction could be costly if the bystander becomes the target of redirected aggression. These examples indicate that animals may have to trade-off several somewhat disparate costs and benefits when gathering information in a communication network. As the costs and benefits are likely to change diurnally and seasonally as well as on much shorter timescales, these trade-offs may be best explored by modelling. Many of the chapters have suggested or implied that more detailed empirical studies, both experimental and observational, are needed to further our understanding. In some sense, this is always going to be true because of the enormous variation in biological systems; however, one value of a communication network perspective is that it suggests what types of information (e.g. mate choice assessment patterns, variation in the form of begging calls) would allow the field to develop. The chapters in this part demonstrate the advantages of a broader view of communication over and above the advantages of a network view: understanding of communication in any particular context can come from contexts other than that under immediate consideration. For example, nestling begging would seem to have little to offer contexts such as mate choice, aggression and resource defence, because the detailed circumstances of the contexts are so different. However, begging behaviour focuses attention on issues that are fundamental to all three contexts and indeed to communication in general, such as the distinction between signals and physical action. Similarly, several explanations for redirection behaviour operate over a timescale encompassing a sequence of several contests, suggesting that some aspects of audience effects and eavesdropping would benefit from consideration over such longer timescales. Difficult as it may be with an everexpanding communication literature, it would seem a good idea to keep an eye on developments in several communication contexts. Because communication networks can be seen to apply to several contexts, adopting a network perspective helps to promote such a breadth of interest.

References Darwin, C. 1872. The Expression of the Emotions in Man and Animals. London: John Murray. Gerhardt, H. C. & Huber, F. 2002. Acoustic Communication in Insects and Anurans: Common Problems and Diverse Solutions. Chicago, IL: Chicago University Press. Lorenz, K. Z. 1952. King Solomon’s Ring; New Light on Animal Ways. New York: Crowell. 1966. On Aggression. New York: Harcourt, Brace and World. Wright, J. & Leonard, M. L. 2002. The Evolution of Begging: Competition, Cooperation, and Communication. Dordrecht: Kluwer.

P1: IYP/... P2: JZZ/... 0521823617c07.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:49

7

Enlightened decisions: female assessment and communication networks ken a. o tter1 & l aurene ratcliffe2 1 2

University of Northern British Columbia, Prince George, Canada Queen’s University, Kingston, Canada

Introduction Asymmetry in parental investment often predicts that females should be choosier about prospective mates than males. It is commonly assumed that females assess male characteristics during mate choice, but which traits are assessed, and how they influence female decision making, is not well understood. Current models of mate choice suggest females may sequentially sample a pool of males, memorizing levels of trait expression among comparison males, or else accept the first male that exceeds some minimum threshold value of mate quality. Recent tests of communication network theory suggest that these models may have to be revised because females can tap into advertising signals broadcast in a network fashion. Such behaviour could reduce costs of mate searching, as signals are perceived simultaneously, allowing instantaneous relative comparisons. In this chapter, we explore the potential of females to extract comparative information on the relative quality of males for use in reproductive decision making. We focus primarily on primary mate choice decisions (i.e. initial selection of a mating partner) and secondary mate choice decisions (i.e. mating decisions that arise after social pairing, which may include extra-pair copulations or ‘divorce’ of the current mate to pair with another male) based on acoustic signals in territorial passerines; however, the ideas that we present should be applicable to other taxa and other sensory modalities. Finally, we discuss the potential impacts of habitat alteration on females’ abilities to use network assessments for mate choice.

Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

133

P1: IYP/... P2: JZZ/... 0521823617c07.xml CU1917B/McGregor

134

0 521 582361 7

April 11, 2005

11:49

K. A. Otter & L. Ratcliffe In the early 1990s, it was realised that bird song and other animal communication took place in a more extensive (network) context than male–female or male–male dyads; several receivers may perceive each signaller simultaneously and, conversely, signallers may direct their signal towards several receivers (McGregor, 1993; McGregor & Dabelsteen, 1996). The idea that females seeking mates may be able to assess several males simultaneously, not only on leks but also in situations where males defend more dispersed territories, complicates the study of female mate choice. For example, we need to rethink how females may be sampling males, especially during initial territorial settlement and any subsequent pursuit of extra-pair copulations. Communication network theory challenges us to reconsider traditional models of mate assessment. Constraints on signal transmission and reception may limit the spread of information to receiving parties. Such constraints should bias assessment to particular kinds of signal and may also help to explain why some signals important in initial mate choice need not necessarily correlate with secondary mate choice (e.g. red plumage in house finches, Carpodacus mexicanus, is selected by females in initial mate choice but does not seem to affect extra-pair success: Hill, 2002). We must also consider the kinds of signal that are received by females; how habitat alteration might influence communication in a network, and how, in turn, that might influence reproductive success. Deciphering which attributes of males are of most interest to females has proved to be a challenging task; a common approach has been to ask whether females pick superior males as partners, where superiority is defined by the expression of condition-dependent signals. In birds, considerable evidence from dyadic mate choice studies suggests that colour and vocal cues believed to be energetically expensive are correlated with mate selection (reviewed by Gil & Gahr, 2002). Whether such dyadic studies accurately reflect female assessment in natural circumstances is still unclear. Considerations of communication in a network context have stimulated new experimental approaches that seek to determine how males evaluate potential rivals (e.g. Naguib et al., 1999; Todt & Naguib, 2000; Ch. 2) and similar types of study may help researchers to decipher how females evaluate potential mates. This chapter reviews data on which traits appear to be important in female choice in songbirds and then suggests ways to model and test female choice using communication network theory. Although work in this field is still limited, we review published and in-progress studies that discuss how signals might spread in networks and how females might assess such information. Finally, we discuss how habitat alteration can affect the propagation of signals, and how this influences the ability of females to assess males in communication networks.

P1: IYP/... P2: JZZ/... 0521823617c07.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:49

Enlightened decisions: female assessment How do females assess males? Traditional models of primary female mate choice assume that females assess males sequentially ( Janetos, 1980; Wittenberger, 1983). In choosing social mates, females move from one male to the next and select mates via either sequential comparison of the current male versus the last male visited, or via a best-of-n model, where females sample all males and then return to the individual of highest quality. Although less-formally modelled, most studies on secondary mate choice assume similar kinds of decision-making strategy; females select a social partner then assess this male relative to other available males to determine whether to engage in extra-pair copulations or divorce (e.g. the ‘better options’ model of divorce (Ens et al., 1993), in which birds select a social mate then assess opportunities to desert and pair with a male of better quality, or similar strategies in the ‘genetic benefits’ models of extra-pair copulation (Kempenaers & Dhondt, 1993)). Recent models of primary mate choice are more realistic because they incorporate costs of mate searching (Real, 1990; Wiegmann et al., 1996) but still assume that males are assessed sequentially (e.g. Fig. 7.1a). However, the signals used by females to assess males may propagate sufficient distance in some circumstances for males to be assessed simultaneously (Gibson & Langen, 1996). In communication networks, female receivers can simultaneously detect the signals of several territorial males. For example, a female moving from one male to the next may still be able to detect the signals of males she has visited previously. Thus, females may be able to assess the relative expression of several males’ signals simultaneously, without having to rely on memory of absolute expression. Females could continue to search in such a fashion until no new male exceeds a preceding male. However, female searching may not be even this constrained. If females use a best-of-n strategy, they need not rely solely on memory of trait expression of each male. All they need do is remember the territory locations of males of perceived high quality and position themselves in a manner that allows simultaneous comparison. If signal transmission is sufficiently long range, females may be able to assess and eliminate a number of males without even closely approaching them, as described in anuran mating aggregations (e.g. Murphy & Gerhardt, 2002; see below). Females positioning themselves strategically within networks and making choices on relative trait expression could decrease search time, maintain safe distances from territory owners and reduce the chance of mistaken decisions (Fig. 7.1b). Field studies typically use female movement patterns to assess the number of males that are sampled by females. Some studies have used radio telemetry to track female movement and have assumed that close approach is evidence

135

P1: IYP/... P2: JZZ/... 0521823617c07.xml CU1917B/McGregor

A

April 11, 2005

A

B iii

D

0 521 582361 7

C

B C

ii

iv

D

E

E

i

(a)

(b)

A

A

B C

B C

D

D E

(c)

11:49

E

(d) Fig. 7.1. Sampling of males by females is often assumed when a female closely approaches a male; however, if females assess males by long-range signals the relationship between sampling and close approach may be rather different: compare (a) with (b), and (c) with (d). Males are represented by capital letters (A–E) and the female’s path is shown as a dashed line. In (a) and (c), dark lines represent their territorial boundaries. In (b) and (d) these lines are shown in grey so as not to obscure the dotted lines representing the range of effective signal transmission. (a) The female travels along path i–iv approaching males E, C, D and then C, with whom the female finally settles. In this scenario, the female may be considered to be using a best-of-n sampling method, having sampled males E, C and D but not A and B. (b) If we consider female sampling in relation to the signal transmission range, we see that the female’s movement may, in fact, also allow her to sample males A and B (i.e. her path lies within their effective signal range) without ever approaching them directly. During the period that she is in the territory of male C, the female could hypothetically sample the signals of males B, A, E and D, and she would be within transmission range of two to three males at any given time along her whole route. (c) In this scenario, the female appears to avoid close contact with all males until approaching male A. This could be interpreted as the female not sampling any of the males prior to making a mating decision, but considering the situation in relation to signal range (d) reveals that the female would be able to assess males E, C and B by their signals en route to male A. Moreover, the female’s movements place her within the signal range of at least two males at any point along her path, allowing her potentially to compare males simultaneously, as well as comparing each male newly encountered with the last male along her pathway. If male A exceeds the traits of the other males, this female may be interpreted as adopting either a threshold-style model or a best-of-n model in mate sampling.

136

P1: IYP/... P2: JZZ/... 0521823617c07.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:49

Enlightened decisions: female assessment of assessment of a signalling male (Bensch & Hasselquist, 1992; Neudorf et al., 1997). The pied flycatcher Ficedula hypoleuca offers an extreme example of such movement, as the nestbox is an assessed resource (Slagsvold et al., 1988; Dale et al., 1990) and assessment requires close inspection, in much the same way that a short-range signal would. However, movement patterns can be difficult to interpret. For example, a female moving along the boundary between territorial males, apparently undetected by them, may still be sampling these males (Fig. 7.1c). If the males are producing long-range signals, then the female could sample males based on this signalling network (Fig. 7.1d). Her movement pattern allows at least two males to be assessed at almost all points along her route. By such surreptitious sampling (i.e. moving silently and apparently remaining undetected by males along the route: Neudorf et al., 1997), females may avoid some of the costs associated with close approach to males. These costs could be harassment from males or aggression from mated females (e.g. Dale & Slagsvold, 1995). After surreptitious sampling, the female may then closely approach the male she has selected. In such a scenario, close approach indicates choice rather than sampling and the amount of sampling is underestimated (in Fig. 7.1d four males have been sampled rather than one). Murphy & Gerhardt (2002) described such a scenario in female barking treefrogs Hyla gratiosa in the field; females approach only a single male in a chorus, suggesting no sampling has occurred and that the first male encountered is selected. However, further anecdotal evidence suggested that females may be assessing several males at a distance and then approaching only the selected male (Murphy & Gerhardt, 2002). Therefore, we urge caution when patterns of female movement are used to infer sampling behaviour. We suggest that it may be more appropriate to determine female movement in relation to the transmission distance of signals used in mate choice. Such considerations will allow us to determine whether females are strategically placing themselves in areas that maximize the number of males that can be simultaneously assessed while concurrently minimizing search costs.

What are females looking for? Research on mate choice since the early 1990s has focused on females selecting males based on perceived quality. But how are such distinctions made? Females are presumably unable to assess male genetic quality directly but can infer this through assessment of traits that tightly correlate to resource-holding potential of the male (Grafen, 1990). Visual signals from plumage are known to be associated with male condition and ability to acquire resources (Hill, 2002). Dominance status and aggressive behaviour also reflect relative male condition, as they predict access to limited

137

P1: IYP/... P2: JZZ/... 0521823617c07.xml CU1917B/McGregor

138

0 521 582361 7

April 11, 2005

11:49

K. A. Otter & L. Ratcliffe resources (Ekman & Askenmo, 1984; Hogstad, 1988; Desrochers, 1989; Ficken et al., 1990; Ekman & Lilliendahl, 1992). Singing behaviour is also indicative of male condition. Repertoire size in several species is related to age (Yasukawa et al., 1980; Lampe & Espmark, 1994; Birkhead et al., 1997; Eens, 1997), providing information on individual survival. Singing behaviour may also reflect a male’s ability to secure access to limited resources (Reid, 1987; Alatalo et al., 1990; Thomas, 1999a,b; Thomas & Cuthill, 2002). Song output is known to be associated with a male’s dominance rank (Otter et al., 1997), level of parasite infestation (Møller, 1991) and immune response (Saino et al., 1997a). Even the fine structure of song may give cues to the survivorship (Forstmeier et al., 2002) or rank (Christie et al., 2004) of males. Many studies have shown that these behavioural and morphological characteristics are important in female choice, both for primary mates (Rades¨ ater et al., 1987; Alatalo et al., 1990; Andersson, 1994; Hoi-Leitner et al., 1995; Buchanan & Catchpole, 1997) and for extra-pair paternity (Smith, 1988; Morton et al., 1990; Houtman, 1992; Wetton et al., 1995; Hasselquist et al., 1996; Kempenaers et al., 1997; Saino et al., 1997b; Møller et al., 1998; Otter et al., 1998; Forstmeier et al., 2002). Which trait is the best indicator of quality and what is meant by a good indicator of quality? Strong correlations between male quality and expression of the trait are assumed for the trait to be ‘reliable’ (Grafen, 1990) and several conditiondependent traits will often be intercorrelated. To determine which of these signals are likely to be assessed by females, however, we should focus on the perception of signals by females rather than the production of signals by males. It does not necessarily follow that a female will be able to discriminate amongst males even if male quality is correlated with the expression of the trait. The traits may all potentially be reliable, but the important question for females in networks is at what distance are they detectable and discriminable (Ch. 20)? The debate on the evolution of multiple signals (Møller & Pomiankowski, 1993; Pomiankowski & Iwasa, 1993; Iwasa & Pomiankowski, 1994; Johnstone, 1996) has focused largely on whether multiple signals are of use to the female in an additive way, or whether their apparent redundancy is used by females to confirm their assessment. The debate assumes that females are able to assess all traits simultaneously. It seems more likely that, in a natural network context, signals that target different sensory modalities may be assessed sequentially in relation to their transmission distance. Distance can have a profound effect on both detection and discrimination (Wiley & Richards, 1982). Many morphological and behavioural traits that require visual inspection can only be discriminated at close range, particularly in habitats with dense vegetation. Tactile signals may be similarly restricted. In contrast, olfactory or auditory signals, particularly song, have evolved to transmit at least the average interterritory spacing within a species (Brenowitz, 1982; Calder, 1990);

P1: IYP/... P2: JZZ/... 0521823617c07.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:49

Enlightened decisions: female assessment thus, detection and discrimination are possible at greater range. Long-range signals may form the basis of initial assessment, which can then be confirmed on closer inspection by assessment of short-range signals. The apparent redundancy of intercorrelated signals may thus reflect the hierarchical way in which they are assessed (Bradbury & Vehrencamp, 1998). Imagine a situation where a female is assessing a prospective male, either as a mate or as an extra-pair partner. In many cases, female movement may be constrained by such factors as mate guarding, aggression of nearby mated females (Slagsvold & Lifjeld, 1994; Dale & Slagsvold, 1995) or risk of predation. Under these circumstances, any mechanism that enables females to narrow down the pool of acceptable males from a position of relative safety would be favoured (Gowaty, 1996). In songbirds, male advertising song provides an ideal signal for assessment at a distance (Fig. 7.2a–c), because of its long transmission range. Initial decisions about males can be made via this single cue; females can then directly approach subsets of males deemed to be of the best quality among the available pool. Further discrimination may then occur by shifting assessment to short-range signals, such as plumage, the expression of which we would expect to correlate with quality indicated by long-range signals. This sequential assessment of signals may increase the certainty of assessment. Females may employ information derived from networks not only to narrow the pool of potential mates; females may also instigate network communication to evaluate male quality during close approach. For example, during intrusions across territorial boundaries, the attraction of neighbouring males may incite competitive interactions between a female’s mate and his neighbours (Fig. 7.2d–f). Montgomerie & Thornhill (1989) suggested that females might incite interactions between males as a mechanism for sperm competition, but it is also possible that such behaviour provides females with more information about the general quality of available males (Sæther, 2002). This possibility is also supported by demonstrations that females use information from male–male singing interactions (i.e. they eavesdrop; Ch. 2) in extra-pair behaviour decisions (Otter et al., 1999; Mennill et al., 2002). Future work on this topic should target species where females readily engage in secondary mate choice, for example the pursuit of extra-pair copulations. Recently, Sæther (2002) demonstrated that such incitement occurs in the great snipe Gallinago media, a species in which females call from the edges of males’ territories within leks. Using playbacks, he showed that female calls from such boundaries increase the competitive interactions between neighbouring males, providing a potential source of information to prospecting females. Similarly, in territorial songbirds, females may exhibit behaviour that draws their mates and neighbours into interactions on territorial boundaries. Ramsay et al. (1999)

139

P1: IYP/... P2: JZZ/... 0521823617c07.xml CU1917B/McGregor

140

0 521 582361 7

April 11, 2005

11:49

K. A. Otter & L. Ratcliffe

(a)

(b)

(c)

(d)

(e)

(f)

Fig. 7.2. Secondary mate choice, such as the decision to engage in extra-pair copulations or divorce a current mate for another male, may be constrained in some territorial species. Unlike primary (i.e. initial) mate choice, females may not be able to move freely through the territories of various males because of the presence of resident females (a). However, long-range signals of males, such as song, transmit beyond the boundaries of the territories and effectively signal presence to neighbouring territories (b). In (a, b, d and e), dark lines represent territorial boundaries with males and females shown as symbols. Dotted lines represent the range of effective signal transmission and some territory boundaries are show in grey in (c) and (f ) to prevent obscuring signal ranges. (c) A female may be able to assess all neighbouring males as well as her mate without having to leave the territory. (d) A female may incite interactions among males by moving (arrow) towards a boundary. (e) Her movement may draw neighbouring males to that boundary (arrows) and the resulting interactions may allow her to assess other signals, including short-range types of display (plumage, direct dominance interactions or fights). (f ) Thus the female could assess a subset of the original males using multiple signals in succession, possibly leading to increased certainty of her assessment.

found that female black-capped chickadees Poecile atricapillus place their nests close to territorial edges, despite evidence that these areas offer no better resources or nesting opportunities than central nest locations. One explanation for this pattern is that females can more easily monitor neighbours in relation to their mates and capitalize on opportunities for secondary mate choice. One observed outcome of this nesting pattern by females is increased numbers of territorial disputes between the resident male and his neighbours (Ramsay et al., 1999), which

P1: IYP/... P2: JZZ/... 0521823617c07.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:49

Enlightened decisions: female assessment may also make it easier for females to make these relative distinctions. Female hooded warblers Wilsonia citrina at the peak of their fertility give characteristic chip calls, which result in increased intrusions by neighbouring males (Neudorf, 1996); female European robins Erithacus rubecula show similar patterns in their rates of seep calls (Tobias & Seddon, 2002). While such ‘fertility announcements’ are often interpreted as mechanisms that increase female choice through potential sperm competition, they may also function more directly in active assessment by females if they increase interactions between mates and other males.

Evidence for female assessment of network information Few studies have asked whether female mate choice incorporates information derived from signals in a communication network, although the potential for females to use such information seems considerable (e.g. Otter et al., 1999; Mennill et al., 2002). K. A. Otter, T. M. Peake, A. M. R. Terry and P. K. McGregor (unpublished data) found that dawn chorus singing of neighbouring male great tits Parus major is clearly recorded by microphones placed within the nestboxes of roosting females; therefore, it is likely that females could assess a network of singing males without leaving the nestbox. Relative song output among males during the dawn chorus is known to correlate with male condition in a number of species (Reid, 1987; Alatalo et al., 1990; Otter et al., 1997; Thomas, 1999a,b; Thomas & Cuthill, 2002), and it is, therefore, a useful cue of quality (Hutchinson et al., 1993). However, surprisingly little work has investigated whether females attend to variation in male dawn song. Otter & Ratcliffe (1993, 1996) suggested that changes in dawn singing of males who have lost their mates might function as useful cues for neighbouring females seeking better mates, and anecdotal evidence in black-capped chickadees suggests that divorces occur soon after the dawn chorus ends. Further studies should be conducted to determine whether the generally higher song output at dawn is used in assessment by females. This could be done by elevating male song output by supplementary feeding (e.g. Reid, 1987; Alatalo et al., 1990; Thomas, 1999a,b) and seeing whether radio-tracked females appear attentive to increased song output of neighbours. To date, the few studies that have investigated female assessment in communication networks have focused on eavesdropping upon dyadic male aggressive singing interactions (Otter et al., 1999; Mennill et al., 2002). These eavesdropping experiments have used interactive playback (Dabelsteen & McGregor, 1996) to manipulate the outcome of aggressive singing interactions between males. Otter et al. (1999) showed that female great tits appear to be aware of the relative ease with which males interact with a ‘strange intruder’ (the interactive playback). The mates of males who lost interactions were more likely to intrude into

141

P1: IYP/... P2: JZZ/... 0521823617c07.xml CU1917B/McGregor

142

0 521 582361 7

April 11, 2005

11:49

K. A. Otter & L. Ratcliffe neighbouring territories in following days than were those whose mates won against the same intruder. As the only difference in singing between the two male treatments was in their song rates relative to the playback, females appear to base their movement patterns on the perceived interaction. Moreover, females visited more frequently the neighbouring male heard to win against the same intruder to which their mate had lost. Our results, however, found no evidence that females produced more young with these males; there was no pattern of extra-pair copulation associated with the playback treatment, although males who had been cuckolded were of lower genetic heterozygosity than males who were not cuckolded (Otter et al., 2001). As genetic heterozygosity appears to be associated with survival and fecundity (Coulson et al., 1998; Hansson et al., 2001), this result might indicate that females drawn to males via song may have found other signals (e.g. colour patterns, which were not assessed during the study) that contradicted the assessment via song. By contrast, Mennill et al. (2002) recently showed that female black-capped chickadees exposed to similar eavesdropping opportunities did modify extra-pair behaviour, although there was no apparent effect on observed intrusions. Males of high and low social rank were exposed to challenges simulated by interactive playback. The challenges either reinforced their rank disparity (e.g. de-escalating playback to a dominant male, escalating to the subordinate), or countered rank disparity (escalate to dominant male, de-escalate to subordinate). While females mated to low-ranking males showed no influence of playback on their decisions to engage in extra-pair copulation, females mated to high-ranking males that had lost against the playback ‘intruder’ were more likely to have extra-pair young in their broods. As females in this species mated to high-ranking males usually forego extra-pair copulation (Otter et al., 1998), this result suggests that the protocols used by Mennill et al. (2002) had a profound impact on female decisions. Normally, if females mated to high-ranking males do engage in extra-pair copulation, they select males of similar or higher rank than their mate (Otter et al., 1998). Yet, D. J. Mennill and colleagues (unpublished data) found that extra-pair males selected by these high-ranking females were nearly random with respect to the relative rank of their mate; further evidence that assessment by eavesdropping in a network can have dramatic influences on behavioural decisions. There is also experimental evidence that female Siamese fighting fish Betta splendens eavesdrop on male–male aggressive visual displays and are more willing to mate with males that they have seen win such interactions (Doutrelant & McGregor, 2000; see also Ch. 2) These initial studies provide impetus for future work. However, a number of fundamental questions still need to be addressed. For example, the nature and accuracy of the information on relative male quality available to females in interactions remains to be determined. Another major issue is the extent to which features of signals used for individual identification are

P1: IYP/... P2: JZZ/... 0521823617c07.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:49

Enlightened decisions: female assessment affected by transmission over long distances. Many ideas related to communication networks assume that individual signallers can be identified; for example, in the context of mate choice in birds, females are assumed to be able to distinguish different males by voice. This has been shown in a number of species (e.g. Davis, 1986; Weary & Krebs, 1992; Lind et al., 1996); however, these studies have not required females to make these discriminations at long distance. Although song is a long-distance signal, all acoustic signals are subject to degradation over distance (Bradbury & Vehrencamp, 1998), which could negatively affect features females use in discrimination and assessment.

Implications for networks of environmental alteration Song transmission is affected not only by distance but also by the medium through which it must travel. Reverberation, differential attenuation and other effects on sound are imposed by habitat characteristics and may shape the songs of species inhabiting different areas (Catchpole & Slater, 1996). Habitat alteration can result in a change in the characteristics of signals to maintain maximum transmission range within new habitats, for example in rufous-collared sparrows Zonotrichia capensis inhabiting forested versus grassland habitats (Tubaro et al., 1993; Tubaro & Segura, 1994). However, it is unknown how long it takes for changes in song to occur in response to habitat change and how females respond to such changes. Most habitat alterations occur over very short timeframes, and unless reproductive isolation occurs between undisturbed and disturbed habitats, selection in response to the altered landscape may be slow (e.g. Dhondt et al., 1992; Dias & Blondel, 1996). Therefore, changes in the structure of male song may not keep pace with changes in the habitats, leading to song structure that is mismatched for transmission in the present environment. In many species, habitat alteration may change sound transmission conditions and also decrease resource availability or breeding success (Blondel, 1985; Blondel et al., 1993; Fort & Otter, 2004). If habitat alteration simultaneously reduces signal transmission and enlarges territory size in response to lowered resources, the extent of communication networks and the information to be gained from them may be seriously reduced. For example, the size of male song networks in blackcapped chickadees appears to be constrained by habitat change. While recording focal males during the dawn chorus, we conducted standard avian point counts at three-minute intervals to determine the number and direction of other males audible at the location of the focal male. The result was that fewer males were audible to male chickadees that occupy early successional forests (characterized by a low canopy and dense understorey) than to those occupying nearby mature, mixed forests (I.-J. Hansen, K. A. Otter & H. van Oort, unpublished data). This is likely a consequence of decreased transmission of song (I.-J. Hansen, K. A. Otter & H. van

143

P1: IYP/... P2: JZZ/... 0521823617c07.xml CU1917B/McGregor

144

0 521 582361 7

April 11, 2005

11:49

K. A. Otter & L. Ratcliffe Oort, unpublished data) and increased territory size (Fort & Otter, 2004) of males occupying this disturbed habitat. Such changes to the transmission of signals following habitat alteration can potentially diminish assessment ability of females. Not only could the networks decline, reducing the number of males a female could assess for secondary mate choice, but, in addition, females in such circumstances could also fail to locate primary mates, or become polygynous. Alternatively, females may mate monogamously with males of lesser quality during primary mate selection, because assessment of neighbouring males is constrained (M. Kasumovic, L. M. Ratcliffe & P. T. Boag, unpublished data). If such assessment is critical in female mating tactics (e.g. Wagner, 1991), females may fail to settle in such altered habitats, even if the resources would support a breeding effort. Another impact of habitat alteration that could affect female assessment in networks is the close relationship between resource access and the ability of males to produce condition-dependent traits. If habitat quality is poor, the absolute expression of traits may be diminished (Hill, 1995); moreover Qvarnstr¨ om and Forsgren (1998) also predict that dominant males may suffer disproportionately in poor habitat. The costs of achieving dominance status are normally countered by the benefit of access to rich resources, but if the habitat is unable to produce these benefits, the high costs paid by dominants may put them in a net metabolic deficit. In support of this, H. van Oort, K. A. Otter, F. Fort & C. I. Holschuh (unpublished data) found that song output in the dawn chorus of black-capped chickadees varies across habitats. As predicted by Otter et al. (1997), birds occupying mature forests in northern British Columbia, Canada had song output that reflected their relative rank: high-ranked birds tended to have higher song output than lower-ranked birds. By comparison, birds settling in neighbouring young, regenerating forests did not show this same trend. Overall, the birds in the disturbed forests had lower song output than birds in the undisturbed forests, as predicted by Hill (1995), but this relationship was driven by abnormally low song output by high-ranking males in the disturbed woods, as predicted by Qvarnstr¨ om and Forsgren (1998). Males of either high or low rank in the disturbed forest could not be differentiated based on song output. Therefore, the transmission of signals may not be the only impact of landscape alterations; the reliability of signals may also be influenced by habitat context and may diminish the ability of females to use long-range signalling networks in assessment.

Summary and future directions The role of communication networks in female mate choice is ripe for study using the techniques that simulate signalling interactions. In territorial

P1: IYP/... P2: JZZ/... 0521823617c07.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:49

Enlightened decisions: female assessment

(a)

(b)

(c) Fig. 7.3. A proposed experiment to test females’ use of information in communication networks in male assessment. The figure shows an aviary in plan view with release points at the corners (solid squares). (a) The aviary contains four cells (squares) each containing a choice stimulus (represented by a loudspeaker symbol here, but they could contain live males. (b) Signals are designed so that, when broadcast, only a position in the middle of the aviary would allow the female to assess all males simultaneously (i.e. the point of overlap of the effective signal ranges, shown as dotted lines). (c) The pattern of female movement observed (dashed lines), for example consistently moving towards the central areas prior to entering a cell, would indicate that several males were being compared.

songbirds, studies of eavesdropping (Ch. 2) may help us to understand female secondary mating tactics. By manipulating the relative signals emanating from neighbouring males, and using radio-tracking and genetics to measure female preferences, we should be able to obtain a clearer idea of how (and perhaps why) socially monogamous females choose secondary partners.

145

P1: IYP/... P2: JZZ/... 0521823617c07.xml CU1917B/McGregor

146

0 521 582361 7

April 11, 2005

11:49

K. A. Otter & L. Ratcliffe Considering communication in a network context should also aid our understanding of primary mate choice, an area of particular interest. It is particularly important to determine at what range females can discriminate between signals. As we argued earlier, measures of close approach to males may be insufficient to determine whether males have been ‘sampled’ by females. This is particularly the case if females are assessing males at long distance, approaching only chosen males. Such cryptic choice is potentially difficult to monitor. One way of addressing this problem is to plot female movements in relation to known signal transmission range. Speaker replacement studies, such as those used with flycatchers (Eriksson & Wallin, 1986) or starlings Sturnus vulgaris (Mountjoy & Lemon, 1991), could simulate clusters of signalling males to determine the effects of altered singing patterns on female assessment routes and strategies. Aviary studies may also be informative in this regard. Rather than the traditional dyadic choices presented to females, aviaries with several males could be presented (Fig. 7.3). The idea that females position themselves to assess males in a network could be investigated by manipulating the transmission range of auditory or visual signals and observing female movements (e.g. Fig. 7.3).

Acknowledgements We thank Peter McGregor, Bart Kempenaers, Dan Mennill, Harry van Oort, Carmen Holschuh, Tania Tripp and David Nordstrom for discussion on the topics in this chapter. IngebjørgJean Hansen, Kevin Fort, Harry van Oort, Peter Christie, Dan Mennill, Mike Kasumovic, Peter McGregor, Tom Peake and Andrew Terry kindly allowed us to cite the results of unpublished, co-authored data. Peter McGregor, Dan Mennill, Marc Naguib, Harry van Oort and an anonymous reviewer also provided useful suggestions on early drafts of the manuscript. Both authors were funded by NSERC (Canada) research grants during the preparation of this work.

References Alatalo, R. V., Glynn, C. & Lundberg, A. 1990. Singing rate and female attraction in the pied flycatcher: an experiment. Animal Behaviour, 39, 601–602. Andersson, M. 1994. Sexual Selection. Princeton, NJ: Princeton University Press. Bensch, S. & Hasselquist, D. 1992. Evidence for active female choice in a polygynous warbler. Animal Behaviour, 44, 301–311. Birkhead, T. R., Buchanon, K. L., de Voogd, T. J. et al. 1997. Song, sperm quality and testes asymmetry in the sedge warbler. Animal Behaviour, 53, 965–971. Blondel, J. 1985. Breeding strategies of the blue tit and coal tit (Parus) in mainland and island Mediterranean habitats: a comparison. Journal of Animal Ecology, 54, 531–556.

P1: IYP/... P2: JZZ/... 0521823617c07.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:49

Enlightened decisions: female assessment Blondel, J., Dias, P. C., Maistre, M. & Perret, P. 1993. Habitat heterogeneity and life-history variation of Mediterranean blue tits (Parus caeruleus). The Auk, 110, 511–520. Bradbury, J. W. & Vehrencamp, S. L. 1998. Principles of Animal Communication. Sunderland, MA: Sinauer. Brenowitz, E. A. 1982. The active space of red-winged blackbird song. Journal of Comparative Physiology, A, 147, 511–22. Buchanan, K. L. & Catchpole, C. K. 1997. Female choice in the sedge warbler, Acrocephalus schoenobaenus: multiple cues from song and territory quality. Proceedings of the Royal Society of London, Series B, 264, 521–526. Calder, W. A. 1990. The scaling of sound output and territory size: are they matched? Ecology, 71, 1810–1816. Catchpole, C. K. & Slater, P. J. B. 1996. Bird Song. Cambridge, UK: Cambridge University Press. Christie, P. J., Mennill, D. J. & Ratcliffe, L. M. 2004. Pitch shifts and song structure indicate male quality in the dawn chorus of black-capped chickadees. Behavioral Ecology and Sociobiology, 55, 341–348. Coulson, T. N., Pemberton, J. M., Albon, S. D. et al. 1998. Microsatellites reveal heterosis in red deer. Proceedings of the Royal Society of London, Series B, 256, 489–495. Dabelsteen, T. & McGregor, P. K. 1996. Dynamic acoustic communication and interactive playback. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca, NY: Cornell University Press, pp. 398–408. Dale, S. & Slagsvold, T. 1995. Female contests for nest sites and mates in the pied flycatcher, Ficedula hypoleuca. Ethology, 99, 209–222. Dale, S., Amundsen, T, Lifjeld, J. T. & Slagsvold, T. 1990. Mate sampling behaviour of female pied flycatchers: evidence for active mate choice. Animal Behaviour, 27, 87–91. Davis, W. J. 1986. Acoustic recognition in the belted kingfisher: cardiac response to playback vocalizations. The Condor, 88, 505–512. Desrochers, A. 1989. Sex, dominance, and microhabitat use in wintering black-capped chickadees: a field experiment. Ecology, 70, 636–645. Dhondt, A. A., Kempenaers, B. & Adriaensen, F. 1992. Density-dependent clutch size caused by habitat heterogeneity. Journal of Animal Ecology, 61, 643–648. Dias, P. C. & Blondel, J. 1996. Local specialization and maladaptation in the Mediterranean blue tit (Parus caeruleus). Oecologia, 107, 79–86. Doutrelant, C. & McGregor, P. K. 2000. Eavesdropping and mate choice in female fighting fish. Behaviour, 137, 1655–1669. Eens, M. 1997. Understanding the complex song of the European starling: an integrated ethological approach. Advances in the Study of Behaviour, 26, 355–434. Ekman, J. & Askenmo, C. 1984. Social rank and habitat use in willow tit groups. Animal Behaviour, 32, 508–514. Ekman, J. B. & Lilliendahl, K. 1992. Using priority to food access: fattening strategies in dominance-structured willow tit (Parus montanus) flocks. Behavioral Ecology, 4, 232–2.

147

P1: IYP/... P2: JZZ/... 0521823617c07.xml CU1917B/McGregor

148

0 521 582361 7

April 11, 2005

11:49

K. A. Otter & L. Ratcliffe Ens, B. J., Safriel, U. N. & Harris, M. P. 1993. Divorce in the long-lived and monogamous oystercatcher, Haematophagus ostralegus: incompatibility or choosing a better option? Animal Behaviour, 45, 1199–1217. Eriksson, D. & Wallin, L. 1986. Male bird song attracts females: a field experiment. Behavioral Ecology and Sociobiology, 19, 297–299. Ficken, M. S., Weise, C. M. & Popp, J. W. 1990. Dominance rank and resource access in winter flocks of black-capped chickadees. Wilson Bulletin, 102, 623–633. Forstmeier, W., Kempenaers, B., Meyer, A. & Leisler, B. 2002. A novel song parameter correlates with extra-pair paternity and reflects male longevity. Proceedings of the Royal Society of London, Series B, 269, 1479–1485. Fort, K. & Otter, K. A. 2004. The impact of habitat disturbance on reproduction of chickadees. The Auk, 121, in press. Gibson, R. M. & Langen, T. A. 1996. How do animals choose their mates? Trends in Ecology and Evolution, 11, 468–470. Gil, D. & Gahr, M. 2002. The honesty of bird song: multiple constraints for multiple traits. Trends in Ecology and Evolution, 17, 133–141. Gowaty, P. A. 1996. Multiple mating by females selects for males that stay: another hypothesis for social monogamy in passerine birds. Animal Behaviour, 51, 482–484. Grafen, A. 1990. Biological signals as handicaps. Journal of Theoretical Biology, 144, 517–546. Hansson, B., Bensch, S., Hasselquist, D., & Åkesson, M. 2001. Microsatellite diversity predicts recruitment of sibling great reed warblers. Proceedings of the Royal Society of London, Series B, 268, 1287–1291. 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. Hill, G. E. 1995. Ornamental traits as indicators of environmental health. BioScience, 45, 25–31. 2002. A Red Bird in a Brown Bag: The Function and Evolution of Colourful Plumage in the House Finch. Oxford: Oxford University Press. Hogstad, O. 1988. Rank-related resource access in winter flocks of willow tit Parus montanus. Ornis Scandinavica, 19, 169–174. Hoi-Leitner, M., Nechtelberger, H. & Hoi, H. 1995. Song rate as a signal for nest site quality in blackcaps (Sylvia atricapilla). Behavioral Ecology and Sociobiology, 37, 399–405. Houtman, A. M. 1992. Female zebra finches choose extra-pair copulation with genetically attractive males. Proceedings of the Royal Society of London, Series B, 249, 3–6. Hutchinson, J. M. C., McNamara J. M. & Cuthill, I. C. 1993. Song, sexual selection, starvation, and strategic handicaps. Animal Behaviour, 45, 1153–1177. Iwasa, Y. & Pomiankowski, A. 1994. The evolution of mate preferences for multiple sexual ornaments. Evolution, 48, 853–867. Janetos, A. C. 1980. Strategies of female mate choice: a theoretical analysis. Behavioral Ecology and Sociobiology, 7, 107–112.

P1: IYP/... P2: JZZ/... 0521823617c07.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:49

Enlightened decisions: female assessment Johnstone, R. A. 1996. Multiple displays in animal communication: ‘backup signals’ and ‘multiple messages’. Philosophical Transactions of the Royal Society of London, Series B, 351, 329–338. Kempenaers, B. & Dhondt, A. A. 1993. Why do females engage in extra-pair copulations? A review of the hypotheses and their predictions. Belgian Journal of Zoology, 123, 93–103 Kempenaers, B., Verheyen, G. R. & Dhondt, A. A. 1997. Extrapair paternity in the blue tit (Parus caeruleus): female choice, male characteristics and offspring quality. Behavioral Ecology, 8, 481–492. Lampe, H. M. & Espmark, Y. O. 1994. Song structure reflects male quality in pied flycatchers, Ficedula hypoleuca. Animal Behaviour, 47, 869–876. Lind, H., Dabelsteen, T. & McGregor, P. K. 1996. Female great tits can identify mates by song. Animal Behaviour, 52, 667–671. McGregor, P. K. 1993. Signalling in territorial systems: a context for individual identification, ranging, and eavesdropping. Philosophical Transactions of the Royal Society of London, Series B, 340, 237–211. McGregor, P. K. & Dabelsteen, T. 1996. Communication networks. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca, NY: Cornell University Press, pp. 409–425. Mennill, D. J., Ratcliffe, L. M. & Boag, P. T. 2002. Female eavesdropping on male song contests in songbirds. Science, 296, 873. Møller, A. P. 1991. Parasite load reduces song output in a passerine bird. Animal Behaviour, 41, 723–730. Møller, A. P. & Pomiankowski, A. 1993. Why have birds got multiple sexual ornaments? Behavioral Ecology and Sociobiology, 32, 167–176. Møller, A. P., Saino, N., Taramino, G., Galeotti, P. & Ferrario, S. 1998. Paternity and multiple signalling: effects of a secondary sexual character and song on paternity in the barn swallow. American Naturalist, 151, 236–242. Montgomerie, R. & Thornhill, R. 1989. Fertility advertisment in birds: a means of inciting male–male competition? Ethology, 81, 209–220. Morton, E. S., Forman, L. & Braun, M. 1990. Extra-pair fertilisations and the evolution of colonial breeding in purple martins. The Auk, 107, 275–283. Mountjoy, D. J. & Lemon, R. E. 1991. Song as an attractant for male and female European starlings, and the influence of song complexity on their response. Behavioral Ecology and Sociobiology, 28, 97–100. Murphy, C. G. & Gerhardt, H. C. 2002. Mate sampling by female barking treefrogs (Hyla gratiosa). Behavioral Ecology, 13, 472–480. Naguib, M., Fitchel, C. & Todt, D. 1999. Nightingales respond more strongly to vocal leaders of simulated dyadic interactions. Proceedings of the Royal Society of London, Series B, 266, 537–542. Neudorf, D. L. 1996. A dual system of female control of extra-pair copulations in the hooded warbler (Wilsonia citrina). Ph.D.; Thesis. York University, North York, Ontario. Neudorf, D. L., Stutchbury, B. J. M. & Piper, W. H. 1997. Covert extraterritorial behavior of female hooded warblers. Behavioral Ecology, 8, 595–600.

149

P1: IYP/... P2: JZZ/... 0521823617c07.xml CU1917B/McGregor

150

0 521 582361 7

April 11, 2005

11:49

K. A. Otter & L. Ratcliffe Otter, K. & Ratcliffe, L. 1993. Changes in the singing behaviour of male black-capped chickadees (Parus atricapillus) following mate removal. Behavioral Ecology and Sociobiology, 33, 409–414. 1996. Female initiated divorce in a monogamous songbird: abandoning mates for males of higher quality. Proceedings of the Royal Society of London, Series B, 263, 351–354. Otter, K., Chruszcz, B. & Ratcliffe, L. 1997. Honest advertisement and singing during the dawn chorus of black-capped chickadees, Parus atricapillus. Behavioral Ecology, 8, 167–173. Otter, K., Ratcliffe, L., Michaud, D. & Boag, P. T. 1998. Do female black-capped chickadees prefer high ranking males as extra-pair partners? Behavioral Ecology and Sociobiology, 43, 25–36. Otter, K., McGregor, P. K., Terry, A. M. R. et al. 1999. Do female great tits (Parus major) assess males by eavesdropping? A field study using interactive song playback. Proceedings of the Royal Society of London, Series B, 266, 1305–1310. Otter, K., Stewart, I. K., Terry, A. M. R., McGregor, P. K. & Burke, T. 2001. Extra-pair paternity in great tits in relation to manipulation of male signals. Journal of Avian Biology, 32, 338–344. Pomiankowski, A & Iwasa, Y. 1993. Evolution of multiple sexual preferences by Fisher’s runaway process of sexual selection. Proceedings of the Royal Society of London, Series B, 253, 173–181. Qvarnstr¨ om, A. & Forsgren, E. 1998. Should females prefer dominant males? Trends in Ecology and Evolution, 13, 498–501. Rades¨ ater, T., Jakobsson, S., Andbjer, N., Bylin, A. & Nystr¨ om, K. 1987. Song rate and pair formation in the willow warbler, Phylloscopus trichilus. Animal Behaviour, 35, 1645–1651. Ramsay, S. M, Otter, K. & Ratcliffe, L. M. 1999. Nest-site selection by female black-capped chickadees: settlement based on conspecific attraction? The Auk, 116, 604–617. Real, L. 1990. Search theory and mate choice I: models of single-sex discrimination. American Naturalist, 136, 376–405. Reid, M. L. 1987. Costliness and reliability in the singing behaviour of Ipswich sparrows. Animal Behaviour, 35, 1735–1743. Sæther, S. A. 2002. Female calls in lek-mating birds: indirect mate choice, female competition for mates or direct mate choice? Behavioral Ecology, 13, 344–352. Saino, N., Galeotti, P., Sacchi, R. & Møller, A. P. 1997a. Song and immunological condition in male swallows (Hirundo rustica). Behavioral Ecology, 8, 364–371. Saino, N., Primmer, C. R., Ellegren, H. & Møller, A. P. 1997b. An experimental study of paternity and tail ornamentation in the barn swallow (Hirundo rustica). Evolution, 51, 562–570. Slagsvold, T. & Lifjeld, J. T. 1994. Polygyny in birds: the role of competition between females for male parental care. American Naturalist, 143, 59–94. Slagsvold, T., Lifjeld, J. T., Stenmark, G. & Breiehagen, T. 1988. On the cost of searching for a mate in female pied flycatchers Ficedula hypoleuca. Animal Behaviour, 36, 433–442.

P1: IYP/... P2: JZZ/... 0521823617c07.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

11:49

Enlightened decisions: female assessment Smith, S. M. 1988. Extra-pair copulations in black-capped chickadees: the role of the female. Behaviour, 107, 15–23. Thomas, R. J. 1999a. Two tests of a stochastic dynamic programming model of daily singing routines in birds. Animal Behaviour, 57, 277–284. 1999b. The effect of variability in the food supply on the daily singing routines of European robins: a test of a stochastic dynamic programming model. Animal Behaviour, 57, 365–369. Thomas, R. J. & Cuthill, I. C. 2002. Body mass regulation and the daily singing routines of European robins. Animal Behaviour, 63, 285–292. Tobias, J. A. & Seddon, N. 2002. Female begging in European robins: do neighbors eavesdrop for extrapair copulations? Behavioral Ecology, 13, 637–642. Todt, D. & Naguib, M. 2000. Vocal interactions in birds: the use of song as a model in communication. Advances in the Study of Behaviour, 29, 247–295. Tubaro, P. L. & Segura, E. T. 1994. Dialect differences in the song of Zonotrichia capensis in the Southern Pampas: a test of the acoustic adaptation hypothesis. The Condor, 96, 1084–1088. Tubaro, P. L., Segura, E. T & Handford, P. 1993. Geographic variation in the song of the rufous-collared sparrow in eastern Argentina. The Condor, 95, 588–595. Wagner, R. H. 1991. Evidence that female razorbills control extra-pair copulations. Behaviour, 118, 157–169. Weary, D. M. & Krebs, J. R. 1992. Great tits classify songs by individual voice characteristics. Animal Behaviour, 43, 283–287. Wetton, J. H., Burke, T., Parkin, D. T. & Cairns, E. 1995. Single-locus DNA fingerprinting reveals that male reproductive success increases with age through extra-pair paternity in the house sparrow (Passer domesticus). Proceedings of the Royal Society of London, Series B, 260, 91–98. Wiegmann, D. D., Real, L. A., Capone, T. A. & Ellner, S. 1996. Some distinguishing features of models of search behavior and mate choice. American Naturalist, 147, 188–204. Wiley, R. H. & Richards, D. G. 1982. Adaptations for acoustic communication in birds: sound transmission and signal detection. In: Acoustic Communication in Birds, Vol. 1, ed. D. E. Kroodsma & E. H. Miller. New York: Academic Press, pp. 131–181. Wittenberger, J. F. 1983. Tactics of mate choice. In: Mate Choice, ed. P. Bateson. Cambridge, UK: Cambridge University Press, pp. 435–447. Yasukawa, K., Blank, J. L. & Patterson, C. B. 1980. Song repertoires and sexual selection in the red-winged blackbird. Behavioral Ecology and Sociobiology, 7, 233–238.

151

P1: IRK/KAA P2: IYP-KOD 0521823617c08.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

13:38

8

Predation and noise in communication networks of neotropical katydids alexander b. l ang, ingeborg teppner, manfred h a r t b a u e r & h e i n e r r O¨ m e r Karl-Franzen University, Graz, Austria

Introduction Intraspecific acoustic communication in grasshoppers or katydids appears to be a very simple and straight forward behaviour: one sex – usually the male – produces an acoustic signal, and the female, once perceiving and recognizing the signal as species specific, shows some kind of response, either an acoustic reply or a phonotactic movement to the male. However, the system is far from being that simple and involves more than just a sender and receiver. First, communication usually takes place in a physically complex environment, where sound signals are subject to attenuation and degradation, depending on the carrier frequencies, which are often in the high-sonic or ultrasonic range because of the small size of the sound radiating structures (Wiley & Richards, 1978; Michelsen, 1992). In addition, the physical conditions of the transmission channel for the sound may vary strongly during day or night and with weather conditions; consequently, the ability to detect and localize a signal undergoes strong variations. Second, insects often aggregate and communicate in areas rich in resources or at periods of the day or night favouring mate attraction. As a result of many signallers calling in close proximity, masking interference will take place at the site of receivers, depending on the spacing, as well as the kind and extent of signal timing. Since such favourable areas and times for signalling are similar for different species, heterospecific choruses may be formed with impressive sound pressure levels of biological background noise, which further complicates the detection of a signal (reviewed for katydids by Schatral (1990)). Third, a female might gain fitness benefits (directly or indirectly) by choosing a male based on variation of Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

152

P1: IRK/KAA P2: IYP-KOD 0521823617c08.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

13:38

Neotropical katydids: predation and noise in networks particular properties of the calling song. Such a female preference may drive the evolution of the signal in a particular direction. For example, in insects, patterns of preferences based on call rate or the duration of long calls are usually highly directional, whereas those based on pulse rate or frequency are stabilizing (reviewed by Gerhardt & Huber, 2002). Female preferences can also result from a sensory bias in the sensory or nervous system of receivers (Ryan, 1990; Ryan & Keddy-Hector, 1992; Ryan & Rand, 1993; Endler & Basolo, 1998). Thus sexual selection by female choice can result in signal traits that enhance the mating success of males. Fourth, traits preferred by females may also decrease male survivorship by increasing exposure to predators. Acoustically orienting predators or parasitoids can use the same signals produced for mate attraction to identify, localize and home in on the signaller (Cade, 1975; Belwood & Morris, 1987; Lehmann & Heller, 1998; Zuk & Kolluru, 1998; Allen, 2000). Male fitness can also be decreased by increased competition resulting from the signal attracting conspecific rivals. Both are cases of interceptive eavesdropping in the sense of Peake (Ch. 2). Fifth, signalling at the long duration and high rate preferred by females may be limited by energetic constraints, as sound production for small animals is rather inefficient and probably costly (Bailey et al., 1993; Wagner & Hoback, 1999). Finally, as ectothermic animals, the motor output of insects depends on the ambient temperature, and in consequence, the temporal properties of calling songs are influenced by environmental temperature. Vertical omer, temperature gradients in a grasshopper’s habitat can be 10 ◦ C in 30 cm (R¨ 2001); therefore, senders and receivers can differ strongly in body temperature. As a consequence, the signaller’s temporal patterning of song may not match the preference function of a receiver. From this short summary, it is clear that some of the factors contributing to the evolution of acoustic communication systems could interact in a complex way. For example, if high predation risk forces a species to communicate acoustically at a time of day or night when acoustic competition with other species is high, the consequence is a high degree of song interference and masking, and the calling activity of one species can inhibit that of other species (Greenfield, 1988; R¨ omer et al., 1989). In this chapter we emphasize the importance of an ecological (integrated) approach to communication networks. By focusing on the intraspecific communication of a subfamily of neotropical katydids, we demonstrate the complex dependency of predation and signalling, nocturnal ambient light levels, masking noise levels and alternative signalling strategies.

Predation and antipredator defences in rainforest katydids A key paper by Belwood & Morris (1987) (see also Belwood, 1990; Morris et al., 1994) suggested that the evolution of specific anti-predator defences in a

153

P1: IRK/KAA P2: IYP-KOD 0521823617c08.xml CU1917B/McGregor

154

0 521 582361 7

April 11, 2005

13:38

A. B. Lang, I. Teppner, M. Hartbauer & H. R¨ omer family of neotropical katydids (Pseudophyllinae) is strongly influenced by antipredator defences. Members of this subfamily differ strongly in appearance, behaviour and hearing from those of the other larger taxon, the Phaneropterinae. The latter live in the canopy, have in general a green, leaf-like appearance and are good flyers. Their ears are about 15 dB more sensitive than the ears of pseudophyllines (unpublished results). By contrast, pseudophyllines live in the rainforest understorey, they have a long, slender, fusiform body and are bad flyers. Katydids are a major source of protein for diurnal predators such as birds (e.g. Formicariidae, Furnariidae and others), rodents and small primates (Nickle & Heymann, 1996; Martins & Setz, 2000). Some of these birds feed almost exclusively on arthropods by searching curled dead leaves that hang from vegetation in the lower understorey (Gradwohl & Greenberg, 1980, 1982, 1984; Remsen & Parker, 1984), thereby counteracting one of the katydids’ primary defence strategies, namely crypsis by general appearance and behaviour (Nickle & Castner, 1995). During the night, foliage-gleaning bats (Micronycteris hirsuta, Lophostoma silvicolum) eat large numbers of Pseudophylline katydids (Belwood, 1988). These bats are attracted by calling songs or other sounds involved in phonotactic activities of their prey. Forest-living katydids exhibit a range of behaviours and signal characters that appear to be adaptations to avoid predation by these bats: a reduction in call redundancy (duty cycles of 3% and less), high carrier frequencies over 20 kHz and the partial (or in one species, complete) replacement of airborne sound signals by substrate-borne vibrations (tremulation) (Belwood & Morris, 1987). In this chapter, we present data about the antipredator behaviour of a neotropical katydid, and its consequences for signal detection in noisy rainforest.

Predation pressure and roost site selection The study was conducted on Barro Colorado Island (BCI), Panama and on nearby peninsulas and small islands. The 1500 ha island is located in central un Lake, part of the Panama Canal. BCI is almost Panama (09◦ 10  N, 79◦ 51 W) in Gat´ totally covered with secondary and primary semideciduous lowland tropical forest (Foster & Brokaw, 1982). The study took place in February/March (dry season) and June/July 2002 (beginning of the rainy season). We studied Docidocercus gigliotosi, a Pseudophylline katydid with a medium-sized, long and slender brown body. Its natural history is only poorly known, although it is one of the most common katydids on the island (Belwood, 1988). M. hirsuta and Micronycteris megalotis are two insectivorous bats that glean highly cluttered spaces (Kalko et al., 1996) and feed on D. gigliotosi (established by identification of remains at bat roosts). This katydid constitutes about 20% of the diet of M. hirsuta (Belwood, 1988; personal observations).

P1: IRK/KAA P2: IYP-KOD 0521823617c08.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

13:38

Neotropical katydids: predation and noise in networks

(a)

(b)

(c)

20 m Fig. 8.1. Roosting of Docidocercus gigliotosi (Pseudophyllinae) in Aechmea magdalenae. (a) The A. magdalenae plant; (b) the 2.5 cm leaf edge spines; (c) the location of A. magdalenae plants in part of field ‘Zetek 15’ on Barro Colorado Island (Panama), mapped with a geographical information system. Each plant is marked with a circle; those occupied with one or more D. gigliotosi are shown with a filled circle.

We regularly found D. gigliotosi roosting during the day in Aechmea magdalenae, a terrestrial bromeliad of the pineapple family that can grow to a height of 2.5 m (Fig. 8.1a). A striking characteristic of these plants are numerous, inch-long spines along the leaf edges (Fig. 8.1b). It is abundant throughout BCI, forming dense stands of sometimes more than 1000 plants. The part of the field ‘Zetek 15’ shown in Fig. 8.1 comprises about 480 plants covering an area of 2600 m2 . The leaves of the plant form a long tube in the centre and that is where most katydids were found roosting. We observed several individuals shortly before sunset and during the night using infrared video cameras. Sunset occurred around 18:30 h and katydids usually became active (climbing and cleaning themselves within the plant) between 19:00 and 19:30 h. Three males were observed tremulating for several seconds. Between 20:00 and 21:00 h they used nearby lianas or trees to climb up into the lower canopy, where they could no longer be observed. We presume that they are active

155

P1: IRK/KAA P2: IYP-KOD 0521823617c08.xml CU1917B/McGregor

156

0 521 582361 7

April 11, 2005

13:38

A. B. Lang, I. Teppner, M. Hartbauer & H. R¨ omer in the canopy throughout the night, because they returned to their host plants around 04:00 h and climbed back into the central tube by 04:30 h, which is about one hour before sunrise. With a mark-and-recapture study, we confirmed this general scheme of daytime inactivity within the roost plant, nocturnal activity in the canopy approximately between 21.00 h and 04.00 h and the return to the plant. We recaptured 35 (out of 65) adults within 17 days; only three recaptures were more distant than 2 m from the marking site. The maximum recapture distance was 10 m and we never found marked individuals in different A. magdalenae fields. Some individuals were found in the same plant for a period of more than two weeks and 66% of katydids were recaptured in the same plant. D. gigliotosi were not randomly distributed among A. magdalenae plants but were found to roost in taller plants in above average condition with leaf-litter-free central tubes close to canopy access ‘walkways’ (A. B. Lang & H. R¨ omer, unpublished data). Figure 8.1c shows a field of such plants (Zetek 15), in which plants occupied by one or more individuals are marked with filled dots. D. gigliotosi roosted in plants that were significantly taller (mean height (± standard deviation) 1.68 ± 0.3 m (n = 32)) than unoccupied plants (1.36 ± 0.36 m (n = 320); (two-tailed MannWhitney U test, p < 0.0001). Similar results were obtained for two study periods in February/March and May/June 2002. We attempted to quantify the quality of roost plants by ranking the condition of each plant on a subjective scale from bad (0) to excellent (3). This ranking included the state of desiccation, number of damaged leaves, and number of fresh, fleshy leaves, in particular those in the centre. A survey performed in July 2002 found that the average condition of A. magdalenae plants occupied by katydids was significantly better than that of unoccupied plants (occupied plants, mean rank (± STD) 2.55 ± 0.69 (n = 36); unoccupied plants, mean rank (± STD) 1.72 ± 0.96 (n = 248); two-tailed Mann–Whitney U test p < 0.0001). Similar results were found for a survey carried out in March 2002. Most (81%) of the plants in which adult D. gigliotosi roosted had direct contact with, or grew within 1 m of a tree or liana reaching at least to the lower canopy. These data indicate that the life history of the Pseudophylline katydid D. gigliotosi is strongly influenced by predators. Two pieces of evidence are consistent with D. gigliotosi attempting to avoid predation. First, by roosting during the day in the spiny bromeliad A. magdalenae they are protected from predatory birds and mammals. Insectivorous birds can have a pronounced effect on populations of their arthropod prey (e.g. Lepidoptera larvae: Holmes et al., 1979) and on BCI Myrmotherula fulviventris (Formicariidae) spends 98% of its foraging time searching aerial leaf litter for arthropods and about 20% of its prey items are crickets and katydids (Gradwohl & Greenberg, 1982). Therefore, the usual pseudophyllines habit of roosting in curled leaves is potentially risky. Heavy predation by birds and

P1: IRK/KAA P2: IYP-KOD 0521823617c08.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

13:38

Neotropical katydids: predation and noise in networks other visually hunting predators selects for various patterns of crypsis in their katydid prey (Belwood, 1990; Castner & Nickle, 1995) and also for their choice of backgrounds on which to hide, and on other life-history traits. Second, D. gigliotosi is not active when gleaning bats show most flight activity (which is likely to be related to foraging activity). I. silvicolum has a peak in flight activity about one hour after sunset, when it flies to its hanging perch (C. D. Weise, E. K. V. Kalko, personal communication), and one hour before sunrise, when it flies back. Radio telemetry showed a similar pattern of activity in M. hirsuta (S. Spehn, personal communication). This correlates with the finding that D. gigliotosi, one of the bat’s common prey species, does not exhibit night-time activity until the major period of bat flight activity is over, and they also return to their bromeliad roost before the bats return to their roosts. Although one may argue that flight activity of bats does not necessarily reflect the time of highest predation pressure for their prey, it is reasonable to assume that bats would home in on katydid song and the noises caused by prey flight or landing activity during this time.

The costs of nocturnal communication: masking interference Many species of insect and anuran communicate acoustically at night and the resulting multispecies choruses have high sound pressure levels (SPL) and complex spectral properties. Figure 8.2a shows measurements of the SPL in the rainforest on BCI over a period of 24 hours. During the day, the SPL was rather low, measuring 40–50 dB. It rapidly increased shortly after sunset by some 20 dB as a result of calling activity of insects and frogs. SPL declined throughout the night (depending on the moon cycle; see below), until it reached daytime levels after sunrise. In such chorus noise, different species occupy different frequency bands. In our recording, the most prominent frequencies were below 8 kHz (calling songs of crickets) (Fig. 8.2b), but frequencies well above 20 kHz were also obvious. We used a ‘biological microphone’ to ‘listen’ through the ears of the biological receiver to analyse the challenging problem of signal detection after sunset for D. gigliotosi at the position of potential receivers. The biological microphone is a small, portable outdoor neurophysiological set-up that records the action potential activity of a single, identified auditory interneuron of a katydid (Rheinlaender & R¨ omer, 1986; R¨ omer & Lewald, 1992). Rather than analysing the properties of signals and noise at the position of potential receivers with conventional microphones, such a method allows one to listen through the ears of the biological receivers and to draw conclusions from the analysis of afferent nervous activity under these natural conditions.

157

P1: IRK/KAA P2: IYP-KOD 0521823617c08.xml CU1917B/McGregor

April 11, 2005

13:38

A. B. Lang, I. Teppner, M. Hartbauer & H. R¨ omer

Background noise level (dB SPL)

(a) 70

Full moon Last quarter New moon

Sunset

Sunrise

60

50

40

06:00

12:00

18:00 24:00 Time of day (h)

06:00

(b) 30 Frequency (kHz)

158

0 521 582361 7

20

10

15

20

25

Time (s)

Fig. 8.2. Measurement of the multispecies chorus on Barro Colorado Island (Panama). (a) Sound pressure level (SPL) in the rainforest over a period of 24 hours, at three moon phases. (b) Sonagram of a 16 second sound recording at the same site after sunset at 19.00 h. Note the different frequency bands between 3 and 7.5 kHz produced predominantly by crickets, and those in the high-frequency and ultrasonic range produced by katydids. The short duration and low redundancy calling song of Docidocercus gigliotosi (frequency range 22–25 kHz) is marked by arrows.

A typical result is shown in Fig. 8.3a. The receiver was placed within the rainforest at 17.00 h, 10 m from a speaker broadcasting a conspecific calling song. Since a female has no a priori knowledge about the presence of a signal, her only information about the presence or absence of a signal is encoded in afferent nervous activity such as shown in Fig. 8.3a. This task is apparently easy before sunset (Fig. 8.3a, upper trace), because each burst of action potential activity (increase in spike frequency) was the result of a conspecific stimulus. A detection criterion based on bursts of action potentials or the corresponding increase in spike rate

P1: IRK/KAA P2: IYP-KOD 0521823617c08.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

13:38

Neotropical katydids: predation and noise in networks

(a)

17.00 h

18.30 h Spike frequency

(b)

150 100 50

5s Fig. 8.3. Outdoor action potential recording of an identified nerve cell of the afferent auditory pathway of Docidocercus gigliotosi (see text). (a) Recording obtained within the rainforest at 17.00 h (upper trace), and at 18.30 h (lower trace). Before sunset, masking noise level is low and each conspecific stimulus elicits bursts of action potentials (hits shown by open dots). After sunset, many bursts of action potentials are elicited by noise (false alarms shown by stars). (b) Longer recording of action potentials after sunset (lower trace) and the corresponding spike rate (upper trace). Note that a signal presented at five second intervals (arrows on upper trace, points between traces) elicits an increase in the spike rate, and noise may result in a stronger increase in the spike rate than the signal.

would give ‘hits’ in terms of signal detection (Green & Swets, 1966). Indeed, in all cases when there was an acoustic signal during the experiment at 17.00 h, there was bursting activity in the nerve cell and there was no such activity when a signal was absent; therefore, there were no ‘misses’ or ‘false alarms’, respectively. However, this ideal situation for signal detection changed completely after sunset, when most katydids and other insects started to communicate acoustically. The same preparation at exactly the same position in the rainforest now exhibited high action potential activity (Fig. 8.3a, lower trace) and only an a priori knowledge of the time of signalling allows correct detection of the stimuli. Using the same detection criterion as in the situation before sunset would result in many false alarms (i.e. identifying background noise as signals: stars in Fig. 8.3a, lower trace). We measured a false alarm rate of more than 1400 in only five minutes, thus exceeding the hit rate dramatically, rendering communication between conspecifics

159

P1: IRK/KAA P2: IYP-KOD 0521823617c08.xml CU1917B/McGregor

160

0 521 582361 7

April 11, 2005

13:38

A. B. Lang, I. Teppner, M. Hartbauer & H. R¨ omer Representation of signal within CNS of receiver

Male signal Fig. 8.4. The amplitude-modulated calling song of a male Thamnobates subfalcata katydid (lower trace) and the corresponding representation of this song in the spike discharge of a nerve cell (omega-neuron) in the afferent auditory pathway of the central nervous system (CNS) in a female Docidocercus gigliotosi receiver.

rather ineffective. Figure 8.3b illustrates this point over a longer time scale and also shows that a signal presented at regular intervals of five seconds elicited a corresponding increase in the spiking rate, but that noise pulses may result in a stronger increase in the spike rate than the signal. Signal detection would be improved by increasing either the duration or the rate of signalling, and indeed we found such an effect with our preparation. Similar experiments to those shown in Fig. 8.3 clearly indicate that the rate of hits increases and the rate of false alarms decreases with increased signal redundancy and duration. However, as pointed out by Belwood & Morris (1987), eavesdropping (interceptive eavesdropping in the sense of Peake (Ch. 2)) by gleaning bats excludes such a solution and illustrates the opposing selection pressures of avoiding predation and signalling effectively. How do the insects solve the problem? A closer look at the signals used by males offers a possible solution. Figure 8.4 shows a typical, short amplitude-modulated signal of a male katydid Thamnobates subfalcata and the corresponding representation of this signal in the spike discharge of a nerve cell (omega-neuron) in the afferent pathway of a receiver. If the parameters of the spike discharge in response to a species-specific call differ from those in bursts elicited by background noise, this difference could be used by the nervous system to discriminate signals from noise. To investigate this further we used the biological microphone to record action potentials in the rainforest at about 21.00 h, when the level of the background noise was still high (Fig. 8.2). A male signal was broadcast every two seconds at 10 m from the preparation. The SPL of the signal was adjusted to a value of approximately 10 dB above the neuron’s masked threshold. We then used a self-developed Delphi-application (Delphi 6, Borland Software Corporation, Scotts Valley, CA 95066-3249, USA) to extract bursts from afferent spike recordings. We

P1: IRK/KAA P2: IYP-KOD 0521823617c08.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

13:38

Neotropical katydids: predation and noise in networks compared several features of bursts (spike variables: the number of spikes in the sequence; the mean, maximum, minimum; variance; average deviation; standard deviation; skewness; kurtosis of the interval between two spikes) between hits (i.e. bursts induced by the playback call) and false alarms (i.e. bursts elicited by background noise). One half of the recording (30 minutes) was used as the training set to compute the decision tree: that is, the program learned the specific spike variables that characterize hits. This tree was then evaluated for the remaining 30 minutes of the recording, the validation set, to see whether it could detect such responses in a noisy background. Machine learning on the basis of decision-tree learning is one of the most widely used and practical data-mining methods to classify very large amounts of data. It is a method for approximating discrete-valued functions, in which the learned function is represented by a decision tree that is robust to noisy data and capable of learning disjunctive expressions (Mitchell, 1997). We used the algorithms J48 (pruned, unpruned) (Quinlan, 1993) and PART (Frank & Witten, 1998) for classification of bursts within spike trains, calculating the results with the Java application WEKA (Trigg et al., 1999). Methods of decision-tree learning such as J48 and PART search a completely expressive hypothesis space and thus avoid the difficulties of restricted hypothesis spaces (Mitchell, 1997). To avoid so-called ‘over fitting’ (see also Mitchell, 1997), we conducted a tenfold cross-validation. The unpruned J48 decision tree classified 95.4% of the bursts correctly, meaning that it was able to distinguish between the double pulse signal and background noise with an error of only 4.6% (Table 8.1). The effect of varying the duration of the playback signal (7, 70 and 700 milliseconds) at a constant signal rate of 0.5/second is shown in Table 8.1. The most obvious effect is the lower detectability of the 7 millisecond signal by all three algorithms; however, there is little increase in detectability between the 70 and 700 millisecond signals. A similar result was found for the grasshopper Chorthippus biguttulus, where signal detection improved with increasing signal duration up to 450 milliseconds but did not improve further with longer signals (Ronacher et al., 2000). Although such results do not tell us that the insect’s nervous system makes use of this information, it does show how bursts of action potentials in response to conspecific song can, in principle, be discriminated from bursts produced by other sound sources. The most important spike variables used by the decision trees in discriminating signals from noise were the kurtosis, number of spikes, mean and variance. The minimum and maximum spike interval was often used in final decisions of the trees. Because some of these parameters (e.g. number of action potentials or the minimum or maximum spike intervals) are also relevant in real nervous systems for processing and discriminating sensory information, one can assume that the insect nervous system can also solve this discrimination task.

161

P1: IRK/KAA P2: IYP-KOD 0521823617c08.xml CU1917B/McGregor

162

0 521 582361 7

April 11, 2005

13:38

A. B. Lang, I. Teppner, M. Hartbauer & H. R¨ omer Table 8.1. The effects of duration (7, 70, 700 milliseconds) of the playback signal (presented at 0.5/second) on signal detection in afferent spike trains of a receiver based on decision tree learning with three different algorithms: J48, unpruned J48, and PART a Algorithm

Correctly classified bursts (%)

Stimulus 7 ms (n = 609) J48

80.13

J48 (unpruned)

79.64

PART

85.22

Stimulus 70 ms (n = 616) J48

94.77

J48 (unpruned)

95.15

PART

94.77

Stimulus 700 ms (n = 108) J48

95.37

J48 (unpruned)

95.37

PART

95.37

a See

text for further explanation.

Using machine-learning procedures to evaluate afferent spike patterns in sensory systems may also enable us to look more closely at the strategic design of signals (Guilford & Dawkins, 1991). For example, given the advantage of shortduration and low-redundancy signalling in the presence of interceptive eavesdropping predators, what degree of amplitude modulation in a signal is necessary to make its representation in afferent channels reliably different from heterospecific signals? This is part of ongoing research on a variety of katydid species on BCI, some of which pose a real challenge for signal detection by using signals of only a few milliseconds duration. Additional behavioural experiments under noisy conditions with the same species are urgently needed in order to show whether the insects’ nervous systems can solve the task.

Variation in ambient light and noise levels, and the use of a conditional communication strategy Two features of neotropical rainforest at night vary with the lunar cycle: insect abundance and noise level. We shall argue that they are related to predation pressure on communication. We quantified the effect of lunar cycle on katydid abundance on BCI by collecting at mercury vapour lights in December 1999 and April 2001 at 21:00 and 24:00 h.

P1: IRK/KAA P2: IYP-KOD 0521823617c08.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

13:38

Neotropical katydids: predation and noise in networks

Full moon

30

20

10

New moon

Collected katydids per night

40

0

1

4

7

10

13

16

19

22

25

28

Day in moonphase

Fig. 8.5. The number of katydids collected at mercury vapour lights on Barro Colorado Island (Panama) in relation to moon phase. Note that the higher variation at new moon may be because of rainfall during these nights, when many katydids hide under leaves and do not fly.

There was a significant relationship between moon phase and the number of katydids collected (Fig. 8.5): at full moon the number of katydids collected approached zero and it reached a maximum at new moon. As these data are similar to results obtained with other insects (e.g. Hardwick, 1972) and are also similar to collections made with suction traps, these cycles in abundance reflect natural activity patterns. Light intensity may vary by three to four orders of magnitude between full moon and new moon (Erkert, 1974). Full-moon ambient light levels are high enough for humans to orient easily in the forest understorey with their dark-adapted eyes. It is, therefore, likely that these light conditions allow a variety of predators to hunt visually and, consequently, their potential prey must adopt a cryptic lifestyle. We quantified the effect of lunar cycle on background noise level on BCI with a continuous recording system. The system consisted of a sound level meter (CEL 414 plus attached CEL-296 digital filter with settings A- weighting and slow time constant) with a condenser microphone (LD 2540, type 4133, range 4–40 kHz). The set-up was protected from humidity and rainfall and heated to 2 ◦ C above ambient temperature with an infrared bulb to prevent fogging of the microphone membrane. The DC output of the sound level meter was monitored at intervals of five seconds with a Maclab/Powerlab 4e data acquisition system (AD Instruments Pty Ltd) connected to a portable computer (Sony PCG-F707). Recordings were made from the end of October to early December 2001, as well as in February, May and June 2002. Figure 8.2a shows representative examples of noise measurements over 24 hours at full moon, new moon and the last quarter of a lunar cycle. The increase

163

P1: IRK/KAA P2: IYP-KOD 0521823617c08.xml CU1917B/McGregor

164

0 521 582361 7

April 11, 2005

13:38

A. B. Lang, I. Teppner, M. Hartbauer & H. R¨ omer

Tremulation

Stridulation

2s

Fig. 8.6. The two types of mating signal produced by a male Docidocercus gigliotosi. Tremulations (vibratory signals, upper trace) were recorded without mechanical contact by laser doppler vibrometry, simultaneously with airborne sound signals produced by elytral stridulation (lower trace). The photograph shows a male on a large leaf.

in noise at sunset and the decrease at sunrise are common to all recordings. However, at full moon, noise levels decrease after sunset and for the rest of the night the noise level is only 10 dB above the daytime level. As a result, the masking noise level between 09:00 and approximately 05:00 h varies cyclically with the moon phase; the amplitude of variation is about 10 dB. Given the fact that the masking noise is the result of acoustic signalling, predominantly by insects, the drop in noise level at full moon is best explained by species and/or individuals reducing or abandoning signalling with airborne sound. We have argued above that the nocturnal lifestyle of many insects avoids predation from visually hunting predators (e.g. rodents) and that light intensity at full moon may allow increased visual predation. Therefore, a cryptic lifestyle may include cryptic signalling. Direct evidence for this hypothesis comes from the signalling behaviour of D. gigliotosi. Males produce airborne sound with the usual elytral stridulation and also tremulations, when the insect rapidly shakes its body up and down or drums with the abdomen on the substratum (Morris et al., 1994). Tremulations are transmitted through the substratum and females respond with tremulations of their own. Although the active space of a tremulation signal is limited to the plant where the tremulation is produced, it may travel 2–3 m (Michelsen et al., 1982; Markl, 1983). Tremulation is an effective way to communicate in the presence of acoustic interceptive eavesdroppers such as gleaning bats because only receivers equipped with sensitive vibration receptors (e.g. spiders) can intercept the signal. Figure 8.6 demonstrates that the duration and redundancy of the tremulation

P1: IRK/KAA P2: IYP-KOD 0521823617c08.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

13:38

Neotropical katydids: predation and noise in networks (a)

60

New moon

Number of events/5 min

50

40 30 20 10 0

(b)

60

Full moon

Number of events/5 min

50

40 30 20

10

03:36

02:24

01:12

00:00

22:48

21:36

20:24

19:12

18:00

0

Time of day Fig. 8.7. The pattern of production of the two types of mating signal (tremulation, open squares; airborne sound, filled squares and grey area) produced by a male Docidocercus gigliotosi at new moon (a) and full moon (b). Males were separated from females a day before data were collected, caged and placed in the rainforest understorey. An infrared camera recorded both signals over a period of about nine hours, from sunset to 03.00 h.

signal of D. gigliotosi is higher than the stridulatory airborne sound signal by orders of magnitude. As expected if male D. gigliotosi vary the relative proportion of these two modes of signalling in relation to the chance of predation from gleaning bats, at new moon, or in the laboratory in complete darkness, males become active about half an hour after sunset and begin signalling with a period of high-rate tremulation, followed by a prolonged period of airborne sound production, often for many hours (Fig. 8.7a). At full moon the onset of signalling after sunset is often delayed, airborne sound signalling is strongly reduced and tremulation is more common. It should also be noted that D. gigliotosi reduces acoustic signalling at full moon despite the fact that the lower noise levels would allow better detection of conspecific signals. The observations are consistent with D. gigliotosi having a conditional strategy of signalling, where fairly cryptic (i.e. short duration, low redundancy) airborne sound production is replaced by the even more private mode of communication with tremulations.

165

P1: IRK/KAA P2: IYP-KOD 0521823617c08.xml CU1917B/McGregor

166

0 521 582361 7

April 11, 2005

13:38

A. B. Lang, I. Teppner, M. Hartbauer & H. R¨ omer Summary In this chapter we have emphasized the importance of predation on the ecology and acoustic communication of a subfamily of neotropical katydids. Pseudophylline katydids switch from airborne sound signalling to tremulation because of high predation by passive listening bats. This results in either decreased active space of their signals or decreased detectability of airborne sound signals in high levels of background nocturnal noise. Neurophysiological experiments and decision-tree learning algorithms nevertheless indicate that information for discrimination between signals and noise is still preserved within afferent spike discharges. Daytime and night-time predation also appears to restrict individuals to roost sites within bromelid plants. D. gigliotosi exhibits a strong site dependency for a particular plant in a field over several weeks; consequently, there is very little horizontal movement of males and females in a population. In conjunction with a reduced active space of acoustic and/or vibrational signalling, and reduced flight capability, this situation may strongly reduce the chances of matings with individuals of neighbouring populations. Current population genetic analysis will show whether genetic exchange between populations is suppressed, despite the lack of a geographical barrier between populations.

Acknowledgements Inspiring discussion with Elisabeth K. V. Kalko, Christa D. Weise, Dina K. N. Dechmann and Sabine Spehn played a major role in maturing the ideas presented here. We acknowledge the logistic support of the Smithsonian Tropical Research Institute, Panama. Barbara Bliem, Franz Kainz, Birgit Roehnfeld and Iris Strauss assisted in the field and in the laboratory. We also thank Peter McGregor and two anonymous reviewers for many suggestions to improve the manuscript. Research was supported by the Austrian Science Fund (FWF P14257-BIO to H. R.) and a Ph.D. scholarship (Austrian Academy of Sciences to A. B. L.).

References Allen, G. R. 2000. Call structure variability and field survival among bushcrickets exposed to phonotactic parasitoids. Ethology, 106, 409–423. Bailey, W. J., Withers, P. C., Endersby, M. & Gaull, K. 1993. The energetic costs of calling in the bushcricket Requena verticalis (Orthoptera: Tettigoniidae: Listroscelidinae). Journal of Experimental Biology, 178, 21–37. Belwood, J. J. 1988. The influence of bat predation on calling behaviour in Neotropical forest katydids (Insecta: Orthoptera: Tettigoniidae). Ph.D. Thesis, University of Florida, Gainesville, FL.

P1: IRK/KAA P2: IYP-KOD 0521823617c08.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

13:38

Neotropical katydids: predation and noise in communication networks 1990. Anti-predator defences and ecology of Neotropical forest katydids, especially the Pseudophyllinae. In: The Tettigoniidae: Behaviour, Systematics, Evolution, ed. W. J. Bailey & D. C. F. Renz. Bathurst, Australia: Crawford House Press, pp. 6–26. Belwood, J. J. & Morris, G. K. 1987. Bat predation and its influence on calling behaviour in Neotropical katydids. Science, 238, 64–67. Cade, W. 1975. Acoustically orienting parasitoids: fly phonotaxis to cricket song. Science, 190, 1312–1313. Castner, J. L. & Nickle, D. A. 1995. Observations on the behaviour and biology of leaf-mimicking katydids (Orthoptera: Tettigoniidae: Pseudophyllinae: Pterochrozini). Journal of Orthopteran Research, 4, 93–98. Endler, J. A. & Basolo, A. L. 1998. Sensory ecology, receiver biases and sexual selection. Trends in Ecology and Evolution, 13, 415–420. Erkert, H. G. 1974. Der Einfluß des Mondlichtes auf die Aktivit¨ atsperiodik nachtaktiver S¨ augetiere. Oecologia, 14, 269–287. Foster, R. B. & Brokaw, N. V. L. 1982. Structure and history of the vegetation of Barro Colorado Island. In: The Ecology of a Tropical Forest, ed. E. G. Leigh, A. S. Rand & D. M. Windsor. Washington, DC: Smithsonian Institute Press, pp. 151–172. Frank, E. & Witten, I. H. 1998. Generating accurate rule sets without global optimization. In: Proceedings of the 15th International Conference on Machine Learning, Madison, WI, ed. J. Shavlik. San Francisco, CA: Morgan Kaufmann, pp. 144–151. Gerhardt, H. C. & Huber, F. 2002. Acoustic Communication in Insects and Anurans. Chicago, IL: University of Chicago Press. Gradwohl, J. A. & Greenberg, R. 1980. The formation of antwren flocks on Barro Colorado Island, Panama. The Auk, 97, 395. 1982. The effect of a single species of avian predators on the arthropods of aerial leaf litter. Ecology, 63, 581–583. 1984. Search behaviour of the checker-throated antwren foraging in aerial leaf litter. Behavioral Ecology and Sociobiology, 15, 281–285. Green, D. M. & Swets, J. A. 1966. Signal Detection Theory and Psychophysics. New York: Wiley. Greenfield, M. D. 1988. Interspecific acoustic interactions among katydids Neoconocephalus: inhibition-induced shifts in diel activity. Animal Behaviour, 36, 684–695. Guilford, T. & Dawkins, M. S. 1991. Receiver psychology and the evolution of animal signals. Animal Behaviour, 42, 1–14. Hardwick, D. F. 1972. The influence of temperature and moon phase on the activity of noctuid moths. Canadian Entomologist, 104, 1767–1770. Holmes, R. T., Schultz, J. C. & Nothnagle, P. 1979. Bird predation on forest insects: an exclosure experiment. Science, 206, 462–463. Kalko, E. K. V., Handley, C. O., Jr & Handley, D. 1996. Organization, diversity, and long-term dynamics of a Neotropical bat community. In: Long-term Studies of

167

P1: IRK/KAA P2: IYP-KOD 0521823617c08.xml CU1917B/McGregor

168

0 521 582361 7

April 11, 2005

13:38

A. B. Lang, I. Teppner, M. Hartbauer & H. R¨ omer Vertebrate Communities, ed. M. L. Cody & J. A. Smallwood. New York: Academic Press, pp. 503–553. Lehmann, G. U. C. & Heller, K. G. 1998. Bushcricket song structure and predation by the acoustically orienting parasitoid fly Therobia leonidei (Diptera: Tachinidae: Ormiini). Behavioral Ecology and Sociobiology, 43, 239–245. Markl, H. 1983. Vibrational communication. In: Neuroethology and Behavioural Physiology, ed. F. Huber & H. Markl. Berlin: Springer, pp. 332–353. Martins, M. M. & Setz, E. F. Z. 2000. Diet of buffy tufted-eared marmosets (Callithrix aurita) in a forest fragment in southeastern Brazil. International Journal of Primatology, 21, 467–476. Michelsen, A. 1992. Hearing and sound communication in small animals: evolutionary adaptations to the law of physics. In: The Evolutionary Biology of Hearing, ed. E. D. Webster, R. R. Fay & A. N. Popper. New York: Springer, pp. 61–77. Michelsen, A., Fink, F., Gogala, M. & Traue, D. 1982. Plants as transmission channels for insect vibrational songs. Behavioral Ecology and Sociobiology, 11, 269–281. Mitchell, T. M. 1997. Machine Learning. New York: McGraw-Hill. Morris, G. K., Mason, A. C., Wall, P. & Belwood, J. J. 1994. High ultrasonic and tremulation signals in Neotropical katydids (Orthoptera: Tettigoniidae). Journal of Zoology (London), 233, 129–163. Nickle, D. A. & Castner, J. L. 1995. Strategies utilized by katydids (Orthoptera: Tettigoniidae) against diurnal predators in rainforests of northeastern Peru. Journal of Orthopteran Research, 4, 93–98. Nickle, D. A. & Heymann, E. W. 1996. Predation on orthoptera and related orders of insects by tamarin monkey, Saguinus mystax and Saquinus fuscicollis (Primates: Callitrichidae), in northeastern Peru. Journal of Zoology, (London), 239, 799–819. Quinlan, J. R. 1993. C.4.5: Programs for Machine Learning. San Francisco, CA: Morgan Kaufmann. Remsen, J. V. & Parker, T. A. 1984. Arboreal dead-leaf-searching birds of the neotropics. The Condor, 86, 36–41. Rheinlaender, J. & R¨ omer, H. 1986. Insect hearing in the field: the use of identified nerve cells as biological microphones. Journal of Comparative Physiology, A, 158, 647–651. Ronacher, B., Krahe, R. & Hennig, R. M. 2000. Effects of signal duration on the recognition of masked communication signals by the grasshopper Chorthippus biguttulus. Journal of Comparative Physiology, A, 186, 1065–1072. R¨ omer H. 2001. Ecological constraints for sound communication: from grasshoppers to elephants. In: Ecology of Sensing, ed. F. G. Barth & A. Schmid. Berlin: Springer, pp. 59–77. R¨ omer, H. & Lewald, J. 1992. High-frequency sound transmission in natural habitats: implications for the evolution of insect acoustic communication. Behavioral Ecology and Sociobiology, 157, 631–642. R¨ omer, H., Bailey, W. & Dadour, I. 1989. Insect hearing in the field. III. Masking by noise. Journal of Comparative Physiology, A, 164, 609–620.

P1: IRK/KAA P2: IYP-KOD 0521823617c08.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

13:38

Neotropical katydids: predation and noise in communication networks Ryan, M. J. 1990. Sensory systems, sexual selection, and sensory exploitation. Oxford Surveys in Evolutionary Biology, 7, 157–195. Ryan, M. J. & Keddy-Hector, A. 1992. Directional patterns of female mate choice and the role of sensory biases. American Naturalist, 139, S4–S35. Ryan, M. J. & Rand, S. A. 1993. Sexual selection and signal evolution: the ghost of biases past. Philosophical Transactions of the Royal Society of London, Series B, 340, 187–195. Schatral, A. 1990. Interspecific acoustic behaviour among bushcrickets. In: The Tettigoniidae: Behaviour, Systematics, Evolution, ed. W. J. Bailey & D. C. F. Renz. Bathurst, Australia: Crawford House Press, pp. 150–165. Trigg, L., Hall, M. & Frank, E. 1999. WEKA (Java collection of machine learning algorithms). Hamilton, New Zealand: University of Waikato; http://www.cs.waikato.ac.nz/ml/weka/. Wagner, W. E., Jr & Hoback, W. W. 1999. Nutritional effects on male calling behaviour in the variable field cricket. Animal Behaviour, 57, 89–95. Wiley, R. H. & Richards, D. G. 1978. Physical constraints on acoustic communication in the atmosphere: implications for the evolution of animal vocalizations. Behavioral Ecology and Sociobiology 3, 69–94. Zuk, M. & Kolluru, G. R. 1998. Exploitation of sexual signals by predators and parasitoids. Quarterly Review of Biology, 73, 415–438.

169

P1: JZZ/... P2: JZZ/... 0521823617c09.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

12:20

9

Nestling begging as a communication network andrew g. horn & mart y l. leonard Dalhousie University, Halifax, Canada

Introduction In many bird species, young beg for care from their parents. A parent arriving at the nest with food is met by begging nestlings, which are waving their wings, calling and stretching to expose brightly coloured gapes, all within the confines of a nest that may contain several other begging nestlings. This mode of parent–offspring communication has become a model for the study of the evolution of biological signalling. Hungrier nestlings beg more intensely, so the parent can use the display to decide which nestling to feed and to decide how soon it should return to the nest with food (reviewed by Budden & Wright, 2001). The fact that the parent can extract information on nestling hunger from such a confusing burst of signalling raises numerous questions. How does each nestling ensure that its own signal of need is received above the din of its nestmates’ displays? How do parents differentiate among these displays to choose which nestling to feed? How much do the displays, as opposed to the physical jostling toward the parent that also goes on in the nest, determine which nestlings are fed? To answer such questions we need to understand how the begging behaviours of whole broods function together. Concepts derived from the new field of communication networks seem well suited to this task but have not yet been explicitly applied to begging. As currently defined (McGregor & Dabelsteen, 1996; McGregor & Peake, 2000), a communication network forms whenever several individuals communicate within transmission range of each other’s signals. Nestlings noisily begging within the confines of a nest clearly fit this definition, since most or all of the nestlings within a brood are within transmission range of each other’s signals. Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

170

P1: JZZ/... P2: JZZ/... 0521823617c09.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

12:20

Nestling begging as a communication network In this chapter, we hope to show that considering begging as a communication network yields new insights, not only into begging behaviour but also into communication networks in general. We begin by briefly summarizing previous research on begging, most of which has treated the display as dyadic communication: that is, as signalling from one individual, the nestling (or the brood considered as sending one joint signal), to one receiver, the parent. We then apply ideas from studies of communication networks to nestling begging, identifying several conceptual issues that we think studies of begging can help to clarify. We also discuss aspects of the design of begging and parental behaviour that may have evolved in response to the network environment and finally we make some suggestions for future work.

Begging as dyadic communication In this section, we summarize the theoretical and empirical work on begging to date, most of which has treated parent–offspring communication as a dyadic communication system. This summary provides background information for the discussion of communication networks that follows, while also illustrating some of the strengths and weaknesses of the dyadic approach to begging behaviour. Theoretical work

Begging has attracted considerable attention from evolutionary biologists largely because of its apparently needless conspicuousness. Because parents are only a few centimetres away from their young, it is not clear why offspring signal for food with such an elaborate display. Perhaps the best-known explanation for this apparent extravagance stems from parent–offspring conflict theory (reviewed by Godfray, 1995; Parker et al., 2002). Natural selection favours parents that distribute resources optimally amongst both their current and future offspring. Each of these offspring, however, is selected to solicit resources so as to benefit its own fitness, rather than the inclusive fitness of its siblings. Therefore, offspring might signal for resources that parents would do better giving to siblings or reserving for future broods. To overcome parental reluctance, offspring may have to send exaggerated signals of need (Trivers, 1974; Godfray, 1995). This basic explanation has been revised or extended in various ways, making the parent–offspring dyad one of the most thoroughly modelled animal communication systems. Some of the most influential models, both for begging and for animal signals in general, have asked how reliable signalling can evolve in the face of conflict between signallers and receivers (reviewed by Godfray & Johnstone, 2000; Johnstone & Godfray, 2002). Specifically, if young are prone to exaggerate, then

171

P1: JZZ/... P2: JZZ/... 0521823617c09.xml CU1917B/McGregor

172

0 521 582361 7

April 11, 2005

12:20

A. G. Horn & M. L. Leonard why would parents respond at all to begging signals? The answer is that, whereas parents might easily be able to assess some aspects of their nestlings, like their size, parents might not be able to assess important aspects of their nestlings’needs, for example their immediate need for food. If begging provides information on these aspects of nestling need, then parents should provision nestlings according to variation in the begging signal. This situation can be evolutionarily stable, however, only if the signal is costly for nestlings to produce. Therefore, in effect, nestlings might have to put on a costly begging display to prove that they really are hungry (Godfray, 1991). These results have been largely responsible for the general acceptance of the idea that reliable signals must be costly if they are to evolve. Some of the complexities of this story are less widely known, however. For instance, recent models have suggested that, in some situations, nestlings might signal their needs accurately without large costs, for example if exaggeration draws so much care away from siblings that the cost to the signaller’s inclusive fitness outweighs the direct benefits of the extra signalling (Maynard Smith, 1994; Bergstrom & Lachmann, 1998; Johnstone, 1999; Price et al., 2002). For the purposes of this chapter, two features of theoretical work on begging particularly stand out. First, these models have focused on fundamental issues in dyadic communication, such as how signalling can evolve despite conflicts of interest between signallers and receivers. Thus they are relevant to our understanding of a wide range of communication systems. Second, the emphasis these models have placed on particular aspects of signalling, such as its honesty and costliness, has led empirical studies to focus on these aspects of begging to the neglect of others (see below). One of these neglected aspects is the communication network in which begging occurs; although recent attempts to model the effects of signalling on nestmates (reviewed by Royle et al., 2002; Johnstone & Godfray, 2002), which we discuss further below, are steps in that direction. Empirical work

The theoretical possibility that begging might be exaggerated led many researchers to test whether begging is indeed a reliable signal of need. Studies in a wide range of species confirm that the intensity of both the visual and vocal aspects of the display increase with food deprivation (Budden & Wright, 2001). In turn, parents use the begging signal in two ways to make provisioning decisions. First, the more intense the begging of the brood as whole, the more often parents return to the nest with food. This level of response has been shown most clearly in experiments in which playback of nestling begging calls stimulates higher provisioning rates (Budden & Wright, 2001). Second, once parents arrive at the nest, nestlings that beg more intensely than their nestmates are more likely to be

P1: JZZ/... P2: JZZ/... 0521823617c09.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

12:20

Nestling begging as a communication network fed. Experiments again provide the clearest demonstrations of this effect: parents are more likely to direct feedings to nestlings with brighter gapes (G¨ otmark & Ahlstr¨ om, 1997; Kilner, 1997; Saino et al., 2000; Saino and Møller, 2002) or to nestlings placed next to speakers playing higher call rates (Leonard & Horn, 2001a; Kilner, 2002a; but see Glassey & Forbes, 2002a). Therefore, begging appears to communicate to parents the requirements both of the brood as a whole and of individual nestlings. Begging is more than a simple cry for food, however, for two reasons. First, food deprivation is not the only aspect of nestling need that the begging display advertises. For example, in some species begging may signal long-term nutritional need as opposed to the short-term hunger described above, with nestlings in poorer condition (e.g. having lower mass than nestmates) begging more than their nestmates (Price et al., 2002). Additionally, some aspects of begging, especially begging calls, can change when nestlings lose heat, thus signalling the need for brooding (Evans, 1994; Leonard & Horn, 2001b; Clotfelter et al., 2003; B. Glassey, personal communication). Finally, gape colour in some species may advertise a nestling’s immunocompetence (Saino & Møller, 2002). Clearly, the message that begging is sending may be more complex than just short-term hunger. A second complicating factor is the effect of siblings on nestling begging. Begging intensity, whether measured by the intensity of the postural display or overall call rate, increases with brood size in many species (Budden & Wright, 2001) and may also increase when nestmates beg (e.g. Leonard & Horn, 1998). Also, nestlings compete physically for access to parents (see below) and their display and its effect on parents may vary according to the nature and intensity of this physical competition (e.g. Price et al., 1996; Cotton et al., 1999). Interest in the effects of both signalling interactions and physical competition among nestmates has mainly focused on how they complicate honest signalling of need (e.g. Rodr´ıguez-Giron´es et al., 2001; Price et al., 2002). We will be discussing them further below because they are clearly central to any discussion of begging as a communication network. Summary

This brief review shows that the main emphasis of work on begging has been on how it functions as a signal of need from nestlings to parents. Begging has been treated mainly as a dyadic signalling system: that is involving one signaller (the nestling or the brood considered as sending one joint signal) and one receiver (the parent). Siblings have been included in the picture, but mainly because they might affect the dyadic signalling of need. Only recently have researchers started to consider the effects of competing signalling by nestmates in any detail, an important step toward treating the begging brood as a communication network.

173

P1: JZZ/... P2: JZZ/... 0521823617c09.xml CU1917B/McGregor

174

0 521 582361 7

April 11, 2005

12:20

A. G. Horn & M. L. Leonard Begging as a communication network If we are to broaden our view of begging to include the communication network in which it occurs, we must first characterize that communication network. By definition, a brood of begging young is a communication network because nestmates are all within range of each other’s signals (McGregor & Dabelsteen, 1996; McGregor & Peake, 2000). Going beyond this definition, however, to characterize the network and explore its implications, raises more conceptual challenges than this simple definition might suggest. In this section, we discuss three of these issues. First, to apply the definition of communication networks at all, we must distinguish signalling from other acts. This can be especially problematic in the case of begging, in which signalling and direct physical competition are tightly linked. Second, to examine some of the more interesting implications of the network, we must carefully consider the nature of signals and signalling interactions – again, a challenging distinction when applied to begging. Third, there are factors, such as the genetic relatedness of nestlings, which are at least as important for characterizing this communication network as the overlapping transmission ranges of signals that define it. While all three of these areas present challenges for studies of begging networks, they also provide opportunities for testing some key concepts in the study of communication networks. Physical competition versus signalling

Nestlings form a communication network because they are within signalling range of each other. Unlike members of many other communication networks, however, nestlings are also in direct physical contact with each other. This tight proximity highlights difficulties that can arise when we try to distinguish between signalling and other acts, in this case physical competition. Since a communication network, by definition, consists of signalling (i.e. of behaviours specialized to communicate information (McGregor & Peake, 2000)), this distinction is fundamental for understanding any communication network. Nestlings jostle with one another for access to parental feeding locations within the nest and their success at reaching the parent strongly affects which nestlings are fed (Budden & Wright, 2001). Nestlings can physically compete in several ways, for example by usurping positions close to where parents arrive at the nest, by blocking parents’ access to other nestlings or, particularly in nonpasserine species, by directly pushing or pecking one another (Mock & Parker, 1997; Budden & Wright, 2001; Drummond, 2002). Much of this physical competition is hard to distinguish from signalling. Jostling for position and direct aggression seem to be non-signalling acts by which

P1: JZZ/... P2: JZZ/... 0521823617c09.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

12:20

Nestling begging as a communication network nestlings get better access to parents. Parents may nonetheless get information about nestling need and quality from these physical interactions, which they then use to choose which nestling to feed (Rodr´ıguez-Giron´es, 1996; Lotem et al., 1999). This informativeness alone does not make them signals. If, however, the interactions are designed to affect that choice, rather than merely to thrust a nestling forward to rob the parent of its choice, then they are signals, by the above definition, despite their outward appearance. Conversely, some features of begging that appear to have been designed partly to convey information and thus are signals by definition (McGregor & Peake, 2000), such as posturing (Kilner, 2002b), seem just as clearly designed for effective jostling toward the parent. Even the design features of begging displays that are adaptations for overcoming interference from nestmates (reviewed below) may be seen either as ways to signal information on need more effectively to parents (Horn & Leonard, 2002) or as scrambles for parental attention (Rodr´ıguez-Giron´es et al., 2001; Royle et al., 2002). In the latter case, their ultimate function would differ little from that of physical competition, since by dominating the parents’ visual and acoustic fields they too would not inform parents so much as reduce the parents’ opportunity to choose which nestling to feed. Therefore, a nest full of begging nestlings is part communication network, part scrum toward the parent. Which view of begging is more accurate depends largely on how parents interpret begging signals and physical competition, a topic we discuss further below. Given that display behaviours ultimately evolve from non-signalling acts, however, we can at least conclude that begging offers an interesting system for studying how social behaviours besides signalling affect communication networks. Signalling interactions versus just signalling

One of the aspects of communication that has become more prominent as a result of the communication network approach is the information content of signalling interactions: the give and take of signals among members of the network (McGregor & Peake, 2000). It is from the interactions between signallers, rather than the signals themselves, that some particularly interesting consequences of communication networks arise, such as signalling to avoid interference (Ch. 13) and eavesdropping (Peake et al., 2002; Ch. 2). Distinguishing signals (directed at the parent) from signal interactions (directed at nestmates) in the case of nestling begging is difficult, however. On the one hand, several lines of evidence show that nestmates’ signals influence how a nestling signals. In many studies, nestlings beg more intensely when in bigger broods or when with nestmates than when alone (Budden & Wright, 2001; but see Cotton et al., 1996). More direct evidence comes from studies in which nestlings

175

P1: JZZ/... P2: JZZ/... 0521823617c09.xml CU1917B/McGregor

176

0 521 582361 7

April 11, 2005

12:20

A. G. Horn & M. L. Leonard increase their postural display when their nestmates do (e.g. Leonard & Horn, 1998) or call more when they hear nestmates calling (e.g. Leonard & Horn, 2001c). On the other hand, it is not clear that these changes in signalling constitute signalling interactions in the sense implied by current discussions of communication networks, especially work on social eavesdropping (McGregor & Peake, 2000; Ch. 2). According to this work, a signalling interaction consists of a sender directing a signal at a receiver, which then responds. To the degree that begging is directed at the parent, then competitive interactions among nestlings to catch the parents’ attention are not signalling interactions in this sense (Royle et al., 2002). By extension, parents that choose to feed nestlings that beg more than their nestmates (Budden & Wright, 2001), like the predators that are attracted to nests whose calling is increased by competition (Haskell, 2002), are interceptive rather than social eavesdroppers, because social eavesdroppers must base their response on signalling interactions not just on signals (Ch. 2). This conclusion may partly reflect our still sketchy understanding of nestling interactions. For example, Roulin (2002) has recently suggested that at least some signalling by nestlings may be directed at nestmates. Nestling barn owls Tyto alba, for example, appear to have calling contests between parental visits, in which nestlings negotiate which of them will receive a feeding when the parent next returns (Roulin, 2002). If nestlings do direct signals to each other in this way, then parents that extract information from these interactions would fit the definition of social eavesdroppers (Ch. 2). In the particular case of barn owls, nestling negotiations occur when the parent is absent and so cannot be overheard by parents. In principle, however, there is no reason why similar interactions between nestlings could not also occur in the parent’s presence, especially in species in which parents spend enough time transferring food to their young that the young have time to interact (e.g. parrots (Psittaciformes); Krebs, 2002). Certainly, if nestlings do direct their signals to each other, the importance of considering nestling begging as a communication network is considerably strengthened. Functional relationships among nestlings and network structure

Communication networks were first defined in the context of communication among territorial songbirds, which are widely separated on different territories but are interconnected by the overlapping transmission ranges of their songs (McGregor & Dabelsteen, 1996). Song is, thus, the main way in which these birds interact; consequently, characterizing interacting songbirds as a communication network captures much of how they affect each other’s signalling behaviour. Nestlings packed together within a nest, however, are interconnected in many ways besides the overlapping ranges of their signals. We have already discussed

P1: JZZ/... P2: JZZ/... 0521823617c09.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

12:20

Nestling begging as a communication network how they interact through physical competition and how that may have strong effects on their signalling behaviour. In this section, we briefly list three other interconnections that are integral to any explanation of how nestmates affect each other’s signalling behaviour. Unlike physical competition, these effects do not present difficulties for defining signals and hence applying the definition of communication networks to begging. They do, however, illustrate that, in some communication networks, signallers are so mutually dependent on one another that the overlapping transmission ranges of their signals are only one way in which their signalling behaviours are interconnected. We will list three such relationships among nestlings: genetic relatedness, shared fate and heat transfer. For each category, we touch briefly on their possible implications for signalling. We then discuss perhaps their most interesting implication, which is how all these relationships might combine to give a structure to the communication network within the brood. Genetic relatedness

Genetic relatedness is perhaps the most important of the relationships among nestlings, because it so heavily influences the fitness consequences of all the other types of relationship. Since nestlings tend to be highly related to one another, relatedness probably affects signalling in this communication network more than in most of the other networks described in this volume. Indeed, for most theoretical models of begging, the main route of sibling effects on begging is through a nestling’s inclusive fitness. In general, theory predicts less-exaggerated or less-costly begging the higher the relatedness among nestmates ( Johnstone & Godfray, 2002; Price et al., 2002). Consistent with such predictions, interspecific brood parasites, whose relatedness with their host nestmates is zero, such as European cuckoos Cuculus canorus, great spotted cuckoos Clamator glandarius and brownheaded cowbirds Molothrus ater, call more loudly and more frequently than their nestmates (Dearborn & Lichtenstein, 2002; Redondo & Zu˜ niga, 2002). Evidence for non-parasitic species, however, is scant. In one comparison across species for which data on genetic parentage were available, begging calls were louder in species with more frequent mixed parentage (Briskie et al., 1994). This result suggests that a species’ average level of relatedness within broods might set its average level of begging. A more relevant result for communication networks, however, would be if nestlings within a species could assess their relatedness to broodmates and adjust their levels of competitive signalling accordingly. Nestlings are generally thought to lack the cues by which their nestmates could assess their relatedness (e.g. Whittingham & Dunn, 2001); indeed there may be selection against such cues ( Johnstone, 1997). As for kin recognition in birds in general

177

P1: JZZ/... P2: JZZ/... 0521823617c09.xml CU1917B/McGregor

178

0 521 582361 7

April 11, 2005

12:20

A. G. Horn & M. L. Leonard (Komdeur & Hatchwell, 1999), addressing this issue directly will require more sophisticated experiments than have been applied to date. Shared fate

Along with relatedness, a fundamental feature underlying nestling interactions is that, like the proverbial eggs in one basket, nestmates often share the same fate. For better or worse, they have the same adults feeding them, share the same local environmental conditions around the nest and, therefore share their chances of survival to a greater degree than participants in most other types of communication network. This shared fate has inevitable consequences for signalling behaviour; if one nestling begs more loudly, for example, the parents might return more often to feed all the nestlings or a predator might be more likely to find the nest and eat all the nestlings. Thus, both the benefits and the costs of begging by any given nestling are at least partially visited on the whole brood. Indeed, Wilson & Clark (2002) went still further and suggested that broods are subject to a form of group selection which may lead nestlings to signal cooperatively. Aspects of begging that are usually presented as competitive, such as signal characteristics that ostensibly serve to circumvent interference (see below), might instead function cooperatively to coordinate nestmates’ signals (Wilson & Clark, 2002). How individual signals fit together to form aggregate brood signals has not been studied yet, but we can safely expect that the shared fate of nestlings will make signalling interactions within their networks differ in interesting ways from those of signallers with more independent fates, such as chorusing frogs. Heat transfer

Nestling birds cannot thermoregulate until partway through the nesting period. Before that point, they rely not only on brooding by parents but also on heat from their nestmates. Nests where young hatch asynchronously may consist of older, heat-producing nestlings and younger, heat-consuming nestlings (e.g. Hill & Beaver, 1982). Such thermal relationships among nestlings may increase the variety of their signals and signalling interactions. Specifically, in several species, some aspects of begging, especially begging calls, change when nestlings lose heat and may signal their need for brooding (see above). Nestlings might, therefore, have to compete for attention from nestmates that are sometimes signalling for food and sometimes for warmth, and they might adopt different signalling strategies for each situation. Thermal relationships might also affect signalling through more direct effects on individual signallers. For example, some evidence suggests that house sparrow Passer domesticus nestlings lose heat when the stretching and gaping of begging increases their surface area (Ovadia et al., 2002). They might,

P1: JZZ/... P2: JZZ/... 0521823617c09.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

12:20

Nestling begging as a communication network therefore, be able to beg more when next to larger nestmates, since any thermal loss during begging would be reduced. Thermal relationships among nestmates are still poorly understood, but, like physical competition and signalling, they are probably readily perceived by nestlings and thus may have immediate and dynamic effects on patterns of signalling within the nest. Network structure

The net result of all the relationships listed above, including the physical competition also discussed, is that they may lend structure to the communication network within the nest. By ‘structure,’ we mean a pattern in which not all nestlings have the same sorts of relationship with one another. Most obviously, physical competition can lead to dominance hierarchies, with larger or stronger nestlings suppressing the begging signals of smaller nestlings or displacing them from positions near the parent where their begging signals would attract the parent’s attention more effectively (Mock & Parker, 1997). Hierarchies, however, are only one of a variety of network architectures that might arise. Speaking more generally, Glassey & Forbes (2002b) noted that nestlings can often be divided into ‘core’ and ‘marginal’ nestlings (Mock & Forbes, 1995). Survival of core nestlings is usually predictable, whereas marginal nestlings, which may be smaller, in poorer condition, younger, subordinate and/or less able to thermoregulate, survive only if ecological conditions are favourable. This ‘structured sibship’ (Glassey & Forbes, 2002b) may yield three different sorts of nestling relationships within the brood: core to core, marginal to marginal, and core to marginal (Glassey & Forbes, 2002b). Variation among species in this underlying structure will affect physical competition and signalling interactions within the nest. For example, one core and one marginal nestling might yield a simple dominance hierarchy, whereas three nestlings in each category might yield two ‘cliques’ of nestlings, between which there is a dominance hierarchy but within which signalling behaviours are similar and physical competition is equitable. In any case, the underlying structure of relationships within the brood, even though they do not consist of signalling relationships, nevertheless may strongly affect the structure of the overlying communication network – no doubt a recurring theme for most communication networks (e.g. Chs. 10 and 25). Summary

We have raised three complexities in applying the concept of communication networks to nestling begging. First, characterizing the communication network entails a difficult distinction between signalling and physical interactions. Second, demonstrating some of the more interesting effects of communication

179

P1: JZZ/... P2: JZZ/... 0521823617c09.xml CU1917B/McGregor

180

0 521 582361 7

April 11, 2005

12:20

A. G. Horn & M. L. Leonard networks entails another difficult distinction: between signalling to the parent and signalling interactions with nestmates. Third, any realistic description of the communication network must include interrelationships among nestmates that do not involve signalling but nevertheless may shape the structure of the network. These particular issues, of course, have less of an impact on communication in some other kinds of network. Territorial birds singing from their song posts, for example, are far beyond the range of physical interaction, are clearly directing their signals at each other (but see Ch. 14), and are generally unrelated to one another. Nonetheless, the issues we have raised are not unique to begging nestlings. Even territorial birds, for example, can engage in close-range interactions that combine signals with direct aggression, sing in ways that can be seen either as signalling interactions or as attempts to overcome interfering signals, and have dominance relationships that structure their communication network. If communication networks are indeed ‘the commonest social environment in which communication occurs’(McGregor & Peake, 2000), then network concepts will inevitably be applied to other systems that do present some of the complications we have discussed to varying degrees. If we are to understand how these networks function, we need to clarify these issues and begging should prove to be a particularly useful system for doing so.

Consequences of the network for begging We now turn from attempting to characterize the communication network within the nest to exploring how it might affect communication, from both signallers’ and receivers’ perspectives. Most discussions of communication networks have emphasized two consequences in particular (e.g. McGregor & Peake, 2000; McGregor et al., 2000; see also other chapters in this volume) and we begin with these. First, from the signaller’s point of view, signals must be designed to catch the receiver’s attention in the face of interference from other signals in the network. Second, receivers, for their part, can more readily compare signallers in a network because they are in transmission range of several signallers at once. A third possible consequence has received less attention: communication networks might reduce error in the information that signals convey. Specifically, as we explain below, nestlings are particularly error prone in deciding when and how intensely to beg. When nestlings partly base these decisions on the behaviour of other nestlings, as they can when signalling within a network, these errors might have less effect on their signals of need. Design to catch receiver attention

McGregor and Peake (2000) suggested that the main effect of networks on signal design arises through competition for receiver attention, as each signaller

P1: JZZ/... P2: JZZ/... 0521823617c09.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

12:20

Nestling begging as a communication network attempts to circumvent the interference caused by competing signals in the network. Perhaps no other communication system is more obviously a competition for receiver attention than a brood of noisy nestlings. Given the interest in the exaggeration of this signal and its role in nestmate competition, however, there are surprisingly few studies that specifically address how begging signals are designed to overcome interference from nestmates. Our understanding of begging and nestling competition might be considerably enhanced by thinking of begging nestlings as a communication network. In particular, we suspect that many of the most striking characteristics of the begging display may be designed for overcoming interference. If so, then the conspicuousness and complexity of the display, which seems unnecessarily extravagant for such a short-range signal, may, in fact, be a proportionate response to signal interference (Dawkins & Guilford, 1997; Horn & Leonard, 2002). Here we briefly discuss how selection for overcoming interference might account for a few of the more obvious features of begging (see also Horn & Leonard, 2002). High output

The most straightforward way to overcome any background noise is to increase the amplitude or duty cycle of one’s signal. There is ample evidence that nestlings respond in this way to signalling by nestmates (Budden & Wright, 2001; but see Cotton et al., 1996). For example, nestlings in some species beg more intensely when placed near a begging nestmate (Leonard & Horn, 1998) and call at higher rates when they can hear a nestmate calling (Leonard & Horn, 2001c). Locatable signals

Surprisingly small apparent angular separation between stimuli can significantly enhance a receiver’s ability to tell them apart (Ch. 20). Thus design features that enhance the locatability of nestlings are likely to enhance how well they stand out from competing signals and so focus parental attention on an individual nestling. The visual components of begging, brightly coloured gapes in particular, seem designed to be readily locatable targets for parental attention. These gapes have particularly bright outlines in species that nest in cavities, most likely so that the location of each nestling’s gape is distinct despite the darkness (Kilner & Davies, 1998; Heeb et al., 2003). Begging calls, in contrast, do not seem as obviously suited for locating nestlings because they are broadcast noisily throughout the nest. Also, there is little evidence so far that their structures are individually distinct in ways that would make them easy for parents to distinguish (Leonard & Horn, 2001c; but see Popp & Ficken, 1991). Indeed, some theoretical models suggest that they should not be individually distinct because that would risk rejection by the parent (Beecher, 1991; Johnstone, 1997).

181

P1: JZZ/... P2: JZZ/... 0521823617c09.xml CU1917B/McGregor

182

0 521 582361 7

April 11, 2005

12:20

A. G. Horn & M. L. Leonard

Fig. 9.1. Spectrograms of three nestling begging calls: tree swallow Tachycineta bicolor, hairy woodpecker Picoides pubescens and white-browed scrubwren Sericornis frontalis. Vertical bar is 10 kHz, horizontal bar is 500 milliseconds and filter bandwidth is 700 Hz.

Nonetheless, many calls do display features thought to enhance locatability, including abrupt onsets and offsets, broad frequency ranges and use of frequencies to which parents are most acutely tuned (Horn & Leonard, 2002; Fig. 9.1). Whether these features really do enhance locatability within the confines of a nest has not been tested directly. Comparative evidence, however, suggests that begging calls do display some of these features, except when subject to counteracting selective pressure from predators that use locatable calls to find and depredate nests (Haskell, 2002; Horn & Leonard, 2002). Multiple components

Which features of signals stand out from the noise of competing signals will depend on the situation, and the multiple components of the begging display may allow nestlings to signal effectively in each of these different situations. For example, a nestling competing with a nestmate in the front of a cavity nest might gain more from gaping wider and posturing more intensely than a nestling stuck in the back of the nest, because the nestling in the front is in plain sight of the parent. In contrast, a nestling in the back of the box cannot be clearly seen by the parent and would probably gain more from large increases in call rate than from any changes in the visual signal (Leonard et al., 2003). Therefore, in addition to the numerous other psychological advantages of multimodal signal components (Rowe, 1999), they may provide nestlings with a toolbox of ways to make their signal stand out despite changing conditions.

P1: JZZ/... P2: JZZ/... 0521823617c09.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

12:20

Nestling begging as a communication network Precedence

Precedence effects, the tendency of receivers to take more notice of signals that occur first, may favour signallers that signal before their competitors do (McGregor et al., 2000). Note that such effects, as shown in insect and frog choruses, for example, may (Greenfield, 2002) or may not (Gerhardt & Huber, 2002) be the result of certain psychological effects also known as precedence (for which Dent & Dooling (2003a,b) provide an avian example). Begging may provide a particularly good example of this effect on signalling. Parents in a wide range of species are more likely to feed nestlings that beg before their nestmates (Budden & Wright, 2001) and nestlings appear to have been selected for hair-trigger responses to the first sign of the parent’s arrival (Leonard & Horn, 2001d). The importance of precedence effects may vary considerably among species, providing interesting opportunities for comparative tests of their effects on signalling. For example, they may be less important in species in which parents spend more time assessing begging signals at each visit (e.g. Krebs, 2002) or in which hasty responses by nestlings might waste energy or attract predators (Leonard & Horn, 2001d). Signal suppression

All the aspects of signal design we have outlined so far can overcome signal competition by enhancing the signaller’s own signal. Signallers might also, however, overcome competition by suppressing the signals of competitors. For example, nestling whydahs Vidua spp. spread their wings to block their parents’ view of nestmate signals (B. Mines, personal communication) and dominant nestlings of many non-passerine species aggressively punish subordinate nestmates that beg in their presence (Drummond, 2002; Roulin, 2002). Subtler versions of such direct approaches to signal competition may be widespread and should be looked for in other species. Comparison among signals

A second consequence of communication networks is that they allow receivers to compare information from several signallers at once. Social eavesdropping, extracting information from a signalling interaction (Ch. 2), is a particularly interesting special case of such comparisons. However, receivers might also benefit from the network simply by being able to compare signals simultaneously rather than having to assess each signaller in succession (Chs. 7 and 14). Surprisingly, how or even whether parents compare begging signals to decide which nestling to feed is still poorly understood. Many studies, using various measures of begging intensity, have shown that more intensely begging nestlings are more likely to be chosen, but such evidence is only correlational. Only a few recent studies have experimented on parental choice in sufficient detail to separate

183

P1: JZZ/... P2: JZZ/... 0521823617c09.xml CU1917B/McGregor

184

0 521 582361 7

April 11, 2005

12:20

A. G. Horn & M. L. Leonard the roles of non-signalling and signalling components of begging, or to demonstrate preferences based on individual components of the begging display (Horn & Leonard, 2002; Kilner, 2002a,b). Demonstrating whether parents use information from signalling interactions among nestlings will require still more refined experiments (see above). Interestingly, recent models suggest that parents must assess interactions among nestlings if begging is to evolve as a signal at all (Rodr´ıguez-Giron´es et al., 2001; Royle et al., 2002). Specifically, if parents simply select the most obvious signal, then the information content of begging becomes irrelevant and begging consists merely of a scramble for the parents’ attention. If, however, parents can calibrate the information in the signals to correct for competitive differences among nestlings, whether those are expressed via signalling (e.g. Roulin, 2002) or physical competition, then begging can indeed convey information on need (Rodr´ıguez-Giron´es et al., 2001; Royle et al., 2002). Under this scenario, a network environment may have been of central importance in the evolution of begging. Error reduction

The last possible consequence of communication networks that we will discuss has received little attention, although it seems simple in principle and broad in implications. Specifically, because information in a network is transferred via not just one but several signals, the impact of error from any given signal might be reduced. To explain this possibility, we first outline some possible sources of error in begging displays and then discuss how the communication network may reduce this error. Begging by individual nestlings may be considerably error prone for at least two reasons (Clark, 2002; Horn & Leonard, 2002). First, nestlings may be poor at assessing their own needs, especially since doing so requires integrating their current condition with their future requirements and their likely returns from begging, both of which are partly under control of their parents, their nestmates and the vagaries of the environment (Clark, 2002). Second, nestlings are often poor at distinguishing the parent’s arrival at the nest from other sights and sounds and, therefore, often beg in response to irrelevant stimuli. In older tree swallow Tachycineta bicolor nestlings, for example, while nestlings often simply start begging after their nestmates do, many of the initial begging responses are to events other than the parent, like the wind blowing through the trees or the bump of another bird species landing by the nest (Leonard & Horn, 2001d; Horn & Leonard, 2002). Conversely, nestlings apparently hold back on begging when they are unsure whether the parent actually has arrived and so may miss the parent’s arrival or may send an inappropriately weak signal (Clark, 2002). From the nestling’spoint of view, these are errors in how they deliver the begging signal. From the parent’s

P1: JZZ/... P2: JZZ/... 0521823617c09.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

12:20

Nestling begging as a communication network point of view, however, such errors corrupt any information that the parent might obtain from the begging signal. Begging in a network may buffer such errors, because each nestling bases its decision of when and how intensely to beg partly on the begging of its nestmates (e.g. Leonard & Horn, 1998, 2001c). This influence of nestmates should reduce the impact of the errors that each nestling would make if it were begging on its own; from the parent’spoint of view, it would provide a more reliable signal of offspring need (Clark, 2002). This argument could, of course, be reversed. Specifically, one might argue that the more links in the information chain from nestlings to parents, the less accurate and reliable information will be (Royle et al., 2002). Determining whether networks reduce or increase error requires modelling of information flow through the network. A nestling’s decision of when to beg, to take the first step in the chain, may be seen as a game of signal detection, in which the nestling can either try to be the first to detect the parent’s arrival, at the risk of more false alarms (as shown above for tree swallows), or can free-ride by eavesdropping on the responses of nestmates, at the risk of begging later than its nestmates (Erev et al., 1995). Notwithstanding the promise of such models, probably the most pressing need for understanding information flow through networks, indeed for all the possible consequences of the begging network surveyed above, is for more empirical research on how parents assess begging signals.

Summary and future directions In this chapter we have tried to show that begging by nestling birds is a promising system for clarifying fundamental aspects of communication networks, particularly the grey but conceptually fruitful areas between physical acts and signals, between signalling competitively and interacting, and between communication and other functional relationships among signallers. Theoretical work on the evolution of begging has already started exploring each of these areas, but it has been inspired more by field workers’ insistence that begging behaviour is complex than by any attempt to treat begging as a communication network. In the future, theoretical work would likely benefit from a more explicit application of network concepts, much as studies of economics and cooperation in humans have benefited from models of social networks (e.g. Slikker & van den Nouweland, 2001). Conversely, those studying other communication networks will likely benefit from staying abreast of theoretical developments in the study of begging. Perhaps the greatest opportunities for future work, however, are in empirical studies that focus on signalling and nestmate interactions in more detail. Despite

185

P1: JZZ/... P2: JZZ/... 0521823617c09.xml CU1917B/McGregor

186

0 521 582361 7

April 11, 2005

12:20

A. G. Horn & M. L. Leonard enormous variation in the form of begging calls within and across species, for example (Popp & Ficken, 1991), only a handful of studies have addressed the function of this variation in any detail (Horn & Leonard, 2002; see also Kilner (2002b) for the display as a whole). Similarly, despite a long history of interest in intrabrood competition in birds (Mock & Parker, 1997), few studies have tried to identify the specific functions of the various behaviours that nestlings use in competition, especially what information they might convey to both parents and nestmates (Clark, 2002; Roulin, 2002). Perhaps most importantly, how, or even whether, parents choose which nestling to feed remains largely unknown because the requisite experiments have not been done (Royle et al., 2002). Hopefully, greater appreciation that nestlings communicate within a network of signallers, with all its attendant challenges and opportunities, will inspire more research on all of these fundamental questions.

Acknowledgements We thank Pete McGregor for the chance to contribute to this volume and for his patience and constructive advice during the preparation of this chapter. We also thank John Bower and an anonymous reviewer for their helpful comments on an earlier draft. Conversations with participants at the Gregynog 2000 Begging Workshop and with authors of the resultant book have been invaluable in developing our ideas about begging, as has our collaboration with Rob Magrath and his students at the Australian National University and with the many students and assistants who have worked on tree swallows with us. We also thank the Coldwell, Hines and Minor families for allowing us to work on their land, and NSERC for financial support.

References Beecher, M. D. 1991. Successes and failures of kin recognition in animals. In: Kin Recognition, ed. P. G. Hepper. Cambridge, UK: Cambridge University Press, pp. 94–124. Bergstrom, C. T. & Lachmann, M. 1998. Signaling among relatives. III. Talk is cheap. Proceedings of the National Academy of Sciences, USA, 95, 5100–5105. Briskie, J. V., Naugler, C. T. & Leech, S. M. 1994. Begging intensity of nestling birds varies with sibling relatedness. Proceedings of the Royal Society of London, Series B, 258, 73–78. Budden, A. E. & Wright, J. 2001. Begging in nestling birds. Current Ornithology, 16, 83–118. Clark, A. B. 2002. Appetite and the subjectivity of hunger. In: The Evolution of Begging: Competition, Cooperation, and Communication, ed. J. Wright & M. L. Leonard. Dordrecht: Kluwer, pp. 173–198.

P1: JZZ/... P2: JZZ/... 0521823617c09.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

12:20

Nestling begging as a communication network Clotfelter, E. D., Shubert, K. A., Nolana, V. Jr & Ketterson, E. D. 2003. Mouth color signals thermal state of nestling dark-eyed juncos ( Junco hyemalis). Ethology, 109, 171–182. Cotton, P. A., Kacelnik, A. & Wright, J. 1996. Chick begging as a signal: are nestlings honest? Behavioral Ecology, 7, 178–182. Cotton, P. A., Wright, J. & Kacelnik, A. 1999. Chick begging strategies in relation to brood hierarchies and hatching asynchrony. American Naturalist, 153, 412–420. Dawkins, M. S. & Guilford, T. 1997. Conspicuousness and diversity in animal signals. Perspectives in Ethology, 12, 55–72. Dearborn, D. C. & Lichtenstein, G. 2002. Begging behaviour and host exploitation in parasitic cowbirds. In: The Evolution of Begging: Competition, Cooperation, and Communication, ed. J. Wright & M. L. Leonard. Dordrecht: Kluwer, pp. 361– 387. Dent, M. K. & Dooling, R. J. 2003a. Investigations of the precedence effect in budgerigars: effects of stimulus type, intensity, duration, and location. Journal of the Acoustical Society of America, 113, 2146–2158. 2003b. Investigations of the precedence effect in budgerigars: the perceived location of auditory images. Journal of the Acoustical Society of America, 113, 2159–2169. Drummond, H. 2002. Begging versus aggression in avian broodmate competition. In: The Evolution of Begging: Competition, Cooperation, and Communication, ed. J. Wright & M. L. Leonard. Dordrecht: Kluwer, pp. 337–360. Erev, I., Gopher, D., Itkin, R. & Greenshpan, Y. 1995. Toward a generalization of signal detection theory to n-person games: the example of two person safety problem. Journal of Mathematical Psychology, 39, 360–375. Evans, R. M. 1994. Cold-induced calling and shivering in young American white pelicans: honest signalling of offspring need for warmth in a functionally integrated thermoregulatory system. Behaviour, 129, 14–34. Gerhardt, H. C. & Huber, F. 2002. Acoustic Communication in Insects and Anurans. Chicago, IL: University of Chicago Press. Glassey, B. & Forbes, S. 2002a. Muting individual nestlings reduces parental feeding for the brood. Animal Behaviour, 63, 779–786. 2002b. Begging and asymmetric nestling competition. In: The Evolution of Begging: Competition, Cooperation, and Communication, ed. J. Wright & M. L. Leonard. Dordrecht: Kluwer, pp. 269–282. Godfray, H. C. J. 1991. The signalling of need by offspring to their parents. Nature, 353, 328–330. 1995. Evolutionary theory of parent–offspring conflict. Nature, 376, 133–138. Godfray, H. C. J. & Johnstone, R. A. 2000. Begging and bleating: the evolution of parent–offspring conflict and sibling rivalry. Philosophical Transactions of the Royal Society of London, Series B, 355, 1581–1591. G¨ otmark, F. & Ahlstr¨ om, M. 1997. Parental preferences for red mouth of chicks in a songbird. Proceedings of the Royal Society of London, Series B, 264, 959–962. Greenfield, M. D. 2002. Signalers and Receivers: Mechanisms and Evolution of Arthropod Communication. Oxford: Oxford University Press.

187

P1: JZZ/... P2: JZZ/... 0521823617c09.xml CU1917B/McGregor

188

0 521 582361 7

April 11, 2005

12:20

A. G. Horn & M. L. Leonard Haskell, D. G. 2002. Begging behaviour and nest predation. In: The Evolution of Begging: Competition, Cooperation, and Communication, ed. J. Wright & M. L. Leonard. Dordrecht: Kluwer, pp. 163–172. Heeb, P., Schwander, T. & Faoro, S. 2003. Nestling detectability affects parental feeding preferences in a cavity-nesting bird. Animal Behaviour, 66, 637–642. Hill, R. W. & Beaver, D. L. 1982. Inertial thermostability and thermoregulation in broods of redwing blackbirds. Physiological Zoology, 55, 250–260. Horn, A. G. & Leonard, M. L. 2002. Efficacy and the design of begging signals. In: The Evolution of Begging: Competition, Cooperation, and Communication, ed. J. Wright & M. L. Leonard. Dordrecht: Kluwer, pp. 127–142. Johnstone, R. A. 1997. Recognition and the evolution of distinctive signatures: when does it pay to reveal identity? Proceedings of the Royal Society of London, Series B, 264, 1547–1553. 1999. Signaling of need, sibling competition, and the cost of honesty. Proceedings of the National Academy of Sciences, USA, 96, 12644–12649. Johnstone, R. A. & Godfray, H. C. J. 2002. Models of begging as a signal of need. In: The Evolution of Begging: Competition, Cooperation, and Communication, ed. J. Wright & M. L. Leonard. Dordrecht: Kluwer, pp. 1–20. Kilner, R. M. 1997. Mouth colour is a reliable signal of need in begging canary nestlings. Proceedings of the Royal Society of London, Series B, 264, 963–968. 2002a. The evolution of complex begging displays. In: The Evolution of Begging: Competition, Cooperation, and Communication, ed. J. Wright & M. L. Leonard. Dordrecht: Kluwer, pp. 87–106. 2002b. Sex differences in canary (Serinus canaria) provisioning rules. Behavioral Ecology and Sociobiology, 52, 400–407. Kilner, R. M. & Davies, N. B. 1998. Nestling mouth colour: ecological correlates of a begging signal. Animal Behaviour, 56, 705–712. Komdeur, J. & Hatchwell, B. J. 1999. Kin recognition: function and mechanism in avian societies. Trends in Ecology and Evolution, 14, 237–241. Krebs, E. A. 2002. Sibling competition and parental control: patterns of begging in parrots. In: The Evolution of Begging: Competition, Cooperation, and Communication, ed. J. Wright & M. L. Leonard. Dordrecht: Kluwer, pp. 319–336. Leonard, M. L. & Horn, A. G. 1998. Need and nestmates affect begging in tree swallows. Behavioral Ecology and Sociobiology, 42, 431–436. 2001a. Begging calls and parental feeding decisions in tree swallows (Tachycineta bicolor). Behavioral Ecology and Sociobiology, 49, 170–175. 2001b. Acoustic signalling of hunger and thermal state by nestling tree swallows. Animal Behaviour, 61, 87–93. 2001c. Dynamics of calling by tree swallow (Tachycineta bicolor) nestmates. Behavioral Ecology and Sociobiology, 49, 170–175. 2001d. Begging in the absence of parents by nestling tree swallows. Behavioral Ecology, 12, 501–505. Leonard, M. L., Horn, A. G. & Parks, E. 2003. The role of calling and posturing in the begging behavior of nestling birds. Behavioral Ecology and Sociobiology, 54, 188– 193.

P1: JZZ/... P2: JZZ/... 0521823617c09.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

12:20

Nestling begging as a communication network Lotem, A., Wagner, R. H. & Balshine-Earn, S. 1999. The overlooked signaling component of nonsignaling behavior. Behavioral Ecology, 10, 209–212. Maynard Smith, J. 1994. Must reliable signals always be costly? Animal Behaviour, 42, 1034–1035. McGregor, P. K. & Dabelsteen, T. 1996. Communication networks. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca, NY: Cornell University Press, pp. 409–425. McGregor, P. K. & Peake, T. M. 2000. Communication networks: social environments for receiving and signalling behaviour. Acta Ethologica, 2, 71–81. McGregor, P. K., Otter, K. & Peake, T. M. 2000. Communication networks: receiver and signaller perspectives. In: Signalling and Signal Design in Animal Communication, ed. Y. Espmark, T. Amundsen & G. Rosenqvist. Trondheim: Tapir Academic Press, pp. 329–340. Mock, D. W. & Forbes, L. S. 1995. The evolution of parental optimism. Trends in Ecology and Evolution, 10, 130–134. Mock, D. W. & Parker, G. A. 1997. The Evolution of Sibling Rivalry. Oxford: Oxford University Press. Ovadia, O., Pinshow, B. & Lotem, A. 2002. Thermal imaging of house sparrow nestlings: the effect of begging behavior and nestling rank. The Condor, 104, 837–842. Parker, G. A., Royle, N. J. & Hartley, I. R. 2002. Intrafamilial conflict and parental investment: a synthesis. Philosophical Transactions of the Royal Society of London, Series B, 357, 295–307. Peake, T. M., Terry, A. M. R., McGregor, P. K. & Dabelsteen, T. 2002. Do great tits assess rivals by combining direct experience with information gathered by eavesdropping? Proceedings of the Royal Society of London, Series B, 269, 1925–1929. Popp, J. & Ficken, M. S. 1991. Comparative analysis of acoustic structure of passerine and woodpecker nestling calls. Bioacoustics, 3, 255–274. Price, K., Harvey, H. & Ydenberg, R. 1996. Begging tactics of nestling yellow-headed blackbirds, Xanthocephalus xanthocephalus, in relation to need. Animal Behaviour, 51, 421–435. Price, K., Ydenberg, R. & Daust, D. 2002. State-dependent begging with asymmetries and costs: a genetic algorithm approach. In: The Evolution of Begging: Competition, Cooperation, and Communication, ed. J. Wright & M. L. Leonard. Dordrecht: Kluwer, pp. 21–42. Redondo, T. & Zu˜ niga, J. M. 2002. Dishonest begging and host manipulation by Clamator cuckoos. In: The Evolution of Begging: Competition, Cooperation, and Communication, ed. J. Wright & M. L. Leonard. Dordrecht: Kluwer, pp. 389–412. Rodr´ıguez-Giron´es, M. A. 1996. Siblicide: the evolutionary blackmail. American Naturalist, 148, 101–122. Rodr´ıguez-Giron´es, M. A., Enquist, M. & Lachmann, M. 2001. Role of begging and sibling competition in foraging strategies of nestlings. Animal Behaviour, 61, 733–745. Roulin, A. 2002. The sibling negotiation hypothesis. In: The Evolution of Begging: Competition, Cooperation, and Communication, ed. J. Wright & M. L. Leonard. Dordrecht: Kluwer, pp. 107–126.

189

P1: JZZ/... P2: JZZ/... 0521823617c09.xml CU1917B/McGregor

190

0 521 582361 7

April 11, 2005

12:20

A. G. Horn & M. L. Leonard Rowe, C. 1999. Receiver psychology and the evolution of multicomponent signals. Animal Behaviour, 58, 921–931. Royle, N. J., Hartley, I. R. & Parker, G. A. 2002. Begging for control: when are offspring solicitation behaviours honest? Trends in Ecology and Evolution, 17, 434–440. Saino, N. & Møller, A. P. 2002. Immunity and begging. In: The Evolution of Begging: Competition, Cooperation, and Communication, ed. J. Wright & M. L. Leonard. Dordrecht: Kluwer, pp. 245–267. Saino, N., Ninni, P., Calza, S. et al. 2000. Better red than dead: carotenoid based mouth coloration reveals infection in barn swallow nestlings. Proceedings of the Royal Society of London, Series B, 267, 57–61. Slikker, M. & van den Nouweland, A. 2001. Social and Economic Networks in Cooperative Game Theory. Dordrecht: Kluwer. Trivers, R. L. 1974. Parent–offspring conflict. American Zoologist, 14, 249–264. Whittingham, L. A. & Dunn, P. 2001. Male parental care and paternity. Current Ornithology, 16, 257–298. Wilson, D. S. & Clark, A. B. 2002. Begging and cooperation: an exploratory flight. In: The Evolution of Begging: Competition, Cooperation, and Communication, ed. J. Wright & M. L. Leonard. Dordrecht: Kluwer, pp. 107–126.

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

14:27

10

Redirection of aggression: multiparty signalling within a network? anahita j. n. kazem1 & filippo aureli2 1 2

University of Wales, Bangor, UK Liverpool John Moores University, UK

Introduction In many species, an individual that finds itself in a losing position may interrupt a contest to harass a bystander (i.e. an apparently uninvolved third party) aggressively or may do so immediately after hostilities with the original opponent have ceased. Such ‘redirection’ of aggression (Bastock et al., 1953) is often interspecific; for example, rollers and chasseur-type kingfishers (Coraciiformes) are reported to dash away frequently during disputes to attack small passerines, doves and plovers (Moynihan, 1998). The scapegoats are typically not ecological competitors but do tend to be smaller and inoffensive individuals, both literally and figuratively, and thus relatively safe targets. In socially living taxa, however, redirection is most commonly directed towards a lower-ranking group member (where available) and, therefore, is usually intraspecific. Both aspects of redirection are conveyed by the description of tensions between spotted hyaenas Crocuta crocuta at a kill, producing a cascade of aggression, in which ‘A chases B, B chases C, C chases D, and D chases vultures’ (Zabel et al., 1992, p. 129). Redirection of aggression has traditionally been explained as a means of reducing the physiological arousal associated with participation in a conflict. The neuroendocrine responses underlying the preparation for ‘fight or flight’, whilst essential in the immediate context, can be detrimental if they remain activated over prolonged periods. Chronically elevated secretion of glucocorticoids, for example, is associated with a range of cardiovascular pathology, depressed immune function and compromised digestion, growth and reproduction (reviewed by Sapolsky, 1998). Any action that prompts the endocrine stress response to Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

191

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

192

0 521 582361 7

April 11, 2005

14:27

A. J. N. Kazem & F. Aureli attenuate to baseline levels more rapidly can, therefore, reduce the physiological costs of being victimized. When an animal is unable to retaliate directly against an aggressor (for example because it is a higher-ranking or otherwise superior competitor), redirecting offers an outlet for ‘frustration’ and allows the actor to perceive a measure of ‘control’over the social situation. In experimental studies on rodents and primates, both psychological effects modulate the elevation in heart rate and glucocorticoid levels that repeated exposure to an unavoidable physical stressor usually produces (reviewed by Levine et al., 1989; Sapolsky, 1998). Furthermore, rats Rattus norvegicus given the opportunity to attack a conspecific when subjected to a mild electric shock subsequently developed fewer gastric lesions than yoked controls that were not provided with this outlet (Weiss et al., 1976). Amongst wild olive baboons Papio anubis, males that readily initiate aggression in appropriate contexts and that frequently respond to losing fights by redirecting aggression against others have significantly lower basal glucocorticoid levels (and a better response to acute challenge) than similarly ranked individuals that do not exhibit these behaviour patterns as frequently (Sapolsky & Ray, 1989; Virgin & Sapolsky, 1997). However, the extent to which this can be attributed to beneficial consequences of redirection per se is not clear, as it was only one of a suite of traits concerning temperamental style in handling male–male competition that characterized these individuals. Nevertheless, recent experimental work has confirmed that redirection does have an inhibitory effect on neuroendocrine stress responses in rainbow trout Oncorhynchus mykiss (Øverli et al., 2004). Here we summarize evidence suggesting there may, in some species, be more to redirection than this. The behaviour is particularly prevalent and has been best demonstrated in primates, notably several members of the genus Macaca (Table 10.1). In these species, harassing a conspecific would seem a rather costly way of gaining stress relief. Aggression carries the risk of injury, attracting predators and, despite a careful choice of target, may still provoke retaliation from its allies. Why harass bystanders rather than quietly chew on wood (which has beneficial effects in rats (Weiss et al., 1976))? Naturally, redirection may confer benefits in terms of resource acquisition or reinforcing one’s status over the target; yet this would not explain why individuals are particularly likely to instigate such aggression immediately after losing a conflict: a time when their energy reserves may be depleted and the former opponent likely to join in coalition against them. Instead, the answer may lie in how redirection influences the behaviour of bystanders, rather than the target. A number of additional hypotheses have been put forward to explain redirection behaviour in macaques: for example that it diverts the aggressor’s attention (Itani, 1963), provides an opportunity for the two opponents to resolve their differences by joining forces against a common foe (de Waal & Yoshihara, 1983) or encourages the former aggressor to participate in a conciliatory

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

14:27

Redirection of aggression: multiparty signalling Table 10.1. Non-human primate studies of intraspecific redirected aggression within intact social groups Species

Holding conditions Redirection Evidencea Source

Barbary macaque,

Captive

Yes

1a

Aureli et al., 1994

Japanese macaque,

Captive

Yes

1c

Aureli et al., 1992b

Macaca fuscata

Captive

Yes

1a

Aureli et al., 1993

Captive

Yes

2

Eaton, 1984; Scucchi et al.,

Wild and captive

Yes

1a

Aureli & van Schaik, 1991a;

Macaca sylvanus

1988 Longtailed macaque,

Aureli, 1992

Macaca fascicularis Pigtail macaque,

Captive

Yes

1b

Judge, 1982b

Free ranging and

Yes

1a

de Waal & Yoshihara,

Macaca nemestrina Rhesus macaque, Macaca mulatta

1983;c Kazem, 1999

captive Captive

Yes

2

Gore, 1994

Stumptail macaque, Captive

Yes

2

Walker Leonard, 1979

Yes

1a

F. Zaragoza &

Macaca arctoides Hamadryas baboon,

Captive

Papio hamadryas

F. Colmenares, unpublished data

Olive baboon, Papio anubis

Captive (females)

No

2

Gore, 1994

Wild (females)

No

1a

Castles & Whiten, 1998a

Wild (males)

Yes

2

Sapolsky & Ray, 1989; Virgin & Sapolsky, 1997

Sooty mangabey,

Captive

Yes

1a

Gust & Gordon, 1993

Wild

Yes

1b,c

Cheney & Seyfarth, 1986,

Cercocebus torquatus atys Vervet monkey,

1989b,c

Cercopithecus aethiops Spectacled langur,

Captive

No

1a

Arnold & Barton, 2001

Wild (males and

Yes

1a

Watts, 1995bd

No

1a

Watts, 1995b

Trachypithecus obscurus Mountain gorilla, Gorilla gorilla beringei

immatures) Wild (females)

(cont.)

193

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

194

0 521 582361 7

April 11, 2005

14:27

A. J. N. Kazem & F. Aureli Table 10.1. (cont.) Species

Holding conditions Redirection Evidencea Source

Common

Captive

Yes

1a

Fuentes et al., 2002e

Wild

Yes

2

S. E. Perry,

chimpanzee Pan troglodytes White-faced capuchin,

unpublished data

Cebus capucinus Black lemur,

Semi free ranging

No

1a

Roeder et al., 2002

Captive

No

1a

Roeder et al., 2002

Eulemur macaco Brown lemur, Eulemur fulvus a The

nature of quantitative evidence varies: 1, comparison of post-conflict period with matched-

controls (1a), with baseline focal observations or pre-conflict period (1b) or with other data (1c); 2, inference from quantitative but uncontrolled data. Type 2 data do not necessarily demonstrate that a significant post-conflict increase in aggression against bystanders exists. For Japanese, longtailed and rhesus macaques, additional anecdotal citations can be found in several other studies. b Analysis

of kin-oriented redirection, not overall incidence against all targets.

c Subject’s role

in previous conflict not distinguished; therefore, analyses potentially include for-

mer aggressors. d Result

for adult males was trend only, owing to sample size (n = 4).

e Conspecific

and human targets were not distinguished in analyses.

reunion with the actor (Aureli & van Schaik, 1991a). Although not explicitly couched in signalling terms, all rely on the principle that the former aggressor perceives the act and responds in ways that indirectly reduce the likelihood of further aggression against the redirecting individual. We develop this notion by proposing that redirection functions as a signal aimed at both the former opponent and other bystanders, which conveys information about the perpetrator’s competitive ability and current state and thus directly reduces challenges from these receivers. Such pre-emptive strikes offer a novel interpretation of the maxim that ‘offence is the best defence’.

The macaque system The results we describe are primarily drawn from three studies examining post-conflict behaviour in both captive and wild longtailed macaques Macaca fascicularis (Aureli & van Schaik, 1991a,b; Aureli, 1992), and in juveniles from two social groups within a free-ranging colony of rhesus macaques Macaca mulatta

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

14:27

Redirection of aggression: multiparty signalling (Kazem, 1999, A. J. N. Kazem, unpublished data). Both species form permanent multimale–multifemale groups and within the genus are regarded as possessing a relatively ‘despotic’ dominance style (de Waal & Luttrell, 1989; Thierry, 2000). Clear-cut dominance relationships are apparent in both sexes, and aggressive disputes are frequent, often injurious and overwhelmingly directed down the social hierarchy. Although individuals will defend relatives and close associates against attack, the most common pattern of intervention is support for the aggressor, with dominant group members receiving assistance from both kin and non-kin against subordinates that attempt to contravene the established hierarchy (Chapais, 1995). Unsurprisingly, direct retaliation by targets against aggressors is uncommon (e.g. less than 9% of conflicts involve counter-aggression in rhesus macaques (de Waal & Luttrell, 1989)). Anecdotal reports of redirected aggression abound in the literature on nonhuman primates, but the phenomenon has been explored and statistically confirmed in rather fewer species (Table 10.1). Demonstrating that a defeat influences the likelihood of initiating aggression against bystanders requires a comparison with the victim’s behaviour during an equivalent period not preceded by a contest. In naturalistic studies within intact social groups, this is typically achieved by comparing the immediate post-conflict period with a ‘matched-control’ observation collected on the next possible day (PC–MC method (de Waal & Yoshihara, 1983)) or, less frequently, with periods selected from a distribution of baseline focal observations on the same individual. Current best-practice protocols entail matching the conditions at the start of these paired samples with respect to factors likely to influence rates of aggression or other social interactions between relevant parties. These include the time of day, prevailing climatic conditions, predominant activity of both the subject and the wider group (if different) and, in analyses of interactions between former adversaries, inter-opponent distance (e.g. Aureli, 1992; Kazem, 1999). Studies differ in whether the demonstration of redirection emphasizes the occurrence of a single critical early post-conflict event (as in the PC–MC method), or simply compares overall rates of aggression initiated within a defined timeframe. Discussion of the relative merits of methods that can be used to demonstrate a significant post-conflict increase and/or operationally identify particular bouts of aggression as ‘redirection’ events, and the statistical issues involved, can be found in Veenema (2000), Das et al. (1997) and Kazem (1999). Characteristic post-conflict aggressive phenomena

Together our three studies provide a database of more than 2550 postconflict samples from 137 individuals, in which the subject had been the victim of unidirectional aggression with a clear outcome (i.e. the initial recipient was always the loser). Despite differences in species, holding conditions, group composition,

195

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

196

0 521 582361 7

April 11, 2005

14:27

A. J. N. Kazem & F. Aureli age of subjects and methods of analysis, the patterns documented were strikingly similar. Subjects threatened or attacked third parties significantly earlier and more frequently in the aftermath of a defeat than under control conditions. In both species, the increase was most pronounced within the first two minutes, quickly declining to baseline levels thereafter (Aureli & van Schaik, 1991a; Kazem, 1999). On average, rhesus victims harassed at least one bystander within the first minute after 22% of contests, corresponding to a 10-fold increase over control levels (an underestimate, as the methodology used excludes bouts initiated during the original conflict and individuals may redirect multiple times). The scapegoats were almost always lower ranking, but lack of a suitable conspecific did not necessarily stop individuals. For example, a young female rhesus victim, finding herself surrounded by members of the alpha matriline, proceeded to energetically and noisily pursue several lizards and a rat before peering intently into and repeatedly threatening a small bush (completely devoid of vertebrates), whilst continually attempting to solicit assistance from her former foe situated some distance away (A. J. N. Kazem, personal observation). Such scenes are not uncommon but were not operationally classified as redirection. By comparison, the post-conflict behaviour of former aggressors has received little attention. In rhesus macaques, victors also exhibit a post-conflict elevation in attacks against bystanders (and disproportionately target relatives of their former victim (Kazem, 1999)), while in longtailed macaques they do not (Das, 1998). Nevertheless, rhesus subjects were significantly more likely to harass bystanders after losing a conflict than after winning one, identifying redirection by former victims as the more distinctive and pervasive phenomenon. Another notable feature is that, having lost one contest, individuals are also liable to receive further aggression. In the aftermath of conflicts where the victim had neither reconciled with its opponent (i.e. engaged in a post-conflict affiliative reunion (Aureli et al., 2002)) nor redirected against a third party, defeated individuals were subjected to significantly elevated rates of threats and attacks in all three studies. The levels received were particularly high for the first three to four minutes (and at least 10 minutes in longtailed macaques), gradually waning toward baseline incidence over time. Interestingly, the renewal of hostilities by the former adversary was not the only cause; in many cases, the aggressor was a previously uninvolved bystander. Such increased receipt of aggression is a common finding, being the predominant sequence of events in triadic interactions between Japanese macaques Macaca fuscata (‘mobbing’; Eaton, 1984) and documented in controlled post-conflict studies of both macaques (e.g. de Waal & Yoshihara, 1983; Cords, 1992; Kutsukake & Castles, 2001) and other cercopithecines (e.g. mountain gorillas Gorilla gorilla beringei (Watts, 1995a) and olive baboons (Castles & Whiten, 1998a)). The effect is specific to former victims; neither the victor nor participants

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

14:27

Redirection of aggression: multiparty signalling in bidirectional contests suffer an enhanced risk (Castles & Whiten, 1998a; Das, 1998; Kazem, 1999). In the rhesus study, the number of aggressive incidents between other group members recorded within a 10 m radius of the focal animal did not differ consistently between post-conflict and control samples (A. J. N. Kazem unpublished data). This rules out the possibility that the increases observed in both initiation and receipt of aggression by victims could have been caused by the post-conflict samples being obtained during periods of generally heightened group aggressivity, and it confirms that control samples effectively matched postconflict conditions even in the more variable free-ranging situation. A consequence of loser effects and eavesdropping?

The most likely explanation for the temporal patterning of challenges from bystanders is that victims undergo some form of ‘loser experience’(cf. Scott & Fredericson, 1951), which renders them more easily beaten than under other circumstances. Research in a broad range of vertebrates has documented a tendency for animals who have suffered a defeat to lose in subsequent interactions against randomly selected and otherwise equally matched individuals (reviewed by Chase et al., 1994; Hsu & Wolf, 1999). In some cases, prior losers are even at a disadvantage against considerably smaller opponents – ones that they would normally be expected to defeat easily in any other encounter. To our knowledge, the standard protocols used to demonstrate this behavioural pattern (staging successive dyadic contests between unfamiliar individuals, with all factors other than the competitors’ prior social experiences held constant) have not been applied in a primate. However, in experiments in which novel triads (trios) of unfamiliar rhesus macaques were convened, the sequences of agonistic interactions observed were consistent with the operation of a loser (and indeed a winner) effect (Mendoza, 1993). Triad members had been matched for size, age, sex and activity level, making it unlikely that the predominance of consecutive losses (or wins) against different opponents was a result of pre-existing differences in intrinsic attributes. These patterns may arise from physiological changes precipitated by an individual’s experiences in a prior contest. During the initial minutes of an encounter, both contestants typically exhibit rapid, and often qualitatively similar, changes in central neurotransmitter activity (serotonergic, dopaminergic and noradrenergic) as well as increased secretion of adrenal axis hormones such as glucocorticoids and testosterone (e.g. van Erp & Miczek, 2000; Summers et al., 2003). However, as a contest progresses and the outcome becomes perceived, the neuroendocrine profiles of winner and loser diverge. Notably, in many vertebrates, central serotonin activity and peripheral glucocorticoid concentrations return to baseline levels relatively rapidly in victors, while greater initial increases and more

197

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

198

0 521 582361 7

April 11, 2005

14:27

A. J. N. Kazem & F. Aureli prolonged elevation are characteristic of defeated individuals (Schuurman, 1980; Hannes et al., 1984; Øverli et al., 1999; Summers et al., 2003). These differences are compounded if the protagonists remain confined together in the longterm, with subordinate individuals often exhibiting chronic elevation of these parameters (Blanchard et al., 1993; Gust et al., 1993; Winberg & Lepage, 1998). Winners are also reported to undergo increases in levels of circulating androgens such as testosterone, while those in losers appear temporarily suppressed (Bernstein et al., 1974; Rose et al., 1975; Hannes et al., 1984; Booth et al., 1989). In keeping with this, dominant animals generally possess higher basal testosterone levels than do subordinates (see Ch. 21), although neither the causal direction nor the cumulative influence of successive contests can be distinguished in such data. High levels of circulating testosterone are known to sharpen concentration and enhance social attention and memory processes (e.g. Andrews, 1991; Cynx & Nottebohm, 1992), and they are associated with both greater risk taking (Kavaliers et al., 2001) and expression of offensive aggression (e.g. Delville et al., 1996; Higley et al., 1996; Ch. 21). In contrast, serotonin generally exerts an inhibitory effect on aggressive behaviour. Experimentally enhancing central serotonin levels reduces an animal’s readiness to initiate aggressive acts (Olivier et al., 1995; Ferris et al., 1997; Perreault et al., 2003), while primates with chronically low serotoninergic functioning exhibit greater impulsivity, perseverance and use of severe unrestrained aggression (i.e. engage in aggression without regard for its consequences (Mehlman et al., 1994; Higley et al., 1996; Fairbanks et al., 2001)). It is, therefore, possible that the physiological changes typical of winners and losers produce transient alterations in factors that affect actual fighting ability. Alternatively, recent evidence in fish suggests that it may simply be an individual’s perception of its own relative ability that is modified (Hsu & Wolf, 2001), hence affecting subsequent decisions to initiate, escalate or withdraw. Either way, bystanders can take advantage of these changes to gain a temporary competitive edge over recent losers, offering an opportunity to reverse or reinforce an existing dominance relationship at relatively low cost. However, the behavioural consequences of a single prior loss are short lived (and species specific). They persist for only a matter of minutes or hours in many taxa, even after the prolonged and intense fights often characteristic of experimentally staged encounters between unfamiliar competitors (Chase et al., 1994; Hsu & Wolf, 1999). Timing may assume additional importance in species where coalitions are common, because a challenger’scosts will be further reduced if, by choosing this moment to attack, its actions are also likely to receive support from the victim’s former opponent (an event that may produce additional dividends by strengthening the assailant’s bonds with the latter individual). One means of being alerted to a possible loser effect is by attending to the outcome of conflicts between other group members. In some cases, simple cues might

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

14:27

Redirection of aggression: multiparty signalling betray a recent defeat (e.g. the dorsal darkening and flank patterns that appear in newly subordinate individuals in many fish), but in primates the only possibility would seem to be specific postural or behavioural changes only apparent after particularly severe losses. Instead, experimental evidence from a growing number of species suggests that third parties ‘eavesdrop’ (McGregor, 1993), extracting information on the relative fighting abilities of conspecifics from interactions they have witnessed (Ch. 2). Eavesdroppers subsequently treat perceived winners and losers differently, in ways consistent with having identified those individuals as relatively strong or weak competitors, respectively (e.g. Hogue et al., 1996; Oliveira et al., 1998; McGregor et al., 2001; Peake et al., 2001; Earley & Dugatkin, 2002). A multitude of evidence suggests that cercopithecines are similarly aware of the nature and outcome of contests in which they are not themselves involved. Macaques and baboons commonly exhibit apparent knowledge of the dominance relations between their groupmates (e.g. Silk, 1999; Ch. 25). In these species, the relative rank of third parties typically cannot be deduced directly from cues such as relative body size (especially in females) and is most likely derived by scrutinizing agonistic interactions between the individuals concerned. As an illustration, female chacma baboons Papio ursinus pay greater attention when presented with manipulated sequences of calls in which the affiliative grunt of a subordinate individual is closely followed by a scream from a higher-ranking female – a situation inconsistent with the existing dominance relationship between those particular group members – than they do to a control (and causally consistent) sequence (Cheney et al., 1995). Eavesdropping may also have contributed, in part, to the pattern of double-wins and double-losses observed in the rhesus experiment described above (Mendoza, 1993), given that the interactions occurred within a triad setting (see Chase et al. (2002) for similar logic). Eavesdropping is often regarded as a means of gathering information on an unknown competitor’s abilities without incurring the costs of directly confronting the party concerned. For example, a bystander might integrate information on how animal A fares against B with prior knowledge (gained via direct interaction) of its own standing relative to A, to extrapolate how it too might fare against the unknown B. Alternatively, even if informed only about the relative prowess of two strangers, this may still provide a probabilistic indication of how the defeated individual might rate relative to oneself and thus be useful in guiding behaviour. We suggest that another function of eavesdropping is simply to detect that an individual (known or unknown) has suffered a loss and, therefore, may be undergoing a loser experience. After all, in the cercopithecine systems being described, individuals are generally well aware of who outranks whom within the group; therefore, when contest outcome is in the expected direction, it may be the information about temporary changes in an individual’s current state that

199

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

200

0 521 582361 7

April 11, 2005

14:27

A. J. N. Kazem & F. Aureli is more important in the decision to challenge. Note that while eavesdropping can account for why known losers are more readily challenged than their victorious counterparts (as in the experiments cited above), it does not in itself explain why an individual should be challenged more frequently following a defeat than at other times (as in our data). Acquiring information on the relative abilities of potential competitors does not necessitate that it be put to use immediately; in contrast, the transient nature of a putative loser effect would require immediate use of the information. Redirection influences the behaviour of bystanders

In both longtailed and rhesus macaques, the amount of aggression a victim received in the minutes following a redirection event was significantly lower than that during comparable periods on occasions when it had lost but had not redirected (Aureli & van Schaik, 1991b; Kazem, 1999; A. J. N. Kazem, unpublished data). In other words, targeting third parties after a defeat appears to confer a protective effect. It is always possible that some feature of the preceding conflict influenced both the likelihood of redirection and of receiving subsequent aggression. Two results counsel against this view. First, in the rhesus study the likelihood of receiving post-conflict harassment in the period before redirection took place did not differ systematically between contests in which the victim did, or did not, go on to redirect (the individual’s mean latency to redirect was used to define the relevant timeframe in the latter). Second, rhesus youngsters redirected aggression more frequently after low-level contests (threats or minor lunges) than after more intense confrontations (involving prolonged pursuits or physical contact) – a point to which we shall return. However, less-intense disputes were not in themselves associated with receipt of low levels of subsequent harassment; victims incurred virtually identical rates of aggression following mild or severe incidents. Therefore, the reduction in harassment documented appears to be associated with the act of redirecting itself. It is not yet known whether the beneficial effect is achieved primarily via an alteration in the behaviour of the former opponent or in that of opportunistic bystanders. A change in the disposition of the former aggressor, at least, seems likely because redirection apparently influences its behaviour in other respects. Among longtailed macaques, redirecting is associated with an increased likelihood that the former aggressor will later participate in an affiliative reunion with the perpetrator (Aureli & van Schaik, 1991a), although a causal connection remains to be demonstrated. Reconciliation between former adversaries is known to have positive consequences in terms of restoring tolerance and reducing subsequent aggression between the protagonists (e.g. Cords, 1992); although note that this effect cannot have been indirectly responsible for the results we report above,

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

14:27

Redirection of aggression: multiparty signalling because post-conflict periods in which reconciliation had occurred were excluded from analyses. An honest indicator of post-conflict condition and motivation?

The evidence that redirection events may influence the disposition and behaviour of individuals other than the target of aggression suggests there is a signalling advantage to be gained by behaving antagonistically in post-conflict contexts. In essence, the primary benefit of the behaviour might derive from it being witnessed by third (or often, fourth) parties. It is often claimed that redirection serves its purpose by focusing attention upon an alternative target, thus cutting short the original contest and/or persuading potential challengers to look elsewhere (Itani, 1963; Gust & Gordon, 1993). Data supporting such an outcome are rarely, if ever, presented. This tactic may be successful in spotted hyaenas, social carnivores that exhibit a high degree of within-group coordination in activities and are liable to join in against any animal that is currently losing – to the point where supporters of an aggressor will often switch sides simply because their target counter-attacked its original opponent (Zabel et al., 1992). It seems less plausible in macaques, where there is little compelling reason why the former aggressor (at least) should so readily divert to a different target. Nor would it account for why aggressors seem more willing to reconcile with victims that have redirected against others. We propose that redirecting does more than merely draw attention to a new stimulus; the act itself may provide bystanders with useful information. As outlined above, the physiological and psychological consequences of a defeat generally reduce the likelihood that an individual will initiate aggression in the ensuing minutes, and they may render it less likely to persist and be less effective in combat when challenged. However, a recent victim that is nevertheless sufficiently confident and capable of rapidly redirecting against a third party thereby demonstrates (to the scapegoat, and more importantly to others) that it has not been unduly compromised by the preceding experience. The act may serve as an unfakeable marker of the perpetrator’spost-conflict state, indicating that it would be ready and/or able to defend itself, thus dissuading renewed or opportunistic challenges from bystanders. It might also make individuals that redirect appear more formidable rivals within the group, both to the former aggressors and their kin (see below; Aureli & van Schaik, 1991a), which could explain why former aggressors become more willing to reconcile. This hypothesis assumes that the neuroendocrine response to a perceived loss is not an all-or-nothing affair: that features of the prior contest (primarily its intensity and duration, and perhaps the opponent’s identity) can modulate the magnitude and/or type of changes undergone and the time course of recovery. The fact that rhesus macaques were much more likely to

201

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

202

0 521 582361 7

April 11, 2005

14:27

A. J. N. Kazem & F. Aureli redirect following receipt of low-intensity as opposed to high-intensity harassment (see above) is consistent with this argument; comparable results are also available in rainbow trout (Øverli et al., 2004). Furthermore, in many taxa, individuals differ consistently not only in their baseline levels of physiological parameters but also in the magnitude and nature of neuroendocrine response (‘reactivity’) produced by stressors such as received aggression (reviewed by Koolhaas et al., 1999). Consequently, the degree to which a victim has been compromised following any specific defeat is not straightforward for bystanders to surmise. Variation in post-conflict state provides both the impetus for conveying this information to potential challengers and the means (motivation or ability) to do so.

Ensuring others notice the event Redirection events possess several features likely to draw these acts to the attention of bystanders. For example, reports often emphasize that victims redirect ‘in front of ’ their former assailant (e.g. de Waal & Yoshihara, 1983; Aureli et al., 1992; but see Watts, 1995b). Bouts of redirection by young rhesus macaques in a free-ranging situation were also more likely to take place within 5 m of (and hence within view and earshot of ) their former opponents than were equivalent bouts instigated under control conditions, even when inter-opponent distance at the start of paired observation periods was statistically controlled (Kazem, 1999). As in other studies, it was not uncommon to observe subjects glancing back at their previous adversary both prior to and while threatening the target, implying that the aggressor’s presence (and perhaps even line of gaze: Emery et al., 1997; Tomasello et al., 1998) was actively taken into account. Experiments have demonstrated that vervet monkeys Cercopithecus aethiops and macaques often take the presence and composition of bystanders into account before behaving antagonistically toward others (Keddy Hector et al., 1989; Cheney & Seyfarth, 1990). Equally, a rule of thumb simply prompting victims to act quickly could ensure that their former adversary was likely to have remained nearby (the mean latency to redirect was 12 and 28 seconds in longtailed and rhesus macaques, respectively, excluding those bouts which occurred within the original conflict (Aureli & van Schaik, 1991b; Kazem, 1999)). In some cases, victims even approached and attempted to enlist their former opponent’s support in the venture with conspicuous head-flagging (a recruitment gesture; see also Cords (1988) for similar behaviour in young longtailed macaques). It has been speculated that such solicitation may be aimed at using partnership in a coalition to achieve some form of ‘reconciliation’ with the former adversary (de Waal & Yoshihara, 1983; Aureli & van Schaik, 1991b), as well as more directly reducing the likelihood of renewed aggression from that quarter. Again, aggression against the target

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

14:27

Redirection of aggression: multiparty signalling appears an almost incidental by-product of communication with one’s former opponent. Redirection is also a particularly noisy affair. These interactions disproportionately often incorporate vocal threats in comparison with equivalent bouts initiated during control observations (Kazem, 1999). Increased incidence of vocal forms of aggression might simply reflect greater arousal in animals that have themselves recently been subjected to attacks, but whatever the proximate mechanism the resulting events will be more effective at alerting bystanders. This is reminiscent of Zahavi’s (1979) observation that many aggressive signals are far louder than actually required for effective information transfer between the two parties directly concerned. He interpreted the probable interception of these signals by several more distant receivers as imposing greater costs upon the signaller, therefore ensuring the reliability of the degree of threat conveyed to the opponent. We suggest that one ‘shouts’ precisely in order to advertise the signal to those more distant receivers, because one gains a benefit by doing so. Furthermore, evidence for individual discrimination by voice exists in macaques, although the extent of individual signatures varies according to call type (rhesus: Gouzoules et al., 1986; Rendall et al., 1996, 1998). Perception of identity in threat calls does not appear to have been tested but, if present, would allow the signaller’sactions to be identified even if the redirection event was not observed. Kin-oriented redirection: a special case?

An intriguing variant is that victims appear specifically to target maternal relatives of their former assailant. Examples have been reported in some cercopithecines: vervet monkeys (Cheney & Seyfarth (1989), although they did not specify whether the actor was victim or aggressor in the original conflict), juvenile longtailed macaques (Aureli & van Schaik, 1991a), Japanese macaques (Aureli et al., 1992) and pigtail macaques Macaca nemestrina ( Judge, 1982). As maternal relatives share the (typically higher) familial rank of the original aggressor, the strategy is not without its risks. Japanese macaques circumvent this issue by selecting younger – and therefore often lower-ranking and more vulnerable – relatives of their former opponent and take advantage of ‘safe’ opportunities to join ongoing coalitions against the target, making it difficult for the target (or the former aggressor) to retaliate (Aureli et al., 1992). Unsurprisingly, kin-oriented threats and attacks typically account for only a small fraction of redirection events and may take place over a longer timescale (minutes or even hours) because of the need to encounter appropriate conditions. In Japanese macaques the majority (74%) of these incidents occurred within view of the former aggressor, leading Aureli et al. (1992) to propose that inflicting indirect fitness costs might serve as a form of social leverage, to deter further aggression from the same individual over the longer

203

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

204

0 521 582361 7

April 11, 2005

14:27

A. J. N. Kazem & F. Aureli term (assuming the latter associates the two events). The suggestion, therefore, relies on a signalling argument, in this case restricted to a particular third-party receiver. It has recently been demonstrated that apparently ‘spiteful’ acts (such as these) can be evolutionarily stable in systems where observers accord each other status on the basis of aggression witnessed, at least under certain simplified social conditions ( Johnstone & Bshary, 2004). Kin-oriented redirection need not be restricted to primates. Appropriate conditions are provided in any system where individuals are often constrained from retaliating directly against aggressors, reside in groups composed of a mixture of related and unrelated conspecifics, and close kin are preferred associates or coalition partners. Macaques and baboons can discriminate kin relationships between third parties (e.g. Dasser, 1988; Cheney & Seyfarth, 1999), possibly by observing the association patterns of other group members, and are often assumed to act on this basis. However, use of simpler proximity-based rules may suffice. If relatives of protagonists tend to cluster at the scene of conflicts (as is often the case in taxa where individuals preferentially support their kin in coalitions), they will be overrepresented among the bystanders present. Therefore, a tendency to strike at any vulnerable individual nearby (other than one’sown close associates) could have the effect of disproportionately targeting the opponent’s kin under post-conflict conditions. Spotted hyaenas and greylag geese Anser anser, both of which are reported to redirect aggression (Table 10.2), might be promising candidates. Spotted hyaenas form large multimale–multifemale clans with high variance in within-group relatedness and a matrilineal structure similar to macaques (Frank, 1986; Mills, 1990). Many geese form cohesive family units (pairs and their immature offspring) that aggregate in large feeding flocks during winter. Both are highly competitive societies with pronounced dominance relationships and a high propensity to intervene aggressively on behalf of relatives in disputes; consequently, rank is highly dependent on social support (hyaenas: Zabel et al., 1992; Engh et al., 2000; geese: Lamprecht, 1986; Black & Owen, 1987; K. Kotrschal, personal communication). Kin are valuable allies and often in spatial proximity, affording competitors the opportunity to learn (or otherwise locate) the habitual associates of others, as well as ‘safe’ opportunities to redirect in coalition.

Intraspecific redirection in other taxa A number of predictions can be made regarding systems in which redirection might operate as a signal. First and foremost, conditions should facilitate eavesdropping: the communication modality, social structure and typical habitat should be such that agonistic signals often transmit further than the average spacing between conspecifics and can, therefore, be detected by several receivers. The

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

14:27

Redirection of aggression: multiparty signalling extent of the communication network will often constrain whether bystanders are likely to be aware of a contestant’s defeat, providing both the impetus to redirect and the means by which its occurrence is detected. Where individuals reside in permanent and cohesive social groups with defined dominance relationships between group members (e.g. many primates, social carnivores, some ungulates), the majority of interactions witnessed are likely to involve animals whose capabilities are already known to the individual and aggression in the expected direction within a dyad (the exceptions being contests involving immigrants and rank reversals occurring as individuals mature or decline in ability). As we have argued for macaques, much of the utility of eavesdropping may then lie in detecting transient loser effects (determining the optimum timing of challenges), and redirection may primarily convey updated information on a victim’s post-conflict state (as well as reaffirming the ‘status quo’ to both target and bystanders). If so, redirection should occur in conjunction with a physiological/behavioural loser effect (at least in the wake of intense contests) in systems where victims often receive several attacks from different individuals in quick succession (coyotes Canis latrans may be an example; Table 10.2). However, the redirection principle may be more broadly applicable. In many taxa, individuals have knowledge of the abilities of a consistent subset of conspecifics yet frequently encounter others whose status is not yet known. Examples include species where kin units coalesce into larger but spatially structured aggregations (e.g. winter feeding flocks in geese and some corvids), or individuals form dominance relationships within a temporary display aggregation (e.g. male ducks). Other possibilities include species where individuals defend nest sites within a breeding colony or form closely spaced territories during the breeding season (as in many birds and fish), thus forming relationships with their immediate neighbours. Where eavesdropping is used mainly to estimate the fighting ability of unfamiliar competitors, redirecting could limit the negative impression conveyed by a loss. Observers should be wary of challenging those losers that have nevertheless demonstrated the motivation and ability still to dominate some opponents ( just as eavesdroppers in some species respond more cautiously toward losers that have exhibited persistent counter-aggression during their previous contest (Earley & Dugatkin, 2002)). In this case, loser effects need not be present for redirection to be worthwhile (although where they are, the incentive for redirecting may be further enhanced). There are also many systems where a signalling aspect to redirection seems unlikely. For example, both eavesdropping and redirection would seem to be of less use when individuals reside in large and continually shifting aggregations (either year-round or during certain seasons), in which they possess little information on conspecifics’ identities. In such cases, ‘badges of status’ may be used to mediate

205

206 resident–intruder tests

Zabel et al., 1992; M. L. East & communicationg

Yeso

H. Hofer, personal

Low

captive cohorts)

E

communicationg

S. Pellis, personal

often females (wild clans and

Adults and juveniles of both sexes,

(captive)n

colonies (triads), in staged

Dominant males in mixed-sex

B

P. J. Drent, personal

April 11, 2005

crocuta

Spotted hyaena, Crocuta

norvegicus

Yesl

personal communicationg

E. J. A. Cunningham,

unpublished datah

K. Kotrschal et al.,

1983g Yesm

Yesl

Yes

Yes

communicationg

C. Mac´ıas Garcia, personal

communicationg

M. R. Gross, personal

communicationg

H. C. Suter, personal

Øverli et al., 2004f

Main sourcee

communication; Drent,

Variable 2

Low

Low

Variable 1

Yes?k

Yes

–

Yesi

Variable 1

challengedd

effectc

aggressionb

Victims

Loser

Counter-

foraging flocks in wild)

D

C

C

A

A

B

B

Systema

flocks (aviary groups and winter

Adults, typically males, in small

pens)

aggregations (wild and outdoor

Drakes within pre-pairing display

(free ranging)

Ganders in winter feeding flocks

groups of seven (captive)

Males in mixed-sex experimental

colony (wild)

Nesting males within breeding

stream) j

sextets in flume (seminatural

Parr held in mixed-sex quintets and

experimental contests (captive)

in successive pairwise

Juvenile triads (unsexed), meeting

redirection observed

Context in which

0 521 582361 7

Laboratory rat, Rattus

Great tit, Parus major

Mallard, Anas platyrhynchos

Greylag goose, Anser anser

multiradiatus

Amarillo fish, Girardinichthys

macrochirus

Bluegill sunfish, Lepomis

Atlantic salmon, Salmo salar

mykiss

Rainbow trout, Oncorhynchus

Species

Table 10.2. Reports of intraspecific redirection of aggression in other species

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor 14:27

E

E −

Low

Yesp Yes

Marcus-Newhall et al., 2000f

communicationg

M. Bekoff, personal

system typical of this context in the wild: A–D, individuals familiar with only a subset of conspecifics regularly encountered; A, defence of nest or

(experimental)

verbal provocation

Both sexes, adults subjected to

sextets (wild and captive litters)

Both sexes, cubs within sibling

between familiar individuals occurs in ≤ 10% of dyadic contests; variable 1, typically low between familiar individuals, but higher in this instance,

207

on ancedotal reports from authors collecting quantitative data on aggressive interactions;

& Dugatkin, 2000.

occasionally received from bystanders, but Wahaj et al., 2001 demonstrate receipt from former opponent.

April 11, 2005

p Bekoff

o Only

known whether the subject was loser or winner of the initial contest.

1946; van der Poll et al., 1982.

n Not

1998; Verbeek et al., 1999.

m Seward,

k Behavioural

0 521 582361 7

l Verbeek,

groups held at relatively high density; redirection not observed in wild. loser effect has been documented in Lepomis gibbosus (Beacham & Newman, 1987; Chase et al., 1994) and L. cyanellus (McDonald et al., 1968).

j Seminatural

et al., 1985.

inferred from quantitative but uncontrolled data.

i Abbott

h or

for redirection demonstrated experimentally.

comparison with control data;

g based

f in

demonstrated using successive pairwise contests, or inferred from sequences of interaction in intact groups.

that victims receive subsequent harassment from bystanders in the minutes following an initial defeat.

e Evidence

d Reports

c Experimentally

likely frequency of counter-aggression decreases from low to variable 2 to variable 1 to experimentally prevented.

e.g. because contestants unfamiliar at start of protocol; variable 2, typically low between familiar individuals; − , prevented by experimental protocol. The

b Low,

between all group members.

non-territorial flock or territorial pair, dominance relationships between subsets of conspecifics; E, membership of cohesive group, dominance relationships

defence of shifting feeding zone around individual or family, dominance relationships between subsets of individuals of one or both sexes; D, membership of

courting site in seasonal aggregation by one sex; B, defence of territory or feeding site, dominance relationships between individuals resident in local area; C,

a Social

Human, Homo sapiens

Coyote, Canis latrans

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor 14:27

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

208

0 521 582361 7

April 11, 2005

14:27

A. J. N. Kazem & F. Aureli access to resources such as food, where individual items are of relatively low value (e.g. certain finches: Rohwer, 1975; Johnstone & Norris, 1993). Displaying a badge means an individual’s relative aggressive propensity and/or ability should be immediately apparent to those it meets and to others who witness its social interactions, reducing the utility of eavesdropping. However, if dominance is not mediated solely via plumage badges (e.g. individuals may still form relationships with a subset of familiar conspecifics), and as the magnitude of any loser effect may still vary, a signalling role for redirection is not ruled out. Finally, a low probability of counter-aggression during contests is thought to promote redirection (e.g. Thierry, 1985), because this factor influences both the necessity and the costs of the behaviour. As mentioned above, redirection offers an outlet for arousal and frustration in situations where animals cannot retaliate aggressively against an initiator, either because they are physically prevented from doing so (in experiments) or because the adversary is perceived as too powerful. A signalling interpretation can also predict a negative association between the two forms of response; they might be alternative methods of ‘proving oneself ’ to pre-empt strikes from conspecifics that had witnessed the defeat (although individuals might sometimes do both to reinforce the message; cf. a multicomponent signal). Second, if there is a high risk of reprisal from the target, redirection may become an excessively costly option. Both arguments have been used to explain the absence of operationally defined redirection in adult female olive baboons and mountain gorillas (Watts, 1995b; Castles & Whiten, 1998b); bidirectional aggression is common between adult females in these species. Potential retaliation from targets might also explain why the majority of ‘redirection’ in a small captive group of chimpanzees Pan troglodytes was directed toward human caretakers, rather than conspecifics (Malone et al., 2000). Suitable targets are expected to be those likely to capitulate immediately. Selection may be based on familiarity: targeting a known subordinate makes sense, especially in species with a ‘strict’ dominance style. However, certain classes of unfamiliar conspecific can also be identified as unlikely to dispute the outcome, for example targeting first-year individuals in birds (identifiable by their size and plumage) or smaller floater and satellite males in fish. Reports of intraspecific redirection

When placed in situations where they are unable to retaliate directly to verbal provocation, humans will readily redirect hostility by verbally or physically punishing substitute targets, both animate and inanimate (reviewed by MarcusNewhall et al., 2000). Experimentally induced ‘displaced aggression’ has also been observed in rodents. Although females are not ordinarily subjected to aggression by males, mates may become the target of ‘redirected’ attacks in the seconds

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

14:27

Redirection of aggression: multiparty signalling following staged contests between resident males and same-sex intruders (mice Peromyscus spp.: Eisenberg, 1962; Simmel & Walker, 1970; montane and prairie voles Microtus montanus and Microtus ochrogaster: S. Pellis, personal communication), especially if continued access to the opponent(s) is suddenly barred. Unfortunately, in rodent studies it is often unclear whether the redirecting individual was, or would eventually have become, the loser in the original contest. A rather artificial situation is also imposed by the research foci of these experiments; the subject may not have the option of retaliating against the original aggressor because s/he is no longer present (often the case in human studies) or may be physically prevented from pursuing the original contest to its conclusion (e.g. because of partitions separating opponents in work on rodents). Consequently the subject is generally presented with a limited range of options and stimuli against which to direct any response. However, apparent redirection of aggression by victims has been reported under more naturalistic social contexts in a wide range of species (a selection, by no means a comprehensive survey, is presented in Table 10.2). In some species, the behaviour may be performed purely for the physiological effects: the benefits of rapidly reducing arousal and perhaps of experiencing victory. However, the occurrence of redirection is likely to be used as a cue by bystanders, because it inevitably carries information about an unknown competitor’s relative ability within a population and, where winner or loser effects are in operation, its current state. If this means bystander behaviour is influenced to the redirecting individual’s advantage, one would expect redirection to have been selected as a signal that is performed more than required for strictly physiological reasons (or even in the absence of any physiological benefits) and adjusted so as to be effectively publicized. The challenge is to identify where particular taxa fall within this spectrum of possibilities. Testing the occurrence and function of redirection

An essential preliminary is verifying whether an identifiable phenomenon exists – i.e. a pronounced increase in initiation of aggression after being victimized in comparison with the incidence under appropriate control conditions. As the primate data illustrate (Table 10.1), redirection may fail to be demonstrated operationally in species (or age–sex classes) where, on the basis of uncontrolled data, it had been assumed to be present. It would also be useful to clarify whether counter-aggression and redirection are dissociated, and what factors favour expression of one tactic over the other. In observational work to date, there is often insufficient data per individual to allow analyses that compare the incidence of redirection after unidirectional versus bidirectional contests, whilst also controlling for systematic differences in aggression intensity, duration and number of participants in the contest. Crucially, the existence of a causal relationship

209

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

210

0 521 582361 7

April 11, 2005

14:27

A. J. N. Kazem & F. Aureli between redirecting and a reduction in the number of challenges subsequently received by the actor requires experimental verification. This could be addressed by experimentally manipulating the access of the former aggressor and other bystanders to visual/acoustic information concerning the initial conflict and its aftermath – how do counter-aggression and redirection compare in terms of their influence upon the behaviour of witnesses? Furthermore, demonstrating that victims are sensitive to the perceived presence and composition of bystanders when redirecting (and not just the availability of a suitable target) would clearly support a signalling interpretation. Given the likely costs of engaging in aggression, such facultative use of redirection is predicted. Although absence of an ‘audience effect’ (i.e. a change in behaviour when in the perceived presence of conspecifics; see Ch. 4) does not preclude a signalling function, one would expect that use of an unconditional rule could evolve only if, under natural circumstances, a relevant audience is almost invariably present. Finally, in taxa that exhibit a loser effect, it should be possible to confirm whether conflicts concluded after differing durations and intensity of aggression produce different physiological changes in the loser. This would provide a basis for exploring how neuroendocrine profiles in the immediate post-conflict period, and the time course of recovery, might differ in cases where the animal did or did not redirect.

Summary We propose that redirection of aggression may function, in part, as a signal, used to pre-empt subsequent challenges from the former aggressor and/or other bystanders that witness the act. In systems where behavioural loser effects exist, such as macaques, this might be achieved because redirection serves as an honest indicator of post-conflict condition and motivation, demonstrating that the perpetrator has not been unduly compromised by its preceding defeat. Alternatively, redirection might simply limit the extent of the negative impression usually conveyed by a loss, reaffirming one’s position within a status hierarchy to bystanders or, where the bystander is unfamiliar with the protagonist, conveying an ability to defeat at least some individuals within the population. This suggestion has a number of broader implications. Being observed to lose a contest incurs social penalties, affecting not only the likelihood of being challenged by others but also the prospects of acquiring or retaining mates (e.g. Doutrelant & McGregor, 2000; Mennill et al., 2002). So far, attention has focused exclusively on how contestants counter these pressures by modifying their displays during the original agonistic exchange if in the perceived presence of an audience (e.g. Doutrelant et al., 2001; Matos & McGregor, 2002; A. J. N. Kazem, R. J. Motos & P. K. McGregor, unpublished data). In reality, there may be several

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

14:27

Redirection of aggression: multiparty signalling points at which the eventual loser can influence others’ perceptions of its ability: in the decision to counter-attack rather than defer, to fight harder (or differently) in the current exchange, or even by redirecting after the contest has concluded. First, this suggests it would be profitable to consider audience effects and eavesdropping over a longer timescale, encompassing sequences of contests. Second, individuals who approach conflicts between others risk becoming targets of redirection (this may explain why, in our study species, low-ranking individuals often flee the scene of incipient or ongoing aggression). Therefore, depending on the signal modality used, the costs of eavesdropping may not be as low as previously assumed – particularly for low-ranking or otherwise poor competitors, which stand to gain the most from acquiring information via this route. Finally, redirection may constitute a rare example of a signal aimed primarily at ‘secondary’ receivers. The communication networks perspective has highlighted how many dyadic displays evolve within the context of, and may even be designed to advertise to, several receivers. However, redirection may be an instance of a signal performed almost entirely for its effect on third parties rather than upon the apparent receiver. There has been substantial theoretical interest in the possibility that apparently altruistic behaviour may be maintained via this route (e.g. Zahavi, 1977; Nowak & Sigmund, 1998; Roberts, 1998; Leimar & Hammerstein, 2001), but few empirical examples have been discovered outside humans. Redirection has the potential to prove a taxonomically widely distributed case, but experimental verification and analyses to explore the evolutionary stability of such a signalling system in aggression are still required. Acknowledgements We are very grateful to the many correspondents who provided access to unpublished data and detailed observations on their study taxa. We would also like to thank Peter McGregor for inviting us to contribute to this volume, and his encouragement and support of the first author’s ongoing work on redirection. The latter’s research on rhesus macaques was financially supported by awards from the H. F. Guggenheim Foundation, L. S. B. Leakey Foundation, Wenner Gren Foundation for Anthropological Research, Zunz Foundation and a University of Durham Research Studentship; the generosity of these organizations is greatly appreciated.

References Abbott, J. C., Dunbrack, R. L. & Orr, C. D. 1985. The interaction of size and experience in dominance relationships of juvenile steelhead trout (Salmo gairdneri). Behaviour, 92, 241–253. Andrews, R. J. 1991. Testosterone, attention and memory. In: The Development and Integration of Behaviour, ed. P. P. G. Bateson. Cambridge, UK: Cambridge University Press, pp. 171–190.

211

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

212

0 521 582361 7

April 11, 2005

14:27

A. J. N. Kazem & F. Aureli Arnold, K. & Barton, R. A. 2001. Postconflict behavior of spectacled leaf monkeys (Trachypithecus obscurus). II. Contact with third parties. International Journal of Primatology, 22, 267–286. Aureli, F. 1992. Post-conflict behaviour among wild long-tailed macaques (Macaca fascicularis). Behavioral Ecology and Sociobiology, 31, 329–337. Aureli, F. & van Schaik, C. P. 1991a. Post-conflict behaviour in long-tailed macaques (Macaca fascicularis): I. The social events. Ethology, 89, 89–100. 1991b. Post-conflict behaviour in long-tailed macaques (Macaca fascicularis): II. Coping with the uncertainty. Ethology, 89, 101–114. Aureli, F., Cozzolino, R., Cordischi, C. & Scucchi, S. 1992. Kin-oriented redirection among Japanese macaques: an expression of a revenge system? Animal Behaviour, 44, 283–291. Aureli, F., Veenema, H. C., van Panthaleon van Eck, C. J. & van Hooff, J. A. R. A. M. 1993. Reconciliation, consolation, and redirection in Japanese macaques (Macaca fuscata). Behaviour, 124, 1–21. Aureli, F., Das, M., Verleur, D. & van Hooff, J. A. R. A. M. 1994. Postconflict social interactions among Barbary macaques (Macaca sylvanus). International Journal of Primatology, 15, 471–484. Aureli, F., Cords, M. & van Schaik, C. P. 2002. Conflict resolution following aggression in gregarious animals: a predictive framework. Animal Behaviour, 64, 325–343. Bastock, M., Morris, D. & Moynihan, M. 1953. Some comments on conflict and thwarting in animals. Behaviour, 6, 66–84. Beacham, J. L. & Newman, J. A. 1987. Social experience and the formation of dominance relationships in the pumpkinseed sunfish, Lepomis gibbosus. Animal Behaviour, 35, 1560–1563. Bekoff, M. & Dugatkin, L. A. 2000. Winner and loser effects and the development of dominance relationships in young coyotes: an integration of data and theory. Evolutionary Ecology Research, 2, 871–883. Bernstein, I. S., Rose, R. M. & Gordon, T. P. 1974. Behavioral and environmental events influencing primate testosterone levels. Journal of Human Evolution, 3, 517–525. Black, J. M. & Owen, M. 1987. Determinants of social rank in goose flocks: acquisition of social rank in young geese. Behaviour, 102, 129–146. Blanchard, D. C., Sakai, R. R., McEwen, B. & Blanchard, R. J. 1993. Subordination stress: behavioral, brain and neuroendocrine correlates. Behavioural Brain Research, 58, 113–121. Booth, A., Shelley, G., Mazur, A., Tharp, G. & Kittok, R. 1989. Testosterone, and winning and losing in human competition. Hormones and Behavior, 23, 556–571. Castles, D. L. & Whiten, A. 1998a. Post-conflict behaviour of wild olive baboons. II. Stress and self-directed behaviour. Ethology, 104, 148–160. 1998b. Post-conflict behaviour of wild olive baboons. I. Reconciliation, redirection and consolation. Ethology, 104, 126–147. Chapais, B. 1995. Alliances as a means of competition in primates: evolutionary, developmental and cognitive aspects. Yearbook of Physical Anthropology, 38, 115– 136.

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

14:27

Redirection of aggression: multiparty signalling Chase, I. D., Bartolomeo, C. & Dugatkin, L. 1994. Aggressive interactions and inter-contest interval: how long do winners keep winning? Animal Behaviour, 48, 393–400. Chase, I. D., Tovey, C., Spangler-Martin, D. & Manfredonia, M. 2002. Individual differences versus social dynamics in the formation of animal dominance hierarchies. Proceedings of the National Academy of Sciences, USA, 99, 5744–5749. Cheney, D. L. & Seyfarth, R. M. 1986. The recognition of social alliances by vervet monkeys. Animal Behaviour, 34, 1722–1731. 1989. Redirected aggression and reconciliation among vervet monkeys, Cercopithecus aethiops. Behaviour, 110, 258–275. 1990. Attending to behaviour versus attending to knowledge: examining monkeys’ attribution of mental states. Animal Behaviour, 40, 742–753. 1999. Recognition of other individuals’ social relationships by female baboons. Animal Behaviour, 58, 67–75. Cheney, D. L., Seyfarth, R. M. & Silk, J. B. 1995. The responses of female baboons (Papio cynocephalus ursinus) to anomalous social interactions: evidence for causal reasoning? Journal of Comparative Psychology, 109, 134–141. Cords, M. 1988. Resolution of aggressive conflicts by immature long-tailed macaques, Macaca fascicularis. Animal Behaviour, 36, 1124–1135. 1992. Post-conflict reunions and reconciliation in long-tailed macaques. Animal Behaviour, 44, 57–61. Cynx, J. & Nottebohm, F. 1992. Testosterone facilitates some conspecific song discriminations in castrated zebra finches Taeniopygia guttata. Proceedings of the National Academy of Sciences, USA, 89, 1376–1378. Das, M. 1998. Conflict management and social stress in long-tailed macaques. Ph.D. Thesis, University of Utrecht, the Netherlands. Das, M., Penke, Z. & van Hooff, J. A. R. A. M. 1997. Affiliation between aggressors and third parties following conflicts in long-tailed macaques (Macaca fascicularis). International Journal of Primatology, 18, 157–179. Dasser, V. 1988. A social concept in Java monkeys. Animal Behaviour, 36, 225–230. Delville, Y., Mansour, K. M., Ferris, C. F. 1996. Testosterone facilitates aggression by modulating vasopressin receptors in the hypothalamus. Physiology and Behavior, 60, 25–29. de Waal, F. B. M. & Luttrell, L. M. 1989. Toward a comparative socioecology of the genus Macaca: different dominance styles in rhesus and stumptail macaques. American Journal of Primatology, 19, 83–109. de Waal, F. B. M. & Yoshihara, D. 1983. Reconciliation and redirected affection in rhesus monkeys. Behaviour, 85, 224–241. Doutrelant, C. & McGregor, P. K. 2000. Eavesdropping and mate choice in female fighting fish. Behaviour, 137, 1655–1669. Doutrelant, C., McGregor, P. K. & Oliveira, R. F. 2001. The effect of an audience on intra-sexual communication in male Siamese fighting fish, Betta splendens. Behavioral Ecology, 12, 283–286. Drent, P. J. 1983. The functional ethology of territoriality in the great tit (Parus major L.). Ph.D. Thesis, University of Groningen, the Netherlands.

213

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

214

0 521 582361 7

April 11, 2005

14:27

A. J. N. Kazem & F. Aureli Earley, R. L. & Dugatkin, L. A. 2002. Eavesdropping on visual cues in green swordtail (Xiphophorus helleri) fights: a case for networking. Proceedings of the Royal Society of London, Series B, 269, 943–952. Eaton, G. G. 1984. Aggression in adult male primates: a comparison of confined Japanese macaques and free-ranging olive baboons. International Journal of Primatology, 5, 145–160. Eisenberg, J. F. 1962. Studies on the behavior of Peromyscus maniculatus gambelli and Peromyscus californicus parasiticus. Behaviour, 19, 177–207. Emery, N. J., Lorincz, E. N., Perret, D. I., Oram, M. W. & Backer, C. I. 1997. Gaze following and joint attention in rhesus monkeys (Macaca mulatta). Journal of Comparative Psychology, 111, 286–293. Engh, A. L., Esch, K., Smale, L. & Holekamp, K. E. 2000. Mechanisms of maternal rank ‘inheritance’ in the spotted hyaena, Crocuta crocuta. Animal Behaviour, 60, 323–332. Fairbanks, L. A., Melega, W. P., Jorgensen, M. J., Kaplan, J. R. & McGuire, M. T. 2001. Social impulsivity inversely associated with CSF 5-HIAA and fluoxetine exposure in vervet monkeys. Neuropsychopharmacology, 24, 370–378. Ferris, C. F., Melloni, R. H., Jr, Koppel, G. et al. 1997. Vasopressin/serotonin interactions in the anterior hypothalamus control aggressive behavior in golden hamsters. Journal of Neuroscience, 17, 4331–4340. Frank, L. G. 1986. Social organization of the spotted hyena (Crocuta crocuta). 2. Dominance and reproduction. Animal Behaviour, 34, 1510–1527. Fuentes, A., Malone, N., Sanz, C., Matheson, M. & Vaughan, L. 2002. Conflict and post-conflict behavior in a small group of chimpanzees. Primates, 43, 223–235. Gore, M. A. 1994. Dyadic and triadic aggression and assertiveness in adult female rhesus monkeys, Macaca mulatta, and hamadryas baboons, Papio hamadryas. Animal Behaviour, 48, 385–392. Gouzoules, H., Gouzoules, S. & Marler, P. 1986. Vocal communication: a vehicle for the study of social relationships. In: The Cayo Santiago Macaques: History, Behavior and Biology, ed. R. G. Rawlins & M. J. Kessler. Albany, NY: State University of New York Press, pp. 111–129. Gust, D. A. & Gordon, T. P. 1993. Conflict resolution in sooty mangabeys. Animal Behaviour, 46, 685–694. Gust, D. A., Gordon, T. P., Hambright, M. K. & Wilson, M. E. 1993. Relationship between social factors and pituitary-adrenocortical activity in female rhesus monkeys (Macaca mulatta). Hormones and Behavior, 27, 318–331. Hannes, R. P., Franck, D. & Liemann, F. 1984. Effects of rank order fights on body and blood concentrations of androgens and corticosteroids in the male swordtail (Xiphophorus helleri). Zeitschrift f¨ ur Tierpsychologie, 65, 53–65. Higley, J. D., Mehlman, P. T., Poland, R. E. et al. 1996. CSF testosterone and 5-HIAA correlate with different types of aggressive behaviors. Biological Psychiatry, 40, 1067–1082. Hogue, M.-E., Beaugrand, J. P. & Lague, P. C. 1996. Coherent use of information by hens observing their former dominant defeating or being defeated by a stranger. Behavioural Processes, 38, 241–252.

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

14:27

Redirection of aggression: multiparty signalling Hsu, Y. & Wolf, L. L. 1999. The winner and loser effect: integrating multiple experiences. Animal Behaviour, 57, 903–910. 2001. The winner and loser effect: what fighting behaviours are influenced? Animal Behaviour, 61, 777–786. Itani, J. 1963. Vocal communication of the wild Japanese monkeys. Primates, 4, 11–66. Johnstone, R. A. & Bshary, R. 2004. Evolution of spite through indirect reciprocity. Proceedings of the Royal Society of London, Series B, 271, 1917–1922. Johnstone, R. A. & Norris, K. 1993. Badges of status and the cost of aggression. Behavioral Ecology and Sociobiology, 32, 127–134. Judge, P. G. 1982. Redirection of aggression based on kinship in a captive group of pigtail macaques. International Journal of Primatology, 3, 301. Kavaliers, M., Choleris, E. & Colwell, D. D. 2001. Brief exposure to female odors ‘emboldens’ male mice by reducing predator-induced behavioral and hormonal responses. Hormones and Behavior, 40, 497–509. Kazem, A. J. N. 1999. Conflict management in free-ranging immature rhesus macaques, Macaca mulatta. Ph.D. Thesis, University of Durham, UK. Keddy Hector, A., Seyfarth, R. M. & Raleigh, M. J. 1989. Male parental care, female choice, and the effect of an audience in vervet monkeys. Animal Behaviour, 38, 262–271. Koolhaas, J. M., Korte, S. M., de Boer, S. F. et al. 1999. Coping styles in animals: current status in behavior and stress-physiology. Neuroscience and Biobehavioral Reviews, 23, 925–935. Kutsukake, N. & Castles, D. L. 2001. Reconciliation and variation in post-conflict stress in Japanese macaques (Macaca fuscata fuscata): testing the integrated hypothesis. Animal Cognition, 4, 259–268. Lamprecht, J. 1986. Structure and causation of the dominance hierarchy in a flock of bar-headed geese (Anser indicus). Behaviour, 96, 28–48. Leimar, O. & Hammerstein, P. 2001. Evolution of cooperation through indirect reciprocity. Proceedings of the Royal Society of London, Series B, 268, 745–753. Levine, S., Coe, C. & Wiener, S. G. 1989. Psychoneuroendocrinology of stress: a psychobiological perspective. In: Psychoendocrinology, ed. S. Levine & R. Brush. London: Academic Press, pp. 341–377. Malone, N., Vaughan, L. & Fuentes, A. 2000. The role of human caregivers in the post-conflict interactions of captive chimpanzees (Pan troglodytes). Laboratory Primate Newsletter, 39, 1–3. Marcus-Newhall, A., Pedersen, W. C., Carlson, M. & Miller, N. 2000. Displaced aggression is alive and well: a meta-analytic review. Journal of Personality and Social Psychology, 78, 670–689. Matos, R. J. & McGregor, P. K. 2002. The effect of the sex of an audience on male–male displays of Siamese fighting fish (Betta splendens). Behaviour, 139, 1211–1221. McDonald, A. L., Heimstra, N. W. & Damkot, D. K. 1968. Social modification of agonistic behaviour in fish. Animal Behaviour, 16, 437–441. McGregor, P. K. 1993. Signalling in territorial systems: a context for individual identification, ranging and eavesdropping. Philosophical Transactions of the Royal Society of London, Series B, 340, 237–244.

215

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

216

0 521 582361 7

April 11, 2005

14:27

A. J. N. Kazem & F. Aureli McGregor, P. K., Peake, T. M. & Lampe, H. M. 2001. Fighting fish, Betta splendens, extract relative information from apparent interactions: what happens when what you see is not what you get. Animal Behaviour, 62, 1059–1065. Mehlman, P. T., Higley, J. D., Faucher, I. et al. 1994. Low CSF 5-HIAA concentrations and severe aggression and impaired impulse control in nonhuman primates. American Journal of Psychiatry, 151, 1485–1491. Mendoza, S. P. 1993. Social conflict on first encounters. In: Primate Social Conflict, ed. W. A. Mason & S. P. Mendoza. Albany, NY: State University of New York Press, pp. 85–110. Mennill, D. J., Ratcliffe, L. M. & Boag, P. T. 2002. Female eavesdropping on male song contests in songbirds. Science, 296, 873. Mills, M. G. L. 1990. Kalahari hyaenas: Comparative Behavioural Ecology of Two Species. London: Unwin Hyman. Moynihan, M. H. 1998. The social regulation of competition and aggression in animals. Washington, DC: Smithsonian Institution Press. Nowak, M. A. & Sigmund, K. 1998. Evolution of indirect reciprocity by image scoring. Nature, 393, 573–577. Oliveira, R. F., McGregor, P. K. & Latruffe, C. 1998. Know thine enemy: fighting fish gather information from observing conspecific interactions. Proceedings of the Royal Society of London, Series B, 265, 1045–1049. Olivier, B., Mos, J., van Oorschot, R. & Hen, R. 1995. Serotonin receptors and animal models of aggressive behavior. Pharmacopsychiatry, 28, 80–90. Øverli, Ø., Harris, C. A. & Winberg, S. 1999. Short-term effects of fights for social dominance and the establishment of dominant-subordinate relationship on brain monoamines and cortisol in rainbow trout. Brain Behavior and Evolution, 54, 263–275. Øverli, Ø., Korzan, W. J., Larson, E. T. et al. 2004. Behavioural and neuroendocrine correlates of displaced aggression in trout: cycle of violence is evolutionarily conserved. Hormones and Behavior, 45, 3224–3349. Peake, T. M., Terry, A. M. R., McGregor, P. K. & Dabelsteen, T. 2001. Male great tits eavesdrop on simulated male-to-male vocal interactions. Proceedings of the Royal Society of London, Series B, 268, 1183–1187. Perreault, H. A. N., Semsar, K. & Godwin, J. 2003. Fluoxetine treatment decreases territorial aggression in a coral reef fish. Physiology and Behavior, 79, 719–724. Rendall, D., Rodman, P. S. & Emond, R. E. 1996. Vocal recognition of individuals and kin in free-ranging rhesus monkeys. Animal Behaviour, 51, 1007–1015. Rendall, D., Owren, M. J. & Rodman, P. S. 1998. The role of vocal tract filtering in identity cueing in rhesus monkey (Macaca mulatta) vocalizations. Journal of the Acoustical Society of America, 103, 602–614. Roberts, G. 1998. Competitive altruism: from reciprocity to the handicap principle. Proceedings of the Royal Society of London, Series B, 265, 427–431. Roeder, J.-J., Fornasieri, I. & Gosset, D. 2002. Conflict and postconflict behaviour in two lemur species with different social organizations (Eulemur fulvus and Eulemur macaco): a study on captive groups. Aggressive Behavior, 28, 62–74.

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

14:27

Redirection of aggression: multiparty signalling Rohwer, S. 1975. The social significance of avian plumage variability. Evolution, 29, 593–610. Rose, R. M., Bernstein, I. S. & Gordon, T. 1975. Consequences of social conflict on plasma testosterone levels in rhesus monkeys. Psychosomatic Medicine, 37, 50–61. Sapolsky, R. M. 1998. Why Zebras Don’t get Ulcers: An Updated Guide to Stress, Stress-related Diseases, and Coping, 2nd edn. New York: Freeman. Sapolsky, R. M. & Ray, J. C. 1989. Styles of dominance and their endocrine correlates among wild olive baboons (Papio anubis). American Journal of Primatology, 18, 1–13. Schuurman, T. 1980. Hormonal correlates of agonistic behaviour in adult male rats. Progress in Brain Research, 53, 415–520. Scott, J. P. & Fredericson, E. 1951. The causes of fighting in mice and rats. Physiological Zoology, 24, 273–309. Scucchi, S., Cordischi, C., Aureli, F. & Cozzolino, R. 1988. The use of redirection in a captive group of Japanese monkeys. Primates, 29, 229–236. Seward, J. P. 1946. Aggressive behavior in the rat. IV. Submission as determined by conditioning, extinction and disuse. Journal of Comparative Psychology, 39, 51–57. Silk, J. B. 1999. Male bonnet macaques use information about third-party rank relationships to recruit allies. Animal Behaviour, 58, 45–51. Simmel, E. C. & Walker, D. A. 1970. Social priming for agonistic behavior in a ‘docile’ mouse strain. American Zoologist, 10, 486–487. Summers, C. H., Summers, T. R., Moore, M. C. et al. 2003. Temporal patterns of limbic monoamine and plasma corticosterone response during social stress. Neuroscience, 116, 553–563. Thierry, B. 1985. Patterns of agonistic interactions in three species of macaque (Macaca mulatta, M. fascicularis, M. tonkeana). Aggressive Behavior, 11, 223–233. 2000. Covariation of conflict management patterns in macaque societies. In: Natural Conflict Resolution, ed. F. Aureli & F. B. M. de Waal. Berkeley, CA: University of California Press, pp. 106–128. Tomasello, M., Call, J. & Hare, B. 1998. Five primate species follow the visual gaze of conspecifics. Animal Behaviour, 55, 1063–1069. van de Poll, N. E., de Jonge, F., van Oyen, H. G. & van Pett, J. 1982. Aggressive behaviour in rats: effects of winning and losing on subsequent aggressive interactions. Behavioural Processes, 7, 143–155. van Erp, A. M. M. & Miczek, K. A. 2000. Aggressive behavior, increased accumbal dopamine and decreased cortical serotonin in rats. Journal of Neuroscience, 15, 9320–9325. Veenema, H. C. 2000. Methodological progress in post-conflict research. In: Natural Conflict Resolution, ed. F. Aureli & F. B. M. de Waal. Berkeley, CA: University of California Press, pp. 21–23. Verbeek, M. 1998. Bold or cautious: behavioural characteristics and dominance in great tits. Ph.D. Thesis, University of Wageningen, the Netherlands. Verbeek, M. E. M., de Goede, P., Drent, P. J. & Wiepkema, P. R. 1999. Individual behavioural characteristics and dominance in aviary groups of great tits. Behaviour, 136, 23–48.

217

P1: GFZ/... P2: JZZ/... 0521823617c10.xml CU1917B/McGregor

218

0 521 582361 7

April 11, 2005

14:27

A. J. N. Kazem & F. Aureli Virgin, C. E., Jr & Sapolsky, R. M. 1997. Styles of male social behavior and their endocrine correlates among low-ranking baboons. American Journal of Primatology, 42, 25–39. Wahaj, S. A., Guse, K. R. & Holekamp, K. E. 2001. Reconciliation in the spotted hyena (Crocuta crocuta). Ethology, 107, 1057–1074. Walker Leonard, J. 1979. A strategy approach to the study of primate dominance behaviour. Behavioural Processes, 4, 155–172. Watts, D. P. 1995a. Post-conflict social events in wild mountain gorillas (Mammalia, Hominoidea). I. Social interactions between opponents. Ethology, 100, 139–157. 1995b. Post-conflict social events in wild mountain gorillas. II. Redirection, side direction, and consolation. Ethology, 100, 158–174. Weiss, J. M., Pohorecky, L. A., Salman, S. & Gruenthal, M. 1976. Attenuation of gastric lesions by psychological aspects of aggression in rats. Journal of Comparative and Physiological Psychology, 90, 252–259. Winberg, S. & Lepage, O. 1998. Elevation of brain 5-HT activity, POMC expression, and plasma cortisol in socially subordinate rainbow trout. American Journal of Physiology, 274, R645–R654. Zabel, C. J., Glickman, S. E., Frank, L. G., Woodmansee, K. B. & Keppel, G. 1992. Coalition formation in a colony of prepubertal spotted hyenas. In: Coalitions and Alliances in Humans and Other Animals, ed. A. H. Harcourt & F. B. M. de Waal. Oxford: Oxford University Press, pp. 113–135. Zahavi, A. 1977. Reliability in communication systems and the evolution of altruism. In: Evolutionary Ecology, ed. B. Stonehouse & C. M. Perrins. London: Macmillan, pp. 253–259. 1979. Why shouting? American Naturalist, 113, 155–156.

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

16:56

11

Scent marking and social communication jane l. hurst University of Liverpool, Liverpool, UK

Introduction The use of chemical scents for communication between individuals is widespread among both vertebrate and invertebrate animals. Scent signals emanating from an animal’s body can be used for intimate and immediate communication when two or more individuals interact at close quarters, but scents can also be deposited in the environment in the form of scent marks. Unlike most visual or acoustic signals used by animals, scent marks persist in the absence of the signaller, often over extended periods. The prolonged duration of signals in deposited scent marks makes them particularly suited for broadcasting information to all conspecifics that visit a scent-marked site. Further, scent marks might be deposited to signal to certain individuals, such as when animals are attempting to attract potential mates or to indicate a territorial boundary to neighbours. Once deposited in the environment, however, the scent is not physically directed towards specific recipients and the information will be available to any other animals in the locality. This ready availability of scent marks to third parties is likely to provide strong selection pressure to ensure that the information deposited in scent marks is appropriate for communication to any individual likely to encounter the scent. Consequently, scent marks are likely to have evolved to be used for network communication rather than as signals between specific individuals. While volatile components of the scent may be detected at some distance from a scent mark, alerting animals to the presence and location of scent signals, nonvolatile components can be detected only by close contact investigation. Typically, animals approach and sniff scent marks very closely; many species will also lick the scent. Scent marks may thus communicate information via both volatile odorants Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

219

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

220

0 521 582361 7

April 11, 2005

16:56

J. L. Hurst and non-volatile components or those of low volatility, and perception may involve more than one chemosensory system. For example, mammals have both a main olfactory system linked to the olfactory epithelium (reviewed by Nibu, 2002) and an accessory olfactory system linked to the vomeronasal organ (reviewed by Brennan, 2001; Takami, 2002; Zufall et al., 2002). Scent marking is particularly common among terrestrial, arboreal and subterranean mammals and plays an important role in a number of social contexts, including recognition of group members and kin, the advertisement of territory ownership and social dominance, assessment of the quality of competitors and potential mates, and the advertisement and control of reproductive status. In each of these contexts, communication characteristically involves all individuals that deposit and detect scent marks within a particular area rather than private communication between two individuals. To illustrate this, I will review the use of scent marks for communication among house mice Mus musculus domesticus, in which both the behavioural and biochemical mechanisms underlying scent communication have been well studied, both in the laboratory and among wild mice in more naturalistic enclosure experiments. In the concluding part of the chapter, I will discuss the advantages of using scent marks as broadcast signals to communicate to a network of conspecifics rather than directing signals to specific individuals.

Scent marking among house mice Urine is the primary source of social odours among house mice and contains ‘fixed’ (genomic) information about the species, sex, individual identity, major histocompatibility complex (MHC) type and other genetic background of the owner, as well as ‘variable’ (metabolic) information on the owner’s current social, reproductive and health status, and its food resources (reviewed by Brown, 1985a, 1995; Hurst et al., 2001a; Malone et al., 2001). Once old enough to leave the nest, mice of both sexes scent mark by depositing urine in small spots and streaks on the substrate in a deliberate pattern of deposition as they move around their home area (Fig. 11.1). All surfaces are rapidly covered with this background scent marking and scent mark density patterns generally correspond with spatial patterns of activity, most marks being deposited around physical edges such as walls, feeding sites and near nest sites (Fig. 11.2). Mice do not appear to urine mark within their nest site. Once mice have scent marked a patch of substrate thoroughly, the rate of further background marking decreases but it persists at a low level so the animals are continually refreshing their scent. Because mice generally live in territorial social groups consisting of one dominant male together with one or more breeding females and their offspring, and variable numbers of non-breeding adult females

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

16:56

Scent marking and social communication

(a) Water stimulus

(b) Intruder urine stimulus

(c) Urine scent posts Fig. 11.1. Scent marks of wild house mice. Urine scent marks deposited by a resident male territory owner on a patch of substrate (15 cm × 15 cm of absorbent paper) introduced into the territory for 30 min. The patch was first treated with 10µl of either water (a) or urine from an intruder male (b). The white box indicates location of the stimulus). Urinary scent marks (white) were visualized under ultraviolet light using a FluorS scanner (Bio-Rad Laboratories Ltd, Hemel Hempsted, UK). (c) Urine scent posts (approx 2 cm tall, see white box) on a wooden batten where mice frequently rested in an infested poultry house.

and subordinate males (reviewed by Barnard et al., 1991), the substrate becomes smothered in urinary scent marks from all mice that share the same home area (Hurst, 1989). Not all mice contribute equally to this background scent marking; the dominant male territory owner, responsible for most defence of the territory, marks at a much higher rate than other individuals. In situations where males

221

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

222

0 521 582361 7

April 11, 2005

16:56

J. L. Hurst

(a)

Male mouse

(b)

Male mouse 1.2 m

Fig. 11.2. Correspondence between activity and scent deposition when an adult male mouse explored a clean enclosure neighbouring his territory for 15 min. Arrow indicates the hole through which the male entered from his own territory. (a) Location sampled three times per second. (b) Urine marks were sampled by swabbing each 15 cm × 15 cm floor section. The darkness of spots shows the amount of urine recovered, visualized by immunoassay of the major urinary proteins. The nestbox (circle) and food station (rectangle) were not swabbed. (Data collected by Karen Sanders.)

have to defend their territory from other males, dominant male territory owners deposit large streaks as well as smaller spots of urine wherever they go (Desjardins et al., 1973; Hurst, 1990), aided by hairs on the end of the adult male prepuce (Maruniak et al., 1975). Adult mice excrete a number of species- and sex-specific volatiles in their urine that are under hormonal control, in addition to a large number of non-specific volatile metabolites (Novotny et al., 1984, 1990; Schwende et al., 1986; Harvey et al., 1989). Male-specific signalling volatiles include two sesquiterpenes, (E , E ) α-farnesene and (E ) β-farnesene, which are secreted into the urine by the preputial glands, and 2-sec-butyl-4,5-dihydrothiazole and 2,3-dehydro-exo-brevicomin, which

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

16:56

Scent marking and social communication are present in the bladder before any additions from accessory glands (Harvey et al., 1989; Novotny et al., 1990). A large number of volatiles in female urine vary over the oestrous cycle (Andreolini et al., 1987) and during pregnancy and lactation ( Jemiolo et al., 1987), although it is not yet known which (if any) of these are chemosignals or metabolic by-products. One urinary constituent, 2,5-dimethylpyrazine, produced under adrenal control by group-housed females is known to inhibit reproduction in other females (Novotny et al., 1986; Jemiolo & Novotny, 1993, 1994). In addition to these sex-specific volatiles, mouse urine is characterized by the presence of a high concentration of protein, over 99% of which is contributed by a group of 18–20 kDa lipocalins known as the major urinary proteins (MUPs). The lipocalins are synthesized in the liver and secreted into the plasma; they subsequently pass through the glomerular filter into the urine (Beynon et al., 2001). Although urine of both sexes contains a substantial quantity of these proteins, adult male urine typically contains 20–30 mg/ml protein, approximately three times as much as female urine (Beynon & Hurst, 2003). This sex difference occurs at puberty when there is an increase in excretion among males (Payne et al., 2001). These lipocalin proteins have a central cavity that binds lipophilic molecules. In males, MUPs bind a number of ligands but principally the two male-specific signalling volatiles 2-sec-butyl-4,5-dihydrothiazole and 2,3-dehydro-exo-brevicomin (Bacchini et al., 1992; Robertson et al., 1993; Novotny et al., 1999a). Once urine is deposited as a scent mark, the binding of signalling volatiles to MUPs greatly slows down their evaporation from the scent mark (Hurst et al., 1998; Robertson et al., 2001), extending the duration over which volatiles can be detected (Hurst et al., 1998; Humphries et al., 1999). No MUP ligands have yet been identified in female urine. Although volatiles may be detected from a distance of several centimetres or more, depending on the amount of scent deposited, the response of mice is almost always to approach and contact the scent mark, investigating it at very close quarters. Consequently, it is likely that mice gain information from both the volatile and non-volatile components in scents (Humphries et al., 1999). Recent evidence suggests that the vomeronasal system only detects scents when these are pumped to the vomeronasal organ after contact with a stimulus (Luo et al., 2003).

Advertising territory ownership and competitive ability The network of scent marks deposited and investigated by all individuals within the local population provides a mechanism by which animals can advertise their competitive ability to both competitors and to potential mates in a manner that makes it difficult, if not impossible, to cheat. This reliable mechanism involves

223

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

224

0 521 582361 7

April 11, 2005

16:56

J. L. Hurst both the spatial and the temporal pattern of competitive scent marking as well as the quality of an individual’s scent. Spatial and temporal distribution of scent

Like many other mammals, dominant male house mice advertise their territory ownership by scent marking throughout their defended area, marking at a much higher rate than other individuals (Ralls, 1971; Desjardins et al., 1973; Hurst, 1990). Territory owners continually refresh these scents at a high rate so that, in frequently used sites, small posts of dried urine can build up like small stalagmites (Fig. 11.1c). Because only animals that dominate a territory can ensure that their marks predominate in that area, the spatial pattern and density of scent marks bearing the owner’s individual identity signature provide physical proof of an individual’s territory ownership (Gosling, 1982; Hurst, 1993). Further, the spatial and temporal pattern of scent marks from other males indicates the success with which an owner dominates its scent-marked territory. Only males that defend their territory effectively can ensure that no other males deposit competing signals that might attract mates (Hurst, 1993; Hurst & Rich, 1999). Conversely, the presence of any competing signals that are as fresh or fresher than those of the owner will indicate that the owner is not stopping competitors from depositing competing marks and, therefore, is not being very successful in dominating the area, even if the area is suffused with the owner’s scent. Accordingly, dominant male mice rapidly counter-mark if they encounter competing scent signals from other males in their territory, as well as attacking and excluding from the territory any competitors that deposit such competing scent marks (Ralls, 1971; Hurst & Rich, 1999). In house mice, counter-marking consists of a rapid increase in the rate of urine scent marking in the vicinity of, but not specifically on top of, a competitor male’s scent (Fig. 11.1b; Hurst, 1989; Humphries et al., 1999). In other species, counter-marking may take the form of over-marking the competitor’s scent to prove which scent was deposited most recently ( Johnston et al., 1997; Johnston, 1999; Ch. 16). House mice do not attempt to over-mark the scent marks of a competitor, perhaps because their urinary scent marks are scattered so widely in numerous spots and streaks. Instead, mice assess the difference in age of nearby scent marks from competing males to determine which male’sscent was deposited most recently (Rich & Hurst, 1999). Because scent marks remain in the environment and are long lived (see p. 232), the spatial and temporal pattern of scents from different individuals provides a continuous record of any challenges for dominance over the area and, crucially, the outcome of those challenges. This record is available for investigation by any other animal in the area. Animals do not need to witness, or to eavesdrop on, individual challenges for dominance while these are occurring, because the outcome will be readily apparent from scent marks for as

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

16:56

Scent marking and social communication long as the marks retain the individual signatures of the depositors (but see Ch. 2 for a discussion of extra information available from interactions). Scent marks provide the ‘minutes’ of a meeting that are made public to all interested parties, although they may not include a transcript of all of the information involved in the detailed arguments expressed during the meeting. Experimental manipulation of the scent marks within male mouse territories has shown that this information is used by third parties to assess the competitive ability of territory owners: (a) by females when selecting high-quality mates, and (b) by other males when deciding whether to avoid a territory owner or to challenge the owner for dominance themselves. These are discussed separately below. Assessment by females

Although female house mice generally nest and raise their offspring within one male’s territory, they often visit or range over several neighbouring male territories (Hurst, 1987) and extraterritorial matings occur frequently (e.g. 43% of all matings in large captive populations occurred when a female travelled to, and mated with, a male owning a nearby territory (Potts et al., 1991)). Interestingly, among wild house mice, resident male territory owners appear to show little or no discrimination against offspring sired by other males that are reared within their territory (Hurst & Barnard, 1992). Therefore, females usually have a choice between several high-quality male territory owners as potential mates in the local population, regardless of where they choose to nest, because there is little risk to their offspring. Most simply, females can compare the scent marks left by two competing males to assess which male’s scent was deposited most recently. Since only a male that is successfully preventing other males from depositing competing scents can ensure that his marks are the freshest in that location, this male must have been the winner of the conflict. Subsequently given a choice between the two signalling males, females generally prefer the owner of the most recently deposited scent (e.g. Johnston et al., 1997; Fig.11.3a). However, females can also use the presence of scent marks from any males in an owner’s territory to make a much more general assessment of each territory owner’s competitive ability when choosing between territory owners. By manipulating the scent marks in equivalent male territories, Rich & Hurst (1998) showed that female house mice prefer the owners of exclusively scent-marked territories (those containing no scents from competitor males) over neighbouring males whose territories contain some competing counter-marks from an intruder male (Fig. 11.3b). After exploring the males’ territories, females spent more time sniffing and chewing a barrier to gain access to the owner of the exclusively scent-marked territory and were more affiliative and invited attempted mounts from this territory owner when they were allowed

225

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

226

0 521 582361 7

April 11, 2005

16:56

J. L. Hurst

Fig. 11.3. Effects of counter-marking between males on female preferences in different types of test. (a) Female encounters a single scent mark from male A that is deposited on top of scent from male B (e.g. Johnston et al., 1997). (b) Females explore neighbouring male scent-marked territories, one of which contains patches of intruder (i) scent counter-marking the owners scent (Rich & Hurst, 1998). (c) Both neighbouring male territories contain patches of intruder (i) scent (Rich & Hurst, 1999).

to interact. When given a choice between two males whose territories both contained some intruder scent marks, but one owner had counter-marked the intruder scents while some of the other owner’s scent had been counter-marked by the intruder, females preferred the owner that had counter-marked the intruder’s scent (Fig. 11.3c; Rich & Hurst, 1999). In both cases, intruder scent marks came from unfamiliar males; therefore, females were not making a simple choice between two interacting males. Neither were females simply responding to the freshest scents encountered because both territories contained fresh scent marks from the

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

16:56

Scent marking and social communication territory owners. Female preferences resulted from their assessment of the pattern of male scent marks and counter-marks as an indicator of each male’s competitive ability. Scent marks were manipulated in such a way that only the females were exposed to the manipulation, while the territory-owning males were temporarily removed from their territories, so responses were not a result of any changes in male behaviour in response to the intruder scent marks introduced into each male’s territory. Female mice appear to distinguish the most recent scent (i.e. the counter-mark) by the age difference between the male’s scents (Rich & Hurst, 1999). When both territory owners’ and intruders’ scents were of very similar age, females failed to show a preference in favour of an owner that had counter-marked intruder scents. This could be because females were unable to discriminate between scent marks and counter-marks without a substantial age difference in the scents (24 hours in these experiments), or because the similar age of the scents indicated that the competition between the males had yet to be resolved. Males can thus gain a reproductive advantage from scent marking their territories and from counter-marking the scents of any competitors to ensure that their own scent marks are those most recently deposited. Although it is often assumed that the very high rates of marking at borders between neighbouring territory owners are signals to warn neighbours to keep out, the main function of frequent scent marking at shared borders may be the need of both territory owners to ensure that their own scent is as fresh as their neighbour’s wherever their scents are in close proximity. This would require both animals continually to refresh their scent at a shared border as an advertisement to females in the locality (including females resident in the male’s territory and those living in neighbouring areas (Hurst & Rich, 1999)). Male territory owners can thus compete with each other simply through their scent-marking behaviour because competitive scent marks are used by females when selecting a mate. Assessment by males

Third-party competitors also appear to use the record of competitive scentmark signals to identify and avoid challenging owners that are defending their territory effectively against other males. Unfamiliar intruders use the scent marks deposited around a territory to identify the territory owner and are much less likely to challenge a male whose individual scent signature matches the local scent marks than a male whose scent does not match (Gosling & McKay, 1990). Adding a small drop of fresh urine from the territory owner onto one of the owners’ scent marking posts increases the frequency with which intruders and resident subordinates spontaneously flee when they encounter the owner, without any attack or pursuit. In contrast, addition of urine from a neighbouring territory owner reduces their evasion and increases challenges against the territory owner,

227

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

228

0 521 582361 7

April 11, 2005

16:56

J. L. Hurst regardless of their own previous experience of the high competitive ability of the territory owner during direct interactions (Hurst, 1993; Hurst & Rich, 1999). Thus, territory owners also appear to gain a strong advantage in competitive interactions with other males through broadcasting their scent around the territory and through advertising their ability to overcome challenges from other males, reducing their need to invest in direct aggression. Other third-party males gain by avoiding challenging a male that is successfully defending his territory against other competitors but will also rapidly detect when the owner is struggling to maintain dominance (for example, if his competitive ability is reduced by ageing, injury or disease). This may be particularly important to resident subordinates, which are likely to be highly familiar with the greater competitive ability of the dominant male and reluctant to challenge if this might result in their exclusion from the territory.

Deposition patterns and scent age

In contrast to isolated or subordinate (non-competitive) mice, males that are advertising territory ownership or competing to establish a territory change their pattern of scent marking by scattering their urine in a much larger number of streaks and small spots. While this helps to ensure that their scent is distributed throughout the territory, numerous scent marks are deposited close together in the same local area (see Fig. 11.1b). When counter-marking another male’s scent, mice do not attempt to deposit a bigger scent mark than that of the competitor, which would contain a greater intensity of volatile signalling molecules. Instead, they deposit many small marks in the vicinity, returning repeatedly to add more scent marks usually over a period of several hours (Humphries et al., 1999). By dribbling out their urine rather than depositing it all in one go, they are maximizing the freshness of their scent marks by increasing the rate of replenishment (Hurst et al., 2001a). Thus, each time they deposit a new scent mark, they are increasing the age difference between their own scent and that of the competitor, while volatiles in their own fresh scent attract the attention of others to the aged scents of a competitor. Notably, males counter-mark both fresh and aged scents from competitors but deposit most marks near to the aged competitor scent where the contrast will be greatest (Humphries et al., 1999). Most male signalling volatiles are lost from scents within a few hours (Hurst et al., 1998; Humphries et al., 1999; Robertson et al., 2001), but non-volatile components of scent marks continue to be detected for at least seven days if males are aware of the presence of scent marks in the area (Humphries et al., 1999, 2001). Both volatile and non-volatile components of a scent mark are likely to be involved in providing a reliable signal of scent-mark age. While volatile components

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

16:56

Scent marking and social communication will be lost as a scent mark ages, the intensity of a volatile signal at any point in time will depend on the amount initially deposited as well as the time since deposition. Receivers will not be able to assess the age of the mark from a volatile signal alone without knowing the amount that was deposited (Hurst et al., 2001b). In contrast, non-volatile components are not lost through time and may provide a measure of the amount of scent deposited. As yet, we do not know the molecular mechanism used to assess scent-mark age, but MUPs that bind and slowly release volatile ligands in mouse urine have the capacity to provide a very reliable indicator of scent-mark age. Because each protein molecule can only bind one ligand molecule, and ligands are slowly released and evaporate from the scent mark, the proportion of protein molecules that contain ligands will decrease with time since deposition. Making one component the ligand of the other defines implicitly the relationship between them. By contrast, the ratio between two unrelated volatile components that have different rates of evaporation requires that the receiving animal knows the ratio between them at the time of deposition. Although MUPs are not odorants, these non-volatile proteins appear to be detected through the vomeronasal system (Brennan et al., 1999; Krieger et al., 1999). Direct contact with the scent source appears to be essential for activation of this system, suggesting that non-volatile components like MUPs are necessary to deliver volatile pheromones to the vomeronasal organ (Luo et al., 2003). The release of volatiles from MUPs also plays an important function in alerting mice to the presence of a scent mark (Hurst et al., 1998; Humphries et al., 1999).

Advertising subordinate status Females are attracted by sex-specific volatiles in the urine of adult male mice ( Jemiolo et al., 1985, 1991), which also act as reproductive priming pheromones to stimulate female oestrus cycling (see below). These same male signalling volatiles elicit aggression from other competitive males (Novotny et al., 1985), or avoidance by subordinates ( Jones & Nowell, 1989; Novotny et al., 1990; Gosling et al., 1996; Mucignat-Caretta et al., 1998). However, subordinate males that live within the territory of another male and are defeated repeatedly by the dominant territory owner reduce their production of these male-specific volatiles ( Jones & Nowell, 1989; Harvey et al., 1989); their preputial glands are smaller than those of dominant males (Hucklebridge et al., 1972; Bronson & Marsden, 1973) and they are much less likely to initiate competitive interactions (Crowcroft & Rowe, 1963; Hurst, 1987). As a consequence of these changes in scent quality, subordinate male urine is no longer attractive to females (Bronson & Caroom, 1971; Jones & Nowell, 1974; Jemiolo et al., 1991) and females will not mate with subordinate males (Wolff, 1980; Hurst, 1987; Potts et al., 1991). In compensation, subordinate

229

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

230

0 521 582361 7

April 11, 2005

16:56

J. L. Hurst male odours elicit less aggression from other males (Mugford & Nowell, 1970; Jones & Nowell, 1973, 1975; Novotny et al., 1985). Subordinate male mice also show a dramatic and immediate reduction in scent-marking behaviour, although this is not completely suppressed and they continue to deposit scent marks around their home area in larger spots and pools (Desjardins et al., 1973; Sandnabba, 1986). Because their urinary scent differs in quality from that of dominant males, these scent marks advertise their subordinate status within the territory to all other animals in the area, including females. Why should subordinates advertise their low quality in such a public manner? Experimental manipulations of these substrate scents indicate that they are critical in determining tolerance of the subordinate by other resident males. Male mice generally attempt to exclude other adult males from their scent-marked territories (relatives or non-relatives) and they are highly aggressive towards unfamiliar mice or familiar neighbours that intrude into the territory even if intruders are of subordinate status (Barnard et al., 1991). However, complete exclusion can be extremely difficult to achieve in complex habitats where persistent males can hide (Crowcroft, 1966; Poole & Morgan, 1976). Familiar males living in the same territory establish a social structure in which one male becomes dominant and maintains dominance over familiar subordinates through brief attacks and aggressive postures, rather than attempting to evict the subordinates from the territory. If these familiar males are housed in separate cages but their soiled cage substrate is regularly mixed to maintain their contact with group scent cues, males continue to be relatively tolerant of each other. However, if a familiar subordinate male is suddenly prevented from contributing fresh scent to the mixed group-marked substrate, although the subordinate itself continues to encounter group substrate scents as if it were still a group member, within 24–48 hours both the resident dominant and other subordinate males in the group start to investigate and to attack and chase this male as if he was no longer a tolerated group member (Hurst et al., 1993). In contrast, scent marks deposited to compete with the signals of the dominant territory owner, for example on territorial scent-marking posts, induce attack against the familiar subordinate (Hurst, 1993). Competitive pressure from dominant male territory owners thus appears to force subordinate males both to change the quality of their urinary scent and to deposit urinary scent marks that advertise their subordinate status to any animals using or visiting the area. Because the scent marks remain in the environment once deposited, and carry the subordinate’s individual identity signature (see below), a subordinate would be unable to cheat by altering the quality of his scent during direct interactions with females. In order to be tolerated and allowed to remain within another male’s territory, subordinates appear to be forced to broadcast honest signals of their low competitive ability at the cost of their reproductive

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

16:56

Scent marking and social communication success, because their scent marks will be encountered by females. The outcome of this is to reduce considerably the risk that subordinate males might compete for females through sneaky matings (mating with subordinate males is very rarely observed in seminaturalistic studies of mouse populations), making it safe for the territory owner to tolerate subordinates within the territory that contribute to group substrate scents. Notably, experiments have revealed that animals of low competitive ability only suppress the production of competitive male scents when they live in very close proximity to a dominant individual; they do not suppress scent signals simply in response to defeat by another male. Jones and Nowell (1989) confirmed that, if males are repeatedly defeated by a higher quality competitor and are kept in continuous olfactory contact with the scent of their victor (as if they lived within his territory), the defeated subordinate male’s scent loses the aversive effect on other males caused by male signalling volatiles. However, if males experience the same frequency of defeat but are housed in separate cages from their victor (as if they could escape to another territory), the defeated male’s scent retains the high levels of male signalling volatiles that are aversive to other males. Not surprisingly, males do not advertise their low quality unless they are forced to do so by the constant threat of attack and displacement from higher quality competitors.

Female reproductive priming Female reproductive physiology is strongly influenced by a number of reproductive priming pheromones in the urine of male or female mice (reviewed by Brown, 1985b; Novotny et al., 1999a). Volatile priming pheromones in the urine of adult male mice have stimulatory effects on female physiology, accelerating puberty in young females (Vandenberg, 1969; Novotny et al., 1999b) and stimulating oestrous cycling (Whitten, 1956; Jemiolo et al., 1986), while urinary odours from pregnant or lactating females have similar though not identical effects (Drickamer & Hoover, 1979; Hoover & Drickamer, 1979). The stimulatory effects of urine from pregnant or lactating females may reflect the preference of house mice for communal nesting, because females raise more offspring when cooperating with another female than they can when breeding alone (Konig, 1994a). In contrast, nestling survival is greatly reduced in overcrowded nest sites, particularly among females of low social status (Southwick, 1955; Hurst, 1987). Accordingly, when non-breeding females live in groups with several other non-breeding females, or have frequent contact with the urine of other non-breeding females, they produce a priming pheromone in their own urine that inhibits oestrous cycling in adults (Champlin, 1971) and delays puberty in prepubertal females (Drickamer, 1977; Jemiolo & Novotny, 1994). However, females of high social status are not affected

231

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

232

0 521 582361 7

April 11, 2005

16:56

J. L. Hurst and continue to breed (Lloyd & Christian, 1969). Lastly, if a female encounters the scent of a novel male within four days of mating, implantation fails and she will abort if she is not protected by continued exposure to urine from the familiar stud male, a phenomenon known as the Bruce effect (Bruce, 1959; Brennan, 1999). This is likely to provide females with the opportunity to mate with a new territory owner if their previous mating partner is displaced. Since male territory owners do not appear to discriminate against the offspring of other males born within their territory (Hurst & Barnard, 1992), this may be a tactic to increase offspring quality. Although priming pheromones are volatile, these are detected through the vomeronasal system and animals must make contact with the scent source to allow chemical stimuli to be pumped to the vomeronasal organ (Brown, 1985b; Luo et al., 2003). Under natural conditions, females are surrounded by urinary scent marks from all individuals using the same sites, exposing them to both stimulatory and inhibitory priming pheromones. Because these urinary scents are very widely distributed and females do not appear deliberately to control their exposure to these cues (Hurst & Nevison, 1994), this network of scent signals from all animals using the same area appears to provide a mechanism for females to adjust their own reproductive physiology appropriately, according to the current local social conditions and to the individual’s own age and social status.

Individual scent signatures Scent marks are deposited in the environment to provide information in the absence of the signaller (unlike other types of signal), so it is essential that they provide stable and persistent information about the donor’s individual identity. Ideally, individuality scents should be ‘hard-coded’ in the individual’s genome, exhibit a high degree of individual polymorphism to uniquely identify the donor and be expressed by all individuals regardless of social status or sex (Beynon et al., 2001). Attention has focused largely on the volatile components of scents as sources of individuality signatures, particularly those associated with the highly polymorphic major histocompatibility complex (MHC) odortypes, although many other genetic loci also influence individual differences in urinary volatile profiles (Boyse et al., 1987; Beauchamp et al., 1990; Eggert et al., 1996). The MHC encodes for glycoproteins involved in individual (self versus non-self) recognition at the cellular level but has also been shown to affect the volatile scent signals produced by animals such as mice, rats, fish and humans (reviewed by Jordan & Bruford, 1998; Singh, 2001; see also Olsen et al., 1998; Reusch et al., 2001; Jacob et al., 2002). Laboratory studies using MHC congenic strains of mice and rats have confirmed that rodents are able to discriminate differences in urinary odours from donors

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

16:56

Scent marking and social communication that differ genetically at alleles within the MHC region, even when donors differ only at a single MHC locus (Yamazaki et al., 1999; Singh, 2001; Carroll et al., 2002). Although the molecular basis of MHC-associated odours is not known, it appears to involve a complex mixture of volatile metabolites bound and released by urinary proteins (Singer et al., 1993, 1997). One hypothesis is that soluble fragments of MHC class I and class II molecules in urine differentially bind volatile metabolites in the antigen-binding groove once the peptide normally bound in this groove is lost (Singh, 2001). Alternatively, MHC haplotype may affect the volatile metabolites that are released into the urine (Singer et al., 1997; Yamazaki et al., 1999). These MHC-dependent volatiles might then be bound and released by the MUPs, which are present at protein concentrations up to a million times higher than MHC class I molecule fragments and possess a large flexible binding pocket for small lipophilic (and thus potentially volatile) molecules (Beynon et al., 2001). Although MHC polymorphism results in differences in scents that are discriminable by mice, these volatile signals appear to be easily disrupted by environmental factors that affect an individual’s metabolite profile, such as changes in food type, bacterial gut flora or social status (reviewed by Brown, 1995; Nevison et al., 2000). This presents a problem for an individual recognition signature, suggesting that MHC-associated odours may not provide sufficient stability or persistence to act as individuality signals in scent marks. Indeed, mice do not use MHC-associated odours to discriminate their own scent marks from those of other males, despite their clear ability to detect differences in their own MHC type and those of other individuals ( J. L. Hurst, unpublished data). Instead, they use the different patterns of MUPs that individual mice express in their urine (Hurst et al., 2001b). MUPs are coded by a multigene family on chromosome 4 and are expressed at high concentration by adult house mice of both sexes, although males invest more than females in both scent marking and MUP production (Beynon et al., 2001; Payne et al., 2001; Beynon & Hurst, 2003). MUPs exhibit a very high level of genetic polymorphism and individual mice express a combination of MUPs (typically at least 7–12) such that the combinatorial diversity of individual MUP profiles among wild mice may be as great as for MHC (Robertson et al., 1997; Beynon et al., 2001, 2002; Payne et al., 2001). These individual-specific patterns of urinary MUPs in the scent marks of wild mice appear to be essential in allowing outbred wild mice to distinguish another individual’s scent marks from their own, regardless of differences at many other genetic loci such as MHC (Hurst et al., 2001b). If a male territory owner encounters a competing scent mark from another male in his territory, he normally investigates closely and then counter-marks and spends time in the vicinity of the scent mark. However, males only recognize an intruder’s scent mark if it carries a different MUP pattern from their own.

233

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

234

0 521 582361 7

April 11, 2005

16:56

J. L. Hurst Scent marks with their own MUP pattern draw initial investigation but no further response (Hurst et al., 2001b; J. L. Hurst, unpublished data). This is not because males avoid competing with a relative (i.e. kin discrimination) as they respond strongly to urinary scents of different MUP type whether from a close relative or an unrelated male. In contrast, if their own urinary MUP profile is altered by adding a recombinant MUP to their urine, males counter-mark as if the scent was from an intruder (Hurst et al., 2001b). There is a high degree of individual variability in MUP patterns expressed by males captured from the same population (Payne et al., 2001) and MUPs show little or no degradation in scent marks (Hurst et al., 2001a). These patterns, therefore, provide a stable and persistent individual scent signature that is hard-coded in the individual’s genome and remains constant throughout the individual’s lifetime.

Kin and group member recongnition Inherited scent signatures are also important in allowing animals to recognize whether others are likely to be close relatives. Many genetic differences appear to contribute to inherited scents used for kin recognition, including MHCassociated odours. For example, female mice prefer to rear their offspring communally with close relatives rather than with unrelated females, and offspring survival is greater when cooperating with a familiar sister (Konig, 1994b). When unfamiliar female mice are mixed together in seminatural populations, they appear to recognize the similar scents of other females of the same MHC type as themselves and are more likely to share nests with these females (Manning et al., 1992). Recognition of close relatives is particularly important to avoid inbreeding. There is considerable evidence that MHC type affects female preference between male scents, with females generally preferring the scents of males of different MHC type to themselves or their parents (e.g. Egid & Brown, 1989; Penn & Potts, 1998). By crossing wild mice with laboratory strains to create wild-type mice of known homozygous MHC type, Potts et al. (1991) confirmed that females in seminatural populations showed a significant preference for MHC-disassortative mating. Interestingly, females showed no MHC bias when mating with the owner of the territory in which they nested (suggesting that MHC type did not influence territory preference despite the genetic information in the owner’s scent marks). However, when females went outside their territory to mate, they preferred owners of neighbouring territories that had a different MHC type to their own familiar MHC-associated odours. In mice, the mechanism of kin recognition appears to be largely through imprinting on scents experienced during early life rather than on cues inherited by

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

16:56

Scent marking and social communication the female herself (D’Udine & Alleva, 1983). Mating preferences can be reversed by cross-fostering. Females fostered onto a different strain will later avoid mating with males related to their foster parents rather than with those related to themselves (e.g. Penn & Potts, 1998; though see Eklund, 1997). Because relatives will carry a much greater range of alleles than the female herself, imprinting on the scents of relatives experienced in their early environment may be more effective than phenotype matching to self for recognizing potential relatives or those from a similar genetic background. In addition to inherited scents, mice acquire scents on their bodies from other group members that influence recognition when mice interact (Aldhous, 1989). All group members are likely to become tainted with the scents of the resident territory owner from the sticky scent marks deposited throughout the area. This acquired group scent may make an important contribution to the ability of mice to recognize their own group members regardless of their inherited scents (e.g. Hurst & Barnard, 1992, 1995).

Scent marking as broadcast signals Animals spontaneously scent mark their territories in the absence of interaction with others, although scent marking is usually significantly enhanced by competitive and sexual interaction with others or with their scent marks. In addition to the chemical information in an individual’s scent, the spatial and temporal patterning of scent marks from all individuals in a locality provides information about their social and genetic relationships. Scent marks are particularly suitable for network communication between many individuals. By their very nature, they are long lasting and readily available for inspection by any individuals that visit a scent-marked site. Conversely, scent marks are not appropriate for private communication (see Ch. 3) unless individual access to the scent marked site is restricted. There are two ways in which scent marking might be used in a signalling network. First, scent marks may be deposited as broadcast signals, designed to communicate information to all other animals in the area. Alternatively, scents may be deposited as signals to specific individuals, with third parties making use of information in scent marks and counter-marks by eavesdropping on the communication between others. Eavesdropping has been defined as ‘extracting information from signalling interactions between others’ (McGregor & Dabelsteen, 1996; McGregor & Peake, 2000; Ch. 2). The implication here is that signals are designed to provide information to one or more interacting individuals (e.g. during the interaction between two competitors), not to provide information to eavesdroppers. However, eavesdroppers make use of this information to their own advantage. It

235

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

236

0 521 582361 7

April 11, 2005

16:56

J. L. Hurst is, therefore, important to establish who gains an advantage from the use of these signals. With respect to competitive counter-marking by territory owners in response to aggressive challenges and scent marking from other males, the main fitness advantage to signallers appears to be the response of females to these scent marks, although females may be viewed as ‘third parties’ responding to signalling interactions between males (see Ch. 7 for parallels with bird song). Evidence from our scent-manipulation studies in house mice indicates that these scent signals have highly significant effects on female preferences, in accordance with our hypothesis that this is a very reliable way to assess the competitive ability of different potential mates. Similarly, territorial scent marking and the counter-marking of any intruder scents by the owner increases avoidance responses and decreases competitive challenges from third-party males (i.e. males that are not the owners of either the scent marks or counter-marks). As this affects the responses of all males in the vicinity, this is likely to have a big impact on the ease with which males defend their territories. Therefore, a successful territory owner gains clear advantages from advertising his territory ownership and competitive ability widely, including his ability to overcome the challenges of competitors. While increased scent marking to counter-mark a competitor’s scent might, at first sight, appear to be a signalling interaction between two competitors, the main selective advantage to successful competitors comes from broadcasting this information to all animals in the vicinity. There is no evidence that counter-marking the scent of an aggressive competitor reduces challenges from the competitor itself without direct aggression, although once defeated, competitors will show a generalized avoidance of competitor scents (e.g. Jones & Nowell, 1989; Hurst et al., 1997). McGregor & Peake (2000) pointed out that there seem to be few demonstrated advantages to signallers of communicating in the social environment of a network. However, competitive scent marking provides an excellent example of the advantages that successful territory owners can gain from depositing competitive scent marks within the network of signals from other males, because this provides a mechanism for the reliable advertisement of their own competitive ability. These scent marks are clearly broadcast signals, designed to communicate information to all other animals in the area. As such, the concept of eavesdropping on signals aimed at others does not seem to be appropriate. The deposition of scent signals that advertise an individual’s subordinate status appears to be enforced by the resident territory owner and other resident subordinate males, and thus subordinate scent marking might be viewed as a signal aimed principally at local competitors to reduce aggression against the subordinate. However, the selective advantage to local competitors of reducing their aggression is that such scent marks broadcast the subordinate status and

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

16:56

Scent marking and social communication identity of a competitor to females. Subordinates that signal their subordinate status through their body scent but not through substrate scent marks are not tolerated. From an evolutionary viewpoint, it again seems more appropriate to view scent marks deposited to advertise subordinate status as broadcast signals, aimed at publicising this information to others despite the immediate reduction in reproductive opportunities to the signaller from doing so. This at least allows unsuccessful males to remain in a suitable habitat with the potential to become a successful territory owner, and gain reproductive success, in the future.

Summary In conclusion, since scent marks persist in the environment and cannot be directed towards specific recipients (unlike most visual and acoustic signals), scent marks are only likely to be used as broadcast signals and are used in social contexts where the signaller can gain an advantage from communicating information to a public audience.

Acknowledgements I am grateful to Rob Beynon for providing Fig. 11.2, for invaluable discussion and for comments on the manuscript. Research carried out in my laboratory was supported by grants from the Biotechnology and Biological Sciences Research Council.

References Aldhous, P. 1989. The effects of individual cross-fostering on the development of intrasexual kin discrimination in male laboratory mice. Animal Behaviour, 37, 741–750. Andreolini, F., Jemiolo, B. & Novotny, M. 1987. Dynamics of excretion of urinary chemosignals in the house mouse (Mus musculus) during the natural estrous cycle. Experientia, 43, 998–1002. Bacchini, A., Gaetani, E. & Cavaggioni, A. 1992. Pheromone binding proteins of the mouse, Mus musculus. Experientia, 48, 419–421. Barnard, C. J., Hurst, J. L. & Aldhous, P. 1991. Of mice and kin: the functional significance of kin bias in social behaviour. Biological Reviews of the Cambridge Philosophical Society, 66, 379–430. Beauchamp, G. K., Yamazaki, K., Duncan, H., Bard, J. & Boyse, E. A. 1990. Genetic determination of individual mouse odour. In: Chemical Signals in Vertebrates 5, ed. D. W. Macdonald, S. Natyncuk & D. Muller-Schwarze. Oxford: Oxford University Press, pp. 244–255. Beynon, R. J. & Hurst, J. L. 2003. Multiple roles of major urinary proteins in the house mouse, Mus domesticus. Biochemical Society Transactions, 31, 142–146.

237

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

238

0 521 582361 7

April 11, 2005

16:56

J. L. Hurst Beynon, R. J., Hurst, J. L., Gaskell, S. J. et al. 2001. Mice, MUPs and myths: structure-function relationships of the major urinary proteins. In: Chemical Signals in Vertebrates, ed. A. Marchelewska-Koj, D. Muller-Schwarze & J. Lepri. New York: Plenum Press, pp. 149–156. Beynon, R. J., Veggerby, C., Payne, C. E. et al. 2002. Polymorphism in major urinary proteins: molecular heterogeneity in a wild mouse population. Journal of Chemical Ecology, 28, 1429–1446. Boyse, E. A., Beauchamp, G. K. & Yamazaki, K. 1987. The genetics of body scent. Trends in Genetics, 3, 97–102. Brennan, P. A. 1999. Bruce effect. In: Encyclopedia of Reproduction, ed. E. Knobil & J. Neill. San Diego, CA: Academic Press, pp. 433–438. 2001. The vomeronasal system. Cellular and Molecular Life Science, 58, 546–555. Brennan, P. A., Schellinck, H. M. & Keverne, E. B. 1999. Patterns of expression of the immediate-early gene egr-1 in the accessory olfactory bulb of female mice exposed to pheromonal constituents of male urine. Neuroscience, 90, 1463–1470. Bronson, F. H. & Caroom, D. 1971. Preputial gland of the male mouse; attractant function. Journal of Reproductive Fertility, 25, 279–282. Bronson, F. H. & Marsden, H. M. 1973. The preputial gland as an indicator of social dominance in male mice. Behavioral Biology, 9. Brown, R. E. 1985a. The rodents II: suborder Myomorpha. In: Social Odours in Mammals, ed. R. E. Brown & D. W. Macdonald. Oxford: Clarendon Press, pp. 345–457. 1985b. The rodents I: effects of odours on reproductive physiology (primer effects). In: Social Odours in Mammals, ed. R. E. Brown & D. W. Macdonald. Oxford: Clarendon Press, pp. 245–344. 1995. What is the role of the immune system in determining individually distinct body odours? International Journal of Immunopharmacology, 17, 655–661. Bruce, H. M. 1959. An exteroceptive block to pregnancy in the mouse. Nature, 184, 105. Carroll, L. S., Penn, D. J. & Potts, W. K. 2002. Discrimination of MHC-derived odors by untrained mice is consistent with divergence in peptide-binding region residues. Proceedings of the National Academy of Sciences, USA, 99, 2187–2192. Champlin, A. K. 1971. Suppression of oestrus in grouped mice: the effects of various densities and the possible nature of the stimulus. Journal of Reproductive Fertility, 27, 233–241. Crowcroft, P. 1966. Mice All Over. London: Foulis. Crowcroft, P. & Rowe, F. P. 1963. Social organization and territorial behaviour in the wild house mouse (Mus musculus L.). Proceedings of the Zoological Society of London, 140, 517–531. Desjardins, C., Maruniak, J. A. & Bronson, F. H. 1973. Social rank in the house mouse: differentiation revealed by ultraviolet visualisation of urinary marking patterns. Science, 182, 939–941. Drickamer, L. C. 1977. Delay of sexual maturation in female house mice by exposure to grouped females or urine from grouped females. Journal of Reproductive Fertility, 51, 77–81. Drickamer, L. C. & Hoover, J. E. 1979. Effects of urine from pregnant and lactating female house mice on sexual maturation of juvenile females. Developmental Psychobiology, 12, 545–551.

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

16:56

Scent marking and social communication D’Udine, B. & Alleva, E. 1983. Early experience and sexual preferences in rodents. In: Mate Choice, ed. P. Bateson. Cambridge, UK: Cambridge University Press, pp. 311–327. Eggert, F., Holler, C., Luszyk, D., Muller-Ruchholtz, W. & Ferstl, R. 1996. MHC-associated and MHC-independent urinary chemosignals in mice. Physiology and Behavior, 59, 57–62. Egid, K. & Brown, J. L. 1989. The major histocompatibility complex and female mating preferences in mice. Animal Behaviour, 38, 548–550. Eklund, A. 1997. The major histocompatibility complex and mating preferences in wild house mice. Behavioral Ecology, 8, 630–634. Gosling, L. M. 1982. A reassessment of the function of scent marking in territories. Zeitschrift f¨ ur Tierpsychologie, 60, 89–118. Gosling, L. M. & McKay, H. V. 1990. Competitor assessment by scent matching: an experimental test. Behavioral Ecology and Sociobiology, 26, 415–420. Gosling, L. M., Atkinson, N. W., Dunn, S. & Collins, S. A. 1996. The response of subordinate male mice to scent marks varies in relation to their own competitive ability. Animal Behaviour, 52, 1185–1191. Harvey, S., Jemiolo, B. & Novotny, M. 1989. Pattern of volatile compounds in dominant and subordinate male-mouse urine. Journal of Chemical Ecology, 15, 2061–2072. Hoover, J. E. & Drickamer, L. C. 1979. Effects of urine from pregnant and lactating female house mice on oestrous cycles of adult females. Journal of Reproductive Fertility, 55, 297–301. Hucklebridge, F. H., Nowell, N. W. & Wouters, A. 1972. A relationship between social experience and preputial gland function in the albino mouse. Journal of Endocrinology, 55, 449–450. Humphries, R. E., Robertson, D. H. L., Beynon, R. J. & Hurst, J. L. 1999. Unravelling the chemical basis of competitive scent marking in house mice. Animal Behaviour, 58, 1177–1190. Humphries, R. E., Robertson, D. H. L., Nevison, C. M., Beynon, R. J. & Hurst, J. L. 2001. The role of urinary proteins and volatiles in competitive scent marking among male house mice. In: Chemical Signals in Vertebrates, ed. A. Marchelewska-Koj, D. Muller-Schwarze & J. Lepri. New York: Plenum Press, pp. 353–360. Hurst, J. L. 1987. Behavioural variation in wild house mice Mus domesticus Rutty: a quantitative assessment of female social organization. Animal Behaviour, 35, 1846–1857. 1989. The complex network of olfactory communication in populations of wild house mice Mus domesticus Rutty: urine marking and investigation within family groups. Animal Behaviour, 37, 705–725. 1990. Urine marking in populations of wild house mice Mus domesticus Rutty. 1. Communication Between Males. Animal Behaviour, 40, 209–222. 1993. The priming effects of urine substrate marks on interactions between male house mice, Mus musculus domesticus Schwarz and Schwarz. Animal Behaviour, 45, 55–81. Hurst, J. L. & Barnard, C. J. 1992. Kinship and social behavior in wild house mice: effects of social group membership and relatedness on the responses of dominant males toward juveniles. Behavioral Ecology, 3, 196–206.

239

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

240

0 521 582361 7

April 11, 2005

16:56

J. L. Hurst 1995. Kinship and social tolerance among female and juvenile wild house mice: kin bias but not kin discrimination. Behavioral Ecology and Sociobiology, 36, 333–342. Hurst, J. L. & Nevison, C. M. 1994. Do female house mice, Mus domesticus, regulate their exposure to reproductive priming pheromones. Animal Behaviour, 48, 945–959. Hurst, J. L. & Rich, T. J. 1999. Scent marks as competitive signals of mate quality. In: Advances in Chemical Communication in Vertebrates, ed. R. E. Johnson, D. Muller-Schwarze & P. W. Sorensen. New York: Plenum Press, pp. 209–226. Hurst, J. L., Fang, J. M. & Barnard, C. J. 1993. The role of substrate odours in maintaining social tolerance between male house mice, Mus musculus domesticus. Animal Behaviour, 45, 997–1006. Hurst, J. L., Gray, S. J., Davey, P. et al. 1997. Social interaction alters attraction to competitor’s odour in the mouse Mus spretus Lataste. Animal Behaviour, 54, 941–953. Hurst, J. L., Robertson, D. H. L., Tolladay, U. & Beynon, R. J. 1998. Proteins in urine scent marks of male house mice extend the longevity of olfactory signals. Animal Behaviour, 55, 1289–1297. Hurst, J. L., Beynon, R. J., Humphries, R. E. et al. 2001a. Information in scent signals of competitive social status: the interface between behaviour and chemistry. In: Chemical Signals in Vertebrates, ed. A. Marchelewska-Koj, D. Muller-Schwarze & J. Lepri. New York: Plenum Press, pp. 43–52. Hurst, J. L., Payne, C. E., Nevison, C. M. et al. 2001b. Individual recognition in mice mediated by major urinary proteins. Nature, 414, 631–634. Jacob, S., McClintock, M. K., Zelano, B. & Ober, C. 2002. Paternally inherited HLA alleles are associated with women’s choice of male odour. Nature Genetics, 30, 175–179. Jemiolo, B. & Novotny, M. 1993. Long-term effect of a urinary chemosignal on reproductive fitness in female mice. Biology of Reproduction, 48, 926– 929. 1994. Inhibition of sexual maturation in juvenile female and male mice by a chemosignal of female origin. Physiology and Behavior, 55, 519–522. Jemiolo, D., Alberts, J., Sochinski-Wiggins, S., Harvey, S. & Novotny, M. 1985. Behavioural and endocrine responses of female mice to synthetic analogs of volatile compounds in male urine. Animal Behaviour, 33, 1114–1118. Jemiolo, B., Harvey, S. & Novotny, M. 1986. Promotion of the Whitten effect in female mice by synthetic analogs of male urinary constituents. Proceedings of the National Academy of Sciences, USA, 83, 4576–4579. Jemiolo, B., Andreolini, F., Wiesler, D. & Novotny, M. 1987. Variations in mouse (Mus musculus) urinary volatiles during different periods of pregnancy and lactation. Journal of Chemical Ecology, 13, 1941–1956. Jemiolo, B., Xie, T. M. & Novotny, M. 1991. Socio-sexual olfactory preference in female mice: attractiveness of synthetic chemosignals. Physiology and Behavior, 50, 1119–1122. Johnston, R. E. 1999. Scent over-marking. How do hamsters know whose scent is on top and why should it matter? In: Advances in Chemical Signals in Vertebrates, ed. R. E. Johnson, D. Muller-Schwarze & P. W. Sorensen. New York: Plenum Press, pp. 227–238.

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

16:56

Scent marking and social communication Johnston, R. E., Sorokin, E. S. & Ferkin, M. H. 1997. Female voles discriminate males’ over-marks and prefer top-scent males. Animal Behaviour, 54, 679–690. Jones, R. B. & Nowell, N. W. 1973. Aversive and aggression-promoting properties of urine from dominant and subordinate male mice. Animal Learning and Behavior, 1, 207–210. 1974. A comparison of the aversive and female attractant properties of urine from dominant and subordinate male mice. Animal Learning and Behavior, 2, 141–144. 1975. Effects of clean and soiled sawdust substrates and of different urine types upon aggressive behaviour in male mice. Aggressive Behavior, 1, 111–121. 1989. Aversive potency of urine from dominant and subordinate male laboratory mice (Mus musculus): resolution of a conflict. Aggressive Behavior, 15, 291–296. Jordan, W. C. & Bruford, M. W. 1998. New perspectives on mate choice and the MHC. Heredity, 81, 239–245. Konig, B. 1994a. Components of lifetime reproductive success in communally and solitarily nursing house mice: a laboratory study. Behavioral Ecology and Sociobiology, 34, 275–283. 1994b. Fitness effects of communal rearing in house mice: the role of relatedness versus familiarity. Animal Behaviour, 48, 1449–1457. Krieger, J., Schmitt, A., Lobell, D. et al. 1999. Selective activation of G protein subtypes in the vomeronasal organ upon stimulation with urine-derived compounds. Journal of Biological Chemistry, 274, 4655–4662. Lloyd, J. A. & Christian, J. J. 1969. Reproductive activity of individual females in three experimental freely growing populations of house mice (Mus musculus L.). Journal of Mammalogy, 50, 49–59. Luo, M., Fee, M. S. & Katz, L. C. 2003. Encoding pheromonal signals in the accessory olfactory bulb of behaving mice. Science, 299, 1196–1201. Malone, N., Payne, C. E., Beynon, R. J. & Hurst, J. L. 2001. Social status, odour communication and mate choice in wild house mice. In: Chemical Signals in Vertebrates, ed. A. Marchelewska-Koj, D. Muller-Schwarze & J. Lepri. New York: Plenum Press, pp. 217–224. Manning, C. J., Wakeland, E. K. & Potts, W. K. 1992. Communal nesting patterns in mice implicate MHC genes in kin recognition. Nature, 360, 581–583. Maruniak, J. A., Desjardins, C. & Bronson, F. H. 1975. Adpatations for urinary marking in rodents: prepuce length and morphology. Journal of Reproductive Fertility, 44, 567–570. McGregor, P. K. & Dabelsteen, T. 1996. Communication networks. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca, NY: Cornell University Press, pp. 409–425. McGregor, P. K. & Peake, T. M. 2000. Communication networks: social environments for receiving and signalling behaviour. Acta Ethologica, 2, 71–81. Mucignat-Caretta, C., Caretta, A. & Baldini, E. 1998. Protein-bound male urinary pheromones: differential responses according to age and gender. Chemical Senses, 23, 67–70.

241

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

242

0 521 582361 7

April 11, 2005

16:56

J. L. Hurst Mugford, R. A. & Nowell, N. W. 1970. Pheromones and their effect on aggression in mice. Nature, 226, 967–968. Nevison, C. M., Barnard, C. J., Beynon, R. J. & Hurst, J. L. 2000. The consequences of inbreeding for recognising competitors. Proceedings of the Royal Society of London, Series B, 267, 687–694. Nibu, K. 2002. Introduction to olfactory neuroepithelium. Microscopy Research and Technique, 58, 133–134. Novotny, M., Schwende, F. J., Wiesler, D. Jorgenson, J. W. & Carmack, M. 1984. Identification of a testosterone-dependent unique volatile constituent of male mouse urine: 7-exo-ethyl-5-methyl-6,8- dioxabicyclo(3.2.1)-3-octene. Experientia, 40, 217–219. Novotny, M., Harvey, S., Jemiolo, B. & Alberts, J. 1985. Synthetic pheromones that promote inter-male aggression in mice. Proceedings of the National Academy of Sciences, USA, 82, 2059–61. Novotny, M., Jemiolo, B., Harvey, S., Wiesler, D. & Marchlewska-Koj, A. 1986. Adrenal-mediated endogenous metabolites inhibit puberty in female mice. Science, 231, 722–725. Novotny, M., Harvey, S. & Jemiolo, B. 1990. Chemistry of male dominance in the house mouse, Mus domesticus. Experientia, 46, 109–113. Novotny, M. V., Ma, W., Zidek, L. & Daev, E. 1999a. Recent biochemical insights into puberty acceleration, estrus induction and puberty delay in the house mouse. In: Advances in Chemical Communication in Vertebrates, ed. R. E. Johnson, D. Muller-Schwarze & P. W. Sorensen. New York: Plenum Press, pp. 99–116. Novotny, M. V., Ma, W., Wiesler, D. & Zidek, L. 1999b. Positive identification of the puberty-accelerating pheromone of the house mouse: the volatile ligands associating with the major urinary protein. Proceedings of the Royal Society of London, Series B, 266, 2017–2022. Olsen, K. H., Grahn, M., Lohm, J. & Langefors, A. 1998. MHC and kin discrimination in juvenile Arctic charr, Salvelinus alpinus (L.). Animal Behaviour, 56, 319–327. Payne, C. E., Malone, N., Humphries, R. E. et al. 2001. Heterogeneity of major urinary proteins in house mice: population and sex differences. In: Chemical Signals in Vertebrates, ed. A. Marchelewska-Koj, D. Muller-Schwarze & J. Lepria. New York: Plenum Press, pp. 233–240. Penn, D. & Potts, W. 1998. MHC-disassortative mating preferences reversed by cross-fostering. Proceedings of the Royal Society of London, Series B, 265, 1299– 1306. Poole, T. B. & Morgan, H. D. R. 1976. Social and territorial behaviour in mice (Mus musculus L.) in small complex areas. Animal Behaviour, 24, 476–480. Potts, W. K., Manning, C. J. & Wakeland, E. K. 1991. Mating patterns in seminatural populations of mice influenced by MHC genotype. Nature, 352, 619–621. Ralls, K. 1971. Mammalian scent marking. Science, 171, 443–449. Reusch, T. B., Haberli, M. A., Aeschlimann, P. B. & Milinski, M. 2001. Female sticklebacks count alleles in a strategy of sexual selection explaining MHC polymorphism. Nature, 414, 300–302.

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

16:56

Scent marking and social communication Rich, T. J. & Hurst, J. L. 1998. Scent marks as reliable signals of the competitive ability of mates. Animal Behaviour, 56, 727–735. 1999. The competing counter-marks hypothesis: reliable assessment of competitive ability by potential mates. Animal Behaviour, 58, 1027–1037. Robertson, D. H. L., Beynon, R. J. & Evershed, R. P. 1993. Extraction, characterization and binding analysis of two pheromonally active ligands associated with major urinary protein of house mouse (Mus musculus). Journal of Chemical Ecology, 19, 1405–1416. Robertson, D. H., Hurst, J. L., Bolgar, M. S., Gaskell, S. J. & Beynon, R. J. 1997. Molecular heterogeneity of urinary proteins in wild house mouse populations. Rapid Communications in Mass Spectrometry, 11, 786–790. Robertson, D. H. L., Marie, A. D., Veggerby, C., Hurst, J. L. & Beynon, R. J. 2001. Characteristics of ligand binding and release by major urinary proteins. In: Chemical Signals in Vertebrates, ed. A. Marchelewska-Koj, D. Muller-Schwarze & J. Lepri. New York: Plenum Press, pp. 169–176. Sandnabba, N. K. 1986. Changes in male odors and urinary marking patterns due to inhibition of aggression in male mice. Behavioral Processes, 12, 349–361. Schwende, F. J., Wiesler, D., Jorgenson, J. W., Carmack, M. & Novotny, M. 1986. Urinary volatile constituents of the house mouse, Mus musculus, and their endocrine dependency. Journal of Chemical Ecology, 12, 277–296. Singer, A. G., Tsuchiya, H., Wellington, J. L., Beauchamp, G. K. & Yamazaki, K. 1993. Chemistry of odortypes in mice: fractionation and bioassay. Journal of Chemical Ecology, 19, 569–579. Singer, A. G., Beauchamp, G. K. & Yamazaki, K. 1997. Volatile signals of the major histocompatibility complex in male mouse urine. Proceedings of the National Academy of Sciences, USA, 94, 2210–2214. Singh, P. B. 2001. Chemosensation and genetic individuality. Reproduction, 121, 529–539. Southwick, C. H. 1955. Regulatory mechanisms of house mouse populations: social behavior affecting litter survival. Ecology, 36, 627–634. Takami, S. 2002. Recent progress in the neurobiology of the vomeronasal organ. Microscopy Research and Technique, 58, 228–250. Vandenberg, J. G. 1969. Male odor accelerates female sexual maturation in mice. Endocrinology, 84, 658–660. Whitten, W. K. 1956. Modification of the oestrous cycle of the mouse by external stimuli associated with the male. Journal of Endocrinology, 13, 399–404. Wolff, R. J. 1980. Mating behaviour and female choice: their relation to social structure in wild caught house mice (Mus musculus) housed in a semi-natural environment. Journal of Zoology (London), 207, 43–51. Yamazaki, K., Singer, A. & Beauchamp, G. K. 1999. Origin, functions and chemistry of H-2 regulated odorants. Genetica, 104, 235–240. Zufall, F., Kelliher, K. R. & Leinders-Zufall, T. 2002. Pheromone detection by mammalian vomeronasal neurons. Microscopy Research and Techniques, 58, 251– 260.

243

P1: JZZ/... P2: JZZ/... 0521823617c11.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

244

16:56

P1: JZZ/... P2: JZZ/... 0521823617agg.xml CU1917B/McGregor

0 521 582361 7

Part III C O M M U N I C A T I O N N E T W O R K S IN DIFFERENT TAXA

245

April 11, 2005

15:10

P1: JZZ/... P2: JZZ/... 0521823617agg.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

246

15:10

P1: JZZ/... P2: JZZ/... 0521823617agg.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

15:10

Introduction

Communication networks can be found in any taxonomic group of animals, all that is required is that their signals travel further than the average distance between individuals. This potential for taxonomically widespread occurrence is one of the reasons that communication networks are likely to be an important concept for the understanding of communication in general. However, taxa vary considerably in several aspects that could affect communication networks, including the senses used by receivers (signal modality), processing power and social organization. The potential insights gained from such taxon-related differences are the reason for grouping chapters into this section. Not all taxa are covered in Part III: for example, fish do not appear, but they do in Parts I and IV (Chs. 4, 5, 21, 22 and 23). Also some taxa are underrepresented: there is a preponderance of endothermic vertebrate groups, which is recognized to be a general feature of the literature (Bonnet et al., 2002), and invertebrates have many fewer chapters than their species richness would seem to require. The invertebrate balance is redressed slightly by the fact that insects are the focus of a chapter elsewhere in this book (Ch. 8) and by recent books on insect communication that deal extensively with chorus behaviour (e.g. Gerhardt & Huber, 2002; Greenfield, 2002). Nevertheless, this part does have chapters ranging from fiddler crabs to humans and that is a sufficiently broad taxonomic coverage to demonstrate common themes and illuminating differences.

Fiddler crabs To the casual observer, fiddler crabs on a mudflat would seem to be a clear example of a communication network because of the density of crabs and Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

247

P1: JZZ/... P2: JZZ/... 0521823617agg.xml CU1917B/McGregor

248

0 521 582361 7

April 11, 2005

15:10

Part III the male behaviour of conspicuously waving an enlarged claw. However, as Denise Pope points out in Ch. 12, that is a human perception; we need to know whether the crabs consider themselves part of a network. After carefully weighing up the evidence, particularly on their visual abilities, she concludes that it is likely that most fiddler crab populations do function as networks. Careful review also indicates that the claw-waving display functions predominantly in male–female contexts, focusing attention on the intriguing receiver behaviour of conspecific interceptive eavesdropping and types of competitive signalling interaction.

Anuran amphibians The frogs and toads (anurans) were one of the first taxonomic groups to be considered from a communication network perspective. This early interest was a result of their habit of communicating in striking choruses: groups of calling individuals that can number several thousand. The calls of male anurans incorporate adaptations to enhance the effectiveness of mate attraction in the noisy environment of a breeding pool. However, calling males also have to repel male competitors and avoid the unwanted attentions of predators and parasites. In Ch. 13, Ulmar Grafe considers how such compromises affect the design of acoustic signals. He also points out that in natural circumstances the precise timing of calls (e.g. whether calls are synchronized or alternated with the calls of neighbouring males) may be as important in determining a male’s reproductive success as the acoustic properties of the calls themselves.

Songbirds Bird song has long excited the interest and admiration of humans, a fact reflected in the large literature on songbirds. Song is a long-range advertising vocalization, so although males defending territories may be widely separated (in contrast to birds on leks, in flocks or at roosts) they can function as a network because of song. In many parts of the world, there is also a distinct dawn chorus, when most individuals of most species are singing at much the same time. Such characteristics explain why this section has two chapters on songbirds and may explain why studies of songbirds from a communication network perspective are becoming more common. Marc Naguib concentrates on vocal interactions between territorial (usually male) songbirds in Ch. 14, integrating such information with aspects of territorial behaviour such as settlement patterns. The spatial and social relationships that define territorial neighbours are mediated by vocal interactions and have implications for general spacing behaviour.

P1: JZZ/... P2: JZZ/... 0521823617agg.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

15:10

Communication networks in different taxa John Burt and Sandy Vehrencamp (Ch. 15) tackle the dawn chorus of songbirds. It is a striking acoustic phenomenon when song rate, singing diversity and song complexity reach their peak. However, the function of the dawn chorus is not readily explained by a single hypothesis. In Ch. 15, the dawn chorus is considered from a network perspective, which seems more likely to reveal its function.

Terrestrial mammals Terrestrial mammals are the second taxonomic grouping represented by two chapters in this section. This reflects the diversity of terrestrial mammals; for example, they span a size range from shrews to elephants, with obvious consequences for the feasibility of laboratory studies, and employ signal modalities such as scent and sound. Bob Johnston deals with scent communication in small terrestrial mammals in Ch. 16. Scent marks may contain information for several weeks and during that time the original scent mark may be over-marked by several individuals. As some species can determine the order of over-marking, this will create a sort of scent bulletin board. Scent marking cannot be detected unaided by human observers; this is in contrast to the loud calls of many large terrestrial mammals but similar to the infrasonic signals of very large mammals. Karen McComb and David Reby, in Ch. 17, consider the loud calls of large mammals that can and cannot be heard by humans. They also point out the implications of social organization for communication networks, particularly how the fluid fission–fusion nature of many large mammal groups is likely to increase opportunities for contact.

Marine mammals Far-carrying acoustic signals and fluid social systems are also characteristics of marine mammals. In Ch. 18, Vincent Janik summarizes the effect of these factors on communication networks of pinnipeds and cetaceans and explains how the nature of sound transmission in water means that sounds can potentially travel much further than in air. He also discusses whether the communication networks of marine mammals have been reduced in size in recent decades as oceans have become noisier (Andrew et al., 2002).

Humans Many of the terms used to discuss communication network behaviours (e.g. eavesdropping, audience effects) have their origin in our everyday human

249

P1: JZZ/... P2: JZZ/... 0521823617agg.xml CU1917B/McGregor

250

0 521 582361 7

April 11, 2005

15:10

Part III experiences. It is, therefore, a surprise to find that studies of human language generally consider dyads, in close parallel to other animal communication. John Locke in Ch. 19 argues that a network perspective is more realistic and that the laws forbidding eavesdropping found in some of the earliest known legal codes show that such behaviour has always been common. He also argues that information gained surreptitiously by eavesdropping may be particularly reliable and hence valuable, repaying the considerable effort and ingenuity often expended by human eavesdroppers.

Future directions There is consensus between the chapters in this section that a communication network perspective is an advance in understanding, but most of the chapters also point out that more information is necessary in order to evaluate fully the utility of this perspective by establishing the costs and benefits of communicating in a network. For this reason, it is premature to contemplate a formal comparative study of features of communication networks at the level of taxa represented by chapters in this section. However, it should be possible to attempt comparative analysis considerably sooner in species-rich groups with diverse features likely to affect networks (e.g. density and habitat) such as fiddler crabs and anurans. A key feature of any communication network addressed by most of the chapters in Part III is the extent of the network: how many individuals are encompassed by a signal? Theoretical estimates of maximum signal transmission distance combined with average separation distances between individuals are a very useful first approximation (for the role of perceptual abilities, see Ch. 20). However, it should be remembered that such maximum estimates of network size might be considerably larger than the actual size to which individuals respond. For example, in many anuran choruses, males adjust their call timing only to immediate neighbours. Similarly, it would be interesting to know the effect of fission–fusion societies on network size. Resolving such issues requires detailed study of actual networks, sometimes involving relatively new techniques (such as passive acoustic location) or features of signalling modalities that have been recognised as important relatively recently (e.g. scent over-marking). Several chapters raise the issue of the cognitive requirements for operating in a communication networks; for example, the extent to which an ability to identify individuals constrains social eavesdropping. Similarly, the nature of the information contained in signals will help to determine information flow through the network: from our human viewpoint, it would seem obvious that human language is a far richer source of information than the long-range advertising signals of other animals. Cognitive aspects of communication networks are also

P1: JZZ/... P2: JZZ/... 0521823617agg.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

15:10

Communication networks in different taxa dealt with by several chapters in Part IV, indicating that such questions are of interest to many.

References Andrew, R. K., Howe, B. M., Mercer, J. A. & Dzieciuch, M. A. 2002. Ocean ambient sound: comparing the 1960s with the 1990s for a receiver off the California coast. Acoustic Research Letters Online, 3, 65–70. Bonnet, X., Shine, R. & Lordais, O. 2002. Taxonomic chauvinism. Trends in Ecology and Evolution, 17, 1–3. Gerhardt, H. C. & Huber, F. 2002. Acoustic Communication in Insects and Anurans: Common Problems and Diverse Solutions. Chicago, IL: Chicago University Press. Greenfield, M. D. 2002. Signalers and Receivers: Mechanisms and Evolution of Arthropod Communication. Oxford: Oxford University Press.

251

P1: IYP/KAB P2: KPB-KOD 0521823617c12.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:36

12

Waving in a crowd: fiddler crabs signal in networks denise s. pope University of Copenhagen, Denmark

Introduction A communication network is formed when more than one receiver can intercept the signal produced by a signaller, and when more than one signal reaches a receiver at the same time (McGregor & Dabelsteen, 1996). Communication network theory broadens the consideration of selection pressures on signallers and receivers to include selection on signallers by receivers other than the primary or target receiver, and selection on receivers when they receive more than one signal simultaneously or intercept a signal that was not targeted at them (McGregor & Dabelsteen, 1996; McGregor & Peake, 2000). A mudflat full of male fiddler crabs (genus Uca, family Ocypodidae), all rhythmically waving their enlarged claw, seems a perfect example of a communication network: there are several signallers and receivers in close proximity, and many signals are being produced simultaneously. However, this human perception of a coordinated network may be partly a product of our excellent visual ability and large size in relation to these small crabs. What about the crabs themselves: how many receivers does a signal reach, and how many signals can individuals receive simultaneously? Is our impression that they form signalling networks simply an illusion caused by our extreme size and high visual acuity? Most importantly, what can we learn about the communication system of fiddler crabs by considering networks of signallers and receivers rather than simple sender–receiver dyads? In this chapter, I will first introduce the biology of fiddler crabs, then review the evidence that groups of displaying males of these species form communication networks; finally, I will examine the implications that such networks may have for our understanding of the fiddler crab communication system and suggest Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

252

P1: IYP/KAB P2: KPB-KOD 0521823617c12.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:36

Fiddler crabs signal in networks possible routes for future investigation. Although I am focusing on a single taxon, my hope is that these ideas may stimulate similar lines of investigation into comparable signalling systems in other taxa. In particular, much of the attention in communication network theory has, to date, focused on the phenomena of social eavesdropping and audience effects (see Chs. 2 and 4). Such social eavesdropping is unlikely in fiddler crabs, for reasons explained below, but in this review I aim to illustrate the utility of the network approach for identifying other consequences for signallers and receivers that are not predicted from the dyadic view of animal communication. In addition, much of the previous work on networks has focused on agonistic interactions between males, while in this chapter I focus on signals used by males to attract and court females.

The biology of fiddler crabs Fiddler crabs are small, deposit-feeding semiterrestrial crabs that inhabit protected shores worldwide in tropical and some warm temperate regions (Crane, 1975). There are 97 recognized species and subspecies in the genus (Rosenberg, 2001). Their intertidal and semiterrestrial existence governs their life in many ways: their activities are confined to low-tide periods and their lives are centred around their individual burrows, which they defend against conspecifics and which serve as shelters during tidal inundation and as refuges from heat, desiccation, and predation during low tide periods. Their reproductive lives are also constrained by an obligate pelagic larval stage. This larval stage often results in lunar or semilunar cycles of reproductive activity (Christy, 1978; Zucker, 1978; Greenspan, 1982; Yamaguchi, 2001a), as egg hatching and larval release are timed to coincide with optimal times for larval transport (Morgan & Christy, 1995). The timing of peak mating activity is, therefore, set by the timing of larval release and the duration of egg incubation (approximately two weeks depending on temperature and species; reviewed in Yamaguchi, 2001b). Among behavioural biologists, fiddler crabs are perhaps best known for their striking sexual dimorphism: males have highly asymmetrical claws, with the major claw greatly enlarged (up to five times in length) relative to both the male’s own minor claw and the female’s two symmetrical small claws (Rosenberg, 2002). Since the minor claw is used for scooping up the sediment for deposit feeding, females have the advantage of two feeding appendages while males have only one and hence have lower intake rates (Weissburg, 1992). The male’smajor claw is used primarily in two types of activity: fighting and signalling. Both males and females defend their burrows and the surrounding area from intruders, and the major claw is a very effective weapon in these disputes. Male–male aggression generally progresses through a series of stereotyped threat postures with the major claw to

253

P1: IYP/KAB P2: KPB-KOD 0521823617c12.xml CU1917B/McGregor

254

0 521 582361 7

April 11, 2005

17:36

D. S. Pope eventual pushing and grappling with interlocked claws, which sometimes results in one opponent being thrown (Crane, 1975; Hyatt & Salmon, 1977; Jennions & Backwell, 1996). When males are displaced from burrows they ‘wander’, either searching for an empty burrow or attempting to take over a burrow from another male (Crane, 1975; Jennions & Backwell, 1996; Backwell et al., 2000). Much of the variation in major claw morphology between species may be related to differences in fighting techniques (Crane, 1975). Fights between males for burrow ownership may be common because male burrows serve as a breeding resource in many species of fiddler crab (Christy, 1982; Backwell & Passmore, 1996). In what has sometimes been regarded as the ‘typical’ fiddler crab mating system, females leave their own burrows to ‘wander’ and sample courting males and their burrows, eventually choosing to stay in one burrow to mate with the male and lay her eggs; she then usually remains there for the duration of egg incubation until the larvae are released (Christy, 1983; Backwell & Passmore, 1996). A second mating system involves copulation on the surface close to the female’s own burrow (Crane, 1975; Salmon, 1984; Christy & Salmon, 1984); in this case, a male may defend his burrow as a base from which to court neighbouring females (Salmon, 1984). Recently, it has been recognized that many species exhibit both of these modes of mating in the same population; hence, they might best be thought of as alternative mating tactics (Koga et al., 1998; de Rivera & Vehrencamp, 2001; de Rivera et al., 2003). Several factors probably contribute to the opportunity for species to engage in surface mating in addition to, or instead of, burrow mating: small clutch size (Christy & Salmon, 1984; de Rivera & Vehrencamp, 2001), anatomical receptivity of females (if gonopore opercula are decalcified to allow copulation throughout the lunar cycle (Salmon, 1984)), predation level (Koga et al., 1998), density (de Rivera et al., 2003) and the spatial overlap between feeding areas where females burrow and breeding areas where males court (Christy, 1982, 1983), because in species where males and females generally do not inhabit adjacent burrows, the opportunity for surface mating is very limited. In addition to fighting, males use their major claw in a variety of movement signals and signalling postures, most conspicuously in the claw-waving display. This display is a species specific (Crane, 1975) and relatively stereotyped (Hyatt, 1977; Doherty, 1982) pattern of claw elevation (and in some species, unflexing), sometimes accompanied by movements of the minor claw, legs and body, and occasionally combined with a stereotyped pattern of locomotion (Crane, 1975). The display is performed only during the breeding season (Salmon, 1965; Crane, 1975; Wolfrath, 1993) and generally only by territorial, burrow-holding males. This context of the display led to suggestions that it may function to attract receptive females to the burrow, repel rival males from it or serve the dual

P1: IYP/KAB P2: KPB-KOD 0521823617c12.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:36

Fiddler crabs signal in networks function of signalling to both types of receiver (reviewed by Crane, 1975; Moriito & Wada, 2000). Recent experimental work has attempted to distinguish between these proposed functions of claw waving by identifying the target receiver of claw waving in several species of fiddler crabs (U. pugilator (Pope, 2000); U. beebei and U. terpsichores (formerly U. musica; Rosenberg, 2001; D. S. Pope, unpublished data); U. annulipes (P. R. Y. Backwell, unpublished data); U. tangeri (D. S. Pope & P. K. McGregor, unpublished data)) and one other species of ocypodid crab, Scopimera globosa (Moriito & Wada, 2000). The details of the experimental design differed between studies, but the general experimental approach involved isolating males either in cages or with temporary fences to control their visual environment, and then exposing them to different categories of potential receiver: neighbouring and/or introduced males, and neighbouring and/or introduced females. The results of these studies are best understood by first explaining that waving can be classified into at least two categories based on the intensity of the display, as has also been pointed out by other authors (von Hagen, 1962; Salmon, 1965; Crane, 1975; Doherty, 1982). High-intensity waving can be differentiated from lowintensity waving both by an increased rate and, in some species, by the addition or deletion of display components. In all six species studied, high-intensity waving was evoked only by the introduction of females, simulating the presence of a mate-searching female in the male’svicinity, strongly implying that high-intensity waving is directed exclusively to wandering females in these species. High-intensity waving is, therefore, clearly part of the courtship sequence in burrow-mating species. During times of peak mating, males in good condition generally wave at the low intensity, or background level, more or less continuously. von Hagen (1962) suggested that low-intensity waving functions to orient matesearching females towards the male from relatively long distances. When a male detects a wandering female near his burrow, he switches to high-intensity waving by increasing the wave rate and adding or subtracting display components. When females approach closely, males often switch to yet another courtship signal (e.g. rapping the major claw against the substrate in U. pugilator (Salmon, 1965) and the raised carpus display in U. beebei (Christy, 1988a)). The rate of waving also varies with temperature (Hyatt, 1977; Doherty, 1982) and male size (Hyatt, 1977; Jennions & Backwell, 1998). While high-intensity waving seems clearly directed to a particular receiver, lowintensity waving may function more as a broadcast signal, in the sense that it is not directed to any particular individual; however, as von Hagen (1962) suggested, it may be targeted at a general class of receiver: the mate-searching female. This function may imply that the presence of other individuals should have no effect on the ‘background’ level of low-intensity waving. The six species studied so far

255

P1: IYP/KAB P2: KPB-KOD 0521823617c12.xml CU1917B/McGregor

256

0 521 582361 7

April 11, 2005

17:36

D. S. Pope differed in the effect of neighbours’ presence on low-intensity waving: in other words, whether the rate of waving differed when males were completely visually isolated and when they could see their near neighbours. S. globosa, U. pugilator and U. annulipes showed no background, or low-intensity, waving in the experiments (Moriito & Wada, 2000; Pope, 2000; P. R. Y. Backwell, unpublished data); U. terpsichores waved at the same rate when alone as when surrounded by neighbours (D. S. Pope, unpublished data), and in both U. beebei and U. tangeri the presence of neighbours increased the wave rate above the background level when visually isolated (D. S. Pope, unpublished data; D. S. Pope & P. K. McGregor, unpublished data). The lack of background waving in the S. globosa and U. pugilator studies may have been experimental artefacts as the males in these cases were translocated to caged areas, so the natural level of background waving in those species remains to be clarified. The U. annulipes study also involved caged males, although in this case the males were not displaced; the fact that males waved only when females approached accords with observations of natural interactions in this species (Backwell et al., 1998). The reasons for lack of low-intensity or background waving in this species deserve further investigation. The different effects of the presence of neighbours on the waving rate in the remaining three species may be attributable to differences among the species in the spatial overlap between the sexes. In U. terpsichores, males and females are generally spatially segregated while in U. beebei and U. tangeri, males and females intermingle in the same microhabitat (personal observation); therefore, U. terpsichores males have only male neighbours while the neighbours of the other two species would include both males and females. This potential correspondence between the presence of female neighbours and an increased rate of low-intensity waving may indicate either that low-intensity waving is simply stimulated by the presence of females in the vicinity, whether they are neighbours or wandering females, or that female neighbours themselves are part of the target receivers of the low-intensity display. Another difference between U. terpsichores and the other two species is that the other species engage in surface mating in addition to burrow mating. In these species, female neighbours are potential mates, and if these females assess the quality of neighbouring males, low-intensity claw waving may be targeted at them (see discussion on p. 268). However, a comparative study of four Panamanian species that exhibit different combinations of mating tactics, including U. terpsichores and U. beebei (D. S. Pope, unpublished data) found no indication that males of any species faced female neighbours while waving, and in every species, males waved most often not facing any individual in the vicinity, suggesting that they may actively avoid facing neighbours while waving. If female neighbours do prove to be part of the target receivers of low-intensity waving, it is likely still more accurate to think of this level of waving as a broadcast display,

P1: IYP/KAB P2: KPB-KOD 0521823617c12.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:36

Fiddler crabs signal in networks directed not at specific individuals but rather targeted at females in general as potential receivers. Despite suggestions based on behavioural observation that claw waving appears to be directed sometimes to males as a territorial or threat display (reviewed by Moriito & Wada, 2000), the experimental evidence indicates that this is not the case in the six species studied so far. From current knowledge of fiddler crab mating systems, it appears that the majority of species engage in burrow mating, either alone or in combination with surface mating and other less-common tactics (Pope, 1998). All of the experimental investigations into the targeted receivers of the display involved species with burrow mating only or mixed tactics; therefore, in these cases, the importance of attracting females for burrow mating is clear. No studies have investigated the function of waving in species that copulate only on the surface, but several authors have noted that surface mating is not preceded immediately by waving (Salmon, 1984; Yamaguchi, 2001c), implying that it does not serve as a courtship signal in this context in the same sense that it does in burrow mating. The fact that the waving display has been retained in these species implies that it continues to serve some function as a communication signal, although phylogenetic comparative evidence suggests that the complexity of the display may be reduced in these species (Pope, 1998; de Rivera & Vehrencamp, 2001). Further investigations are warranted into the possibilities that neighbouring females assess males for surface mating by their waving display (see discussion below) or that waving is used more in male–male interactions in these species. In addition to claw waving, fiddler crabs have a rich repertoire of other signals, not all of which involve the major claw. Using the major claw held outstretched, males produce threat signals to other males (similar to the threat displays of many other brachyuran crabs (Wright, 1968)). By rapping the claw against the substrate or through stridulation of body parts, males also produce vibration signals (Salmon & Horch, 1972; Popper et al., 2001). These vibration signals are most commonly produced at night by species that are nocturnally active, although they may also play a role in the final sequence of diurnal courtship, as described above. The other major class of signals in fiddler crabs are structures constructed from the sediment (sand or mud), generally in close proximity to the signaller’s burrow. Courting males of 16 species build structures that range from the small semidomes of U. pugilator to the elaborate pillars of U. beebei and hoods of U. terpsichores (reviewed by Christy, 1988b; Christy et al., 2002). Hoods and pillars increase the attractiveness of the males that build them to mate-searching females (Christy, 1988a; Christy et al., 2002). While I will focus on claw waving in this review, some of the conclusions may also be applicable to the other potentially long-range signals of fiddler crabs, specifically vibration signals and structures such as pillars and hoods.

257

P1: IYP/KAB P2: KPB-KOD 0521823617c12.xml CU1917B/McGregor

258

0 521 582361 7

April 11, 2005

17:36

D. S. Pope Much of the preceding descriptions of the crabs’ biology, as well as many of the suggestions to come in this chapter, may also apply to other members of the family Ocypodidae. There are five subfamilies (Kitaura et al., 1998) and fiddler crabs belong to the subfamily Ocypodinae, along with their closest relatives, the ghost crabs (genus Ocypode). This close relationship with ghost crabs can be misleading as ghost crabs are unusual within the family (many ghost crabs are specialized predators; many are nocturnal; they often inhabit exposed shorelines; and they are, on average, larger than other members of the family), and many of the other species more closely resemble fiddler crabs in ecology and behaviour than do ghost crabs. Crabs in the genera Macrophthalmus, Ilyoplax and Scopimera, in particular, resemble fiddler crabs in that they are small, deposit-feeding crabs inhabiting protected shores, which show a rich repertoire of signalling behaviour including claw waving (e.g. Wada, 1991; Kosuge et al., 1994; Moriito & Wada, 1997) and structure building (e.g. Wada, 1994, Kitaura et al., 1998). They also show both surface- and burrow-mating tactics, with many species exhibiting both tactics (e.g. Wada, 1984; Henmi et al., 1993).

Do fiddler crabs signal in networks? All signals produced by fiddler crabs have four classes of potential conspecific adult receiver: wandering females, burrow-holding females, wandering males and burrow-holding males. At least in species with substantial overlap in microhabitat use by males and females, all four classes of individual are potentially within receiving range of the display at any given time, creating the conditions necessary for the formation of communication networks: that is, the active space of a signal exceeds the average spacing between individuals (McGregor, 1993; McGregor & Dabelsteen, 1996). What is the evidence that these conditions are met within fiddler crabs? Interindividual spacing

Fiddler crab colonies are often described as dense, but this is of course from the human observer’s point of view. In fact, average densities vary both between and within species. Larger species are in general less densely distributed than smaller species (de Rivera & Vehrencamp, 2001). The density of fiddler crab colonies can be estimated by counting surface-active individuals, by counting the number of open burrows or by excavating the sediment and counting all crabs uncovered in a given area; the last gives the best estimate of true population density (Macia et al., 2001). Counts of surfacing individuals underestimate the true population density in U. annulipes while burrow counts tend to overestimate the number of excavated crabs because some burrows are empty (Macia et al., 2001). However,

P1: IYP/KAB P2: KPB-KOD 0521823617c12.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:36

Fiddler crabs signal in networks when considering the social environment available to waving male crabs, the density of individuals on the surface at any given time is the best estimate of the potential density of interactants in the putative communication network. We can get an estimate of the range of interindividual distances by considering two well-studied species at opposite ends of the size spectrum: U. beebei and U. tangeri (with average adult male body sizes, measured as carapace width or the distance between the outer edge of the eye sockets, of approximately 0.9 cm and 3.0 cm, respectively; D. S. Pope, unpublished data). Density measures can be transformed into estimates of interindividual distances by taking the reciprocal of the density to yield the average space per individual; if it is then assumed that the area is circular, the radius of that circle can be calculated from the area. Multiplying the radius by two gives the average distance between two individuals. U. beebei is found at high densities of 49 active individuals per m2 on average (D. S. Pope, unpublished data), which translates into 16 cm between individuals. The larger species U. tangeri is more widely spaced, at an average of 4.6 active individuals per m2 , giving an average interindividual distance of 53 cm (D. S. Pope, unpublished data). Each of these densities was estimated from areas of high crab activity, but crab distribution is patchy within the colonies of both species (U. beebei (de Rivera et al., 2003), U. tangeri (D. S. Pope, unpublished data)), so there will be areas with larger spacing between individuals. These estimates may, therefore, be regarded as the optimal conditions for communication networks in these species. In addition, other species are found at lower densities. For example, U. terpsichores is in the same size class as U. beebei but is often found in sandier areas with densities of 17.5 individuals per m2 , or 27 cm between neighbours. U. stylifera is a larger crab (approximately 2.4 cm body size (Crane, 1975)) and is found at 2.6 individuals per m2 (70 cm spacing) at high local densities (D. S. Pope, unpublished data). These estimated interindividual distances represent the average spacing of burrow-holding individuals, but the distance of a wandering individual from conspecifics will be approximately half this distance, assuming that a wanderer maintains an equal distance between conspecific burrows. Therefore, wandering males or wandering mate-searching females of U. beebei would be, on average, 8 cm from the closest burrow-holding conspecifics and U. tangeri wanderers would be, on average, 26.5 cm from the closest burrow-holding conspecifics. Detection distance of conspecifics

Given that we know the average spacing of individuals, we now need to know at what distance fiddler crabs are likely to perceive a waving male conspecific. On an absolute level, the vision of fiddler crabs is constrained by the size of the ommatidia in the crabs’ eyes, as its size determines the resolution of the eye. The ability of the eye to resolve an object depends on both the angular size

259

P1: IYP/KAB P2: KPB-KOD 0521823617c12.xml CU1917B/McGregor

260

0 521 582361 7

April 11, 2005

17:36

D. S. Pope of the object and the object’s contrast with the background. Therefore, a conservative estimate of the limit of detectability is that an object with an angular size less than the size of a single ommatidium is unlikely to be resolved from the background unless it contrasts strongly with it (Land & Nilsson, 2001; Zeil & Hofmann, 2001). However, ommatidial size varies throughout the eye. Fiddler crabs have eyes well adapted to their primarily flat visual environment: like all ‘flat world’ crabs, they have a band of high vertical resolution around the horizon of the eye, which they align with the visual horizon (Zeil et al., 1986; Land & Layne, 1995a; Zeil & Al-Mutairi, 1996). Since fiddler crabs carry their eyes on long stalks, objects below the height of their eyes, including the bodies of most conspecifics, will be seen below the horizon line. Several authors have argued that this visual horizon can allow crabs easily to categorize stimuli into either ‘conspecific’ or ‘predator’, as predators, being larger than the crab, would be imaged above the horizon line (Land & Layne, 1995a; Layne, 1998). In this high-resolution zone close to the horizon, the theoretical resolution threshold is approximately 0.5–1◦ . Given this resolution and assuming an eye height of 2.5 cm for crabs of 1.0 cm average body size, a 1 cm conspecific should be easily resolvable at 57 cm; this distance doubles to 114 cm for a 2 cm conspecific ( J. Zeil, personal communication). Larger species, such as U. tangeri, should theoretically be able to detect conspecifics at greater distances, both because of the larger stimulus size and because of the increased height of the eyes above the ground (estimated to be 4 cm in U. tangeri (D. S. Pope, unpublished data)), which expands the range over which distance judgements can be made based on retinal elevation ( J. Zeil, personal communication). Therefore, the likely detection distance for conspecifics exceeds the average interindividual spacing in both large and small species of fiddler crab. Unfortunately, we are technically constrained in our measurements of what a crab can see by what its behaviour tells us. In other words, a field measurement of the distance at which a crab reacts to a specific stimulus is a combined measure of both the crab’s ability to detect it and the relevance of the stimulus. A given stimulus may well be detectable at much greater distances but it only elicits a response once it enters a specific zone around the individual’s burrow. Given these caveats, what reaction distances have been reported for fiddler crabs? Reaction distances to conspecifics and conspecific-sized stimuli have been tested in only a few species. Land & Layne (1995a) reported 30 cm as the distance at which courting male U. pugilator (large male carapace width 1.6 cm (de Rivera & Vehrencamp, 2001)) seemed to notice a wandering conspecific, based on an increase in wave rate. They reported 10–15 cm as the distance at which males apparently discriminated the sex of an approaching individual (presumably by the presence or absence of a major claw), based on the approach distance at which males switched from waving to threat behaviour to males. The 30 cm distance corresponds to an angular

P1: IYP/KAB P2: KPB-KOD 0521823617c12.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:36

Fiddler crabs signal in networks size of two to three interommatidial angles, which is two to three times larger than the theoretical detection threshold. These distances should be interpreted as minimum estimates of detection distance, however, because they are probably modulated by the size of a male’s territory, as illustrated by the fact that these authors also measured responses to predator-like stimuli above the horizon (to which the motivation to respond is presumably higher) at much smaller angular sizes. Hemmi & Zeil (2003) performed experiments on burrow surveillance by male U. vomeris by moving crab-sized dummies (2.25 cm wide and 1.2 cm high) across the surface towards the male’s burrow and measuring the distance at which the crabs responded by returning to their burrow to defend it. Male response was better predicted by the distance of the dummy to the burrow than by the distance of the dummy to the crab: they consistently responded when the dummy was an average of 23.8 cm from the burrow, suggesting that this distance might represent the radius of their defended territory. However, the distance between the dummy and the crab when the crab reacted varied greatly, since this depended on the male’s distance to his burrow and the orientation of approach of the dummy towards the burrow. Crabs responded to dummies at distances of up to 80 cm ( J. Zeil, personal communication), indicating that these conspecific-sized objects were clearly resolvable at that distance. Finally, in U. tangeri (3 cm, as above), response distances to wandering females have been estimated as 150–200 cm (von Hagen, 1962). In a preliminary experiment on the same species, I attempted to determine a maximum reaction distance by controlling a male’s visual environment with fences and then providing him with a stimulus to which he should be highly motivated to respond, a tethered but realistically moving female. In these preliminary measurements, I estimated a reaction threshold distance of 100–150 cm. More field experiments and observations, in addition to neurobiological work on crab vision, will help us to understand better how fiddler crabs process and respond to their visual environment. All of these measured reaction distances were reactions to non-waving conspecifics, or conspecific-sized objects. There are several reasons to suspect that the claw-waving display makes a male more detectable to conspecifics than a motionless male. First, at adult male body sizes, the claw length exceeds the carapace width of the male (Crane, 1975), increasing the size of the visual stimulus. Second, the claw itself is probably the most detectable part of the male: in displaying males it is often bleached white to contrast strongly with the body (Crane, 1975) and it is the part of the body that contrasts most strongly with the substrate in terms of spectral reflectance and polarization (Zeil & Hofmann, 2001). There is, as yet, no definitive evidence that fiddler crabs have colour vision, but recent findings are suggestive of a dual-pigment system (Horch et al., 2002), and the smooth wet cuticle of fiddler crabs generate ample specular and ultraviolet reflectance contrasts

261

P1: IYP/KAB P2: KPB-KOD 0521823617c12.xml CU1917B/McGregor

262

0 521 582361 7

April 11, 2005

17:36

D. S. Pope with the mudflat background (Zeil & Hofmann, 2001). The movement of the claw provides motion signatures that should also make the crab more detectable (Zeil & Zanker, 1997). In addition, during claw waving, the claw itself is elevated above the level of the crabs’ eyestalks and crosses the visual horizon of viewing conspecifics, hence entering the zone above the horizon that is used to detect predators (Land & Layne, 1995a; Zeil & Zanker, 1997). Some authors have suggested that claw waving thus taps into the predator-escape responses of conspecifics (e.g. Land & Layne, 1995a). Clearly, even if the waving display did initially exploit the female’sreceiver biases for detection of moving objects in the zone above the horizon, selection has since modified the response from a generalized negative or inhibitory response towards threatening stimuli to a more specialized positive or attractive response towards preferred male conspecifics (see discussion of exploitation of antipredator receiver biases in Greenfield (2002)). In summary, it is likely that several aspects of signal design work together to increase the detectability of displaying male fiddler crabs to conspecifics, although we do not yet know the absolute ‘signal space’ of a male’s claw-waving display. Inferences and assumptions

The evidence reviewed above suggests that both the distances at which crabs can theoretically detect conspecific-sized objects and the distances at which they have been shown to react to conspecifics exceed the average spacing between individuals, especially the distances between a wandering individual and the closest burrow owner, thus setting the stage for potential communication networks to exist, at least in species or populations with relatively high densities of individuals. Clearly more research is needed on the perceptual ability of crabs, including potential differences between species resulting from factors such as phylogeny (fiddler crabs are traditionally divided into broad-front and narrow-front species, based on the space between their eyes and hence a difference in relative eyestalk length between species – what effect does this have on their vision and behaviour?), crab size (does ommatidial size and number scale with body size?) and visual environment (do mangrove-dwelling fiddler crabs have any special adaptations to deal with their more complex visual environment?). All of these considerations are likely to mean that the extent of communication networks will vary both between and within species of fiddler crabs, and this variation could be a fruitful avenue for further research.

Implications of network signalling for fiddler crab communication I will now examine the consequences of assuming that fiddler crabs signal in networks. What are the possible behavioural effects this will have on how fiddler

P1: IYP/KAB P2: KPB-KOD 0521823617c12.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:36

Fiddler crabs signal in networks crabs produce and respond to signals, as distinct from what would be predicted from a traditional ‘dyadic’ signalling scenario? Production of signals by fiddler crabs in a network: strategies for signal competition

In a communication network, signallers are faced with a more complex problem than in a simple sender–receiver dyad in the sense that they must necessarily compete with other signallers for the attention of the targeted receiver. This is self-evident in many cases of sexual signalling, where males are competing to attract females. In most situations in fiddler crabs, as in many chorusing species of insects and anurans, females will be in a position to receive more than one signal simultaneously, so the environment is by default a network (Ch. 2). Such signal competition probably has effects on at least two timescales: in terms of gross signal timing (whether to signal or not), and in terms of the fine-scale patterns of signal timing among neighbouring males. I will term these two timescales bout timing and signal timing, respectively (Gerhardt & Huber, 2002; Greenfield, 2002). Bout timing

At the level of bout timing, males undoubtedly use other males as cues for when to start signalling. Presumably because of energetic constraints, males do not constantly wave their claws, so they should time their signalling to coincide with the maximum likelihood of attracting receptive females, assuming that, at least in burrow-mating species, the display functions primarily in attracting females to the male’s burrow and persuading them to mate. As noted above, male fiddler crabs track the timing of receptive females on the scale of the breeding season (Salmon, 1965; Crane, 1975; Wolfrath, 1993), the lunar cycle (Christy, 1978; Zucker, 1978) and also the daily cycle (Christy et al., 2001). In addition, males are probably selected to signal whenever their neighbours are signalling lest they miss mating opportunities; hence males are probably likely to begin signalling if another male does so. There is some evidence that males do respond this way in U. pugilator, both to waving neighbours and acoustically signalling neighbours (Salmon, 1965; Pope, 1998). This effect may also be inferred in U. beebei and U. tangeri, as males of these species show an increase in waving rate with an increasing number of male neighbours (D. S. Pope, unpublished data). There is good evidence that males use other males as cues to begin signalling bouts in acoustic insects (Greenfield, 2002) and so this should be a fruitful avenue for research in fiddler crabs as well. Signal timing

Choruses of acoustically signalling anurans and insects often exhibit group coordination of signals such as synchronous or alternating calling, which must be accomplished through fine-scale timing adjustments of individual males

263

P1: IYP/KAB P2: KPB-KOD 0521823617c12.xml CU1917B/McGregor

264

0 521 582361 7

April 11, 2005

17:36

D. S. Pope to the calls of other calling males (Gerhardt & Huber, 2002; Greenfield, 2002; Ch. 13). These chorusing interactions are thought to be epiphenomena resulting from the preference of females for leading signals (Greenfield, 1994, 2002). Recent experiments using acoustic playback have revealed in many anuran and insect species that if two signals overlap in time but are otherwise equal, females will prefer the leading of the two signals (reviewed by Gerhardt & Huber, 2002; Greenfield, 2002). Such a preference is thought to result from psychoacoustic constraints such as masking (Gerhardt & Huber, 2002) or the precedence effect (a phenomenon by which two signals, overlapping in time, are perceived as a single acoustic object by the receiver (Greenfield et al., 1997; Greenfield, 2002)). Males have thus been selected to avoid being the following male and hence have evolved timing mechanisms by which they delay their calling if a rival male calls within a certain critical interval following the male’s own call (Greenfield et al., 1997). As a consequence, males subtly adjust their timing in an effort always to be the first to call of a pair of calling males. Such mutual adjustment leads to either synchrony or call alternation, depending on the call timing of the species. There is no evidence that either synchronous or alternating patterns of calling are cooperative in the sense that there is any benefit owing to greater attraction of females per capita by grouped versus solitary signallers (Gerhardt & Huber, 2002; Greenfield, 2002); hence signal competition is the most parsimonious explanation for these chorusing phenomena. In a few species of fiddler crabs and a related ocypodid, there is some very good evidence that males adjust the timing of their signals in relation to their neighbours, resulting in the production of synchronous signals: U. annulipes (Gordon, 1958; Backwell et al., 1998, 1999), U. perplexa and U. saltitanta (P. R. Y. Backwell, M. D. Jennions, K. Wada, M. Murai & J. H. Christy, unpublished data), and Ilyoplax pusilla (Aizawa, 1998). The phenomenon of synchronous waving has been most thoroughly documented in U. annulipes (Backwell et al., 1998, 1999), an Indo-West Pacific broad-fronted species. U. annulipes is unusual in that males apparently do not produce low-intensity, or background, waves but only wave in the presence of a female (Backwell et al., 1998). When a mate-searching female approaches, males cluster around her, and males within this cluster synchronize their waves with each other. As in most synchronously calling insects and anurans, Backwell et al. (1998) also found a female preference for the leading male of a group of synchronous wavers. These males also signal at a faster rate than their neighbours (Backwell et al., 1999), producing some waves that are not overlapped by other males. It is not clear, therefore, whether the female preference for the leading male is the result of a perceptual constraint as it appears to be in insects and anurans, or whether it is simply a consequence of females preferring males that signal at the fastest rate, which is condition-dependent in this species (Jennions & Backwell, 1998). The fact that

P1: IYP/KAB P2: KPB-KOD 0521823617c12.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:36

Fiddler crabs signal in networks synchronous waving results, however, would imply that perhaps the overriding preference is for the leader (Backwell et al., 1999), and the fastest waving male is simply the one who least needs to adjust his signal to those of his neighbours (Greenfield, 2002). In both U. perplexa and U. saltitanta, males produce synchronous waves in both the presence and absence of mate-searching females and increase their wave rate in the presence of females (P. R. Y. Backwell, M. D. Jennions, K. Wada, M. Murai & J. H. Christy, unpublished data), as expected if high intensity waving serves to court females and persuade them to mate, as discussed above. The waves of U. perplexa become less synchronous in the presence of females, perhaps because the increased rate of high-intensity waving means that some males are not able to produce waves at such a high rate and fall out of synchrony with their neighbours (P. R. Y. Backwell, M. D. Jennions, K. Wada, M. Murai & J. H. Christy, unpublished data). In these two species, females also prefer the male waving at the fastest rate, but only in U. perplexa is that male also most often the leading male. This would suggest that the overwhelming preference is for high-quality males, waving at the fastest rate. The fact that U. perplexa and U. saltitanta continue to wave in synchrony even when mate-searching females are not nearby suggests other aspects of signal competition: perhaps it results from a female preference for males with the highest wave rate even at a distance, a preference for groups of synchronous wavers, or because such signal competition allows males to compete better with their male neighbours in other contexts such as territory acquisition and maintenance (P. R. Y. Backwell, M. D. Jennions, K. Wada, M. Murai & J. H. Christy, unpublished data). The results of these studies clearly demonstrate that the network phenomenon of signal competition for receivers’ responses, resulting in signal synchrony, occurs in these species. If a high rate of waving indicates male quality in other species of fiddler crabs, females may be expected to prefer males that wave at the highest rate in these species as well, which leads to the question of why signal competition resulting in synchronous signalling is not more widespread in fiddler crabs. Future research should be directed at investigating what factors may have promoted the evolution synchronous waving in some species of Uca and not others. While synchronous waving has been observed in a few other Uca species (P. R. Y. Backwell, personal communication), bouts of synchrony do not appear to be as sustained or as tightly timed as they are in U. annulipes, U. perplexa and U. saltitanta. In the vast majority of species (i.e. those that do not signal synchronously), it is not clear whether males adjust the timing of their signals in relation to each other, other than at the gross level of initiating bouts when neighbours do. However, results from the ocypodid I. pusilla suggest that males may adjust their signal timing in more subtle ways. Aizawa (1998) found that males delayed the timing of their waves so that they overlapped with both live and videotaped male neighbours in a laboratory setting, agreeing with the observations that neighbouring males

265

P1: IYP/KAB P2: KPB-KOD 0521823617c12.xml CU1917B/McGregor

266

0 521 582361 7

April 11, 2005

17:36

D. S. Pope of this species signal synchronously. More surprisingly, field measurements also demonstrated that male I. pusilla adjusted the timing of their signals to match that of the much larger fiddler crab U. lactea, whereas there was no evidence that U. lactea adjusted their timing reciprocally (Aizawa, 2000). These results suggest that males of other species of fiddler crab (and other waving ocypodids) may make finescale adjustments to their signal timing in ways that are not immediately obvious to the naked eye and will only be uncovered by careful analysis. Burford et al. (1998) found some suggestion of an adjustment of wave rate by large males of U. tangeri when signalling in the presence of smaller neighbours, but the comparisons were made only at the level of wave rate, and not the males’ timing of signal initiation in relation to each other’s signals. The considerations above lead to the conclusion that males in many fiddler crab species are easily able to perceive the signals of at least their immediate neighbours, implying that such timing adjustments are likely. Reception of signals by fiddler crabs in a network: strategies for information gathering

Some of the most intriguing possible consequences of fiddler crab communication networks relate to how receivers may use signals to gather information in ways not traditionally considered in a dyadic framework. This section is somewhat speculative as, to date, there is little evidence for these effects. However, given the clear existence of network effects on the production of signals by male fiddler crabs (reviewed above), the network environment is likely to have consequences for receivers as well. As outlined in the introduction, the experimental evidence from the five fiddler crab species so far studied points to wandering, receptive females as the primary receivers to which claw waving is directed. In this section, I will work under the assumption that wandering females are the target receivers. As all of these five species engage in burrow mating, and since less research to date has been focused on species that only mate on the surface, the discussion is biased towards species that mate in burrows. Our understanding of communication networks in fiddler crabs will benefit from more in-depth study into the function and use of the clawwaving display and other signals in surface-mating species, and such study will provide a useful test of the generality of the following inferences about fiddler crab networks. While I assume here that wandering females are the primary receivers of the claw-waving display, I reiterate that by ‘receivers’ here I mean not only the ‘intended’ receivers, or primary targets, of the signal but also other individuals that may be paying attention to the signal. As described above, in most species there are at least two other classes of adult conspecifics within receiving range of the

P1: IYP/KAB P2: KPB-KOD 0521823617c12.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:36

Fiddler crabs signal in networks display: neighbouring males and wandering males. In addition, in species with extensive spatial overlap between males and females, neighbouring females will also be within receiving range. Such potential receivers may be considered ‘conspecific interceptive eavesdroppers’,that is, conspecifics that benefit from intercepting signals targeted at another individual (see Ch. 2). In addition, other receivers such as bird predators, juvenile conspecifics and heterospecific crabs may also benefit from intercepting the signals of males (by better locating or avoiding waving males, depending on the context), but I will not consider these additional receivers further. I will consider each of the four categories of adult conspecific receivers in separate sections. Wandering females: the target receivers

Females may be attracted to groups of signalling males because the group as a whole is more detectable or because it provides the female with an enhanced opportunity to compare mates. Such opportunities might benefit a female because they reduce search time (presumably also reducing search risk) and perhaps result in a higher overall mate quality if signalling in some way allows her to assess male quality. There is mixed evidence from acoustically signalling insects and anurans that females are differentially attracted to groups of signalling males. Studies testing whether choruses themselves attract females better than lone signallers have found primarily negative results (reviewed by Gerhardt & Huber, 2002; Greenfield, 2002). Yet, there is fairly good evidence that females may be attracted to some groups of males over others within choruses, based on the size or density of the group (reviewed by Gerhardt & Huber, 2002). As resources in fiddler crabs are not as clumped as in these acoustic species, the situation in fiddler crabs may be more analogous to the second case, of females moving within choruses; as such, the potential for similar effects exist: females may be more attracted to areas with a higher density of signalling males. In the eastern Pacific species U. beebei, higher density increased the likelihood of mate searching by females in experimentally manipulated areas, and wandering females more frequently entered areas of naturally higher density than low-density areas (de Rivera et al., 2003). If females are directly attracted to higher densities of waving males, this can be tested by experimentally offering females a choice of patches of males that differ in density of signallers. Factors affecting the attractiveness of claw waving to females have been best studied in synchronously waving species (Backwell et al., 1998, 1999; P. R. Y. Backwell, M. D. Jennions, K. Wada, M. Murai & J. H. Christy, unpublished data). Further work is needed on other species to determine what aspects of waving are most attractive to females and how these factors correlate with male quality and resource (burrow) quality. More information on what attracts females to claw waving

267

P1: IYP/KAB P2: KPB-KOD 0521823617c12.xml CU1917B/McGregor

268

0 521 582361 7

April 11, 2005

17:36

D. S. Pope may provide insight into the potential exploitation of receiver biases and how males may use waving to enhance their detectability and attractiveness to females, as Christy and colleagues have done for male courtship structures (Christy et al., 2001, 2002). Neighbouring females

As reviewed above, there is as yet no evidence that claw waving functions as a direct prelude to surface copulation. However, in species that have both mixedsex colonies and either mixed mating tactics (both surface and burrow mating) or surface mating only, the possibility exists that females may assess and compare their male neighbours (who are their most likely surface copulation partners (Yamaguchi, 2001c)) for potential future copulations. In an unpublished study of four Uca species in Panama, I found no evidence that males directed their displays to female neighbours, but this does not preclude the possibility that females may still use the displays to gain information. A recent study by Murai et al. (2002) found that in U. paradussimieri, a species with a unique mating tactic in which males enter female’s burrows for mating, territorial males showed evidence of being able to assess the reproductive state of their female neighbours and directed non-waving courtship at ones that were close to being receptive, mating with them up to three days later. Therefore, if males are able to assess and integrate such information over such time periods, females may in a parallel fashion be able to assess and compare male neighbours and integrate such information into their decisions about surface-mating partners. The potential for longer-term assessment by neighbouring females in species that surface mate only should be investigated in concert with experimental manipulations to assess the target receivers of waving in these species; this would improve our understanding of how waving functions in these species. Such knowledge will also help us to assess how waving has evolved across the genus and family as a whole. Wandering males

The traditional view of claw waving as a dual function territorial signal holds that claw waving would repel a wandering male from approaching the burrow of a waving male, because the display would in this case signal the male’s ownership of the burrow and his willingness to defend it. Given that there is no evidence that males address their waving to wandering males, and the fact that they switch to distinct threat displays when another male intrudes (Salmon & Stout, 1962; Land & Layne, 1995a; D. S. Pope & P. K. McGregor, unpublished data), claw waving may not have this function in fiddler crabs. Conversely, the network view suggests that the display might have the opposite effect: that of attracting a wandering male to the vicinity. If claw waving is directed at females, wandering males might use it to gather information about potential areas to establish a new

P1: IYP/KAB P2: KPB-KOD 0521823617c12.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:36

Fiddler crabs signal in networks burrow or attempt a take over of an occupied one. If the goal of burrow ownership is to attract females for mating (over and above the necessity of a burrow as a shelter and refuge), then areas of high waving activity might indicate to the male an area of high likelihood of mate attraction. This effect may be magnified in the case of high-intensity waving, as it indicates the actual presence of a female. A wandering male’s response to a waving male may be modulated by the wanderers’ size relative to the resident, since larger crabs usually win fights, although there is also a resident advantage ( Jennions & Backwell, 1996). Thus, waving may actually repel smaller males, whereas equally sized or larger males may be attracted by it. Male attraction to rival males’ advertisement signals has been suggested in a few taxa. Stamps (1988) found that juvenile Anolis aeneus lizards showed ‘conspecific attraction’ when settling in new habitat and suggested that they might use the territorial advertisement displays of head-bobbing to assess conspecific density. Alatalo et al. (1982) demonstrated that broadcasting the song of the pied flycatcher Ficedula hypoleuca attracted settlement by conspecifics in nearby nest boxes. Playback of male advertisement calls in laboratory phonotaxis experiments attracted conspecific males in the spadefoot toad Spea multiplicata (Pfennig et al., 2000) and the house cricket Acheta domesticus (Kiflawi & Gray, 2000). In both of these cases, smaller males in particular differentially approached calls that were most attractive to females. Such behaviour can result in male aggregations that are independent of resource distribution, and it may also be a prerequisite for satellite male-calling behaviour (reviewed by Gerhardt & Huber, 2002). Conspecific interceptive eavesdropping of this type, involving male attraction to signals of their rivals, may be much more common than is widely recognized, and the dearth of examples may simply result from an absence of studies. Observational studies of the movements of wandering males and experimental studies testing whether males respond aversively or positively to other males’ signals would help to clarify the existence of this phenomenon in fiddler crabs. Neighbouring males

One final possible consequence of fiddler crab communication networks is another form of interceptive eavesdropping, in this case by neighbouring territorial males. The fact that males are more likely to switch to high-intensity waving when wandering females approach (D. S. Pope & P. K. McGregor, unpublished data) suggests the possibility that other males could use their neighbours as ‘female detectors’. By monitoring the signalling of neighbouring males, in particular noting when they switch to high-intensity waving, they might effectively expand the distance at which they can detect a wandering female. This monitoring would presumably need to be upregulated by the actual detection of the female herself, so that males do not end up wasting energy unnecessarily. The number of

269

P1: IYP/KAB P2: KPB-KOD 0521823617c12.xml CU1917B/McGregor

270

0 521 582361 7

April 11, 2005

17:36

D. S. Pope neighbours that are high-intensity waving may also have a synergistic effect, such that a male is more likely to pay attention if more than one neighbour switches to high intensity, because this is a more reliable indicator of female presence. Recent experiments on U. tangeri suggest that this effect may indeed be occurring (D. S. Pope & P. K. McGregor, unpublished data). The high-intensity waving display of this species is easily differentiated from the low-intensity display by the addition of an introductory curtsey, and males reliably switch to high-intensity waving when a female approaches (D. S. Pope & P. K. McGregor, unpublished data). The visual environment of neighbouring pairs of males was controlled with an opaque fence so that the males could view each other but females could be introduced on one side of the fence in such a way that one male and not the other could see them. Males that could view the introduced females significantly increased their rate of high-intensity waving, as expected. Males that could see their neighbour, but not the introduced female, significantly increased their rate of low-intensity waving; the rate of high intensity waving also increased, but not significantly. Thus these ‘interceptor’ males showed an intermediate level of waving between the lower background level and the higher level in the presence of a female. These results suggest that males do attend to the waves of their neighbours and use the information to compete to attract females, but they are more responsive to the actual detection of the female nearby. A similar result was found in the synchronously waving species U. annulipes (M. D. Jennions, unpublished data): males began to wave in synchrony with a neighbouring male even when their view of the female was blocked with a fence. This ‘female detector’ effect might be common in fiddler crab species but may not normally be noted by a human observer because it is often not clear whether a male is responding to seeing the female herself or to his neighbour. The possibility of males using rival males’ signals as ‘female detectors’ may be extended to other taxa as well. Such interceptive eavesdropping by rival males is a type of socially acquired information (Giraldeau et al., 2002) in which animals collect information from other conspecifics on resource location and quality. When the probability of acting on socially acquired information goes up when the same information is obtained from more than one individual, information cascades can result in which many individuals make the same behavioural decision without obtaining direct information themselves (Giraldeau et al., 2002; Watts, 2002). Such information cascades can result in information being transmitted faster than the rate of direct information acquisition, for example when escape responses are transmitted throughout a group faster than the approach of a model predator (Treherne & Foster, 1981), which the authors termed the Trafalgar effect. These information cascades can sometimes lead to suboptimal behaviour if the behavioural decisions of the initiators of the cascades were erroneous (e.g. escape response to a sudden movement of vegetation rather than a predator (Giraldeau

P1: IYP/KAB P2: KPB-KOD 0521823617c12.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:36

Fiddler crabs signal in networks et al., 2002)). The experiments described above on U. tangeri and U. annulipes involved only exposure to the behavioural decision (in this case, high-intensity waving) of a single conspecific; it remains to be tested if the effect is stronger if males witness the waving of more than one individual, as would be predicted if the situation is a true information cascade (Giraldeau et al., 2002). The potential also exists that males may track the movement of a female through a group of males by monitoring the signals of other males, as suggested by McGregor & Dabelsteen (1996) for territorial intrusions in songbirds.

Summary and future directions The ecology of fiddler crabs (typically high densities and open habitat) as well as their sensory physiology (good visual resolution) argue that most fiddler crab colonies operate as communication networks. The fact that their most common and conspicuous signal, the claw-waving display, does not appear to be used in male–male interactions makes it unlikely that the network effect of social eavesdropping (i.e. eavesdropping on signalling interactions; Ch. 2) occurs in these crabs. However, many other consequences of communication networks suggest themselves, particularly involving competitive signalling interactions and novel information-gathering strategies, including forms of conspecific interceptive eavesdropping. Such possibilities should provide fruitful avenues for future investigation. The best-studied cases of interceptive eavesdropping, which involve interception by predators or parasites, clearly harm the signallers, but the situation is less clear for conspecific interceptive eavesdropping (reviewed in Ch. 2). In fiddler crabs, males are likely to benefit from any interception of their signals by females, whether or not a given individual female was the target of the signal. Even in cases where a male’s signal compares unfavourably with that of a neighbour because it is given at a lower rate or is not a leading signal in a synchronous species, males are still likely to benefit more from waving than from not signalling at all. However, males may be more likely to suffer costs, in terms of missed mating opportunities, from the interception of their signals by other males. Are there any strategies that males can use to minimize these potential costs? The broadcast nature of the claw-waving display (at least at low-intensity, or background, levels) make it unlikely that males can eliminate the potential for eavesdropping (cf. in songbirds; Ch. 3); however, they may be able to target specific individuals when signalling by orienting either their front or back sides towards the approaching individual (female). This would present the largest visual stimulus to the female while reducing the view for neighbouring males situated off-axis to the signalling male. There is some suggestion that males orient in this way towards approaching females in U. pugilator (Land & Layne, 1995b), although such orientation was not in evidence

271

P1: IYP/KAB P2: KPB-KOD 0521823617c12.xml CU1917B/McGregor

272

0 521 582361 7

April 11, 2005

17:36

D. S. Pope in U. beebei (Christy, 1988a). The effectiveness of such a strategy for minimizing the information available to neighbours is unknown and more information is needed on how widespread this orientation strategy is in fiddler crabs. Future research should be directed towards investigating not only the existence of the potential conspecific interceptive eavesdropping described here but also whether males suffer substantial costs from such eavesdropping, and if so, what counter-strategies they might employ for minimizing these costs. Only further work can illuminate the costs and benefits to fiddler crab signallers and receivers of operating in a network environment and uncover behavioural strategies each party might use to exploit the situation to their best advantage. The diversity among fiddler crab species, in terms of habitats (mudflat versus mangrove), mating tactics and display form and function will create variation in the extent to which the network perspective is applicable to fiddler crabs and could serve as useful tests of the predicted consequences to these crabs of signalling in a network environment.

Acknowledgements I would like to thank Peter McGregor and the ‘communication crew’ at the University of Copenhagen for welcoming me into their network and stimulating and clarifying my thinking on the communication networks of fiddler crabs. Michael Jennions, Giuliano Matessi, Ricardo Matos, Peter McGregor, Tom Peake, Andrew Terry and an anonymous reviewer all provided very helpful feedback on earlier versions of this chapter. I would like to thank Jochen Zeil in particular for his comments and insight into the visual world of fiddler crabs. Patricia Backwell and Catherine de Rivera generously shared their results with me before they were published. My own research described here was funded by the US National Science Foundation and the Danish Natural Science Research Council, which also supported me during the writing of this chapter.

References Aizawa, N. 1998. Synchronous waving in an ocypodid crab, Ilyoplax pusilla: analyses of response patterns to video and real crabs. Marine Biology, 131, 523–532. 2000. Heterospecific timing of waving display in ocypodid crabs Ilyoplax pusilla and Uca lactea. Journal of the Marine Biological Association of the UK, 80, 947–948. Alatalo, R. V., Lundberg, A. & Bj¨ orkland, M. 1982. Can the songs of male birds attract other males? An experiment with the Pied Flycatcher Ficedula hypoleuca. Bird Behavior, 4, 42–45. Backwell, P. R. Y. & Passmore, N. I. 1996. Time constraints and multiple choice criteria in the sampling behavior and mate choice of the fiddler crab, Uca annulipes. Behavioral Ecology and Sociobiology, 38, 407–416. Backwell, P., Jennions, M., Passmore, N. & Christy, J. 1998. Synchronized courtship in fiddler crabs. Nature, 391, 31–32.

P1: IYP/KAB P2: KPB-KOD 0521823617c12.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:36

Fiddler crabs signal in networks Backwell, P. R. Y., Jennions, M. D., Christy, J. H. & Passmore, N. 1999. Female choice in the synchronously waving fiddler crab, Uca annulipes. Ethology, 105, 415–421. Backwell, P. R. Y., Christy, J. H., Telford, S. R., Jennions, M. D. & Passmore, N. I. 2000. Dishonest signalling in a fiddler crab. Proceedings of the Royal Society of London, Series B, 267, 719–724. Burford, F. R. L., McGregor, P. K. & Oliveira, R. F. 1998. Chorusing by male European fiddler crabs, Uca tangeri: a study of visual communication networks. Acta Ethologica, 1, 33–41. Christy, J. H. 1978. Adaptive significance of reproductive cycles in the fiddler crab Uca pugilator: a hypothesis. Science, 199, 453–455. 1982. Burrow structure and use in the sand fiddler crab, Uca pugilator (Bosc). Animal Behaviour, 30, 687–694. 1983. Female choice in the resource-defense mating system of the sand fiddler crab, Uca pugilator. Behavioral Ecology and Sociobiology, 12, 169–180. 1988a. Pillar function in the fiddler crab Uca beebei (II): competitive courtship signaling. Ethology, 78, 113–128. 1988b. Pillar function in the fiddler crab Uca beebei (I): effects on male spacing and aggression. Ethology, 78, 53–71. Christy, J. H. & Salmon, M. 1984. Ecology and evolution of mating systems of fiddler crabs (genus Uca). Biological Reviews 59, 483–509. Christy, J. H., Backwell, P. R. Y. & Goshima, S. 2001. The design and production of a sexual signal: hoods and hood building by male fiddler crabs Uca musica. Behaviour, 138, 1065–1083. Christy, J. H., Backwell, P. R. Y., Goshima, S. & Kreuter, T. 2002. Sexual selection for structure building by courting male fiddler crabs: an experimental study of behavioral mechanisms. Behavioral Ecology, 13, 366–374. Crane, J. 1975. Fiddler Crabs of the World (Ocypodidae: Genus Uca). Princeton, CT: Princeton University Press. de Rivera, C. E. & Vehrencamp, S. L. 2001. Male versus female mate searching in fiddler crabs: a comparative analysis. Behavioral Ecology 12, 182–191. de Rivera, C. E., Backwell, P. R. Y., Christy, J. H. & Vehrencamp, S. L. 2003. Density affects female and male mate searching in the fiddler crab, Uca beebei. Behavioral Ecology and Sociobiology, 53, 72–83. Doherty, J. A. 1982. Stereotypy and the effects of temperature on some spatio-temporal subcomponents of the ‘courtship wave’ in the fiddler crabs Uca minax (Le Conte) and Uca pugnax (Smith) (Brachyura, Ocypodidae). Animal Behaviour, 30, 352–363. Gerhardt, H. C. & Huber, F. 2002. Acoustic Communication in Insects and Anurans: Common Problems and Diverse Solutions. Chicago, IL: University of Chicago Press. Giraldeau, L.-A., Valone, T. J. & Templeton, J. J. 2002. Potential disadvantages of using socially acquired information. Philosophical Transactions of the Royal Society of London, Series B, 357, 1559–1566. Gordon, H. R. S. 1958. Synchronous claw-waving of fiddler crabs. Animal Behaviour, 6, 238–241. Greenfield, M. D. 1994. Cooperation and conflict in the evolution of signal interactions. Annual Review of Ecology and Systematics, 25, 97–126.

273

P1: IYP/KAB P2: KPB-KOD 0521823617c12.xml CU1917B/McGregor

274

0 521 582361 7

April 11, 2005

17:36

D. S. Pope 2002. Signalers and Receivers: Mechanisms and Evolution of Arthropod Communication. Oxford: Oxford University Press. Greenfield, M. D., Tourtellot, M. K. & Snedden, W. A. 1997. Precedence effects and the evolution of chorusing. Proceedings of the Royal Society of London, Series B, 264, 1355–1361. Greenspan, B. N. 1982. Semi-monthly reproductive cycles in male and female fiddler crabs, Uca pugnax. Animal Behaviour, 30, 1084–1092. Hemmi, J. M. & Zeil, J. 2003. Robust judgement of inter-object distance by an arthropod. Nature, 421, 160–163. Henmi, Y., Koga, T. & Murai, M. 1993. Mating behavior of the sand bubbler crab Scopimera globosa. Journal of Crustacean Biology, 13, 736–744. Horch, K., Salmon, M. & Forward, R. 2002. Evidence for a two pigment visual system in the fiddler crab, Uca thayeri. Journal of Comparative Physiology A, 188, 493–499. Hyatt, G. W. 1977. Quantitative analysis of size-dependent variation in the fiddler crab wave display (Uca pugilator, Brachyura, Ocypodidae). Marine Behavior and Physiology, 5, 19–36. Hyatt, G. W. & Salmon, M. 1977. Combat in the fiddler crabs Uca pugilator and U. pugnax: a quantitative analysis. Behaviour, 65, 182–211. Jennions, M. D. & Backwell, P. R. Y. 1996. Residency and size affect fight duration and outcome in the fiddler crab Uca annulipes. Biological Journal of the Linnean Society, 57, 293–306. 1998. Variation in courtship rate in the fiddler crab Uca annulipes: is it related to male attractiveness? Behavioral Ecology, 9, 605–611. Kiflawi, M. & Gray, D. A. 2000. Size-dependent response to conspecific mating calls by male crickets. Proceedings of the Royal Society of London, Series B, 267, 2157–2161. Kitaura, J., Wada, K. & Nishida, M. 1998. Molecular phylogeny and evolution of unique mud-using territorial behavior in ocypodid crabs (Crustacea: Brachyura: Ocypodidae). Molecular Biology and Evolution, 15, 626–637. Koga, T., Backwell, P. R. Y., Jennions, M. D. & Christy, J. H. 1998. Elevated predation risk changes mating behaviour and courtship in a fiddler crab. Proceedings of the Royal Society of London, Series B, 265, 1385–1390. Kosuge, T., Poovachiranon, S. & Murai, M. 1994. Male courtship cycles in three species of tropical Ilyoplax crabs (Decapoda, Brachyura, Ocypodidae). Hydrobiolgia 285, 93–100. Land, M. & Layne, J. 1995a. The visual control of behavior in fiddler crabs: I. Resolution, thresholds and the role of the horizon. Journal of Comparative Physiology A, 177, 81–90. 1995b. The visual control of behavior in fiddler crabs: I1. Tracking control systems in courtship and defence. Journal of Comparative Physiology A, 177, 91–103. Land, M. F. & Nilsson, D.-E. 2001. Animal Eyes. Oxford: Oxford University Press. Layne, J. E. 1998. Retinal location is the key to identifying predators in fiddler crabs (Uca pugilator). Journal of Experimental Biology, 201, 2253–2261. Macia, A., Quincardete, I. & Paula, J. 2001. A comparison of alternative methods for estimating population density of the fiddler crab Uca annulipes at Saco Mangrove, Inhaca Island (Mozambique). Hydrobiologia, 449, 213–219.

P1: IYP/KAB P2: KPB-KOD 0521823617c12.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:36

Fiddler crabs signal in networks McGregor, P. K. 1993. Signalling in territorial systems: a context for individual identification, ranging and eavesdropping. Philosophical Transactions of the Royal Society, Series B, 340, 237–244. McGregor, P. K. & Dabelsteen, T. 1996. Communication networks. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca, NY: Cornell University Press, pp. 409–425. McGregor, P. K. & Peake, T. M. 2000. Communication networks: social environments for receiving and signalling behaviour. Acta Ethologica, 2, 71–81. Morgan, S. G. & Christy, J. H. 1995. Adaptive significance of the timing of larval release by crabs. American Naturalist 145, 457–479. Moriito, M. & Wada, K. 1997. When is waving performed in the ocypodid crab Scopimera globosa? Crustacean Research, 26, 47–55. 2000. The presence of neighbours affects waving display frequency by Scopimera globosa (Decapoda, Ocypodidae). Journal of Ethology, 18, 43–45. Murai, M., Koga, T. & Yong, H.-S. 2002. The assessment of female reproductive state during courtship and scramble competition in the fiddler crab, Uca paradussumieri. Behavioral Ecology and Sociobiology, 52, 137–142. Pfennig, K. S., Rapa, K. & McNatt, R. 2000. Evolution of male mating behavior: male spadefoot toads preferentially associate with conspecific males. Behavioral Ecology and Sociobiology 48, 69–74. Pope, D. S. 1998. The fiddler crab claw waving display: function and evolution of a sexually selected signal. Ph.D. thesis, Duke University, USA. 2000. Testing function of fiddler crab claw waving by manipulating social context. Behavioral Ecology and Sociobiology, 47, 432–437. Popper, A. N., Salmon, M. & Horch, K. W. 2001. Acoustic detection and communication by decapod crustaceans. Journal of Comparative Physiology A, 187, 83–89. Rosenberg, M. S. 2001. The systematics and taxonomy of fiddler crabs: a phylogeny of the genus Uca. Journal of Crustacean Biology, 21, 839–869. 2002. Fiddler crab claw shape variation: a geometric morphometric analysis across the genus Uca (Crustacea: Brachyura: Ocypodidae). Biological Journal of the Linnean Society, 75, 147–162. Salmon, M. 1965. Waving display and sound production in the courtship behavior of Uca pugilator, with comparisons to U. minax and U. pugnax. Zoologica, 50, 123–150. 1984. The courtship, aggression and mating system of a ‘primitive’ fiddler crab (Uca vocans: Ocypodidae). Transactions of the Zoological Society of London, 37, 1–50. Salmon, M. & Horch, K. 1972. Acoustic signaling and detection by semi-terrestrial crabs of the family Ocypodidae. In: Behavior of Marine Animals, vol. 1, ed. H. E. Winn & B. L. Olla, New York: Plenum Press, pp. 60–96. Salmon, M. & Stout, J. F. 1962. Sexual discrimination and sound production in Uca pugilator Bosc. Zoologica 47: 15–19. Stamps, J. A. 1988. Conspecific attraction and aggregation in territorial species. American Naturalist 131, 329–347. Treherne, J. E. & Foster, W. A. 1981. Group transmission of predator avoidance behaviour in a marine insect: the Trafalgar effect. Animal Behaviour 29, 911–917.

275

P1: IYP/KAB P2: KPB-KOD 0521823617c12.xml CU1917B/McGregor

276

0 521 582361 7

April 11, 2005

17:36

D. S. Pope von Hagen, H.-O. 1962. Frielandstudien zur sexual- und forpflanzungs-biologie von Uca ¨ kologie der Tiere, 51, 611–725. tangeri in Andalusien. Zietschrift f¨ ur Morphologie und O Wada, K. 1984. Pair formation in the two forms of Macrophthalmus japonicus De Haan (Crustacea: Brachyura) at a co-occurring area. Journal of Ethology, 2, 7–10. 1991. Biogeographic patterns in waving display, and body size and proportions of Macrophthalmus japonicus species complex (Crustacea: Brachyura: Ocypodidae). Zoological Science, 8, 135–146. 1994. Earthen structures built by Ilyoplax dentimerosa (Crustacea, Brachyura, Ocypodidae). Ethology, 96, 270–282. Watts, D. J. 2002. A simple model of global cascades on random networks. Proceedings of the National Academy of Sciences, USA, 99, 5766–5771. Weissburg, M. 1992. Functional analysis of fiddler crab foraging: sex-specific mechanics and constraints in Uca pugnax (Smith). Journal of Experimental Marine Biology and Ecology, 156, 105–124. Wolfrath, B. 1993. Observations on the behaviour of the European fiddler crab Uca tangeri. Marine Ecology Progress Series, 100, 111–118. Wright, H. O. 1968. Visual displays in brachyuran crabs: field and laboratory studies. American Zoologist 8, 655–665. Yamaguchi, T. 2001a. The breeding period of the fiddler crab, Uca lactea (Decapoda, Brachyura, Ocypodidae) in Japan. Crustaceana 74, 285–293. 2001b. Incubation of eggs and embryonic development of the fiddler crab, Uca lactea (Decapoda, Brachyura, Ocypodidae). Crustaceana 74, 449–458. 2001c. The mating system of the fiddler crab, Uca lactea (Decapoda, Brachyura, Ocypodidae). Crustaceana 74, 389–399. Zeil, J. & Al-Mutairi, M. M. 1996. Variation of resolution and of ommatidial dimensions in the compound eyes of the fiddler crab Uca lactea annulipes (Ocypodidae, Brachyura, Decapoda). Journal of Experimental Biology, 199, 1569–1577. Zeil, J. & Hofmann, M. 2001. Signals from ‘crabworld’: cuticular reflections in a fiddler crab colony. Journal of Experimental Biology, 204, 2561–2569. Zeil, J. & Zanker, J. M. 1997. A glimpse into crabworld. Vision Research, 37, 3417–3426. Zeil, J., Nalbach, G. & Nalbach, H.-O. 1986. Eyes, eye stalks and the visual world of semi-terrestrial crabs. Journal of Comparative Physiology A, 159, 801–811. Zucker, N. 1978. Monthly reproductive cycles in three sympatric hood-building tropical fiddler crabs (genus Uca). Biological Bulletin, 155, 410–424.

P1: JZZ/... P2: JZZ/... 0521823617c13.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

16:26

13

Anuran choruses as communication networks t. u l m a r g r a f e University of W¨ urzburg, Germany

Introduction It is becoming more and more apparent that communication often takes place in a network of several signallers and receivers (as shown by most of the chapters in this volume and reviewed by McGregor & Peake (2000)). The network view of communication stresses that signallers and receivers have additional costs and benefits to those usually found in dyadic interactions. For example, in communication networks signallers often face the problem of intense intra- and interspecific competition whereas receivers must discriminate information from individuals under conditions of high background noise. In many frogs and toads, males aggregate in large choruses to advertise for females. The signals they use are conspicuous and long range; therefore, choruses constitute a classic example of a communication network. The challenge of communicating in such large choruses is to balance the costs and benefits of attracting a mate, repelling rivals and avoiding predators and/or parasites. Consequently, advertising in choruses will have far-reaching effects on vocal behaviour. If we want to understand signal design and signalling behaviour in such aggregations, we need to look at communication in the network context in which these different selective pressures operate. In this chapter, I will review why it is important to investigate communication in chorusing anurans within the network environment. I will focus on the behaviour of both signallers and receivers. First, I will discuss patterns of male–male vocal competition that can best be understood within the network environment. In particular, I will discuss how the timing of signals within the chorus determines mating success. The variation in signal timing between anurans suggests that they Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

277

P1: JZZ/... P2: JZZ/... 0521823617c13.xml CU1917B/McGregor

278

0 521 582361 7

April 11, 2005

16:26

T. U. Grafe are fine-tuned to both the level of competition and the receiver biases. Furthermore, high chorus density necessitates that males interact with only a subset of males. Determining the degree of selective attention (i.e. the number of individuals males interact with) is, therefore, an important parameter of connectivity in the network. Aggregations of chorusing males allow females to monitor male– male interactions prior to mate choice. What evidence from anurans is there that females eavesdrop? Similarly, both calling and satellite males may benefit from monitoring male–male interactions and position themselves within the chorus to increase their active space, associate with attractive callers or gain information about the location of gravid females. Signaller behaviour Patterns of male–male vocal competition

Many anurans form large and conspicuous aggregations in which high levels of background noise reduce their effectiveness in attracting females (Gerhardt & Klump, 1988; Narins & Zelick, 1988; Gerhardt & Schwartz, 1995). Similar effects are found in insects (e.g. R¨ omer et al., 1989) and birds (Klump, 1996). Chorusing intensifies competition between males over females or over resources of interest to females. Temporal segregation of calling activity or the partitioning of calling frequencies are solutions available to some species in some situations. However, in many cases, males cannot avoid calling in dense aggregations and are subject to high levels of intra- and interspecific acoustic interference. This is especially evident in species that breed in temporary ponds where reproductive success depends on laying eggs as early as possible to ensure that larval development is completed before the breeding sites dry out. Males that vocalize in dense aggregations deal with high levels of acoustic competition in several ways (reviewed by Wells, 1988; Gerhardt & Schwartz, 1995; Gerhardt & Huber, 2002). They may increase their call repetition rate, increase the complexity of their calls or defend calling sites and acoustic space against other males. Many species show all of these adaptations. The classic example is that of the t´ ungara frog Physalaemus pustulosus where males add chuck notes to their whines to form complex advertisement calls and increase their call rate depending on the social milieu (Ryan, 1985). Some frogs even increase call intensity in response to playbacks of conspecific advertisement calls (Lopez et al., 1988) or lower the dominant frequency of their advertisement calls during aggressive interactions with other males (reviewed by Bee & Bowling, 2002). In addition, chorusing males alternate or synchronize their calls with neighbouring males (e.g. Zelick & Narins, 1983; Forester & Harrison, 1987; Klump & Gerhardt, 1992; Greenfield & Roizen, 1993; Grafe, 1999; Bosch & M´ arquez, 2002). The precise timing of calls may, in

P1: JZZ/... P2: JZZ/... 0521823617c13.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

16:26

Anuran choruses as communication networks fact, be as important in determining a male’s mating success as the acoustic call properties that, traditionally, are investigated. I discuss this in more detail in following sections. Repertoire sizes can be large in some anurans, reflecting the potential to modulate social interactions on a fine scale. Repertoires appear especially impressive in Boophis madagascariensis, with 28 distinct calls reported (Narins et al., 2000) and Amolops tormotus, with an amazing variety of calls that have defied categorization (Feng et al., 2002). Such call variety may reflect the need to signal to males and females simultaneously (see below). Signals used during aggressive interactions

Aggressive calls play an important role in mediating the spacing between male anurans within a chorus and there is tremendous variation in their use (Schwartz, 2001; Gerhardt & Huber, 2002). In fact, many species never produce distinct aggressive calls. In the genus Kassina, for example, aggressive calls have only rarely been reported (Fleischack & Small, 1978), despite extensive playback experiments in most species (T. U. Grafe, unpublished data). In some species, the advertisement call grades into the aggressive call (e.g. Wells, 1989; Grafe, 1995a); in others the advertisement call is distinct from the aggressive call, with the aggressive call being graded (e.g. Schwartz, 1989). Some species have several kinds of aggressive call (e.g. Given, 1987). Much discussion revolves around the functional significance of graded versus discrete signalling systems and how cheating can be prevented. Handicap models predict that aggressive signals should be graded to convey best information about the probability that the signaller will attack (Grafen, 1990). In contrast, conventional signalling (Enquist et al., 1998) and discrete handicap models (Johnstone, 1994) both predict that signals should be discrete. This discussion has not considered the necessity for chorusing males simultaneously to repel rivals and attract females. Here, signals need to reach two different classes of receiver (males and females) and may need to be designed differently. In general in anurans, aggressive calls are less attractive to females than advertisement calls (e.g. Brenowitz & Rose, 1999). Therefore, signals that grade between advertisement and aggressive calls or discrete aggressive call variants may allow males to increase the aggressive content of a signal gradually or discretely, while only partially reducing a male’s attractiveness to females. This may be a general solution to signalling when trying to reach different classes of receiver and a resolution to a signalling conflict that is typical of a network environment. The variation in the use of aggressive calls in anurans may depend on the degree of flexibility needed in dealing with the conflicting demands of signalling to males and females simultaneously (Brenowitz et al., 2000; Marshall et al., 2003). An elegant solution to this problem is the use of two-part advertisement calls, in which

279

P1: JZZ/... P2: JZZ/... 0521823617c13.xml CU1917B/McGregor

280

0 521 582361 7

April 11, 2005

16:26

T. U. Grafe one part is directed to females and the other to males. In Geocrinia victoriana, for example, males produce an introductory note that is directed towards males and a series of shorter repeated notes that are directed towards females (Littlejohn & Harrison, 1985). A similar case can be made for the diphasic call of Eleutherodactylus coqui (Narins & Capranica, 1978). Viewed from an energetic perspective, such a solution appears wasteful because only every second note is directed to females. This may explain why such a solution is not very common. Fine-scale patterns of signal timing

Signal timing can be an important feature influencing mating success in chorusing anurans and, therefore, males need to monitor the call timing of competitors (reviewed by Greenfield, 1994a; Gerhardt & Huber, 2002). The finescale patterns of signal timing are often described by taking the calling period of one male as a reference point and relating the timing of the second individual to this reference (reviewed by Klump & Gerhardt, 1992; Greenfield, 1994a). At one end of the continuum of call-timing patterns, individuals signal in perfect synchrony (relative phase of 0◦ ); at the other end signals are spaced with equal intervals between them resulting in perfect alternation (relative phase of 180◦ ). Most anurans do not show either of these extreme patterns on a regular basis. Synchrony can, therefore, refer to signalling patterns in which signals overlap, whereas alternation characterizes patterns in which signals regularly are more or less evenly spaced in time (Gerhardt & Huber, 2002). Examples of different calltiming patterns are shown in Fig. 13.1. I use the term entrainment to refer to call-timing patterns with relative phase angles below 45◦ but not overlapping (a similar classification is used in insects (Greenfield, 1994a)). This represents an operational definition that helps to classify a continuously varying parameter. Alternating would then refer to call-timing patterns with relative phase angles above 45◦ . In dyadic interactions of entrained calling, the calls of one male will lead and the calls of the other male will follow. Leading and lagging roles often switch between individuals (Fig. 13.1; Gerhardt & Huber, 2002; Grafe, 2003) are less discernible in alternating species (Klump & Gerhurdt 1992; Bosch & M´ arquez, 2002). Levels of overlap rise dramatically in aggregations of three and four males (e.g. Schneider et al., 1988; Schwartz, 1993; Grafe, 1996a). However, overlap remains lower than expected if frogs were calling at random. This suggests that males are interacting in ways that prevent high levels of overlap. In most cases, males will attempt to place their calls in relatively silent gaps, thus avoiding masking interference by other males. Zelick & Narins (1985), in their pioneering work with the Puerto Rican treefrog E. coqui, documented experimentally that males were able to place their calls in unpredictable gaps of silence. Males were triggered to

P1: JZZ/... P2: JZZ/... 0521823617c13.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

16:26

Anuran choruses as communication networks

Fig. 13.1. Spectrograms of call-timing patterns (defined in text) in five different species of running frogs (Kassina). K. cassinoides shows alternation, K. schioetzi entrainment, K. senegalensis entrainment with occasional overlap, K. fusca synchronous calling and K. kuvangensis synchronous call groups with alternating calls. The letters (A, B) designate individual males. Note the switching of leader and follower roles between males in the top four spectrograms. These species call syntopically (i.e. in the same pond and at the same time) and were recorded in the Guinea savannah region of the Como´e National Park, Ivory Coast. K. kuvangensis was recorded at Hillwood Farm, northwestern Zambia.

281

P1: JZZ/... P2: JZZ/... 0521823617c13.xml CU1917B/McGregor

282

0 521 582361 7

April 11, 2005

16:26

T. U. Grafe call by the rapid offset of background sound. Similar studies on other anurans have supported these basic results and the underlying mechanisms of call initiation (Schwartz, 1993; Grafe, 1996a). Interestingly, constant high noise levels do not prevent calling, suggesting that call inhibition subsides with time. Like E. coqui, Broadley’s painted reed frogs Hyperolius marmoratus broadleyi initiate their calls in response to drops in background noise levels, showing an ‘off-response’ (Grafe, 1996a). The modal response latencies in the reaction of males were between 40 and 80 milliseconds (in some bushcrickets it is less than 20 milliseconds (Robinson, 1990)), suggesting that higher auditory centres such as the thalamus are not involved in processing such a fast response (Walkowiak, 1992). Although modal response latencies are very short, most calls are given with much longer latencies. Males avoid call overlap by calling within windows of low noise levels (‘silence’) by selectively attending to near and thus loud neighbours. The flexibility with which males can adjust their calls even on a note-by-note basis is remarkable. Schwartz (1993), studying Hyla microcephala, found that males increased the spacing between their calls when interrupted, thus avoiding further overlap of subsequent notes in their call. I found a similar response in the Central African frog Kassina kuvangensis (Grafe, 2003). Calling in this species is characterized by synchronizing call groups while at the same time alternating advertisement calls with those of neighbouring males. As in H. microcephala, males readjusted their inter-call intervals within milliseconds in response to the playback stimulus. In contrast to alternating or entrained calling, synchronous calling is unusual in anurans. It has been reported for only a handful of species: the neotropical hylids Smilisca sila (Ryan, 1986), Hyla ebraccata (Wells & Schwartz, 1984) and Centrolenella ˜ ez, 1993) as well as the African running frogs Kassina fusca and granulosa (Ib´ an Kassina senegalensis (Grafe, 1999; T. U. Grafe & H. L¨ ussow, unpublished data). In pairwise interactions of the savannah running frog K. fusca, 81.5% (overall median) of calls overlapped (relative phase of 8.6 ± 4.4◦ ) with a median degree of overlap of 20.8%. In the Senegal running frog K. senegalensis, only 21.6% of calls between neighbouring males overlapped on average (T. U. Grafe & H. L¨ ussow, unpublished data). Synchrony in anurans, in contrast to many insects (reviewed by Greenfield, 1994a), is achieved by a rapid acoustic response to the onset of a concurrent signal produced by a neighbour. This ‘on-response’ mechanism is thus fundamentally different from the ‘off-response’ found in the alternating and entrained response type. It should be stressed, however, that at least one anuran, K. fusca, can vary its response type. When presented with playbacks of conspecific advertisement calls, males showed a synchronous response whereas they entrained calls to heterospecific calls or white noise stimuli, thus showing both on- and off-responses (Grafe, 1999).

P1: JZZ/... P2: JZZ/... 0521823617c13.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

16:26

Anuran choruses as communication networks There is considerable variation in the patterns of call timing between species. In insects and anurans, call alternation is found predominantly in species that show low call rates, whereas synchrony is more common in species that call at high rates (Narins, 1982; Greenfield, 1994a). An exception to this rule is the genus Kassina. Males show synchronous or entrained calling at both low and high call rates independent of density, suggesting that call rate alone is not the prime determinant of signal timing (Grafe, 1999). Moreover, some species show both synchrony and alternation (e.g. Moore et al., 1989). In K. kuvangensis, males synchronize call groups while at the same time alternating calls within call groups. Such synchronized interdigitated calling may serve to reduce predation while maintaining speciesspecific temporal information important to females (Grafe, 2003). This may be a solution to reducing the costs associated with synchrony and alternation. Adaptive significance of call timing

Both cooperative and competitive hypotheses can explain the evolution and maintenance of call timing in anurans and insects. Cooperative explanations for synchrony include (a) confusion of predators by decreasing the locatability of signals; (b) enhancement of detection by females by increasing the peak amplitude of signals; and (c) improving the detection of female acoustic responses. Several authors have noted that there is little support for these cooperative explanations (Greenfield, 1994a; McGregor & Peake, 2000). Only in one species, S. sila, may synchronous calling provide protection against frog-eating bats Trachops cirrhosus, since bats were shown to be more attracted to alternating than to synchronous playback of calls (Tuttle & Ryan, 1982). There is no evidence for enhanced detection of overlapping calls by females. Peak amplitude of synchronous calls or choruses of males is not much higher than that of individual signallers (Bradbury, 1981), thus providing only a marginal increase in the active space of males. Although females are attracted to larger choruses in a number of species, the mating success per male often declines as lek size increases (Deutsch, 1994; Widemo & Owens, 1995). Similarly, playback experiments with the grasshopper Ligurotettix coquilletti and the treefrog H. microcephala showed that females were attracted to larger arrays of speakers but the attractiveness per speaker was not higher for larger arrays (Shelly & Greenfield, 1991; Schwartz, 1993). Finally, acoustic responses to male advertisement calls are given in only a few species, most notably in midwife toads Alytes spp. (Bosch, 2001). Since calling effort (duty cycle) is generally not very high in midwife toads, there is little gained by synchronizing calls to improve the detection of female acoustic responses. There is strong evidence that competition to produce leading calls often explains call timing in anurans (and insects) because there is a strong preference by females for leading signals (e.g. Grafe, 1996a; Greenfield et al., 1997; Bosch &

283

P1: JZZ/... P2: JZZ/... 0521823617c13.xml CU1917B/McGregor

284

0 521 582361 7

April 11, 2005

16:26

T. U. Grafe M´ arquez, 2002). This preference for leading calls has been termed the precedence effect (reviewed by Gerhardt & Huber, 2002). In species in which females prefer leading calls, males that produce trailing calls will benefit by delaying the onset of their calls to avoid overlap or by speeding up calls to produce leading calls. Selective attention

In the previous section, I have outlined why chorusing anurans should generally avoid interference with other signalling males. The problem in a large aggregation is that if a male is to avoid interference with all males in the chorus, he would have to stop calling. The solution is to attend only to a subset of nearby males. Brush & Narins (1989) were the first to investigate systematically the question of to how many chorus members an individual male attends. They found that E. coqui males typically avoided overlap with just two neighbours, only rarely with three individuals. Monitoring small choruses of four to six male H. microcephala, Schwartz (1993) found that males generally responded to their loudest rivals. In addition, more centrally located males typically attended to more males (one to four) than those at the periphery of the chorus (one). Greenfield & Rand (2000) have further demonstrated the plasticity involved in selective attention. Male P. pustulosus responded to two to three neighbouring males depending on the chorus structure and intensity of those males’ calls. These results suggest that selective attention is a dynamic process that will vary as males enter the chorus and move within it. Selective attention to nearest neighbours may occur even in anurans that do not avoid call overlap. In K. fusca, call overlap itself is a measure of the attention paid to other males. Preliminary work (T. U. Grafe, unpublished data) suggests that the spatial distribution of males in the chorus is an important factor in determining the number of males that are paid attention. Another interesting case is the gray treefrog Hyla versicolor (Schwartz, 2001; Schwartz et al., 2001). Although females discriminate against overlapping calls, neighbouring males in small choruses do not avoid overlap. Avoidance of overlap appears to be overridden by female preference for longer calls and a step-like decrease in attractiveness of short calls even if they are unmasked. Given the preferences of females, the best strategy for male gray treefrogs appears to be the production of long calls that partly overlap with those of neighbours. The plasticity in response to acoustic competition and the differences in auditory perception of receivers suggest the absence of a unifying general rule governing selective attention. Energetics of calling

Important determinants of the interactions of calling male anurans are the energetic constraints of calling. Calling is the most energetically expensive

P1: JZZ/... P2: JZZ/... 0521823617c13.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

16:26

Anuran choruses as communication networks behaviour of ectothermic vertebrates (reviewed by Wells, 2001). In the European treefrog Hyla arborea, for example, instantaneous rates of oxygen consumption can reach 41 times the resting rate (Grafe & Thein, 2001). Such high levels of expenditure cannot be maintained for long and they set an upper limit to the rate and complexity of calling. Female anurans prefer to mate with males calling at high rates in all species tested (Ryan & Keddy-Hector, 1992; Gerhardt & Huber, 2002). Consequently, males need to adjust the rate and complexity of calls to the levels of competition or risk having to drop out of the chorus prematurely because they have run out of energy. In species that vary the duration of calls, such as H. versicolor, the best predictor of energy expenditure is the product of call rate and call duration (calling effort or duty cycle; reviewed by Wells, 2001). In most anurans, calling effort increases with chorus density (e.g. Taigen et al., 1985; Grafe, 1996b). In H. versicolor, however, males reduce the rate of calling while increasing call duration as chorus density increases, thus maintaining a constant level of energy expenditure (Wells & Taigen, 1986; Grafe, 1997a). Females prefer long calls at low rates to short calls at high rates (Klump & Gerhardt, 1987). These studies show that the energetic constraints of calling require males to monitor the behaviour of others to maintain their attractiveness towards females and that the way males partition their energy depends on female preferences. Receiver behaviour Mechanisms of female preferences

Important selection pressures on signal design and signalling behaviour are the sensory and neuronal abilities of females. In anurans, acoustic communication plays a central role in mate choice. The wide range in the threshold of auditory neurons and the sensitivity for narrow frequency bands in the peripheral auditory system are important in allowing females to choose between conspecific males in the presence of background noise (reviewed by Narins & Zelick, 1988). Numerous studies have shown that females prefer males that produce loud and conspicuous signals with large active space (Ryan & Keddy-Hector, 1992). Consequently, receivers generally exert strong selection for loud and ritualized signals. In recent years, it has become clear that the fine-scale patterns of signal timing have a large influence on female choice. Females of many taxonomic groups show a preference for leading, but not necessarily overlapping, signals in the olfactory (voles: Johnston et al., 1997), visual (fiddler crabs: Backwell et al., 1998; fireflies: Vencl & Carlson, 1998) and auditory modalities (field crickets: Wyttenbach & Hoy, 1993; katydids: Greenfield, 1994b; Greenfield et al., 1997; frogs: Gerhardt & Huber, 2002; rats: Kelly, 1974; cats: Cranford & Oberholzer, 1976). This preference for leading signals has entered the literature under the term precedence effect. It was originally described by Wallach et al. (1949) in humans and describes the

285

P1: JZZ/... P2: JZZ/... 0521823617c13.xml CU1917B/McGregor

286

0 521 582361 7

April 11, 2005

16:26

T. U. Grafe observation that when two spatially separated sounds are presented with a brief delay in onset the leading sound dominates localization. However, apart from the work on mammals, the experimental designs of the studies listed above do not distinguish between masking of the trailing call and the inability to locate the trailing call even though it is easily heard (i.e. it is not masked). To distinguish between these alternatives, psychoacoustic studies will be necessary. From an evolutionary perspective, however, the selection pressures on males to produce leading calls will be strong irrespective of the underlying mechanisms. I was able to demonstrate a strong preference for leading calls in Broadley’s painted reed frog H. m. broadleyi (Grafe, 1996a) and the savannah running frog K. fusca (Grafe, 1999). In reed frogs, the preference by females for the leading call was largely independent of sound pressure, underscoring the robustness of this preference (see also Dyson & Passmore, 1988a). In the synchronously calling running frog, preference for both leading and trailing calls was observed depending on the degree of overlap (Grafe, 1999). Females preferred leading calls when calls overlapped by 75% and 90% but switched their preference to trailing calls at 10% and 25% of overlap. Thus, males are selected to overlap the calls of neighbours; however, they should not do so with high degrees of overlap. Interestingly, playback experiments also showed that males were able to initiate their calls sooner than they actually do, suggesting that special mechanisms are involved that inhibit males from calling with high degrees of overlap. For the savannah running frog, the adaptive significance of synchronous calling is explained, at least in part, by the auditory preferences of females. Whereas the preference for leading signals appears to be a basic design feature of nervous systems and thus a constraining feature of receivers that males need to attend to, the preference of females for trailing signals is likely to be a more fine-tuned adaptation by receivers to specific signalling environments. It remains unclear why female savannah running frogs prefer trailing calls at low degrees of overlap. A comparative analysis within the genus Kassina may provide some answers as to how a species’environment, in particular the communication network in which a population finds itself, influences signal design and signalling behaviour. Preliminary female choice experiments suggest that the call-timing pattern of males is also tuned to the respective preference functions of females in K. senegalensis (T. U. Grafe & H. L¨ ussow, unpublished data). However, call-timing patterns do not correlate with habitat characteristics such as degree of cover or calling site. Further comparative work needs to be done to elucidate the environmental correlates of call timing in anurans. It should be noted that the physical characteristics of the communication channel, the transmission properties of the environment and the network structure

P1: JZZ/... P2: JZZ/... 0521823617c13.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

16:26

Anuran choruses as communication networks (i.e. distance between network members) might select for different signals (Wiley & Richards, 1978; Ryan et al., 1990; Staaden & R¨ omer, 1997). Signal attenuation and degradation through reverberations or irregular amplitude fluctuations are factors that limit the active space of signals (Wiley & Richards, 1978; see also Ch. 20). These effects vary between habitats, with distance between sender and receivers and with height above ground. Spectral components of the call are degraded least, for example, if either sender or receivers are elevated (Wiley & Richards, 1978). Ryan & Wilczynski (1991) demonstrated that differences in habitat characteristics explained a large part of the variance in the frequency of advertisement calls between populations of the chorus frog Acris crepitans (see Wiley (1991) for bird examples). In several recent studies with anurans, however, none of the predicted differences in call features was found between species from different habitats (Penna & Solis, 1998; Kime et al., 2000). The evidence for the influence of transmission properties of the environment and network structure on signal design and signalling behaviour in anurans is equivocal at best. Comparing female choice in two-choice trials and in natural choruses

To identify which acoustic parameters are important in determining male mating success, researchers traditionally use two-choice trials in which female anurans are given the choice between two acoustic stimuli. Gravid female frogs and toads readily phonotactically approach one of the speakers and will search for the male on or in the speaker. Typically population-wide preferences of females are then noted and inferences drawn about the importance of male acoustic traits (Gerhardt, 1994). Regarded from a network perspective, such experiments should be viewed with caution because females are being tested in very simplified environments in which background noise is reduced to a minimum (see also Sullivan et al., 1995). To illustrate this point, I will review three examples of how preferences demonstrated in simple arena trials may not translate into sexual selection in natural populations. The first example is from the detailed studies on the South African painted reed frog Hyperolius marmoratus marmoratus by Neville Passmore and his colleagues. They showed that females preferred lower frequency calls, suggesting that larger males should have a mating advantage. This, however, was only the case when comparing the mating success and body size of males in small choruses (Telford et al., 1989). In large choruses, large males no longer had an advantage. Instead of preferring large males, females tested in the field preferred males calling at high rates (Passmore et al., 1992). Therefore, under noisy field conditions, female preference for calls of lower frequency were overridden by preferences for call rate. I made a similar observation when studying mate choice in Broadley’s painted reed frog H. m. broadleyi. To determine female preferences in the small choruses

287

P1: JZZ/... P2: JZZ/... 0521823617c13.xml CU1917B/McGregor

288

0 521 582361 7

April 11, 2005

16:26

T. U. Grafe in the field, I monitored male calling behaviour using an array of microphones set around the periphery of small natural ponds (Grafe, 1997b). This technique of passive sound localization (McGregor & Dabelsteen, 1996) enabled me to match precisely each call on the recording to an individual male in the chorus. I monitored both the spectral and the temporal components of each male’s calls while a female, released from the edge of the pond, was choosing among males. The analysis showed that call rate and proximity to the female release site were the best predictors of male mating success. These preferences were corroborated in traditional two-choice trials. Interestingly, females also preferred medium-frequency calls to high or low frequencies in two-choice trials, a preference not manifested under noisy field conditions. Furthermore, call parameters of interest to females are often intensity dependent: that is, preference for a call trait can be reversed by increasing the intensity of an alternative stimulus (e.g. Arak, 1988; Gerhardt, 1988). The third example showing the importance of testing females under natural conditions is the work on H. versicolor by Schwartz (2001). As mentioned above, females tested in arena choice trials prefer males producing longer calls even if they call at a lower rate as long as calling effort remains the same. The preference for long calls is non-linear, with strong discrimination against very short calls. Under quiet conditions, the discrimination was remarkable. On average, females discriminated in favour of calls on the basis of 1.5 pulses (out of 20). Background noise played to females over an additional speaker reduced the ability to discriminate to 2.3 pulses. Field experiments supported the importance of call duration in mating success; however, it was limited. An array of eight speakers was placed along the perimeter of a pond and calls of varying duration and call rate were broadcast over many nights. Naturally arriving females were trapped at the speaker of their choice. The extent of the preference for call duration was quite restricted, with only the shortest call being discriminated against. Overall, the preference for long calls explained less than 10% of the variation in male mating success in the field. These examples do not argue against the utility of two-choice trials but suggest that they should not stand alone. In chorusing anurans, females have to make decisions under acoustically unfavourable environments, often under the risk of predation, and must, therefore, limit their choosiness. The utility of two-choice trials comes into play when testing hypotheses generated by field observations or experiments, as recently demonstrated by Murphy & Gerhardt (2002). They combined field observations of mate sampling by female barking treefrogs Hyla gratiosa with two- and three-speaker choice trials to determine the influence of distance of calling males on female choice. As in Broadley’s painted reed frog, most female barking treefrogs approached the closest male and mated with him. Evidence suggests that these species simultaneously sample males. Such sampling is especially

P1: JZZ/... P2: JZZ/... 0521823617c13.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

16:26

Anuran choruses as communication networks vulnerable to background noise and will be limited by the perceptual abilities of females (Murphy & Gerhardt, 2002). It should be noted, however, that females of other anurans probably use sequential sampling techniques and, therefore, are less influenced by background noise because they closely approach several males (reviewed by Murphy & Gerhardt, 2002). Eavesdropping by females

Do females extract information from male–male interactions that influences their mating decisions? Rephrasing this question in communication network terminology: are females eavesdropping on male–male interactions (i.e. indulging in social eavesdropping; Ch. 2)? The evidence that they are comes from observing female choice for the relative timing of male advertisement calls. Like bird song, the advertisement call of anurans is directed to both males and females (i.e. it is not just a mating call). Females can potentially extract information from how males interact. Chorusing male anurans adjust the timing of their advertisement calls to that of neighbouring males in a competitive way (Klump & Gerhardt, 1992; Schwartz, 2001). As outlined above, in most species, females prefer the calls of leading males, thereby often overriding their preferences for other call parameters (e.g. Dyson & Passmore, 1988b; Grafe, 1996a). In a few cases, females prefer follower calls (Schwartz & Wells, 1984; Grafe, 1999). Evidence that females eavesdrop on male–male interactions requires simultaneous monitoring of males, i.e. that they show simultaneous mate choice. Good evidence for simultaneous mate choice comes from species that approach males only after spending some time, often several minutes, at the edge of breeding ponds and from the relative preferences of females tested in two-choice trials (Grafe, 1997b; Murphy & Gerhardt, 2002). However, a convincing study of social eavesdropping would need to show that females are not just approaching the first male they can distinguish from the background noise, in most cases this would be the nearest male. In H. m. broadleyi, females based their choice not only on nearby males but also on male call rate (Grafe, 1997b). Likewise, female Hyla gratiosa did not just approach the first male they could distinguish from the background noise (Murphy & Gerhardt, 2002). Such observations and experimental evidence suggests that females monitor the calling behaviour of more than one male. Two-choice trials have shown that females prefer males that jam the calls of other males. For social eavesdropping to occur, one need not assume high cognitive abilities. In fact, the proximate mechanisms underlying female preference for leading or follower calls may not even require the involvement of higher auditory centres. Eavesdropping by males

Potentially, anuran choruses provide ample opportunities for males to eavesdrop on the interactions between other males or between males and females.

289

P1: JZZ/... P2: JZZ/... 0521823617c13.xml CU1917B/McGregor

April 11, 2005

16:26

T. U. Grafe 30 25

Call rate (calls/0.5 min)

290

0 521 582361 7

20 15 10 5 0

0

2.5

5

7.5

10

12.5

15

17.5

20

Time (min) Fig. 13.2. Call rates of seven males (open and closed symbols) in a small chorus over 20 minutes. Note that the ‘opportunistic male’ (closed symbol) is not calling until 1.25 minutes before he went into amplexus (at 19.8 minutes). Since call rates were high at other times during the recording, other cues, such as seeing the female, in addition to high call rates are likely to have directed the attention of the ‘opportunistic male’ to the female.

Males of many species increase their call rate substantially when approached by a female and this can lead to local interactions with heightened activity between males. Other silent or satellite males in the vicinity may make use of such information by approaching this chorus area and attempting to intercept the female or attract females by vigorously starting to call themselves. I recorded three cases of such ‘opportunistic’ calling in Broadley’s painted reed frog using a microphone array (Grafe, 1995b). These males were silent throughout most of the recording period (3, 19, 20 minutes) and started calling vigorously only after a female had been introduced into the chorus and was moving towards other calling males. In all three cases, males approached the area of heightened activity and were successful in attracting the female’s attention and mating with her. One case is shown in Fig. 13.2. Documenting the behaviour of non-signalling individuals is particularly difficult and may be one of the reasons why little is known about ‘silent’ (eavesdropping) strategies. This underscores the utility of using microphone arrays to record chorusing activity because the absence of calling can be documented using this technique. Interceptive eavesdropping (Ch. 2) is known to occur in some anurans. In spadefoot toads and green frogs, for example, satellite males associate with speakers that produce attractive advertisement calls (Pfennig et al., 2000; Gerhardt & Huber, 2002), suggesting that these males are monitoring the activity of other

P1: JZZ/... P2: JZZ/... 0521823617c13.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

16:26

Anuran choruses as communication networks males and making adaptive decisions. Such behaviour is probably more common than reported. In anurans, in which females give acoustic responses to male advertisement calls (e.g. midwife toads, see above), it would be interesting to see if either sex uses these male–female interactions to interfere with the courting pair. For example, female Australian bushcrickets Elephantodeta nobilis give acoustic responses to male advertisement calls (Bailey & Field, 2000). Males that are probably satellites are attracted by these duets and produce advertisement calls, thereby occasionally attracting these females themselves. Eavesdropping on male–female vocal interactions may occur in Alytes spp. as well. Two studies report eavesdropping in captive Alytes obstreticans in which females competed for male parental care by approaching vocalizing pairs, attempting to block the path of other females and displacing amplectant males (Verrell & Brown, 1993; Grafe et al., 1999). Indirect evidence for eavesdropping comes from observing the signal type and signalling intensity used during communication in a network environment. Privatizing an interaction is a likely consequence of eavesdropping (Ch. 3). Many male anurans have distinct courtship calls that are quieter than their advertisement calls (reviewed by Wells, 1988). In E. coqui, for example, males use these quiet courtship calls to lead females to oviposition sites on the forest floor (Townsend & Stewart, 1986). Since females are not being mate guarded (i.e. are not in amplexus), it is important for the male to prevent interference by other males. Another point of interest is that anuran advertisement signals are generally omnidirectional (e.g. Passmore, 1981), probably because males cannot predict from where females approach. Spherical spreading, however, facilitates eavesdropping. One would predict that courtship calls should be more directional; however, sound fields of courtship calls have not been measured in anurans.

Summary and future directions Aggregations of calling frogs and toads are characterized by high levels of background noise. The common problems of communicating in a noisy environment, with its variety of conflicting selection pressures that act on both signallers and receives, have led to a diversity of solutions. Masking interference, for example, is reduced in most species by adjusting the timing of signals to avoid overlap. In some species, however, overlap increases as males compete to become more attractive to females. Many features of communicating in networks, such as signal timing interactions, selective attention and simultaneous mate choice, have been relatively well studied in anurans. Other specific signalling behaviours, predicted from communication network theory, such as social eavesdropping (Ch. 14), audience effects

291

P1: JZZ/... P2: JZZ/... 0521823617c13.xml CU1917B/McGregor

292

0 521 582361 7

April 11, 2005

16:26

T. U. Grafe (Ch. 4) and victory displays (Ch. 6), have generally not been considered or investigated explicitly. Such behaviours would be predicted to occur in species that have several encounters with the same individuals, such as in many territorial ranid frogs that show individual recognition (e.g. Bee & Gerhardt, 2002) or in dendrobatid frogs with year-round territorial behaviour and complex patterns of parental care (e.g. Summers, 1992). Social eavesdropping is likely to be a general phenomenon of anuran choruses. Identifying additional cases of social eavesdropping in anurans, as outlined above, would provide a fruitful avenue for future studies and would further highlight that anuran choruses are complex communication networks. An open question in this context is the functional significance of female preferences for leading (or lagging) advertisement calls. Determining any indirect and direct benefits females may obtain from their choice of leading males would be highly desirable. In addition, studies that demonstrate social eavesdropping through aggressive interactions would be of interest. The relative ease with which phonotaxis can be induced in females in the laboratory and the availability of sound synthesis software for the production of synthetic signals that can be constructed with signal parameters varying independently of each other have diverted attention from testing female preferences in the chorus. New techniques, such as multiple channel recordings as well as microphone and speaker arrays, will undoubtedly contribute to our understanding of patterns of male vocal competition and female choice. More observational data on female sampling behaviour would also contribute to revealing how receivers deal with complex signalling networks (e.g. Murphy & Gerhardt, 2002). The perceptual basis of communication in noisy environments remains largely unexplored (Ch. 20). It seems likely that receivers group sounds into auditory streams in order to improve recognition and to assign them to individual signallers (i.e. auditory scene analysis: Feng & Ratnam 2000; Hulse 2002). A recent study by Farris et al. (2002) demonstrated auditory grouping in female t´ ungara frogs in which the whine and chuck are grouped together even when presented from widely different directions. Understanding how auditory systems group incoming signals or signal components will help to explain how animals communicate in noisy networks and how they achieve selective attention; it also has the potential to provide a mechanistic basis for understanding the evolution of multicomponent signals. Anurans offer unique opportunities to study communication networks. Males of many species aggregate in choruses of varying size and complexity. Advertisement calls of anurans are long-range signals used to attract females and repel rival males. Investigations are facilitated by the species-specific and highly stereotyped signals as well as by the generally small signal repertoires. The consequences of

P1: JZZ/... P2: JZZ/... 0521823617c13.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

16:26

Anuran choruses as communication networks strategic decisions can readily be observed because fertilization is generally external, making paternity analysis unnecessary. Consequently, the evolutionary consequences of communicating in network environments can be assessed with relative ease. In addition, anurans occupy a wide variety of habitats and use a variety of reproductive strategies. This diversity provides numerous opportunities for comparative analyses.

Acknowledgements I thank Peter McGregor and two anonymous reviewers for their helpful comments on a previous version of this chapter. My work on running frogs in West Africa was supported by the Deutsche Forschungsgemeinschaft (Gr 1584).

References Arak, A. 1988. Female mate selection in the natterjack toad: active choice or passive attraction? Behavioral Ecology and Sociobiology, 22, 317–327. Backwell, P., Jennions, M., Passmore, N. & Christy, J. 1998. Synchronized courtship in fiddler crabs. Nature, 391, 31–32. Bailey, W. J. & Field, G. 2000. Acoustic satellite behaviour in the Australian bushcricket Elephantodeta nobilis (Phaneropterinae, Tettigonidae, Orthoptera). Animal Behaviour, 59, 361–369. Bee, M. A. & Bowling, A. C. 2002. Socially mediated pitch alteration by territorial male bullfrogs, Rana catesbeiana. Journal of Herpetology, 36, 140–143. Bee, M A. & Gerhardt, H. C. 2002. Individual voice recognition in a territorial frog (Rana catesbeiana). Proceedings of the Royal Society of London, Series B, 269, 1443–1448. Bosch, J. 2001. Female reciprocal calling in the Iberian midwife toad (Alytes cisternasii) varies with male call rate and dominant frequency: implications for sexual selection. Naturwissenschaften, 88, 434–437. Bosch, J. & M´ arquez, R. 2002. Female preference function related to precedence effect in an amphibian anuran (Alytes cisternasii): tests with non-overlapping calls. Behavioral Ecology, 13, 149–153. Bradbury, J. W. 1981. The evolution of leks. In: Natural Selection and Social Behavior, ed. R. D. Alexander & D. W. Tinkle. New York: Chiron Press, pp. 138–169. Brenowitz, E. A. & Rose, G. J. 1999. Female choice and plasticity of male calling behaviour in the Pacific treefrog. Animal Behaviour, 57, 1337–1342. Brenowitz, E. A., Rose, G. J. & Alder, T. 2000. The neuroethology of acoustic communication in Pacific treefrogs. In: Anuran Communication, ed. M. J. Ryan. Washington, DC: Smithsonian Institute Press, pp. 145–155. Brush, J. S. & Narins, P. M. 1989. Chorus dynamics of a Neotropical amphibian assemblage: comparison of computer simulation and natural behaviour. Animal Behaviour, 37, 33–44.

293

P1: JZZ/... P2: JZZ/... 0521823617c13.xml CU1917B/McGregor

294

0 521 582361 7

April 11, 2005

16:26

T. U. Grafe Cranford, J. & Oberholzer, M. 1976. Role of neocortex in binaural hearing in the cat. II: The ‘precedence effect’ in sound localization. Brain Research, 111, 225–239. Deutsch, J. C. 1994. Uganda kob mating success does not increase on larger leks. Behavioral Ecology and Sociobiology, 34, 451–459. Dyson, M. L. & Passmore, N. I. 1988a. Two-choice phonotaxis in Hyperolius marmoratus (Anura: Hyperoliidae): the effect of temporal variation in presented stimuli. Animal Behaviour, 36, 648–652. 1988b. The combined effect of intensity and the temporal relationship of stimuli on phonotaxis in female painted reed frogs Hyperolius marmoratus. Animal Behaviour, 36, 1555–1556. Enquist, M., Ghirlanda, S. & Hurd, P. L. 1998. Discrete conventional signalling of a continuous variable. Animal Behaviour, 56, 749–754. Farris, H. E., Rand, A. S. & Ryan, M. J. 2002. The effects of spatially separated call components on phonotaxis in t´ ungara frogs: evidence for auditory grouping. Brain, Behavior and Evolution, 60, 181–188. Feng, A. S. & Ratnam, R. 2000. Neural basis of hearing in real-world situations. Annual Review of Psychology, 51, 699–725. Feng, A. S., Narins, P. M. & Xu, C.-H. 2002. Vocal acrobatics in a Chinese frog, Amolops tormotus. Naturwissenschaften, 89, 352–356. Fleischack, P. C. & Small, C. P. 1978. The vocalizations and breeding behaviour of Kassina senegalensis (Anura, Rhacophoridae) in summer breeding aggregations. Koedoe, 21, 91–99. Forester, D. C. & Harrison, W. K. 1987. The significance of antiphonal vocalisation by the spring peeper, Pseudacris crucifer (Amphibia, Anura). Behaviour, 103, 1–15. Gerhardt, H. C. 1988. Acoustic properties used in call recognition by frogs and toads. In: The Evolution of the Amphibian Auditory System, ed. B. Fritzsch, M. J. Ryan, W. Wilczynski, T. E. Hetherington & W. Walkowiak. New York: John Wiley, pp. 455–483. 1994. The evolution of vocalization in frogs and toads. Annual Review of Ecology and Systematics, 25, 293–324. Gerhardt, H. C. & Huber, F. 2002. Acoustic Communication in Insects and Anurans: Common Problems and Diverse Solutions. Chicago, IL: Chicago University Press. Gerhardt, H. C. & Klump, G. M. 1988. Masking of acoustic signals by the chorus background noise in the green treefrog: a limitation on mate choice. Animal Behaviour, 36, 1247–1249. Gerhardt, H. C. & Schwartz, J. J. 1995. Interspecific interactions in anuran courtship. In: Amphibian Biology: Social Behaviour, ed. H. Heatwole & B. K. Sullivan. Chipping Norton, UK: Surrey Beatty, pp. 601–632. Given, M. F. 1987. Vocalizations and acoustic interactions of the carpenter frog, Rana virgatipes. Herpetologica, 43, 467–481. Grafe, T. U. 1995a. Graded aggressive calls in the African painted reed frog Hyperolius marmoratus (Hyperoliidae). Ethology, 101, 67–81. 1995b. Sexual selection and communication in the African painted reed frog (Hyperolius marmoratus). Ph.D. Thesis, Cornell University, Ithaca, NY.

P1: JZZ/... P2: JZZ/... 0521823617c13.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

16:26

Anuran choruses as communication networks 1996a. The function of call alternation in the African reed frog Hyperolius marmoratus: precise call timing prevents auditory masking. Behavioral Ecology and Sociobiology, 38, 149–158. 1996b. Energetics of vocalization in the African reed frog (Hyperolius marmoratus). Comparative Biochemistry and Physiology A, 114, 235–243. 1997a. Use of metabolic substrates in the gray treefrog, Hyla versicolor: implications for calling behavior. Copeia, 1997, 356–362. 1997b. Costs and benefits of mate choice in the lek-breeding reed frog, Hyperolius marmoratus. Animal Behaviour, 53, 1103–1117. 1999. A function of synchronous calling and a novel female preference shift in an anuran. Proceedings of the Royal Society of London, Series B, 266, 2331–2336. 2003. Synchronised interdigitated calling in the Kuvangu running frog (Kassina kuvangensis). Animal Behaviour, 66, 127–136. Grafe, T. U. & Thein, J. 2001. Energetics of calling and metabolic substrate use during prolonged exercise in the European treefrog Hyla arborea. Journal of Comparative Physiology B, 171, 69–76. Grafe, T. U., Spieler, M. & K¨ onig, B. 1999. Soziobiologische Erkl¨ arungsans¨ atze des Verhaltens von Amphibien und Reptilien. Zeitschrift f¨ ur Feldherpetologie, 6, 1–42. Grafen, A. 1990. Biological signals as handicaps. Journal of Theoretical Biology, 144, 517–546. Greenfield, M. D. 1994a. Cooperation and conflict in the evolution of signal interactions. Annual Review of Ecology and Systematics, 25, 97–126. 1994b. Synchronous and alternating choruses in insects and anurans: common mechanisms and diverse functions. American Zoologist, 34, 605–615. Greenfield, M. D. & Rand, A. S. 2000. Frogs have rules: selective attention algorithms regulate chorusing in Physalaemus pustulosus (Leptodactylidae). Ethology, 106, 331–347. Greenfield, M. D. & Roizen, I. 1993. Katydid synchronous chorusing is an evolutionary stable outcome of female choice. Nature, 364, 618–620. Greenfield, M. D., Tourtellot, M. K. & Snedden, W. A. 1997. Precedence effects and the evolution of chorusing. Proceedings of the Royal Society of London, Series B, 264, 1355–1361. Hulse, S. 2002. Auditory scene analysis in animal communication. Advances in the Study of Behavior, 31, 163–200. ˜ ez, D. R. 1993. Female phonotaxis and call overlap in the Neotropical glassfrog Ib´ an Centrolenella granulosa. Copeia, 1993, 846–850. Johnston, R. E., Sorokin, E. S. & Ferkin, M. H. 1997. Female voles discriminate males’ over-marks and prefer top-scent males. Animal Behaviour, 54, 679–690. Johnstone, R. A. 1994. Honest signalling, perceptual error and the evolution of ‘all-or-nothing’ displays. Proceedings of the Royal Society of London, Series B, 256, 169–175. Kelly, J. B. 1974. Localization of paired sound sources in the rat: small time differences. Journal of the Acoustical Society of America, 55, 1277–1284.

295

P1: JZZ/... P2: JZZ/... 0521823617c13.xml CU1917B/McGregor

296

0 521 582361 7

April 11, 2005

16:26

T. U. Grafe Kime, N. M., Turner, W. R. & Ryan, M. J. 2000. The transmission of advertisement calls in Central American frogs. Behavioral Ecology, 11, 71–83. Klump, G. M. 1996. Bird communication in the noisy world. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca, NY: Cornell University Press, pp. 321–338. Klump, G. M. & Gerhardt, H. C. 1987. Use of non-arbitrary acoustic criteria in mate choice by female gray tree frogs. Nature, 326, 286–288. 1992. Mechanisms and function of call-timing in male–male interactions in frogs. In: Playback and Studies of Animal Communication, ed. P. K. McGregor. New York: Plenum Press, pp. 153–174. Littlejohn, M. J. & Harrison, P. A. 1985. The functional significance of the diphasic advertisement call of Geocrinia victoriana (Anura: Leptodactylidae). Behavioral Ecology and Sociobiology, 16, 363–373. Lopez, P. T., Narins, P. M., Lewis, E. R. & Moore, S. W. 1988. Acoustically induced call modification in the white-lipped frog, Leptodactylus albilabris. Animal Behaviour, 36, 1295–1308. Marshall, V. T., Humfeld, S. C. & Bee, M. A. 2003. Plasticity of aggressive signalling in male spring peepers (Pseudacris crucifer). Animal Behaviour, 65, 1223–1234. McGregor, P. K. & Dabelsteen, T. 1996. Communication networks. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca, NY: Cornell University Press, pp. 409–425. McGregor, P. K. & Peake, T. M. 2000. Communication networks: social environments for receiving and signalling behaviour. Acta Ethologica, 2, 71–81. Moore, S. W., Lewis, E. R., Narins, P. M. & Lopez, P. T. 1989. The call-timing algorithm of the white-lipped frog, Leptodactylus albilabris. Journal of Comparative Physiology A, 164, 309–319. Murphy, C. G. & Gerhardt, H. C. 2002. Mate sampling by female barking treefrogs (Hyla gratiosa). Behavioral Ecology, 13, 472–480. Narins, P. M. 1982. Behavioral refractory period in neotropical treefrogs. Journal of Comparative Physiology A, 148, 337–344. Narins, P. M. & Capranica, R. R. 1978. Communicative significance of the two-note call of the treefrog Eleutherodactylus coqui. Journal of Comparative Physiology A, 127, 1–9. Narins, P. M. & Zelick, R. 1988. The effects of noise on auditory processing and behavior in amphibians. In: The Evolution of the Amphibian Auditory System, ed. B. Fritzsch, T. Hetherington, M. Ryan, W. Wilczynski & W. Walkowiak. New York: John Wiley, pp. 455–483. Narins, P. M., Lewis, E. R. & McClelland, B. E. 2000. Hyperextended call note repertoire of the endemic Madagascar treefrog Boophis madagascariensis (Rhacophoridae). Journal of Zoology, 250, 283–298. Passmore, N. I. 1981. Sound levels of mating calls of some African frogs. Herpetologica, 37, 166–171. Passmore, N. I., Bishop, P. J. & Caithness, N. 1992. Calling behaviour influences mating success in male painted reed frogs, Hyperolius marmoratus. Ethology, 92, 227–241.

P1: JZZ/... P2: JZZ/... 0521823617c13.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

16:26

Anuran choruses as communication networks Penna, M. & Solis, R. 1998. Frog call intensities and sound propagation in the South American temperate forest region. Behavioral Ecology and Sociobiology, 42, 371–381. Pfennig, K. S., Rapa, K. & McNatt, R. 2000. Evolution of male mating behavior: male spadefoot toads preferentially associate with conspecific males. Behavioral Ecology and Sociobiology, 48, 69–74. Robinson, D. 1990. Acoustic communication between the sexes in bushcrickets. In: The Tettigoniidae: Behavior, Systematics, Evolution, ed. W. J. Bailey & D. C. F. Rentz. Bathurst, Australia: Crawford House Press, pp. 110–129. R¨ omer, H., Bailey, W. J. & Dadour, I. 1989. Insect hearing in the field. III. Masking by noise. Journal of Comparative Physiology A, 164, 609–620. Ryan, M. J. 1985. The T´ ungara Frog. Chicago, IL: University of Chicago Press. 1986. Synchronized calling in a treefrog (Smilisca sila). Brain, Behavior and Evolution, 29, 196–206. Ryan, M. J. & Keddy-Hector, A. 1992. Directional patterns of female mate choice and the role of sensory biases. American Naturalist, 139, S4–S35. Ryan, M. J. & Wilczynski, W. 1991. Evolution of intraspecific variation in the advertisement call of a cricket frog (Acris crepitans). Biological Journal of the Linnean Society, 44, 249–271. Ryan, M. J., Cocroft, R. B. & Wilczynski, W. 1990. The role of environmental selection in intraspecific divergence of mate recognition signals in the cricket frog, Acris crepitans. Evolution, 44, 1869–1872. Schneider, H., Hoermann, G. & H¨ odl, W. 1988. Calling and antiphonal calling in four Neotropical anuran species of the family Leptodactylidae. Zoologische Jahrb¨ ucher Physiologie, 92, 77–103. Schwartz, J. J. 1989. Graded aggressive calls of the spring peeper, Pseudacris crucifer. Herpetologica, 45, 172–181. 1993. Male calling behavior, female discrimination and acoustic interference in the Neotropical treefrog Hyla microcephala under realistic acoustic conditions. Behavioral Ecology and Sociobiology, 32, 401–414. 2001. Call monitoring and interactive playback systems in the study of acoustic interactions among male anurans. In: Anuran Communication, ed. M. J. Ryan. Washington, DC: Smithsonian Institute Press, pp. 183–204. Schwartz, J. J. & Wells, K. D. 1984. Interspecific acoustic interactions of the neotropical treefrog Hyla ebraccata. Behavioral Ecology and Sociobiology, 14, 211–224. Schwartz, J. J., Buchanan, B. W. & Gerhardt, H. C. 2001. Female mate choice in the gray treefrog (Hyla versicolor) in three experimental environments. Behavioral Ecology and Sociobiology, 49, 443–455. Shelly, T. E. & Greenfield, M. D. 1991. Dominions and desert clickers (Orthoptera: Acrididae): influences of resources and male signaling on female settlement patterns. Behavioral Ecology and Sociobiology, 28, 133–140. Staaden, M. J. v. & R¨ omer, H. 1997. Sexual signalling in bladder grasshoppers: tactical design for maximizing calling range. Journal of Experimental Biology, 200, 2597–2608.

297

P1: JZZ/... P2: JZZ/... 0521823617c13.xml CU1917B/McGregor

298

0 521 582361 7

April 11, 2005

16:26

T. U. Grafe Sullivan, B. K., Ryan, M. J. & Verrell, P. A. 1995. Female choice and mating system structure. In: Amphibian Biology: Social Behaviour, ed. H. Heatwole & B. K. Sullivan. Chipping Norton, UK: Surrey Beatty, pp. 469–517. Summers, K. 1992. Dart-poison frogs and the control of sexual selection. Ethology, 91, 89–107. Taigen, T. L., Wells, K. D. & Marsh, R. L. 1985. The enzymatic basis of high metabolic rates in calling frogs. Physiological Zoology, 58, 719–726. Telford, S. R., Dyson, M. L. & Passmore, N. I. 1989. Mate choice occurs only in small choruses of painted reed frogs (Hyperolius marmoratus). Bioacoustics, 2, 47–53. Townsend, D. S. & Stewart, M. M. 1986. Courtship and mating behavior of a Puerto Rican frog, Eleutherodactylus coqui. Herpetologica, 42, 165–170. Tuttle, M. D. & Ryan, M. J. 1982. The role of synchronized calling, ambient light, and ambient noise, in anti-bat-predator behavior of a treefrog. Behavioral Ecology and Sociobiology, 11, 125–131. Vencl, F. V. & Carlson, A. D. 1998. Proximate mechanisms of sexual selection in the firefly Photinus pyralis (Coleoptera: Lampyridae). Journal of Insect Behavior, 11, 191–207. Verrell, P. A. & Brown, L. E. 1993. Competition among females for mates in a species with male parental care, the widwife toad Alytes obstetricans. Ethology, 93, 247–257. Walkowiak, W. 1992. Acoustic communication in the fire-bellied toad: an integrative neurobiological approach. Ethology, Ecology and Evolution, 4, 63–74. Wallach, H., Newman, E. B. & Rosenzweig, M. R. 1949. The precedence effect in sound localization. American Journal of Psychology, 62, 315–336. Wells, K. D. 1988. The effect of social interactions on anuran vocal behavior. In: The Evolution of the Amphibian Auditory System, ed. B. Fritzsch, M. J. Ryan, W. Wilczynski, T. E. Hetherington & W. Walkowiak. New York: John Wiley, pp. 433–454. 1989. Vocal communication in a Neotropical treefrog, Hyla ebraccata: responses of males to graded aggressive calls. Copeia, 1989, 461–466. 2001. The energetics of calling in frogs. In: Anuran Communication, ed. M. J. Ryan. Washington, DC: Smithsonian Institute Press, pp. 45–60. Wells, K. D. & Schwartz, J. J. 1984. Vocal communication in a Neotropical treefrog, Hyla ebraccata: advertisement calls. Animal Behaviour, 32, 405–420. Wells, K. D. & Taigen, T. L. 1986. The effect of social interactions on calling energetics in the gray treefrog (Hyla versicolor). Behavioral Ecology and Sociobiology, 19, 9–18. Widemo, F. & Owens, I. P. F. 1995. Lek size, male mating skew and the evolution of lekking. Nature, 373, 148–151. Wiley, R. H. 1991. Associations of song properties with habitat for territorial oscine birds of eastern North America. American Naturalist, 138, 973–993. Wiley, R. H. & Richards, D. G. 1978. Physical constraints on acoustic communication in the atmosphere: implications for the evolution of animal vocalizations. Behavioral Ecology and Sociobiology, 3, 69–94. Wyttenbach, R. A. & Hoy, R. R. 1993. Demonstration of the precedence effect in an insect. Journal of the Acoustical Society of America, 94, 777–784.

P1: JZZ/... P2: JZZ/... 0521823617c13.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

16:26

Anuran choruses as communication networks Zelick, R. D. & Narins, P. M. 1983. Intensity discrimination and the precision of call timing in two species of neotropical treefrogs. Journal of Comparative Physiology A, 153, 403–412. 1985. Characterization of the advertisement call oscillator in the frog Eleutherodactylus coqui. Journal of Comparative Physiology A, 156, 223–229.

299

P1: IRK/KFO P2: IRK-KOD 0521823617c14.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:1

14

Singing interactions in songbirds: implications for social relations and territorial settlement marc naguib University of Bielefeld, Germany

Introduction Interactions between individuals make up a significant part of life in social animals. They form a crucial behavioural mechanism establishing and maintaining particular spacing patterns among individuals and groups of individuals and are inherent in the regulation of social relations. Animals interact with each other in a broad range of contexts, such as during intersexual competition, mate choice, or parent–offspring communication, but still many of the underlying principles share common ground (Hauser, 1996; Bradbury & Vehrencamp, 1998). It is well documented that the performance of individuals in interactions has profound implications for the resolution of conflicts over resources, such as mates, food or space. Interactions may consist of complex behavioural displays or may be based exclusively on signals in either one or several signalling modalities. Vocal interactions are among the most conspicuous forms of interactions and have been well studied in several taxonomic groups, such as insects, anurans and birds (Bradbury & Vehrencamp, 1998). In birds, vocal interactions are most evident in parent– offspring communication (Kilner & Johnstone, 1997; Ch. 9), calling and singing in group-living species (Farabaugh & Dooling, 1996; Zann, 1996), duetting in tropical songbirds (von Helversen, 1980; Farabaugh, 1982) and in singing interactions between male territorial songbirds (Todt & Naguib, 2000). In this review, I will focus on singing interactions in male territorial songbirds. Their vocal interactions are among the most striking examples of bird vocal communication and are an established model for studies on territoriality and communication networks (McGregor, 1993; McGregor & Dabelsteen, 1996; Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

300

P1: IRK/KFO P2: IRK-KOD 0521823617c14.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:1

Singing interactions in songbirds Todt & Naguib, 2000). Male songbirds commonly hold adjacent territories forming neighbourhoods; so usually several males sing within signalling range of each other. The typical pattern of settlement and spacing of individuals of the same species thus sets the framework for the evolution of communication behaviour, including the evolution of complex patterns of vocal interactions and the strategies for gathering information on conspecifics. Singing interactions between males take place in a variety of different contexts and the information exchanged may strongly depend on the singers’ social and spatial relations. They take place during immediate competition over resources such as space or mates in addition to being a conspicuous component in the social relations between established territorial neighbours. In general, vocal interactions differ from the classical song traits such as singing activity (Kempenaers et al., 1997; Gil et al., 1999; Amrhein et al., 2002, 2004), singing versatility (Hasselquist et al., 1996; Searcy & Yasukawa, 1996) or other performance-related traits (Podos, 1996; Forstmeier et al., 2002). These classical traits can be regarded as ‘individual’ traits that are present regardless of the social context in which a male is singing (Fig. 14.1). Vocal interactions, in contrast, have an additional interactive dimension as the message conveyed depends strongly on the pattern of song interchange between the interacting singers (Todt & Naguib, 2000). This interactive dimension has resulted in considerable current research interest in vocal interactions, as the performance of singers during an interaction provides immediate information on relative differences between them: information that also is used by eavesdropping individuals in a communication network (McGregor & Dabelsteen, 1996). My principal goal in this chapter is to integrate current knowledge on strategies of vocal interactions in territorial songbirds with concepts of territorial behaviour and territorial settlement (Waser & Wiley, 1980; Stamps, 1994; Stamps & Krishnan, 2001) and to explore how recent advances in studies on vocal interactions contribute to our understanding of the social relations among neighbouring territorial songbirds. The social and spatial relationships among neighbouring males can be mediated by their vocal interactions; consequently, vocal interactions can have profound implications for the evolution of strategies for territorial settlement and spacing behaviour in general. I will also evaluate how principles of vocal interactions contribute to our understanding of the evolution of singing strategies and the evolution of receivers’adaptations to gather information from conspecific signalling behaviour. A key trait of vocal interactions responsible for much of the interest is that they are commonly asymmetric in the sense that each of the singers involved uses its songs differently in relation to those of its counterpart. These asymmetries can reflect differences in motivation or quality among singers and consequently provide information about the relationship between them (Todt & Naguib, 2000). Therefore, this chapter will consider the internal aspects of

301

P1: IRK/KFO P2: IRK-KOD 0521823617c14.xml CU1917B/McGregor

302

0 521 582361 7

April 11, 2005

17:1

M. Naguib Pathways of information gathering

Extracting information from individual signals

Individual song traits

Extracting information from interactions and their asymmetries (eavesdropping)

Social traits

Song structure Singing versatility

Vocal interactions

Singing rate

(e.g. asymmetries in vocal matching, or in timing of songs)

Diurnal singing activity

Fig. 14.1. Singing traits and pathways of information gathering. Receivers can gather information from individual traits during and in the absence of interactions. Information gathering by extracting information from signalling interactions (eavesdropping) provides additional information that cannot be extracted from individual song traits. In interactions, additional social factors come into play such as the relative relation of the songs, which can provide immediate information on differences between singers.

interactions (i.e. their function in the interaction between the singers) as well as their external implications (i.e. their wider importance as a source of information for other listening (eavesdropping) individuals; see also Ch. 2). After reviewing some recent studies of vocal interactions and by drawing several examples from our own studies on nightingales Luscinia megarhynchos (reviewed in more detail by Todt & Naguib, 2000), I will evaluate more closely the general association of social and spatial relationships between males and their strategies in vocal interactions, as well as their strategies in gathering information by attending to others’ vocal interactions.

Nature of interactions Interactions in communication can be defined as the exchange of information through signals by at least two individuals where the signals of both signallers have some direct relation to each other. In acoustic communication, interactions can be best determined when there are two individuals signalling; such interactions are referred to commonly as dyadic interactions. When several individuals are signalling, they may form a communication network with highly complex modes of interactions (see Chs. 13 and 15) but this need not be the case

P1: IRK/KFO P2: IRK-KOD 0521823617c14.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:1

Singing interactions in songbirds (Shackleton & Ratcliffe, 1994). Although interactions pragmatically can be defined broadly as instances when two or more singers are singing at the same time, true interactions are best identified by demonstrating that the singers influence each other in their choice and timing of song types or in other aspects of their singing strategy. Such a definition is analogous to our understanding of dialogues in human speech and allows us to extract and define specific strategies and to study their causations and evolutionary implications. Song matching

Most songbirds have song repertoires, allowing them flexibility in decisions on the next song to be sung (Kroodsma, 1982). The most conspicuous way of using specific song types during an interaction is song matching, a situation in which a singer replies with a song of the same type as the preceding song sung by the opponent. Such matching of signals is also found in other taxa (Ch. 18) and can give insights into the functions of signalling strategies as well as addressing mechanistic questions such as how animals perceive and categorize signals (Falls et al., 1982, 1988; Weary et al., 1990; Naguib et al., 2002). Song matching is known to be used to address a particular rival and often it appears to function as an aggressively directed signal to increase the level of threat towards a specific rival (Krebs et al., 1981; McGregor et al., 1992; Nielsen & Vehrencamp, 1995; Vehrencamp, 2001). However, several studies have shown that song matching is not always a strong aggressive signal and may be used as a graded signal of intent (Searcy et al., 2000; Burt et al., 2002; Naguib et al., 2002). Moreover, males may sometimes match some, but not all, features of songs. Males may match only parts of a song or certain song parameters such as the frequency of the full song (Otter et al., 2002) or specific song components (Burt et al., 2002; Naguib et al., 2002), duration of songs (Weary et al., 1990) or categories of song (Wiley et al., 1994; Naguib et al., 2002). Neighbouring song sparrows Melospiza melodia have been shown to match repertoires by replying with non-matching songs that are shared with the singing opponent (Beecher et al., 1996). The meaning of matching also may vary with the distance between singers, the general context, the song type or even the specific song component that is matched. In nightingales, we discovered patterns of song matching that clearly differed from the most widespread principle that song matching increases with the level of perceived threat (Naguib et al., 2002). In playbacks conducted on nightingales’ nocturnal song, males increased the precision of matching the pitch of whistles in so-called whistle songs (Fig. 14.2) with increasing distance to the simulated unfamiliar opponent but not the overall rate of matching the song category. Matching whistle songs and specifically the pitch of the whistles may have a different biological significance than matching ‘normal’ songs. The narrow spectral bandwidth of the whistles implies that they transmit with less

303

P1: IRK/KFO P2: IRK-KOD 0521823617c14.xml CU1917B/McGregor

304

0 521 582361 7

April 11, 2005

17:1

M. Naguib (a) 1

2

1

kHz 10 8 6 4 2 2

4

6

s

(b) kHz 10 8 6 4

2 1

2 2

4 s

Fig. 14.2. Spectrographic examples of whistle song matching and overlapping in nightingales. Hatched arrows indicate beginning of songs, numbers indicate different males. (a) The whistle song in the centre from male 2 is a non-overlapping, full song type match of the first song of male 1, i.e. the whistle part and terminal sections are the same in both songs and there is no noticeable overlap of songs. Male 1 replied immediately to match and clearly overlap his opponent. (b) The two whistle songs match but also differ in frequency so that the whistle parts do not mask each other.

spectral degradation over long distances and thus may function particularly in long-range signalling (Wiley & Richards, 1978; Slabbekoorn et al., 2002). At first glance, more matching at long distance is puzzling as the social importance and, therefore, the urgency of addressing a rival is assumed to decrease with interindividual spacing. However, even distant neighbours that do not share a territorial boundary are part of the same neighbourhood, in which males have to establish and maintain social and spatial relations that are likely to be regulated through long-range vocal interactions. Radio-tracking data have shown that males make substantial excursions into territories of direct neighbours and even more distant ones within auditory range from the own territory (Hanski, 1992; Chandler et al., 1997; Pitcher & Stutchbury, 2000; Naguib et al., 2001), indicating that the social and spatial relations between males go beyond their immediate neighbours with whom territorial boundaries are shared. Long-distance matching, therefore, may be a mechanism involved directly in establishing and maintaining spatial relations between males.

P1: IRK/KFO P2: IRK-KOD 0521823617c14.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:1

Singing interactions in songbirds In summary, the extensive research on song matching (reviewed more fully by Todt & Naguib, 2000) shows that matching can do more than simply address a rival; it can have several functions, depending on the social context, the territorial relationships of the interacting males and the song type or its specific parameters that are matched. It will be interesting to see how the refinements of playback design and playback technologies (e.g. Dabelsteen, 1992) will address more subtle questions on the function of song matching in different contexts and thus provide more detailed insights into the kinds of message conveyed. Timing of songs

In a singing interaction, the relative timing of song production by the interactants can signal specific information on the singer’s state and intention, such as its readiness to escalate the contest (Todt & Naguib, 2000). The relative timing of songs during an interaction differs from song matching in two ways. First, song sharing or detailed knowledge of a rival’s repertoire is not required. Second, relative song timing can vary continuously whereas song matching is more categorical (i.e. matching occurs or does not occur) even though recent research has emphasized that song matching can be subtle with graded components and that it is not restricted to matching full song types, as discussed above. Despite the continuous nature of relative song timing, two categories of timing of songs have been shown to occur to date (Hultsch & Todt, 1982) and to have functionally different signal value: song alternating and song overlapping (Brindley, 1991; Dabelsteen et al., 1996; Naguib et al., 1999; Langemann et al., 2000). Song alternating is a common strategy where males take turns in delivering their songs. This strategy can also be observed in concurrent interspecific singing where males avoid acoustic competition (Ficken et al., 1974). During song overlapping, in contrast, males begin to sing a song before the opponent has ended its song. Song overlapping has been shown in several species to function as a directed agonistic signal (Brindley, 1991; McGregor et al., 1992; Dabelsteen et al., 1996, 1997; Naguib, 1999) whereas song alternating is the seemingly predominating singing strategy during less-intense contexts. Interestingly, song overlapping is not only treated as an agonistic signal by the singer whose songs are overlapped but also is used by eavesdropping males and females to assess differences in the relative quality or motivation in two interacting males (Naguib & Todt, 1997; Naguib et al., 1999; Otter et al., 1999; Peake et al., 2001, 2002; Mennill et al., 2002), as discussed below (see also Ch. 2). Song overlapping can result in considerable masking of part of the rival’s song, but this effect is not inherent in song overlapping. Masking effects will depend on the distance between singers (i.e. their relative difference in amplitude), the amount of song overlapped in terms of duration and the similarity in phonology between the overlapping parts of the two songs (Fig. 14.2; see also

305

P1: IRK/KFO P2: IRK-KOD 0521823617c14.xml CU1917B/McGregor

306

0 521 582361 7

April 11, 2005

17:1

M. Naguib Ch. 20). In addition, the extent of perceived masking will vary with whose perception is considered: the singer whose songs are overlapped, the overlapper or other individuals, which may be at any relative location to the two singers (see below). As a result, benefits through masking the opponent’s songs may be limited to certain conditions and locations and, despite the attractiveness of the argument, it is unlikely to be the primary consequence that led to the evolution of song overlapping as a singing strategy in long-distance interactions. Although overlap will affect detection and recognition of subtle sound features (Wiley, 1983, 1994) by other listening conspecifics, spatial release from masking (Klump, 1996) is a compensatory mechanism that can help in coping with problems resulting from masking. Todt and Naguib (2000) further suggested that a male that is overlapping the songs of its opponent might benefit by shifting the attention of eavesdroppers to the overlapper. Finally, despite such a range of basic effects of song overlap on signal perception, song overlapping may also be a conventional signal of dominance that is maintained by retaliation costs if overlapping increases the probability to escalate a contest. Song overlapping and song alternating have been shown to be of biological significance, but the issue of how much overlap needs to be achieved in order to accomplish a certain function remains to be studied. Similarly, the precise timing of songs during alternating singing and also during song overlapping in natural conditions may have specific signal value (McGregor et al., 1992), another issue that deserves to be explored in more detail in future studies. For instance, some studies suggest that song alternating is not a homogeneous strategy but that the exact timing during alternating is also of functional significance. Specifically, leader–follower relationships, in which the follower sings soon after the leader, have been interpreted as the leader representing the more dominant singer (Smith & Norman, 1979; Popp, 1989; Naguib et al., 1999). Gathering information on very fine temporal differences in timing of songs will require knowledge of the distance to the opponent (Naguib & Wiley, 2001) and eavesdroppers will require to know the distance to each interactant because of the different time delay of the songs originating from two sources at unequal distance. However, confusion over whether or not a song is overlapping can occur only in long-range interactions in two extremes: when both singers start their song at about the same time (so that each singer will perceive the opponent’s song as overlapping) or when an overlapping song sets on late during the song that is overlapped (so that the opponent may receive it as non-overlapping). It remains to be studied if overlapping events that fall into this ‘confusion range’ are interpreted differently from unambiguous events. In nightingales, overlapping songs most commonly fall outside these confusion ranges (Hultsch & Todt, 1982; Naguib, 1999); consequently, in most cases overlapping is an unambiguous event.

P1: IRK/KFO P2: IRK-KOD 0521823617c14.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:1

Singing interactions in songbirds Matching and timing

In the preceding sections, we have seen that both song matching and song timing are characteristic of bird vocal interactions and in many cases are related to a specific meaning of singing. However, the possibility that both may be dependent on each other (Todt, 1981; Wolffgramm & Todt, 1982) has been little explored from a functional perspective. There are several possible combinations of these two aspects. For example, males may match and overlap a rival’s song, as nightingales frequently do when matching whistle songs (Fig. 14.2), or they may match a song with varying delays during alternating singing. Matching songs during boundary disputes often occurs in interactions with high song rates and thus short delays in responses. Matching with short delays may be of specific value in signalling the willingness to escalate a contest. When males interact where no immediate dispute is apparent, such as during the dawn chorus or with low song rates in long-range interactions between neighbours that sing at the same time of day or night, matching may be timed differently and, therefore, may have a different meaning and consequence for the social relations between singers. One possibility is that frequent matching between established territory holders acts to repel non-territorial males seeking to establish a territory (Amrhein et al., 2004) (thus benefiting both territorial males) by signalling longterm territory tenure and an established spatial and social relationship. There is evidence that males sharing song types have longer territory tenure (Beecher et al., 2000) and neighbouring males often share more songs than non-neighbouring males (Kroodsma, 1974; Hultsch & Todt, 1981; McGregor & Krebs, 1982; Schroeder & Wiley, 1983; Beecher, 1996; Payne, 1996; Beecher et al., 2000; Griessmann & Naguib, 2002): both features that make song matching more likely. An increase in song sharing over the season, as shown for thrush nightingales Luscinia luscinia (Sorjonen, 1987), may increase the probability of matching during vocal interactions as the season progresses. If so, matching during long-range vocal interactions such as the thrush nightingales’ nocturnal song may function to strengthen territorial residency in neighbouring males rather than being an agonistic signal. Vocal interactions and social relationships among singers Songbirds interact with song in at least five different social contexts that need to be considered when singing strategies and their evolutionary implications are studied (Fig. 14.3). Territorial males interact by song with; (a) neighbours over long distances when there is no immediate dispute noticeable (Fig. 14.3a); (b) neighbours in immediate disputes over territorial boundaries or possibly over access to females (Fig. 14.3b); (c) neighbours that have crossed the shared territorial boundary (Fig. 14.3c); (d) unfamiliar rivals that have intruded into the territory

307

P1: IRK/KFO P2: IRK-KOD 0521823617c14.xml CU1917B/McGregor

308

0 521 582361 7

April 11, 2005

17:1

M. Naguib

Fig. 14.3. Five contexts of vocal interactions between male songbirds. Territories are represented by elliptical shapes. (a) Long-range interaction between established neighbours (filled and open circles); (b) boundary interaction between established neighbours with both males singing in their own territory; (c) interaction in which the resident male of the right-hand territory has intruded; (d) interaction between a territory holder and a non-territorial stranger (hatched circle) that has intruded into the territory; (e) interaction between a territory holder and a distant stranger (hatched circle) that may attempt to establish a territory nearby.

P1: IRK/KFO P2: IRK-KOD 0521823617c14.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:1

Singing interactions in songbirds and start vocally claiming part of it (Fig. 14.3d); or (e) unfamiliar rivals that attempt to establish a territory nearby without directly threatening the resident’s territory (Fig. 14.3e). Singing strategies in these different contexts are under different selection pressures as the social context differs; accordingly, the implications of specific singing strategies will vary in these situations. In encounters between a resident male and an intruder, asymmetries in site-specific dominance are inherent, as the payoff for each male differs because of the prior investment of the resident male in establishing and maintaining a territory before the contest (Maynard Smith & Parker, 1976; Waser & Wiley, 1980). Interactions with strangers are often single and time-limited events and, therefore, males clearly have to signal their strength and should signal a higher readiness to escalate the contest. Territorial residents are more likely to win a contest than intruders, so it is adaptive for residents to invest more in the interaction (Pusey & Packer, 1997). Interactions among neighbours, in contrast, are repeated and it may pay males to use a different singing strategy. Moreover, the asymmetries in site-specific dominance that is evident in all encounters between residents and intruders does not apply to neighbour–neighbour interactions, provided that both are singing from within their territories at locations that are not under direct dispute. The exact ways males interact with their neighbours will depend, therefore, on their locations and on previous experience of the dyad (Wiley & Wiley, 1980). Interactions between neighbours, as between residents and intruders, are still characterized by asymmetries, presumably as a result of inherent differences among males, such as differences in age, duration of prior residency or mating status. Remaining asymmetries in status between territorial neighbours can then well be reflected in the way they interact with each other vocally. Moreover, males may develop specific expectations when interacting with specific neighbours because of the specific ontogenetic trajectory of their relationship; consequently a given singing strategy may have a different functional significance when used with neighbours than when it is used with strangers. Biologically, the interactions among neighbours are particularly interesting as they presumably reflect the social relationships between them and, therefore, provide deeper insights into the territorial and social system of songbirds. Experiments that focused explicitly on vocal interactions have simulated a stranger’s or neighbour’s intrusion into a territory (or appearance near the territory), simulating the contexts illustrated in Fig. 14.3d,e (Todt & Naguib, 2000). Vocal interactions in such high-intensity contexts have served as an important experimental model to unravel the functions of specific singing strategies during an interaction. In general, such situations are characterized by high song rates, song overlap, song matching and song switching (in species in which males usually repeat the same song type several times before switching to a different one).

309

P1: IRK/KFO P2: IRK-KOD 0521823617c14.xml CU1917B/McGregor

310

0 521 582361 7

April 11, 2005

17:1

M. Naguib Depending on locations of simulated intrusions, territorial males distinguish between neighbours and strangers in their responses (Falls, 1982; Stoddard, 1996). Such individual recognition is the prerequisite for different response strategies with specific neighbours and new rivals and raises issues about how interactions vary in dynamics and functions with the familiarity of the singers and their spatial and social relationship. The information on functions and principles of interactions obtained from these playback experiments simulating intrusions are likely to be of general value and applicability also to long-range interactions among established neighbours that are interacting when singing on their own territory. However, it is important to consider that neighbours can be expected to exchange much more subtle information in vocal interactions in the absence of an immediate dispute. Neighbours having prior experience with each other may be better at using nuances in variation of singing patterns, such as song rate, quality of sound production or use of specific song variants or song types. Communication among established males may thus reach a much higher level of complexity with higher cognitive demands than communication among unfamiliar males, where disputes are driven by more immediate contests in specific contexts over specific resources. More descriptive and experimental studies on long-range vocal interactions in the absence of immediate disputes will be needed to test these ideas further. Functions of vocal interactions in territorial defence against intruders

The functions of vocal interactions among residents and intruders are obvious as there is an immediate conflict over space. Vocal interactions here make up a significant fraction of the behaviour during such conflicts, underlying their importance in spacing behaviour. Asymmetries in these interactions may be an important predictor for subsequent behaviour over the spatial conflict and, therefore, may set the stage for the occurrence of subsequent and intermittent movements and the probability of physical encounters. Vocal interactions commonly escalate in intensity in immediate disputes over territorial boundaries. Consequently, most intense interactions can be observed early in the season when territories are established, whenever boundaries are violated by neighbours, and when males attempt to establish a new territory in an area with males that have been resident for some time. In these situations, territory holders are highly aroused and attempt to drive the rival from the disputed area by intense singing and an interactive strategy that signals high readiness for escalation. Commonly, males then sing at a high song rate and interactions are characterized by frequent song overlapping and high rates of song matching or song switching, as discussed above. Most experimental research on the function of singing during vocal contests has used playback simulating an unfamiliar

P1: IRK/KFO P2: IRK-KOD 0521823617c14.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:1

Singing interactions in songbirds intruder and demonstrated that song matching, song switching (not reviewed here) and song overlapping are strategies used and perceived as an agonistic signalling behaviour, as discussed above. Therefore, these singing variables are likely to determine the outcome of a contest over space. Future playback experiments that focus on consequences of song overlap and song matching in terms of choice of song posts by the opponent will be important to address this issue in more detail. Males may avoid singing at posts where they provoke intense responses, for example being overlapped (e.g. Todt, 1981) and challenged by high song rates, and retreat earlier when their singing evokes such responses by resident males. Functions of vocal interactions among neighbouring males

Songbirds sing extensively in the phase when territories are established but continue to interact vocally when conflicts over space become less intense, that is when territories appear to be established. The functions of these continuing vocal interactions must have a different evolutionary significance, as their outcome is less likely to have drastic effects in conflicts over space. Territorial neighbours exhibit site-specific dominance; therefore, their vocal interactions are not associated with spatial asymmetries unless intrusions take place. In these situations, when vocal interactions are unlikely to function to resolve immediate conflicts over space, such interactions are more likely to function in maintaining a spacing pattern and keeping remaining asymmetries of the territorial neighbours at an equilibrium that avoids conflicts in which no clear winner is likely to emerge. Information on conspecifics will be imperfect; consequently, neighbouring males may need to continuously update their information on neighbours to refine their assessment. Therefore, after the spatial arrangement in a territorial system becomes established, vocal interactions occur most frequently between familiar neighbouring males singing from their own territories in the absence of immediate boundary disputes. However, long-range interactions among neighbours are much less well studied than interactions in high-intensity contexts (Kramer & Lemon, 1983; Kramer et al., 1985). Factors that are important to consider when assessing the function of specific singing strategies are that basic principles of singing are likely to depend on (a) the specific prior relationship of the singers, (b) the current singers’ relationship and, (c) the expected future relationship. Many of the basic principles of singing, such as song matching and strategies of song timing, may have the same function regardless of the singers’ specific relationship as long as they reflect the singers’ internal states. However, the interpretation of a singing strategy may differ depending on whether or not the singers are familiar with each other: whether or not they are territorial neighbours. Neighbouring males often already have substantial previous experience with each other and have to expect a long-term relationship. Singing

311

P1: IRK/KFO P2: IRK-KOD 0521823617c14.xml CU1917B/McGregor

312

0 521 582361 7

April 11, 2005

17:1

M. Naguib strategies during their interactions may be different as the strategy is an integral component of the previous relationship and that expected in the future. Thus, males may be able to code information more subtly in choice of song patterns or in the timing of songs. Males who have established dominance over rivals through previous interactions may not escalate substantially during subsequent interactions. Rather a few occurrences of song overlap (for instance) may suffice to signal alertness or the readiness to escalate a contest. Therefore, the function and the consequences of singing strategies during vocal interactions among neighbours may depend not only on the current singing strategy but also on how it relates to the strategy in previous interactions. By changing song posts during vocal interactions with neighbours, males may probe each other. If males avoid song posts at which they encounter repeated vocal aggression, not only territory boundaries but also the choice of song posts may be determined by neighbours’ singing behaviour during vocal interactions. Overall, more descriptive studies on the nature of vocal interactions between established males in relation to their spatial relationships will be needed to answer questions on how vocal interactions reflect and affect the social relations between males in long-term spatial relationships.

Vocal interactions and territorial settlement The role of vocal interactions in territorial settlement is particularly interesting when singing strategies during vocal interactions reflect the males’qualities or their motivation to defend a particular space. Although the function of song in territorial defence is well established once a male has an established territory (Krebs et al., 1978; Nowicki et al., 1998; Naguib et al., 2001), there has been little discussion of how song and vocal interactions determine spatial relations of males during territorial settlement or when territories shift in the course of the breeding season. Interpretations that particular singing strategies reflect a winner (Peake et al., 2001, 2002) can apply to single interactions, but single interactions may not necessarily reflect the overall relationship between the singers. For example, the singing behaviour of individuals establishing territories may differ from that when they are defending an established territory. The terms winner and loser in competition over resources usually imply that one individual gets all (or has first access) and the other gets nothing (retreats). Stamps and Krishnan (1997, 1999, 2001) pointed out that during territorial settlement the consequences of winning and losing contests are much more complex. During territorial establishment, individuals are dividing up space and such division of space usually involves repeated interactions from different locations. Therefore, the outcome of vocal interactions may be determined by the net outcome of repeated interactions at different locations rather than a single interaction. This situation presumably predominates

P1: IRK/KFO P2: IRK-KOD 0521823617c14.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:1

Singing interactions in songbirds during establishment of territories in areas that are not yet saturated and where both interacting males can expect to succeed in establishing a territory in a particular area. In some instances, males may even divide up space without a definite winner. By shifting song posts, space may be divided up passively when males avoid song posts in which they elicit high-intensity interactions with their rival. In this way, losing an interaction in territorial conflicts does not necessarily mean that losers fully retreat but rather that they shift song posts and they may be the winner at a different location. The final spatial arrangement of singing territories may be determined by the pattern of repeated vocal interactions with males singing at different song posts. Vocal interactions between established neighbours may reveal information on remaining differences among them that does not result from asymmetries in site-dependent dominance, where each male may be the winner at a certain location. In nightingales, males may systematically ‘win’ repeated interactions with a particular neighbour, but this pattern is not true for all neighbouring males (M. Naguib, unpublished data). If recurring asymmetries exist in the interactions between particular males, this may reflect a stable dominance relationship that is maintained after space is divided up.

Vocal interactions in communication networks Vocal interactions have received particular attention in recent years as their asymmetries have been shown to be used by other individuals as a source of information. Vocal interactions in songbirds are a clearly defined signalling context and so have become one of the main models in studies of communication networks (McGregor & Dabelsteen, 1996; McGregor & Peake, 2000; Whitfield, 2002). To date, several studies have shown that male songbirds eavesdrop on rival vocal interactions and have expanded the understanding of information gathering when several individuals are within signalling rage of each other (Naguib & Todt, 1997; Naguib et al., 1999; Otter et al., 1999; Peake et al., 2001, 2002; Mennill et al., 2002). In our own studies, we showed in two-loudspeaker experiments that territorial nightingales discriminated between asymmetries in vocal interactions: subjects responded significantly more strongly to a simulated rival that was overlapping the songs of the opponent, i.e. was the more aggressive intruder (Naguib & Todt, 1997). When songs in the interaction simulated by the loudspeakers did not overlap each other but were played in an alternating order with songs of one loudspeaker leading (closely followed by the songs of the other loudspeaker), subjects responded more strongly to the loudspeakers playing the leading songs (Naguib et al., 1999). These combined experiments indicate that different proximate cues were used depending on the kind of asymmetry simulated. The use of opposing proximate cues by subjects depending on the kind of asymmetry perceived

313

P1: IRK/KFO P2: IRK-KOD 0521823617c14.xml CU1917B/McGregor

314

0 521 582361 7

April 11, 2005

17:1

M. Naguib provides evidence that subjects use an adaptive strategy that is independent of a putative general proximate shift of attention to a first or last heard stimuli. Peake and coworkers (2001) elegantly expanded this playback design and showed that great tits Parus major attended to such asymmetries in interactions played to them from outside their territory and then responded differently to subsequent intrusions, depending on which previous singer was simulated as intruder. They further showed that males varied their song output in responses to intruders depending on the kind of experience they had with the intruder prior to an interaction with another male (Peake et al., 2002). This suggests more complex ways of gathering information than shown in any of the previous experiments on eavesdropping. Studies by Otter et al. (1999) and Mennill et al. (2002) indicated that females also use asymmetries in male–male interactions as sources of information in their responses and seemingly even in reproductive decisions (see also Ch. 7). Therefore, it is well documented that songbirds not only attend to vocal interactions between males but also extract information coded in the asymmetries in singing strategies and use the information adaptively. An interesting question to be answered in order to understand the further implications of eavesdropping in communication networks with widely spaced individuals is whether birds are able accurately to extract meaningful asymmetries in an interaction occurring at a distance, or whether they are only able to do so when they are close to, or at equal distance from, the singers. We already have clear evidence that vocal interactions have much wider implications than the exchange of information between interacting males. Strategies of singing during vocal interactions then are likely not only to evolve through responses of the opponent but also through effects on other listening individuals (Ch. 2). Viewing communication from the perspective of communication networks considerably broadens the view of social implications of the prevalence of vocal interactions and the functional implications of certain singing strategies during vocal interactions. Most experimental studies of the functional significance of different singing strategies during an interaction have used different playback protocols and there is a need to complement these studies by more descriptive studies on natural interactions.

Summary and future directions Vocal interactions in songbirds have many facets that need to be integrated into models of information gathering in communication. Because of the accessibility of song, vocal interactions are a suitable model to address wider concepts in communication, to address questions in cognitive ecology (Ch. 24) and to obtain new insights into the social relationships between territorial neighbours. Singing strategies during vocal interactions and strategies for information gathering from vocal interactions by participants and by eavesdroppers have

P1: IRK/KFO P2: IRK-KOD 0521823617c14.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:1

Singing interactions in songbirds evolved under ecological constraints, such as the spacing patterns of conspecifics or more specifically the abundance and distribution of signallers in space and time. As we continue to gain more information on strategies of vocal interactions in natural settings among established neighbours and during establishment of territories, we are likely to obtain new insights into how social relations are mediated by song and how strategies of singing interactions are related to spatial ecology.

Acknowledgements I thank Peter McGregor, Ken Otter and an anonymous referee for helpful comments on a previous version of the manuscript

References Amrhein, V., Korner, P. & Naguib, M. 2002. Nocturnal and diurnal singing activity in the nightingale: correlations with mating status and breeding cycle. Animal Behaviour, 64, 939–944. Amrhein, V., Kunc, H. P. & Naguib, M. 2004. Non-territorial nightingales prospect territories during the dawn chorus. Proceedings of the Royal Society of London, Series B, (Supplement), 271, S167–S169. Beecher, M. D. 1996. Bird song learning in the laboratory and field. In: Ecology and Evolution in Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca, NY: Cornell University Press, pp. 61–78. Beecher, M. D., Stoddard, P. K., Campbell, S. E. & Horning, C. L. 1996. Repertoire matching between neighbouring song sparrows. Animal Behaviour, 51, 917–923. Beecher, M. D., Campbell, S. E. & Nordby, J. C. 2000. Territory tenure in song sparrows is related to song sharing with neighbours, but not to repertoire size. Animal Behaviour, 59, 29–37. Bradbury, J. W. & Vehrencamp, S. L. 1998. Principles of Animal Communication. Sunderland, MA: Sinauer. Brindley, E. L. 1991. Response of European robins to playback of song: neighbour recognition and overlapping. Animal Behaviour, 41, 503–512. Burt, J. M., Bard, S. C., Campbell, S. E. & Beecher, M. D. 2002. Alternative forms of song matching in song sparrows. Animal Behaviour, 63, 1143–1151. Chandler, C. R., Ketterson, E. D. & Nolan, V. 1997. Effects of testosterone on use of space by male dark-eyed juncos when their mates are fertile. Animal Behaviour, 54, 543–549. Dabelsteen, T. 1992. Interactive playback: a finely tuned response. In: Playback and Studies of Animal Communication, ed. P. K. McGregor. New York: Plenum Press, pp. 97–109. Dabelsteen, T., McGregor, P. K., Shepherd, M., Whittaker, X. & Pedersen, S. B. 1996. Is the signal value of overlapping different from that of alternating during matched singing in great tits? Journal of Avian Biology, 27, 189–194.

315

P1: IRK/KFO P2: IRK-KOD 0521823617c14.xml CU1917B/McGregor

316

0 521 582361 7

April 11, 2005

17:1

M. Naguib Dabelsteen, T., McGregor, P. K., Holland, J., Tobias, J. A. & Pedersen, S. B. 1997. The signal function of overlapping singing in male robins. Animal Behaviour, 53, 249–256. Falls, J. B. 1982. Individual recognition by sounds. In: Acoustic Communication in Birds, Vol. 2, ed. D. E. Kroodsma & E. H. Miller. New York: Academic Press, pp. 237–278. Falls, J. B., Krebs, J. R. & McGregor, P. K. 1982. Song matching in the great tit (Parus major): the effect of similarity and familiarity. Animal Behaviour, 30, 997–1009. Falls, J. B., Horn, A. G. & Dickinson, T. E. 1988. How western meadowlarks classify their songs: evidence from song matching. Animal Behaviour, 36, 579–585. Farabaugh, S. M. 1982. The ecological and social significance of duetting. In: Acoustic Communication in Birds, Vol. 2, ed. D. E. Kroodsma & E. H. Miller. New York: Academic Press, pp. 85–124. Farabaugh, S. M. & Dooling, R. J. 1996. Acoustic communication in parrots: laboratory and field studies of budgerigars, Melopsittacus undulatus. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca, NY: Cornell University Press, pp. 97–117. Ficken, R. W., Ficken, M. S. & Hailman, J. P. 1974. Temporal pattern shifts to avoid acoustic interference in singing birds. Science, 183, 762–763. Forstmeier, W., Kempenaers, B., Meyer, A. & Leisler, B. 2002. A novel song parameter correlates with extra-pair paternity and reflects male longevity. Proceedings of the Royal Society of London, Series B, 269, 1479–1485. Gil, D., Graves, J. A. & Slater, P. J. B. 1999. Seasonal patterns of singing in the willow warbler: evidence against the fertility announcement hypothesis. Animal Behaviour, 58, 995–1000. Griessmann, B. & Naguib, M. 2002. Song sharing in neighboring and non neighboring thrush nightingales (Luscinia luscinia) and its implications for communication. Ethology, 108, 377–387. Hanski, I. K. 1992. Territorial behaviour and mixed reproductive strategy in the chaffinch. Ornis Scandinavica, 23, 475–482. 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. Hauser, M. D. 1996. The Evolution of Communication. Cambridge, MA: MIT Press. Hultsch, H. & Todt, D. 1981. Repertoire sharing and song-post distance in nightingales (Luscinia megarhynchos B.). Behavioral Ecology and Sociobiology, 8, 183–188. 1982. Temporal performance roles during vocal interactions in nightingales (Luscinia megarhynchos). Behavioral Ecology and Sociobiology, 11, 253–260. Kempenaers, B., Verheyren, G. R. & Dhondt, A. A. 1997. Extrapair paternity in the blue tit (Parus caeruleus): female choice, male characteristics, and offspring quality. Behavioral Ecology, 8, 481–492. Kilner, R. & Johnstone, R. A. 1997. Begging the question: are offspring solicitation behaviours signals of need? Trends in Ecology and Evolution, 12, 11–15. Klump, G. 1996. Bird communication in the noisy world. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca, NY: Cornell University Press, pp. 321–338.

P1: IRK/KFO P2: IRK-KOD 0521823617c14.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:1

Singing interactions in songbirds Kramer, H. G. & Lemon, R. E. 1983. Dynamics of territorial singing between neighboring song sparrows (Melospiza melodia). Behaviour, 85, 198–223. Kramer, H. G., Lemon, R. E. & Morris, M. J. 1985. Song switching and agonistic stimulation in the song sparrow (Melospiza melodia). Animal Behaviour, 33, 135–149. Krebs, J. R., Ashcroft, R. & Webber, M. 1978. Song repertoires and territory defense in the great tit Parus major L. Nature, 271, 539–542. Krebs, J. R., Ashcroft, R. & van Orsdol, K. 1981. Song matching in the great tit Parus major L. Animal Behaviour, 29, 918–921. Kroodsma, D. E. 1974. Song learning, dialects and dispersal in the Bewick’s wren. Zeitschrift f¨ ur Tierpsychologie, 35, 352–380. 1982. Song repertoires: problems in their definitions and use. In: Acoustic Communication in Birds, Vol. 1, ed. D. E. Kroodsma & E. H. Miller. New York: Academic Press, pp. 125–146. Langemann, U., Tavares, J. P., Peake, T. M. & McGregor, P. K. 2000. Response of great tits to escalating patterns of playback. Behaviour, 137, 451–471. Maynard Smith, J. & Parker, G. A. 1976. The logic of asymmetric contests. Animal Behaviour, 24, 159–175. McGregor, P. K. 1993. Signalling in territorial systems: a context for individual identification, ranging and eavesdropping. Philosophical Transactions of the Royal Society of London, Series B, 340, 237–244. McGregor, P. K. & Dabelsteen, T. 1996. Communication networks. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca, NY: Cornell University Press, pp. 409–425. McGregor, P. K. & Krebs, J. R. 1982. Mating and song types in the great tit. Nature, 297, 60–61. McGregor, P. K. & Peake, T. M. 2000. Communication networks: social environments for receiving and signalling behaviour. Acta Ethologica, 2, 71–81. McGregor, P. K., Dabelsteen, T., Shepherd, M. & Pedersen, S. B. 1992. The signal value of matched singing in great tits: evidence from interactive playback experiments. Animal Behaviour, 43, 987–998. Mennill, D. J., Ratcliffe, L. M. & Boag, P. T. 2002. Female eavesdropping on male song contests in songbirds. Science, 296, 873–873. Naguib, M. 1999. Effects of song overlapping and alternating on nocturnally singing nightingales. Animal Behaviour, 58, 1061–1067. Naguib, M. & Todt, D. 1997. Effects of dyadic vocal interactions on other conspecific receivers in nightingales. Animal Behaviour, 54, 1535–1543. Naguib, M. & Wiley, R. H. 2001. Estimating the distance to a source of sound: mechanisms and adaptations for long-range communication. Animal Behaviour, 62, 825–837. Naguib, M., Fichtel, C. & Todt, D. 1999. Nightingales respond more strongly to vocal leaders in simulated dyadic interactions. Proceedings of the Royal Society of London, Series B, 265, 537–542. Naguib, M., Altenkamp, R. & Griessmann, B. 2001. Nightingales in space: song and extra-territorial forays of radio tagged song birds. Journal f¨ ur Ornithologie, 142, 306–312.

317

P1: IRK/KFO P2: IRK-KOD 0521823617c14.xml CU1917B/McGregor

318

0 521 582361 7

April 11, 2005

17:1

M. Naguib Naguib, M., Mundry, R., Hultsch, H. & Todt, D. 2002. Responses to playback of whistle songs and normal songs in male nightingales: effects of song category, whistle pitch, and distance. Behavioral Ecology and Sociobiology, 52, 216–223. Nielsen, B. M. B. & Vehrencamp, S. L. 1995. Responses of songs sparrows to song-type matching via interactive playback. Behavioral Ecology and Sociobiology, 37, 109–117. Nowicki, S., Searcy, W. A. & Hughes, M. 1998. The territory defense function of song in song sparrows: a test with the speaker occupation design. Behaviour, 135, 615–628. Otter, K., McGregor, P. K., Terry, A. M. R. et al., 1999. Do female great tits (Parus major) assess males by eavesdropping? A field study using interactive song playback. Proceedings of the Royal Society of London, Series B, 266, 1305–1309. Otter, K. A., Ratcliffe, L., Njegovan, M. & Fotheringham, J. 2002. Importance of frequency and temporal song matching in black-capped chickadees: evidence from interactive playback. Ethology, 108, 181–191. Payne, R. B. 1996. Song traditions in indigo buntings: origin, improvisation, dispersal, and extinction in cultural evolution. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca, NY: Cornell University Press, pp. 198–220. Peake, T. M., Terry, A. M. R., McGregor, P. K. & Dabelsteen, T. 2001. Male great tits eavesdrop on simulated male-to-male vocal interactions. Proceedings of the Royal Society of London, Series B, 268, 1183–1187. Peake, T. M., Terry, A. M. R., McGregor, P. K. & Dabelsteen, T. 2002. Do great tits assess rivals by combining direct experience with information gathered by eavesdropping? Proceedings of the Royal Society of London, Series B, 269, 925–1929. Pitcher, T. E. & Stutchbury, B. J. M. 2000. Extraterritorial forays and male parental care in hooded warblers. Animal Behaviour, 59, 1261–1269. Podos, J. 1996. Motor constraints on vocal development in a songbird. Animal Behaviour, 51, 1061–1070. Popp, J. W. 1989. Temporal aspects of singing interactions among territorial ovenbirds (Seiurus aurocapillus). Ethology, 82, 127–133. Pusey, A. E. & Packer, C. 1997. The ecology of relationships. In: Behavioural Ecology, an Evolutionary Approach, ed. J. R. Krebs & N. B. Davies. Oxford: Blackwell Scientific, pp. 254–283. Schroeder, D. J. & Wiley, R. H. 1983. Communication with shared song themes in tufted titmice. The Auk, 100, 414–424. Searcy, W. A. & Yasukawa, K. 1996. Song and female choice. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca, NY: Cornell University Press, pp. 454–473. Searcy, W. A., Nowicki, S. & Hogan, C. 2000. Song type variants and aggressive context. Behavioral Ecology and Sociobiology, 48, 358–363. Shackleton, S. A. & Ratcliffe, L. 1994. Matched counter-singing signals escalation of aggression in black-capped chickadees (Parus atricapillus). Ethology, 97, 310–316. Slabbekoorn, H., Ellers, J. & Smith, T. B. 2002. Bird song and sound transmission: the benefits of reverberations. The Condor, 104, 564–573. Smith, D. G. & Norman, D. O. 1979. ‘Leader–follower’ singing in red-winged blackbirds. The Condor, 81, 83–84.

P1: IRK/KFO P2: IRK-KOD 0521823617c14.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:1

Singing interactions in songbirds Sorjonen, J. 1987. Temporal and spatial differences in traditions and repertoires in the song of the thrush nightingale (Luscinia luscinia) birds. Behaviour, 102, 196–212. Stamps, J. 1994. Territorial behavior: testing the assumptions. Advances in the Study of Behavior, 23, 173–232. Stamps, J. A. & Krishnan, V. V. 1997. Functions of fights in territory establishment. American Naturalist, 150, 393–405. 1999. A learning-based model of territory establishment. Quarterly Review of Biology, 74, 291–318. 2001. How territorial animals compete for divisible space: a learning-based model with unequal competitors. American Naturalist, 157, 154–169. Stoddard, P. K. 1996. Vocal recognition of neighbors by territorial passerines. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca, NY: Cornell University Press, pp. 356–376. Todt, D. 1981. On functions of vocal matching: effects of counter-replies on song post choice and singing. Zeitschrift f¨ ur Tierpsychologie, 57, 73–93. Todt, D. & Naguib, M. 2000. Vocal interactions in birds: the use of song as a model in communication. Advances in the Study of Behaviour, 29, 247–296. Vehrencamp, S. L. 2001. Is song-type matching a conventional signal of aggressive intentions? Proceedings of the Royal Society of London, Series B, 268, 1637–1642. von Helversen, D. 1980. Structure and function of antiphonal duets. Acta XVIIIth International Ornithological Congress, Berlin, 682–688. Waser, P. M. & Wiley, R. H. 1980. Mechanisms and evolution of spacing in animals. In: Handbook of Behavioral Neurobiology: Social Behavior and Communication, ed. P. Marler & J. G. Vandenbergh. New York: Plenum Press, pp. 159–233. Weary, D., Falls, J. B. & McGregor, P. K. 1990. Song matching and perception of song types in great tits, Parus major. Behavioral Ecology, 1, 43–47. Wiley, R. H. 1983. The evolution of communication: information and manipulation. In: Animal Behaviour : Communication, Vol. 2, ed. T. R. Halliday & P. J. B. Slater. Oxford: Blackwell Scientific, pp. 156–189. 1994. Errors, exaggeration and deception in animal communication. In: Behavioral Mechanisms in Ecology, ed. L. Real. Chicago, IL: University of Chicago Press, pp. 157–189. Wiley, R. H. & Richards, D. G. 1978. Physical constraints on acoustic communication in the atmosphere: implications for the evolution of animal vocalizations. Behavioral Ecology and Sociobiology, 3, 69–94. Wiley, R. H. & Wiley, M. S. 1980. Territorial behavior of a blackbird: mechanisms of site dependent dominance. Behaviour, 73, 130–154. Wiley, R. H., Godard, R. & Thompson, A. D. 1994. Use of two singing modes by hooded warblers as adaptations for signaling. Behaviour, 129, 243–278. Whitfield, J. 2002. Nosy neighbours. Nature, 419, 242–243. Wolffgramm, J. & Todt, D. 1982. Pattern and time specificity in vocal responses of blackbirds Turdus merula. Behaviour, 81, 264–286. Zann, R. A. 1996. The Zebra Finch: a Synthesis of Field and Laboratory Studies. Oxford: Oxford University Press.

319

P1: JZZ/... P2: JZZ/... 0521823617c15.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:8

15

Dawn chorus as an interactive communication network john m. burt & sandra l. vehrenc amp Cornell Laboratory of Ornithology, Ithaca, USA

Introduction Dawn chorus singing is a striking behaviour pattern, performed by some temperate-zone and tropical songbird species, as well as a few non-passerine and non-avian species. In a typical chorusing songbird species, all territorial males in a neighbourhood synchronously start singing 30 to 90 minutes before sunrise. During the ensuing chorus period, song rate, singing diversity and song complexity reach maximal levels, and often birds do not seem to be interacting with any one particular neighbour (Hultsch & Todt, 1982). Then, as the light level increases around sunrise, this mode of singing usually abruptly ends. Soon after dawn chorus is over, birds begin to forage and patrol their borders, and they switch to courtship singing or dyadic (i.e. paired) counter-singing with nearby neighbours. Post-chorus singing is typically more sporadic and overall song rates tend to be lower and much more variable than they are at dawn chorus (for a review of dawn chorus behaviour, see Staicer et al. (1996)). Numerous hypotheses have been proposed to explain dawn chorus singing. In an insightful review, Staicer et al. (1996) outlined 12 non-exclusive hypotheses and compared their predictions against the existing empirical evidence across many songbird and non-songbird species. The 12 hypotheses were grouped into three categories: intrinsic, environmental and social. Intrinsic explanations such as circadian cycles of testosterone and self-stimulation are likely proximate mechanisms for dawn singing (Wingfield & Farner, 1993; Goodson, 1998). Environmental explanations, such as low predation risk, good acoustic transmission, inefficient foraging with low ambient light and unpredictable night-time conditions leading to excess energy reserves on most mornings, provide reasons why singing at Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

320

P1: JZZ/... P2: JZZ/... 0521823617c15.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:8

Dawn chorus as an interactive network dawn might be less costly than singing at other times (Henwood & Fabrick, 1979; Kacelnik & Krebs, 1982; Mace, 1987; McNamara et al., 1987; Hutchinson, 2002; Dabelsteen & Mathevon, 2002). However, none of these intrinsic or environmental hypotheses provide a functional explanation of the selective advantage for vigorous, continuous, complex vocal displays prior to sunrise, nor do they explain why only some species exhibit this phenomenon. The social hypotheses outlined by Staicer et al. (1996) that do attempt to provide functional explanations for dawn chorus singing include mate attraction and/or stimulation, territory defence and resolution of social dynamics. Predictions for these alternatives have been used to support or reject certain hypotheses in several species. For example, mate attraction and mate stimulation can be ruled out as primary explanations in species where dawn chorus singing is uniformly high across the breeding season, rather than being concentrated during periods of mate attraction and mate fertility as these hypotheses predict (Kroodsma et al., 1989; P¨ art, 1991; Slagsvold et al., 1994). Moreover, males have generally been observed to stop dawn chorus singing as soon as their mates emerge from their night roost, and males of some species drop out of the dawn chorus altogether on days of peak mate fertility, findings that do not support mate stimulation hypotheses for those species (Mace, 1986; Cuthill & Macdonald, 1990; P¨ art, 1991; Otter & Ratcliffe, 1993). The territory defence hypothesis, suggesting that dawn song is an extra vigorous keep-out signal, is contradicted in some sedentary species because males cease dawn chorus singing during the non-breeding season even though they continue to occupy and defend a territory (Staicer et al., 1996). Staicer et al. concluded that the social hypothesis that best fits the existing evidence is the social dynamics hypothesis, which proposes that the function of dawn chorus is the interactive communication and adjustment of social relationships among males. Their conclusion is based partly on findings in numerous species that dawn chorus singers use signals and modes of signalling that are specifically associated with male–male interaction and that dawn chorus singers of some species appear to be listening to and directing their songs towards particular neighbours (Kroodsma et al., 1989; Staicer, 1989; Nelson & Croner, 1991; Spector, 1991; Dabelsteen, 1992; Staicer et al., 1996). Females of at least some species may acquire information about mate quality from dawn singing. In several species, individual differences in dawn song output are correlated with male age and/or dominance and with female laying date and/or fecundity (Welling et al., 1995; Otter et al., 1997; Poesel et al., 2001; Ballentine et al., 2003). Females could possibly assess the quality of singing males at dawn chorus either by attending to song traits that are directly associated with indicators of fitness, such as stamina, age or dominance, or by eavesdropping to

321

P1: JZZ/... P2: JZZ/... 0521823617c15.xml CU1917B/McGregor

322

0 521 582361 7

April 11, 2005

17:8

J. M. Burt & S. L. Vehrencamp acquire information about relative quality if males are interacting (Otter et al., 2001; Mennill et al., 2002; Ch. 7). Taken as a whole, the pattern of dawn chorus behaviour across species does not seem to fit any single currently hypothesized function. Furthermore, the social hypotheses are not mutually exclusive, suggesting that the dawn chorus might have multiple functions that may differ in relative importance, depending on species. If that is the case, then it may be necessary to examine more closely (or re-examine) the singing behaviour of each species that has a dawn chorus, paying attention to such factors as the intended receiver(s) and whether interactions are occurring among singers within the chorus. Staicer et al. (1996) provided a set of predictions for such evidence that could be helpful in distinguishing which hypotheses might apply to a particular species (cf. Table 24.1 in Staicer et al., 1996). Investigation of these target and interaction issues would benefit from an approach that considers the neighbourhood of singing males as a communication network. Network communication is broadly defined as the involvement of at least three individuals, one or more of them signalling and all receiving (McGregor & Dabelsteen, 1996; see other chapters in this volume). The dawn chorus, with many simultaneous signallers (possibly interacting) and many potential receivers, certainly fits the broad definition of a communication network. However, to date, no study has examined dawn chorus from a network perspective. We suggest that a study of the characteristics of the communication network that occurs at dawn chorus for a given species could provide further information about its function. In this chapter, we discuss what kinds of communication network the different functional hypotheses might predict for dawn chorus and how one might go about testing for them. We then test some of these ideas using a dawn chorus recording of a neighbourhood of banded wrens Thryothorus pleurostictus as an example.

Communication network structures at dawn chorus The detailed structure and complexity of communication networks can vary and depend largely on the degree to which the signalling ‘links’between communicating individuals are one way or are interactive (i.e. signals flow both ways between individuals). Using three hypothetical individuals (Fig. 15.1), it is possible to define three basic network structures: broadcast networks, in which one sender broadcasts a one-way omnidirectional signal to two receivers (Fig. 15.1a); eavesdropping networks, where two senders interact and a third receiver eavesdrops on the interaction to obtain information about the interactants (Fig. 15.1b); and interactive networks, with three senders all interacting with each other (Fig. 15.1c).

P1: JZZ/... P2: JZZ/... 0521823617c15.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:8

Dawn chorus as an interactive network

(a)

(b)

(c)

Fig. 15.1. Three basic network components may occur within a communication network, either singly or in combination: (a) broadcast networks, in which at least one sender produces undirected one-way signals that are received by potentially many receivers; (b) eavesdropping networks, in which two signallers interact and eavesdroppers obtain relative information about each interactant; (c) interactive networks containing three or more individuals signalling interactively to one another and eavesdropping on nearby interactions.

Any real-world communication network is likely to include many more than just three individuals, and networks can theoretically consist of one or any combination of these basic components, adding more potential variety and complexity to network structure. For example, it is likely that any bird-song network with an interactive component probably also has an eavesdropping component in the form of non-interacting listeners such as females and floater males, as well as interacting males who are eavesdropping on their neighbours’ interactions.

323

P1: JZZ/... P2: JZZ/... 0521823617c15.xml CU1917B/McGregor

324

0 521 582361 7

April 11, 2005

17:8

J. M. Burt & S. L. Vehrencamp Investigating the structure of a dawn chorus communication network might provide useful clues about its function, with a key variable being the degree to which signallers within the network are interactive. For example, omnidirectional non-interactive signalling at dawn would indicate a broadcast network. A broadcast network at dawn chorus would support male keep-out, female attraction or direct male quality-assessment functions, since these hypotheses do not strictly require male–male interaction. Conversely, evidence of two-way or multi-way signalling between singers at dawn would indicate an interactive network, supporting the notion that the dawn chorus serves an inter-male communication role (the social dynamics hypothesis). Evidence of eavesdropping would also support a relative male quality-assessment function for dawn chorus.

Searching for interactions at dawn The presence or absence of an interactive network may be a key diagnostic for the function of dawn chorus, but on a practical level how do we go about looking for interactions at dawn? One possible first step would be to examine what we know about how males in a given dawn-chorusing species use their songs to communicate during the daytime. Indeed, much effort has been devoted to analysing daytime counter-singing interactions between pairs of adjacent males. Observational studies of daytime dyadic counter-singing between focal birds and a neighbour or intruder are relatively easy to conduct, and playback experiments have been used to test hypotheses for the function of male song interactions. In territorial species, song has generally been found to function as a keep-out signal to other males (Krebs et al., 1978; Yasukawa et al., 1982; Nowicki et al., 1998). Studies of daytime singing have found that song can also be used in complex ways to mediate aggression between neighbours. For example, in populations with high levels of song-type sharing between neighbours, birds can match their neighbour’s song with their own version of that type. Matching is particularly useful as a directed signal, since by replying with the same type a bird can unambiguously address a rival. Post-chorus, birds have been shown to use song matching as a directed threat to indicate subsequent aggressive intentions (Krebs et al., 1981; McGregor et al., 1992; Burt et al., 2001; Vehrencamp, 2001). The rate of switching between song types, temporal overlapping of songs, duration matching and pitch matching are additional potential directed signalling strategies that vary with the intensity of agonistic interaction and serve to mediate aggression between neighbours (reviewed by Vehrencamp, 2000). The directional properties of signals such as song matching make them potentially useful for detecting interactions at dawn since an audio recording of both singers can determine the singer and target. However, conventional one- or twochannel recording methods will not usually be adequate for this task for three

P1: JZZ/... P2: JZZ/... 0521823617c15.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:8

Dawn chorus as an interactive network reasons. First, a dawn chorus interactive network would span a large area of adjacent territories and birds could interact with any neighbour, thereby reducing the effectiveness of focal recording. Second, identification and location of singers by visual means is very difficult in the twilight before dawn; third, the sheer quantity of vocalizations that occur at dawn in many locations tends to mask easy identification of specific interacting participants. In fact, these difficulties may be the reason why other researchers have not detected interactive networks at dawn chorus, even in species where males clearly interact with their neighbours during daylight. New methods for studying dawn chorus

One solution to the problem of detecting interactions at dawn is the use of distributed microphone arrays, which have been proposed as an ideal method for monitoring communication networks in territorial neighbourhoods (McGregor & Dabelsteen, 1996). Such systems are ideal for studying dawn chorus because they can simultaneously record the songs of many singers in a large area. Another advantage to using microphone arrays is the ability to determine the location (and, therefore, in many cases the identity) of each singer using sound arrival time differences (Watkins & Schevill, 1972; Speisberger & Fristrup, 1990). The specific details of how acoustic location systems work is reviewed more thoroughly in McGregor et al. (1997). Array recording is a particularly good method for detecting interactive networks in species that are known to use some form of directed signal, such as immediate matching or overlapping during vocal exchanges. Changes in singing behaviour associated with movement to different parts of the territory may also indicate interaction, most likely of a dyadic nature. Additionally, microphone array recordings can be used to document changes in the singing behaviour of non-interacting individuals before, during and after an intense interaction between two other individuals in a neighbourhood (Eason & Stamps, 1993; Bower, 2000). Evidence that non-vocalizing receivers act on information gained by eavesdropping (i.e. eavesdropping networks) is best acquired with carefully designed playback experiments (e.g. Naguib & Todt, 1997; Oliveira et al., 1998; Naguib et al., 1999; McGregor et al., 2001; Peake et al., 2001; see also Ch. 2).

Banded wren song behaviour In this chapter, we describe one of the first attempts to study a dawn chorus communication network using a microphone array system (see also Bower, 2000). Our study species is the banded wren, in which males possess repertoires of discrete, distinctive song types and usually switch to a different type after each consecutive song. Neighbouring males share many of the same song types

325

P1: JZZ/... P2: JZZ/... 0521823617c15.xml CU1917B/McGregor

326

0 521 582361 7

April 11, 2005

17:8

J. M. Burt & S. L. Vehrencamp and prior studies have documented the use of type matching and other singing patterns during aggressive encounters (Molles & Vehrencamp, 1999, 2001). This species is also a vigorous dawn chorus singer. Here we present an initial analysis of the singing behaviour and interactions among four neighbouring male wrens recorded during a single morning encompassing dawn chorus and the subsequent hour of post-dawn chorus. We look for evidence of network communication interactions involving two, three or more birds by searching for the presence of directed signals such as matching and overlapping. The banded wren is a common and vocally active species that inhabits the tropical dry deciduous forest of the Pacific slope of Central America. It breeds only during the first half of the rainy season (May–August) but remains resident and paired on the same territory during the rest of the year. The mating system is socially monogamous and each pair defends an all-purpose territory approximately 0.4 ha in area. Although not a true duet, a female occasionally sings short malelike songs following or overlapping her mate’s songs. Males possess a repertoire of 15 to 30 discrete song types, which may be delivered with a high rate of switching between song types (immediate variety mode) or in a more repetitive fashion (eventual variety mode). Young males tend to copy whole song types from nearby males and generally do not disperse very far from their natal territory, so established adjacent neighbours share between 50 and 90% of their song-type repertoire (Molles & Vehrencamp, 1999). In the course of our research, we have identified a variety of song-delivery patterns that banded wrens use to communicate with their neighbours in the daytime during bouts of counter-singing. Song matches appear to be a threat signal; repertoire matches (singing a song shared with but not currently sung by the neighbour) are used as a low-threat directed signal maintaining the interaction, and switches to non-shared song types indicate a desire to deescalate (Molles & Vehrencamp, 2001). Finally, banded wrens also appear to use song overlapping during escalated interactions. Although the function of overlapping is not well understood, it is often associated with (and often simultaneously combined with) high rates of song matching, suggesting it also has a threat function. Overlapping is apparently avoided during low-intensity counter-singing interactions between distant males (Molles & Vehrencamp, 1999). Banded wrens have a pronounced dawn chorus during the breeding season months and are relatively silent at dawn during the rest of the year, despite remaining on their territories. The chorus starts at twilight (approximately 05:00 h) and lasts about 30 minutes. During this period, males sing vigorously and loudly. They initially perch high in an emergent tree (10–20 m) and sing without pause from one location for several minutes while constantly changing their body orientation. As light levels increase, dawn chorus singers usually shift to other high perches

P1: JZZ/... P2: JZZ/... 0521823617c15.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:8

Dawn chorus as an interactive network in other parts of their territories. Prior to the array study, we had anecdotally observed numerous type matching events at dawn, but, given the conditions, it was unknown to what extent, and with whom, the birds were matching. These preliminary observations of song matching suggested to us that dawn chorus might be an interactive phenomenon. A male usually abruptly ceases dawn chorusing behaviour when his mate approaches and interacts with him. She may join him in a brief, uncoordinated duet. Occasionally a male forgoes the dawn chorus completely to interact with his (presumably fertile) mate. Unmated males continue to sing at a high rate for another 30 to 60 minutes. After the dawn chorus, males begin foraging, interspersed with bouts of singing, and patrol the borders of their territories more actively. At this time, males seem to shift to more focused counter-singing with nearby neighbours. Each territorial male has three to four adjacent neighbours with whom he regularly interacts. Males construct bulky covered nests and are constantly initiating new ones because of high nest-predation rates. Females appear to select the nest location and sometimes choose a site near a territorial border, which forces the male to renegotiate that boundary with his neighbour through closerange counter-singing and fights. The dynamic nature of territory boundaries in this species could be one source of changes in social status hypothetically being signalled during the dawn chorus. Recording methods and subjects

As part of our research project studying the function of banded wren song, we developed a microphone array recording system as a tool for acoustically monitoring several vocalizing individuals. The technique involves placing an array of many microphones at strategic locations within and around a small neighbourhood of banded wren territories and simultaneously recording all song interactions picked up by the microphones on a central multiple-channel receiving unit. With these array recordings, we can quantify neighbourhood-wide singing patterns and also focus on individual birds to gain a more complete picture of their interactions with all of their neighbours. The array data presented in this chapter were taken from an analysis of a recording made on 20 June, 2001 at one of our study sites in Santa Rosa Park, Costa Rica. This recording was selected from our library of daily recordings made during the 2000, 2001 and 2002 May–July field seasons as a representative example of dawn chorus singing in our study population. The chapter dataset runs from 05:02 h (Central American time zone) to 06:42 h, a total of 100 minutes. On that day, civil twilight occurred at 04:58 h and sunrise at 05:21 h. Behaviourally, the recording covered the first song sung in the neighbourhood at 05:04 h, the entire dawn chorus and about one hour of post-chorus singing when dyadic counter-

327

P1: JZZ/... P2: JZZ/... 0521823617c15.xml CU1917B/McGregor

April 11, 2005

17:8

J. M. Burt & S. L. Vehrencamp 450

400

*

350

PU

*

LE 300

Distance (m)

328

0 521 582361 7

* 250

OH

* BO

*

N

N

*

*

150

WB

* N

* WE

*

100

*YO

N

*

200 UB

*

50 50

100

150

200

250

300

350

400

Distance (m) Fig. 15.2. Map of banded wren territorial neighbourhood recorded by the array. Solid lines show the boundaries of the focal neighbours, while dotted lines are the boundaries of adjacent neighbours. An asterisk indicates a microphone position and N indicates current nest sites for each focal male. The shaded region in bird OH’s territory indicates the area recently annexed from bird YO by OH.

singing predominated. The array consisted of 13 microphones, situated among and surrounding four focal neighbours (males OH, YO, BO and WB; Fig. 15.2 shows array configuration and territories). Since the focal birds were usually within the array, nearly all of the songs sung by these four birds could be identified and located. Songs of four other neighbours adjacent to the central four often could be detected and identified if the bird was close to the array, but usually not located (birds WE, PU, UB and LE; Fig. 15.2). Songs of these outlying birds were included in calculations of overlapping and song matching to create more accurate song statistics for the central focal birds. During the recording, four observers were posted near each focal bird to take behavioural notes so that later we could reconstruct patterns of interaction.

P1: JZZ/... P2: JZZ/... 0521823617c15.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:8

Dawn chorus as an interactive network The recording on 20 June 2001 was typical for a mid-breeding season dawn chorus and post-dawn chorus day of singing. The focal birds varied in age and stage of nesting. OH had recently built a nest near his border with YO and had extended his territory into that of YO (shaded area on OH’s territory in Fig. 15.2). This boundary shift was a source of ongoing aggression between OH and YO, as well as with BO, another of OH’s neighbours who was also affected by the shift. OH was a long-term resident of the field site (at least seven years old). YO was banded as an adult two years earlier (so was three or more years of age) and on the recording day had just started building a new nest, which was located approximately 10–20 m from OH’s border and approximately 50 m from OH’s nest. WB was an offspring of the previous owner of PU’s territory, hatched in 1999 (two years old) and had an active nest with nestlings. BO was a newly banded bird (probably first year), whose nest was predated the day before by capuchin monkeys. During the recording, BO was observed to be building a new nest. Temporal patterns of singing behaviour

In this section, we provide a quantitative description of changes in several key singing behaviours over time, based on the four focal birds in our recording. To help the reader to visualize these dynamic patterns, we have adopted a presentation format that plots a running mean and standard error of the mean (SEM) of the four focal birds’scores, starting at the time of the first song of the morning (05:04 h) over a series of overlapping five minute intervals that move forward in time in one minute increments (Fig. 15.3). Since the numbers of matches and overlaps are highly dependent on the number of songs delivered in each interval, percentages are given for these measures. Table 15.1 shows the correlations among these measures. For matching and overlapping analyses, all focal birds were considered to be ‘adjacent’ neighbours (see Fig. 15.1), because all focal birds could easily hear all the other focals and we had previously observed matched counter-singing between all combinations of focal neighbours during the daytime. Bout structure

Intersong interval (ISI), measured as the time from the end of a song to the beginning of the next song, was used as an index of bout structure. This variable is particularly sensitive to shifts between continuous singing and bout singing (periods of relatively high song rates interspersed with pauses in singing) and can be used to indicate the point in time that dawn chorus ends and postchorus bout singing begins. When all birds sing continuously at high rates, as in the classic dawn chorus, the mean and variability of ISI will be small. When birds shift to singing in asynchronous bouts, both the mean and the variability of ISI will increase. Figure 15.3a shows the running mean and SEM of ISI over the course

329

P1: JZZ/... P2: JZZ/... 0521823617c15.xml CU1917B/McGregor

April 11, 2005

17:8

J. M. Burt & S. L. Vehrencamp

(a) 200

twilight

150 ISI (s)

100 50 0 05:00

(b)

05:15

05:30

05:45

06:00

06:15

06:30

06:45

05:15

05:30

05:45

06:00

06:15

06:30

06:45

05:15

05:30

05:45

06:00

06:15

06:35

06:45

05:15

05:30

05:45

06:00

06:15

06:30

06:45

05:15

05:30

05:45

06:00

06:15

06:30

06:45

Songs/bird

20

10 5

100

Movement (m)

(c)

15

0 05:00

(d)

50

0 05:00

Matches (%)

100

(e)

50

0 05:00 100

Overlaps (%)

330

0 521 582361 7

50

0 05:00

Time of day Fig. 15.3. Changes in song behaviour and movement patterns over the course of the recording, averaged across the four focal birds OH, YO, BO and WB. Values are measures calculated over a series of overlapping five minute intervals that move forward in time in one minute increments. Values are plotted at the centre time for each segment. Means are shown as a solid line and the grey region represents ± SEM. (a) Intersong interval (ISI); (b) number of songs per bird; (c) distance moved from each bird’s average position at the previous to the current time segment; (d) percentage of songs within a segment that were matches; (e) percentage of songs within a segment that were overlaps.

P1: JZZ/... P2: JZZ/... 0521823617c15.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:8

Dawn chorus as an interactive network Table 15.1. Pearson correlation coefficients among the key singing behaviours above the diagonal, p values below the diagonala Intersong Song rate Song rate

interval

Movement

0.046

−0.222 0.280

Intersong interval

0.657

Movement

0.029

0.005

Matches (%)

0.322

0.211

0.147

Overlaps (%)

0.0001

0.549

0.111

a

Matches (%) 0.102

Overlaps (%) 0.378

−0.0128

−0.062

−0.149

−0.163 0.251

0.013

Significant correlation coefficients are shown in bold. The time series data were separated into one minute bins. Each variable was corrected for autocorrelation by regressing it against its lagged values and the residuals were used for the correlations (n = 97).

of the recording. Song bout structure for these banded wrens clearly differed between dawn chorus and later. From the start of dawn chorus and for 20 minutes into the recording, the mean ISI was very brief (around 10 seconds) and SEM was very low, indicating that all birds were singing more or less continuously. At around 05:21 h, the ISI measure exhibits a noticeable break from the previous trend and both the mean and SEM increase and become slightly more variable over time. At 05:45 h, a major break in ISI occurred, reflecting the fact that birds had begun to sing in asynchronous bouts with variable interbout pauses. From that point on, each bird stopped singing at least once for five minutes or longer. These pauses are shown as spikes in mean ISI in Figure 15.3a, which occurred at a different time for each bird (WB: 05:44–05:50 and 06:16–06:23 h; BO: 06:06–06:12 h; YO: 06:23–06:29 h; OH: 06:25–06:32 h). Song rates

In this recording, birds sang at uniformly high rates throughout dawn chorus (Fig. 15.3b). OH began singing two minutes before the other birds, then YO and BO began to sing, and two minutes later WB finally joined the chorus. At 05:22 h, WB began to sing at a lower rate, while OH, BO and YO continued to sing at high but more variable rates (shown as an increase in SEM in Fig. 15.3b). At about 05:40 h, all four birds began a bout of intense counter-singing, and OH continued to sing at an exceptionally high rate. After this synchronized bout, song rates declined again but were punctuated by four more peaks, which reflect brief bouts of countersinging between different sets of neighbours. Individually, the four birds clearly differed in their overall song output, with WB in particular being consistently lower in song rate, starting later and quitting earlier compared with the others.

331

P1: JZZ/... P2: JZZ/... 0521823617c15.xml CU1917B/McGregor

332

0 521 582361 7

April 11, 2005

17:8

J. M. Burt & S. L. Vehrencamp Movement patterns

Patterns of movement were analysed using passive acoustic location to calculate the position of each bird whenever he sang. Figure 15.3c plots distances between mean locations for successive time segments. This measure has low values when birds are singing continuously from one position and increases when birds move to new positions. The focal birds showed an initial spike of movement at the beginning of dawn chorus (05:04–05:07 h) and then remained relatively stationary until about 05:22 h. The initial movement spike was caused by several birds (BO, OH and WB), who appear to have sung for a brief period from positions near their sleeping nests and then moved to a more centrally located high song post for the bulk of their dawn chorus singing. Between 05:27 and 05:43 h, birds made two or three short distance movements to different parts of their territories while still singing fairly vigorously and continuously. After 05:43 h, movements became larger and more frequent. At this time, birds moved relatively quickly between song perches and tended to stay at each location for several minutes before moving again (these movements are seen as a pattern of brief spikes in Fig. 15.3c).

Song matching

Figure 15.3d plots the mean percentage across birds of songs sung that were matches. A song was judged to be a match if it was the same type as an adjacent neighbour’s recent song (either a song the neighbour had just sung, or the one previous) and occurred less than 30 seconds after the matched song. Matching was initially lower because OH had sung alone for two minutes, but as soon as the other birds began to sing, matching rates quickly increased. Song matching peaked at 70% at 05:14 h (12 minutes into dawn chorus). After the first peak, matching declined slightly (although it remained quite high at around 40%) and peaked again at 05:30 h (52%). Matching declined thereafter, but peaked again at 05:56 (43%), 06:07 (30%) and 06:26 h (25%). The four focal birds sang an average of 22 different song types during the recording (range, 20–24) and shared 82% of their song types with any given neighbour (range, 77.3–85.1). With high singing rates, frequent song type switches and high levels of song type sharing, the possibility of birds matching by chance will be higher at dawn chorus. We created a model to test whether the observed rates of matching were higher than that expected through chance. To estimate chance matching rates, we generated new datasets using the observed singing data for all males with randomly shuffled song-type assignments within each bird’s repertoire. By averaging the matching rates obtained over multiple permutations of shuffled song types, we could estimate the probability of chance matching if the birds had chosen their songs without regard to other singers. Figure 15.4 compares observed matching versus expected chance matching calculated on the basis of

P1: JZZ/... P2: JZZ/... 0521823617c15.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:8

Mean percentage matches per 5 min

Dawn chorus as an interactive network

Fig. 15.4. Observed (± SEM) and expected percentage matching rate per five minute interval. Expected values were calculated by averaging matching outcomes of 100 random permutations of song types within focal birds. Observed matching rates that were significantly higher than expected are marked with an open circle (two-tailed binomial tests, Holm corrected for multiple comparisons; criterion p < 0.05).

100 permutations of random song-type shuffling. Observed rates were significantly higher than expected throughout the first 15 minutes of dawn chorus and during the four subsequent peaks later on. The peak at 05:28 h was caused by an intense interaction between the four focal birds, with a three-way matching interaction between OH, YO and BO, and a separate matching interaction between WB and BO (conclusions drawn from analysing individual bird data not shown in the figure). The three later peaks are attributable to further intense bouts of matched counter-singing between pairs and trios of neighbours (OH, YO and PU at 05:56 h; WB, UB and WE at 06:07 h; and WB, BO and LE at 06:27 h). If song matching is an indicator of conflict, knowing who is matching with whom can provide useful information about what is going on in a neighbourhood. Bird OH was involved in many of the interactions that morning. We think that his singing behaviour and his neighbours’ responses to him were related to his recent annexation of space at the corner of his, YO’sand BO’sboundaries to defend a newly active nest in that area (Fig. 15.2). Figure 15.5 shows the patterns of matching by

333

P1: JZZ/... P2: JZZ/... 0521823617c15.xml CU1917B/McGregor

334

0 521 582361 7

April 11, 2005

17:8

J. M. Burt & S. L. Vehrencamp

Fig. 15.5. Patterns of matching by and to bird OH throughout the recording period. Grey squares indicate OH’s song matches to one of his four neighbours. Black squares indicate that OH sang a non-matching song. Open squares indicate a match by a particular neighbour to OH.

and to bird OH, giving us a finer picture of the interactions between him and the other birds that morning. OH’s bout of solo singing two minutes before his other neighbours accounts for the lower mean song matching at the very beginning of dawn chorus. During the initial peak of song at 05:12 h, OH was matched by and was matching the neighbours at the disputed corner (BO and YO). At this time, the majority of matches were initiated by the neighbours (each matched a different one of OH’s songs) and OH replied with his own matches to those neighbours (i.e.

P1: JZZ/... P2: JZZ/... 0521823617c15.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:8

Dawn chorus as an interactive network Table 15.2. Percentage of matches by and to each focal bird Target bird

Matching bird BO

OH

YO

WB

Other neighbours

BO

–

25

38

20

17

OH

32

–

43

16

8

YO

22

49

–

11

17

WB

21

38

24

–

17

Other neighbours

14

8

16

25

–

in Fig. 15.5 the open squares indicating a neighbour match usually preceded OH’s own match to that neighbour, so the neighbour matched first). OH’s border shift had particularly affected YO, and YO and OH matched most often throughout the morning (43% of YO’s matches were to OH, while 49% of OH’s matches were to YO; Table 15.2), as would be predicted by an agonistic function for dawn chorus singing. Another trend is visible in Fig. 15.5, primarily during the dawn chorus (05:07–05:32 h) and briefly later (05:52–05:57 h): OH, as well as the other birds not shown in this figure, frequently alternately matched multiple neighbours within a short period of time. Often these multiple matches involved using a different song to match each neighbour. For example, at the start of dawn chorus, OH had been switching between two song types. The first song sung by YO was a match to one of OH’s types, while BO first sang a match to the other type. Matching was not merely isolated between the OH/YO/BO trio – all of these birds were often matching other birds at the same time. Based on our observations that morning, we know that WB spent most of his time interacting with two other neighbours (WE and UB) but, nevertheless, at various times did match and was matched by OH, BO and YO. Overlapping

In prior work, we had noticed many occurrences of what appeared to be deliberate overlapping during close-range counter-singing interactions between neighbouring males. Furthermore, overlapping often occurred in conjunction with a song match, timed to cover the majority of the other bird’s song as if the matcher were trying to ‘jam’ the other singer. The context of the ‘overlapping match’ phenomenon suggested that overlapping is used as an aggressive signal. Figure 15.3e plots the mean percentage across birds of songs that overlap another song. A song was considered to be an overlap if it ‘covered’ an adjacent neighbour’s song by 50% or more of its duration. Songs are sufficiently long in this species (mean duration is 3.4 seconds) and the territories relatively small (centres 120 m apart) that errors in perceived overlapping caused by the slow speed of

335

P1: JZZ/... P2: JZZ/... 0521823617c15.xml CU1917B/McGregor

336

0 521 582361 7

April 11, 2005

17:8

J. M. Burt & S. L. Vehrencamp sound are small for this high overlap criterion (Dabelsteen, 1992). Overlapping was strongly correlated with mean song rate (Table 15.1). Our overlapping measure cannot distinguish between accidental and deliberate overlapping and it is likely that much of the trend in overlapping resulted from an increase in the probability of accidental overlapping during times of increased song rates. Overlapping may still be used deliberately, as we have observed, but much less frequently or under very specific contexts. We have some evidence for this: overlapping rate was significantly correlated with matching rate, even after controlling for the effects of song rate (partial r = 0.215; p = 0.024), a pattern that would occur if birds occasionally deliberately combined overlapping with song matching. There were 27 occurrences of overlapping matches out of 924 songs delivered by the focal birds during the recording.

Overall patterns of banded wren singing at dawn chorus The structure of the banded wren dawn chorus follows the pattern commonly described for other species: males begin singing at twilight and sing continuously at high rates. Then, coincident with increasing light levels and female emergence, males change to a daytime pattern of interacting individually with nearby neighbours. Banded wrens also show high rates of song matching to neighbouring singers at dawn chorus. Although dawn chorus matching has been noted in several other species (Todt, 1970; Spector, 1991), it had not been quantified before our study, making it difficult to know whether high-rate matching at dawn occurs in many other species. Therefore, to the degree that dawn chorus has been characterized in other birds, banded wrens appear to behave similarly to other chorusing species. We found clear differences among the four focal males recorded on this single morning in their rate and duration of singing, which could be caused by variation in male quality, condition, dominance status or territory quality, as described for several species (Cuthill & Macdonald, 1990; Otter et al., 1997; Poesel et al., 2001). However, there was evidence that these differences could be caused by short-term variation in motivation arising from differences in nesting stage or territory boundary disputes. The lowest-rate singer in this recording, WB, was a highly successful male who was feeding nestlings, whereas the other three males had recently lost their nests and were engaged in boundary disputes resulting from OH’s incursion into YO’s and BO’s territories. Analysis of additional recordings would clearly be needed to determine whether individual differences in song output are consistent over time or vary with breeding conditions and/or short-term motivation.

P1: JZZ/... P2: JZZ/... 0521823617c15.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:8

Dawn chorus as an interactive network Evidence for an interactive network

Our strongest evidence for an interactive network during the banded wren dawn chorus is the high level of song matching that takes place during the first half hour of singing. Matching at rates significantly higher than chance indicates that song types delivered by one bird are affecting song types delivered by other birds and that implies, by definition, that they are interacting. The dawn matching was likely of a competitive nature too, given that bouts of matching also occur in the daytime during synchronized counter-singing between dyads or trios of neighbouring males and when males approach each others’ boundaries for closer interaction. Our simultaneous recordings showed that males alternately matched different neighbours in rapid succession with different song types, the key type of evidence for a fully multi-way interactive communication network. Furthermore, matching was strongly directed toward one male, OH, by his neighbours YO and BO at a time when OH was expanding his territory into mostly YO’s territory to accommodate a new nest site. We thus see particularly clear evidence of multi-way competitive interactions among these three males. An observational study conducted on the same population in 2000 indirectly corroborates our claim of male–male interaction at dawn. In that study, we found that banded wren males use their most vigorous song type renditions during the most intense period of singing at dawn, for example longer song-type variations, compound songs (two or more types sung together) and song types with longer trills, wider bandwidth and rattle and buzz elements, all of which are associated with intense male–male interactions at other times (S. L. Vehrencamp & A. Trillo, unpublished data). Similar patterns have been described for the European blackbird Turdus merula and yellow warbler Dendroica petechia, which use louder, longer songs of higher intensity during the dawn chorus and when counter-singing from a distance with other males (Dabelsteen, 1992; Lowther et al., 1999).

Comparing song patterns with aggression Patterns of matching and switching in this recording could also reveal short-term changes in motivation and the outcomes of recent interactions. For example, OH’s prior aggressive behaviour (annexing portions of two of the focal neighbours’ territories) appears to have had a strong effect on the behaviour observed in this recording. From the start of dawn chorus, the pattern of matching toward individual birds was strongly asymmetrical, with most of the matching directed toward OH (Table 15.2). OH’s boundary shift had probably involved prolonged bouts of escalated counter-singing and physical fighting with the affected neighbours (YO and BO), making OH a particularly threatening neighbour. OH’s

337

P1: JZZ/... P2: JZZ/... 0521823617c15.xml CU1917B/McGregor

338

0 521 582361 7

April 11, 2005

17:8

J. M. Burt & S. L. Vehrencamp early start to dawn chorus, two minutes before the others, may have been intended to reinforce his tentative ownership of the disputed property. The matching response directed toward OH by the affected neighbours may, in turn, have been retaliatory threats in response to OH’s announcement of continued occupation. Frequent type switching during a matching interaction may enable a bird to assess more easily which neighbour is feeling most threatened or to challenge each neighbour with a distinctive signal. Dawn chorus singing as an indicator of male quality

In addition to revealing information about short-term changes in motivation, dawn singing could also give eavesdropping receivers information about longer-term or intrinsic differences among males related to dominance, condition and age. Montgomerie (1985) suggested that energy reserves should be lowest at dawn, imposing a handicap such that the vigour and amount of singing honestly reflects a male’s condition or territory quality. Food supplementation was shown to increase the amount of song in blackbirds (Cuthill & Macdonald, 1990). Peak song rate during the dawn chorus was correlated with winter dominance at feeders in black-capped chickadees Parus atricapillus (Otter et al., 1997) and with earlier female laying date in the blue tit Parus caeruleus (Poesel et al., 2001). Banded wrens attain peak daily song rates during the dawn chorus. In addition, their songs seem to be especially loud at this time, although this impression could be caused by the high song perches (Dabelsteen & Mathevon, 2002). One drawback to array recording is that individuals who do not vocalize are ‘invisible’ to the analysis and so we have not been able to show any evidence for eavesdropping with this dataset. However, with more recordings and more detailed analysis, it may be possible to show some direct effects of eavesdropping on vocalizing interactants. The value of matching at dawn chorus

A defining feature of the dawn chorus is a continuous high rate of singing by all territorial neighbours. To a listener in a forest at dawn chorus, there is a confusingly high density of song coming from many directions. This unique ‘song environment’ poses a challenge to singers, who may be trying to direct signals to specific neighbours and simultaneously listen for signals directed at them from all of their neighbours. For this reason, highly directional signals may increase in value and birds would be predicted to shift their singing strategies to favour more directional signals and avoid using signals that might create ambiguity as to the singer’s intended target. In particular, dawn chorusing birds may avoid using signals that rely on song rate and timing such as song overlapping,

P1: JZZ/... P2: JZZ/... 0521823617c15.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:8

Dawn chorus as an interactive network synchronized song rates and synchronized song-type switching, because these signals are more likely to be masked when many individuals are singing nearby at a high rate. Song matching is one of the few signals that retains its usefulness as a directional signal at dawn chorus, because it is based on song-type selection, rather than song timing or rate. In addition, the higher rates of singing and song-type switching often seen at dawn chorus provide birds with more opportunities for directed song matching than at other times. For these reasons, it is possible that song matching accompanied by rapid switching rates will be a common occurrence in dawn chorusing species that engage in neighbour–neighbour interactive networks.

Summary The single array recording presented here gives the reader a glimpse into the behaviour of banded wrens at dawn chorus. Our observations provide evidence of a highly interactive communication network, which is most consistent with the social dynamics hypothesis for the function of dawn chorus, as presented in Staicer et al. (1996). We are currently analysing a number of similar recordings, made between 2000 and 2002, on the same focal birds, as well as on different sets of focal neighbours. With more recordings and more birds, we intend to test more rigorously the trends we saw in the recording presented in this chapter. In particular, evidence that the bird who is the focus of matching changes on different days, in relation to current patterns of boundary and nest-site movements, would greatly strengthen the argument for dawn chorus mediating changes in social status. Repeated observations on the same birds will also be necessary to determine whether any of the dawn chorus behaviours are related to male age, repertoire size, sharing level or quality, which would indicate an additional male quality-assessment role for the banded wren dawn chorus, and the existence of eavesdropping. Based on our findings, we think that the microphone array recording technique is an ideal method for studying the details of dawn chorus song behaviour. In particular, multimicrophone recording in some form is possibly the only feasible way to detect and study multi-way interactive communication networks, such as we found in the banded wren dawn chorus. It is our hope that other species can be studied using techniques similar to ours so that cross-species comparisons can be made of dawn chorus communication networks. Until that time, it will not be known whether the highly interactive dawn chorus network we have documented in banded wrens is unique or common among dawn chorusing species.

339

P1: JZZ/... P2: JZZ/... 0521823617c15.xml CU1917B/McGregor

340

0 521 582361 7

April 11, 2005

17:8

J. M. Burt & S. L. Vehrencamp Acknowledgements Logistical support was provided by the staff of the Area de Conservaci´ on Guanacaste. We thank Alex Trillo, Liz Campbell, Carlos Botero, Richard Mills, Dan Pendleton and Harold Mills for helping us set up the array and make observations during recording sessions, and Cary Leung for analysing much of the array data. This research was funded by NIH grant R01-MH60461.

References Ballentine, B., Badyaev, A. & Hill, G. 2003. Changes in song complexity correspond to periods of female fertility in Blue Grosbeaks (Guiraca caerulea). Ethology, 109, 55–66. Bower, J. L. 2000. Acoustic interactions during naturally occurring territorial conflict in a song sparrow neighbourhood. Ph.D. Thesis, Cornell University, New York. Burt, J. M., Campbell, S. E. & Beecher, M. D. 2001. Song type matching as threat: a test using interactive playback. Animal Behaviour, 62, 1163–1170. Cuthill, I. C. & MacDonald, W. A. 1990. Experimental manipulation of the dawn and dusk chorus in the blackbird Turdus merula. Behavioral Ecology and Sociobiology, 26, 209–216. Dabelsteen, T. 1992. Interactive playback: a finely tuned response. In: Playback and Studies of Animal Communication, ed. P. K. McGregor. New York: Plenum Press, pp. 97–109. Dabelsteen, T. & Mathevon, N. 2002. Why do songbirds sing intensively at dawn? A test of the acoustic transmission hypothesis. Acta Ethologica, 4, 65–72. Eason, P. K. & Stamps, J. A. 1993. An early-warning system for detecting intruders in a territorial animal. Animal Behaviour, 46, 1105–1109. Goodson, J. L. 1998. Territorial aggression and dawn song are modulated by vasotocin and vasoactive intestinal polypeptide in male field sparrows (Spizella pusilla). Hormones and Behavior, 34, 66–77. Henwood, K. & Fabrick, A. 1979. A quantitative analysis of the dawn chorus: temporal selection for communicatory optimization. American Naturalist, 114, 260–274. Hultsch, H. & Todt, D. 1982. Temporal performance roles during vocal interactions in nightingales (Luscinia megarhynchos). Behavioral Ecology and Sociobiology, 11, 253–260. Hutchinson, J. M. C. 2002. Two explanations of the dawn chorus compared: how monotonically changing light levels favour a short break from singing. Animal Behaviour, 64, 527–539. Kacelnik, A. & Krebs, J. R. 1982. The dawn chorus in the great tit (Parus major): proximate and ultimate causes. Behaviour, 83, 287–309. Krebs, J. R., Ashcroft, R. & Webber, M. 1978. Song repertoires and territory defence in the great tit. Nature, 271, 539–542. Krebs, J. R., Ashcroft, R. & van Orsdol, K. 1981. Song matching in the great tit Parus major. Animal Behaviour, 29, 918–923.

P1: JZZ/... P2: JZZ/... 0521823617c15.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:8

Dawn chorus as an interactive network Kroodsma, D. E., Bereson, R. C., Byers, B. E. & Minear, E. 1989. Use of song types by the chestnut-sided warbler: evidence for both intra- and inter-sexual functions. Canadian Journal of Zoology, 67, 447–456. Lowther, P. E., Celada, C., Klain, N. K., Rimmer, C. C. & Spector, D. A. 1999. Yellow warbler. In: The Birds of North America, no. 454, ed. A. Poole & F. Gill. Philadelphia: The Birds of North America, Inc. Mace, R. H. 1986. The importance of female behaviour in the dawn chorus. Animal Behaviour, 34, 621–622. 1987. Why do birds sing at dawn? Ardea, 75, 123–132. McGregor, P. K. & Dabelsteen, T. 1996. Communication networks. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca, NY: Cornell University Press, pp. 409–425. McGregor, P. K., Dabelsteen, T., Shepherd, M. & Pedersen, S. B. 1992. The signal value of matched singing in great tits: evidence from interactive playback experiments. Animal Behaviour, 43, 987–998. McGregor, P. K., Dabelsteen, T., Clark, C. W. et al. 1997. Accuracy of a passive acoustic location system: empirical studies in terrestrial habitats. Ethology, Ecology and Evolution, 9, 269–286. McGregor, P. K., Peake, T. M. & Lampe, H. M. 2001. Fighting fish Betta splendens extract relative information from apparent interactions: what happens when what you see is not what you get? Animal Behaviour, 62, 1059–1065. McNamara, J. M., Mace, R. H. & Houston, A. I. 1987. Optimal daily routines of singing and foraging in a bird singing to attract a mate. Behavioral Ecology and Sociobiology, 20, 399–405. Mennill, D. J., Ratcliffe, L. M. & Boag, P. T. 2002. Female eavesdropping on male song contests in songbirds. Science, 296, 873. Molles, L. E. & Vehrencamp, S. L. 1999. Repertoire size, repertoire overlap, and singing modes in the banded wren, Thryothorus pleurostictus. The Auk, 116, 677–689. 2001. Songbird cheaters pay a retaliation cost: evidence for auditory conventional signals. Proceedings of the Royal Society of London, Series B, 268, 2013–2019. Montgomerie, R. D. 1985. Why do birds sing at dawn? In: Proceedings of the XIX International Ethology Congress, p. 242. Naguib, M. & Todt, D. 1997. Effects of dyadic vocal interactions on other conspecific receivers in nightingales. Animal Behaviour, 54, 1535–1543. Naguib, M., Fitchel, C. & Todt, D. 1999. Nightingales respond more strongly to vocal leaders of simulated dyadic interactions. Proceedings of the Royal Society of London, Series B, 266, 537–542. Nelson, D. A. & Croner, L. J. 1991. Song categories and their functions in the field sparrow (Spizella pusilla). The Auk, 108, 42–52. Nowicki, S., Peters, S. & Hughes, M. 1998. The territory defense function of song in song sparrows: a test with the speaker occupation design. Behaviour, 135, 615–628.

341

P1: JZZ/... P2: JZZ/... 0521823617c15.xml CU1917B/McGregor

342

0 521 582361 7

April 11, 2005

17:8

J. M. Burt & S. L. Vehrencamp Oliveira, R. F., McGregor, P. K. & Latuffe, C. 1998. Know thine enemy: fighting fish gather information from observing conspecific interactions. Proceedings of the Royal Society of London, Series B, 265, 1045–1049. Otter, K. & Ratcliffe, L. 1993. Changes in singing behaviour of male black-capped chickadees (Parus atricapillus) following mate removal. Behavioral Ecology and Sociobiology, 33, 409–414. Otter, K. A., Chruszcz, B. & Ratcliffe, L. 1997. Honest advertisement and song output during the dawn chorus of black-capped chickadees. Behavioral Ecology, 8, 167–173. Otter, K. A., Stewart, I. R. K., McGregor, P. K. et al. 2001. Extra-pair paternity among great tits Parus major following manipulation of male signals. Journal of Avian Biology, 32, 338–344. P¨ art, T. 1991. Is dawn singing related to paternity insurance? The case of the collared flycatcher. Animal Behaviour, 41, 451–456. Peake, T. M., Terry, A. M. R., McGregor, P. K. & Dabelsteen, T. 2001. Male great tits eavesdrop on simulated male-to-male vocal interactions. Proceedings of the Royal Society of London, Series B, 268, 1183–1187. Poesel, A., Foerster, K. & Kempenaers, B. 2001. The dawn song of the blue tit Parus caeruleus and its role in sexual selection. Ethology, 107, 521–531. Slagsvold, T., Dale, S. & Saetre, G.-P. 1994. Dawn singing in the great tit (Parus major): mate attracting, mate guarding, or territorial defence? Behaviour, 131, 115–138. Spector, D. A. 1991. The singing behaviour of yellow warblers. Behaviour, 117, 29–52. Speisberger, J. L. & Fristrup, K. M. 1990. Passive location of calling animals and sensing of their acoustic environment using acoustic tomography. American Naturalist, 135, 107–135. Staicer, C. A. 1989. Characteristics, use, and significance of two singing behaviours in Grace’s Warbler (Dendroica graciae). The Auk, 106, 49–63. Staicer, C. A., Spector, D. A. & Horn, A. G. 1996. The dawn chorus and other diel patterns in acoustic signaling. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca, NY: Cornell University Press, pp. 426–453. Todt, D. 1970. Gesang und gesangliche Korrespondenz der Amsel. Naturwissenschaften, 57, 61–66. Vehrencamp, S. L. 2000. Handicap, index, and conventional signal elements of bird song. In: Signalling and Signal Design in Animal Communication, ed. Y. Espmark, T. Amundsen & G. Rosenqvist. Trondheim: Tapir Academic Press, pp. 277–300. 2001. Is song-type matching a conventional signal of aggressive intentions? Proceedings of the Royal Society of London, Series B, 268, 1637–1642. Watkins, W. A. & Schevill, W. E. 1972. Sound source location by arrival times on a non-rigid three-dimensional hydrophone array. Deep-Sea Research, 19, 691–706. Welling, P., Koivula, K. & Lahti, K. 1995. The dawn chorus is linked with female fertility in the willow tit Parus montanus. Journal of Avian Biology, 26, 241–246.

P1: JZZ/... P2: JZZ/... 0521823617c15.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

17:8

Dawn chorus as an interactive network Wingfield, J. C. & Farner, D. S. 1993. Endocrinology of reproduction in wild species. In: Avian Biology, Vol. 9, ed. D. S. Farner, J. R. King & K. C. Parkes. London: Academic Press, pp. 163–327. Yasukawa, K., Bick, E. I., Wagman, D. W. & Marler, P. 1982. Playback and speaker-replacement experiments on song-based neighbour, stranger, and self discrimination in male red-winged blackbirds. Behavioral Ecology and Sociobiology, 10, 211–215.

343

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:57

16

Eavesdropping and scent over-marking robert e. johnston Cornell University, Ithaca, USA

Introduction Compared with communication in other sensory domains and with scents that are released into the air, scent marking is unusual because the signal remains long after the signalling behaviour; for example, the flank gland marks of male golden hamsters Mesocricetus auratus deposited on glass in the laboratory are detected by other hamsters 40 days later and vaginal secretion marks are detected at least 100 days after deposition (Johnston & Schmidt, 1979). In the field, the paste scent marks deposited by brown hyaenas Crocuta crocuta can be detected by humans for at least 30 days (Gorman, 1990) and klipspringers Oreotragus oreotragus respond to preorbital gland marks that have been exposed to direct sun for at least seven days by an increase in scent marking (Roberts, 1998). In many species, especially those that live solitarily, there is often no receiver present when the marks are deposited. Consequently, scent marks are necessarily general broadcast signals that usually have several functions, depending on the age, sex, reproductive status, social status and individual identities of both senders and receivers. One type of marking, scent counter-marking, is directed at the scent marks of other individuals, but again these individuals are often not present to observe the signalling behaviour. I consider the term scent counter-marking to include two different types of behaviour: (a) over-marking, in which the second individual’s scent at least partially overlaps that of the first individual; and (b) adjacent marking, in which the second individual’s scent is close to that of the first individual but does not overlap it. In the species with which I am most familiar (golden hamsters and meadow voles Microtus pennsylvanicus), both types of Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

344

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:57

Eavesdropping and scent over-marking counter-marking are usually intermixed, but it is possible that some species would engage in one type of counter-marking but not the other, or would vary the type of counter-marking or the proportions of different types of counter-mark in different contexts. Based on the variety of situations in which counter-marking is observed, it no doubt has a variety of functions (Brown & Macdonald, 1985). Counter-marking often occurs only between adult males, suggesting a sexually selected trait related to competition for females (Hurst & Rich, 1999; Johnston, 1999), but in some species adult males mark over the marks of their mates, suggesting mate guarding (Mertl, 1977; Moore & Byers, 1989; Kappeler, 1998; Roberts, 2000; Powzyk, 2002). Females may counter-mark other females (e.g. golden hamsters (Johnston, 1977), house mice Mus musculus (Hurst, 1990)) or mark over the marks of males (e.g. moustached tamarin Saguinus mystax (Heymann, 1998)). Among species that live in social groups, most or all of the members of a group may mark in the same place sequentially, producing a group counter-mark (e.g. Rasa, 1973; Mills et al., 1980; Gorman & Mills, 1984); sometimes dominant individuals in the group mark most often (Peters & Mech, 1975), but sometimes subordinate, subadult individuals mark most (Lazaro-Perea et al., 1999). Groups may also counter-mark during territorial encounters (Jolly, 1966). There is a great need for more observational and experimental field studies focused on the functions of over-marking. Several years ago, I proposed that the functions of over-marking could be approached from the question of what kinds of information third parties could obtain about the marking individuals from scents in over-marks (Johnston et al., 1994). I suggested that there were three different types of effect that might occur: (a) masking, in which the most recently deposited scent covers previous scent marks and thereby eliminates access to the information they contain (e.g. individual identity, sex, reproductive state); (b) mixing, in which the scents of different individuals become a chemical mixture, thus producing a new odour quality (e.g. a group odour) and thereby eliminating information about particular individuals; and (c) posting, in which each scent mark to some extent remains separate or distinguishable, thus producing a bulletin board at which information about all individuals that marked there can be obtained. Subsequent research suggests that an important fourth possibility is that scent over-marks can also provide information about the relationships between the odours of different individuals, perhaps including the relative freshness, amount of scent, number of marks, top or bottom position, or geometric layout (Wilcox & Johnston, 1995; Johnston et al., 1997a; Johnston & Bhorade, 1998; Hurst & Rich, 1999). Although these early experiments addressed the perception of over-marks and the subsequent memory for individual odours, they also can be viewed as indicating the kinds of information that are available to eavesdroppers.

345

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

346

0 521 582361 7

April 12, 2005

11:57

R. E. Johnston Scent marking, scent over-marking and eavesdropping

Regardless of the different functions that counter-marking might have, eavesdropping can occur when a third individual investigates a place marked by two or more other individuals. I use eavesdropping in the broad sense of an animal witnessing some type of interaction or an exchange of information between two other individuals (similar to ‘social eavesdropping’ sensu Peake, Ch. 2). This does not imply conscious intent or secrecy on the part of the animal doing the observing nor any awareness or attempt to conceal the interaction by the animals whose interaction is observed. Eavesdropping on scent counter-marking may be particularly common because the marks are so long lasting, allowing many individuals to investigate them, not just the individuals present when the marking behaviour was performed. Furthermore, when eavesdropping on scent counter-marks, individuals can more easily avoid a potential cost of eavesdropping on vocalizations or visual displays, namely being detected and threatened, chased or attacked. Among vertebrates, neither scent marking nor other aspects of chemical communication have been explicitly analysed using the framework of eavesdropping. Among invertebrates, however, numerous cases have been described using these concepts, but authors studying insects seem to have adopted a different meaning for the term. All of the examples I have found involve predators, parasites or parasitoids using the chemical signals of the host or prey as a means of locating them (e.g. Stowe et al., 1995). This is a different set of phenomena (‘interceptive eavesdropping’; see Chs. 2 and 23) and will not be considered here. What evidence is there for eavesdropping based on scent counter-marking among mammals? I will discuss three different types of evidence, primarily from our own work: (a) apparent sensory and perceptual specializations for the evaluation of scent over-marks; (b) evidence that the information in over-marks leads to differential responses by the perceiver towards individuals whose marks are in different positions in over-marks (top versus bottom); and (c) specialized mechanisms for the production of over-marks.

Scent marking in golden hamsters

There are two types of scent marking in golden hamsters. Flank marking is carried out by both males and females and scent is deposited from the flank gland, which is a region of specialized, pigmented sebaceous glands on the posterior flank. This glandular field is larger in males than in females and is testosterone dependent (Vandenbergh, 1973). Flank marking is performed as a part of general maintenance activities (e.g. shortly after animals wake up and groom themselves) and is especially prevalent in potentially agonistic situations (e.g. in the presence of odours of other individuals). Hamsters do not usually flank mark during actual

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:57

Eavesdropping and scent over-marking interactions (Johnston, 1975a,b,c, 1977, 1985). In seminatural laboratory environments, hamsters mark just inside the tunnel to their burrow, in the vicinity of the burrow entrance and in other locations (Johnston, 1975c). Subordinate males in these environments have been observed marking in their nest when a dominant male is attempting to enter (Johnston, 1975c). Therefore, it seems likely that flank marking is involved in defending the burrow and food hoard by both males and females (Johnston, 1975c). The second type of scent marking is a type of anogenital marking, called vaginal marking, that deposits vaginal secretions. Hamsters have a specialized pouch surrounding the distal vagina that produces and collects this secretion. The frequency of vaginal marking is related to the female’sreproductive state, peaking during the night 12–24 hours before receptivity (Johnston, 1977, 1985). This secretion is highly attractive to males; it stimulates copulatory behaviour, reduces aggressive behaviour and causes increases in circulating luteinizing hormone and testosterone (Johnston, 1985, 1990). In a study in seminatural enclosures in the laboratory, females attracted males the night before receptivity, slept with them during the day, mated early the next day and then drove the male away (Lisk et al., 1983). Therefore, one primary function of this type of marking is to advertise sexual receptivity and to attract males. In addition, females may over-mark the vaginal marks of other females (Fischer & McQuiston, 1991), perhaps as a means of competing for the attention of males or as a secondary aspect of defence of the burrow and food hoard against other females.

Specialized mechanisms for evaluation of scent over-marks The first evidence suggesting special mechanisms for evaluation of scent over-marks came from experiments aimed at understanding the information obtained about individuals from the scents in an over-mark, as described above. In particular, these experiments were designed to discover what golden hamsters would remember after investigating an over-mark consisting of the scents of two individuals (Johnston et al., 1994). Corresponding to the idea that an over-mark by one individual might mask, mix or remain distinguishable from the underlying individual’s scent, would hamsters remember just the top scent, neither scent, or both scents? We used an habituation technique in which male subjects were first exposed to a newly deposited, flank-gland over-mark from two donor males (male B always on top of male A) on four or five successive trials and then were tested on a final trial with the scent of one of these donors and the scent of a novel donor. In the first experiment, we wanted to be sure that the scent of male B covered the scent of male A, so we simulated natural scent marks by picking up the animals and rubbing their flank gland region against the substrate, a glass plate. We placed a paper cardstock template over the glass so that scent was deposited only in a

347

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

348

0 521 582361 7

April 12, 2005

11:57

R. E. Johnston limited area on the plate (the exact size of the area corresponded to the usual size of flank or vaginal scent marks (Johnston et al., 1994)). Across repeated habituation trials, the investigation of the scent over-mark decreased, as it would to a single individual’s scent. On the test trial, subjects should investigate the familiar scent significantly less than the novel scent, indicating memory for the familiar scent (e.g. Johnston et al., 1993). The results were quite interesting: subjects treated the flank gland odour of the top-scent individual as familiar (investigated it less than that of a novel individual) but investigated the flank odour of the bottom scent male the same amount as that of the novel individual (Johnston et al., 1994). We obtained similar results when males were habituated to vaginal secretion overmarks of females and then tested with each scent individually compared with a novel vaginal secretion. Because the scent of the second donor was placed on top of that of the first donor, our interpretation was that the top scent masked the bottom scent and, therefore, the bottom scent actually was novel to the subjects (Johnston et al., 1994). This first experiment was, however, somewhat unrealistic in that the scent of the second individual completely covered the scent of the first individual; when hamsters deposit their own scent they are usually not so thorough or precise. In a second experiment, we placed the top and bottom scents at right angles to one another such that they formed a cross; that is, there was a region of overlap of the two individuals’ scent marks, but also regions where each individual’s scent was by itself. The same results were obtained as in the first experiment: male subjects treated the vaginal scent of donor B as familiar but treated the scent of the bottom-scent individual (donor A) the same as that of a novel individual (Fig. 16.1a). Results using flank-gland secretions showed the same pattern (Johnston et al., 1995; Johnston, 1995). These results indicate that the subjects had a preferential memory for the top scent compared with the bottom scent, despite being able to investigate both scents during the habituation trials. This preferential memory suggests that the subjects have a mechanism for evaluating over-marks and either preferentially remember the top scent, selectively forgetting the bottom scent, or tag the memory of the top scent so that it is more salient than that of the bottom scent. In further experiments of this type, we found evidence for preferential memory for the top scent even if, during the habituation trials, there was an additional scent mark of the bottom-scent donor that was not marked over (Fig. 16.1b; Wilcox & Johnston, 1995). These experiments suggest that hamsters extracted information about the relative position (top or bottom) or the relative freshness of the two individuals’ scent marks and either selectively remembered just one of them or had placed a different value on the memory of the top and bottom odours. The existence of either of these mechanisms suggests that the relative position of an individual’s scent in an over-mark is important to the perceiver and that this

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:57

Mean investigating time (s)

Eavesdropping and scent over-marking

0.005

0.0025

Fig. 16.1. The time that male hamsters spent investigating the vaginal scents of females on the test trial after habituation to a pattern of vaginal scent-marks as shown above the graphs. In both (a) and (b), males investigated the scent from the top-scent female significantly less than the novel female’s scent (n = 9 in (a); n = 10 in (b)) whereas there was no significant difference in time spent investigating the scent of the bottom-scent male and the novel scent (n = 10 in both groups), thus indicating a preferential memory for the top scent of the over-mark. The bar indicates the standard error; p values derived from t-tests. (From Wilcox & Johnston, 1995.)

information may influence subsequent social interactions between the perceiver and the individuals that deposited the scent marks. Did the subjects in these experiments actually forget the bottom scent in an over-mark or did they just attach less value or importance to it? The latter seems more likely. First, it is difficult to believe that hamsters would not remember one of two individually distinctive odours after investigating them four or five times, since a single scent is remembered at least 10 days after such exposures (Johnston, 1993). Second, later experiments showed that hamsters would remember two adjacent scents (see below). Third, in another experiment, we obtained evidence that males did have some memory of the bottom scent from an over-mark. Male hamsters were first exposed to experimenter-produced over-marks of male flank glands in the pattern of a cross during four habituation trials with 15 minutes between trials. Investigation of the crossed scents decreased significantly in all three groups (10 per group) across the habituation trials: the results for trial 1 and trial 4, respectively were 16.8 and 3.8 seconds in group 1 (t = 10.216; p < 0.001), 21.1 and 3.2 seconds in group 2 (t = 14.802; p < 0.001), and 19.3 and 3.7 seconds in group 3 (t = 7.156; p < 0.001). These groups were then tested 15 minutes after

349

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

April 12, 2005

11:57

R. E. Johnston

0.01 Mean investigating time (s)

350

0 521 582361 7

0.0007

Fig. 16.2. The time male hamsters spent investigating the flank scent of one male in the test trial after habituation to a scent over-mark in the pattern of a cross (shown above) in a series of trials. The top scent from the over-mark was investigated least, indicating habituation to and memory for this odour compared with the novel odour. The bottom scent from the scent mark was investigated an intermediate amount, suggesting some memory for this odour but one that was significantly different than that for the top scent. The bar indicates the standard error; p values derived from t-tests; 10 animals in each group. (R. E. Johnston & M. Schiller, unpublished data.)

the last habituation trial with the flank scent from just one male on the test trial: the flank scent of the top-scent male for group 1, the bottom-scent male for group 2 and a novel male for group 3. The novel flank scent was investigated most and the top scent was investigated significantly less, as in the previous tests in which there were two stimuli present in the test trial (Fig. 16.2; R. E. Johnston & M. Schiller, unpublished data). The bottom scent, however, was investigated an intermediate amount, significantly more than the top scent but significantly less than the novel scent (Fig. 16.2). This experiment suggests that hamsters do remember the scent of the bottom-scent male but the memory is not as strong or as salient, or that the behaviour based on the memory is different from that for the top-scent male. We do not know exactly in what way it is different, but we suspect that the bottom scent is less important to the subjects (see p. 359–360). Experiments on meadow voles support the notion that the odour of the bottom-scent male is devalued relative to top-scent males or novel males (Woodward et al., 2000). Before describing our experiments on how hamsters and voles determine which scent is on top, it is useful to review what we know about the mechanisms underlying discrimination between odours from different individuals. As in any other

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:57

Eavesdropping and scent over-marking sensory domain, individual recognition is accomplished by pattern recognition mechanisms: perceptual mechanisms by which animals discriminate between similar, complex stimuli (e.g. faces, voices, odours). In such processes, specific features are generally much less important than the relationships between features. In the case of odours, the pattern is generated by a large number of individual chemical compounds that occur in differing proportions across individuals (Gorman, 1976; Bagneres et al., 1991; Gamboa et al., 1996; Singer et al., 1997; Smith et al., 2001). It is these differences in proportions of chemical compounds that give each individual its distinctive odour quality. The particular chemicals that differ in proportion vary across pairs of individuals; that is, there does not seem to be a particular set of chemical compounds that are used for this purpose (Smith et al., 2001). Consequently, one might expect that a mixture of two scents would be created when one individual over-marks another’s scent that was different from either of the two original mixtures and that this new mixture would produce a new odour quality. The results of the experiments reported in the preceding paragraphs, in which subjects were habituated to over-marks, argue against this hypothesis because in the test trial hamsters showed memory for one individual but not the other (Johnston et al., 1994, 1995; Wilcox & Johnston, 1995). We have undertaken a series of experiments to try to characterize the mechanisms used to distinguish top and bottom scents in an over-mark. The strategy in these experiments was to determine if a particular kind of information in overmarks was sufficient, by itself, to promote differential responses to the odour of one donor versus the other donor. One possible cue is the relative freshness of the two scents, because the top scent of an over-mark is necessarily fresher than the bottom scent. In the experiments discussed above, the difference in the age of the two scent marks in over-marks was about 30 seconds (not more than 60 seconds); however, in nature, scent marks are likely to differ more than this in freshness (e.g. by at least tens of minutes and often by hours or days). Therefore, a series of experiments were carried out to see if differences in freshness alone would cause the differential memory effect. In addition, we wanted to be sure that animals could detect and remember the scents of two individuals that were in close proximity. In the first experiment, subjects were habituated to vaginal secretion scents from two individuals (A and B) that were placed adjacent to one another in an open cross pattern, as shown in Fig. 16.3a, and then were tested for their responses to scent A or B versus a novel scent. Males investigated both scents (A and B) less than the novel scent, indicating that they could remember the scents of two individuals from habituation trials (for flank gland scents, see Cohen et al. (2001)). Then we tested males to determine if differences in freshness between the two scents would result in differential memory for the fresher scent. The scents were again presented in the

351

352 Scent 1 group

0.001

Scent 2 group

0.005

Scent 2

Scent 1

0.0025

scent

0.025

novel

scent

0.025

Bhorade, 1998.)

April 12, 2005

scent. The bar indicates the standard error; p values derived from t-tests; 10 animals in each group. (From Johnston &

the interrupted scent and the novel scent, suggesting poorer memory for the interrupted scent than for the continuous

investigated significantly less than the novel scent but there was no significant difference in time spent investigating

habituation to scents in an ‘apparent cross’ (one continuous scent and one interrupted scent), the continuous scent was

scents, indicating that a difference in freshness does not lead to differential memory for one of them. (c) After

0 521 582361 7

habituation to two scents of differing degrees of freshness, both scents are investigated significantly less than the novel

cross, both scents were investigated significantly less than a novel scent, indicating memory for both. (b) After

secretion marks in the pattern shown above the graph. (a) After habituation to two scent marks in the pattern of an open

Fig. 16.3. The time spent by male hamsters investigating vaginal secretion scents following habituation to vaginal

Mean investigating time (s)

Habituation (a) stimuli

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor 11:57

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:57

Eavesdropping and scent over-marking open cross pattern during habituation trials, but one vaginal secretion scent was 30 seconds old and the other was 24 hours old. On the test trial, male hamsters investigated novel scents more than familiar scents of the same age, indicating memory of both of the familiar scents. The investigation of the familiar, 30 second scent did not differ from that of the familiar 24 hour scent, indicating no preferential memory for the fresher scent (Fig. 16.3b; Johnston & Bhorade, 1998; for similar data on male flank gland scent, see Cohen et al. (2001)). Similar experiments with meadow voles exposed to adjacent anogenital marks differing in age by 60 minutes yielded no significant difference in response to the donor of the fresh scent versus the donor of the 60-minute-old scent (Ferkin et al., 1999). Therefore, freshness by itself did not lead to preferential treatment of the scent of one individual. It is worth noting that four different types of odour were used in these experiments: vaginal secretions and flank glands in hamsters (Johnson & Bhorade, 1998), anogenital area scent in voles (Ferkin et al., 1999) and urine in voles (M. H. Ferkin, J. Dunsavage & R. E. Johnston, unpublished data). This suggests that the lack of an effect of freshness is not caused by the chemistry of one particular type of scent (such as the sebaceous scent from flank glands) that might change little over 24 hours. We do not know how long individually specific information lasts, but hamster flank and vaginal scents deposited on glass in the laboratory are detected and investigated after 40 and 100 days, respectively (Johnston & Schmidt, 1979). A second possible cue that might be used as a guide for preferential responses to one individual over another is the relative amount of scent or number of marks deposited by two individuals. Among many species, dominant or high-ranking individuals mark more often than subordinate or low-ranking individuals (Ralls, 1971; Eisenberg & Kleiman, 1972; Johnson, 1973; Johnston, 1975a,c; Bronson, 1976; Hurst & Rich, 1999) and the relative amount of over-marking is probably correlated with the frequency of marking. Therefore, the number of marks or over-marks could be an indication of dominance status. Even when there was a single over-mark, it might be the case that subjects would perceive that the top-scent individual deposited more scent if the top scent masked some of the bottom-scent individual’s mark. The ‘amount of scent’ and ‘number of marks’ hypotheses were, however, ruled out as explanations of the differential effects in hamsters and meadow voles. That is, hamsters still remembered the top-scent individual and meadow voles investigated the top-scent individual more even when, during the exposure to over-marks, there were more marks or more area covered by the bottomscent individual (e.g. Fig. 16.1b; Wilcox & Johnston, 1995; Johnston et al., 1997a,b; Johnston & Bhorade, 1998; Ferkin et al., 1999). In one experiment with meadow voles, for example, female subjects investigated the home cage of a male that had been briefly investigated and marked by an ‘intruder’ male. When tested in the Y-maze, females spent more time investigating the whole-body odour of the

353

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

April 12, 2005

11:57

R. E. Johnston

0.01

Mean investigating time (s)

354

0 521 582361 7

Fig. 16.4. The time that female meadow voles spent investigating whole-body odours from cotton bedding in a Y-maze after exposure to the home cage of one male that had been briefly investigated and scent marked by another male, as represented by the drawing on the left. The bar indicates the standard error; p values derived from Wilcoxon matched-pairs signed-rank test; n = 12. (From Johnston et al., 1997b.)

intruder male than that of the home-cage male, even though the home-cage male must have had many more marks and covered more area with his marks than the intruder did (Fig. 16.4; Johnston et al., 1997b). Therefore, neither hamsters nor voles seem to use the relative amount of scent or number of marks, by themselves, as a means of selective responses to other individuals. Other experiments suggest that it is some type of information from the area of overlap, or the area of overlap compared with adjacent areas with non-overlapped scents, that leads to preferential responses to the top-scent individual. One line of evidence for this conclusion is that among all of the experiments that we have done with hamsters and meadow voles, using a variety of testing methods, we have found differential responses in the test phase only when there were scent marks that overlapped during the exposure phase (with one exception, see below) (Johnston et al., 1994, 1995, 1997a,b; Johnston & Bhorade, 1998; Wilcox & Johnston, 1995; Ferkin et al., 1999; Cohen et al., 2001). A second line of evidence for the importance of scent overlap comes from experiments specifically designed to test whether overlap of two individuals’ scent marks was necessary for differential responses to the two animals or to their odours. For example, we first habituated

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:57

Eavesdropping and scent over-marking

Fig. 16.5. The time that female meadow voles spent investigating whole-body odours from cotton bedding in a Y-maze after an exposure to scent marks in the patterns shown above the graph. (a) After exposure to a large area of scent from one male with a small spot of scent from another male on top, females (n = 12) spent significantly more time investigating the top-scent male. (b) When the small spot was placed in a clean area of the slide, so that there is no overlap, females (n = 10) show no significant difference in investigation time. Bar indicates the standard error; p values derived from Wilcoxon test. (From Johnston et al., 1997a.)

male golden hamsters either to two scent marks in the pattern of a cross or to two scent marks in a pattern of an ‘open cross’ – that is, there was no overlap in the middle, just an unscented space. On the test trial, males exposed to this latter pattern showed an equivalent response to the scent of the two donors, treating them both as familiar (e.g. Fig. 16.3a; Cohen et al., 2001), but males exposed to the crossed scents showed memory for the top scent but not for the bottom scent (Fig. 16.1a; Johnston et al., 1995; Johnston & Bhorade, 1998; Cohen et al., 2001). In another example, one group of female meadow voles were exposed for 15 minutes to a microscope slide largely covered by the anogenital scent of one male but with a small spot of scent from a second male placed on top. During the test trial, these females spent significantly more time investigating the whole-body odours of the small-spot, top-scent male (Fig. 16.5a). If, however, the second male’s scent was placed in a clean ‘hole’ surrounded by scent from the first male, female meadow voles showed no significant difference in response to the odours of the two males (Fig. 16.5b; Johnston et al., 1997a). In both hamsters and voles, an area of overlap was necessary to produce a differential response to their odours. We do not know

355

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

356

0 521 582361 7

April 12, 2005

11:57

R. E. Johnston exactly how hamsters or voles use the area of overlap to determine which scent is on top, but one hypothesis is that they compare the odour quality of the overlap region with that in the adjacent areas. Assuming that there is partial mixing in the area of overlap, the odour in this region should smell more like the top-scent male than the bottom-scent male. In the hamster experiment, animals could compare the area of overlap with the two adjacent, non-overlapped scents and determine which was the top-scent by which scent was closest in odour quality to the region of overlap. This mechanism would not work for the vole experiment, however, because in this situation there is just an area of overlap and an area of scent from the bottom-scent male. Finally, there is some evidence that hamsters may be able to use the geometric relationships between the two scent marks (interrupted versus uninterrupted streaks) to determine which is the top scent. We reasoned that, by analogy with depth perception in vision, if one scent occludes another, it must be on top. If, when investigating scent over-marks, hamsters or other animals develop a representation of the geometrical layout of the marks, they might be able to determine which of two scents was on top by determining which one occluded the other. To test this possibility while at the same time eliminating cues from a region of overlap, we investigated how male hamsters would respond after being habituated to a pattern of scent marks in which it might appear that one individual’s scent occluded the other but in fact there was no region of overlap; rather, there was one continuous scent and, at right angles to it, two scent marks that approached this scent closely but did not touch it (Fig. 16.3c). Male hamsters showed a preferential memory for the continuous scent compared with the interrupted scent for vaginal scent marks (Fig 16.3c; Johnston & Bhorade, 1998) and male flank marks (Cohen et al., 2001). This is the only case in which a region of overlap was not necessary to obtain a differential response to one animal or its odours after exposure to scent marks of two individuals. In contrast, meadow voles show no evidence of using the same kind of spatial information (Ferkin et al., 1999). Perhaps the primary reason for this species difference is that, whereas hamsters deposit marks in linear streaks, voles’ marks are more often small spots or larger irregularly shaped areas (especially urine marks) and it is not obvious how spots and blobs could provide spatial cues about which scent was on top. The results summarized above indicate that at least two species of rodents have evolved specialized mechanisms for the perception and analysis of scent overmarks. These abilities are quite striking – indeed, amazing – and we do not yet fully understand them. Many other species have probably evolved similar mechanisms, but, because of the diversity of functions served by scent marking, it is not likely that all species have evolved such mechanisms. The existence of these abilities,

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:57

Eavesdropping and scent over-marking however, should stimulate us to think about other ways in which animals might extract information from arrays of scent marks.

Specialized mechanisms for production of over-marks If over-marking has functions that are distinct from the functions of regular scent marking, specialized mechanisms should have evolved to promote accurate placement of scent on top of that of another individual: that is, mechanisms to target the scent marks of others. Numerous observations of a variety of mammalian species in nature indicate that such targeted over-marking does occur (Ralls, 1971; Eisenberg & Kleiman, 1972; Johnson, 1973; Brown & Macdonald, 1985). Several recent field studies have provided quantitative data on the extent of over-marking in natural environments (Kappeler, 1998; Lazaro-Perea et al., 1999). In perhaps the most dramatic case, male diademed sifakas Propithecus diadema overmark 94% of the scent marks deposited by their mates (Powzyk, 2002). Observations of over-marking in nature are convincing evidence for a targeting mechanism because the probability of even one over-mark occurring by chance is extremely low, given that there are many possible places to mark. In addition, most animals that over-mark also engage in other related activities when encountering scent marks from another individual, such as careful investigation, pawing or scratching the ground, biting the bark of the tree on which the marks are deposited or becoming visibly aroused or agitated, indicating that they are reacting to this odour and are focused on it. Virtually all experimental studies of the mechanisms underlying scent marking and over-marking, however, have been carried out in captivity or in laboratory settings, where space is limited. In such circumstances, it is much more difficult to determine if over-marks occur by targeting mechanisms or occur by chance. For example, a common (but not universal) observation is that a particular odour (e.g. urine, flank gland) will increase the frequency of scent marking with the same scent (Ralls, 1971; Johnston, 1975a, 1977; Hurst, 1990). Some of these scent marks will be over-marks. In the limited spaces used in the laboratory, however, it is difficult to know whether the increase in the number of over-marks occurs just because the rate of marking increased and, by chance, some were deposited in places that had been marked previously, or because a specific targeting mechanism exists. Similarly, other mechanisms that increase marking frequency, such as changes in hormone levels with reproductive state, can increase the frequency of over-marking. Furthermore, many species have preferred types of site for marking, such as particular topological features (visually prominent rocks or vegetation, intersections of trails, particular types of plant or tree, water holes or other rare

357

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

358

0 521 582361 7

April 12, 2005

11:57

R. E. Johnston resources). Hamsters in laboratory environments, for example, prefer to mark in confined spaces, such as the corners of a rectangular arena and just outside of their burrow entrance (Johnston, 1975c). An increase in the number and the percentage of over-marks could occur if the rate of marking is increased by stimulation and there are a limited number of preferred places in which to mark. We designed a simple method to determine if individuals have a specific mechanism for targeting another individual’sscent mark to produce an over-mark, which we have applied to both flank marking and vaginal marking by hamsters (R. E. Johnston, S. Szmuilowicz, D. J. Mayeaux, S. K. Barot, & N. S. Schwarz, unpublished data). This involves comparing the number of marks that are deposited over scent marks to the number of marks deposited over mirror-symmetric locations in the same arena that have no scent (imaginary marks). Since the types of locations are identical and are in the same arena, this method controls for both the problem of preferred locations and the problem of odour-stimulated increase in overall marking frequency. In one recent experiment of this type (S. K. Barot, N. S. Schwarz & R. E. Johnston, unpublished data), we found that the mean number of flank marks by 12 male hamsters that overlapped another male’s flank marks was 6.0, whereas the number of flank marks that overlapped imaginary marks (symmetric locations but clean) was 2.8, (t = 3.171; p < 0.01). Females, however, did not selectively flank mark over the flank scents of other females or males. We are currently replicating and refining these experiments, but these initial experiments suggest the existence of a specific mechanism in adult, male golden hamsters that targets the flank marks of other adult males but not the marks of juvenile males or females. These results suggest that flank over-marking by males is a sexually selected characteristic related to male–male competition, but that flank marking by females is a more generalized kind of broadcast signalling.

Functions of over-marking and eavesdropping on over-marks What is the function of scent over-marking, and why have hamsters and meadow voles evolved specialized mechanisms for evaluating over-marks? We have hypothesized that in meadow voles and golden hamsters, both of which live solitarily, scent over-marking by males may be a type of intrasex competition in which each male targets its male neighbours (Johnston et al., 1997a,b; Johnston & Bhorade, 1998; Johnston, 1999; Cohen et al., 2001). This mutual over-marking presumably reflects a struggle for dominance between males. If each male is, at least during the breeding season, continuously trying to keep its scent on top of the scents of its male neighbours, over-marking should be an energetically costly activity. I am not aware of any data on the energetic costs of over-marking in natural settings, but the cost of marking for defence of territory is important for

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:57

Eavesdropping and scent over-marking some species (Mills et al., 1980; Gorman & Mills, 1984; Gorman, 1990; Gosling et al., 2000; Gosling & Roberts, 2001). The degree to which an individual is successful at keeping its marks on top of those of its neighbours should be an honest indicator of phenotypic vigour and quality and, to the extent that these characteristics are dependent on genotype, genetic quality. Information gathered by third parties from over-marks about ‘whose scent is on top’ should, therefore, be valuable information for mate-choice decisions by females and it may influence interactions between like-sex rivals as well. Is there evidence to support these hypotheses? The evidence suggests that analysis of scent over-marks does affect preferences for opposite-sexed individuals and that such preferences are likely to be important in mate-choice decisions. Most of the tests we carried out with meadow voles described above were preference tests rather than habituation tests: female voles were first exposed to anogenital or urine over-marks from two males for 15 minutes; 10 minutes later they were tested for their preferences for the whole-body odours (cotton bedding material, that contained additional body odours and possibly urine odours) of donor males in a Y-maze. Females spent more time close to and investigating the whole-body odour of the top-scent male than that of the bottom-scent male in a variety of experiments (Figs. 16.4, 16.5a; Johnston et al., 1997a,b). Males and females were housed in long-day light cycles and were in reproductive condition, so this preferential behaviour suggests that females would prefer the top-scent males as mates (Johnston et al., 1997b). In addition, we also exposed female meadow voles to naturally deposited over-marks of males and found that females again preferred the whole-body odours (cotton bedding) from top-scent males over that of bottom-scent males (Fig. 16.4; Johnston et al., 1997b). It is worth noting that in several of these experiments there was far more of the bottom-scent male’s odour present during the exposure phase than of the top-scent male’s odour, but females still preferred the top-scent male (Figs. 16.4 and 16.5; Johnston et al., 1997a,b). We interpret these results as a preference for the top-scent male because the test stimulus (whole-body odours in bedding material) incorporated whole set of body odours and was a different stimulus from the one odour collected by us and presented during the exposure phase. Furthermore, the test was carried out in a different environment from the one used for the exposure to the over-marks (Y-maze rather than subject’s home cage). These results cannot be explained as merely the result of habituation for two reasons. First, females were exposed to the odours of both stimulus animals during the presentation of the over-marks. Second, our studies using a habituation paradigm showed that after habitation to an over-mark subjects spent less time investigating the scent from the top-scent male than that from the bottom-scent male, which is the opposite of what we found in our Y-maze experiments.

359

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

360

0 521 582361 7

April 12, 2005

11:57

R. E. Johnston Pilot experiments with hamsters suggest that after females eavesdrop on males’ over-marks they show preferences for the top-scent over the bottom-scent male. Female hamsters living in a seminatural enclosure in the laboratory were allowed to explore regions of the environment that had been explored and marked by two males, but one male was always first (male A) and the other was always second (male B); consequently, when the female explored the arena, the marks of male B should have been on top. (Although we did observe males marking, we did not have a method of determining the location of the marks with sufficient accuracy to be certain for all cases that they overlapped or not.) Females were tested in the seminatural environment approximately 18 hours prior to receptivity (when they are soliciting males) and also when they were receptive. The stimulus males (A and B) were confined in small, wire-mesh enclosures. Females spent more time investigating male B (that had explored and marked the arena second) than male A (there first): combined investigation time for day before oestrous was 238.1 ± 39.8 seconds and for day of oestrous was 159.1 ± 24.6 seconds (degrees of freedom (df ) = 9; t = 2.39; p = 0.04). Females also vaginal marked more in the vicinity of male B (7.4 ± 1.4 seconds for male B and 4.9 ± 1.0 seconds for male A; df = 9; t = 2; p = 0.057 (S. K. Barot, N. S. Schwarz & R. E. Johnston, unpublished data)). In other experiments with hamsters, males explored an arena in which there were six vaginal secretion marks of female B overlapping those of female A and two marks of female A by themselves. After 45–60 minutes, males were tested for their preference in a simultaneous choice apparatus (Steel, 1984). Males spent more time sniffing the top-scent female B (108 ± 4.8 seconds) than female A (88.7 ± 5.4 seconds; df = 15; t = 2.956; p = 0.01 (R. E. Johnston & C. Lee, unpublished data)). These results suggest that males may prefer as mates females that successfully over-mark other females with vaginal secretions; such over-marking by females may reflect their vigour and ability to defend their burrow and food hoard from other females. Maintaining a safe burrow and a food hoard could lead to greater pup survival and if so it would be advantageous for males to mate preferentially with such females. Experiments with other species also indicate that information gained by investigation of scent marks influences mate preferences. After female house mice explored areas in which the marks of only one male were present (exclusively marked territory) and another area in which a second male had also marked (and overmarked), females preferred the males that had the exclusively marked areas (Rich & Hurst, 1998). Females also preferred a male that had counter-marked another male to one that had been counter-marked (Rich & Hurst, 1999; see also Ch. 11). Similar supporting evidence comes from studies with a primate, the pygmy loris Nycticebus pygmaeus (Fisher et al., 2003). Male lorises over-mark the urine marks of other males. Females were exposed to naturally produced male over-marks over a period of 14–20 weeks, until the females came into oestrous. Several measures

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:57

Eavesdropping and scent over-marking of female behaviour (proximity and orientation to males, investigation of odours and socio-sexual behaviour) all indicated a preference for the top-scent male by oestrous females. The evidence so far suggests that in several species eavesdropping on scent over-marks influences mate preferences, perhaps because over-marks usually reflect the relative phenotypic and possibly genotypic quality of individuals. Much more research needs to be done to determine how widespread this phenomenon is, the degree to which eavesdropping on over-marks influences mate choice in natural settings, and the factors leading to the evolution of over-marking as a sexually selected trait that provides useful information about potential mates. It would also be valuable to determine the degree to which over-marking correlates with other measures of behaviour or physiology that are related to the genotypic or phenotypic quality of individuals within a population. Although I have stressed the usefulness of over-marks as a means by which third parties might assess opposite-sexed individuals, analysis of over-marks may also influence competitive interactions between third parties of the same sex as those that deposited the over-marks (e.g. in competition for territory, food or water resources, nesting sites or mates) could be seen as more likely to dominate in confrontations. Also yet to be investigated are the effects of over-marks on the individuals that are engaged in over-marking contests. No doubt there are many interesting phenomena yet to be discovered.

Field tests of specific hypotheses about over-marking

Both field and laboratory studies are needed to test specific hypotheses about the functions of over-marking. If over-marking is a type of advertising contest between like-sex individuals for mates for example, one would expect that over-marking the scent of potential rivals (adults of the same sex in reproductive condition) would be especially prevalent. There are many studies that have shown that scent marking in general is stimulated by like-sex adults, but relatively few studies that have measured over-marking and even fewer that have compared over-marking towards rivals and non-rivals. Likewise, if over-marking is a means of mate-guarding, one would expect it to be predominantly done by the sex that is most actively competing for mates (usually males) and that males would mark over the scent of their mate, as has been observed in the pronghorn antelope Antilocapra americana (Byers & Bekoff, 1986). Particularly valuable should be field experiments in which scent marks are manipulated in ways similar to song playbacks (Sliwa & Richardson, 1998), for example, by experimentally over-marking some residents but not others, or by using the scent of one resident to over-mark the scents marks of rivals.

361

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

362

0 521 582361 7

April 12, 2005

11:57

R. E. Johnston Audience effect on scent marking? Audience effects refer to an alteration in behaviour because of the presence of specific individuals or classes of individuals: the audience (see Ch. 4). Generally, the presence of other animals is defined as the subject being able to see or hear these other animals. Scent provides interesting possibilities that have not been systematically explored. For example, might the behaviour of one individual toward a second individual be altered by the presence of scent from some individuals but not by scent from others? If so, would fresh scent be more effective than older scent? Although fresh scent is not exactly the same as being observed or heard by another individual, it could indicate the recent presence of an individual, and thus it could indicate a high probability of being observed or discovered by the animal that deposited the scent. One can imagine that very fresh scent of a dominant individual, for example, could inhibit some types of aggressive behaviour of a subordinate or that fresh scent of a male paired with that of a female might inhibit another male from courting her. In a more traditional sense, do animals alter their scent marking behaviours based on the social environment? Both of the species that we have studied in detail (golden hamsters and meadow voles) live solitarily and in nature one would predict that they usually mark when alone. In laboratory settings, golden hamsters do not usually flank mark in the presence of another hamster but often mark vigorously just after social encounters (Johnston, 1975a). This could be considered a type of audience effect, but not one that depends on the presence or absence of a specific audience. The one exception that I have observed to this pattern of marking when alone is that subordinate males do sometimes mark just inside their burrow entrance when they defend their burrows from a dominant male (Johnston, 1975c). This behaviour, however, seems to be readily explained by flank marking as an agonistic behaviour involved in the defence of the burrow or food hoard: the notion of an audience effect does not aid our understanding. Female hamsters are stimulated to vaginal mark in the presence of a male or a male’s odours but not in the presence of a female or her odours: they mark most frequently in the period 12–24 hours before sexual receptivity (Johnston, 1977, 1979). Once again, this pattern seems to be primarily related to the function of sexual advertisement and is not a modulation of behaviour based on specific relationships between individuals. Perhaps the most likely situations in which audience effects might be observed would be in gregarious species. For example, in the ring-tailed lemur, Lemur catta, individuals are attentive to the marking behaviour of others and 62% of all marks are investigated within 30 seconds of deposition; in 89% of these cases the original mark is over-marked (Kappeler, 1998). In this study, however, no evidence was

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:57

Eavesdropping and scent over-marking found for an influence of social context on the likelihood of scent marking. In species that live in groups that have a clear dominance hierarchy, such as wolves Canis lupus or free-ranging packs of domestic dogs, dominant individuals mark much more than subordinates (Peters & Mech, 1975; Bekoff, 1979). One could imagine that subordinates might alter their marking based on the presence or absence of a more dominant individual, but I am not aware of any data that explicitly supports this speculation. In some group-living species, such as dwarf mongooses Helogale undulata, all members of the group may mark in the same place, especially when first emerging from the burrow in the morning (Rasa, 1973). Marking by some individuals could be stimulated by observing others mark (social facilitation) or by the mere presence of others in the group at that time of day (a possible audience effect). Another situation in which an audience effect might be observed is in cases in which the scent-marking behaviour serves as a visual signal as well as a means of depositing a chemical signal. For example, males among all species of gazelles engage in stereotyped visual displays when marking their territories. This type of marking has been called ‘demonstrative marking’ because of its obvious nature and probable value as a simultaneous visual and olfactory signal (Estes, 1967). Again, one could hypothesize that the occurrence or vigour of such marking displays might vary depending on the relationships between the marker and specific males that were present or the presence, or absence, of females. I am not aware of any data that have been analysed in this context, however. The alpha male and female dogs of a pack are more likely to mark with a raised leg urination after observing another dog do so than are lower-ranking individuals (Bekoff, 1979), but again it is not clear if this is a competitive reaction to observing the mark or is dependent on the presence of the audience of the other members of the pack. Networks, cognition and individual recognition: speculations on species differences in underlying mechanisms The notion of a communication network is important because it emphasizes that individuals are a part of a community of interacting individuals (for an early version of this view, see Estes (1969)). This is true even in species in which individuals spend most of their time by themselves. The concepts of eavesdropping and audience effects draw attention to two specific ways that the social context can modulate the behaviour of individuals. Although evidence demonstrating eavesdropping or audience effects does not imply anything specific about the mechanisms underlying these effects, most researchers working in this area seem to assume that the animals they work with treat one another as unique individuals with unique sets of distinctive characteristics. This is an inference about cognitive

363

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

364

0 521 582361 7

April 12, 2005

11:57

R. E. Johnston and neural processes that I will refer to as true individual recognition or having a concept of other individuals. One important aspect of having such an attribute is that animals should have memories consisting of several types of information about other individuals and these memories should be stored as organized units or representations of individuals (Johnston & Jernigan, 1994; Johnston & Bullock, 2001). This type of representation contrasts with other, simpler mechanisms that are nonetheless sufficient to explain many of the findings that are referred to as individual recognition, neighbour recognition, etc. Although the mechanisms involved in recognition and memory are not directly observable, it is possible to characterize these mechanisms by appropriate behavioural and physiological measurements. The fields of cognitive psychology and cognitive neuroscience, for example, depend on this kind of analysis. With regard to recognition of individuals by non-human animals, for example, it is possible to discover the kinds of information that one animal knows about another (Johnston & Bullock, 2001). Further, this information can help to explain why individuals behave the way they do. Indeed, the complexity of an individual’s knowledge about other individuals is likely to have a profound influence on how that individual behaves in social interactions with them (e.g. Chs. 24 and 25). I have previously argued that many demonstrations of individual recognition in the field and in the laboratory do not allow us to discriminate between merely recognizing familiar versus unfamiliar cues or combinations of cues and recognizing individuals as unique entities (Johnston & Jernigan, 1994; Johnston & Bullock, 2001). For example, in the literature on neighbour recognition by song in birds, the majority of studies merely demonstrate that a territorial male responds more strongly to a novel song than a familiar song, or a novel song–direction combination than to a familiar song–direction combination. They do not provide proof that the birds recognize their neighbours as individuals. A slightly more complex mechanism might be categorization of information into heterogeneous categories (e.g. a frequently heard song versus a rarely heard song) but again, not a categorization based on individuals as the unit of analysis (Barrows et al.,1975; Caldwell, 1985). Cases in which birds engage in repertoire matching, in contrast, suggest the existence of true individual recognition. That is, when a song sparrow Melospiza melodia hears a neighbour sing one song type and then responds with a song type that he shares with this neighbour but that is different from the one that the neighbour just sang (Beecher et al., 1996), this suggests that the male knows several characteristics of his neighbour and that they differ for different neighbours. This type of observation provides evidence that song sparrows respond to neighbours as unique individuals. Another example that demonstrates complex, multicomponent representations of individuals comes from work with golden hamsters, in which we showed that after males had interacted with several females they had

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:57

Eavesdropping and scent over-marking memories of these females that incorporated at least three different odours. Using an habituation task, we showed that habituation to one of these odours resulted in habituation to the other odours as well. That is, when becoming habituated to a particular stimulus the males also became habituated to the individual and thus other features of this individual (Johnston & Jernigan, 1994; Johnston & Bullock, 2001). These effects do not occur if the males have not interacted with the stimulus females and, therefore, have not had an opportunity to learn about the features of these females; that is, the effects are not a result of inherent similarities across odours and consequent generalization across odours. Similarly, an experimental demonstration of eavesdropping may result from true individual recognition or it could result from a simpler mechanism. For example, if one male fighting fish observes a fight between two other males of the species, he could remember that fish A with the purple fringe on this dorsal fin and red stripes on his tail and a distinctive wiggle in his display is much more aggressive than fish B, characterized by all red fins and tail but a purple spot in the middle of the tail fin. The observer fish could have memories of these two individuals, and each memory would consist of an integrated memory of that male’s physical and behavioural characteristics. A simpler type of memory would, however, also be sufficient to explain the effects. The observer fish might associate ‘purple edge above and red stripes in back’ with fear and ‘all red’ with lack of fear. He could learn an association between a few specific cues and fear or danger but not have memories of individuals as such. Both types of memory would result in differential responses to the two individuals, but the mechanisms underlying the responses would be different and indicate different levels of neural and cognitive complexity. Additional experiments could provide evidence for or against the existence of true individual recognition. For example, does the subject react differently to two individuals that were observed to have similar experiences (e.g. lost a fight) but some aspects of the interaction were different (e.g. how quickly the animal lost)? Does the subject react differently to the two individuals in a non-aggressive context? Does the subject cross-habituate to different features (e.g. odours, sounds, other visual information) from the same individuals (Johnston & Jernigan, 1994; Johnston & Bullock, 2001)? There is probably a continuum of complexity in the types of representation that animals have of other individuals, and these differences in complexity should correlate roughly with the complexity of the social behaviour of different species and taxonomic groups. At higher levels of complexity are species that not only remember individuals as such but also remember something about the relationships between different individuals (Cheney & Seyfarth, 1990a, b; Chs. 24 and 25). At the most complex end of the continuum would be species, such as humans and perhaps some other highly social animals, that have partially or well-developed

365

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

366

0 521 582361 7

April 12, 2005

11:57

R. E. Johnston abilities to understand that other individuals have different knowledge or intentions than themselves; that is, they have a so-called ‘theory of mind’ (Premack & Woodruff, 1978; Wimmer & Perner, 1983; Cheney & Seyfarth, 1990a,b; Ch. 25). One crucial task is to identify the simplest level of representation that can explain a particular phenomenon because any phenomena could be explained by a more complex mechanism. Identifying the simplest mechanism that could be used can provide a starting place for a taxonomy of the kinds of representation that individuals have of other individuals, species differences in these representations and hypotheses about the neural mechanisms underlying them. Ultimately, it may help us to understand the evolution of social behaviour. The complexity of representations that individuals have of other individuals is important in the context of communication networks because these representations will influence the way in which individuals interact with others, the kinds of information they extract from observing interactions between others and the effects that such information has on their own behaviour. This, in turn, will influence the nature of the networks that develop and the ways that individuals interact within those networks (Chs. 24 and 25). I suggest that animals with complex, integrated representations of individuals will have a number of advantages over those with simpler mechanisms. For example, complex representations contain more information and should reduce errors in recognition, especially recognition over long intervals during which changes in relevant cues may have occurred through age, injury, nutrition or hormonal status. More complex representations may also facilitate the evaluation of relationships between two or more individuals obtained via observation of interactions.

Summary and future directions The observations described in this chapter suggest that individuals in some species obtain information from over-marks about the relative quality of the individuals that deposited these marks and that this information influences subsequent responses to those individuals, thus providing evidence for eavesdropping (but see Ch. 11). In addition, some species appear to have specialized sensory mechanisms for the evaluation of scent over-marks and specialized mechanisms that promote accurate placement of a scent mark over the scent of another individual (targeting). At present there is little evidence for audience effects on scent marking, but this may be because few observers have looked for such effects. More studies are desperately needed on the functions of scent over-marking, eavesdropping on over-marks and the effects of such eavesdropping, especially observations and experiments in natural settings. Also important are comparative studies on a set of related species with a rich, diverse repertoire of scentmarking behaviours. Little is understood about the variety of functions that scent

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:57

Eavesdropping and scent over-marking over-marking has or about the kinds of information that may be obtained from such over-marks. Studies of eavesdropping on singing interactions in birds suggest some interesting possibilities for further experiments with scent over-marking. For example, use of information gathered by eavesdropping to direct extra-pair behaviour (Mennill et al., 2002; Chs. 2 and 7). Among mammals, similar behaviour may be found in monogamous pairs. In species in which females usually mate with several partners, individual females might reduce their interactions with additional males if the first mating partner was known to be highly successful in over-mark competitions; in contrast, females that live near males that are less successful in over-marking competitions might be more vigorous in advertising for, or in seeking out, other males. That is, a female might be influenced by knowledge of the history of over-marking interactions between numerous males in her vicinity.

Acknowledgements Thanks to E. Regan, Peter McGregor, Jane Hurst and two anonymous reviewers for comments on the manuscript. Thanks to Joan Johnston for help with graphics and other technical help.

References Bagneres, A.-G., Killian, A., Clement, J.-L. & Lange, C. 1991. Interspecific recogntion among termites of the genus Reticulitermes: evidence for a role for the cuticular hydrocarbons. Journal of Chemical Ecology, 17, 2397–2420. Barrows, E. M., Bell, W. J. & Michner, C. D. 1975. Individual odor differences and their social functions in insects. Proceedings of the National Academy of Sciences, USA, 72, 2824–2828. Beecher, M. D., Stoddard, P. K., Campbell, S. E. & Horning, C. L. 1996. Repertoire matching between neighboring song sparrows. Animal Behaviour, 51, 917–923. Bekoff, M. 1979. Scent marking by free-ranging domestic dogs: olfactory and visual components. Biology of Behavior, 4, 123–139. Bronson, F. H. 1976. Urine marking in mice: causes and effects. In: Mammalian Olfaction, Reproductive Processes and Behavior, ed. R. L. Doty. New York: Academic Press, pp. 119–141. Brown, R. E. & Macdonald, D. W. 1985. Social Odours in Mammals, Vols. 1 & 2. Oxford: Oxford University Press. Byers, J. A. & Bekoff, M. 1986. What does ‘kin recognition’ mean? Ethology, 72, 342–345. Caldwell, R. L. 1985. A test of individual recognition in the stomatopod, Gonodactylus festae. Animal Behaviour, 33, 101–106. Cheney, D. L. & Seyfarth, R. M. 1990a. How Monkeys See the World. Chicago, IL: University of Chicago Press. 1990b. The representation of social relations by monkeys. Cognition, 37, 167–196.

367

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

368

0 521 582361 7

April 12, 2005

11:57

R. E. Johnston Cohen, A. B., Johnston, R. E. & Kwon, A. 2001. How golden hamsters (Mesocricetus auratus) discriminate top from bottom flank scents in over-marks. Journal of Comparative Psychology, 115, 241–247. Eisenberg, J. F. & Kleiman, D. 1972. Olfactory communication in mammals. Annual Review of Ecology and Systematics, 3, 1–31. Estes, R. D. 1967. The comparative behavior of Grant’s and Thomson’s gazelles. Journal of Mammalogy, 48, 189–209. 1969. Territorial behavior of the wildebeest (Connochaetes taurinus Bruchell, 1823). Zeitschrift f¨ ur Tierpsychologie, 26, 284–370. Ferkin, M. H., Dunsavage, J. & Johnston, R. E. 1999. What kind of information do meadow voles, Microtus pennsylvanicus, use to distinguish between the top and bottom scent of an overmark? Journal of Comparative Psychology, 113, 43–51. Fischer, R. B. & McQuiston, J. 1991. A possible role for Syrian hamster, Mesocricetus auratus, vaginal secretion in inter-female competition. Animal Behaviour, 42, 949–954. Fisher, H. S., Swaisgood, R. R. & Fitch-Snyder, H. 2003. Countermarking by male pygmy lorises (Nycticebus pygmaeus): do females use odor cues to select mates with high competitive ability? Behavioral Ecology and Sociobiology, 53, 123–130. Gamboa, G. J., Grudzien, T. A., Espelie, K. E. & Bura, E. A. 1996. Kin recognition pheromones in social wasps: combining chemical and behavioural evidence. Animal Behaviour, 51, 625–629. Gorman, M. L. 1976. A mechanism for individual recognition by odour in Herpestes auropunctatus (Carnivora: Viverridae). Animal Behaviour, 24, 141–145. 1990. Scent marking strategies in mammals. Revue suisse de Zoologie, 97, 3–29. Gorman, M. L. & Mills, M. G. L. 1984. Scent marking strategies in hyaenas (Mammalia). Journal of Zoology, 202, 535–547. Gosling, L. M. & Roberts, S. C. 2001. Scent-marking by male mammals: cheat-proof signals to competitors and mates. Advances in the Study of Behavior, 30, 169–217. Gosling, L. M., Roberts, S. C., Thorton, E. A. & Andrew, M. J. 2000. Life history costs of olfactory status signalling in mice. Behavioral Ecology and Sociobiology, 48, 328–332. Heymann, E. W. 1998. Sex differences in olfactory communication in a primate, the moustached tamarin, Saguinus mystax (Callitrichinae). Behavioral Ecology and Sociobiology, 43, 37–45. Hurst, J. L. 1990. Urine marking in populations of wild house mice Mus domesticus Rutty. II. Communication between females. Animal Behaviour, 40, 223–232. Hurst, J. L. & Rich, T. J. 1999. Scent marks as competitive signals of mate quality. In: Advances in Chemical Signals in Vertebrates, ed. R. E. Johnston, D. M¨ uller-Schwarze & P. W. Sorensen. New York: Kluwer Academic/Plenum, pp. 209–225. Johnson, R. P. 1973. Scent marking in mammals. Animal Behaviour, 21, 521–535. Johnston, R. E. 1975a. Scent marking by male hamsters: I. Effects of odors and social encounters. Zeitschrift f¨ ur Tierpsychologie, 37, 25–98. 1975b. Scent marking by male hamsters II. The role of flank gland odor in the causation of marking. Zeitschrift f¨ ur Tierpsychologie, 37, 138–144.

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:57

Eavesdropping and scent over-marking 1975c. Scent marking by male hamsters III. Behavior in a semi-natural environment. Zeitschrift f¨ ur Tierpsychologie, 37, 213–221. 1977. The causation of two scent marking behaviour patterns in female golden hamsters. Animal Behaviour, 25, 317–327. 1979. Olfactory preferences, scent marking and ‘proceptivity’ in female hamsters. Hormones and Behavior, 13, 21–39. 1985. Communication. In: The Hamster: Reproduction and Behavior, ed. H. I. Siegel. New York: Plenum Press, pp. 121–124. 1990. Chemical communication golden hamsters: from behavior to molecules and neural mechanisms. In: Contemporary Issues in Comparative Psychology, ed. D. A. Dewsbury. Sunderland, MA: Sinauer, pp. 381–409. 1993. Memory for individual scent in hamsters (Mesocricetus auratus) as assessed by habituation methods. Journal of Comparative Psychology, 107, 201–207. 1995. Responses to individual signatures in scent counter-marks: a striking new phenomenon. In: Chemical Signals in Vertebrates 7, ed. R. Apfelbach, D. M¨ uller-Schwarze, K. Reutter & E. Weiler. Oxford: Elsevier, pp. 361–369. 1999. Scent over-marking: how do hamsters know whose scent is on top and why should it matter? In: Advances in Chemical Communication in Vertebrates, ed. R. E. Johnston, D. M¨ uller-Schwarze & P. W. Sorensen. New York: Kluwer Academic/Plenum, pp. 227–238. Johnston, R. E. & Bhorade, A. 1998. Perception of scent over-marks: novel mechanisms for determining which individual’s mark is on top. Journal of Comparative Psychology, 112, 230–243. Johnston, R. E. & Bullock, T. A. 2001. Individual recognition by use of odours in golden hamsters: the nature of individual representations. Animal Behaviour, 61, 545–557. Johnston, R. E. & Jernigan, P. 1994. Golden hamsters recognize individuals, not just individual scents. Animal Behaviour, 48, 129–136. Johnston, R. E. & Schmidt, T. 1979. Responses of hamsters to scent marks of different ages. Behavioral and Neural Biology, 26, 64–75. Johnston, R. E., Derzie, A., Chiang, G., Jernigan, P. & Lee, H.-C. 1993. Individual scent signatures in golden hamsters: evidence for specialization of function. Animal Behaviour, 45, 1061–1070. Johnston, R. E., Chiang, G. & Tung, C. 1994. The information in scent over-marks of golden hamsters. Animal Behaviour, 48, 323–330. Johnston, R. E., Munver, R. & Tung, C. 1995. Scent counter marks: selective memory for the top scent by golden hamsters. Animal Behaviour, 49, 1435–1442. Johnston, R. E., Sorokin, E. S. & Ferkin, M. H. 1997a. Female voles discriminate males’ over-marks and prefer top-scent males. Animal Behaviour, 54, 679–690. 1997b. Scent counter-marking by male meadow voles: females prefer the top-scent male. Ethology, 103, 443–453. Jolly, A. 1966. Lemur Behavior. Chicago, IL: University of Chicago Press. Kappeler, P. M. 1998. To whom it may concern: the transmission and function of chemical signals in Lemur catta. Behavioral Ecology and Sociobiology, 42, 411–421.

369

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

370

0 521 582361 7

April 12, 2005

11:57

R. E. Johnston Lazaro-Perea, C., Snowdon, C. T. & Arruda, M. F. 1999. Scent-marking behavior in wild groups of common marmosets (Callithrix jacchus). Behavioral Ecology and Sociobiology, 46, 313–324. Lisk, R. D., Ciaccio, L. A., & Catanzaro, C. 1983. Mating behaviour of the golden hamster under seminatural conditions. Animal Behaviour, 31, 659–666. Mennill, D. J., Ratcliffe, L. M. & Boag, P. T. 2002. Female eavesdropping on male song contests in songbirds. Science, 296, 873. Mertl, A. S. 1977. Habituation to territorial scent marks in the field by Lemur catta. Behavioral Biology, 21, 500–507. Mills, M. G. L., Gorman, M. L. & Mills, M. E. J. 1980. The scent marking behaviour of the brown hyena Hyaena brunnea. South African Journal of Zoology, 15, 240–248. Moore, J. D. & Byers, J. A. 1989. The function of scent marking by males on female urine in pronghorns. Journal of Mammalogy, 70, 812–814. Peters, R. & Mech, L. D. 1975. Scent marking in wolves. American Scientist, 63, 628–637. Powzyk, J. A. 2002. Scent wars. Natural History, 111, 82–85. Premack, D. & Woodruff, G. 1978. Does the chimpanzee have a theory of mind? Behavioral and Brain Sciences, 4, 515–526. Ralls, K. 1971. Mammalian scent marking. Science, 171, 443–449. Rasa, O. A. E. 1973. Marking behaviour and its social significance in the African dwarf mongoose, Helogale undulata rufula. Zeitschrift f¨ ur Tierpsychologie, 32, 293–318. Rich, T. J. & Hurst, J. L. 1998. Scent marks as reliable signals of the competitive ability of mates. Animal Behaviour, 56, 727–735. 1999. The competing countermarks hypothesis: reliable assessment of competitive ability by potential mates. Animal Behaviour, 58, 1027–1037. Roberts, S. C. 1998. Behavioural responses to scent marks of increasing age in klipspringer Oreotragus oreotragus. Ethology, 104, 585–592. 2000. Female territoriality and the function of scent-marking in a monogamous antelope (Oreotragus oreotragus). Behavioral Ecology and Sociobiology, 47, 417–423. Singer, A. G., Beauchamp, G. K. & Yamazaki, K. 1997. Volatile signals of the major histocompatibility complex in male mouse urine. Proceedings of the Naional Academy of Sciences, USA, 94, 2210–2214. Sliwa, A. & Richardson, P. R. K. 1998. Responses of aardwolves, Proteles cristatus, Sparrman 1783, to translocated scent marks. Animal Behaviour, 86, 137–146. Smith, T. E., Tomlinson, A. J., Mlotkiewicz, J. A. & Abbott, D. H. 2001. Female marmoset monkeys (Callithrix jaccus) can be identified from the chemical composition of their scent marks. Chemical Senses, 26, 449–458. Steel, E. 1984. Effect of the odour of vaginal secretion on non-copulatory behaviour of male hamsters (Mesocricetus auratus). Animal Behaviour, 32, 597–608. Stowe, M. K., Turlings, T. C. J., Loughrin, J. H., Lewis, W. J. & Tumlinson, J. H. 1995. The chemistry of eavesdropping, alarm and deceit. Proceedings of the National Academy of Sciences, USA, 92, 23–28. Vandenbergh, J. G. 1973. Effects of gonadal hormones on the flank gland of the golden hamster. Hormone Research, 4, 28–33.

P1: GFZ/... P2: JZZ/... 0521823617c16.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:57

Eavesdropping and scent over-marking Wilcox, R. M. & Johnston, R. E. 1995. Scent counter-marks: specialized mechanisms of perception and response to individual odors in golden hamsters, Mesocricetus auratus. Journal of Comparative Psychology, 109, 349–356. Wimmer, H. & Perner, J. 1983. Beliefs about beliefs: representation and constraining function of wrong beliefs in young childrens’s understanding of deception. Cognition, 13, 103–128. Woodward, R. L. J., Bartos, K. & Ferkin, M. H. 2000. Meadow voles (Microtus pennsylvanicus) and prairie voles (M. ochrogaster) differ in their responses to over-marks from opposite- and same-sex conspecifics. Ethology, 106, 979–992.

371

P1: JZZ/... P2: JZZ/... 0521823617c17.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:14

17

Vocal communication networks in large terrestrial mammals k a r e n m c c o m b & dav i d r e b y University of Sussex, Falmer, Brighton, UK

Introduction Many mammals give long-range calls that can be received over wide areas, often containing large numbers of receivers. In the case of mammals with fluid social systems, opportunities for exposure to the calls of others are further enhanced by the movement of individuals with respect to one another. In our chapter, we discuss the relevance of eavesdropping and communication networks in a range of mammal species, first considering how these concepts apply in cases where loud calls are used to exchange social information in static territorial and fluid fission–fusion societies, and then exploring their potential importance where mammals use loud sexual calls to broadcast information about resource-holding potential. We also outline the mechanisms by which information in mammalian calls is encoded, broadcast and acquired, and we consider the possible fitness consequences that attending to calling interactions can confer. Finally, we evaluate how the vocal communication networks described for non-human mammals differ from human communication networks and discuss possible explanations for these differences. When mammals give loud calls, the area over which the signal can be received is potentially extensive. Such calls are typically emitted at high soundpressure levels (greater than 100 dB at 1 m) and while spherical spreading and excess attenuation from the environment eventually result in the signal being engulfed in background noise, it often remains intelligible over distances of several kilometres from the source: for example the calls of lions Panthera leo (Ogutu & Dublin, 1998; Funston, 1999; K. McComb, unpublished data), hyaenas Crocuta crocuta (Ogutu & Dublin, 1998) and elephants Loxodonta africana (McComb Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

372

P1: JZZ/... P2: JZZ/... 0521823617c17.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:14

Vocal networks in large terrestrial mammals et al., 2003). Many mammals occur at relatively high densities; therefore, the active space may contain a large number of potential receivers. In addition, some mammals live in fluid social systems where there are unusual opportunities for exposure to vocal signals from others in the population. In such social systems, where the identity of immediate neighbours constantly changes because of the movement of individuals and groups in relation to one another, an individual may broadcast to and receive signals from a larger section of the overall population than is normally possible in territorial systems. These two important characteristics of mammal signalling systems have the potential to generate a much wider audience for the signaller. On the basis of the characteristics outlined above, it seems appropriate to conceptualize the production and perception of mammal loud calls in the context of an array of several receivers (after McGregor & Dabelsteen, 1996). Indeed we will argue, based on the examples presented in our review, that this is the best way to view communication involving loud calls. There is direct evidence that mammals attend to vocal signals that are not explicitly directed at them. For example, elephant contact calls, although directed at family and bond group members, are attended to by others in the population, who exhibit knowledge of these calls and adjust their social behaviour on the basis of them (McComb et al., 2000). Diana monkeys Cercopithecus diana attend to the alarm calls of another primate (Campbell’s monkey Cercopithecus campbelli) and appear to obtain functionally relevant information from the detailed combination of different vocalizations used (Zuberbuhler, 2002, see also below). However, in considering the extent to which this form of audience effect (Ch. 4) in loud-calling mammals involves ‘eavesdropping’ or constitutes a ‘communication network’, three important issues need to be considered. First, eavesdropping in the context of animal signalling has been defined as ‘extracting information from an interaction between other individuals’(McGregor & Dabelsteen, 1996; see also Ch. 2). This is a technical definition of a term that in colloquial usage implies more specifically that receivers ‘secretly listen to a conversation’(Concise Oxford English Dictionary). While behavioural acts constituting secrecy or deception are notoriously difficult to identify in mammals (Semple & McComb, 1996), it seems important to distinguish between cases of eavesdropping in which transmission of information to receivers other than the main recipient would be selected for and those where it would not. This can be achieved by conducting cost–benefit analyses of particular signalling interactions (e.g. Ch. 3). Second, an important aspect of the McGregor & Dabelsteen (1996) definition of eavesdropping is that it involves extracting information from ‘an interaction’ rather than simply attending to the call itself. At this stage, evidence that mammals attend to the signalling interaction (rather than just the signal) is very sparse. An

373

P1: JZZ/... P2: JZZ/... 0521823617c17.xml CU1917B/McGregor

374

0 521 582361 7

April 11, 2005

18:14

K. McComb & D. Reby isolated example is provided by the studies of Cheney and colleagues (Cheney et al., 1995; Ch. 25) on baboons Papio cynocephalus ursinus, which demonstrated that receivers distinguish between appropriate and anomalous vocal exchanges between dominants and subordinates in their group. However, lack of evidence for attention to the vocal interaction itself may be more apparent than real. Few researchers other than Cheney and colleagues have conducted the appropriate experiments specifically to examine this phenomenon. Finally, while it is certainly the case that a system of mammal communication linking signallers to several receivers has some properties of a network, it lacks others. A network can be viewed simply as ‘a system of interconnected people or things’. However, advanced networks such as human social networks or the Internet are generally understood to involve the passage of information from one remote part of the network to another, via intermediate recipients that pass on information to other individuals. Non-human vocal communication (including bird and mammal communication systems) appears to fall short of this and we will consider possible explanations for this in the course of the review. Loud calls and social behaviour Availability of information in a simple territorial system

The typical nature of mammal loud calls that are used to mediate social behaviour suggests that selection for concealing information from unwanted receivers has not been paramount. Their high sound-pressure level, abrupt onset and broadband (often noisy) nature are properties that would be expected to make them easy to detect and locate by listeners (e.g. Brown et al., 1979, 1980). Given that such calls often function to advertise territory ownership, it would usually be advantageous for them to attract the attention of any conspecifics in the vicinity. There is evidence from a range of mammals that individuals can distinguish between familiar and unfamiliar callers, even where receivers are separated from signallers by large distances relative to the size of the animal: for example pikas Ochotona princeps (Conner, 1985), cotton-top tamarins Sanguinus oedipus (Snowdon et al., 1983) mangabeys Cercocebus albigena (Waser, 1977), rhesus macaques Macaca mulatta (Rendall et al., 1996), wolves Canis lupus (Tooze et al., 1990), and lions (McComb et al., 1993; Grinnell & McComb, 2001). Where calls characteristics are adapted for long-distance transmission and are easy to locate, the only mechanism for withholding information from unwanted receivers would be to suppress calling altogether. The potential for loud calls to attract unwanted attention may well be considerable in social mammals that use loud calls not only in territorial defence but also to maintain contact with widely spaced social companions: for example wolves (Harrington & Mech, 1979), lions

P1: JZZ/... P2: JZZ/... 0521823617c17.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:14

Vocal networks in large terrestrial mammals (McComb et al., 1994; Grinnell & McComb, 1996), and chimpanzees Pan troglodytes (Mitani & Nishida, 1993). Here particular classes of individual that cannot afford to risk escalated encounters with competitors that might hear their calling could benefit by remaining silent even though, by doing so, they would forfeit the benefits of coordinating their movements with members of their own social group. The behaviour of free-ranging nomadic male lions in the Serengeti National Park is consistent with these predictions (Grinnell & McComb, 2001). Prides of African lions consist of matrilineal kin groups of females, their dependent offspring and a coalition of resident males that enter the pride from outside (Packer et al., 1988). In the pride, both sexes use loud calls (roaring) to advertise ownership of a territory and to stay in contact with other members of their social group (Schaller, 1972; McComb et al., 1994; Grinnell et al., 1995; Grinnell & McComb, 1996). At any one time, however, a high proportion of male lions in the population are not in possession of a pride. These ‘nomads’wander widely, passing through pride ranges singly or in coalitions until they are successful in taking over a pride of their own (Hanby & Bygott, 1987; Pusey & Packer, 1987). It is crucial for nomadic males to gain and maintain social bonds with their companions while they wander because success in competition for prides is strongly dependent on group size (Bygott et al., 1979; Grinnell et al., 1995). Roaring provides a means by which nomadic males might coordinate their movements with coalition partners or recruit new ones (see also McComb et al., 1994; Grinnell et al., 1995). However, if nomads used this loud, long-distance signal to communicate with social companions, they would also advertise their position to resident males in the area. Nomadic males are likely to pay high costs if they attract the attention of resident males in the area. Resident males have been consistently shown to approach aggressively playbacks of roaring from strange males that are broadcast in their territories (Grinnell et al., 1995; Grinnell & McComb, 2001) and intercoalition encounters can be fatal (Schaller, 1972; Grinnell et al., 1995). Given these costs, nomads might benefit by reducing their rate of roaring or even abandoning roaring altogether and concealing their presence – despite the detrimental effects that this would be likely to have on their ability to maintain contact with coalition partners and attract potential mates. Grinnell & McComb (2001) found that in the Serengeti population only male lions that were resident in a pride ever roared. Nomadic males were never observed roaring when they were followed at night, even when they became separated from their coalition partners. They also failed to roar when played recordings of unfamiliar males roaring. In contrast, resident males maintained a high rate of roaring in both these circumstances. There are two possible explanations for why nomadic male lions fail to roar: first, non-resident males gain no benefits from roaring and so never do so; second, non-resident males could benefit from roaring, particularly by enhancing their

375

P1: JZZ/... P2: JZZ/... 0521823617c17.xml CU1917B/McGregor

376

0 521 582361 7

April 11, 2005

18:14

K. McComb & D. Reby ability to recruit and maintain contact with coalition partners (McComb et al., 1994; Grinnell et al., 1995), but the costs of engaging in this behaviour outweigh the benefits. If the first explanation is true, then nomadic males should never roar under any circumstances, whereas the second explanation predicts that nomadic males will only roar when the probability of incurring costs, specifically attracting the attention of resident males in the area, is low. Observational studies at other field sites suggest that the second explanation is correct (Grinnell & McComb, 2001). Funston (1999), working on nomadic male lions in Kruger National Park, South Africa, found that nomadic males do sometimes roar, but at greatly reduced rates in comparison with resident males. Of the three nomadic coalitions that he followed, one was explicitly noted to spend most of their time in an area without resident males and thus where the social costs to roaring would be reduced. In addition, observations by Grinnell in Pilanesberg National Park, South Africa suggested that here, too, nomadic males roar when local resident males are unlikely to hear them. Pilanesberg is an ancient volcanic caldera that contains valleys which are acoustically isolated from each other by mountainous ridges. A non-resident male coalition was observed roaring in one of these valleys that was not occupied by resident males or females (Grinnell & McComb, 2001). It is also important to note that, while nomadic males in the Serengeti study did not roar, males that had been nomadic were seen to begin roaring as soon as they launched a challenge for ownership of a pride (Grinnell & McComb, 2001). This emphasizes that roaring is a flexible behaviour that signallers may have been selected to adjust according to the potential costs and benefits of revealing information on location to listeners. There are reports from other species of low signalling rates, or suppression of signals altogether, in situations in which conspicuous signals could attract the attention of potential aggressors. Chimpanzees have been observed to remain unusually quiet during excursions into the territories of other communities (Goodall, 1986) and, when they hunt monkeys (Colobus and Cercopithecus spp.), are reported to fall silent on hearing the prey’s calls (Boesch & Boesch-Achermann, 2000). Lone wolves howl less than do territorial pairs and packs (Harrington & Mech, 1979). Similarly, transient coyotes Canis latrans howl at greatly reduced rates compared with residents when passing through others’ territories (Gese & Ruff, 1998). It is interesting that resident male lions may also adopt an apparently stealthy strategy when ranging outside their territory. Grinnell & McComb (2001) noted that resident males that had ventured well beyond their territory boundaries never roared even when missing male companions. Long-distance signalling may well be controlled in similar ways in other social species where eavesdroppers can impose high costs on signalling (see also Ch. 4). Recent work on transient killer whales (Orcinus orca) suggests that these animals adjust their calling behaviour to

P1: JZZ/... P2: JZZ/... 0521823617c17.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:14

Vocal networks in large terrestrial mammals minimize the costs of being detected by their acoustically sensitive mammalian prey (Deecke, 2003). Availability of information in mammals with fluid social systems

Above we have presented evidence that information on caller characteristics such as identity is potentially available over quite long distances in species that use loud calls for social communication within territorial systems. Nonetheless, degradation of calls with distance from the source will eventually result in such information being engulfed in background noise and lost to receivers. In certain mammal social systems, however, receivers are not limited to learning only the calls of individuals in their own group or of particular territorial neighbours within their hearing range. Some mammal societies are highly fluid, with individuals and social units moving freely with respect to each other and ranging widely. In these fluid societies, individuals pass through the signalling ranges of a much larger number of conspecifics and are provided with opportunities to learn to recognize the vocalizations of many more signallers than just their immediate neighbours (see also discussions in Chs. 20 and 25). If mental capacities for storing information on the identity of conspecifics’ signals are adequate, these circumstances would provide individuals with opportunities to become familiar with the signals of many different conspecifics that form part of a widespread population. Thus in mammals with fluid social systems, the unusually high encounter rates that individuals have with conspecifics should interact with long-distance signalling abilities to increase greatly the opportunities that receivers have for learning to recognize the vocalizations of other individuals in the population. A number of large mammals, including some primates (e.g. chimpanzees: Boesch & Boesch-Achermann, 2000), cetaceans (e.g. sperm whales Physeter macrocephalus: Whitehead et al., 1991) and African elephants use long-distance signals for social communication and have fluid social systems. In African elephants, the closest social relationships exist between members of a family unit, typically composed of adult females that are matrilineal relatives and their immature offspring, and between bond groups of families that associate frequently and often greet one another when they meet (Moss & Poole, 1983). However, individual family units move freely with respect to one another and range widely, frequently coalescing with other family units in the population as they move and feed, thus forming highly fluid fission–fusion societies (Moss & Poole, 1983). In a population of elephants in Amboseli National Park, Kenya, with known life histories and ranging patterns, the extent to which female subjects were capable of recognizing others in the population through long-distance contact calls was evaluated from playback experiments (McComb et al., 2000). These experiments demonstrated that female African elephants not

377

P1: JZZ/... P2: JZZ/... 0521823617c17.xml CU1917B/McGregor

378

0 521 582361 7

April 11, 2005

18:14

K. McComb & D. Reby only give a characteristic reaction to the contact calls of family or bond group members but also can discriminate between the calls of less-frequent associates, distinguishing the calls of individuals in this category with whom they have higher association indices from those with whom they have lower association indices. Based on the association indices involved, McComb et al. (2000) estimated that subjects would have to be familiar with the contact calls of a mean of 14 different families (including about 100 adult females) in order to perform this discrimination. Empirical studies of the extent of networks of vocal recognition in cetaceans and primates, which are currently lacking, may reveal similar patterns. Networks of vocal recognition are likely to be particularly extensive where individuals are long lived and social knowledge can be accumulated over considerable time periods.

Loud calls and sexual behaviour Many large mammals have loud calls that function to attract individuals of the opposite sex and advertise resource-holding potential to competitors in the vicinity (e.g. Clutton-Brock & Albon, 1979; Tyack, 1981 ; McElligott et al., 1999). These calls are often very conspicuous and seem specifically adapted for attracting the attention of a wide audience. The loud reproductive calls of polygynous deer, which typically serve several functions, provide some of the best examples. Loud mating calls in deer

Male red deer Cervus elaphus roar at high rates during the autumn breeding season or rut and these loud vocalizations are known not only to affect the outcome of contests between males (Clutton-Brock & Albon, 1979; Reby & McComb, 2003b) but also to influence mate attraction (McComb, 1991) and advance ovulation in females (McComb, 1987). There are consequently several receivers to whom male roars might be relevant, including other males, the signaller’s own harem of females and other potential mates within hearing range. Video footage of red deer stags orientating their responses to the roars of several neighbours with distinct spatial locations clearly indicates that they take the complex spatial distribution of callers around them into account (D. Reby & K. McComb, personal observation). Moreover, it has been shown that female red deer are able to discriminate between the roars of their own stag and those of surrounding males (Reby et al., 2001). Finally, there is some evidence that when signallers would benefit by advertising the outcome of their interactions, they use particularly conspicuous vocalizations. Roaring bouts given during roaring competitions with rival stags and after chasing hinds tend to contain a high proportion of ‘harsh roars’, which are unusually loud

P1: JZZ/... P2: JZZ/... 0521823617c17.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:14

Vocal networks in large terrestrial mammals and easy to locate and have an acoustic structure that emphasizes the caller’s body size (Reby & McComb, 2003b). Groaning in fallow deer bucks Dama dama also appears adapted for more than one category of receiver and each of these must be considered when modelling the vocal behaviour of callers (McElligott & Hayden, 1999, 2001; McElligott et al., 1999). McElligott et al. (1999) found that the bucks that achieved most matings were those who had initiated vocal activity early in the season and who had remained vocal on most days. This led the authors to conclude that females may discriminate between males on the basis of long-term cumulative investment in vocal activity. However, although rates of groaning were higher when females were present, males with females exhibited higher groaning rates in the presence of nearby vocal males, suggesting that the signal was also a threat aimed at male rivals (McElligott & Hayden, 1999). Given the several functions of deer vocalizations, it is clear that there are situations in which signals that would be beneficial in one context may be costly in another: for example, when an individual male could gain reproductive advantages by signalling to attract mates but in doing so would invite escalated contests with male competitors. We have observed that young red deer stags (four to five year old) who have gained access to a harem of females while the mature harem holder is temporarily absent, and who have started to roar, will rapidly fall silent when the harem holder returns, often dropping their heads to feed as he approaches (K. McComb & D. Reby, personal observation). Similarly, playback experiments on fallow deer (Komers et al., 1997) have shown that immature males decrease their rate of groaning in response to playbacks of groans from mature males, whereas mature males increase their groaning rates in this situation. Fallow deer bucks may, therefore, adjust groaning rate in relation to several receivers, responding to the complex balance between the benefits of deterring other males and displaying to females and the costs of inviting contests with potentially stronger males in the vicinity. Since red and fallow deer rutting calls are individually distinct (McComb, 1988; Reby, 1998; Reby et al., 1998), females and males may be able to recognize individual callers from their vocalizations and accumulate knowledge on both a signaller’s short-term vocal interactions with others and it’s long-term calling behaviour. Research on red deer has revealed that females can discriminate between the roars of their own stag and those of neighbouring harem holders (Reby et al., 2001). It is possible that red deer hinds could receive information from roaring exchanges and move between harems accordingly. Stags may also attend to contests between other males for information on the body size and motivational state and adjust their decisions to challenge harem holders on this basis. In this context, all receivers, whether they are directly involved in an interaction with the caller or not,

379

P1: JZZ/... P2: JZZ/... 0521823617c17.xml CU1917B/McGregor

380

0 521 582361 7

April 11, 2005

18:14

K. McComb & D. Reby would benefit from attending to cues to resource-holding potential madeavailable in this way. What is now required is empirical work to investigate directly the extent to which receivers attend to interactions in which they are not themselves involved. Loud mating calls in other mammals

Loud and acoustically complex sexual songs produced by humpback whales Megaptera novoeangliae (Payne & McVay, 1971) and fin whales Balaenoptera physalus (Croll et al., 2002) have the potential to travel unprecedented distances underwater (e.g. Croll et al., 2002). While it is known that the individuals who give these songs are male, the intended receivers have not yet been unambiguously identified; they may be rival males, potential mates or both (Tyack, 1983; Mobley et al., 1988; Noad et al., 2000; Croll et al., 2002). What is clear is that such vocalizations are detectable over vast tracts of ocean and may reach a much larger audience than the sexual calls of terrestrial mammals discussed above. Male pinnipeds also have loud sexual advertisement calls (e.g. Northern elephant seals Mirounga angustirostris (Shipley et al., 1981, and common seals Phoca vitulina (van Parijs et al., 2000)) and calling interactions between males on land or underwater may be attended to by rival males, potential mates or both. It remains to be seen whether receivers alter their subsequent behaviour on the basis of which male dominates in a calling interaction (see fuller discussion in Ch. 18).

Mammal anti-predator calls In contrast to long-distance social and sexual calls given by large mammals, alarm calls typically have acoustic features that would be expected to make them difficult to locate. While these calls may be delivered at moderate amplitudes, the information that they contain is likely to be available over shorter distances. Despite this, they are clearly attended to by a range of receivers, including members of other mammal species (Schaller, 1967; Oda, 1997; Zuberbuhler, 2002). In responding to the alarm calls of Campbell’smonkeys, Diana monkeys attend not only to the referent of the alarm call, responding with their own species-specific alarm call for the same predator, but also appear sensitive to the detailed composition of the alarm-calling sequence. In situations where the presence of a predator is less threatening, Campbell’s monkeys emit a pair of ‘boom’ calls before their alarm calls. Playbacks of Campbell’s alarm calls with booms did not elicit alarm calls from Diana monkey subjects (Zuberbuhler, 2002). Some anti-predator calls may have an even wider audience. A study of roe deer Capreolus capreolus revealed that barks, previously identified as ‘alarm calls’, in fact function to elucidate the cause of disturbance (Reby et al., 1999a). In this

P1: JZZ/... P2: JZZ/... 0521823617c17.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:14

Vocal networks in large terrestrial mammals communication system, calls inform any predator that might be present that it has been detected and simultaneously reveal the caller’sidentity and status to any conspecific (whether the latter is the cause of the disturbance or not). The likelihood of barking in response to a predator-like disturbance is independent of the presence of (related or unrelated) conspecifics in the close vicinity, demonstrating that it is not an alarm call (Reby et al., 1999a). However, barking is contagious, with one individual’s barks often being followed by antiphonal calling behaviour from up to seven neighbouring individuals of both sexes (Reby et al., 1999a). Since the acoustic structure of the vocalization carries information on the sex, age and identity of the caller (Reby et al., 1999b), barking may enable roe deer to identify and locate each other, and possibly assess dominance status (particularly during counterbarking sessions involving several animals). Playback experiments supported the hypothesis that although barking may have initially evolved as an anti-predator strategy it is also a signal attended to by conspecific receivers, in particular other males during the territorial period (Reby et al., 1999a). Therefore, when a roe deer barks, irrespective of the stimulus that elicits it (predator or conspecific), it reveals its location, identity and status to a diverse audience of receivers, the composition of which will have marked effects on the costs and benefits associated with calling.

Encoding of information on individuality and size Within a network, the ability of individuals to determine each other’s identity, physical status or internal state from signals dramatically increases the level of functionally relevant information that is potentially exchanged. Whereas in some cases identity may be inferred from the location of the caller or by using visual or olfactory signals, acoustic cues are likely to be of primary importance when individuals range widely. Such cues can provide receivers with instantaneous information on the location and attributes of the caller and may represent the only effective signalling modality in nocturnal or forest-dwelling species. There is a considerable body of evidence indicating that the vocalizations of terrestrial mammals contain information on the identity and physical attributes of the caller (see below). In principle, individual differences can be present at several levels in the acoustic structure of the call. When mammals give voiced calls, the resultant sound is the product of a source signal, generated in the larynx, that is subsequently filtered in the cavities of the vocal tract (Fant, 1960). The source–filter theory of voice production separates the source components, generated by the vibration of the vocal folds, from the filter components, generated when certain frequencies in the source spectrum are selectively amplified or filtered as the signal passes

381

P1: JZZ/... P2: JZZ/... 0521823617c17.xml CU1917B/McGregor

382

0 521 582361 7

April 11, 2005

18:14

K. McComb & D. Reby through the supralaryngeal vocal tract. The characteristics of the source include the duration of the call, its fundamental frequency, the periodicity of the signal, its spectral slope and the presence of phenomena associated with non-linear dynamics, such as subharmonics, biphonation and deterministic chaos (Wilden et al., 1998). Differences in these characteristics of call structure arise from variation in subglottal pressure and in the length and shape of the vocal folds and their stress and tension. All of these parameters can vary between individuals, either as a result of differences in the way the larynx is operated or simply because of random variation in the morphology of callers. In comparison, the key characteristics of the filter are the position and bandwidths of the formant frequencies, which describe the shape of the spectral envelope. Formant frequencies are determined by the length and shape of the cavities of the vocal tract, namely the pharynx, mouth and nasal cavities. Individual differences in formant frequencies can arise from differences in vocal tract morphology or from variation in the way the shape of the vocal tract is actively modified during vocalization (e.g. the extent of mouth opening, lip rounding and vocal tract extension). Variation in source and filter characteristics both appear to be important in encoding individual identity in a range of large mammals. Differences in the fundamental frequency contour have been identified as important in broadcasting information on identity in wolves (Tooze et al., 1990) and elephants (McComb et al., 2003), while individuality in formant frequencies has been demonstrated in fallow deer (Reby et al., 1998), roe deer (Reby et al., 1999b), red deer (McComb, 1988; Reby, 1998), elephants (McComb et al., 2003) and rhesus macaques (Rendall et al., 1998). Filter characteristics, in particular the frequency spacing between successive formants, provide the most reliable cues to body size (Fitch, 1997; Riede & Fitch, 1999; Reby & McComb, 2003a). In contrast, source characteristics, in particular fundamental frequency values, provide relatively poor information on size (Masataka, 1994; Reby & McComb, 2003b) but are better indicators of age and sex and may, therefore, reflect important variation in vocal fold length between sexes and throughout the lifetime (Reby & McComb, 2003b). It is important to appreciate that source and filter characteristics that have the potential to provide receivers with information on caller identity can be distorted or lost as distance from the signaller increases. Even where calls can theoretically be transmitted over long distances because they possess acoustic characteristics that are well adapted for sound transmission in a particular environment, it is unsafe to conclude that receivers can extract socially relevant information from degraded calls at these distances. In female African elephants, playback experiments and re-recordings indicate that abilities for social recognition through long-distance contact calls become limited when frequency components around 115 Hz become immersed in background noise (McComb et al., 2003). This typically

P1: JZZ/... P2: JZZ/... 0521823617c17.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:14

Vocal networks in large terrestrial mammals occurs at distances of 1–2 km from the caller, which is considerably shorter than the propagation distances that have been proposed for calls with infrasonic fundamental frequencies (McComb et al., 2003). This finding highlights the importance of considering the distances over which vocal signals within communication networks can propagate without losing their intelligibility to receivers, which are not necessarily equivalent to the distances over which such signals are physically detectable (see also Ch. 20).

Acquiring and storing information on vocal characteristics Little is known about the factors that influence how effectively individuals acquire and store information about their social companions, although it is known that social knowledge, particularly that used in vocal recognition between mothers and offspring, can be retained for several years even when individuals do not encounter each other (Insley, 2000; McComb et al., 2000). In African elephants, where adult females are familiar with the contact calls of a large proportion of the population around them (see above), the key factor that affects social discrimination abilities is the age of the oldest female in the group (McComb et al., 2001). Playback experiments revealed that families with older matriarchs were significantly better at discriminating the contact calls of genuine strangers from those of more familiar associates than were families with younger matriarchs (McComb et al., 2001). While families with older matriarchs were several thousand times more likely to bunch into defensive formation when played the calls of families they have encountered only rarely than when played the calls of families they frequently associate with, families with younger matriarchs were only marginally more likely to bunch (McComb et al., 2001). Log-linear analysis revealed that variables such as the number of other females present in the group, and their respective ages, did not affect vocal discrimination abilities. An additional factor that did appear to be important was the rate at which subjects encountered other families in the population. An elephant family unit directly encounters, on average, 25 other families over the course of the year in Amboseli National Park, and passes within 1–2 km of 35, providing family members with plenty of opportunity to become familiar with the calls of others. Recent analyses suggest that having a high encounter rate with others in the population can enhance a family’sability to identify the calls of genuine strangers, and that this may be particularly beneficial for families with younger matriarchs (K. McComb, unpublished data). The above results suggest that the age of one crucial individual, the oldest female or matriarch, can affect the social knowledge of the group as a whole. Age and experience are likely to affect abilities to acquire and store information on vocal signals in other societies where animals are long lived and remain part of

383

P1: JZZ/... P2: JZZ/... 0521823617c17.xml CU1917B/McGregor

384

0 521 582361 7

April 11, 2005

18:14

K. McComb & D. Reby a social network for many years. The social systems of some whales have strong parallels with those of elephants (e.g. Pennisi, 2001) and Ford et al. (1994) noted that in killer whales the death of the oldest female, from whom many of the individuals are usually descended, may destabilize a pod. The effects of age and other factors on abilities to recognize the vocalizations of conspecifics has not been investigated for species that defend individual territories (rather than sharing a range with matrilineal relatives as in elephants and some whales) and studies of this sort are now required. Moreover, we as yet know nothing of the extent to which large mammals develop knowledge of the mating calls of others in the population and the factors that affect the acquisition of this knowledge.

The fitness consequences of attending to the calls of others There is some evidence that attending to the exchange of social calls between other individuals can confer fitness benefits on receivers. In African elephants, where the matriarch appears to act as a repository for information on the calls of others in the population (see above), families with older matriarchs have greater reproductive success, at least some of which appears to derive from superior social knowledge (McComb et al., 2001). The fitness consequences of attending to vocal interactions involving sexual calls have never been quantified for mammals but are likely to be highly significant. Acquiring information on resource-holding potential by monitoring the outcome of vocal contests may allow receivers to assess rivals and potential mates much more accurately, and to benefit from better decisions made as a consequence.

Summary and back to definitions In light of the examples discussed above, the term communication network can be usefully employed to identify sets of links between individuals (not necessarily contiguous) that are known to each other through vocal signals or that acquire information about each other’s interactions through vocalizations. However, animal communication networks appear to be limited in a number of important respects (see the Introduction to this chapter). In their typical form, they describe overlapping lattices each consisting of three individuals: the signaller, the intended receiver and an extraneous listener. In such systems, extraneous listeners do not normally interact – in particular they do not pass on information that they gain from attending to interactions. This is in stark contrast with human communication networks, where information can be transferred from one remote part of the network to another and where intermediate recipients may not be the ultimate receivers. Several key constraints on vocal communication

P1: JZZ/... P2: JZZ/... 0521823617c17.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:14

Vocal networks in large terrestrial mammals may prevent non-human mammals from sharing information in this more complex way. These include the limited size of mammalian vocal repertoires, the rarity of fully referential calls and the limited productivity possible in the absence of duality of patterning – a unique feature of human language whereby phonemes can be combined into words and words into sentences (Hockett, 1960; Pinker, 1994). These characteristics are likely to have been selected for in the course of the massive expansion in sociality and social fluidity that occurred during human evolution, creating an environment where the ability to use symbolism and syntax to communicate about displaced activities would be of great importance. It is important to appreciate that once these abilities had evolved, the fitness benefits of attending to the calls of others would increase by orders of magnitude.

Acknowledgements The research described was funded by grants from BBSRC, NERC, the Royal Society (all to K.M.) and INRA (to D.R.). David Reby was supported by Fyssen and E. U. Marie Curie Fellowships and the University of Sussex. We also thank Vincent Janik, Peter McGregor and an anonymous referee for helpful comments on the first draft of the chapter.

References Boesch, C. & Boesch-Achermann, H., 2000. The Chimpanzees of the Tai Forest. Oxford: Oxford University Press. Brown, C. H., Beecher, M. D., Moody, D. B. & Stebbins, W. C. 1979. Locatability of vocal signals in Old World monkeys: design features for the communication of position. Journal of Comparative Physiological Psychology, 5, 806–819. 1980. Localization of noise bands by Old World monkeys. Journal of the Acoustical Society of America, 68, 127–132. Bygott, J. D., Bertram, B. C. R. & Hanby, J. P. 1979. Male lions in large coalitions gain reproductive advantages. Nature, 282, 839–841. Cheney, D. L., Seyfarth, R. M. & Silk, J. B. 1995. The responses of female baboons (Papio cynocephalus ursinus) to anomalous social interactions: evidence for causal reasoning? Journal of Comparative Psychology, 109, 134–141. Clutton-Brock, T. H. & Albon, S. D. 1979. The roaring of red deer and the evolution of honest advertisement. Behaviour, 69, 145–169. Conner, D. A. 1985. Analysis of the vocal repertoire of adult pikas: ecological and evolutionary perspectives. Animal Behaviour, 33, 124–134. Croll, D. A., Clark C. W., Acevedo, A. et al. 2002. Only male fin whales sing loud songs. Nature, 417, 809. Deecke, V. B. 2003. The vocal behaviour of transient killer whales (Orcinus orca): communicating with costly calls. Ph.D. Thesis, University of St Andrews.

385

P1: JZZ/... P2: JZZ/... 0521823617c17.xml CU1917B/McGregor

386

0 521 582361 7

April 11, 2005

18:14

K. McComb & D. Reby Fant, G. 1960. Acoustic Theory of Speech Production. The Hague: Mouton. Fitch, W. T. 1997. Vocal tract length and formant frequency dispersion correlate with body size in rhesus macaques. Journal of the Acoustical Society of America, 102, 1213–1222. Ford, J. K. B., Ellis, G. M. & Balcomb, K. C. 1994. Killer Whales. Vancouver: UBC Press; Seattle, WA: University of Washington Press. Funston, P. J. 1999. Predator–prey relationships between lions and large ungulates in the Kruger National Park. Ph.D. Thesis, University of Pretoria. Gese, E. M. & Ruff, R. L. 1998. Howling by coyotes (Canis latrans): variation among social classes, seasons, and pack sizes. Canadian Journal of Zoology, 76, 1037–1043. Goodall, J. 1986. The Chimpanzees of Gombe. Cambridge, MA: Belknap. Grinnell, J. & McComb, K. 1996. Maternal grouping as a defense against potentially infanticidal males: evidence from field playback experiments on African lions. Behavioral Ecology, 7, 55–59. 2001. Roaring and social communication in African lions: the limitations imposed by listeners. Animal Behaviour, 62, 93–98. Grinnell, J., Packer, C. & Pusey, A. E. 1995. Co-operation in male lions: kinship, reciprocity or mutualism? Animal Behaviour, 49, 95–105. Hanby, J. P. & Bygott, J. D. 1987. Why do subadult lions leave their natal pride? Animal Behaviour, 35, 161–169. Harrington, F. H. & Mech, L. D. 1979. Wolf howling and its role in territory maintenance. Behaviour, 68, 207–249. Hockett, C. F. 1960. The origin of speech. Scientific American, 203, 88–96. Insley, S. J. 2000. Long-term vocal recognition in the northern fur seal. Nature, 406, 404–405. Komers, P. E., P´elabon, C. & Stenstr¨ om, D. 1997. Age at first reproduction in male fallow deer: age-specific versus dominance-specific behaviors. Behavioral Ecology, 8, 456–462. Masataka, N. 1994. Lack of correlation between body size and frequency of vocalisations in young female Japanese macaques (Macaca fuscata). Folia Primatologica, 63, 115–118. McComb, K. 1987. Roaring by red deer stags advances the date of oestrus in hinds. Nature, 330, 648–649. 1988. Roaring and reproduction in red deer, Cervus elaphus. D. Phil. Thesis, University of Cambridge. 1991. Female choice for high roaring rates in red deer, Cervus elaphus. Animal Behaviour, 41, 79–88. McComb, K., Pusey, A., Packer, C. & Grinnell, J. 1993. Female lions can identify potentially infanticidal males from their roars. Proceedings of the Royal Society of London, Series B, 252, 59–64. McComb, K., Packer, C. & Pusey, A. E. 1994. Roaring and numerical assessment in contests between groups of female lions, Panthera leo. Animal Behaviour, 47, 379–387.

P1: JZZ/... P2: JZZ/... 0521823617c17.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:14

Vocal networks in large terrestrial mammals McComb, K., Moss, C., Sayialel, S. & Baker, L. 2000. Unusually extensive networks of vocal recognition in African elephants. Animal Behaviour, 59, 1103–1109. McComb, K., Moss, C., Durant, S. M., Baker, L. & Sayialel, S. 2001. Matriarchs act as repositories of social knowledge in African elephants. Science, 292, 491–494. McComb, K., Reby, D., Baker, L., Moss, C. & Sayialel, S. 2003. Long-distance communication of social identity in African elephants. Animal Behaviour, 65, 317–329. McElligott, A. G. & Hayden, T. J. 1999. Context-related vocalization rates of fallow bucks, Dama dama. Animal Behaviour, 58, 1095–1104. 2001. Postcopulatory vocalisations of fallow bucks: who is listening? Behavioral Ecology, 12, 41–46. McElligott, A. G., O’Neill, K. P. & Hayden, T. J. 1999. Cumulative long-term investment in vocalization and mating success of fallow bucks, Dama dama. Animal Behaviour, 57, 1159–1167. McGregor, P. K. & Dabelsteen, T. 1996. Communication networks. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma and E. H. Miller. Ithaca, NY: Cornell University Press, pp. 409–425. Mitani, J. C. & Nishida, T. 1993. Contexts and social correlates of long-distance calling by male chimpanzees. Animal Behaviour, 45, 735–746. Mobley, J. R., Herman L. M. & Frankel A. S. 1988. Responses of wintering humpback whales (Megaptera novaeangliae) to playback of recordings of winter and summer vocalizations and of synthetic sound. Behavioral Ecology and Sociobiology, 23, 211–223. Moss, C. J. & Poole, J. H. 1983. Relationships and social structure of African elephants. In: Primate Social Relationships: An Integrated Approach, ed. R. A. Hinde. Oxford: Blackwell Scientific, pp. 315–325. Noad, M. J., Cato, D. H., Bryden, M. M. et al., 2000. Cultural revolution in whale songs. Nature, 408, 537. Oda, R. 1997. The response of Verreaux’s sifakas to anti-predator alarm calls given by sympatric ring tailed lemurs. Folia Primatologica, 69, 357–360. Ogutu, J. O. & Dublin, H. T. 1998. The response of lions and spotted hyenas to sound playbacks as a technique for estimating population size. African Journal of Ecology, 36, 83–95. Packer, C., Herbst, L., Pusey, A. E. et al. 1988. Reproductive success of lions. In Reproductive Success, ed. T. H. Clutter-Brock. Chicago, IL: University of Chicago Press, pp. 363–383. Payne, R. S. & McVay, S. 1971. Songs of humpback whales. Science, 173, 585–597. Pennisi, E. 2001. Elephant matriarchs tell friend from foe. Science, 292, 417–418. Pinker, S. 1994. The Language Instinct. New York: William Morrow. Pusey, A. E. & Packer, C. 1987. The evolution of sex-biased dispersal in lions. Behaviour, 101, 275–310. Reby, D. 1998. Etude de la communication sonore chez les cervid´es. Prise en compte de la dimension individuelle. Ph.D. Thesis, Universit´e Paul Sabatier, Toulouse, France.

387

P1: JZZ/... P2: JZZ/... 0521823617c17.xml CU1917B/McGregor

388

0 521 582361 7

April 11, 2005

18:14

K. McComb & D. Reby Reby, D. & McComb, K. 2003a. Anatomical constraints generate honesty: acoustic cues to age and weight in the roars of red deer stags. Animal Behaviour, 65, 519–530. 2003b. Vocal communication and reproduction in deer. Advances in the Study of Behaviour, 33, 231–264. Reby, D., Joachim, J., Lauga, J., Lek, S. & Aulagnier, S. 1998. Individuality in the groans of fallow deer (Dama dama) bucks. Journal of Zoology (London), 245, 79–84. Reby, D., Cargnelutti, B. & Hewison, A. J. M. 1999a. Contexts and possible functions of barking in roe deer. Animal Behaviour, 57, 1121–1128. Reby, D., Cargnelutti, B., Joachim, J. & Aulagnier, S. 1999b. Spectral acoustic structure of barking in roe deer: sex-, age-, and individual-related variations. Comptes Rendus de l’Acad´emie des Sciences, 322, 271–279. Reby, D., Izquierdo, M., Hewison, A. J. M. & Pepin, D. 2001. Red deer (Cervus elaphus) hinds discriminate between the roars of their current harem holder stag and those of neighbouring stags. Ethology, 107, 951–959. Rendall, D., Peter S. R. & Emond R. E. 1996. Vocal recognition of individual and kin in free ranging rhesus monkeys. Animal Behaviour, 51, 1007–1015. Rendall, D., Owren, M. J. & Rodman, P. S. 1998. The role of vocal tract filtering in identity cueing in rhesus monkey (Macaca mulatta) vocalizations. Journal of the Acoustical Society of America, 103, 602–614. Riede, T. & Fitch, W. T. 1999. Vocal tract length and acoustics of vocalization in the domestic dog, Canis familiaris. Journal of Experimental Biology, 202, 2859–2869. Schaller, G. B. 1967. The Deer and the Tiger: A Study of Wildlife in India. Chicago, IL: University of Chicago Press. 1972. The Serengeti Lion. Chicago, IL: University of Chicago Press. Semple, S & McComb, K. 1996. Behavioural deception. Trends in Ecology and Evolution, 11, 434–437. Shipley, C., Hines, M. & Buchwald J. S. 1981. Individual differences in threat calls of northern elephant seal bulls. Animal Behaviour, 29, 12–18. Snowdon, C. T. Cleveland, J. & French, J. A. 1983. Responses to context- and individual-specific cues in cotton top tamarin long-calls. Animal Behaviour, 31, 92–101. Tooze, Z. J., Harrington, F. H. & Fentress, J. C. 1990. Individually distinct vocalisations in timber wolves, Canis lupus. Animal Behaviour, 40, 723–730. Tyack, P. 1981. Interactions between singing Hawaiian humpback whales and conspecifics nearby. Behavioral Ecology and Sociobiology, 8, 105–116. 1983. Differential response of humpback whales, Megaptera novaeangliae, to playback of song or social sounds. Behavioral Ecology and Sociobiology, 13, 49–55. van Parijs, S. M., Hastie, G. D. & Thompson, P. M. 2000. Individual and geographic variation in display behaviour of male harbour seals in Scotland. Animal Behaviour, 59, 559–568. Waser, P. M. 1977. Individual recognition, intragroup cohesion and intergroup spacing: evidence from sound playback to forest monkeys. Behaviour, 60, 28–74.

P1: JZZ/... P2: JZZ/... 0521823617c17.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:14

Vocal networks in large terrestrial mammals Whitehead, H., Waters, S. & Lyrholm, T. 1991. Social organisation of female sperm whales and their offspring: constant companions and casual acquaintances. Behavioral Ecology and Sociobiology, 29, 385–389. Wilden, I., Herzel, H., Peters, G. & Tembrock, G. 1998. Subharmonics, biphonation and deterministic chaos in mammal vocalisation. Bioacoustics, 9, 171–196. Zuberbuhler, K. 2002. A syntactic rule in forest monkey communication. Animal Behaviour, 63, 293–299.

389

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:33

18

Underwater acoustic communication networks in marine mammals vincent m. janik University of St Andrews, St Andrews, UK

Introduction Communication in networks has received considerable research attention over the last few years (Naguib & Todt, 1997; Otter et al., 1999; Peake et al., 2001; Mennill et al., 2002; Ch. 1). This is true for two types of network interaction, one in which several receivers react to the signal of just one individual and a more complex one in which receivers eavesdrop on the signal exchange of two or more individuals and use the information they gained in their own decision making (McGregor & Dabelsteen, 1996; Ch. 2). If we think about communication, the fact that there often is more than just one individual receiving any given signal is not surprising. One reason that many studies on more transient signals concentrated on only one signaller and one receiver was a methodological problem. It is notoriously difficult to sample behaviour from more than one or two individuals at a time, especially if interactions are rapid and involve movements of individuals. Recently, the simultaneous tracking of several individuals in a large group has become feasible. This has led to an increase in studies investigating the effects of signals on several receivers in rapid communication interactions. Many such studies have concentrated on the acoustic domain, a modality that is inherently transient. Signals rarely last more than a few seconds and usually provide a variety of different messages within a single signal. While there is ample evidence from terrestrial environments that more than one individual can receive and use information from a call or a calling interaction, data on acoustic communication networks in marine environments are sparse. The marine environment imposes constraints on, and presents opportunities for, communication in networks that are different from those found in the terrestrial Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

390

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:33

Underwater acoustic communication in marine mammals world (Janik, 1999). Light only penetrates a few tens of metres into the ocean from the surface. This limits the use of colouration or movement signals in the ocean. Signals that use active light sources are more successful but also cost more to develop. Sound is a much cheaper option and travels much further than in air. However, a variety of parameters, especially pressure and temperature gradients, influence sound propagation at different depths. The loudest and best-known biological sound sources in the ocean are marine mammals. Pinnipeds and cetaceans use underwater sound in very similar ways. Many features of their social lives make them interesting subjects for the study of communication networks. For example, sound transmission characteristics of the sea allow individuals to stay in acoustic contact over very long distances (Tyack, 1998). Several species live in large aggregations or fission–fusion societies in which group composition changes frequently while individuals maintain preferences for certain associates (Wells et al., 1999). Many marine mammals are capable of vocal learning, which helps them to produce a variety of different sounds (Janik & Slater, 1997). Finally, many of their cognitive abilities rival those of the great apes (Herman, 1987; Kako, 1999; Schusterman & Kastak, 2002). This combination of environmental conditions and social skills is unique among mammals. Each of them affects how animals communicate and will have profound effects on the characteristics of communication networks. In this chapter, I summarize what we know about acoustic communication networks in marine mammals.

Size and characteristics of marine mammal communication networks at sea Payne & Webb (1971) suspected that cetacean communication networks are among the largest in the world. To identify the potential size of a communication network, we need information on the density of receivers and the active space of a signal. The active space is the area in which another individual can perceive the calls of a conspecific (Brenowitz, 1982). Active space can either be measured directly by playing back calls of known source level and observing a predicted reaction of the receiver, or through theoretical calculations using call source levels, perception thresholds of individuals and models of sound propagation. Direct measurement and theoretical calculation have advantages and disadvantages. For example, an advantage of direct measurement using playback is that the sound reached the animal through its actual environment and no assumptions about propagation loss are necessary. Such assumptions can be a weak point of theoretical calculations since propagation is influenced by several parameters, many of which can change from one minute to the next. A disadvantage of direct measurement using playback is that the active space is likely to be underestimated because

391

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

392

0 521 582361 7

April 11, 2005

18:33

V. M. Janik receivers may extract information from a call but not show an immediate reaction. An advantage of theoretical calculations of active space is that empirical data on sound perception thresholds in different noise conditions can be used to predict when an animal can perceive a sound. However, an ability to perceive a sound of a particular frequency does not necessarily mean that the signal is recognizable as a call of a conspecific. Marine mammal sound propagation in the sea

Sound propagation in the sea differs greatly from that in air. In addition to normal spreading loss, underwater a sound of 1 kHz loses around 0.04 dB/km through absorption while the same sound in air loses 4 dB/km (Richardson et al., 1995). The result is that marine mammal calls have a much larger active space than those of most terrestrial animals. Given that animals can usually detect signals at, or greater than, the level of background noise (Ch. 20), as can we using microphones, it is safe to assume that animals can perceive sounds if we are able to record them. If such recordings are made with passive acoustic localization techniques (e.g. Watkins & Schevill, 1972; Clark & Ellison, 2000; Janik et al., 2000), we are able to determine the distance to the sound source. Studies using such equipment have confirmed that many marine mammals produce signals that can be detected from more than 10 km away (Table 18.1). These distances coincide roughly with those at which reactions to calls have been observed. A fin whale Balaenoptera physalus has been observed to start swimming towards a vocalizing group 20–25 km away (Watkins, 1981). Fin whales also refrain from using certain sound types if there are no other whales within a 20 km radius (Watkins, 1981). Humpback whales Megaptera novaeangliae respond to sounds 9 km away (Tyack & Whitehead, 1983). Such responses are indications that the communication network includes animals at such distances. However, the actual network may be much larger. Recent use of ocean-wide microphone arrays offshore has allowed researchers to record baleen whales over several hundred kilometres (Table 18.1). At large distances, it is unlikely that an animal shows an immediate reaction to a single call. Nevertheless, marine mammals may use distant sounds that indicate the location of other individuals to find breeding or foraging grounds. Detailed theoretical calculations of active space have been conducted for bottlenose dolphins Tursiops truncatus, sperm whales Physeter macrocephalus and killer whales Orcinus orca. Figure 18.1 shows the active space of bottlenose dolphin whistles. Given the fact that most bottlenose dolphin whistles do not have much energy below 3 kHz, whistles should be detected over distances of 20–25 km at maximum source levels. The average source level of 158 dB re 1 µPa measured by Janik (2000a) still gives an active space of 9–16 km in calm seas. However, transmission loss increases with frequency, which means that high-frequency whistles or parts of

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:33

Underwater acoustic communication in marine mammals Table 18.1. Maximum distances from which marine mammal calls can be detected Species

Bearded seal, Erignathus

Frequency range of

Recorded from

vocalizations (kHz)a

distance (km)

0.02–6

Source

25

Cleator et al., 1989

30

Watkins & Schevill,

berbatus Harp seal, Phoca

< 0.1–16+

groenlandica Ringed seal, Phoca

1979 0.4–16

1

Richardson et al., 1995

hispida Delphinids

0.1–27.3

Peale’s dolphin,

0.3–12

16 0.02

Barlow et al., 2001 Schevill & Watkins, 1971

Lagenorhynchus australis Sperm whale, Physeter

0.1–30

37

Barlow & Taylor, 1997

0.025–3.5

17

Clark et al., 1986

15

Helweg et al., 1992

macrocephalus Bowhead whale, Balaena mysticetus Humpback whale,

0.02–8.2

Megaptera

160

Clark, 1995

0.01–0.75

> 20

Watkins, 1981;

0.012–0.39

600

novaeangliae Fin whale, Balaenoptera physalus Blue whale,

Watkins et al., 1987

Balaenoptera musculus a Taken

1600

Stafford et al., 1998 Clark, 1995

from overview in Richardson et al., 1995.

whistles would not have the same active space as low-frequency components. Currently, we know little about how this would affect the information that is available to the receiver. With experience, bottlenose dolphins can identify individually distinctive signature whistles of conspecifics even if they only hear parts of the whistle (Caldwell et al., 1990). However, subtle variations in whistle parameters can carry additional information (Janik et al., 1994), which could be lost in such cases. Using similar methods, Miller (2004) found that killer whale calls have an active space of up to 26 km in calm seas and Madsen et al. (2002) calculated 60 km for slow clicks and 16 km for usual clicks of sperm whales. Again, the maximum source levels were used for these calculations and the same restrictions for high frequencies apply. While these active spaces seem very large, there are some marine mammal signals that are much quieter and do not travel nearly as far (Table 18.1). Most marine mammals produce sounds at various different source levels many of which

393

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

April 11, 2005

18:33

V. M. Janik (a)

40

Range (km)

30

20

10

0 1

2

3

4

5

6

7

8

9

10

11

12

Frequency (kHz)

(b)

40

30 Range (km)

394

0 521 582361 7

20

10

0 1

2

3

4

5

6

8

10

12

Frequency (kHz)

Fig. 18.1. The estimated radius of the active space of dolphin whistles without frequency modulation at different frequencies for sea state 0 (
, calm seas, no wind) and sea state 4 (, moderate breeze of 13–18 miles/h). (a) Whistles produced at maximum source level of 169 dB re 1 µ Pa. (b) Whistles produced at mean source level of 158 dB re 1 µ Pa. Transmission loss in a habitat of homogeneous temperature and 10 m depth (source and receiver at 5 m depth) was calculated following Marsh & Schulkin (1962) and Urick (1983). Ambient noise was taken from Knudsen et al. (1948). Data for auditory thresholds and critical ratios of Tursiops truncatus were taken from Johnson (1967, 1968). (After Janik, 2000a.)

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:33

Underwater acoustic communication in marine mammals would only be audible to conspecifics within a 100 m or less. Furthermore, several species like harbour porpoises Phocoena phocoena (Busnel & Dziedzic, 1966) and Hector’s dolphins Cephalorhynchus hectori (Dawson, 1991), rarely use low-frequency sounds (i.e. <20 kHz) but instead use clicks to communicate. Their signals are subject to much larger transmission loss. Communicative clicks are very similar to echolocation clicks and travel only a few hundred metres, making the active space of these species relatively small. Communication networks in these species are, therefore, much smaller and more comparable to those found in some terrestrial species. The active space calculations for bottlenose dolphins and killer whales used empirical models of sound propagation to predict transmission loss in shallow water. While this is a useful method for this estimate, several other factors influence transmission in different parts of the water column. Shallow water transmission is greatly influenced by reflections off the surface and the bottom. This leads to reverberation that can make acoustic signals unrecognizable. Perception experiments using degraded signals would help us to understand how degradation affects signal detection and recognition in marine mammals. Furthermore, other parameters like transmitter depth or frequency can have a strong effect on active space (Mercado & Frazer, 1999). Consequently, the loudest marine mammal signals are not necessarily the ones that transmit the furthest. In deep water, temperature and pressure profiles give the propagation path a unique shape that is very different from those found in terrestrial environments (Richardson et al., 1995). The speed of sound increases with depth and temperature. In summer, when the surface layer is warmer than the water below, sound is refracted downwards, leading to a shadow zone ahead of the sound source. As the sound travels deeper, temperature does not change much but pressure increases. This leads to refraction towards the surface. As a result, sound travels up and down through the water column as it travels away from the source (Fig. 18.2). If the surface layer is mixed or shows little temperature layering, as is often the case in winter, sound travels more easily through the upper layers. However, some energy still leaks into lower layers and travels in the same ray pattern as in summer. The result of these conditions is that animals at the surface enter and leave convergence zones of the ray paths of a sound produced at great distance. Therefore, to locate a calling animal, an individual needs to consider the special propagation path. If it listened at the surface in one of the convergence zones, it may encounter an area that appears to have a caller in its centre with sound energy decreasing in all directions from it. However, the centre does not have the calling animal in it. Instead, the sound was produced several kilometres away and has travelled through deep waters before returning to the surface. The result is that animals cannot use changes in received levels at the surface to locate a distant caller. However,

395

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

(a)

0 521 582361 7

April 11, 2005

18:33

0

500

1000

1500

Depth (m)

2000

2500

3000

3500

4000

4500

5000 1500

1510

1520

1530 Sound speed (m/s)

1540

1550

1560

(b)

Fig. 18.2. Changes in the speed of sound and ray paths with depth. (a) Typical profile of speed of sound versus depth for temperate or tropical seas. (b) Calculated ray paths for a 20 Hz signal produced at a depth of 50 m in an environment with the speed of sound profile of (a). Ray paths were calculated using a parabolic equation model. White blocks indicate attenuation of ≤ 60 dB; black blocks indicate attenuation of ≥100 dB. Note the convergence zone (shown by arrows) near the surface at ranges of 6.5, 130 and 190 km.

396

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:33

Underwater acoustic communication in marine mammals following parts of the actual sound path underwater may allow an individual to recognize that it is listening in a convergence zone rather than being close to the original caller. Another interesting aspect of sound transmission at sea is the deep sound channel or SOFAR channel. This is a layer at approximately 600–1200 m in which sound is trapped and travels almost horizontally with much less transmission loss (because of the shorter travel path and no losses from surface or bottom reflections). It can be found in the layer with minimum sound speed. Little is known on whether marine mammals use this channel, but it has been suggested that whales may use it for long-distance communication (Payne & Webb, 1971). This is only possible for a few species that travel routinely to this depth, for example elephant seals Mirounga spp. (Le Boeuf et al., 1989; Hindell et al., 1991) or northern bottlenose whales Hyperoodon ampullatus (Hooker & Baird, 1999). However, in Arctic waters, where the minimum sound speed (and thus the SOFAR channel) can be at much shallower depths, it may be within reach of more species. The conditions described here are idealizations assuming little variation in other parameters. They describe general patterns but the actual situation faced by a marine mammal changes with location and time. One conclusion from these patterns is that it must be difficult to estimate range from a caller using parameters such as sound intensity. However, they may be able to use other parameters to determine their distance from a caller. Premus & Spiesberger (1997) analysed fin whale sounds recorded in the Gulf of California. They found that the signal arrived several times at each hydrophone, which is typical if the sound takes several different paths to reach the receiver. Longer paths result in later arrivals, and at great distances these time delays can be substantial. However, Premus & Spiesberger (1997) noted that the first arrival of a fin whale call was much sooner than expected even if it was taking the shortest route available through the water. This fast sound transmission could only be explained if the first arrival represented sound energy that entered and travelled in the sediment, where sound speed is much higher than in water. If such multipath arrivals through different media are common, whales may be able to tell the distance of the caller by listening to the differences in the time of arrival of the sound travelling through the sediment and that travelling through water. This sound path may even allow them to listen to individuals on the other side of an island. Another way in which distance information could be extracted is by listening to the extent of sound degradation. Again, we know little about the abilities of marine mammals to use such features to judge distance to the caller. The number of animals in a communication network

The second variable that determines communication network size is the number of animals within the transmission range of a signal. This varies greatly

397

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

398

0 521 582361 7

April 11, 2005

18:33

V. M. Janik Table 18.2. Examples of animal density and average group sizes for selected sites Species

Bottlenose dolphin Harbour

Location

Gulf of Mexico

Density

Average

(animals/100 km2 )

group size

6–480

2–15

Shane et al., 1986

1.49

Hammond et al.,

(US coast) Northwest

porpoise

Europe

White-beaked

Northwest

dolphin

Europe

Minke whale

Northwest

Vocalizing fin

Hawaii

10–80

2002 0–5

3.78

Hammond et al.,

0–3

1.04

Hammond et al.,

1

McDonald &

2002

Europe whales

Source

2002 0.0027; maximum 0.0081

Fox, 1999

and depends on the area, species and behaviour of a marine mammal. Bottlenose dolphins, for example, can be found in groups of hundreds (Saayman et al., 1973) or even thousands (Scott & Chivers, 1990) offshore, while individuals in coastal areas may at times find themselves acoustically isolated from all conspecifics if they enter small inlets in which sound is blocked by land. Furthermore, many marine mammal species, especially delphinids, live in fission–fusion societies where group composition and size can change rapidly. Finally, if we consider that individuals are capable of restricting signal transmission to specific receivers (see below), it becomes clear that network size is difficult to assess. On an evolutionary scale, however, it is interesting to look at how many potential receivers there are for any given signal. This might help us to understand the relationship between network size and specific strategies to direct or restrict signals. Because of the lack of information on average transmission distances of marine mammal sounds, we can currently only look at data on the number of animals in an area rather than calculate network sizes. Ultimately, to calculate network sizes, population densities and the average active space of a signal from the same area need to be combined. The average group size of a marine mammal species is a good indicator of the most commonly found minimum network size (Table 18.2). If group size is very large, as in some oceanic dolphin species, the transmission range of a signal can be limited by masking noise from conspecifics and the actual network would contain fewer animals than are in the group (for similar considerations in anurans, see Ch. 13). In most cases, however, the network will be larger than the average group size because of the large active space of marine mammal calls. Population-density data can be used to estimate average network sizes for unrestricted signals that

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:33

Underwater acoustic communication in marine mammals travel beyond the group’s boundaries. Table 18.2 gives population densities and average group sizes for a few selected marine mammal species. Some studies have assessed animal density using acoustic surveys. This means that only vocalizing animals are registered. While this gives a good density estimate of signallers, it represents a minimum rather than a representative average estimate of network size. However, population density studies usually look at very large areas that are less relevant for estimates of communication network sizes. One tight group of 20 dolphins in 1000 km2 , for example, will yield a very low population density but still represents a communication network of 20 animals. Like sound propagation in the sea, population density is a highly dynamic variable. Therefore, the size of marine mammal communication networks is likely to vary greatly on a temporal as well as spatial scale. We can expect that territorial and, therefore, relatively stationary species display more stability in network size, but studies on the dynamics of such network sizes are still lacking.

Directing and restricting signals In a network, we can expect to find two different kinds of signal: those that are directed at all receivers within range and others aimed at only one or a few. Callers directing signals at specific individuals benefit from adding information that indicates who they are addressing. This is even more important if, like in marine mammals, the network can be large, locations of individuals are difficult to predict (e.g. if animals are not territorial) and if only one sensory modality is available. Concurrently, such conditions render it more difficult for the sender to identify who is within range as a potential receiver that it is worth calling to. One way of solving this problem for the caller is to give unequivocal information about its own identity or group membership. This makes it more likely to be recognized by other group members or close associates within range. While it might be disadvantageous to broadcast one’s identity or location if predators use such cues to find prey, signalling this information can be evolutionarily stable if it improves information transmission for the sender to the required receiver (Johnstone, 1997). Most animal species cannot avoid providing identity information through individually specific voice cues. Such cues result from individually specific genetic and environmental influences on the morphology of the vocal apparatus during development. Similarly, genetically related individuals may share a voice feature that can be used in kin or even group recognition if related individuals stay together. However, voice cues are relatively subtle and can be difficult to decode over long distances or in high background noise. There are several ways to improve the encoding of information on identity. First, if groups are genetically isolated, genetic

399

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

400

0 521 582361 7

April 11, 2005

18:33

V. M. Janik drift can increase group differences. Even between sympatric groups, this is possible if their members do not interbreed. Another solution is to use call types that are shared by all animals in a population at higher or lower rates than the rest of the population. Finally, animals can develop group- or individual-specific call types. This last solution can usually only be achieved through vocal learning or invention (Janik & Slater, 2000). Nevertheless, several of these influences can act together to create individual differences. For example, limited skills in vocal learning that only allow a slight change in the fundamental frequency of a given signal may be used to enhance individual differences caused by environmental influences during development. Many marine mammal species show pronounced differences between groups of animals. Weddell seals Leptonychotes weddelli in breeding colonies only 20 km apart have been found to use colony-specific call types and show differences in usage of shared call types (Morrice et al., 1994). Similar geographic variation over much larger distances has been described for leopard seals Hydrurga leptonyx (Thomas & Golladay, 1995), bearded seals Erignathus berbatus (Cleator et al., 1989), harp seals Phoca groenlandica (Terhune, 1994) and harbour seals Phoca vitulina (van Parijs et al., 2000a). However, in these cases it is possible that individuals from different sites are geographically isolated. Humpback whales in the Atlantic and the Pacific, for example, sing very different songs (Winn et al., 1981). Since they cannot encounter each other, these differences are not necessary for group recognition. Thus, the occurrence of differences between the calls of groups of animals is not evidence for a specific adaptation for group recognition. Killer whales (Ford & Fisher, 1983), sperm whales (Rendell & Whitehead, 2003) and blue whales Balaenoptera musculus (Stafford et al., 2001) also have distinctive group calls, but here these groups overlap in their geographic ranges. In these cases, the distinctiveness in the repertoire may be more important for directing signals than in geographically isolated groups. However, the calls in these examples are not individually specific. Individual specificity may not be necessary for animals that live in stable family groups like killer whales. If group composition is less stable though, more unequivocal signals may be required for individual recognition. Bottlenose dolphins, for example, associate preferentially with specific individuals, but their daily ranging behaviour results in regular changes of group composition and short-term associations (Wells et al., 1987). This organization is often referred to as a fission–fusion society. Bottlenose dolphins develop individually distinctive signature whistle types (Fig. 18.3) that are used while animals are out of visual contact (Caldwell et al., 1990; Janik & Slater, 1998). These signals have a much larger interindividual variability than isolation calls of other animal species and thus transmit individual identity more reliably (Tyack, 2000). Signature whistles

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:33

Underwater acoustic communication in marine mammals

Fig. 18.3. Three randomly chosen spectrograms (columns; FFT size, 1024; time resolution, 20.5 milliseconds; frequency resolution, 50 Hz; number of FFT steps, 200; weighting function, Hanning window) of signature whistles from each of four different individual bottlenose dolphins (rows). Background noise and harmonics have been removed on all spectrograms to show the pronounced difference in the shape of the fundamental frequency of signature whistles of the different individuals. (After Janik & Slater, 1998).

401

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

402

0 521 582361 7

April 11, 2005

18:33

V. M. Janik have been reported to be stable over more than 12 years in female bottlenose dolphins in the Bay of Sarasota, Florida, USA (Sayigh et al., 1990). Tyack (1997) found that vocal learning influences whistle development. In fact, learning or innovation may be the only way to develop such signals that have to be different from those of a large number of conspecifics within a fission–fusion society. Janik & Slater (1997) suggested that individual or group recognition might have been one of the main selection pressures on the evolution of vocal learning in cetaceans. Evidence for similar individually distinctive signature signals exists for common dolphins Delphinus delphis (Caldwell & Caldwell, 1968), Pacific white-sided dolphins Lagenorhynchus obscurus (Caldwell & Caldwell, 1971), spotted dolphins Stenella plagiodon (Caldwell et al., 1973), Pacific humpback dolphins Sousa chinensis (van Parijs & Corkeron, 2001) and sperm whales (Watkins & Schevill, 1977). While such shared calls may facilitate recognition in general, calls can also be directed at specific individuals through vocal matching. In vocal matching, an individual responds to the signal of a caller by producing a signal of the same type. Vocal matching can be used without the existence of individual-specific calls as long as other individuals can copy calls or if they have a repertoire of shared calls. Many species of cetaceans have been observed to produce calls of the same kind in response to a call of a conspecific, but such anecdotal reports cannot exclude the possibility of matching occurring by chance. If individuals share a repertoire and produce sounds independently, by chance alone two different individuals can produce signals of the same type in close succession. However, this does not necessarily mean they interact vocally. The proportion of such interactions has to be larger than expected by chance to represent evidence for vocal matching. True vocal matching has been demonstrated for bottlenose dolphins (Janik, 2000b) and for killer whales (Miller et al., 2004). In bottlenose dolphins, signature whistles can be copied by another individual in such matching interactions (Janik & Slater, 1998). Tyack (1991) raised the interesting possibility that bottlenose dolphins may use signature whistles of other individuals to initiate contact with the ‘owner’ of the signature whistles. However, in all reported cases in which signature whistle matching has been observed and the identities of the calling individuals were known, the ‘owner’ of the signature whistle called first (Janik & Slater, 1998). A different strategy is changing the directionality of calls. The fundamental frequencies of most marine mammal calls are usually transmitted in a relatively omnidirectional pattern (Evans et al., 1964; Lammers & Au, 2003). However, clicks are highly directional (reviewed by Au, 1993) and have the potential to be used in addressing specific individuals. Dolphins use clicks in echolocation as well as communication. Several species (see above) rely on clicks for communication and do not produce any whistles at all (Dawson, 1991; reviewed by Herman & Tavolga, 1980). It is possible that these species use the directionality of their clicks to direct

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:33

Underwater acoustic communication in marine mammals or restrict signals in social interactions. Even in whistles, the high-frequency parts are highly directional (Lammers & Au, 2003) so that the same signal may carry different types of information with some of it only available to animals ahead of the caller. This could be achieved by filtering high-frequency components in specific patterns, similar to formants in human signals. Such filtering would not be discernible from listening to the low-frequency component alone. Dolphins have control over the filtering of higher-frequency harmonics (Fig. 1 in Janik et al., 1994), but the significance of such changes is unclear. Miller (2002) found that killer whales have call types with high-frequency components that show higher directionality than the low-frequency parts of the same call or calls without these components. High- and low-frequency components of the same killer whale call are not harmonically related. Therefore, the modulation pattern of one component cannot be discerned from the other one. This makes withholding information even easier. Miller (2002) suggested that killer whales might use calls with high-frequency components to indicate their direction of movement. Alternatively, they may be used to direct signals at specific individuals and withhold information from others. Another way of restricting the spread of signals through a communication network is by decreasing the source level so that they do not carry as far. Pinnipeds and cetaceans produce the same call types at a variety of different source levels (reviewed by Richardson et al., 1995). Many species of odontocetes also have signals of very different frequency in their repertoires. High-frequency signals are attenuated much more rapidly than low-frequency sounds. Odontocetes may be able to restrict transmission range by choosing high-frequency clicks rather than lower-frequency whistles even though they are produced with the same source level. However, calling depths have different optimal frequencies for long-range signal transmission (Mercado & Frazer, 1999). Higher frequencies can sometimes travel further than lower ones, especially in relatively shallow water (i.e. less than 100 m deep). Further studies are needed to explore the possible use of source-level adjustments and frequency selection in directing and restricting signals.

Eavesdropping Peake (Ch. 2) has distinguished two types of eavesdropping; interceptive eavesdropping (e.g. predators locating prey by listening to prey vocalizations) and social eavesdropping (extracting information from a signalling interaction). Bradbury & Vehrencamp (1998) also used the term ‘cue’ for prey signals that are used by predators to locate prey. While it is arguable whether such interactions can be called communication, effects of calls on predators and their prey are an interesting ecological variable that can influence the design of communication systems.

403

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

404

0 521 582361 7

April 11, 2005

18:33

V. M. Janik Several studies have looked at the impact of killer whale signals on other marine mammal species that are potential prey. The diet of killer whales can vary considerably from one location to another. In British Columbia, Canada, some killer whales only eat fish while others take marine mammals as prey. Fish eaters are known as resident killer whales since they have smaller ranging patterns than the so-called transient killer whales that feed on marine mammals. Deecke et al. (2002) conducted playback experiments and inferred from diving patterns that harbour seals in British Columbia avoided sounds made by transient killer whales but they did not react to sounds of resident killer whales. Transients and residents use different call types and individual killer whale pods have repertoires of up to 17 call types (Ford, 1989). Deecke et al. (2002) carefully selected specific sound types for each comparison to ensure that the discrimination performed by the seals could not be based on just one or two call types. Harbour seals also avoided playbacks of sounds from Norwegian killer whales. These whales concentrate on herring as prey for at least part of the year, which makes it unlikely that fisheating killer whales share voice features that identify them as harmless to seals. It is unclear how harbour seals distinguish between known residents and other killer whales. There are genetic differences between killer whale populations and even between sympatric residents and transients of British Columbia (Hoelzel et al., 1998). Perhaps residents share a voice feature that affects all their calls and makes them recognizable. Alternatively, the seals may have learned all call types used by resident killer whales and avoid all other call types. Grey whales Eschrichtius robustus (Cummings & Thompson, 1971) and beluga whales Delphinapterus leucas (Fish & Vania, 1971) have been found to avoid locations from which killer whale sounds had been played. Unfortunately, it is not clear whether the sounds used in these studies came from mammal-eating or fisheating killer whales. Belugas (Schevill 1964; Fish & Vania, 1971) and grey whales (Cummings & Thompson, 1971) also ceased vocalizing when exposed to killer whale sounds, another well-known reaction of cetaceans to any unusual stimulus (Herman & Tavolga, 1980). Other examples are pilot whales Globicephala melaena falling silent when hunted (Schevill, 1964) and bottlenose dolphins (Caldwell & Caldwell 1967) and Peale’s dolphins Lagenorhynchus australis (Schevill & Watkins, 1971) falling silent when captured or when approached by a boat. Interestingly, transient killer whales appear to use fewer echolocation clicks than resident whales while they forage (Barrett-Lennard et al., 1996). This may be a counterstrategy to avoid early detection by their prey. Several aspects of marine mammals make it difficult to establish whether social eavesdropping occurs in this group (i.e. whether information has been extracted from a signalling interaction). Individuals often approach callers, for example. Groups of surface-active humpback whales produce a variety of sounds that can

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:33

Underwater acoustic communication in marine mammals attract males that are several kilometres away (Tyack, 1983) and similar results have been found for southern right whales Eubalaena glacialis (Clark & Clark, 1980). However, it is difficult to determine who is interacting (i.e. whether individuals in such groups signal to each other or to individuals outside the group) and, therefore, whether there is potential for eavesdropping. Distant individuals may extract information from the calls of animals interacting in the group and decide to approach (which would qualify as social eavesdropping), or they may be attracted by calls that are directed at distant animals (which is a good example of communication in a network but not for eavesdropping). It will be difficult to distinguish between these possibilities experimentally; furthermore, these two scenarios are not mutually exclusive. One context in which social eavesdropping has been demonstrated is in song interactions between birds (Naguib & Todt, 1997; Otter et al., 1999; Peake et al., 2001; Mennill et al., 2002). Many marine mammal species also produce song during the mating season and some, like Weddell seals (Bartsh et al., 1992) and harbour seals (van Parijs et al., 2000b), establish underwater territories. By analogy with songbirds, social eavesdropping by marine mammals may be found in such circumstances. However, other singing species of marine mammals are less stationary. For example, while singing humpback whales are spaced further apart than non-singers and singers often avoid each other (Frankel et al., 1995), individuals can rarely be found in the same location from one day to the next (Clapham, 2000). Clapham termed this arrangement a floating lek, in which females are able to listen to several males but males are not stationary. Given the apparent lack of direct vocal interactions outside of the surface-active groups that form when several males start to escort a female, eavesdropping is less likely to be of importance here. However, further studies relating vocal displays to movement of individuals are needed before we can assess the relevance of eavesdropping in this context. Another very different context in which the term eavesdropping has been used is echolocation (Xitco & Roitblat, 1996). These authors found that a bottlenose dolphin could extract information about the location and shape of an object without having to produce echolocation sounds itself; it did so by listening to the echoes of echolocation clicks produced by another individual. This might be a common feature of echolocating animals. Bats have been found to be attracted by feeding buzzes of conspecifics (Barclay, 1982; Balcombe & Fenton, 1988). However, the studies on bats could not determine whether feeding buzzes are generally attractive, like food calls of non-echolocating animals, or whether they can provide information about the exact location and shape of the target to the eavesdropper. In the study of Xitco & Roitblat (1996), the eavesdropping animal was very close to the echolocating one and such close proximity may be a prerequisite for gathering such target-specific information. This form of eavesdropping can be defined as

405

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

406

0 521 582361 7

April 11, 2005

18:33

V. M. Janik extracting information from an interaction between another individual’s echolocation signal and an echolocation target. While this might explain some of the swimming formations that dolphins use during foraging, it is of less relevance to the topic of communication networks.

Conservation implications A major concern in marine mammal conservation is the impact of noise made by human activity in the sea. There has been an increasing amount of industrial, shipping and seismic survey noise over the last century. For example, engine noise of ships in the busy shipping lanes of the North Atlantic increases the average ambient noise levels below 500 Hz by 10–40 dB (Urick, 1983). Ross (1976) estimated that shipping led to a 10 dB increase in ambient noise in these areas from 1950 to 1975. The issue of such noise has been discussed recently in the context of the Acoustic Thermometry of Ocean Climate (ATOC) study and the low-frequency active sonar systems deployed by the military (Richardson et al., 1995). These techniques can potentially harm marine mammals because of high source levels and signals that are similar to those of some marine mammal species. The main concern in noise exposure has been potential physical damage to the animals. For example, several Cuvier’s beaked whale Ziphius cavirostris strandings occurred at the same time as military exercises (e.g. Frantzis, 1998; Balcomb & Claridge, 2001) and Jepson et al. (2003) reported acute and chronic tissue damage caused by gas bubbles in whales stranded during such exercises. Weddell seals exposed to underwater blasts showed severe damage to their inner ears (Bohne et al., 1986). Another form of impact is a change in the animal’s behaviour. This can have the same consequences as physical damage since isolation from group members or the exclusion from feeding grounds can easily lead to the death of an animal. There are many studies showing short-term avoidance by marine mammals of sound sources (review in Richardson et al., 1995). Examples are killer whales (Morton & Symonds, 2002) and harbour porpoises (Johnston, 2002) avoiding areas ensonified by acoustic harassment devices deployed to reduce seal predation on fish farms; beluga whales avoiding ice-breakers by as much as 80 km for up to 48 hours (Finley et al., 1990; Foote et al., 2004); and bottlenose dolphins in Florida avoiding specific feeding grounds on weekends when boat activity is highest (Allen & Read, 2000). Another response to noise of human origin is a change in calling behaviour. Such responses can involve a change in temporal or structural parameters of a call (e.g. Au et al., 1985; Foote at al., 2004) or lead to animals changing call rates or ceasing to vocalize (Terhune et al., 1979; Bowles et al., 1994). Such changes could either indicate a direct disruption of communication or be a by-product of a

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:33

Underwater acoustic communication in marine mammals general change in behaviour if animals stop activities that involve specific calling rates or types of signal to avoid a sound source. A source of human-derived noise may also affect the behaviour of an animal by masking marine mammal sounds, thereby disrupting communication. This could be a serious problem for animals in a number of circumstances. First, many marine mammals use acoustic signals to maintain contact between mothers and calves (e.g. Renouf, 1984; Smolker et al., 1993) and noise can shorten the range over which they are able to hear each other. Second, if information is gathered by eavesdropping on interactions of more distant individuals, noise could mask such interactions. Erbe (2002) found that the noise of a fast-moving boat can mask quiet killer whale sounds if the vessel is 14 km from the listening animal. Similar calculations predict that icebreaking noise can mask quiet beluga sounds if the icebreaker is up to 71 km from the animal that is listening (Erbe & Farmer, 2000). While we do not know to what extent information gathered through eavesdropping is used by marine mammals, masking certainly has an effect on signals designed to reach more distant receivers, as in marine mammal song. Therefore, apart from inflicting physical damage, noise could have a severe effect by disrupting acoustic contact between individuals.

Summary and future directions While we have data on maximum transmission distances for some marine mammal sounds, it is still unclear to what extent acoustic signals from distant animals provide valuable information to a conspecific. If the active space of the signal is particularly large, as seems to be the case for many marine mammals, the information from distant animals may not be of much use. For example, it is of only limited value for a predator to know that an animal is foraging 20 km away if a long time is needed to travel that distance. For marine mammals, most aggregations of prey species are very dynamic and either move quickly or only last for brief periods of time. Consequently, a large active space may just increase noise for distant receivers and could have contributed to the evolution of redundancy and distinctiveness in communication signals: two features that can help to improve information transmission and that are pronounced in marine mammal communication systems. One way of addressing the question of the value of distant signals would be to compare reactions to distant marine mammal calls with reactions to artificial broadband noise at similar levels. If conspecifics only add noise to the communication channel, responses should be the same. Only at a closer, more relevant distance should reactions differ. Yet communication over large distances may help in mate attraction or coordination of behaviour patterns. In that case, the

407

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

408

0 521 582361 7

April 11, 2005

18:33

V. M. Janik animals should show specific reactions to distant signals of conspecifics, for example specific changes in movement direction. Any studies investigating how marine mammals react to distant signals would be extremely valuable. Two related issues are the effect of degradation and how marine mammals judge the distance to a caller. The distinctiveness of signature whistles in bottlenose dolphins, for example, suggests that at least the identity information encoded is relatively resistant to degradation. However, what happens to more subtle cues? The auditory system of marine mammals is adapted to detect and identify marine mammal signals. Therefore, it would be difficult to make predictions from experiments with artificial test signals. How receivers estimate their distance from a sound source has been studied extensively in birds and humans (reviewed by Naguib & Wiley, 2001). The most important parameter appears to be the degree of reverberation. However, other parameters such as overall and frequency-dependent attenuation or amplitude fluctuations can also help in the assessment of distances if the receiver has some experience with the source signal and the environment. Marine mammals may also use additional cues like time delays of multiple arrivals via different sound paths (Premus & Spiesberger, 1997) or changes in signal composition of the same received signal at different receiver depths (Mercado & Frazer, 1999). Whether and how such information is used by marine mammals is still unknown. Eavesdropping on interactions of conspecifics in marine mammals is still virtually unstudied. Territorial seal species would probably be the best starting point for such studies as interactions between neighbours and intruders are the most likely source of relevant information that could be obtained through eavesdropping. However, to simulate such interactions experimentally we need to know the acoustic parameters that identify a successful or unsuccessful animal in such contests. Furthermore, we need to investigate whether individuals can recognize other individuals by general voice features. Without voice recognition, it is difficult to explain how an animal would recognize an individual that it previously eavesdropped on. Studying how marine mammals address specific individuals can also help to understand how relevant eavesdropping is. If marine mammals not only address specific individuals by matching or the use of signature signals but also actively exclude potential receivers through the selective use of highly directional signals, eavesdropping might have been a factor in the evolution of such strategies. Theoretical estimates of maximum signal transmission distance and communication network sizes are useful but they need verification in the real world. Most likely such extremes are rarely relevant for communicating in everyday life. Nevertheless, marine mammal communication networks are clearly among the largest that can be found. As we have seen, this opens up interesting opportunities

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:33

Underwater acoustic communication in marine mammals but also imposes further constraints by increasing background noise. Future studies that investigate the dynamics of marine mammal signalling will improve our understanding of how underwater sound transmission helped to shape their communication systems and to what extent marine mammals use the extra information provided by such large active spaces in their communication networks.

Acknowledgements I would like to thank Peter McGregor for valuable comments on earlier drafts of this chapter. The chapter was written with support from a Royal Society University Research Fellowship. Figure 18.1 has been reprinted from Janik (2000a) with permission from Springer Verlag. Figure 18.3 has been reprinted from Janik & Slater (1998) with permission from Elsevier Science.

References Allen, M. C. & Read, A. J. 2000. Habitat selection of foraging bottlenose dolphins in relation to boat density near Clearwater, Florida. Marine Mammal Science, 16, 815–824. Au, W. W. L. 1993. The Sonar of Dolphins. New York: Springer. Au, W. W. L., Carder, D. A., Penner, R. H. & Scronce, B. L. 1985. Demonstration of adaptation in beluga whale echolocation. Journal of the Acoustical Society of America, 77, 726–730. Balcomb, K. C. & Claridge, D. E. 2001. A mass stranding of cetaceans caused by naval sonar in the Bahamas. Bahamas Journal of Science, 8, 2–12. Balcombe, J. P. & Fenton, M. B. 1988. Eavesdropping by bats: the influence of echolocation call design and foraging strategy. Ethology, 79, 158–166. Barclay, R. M. R. 1982. Interindividual use of echolocation calls: eavesdropping by bats. Behavioral Ecology and Sociobiology, 10, 271–275. Barlow, J. & Taylor, B. 1997. Acoustic census of sperm whales in the eastern temperate North Pacific. Journal of the Acoustical Society of America, 102, 3213. Barlow, J., Rankin, S., Oswald, J. N. & Mellinger, D. K. 2001. Long-range detection of delphinid whistles using a towed hydrophone array. In Proceedings of the 14th Biennial Conference on the Biology of Marine Mammals, Vancouver, p. 17. Barrett-Lennard, L. G., Ford, J. K. B. & Heise, K. A. 1996. The mixed blessing of echolocation: differences in sonar use by fish-eating and mammal-eating killer whales. Animal Behaviour, 51, 553–565. Bartsh, S. S., Johnston, S. D. & Siniff, D. B. 1992. Territorial behaviour and breeding frequency of male Weddell seals (Leptonychotes weddelli) in relation to age, size, and concentration of serum testosterone and cortisol. Canadian Journal of Zoology, 70, 680–692. Bohne, B. A., Bozzay, D. G. & Thomas, J. A. 1986. Evaluation of inner ear pathology in Weddell seals. Antarctic Journal, 21, 208.

409

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

410

0 521 582361 7

April 11, 2005

18:33

V. M. Janik Bowles, A. E., Smultea, M., W¨ ursig, B., DeMaster, D. P. & Palka, D. 1994. Relative abundance and behavior of marine mammals exposed to transmissions from the Heard Island feasibility test. Journal of the Acoustical Society of America, 96, 2469–2482. Bradbury, J. W. & Vehrencamp, S. L. 1998. Principles of Animal Communication. Sunderland, MA: Sinauer. Brenowitz, E. A. 1982. The active space of red-winged blackbird song. Behavioral Ecology and Sociobiology, 147, 511–522. Busnel, R.-G. & Dziedzic, A. 1966. Acoustic signals of the pilot whale Globicephala melaena and of the porpoises Delphinus delphis and Phocoena phocoena. In: Whales, Dolphins, and Porpoises, ed. K. S. Norris. Berkeley, CA: University of California Press, pp. 607–646. Caldwell, M. C. & Caldwell, D. K. 1967. Intraspecific transfer of information via the pulsed sound in captive odontocete cetaceans. In: Animal Sonar Systems: Biology and Bionics, ed. R. G. Bullock. Jouy-en-Josas: Laboratoire Physiologie Acoustique, pp. 879–936. 1968. Vocalization of naive captive dolphins in small groups. Science, 159, 1121–1123. 1971. Statistical evidence for individual signature whistles in Pacific whitesided dolphins, Lagenorhynchus obliquidens. Cetology, 3, 1–9. Caldwell, M. C., Caldwell, D. K. & Miller, J. F. 1973. Statistical evidence for individual signature whistles in the spotted dolphin, Stenella plagiodon. Cetology, 16, 1–21. Caldwell, M. C., Caldwell, D. K. & Tyack, P. L. 1990. Review of the signaturewhistle-hypothesis for the Atlantic bottlenose dolphin. In: The Bottlenose Dolphin, ed. S. Leatherwood & R. R. Reeves. San Diego, CA: Academic Press, pp. 199–234. Clapham, P. J. 2000. The humpback whale: seasonal feeding and breeding in a baleen whale. In: Cetacean Societies: Field Studies of Dolphins and Whales, ed. J. Mann, R. C. Connor, P. L. Tyack & H. Whitehead. Chicago, IL: University of Chicago Press, pp. 173–196. Clark, C. W. 1995. Application of US Navy underwater hydrophone arrays for scientific research on whales. Reports of the International Whaling Commission, 45, 210–212. Clark, C. W. & Clark, J. M. 1980. Sound playback experiments with southern right whales (Eubalaena australis). Science, 207, 663–665. Clark, C. W. & Ellison, W. T. 2000. Calibration and comparison of the acoustic location methods used during the spring migration of the bowhead whale, Balaena mysticetus, off Pt Barrow, Alaska, 1984–1993. Journal of the Acoustical Society of America, 107, 3509–3517. Clark, C. W., Ellison, W. T. & Beeman, K. 1986. A preliminary account of the acoustic study conducted during the 1985 spring bowhead whale, Balaena mysticetus, migration off Point Barrow, Alaska. Reports of the International Whaling Commission, 36, 311–316. Cleator, H. J., Stirling, I. & Smith, T. G. 1989. Underwater vocalizations of the bearded seal (Erignathus barbatus). Canadian Journal of Zoology, 67, 1900–1910. Cummings, W. C. & Thompson, P. O. 1971. Gray whales, Eschrichtius robustus, avoid the underwater sounds of killer whales, Orcinus orca. Fishery Bulletin, 69, 525–530.

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:33

Underwater acoustic communication in marine mammals Dawson, S. M. 1991. Clicks and communication: the behavioural and social contexts of Hector’s dolphin vocalizations. Ethology, 88, 265–276. Deecke, V. B., Slater, P. J. B., Ford, J. K. B. 2002. Selective habituation shapes predator recognition in harbor seals. Nature, 420, 171–173. Erbe, C. 2002. Underwater noise of whale-watching boats and potential effects on killer whales (Orcinus orca), based on an acoustic impact model. Marine Mammal Science, 18, 394–418. Erbe, C. & Farmer, D. M. 2000. Zones of impact around icebreakers affecting beluga whales in the Beaufort Sea. Journal of the Acoustical Society of America, 108, 1332–1340. Evans, W. E., Sutherland, W. W. & Beil, R. G. 1964. The directional characteristics of delphinid sounds. In: Marine Bio-acoustics, ed. W. N. Tavolga. Oxford: Pergamon Press, pp. 353–370. Finley, K. J., Miller, G. W., Davis, R. A. & Greene, C. R. 1990. Reactions of belugas, Delphinapterus leucas, and narwhals, Monodon monoceros, to ice-breaking ships in the Canadian high Arctic. Canadian Bulletin for Fisheries and Aquatic Sciences, 224, 97–117. Fish, J. F. & Vania, J. S. 1971. Killer whale, Orcinus orca, sounds repel white whales, Delphinapterus leucas. Fishery Bulletin, 69, 531–535. Foote, A. D., Osborne, R. W. & Hoelzel, A. R. 2004. Whale-call response to masking boat noise. Nature, 428, 910. Ford, J. K. B. 1989. Acoustic behaviour of resident killer whales (Orcinus orca) off Vancouver Island, British Columbia. Canadian Journal of Zoology, 67, 727–745. Ford, J. K. B. & Fisher, H. D. 1983. Group-specific dialects of killer whales (Orcinus orca) in British Columbia. In: Communication and Behavior of Whales, ed. R. Payne. Boulder, CO: Westview Press, pp. 129–161. Frankel, A. S., Clark, C. W., Herman, L. M. & Gabriele, C. M. 1995. Spatial distribution, habitat utilization, and social interactions of humpback whales, Megaptera novaeangliae, off Hawai’i, determined using acoustic and visual techniques. Canadian Journal of Zoology, 73, 1134–1146. Frantzis, A. 1998. Does acoustic testing strand whales? Nature, 392, 29. Hammond, P. S., Berggren, P., Benke, H. et al. 2002. Abundance of harbour porpoise and other cetaceans in the North Sea and adjacent waters. Journal of Applied Ecology, 39, 361–376. Helweg, D. A., Frankel, A. S., Mobley, J. R. & Herman, L. M. 1992. Humpback whale song: our current understanding. In: Marine Mammal Sensory Systems, ed. J. Thomas, R. Kastelein & S. Supin. New York: Plenum Press, pp. 459–483. Herman, L. M. 1987. Receptive competencies of language-trained animals. Advances in the Study of Behavior, 17, 1–60. Herman, L. M. & Tavolga, W. N. 1980. The communication systems of cetaceans. In: Cetacean Behavior: Mechanisms and Functions, ed. L. M. Herman. New York: John Wiley, pp. 149–209. Hindell, M. A., Slip, D. J. & Burton, H. R. 1991. The diving behaviour of adult male and female southern elephant seals, Mirounga leonina (Pinnipedia: Phocidae). Australian Journal of Zoology, 39, 595–619.

411

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

412

0 521 582361 7

April 11, 2005

18:33

V. M. Janik Hoelzel, A. R., Dahlheim, M. & Stern, S. J. 1998. Low genetic variation among killer whales (Orcinus orca) in the Eastern Pacific and genetic differentiation between foraging specialists. Journal of Heredity, 89, 121–128. Hooker, S. K. & Baird, R. W. 1999. Deep-diving behaviour of the northern bottlenose whale, Hyperoodon ampullatus (Cetacea: Ziphiidae). Proceedings of the Royal Society of London, Series B, 266, 671–676. Janik, V. M. 1999. Origins and implications of vocal learning in bottlenose dolphins. In: Mammalian Social Learning: Comparative and Ecological Perspectives, ed. H. O. Box & K. R. Gibson. Cambridge, UK: Cambridge University Press, pp. 308–326. 2000a. Source levels and the estimated active space of bottlenose dolphin (Tursiops truncatus) whistles in the Moray Firth, Scotland. Journal of Comparative Physiology A, 186, 673–680. 2000b. Whistle matching in wild bottlenose dolphins (Tursiops truncatus). Science, 289, 1355–1357. Janik, V. M. & Slater, P. J. B. 1997. Vocal learning in mammals. Advances in the Study of Behavior, 26, 59–99. 1998. Context-specific use suggests that bottlenose dolphin signature whistles are cohesion calls. Animal Behaviour, 56, 829–838. 2000. The different roles of social learning in vocal communication. Animal Behaviour, 60, 1–11. Janik, V. M., Dehnhardt, G. & Todt, D. 1994. Signature whistle variations in a bottlenosed dolphin, Tursiops truncatus. Behavioral Ecology and Sociobiology, 35, 243–248. Janik, V. M., van Parijs, S. M. & Thompson, P. M. 2000. A two-dimensional acoustic localization system for marine mammals. Marine Mammal Science, 16, 437–447. Jepson, P. D. Arbelo, M., Deaville, R. et al. 2003. Gas-bubble lesions in stranded cetaceans. Nature, 425, 575. Johnson, C. S. 1967. Sound detection thresholds in marine mammals. In: Marine Bio-acoustics, Vol. 2, ed. W. N. Tavolga. Oxford: Pergamon Press, pp. 247–255. 1968. Masked tonal thresholds in the bottlenosed porpoise. Journal of the Acoustical Society of America, 44, 965–967. Johnston, D. W. 2002. The effect of acoustic harassment devices on harbour porpoises (Phocoena phocoena) in the Bay of Fundy, Canada. Biological Conservation, 108, 113–118. Johnstone, R. A. 1997. Recognition and the evolution of distinctive signatures: when does it pay to reveal identity? Proceedings of the Royal Society of London, Series B, 264, 1547–1553. Kako, E. 1999. Elements of syntax in the systems of three language-trained animals. Animal Learning and Behavior, 27, 1–14. Knudsen, V. O., Alford, R. S. & Emling, J. W. 1948. Underwater ambient noise. Journal of Marine Research, 7, 410–429. Lammers, M. O. & Au, W. W. L. 2003. Directionality in the whistles of Hawaiian spinner dolphins (Stenella longirostris): a signal feature to cue direction of movement? Marine Mammal Science, 19, 249–264.

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:33

Underwater acoustic communication in marine mammals Le Boeuf, B. J., Naito, Y., Huntley, A. C. & Asaga, T. 1989. Prolonged, continuous, deep diving by northern elephant seals. Canadian Journal of Zoology, 67, 2514–2519. Madsen, P. T., Wahlberg, M. & Møhl, B. 2002. Male sperm whale (Physeter macrocephalus) acoustics in a high-latitude habitat: implications for echolocation and communication. Behavioral Ecology and Sociobiology, 53, 31–41. Marsh, H. W. & Schulkin, M. 1962. Shallow-water transmission. Journal of the Acoustical Society of America, 34, 863–864. McDonald, M. A. & Fox, C. G. 1999. Passive acoustic methods applied to fin whale population density estimation. Journal of the Acoustical Society of America, 105, 2643–2651. McGregor, P. K. & Dabelsteen, T. 1996. Communication networks. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca, NY: Comstock, pp. 409–425. Mennill, D. J., Ratcliffe, L. M. & Boag, P. T. 2002. Female eavesdropping on male song contests in songbirds. Science, 296, 873. Mercado III, E. & Frazer, L. N. 1999. Environmental constraints on sound transmission by humpback whales. Journal of the Acoustical Society of America, 106, 3004–3016. Miller, P. J. O. 2002. Mixed-directionality of killer whale stereotyped calls: a direction of movement cue? Behavioral Ecology and Sociobiology, 52, 262–270. Miller, P. J. O. 2004. Estimated active space of killer whale sounds. Journal of Comparative Physiology A, in press. Miller, P. J. O., Shapiro, A. D., Tyack, P. L. & Solow, A. R. 2004. Call-type matching in vocal exchanges of free-ranging resident killer whales (Orcinus orca). Animal Behaviour, 67, 1099–1107. Morrice, M. G., Burton, H. R. & Green, K. 1994. Microgeographic variation and songs in the underwater vocalization repertoire of the Weddell seal (Leptonychotes weddelli) from the Vestfold Hills, Antarctica. Polar Biology, 14, 441–446. Morton, A. B. & Symonds, H. K. 2002. Displacement of Orcinus orca (L.) by high amplitude sound in British Columbia, Canada. ICES Journal of Marine Science, 59, 71–80. Naguib, M. & Todt, D. 1997. Effects of dyadic interactions on other conspecific receivers in nightingales. Animal Behaviour, 54, 1535–1543. Naguib, M. & Wiley, R. H. 2001. Estimating the distance to a source of sound: mechanisms and adaptations for long-range communication. Animal Behaviour, 62, 825–837. Otter, K. A., McGregor, P. K., Terry, A. M. R. et al. 1999. Do female great tits (Parus major) assess males by eavesdropping? A field study using interactive song playback. Proceedings of the Royal Society of London, Series B, 266, 1305–1309. Payne, R. & Webb, D. 1971. Orientation by means of long range acoustic signaling in baleen whales. Annals of the New York Academy of Sciences, 188, 110–140. Peake, T. M., Terry, A. M. R., McGregor, P. K. & Dabelsteen, T. 2001. Male great tits eavesdrop on simulated male-to-male vocal interactions. Proceedings of the Royal Society of London, Series B, 268, 1183–1187. Premus, V. & Spiesberger, J. L. 1997. Can acoustic multipath explain finback (B. physalus) 20-Hz doublets in shallow water? Journal of the Acoustical Society of America, 101, 1127–1138.

413

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

414

0 521 582361 7

April 11, 2005

18:33

V. M. Janik Rendell, L. E. & Whitehead, H. 2003. Acoustic clans in sperm whales (Physeter macrocephalus). Proceedings of the Royal Society of London, Series B, 270, 225–231. Renouf, D. 1984. The vocalization of the harbour seal pup (Phoca vitulina) and its role in the maintenance of contact with the mother. Journal of Zoology, 202, 583–590. Richardson, W. J., Greene, C. R., Jr, Malme, C. I. & Thomson, D. H. 1995. Marine Mammals and Noise. San Diego, CA: Academic Press. Ross, D. 1976. The mechanics of underwater noise. New York: Pergamon Press. Saayman, G. S., Taylor, C. K. & Bower, D. 1973. Diurnal activity cycles in captive and free-ranging Indian Ocean bottlenose dolphins (Tursiops truncatus Ehrenburg). Behaviour, 44, 212–233. Sayigh, L. S., Tyack, P. L., Wells, R. S. & Scott, M. D. 1990. Signature whistles of free-ranging bottlenose dolphins, Tursiops truncatus: mother–offspring comparisons. Behavioral Ecology and Sociobiology, 26, 247–260. Schevill, W. E. 1964. Underwater sounds of cetaceans. In: Marine Bio-Acoustics, Vol. 1, ed. W. N. Tavolga. Oxford: Pergamon Press, pp. 307–316. Schevill, W. E. & Watkins, W. A. 1971. Pulsed sounds of the porpoise Lagenorhynchus australis. Breviora, 366, 1–10. Schusterman, R. J. & Kastak, D. 2002. Problem solving and memory. In: Marine Mammal Biology: an Evolutionary Approach, ed. A. R. Hoelzel. Oxford: Blackwell, pp. 371–387. Scott, M. D. & Chivers, S. J. 1990. Distribution and herd structure of bottlenose dolphins in the eastern tropical Pacific Ocean. In: The Bottlenose Dolphin, ed. S. Leatherwood & R. R. Reeves. San Diego, CA: Academic Press, pp. 387–402. Shane, S. H., Wells, R. S. & W¨ ursig, B. 1986. Ecology, behavior and social organization of the bottlenose dolphin: a review. Marine Mammal Science, 2, 34–63. Smolker, R. A., Mann, J. & Smuts, B. B. 1993. Use of signature whistles during separations and reunions by wild bottlenose dolphin mothers and infants. Behavioral Ecology and Sociobiology, 33, 393–402. Stafford, K. M., Fox, C. G. & Clark, D. S. 1998. Long-range acoustic detection and localization of blue whale calls in the northeast Pacific Ocean. Journal of the Acoustical Society of America, 104, 3616–3635. Stafford, K. M., Nieukirk, S. L. & Fox, C. G. 2001. Geographic and seasonal variation of blue whale calls in the North Pacific. Journal for Cetacean Research and Management, 3, 65–76. Terhune, J. M. 1994. Geographical variation of harp seal underwater vocalizations. Canadian Journal of Zoology, 72, 892–897. Terhune, J. M., Stewart, R. E. A. & Ronald, K. 1979. Influence of vessel noises on underwater vocal activity of harp seals. Canadian Journal of Zoology, 57, 1337–1338. Thomas, J. A. & Golladay, C. L. 1995. Geographic variation in leopard seal (Hydrurga leptonyx) underwater vocalizations. In: Sensory Systems of Aquatic Mammals, ed. R. A. Kastelein, J. A. Thomas & P. E. Nachtigall. Woerden: De Spil, pp. 201–221. Tyack, P. 1983. Differential response of humpback whales, Megaptera novaeangliae, to playback of song or social sounds. Behavioral Ecology and Sociobiology, 13, 49–55. 1991. If you need me, whistle. Natural History, 8, 60–61.

P1: IYP/... P2: JZZ/... 0521823617c18.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:33

Underwater acoustic communication in marine mammals 1998. Acoustic communication under the sea. In: Animal Acoustic Communication: Sound Analysis and Research Methods, ed. S. L. Hopp, M. J. Owren & C. S. Evans. Berlin: Springer, pp. 163–220. 1997. Development and social functions of signature whistles in bottlenose dolphins Tursiops truncatus. Bioacoustics, 8, 21–46. 2000. Dolphins whistle a signature tune. Science, 289, 1310–1311. Tyack, P. & Whitehead, H. 1983. Male competition in large groups of wintering humpback whales. Behaviour, 83, 132–154. Urick, R. J. 1983. Principles of Underwater Sound, 3rd edn. New York: McGraw-Hill. van Parijs, S. M. & Corkeron, P. J. 2001. Evidence for signature whistle production by a Pacific humpback dolphin, Sousa chinensis. Marine Mammal Science, 17, 944–949. van Parijs, S. M., Hastie, G. D. & Thompson, P. M. 2000a. Individual and geographic variation in display behaviour of male harbour seals in Scotland. Animal Behaviour, 59, 559–568. van Parijs, S. M., Janik, V. M. & Thompson, P. M. 2000b. Display area size, tenure and site fidelity in the aquatic mating male harbour seal. Canadian Journal of Zoology, 78, 2209–2217. Watkins, W. A. 1981. Activities and underwater sounds of fin whales. Scientific Reports of the Whales Research Institute, 33, 83–117. Watkins, W. A. & Schevill, W. E. 1972. Sound source location by arrival times on a non-rigid three-dimensional hydrophone array. Deep-Sea Research, 19, 691–706. 1977. Sperm whale codas. Journal of the Acoustical Society of America, 62, 1485–1490. 1979. Distinctive characteristics of underwater calls of the harp seal, Phoca groenlandica, during the breeding season. Journal of the Acoustical Society of America, 66, 983–988. Watkins, W. A., Tyack, P., Moore, K. E. & Bird, J. E. 1987. The 20-Hz signals of finback whales (Balaenoptera physalus). Journal of the Acoustical Society of America, 82, 1901–1912. Wells, R. S., Scott, M. D. & Irvine, A. B. 1987. The social structure of free-ranging bottlenose dolphins. In: Current Mammalogy, ed. H. H. Genoways. New York: Plenum Press, pp. 247–305. Wells, R. S., Boness, D. J. & Rathbun, G. B. 1999. Behavior. In: Biology of Marine Mammals, ed. J. E. Reynolds III & S. A. Rommel. Washington, DC: Smithsonian Institute Press, pp. 324–422. Winn, H. E., Thompson, T. J., Cummings, W. C. et al. 1981. Song of the humpback whale: population comparisons. Behavioral Ecology and Sociobiology, 8, 41–46. Xitco, M. J. & Roitblat, H. L. 1996. Object recognition through eavesdropping: passive echolocation in bottlenose dolphins. Animal Learning and Behavior, 24, 355–365.

415

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:41

19

Looking for, looking at: social control, honest signals and intimate experience in human evolution and history john l. locke City University of New York, USA

Introduction Recently, Hauser et al. (2002) argued that if we are to understand human language, several disciplines must work cooperatively. Predictably, these include linguistics and certain areas within psychology and anthropology as well as some relative newcomers: biology and animal behaviour. However, if collaboration can facilitate the investigation of language, long held to be a uniquely human faculty, it is surely indispensable to the study of human communication, for which a number of homologous or analogous processes exist in other species. In the case of language, a behaviour with countless social benefits, researchers have tended to focus on dyadic interactions. In the typical model, the ‘sender’ is a rational human being who has information. As a social being, the sender wishes to share it. The ‘receiver’, equally rational and social, wants to hear it; so the receiver listens and makes an appropriate response. ‘Communication occurs,’ according to one authoritative source, ‘when one organism (the transmitter) encodes information into a signal which passes to another organism (the receiver) which decodes the signal and is capable of responding appropriately’ (Ellis & Beattie, 1986, p. 3). Dyadic interactions such as these occur, of course, and deserve linguists’ theoretical attention. However, in a gregarious species such as ours – and this is a major point of divergence between social communication and linguistic interaction – dyads are often embedded in aggregations of individuals, in various arrangements (communication networks in the sense of this book), and these will usually include one or more perceptual bystanders. If thought to be unobserved, dyads tend to behave in an unguarded way, making them unusually interesting, and Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

416

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:41

Humans: control, signalling and intimate experience their behaviour unusually informative, to uninvited viewers and listeners. Of course, the perceptual target may be alone, acting without recourse to the displays or material objects that are normally used to project his more public self. These solitary behaviours will be less veiled than the dyadic interactions, making them especially useful to anyone who stands to benefit from prediction-grade social knowledge. What I am suggesting is that behaviour which is neither donated by the actor nor observed with his awareness is likely to be unusually high in reliability. It is also likely to be intimate: that is, sufficiently personal that the actor might like to shield it from prying eyes and ears. Reliability and intimacy give prospective observers two rather compelling reasons to sample such behaviour, but it will usually be impossible for them to do so overtly. This gives rise to eavesdropping, a form of information gathering that in humans occurs only by stealth. In animals, research has addressed two broad areas of observation. One relates to the information that is obtained when animals look for other animals. When non-human primates do this type of looking, their focus is typically on the location and activities of outsiders, including predators and competitors from other groups. Typically, this is referred to as vigilance. Other types of information are obtained when animals look at the constituents of their own groups. This type of looking, social observation, is addressed later in the chapter and elsewhere in this volume (e.g. Ch. 25). Predictably, these discriminable functions are associated with different benefits. Looking for Animal vigilance

When animals look for other animals, their tendency is to scan territorial boundaries in order to detect encroachment of predators or competitors. Early detection alerts individuals and, through their reactions, other group members to the need for evasive or defensive action. The perceptual act is performed from the naturally exposed position of group-living animals. Vigilance appears to be a form of perceptual alertness that occurs in anticipation of important events, rather than a form of observation per se, and may even be discontinued when those events occur. The observing itself is performed as sporadic interruption of other activities, rather than a circumscribed commitment of looking time (an exception, ‘sentinels,’ will be discussed below). Much of the research on vigilance involves non-human primates. They, like other animals, need to look out for predators. Vigilance thus produces valuable information, but it comes at a price. Red colobus Procolobus badius tephrosceles and redtail monkeys Cercopithecus ascanius schmidtii typically spend over 50% of their

417

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

418

0 521 582361 7

April 11, 2005

18:41

J. L. Locke time just visually scanning (Treves, 1998). Similar figures have been obtained for chacma baboons Papio cynocephalus ursinus (Cowlishaw, 1998). This does not count all the time these primates spend in a vigilant state, a figure that may approach the totality of free-ranging animals’ waking time. In several different species of monkey, animals that were physically isolated spent more time looking than monkeys that were near a group member (Steenbeek et al., 1999; Treves et al., 2001). This difference may reflect a greater fear of predation on the part of solitaires. Where attacks are unlikely, however, animals might be expected to devote more of their attention to members of their group. Research on brown capuchin monkeys Cebus apella in Iguazu Falls National Park in Argentina, where the annual rate of predation is extremely low, suggests that this may be so (Hirsch, 2002). Predictably, there are also variations in time spent looking among the members of a single group. Subordinate animals tend to look more than dominant ones, largely because they spend a great deal of time watching the dominant animals themselves. These rank differences were suggested some years ago by Chance (Chance, 1967; Chance & Jolly, 1970) and have since been confirmed in a number of species, including long-tailed macaques Macaca fascicularis (Pitcairn, 1976), talapoins Miopithecus talapoin (Dixson et al., 1975; Keverne et al., 1978) and brown capuchins (Hirsch, 2002). In each group, subordinates more often look at dominant animals than the reverse arrangement. In many studies, there have also been sex effects. A male vigilance bias has been witnessed in various primate groups in at least seven different studies (Cheney & Seyfarth, 1981; Fragaszy, 1990; Baldellou & Henzi, 1992; Rose, 1994; Rose & Fedigan, 1995; Gould et al., 1997; also see reviews by Quenette, 1990; Steenbeek et al., 1999). A great deal of male vigilance appeared to be directed outside the group, presumably to predators or sexual competitors, but perceptual targets are notoriously difficult to identify in free-ranging animals. There is one exception to the usual hierarchical pattern. High-ranking males frequently assume unusual responsibilities for vigilance (Rose & Fedigan, 1995). In cooperative groups such as vervets and marmosets, as well as baboons, they – or some other large male – may even take on the role of sentinel and adopt a superior vantage point. These individuals then become the focus of attention for group members, who monitor the sentinel instead of looking for predators themselves (Hall, 1960; Horrocks & Hunte, 1986; Koenig, 1994). The behaviour of sentinels has been described in detail. The tendency is for the sentinel to ascend a tree or rock, mainly so foraging animals can achieve visual contact, making auditory warnings unnecessary (Horrocks & Hunte, 1986). In green monkeys Cercopithecus aethiops sabaeus, this works so well that when a sentinel detects approaching humans, his quietly visible movements may enable the troop

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:41

Humans: control, signalling and intimate experience to disperse without detection (Poirier, 1972). Therefore, in this species at least, the sentinel seemed to ‘tip off ’ the foragers with subtle visual behaviours instead of warning them with loud barks. In the case of baboons, the vigilance itself involves repeated 180◦ head turns, which take about five seconds (Hall, 1960), and in this group detection produces barks. In every report I have read, the only ‘predators’ of possible concern to the sentinel were human and the ongoing feeding behaviours best characterized as ‘raids’ on a plantation (Maples et al., 1976). Consequently the use of sentinels may be predator and context specific. Much of the time, the targets of vigilant males are other males, including interlopers from outside conspecific groups (Rose & Fedigan, 1995). Therefore, the motivation for male vigilance may be more closely linked to a self-oriented control function than a contribution to group welfare. This issue will assume a broader significance below when we see some related sex differences in our own species. Levels of vigilance vary, affected by a variety of internal factors such as mating periods, births and infant excursions away from their mothers. In vervet monkeys Cercopithecus aethiops, males are more vigilant than females, especially during the breeding season (Baldellou & Henzi, 1992). In a study of black howler monkeys Alouatta pigra, female vigilance rates increased after the birth of infants (Treves et al., 2001). In squirrel monkeys Saimiri boliviensis, tape recordings of infant vocalizations increased the time that adult females spent looking for predators fivefold (Biben et al., 1989). In several species, it has been shown that visual obstruction alters the usual benefits of herding and flocking. The first to study this effect was Underwood (1982). He noted that African antelopes frequently interrupted their foraging to look around, but when grazing in tall grass they spent even more time lifting their heads to look at distant areas. Metcalfe (1984a,b) observed a similar pattern in two different species of shorebirds in western Scotland. He found that in both species the time devoted to vigilance rose with increases in the density of obstructions such as rocks, boulders and banks of seaweed. He also noticed that obstructions broke up the usual relationship between flock size and vigilance. Metcalfe reasoned that obstructed animals were in a vulnerable position, unable to see if predators or potentially protective neighbours were nearby. Recapitulating, the primary functions of primate vigilance appear to be defensive when looking is externally directed and the threat of predation is high, and social when looking is internally directed and the threat of predation is low. In the latter, vigilance enables animals to evaluate dominance relations – a critical function in primate societies – and resource-holding potential. In the species studied, the primary sensory modality has been visual. In non-human primates, males and subordinates generally devote more time to vigilance than females and dominants. The target of review is a physical area, such as the perimeter of an

419

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

420

0 521 582361 7

April 11, 2005

18:41

J. L. Locke occupied territory, or other animals that are either within or outside the group. The observer may make no attempt either to conceal or to expose his position. Human vigilance

Humans have not been reluctant to engage in the ancient and deeply ingrained behaviour of vigilance. Some of the benefits derived from these activities resemble those enjoyed by our evolutionary ancestors, broadly understandable as social knowledge and social control. As we will see, our species critically relies on information that can only be obtained through these one-way processes, and yet there is little record of empiricism. To be sure, there have been psychological studies of vigilance – usually defined as the detection of prespecified perceptual targets that occur infrequently, irregularly and weakly – in relation to a range of military and industrial issues. Currently, there is concern with baggage scanning in relation to airport security screening. It is difficult, however, to find reports of research conducted within an ethological framework. Few investigators have asked how humans exercise vigilance with respect to strangers or potentially aggressive intruders. Yet, in societies wishing to guard against crime and terrorism, citizens are concerned with precisely this issue. Until about 20 000 years ago, our ancestors spent much of their time following herds of large animals from place to place. But when the herds dissipated, nomads began to hunt smaller game, to fish and to gather. This shift enabled the new sedentists to spend more time in their resting places before seasonal changes precipitated the next round of migration. At this stage in history, one assumes that human and non-human primates behaved rather similarly with respect to vigilance. Since human groups were several times larger, there were more individuals that had to be monitored, but this was obviously manageable as our premodern ancestors lived almost as openly as the other primates. The nature of their encampments is implied both by archaeological evidence and the behaviour of an existing group whose way of life is thought to replicate ancient patterns of living (Lee, 1979). This group is the !Kung, a population of largely egalitarian hunter–gatherers who inhabit the Kalahari Desert of Botswana and southwest Africa. Although their way of life is changing, in the mid- to latetwentieth century, when they were studied fairly intensively, most of the !Kung lived in bands of 50 or 60. These bands periodically dispersed into still smaller groups or concentrated into larger ones, as suited their needs. The typical camp was laid out in concentric circles. In the centre was a public gathering place. Rimming this plaza were the bandsmen’s grass huts, which were used mainly for storage. These were packed very closely together, enabling bandsmen to perceive and react to the earliest and subtlest acts of an antisocial nature.

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:41

Humans: control, signalling and intimate experience ‘If a person is angry,’ wrote Draper (1978, p. 47), ‘someone, if not everyone, will soon know about it.’ Significantly, the !Kung rarely if ever entered their huts to escape scrutiny. The reason is that it was considered improper for anyone to withdraw from the sociality of camp life, either physically or psychologically. ‘To seek solitude,’ according to Lee (1979, p. 32), ‘is regarded as bizarre behaviour.’But this attitude toward privacy was not unique to the !Kung. There are several other openly living groups. 1.

2.

3.

4.

5.

6.

The Baktaman of New Guinea. ‘There are no recognised and respected ways in which the public gaze can be cut off, no way of separating oneself out from others present’ (Barth, 1975, p. 24). The Mehinacu of Central Brazil. ‘Wherever a person goes in the village he can be seen or heard. When he speaks there is a chance that a third person is listening, and that in a short time everyone else will know what he said. Even the most intimate details of his sex life often become a matter of public knowledge’ (Gregor, 1970, p. 238). The Nayaka of southern India. ‘They remain sited by their respective fire-places, and talk across space from fire to fire . . . they rarely try to conceal their domestic activities’ (Bird-David, 1994, pp. 590–591). The Samoans. They ‘live most of their lives in a very public arena. The more private aspects of experience are strongly discouraged by the absence of walls in a Samoan house, and by powerful norms of social life, which keep people in almost constant social interaction’ (Shore, 1982, p. 148). The Sakalava people of Madagascar. ‘To stay alone in the house is considered a sure sign of evil intent.’ (Feeley-Harnik, 1980, p. 568). A house with curtains on the outside doors, or fences and walls, was also seen as a threat to normal sociality. Even the house itself could pose problems, Feeley-Harnik wrote, since it is meant to remove the occupants from the larger social order. ‘Secrecy and separation,’ she continued, ‘indicate at best a lack of generosity, a suspiciously anti-social striving for distinction’ (Feeley-Harnik, 1980, p. 581). Villagers in the mountainous Zinacantan region of southern Mexico. They too have also been suspicious of too much domestic privacy. The typical home is fenced in, and village folk are forbidden from passing through the fence without prior approval. However, staying indoors, or closing the house door, is considered ‘a gross and open admission of being up to no good’ (Haviland & Haviland, 1983, p. 347).

Many of these cultures that have opposed privacy and favoured social visibility were egalitarian; according to Bailey (1971, p. 19), ‘equality is the reward for

421

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

422

0 521 582361 7

April 11, 2005

18:41

J. L. Locke constant vigilance’. In parallel, there was also a suspiciousness of structures and behaviours that reduced visibility, since these would surely foil the only proven means of keeping the group together and under control. I consider that there are three benefits of vigilance and eavesdropping: social control, honest signals and intimate experience. Let us now examine the first of these, which is closely linked to vigilance. Honest signals and intimate experience are tied to privacy, thus to eavesdropping, and we will address these benefits in that section. Since there is little in the way of relevant research, my treatment of social control will necessarily be historical, discursive and somewhat speculative. Social control

Some things that occur in private are intended to be secret. They may be offensive, morally wrong or even criminal. There are good reasons for humans to observe this activity, too, but in some cases there may be little benefit in doing so covertly. The reason is that looks, if interpreted as gazes, can also send messages of their own. Some are confrontations that vary in intensity from ‘I see you’ to ‘I’m keeping an eye on you’ and, in the extreme case, ‘Back off ’. On the community level, the protective function of surveillance has always been clear. ‘If by chance some good-for-nothing appeared in the neighbourhood,’ wrote Yves Castan (1989a, p. 49) in reference to French villages, ‘there were plenty of eyes to survey his movements’. In Victorian England, the rich and powerful lived side by side with the poor and powerless. This made it possible for each group to observe the other and particularly for establishment figures to keep an eye on potentially troublesome subordinates. ‘The middle-classes desired privacy for themselves,’ wrote Olsen (1974, pp. 275–276) ‘but wished the lives of the lower orders to be lived in the full blaze of publicity. Street improvements and slum clearance schemes were designed to bring the poor out into the open, where they could be observed, reproved and instructed by their superiors.’ On an individual level, vigilance also enables humans to avoid quarrelsome or dangerous people, our equivalent of predators. In large cities, one is forced to acquire ‘street smarts’, an awareness of menacing strangers in relation to oneself, and the relation of one’s own location to places of safety. Predictably, the best security – as criminologists have shown – is the presence of some reasonable number of non-predatory people on the street. For a city to be safe, ‘there must be eyes upon the street,’ wrote Jacobs (1961, p. 45), ‘eyes belonging to those we might call the natural proprietors of the street’. Research in the ensuing years has been supportive of this view (Kelling & Coles, 1996). Primates in hierarchically organized groups spend more time looking at each other when they could be looking for predators or food (Caine & Marra, 1988) and animals do more social looking within mixed than in homogeneous groups (Treves, 1999). These findings are relevant to Putnam’s (1993) study of provincial

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:41

Humans: control, signalling and intimate experience self-government in Italy. Provinces that lacked trust, he found, spent a great deal of time keeping an eye on each other. Watching humans

With vision alone, non-human primates pick up cues to sex, age and rank, the last inferred from dominance and submissive displays. We humans care about these things, too, and also transmit much of our information visually, through physical alterations and adornments. Among the Kayapo of the Amazon forest, visible affectations include pierced ears, lip plugs, penis sheaths and body painting (Turner, 1980). These adornments convey messages about status as well as personal roles and significance and do so just as surely as verbal signals. In modern societies, hairstyle, cosmetics, jewellery, eyeglasses, tattoos and body rings – to say nothing of cell phones, water bottles, clothes, shoes, handbags, briefcases, fanny packs, shopping bags and backpacks – send visual signals about who we are or how we wish to be perceived. The desire to enhance personal images goes back at least 28 000 years. Studies of the ‘Venus’ figurines and burial sites indicate that women many millennia ago were already wearing hats, dresses and various bodily adornments (Soffer et al., 2000). This suggests that, before they were securely and privately housed, our historical ancestors already had some sense of self, a matter to which we will return shortly. We are not, of course, merely intelligible through our clothes and other objects of material culture. Like other species, humans have a number of ritualized action patterns that presuppose visualization (cf. Smith, 1977). These include the facial and bodily displays that emerge in infancy, are seemingly universal (Schiefenh¨ ovel, 1997) and occur in blind as well as sighted infants (Eibl-Eibesfeldt, 1973). Under the influence of culture, humans take on additional gestures – some functioning as salutations, others signalling transition points in verbal engagements (Kendon, 1990) – and learn rules of proxemics that suggest possible ranges of interpersonal distance (Hall, 1966). Personal status and relational intimacy are also revealed by touching (Hall, 1996) and, in the case of single women in America, hair flips and head tosses (Moore, 1985). The eyes send many different types of social and emotional signal. We saw earlier that in primates socially dominant individuals receive more gazes than subordinate ones. This relationship also holds in humans. At any given moment in time, the person who is being looked at is usually the person who is talking and that will typically be the person with the highest status (Bales et al., 1951; Fisek & Ofshe, 1970; Exline et al., 1975; Abramovitch, 1976; Kalma, 1991). With all these visible signals, it would be surprising indeed if people did not create opportunities to be looked at and to do so on their own terms. In 1800, Parisians began to put their public selves on parade. That is when pavements

423

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

424

0 521 582361 7

April 11, 2005

18:41

J. L. Locke came to Paris and merchants adaptively repositioned their shops and displays. In the new pavement caf´es, the chairs were ‘always placed towards the street,’ wrote an urbanologist, ‘as the chairs in a theatre are placed towards the stage’ (Oosterman, 1992, p. 161). Promenades were once expected to achieve an instructive or regulatory function. When a family went out for an evening stroll, it was assumed that the husband would ‘see himself as others saw him,’ according to Cranz (1980, p. S80), ‘the head of a family, wife on arm, children in tow, all in Sunday best. Reformers reasoned that he would experience this as pleasurable and resolve to make it the mainstay of his life.’ In 1890, the commissioners of Boston’s parks department saw public viewings as a course of moral instruction. The mere sight of families was expected to exert ‘a wholesome influence’on other patrons, and to do so far more effectively than laws and police ever could (Cranz, 1980, p. 581). In contemporary America, recreational vigilance is largely carried out in parks and malls. In a survey conducted in the early 1970s, a fifth or fewer of the patrons of two parks in Portland, Oregon said they went to the parks to walk, eat, talk, read, engage in crafts or hobbies, or exercise. Far more patrons, fully 55%, said they went to the parks in order to watch other people (Love, 1973). In a survey conducted in the Los Angeles area a decade later, adolescents said that the main reason they went to a particular mall, after shopping, was to look for members of the opposite sex (Anthony, 1985). There are several circumstances in which vigilance is exercised in relation to intimate relationships. Buss (1988, 1997) surveyed American couples to see how frequently they reported the use of vigilance in order to control intimate relationships. Items in the survey included unexpectedly calling and dropping by a place to see if the partner was there and remaining nearby, or at least in visual contact, during social engagements. Men and women reported equal levels of vigilance, but there was a significant correlation for men, and not women, between levels of vigilance and ratings of partner attractiveness. Control and intimacy are also conjoined in many cases of the crime ‘stalking’. Since stalking is usually defined as an unwelcome act of ‘perceptual following’ that is overt or blatant, it qualifies as vigilance. In one study, 57% of stalkers had previously been in an intimate relationship with the victim (Hall, 1998). In another study, approximately a third of all stalkers were considered intimacy seekers. Most lived alone and had never had a romantic partner (Mullen et al., 2000). Ethological studies of human vigilance

In non-human primates, individuals tend to look up less often when a group member is nearby (Hirsch, 2002). A similar trend has been found in humans. Observing students in a university snack bar, Barash (1972) found that cumulative

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:41

Humans: control, signalling and intimate experience looking-up frequency was significantly higher in solitaires than individuals in groups. In a similar study conducted in Germany, Wirtz & Wawra (1986) observed university students having lunch in a refectory. Each sat alone or with one to four other students. In these subjects, it was found that the time spent looking away from the table steadily decreased as the number of number of people at the table increased, possibly because this increased the proximity of others. In Wirtz & Wawra (1986), male students spent significantly more time looking away from their table than females. This fits with primate research, reviewed earlier, that revealed a male looking bias, particularly for distant areas. It also agrees with Aiello (1972, 1977), who found that men looked significantly longer at each other than women did when seated 10 feet (3.2 m) or more apart, a trend that was reversed for shorter distances. Paradoxically, most of the work on vigilance in humans involves detection of signals, whereas in animals, vigilance involves attention to the existence and behaviour of individuals (also see studies of social monitoring and comparison). The disposition of females to look longer at near individuals may be linked to a tendency to rely on the support of group members, while the disposition of males to look longer at distant individuals may be associated with the need to address the threats posed by strangers. Stripped to the basics, here are two issues – intimacy (the network ‘glue’) and control – that concern human women and men. These issues, as we will see, have been connected to sex differences in social monitoring for the last six or seven centuries of recorded history.

Looking at Eavesdropping in animals

Much information is acquired by social observation: looking at conspecifics. For example, male Mallee dragon lizards Ctenophorous fordi produce significantly more ejaculate and spend 60% more time copulating with a female previously seen copulating with another male than do males not having this prior perceptual experience (Olsson, 2001). In this example, the source of information did not involve signals. However, an important subset of observational information comes from the signals of others. Such information is gathered by eavesdropping, a behaviour that is defined in animals as ‘the use of information in signals by individuals other than the primary target’ (Ch. 2). The context for eavesdropping is a communication network (Ch. 1); therefore it is not surprising that this volume discusses at length the evidence for eavesdropping by animals (e.g. Chs. 2 and 5). There have also been a number of recent reviews of eavesdropping (e.g. McGregor & Peake, 2000).

425

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

426

0 521 582361 7

April 11, 2005

18:41

J. L. Locke Among the primates, maintenance of societies – including kin and power relations – requires that individuals spend a certain amount of time gathering information about other group members. These internal appraisals, which in human research are usually called social comparison and in non-human primates are considered a form of social vigilance (Hirsch, 2002), are required if animals are to alter or maintain their status. Since Cheney & Seyfarth (Ch. 25) describe research in this area, I will limit my own review to studies that expose links to our own species. Primates’ resources include cooperative relationships, which may involve highranking animals. As Whiten (1993, p. 719, italics his) has pointed out, ‘simply to be seen by others grooming with high-ranking A, or chatting with high-status B, is worth something to the individual because of what this advertises with respect to future coalition.’ An individual that has these kinds of social resource is considered to be rich in ‘social attention holding potential’ (Gilbert, 1989). Animals may also look within the group for individuals with valued physical resources. For example, a perceptual target of so-called ‘scroungers’ is the foraging success of other animals (Beauchamp, 2001). There are variations between species in social vigilance, partly because of differences in social organization. Consider squirrel monkeys, which live in large groups that are characterized by cliques, subgroups and dominance hierarchies, and cotton-top tamarins Saguinus oedipus, who live in more egalitarian family groups known for cooperation, sharing and relative peacefulness. In a comparison of social looking during foraging, the congenial tamarins devoted 17% of their time to within-group vigilance. The more competitive squirrel monkeys, by contrast, devoted 45% of their time attending to group members (Caine & Marra, 1988). A feature of eavesdropping by animals is that it is usually carried out by isolated individuals who do not subsequently share their perceptual intake with others, although in a densely populated area other observers may individually sample the same activity on their own. The perceptual target of eavesdropping is often a pair or small group of individuals, which provides the observer with interactive or relational information connected with fighting or sex, and it does so with minimal risk or expenditure of effort. Eavesdropping in humans

In this section, I follow accepted semantic practice and use the term eavesdropping only where the act of observation occurs surreptitiously. I also include cases of social vigilance that do not involve signals or interactions (as do Cheney & Seyfarth in Ch. 25). I noted earlier that domestic vigilance has gone largely unstudied ethologically, but when it comes to human eavesdropping there is no record of empiricism

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:41

Humans: control, signalling and intimate experience whatsoever. To be sure, there are publications that use the word eavesdropping, but these typically describe government anti-crime programmes that include wiretapping and surveillance. What is missing is research on the behaviour of social eavesdropping in a naturalistic context. The lack of research on human eavesdropping seems odd, since the practice is neither rare nor lacking in benefits. When asked, people usually admit that they have eavesdropped in the past, or even do so habitually. Frequently, the admissions are offered shyly, occasionally with embarrassment, but I have yet to find anyone who denies ever having engaged in this practice. This is not to say that everyone peeks through keyholes. Most of us ‘tune in’ less adventurously. When in a restaurant or waiting room, for example, we tend to accomplish our perceptual business in a number of optical stabs, interrupted by bogus glances at other features of the physical or social landscape. If the subject suddenly looks up, the invasion may be disguised by a slow and smooth deflection, as though a continuous sweep was in progress when the ‘interruption’ occurred. If people are naturally inclined to penetrate the private spaces of others, and just as naturally resist such intrusions themselves, one might expect historical evidence of these dispositions, perhaps in art or literature. In fact, there was activity in both media in the seventeenth century, from the paintings of Dutch artist Nicolaes Maes to the novels of Le Sage and Hawthorne and the plays of Marivaux. These depictions suggest that our historical ancestors were acutely aware of eavesdropping and may even have approved of it. However, there are also church and court documents going back three centuries earlier, in several different cultures, and these tell us something about the relative frequency of eavesdropping as a behaviour, and a crime. The !Kung hunter–gatherers, as we have seen above, welcomed round-the-clock surveillance and intentionally subjected themselves to a panoptical living arrangement (cf. Bentham, 1791). This made vigilance easy, but for the same reason it made covert eavesdropping impossible (P. Wiessner, personal communication). One assumes the !Kung’s residential arrangement was somewhat representative of historically earlier ways of living, when variations in the availability of food required individuals constantly to relocate. With the advent of agriculture, however, the new sedentists departed from the hunter–gatherer pattern, building huts that could be lived in – not just used for storage – and spacing them more widely. This necessitated an aggressive form of perceptual intrusion. At the same time, groups began to expand, and strangers grew more numerous. In a brief space of time, a lifestyle that was two million years old – open living, with visual monitoring – began to unravel. Before the development of structural privacy as it is enjoyed in modern societies, some degree of solitude was achieved behaviourally. Bird (1983) reported that

427

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

428

0 521 582361 7

April 11, 2005

18:41

J. L. Locke the Naiken people of India exhibit nachika, literally a ‘shyness’ or ‘reticence’ that protects them from direct encounters with others. Since the Naiken live openly, reticence provides relief against what Bird called ‘involuntary intimacy.’ Similar observations were made by Fejos (1943), who studied the Yagua people of northeastern Peru in the early 1940s. All the families of a clan, which ranged from 25 to 50 members, lived communally in one large house. Fejos noted that although there were no partitions, members could achieve privacy at any time simply by turning away. ‘No one in the house,’ wrote Fejos (1943, p. 87), ‘will look upon, or observe, one who is in private facing the wall, no matter how urgently he may wish to talk to him’. Note that the privacy achieved by these individuals was in each case negotiated with, and conferred by, others. It began when the privacy-seeking individual gave an observable sign. The observers, out of respect for the person, then reduced or suspended evaluation. Goffman (1963) called this ‘civil inattention’. Perhaps these behavioural means of securing privacy were sufficient, for even with inclement weather, social competitors and wild animals, little interest was shown in domestic walls (Carpenter, 1966; Rapoport, 1969). Consequently one is curious about the residents of more hospitable climes who nonetheless chose to live behind walls. The reasons for these exceptions to climatic determinism, Rapoport (1969) pointed out, may have had something to do with religion, status or some ‘other’ factor. One candidate for the ‘other’ factor, according to Wilson (1988), would have been the desire to escape constant scrutiny. But there is another possibility. We have already seen that the !Kung sat in full view of each other during their time in camp. If new members continually join such arrangements, eventually something has to give. Individuals who cut back on their looking time will discover that the machinations of an increasingly complex, if not Machiavellian, society have left them frightfully unaware and out of step. Alternatively, those who continue to crank up their looking time in step with population growth will soon have no time to do anything but look. Therefore, the critical factor may have been the need to minimize the time that they, as members of burgeoning groups, had to spend surveying the social landscape. Little wonder that the desire for privacy grew as people became accustomed to domestic life. In the 1960s, the Sarakatsani were a small group of shepherds who alternately, by season, inhabited the Zagori Mountains and plains of Greece. To them, a hut was inviolate. ‘Whatever takes place within the sanctuary of its walls is private and sacred to the members of the family’ wrote Campbell (1964, p. 292). ‘No stranger may invade it without an invitation.’ Occupants could only be safe by assuming that they, like birds (e.g. Metcalfe, 1984a,b), grazing animals (e.g. Underwood, 1982) or isolated monkeys (e.g. Treves

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:41

Humans: control, signalling and intimate experience et al., 2001) were ineligible for assistance. Therefore, the need was for walls that were not merely visually obstructive but also secure. Eavesdropping became the only way to restore information that had once been available. However, eavesdropping was not merely restorative, for with increasing privacy, this very penetrant means of observation became the only way to obtain the newest and highest grade of personal information, one that had never been available previously. Honest signals

As people spent more time behind walls, direct sensory information about them became less available. This posed problems for the community, but the experience of privacy also altered the people themselves. In time, the most honest and reliable information about individuals was only available to those who were behind the walls, or in personal relationships, with them. Others were excluded. Therefore, behaving as trained ethologists, eavesdroppers attempted to conceal themselves in order to avoid detection, which would alter or discontinue the flow of desired information. People behave differently when they believe others are unable to see them. When shielded from public view, they have the opportunity literally to compose themselves – to decide who they are and how they would like to be perceived by others. When they plunge into the social world again, they may then do so appropriately dressed and ornamented, presenting others with the image they would most like to convey. Earlier I referred to a low-grade sense of self that antedated, or occurred early in the development of, domestic life. While the hominids may have had some level of self-awareness, along with the other primates (Hauser, 2000), every domestically living human now has two selves: one public, the other private. The public self is the way we are in the presence of others. Our private self is on view only when individuals are alone or with intimate friends. There is a telling fact about the private self, in connection with the process of perceptual theft. It is, as Baumeister (1986, p. v, italics mine) said, ‘the way the person really is’. The dishonest signals that are issued in public are not worthless, of course. These may provide information as to the way a person really is not. For example, a person who is making a conspicuous display of wealth may be ‘financially strapped’ – not wealthy at all – and also seeking to hide this fact for a reason, one that with further analysis may be discovered. Still, people in private are likely to act in ways that are, as Baumeister said, more ‘them’. This fact, by itself, increases the reliability of private behaviour. However, private behaviour is also privileged. This gives others reasons to want it, for as humans they have the inherited dispositions of evolutionary ancestors whose survival was dependent on the ability to observe

429

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

430

0 521 582361 7

April 11, 2005

18:41

J. L. Locke behaviour and infer intentions, and they enjoy vicariously experiences that are not issued for the benefit of observers. Intimate experience

After nearly two million years of watchfulness, walls enabled the chronically wary Homo sapiens to cast its senses inwards for a change. Dwellers could pay undivided attention to compelling tasks not just for three seconds, but for three hours. Free from the stares and queries of villagers, the new residents could begin to examine their own lives and think about how they differed from others. With shielding, they could create or discover the existence of a deeper and more reflective form of themselves and begin to contrast this with their public presentation. The time that domestication liberated from external vigilance could be devoted to matters that were occurring – or with additional attention could be initiated – on the inside. These would have included personal and communal activities. Family members could, at last, devote their undivided attention to each other. By creating an ‘outside’, individuals found ways in which more intimate relations could be developed with members of the family on the ‘inside’. Consequently, to look inside a house was to get unprecedented glimpses of intimate behaviour. If someone peeks through a crack or keyhole, how will this come to the attention of researchers? Eavesdroppers are no more likely to be detected by an ethologist than by their perceptual prey, nor would they be likely to describe their activities truthfully to an interviewer. The situation seems hopeless, and yet we do know something about eavesdropping, especially the kind that occurred many centuries ago when the threat of privacy was new and attempts to breech it were frequent, adaptive and perhaps even honourable. In sixteenth century England, there was a law against adultery and it required eye-witness testimony. Court records have been preserved, revealing the testimony and identity of witnesses. Frequently, the lead or sole witness was a woman who had peeked through a door from within the house, or crack in the wall from an adjoining house. In a case that occurred in London in 1598, a housewife named Margaret Browne watched a tryst involving the woman who lived next door, her looking ‘bout’ – like the adulterous activity itself – lasting for an entire afternoon (Crawford & Gowing, 2000). I have inspected many cases involving this sort of domestic eavesdropping. Although I kept no detailed count, it is clear that the typical perpetrator was female. One might suppose that this is because women were merely home more often, but in the sixteenth and seventeenth centuries, the husband was often somewhere about the house, too. When Margaret Browne saw what was happening next door, she called her husband to the crack to confirm her observations. Mr Browne took a brief look and left, but Margaret remained at the crack, taking mental notes. Her courtroom testimony two weeks later was extraordinarily detailed, down to the exact words and phrases of the lovers and details of their

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:41

Humans: control, signalling and intimate experience various sex acts, as well as the colour of her neighbour’s underwear. Clearly, this was a memorable experience for Mrs Browne. In early fourteenth century France, they had something more ominous than adulterers: heretics. In Montaillou, a small village in the Pyrenees, there was a group of Cathars that were actively working to oppose the Catholic Church. In order to get needed evidence, the church requested parishioners to bring in eyewitness testimony, which was subsequently used in court (Le Roy Ladurie, 1978). These church records were preserved, testimony revealing that the women of Montaillou, in general, were unusually active in the more subtle form of eavesdropping that involves listening at keyholes and looking through holes in domestic doors and walls. The men ‘were inquisitive enough,’ wrote Le Roy Ladurie (1978, p. 257), ‘but their curiosity was nothing beside that of the women’ (italics mine). Centuries later, in a completely different context, a similar comment was made about the women of Italy. These women, too, were ‘curious by nature’ according to Nicole Castan (1989b, p. 417). ‘Women of the lower orders shamelessly admitted it.’ One confessed that ‘she was “obliged” to follow the movements of a passer-by, another that she could not help overhearing a conversation or lying in wait for a neighbor’. While the courts welcomed eavesdropping as eyewitness testimony, they punished cases of eavesdropping when it proved to be disruptive to community life. In England, it was a crime to ‘listen under walls or windows, or the eaves of a house to hearken after discourse, and thereupon to frame slanderous and mischievous tales’. Data analysed by McIntosh (1998) revealed that for a good 200 years, beginning in the 1370s, eavesdropping made up about 8% of all social crimes. But here we find a sex reversal; during this period, about 80% of the courts having some incidence of eavesdropping happened to hear male cases only. Why such a high percentage of men? McIntosh (1998) suggested that the men who were caught listening under eaves were actually attempting to control their communities by investigating the possibility of domestic misbehaviour. If so, many of the arrests for eavesdropping may well have been instances of vigilance, an activity that in other primates also favours males. The irony is that much of the eavesdropping – a misdemeanour – was undertaken by people who may have been attempting to prevent domestic misbehaviour. When the English eavesdroppers witnessed moral transgressions, the obvious next step was to broadcast what they had seen and this is what many did. However, recall that the second part of the eavesdropping law involved framing ‘slanderous and mischievous tales’. The fourteenth and fifteenth century English were still in the process of privatizing and so felt ambivalent about publicizing the results of perceptual invasion. The male eavesdroppers may have been attempting to police their communities, but they could also have been attempting to control individuals. I base this,

431

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

432

0 521 582361 7

April 11, 2005

18:41

J. L. Locke in part, on the fact that stalking, as discussed earlier, is widely understood as a means of controlling the life of another and it also has a near-identical sex bias to medieval English eavesdropping. In a large American survey, 87% of the stalkers were male (Tjaden & Thoennes, 2000) and similar statistics are available for other cultures (Mullen et al., 2000). If honest signals, social control and intimate experience are interconnected, it would not be surprising if visual monitoring sometimes leads to eavesdropping. In Mineville, a town of 1000 inhabitants in America’s Rocky Mountains, Blumenthal (1932, p. 103) noted that peoples’attempts to live privately merely inflamed the curiosities of others. Some, he wrote, became ‘more thoroughly known than would have been the case had they not tried so obviously to guard their privacies, for in doing so they made themselves mysterious, and thus stimulated the curiosity of the people so that more than ordinary attention was given to discovering something about them’. Because of such interconnections, it may be difficult to carry out a motivational analysis on anything but the initial bout of observation. In the daily parade of public selves, people in search of honest signals have been forced to invade private spaces, thereby accessing the intimate experience that occurs there, finally finding themselves in possession of knowledge of the kind that leads to social control.

Looking at and for: a functional comparison Cheney & Seyfarth (Ch. 25) have described the need of primates to monitor their fellow group members, but do non-human primates actually engage in dictionary-definition eavesdropping: that is, observe under conditions of stealth? It is not clear that researchers have asked this question, and yet it appears that animals sometimes secure conditions of perceptual privacy – a circumstance that favours eavesdropping – before undertaking certain behaviours. For example, when subordinate males approach females in oestrus, they look around, evidently to see if they and their intended partners are under review. This is evidently because the sight of a presenting female is arousing and may produce unwanted competition (Hall & De Vore, 1965). Females do the same. Kummer (1968, p. 41) described the attempt by adult female baboons to copulate with young males ‘behind the backs of their leaders’ and Smuts (1987) presented photographic evidence of a rhesus female checking to see if she and an extragroup male were being watched before they commenced mating activities. Whether primates ever undertake within-group evaluation from obscure positions remains to be demonstrated, but some males take measures that make this unlikely. I refer to consortship, the practice whereby a male browbeats a female into following his exodus from the group, for mating purposes, sometimes over considerable distances and for extended periods of time (Goodall, 1986; McGinnis,

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:41

Humans: control, signalling and intimate experience 1979; Tutin, 1979). These bouts of absenteeism are of particular interest in relation to eavesdropping, since they suggest an awareness, on some level, that some activities are better pursued in perceptual privacy. As groups enlarge beyond some optimal size, any savings in predator vigilance may be mitigated by new observational needs within the group. Other things being equal, the larger the group the more competition there will be for food and other resources (van Schaik et al., 1983). This increases conspecific threat and, with it, alliances, which also must be visually monitored (Treves, 2000) and personally serviced (Dunbar, 1993). In red colobus and redtail monkeys, Treves (1999) found that if vigilance was needed for external activity it came out of the time that would otherwise be devoted to within-group looking. This, he speculated, might explain the fact that in primates there has been little evidence for the hypothesis that looking time decreases as group size increases (Pulliam, 1973; Elgar, 1989). For it is difficult to see how total looking time could decrease if individuals are forced to keep an eye on individuals in their own group. Indeed, there is a tension between the time devoted to vigilance outside one’s group and the time spent looking within it. In white-faced capuchin males, as indicated above, external competition increases males’attention to outside males, at the expense of internal vigilance (Rose & Fedigan, 1995), although presumably reducing cuckoldry (Gould et al., 1997). In one study, redtail monkeys glanced at associates more often when in the presence of red colobus monkeys than in purely conspecific groups (Treves, 1999). In various species, focus of attention is susceptible to rapid and dynamic shifts from family and alliance members to strangers and predators. Some types of monitoring of the physical and social environment are carried out openly – even demonstratively – while other types may be effected with stealth. These shifts require a dynamic model that recognizes the continuous interplay of multiple variables. Indeed, the optimal paradigm would seem to be one that flexibly admits all types of observation.

Toward a unified model Is it possible to achieve a model that accounts for core principles associated with vigilance, social observation and eavesdropping? In both human and non-human primates, individuals appear to spend less time looking for predators and competitors, and more time looking at each other, if group members are nearby. In hierarchically organized societies – whether inherently complex squirrel monkey groups or highly politicized human societies – a great deal of internal attention appears to be needed if individuals are to keep or to feel adequately informed. In Machiavellian societies, inference would seem to play an exaggerated

433

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

434

0 521 582361 7

April 11, 2005

18:41

J. L. Locke role. In such societies, there may be greater need of observation, particularly of the surreptitious kind. It should be noted that when one looks at fellow group members, information becomes available not only about them but also about the events to which they may be reacting. In our own species, for example, when the young enter novel situations they often monitor caregivers’ facial activity, which serves as a reliable index of danger. Thus, internal observation offers both within-group information needed for social comparison and extra-group information about predators and competitors. In both human and non-human primates, subordinate individuals spend more time watching dominant individuals than the reverse arrangement. In Marivaux’s plays, according to Trapnell (1987), keyholes enabled the young social climber ‘to distinguish between his ally and his enemies’, to ‘observe the terrain on which he must manoeuver, assess the efforts his ambition will require, determine the appropriate strategy and gauge his chances of success’. Were it not for eavesdropping, Trapnell (1987, p. 109) wrote, the world would be ‘inaccessible and even unknown’ to people born without special advantages and privileges. In Victorian England, as we saw above, the upper classes in many instances lived beside the lower ones. This arrangement gave the ruling classes unobstructed vision of the individuals they wished to control, but it also gave the lower classes a regular view of behaviours they had reason to emulate. In nineteenth century America, upwardly mobile men and women had limited perceptual access to the upper class behaviours they needed to absorb. To compensate, they used biographies as ‘handbooks’. They did so, according to Casper (1999), in the belief that the difference between public success and failure lay in the private habits that defined one’s character, or true self: the stuff of which compelling biographies are made. Earlier, we saw that, in non-human primates, males do more looking than females, presumably in an attempt to detect competition and danger. This vigilance is very clearly tied to control and defence. In our own species, too, males seem to have performed in a vigilant capacity more than females, boldly standing under domestic eaves and then broadcasting the perceptual ‘take’. Female networks are more extensive and stronger both in non-human primates (Dunbar, 1988) and in our own species. There is evidence of a strong female advantage in human grooming (Sugawara, 1984, 1990) as well as touching ( Jones & Yarbrough, 1985) and concerted social action (Motz, 1983). There also is evidence of a female preference for gossip – the use of speech to discuss mutual acquaintances not physically present – that spans cultures and most decades of the twentieth century (Bischoping, 1993). There are indications, additionally, that when peers offend young women in various cultures, the victims respond by working

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:41

Humans: control, signalling and intimate experience indirectly through female friends (Bj¨ orkqvist et al., 1994; Galen & Underwood, 1997; Crick & Bigbee, 1998). We saw earlier that vigilance levels in non-human primates are affected by transient factors such as mating and birth. In humans, too, vigilance levels are clearly influenced by environmental change, including increases in population, economic competition and terrorism. Surveillance cameras proliferated in the USA after 11 September 2001. Although there had been a long-standing fear of surveillant societies of the type envisaged in Orwell’s book 1984, the terrorist attacks on New York and Washington seemed to have had the opposite effect. The lack of objection noted by several newspaper columnists suggests that citizens may have derived solace from the knowledge that government officials were looking for and screening out ‘predators’. In non-human primates, as we have seen, animals may devote as much as half their waking hours to looking. Dunbar (1993) reported that animals spend as much as 20% of their time grooming in some primate groups. There are indications that crowding increases grooming (Nieuwenhuijsen & de Waal, 1982; Novak et al., 1992; Judge & de Waal, 1997) and grooming has been found to decrease withingroup monitoring (Maestripieri, 1993; Hirsch, 2002). In future work, it would be interesting to look at within-group vigilance and grooming in the same animals as a function of density and predational threat. Merely by comparing species, certain common patterns emerge, but more can be done, beginning with the resolution of definitional issues. If interdisciplinary collaboration facilitates the study of processes by which ‘senders’ use language to donate information, as Hauser et al. (2002) have argued, it will surely affect the investigation of processes by which ‘receivers’ of widely ranging communicative abilities use their senses to extract it.

Acknowledgements This chapter developed from a paper delivered to the Konrad Lorenz Institute in Altenberg, Austria in December of 2001. Portions coevolved with a larger work in progress about eavesdropping. The author wishes to acknowledge helpful comments by Adrian Treves, Eric Salzen, Michael Studdert-Kennedy, Polly Wiessner and Ben Hirsch.

References Abramovitch, R. 1976. The relation of attention and proximity to rank in preschool children. In: The Social Structure of Attention, ed. M. R. A. Chance & R. R. Larsen. London: John Wiley, pp. 153–176.

435

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

436

0 521 582361 7

April 11, 2005

18:41

J. L. Locke Aiello, J. R. 1972. A test of equilibrium theory: visual interaction in relation to orientation, distance and sex of interactants. Psychonomic Science, 27, 335–336. 1977. Visual interaction at extended distances. Personality and Social Psychology Bulletin, 3, 83–86. Anthony, K. H. 1985. The shopping mall: a teenage hangout. Adolescence, 20, 307–312. Bailey, F. G. 1971. Gifts and poison. In: Gifts and Poison: The Politics of Reputation, ed. F. G. Bailey. Oxford: Basil Blackwell, pp. 1–25. Baldellou, M. & Henzi, P. 1992. Vigilance, predator detection and the presence of supernumerary males in vervet monkey troops. Animal Behaviour, 43, 451–461. Bales, R. F., Strodtbeck, F. L., Mills, T. M. & Roseborough, M. E. 1951. Channels of communication in small groups. American Sociological Review, 16, 461–468. Barash, D. P. 1972. Human ethology: the snack-bar security syndrome. Psychological Reports, 31, 577–578. Barth, F. 1975. Ritual and Knowledge Among the Baktaman of New Guinea. New Haven, CT: Yale University Press. Baumeister, R. F. 1986. Preface. In: Public Self and Private Self, ed. R. F. Baumeister. New York: Springer, pp. v–viii. Beauchamp, G. 2001. Should vigilance always decrease with group size? Behavioral Ecology and Sociobiology, 51, 41–46. Bentham, J. 1791. Panopticon: Postscript; Part I: Containing Further Particulars and Alterations Relative to the Plan of Construction Originally Proposed; Principally Adapted to the Purpose of a Panopticon Penitentiary-House. London: Printed for T. Payne, at the Mews-Gate. Biben, M., Symmes, D. & Bernhards, D. 1989. Vigilance during play in squirrel monkeys. American Journal of Primatology, 17, 41–49. Bird, N. 1983. Conjugal units and single person: an analysis of the social system of the Naiken of the Nilgiris (South India). D. Phil. Thesis, University of Cambridge. Bird-David, N. 1994. Sociality and immediacy: or, past and present conversations on bands. Man, 29, 583–603. Bischoping, K. 1993. Gender differences in conversation topics, 1922–1990. Sex Roles, 28, 1–18. ¨ Bj¨ orkqvist, K., Osterman, K. & Lagerspetz, K. M. J. 1994. Sex differences in covert aggression among adults. Aggressive Behaviour, 20, 27–33. Blumenthal, A. 1932. Small Town Stuff. Chicago, IL: University of Chicago Press. Buss, D. M. 1988. From vigilance to violence: tactics of mate retention in American undergraduates. Ethology and Sociobiology, 9, 291–317. 1997. From vigilance to violence: mate retention tactics in married couples. Journal of Personality and Social Psychology, 72, 346–361. Caine, N. G. & Marra, S. L. 1988. Vigilance and social organization in two species of primates. Animal Behaviour, 36, 897–904. Campbell, J. K. 1964. Honour, Family and Patronage: A Study of Institutions and Moral Values in a Greek Mountain Community. Oxford: Clarendon Press. Carpenter, E. 1966. Image making in Arctic art. In: Sign, Image, Symbol, ed. G. Kepes. New York: George Braziller, pp. 206–225. Casper, S. E. 1999. Constructing American Lives: Biography and Culture in Nineteenth-Century America. Chapel Hill, NC: University of North Carolina Press.

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:41

Humans: control, signalling and intimate experience Castan, N. 1989b. The public and the private. In: A History of Private Life. III. Passions of the Renaissance, ed. R. Chartier. Cambridge, MA: Harvard University Press, pp. 403–446. Castan, Y. 1989a. Politics and private life. In: A History of Private Life. III. Passions of the Renaissance, ed. R. Chartier. Cambridge, MA: Harvard University Press, pp. 21–67. Chance, M. R. A. 1967. Attention-structure as the basis of primate rank orders. Man, 2, 503–518. Chance, M. R. A. & Jolly, C. J. 1970. Social Groups of Monkeys, Apes and Men. London: Jonathan Cape. Cheney, D. L. & Seyfarth, R. M. 1981. Selective forces affecting predator alarm calls of vervet monkeys. Behaviour, 76, 25–61. Cowlishaw, G. 1998. The role of vigilance in the survival and reproductive strategies of desert baboons. Behaviour, 135, 431–452. Cranz, G. 1980. Women in urban parks. Signs: Journal of Women in Culture and Society, 5, S79–S95. Crawford, P. & Gowing, L. 2000. Women’s World in Seventeenth-Century England: A Sourcebook. London: Routledge. Crick, N. R. & Bigbee, M. A. 1998. Relational and overt forms of peer victimization: a multiinformant approach. Journal of Consulting and Clinical Psychology, 66, 337–347. Dixson, A. F., Scruton, D. & Herbert, J. 1975. Behaviour of the talapoin monkey (Miopithecus talapoin) studied in groups in the laboratory. Journal of Zoology, (London), 176, 177–210. Draper, P. 1978. The learning environment for aggression and anti-social behavior among the !Kung. In: Learning Non-Aggression: The Experience of Non-Literate Societies, ed. M. F. A. Montagu. Oxford: Clarendon Press, pp. 31–53. Dunbar, R. I. M. 1988. Primate Social Systems. Ithaka, NY: Cornell University Press. 1993. On the origins of language: a history of constraints and windows of opportunity. Behavioral and Brain Sciences, 16, 720–735. Eibl-Eibesfeldt, I. 1973. The expressive behaviour of the deaf- and -blind-born. In: Social Communication and Movement: Studies of Interaction and Expression in Man and Chimpanzee, ed. M. von Cranach & I. Vine. London: Academic Press, pp. 163–194. Elgar, M. A. 1989. Predator vigilance and group size in mammals and birds: a critical review of the empirical evidence. Biological Review, 64, 13–33. Ellis, A. & Beattie, G. 1986. The Psychology of Language and Communication. New York: Guilford Press. Exline, R. V., Ellyson, S. L. & Long, B. 1975. Visual behaviour as an aspect of power role relationships. In: Nonverbal Communication of Aggression 2, ed. P. Pliner, L. Krames & T. Alloway. New York: Plenum press, pp. 21–52. Feeley-Harnik, G. 1980. The Sakalava house (Madagascar). Anthropos, 75, 559–585. Fejos, P. 1943. Viking Fund Publications in Anthropology, No. 1: Ethnography of the Yagua. New York: Viking Fund. Fisek, M. H. & Ofshe, R. 1970. The process of status evolution. Sociometry, 33, 327–346. Fragaszy, D. M. 1990. Age and sex differences in the organization of behavior in wedge-capped capuchins, Cebus olivaceus. Behavioral Ecology, 1, 81–94. Galen, B. R. & Underwood, M. K. 1997. A developmental investigation of social aggression among children. Developmental Psychology, 33, 589–600.

437

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

438

0 521 582361 7

April 11, 2005

18:41

J. L. Locke Gilbert, P. 1989. Human Nature and Suffering. London: Erlbaum. Goffman, E. 1963. Behavior in Public Places. New York: Free Press. Goodall, J. 1986. The Chimpanzees of Gombe: Patterns of Behavior. Cambridge, MA: Harvard University Press. Gould, L., Fedigan, L. M. & Rose, L. M. 1997. Why be vigilant? The case of the alpha animal. International Journal of Primatology, 18, 401–414. Gregor, T. 1970. Exposure and seclusion: a study of institutionalized isolation among the Mehinacu Indians of Brazil. Ethnology, 9, 234–250. Hall, D. M. 1998. The victims of stalking. In: The Psychology of Stalking: Clinical and Forensic Perspectives, ed. J. R. Meloy. San Diego, CA: Academic Press, pp. 115–137. Hall, E. T. 1966. The Hidden Dimension. New York: Doubleday. Hall, J. A. 1996. Touch, status, and gender at professional meetings. Journal of Nonverbal Behavior, 20, 23–44. Hall, K. R. L. 1960. Social vigilance, behaviour of the chacma baboon, Papio ursinus. Behaviour, 16, 261–294. Hall, K. R. L. & De Vore, I. 1965. Baboon social behavior. In: Primate Behavior: Field Studies of Monkeys and Apes, ed. I. De Vore. New York: Holt, Rinehart & Winston. pp. 53–110. Hauser, M. 2000. Wild Minds: What Animals Really Think. London: Allen Lane. Hauser, M., Chomsky, N. & Fitch, W. T. 2002. The faculty of language: what is it, who has it, and how did it evolve? Science, 298, 1569–1579. Haviland, L. K. & Haviland, J. B. 1983. Privacy in a Mexican Indian village. In: Public and Private in Social Life, ed. S. I. Benn & G. F. Gaus. London: Croom Helm, pp. 341–361. Hirsch, B. T. 2002. Social monitoring and vigilance behavior in brown capuchin monkeys (Cebus apella). Behavioral Ecology and Sociobiology, 52, 458–464. Horrocks, J. A. & Hunte, W. 1986. Sentinel behaviour in vervet monkeys: who sees whom first. Animal Behaviour, 34, 1566–1567. Jacobs, J. 1961. The Death and Life of Great American Cities. London: Penguin Books. Jones, S. E. & Yarbrough, A. E. 1985. A naturalistic study of the meanings of touch. Communication Monographs, 52, 19–56. Judge, P. G. & de Waal, F. B. M. 1997. Rhesus monkey behaviour under diverse population densities: coping with long-term crowding. Animal Behaviour, 54, 643–662. Kalma, A. 1991. Hierarchisation and dominance assessment at first glance. European Journal of Social Psychology, 21, 165–181. Kelling, G. L. & Coles, C. M. 1996. Fixing Broken Windows: Restoring Order and Reducing Crime in our Communities. New York: The Free Press. Kendon, A. 1990. Conducting Interaction: Patterns of Behavior in Focused Encounters. Cambridge: Cambridge University Press. Keverne, E. B., Leonard, R. A., Scruton, D. M. & Young, S. K. 1978. Visual monitoring in social groups of talapoin monkeys (Miopithecus talapoin). Animal Behaviour, 26, 933–944. Koenig, A. 1994. Random scan, sentinels or sentinel system? A study in captive common marmosets (Callithrix jacchus). In: Current Primatology, Vol. 11: Social Development, Learning and Behaviour, ed. J. J. Roeder, B. Thierry, J. R. Anderson & N. Herrenschmidt. Strasbourg: University Louis Pasteur, pp. 69–76.

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:41

Humans: control, signalling and intimate experience Kummer, H. 1968. Social Organization of Hamadryas Baboons. Chicago, IL: University of Chicago Press. Lee, R. B. 1979. The Dobe !Kung. New York: Holt, Rinehart & Winston. Le Roy Ladurie, E. 1978. Montaillou: Cathars and Catholics in a French Village. London: Scolar Press. Love, R. L. 1973. The fountains of urban life. Urban Life and Culture, 2, 161–209. Maestripieri, D. 1993. Vigilance costs of allogrooming in macaque mothers. American Naturalist, 141, 744–753. Maples, W. R., Maples, M. K., Greenhood, W. F. & Walek, M. L. 1976. Adaptations of crop-raiding baboons in Kenya. American Journal of Physical Anthropology, 45, 309–316. McGinnis, P. R. 1979. Sexual behavior in free-living chimpanzees: consort relationships. In: The Great Apes, ed. D. A. Hamburg & E. R. McCown. Menlo Park, CA: Benjamin/Cummings, pp. 429–439. McGregor, P. K. & Peake, T. M. 2000. Communication networks: social environments for receiving and signalling behaviour. Acta Ethologica, 2, 71–81. McIntosh, M. K. 1998. Controlling Misbehavior in England, 1370–1600. Cambridge: Cambridge University Press. Metcalfe, N. B. 1984a. The effects of mixed-species flocking on the vigilance of shorebirds: who do they trust? Animal Behaviour, 32, 986–993. 1984b. The effects of mixed-species flocking on the vigilance of shorebirds: is visibility important? Animal Behaviour, 32, 981–995. Moore, M. M. 1985. Nonverbal courtship patterns in women: context and consequences. Ethology and Sociobiology, 6, 237–247. Motz, M. F. 1983. True Sisterhood: Michigan Women and their Kin 1820–1920. Albany, NY: State University of New York Press. Mullen, P. E., Path´e, M. & Purcell, R. 2000. Stalkers and their Victims. Cambridge: Cambridge University Press. Nieuwenhuijsen, K. & de Waal, F. B. M. 1982. Effects of spatial crowding on social behavior in a chimpanzee colony. Zoo Biology, 1, 5–28. Novak, M. A., O’Neill, P. & Suomi, S. J. 1992. Adjustments and adaptations to indoor and outdoor environments: continuity and change in young adult rhesus monkeys. American Journal of Primatology, 28, 125–138. Olsen, D. J. 1974. Victorian London: specialization, segregation, and privacy. Victorian Studies, 17, 265–278. Olsson, M. 2001. ‘Voyeurism’ prolongs copulation in the dragon lizard Ctenophorous fordi. Behavioral Ecology and Sociobiology, 50, 378–381. Oosterman, J. 1992. Welcome to the pleasure dome: play and entertainment in urban public space – the example of the sidewalk caf´e. Built Environment, 18, 155–164. Pitcairn, T. K. 1976. Attention and social structure in M. fascicularis. In: The Social Structure of Attention, ed. M. R. A. Chance & R. R. Larsen. London: John Wiley pp. 51–81. Poirier, F. E. 1972. The St Kitts green monkey (Cercopithecus aethiops sabaeus): ecology, population dynamics, and selected behavioral traits. Folia primatologica, 17, 20–55.

439

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

440

0 521 582361 7

April 11, 2005

18:41

J. L. Locke Pulliam, H. R. 1973. On the advantages of flocking. Journal of Theoretical Biology, 38, 419–422. Putnam, R. D. 1993. Making Democracy Work: Civic Traditions in Modern Italy. Princeton, NJ: Princeton University Press. Quenette, P.-Y. 1990. Functions of vigilance behaviour in mammals: a review. Acta Œcologica, 11, 801–818. Rapoport, A. 1969. House Form and Culture. Englewood Cliffs, NJ: Prentice-Hall. Rose, L. M. 1994. Benefits and costs of resident males to females in white-faced capuchins (Cebus capucinus). American Journal of Primatology, 32, 235–248. Rose, L. M. & Fedigan, L. M. 1995. Vigilance in white-faced capuchins, Cebus capucinus, in Costa Rica. Animal Behaviour, 49, 63–70. Schiefenh¨ ovel, W. 1997. Universals in interpersonal interactions. In: Nonverbal Communication: Where Nature Meets Culture, ed. U. Segerstr˚ ale & P. Moln´ ar. Mahwah, NJ: Lawrence Erlbaum, pp. 61–79. Shore, B. 1982. Sala ‘Ilua: A Samoan Mystery. New York: Columbia University Press. Smith, W. J. 1977. The Behavior of Communicating: An Ethological Approach. Cambridge, MA: Harvard University Press. Smuts, B. B. 1987. Sexual competition and mate choice. In: Primate Societies, ed. B. B. Smuts, D. L. Cheney, R. M. Seyfarth, R. W. Wrangham & T. T. Struhsaker. Chicago, IL: University of Chicago Press, pp. 385–309. Soffer, O., Adovasio, J. M. & Hyland, D. C. 2000. The ‘Venus’ figurines: textiles, basketry, gender, and status in the Upper Paleolithic. Current Anthropology, 41, 511–537. Steenbeek, R., Piek, R. C., van Buul, M. & van Hooff, J. A. R. A. M. 1999. Vigilance in wild Thomas langurs (Presbytis thomasi): the importance of infanticide risk. Behavioral Ecology and Sociobiology, 45, 137–150. Sugawara, K. 1984. Spatial proximity and bodily contact among the Central Kalahari San. African Study Monographs, 3, 1–43. 1990. Interactional aspects of the body in co-presence: observations on the Central Kalahari San. In: Senri Ethnological Studies 27: ‘Culture Embodied,’ ed. M. Moerman & M. Nomura. Osaka, Japan: National Museum of Ethnology, pp. 79–122. Tjaden, P. & Thoennes, N. 2000. Prevalence and consequences of male-to-female and female-to-male intimate partner violence as measured by the National Violence Against Women Survey. Violence Against Women, 6, 142–161. Trapnell, W. H. 1987. Eavesdropping in Marivaux. Geneve: Librairie Droz. Treves, A. 1998. The influence of group size and neighbors on vigilance in two species of arboreal monkeys. Behaviour, 135, 453–481. 1999. Within-group vigilance in red colobus and redtail monkeys. American Journal of Primatology, 48, 113–126. 2000. Theory and method in studies of vigilance and aggregation. Animal Behaviour, 60, 711–722. Treves, A., Drescher, A. & Ingrisano, N. 2001. Vigilance and aggregation in black howler monkeys (Alouatta pigra). Behavioral Ecology and Sociobiology, 50, 90–95. Turner, T. S. 1980. The social skin. In: Not Work Alone: A Cross-Cultural View of Activities Superfluous to Survival, ed. J. Cherfas & R. Lewin. London: Temple Smith, pp. 112–140.

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

18:41

Humans: control, signalling and intimate experience Tutin, C. E. G. 1979. Mating patterns and reproductive strategies in a community of wild chimpanzees (Pan troglodytes schweinfurthii). Behavioral Ecology and Sociobiology, 6, 29–38. Underwood, R. 1982. Vigilance behaviour in grazing African antelopes. Behaviour, 79, 81–107. van Schaik, C. P., van Noordwijk, M. A., Warsono, B. & Sutriono, E. 1983. Party size and early detection of predators in Sumatran forest primates. Primates, 24, 211–221. Whiten, A. 1993. Social complexity: the roles of primates’ grooming and people’s talking. Behavioral and Brain Sciences, 16, 719. Wilson, P. J. 1988. The Domestication of the Human Species. New Haven, CT: Yale University Press. Wirtz, P. & Wawra, M. 1986. Vigilance and group size in Homo sapiens. Ethology, 71, 283–286.

441

P1: JZZ/... P2: JZZ/... 0521823617c19.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

442

18:41

P1: KPB/... P2: KOD 0521823617agg.xml

CU1917B/McGregor

0 521 582361 7

Part IV I N T E R F A C E S W I T H O T H E R DISCIPLINES

443

April 12, 2005

12:11

P1: KPB/... P2: KOD 0521823617agg.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

444

12:11

P1: KPB/... P2: KOD 0521823617agg.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:11

Introduction

Communication has a history of addressing topics of interest to other disciplines, both in biology and more generally. The interface between disciplines has been long recognized to generate paradigm shifts and the same has been true of interfaces with communication. The interface between communication and neurobiology provides a good example. The discovery that the brain nuclei controlling song production varied in size seasonally (Nottebohm, 1981) was a finding that overturned accepted notions of the stability of brain architecture and triggered studies of evolutionary plasticity in brain structure (e.g. Sherry, 1998). An important question then is whether the communication network approach enhances communication’s interest to, and interfaces with, other disciplines. This section shows that the answer is an emphatic yes, it does. In part, this is shown by the wide range of topics addressed: from perception and physiology, through aspects of cognition to the evolution of altruism. However, it is in the details of the chapters that the value of the approach becomes apparent, as does an enthusiasm about the further research possibilities.

Perception The extent of a communication network is often an important issue and is discussed by several chapters in this book. Network size is related to the distance at which signals can be received and this distance is influenced by several factors. These factors include the distorting and attenuating effects of the environment through which the signal travels and the level of interference from the signals of others. A key factor that is often overlooked is the sensory abilities of the receiver. Such abilities can be extraordinary; for example, some bird species respond to a Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

445

P1: KPB/... P2: KOD 0521823617agg.xml

446

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:11

Part IV signal even when that signal is embedded in noise that is louder than the signal. Ulli Langemann and Georg Klump (Ch. 20) discuss such perceptual abilities and how they relate to aspects of signal structure and transmission in the acoustic modality. Their chapter, therefore, covers the interface between communication networks, psychophysics and physics.

Endocrinology Hormones play an established role in determining when animals communicate, by controlling annual and circadian rhythms. They are also known to be involved in mediating the response to signals; for example, raising the level of oestradiol can induce female songbirds to perform copulation solicitation displays in response to song in the absence of the male singer (Searcy & Yasukawa, 1996). In Ch. 21, Rui Oliveira uses examples mainly from bony fishes to show that androgens can be affected by the social environment as well as modulating behaviours that partly create the social environment. Establishing this reciprocal link between hormones and behaviour (comparable to that for stress hormones) has obvious consequences for both endocrinology and communication behaviour.

Cooperation and altruism Altruism between unrelated individuals (that can include cooperation) has long been considered an evolutionary puzzle. Recent mathematical models and experiments with humans have shown that altruism can evolve through an increase in the altruist’s ‘prestige’ or ‘image’ in the eyes of others (e.g. Wedekind & Milinski, 2000). Redouan Bshary and Arun D’Souza point out in Ch. 22 that these recent advances in cooperation theory are a specific instance of a communication network because others not directly involved in the altruistic interaction must observe it. They then investigate the evolution and maintenance of altruistic behaviour, tactical deception and spiteful behaviour using data gathered in the field from interactions between coral reef cleaner fish and their clients.

Semiochemicals Information is obviously a key concept in communication, but as Brian Wisenden and Norm Stacey make clear in Ch. 23, communication is a subset of information. Aquatic animals can obtain information to guide reproductive and predator-avoidance behaviour from chemicals released as a by-product of other processes. Information in such semiochemicals can have striking effects; for example chemicals released by predators attacking or digesting their prey can change

P1: KPB/... P2: KOD 0521823617agg.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:11

Interfaces with other disciplines the behaviour, life history and morphology of potential prey (e.g. Brancelj et al., 1996). After considering the use of chemical information by fishes in predator– prey and sexual interactions, Wisenden and Stacey set out the case for communication networks as a subset of information networks and the potential evolutionary routes for the origins of network behaviours such as eavesdropping. Their suggestions include the possibility that a process analogous to eavesdropping can precede the origin of communication, rather than such information gathering following the development of communication interactions.

Cognition The role of communication in cognition has ensured a lively interface with the cognitive sciences, often concerning the extent to which non-human communication can be considered a language (e.g. Hauser, 1996). Irene Pepperberg considers the cognitive abilities of birds in Ch. 24, particularly the issue of transitive inference, using information from two different approaches. The first approach uses communication to explore the cognitive abilities of parrots, with human speech being used as the tool in much the same way as it is in explorations of human cognitive abilities. The second approach uses the results of field experiments investigating social eavesdropping to indicate the cognitive abilities of territorial songbirds. Several chapters in other sections of this book touch upon the cognitive abilities of animals communicating in networks (e.g. the extent to which individual identification is a prerequisite of eavesdropping in Ch. 16). Dorothy Cheney and Robert Seyfarth expand this theme in Ch. 25. They review the evidence for eavesdropping in primates, concentrating on species living in large, permanent social groups, often with complex social relationships. They then suggest a framework for assessing the occurrence in other animal groups of social intelligence (i.e. mechanisms such as transitive inference used to gather information relevant to social interrelationships). This framework, together with the information presented in other chapters of this book, provides an opportunity for a taxonomically wide-ranging comparative approach to the issue of social intelligence.

Mathematical models Mathematical models have provided insights in many areas of biology, including communication. However, most models of communication have not dealt with networks, at least partly because of the difficulty in applying tractable analytical models to networks. In Ch. 26, Andrew Terry and Rob Lachlan describe models that capture two important aspects of networks by being spatially explicit

447

P1: KPB/... P2: KOD 0521823617agg.xml

448

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:11

Part IV and based on individual behaviour. These simulation models of anuran acoustic choruses and eavesdropping strategies generate different results from more traditional models applied to the same questions. Incorporating strategic communication decisions into such spatially realistic individually based models seems an approach that is likely to generate the kind of testable predictions about communication in networks that are needed to stimulate further research.

Applied aspects Communication is central to animals’ lives and as such can be used to modify their behaviour for human ends, such as in pest control, animal welfare and conservation. A chapter jointly written with Tom Peake was planned for this book; it would have explored the possible interfaces between communication networks and aspects of applied biology. However, the relative newness of the network perspective means that there are, as yet, few concrete examples. We decided, therefore, to outline here some of the interfaces that we think are promising and to draw attention to chapters elsewhere in the book that have mentioned applications. One way to judge the welfare of animals is to assess the extent to which animals in captivity, including those on public display in zoos and aquaria, are able to display the full range of behaviours shown by free-ranging animals under natural conditions. Many of the chapters in this book have argued that communication in a network is such a natural condition and, therefore, the ability to communicate as part of a network could be regarded as a feature of adequate captive provision. Adverse effects on breeding performance have been noted when group-living species are kept in small groups, and communication networks that are much smaller than occur in the wild could underlie this effect. Attempts to increase apparent group size (e.g. use of mirrors with captive flamingos (Whitfield, 2002)) have met with mixed success, perhaps because the manipulations did not adequately mimic communication networks. It may be easier to create an apparent communication network for species that are widely spaced and possibly territorial, because at long range it is only signals that are detected and signal playback is straightforward, at least with acoustic signals. Tom Peake has suggested that interactive acoustic signals could be provided as a type of environmental enrichment in communication for captive animals. For example, zoo visitors could interact with vocal species such as gibbons Hylobates spp. via playback from remotely sited loudspeakers triggered from a control panel in the cage’s viewing area (obviously there would have to be safeguards to ensure that the nature and extent of interaction did not exceed natural levels). Communication in a network may be equally difficult if a population is held at an unnaturally high density for production reasons (e.g. fish farms),

P1: KPB/... P2: KOD 0521823617agg.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:11

Interfaces with other disciplines because the network is overloaded. From the discussions above, it can be expected that removing the ability to signal will have adverse effects on welfare (e.g. removal of rodent scent marks by cage cleaning (Gray & Hurst, 1995)). Signal removal can also have additional network-wide effects. Removing the major chela of male fiddler crabs Uca tangeri clearly prevents individual males from signalling visually in their usual manner. In addition, it may skew the sex ratio apparent to males and females because males without a major chela appear female to both sexes, and the response to the apparent sex ratio could accelerate population declines (Oliveira et al., 2000). There are several established applications of communication in conservation, such as identifying individuals from features of their vocalizations (e.g. McGregor et al., 2000). The network perspective emphasizes the possibility that anthropogenic noise could have adverse effects by disrupting or restricting the size of acoustic communication networks (e.g. McGregor & Dabelsteen, 1996). Vincent Janik considers this applied aspect in more detail in relation to marine mammals in Ch. 18. In the terrestrial environment, road noise may similarly restrict the acoustic communication networks of other taxa such as songbirds. Habitat fragmentation is another way of disrupting songbird communication networks and it is discussed by Ken Otter and Laurene Ratcliffe in Ch. 7. In Ch. 8, Alexandra Lang and colleagues discuss how the need to use signals with limited range and detectability to predators may have the side effect of reducing the effective population size of katydids, with associated increased susceptibility to random extinction processes. We think that these brief examples illustrate that communication networks can have relevance to applied biology and that often the implications are not straightforward. We suggest that those researching communication networks have an obligation to explore the applications of their findings. Arguably the best way to make applied biologists aware of the relevance of network research is to make suggestions on how best to modify current practice to incorporate new findings.

Future directions The chapters in this section clearly show how several areas of research interface with communication networks. An obvious question is whether this will also be true for interfaces with disciplines that are not represented in this section. A number of chapters (e.g. Chs. 12, 14 and 26) mention the possibility of fruitful links with ecology, or more specifically spatial ecology, suggesting that this would seem to be a good interface to explore. Also, given that we humans consider ourselves to be supreme communicators and often do so in a network environment, it is possible that many aspects of social psychology and sociology could interface

449

P1: KPB/... P2: KOD 0521823617agg.xml

450

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:11

Part IV fruitfully with a communication network approach. These brief considerations suggest that the answer to the question posed above is that interfaces between communication networks and many other disciplines can be sources of insight and inspiration in the future.

References Brancelj, A., Celhar, T. & Sisko, M. 1996. Four different head shapes in Daphnia hyalina (Leydig) induced by the presence of larvae of Chaoborus flavicans (Meigen). Hydrobiologia, 339, 37–45. Gray, S. & Hurst, J. L. 1995. The effects of cage cleaning on aggression within groups of male laboratory mice. Animal Behaviour, 49, 821–826. Hauser, M. D. 1996. The Evolution of Communication. Cambridge, MA: MIT Press. McGregor, P. K. & Dabelsteen, T. 1996. Communication networks. In Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca, NY: Cornell University Press, pp. 409–425. McGregor, P. K., Peake, T. M. & Gilbert, G. 2000. Communication behaviour and conservation. In: Behaviour and Conservation, ed. L. M. Gosling & W. J. Sutherland. Cambridge, UK: Cambridge University Press, pp. 261–280. Nottebohm, F. 1981. A brain for all seasons: cyclical anatomical changes in song control nuclei of the canary brain. Science, 214, 1368–1370. Oliveira, R. F., Machado, J. L., Jord˜ ao, J. M. et al. 2000. Human exploitation of male fiddler crab claws: behavioural consequences and implications for conservation. Animal Conservation, 3, 1–5. Searcy, W. A. & Yasukawa, K. 1996. Song and female choice. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca, NY: Cornell University Press, pp. 454–473. Sherry, D. F. 1998. The ecology and neurobiology of spatial memory. In: Cognitive Ecology: The Evolutionary Ecology of Information Processing and Decision Making, ed. R. Dukas. Chicago, IL: Chicago University Press, pp. 261–296. Wedekind, C. & Milinski, M. 2000. Cooperation through image scoring in humans. Science, 288, 850–852. Whitfield, J. 2002. Mirrors to help birds mate. Nature Science Update, 19 March: http://www.nature.com/news/2002/020318/full/020318-2.html.

P1: KPB/... P2: KOD 0521823617c20.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

20

Perception and acoustic communication networks ulrike l angemann & georg m. klump University of Oldenburg, Germany

Introduction Traditionally, the analysis of acoustic communication has been based on a model system composed of a sender, the transmission channel and a receiver (Shannon & Weaver, 1949). Since the early 1990s, this view has been extended to communication networks, in which several signallers and receivers are involved (e.g. McGregor & Peake, 2000). Two general approaches have been adopted in order to investigate communication behaviour. First, measurements of physical modifications to the signal during transmission (e.g. Wiley & Richards, 1978; Dabelsteen et al., 1993; Holland et al., 1998) have been used to assess the feasibility of communication (e.g. estimating maximum communication distances) or to evaluate which features of signals might be adaptive in a certain context. Second, playback studies have been used to conclude which features may be of importance for signal discrimination: different behavioural responses can be elicited by playback of signals that have been modified by physical properties of the environment or by the experimenter. Often the physical properties of signals are manipulated in ways that are informed by studies of signal transmission in the animal’s environment. However, behavioural responses can be understood more fully if the animal’s perceptual abilities are taken into account (Wiley & Richards, 1982; Klump, 1996). Perception includes the transduction process by the animal’s sensory organs and the subsequent processing by the nervous system. However, perception can only be inferred indirectly from the animal’s responses. An animal’s failure to respond differentially to playback can either mean that the animal was not motivated to discriminate within the experimental context (e.g. because the modified signal deviated too far from species-specific signals) or that the animal’s auditory Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

451

P1: KPB/... P2: KOD 0521823617c20.xml

452

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

U. Langemann & G. M. Klump system could not ‘resolve’ modifications of the signal and, therefore, they were not perceived. In the latter case, perceptual constraints render a behavioural response impossible. If more than one signal parameter was modified in playback experiments, the results might be even more difficult to interpret. Studies in the laboratory focusing on perceptual mechanisms allow us to control for motivation of an animal, help to conclude which modifications can be exploited by the animal and make it possible to determine the perceptual resolution of the animal’s sensory system. In this chapter, we will explain how the current knowledge on perceptual mechanisms offers a better understanding of animal communication, especially in the context of a communication network. For example, new results from perceptual studies now give us a more complete and accurate picture of how animals parse signals from noise and from different sources in acoustic scenes (e.g. Feng & Ratnam, 2000; Hulse, 2002). We will demonstrate this by showing how knowledge on perceptual masking can aid in explaining signal assessment by the animal. An understanding of how animals segregate several sources of signals in a communication network requires knowledge of how accurately signals are localized and how the spatial arrangement of sources affects masking (e.g. Klump, 2000). Finally, the behaviour of an animal in a communication network often requires an ability to range a signal, that is, assess the distance of a sound source (e.g. Naguib & Wiley, 2001). Explaining results from playback studies of ranging requires an understanding of the perception of degradation. Providing a comprehensive review of all the topics mentioned here is beyond the scope of this chapter. Instead, we will present specific examples from the animal behaviour literature and discuss them in light of knowledge of the physiology of perception. Detection and recognition Spectral aspects of masking and signal detection

Communication in any context requires signal detection. This would hardly be a problem in a silent world. However, the environment is noisy, perhaps particularly so in a communication network. Therefore, signal detection needs to be considered in relation to the level of the background noise. Environmental noise originates from biotic sources such as calling insects (e.g. grasshoppers and cicadas (Waser & Waser, 1977; Ryan & Brenowitz, 1985)) and calling frogs (Wollerman & Wiley, 2002) or singing birds (especially during the dawn chorus, e.g. Staicer et al. (1996)). Biotic environmental noise constitutes an especially severe problem in large assemblies of individuals of the same species (e.g. roosts, breeding colonies, choruses) since the masking noise matches the frequency spectrum of the signal (as the signals of conspecifics constitute most of the masking noise

P1: KPB/... P2: KOD 0521823617c20.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

Perception and acoustic communication networks experienced by an individual). There are other biotic sources that may contribute to background noise. For example, the rustling of the foliage and the movement of twigs and branches in a deciduous wood provide a substantial level of background noise (e.g. Klump, 1996) that increases with wind speed (F´egeant, 1999). A similar increase in background noise with wind speed can be observed in open grassland habitats. For example, wind with a moderate speed of 5 m/s produces sound pressure levels of more than 60 dB in the one-third-octave band at 20 Hz for at least 95% of the time (Boersma, 1997). Much lower levels of background noise are observed at frequencies above 500 Hz. Coastal environments and running waters may produce a substantial amount of non-biotic noise (Dubois & Martens, 1984; Douglas & Conner, 1999). The difference between the frequency spectrum of the signal and the background noise and their relative amplitudes determine the amount of masking. The amplitude of sounds can either be specified in terms of their overall soundpressure level or in terms of the amplitude contained in the individual components. Commonly, the components are resolved into frequency bands 1 Hz wide and the sound pressure level (which is a measure of the amplitude relative to the standard reference pressure of 2 × 10−5 Pa) in each 1 Hz band is determined. This amplitude measure is conventionally called the spectral density or the spectrum level (e.g. Moore, 2003). The difference between the signal amplitude and the amplitude of the background can be described by the signal-to-noise (S/N) ratio. The S/N ratio can be either expressed as the difference between the overall signal level and the overall level of the noise or as the difference between the signal spectral density and the spectral density of the noise. The S/N ratio allows us to estimate whether a signal can be detected or not (see Box 20.1). The same rules for signal detection apply for single receivers as well as for individuals in a communication network. Each ‘node’ in a network (a receiver sitting at a different place), however, might experience quite different S/N ratios for the same signal. Behavioural experiments in the laboratory can determine how random background noise affects absolute auditory sensitivity for tonal signals or signals with a distinctive peak in the spectrum. The value that denotes the shift in auditory sensitivity when random (wideband) noise is present is called the critical masking ratio or critical ratio (CR). The CR is simply the S/N ratio at detection threshold in random wideband noise expressed as the difference between the level of a tonal signal (which is identical to its spectral density) and the spectral density of the noise (N 0 ). The CR is usually independent of the level of random background noise. However, the CR is frequency dependent, increasing at about 3 dB per octave (overview in Fay, 1988). The CR also provides a rough estimate of the bandwidth of auditory analysis filters (CR filter bandwidth in Hz is given by 10CR/10 ; e.g. Yost, 1994). There are numerous

453

P1: KPB/... P2: KOD 0521823617c20.xml

454

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

U. Langemann & G. M. Klump Box 20.1 Determining the signal-to-noise (S/N) ratio for signal detection Calculating the S/N ratio for signal detection in animal studies is relatively straightforward in the laboratory environment with suitable equipment and all parameters well controlled. Field measurements are more difficult. Nevertheless, as shown here, it is possible to get an estimate of the S/N ratio for signal detection in the natural environment. Which equipment to use Preferably, a sound-level meter should be used in the field to determine signal level directly. Often it is possible to record sounds through the microphone of the sound-level meter (e.g. on a DAT recorder or directly with a note-book computer) for subsequent analysis, but any microphone and recording equipment that allows signals to be recorded with an accurately defined gain is suitable. The directional characteristics of the microphone should be adapted to the question of interest. For example, an omnidirectional microphone is the best choice when the general level of background noise is being measured. If the goal is to measure a signal originating from a specific source, it is recommended to use a directional microphone and approach the source as close as possible. The frequency-transfer function of the microphone should be as flat as possible to avoid a bias in the later analysis. Recording a calibration signal of known sound-pressure level with the signal of interest makes it subsequently possible to compute the absolute power spectra of the recorded signals (the calibration signal can be recorded at 1 m from the microphone, at the location from which the animal was recorded or by using a calibrator placed on the microphone). Even inexpensive sound-level meters can be used for accurate field measurements if they are calibrated against a high-quality instrument in the laboratory. Spectral analysis can be carried out with specialized spectrum analyser hardware or (if the signals are digitized and stored on a computer) with suitable software, e.g. freeware. Computing signal-to-noise ratios First, one needs to decide which of the two common measures of the S/N ratio should be used: comparing overall sound-pressure levels or comparing the sound-pressure levels with reference to spectral density (i.e. compute the ratio between signal and noise spectrum level N 0 ; for further details, see text). The frequency spectrum and the sound-pressure level of

P1: KPB/... P2: KOD 0521823617c20.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

Perception and acoustic communication networks both the signal of interest and the background noise have to be known in order to determine the S/N ratio correctly. Step 1: measuring sound-pressure levels of signal and background noise

It is important that the sound-pressure level of the signal of interest is at least 10 dB above the sound pressure level of other sound making up the background noise. If the difference is less than 10 dB, the sound-pressure level of the signal cannot be measured independently from the sound-pressure level of the background noise and, therefore, a S/N ratio cannot be computed accurately (the ratio would then be the level of signal plus noise divided by the level of the noise). Strategies to achieve the 10 dB minimum difference include getting close enough to the signal source, getting as far away as possible from sources of noise or using a highly directional microphone. The distance between the microphone and the sound source(s) should be reported. This allows an estimate of the signal level at a specified distance from the source (e.g. by using the rule of thumb that the signal level is reduced by 6 dB for every doubling of the distance and, if more accurate estimates are required, the estimates should include effects of excess attenuation). Sound-level meters frequently offer at least two types of filter setting: the A and C settings. The A filter has a low-frequency cut-off of about 800 Hz, a high-frequency cut-off of about 9 kHz and emphasizes the intermediate frequencies in this range. The C filter has a low-frequency cut-off of approximately 30 Hz, a high-frequency cut-off of approximately 8 kHz and has a flat unbiased frequency response. To be sure that an appropriate filter is used, the signals of interest must fall within the frequency range of the filter. For example, measuring ambient background noise over a wide frequency range is only possible with a C filter setting. The integration time constant of the sound-level meter ideally should match the integration time of the auditory system of the study species or at least the duration of its signals. The ‘fast’ integration time is suitable for measuring the sound-pressure level of animal signals of 125 milliseconds or longer duration. Using a very long integration time (1 second at the ‘slow’ setting of the sound-level meter) will underestimate the level of signals that are composed of brief components. If shorter signals than 125 milliseconds need to be analysed, the ‘impulse’ or ‘peak’ settings are more suitable. These provide integration times of 35 and 0.05 milliseconds, respectively, for fast-rising signal levels and a very slow decay. The integration times used to determine root mean square sound pressure with a sound-level meter allow only approximations of the real sound pressure (that can be calculated from

455

P1: KPB/... P2: KOD 0521823617c20.xml

456

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

U. Langemann & G. M. Klump the calibrated digitized signal) or the perceived sound pressure (for which one should apply the integration time of the animal’s auditory system). When digitizing the recordings for further analysis with a computer, the sampling rate must be at least twice the highest frequency of the signals to avoid serious sampling errors. Step 2: calculating signal-to-noise ratios

If the signal and the background noise have the same frequency spectra (e.g. frog calls in a dense chorus of conspecifics) the S/N ratio is simply the difference between the overall signal level and the overall level of the background noise, both measured in decibels. If the signal and the background noise have different bandwidths and thus different frequency spectra (e.g. a tonal bird vocalization in a wideband background noise), a S/N ratio based on measurements of the spectrum level is easier to interpret. The spectrum level can be calculated from the overall level by subtracting the bandwidth (in decibels) of the frequency filter used in the measurement from the overall sound-pressure level measured through this filter, i.e. spectrum level (dB) = overall level (dB) − 10 × log10 (bandwidth) If the signal has a smaller bandwidth than the frequency filter through which its sound-pressure level was determined, the bandwidth of the signal must be used instead of the bandwidth of the filter. Then the S/N ratio is calculated as the difference between the spectrum level of the signal and of the background noise measured in decibels. It must be noted that this engineering-type measure of the S/N ratio provides only an approximation of stimulus characteristics relevant for perception. The filters relevant for the perception of the animal are the auditory analysis filters. Furthermore, the measure of the S/N ratio as calculated here does not incorporate temporal aspects of the structure of the signal and background noise. The physiology of the auditory system will determine how temporal aspects affect the perceived S/N ratio. If one wants to know more accurately how the acoustic environment is perceived by an animal, the measurements should be interpreted using a model of the physiology of the animal’s auditory system. studies determining the CR values at various frequencies of the hearing range in fish, amphibians, birds and mammals (summarized by Fay, 1988). The average CR in birds is about 23 dB at 1 kHz, 24.5 dB at 2 kHz, 27.5 dB at 4 kHz and 36 dB at 8 kHz (Klump, 1996). Typical CR values in mammals are 22.5 dB at 1 kHz, 25.5 dB at 2 kHz, 29.5 dB at 4 kHz and 32 dB at 8 kHz (average for cat, rat, chinchilla and human, respectively (Fay, 1988)). The average slope of 3 dB per octave that is often described probably does not hold for frequencies below 500 Hz and the CR will decrease much

P1: KPB/... P2: KOD 0521823617c20.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

Perception and acoustic communication networks less for such frequencies (e.g. Moore & Glasberg, 1983). There are some exceptions from the usual pattern of an increasing values of CR with increasing frequency. In mammals, the greater horseshoe bat Rhinolophus ferrumequinum has the lowest CR values at the frequencies of their ultrasonic echolocation calls (Long, 1977). In birds, great tits Parus major show relatively little change in CR with frequency. At high frequencies, CR values for great tits are much lower than those of other bird species (e.g. 25.9 dB at 8 kHz). This may be an adaptation that makes the highpitched communication sounds of great tits much less susceptible to masking by environmental noise in their deciduous forest habitat (Klump & Curio, 1983; Langemann et al., 1998). Signal-detection thresholds in background noise, called masked thresholds, thus depend on the level of background noise and on the CR. Estimates of masked auditory thresholds (MAT) are commonly calculated by adding the CR to the spectrum level of the noise (N 0 + CR = MAT). Knowing both CR and the level of the background noise allows us to estimate approximately how background noise in a specific communication setting will influence the auditory sensitivity of animals. We will use an example to demonstrate how such calculations can estimate the distance over which communication is possible: that is, the extent of the communication network (see also Box 20.1). The detection distance for a typical great tit song element with a spectral peak frequency of 2 kHz and signal amplitude of 90 dB can be estimated to be 331 m (Table 20.1). This estimate takes into account the great tit’s absolute threshold at 2 kHz (6.5 dB), spherical spreading of the song (6 dB per doubling of distance for all frequencies) and the habitat-dependent signal attenuation (excess attenuation, e.g. Marten & Marler, 1977; Dabelsteen et al., 1993). However, the main acoustic energy of background noise in a deciduous forest occurs at lower frequencies (Klump, 1996; F´egeant, 1999) and the 2 kHz signal will be masked by background noise of 10 dB, and the perceptual threshold of a great tit receiver in noise (masked auditory threshold) will be much worse than 6.5 dB: in fact the masked threshold of a 2 kHz signal would be 35.6 dB (N 0 + CR, i.e. 10 + 25.6 dB; Table 20.1). Hence, a great tit can only perceive song elements of 2 kHz as long as the sound level does not drop below the great tit’s masked threshold, giving a maximum detection distance of 124 m. At this distance, the sound level of the element equals the value of the masked threshold (35.6 dB). In contrast, a ‘seeet’ alarm call indicating the presence of an aerial predator (peak frequency around 8 kHz, mean amplitude 60.1 dB) experiences much less masking by the forest background noise. At 8 kHz, the background noise spectrum level is −5.2 dB and since great tits’ CR at 8 kHz is similar to the CR at 2 kHz, the masked threshold is very close to great tits’ absolute threshold. In consequence, the perception of this alarm call is mainly limited by great tits’ absolute auditory sensitivity. Infrasound communication in elephants is assumed to be a communication network extending over several kilometres (Larom et al., 1997; Ch. 17).

457

P1: KPB/... P2: KOD 0521823617c20.xml

458

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

U. Langemann & G. M. Klump Table 20.1. An example of estimates of masked thresholds (i.e. the signal-detection threshold in background noise) and perceptual distances (i.e. the maximum distance where behavioural responses would be expected) of single great tit song elements with spectral peak frequencies between 2 and 8 kHz or the aerial predator ‘seeet’ alarm call (of approximately 8 kHz) Signal frequency (kHz) 2 Spherical spreading (dB/dd)

4

6.3

8

‘Seeet’ call

6

6

6

6

6

Excess attenuation (dB/100 m)

10.0

13.7

18.0

21.1

21.1

Source level (dB)

90

90

90

90

60.1

12.8

18.1

18.1

Absolute hearing threshold (dB) How far in quiet (m)

6.5 331

9.1 242

179

137

56

Noise level N0 (dB)

10.0

4.3

−1.7

−5.2

−5.2

Critical ratio (dB)

25.6

23.8

25.9

25.9

25.9

Masked auditory threshold (dB)

35.6

28.1

24.2

20.7

20.7

Detection distance in noise (m)

124

Recognition threshold (dB) Perceptual distance in noise (m)

38.6 107

138 31.1 124

130 27.2 118

128 23.7 117

49 23.7 41

dd, doubling of distance; CR, critical ratio; Signal detection is a function of physical signal properties (signal frequency, source level), of the environment (background noise level N 0 , intensity loss from excess attenuation), of physics (intensity loss from spherical spreading) and of physiological constraints set by the auditory system. The critical ratio denotes the shift in auditory sensitivity from the absolute hearing threshold (in quiet) to the masked auditory threshold when random wideband noise is present. For calculating perceptual distances, random noise is used as an approximation of the ambient background noise. Note that the background noise level N0 is expressed with reference to spectral density of the noise in order to calculate masked thresholds (N 0 + CR) Spectral density is the soundpressure level of each 1 Hz wide frequency component relative to the standard reference pressure of 2 × 10−5 Pa). Detection is only the first step in perception; the second step is recognition (or discrimination), which is also influenced by noise. Recognition thresholds may be estimated to be an additional 3 dB higher in signal-to-noise ratio than detection thresholds. Values for thresholds in great tits are from Langemann et al. (1998), noise level from Klump (1996), source level of ‘seeet’ calls from Klump & Shalter (1984). An excess attenuation of 10 dB/100 m was assumed for 2 kHz; the excess attenuation above 2 kHz was increased by 1.85 dB/100 m for every additional 1 kHz. The amount (A) to which the original sound pressure level of the source (S) can drop to be just detectable in quiet (i.e. to the absolute threshold AAT ), in noise (the masked auditory threshold AMAT ) or recognized in noise (the recognition threshold ART ) is one variable used for estimating the perceptual distance (m). The other components include the amplitude decrease from spherical spreading (20 log10 m) and the decrease from excess attenuation (EA/100 m). Solving the following equation for m yields the perceptual distance (see also Marten & Marler, 1977): A(dB) = S(dB) − 20 log10 m − m(E A/100).

P1: KPB/... P2: KOD 0521823617c20.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

Perception and acoustic communication networks Communication distances of nearly 10 km are obtained by assuming that absolute thresholds are limiting and that wind speeds are low at night. However, during the day, wind speed is much higher and it is likely that wind-induced noise provides sufficient masking to reduce communication distances. Using measurements of the sound-pressure level of wind-induced noise in grassland (Boersma, 1997) and assuming a CR of 10 dB (low-frequency CR values can only be extrapolated from studies in humans, e.g. Moore & Glasberg, 1983) gives a masked threshold of 73 dB. This is much higher than the absolute threshold of 50 dB that was used in the calculation by Larom et al. (1997) and suggests that infrasound communication networks may be less extensive than previously thought. Some caveats must be borne in mind when estimating detection distances or transmission distances from CR values. First, CR is measured by presenting a narrowband signal in wideband masking noise. If the frequency spectrum of the signal and the noise are rather similar (e.g. detection of an individual calling frog in the masking noise provided by a chorus of thousands of conspecific frogs), then the masked threshold calculated from the CR will be overestimated. With signals of similar frequency spectra, the S/N ratio is probably close to, or even below, 0 dB, i.e. signals can be detected when their level is equal to, or even below, the level of the background noise. In this case, the task resembles an increment detection in overall sound amplitude when the signal is added to the background noise (e.g. Miller, 1947). Second, the CR may not provide a good estimate of masked thresholds when the temporal structure of the background noise has very pronounced slow envelope fluctuations (see below). In this case, masked thresholds may be up to 20 dB more sensitive than would be expected from the CR. Third, it should be remembered that communication signals are often broadcast repeatedly, or at least some signal elements are repeated, whereas in laboratory studies the test signals for detection will often be presented only once. Detection sensitivity is known to improve by the square root of the number of independent observations (e.g. Swets et al., 1950), suggesting that repetitive signals may be detected more readily. Separating sounds by exploiting temporal patterns

Environmental background noise will usually not resemble the random noise with a steady-state envelope that is often employed in the laboratory as a masker. Animals in communication networks appear to be adapted to exploit temporal envelope fluctuations in the background noise that are typical of the natural environment (Klump, 1996; Nelken et al., 1999). Signallers have often been observed to call or sing during periods of reduced amplitude of the background noise. For example, many frog species call when the nearest signalling neighbours are silent (Klump & Gerhardt, 1992; Ch. 13). Some frog species go even further and

459

P1: KPB/... P2: KOD 0521823617c20.xml

460

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

U. Langemann & G. M. Klump (a)

(b)

(c)

Time (s)

Fig. 20.1. Waveforms illustrating different noise envelopes. (a) Unmodulated wideband random noise exhibiting a Gaussian distribution of amplitude values. This type of noise exhibits little variation in the temporal envelope. Unmodulated wideband noise is normally used in determining critical masking ratios. (b) Coherently amplitude modulated noise. Note the slow envelope fluctuations. This noise was synthesized by multiplying random wideband noise by a low-passed noise with a bandwidth of 12.5 Hz. It has the same overall bandwidth as the unmodulated noise in (a). This type of noise resembles more closely the structure of environmental noise. (c) A dawn chorus recorded in a European deciduous forest. It has a waveform with pronounced slow amplitude fluctuations that are more similar to the fluctuations in the envelope of coherently modulated noise (b) than of unmodulated noise (a).

will use call timing strategies to mask their neighbours’ calling (e.g. Gerhardt & Huber, 2002). Receivers may also benefit from temporal patterns in signals and in background noise because the separation of sounds from different sources is improved if their amplitude patterns differ considerably. The separation of sounds originating from different sources into different ‘auditory streams’ is also known as sound segregation (Bregman, 1990). The ‘unmasking’ effect associated with sound segregation is well documented in laboratory studies with humans and other animals (e.g. Moore, 1990; Klump & Langemann, 1995; Nelken et al., 1999; Klump & Nieder, 2001; Pressnitzer et al., 2001). It should be noted that unmasking

P1: KPB/... P2: KOD 0521823617c20.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

Perception and acoustic communication networks Table 20.2. Masking release observed in the laboratory for detection of pure tone signalsa Species

Masking

Amplitude

release (dB)

factor

Study

Human, Homo sapiens

12

4

e.g. Schooneveldt & Moore, 1989

European starling, Sturnus

12

4

Klump & Langemann, 1995

Gerbil, Meriones unguiculatus

17

7

Klump et al., 2001

Chinchilla, Chinchilla laniger

6

2

Niemiec, 2001; A. J. Niemieac,

Cat, Felis catus

5

2

Budelis et al., 2002; b B. J. May,

vulgaris

personal communication personal communication a Masking

release is the masked threshold in wide-band coherently amplitude-modulated back-

ground noise compared with the masked threshold in unmodulated background noise of the same overall signal energy and bandwidth. The ‘gain’ in signal detection can be expressed either in decibels or as an amplitude factor (10dB/20 ). Similar masker envelopes were used in all species, i.e. the noise bands had dominant frequencies of envelope fluctuations below 50 Hz. b Taking

into account only the masking release across auditory analysis filters.

does not mean that masking is absent, rather it means that a partial release from masking can be observed. For example, this unmasking effect is shown by the ability of European starlings Sturnus vulgaris to detect a tone (i.e. a signal of a particular frequency) in noise (e.g. Klump & Langemann, 1995). Their tone-detection threshold in the type of wideband random noise commonly used in studies of the CR (Fig. 20.1a) is up to 20 dB worse than in noise of the same overall signal energy and bandwidth that has been coherently amplitude modulated (Fig. 20.1b) to resemble more closely typical environmental noise (e.g. bird song dawn chorus; Fig. 20.1c). The amount of masking release resulting from the modulation of noise depends on the bandwidth of the masking noise: masking release decreases with decreasing bandwidth of the masking noise. For bandwidths of the size of an auditory analysis filter (about 10–20% of the centre frequency), the masking release is reduced. For example, in the starling, the masking release is reduced to about 5 dB if 2 kHz signals have to be detected in 200 Hz wide noise centred at the signal frequency (Klump et al., 1998). Slow rates of envelope fluctuation result in a larger masking release than fast rates of fluctuation (Klump & Langemann, 1995). Masking release in amplitude-modulated background noise has been observed in all species studied so far, although to a different extent (Table 20.2). In humans, an unmasking effect is also observed if the signal and the masking background noise are both amplitude-modulated noise bands (no other species has been tested with

461

P1: KPB/... P2: KOD 0521823617c20.xml

462

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

U. Langemann & G. M. Klump this paradigm). If the noise-band signal has a pattern of envelope fluctuation that is different from the masking noise bands, the detection of the noise band signal is improved in comparison with stimuli that have similar correlated envelope fluctuations in both signal and masking bands. For signals with durations of more than 200 milliseconds, detection differences of about 8 dB have been observed (e.g. McFadden & Wright, 1990). The unmasking effects described in the previous paragraph are likely to be relevant in animal communication systems (e.g. Klump, 1996). Signals that are broadcast from a sender will often be amplitude modulated during transmission (Richards & Wiley, 1980). The same signal travelling along different paths will experience different modulation patterns since the modulation results from local turbulence in the atmosphere (arising from temperature gradients or wind). This applies also for signals from different sources that travel along different paths to a receiver. Thus, receivers should be able to exploit the different modulation patterns imposed on signals by the natural environment to gain an advantage in signal detection. In addition, signallers themselves create amplitude modulation patterns when broadcasting calls or songs. For example, king penguins Aptenodytes patagonicus appear to be able to utilize amplitude modulations in their calls to improve their sensitivity when searching for their mate or chick in an assembly of hundreds of individuals (e.g. Aubin & Jouventin, 1998). The contact calls of the emperor penguin Aptenodytes forsteri and of the king penguin exhibit a syllable structure with pronounced amplitude modulations, and their two-voice mode of call generation creates amplitude beats (Aubin et al., 2000; Lengagne et al., 2001). Both species of penguin thus produce calls with a distinctive envelope-modulation pattern that facilitates unmasking effects. The observation by Aubin & Jouventin (1998) that king penguin chicks are able to detect their parents’ calls within the colony background noise with a S/N ratio of about −6 dB can be adequately explained by masking release in amplitude-modulated background noise. Spatial release from masking

Signal detection also depends on the spatial arrangement of the sound sources. If the sources of signal and masking noise are well separated, such as by territorial songbirds, a considerable masking release may be observed. Hine et al. (1994) measured the detection thresholds of ferrets Mustela putorius for 500 Hz tones masked by narrowband noise. Signal and noise were presented either from the same direction (+90◦ , i.e. from the subject’s right side) or the signal was presented from −90◦ and the noise bilaterally from +90◦ and −90◦ . Signal detection was improved by about 10 dB in the bilateral case. Signal detection did not improve in animals that were only allowed monaural listening (Hine et al., 1994), indicating

P1: KPB/... P2: KOD 0521823617c20.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

Perception and acoustic communication networks that the release from masking was a result of binaural processing of signal and noise. Dent and colleagues (1997) replicated the experiment with budgerigars Melopsittacus undulatus using wideband noise as the masker and tone signals of different frequencies. The amount of masking was, on average, 7.5 dB less for bilaterally presented masking noise versus unilateral masking noise presented from the same direction as the tone signal (Dent et al., 1997). In an additional experiment, the authors demonstrated that the directional characteristics of the birds’ auditory system are sufficient to explain the amount of unmasking. By keeping the masking noise source constant at 0◦ azimuth and moving the signal source around the animal in 30◦ steps, they observed a masking release of up to 10 dB (Dent et al., 1997; see also the review by Klump, 1996). Binaural processing will thus contribute to signal detection in communication networks as well as in other natural situations. However, such a large masking release has not been found in all species that have been tested. For example, in the green treefrog Hyla cinerea, spatial unmasking of up to 3 dB has been observed by phonotaxis experiments with separate noise and signal sources (Schwartz & Gerhardt, 1989). Recognition of signals

Detecting a signal is the first step in perception. Individuals in an acoustic communication network may become alert when they detect a signal, but further reaction will depend on the specific message: that is, the signal needs to be recognized. Signal recognition in the sense of statistical separation of signals may be explained best by an everyday example. When listening to the radio while driving a car, an individual may tune to a programme and can just detect ‘some signal’ or even ‘human speech’. By turning up the volume, and thus increasing the S/N ratio between the speech and the engine noise of the car, words and, therefore, the meaning of the message can be recognized. More theoretically, detecting a signal means observing the occurrence of a signal, i.e. the addition of some signal to a (possibly noisy) background. However, recognition implies the ability to classify a detected signal as a member of a set of many (e.g. Green et al., 1977; Wiley & Richard, 1982). We have little direct evidence on how the S/N ratio for recognition compares with the S/N ratio required for detection of a signal. Lohr et al. (2003) presented budgerigars and zebra finches Taeniopygia guttata with contact calls of three species (zebra finch, budgerigar, canary Serinus canarius) in order to determine thresholds for the detection of the calls in noise. They also determined each species’ ability to discriminate between different call types of zebra finches or budgerigars in the same masking noise. Birds were thus forced in noisy background conditions to ‘hear out’ and recognize a deviant call in a series of repeating reference calls. Thresholds for the discrimination of both conspecific

463

P1: KPB/... P2: KOD 0521823617c20.xml

464

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

U. Langemann & G. M. Klump and heterospecific calls in noise were about 3 dB worse than thresholds for call detection: the S/N ratio for recognition had to be, on average, 3 dB higher than the S/N ratio for mere detection (Lohr et al., 2003). There is no information from field studies that allows us to compare the S/N ratio necessary for detection of signals with that for recognition of signals in the natural environment. This is because field studies use playback to elicit natural species-specific responses and, therefore, the subjects have both detected and recognized the signal (e.g. Brenowitz, 1982a,b; Aubin & Jouventin, 1998). Brenowitz (1982a,b) studied the reaction of territorial male red-winged blackbirds Agelaius phoeniceus to song signals that were either played alone or with wideband random noise added to the playback song. Playback elicited more high-intensity song and visual display when the S/N ratio was increased from 0 to 3 dB (as measured in the 4 kHz octave band that contained most of the spectral energy of the song). Since this male response required the recognition of the signal, one can conclude that 3 dB is a conservative estimate of the S/N ratio necessary for signal recognition. Using data on auditory signal detection in red-winged blackbirds obtained in the laboratory (Hienz & Sachs, 1987), Klump (1996) calculated that the S/N ratio necessary for detection should be about 8 dB less than the S/N ratio necessary for recognition. In king penguins, the S/N ratio necessary for recognition appears to be lower than in the red-winged blackbird (Aubin & Jouventin, 1998). However, as suggested above, this may be because of the distinctive envelope-modulation pattern of penguin calls. The sender’s adaptations for maximizing signal transmission

Behavioural observations of signalling birds suggest that senders have evolved mechanisms to modify signal production in order to improve detection by the receiver. Holland et al. (1998) concluded from measurements of broadcast song of the wren Troglodytes troglodytes that higher song posts could optimize sound transmission. Broadcasting song from high perches where the vegetation is less dense also possibly improves the ability to detect responses of conspecifics in wrens (Holland et al., 1998) and blackbirds Turdus merula (Dabelsteen et al., 1993). However, it should be noted that changing location alone may improve perception by allowing sequential integration of acoustic information, for example in the context of sound localization. Another strategy employed by a sender to increase information transfer within a communication network is to adapt its sound output to the level of the background noise. This requires that a sender constantly monitors the level and spectral composition of the background noise interfering with its own vocalization. The increase in sound-pressure level of vocal output at times of increased levels of

P1: KPB/... P2: KOD 0521823617c20.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

Perception and acoustic communication networks background noise is called the Lombard effect and it has been well investigated in humans (e.g. Pick et al., 1989). Sinnott et al. (1975) demonstrated that trained monkeys Macaca fascicularis and M. nemestrina spontaneously increased their call amplitude if noise bands of the same fundamental frequencies as their call masked them, but not if noise bands were of much higher frequencies. The monkeys’ voice amplitude increased by about 2 dB for every 10 dB of masking noise. Also budgerigars significantly increase the level of their vocalization in response to noise of the same frequency spectrum as their contact calls but not to noise outside this spectral range (Manabe et al., 1998). The Lombard effect has also been reported in zebra finches (Cynx et al., 1998) and nightingales Luscinia megarhynchos (Brumm & Todt, 2002). The first animal species in which the Lombard effect has been shown under field conditions is the blue-throated hummingbird Lampornis clemenciae (Pytte et al., 2003). The authors observed that naturally occurring and experimentally controlled amplitude changes of the ambient noise level induced change in amplitude of the birds’ territorial advertisement call. Localization and distance perception Perceiving the direction of a sound source

For participants in communication networks, it is advantageous to be able to identify the location of the signal source. For example, the pattern of alarm calls in a bird community could provide a good estimate of the path taken by a predator (e.g. McGregor & Dabelsteen, 1996). Also, in territorial interactions, birds appear to combine information from the song signal and the direction from which it is heard to evaluate the potential threat by a competitor (e.g. McGregor & Avery, 1986). There is considerable variation in the accuracy of sound localization between species. Furthermore, each species’ ability to localize sound depends on the physical characteristics of the sound, such as the frequency spectrum of the sound or its temporal characteristics. The accuracy and mechanisms of sound localization have been reviewed in frogs (e.g. Rheinlaender & Klump, 1988), in birds (e.g. Klump, 2000) and in terrestrial mammals (e.g. Gourevitch, 1987; Brown, 1994). Two cues are used in sound localization: the difference in the time of arrival (or the phase difference) and the intensity differences between the spectral components of the sound impinging on the two ears. Figure 20.2a shows an example of interspecific variation in the accuracy of the localization of tones in the horizontal plane (azimuth) for two birds of prey, barn owl Tyto alba (Knudsen & Konishi, 1979) and sparrowhawk Accipiter nisus (G. M. Klump & E. Kretzschmar, unpublished data), and for four species of small birds, great tit (Klump et al., 1986), zebra finch, budgerigar and canary

465

P1: KPB/... P2: KOD 0521823617c20.xml

466

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

U. Langemann & G. M. Klump (a)

(b)

Fig. 20.2. The accuracy of azimuth sound localization. (a) Minimum detectable angles of pure-tone stimuli in relation to frequency as determined in the laboratory in two avian predators (open symbols, Knudsen & Konishi, 1979; G. M. Klump & E. Kretzschmar, unpublished data) and four species with smaller interaural distances than the raptors (filled symbols; Klump et al., 1986; Park & Dooling, 1991).

P1: KPB/... P2: KOD 0521823617c20.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

Perception and acoustic communication networks (Park & Dooling, 1991). The superior accuracy of the two birds of prey may be explained at least partly by the physical properties of their auditory system providing larger interaural cues and by specializations in processing the interaural sound differences in the auditory pathway (see Klump, 2000). When discussing the accuracy of sound localization in the natural environment, data from field studies are relevant (e.g. Nelson & Stoddard, 1998). Figure 20.2b shows the sound localization accuracy of a trained male sparrowhawk in the laboratory and in the field (G. M. Klump & E. Kretzschmar, unpublished data). The data in both sets were obtained with the same stimulus paradigm and operant procedure and so can be compared directly. The sparrowhawk’ssound localization accuracy was considerably reduced in the field. Localization accuracy for natural sounds was similar to the accuracy for tones of comparable frequency. The data shown in Fig. 20.2 were obtained by forcing the bird to localize a single signal presented at an unpredictable time. This procedure ensures that the bird is only using open-loop sound localization (Klump, 1995) and cannot use strategies to integrate information over several signal presentations or maximize binaural cues in some other way. It is to be expected that field studies presenting several signals before the subject responds will result in more accurate sound localization (e.g. Nelson & Stoddard, 1998). Perceiving the distance of a sound source

To assess the location of a signal source, knowledge of distance is as important as information about the direction from which the signal is heard. Assessing distance information is often referred to as ranging (Morton, 1982). In the past, distance assessment has almost exclusively been investigated in birds by simulating territorial intrusions and studying the behavioural response of the territory owner (reviewed by Naguib & Wiley, 2001). There are also many other common contexts in which distance is assessed, for example animals maintaining contact with mates or flock members when moving through dense vegetation.

Fig. 20.2 (cont.) (b) Minimum angles detectable by a sparrowhawk Accipiter nisus determined either in an anechoic chamber in the laboratory (open symbols, thick grey lines) or in a natural deciduous forest (closed symbols, thick black lines) using a two-alternative forced-choice procedure that had been established for measuring the accuracy of sound localization in small birds (Klump, 1995). Circles show the minimum detectable angle for pure tones (open circles are the same data as in (a). Diamonds indicate the minimum detectable angle for the ‘seeet’ aerial predator call. Horizontal lines represent the frequency range and the sparrowhawk’s localization accuracy for great tit mobbing calls (dashed thick lines) and scolding calls (solid thick lines). (G. M. Klump & E. Kretzschmar, unpublished data.)

467

P1: KPB/... P2: KOD 0521823617c20.xml

468

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

U. Langemann & G. M. Klump Two recent studies have provided data on the accuracy of distance perception as revealed by the approach responses of birds in the natural environment. Nelson & Stoddard (1998) measured the accuracy of the approach by Eastern towhees Pipilo erythrophthalmus to a loudspeaker playing back the species’ calls. The error in distance assessment was determined from the birds’ closest approach to the loudspeaker. The birds’ initial distance to the sound source was approximately 10, 20 or 30 m and resulted in average distance errors of 2.3, 3.5 and 3.4 m, respectively. In additional experiments, the actual distance to the sound source did not match the characteristics of the playback signal. For example, a signal rerecorded after being transmitted over a distance of 20 m was played from an actual distance of 10 m resulting in a simulated distance of 30 m. About half of the experimental birds responded by flying the actual distance towards the loudspeaker and the other half by flying the simulated distance, suggesting that some cues for distance assessment are derived from signal characteristics and some from the actual location of the sound source (Nelson & Stoddard, 1998). Simulated distances have also been used by Naguib et al. (2000) to study distance assessment by chaffinches Fringilla coelebs. Unlike previous field studies, Naguib et al. (2000) manipulated song signals by simulating their transmission in a virtual forest with the help of a computer. This allowed control over the amount of reverberation imposed and control over frequency-dependent attenuation. Simulated transmission distances ranged from 0 (original source signal) to 120 m. The approach response to playback of a single song with different virtual distances showed that chaffinches mainly discriminated between playback signals simulating shorter distances (0, 20 and 40 m) and playback signals simulating longer distances (80 and 120 m) but did not discriminate within short- or long-distance categories. This means that chaffinches exhibit a categorical response to simulated intruders close to, versus more distant from, their territory, rather than gauging their approach to the virtual distance of the sound source (Naguib et al., 2000). Laboratory studies of birds’ perception allow us to evaluate the salience of the cues that may be used for distance assessment. Cues suggested by field studies include the overall signal amplitude, the frequency-dependent excess attenuation, the amplitude modulation of the signal envelope imposed by atmospheric turbulence along the transmission path, the addition of noise to the signal, and reverberation resulting from echoes overlapping or trailing the signal (e.g. Dabelsteen et al., 1993; McGregor, 1994). Overall signal amplitude has been identified as a useful cue for distance assessment in psychoacoustic studies. Phillmore et al. (1998) trained zebra finches and black-capped chickadees Poecile atricapillus with operant procedures to distinguish calls and songs from either species that were recorded in a woodland habitat at distances of 5, 25, 50 and 75 m from the

P1: KPB/... P2: KOD 0521823617c20.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

Perception and acoustic communication networks loudspeaker. After training with a set of signals from all four distances, the birds could discriminate between the signals of the set and also between unknown songs from these recording distances. The stimuli representing various distances differed in amplitude at least as much as the known intensity-difference limen (the just-noticeable intensity difference) in birds (Dooling & Saunders, 1975; Klump & Baur, 1990). Removing amplitude cues made discrimination considerably worse (Phillmore et al., 1998), indicating its potential role in distance assessment. However, in the field, amplitude alone may not be a reliable indicator of distance, since head movements of the singing bird can lead to amplitude differences at the receiver’s position (e.g. Larsen & Dabelsteen, 1990). A number of field studies suggest reverberation as an important cue for distance assessment (e.g. McGregor, 1994; Naguib & Wiley, 2001). Echos that are imposed on each signal element during transmission ‘degrade’ (i.e. distort) its original amplitude and time pattern: the study by Holland et al. (2001) suggested that wrens can extract cues that allow distance assessment from the echo tail trailing the signal. Wrens responded to songs consisting of undegraded elements with added trailing echo tails (from degraded elements) in the same way as to degraded songs (i.e. degraded element and echo tail). Songs consisting of degraded elements without echo tails elicited a response that was intermediate between that to an undegraded and that to a degraded song with echo tail, stressing the salience of the echo tail as a cue. Psychoacoustic studies of humans also indicate that echo tails provide an important cue for distance assessment; we appear to use the directto-reverberant energy ratio (Zahorik, 2002). This cue may allow us to estimate sound-source distance independent of the sound level, which in turn may allow the loudness of the source to be inferred (Zahorik & Wightman, 2001). SINDSCAL: an analysis method for perceptual distances

Laboratory studies in the great tit also indicate that echoes alone may provide a sufficient cue for distance assessment. In this final part, we would like to describe how perceptual differences can be examined in trained animals. We present a multidimensional scaling procedure (SINDSCAL: symmetric individual difference scaling; e.g. Arabie et al., 1987) that is especially suited to investigate which signal modifications are salient to the animals. A virtual forest (Naguib et al., 2000) was used to impose reverberation on synthetic great tit song signals equivalent to sound transmission distances of between 5 and 320 m (Fig. 20.3). The sound pressure of the song signals was then adjusted to the same overall root mean square amplitude so that the reverberation pattern imposed on the signal remained the only possible cue to distance. Great tits were then trained in an operant Go/NoGo procedure with repeating background to discriminate between song

469

P1: KPB/... P2: KOD 0521823617c20.xml

470

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

U. Langemann & G. M. Klump

Fig. 20.3. An example of stimuli used to estimate perceptual differences in laboratory experiments with great tits. The spectrogram shows a two-element great tit song at a virtual distance of 5 m and the waveforms show the same song elements at virtual distances of 5, 80 and 160 m. Further details in text.

P1: KPB/... P2: KOD 0521823617c20.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

Perception and acoustic communication networks elements in which echo patterns alone indicated different virtual distances (e.g. Dooling & Okanoya, 1995). The response latencies of the great tits to all possible reference–test differences were recorded (Fig. 20.4a) and analysed by a SINDSCAL (e.g. Arabie et al., 1987) model with log-transformed data. The result of SINDSCAL is a three-dimensional object space (Fig. 20.4b), providing a kind of ‘perceptual map’ in which the response latencies are translated into relative distances between data points. The distance between data points in the perceptual space provides a measure of the perceived similarity of the acoustic signals. Data points that are close to each other indicate that large response latencies were observed in the discrimination (i.e. the difference was not salient to the birds). Short response latencies indicate a salient differences that will lead to a large spread of the data points. A three-dimensional SINDSCAL model accounts for 86–90% of the variance in response latencies. Since the first dimension explained most of the total variance, the latency data for the different experimental songs were reanalysed with a one-dimensional SINDSCAL model to allow easier comparison between song types. Perceptual space coordinates (from the one-dimensional model) of the four different two-note songs that were tested are shown in Fig. 20.5 as a function of the virtual distance of each signal. As in Fig. 20.4b, close perceptual space coordinates indicate signals that have been perceived as being more similar. There was no significant difference between the different song types tested (two-way ANOVA with distance and song type as factors and subsequent Tukey-tests, F 3,24 < 0.001; p = 1). However, within the factor ‘virtual distance’, significant differences were obvious (F 6,21 = 73.7; p ≤ 0.001). Space coordinates for ‘long’ virtual distances (160 and 320 m) differed significantly from those for the other distances but not from each other (i.e. ‘long’ virtual distances were similar for all the birds’). Space coordinates for the virtual distance of 80 m significantly differed from those of the three ‘short’ distances (5, 10 and 20 m). ‘Short’ virtual distances of 5, 10 and 20 m were treated as similar by the birds (i.e. no significant difference was found). Therefore, great tits in the laboratory provided with echo patterns as the only available distance assessment cue showed a categorical response that was similar to the response observed in field experiments with chaffinches (Naguib et al., 2000).

Auditory scene analysis So far we have discussed basic perceptual mechanisms involved in relatively simple auditory detection and discrimination tasks. Real-world situations, however, require receivers to analyse sounds from a mixture of simultaneously active sources: that is, to perform auditory scene analysis (Bregman, 1990). For example, this applies to communication networks of birds during the dawn chorus, when a receiver has to analyse streams of song elements from each individual

471

P1: KPB/... P2: KOD 0521823617c20.xml

0 521 582361 7

April 12, 2005

U. Langemann & G. M. Klump (a)

2500 Song 5

Response latency (ms)

472

CU1917B/McGregor

rs = −0.839 p < 0.001

2000

1500

1000

500 0

50

100

150

200

250

300

350

Virtual distance difference (m) Perceptual space (song 5)

(b)

80 m 40 m 160 m −

− −

20 m

5m

−

10 m

−

320 m

− −

− − − − −

− − − −

Fig. 20.4.

12:27

P1: KPB/... P2: KOD 0521823617c20.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

Perception and acoustic communication networks

Fig. 20.5. One-dimensional perceptual space coordinates as a function of the virtual distance (i.e. the distance simulated by imposing reverberations upon the signal) for four test songs. Similar perceptual space coordinates indicate that the differences in the echo pattern of the signals are not very salient to the great tits, and large differences between data points indicate salient differences have been detected between the respective echo modifications. There are no salient differences between signals of ‘short’ virtual distances of 5, 10 and 20 m or between signals of ‘long’ virtual distances of 160 and 360 m. Differences between signals of ‘short’ virtual distances and ‘long’ virtual distances are very salient to the birds. Virtual distances of 40 and 80 m lie in a transition range. This pattern occurs for all songs that were tested (see ANOVA results in the text).

Fig. 20.4 (cont.) Results of great tits scaling differences between songs manipulated to represent various virtual distances. The signal variants consisted of the same song elements that differed only in echo pattern, which simulated sound transmission distances of 5 m to 320 m. (a) Average response latencies of four great tits as a function of the difference in virtual distance (e.g. the virtual-distance difference between signals of simulated transmission distances of 320 and 80 m is 240 m). Larger response latencies indicate less-salient differences in the cues; shorter response latencies indicate more-salient differences. (b) Three-dimensional object space or ‘perceptual map’ (SINDSCAL model with log-transformed data, see text; Arabie et al., 1987) demonstrating the salience of reverberation for great tits. Distances between data points in perceptual space reflect response-latency differences in discriminating between the signal variants. Small distances indicate that signals are treated as being similar, and large distances indicate that salient differences between signals have been perceived. (U. Langemann, U. Pander & G. M. Klump, unpublished data.)

473

P1: KPB/... P2: KOD 0521823617c20.xml

474

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

U. Langemann & G. M. Klump singer. Laboratory experiments have shown that animals form auditory streams and analyse auditory scenes in a similar way to humans (e.g. Hulse et al., 1997; Feng & Ratnam, 2000; Moss & Surlykke, 2001; Hulse, 2002). However, some of the basic mechanisms of masking discussed above also contribute to an auditory scene analysis that is characterized by an improved segregation of overlapping signals. For example, we have shown that common modulation of components of sounds aids signal segregation and results in reduced masking of one sound by another. Similarly, common onsets or offsets of signal components lead to the formation of auditory objects (e.g. Geissler & Ehret, 2002) that can be analysed separately from other objects in the same auditory scene. Spatial separation of sources will also aid auditory object formation, and the spatial release from masking discussed above may be partly a result of improved signal segregation from the background.

Summary In this chapter, we have illustrated how sensory abilities of individuals affect auditory perception and thus acoustic communication. How does this relate to communication between individuals in networks? In a network, individuals are distributed in space. The relative position of any ‘node’ in this network, the distance between individuals, profoundly determines a receiver’s ability to detect and recognize acoustic signals. Because of the spatial distribution of signallers and receivers, the same propagated signal may result in quite different perception at different places in a communication network. Acoustic signals will be modified along their transmission path and will be masked by acoustical energy from other sources. On the one hand, masking is certainly the most important factor severely impairing the detection of acoustic communication signals. Spectral aspects and the temporal patterns of masking sounds affect the amount of masking that is exerted and the spatial distribution of concurrent sound sources (or individuals) contributes to masking efficiency. On the other hand, receivers may exploit changes imposed on a signal during transmission. For example, reverberation patterns will allow distance assessment of sound sources and, together with binaural cues, render it possible to gain insights into the spatial distribution of the individuals in a network. Our current knowledge from perceptual studies will provide a better understanding of animal behaviour within acoustically complex communication networks. In addition, an approach that takes the receiver’s perception into account will allow a better evaluation of communication behaviour in the field than approaches that rely mainly on physical properties of signals and their transmission (McGregor et al., 2000).

P1: KPB/... P2: KOD 0521823617c20.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

Perception and acoustic communication networks Acknowledgements The research was supported by grants from the Deutsche Forschungsgemeinschaft (SFB 204, FOR 306). Ulrike Pander provided data from experiments with great tits that were reanalysed for this study.

References Arabie, P., Carroll, J. D. & DeSarbo, W. S. 1987. Three-Way Multidimensional Scaling and Related Techniques. Newbury Park, CA: Sage. Aubin, T. & Jouventin, P. 1998. Cocktail-party effect in king penguin colonies. Proceedings of the Royal Society of London, Series B, 265, 1665–1673. Aubin, T., Jouventin, P. & Hildebrand, C. 2000. Penguins use the two-voice system to recognize each other. Proceedings of the Royal Society of London, Series B, 267, 1081–1087. Boersma, H. F. 1997. Characterization of the natural ambient sound environment: measurements in open agricultural grassland. Journal of the Acoustical Society of America, 103, 2104–2110. Bregman, A. S. 1990. Auditory Scene Analysis: The Perceptual Organization of Sound. Cambridge, MA: MIT Press. Brenowitz, E. A. 1982a. The active space of red-winged blackbird song. Journal of Comparative Physiology A, 147, 511–522. 1982b. Long-range communication of species identity by song in the red-winged blackbird. Behavioral Ecology and Sociobiology, 10, 29–38. Brown, C. H. 1994. Sound localization. In: Comparative Hearing: Mammals, ed. R. R. Fay & A. N. Popper. New York: Springer, pp. 57–96. Brumm, H. & Todt, D. 2002. Noise-dependent song amplitude regulation in a territorial songbird. Animal Behaviour, 63, 891–897. Budelis, J., Fishbach, A. & May, B. J. 2002. Behavioral assessment of comodulation masking release in cats. Abstracts of the Association for Research in Otolaryngology, 25, 211. Cynx, J., Lewis, R., Tavel, B. & Tse H. 1998. Amplitude regulation of vocalizations in noise by a songbird, Taeniopygia guttata. Animal Behaviour, 56, 107–113. Dabelsteen, T., Larsen, O. N. & Pedersen, S. B. 1993. Habitat-induced degradation of sound signals: quantifying the effects of communication sounds and bird location on blur ratio, excess attenuation, and signal-to-noise ratio in blackbird song. Journal of the Acoustical Society of America, 93, 2206–2220. Dent, M. L., Larsen, O. N. & Dooling, R. J. 1997. Free-field binaural unmasking in budgerigars (Melopsittacus undulatus). Behavioral Neuroscience, 111, 590–598. Dooling, R. J. & Okanoya, K. 1995. Psychophysical methods for assessing perceptual categories. In: Methods in Comparative Psychoacoustics, ed. G. M. Klump, R. J. Dooling, R. R. Fay & W. C. Stebbins. Basel: Birkh¨ auser, pp. 307–318. Dooling, R. J. & Saunders, J. C. 1975. Auditory intensity discrimination in the parakeet (Melopsittacus undulatus). Journal of the Acoustical Society of America, 58, 1308–1310.

475

P1: KPB/... P2: KOD 0521823617c20.xml

476

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

U. Langemann & G. M. Klump Douglas, H. D. & Conner, W. E. 1999. Is there a sound reception window in coastal environments? Evidence from shorebird communication systems. Naturwissenschaften, 86, 228–230. Dubois, A. & Martens, J. 1984. A case of possible vocal convergence between frogs and a bird in Himalayan torrents. Journal f¨ ur Ornithologie, 125, 455–463. Fay, R. R. 1988. Hearing in Vertebrates: a Psychophysics Databook. Winnetka, IL: Hill-Fay. F´egeant, O. 1999. Wind-induced vegetation noise. Part II: field measurements. Acustica/Acta Acustica, 85, 241–249. Feng, A. S. & Ratnam R. 2000. Neural basis of hearing in real-world situations. Annual Review of Psychology, 51, 699–725. Gerhardt, H. C. & Huber, F. 2002. Acoustic Communication in Insects and Anurans. Chicago, IL: University of Chicago Press. Geissler, D. B. & Ehret, G. 2002. Time-critical integration of formants for perception of communication calls in mice. Proceedings of the National Academy of Sciences, USA, 99, 9021–9025. Gourevitch, G. 1987. Binaural hearing in land mammals. In: Directional Hearing, ed. W. A. Yost & G. Gourevitch. New York: Springer, pp. 226–246. Green, D. M., Weber, D. L. & Duncan, J. E. 1977. Detection and recognition of pure tones in noise. Journal of the Acoustical Society of America, 62, 948–954. Hienz, R. D. & Sachs, M. B. 1987. Effects of noise on pure-tone thresholds in blackbirds (Agelaius phoeniceus and Molothrus ater) and pigeons (Columba livia). Journal of Comparative Psychology, 101, 16–24. Hine, J. E., Martin, R. L. & Moore, D. R. 1994. Free-field binaural unmasking in ferrets. Behavioral Neuroscience, 108, 196–205. Holland, J., Dabelsteen, T., Pedersen, S. B. & Larsen, O. N. 1998. Degradation of wren Troglodytes troglodytes song: implications for information transfer and ranging. Journal of the Acoustical Society of America, 103, 2154–2166. Holland, J., Dabelsteen, T., Bjørn, C. P. & Pedersen, S. B. 2001. The location of ranging cues in wren song: evidence from calibrated interactive playback experiments. Behaviour, 138, 189–206. Hulse, S. H. 2002. Auditory scene analysis in animal communication. Advances in the Study of Behaviour, 31, 163–200. Hulse, S. H., MacDougall-Shackleton, S. A. & Wisniewski, A. B. 1997. Auditory scene analysis by songbirds: stream segregation of birdsong by European starlings (Sturnus vulgaris). Journal of Comparative Psychology, 111, 3–13. Klump, G. M. 1995. Sound localization studies in nonspecialized birds. In: Methods in Comparative Psychoacoustics, ed. G. M. Klump, R. J. Dooling, R. R. Fay & W. C. Stebbins. Basel: Birkh¨ auser, pp. 171–182. 1996. Bird communication in the noisy world. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca NY: Cornell University Press, pp. 321–338. 2000. Sound localization in birds. In: Comparative Hearing: Birds and Reptiles, ed. R. J. Dooling, R. R. Fay & A. N. Popper. New York: Springer, pp. 247–305. Klump, G. M. & Baur, A. 1990. Intensity discrimination in the European starling (Sturnus vulgaris). Naturwissenschaften, 77, 545–548.

P1: KPB/... P2: KOD 0521823617c20.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

Perception and acoustic communication networks Klump, G. M. & Curio, E. 1983. Warum liegen viele Laut¨ außerungen von Singv¨ ogeln oberhalb des Bereichs besten H¨ orens? Verhandlungen der Deutschen Zoologischen Gesellschaft, 76, 182. Klump, G. M. & Gerhardt, H. C. 1992. Mechanisms and function of call-timing in male–male interactions in frogs. In: Playback and Studies of Animal Communication, ed. P. K. McGregor. London: Plenum Press, pp. 153–174. Klump, G. M. & Langemann, U. 1995. Comodulation masking release in a songbird. Hearing Research, 87, 157–164. Klump, G. M. & Nieder, A. 2001. Release from masking in fluctuating background noise is represented in a songbird’s auditory forebrain. Neuroreport, 12, 1825–1829. Klump, G. M. & Shalter, M. D. 1984. Acoustic behaviour of birds and mammals in the predator context: I. Factors affecting the structure of alarm calls. II. The functional significance of alarm signals and their evolution. Zeitschrift f¨ ur Tierpsychologie, 66, 189–226. Klump, G. M., Windt, W. & Curio E. 1986. The great tit’s (Parus major) auditory resolution in azimuth. Journal of Comparative Physiology, 158, 383–390. Klump, G. M., Langemann, U. & Nieder, A. 1998. Mechanisms that improve signal detection in noise: a study of comodulation masking release in a songbird. In: Psychophysical and Physiological Advances in Hearing, ed. A. R. Palmer, A. Rees, A. Q. Summerfield & R. Meddis. London: Whurr, pp. 270–276. Klump, G. M., Kittel, M., Wagner, E. 2001. Comodulation masking release in the Mongolian gerbil. Abstracts of the Association for Research in Otolaryngology, 24, 84. Knudsen, E. I. & Konishi, M. 1979. Mechanisms of sound localization in the barn owl (Tyto alba). Journal of Comparative Physiology, 133, 13–21. Langemann, U., Gauger, B. & Klump, G. M. 1998. Auditory sensitivity in the great tit: perception of signals in the presence and absence of noise. Animal Behaviour, 56, 763–769. Larom, D., Garstang, M., Lindeque, M. et al. 1997. Meteorology and elephant infrasound at Etosha National Park, Namibia. Journal of the Acoustical Society of America, 101, 1710–1717. Larsen, O. N. & Dabelsteen, T. 1990. Directionality of blackbird vocalization. Implications for vocal communication and its further study. Ornis Scandinavica, 21, 37–45. Lengagne, T., Lauga, J. & Aubin, T. 2001. Intra-syllabic acoustic signatures used by the king penguin in parent-chick recognition: an experimental approach. Journal of Experimental Biology, 204, 663–672. Lohr, B., Wright T. F. & Dooling, R. J. 2003. Detection and discrimination of natural calls in masking noise by birds: estimating the active space of a signal. Animal Behaviour, 65, 763–777. Long, G. R. 1977. Masked auditory thresholds from the bat, Rhinolophus ferrumequinum. Journal of Comparative Physiology A, 116, 247–255. Manabe, K., Sadr, E. I. & Dooling, R. J. 1998. Control of vocal intensity in budgerigars (Melopsittacus undulatus): differential reinforcement of vocal intensity and the Lombard effect. Journal of the Acoustical Society of America, 103, 1190–1198.

477

P1: KPB/... P2: KOD 0521823617c20.xml

478

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

U. Langemann & G. M. Klump Marten, K. & Marler, P. 1977. Sound-transmission and its significance for animal vocalization. I. Temperate habitats. Behavioral Ecology and Sociobiology, 2, 271–290. McFadden, D. & Wright, B. A. 1990. Temporal decline of masking and comodulation detection differences. Journal of the Acoustical Society of America, 88, 711–724. McGregor, P. K. 1994. Sound cues to distance: the perception of range. In: Perception and Motor Control in Birds: An Ecological Approach, ed. M. N. O. Davies & P. R. Green. Berlin: Springer, pp. 74–94. McGregor, P. K. & Avery, M. I. 1986. The unsung songs of great tits (Parus major): learning neighbours’ songs for discrimination. Behavioral Ecology and Sociobiology, 18, 311–316. McGregor, P. K. & Dabelsteen, T. 1996. Communication networks. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca NY: Cornell University Press, pp. 409–425. McGregor, P. K. & Peake, T. M. 2000. Communication networks: social environments for receiving and signalling behaviour. Acta Ethologica, 2, 71–81. McGregor, P. K., Otter, K. A. & Peake, T. M. 2000. Communication networks: receiver and signaller perspectives. In: Animal Signals. Signalling and Signal Design in Animal Communication, ed. Y. Espmark, T. Amundsen & G. Rosenqvist. Trondheim: Tapir Academic Press, pp. 405–416. Miller, G. A. 1947. Sensitivity to changes in the intensity of white noise and its relation to masking and loudness. Journal of the Acoustical Society of America, 19, 609–619. Moore, B. C. J. 1990. Co-modulation masking release: spectro-temporal pattern analysis in hearing. British Journal of Audiology, 24, 131–137. 2003. An Introduction to the Psychology of Hearing. Amsterdam: Academic Press. Moore, B. C. J. & Glasberg, B. R. 1983. Suggested formulae for calculating auditory-filter bandwidths and excitation pattern. Journal of the Acoustical Society of America, 74, 750–753. Morton, E. S. 1982. Grading, discreteness, redundancy, and motivation–structural rules. In: Acoustic Communication in Birds, Vol. I, ed. D. E. Kroodsma, E. H. Miller & H. Ouellet. New York: Academic Press, pp.183–212. Moss, C. F. & Surlykke, A. 2001. Auditory scene analysis by echolocation in bats. Journal of the Acoustical Society of America, 110, 2207–2226. Naguib, M. & Wiley, R. H. 2001. Estimating the distance to a source of sound: mechanisms and adaptations for long-range communication. Animal Behaviour, 62, 825–837. Naguib, M., Klump, G. M., Hillmann, E., Griessman, B. & Teige, T. 2000. Assessment of auditory distance in a territorial song bird: accurate feat or rule of thumb? Animal Behaviour, 59, 715–721. Nelken, I., Rotman, Y. & Bar Yosef, O. 1999. Responses of auditory-cortex neurons to structural features of natural sounds. Nature, 397, 154–157. Nelson, B. S. & Stoddard, P. K. 1998. Accuracy of auditory distance and azimuth perception by a passerine bird in natural habitat. Animal Behaviour, 56, 467–477.

P1: KPB/... P2: KOD 0521823617c20.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

Perception and acoustic communication networks Niemiec, A. J. 2001. The effects of increasing masker temporal regularity on co-modulation masking thresholds in chinchillas. Abstracts of the Association for Research in Otolaryngology, 24, 85. Park, T. J. & Dooling, R. J. 1991. Sound localization in small birds: absolute localization in azimuth. Journal of Comparative Psychology, 105, 125–133. Pick, H. L., Siegel, G. M., Fox, P. W., Garber, S. R. & Kearney, J. K. 1989. Inhibiting the Lombard effect. Journal of the Acoustical Society of America, 85, 894–900. Phillmore L. S., Sturdy, C. B., Ramsay, S. M. & Weisman, R. G. 1998 Discrimination of auditory distance cues by black-capped chickadees (Poecile atricapillus) and zebra finches (Taeniopygia guttata). Journal of Comparative Psychology, 112, 282–291. Pressnitzer, D., Meddis, R., Delahaye, R. & Winter, I. M. 2001. Physiological correlates of comodulation masking release in the mammalian ventral cochlear nucleus. Journal of Neuroscience, 21, 6377–6386. Pytte, C. L., Rusch, K. M. & Ficken, M. S. 2003. Regulation of vocal amplitude by the blue-throated hummingbird, Lampornis clemenciae. Animal Behaviour, 66, 703–710. Rheinlaender, J. & Klump, G. M. 1988. Behavioural aspects of sound localization. In: The Evolution of the Amphibian Auditory System, ed. B. Fritsch, M. J. Ryan, W. Wilczynski, T. E. Hetherington & W. Walkowiak. New York: John Wiley, pp. 297–305. Richards, D. G. & Wiley, R. H. 1980. Reverberations and amplitude fluctuations in the propagation of sound in a forest: implications for animal communication. American Naturalist, 115, 381–399. Ryan, M. J. & Brenowitz, E. A. 1985. The role of body size, phylogeny and ambient noise in the evolution of bird song. American Naturalist, 126, 87–100. Schooneveldt, G. P. & Moore, B. C. J. 1989. Comodulation masking release (CMR) as a function of masker bandwidth, modulator bandwidth, and signal duration. Journal of the Acoustical Society of America, 85, 273–281. Schwartz, J. J. & Gerhardt, H. C. 1989. Spatially mediated release from auditory masking in an anuran amphibian. Journal of Comparative Physiology A, 166, 37–41. Shannon, C. E. & Weaver, W. 1949. The Mathematical Theory of Communication. Urbana, IL: University of Illinois Press. Sinnott, J. M., Stebbins, W. C. & Moody, D. B. 1975. Regulation of voice amplitude by the monkey. Journal of the Acoustical Society of America, 58, 412–414. Staicer, C. A., Spector, D. A. & Horn, A. G. 1996. The dawn chorus and other dial patterns in acoustic signalling. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca NY: Cornell University Press, pp.426–453. Swets, J. A., Shipley, E. F., McKey, M. J. & Green, D. M. 1950. Multiple observations of signals in noise. Journal of the Acoustical Society of America, 31, 514–521. Waser, P. M. & Waser, M. S. 1977. Experimental studies of primate vocalization: specializations for long-distance propagation. Zeitschrift f¨ ur Tierpsychologie, 43, 239–263. Wiley, R. H. & Richards, D. G. 1978. Physical constraints on acoustic communication in the atmosphere: implications for the evolution of animal vocalizations. Behavioral Ecology and Sociobiology, 3, 69–94.

479

P1: KPB/... P2: KOD 0521823617c20.xml

480

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:27

U. Langemann & G. M. Klump 1982. Adaptations for acoustic communication in birds: sound transmission and signal detection. In: Acoustic Communication in Birds, Vol. I, ed. D. E. Kroodsma, E. H. Miller & H. Ouellet. New York: Academic Press, pp. 131–181. Wollerman, L. & Wiley, R. H. 2002. Background noise from a natural chorus alters female discrimination of male calls in a Neotropical frog. Animal Behaviour, 63, 15–22. Yost, W. A. 1994. Fundamentals of Hearing: An Introduction. San Diego CA: Academic Press. Zahorik, P. 2002. Direct-to-reverberant energy ratio sensitivity. Journal of the Acoustical Society of America, 112, 2110–2117. Zahorik, P. & Wightman, F. L. 2001. Loudness constancy with varying sound source distance. Nature Neuroscience, 4, 78–83.

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:12

21

Hormones, social context and animal communication r u i f. o l i v e i r a Instituto Superior de Psicologia Aplicada, Lisbon, Portugal

Introduction The views on the role that hormones play in the control of behaviour have changed progressively with time. Hormones were classically seen as causal agents of behaviour, acting directly on the display of a given behaviour. This view was mainly supported by early studies of castration and hormone-replacement therapy, which showed that some behaviours were abolished by castration and restored by exogenous administration of androgens (Nelson, 2001). Later this view shifted towards a more probabilistic approach and hormones started to be seen more as facilitators of behaviour than as deterministic factors (Simon, 2002). According to this new view, hormones may increase the probability of the expression of a given behaviour by acting as modulators of the neural pathways underlying that behavioural pattern. For example, the effects of androgens on the expression of aggressive behaviours in mammals are mediated by modulatory effects on central serotonergic and vasopressin pathways (Simon, 2002). Yet, it is also known that the social environment (i.e. network of interacting individuals) also feeds back to influence hormone levels (Wingfield et al., 1990), suggesting a twoway type of interaction between hormones and behaviour. In this chapter, I will develop the hypothesis that social modulation of androgens is an adaptive mechanism through which individuals adjust their motivation according to the social context that they are facing. Thus, the social interactions within a given social network would stimulate the production of androgens in the individuals and the individual levels of androgens would be a function of the perceived social status and the stability of the social environment in which the animal is living. According to this view, androgens may play a key role as endocrine mediators of the effects Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

481

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

482

0 521 582361 7

April 11, 2005

19:12

R. F. Oliveira

Fig. 21.1. Interplay between androgens, social behaviour and social context. Androgens influence the production of a number of social behaviours involved in communication interactions between animals. In turn, these social interactions among a network of individuals will shape the social context in which these animals live, which subsequently will modulate their androgen levels.

of social context on the expression of social behaviour, allowing the animal to adjust its social behaviour to the context by modulating sensory, cognitive and motor neural mechanisms underlying animal communication (Fig. 21.1). In this chapter, I only consider vertebrates because they have a remarkably similar endocrine system, whereas that of invertebrates can be very different (e.g. the androgenic hormone in crustaceans is a peptide not a steroid as in vertebrates (Hasegawa et al., 2002)). Within the vertebrates, I mainly use examples from bony fishes, a group with wide diversity in mating and parental care systems that makes up about half the existing vertebrate species (Nelson, 1994). I have also concentrated on androgens and male behaviour because I argue that the social network in which the individual lives modulates its neuroendocrine system, which, in turn, adjusts the expression of behaviour according to social context. Stress hormones are, by definition, affected by the social environment and a number of reviews on social stress and hypothalamus–pituitary–adrenal axis have been published since the early 1990s (e.g. Sapolsky, 1992). Consequently, an additional benefit of this chapter is to claim that, like stress hormones, androgens (and perhaps also other hormones) respond in an adaptive way to the social context, preparing the animal for the social interactions that it has to face in its everyday life.

Hormones and communication I: the dyadic view Conceptually, the neurochemical pathways modulated by hormones can be part of one of three major functional compartments of the nervous system: sensory, central processing and effector systems (Nelson, 2001). If we translate this rationale to the communication paradigm, one can consider that hormones may affect communication by modulating the production of the signal in the

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:12

Hormones, social context and animal communication

Sender

Receiver Input/ perception

Input/ perception Central processing

Signal Motor output

Central processing

Motor output

Hormones

Hormones

Somatic releasers Fig. 21.2. Flow of information in a communication dyad. The arrows indicate the direction of circulation of the information within and between individuals. In the sender, sensory information received will influence central processing mechanisms in the central nervous system (CNS), which control, at a higher level, the behavioural motor output systems that produce the signals. Hormones may modulate signal production by the sender by acting on central mechanisms, on motor output mechanisms or by modifying somatic structures that affect the emission of the signal (i.e. somatic releasers). In the receiver, the signal will be detected by sensory systems and after peripheral processing will be forwarded to central processing systems in the CNS. Hormones may affect signal reception and processing in the receiver by acting directly on the sensory systems that perceive it and/or by acting at a higher level on the central processing mechanism of the CNS. The central processing of the signal by the limbic system (and other structures involved in motivational mechanism) may feed back on hormone levels. The boxes delimit the two organisms and within the boxes the grey elliptical areas represent the nervous system.

sender, the perception of the signal by the receiver or the central processing of the message in both senders and receivers (Fig. 21.2). Hormonal modulation of effector pathways

In senders, hormones may modulate the effector pathways that are involved in the motor circuits underlying the production of the signal. In this way, hormones can affect the expression of visual displays, vocalizations or pheromone production and/or release. From the numerous examples in the literature, I have selected the following, which are intended to cover different communication channels in different vertebrate taxa. Androgens and the production of acoustic signals

In songbirds, circulating levels of testosterone are higher at the peak of the breeding season when singing behaviour reaches its maximum (e.g. Rost, 1990,

483

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

484

0 521 582361 7

April 11, 2005

19:12

R. F. Oliveira 1992; Smith et al., 1997). Moreover, song production is substantially reduced after castration and is restored after androgen-replacement therapy (e.g. Arnold, 1975; Heid et al., 1985). Finally, both androgen and oestrogen receptors have been localized in the song control nuclei of the bird brain: the former in the high vocal centre, the robust nucleus of the archistriatum, the lateral part of the magnocellular nucleus of the anterior neostriatum and the dorsomedial part of the intercollicular nucleus; the latter in the high vocal centre and the intercollicular nucleus (Balthazart et al., 1992; Brenowitz & Arnold, 1989, 1992; Gahr et al., 1987, 1993). Song is produced by the coordinated contraction of respiratory, syringeal and craniomandibular muscles (Suthers et al., 1999). The activity of syringeal muscles regulates both the timing and the fundamental frequency of the sound (Suthers et al., 1999). Therefore, by acting directly on the activity and development of syringeal muscles, hormones may affect song production. For example, in zebra finches Taenopygia guttata, androgens inhibit the activity of the enzyme cholinesterase, which breaks down the neurotransmitter acetylcholine in the neuromuscular junctions of the syrinx. This results in a longer lifetime for the neurotransmitter in the synaptic cleft, which will affect the syringeal contraction pattern and, subsequently, song output and/or structure (Luine et al., 1980). Testosterone is also known to increase both syringeal muscle mass (Luine et al., 1980) and the density of acetylcholine receptors in syringeal muscles, suggesting that circulating levels of testosterone may increase the size and number of endplates in neuromuscular junctions (Bleisch et al., 1984). Also, in non-oscine birds, testosterone is effective in inducing changes in call structure by acting on the motor vocal structure underlying these calls. In grey partridges Perdix perdix, male mating calls used by females in mate choice are affected by testosterone treatment, which induces a thickening of the external tympanic membranes that are known to be the main sound source in galliforms (Beani et al., 1995). These effects of androgens on motor systems underlying the production of vocal signals are not exclusive to birds. Many fish species also use sounds to communicate. Male toadfish are among the most vocal fish, producing loud humming calls to attract females to their nest site (e.g. plainfin midshipman Porichthys notatus; Brantley & Bass, 1994). Also in toadfish, the exogenous administration of androgens promotes the development of the sonic muscles, for example the oyster toadfish Opsanus tau (Fine & Pennymaker, 1986) and the plainfin midshipman (Brantley et al., 1993). Another example comes from amphibians, in which vocal behaviour is sexually dimorphic in most species and thus potentially androgen dependent (Kelley, 2002). In the African frog Xenopus laevis, males produce mating calls characterized by fast trills that attract females (Wetzel & Kelley, 1983; Kelley, 2002). The call-production organ of X. laevis is the larynx and all sounds are produced underwater (Kelley, 2002). The sex differences in vocal behaviour observed in this species

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:12

Hormones, social context and animal communication are mostly a result of sex differences in adult laryngeal synapses. Male larynx motor neurons release less neurotransmitter, which will produce lower postsynaptic potentials than in females, allowing the fibres to reach a spike threshold (Tobias et al., 1995). This synaptic facilitation in male motor neurons allows modulation of the amplitude of the trills, a characteristic of the call that is used by females when assessing the males as potential mates (Tobias et al., 1995). Contrary to most cases of sexual dimorphism in which the default situation is female, the sex differences in postsynaptic response emerge in females under the influence of oestradiol, with the default being the slow neurotransmitter release typical of males (Tobias & Kelley, 1995). However, other sex differences in this vocal system are androgen dependent, namely the differentiation of laryngeal motor neurons, muscles fibres and laryngeal cartilage (Kelley, 2002). Androgens and pheromone production and/or release

A very large number of mammals use chemical signals (i.e. pheromones) in intraspecific communication. These pheromones can be produced by specific scent glands or are released into the environment in the urine or in other body fluids (Bradbury & Vehrencamp, 1998). Most mammals use marking behaviour to release these pheromones, a behaviour that is sexually dimorphic ( Johnson, 1973; Brown & McDonald, 1985; Chs. 11 and 16). There are classic examples of marking behaviour, such as the scent marking of reindeer Rangifer tarandus, with preorbital, caudal and tarsal glands as well as with urine (see Brown & MacDonald, (1985) for other examples and detailed references). Scent marks are also widespread in rodents such as mice Mus musculus, hamsters Mesocricetus auratus and rats Ratus spp. (Hurst, 1990 Chs. 11 and 16). In general, both pheromone production and its release (i.e. scent marking) are androgen dependent in males, as shown by castration and testosteronereplacement therapy experiments for example hamsters Mesocricetus auratus (Gawienowsky et al., 1976), meadow voles Microtus pennsylvanicus (Ferkin & Johnston, 1993), tree shrews Tupaia belangeri (Holst & Eichman, 1998) and Wistar rats Manzo et al., 2002); however, see Lepri & Randall (1983) and Randall (1986) for an exception regarding the endocrine control of sandbathing in male kangaroo rats Dipodomys spp. The scent-marking behaviour decreases after castration and is restored after treatment with testosterone (e.g. rats: Brown, 1978; Taylor et al., 1987; Manzo et al., 2002). Interestingly, in many species, testosterone is the prohormone for this effect, because it needs to be metabolized in specific brain areas into oestradiol or dihydrotestosterone in order to become biologically active, for example rabbits Oryctolagus cuniculus (Gonz´ alez-Mariscal et al., 1993), gerbils Meriones unguiculatus (Yahr & Stephens, 1987) and Wistar rats (Manzo et al., 2002).

485

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

486

0 521 582361 7

April 11, 2005

19:12

R. F. Oliveira Chemical communication is also widespread in urodeles, playing a major role in sex recognition and mate attraction (e.g. European newts Triturus spp.: Cedrini & Fasolo, 1970; Malacarne et al., 1984; Belvedere et al., 1988). One of the best-studied species is the Japanese red-bellied newt Cynops pyrrhogaster. In this species, males produce a female-attracting pheromone (sodefrin) with the abdominal glands, which is released by the cloaca of the male (Kikuyama et al., 1995, 1997). Both castration and hypophysectomy reduced the sodefrin content of the abdominal glands and testosterone administration restored it (Yamamoto et al., 1996). Androgens and visual displays

Many species of vertebrates use complex visual displays in intraspecific communication, both in the context of conflict resolution (i.e. aggressive displays) and for mate attraction (i.e. courtship displays) (Bradbury & Vehrencamp, 1998). The evolution of stereotypic species-specific movements suggests that specific neuromuscular systems (i.e. motoneurons and their target muscles) may have evolved specifically for the production of these behaviours. In some bird species, courtship displays involve coordinated wing and leg movements with the individuals on the ground, on perches or in the air (Schlinger et al., 2001). These visual displays are usually sexually dimorphic. Because sex steroids, including androgens, have been shown to play a major role in secondary sex differentiation in most vertebrate species studied so far, they are also potential candidates for a key role in the control of these displays. In wild golden-collared manakins Manacus vitellinus, a tropical arena bird, males perform a courtship display that consists of a sequence of jumps and wing snaps (i.e. upward flips of the bird’s wings that produce an acoustic signal). The feathers involved in the production of these wing snaps are the primary and secondary wing feathers (Schlinger et al., 2001), which are sexually dimorphic (Chapman, 1935). Also the muscles controlling the wing movements and/or feather position and the jump often associated with the wing snap are hypertrophied in male manakins (Lowe, 1942). The muscles involved in the wing-snap movement also show sex differences when examined in more detail (e.g. in fibre diameter, metabolic enzyme activity and myosin isoform expression), which suggests that they are specialized for greater force generation and speed of contraction (Schultz et al., 2001). These sex differences in this neuromuscular system are not present in species in which males do not use these muscles in courtship displays (e.g. zebra finch), although they are still functional for other activities (e.g. for raising and lowering of the wings during flying). These muscles are innervated by motor neurons that accumulate [3 H]-testosterone in their soma in the spinal cord, suggesting a role for androgens in the control of these behavioural mechanisms (Schultz & Schlinger, 1999).

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:12

Hormones, social context and animal communication Another example of an androgen-dependent display is the amplexus behaviour displayed by amphibian males to clasp females during mating. The forelimb muscle involved in this behaviour (i.e. the flexor carpi radiali), is androgen sensitive (Dorl¨ ochter et al., 1994). Castration induces atrophy and testosterone treatment of castrated males causes hypertrophy of some regions of this muscle; immunocytochemistry techniques have identified the presence of androgen receptors (Dorl¨ ochter et al., 1994). Adult males have slower acetylcholine receptor kinetics than females, which facilitates slow and tonic muscle contractions appropriate for the function of this behaviour (Brennan & Henderson, 1995). Moreover, testosterone has been shown to act both at the pre- and postsynaptic level in these neuromuscular junctions, which may be viewed as an adaptation for a more flexible modulation of this behaviour (Nagaya & Herrera, 1995). Finally in fish, androgens induce the development of somatic structures used in visual signalling such as the elongation of the dorsal and anal fins used in lateral displays and the thickening of the jaw used in mouthfighting (e.g. Mozambique tilapia Oreochromis mossambicus: Oliveira & Almada, 1998). Androgens and electrocommunication signals

There are two orders of fish that produce weak electric signals with an electric organ located in their tails: the Gymnotiformes from South America and the Mormyriformes from Africa (Zakon & Smith, 2002). The evolution of these weak electric signals most probably occurred independently in the two orders because they are phylogenetically distant (Alves-Gomes, 1999). Nevertheless, in both orders, this electric sense is used for the same two functions: electrolocation (i.e. locating objects in the environment) and intraspecific communication (Zakon & Smith, 2002). Electrical signals are perceive by the receivers with specialized electroreceptors mainly located in the midline of the fish (Zakon & Smith, 2002). There are two types of electric organ discharges: pulse type and wave type. Each species only produces one or the other (Zakon & Smith, 2002). Within species, there are marked sex differences in the electric organ discharge. In most Gymnotiform species that generate wave-type discharges, the males produce signals of lower frequency than females. For example, in Sternopygus macrurus males produce an electric discharge of 50–90 Hz while female signals range from 100 to 150 Hz (Hopkins, 1972). Sex steroids, in particular androgens, seem to be important in the determination of electric organ discharge frequency. In male S. macrurus, circulating levels of androgens are negatively correlated with frequency (Zakon et al., 1991) and when their reproductive axis was challenged with human chorionic gonadotrophin, they responded with an increase in circulating 11-ketotestosterone levels and a decrease in the frequency of the discharge from their electric organs

487

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

488

0 521 582361 7

April 11, 2005

19:12

R. F. Oliveira (Zakon et al., 1990). Moreover, treatment of wave gymnotiforms with androgens induces a masculinization of the waveform (i.e. higher wave frequency and increased duration (Meyer, 1983; Mills & Zakon, 1987; Dunlap & Zakon, 1998)). Interestingly, 11-ketotestosterone increased the frequency of electric organ discharge (Meyer et al., 1987) in species in which the discharge pattern is sex reversed, that is males generate higher-frequency discharges than females (e.g. brown ghost, Apternotus leptorhynchus, Hagedorn & Heiligenberg, 1985). In all pulse-type species, both mormyriforms and gymnotiforms, the treatment of juveniles, females, castrated males or non-reproductive males with androgens masculinizes the pulse form (Bass & Hopkins, 1983, 1985; Hagedorn & Carr, 1985; Bass & Volman, 1987; Landsman & Moller, 1988; Freedman et al., 1989; Landsman et al., 1990; Herfeld & Moller, 1998). The effects of androgens on the frequency and/or duration of electric organ discharges may be mediated by their effects on the morphology of the electric organ (i.e. size and/or shape of electrocytes) or by an influence on the ionic currents of the electromotor system (e.g. Bass et al., 1986; Bass & Volman, 1987; Mills & Zakon, 1991). Apart from its influence on electric organ discharge parameters, testosterone also activates the onset of electric signalling in weakly electric fish (Landsman & Moller, 1988). Hormonal effects on signal reception

A literature search revealed fewer studies of androgen modulation of sensory perception than of the effects of androgens on effector mechanisms. The four studies below are examples of effects on perception. In many cyprinid fishes, females produce a sex pheromone that elicits male courtship behaviour. The response of males to the female pheromones can be measured either behaviourally or electrophysiologically, by placing electrodes in the olfactory epithelium and measuring the potentials evoked by the exposure of the epithelium to different odorants (i.e. electroolfactograms: Stacey & Sorensen, 2002). In the tinfoil barb Puntius schwanenfeldi, females release a sex pheromone (15-ketoprostaglandin-2α) that stimulates male courtship behaviour (Cardwell et al., 1995). This response is greatest during the breeding season in sexually mature males; such males have visible breeding tubercules, dermal structures that are known to be androgen dependent (Smith, 1974). Moreover, juveniles implanted with androgens (either 11-ketotestosterone or methyltestosterone) show both an increased electroolfactogram response to 15-ketoprostaglandin-2α and increased sexual behaviours directed towards stimuli fish (i.e. juveniles injected with 15-ketoprostaglandin-2α (Cardwell et al., 1995)). These results clearly demonstrate a peripheral effect of androgens on olfactory sensitivity. Other species also show increased olfactory sensitivity to such stimuli during the breeding season when androgen levels are also higher (e.g. electroolfactogram responsiveness to

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:12

Hormones, social context and animal communication testosterone in the Atlantic salmon Salmo salar (Moore & Scott, 1991)), which suggests that the effect described above may be a general phenomenon in fish olfaction. In addition, electroreception in weakly electric fish seems to be modulated by androgens (Keller et al., 1986; Sisneros & Tricas, 2000). Testosterone not only affects the frequency of discharge from electric organs (as described above) but also shifts the maximum receptivity of the electroreceptor to the new frequency produced (Meyer & Zakon, 1982; Bass & Hopkins, 1984). Thus, androgens keep the electroreceptors of a given individual fine-tuned to its own electric organ discharge, which might be viewed as an adaptation for electrolocation. A third example comes from studies of auditory sensitivity in the plainfin midshipman. As mentioned above, in this species type I males produce a humming call during the breeding season that is used to attract spawning females to their nests (Ibara et al., 1983; Brantley & Bass, 1994). Male reproductive success must depend heavily on their calling behaviour because females are choosy regarding call parameters of the ‘hum’ signal (McKibben & Bass, 1998). Female reproductive success is also expected to depend on their ability to locate and choose males based on their acoustic signals. Recently, it has been demonstrated that, during the summer when females need to exert their mate choice preferences based on the male call, the auditory saccular units in the females increase their temporal encoding capacity up to 340 Hz, compared with only 100 Hz in winter females (Sisneros & Bass, 2003). This seasonal plasticity of the peripheral auditory system is most probably driven by sex steroids, because it follows the seasonal variation in steroid profiles (Forlano et al., 2003) and because expression of the oestrogen receptor β has been identified recently in auditory hair cells (P. M. Forlano & A. H. Bass, unpublished data). Therefore, an increase in sex steroids at the beginning of the breeding season may induce changes in the frequency sensitivity of these hair cells in a similar way to androgen-dependent changes in electroreceptor tuning described above. Finally, there are suggestions that sex steroids may also be involved in the modulation of visual perception in teleost fish. In the three-spined stickleback Gasterosteus aculeatus, sexually active females prefer to mate with males with redder bellies (e.g. Milinski & Bakker, 1990). Using optomotor responses, Cronley-Dillon & Sharma (1968) have demonstrated that the sensitivity of the female visual system to red wavelengths increases during the breeding season, suggesting a potential role for female sex hormones. In this example, it can be argued that the effect found could be acting either at the level of the sensory organ or at the level of visual information processing by the central nervous system (i.e. optic tectum). Interestingly, aromatase activity has been found in fish retina, indicating that these cells are actively metabolizing sex steroids (Callard et al., 1993) and supporting

489

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

490

0 521 582361 7

April 11, 2005

19:12

R. F. Oliveira the idea that the steroid modulation of visual sensitivity to key colours may occur in the periphery. These four studies taken together suggest that sex steroid modulation of sensory perception is a common phenomenon in different sensory modalities. Hormonal modulation of motivational and memory mechanisms

Androgens can also affect central mechanisms of information processing both in senders and receivers. At this level, the modulatory action of hormones may affect signalling behaviour by acting either on motivational neural circuits underlying decision-making mechanisms, or on learning and memory systems (Schulkin, 2002; Dohanich, 2002). By acting on central mechanisms, androgens may set up the subject to perceive stimuli and to behave in particular ways, for example by increasing the likelihood of the expression of a given behaviour, ranging from food ingestion to maternal behaviour or aggression. For example, androgens modulate central mechanisms of chemical perception in male hamsters. In this species, vaginal secretions stimulate male sexual behaviour after male anogenital investigation of the female ( Johnston, 1975; Ch. 16). These secretions are detected by two different sensory systems, the olfactory mucosa and the vomeronasal organ, that use different neural pathways converging in three central areas: the medial nucleus of the amygdala, the bed nucleus of the stria terminalis and the medial preoptic area (Scalia & Winans, 1975). Androgen receptors are found in all these three areas (Wood et al., 1992) and direct androgen implantation here restores sexual behaviour in castrated males (Lisk & Bezier, 1980). Usually the effects of steroids, including androgens, on motivational mechanisms involves the regulation of neuropeptide gene expression in the limbic system, namely of arginine-vasopressin (or its homologue arginine-vasotocin in nonmammalian vertebrates), which subsequently influence central states that control the behavioural output (Herbert, 1993). There are numerous examples of this principle. In hamsters, testosterone enhances the effects of arginine-vasopressin infused in the bed nucleus of the stria terminalis on scent-marking behaviour (Albers et al., 1988). In male prairie voles, testosterone also promotes the expression of parental behaviour by increasing arginine-vasopressin synthesis and by preventing the apoptosis of responsive neurons (De Vries, 1995). Finally, in amphibians, sex steroids control both female egg-laying behaviour and male courtship via arginine-vasotocin modulation (Moore et al., 1992). The potential effects of sex steroids on learning, memory and other cognitive functions have been addressed using two main approaches: (a) by documenting the distribution of androgen and oestrogen receptors in brain areas known to be involved in these functions, and (b) by testing hormone-treated subjects in

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:12

Hormones, social context and animal communication cognitive tasks. There is a much larger body of literature on oestrogens than on androgens regarding this topic. The available data on androgens will be summarized below. Androgen receptors are found in the hippocampus of mammals and birds (Kerr et al., 1995; Saldanha et al., 1999) and in the homologue dorsolateral telencephalon of fish (Northcutt & Davis, 1983; Gelinas & Callard, 1997). These brain areas are involved in relational memory processes, namely in spatial memory (Eichenbaum et al., 1992; Squire, 1992). Androgen receptors have also been found in pyramidal cells of the cortex in rats, monkeys and humans (Pomerantz & Sholl, 1987; Kerr et al., 1995; Tohgi et al., 1995). These results set the stage for a potential functional direct effect of androgens on memory mechanisms. The occurrence of oestrogen receptors together with aromatase (an enzyme that metabolizes androgens into oestrogens) also suggests a potential alternative route for aromatizable androgens to affect cognitive function (e.g. Gelinas & Callard, 1997). There are numerous examples of sex differences in spatial memory tasks, with males outperforming females, which suggests a role for sex steroids in spatial memory mechanisms (reviewed by Dohanich, 2002). Early androgen exposure apparently has organizational effects on adult spatial abilities, and the masculinization of spatial learning involves the aromatization of androgens into oestrogens in rodents (Williams et al., 1990; Roof & Havens, 1992; Roof, 1993). In humans, early exposure to androgens masculinizes spatial function, as is suggested by data on girls suffering from congenital adrenal hyperplasia. These girls are exposed to androgens in utero as a result of hypertrophy of the adrenal glands and are born with virilized genitalia. When compared with their unaffected sisters, girls with congenital adrenal hyperplasia have better performances in mental object-rotation tests designed to measure spatial ability (Resnick et al., 1986; see Kimura (1996) for further references). In adults, the relationship between circulating androgen levels and spatial ability is not linear. Lower testosterone levels in males, and higher testosterone levels in females, are associated with better performances in an object-rotation task, which suggests an optimum circulating level of testosterone to excel in this task (Moffat & Hampson, 1996). As regards other cognitive mechanisms, in general the administration of androgens to birds and mammals outside the critical period of development fails to affect learning and memory tasks (Dohanich, 2002). However, social memory is an exception to this rule in rats and zebra finches (Sawyer et al., 1984; Cynx & Nottebohm, 1992). Hormones and somatic releasers

There are a number of somatic structures that act as sign stimuli (sensu Tinbergen, 1951) evoking a behavioural response in conspecifics. The classic

491

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

492

0 521 582361 7

April 11, 2005

19:12

R. F. Oliveira example of these releasers is the red belly of the male three-spined stickleback, which elicits aggressive responses in other male sticklebacks (Tinbergen, 1951). Since initially proposed by Tinbergen, these social releasers have been described in many other species and can range from nuptial colouration patterns in fish and birds to dermal appendages in fish (e.g. dermal tubercules), birds (e.g. combs, elongated tail feathers) and reptiles (e.g. dewlap membrane in Anolis spp.). The development of at least some of these somatic structures with a releaser function is under hormonal control. There are various examples in the teleosts. First, male nuptial colouration in African cichlids is suppressed in castrated males and restored in castrates and females by exogenous administration of testosterone (Levy & Aronson, 1955; Reinboth & Rixner, 1972; Wapler-Leong & Reinboth, 1974; Fernald, 1976). Also in male sticklebacks, the nuptial colouration can be suppressed by castration (Ikeda, 1933) or by the exogenous administration of an anti-androgen (cyproterone acetate) (Rouse et al., 1977). Finally, in the sexrole-reversed peacock blenny Salaria pavo, in which some ‘sneaker’ males mimic female nuptial colouration, androgens (i.e. 11-ketotestosterone) inhibit the expression of female nuptial colouration in these sneaker males (Oliveira et al., 2001a). However, nuptial colouration is not the only releaser to be androgen dependent in fish. The development of the sword as an extension of the caudal fin in male swordtail fish Xiphophorus helleri and the development of the dermal breeding tubercules in male cyprinids are both also induced by testosterone (Baldwin & Goldin, 1939; Smith, 1974). Therefore, another way for hormones to affect communication is by affecting the expression of somatic releasers in senders. Social modulation of androgen levels

As shown above, androgens can be viewed, on the one hand, as causal agents of behaviour, including signalling behaviour among animals in a communication network. On the other hand, the endocrine system is responsive to the network of social relationships in which the animal is involved. Several studies have shown the effects of social interactions on the short-term modulation of androgen levels. In the early 1940s, it was established that male mice that lost an agonistic interaction had lower levels of androgens than winners (Ginsberg & Allee, 1942). This pattern has been found repeatedly in other vertebrate taxa from fish (e.g. Hannes, 1984, 1986) to primates, including humans (e.g. Rose et al., 1971, 1975; Bernstein et al., 1974; Booth et al., 1989; see Mazur & Booth, 1998 for more references). This set of results led to the proposal of the ‘challenge hypothesis’ by John Wingfield and co-workers (Wingfield, 1984; Wingfield et al., 1987, 1990), according to which the social interactions involving the subject determine androgen levels. This hypothesis gives a conceptual framework for the study of the interplay

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:12

Hormones, social context and animal communication between social factors and endocrine responses and generates a number of testable predictions. 1.

2.

3.

Androgen levels should be higher during periods of social instability when social interactions are more frequent and more intense. In fact, in bird species in which a clear breeding cycle can be recognized, testosterone levels are higher during the period of territory establishment than when territories are established (Hegner & Wingfield, 1987a; see Wingfield et al., 1999, 2000 for more examples). Territorial and dominant males are expected to show higher androgen levels than non-territorial or subordinate males because territorial males have to defend their territories from intruders and dominant males have actively to maintain their status. Again the available evidence supports this hypothesis (e.g. see Oliveira et al. (2002) for a review of teleost fish and Wingfield et al. (1999, 2000) for reviews of birds). Populations of the same species breeding under different population-density regimes should also show differences in the average androgen levels of breeding males as a result of a different probability of territory intrusions. This prediction should be taken with caution because in a population with increased density, physiological and/or behavioural mechanisms may be present to avoid aggression. Nevertheless, positive correlations have been found between density of breeding territories and androgen levels both in fish and in birds (e.g. Ball & Wingfield, 1987; Beletsky et al., 1990, 1992; Pankhurst & Barnett, 1993).

Interestingly, during periods of social inertia, the levels of social interaction fall to a baseline and androgen levels become decoupled from social behaviour. These results have been interpreted as an adaptation (or an exaptation sensu Gould & Vrba (1982), depending on the underlying historical evolutionary pathway) for the individuals to adjust their behaviour (motivation) to the social milieu that they are currently experiencing. Thus, social interactions would stimulate the production of androgens and androgen levels would be a function of the stability of the social environment in which the animal is living (Wingfield et al., 1990, 1999, 2000; Oliveira et al., 2002). It is interesting to note here that it is the perception that the individual has of the interaction in which it is involved or which it is observing that activates the endocrine response and not the objective structure of the situation per se. To investigate this idea we have recently tested the effect of mirror-elicited aggression on androgen levels in a cichlid fish (L. Carneiro & R. F. Oliveira, unpublished data). The mirror image stimulation test is widely used in fish ethology to assess

493

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

494

0 521 582361 7

April 11, 2005

19:12

R. F. Oliveira aggressiveness (Rowland, 1999), but some inconsistencies have been found in the relationship between social status and the aggressive score of an individual in this test (Ruzzante, 1992). In Mozambique tilapia, we showed that androgen levels before the fish were grouped were not good predictors of social status, but androgen levels at the end of the time spent in a group were highly correlated with the social status of each individual, suggesting that androgens are being modulated by the social interactions experienced by the grouped individuals (Oliveira et al., 1996). In the mirror image stimulation test, the individual is placed in a very peculiar situation. Because fish do not recognize as themselves the image reflected by the mirror, they respond to it as an intruder and attack. In our experiment, males reacted aggressively to their own images in the mirror and escalated the interaction using more overt aggressive behaviours (e.g. biting) as time went by. However, because the mirror reflects exactly the same behaviours that the experimental fish is displaying, the interaction has no outcome (winning versus losing). Therefore, if the endocrine response to the social interaction is triggered by the behavioural output during the interaction (e.g. number of displays or time spent displaying), a variation in androgen levels is predicted. However, if it depends on behavioural feedback received from the opponent, then no androgen variation is predicted in the test. In our mirror image stimulation experiment with Mozambique tilapia, we found a strong behavioural response but a complete lack of an androgen response (L. Carneiro & R. F. Oliveira, unpublished data), which suggests that the endocrine system responds to a clear perception of the outcome of the social interaction. This result is also interesting because it shows that it is the communication component of the social interaction that may affect hormone levels.

Hormones and communication II: the network view In the previous section, the interrelationship between hormones (i.e. androgens) and social behaviour was considered at the dyadic level. First, the mechanisms through which androgens may affect communication between a pair of individuals were described. Second, the ways in which these androgen levels might be modulated by the social environment (network) in which the animal is living were described. If we now consider an interaction that occurs within a social network (or a communication network sensu (McGregor, 1993)), with possibilities for other individuals in the network to eavesdrop on the interaction (e.g. Ch. 2) and for the interacting pair to adjust their behaviour according to the presence of an audience (Ch. 4), the complexity of the interrelationship between hormones and behaviour/communication mechanisms could increase substantially. The presence

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:12

Hormones, social context and animal communication

Receiver Input/ perception

Sender

Central processing

Motor output Input/ perception

Central processing

Hormones

Signal Motor output

Bystander

Hormones Somatic releasers

Input/ perception

Central processing

Motor output

Hormones Fig. 21.3. Flow of information in a communication network. This figure is similar to Fig. 21.2 with the following extra elements. The presence of a third individual (the bystander) may be perceived by both the sender and the receiver and affect their signal production and signal reception mechanisms, respectively. The perception of the presence of the bystander may also affect central (motivational) mechanisms in the central nervous system in both senders and receivers, which may, in turn, modulate hormone levels in both individuals. The perception of the signal by the bystander would also affect its central processing of information at the level of motivational mechanisms, which could affect its hormone levels.

of a bystander that could act both as an eavesdropper and as an audience (Fig. 21.3), may affect androgen levels in both the sender and the receiver and subsequently affect their androgen-modulated communication behaviour in the same ways as described above (see also Fig. 21.2). The androgen levels of the bystander itself may respond to the observed interaction, which, in turn, will affect its own subsequent social behaviour. Consequently, androgens may play a key role as physiological mediators of the modulation of behaviour by the social context. A number of social phenomena (e.g. winner–loser effects, bystander effect, dear enemy effect) that have been described in social networks may then be physiologically mediated by changes in hormone levels, especially androgens.

495

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

496

0 521 582361 7

April 11, 2005

19:12

R. F. Oliveira Adjusting behaviour to the social context: a role for androgens? Winner–loser effects

There is an extensive literature (including most of the chapters of this book) that clearly shows that animals use information on relative competitive abilities in the social network in which they are placed to adjust their behaviour accordingly. They may obtain this information by direct assessment of their peers by interacting with one another in a dyadic fashion and then adjusting their behaviour in subsequent interactions depending on the outcome of previous interactions. For example, individuals that win an interaction increase their probability of winning a subsequent interaction and vice versa for losers. In this case, although only two individuals have to be present during the initial interaction, unless there were other individuals with whom the interactants subsequently interacted, there would be no winner–loser effect. Therefore, this effect is better understood within the framework of social networks than with a dyadic approach. This winner–loser effect may last from a few minutes up to several hours or even days and has been reported for several taxa, for example invertebrates (Alexander, 1961; Otronen, 1990; Whitehouse, 1997), fish (McDonald et al., 1968; Frey & Miller, 1972; Bakker & Sevenster, 1983; Francis, 1983, 1987; Abbott et al., 1985; Beaugrand & Zayan, 1985; Beacham & Newman, 1987; Franck & Ribowski, 1987; Beacham, 1988; Bakker et al., 1989; Beaugrand et al., 1991, 1996; Chase et al., 1994; Hsu & Wolf, 1999), reptiles (Schuett, 1997) and birds (Drummond & Os´ orio, 1992). The winner effect is usually of shorter duration than the loser effect (e.g. Chase et al., 1994), and when integrating prior social experiences more recent outcomes are more effective in predicting the probability of winning a subsequent interaction than previous ones (Hsu & Wolf, 1999). Another interesting characteristic of this effect is that it is more effective when winning or losing against a well-matched opponent than when there is a large asymmetry in resource-holding potential (sensu Parker, 1974) between the two individuals (Beaugrand & Goulet, 2000). The behavioural mechanism proposed to explain the winner effect is based on the fact that initiators of interactions have higher probabilities of winning and that winners of recent encounters become more likely to initiate future interactions ( Jackson, 1991). This is especially true for the initiators of attacks (Hsu & Wolf, 2001). It is conceivable that by winning an interaction an individual raises its androgen levels, which, in turn, increases its willingness to initiate future interactions and the probability of winning the next interaction in which it participates. The reverse would be predicted for losers. We are conducting an ongoing literature survey to collect data on the two steps of this endocrine hypothesis for the winner–loser effect: (a) that winners

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:12

Hormones, social context and animal communication Table 21.1. Literature survey of reported differences in male androgen levels among vertebrates according to social status and the phase of the sexual cycle and of the effects of androgen treatment on aggressive behaviour Taxa

Androgens and

Androgens and

Effect of androgen

social status

phase of

treatment on

sexual cycle

aggressive behaviour

D = S or No. Fish Amphibians Reptiles

D>S

D<S

MP < PP or No.

MP > PP

No

MP = PP

No.

Effect

effect

12

9

3

9

8

1

12

7

5

1

0

1

2

2

0

0

–

–

3

2

1

0

–

–

3

3

0

Birds

10

4

6

48

47

1

10

6

4

Mammals

18

15

3

4

3

1

9

5

4

Total

44

30

14

63

58

5

34

21

13

Male androgen levels in; D, dominant; S, subordinate; MP, mating phase; PP, parental care phase. K. Hirschenhauser & R. F. Oliveira, unpublished data.

have higher androgen levels than losers; (b) that androgens increase aggressive behaviour and hence the probability of victory in a subsequent interaction. As there are not enough studies that we can find that measured the androgen variations in response to a social interaction to address the first step, it was decided to search for correlational data in the form of reported androgen differences between dominant and subordinate individuals. We found 44 published studies, 68% of which confirmed that androgen levels were higher in dominants than in subordinates (Table 21.1). Our literature survey revealed that 62% of the studies confirmed that administration of androgens increased aggressive behaviour in different taxa (Table 21.1), thereby supporting the second step of the endocrine hypothesis. Although the majority of the studies supported the assumptions of the proposed hypothesis, the percentages do not provide overwhelming support and so we decided to test this hypothesis experimentally with the Mozambique tilapia (A. Silva & R. F. Oliveira, unpublished data). After staging a first fight between two males, the winner and the loser fought two independent, naive individuals (i.e. males that have not been involved in social interactions recently) (Fig. 21.4a). As expected, our preliminary data showed that winners of the first encounter won the majority of the interactions with the naive fish and vice versa for losers (Fig. 21.4b). When winners were treated with an anti-androgen (cyproterone acetate) between the two interactions (which were two hours apart), the winner effect was no longer detectable in the second fight with the neutral fish, suggesting an involvement

497

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

April 11, 2005

19:12

R. F. Oliveira

(a)

(b)

t1

Neutral male

Winner

t2

Loser

t2

Neutral male

100 90 80 70 60 50 40 30 20 10 0

Percentage

498

0 521 582361 7

control treated

Winner

Loser

Fig. 21.4. Androgens and the winner–loser effect. (a) Experimental-set up: four Mozambique tilapia Oreochromis mossambicus males were introduced to individual compartments separated by opaque partitions in an aquarium. After a period of acclimation (t1 ), the central partition was removed and the two individuals in the central compartments were allowed to interact until a winner and a loser could be recognized. Then the partition was put back in place. Two hours later (t2 ), the two lateral partitions were simultaneously removed and the winner and the loser of the previous interaction were allowed to interact with the neutral males that were placed in the end compartments. This second interaction went on until a winner and a loser could be recognized. (b) Percentage of second interactions won by winners or losers of the first interaction. Four groups of experimental animals were compared: t1 winners treated with an androgen inhibitor (cyproterone acetate); control t1 winners (treated with a placebo saline solution); t1 losers treated with an androgen (11-ketotestosterone); control t1 losers (treated with a placebo saline solution). Twelve replicates of the experiment were run. (A. Silva & R. F. Oliveira, unpublished data.)

of androgens in the winner effect (Fig. 21.4b). However, the loser effect was not inhibited in the second interaction by treating losers with exogenous androgens, which suggests that, although a fall in androgens is observed in losers, it is not the underlying causal mechanism for the loser effect. Although this result may seem paradoxical at first sight, it makes some sense; androgen variations induced by social interactions occur in the short term and so do winner effects; however, the loser effect may last up to several days depending on a number of factors. Consequently, other neuroendocrine mechanisms must be involved in the loser effect. One of the best candidates for this role is the serotonergic system. The following evidence from studies using different fish species seems to support this hypothesis: (a) losers experience increased brain levels of serotonin and subordinate individuals have chronically elevated brain levels of serotonin (Winberg &

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:12

Hormones, social context and animal communication Nilsson, 1993a,b; Winberg et al., 1997; Winberg & Lepage, 1998); and (b) serotonin appears to be inhibitory to behavioural responsiveness in general and to inhibit aggressive behaviour in particular (Winberg & Nilsson, 1993a,b; Adams et al., 1996; Edwards & Kravitz, 1997). Therefore, losers would display a marked behavioural inhibition, with increased attack latencies in subsequent interactions, which would prevent them from winning these interactions and would reinforce their subordinate role. Interestingly, the administration of a precursor of dopamine (L-dopa) to individuals that had lost an interaction two days before induced lower serotonergic activity and reduced the attack latency in subsequent interactions, suggesting that the dopaminergic system counteracts the serotonin-mediated effects of social subordination (H¨ oglund et al., 2001). Bystander effects

Information on the relative competitive ability of conspecifics within a social network can also be gathered using indirect methods, namely by extracting information from watching conspecific interactions that the subject uses in subsequent interactions with the observed individuals (eavesdropping (McGregor, 1993; McGregor & Peake, 2000) and social eavesdropping sensu Peake, Ch. 2). This sort of information gathering on the relative ability of conspecifics has been demonstrated in a number of species (see Ch. 2), for example fish ( Johnsson & Åkerman, 1998; Oliveira et al., 1998; Earley & Dugatkin, 2002; Ch. 5) and birds (Hogue et al., 1996; McGregor et al., 1997; Naguib et al., 1999; Peake et al., 2001), and has the advantage of avoiding the costs associated with fighting (e.g. McGregor, 1993; McGregor & Peake, 2000; Dugatkin, 2001). Some authors consider that there is a difference between bystander and eavesdropping effects: eavesdropping implies an active gathering of information by bystander individuals that will be used in future interactions within the social network (McGregor, 1993; McGregor & Peake, 2000), while the bystander effect was originally described as a priming of aggressive motivation in bystanders of agonistic interactions (Hogan & Bols, 1980; Bronstein, 1989). Therefore, from the point of view of the required cognitive abilities, eavesdropping is expected to be more demanding than a mere priming response. However, both phenomena are adaptive because they might increase the probability of eavesdroppers/bystanders of winning their next social interaction (Clotfelter & Paolino, 2003; Hollis et al., 1995; Peake & McGregor, 2004). The priming response associated with the bystander effect is another phenomenon that could be mediated by androgens. To investigate if bystanders experience an increase in their androgen levels, we conducted an experiment with Mozambique tilapia in which a bystander fish had visual access through a one-way mirror to two conspecific neighbours separated by an opaque partition

499

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

April 11, 2005

19:12

R. F. Oliveira

(a)

Male A Male C Male B One way mirror

(b) 11-Ketotestosterone variation (ng/ml)

500

0 521 582361 7

6 4 2 0 −2 −4 −6 −8 −10 −12

p < 0.05 p < 0.01

30 min

p < 0.01

2h

6h

Time after exposure to stimuli Fig. 21.5. Social modulation of androgen levels in bystander male Mozambique tilapia Oreochromis mossambicus. (a) Experimental-set up: three males were introduced to individual compartments in an aquarium. Males A and B were separated by an opaque partition and male C was separated from the other two males by a one-way mirror that allowed it to observe the other two males without being observed. After a period of acclimation, two conditions were created. (i) In the experimental group, the opaque partition separating males A and B was removed (grey arrow) and the two individuals were allowed to interact for 20 minutes while male C observed the interaction (bystander). (ii) In the control group, the opaque partition separating males A and B was not removed and the bystander individual observed its two neighbours resting or swimming around for 20 minutes. Urine samples were collected from male C at regular intervals (just before the start of the test, 30 minutes after, two hours after and six hours after the experiential situation) and assayed for androgens using radioimmunoassays. (b) Androgen (11-ketotestosterone) variation (i.e. urine concentrations after the experiential situation minus the urine concentrations just before the experiential situation) in bystander males of the experimental (black bars) and control (white bars) groups. (Adapted from Oliveira et al., 2001b.)

(Fig. 21.5; Oliveira et al., 2001b). After a period of familiarization, in the experimental treatment the opaque partition between neighbours was removed and the bystander was allowed to observe the agonistic interaction between its neighbours. In the control group after the same period of familiarization, the opaque partition between neighbours remained in place and the bystander could see its

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:12

Hormones, social context and animal communication two neighbours resting or swimming around in their respective compartments. As predicted, androgen levels (both 11-ketotestosterone and testosterone) increased significantly in the experimental group of bystanders after watching their neighbours fighting and no effect was detected in the control group (Oliveira et al., 2001b). This result has an interesting parallel in humans. It has been demonstrated that sports fans experience variations in testosterone levels depending on the outcome of the game they have attended, both for college basketball and for soccer. Fans of the winning team display an increase in salivary testosterone levels and there is a decrease in testosterone levels in fans of the losing team (Bernhardt et al., 1998). Audience effects

The term audience effect was first used in the ethological literature to describe the facilitation effect of the presence of other individuals on the production of food calls or alarm calls in response to food items or a predator, respectively (Gyger et al., 1986; Marler et al., 1986; Evans & Marler, 1994). Here the term will be used in a more restricted way, following the definitions provided by McGregor & Peake (2000) and by Matos (2002) (see also Ch. 4): individuals participating in an interaction may also manipulate the information available to others and adjust their signalling behaviour according to the presence and composition of an audience of conspecifics. These audience effects have been demonstrated in different vertebrate taxa, including fish (e.g. Doutrelant et al., 2001; Matos & McGregor, 2002), birds (e.g. Searcy et al., 1991; Baltz & Clark, 1997) and mammals (e.g. Hector et al., 1989), and have involved different social contexts, from agonistic interactions in Siamese fighting fish Betta splendens (Doutrelant et al., 2001; Matos & McGregor, 2002) to extra-pair copulations in male budgerigars Melopsittacus undulatus (Baltz & Clark, 1997). The audience effect has been interpreted as a way for the individual to manipulate the information broadcast to its social network, which may influence subsequent social interactions in which it will have to participate. Therefore, it can be predicted that subjects behave more promptly and aggressively towards an intruder when a male audience is present. Again it is predicted that this effect may be mediated by increased androgen levels in the interacting individuals induced by the presence of the audience. This aggressive priming effect of an audience has already been established in Siamese fighting fish (Matos, 2002) but its androgen-mediation remains to be tested. Dear enemy effects

In territorial systems, residents react less aggressively towards familiar opponents than to intrusions by strangers, a phenomenon called the dear

501

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

502

0 521 582361 7

April 11, 2005

19:12

R. F. Oliveira enemy effect (Ydenberg et al., 1988; Temeles, 1994). In evolutionary terms, this phenomenon can be viewed as an adaptation for the individual to adjust its territorial behaviour according to the threat posed by the intruder (Temeles, 1994): having a dear enemy neighbour allows the resident individual to defend its territory against unfamiliar intruders with the same efficiency as if they were the only competitors in the area, which reduces the costs of territory defence (Leiser & Itzkowitz, 1999; Whiting, 1999). The dear enemy phenomenon can be explained in terms of proximate mechanisms by an ability of the resident male to discriminate between familiar and unfamiliar intruders together with a habituation to the neighbours, which would explain the lower response that they elicit, for example visual habituation to neighbours in Siamese fighting fish (Bronstein, 1994) and habituation to neighbours’ calls in frogs (Owen & Perrill, 1998). Therefore, it can be predicted that resident males will react more aggressively towards strangers than towards familiar intruders and that the increase in androgen levels expected from the social challenge experienced will be higher in the case of intrusions by strangers. Moreover, it is also predicted that, for repeated intrusions by neighbouring males, the androgen response should be higher in the first trials and decrease with the number of trials (i.e. habituation). These two predictions remain to be tested.

The adaptive value of social modulation of hormones: a cost–benefit analysis of androgen levels As stated above, the main adaptive reason for androgens to respond to the social environment is to allow individuals to fine-tune the expression of their behaviours in a context-dependent fashion. For example, this mechanism would allow subordinate individuals to downregulate the expression of their aggressive behaviour and thus avoid the initiation of agonistic encounters that they have low probabilities of winning. In the long run, this mechanism can be seen as an opportunity for individuals to adopt a behavioural tactic that suits best their relative competitive ability. As a result, androgen-mediated behavioural tuning to the social environment may result in either a continuous or a discrete variation of behavioural phenotypes. For example, even small changes in androgen levels induced by social interactions in electric fish, can affect the pulse duration, resulting in dominant males with more masculinized discharges from their electric organs than subordinates (e.g. Brienomyrus brachyistius: Carlson et al., 2000). Also, in the Mozambique tilapia, the acquisition of dominant status induces the exaggeration of male morphological traits, an effect that has been shown to be mediated by androgens (Oliveira & Almada, 1998). By comparison, in a number

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:12

Hormones, social context and animal communication of teleost species, individuals of lower competitive ability can adopt frequencyor condition-dependent alternative reproductive tactics (Taborsky, 1994) or even change sex (Grober, 1998). However, it can be argued that, instead of having their androgen levels open to social influences, selection could have favoured animals that permanently keep their androgen levels at an optimum high value in order to optimize their social behaviour at all times. It follows that there must be costs associated with maintaining high levels of androgens that counteract the social benefits of high androgen levels. Therefore, a cost–benefit analysis is needed to establish the adaptive value of the social modulation of androgens. Potential benefits of high androgen levels

Among the potential benefits of increasing androgen levels at periods of social challenge, one can think of androgen effects both on aggressive motivation and on cognitive tasks that would promote the success of the animals in social interactions. The available data on the effects of androgens on aggressive motivation has already been review above, and in most studies an effect has been found (Table 21.1). Sex steroids, including androgens, are known to play a major role in cognitive processes such as social attention, learning and memory in a variety of vertebrate taxa (e.g. Andrew, 1991; Cynx & Nottebohm, 1992) and so they may help the animal to be prepared for a competitive context (see text above for more references). We have recently tested the effects of androgens on social attention in Siamese fighting fish. Eavesdropping has already been demonstrated in this species and male Siamese fighting fish are known to spend time observing conspecific interactions (Oliveira et al., 1998). So we designed an experiment to assess the effect of the administration of exogenous androgens on the time males spend observing social interactions between conspecific males. Not surprisingly, androgen-treated males spent more time observing social interactions than controls, suggesting an effect of androgens on selective attention to the social environment (R. F. Oliveira & L. Carneiro, unpublished data). Another potential benefit that androgens may convey in a competitive situation is an increased probability of the expression of risk-taking behaviours, which might be adaptive in a competitive situation. A nice example of this phenomenon has recently been published (Kavaliers et al., 2001). Male mice were pre-exposed to the odour of an oestrous female and subsequently exposed to the odours of predators (cat and weasel). Mice that were only exposed to the predator odour, simulating a situation of increased predation risk, showed increased circulating

503

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

504

0 521 582361 7

April 11, 2005

19:12

R. F. Oliveira levels of corticosterone and decreased levels of testosterone. The pre-exposure to the female odour attenuated this response to predator odour, which might reflect a greater tendency for risk taking in the presence of predators (Kavaliers et al., 2001). Potential costs of high androgen levels

Elevated androgen levels have been shown to have associated costs; consequently, one would expect high circulating levels to be restricted to periods of social challenge. The following potential costs have been discussed in the literature: (a) increased energy consumption; (b) impairment of immunocompetence; (c) higher rates of injuries and reduced survival; (d) interference with parental care; and (e) potential oncogenic effects (Wingfield et al., 1999). Of all these potential costs, two will be analysed in more detail below: the potential negative effects of androgens on metabolic rate and on parental behaviour. Metabolic costs of high androgen levels

Studies of the metabolic effects of androgens have produced contradictory results. In bird species, testosterone treatment increased the basal metabolic rate in house sparrows Passer domesticus (Buchanan et al., 2001), reduced it in whitecrowned sparrows Zonotrichia leucophrys (Wikelsky et al., 1999) and had no effect in dark-eyed juncos Junco hyemalis (Deviche, 1992). However, in juncos, an independent study found an association between high testosterone levels and increased lipid catabolism and nocturnal body temperature (Vezina & Thomas, 2000). In the lizard Sceloporus jarrovi, testosterone treatment increased the maximal metabolic rate but had no effect on basal metabolic rate (Marler et al., 1995) and male tilapia treated with 11-ketotestosterone showed an increase in the resting metabolic rate and in metabolic scope but a non-significant increase in the basal metabolic rate (Ros et al., 2004). This discrepancy in the results can be attributed to methodological variations among studies, including the choice of the measures taken, the timespan of the experiment, the season, etc. Nevertheless, androgens failed to affect the metabolic measure used in only two studies, and in one case the data are contradicted by a subsequent study on the same species. In the other four studies, androgens affected different metabolic measures. Consequently, it can be said that androgens may affect metabolism in a non-linear way and a metabolic cost associated with higher levels of androgens should not be excluded. Parental care trade-off with androgens

One of the predictions of the challenge hypothesis is that male androgen levels above a breeding baseline are incompatible with male parental care

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:12

Hormones, social context and animal communication (Wingfield et al., 1990). If androgen levels increase as a result of social challenges, males will invest less time in paternal activities, and thus a trade-off between social interactions and paternal care, mediated by androgens emerges. In many bird species with male parental care, the experimental increase of circulating testosterone in parental males suppressed paternal behaviour and promoted agonist interactions (Silverin, 1980; Hegner & Wingfield, 1987b; Ketterson et al., 1992; Beletsky et al., 1995). Moreover, several studies on the seasonal variation of androgen levels in birds show that during the breeding season male androgen levels are higher during the mating phase than during the parental phase (Wingfield et al., 1987). To document this trade-off further, we have gathered published data on androgen levels in vertebrate species with respect to paternal care: out of the 63 species of vertebrates for which data are available, 92% show the expected pattern of lower circulating androgen concentrations during the parental phase (Table 21.1). In summary, keeping high levels of androgens at all times is detrimental to the individual in many ways and so the stage is for the evolution of a flexible system modulated by the social environment.

Summary and future directions Androgen modulation by social context and the subsequent role of androgens in the activation of expression of social behaviour have been proposed in this chapter to explain the mechanisms underlying experiential effects. However, this hypothesis does not exclude explanations of the phenomena described, in terms of associative learning mechanisms. Cognitive abilities such as individual recognition and discrimination would explain some of the described behavioural responses to social context (e.g. McDonald et al., 1968) and winning or losing can be seen as having reinforcing properties. For instance, male Siamese fighting fish will perform an operant response to have access to an opponent that they can subsequently fight (Hogan, 1967; Bols, 1977). Similar results have been reported for mice (Tellegen et al., 1969), suggesting that the opportunity to interact with an opponent may be a universal positive reinforcer in vertebrates. However, the two explanations (i.e. endocrine modulation and associative learning) should not be seen as mutually exclusive but as complementary, and it is even possible that they represent two levels of analysis that are tightly interconnected. Conditioning of the endocrine response by social stimuli is a possibility that remains to be tested, and there are already examples of androgen modulation of learning mechanisms (e.g. in castrated zebra finches testosterone facilitates conspecific song discrimination (Cynx & Nottebohm, 1992)). Therefore, the interrelationship between androgens and associative learning mechanisms is certainly a key topic for future research in this area.

505

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

506

0 521 582361 7

April 11, 2005

19:12

R. F. Oliveira Acknowledgements I wish to thank all the present and past members of the ‘Mackerel Academy’ (what other name could a fish ethology group have in a Psychology School?) for all the reading club discussions and the Monday seminars that generated hypotheses, shaped experimental designs and dissected the results of some of the research reported here. They are in alphabetical order: R. Andrade, K. Becker, L. Carneiro, N. Castro, T. Fagundes, D. Gonc¸alves, K. Hirschenhauser, J. Jord˜ ao, M. Lopes, T. Oliveira, A. Ros, J. Saraiva and A. Silva. I also thank P. McGregor, R. Bshary and two anonymous reviewers for providing helpful comments on an earlier version of the manuscript. I would like to dedicate this chapter, as a posthumous expression of thanks, to the memory of the late Luis Carneiro (b. 18 January 1969, d. 12 September 2002). More than a PhD student with a promising career, Luis was a beloved friend and his humour and attitude towards life made him an example of how intensely life can, and should, be lived. The unpublished studies reported ˜o para a Ciˆencia e a Tecnologia – here were funded by two ongoing research grants from Fundac¸a FCT (PRAXIS XXI/P/BIA/10251/1998 and POCTI/BSE/38395/2001). The writing of this chapter was partially funded by the Plurianual Program from FCT (UI&D 331/94).

References Abbott, J. C., Dunbrack, R. L. & Orr, C. D. 1985. The interaction of size and experience in dominance relationships of juvenile steelhead trout (Salmo gairdneri). Behaviour, 92, 241–253. Adams, C. F., Liley, N. R. & Gorzalka, B. B. 1996. PCPA increases aggression in male firemouth cichlids. Pharmacology, 53, 328–330. Albers, H. E., Liou, S. Y. & Ferris, C. F. 1988. Testosterone alters the behavioural response of the medial preoptic–anterior hypothalamus to microinjection of arginine vasopressin in the hamster. Brain Research, 456, 382–386. Alexander, R. D. 1961. Aggressivness, territoriality and sexual behaviour in field crickets. Behaviour, 17, 130–223. Alves-Gomes, J. 1999. Systematic biology of gymnotiform and mormyriform electric fishes: phylogenetic relationships, molecular clocks and rates of evolution in the mitochondrial rRNA genes. Journal of Experimental Biology, 202, 1167–1183. Andrew, R. J. 1991. Testosterone, attention and memory. In: The Development and Integration of Behaviour, ed. P. Bateson. Cambridge, UK: Cambridge University Press, pp. 171–190. Arnold, A. P. 1975. The effects of castration and androgen replacement on song, courtship and aggression in zebra finches (Poephilia guttata). Journal of Experimental Zoology, 191, 309–326. Bakker, T. C. M. & Sevenster, P 1983. Determinants of dominance in male sticklebacks (Gasterosteus aculeatus L.). Behaviour, 86, 55–71. Bakker, T. C. M., Bruijn, E. & Sevenster, P. 1989. Asymmetrical effects of prior winning and losing on dominance in sticklebacks (Gasterosteus aculeatus). Ethology, 82, 224–229.

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:12

Hormones, social context and animal communication Baldwin, F. M., Goldin, H. S. 1939. Effects of testosterone propionate on the female viviparous teleost, Xiphophorus helleri Heckel. Proceedings of the Society for Experimental Biology and Medicine, 42, 813–819. Ball, G. F. & Wingfield, J. C. 1987. Changes in plasma luteinizing hormone and sex steroid hormones in relation to multiple-broodness and nest-site density in male starlings. Physiological Zoology, 602, 191–199. Balthazart, J., Foidart, A., Wilson, E. M. & Ball, G. F. 1992. Immunocythochemical localization of androgen receptors in the male songbird and quail brain. Journal of Comparative Neurology, 317, 407–420. Baltz, A. P. & Clark, A. B. 1997. Extra-pair courtship behaviour of male budgerigars and the effect of an audience. Animal Behaviour, 53, 1017–1024. Bass, A. H. & Hopkins, C. D. 1983. Hormonal control of sexual differentiation: changes in electric organ discharge wave-form. Science, 220, 971–974. 1984. Shifts in frequency tuning in electroreceptors in androgen-treated mormyrid fish. Journal of Comparative Physiology A, 155, 713–724. 1985. Hormonal control of sex differences in the electric organ discharge (EOD) of mormyrid fishes. Journal of Comparative Physiology A, 156, 587–604. Bass, A. H. & Volman, S. F. 1987. From behavior to membranes: testosterone-induced changes in action potential duration in electric organs. Proceedings of the National Academy of Sciences, USA, 84, 9295–9298. Bass, A. H., Denizot, J. P. & Marchaterre, M. A. 1986. Ultrastructural features and hormone-dependent sex differences of mormyrid electric organs. Journal of Comparative Neurology, 254, 511–528. Beacham, J. L. 1988. The relative importance of body size and aggressive experience as determinants of dominance in pumpkinseed sunfish, Lepomis gibbosus. Animal Behaviour, 36, 621–623. Beacham, J. L. & Newman, J. A. 1987. Social experience and the formation of dominance relationships in the pumpkinseed sunfish, Lepomis gibbosus. Animal Behaviour, 35, 1560–1562. Beani, L., Panzica, G., Briganti, F., Persichella, P. & Dess`ı-Fulgheri, F. 1995. Testosterone-induced changes of call structure, midbrain and syrinx anatomy in partridges. Physiology and Behavior, 58, 1149–1157. Beaugrand, J. P. & Goulet, C. 2000. Distinguishing kinds of prior dominance and subordination experiences in males of green swordtail fish (Xiphophorus helleri). Behavioural Processes, 50, 131–142. Beaugrand, J. P. & Zayan, R. 1985. An experimental model of aggressive dominance in Xiphophorus helleri (Pisces, Poeciliidae). Behavioural Processes, 10, 1–52. Beaugrand, J. P., Goulet, C. & Payette D 1991. Outcome of dyadic conflict in male green swordtail fish, Xiphophorus helleri: effects of body size and prior dominance. Animal Behaviour, 41, 417–424. Beaugrand, J. P., Payette, D. & Goulet, C. 1996. Conflict outcome in male green swordtail fish dyads (Xiphophorus helleri): interaction of body size, prior dominance/subordination experience, and prior residency. Behaviour, 133, 303–319.

507

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

508

0 521 582361 7

April 11, 2005

19:12

R. F. Oliveira Beletsky, L. D., Orians, G. H. & Wingfield, J. C. 1990. Steroid hormones in relation to territoriality, breeding density, and parental behavior in male yellow-headed blackbirds. The Auk, 107, 60–68. 1992. Year-to-year patterns of circulating levels of testosterone and corticosterone in relation to breeding density, experience, and reproductive success of the polygynous red-winged blackbird. Hormones and Behavior, 26, 420–432. Beletsky, L. D., Gori D. F., Freeman, S. & Wingfield, J. C. 1995. Testosterone and polygyny in birds. In: Current Ornithology, Vol. 12, ed. D. M. Power. New York: Plenum Press, pp. 1–41. Belvedere, P., Colombo, L., Giacoma, C., Malacarne G. & Andreoletti, G. E. 1988. Comparative ethological and biochemical aspects of courtship pheromones in European newts. Monitore Zoologico Italiano, 22, 397–403. Bernhardt, P. C., Dabbs, J. M., Fielden, J. A. & Lutter, C. D. 1998. Testosterone changes during vicarious experiences of winning and losing among fans at sports events. Physiology and Behavior, 65, 59–62. Bernstein, I. S., Rose, R. M. & Gordon T. P. 1974. Behavioral and environmental events influencing primate testosterone levels. Journal of Human Evolution, 3, 517–525. Bleisch, W., Luine, V. N. & Nottebohm, F. 1984. Modification of synapses in androgen-sensitive muscle. I. Hormonal regulation of acetylcholine receptor number in the songbird syrinx. Journal of Neuroscience, 4, 786–792. Bols, R. J. 1977. Display reinforcement in the Siamese fighting fish, Betta splendens: aggressive motivation or curiosity? Journal of Comparative and Physiological Psychology, 91, 233–244. Booth, A., Shelley, G., Mazur, A., Tharp, G. & Kittok, R. 1989. Testosterone, and winning and losing in human competition. Hormones and Behavior, 23, 556–571. Bradbury, J. W. & Vehrencamp, S. L. 1998. Principles of Animal Communication. Sunderland, MA: Sinauer. Brantley, R. K. & Bass, A. H. 1994. Alternative male spawning tactics and acoustic signals in the plainfin midshipman fish, Porichthys notatus (Teleostei, Batrachoididae). Ethology, 96, 213–232. Brantley, R. K., Marchaterre, M. A. & Bass, A. H. 1993. Androgen effects on vocal muscle structure in a teleost fish with inter- and intra-sexual dimorphisms. Journal of Morphology, 216, 305–318. Brenowitz, E. A. & Arnold, A. P. 1989. Accumulation of estrogen in a vocal control brain region of a duetting songbird. Brain Research, 359, 364–367. 1992. Hormone accumulation in song regions of the canary brain. Journal of Neurobiology, 23, 871–880. Brennan, C. & Henderson, L. P. 1995. Androgen regulation of neuromuscular junction structure and function in a sexually dimorphic muscle of the frog Xenopus laevis. Journal of Neurobiology, 27, 172–188. Bronstein, P. M. 1989. The priming and retention of agonistic motivation in male Siamese fighting fish, Betta splendens. Animal Behaviour, 37, 165–166.

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:12

Hormones, social context and animal communication 1994. On the predictability, sensitisation, and habituation of aggression in male bettas (Betta splendens). Journal of Comparative Psychology, 98, 421–431. Brown, R. E. 1978. Hormonal control of odor preferences and urine-marking in male and female rats. Physiology and Behavior, 20, 21–24. Brown, R. E. & MacDonald, D. W. 1985. Social Odours in Mammals. Oxford: Clarendon Press. Buchanan, K. L., Evans, M. R., Goldsmith, A. R., Bryant, D. M. & Rowe, L. V. 2001. Testosterone influences basal metabolic rate in male house sparrows: a new cost of dominance signalling? Proceedings of the Royal Society of London, Series B, 268, 1337–1344. Callard, G. V., Drygas, M. & Gelinas, D. 1993. Molecular and cellular physiology of aromatase in the brain and retina. Journal of Steroid Biochemistry and Molecular Biology, 44, 541–547. Cardwell, J. R., Stacey N. E., Tan, E. S. P., McAdam, D. S. O. & Lang, S. L. C. 1995. Androgen increases olfactory receptor response to a vertebrate sex pheromone. Journal of Comparative Physiology A, 176, 55–61. Carlson, B. A., Hopkins, C. D. & Thomas, P. 2000. Androgen correlates of socially induced changes in the electric organ discharge waveform of a mormyrid fish. Hormones and Behavior, 38, 177–186. Cedrini, L. & Fasolo, A. 1970. Olfactory attractants in sex recognition of the crested newt: an electrophysiological research. Monitore Zoologico Italiano, 5, 223–229. Chapman, F. M. 1935. The courtship of Gould’s manakin (Manacus vitellinus vitellinus) on Barrado Colorado Island, Canal Zone. American Museum of Natural History Bulletin, 68, 472–521. Chase, I. D., Bartolomeo, C. & Dugatkin, L. A. 1994. Aggressive interactions and inter-contest interval: how long do winners keep winning? Animal Behaviour, 48, 393–400. Clotfelter, E. & Paolino, A. 2003. Winner effect is induced in bystanders to aggressive contests between male fighting fish. Animal Behaviour, 66, 343–347. Cronley-Dillon, J. & Sharma, S. C. 1968. Effect of season and sex on the photopic spectral sensitivity of the three-spined stickleback. Journal of Experimental Biology, 49, 679–687. Cynx, J. & Nottebohm, F. 1992. Testosterone facilitates some conspecific song discriminations in castrated zebra finches Taeniopygia guttata. Proceedings of the National Academy of Sciences, USA, 89, 1376–1378. De Vries, G. J. 1995. Studying neurotransmitter systems to understand the development and function of sex differences in the brain: the case of vasopressin. In: Neurobiological Effects of Sex Steroid Hormones, ed. P. E. Micevych & R. P. Hamer Jr. Cambridge, UK: Cambridge University Press, pp. 254–278. Deviche, P. 1992. Testosterone and opioids interact to regulate feeding in a male migratory songbird. Hormones and Behavior, 26, 394–405. Dohanich, G. 2002. Gonadal steroids, learning and memory. In: Hormones, Brain and Behavior, Vol. 2, ed. D. W. Pfaff, A. P. Arnold, A. M. Etgen, S. E. Fahrbach & R. T. Rubin. New York: Academic Press, pp. 265–327.

509

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

510

0 521 582361 7

April 11, 2005

19:12

R. F. Oliveira Dorl¨ ochter, M., Astrow, S. H. & Herrera, A. A. 1994. Effects of testosterone on a sexually dimorphic frog muscle: repeated in vivo observations and androgen receptor distribution. Journal of Neurobiology, 25, 897–916. Doutrelant, C., McGregor, P. K. & Oliveira, R. F. 2001. The effect of an audience on intra-sexual communication in male Siamese fighting fish, Betta splendens. Behavioral Ecology, 12, 283–286. Drummond, H. & Os´ orio, J. L. 1992. Training siblings to be submissive losers: dominance between booby nestlings. Animal Behaviour, 44, 881–893. Dugatkin, L. A. 2001. Bystander effects and the structure of dominance hierarchies. Behavioral Ecology, 12, 348–352. Dunlap, K. D. & Zakon, H. H. 1998. Behavioral actions of androgens and androgen receptor expression in the electrocommunication system of an electric fish, Eigenmannia virescens. Hormones and Behavior, 34, 30–38. Earley, R. L. & Dugatkin, L. A. 2002. Eavesdropping on visual cues in green swordtail (Xiphophorus helleri) fights: a case for networking. Proceedings of the Royal Society of London, Series B, 269, 943–952. Edwards, D. H. & Kravitz, E. A. 1997. Serotonin, social status and aggression. Current Opinion in Neurobiology, 7, 812–819. Eichenbaum, H., Otto, T. & Cohen, N. J. 1992. The hippocampus: what does it do? Behavioral and Neural Biology, 57, 2–36. Evans, C. S. & Marler, P. 1994. Food calling and audience effects in male chickens, Gallus gallus: their relationships to food availability, courtship and social facilitation. Animal Behaviour, 47, 1159–1170. Ferkin, M. H. & Johnston, R. E. 1993. Roles of gonadal hormones in control of five sexually attractive odors in meadow voles (Microtus pennsylvanicus). Hormones and Behavior, 27, 523–538. Fernald, R. D. 1976. The effect of testosterone on the behavior and coloration of adult male cichlid fish, Haplochromis burtoni. Hormone Research, 7, 172–178. Fine, M. L. & Pennymaker, K. R. 1986. Hormonal basis for sexual dimorphism of the sound-producing apparatus of the oyster toadfish. Experimental Neurology, 92, 289–298. Forlano, P. M., Sisneros, J. A., Knapp, R. & Bass, A. H. 2003. Reproductive periodicity of gonadal steroid hormone levels in the vocal plainfin midshipman, Porichthys notatus. Hormones and Behavior, 44: 49. Francis, R. C. 1983. Experiential effects on agonistic behavior in the paradise fish, Macropodus opercularis. Behaviour, 85, 292–313. 1987. The interaction of genotype and experience in the dominance success of paradise fish (Macropodus opercularis). Biology of Behaviour, 12, 1–11. Franck, D. & Ribowski, A. 1987. Influences of prior agonistic experiences on aggression measures in the male swordtail (Xiphophorus helleri). Behaviour, 103, 217–240. Freedman, E. G., Olyarchuk, J., Marchaterre, M. A. & Bass, A. H. 1989. A temporal analysis of testosterone-induced changes in electric organs and electric organ discharges of mormyrid fishes. Journal of Neurobiology, 20, 619–634.

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:12

Hormones, social context and animal communication Frey, D. F. & Miller, R. J. 1972. The establishment of dominance relationships in the blue gourami Trichogaster trichopterus (Pallas). Behaviour, 42, 8–62. Gahr, M., Fl¨ ugge, G. & G¨ uttinger, H. R. 1987. Immunocytochemical localization of estrogen-binding neurons in the songbird brain. Brain Research, 402, 173–177. Gahr, M., G¨ uttinger, H. R. & Kroodsma, D. A. 1993 Estrogen receptors in the avian brain: survey reveals general distribution and forebrain areas unique to songbirds. Journal of Comparative Neurology, 327, 112–122. Gawienowsky, A. M., De Nicola, D. B. & Stacewicz-Sapuntzakis, M. 1976. Androgen dependence of a marking pheromone in rat urine. Hormones and Behavior, 7, 401–405. Gelinas, D. & Callard, G. V. 1997 Immunolocalization of aromatase and androgen receptor-positive neurons in the goldfish brain. General and Comparative Endocrinology, 106, 155–168. Ginsberg, B. & Allee, W. C. 1942. Some effects of conditioning of social dominance and subordination in inbred strains of mice. Physiological Zoology, 15, 485–506. Gonz´ alez-Mariscal, G., Melo, A. E., Zavala, A. & Beyer, C. 1993. Sex steroid regulation of chin-marking behavior in male New Zealand rabbits. Physiology and Behavior, 54, 1035–1040. Gould, S. J. & Vrba, E. S. 1982. Exaptation: a missing term in the science of form. Paleobiology, 8, 4–15. Grober, M. S. 1998. Socially controlled sex-change: integrating ultimate and proximate levels of analysis. Acta Ethologica, 1, 3–17. Gyger, M., Karakashian, S. J. & Marler, P. 1986. Avian alarm calling: is there an audience effect? Animal Behaviour, 34, 1570–1572. Hagedorn, M. & Carr, C. 1985. Single electrocytes produce a sexually dimorphic signal in South American electric fish Hypopomus occidentalis (Gymnotiformes, Hypopomidae). Journal of Comparative Physiology A, 156, 511–523. Hagedorn, M. & Heiligenberg, W. 1985. Court and spark: electric signals in the courtship and mating of gymnotid fish. Animal Behaviour, 32, 254–265. Hannes, R. P. 1984. Androgen and corticoid levels in blood and body extracts of high and low ranking swordtail males (Xiphophorus helleri) before and after social isolation. Zeitschrift f¨ ur Tierpsychologie, 66, 70–76. 1986. Blood and whole-body androgen levels of male swordtails correlated with aggression measures in a standard-opponent test. Aggressive Behavior, 12, 249–254. Hasegawa, Y., Okuno, A. & Nagasawa, H. 2002. Immunohistochemical study of androgenic gland hormone: localization in the male reproductive system and species specificity in the terrestrial isopods. General and Comparative Endocrinology, 125, 218–225. Hector, A. C. K., Seyfarth, R. M. & Raleigh, M. 1989. Male parental care, female choice and the effect of an audience in vervet monkeys. Animal Behaviour, 38, 262–271. Hegner, R. E. & Wingfield, J. C. 1987a. Social status and circulating levels of hormones in flocks of house sparrows, Passer domesticus. Ethology, 76, 1–14. 1987b. Effects of experimental manipulations of testosterone levels on parental investment and breeding success in male house sparrows. The Auk, 104, 462–469.

511

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

512

0 521 582361 7

April 11, 2005

19:12

R. F. Oliveira Heid, P., G¨ uttinger, H. R. & Prove, E. 1985. The influence of castration and testosterone replacement on the song architecture of canaries (Serinus canarius). Zeitschrift f¨ ur Tierpsychologie, 69, 224–236. Herbert, J. 1993. Peptides in the limbic system: neurochemical codes for coordinated adaptive responses to behavioural and physiological demand. Neurobiology, 41, 723–791. Herfeld, S. & Moller, P. 1998. Effects of 17-alfa-methyltestosterone on sexually dimorphic characters in the weakly electric fish Brienomyrus niger (Gunther, 1866) (Mormyridae): electric organ discharge, ventral body wall indentation, and anal-fin ray bone expansion. Hormones and Behavior, 34, 303–319. Hogan, J. A. 1967. Fighting and reinforcement in Siamese fighting fish Betta splendens. Journal of Comparative and Physiological Psychology, 64, 356–359. Hogan, J. A. & Bols, R. J. 1980. Priming of aggressive motivation in Betta splendens. Animal Behaviour, 28, 135–142. H¨ oglund, E., Kolm, N. & Winberg, S. 2001. Stress-induced changes in brain serotonergic activity, plasma cortisol and aggressive behavior in Arctic charr (Salvelinus alpinus) is counteracted by L-DOPA. Physiology and Behavior, 74, 381–389. Hogue, M.-E., Beaugrand, J. P. & Lagu¨e, P. C. 1996. Coherent use of information by hens observing their former dominant defeating or being defeated by a stranger. Behavioural Processes, 38, 241–252. Hollis, K. L., Dumas, M. J., Singh, P. & Fackelman, P. 1995. Pavlovian conditioning of aggressive behavior in blue gourami fish (Trichogaster trichopterus): winners become winners and losers stay losers. Journal of Comparative Psychology, 109, 123–133. Holst, D. V. & Eichmann, F. 1998. Sex-specific regulation of marking behavior by sex hormones and conspecifics scent in tree shrews (Tupaia belangeri). Physiology and Behavior, 63, 157–164. Hopkins, C. D. 1972. Sex differences in signalling in an electric fish. Science, 176, 1035–1037. Hsu, Y. & Wolf, L. L. 1999. The winner and loser effect: integrating multiple experiences. Animal Behaviour, 57, 903–910. 2001. The winner and loser effect: what fighting behaviours are influenced? Animal Behaviour, 61, 777–786. Hurst, J. L. 1990. Urine marking in populations of wild mice Mus domesticus Rutty. III. Communication between the sexes. Animal Behaviour, 40, 233–243. Ibara, R. M., Penny, L. T., Ebeling, A. W., van Dykhuizen, G. & Cailliet, G. 1983. The mating call of the plainfin midshipman fish, Porichthys notatus. In: Predators and Prey in Fishes, ed. D. G. L. Noakes, D. G. Lindquist, G. S. Helfman & J. A. Ward. The Hague: Junk Press, pp. 205–212. Ikeda, K. 1933. Effect of castration on the sexual characters of anadromous three-spined stickleback, Gasterosteus aculeatus aculeatus (L.). Japanese Journal of Zoology, 5, 135–157. Jackson, W. 1991. Why do winners keep winning? Behavioral Ecology and Sociobiology, 28, 271–276. Johnson, R. P. 1973. Scent marking in mammals. Animal Behaviour, 21, 521–535.

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:12

Hormones, social context and animal communication Johnsson, J. I. & Åkerman, A. 1998. Watch and learn: preview of the fighting ability of opponents alters contest behaviour in rainbow trout. Animal Behaviour, 56, 771–776. Johnston, R. E. 1975. Sexual excitation function of hamster vaginal secretions Animal Learning and Behavior, 3, 161–166. Kavaliers, M., Choleris, E. & Colwell, D. D. 2001. Brief exposure to female odors ‘emboldens’ male mice by reducing predator-induced behavioural and hormonal responses. Hormones and Behavior, 40, 497–509. Keller, C. H., Zakon, H. H. & Sanchez, D. Y. 1986. Evidence for a direct effect of androgens on electroreceptor tuning. Journal of Comparative Physiology A, 158, 301–310. Kelley, D. C. 2002. Hormonal regulation of motor output in amphibians: Xenopus laevis vocalizations as a model system. In: Hormones, Brain and Behavior, Vol. 2, ed. D. W. Pfaff, A. P. Arnold, A. M. Etgen, S. E. Fahrbach & R. T. Rubin. New York: Academic Press, pp. 445–468. Kerr, J. E., Allore, R. J., Beck, S. G. & Handa, R. J. 1995. Distribution and hormonal regulation of androgen receptors (AR) and AR messenger ribonucleic acid in the rat hippocampus. Endocrinology, 136, 3213–3221. Ketterson, E. D., Nolan, V., Jr, Wolf, L. & Ziegenfus, C. 1992. Testosterone and avian life histories: effects of experimentally elevated testosterone on behavior and correlates of fitness in the dark-eyed junco, Junco hyemalis. American Naturalist, 140, 980–999. Kikuyama, S., Toyoda, F., Ohmiya, Y. et al. 1995. Sodefrin: a female-attracting pheromone in newt cloacal glands. Science, 267, 1643–1645. Kikuyama, S., Toyoda, F., Yamamoto, K., Tanaka, S. & Hayashi, H. 1997. Femaleattracting pheromone in newt cloacal glands. Brain Research Bulletin, 44, 415–422. Kimura, D. 1996. Sex, sexual orientation and sex hormones influence human cognitive function. Current Opinion in Neurobiology, 6, 259–263. Landsman, R. E. & Moller, P. 1988. Testosterone changes the electric organ discharge and external morphology of the mormyrid fish Gnathonemus petersii (Mormyriformes). Experientia, 44, 900–903. Landsman, R. E., Harding, C. F., Moller, P. & Thomas, P. 1990. The effects of androgens and estrogen on the external morphology and electric organ discharge waveform of Gnathonemus petersii (Mormyriformes, Teleostei). Hormones and Behavior, 24, 532–553. Leiser, J. K. & Itzkowitz, M. 1999. The benefits of dear enemy recognition in three-contender convict cichlid (Cichlasoma nigrofasciatum) contests. Behaviour, 136, 983–1003. Lepri, J. & Randall, J. A. 1983. Hormonal regulation of sandbathing in male kangaroo rats (Dipodomys merriami). Behavioral and Neural Biology, 37, 256–264. Levy, M. & Aronson, L. R. 1955. Morphological effects of castration and hormone administration in the male cichlid fish Tilapia macrocephala. Anatomical Records, 122, 450–451. Lisk, R. D. & Bezier, J. L. 1980. Intrahypothalamic hormone implantation and activation of sexual behavior in the male hamster. Neuroendocrinology, 30, 220–222.

513

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

514

0 521 582361 7

April 11, 2005

19:12

R. F. Oliveira Lowe, P. 1942. The anatomy of Gould’s manakin (Manacus vitellinus) in relation to its display. Ibis, 6, 50–83. Luine, V., Nottebohm, F., Harding, C. & McEwen, B. S. 1980. Androgen affects cholinergic enzymes in syringeal motor neurons and muscle. Brain Research, 192, 89–107. Malacarne, G., Bottoni, L., Massa, R. & Vellano, C. 1984. The abdominal gland of the crested newt: a possible source of courtship pheromones. Monitore Zoologico Italiano, 18, 33–39. Manzo, J., Garcia, L. I., Hernandez, M. E., Carrillo, P. & Pacheco, P. 2002. Neuroendocrine control of urine-marking behavior in male rats. Physiology and Behavior, 75, 25–32. Marler, C. A., Walsberg, G., White, M. I. & Moore, M. 1995. Increased energy expenditure due to increased territorial defense in male lizards after phenotypic manipulation. Behavioral Ecology and Sociobiology, 37, 225–231. 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? Animal Behaviour, 34, 194–198. Matos, R. J. 2002. Social influences on signalling: audience effects and communication networks. Ph.D. Thesis, University of Copenhagen. Matos, R. J. & McGregor, P. K. 2002. The effect of the sex of an audience on male–male displays of Siamese fighting fish (Betta splendens). Behaviour, 139, 1211–1221. Mazur, A. & Booth, A. 1998. Testosterone and dominance in men. Behavioral and Brain Sciences, 21, 353–397. McDonald, N. W., Heimstra, A. L. & Damkot, D. K. 1968. Social modification of agonistic behaviour in fish. Animal Behaviour, 16, 437–441. McGregor, P. K. 1993. Signalling in territorial systems: a context for individual identification, ranging and eavesdropping. Philosophical Transactions of the Royal Society of London, Series B, 340, 237–244. McGregor, P. K. & Peake, T. 2000. Communication networks: social environments for receiving and signalling behaviour. Acta Ethologica, 2, 71–81. McGregor, P. K., Dabelsteen, T. & Holland, J. 1997. Eavesdropping in a territorial songbird communication network: preliminary results. Bioacoustics, 8, 253–254. McKibben, J. R. & Bass, A. H. 1998. Behavioral assessment of acoustic parameters relevant to signal recognition and preference in a vocal fish. Journal of the Acoustic Society of America, 104, 3520–3533. Meyer, J. H. 1983. Steroid influences upon the discharge frequency of a weakly electric fish. Journal of Comparative Physiology A, 153, 29–38. Meyer, J. H. & Zakon, H. H. 1982. Androgens alter the tuning of electroreceptors. Science, 217, 635–637. Meyer, J. H., Leong, M. & Keller, C. H. 1987. Hormone-induced and ontogenetic changes in electric organ discharge and electroreceptor tuning in the weakly electric fish Apteronotus. Journal of Comparative Physiology A, 160, 385–394. Millinski, M. & Bakker, T. C. M. 1990. Female sticklebacks use male coloration in mate choice and hence avoid parasitised males. Nature, 344, 330–333.

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:12

Hormones, social context and animal communication Mills, A. C. & Zakon, H. H. 1987. Coordination of EOD frequency and pulse duration in a weakly electric wave fish: the influence of androgens. Journal of Comparative Physiology A, 161, 417–430. 1991. Chronic androgen treatment increases action potential duration in the electric organ of Sternopygus. Journal of Neuroscience, 11, 2349–2361. Moofat, S. D. & Hampson, E. 1996. A curvilinear relationship between testosterone and spatial cognition in humans: possible influence of hand preferences. Psychoneuroendocrinology, 21, 323–337. Moore, A. & Scott, A. P. 1991. Testosterone is a potent odorant in precocious male Atlantic salmon (Salmo salar L.) parr. Philosophical Transactions of the Royal Society of London, Series B, 332, 241–244. Moore, F. L., Wood, R. E. & Boyd, S. K. 1992. Sex steroids and vasotocin interact in a female amphibian (Taricha granulosa) to elicit female-like egg-laying behavior or male-like courtship. Hormones and Behavior, 26, 156–166. Nagaya, N. & Herrera, A. A. 1995. Effects of testosterone on synaptic efficacy at neuromuscular junctions in a sexually dimorphic muscle of male frogs. Journal of Physiology, 483, 141–153. Naguib, M., Fichtel, C. & Todt, D. 1999. Nightingales respond more strongly to vocal leaders of simulated dyadic interactions. Proceedings of the Royal Society of London, Series B, 266, 537–542. Nelson, J. S. 1994. Fishes of the World, 3rd edn. New York: John Wiley. Nelson, R. 2001. An Introduction to Behavioural Endocrinology, 2nd edn. Sunderland, MA: Sinauer. Northcutt, R. G. & Davis, R. E. 1983. Telencephalic organization in ray-finned fishes. In: Fish Neurobiology, Vol. 2: Higher Brain Areas and Functions, ed. R. E. Davis & R. G. Northcutt. Ann Arbor, MI: University of Michigan Press, pp. 203–236. Oliveira, R. F. & Almada, V. C. 1998. Androgenization of dominant males in a cichlid fish: androgens mediate the social modulation of sexually dimorphic traits. Ethology, 104, 841–858. Oliveira, R. F., Almada, V. C. & Can´ ario, A. V. M. 1996. Social modulation of sex steroid concentrations in the urine of male cichlid fish Oreochromis mossambicus. Hormones and Behavior, 30, 2–12. Oliveira, R. F., McGregor, P. K. & Latruffe, C. 1998. Know thine enemy: fighting fish gather information from observing conspecifics interactions. Proceedings of the Royal Society of London, Series B, 265, 1045–1049. Oliveira, R. F., Carneiro, L. A., Gonc¸alves, D. M., Can´ ario, A. V. M. & Grober, M. S. 2001a. Effects of 11-ketotestosterone in sneaker males of the peacock blenny, a species with male alternative mating tactics. Brain, Behavior and Evolution, 58, 28–37. Oliveira, R. F., Lopes, M., Carneiro, L. A. & Can´ ario, A. V. M. 2001b. Watching fights raises fish hormone levels. Nature, 409, 475. Oliveira, R. F., Hirschenhauser, K., Carneiro, L. & Can´ ario, A. V. M. 2002. Social modulation of androgen levels in male teleost fish. Comparative Biochemistry and Physiology B, 132, 203–215.

515

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

516

0 521 582361 7

April 11, 2005

19:12

R. F. Oliveira Otronen, M. 1990. The effect of prior experience on the outcome of fights in the burying beetle, Nicrophorus humatio. Animal Behaviour, 40, 980–1004. Owen, P. C. & Perrill, S. A. 1998. Habituation in the green frog, Rana clamitans. Behavioral Ecology and Sociobiology, 44, 209–213. Pankhurst, N. W. & Barnett, C. W. 1993. Relationship of population density, territorial interaction, and plasma levels of gonadal steroids in spawning male demoiselles Chromis dispilus (Pisces, Pomacentridae). General and Comparative Endocrinology, 90, 168–176. Parker, G. A. 1974. The assessment strategy and the evolution of fighting behaviour. Journal of Theoretical Biology, 47, 223–243. Peake, T. M. & McGregor, P. K. 2004. Information and aggression in fishes. Learning and Behavior 32, 114–121. Peake, T. M., Terry, A. M. R., McGregor, P. K. & Dabelsteen, T. 2001. Male great tits eavesdrop on simulated male-to-male vocal interactions. Proceedings of the Royal Society of London, Series B, 268, 1183–1187. Pomerantz, S. M. & Sholl, S. A. 1987. Analysis of sex and regional differences in androgen receptors in fetal rhesus monkey brain. Brain Research, 433, 151–154. Randall, J. A. 1986. Lack of gonadal control of the dorsal gland and sandbathing in male and female Bannertail kangaroo rats (Dipodomys spectabilis). Hormones and Behavior, 20, 95–105. Reinboth, R. & Rixner, W. 1972. Verhalten des Kleinen Maulbr¨ uters Hemihaplochromis multicolor nach Kastration und Behandlung mit Testosteron. [Film C 1019/, 1970]. G¨ ottingen: Institut f¨ ur den Wissenschaftlichen Film. Resnick, S. M., Berenbaum, S. A., Gottesman, I. J. & Bouchard, T. J. 1986. Early hormonal influences on cognitive functioning in congenital adrenal hyperplasia. Developmental Psychology, 22, 191–198. Roof, R. L. 1993. Neonatal exogenous testosterone modifies sex difference in radial arm and Morris water maze performance in prepubescent and adult rats. Behavioral and Brain Research, 53, 1–10. Roof, R. L. & Havens, M. D. 1992. Testosterone improves maze performance and induces development of a male hippocampus in females. Brain Research, 572, 310–313. Ros, A. F. H., Becker, K., Canario, A. V. M. & Oliveira, R. F. 2004. Androgen levels and energy metabolism in Oreochromis mossambicus. Journal of Fish Biology, 65, 895–905. Rose, R. M., Bernstein, I. S. & Holaday, J. W. 1971. Plasma testosterone, dominance rank and aggressive behavior in a group of male rhesus monkeys. Nature, 231, 366. Rose, R. M., Berstein, I. S. & Gordon, T. 1975. Consequences of social conflict on plasma testosterone levels in rhesus monkeys. Psychosomatic Medicine, 37, 50–61. Rost, R. 1990. Hormones and behavior: a joint examination of studies on seasonal variation in song production and plasma levels of testosterone in the great tit Parus major. Journal of Ornithology, 131, 403–411. 1992. Hormones and behavior: a comparison of studies on seasonal changes in song production and testosterone plasma levels in the Willow tit Parus montanus. Ornis Fennica, 69, 1–6.

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:12

Hormones, social context and animal communication Rouse, E. F., Coppinger, C. J. & Barnes, P. R. 1977. The effect of an androgen inhibitor on behavior and testicular morphology in the stickleback Gasterosteus aculeatus. Hormones and Behavior, 9, 8–18. Rowland, W. J. 1999. Studying visual cues in fish behavior: a review of ethological techniques. Environmental Biology of Fishes, 56, 285–305. Ruzzante, D. E. 1992. Mirror image stimulation, social hierarchies, and population differences in agonistic behavior: a reappraisal. Canadian Journal of Fisheries and Aquatic Sciences, 49, 1966–1968. Saldanha, C. J., Clayton, N. S. & Schlinger, B. A. 1999. Androgen metabolism in the juvenile oscine forebrain: a cross-species analysis at neural sites implicated in memory function. Journal of Neurobiology, 40, 397–406. Sapolsky, R. M. 1992. Neuroendocrinology of the stress-response. In: Behavioral Endocrinology, ed. J. B. Becker, S. M. Breedlove & D. Crews. Cambridge, MA: MIT Press, pp. 287–324. Sawyer, T. F., Hengehold, A. K. & Perez, W. A. 1984. Chemosensory and hormonal mediation of social memory in male rats. Behavioral Neuroscience, 98, 908–913. Scalia, F. & Winans, S. S. 1975. The differential projections of the olfactory bulb and accessory olfactory bulb in mammals. Journal of Comparative Neurology, 161, 31–56. Schlinger, B. A., Schultz, J. D. & Hertel, F. 2001. Neuromuscular and endocrine control of an avian courtship behavior. Hormones and Behavior, 40, 276–280. Schuett, G. W. 1997. Body size and agonistic experience affect dominance and mating success in male copperheads. Animal Behaviour, 54, 213–224. Schulkin, J. 2002. Hormonal modulation of central motivational states. In: Hormones, Brain and Behavior, Vol. 1, ed. D. W. Pfaff, A. P. Arnold, A. M. Etgen, S. E. Fahrbach & R. T. Rubin. New York: Academic Press, pp. 663–657. Schultz, J. D. & Schlinger, B. A. 1999. Widespread accumulation of [3 H]testosterone in the spinal cord of a wild bird with an elaborate courtship display. Proceedings of the National Academy of Sciences, USA, 96, 10428–10432. Schultz, J. D., Hertel, F., Bauch, M. & Schlinger, B. A. 2001. Adaptations for rapid and forceful contraction in wing muscles of the male golden-collared manakin: sex and species comparisons. Journal of Comparative Physiology A, 187, 677–684. Searcy, W. A., Eriksson, D. & Lundberg, A. 1991. Deceptive behavior in pied flycatchers. Behavioral Ecology and Sociobiology, 29, 167–175. Silverin, B. 1980. Effects of long-acting testosterone treatment on free-living pied flycatchers, Ficedula hypoleuca, during the breeding period. Animal Behaviour, 28, 906–912. Sisneros, J. A. & Bass, A. H. 2003. Seasonal plasticity of peripheral auditory frequency sensitivity. Journal of Neuroscience, 23, 1049–1058. Sisneros, J. A. & Tricas, T. C. 2000. Androgen-induced changes in the response dynamics of ampullary electrosensory primary afferent neurons. Journal of Neuroscience, 20, 8586–8595. Simon, N. G. 2002. Hormonal processes in the development and expression of aggressive behavior. In: Hormones, Brain and Behavior, Vol. 1, ed. D. W. Pfaff, A. P. Arnold, A. M. Etgen, S. E. Farbach & R. T. Rubin. New York: Academic Press, pp. 339–392.

517

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

518

0 521 582361 7

April 11, 2005

19:12

R. F. Oliveira Smith, G. T., Brenowitz, E. A., Beecher, M. D. & Wingfield, J. C. 1997. Seasonal changes in testosterone, neural attributes of song control nuclei, and song structure in wild songbirds. Journal of Neuroscience, 17, 6001–6010. Smith, R. J. F. 1974. Effects of 17-methyltestosterone on the dorsal pad and tubercles of fathead minnows (Pimephales promelas). Canadian Journal of Zoology, 52, 1031–1038. Squire, L. R. 1992. Memory and the hippocampus: a synthesis from findings with rats, monkeys, and humans. Psychological Reviews, 9, 195–231. Stacey, N. & Sorensen, P. 2002. Hormonal pheromones in fish. In: Hormones, Brain and Behavior, Vol. 2, ed. D. W. Pfaff, A. P. Arnold, A. M. Etgen, S. E. Fahrbach & R. T. Rubin. New York: Academic Press, pp. 375–434. Suthers, R. A., Goller, F. & Pytte, C. 1999. The neuromuscular control of bird song. Philosophical Transactions of the Royal Society of London, Series B, 354, 927–939. Taborsky, M. 1994. Sneakers, satellites, and helpers: parasitic and cooperative behaviour in fish reproduction. Advances in the Study of Behaviour, 23, 1–100. Taylor, G. T., Bartko, G. & Farr, S. 1987. Gonadal hormones and conspecific marking in male rats. Hormones and Behavior, 21, 234–244. Tellegen, A., Horn, J. M. & Legrand, R. G. 1969. Opportunity for aggression as a reinforcer for mice. Psychonomic Science, 14, 104–105. Temeles, E. J. 1994. The role of neighbours in territorial systems: when are they ‘dear enemies’? Animal Behaviour, 47, 339–350. Tinbergen, N. 1951. The Study of Instinct. Oxford: Clarendon Press. Tobias, M. L. & Kelley, D. B. 1995. Sexual differentiation and hormonal regulation of the laryngeal synapse in Xenopus laevis. Journal of Neurobiology, 28, 515–526. Tobias, M. L., Kelley, D. B. & Ellisman, M. 1995. A sex-difference in synaptic efficacy at the laryngeal neuromuscular junction of Xenopus laevis. Journal of Neuroscience, 15, 1660–1668. Tohgi, H., Utsugisawa, K., Yamagata, M. & Yoshimura, M. 1995. Effects of age on mRNA expression of glucocorticoid, thyroid hormone, androgen, and estrogen receptors in postmortem human hippocampus. Brain Research, 700, 245–253. Vezina, M. & Thomas, D. W. 2000. Social status does not affect resting metabolic rate in wintering dark-eyed junco (Junco hyemalis). Physiological and Biochemical Zoology, 73, 231–236. Wapler-Leong, D. C. Y. & Reinboth, R. 1974. The influence of androgenic hormone on the behaviour of Haplochromis burtoni (Cichlidae). Fortschritte der Zoologie, 22, 334–339. Wetzel, D. M. & Kelley, D. B. 1983. Androgen and gonadotrophin effects on male mate calls in South African clawed frogs, Xenopus laevis. Hormones and Behaviour, 17, 388–404. Whitehouse, M. A. 1997. Experience influences male–male contests in the spider Argyrodes antipodiana (Theridiidae: Aranea). Animal Behaviour, 53, 913–923. Whiting, M. J. 1999. When to be neighbourly: differential agonistic responses in the lizard Platysaurus broadleyi. Behavioral Ecology and Sociobiology, 46, 210–214. Wikelsky, M., Lynn, S., Breumer, C., Wingfield, J. C. & Kenagy, G. J. 1999. Energy metabolism, testosterone and corticosterone in white-crowned sparrows. Journal of Comparative Physiology A, 185, 463–470.

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:12

Hormones, social context and animal communication Williams, C. L., Barnett, A. M. & Meck, W. H. 1990. Organizational effects of early gonadal secretions on sexual differentiation in spatial memory. Behavioral Neuroscience, 104, 84–97. Winberg, S. & Lepage, O. 1998. Elevation of brain 5-HT activity, POMC expression, and plasma cortisol in socially subordinate rainbow trout. American Journal of Physiology, 274, R645–R654. Winberg, S. & Nilsson, G. E. 1993a. Roles of brain monoamine neurotransmitters in agonistic behaviour and stress reactions, with particular reference to fish. Comparative Biochemistry and Physiology C, 10, 597–614. 1993b. Time course of changes in brains serotonergic activity and brain tryptophan levels in dominant and subordinate juvenile Arctic charr. Journal of Experimental Biology, 179, 181–195. Winberg, S., Winberg, Y. & Fernald, R. D. 1997. Effect of social rank on brain monoaminergic activity in a cichlid fish. Brain, Behavior and Evolution, 49, 230–236. Wingfield, J. C. 1984 Environmental and endocrine control of reproduction in the song sparrow, Melospiza melodia. II. Agonistic interactions as environmental information stimulating secretion of testosterone. General and Comparative Endocrinology, 56, 417–424. Wingfield, J. C., Ball, G. F., Dufty, A. M., Hegner, R. E. & Ramenofsky, M 1987. Testosterone and aggression in birds: tests of the challenge hypothesis. American Scientist, 75, 602–608. Wingfield, J. C., Hegner, R. E., Dufty, A. M. & Ball, G. F. 1990. The ‘challenge hypothesis’: theoretical implications for patterns of testosterone secretion, mating systems, and breeding strategies. American Naturalist, 136, 829–846. Wingfield, J. C., Jacobs, J. D., Soma, K. et al. 1999. Testosterone, aggression, and communication: ecological bases of endocrine phenomena. In: The Design of Animal Communication, ed. M. D. Hauser & M. Konishi. Cambridge, MA: MIT Press, pp. 257–283. Wingfield, J. C., Jacobs, J. D., Tramontin, A. D. et al. 2000. Toward an ecological basis of hormone–behavior interactions in reproduction of birds. In: Reproduction in Context, ed. K. Wallen & J. E. Schneider. Cambridge, MA: MIT Press, pp. 85–128. Wood, R., Brabec, R. K., Swann, J. M. & Newman, S. W. 1992. Androgen and estrogen concentrating neurons in chemosensory pathways in the male Syrian hamster brain. Brain Research, 596, 89–98. Yahr, P. & Stephens, D. R. 1987. Hormonal control of sexual and scent marking behaviors of male gerbils in relation to the sexually dimorphic area of the hypothalamus. Hormones and Behavior, 21, 331–346. Yamamoto, K., Toyoda, F., Tanaka, S., Hayashi, H. & Kikuyama, S. 1996. Radioimmunoassay of a newt sex pheromone, sodefrin, and the influence of hormones on its level in the abdominal gland. General and Comparative Endocrinology, 104, 356–363. Ydenberg, R. C., Giraldeau, L. A. & Falls, J. B. 1988. Neighbours, strangers and the asymmetric war of attrition. Animal Behaviour, 36, 343–347.

519

P1: GFZ-IYP/... P2: KOD 0521823617c21.xml CU1917B/McGregor

520

0 521 582361 7

April 11, 2005

19:12

R. F. Oliveira Zakon, H. H. & Smith, G. T. 2002. Weak electric fish: behavior, neurobiology and neuroendocrinology. In: Hormones, Brain and Behavior, Vol. 2, ed. D. W. Pfaff, A. P. Arnold, A. M. Etgen, S. E. Fahrbach & R. T. Rubin. New York: Academic Press, pp. 349–374. Zakon, H. H., Yan, H. Y. & Thomas, P. 1990. Human chorionic gonadotropin-induced shifts in the electrosensory system of the weakly electric fish, Sternopygus. Journal of Neurobiology, 21, 826–833. Zakon, H. H., Thomas, P. & Yan, H. Y. 1991. Electric organ discharge frequency and plasma sex steroid levels during gonadal recrudescence in a natural population of the weakly electric fish Sternopygus macrurus. Journal of Comparative Physiology A, 169, 493–499.

P1: JZZ/... P2: JZZ/... 0521823617c22.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:30

22

Cooperation in communication networks: indirect reciprocity in interactions between cleaner fish and client reef fish r e d o u a n b s h a r y 1 & a r u n d ’s o u z a 2 1 2

University of Cambridge, Cambridge, UK University of W¨ urzburg, W¨ urzburg, Germany

Introduction The aim of this chapter is twofold. First, to outline how recent developments in cooperation theory are so similar to the communication network concept (McGregor, 1993) that a unified terminology would be useful to facilitate exchange of ideas. Second, we argue that the communication network concept provides an evolutionary framework to predict the widespread occurrence of phenomena that until now have been discussed in the context of highly developed cognitive abilities. This creates a problem: as it stands, there appear to be no words in cooperation theory that were not developed in the human context and hence do not include a cognitive component. We have to use definitions that only constitute the functional aspects of phenomena (like tactical deception and indirect reciprocity) and ignore the mechanistic aspects (i.e. theory of mind, intentionality) that are commonly part of the definitions. We ask readers always to keep in mind that our definitions never imply any specific cognitive abilities. We illustrate our ideas with data on the mutualism between the cleaner wrasse Labroides dimidiatus and its ‘client’ reef fish, which visit cleaners to have ectoparasites removed from their surface, gills and mouth (reviewed by Losey et al., 1999; Cˆ ot´e, 2000). Cooperation provides a challenge to evolutionary theory because it often involves apparently altruistic behaviour. Hamilton (1964) provided a framework to understand why altruism between kin can be evolutionary stable; specifically, an altruist gains indirect fitness benefits from its action. However, there are plenty of Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

521

P1: JZZ/... P2: JZZ/... 0521823617c22.xml CU1917B/McGregor

522

0 521 582361 7

April 11, 2005

19:30

R. Bshary & A. D’Souza examples where recipients of altruistic acts are unrelated to the helping individual (Dugatkin, 1997). Trivers (1971) pointed out that such altruism might be evolutionarily stable if the recipient later reciprocates, an idea that was formalized by Axelrod & Hamilton (1981). They used an iterated version of the prisoner’s dilemma game, a two-player game in which opponents can either cooperate or defect. Defection yields a higher payoff than cooperation independently of the partner’s action, but if both players cooperate they receive a higher payoff than if both defect, hence the dilemma. In a computer tournament in which several strategies competed with each other, a simple strategy called ‘tit-for-tat’ emerged as a cooperative solution to the game. Tit-for-tat players start by being cooperative; then in the next round they play what their opponent did in the previous round. Thus, tit-for-tat players can reap the benefits of mutual cooperation while avoiding exploitation by an uncooperative opponent (except for the first occasion on which an opponent defected). Several new strategies that are similar but apparently superior to tit for tat have been tested since the first computer tournament (reviewed by Dugatkin, 1997). Despite the intuitive appeal of reciprocal altruism and behavioural strategies similar to tit for tat to ensure cooperative behaviour, few empirical examples have been reported and some that have are contentious (Dugatkin, 1997). In addition, Alexander (1987) pointed out that many examples of human altruistic behaviour do not fit an iterated prisoner’s dilemma game: humans often help individuals who are highly unlikely to ever reciprocate. He proposed that humans might help others simply to increase their own image within the society. Nowak & Sigmund (1998) explored this idea by developing a game theory model in which direct reciprocity on altruistic acts was excluded. Instead, an altruist gained an increase in his image score. If an individual’s image score was linked to the probability that others were willing to provide help when needed, cooperation readily emerged and was evolutionarily stable (Nowak & Sigmund, 1998). Individuals helped in order to be helped themselves during future interactions with current bystanders (for further theoretical developments see Lotem et al. (1999, 2003) and Leimar & Hammerstein (2001)). An experiment with first-year students confirmed a crucial prediction of the models: students that helped more than average received more help and, therefore, a final payoff that was above the average (Wedekind & Milinski, 2000). This new approach to the evolution of cooperation is a specific instance of a communication network: the interactions between individuals do not happen in a social vacuum but in the presence of other individuals who eavesdrop and thereby extract relevant information for own future interactions with the actors (McGregor, 1993). As a consequence of eavesdropping, it pays individuals to alter their behaviour, either as a general unconditional response (Johnstone, 2001) or specifically in situations where bystanders are present (audience effects; see Ch. 4).

P1: JZZ/... P2: JZZ/... 0521823617c22.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:30

Indirect reciprocity in interactions in fish Nowak & Sigmund (1998) proposed that altruism based on indirect reciprocity is a hallmark of human evolution. Although they do not specify why this should be so, their proposal implies that non-human animals either do not live in social environments that would favour the evolution of altruism through indirect reciprocity or lack some of the cognitive abilities required (but see Zahavi, 1995; Roberts, 1998). In contrast, we suggest that the communication network concept has the advantage of coming from a purely functional perspective rather than trying to explain apparently maladapted human behaviour. Early discussions of communication networks (McGregor, 1993) argued that behaviours such as eavesdropping and audience effects should evolve in the context of a network, without detailed consideration of the underlying cognitive mechanisms. By ignoring mechanisms, we use the communication network concept to predict eavesdropping and audience effects in potentially cooperative contexts without worrying about cognitive constraints. Experimental evidence for the existence of eavesdropping has been provided for a wide variety of taxa (Ch. 2). Whereas human subjects can be asked about their behaviour, eavesdropping in other animals has to be inferred from the eavesdroppers’ subsequent behaviour towards individuals observed interacting. Differences in individuals’ roles must elicit differences in subsequent eavesdropper behaviour towards them. It was thus an implicit assumption of communication-network studies that eavesdroppers attribute some sort of image score to observed individuals and that this score governs their own future behaviour towards those individuals. Scoring an individual’s tendency to help is just one type of image score. Fighting ability, aggressiveness and mating success with regard to female choice are the image scores typically studied in communication networks (e.g. Ch. 5). That eavesdroppers adjust their own behaviour to what they have witnessed has important implications for the behaviour of individuals that are observed. While a classical approach would suggest that individuals maximize payoffs in each single interaction (with the exception of reciprocal altruism and punishment, where benefits are delayed), any occurrence of eavesdropping implies that selection favours individuals that optimize current actions by integrating both immediate payoffs and future consequences of their behaviour. Within the framework of cooperation theory, it may pay individuals to be altruistic if this increases the probability of meeting more cooperative eavesdroppers in the future; in contest theory, it may pay to be more aggressive in the presence of potential challengers if winning a fight results in fewer attacks from eavesdroppers. Individuals can respond to eavesdropping in two ways. First, they can alter their behaviour in any interaction in relation to the average probability that eavesdroppers are present. In this case, all individuals behave in the same way (with respect to eavesdroppers) in all interactions. Second, individuals can pay attention to specific cues that eavesdroppers

523

P1: JZZ/... P2: JZZ/... 0521823617c22.xml CU1917B/McGregor

524

0 521 582361 7

April 11, 2005

19:30

R. Bshary & A. D’Souza are present for a particular interaction and alter their behaviour accordingly. In the latter, individuals show a flexible behavioural pattern. Communication network models have until now dealt with the first scenario (Nowak & Sigmund, 1998; Johnstone, 2001; Leimar & Hammerstein, 2001). There is, however, increasing evidence that animals adjust their behaviour in a particular interaction according to the presence or absence of eavesdroppers (Doutrelant et al., 2001; Bshary, 2002; Ch. 4). Altruism towards unrelated individuals has been linked to positive reciprocity, be it direct (Trivers, 1971) or indirect (Nowak & Sigmund, 1998): individuals help because they will receive help in return. However, this is not necessarily the case. An alternative is that an individual helps in order to raise its image and uses its image to exploit recipients or eavesdroppers, which will behave cooperatively because of this high image. If all individuals in a population exploited eavesdroppers, then image scoring would break down. However, as long as image scoring yields an overall benefit, either because most altruistic acts are honest in that an individual’s willingness to cooperate is revealed or because the benefits of cooperation exceed the costs of being exploited, altruism may be used both as an honest and as a deceptive signal. Therefore, altruism may sometimes serve as a signal out of context, causing other individuals to react in the signaller’s favour and to their own disadvantage. This is the functional definition of tactical deception (Hauser, 1998). In communication-network terms, it may pay individuals to be altruistic if this signal is misinterpreted by eavesdroppers in a way that allows future exploitation of them. Such a functional approach to tactical deception is in strong contrast to the traditional cognitive approach. Though such behaviour has been described, for example in birds (Munn, 1986), tactical deception is often seen as a hallmark of primate ‘Machiavellian intelligence’ (Byrne & Whiten, 1988): the notion that most primate species have been strongly selected for the cognitive abilities to cope with their social environment (see references in Byrne & Whiten (1988) and Ch. 25). The ability to use tactical deception has, therefore, been linked to the concept of theory of mind (Premack & Woodruff, 1978): the ability to speculate how another individual might perceive a certain situation. However, Heyes (1998) cautioned that any observations of tactical deception do not imply the existence of particularly high cognitive abilities. Instead, originally animals might have made an error (i.e. produced a signal out of context) but it may have had a favourable outcome for the signaller. As a result the signaller may associate this error with a reward and consequently would be more likely to produce the signal again in this context. The notion that simple associative learning might suffice to produce signals that fit the functional definition of tactical deception offers the possibility of using a much more functional, rather than mechanistic, approach to the topic

P1: JZZ/... P2: JZZ/... 0521823617c22.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:30

Indirect reciprocity in interactions in fish (see also Hauser, 1998) – and the appropriate framework for the study of tactical deception is communication networks.

Interactions between cleaner fish and clients In the remainder of this chapter, we present data on interactions between the cleaner wrasse and client reef fish to illustrate the arguments outlined above. Data were collected in the Red Sea, at Ras Mohammed National Park, Egypt. Methods of data collection are described in detail elsewhere (Bshary, 2001, 2002), so here we will keep this kind of information to a minimum. All data are field observations; therefore, experimental proof is still lacking. However, these data illustrate that it is worthwhile searching for potential examples of positive indirect reciprocity and tactical deception with a functional perspective rather than worrying about cognitive constraints. Clients regularly visit the cleaners at their small territories called ‘cleaning stations’ (cleaning mutualism reviewed by Losey et al., 1999; Cˆ ot´e, 2000). As individual cleaner wrasse may have more than 2000 interactions per day (Grutter, 1995), interactions often take place in the presence of other potential visitors. Such bystanders can eavesdrop and evaluate the cleaner’s service quality. While the cleaner fish eat parasites, in particular gnathiid isopods (Grutter, 1996), they also feed on client mucus and scales (Randall, 1958; Grutter, 1997). Feeding on healthy client tissue is correlated with the occurrence of client ‘jolts’, an observable short shake of the client’s body, in response to mouth contact by the cleaner fish (Bshary & Grutter, 2002a). The frequency of client jolts correlates negatively with parasite load; therefore, client jolts are not a byproduct of parasite removal. Rather, jolts are an easily observable correlate of cleaner fish cheating (Bshary, & Grutter, 2002a). Note that only non-predatory clients (i.e. species that could not eat cleaner fish) jolt on a regular basis, while jolts of predatory clients are infrequent (Bshary, 2001). Therefore, we will only present data on non-predatory clients. In response to a jolt, clients often dart off or chase the cleaner, depending on their strategic options. Client species with large home ranges that cover several cleaning stations (‘choosy clients’) usually make use of their choice options and swim off and visit a different cleaning station for their next inspection (Bshary & Sch¨ affer, 2002), as predicted by biological market theory (No¨e et al., 1991; reviewed by No¨e, 2001). In contrast, client species with small territories or home ranges, and hence with access to only one cleaning station (‘resident clients’),tend to punish cleaners by chasing them (Bshary & Grutter, 2002a). Both darting off and punishment could readily provide bystanders with the information that a cleaning service was bad. In contrast, if an observed interaction ends without apparent conflict, then the service had probably been good. Therefore, clients could easily attribute an image

525

P1: JZZ/... P2: JZZ/... 0521823617c22.xml CU1917B/McGregor

526

0 521 582361 7

April 11, 2005

19:30

R. Bshary & A. D’Souza score to a cleaner fish, and cleaners could adjust their behaviour to the presence of eavesdropping potential clients. Why should clients attribute an image score to cleaners?

Attributing an image score to an individual and basing behavioural decisions during interactions with that individual on that score only makes sense if the score has some predictive power about how that individual will behave. In potentially cooperative interactions, a positive image score should be attributed to an individual only if cooperation on one occasion is usually followed by cooperation on the next occasion. In the context of cleaning mutualism, this implies that there must be consistent variation in cheating rates either between individual cleaners or within individual cleaners. Indeed, Bshary (2002) found that a minority of cleaners cheated more frequently than the rest. These ‘biting cleaners’, compared with normal cleaners, specifically targeted larger non-predatory clients, both residents (median client jolt rate was 12/100 seconds in interactions with biting cleaners compared with 2/100 seconds in interactions with normal cleaners) and choosy clients (18/100 seconds compared with 3/100 seconds), while there was no evidence for increased cheating of predatory clients (0/100 seconds compared with 0/100 seconds) or small resident clients (6/100 seconds compared with 6/100 seconds) (Bshary, 2002). These data suggest that it would pay larger non-predatory clients to avoid interactions with such biting cleaners. One way they could do this is to extract information from ongoing interactions and attribute an image to a cleaner. There is another reason why constant image scoring of cleaner behaviour is advantageous for clients. Data from one biting cleaner fish revealed that cheating rates changed considerably over a period of six weeks. Some of the 99% confidence intervals around observed daily jolt rates of choosy clients did not overlap, suggesting that the variation is significant (Fig. 22.1). This individual was a female, as were all other biting cleaners that have been observed (n = 7). Cheating of nonpredatory choosy clients peaked at the two periods of full moon that occurred during the observation period, and full moon coincided with repeated spawning with her male partner. After the second spawning period, the male disappeared and cheating rates fell to very low values. The cleaner wrasse is a protogynous hermaphrodite; that is, individuals start their reproductive career as females and eventually switch sex to become males (Robertson, 1972). Males have a larger reproductive output because they often have a harem. Therefore, females face a trade-off between investing in current reproductive effort through the production of eggs and investing in growth to become a male. If the energy requirements for egg production are maximal close to spawning, females needed extra energy in order to avoid compromising growth too much. We suggest that the females’ switch to a

P1: JZZ/... P2: JZZ/... 0521823617c22.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:30

Choosy client jolts/100 s

Indirect reciprocity in interactions in fish

Fig. 22.1. Jolt rates (with 99% confidence intervals) of choosy clients (species with large home ranges that cover several cleaning stations) when interacting with one particular female cleaner on nine different days, based on one hour of observations on each day.

temporarily deceptive strategy yields short-term energetic advantages. In aquaria, clients jolt more frequently when interacting with hungry cleaners that when interacting with satiated cleaners (A. S. Grutter, unpublished data). The benefits of cheating, therefore, seem to be condition dependent, and the client control mechanisms like punishment (Bshary & Grutter, 2002a) and partner switching (Bshary & Sch¨ affer, 2002) only work most of the time. It even appears that the same individual can switch back and forth between a cooperative and a biting strategy within seconds. Another biting female, observed over a six-week period, cheated clients frequently during the spawning period but client jolt levels remained high after that. Her male partner tolerated her presence at his cleaning station only during spawning but not thereafter and chased her off repeatedly. The female spent about equal amounts of time at her own cleaning station on the other side of the reef patch and on excursions to the male’s cleaning station. When the female was at the male’s cleaning station, her resident and choosy clients jolted significantly more frequently than when interacting with the female at her own cleaning station (residents (n = 8): t = 1.5, p = 0.021; choosy clients (n = 7): t = 0, p = 0.016; Wilcoxon matched-pair signedranks tests; Fig. 22.2). The variation both within and between cleaners apparent from these examples means that client image scoring is a profitable strategy to avoid (temporarily) cheating cleaners. Do clients attribute image scores to cleaners?

As shown elsewhere (Bshary, 2002), clients use information about the outcome of ongoing interactions when visiting a cleaning station. To appreciate fully what is happening, it is important to note that clients usually do not ‘hang out’ at cleaning stations but visit them only when they seek an inspection by a cleaner.

527

P1: JZZ/... P2: JZZ/... 0521823617c22.xml CU1917B/McGregor

April 11, 2005

19:30

R. Bshary & A. D’Souza p

p

Jolts/100 s

528

0 521 582361 7

C

C

Fig. 22.2. Jolt rates (medians and interquartile ranges) of eight resident species (clients without choice; n = 8) and seven client species with choice (defined in text; n = 7) during interactions with one particular female cleaner depending on the location of the interaction (at her own station (•) or at the male’s station (
); see text for further details).

Therefore, clients can gather information on how cleaners treat other clients only when they visit the station themselves and only if a cleaner is busy inspecting another client while they approach. So visiting clients can base their decision to invite inspection on current information only if they can observe another client being inspected. The newly arrived individual can attribute a positive image score to the cleaner if the current interaction ends without apparent conflict and a negative image score if the current interaction ends with the client darting off or chasing the cleaner. If another client is not present when the client arrives, no current information is available and the image score might be neutral. This is what the data suggest. If an ongoing interaction ended without apparent conflict, clients that had arrived during the interaction invited inspection by the cleaner in almost 100% of observed interactions. In contrast, if the interaction ended with an apparent conflict, clients hardly ever invited inspection (Bshary, 2002). When no information about a cleaner’s previous interaction was available, clients invited inspection with intermediate probability and the actual outcome of the previous interaction (that was unobserved by the client) had no significant effect. When clients do not invite inspection, they often exhibit an ambiguous response; they let the cleaner approach and inspect but do not stop coordinated swimming movements before the interaction starts (they may stop afterwards). They may also flee from the approaching cleaner. Fleeing most often results in no inspection and frequently happens when clients are approached immediately after an interaction has ended with a conflict. In contrast, fleeing hardly ever occurs after a positive interaction had just ended or if the previous interaction had ended a while ago (Fig. 22.3). These observations are consistent with the statement above that clients only visit cleaning stations to seek an inspection. The decision to invite inspection is only altered if they observe a negative interaction.

P1: JZZ/... P2: JZZ/... 0521823617c22.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:30

Indirect reciprocity in interactions in fish P N

C

P

Fig. 22.3. The frequency of fleeing (accelerating away from approaching cleaners) by resident client species (•) and choosy client species (defined in text,
) arriving at a cleaning station in four different situations: the previous interaction ended either ≤ 5 s or ≤ 5 s ago and had positive (without conflict) or negative (with conflict) outcomes. Values are median and interquartile ranges. The letters b and b above the values for fleeing in the situation where the previous interaction had ended negatively ≤ 5 s ago indicate a significant difference to the other three situations, which are not statistically significant between each other, as indicated by using the same letters a and a .

Response of ‘normal’ cleaner fish to image-scoring clients

If clients attribute image scores to cleaners, one would expect that cleaners adjust their behaviour and cheat current clients less frequently if bystanders are present than when no bystanders are present. Such audience effects should be particularly common if bystanders have access to several cleaning stations, as these species (see above) might not only delay their interaction with the cleaner but also swim to another cleaning station. Resident bystanders can only delay their interaction or avoid interactions altogether and remain uncleaned. To look for such effects, we assumed that all individuals within 50 cm of cleaner–client interactions at the beginning and at the end of each interaction were able to collect information about the ongoing interaction and all individuals ≥ 10 cm total length were potential next clients. We quantified the number of all such individuals and their species identity for 12 cleaners. For each client species and cleaner station, we calculated correlations between the frequency of jolts and the number of bystanders. We analysed four (partly overlapping) classes of bystander; conspecific, heterospecific, resident species and choosy species. We only calculated correlations for observations where only one of these classes of bystander was present. For each client species and bystander category, we compared the number

529

P1: JZZ/... P2: JZZ/... 0521823617c22.xml CU1917B/McGregor

April 11, 2005

19:30

R. Bshary & A. D’Souza p

p

No. species

530

0 521 582361 7

C

Heterospecifics

R

C

Fig. 22.4. The influence of the presence (within 50 cm at the beginning and the end of an interaction) of four bystander classes on the jolt rate of clients during interactions with cleaners. The histograms show the number of client species for which the correlation between jolt rate and the number of bystanders of a category was either negative (black) or positive (white).

of positive and negative correlations and scored a plus if the majority was positive and a minus if the majority was negative. Thus, we had one data point for each client species and bystander category and evaluated any significant impacts of bystander categories on client jolt rates using sign tests. We did not find a significant effect of conspecific bystanders on client jolt rates (n = 13; x = 6; NS) while the presence of heterospecific bystanders had a significantly negative effect (n = 23; x = 6; p = 0.034; Fig. 22.4). The effect of heterospecific bystanders was mainly owing to choosy bystanders (n = 17; x = 1; p < 0.001) while resident bystanders did not have a significant effect on client jolt rates (n = 15; x = 6; NS; Fig. 22.4). While the data presented above are in line with the hypothesis that client image scoring influences cleaner fish behaviour, there is an alternative explanation. It could be that when more clients are present it is easier for cleaners to pick the few obvious parasites from each of them and the reduction in client jolt rate is a result of an optimal foraging decision of cleaners rather than caused by bystander image scoring. In favour of the optimal foraging interpretation, it is known that choosy clients are, on average, larger than resident clients (Bshary 2001), which could explain the stronger effect of their presence on the current clients’ jolt rates than the smaller residents. However, optimal foraging cannot explain our observation that choosy bystanders have different effects on cleaner fish behaviour, depending on whether they are the same species as the interacting client or whether they are a different species. We can explore this effect further by considering only interactions in which individuals of one choosy species, the sergeant major Abudefduf vaigiensis, were bystanders. We picked 13 cleaning stations for data collection on the basis that these clients were frequent visitors. Sergeant majors may visit as single individuals or as large shoals of 20–50 individuals. For 11 out of 12 client species, we found more negative than positive correlations between the number of sergeant majors present and client jolt rates (sign test: n = 12; x = 1; p < 0.01). This result is the opposite of the effects of sergeant major bystanders on the jolt

P1: JZZ/... P2: JZZ/... 0521823617c22.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:30

Indirect reciprocity in interactions in fish p

Client (jolts/100 s)

p

0

1–4 Bystanders

≥5

Fig. 22.5. The influence of the number of sergeant major Abudefduf vaigiensis bystanders present (within 50 cm at the beginning and the end of an interaction) on the jolt rate of the sergeant major client.

rates of sergeant majors being cleaned: in comparison to interactions in which no bystanders were present, small numbers of bystanders did not have any detectable effect and the presence of large numbers led to an increase in jolt rates (Friedman test: n = 12; χ 2 = 9.9; df = 2; p < 0.01; Fig. 22.5). Response of biting cleaners to image scoring clients

Biting cleaners have more interactions that end with a conflict and clients approaching their cleaning station more often avoid them than normal cleaners (Bshary, 2002). As explained above, it seems likely that the latter observation is the result of client image scoring rather than previous direct experience of clients. Do biting cleaners still have some means to improve their image? In this respect, it is important to note that biting cleaners behave very differently from normal cleaners, not only with respect to jolt rates of large clients but also with respect to their behaviour towards small resident clients. Biting cleaners often ride above the small residents’dorsal area and provide tactile stimulation with their pectoral and pelvic fins. While this behaviour is part of every cleaner’s repertoire, about 50% of the interactions between biting cleaners and small residents consisted of tactile stimulation only (Bshary, 2002). Providing tactile stimulation is incompatible with foraging; hence interactions that consist of tactile stimulation only are clearly costly to cleaners. Usually, cleaners provide tactile stimulation in response to the behaviour of the client; for example, manipulating clients that are unwilling to interact. The manipulation serves to slow down the clients, allowing the cleaners to forage on the clients’ surface (Bshary & W¨ urth, 2001). As tactile stimulation of small residents did not appear to provide the cleaners with any direct benefits from the recipients, Bshary (2002) proposed that it may serve as a signal to attract imagescoring clients, which can then be exploited. In line with this argument, it was found that interactions that consisted of tactile stimulation only were followed

531

P1: JZZ/... P2: JZZ/... 0521823617c22.xml CU1917B/McGregor

April 11, 2005

19:30

R. Bshary & A. D’Souza

p

C

532

0 521 582361 7

B

N

Fig. 22.6. Frequencies with which biting and normal cleaners ignored the invitations for inspection of small resident clients. Values are the median and interquartile ranges for five biting and eleven normal cleaners. (The p value is derived from Mann–Whitney U-test.)

by interactions ending with a conflict immediately after a client jolt more often than expected. It appears that tactile stimulation of small residents is a signal out of context that attracts image-scoring clients to their own disadvantage (they will be cheated) and to the cleaners’ advantage, fulfilling the functional definition of tactical deception (Hauser, 1998). The presence or absence of bystanders was not noted, so it remains unclear whether biting cleaners seek small residents in particular when larger clients are nearby or whether they start such interactions independently of the presence of bystanders. The latter scenario is more plausible, as larger clients, in particular the choosy ones, are not willing to queue for inspection (Bshary & Sch¨ affer, 2002) and would, therefore, swim off despite the cleaner’spositive image. So, cleaners appear unable to time interactions with small residents for maximal effects. However, some evidence suggests that tactile stimulation of small residents is part of biting cleaners’ strategies to improve their image. Cleaner fish sometimes ignore clients that invite inspection, in particular small resident clients (Bshary & W¨ urth, 2001). These clients do not offer a large food source and do not have the option to visit another cleaner if ignored. When approached by small residents, biting cleaners ignore them significantly less frequently than normal cleaners (Mann-Whitney U-test: n = 11; m = 5; U = 2; p = 0.004; Fig. 22.6).

Discussion We have provided a description of behavioural patterns in interactions between cleaner fish and client reef fish that emphasizes the importance of the communication-network framework in understanding the dynamics of cooperative interactions and the occurrence of tactical deception. Cleaning interactions often occur in the presence of other potential clients of cleaners. These bystanders eavesdrop on ongoing interactions, and the information that they collect appears

P1: JZZ/... P2: JZZ/... 0521823617c22.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:30

Indirect reciprocity in interactions in fish to be crucial for their decision to invite inspection or to avoid the cleaner. In response, it appears that normal cleaners reduce cheating frequencies in the presence of eavesdroppers, in particular if these eavesdroppers have access to several cleaning stations. Data of this kind are still missing for biting cleaners. Biting cleaners frequently engage in costly (or at least non-profit) interactions with small residents that appear to serve to attract larger image-scoring clients, which can then be exploited. The results have important implications for theoretical approaches to indirect reciprocity. Existing models predict that image scoring drives altruistic behaviour towards fixation (Nowak & Sigmund, 1998; Lotem et al., 1999; Leimar & Hammerstein, 2001). Cheating individuals can only reinvade an image-scoring population after genetic drift has led to an increase in non-discriminatory altruists. This scenario does not fit the cleaner fish mutualism very well. Image scoring of clients mediates cooperative behaviour of cleaners, but this cooperative behaviour may be an honest or a deceptive signal. Cheating individual cleaners use one class of clients for altruistic behaviour to produce a signal that allows them to exploit, through image scoring, another class of clients. Image scoring thus works for the receiver of altruistic behaviour but it does not always work for the eavesdropper. The major difference between the cleaner fish system and the models might concern the payoff matrix. While it is assumed in the models that payoffs are the same in every interaction, payoffs are variable for cleaners. First, an advantage of cheating clients with access to several cleaning stations rather than resident clients is that the former just swim off after being cheated, while the latter chase the cleaner fish around (Bshary & Grutter, 2002a), so the cleaner loses some of the energy it has just gained. Second, cleaners can probably gain very little from interactions with small clients anyway, no matter whether they cooperate or cheat. This contrasts with interactions with large clients, which have more parasites but also more mucus and a larger surface for the cleaner to scrape along with its lower jaw. This gives the opportunity for cleaners to behave altruistically when payoffs are low and to be exploitative when payoffs are high, as long as the altruism increases the frequency of high payoff interactions. Image scoring would not persist if it did not yield a benefit to its performer, but it also provides an opportunity for individuals to perform altruistic acts in order to gain access to and exploit image-scoring individuals. Therefore, image scoring in communication networks may explain both the evolution of altruistic behaviour and the occurrence of tactical deception. The commonness of dishonest signals that nevertheless still fool observers has yet to be evaluated. While verbal arguments predicted low frequencies (Dawkins & Krebs, 1978), game theoretic models indicate that this is not necessarily the case (Johnstone & Grafen, 1993; Szamado, 2000). In particular, if the benefits of finding a cooperative partner largely outweigh the cost of

533

P1: JZZ/... P2: JZZ/... 0521823617c22.xml CU1917B/McGregor

534

0 521 582361 7

April 11, 2005

19:30

R. Bshary & A. D’Souza interacting with a cheating partner, tactical deception may occur at quite high frequencies. Future work with the cleaner system

Several important points of the cleaner fish mutualism still have to be clarified. First, we need experimental evidence for both client image scoring and cleaner fish audience effects. There is increasing evidence that a reduction in client jolt rates in the presence of bystanders reflects a more cooperative behaviour by cleaners (Bshary & Grutter, 2002b) and that such behaviour is indeed more altruistic. Grutter & Bshary (2003) offered cleaners the choice between equal amounts of mucus, gnathiid isopods and monogeneans attached to plexiglas plates and found that cleaners ate mucus more often than parasites, in particular gnathiids. Assuming that the results reflect the items’quality as a food source, cleaners profit even more from feeding on mucus when interacting with real clients as mucus is abundantly spread over the clients’ surface whereas parasites have to be searched for. In conclusion, while the experiment did not quantify energy intake, it makes it very plausible that feeding on mucus yields a higher energy gain than feeding on parasites. Another point that needs to be addressed is the biting cleaners’ behaviour with respect to small and large clients. Is it really true that interactions with small clients generally offer low payoffs compared with interactions with larger clients, and that the margin between the benefits from cooperation and cheating increase with client size? Does image scoring of clients indeed inhibit the cleaners’ tendency to cheat in low-payoff interactions but not in high-payoff interactions? Currently, no data are available to evaluate these questions. Finally, one might expect that clients should respond to the biting cleaners’ behaviour by fine-tuning their image scoring, paying less attention to the outcome of interactions between cleaners and small residents. Pooling of existing data indicate that this is indeed the case. Invitation for inspection (i.e. spreading the pectoral fins and stopping coordinated swimming movements) occurred more frequently if cleaners interacted with choosy clients than with small residents. This preliminary result has to be tested with a larger data set that allows statistical analysis based on behaviour of individual client species. Cognitive aspects

While indirect reciprocity and tactical deception were considered to be a hallmark of human evolution (Nowak & Sigmund, 1998) and primate Machiavellian intelligence (Byrne & Whiten, 1988), the data on cleaner–client interactions suggest that, on a purely descriptive level ignoring underlying mechanisms, these phenomena are more widespread. We have argued that they should occur

P1: JZZ/... P2: JZZ/... 0521823617c22.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:30

Indirect reciprocity in interactions in fish frequently in social networks. These phenomena should occur if it pays to alter the optimal behaviour in a situation in order to alter one’s image, which will, in turn, produce benefits during future interactions with bystanders that exceed the momentary costs. With respect to aggression, a game theoretic model shows that it may even pay individuals to act spitefully towards a partner (in the sense that the spiteful act will not lead to any benefits gained from future behaviour of the recipient) if this spiteful act reduces, for example, the threat of attack from bystanders (Johnstone & Bshary, 2004; Ch. 10). With respect to cognition, the data generally support the view of Heyes (1998) that we need to establish what kind of information animals use for their decision making to find out what cognitive abilities are involved in a given phenomenon. Indirect reciprocity and tactical deception may be something ‘smart’in some species and simple conditioning in others. Cleaners certainly have ideal conditions to develop their behaviour through conditioning. They have more than 2000 interactions per day (Grutter, 1995), making it easy to connect altruistic behaviour with reward (i.e. the invitation from bystanders to inspect) and cheating with punishment (i.e. evasive actions of bystanders when approached by the cleaner). In the absence of decisive experiments, it could even be possible that parts of cleaner fish and client behaviour may be governed by endocrine responses rather than through learning (Ch. 21). A good candidate for an endocrine-mediated behaviour might be the good service that cleaners provide to predatory clients: there might be an innate programme to recognize predators, and the presence of a predator might trigger a stress response that, in turn, may inhibit cheating behaviour. Alternatively, one could also generalize the Machiavellian intelligence hypothesis and predict that a complex social network should have similar effects on cognitive abilities in all species (Byrne & Whiten, 1988). It may turn out that cleaners have relatively high cognitive abilities, as their large interspecific social network is at least in part based on individual recognition (Tebbich et al., 2002) and demands the solving of a variety of problems (Bshary et al., 2002). In this context, it is worth pointing out that both the biting females which were observed over longer time periods showed considerable variation in their behaviour, as so did their male partners. One male often prevented his female from interacting with clients while the other did not (Fisher test: n = 34; p = 0.003; Table 22.1). The preventive male was also almost significantly more likely to chase his female when their client darted off after a jolt, while the other male often followed the client to provide tactile stimulation (Fisher test: n = 12; p = 0.053; Table 22.1). This observation of flexibility of both males and females is important as it was the careful description of individual-specific strategies in primates that eventually led to a cognitive, rather than genetic, approach towards behaviour (Strum et al., 1998).

535

P1: JZZ/... P2: JZZ/... 0521823617c22.xml CU1917B/McGregor

536

0 521 582361 7

April 11, 2005

19:30

R. Bshary & A. D’Souza Table 22.1. The behaviour of two males that were partners of biting females Male responses to female biting Tactile

Preventative

stimulation

Chasing

No obvious

chasing of

clienta

femalea

reactiona

femaleb

of First male

5

2

3

1

Second male

0

5

2

16

a

Reaction to clients darting off following cheating by female.

b

Keeping female away from clients.

Summary In summary, we think that the concept of communication networks has major implications for our understanding of the evolution and maintenance of altruistic behaviour, tactical deception and spiteful behaviour. Because of its functional approach, the communication-network framework may help to demystify phenomena that are often considered to demand high cognitive abilities, opening the way to focus on the underlying mechanisms and the complexity of information processing and decision rules in order to illuminate cognitive differences between species (for a parallel discussion, see Ch. 24). Game theory models should help to generate testable predictions of the circumstances in which altruism, tactical deception and spiteful behaviour may yield fitness benefits within communication networks. In particular, it is time to develop cognitive models rather than genetic models, allowing individuals to process information about their social environment before making a behavioural decision (see Stephens & Clements (1998) for a first approach towards cognitive game theory).

Acknowledgements We thank Peter McGregor for inviting us to write this chapter. We are grateful to the EEAA in Cairo for the permit to work in the Park and to Alain de Grissac, the Park rangers and Ingo Riepl for their support at the Park. The study was supported by the Deutsche Forschungsgemeinschaft (grants BS 2/2-1 to BS 2/2-4) and written while RB was on a Marie Curie Fellowship of the EU. We want to thank Wolfgang Wickler, Karin Bergmann and Barbara Knauer for additional support. The chapter was greatly improved by comments from Peter McGregor, Rui Oliveira, Alexandre Roulin, Sabine Tebbich and an anonymous referee.

P1: JZZ/... P2: JZZ/... 0521823617c22.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:30

Bibliography References Alexander, R. D. 1987. The Biology of Moral Systems. New York: Aldine de Gruyter. Axelrod, R. & Hamilton, W. D. 1981. On the evolution of co-operation. Science, 211, 1390–1396. Bshary, R. 2001. The cleaner fish market. In: Economics in Nature, ed. R. No¨e, J. A. R. A. M. van Hooff & P. Hammerstein. Cambridge, UK: Cambridge University Press, pp. 146–172. 2002. Biting cleaner fish use altruism to deceive image scoring clients. Proceedings of the Royal Society of London, Series B, 269, 2087–2093. Bshary, R. & Grutter, A. S. 2002a. Experimental evidence that partner choice is the driving force in the payoff distribution among cooperators or mutualists: the cleaner fish case. Ecology Letters, 5, 130–136. 2002b. Asymmetric cheating opportunities and partner control in a cleaner fish mutualism. Animal Behaviour, 63, 547–555. Bshary, R. & Sch¨ affer, D. 2002. Choosy reef fish select cleaner fish that provide high service quality. Animal Behaviour, 63, 557–564. Bshary, R. & W¨ urth, M. 2001. Cleaner fish Labroides dimidiatus manipulate client reef fish by providing tactile stimulation. Proceedings of the Royal Society of London, Series B, 268, 1495–1501. Bshary, R., Wickler, W. & Fricke, H. 2002. Fish cognition: a primate’s eye view. Animal Cognition, 5, 1–13. Byrne, R. W. & Whiten, A. 1988. Machiavellian Intelligence. Oxford: Clarendon Press. Cˆ ot´e, I. M. 2000. Evolution and ecology of cleaning symbioses in the sea. Oceanography and Marine Biology Annual Review, 38, 311–355. Dawkins, R. & Krebs, J. R. 1978. Animal signals: information or manipulation? In: Behavioural Ecology: An Evolutionary Approach, ed. J. R. Krebs & N. B. Davies. Oxford: Blackwell, pp. 282–309. Doutrelant, C., McGregor, P. K. & Oliveira, R. F. 2001. The effect of an audience on intrasexual communication in male Siamese fighting fish, Betta splendens. Behavioral Ecology, 12, 283–286. Dugatkin, L. A. 1997. Cooperation among Animals: A Modern Perspective. Oxford: Oxford University Press. Grutter, A. S. 1995. Relationship between cleaning rates and ectoparasite loads in coral reef fishes. Marine Ecology Progress Series, 118, 51–58. 1996. Parasite removal rates by the cleaner wrasse Labroides dimidiatus. Marine Ecology Progress Series, 130, 61–70. 1997. Spatio-temporal variation and feeding selectivity in the diet of the cleaner fish Labroides dimidiatus. Copeia, 1997, 346–355. Grutter, A. S. & Bshary, R. 2003. Cleaner wrasse prefer client mucus: support for partner control mechanisms in cleaning interactions. Proceedings of the Royal Society of London, Series B, Biology Letters Supplement, 2, 242–244. Hamilton, W. D. 1964. The genetical evolution of social behaviour. Journal of Theoretical Biology, 7, 1–52.

537

P1: JZZ/... P2: JZZ/... 0521823617c22.xml CU1917B/McGregor

538

0 521 582361 7

April 11, 2005

19:30

R. Bshary & A. D’Souza Hauser, M. D. 1998. Minding the behaviour of deception. In: Machiavellian Intelligence II, ed. A. Whiten & D. W. Byrne. Cambridge, UK: Cambridge University Press, pp. 112–143. Heyes, C. M. 1998. Theory of mind in non human primates. Behavioral and Brain Sciences, 21, 101–148. Johnstone, R. A. 2001. Eavesdropping and animal conflict. Proceedings of the National Academy of Sciences, USA, 98, 9177–9180. Johnstone, R. A. & Bshary, R. 2004. The evolution of spite through indirect reciprocity. Proceedings of the Royal Society of London, Series B, 271, 1917–1922. Johnstone, R. A. & Grafen, A. 1993. Dishonesty and the handicap principle. Animal Behaviour, 46, 759–764. Leimar, O. & Hammerstein, P. 2001. Evolution of cooperation through indirect reciprocity. Proceedings of the Royal Society of London, Series B, 268, 745–753. Losey, G. C., Grutter, A. S., Rosenquist, G., Mahon, J. L. & Zamzow, J. P. 1999. Cleaning symbiosis: a review. In: Behaviour and Conservation of Littoral Fishes, ed. V. C. Almada, R. F. Oliveira & E. J. Goncalves. Lisbon: Instituto Superior de Psicologia Aplicada, pp. 379–395. Lotem, A., Fishman, M. A. & Stone, L. 1999. Evolution of cooperation between individuals. Nature, 400, 226–227. 2003. From reciprocity to unconditional altruism through signaling benefit. Proceedings of the Royal Society of London, Series B, 270, 199–205. McGregor, P. K. 1993. Signalling in territorial systems: a context for individual identification, ranging and eavesdropping. Philosophical Transactions of the Royal Society of London, Series B, 340, 237–244. Munn, C. A. 1986. Birds that ‘cry wolf’. Nature, 319, 143–145. No¨e, R. 2001. Biological markets: partner choice as the driving force behind the evolution of cooperation. In: Economics in Nature, ed. R. No¨e, J. A. R. A. M. van Hooff & P. Hammerstein. Cambridge, UK: Cambridge University Press, pp. 92–118. No¨e, R., van Schaik, C. P. & van Hooff, J. A. R. A. M. 1991. The market effect: an explanation for pay-off asymmetries among collaborating animals. Ethology, 87, 97–118. Nowak, M. A. & Sigmund, K. 1998. Evolution of indirect reciprocity by image scoring. Nature, 393, 573–577. Premack, D. & Woodruff, G. 1978. Does the chimpanzee have a theory of mind? Behavioral and Brain Sciences, 4, 515–526. Randall, J. E. 1958. A review of labrid fish genus Labroides, with descriptions of two new species and notes on ecology. Pacific Scientist, 12, 327–347. Roberts, G. 1998. Competitive altruism: from reciprocity to the handicap principle. Proceedings of the Royal Society of London, Series B, 265, 427–431. Robertson, D. R. 1972. Social control of sex reversal in a coral-reef fish. Science, 177, 1007–1009. Stephens, D. W. & Clements, K. C. 1998. Game theory and learning. In: Game theory and Animal Behaviour, ed. L. A. Dugatkin & H. K. Reeve. Oxford: Oxford University Press, pp. 239–260.

P1: JZZ/... P2: JZZ/... 0521823617c22.xml CU1917B/McGregor

0 521 582361 7

April 11, 2005

19:30

Indirect reciprocity in interactions in fish Strum, S. C., Forster, D. & Hutchins, E. 1998. Why machiavellian intelligence may not be machiavellian. In: Machiavellian Intelligence II, ed. A. Whiten & D. W. Byrne. Cambridge, UK: Cambridge University Press, pp. 50–85. Szamado, S. 2000. Cheating as a mixed strategy in a simple model of aggressive communication. Animal Behaviour, 59, 221–230. Tebbich, S., Bshary, R. & Grutter, A. S. 2002. Cleaner fish Labroides dimidiatus recognise familiar clients. Animal Cognition, 5, 139–145. Trivers, R. L. 1971. The evolution of reciprocal altruism. Quarterly Review of Biology, 46, 35–57. Wedekind, C. & Milinski, M. 2000. Cooperation through image scoring in humans. Science, 288, 850–852. Zahavi, A. 1995. Altruism as a handicap : the limitations of kin selection and reciprocity. Journal of Avian Biology, 26, 1–3.

539

P1: IYP/... P2: IYP 0521823617c23.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

23

Fish semiochemicals and the evolution of communication networks brian d. wisenden1 & norman e. stacey2 1 2

Minnesota State University, Moorhead, USA University of Alberta, Edmonton, Canada

Introduction The concept that animals typically communicate in networks (involving at least one signaller and more than one receiver) derives from the active space of signals and social spacing of conspecific and heterospecific receivers (McGregor & Peake, 2000; Ch. 1). The ecological and evolutionary consequences of such networks have been explored most thoroughly for visual (e.g. Ch. 12) and acoustic signals (e.g. Otter et al., 1999; Ch. 2), although it is clear that chemical signalling also can involve networks (Chs. 11 and 16). Research on aquatic communication networks has so far been limited to the context of visual and acoustic signalling (e.g. Oliveira et al., 1998; Chs. 5 and 18). Semiochemicals (i.e. chemicals that transfer information within and/or between species) exert important and diverse effects on the behaviour and physiology of aquatic animals (Liley, 1982; Chivers & Smith, 1998; Kats & Dill, 1998; Sorensen & Stacey, 1999; Stacey & Sorensen, 2002; Wisenden, 2003). Studies of two key aspects of fish chemical ecology (predator– prey and reproductive interactions) have revealed great differences in the sources and nature of the semiochemicals released, their active spaces and their biological functions. These studies also provide sufficient information to assess, in fish, the existence and function of semiochemical information networks, which we define more fully below as a general category of network that includes not only communication networks employing specialized signals but also other networks employing unspecialized cues. Here, we briefly describe well-studied examples of predator–prey and reproductive semiochemicals to explore the applicability of current communication network theory to aquatic chemical information networks

Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

540

P1: IYP/... P2: IYP 0521823617c23.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

Fish semiochemicals and consider how their function and evolution might differ from those employing other sensory modalities and information-transmission media. Research on intra- and interspecific transfer of chemical information in terrestrial species has generated a bewildering terminology related not only to the nature, actions and functions of the chemicals but also to the concept of communication (Hauser, 1996; Beauchamp, 2000; Hasson, 2000; McClintock, 2002). Although we do not presume to clarify such a complex terminological problem in this brief paper, it is imperative that we begin by clearly defining key terms, particularly as much of our subject matter appears to be distinct from that typically discussed in the context of communication networks. Definition of terminology

Semiochemicals include allomones and pheromones that, respectively, transmit interspecific and intraspecific information. We will consider fish pheromones and allomones involved in predator–prey interactions (p. 544) separately from those involved in reproduction (p. 549). We define a pheromone as ‘a substance, or mixture of substances, which is released by an individual and that evokes a specific and adaptive response in conspecifics’ (Stacey & Sorensen, 2002). This definition is more inclusive than the original definition of pheromone (Karlson & L¨ uscher, 1959) because, for reasons explained below, it omits any requirement that pheromones be involved in communication. We use the terms releaser and primer not to classify pheromones but only to describe their rapid behavioural and slower physiological actions, respectively, for the simple reason that ‘it is quite possible for the same pheromone to be both a releaser and a primer’ (Wilson & Bossert, 1963), as is the case for sex pheromones of goldfish Carassius auratus (p. 549). Central to our terminological schema is the concept (Sorensen & Stacey, 1999; Stacey & Sorensen, 2002) that evolution of chemical communication progresses through a series of three functional phases: ancestral, spying and communication. In the ancestral phase, individuals (originators) release a chemical(s) that does not influence receivers (Fig. 23.1). This primitive, prepheromonal condition progresses to spying if receivers evolve the ability to detect and respond adaptively to the originator’s released chemical(s), now termed a pheromonal or allomonal cue(s). In spying, originators may or may not benefit from the receiver’s response but, importantly, remain in an unspecialized state with respect to production and release of pheromonal cues. Finally, spying progresses to communication if there is a mechanism for receiver responses to select for specialization in production and/or release of the detected cue(s), now termed a pheromonal or allomonal signal(s) and released by a signaller. Signals evolve through natural selection because of fitness benefits the signaller receives by manipulating the behaviour or physiology of receivers. In many cases, signal senders and receivers form a mutualism in which

541

P1: IYP/... P2: IYP 0521823617c23.xml

542

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

B. D. Wisenden & N. E. Stacey

Fig. 23.1. The evolution of communication from the ancestral state, where the originator does not possess specializations for synthesis and release of semiochemicals, to spying, where receivers possess specializations for detecting semiochemicals but originators do not possess specializations, to true communication, in which both originator (now signaller) and receiver possess specializations for semiochemical exchange of information.

signals coevolve with the sensory biology of receivers. Once established, however, mutually beneficial communicative relationships could be susceptible to deceitful signal manipulations by signallers, which reduce the receiver fitness, as seen in visual, acoustic and chemical signals (e.g. Lloyd, 1965; Møller, 1989; Paxton & Tengo, 2001). The ancestral state applies to released chemicals not currently functioning in spying or communication. We restrict the terms signal and communication to those situations in which there is clear evidence for signal specialization, such as tissue hypertrophy or discrete structures for signal production; in contrast to the situation in terrestrial insects and vertebrates, where pheromone-producing

P1: IYP/... P2: IYP 0521823617c23.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

Fish semiochemicals glandular structures are common, such specializations in fish appear to be the exception rather than the rule (e.g. Laumen et al., 1974; Colombo et al., 1980; van den Hurk & Resink, 1992). Consequently, we regard the great majority of fish predator– prey and reproductive semiochemicals to function in spying, which is, in effect, the default condition for cases where there is no evidence for specialization in semiochemical production or release and/or where the social system apparently precludes selection for signal specialization. It is to be expected that future research may reveal some putative examples of spying to be true communication because they involve previously undetected signal specialization. Although we believe the distinction between cues and signals is fundamental to an understanding of the function and evolution of semiochemical systems, fish olfactory systems evidently do not make this distinction and process cues and signals through similar mechanisms, which differ considerably from those processing food odours (amino acids). Therefore, in comparison with food odours, semiochemical cues and signals are detected by more sensitive and specific olfactory receptor mechanisms and generate neuronal activity that is processed in distinct arrays (glomeruli) in the olfactory bulbs, is conducted to the brain by distinct nerve bundlets (olfactory tracts), and is projected to distinct brain areas (Sorensen et al., 1998; Hamdani et al., 2000, 2001; Brown et al., 2001; Stacey & Sorensen, 2002). The distinction between chemical spying (via cues) and chemical communication (via signals) highlights a dichotomy relevant not only to our understanding of semiochemicals (the functional relationships among originators, signallers and receivers; evolutionary origins of species-specific cues and signals: Sorensen & Stacey (1999)) but also to the concept of communication networks (McGregor & Peake, 2000). In particular, first, how might networks involving communication differ from those involving spying and, second, can the concept of eavesdropping, defined as ‘extracting information from signalling interactions between others’ (McGregor and Peake, 2000) be applied to information networks that do not involve signalling?

Transfer of chemical information Propagation of chemical information differs fundamentally from propagation of visual and acoustic information. In general, visual and acoustic signals are propagated with predictable speed and direction, and they generate predictable active spaces throughout which much of the temporal information contained in the signal’s initial pattern can be retained. In contrast, semiochemicals are released into fluid media (air or water) in which local variation in flow typically creates turbulent odour plumes, which not only distort or destroy temporal pattern but also make the position, shape and size of the chemical’s active

543

P1: IYP/... P2: IYP 0521823617c23.xml

544

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

B. D. Wisenden & N. E. Stacey space highly unpredictable (Weissburg, 2000). Moreover, semiochemicals can persist in the environment for considerable time (e.g. Wisenden et al., 1995; Sorensen et al., 2000; Polkinghorne et al., 2001), and thus can become disassociated from originators/signallers either when currents carry away a transiently released semiochemical or when the originator/signaller moves to a new location. Although there is considerable information on the mechanisms by which some invertebrates (e.g. crustaceans and moths) navigate in physically characterized odour plumes, this complex issue is poorly understood in fish (Vickers, 2000). Finally, it is important to realise that semiochemical function in water also can be influenced by additional solutes that affect olfactory response, such as heavy metals (Hansen et al., 1999) and organics (Hubbard et al., 2002). The olfactory system is similar to other sensory systems in being functionally delimited by the sensitivity and specificity of its sensory neurons, but it differs in the nature of the information it processes. Visual and acoustic systems process linear arrays of light and sound frequencies in spectra common to many species, particularly if they are related; olfactory systems process information from odorants that cannot be arranged in a linear dimension by means of receptors that are sensitive to one or a few chemicals. These differences have two important implications for the nature and evolution of semiochemicals. First, whereas visual and acoustic signals usually encode species-typical information in frequency and temporal pattern, semiochemicals encode this information through the presence, absence or ratio of specific odorants. Second, whereas visual and acoustic signals are potentially detectable by all individuals and species sensitive to the emitted spectra, semiochemical detection will be restricted to individuals with olfactory receptors sensitive to the odorant(s). Thus, large differences in semiochemical production and detection can occur with only small changes either in chemical metabolism and release or in olfactory receptor specificity.

Assessment of predation risk Natural selection strongly promotes attendance to cues that reduce the probability of predation. Consequently, temporal and spatial variation in predation risk governs much of animal behaviour. Chemicals reliably inform about predation risk because they are carried well in water, persist for ecologically appropriate amounts of time, transmit information through turbid or highly structured habitat and darkness, and provide types of information not contained in visual and acoustic modalities. To apply communication-network theory to chemical assessment of predation risk, we must first determine whether use of chemical information for the purposes of risk assessment involves spying (via cues) or communication (via signals).

P1: IYP/... P2: IYP 0521823617c23.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

Fish semiochemicals

Fig. 23.2. Semiochemicals associated with predation. Predation escalates from initial detection (top) to attack (middle) and, finally, ingestion (bottom). Chemical cues (solid arrows) released at each stage inform nearby prey (conspecific and heterospecific) and predators of the presence and extent of the interaction between predator and prey. Known and hypothesized benefits are indicated by dotted arrows.

We conclude in the discussion below that, despite a plethora of semiochemically mediated mechanisms for predator avoidance, evidence for signals is not compelling. Although these information networks may not be communication networks per se, there is evolutionary opportunity for communication networks and eavesdropping to evolve because receivers have evolved the ability to detect and respond to many types of semiochemical (see below). Chemicals correlated with predation

The literature concerning chemicals linked to predation has been reviewed elsewhere (Smith, 1992; Chivers & Smith, 1998; Kats & Dill, 1998; Wisenden, 2000, 2003; Chivers & Mirza, 2001) and only a brief overview will be presented here. Several classes of chemical compound inform prey about predation risk. Generally, these cues are released passively before, during and after a predation event (Fig. 23.2). Before an attack is initiated, prey can detect and respond to three types of chemical cues: (a) odour of disturbed (startled but uninjured) prey (Chivers & Smith, 1998; Wisenden, 2003), (b) species-specific kairomones (a predator’s natural odour) (Kats & Dill, 1998) and (c) injury-released alarm cues of prey that leak from

545

P1: IYP/... P2: IYP 0521823617c23.xml

546

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

B. D. Wisenden & N. E. Stacey the gut of the predator (Chivers & Mirza, 2001). When a predator attacks and injures a prey organism, damaged prey tissues release chemical compounds that are released only in this context; consequently, these cues reliably indicate risk and elicit intense anti-predator behaviour (Chivers & Smith, 1998). These are alarm cues. Most aquatic taxa exhibit anti-predator behaviour in response to alarm cues (Chivers & Smith, 1998; Wisenden, 2003) in ways that reduce the probability of predation (Hews, 1988; Mathis & Smith, 1993; Wisenden et al., 1999; Gazdewich & Chivers, 2002). Ingested prey release alarm cues, or their metabolites, from the gut of their predators (Chivers & Mirza, 2001). The ecological reality is undoubtedly more complex than the interactions depicted in Fig. 23.2. Additional interactions arise from variation in (a) diet breadth and overlap among predators, (b) relative threat from each predator species over time and space, (c) interacting ontogenies of prey and predator species, and (d) learned behavioural responses to correlates of alarm cues. The vast majority of chemical information used by aquatic prey to assess predation risk appears to be opportunistic use of chemical information mediated by unspecialized chemical cues. This information is of great fitness benefit to receivers, but receiver response, with one notable exception discussed below, generally has not been shown to accrue benefit to the originator/signaller. Chemically mediated predation risk might be described most parsimoniously as an information network, where a suite of prey species spy on the foraging activities of a suite of predator species. Ostariophysan alarm substance cells

For passively released chemical cues to qualify as signals, specializations for their synthesis and/or release must occur that plausibly have been selected for by benefits accruing to the originator/signaller. This condition appears to be met in fishes of the superorder Ostariophysi (minnows, tetras, catfishes, suckers and sundry others). This large group of vertebrates (> 5500 species) makes up approximately 27% of the global ichthyofauna and 64% of all freshwater fish species (Nelson, 1994). In addition to successful occupation of a diverse array of habitats, they are often the numerically dominant vertebrates in aquatic ecosystems. Ostariophysans possess specialized epidermal cells that contain a potent alarm chemical(s), termed schreckstoff or alarm substance (von Frisch, 1941; Pfeiffer, 1977; Smith, 1992); this appears to activate components of the olfactory system that also are activated by sex pheromones (Hamdani et al., 2000). It is not known how much skin area is typically damaged during a predatory attack, but homogenates of 1 cm2 skin can create active spaces of 10 000 litres (zebrafish Danio rerio; Gandolfi et al., 1968) to 58 000 litres (fathead minnow Pimephales promelas; Lawrence & Smith, 1989), equivalent to spheres 2.6–4.8 m in diameter. The active ingredient in

P1: IYP/... P2: IYP 0521823617c23.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

Fish semiochemicals ostariophysan alarm cells is likely, at least in part, to be hypoxanthine 3N-oxide, a compound first isolated from European minnows Phoxinus phoxinus (Argentini, 1976; after Smith, 1999). Subsequent work demonstrated that hypoxanthine 3Noxide elicits anti-predator behaviour from a characin (Pfeiffer et al., 1985) and from fathead minnows at concentrations as low as 0.4 nmol/l (Brown et al., 2001). However, efforts to detect hypoxanthine 3N-oxide in fathead minnow skin with high performance liquid chromatography have not been successful (Smith, 1999) and fractionation of skin extract indicates the biologically active component is found with the polypeptides with molecular weights greater than 1100 (Kasumyan & Ponomarev, 1987) rather than with the small molecules such as hypoxanthine 3Noxide. It is possible that hypoxanthine 3N-oxide is associated with protein while within skin cells and remains associated with protein once released. Heat-treated skin extract of fathead minnows loses 70% of its protein and its ability to elicit alarm (N. L. Korpi, L. D. Louisiana, J. J. Provost & B. D. Wisenden, unpublished data). A protein–hypoxanthine association would be consistent with cross-species alarm reactions that decline with phylogenetic distance (Schutz, 1956). Whatever the active ingredient(s) of alarm cue might be, their biological potency (Lawrence & Smith, 1989; Brown et al., 2001) suggests selection for olfactory sensitivity similar to that seen with sex pheromones (Stacey & Sorensen, 2002). Is ostariophysan alarm substance a passively released cue or a specialized signal? Although the epidermal cells appear to be structures specialized for information transfer of alarm, selection for signal specialization via benefits to the originator (i.e. the individual that released the substance) is not immediately apparent. There has been much speculation over the historical and current selection benefits to individuals that invest in these cells (Smith, 1992, 1997; Williams, 1992; Magurran et al., 1996; Henderson et al., 1997). Smith (1992) summarized 16 hypotheses by which signallers may benefit from alarm signalling. One of these, attraction of secondary predators (Fig. 23.2), has empirical support. Laboratory and field experiments have demonstrated that predators are attracted to minnow skin extract containing alarm substance cells over minnow skin lacking alarm substance cells or skin extract from non–ostariophysan species (Mathis et al., 1995; Wisenden & Thiel, 2002). Interruption of a predation event by the arrival of a second predator allows prey an opportunity to escape (Chivers et al., 1996), a benefit that elevates passively released cues to signal status. In this context, however, this signal is not an alarm signal, but an attractant signal because the signaller benefits because the from the responses of secondary predators, rather than from responses of conspecific and heterospecific members of the prey community. Is this a communication network? Are members of the prey community eavesdroppers on the predator-attractant signal? Assessment of predation risk via chemical cues does not depend on receiver (secondary predators) response (i.e. not social

547

P1: IYP/... P2: IYP 0521823617c23.xml

548

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

B. D. Wisenden & N. E. Stacey eavesdropping) but interceptive eavesdropping (defined in Ch. 2) on this attractant signal provides highly salient temporal and contextual information about predation risk. Therefore, from the perspective of the general non-ostariophysan prey community, detection of injury-released chemical compounds may be considered as (interceptive) spying; however for the ostariophysan fishes, spying on the predator-attractant signal might best be considered as a case of interceptive eavesdropping (Ch. 2).

Evolutionary opportunities for communication networking

Several lines of evidence suggest potential for communication networks in the ostariophysan alarm cue system. First, fishes frequently survive predatory attacks (Smith & Lemly, 1986). An originator/signaller that survives an attack may benefit from group behavioural responses of the prey community (Smith, 1992; Fig. 23.2) increased shoal cohesion and dashing or skittering behaviour that confuse predators and reduce attack efficiency. Second, minnows associate alarm cues with correlates of predation such as predator appearance and odour (reviewed by Chivers & Smith, 1998). Alarm cues enable conspecifics and heterospecifics to acquire predator recognition after a single simultaneous or non-simultaneous encounter with a novel indicator of risk (Suboski, 1990; Suboski et al., 1990; Chivers & Smith, 1994; Hall & Suboski, 1995; Korpi & Wisenden, 2001). Therefore, a second benefit to the signaller could be providing shoalmates with an opportunity to learn predator identity, as a shoalmate trained in this way may detect that predator in the future and alert the signaller. If a shoal contains individuals related to the signaller, then a third benefit might accrue to the signaller’s inclusive fitness through kin selection. In this context, eavesdropping on signaller–group communication in the ostariophysan system could occur when minnows observe the visual stimulus of anti-predator behaviour of an alarmed shoal (Verheijen, 1956; Magurran, 1989; Suboski et al., 1990; Brown et al., 1999). Although evolutionary ecologists have focused on cells producing alarm substances in ostariophysans, these fishes are not unique in possessing specialized epidermal cells (Smith, 1992). The epidermal layer of freshwater perch, walleye and darters (superorder Acanthopterygii, order Perciformes, family Percidae: Smith, 1979, 1982; Wisenden, 2003), Australian bullies (order Perciformes, family Eleotridae: Kristensen & Closs, 2004) and poeciliids (superorder Acanthopterygii, Order Cyprinodontiformes, family Poeciliidae: Bryant, 1987) all possess epidermal club cells with similar histological properties. The tropical marine and freshwater fishes in the Gobiidae (order Perciformes, 1875 species) possess epidermal vacuolate cells but have an inconsistent behavioural response to skin extract (Smith,

P1: IYP/... P2: IYP 0521823617c23.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

Fish semiochemicals 1992). This leaves open the possibility that analogous communication networks for assessment of predation risk occur among other fish taxa. In summary, it is parsimonious, based on current knowledge of chemical alarm cues, to conclude that for aquatic taxa, including most fishes, chemically mediated risk assessment does not constitute a true communication network because it is not based on specialized signals. However, the ostariophysan alarm semiochemical system appears to be a good candidate for an incipient communication system (see p. 558). Future research may reveal signaller–group communication of alarm and potentially uncover communication networks and eavesdropping.

Sex pheromones in information networks In addition to alarm responses discussed in the previous section, pheromonal cues and signals of teleost fish influence many diverse nonreproductive (migration, parent–young interactions, schooling and related social behaviours: Liley, 1982) and reproductive (Stacey et al., 1986; Stacey & Sorensen, 2002) phenomena. Best understood are those cases (the great majority being reproductive) in which chemical identification has allowed study of pheromone production, detection and biological effects under controlled and repeatable conditions. Since Colombo et al. (1980) first proposed that the male black goby, Gobius niger, releases a conjugated steroid (etiocholanolone glucuronide) to function as a pheromone that attracts the female to his nest for spawning, many studies have reported putative pheromonal roles for steroid and prostaglandin hormones, and their precursors and metabolites (hereafter termed hormonal pheromones) in a variety of fish (reviewed by Sorensen & Stacey, 1999; Stacey & Sorensen, 2002). Indeed, we expect the use of hormonal pheromones might be universal among fish, given that information-rich hormones and hormonal metabolites are necessarily released into the same water medium bathing the olfactory systems of conspecifics. Here, we briefly discuss two species in which identification of distinctly different reproductive pheromones has led to an understanding of pheromone function germane to concepts of chemically mediated information networks. Goldfish

The hormonal pheromones of goldfish are currently the best understood of any fish and have recently been reviewed in detail (Sorensen & Stacey, 1999; Kobayashi et al., 2002; Stacey & Sorensen, 2002); therefore, we provide a brief summary before considering aspects that appear directly related to concepts of semiochemical information networks. Goldfish live in mixed-sex, apparently unstructured, groups, undergoing gonadal growth during the winter and spawning a number of times in spring and summer. At ovulation, which occurs

549

P1: IYP/... P2: IYP 0521823617c23.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

Fig. 23.3. Nature and actions of goldfish hormonal pheromones released by periovulatory females (see Stacey & Sorensen (2002) for additional details and original sources). (a) Female periovulatory events. An afternoon surge of pituitary (P) gonadotrophin II (GTH-II) release induces follicular synthesis of 17α,20ß-dihydroxy-4-pregnen-3-one (17,20ß-P), which induces final maturation (completion of arrested meiosis) in mature oocytes. When ovulation occurs approximately 12 hours later, oocytes in the oviduct stimulate synthesis of prostaglandin F2α (PGF2α), which remains at high concentrations in the blood until ovulated oocytes are shed. (b) Preovulatory pheromone. During the GTH-II surge, females release a changing mixture of three steroids: 17,20ß-P and androstenedione (AD), which are released together across the gills, and a sulphated 17,20ß-P metabolite (17,20ß-P-S), which is released in urine pulses. Peak release of AD (which inhibits endocrine response to 17,20ß-P) occurs early in the GTH-II surge, followed by peaks of 17,20ß-P and 17,20ß-P-S release. The preovulatory steroid acts on specific and sensitive (picomolar detection threshold) olfactory receptors, both inducing male behavioural responses and, by the time of ovulation, increasing the quantity and quality of sperm

550

P1: IYP/... P2: IYP 0521823617c23.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

Fish semiochemicals near dawn, groups of males vigorously compete for spawning access as females repeatedly enter aquatic vegetation to oviposit adhesive, undefended eggs over a period of several hours. In such a promiscuous mating system, where male reproductive success likely depends only on the number of eggs fertilized, we believe sperm competition has been a major selective force in the evolution of male reproductive tactics. The cascade of events leading to spawning begins when exogenous factors (increased water temperature and aquatic vegetation) trigger an afternoon surge release of pituitary gonadotrophin II (GTH-II), which stimulates follicular synthesis of the oocyte maturation-inducing steroid 17α,20ß-dihydroxy-4-pregnen3-one (17,20ß-P) (Fig. 23.3a). Ovulation occurs approximately 12 hours later; at which point females become sexually active for the several hours that eggs in the oviduct stimulate synthesis of prostaglandin F2α (PGF2α), a behavioural hormone (Fig. 23.3c). During the approximately 15 hours between the onset of the GTH-II surge and completion of spawning, females sequentially release a preovulatory steroid pheromone (Fig. 23.3b) and a postovulatory prostaglandin pheromone (Fig. 23.3c), which dramatically affect male physiology and behaviour. The preovulatory steroid pheromone (Fig. 23.3b) is a dynamic mixture in which the primary components appear to be 17,20ß-P, its sulphated metabolite (17,20ßP-S) and androstenedione (a testosterone precursor). Although the nature and actions of the preovulatory pheromone are complex (Stacey & Sorensen, 2002), it induces in males both releaser effects on socio-sexual behaviours (e.g. Poling et al., 2001) and a dramatic primer effect: a rapid increase in blood GTH-II that increases both the quantity and quality of releasable stores of milt (sperm and seminal fluids) in the sperm ducts prior to ovulation and spawning (e.g. Zheng et al., 1997). At ovulation, females terminate release of the preovulatory steroid pheromone and begin to release the postovulatory prostaglandin pheromone (PGF2α and its more potent metabolite 15-keto-PGF2α) (Fig. 23.3c). The prostaglandin pheromone not only triggers male courtship and attracts the male to the ovulated female (anosmic males do not spawn) but also activates non-endocrine and endocrine

Fig. 23.3 (cont.) stores in the ducts (inducing GTH-II release, which stimulates testicular 17,20ß-P synthesis). (c) Postovulatory pheromone. Entry of ovulated oocytes to the oviduct stimulates synthesis of PGF2α, which acts in the brain (B) to stimulate female sexual behaviours. PGF2α and its more potent metabolite 15-keto-PGF2α are released in urinary pulses and act on olfactory receptors to trigger male sexual behaviours. Sexual interactions then stimulate movement of sperm to the ducts by two mechanisms: an endocrine mechanism distinct from that mediating testicular response to the preovulatory pheromone; and a rapid and apparently non-endocrine mechanism that begins to increase sperm stores within 15 minutes.

551

P1: IYP/... P2: IYP 0521823617c23.xml

552

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

B. D. Wisenden & N. E. Stacey mechanisms (different from those mediating responses to the preovulatory pheromone) that further increase the volume of releasable milt. In summary, male goldfish first increase their potential fertility through endocrine responses to reliable chemical indicators of imminent ovulation (17,20ß-P and 17,20ß-P-S) and then use reliable indicators that ovulation has occurred (PGF2α and 15-keto-PGF2α) to locate the female and maintain sperm stores. For a number of reasons (Stacey & Sorensen, 2002), most notably a lack of evidence for specialized pheromone production and release, we regard these components of the goldfish hormonal pheromone system as an example of male spying on female chemical cues. Although it is difficult to exclude the possibility that domestication has influenced the goldfish pheromone system, it appears remarkably similar to those of the closely related Crucian carp Carassius carassius and common carp Cyprinus carpio (Irvine & Sorensen, 1993; Stacey et al., 1994; Bjerselius et al., 1995). Furthermore, it is likely that other cyprinids (Family Cyprinidae; > 2000 species) possess similar pheromone systems given that olfactory detection of 17,20ß-P-like steroids and prostaglandins is widespread among this taxon (Stacey & Sorensen, 2002). The effects of goldfish pheromones described above have been studied in the context of dyadic interactions between female originators of hormonal pheromone cues and their male receivers (Fig. 23.3b,c). However, given the proximity of individuals in aggregations, and the size of pheromonal active spaces estimated from release rates and olfactory detection threshold (Sorensen et al., 2000), it is obvious that these ovulatory cues normally operate in an information network, where a female’s preovulatory steroids can potentially be detected by many males and her postovulatory prostaglandins are the proximate trigger promoting sperm competition at spawning. Moreover, the network activated by the preovulatory pheromone evidently includes not only the ovulatory female and her potential spawning partners but also additional females and males not directly exposed to her preovulatory cues (Fig 23.4.). The evidence for female interactions is based on the finding that low concentrations of water-borne 17,20ß-P induced ovulation in goldfish (Kobayashi et al., 2002), suggesting a mechanism for the ovulatory synchrony observed in the field and laboratory. The female benefit(s) of ovulatory synchrony is not known but may involve predator swamping, amplification of preovulatory cues that stimulate male fertility, or (perhaps counter-intuitively) reduction of male to female ratios at spawning (high ratios can result both in ‘forced’ egg release away from suitable spawning substrate and skin damage through abrasion by the male’s breeding tubercles or ‘pearl organs’). Interactions among males appear more complex because they both decrease (Fig. 23.4.a) and increase sperm stores in response to unidentified cues from other males (Stacey et al., 2001; Fraser & Stacey, 2002). For example, males isolated from

f f

m m

m

Temperature + substrate

(b)

F i

1

ii

ii

2

2

?

?

iii

3

553 in turn influence additional males (3 and iii).

0 521 582361 7

indirectly (ii) stimulated males. It is not known if males (2 and ii) that receive indirect information that ovulation is imminent

ovulation in additional females (F), whose preovulatory pheromone release will activate additional networks of directly (i) and

releases unknown cues inducing similar changes in additional males (2). In addition, the preovulatory pheromone can induce

the preovulatory pheromone can directly stimulate endocrine changes in one male (1) that increase his sperm store; he than

water temperature and aquatic vegetation) induce an ovulatory gonadotrophin II surge in one female (shaded grey), release of

remain in a ‘basal’ endocrine state (m) and release unknown cues that suppress sperm stores. (b) If exogenous stimuli (increased

CU1917B/McGregor

Fig. 23.4. Network activation by pheromones. (a) In the presence of mature females that are not ovulatory (f), mature males

(a)

P1: IYP/... P2: IYP 0521823617c23.xml April 12, 2005 12:49

P1: IYP/... P2: IYP 0521823617c23.xml

554

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

B. D. Wisenden & N. E. Stacey a male group dramatically increase sperm stores within 24 hours, indicating they normally suppress their potential for milt production in response to an inhibitory male cue(s). However, if sperm stores of one of a group of males are increased (either by gonadotrophin injection or exposure to 17,20ß-P), untreated males in the group also increase their stores. Therefore, it appears that, in the absence of cues from preovulatory or ovulated females, a mature male goldfish is both originator and receiver of unknown cues that suppress sperm stores in other males by maintaining basal GTH-II and steroids (Fig. 23.4.a). This stable, negative-feedback situation is rapidly and transiently perturbed, however, when exogenous stimuli trigger a preovulatory GTH-II surge in females, resulting in release of the preovulatory steroid pheromone. Males and non-ovulatory females encountering this stimulatory cue in turn increase their GTH-II, amplifying and disseminating the original cue(s) and promoting synchronous final maturation (ovulation and increased sperm stores) of individuals within the network (Fig. 23.4.b). Numerous unresolved questions make it difficult to compare the complex reproductive interactions of goldfish with the classical visual and acoustic communication networks that have been studied in terrestrial species. Perhaps the key issue is whether the pheromonal interactions known among goldfish involve only responses of receivers to unspecialized cues, or whether some are mediated by specialized signals. There is no evidence that female preovulatory and postovulatory pheromones are specialized signals to males or other females (Sorensen & Stacey, 1999; Stacey & Sorensen, 2002). Nor is it obvious how the male’s adaptive endocrine–testicular response to female preovulatory cues (Fig. 23.3b) would also be shaped by selection to include the release of a specialized signal that evidently increases the fertility of his competitors. Indeed, it seems more probable that the indirect responses of males (2 and ii in Fig. 23.4.b) to female preovulatory cues are mediated by spying on unspecialized cues released as by-products of the endocrine responses of males (1 and i in Fig. 23.4.b) directly stimulated by a preovulatory female. In a species where males are territorial and where females mate with several males on their territories, the interactions depicted in Fig. 23.4.b might be expected to have arisen from female tactics to promote sperm competition. However, given that the female goldfish cannot control the number of males competing for fertilization attempts, that virtually all her eggs can be fertilized by a single male (Zheng et al., 1997) and that, as noted above, additional males may disrupt spawning activity, we feel it most probable that the interactions depicted in Fig. 23.4.b result solely from male competition. We hope that our proposal that the hormonal pheromones of goldfish function in spying interactions will stimulate discussion of how this and similar systems can be integrated into current theoretical concepts of information networks based

P1: IYP/... P2: IYP 0521823617c23.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

Fish semiochemicals on true communicative interactions. To open such a discussion, can we speak of eavesdropping in spying networks (e.g. males 2 and ii in Fig 23.4.), given that eavesdropping appears to be restricted to interactions mediated by specialized signals (McGregor & Peake, 2000; Ch. 2) or do we require new terms and/or new definitions? Sea lamprey Petromyzon marinus

Anadromous sea lamprey Petromyzon marinus spend most of their life as stream-dwelling, filter-feeding ammocoete larvae before undergoing a dramatic metamorphosis, migrating to the ocean or large lakes and feeding parasitically on large fish, whose unpredictable movements can carry the lamprey far from their natal streams. After approximately a year, the parasites cease feeding, begin to mature sexually and search for a spawning stream, guided by a potent pheromone that serves as a reliable indicator of suitable larval habitat. Since gaining access to the American Great Lakes from the Atlantic Ocean about a century ago, the sea lamprey has seriously depleted many of these lakes’ fisheries. Based on preliminary evidence (Teeter, 1980) that larvae release a pheromone attracting migrating adults and that spawning adults employ sex pheromones, sea lamprey pheromones have been extensively studied in the hope of identifying semiochemicals for use in biological control, as has successfully been achieved for many insects (Chapman, 2000). Larval pheromone attracting migratory adults

Both field and laboratory studies provide compelling evidence that migratory adult lamprey do not return preferentially to natal streams but instead locate suitable spawning habitat by responding to a pheromone released by streamdwelling larvae. Historical capture records show that estimated numbers of migrating adults fall by up to 50% following application of larvicides to remove larval populations (reviewed by Sorensen & Vrieze, 2003). Evidence that such reductions in migrant numbers result from removal of larval odour comes from studies of captive migrants in large two-choice mazes (Vrieze & Sorensen, 2001). Water from streams without larvae is much less attractive to adults than is water from larva-bearing streams, but water from streams without larvae becomes attractive following addition of low concentrations of larval odour. The potency of larval odour is such that a single larva (weighing only several grams) creates an active space of 400–4000 l/h, sufficient to account for the attractive properties of streams with larvae. These studies also reveal that spawning-stream selection is based on more than larval odour alone: migratory adults prefer stream water (even

555

P1: IYP/... P2: IYP 0521823617c23.xml

556

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

B. D. Wisenden & N. E. Stacey without larval odour) to lake water, suggesting the presence of unknown stream odorants that act synergistically with larval odour (Vrieze & Sorensen, 2001). The larval pheromone attracting adult migrants has been fully characterized and shown to be a mixture of chemicals (Vrieze & Sorensen, 2001; P. W. Sorensen, personal communication); two of the primary components are the novel bile acids, allocholic acid (ACA; 3α,7α,12α-trihydroxy-5α-cholan-24-oic acid) and petromyzonol sulphate (PS; 3α,12α,24-trihydroxy-5α-cholan-24-sulphate). PS may be a unique lamprey product and is synthesized by the liver of larvae but not by the parasitic or adult phases (Polkinghorne et al., 2001). Because larvae undergo gall bladder and bile duct atrophy at metamorphosis and also cease synthesis of PS and ACA (Polkinghorne et al., 2001), these compounds should be specific indicators of streams containing favourable spawning and nursery habitat. PS and ACA, which are released primarily in larval faeces (Polkinghorne et al., 2001), are detected by the olfactory organ of migratory adults (Li & Sorensen, 1997) not only with great specificity, but also with a sensitivity (1 pmol/l olfactory detection threshold) that would account for behavioural responsiveness at the low concentrations estimated to occur in spawning streams (Polkinghorne et al., 2001). Furthermore, these bile acids attract migratory adults (but not parasites) in maze tests (Bjerselius et al., 2000; Vrieze & Sorensen, 2001). Taken together, the results indicate that a suite of conspecific cues regulate stream selection and upstream migration of maturing adult lamprey, and that response to larval odour is adaptive in so far as it increases the likelihood of locating habitat suitable for larval growth. Moreover, because there is no evidence at this time that larval production and release of PS and ACA are specialized for functions other than digestion (Polkinghorne et al., 2001), and no evident mechanism whereby adult response could select for specialized signalling functions for these compounds, we regard these components of the migratory pheromone as cues involved in chemical spying. Unlike the transient pheromonal steroid and prostaglandin cues of goldfish, which are released only at specific stages of reproduction, however, the bile acid cues of lamprey appear to be released not only during the period of peak adult migration in May but throughout the extended period (April–August) of larval feeding (Sutton & Bowen, 1994; Polkinghorne et al., 2001). In addition, whereas the transient pheromonal cues of female goldfish are estimated to generate only small active spaces (Sorensen et al., 2000), PS and ACA released by larval lamprey are estimated to create very large active spaces sufficient to serve effectively as an upstream attractant given that larval populations can contain hundreds of thousands of individuals (Polkinghorne et al., 2001). Perhaps the greatest departure from the goldfish situation, however, is that, whereas goldfish pheromonal cues promote interactions of originators and receivers within a small social unit, the lamprey larval pheromone functions in

P1: IYP/... P2: IYP 0521823617c23.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

Fish semiochemicals a vast network of dispersed originators and receivers that do not interact behaviourally. Sex pheromones

During upstream migration, adult male and female lamprey undergo final maturation (spermiation and ovulation), lose behavioural responsiveness to the larval pheromone and develop behavioural responsiveness to the odour of mature conspecifics of the opposite sex (Bjerselius et al., 2000; Li et al., 2002). Although the described behavioural responses of mature adults (positive rheotaxis, increased locomotory behaviours) are rather non-specific, they are appropriate to mediate upstream movement to spawning grounds and facilitate male–female interactions, although this has not been demonstrated experimentally. However, the traditional use of mature males to trap females (Fontaine, 1938; discussed in Teeter, 1980) supports the existence of a potent male attractant, which is the only lamprey sex pheromone to be studied intensively. This pheromone, estimated to have a large active space (> 106 l/h per adult male (Li et al., 2002)), is proposed to function in attracting females to mature males, which are reported to precede females to the spawning grounds. Major components of the pheromone released by spermiated male lamprey are proposed to be 3-keto-petromyzonol-sulphate (3-keto-PS; 7α,12α,24trihydroxy-3-one-5α-cholan-24-sulphate) and 3-keto-allocholic acid (3-keto-ACA; 7α,12α-dihydroxy-5α-cholan-3-one-24-oic acid) (Li et al., 2002; Yun et al., 2003). Although both these compounds are detected by the lamprey olfactory system, only 3-keto-PS has been investigated for pheromonal activity. As with the odour of spermiated males, 3-keto-PS when added to a two-choice maze both attracts ovulated females (but not preovulatory females or males) and stimulates their searching behaviours (Li et al., 2002). Moreover, whereas non-spermiated males (whose odour does not attract ovulated females in the maze) do not release appreciable quantities of 3-keto-PS, spermiated males release large quantities of 3-keto-PS (approximately 500 g/h) (Li et al., 2002; Yun et al., 2002). As with the bile acid pheromone of larval lamprey (Polkinghorne et al., 2001), 3-keto-PS has been found in the liver of spermiated males (Li et al., 2002). However, unlike the larval pheromone, which is released primarily in faeces, the pheromone from spermiated males appears to be released by the gills, which in mature males (but not females) develop glandular cells (Pickering, 1977) that evidently are specialized for pheromone release (Siefkes et al., 2003). The current information on male lamprey pheromone suggests its synthesis occurs through a subtle shift in bile acid metabolism that results in the larval pattern of PS and ACA production changing to 3-keto-PS and 3-keto-ACA in spermiating males (presence of the 3-keto acids in livers of ovulating females appears not

557

P1: IYP/... P2: IYP 0521823617c23.xml

558

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

B. D. Wisenden & N. E. Stacey to have been examined). Furthermore, because fully mature adults are exposed to larval and adult bile acids in spawning streams, it is expected that the lamprey olfactory system has been selected to discriminate larval (3-hydroxy acid) and adult (3-keto acid) odours, although this remains to be examined. The identified sex pheromones of lamprey and goldfish are similar in that they operate within a complex network of originators/signallers and receivers, although they differ fundamentally both in the interactions between genders and in the ancestral (prepheromonal) functions of the cues and signals. Moreover, the evidence for signal specialization in production and release of male lamprey pheromone suggests a true communicatory interaction, which is unlikely in goldfish.

Synthesis Current theory about the function of animal communication networks (e.g. McGregor & Peake, 2000) has been heavily influenced by studies of acoustic and visual systems, where it seems clear that true communication between specialized signallers and receivers has arisen through the bilateral benefits resulting from their reciprocal interactions. Although studies of fish semiochemicals also provide evidence of specializations indicative of communication, the specific functions of such specialized semiochemicals within networks are not well understood. In sea lamprey, for example, both the large active space of the proposed male sex pheromone 3-keto-PS and apparent male-specific gill structure facilitating its release (Li et al., 2002) suggest specializations for increased amplitude of a specialized tonic signal. The proposed function of this male lamprey signal appears analogous to the aggregate signal produced by chorusing male anurans (Ch. 13), in so far as the combined odour of many males induces the upstream movement of many females. However, it remains to be determined if attracted female lamprey also use the male pheromone in mate choice and if this might have been the pheromone’s original function. Also, in the black goby, non-spermatogenic portions of the testes appear specialized for synthesis of a steroid pheromone, etiocholanolone glucuronide, originally proposed simply to attract ovulated females to the male’s nest (Colombo et al., 1980). In the round goby Neogobius melanostomus, however, both males and females respond behaviourally to etiocholanolone glucuronide (Murphy et al., 2001), suggesting that the pheromone functions in a more complex network involving both intra- and intersexual communication. Given that semiochemical communication appears to have evolved in sea lamprey and gobies, and perhaps in some other fish such as blennies (Laumen et al., 1974; Gonc¸alves et al., 2002) and African catfish Clarias gariepinus (van den Hurk &

P1: IYP/... P2: IYP 0521823617c23.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

Fish semiochemicals Resink, 1992), these species may communicate in semiochemical networks analogous to those seen in terrestrial systems involving acoustic and visual signals. However, other fish semiochemicals, such as the alarm cues of ostariophysans and the sex pheromones of goldfish, appear to function not in communication but rather in spying, where specialization for information transfer evidently is restricted to receivers. Nonetheless, these semiochemical cues also operate in complex information networks in which semiochemicals can influence several conspecifics both directly (through exposure) and indirectly (through changes induced in exposed individuals) (e.g. Figs. 23.2 and 23.4.b) Because such fish semiochemical networks based on unspecialized cues have the potential to give rise to true communication networks, they should not only extend the scope of current network theory but also raise important issues relevant to the evolutionary processes by which such communicatory networks evolve. To cite just one example, when discussion of information networks is restricted to those that involve communication, it might seem reasonable to assume that eavesdropping arises only after communicative interaction has been established. However, the ability of male goldfish to derive information indirectly about female cues by spying on the responses of exposed males (e.g. Figs. 23.2 and 23.4.b) demonstrates that a process analogous (and possibly homologous) to eavesdropping can precede the origin of communication. To promote discussion of the functional and evolutionary relationships among spying, eavesdropping and communication, we propose two hypothetical schemes. One is based on the intraspecific interactions induced by the goldfish preovulatory steroid pheromone (Fig. 23.5a); the second involves both intra- and interspecific predator–prey interactions in ostariophysan fishes (Fig. 23.5b), and both are derived from our general model for the evolution of communication (Fig. 23.1). In goldfish, spying by male receivers (R) on an unspecialized steroid cue released by female originators (O; Fig. 23.5a1) could lead to communication (Fig. 23.5a2) if male response to heritable variation in cue production leads to differential female fitness. If this occurs, females would then be signallers (S) releasing a specialized pheromonal signal and the male’s role would change (R1 ), as he now influences, and is influenced by, signal evolution. As we emphasize in this chapter, however, the goldfish preovulatory pheromone mediates more than the simple dyadic spying event depicted in Fig. 23.5a1. The pheromone directly stimulates behavioural and endocrine–testicular responses in more than one male (R) and also induces a distinct response (ovulation) in females (Fig. 23.5a3). In addition, the pheromone indirectly stimulates males (R2 ) via cues released by pheromoneexposed males (Fig. 23.5a4). In the ancestral condition of predator-induced prey chemical alarm cues, predator (P) attack releases general cues from the originator (O) that can be received both

559

P1: IYP/... P2: IYP 0521823617c23.xml

560

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

B. D. Wisenden & N. E. Stacey

(a) 3

R

4

R2

1

O

R

spying communication

O

R

S

R1

O O

R

2

IE

6

5

S

S

R1

(b)

3

+

R1 4

1

O C

O

SE

R

R

O

R

P

O

P2

P

R r P2

P

spying communication

+

S P

5

O C R1

2

6

IE

S P

P3

IE

SE

S

P3

P

Fig. 23.5. Theoretical evolutionary pathways of the transition between spying and communication networks involving semiochemicals used in reproductive (a) and predator–prey (b) interactions. Thin solid and dashed arrows indicate spying functions; thick, opposed, black and white arrows indicate communicative functions, and large white arrows indicate transitions between proposed stable states. O, originator; S, signaller; R, receiver (r, heterospecific receiver); IE, interceptive eavesdropper; SE, social eavesdropper; P, predator; C, alarm cue. See text for further explanation.

as an alarm cue by conspecific prey (R) and as a feeding cue by secondary predators (P2 ; Fig. 23.5b1). If interference by secondary predators benefits originators and leads to alarm cue specialization, originators become signallers (S), the secondary predator’s role changes (P3 ), and receiving conspecific prey become interceptive eavesdroppers (IE) in a communication network (Fig. 23.5b2). As with the goldfish pheromone (Fig. 23.5a3,4), predator-induced alarm cues can exert complex effects prior to the evolution of communication. For example, alarm cues are used to

P1: IYP/... P2: IYP 0521823617c23.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

Fish semiochemicals associate risk with stimuli (C) correlated with predation, which later serve as indicators of predation risk (Fig. 23.5b3). This latter system may become elevated to that of a communication network without involvement of a secondary predator if an originator’s shoalmates learn to recognize a novel indicator of risk and later alert the surviving originator to the presence of risk through early response to danger (Fig. 23.5b5). In direct relevance to the evolution of eavesdropping, alarm cues can also affect predator–prey interactions indirectly through social facilitation (social spying?) of alarm behaviour both in conspecifics (R2 ) and in heterospecifics (r; Fig. 23.5b4). If it is reasonable to assume that sex and alarm pheromone communication evolves from spying, as depicted in Figs. 23.1, 23.5a1,2 and 23.5b1,2, then it also seems reasonable to ask whether and how communication networks evolve from spying networks. We, therefore, propose two general scenarios, which differ primarily in the evolutionary origins of eavesdropping. In the first scenario, a simple dyadic communication (Figs. 23.5a2 and 23.5b2) could lead to the evolution of interceptive or social eavesdropping (Ch. 2) if receivers evolve adaptive responses either to the signalling behaviour per se (interceptive eavesdropper (IE): Figs. 23.5a5 and 23.5b6) or to the signalling interaction (social eavesdropper (SE): Figs. 23.5a6 and 23.5b6). In this scenario, where the evolution of communication precedes that of eavesdropping, eavesdropper functions (interceptive and social) are analogous to the various receiver functions in spying networks (Figs. 23.5a3,4 and 23.5b3,4). In the second scenario, incipient eavesdropping arises in spying networks, either as direct (Figs. 23.5a3 and 23.5b4) or indirect (Figs. 23.5a4 and 23.5b4) spying by receivers on originators and is retained as interceptive and social eavesdropping, respectively, following the evolution of communication. In this scenario, receivers in spying networks are homologous to eavesdroppers in communication networks. In all the scenarios shown in Fig. 23.5, we depict eavesdropping in its proposed initial state: that is, spying via a cue that is not specialized for transmission to eavesdroppers, despite being a signal specialized for information transfer to the primary target (Ch. 2). At this early stage, the network functions of eavesdropper and primary target differ in kind. However, if subsequent selection by eavesdroppers leads to signal specialization specific to the eavesdropping interaction, and thus forming a communicative relationship between eavesdropper and signaller, functions of eavesdroppers and receivers will come to differ only in degree. Studied examples of eavesdropping (Ch. 2) typically appear to involve costs or benefits to signallers that would be expected to modify signal function; consequently, it will be important to determine whether, as has been suggested for sex pheromone function in fish (Fig. 23.5a1), various forms of eavesdropping in communication networks (Figs. 23.5a5,6 and 23.5b2,5,6) can persist as spying. Moreover, it will be important to document covariance in the relative proportions of spying versus

561

P1: IYP/... P2: IYP 0521823617c23.xml

562

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

B. D. Wisenden & N. E. Stacey communicative eavesdropping and the ecological and social factors that lead to the spying–communication transition.

Acknowledgements The authors gratefully acknowledge support from MSUM College of Social and Natural Sciences, MSUM Dille Fund for Excellence, MSUM Alumni Foundation, MnSCU Learning by Doing (B. Wisenden) and the Natural Sciences and Engineering Research Council of Canada (N. Stacey).

References Argentini, M. 1976. Isoleirung des Schreckstoffes aus der Haut der Erlitz Phoxinus phoxinus (L.). Ph.D. Thesis, University of Zurich. Beauchamp, G. K. 2000. Defining pheromones. In: The Monell Connection, Fall 2000, ed. L. J. Stein. Philadelphia, PA: Monell Chemical Senses Center, p. 2. Bjerselius, R., Olsen, K. H. & Zheng, W. 1995. Endocrine, gonadal and behavioral responses of male crucian carp (Carassius carassius) to the hormonal pheromone 17α,20β-dihydroxy-4-pregnen-3-one. Chemical Senses, 20, 221–230. Bjerselius, R., Li, W., Teeter, J. H. et al. 2000. Direct behavioral evidence that unique bile acids released by larval sea lamprey (Petromyzon marinus) function as a migratory pheromone. Canadian Journal of Fisheries and Aquatic Sciences, 57, 557–569. Brown, G. E., Godin, J.-G. J. & Pedersen, J. 1999. Fin-flicking behaviour: a visual anti-predator alarm signal in a characin fish, Hemigrammus erythrozonus. Animal Behaviour, 58, 469–475. Brown, G. E., Adrian, J. C., Jr & Shih, M. L. 2001. Behavioural responses of fathead minnows to hypoxanthine-3N-oxide at varying concentrations. Journal of Fish Biology, 58, 1465–1470. Bryant, P. B. 1987. A study of the alarm system in selected fishes of Northern Mississippi. Ph.D. Thesis, University of Mississippi. Chapman, R. F. 2000. Entomology in the twentieth century. Annual Reviews of Entomology, 45, 261–285. Chivers, D. P. & Mirza, R. S. 2001. Predator diet cues and the assessment of predation risk by aquatic vertebrates: a review and prospectus. In Chemical Signals in Vertebrates, ed. A. Marchlewska-Koj, J. Lepri & D. M¨ uller-Schwarze. New York: Plenum Press, pp. 277–284. Chivers, D. P. & Smith, R. J. F. 1994. The role of experience and chemical alarm signalling in predator recognition by fathead minnows, Pimephales promelas Rafinesque. Journal of Fish Biology, 44, 273–285. 1998. Chemical alarm signalling in aquatic predator–prey systems: a review and prospectus. Ecoscience, 5, 338–352. Chivers, D. P., Brown, G. E. & Smith, R. J. F. 1996. The evolution of chemical alarm signals: attracting predators benefits alarm signal senders. American Naturalist, 148, 649–659.

P1: IYP/... P2: IYP 0521823617c23.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

Fish semiochemicals Colombo, L., Marconato, A., Belvedere, P. C. & Frisco, C. 1980. Endocrinology of teleost reproduction: a testicular steroid pheromone in the black goby, Gobius jozo L. Bolletino di Zoologia, 47, 355–364. Fontaine, M. 1938. La lamproie marine. Sa peche et son importance economique. Bulletin de la Soci´et´e Oc´eanographique de France, 17, 1681–1687. Fraser, E. J. & Stacey, N. 2002. Isolation increases milt production in goldfish. Journal of Experimental Zoology, 293, 511–524. Gandolfi, G., Classon, L. J. & Rossi, A. C. 1968. The fright reaction of zebra fish. Atti Societa Italiana di Scienze Naturali, 107, 74–88. Gazdewich, K. J. & Chivers, D. P. 2002. Acquired recognition by fathead minnows: influence of habitat characteristics on survival. Journal of Chemical Ecology, 28, 439–445. Gonc¸alves, D. M., Barata, E. N., Oliveira, R. F. & C´ anario, A. V. M. 2002. The role of male visual and chemical cues on the activation of female courtship behaviour in the sex-role reversed peacock blenny. Journal of Fish Biology, 61, 90–105. Hall, D. & Suboski, M. D. 1995. Visual and olfactory stimuli in learned release of alarm reactions by zebra danio fish (Brachydanio rerio). Neurobiology of Learning and Memory, 63, 229–240. Hamdani, E. H., Stabell, O. B., Alexander, G. & Døving, K. B. 2000. Alarm reaction in the crucian carp is mediated by the medial bundle of the medial olfactory tract. Chemical Senses, 25, 103–109. Hamdani, E. H., Kasumyan, A. & Døving, K. B. 2001. Is feeding behaviour in crucian carp mediated by the lateral olfactory tract? Chemical Senses, 26, 1133–1138. Hansen, J. A., Rose, J. D., Jenkins, R. A., Gerow, K. G. & Bergman, H. L. 1999. Chinook salmon (Oncorhynchus tshawytscha) and rainbow trout (Oncorhynchus mykiss) exposed to copper: neurophysiological and histological effects on the olfactory system. Environmental Toxicology and Chemistry, 18, 1979–1991. Hasson, O. 2000. Knowledge, information, biases and signal assemblages. In: Animal Signals: Signalling and Signal Design in Animal Communication, ed. Y. Espmark, T. Amundsen & G. Rosenqvist. Trondheim: Tapir Academic Press, pp. 119–139. Hauser, M. 1996. The Evolution of Communication. Cambridge, MA: MIT Press. Henderson, P. A., Irving, P. W. & Magurran, A. E. 1997. Fish pheromones and evolutionary enigmas: a reply to Smith. Proceedings of the Royal Society of London, Series B, 264, 451–453. Hews, D. K. 1988. Alarm response in larval western toads, Bufo boreas: release of larval chemicals by a natural predator and its effect on predator capture efficiency. Animal Behaviour, 36, 125–133. Hubbard, P. C., Barata, E. N. & Can´ ario, A. V. M. 2002. Possible disruption of pheromonal communication by humic acid in the goldfish, Carassius auratus. Aquatic Toxicology, 60, 169–183. Irvine, I. A. S. & Sorensen, P. W. 1993. Acute olfactory sensitivity of wild common carp, Cyprinus carpio, to goldfish sex pheromones is influenced by gonadal maturity. Canadian Journal of Zoology, 71, 2199–2210. Karlson, P. & L¨ uscher, M. 1959. ‘Pheromones’: a new term for a class of biologically active substances. Nature, 183, 55–56.

563

P1: IYP/... P2: IYP 0521823617c23.xml

564

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

B. D. Wisenden & N. E. Stacey Kasumyan, A. O. & Ponomarev V. Y. 1987. Biochemical features of alarm pheromone in fish of the order Cypriniformes. Journal of Evolutionary Biochemistry and Physiology, 23, 20–24. Kats, L. B. & Dill, L. M. 1998. The scent of death: chemosensory assessment of predation risk by prey animals. Ecoscience, 5, 361–394. Kobayashi, M., Sorensen, P. W. & Stacey, N. E. 2002. Hormonal and pheromonal control of spawning behaviour in the goldfish. Fish Physiology and Biochemistry, 26, 71–84. Korpi, N. L. & Wisenden, B. D. 2001. Learned recognition of novel predator odour by zebra danios, Danio rerio, following time-shifted presentation of alarm cue and predator odour. Environmental Biology of Fishes, 61, 205–211. Kristensen, E. A. & Closs, G. P. 2004. Anti-predator response of naive and experienced common bully to chemical alarm cues. Journal of Fish Biology, 64. 643–652. Laumen, T. J., Pern, U. & Blum, V. 1974. Investigations on the function and hormonal regulation of the anal appendices in Blennius pavo (Risso). Journal of Experimental Zoology, 190, 47–56. Lawrence, B. J. & Smith, R. J. F. 1989. Behavioral response of solitary fathead minnows, Pimephales promelas, to alarm substance. Journal of Chemical Ecology, 15, 209–219. Li, W. & Sorensen, P. W. 1997. Highly independent olfactory receptor sites for conspecific bile acids in the sea lamprey, Petromyzon marinus, Journal of Comparative Physiology A, 180, 429–438. Li, W., Scott, A. P., Siefkes, M. J. et al. 2002. Bile acid secreted by male sea lamprey that acts as a sex pheromone. Science, 296, 138–141. Liley, N. R. 1982. Chemical communication in fish. Canadian Journal of Fisheries and Aquatic Sciences, 39, 22–35. Lloyd, J. E. 1965. Aggressive mimicry in Photuris: firefly femme fatales. Science, 149, 653–654. Magurran, A. E. 1989. Acquired recognition of predator odour in the European minnow (Phoxinus phoxinus). Ethology, 82, 216–223. Magurran, A. E., Irving, P. W & Henderson, P. A. 1996. Is there a fish alarm pheromone? A wild study and critique. Proceedings of the Royal Society of London, Series B, 263, 1551–1556. Mathis, A. & Smith, R. J. F. 1993. Chemical alarm signals increase the survival time of fathead minnows (Pimephales promelas) during encounters with northern pike (Esox lucius). Behavioral Ecology, 4, 260–265. Mathis, A., Chivers, D. P. & Smith, R. J. F. 1995. Chemical alarm signals: predator deterrents or attractants? American Naturalist, 146, 994–1005. 1996. Cultural transmission of predator recognition in fishes: intraspecific and interspecific learning. Animal Behaviour, 51, 185–201. McClintock, M. 2002. Pheromones, odors, and vasanas: the neuroendocrinology of social chemosignals in humans and animals. In: Hormones, Brain, and Behavior, Vol. 2, ed. D. Pfaff, D. Arnold, A. Etgen, S. Fahrbach & R. Rubin. New York: Academic Press, pp. 797–870. McGregor, P. K. & Peake, T. M. 2000. Communication networks: social environments for receiving and signalling behaviour. Acta Ethologica, 2, 71–81.

P1: IYP/... P2: IYP 0521823617c23.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

Fish semiochemicals Møller, A. P. 1989. Deceptive use of alarm calls by male swallows, Hirundo rustica: a new paternity guard. Behavioral Ecology, 1, 1–6. Murphy, C. A., Stacey, N. E. & Corkum, L. D. 2001. Putative steroidal pheromones in the round goby, Neogobius melanostomus: olfactory and behavioural responses. Journal of Chemical Ecology, 27, 443–470. Nelson, J. S. 1994. Fishes of the World, 4th edn. New York: Wiley-Interscience. Oliveira, R. F., McGregor, P. K. & Latruffe, C. 1998. Know thine enemy: fighting fish gather information from observing conspecific interactions. Proceedings of the Royal Society of London, Series B, 254, 1045–1049. Otter, K. A., McGregor, P. K., Terry, A. M. R. et al. 1999. Do female great tits (Parus major) assess males by eavesdropping? A field study using interactive song playback. Proceedings of the Royal Society of London, Series B, 266, 1305–1310. Paxton, R. J. & Tengo, J. 2001. Double duped males: the sweet and sour of the orchid’s bouquet. Trends in Ecology and Evolution, 16, 167–169. Pfeiffer, W. 1977. The distribution of fright reaction and alarm substance cells in fishes. Copeia, 1977, 653–665. Pfeiffer, W., Riegelbauer, G., Meir, G. & Scheibler, B. 1985. Effect of hypoxanthine-3(N)-oxide and hypoxanthine-1(N)-oxide on central nervous excitation of the black tetra Gymnocorymbus ternetzi (Characidae, Ostariophysi, Pisces) indicated by dorsal light response. Journal of Chemical Ecology, 11, 507–524. Pickering, A. D. 1977. Sexual dimorphism in the gills of the spawning river lamprey, Lampetra fluviatilis L. Cell and Tissue Research, 180, 1–10. Poling, K. R., Fraser, E. J. & Sorensen, P. W. 2001. The three steroidal components of the goldfish preovulatory pheromone signal evoke different behaviors in males. Comparative Biochemistry and Physiology B, 129, 645–651. Polkinghorne, C. N., Olson, J. N., Gallaher, D. G. & Sorensen, P. W. 2001. Larval sea lamprey release two unique bile acids to the water at a rate sufficient to produce detectable riverine pheromone plumes. Fish Physiology and Biochemistry, 24, 15–30. Schutz, F. 1956. Vergleichende Untersuchungen uber die Schreckreacktion bei Fischen und Verbreitung. Zeitschrift f¨ ur Vergleichende Physiologie, 38, 84–135. Siefkes, M. J., Scott, A. P., Zielinski, B. et al. 2003. Male sea lampreys Petromyzon marinus, excrete a sex pheromone from gill epithelia. Biology of Reproduction, 69, 125–132. Smith, R. J. F. 1979. Alarm reaction of Iowa and johnny darters (Etheostoma, Percidae) to chemicals from injured conspecifics. Canadian Journal of Zoology, 57, 1278–1282. 1982. Reaction of Percina nigrofasciata, Ammocrypta beani, and Etheostoma swaini (Percidae, Pisces) to conspecific and intergeneric skin extracts. Canadian Journal of Zoology, 17, 2253–2259. 1992. Alarm signals in fishes. Reviews in Fish Biology and Fisheries, 2, 33–63. 1997. Does one result trump all others? A response to Magurran, Irving and Henderson. Proceedings of the Royal Society of London, Series B, 264, 445–450. 1999. What good is smelly stuff in the skin? Cross function and cross taxa effects in fish ‘alarm substances’. In: Advances in Chemical Signals in Vertebrates, ed. R. E. Johnston, D. M¨ uller-Schwarze & P. W. Sorensen. New York: Kluwer Academic/Plenum Press, pp. 66–84.

565

P1: IYP/... P2: IYP 0521823617c23.xml

566

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

B. D. Wisenden & N. E. Stacey Smith, R. J. F. & Lemly, A. D. 1986. Survival of fathead minnows after injury by predators: the possible role in the evolution of alarm signals. Environmental Biology of Fishes, 15, 147–149. Sorensen, P. W. & Stacey, N. E. 1999. Evolution and specialization of fish hormonal pheromones. In: Advances in Chemical Signals in Vertebrates, ed. R. E. Johnston, D. Muller-Schwarze & P. W. Sorensen. New York: Kluwer Academic/Plenum Press, pp. 15–47. Sorensen, P. W. & Vrieze, L. A. 2003. The chemical ecology and potential application of the sea lamprey migratory pheromone. Journal of Great Lakes Research, 29 (Suppl. 1), 66–84. Sorensen, P. W., Christensen, T. A. & Stacey, N. E. 1998. Discrimination of pheromonal cues in fish: emerging parallels with insects. Current Opinion in Neurobiology, 8, 458–467. Sorensen, P. W., Scott, A. P. & Kihslinger, R. L. 2000. How common hormonal metabolites function as relatively specific pheromonal signals in goldfish. In: Proceedings of the Sixth International Symposium on the Reproductive Physiology of Fish, ed. B. Norberg, O. S. Kjesbu, G. L. Taranger, E. Andersson & S. O. Stefansson. Bergen: John Grieg, pp. 125–128. Stacey, N. E. & Sorensen, P. W. 2002. Hormonal pheromones in fish. In: Hormones, Brain, and Behavior, Vol. 2, ed. D. Pfaff, D. Arnold, A. Etgen, S. Fahrbach & R. Rubin. New York: Academic Press, pp. 375–435. Stacey, N. E., Kyle, A. L. & Liley. N. R. 1986. Fish reproductive pheromones. In: Chemical Signals in Vertebrates, Vol. 4, ed. D. Duvall, D. Muller-Schwarze & R. M. Silverstein. New York: Plenum Press, pp. 117–133. Stacey, N. E., Zheng, W. B. & Cardwell, J. R. 1994. Milt production in common carp (Cyprinus carpio): stimulation by a goldfish steroid pheromone. Aquaculture, 127, 265–276. Stacey, N. E., Fraser, E. J., Sorensen, P. W. & van der Kraak, G. J. 2001. Milt production in goldfish: regulation by multiple social stimuli. Comparative Biochemistry and Physiology C, 130, 467–476. Suboski, M. D. 1990. Releaser-induced recognition learning. Psychology Review, 97, 271–284. Suboski, M. D., Brian, S., Carty, A. E. et al. 1990. Alarm reaction in acquisition and social transmission of simulated predator recognition by zebra danio fish (Brachydanio rerio). Journal of Comparative Physiology, 104, 101–112. Sutton, T. M. & Bowen, S. H. 1994. Significance of organic detritus in the diet of larval lampreys in the Great Lakes basin. Canadian Journal of Fisheries and Aquatic Sciences, 51, 2380–2387. Teeter, J. 1980. Pheromone communication in sea lampreys (Petromyzon marinus): implications for population management. Canadian Journal of Fisheries and Aquatic Sciences, 37, 2123–2132. van den Hurk, R. & Resink, J. W. 1992. Male reproductive system as sex pheromone producer in teleost fish. Journal of Experimental Zoology, 261, 204–213. Verheijen, F. J. 1956 Transmission of fright reaction amongst a school of fish and the underlying sensory mechanisms. Experientia, 12, 202–204.

P1: IYP/... P2: IYP 0521823617c23.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

12:49

Fish semiochemicals Vickers, N. J. 2000. Mechanisms of animal navigation in odor plumes. Biological Bulletin 198, 203–212. ¨ ber einen Schreckstoff der Fischaut and seine biologische von Frisch, K. 1941. U Bedeutung. Zeitschrift F¨ ur Vergleichende Physiologie, 29, 46–145. Vrieze, L. A. & Sorensen, P. W. 2001. Laboratory assessment of the role of a larval pheromone and natural stream odor in spawning stream localization by migratory sea lamprey (Petromyzon marinus). Canadian Journal of Fisheries and Aquatic Sciences, 58, 2374–2385. Weissburg, M. J. 2000. The fluid dynamical context of chemosensory behavior. Biological Bulletin, 198, 188–202. Williams, G. C. 1992. Natural Selection: Domains, Levels and Challenges. Oxford: Oxford University Press. Wilson, E. O. & Bossert, W. H. 1963. Chemical communication among animals. In: Recent Progress in Hormone Research, Vol. 19, ed. G. Pincus. New York: Academic Press, pp. 673–716. Wisenden, B. D. 2000. Scents of danger: the evolution of olfactory ornamentation in chemically mediated predator–prey interactions. In: Animal Signals: Signalling and Signal Design in Animal Communication, ed. Y. Espmark, T. Amundsen & G. Rosenqvist. Trondheim: Tapir Academic Press, pp. 365–386. 2003. Chemically mediated strategies to counter predation. In: Sensory Assessment of the Aquatic Environment, ed. S. P. Collin & N. J. Marshall. New York: Springer, pp. 236–251. Wisenden, B. D. & Thiel, T. A. 2002. Field verification of predator attraction to minnow alarm substance. Journal of Chemical Ecology, 28, 433–438. Wisenden, B. D., Chivers, D. P., Brown, G. E. & Smith, R. J. F. 1995. The role of experience in risk assessment: avoidance of areas chemically labelled with fathead minnow alarm pheromone by conspecifics and heterospecifics. Ecoscience, 2, 115–122. Wisenden, B. D., Cline, A. & Sparkes, T. C. 1999. Survival benefit to anti-predator behavior in the amphipod Gammarus minus in response to injury-released chemical cues from conspecifics and heterospecifics. Ethology, 105, 407–414. Yun, S.-S., Scott, A. P., Siefkes, M. & Li, W. 2002. Development and application of an ELISA for a sex pheromone released by the male sea lamprey (Petromyzon marinus L.). General and Comparative Endocrinology, 129, 163–170. Yun, S.-S., Scott, A. P. & Li, W. 2003. Pheromones of the male sea lamprey, Petromyzon marinus L: structural studies on a new compound, 3-keto allocholic acid, and 3-keto petromyzonol sulfate. Steroids, 68, 297–304. Zheng, W., Strobeck, C. & Stacey, N. E. 1997. The steroid pheromone 17α,20β-dihydroxy-4-pregnen-3-one increases fertility and paternity in goldfish. Journal of Experimental Biology, 200, 2833–2840.

567

P1: IYP/... P2: IYP 0521823617c24.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

13:31

24

Cognitive aspects of networks and avian capacities irene m. pepperberg Brandeis University, Waltham, USA

Introduction The natural world is an extremely complicated place of myriad interactions – some obvious, some hidden – but all of critical importance if one is to understand its workings. Information must be processed, sorted, ignored or acted upon by all creatures, even though the levels of processing ability vary across species. Scientists, although well aware of these complexities and eager to make sense of them, often begin by reducing interactions to their simplest form, under the assumption that one can gain an understanding of more complex issues by first gaining full knowledge of the simplest. Consequently, in most scientific endeavours, initial studies examine the effect of a single stimulus on an entity: in physics, how light waves interact with a single atom, or how two atoms might interact; in child psychology, the reaction of an infant to a caretaker’s smile or to a novel toy; in animal behaviour laboratories, the effect of a shock on the behaviour of a rat’s movement in a simple laboratory maze or the effect of a tape loop of song on a bird in a sound isolation box. In each instance, however, the data obtained provide only a small glimmer of the complexity that exists in the real world, and in many cases inferences drawn from data in such experiments truly explain only the specific laboratory situation being studied. To expand to a larger system and a broader base often requires – and triggers – the development of more sophisticated tools, be they mathematical theories (e.g. the Nash equilibrium), more powerful computers for handling data or more sophisticated equipment for gathering data (e.g. complex recording arrays). Sometimes, however, what is first needed is simply the capacity to think outside of traditional forms of experimentation.

Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

568

P1: IYP/... P2: IYP 0521823617c24.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

13:31

Cognitive aspects of networks and avian capacities The study of animal (and particularly avian) cognition was one such paradigm shift, and the consequent realization that animals needed to – and could – process several sets of information (e.g. Pepperberg, 1990) was a logical outcome of such a shift. The story begins with the so-called cognitive revolution (Hulse et al., 1968), when researchers began to accept that levels and types of intelligence in nonhumans formed a continuum with those of humans, and to investigate a wide range of behaviour and its development through many techniques in various species. Most resulting studies, however, simply adapted material from human cognitive experiments, often focused on a single set of tasks (i.e. not forcing subjects to choose an appropriate set of responses from several possibilities, as they must do in nature), continued to use a small number of species (predominantly monkeys, rats, and pigeons) and made fairly sweeping conclusions as to the relative intelligence of all animals on the basis of these data (review in Pepperberg, 2001). Researchers, mostly in the laboratory but even in the field, initially failed to examine species’ innate predispositions, evolutionary histories or ecological constraints and, possibly most important of all, focused on how mostly social animals reacted in situations of social isolation. Although laboratory tasks presented to animals may indeed have been cognitively complex, many animals failed to demonstrate advanced capacities because of the specific nature of the task and the situation in which the task was presented (for examples see Menzel & Juno, 1982). Moreover, researchers often allowed their prejudices about animal capacities to influence their hypotheses. The phylogenetic closeness of primates to humans (e.g. Sarich & Cronin, 1977) and the large brains of cetaceans (e.g. Russell, 1979) led scientists to anticipate and accept that their communicative and cognitive capacities would be comparable to those of humans. (It should be noted however, that Morgane et al. (1986) expressed concern that the dolphin brain, although large even with respect to body size, may lack some of the complexity found in primates.) Yet experimenters rarely expected analogous abilities in birds and failed to search for such capacities. For many years, researchers argued that cognitive capacity was likely a consequence of relative cortical size, and that birds, lacking much in the way of cortical development, had to be inferior to mammals and primates (e.g. Premack, 1978). My own research on the cognitive and communicative capacities of the African grey parrots Psittacus erithacus is a particularly striking example of these issues (Pepperberg, 1999): birds once thought to be capable merely of mindless mimicry have demonstrated, under appropriate experimental conditions, referential use of elements of English speech and cognitive abilities (e.g. concepts of category, number, bigger/smaller, same/different, absence) comparable to those of a human child aged four to six years. Similarly, songbirds once thought merely to be emitting sets of innately predisposed vocal patterns acquired during

569

P1: IYP/... P2: IYP 0521823617c24.xml

570

CU1917B/McGregor

0 521 582361 7

April 12, 2005

13:31

I. M. Pepperberg a brief stage in their lives, have demonstrated vocal behaviour indicating various levels of cognitive processing and extensive memory. There are a number of good examples. White-crowned sparrows Zonotrichia leucophrys, assumed to acquire only their species-specific song during a limited sensitive period (Marler, 1970), show flexibility in learning elements of other species’songs when living in a complex social environment, not only in the laboratory but also in areas of sympatry (Baptista & Catchpole, 1989). Marsh wrens Cistothorus palustris actively choose which song in their 100–400 song repertoire to use in competitive countersinging so as to match, anticipate and possibly ‘jam’ the next song in their neighbours’ 100–400 song series (Kroodsma, 1979; Kroodsma & Byers, 1998). Nightingales Luscinia megarhynchos learn their 100 or so songs in chunks, much like humans learning long lists (Todt & Hultsch, 1998). Numerous avian species recognize subtle variations that differentiate their neighbours’ songs from those of strangers (Stoddard, 1996) and some even remember neighbours’ songs from one year to the next (Godard, 1991). Finally, if a song sparrow Melospiza melodia does not have an appropriate song type in its repertoire for an exact match in a countersinging bout, it selects one that is most similar (e.g. with the same introductory section,showing some level of same/different comprehension (Burt et al., 2002)). However, even these examples generally have focused on a single individual or one-on-one interactions and have, therefore, to some extent ignored the real world: that these birds are actually part of a larger network and that a countersinging bird, for example, would interact over time with usually at least two or three individuals (i.e. all its territorial neighbours), processing and storing all that information. Interestingly, advances in field techniques, both in recording and playback (e.g. McGregor et al., 1992; Naguib & Todt, 1997; Burt 2000), have not only allowed researchers to examine all the information available to their subjects but have also led these researchers to appreciate the complex cognitive processes that birds must be using to make sense of this information. A specific avian case involves the relationship between the complex cognitive task called ‘transitive inference’ and the natural situational behaviour of ‘eavesdropping’ among networks of songbirds (Dabelsteen et al., 1997; McGregor et al., 1997; Naguib & Todt, 1997; Naguib et al., 1999; Otter et al., 1999; Peake et al., 2001, 2002; Mennill et al., 2002). A discussion of the complexity of transitive inference, including the advances and pitfalls of laboratory work on the topic and a brief review of data on eavesdropping (for details, see Ch. 2 and several other chapters in this volume) will demonstrate some extent of avian cognitive capacities.

Transitive inference Transitive inference is one of several psychological tasks that engender incredible amounts of discussion both as to the actual mechanism by which it is

P1: IYP/... P2: IYP 0521823617c24.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

13:31

Cognitive aspects of networks and avian capacities performed and as to whether animals (particularly birds) are indeed capable of its performance (reviewed by Zentall, 2001). The task, as originally stated for human children, goes something like ‘Sam is taller than Bob. Bob is taller than Jack. Is Jack shorter than Sam?’ (e.g. McGonigle & Chalmers, 1984a); sometimes the problem is given as ‘Bob is taller than Jack. Sam is taller than Bob. Who is shortest?’. No training or rewards are involved when the task is given to children and the child is assumed to understand the concept of taller (or bigger, stronger, etc.). The child is thought to succeed by being able to integrate the two different vocal pieces of information, including the reversal from taller to shorter, in a conscious, cognitive manner; the complexity of the task derives from this integration and reversal and the strong likelihood that the process requires a mental representation of the integrated pieces of information for success. In one instance, adult subjects were given only five seconds to solve each of a series of transitive inference problems of various forms (‘Triangle is above circle. Square is below circle. Is triangle above square?’ ‘Circle is darker than square. Circle is lighter than triangle. Is triangle darker than square?’) and were then asked to report their reasoning patterns (Egan, 1983). Subjects who used different reasoning strategies (e.g. ordering the objects on a linear scale, which they then ‘scanned’, versus making individual images of the objects, which they then compared) made different amounts and different types of reasoning error (the linear thinkers had approximately 10% errors whereas the imaging group had approximately 38% errors). Use of a particular reasoning strategy was affected by aptitude for visualizing spatial transformations of figures and the context in which reasoning problems were posed, but each strategy involved some form of representation and integration: that is, cognitive processing. Whatever the strategy, the connection is evident between such a task and real-world knowledge of dominance hierarchies for any species living in a network of individuals; therefore, researchers assumed that demonstrating this understanding in non-human animals would be straightforward. Such, however, has not been the case. The task, as presented to non-humans, usually differs in a number of ways. First, the number of contrasting pairs involved usually is at least five (note that some studies such as those of McGonigle & Chalmers (1984b) use comparable numbers for children). Second, a hungry animal undergoes extensive training on pairwise comparisons where one of the pair is reinforced by a food reward (designated by +); the other is not reinforced (designated by −); the amount and type of food reward never varies. So the animal is trained to criterion on one pair (A+/B−), then to criterion on the next pair (B+/C−), likewise for subsequent pairs (C+/D−, D+/E−), and finally tested on an internal novel pair such as B/D to see which it will choose. The elements of the pair to be tested have never individually been shown to be the best or the worst and their relative worth has never been trained. Third, the animal is not specifically cued, as humans generally are, that the task

571

P1: IYP/... P2: IYP 0521823617c24.xml

572

CU1917B/McGregor

0 521 582361 7

April 12, 2005

13:31

I. M. Pepperberg involves relative judgements. Remember, humans are given specific verbal cues that could be seen as the equivalent of ‘four chocolates if you chose A, three if you choose B, two if you choose C, one if you choose D and none if you choose E; now, do you prefer B over D?’ So in some sense the animals are given a task that is more difficult than that initially given to children: the animal has not only to learn and remember a series of comparisons but also to understand the point of the query when presented with a novel pair, each member of which had actually led to reward in some instances. Rules that might have initially assisted during training (e.g. ‘choose the familiar item in a new pair even if it was not previously rewarded’) are of no use during testing, and additional rules developed during pairwise training (‘choose what was most recently rewarded’) would be misleading. Moreover, because the tests are not rewarded, the subject may not even be able to learn through successive iterations. Note, however, that the animal does not need to engage in reversal (i.e. the taller-to-shorter change mentioned above) and whether the animal is indeed cognitively engaged is unclear. Several researchers argue that non-cognitive mechanisms based simply on reward might be sufficient to explain the results of the pigeon subjects tested, which received only a reward/no-reward condition within pairs of items (i.e. given no reason to expect explicit relative relationships; see Couvillon & Bitterman, 1992; Wynne, 1997). At issue is the fact that the speed of acquisition and thus the number of trials would differ somewhat for each of the pairs, and consequently more errors would be made to some elements than others (see Zentall, 2001). Other researchers argue for cognitive processes based on spatial mapping (e.g. Weaver et al., 1997), such that the animals form some kind of linear set or mental representation to which they can retrospectively refer during testing. Interestingly, pigeons given explicit size cues to assist in forming a linear hierarchy did not learn any faster nor were they more accurate than those without such cues (von Ferson, 1989), suggesting that a linear model was not necessary for success. Possibly a form of ‘value transfer’ is involved, in that B, although never rewarded when given with A, accrues some of A’s value of 100% reward, whereas D, although always rewarded with E, loses some value by being connected to C, which accrues only 50% reward, and E, which is never rewarded (Weaver et al., 1997). Although a number of different species, from pigeons (Zentall, 2001) to chimpanzees (Gillan, 1981; Boysen et al., 1993), appear to succeed, the mechanisms used by animals and humans might differ. The transitive inference studies my students and I are attempting with African grey parrots, although not involving problems faced in the natural world, should avoid these issues; I present the material to clarify how an experiment on transitive inference could be performed in a laboratory to test whether a bird is using a representational, cognitive mechanism. The oldest parrot, Alex, already vocally

P1: IYP/... P2: IYP 0521823617c24.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

13:31

Cognitive aspects of networks and avian capacities designates the bigger/smaller of an object pair with respect to mass (Pepperberg & Brezinsky, 1991) and can quantify collections of up to six objects with vocal English number labels (Pepperberg, 1994); he is learning to label Arabic numerals so that we can determine if he can combine these abilities to rank order number symbols using transitive inference. Will he understand, without specific training to associate Arabic numerals directly with their physical values, that the symbol ‘5’ is greater than the symbol ‘3’? Only a task using equivalence relations and transitive inference can test this ability: he must use the commonality of English to correlate (form equivalence relations between) quantity and Arabic numerals, then use a form of transitive inference to identify the colour of one of a pair of Arabic numbers that is bigger or smaller (e.g. a blue 3, red 5). To succeed, he must base choice of ‘Arabic numeral X bigger/smaller than Arabic numeral Y’ on deductions and on inferences: deduce that an Arabic numeral has the same value as a vocal label, compare representations of quantity (mass) for which the numeral stands, infer rank ordering based on these representations (transitive inference) and then vocally report the result. Specific stimuli within pairs are not associated with reward (Wynne, 1997), and by requiring colour, not number label responses, rote replies cannot be used for a given pair. He has had no explicit training on ‘more/less than’rankings of individual elements (Arabic numerals) to be tested. The task involves use of both working and long-term memory (Geary et al., 2000). Note, too, that Alex was not trained to associate numbers with quantities sequentially: He first learned 3 and 4, then 2 and 5, then 6 (Pepperberg, 1987), and he does not produce vocal strings of number labels (i.e. does not say ‘one, two, three . . . ’); therefore, he has had no training in rank ordering numbers whatsoever. Initial trials are encouraging; he scored 15/18 in probes. Although Alex’s task will provide some intriguing information on transitive inference, the procedure still does not allow us to equate animal and human studies. Consequently, given the differences between the standard animal and human tasks, of particular interest are data collected when tasks more like those given to the animals were given to adult humans. When adult humans were given a non-vocal transitive task based on a computer game and not told that the game involved transitive inference, only 70% succeeded (Siemann, 1993). In a different experiment, when adult college students were given the exact same task as the animals (Werner et al., 1992), their accuracy for the transitive pair was impressive (approximately 95%), but only about two-thirds could explicitly state how they solved the problem. When two groups of humans were given the non-vocal animal task, with only one having been told that the task was inferential (Greene et al., 2001), both groups succeeded, again suggesting that the processing need not be conscious. Nevertheless, the informed group performed slightly better than those in the uninformed group, many of whom, after testing, indicated that they had

573

P1: IYP/... P2: IYP 0521823617c24.xml

574

CU1917B/McGregor

0 521 582361 7

April 12, 2005

13:31

I. M. Pepperberg inferred the hierarchical relationship. Finally, when human subjects are given the vocal human task with five elements even without reversal (‘John is taller than Bob. Bob is taller than Jim. Jim is taller than Richard. Richard is taller than David. Who is taller, Bob or Richard?’), humans often fail because the information given in such a task exceeds their memory capacity; many repetitions of the information are necessary before they succeed (Delius & Siemann, 1998; see also Woocher et al., 1978). Therefore, one can argue that the emphasis on overt hierarchical presentation in the standard human vocal task (A > B, B > C, A ? C) provides strong cues for its solution, and that in the longer case (A > B, B > C, C > D, D > E, B ? D) some form of learning is likely involved. When the task involves specific sequential training on abstract pairs and trial-and-error learning, however, subjects – be they pigeons (and by inference birds in general) or humans – do not truly demonstrate transitive inference; rather their results appear to be merely an artefact of the training and reward situation. Furthermore, transitive inference involving groups larger than three, even with explicit instruction as to the hierarchical nature of the task, appears to be more difficult than expected. The real issue, then, is not whether a pigeon can be taught something that has the surface appearance of transitive inference, but whether birds (a) are indeed capable of a task that has the same cognitive complexity of the human task and, probably more importantly, (b) are faced with, and able to solve, such a task in the real world. Although my research on the former issue is only in the earliest stages, my findings suggest that a parrot, with vocal and cognitive capacities that resemble those of very young children, is a good candidate for such a task (Pepperberg, 1999; see comment in Delius et al., 2000). Some data on chickens (Hogue et al., 1996) and recent field research on avian song (e.g. Peake et al., 2002), however, have provided intriguing information that birds use a network of information to solve at least simple transitive inference problems in their daily lives. Such data demonstrate a level of cognitive processing unexpected in a creature with a brain not much bigger than the size of a pea, or at most a shelled walnut, and a brain that is organized so differently from that of humans (cf. Jarvis & Mello, 2000). Excluding the chickens because they watched actual physical interactions, I will concentrate on songbirds, where the data involve decisions based on vocalizations.

Songbirds and transitive inference Songbirds, as noted above, live in a noisy environment of vocalizations and other sounds of numerous species. Even if we make simplifying assumptions that they rarely need to attend to the songs of other species or that they may not need to learn to ignore such songs, they still need to learn the repertoires of their neighbours. By so doing, they can determine, for example, whether, a territorial

P1: IYP/... P2: IYP 0521823617c24.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

13:31

Cognitive aspects of networks and avian capacities encroachment is by a neighbour with whom they can resolve the boundary dispute fairly quickly (e.g. Kroodsma, 1979; Beecher et al., 1996; Stoddard, 1996) or by a stranger who may pose a serious threat (Stoddard, 1996). Recent research suggests that songbirds also process and remember how their neighbours fare in territorial disputes with other neighbours and strangers and, by eavesdropping, they become aware of the relative dominance hierarchy of these birds and react with respect to that information, a case of transitive inference. Note that, for example, an unfamiliar floater male passing through a given area is quite likely to challenge several residents; knowing one’s relationship to one’s neighbours and how one’s neighbours have fared in such interactions could be advantageous. Now, eavesdropping by itself does not provide direct evidence for transitive inference, but it clearly sets the stage. Female black-capped chickadees Poecile atricapillus, for example, attend to the vocal duels between males and make their reproductive choices based on the outcomes (Mennill et al., 2002); they actively seek extra-pair copulations when their high-ranking mate has lost an interaction with a simulated intruder, and rarely if he has won. Data from such experiments (Ch. 7) show that the females are capable of processing information from at least two sources and making comparisons: A has beaten B, so B is less appealing. However, such data merely suggest that rankings can be made on the basis of several comparisons; for transitive inference, the question is whether females rank a number of different males and, if so, must they use overt interactions or can they interpolate (true transitive inference)? Note, too, that several females will be competing for the winning male in nature and female quality must also be taken into account. An interesting study would be to see how females judge their relative quality and whether they use some form of transitive inference. At a different level are results from nightingales using a simulated playback between two rivals; the target male noted which rival was overlapping the other (a sign of dominance) and proceeded to respond more strongly to the overlapper (Naguib & Todt, 1997; Naguib et al., 1999). The targeted male processed the interaction it heard, apparently viewed the overlapper as the greater threat and chose to react as though he needed to establish dominance over the overlapper: that is, given that A has beaten B, I had better beat A because he poses the greater threat; the untested inference is that B will be attending and will not have to be dealt with independently. Additional compelling data come from a study on male great tits Parus major (Peake et al., 2002; Ch. 2), which also appear to base their interactions with a simulated stranger on how that stranger has fared with a neighbour of known rank with respect to themselves. Another interesting case involves female great tits, who appear to decide whether to enter a male neighbour’s territory based on eavesdropping upon experimentally manipulated interactions between a stranger and her mate and the

575

P1: IYP/... P2: IYP 0521823617c24.xml

576

CU1917B/McGregor

0 521 582361 7

April 12, 2005

13:31

I. M. Pepperberg same stranger and said neighbouring male. The female makes her decision by inferring the ranking of the two resident males based on their respective abilities in dealing with the same intruder and is much more likely to enter the territory of the neighbour if he is inferred to be dominant to her mate (Otter et al., 1999; Chs. 2 and 7; see also Fig. 2.1b, p. 20). Interestingly, because the relative ranking of males chosen for the experiment was unknown and the choice was random as to whether a given playback would simulate a dominant or a subordinate interaction, the information might counter what she knows about previous interactions between her mate and her neighbour. Even if her response was merely to obtain more information, the experiment shows how much attention is paid to such interactions. Of course, her mate might act differently overall after losing a simulated encounter, and no one has yet observed natural interactions of this type, although experiments are underway to examine this possibility (K. A. Otter, personal communication; note that Mennill et al. (2002) did not observe any post-playback behavioural differences in chickadees). However, how do these interactions demonstrate cognitive complexity? Is the level of complexity as great as it is in the human case? Is cognition involved at all, or are some other mechanisms at play, as in the case of the trained pigeons?

Discussion These field studies did not test several levels of inference as did occur in the laboratory studies: that is, the birds were not exposed to a large number of different interactions among a simulated intruder and several different neighbours whose rankings were known and asked to rank the simulated intruder with respect to an untested comparison with these birds. Yet the field studies did demonstrate an interesting level of cognitive complexity. The male birds were asked to place themselves inside the rankings and to determine how they were likely to fare in a previously untested situation; the female birds appeared to act on an inference based on their observations. To determine the specific complexity of the situation, let us deconstruct two of the field tasks. In one task involving male responses, the bird judges the relative worth of the intruders in order to decide how to respond (Peake et al., 2002; see also Fig. 2.1e, p. 20). First, the bird must recognize that stranger A is in its territory, duel with it, then store its own rank with respect to that stranger. It must then attend to an interaction outside its territory between stranger A and a second stranger, B, and determine which has the higher rank. Subsequently, it must listen to one of those strangers, determine that it was B and not A (with whom it had previously interacted), remember this stranger’s rank with respect to the bird A with whom

P1: IYP/... P2: IYP 0521823617c24.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

13:31

Cognitive aspects of networks and avian capacities it did interact, remember its rank with respect to A, and then infer whether it has a chance against B or whether it should avoid B. In the task involving female responses, a bird judges the relative worth of her mate and neighbours based on the males’ interactions with a simulated intruder in order to decide how to respond (Otter et al., 1999; see also Fig. 2.1b, p. 20). First, the female has to distinguish her various neighbours from her mate and has likely stored the relative worth of her mate M, and each neighbour N. She must identify a new male, S, listen and determine his rank in a contest versus M, and then in another contest versus N. She must store and compare these two rankings and then infer the relative ranking of M and N based on their rankings with S, possibly updating her stored original memory. Although not examined, of interest would be whether she would try to search out S if he beat both M and N. Conceivably, birds, like humans (Duchaine et al., 2001), have an easier time making decisions that involve a social, familiar setting than they do if the same decision is required in an abstract context (Greene et al., 2001). That is, these tasks are somewhat simplified by being explicitly important to the bird’s survival and its reproductive success. Nevertheless, both situations may involve reversal and neither involves specific pairwise rewards; consequently, the likelihood of the results being merely some experimental artefact as in the case of, for example, the pigeon studies is unlikely. Clearly, of future interest would be the addition of simulated interactions by more intruders, C and D with A or B, and with the subjects’ other neighbours to determine how many dominance relationships a bird might encode. Chickadees may present an interesting case: their sense of overall ranking at winter feeding stations may be settled well in advance of their daily interactions, because the signals between any two birds landing at a feeding station are somewhat cursory (Popp et al., 1990). Given that such flocks involve approximately a dozen birds, the data suggest that individual birds may have some general understanding of their rank on a global basis, that is, via transitive relationships (K. A. Otter, personal communication). Could an experiment be designed to test whether (or how) a subject could rank others independent of their relationships to himself, or would a bird be able to rank others only in relationship to its need to avoid or engage in a direct confrontation? Possibly the rankings with respect to self actually complicate the issue, in that the bird must demonstrate some level of self-awareness as to where it fits into the hierarchy. Self-awareness is a separate but related issue in terms of animal abilities (e.g. Griffin, 1998) and merits some discussion in the present case, at least for clarification. Self-awareness, as used here, is distinguishable from ‘consciousness’: the full-blown central monitoring of sensory inputs and mental states, executive

577

P1: IYP/... P2: IYP 0521823617c24.xml

578

CU1917B/McGregor

0 521 582361 7

April 12, 2005

13:31

I. M. Pepperberg control of decision making and voluntary action, awareness of one’sown thoughts (being aware that one is aware (Carruthers, 1992)) and attribution of mental states to others (see discussion in Pepperberg & Lynn, 2000). Here, awareness describes a state of higher-order cognition in which information is represented, processed and used to control behaviour (e.g. Pepperberg, 1992). Thus the male tit is aware of its relationship to A, aware of A’s relationship to B, and makes a decision based on processing of these pieces of information; likewise, the female tit is aware of its relationships to M and N, the relationships between these males and S, and makes a decision based on the processing of these pieces of information. The point of evoking awareness is that the tits will probably have a mental representation of these pieces of information (e.g. Saidel, 2002) and use this representation in a ‘mind’s eye’ view to make a decision; a researcher would be hard pressed to characterize the tits’ processing in any of the non-cognitive, non-representational manners used to characterize the laboratory-based pigeon studies described above. The tits, however, may not be consciously aware of their use of these representations (e.g. be reacting to the specific situation in which, for example, its relationship to B is unclear by consciously weighing all the possible risks and future benefits on a personal basis and imputing the same to B, rather than reacting by chance based on lack of information). Devising a test to uncover conscious processing would be difficult.

Summary In sum, at least some birds appear capable of solving transitive inference tasks when dealing with a network of information, thus demonstrating complex cognitive processing requiring the formation of several representations, extensive memory for these various representations, and the ability to make inferences based on a hierarchical organization of these representations. The situations presented to great tits (Otter et al., 1999; Peake et al., 2002) are at least as complicated as those presented vocally to young children, and the results are not likely artefacts of experimental manipulation. Should Alex’s preliminary data hold, African grey parrots will also have succeeded in transitive inference. Although many objections exist to evaluating animal intelligence and cognition based on human tasks (see Pepperberg, 2001), the issue of importance here is that animals in nature are often faced with the same types of task as their human counterparts (both at present and historically) and so have been faced with the same evolutionary pressures on cognitive development. In such circumstances, evaluating their competence on what at first appears to be a human task does not ignore their natural behaviour, their motivations, their ecological niches or their sensorimotor competence. Rather, when animals are given tasks that are fully comparable to those given to humans

P1: IYP/... P2: IYP 0521823617c24.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

13:31

Cognitive aspects of networks and avian capacities with respect both to ecological validity and experimental control, we can make clear comparisons of animal and human abilities.

Acknowledgements Writing of this manuscript was supported by the MIT School of Architecture and Planning and a grant from the American Foundation. Research on African grey parrots was supported by NSF (IBN 96–03803) and REU supplements, the John Simon Guggenheim Foundation, the Kenneth A. Scott Charitable Trust, the Pet Care Trust, the University of Arizona Undergraduate Biology Research Program and many donors to the Alex Foundation.

References Baptista, L. F. & Catchpole, C. K. 1989. Vocal mimicry and interspecific aggression in songbirds: experiments using white-crowned sparrow imitation of song sparrow song. Behaviour, 109, 247–257. Beecher, M. D., Stoddard, P. K., Campbell, S. E. & Horning, C. L. 1996. Repertoire matching between neighbouring song sparrows. Animal Behaviour, 51, 917–923. Boysen, S. T., Berntson, G. G., Shreyer, T. A. & Quigley, K. S. 1993. Processing of ordinality and transitivity by chimpanzees (Pan troglodytes). Journal of Comparative Psychology, 107, 208–215. Burt, J. M. 2000. Use of a radio microphone array to study banded wren song interactions at the neighborhood level. Journal of the Acoustical Society of America, 108, 2583. Burt, J. M., Bard, S. C., Campbell, S. E. & Beecher, M. D. 2002. Alternative forms of song matching in song sparrows. Animal Behaviour, 63, 1143–1151. Carruthers, P. 1992. The Animals Issue: In Search of a Fundamental Theory. New York: Oxford University Press. Couvillon, P. A. & Bitterman, M. E. 1992. A conventional conditioning analysis of ‘transitive inference’ in pigeons. Journal of Experimental Psychology: Animal Behavior Processes, 18, 308–310. Dabelsteen, T., McGregor, P. K., Holland, J., Tobais, J. & Pedersen, S. B. 1997. The signal value of overlapping singing in male robins (Erithacus rubecula). Animal Behaviour, 53, 249–256. Delius, J. D. & Siemann, M. 1998. Transitive responding in animals and humans: exaptation rather than adaptation? Behavioural Processes, 42, 107–137. Delius, J. D., Jitsumori, J. & Siemann, M. 2000. Stimulus equivalencies through discrimination reversals. In: The Evolution of Cognition, ed. C. Heyes & L. Haber. Cambridge, MA: MIT Press, pp. 103–122. Duchaine, B., Cosmides, L. & Tooby, J. 2001. Evolutionary psychology and the brain. Current Opinion in Neurobiology, 11, 225–230. Egan, D. E. 1983. Retrospective reports reveal differences in people’s reasoning. Bell System Technical Journal, 62, 1675–1697.

579

P1: IYP/... P2: IYP 0521823617c24.xml

580

CU1917B/McGregor

0 521 582361 7

April 12, 2005

13:31

I. M. Pepperberg Geary, D. C., Hamson, C. O. & Hoard, M. K. 2000. Numerical and arithmetical cognition: a longitudinal study of process and concept deficits in children with learning disability. Journal of Experimental Child Psychology, 77, 236–263. Gillan, D. J. 1981. Reasoning in the chimpanzee: II. Transitive inference. Journal of Experimental Psychology: Animal Behavior Processes, 7, 150–164. Godard, R. 1991. Long-term memory of individual neighbours in a migratory songbird. Nature, 350, 228–229. Greene, A. J., Spellman, B., Dusek, J. A., Eichenbaum, H. B. & Levy, W. B. 2001. Relational learning with and without awareness: transitive inference using nonverbal stimuli in humans. Memory & Cognition, 29, 893–902. Griffin, D. R. 1998. From cognition to consciousness. Animal Cognition, 1, 3–16. Hogue, M.-E., Beaugrand, J. P. & Lagu¨e, P. C. 1996. Coherent use of information by hens observing their former dominant defeating or being defeated by a stranger. Behavioural Processes, 38, 241–252. Hulse, S. H., Fowler, H. & Honig, W. K. (eds.) 1968. Cognitive Processes in Animal Behavior. Hillsdale, NJ: Erlbaum. Jarvis, E. D. & Mello, C. V. 2000. Molecular mapping of brain areas involved in parrot vocal communication. Journal of Comparative Neurology, 419, 1–31. Kroodsma, D. E. 1979. Vocal dueling among male marsh wrens: evidence for ritualized expressions of dominance/subordinance. The Auk, 96, 506–515. Kroodsma, D. E. & Byers, B. E. 1998. Songbird song repertoires: an ethological approach to studying cognition. In: Animal Cognition in Nature, ed. R. P. Balda, I. M. Pepperberg & A. C. Kamil. London: Academic Press, pp. 305–336. Marler, P. 1970. A comparative approach to vocal learning: song development in white-crowned sparrows. Journal of Comparative and Physiological Psychology, 71, 1–25. McGonigle, B. & Chalmers, M. 1984a. The selective impact of question form and input mode on the symbolic distance effect in children. Journal of Experimental Child Psychology, 37, 525–554. 1984b. Are children any more logical than monkeys on the five-term series problem? Journal of Experimental Child Psychology, 37, 355–377. McGregor, P. K., Dabelsteen, T., Shepherd, M. & Pedersen, S. B. 1992. The signal value of matched singing in great tits: evidence from interactive playback experiments. Animal Behaviour, 43, 987–998. McGregor, P. K., Dabelsteen, T. & Holland, J. 1997. Eavesdropping in a territorial songbird communication network: preliminary results. Bioacoustics, 8, 253–254. Mennill, D. J., Ratcliffe, L. M. & Boag, P. T. 2002. Female eavesdropping on male song contests in songbirds. Science, 296, 873. Menzel, E. W., Jr & Juno, C. 1982. Marmosets (Saguinus fuscicollis): are learning sets learned? Science, 217, 750–752. Morgane, P. J., Jacobs, M. S. & Galaburda, A. 1986. Evolutionary morphology of the dolphin brain. In: Dolphin Cognition and Behavior: A Comparative Approach, ed. R. J. Schusterman, J. A. Thomas & F. G. Wood. Hillsdale, NJ: Erlbaum, pp. 5–29. Naguib, M. & Todt, D. 1997. Effects of dyadic vocal interactions on other conspecific receivers in nightingales. Animal Behaviour, 54, 1535–1543.

P1: IYP/... P2: IYP 0521823617c24.xml

CU1917B/McGregor

0 521 582361 7

April 12, 2005

13:31

Cognitive aspects of networks and avian capacities Naguib, M., Fitchel, C. & Todt, D. 1999. Nightingales respond more strongly to vocal leaders of simulated dyadic interactions. Proceedings of the Royal Society London, Series B, 266, 537–542. Otter, K. A., McGregor, P. K., Terry, A. M. R. et al. 1999. Do female great tits Parus major assess males by eavesdropping? A field study using interactive song playback. Proceedings of the Royal Society London, Series B, 265, 1045–1049. Peake, T. M., Terry, A. M. R., McGregor, P. K. & Dabelsteen, T. 2001. Male great tits eavesdrop on simulated male-to-male interaction. Proceedings of the Royal Society London, Series B, 268, 1183–1187. 2002. Do great tits assess rivals by combining direct experience with information gathered by eavesdropping? Proceedings of the Royal Society London, Series B, 269, 1925–1929. Pepperberg, I. M. 1987. Evidence for conceptual quantitative abilities in the African grey parrot: labeling of cardinal sets. Ethology, 75, 37–61. 1990. Cognition in an African grey parrot (Psittacus erithacus): further evidence for comprehension of categories and labels. Journal of Comparative Psychology, 104, 41–52. 1992. Proficient performance of a conjunctive, recursive task by an African grey parrot (Psittacus erithacus). Journal of Comparative Psychology, 106, 295–305. 1994. Numerical competence in an African grey parrot. Journal of Comparative Psychology, 108, 36–44. 1999. The Alex Studies: The Cognitive and Communicative Abilities of Grey Parrots. Cambridge, MA: Harvard University Press. 2001. Evolution of avian intelligence. In: The Evolution of Intelligence, ed. R. Sternberg & J. Kaufman. Mahwah, NJ: Erlbaum, pp. 315–337. Pepperberg, I. M. & Brezinsky, M. V. 1991. Relational learning by an African grey parrot (Psittacus erithacus): discriminations based on relative size. Journal of Comparative Psychology, 105, 286–294. Pepperberg, I. M. & Lynn, S. K. 2000. Possible levels of animal consciousness with reference to grey parrots (Psittacus erithacus). American Zoologist, 40, 893–901. Popp, J. W., Ficken, M. S. & Weise, C. M. 1990. How are agonistic encounters among black-capped chickadees resolved? Animal Behaviour, 39, 980–986. Premack, D. 1978. On the abstractness of human concepts: why it would be difficult to talk to a pigeon. In: Cognitive Processes in Animal Behavior, ed. S. H. Hulse, H. Fowler & W. K. Honig. Hillsdale, NJ: Erlbaum, pp. 421–451. Russell, I. S. 1979. Brain size and intelligence: a comparative perspective. In: Brain, Behavior and Evolution, ed. D. A. Oakley & H. C. Plotkin. London: Methuen, pp. 126–153. Saidel, E. 2002. Animal minds, human minds. In: The Cognitive Animal, ed. M. Bekoff, C. Allen & G. M. Burghardt. Cambridge, MA: Bradford Books and MIT Press, pp. 53–57. Sarich, V. M & Cronin, J. E. 1977. Generation length and rates of hominid evolution. Nature, 269, 354–355. Siemann, M. 1993. Transitive Inferenz: Experimentelle Untersuchung einer kognitiven Leistund. [Transitive inference: experimental investigation of a

581

P1: IYP/... P2: IYP 0521823617c24.xml

582

CU1917B/McGregor

0 521 582361 7

April 12, 2005

13:31

I. M. Pepperberg cognitive performance.] Ph.D. Thesis, University of Konstanz. [Cited in Delius & Siemann, 1998.] Stoddard, P. K. 1996. Vocal recognition of neighbors by territorial passerines. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca, NY: Cornell University Press, pp. 356–374. Todt, D. & Hultsch, H. 1998. Hierarchical learning, development and representation of song. In: Animal Cognition in Nature, ed. R. Balda, I. M. Pepperberg & A. C. Kamil. London: Academic Press, pp. 275–303. von Ferson, L. 1989. Kognitive Prozesse bei Tauben (Columba livia). [Cognitive processes in pigeons (Columba livia).] Pfaffenweiler: Centaurus. Weaver, J. E., Steirn, J. N. & Zentall, T. R. 1997. Transitive inference in pigeons: control for differential value transfer. Psychonomic Bulletin & Review, 4, 113–117. Werner, U. B., Koeppl, U. & Delius, J. D. 1992. Transitive Inferenz bei nicht-verbaler Aufgabendarbietung. [Transitive inference in nonverbal task presentation.] Zeitschrift f¨ ur Experimentelle und Angewandte Psychologie, 39, 662–683. Woocher, F. D., Glass, A. L. & Holyoak, K. J. 1978. Positional discriminability in linear orderings. Memory & Cognition, 6, 165–173. Wynne, C. D. L. 1997. Pigeon transitive inference: tests of simple accounts of a complex performance. Behavioural Processes, 39, 95–112. Zentall, T. R. 2001. The case for a cognitive approach to animal learning and behavior. Behavioural Processes, 54, 65–78.

P1: JZZ/... P2: JZZ/... 0521823617c25.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

10:36

25

Social complexity and the information acquired during eavesdropping by primates and other animals doro thy l. cheney & robert m. seyfarth University of Pennsylvania, Philadelphia, USA

Introduction In many of the studies reviewed in this book, eavesdropping takes the following form: a subject has the opportunity to monitor, or eavesdrop upon, an interaction between two other animals, A and B. The subject then uses the information obtained through these observations to assess A’s and B’s relative dominance or attractiveness as a mate (e.g. Mennill et al., 2002; Ch. 2). For example, Oliveira et al. (1998) found that male fighting fish Betta splendens that had witnessed two other males involved in an aggressive interaction subsequently responded more strongly to the loser of that interaction than the winner. Subjects’behaviour could not have been influenced by any inherent differences between the two males, because subjects responded equally strongly to the winner and the loser of competitive interactions they had not observed. Similarly, Peake et al. (2001) presented male great tits Parus major with the opportunity to monitor an apparent competitive interaction between two strangers by simulating a singing contest using two loudspeakers. The relative timing of the singing bouts (as measured by the degree of overlap between the two songs) provided information about each ‘contestant’s’ relative status. Following the singing interaction, one of the ‘contestants’ was introduced into the male’s territory. Males responded significantly less strongly to singers that had apparently just ‘lost’ the interaction (see also McGregor & Dabelsteen, 1996; Naguib et al., 1999; Ch. 2). What information does an individual acquire when it eavesdrops on others? In theory, an eavesdropper could acquire information of many different sorts: about A, about B, about the relationship between A and B, or about the place of Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

583

P1: JZZ/... P2: JZZ/... 0521823617c25.xml CU1917B/McGregor

584

0 521 582361 7

April 12, 2005

10:36

D. L. Cheney & R. M. Seyfarth A’s and B’s relationship in a larger social framework. The exact information acquired will probably reflect the particular species’ social structure. For example, songbirds like great tits live in communities in which six or seven neighbours surround each territory-holding male. Males appear to benefit from the knowledge that certain individuals occupy specific areas (e.g. Brooks & Falls, 1975), that competitive interactions between two different neighbours have particular outcomes, and that these outcomes are stable over time. We would, therefore, expect an eavesdropping great tit not only to learn that neighbour A was dominant to neighbour B, for example, but also to form the expectation that A was likely to defeat B in all future encounters. More speculatively, because the outcome of territorial interactions are often site specific (reviewed by Bradbury & Vehrencamp, 1998), we would expect eavesdropping tits to learn further that A dominates B in some areas but B dominates A in others. In contrast, the information gained from monitoring neighbours’ interactions would unlikely be sufficient to allow the eavesdropper to rank all of its neighbours in a linear dominance hierarchy, because not all neighbouring males would come into contact with one another. Such information would be difficult if not impossible to acquire; it might also be of little functional value. In contrast, species that live in large, permanent social groups have a much greater opportunity to monitor the social interactions of many different individuals simultaneously. Monkey species such as baboons Papio cynocephalus, for example, typically live in groups of 80 or more individuals, which include several matrilineal families arranged in a stable, linear dominance rank order (Silk et al., 1999). Offspring assume ranks similar to those of their mothers, and females maintain close bonds with their matrilineal kin throughout their lives. Cutting across these stable long-term relationships based on rank and kinship are more transient bonds: for example, the temporary associations formed between unrelated females whose infants are of similar ages, and the ‘friendships’ formed between adult males and lactating females as an apparent adaptation against infanticide (Palombit et al., 1997, 2001). In order to compete successfully within such groups, it would seem advantageous for individuals to recognize who outranks whom, who is closely bonded to whom, and who is likely to be allied to whom (Harcourt, 1988, 1992; Cheney & Seyfarth, 1990; see below). The ability to adopt a third party’s perspective and discriminate among the social relationships that exist among others would seem to be of great selective benefit. In this chapter, we review evidence for eavesdropping in selected primate species and we consider what sort of information is acquired when one individual observes or listens in on the interactions of others. We then compare eavesdropping by primates with eavesdropping in other animal species, focusing on both potential differences and directions for further research.

P1: JZZ/... P2: JZZ/... 0521823617c25.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

10:36

Social complexity and eavesdropping Playback sequence

Subjects

B1 threat-grunt + E1 scream

B and E

Control 1 (dominant’s kin)

B1 threat-grunt + D scream

B and E

Control 2 (no kin)

A threat-grunt + C scream

B and E

Test

Fig. 25.1. The protocol for playback experiments testing baboon females’ recognition of other individuals’ kin. B, the more dominant of the subjects; E, the more subordinate; B1 and E1 , the subjects’ close kin; A, C and D, signallers unrelated to either subject.

Knowledge about other animals’ kin Some of the first evidence that monkeys recognize other individuals’ social relationships emerged as part of a relatively simple playback experiment designed to document individual vocal recognition in vervet monkeys Cercopithecus aethiops (Cheney & Seyfarth, 1980). We had noticed that mothers often ran to support their juvenile offspring when these individuals screamed during aggressive interactions. This observation, like many others (e.g. Hansen, 1976; Gouzoules et al., 1984), suggested that mothers recognized the calls of their offspring. To test this hypothesis, we designed a playback experiment in which we played the distress scream of a juvenile to a group of three adult females, one of whom was the juvenile’s mother. As expected, mothers consistently looked toward the loudspeaker for longer durations than did control females. Even before she had responded, however, a significant number of control females looked at the mother. In so doing, they behaved as if they recognized not only the identity of signallers unrelated to themselves but also associated those individuals with specific adult females (Cheney & Seyfarth, 1980, 1982). In an attempt to replicate these results, we carried out a similar set of experiments on free-ranging baboons in the Okavango Delta of Botswana. In these experiments, two unrelated female subjects were played a sequence of calls that mimicked a fight between their close relatives (Fig. 25.1). The females’ immediate responses to the playback were videotaped and both subjects were followed for 15 minutes after the playback to determine whether their behaviour was affected by the calls they had heard. In separate trials, the same two subjects also heard two control sequences of calls (Fig. 25.1). The first sequence mimicked a fight involving the dominant subject’s relative and an individual unrelated to either female; the second mimicked a fight involving two individuals who were both unrelated to either female (for details see Cheney & Seyfarth, 1999).

585

P1: JZZ/... P2: JZZ/... 0521823617c25.xml CU1917B/McGregor

April 12, 2005

10:36

D. L. Cheney & R. M. Seyfarth 4

Duration of looking (s)

586

0 521 582361 7

Both kin Dominant kin No kin

3

2

1

0

Dominant female (B)

Subordinate female (E)

Fig. 25.2. The duration that the subject looked at the other female following each type of playback sequence. Histograms show means for 26 dyads in each of the three conditions.

After hearing the test sequence, a significant number of subjects looked toward the other female (Fig. 25.2), suggesting that they not only recognized the calls of unrelated individuals but also associated these individuals with their kin (or close associates). Females’ responses following the test sequence differed significantly from their responses following control sequences. Following the first control sequence, when only the dominant subject’s relative appeared to be involved in the fight, only the subordinate subject tended to look at her partner (Fig. 25.2). Following the second control sequence, when neither of the subjects’ relatives was involved, neither subject looked at the other (Fig. 25.2). Finally, following a significant proportion of test sequences, the dominant subject approached and supplanted (a mild form of aggression) the subordinate (Fig. 25.3). In contrast, when the two subjects approached each other following the two control sequences, the dominant rarely supplanted the subordinate (Fig. 25.3). Taken together, these experiments suggest that baboons and vervet monkeys recognize the individual identities of group members unrelated to themselves and that they recognize the social relationships that exist among these animals. Such knowledge can only be acquired by observing, or eavesdropping, on social interactions in which the observer is not involved and making the appropriate deductions. Other studies provide additional evidence of monkeys’ability to distinguish the close associates of other individuals. For example, in an experiment performed on captive long-tailed macaques Macaca fascicularis, Dasser (1988a) trained a female subject to choose between slides of one mother–offspring pair from her social

P1: JZZ/... P2: JZZ/... 0521823617c25.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

10:36

Social complexity and eavesdropping

Percentage of trials

50

Both kin Dominant kin No kin

40

30

20

10

0

Dominant supplants

Dominant approaches

Subordinate approaches

Fig. 25.3. The percentage of subjects’ first interactions with each other that took various forms following each playback sequence. Histograms show means for 26 dyads in each condition. Dominant supplants indicates that the dominant subject approached and supplanted the more subordinate subject. Dominant approaches indicates that the dominant subject approached the subordinate subject without supplanting her and/or interacted with her in a friendly manner. Subordinate approaches indicates that the subordinate subject approached the dominant subject and/or interacted with her in a friendly manner.

group and slides of two unrelated individuals from her group. Having learned to respond to one mother–offspring pair, the subject was then tested with 14 novel slides of familiar mothers and offspring paired with an equal number of novel slides of familiar unrelated animals matched for age and sex. In all tests, she correctly selected the mother–offspring pair. In so doing, she appeared to use an abstract category to classify pairs of individuals that was analogous to our concept of ‘mother–child affiliation’. Dasser (1988a) was able to exclude the possibility that mothers and offspring were matched according to physical resemblance, because subjects were unable to match unfamiliar mothers and offspring. Instead, individuals appeared to be classified according to their degree of association. Again, such knowledge of other individuals’ close associates can only be obtained by monitoring, or eavesdropping upon, their social interactions. Under natural conditions, it is difficult to determine whether animals distinguish between different categories of social relationships. Do monkeys recognize, for example, that mother–offspring bonds are distinct from sibling bonds or friendships even when all are characterized by high rates of interaction? In perhaps the only test of monkeys’ ability to recognize different categories of social affiliation, Dasser (1988b) trained a long-tailed macaque to identify a pair of siblings from

587

P1: JZZ/... P2: JZZ/... 0521823617c25.xml CU1917B/McGregor

588

0 521 582361 7

April 12, 2005

10:36

D. L. Cheney & R. M. Seyfarth her social group and then tested her ability to distinguish novel slides of familiar sibling pairs from familiar mother–offspring pairs, familiar pairs of less-closelyrelated matrilineal kin and familiar unrelated pairs. Although the subject did distinguish siblings from unrelated pairs and pairs of less-closely-related individuals, she was unable to discriminate between siblings and mothers and offspring. This failure may have occurred because the same female had previously been rewarded for picking the mother–offspring pair. It is also possible, however, that she did not distinguish between different kinship categories and simply chose the pair that was more closely affiliated. Natural patterns of aggression also reflect the knowledge that monkeys have of their group’s social network. In many monkey species, an individual who has just threatened or been threatened by another animal will often ‘redirect aggression’ by threatening a third, previously uninvolved, individual. Judge (1982) was the first to note that redirected aggression in rhesus macaques Macaca mulatta does not always occur at random. Rather than simply threatening any nearby individual, animals will instead specifically target a close matrilineal relative of their recent opponent. Similar kin-biased redirected aggression occurs in Japanese macaques Macaca fuscata (Aureli et al., 1992) and vervets (Cheney & Seyfarth, 1986, 1989). Kazem & Aureli (Ch. 10) further discuss the relationship between redirected aggression and communication networks.

Knowledge about other animals’ dominance ranks Dominance ranks offer another opportunity to test whether non-human primates gain information about other animals’relationships by eavesdropping on their social interactions. Like matrilineal kinship, linear, transitive dominance relations are a pervasive feature of social behaviour in groups of Old World monkeys. A linear, transitive rank order might emerge because individuals simply recognize who is dominant or subordinate to themselves. In this case, a linear hierarchy would occur as an incidental outcome of paired interactions and there would be no evidence to suggest that animals eavesdropped on others’interactions. Alternatively, a linear hierarchy might emerge because individuals genuinely recognize the transitive dominance relations that exist among others: a middle-ranking individual, for example, might know that A is dominant to B and B is dominant to C and, therefore, conclude that A must be dominant to C. Like knowledge of matrilineal kin, such knowledge could only be acquired through eavesdropping on the interactions of others. In many species of Old World monkeys, female dominance ranks are determined by the rank of an individual’s matriline (Walters & Seyfarth, 1987; Chapais, 1988). Knowledge of another female’s rank cannot, therefore, be obtained by

P1: JZZ/... P2: JZZ/... 0521823617c25.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

10:36

Social complexity and eavesdropping attending to absolute attributes such as age or size; instead, it demands the monitoring of other individuals’ interactions. Several observations and experiments suggest that monkeys do recognize the rank relations that exist among other females in their group. For example, dominant female baboons often grunt to mothers with infants as they approach the mothers and attempt to handle or touch their infants. Grunts seem to function to facilitate social interactions by appeasing anxious mothers, because an approach accompanied by a grunt is significantly more likely to lead to subsequent friendly interaction than is an approach without a grunt (Cheney et al., 1995a). Occasionally, however, a mother will utter a submissive call, or ‘fear bark’, as a dominant female approaches. Fear barks are an unambiguous indicator of subordination; they are never given to lower-ranking females. To test whether baboons recognize that only a more dominant animal can cause another individual to give a fear bark, we designed a playback experiment in which adult female subjects were played a causally inconsistent call sequence in which a low-ranking female apparently grunted to a high-ranking female and the higher-ranking female apparently responded with fear barks. As a control, the same subjects heard the same sequence of grunts and fear barks made causally consistent by the inclusion of additional grunts from a third female who was dominant to both of the other signallers. For example, if the inconsistent sequence was composed of female 6’sgrunts followed by female 2’sfear barks, the corresponding consistent sequence might begin with female 1’s grunts, followed by female 6’s grunts and ending with female 2’s fear barks. Some subjects were higher-ranking than the signallers; others were lower ranking. Regardless of their own relative ranks, subjects responded significantly more strongly to the causally inconsistent sequences, suggesting that they recognize not only the identities of different signallers but also the rank relations that exist among others in their group (Cheney et al., 1995b). Further suggestion that monkeys recognize other individuals’ ranks comes from observations on competition among adult female vervet monkeys for access to a grooming partner (Seyfarth, 1980). Such competition occurs when one female approaches two that are grooming, supplants one of them and then grooms with the female that remains. Interestingly, in those cases when a female approaches two groomers who are both subordinate to her, the lower ranking of the two groomers typically moves away, while the higher ranking remains (Cheney & Seyfarth, 1990). By remaining seated, the higher ranking of the two groomers acts as if she recognizes that, although they are both lower ranking than the approaching female, she is the higher ranking. Though not definitive, these observations suggest that females recognize not only their own status relative to other individuals but also other individuals’ status relative to each other.

589

P1: JZZ/... P2: JZZ/... 0521823617c25.xml CU1917B/McGregor

590

0 521 582361 7

April 12, 2005

10:36

D. L. Cheney & R. M. Seyfarth In other words, they appear to recognize a rank hierarchy (Cheney & Seyfarth, 1990). The ability to rank other group members is perhaps not surprising, given the evidence that captive monkeys and apes can be taught to rank objects according to an arbitrary sequential order (D’Amato & Colombo, 1989; Treichler & van Tilberg, 1996), the amount of food contained within a container (Gillan, 1981), their size or the number of objects contained within an array (e.g. Matsuzawa, 1985; Hauser et al., 1996; Brannon & Terrace, 1998). What distinguishes the social example, however, is the fact that, even in the absence of human training, female monkeys seem able to construct a rank hierarchy and then place themselves at the appropriate location within it.

Knowledge about more transient social relationships All of the studies discussed so far focus on interactions among females in groups where matrilineal kin usually retain close bonds and similar ranks throughout their lives. It might seem, therefore, that an individual could simply memorize the close associates and relative ranks of other females and thereafter navigate easily through a predictable network of social relationships. Not all social and rank relationships, however, are as stable as those among matrilineal kin. Some types of social bond are relatively transient, and some rank relationships – particularly among adult males – change often. Nonetheless, there is evidence that non-human primates also recognize these more transient associations. For example, under natural conditions, male and female hamadryas baboons Papio hamadryas form close, long-term bonds that can last for a number of years. Potential rivals appear to recognize the ‘ownership’ of specific females by other males and refrain from challenging those males for their females (Kummer et al., 1974). Experiments conducted in captivity have shown that rival males assess the strength of other males’ relationships with their females before attempting to challenge them. They do not attempt to take over a male’s female if the pair appears to have a close social bond (Bachmann & Kummer, 1980). Although similar experiments have not yet been conducted with savannah baboons, observational data suggest that these baboons, too, recognize the temporary bonds, or ‘friendships’, that are formed between males and lactating females (Palombit et al., 1997). For example, Smuts (1985) observed that males who had recently been threatened by another male often redirected aggression toward the female friends of their opponent (see Dunbar (1983) for similar observations on gelada baboons Theropithecus gelada). Monkeys also seem to recognize the bonds that exist between males and particular infants. In Tibetan macaques Macaca thibetana, males are often closely affiliated

P1: JZZ/... P2: JZZ/... 0521823617c25.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

10:36

Social complexity and eavesdropping with a particular infant in the group. Competitive interactions between males are mediated by the carrying of infants and a male will frequently carry an infant and present it to another male. In a study of such carrying (or ‘bridging’) behaviour, Ogawa (1995) observed that males more frequently provided other males with those males’ affiliated infants than with other, non-affiliated infants. Finally, there is evidence that monkeys recognize even very transient dominance relations among others. Dominance among male vervets, baboons and macaques is determined primarily by age, fighting ability, and, in some populations, the presence of alliance partners. As a result, rank relations among males are considerably less stable than they are among females (Walters & Seyfarth, 1987). In a study of a large social group of captive bonnet macaques Macaca radiata, Silk (1993, 1999) found that males formed linear, transitive dominance hierarchies that remained stable for only short periods of time. As in other primate species, males occasionally attempted to recruit alliance support during aggressive interactions (approximately 12% of all aggressive encounters). Significantly, males consistently solicited allies that outranked both themselves and their opponents. Males did not simply solicit the highest-ranking individual in the group or choose allies that outranked only themselves. Instead, soliciting males seemed to recognize not only their own rank relative to a potential ally but also the rank relation between the ally and their opponent. If dominance ranks remained stable, this might not have been a difficult task. However, over the course of one year, approximately half of the 16 males changed dominance rank each month (data from Table 3 in Silk, 1993). The males’ apparent ability to keep track of such highly transient rank relations suggests that they carefully monitored all aggressive interactions among other males, constantly updated their list of relative ranks and placed themselves accurately into each new list.

Eavesdropping by other mammals Data from dolphins Tursiops truncatus and hyaenas Crocuta crocuta suggest that non-human primates are not the only mammals in which individuals acquire information about many different individuals’social relationships (for other mammals see Chs. 17 and 18). When competing over access to females, male dolphins form dyadic and triadic alliances with selected other males, and allies with the greatest degree of partner fidelity are most successful in acquiring access to females (Connor et al., 1992, 1999, 2001). The greater success of high-fidelity alliances raises the possibility that males in newly formed alliances, or in alliances that have been less stable in the past, recognize the strong bonds that exist among others and are more likely to retreat when they encounter rivals with a long history of cooperative interaction.

591

P1: JZZ/... P2: JZZ/... 0521823617c25.xml CU1917B/McGregor

592

0 521 582361 7

April 12, 2005

10:36

D. L. Cheney & R. M. Seyfarth Like many species of Old World monkeys, hyaenas live in social groups comprising matrilines in which offspring inherit their mothers’ dominance ranks (Smale et al., 1993; Engh et al., 2000). Holekamp et al. (1999) played recordings of cubs’ ‘whoop’ calls to mothers and other breeding females. As with vervet monkeys and baboons, hyaena females responded more strongly to the calls of their own offspring and those of close relatives than to the calls of unrelated cubs. In contrast to vervets and baboons, however, unrelated females did not look at the cubs’ mothers. One explanation for these negative results is that hyaenas are unable to recognize third-party relationships, despite living in social groups that are superficially similar to those of many primates. It also remains possible, however, that hyaenas are simply uninterested in the calls of unrelated cubs. In fact, hyaenas’ patterns of alliance formation suggest that they do monitor other individuals’ interactions and extrapolate information about other animals’ relative ranks from their observations. During competitive interactions over meat, hyaenas often solicit alliance support from other, uninvolved individuals. When choosing to join ongoing skirmishes, hyaenas that are dominant to both of the contestants almost always support the more dominant of the two individuals (Engh et al., 2004). Similarly, when the ally is intermediate in rank between the two opponents, it inevitably supports the dominant individual. These data provide the first evidence in a non-primate species that alliance partners may be chosen on the basis of both the allies’and the opponents’relative ranks (Harcourt, 1988, 1992). They are consistent with the hypothesis that hyaenas are able to infer transitive rank relations among other group members.

Possible differences between primates and other animals Do primates differ from other animals in their ability to infer third-party social relationships through eavesdropping? We can identify at least three competing hypotheses. The first hypothesis argues that primates are in fact more intelligent than nonprimates. This intelligence is reflected not only in tests of captive animals but also in primates’ superior ability to keep track of complex social relationships. The difference between primates and non-primates is qualitative and fundamental and will be corroborated by future research. The second hypothesis maintains that selection has favoured the ability to recognize other individuals’ relationships in all species that live in large, complex social groups. According to this hypothesis, monkeys only appear to have a greater capacity to recognize third-party social relationships because they have received more attention than non-primates living in similarly large groups. Once this imbalance in research has been redressed, differences between primates and other

P1: JZZ/... P2: JZZ/... 0521823617c25.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

10:36

Social complexity and eavesdropping animals will disappear, to be replaced by a difference that depends primarily on group size and composition. The third hypothesis claims that neither phylogeny nor group size and composition have influenced animals’ ability to gain information about other individuals’ social relationships. It argues, in effect, that there are no species differences in ‘social intelligence’. Monkeys and hyaenas, for example, only appear to excel in their ability to recognize the relative ranks of allies and opponents because their large social groups allow them to display this knowledge. In contrast, studies of species that live in small social groups have to date focused primarily on observers’ ability to assess the dominance of only two individuals. Once monogamous and even solitary species have been given the opportunity to reveal what they know about the social relationships of many different individuals, they will be shown to possess a level of social intelligence that is no different from that found among animals living in large social groups. At present, it is difficult to test these alternative hypotheses; below we review some information that may be relevant. Hypothesis 1: primates have greater social intelligence than other species

Primates have larger brains for their body size than other vertebrates (Martin, 1983). Dunbar (2000) argued that this arises because primate social groups are not only larger but also more complex than those of other taxa. Primate groups are typically composed of many reproductively active males and females, and individuals interact regularly with both kin and non-kin, with whom they must simultaneously cooperate and compete for resources. Such social complexity may place strong selective pressure on the ability to recognize close associates of other individuals. To date, only monkeys and possibly dolphins have been shown to recognize the affiliative relationships that exist among other group members. In monkey groups, closely bonded individuals are usually matrilineal kin, but this is not always the case. The ability to classify other individuals into matrilineal or closely bonded subgroups is likely to be relatively complex, for several reasons. Matrilineal kin groups vary in size and not all individuals within a kin group interact at the same rate or in the same way. Moreover, no single behavioural measure underlies the associations between individuals and there is no threshold or defining criterion for a ‘close’ social bond. For example, females in many monkey species form the majority of their alliances with matrilineal kin, and high-ranking kin usually form alliances at higher rates than low-ranking kin (reviewed by Silk, 1987; Walters & Seyfarth, 1987). There is no evidence, however, that other group members more easily recognize the kin (or close associates) of high-ranking individuals than the kin of low-ranking individuals. Similarly, female kin usually

593

P1: JZZ/... P2: JZZ/... 0521823617c25.xml CU1917B/McGregor

594

0 521 582361 7

April 12, 2005

10:36

D. L. Cheney & R. M. Seyfarth occupy adjacent dominance ranks. This rule of thumb, however, cannot reliably be used to classify females into kin groups, because not all adjacently ranked females are kin. We do not yet know whether monkeys discriminate among different types of social bond: whether they distinguish, for example, among the bonds formed by mothers and offspring, sisters, or friends. Moreover, the degree to which there is a quantitative or qualitative threshold for learning to recognize that two other individuals share a close bond is not known. Furthermore, some social relationships among monkeys are transitive, while others are not. For example, if infant A1 and juvenile A2 both associate at high rates with a particular adult female A, it is usually correct to infer that the juvenile and infant are also closely bonded. Similarly, if A is dominant to B and B is dominant to C, it is usually true that A is dominant to C. In other cases, however, transitivity cannot be assumed. If infant baboon A1 and juvenile baboon A2 both associate at high rates with the same adult female and she associates with an adult male ‘friend’, we can infer that the male is probably also closely allied to the infant. However, it would incorrect to assume that he is equally closely allied to the juvenile, who may instead be more closely allied to another male who was previously the mother’s friend (Seyfarth, 1978; Smuts, 1985; Palombit et al., 1997). Baboon females from the same matriline often form friendships with different males; conversely, the same male may form simultaneous friendships with females from two different matrilines. In the latter case, the existence of a close bond between a male and two females does not predict a close bond between the two females. In fact, their relationship is likely to be as competitive as it is friendly (Palombit et al., 2001). Finally, as group size increases, the challenge of monitoring other individuals’ social relationships and dominance ranks increases exponentially. In a group of 80 animals (not an unusual size for many monkey species), each individual confronts 3160 different possible dyadic combinations and 82 160 different triadic combinations of individuals: numbers that may place considerable demands on the observer’s memory and inferential abilities. Preliminary evidence suggests that monkeys are able to monitor and remember the social ranks and relationships of many individuals simultaneously. Despite the lack of a consistent criterion for determining which individual is likely to be closely bonded with which others, monkeys appear to be able to distinguish the close associates of other group members. They appear to view their social groups not just in terms of the individuals that constitute them but also in terms of a web of social relationships in which certain individuals are linked with several others. Some learning experiments with captive animals support the view that primates are generally more adept than non-primates at classifying items according

P1: JZZ/... P2: JZZ/... 0521823617c25.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

10:36

Social complexity and eavesdropping to their relative relations. In oddity tests, for example, a subject is presented with three objects, two of which are the same and one of which is different, and asked to choose the object that is different. Monkeys and apes achieve high levels of accuracy in such tests even when tested with novel stimuli (Harlow, 1949; D’Amato et al., 1985; see also reviews by Tomasello & Call, 1997; Shettleworth, 1998). Baboons and chimpanzees can also learn to make abstract discriminations about relations between relations, matching patterns containing repeated samples of the same item with similar ‘same’ patterns (Premack, 1983; Oden et al., 1988; Fagot et al., 2001). In all cases, subjects’ performances suggest the use of an abstract hypothesis, because concepts like ‘odd’ specify a relation between objects independent of their physical features. In a similar manner, the concept ‘closely bonded’ can be applied to any two individuals and need not be restricted to specific pairs that look alike. Judgements based on relations among items have been demonstrated more often in non-human primates than in other taxa, and primates seem to recognize abstract relations more readily than at least some other animals. Although it is possible, for example, to train pigeons to recognize relations such as ‘same’, the procedural details of the test appear more critical for pigeons than they are for monkeys, and relational distinctions can easily be disrupted (Herrnstein, 1985; Wright et al., 1988; Wasserman et al., 1995). Rather than attending to the relations among stimuli, pigeons seem predisposed to focus on absolute stimulus properties and to form item-specific associations (reviewed by Shettleworth, 1998). Similarly, in tests of transitive inference, monkeys and apes appear to acquire a representation of series order that allows them to rank items even when some items in the list are missing. In contrast, pigeons seem to attend primarily to the association between adjacent pairs, which limits their ability to add or delete items from a list (D’Amato & Colombo, 1989; von Fersen et al., 1991; Treichler & van Tilberg, 1996; Zentall et al., 1996). Hypothesis 2: differences in ‘social intelligence’ are related to group size and complexity

If, as has been hypothesized, the recognition of third-party relationships confers a selective advantage because it allows individuals to remember who associates with whom, who outranks whom and who is allied to whom, we should expect to find evidence for this ability not just in non-human primates but also in any animal species that lives in large social groups composed of individuals of varying degrees of dominance rank and genetic relatedness. We would also predict that selection should have acted less strongly on this ability in solitary species and species living in small, egalitarian groups that are composed primarily either of close kin or of unrelated individuals. Thus, the ability to recognize the close

595

P1: JZZ/... P2: JZZ/... 0521823617c25.xml CU1917B/McGregor

596

0 521 582361 7

April 12, 2005

10:36

D. L. Cheney & R. M. Seyfarth associates of others should be evident in non-primate species such as hyaenas and lacking or less evident in some ape species, including gorillas Gorilla gorilla and orangutans Pongo pygmaeus. Although recent evidence that hyaenas recognize other individuals’ relative ranks lends support to this hypothesis, other comparative data are lacking. For example, no study has yet attempted to determine the extent to which any ape species is able to recognize the social relationships of other group members. Within the Primate order, species that live in large groups have a relatively larger neocortex than those that are solitary or live in small groups (Barton & Dunbar, 1997). A similar relation is found in carnivores (Barton & Dunbar, 1997) and toothed whales (Connor et al., 1998a,b; Marino, 1998), supporting the hypothesis that sociality has favoured the evolution of large brains (see also Jolly, 1966; Humphrey, 1976; Cheney & Seyfarth, 1990). Indeed, differences in social complexity may exert their effect even in species that lack a cortex entirely. In paper wasps Polistes dominulus, for example, there is a significant increase in the size of the antennal lobes and collar (a substructure of the calyx of the mushroom body) in females that nest colonially – with other queens – as opposed to solitary breeders (Ehmer et al., 2001). This increase in neural volume may be favoured because sociality places increased demand on the need to discriminate between familiar and unfamiliar individuals and to monitor other females’ dominance and breeding status. Clearly, therefore, neural correlates of sociality need not be restricted to higher mammals. Further supporting this argument are data from some other laboratory studies suggesting fewer differences between primates and other animals in the ability to make relational distinctions. For example, Alex, an African grey parrot Psittacus erithacus, is reported to make explicit same/different judgements about sets of objects (Pepperberg, 1992, Ch. 24). Similarly, sea lions Zalophus californianus (Schusterman & Krieger, 1986; Schusterman & Gisiner 1988) and dolphins (Herman et al. 1993; Mercado et al. 2000) have been taught to respond to terms such as ‘left’ and ‘bright’, which require the animals to assess relations among a variety of different objects. Finally, a number of species, including parrots (Pepperberg, 1994) and rats (Church & Meck, 1984; Capaldi, 1993), are able to assess quantities, suggesting that relatively abstract concepts of numerosity and transitivity may be pervasive among animals (reviewed by Shettleworth, 1998). Hypothesis 3: there are few differences in ‘social intelligence’ across species

Recent research on social eavesdropping (Ch. 2) by birds and fish indicates that even animals living in small social groups are capable of acquiring detailed information about other individuals’ relative dominance or attractiveness as a mate. Often, this information is of necessity restricted to a few other

P1: JZZ/... P2: JZZ/... 0521823617c25.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

10:36

Social complexity and eavesdropping individuals. For territorial species living in small family groups, questions about the ability to track social relationships among many other individuals are largely moot, because the opportunity to monitor interactions among all possible neighbours rarely arises. Eavesdropping on the competitive singing duets of strangers, for example, allows territorial songbirds to extract information about the two contestants’ relative dominance. Whether these birds would also be capable of recognizing a dominance hierarchy involving numerous individuals remains unclear. Although many species of songbirds form flocks during the winter, little is known about the social interactions that take place within such flocks, or the degree to which flock members recognize other individuals’ relative ranks (but see Popp, 1987). Recently, Bond et al. (2003) tested the prediction that socially living birds will display enhanced abilities to make transitive inferences by comparing the performance of highly social pinyon jays Gymnorhinus cyanocephalus with relatively non-social western scrub jays Aphelocoma californica. Using operant procedures, subjects were required to order a set of arbitrary stimuli by inference from a series of dyadic comparisons. Subjects of both species learned the sequence order, but pinyon jays did so more rapidly and more accurately than scrub jays. Although not conclusive, these results lend support to the hypothesis that social complexity may be correlated with superior performance in tasks involving the ranking of multiple stimuli (see also Hogue et al. (1996) for experiments with flock-dwelling domestic chickens Gallus domesticus). As yet, very little is known about the ability of non-primate mammals or birds to recognize social relationships of other individuals. Colonial white-fronted beeeaters Merops bullockoides offer one example of an avian society in which there would appear to be strong selective pressure for the recognition of the kin groups of other individuals. Observational evidence suggests that bee-eaters may recognize other individuals and kin groups and associate these groups with specific feeding territories (Emlen et al., 1995), although this has not yet been tested experimentally. Clearly, more data are needed from both natural and laboratory studies before we can draw any definitive conclusions about cognitive differences between primates and other animals, or between species living in large as opposed to small groups. It remains entirely possible that apparent species differences between primates and other animals in the recognition of third-party social relationships result more from differences in the social context in which eavesdropping occurs than from any cognitive differences in the ability to monitor social interactions. Given the opportunity to evaluate the social relationships of many different individuals, species living in small family groups and even primarily solitary species may well be shown to have similar abilities to those living in large social groups. It

597

P1: JZZ/... P2: JZZ/... 0521823617c25.xml CU1917B/McGregor

598

0 521 582361 7

April 12, 2005

10:36

D. L. Cheney & R. M. Seyfarth is to be hoped that future research will attempt to investigate the extent to which gregarious species in taxa other than primates are capable of recognizing the close associates and allies of other group members, and to determine the neural correlates, if any, of this ability.

Summary Non-human primates are skilled voyeurs. By observing or listening to the interactions of others, they acquire information about the social relationships of other individuals and learn to place these relationships within a larger social framework, such as a group of ranked, matrilineal families. Given the large, complex societies in which monkeys cooperate and compete, the adaptive value of such eavesdropping seems clear. At present, however, we do not know whether the information acquired by eavesdropping in primates differs significantly from the information acquired by individuals in other species. Primates (and a few other mammals) may be qualitatively different from other species in their ability to monitor the social relationships of many other individuals. Alternatively, the societies of birds, fish and other non-primate species – often superficially simpler than those of primates – may have led us to underestimate the information that individuals acquire about others. Finally, both hypotheses may have some validity. There may be qualitative differences in social intelligence between different taxonomic groups, but within each group the information acquired from eavesdropping may increase in sophistication with increasing social complexity. The chapters in this volume demonstrate that eavesdropping is widespread among animals. They set the stage for comparative research that examines differences between species in the information acquired about others.

References Aureli, F., Cozzolino, R., Cordischi, C. & Scucchi, S. 1992. Kin-oriented redirection among Japanese macaques: an expression of a revenge system? Animal Behaviour, 44, 283–291. Bachmann, C. & Kummer, H. 1980. Male assessment of female choice in hamadryas baboons. Behavioral Ecology and Sociobiology, 6, 315–321. Barton, R. A. & Dunbar, R. 1997. Evolution of the social brain. In: Machiavellian Intelligence II: Extensions and Evaluations, ed. A. Whiten & R. W. Byrne. Cambridge, UK: Cambridge University Press, pp. 240–263. Bond, A. B., Kamil, A. C. & Balda, R. P. 2003. Social complexity and transitive inference in corvids. Animal Behaviour, 65, 479–487. Bradbury, J. W. & Vehrencamp, S. L. 1998. Principles of Animal Communication. Sunderland, MA: Sinauer.

P1: JZZ/... P2: JZZ/... 0521823617c25.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

10:36

Social complexity and eavesdropping Brannon, E. & Terrace, H. 1998. Ordering of the numerosities 1 to 9 by monkeys. Science, 282, 746–749. Brooks, R. J. & Falls, J. B. 1975. Individual recognition by song in white-throated sparrows. III. Song features used in individual recognition. Canadian Journal of Zoology, 53, 1749–1761. Capaldi, E. J. 1993. Animal number abilities: implications for a hierarchical approach to instrumental learning. In: The Development of Numerical Competence, ed. S. T. Boysen & E. J. Capaldi. Hillsdale, NJ: Lawrence Erlbaum, pp. 191–209. Chapais, B. 1988. Rank maintenance in female Japanese macaques: experimental evidence for social dependency. Behaviour, 104, 41–59. Cheney, D. L. & Seyfarth, R. M. 1980. Vocal recognition in free-ranging vervet monkeys. Animal Behaviour, 28, 362–367. 1982. Recognition of individuals within and between groups of free-ranging vervet monkeys. American Zoologist, 22, 519–529. 1986. The recognition of social alliances among vervet monkeys. Animal Behaviour, 34, 1722–1731. 1989. Reconciliation and redirected aggression in vervet monkeys. Behaviour, 110, 258–275. 1990. How Monkeys See the World: Inside the Mind of Another Species. Chicago, IL: University of Chicago Press. 1999. Recognition of other individuals’ social relationships by female baboons. Animal Behaviour, 58, 67–75. Cheney, D. L., Seyfarth, R. M. & Silk, J. B. 1995a. The role of grunts in reconciling opponents and facilitating interactions among adult female baboons. Animal Behaviour, 50, 249–257. 1995b. The responses of female baboons (Papio cynocephalus ursinus) to anomalous social interactions: evidence for causal reasoning? Journal of Comparative Psychology, 109, 134–141. Church, R. M. & Meck, W. H. 1984. The numerical attribute of stimuli. In: Animal Cognition, ed. H. L. Roitblat, T. G. Bever & H. S. Terrace. Hillsdale, NJ: Lawrence Erlbaum, pp. 445–464. Connor, R. C., Smolker, R. A. & Richards, A. F. 1992. Dolphin alliances and coalitions. In: Coalitions and Alliances in Humans and other Animals, ed. A. H. Harcourt & F. de Waal. Oxford: Oxford University Press, pp. 415–443. Connor, R. C., Mann, J., Tyack, P. L. & Whitehead, H. 1998a. Social evolution in toothed whales. Trends in Ecology and Evolution, 13, 228–232. 1998b. Reply. Trends in Ecology and Evolution, 13, 408. Connor, R. C., Heithaus, R. M. & Barre, L. M. 1999. Superalliance of bottlenose dolphins. Nature, 371, 571–572. 2001. Complex social structure, alliance stability, and mating access in a bottlenose dolphin ‘super-alliance’. Proceedings of the Royal Society of London, Series B, 268, 263–267. D’Amato, M. & Colombo, M. 1989. Serial learning with wild card items by monkeys (Cebus apella): implications for knowledge of ordinal rank. Journal of Comparative Psychology, 103, 252–261.

599

P1: JZZ/... P2: JZZ/... 0521823617c25.xml CU1917B/McGregor

600

0 521 582361 7

April 12, 2005

10:36

D. L. Cheney & R. M. Seyfarth D’Amato, M., Salmon, P. & Colombo, M. 1985. Extent and limits of the matching concept in monkeys (Cebus apella). Journal of Experimental Psychology: Animal Behavior Processes, 11, 35–51. Dasser, V. 1988a. A social concept in Java monkeys. Animal Behaviour, 36, 225–230. 1988b. Mapping social concepts in monkeys. In: Machiavellian Intelligence: Social Expertise and the Evolution of Intellect in Monkeys, Apes, and Humans, ed. R. W. Byrne and A. Whiten. Oxford: Oxford University Press, pp. 85–93. Dunbar, R. 1983. Structure of gelada baboon reproductive units. III. The male’s relationship with his females. Animal Behaviour, 31, 565–575. 2000. Causal reasoning, mental rehearsal, and the evolution of primate cognition. In: Vienna Series in Theoretical Biology: The Evolution of Cognition, ed. C. Heyes & L. Huber. Cambridge, MA: MIT Press, pp. 205–219. Ehmer, B., Reeve, H. K. & Hoy, R. R. 2001. Comparison of brain volumes between single and multiple foundresses in the paper wasp Polistes dominulus. Brain, Behavior and Evolution, 57, 161–168. Emlen, S. T., Wrege, P. H. & Demong, N. J. 1995. Making decisions in the family: an evolutionary perspective. American Scientist, 83, 148–157. Engh, A. L., Esch, K., Smale, L. & Holekamp, K. E. 2000. Mechanisms of maternal rank ‘inheritance’ in the spotted hyaena, Crocuta crocuta. Animal Behaviour, 60, 323–332. Engh, A. L., Siebert, E., Greenberg, D. & Holekamp, K. 2004. Primate-like cognition in a social carnivore. Animal Behaviour, in press. Fagot, J., Wasserman, E. A. & Young, M. 2001. Discriminating the relation between relations: the role of entropy in abstract conceptualization by baboons (Papio papio) and humans (Homo sapiens). Journal of Experimental Psychology: Animal Behavior Processes, 27, 316–328. Gillan, D. J. 1981. Reasoning in the chimpanzee, II. Transitive inference. Journal of Experimental Psychology: Animal Behavior Processes, 7, 150–164. Gouzoules, S., Gouzoules, H. & Marler, P. 1984. Rhesus monkey (Macaca mulatta) screams: representational signalling in the recruitment of agonistic aid. Animal Behaviour, 32, 182–193. Harlow, H. F. 1949. The formation of learning sets. Psychological Review, 56, 51–65. Hansen, E. W. 1976. Selective responding by recently separated juvenile rhesus monkeys to the calls of their mothers. Developmental Psychobiology, 9, 83–88. Harcourt, A. H. 1988. Alliances in contests and social intelligence. In: Machiavellian Intelligence: Social Expertise and the Evolution of Intellect in Monkeys, Apes, and Humans, ed. R. W. Byrne & A. Whiten. Oxford: Oxford University Press, pp. 132–152. 1992. Coalitions and alliances: are primates more complex than non-primates? In: Coalitions and Alliances in Humans and Other Animals, ed. A. H. Harcourt & F. de Waal. New York: Oxford University Press, pp. 86–99. Hauser, M. D., MacNeilage, P. & Ware, M. 1996. Numerical representations in primates. Proceedings of the National Academy of Sciences, USA, 93, 1514–1517. Herman, L. M, Pack, A. A & Morrel-Samuels, P. 1993. Representational and conceptual skills of dolphins. In: Comparative Cognition and Neuroscience, ed. H. L. Roitblat, L. M. Herman & P. E. Nachtigall. Hillsdale, NJ: Lawrence Erlbaum, pp. 403–442.

P1: JZZ/... P2: JZZ/... 0521823617c25.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

10:36

Social complexity and eavesdropping Herrnstein, R. J. 1985. Riddles of natural categorization. Philosophical Transactions of the Royal Society of London, Series B, 308, 129–144. Hogue, M. E., Beaugrand, J. P. & Laugue, P. C. 1996. Coherent use of information by hens observing their former dominant defeating or being defeated by a stranger. Behavioural Processes, 38, 241–252. Holekamp, K. E., Boydston, E. E., Szykman, M. et al. 1999. Vocal recognition in the spotted hyaena and its possible implications regarding the evolution of intelligence. Animal Behaviour, 58, 383–395. Humphrey, N. 1976. The social function of the intellect. In Machiavellian Intelligence: Social Expertise and the Evolution of Intellect in Monkeys, Apes, and Humans, ed. R. W. Byrne and A. Whiten. Oxford: Oxford University Press, pp. 13–26. Jolly, A. 1966. Lemur Behavior. A Madagascar Field Study. Chicago, IL: University of Chicago Press. Judge, P. 1982. Redirection of aggression based on kinship in a captive group of pigtail macaques. International Journal of Primatology, 3, 301. Kummer, H., Goetz, W. & Angst, W. 1974. Triadic differentiation: an inhibitory process protecting pair bonds in baboons. Behaviour, 49, 62–87. Marino, L. 1998. Quantifying brain–behavior relations in cetaceans and primates. Trends in Ecology and Evolution, 13, 408. Martin, R. D. 1983. Primate Origins and Evolution: A Phylogenetic Reconstruction. Princeton: Princeton University Press. Matsuzawa, T. 1985. Use of numbers by a chimpanzee. Nature, 315, 57–59. McGregor, P. K. & Dabelsteen, T. 1996. Communication networks. In: Ecology and Evolution of Acoustic Communication in Birds, eds. D. E. Kroodsma & E. H. Miller. Ithaca, NY: Cornell University Press, pp. 409–425. Mennill, D. J., Ratcliffe, L. M. & Boag, P. T. 2002. Female eavesdropping on male song contests in songbirds. Science, 296, 873. Mercado, E., Killebrew, D. A., Pack, A. A., Macha, I. & Herman, L. M. 2000. Generalization of ‘same-different’ classification abilities in bottlenosed dolphins. Behavioural Processes, 50, 79–94. Naguib, M., Fichtel, C. & Todt, D. 1999. Nightingales respond more strongly to vocal leaders of simulated dyadic interactions. Proceedings of the Royal Society of London, Series B, 266, 537–542. Oden, D. Thompson, R. & Premack, D. 1988. Spontaneous transfer of matching by infant chimpanzees (Pan troglodytes). Journal of Experimental Psychology: Animal Behavior Processes, 14, 140–145. Ogawa, H. 1995. Recognition of social relationships in bridging behavior among Tibetan macaques (Macaca thibetana). American Journal of Primatology, 35, 305–310. Oliveira, R. F., McGregor, P. K. & Latruffe, C. 1998. Know thine enemy: fighting fish gather information from observing conspecific interactions. Proceedings of the Royal Society of London, Series B, 265, 1045–1049. Palombit, R. A., Seyfarth, R. M. & Cheney, D. L. 1997. The adaptive value of ‘friendships’ to female baboons: experimental and observational evidence. Animal Behaviour, 54, 599–614.

601

P1: JZZ/... P2: JZZ/... 0521823617c25.xml CU1917B/McGregor

602

0 521 582361 7

April 12, 2005

10:36

D. L. Cheney & R. M. Seyfarth Palombit, R. A., Cheney, D. L. & Seyfarth, R. M. 2001. Female–female competition for male ‘friends’ in wild chacma baboons (Papio cynocephalus ursinus). Animal Behaviour, 61, 1159–1171. Peake, T. M., Terry, A. M. R., McGregor, P. K. & Dabelsteen, T. 2001. Male great tits eavesdrop on simulated male–male vocal interactions. Proceedings of the Royal Society of London, Series B, 268, 1183–1187. Pepperberg, I. M. 1992. Proficient performance of a conjunctive, recursive task by an African grey parrot (Psittacus erithacus). Journal of Comparative Psychology, 106, 295–305. 1994. Numerical competence in an African gray parrot (Psittacus erithacus). Journal of Comparative Psychology, 108, 36–44. Popp, J. W. 1987. Choice of opponents during competition for food among American goldfinches. Ethology, 75, 31–36. Premack, D. 1983. The codes of man and beast. Behavioral and Brain Sciences, 6, 125–167. Schusterman, R. J. & Gisiner, R. 1988. Artificial language comprehension in dolphins and sea lions: the essential cognitive skills. Psychological Record, 38, 311–348. Schusterman, R. J. & Krieger, K. 1986. Artificial language comprehension and size transposition by a California sea lion (Zalophus californianus). Journal of Comparative Psychology, 100, 348–355. Seyfarth, R. M. 1978. Social relationships among adult male and female baboons. II. Behavior throughout the female reproductive cycle. Behaviour, 64, 227–247. Seyfarth, R. M. 1980. The distribution of grooming and related behaviors among adult female vervet monkeys. Animal Behaviour, 28, 798–813. Shettleworth, S. 1998. Cognition, Evolution, and Behaviour. Oxford: Oxford University Press. Silk, J. B. 1987. Social behavior in evolutionary perspective. In: Primate Societies, ed. B. B. Smuts, D. L. Cheney, R. M. Seyfarth, R. W. Wrangham & T. Struhsaker. Chicago, IL: University of Chicago Press, pp. 318–329. 1993. Does participation in coalitions influence dominance relationships among male bonnet macaques? Behaviour, 126, 171–189. 1999. Male bonnet macaques use information about third-party rank relationships to recruit allies. Animal Behaviour, 58, 45–51. Silk, J. B., Seyfarth, R. M. & Cheney, D. L. 1999. The structure of social relationships among female savanna baboons. Behaviour, 136, 679–703. Smale, L., Frank, L. G. & Holekamp, K. E. 1993. Ontogeny of dominance in free-living spotted hyaenas: juvenile rank relations with adult females and immigrant males. Animal Behaviour, 46, 467–477. Smuts, B. 1985. Sex and Friendship in Baboons. Chicago, IL: Aldine. Tomasello, M. & Call, J. 1997. Primate Cognition. Oxford: Oxford University Press. Treichler, F. & van Tilburg, D. 1996. Concurrent conditional discrimination tests of transitive inference by macaque monkeys: list linking. Journal of Experimental Psychology, Animal Behavior Processes, 22, 105–117. von Fersen, L., Wynne, C., Delius, J. & Staddon, J. 1991. Transitive inference formation in pigeons. Journal of Experimental Psychology: Animal Behavior Processes, 17, 334–341.

P1: JZZ/... P2: JZZ/... 0521823617c25.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

10:36

Social complexity and eavesdropping Walters, J. R. & Seyfarth, R. M. 1987. Conflict and cooperation. In: Primate Societies, ed. B. B. Smuts, D. L. Cheney, R. M. Seyfarth, R. W. Wrangham & T. Struhsaker. Chicago, IL: University of Chicago Press, pp. 306–317. Wasserman, E. A., Hugart, J. A. & Kirkpatrick-Steger, K. 1995. Pigeons show same-different conceptualization after training with complex visual stimuli. Journal of Experimental Psychology: Animal Behavior Processes, 21, 248–252. Wright, A., Cook, R. & Rivera, J. 1988. Concept learning by pigeons: matching to sample with trial-unique video picture stimuli. Animal Learning and Behavior, 16, 436–444. Zentall, T., Weaver, J. & Sherburne, L. 1996. Value transfer in concurrent-schedule discriminations by pigeons. Animal Learning and Behavior, 24, 401–409.

603

P1: JZZ/... P2: JZZ/... 0521823617c26.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

10:42

26

Communication networks in a virtual world andrew m.r. terry1 & robert l achl an2 1 2

University of Copenhagen, Denmark University of North Carolina, Chapel Hill, USA

Introduction When one individual is signalling, or two individuals are interacting, they do so within a network of potential receivers (see McGregor, 1993; McGregor & Dabelsteen, 1996; Ch. 1). As the other chapters in this book show, the decisions that both signallers and receivers make about their future behaviour are thus contingent not only on each other’s behaviour but also on a wider network of individuals (McGregor & Peake, 2000). This view is finding support in empirical studies showing that individuals use information that could only be extracted from network interactions (e.g. Oliveira et al., 1998; Peake et al., 2001, 2002; Ch. 2). These empirical findings also have implications for the theoretical study of signalling strategies (e.g. Nowak & Sigmund, 1998; Johnstone, 2001). For example, an individual’s signalling strategy may no longer be predicted solely from the responses of an opponent. We consider that the signalling strategies of individuals will only be explored realistically by models that include the potential responses of signallers to other individuals. In this chapter, we ask whether current modelling approaches can be adapted to include networks or whether new modelling techniques need to be considered. The aim of creating a model is to advance our conceptual understanding of a system and create empirically testable hypotheses (Wilson, 2000; Hemelrijk, 2002) by simplifying the real world using words or mathematical expressions. Most hypotheses start with a verbal model and develop into mathematical models, which more precisely specify limiting conditions and assumptions and often provide a deeper understanding of the logic underlying the hypothesis. Models vary in Animal Communication Networks, ed. Peter K. McGregor. Published by Cambridge University Press.
c Cambridge University Press 2005.

604

P1: JZZ/... P2: JZZ/... 0521823617c26.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

10:42

Communication networks in a virtual world their complexity and the level of explanatory power they provide: often, simple models provide a deeper and general understanding of underlying dynamics but also contain more restrictive (or unrealistic) assumptions. For example, simple biological models, such as the Lotka–Volterra equations of predator–prey interactions, provided a powerful heuristic insight into the cyclic nature of population abundances but contained highly restrictive assumptions, primarily concerning the heterogeneity of populations and environments, that limited their application to specific cases (Maynard Smith, 1982; Begon et al., 1990; Wilson, 2000). In population genetics models, assumptions are typically made about population sizes and the pattern of distribution of traits. Nevertheless, such assumptions often do not qualitatively affect the conclusion (e.g. Turelli & Barton, 1994). Simpler models are likely to be mathematically tractable, allowing a more complete analysis and understanding of the processes underlying the system being investigated. Complexity increases as more realistic assumptions are incorporated and the possibility of mathematical analysis becomes more remote. However, in recent years, the rapid rise in computer power has allowed theoretical methods to acquire new levels of complexity, mostly through the use of simulation modelling (Grimm, 1999). Simulation models allow as many variables to be included as the investigators have imagination, programming skills and time. The downside to this is that the models are harder to generalize (it is harder to ensure that all the variables are realistic for a wide range of conditions) and that it is more difficult to isolate precisely the factors that are causing an effect of interest (Wilson, 2000). An attractive solution to this dilemma is to use a variety of modelling techniques, differing in how many assumptions are required (Dieckman, 1997), with the specific aim of identifying the parameters, variables and assumptions that are critical in explaining the behaviour of the system under investigation (Wilson, 2000). In this chapter, we discuss the role modelling has played in the conceptual development of communication networks. In doing this, we also examine which features of communication network models are especially important, often by identifying which unrealistic assumptions are likely to change qualitatively the conclusions reached about behaviour in networks. Finally, we examine how theories of communication that include networks are likely to differ from those that do not. Conceptualizing networks Network structure in existing models

A network is an association of nodes connected to each other by some means. In animal communication, the nodes are individual animals and the connections (or links) are patterns of communication between them. For every signal,

605

P1: JZZ/... P2: JZZ/... 0521823617c26.xml CU1917B/McGregor

606

0 521 582361 7

April 12, 2005

10:42

A. M. R. Terry & R. Lachlan there are likely to be several possible receivers, leading to the concept of the communication network. This concept modifies our understanding of the costs and benefits of signalling (e.g. McGregor & Peake, 2000). For example, individuals winning an interaction may be more willing to publicize the interaction than those losing it. How does this view of communication fit with standard methods used to model animal signalling and do different methods need to be developed? Most current models of animal communication consider only dyadic encounters between individuals. For example, game theory models of communication normally study the evolution of strategies through the responses of two interactants to each other (examples in Maynard Smith, 1982). When analytical models have been used to study interactions within large groups, they typically make restrictive assumptions that may limit their ability to assess the evolutionary pressures of communicating within networks. For example, individuals may be drawn at random from the population to interact or they may have perfect knowledge about the behaviour of all other individuals (e.g. Nowak & Sigmund, 1998; Johnstone, 2001). It seems that, whereas traditional evolutionary game theory models are well suited to the study of dyadic encounters and contests, a network of individuals all gathering information from each other and using this in signalling interactions, which may or may not be directed at other receivers as well as the primary receiver, may prove too complicated to be tractable (although see p. 617). It is with this in mind that we consider the role of simulation modelling as a tool for studying communication in networks. Simulation models are widely used in ecology as they allow the user to incorporate an unlimited number of variables and parameters in the model, and they are being increasingly used in behavioural studies. A recent trend has favoured individually based and spatially explicit models, which contain a discrete population of individual animals within a defined spatial environment. Individuals are governed by a series of movement and behavioural rules (Grimm, 1999). These models, therefore, allow an increase in biological realism (Wilson, 2000) by replacing the inaccurate assumptions about the random or structured patterns of interaction within a population found in simpler models with assumptions that better capture spatial structuring. These types of model have been used to show how complex collective behaviours can arise from the interaction of individuals obeying simple behavioural rules (Hemelrijk, 2002). For example, Hemelrijk (2000) created an individually based model of dominance interactions within a social group to show that increased aggression caused the emergence of ‘selfish herd’ organization in the group. Previous theoretical studies of selfish herding had difficulties equating the complex movement rules needed to make individuals aggregate in the models (i.e. with tight clustering of dominants surrounded by subordinate individuals (Hamilton, 1971)) with observations of this herding

P1: JZZ/... P2: JZZ/... 0521823617c26.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

10:42

Communication networks in a virtual world behaviour in the field (Morton et al., 1994; Viscido et al., 2002). Individuals in the model of Hemelrijk (2000) were governed by two rules; the tendency to aggregate and the tendency to enter into dominance interactions. When an individual lost a dominance interaction, it would flee and the winner would move to occupy its space. In the model, it became possible for weak individuals to get ‘caught in a rut’ and lose several interactions in a row causing them to move to the periphery of the group, whereas the strongest individuals would remain in the centre. This is an example of a model that made no assumptions about the way in which individuals chose to aggregate and yet patterns of social organization that were similar to those observed in the wild emerged as products of aggressive interactions and individual differences. Small-world network analysis

As explained above, most models of communication have not considered the role of individuals other than the immediate interactants; when they have, they placed unrealistic assumptions on the information gathered. A further consideration in the theoretical analysis of communication and information gathering in a network is whether the way in which the network is organized affects its function. In recent years, considerable research has focused on the structure and organization of networks. Most networks associated with social interactions may be physically limited. For example, the interactions in a territorial system are mostly restricted to neighbouring individuals. However, a consideration of communication networks means that a larger group of individuals must be considered in the analysis as the long-range signals most commonly used for advertising or aggressive interactions usually travel much further than an individual’s immediate neighbours (McGregor & Peake, 2000). The analysis of networks in fields as diverse as metabolic pathways in eukaryote cells (Jeong et al., 2000), food webs (Williams et al., 2000; but see Dunne et al., 2002) and links in the World Wide Web (Albert et al., 1999) has found that they show a number of similar structural properties. As a result of these similarities, such networks are referred to as ‘small-worlds’ or ‘scale-free’ networks (Watts & Strogatz, 1998; Barabasi & Albert, 1999). Fundamentally, these networks are dynamic: a new node joining the network is likely to attach preferentially to certain existing nodes. As a result, networks arise that contain tight clustering around some highly connected nodes, called ‘hubs’. There is no one centralized dominant node; consequently, organization is spread between the few highly connected hubs. This organization is both the network’s main asset and its Achilles heel, as it means that the behaviour of most nodes has little impact on the network at large, but removing one of the hubs can have a critical impact on the flow of information through the network (Albert & Barabasi, 2002).

607

P1: JZZ/... P2: JZZ/... 0521823617c26.xml CU1917B/McGregor

608

0 521 582361 7

April 12, 2005

10:42

A. M. R. Terry & R. Lachlan A key consequence of this structure is that these networks have short path lengths between any two nodes (i.e. any two individuals within the population can be linked by a small number of connections). For example, Kudo & Dunbar (2001) found that primate groups remained socially interconnected despite being fragmented into small cliques, with males possibly serving the role as hubs, connecting female groups. The implication of this phenomenon is that information exchanged between individuals may spread widely and rapidly throughout the network. Watts & Strogatz (1998) considered a simple model of a contagion that examined how a disease might spread through a small-world network. They found, rather frighteningly, that only a few hubs were required for the disease to spread rapidly through the population. This model can be directly compared to a communication situation, where the probability of an individual producing a signal is affected by the number of others it perceives signalling. The conclusion would be that, in a small-world network, a bout of signalling could spread very rapidly throughout an entire population (see Ch. 12 on information cascades). In summary, small-world analyses suggest that, when studying the pattern of communication within a population, it may be important to identify how asymmetric the communication networks of individuals are. In this context, asymmetry means the number of connections each individual has with other individuals. In a random network, all individuals would have, on average, the same number of connections. However, in a small-world network, the few hub individuals have the majority of connections, while the rest of the population has very few. Asymmetry could be imposed by environmental features (a hub individual could occupy a more central position within a forest or a more open spot from which his signal could propagate further) or by social roles (e.g. male primates connecting female groups; or possibly ‘floating’ juveniles connecting territorial adults in bird species).

Models of communication network dynamics from signallers’ perspectives There are a number of potential costs and benefits associated with signalling in a network. Signallers must compete with each other to make their signals detectable by receivers. They must also balance the benefits of the intended receiver perceiving the signal with the costs of other receivers doing the same. These costs can range from heterospecific predators or parasites to competing conspecifics. Signalling within a network may also coordinate behaviour among individuals within larger social groups. Although many experimental studies have shown the effects of signalling within networks of several individuals, there have been far fewer studies modelling the effects of networks on signals and signalling dynamics (but see Chs. 2, 5 and 13). Here we highlight some of the benefits of

P1: JZZ/... P2: JZZ/... 0521823617c26.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

10:42

Communication networks in a virtual world modelling signalling within groups with three case studies: the chorus behaviour of acoustic insects and frogs, the social coordination of ant foraging, and territory establishment. The case studies share similarities in the methods they employ and the conclusions they draw. In each case, the examples show that complex patterns of behaviour can arise from simple individual decision patterns. Signal dynamics in acoustic choruses

Choruses are of interest when studying behaviour in communication networks from both experimental and theoretical perspectives as they can show tightly coordinated patterns of signalling within large groups (Chs. 12 and 13). Individuals within an acoustic chorus must deal with a complex acoustic background generated by other signallers within which they must maximize the efficacy of their signals in terms of transmission and female attraction. Females can show preferences for specific temporal features of male signals; for example, they may respond less to calls that are overlapped (e.g. gray tree frogs Hyla versicolor: Schwartz et al., 2001) or may prefer leading calls (Snedden & Greenfield, 1998; Greenfield, 2002). Thus, female preferences for certain temporal features may have led to patterns of synchronous and alternating choruses in anuran and insect species. In general, when a species calls at a rapid rate, choruses tend towards synchrony; as the call period becomes longer, choruses are more likely to alternate (Grafe, 1999). Anuran and insect choruses are amenable to experimental studies of network behaviour because the whole network can be controlled and manipulated (Schwartz et al., 2002) and individual behaviour within the network can be measured. This level of experimental control allows the predictions of models to be tested. However, to date, there have been few theoretical considerations of signalling dynamics within choruses. We discuss two different approaches that have been used to model signalling within choruses (Brush & Narins, 1989; Greenfield et al., 1997). Both models consider mechanisms individuals may use to control their call timing and hence avoid interference in a chorus. The mechanism used in each case is a form of inhibitory resetting. Each individual has an internal mechanism that increases from a basal state to a peak where it initiates a call. If, before calling, the individual perceives another individual’s call, the mechanism is reset to its basal level and it begins to increase again. Such mechanisms have been shown to exist for many anuran and insect species (Zelick & Narins, 1985; Greenfield et al., 1997). Brush & Narins (1989) adapted models of computer networks to study whether choruses in the Puerto Rican treefrog Eleutherodactylus coqui were controlled by this inhibitory resetting mechanism. Computer-network models simulate the flow of data between interconnected computers and study how a shared resource (i.e. bandwidth) is partitioned between them. Individual computers are linked to each other via data lines and send packets of information through the network. Before

609

P1: JZZ/... P2: JZZ/... 0521823617c26.xml CU1917B/McGregor

610

0 521 582361 7

April 12, 2005

10:42

A. M. R. Terry & R. Lachlan sending a packet, each computer checks if the data line is clear. If it is not, the computer waits for a random time period before checking the line again (Brush & Narins, 1989). However, even when the line is free, computers have a probability of deferring the transfer. Also in some cases, the terminals must transmit information even if the lines are busy. In the model, frogs represented the terminals; the data lines represented the communication network between them, and their calls were the packets of information. Although their model was an individually based simulation, it was not spatially explicit and, therefore, distance between individuals was not a factor in the analysis. The frogs in this model used their inhibitory resetting to avoid being jammed by other individuals; if they detected another frog calling during their refractory period, they delayed the next call by a randomly chosen period and then returned to the standard refractory period for the next call (Brush & Narins, 1989). Using this model, Brush & Narins (1989) showed that this mechanism would lead to fewer calls being overlapped and that there was an optimal chorus size of between three and four individuals at which information transfer was maximized (i.e. overlap was minimized). The results of the model were also corroborated by field data collected on the treefrog, which showed that choruses occurred in small groups and when group size was large (five or six individuals) males showed selective attention to one or two neighbours. These results are similar to those of Greenfield & Rand (2000) who show that tungara frogs Physalaemus pustulosus paid attention to a subset of the potential signallers in artificially generated choruses. Greenfield et al. (1997) developed a linear model of chorus signalling in the katydid Neoconocephalus spiza, which initially was based on dyadic interactions. Individuals in their model used an inhibitory resetting mechanism similar to that of Brush & Narins (1989). They modelled two individuals signalling at the same time and measured signal overlap from a receiving female’s perspective. They showed that individuals overlapped far less and avoided producing following calls (i.e. calls that were initiated after the onset of another male’s call) when using an inhibitory resetting mechanism, and that signals tended towards synchrony or alternation depending on the speed at which the mechanism reached its peak level (i.e. at call initiation). If males within a given call period could return to their peak level quickly from inhibition (‘rebound’), alternation would arise as males could quickly begin calling again. However, if the rebound took almost as long as the call cycle, males would fall into bouts of synchrony (Greenfield et al., 1997). They also considered the case of a larger number of calling males. This model was an individually based and spatially explicit simulation, with males randomly placed on a 20 m × 20 m grid. Simulated females were also randomly located around this grid; the model assessed the level of call overlapping from the female’s perspective. Male attractiveness was assessed as the number of their calls

P1: JZZ/... P2: JZZ/... 0521823617c26.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

10:42

Communication networks in a virtual world that were overlapped. When they ran simulations for a number of males (2–10), Greenfield et al. (1997) found that males had to pay attention to a subset of the chorus for inhibitory resetting to function as a chorus mechanism, but when they did, patterns of synchrony and alternation emerged. The theoretical analyses of call timing in choruses of signalling anurans and acoustic insects have shown that simple mechanisms that control signal timing, which may have arisen in response to female preferences for leading signals, can generate the complex patterns of alternation and synchrony observed in nature. An emergent feature of these models was that, given the existence of such a resetting mechanism, males could only pay attention to a subset of the individuals in their chorus when determining their call timing. Choruses, in general, would seem to represent a tractable means of studying signalling dynamics in networks. Although the individuals within the chorus are signalling to females, males indirectly compete with each other for acoustic space and directly compete over actual resources, all within the scope of the chorus. Therefore, it is surprising that chorus dynamics have received such limited theoretical attention given that they represent such an amenable study system for simulation models. Swarm intelligence and self-organization in social insects

Social insects are not noted for their individual cognitive abilities, yet they are famous for their ability collectively to ‘solve’ problems of how best to exploit food resources in their environment. Such coordination requires communication between individuals; for example, the honeybee Apis mellifera uses a waggle dance that indicates the location of food sources to other individuals (von Frisch, 1967). Similarly, ants use pheromone trails to lead colony members to food. The collectively adaptive processes that arise out of these interactions are examples of self-organization. Key ‘ingredients’ of self-organization (Bonabeau et al., 1999, p. 9) are positive feedback (e.g. one forager recruits more bees to a food source by dancing) and multiple communicative interactions. The field of foraging strategies in social insects has a rich empirical background. Moreover, several recent models of such behaviour (Deneubourg & Goss, 1989; Deneubourg et al., 1990; Camazine & Sneyd, 1991; Seeley et al., 1991; reviewed by Bonabeau et al., 1999) have shed light on the much simpler underlying individual behaviour patterns. In these models, the type of foraging problem that is faced typically structures the social and communication networks of the population. For example, individuals that are laying a pheromone trail for a food source that is nearby will overlay that trail more frequently than those laying a trail for a food source further away, simply because they move along it more frequently. Such a model can explain the ability of ants preferentially to use the nearest food source first (Resnick, 1994).

611

P1: JZZ/... P2: JZZ/... 0521823617c26.xml CU1917B/McGregor

612

0 521 582361 7

April 12, 2005

10:42

A. M. R. Terry & R. Lachlan Most of these models have implicitly incorporated communication networks. For example, Deneubourg et al. (1990) investigated the ability of ants to choose one of two paths to a food source. Experimental evidence suggests that Argentine ants Linepithema humile quickly decide to use one of the two paths. Deneubourg et al. (1990) fitted this behaviour to a simple positive-feedback model in which the more ants had travelled along a path, the more likely it was to be chosen by another ant. Interestingly, the best fit between model parameters and empirical data occurred when each extra ant travelling along a path had a greater impact on others’ behaviour (e.g. the twentieth ant had a disproportionately larger impact than the first). This is yet another example of asymmetry in networks, where the most recent information has a greater impact than old, and possibly outdated, information. A similar situation is found in the chemical over-marking signals of mice (e.g. Rich & Hurst 1999; Ch. 11). Territory establishment

To compare and contrast the different approaches to modelling behaviour and communication, we use an example from models of territory establishment. Models of territory establishment are predominantly based on game theory, where a series of dyadic encounters occur over indivisible areas. These models contain ‘winner-takes-all’ assumptions, in that whoever wins the contest takes the resource, and contests cannot end in draws or with division of the area. Stamps & Krishnan (1999) developed an individually based spatial simulation model of territory establishment. In their model, individuals moved around a spatially heterogeneous area containing patches of different size. At each time step, they assessed the attractiveness of the patches around them and moved into the one with the highest attractiveness. The attractiveness of an area was based on two key parameters: positive and aggressive experiences. Positive experiences occurred when an individual entered a patch and did not become involved in an aggressive interaction. Positive experiences increased the attractiveness of the patch and thus increased the likelihood that the individual would return to the patch in the future. Aggressive experiences occurred when two individuals entered the same patch. They would then enter into a costly interaction, which, for this model, would end in a draw (i.e. there would be no clear winner). Aggressive encounters would discourage individuals from returning to that patch. The model predicted that individuals would gradually build territories by incorporating novel patches into their home range. Individuals would show periods of sustained aggression when fighting over familiar (i.e. repeatedly visited) sites and they could also take over sites by repeatedly entering into aggressive interactions. The net result of the model was that a pattern of stable territories would be generated through repeated interactions where there was no clear winner taking a resource.

P1: JZZ/... P2: JZZ/... 0521823617c26.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

10:42

Communication networks in a virtual world Game theoretic models make similar predictions to those of Stamps & Krishnan (1999, 2001), albeit for different reasons (Sih & Mateo, 2001). Similar predictions include the increased benefits of being a resident, prolonged encounters between two residents and the more desperate attempts of newcomers to claim territorial space when they have few remaining options. In game theory models, the territory has an intrinsic value, which is greater for residents than newcomers, and the stable strategy that evolves is one of territory choice (i.e. when to stay and when to move on). As with the previous case studies, the models of Stamps & Krishnan (1999, 2001) do not explicitly study networks; however, they contain network-like features, as the behaviour of individuals are affected by their previous encounters with several other individuals. The model could be extended to include some explicit network effects such as eavesdropping (e.g. Ch. 2). For example, an individual’s decision to enter an area could, in part, be based on previous encounters it had observed take place there. This would allow it to discriminate between two novel areas, one hotly contested by other individuals and one that was not. This form of modelling shows that individually based simulation models can create predictions similar to those of game models but without the same restrictive assumptions. As in this case, they can also extend the predictions that the models can make. Summary

In this section, we have chosen examples of communication within networks that are very different. While these studies have not explicitly used the concept of communication networks, they would not be possible without such a viewpoint. They also show how the network may influence a signaller’s behaviour and show that complex patterns of organization and behaviour within networks may be possible through the implementation of simple rules. The territory establishment models also show how the simulation models can be compared with game theoretic models and generate similar predictions while making fewer restrictive assumptions.

Models of communication networks from receivers’ perspectives In the previous section, we emphasized the consequences of communication networks for signalling behaviour and how it influenced population level patterns of signalling and organization. In this section, we examine how receivers can use the network environment to extract information about others and modify their own behaviour. We deal primarily with the role of eavesdropping (Ch. 2) as a means of information gathering. Eavesdropping is one source of information that becomes available to receivers within a network and it is the one that has received

613

P1: JZZ/... P2: JZZ/... 0521823617c26.xml CU1917B/McGregor

614

0 521 582361 7

April 12, 2005

10:42

A. M. R. Terry & R. Lachlan the most empirical study. Here we describe two case studies that reflect the new approaches being taken to the theoretical analysis of communication networks; the game theory simulation model (R. Lachlan & A. M. R. Terry, unpublished data) studies the evolution of eavesdropping as a strategy and the spatial simulation model (R. L. Earley, S. Brosnan & J. Bragg, unpublished data) determines how eavesdropping may shape the formation of linear dominance hierarchies. Both models are spatial simulations and emphasize the point made throughout this chapter: that consideration of the spatial nature of communication is required for the successful modelling of communication networks. In this section, we also discuss how game theory has been used to model communication in groups and why traditional game theory methods may not be best suited to the study of networks. We conclude that significant advances may be made through the combined use of both proximate-based simulation models and strategic decision making through game theory models. Game theory and eavesdropping

Evolutionary game theory encompasses a well-established set of techniques for determining which strategies prove most effective in interactions between individuals. The aim is to establish whether a given strategy can be invaded by any other strategy; if not, it can be called an evolutionarily stable strategy (ESS, Maynard Smith, 1982). There have been several studies using traditional game theory in which individuals use communication networks to predict the behaviour of others. The studies adapted well-established game scenarios and investigated how a given eavesdropping strategy would fare in the game. Johnstone (1998) developed a model of signal detection that aimed to determine whether ‘conspiratorial whispers’ (low-cost and inconspicuous signals (Krebs & Dawkins, 1984)) could be evolutionarily stable in cooperative communication systems. The idea of conspiratorial whispers implicitly recognizes the role of eavesdroppers and hence communication networks. Here eavesdroppers were modelled in the most general sense of the word (i.e. both conspecific and heterospecifics receivers, see Ch. 2) and it was assumed that it was generally costly to be overheard by an eavesdropper. Johnstone’s (1998) model maintained that even when signalling was cooperative, expenditure was required to make the signal detectable for a receiver and that this creates a conflict of interest between signallers and receivers in the face of the costly eavesdroppers. However, in many situations it may not be costly to be overheard. Pollock & Dugatkin (1992) investigated eavesdropping in the famous Prisoner’s Dilemma game. The Prisoner’s Dilemma is an extreme abstraction of many cooperative situations where individuals have the option of either cooperating or defecting in any given round of the game (for a review, see Dugatkin, 1998). Cooperators benefit if they play one another, compared with two defectors

P1: JZZ/... P2: JZZ/... 0521823617c26.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

10:42

Communication networks in a virtual world playing one another. However, if a cooperator plays a defector, the former does badly, while the latter does well (i.e. the defector exploits the cooperator’s beneficence). Therefore, in a single round of the game, individuals should always defect as they will benefit the most; if, however, the game is repeated over a number of rounds, a strategy of reciprocation or ‘tit-for-tat’ becomes most beneficial and stable (Axelrod & Hamilton, 1981; Stephens et al., 2002). Since the publication of the classic paper by Axelrod and Hamilton (1981), many revisions have been published, but versions of tit-for-tat are still regarded as the most successful strategy. Although individuals playing tit-for-tat copy the strategy their opponent played in the previous round, in reality this information may not be present (if they only play each other once) or might be unreliable (if individuals update their strategy frequently). In these situations, individuals could use eavesdropping as a way of obtaining up-to-date information about their opponents’ strategies. Pollock & Dugatkin (1992) found that their so-called ‘observer tit-for-tat’ strategy was sometimes successful when tit-for-tat itself was not an evolutionarily stable strategy (although observer tit-for-tat was out-competed by tit-for-tat under many conditions). Sigmund & Nowak (1998) examined a similar situation and found that indirect reciprocity through ‘image scoring’ (i.e. cooperating with individuals that had a record of cooperation) was a successful strategy. In this case, individuals gain an increase to their image score each time they cooperate and a decrease when they do not (for more details, see Ch. 22). One of the main criticisms of image scoring was that observing an individual’s image score did not take into account the behaviour of that individual’s opponent. Thus, an observer would react the same way to an individual that defected against a notoriously uncooperative opponent as to one that defected against a good cooperator (Leimar & Hammerstein, 2001). To remedy this, Leimar & Hammerstein (2001) investigated the evolutionary stability of the ‘good standing’ strategy (Sugden, 1986), in which individuals strive to maintain their good social standing. Under this strategy, individuals could improve their standing by cooperating and could have it damaged by defecting, but defecting against a player that was uncooperative was not punished. Leimar & Hammerstein (2001) found that this strategy was very successful. Milinski et al. (2001) investigated whether humans used good standing in cooperative games but found that the simpler image scoring tended to be used. They concluded that a strategy using good standing might ask too much of working memory as individuals would have to remember each opponent’s previous interactions and when there were errors in perception of the roles adopted in the interaction, image scoring predominated (Milinski et al., 2001). The models discussed above studied the evolution of cooperation through indirect reciprocity, something which takes place within communication networks (e.g. Ch. 22); however, they do not really examine how individuals interpret

615

P1: JZZ/... P2: JZZ/... 0521823617c26.xml CU1917B/McGregor

616

0 521 582361 7

April 12, 2005

10:42

A. M. R. Terry & R. Lachlan interactions between others, a prerequisite for eavesdropping. In a more explicit model of eavesdropping, Johnstone (2001) adapted the hawk–dove game to assess the use of eavesdropping within such a framework of contests. The hawk–dove model is a simple game that models some of the essential features of animal contests (Maynard Smith, 1982; Riechert, 1998). In this game, individuals can play one of two strategies; aggressive hawks escalate contests into fights while peaceful doves rely on non-aggressive displays. Hawks, therefore, always beat doves but risk damage if they face another hawk. Doves always lose to hawks but suffer no cost to meeting another dove. With a high risk of damage from fights, the evolutionarily stable strategy of this game is a mixture of hawks and doves (Maynard Smith, 1982). In Johnstone’s (2001) model, the success of a third strategy, eavesdropper was investigated in comparison to the pure strategies. The eavesdropper chose to play dove if its opponent had won its previous encounter, and hawk otherwise. To replicate ‘error’ in an eavesdropper’s assessment of the outcome of an interaction, there was a chance that individuals could misinterpret their eavesdropping, for example, by playing hawk against a winner. The model showed that eavesdropper was most common when there was a high cost to fighting. When eavesdropping errors were more frequent, eavesdroppers reached their peak frequency at a lower fighting cost. However, eavesdropping never spread to fixation within the population. Johnstone (2001) suggests that this is because eavesdroppers are unable to assess the strategy that other eavesdroppers will adopt, because it may change each round. Therefore, eavesdroppers are reducing the fitness benefit upon which they are based. When rare, they have the advantage of predicting the correct role in most cases. Consequently, a mixed equilibrium evolves with eavesdroppers at low frequencies. A surprising result of the model was that eavesdropping promoted increased aggression. This is because the model includes a form of ‘winner–loser’ effect: an individual that won in one round is more likely to win in the next as eavesdroppers will chose a submissive role to play. In the models described above, individuals were restricted to obtaining information about only a small part of the interaction between two others. For example, Johnstone’s (2001) model focused on the outcomes of interactions (who won and who lost) and not on the roles adopted in the interactions themselves. This represents one form of eavesdropping (interceptive eavesdropping, see Ch. 2); however, it is likely that real eavesdroppers also consider the roles that the interactants play (who played hawk or dove). Finally, except for the simple viscosity factor of Pollock & Dugatkin (1992), which defined a probability that an individual would have eavesdropped on his opponent’s last contest (Nowak & Sigmund (1998), include a similar factor), the models did not consider how structuring the communication networks might be important. However, as stressed throughout this chapter,

P1: JZZ/... P2: JZZ/... 0521823617c26.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

10:42

Communication networks in a virtual world spatial structuring of populations is likely to have a significant effect on the predictions of theoretical models. A spatial simulation model of the hawk–dove game, in which strategies could evolve in response to all aspects of the interaction, has been developed by R. Lachlan and A. M. R. Terry (unpublished data) to analyse both these factors. In a hawk–dove game between two individuals, A and B, player A could be involved in one of six different types of interaction: 1. 2. 3. 4. 5. 6.

A plays hawk, defeats B, which also plays hawk A plays hawk, loses to B, which also plays hawk A plays dove, defeats B, which also plays dove A plays dove, loses to B, which also plays dove A plays hawk, defeats B, which plays dove A plays dove, loses to B, which plays hawk.

An eavesdropper on the interaction, who would eventually play A, could, therefore, obtain one of six pieces of information about A. The spatial simulation model (R. Lachlan & A. M. R. Terry, unpublished data) examines how eavesdroppers should respond to obtaining one of these pieces of information. The model is a spatially explicit game theoretic simulation: individuals within the population were placed in a 40 × 40 two-dimensional grid (i.e. population size of 1600). In each round of the simulation, individuals engaged in one contest with each of their four neighbours. The strategy adopted by each individual was determined by a vector containing six values (varying between 0 and 1). Each value represented the probability of playing either hawk or dove having just witnessed the opponent in one of the six situations mentioned above. For example an individual with the vector {0.9 0.8 0.2 0 0.7 0.2} would follow an image-scoring strategy as it would tend to play hawk, having observed its opponent play hawk in a previous round (refer to the six types of information listed above). Mortality and reproduction were modelled by having individuals periodically update their strategy by choosing one of their neighbours’ strategies. This choice was determined by the success of individuals: that is, their total payoff after playing the game during the previous round. We investigated two types of inheritance; either the strategy of the most successful neighbour was inherited, or the probability of inheritance was directly proportional to the neighbours’relative success. The difference between these conditions was that only the most successful strategies were rewarded in the first, whereas in the second, moderately successful strategies could also be inherited. A mutation rate was also included, which would create novel strategies in new individuals. Over a range of conditions, a variety of eavesdropping strategies evolved, but only two groups of strategies were found to arise regularly. The first strategy is

617

P1: JZZ/... P2: JZZ/... 0521823617c26.xml CU1917B/McGregor

618

0 521 582361 7

April 12, 2005

10:42

A. M. R. Terry & R. Lachlan

Fig. 26.1. The pattern of aggression in a simulated population of 1600 individuals. Each square represents an individual and the square’s shade of grey represents how many times the individual was attacked over the previous 20 rounds of interaction: attacked every time (black); never attacked (white). (a) The pattern for an image-scoring strategy (individuals copy the hawk or dove behaviour of their opponents) is a similar level of aggression to that of neighbours (the pattern is a fairly even shade of grey) but darker areas show local waves of aggression. (b) The pattern for a reputation-scoring strategy, in which individuals copied the behaviour of their opponent’s opponent (i.e. if an individual was witnessed being attacked, it was subsequently more likely to be attacked), shows that squares are usually either black or white.

somewhat similar to the image-scoring (or observer tit-for-tat) strategy: play hawk if you witness your opponent playing hawk, and dove if he played dove. However, under a wider range of conditions, especially when only the most successful individuals could pass their strategies on, a second group of strategies was most successful. This group consisted of two extreme strategies and a range of intermediates between them. The first extreme corresponded to Johnstone’s (2001) eavesdropper strategy (i.e. play hawk if you eavesdrop on situations 2, 4 or 6; play dove otherwise, see Fig. 26.1a). The more common extreme (which we call reputation scoring), however, consisted of a novel strategy in which individuals essentially copied the behaviour of their opponent’s opponent (i.e. play hawk in situations 1, 2 and 6; play dove otherwise). An anthropomorphism of this strategy would be that individuals paid attention to an individual’s reputation rather than its deeds. The reputationscoring strategy was successful because it tended to lead to neighbours ‘ganging up’ on the same individuals (Fig. 26.1b); as a result, some individuals were highly successful, and others were very unsuccessful (Fig. 26.2). The overall mean level of success for the image-scoring and reputation-scoring strategies was actually rather

P1: JZZ/... P2: JZZ/... 0521823617c26.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

10:42

Communication networks in a virtual world

(a) 900

Frequency

600

300

0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7

3

Success (arbitrary units)

(b)

Frequency

900

600

300

0 0.3 0.6

0.9

1.2

1.5 1.8 2.1

2.4 2.7

3

Success (arbitrary units)

Fig. 26.2. The distribution of individual success after five rounds of interaction (five lines on graph). Ten runs of a simulation of 1600 individuals were carried out for each case. (a) The distribution for an image-scoring strategy shows one sharp peak, indicating that most individuals have a similar, moderate level of success. (b) The distribution for a reputation-scoring strategy has two peaks, indicating that individuals were either successful or unsuccessful.

similar (1.78 versus 1.73), but the higher frequency of very successful individuals means that the reputation-scoring strategy out-competes the image-scoring strategy if the most successful individuals within a local neighbourhood monopolize their success at replication (Fig. 26.2). The reputation-scoring strategy is the closest

619

P1: JZZ/... P2: JZZ/... 0521823617c26.xml CU1917B/McGregor

620

0 521 582361 7

April 12, 2005

10:42

A. M. R. Terry & R. Lachlan approximation to social eavesdropping (Ch. 2), because it requires eavesdroppers to pay attention to the interaction and the role of each individual. The model’s combination of a spatial simulation limiting individuals to information gathered from the interactions going on around them, with a cost–benefit approach of determining the optimal strategy to follow should provide a deeper understanding of the evolution of eavesdropping as a strategy for gathering information. Network effects on the formation of linear dominance hierarchies

The formation of dominance hierarchies represents another area where information from interactions between individuals can be used by observers to influence future encounters (e.g. Dugatkin, 2001). A linear dominance hierarchy is defined by the number of component triads (trios) within the group that form transitive relationships (i.e. if A beats B, and B beats C, then A also beats C); as the number of transitive relationships increases, linearity increases. Here we consider a simulation model of the effects of communication networks on the formation of linear dominance hierarchies (R. L. Earley, S. Brosnan & J. Bragg, unpublished data). Linear dominance hierarchies are established via overt aggressive interactions and their establishment leads to the unequal distribution of resources among dominant and subordinate individuals. Hierarchies also cause a general decrease in the overall aggression levels within the group. Dominance hierarchy formation has been studied in a wide range of taxa; however, the factors involved in their formation remain contentious. Conceptual models attribute the formation of linear hierarchies either to some aspect related to fighting ability (e.g. Slater, 1986; Jackson & Winnegrad, 1988) or to social effects such as winner effects, loser effects and eavesdropping (e.g. Chase, 1980; Bonabeau et al., 1996; Dugatkin, 1997, 2001). The individual-based spatially explicit simulation model is being developed (R. L. Earley, S. Brosnan & J. Bragg, unpublished data) to study how social eavesdropping may influence the dynamics of hierarchy formation in groups of virtual animals. In each simulation, a group of 10 individuals are allowed to interact for a predetermined period of time. At each time step, one individual can initiate an interaction with another and, if the other individual responds an aggressive interaction begins. Individuals can interact through displays or they can escalate the contest to fighting. At the conclusion of the contest, individuals update their estimate of their own fighting ability. These updates mimic the winner and loser effects where dominant animals increase and subordinate animals decrease their perception of their own fighting ability (Hsu & Wolf, 2001). A certain proportion of individuals close to the interaction (eavesdroppers) can observe it and, in consequence, update their estimates of the interactants’ fighting abilities. The estimate

P1: JZZ/... P2: JZZ/... 0521823617c26.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

10:42

Communication networks in a virtual world of an opponent’sfighting ability determines whether an individual will initiate an interaction and to what extent it will be pursued. The outcome of the simulation is a dominance matrix and the degree of linearity is determined using Landau’s index (0 < h < 1; Landau, 1951). If Landau’s h > 0.9, the hierarchy is considered to be linear (Chase, 1974). In an initial assessment of the model, the social effects were kept symmetrical (i.e. winner effects = loser effects); there was no initial variation in fighting abilities, and all individuals could eavesdrop on each interaction within their network. When social factors were excluded from the model, display and escalated interactions occurred with equal frequency and non-linear hierarchies emerged. Also when winner or loser effects operated alone, the linearity index remained low and did not increase greatly when the magnitude of the winner/loser effects was increased. However, the model showed that eavesdropping, when it was included, acted to increase the estimates of the fighting abilities of winning individuals and to decrease those of losers. When included with winner–loser effects, eavesdropping caused the formation of strongly linear hierarchies. The most important factor was the inflation of the winner’s estimated fighting ability (decreasing the loser’s estimate had less effect). As with our model of a spatial hawk–dove game, this model provided a simulation of social eavesdropping (Ch. 2) as individuals paid attention to both the interactants. The resulting modification of the estimates of fighting ability depended not only on each of the interactants but also on the relative differences in fighting ability when the two individuals met in the interaction. This model is a first attempt to study the implications of networks in the formation of dominance hierarchies. Future studies could investigate the relationship between asymmetries in both fighting abilities and eavesdropping in promoting or hindering linear dominance hierarchies. Summary

The models detailed here have provided the first theoretical studies of the role of eavesdropping in communication networks. Eavesdropping is one of the potential sources of information available to receivers and has received the most experimental attention (McGregor & Peake, 2000, Ch. 2). Experimental data from different taxa have shown that individuals can and do pay attention to the interactions of conspecifics and that these interactions will modify the behaviour of individuals in future encounters (e.g. Oliveira et al., 1998; Peake et al., 2001, 2002). It is likely that eavesdropping is a common behaviour. Although originally thought to be a cost-free source of information (McGregor, 1993), it is likely that eavesdropping has costs associated with the partitioning of cognitive processes required to follow interactions and the fact that individuals may have to abstain from performing other behaviours to witness interactions. Initial theoretical treatments

621

P1: JZZ/... P2: JZZ/... 0521823617c26.xml CU1917B/McGregor

622

0 521 582361 7

April 12, 2005

10:42

A. M. R. Terry & R. Lachlan of eavesdropping showed that it was only stable as a minority strategy because eavesdroppers could not predict how other eavesdroppers would behave and, as a result, it would lead to more aggressive encounters (Johnstone, 2001). However, when eavesdroppers were able to follow the interactions of individuals around them, as opposed to being randomly drawn into interactions, it became apparent that eavesdropping was a viable strategy (R. Lachlan & A. M. R. Terry, unpublished data). Further analyses will address how individuals balance the costs and benefits of eavesdropping against other sources of information or other behaviours.

Summary When communication is considered to occur within a network, new possibilities emerge for individuals to broadcast and receive information that will affect their behaviour in future interactions. Networks also extend the consideration of the costs and benefits of signalling to include other signallers and receivers that may or may not be apparent to the respective interactants. Current models of animal communication have been dominated by game theory, which is well suited to the analysis of strategies used by individuals in dyadic encounters. However, when applied to networks, pure game theory models allow only a superficial analysis of the costs and benefits of signalling. While consideration of the strategic nature of communication in networks remains important, we feel that it must be combined with more process-based approaches that place fewer restrictions on individual behaviour. Individually based simulation models are becoming increasingly common in behavioural ecology and a combination of these more proximate level models with game theory approaches will give a greater understanding of the evolution of communication networks. In particular, in the examples we have described (anuran acoustic choruses and eavesdropping strategies), spatially realistic individually based models have generated different results from more traditional techniques applied to the same question. In this chapter, we have emphasized the importance of studying realistically structured populations and have identified the spatial nature of communication in networks as an important feature of any theoretical consideration. The importance of the spatial nature of networks may also extend to network organization. There are likely to be asymmetries in the extent to which individuals contribute to the flow of information in a network: some individuals providing more information than others. One future avenue for both theoretical and experimental research may be to determine how communication networks are organized and whether hub individuals act as routers through which most information flows. The nature of the flow of information through networks, whether it is a signal spreading through a chorus or an alarm

P1: JZZ/... P2: JZZ/... 0521823617c26.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

10:42

Communication networks in a virtual world call spreading through a population, will undoubtedly have implications for the evolution of signalling within that system. Although there is a wide literature on the theoretical aspects of strategic decisions when communicating and cooperating in groups, there is little experimental evidence to support these models (for exceptions see Wedekind & Milinski, 2000; Milinski, et al., 2001; Stephens et al., 2002). Work on cooperation in groups has shown that modelled strategies may not consider the complex combination of assumptions, working rules and limitations that individuals face when deciding how to respond in interactions with known individuals (Milinski et al., 2001; Stephens et al., 2002). We suggest that future models incorporating both strategic and simulation aspects will be able to model more closely the dynamics involved in observing and taking part in repeated interactions, and this, in turn, will lead to a better understanding of the strategies and behaviours that individuals employ when communicating in networks.

Acknowledgements We are very grateful to Ryan Earley for allowing us to discuss and present his model. We would also like to thank several people whose comments helped make this chapter more readable; Ryan Earley, Ricardo Matos, Tom Peake and Denise Pope. A. T. was funded by the Zoological Institute at the University of Copenhagen during the writing of this chapter.

References Albert, R. & Barabasi, A. L. 2002. Statistical mechanics of complex networks. Reviews of Modern Physics, 74, 47–97. Albert, R., Jeong, H. & Barabasi, A. L. 1999. Diameter of the World-Wide Web. Nature, 401, 130–131. Axelrod, R. & Hamilton, W. D. 1981. The evolution of cooperation. Science, 211, 1390–1396. Barabasi, A. L. & Albert, R. 1999. Emergence of scaling in random networks. Science, 286, 509–512. Begon, M., Harper, J. L. & Townsend, C. R. 1990. Ecology: Individuals, Populations and Communities, 2nd edn. Oxford: Blackwell. Bonabeau, E., Theraulaz, G. & Deneubourg, J.-L. 1996. Mathematical model of self-organizing hierarchies in animal societies. Bulletin of Mathematical Biology, 58, 661–717. Bonabeau, E., Dorigo, M. & Theraulaz, G. 1999. Swarm Intelligence: From Natural to Artificial Systems. Oxford: Oxford University Press.

623

P1: JZZ/... P2: JZZ/... 0521823617c26.xml CU1917B/McGregor

624

0 521 582361 7

April 12, 2005

10:42

A. M. R. Terry & R. Lachlan Brush, J. S. & Narins, P. M. 1989. Chorus dynamics of a neotropical amphibian assemblage: comparison of computer-simulation and natural behavior. Animal Behaviour, 37, 33–44. Camazine, S. & Sneyd, J. 1991. A model of collective nectar source selection by honey-bees: self-organization through simple rules. Journal of Theoretical Biology, 149, 547–571. Chase, I. D. 1974. Models of hierarchy formation in animal societies. Behavioral Science, 19, 374–382. 1980. Social process and hierarchy formation in small groups: a comparative perspective. American Sociology Review, 45, 905–924. Deneubourg, J. L. & Goss, S. 1989. Collective patterns and decision making. Ethology, Ecology and Evolution, 1, 295–311. Deneubourg, J. L., Aron, S., Goss, S. & Pasteels, J. M. 1990. The self-organizing exploratory patterns of the Argentine ant. Journal of Insect Behaviour, 3, 159–168. Dieckman, U. C. 1997. Can adaptive dynamics invade? Trends in Ecology and Evolution, 12, 128–131. Dugatkin, L. A. 1997. Winner and loser effects and the structure of dominance hierarchies. Behavioral Ecology, 8, 583–587. 1998. Game theory and cooperation. In: Game Theory and Animal Behavior, ed. L. A. Dugatkin & H. K. Reeve. Oxford: Oxford University Press, pp. 38–63. 2001. Bystander effects and the structure of dominance hierarchies. Behavioral Ecology, 12, 348–352. Dunne, J. A., Williams, R. J. & Martinez, N. D. 2002. Food-web structure and network theory: the role of connectance and size. Proceedings of the National Academy of Sciences, USA, 99, 12917–12922. Grafe, T. U. 1999. A function of synchronous chorusing and a novel female preference shift in an anuran. Proceedings of the Royal Society of London, Series B, 266, 2331–2336. Greenfield, M. D. 2002. Signalers and Receivers: Mechanisms and Evolution of Arthropod Communication. Oxford: Oxford University Press. Greenfield, M. D. & Rand, A. S. 2000. Frogs have rules: selective attention algorithms regulate chorusing in Physalaemus pustulosus (Leptodactylidae). Ethology, 106, 331–347. Greenfield, M. D., Tourtellot, M. K. & Snedden, W. A. 1997. Precedence effects and the evolution of chorusing. Proceedings of the Royal Society of London, Series B, 264, 1355–1361. Grimm, V. 1999. Ten years of individual-based modelling in ecology: what have we learned and what could we learn in the future? Ecological Modelling, 115, 129–148. Hamilton, W. D. 1971. Geometry of the selfish herd. Journal of Theoretical Biology, 31, 295–311. Hemelrijk, C. K. 2000. Towards the integration of social dominance and spatial structure. Animal Behaviour, 59, 1035–1048. 2002. Understanding social behaviour with the help of complexity science. [Invited article.] Ethology, 108, 655–671.

P1: JZZ/... P2: JZZ/... 0521823617c26.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

10:42

Communication networks in a virtual world Hsu, Y. Y. & Wolf, L. L. 2001. The winner and loser effect: what fighting behaviours are influenced? Animal Behaviour, 61, 777–786. Jackson, W. M. & Winnegrad, R. L. 1988. Linearity in dominance hierarchies: a second look at the individual attributes model. Animal Behaviour, 36, 1237–1240. Jeong, H., Tombor, B., Albert, R., Oltval, Z. N. & Barabasi, A. L. 2000. The large-scale organization of metabolic networks. Nature, 407, 651–654. Johnstone, R. A. 1998. Conspiratorial whispers and conspicuous displays: games of signal detection. Evolution, 52, 1554–1563. 2001. Eavesdropping and animal conflict. Proceedings of the National Academy of Sciences, USA, 98, 9177–9180. Krebs, J. R. & Dawkins, R. 1984. Animal signals: mind-reading and manipulation. In: Behavioural Ecology: An Evolutionary Approach, 2nd edn, ed. J. R. Krebs & N. B. Davies. Oxford: Blackwell, pp. 380–402. Kudo, H. & Dunbar, R. I. M. 2001. Neocortex size and social network size in primates. Animal Behaviour, 62, 711–722. Landau, H. G. 1951. On dominance relations and the structure of animal societies: I. Effect of inherent characteristics. Bulletin of Mathematical Biophysics, 13, 1–19. Leimar, O. & Hammerstein, P. 2001 Evolution of cooperation through indirect reciprocity. Proceedings of the Royal Society of London, Series B, 268, 745–753. Maynard Smith, J. 1982. Evolution and the Theory of Games. Cambridge, UK: Cambridge University Press. McGregor, P. K. 1993. Signalling in territorial systems: a context for individual identification, ranging and eavesdropping. Philosophical Transactions of the Royal Society, Series B, 340, 237–244. McGregor, P. K. & Dabelsteen, T. 1996. Communication networks. In: Ecology and Evolution of Acoustic Communication in Birds, ed. D. E. Kroodsma & E. H. Miller. Ithaca, NY: Cornell University Press, pp. 409–425. McGregor, P. K. & Peake, T. M. 2000. Communication networks: social environments for receiving and signalling behaviour. Acta Ethologica, 2, 71–81. Milinski, M., Semmann, D., Bakker, T. C. M. & Krambeck, H.-J. 2001. Cooperation through indirect reciprocity: image scoring or standing strategy? Proceedings of the Royal Society of London, Series B, 268, 2495–2501. Morton, T. L., Haefner, J. W., Nugala, V. & Decino, R. D. 1994. The selfish herd revisited: do simple movement rules reduce relative predation risk? Journal of Theoretical Biology, 167, 73–79. Nowak, M. A. & Sigmund, K. 1998. Evolution of indirect reciprocity by image scoring. Nature, 393, 573–577. Oliveira, R. F., McGregor, P. K. & Latruffe, C. 1998. Know thine enemy: fighting fish gather information from observing conspecific interactions. Proceedings of the Royal Society of London, Series B, 265, 1045–1049. Peake, T. M., Terry, A. M. R., McGregor, P. K. & Dabelsteen, T. 2001. Male great tits eavesdrop on simulated male-to-male vocal interactions. Proceedings of the Royal Society of London, Series B, 268, 1183–1187.

625

P1: JZZ/... P2: JZZ/... 0521823617c26.xml CU1917B/McGregor

626

0 521 582361 7

April 12, 2005

10:42

A. M. R. Terry & R. Lachlan 2002. Do great tits assess rivals by combining information gathered by eavesdropping? Proceedings of the Royal Society of London, Series B, 269, 1925–1929. Pollock, G. & Dugatkin, L. A. 1992. Reciprocity and the evolution of reputation. Journal of Theoretical Biology, 159, 25–37. Resnick, M. 1994. Turtles, Termites and Traffic Jams. Cambridge, MA: MIT Press. Rich, T. J. & Hurst, J. L. 1999. The competing countermarks hypothesis: reliable assessment of competitive ability by potential mates. Animal Behaviour, 58, 1027–1037. Riechert, S. E. 1998. Game theory and animal contests. In: Game Theory and Animal Behaviour, ed. L. A. Dugatkin & H. K. Reeve. Oxford: Oxford University Press, pp. 64–93. Schwartz, J. J., Buchanan, B. W. & Gerhardt, H. C. 2001. Female mate choice in the gray treefrog (Hyla versicolor) in three experimental environments. Behavioral Ecology and Sociobiology, 49, 443–455. 2002. Acoustic interactions among male gray treefrogs, Hyla versicolor, in a chorus setting. Behavioral Ecology and Sociobiology, 53, 9–19. Seeley, T. D., Camazine, S. & Sneyd, J. 1991. Collective decision making in honey-bees: how colonies choose among nectar sources. Behavioral Ecology and Sociobiology, 28, 277–290. Sigmund, K. & Nowak, M. A. 1998. The complexity of cooperation: agent-based models of competition and collaboration. Nature, 392, 457–457. Sih, A. & Mateo, J. 2001. Punishment and persistence pay: a new model of territory establishment and space use. Trends in Ecology and Evolution, 16, 477–479. Slater, P. J. B. 1986. Individual differences and dominance hierarchies. Animal Behaviour, 34, 1264–1265. Snedden, W. A. & Greenfield, M. D. 1998. Females prefer leading males: relative call timing and sexual selection in katydid choruses. Animal Behaviour, 56, 1091–1098. Stamps, J. A. & Krishnan, V. V. 1999. A learning-based model of territory establishment. Quarterly Review of Biology, 74, 291–318. 2001. How territorial animals compete for divisible space: a learning-based model with unequal competitors. American Naturalist, 157, 154–169. Stephens, D. W., McLinn, C. M. & Stevens J. R. 2002. Discounting and reciprocity in an iterated prisoner’s dilemma. Science, 298, 2216–2218. Sugden, R. 1986. The Economics of Rights, Co-operation and Welfare. Oxford: Blackwell. Turelli, M. & Barton, N. H. 1994. Genetic and statistical analyses of strong selection and polygenic traits: what, me normal? Genetics, 138, 913–941. Viscido, S. V. & Wethey, D. S. 2002. Quantitative analysis of fiddler crab flock movement: evidence for ‘selfish herd’ behaviour. Animal Behaviour, 63, 735–741. von Frisch, K. 1967. The Dance Language and Orientation of Bees. Cambridge, MA: Harvard University Press. Watts, D. J. & Strogatz, S. H. 1998. Collective dynamics of ‘small-world’ networks. Nature, 393, 440–442. Wedekind, C. & Milinski, M. 2000. Cooperation through image scoring in humans. Science, 288, 850–852.

P1: JZZ/... P2: JZZ/... 0521823617c26.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

10:42

Communication networks in a virtual world Williams, R. J. & Martinez, N. D. 2000. Simple rules yield complex food webs. Nature, 404, 180–183. Wilson, W. 2000. Simulating Ecological and Evolutionary Systems in C. Cambridge, UK: Cambridge University Press. Zelick, R. & Narins, P. M. 1985. Characterization of the advertisement call oscillator in the frog Eleutherodactylus coqui. Journal of Comparative Physiology A: Sensory Neural and Behavioral Physiology, 156, 223–229.

627

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:14

Index

Abudefduf vaigiensis see sergeant major fish Acheta domesticus see house cricket Accipiter nisus see sparrowhawk acoustic choruses, models of signalling dynamics 609–611 see also anuran choruses; insect choruses acoustic communication caller identity and status information 381

experience 383–384 African grey parrot (Psittacus erithacus), cognitive and communicative capacities 569, 572–573 Agelaius phoeniceus see red-winged blackbird Agelenopsis aptera see funnel-web spider aggression 2

factors affecting evolution of 152–153

and social instability in fishes 96

problem of masking interference 157–162

androgens and 481–482

see also signal production; sound

in models of territory establishment

transmission Acrocephalus arundinaceus see great reed warbler advertising signals, to facilitate eavesdropping 48–49, 52 African catfish (Clarias gariepinus), semiochemicals 558–559 African elephant (Loxodonta africana) benefits of attending to others’ social calls 384 contact call discrimination and memory 383–384 fluid social systems and long-distance signalling 377–378 group social knowledge and age of matriarch 383–384 individuality in fundamental frequency contour in vocalizations 382 infrasound communication 457–459 intelligible distance of calls 372, 382–383

628

social knowledge related to age and

612–613 modelling linear dominance hierarchy formation 620–621 physiological costs 191–192 song overlapping in songbirds 304, 305–306 victory displays 11 see also redirected aggression aggressive calls, variety in male anurans 279–280 aggressive priming, audience effects 75–78 aggressive signal exchanges see eavesdropping; signalling interactions agonistic contests see aggression Alcelaphus buselaphus see hartebeeste allomones see fish semiochemicals Alouatta pigra see black howler monkey alpine accentor (Prunella collaris), quiet singing 53

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:14

Index altruism and cooperation in communication networks 446 audience effects 30–31 dishonest signals used for tactical deception 524–525 eavesdropping and indirect reciprocity 522–523 evolution and maintenance 536 functional rather than cognitive approach 522–523 image scoring and evolution of 533–534 reciprocal altruism and behavioural strategies 521–522 Alytes spp. see midwife toads Amolops tormotus see Chinese frog androgens adaptive value of social modulation 497, 502–505 and aggressive behaviour 481–482, 496–498, 499 and bystander priming response 499–500, 501 and electrocommunication signals 487–488 and pheromone production and/or release 485–486 and sex differences in spatial memory 490–491 brain receptors 490–491 costs and benefits of high levels 497, 502–505 effects of early exposure (critical period) 490–491 effects of high levels on male parental care 497, 504–505

effects on visual displays in vertebrates 486–487 effects on vocal structures of amphibians 484–485 effects on vocal structures of birds 483–484 effects on vocal structures of toadfish 484 interactive effects with social environment 482 interrelationship with associative learning mechanisms 505 levels during social instability 492–494 levels in dominant and subordinate males 492–494, 496–498, 499 modulation of behaviour in communication networks 494–502 modulation of central mechanisms affecting motivation 490 modulation of sensory perception 488–489 possible mediation of audience effects 501 possible mediation of dear enemy effects 501–502 role in winner–loser effects 198, 492–494, 496–498, 499 social modulation of androgen levels 492–494 social modulation of behavioural effects 481–482 stimulation by social interactions 492–494 Anolis aeneus see lizard anonymity, used to counter eavesdropping 56–57 Anser anser see greylag goose antbirds (Thamnophilidae), sound characteristics 49 Antilocapra americana see pronghorn antelope anuran amphibians

effects of population density 492–494

communication networks 2, 248

effects on cognitive functions 490–491

effects of androgens 484–485, 486–487

effects on expression of somatic releasers

use of private signalling 291

491–492 effects on pheromone production in urodeles 485–486 effects on singing behaviour of songbirds 483–485

victory displays 118, 122 anuran choruses 263–264 adaptations for acoustic competition 278–279 alternation of signalling 280–282

629

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

630

0 521 582361 7

April 12, 2005

11:14

Index anuran choruses (cont.)

Aptenodytes forsteri see emperor penguin

auditory systems and background noise 292

Aptenodytes patagonicus see king penguin

call-timing, evolution and maintenance

Apternotus leptorhynchus see brown ghost

282–284 conflicting demands of signalling 279–280 entrained calling 280–282 features of network communications 291–292

Argentine ant (Linepithema humile), foraging strategies 611–612 arthropods, victory displays 117–118, 122 associative learning

female acoustic responses 283

and tactical deception 524–525

female attraction to larger choruses 283

interrelationship with androgen effects

female call preferences 264, 283–284, 285–286, 292 female choices in trials and natural choruses 287–289 female cognitive requirements for social eavesdropping 289 female eavesdropping on male–male interactions 289

505 Atlantic salmon (Salmo salar), androgen effects on olfactory sensitivity 488–489 audience effects 10, 30–31 and altruism 522–523 and eavesdropping 66 and victory displays 122–123 clarification of terminology 66–67, 68

female mate sampling techniques 288–289

definition 65–66

female sensory abilities and signal

human behaviour 74–75

discrimination 285

in redirected aggression 209–210

fine-scale patterns of signal timing 280–283

male–male aggressive signalling 68–72

graded or discrete calls 279

male parental behaviour 72–74

leading and lagging roles 280–282

on cheating behaviour 529–530, 531

male aggressive calls 279–280

on signalling 65–66

male calling energy costs 284–285

possible influences on scent marking

male interceptive eavesdropping 289–291

362–363

male–male vocal competition 278–279

possible mediation by androgens 501

males spatial distribution and selective

pre-exposure and male aggression 75–78

attention 284 models of signalling dynamics 609–610, 611

review of evidence for 67–75 audiences

network view 277–278

and signal evolution 79

‘off response’ call initiation 280–282

apparent 65

repertoire size of signallers 279

definition of 64

selection pressures on signals 277, 283–284

evolutionary 64–65

signal overlap avoidance in males 280–282

terminology 66–67, 68

suitability for studying communications networks 292–293

see also eavesdropping auditory systems

synchronous calling 280, 281, 282

analysis by receivers 471–474

two-part advertising call 279–280

coping with background noise 292

use of both ‘on-response’ and ‘off-response’ calling 282–283 Aphelocoma californica see western scrub jay Apis mellifera see honeybee

hormonal modulation of sensitivity 489 Australian bushcricket (Elephantodeta nobilis), evidence for eavesdropping 291 autocommunication, and eavesdropping 18

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:14

Index avian cognition, research history and prejudices 568–570

song repertoire 326 territorial conflict and song matching 333–335

baboon (Papio cynocephalus) knowledge of other animals 585–586, 587, 589 social structure and relationship recognition 584 background noise and functioning of auditory systems 292 effects of lunar phase 158, 163–164 see also masking interference; masking release Baktaman people (New Guinea) 421 Balaenoptera musculus see blue whale Balaenoptera physalus see fin whale banded wren (Thryothorus pleurostictus) behaviour after dawn chorus 327 bout structure during and after dawn chorus 329–331 dawn chorus as an interactive network 322, 337 dawn chorus features 326–327 dawn chorus singing and male quality 338

territorial behaviour 325–326, 327 variation in dawn chorus singing of individual males 336 barking treefrog (Hyla gratiosa), female assessment of males 137, 288–289 barn owl (Tyto alba) nestling interactions 176 sound localization ability 465–467 barnacle goose (Branta leucopsis), victory display 115–116 bats (various species) attraction to echolocation signals 405 predation on katydids 154, 156–157 bearded seal (Erignathus berbatus), distinctive group calls 399–400 behaviour, reciprocal link with hormones 481–482 beluga whale (Delphinapterus leucas) avoidance of ice-breakers 406–407 reactions to killer whale sounds 403–404 victory display 118

dawn chorus structure 336

Betta splendens see Siamese fighting fish

daytime song-delivery patterns 326

binaural processing, and acoustic signal

male interactions at dawn chorus 335, 337–338 movement patterns around dawn chorus 330, 332 multiple song matching 335 overlapping and matching songs 331, 334, 335–336

masking release 462–463 birds effects of androgens on vocal structures 483–484 equating human and avian cognitive studies 573–574 signal distance assessment 467–468

recording methods 327–329, 330, 339–340

victory displays 115–117

singing behaviour during and after dawn

visual displays and effects of androgens 486

chorus 323, 329–336 song behaviour 325–327 song matching as indicator of conflict 333–335 song matching during and after dawn chorus 330, 332–335 song overlapping as an aggressive signal 331, 334, 335–336 song rates during dawn chorus 330, 331

see also songbirds; individual species black-capped chickadee (Poecile atricapillus) dawn chorus and male condition 141, 338 eavesdropping and transitive inference (TI) 574–576 female mate choice 140–141, 142 habitat change and song networks 143–144 habitat quality and song output 144 ranking process 577

631

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

632

0 521 582361 7

April 12, 2005

11:14

Index black-capped chickadee (cont.) signal amplitude as a distance cue 468–469 use of social eavesdropping 21–22 black goby (Gobius niger)

recognition of others’ social relationships 591 silence when captured or near a boat 403–404

male sex pheromone 549–558

victory display 118

semiochemical communication 558–559

vocal learning and individual whistle

black howler monkey (Alouatta pigra), factors affecting levels of vigilance 419 black swan (Cygnus atratus), victory display 115–116 black-throated diver (Gavia immer), victory

development 400–402 vocal matching to signal a specific individual 402 boubou (Laniarius aethiopicus), victory display 116

display 116–117

Branta canadensis see Canada goose

blackbird (Turdus merula)

Branta leucopsis see barnacle goose

courtship interruption by neighbours 41

Brienomyrus brachyistius see electric fish

dawn chorus singing 337

Broadley’s painted reed frog (Hyperolius

male quality and dawn chorus singing 338 sound characteristics and attenuation 49–50 use of high perches 43, 44, 464–465 use of quiet song 50, 53–55, 56 blackcap (Sylvia atricapilla) sound characteristics and attenuation 49 use of high perches 43, 44 blenny semiochemical communication 558–559 blue-throated humming bird (Lampornis clemenciae), Lombard effect 464–465 blue tit (Parus caeruleus), dawn chorus singing and male quality 338 blue whale (Balaenoptera musculus), distinctive group calls 400 bonnet macaque see macaques Boophis madagascariensis see Madagascar treefrog bottlenose dolphin (Tursiops truncatus) active space of signals 392–394, 395 avoidance of feeding grounds with boat noise 406–407 eavesdropping on other’s echolocation clicks 18, 405–406 fission–fusion societies 400–402 individual signature whistle types 392–395, 400–402 numbers of animals in networks 397–399

marmoratus broadleyi) female call preferences 286, 287–288 ‘off-response’ call initiation 280–282 simultaneous mate choice in females 289 brown capuchin monkey (Cebus apella) demands of social monitoring 32 predation levels and vigilance 417–418 time spent looking by subordinates 417–418 brown ghost (Apternotus leptorhynchus), androgen effects on electric signals 487–488 brown-headed cowbird (Molothrus ater), nestling begging 177 budgerigar (Melopsittacus undulatus) male parental behaviour 73–74 masking release and binaural processing 462–463 signal-to-noise ratios for recognition and for detection 463–464 sound localization ability 465–467 bystanders and redirected aggression 192–194, 200, 201–203 costs and benefits of information gathering 131–132 distinction from eavesdroppers 84–85

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:14

Index effects on dominance hierarchies 97–99 influence of eavesdropping on behaviour 86–91, 93 influence of previous experience on behaviour 88–89 postconflict attacks on losers 196–197 priming response and androgens 499–500, 501 social eavesdropping 10–11 taking advantage of ‘loser effects’ 198–200 see also eavesdroppers

chaffinch (Fringilla coelebs), signal distance assessment 467–468 chemical communication correlation with predation 545–546 evolution through different functional phases 541–542, 543 predation risk assessment 544–549 scent marking 131 signal propagation 543–544 see also fish semiochemicals; scent marking chimpanzee (Pan troglodytes) choice of targets for aggression 208

Canis latrans see coyote Canis lupus see wolf Campbell’s monkey (Cercopithecus campbelli), attention to alarm calls by another species 373, 380 Canada goose (Branta canadensis), victory display 115–116 canary (Serinus canarius), sound localization ability 465–467

fluid social systems and long-distance signalling 377 situations where calling is suppressed 374–375, 376–377 Chinese frog (Amolops tormotus) acoustic signal repertoire 279 chorusing interactions and female preference for leading signals 264

Capreolus capreolus see roe deer

and signal competition 263–264

Carassius auratus see goldfish

in anurans and insects 263–264

Carassius carassius see Crucian carp

precedence effect 264

cat, feral ( Felis catus), victory display 118–119,

to avoid signal masking 264

121–122

see also anuran choruses

Cebus apella see brown capuchin monkey

Cichlasoma nigrofasciatum see convict cichlid

Cebus capucinus see white-faced capuchin

cichlid see Mozambique tilapia

monkey

Cistothorus palustris see marsh wren

Cephalorhyncus hectori see Hector’s dolphin

Clamator glandarius see great spotted cuckoo

Cercocebus albigena see mangabey

Clarias gariepinus see African catfish

Cercopithecus aethiops see vervet monkey

claw waving see fiddler crabs

Cercopithecus ascanius schmidtii see redtail

cleaner wrasse (Labroides dimidiatus)

monkey

behaviour towards clients 534

Cercopithecus campbelli see Campbell’s monkey

cheating behaviour 525–526

Cercopithecus diana see Diana monkey

client image scoring and wrasse behaviour

Cervus elaphus see red deer cetaceans, vocal matching to signal a specific individual 402 chacma baboon (Papio cynocephalus ursinus) attending to signal interactions 373–374 awareness of dominance relationships 199 time spent visually scanning 417–418

526–528, 529–530, 531 cognitive abilities used in tactical deception 535, 536 effects of bystander types on cheating behaviour 529–530, 531 interactions with client reef fish 525–526 mutualism with client reef fish 521 possible endocrine-mediated response 535

633

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

634

0 521 582361 7

April 12, 2005

11:14

Index cleaner wrasse (cont.) preferred food 534 response of biting cleaners to image scoring clients 531–532

concept of true individual recognition 363–366 context effects 1, 78–79, 129–132 cooperation and altruism 446, 536

tactile stimulation 531–532

defining properties 13

variable payoffs from different clients

effects on signalling and receiving 180–185

533–534 variations in cooperative and cheating behaviour 526–527, 528 cognition in animals, research history and prejudices 568–570 cognitive aspects of communication networks

evolution of spiteful behaviour 536 evolution of tactical deception 536 evolutionary process 558–559 game theory models 536 habitat alteration effects 143–144 hormones and communication 494–502

250–251

impact on signalling errors 184–185

cognitive capacity

implications for theoretical study of

and cortical size 569

signalling strategies 604

of birds 568–570

individual recognition mechanism 363–366

requirements for tactical deception 524–525

information cascades 270, 607–608

cognitive sciences, interface with communication networks 447 common carp (Cyprinus carpio), pheromone system 552 common dolphin (Delphinus delphis), individual signature signals 400–402 common seal (Phoca vitulina) avoidance of killer whale sounds 403 distinctive group calls 399–400 loud sexual advertisement calls 380 territorial behaviour 404–406

interfaces with other disciplines 1, 445–450 mathematical modelling 447–448 models of dominance hierarchy formation 620–621 models of eavesdropping by receivers 613–622 models of effects on signallers and signalling dynamics 608–613 models of structure and organization 607–608 ostariophysan alarm system 546, 547–549

communication, dyadic view 2

possible links with applied biology 448–450

and hormones 482–483, 494

receiver’s perception 445–446, 451–452, 474

limitations of 9

semiochemicals 446–447, 540–541

nestling begging 171–173

size and extent 250

social context 78–79

spatial distribution effects on acoustic

communication networks 2, 9 androgen-modulated behaviour in 494–502 animal versus human 384–385

signals 474 structure influences signals and signalling 286–287

application to welfare of captive animals

structure within a nest 179

assessing social intelligence 447

use of nestling begging to study 179–180

banded wren dawn chorus 337 chemical assessment of predation risk 544–549 cognitive requirements for participants 250–251, 447

see also modelling communication contest behaviour see aggression; redirected aggression; victory displays contest behaviour in fishes 85 effects of eavesdropping 85–86, 94–95 environmental influences 95–97

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:14

Index ‘loser effect’ 88–89 opportunities to eavesdrop 86, 87

Cyprinus carpio see common carp system 552

physical effects of observing fights 90–91, 93–94

Dama dama see fallow deer

potential costs 86

Danio rerio see zebrafish

‘winner effect’ 88–89

dark-eyed junco (Junco hyemalis), metabolic

convict cichlid (Cichlasoma nigrofasciatum), eavesdropping 17 cooperation and altruism, in communication networks 446, 521–525 cooperation theory 521 cotton-top tamarin (Saguinus oedipus) ability to distinguish unfamiliar callers 374 courtship interactions, interruption by eavesdroppers 40–41 ‘social organization and vigilance 426 coyote (Canis latrans) potential for redirected aggression 205, 206 situations where calling is suppressed 376–377 victory display 118–119, 122 crabs (various species), similarities to fiddler crabs 258 crickets (various species), victory displays 118, 122

effects of high androgen levels 504 dawn chorus characteristics 320 communication network view 322 comparison with known daytime singing interactions 324 directed song matching 338–339 environmental explanations for 320–321 female eavesdropping to assess males 321–322 hypotheses to explain 320–322 indicator of male condition 141 interactions in relation to functions 324–325 meaning of song matching and timing 307 mediating changes in social status 339 multi-way male interaction 335, 337–338 possible network structures 322–323 recording methods 324–325, 327–329 singing and male quality 338 social dynamics hypothesis 321, 339

critical (masking) ratio 459

structure in relation to functions 324

Crocuta crocuta see hyaena

temporal patterns of singing behaviour 323,

Crucian carp (Carassius carassius), pheromone system 552 Ctenophorous fordii see Mallee dragon lizard Cuculus canorus (European cuckoo), nestling begging 177 Cuvier’s beaked whale (Ziphius cavirostris), strandings and noise pollution 406–407 Cygnus atratus see black swan cyprinid fishes androgen effects on olfactory sensitivity 488–489 androgen effects on somatic releasers 491–492 pheromone systems 552 Cynops pyrrhogaster see Japanese red-bellied newt

329–336 see also banded wren dear enemy effects, possible androgen mediation 501–502 Delphinapterus leucas see beluga whale Delphinus delphis see common dolphin diademed sifaka (Propithecus diadema), scent over-marking 357 Diana monkey (Cercopithecus diana), attention to alarm calls of another species 373, 380 Dipodomys spp. see kangaroo rats Docidocercus gigliotosi see katydid dolphins filtering of high–frequency signal components 402–403

635

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

636

0 521 582361 7

April 12, 2005

11:14

Index dolphins (cont.) use of directionality of clicks to target signals 402–403 see also bottlenose dolphin domestic fowl (Gallus domesticus) audience effects on calls 66–67, 68 eavesdropping and dominance 28 dominance hierarchies among nestlings 179 and social eavesdropping 28 eavesdropping effects on 97–99, 620–621 effects of prior experience (winner/loser effects) 97–99

information from scent counter-marking 345, 346 interruption of courtship interactions 40–41 selection pressures caused by 30–31, 523–524 strategies for effective eavesdropping 42 use of song overlapping information 304, 305–306 see also bystanders, eavesdropping eavesdropping and altruism 522–523 and audience effects 66

in non-human primates 588–589

and autocommunication 18

modelling formation of 620–621

and dominance hierarchies 97–99

dominance interactions, simulation modelling 606–607 dominance status, problems in poor habitats 144 dunnock (Prunella modularis) courtship interruption by neighbours 41 quiet singing 53 dwarf mongoose (Helogale undulata), group scent marking 363 dyadic view see communication, dyadic view

and predation risk 45 and secrecy 13–14 and transitive inference 574–576 as a type of bystander effect 499 at dawn chorus 338 bystander behaviour and social instability 96 comparison of primates with other animals 592–598 costs and benefits for signallers 30–31, 40–41, 48–49

Eastern towhee (Pipilo erythrophthalmus), signal distance assessment 467–468 eavesdroppers alternative terms for 14

countering with private signalling 52–53, 55 definitions of 3, 10, 13–15 effects on bystanders’ behaviour 85, 86–91, 93, 94–95

and mate choice 142–143

effects on female mate choice 100–103

attending to outcomes of conflicts 198–200

effects on interactions 30–31

attention to asymmetries in songbird vocal

environmental influences on 95–97, 107

interactions 313–314 awareness of dominance relationships 198–200 behavioural responses of those observed 523–524

evidence for 39 evolution in semiochemical communication 542, 559–560, 562 facilitation by interactants 51–52 factors influencing 97, 107–108

costs and benefits 31–32, 39–40, 45–47, 211

for song repertoire information 45–47, 76

distinction from bystanders 84–85

identifying different types 28–29

image scoring allows exploitation by cheats

image scoring and dishonest signals

532–533 image scoring of observed individuals 523

524–525 image scoring in client reef fish 526–528, 529

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:14

Index implications and future research 57–58

used to assess fighting ability 85–86

in anuran choruses 289

using anonymity to counter 56–57

in communication networks 2–3

see also audiences; interceptive

in non-primate mammals 591, 592 in relation to scent marking 235–237 indicators of dominance 52–53

eavesdropping; social eavesdropping; mate copying electric fish (Brienomyrus brachyistius),

individual differences 106–107

androgen levels and dominance signals

information not shared in animals 426

502–505

knowledge about other animals’ dominance ranks 588–590 knowledge about transient social relationships 590–591 knowledge of other animals’ kin 585–587, 588 models of how receivers use networks 613–622 ostariophysan alarm system 545, 547–549 physical effects of observing fights 90–91, 93–94 predation pressures 95–96 quiet song as response to risks 55 reliability and intimacy of information 416–417 risks for interactants 52 scent over-marking and mate choice 359, 360–361 signalling in different modalities 14, 28–29

electrocommunication signals, effects of androgens 487–488, 489 elephant see African elephant elephant seals (Mirounga spp.) loud sexual advertisement calls 380 use of deep sound channel 397 Elephantodeta nobilis see Australian bushcricket Eleutherodactylus coqui see Puerto Rican treefrog emperor penguin (Aptenodytes forsteri), amplitude modification in calls 462 endocrine response, importance of individual’s perception of event 493–494 endocrine system interface with communication networks 446 vertebrates 482 environment, influences on signals and signalling 286–287 see also habitat alteration

social observation in animals 425

Erignathus berbatus see bearded seal

social structure and types of information

Erithacus rubecula see robin

available 87, 583–584 sound transmission in natural habitats 42–43, 44 state dependency of bystander effects 95 strategies for effective eavesdropping 44–48 strategies for private signalling 52–53 to assess potential mates 141–142 to take advantage of ‘loser effects’ 198–200 true recognition or simple association 363–366 use of advertising to facilitate 48–49 use of high perches to improve reception 43, 44

Eschrichtius robustus see grey whale Eubalaena glacialis see southern right whale Eudyptula minor see little blue penguin European cuckoo (Cuculus canorus), nestling begging 177 European minnow (Phoxinus phoxinus), alarm substance 546–548 European newts (Triturus spp.), chemical communication and mate attraction 485–486 European starling (Sturnus vulgaris), unmasking effect of sound segregation 460, 461 European treefrog (Hyla arborea), energy costs of male calling 284–285

637

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

638

0 521 582361 7

April 12, 2005

11:14

Index fallow deer (Dama dama)

state-dependent influences 106–107

costs and benefits of groaning 379

strategies for assessing males 135–136, 137

discrimination of individual male callers

transmission distance and multiple signals

379–380

138–140

functions of groaning in males 379

use of dawn chorus to assess males 321–322

individuality in formant frequencies in

use of network information 141–143

vocalizations 382 fathead minnow (Pimephales promelas), active space of alarm substance 546–548

see also mate choice; mate copying ferret (Mustela putorius), masking release and binaural processing 462–463

Felis catus see cat, feral

Ficedula hypoleuca see pied flycatcher

female mate choice

fiddler crabs (Ocypodidae, Uca spp.)

assessment by eavesdropping 141–142

biology 253–256, 257, 258

call preferences in anurans 264, 283–284,

claw waving display 254–256, 261–262

285–286, 292 desirable male attributes 134

communication networks 247–248, 252–253, 258

effects of observing interactions 78, 100–103

competition for burrow ownership 268–269

effects on male trait distribution 99–100,

conspecific interceptive eavesdroppers

101, 105–106 emergence of strategies 106 environmental influences 107

266–267, 268–271 costs and benefits of interceptive eavesdropping 271–272

future work on 144–145, 146

costs and benefits of signalling 271–272

genetic-based preferences 99–100,

courtship displays 254–256

101 in communication networks 129, 133–134 inferences from female movement patterns 135–136, 137 influence of differences in females 106–107

detection distance for conspecifics 259–260 estimating density 258–259 female assessment of male quality 264–266, 267

influence of particular signals 134

gross signal timing among males 263

influence of predation risk 107

male response to rivals’ waving 268–271

information from dawn chorus 141

neighbouring eavesdroppers 268, 269–271

instigation of male–male interactions

primary and secondary receivers 266–271

139–140, 141 male trait preference versus mate copying 103–105

range and functioning of visual system 259–260 reaction distances for conspecifics 260–261

measures of male quality 137–141

signals other than claw waving 257

movement patterns infer assessment of

strategies for information gathering

males 135–136, 137

266–271

preference for leading signals 264, 285–286

strategies for signal competition 263–266

scent mark assessment in house mice

synchronous waving in males 264–266

225–227 secondary mate choice 133, 135

wandering females as target receivers 266, 267

simultaneous assessment 135, 136

wandering males as eavesdroppers 268–269

song preferences, costs and benefits for

waving rate as indication of male quality

males 152–153

264–266

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:14

Index field cricket (Gryllus bimaculatus), audience effects on males 72 fin whale (Balaenoptera physalus) call response distances 392–395 loud sexual songs 380 use of multipath signal arrivals to locate callers 397

Gallinago media see great snipe Gallus domesticus see domestic fowl game theory models 536, 606, 612–613, 614–618, 619, 620 Gasterosteus aculeatus see three-spined stickleback Gavia immer see black-throated diver genetic relatedness and nestling begging

fish androgen-induced development of somatic display structure 487 eavesdropping on visual interactions 24–27, 28

177–178 gerbil (Meriones unguiculatus), effects of androgens on scent marking behaviour 485

information transfer within shoals 84

ghost crabs (Ocypode spp.) 258

vocal sounds 484

Globicephala melaena see pilot whale

fish semiochemicals ancestral phase 541–542, 543 correlation with predation 545–546 distinction between cues and signals 541–542, 543 eavesdropping 543 evolution into communication networks 541–542, 543, 558–559

Gobius niger see black goby golden-collared manakins (Manacus vitellinus), effects of androgens on visual display 486 golden hamster (Mesocricetus auratus) androgen effects on central motivational mechanisms 490 androgen effects on scent marking behaviour 485

evolution into spying 541–542, 543

female counter-marking

evolution through different functional

female preference for top-scent males 360

phases 541–542, 543 fish taxa with specialized epidermal cells 548–549 fitness benefits for alarm signallers 545, 548–549 hormonal pheromones in information networks 549–558

flank marking 346–347 male preference for top-scent females 360 mechanisms to distinguish top and bottom scents 347–357 over-marking and territory defence 358–359 persistence of scent marks 344 preferential memory for top scents 347–350

hypoxanthine N-oxide 540–541

process of counter-marking 344–345

occurrence of true communication

scent marking and social environment 362

networks 545, 547–549

targeted over-marking in males 358

ostariophysan alarm substance 546–548

true individual recognition 364–365

processing by fish olfactory system 543–544

use of geometric relationships to determine

signal propagation 543–544 spying and communication in information networks 542, 545, 553, 558–560, 562 terminology 541–543 Fringilla coelebs see chaffinch frogs, habitat influence on signalling 286–287 funnel-web spider (Agelenopsis aptera), victory display 117–118, 122

top scent 352, 356 vaginal marking 347 goldfish (Carassius auratus) hormonal pheromones and spawning 549–551, 552 hormonal pheromones in an information network 550–551, 552–553, 555 Gorilla gorilla beringei see mountain gorilla

639

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

640

0 521 582361 7

April 12, 2005

11:14

Index grasshopper (Ligurotettix coquilletti), female attraction to larger choruses 283 great reed warbler (Acrocephalus arundinaceus), female mate assessment 47–48 great snipe (Gallinago media), female assessment of males 139 great spotted cuckoo (Clamator glandarius), nestling begging 177 great tit (Parus major) attention to asymmetries in vocal interactions 314 calculation of maximum detection distance 454–456, 457, 458, 459

grey whale (Eschrichtius robustus), reactions to killer whale sounds 403–404 greylag goose (Anser anser) kin-oriented redirected aggression 204 victory display 115–116 Gryllus bimaculatus see field cricket gulf toadfish (Opsanus beta), acoustic signalling 99 guppy (Poecilia reticulata) eavesdropping and mate choice 100–103 mate copying versus male trait preference 103–105 social interactions 84, 85

critical ratios 457

Gymnorhinus cyanocephalus see pinyon jay

dawn chorus and male condition 141

Gymnotiformes, androgen effects on weak

eavesdropping and transitive inference

electric signals 487–488

575–576 female mate choice and eavesdropping 141–142 information acquired from eavesdropping 19–21, 22–24, 583–584 signal reverberation as a distance cue 469–471, 472–473 sound localization ability 465–467 use of unmodulated sounds 49 greater horseshoe bat (Rhinolophus ferrumequinum), critical ratios 457 green frog (Rana clamitans), victory display 118, 122 green swordtail (Xiphophorus helleri) eavesdropping on visual displays 25, 26–27 social interactions 84, 85 influence of eavesdropping on bystander behaviour 86–91, 93 opportunities to eavesdrop 86, 87 potential costs of combat 86 green treefrog (Hyla cinerea), spatial unmasking of signals 462–463 grey partridge (Perdix perdix), effects of androgens on vocal structures 483–484 grey treefrog (Hyla versicolor) energy costs of male calling 284 female call preferences 284, 288

habitat alteration and female mate assessment 144 effects on song transmission 143–144 see also environment habitat quality, and dominance effects 144 hamadryas baboon (Papio hamadryas), knowledge about transient social relationships 590 hamster see golden hamster harbour porpoise (Phocoena phocoena) avoidance of noise of human activity 406–407 use of clicks for communication 395 harbour seal see common seal harp seal (Phoca groenlandica), distinctive group calls 399–400 hartebeeste (Alcelaphus buselaphus), victory display 118 Hector’s dolphin (Cephalorhyncus hectori), use of clicks for communication 395 Helogale undulata see dwarf mongoose Hemideina spp. see wetas honeybee (Apis mellifera), communication and social coordination 611–612 hooded warbler (Wilsonia citrina), female assessment of males 141

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:14

Index hormones and communication adaptive value of social modulation 497, 502–505 dyadic view 482–483, 494

scent mark assessment of males by other males 227–228 scent mark detection and the vomeronasal organ 219–220, 223, 229, 232

effects on cognitive functions 490–491

scent marking patterns 220–222, 228

effects on communication 482–483

scent marks of subordinate males 229–231

effects on learning and memory 490–491

scent over-marking and mate choice 360

effects on signal reception 488–490

urine scent marking 220–237

expression of somatic releasers 491–492

volatile and non-volatile components of

modulation of central mechanisms affecting motivation 490 modulation of effector pathways 483–488 network view 483, 494–495, 502 pheromones 549–558 reciprocal link with behaviour 481–482 role in control of behaviour 481–482 social modulation of androgen levels 492–494 house cricket (Acheta domesticus)

urine 222–223, 228–229 see also mice house sparrow (Passer domesticus), metabolic effects of high androgen levels 504 human communication dyadic view 416–417 network view 416–417 human eavesdropping 426–432 achieving privacy by behavioural means 427–428

eavesdropping 17

and female curiosity 430–431

male attraction to rivals’ signals 269

as a result of increasing privacy 428–429

house mouse (Mus musculus domesticus)

caused by the need for privacy 428–430, 432

effect of androgens on scent marking 485

everyday occurrence of 427

female reproductive priming through scent

historical evidence for 427, 430–432

231–232 genetic sources of individual scent signatures 232–234 information from age of scent marks 228–229

honest signals and private behaviour 416–417, 429–430 in communication networks 249–250 lack of research on 426–427 male attempts to control others 431–432

information in urine about owner 220–222

privacy and intimate experience 430

kin and group member recognition by scent

social observation time costs in large

234–235 major urinary proteins present in urine 222–223, 233–234 male dominance structures 229–231 male territorial scent marking 221, 223–225, 227–228 MHC odour types and individual scent signatures 232–233, 234–235 pheromones in urine 231–232 reproductive priming and the Bruce effect 232 scent mark assessment of males by females 225–227

groups 428–429 stalking as a means of control 431–432 humans altruism and indirect reciprocity 522–523 ancestral way of life 420 androgen effects on spatial memory 490–491 animal networks compared with human networks 384–385 audience effects on behaviour 74–75 benefits from vigilance 420 common behaviour patterns with other primates 433–435

641

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

642

0 521 582361 7

April 12, 2005

11:14

Index humans (cont.)

humpback whale (Megaptera novaeangliae)

community benefits of vigilance 422–423

call response distances 392–395

congenital adrenal hyperplasia 490–491

distinctive group calls 399–400

creating opportunities to be looked at

loud sexual songs 380

423–424 demands of social monitoring 31–32 domestication and freedom from external vigilance 430

possible social eavesdropping 404–406 hyaena (Crocuta crocuta) ability to infer rank among other group members 592

domestication and intimate behaviour 430

intelligible distance of loud calls 372

enhancing personal image with visual cues

kin-oriented redirected aggression 204

423 equating human and non-human cognitive studies 573–574 ethological studies of vigilance 424–425

persistence of scent marks 344 redirected aggression 191 redirected aggression and target diversion 201

factors affecting vigilance 435

Hydrurga leptonyx see leopard seal

individual benefits of vigilance 422

Hyla arborea see European treefrog

information gained from observation 422,

Hyla cinerea see green treefrog

433–434

Hyla gratiosa see barking treefrog

Lombard effect 464–465

Hyla microcephala see neotropical treefrog

male vigilance and control 434

Hyla versicolor see grey treefrog

parading in front of other people 423–424

Hyperolius marmoratus broadleyi see Broadley’s

privacy and self-awareness 429–430 redirected aggression 208–209 sex differences in social monitoring 424–425 sex differences in spatial memory 490–491

painted reed frog Hyperolius marmoratus marmoratus see South African painted reed frog Hyperoodon ampullatus see northern bottlenose whale

signal echo tail as a distance cue 469–471, 472–473

image scoring

social comparisons 425–426

and tactical deception 531–532, 533–534

social control and vigilance 421–423, 424

and evolution of altruistic behaviour

stalking 424, 431–432 strength and features of female networks 434–435 suspicion of private behaviour in openly living groups 420–422

533–534 benefits for client reef fish 526–527, 528 cheats exploit eavesdroppers 532–533 evidence in client reef fish 527–528, 529 indirect reciprocity

transitive inference task 570–571

and cheating behaviour 532–534

unifying model for vigilance, social

cognitive abilities involved 535

observation and eavesdropping 433–435 unmasking effects of amplitude-modulated background noise 461–462 using vigilance to control 424 victory displays 118–119

occurrence in social networks 534–535 information cascades 270, 607–608 information gathering see bystanders; eavesdroppers information networks

vigilance 420–425

and sex pheromones 549–558

visual cues in movements and gestures 423

fish semiochemicals 540–541

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:14

Index information transfer hierarchies 18–19 within fish shoals 84 infrasound communication, in African elephants 457–459 insect choruses 2, 263–264

masking interference in acoustic communication 157–162 predation of 154, 156–157 predator avoidance 154–157, 164–165 Kayapo people physical adornments 423 killer whale (Orcinus orca)

female preference for leading signals 264

active space of signals 392–395

models of signalling dynamics 609, 610–611

avoidance of human noise 406–407

insects

call suppression near prey 376–377

foraging strategies 611–612

distinctive group calls 400

self-organization among social insects

filtering of high-frequency signal

611–612 interceptive eavesdropping

components 403 importance of oldest female 384

among invertebrates 346

mammal- and fish-eating groups 403

among marine mammals 403

resident and transient groups 403

compared with social eavesdropping 14–15

sound avoidance by potential prey 403

interspecific 16

vocal matching to signal a specific

intraspecific 17–18 ostariophysan alarm system 547–548

individual 402 king penguin (Aptenodytes patagonicus)

signaller payoff 15–16, 18

amplitude modifications in calls 462

see also eavesdropping

signal-to-noise ratio for recognition

invertebrates, interceptive eavesdropping 346

463–464 kingfishers (Coraciiformes), interspecific aggression 191

Japanese macaque see macaques

kin-oriented redirected aggression 203–204

Japanese medaka (Oryzias laticeps), mate

klipspringer (Oreotragus oreotragus), persistence

copying 103 Japanese red-bellied newt (Cynops pyrrhogaster), androgens and pheromone production 485–486 Junco hyemalis see dark-eyed junco

of scent marks 344 !Kung people, openly living groups 420–421, 427 Kuvanga running frog (Kassina kuvangensis) adjustment of male call 282 call response types 282–283

kangaroo rats (Dipodomys spp.), endocrine control of sandbathing in males 485–486

Labroides dimidiatus see cleaner wrasse

Kassina fusca see savannah running frog

Lagenorhynchus australis see Peale’s dolphin

Kassina kuvangensis see Kuvangu running frog

Lagenorhynchus obscurus see Pacific white-sided

Kassina senegalensis see Senegal running frog katydids (various species) acoustic communication 152–153 activity patterns and predator avoidance 153–154

dolphin Lampornis clemenciae see blue-throated humming bird Laniarius aethiopicus see boubou Lemur catta see ring-tailed lemur

chorus signalling dynamics 610–611

Leptonychotes weddelli see Weddell seal

cryptic signalling mode 164–165

leopard seal (Hydrurga leptonyx), distinctive

lunar phase effects 158, 162–164

group calls 399–400

643

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

644

0 521 582361 7

April 12, 2005

11:14

Index Ligurotettix coquilletti see grasshopper Linepithema humile see Argentine ant lion (Panthera leo) ability to distinguish unfamiliar callers 374 benefits and costs of loud calling 374–376 intelligible distance of loud calls 372 male suppression of loud calling 374–376, 376–377 little blue penguin (Eudyptula minor), victory display 116, 121–122 lizards (various species) male attraction to rivals’ signals 269 metabolic effects of high androgen levels 504 Lombard effect 464–465 longtailed macaque (Macaca fascicularis) ability to distinguish social relationships 586–588 kin-oriented redirected aggression 203–204 Lombard effect 464–465 postconflict behaviour 194–204 time spent looking by subordinates 417–418 losers see also winner–loser effects increased receipt of aggression 196–197 loser effects in victims 197–198 physiological changes from conflict 197–198

kin-orientated redirected aggression 203–204, 588 knowledge about transient social relationships 586–588, 590–591 Lombard effect 464–465 postconflict behaviour 194–204 see also longtailed and rhesus macaques Madagascar treefrog (Boophis madagascariensis), acoustic signal repertoire 279 male parental behaviour, audience effects 72–74 male signal traits and exposure to predators 152–153 and female preferences 152–153 male traits distribution effects of mate copying 105–106 effects of female mate choice 99–100, 101 preference versus mate copying 103–105 male-male aggressive signalling, audience effects 68–72 Mallee dragon lizard (Ctenophorous fordii), social observation affects sexual behaviour 425 mammals acquiring and storing social knowledge 383–384 animal versus human networks 384–385

postconflict changes 195–196, 201–202

anti-predator calls 380–381

role of androgens 496–498, 499

benefits of attending to others’ social calls

serotonin-related behavioural inhibition 498–499 Luscinia megarhynchos see nightingale

384 caller identity and status information in acoustic signals 381

Loxodonta africana see African elephant

communication network view of loud

Macaca fascicularis see longtailed macaque

contact call discrimination and memory

calling 373–374 Macaca fuscata see macaques

383–384

Macaca mulatta see rhesus macaque

distinguishing unfamiliar callers 374

Macaca nemestrina see macaques

effects of androgens on scent marking

Macaca radiata see macaques

behaviour 485

Macaca thibetana see macaques

filter characteristics of vocalizations 382

macaques (various species)

fluid social systems increase receivers of

dominance relationships 194–195

loud calls 373

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:14

Index high densities increase receivers of loud calls 373 high encounter rates and long-distance signalling 377–378 information availability in fluid social systems 377–378 information availability in territorial systems 374–377 intelligible distance of calls 372, 382–383 loud calls 249, 372, 374–376, 378–380 recognition of vocalizations from conspecifics 377–378 situations where calling is suppressed 374–377 scent marking in small terrestrial mammals 249 source characteristics of vocalizations 382 source-filter theory of voice production 381–382 victory displays 118–119 see also marine mammals Manacus vitellinus see golden-collared manakins mangabey (Cercocebus albigena), ability to distinguish unfamiliar callers 374 marine environment and acoustic communications 390–391 marine mammals acoustic communication networks 249

numbers of animals in networks 397–399 population density and network size 397–399 predator–prey interceptive eavesdropping 403 restricting range by selecting high-frequency signals 403 restricting signal by decreasing source level 403 size of communication networks 391–394, 399, 408–409 sound propagation in the sea 392–394, 396, 397, 408 value of distant signals 407–408 vocal learning in cetaceans 400–402 vocal matching to signal a specific individual 402 see also mammals marsh wren (Cistothorus palustris), cognitive abilities 569–570 masking interference 157–162 masking release and binaural processing 462–463 and spatial separation of sound sources 462–463 in amplitude-modulated background noise 459–461, 462 mate choice see also female mate choice

active space of signals 391–395

and eavesdropping 141–143

caller identity information in signals

and scent over-marking 359, 360–361

399–400 determining the distance of a caller 395–396, 397 disruption by human noise 406–407 directional high-frequency signal components 402–403

in communication networks 129 mate copying 100–103 costs and benefits for different females 106–107 effects on male trait distribution 105–106 emergence of strategies 106

eavesdropping 403–406, 408

environmental influences 107

features of communications 391

evolutionary consequences 105–106

fission–fusion societies 397–399

influence of predation risk 107

group identity information in signals

state-dependent influences 106–107

399–400 maximum call detection distances 393 methods of restricting and directing signals 402–403

versus male trait preference 103–105 mathematical modelling, interface with communication networks 447–448 see also modelling communication

645

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

646

0 521 582361 7

April 12, 2005

11:14

Index meadow vole (Microtus pennsylvanicus) effects of androgens on scent marking behaviour 485 female preference for top-scent males 354, 355, 359 mechanism to distinguish top and bottom scents 351–356 over-marking and territory defence 358–359 overlap of scent marks to determine top scent 349, 352, 354–356 preferential memory for top scents 350 process of counter-marking 344–345 Megaptera novaeangliae see humpback whale Mehicacu people, openly living groups 421 Melopsittacus undulatus see budgerigar Melospiza melodia see song sparrow Meriones unguiculatus see gerbil Mesocricetus auratus see golden hamster mice (Peromyscus spp.), studies of redirected aggression 208–209 Micronycteris spp. see bats Microtus montanus see montane vole Microtus ochrogaster see prairie vole Microtus pennsylvanicus see meadow vole midwife toads (Alytes spp.)

effects of networks on signal and signalling dynamics 608–613 emergence of ‘selfish herd’ organization 606–607 game theory and eavesdropping 614–618, 619, 620 game theory compared with simulations 612–613 game theory models of dyadic encounters 606 hawk–dove game 615, 618, 619 implications of communication networks 604 individually based spatially explicit simulations 606–607, 612–613 information cascades in networks 607–608 mechanisms to control call timing in choruses 611 network effects on linear dominance hierarchy formation 620–621 network structure in existing models 605–607 new possibilities with network approach 622–623 signal dynamics in acoustic choruses 609–611

evidence for eavesdropping 291

simulation modelling 605

female acoustic responses 283

small-world (scale-free) network analysis

Miopithecus talapoin see talapoin 417–418 Mirounga spp. see elephant seals modelling communication comparing eavesdropping strategies 614–618, 619, 620 complex behaviour from simple rules 606–607 conceptualizing networks 605–608 cooperation strategies 614–615 development and assumptions 604–605 eavesdropping by receivers in networks 613–622 effects of female preferences on signalling in choruses 611 effects of hubs in a network 607–608

607–608 swarm intelligence and self-organization in social insects 611–612 territory establishment 612–613 Molothrus ater see brown-headed cowbird montane vole (Microtus montanus), studies of redirected aggression 208–209 Morymyriformes, androgen effects on weak electric signals 487–488 mountain gorilla (Gorilla gorilla beringei) absence of redirected aggression in females 208 postconflict attacks on losers 196–197 mouse see house mouse

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:14

Index Mozambique tilapia (Oreochromis mossambicus) androgen effects 487, 491–492, 499–500, 501, 502–505

parents’ responses to 172–173, 188 physical competition and dominance hierarchies 179

effects of eavesdropping 27–28

reliability as a signal of need 172–173

endocrine response to social interaction

signal costs and reliability 171–172

493–494 hormonal effects of aggressive priming 78 metabolic effects of high androgen levels 504 winner–loser effects 496–498, 499

signalling errors 184–185 signalling to catch receiver attention 180–183 suppressing competitors’ signals 183 use of locatable signals 181–182 nestling gape 181

Mus musculus see house mouse

network view of communication 2

Mustela putorius see ferret

nightingale (Luscinia megarhynchos)

Myotis lucifugus see brown bat

background noise and song output level

Naiken people, achieving privacy by

cognitive abilities 569–570

51 behavioural means 427–428 Nayaka people, openly living groups 421

eavesdropping and transitive inference (TI) 574–576 Lombard effect 464–465

Neoconocephalus spiza see katydids

male vocal interactions 302

Neogobius melanostomus see round goby

solo and interactive singing 51

neotropical tree frog (Hyla microcephala)

song matching 303–304, 307

female attraction to larger choruses 283 fine adjustment of male calls 282 male selective attention to neighbours 284 nestling begging and genetic relatedness 177–178 as communication network 179–180, 185–186 costs and benefits to the brood 178 distinguishing signalling from physical competition 174–175 dyadic communication approach 171–173 evolution of 171–172, 184 future work 185–186 heat loss and signalling behaviour 178–179 importance of signalling first 183 in interspecific brood parasites 177

use of social eavesdropping 20–21, 22 non-human primates common behaviour patterns with humans 433–435 consortship behaviour 432–433 defensive and social functions of vigilance 417, 419–420, 425–426 factors affect levels of vigilance 419 male bias for vigilance outside the group 417–418 male vigilance and social control 419 proportion of time spent in vigilance 417–418 securing perceptual privacy for some interactions 432–433 sentinel behaviour among high-ranking males 418–419

influence of nestmates 173, 175–176, 178

time spent looking 417–418, 433

locatability of calls 181–182

unifying model for vigilance, social

nestling signalling interactions 175–176 nestlings as a communication network 130, 170–171, 174–179, 180

observation and eavesdropping 433–435 vigilance 417 see also primates

647

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

648

0 521 582361 7

April 12, 2005

11:14

Index northern bottlenose whale (Hyperoodon ampullatus), use of deep sound channel 397 Nycticebus pygmaeus see pygmy loris

Papio hamadryas see hamadryas baboon parent birds comparison of nestlings’ signals 183 information from behaviour of nestlings 174–175

Ochotona princeps see pika odontocetes, restricting range by selecting high-frequency signals 403 oestrogens, receptors in the brain 490–491 olfactory system hormonal modulation of sensitivity 488–489 signal detection and encoding 543–544 olive baboon (Papio anubis) absence of redirected aggression in females 208 effects of redirected aggression 191–192 postconflict attacks on losers 196–197 Onchorhynchus mykiss see rainbow trout

parental care, effects of high androgen levels in males 497, 504–505 Paroaria gularis see red-capped cardinal parrots (Trichoglossus spp.), victory displays 116 Parus atricapillus see black-capped chickadee Parus caeruleus see blue tit Parus major see great tit Passer domesticus see house sparrow peacock blenny (Salaria pavo), androgen effects on somatic releasers 491–492 Peale’s dolphin (Lagenorhynchus australis), silence when captured or near a boat 403

Opsanus beta see gulf toadfish

Perdix perdix (grey partridge), effects of

Opsanus tau see oyster toadfish

androgens on vocal structures

Orcinus orca see killer whale Oreochromis mossambicus see Mozambique tilapia

483–484 Petromyzon marinus see sea lamprey pheromones

Oreotragus oreotragus see klipspringer

fish semiochemicals 541–543

Oryctolagus cuniculus see rabbit

hormonal pheromones 549–558

Oryzias laticeps see Japanese medaka

releaser and primer effects 541

oyster toadfish (Opsanus tau), effects of

reproductive priming effects of mouse

androgens on vocal structures 484

urine 231–232 Phoca groenlandica see harp seal

Pacific humpback dolphin (Sousa chinensis), individual signature signals 400–402 Pacific white-sided dolphin (Lagenorhynchus

Phoca vitulina see common seal Phocoena phocoena see harbour porpoise Phoxinus phoxinus see European minnow

obscurus), individual signature signals

Physalaemus pustulosus see t´ ungara frog

400–402

Physeter macrocephalus see sperm whale

Pan troglodytes see chimpanzee Panthera leo see lion paper wasp (Polistes dominulus), neural development in colonial females 595–596 Papio anubis see olive baboon

physics, interface with communication networks 445–446 pied flycatcher (Ficedula hypoleuca) female assessment of males 135–136, 137 male attraction to rivals’ signals 269

Papio cynocephalus see baboon

pigtail macaque see macaques

Papio cynocephalus ursinus see chacma

pika (Ochotona princeps), ability to distinguish

baboon

unfamiliar callers 374

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:14

Index pilot whale (Globicephala melaena), falling silent when hunted 403 Pimephales promelas see fathead minnow pinyon jay (Gymnorhinus cyanocephalus), ability to rank multiple stimuli 596–598 Pipilo erythrophthalmus see Eastern towhee plainfin midshipman (Porichthys notatus) auditory sensitivity modulation by sex steroids 489 effects of androgens on vocal structures 484 Poecile atricapillus see black-capped chickadee Poecilia latipinna see sailfin molly Poecilia reticulata see guppy Polistes dominulus see paper wasp Porichthys notatus see plainfin midshipman prairie vole (Microtus ochrogaster) androgen effects on central motivational mechanisms 490 studies of redirected aggression 208–209 precedence effect, female preference for leading signals 285–286 predation risk and eavesdropping 14, 16, 17, 45, 95–96 and katydid activity patterns 154–157 and katydid roost site selection 154–157

primates ability to classify objects based on abstract concepts 594–595 ability to rank objects 590 awareness of social rank relationships 28, 588–591 benefits of associations with high-ranking animals 426 complexities of recognizing affiliative relationships 593–594 eavesdropping abilities versus that in other animals 592–598 effects of social organization on vigilance 426 ‘greater intelligence’ hypothesis for eavesdropping abilities 592, 593–595, 598 interconnected groups with males as ‘hubs’ 608 knowledge of other animals’ kin 585–587, 588 ‘large social groups’ hypothesis for eavesdropping abilities 592–593, 595–596, 598 monitoring social relationships as group size increases 594 ‘no species difference’ hypothesis about eavesdropping abilities 593, 596–598

and male signal traits 152–153

social resources 426

and signal detection 159, 160

types of information acquired from

avoidance in communication networks 130 chemical assessment of 544–549

eavesdropping 584 see also non-human primates priming response

influence on female mate choice 107

and male aggression 75–78

influence on mate copying 107

androgen effects on bystanders 499–500,

interceptive eavesdropping among marine mammals 403 mammal anti-predator calls 380–381 quiet song as response to 55 predators

501 private signalling, in male anurans 291 see also quiet song pronghorn antelope (Antilocapra americana), scent over-marking 361

information from prey alarm calls 380–381

Prunella collaris see alpine accentor

use of locatable calls to find prey 181–182

Prunella modularis see dunnock

Procolobus badius tephrosceles see red colobus monkey Propithecus diadema see diademed sifaka

Psittacus erithacus see African grey parrot psychophysics, interface with communication networks 445–446

649

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

650

0 521 582361 7

April 12, 2005

11:14

Index Puerto Rican treefrog (Eleutherodactylus coqui) diphasic advertisement call 280 male selective attention to neighbours 284 models of chorus signalling dynamics 609–610 ‘off response’ call initiation 280–282 Puntius schwanenfeldi see tinfoil barb pygmy loris (Nycticebus pygmaeus), scent over-marking and mate choice 360–361 pygmy Mulga monitor lizard (Varanus gilleni), victory display 118

red deer (Cervus elaphus) discrimination of individual male callers 379–380 individuality in formant frequencies in vocalizations 382 loud mating calls 378–379 red-winged blackbird (Agelaius phoeniceus), signal-to-noise ratios for recognition and for detection 463–464 redirected aggression and postconflict attacks on losers 196–197, 200 and reconciliation with aggressor 192–194,

quiet song

200, 202–203

as aid to anonymity 56–57

and the ‘fight or flight’ response 191–192

as response to predation risks 55

as an outlet for ‘frustration’ 191–192

reasons for variability 56–57

as possible target diversion 201

used for private signalling 53–55

as audience effect 202–203, 209–211 benefits for losers 200, 201–202

rabbit (Oryctolagus cuniculus), effects of

in communication networks 130–131

androgens on scent marking behaviour

in non-human primates 192–194, 195

485

in species other than primates 204–205, 206

rainbow trout (Onchorhynchus mykiss) eavesdropping and dominance 28 effects of redirected aggression 191–192

in winners 196 influence on bystanders’ behaviour 192–194, 200

Rana clamitans see green frog

intraspecific aggression 206, 208–209

Rangifer tarandus see reindeer

kin-oriented 203–204, 588

rank-order fights see contest behaviour

‘loser effects’ 195–196, 198–200

rat (Rattus norvegicus), effects of redirected aggression 191–192 receivers auditory scene analysis 471–474 comparison of signals in a network 183–184 precedence effects 183 signals designed to catch attention 180–183 see also signal detection reciprocal altruism, and behavioural strategies 521–522 red-capped cardinal (Paroaria gularis), territorial behaviour 17–18 red colobus monkey (Procolobus badius tephrosceles)

possible use a signal 201–202, 204–205, 208 summary of functions 210–211 testing occurrence and function 193, 209–210 to attenuate endocrine stress response 191–192 to signal postconflict condition to bystanders 201–202 redtail monkey (Cercopithecus ascanius schmidtii) time spent looking 417–418, 433 redwing (Turdus iliacus), quiet singing 53 reindeer (Rangifer tarandus), androgen effects on scent marking behaviour 485 relational distinction

time spent looking and group size 433

abilities of non-primates 596–598

time spent looking 417–418

abilities of primates 594–595

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:14

Index reptiles, victory displays 118 rhesus macaque (Macaca mulatta) ability to distinguish unfamiliar callers 374

amount of scent and top scent discrimination 349, 353–354 and communication networks 131 and mate choice 359, 360–361

kin-biased redirected aggression 588

androgen effects 485

individuality of frequencies in vocalizations

as a visual signal 362–363

382 postconflict behaviour 194–204 see also macaques Rhinolophus ferrumequinum see greater horseshoe bat ring-tailed lemur (Lemur catta), scent over-marking 362–363 robin (Erithacus rubecula) courtship interruption by neighbours 41 eavesdropping 17

as broadcast signals 235–237, 344 discriminating individual odours 350–351 functions of 345, 354, 355, 358–361 for network communication 219, 362–363 in communication networks 366–367 information available to eavesdroppers 345, 346 information in spatial and temporal distributions 221, 223–225

female assessment of males 141

in social contexts 220

quiet singing 53

mechanisms for targeted over-marking

rodents, redirected aggression 208–209 roe deer (Capreolus capreolus) barking call 380–381 individuality of vocal frequencies 382 round goby (Neogobius melanostomus), semiochemical communication 558–559 rufous-collared sparrow (Zonotrichia capensis), song changes with habitat 143–144

357–358 mechanisms to distinguish top and bottom scents 347–357 olfactory detection 219–220, 223 persistence of signals 219, 344 possible olfactory consequences 345 relative freshness and top scent discrimination 351–353 reliability of signals 221, 223–225

Saguinus oedipus see cotton-top tamarin sailfin molly (Poecilia latipinna), mate copying 103, 104–105 Saimiri boliviensis see squirrel monkey Sakalava people, openly living groups 421 Salaria pavo see peacock blenny Salmo salar see Atlantic salmon Samoan people, openly living groups 421 Sarakatsani people, privacy of the hut 428 savannah baboon, knowledge about transient social relationships 590 savannah running frog (Kassina fusca) female call preference 286 variation in call response types 282–283 Sceloporus jarrovi see lizards scent marking adjacent marking 344–345

to advertise competitive ability 221, 223–225, 227–228 to advertise territory ownership 221, 223–225, 227–228 use of geometric relationships to determine top scent 352, 356 volatile and non-volatile components 219–220 sea lamprey (Petromyzon marinus) larval pheromone attracts migrating adults 555–557 life cycle 555 male sex pheromone in communication networks 557–558 possible specialization in signal production and release 557–559 search for biological control 555

651

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

652

0 521 582361 7

April 12, 2005

11:14

Index selection pressures

high background noise 157–160

female song preferences 152–153

hormonal effects on 488–490

from eavesdroppers 14, 16, 30–31

in a complex environment 152

imposed by audiences 10, 64–65, 79

perceptual mechanisms of receivers

on interactive singing strategies 309

451–452

on primate brain size 28

precedence effect 183, 285–286

on signalling interactions 18–19, 79

reverberation as a distance cue 469–471,

on victory displays 122–123 self-awareness in animals 577–578 semiochemicals

472–473 signal ranging 467–471, 472–473 signal-to-noise ratio 453–456, 457

information from 446–447

simultaneous comparison by receiver 183

see also fish semiochemicals

spatial separation of sound sources 462–463

Senegal running frog (Kassina senegalensis), synchronous calling 282 sergeant major fish (Abudefduf vaigiensis),

use of ‘biological microphone’ 157–160 see also receivers signal production

bystander effects on cleaner fish

amplitude modification 462

behaviour 530–531

anonymous signalling 56–57

Serinus canarius see canary

audience effects 64, 65–66

serotonin

broadcasting from high perches 464–465

and aggressive behaviour 197–198

catching receiver attention 180–183

and behavioural inhibition in losers

competition and cooperation with other

498–499 sex steroids, modulation of sensory perception 488–490 Siamese fighting fish (Betta splendens) androgen effects on eavesdropping 503–504 audience effects 65, 68–72

signallers 63 competition in chorusing interactions 263–264 cryptic signalling 164–165 environmental temperature effects on signallers 153

eavesdropping on visual displays 24–27

high output to overcome interference 181

female mate choice and eavesdropping 142

impact of signal errors in communication

priming and male aggression 75–78 signal detection

networks 184–185 Lombard effect 464–465

amplitude as a distance cue 468–469

predation risk 159, 160

auditory scene analysis by receivers 471–474

sender adaptations to maximize

biological background noise 152 calculation of maximum detection distance 454–456, 457, 458, 459

transmission 464–465 signal evolution and selection pressures 79

compared with signal recognition 463–464

suppressing competitors’ signals 183

critical (masking) ratio for various animals

targeting a specific receiver 63–64

453–457, 459 decision tree learning to discriminate signals 158, 160–162 distance assessment 467–468, 470, 471, 472–473 female song preferences 152–153

use of multimodal components 182 using locatable signals 181–182 signal transmission background noise sources and levels 52–53, 452–453 environmental influences on 286–287

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:14

Index frequency and amplitude masking effects 49–50, 453 influence of communication network structure 286–287 information transfer hierarchies 29 in natural habitats 42–43, 44

small-world (scale-free) network analysis 607–608 Smilisca sila see treefrog social complexity, possible selection pressure on brain development 595–596 social context

masked auditory thresholds 457

influence on eavesdropping effects 97

masking release in amplitude-modulated

influence on hormone levels 481–482,

background noise 459–460, 461, 462

492–494

private signalling 32–53

of communication 1

signaller proximity and masking

of communication networks 78–79

interference 152 sound segregation unmasking effect 460–461

of dyadic interactions 78–79 social eavesdropping 10, 18–32 and dominance hierarchies 28

sound types 49–51

and victory displays 122–123

use of high perches 44, 50

cognitive requirements in female anurans

using background noise fluctuations to reduce masking 459–460, 461, 462 signal recognition, compared with signal detection 463–464 signal-to-noise ratio

289 compared with interceptive eavesdropping 14–15 form of bystander behaviour 10–11 in communication networks 2–3

and signal detection 453–456, 457

in territorial songbirds 19–21, 24

determination 454–456

in territorial systems 18–19

for detection and for recognition 463–464

information gathered from 38–39

signalling interactions between nestlings 175–176 influence of shared fate of nestlings 178 information gathered from 29–30, 38–39 models of effects of networks on signalling dynamics 608–613 selection pressures on 18–19 see also eavesdropping signalling modalities

visual interactions in fish 24–27, 28 see also bystanders; eavesdropping social instability, and bystander decisions 96 social intelligence, assessment of 447 social modulation of androgen levels 492–494 song matching during and after dawn chorus 330, 332–335 value as directional signal at dawn chorus 338–339

and audience effects 10

song output, and habitat quality 144

acoustic signalling in 99

song overlapping, as an aggressive signal 331,

multimodal signalling 99, 182 use of different modalities 28–29 signals, distinguishing from physical competition in nestlings 174–175 simulation models see modelling communication SINDSCAL multidimensional scaling analysis 469–471, 472–473

334, 335–336 song repertoire size, and individual identification 56–57 song sparrow (Melospiza melodia) cognitive abilities 569–570 individual recognition of neighbours 364 song repertoire matching 303 victory display 116, 117, 121–122

653

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

654

0 521 582361 7

April 12, 2005

11:14

Index song transmission, effects of habitat on 143–144 see also sound transmission songbirds eavesdropping and female mate assessment 19–21, 22, 47–48, 314, 577 attention to interaction asymmetries 313–314 cognitive complexity of male ranking judgements 576–577 courtship interruptions by eavesdroppers 40–41 for repertoire size information 45–47 male eavesdropping 313–314 use of quiet singing 53–55 songbirds vocal interactions among neighbours 309–310, 311–312

social contexts of vocal interactions 307–308, 310 solo versus interactive singing 51–52 song alternating 305, 306 song matching 303–305, 307 song overlapping 304, 305–306 song production relative timing 304, 305–306 territorial settlement by multiple interactions 312–313 territories as communication networks 248–249 sound localization cues used for 465 variation in ability between species 465–467

among territorial males 301–302

sound transmission see signal transmission

and territorial behaviour 301–302, 312–313

source-filter theory of voice production

and transitive inference 574–576

381–382

androgen effects 483–485

Sousa chinensis see Pacific humpback dolphin

asymmetries in vocal interactions 301, 309,

South African painted reed frog (Hyperolius

310–311, 313–314 cognitive processes in a communication network 569–570, 574–576 evolution of vocal interaction strategies 314–315 function of the dawn chorus 248–249 functions of specific singing strategies 308, 309–310 in dialogues 302–303 in various contexts 301 interactive dimension of vocal interactions 301, 302–303 memory capacity 569–570 maintenance of territorial spacing pattern 311–312 precise timing and interpretation of interactions 306 resident–intruder vocal interactions 309–311 selection pressures on singing strategies 309 self-awareness 577–578 site-specific dominance in neighbours 311–312

marmoratus marmoratus), female call preferences 287 southern right whale (Eubalaena glacialis), possible social eavesdropping 404–406 spadefoot toad (Spea multiplicata), male attraction to rivals’ signals 269 sparrowhawk (Accipiter nisus), sound localization ability 465–467 spatial memory, sex differences and effects of androgens 490–491 Spea multiplicata see spadefoot toad sperm whale (Physeter macrocephalus) active space of signals 392–395 distinctive group calls 400 fluid social systems and long-distance signalling 377 individual signature signals 400–402 spiteful behaviour, evolution and maintenance 534–535, 536 spotted dolphin (Stenella plagiodon), individual signature signals 400–402 spotted hyaena see hyaena spying see fish semiochemicals

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:14

Index squirrel monkey (Saimiri boliviensis) factors affect levels of vigilance 419 social organization and vigilance 426

modelling of territory establishment 612–613 social eavesdropping in 18–19

Stenella plagiodon see spotted dolphin

testosterone levels, increase in winners 198

Sternopygus macrurus see weakly electric fish

three-spined stickleback (Gasterosteus aculeatus)

stress hormones, effects of social environment 482 Sturnus vulgaris see European starling swordtail (Xiphophorus helleri) androgen effects on sword development 491–492

androgen effects on somatic releasers 491–492 visual perception modulation by sex steroids 489–490 Thryothorus pleurostictus see banded wren Tibetan macaque see macaques

audience effects on males 70

tilapia see Mozambique tilapia

communication via private channels 79

Tinbergen, Niko 1

Sylvia atricapilla see blackcap Sylvia communis see whitethroat

tinfoil barb (Puntius schwanenfeldi), androgen effects on olfactory sensitivity 488–489 toadfish, effects of androgens on vocal

Tachycineta bicolor see tree swallow tactical deception cognitive abilities required 524–525, 535

structures 484 Tonatia silvicola see bats transitive inference

concept of theory of mind 524–525

cognitive abilities of birds 578–579

evolution and maintenance 536

equating human and avian studies 573–574,

exploitation of eavesdroppers 524–525

578–579

image scoring and evolution of 533–534

mechanisms used by animals 572

occurrence in social networks 534–535

presentation to humans 570–571

Taeniopygia guttata see zebra finch

presentation to non-humans 571–572

talapoin (Miopithecus talapoin), time spent

songbirds in communication networks

looking by subordinates 417–418 terrestrial mammals see mammals territorial behaviour and eavesdropping 17–18 and transitive inference 574–576 relative timing of song production 305–306 resident–intruder vocal interactions 309–311 scent marking in house mice 220–222, 223–225, 227–228 song matching 303–305, 307 vocal interactions among neighbours 309–310, 311–312 victory displays 122–123 territorial systems and communication networks 143–144 availability of information from loud calls 374–377 dear enemy effects 501–502

574–576 studies with parrots 572–573 tree shrew (Tupaia belangeri), effects of androgens on scent marking behaviour 485 tree swallow (Tachycineta bicolor), nestling signalling errors 184–185 treefrog (Smilisca sila), synchronous calling 282, 283–284 tremulation, as cryptic signalling mode 164–165 Trichoglossus spp. see parrots Triturus spp. see European newts triumph ceremonies see victory displays Troglodytes troglodytes see wren t´ ungara frog (Physalaemus pustulosus) adaptations for acoustic competition 278 male selective attention to neighbours 284 Tupaia belangeri see tree shrew

655

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

656

0 521 582361 7

April 12, 2005

11:14

Index Turdus iliacus see redwing Turdus merula see blackbird Tursiops truncatus see bottlenose dolphin Tyto alba see barn owl urodeles, androgen effects on pheromone production 485–486 Uca spp. see fiddler crabs

vocal matching, by cetaceans to signal a specific individual 402 vomeronasal organ 219–220, 223, 229, 232 weakly electric fish (Sternopygus macrurus), androgen effects on electroreception 489 Weddell seal (Leptonychotes weddelli) colony-specific call types 399–400

Varanus gilleni see pygmy Mulga monitor lizard vertebrate endocrine systems 482 vervet monkey (Cercopithecus aethiops) bystander effects on aggression 202–203 factors affecting levels of vigilance 419

ear damage from underwater noise 406–407 territorial behaviour 404–406 western scrub jay (Aphelocoma californica), ability to rank multiple stimuli 596–598 wetas (Hemideina spp.), victory displays 118, 122 white-crowned sparrow (Zonotrichia leucophrys)

kin-oriented redirected aggression 203–204

cognitive abilities 569–570

knowledge of other animals 585, 589–590

metabolic effects of high androgen levels

male parental behaviour 72–73 victory displays 11, 114 amphibians 118

504 white-faced capuchin monkey (Cebus capucinus)

androgen effects on vertebrates 486–487

male vigilance and social control 419

arthropods 117–118, 122

time spent looking as group size increases

birds 115–117

433

categories of 121–122

white whale see beluga whale

distinctive features of 120

whitethroat (Sylvia communis)

distinguished by context 119–120

courtship interruption by neighbours 41

effects within the communication network

private signalling 55

120–121, 122–123 functions within the winner–loser dyad 120–122 humans 118–119 mammals 118–119

signalling in different modalities 29 use of high perches to improve reception 42–43 whydahs (Vidua spp.), nestmate signal suppression 183

occurrences of 119

Wilsonia citrina see hooded warbler

reptiles 118

winner–loser effects

Vidua spp. see whydahs

role of androgens 496–498, 499

vigilance

serotonin-related behavioural inhibition in

in animals 417–420 in humans 420–425 visual interactions in fish, eavesdropping on 24–27, 28 visual system, hormonal modulation 489–490 vocal learning in cetaceans 400–402

losers 498–499 see also losers; winners winners increase in androgen levels 198 physiological changes following victory 198 postconflict aggression 196 ‘winner effects’ following victory 198

P1: IRK/KFO P2: KOD 0521823617ind.xml CU1917B/McGregor

0 521 582361 7

April 12, 2005

11:14

Index Wistar rat, effects of androgens on scent marking behaviour 485

Yagua people, achieving privacy by behavioural means 427–428

wolf (Canis lupus) ability to distinguish unfamiliar callers 374 individuality in vocal frequency 382

zebra finch (Taeniopygia guttata) effects of androgens on vocal structures 483–484

scent marking 363

Lombard effect 464–465

situations where calling is suppressed

signal-to-noise ratios 463–464

376–377 victory display 118–119 wren (Troglodytes troglodytes) signal echo tail as distance cue 469–471, 472–473 use of high perches 43, 44, 464–465 use of low-frequency sounds 49

signal amplitude as a distance cue 468–469 sound localization ability 465–467 zebrafish (Danio rerio), active space of alarm substance 546–548 Zinacantan people, suspicion of private behaviour 421–422 Ziphius cavirostris see Cuvier’s beaked whale Zonotrichia capensis see rufous-collared

Xenopus laevis see African frog Xiphophorus helleri see green swordtail; swordtail

sparrow Zonotrichia leucophrys see white-crowned sparrow

657