Advances in the Study of Behavior, Volume 27 (Advances in the Study of Behavior)

Advances in

THE STUDY OF BEHAVIOR VOLUME 27

Advances in THE STUDY OF BEHAVIOR Edited by

PETER J. B. SLATER JAY S. ROSENBLATT

CHARLES T. SNOWDON MANFRED MILINSKI

Stress and Behavior A Volume in

Advances in THE STUDY OF BEHAVIOR VOLUME 27 Edited by ANDERS PAPEMOLLER Lahoratoire d’Ecologie UniversitP Pierre et Marie Curie Paris, France

MANFRED MILINSKI Zoologisches Institut Ahteilung Verhalten-sokologie Universitat Bern Hinterkappelen, Switzerland

PETER J . B. SLATER School of Environmental and Evolutionary Biology University of St. Andrews Fife, United Kingdom

ACADEMIC PRESS San Diego London Boston New York

Sydney Tokyo Toronto

T h k book Copyright

ir, p r i n t e d

on

acid-free paper.

@

0 1998 by A C A D E M I C PRESS

A l l Rights Reserved. No part o f this publication may be reproduced or transmitted in any form or by any iiicans. electronic or mechanical. including photocopy. recording, or any information \torage and retrieval system. without permission i n writing from the Publisher. The appearance of the code at the bottom o f the first page of a chapter in this book indicates the Publisher's consent that copies of the chapter may be made for personal or internal use of specific clients. This consent i s given on the condition. however. that the copier pdy the stated per copy fee through the Copyright Clearance Center. Inc. ( 2 2 2 Rosewood Drive. Danvers. Massachusetts 0 1923). for copying heyond that permitted by Sections 107 or 108 o f the U . S . Copyright Law. This consent does not extend to other kinds o f copying. such 3 s copying for general distribution. for advertising or promotional purpoes. for creating new collective works. or for resale. Copy fee\ for pre-199X chapters are as shown on the t i t l e pages. If no fee code appears on the title page, the copy fee i s the hame as for current chapters.

0065-3454/9X $25.00

Academic Press (I dii,i.\;mi

o/'Iioi.c~orrrt Brui.i. & C'onilitrii?

525 B Street. Suite 1900. San Diep, California 97_101-4495.USA http:/Jwww.apnet.com Acadeniic Press Limited 24-2X Oval Road, London N W I 7DX. U K

http://www.Iibuk.co.uk/ap/ International Standard Book Number: 0-1 2-004527-3

PRINTED IN THE UNITED STATES OF AMERICA 9X 99 00 01 01 03 Q W 9 8 7 6

5

4

3

2

I

Contents

Contributors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ix xi ...

Xlll

The Concept of Stress and Its Relevance for Animal Behavior DIETRICH VON HOLST

I. I1. I11. IV .

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Concept of Stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Social Stress in Mammals . . . . . . . . . . . . . . . . . . . . . . . . . . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 2 42 106 109

Stress and Immune Response VICTOR APANIUS 1. I1. I11. IV. V. Vl . VII .

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Nature of Stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Nature of Immunocompetence . . . . . . . . . . . . . . . . . . Neurological Linking of Stress and Immunocompetence . . Endocrine Linkage of Stress and Immunocompetence . . . . Why Stress Alters Immunocompetence . . . . . . . . . . . . . . . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

133 134 136 140 142 145 149 150

Behavioral Variability and Limits to Evolutionary Adaptation under Stress P. A . PARSONS 1. I1. I11. IV . V.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Energy Limits to Adaptation . . . . . . . . . . . . . . . . . . . . . . . Variability and the Survival of Variants . . . . . . . . . . . . . . . Extending the Limits of Adaptation . . . . . . . . . . . . . . . . . . From Stress-Resistance Genotypes to a Connected Metabolism . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V

155 158 164 165 169 174 174

vi

CONTENTS

Developmental Instability as a General Measure of Stress ANDERS PAPE M@LLER I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I1. Genetic and Environmental Determinants of Developmental Instability . . . . . . . . . . . . . . . . . . . . . . . . . . 111. Directional Selection and Developmental Instability . . . . . IV . Fitness Correlates of Developmental Instability . . . . . . . . . V . Practical Uses of Developmental Instability . . . . . . . . . . . . VI . Conclusions and Prospects for Future Studies . . . . . . . . . . VII . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

181

186 190 192 193 206 208 208

Stress and Decision-Making under the Risk of Predation: Recent Developments from Behavioral. Reproductive. and Ecological Perspectives STEVEN L . LIMA I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I1. Behavior of Feeding Animals: Classical Motivations . . . . . 111. Patterns of Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV . After an Encounter with a Predator . . . . . . . . . . . . . . . . . . V . Social Situations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI . Reproduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII . Long-Term Consequences of Decision Making . . . . . . . . . VIII . Ecological Influences and Implications . . . . . . . . . . . . . . . . IX. Additional Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . X . Conclusions and Summary . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

215 217 225 235 239 241 245 248 261 264 265

Parasitic Stress and Self-Medication in Wild Animals G. A . LOZANO I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I1. Self-Medication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111. Prophylactic Se1f.Medication . . . . . . . . . . . . . . . . . . . . . . . . IV. Therapeutic Self-Medication . . . . . . . . . . . . . . . . . . . . . . . . V . Skepticism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI . Behavioral Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . .

291 293 294 298 303 304

CONTENTS

VII . Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VIII . Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

vii 308 310 311

Stress and Human Behavior: Attractiveness. Women’s Sexual Development. Postpartum Depression. and Baby’s Cry RANDY THORNHILL AND F. BRYANT FURLOW I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I1. Human Attraction and Attractiveness . . . . . . . . . . . . . . . . I11. Parent-Daughter Relations and Women’s Sexual Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV . Postpartum Depression . . . . . . . . . . . . . . . . . . . . . . . . . . . . V . Infant Crying as a Signal of Phenotypic Quality . . . . . . . . . VI . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

319 321 332 341 352 356 359

Welfare. Stress. and the Evolution of Feelings DONALD M . BROOM I . Feelings. Their Role and Their Evolution . . . . . . . . . . . . . I1. Welfare. Stress. and Feelings . . . . . . . . . . . . . . . . . . . . . . . . I11. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

371 394 400 401

Biological Conservation and Stress HERIBERT HOFER AND MARION L . EAST I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I1. Stress in a Conservation Biology Context . . . . . . . . . . . . . . I11. Designing a Conservation Study to Measure Stress and Its Impact . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV. The Natural History of Stress . . . . . . . . . . . . . . . . . . . . . . . V . Effects of Anthropogenic Stressors . . . . . . . . . . . . . . . . . . . VI . Conservation Research and Management Activities as Stressors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII . The Equivalence of Natural and Anthropogenic Stressors

405 407 420 428 452 473 486

...

Vlll

CONTENTS

VIII . Minimizing Occurrence and Impact of Stress in Conservation Research and Management . . . . . . . . . . . . . . IX . Conclusions: How Important Is Stress in Biological Conservation? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . X . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

488 494 496 497

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

527

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

549

Contributors

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

VICTOR APANIUS (133),Department of Biological Sciences, Florida International University, University Park, Miami, Florida 33199 DONALD M. BROOM (371), Department of Clinical Veterinary Medicine, University of Cambridge, Madingley Road, Cambridge CB3 OES, United Kingdom MARION L. EAST (405), Max -Planck -1nstitut f i r Verhaltensphysiologie, 0-82319 Seewiesen Post Starnberg, Germany BRYANT FURLOW (319). Department of Biology, University of New Mexico, Albuquerque, New Mexico 87131 HERIBERT HOFER (405). Max -P/anck -1nstitut far Verhaltensphysiologie, D 82319 Seewiesen Post Starnberg, Germany DIETRICH VON HOLST (1). Department of Animal Physiology, University of Bayreuth, 95440 Bayreuth, Germany STEVEN L. LIMA (215), Department of Life Sciences, Indiana State University, Terre Haute, Indiana 47809 G. A. LOZANO (291), Department o,f Biology, University of California, Riverside, Calljornia 92522

ANDERS PAPE M0LLER (181), Laboratoire d’Ecologie, UniversitP Pierre et Marie Curie, Paris Cedex 5, France

P. A. PARSONS (155), School of Genetics and Human Variation, La Trobe University, Bundoora, Victoria 3083, Australia RANDY THORNHILL (31Y), Department of Biology, University of New Mexico, Albuquerque, New Mexico 87131

ix

This Page Intentionally Left Blank

Preface

The aim of the Advances in the Study of Behavior series remains as it has been since the series began: to serve the increasing number of scientists who are engaged in the study of animal behavior by presenting their theoretical ideas and research to their colleagues and to those in neighboring fields. We hope that the series will continue its “contribution to the development of cooperation and communication among scientists in our field,” as its intended role was phrased in the preface to the first volume in 1965. Since that time, traditional areas of animal behavior research have been given new vigor through the links they have formed with related fields and by the closer relationship that now exists between those studying animal and human subjects. Beginning with Volume 25, which was on the special topic of Parental Care, we departed from the previous policy of publishing articles on varied subjects in every volume. This volume, titled “Stress and Behavior,” is the second thematic volume. The next volume will again be a broad-ranging one, as was the last, and it is our intention to continue the series with this mixture of wide-ranging volumes of eclectic interest and occasional volumes focusing on particular themes that appear timely to us. Although volumes such as this do represent a new initiative, they do not, we believe, violate the basic principles underlying the series. The specific theme of Stress and Behavior was chosen because it is an especially exciting and active area of research at present, and one to which researchers with a wide range of approaches and backgrounds are making important contributions. We have invited as contributors leading experts across this range, thus giving a truly multidisciplinary perspective on the topic. For this volume we have been fortunate to be joined by Dr. Anders Pape Mflller as guest editor, and his expertise in the area has been of immense help to us. Sadly, this will also be the last volume for which Dr. Manfred Milinski will act as an associate editor: he has made a valuable contribution to the series, for which we are very grateful. His place will be taken by Dr. Tim Roper, of the University of Sussex, and we look forward to benefiting from his broad interests and well-honed editorial skills in future volumes.

P. J. B. Slater

xi

This Page Intentionally Left Blank

Stress & Behavior Introduction

All organisms suffer from a deficiency of one or more resources during their lifetime, and conditions for development, growth, survival, and reproduction are rarely, if ever, optimal. This was realized by Charles Darwin, who used competition among conspecifics for limiting resources as a cornerstone of his theory of natural selection. Although suboptimal living conditions are widespread, the relationship between environmental conditions and behavior has not attracted much attention from scientists (for exceptions, see Hoffmann and Parsons, 1989; Maller and Swaddle, 1997). The conditions under which organisms live are frequently suboptimal, and the difference between suboptimal and optimal conditions is often perceived by individuals as causing a change in their state. Stress is an appealing but illusive concept in biology, with definitions being almost as numerous as the different fields of research. Although stress can be defined explicitly in terms of concentrations of biochemicals involved in metabolism (Ivanovics and Wiebe, 1982), a general but still operational definition is that provided by Hoffmann and Parsons (1989): the state caused by any agent that results in suboptimal performance and potentially causes permanent damage to an individual. Life has evolved under stressful conditions, although exceptions exist, such as the relatively constant environments experienced by organisms living in caves and hot springs. The fact that organisms generally live and reproduce under adverse environmental conditions has not been well appreciated by the majority of the community of biologists. Theoretical evolutionary biology has not considered stress to be of overriding importance. For example, Sewall Wright’s famous shifting balance theory is based on the concept of a fitness landscape with multiple valleys and peaks. The unappreciated fact that different phenotypes may be expressed under poor and optimal conditions may have enormous effects on the ability of species to evolve from one peak to another. Empirical evolutionary biologists have performed most of their studies on fruit flies, mice, and rats under benign laboratory conditions where stress is much reduced compared to natural situations. Interestingly, phenotypic and genetic parameters are not congruent or even comparable in such situations, as revealed by recent studies of animals reared under more and less benign conditions (Bijlsma and Loeschcke, 1997; Maller and Swaddle, 1997). This discrepancy between conditions under which animals are studied and conditions under which they live also applies to behavioral research. ...

Xlll

xiv

INTRODUCTION

Laboratory studies are generally performed under benign conditions with ad libitum availability of food and a virtual absence of predators and para-

sites. Field studies are predominantly conducted in optimal habitats with high population densities, whereas much less is known about behavior of animals in low-quality, marginal habitats where stress is predominant. Behavior under adverse environmental conditions is a topic of general interest for two reasons. First, such studies may provide us with a better understanding of the conditions under which most evolution has taken place. Second, studies of behavior under stressful conditions may give us a better understanding of the consequences of global change, as well as provide us with important information on conservation biology. In the first chapter, von Holst describes how social relationships, especially aggressive ones, can influence the physiological state of individuals in many positive and negative ways. This chapter introduces the neuroendocrinology of the stress response. Stressed organisms are easy targets of infectious diseases. The complex strategic use of the immune system sometimes affords suppression of immune defenses under stress as discussed by Apanius in the second chapter. In the third chapter, Parsons presents an energetic approach to the fitness of organisms that are challenged by biotic and abiotic stress. Under stressed free-living conditions, his environmental model suggests that favored “good genotypes” are likely to be stress resistant and heterozygous. In the fourth chapter, Mgller discusses developmental stability as a reflection of the ability of organisms to buffer their developmental trajectories against disturbance. The inability to fulfill this goal can be assessed in terms of fluctuating asymmetry, i.e., random deviations from perfect symmetry, and used for studies of, for example, environmental monitoring, animal welfare, and human medicine. In the fifth chapter, Lima discusses the many ways in which behavioral decision-making alters the nature of predator-induced stress. In the sixth chapter, Lozano discusses the growing evidence that wild animals use self-medication in response to parasitic stress. In chapter 7, Thornhill and Furlow examine how stress interacts with human behavior in relation to physical attractiveness, the development of women’s sexuality, and parental investment in babies. The biological function of feelings such as pain and fear may be to affect an organism’s behavior in such a way that it maximizes the chances that good things will happen and minimizes the chances that bad things will happen. Broom discusses welfare, stress, and the evolution of feelings in chapter 8. The last chapter, by Hofer and East, reviews why stress has important implications for biological conservation and considers practical ways in which conservationists can identify and tackle problems due to stress.

INTRODUCTION

xv

We believe that a concerted research effort concerning behavior of animals living under adverse environmental conditions will add considerably to our understanding not only of the role that behavior plays in the general coping strategies of animals, but also of the flexibility of behavioral strategies under variable environmental conditions. The authors of the chapters of this thematic volume of Advances in the Study of Behavior have shown some ways in which this may be achieved.

A. P. M@ller M. Milinski

References

Biljsma, R. and Loeschcke, V. (eds.) (1997). “Stress. Adaptation. and Evolution.” Birkhauser. Basle. Hoffmann, A. A. and Parsons, P. A. (1989). “Evolutionary Genetics and Environmental Stress.” Oxford University Press, Oxford. Ivanovics, A. M. and Wiebe, W. J. (1982). Toward a working definition of stress: A review and critique. In “Stress Effects on Natural Ecosystems” (G. W. Barrett and R. Rosenberg, eds.), pp. 13-27. Wiley, New York. M@ller,A. P. and Swaddle, J . P. (1997). “Asymmetry, Developmental Stability and Evolution.” Oxford University Press, Oxford.

This Page Intentionally Left Blank

ADVANCES IN THE STUDY OF BEHAVIOR. VOL 21

The Concept of Stress and Its Relevance for Animal Behavior DIETRICH VON HOLST DEPARTMENT OF ANIMAL PHYSIOLOGY UNIVERSITY OF BAYREUTH

95440 BAYREUTH,

GERMANY

1. INTRODUCTION

Mammals live in social systems, which differ from species to species but are relatively constant for any species, although some variation as a function of the ecological situation is possible. These social systems are maintained by constant contact between the animals, which not only affects the behavior of the individuals, but may also positively or negatively influence their fertility and health. The negative consequences of social interactions are usually explained by the stress concept as shown in a particularly impressive way in the Australian dasyurid marsupials of the genus Antechinus. This genus is widely distributed in Australia and feeds mainly on insects and small vertebrates. All species examined so far exhibit an extremely synchronous life cycle: At the end of September-during the Australian spring-the females give birth to their young, which are weaned in January, but continue to live in harmony with their mothers for a few more months. At the end of May, the young leave their birthplace and spread out within their habitat. The short reproductive season commences in August, during the Australian winter. During the search for females, the males roam their territory and are continually involved in vehement fights with other males. Following the 2- to 3-week reproductive season and before the end of the first year of their life, virtually all the males “die off.” The females survive and after a 1-month gestation period they give birth to their young. A new cycle ensues (Woolley, 1966). The death of the males is due to typical stress reactions characterized by a tenfold increase in the plasma levels of free glucocorticosteroids and a simultaneous breakdown of the immune and inflammatory responses. As a consequence, gastrointestinal hemorrhaging associated with gastroduodenal ulcers, bacterially induced hepatic necrosis, heavy parasitic diseases, and other infections cause the death of all males 1

Copyright 0 1998 by Academic Press All rights of reproduction in any form reserved 0065-3454198 $ZS.MI

2

DIETRICH VON HOLST

within a short period of time (Barnett, 1964; Bradley et al., 1980; McDonald et nl., 1981, 1986). The physiological changes causing death are based mainly on the increased levels of aggression between the males. Accordingly, if males are captured before the breeding seasons and housed singly, they may live to about 2 years of age as do females under natural conditions. This means that the males die of stress mainly due to their enhanced aggression and persistent sexual activity. In this chapter, the significance of the stress concept in gaining a better understanding of social mechanisms in nonhuman mammals will be examined. In the second section the development of this concept during the last 50 years and the resulting current understanding of different stress reactions are described. The triggers of stress reactions are mainly psychical processes resulting from the assessment of a situation by an individual. Dependent on its coping behavior, these processes lead to different physiological response patterns, which can result in a number of pathophysiological effects. In the third section the most important currently applied methods in assessing stress levels in animals are introduced. Particular attention is paid to methodological problems as well as to the limits of interpretation. Focal points are the sympathetico-adrenomedullary and pituitary-adrenocortical systems, the pituitary-gonadal axis, and the immune system. In the fourth section an overview is provided of the relationships between social situations and stress responses, in which I concentrate mainly on our research on the monogamous and territorial tree shrews and the polygamous and territorial European wild rabbits. In these cases the social rank of an individual, as well as its sociopositive interactions with conspecifics, and the stability of the social system are determinants in the effects of a social situation on the individual’s vitality and fertility.

OF STRESS 11. THECONCEPT

A. INTRODUCTION

Few biomedical terms are as popular as stress. However, its definition is as inconsistent as the research strategies of the scientists from a variety of disciplines (biomedicine, psychology, or sociology) working on stressrelated topics (Lazarus and Folkman, 1984;Levine and Ursin, 1991; Weiner, 1991). Although it is probably impossible to find a definition that the majority of researchers will agree upon, and some authors even suggest that the concept is meaningless (e.g., Engel, 1985), the concept of stress has a long history that goes back to the ancient Greeks. As early as the

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

3

year 400 B.c., Hippocrates ascribed the causes of disease to disturbing forces of nature and referred to the adaptive responses of the body as the “healing power of nature.” One hundred years later psychological stress was mentioned by Epicurus, who suggested that coping with emotional challenges is a way of improving the quality of life (cited from Chrousos et al., 1988). All recent stress concepts deal with the daily social and nonsocial stimuli that are challenging or threatening to the survival, health, and reproductive success of animals and that are, therefore, an essential part of natural selection. OF THE STRESS CONCEPT B. DEVELOPMENT

Modern biomedical stress research is based in particular on the work of the American physiologist Walter B. Cannon and his colleagues, and the work of the Canadian physician Hans Selye. 1.

Cannon’s Fight or Flight Syndrome

In 1929, Cannon published an important monograph entitled “Bodily Changes in Pain, Hunger, Fear and Rage,” in which he summarized the results of decades of research into the effects of emotional challenges on physiological processes. Cannon did not regard emotions as purely subjective sensations, but as all-encompassing phenomena that also embrace objective physiological and ethological components and could, therefore, be analyzed scientifically. This opinion is still held today (Buck, 1988a). Cannon found a multifarious mosaic of changes in bodily functions in both animals and humans in emotionally stimulating situations: a reduction in gastric and intestinal function; an increase in heart rate, blood pressure, and breathing rate as well as in the number of red blood cells and the sugar content in the blood; and accelerated blood clotting. All these effects were attributed by Cannon to the increased activity of the sympathetic nervous system (Fig. 1). Cannon not only concentrated on the explanation of these causal controlling mechanisms, but also questioned the adaptive value of this variety of reactions. His conclusion was the following: All these effects increase the capability of an individual to react actively to critical situations in its environment-to prepare it for fight or flight. However, Cannon also realized that not every emotional process results in the activation of the organism. A difficult situation that cannot be changed by action can trigger apathetic, inactive behavior and lead to, among other things, a reduction in pulse rate and blood pressure. An early and detailed description of this reaction is given by Charles Darwin in 1872 in his book

Release of

Decreased clotting time

I I I

I

1

ventilation increased

Increased blood flow to brain, heart, and skeletal muscles

Lipolysis

I

Glycolysis

Increased plasma levels of

FIG. 1. Cannon’s fight or flight response: Activation of the sympathetic nervous system and release of the adrenomedullary hormones epinephrine and norepinephrine. Their effector organs and the effects on the whole organism are shown (see also Fig. 11).

I

STRESS A N D ITS RELEVANCE FOR ANIMAL BEHAVIOR

5

“The Expression of the Emotions in Man and Animals,” in which he spoke of the feeling of despair or grief.

2. Selye’s General Adaptation Syndrome Cannon and associates were concerned mainly with the acute responses of an individual to potentially dangerous stimuli, while recognizing that repeated exposure to such stimuli results in adaptive changes in the organism that make it more resistent to challenge. The adaptation of an organism to chronic challenges, on the other hand, was the main interest in the research by Hans Selye, who also introduced the term stress into biomedical research (Selye, 1950). Contrary to the use of this term in everyday language and in other scientific disciplines, he designated stress as the response of an organism to any strong and potentially damaging stimulus, while for the damaging stimulus he introduced the term stressor. a. General Adaptation Syndrome. In 1936 Selye published a short paper “A Syndrome Produced by Diverse Nocuous Agents,” describing for the first time a pattern of physiological reactions in response to various damaging agents or critical situations, such as injuries, cold, infections, intoxications, burns, o r strong muscular exercise. An organism responds to these different stressors with stimulus-specific responses, such as with immunological responses to infections or with increased erythrocyte numbers to oxygen deficits. However, no matter how variable the nature of these stressors, according to Selye, they always elicit the same pattern of physiological responses, which seem to represent a generalized effort of the organism to adapt itself to the new situation. The response of the organism to stressors is accomplished in three stages, which Selye called the general adaptation or stress syndrome. 1. Alarm Reaction. The initial responses to physiological changes induced by a stressor are thymolymphatic involution, gastrointestinal ulceration, and loss of cortical lipids and medullary chromaffin substances from the adrenals, indicating an activation of the sympathetico-adrenomedullary and pituitary-adrenocortical systems. If the stressor is too strong (severe burns, extreme temperatures), death may result within a few hours. However, if the stressor is not too strong and has only a brief effect, then it usually has no further consequences for the organism, which quickly regains its original state. 2. Stage of Resistance. If the challenge persists, the body adapts itself to tolerable stressors, such as very low temperatures or unavoidable physical exertion, by changing its entire physiological state. According to Selye, the increased activity of the adrenal cortex during this stage of defense or

6

DIETRICH VON HOLST

resistance is of particular importance. The adrenal cortex adapts to the increased production and secretion of its hormones by markedly increasing its size. Concomitant with this, those functions unnecessary to coping with the stressor, such as growth, gonadal activity, and immunological resistance are suppressed.

3. Stage of Exhaustion. If the stressor is sufficiently severe and prolonged, the adaptation mechanisms will finally fail and lead to the death of the individual. Long-term, tolerable stress therefore impairs fertility and vitality in animals. Simultaneously, the initial advantageous physical adaptive reactions (particularly the increased production of adrenocortical hormones) were thought by Selye to lead to a number of diseases (referred to as “diseases of adaptation”), ranging from high blood pressure and gastric ulcers to diabetes and cancer (Selye, 1950, 1976, 1981). This concept of stress had a lasting effect on research. Ever since the 1950s, hundreds of scientific publications with the term stress in their titles have been published each year. Due to the central role of the adrenocortical system in the Selyean concept, research on stress centered to a large extent around the adrenal cortex and its hormones, while other endocrine responses or systems such as the gastrointestinal or the adrenomedullary system were largely neglected, even though changes in these systems were clearly recognized. As a consequence, it became common practice to equate stress with adrenocortical activity: Increased serum levels or excretion rates of glucocorticosteroids, such as cortisol and/or corticosterone, or other indications of heightened adrenocortical activity were used as an index of the adaptation of an organism to a stressful situation or to the intensity of a stressor. Although, even by today’s standards, this approach may appear attractive methodologically, it is important not to equate stress with adrenocortical function, as the responses of an organism to new and sudden demands comprise almost all physiological systems. Heightened adrenocortical activity constitutes only one part of this response pattern and is in no way sufficient to characterize the stress state of an animal, especially because adaptive responses to stressful situations may occur without any heightened adrenocortical activity (discussed later). b. Physical versus Psychical Stress. While originally the adrenocortical activity was assessed by changes in adrenal gland weight and morphology, developments in the late 1950s yielded the first biochemical methods enabling determination of adrenocortical activity by measurement of their hormone levels in plasma or urine. This led to a growing interest of psychologists and physiologists in emotionally induced adrenocortical activation. By 1956, Mason and Brady had demonstrated for the first time increased

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

7

17-hydroxycorticosteroid plasma levels in rhesus monkeys in an emotionally distressing situation (“conditioned anxiety”), and at the same time previously impossible studies on humans began. In the following years, countless studies on mammals (including humans) demonstrated strong adrenocortical activation not only during acute emotional arousal, but also in longlasting emotionally disturbing situations. Nowadays, emotional “loads” are cited as the most common reasons for stress in humans and, as pointed out by Ursin and Olff in a recent review (1993), emotional processes are also the most commonly used stressors in animal research. These emotional processes must be considered even when the experimenter assumes he or she is dealing with physical stressors. It is necessary to bear in mind, however, that an activation of the adrenocortical system can also be induced without any concomitant emotional arousal (such as during surgery under deep anesthesia or during infections and the resulting release of mediators by the immune system). A most important contribution to the modern stress concept is the work of Mason and associates on the effects of psychological influences on the general endocrine response pattern (Mason, 1968a,c). His own work, as well as the results of the relevant literature, led Mason to the conclusion that situations of novelty, uncertainty, or unpredictability are especially potent in inducing heightened adrenocortical activity. Today it is generally accepted that unpredictability is most effective in stimulating adrenocortical activity in a variety of situations. Correspondingly, if an individual is given information about the occurrence of an adverse stimulus, its predictability leads to a reduction of the adrenocortical response. One illustration of this is the study by Dess and associates (1983) on dogs that were subjected to a series of either predictable or unpredictable electric shocks. In the predictable condition the animals were presented with a tone prior to the onset of shock, while in the unpredictable condition, no tone was presented. Dogs that did not have the signal preceding the shock showed an adrenocortical response two to three times that observed in animals with the predictable shock experience. Furthermore, as shown by Mason, even subtle everyday changes in the environment, usually not considered as stressful, such as presence or absence of familiar persons in a room in which monkeys were kept in cages, can result in measurable changes of adrenocortical activity. These results suggest “that the central nervous system exerts a constant ‘tonicity’ upon this endocrine system, in much the same fashion as has been previously demonstrated for the autonomic and skeletal muscular effector systems” (Mason, 1968b). c. UnspeciJicity of the Stress Response. Mason also questioned the basic premise of the Selyean stress concept of a nonspecific response by an

8

DIETRICH VON HOLST

organism to many different stimuli or agents (stressors). Instead, he considered the Selyean stress response to be a specific physiological response to its corresponding psychological reaction, which is probably induced by the different Selyean stressors. Whether an animal is immobilized, subjected to unavoidable electric shocks or extreme temperatures, or whether it is forced to swim to exhaustion, it is always in a hopeless situation that is out of its control and that may be responsible for the adrenocortical activation (Mason, 1968b). The same opinion was held by Bush (1962, p. 321), who stated in a review: It is probable that very severe burns, and large doses of certain agents such as bacterial pyrogens, histamine, and peptones. cause a brisk release of ACTH that is independent of any emotional concomitants; but . . . severe exercise, cold, and fasting produce little or no effect on the secretion and metabolism of cortisol in man unless they are part of a situation that provokes emotion. (1962, p. 321)

d. Predictability and Control. As mentioned above, the typical Selyean stress response occurs in those situations that are characterized by uncertainty or unpredictability. Prolonged stress responses can incur a high biological cost, leading to a number of immunological, gastrointestinal, and cardiovascular changes that may reduce the vitality of the animals. Therefore, mechanisms have evolved whereby the animals can reduce excessive adrenocortical activation. The most important factor involved in reducing hormonal responses to adverse stimuli is control. Control can be defined as the capacity of an animal to produce active responses during the presence of an adverse stimulus. These responses may allow the animal to avoid or escape from the stimulus, but they may also provide the animal with the opportunity to change from one set of stimulus conditions to another, rather than to escape the adverse situation entirely. In both cases control reduces an animal’s physiological stress response. Particularly impressive support for this is provided by the research conducted by Jay M. Weiss on the development of gastric ulceration in laboratory rats (Weiss, 1972). In one of the earliest experiments, two rats were restrained in an apparatus for 21 hours with identical electrodes on their tails attached to the same shock-delivering device. Every minute a tone was presented to the rats for 10 s, which was followed by a light electric shock. One of the animals (“avoidance-escape rat”) was given the possibility of avoiding the shock by touching a panel with its nose during the presentation of the signal, or of terminating it after the beginning of the shock; the other rat had no possibility of influencing the shock outcome (“yoked rat”). Every time the avoidance-escape rat received a shock the helpless yoked rat was given exactly the same shock. Thus, the two animals received exactly the same physical stressor, but they differed in their control over

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

9

the situation. The two most important results of this study were (1) Simply receiving shock itself is not in particular responsible for the production of ulcers, but rather whether or not the animal is able to control the shock. The helpless yoked rats developed much more ulceration than did their partners; and (2) The more often an animal terminated the signal andlor the shock by its behavior, the less ulceration developed. That is, animals that can exercise control over a stressful situation do receive the relevant feedback when they respond by getting the information that they are “doing the right thing.” This is never the case in the helpless yoked animals. These results led Weiss to the conclusion that the most important aspect of an animal coping in a stressful situation is whether or not it can predict the consequences of its behavior. This conclusion was elegantly confirmed by experiments in which avoidance-escape animals with control over shock were given a brief electric shock every time they performed the previously correct response. Thus, each avoidance-escape response now produced precisely the wrong kind of feedback stimulus, a shock. In this “negative” feedback situation, the animals developed even more ulceration than did their helpless yoked partners (Fig. 2). The results of this research have been confirmed many times over by experiments designed to embrace endocrinological parameters and carried out on other species: Animals that are allowed to control the stimulus or

I

t-7

Nonshock

Signal

p < .05

Signal + punishment

FIG.2. Length of gastric lesions (medians) of nonshock, avoidance-escape, and yoked groups of rats exposed to shock pulses that were preceded by a warning signal (lefi) and of groups that perceived a shock pulse whenever they performed an avoidance-escape response (right). Significant differences between the two shock groups and nonshock groups are indicated: *p < .05; **p < .01; ***p < .001. For details see text. Adapted from Weiss (1971). with kind permission from American Psychological Association, Washington, D.C.

10

DIETRICH VON HOLST

situation show less (and in some cases no) physiological stress responses (e.g., glucocorticosteroid levels not different from those of undisturbed controls), whereas their yoked counterparts exhibit extremely high levels of glucocorticosteroids and other signs of stress (e.g., Davis et al., 1977; Hanson et al., 1976; Seligman, 1975; Weiss, 1984). Hence, current opinion links Selye’s stress response or the activation of the pituitary-adrenocortical system to psychological processes, resulting from uncertainty to loss of control and helplessness.

3. Active and Passive Stress Responses An important modification of the original Selyean stress concept was made in 1977 by James P. Henry and Patricia M. Stephens. In their monograph “Stress, Health, and the Social Environment” they summarized the results of zoological, psychological, sociological, and medical research into stress and concluded that two independent chronic stress reactions needed to be distinguished from each other: active and passive stress. The central theme of this concept of two different stress responses is the relationship between styles of coping; limbic (emotional) processes and neuroendocrine stress responses. Every threatening stimulus or challenge to control immediately induces Cannon’s fight or flight response, followed within a few minutes by adrenocortical activation as the animal makes a behavioral effort to ensure that control over a conspecific or a situation is retained. If control is not possible, different types of coping are seen in nonhuman animals and humans alike, which clearly differ behaviorally and physiologically (Henry, 1986, 1992). a. Active Chronic Stress. If an animal reacts with a style of coping characterized by active attempts to control the situation, for example, by fighting to maintain or defend a social position or a territory or by fleeing to avoid the situation, Cannon’s sympathetico-adrenomedullary system is chronically activated; the activity of the adrenocortical axis can, but may not necessarily, be increased in this response. According to this concept, this active stress response is characterized by subjective feelings of anger or fear, depending on the context. Chronic active stress or the constantly heightened sympathetico-adrenomedullary activity may lead to arteriosclerosis and cardiovascular diseases. Recent studies have even shown a distinct response pattern, activated by the brain in differing emotional states within the sympathetico-adrenomedullary system. The neurosympathetic outflow of norepinephrine, the “fight hormone,” can be independently activated by the “flight hormone,” epinephrine, that is released from the adrenal medulla (Hucklebridge et al., 1981; de Boer et al., 1990). b. Passive Chronic Stress. When active coping (e.g., by flight) is not feasible, a state of helplessness emerges, characterized mostly by immobility

11

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

and symptoms indicative of depression. This passive stress response is characterized by greatly enhanced activity of the pituitary-adrenocortical system, while the activity of the sympathetico-adrenomedullary system remains more or less unchanged (Fig. 3).

Perceived stimulus

Threat to cont

ss of control

Active behavior @ -,

Passive behavior

(Fight -flight)

(Nonaggressive)

Behavioral arousal with challenge to status

Inhibition of spatially organized behaviors and status

Defense reaction Territorial or status control with mobility, display, and

Defeat reaction Loss of territorial or status control, low sex drives

I

I

YAnger")

Control

Striving ("Fear")

Loss of control

("Depression")

Norepinephrine Epinephrine

t

t

tc

tc

t

cf

Glucocorticosteroids

cc

t

Sex hormones

t

4

t

4

FIG. 3. Schematic diagram contrasting the active and passive stress responses. The sympathetico-adrenomedullary system is divided into two branches: one of fight, anger, and norepinephrine; another of flight, fear, and epinephrine. Adapted from Henry et al. (1995). with kind permission from Lippincott-Raven Publishers, Philadelphia.

12

DIETRICH VON HOLST

c. Coping Behavior and Appraisal. These two styles of coping depend on the appraisal of the situation by the animal and on the quality and/or duration of the stressor. They thus may represent alternative strategies to the solution of a problem. For example, to avoid the constant attacks of a dominant rival, active escape may in the short term have the same result as hiding quietly in the corner of a cage, especially if escape is impossible. Furthermore, van Oortmerssen and associates (1985) demonstrated in a study of wild house mice (Mus rnusculus) that the two coping patterns may play different roles depending on the social structures and dynamics of populations: Aggressive mice do better in settled stable demes, whereas nonaggressive mice fare better in growing colonies. Appraisal of a stimulus or situation as well as the resulting coping behavior are basically psychological processes. There are, therefore, no clear relationships between stimuli imposed on individuals and their physiological responses. It is the behavioral, psychological, and thus the physiological responses of individuals to stimuli that differ depending on their genetics, prenatal influences, and especially postnatal learning processes (e.g., Fokkema et al., 1988; Henry et al., 1993). A striking example for the significance of social experiences on stress responses is provided by the work of Sachser and associates on guinea pigs (Cavia aperea f porcellus). Male and female guinea pigs can be kept in large groups without any behavioral or physiological signs of stress. When two adult males from different colonies are confronted in an experimental arena in the presence of an unfamiliar female, they arrange themselves in a dominance order within a short time in the absence of any serious fighting (Sachser, 1986; Sachser and Lick, 1991). However, if two males, each reared with only a single female, are confronted in the same way, both display continuously very high levels of aggressive behavior and extreme stress responses and die within a few days unless separated (Sachser and Lick, 1989). The ability to come to a peaceful arrangement with conspecifics is dependent on social experiences with male conspecifics around and shortly after puberty (Sachser, 1993). A few low-key confrontations with an unfamiliar male, introduced into their enclosure 5 times for 10 min between the age of 90 and 138days, were sufficient to reduce the fights with unknown males in later chronic confrontations to the same low levels of animals raised in colonies. That is, only 50 min of aggressive experience around puberty is required to enable adult male guinea pigs to come to a stressfree arrangement with conspecifics (Sachser et al., 1994). The crucial role of social experiences for behavior and stress responses was confirmed in a further approach (Sachser and Renninger, 1993). Colony- and individually reared males were singly introduced into unfamiliar colonies of conspecifics for a period of 30 days. Colony-reared males

13

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

easily adjusted to the new social situation: On the first day they only explored the new environment but did not court any female, thereby avoiding attacks from the male residents. In the course of the following days they gradually integrated into the social network of the established colonies. Changes could not be determined in either their body weights or their plasma concentrations of cortisol, androgens, or catecholamines. In contrast, individually reared males were involved from the beginning in courtship behavior, threat displays, and fighting. As a consequence, they responded to the new situation with substantial body weight loss as well as with extreme increases in plasma cortisol levels (Fig. 4). These data from guinea pigs clearly demonstrate the causal relationship between social experience around puberty, behavior of the individuals as adults, and the degree of their stress in unfamiliar social situations. d. General Physiological Response Pattern, Physiological studies of recent decades have revealed that many, if not all, neuroendocrine systems respond to stressors. In his comprehensive treatise on motivation and emotions, Buck (1988b) differentiates between behavior involved in selfpreservation (offensive and defensive behavior) and behavior concerned with reproduction. These are accompanied by the arousal of different parts of the limbic system and the hypothalamus as well as by patterns of neuroendocrine response, each peculiar to the particular emotion involved. Apart from modifications in the sympathetic nervous system and the adrenal

700

1

I I

Males raised -f in colonies -0 individually

-

m

s

m -

a

.500 E

c 0

100

-1 0

-5

0

5

10

15

20

Days before and after introduction into an unfamiliar colony FIG. 4. Plasma cortisol levels ( M 2 SEM) of 6 colony- and 6 individually reared male guinea pigs before and after transfer into an unfamiliar colony. Significant differences between the two groups: **p < .01; ***p < ,001. Adapted from Sachser and Renninger (1993), with kind permission from II Sedicesimo, Florence, Italy.

14

DIETRICH VON HOLS’I

glands, examples are to be found in the modification of neuroendocrine systems that are involved in the regulation of reproduction (e.g., folliclestimulating hormone [FSH], luteinizing hormone [LH], testosterone, estrogen, prolactin), in metabolism (e.g., growth hormone, thyroid-stimulating hormone [TSH], thyroxine, insulin), in osmoregulation and regulation of blood pressure (e.g., aldosterone, vasopressin, renin), and in immune response. In accordance with the Selyean hypothesis, it appears that gonadal activity is always inhibited by passive stress, whereas, dependent on the context, active stress can have an inhibiting or activating effect (see Section 111,BJ). However, the immune system appears, at least in the long term, to be more or less inhibited by all stress reactions. Divergent hormonal patterns probably have differential effects on the function of the immune system. Current knowledge is insufficient regarding the other systems and prevents any general statements on their participation in a given stress reaction or on their long-term effects. Future findings will most certainly lead to further differentiation of the Henry-Stephens concept, particularly regarding the participation of the immune system in stress reactions. However, this concept of two independent stress axes has proved durable in zoology as well as in medicine and psychology over the past 20 years (e.g., Bohus et al., 1987; Henry and Meehan, 1981; von Holst, 1986a,b; Lemaire et al., 1993; Lundberg and Frankenhaeuser, 1980; Mormkde et al., 1990). 4. Stress-A

Useful Concept for Behavioral Research

In a very general form, Selye (1952) defined stress as “a non-specific response of the body to any demand made on it.” It is only in this sense that the term stress can be employed usefully today. In contrast to the original Selyean assumption, “nonspecific” must be interpreted as those reactions triggered within the body that are not a result of peripheral changes (e.g., a drop in blood sugar content or blood pressure) and, therefore, represent correction mechanisms of homeostatic processes. These are neuroendocrine response patterns induced by the central nervous system, which change the organism’s physiological state and thereby generally lead to its activation. These neuroendocrine stress reactions differ depending on the situation as well as the behavior of the animals and concomitant emotional processes. This definition of stress in no way implicates definitive reaction patterns or the participation of specific endocrine systems. However, it does assume physiological reactions that could be detrimental to the individual if they reached sufficient intensity or were of long duration. Social stress or psychosocial stress describes the state of an animal, in which interactions with

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

15

conspecifics trigger central nervous processes, which themselves induce physiological reactions that can lead to detrimental effects on the animal’s vitality in the long term. Reduced fertility is not automatically also a consequence of these stress reactions. As the underlying factor in stress reactions is usually to be found in emotional processes, a given situation or stimulus can vary in its effect from one individual to the next by acting as an extremely damaging stressor, a harmless influence, or even a positive trigger. Hence, the recording of physiological stress reactions is-independent of the relevance of these reactions to diseases-a methodological basis used to rate the appraisal of a given situation by an individual and, therefore, to evaluate the biological significance of social interactions and situations as well as that of nonsocial influences (such as climatic factors or housing conditions).

C. ASSESSMENT OF STRESS

1. General Methodological Considerations Basically the effects of stressors can be measured on two levels, which, of course, are not exclusive. a. Epidemiological Approach. This approach is often taken in medicosociologically oriented research; the actual aim of research is to clarify the role of stressors in the development of malfunctions and diseases, from cardiovascular, gastrointestinal, or renal diseases to tumor growth and infertility. These mainly epidemiological studies usually focus on a few variables relevant for the respective organ or system, without paying much attention to the underlying physiological (regulatory) mechanisms (e.g., Ader et al., 1991; Adler et al., 1986; Dohrenwend and Dohrenwend, 1974; Friedman and Rosenman, 1974; Levi, 1971; Price, 1982). b. Physiological Approach. This approach mainly investigates those structures of the central nervous system involved in stress reactions and their effects on peripheral neuroendocrine and immunological parameters. It is not the aim of this chapter to discuss those central nervous structures involved in controlling neuroendocrine processes, but rather to focus on changes in peripheral parameters. These are of particular importance, as they not only indicate the presence of a stressful situation, but also allow limited statements on possible pathophysiological consequences of the situation for the individual. Hence, this is the preferred approach in research into stress (for details, see Section 111,B). A multitude of very different physiological parameters must be assessed, as individuals can react to the same stimuli in very different ways. However, even today this is not the case in most studies. Measures are usually selected on the basis of methodological constraints rather than based on present

16

DIETRICH VON HOLST

knowledge, which makes interpretation of apparently contradictory results often difficult or impossible. Although the determination of many different hormones, immunological mediators, and other clinically relevant parameters present few problems today, the interpretation of data is complicated by numerous possibilities for methodological errors. These are described briefly in the following section (for details, see textbooks of endocrinology). 2. Methodological Problems a. Animal Housing and Handling. Every organism responds to challenges with arousal responses of varying intensity (acute stress response). Moving an animal that is accustomed to a specific laboratory environment into a new cage or an unfamiliar room is sufficient disruption to act as a strong stressor for hours (e.g.. laboratory rats: Schuurman, 1981) or even days depending on the species. This can result in a total masking of actually interesting social situations or interactions. Thus, in tree shrews, transfer into a new cage within an experimental room results in increased serum levels of glucocorticosteroids and catecholamines for about 1week. Interestingly, serum levels of testosterone follow a biphasic pattern; after an initial decrease over a few days, the testosterone levels increase significantly above initial levels as the animals become habituated to the new situation (Fig. 5). This same biphasic testosterone response was found already in 1968 by Mason and co-workers (1968a) in rhesus monkeys during and after a 3day avoidance session. All handling of animals (e.g., weighing or the taking of blood samples) also functions as a stressor that acts on the corresponding variables within

-a, -a,

150

12 males per sampling point

u)

Q

125 .*0

..

-0-

* *

75

-: 0

--c

COrtlSOl

- Norepinephrine

.-m.

Testosterone

5 10 15 Days after transfer into a new room

20

FIG. 5. Some endocrine responses of male tree shrews after a transfer into an unfamiliar room. Significant differences to initial values: * p < .05.

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

17

seconds (catecholamines), minutes (glucocorticosteroids, thyroid and gonadal hormones), or a few hours (some immunological parameters); the repeated taking of blood samples on consecutive days can therefore lead to extensive stress reactions. After withdrawal of larger quantities of blood, a physiological stimulus is added to the psychological one, leading among other things to a sustained stimulation of the sympatheticoadrenomedullary system of variable duration. Finally, even in laboratory conditions, changes in climatic conditions or in food and introduced (nonvirulent) germs are capable of influencing many different endocrine as well as other parameters that indicate the presence or absence of stress. These problems can largely be avoided by standardized laboratory conditions. b. Diurnal and Other Variations of Physiological Parameters. There are a number of other problems associated with the measurement of hormone levels. In the majority of hormones secretion is not continuous, but occurs in a pulsatile fashion. The pulse amplitude is usually highest at the beginning of the activity period and lowest at the end, resulting in a marked diurnal rhythmicity of the hormone output (Fig. 6). Due to this pulsatile secretion, the hormone concentrations in the blood can change by a factor of 10 or more within minutes. Hence, baseline values exhibit much intra- and interindividual variation, even if great care is taken to exclude all potential interfering factors. This prevents interpretation on the level of the individual of most endocrine parameters based on single blood samples. Although it is possible to obtain blood samples from larger laboratory animals, over several hours and up to a few days, by insertion of cannulas into the blood vessels, this method is generally stressful to the animals, inhibits freedom of movement, and is therefore only of limited application in laboratory experiments (Fagin et al., 1983; Schuurman, 1981). Furthermore, many physiological parameters also show annual rhythms or other periodicities, which may influence hormone values. It is possible to circumvent the general influence of such rhythms in controlled laboratory conditions by, for example, always taking blood samples for endocrinological investigations at the same time of day. Nevertheless, because of the pulsatile secretory pattern of hormone release, marked individual variations will remain. Furthermore, no information is available on the influence of changing day length and other naturally occurring factors on such daily rhythms and, therefore, how comparable such values are even if they are collected at the same time each day. It is also more or less unknown whether stressful situations lead always to the same changes in the levels of different parameters during the day or not. A case in point are our investigations carried out on tree shrews with the aid of telemetry, which reveal a particular increase in heart rate during the night (sleep

18

DIETRICH VON HOLST

Body temperature ("C)

Heart rate (beatshin)

41

360

39

280

37

20 0

35

120

Triiodothyronine (nglrnl serum)

Glucose (mg/100 ml blood)

06

140

05

120

04

100

03

80

Cortisol (ng/ml serum)

Testosterone (ng/ml serum)

12

24

0

18

4

12

0

6 0

4

8

12

Time

16

20

24

0

4

8

12

16

20

24

Time

FIG. 6. Diurnal rhythms of some physiological parameters in male tree shrews. Night periods are characterized by dark color. Depending on the blood parameter, each point represents the mean (t SEM) of 20-80 males: heart rate and body temperature are hourly means ( 2 SEM) measured with implanted radio transmitters from 12 males and females.

periods) if the animals are in a stressful situation, even though they appear to be sleeping normally (e.g., Figs. 20 and 40). In summary, in order to gain relatively reliable data on endocrine and other physiological processes on the individual level, it is necessary to collect these data only on individuals that are kept in a controlled laboratory environment; in natural conditions in the field it is usually possible to detect only strong effects.

3. Physiological Markers of Stress In this section, I shall briefly discuss those methods that appear to me to be the most important or those that are most commonly employed in assessing the level of stress in an individual, as well as their application

STRESS A N D ITS RELEVANCE FOR ANIMAL BEHAVIOR

19

and power. To begin with, those systems will be briefly introduced that are necessary to understand the choice of variables that are measured. a. Pituitary-Adrenocortical System. Each of the paired adrenal glands of higher vertebrates is composed of two distinct and functionally different tissues-the adrenal cortex and the adrenal medulla. The cortex forms the outer part of the adrenal gland and consists of three zones: the outer glomerulosa; the zona fasciculata; and the inner zona reticularis, all of which produce large numbers of different steroid hormones. Glucocorticosteroids. For the present discussion, the hormones of the zona fasciculata are of special relevance, as they are released immediately in stressful situations. The most important and biologically relevant hormones are cortisol and corticosterone, the presence of which varies from species to species. For example, it is corticosterone almost exclusively that is found in the blood of rats and mice; in primates and guinea pigs cortisol is predominant; whereas other species (such as rabbits, hamsters, or tree shrews) exhibit both hormones, although they are liable to differential change during stress (e.g., rabbit: Kass et af., 1954; Krum and Glenn, 1965; hamster: Ottenweller etal, 1985). Because of their strong effects on carbohydrate and protein metabolism, all hormones of the zona fasciculata are grouped together as glucocorticosteroids: They increase the production of glucose from protein resources, and this is then stored in the liver as glycogen (a process referred to as gluconeogenesis), thus increasing the available glucose necessary for energetic processes during stress. Furthermore, they inhibit inflammatory processes and suppress many immunological responses by directly acting on receptors of the thymus and blood cells. Finally, glucocorticosteroids are required for the action of catecholamines such as for the induction of vasoconstriction by norepinephrine (e.g., Beato and Doenecke, 1980; Munck et af., 1984). Long-term increased glucocorticosteroid levels selectively reduce glucocorticosteroid receptors in the hippocampus (Brooke et af., 1994). Furthermore, high levels of glucocorticosteroids, such as are found in individuals suffering from chronic stress, are known to cause severe dendritic atrophy. This atrophy is particularly notable in hippocampal neurons and may contribute to the cognitive impairment found in persistently challenged individuals (e.g., Aus der Muhlen and Ockenfels, 1969; Magarinos et af., 1996; McEwen et al., 199.5; Uno et af., 1994; Sapolsky, 1991, 1992) (Table I). The synthesis and release of glucocorticosteroids are controlled by the pituitary hormone ACTH (adrenocorticotrophic hormone), which itself is controlled by the hypothalamic corticotrophin-releasing hormone (CRH). In emotionally induced stress reactions the release of corticosteroids appears to be largely controlled by CRH, whereas physical pressures can result in an increase in ACTH and, therefore, also in an increase in glucocor-

20

DIETRICH VON HOLST

TABLE I ACUTEA N D POTENTIAL LONG-TERM EFFECTS OF GLUCOCORTICOSTEROIDS Elevated levels of glucocorticosteroid hormones Acute effects Chronic effects Mobilization of energy (Gluconeogenesis) Lipolysis (synergistic with catecholamines) Raised muscle contractibility (permissive to catecholamines) Sodium retention and diuresis Release of calcium from bones Elevated release of hydrochloric acid and pepsinogen in stomach Antiinflammatory and immunosuppressive actions Suppression of gonadal activity ?

Neural responses, including altered cognition and sensory threshold

Loss of muscle mass, fatigue, steroid diabetes Arteriosclerosis Hypertension Hypertension Osteoporosis Ulcerat ion Decreased wound healing, increased disease susceptibility Decreased sexual behavior, sterility Dendritic atrophy (especially of hippocampal neurons) Psychoses and depression

ticosteroids through other mechanisms (such as through direct action of interleukin 1 during infections or through vasopressin during disturbances of the electrolyte balance: e.g., Aguilera et al., 1992; Berkenbosch et al., 1992; Brown, 1991; Dallman, 1991; Dempsher and Gann, 1983; Lilly et al., 1983; Rivier, 1991; Smelik and Vermes, 1980) (Fig. 7). While the release of glucocorticosteroids is usually controlled by ACTH, an additional possibility for the modification of the adrenocortical activity that has so far largely been ignored, is by its innervation. Henry and associates (1976) discussed the morphological and physiological evidence and presented their own data indicating that the activation of the adrenal cortex in dominant and aggressive fighting mice is due to direct sympathetic nervous stimulation, while in subordinate and repeatedly defeated mice the normal hormonal ACTH pathway is involved (see also Hucklebridge et al., 1981). Apart from its effect on the adrenal cortex, ACTH acts directly on the central nervous system. This indicates that ACTH may play an important role in the establishment and maintenance of social hierarchies, as was pointed out by Brain and Poole in 1974: Injection of ACTH suppresses defensive fighting behavior in mice pitted against trained fighters. In addition, acquisition of both actively and passively conditioned avoidance responses is enhanced by ACTH and the disappearance of these responses

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

21

-u=Stressor

Interleukin 1

I

Epinephrine

]

Vasopressin

I

Adrenal cortex

I

I

I Effector cells I FIG. 7. Schematic diagram of the hypothalamo-pituitary-adrenocorticalaxis. Stimulating (+) and inhibiting (-) influences are indicated.

is delayed. On the basis of these results, Brain and Poole proposed that subordination in a dominance hierarchy may be a form of conditioned avoidance response, which causes subordinates to avoid further attacks by dominants either by fleeing or by signaling subordination. Corticosterone treatment has no apparent effect on offensive, aggressive behavior, but increases submissiveness in mice. Evidence for this is found in the occurrence of “the rigid upright posture” and the failure to defend themselves when attacked by an opponent (Leshner, 1981; Leshner and Politch, 1979; Leshner et al., 1980; Politch and Leshner, 1977). The authors

22

DIETRICH VON HOLST

conclude that, whereas ACTH may be important in the regulation of aggression, corticosterone regulates submission. As already mentioned, the influence of stressors induces an increased production and release of glucocorticosteroids. Long-term stress can therefore lead to ACTH-induced hypertrophy and hyperplasia in cells, resulting in a substantial enlarging of the zona fasciculata and hence of the entire adrenal gland. Minerafocorticosteroids. The second group of adrenocortical hormones are produced in the zona glomerulosa and are called mineralocorticosteroids after their function. The only physiologically relevant hormone is aldosterone, which affects sodium reabsorption in the distal tubuli of the kidneys and is hence involved in water and electrolyte metabolism. Its secretion is regulated by several factors (mainly by the concentrations of potassium and/or angiotensin I1 in the serum). Although a participation of aldosterone in stress responses (“conditioned anxiety”) has been demonstrated in rhesus monkeys, the direction of the initial change varies between the animals (Mason et al., 1968b). Because such studies on aldosterone and stress are few and far between, it will not be considered here. Sex steroids. The third group of adrenocortical hormones are sex steroids, particularly androgens such as dehydroepiandrosterone and androstenedione, which are normally released in considerable amounts by ACTH. However, in certain states (e.g., puberty, aging, and stress) there is a divergence between the stimulation of cortisol release on one hand and adrenal androgens on the other, which indicates the additional release of adrenal androgens by other, probably pituitary, factors (for details, e.g., Labhart, 1986). Compared to the biologically relevant testicular androgen testosterone, the biological effectiveness of adrenal androgens is very weak and little is known about their physiological role under normal conditions. However, the possibility cannot be ruled out that female reproduction (inclusive of fetal development) may be impaired by increased androgen concentrations in stressful situations (for a recent review, see Collaer and Hines, 1995). In contrast to the mineralocorticoids, glucocorticosteroids and sex hormones in the blood are mainly bound to transport proteins (corticoidbinding protein (CBP) and albumin) and free and bound fractions are at equilibrium. Only the free fractions exhibit biological activity. Concentration of these transport proteins is variable (e.g., increase during pregnancy: decrease during starvation). Assessment of adrenocortical activity

In the simplest case, changes in the size and weight of the adrenals can be useful to infer changes in activity. Adrenal weight and histology were the first parameters used to asess the

A D R E N A L G L A N D WEIGHT A N D HISTOLOGY.

STRESS A N D ITS RELEVANCE FOR ANIMAL BEHAVIOR

23

extent of adrenocortical activity by Selye and other scientists up to the end of the 1950s. Its usefulness is restricted, as the animals have to be killed to gain access to the organs and it is not possible to follow responses of the adrenocortical system in animals on the individual level. Even so, weighing is still often the only way of assessing adrenal activity in smaller animals in the wild, particularly as the weight of the adrenals is not affected by the capture and killing of the animal. Adrenal weights do not, however, provide information either on current hormone levels or on short-term changes in adrenal activity, as changes in size require several days of continual stress. Therefore, weights are more indicative of the adaptive state of the adrenal cortex. In addition, they provide little more than semiquantitative indications of current hormone concentrations. In many cases of less recent laboratory and field research on small rodents no absolute adrenal weights were given, and only results of calculations relative to the body weights of the animals were supplied. These relative adrenal weights are aimed at highlighting developmental differences between individuals and should compensate for differences in the size of their organs. In my opinion, these values are not, however, satisfactory measures of the adrenal activity in individuals, as body weights are particularly prone to rapid change if animals are stressed. Although relative adrenal weights do not allow any conclusions as to the adrenocortical activities or serum glucocorticosteroid concentrations in animals, increased relative adrenal weights do indicate stress. Chemical, histological, or histochemical studies of adrenal glands, as used to determine adrenocortical function in the initial research into stress, have since lost all relevance because of the development of direct methods in the determination of hormone concentrations. HORMONE MEASUREMENTS. The direct measurement of glucocorticosteroid concentrations in blood samples (serum or plasma) by radioimmunoassays and other methods is quite simple. Since, however, glucocorticosteroid concentrations increase after only 3 min subsequent to the beginning of the blood sampling procedure, “true” baseline levels can usually be obtained only under laboratory conditions. In the controlled laboratory environment, this methodology is applicable to assessing the effects of social and other stimuli on adrenal activity in ) individuals, as the necessary blood sample size is so small ( 4 0 ~ 1 that, even in small animals with body weights far below 100 g, blood sampling at 1- to 2-day intervals over long time spans is possible without detrimental effects due to blood loss. As mentioned previously, though, depending on the species and its emotional reaction and resulting psychological processes, an insufficient time lapse between each blood sampling procedure may result in typical stress responses with heightened glucocorticosteroid levels

24

DIETRICH VON HOLST

in the serum. For example, regular blood sample collection at 7-day intervals over several months induces no quantifiable physiological changes in tree shrews, whereas sampling at 4-day intervals o r less induces clear stress reactions after only two to three blood sample collections. One largely neglected aspect that may be particularly relevant to stress research is the relationship between free and protein-bound hormone levels. In the majority of studies only the total amount of the hormones (bound and unbound) is determined. As mentioned earlier, the biologically active fraction of the glucocorticosteroids are the free hormones: They affect tissues and regulate the release of glucocorticosteroids from the adrenal glands by their negative feedback effects on hypothalamic and hypophyseal structures. The concentration in the blood of these proteins, that bind to and transport the hormones, is usually restricted and can be saturated by increased hormone concentrations. Dependent on the concentration of transport proteins in the blood, this means that a small increase in total hormone concentration can lead to a substantial increase in the concentration of biologically active free hormones, as shown in laboratory mice (Bronson and ElefthCriou, 1965a). However, there appear to be substantial interspecific differences: Serum concentrations of both cortisol and corticosterone in tree shrews in acute stressful situations can rise by factors of 4-5 within 30 min, without affecting the ratio of free to bound hormones (correlation between initial values and stress values of free and proteinbound glucocorticosteroids is always >.90). On the other hand, a decrease in concentration of transport proteins, as the consequence of a glucocorticosteroid-induced general protein mobilization, body weight loss during stress, and/or as the consequence of increased testosterone levels, can increase the free hormone levels, although the total concentration of hormones remains the same or even decreases (e.g., Blanchard et al., 1993; Bradley et af., 1980). In mammals, free (non-protein-bound) hormones and their metabolites are largely excreted with the urine. The determination of the excretion rates of glucocorticosteroids (as well as those of other hormones) should therefore be especially appropriate in making statements on hormonal changes in mammals in stressful situations. The main limitation associated with the measurement of hormone levels in the urine is the considerable time lag between the appearance of the hormones in the blood and their excretion with the urine. Furthermore, the concentration of hormones in the urine varies according to the amount of urine produced, and both urine production and the drinking behavior of animals are influenced by stressors. Already in 1859, Claude Bernard reported the occurrence of oliguria in association with pain or emotional reactions, and the antidiuretic effect of emotional stimuli has repeatedly been confirmed by numerous subsequent

STRESS A N D ITS RELEVANCE FOR ANIMAL BEHAVIOR

25

studies on mammals including human beings. There are, however, also reports of diuretic responses to psychological stimuli (Mason et al., 1968b). It is not yet known what situations or specific stress responses are associated with these divergent effects on urine excretion. Conclusions on the activity of the adrenocortical (or other) endocrine systems, drawn from the concentrations of specific hormones or their metabolites in individual urine samples (e.g., collected at the beginning of the activity period of an animal), are therefore not particularly reliable. This problem is aggravated by the fact that reliable internal standards correcting for changes in urine concentration and/or loss of urine are not available. The creatinine concentration in urine samples is often used as an internal standard, but is subject to substantial change under stress conditions, as are all other urinary parameters, thus making it unsuitable as an internal standard. This means that statements on the activity of endocrine systems are usually possible only if the urine is collected quantitatively, throughout the entire day, in either one or several samples. With the exception of humans, this is possible only in laboratory conditions and requires the isolation of the animals. Very often the animals have to be kept in specialized urine collection or metabolic cages (or, in psychological research on monkeys, often in restraint chairs), which results in extreme restriction of freedom of movement and hence usually in stress reactions or the requirement of long habituation periods of the animals to the situation. This usually precludes any research into the effects of social interactions or factors on endocrine processes. In many species that can be kept in cages on lattices, the influence of specific social stimuli (e.g., sight or smell of rivals or sex partners or separation of mother and its infant) can be investigated by collection of the urine in a basin beneath the cage. As steroid hormones (such as the glucocorticosteroids and sex hormones) are not destroyed by delayed collection or by drying out, the urine can be collected at fixed time intervals (e.g., 24 hours), dissolved in distilled water and the hormone concentrations can be determined (Fenske, 1989). In this manner acute changes in hormone secretion due to stressors, as well as chronic effects of social input, can be assessed (e.g., Fig. 8). In field studies, some groups have also used concentrations of steroid hormones in samples of feces as indications of adrenocortical activity (e.g., Miller et al., 1991). Although the quantitative collection of feces of individuals is sometimes possible, even under natural conditions usually only single samples are collected (e.g., morning feces). The interpretation of these fecal hormone concentrations is subject to the same methodological constraints as those of single urinary samples, even when sampled at predetermined times of the day. For the past few years salivary steroid hormone levels have been used to analyze the stressful effects of different social situations. The hormone

26

DIETRICH VON HOLST

m c .-c

15

e3

10

-

c 0

2 L

Male 475 dominant

Daily confrontations and visual contact

5

0) P

s o

.-

c

-6

-4

-2

C C C

2

4

6

Days before resp. after confrontation FIG. 8. Daily urinary cortisol excretion of two male tree shrews that lived in a cage separated by a nontransparent partition. After habituation to the new cage the partition between the animals was removed on 3 days daily for 10 min, which resulted in slight fights and the establishment of a dominance order (C). For the rest of the days both animals were separated by a wire mesh partition to allow visual contact between the rivals. As evident from the figure, cortisol excretion increased in the subordinate male during the period of visual contact with its rival and returned to initial values after separation by a nontransparent partition, while the opposite was found in the dominant individual. Horizontal dashed line: Mean cortisol excretion during the 6 days before the confrontation.

concentrations in the saliva correspond in most species to those in the blood and are independent of saliva production. This method has many advantages compared to blood sampling procedures, as it is noninvasive, very fast (about 1 ml of saliva is needed), and it measures only the biologically relevant free (non-protein-bound) fraction of the hormones (RiadFahmy et af., 1982; Vining and McGinley, 1986; Wade, 1991). For these reasons, this method is widely used in psychological studies on humans and in some animal welfare studies on larger mammals, such as dogs or pigs (e.g., Beerda et al., 1996; De Jonge et al., 1996; van Eck et al., 1996; Ekkel et al., 1996; Kirschbaum and Hellhammer, 1989); recently Fenske (1996) showed that this method is also applicable to small mammals. In studies with guinea pigs he found a good correlation between saliva and plasma concentrations of cortisol, but, in contrast to studies in humans, not of testosterone. The reason for this discrepancy cannot be explained so far. CHALLENGE TESTS. Due to the time required to catch and handle the animals for blood sampling, in many cases it is impossible, even in the laboratory,

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

27

to obtain “real” baseline hormonal levels; therefore, some authors have used challenge tests to gain information on the state of the adrenocortical system. The most common challenge used in studies on small rodents in the laboratory is the “open field test.” Animals transferred to an open field respond to this situation with acute stress reactions, for example, an increase of glucocorticosteroid concentrations in their serum, which can be determined by taking a blood sample after a predetermined time (e.g., 15 rnin). As many studied have shown, animals living in a stressful situation show higher glucocorticosteroid levels in subsequent challenge tests than do control animals. The transfer of laboratory animals to an open field is not always a sufficiently strong stressor to elicit maximal release of glucocorticosteroids. Thus, following maximal stimulation of their release by high amounts of ACTH, serum cortisol values in male guinea pigs are about three times higher after 240 min than they are 240 rnin after the transfer to an open field (Sachser, 1994a). Nevertheless, there is a very good correlation between the cortisol values of the individuals in the open field and those in the ACTH test. This indicates that challenge tests, such as transfer to a novel room or cage, give a suitable measurement of an animal’s adrenocortical activity or secretory capacity. In the field, the procedure of catching the animals or of the anesthesia necessary for blood sampling in larger species has been successfully used as a standard challenge to determine the adrenocortical capacity of individuals. A particularly simple challenge for the determination of the secretory capacity of the adrenal cortex in tree shrews in the laboratory, which so far has not been used by other researchers, is the blood sampling procedure itself (von Holst, 1986b). To this aim the animals were brought in their sleeping boxes to a laboratory and blood samples were taken 1, 5, 15, and 30 min after the room had first been entered. Between sampling the animals were returned to their sleeping boxes, but remained in the laboratory for the entire period. This challenge test is a strong stressor for all tree shrews: Their heart rates are elevated for the entire experimental period and the levels of catecholamines, glucocorticosteroids, and glucose in the blood increase greatly. Shortly after the test and transfer to the home cage, heart rate and all other parameters return to the initial levels. On average, the cortisol values of control animals increase within 30 min from less than 10 ng/ml serum to approximately 60 ng/ml serum, with substantial differences being observed between the individuals (for examples, see Fig. 9). As long as the animals live under constant conditions, repetitions of these challenge tests after periods ranging from 1 week to several months result in almost identical (and individually different) values for all animals ( r > .92; p < .001;

28

DIETRICH VON HOLST

150

120

.-C

30 Male 21

d

0 ' , 0

10

20

30

Minutes after first disturbance

FIG. 9. Blood sampling challenge tests (BSCT): Adrenocortical responses of 3 males to 3 challenge tests separated by 3-5 months.

based on data from several experiments with more than 150 animals; see also Fig. 9). The blood sampling procedure elicits a maximal glucocorticosteroid release in tree shrews within the first 15 min, which cannot be further increased by injection of higher doses of ACTH. Accordingly, in vitro superfusion analyses of the adrenals from controls and stressed animals show an extremely good correlation between the in vitro cortisol production of the adrenals after maximal stimulation of their secretion through addition of ACTH to the superfusion medium, and the serum cortisol values obtained from the individuals earlier in a challenge test (Fig. 10). The individual differences of cortisol challenge values are therefore due to corresponding differences in the adrenal capacities of the individuals to synthesize and release glucocorticosteroids after stimulation. Chronic stress (e.g., transfer to a new room or a confrontation with a dominant rival) always leads to an increase in these challenge values by up to 200%. O n the other hand, the opposite is found when males habituate to a new room or become dominant in a confrontation (see also Fig. 22). It must be emphasized here, that an alteration of a challenge value must not be taken as an indication of an equivalent alteration of serum baseline levels of the glucocorticosteroid hormones. This is due to the fact that the secretory activity of the adrenal glands is dependent on the nominal value

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

29

120 1

Correlation coefficient r = .91; p<.oo1 v = 12.13 + 6 . 3 7 ~

20

7

0

,

3

,

,

6

,

,

9

,

,

12

,

,

15

,

I

18

In vitro cortisol secretion of adrenals (nglmin)

FIG. 10. Correlation between the serum cortisol values of 12 males 15 min after beginning of a BSCT and the in vitro cortisol release of their adrenals 15 min after ACTH has been added to the superfusion medium. The in v i m tests were performed 8 days after the challenge tests.

of the hormone levels as well as on the metabolism and clearance of the hormones, all of which are factors that may be influenced by stressors. Therefore, increased secretory capacities of the adrenal glands in animals are found especially during active stress, without simultaneously increased baseline adrenocortical hormone levels. In summary, while an increased adrenocortical capacity (as measured with a challenge test) does not necessarily have to be associated with increased levels of glucocorticosteroids in the blood, it is a sensitive measure of stress, which can also be used in field studies. b. Sympathetico-Adrenomedullary System. In contrast to the cortex, the adrenal medulla arises embryologically from neural tissue, and remains a functional part of the autonomic (vegetative) nervous system. It may in fact be considered as a specialized sympathetic ganglion that, on activation of the sympathetic nervous system, discharges epinephrine and norepinephrine directly into the blood (Fig. 11). Epinephrine in the periphery is derived primarily or wholly from the adrenal medulla; norepinephrine, however, may be secreted from the adrenal medulla, or its presence may be due to the overflow of neurotransmitters from the noradrenergic sympathetic nerves into the circulation. In rats and cats, for example, approximately 60-70% of the peripheral serum norepinephrine is derived from sympathetic nerves (Kvetnansky et af., 1979; Stoddard, 1991). Both the activation of the sympathetic nervous system, as well as the release of adrenomedullary

30

DIETRICH VON HOLST

Autonomic nervous system Sympathetic system

Parasympathetic syster

nal rd Cervical Hair follicle muscle Blood vessel

Thoracic

Lumbar

Sacral

FIG. 11. Schematic diagram of the autonomic nervous system with the effector organs of its two subdivisions-the sympathetic and the parasympathetic nervous system.

hormones, therefore act upon different bodily functions in very similar ways and generally increase the capabilities of the organism (Cannon’s fight or flight response): Hence the sympathetic nervous system and adrenal medulla are usually summarily referred to as the sympatheticoadrenomedullary system. Apart from its catabolic effects on peripheral organs, peripherally released epinephrine (but not norepinephrine) also affects the central nervous system by eliciting general arousal and subjective feelings ranging from restlessness to anxiety. Furthermore, there is also evidence that the release of epinephrine in situations eliciting fear enhances

STRESS A N D ITS RELEVANCE FOR ANIMAL BEHAVIOR

31

the learning of avoidance behavior, while high doses apparently produce the reverse effects (McGaugh and Gold, 1989). The biosynthesis of the catecholamines epinephrine (E) and norepinephrine (NE) from tyrosine progresses in several steps in both the noradrenergic neurons of the sympathetic nervous system and in the adrenal medulla; the first step-the conversion of tyrosine to dopa-is catalyzed by the enzyme tyrosine hydroxylase (TH), which is the pacemaking step of the catecholamine biosynthesis. For the subsequent two steps-the formation of dopamine and norepinephrine-enzyme levels are not usually limited. The final conversion of norepinephrine to epinephrine is catalyzed by the enzyme phenylethanolamine-N-methyltransferase(PNMT). PNMT activities in sympathetic neurons are very low, thus norepinephrine is the main end product of catecholamine biosynthesis (the transmitter), while in adrenal medullary tissue PNMT activities are high, resulting in epinephrine as the main adrenomedullary hormone. The TH levels in the medulla are controlled mainly by neuronal influences from the sympathetic nervous system (Thoenen et al., 1969; Ungar and Phillips, 1983). Repeated stimulation of catecholamine release in stressful situations leads to an adaptive increase of the TH levels in the adrenal medullary tissue, which in turn increases the capacity of the adrenal gland to synthesize and release its hormones. In laboratory rats, for example, there is a doubling of adrenal T H activities within less than one week of daily immobilization stress (Fukuhara et al., 1992; Kopin, 1980; Kopin et al., 1988). PNMT levels, on the other hand, are usually not influenced by sympathetic (splanchnic) nerves, but by the glucocorticosteroid hormones of the adrenal cortex, leading to a 50% increase under persistent stress (Ciaranello, 1978; Kopin, 1980; Wurtman and Axelrod, 1966). There are, however, differences in the regulation of PNMT activities even between different strains of rats, as has been demonstrated recently by Lemaire and colleagues (1993). According to these authors, the increase of PNMT activity in male Wistar rats, kept in unstable mixed-sex groups, is dependent on the activation of the sympathetic nervous system, as it can be completely abolished by denervation of the adrenals. Assessment of sympathetico-adrenomedullary activity. Each capture and handling of an animal for the collection of blood samples immediately activates the sympathetico-adrenomedullary system and triggers the release of catecholamines into the blood, thus making the collection of baseline values impossible without prior introduction of cannulas into the blood vessels (Stoddard, 1991). Therefore, with one exception (Fokkema, 1985), I am not aware of investigations dealing with catecholamine baseline values in animals in social situations.

32

DIETRICH VON

nouT

CHALLENGE TESTS. It appears that challenge tests are useful techniques to gain information on the adaptive state of the sympathetico-adrenomedullary system. As already mentioned, the adrenal medulla adapts to an increased production and secretion of its hormones by increasing the activity of its pacemaking enzyme, tyrosine hydroxylase. Accordingly, the stimulated adrenal gland increases its secretion of the two hormones epinephrine and norepinephrine. In our research on tree shrews, we use their emotional response to handling during blood sampling as the challenge. We found that catecholamine levels in the blood of tree shrews are maximally raised as rapidly as 20 s after the first disturbance. When we take several blood samples from an individual within a period of 1-15 min, serum catecholamine levels remain more or less the same, while differences of several hundred percent are found between different individuals. The mean value of the norepinephrine and epinephrine concentrations in several blood samples from a given individual taken during a challenge test is therefore considered to be an indication of the adaptation state of its sympathetico-adrenomedullary system. This suggestion is supported by the following results: Between repeated challenge tests separated by at least 1 week, the individually different values of norepinephrine (as well as of epinephrine) are highly correlated (Fig. 12). Stressful situations, such as transfer to a new cage or a confrontation with a rival, increase the norepinephrine and epinephrine values by up to

12 0

55

-

0

z 10

. F a E

f

.-C a,

r = .85

r = .91

2

2

6 8 1 0 1 2 NE (nglml serum)

4

2

3

4

5

6

7

TH (nmollh adrenals)

FIG. 12. Correlations between BSCT values of serum norepinephrine (NE) of male tree shrews at two sampling points separated by 8 days (left) as well as between serum norepinephrine values and adrenal tyrosine hydroxylase activities (TH) of animals. Unpublished data: after Hutzelmeyer (1987).

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

33

100% within 1-3 days. After habituation to the new situation (room or cage) or removal of the rival, challenge values approximate the initial individual levels (see also Fig. 5). Finally, a very good correlation is found between the norepinephrine and epinephrine challenge values of tree shrews and their adrenal tyrosine hydroxylase activities (Fig. 12). Thus, challenge levels of catecholamines are a relevant measure of the activity or adaptive state of the sympathetico-adrenomedullary system, which can be used in the laboratory as well as under more natural conditions in large enclosures, as has been shown in studies on tree shrews, guinea pigs, and wild rabbits. In guinea pigs, for example, catecholamine challenge values are clearly related to individual behavioral strategies in situations of social conflict: Plasma norepinephrine levels were significantly higher weeks before as well as directly after a 10-min agonistic encounter in offensive “fighters” compared to nonaggressive “nonfighters” (Sachser, 1987). Thus, norepinephrine responsiveness is of predictive value for the behavior of guinea pigs in contest situations. Similar results are also known from tree shrews (e.g., Fig. 16). Excretion rates of catecholamines and their metabolites are frequently used in psychosocial stress studies in humans, and less frequently in psychophysiological studies in monkeys, but not in behaviorally oriented studies on other animals. This has many reasons, some of which have been already mentioned (see discussion on glucocorticosteroid excretion). Even more important is, however, the fact that catecholamines and their metabolites are rapidly metabolized and/or destroyed after urination. To avoid degradation, the urine must therefore be strongly acidified, which makes its collection more or less impossible, except from animals kept in a restraint chair. ENZYME ACTIVITIES. As mentioned above, TH and PNMT levels in the adrenal medulla adapt to stimulation with an increase in their activities. The determination of these two enzymes as a measure of the activity of the sympathetico-adrenomedullary system was introduced by Henry and colleagues (1972; Kvetnansky et al., 1970) in their work on social stress in mice. In the meantime, this method is widely established and is being applied to many different animal species by other research groups. As the increased change in enzyme activity appears only several hours after the application of stress, these methods are equally applicable in the field and the laboratory. However, since the animals have to be sacrificed, the range of application of this method is restricted. HEART RATE AND BLOOD PRESSURE. As the sympathetico-adrenomedullary system induces strong physiological reactions in the organism, indirect measures are sometimes used as indicators of sympathetico-adrenomedullary activity. A particularly important measure in psychobiological research is the heart rate and-mainly for medical purposes-blood pressure. Both

34

DIETRICH VON HOLST

parameters usually increase rapidly with every activation of the sympathetic nervous system. While the heart rate is activated by the sympathetic nervous system (including the catecholamines of the adrenal medulla) and inhibited by the parasympathetic nervous system, and correspondingly changes in both parts of the autonomic nervous system can influence heart rate, blood pressure is activated only in the short term by the sympatheticoadrenomedullary system and the renin-angiotensin system. Chronic emotional responses, however, can result in structural (arteriosclerotic) changes of the cardiovascular system, which may lead to persistent hypertension (Folkow et al., 1958, 1973). The heart rate is particularly well suited for the detection of acute activations of the sympathetico-adrenomedullary system in socially or otherwise stressful situations, and for the monitoring of chronic stressors and their potential pathophysiological consequences. Implantable radio transmitters have been used for many years to record heart rate telemetrically. Since the working life and range of transmitter systems depend on battery size, the weight of the transmitters usually determines their working life. An extremely small radio transmitter (weight including battery <1 g) with a range of above 30 m and a working life of 4-6 months has been developed at our institute (Stohr, 1988). This transmitter enabled us to record heart rate (and body temperature) continuously in animals differing greatly in size (Mongolian gerbils, laboratory mice, tree shrews, and wild rabbits) (Eisermann, 1992; Eisermann and Stohr, 1992). The weights of most commercially available transmitter systems used in stress research in nonhuman mammals are in the range of 5-10 g, with a lifetime of 1-2 weeks and a range of a few meters. This usually restricts their application to laboratory conditions and animals with body weights of over 200 g. The working life is also usually too short to allow any more than the acute effects of specific stressors to be recorded. This problem is exaggerated by the fact that, depending on the species and the size of the transmitters, a period of 1-2 weeks is necessary after implantation before the animals have recovered and stable, low levels of their heart rate can be recorded. In technically advanced systems, a magnetic on/off switch is used to transmit the signals only at certain times, which conserves power and may extend the working life up to several months. Blood pressure has mainly been measured in animals in chronic emotionally arousing situations, by direct arterial or indirect tail-cuff techniques at weekly time intervals. For both techniques, the animals must be handled and restrained before measurements can be taken. To reduce emotional stress responses, the measurements are often performed after slight narcosis. Nevertheless, these procedures elicit stress responses in the animals, which may influence the measurement. In laboratory animals, however,

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

35

the tail-cuff technique in particular has been proven to provide reliable results, when the animals are used to this procedure (Bunag, 1984). Over the past few years, the direct and continuous measurement of blood pressure has also become possible in small mammals using implanted radio transmitters; in one of the first studies using this method, Henry and co-workers showed parallel blood pressure changes in rats during chronic stress, measured by radiotelemetry and by indirect tail-cuff techniques. The latter, however, revealed levels 10% higher than the former (Henry et al., 1995). c. Pituitary-Gonadal System. Apart from the generative testicular and ovarian functions (the development of spermatozoa and ova, respectively), these organs also have endocrine functions. They produce sex hormones, the production of which is independent of spermatogenesis in males, but closely linked to oogenesis in females. Testes. The testes secrete hormones, which are collectively termed androgens, of which testosterone is the main physiologically active one. Depending on the species and organ, testosterone sometimes has to be converted enzymatically to dihydrotestosterone (17-DHT) in target cells in order to become biologically effective. Before birth, the testes of mammals release large amounts of androgens, which are responsible for the development of male-specific sex organs during fetal life, and for the organization of the central nervous structures responsible for the expression of male-specific behavior in later life (e.g., Breedlove, 1994; Collaer and Hines, 1995; Goy et al., 1988; Gustafson and Donahoe, 1994; Huhtaniemi, 1994; McCarthy, 1994; Phoenix et al., 1959; Turner, 1984). In the absence of the male sex hormones, female-specific sex organs and behavior patterns develop. These pre- or perinatal effects of androgens are usually characterized as “organizing effects” (Phoenix et al., 1959). Following puberty, testosterone (and DHT) are necessary for the differentiation and activation of spermatozoa during spermatogenesis, induce growth and function of accessory sexual glands, and modulate sexual, aggressive, and other testosterone-dependent behaviors. This occurs to a greater or lesser extent depending on the species (e.g., Arnold and Breedlove, 1985; Baum, 1992; Beach, 1975; Matochik and Barfield, 1991; Monaghan and Glickman, 1992; Thiessen and Rice, 1976). These peri- and postpubertal effects of androgens on morphology, physiology, and behavior are usually called “activating effects.” Furthermore, marking behavior and the secretory activity of glands, involved in marking or the characterization of the hormonal state of a male, are modulated by testosterone. Thus, decreased testosterone levels in a subordinate male can reduce the intensity of its aggression-eliciting signals and thus reduce attacks against it.

36

DIETRICH VON HOLST

Production and release of androgens by the testes is controlled by the hypothalamus. Gonadotropin-releasing hormones (GnRH) induce the release of follicle-stimulating hormone (FSH), which stimulates spermatogenesis, and luteinizing hormone (LH), which induces synthesis and the release of androgens from the Leydig cells of the testes. Ovaries. In contrast to the male androgens, female sex hormones have no organizational functions in mammals. Following puberty, dependent on the development of the follicles and subject to the influence of the two gonadotropins LH and FSH, the ovaries produce two main classes of sex hormones-estrogens and progestins, which are responsible in female mammals for their sex-specific morphology and physiology (e.g., Carter, 1992; Takahashi, 1990). Both classes of female sex hormones are also produced placentally in gravid females, as well as by the adrenal cortex in both sexes. Estrogens (especially the most effective estradiol) stimulate growth of the uterine wall and have a variety of other functions contributing to the maintenance of the condition of the female reproductive system. The biologically relevant progestin is progesterone, which is secreted in most mammalian species after ovulation by the corpus luteum of the ovary, and is necessary for maintaining pregnancy. In small rodents no corpus luteum develops during the estrous cycle and progesterone is synthesized by the interstitial tissue of the ovary. Estrogens (in many rodents in combination with progesterone) are responsible for female sexual receptivity, although their effect differs very much between species. In addition, estrogens can also enhance the attractivity of females by inducing the production of odors (sex pheromones) or other signals. Most stress situations apparently inhibit the release of GnRH, thus modifying fertility and the sex-hormone-dependent behavior of both sexes (e.g., Kime et al., 1980; Moberg, 1987; Orr and Mann, 1992; Rabin et al., 1988). Because of their specific receptors, the Leydig cells of the testes can also directly be influenced by glucocorticosteroid hormones (e.g., Evain et al., 1976; Stalker et af., 1989). Thus, 2 hr of immobilization induces a fall in serum androgen concentrations of rats without detectable changes in serum LH values. The chronic stress of a daily 2-hr immobilization for 10 days, however, results in decreased serum levels of both testosterone and LH (Maric et al., 1996). Correspondingly, in vitro corticosterone directly inhibits testosterone production by purified Leydig cells of rats, although only at high concentrations (Gao et al., 1996). In contrast to long-lasting stressful situations, acute stressors sometimes cause a transient elevation of plasma testosterone concentrations, despite the suppressed LH levels that precede the subsequent decline in testosterone levels. The reason for this effect has not been clarified yet, although increased testicular blood flow as a general consequence

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

37

of heightened sympathetico-adrenomedullary activity or of testicular nerves, as well as testicular pro-opiomelanocortin-derived peptides, may be involved in this function (e.g., Mann and Orr, 1990). Assessment of gonadal activity MORPHOLOGICAL AND HISTOLOGICAL DATA. As mentioned before, the testes have endocrine as well as generative functions that are not necessarily closely correlated. In order to obtain information on the generative functions, data collection in animal studies on social stress focuses on the weights of the testes, epididymides, and secondary sex glands, and in particular on histological investigations on the testes and epididymides for the proof of normal spermatogenesis, although the presence of intact spermatozoa does not rule out functional damage. In females, the weights of the ovaries provide a rather more superficial indication of their function; histological investigations on the ovaries, however, provide relatively differentiated information on maturing follicles, the number of ova released into the uterus, or impeded follicular maturation (e.g., increased follicular atresia). In addition, investigations on the uterus can provide information on implantation and also on abortions. All of these methods, however, are unsatisfactory in connection with studies on social stress and its consequences for the gonadal system, as they necessitate the sacrifice of the animals. HORMONE MEASUREMENTS.The determination of sex hormones in the blood, urine, or feces is possible in both sexes. The problems involved in the collection of these data (methods of blood sampling, time factors, methods of sample collection) are equivalent to those involved in the collection of glucocorticosteroids. Relatively reliable data on the influence of social and other factors on endocrine activity in the testes can be collected by blood sampling in males. In females, this is usually not possible except in extreme conditions, as the changes in hormone values are dependent on the stage of the cycle or pregnancy and daily hormone determination would be required in order to pinpoint stress-related changes, especially in animals with estrous cycles of a few days. This would be possible by determining selected hormones and/or their metabolites in the urine or feces. However, due to the many difficulties involved, very little information is as yet available on the quantitative influences of social and other stressors on the female endocrine system. Extreme changes, such as those during estrus and pregnancy can, however, be assessed by determination of hormones in urine or feces (e.g., Schaftenaar et al., 1992). d. Immune System. The immune system has two functional divisions: the innate (unspecific) system and the adaptive (specific) system. The innate immune system acts as a first line of defense against infectious agents

38

DIETRICH VON HOLST

(viruses, bacteria, fungi, and parasites) and clears them before they establish an overt infection. The adaptive immune system produces a specific reaction to each pathogen (antigen), which normally eradicates that particular agent. The adaptive immune system also remembers the infectious agent with the aid of B and T memory cells, causing an immunity of variable duration against the pathogen by inducing a much enhanced specific response at the next contact with this antigen. Both divisions of the immune system consist of a large number of humoral (molecules) and cellular (leucocytes) factors, which circulate with the bloodstream and are distributed throughout the entire body (Table 11). During infections, both systems are usually activated and combat the infectious agents in an integrated way: Following clonal activation after contact with an antigen, T lymphocytes produce several soluble molecules (cytokines), which stimulate the phagocytes to destroy the infectious agents more effectively, and also stimulate antibody production by the B lymphocytes. These antibodies then also help the phagocytes to recognize their targets. Depending on the antigen, the various parts of the immune system are differently involved in the response pattern. Although the immune system displays a certain degree of autonomy, the research of the last two decades has revealed multiple channels of communication between the central nervous system and the immune system. Emotionally stressful situations are particularly associated with altered immune function, and in some instances with altered health status, although these two processes have not been linked causally (e.g., Adams, 1994; Ader and Cohen, 1985; Ader et al., 1991; Dunn, 1989; Glaser and Kiecolt-Glaser, 1994; Kelley et al., 1994; Laudenslager and Fleshner, 1994; Monjan, 1981; Solomon and Amkraut, 1981). In his first publication, Selye (1936) described thymolymphatic involution as one of the most conspicuous signs of stress,

TABLE I1 MAJORCOMPONENTS OF THE INNATEA N D THE ADAPTIVE I M M U N ESYSTEM Innate (unspecific) system Humoral factors Lysozymes Complement system Acute phase proteins (e.g., Creactive protein) Interferons Cellular factors Phagocytes (polymorphs and monocytes) Natural killer (NK) cells

Adaptive (specific) system

Antibodies (produced by B lymphocytes)

T lymphocytes (e.g.. cytotoxic cells, helper cells)

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

39

and the connections between the hypothalamo-pituitary adrenocortical axis and the immune system are among the best examined so far (e.g., Dhabhar et al., 1995; Heinjen et al., 1991; Keller et al., 1991; Munck and Guyre, 1991; Zwilling, 1994). A strong immunosuppressive effect is also exerted by the catecholamines, but virtually every hormone investigated so far has some effect on the immune system (such as growth hormone, prolactin, and sex hormones), although not all of the effects are direct (e.g., Bernton et al., 1991; Grossman, 1984; Kelley, 1991; McCruden and Stimson, 1991; Olsen and Kovacs, 1996; Paavonen, 1994; Rabin et al., 1994). Furthermore, lymphoidal tissues (e.g., bone marrow, thymus, spleen, and lymph nodes) are innervated by sympathetic and sciatic fibers and hence the central nervous system can directly influence immune function (e.g., Ackerman et al., 1991; Felten and Felten, 1991; Madden and Livnat, 1991). On the other hand, the immune system is capable of modulating both neuroendocrine responses and the behavior of mammals by nervous influences from lymphoidal structures and by lymphokines produced by leucocytes, which act on hypothalamic and other central nervous structures (e.g., Anisman et al., 1993; Bateman et al., 1989; Besedovsky and del Rey, 1991; Blalock, 1988; Carr and Blalock, 1991; Hall et al., 1991; Madden and Felten, 1995; O’Grady and Hall, 1991; Sternberg, 1988). These responses may improve the defense reaction of the body against pathogens. Thus, increased glucocorticosteroid levels may help to suppress an overly strong immune response, which could in itself be dangerous (as can be seen in allergic reactions) (e.g., Munck et al., 1984). Behavioral changes, such as lethargy, anorexia, or reduced grooming, which are typically observed in sick animals, can also be elicited by cytokines produced during the immune response of an organism against an infection (e.g., Crnic, 1991; Kelley et al., 1994; Myers and Murtaugh, 1995). Assessment of immunological functions. There are two fundamentally different-but not mutually exclusive-approaches to gaining information on the activity and capacity of the immune system in stressful situations: RELATIONSHIP BETWEEN STRESS A N D DISEASES

Diseases in natural populations. The first indications of the influence of psychosocial factors on the functions of the immune system were provided by investigations that demonstrated a relationship between certain social situations and the outbreak of diseases (e.g., at high population densities in rodents or after the death of a partner in human beings). This epidemiological approach is of particular relevance for the human situation, as it provides the only means of assessing the relevance of social influences on disease. However, this approach usually requires large amounts of data and provides no information on the specific underlying immunological

40

DIETRICH VON HOLST

processes associated with the increased morbidity (e.g., Gentry, 1984; Weiner, 1977; Wenar, 1983). Induction of diseases. In this case, pathogens (such as parasites, bacteria, or tumor cells) are injected into animals housed under different stress conditions and the outbreak and progress of the disease is correlated with the intensity of the stress (in some cases also the rejection of skin transplants is used for the characterization of immune capacity). In general today, those immunological parameters are also assessed that are capable of providing information on the action of the stress-induced changes in resistance to disease. This approach is of undeniable importance to the explanation of the relationship between stress and resistance to disease, but is methodologically complicated ( e g , knowlege of suitable pathogens, complicated housing conditions) and hence has so far been used only in a small number of investigations on standard laboratory animals (especially rats and mice) (for details see Section III,B,Z). Direct assessment of immunological parameters. No general statements can as yet be made on the effect of social and psychological influences on the immune system, as it is composed of many different subunits that can exhibit synchronous or antagonistic changes, depending on the situation or subsystem. We also have very little current understanding of the relevance of changes in specific immunological subsystems to disease in an individual. Therefore, the monitoring of many different parameters is necessary, in order to obtain a picture as informative as possible of the reaction of the immune system. This is, however, not the case in most experimental research carried out on animals. Immune measures are usually selected on the basis of financial or methodological constraints (e.g., availability of antibodies and laboratory facilities). Studies based on a limited selection of immune parameters may, however, result in a failure to detect immunological changes. Premature conclusions that a situation has no immunological effects must therefore be avoided. In the rest of this section, the most common immunological parameters used in animal behavior studies are mentioned. A reduction of one or several of these parameters is usually interpreted as a measurement of a reduced immune function. HUMORAL FACTORS AND LEUCOCYTE NUMBERS IN THE BLOOD. Very little blood is required for the determination of humoral factors (e.g., serum concentrations of immunoglobulins or C-reactive protein) as well as of the numbers and types of leucocytes in the blood (e.g., total leucocyte number, leucocyte subcategories such as neutrophils, eosinophils, basophils monocytes, lymphocytes), and subsets of lymphocytes (e.g., T and B lymphocytes, cytotoxic T lymphocytes, T helper cells, natural killer cells). All these parameters can therefore be determined at regular intervals from blood samples of

STRESS A N D ITS RELEVANCE FOR ANIMAL BEHAVIOR

41

individuals. For the differentiation of the lymphocyte subsets and the determination of the different immunoglobulins specific antibodies are necessary, which are currently available only for human beings and a few experimental animal species. The activity of the complement system in serum samples can easily be measured by in v i m bioassays (e.g., CH5[)hemolysis test), by determining the volume of serum necessary to induce hemolysis in 50% of red blood cells (usually of sheep) used as antigens. The measurement can be carried out repeatedly even in small mammals, as this test requires very little serum. IN VITRO TESTS. The in vitro proliferative response of lymphocytes after stimulation with mitogens is widely used as a functional test of the activity of the cellular part of the immune system. Different mitogens can be used, such as concanavalin A (Con A), that stimulates in most species predominantly T lymphocytes, or pokeweed, that is more specific to B lymphocytes. These in vitro tests can be performed on cell cultures of lymphocytes from the blood or from the spleen, as well as on unmanipulated blood samples. In addition to the proliferation rate, the production of the different cytokines (interferons, interleukins) by the lymphocytes after stimulation with mitogens can be determined in the supernatant of the cell suspension. Phagocytic activity of blood cell suspensions (mainly of monocytes [macrophages] and neutrophil leucocytes) can be determined after contact with an antigen (e.g., cymosan A) and the hereby induced release of reactive oxygen molecules. All tests mentioned so far can be performed on the cells of all animal species, so long as the laboratory facilities are available. Most studies so far have been done on animals living under constant laboratory conditions: the application of some of these methods under field conditions is, however, also possible. Natural killer (NK) cell activity is usually determined in v i m by bioassays using the destruction of certain tumor cell lines by the NK cells; the currently available tumor cell lines are, however, specific to only a few laboratory animal species. Although originally most functional tests in animals (especially mice and rats) used lymphocyte cultures received from the spleen of the animals, studies on human beings are naturally based on cells from the blood. The relationship between the activities of these cells of different origin is not clarified. It is, however, accepted that changes in the blood reflect changes in other immunological organs (such as spleen or lymph nodes). Of course, tests on blood cells have the great advantage that changes in immunological functions under different conditions of stress can be monitored on an individual level. These tests are therefore being used increasingly in animal research. For most of these tests, however, large numbers of blood cells

42

DIETRICH VON HOLST

are needed, which means that sacrifice is necessary for animals below the size of a rat. I N VIVO ANTIBODY PRODUCTION AFTER A N ANTIGEN CHALLENGE. In these tests the capacity of an intact organism to produce specific antibodies against antigens (e.g., sheep erythrocytes or keyhole limpet hemocyanin [KLH]) is determined. The specific antibody concentrations can be measured indirectly (by bioassays such as in sheep erythrocytes) or directly by immunological methods, when specific antibodies against the antibodies produced by the individual are available. For antibody determination very little blood is needed, and therefore these functional tests can be performed even in small animals. Since antibody concentrations do not change quickly due to handling processes, these tests can be used in laboratory and field conditions. To determine the time course of antibody production, several blood samples must usually be taken over a period of several weeks, which may raise some problems in field studies. IN MAMMALS 111. SOCIAL STRESS

A. INTRODUCTION While the research of Selye and his contemporaries was mostly concerned with the effects of physical stressors, after 1950 many ecologically oriented studies were published, indicating that social factors participate in pituitaryadrenocortical regulation. This work originated from the hypothesis, advanced by John J. Christian in 1950, that regulation of population densities of small mammals might be achieved by mechanisms intrinsic to the population itself: Increasing population density, according to his hypothesis, results in qualitative and quantitative changes in the behavior of the animals, which in turn stimulate pituitary-adrenocortical activity and decrease pituitarygonadal activity. As a consequence of these endocrine stress responses, the mortality of the animals increases and their natality decreases, thus counteracting the increase of population density. In his first paper (1950), Christian suggested adrenocortical exhaustion and, as a direct consequence, mortality as the major cause of cyclical fluctuations in the population numbers of small mammals. In subsequent years, however, it became evident that increased susceptibility to infectious and parasitic diseases, brought about by increased adrenocortical activity and decreased reproduction, is much more important (e.g., Christian, 1963, 1971, 1975; Christian et al., 1965). 1. Self-Regulation of Mammalian Populations

The growth of mammalian populations usually stops at a more or less stable level below the environmental capacity. Exceptions are the popula-

STRESS A N D ITS RELEVANCE FOR ANIMAL BEHAVIOR

43

tions of many small rodents of the arctic regions, which undergo marked fluctuations with a periodicity of 3-4 years. The causes of these cycles have interested ecologists since the beginning of this century (Elton, 1942). Some authors proposed extrinsic factors such as weather, food, predation, and diseases as the predominant or sole factors influencing animal numbers in natural populations. Without doubt all these extrinsic factors act on populations, either alone or in combination, and can cause a population decline or even an extinction in particular instances, but it is unclear whether they can limit population increases under natural conditions (Krebs and Myers, 1974). Therefore, alternative hypotheses were proposed, claiming population regulation by intrinsic factors. According to a hypothesis proposed by Chitty (1958,1960), the behavior of animals changes with density as a consequence of selection on genetically different behavioral types. Aggressive individuals might be at a selective advantage as the population increases. Aggressive behavior among individuals could then be the direct cause of the population decline, even though the exact reasons for the mortality during the decline are so far unknown. Several studies have demonstrated genetic changes associated with density changes (using mostly gel electrophoretic variants as markers), but the question still remains as to whether these genetic changes are the cause or the result of the demographic changes (Krebs, 1996; Krebs and Myers, 1974). The hypothesis that they are the cause has never gained wide acceptance. In contrast, from the outset, Christian’s hypothesis of a self-regulation of mammalian populations by social stress was a subject of wide debate (e.g., Krebs, 1996; Krebs and Myers, 1974), resulting in intensive stress research in the laboratory as well as in the field (mostly on voles, mice, rats, and rabbits). The strength of this concept is based on the results of experimental crowding of small mammal populations in the laboratory, which demonstrated all the changes in reproduction, growth, and mortality typically found in natural populations of high densities. The relationships between behavior, stress, and density under natural conditions, however, remain far from clear. This is for many different and often methodological reasons. In my opinion, the most important of these is the lack of detailed behavioral studies. This is not surprising in the case of the generally cryptic small mammals. Most studies have collected data only on density-dependent changes in agonistic behavior, and even then these data are often inferred indirectly by counting skin wounds on animals living in different population densities. Such an approach to the determination of relevant behavioral changes may be misleading, however, as some studies have demonstrated that no relationship exists between the number of wounds and population densities in rodents (e.g., Batzli and Pitelka, 1971; Christian, 1971; Krebs,

44

DIETRICH VON HOLST

1964). As a consequence, Christian (1963) introduced the term “social density” to explain discrepancies between animal numbers per unit space and stress responses. This term, however, is far from clear and may include any combination of aggressive behavior, social interactions, space, and number of individuals. In addition, population densities in the field have usually been estimated on the basis of trappings, which are subject to a great number of potential shortcomings, such as differences in their success rate, depending on the individuals’ trap experience, their social status, or their food situation (e.g., Krebs and Myers, 1974). Statements on population densities and changes therein are therefore affected by gross and possibly density-dependent errors. Furthermore, the indices used in the measurement of stress in individuals have usually been provided by data on adrenal weights or by other indirect measurements of adrenocortical activity, both of which give only rough indications of the endocrine state of the animals. Finally, data on the sympathetico-adrenomedullary activities are generally lacking. Thus, many contradictory conclusions on relationships between population densities and stress may have been based on inadequate measurements or interpretations of the data. For these reasons I will not go any further into the question of selfregulation of densities of individuals in mammals. In addition, several reviews on the causes of population cycles in small mammals have been written over the past 30 years and the understanding of these causes has not subsequently improved (e.g., Christian, 1978; Krebs, 1996; Myers et al., 1971; Nowell, 1980; Snyder, 1968; Watson and Moss, 1970).

2. Behavioral Stress Research in Biomedicine While initially ecological questions were central to research on stress, and factors relevant to populations such as fertility and mortality were investigated, since the 1960s, interest has shifted into the biomedical area. Three developments were responsible for this shift. 1. The development of modern techniques in hormone analysis, allowing the recording of endocrine stress reactions in humans and revealing the general correspondence between many different animal species including humans.

2. The proof that in animals, as well as in humans, psychological processes are decisive in the triggering of stress reactions. Psychosocial stressors are capable of long-term modification of both the sympatheticoadrenomedullary and the adrenocortical systems. 3. The epidemiological evidence that social challenges (life events) and individual traits in personality (e.g., type A personality, characterized

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVlOR

45

after Rosenman, 1986, by enhanced competitiveness, aggressiveness, impatience, and a chronic sense of time urgency) in humans are involved in the development or outbreak of specific diseases. Experimental investigations on animals are now primarily directed at recording physiological processes potentially involved in the development of disease in humans. The relevance of heart and circulatory diseases to our society has resulted in most research focusing on this area. In recent years, psychoneuroimmunological research has increasingly analyzed the importance of social stress reactions in the development of infectious diseases and tumors. Up to the present, research has concentrated almost entirely on laboratory animals kept in highly standardized conditions. It is only in recent years that a few ecologically oriented field research projects have reappeared, such as on dasyurid marsupials in Australia, rabbits in Europe and elsewhere, and mongooses, cheetahs, and several monkey species in Africa (see Section 111,BJ). Although the relevance of social stress for the “regulation” of populations is still controversial, nowadays it is generally accepted that social interactions or situations may result in strong stress responses in mammals (including humans), which in certain circumstances can greatly reduce the vitality and fertility of individuals and even lead to their death. Based on results of our studies on tree shrews and wild rabbits, I shall present in the following section my view of the relevance of recent stress research for the understanding of animal behavior. B. SOCIAL CHALLENGES THAT MAYRESULT IN STRESS RESPONSES 1. Social Conpict

a. Adrenocortical Activity and Fights. Fights over limited resources such as food, shelter, territories, and/or rank are among the most conspicuous behaviors in animal societies. All mammals react to acute stress, such as is certainly induced by a fight, by immediate activation of their sympatheticoadrenomedullary and hypothalamo-pituitary-adrenocortical systems. If such fights occur only infrequently, they have no consequences for the fertility and health of an individual: For instance, tree shrews can be subjected to a 30-min confrontation daily over a period of several weeks without suffering any detrimental physiological effects. If these daily fights occur more frequently, however, they can have grave detrimental effects on the fertility and health of the animal and even-as described at the beginning of this chapter in Antechinus-rapidly result in death. One or two aggressive conflicts per hour, not unusual in a natural environment, may be sufficient to have a damaging effect, as the adrenocortical hormones require one to

46

DIETRICH VON HOLST

several hours, depending on the species and the interaction, before they (and hence many other parameters) revert to their original concentrations. As is the case in Antechinus, and also in many other species with seasonally restricted reproduction, this phase of increased aggression is characterized by correspondingly increased adrenocortical activity in individuals, as was shown in the early 1960s by Bronson (1963, 1964) in his studies on natural populations of woodchucks (Marmora monax). The same is also true for other species (Saad and Bayle, 1985; Saboureau eta/., 1977; Schiml el al., 1996), as shown later in more detail, based on our investigations on European wild rabbits. European wild rabbits (0ryctoIagu.y cnniculus) live in small territorial groups of 1-3 males and approximately double the number of females. The territories are intensively defended by the males against external rivals during the reproductive period. Within the groups both sexes exhibit separate intrasexual linear ranking systems. Aggression in both sexes is particularly high at the beginning of the reproductive period, whereas for the rest of the year wild rabbits live together largely in peace (e.g., Brambell, 1944; Cowan, 1987; Lockley, 1961; Marsden and Holler, 1964; Myers and Poole, 1959; Mykytowycz, 1958; Southern, 1940). In order to gain information on the influence of agonistic behavior on physiological parameters, adult wild rabbits of a large natural population on the North Sea island of Sylt were investigated at the beginning of the reproductive period, when aggression was maximal (end of March), and again after the end of the reproductive period (October/November). At these two times of the year a total of 500 animals were shot between 18:OO and 19:OO hours over a period of 3 years, and within less than 3 min of their death blood samples were taken and different organs extracted for endocrinological and other investigations. As food availability, temperature, and day length were more or less equivalent during the two hunting seasons in spring and late autumn, differences in physiological stress parameters should be due largely to differences in aggression. As expected, we found greatly increased adrenocortical and sympathetico-adrenomedullary activities as well as many other changes in both sexes in spring, indicating high levels of stress. That is, under natural conditions, wild rabbits of both sexes show endocrine stress responses of the same magnitude as those demonstrated mainly in rodents under laboratory conditions (Fig. 13). We have closely investigated these relationships between social behavior and physiological stress responses over a period of 10 years in a population of wild rabbits living in a seminatural environment in an enclosure covering an area of approximately 22,000 m2 (e.g., Eisermann et a/., 1993; Kiinkele and von Holst, 1996). In these conditions, both sexes also start fighting

STRESS AND ITS RELEVANCE €OR ANIMAL BEHAVIOR

Adrenal gland

Corticosterone

400

I

(ng/ml serum)

Cholesterol 32

300

24

200

16

100

8

0

1

(mall00 ml . serum)

0 TH activity

7.5

47

(nmollh adrenals)

Adrenal medulla 500

(Cell nucleus volume in LP)

PNMT activity

36

6.0

400

27

4.5

300

18

3.0

200

9

1no

1.5

Males Females

1

(nmollh adrenals)

n

I

Males Females

Spring: 89 males, 76 females

Males Females

Autumn. 25 males, 20 females

FIG. 13. Indices of adrenocortical and sympathetico-adrenomedullary activities of adult male and female wild European rabbits in spring and autumn. The volume of the cell nuclei of adrenal medullary cells was histologically determined from 10 males and females in each season; their enlargement indicates a markedly increased adrenomedullary activity in spring. All data means ( + SEM). Significant differences between the spring and autumn data: * p < .05, * * p < .01.

heavily in spring and many injuries result (Fig. 14). In the males these fights are over territories and ranking positions and decrease only slightly in the course of the reproductive period. Females exhibit a bimodal pattern of aggression with maxima at the beginning and the end of the reproductive period. At the beginning of the reproductive period, fights usually occur over ranking positions within their groups as well as with external females, while in autumn most aggression is directed against young animals attempting to enter the groups (Fig. 15). In concert with these behavioral changes, there are marked changes in adrenocortical activity in both sexes. Conflict avoidance by the restriction of aggression mainly to the reproductive period is therefore a useful means of avoiding such stress reactions.

48

DIETRICH VON HOLST

4.0

1

[7 sexual following

Male sexual behavior (interactionshour)

urination

3.0 2.0 1.o

0.0 2.4

Fresh wounds (number)

1.8

1.2

0.6 0.0

250

Corticosterone

200 150 100

50 Dec

Feb

Apr

Jun

Aug

Oct

Dec

FIG. 14. Sexual behavior (medians), wounds, and serum corticosterone ( M _f SEM) of wild European rabbits, kept in a 22,000-m2 enclosure under natural conditions. Data from about 25 males and 50 females, observed over a 4-year period. Blood samples forcorticosterone determination were taken monthly from the animals 1 hr after maximal stimulation of their corticosterone release by an injection of ACTH. For behavior analysis each individual was observed about 8 hr per month. Unpublished data; after Schonheiter (1992).

b. Physical versus Psychosocial Processes. At the onset of mammalian stress research some authors suggested that intensive physical effort during fighting and wounds resulting from these fights might be the principal factor in adrenal enlargement; however, most studies found no correlation between injuries from fights and adrenal weight, and suggested psychological (psychosocial) factors as underlying causes (e.g., Christian, 1963). One of the first studies supporting this hypothesis was carried out by Davis and Christian (1957), who demonstrated a significant correlation between

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

49

Aggressive behavior (interactions per hour) 9.0

Males

6.0

3.0

0.0

Females

0adult group members

Jan

Mar

May

Jul

Sep

Nov

Jan

FIG. 15. Aggressive behavior of adult male and female wild European rabbits toward juveniles and adults (medians; for animal numbers see Fig. 14).

adrenal weight and social rank in groups of male house mice. Another study by Barnett (1958) found increased adrenal weights only in subordinate wild Norway rats (Rattus norvegicus) in a resident intruder paradigm. Furthermore, Bronson and ElefthCriou (1965b) showed that even the mere exposure without physical contact of subordinate mice to fighters-if they had previously been defeated in a confrontation-produces adrenocortical responses of the same magnitude as those observed in mice actually attacked and defeated. Similar findings have been obtained from research on Syrian hamsters (Mesocricetus auratus) (Human et al., 1992). Furthermore, research by Fokkema and Koolhaas, using chronically catheterized laboratory rats, showed that defeated males exhibited over twice the increase in blood pressure during brief dyadic encounters than their superior rivals (Fokkema, 1985; Fokkema and Koolhaas, 1985). If an animal had previously been defeated and was then threatened by exposure to the victor, while penned in a small wire mesh cage, the mere presence of the victor raised the former victim’s blood pressure to the same level as during the direct defeat. Extreme psychosocial stress can even rapidly cause death, as has been demonstrated in wild Norway rats (Barnett, 1958, 1964, 1988; Barnett et al., 1975), tree shrews (von Holst, 1972a,b, 1985a), rhesus monkeys (Hamil-

50

DIETRICH VON HOLST

ton and Chaddock, 1977), and humans (Stumpfe, 1973). Death in these cases is always associated with behavioral impairment, indicating a state of helplessness or loss of control, and extremely heightened adrenocortical activity (as shown in rats and tree shrews). An example of these mechanisms is given in the following section. c. Social Stress in Tree Shrews. Tree shrews (Tupaia belangeri, order Scandentia) are small diurnal mammals distributed throughout Southeast Asia. In the wild, tree shrews live in pairs in territories that they defend vigorously against intruders of their own sex. In the laboratory, adult males (and females) also immediately attack intruders of their own sex and normally defeat them within a few minutes. A short time after the fight, the winners show no further signs of arousal and pay virtually no attention to the defeated animals. The losers, in contrast, creep into any hiding place, which they leave only to eat and drink. During the following days, fights between the animals are extremely rare or nonexistent. Nevertheless, the losers die within a few days. Death is not a result of physical exertion during the fights, nor are wounds the cause of death, as the animals usually inflict only superficial scratches and bites on one another. Death is rather more the consequence of the continual presence of the winner, as was shown by the following experiments: An adult male was placed in the cage of a male conspecific (an experienced fighter), which usually immediately attacked the intruder and subdued him in less than 2 min. Afterward, both animals were separated either by a nontransparent partition or by a wire mesh partition, so that the loser could no longer be attacked but could continually see the threatening winner. Short fights were repeated every 1-3 days. Losers separated by a nontransparent partition from the winner recovered from the fights just as fast as the winners and did not die prematurely, even when they were subjected to short daily fights over weeks. The situation of the losers within sight of the winners, however, was completely different: As from the first subjugation, all the submissive animals sat in a corner of their part of the cage or in the sleeping box attached to the cage for practically the whole day, hardly responding to external stimuli. During confrontations, they did not even attempt to escape the attacks of the dominant animals, but usually suffered them without any attempt t o defend themselves or to flee. After the first subjugation their body weight decreased daily at an individually stable rate of 2-8% of their initial weight and all animals died within 2-20 days, if not separated earlier. As these results show, death of the submissive animals is not a direct result of the fights and their physiolgoical consequences, but is rather a result of central nervous (emotional) processes in the defeated animal, induced by the constant presence of the threatening winner.

STRESS A N D ITS RELEVANCE FOR ANIMAL BEHAVIOR

51

In all submissive animals dramatic stress responses were seen: Among other responses, serum glucocorticosteroid levels rose to more than six times their initial levels, serum concentrations of both thyroidal hormones (thyroxine and triiodothyronine) decreased to less than 40%, and those of testosterone to less than 10% within a few days. As shown by histological examinations, within a few days spermatogenesis ceased completely and all animals became sterile. In addition, a dramatic drop in the numbers of lymphocytes as well as of basophil and eosinophil granulocytes to less than 20% of their initial values indicates strong immunosuppression (von Holst, 1985a). The physiological cause of death differed among the animals. In animals dying within 8 days of the confrontation, urea nitrogen (and creatinine) levels in the serum rose to more than ten times their initial values, leading to death by uremia, while in those animals surviving longer, these increased only slightly. The cause of death in the latter is not known (von Holst, 1972a,b, 1985a). (Of course, the introduction of an individual into the cage, or “territory,” of an experienced fighter without any possibility of avoiding the rival is an unnatural and extremely stressful situation, especially in a territorial animal such as the tree shrew.) To obtain information on a less severe form of stress or even on adaptation to it, two male tree shrews unknown to each other were put together in a cage with two separate sleeping boxes, water bottles, and feeding dishes. In this situation the animals did not start to fight immediately, but first hesitantly explored and marked the cage. Slight fights usually began within the first few hours, leading in most cases to clear dominance relationships within 1-3 days. While the behavior of all males was more or less comparable before the fights, it changed considerably after the dominance relationship was established, depending on the social position of the animals. Although both animals continued to live together in the relatively small cage, the winners more or less ignored the losers and attacks on the latter were rare or even completely absent. The losers, on the other hand, drastically moderated their behavior. On the basis of behavioral differences, two types of losers could be distinguished: subdominant and submissive animals. Submissive animals corresponded to those of the first experiment: They crouched in a corner of the cage or in a sleeping box and left their hiding place only to drink and eat hastily. They even tolerated the infrequent attacks of the winners without any attempts to defend themselves or to flee. They ceased grooming completely and their fur became rough and dirty. To the human observer they gave an apathetic or depressive impression. All submissive animals died within 2 weeks.

52

DIETRICH VON HOLST

Sudominant animals, in contrast, showed greatly increased locomotor activity, watching the movements of the dominants continually and attempting to avoid possible confrontations by giving way or fleeing. If a confrontation could not be avoided, they even defended themselves. Under these conditions, subdominants were capable of living in the presence of the dominants for weeks, albeit with a reduced freedom to move. In concert with these behavioral changes, body weights and many different physiological parameters in the animals changed drastically. In accordance with the stress concept, the confrontation had the immediate effect of activating the sympathetico-adrenomedullary and the pituitaryadrenocortical systems in both rivals: Accordingly, the serum concentrations of catecholamines and glucocorticosteroids as well as the heart rates of all animals were greatly increased (Fig. 16). As soon as the dominance relationship was clearly recognizable in the behavior of the tree shrews, all stress reactions in the dominant animals disappeared in spite of continued occasional fights. Moreover, glucocorticosteroid concentrations in the blood dropped marginally below the original values, their body weights increased, and their gonadal functions improved: After about 3 weeks, the dominant animals were significantly heavier than before the confrontation and serum testosterone concentrations had increased by approximately 100% (Figs. 17 and 18).

Cortisol

1

6.0

Epinephrine

Norepinephrine

(ng/ml serum)

I

20

5.0

8.0

15

4.0

6.0

10

3.0

4.0

5

7

2.0 D

SD

SM

(nglml serum)

10.0

D

SD

SM

D

SD

SM

FIG. 16. Serum concentrations ( M -+ SEM) of cortisol and catecholamines o f , minant (D), subdominant (SD), and submissive (SM) tree shrews 8 days before (light bars) and two days after (cross-hatched bars) start of a confrontation. Blood samples were always taken 2 hr before the activity period. Cortisol data (20-30 animals per group) are initial values, catecholamine data (10 animals per group) are BSCT values. Significant differences to initial values: * p < .05; * * p < .01; ***p < ,001.

53

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

Dominants

Controls 40

cortisol (nglml serum)

20 301

B 2 10 20

Subdominants

Submissives

! B 2 10 20

B 2 10 20

6 2 10 20

FIG. 17. Serum concentrations (M t SEM) of cortisol and testosterone of male tree shrews before (B) and at different days after start of the experiment. 20 controls remained in their living rooms, but were treated as the other animals (handling for weighting and blood sampling) and 25-40 males per group that confronted each other. See text for further details.

Subdominants Submissives

0

6

12

18

24

Days after start of experiment FIG. 18. Body weight changes of the animals after the start of the experiment ( M for animal numbers see Fig. 17. Initial body weights of the males: 190-220 g.

?

SEM):

54

DIETRICH VON HOLST

Defeated tree shrews were characterized by reduced body weights and a number of hormonal and other physiological changes; among other effects, the serum concentrations of testosterone, insulin, and thyroid hormones decreased drastically. Overall, these effects were the same in subdominant and submissive animals, differing only in their extent (Fig. 19). Qualitative differences were, however, to be found in their sympatheticoadrenomedullary and pituitary-adrenocortical systems. Active and passive stress reponses in tree shrews. In dominant animals, the activity of the sympathetico-adrenomedullary system reverted back to

Testosterone 9.0

j

30

8.0

4

20

7.0

10

6.0

0

5.0

Protein (g/100mlserum)

*

Triiodthyronine 1.0

1

(ng/ml serum)

Kidneys

Insulin

36

1

(ng/ml serum)

150

0.8

27

0.6

18

100

0.4

9

75

0

50

0.2

125

1 i

Hemoglobin

20

1 (g/lDo

!joo

400

16

300

14

200

c

D SD SM

I

I

Epididymides

ml blood)

18

3I 7L

Triglycerides (mg/l00 ml serum)

1

;mg)

ion C

D SD SM

C

D SD SM

FIG. 19. Several physiological measures and organ weights ( M ? SEM) of controls and experimental male tree shrews 10 days after start of the experiment. Animal numbers as in Fig. 17. Significant differences to controls: * p < .05: **p < .01; ***p < ,001. C: controls: D: dominants: SD: subdominants: SM: submissives.

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

55

its original state after the dominance relationship had been established, and the serum catecholamine levels even showed a tendency to levels lower than those before the confrontation. In contrast, subdominant animals exhibited a continually increased sympathetico-adrenomedullary activity. While, correspondingly, the heart rate in the dominant animals returned to normal once dominance was established, it remained high in subdominant animals, not only throughout the day (when an attack by the dominant animal was always possible), but also at night when they were sleeping in their own sleeping box. Nocturnal heart rates in subdominant animals were almost equivalent to diurnal values, thus abolishing the original day-night rhythm (Fig. 20). The tyrosine hydroxylase activity in their adrenal glands-an index of sympathetic activity-also increased on average by loo%, in comparison to dominant or control animals (Fig. 21). Contrary to the Selyean concept, the serum levels of the glucocorticosteroids decreased to initial levels; their adrenocortical secretory capacity, however,

500

-

400

E6=

200

d

n

Male dominant

300

100

u)

iii

500

a E u) c

400

al

m

E 3

300 200

L

g 2

100

0)

5

500

1

Male submissive

72

400 I

300 200 100

Days before and

I

after start of confrontation

FIG. 20. Heart rates of a dominant, a subdominant, and a submissive male tree shrew before and after the start of the confrontation.

56

DIETRICH VON HOLST Cortisol

40

(nglmlserum) *** I

60

Corticosterone

Adrenal weight

**

***

nglrnl serum)

I

I

30

45

36

20

30

32

10

15

28

0

0 Urea nitrogen

100

(mg/l00 ml serum)

***

I

80

24 irosine hydroxylase

10.0

:nmol/h adrenal)

PNMT

* I

adrenal)

7.5

2.4

5.0

1.6

2.5

0.8

*

60 40 20 0

0.0

0.0 Eosinophils

Lymphocytes

Erythroblasts

2800

600

2100

450

540

1400

300

360

700

150

180

0

(n/ pl blood)

0 C

D SD SM

0 C

D SD SM

C

D SD SM

Data 10 days after start of experiment

FIG.21. Several measures of the adrenocortical and adrenomedullary systems as well as some blood cell numbers ( M i SEM) of controls and confronted male tree shrews 10 days after start of the experiment. Animal numbers and abbreviations as in Figs. 17 and 19. Significant differences to controls: * p < .05: **p < .01: ***p < ,001. See text for further details.

increased to the same level as in submissives, indicating a heightened cortisol release and elimination under active stress (Figs. 17, 21, and 22). Submissive animals exhibited the opposite reactions: a tendency to a decreased sympathetico-adrenomedullary activity (Fig. 21), as indicated by their lowered adrenal tyrosine hydroxylase activities; and a substantial increase in their serum levels of glucocorticosteroids, as well as of their adrenal capacities (Figs. 17, 21, and 22), which was probably responsible

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

***

57

*** .L

D

SD

SM

D

SD

SM

FIG.22. Serum cortisol levels ( M t SEM) 15 min after start of blood sampling challenge tests of dominants (n = 11). subdominants (n = 11) and submissives (n = 6 ) before (left) and 10 days after (right) start of the confrontation. Significant differences: * p < .05: ***p < .001. Note the significantly lower initial values of the males that became submissive during the confrontation.

for the increased loss of muscle and adipose tissue, leading to a dramatic loss of weight averaging 5% daily (von Holst, 1986a, 1994; Stohr, 1986). Interestingly, the adrenal capacities of prospective submissive animals were significantly lower before the confrontation than those of the other two groups (Fig. 22). Confrontation also triggered marked immunological changes: No significant change was observed in leucocyte numbers and subsets in the blood of dominant and subdominant animals. In contrast to dominant animals, however, the efficiency of lymphocytes and phagocytes in the subdominant animals was clearly reduced (Fig. 23). In submissive animals, substantial changes were also found in those types of leucocytes indicative of strong immune suppression (e.g., Fig. 21: lymphocytes and eosinophil granulocytes). Accordingly, the proliferation capacity of their lymphocytes was reduced to less than 20% of original values following a 10-day confrontation period. Subordination and presence of the dominant animal therefore affected the immune system in subdominant and submissive animals. These effects were qualitatively equivalent in both categories of animals as far as their inhibiting effect on proliferation of lymphocytes was concerned, but different regarding the distribution of the different types of leucocytes in the blood: General statements on the possible qualitative differences in immunological reactions in these ethologically and physiologically distinct subordinates are not yet possible based on available data.

58

DIETRICH VON HOLST

Before

After

Before

After

FIG.23. In vitro lymphocyte proliferation (LP) after stimulation with the mitogen concanavalin A (Con A) and in vitro phagocytosis of monocytes and granulocytes ( M ? SEM) of 16 subdominant male tree shrews before and 10 days after start of the confrontation (for details see Section 11,C: direct assessment of immunological parameters). While their lymphocyte numbers during the confrontation were not different from initial values, their proliferation capacity decreased markedly ( p < ,001).

d. Assessment of Dominance between Rivals. As the results of research on tree shrews show, prospective winners and losers exhibited differences in body weights and cortisol values after only 2 days (Figs. 17 and 18), even though it was not usually possible to predict the outcome of the confrontation based on the animals’ behavior at this point. Similiar results were found in domestic guinea pigs (Sachser and Lick, 1989): At the age of 30 days, the authors removed juvenile males from their breeding colonies and housed each of them with a female in a 1-m2 enclosure. At the age of about 8 months, two males were confronted by removing the partition between their enclosures. In all cases these confrontations escalated into fights between the two opponents, which declined to low levels after the first day. However, dominance relationships between the rivals could not be distinguished before day 4 of confrontation. Nevertheless, at day 2-3, prospective winners and losers already differed significantly in nearly all physiological parameters measured: Prospective losers exhibited a higher loss of body weight, their serum concentrations of cortisol and catecholamines were two to three times higher than those of prospective winners, while their testosterone levels decreased to about 30% of their original value. Commencing with the fourth day of controntation, the physiological changes became so dramatic that the losers died within a few days. As these data show, the outcome of the agonistic encounters could be

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

59

confidently predicted from physiological data 1-2 days before it could be derived from the behavior of the individuals. These results of research on tree shrews and guinea pigs demonstrate the animals’ capability of predicting the outcome of a confrontation while it is taking place and of producing the corresponding physiological reactions. How quickly endocrine changes occur during the changing assessment of the situation by the animals is demonstrated by Schuurman’s results (1981) with laboratory rats, which had been provided with chronic jugular vein cannulae for frequent blood sampling. Schuurman introduced a male rat into the home cage of an aggressive male conspecific. At first both animals were separated by a wooden partition, which was removed after a habituation period of 2.5 hr for a l-hr confrontation. During the first quarter of the confrontation fierce fighting took place, but with no apparent dominance relationship between the two rivals: Both showed offensive as well as defensive behavior, and their plasma corticosterone levels rose to about five times their initial values. After about 15 min a dominance relationship developed, which resulted in distinctly different adrenocortical responses in both animals: Although the victors continued offensive fighting, their plasma corticosterone levels declined. On the other hand, plasma corticosterone levels of losers continued to rise and the animals exhibited much defensive and submissive behavior, but no longer any offensive behavior. At the end of the encounter, plasma corticosterone level in losers was more than twice the level found in victors. Once a dominance relationship has been established in mammals, fights between rivals usually decrease. Nevertheless, as shown for tree shrews (see earlier discussion), the permanent presence of a dominant rival can cause dramatic stress responses in subordinate animals and ultimately even lead to their death within a few days. In spite of these highly negative consequences of subordination, fights for dominance are often astonishingly slight or even completely absent in tree shrews. This is apparently due to the fact that male tree shrews are able to recognize the potential “strength” or “dominance” of a rival before the first physical interaction, as is shown by the following experiments. Two animals were transferred to an experimental room and housed in one cage, but separated for 20 days by a wooden partition. Before the transfer into the experimental room, and on days 10 and 20, blood samples were taken from all animals for the determination of several endocrine and immunological parameters. On days 21-23, the two animals confronted each other daily for 10 min, while the rest of the time they were separated by the wooden partition. Depending on the results of these confrontations, the animals were designated prospective “dominants” and “subordinates.”

60

DIETRICH VON HOLST

From the first day in the experimental cage both animals apparently recognized the presence of the rival: They sniffed intensively at the wooden partition and marked it. Animals that later turned out to be subordinate in the confrontation seemed more alert and exhibited more locomotor activity, as was also evident from their slightly decreased daily resting times compared to their prospective dominant rivals. While few overt behavioral differences were observed between the two groups, there were significant immunomodulatory effects (Fig. 24). These effects were in opposite directions in the animals of the two groups: Tree shrews that later became subordinate showed indications of an immunosuppression, while in prospective dominants the activity of the immune system improved. In contrast to these immunomodulatory effects, we found only very slight changes in adrenocortical or sympathetico-adrenomedullary activities. Amazingly, the immunosuppressive effects in subordinate animals before any physical contact between the rivals were of the same magnitude as those in direct confrontations with constant physical presence of the rival (Fig. 25). To exclude the possibility that these opposite immunomodulatory reactions were consequences of differences in the constitution of the rivals, the

200

Lymphocyte proliferation

Interleukin 1

150

I 3

100

7

50

m ._ c ._ ._ c

o

-

-

0

Interferon gamma

400

c ._

***

111 300 0

5 200 100

0 H

10 20

H

10 20

Dominants Subdominants

H

10 20

H

10 20

Dominants Subdominants

FIG.24. Several immunological measures taken from male tree shrews housed in the same room with a potential rival ( M 2 SEM); data 8 days before (H) as well as 10 and 20 days after transfer into the experimental room: 20 prospective dominant and 20 prospective subdominant males. Significant differences: **p < .01; ***p < ,001. See text for further details.

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

Lymphocyte proliferation 120 -

Interferon gamma

(lo3 cpm/lo5 cells)

+ r * 80 -

6.0

-

3.0

40

0-

61

1

00 10

Home

2 12 Exp

2 12 Conf

10 Home

2

12 Exp

2 12 Cod

Experimental conditions and days of blood sampling FIG.25. In v i m lymphocyte proliferation after stimulation with the mitogen Con A and interferon production of 20 male subdominant tree shrews in their home room (Home: days 0 and lo), 2 and 12 days after transfer into an experimental room with a rival (Exp), and 2 and 12 days after start of the confrontation (Conf); significant differences to initial values (= Home 10): * p < .05: * * p < .01. See text for further details. Unpublished data from Vitek (1996).

tests were repeated 3 months later with all experimental animals in different combinations: Previous dominants were confronted with previous dominants, and previous subordinates with previous subordinates. In all cases, the animals formed new dominance relationships, with identical immunological consequences for the final winner and loser, just as in the first experiment. Thus, the differences in these immunomodulatory reactions of prospective dominants and subordinates were only dependent on the situation, that is, on the “quality” of the prospective rival. As preliminary data show, male tree shrews are capable of obtaining information on the potential strength of a rival through olfactory signals, which originate from urine and glandular areas used for territorial marking. It is entirely unknown, however, what determines this olfactorily communicated “strength” of a rival. As our experiments with tree shrews have shown, body size or body weight are of no predictive value for the outcome of a fight between unknown rivals; the same is true for differences between animals in testosterone serum concentrations or excretion rates.

2. Dominance Relations and Stress Responses in Other Mammals a. Pituitary-Adrenocortical and Sympathetico-Adrenomedullary System. As shown in the previous section, subordinate tree shrews exhibit behaviorally dependent differences in their stress responses, with only the behaviorally inactive and apathetic submissive animals corresponding physiologically to the Selyean concept. These findings strongly support the

62

DIETRICH VON HOLST

concept of Henry and associates, which proposes that mammals exhibit two types of stress with differing neuroendocrine responses and differing ultimate disease states (e.g.? Ely and Henry, 1978; Henry and Meehan, 1981; Henry and Stephens, 1977): Cannon’s flight or fight response, accompanied by heightened sympathetico-adrenomedullary activity, which may eventually lead to cardiovascular deterioration (“active stress”); and Selye’s stress response, characterized in particular by heightened pituitaryadrenocortical activity (“passive stress”). In species such as the tree shrews, which live in territorial pairs in natural conditions, adult individuals of the same sex cannot be kept together for any length of time, since, depending on their passive or active coping behavior, this may result in the death of subordinate animals within a few days to weeks. For this reason, our confrontation experiments with males were always terminated after 3 weeks. Following this time span, subdominant animals also succumbed to apparent signs of coronary insufficiency. The situation is different in animals that normally live in groups with idiosyncratic yet relatively stable social roles. In most species, hierarchical systems with dominant and subordinate animals develop in such a way that subordinates can usually live in the presence of dominants with only very slight or no signs of stress at all. In situations of continuing conflict that result from social instability within the group, life may be stressful for subordinates as well as for dominants, even in the absence of overt aggression, as was shown by Henry and associates in their extensive studies on laboratory mice (e.g., Ely, 1981; Ely and Henry, 1978; Henry and StephensLarson, 1985). After weeks of isolation, the authors introduced several male and female laboratory mice into colony cages consisting of small boxes connected by tubular runways and designed to induce frequent social interactions. This always resulted in intensive fights among the males, which led within a few weeks to a relatively stable social system with one dominant male and several subordinates in each colony. During the time of colony formation the physiological responses of all animals were typical of general nonspecific arousal, with increased activities of their adrenocortical and adrenomedullary systems. Once the social hierarchy was established, overt aggression and strong stress responses decreased. Nevertheless, the behavior and physiological response patterns of dominants and subordinates differed: Dominant animals were significantly more active and vigilant than the subordinate animals; they constantly visited the boxes of the cage system and tried to exert control over all other animals. The subordinate males, on the other hand, were restricted to very small areas of the population cage and showed behavioral withdrawal, which according to the authors minimized aggressive encounters with the dominants. As demonstrated previously by other au-

STRESS A N D ITS RELEVANCE FOR ANIMAL BEHAVIOR

63

thors in crowding experiments on mice and other species (e.g., Bronson, 1973; Chapman et al., 1969; Louch and Higginbotham, 1967; Popova and Naumenko, 1972), even in this more naturalistic experimental design subordinate mice maintained slightly elevated adrenocortical activities. In contrast, dominant mice developed heightened adrenomedullary activities and a moderate hypertension, which apparently permitted the animals to maintain the vigilant patrolling behavior necessary to the control and stabilization of the colony. Testosterone levels in subordinate mice decreased to about 30% of that of controls kept in standard laboratory cages, while the levels in the dominants did not differ from those in controls, in spite of their heightened sympathetico-adrenomedullary stress responses. All these physiological response patterns typical of dominants and subordinates could be reversed by experimentally induced changes of the social positions of individuals (e.g., removal of dominant males). When the animals were separated before 5 months of social conflict, all these stress values returned more or less to those of control animals. If, however, the situation persisted, within a few months pathophysiological changes began to develop in all animals: Fixed hypertension and increased PNMT values, increased heart weights, and histopathological deterioration, such as interstitial nephritis, aortic arteriosclerosis, intramural coronary arteriosclerosis, and myocardial fibrosis developed, which after 9 months of colony living remained irreversible even after months of isolation, and led to the premature death of many individuals (Henry and StephensLarson, 1985; Henry et al., 1971; Vander et al., 1978). As shown by these data, in mice kept in complex population conditions, dominant individuals show predominantly active stress responses, while the subordinate animals show slight passive stress responses. These results are in contrast to our data on dominant tree shrews, which show no sympathetico-adrenomedullary activation. However, this is probably due to the striking differences in the social organization of these two species. In tree shrews, which live in territorial pairs in the wild, a defeated rival is apparently no longer threatening, and therefore, even in small cages, does not elicit any behavioral or physiological arousal. In contrast to dominant mice, dominant tree shrews even show slightly decreased adrenocortical activity. Lundberg and Frankenhaeuser (1980) particularly emphasize this “bidirectional nature of the pituitary-adrenal response.” In their study on humans, adrenocortical suppression was demonstrated in conditions characterized by high levels of control and predictability. According to the authors, this is consistent with the conclusion by Levine and associates (1979), that reinforcement is an important cognitive factor mediating suppression of the adrenocortical system. This also seems to be the case in dominant tree shrews. Wild mice in natural conditions, however, live in

64

DIETRICH VON HOLS’I

groups containing one dominant and many subordinate males. The dominant male must constantly control and keep in check the subordinate males and this apparently necessitates heightened adrenomedullary activity (Crowcroft, 1955; Lloyd, 1973). Thus, it is not the social position that determines the physiological state of an animal, but the effort of achieving and maintaining the status (i.e., whether the position is endangered or not). This has been clearly demonstrated by Lundberg and Frankenhaeuser (1980) in the experiments on humans, which demonstrate that pituitary-adrenocortical activation is associated with negative feelings of distress, and sympathetico-adrenomedullary activation with feelings of alertness and a readiness to act. A similar conclusion was drawn by Ursin and associates (1978), who identified a “cortisol factor” and a “catecholamine factor” in their analysis of data collected in a study of trainee parachutists. It has to be emphasized, however, that in these and many similar studies with mice chronic high levels of stress were induced by using males that were housed singly after weaning for many weeks before colony formation (“unstable social situation”). In contrast to these socially deprived individuals, mice that had been raised in groups were able to live together in stable social groups without overt aggression and stress responses (Fig. 26). As pointed out already (Section II,B,3,c), these results again demonstrate the crucial role of social experience after weaning for an animal to cope with social conflict in a more or less stress-free way. These results lead to the conclusion that, depending on species, group composition, and group stability, dominant individuals may be characterized by lower or higher hypophyseo-adrenocortical and/or sympatheticoadrenomedullary activities than their subordinate conspecifics. Our knowledge of the effects of social positions and behavior on differing endocrine stress responses in mammalian species, however, is scanty, as most researchers have chosen to work with one system only (usually the adrenocortical). Accordingly, countless publications on many different species have demonstrated that repeated subjugations or a subordinate social position result in increased adrenocortical and decreased gonadal activities. However, little is known about the species-specific and context-dependent positive or negative consequences of dominant positions for the sympatheticoadrenomedullary system (including heart rate and blood pressure). An exception to this is provided by several studies on rats and monkeys (mainly on the effects of psychosocial stress on blood pressure) and, as mentioned earlier, to some extent also by studies on wild rabbits, tree shrews, and guinea pigs. An overall analysis of our wild rabbits housed in large enclosures revealed a distinct relationship between the social rank of males and females and

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

Systolic blood pressure I

65

Tyrosine hydroxylase 240

1

(oh of controls)

,

140

120

100

Renin

Corticosterone

1 (nglml plasma)

1(nglml plasma x hr) 10

150

5

100

0

50

FIG.26. Effects of chronic psychosocial stress on several physiological measurements of male mice. Systolic blood pressure, adrenal tyrosine hydroxylase activity, and plasma values of renin and corticosterone of mice housed singly (light bars) in small cages as well as of mice housed in mixed-sex groups for 6 months. Significant differences between animals housed in unstable (cross-hatched bars) social groups and the other housing conditions (stable groupsstriped bars) are indicated: **, < .01; ***, < .001. See text for further details. Adapted from Henry (1992). with kind permission from Transaction Publishers, New Brunswick.

their adrenocortical activities during the reproductive season (Fig. 27). At the group level, however, marked differences existed (Fig. 28): Generally, males living without rivals within their groups had the lowest corticosterone challenge test values. Depending on the number of subordinate rivals in a territory (and hence the social instability within their groups), the adrenocortical activities of dominant males rose to the values of subordinate males; thus no rank-related differences were evident. Furthermore, the number of females living within a group had some influence on the adrenocortical activities, especially of the dominant males. Dominant females, in contrast, had generally the lowest adrenocortical activities of all animals within their groups, although their activities differed to some extent among the groups. Taken together, these data indicate that the adrenocortical activities of dominant wild rabbits depend on the composition and stability of their

66

DIETRICH VON HOLST

Corticosterone values

Heart rates (bpm)

I

250

240

200

200 monthly mean)

1

2

3

>3

1

=.I

Social rank of animals

FIG.27. Relationship between social ranks, adrenocortical activities, and heart rates of wild European rabbits, which lived in large field enclosures. Corticosterone measures of four reproductive seasons from animals from a 22,000-m2 field enclosure. Blood samples for hormone analysis were taken from all animals once every month 1 hr after maximal stimulation of their corticosterone release by an injection of ACTH; corticosterone values of the males are absolute serum concentrations; since corticosterone measures of females show marked variations during the reproductive season (see Fig. 14). their values are given as deviations from the monthly mean of all females. All data are means ( 2 SEM) of the mean of 4-6 values of each animal from one reproductive season. Heart rates of the animals were determined telemetrically by transmitters implanted ( M 2 SEM of 30-60 days of measurement per individual) in animals that lived in about 150-m2 enclosures. Social ranks were determined by behavioral observations (> 40 hr per animal). Animal numbers are shown at the bottom of the bars.

groups, which may explain that, especially under unstable social conditions in wild rabbits as well as in other species, no rank-dependent differences in adrenocortical activities are found. The heart rate in dominant individuals of both sexes, living in smaller enclosures (about 150 m2) in groups of 2-3 males and as many females, was also lower compared to that in subordinates, and every change of rank resulted in a corresponding change of heart rate (Fig. 27; see also Eisermann, 1992). Most stress research has been performed on laboratory rats. Rats are highly social and intensive fighting is present only for as long as the animals are unfamiliar with each other and no stable hierarchies have developed

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

67

Two or more females per group 1-male groups

5

I

2-3 males per group

280

b

In

E .

c 0) ._ (1

240 200

In

W -

-

160

L

0

I

al

> 3 males per group I

280 0

& + In

._ 8

5

0

240 200

160

alone

1

2

3

Social rank of males

FIG. 28. Relationship between social ranks and corticosterone challenge values of wild European rabbits in relation to their group composition. Data from animals living without male rivals in their territories are shown in top and bottom parts. See Fig. 27 and text for further details.

(Barnett, 1975; Calhoun, 1963). During this period all animals show typical stress responses including elevated blood pressure. Once a stable hierarchy with a dominant male has developed, fighting more or less ceases and blood pressure decreases (Henry et al., 1993). Nevertheless, differences between dominants and subordinates persist, as has been shown by Dijkstra and colleagues (1992) in male Wistar rats housed in mixed-sex groups in complex colony cages. Compared to pairwise housed controls, the dominant animals exhibited significantly heightened testosterone plasma levels, while those of the subordinates were in the range of the controls; corticosterone plasma levels were increased in both ranks, but in subordinates the increase was about 150%, three times higher than in dominants. If the composition of a mixed-sex group of rats is changed regularly, thus preventing the establishment of a hierarchical social system, this persistent stress can result in a progressive rise of systolic blood pressure over a period of months (Fig. 29). There are, however, marked differences between different strains of rats in their cardiovascular response to chronic

68

DIETRICH VON HOLST

h

160 -

0

E

5

Long-Evans rats

.+ . pairwise housed (n = 15)

I

150

-

4

H*

. stable colony ( n = 14)

6 unstable colony (n= 14)

2 2 ln u)

g

140 -

*

0

1

2

3

4

5

6

Months after start of experiment

FIG. 29. Effects of housing conditions on the systolic blood pressure of male Long-Evans rats. Significant differences to initial levels: *p < .OS; **p < .01; ***p < ,001 (see text for further details). Adapted from Henry er af. (1995). with kind permission from LippincottRaven Publishers, Philadelphia.

stress, which is found to correlate with their aggressiveness: The very aggressive Long-Evans rats show a great increase of blood pressure, the less aggressive Sprague-Dawley rat, a modest increase, and no change is observed in the peaceable Wistar-Kyoto (hyperactive) strain (Henry ef al., 1993). The same relationships found between aggressive behavior and blood pressure responses also seem to apply to individual differences within a strain (Bohus et al., 1987; Fokkema, 1985, Fokkema and Koolhaas, 1985; Fokkema et al., 1988; 1995). These authors tested the aggressive behavior of male laboratory rats (strain TMD-S3) in several resident-intruder tests. Following these precolony tests, 10 males together with 5 sterilized females were transferred into a large colony cage, which was fitted with small boxes in which the animals could find shelter. Cannulas were attached to most males for intermittent direct blood pressure measurements and blood sampling. In this seminatural situation, the levels of aggressive behavior in individuals correlated with those levels determined in the precolony resident-intruder tests: The more aggressive in the precolony tests the more competitive were the rats during confrontations in the colony, whether they became dominant animals exhibiting offensive behavior, or subdominant animals exhibiting defensive behavior or flight. However, blood pressure as well as plasma corticosterone levels in dominant animals tended to be lower than those in equally competitive subdominant animals.

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

69

In contrast to the actively competing dominant and subdominant animals, lower blood pressure was observed in nonaggressive rats (“subordinates”) as well as in formerly dominant animals that had lost their position after severe defeat (“outcasts”) and made no further attempts to defend themselves. Barnett (1975) termed such frequently defeated males “omegas”; behaviorally they correspond to submissive tree shrews: They are inactive, socially withdrawn, and die after a few weeks without having suffered any wounds. This was demonstrated by the study by Fokkema (1985), mentioned earlier, as well as in a life span study of laboratory rats under seminatural conditions (Blanchard et al., 1988). An even more dramatic decline in blood pressure in the course of chronic conflict has been reported by Adams and Blizard (1987) in S/JR rats (a salt-sensitive strain), which were repeatedly exposed to the presence of a trained fighter rat (Long-Evans) and subjugated by it. Thus, chronic high blood pressure is the result of continuous attempts of socially active animals to adapt to an environment that is both threatening and demanding; on the other hand, loss of control (as also seen in submissive tree shrews) results in decreased blood pressure. The relevance of the stability of a social position for the physiological status of an animal is also evident from studies by Russian scientists on hamadryas baboons (Papio hamadryas) and rhesus monkeys (Macaca mulatfa).The heart rates of these monkeys were recorded telemetrically using a transmitter placed on the monkeys’ backs in the pocket of a jacket. Dominant males of both species, kept in groups of 2 males and 1-2 females, always had lower heart rates than the subordinates and these differences could be reversed after experimentally induced changes of the social positions of the individuals. The higher heart rates in subordinate monkeys were not related to increased locomotor activity, but, according to the authors, reflected the degree of emotional tension (Cherkovich and Tatoyan, 1973). Any challenge to the stable position of a dominant male results in dramatic cardiovascular stress responses. An example of this was shown in dominant hamadryas baboons, which had lived for months with a harem of several females and their young. If the dominant male was not allowed access to his former group, which now lived with a rival male in an adjacent enclosure, the former harem owner at first tried fiercely to attack the new harem owner through the bars again and again. This behavior, however, ceased after a few weeks. Nevertheless, over a period of several months, hypertension, coronary insufficiency, myocardial infarction, and other somatic diseases developed, leading to the death of many former harem owners (Lapin and Cherkovich, 1971).

70

DIETRICH VON HOLST

Social stress as a determining factor in coronary artery disease has also been implicated by Hamm and associates (1983) in Java monkeys (Macaca fascicularis), which were kept in groups of 5 males for nearly 2 years. Subordinate individuals in stable groups had significantly heavier adrenal glands and more extensive coronary artery stenosis than did their dominant counterparts. In repeatedly reorganized “unstable groups,” dominant males developed greater blood pressure and arteriosclerosis of coronary arteries, but this occurred only in the more aggressive and highly competitive individuals, which retained dominant status over the whole study (Kaplan et al., 1982, Manuck et af., 1983; Shively and Kaplan, 1984). Subsequent studies have demonstrated that individuals of both sexes, exhibiting a heightened cardiac response to a standard stressor (threat of capture), probably sympathetic in origin, also develop the most extensive coronary lesions (Manuck et al., 1986; 1989, 1995). This is consistent with observations in humans on relationships between behavioral reactivity (“type A behavior”), sympathetic arousal, and cardiovascular disease (Dembroski et al., 1983; Houston, 1992). Similar results have also been found in field studies on monkeys. Dominant male olive baboons (Pupio anubis) living in stable groups in the East African savannah exhibited lower “initial” levels of cortisol (10 min after darting), but responded relatively faster and more strongly following stress due to anesthesia. In this way, differences between high- and low-ranking males were compensated (Sapolsky, 1982), as with findings in rhesus monkeys (Sassenrath, 1970). Additionally, subordinate olive baboons were less responsive to dexamethasone-induced cortisol suppression than were dominant males, which was due to a selective decrease of glucocorticosteroid receptors in the hippocampus (Brooke et al., 1994; Sapolsky, 1983, 1990). Since dexamethasone resistance is a typical indicator for reactive depression in humans, these results may indicate a similar state in animals with a long history of social instability and lack of control. Finally, there were indications of cardiovascular pathologies following prolonged periods of subordination. Compared to dominant individuals, subordinate animals exhibited significant reductions in high-density cholesterol, which can promote arteriosclerosis and coronary heart disease (Sapolsky and Mott, 1987). In studies on male rhesus monkeys, placed in groups of four in large cages for several months, Hamilton and Chaddock (1977) even demonstrated death of apathetic (“submissive”) individuals after a rank order had developed among the males. In contrast to all other males, the two animals concerned neither battled for dominance nor did they flee from attack. In general they crouched in the corner of the cage and, although attacked on occasion, they were not grossly maltreated. They appeared

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

71

helpless in dealing with the situation and did not display the “fight or flight” syndrome. In New World monkeys, rank-related differences have also been demonstrated. Dominant squirrel monkeys (Saimiri sciureus) in stable heterosexual colonies as well as in newly formed groups have lower plasma cortisol levels than subordinate individuals (Candland and Leshner, 1974; Manogue et al., 1975). Furthermore, during group formation, concentrations of urinary catecholamines increased only in the midranking individuals who successfully fought to maintain their status, but decreased in those animals who were unsuccessful and became further subordinated. Similar endocrine findings have also been described for the African talapoin (Miopifhecustalapoin) (Eberhart et al., 1983; 1985). Sandra Vellucci (1990) manipulated the behavior of dominant and subordinate male talapoin monkeys with drugs that are used in the treatment of human psychiatric disorders, such as anxiety and depression. In order to maximize the number of interactions between dominant and subordinate animals, groups of males were allowed to interact daily for a period of 50 min with females. This led to intense fights for control between dominant individuals, while subordinate individuals retreated, huddled in corners, moved very little, and showed high levels of visual monitoring. As her results indicate, the behavior of dominant individuals is more susceptible to drugs that are known to decrease levels of anxiety in humans, whereas subordinate individuals appear more susceptible to treatment with antidepressant drugs. This clearly demonstrates different emotional states in the individuals, depending on their social position within this stressful situation. Overall, these results indicate lowered adrenocortical and sympatheticoadrenomedullary activities in dominant monkeys living in stable groups, while in unstable situations, heightened activities of both stress systems are present in high-ranking individuals actively trying to attain control and/ or dominance. There are, however, contradictory results even in closely related species of primates. Thus, Shively and Kaplan (1984) found that in Java monkeys, dominant males in well-established mixed-sex groups exhibited higher blood pressure and more advanced arteriosclerosis than subordinates, while the latter had heavier adrenal glands, indicative of heightened adrenocortical activity. Furthermore, McGuire and associates (1986) failed to detect a clear relationship between cortisol levels and dominance status in established colonies of vervet monkeys (Cercopifhecus aethiops sabaeus), while during competition for dominance, plasma cortisol increased in all males. The same has been demonstrated for baboons living in natural conditions (Alberts et al., 1992; Sapolsky, 1990).

72

DIETRICH VON HOLST

In contrast to males, data on relationships between social status and adrenocortical and adrenomedullary stress responses in females are largely missing. As shown by Christian (1980) in an extensive review, subordinate females in most rodent species are characterized by larger adrenal glands, indicating an increased adrenocortical activity. This was also demonstrated by Schuhr (1987) through direct measurement of plasma corticosterone in female laboratory mice housed in groups. As mentioned previously, in our study on wild rabbits, we found lower adrenocortical activities and heart rates in high-ranking males and females, but only when stable group compositions and a sex ratio of about 1 male to 1-2 females prevailed. In one of the few studies on female monkeys, Gust and colleagues (1993b) examined the relationship between specific social behavior and serum cortisol concentrations in rhesus monkeys. The subjects were 9 females living in an established long-term (8 years) mixed-sex group with their young, while a second group of 9 females was formed 5 months prior to the onset of the study and made up of animals initially unfamiliar to each other. During the 1-year study, the rank of the females correlated significantly with cortisol levels in the established group, with higher serum levels exhibited by subordinate individuals. This was not the case in the recently formed group. In addition, the authors demonstrated that cortisol levels were not only negatively influenced by aggressive interactions, such as receiving bites, but also positively influenced by sociopositive interactions, such as being groomed. In contrast to most studies, Creel and associates (1996) described higher fecal glucocorticosteroid levels in dominant African wild dogs (Lycaon pictus) of both sexes, as well as in samples of urine of dominant female dwarf mongooses (Helogale parvula), both living under natural conditions in the wild. Although measurements made on urine and feces samples must be interpreted cautiously, these data indicate higher adrenocortical activities in dominant females in both species, and also in males in African wild dogs. Details on group composition and stability are, however, not provided by the authors. To summarize, group formation in primates as well as in other species requires the establishment of a social structure. This process is typically characterized by high levels of aggression, particularly among males, with a return to baseline levels within a few days or weeks (e.g., Bernstein and Mason, 1963). As demonstrated by the data given above, the process of establishing a dominance hierarchy represents a potent psychosocial stressor in all mammalian species, and usually affects lower ranking animals more greatly. Social subordination and defeat in aggressive encounters usually leads to increased adrenocortical activity and this relationship has been found in both recently formed and established social groups. Further-

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

73

more, active coping with social subordination in stable social systems, or active efforts to maintain a dominant rank in unstable groups, may eventually have health-impairing consequences when effects extend to the cardiovascular system. Such stress-related pathological conditions are evident even in those primate species in which overt fighting and injury are infrequent in the maintenance of dominant-subordination hierarchies. It must be emphasized once more that social hierarchies do not always result in rank-dependent stress states. In some species, life in well-established social systems is possible without any negative physiological stress effects on any group members, as has been shown by Sachser (1994b) in guinea pigs, and the same may apply to other species as well. b. Gonadal System. As already mentioned in the earlier sections, the effects of social stress on reproduction are profound and in extreme instances can result in sterility in both sexes within a few days. In their classic studies on small mammals housed in stationary and in freely growing populations, Christian and many others have demonstrated that the various endocrine responses of animals in crowded situations decrease natality at every possible physiological level: Crowding inhibits growth and development of reproductive organs in males and females; in addition to inhibiting spermatogenesis in males, it inhibits estrus and ovulation in females and it may delay or inhibit implantation. induce fetal reabsorption, and cause damage to or the loss of litters. Furthermore, changes in the endocrine state of stressed females may influence the physiology and behavior of their progeny (Sachser and Kaiser, 1996). Thus, prenatal stress has been associated with feminized sexual behavior in males and altered behavior in females, as well as with various changes in exploratory behavior, cognitive performance, and aggression. In 1958, Christian and LeMunyan described the effects of crowding of pregnant female laboratory mice on two generations of their offspring! These results have been repeatedly confirmed in recent years. Furthermore, after birth, reduced lactation and the retarded growth of progeny may delay their maturation and increase their morbidity. As these data have been reviewed in many excellent papers (e.g., Christian, 1978, 1980; Christian et af., 1965; Collaer and Hines, 1995; Krebs, 1978; Krebs and Myers, 1974; Lee and McDonald, 1985; Ward, 1984), I shall deal here only with some more recent results. Dominant male sugar gliders (Petaurus breviceps), living in stable colonies consisting of four males and one female, are heavier than socially subordinate males, have significantly higher plasma testosterone and lower cortisol levels, are more active, and are the only males that exhibit scentmarking behavior. When transferred into a foreign stable colony, former dominant males became subordinate and exhibited a reduction or loss of behavioral measures associated with dominance and a concomitant de-

74

DIETRICH VON HOLST

crease in plasma testosterone and rise in cortisol over a period of 3 weeks (Mallick et al., 1994). The same relationships between high plasma levels of testosterone and dominance have also been observed in males of many other species (e.g., humans: Booth et al., 1989; Elias, 1981; Mazur and Lamb, 1980; McGrady, 1984; nonhuman primates: Alberts et al., 1992; Coe et al., 1979; Keverne et al., 1982; Leshner and Candland, 1972; Mendoza et al., 1979; Rose et al., 1971, 1974, 1975; Sapolsky, 1982, 1983, 1985a,b; Schiml et al., 1996; rats, voles, and mice: review, Christian, 1980; guinea pig: Sachser, 1994a; Sachser and Prove, 1986; tree shrews: von Holst, 1969; see also Fig. 17). As plasma testorterone levels as well as the weights of testosteronedependent organs are usually correlated with dominance rank and sometimes also with the frequency of aggressive behavior in stable social systems (e.g., monkeys: Alberts et al., 1992; Rose et al., 1971; laboratory rats: Koolhaas et al., 1980; Monder et al., 1994), it is sometimes assumed that individuals with higher initial testosterone levels and therefore heightened levels of aggression will gain dominant rank positions. This conclusion is, however, not justified. Mendoza and associates (1979) housed male squirrel monkeys either alone or in groups of three males with or without a female. While prospective dominant males housed alone had the lowest plasma testosterone levels compared to the subordinate individuals, their testosterone levels were highest in all-male groups, and this effect became even more pronounced in the presence of females (Fig. 30). Gonadal endocrine activity changes very quickly during dominance interactions, as was shown already in 1973 by Bronson and associates in their

0Males alone

;ii 280

5m -

_n

Male groups Male-female groups

. E

m c

210 140

c

2 Q) 5 0

70

I

u)

$

0

Rank 1

Rank 2

Rank 3

FIG. 30. Relationships between social rank, housing conditions, and plasma levels of testosterone in male squirrel monkeys; 3 males per rank. Increase of testosterone levels in dominant and decrease in subordinate males in the different test situations significant a t p < .01. Adapted from Mendoza er al. (1979), with kind permission from Elsevier Science Ltd, The Boulvard, Langford Lane, Kidlington OX5 IGB. United Kingdom.

STRESS A N D ITS RELEVANCE FOR ANIMAL BEHAVIOR

75

studies on laboratory mice (Bronson, 1973; Bronson and Marsden, 1973; Bronson et al., 1973). The authors grouped 4 adult males per cage for periods of time ranging from 1 hr to 14 days. During the first hour all males fought intensively to establish dominance orders; at the same time, their plasma levels of corticosterone increased by about a factor of five and their gonadotropin levels decreased by about 20% for FSH and by more than 90% for LH. Plasma corticosterone levels returned to baseline levels between days 1 and 3 in dominants and between days 3 and 6 in the case of subordinates. In dominant mice, the most conspicuous effect was the increase in weight of their preputial glands, which produce an aggressionprovoking pheromone, while preputial glandular weight decreased in subordinates by about 30% within 14 days. The same relationship between preputial glandular weight and rank has also been found in laboratory rats (Dijkstra et af.,1992). In this study the authors also demonstrated a strong increase in testosterone plasma levels in individuals as a consequence of a successful fight for a dominant position. In their elegant studies on laboratory rats, Koolhaas and associates (1980) followed the endocrine changes in males during and after confrontations, by repeated blood sampling using cannulas inserted into blood vessels. During the 1-hr encounters, plasma testosterone concentrations rose in victors as well as in losers, but the rise was significantly greater in victors than in losers. About 30 min after the start of the confrontation, plasma testosterone concentration in both victors and losers started to decrease. Victors regained their original baseline levels about 90 min after the end of the confrontation, whereas testosterone levels in losers continued to decline, reaching about 20% of initial levels 4 hr after the end of the confrontation, and most defeated rats maintained these lowered baseline levels for several days (Schuurman, 1981). In summary, increased testosterone levels, such as are usually found in high-ranking males, are the consequences rather than the cause of high rates of aggression, as exogenous manipulations of testosterone concentrations within the physiological range do not cause parallel changes in rates of aggression or other testosterone-modulated behaviors (e.g., Booth et af., 1989; Dixon, 1979 Mendoza ef al., 1979; Monaghan and Glickman, 1992; Rose, 1985). On the other hand, loss of control as evident in subordinate individuals or loss of a dominant status is associated with suppressed testosterone levels, and can even lead to sterility within a few days (see also Rose 1985; Rose et al., 1972; 1974). As is the case in the adrenocortical and sympathetico-adrenomedullary systems, the gonadal endocrine system is not activated by the physical exertions of successful fighting but by the emotional processes induced by it. Accordingly, Mazur and Lamb (1980) have shown in human males that

76

DIETRICH VON HOLST

only those that win a contest (leading them to perceive that their status is thereby improved) show an elevation of testosterone levels. This is apparently also the case in nonhuman mammals, as can be deduced from the findings of our research on tree shrews (Kaiser, 1996). In order to differentiate between the physical and psychological effects of confrontations on dominant and subordinate male tree shrews, 2 males confronted each other for 10 min daily over a period of 14 days, in an experimental cage that could be divided into two identical subdivisions by a wall. The confrontations always led to low-key fights, which resulted right from the beginning in definite dominance relationships. Outside the confrontation periods each animal was separated from its rival by a wooden partition (“without visual contact”) or a wire mesh partition (“visual contact”). Outside the confrontation periods, dominant individuals in visual contact with their rivals were less active and rested more compared to the days before the confrontation period. In addition, their daily excretion rates of cortisol decreased after the start of confrontations, while the excretion of testosterone and the in vitro proliferation rate of their lymphocytes increased (Fig. 31). By contrast, dominants without visual contact with their rivals showed no changes in behavior or physiological parameters in comparison with initial values. Thus, only constant visual contact with the subordinate opponent and the emotional process of “elation” probably thereby induced modulated the behavior and physiology of dominant individuals. As expected, subordinates in visual contact with their dominant rivals showed opposite reactions to those of their opponents: a slightly increased locomotor activity and urinary cortisol excretion, as well as a decrease in testosterone excretion and the in vitro proliferation rates of their lymphocytes (Fig. 31). Surprisingly, subordinate animals without visual contact with their rivals showed qualitatively similar reactions to dominant animals in visual contact with their subordinate rivals (Fig. 31). Thus, low-key fights during the daily confrontations had no negative effects on the behavior or physiological parameters of the subordinates. Our data even point to an improved physiological state of these individuals, which may be due to the high level of control and predictability which these animals perceive in this situation (Fig. 31). In female mammals, social subordination is associated with a diminished number of ovulatory cycles and hence also with impaired reproductive success (Dittus, 1979; Drickhammer, 1974; Sade et al., 1976; Silk etal., 1981; Walker et al., 1983; Wilson et al., 1978; Wise et al., 1985). Most studies have been carried out on rodents and, as they have been reviewed extensively,

STRESS A N D ITS RELEVANCE FOR ANIMAL BEHAVIOR

77

Locomotory activity

0

-5 -10 -15

-20 Testosterone excretion with urine

30 15

0 -15

-30

’

I

I

Lymphocyte proliferation

I * 100 50

0 -50

-100

Visual contact

Without visual contact

FIG.31. Effect of visual contact on locomotory behavior (medians), testosterone excretion ( M 2 SEM), and in vitro lymphocyte proliferation after Con A stimulation ( M 2 SEM) of 14 dominant and 14 subdominant male tree shrews. After 10 days of habituation to the experimental room all animals were daily confronted for 10 min over a 2-week period; at other times they were separated by either a wooden partition (without visual contact) or a wire mesh partition (visual contact). Blood samples were taken 1day before the first confrontation and 1 day after the last confrontation. Urine was collected over the whole period and the individual means of each animal’s excretion rates over 8 days before the confrontation were used as initial values. Locomotory behavior was determined daily for 3 hr. Significant differences: *p < .05; **p < .01.

they will not be covered here (e.g., Christian, 1978, 1980; Christian et a!., 1965; Krebs, 1978; Krebs and Myers, 1974; Lee and McDonald, 1985). The same relationships between rank and reproductive success were demonstrated in primates as well as in other mammalian groups. Among macaques

78

DIETRICH VON HOLST

living in natural or seminatural environments, subordinate females are less likely to become pregnant, and their pregnancies are more likely to result in abortion, stillbirths, or neonatal death than are those of dominant females. Furthermore, in free-ranging populations of rhesus monkeys at Cay0 Santiago and La Parguera and in wild populations of toque macaques, subordinate female genealogies were found to exhibit a lower intrinsic rate of natural increase than those of dominant females (Dittus, 1979; Drickhammer, 1974; Sade et al., 1976). This was also demonstrated in groups of adult female Java monkeys housed in harem groups consisting of one adult male and five to six females (Adams et al., 1985). To induce social instability and social disruption in three groups, the females were redistributed every 12 weeks for a period of 24 months (unstable groups), while the remaining groups served as stable controls for the duration of the study. Compared to socially dominant females, subordinate individuals had fewer ovulatory menstrual cycles, more cycles with deficient luteal plasma progesterone concentrations, increased adrenal weights, and increased heart weights. Social instability, however, influenced none of these variables. These results indicate that impaired reproductive success observed in subordinate female macaques may be related, at least in part, to changes in ovarian function. The same relationship between rank and reproductive success has been found in many studies on European wild rabbits under seminatural conditions (e.g., Garson, 1979; Myers and Poole, 1962; Mykytowycz, 1959a,b). In our studies on wild rabbits we also found higher reproductive success and individual fitness in females, depending on their rank at the time of their insemination and pregnancy: Compared to subordinate females, dominant individuals gave birth to more litters per year, the weight of the young was higher at birth and at weaning, and mortality during the nest period was lower (Fig. 32). This last feature is due mainly to decreased milk production in females of subordinate ranks, which leads to the starvation of their young. The lower number of litters produced by subordinate females is apparently not due to sterility or delayed implantation, but results from a high rate of resorption and abortion of entire litters during pregnancy, as was verified by hormone analysis. As a consequence of this higher reproductive success in dominant females, there are also rank-dependent differences in the fitness of the individuals (Fig. 33). A particularly interesting effect of dominance on reproductive success has been demonstrated in dwarf mongooses. In these group-living carnivores, the oldest male and female dominate reproduction, while the younger and subordinate group members are reproductively suppressed and provide care for the offspring of the oldest pair. Nevertheless, subordinate males from several wild populations in Tanzania exhibited urinary testosterone

STRESS A N D ITS RELEVANCE FOR ANIMAL BEHAVIOR

79

Litters per female @/year)

4.5

1

I

3.6

2.7 Birth weight of young (9) 50 1

45

40 Mortality of young before weaning (%) 30

1I

.

15

0

1

2 3 Social rank of females

>3

FIG.32. Relationships between social rank and reproduction of female wild European rabbits living in a 22.000-m2 field enclosure. Data from about 50 females and 4 years; the numbers of young are indicated in the bars of the bottom figure. Data in figure “Mortality” are means; other figures means ? SEM. See text for further details. Unpublished data from H. Draxler. 1996.

levels corresponding to those of dominants. They were, however, apparently prevented from mating by dominant male aggression. In contrast, subordinate females exhibit a decreased ovarian function (Creel et al., 1992). As Keane and associates have shown in a subsequent paper (1994), subordinates of both sexes mate and about 20% of all young had subordinate mothers or fathers. Those subordinates that reproduced were of higher rank than those that did not. Among the primates, marmoset monkeys and tamarins demonstrate an extreme form of rank-dependent fertility. In laboratory colonies, as well

80

DIETRICH VON HOLST

Social rank and reproductive success of males and females in % 100 139 animals in 7 years

75 50 25

0 1

2 3 Rank of females

1 >I Rank of males

FIG. 33. Relationships between social rank and reproductive success of male and female European rabbits. Data are percentages of all young that survived until the reproductive season following the year of their birth. The mothers of the litters were determined by observations, and the fathers were determined by multilocus DNA fingerprinting. See text for further details. Unpublished data: after Zobelein (1996).

as under natural conditions, only the socially dominant female of each group reproduces, while ovulation in subordinate females is always suppressed. This infertility is immediately reversed when subordinate females are removed from their group and housed singly. As shown by Abbott and associates (1988), the social suppression of fertility in the subordinate females is apparently mediated by impaired hypothalamic GnRH secretion. The most impressive example of socially induced contraception is known from naked mole rats (Heterocephalus glaber), which live in colonies of up to 300 animals entirely underground in the semiarid regions of East Africa. In the wild as well as in captivity, there is only 1 breeding female, the “queen,” and 1-2 breeding males in each colony, while all other animals are infertile workers or play defensive roles within the colonies. Suppression of reproduction in nonreproductive females appears to be induced by ovulatory failure due to insufficient gonadotropin secretion from the anterior pituitary gland and the same suppression of gonadotropin release is also evident in the nonreproductive (subordinate) males (Abbott et al., 1989; Sherman et al., 1991). It is probable that in all of the cases mentioned so far, neural responses associated with psychosocial stress operate through the hypothalamopituitary-gonadal axis, to induce ovarian dysfunction and subsequent infertility or pregnancy failure, such as has been demonstrated in small mammals (Christian, 1980). Support for this is provided by the findings of elevated plasma prolactin concentrations and failure of the estrogen-induced LH

STRESS A N D ITS RELEVANCE FOR ANIMAL BEHAVIOR

81

surge in socially subordinate ovariectomized talapoins (Bowman et al., 1978; Keverne et al., 1982), and lower fecal estrogen levels found in subordinate wild yellow baboons during the luteal phase of ovarian cycles, when compared to females of high rank (Wasser, 1996). Furthermore, Packer and associates (1995) found some of the strongest evidence for the advantage of high rank in primates in their 30-year study of olive baboons (Papio cynocephalus anubis) at the Gombe National Park: Dominant females had shorter interbirth intervals, improved infant survival, and accelerated maturation of their daughters. It must be mentioned in this context, however, that the authors also reported negative effects on several aspects of female reproductive success, which they interpreted as the reproductive cost of dominance. However, as pointed out by Altmann and colleagues (1995), this conclusion is not supported by their data and might have resulted from inadequate interpretation of external signs of early pregnancy. c. Zmmune System. Most earlier research into the relationship between the social behavior of animals and their immune system and resistance to disease stemmed from crowding experiments and was conducted mainly on mice. Davis and Read (1958) conducted a series of experiments on the influence of daily fighting on wild-stock house mice that were infected parenterally with about 125 Trichinella larvae. Each mouse was housed in a separate cage and from day 3 through 11 after infection half of the mice were placed in groups of 5-6 animals for 3 hr a day, while the other half were left separated. All mice were killed on the 15th day after infection. About 25% of the singly housed mice were infected with an average of 9 worms apiece, whereas all grouped mice were infected and had an average of 32 worms. Furthermore, severe and prolonged fighting among crowded male mice impeded the development of acquired immunity to the dwarf tapeworm and also increased the reinfection rate in mice with wellestablished acquired immunity (Weinmann and Rothman, 1967). The effects were clearly rank dependent: Four days after a second dose of tapeworm eggs (3500 eggdmouse) the dominant mice had an intestinal cysticercoid count of 27, comparable to that in nonstressed mice exposed to the same infection; however, the counts in subordinate individuals ranged from 108 (rank 2) to 685 cysticercoids (lowest rank of the 8 males). In a similar study, Tobach and Bloch (1958) demonstrated a significantly reduced survival time to an acute tuberculosis infection in socially stressed mice (20 individuals per cage) in comparison to singly housed controls. Furthermore, Edwards and Dean (1977) found that laboratory mice of both sexes kept at high animal numbers (30 and 60 animals per cage) exhibited reduced antibody production (against typhoid paratyphoid vaccine) and reduced resistance to disease. This was evident from the significantly higher mortality rate following an injection of Salmonella typhimurium, in compar-

82

DIETRICH VON HOLST

ison to groups with lower animal numbers (2 or 10 animals per cage). The inflammatory response to subcutaneous implants of cotton pellets moistened with turpine as well as the formation of granulation tissue is also reduced in grouped mice (Christian and Williamson, 1958). Finally, Temoshok and Peeke (1988) found differences in induced tumor growth in two experiments on adult female Syrian hamsters placed in groups of ten: Females ranked as dominant by the authors exhibited reduced tumor growth compared to the subordinate individuals. These results indicate an influence of social disturbances on disease susceptibility due to immunomodulatory processes. One of the earliest experimental proofs of this stems from research carried out by Vessey (1964), who examined the antibody production against bovine serum in male laboratory mice. Previously isolated mice were placed together in groups of 6 each for 4 hr daily. They were injected with bovine serum 5 days after grouping and were found to have significantly lower titers of circulating antibodies than isolated control mice. Vessey (1964) also provided the first indication of rank-dependent immunological changes: The winners of confrontations exhibited substantially higher titers of antibodies than did the losers; likewise, T lymphocytes of subordinate mice showed a distinctly reduced in v i m response to mitogenic stimulation and reduced interleukin 2 production compared to their dominant counterparts (Hardy et al., 1990). Correspondingly, Ebbesen and associates (1991) found a lower incidence of virus-induced leukemia in dominant mice compared with subordinates. Similar suppressive effects of defeat or subordinate social rank on immunological parameters have also been found in many other species. In one of the earliest studies on rats, Raab and associates (1986) found higher tyrosine hydroxylase activities in both dominants and subordinates compared to individually or pair-housed rats (controls) after 10 days of chronic cohabitation. Only subordinates, however, lost body weight and they exhibited plasma corticosterone levels more than twice as high as those in dominants and controls. In addition, they had smaller thymus glands and a reduced lymphocyte response to in vitro mitogenic stimulation, while the values of dominants did not differ from those of controls. In laboratory rats housed in colonies, rank-dependent alterations in various components of the cellular and humoral systems have also been demonstrated (Bohus et al., 1992). Taken together, the results of these authors indicate an improvement of the immune system in dominant rats and to a lesser degree also in actively coping subdominants, while most immunological parameters in subordinates and outcasts are clearly suppressed. In piglets housed in mixed-sex groups in large pens for 80 days, Hessing and associates (1994) demonstrated clear relations between rank and sus-

STRESS A N D ITS RELEVANCE FOR ANIMAL BEHAVIOR

83

ceptibility to disease and immune reactivity. Based on dominance fights and a food competition test, piglets were divided into high-, middle-, and low-ranking groups. Dominant individuals showed a higher in vitro lymphocyte response to an Aujeszky disease virus, less severe clinical signs of disease, and threefold lower mortality rates compared to the individuals of lower rank. Similar findings have also been described in farmed red deer hinds (Hanlon et al., 1995). Corresponding results have also been demonstrated for primates. Gust and associates (1991) studied the immunological consequences of social stress associated with the formation of a new group of 8 unfamiliar adult female rhesus monkeys, introduced into an outdoor enclosure along with 1 adult male. The establishment of a stable dominance hierarchy, apparent within 48 hr, was accomplished without serious fighting and in a complete absence of wounding. While humoral components of the immune system (IgG, IgA, IgM) were not significantly influenced over the period of colony formation, within 24 hr all females generally showed a significant increase in cortisol plasma levels and a 30% decrease in absolute numbers of total lymphocytes as well as CD4+ and CD8+ T cells. These changes were significantly greater in the 4 lowest ranking females compared to those with higher ranks. After 1 week the T cell subsets of the high ranking females had returned to initial levels or exhibited even higher levels; however, values of the low-ranking females returned more slowly to baseline levels and were still low 9 weeks after group formation. This was in spite of the fact that there were no significant differences in aggressive (offensive or defensive) or affiliative behaviors between the two groups, with the exception of grooming: High-ranking subjects were groomed significantly longer than the subordinates. Recent studies by Gust et al. (1996) on female pigtail macaques (Macaca nernestrina) confirmed these results. Furthermore, highranking males in small stable groups of male rhesus monkeys exhibited significantly higher lymphocyte proliferation than middle- or low-ranking individuals. Regrouping of the animals led to an increase in aggressive behavior and plasma cortisol levels and a decrease in the lymphocyte proliferation response to a mitogen (Clarke et aZ., 1996). The same effects have been found in male Java monkeys, with particularly strong immunosuppressive effects among those monkeys showing high levels of fear behavior (Line et al., 1996). In a field study, Alberts and colleagues (1992) found significantly lower lymphocyte counts and a higher basal cortisol concentration in an adult male that had entered a stable group of olive baboons, as well as in those individuals that were victims of the intruder’s aggression, than in noninvolved individuals. The same suppressive effects of crowding stress, repeated regrouping or defeat, and social subordination on resistance against

84

DIETRICH VON HOLST

parasites, bacterial and viral diseases, tumor growth, as well as on humoral and cellular immunological parameters have also been found in many other studies, conducted mainly on mice, rats, and rabbits (e.g., 1994; Brayton and Brain, 1974a,b; Edwards and Dean, 1977; Edwards et al., 1980; Fleshner et al., 1989; Hardy et al., 1990; Hoffman-Goetz et al., 1991; Mykytowycz, 1961; Plaut et al., 1969; Stefanski and Ben-Eliyahu, 1996; Stefanski and Hendrichs, 1996; Stefanski et al., 1996; Tecoma and Huey, 1985; for reviews, see also Ader and Cohen, 1985; Monjan, 1981; Plaut and Friedman, 1981; Riley, 1981). Amazingly, the odor from stressed laboratory mice alone can induce an altered immune function in conspecifics after 24 hr of odor exposure, and lead to a decrease in the number of cells forming antibodies to sheep red blood cells (Zalcman et al., 1991), as well as to a decrease in production of interleukin 2 by Con A-stimulated spleen cells, and decreased activity of natural killer cells (Cocke et al., 1993). In contrast to the observed suppression in cell-mediated responses, stress-odor exposed mice had an enhanced humoral immune response to KLH. Thus, even in a given strain, stressors do not necessarily affect all immune measures unidirectionally, which cautions against premature conclusions based on a limited selection of cellular or humoral immune parameters. The general conclusion can be drawn from these and many other studies, that “stress” as determined by adrenocortical activation can increase susceptibility to infectious diseases. However, there are exceptions to this generalization (e.g., Moynihan et al., 1994). In their study on male and female laboratory rats that were submitted for 4 weeks to different forms of regrouping, Klein and associates (1992) found clear indications of heightened adrenocortical activity (adrenal enlargement and increased basal corticosterone levels) and thymus involution. However, compared to undisturbed controls, neither natural killer cell activity, splenocyte reactivity to mitogens, nor the rate of spontaneous development of antibodies against a common pathogen of the respiratory tract of mice (Mycoplasma pulmonis) were changed in the stressed animals. This could be due to genetic differences in immunological responsiveness, as was demonstrated in 1955 by Tobach and Bloch. These authors used strains of rats and mice that varied in degree of “emotionality,” and found that the most “emotional” strains had the shortest mean survival times after a standard dose of intravenously administered tuberculosis bacilli. Similarly, Friedman and Glasgow (1973) found, depending on mouse strain, that grouped laboratory mice are more susceptible to Plasmodium berghei than individually housed animals. Likewise, Fauman (1987) demonstrated that, relative to subordinate animals and isolated controls, dominant laboratory mice have a reduced anti-

STRESS A N D ITS RELEVANCE FOR ANIMAL BEHAVIOR

85

body response to an antigen (keyhole limpet hemocyanin). The extent of the reduction of the antibody response in dominant mice was related to the intensity of their aggressive behavior during confrontations. Similar results have also been found in male laboratory rats (Bohus et al., 1993) as well as in female chimpanzees (Pun trogfodytes) living in captive colonies (Masataka et al., 1990). Furthermore, according to Hausfater and Watson (1976), high-ranking individuals in groups of free-living yellow baboons (Papio cynocephafus) of both sexes exhibited a higher fecal parasite ova emission than more subordinate individuals. Subadult individuals generally occupy lower ranks and have lower egg counts than older ones. However, examination of the mean egg counts in adult individuals only continued to show a correlation between egg output and dominance rank in males, but not in adult females. There are many possible reasons for these and other contradictions. In relation to investigations carried out on laboratory animals, some of these contradictions may be due to differences between the various strains. They could also, however, be due to differences in housing conditions and in the duration and type of stress. Very often the necessary information needed to assess this question is missing, as are appropriate control groups. In addition, the detailed observations required to supply reliable information on social ranking and stress levels in individuals have often not been carried out (see also Bohus and Koolhaas, 1991). Furthermore, data on the activity of the sympathetico-adrenomedullary system, which could indicate the presence of social tension, are completely lacking. Finally, all of these studies are based on a rather limited selection of immune parameters, which makes general conclusions on the function of the immune system impossible. In spite of these shortcomings, all investigations do indicate a strong response of immune parameters to socially stressful situations. d. Physiological Costs of Male Dominance. In general, engaging in social conflict exposes individuals to the risk of injury and attacks by predators, diverts precious energy from reproductive activities and feeding opportunities, and may enhance vulnerability to disease. In the long term these costs are weighed against potential benefits for the dominant individual of ready access to mates with high reproductive success (e.g., Huntingford and Turner, 1987; Maynard Smith and Price, 1973; Riechert, 1988). In fact, one of the most prominent views on subordinate animals is that they have less access to mates and consequently leave fewer offspring than do dominant animals, an idea that was advanced by Zuckerman (1932) and Maslow (1936) in the 1930s for primates. This concept is widely accepted today and its validity has been demonstrated for many species in the wild as well as in captivity (e.g., Ellis, 1995; Miczek eta!., 1991). Several studies in laboratory conditions have also shown that females, when given a choice, tend to

86

DIETRICH VON HOLST

associate and mate with dominant males (e.g., lemmings: Huck and Banks, 1982; rats: Carr et al., 1982; bank voles: Hoffmeyer, 1982; Shapiro and Dewsbury, 1986; hamsters: Brown et al., 1988; White, 1986; vervet monkeys: Keddy, 1986). The relationship between social status and susceptibility to disease is currently of great interest among some evolutionary biologists, as parasite burdens may influence several aspects of social and sexual behavior (e.g., Barnett and Sanford, 1982; Dobson and Hudson, 1986; Edwards, 1988; Edwards and Barnard, 1987; Freeland, 1981; Kavaliers and Colwell, 1995; Moore and Gotelli, 1990; Rau, 1983; Read, 1990; Toft and Karter, 1990; Wedekind, 1994). In 1982, Hamilton and Zuk proposed that because of the genetically based interactions between parasites and their hosts, females are expected to choose mates based on their resistance to pathogens. Male secondary sex characters or ornaments were supposed to have evolved at least in part as indicators of this resistance. According to the authors, females should prefer males with fewer parasites, an indication of which is given by the degree of the development of secondary sex characters. Many, though not all, tests designed to prove this hypothesis have been supportive. On the basis of the higher susceptibility of human males compared with females to a variety of bacterial, viral, and parasitic diseases, Marlene Zuk proposed in 1994 that high-ranking males are more vulnerable to diseases: “The deleterious effects of testosterone may be an unavoidable price paid by males for achieving reproductive success in a competitive environment.” This hypothesis, however, seems rather unlikely, at least for mammals. As shown in the previous sections, it is absolutely possible that under conditions of social instability, dominant individuals fighting actively for control may develop cardiovascular diseases that may shorten their life. Their immunological resistance and therefore their resistance to bacterial, viral, and parasitic diseases is, however, usually higher than that of subordinates, especially in stable social situations. Furthermore, the hypothesis is based on rather doubtful premises. The author writes: “Social dominance has also been demonstrated to be testosterone dependent, with experimental castration generally reducing aggression and subsequent testosterone injections usually causing its return.” Although these effects of castration and subsequent testosterone replacement have been demonstrated in several mammalian species, the results are of limited relevance to the hypothesis put forward by Zuk. Behavioral endocrinology has shown that aggressive behavior in male mammals is predominantly determined by genetic influences, which apparently modify prenatally through testosterone those central nervous structures involved in the expression of aggressive behavior (e.g., de Ruiter et

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

87

al., 1993). Postnatally, these modifications are based on experience, so long as certain minimal testosterone levels are present, and this is probably the case in most mammals in natural conditions (e.g., Monaghan and Glickman, 1992; Sachser et al., 1994; Scott and Fredericson, 1951). Testosterone levels are indeed usually higher in dominant males, but it must be emphasized again that these increased testosterone levels are the consequence, rather than the cause of the high-ranking positions. There are, at least to my knowledge, no conclusive published data indicating that intact males of a given species (or strain) with higher initial testosterone levels are more aggressive and successful in confrontations than rivals with lower levels. Rather, the opposite seems to be true (see also Figs. 17 and 30) and, accordingly, it is not possible to increase aggression or other testosteronemodulated behaviors in male mammals with normal serum androgen levels by means of testosterone treatment (e.g., Clarke et al., 1996; Leshner, 1981; Monaghan and Glickman, 1992; van Oortmerssen et al., 1987; Rose et al., 1972; 1975). Additionally, Zuk states in this context: “No one could be more stressed than the males of many vertebrate species during the mating season, when courtship displays are exhausting, the environment must be constantly scrutinized for competitors and those competitors fought off. . . .” Although this statement applies to most species, a decrease in the response of immune parameters during the mating season has not been demonstrated so far, apart from in the highly stressed marsupial, Antechinus, mentioned in this chapter’s introduction (for a recent review, see also Nelson and Demas, 1996). In our wild rabbits there is definite evidence of an improved immunological state during the mating season compared to the nonmating season (Fig. 34). The same objections have to be raised to the postulation by Zuk of a relationship between testosterone levels of fertile males and their immunological resistance. In general, increased testosterone levels during fetal life as well as in adult males after puberty are thought to reduce cell-mediated immunological resistance, although it must be pointed out that these conclusions are based on studies on very few laboratory animal species. Furthermore, there is considerable controversy concerning the effects of sex hormones on antibody formation and unspecific biological resistance (e.g., Grossman, 1984; Madden and Felten, 1995; McCruden and Stimson, 1991; Olsen and Kovacs, 1996; Schuurs and Verheul, 1990). Nevertheless, castration of adult males of those species examined so far does at least lead to increases in their cellular immune resistance compared to that of fertile males. However, Zuk’s conclusion that this relationship also applies to fertile individuals is not supported by the literature. In contrast, most data pre-

88

DIETRICH VON HOLST

Testosterone 0.5

1

I

(ng/ml

Lymphocyte proliferation

immunoglobulin G

36 32 20

24

20 Reproductive season

Nonreproductive season

FIG. 34. Serum testosterone levels (40 males) and immunological parameters (117 males) ( 2 SEM) from the reproductive period (April-September) and the nonreproductive period (October-March). Differences between the seasons were always significant at p < ,001.

of wild European rabbits living in a 22,000-m2 field enclosure. Means

sented in the previous section indicate an improved immunological resistance for dominant males, along with increased testosterone plasma levels. Preliminary data from our laboratory collected on European rabbits, tree shrews, and Long-Evans laboratory rats also contradict this hypothesis: Injection of fertile males with physiological doses of testosterone over a period of 2 weeks had no recognizable immunosuppressive effects in any of these 3 species; on the contrary it even increased several cellular immune parameters in rabbits that were kept under constant laboratory conditions. e. Summary. In a stable dominance hierarchy, the dominant individuals can predict and actively control the outcome of social interactions, they have priority of access to food, mates, and other resources. On the whole, this situation increases the fertility and health of dominant individuals, while the opposite is usually true for subordinate individuals. This endocrinological advantage of a dominant social position may, however, be very small or even nonexistent, depending on the social system and the species (Table 111). The contrasting situation of instability occurs in the wild, when new animals migrate into a social group and destabilize the status quo, or when individuals die. In captivity, such instability is evident when social groups are first formed. In this case, the situation is very different for dominants when compared to stable systems. Typically, the rates of aggressive interactions are elevated, and are focused on animals in high-ranking positions. Rank shifts may occur repeatedly and unexpectedly. This is a situation that

89

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

TABLE I11 RELATIONSHIP BBI'WEEN SOCIAL STATUS, COPING STYLE, A N D PHYSIOLOGICAL RESPONSE PATTERN Dominant Control of situation or position Coping style Pituitaryadrenocortical function Sympathicoadrenomedullary function Pituitary-gonadal function Immune function

Subordinate

low, rank threatened active (offensive) slightly elevated

low

loss of control

active (defensive/ offensive) slightly elevated

passive and apathetic markedly elevated

unchanged

markedly elevated

greatly elevated

elevated

elevated

reduced

unchanged or often improved

?

reduced

unchanged or even reduced markedly reduced greatly reduced

high active (fights usually not necessary) unchanged or decreased

exhibits anything but control and predictability. All individuals, especially the top-ranking individuals, are therefore experiencing stress (Table 111). These relationships between control and predictability of a social situation and the physiological response pattern of an individual may explain most of the discrepancies between studies on social rank, physiology, and stress-related diseases in different mammalian species. It must be emphasized again that, in order to gain reliable information on the physiological state of an individual, it is not sufficient to focus on one stress system only, as is the case in most studies so far conducted, but information has to be collected at least on pituitary-adrenocortical as well as on sympatheticoadrenomedullary systems. Furthermore, conclusions about the influence of social situations on the immune status of an individual must be based on a large number of different immune parameters, to avoid the premature conclusion that a situation is without immunological effects.

3. Disruption of Social Bonds a. Introduction. Dominance has been used in the preceding text in a very general way, as a shorthand term that indicates the outcome of agonistic or competitive interactions between two individuals. Dominance relationships are usually more pronounced in males and are of overwhelming importance to all aspects of the life of group-living mammals-they influence their behavior, reproductive success, and health.

90

DIETRICH VON HOLST

Dominance hierarchy has often tacitly been assumed to be an equivalent term to social organization. However, dominance hierarchies based on agonistic behavior are only one aspect of social systems. Although much less conspicuous, social bonds, usually based on attachment between individuals, are at least as necessary for the establishment and stability of social systems as are dominance relationships. Attachment between mothers and their infants is usually a precondition for the survival of infants in mammals, and later social bonds to peers and mates, as well as to adults of the same sex, may develop, which profoundly influence the behavior of the animals. Such bonds, in terms of social support, can also play a positive role in health by presumably altering the way in which a potentially stressful situation is perceived (House et al., 1988; Levine, 1993a; Unden et al., 1991). It is, therefore, not surprising that the loss of a social bond and/or lack of social support may result in strong stress responses and an increased risk of mortality, as is indicated by numerous epidemiological studies in humans (e.g., Berkman and Syme, 1979; Broadhead et al., 1983; Dyer et al., 1980; Gilman et al., 1982; House et al., 1982; Kannel et al., 1987; Perrson et al., 1994; Schoenbach et al., 1986). Although social relationships play an important role in most mammalian societies, research into the relevance of social bonds and the stress-buffering effects of sociopositive interactions are usually neglected in stress research on nonhuman mammals. b. Mother-Infant Bond. Mammalian young are born defenseless and are highly dependent on their mothers for a relatively long period of time. Therefore, they have to learn to bond to their mothers, which is essential for their nurture and social development and has been dramatically demonstrated in rhesus monkeys in Harlow’s classical studies (e.g., Harlow and Suomi, 1974; Harlow et al., 1971; Rosenblum and Plimpton, 1981; Suomi, 1976). The mother-infant bond is crucial for the infant to learn to overcome fear of novel stimuli and to control aggression in social settings in later life. Maternal separation from infants is one of the most profound stressors for monkeys, and usually results in the death of the young in the wild (e.g., Thierry et al., 1984). In the laboratory, it has been used as an animal model for separation and depression. Extreme passive stress responses characterized by increased excretion of urinary 17-hydroxycorticosteroids and plasma cortisol levels, as well as strong immunomodulatory responses occur in infant rhesus and squirrel monkeys in response to separation from their mothers (e.g., Coe and Scheffler, 1989; Hrdina and Henry, 1981; Levine, 1993b; Levine et al., 1985; Wiener et al., 1992). Infants (age around 30 weeks) of macaques also showed behavioral changes and depressed in vitro lymphocyte proliferative responses to T-cell-specific mitogens over the 14-day separation period, while response to a B-cell-selective mitogen was not significantly affected. In addition, decreased natural killer cell

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

91

activity and significant alterations of different lymphocyte populations occurred (Lubach et al., 1996). Following reunion, both behavior and immunological parameters returned to initial values. Maternal responses to separation were usually similar to those of their infants (Laudenslager et al., 1982; Reite et al., 1981). The behavioral responses of infants to separation from their mothers differ among species, even when they are as closely related as the pigtail and bonnet macaques. Both species exhibit an initial agitation phase, characterized by distress vocalizations, high levels of locomotion, and other behavioral attempts by the infant to relocate and reestablish contact with its mother. Only pigtail macaques, however, exhibit a second depressive phase in this response (Boccia et al., 1995). This difference is apparently due to the different social environments of the young of these two species. Because bonnet macaques exhibit lower levels of aggression and higher levels of social contact than pigtail macaques, mothers are less restrictive and permit their infants to freely interact with other group members. As a consequence, when separated from their mothers, these infants are adopted by one of the other females. Boccia and collegues (1995) tested the effect of this different socialization directly. They examined the behavior of two infant pigtail macaques, who grew up in an environment of elevated aggression induced by a feeding paradigm, which allowed the individuals to feed only one after another. As a consequence, the high-ranking mother always had free access to the food, while the subordinate mother became restricted. This situation had striking effects on the social relationships of the infants of these two mothers. The infant of the unrestricted dominant female exhibited close attachments to four other group members, representing over 80% of the social interactions, whereas the infant of the subordinate mother restricted his social interactions to the mother. When the two infants were separated from their mothers, the second infant without alternative attachments, but not the first, became profoundly depressed and spent nearly 50% of its time during the separation exhibiting a depressive slouched posture (Boccia et al., 1991). Furthermore, the authors showed that social support from older peers can be protective. In a social group containing six infants, three had significant attachments with three older juveniles in the group, while the others were attached only to their mothers. The authors removed all mothers and juveniles from the group except for the three previously identified juveniles. Thus, out of the six remaining infants, only three retained social relationships with the juveniles with whom they had already had relationships prior to the separation. As a measure of social support, the authors took the number of affiliative behaviors directed by the juveniles to each infant. This measure demonstrated the strong protective effects of social

92

DIETRICH VON HOLST

support after mother-infant separation. Behaviorally, infants with social support showed less evidence of depression, as was reflected in play and eating behaviors. They also exhibited no change from baseline function in natural killer cells, while infants without support showed a 40% decrease from baseline function 2 hr after separation. Similar studies have been performed on only a few nonprimate species, such as the laboratory rat. Removal of mothers of laboratory rats at an age of 2 weeks markedly decreased the heart rates of their infants to about 60% of the normal rate during the following 2 days, which was followed by leveling off and recovery during the next few days (e.g., Hofer, 1981, 1994). In addition, increased plasma corticosterone baseline levels and adrenal responsivity to acute stressors were evident even several days after a single 24-hr period of maternal deprivation (Rosenfeld et al., 1992; Takahashi, 1991). The social bonds between mothers and their infants have also been evaluated in guinea pigs. Guinea pigs are capable of coordinated locomotion almost immediately following birth. Maternal care is minimal. When infants at an age of about 2 weeks were separated from their mothers for 30 min and transferred to an unfamiliar cage in an unfamiliar room, they exhibited high rates of vocalization and almost doubled plasma cortisol levels. The presence of their mothers reduced vocalization and plasma cortisol levels significantly, while the presence of an unfamiliar lactating female produced no effect over a period of hours (Hennessy and Ritchey, 1987). c. Bonds between Juveniles and Adult Individuals. The separation from peers can also result in strong physiological responses. The removal of squirrel monkeys from their companions resulted in a strong decrease of the lymphocyte proliferation to the mitogen Con A. The decrease reached significance within the first day, was maximal on the second day, and returned to initial levels within 7 days. As in common marmosets (Johnson et al., 1996) and most other primate species, plasma cortisol levels in squirrel monkeys peaked during the first day after separation, but took much longer to return to baseline levels than did the immune parameter (Coe, 1993; Friedman et al., 1991;Levine et al., 1989). The independence of adrenocortical activation and immune responses has also been demonstrated in juvenile rhesus monkeys, which were removed from their natal social group to peer housing at the age of 2 years. The highest plasma cortisol levels and greatest decrease of total blood lymphocytes and several T cell subsets (CD4+ and CD8+) were observed on the first day. While their adrenocortical activities returned to baseline levels within about 2 weeks, immune measures remained decreased for up to 2 months (Gust ef al., 1992). In contrast to juveniles, adult male rhesus monkeys showed no stress response to separation from their group (Gust et al., 1993a), while separation of adult individu-

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

93

als from their group in rats and sheep causes a strong activation of the pituitary-adrenocortical system (cited from Toates, 1987). Ratcliffe and associates (1969) studied the psychological response of swine to separation after the social bonds of grouped animals had been established. Swine housed pairwise or in groups, responded to human visitors with grunts and squeals for a handout. Competition among the males was very low and was limited to pushing and shoving. By contrast, separated swine, especially the normally sociable females, failed to respond to visitors, lying unresponsive and refusing offers of added food. After a year of isolation the separated females showed a significantly greater development of arteriosclerosis than those that were grouped. These data suggest that the lack of social bonds may result in sustained emotional disturbance and pathophysiological changes. Kaplan and associates (1991) examined the relationship between the aggressive and affiliative behavior and cellular immune parameters in adult male Java monkeys living in small groups, whose members were periodically redistributed over months. While the authors did not find any influence of social status on the immune parameters, the in vitro lymphocyte proliferation in reaction to the two T-cell-selective mitogens concanavalin A (Con A) and phytohemagglutinin (PHA) was greatest in individuals that were both highly affiliative and exhibited low levels of aggression. Furthermore, natural killer cell activity was highest among highly affiliative males, regardless of their levels of aggression. These findings indicate that the cellular immune competence may be enhanced among monkeys that, in response to a disrupted social environment, spend large amounts of time in affiliation with other males, or in males that seek and find social support. 4. Social Support and Its Stress-Reducing Effects As these findings show, numerous factors influence the magnitude of the physiological stress response, and one of the most important variables in pairwise or group-living mammals appears to be the presence of a familiar social partner. Social support generally reduces the magnitude of stress responses and it has a stress-buffering effect, as was impressively demonstrated by Levine and associates in their studies on squirrel monkeys (Levine, 1993a,b). The authors exposed a well-established group of adult squirrel monkeys to a live Boa constrictor that was confined in a plastic box. Although direct physical contact between the monkeys and the snake was prevented, all monkeys showed increased levels of vigilance, agitation, and avoidance behavior. A strong adrenocortical activation, however, was observed only when the monkeys were tested individually, but not when tested together as a group. Surprisingly, this stress-buffering effect appears only when adult squirrel monkeys are exposed to this situation together

94

DIETRICH VON HOLST

with multiple partners. In pair-housed individuals no social buffering was evident, although the behavioral signs of arousal were reduced (Coe et al., 1982). In this context, the findings of Mendoza and Mason (1986) are of special interest. They compared the effects of intruders on behavior and adrenocortical activities of polygynous squirrel monkeys and monogamous titi monkeys (Cullicebus rnoloch), housed as heterosexual pairs. In titi monkeys, the presence of an intruder resulted in marked behavioral signs of agitation, especially in the subjects of the same sex as the intruders. Plasma cortisol levels of females showed no consistent changes to intruders of either sex, while those of males were always increased in the presence of a male rival. Squirrel monkeys of both sexes, on the other hand, responded to female intruders with a reduction in plasma cortisol to below baseline levels, whereas a male intruder had no effect. Maintenance of a monogamous social structure, such as in titi monkeys, is presumably based on a bond between the male and female of the pair and the exclusion of male rivals by the male. In squirrel monkeys, which usually live in large groups of both sexes, life as a pair is lacking in the usual companionship and could be improved by new individuals. Although this interpretation is not without contradiction, these results nevertheless demonstrate impressively that the social system of a species influences the behavioral and physiological responses of the individuals to conspecifics in very different ways. The relevance of the quality of the relationships between individuals on their stress-reducing effects are also evident from guinea pigs (Sachser et ul., 1998). In mixed-sex colonies male guinea pigs develop long-lasting and strong bonds to some females, while no such social ties exist to other females. When male guinea pigs are taken from such colonies and placed singly into unfamiliar cages their plasma cortisol levels increase for hours by about 100%compared to initial levels. Presence of an unfamiliar female from a different colony or a familiar female from their own colony to which no social bonds exist has no stress-reducing effects. There is, however, a sharp reduction in the endocrine stress response when each male is transferred into an unfamiliar cage together with a female with whom a social bond exists (Fig. 35). The relevance of social integration and social bonds to members of a group is especially evident in our work on wild rabbits. Throughout the reproductive phase, each female usually produces 5-6 litters at monthly intervals, resulting in up to 30 progeny per year. Depending on the number of adult females in our enclosure, up to 1000 animals are born each year. However, about 70% of the young are taken by predators (e.g., cats, martens, weasels, hawks) before the onset of the winter season and only an average of 5% of the original number actually survive the winter. The

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

6 m

95

225

c

c

200

s -

175

u) 0 ._

r

8

150

E

$ 125 h 100

familiar

bonded

unfamiliar

Females

FIG. 35. Cortisol values ( M -C SEM) of 10 male guinea pigs 2 hr after transfer into an unfamiliar enclosure alone (dotted horizontal line), together with an unfamiliar female, a familiar but unbonded female, or a bonded female. All data are percentages of initial levels. Significant differences between the effects of the presence of bonded and unbonded conspecifics: *p < .05; **p < .01. Adapted from Sachser et al. (1997).

number of surviving juveniles varies from one year to the next (from 0 to more than 40 individuals) and is approximately equivalent to the number of adults that have died. Consequently, the number of adult rabbits at the beginning of each reproductive season has remained surprisingly constant over the past 10 years (at 50-95 individuals). Predation plays only a minor role in mortality during the winter months (November to February). Death is usually due to an extreme loss of weight, based on the breakdown of all fat reserves as well as large quantities of muscle tissue, culminating in hypoglycemic shock. Although these findings point to starvation, a general lack of food cannot be the reason for death, as all the adults as well as those juveniles that survive the winter show no loss in body weight. Rather, the moribund juveniles are incapacitated, in spite of increased food intake, by extensive parasitic damage to their intestinal epithelium, which prevents the digestion and/or resorption of food. In addition, toxins produced by the changed intestinal flora probably also contribute to the death of the animals. Within the last few weeks prior to death, the number of oocysts and nematode ova in the feces of the moribund juveniles increases dramatically and parallel to the loss in weight (Fig. 36). In comparison to the surviving individuals, this parasitic infestation is probably due to a reduced immune resistance against the parasites, as indicated by a reduced in v i m lymphocyte proliferation (Fig. 37), a de-

96

DIETRICH VON HOLST

Body weight

Nematodes

Coccidia

1200

900

-

600

6

2

0

6

2

0

6

2

0

Weeks before death

FIG. 36. Changes of body weight and numbers of nematode eggs (predominantly Trichostrongylus retorfaefornzins and Gruphidium strigosum) and oocysts of several Elimera species in the feces of 20 subadult European wild rabbits during the last weeks before their death in the winter period. All data are means ( 2 SEM). See text for details.

creased number of T lymphocytes in the blood, and a 50% reduction in the phagocytic capacity of the leucocytes. Based on our current findings, mortality during the winter months appears to be a result of socially induced immune suppression: Young animals usually leave their native groups in autumn and attempt to join other groups (Kunkele and von Holst, 1996). In this process, all immigrants are initially attacked and chased away by members of the group. However, some juveniles are tolerated after a while and integrated into the group, although most do not achieve this social integration. The successful integration of a juvenile is indicated by its spatial position within the group and by its behavior toward the adults: Integrated animals restrict their whereabouts more or less exclusively to the existing territory of a group of adults, while nonintegrated animals tend to roam over a wide area and from one group to the next. In addition, integrated animals are observed either in close proximity to or in direct contact with individuals of a group during 30% of observation time, while this is seldom the case in nonintegrated individuals (Fig. 38: Spatial integration). Although aggressive reactions by adults are directed against integrated and nonintegrated juveniles with almost equal frequency, integrated animals are more often involved in friendly interactions with adults. While in integrated animals two out of three interactions with adults are of a friendly nature, nonintegrated animals are only involved in one sociopositive interaction for every three aggressive ones (Fig. 38: Social integration). Finally, both groups also differ significantly from each other in immunological measures: nonintegrated individuals clearly exhibit lower values than integrated animals (Fig. 38: Immune measures).

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

Body weight

97

LP after Con A

Food intake 6.0

4.0

2.0 0.0 Nematodes

Coccidia

Intestinal villi

450

800

300

400

150

0

0

Surv

Died

Surv

Died

Surv

Died

FIG. 37. Body weight, parasites in feces, and length of the intestinal villi of 20 subadult wild rabbits at their death during the winter (Died) as well as their food intake and in vitro lymphocyte proliferation (LP) after Con A stimulation 2-6 weeks before their death. All measures were also determined at corresponding times from 20 animals of about the same age that survived the winter period (Surv). All data are means (rf- SEM); significant differences: **p < .01; * * * y < ,001.

The change in integration state was followed during the winter season in several juveniles: Individuals that were first more or less integrated within a group were expelled from it, and nonintegrated juveniles were accepted. In all cases this also involved changes in immunological parameters: If integration status deteriorated, then lymphocyte proliferation was reduced; if integration status improved, that is, in the case of successful integration into a group, proliferation increased (Fig. 39). Based on these findings, an improved immune state and a reduced parasitic infestation in juveniles surviving the winter would appear to be the result of successful integration into the existing social group. Accordingly, out of more than 100 animals observed in detail over 5 years, only those animals capable of successful integration into groups during the autumn and winter months actually survived the winter. The number of juvenile wild rabbits is therefore regulated during the winter season, by giving only those individuals that have achieved integra-

98

DIETRICH VON HOLST

Local attachment

Being attacked 4.8

Cellular (LP)

Interactions per hour

3.6 2.4

1.2 0.0

Distance to adults 4o

1

Sights < 2m in %

Sociopositive behavior 4.8

lnteractmns per hour I

Humoral (IgG) 160

30

3.6

120

20

2.4

80

10

1.2

40

0

n

0.0

IN NI Spatial integration

Deviation In %of the mean of all subadults

IN NI Social integration

IN NI Immune measures

FIG. 38. Spatial integration. social integration, and immune measures of about 20 integrated ( I N ) and 30 nonintegrated (NI) subadult wild rabbits during the winter period ( M 2 SEM). Immunological parameters: means of 1-3 measurements per animal; behavioral data: means of 8-24 hr of observations per animal. Significant differences between IN and NI: **p < .01; ***p < ,001. See text for further details. Unpublished data from M. Kaschei (1996).

tion into an existing group a chance of surviving the winter. As the acceptance of juveniles into an existing group of adults is apparently dependent on the size and composition of the group, this mechanism results in optimal group composition before the onset of the reproductive season. As mentioned previously, numerous epidemiological studies on humans indicate that social bonds, in terms of social support, can play a positive role in the health of an individual. The direct physiological mechanisms are, however, far from being clear. As shown by the various studies on nonhuman mammals described above, a breakdown of social bonds elicits strong passive stress responses, while the presence of a bonded partner or group has some stress-buffering effects. Furthermore, the development of a bond can exert strong physiological consequences even in individuals that apparently beforehand had lived an unstressed life. Thus, tree shrews can be housed singly for more than 10 years in captivity and be in excellent condition without any apparent signs of stress. Neverthe-

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

E

99

Change of social integration

400

J

U m

2

-m

300

v)

’is 200 c

m

100 a,

5 $

s

0

1

5

3

4

6 7 Number of animal

0

2

FIG. 39. Changes of the in virro lymphocyte proliferation after Con A stimulation of 8 subadult wild rabbits before and within 4 weeks after change of their social integration. Unpublished data from M. Kaschei (1996).

less, the formation of a pair bond greatly improves their well-being, as indicated by physiological data. Tree shrews usually live in pairs in the wild. Putting a male and a female together, however, does not inevitably lead to the formation of a pair bond. In some instances it can result in intensive fights and-unless the animals are separated-in the death of one of the opponents (male or female). In most cases, however, especially in large enclosures, tree shrews of both sexes can coexist, although they suffer from a certain amount of social tension, as evident from occasional fights and avoidance behavior. At estrus, successful copulations may even occur, but the offspring are always cannibalized by the parents shortly after birth (von Holst, 1969). In all these unharmonious pairings, even if overt aggression is not evident, the heart rates of the animals are constantly increased, as is the case in subdominant males living together with a dominant male under constant active stress (Fig. 40). In about 20% of all pairings, however, contact between an unfamiliar male and female is characterized from the outset by amicable behavior, which conveys the strong impression of “love at first sight.” Both individuals “greet” each other frequently with

100

DIETRICH VON HOLST

long bouts of mouth licking (up to 90 min per 12-hr observation day), they move around in close contact, and mostly rest together. Copulations may occur on the first day, but are not a necessary prerequisite for such a harmonious pair bond. During the nights both animals always sleep together in the same nest box, which is never the case in the previously mentioned unharmonious pairs. In the laboratory harmonious pairs can live together for more than 10 years and breed successfully and regularly in the absence of any aggression. In all harmonious pairs we found a drastic reduction in serum levels of glucocorticosteroids and adrenocortical reactivity to standard stressors, and-even more surprising-a reduction in heart rates (Fig. 40). Furthermore, all immunological parameters that were measured indicate an improvement of the immunological state of both individuals. The opposite is true for unharmonious pairings. Amazingly, the quality of a pairing depends on personal “sympathy” or “antipathy” between the individuals. Thus, a male that has been fiercely rejected by one female can be accepted as a “loved” partner by another female. Accordingly, the physiological status of tree shrews kept as pairs changes depending on the quality of their pair bond, as shown in studies in which females were paired with different males. In standard tests females respond to males that they will accept as partners with high marking responses, and to those that they will not accept with low marking responses. Hence it was possible to pair females once with males that they accepted as partners, and once with males that they rejected (Fig. 41). As the results of these pairings demonstrate, both sexes exhibited low levels of aggression and high levels of sociopositive behaviors when females were combined with males to whose scent they had shown the highest marking responses (harmonious pairs). The opposite was the case when the females were unharmoniously paired with males whose scent stimulated their marking behavior very little (Fig. 41). Furthermore, harmonious pairings decreased serum levels of glucocorticosteroids and epinephrine, while increasing those of gonadal hormones as well as improving cellular and humoral immune measures. The opposite was true in the same individuals in unharmonious pairings (Fig. 42). Unfortunately, little is known of the physiological effects of pair formation in other species, with the exception of several studies on monogamous and polygamous species of vole. In the prairie vole (Microtus ochrogaster), long-term heterosexual pair bonds are formed, which are characterized by affiliative behaviors, such as side-by-side contact, and are independent of sexual behavior (Carter et al., 1988, 1995; Winslow et al., 1993). In contrast to tree shrews, however, prolonged mating of naive females with an unfamiliar male is necessary for the induction of a pair bond in this species (Insel et al., 1995). The

-1 500

disharmonious pairing I

single

harmonious pairing

0,

+

single

1

-

m

il

5 f

300

100

0

5

10

15

Days after beginning of pairing experiment FIG.40. Effects of an unharmonious and a harmonious pairing on the heart rate of a female tree shrew. Night periods are striped. Data higher and lower (unharmonious and harmonious pairing, respectively) than the mean of the last 3 days before the pairings are accentuated by black. Adapted from von Holst (1987). with kind permission from Gustav Fischer Verlag, Stuttgart, Germany.

102

DIETRICH VON HOLST

Marking behavior

Sociopositive behavior

Defensive behavior

90 u) (u

-

E

C 60

6.0

’c

0

.b >

m

$ 30

3.0

m

0

Harm Unharm

-

0.0

Harm Unharm

Harm Unharm

FIG. 41. The marking activities of females in response to the scent of different males were used to create harmonious (Harm) and unharmonious (Unharm) pairings (for details, see von Holst, 1985b). The more that the scent of a male stimulates the marking activity of a female the greater is the probability that the pairing with the female will result in a harmonious pair bond. Each female was therefore paired for 14 days with that male whose scent elicited the highest, and after 4 weeks of single housing with that whose scent elicited the lowest marking response (“Marking behavior”). During the pairings the behavior of each male and female was recorded for a total of 12 hr (“Sociopositive behavior” and “Defensive behavior”). All data are means ( 2 SEM); significant differences are indicated: ***p < .001.

Tes

Cor

Nor

Epi

LTT

IgG

FIG.42. The effects of harmonious (striped bars) and unharmonious (cross-hatched bars) pairings on several physiological measures of 12 male tree shrews. All data ( M ? SEM) are given as deviations from the initial levels of the males before the pairings. Abbreviations: serum levels of testosterone (Tes), cortisol (Cor), norepinephrine (Nor), epinephrine (Epi), and immunglobulin G (IgG); in vitro lymphocyte proliferation after Con A stimulation (LTT). Significant differences are indicated: ***p < ,001. See text for further details.

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

103

results of many studies indicate that centrally released oxytocin during mating may be critical to the formation of partner preferences in female prairie voles, while vasopressin appears to be more important to pair bonding in the males of this species (e.g., Carter er al., 1992; Insel and Hulihan, 1995; Williams et al., 1992,1994). Furthermore, plasma glucocorticosteroid levels may respond to and influence the development of social attachments. In naive female prairie voles, cohabitation with a male resulted in a dramatic decline in serum corticosterone levels, which facilitated pair bonding. When corticosterone levels were reduced via adrenalectomy, females developed partner preferences after 1 hr of cohabitation, while sham-operated and untreated females required 3 hr or more of cohabitation to establish partner preferences (De Vries et al., 1995). The role of oxytocin in the development of social bonds was first proposed by Klopfer (1971), who suggested that the increased oxytocin levels after birth facilitate the mother-infant attachment. In subsequent studies, the relevance of centrally released oxytocin for the development of maternal behavior and the development of mother-infant bonds was shown in sheep, rats, and some other species (e.g., Da-Costa et al., 1996; Kendrick et al., 1987; Pedersen and Prange, 1985; Uvnas-Moberg, 1994; Yu et al., 1996). Furthermore, many studies demonstrated the relevance of oxytocin for sexual behavior (in addition to sex hormones). Thus, injections of oxytocin in estrous rats stimulates sexual behavior in female rats, reduces aggression, and increases physical contact with the males (Arletti and Bertolini, 1985; Caldwell et al., 1986); similar results were also found in female Syrian hamsters (Whitman and Albers, 1995). These data indicate that oxytocin may be involved in the formation of social bonds between mothers and their infants as well as between males and females in mammals (Carter et al., 1990; Keverne, 1988). Shared sexual experience and the concomitant oxytocin release usually found in both sexes may thereby facilitate social bonds in mammals including human beings. In addition, stress-buffering effects of intracerebroventricular injections of oxytocin have been described. In laboratory rats, the development of gastric lesions induced by cold and restraint stress or by the administration of cysteamine was reduced by oxytocin treatment (Grassi and Drago, 1993). The physiological mechanisms involved are unsolved, but nevertheless, these results indicate that social bonds may improve the health of individuals by reducing their response to stressors (probably by reducing the stress-inducing properties of stressors, due to the presence of a security-providing partner), as well as by influencing the physiological state of the individual.

104

DIETRICH VON HOLST

C. CONCLUSIONS Social relationships based on agonistic and sociopositive behaviors play an important role in most mammalian societies. They determine not only the stability of social systems but greatly influence almost all behavioral elements of the individuals as well as their fertility and health. Disturbances of social relationships may lead to stress responses that differ greatly in quality and intensity, depending on the stressor and the coping behavior of the individuals. Disruption of social bonds usually initially elicits an alarm response characterized by heightened physiological and behavioral arousal. Particularly in infants separated from their mothers, this is followed by passive stress responses characterized by apathetic behavior, withdrawal, and eventually death. Information on the long-term consequences of disruption of social bonds for the health of adults is, however, lacking for nonhuman mammals. In human beings, the loss of partners can have very strong healthimpairing effects. Social conflict elicits an immediate acute alarm response in all animals, characterized by increased sympathetico-adrenomedullary and pituitaryadrenocortical activation. There is evidence that norepinephrine (the fight hormone) predominates in this first response to challenges, which is probably characterized by the feeling of anger. If the stressful situation cannot be resolved by behavioral responses (e.g., by fight or flight), differing chronic stress responses may result, the degree of which is dependent on the perception of the amount of control over a social situation, and which are therefore almost exclusively psychological phenomena. The perception that loss of control is either possible or probable appears to lead to a change from anger to fear, as is indicated by an increasing production of epinephrine (the flight hormone) and mainly active subordinate behavior. As the threatening situation continues, this active coping can shift to a more passive, apathetic mode, accompanied by greatly increased adrenocortical activity and associated with the feelings of helplessness and depression. The adrenomedullary epinephrine release can remain high or decrease by comparison to actively coping individuals. Gonadal activity (at least in males) may actually increase in early phases of successful responses to challenge (during anger), but eventually declines as loss of control threatens. The immune system responds extremely sensitively to social challenges. Every challenge to control (feelings of anger, fear, and depression) is usually accompanied by profound indications of immunosuppression. The relationships between social rank and stress response depend mostly on the stability and predictibility of the social relationships. In stable social systems the dominants are usually not, or only slightly, stressed: Com-

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

105

pared to subordinate individuals, they exhibit lower adrenocortical- and sympathetico-adrenomedullary activities as well as higher gonadal activities and immune resistance. Subordinate individuals usually show active stress responses, but in some cases also passive stress responses, which usually lead to death within a short period of time. In socially unstable systems, which are characterized by immigration processes or dominance conflicts, all individuals usually show active stress responses of varying degrees of strength as they fight to regain high control and predictability. The effects on dominants and subdominants, however, may differ depending on the species and the social situation. In many species, the highest active stress responses are found in those individuals that are dominant and fight actively to maintain their high social ranks. This may even lead to higher incidences of cardiovascular diseases and premature death, in comparison to subordinate individuals. The neuroendocrine stress responses accompanying these subjective feelings have a bipolar aspect: According to a concept proposed by Henry in 1986, the anger-fear (fight-flight) response is opposed by the serenityrelaxation state, which is characterized by enhanced grooming and resting. The opposite pole to the depression, loss-of-control, and loss-of-attachment axis, is probably a subjective feeling of elation, such as in dominant tree shrews in the presence of clear subordinate individuals-and probably also in animals with strong bonds to partners (Fig. 43). Because social relationships can influence the physiological state of individuals in so many positive or negative ways, it is not surprising that social status alone cannot always predict stress-related measures in individuals, particularly in natural conditions containing many uncontrollable social influences. This means that, in order to understand the physiological consequences of social interactions, an integrated approach is required to assess what factors, including rank and social bonds, interact to affect an individual’s fertility and health (see also Fig. 44). O r as Sapolsky (1988) puts it in his discussion of individual differences in olive baboons and their stress responses: Thus, among these primates, who you are. what your place is in your society, and what sort of society it is appear to have everything to do with your physiology, both under basal and stressed circumstances. Furthermore, one may argue at this stage that these rank-related differences in physiology are of consequence, that the pattern observed in dominant males seems to be the most adaptive.

And these comments on baboons apply to all mammalian species, from rodents to humans, as pointed out as long ago as 1977 by Henry and Stephens.

106

DIETRICH VON HOLST

FIG.43. Schematic diagram of the stress-buffering emotional processes and their physiological consequences. Adapted from Henry (1986). with kind permission from Academic Press, Inc., New York.

IV. SUMMARY Contact with conspecifics not only influences the behavior of individuals, but is also associated with marked physiological changes, which can influence their vitality and fertility in positive or negative ways, depending on the type of interaction. The term used to describe the negative effects is

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

107

+pq dominant

Acute stress Chronic passive stress

FIG. 44. Schematic diagram of the two stress axes (PAS, pituitary-adrenocortical system, and SAS, sympathetico-adrenomedullarysystem) and their activation (+) or inactivation (-) depending on the social control perceived by the individuals and the associated subjective emotions. As an example, the physiological states of tree shrews in the differing social situation described in this paper are given in circles. See text for further details.

social stress. In this chapter, the many negative physiological consequences of social stress are addressed and the stress-reducing effects of sociopositive contacts with conspecifics are described. The changes that have taken place over the past 30 years in the concept of stress are an important prerequisite to understanding the effects of social interactions on the physiology of individuals. As the current concept of stress has been developed within the boundaries of psychology and medicine, largely excluding the field of zoology, a short synopsis of this development is given. According to this concept, physiological stress reactions are generally triggered by central nervous processes (emotions or feelings), which always occur when a situation is characterized by uncertainty or unpredictability, that is, when the individual’s control over the situation is endangered or impossible. Differing physiological stress reactions are induced, depending on the behavioral strategy used by an animal to either obtain control or to cope with the situation. Active attempts at obtaining control over a situation (e.g., fight or flight) are characterized initially by

108

DIETRICH VON HOLST

activation of the sympathetico-adrenomedullary system and in the long term by cardiovascular disease (active stress). Passive perception of defeat or loss of control is characterized initially by pituitary-adrenocortical activation and in the long term by negative effects on almost all bodily functions (passive stress). In a second section, such peripheral physiological processes are described as are essential to the understanding of stress reactions. Since various stress reactions differ depending on the situation and the coping behavior of the individuals, studies based on one or only a few measures can lead to misleading or false conclusions. An introduction is also given into the most frequently used methods in obtaining indications of the activities in the pituitary-adrenocortical and sympathetico-adrenomedullary systems, as well as in gonadal and immune functions. Particular attention is paid to the limitations of these methods. In a third section, based on our research on tree shrews, an overview of the relationships between the social position of an animal and its physiological state is given. General statements cannot be made due to the close relationships between social system, social rank and rank stability, and the impact of positive relationships with other conspecifics. However, in stable social systems, an overall dominant position can improve the fertility and vitality of the individual, while subordinate individuals exhibit no or only slight active or passive stress reactions, dependent on the species. In unstable social systems, dominant individuals are characterized by particularly high active stress reactions, due to their efforts of improving or retaining their position through increased levels of aggression; subordinate animals exhibit active or passive stress reactions of variable intensities. Finally, the importance of social bonds for the health of individuals is assessed: The loss of social bonds can provoke long-term stress reactions, while the presence of bonded partners has stress-reducing effects. Because social relationships can influence the physiological status of individuals in so many positive or negative ways, social status alone cannot always predict stress-related measures in individuals, especially in natural conditions containing many uncontrollable social influences. Consequently, in order to understand the physiological consequences of social interactions, an integrated approach is required to assess which factors, including rank and social bonds, interact to affect an individual’s fertility and health. Acknowledgments

This chapter is dedicated to my late friend James Henry whose scientific results and concepts were twenty years ahead of his time. I wish to thank Norbert Sachser for his critical comments, which greatly improved this manuscript. Thanks are also due to Debby Curtis for her most

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

109

valuable help with the English version on this paper. 1 also appreciate the editors’ helpful remarks on the manuscript.

References

Abbott, D. H.. Hodges, J. K., and George, L. M. (1988). Social status controls LH secretion and ovulation in female marmoset monkeys (Callifhrixjacchus). J. Endocrinof. 117,329-339. Abbott, D. H., Barrett, J., Faulkes, C. G . , and George, L. M. (1989). Social contraception in naked mole-rats and marmoset monkeys. J. Zool. 219,703-710. Ackerman, K. D., Bellinger, D. L., Felten, S. Y., and Felten, D. L. (1991). Ontogeny and senescence of noradrenergic innervation of the rodent thymus and spleen. In “Psychoneuroimmunology” (R. Ader. D. L. Felten, and N. Cohen, eds.). pp. 71-125. Academic Press, San Diego, CA. Adams, D. 0. (1994). Molecular biology of macrophage activation: A pathway whereby psychosocial factors can potentially affect health. Psychosom. Med. 56, 316-327. Adams, M. R.. Kaplan, J. R.. and Koritnik, D. R. (1985). Psychosocial influences on ovarian endocrine and ovulatory function in Macaca fascicularis. Physiol. Behav. 35, 935-940. Adams. N., and Blizard, D. A. (1987). Defeat and cardiovascular response. Psychol. Rec. 37,349-368. Ader, R., and Cohen, N. (1985). CNS-immune system interactions: Conditioning phenomena. Behav. Brain Sci. 8, 379-394. Ader, R.,Felten, D. L., and Cohen, N., eds. (1991). “Psychoneuroimmunology ” Academic Press, New York. Adler, R., Herrmann, J. M., Kohle, K., Schonecke, 0. W., von Uexkull. T., and Wesiak, W., eds. (1986). “Psychosomatische Medizin.” Urban & Schwarzenberg, Munich. Aguilera, G., Kiss, A,, Hauger, R., and Tizabi, Y. (1992). Regulation of the hypothalamicpituitary-adrenal axis during stress: Role of neuropeptides and neurotransmitters. In “Stress: Neuroendocrine and Molecular Approaches” (R. Kvetnansky, R. McCarty, and J. Axelrod, eds.), pp. 365-381. Gordon & Breach, New York. Alberts, S. C., Sapolsky, R. M.. and Altmann, J. (1992). Behavioral, endocrine, and immunological correlates of immigration by an aggressive male into a natural primate group. Horm. Behav. 26, 167-178. Altmann, J., Sapolsky, R., and Licht, P. (1995). Baboon fertility and social status. Nature (London) 377,688-689. Anisman, H., Zalcman, S., and Zacharko, R. M. (1993). The impact of stressors on immune and central neurotransmitter activity: Bidirectional communication. Rev. Neurosci. 4,147-180. Arletti, R., and Bertolini, A. (1985). Oxytocin stimulates lordosis behavior in female rats. Neuropeprides (Edinburgh) 6, 247-253. Arnold, A. S., and Breedlove, S. M. (1985). Organizational and activational effects of sex steroids on brain and behavior: A reanalysis. Horm. Behav. 19, 469-498. Aus der Muhlen. K., and Ockenfels, H. (1969). Morphologische Veranderungen im DiencephaIon und Telencephalon nach Storungen des Regelkreises Adenohypophyse-Nebennierenrinde. 111. Ergebnisse beim Meerschweinchen nach Verabreichung von Cortison und Hydrocortison. Z. Zeflforsch. Mikrosk. Anat. 91, 126-138. Barnett, S. A. (1958). Physiological effects of “social stress” in wild rats: I. The adrenal cortex. J. Psychosom. Res. 3, 1-I 1. Barnett, S. A. (1964). Social stress. Viewpoints Biol. 3, 170-218. Barnett, S. A. (1975). “The Rat: A Study in Behavior.” University of Chicago Press, Chicago.

110

DIETRICH VON HOLST

Barnett, S. A. (1988). Enigmatic death due to “social stress.” A problem in the strategy of research. ISR, Inrerdiscip. Sci. Rev. 13, 40-51. Barnett, S. A., and Sanford, M. H. R. (1982). Decrements in “social stress” among wild Rattus rattus treated with antibiotic. Physiol. Behav. 28, 483-487. Barnett. S. A.. Hocking, W. E.. Munro, K. M. H., and Walker, K. Z . (1975). Socially induced renal pathology of captive wild rats (Ratrus villosissirnirs).Aggressive Behav. 1, 123-133. Bateman, A., Singh. A., Krai, T., and Solomon, S. (1989).The immune-hypothalamic-pituitaryadrenal axis. Endocr. Rev. 10, 92- 112. Batzli. G. 0..and Pitelka, F. A. (1971). Influence of meadow mouse populations on California grassland. Ecology 51, 1027-1039. Baum. M. J. (1992). Neuroendocrinology of sexual behavior in the male. In “Behavioral Endocrinology” (J. B. Becker, S. M. Breedlove, and D. Crews, eds.). pp. 97-130. MIT Press, Cambridge, MA. Beach. F. A. (1975). Hormonal modification of sexually dimorphic behavior. fsychoneuroendocrinology 1, 3-23. Beato, M., and Doenecke. D. (1980). Metabolic effects and modes of action of glucocorticoids. In “General, Comparative and Clinical Endocrinology of the Adrenal Cortex” (1. Chester Jones and 1. W. Henderson, eds.), pp. 117-181. Academic Press, London. Beerda, B.. Schilder, M. B. H., Janssen. N. S. C. R. M., and Mol, J. A. (1996). The use of salivary cortisol, and catecholamine measurements for a noninvasive assessment of stress responses in dogs. Horm. Behav. 30, 272-279. Berkenbosch. F., de Rijk, R.. Schotanus, K., Wolvers, D., and van Dam, A. (1992). The immune-hypothalamo-pituitary adrenal axis: Its role in immunoregulation and tolerance to self-antigens. In “Interleukin-1 in the Brain. Pergamon Studies in Neuroscience” (N. Rothwell and R. Dantzer, eds.), Vol. 5 , pp. 75-91. Pergamon, Oxford. Berkman. L. F., and Syme, S. L. (1979). Social networks, host resistance and mortality: A nine-year follow-up study of the Almameda County residents. A m . J . Epidemiol. 109, 186-204. Bernard, C. (1859). “Leqon sur les ProprietCs Physiologiques et les AltCrations Pathologiques des Liquides de I’Organisme,” Vol. 1. Baillibre. Paris. Bernstein, I. S., and Mason, W. A. (1963). Group formation by rhesus monkeys. Anim. Behav. 11, 28-31. Bernton, E. W., Bryant, H. U., and Holaday, J. W. (1991). Prolactin and immune function. In “Psychoneuroimmunology” (R. Ader. D. L. Felten, and N. Cohen, eds.). pp. 403-428. Academic Press, San Diego, CA. Besedovsky, H. O., and del Rey, A. (1991). Physiological implications of the immuneneuro-endocrine network. In “Psychoneuroimmunology” (R. Ader. D. L. Felten, and N. Cohen, eds.). pp. 589-608. Academic Press, San Diego, CA. Blalock, J. E. (1988). Immunologically-mediated pituitary-adrenal activation. A h . Exp. Med. B i d . 245, 217-223. Blanchard, D. C., Sakai. R. R., McEwen. B.. Weiss. S. M., and Blanchard. R. J. (1993). Subordination stress: Behavioral. brain, and neuroendocrine correlates. Behav. Brain Rex 58, 113-121. Blanchard, R. J., Flannelly. K. J.. and Blanchard, D. C. (1988). Life span studies of dominance and aggression in established colonies of laboratory rats. Physiol. Behav. 43, 1-7. Boccia, M. L., Reite. M. L. and Laudenslager, M. L. (1991). Early social environment may alter the development of attachment and social support: Two case reports. Infanf Behav. Dev. 14,253-260. Boccia. M. L., Laudenslager, M. L., and Reite, M. L. (1995). Individual differences in macacques’ responses to stressors based on social and physiological factors: Implications for primate welfare and research outcomes. Lab. Anim. 29, 250-257.

STRESS A N D ITS RELEVANCE FOR ANIMAL BEHAVIOR

111

Bohus. B., and Koolhaas, J. M. (1991). Psychoimmunology of social factors in rodents and other subprimate vertebrates. In “Psychoneuroimmunology” (R. Ader, D. L. Felten. N. Cohen, eds.), pp. 807-830. Academic Press, San Diego, CA. Bohus, B., Benus, R. F., Fokkema, D. S., Koolhaas, J . M., Nyakas, C., van Oortmerssen, G. A., Prins, A. J . A.. de Ruiter, A. J. H., Scheuring, A. J . W., and Steffens. A. B. (1987). Neuroendocrine states and behavioral and physiological stress responses. Prog. Brain Res. 72, 57-70. Bohus, B.. Koolhaas, M.. de Ruiter, A. J. H., and Heijnen, C. J. (1992). Psycho-social stress. differential alterations in immune system functions and tumor growth. In “Stress: Neuroendocrine and Molecular Approaches” (R. Kvetnansky, R. McCarty, and J. Axelrod, eds.), pp. 607-621. Gordon & Breach, New York. Bohus, B., Koolhaas. J. M., Heijnen, C. J., and de Boer, 0. (1993). Immunological responses to social stress: Dependence on social environment and coping ab es. Neuropsychobiology 28, 95-99. Booth, A., Shelley, G., Mazur, A,, Tharp, G.. and Kittok, R. (1989). Testosterone, and winning and losing in human competition. Horm. Behuv. 23,556-571. Bowman, L. A., Dilley. S. R., and Keverne. E. B. (1978). Suppression of oestrogen-induced LH-surge by social subordination in talapoin monkeys. Nature (London) 275, 56-58. Bradley, A. J., McDonald, I. R., and Lee, A. K. (1980). Stress and mortality in a small marsupial (Anfechinusstuarfii, Mackuy). Gen. Comp. Endocrinol. 40, 188-200. Brain, P. F., and Poole, A. E. (1974). The role of endocrines in isolation-induced intermale fighting in albino laboratory mice. 1. Pituitary-adrenocortical influences. Aggressive Behav. 1,39-69. Brambell, F. R. (1944). The reproduction of the wild rabbit Orycfolugus cuniculus (L.). Proc. Zool. Soc. London 114, 1-45. Brayton, A. R., and Brain, P. F. (1974a). Studies on the effects of differential housing on some measures of disease resistance in male and female laboratory mice. J . Endocrinol. 61,48-49. Brayton, A. R., and Brain, P. F. (1974b). Effects of differential housing and glucocorticoid administration on immune responses to sheep red blood cells in albino “TO” strain mice. J . Endocrinol. 64, 4-5. Breedlove, M. (1994). Sexual differentiation of the human nervous system. Annu. Rev. Psychol. 45,389-418. Broadhead. W. E., Kaplan. B. H., James, S. A,. Wagner, E. H.. Schoenbach, V. J., Crimson, R., Heyden, S.. Tibblin. G., and Gehlbach. S. H. (1983). The epidemiological evidence for a relationship between social support and health. Am. J. Epidemiol. 117,521-537. Bronson, F. H. (1963). Some correlates of interaction rate in natural populations of woodchucks. Ecology 44,637-643. Bronson, F. H. (1964). Agonistic behaviour in woodchucks. Anim. Behav. 12,470-478. Bronson, F. H. (1973). Establishment of social rank among grouped mice: Relative effects on circulating FSH, LH, and corticosterone. Physiol. Behav. 10, 947-951. Bronson, F. H., and Eleftheriou, B. E. (1965a). Relative effects of fighting on bound and unbound corticosterone in mice. Proc. Soc. Exp. Biol. Med. 118, 146-149. Bronson, F. H.. and ElefthCriou, B. E. (1965b). Adrenal responses to fighting in mice: Separation of physical and psychological causes. Science 147, 627-628. Bronson, F. H., and Marsden, H. M. (1973). The preputial gland as an indicator of social dominance in male mice. Behuv. Biol. 9, 625-628. Bronson, F. H., Stetson, M. H.. and Stiff, M. E. (1973). Serum FSH and LH in male mice following aggressive and nonaggressive interactions. Physiol. Behav. 10, 369-372.

112

DIETRICH VON HOLST

Brooke. S. M.. de Haas-Johnson, A. M., Kaplan, J. R.. Manuck, S. B.. and Sapolsky, R. M. (1994). Dexamethasone resistance among nonhuman primates associated with a selective decrease of glucocorticoid receptors in the hippocampus and a history of social instability. Neuroendocrinology 60, 134-140. Brown, M. R. (1991). Brain peptide regulation of autonomic nervous and neuroendocrine functions. In “Stress, Neurobiology and Neuroendocrinology” (M. R. Brown, G . F. Koob. and C. Rivier. eds.), pp. 193-215. Dekker, New York. Brown, P. S., Humm, R. D., and Fischer, R. B. (1988). The influence of a males’s dominance status on female choice in Syrian hamsters. Horm. Behav. 22, 143-149. Buck, R. (1988a). Introduction. In “Human Motivation and Emotion” (R. Buck. ed.), pp. 1-89. Wiley, New York. Buck, R. (1988b). Central nervous system mechanisms of motivation and emotion. In “Human Motivation and Emotion” (R. Buck, ed.), pp. 90-151. Wiley, New York. Bunag, R. D. (1984). Measurement of blood pressure in rats. In “Handbook of Hypertension” (W. de Jong, ed.), Vol. 4, pp. 1-12. Elsevier, New York. Bush, I. E. (1962). Chemical and biological factors in the activity of adrenocortical steroids. Pharmacol. Rev. 14,317-445. Caldwell, J. D., Prange, A. J., Jr., and Pedersen, C. A. (1986). Oxytocin facilitates the sexual receptivity of estrogen-treated female rats. Neuropeptides (Edinburgh) 7 , 175-189. Calhoun, J. B. (1963). “The Ecology and Sociology of the Norway Rat.” U . S. Public Health Service Publ., Bethesda, MD. Candland, D. K.. and Leshner. A. I. (1974). A model of agonistic behavior: Endocrine and autonomic correlates. In “Limbic and Autonomic Nervous Systems Research” (L. V. DiCara, ed.), pp. 137-163. Plenum, New York. Cannon, W. B. (1929). “Bodily Changes in Pain, Hunger, Fear and Rage.” Branford, Boston. Carr, D. J., and Blalock, J. E. (1991). Neuropeptide hormones and receptors common to the immune and neuroendocrine systems: Bidirectional pathway of intersystem communication. In “Psychoneuroimmunology” (R. Ader, D. L. Felten, and N. Cohen, eds.), pp. 573-588. Academic Press, San Diego, CA. Carr, W. J., Kimmel, K. R., and Anthony, S. L. (1982). Female rats prefer to mate with dominant rather than subordinate males. Bull. Psychon. 20, 89-9 1. Carter, C. S. (1992). Neuroendocrinology of sexual behavior in the female. In “Behavioral Endocrinology” (J. B. Becker, S. M., Breedlove, and D. Crews, eds.), pp. 71-96. MIT Press, Cambridge, MA. Carter, C. S., Witt, D. M., Thompson, E. G . , and Carlstead, K. (1988). Effects of hormonal, sexual, and social history on mating and pair bonding in prairie voles. Physiol. Behav. 44,691-697. Carter, C. S., Williams, J. R., and Witt, D. M. (1990). The biology of social bonding in a monogamous mammal. In “Hormones, Brain and Behaviour in Vertebrates” (J. Balthazart, ed.). pp. 154-164. Karger, Basel. Carter, C. S., Williams, J. R., Witt, D. M., and Insel, T. R. (1992). Oxytocin and social bonding. Ann. N . Y. Acad. Sci. 652, 204-211. Carter, C. S.. DeVries, A. C., and Getz, L. L. (1995). Physiological substrate of mammalian monogamy: The prairie vole model. Nrurosci. Biobehav. Res. 19, 303-314. Chapman, V., Desjardins, C., and Bronson, F. (1969). Social rank in male mice and adrenocortical response to open field exposure. Proc. Soc. Exp. Biol. Med. 130,624-627. Cherkovich, G . M.. and Tatoyan, S. K. (1973). Heart rate (radiotelemetrical registration) in macacques and baboons according to dominant-submissive rank in a group. Folia Primatol. 20,265-273.

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

113

Chitty, D. (1958). Self-regulation of numbers through changes in viability. Cold Spring Harb. Symp. Quant. Biol. 22, 277-280. Chitty, D. (1960). Population processes in the vole and their relevance to general theory. Can. J. Zool. 38, 99-113. Ciaranello, R. D. (1978). Regulation of phenylethanolamine N-methyltransferase synthesis and degradation. I. Regulation by adrenal glucocorticoids. Mol. Pharmacol. 14,478-489. Christian, J. J. (1950). The adreno-pituitary system and population cycles in mammals. J. Mammal. 31, 247-259. Christian, J. J. (1963). Endocrine adaptive mechanisms and the physiologic regulation of population growth. I n “Physiological Mammalogy” (W. V. Mayer and R. G. van Gelder, eds.), Vol. 1, pp. 189-353. Academic Press, New York. Christian, J. J. (1971). Population density and reproductive efficiency. Biol. Reprod. 4,248-294. Christian, J. J. (1975). Hormonal control of population growth. I n “Hormonal Correlates of Behavior” (B. E. ElefthCriou and R. L. Sprott, eds.), pp. 205-274. Plenum, New York. Christian, J. J. (1978). Neurobehavioral endocrine regulation in small mammal populations. I n “Populations of Small Mammals under Natural Circumstances” (D. P. Snyder, ed.), pp. 143-158. University of Pittsburg Press, Pittsburg. Christian, J. J. (1980). Endocrine factors in population regulation. In “Biosocial Mechanisms of Population Regulation” (N. N. Cohen, R. S. Malpass, and H. G. Klein, eds.), pp. 55-1 16. Yale University Press, New Haven, CT. Christian, J. J., and LeMunyan, C. D. (1958). Adverse effects of crowding on lactation and reproduction of mice and two generations of their offspring. Endocrinology (Ba/timore) 63,517-529. Christian, J. J.. and Wiliamson, H. 0. (1958). Effect of crowding on experimental granuloma formation in mice. Proc. Soc. Exp. Biol. Med. 99,385-387. Christian, J. J.. Lloyd, J. A,, and Davis, D. E. (1965). The role of endocrines in the selfregulation of mammalian populations. Recent Prog. Horm. Res. 21, 501-568. Chrousos, G . P., Loriaux. L. D., and Gold, P. W. (1988). The concept of stress and its historical development. Adv. Exp. Med. Biol. 245, 3-7. Clarke, M. R.. Harrison, R. M., and Didier, E. S. (1996). Behavioral, immunological, and hormonal responses associated with social change in rhesus monkeys (Macaca mularra). Am. J. Primatol. 39, 223-233. Cocke, R., Moynihan. J. A,, Cohen, N., Grota, L. J., and Ader, R. (1993). Exposure to conspecific alarm chemosignals alters immune function in BALB/c mice. Brain, Behav., Immunol. 7 , 36-46. Coe, C. L. (1993). Psychosocial factors and immunity in nonhuman primates: A review. Psychosom. Med. 55,298-308. Coe, C. L., and Scheffler. J. (1989). Utility of immune measures for evaluating psychological well-being in nonhuman primates. Zoo B i d 1, 89-99. Coe, C. L., Mendoza, S. P., and Levine, S. (1979). Social status constrains the stress response in the squirrel monkey. Physiol. Behav. 23, 633-638. Coe, C. L., Franklin, D., Smith, E. R., and Levine, S. (1982). Hormonal responses accompanying fear and agitation in the squirrel monkey. Physiol. Behav. 29, 1051-1057. Collaer. M. L.. and Hines, M. (1995). Human behavioral sex differences: A role for gonadal differences during early development. Psychol. Bull. 118, 55-107. Cowan, D. P. (1987). Aspects of the social organisation of the European wild rabbit (Oryctolagus cuniculus). Efhology 75, 197-210. Creel, S., Creel, N., Wildt, D. E., and Montfort, S. L. (1992). Behavioural and endocrine mechanisms of reproductive suppression in Serengeti dwarf mongooses. Anim. Behav. 43,231 -245.

114

DIETRICH VON HOLST

Creel, S., Creel, N. M., and Monfort, S. L. (1996). Social stress and dominance. Nariire (London) 379,212. Crnic. L. S . (1991). Behavioral consequences of viral infection. In “Psychoneuroimmunology” (R. Ader, D. L. Felten, and N. Cohen, eds.), pp. 749-787. Academic Press, San Diego, CA. Crowcroft, P. (1955). Territoriality in wild house mice. J. Mammal. 36, 299-301. Da Costa. A. P. C., Guevara-Guzman, R. G., Ohkura. S.. Goode, J. A,, and Kendrick, K. M. (1996). The role of oxytocin release in the paraventricular nucleus in the control of maternal behaviour in the sheep. J. Neiiroendocrinol. 8, 163-177. Dallman. M. F. (1991). Regulation of adrenocortical function following stress. In “Stress, Neurobiology and Neuroendocrinology” (M. R. Brown, G. F. Koob, and C. Rivier, eds.), pp. 173-192. Dekker. New York. Darwin, C. (1872). “The Expression of the Emotions in Man and Animals.” Murray, London. Davis, D. E., and Christian, J. J. (1957). Relation of adrenal weight to social rank of mice. Proc. Soc. Exp. Biol. Med. 94, 728-731. Davis, D. E.. and Read, C. P. (1958). Effects of behavior on development of resistance in Trichinosis. Proc. Soc. Exp. Biol. Med. 99, 269-272. Davis, H., Porter, J. W., Livingstone, J., Herrmann, T., MacFadden, L., and Levine. S. (1977). Pituitary-adrenal activity and lever press shock escape behavior. Physiol. Psycho/. 5, 280-284.

de Boer. S. F.. de Beun, R., Slagen, J. L., and van der Gugten. J. (1990). Dynamics of plasma catecholamine and corticosterone concentrations during reinforced and extinguished operant behavior in rats. Physiol. Behav. 47, 691-698. De Jonge. F. H., Bokkers, E. A. M.. Schouten, W. G. P., and Helmond. F. A. (1996). Rearlin piglets in a poor environment: Developmental aspects of social stress in pigs. Physiol. Behav. 60,389-396. Dembroski, T . M., Schmidt, T. H., and Blumchen, G., eds. (1983). “Biobehavioral Bases of Coronary Heart Disease.” Karger, Basel. Dempsher, D. P., and Gann, D. S. (1983). Increased cortisol secretion after small hemorrhage is not attributable to changes in adrenocorticotropin. Endocrinology (Balrimore) 113,86-93. de Ruiter. A. J. H., Feitsma, L. E., Keijser, J. N., Koolhaas, J. M., van Oortmerssen. G. A., and Bohus, B. (1993). Differential perinatal testosterone secretory capacity of wild house mice testes is related to aggressiveness in adulthood. Horm. Behav. 27, 231-239. Dess. N. K., Linwick, D., Patterson, J.. Overmier, J. B., and Levine, S. (1983). Immediate and proactive effects of controllability and predictability on plasma cortisol responses to shocks in dogs. Behav. Neiirosci. 97, 1005-1016. DeVries. A. C.. DeVries, M. B., Taymans. S., and Carter, C. S. (1995). Modulation of pair bonding in female prairie voles (Microtiis ochrogasrer) by corticosterone. Proc. Narl. Acad. Sci. U. S. A. 92, 7744-7748. Dhabhar, F. S., Miller, A. H., McEwen, B. S., and Spencer, R. L. (1995). Effects of stress on immune cell distribution. Dynamics and hormonal mechanisms. J. Immunol. 154, 551 15527.

Dijkstra, H., Tilders. F. H. J.. Hiehle, M. A,. and Smelik, P. G. (1YY2). Hormonal reactions to fighting in rat colonies: Prolactin rises during defense, not during offense. Physiol. Behav. 51,961 -968. Dittus, W. P. J. (1979). The evolution of behaviors regulating density and age-specific sex ratios in a primate population. Behavioiir 69, 265-302. Dixon, A. (1979). Androgens and aggressive behavior in primates: A review. Aggressive Behav. 6,37-52. Dobson, A. P., and Hudson, P. J. (1986). Parasites, disease and the structure of ecological communities. Trends Ecol. Evol. 1, 11-15.

STRESS A N D ITS RELEVANCE FOR ANIMAL BEHAVIOR

115

Dohrenwend. B. S.. and Dohrenwend, B. P., eds. (1974). “Stressful Live Events: Their Nature and Effects.” Wiley. New York. Drickhammer. L. C. (1974). A ten-year summary of reproductive data for free-ranging Macaca mulatfa. Folia Primafol. 21, 61 -80. Dunn, A. J. (1989). Psychoneuroimmunology for the psychoneuroendocrinologist: A review of animal studies of nervous system-immune system interactions. PsychoneicroendocrinolOgy 14,25 1-274. Dyer, A. R., Persky, V., Stamler, J., Oglesby, P., Shekelle. R. B.. and Berkson, D. M. (1980). Heart rate as a prognostic factor for coronary heart disease and mortality: Findings in three Chicago epidemiologic studies. Am. J . Epidemiol. 112, 736-749. Ebbesen. P., Villadsen, J. A,, Villadsen, D., and Heller, K. E. (1991). Effect of subordinance. lack of social hierarchy and restricted feeding on murine survival and virus leukemia. Exp. Gerontol. 26, 479-486. Eberhart, J. A,. Keverne, E. B., and Meller, R. E. (1983). Social influences on circulating levels of cortisol and prolactin in male talapoin monkeys. Physiol. Behav. 30, 361-369. Eberhart. J. A,, Yodyingyuad, U.. and Keverne. E. B. (1985). Subordination in male talapoin monkeys lowers sexual behaviour in the absence of dominants. Physiol. Behav. 35, 673-677. Edwards, E. A,, and Dean. L. M. (1977). Effects of crowding of mice on humoral antibody formation and protection to lethal antigenic challenge. Psychosom. Med. 39, 19-24. Edwards, E. A,, Rahe, R. H., Stephens, P. M., and Henry. J. H. (1980). Antibody response to bovine serum albumin in mice: The effects of psychosocial environmental change. Proc. Soc. Exp. Biol. Med. 164,478-481. Edwards, J. C. (1988). The effects of Trichinella spiralis infection on social interactions in mixed groups of infected and uninfected male mice. Anim. Behav. 36, 529-540. Edwards. J. C.. and Barnard. C. J. (1987). The effects of Trichinella infection on intersexual interactions between mice. Anim. Behav. 35, 533-540. Ekkel, E. D., Dieleman, S. J.. Schouten, W. G. P., Portela, A,, Cornilissen, G., Tielen. M. J. M.. and Halberg, F. (1996). The circadian rhythm of cortisol in the saliva of young pigs. Physiol. Behav. 60, 985-989. Eisermann. K. (1992). Long-term heart responses to social stress in wild European rabbits: Predominant effect of rank position. Physiol. Behav. 52, 33-36. Eisermann, K.,and Stohr, W. (1992). Diurnal heart rate rhythms in small mammals: Speciesspecific patterns and their environmental modulation. In “Temporal Variations in Cardiac Rhythms in Rodents” (T. H. Schmidt, B. T. Engel, and G. Bliimchen, eds.), pp. 87-91. Springer, Berlin. Eisermann, K., Meier. B., Khaschei, M., and von Holst, D. (1993). Ethophysiological responses to overwinter food shortage in wild European rabbits. Physiol. Behav. 54, 983-980. Elias, M. (1981). Cortisol, testosterone and testosterone-binding globulin responses to competetive fighting in human males. Aggressive Behav. 7 , 215-222. Ellis, L. (1995). Dominance and reproductive success among nonhuman animals: A crossspecies comparison. Ethol. Sociobiol. 16, 257-333. Elton, C. S. (1Y42). “Voles, Mice and Lemmings.” Clarendon Press. Oxford. Ely, D. L. (1981). Hypertension, social rank, and aortic arteriosclerosis in CBA/J mice. Physiol. Behav. 26, 655-661. Ely, D. L., and Henry, J. P. (1978). Neuroendocrine response patterns in dominant and subordinate mice. Horrn. Behav. 10, 156-169. Engel, B. T. (1985). Stress is a noun! No, a verb! No, an adjective. In “Stress and Coping” (T. M. Field. P. M. McCabe, and N. Schneidermann, eds.), pp. 3-12. Erlbaum, Hillsdale, NJ.

116

DIETRICH VON HOLS?

Evain, D., Morera. A. M., and Saez, J. M. (1976). Glucocorticoid receptors in interstitial cells of rat testis. J. Steroid Biochem. 7, 1135-1 139. Fagin, K. D.. Shinsako, J., and Dallman, M. F. (1983). Effects of housing and chronic cannulation on plasma ACTH and corticosterone in the rat. Am. J . Physiol. 245, E515-E520. Fauman, M. A. (1987). The relation of dominant and submissive behavior to the humoral immune response in BALBlc mice. Biol. Psychiatry 22, 771-776. Felten, S. Y., and Felten, D. L. (1991). Innervation of lymphoid tissue. In “Psychoneuroimmunology” (R. Ader, D. L. Felten, and N. Cohen, eds.), pp. 27-61. Academic Press, San Diego, CA. Fenske, M. (1989). Application of a new, simple method for quantitative collection of 24hour urines in small laboratory animals: Determination of basal excretion of proteins, creatinine, urea, electrolytes, and free steroids. Z . Versuchstierk. d. 32, 65-70. Fenske, M. (1996). Saliva cortisol and testosterone in the guinea pig: Measures for the endocrine function of adrenals and testes? Steroids 61, 647-650. Fleshner. M., Laudenslager, M. L., Simons, L., and Maier, S. F. (1989). Reduced antibodies associated with social defeat in rats. Physiol. Behav. 45, 1183-1187. Fokkema, D. S. (1985). Social behavior and blood pressure: A study of rats. Unpublished Doctoral Dissertation, Rijksuniversiteit, Groningen, Netherland. Fokkema, D. S., and Koolhaas, J. M. (1985). Acute and conditioned blood pressure changes in relation to social and psychosocial stimuli in rats. Physiol. Behav. 34, 33-38. Fokkema, D. S., Koolhaas, J. M., and van der Gugten, J. (1995). Individual characteristics of behavior, blood pressure, and adrenal hormones in colony rats. Physiol. Behav. 57, 857-862. Fokkema, D. S.. h i t , K., van der Gugten, J., and Koolhaas, J. M. (1988). A coherent pattern among social behavior, blood pressure, corticosterone and catecholamine measures in individual male rats. Physiol. Behav. 42, 485-489. Folkow, B., Grimby, G., and Thulesius, 0.(1958). Adaptive structural changes of the vascular walls in hypertension and their relation to the control of the peripheral resistance. Acta Physiol. Scand. 44,255-272. Folkow, B., Halback, M., and Lundgren, Y. (1973). Importance of adaptive changes in vascular design for establishment of primary hypertension: Studies in man and in spontaneously hypertensive rats. Circ. Res. 32/33, suppl. I, 2-13. Freeland, W. J. (1981). Parasitism and behavioral dominance among male mice. Science 213,461-462. Friedman, E., Coe. C. L., and Ershler, W. B. (1991). Time-dependent effects of peer separation on lymphocyte proliferatin response in juvenile squirrel monkeys. Dev. Psychobiol. 24, 159-173. Friedman, M., and Rosenman, R. H. (1974). “Type A Behavior and your Heart.” Alfred A. Knopf, New York. Friedman. S. B., and Glasgow, L. A. (1973). Interaction of mouse strain and differential housing upon resistance to Plasmodium berghei. J. Parsitol. 59, 851-854. Fukuhara, K., Kvetnansky, R., Weise, V. K., Ohara, H., Yoneda, R., and Kopin, I. J. (1992). Correlation of plasma catecholamine levels with tissue tyrosine hydroxylase activity in sart-stressed rats. In “Stress: Neuroendocrine and Molecular Approaches” (R. Kvetnansky, R. McCarty, and J. Axelrod, eds.). pp. 881-889. Gordon & Breach, New York. Gao, H. B., Shan, L. X., Monder, C., and Hardy, M. P. (1996). Suppression of endogenous corticosterone level in vivo increases the steroigenic capacity of purified rat Leydig cells in vitro. Endocrinology (Baltimore) 137, 1714-1718. Carson. P. J. (1979). Social organization and reproduction in the rabbit: A review. In “Proceedings of the World Lagomorph Conference 1979” (K. Myers and C. D. Maclnnes, eds.), pp. 256-270. University of Guelph Press, Guelph.

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

117

Gentry, W. D., ed. (1984). “Handbook of Behavioral Medicine.” Guilford, New York. Gilman, S. C., Schwarz, J. M., Mulner, R. J.. Bloom, F. E., and Feldmann, J. D. (1982). Betaendorphin enhances lymphocyte proliferative responses. Proc. Natl. Acad. Sci. U.S.A. 79,4426-4430. Glaser, R., and Kiecolt-Glaser, J., eds. (1994). “Human Stress and Immunity.” Academic Press, San Diego, CA. Goy, R. W.. Bercowitch. F. B., and McBrair, M. C. (1988). Behavioral masculinization is independent of genital masculinization in prenatally androgenized female rhesus macaques. Horm. Behav. 22,552-571. Grassi, M., and Drago, F. (1993). Effects of oxytocin on emotional stress and stress-induced gastric lesion. J. Physiol. (London) 87, 261 -264. Grossman, C. J. (1984). Regulation of the immune system by sex steroids. Endocr. Rev. 5,435-455. Gust, D. A., Gordon, T. P., Wilson, M. E., Ahmed-Ansari, A.. Brodie, A. R., and McClure, H. M. (1991). Formation of a new group of unfamiliar female rhesus monkeys affects the immune and pituitary adrenocortical systems. Brain, Behav., Immunol. 5, 296-307. Gust, D. A,, Gordon, T. P.. Wilson, M. E.. Brodie. A. R., Ahmed-Ansari, A,, and McClure, H. M. (1992). Removal from natal social group to peer housing affects cortisol levels and absolute numbers of T cell subsets in juvenile rhesus monkeys. Brain, Behav., Immunol. 6, 189- 199. Gust, D. A,, Gordon, T. P., and Hambright, M. K. (1993a). Response to removal from and return to a social group in adult male rhesus monkeys. Physiol. Behav. 53, 599-602. Gust, D. A.. Gordon, T. P.. Hambright, M. K., and Wilson, M. E. (1993b). Relationship between social factors and pituitary-adrenocortical activity in female rhesus monkeys (Macaca mulatta). Horni. Behav. 27, 318-331. Gust, D. A,, Gordon, T. P., Wilson, M. E., Brodie, A. R., Ahmed-Ansari, A,, and McClure. H. M. (1996). Group formation of female pigtail macacques (Macaca nemestrina). Am. J. Primatol. 39, 263-273. Gustafson, M. L., and Donahoe. P. K. (1994). Male sex determination-current topics of male sexual differentiation. Annu. Rev. Med. 45, 505-524. Hall, N. S., O’Grady, M . P., and Farah, J. M. (1991). Thymic hormones and immune function: Mediation via neuroendocrine circuits. In “Psychoneuroimmunology” (R. Ader, D. L. Felten, and N. Cohen, eds.), pp. 515-528. Academic Press, San Diego, CA. Hamilton. C. L., and Chaddock, T. (1977). Social interaction and serum insulin values in the monkey (Macaca mulatto). Psychosom. Med. 39,444-449. Hamilton, W. D., and Zuk, M. (1982). Heritable true fitness and bright birds: A role for parasites? Science 218, 384-387. Hamm, T. E., Kaplan. J . R., Clarkson, T. B., and Bullock. B. C. (1983). Effects of gender and social behavior on the development of coronary artery atherosclerosis in cynomolgus monkeys. Atherosclerosis (Shannon, Irel.) 48, 221-233. Hanlon. A. J.. Rhind, S. M., Reid, H. W., Burrells, C., and Lawrence, A. B. (1995). Effects of repeated changes in group composition on immune response, behaviour, adrenal activity and lifeweight gain infarmed red deer yearlings. Appl. Anim. Behav. Sci. 44,57-64. Hanson, J. D., Larson, M. E., and Snowdon, C. T. (1976). The effects of control over high intensity noise on plasma cortisol levels in rhesus monkeys. Behav. Biol. 16,333-340. Hardy, C. A., Quay, J., Livnat. S.. and Ader, R. (1990). Altered lymphocyte response following aggressive encounters in mice. Physiol. Behav. 47, 1245-1251. Harlow, H. F., and Suomi. S. J. (1974). Induced depression in monkeys. Behav. Biol. 12, 273-296.

118

DIETRICH VON HOLST

Harlow, H. F., McGaugh, J. L., and Thompson, R. F. (1971). “Psychology.” Albion Publ., San Francisco. Hausfater, G., and Watson, D. F. (1976). Social and reproductive correlates of parasite ova emissions by baboons. Nature (London) 262, 688-689. Heinjen, C. J., Kavelaars, A., and Ballieux, R. E. (1991). Corticotropin-releasing hormone and the proopiomelanocortin-derived peptides in the modulation of immune function. In “Psychoneuroimmunology” (R. Ader, D. L. Felten, and N. Cohen, eds.), pp. 429-513. Academic Press, San Diego, CA. Hennessy, M. B., and Ritchey, R. L. (1987). Hormonal and behavioral attachment responses in infant guinea pigs. Dev. Psychohiol. 20, 613-625. Henry, J. P. (1986). Neuroendocrine patterns of emotional response. I n “Emotion: Theory. Research and Experience” (R. Plutchik, ed.), Vol. 3. pp. 37-60. Academic Press, New York. Henry, J. P. (1992). Biological basis of the stress response. Integr. Physiol. Behav. Sci. 27,66-83. Henry, J. P., and Meehan, J. P. (1981). Psychosocial stimuli, physiological specificity, and cardiovascular disease. 111 “Brain, Behavior, and Bodily Disease” (H. Weiner, M. A. Hofer, and A. J. Stunkard. eds.), pp. 305-333. Raven Press, New York. Henry, J. P., and Stephens, P. M. (1977). “Stress, Health. and the Social Environment. A Sociobiologic Approach to Medicine.” Springer, New York. Henry, J. P., and Stephens-Larson, P. M. (1985). Specific effects of stress on disease processes. In “Animal Stress” (G. P. Moberg, ed.), pp. 161-175. Am. Physiol. SOC.,Bethesda, MD. Henry, J. P., Ely, D. L., Stephens, P. M., Ratcliffe, H. L., Santisteban, G. A,, and Shapiro, A. P. (1971). The role of psychosocial factors in the development of arteriosclerosis in CBA mice: Observations on the heart, kidney. and aorta. Atherosclerosis (Shannon Ire/.) 14, 203-218. Henry, J. P., Ely, D. L.. and Stephens, P. M. (1972). Changes in catecholamine-controlling enzymes in response to psychosocial activation of the defense and alarm reactions. In “Physiology, Emotion and Psychosomatic Illness” (R. Porter and J. Knight, eds.), Ciba Found. Symp. 8, pp. 225-25 I . Elsevier, Amsterdam. Henry, J. P., Kross, M. E.. Stephens, P. M., and Watson, F. M. C. (1976). Evidence that differing psychosocial stimuli lead to adrenal cortical stimulation by autonomic or endocrine pathways. In “Catecholamines and Stress” (E. Ursdin, R. Kvetnansky. and I. J. Kopin, eds.), pp. 457-468. Pergamon, Oxford. Henry, J. P.. Liu, Y. Y.. Nadra. W. E., Quian, C., Mormede, P., Lemaire, V.. Ely, D., and Henley, E. D. (1993). Psychosocial stress can induce chronic hypertension in normotensive strains of rats. Hypertension (Dallas) 21, 714-723. Henry, J. P., Liu, J., and Meehan. W. P. (1995). Psychosocial stress and experimental hypertension. I n “Hypertension: Physiology, Diagnosis, and Management” (J. H. Laragh and B. M. Brenner, eds.), pp. 905-921. Raven Press, New York. Hessing. M. J. C., Scheepens. C. J. M.. Schouten. W. G. P., Tielen, M. J. M., and Wiepkema, P. R. (1994). Social rank and disease susceptibility in pigs. Vet. Immunol. Immunopathol. 43, 373-387. Hofer, M. A. (1981). Toward a developmental basis for disease predisposition: The effects of early maternal separation on brain, behavior, and cardiovascular system. I n “Brain. Behavior, and Bodily Disease” (H. Weiner. M. A. Hofer. and A. J . Stunkard, eds.). pp. 209-228. Raven Press, New York. Hofer. M. A. (1994). Early relationships as regulators of infant physiology and behavior. Acta Paediatr., Suppl. 397, 9-18. Hoffman-Goetz, L., Simpson, J. R.. and Arumugam, Y. (1991). Impact of changes in housing condition on mouse natural killer cell activity. Physiol. Behav. 49, 657-660.

STRESS A N D ITS RELEVANCE FOR ANIMAL BEHAVIOR

119

Hoffmeyer, I. (1982). Responses of female bank voles (Clethrionomys glareolus) to dominant vs subordinate conspecific males and to urine odors from dominant vs subordinate males. Behav. Neural Biol. 36, 178-188. House, J. S., Robbins, C.. and Metzner, H. L. (1982). The association of social relationships and activities with mortality: Prospective evidence from the Tecumseh community health study. A m . J . Epidemiol. 116, 123-140. House, J. S., Landis. K. R., and Umberson, D. (1988). Social relationships and health. Science 241,540-544. Hrdina, P. D., and Henry, M. E. (1981). Experimental models of mental illness: Separationinduced depression in primates. In “Neuroendocrine Regulation and Altered Behavior” (P. D. Hrdina and R. L. Singhal, eds.), pp. 263-275. Croom Helm, London. Huck, W. W., and Banks, E. M. (1982). Male dominance status, female choice and mating success in the brown lemming, Lemmus trimucronatus. Anim. Behav. 30,665-675. Hucklebridge, F. H., Gamal-el-Din, L., and Brain, P. (1981). Social status and the adrenal medulla in the house mouse ( M u s musculus, L.) Behav. Neural Biol. 33, 345-363. Huhtaniemi, 1. (1994). Fetal testis: A very special endocrine organ. Eur. J. Endocrinol. 130, 25-31. Human, K. L., Moore, T. O., Mougey. E. H., and Meyerhoff, J. L. (1992). Hormonal responses to fighting in hamsters: Separation of physical and psychological causes. Physiol. Behav. 51, 1083-1086. Huntingford, F. A., and Turner, A. K. (1987). “Animal Conflict.” Chapman & Hall, London. Hutzelmeyer, H. (1987). Sympathicus-Nebennierenmarkaktivitatmannlicher Tupaia belangeri: Ein Mass zur Charakterisierung von Individuen mit unterschiedlichem Verhalten. Unpublished Doctoral Dissertation, University of Bayreuth, Bayreuth, Germany. Insel. T. R.. and Hulihan. T. J. (1995). A gender-specific mechanism for pair bonding: Oxytocin and partner preference formation in monogamous voles. Behav. Neurosci. 109,782-789. Insel, T. R., Preston, S.. and Winslow. J. T. (1995). Mating in the monogamous male: Behavioral consequences. Physiol. Behav. 57, 615-627. Johnson, E. O., Kamilaris, T. C., Carter, C. S., Calogero, A. E., Gold, P. W., and Chrousos. G. P. (1996). The biobehavioral consequences of psychogenic stress in a small social primate (Callithrix jacchus jacchus). Biol. Psychiatry 40,317-337. Kaiser, C. (1996). Trennung physischer und psychischer Anteile sozialer Stressreaktionen bei mannlichen Tupaia belangeri. Unpublished Doctoral Dissertation, University of Bayreuth. Bayreuth, Germany. Kannel, W. B., Kannel, C., Pfaffenbarger. R. S., and Cupples, L. A. (1987). Heart rate and cardiovascular mortality. The Framingham Study. A m . Heart J. 116, 1369-1373. Kaplan, J. R., Manuck, S. B.. Clarkson, T. B., Lusso, F. M., and Taub, D. M. (1982). Social status, environment, and atherosclerosis in cynomolgus monkeys. Arteriosclerosis (Dallas) 2,359-368. Kaplan. J. R., Heise, E. R., Manuck, S. B., Shively, C. A,, Cohen, S., Rabin, B. S.. and Kasprowicz, A. L. (1991). The relationship of agonistic and affiliative behavior patterns to cellular immune function among cynomolgus monkeys Macaca fascicularis living in unstable social groups. Am. J. Primafol. 25, 157-174. Kass, E. M., Hechter, 0..Macci, I. A,, and Mou, T. W. (1954). Changes in patterns of secretion in rabbits after prolonged treatment with ACTH. Proc. Soc. Exp. Biol.Med. 85,583-587. Kavaliers. M., and Colwell, D. D. (1995). Odours of parasized males induce aversive response in female mice. Anim. Behav. 50, 1161-1169. Keane, B., Waser, P. M., Creel, S. R.. Elliott, L. F., and Minchella, D. J. (1994). Subordinate reproduction in dwarf mongooses. Anim. Behav. 4 7 , 6 5 7 5 .

120

DIETRICH VON HOLST

Keddy, A. C. (1986). Female mate choice in vervet monkeys (Cercopithecus aefhivps sabaeus). Am. J. Primatol. 10, 125-134. Keller. S. E., Schleifer, S. J . , and Demetrikopoulos, M. K. (1991). Stress-induced changes in immune function in animals: hypothalamo-pituitary-adrenal influences. In “Psychoneuroimmunology” (R. Ader, D. L. Felten. and N. Cohen, eds.). pp. 771-787. Academic Press, San Diego, CA. Kelley, K. W. (1991 ). Growth hormone in immunobiology. I n “Psychoneuroimmunology” (R. Ader, D. L. Felten, and N. Cohen. eds.), pp. 377-402. Academic Press, San Diego. CA. Kelley, K. W., Johnson, R. W., and Dantzer. R. (1994). Immunology discovers physiology. Vet. Immunol. Immunopathol. 43, 157-165. Kendrick, K. M., Keverne, E. B., and Baldwin. B. A. (1987). Intracerebroventricular oxytocin stimulates maternal behavior in the sheep. Neuroendocrinology 46, 56-61. Keverne, E. B. (1988). Central mechanisms underlying the neural and neuroendocrine determinants in maternal behavior. Psychoneurvendocrinology 13, 127-141. Keverne, E. G., Meller, R. E.. and Eberhart, J . A. (1982). Dominance and subordination: Concepts and physiological states. I n “Advanced Views in Primate Biology” (A. B. Chiarelli and R. S. Corrucini, eds.), pp. 81-94. Springer, New York. Kime, D. E., Vinson. G. P., Major, P. W., and Kilpatrick, R. (1980). Adrenal-gonad relationships. In “General. Comparative and Clinical Endocrinology of the Adrenal Cortex” (I. Chester Jones and 1. W. Henderson, eds.), 183-264. Academic Press, London. Kirschbaum, C., and Hellhammer, D. H. (1989). Salivary cortisol in psychobiological research: An overview. Neurnpsychvbiology 22, 150-169. Klein, F., Lernaire, V., Sandi. C., Vitiello, S.. Van der Logt. J., Laurent, P. E., Neveu, P., Le Moal, M., and Mormede, P. (1992). Prolonged increase of corticosterone secretion by chronic social stress does not necessarily impair immune function. Life Sci. 50,723-731. Klopfer, P. H. (1971). Mother love: What turns it on? Am. Sci. 59,404-407. Koolhaas, J., Schuurman. T., and Wiepkema, P. R. (1980). The organization of intraspecific agonistic behaviour in the rat. Prvg. Neurobiol. 15, 247-268. Kopin, I. J . (1980). Catecholamines, adrenal hormones and stress. I n “Neuroendocrinology” (D. T. Krieger and J. C. Hughes, eds.), pp. 159-166. Sinauer, Sunderland, MA. Kopin. 1. J., Eisenhofer, G., and Goldstein, D. (1988). Sympathoadrenal medullary system and stress. In “Mechanisms of physical and emotional stress: Advances in experimental medicine and biology, Volume 245” (G. P. Chrousos, L. D. Loriaux, and P. W. Gold, eds.), pp. 11-23. Plenum Press, New York. Krebs, C. J . (1964). The lemming cycle at Baker Lake, Northwest Territories, during 1959-1962. Tech. Pap.-Arct. Inst. North Am. 15. Krehs, C. J. (1978). “Ecology-The Experimental Analysis of Distribution and Abundance.” Harper & Row, New York. Krebs, C. J . (1996). Population cycles revisited. J. Mammal. 77, 8-24. Krebs. C. J., and Myers, J . H. (1974). Population cycles in small mammals. Adv. Ecol. Res. 8,267-399. Krum, A. A., and Glenn, R. E. (1965). Adrenal steroid secretion in rabbits following prolonged ACTH administration. Proc. SOC. Exp. Biol. Med. 118, 225-258. Kunkele, J.. and von Holst, D. (1996). Natal dispersal in the European wild rabbit. Anim. Behav. 51, 1047-1059. Kvetnansy. R., Weise, V. K., and Kopin, I. J . (1970). Elevation of adrenal tyrosine hydroxylase and phenylethanolamine-N-methyltransferase by repeated immobilization of rats. Endocrinology (Baltimore) 87, 744-749. Kvetnansy. R.. Weise, V. K.. Thoa. N. B., and Kopin, I. J. (1979). Effects of chronic guanethidine treatment and adrenal medullectomy on plasma levels of catecholamines and corticosterone in forcibly immobilized rats. J. Pharmacol. Exp. Ther. 209, 287-291.

STRESS A N D ITS RELEVANCE FOR ANIMAL BEHAVIOR

121

Labhart, A,, ed. (1986). “Clinical Endocrinology. Theory and Practice.” Springer, Berlin. Lapin, B. A.. and Cherkovich, G. M. (1971). Environmental changes causing the development of neuroses and corticovisceral pathology in monkeys. I n “Society, Stress and Disease” (L. Levi, ed.), Vol. 1, pp. 266-279. Oxford University Press, London. Laudenslager, M. L., and Fleshner. M. (1994). Stress and immunity: Of mice, monkeys, models, and mechanisms. I n “Handbook of Human Stress and Immunity” (R. Glaser and J. K. Kiecolt-Glaser. eds.), pp. 161-181. Academic Press, San Diego, CA. Laudenslager, M. L., Reite. M., and Harbeck, R. J. (1982). Suppressed immune response in infant monkeys associated with maternal separation. Behav. Neural Biol. 36, 40-48. Lazarus, R. S.. and Folkman, S. (1984). Coping and adaptation. I n “Handbook of Behavioral Medicine” (W. D. Gentry, ed.), pp. 282-325. Guilford. New York. Lee, A. K., and McDonald, I. R. (1985). Stress and population regulation in small mammals. Oxford Rev. Reprod. Biol. 7 , 261-304. Lemaire, V., Le Moal, M., and Mormkde. P. (1993). Regulation of catecholamine-synthesizing enzymes in adrenals of Wistar rats under chronic stress. Am. J. Physiol. 264, R957-R962. Leshner, A. I. (1981). The role of hormones in the control of submissiveness. I n “Multidisciplinary Approaches to Aggression Research” (P. F. Brain and D. Benton, eds.), pp. 309-322. Elsevier, Amsterdam. Leshner. A. I., and Candland, D. K. (1972). Endocrine effects of grouping and dominance rank in squirrel monkeys. Physiol. Behav. 8, 441-446. Leshner. A. I., and Politch. J. A. (1979). Hormonal control of submissiveness in mice: Irrelevance of the androgens and relevance of the pituitary-adrenal hormones. Physiol. Behav. 22,53 1-534. Leshner, A. I.. Korn, S. J., Mixon, J. F., Rosenthal, C., and Besser, A. K. (1980). Effects of corticosterone on submissiveness in mice: some temporal and theoretical considerations. Physiol. Behav. 24,283-288. Levi. L.. ed. (1971). “Society. Stress and Disease,” Vol. 1. Oxford University Press, London. Levine, S. (1993a). The psychoendocrinology of stress. Ann. N.Y. Acad. Sci. 697, 61-69. Levine. S. (1993b). The influence of social factors on the response to stress. Psychofher. Psychosom. 60, 33-38. Levine, S., and Ursin, H. (1991).What is stress? I n “Stress,Neurobiology and Neuroendocrinology” (M. R. Brown. G. F. Koob, and C. Rivier, eds.), pp. 3-21. Dekker, New York. Levine, S., Weinberg, J., and Brett, L. P. (1979). Inhibition of pituitary-adrenal activity as a consequence of consummatory behavior. Psychonellroendocrinology 4,275-286. Levine, S., Johnson, D. F., and Gonzales, C. A. (1985). Behavioral and hormonal responses to separation in infant rhesus monkeys and mothers. Behav. Neurosci. 99, 399-410. Levine, S., Coe, C. L., and Wiener, S. G. (1989). Psychoneuroendocrinology of stress: A psychobiological perspective. In “Psychoendocrinology” (F. R. Brush and S. Levine, eds.), 341-378. Academic Press, San Diego, CA. Lilly, M. P., Engeland, W. C., and Gann, D. S. (1983). Responses of cortisol secretion to repeated hemorrhage in the anesthetized dog. Endocrinology (Balfimore)112, 681-688. Line, S. W., Kaplan, J. R., Heise, E. R., Hilliard. J. K., Cohen, S.. Rabin, B. S., and Manuck, S. B. (1996). Effects of social reorganization on cellular immunity in male cynomolgus monkeys. Am. J. Primatol. 39,235-249. Lloyd, J. A. (1973). Frequency of activity and endocrine response among male house mice (Mus musculus) in freely growing populations. Proc. SOC. Exp. Biol. Med. 142,784-786. Lockley, R. M. (1961). Social structure and stress in the rabbit warren. J. h i m . Ecol. 30, 385-423. Louch, C. D.. and Higginbotham, M. (1967). The relation between social rank and plasma corticosterone levels in mice. Gen. Comp. Endocrinol. 8, 441-444.

122

DIETRICH VON HOLST

Lubach, G. R., Coe. C. L., Karaszewski, J. W., and Ershler, W. B. (1996). Effector and target cells in the assessment of natural cytotoxic activity of rhesus monkeys. Am. J. Primatol. 39,275-287. Lundberg, U., and Frankenhaeuser, M. (1980). Pituitary-adrenal and sympathetic-adrenal correlates of distress and effort. J. Psychosom. Res. 24, 125-130. Madden, K. S., and Felten, D. L. (1995). Experimental basis for neural-immune interactions. Physiol. Rev. 75, 77-106. Madden, K. S., and Livnat, S. (1991). Catecholamine action and immunological reactivity. In “Psychoneuroimmunology” (R. Ader, D. L. Felten, and N. Cohen, eds.), pp. 283-310. Academic Press, San Diego, CA. Magarinos, A. M., McEwen, B. S., Flugge, G., and Fuchs, E. (1996). Chronic psychosocial stress causes apical dendrite atrophy of hippocampal CA3 pyramidal neurons in subordinate tree shrews. J . Neitrosci. 16, 3534-3540. Mallick, J., Stoddart, D. M., Jones, 1.. and Bradley, A. J. (1994). Behavioral and endocrinological correlates of social status in the male sugar glider (Petaurus breviceps Marsupialia: Petauridae). Physiol. Behav. 55, 1131-1134. Mann, D. R.. and Orr. T. E. (1990). Effect of restraint stress on gonadal proopiomelanocortin peptides and the pituitary-testicular axis in rats. Life Sci. 46, 1601-1609. Manogue, K. R., Leshner, A. I., and Candland, K. (1975). Dominance status and adrenocortical reactivity to stress in squirrel monkeys (Saimiri sciureus). Primates 16, 457-463. Manuck, S. B., Kaplan. J. R., and Clarkson. T. B. (1983). Social instability and coronary artery atherosclerosis in cynomolgus monkeys. Neurosci. Biobehav. Rev. 7,485-491. Manuck, S . B., Kaplan, J. R., and Matthews, K. A. (1986). Behavioral antecedents of coronary heart disease and atherosclerosis. Arteriosclerosis (Dallas) 6, 2-14. Manuck, S. B., Kaplan, J. R., Adams, M. R.. and Clarkson, T. B. (1989). Behaviorally elicited heart rate reactivity and atherosclerosis in female cynomolgus monkeys (Macaca fascicularis). Psychosom. Med. 51, 306-318. Manuck, S. B., Marsland, A. L., Kaplan. J. R., and Williams, J. K. (1995). The pathogenicity of behavior and its neuroendocrine mediation: An example from coronary heart disease. Psychosom. Med. 57, 275-283. Maric, D., Kostic, T., and Kovacevic, R. (1996). Effects of acute and chronic immobilization stress on rat Leydig cell steroidogenesis. J. Steroid Biochem. Mol. Biol. 58, 351-355. Marsden, H. M., and Holler, N. R. (1964). Social behavior in confined populations of the cottontail and the swamp rabbit. Wildl. Monogr. 13, 1-39. Masataka, N., Ishida, T., Suzuki, J., Matsumura, S., Udono, S.. and Sasaoka, S. (1990). Dominance and immunity in chimpanzees (Pan troglodytes). Ethology 85, 147-155. Maslow, A. H. (1936). The role of dominance in the social and sexual behavior of infrahuman primates: 111. A theory of sexual behavior of infra-human primates. J . Genet. Psychol. 48, 310-338. Mason, J. W. (1968a). Organization of psychoendocrine mechanisms. Psychosom. Med. 30, 565-808. Mason, J. W. (1968b). A review of psychoendocrine research on the pituitary-adrenal cortical system. Psychosom. Med. 30, 576-607. Mason, J. W. (1968~).“Over-all’’ hormonal balance as a key to endocrine organization. Psychosom. Med. 30, 791-808. Mason, J. W., and Brady, J. V. (1956). Plasma 17-hydroxycorticosteroidchanges related to reserpine effects on emotional behavior. Science 124, 983-984. Mason, J. W., Kenion, C. C., Collins, D. R., Mougey, E. H., Jones. J. A,, Driver, G. C., Brady. J . V., and Beer, B. (1968a). Urinary testosterone response to 72-hr. avoidance sessions in the monkey. Psychosom. Med. 30, 721 -732.

STRESS A N D ITS RELEVANCE FOR ANIMAL BEHAVIOR

123

Mason, J. W., Jones, J. A,, Ricketts, P. T., Brady, J. V.. and Tolliver, G. A. (1968b). Urinary aldosterone and urine volume responses to 72-hr. avoidance sessions in the monkey. Psychosom. Med. 30, 733-745. Matochik. J. A,, and Barfield. R. J. (1991). Hormonal control of precopulatory sebaceous scent marking and ultrasonic mating vocaliztion in male rats. Horm. Behav. 25,445-460. Maynard Smith, J., and Price, G. R. (1973). The logic of animal conflict. Nature (London) 246,lS-18. Mazur, A,, and Lamb, T. A. (1980). Testosterone. status and mood in human males. Horm. Behav. 14,236-246. McCarthy. M. M. (1994). Molecular aspects of sexual differentiation of the rodent brain. Physiol. Brhav. 19,415-427. McCruden, A. B., and Stimson. W. H. (1991). Sex hormones and immune function. I n “Psychoneuroimmunology” (R. Ader, D. L. Felten, and N. Cohen. eds.). pp. 475-493. Academic Press, San Diego, CA. McDonald, I. R., Lee, A. K., Bradley, A. J.. and Than, K. A. (1981). Endocrine changes in dasyurid marsupials with differing mortality patterns. Gen. Comp. Endocrinol. 44, 292-301. McDonald, I. R., Lee, A. K., Than, K. A., and Martin, R. W. (1986). Failure of glucocorticoid feedback in males of a population of small marsupials (Antechinus swainsonii) during period of mating. J . Endocrinol. 108, 63-68. McEwen, B. S., Albeck, D., Cameron, H., Chao, H. M., Could. E., Hastings, N., Kuroda, Y., Luine, V., Magarinos, A. M., McKittrick, C. R., Orchinik, M.. Pavlides, C., Vaher. P., Watanabe, Y., and Weiland, N. (1995). Stress and the brain: A paradoxical role for adrenal steroids. Vitam. Horm. ( N .Y . ) 14, 371-402. McGaugh, J. L.. and Gold, P. E. (1989). Hormonal modulation of memory. In “Psychoneuroendocrinology” (F. R. Brush and S. Levine, eds.). pp. 305-340. Academic Press, San Diego. CA. McGrady. A. V. (1984). Effects of psychological stress on male reproduction: A review. Arch. Androl. 13, 1-7. McGuire. M. T., Brammer, G. L.. and Raleigh, M. J . (1986). Resting cortisol levels and the emergence of dominance status among male vervet monkeys. Horm. Behav. 10,285-292. Mendoza, S. P., and Mason, W. A. (1986). Contrasting responses to intruders and to involuntary separation by monoganous and polygynous New World monkeys. Physiol. Behav. 38, 795-801. Mendoza, S. P.. Coe, C. L., Lowe, C. L., and Levine, S. (1979). The physiological response to group formation in adult male squirrel monkeys. Psychoneuroendocrinology 3,221-229. Miczek, K. A,, Thompson, M. L., and Tornatzky, W. (1991). Subordinate animals. Behavioral and physiological adaptations and opiod telerance. In “Stress, Neurobiology and Neuroendocrinology” (M. R. Brown, G. F. Koob, and C. Rivier, eds.), pp. 323-357. Dekker, New York. Miller, M. W., Hobbs. N. T., and Sousa, M. C. (1991). Detecting stress responses in Rocky Mountain bighorn sheep (Ovis canadensis canadensis): Reliability of cortisol concentrations in urine and feces. Can. J. Zool. 69, 15-24. Moberg, G. P. (1987). Influence of the adrenal axis upon the gonads. Oxford Rev. Reprod. Biol. 9, 456-496. Monaghan, E. P., and Glickman, S. E. (1992). Hormones and aggressive behavior. In “Behavioral Endocrinology” (J. B. Becker, S. M.. Breedlove, and D. Crews, eds.), pp. 261-286. MIT Press, Cambridge. MA. Monder, C., Sakai, R. R., Miroff, Y..Blanchard, D. C., and Blanchard, R. J. (1994). Reciprocal changes in plasma corticosterone and testosterone in stressed male rats maintained in a

124

DIETRICH VON HOLST

visible burrow system: Evidence for a mediating role of testicular 11 beta-hydroxysteroid dehydrogenase. Endocrinology (Baltimore) 134, 1193-1 198. Monjan, A. A. (1981). Stress and immunological competence: Studies in animals. In “Psychoneuroimmunology” (R. Ader, ed.). pp. 185-228. Academic Press, New York. Moore, J., and Gotelli, N. J. (1990). Phylogenetic perspective on the evolution of altered host behaviours: A critical look at the manipulation hypothesis. In “Parasitism and Host Behaviour” (C. J. Barnard and J. M. Behnke, eds.), pp. 193-233. Taylor & Francis, London. Mormtde, P., Lemaire, V., Castanon, N., Dulluc. J., Laval, M., and Le Moal, M. (1990). Multiple neuroendocrine responses to chronic social stress: Interaction between individual and characteristic situational factors. Physiol. Behuv. 47, 1099-1 105. Moynihan, J. A., Brenner, G. J., Cocke, R., Karp, J. D., Breneman, S. E.. Dopp, J. M., Ader, R., Cohen, N., Grota, L. J., and Felten, S. Y. (1994). Stress-induced modulation of immune function in mice. In “Handbook of Human Stress and Immunity” (R. Glaser and J. K. Kiecolt-Glaser, eds.), pp. 1-22. Academic Press, San Diego, CA. Munck, A,. and Guyre, P. M. (1991). Glucocorticois and immune function. In “Psychoneuroimmunology” (R. Ader. D. L. Felten, and N. Cohen, eds.). pp. 429-474. Academic Press, San Diego, CA. Munck, A., Guyre, P. M., and Holbrook, N. J. (1984). Physiological functions of glucocorticoids in stress and their relation to pharmacological actions. Endocr. Rev. 5, 25-45. Myers, K., and Poole, W. E. (1959). A study of the biology of the wild rabbit, Orycfolugus cuniculus (L.), in confined populations. CSIRO Wildl. Res. 4, 14-27. Myers, K., and Poole, W. E. (1962). A study of the biology of the wild rabbit, Orycfolagus cuniculus (L.) in confined populations. 111. Reproduction. Aust. J. 2001.10, 225-267. Myers, K., Hale, C. S., Mykytowycz, R.. and Hughes, R. L. (1971). The effects of varying density and space on sociality and health in animals. In “Behavior and Environment. The Use of Space by Animals and Men” (A. H. Esser. ed.). pp. 148-187. Plenum, New York. Myers, M. J., and Murtaugh, M. P. (1995). “Cytokines in Animal Health and Disease.” Dekker, New York. Mykytowycz, R. (1958). Social behaviour of an experimental colony of wild rabbits, Oryclolugus cuniculus (L.). I. Establishment of the colony. CSIRO Wildl. Res. 3, 7-25. Mykytowycz, R. (1959a). Social behaviour of an experimental colony of wild rabbits, Orycfolugus cuniculus (L.). 11. First breeding season. CSIRO Wildl. Res. 4, 1-13. Mykytowycz, R. (1959b). Social behaviour of an experimental colony of wild rabbits, Orycfolugus cuniculus (L.). 111. Second breeding season. CSIRO Wildl. Res. 5, 1-20. Mykytowycz, R. (1961). Social behaviour of an experimental colony of wild rabbits, Orycfolugus cuniculus (L.). IV. Conclusions: Outbreak of myxomatosis, third breeding season, and starvation. CSIRO Wildl. Res. 6 , 142-155. Nelson, R. J., and Demas, G. E. (1996). Seasonal changes in immune function. Q. Rev. Biol. 71, 511-548. Nowell, N. W. (1980). Adrenocortical function in relation to mammalian population densities and hierarchies. In “General. Comparative and Clinical Endocrinology of the Adrenal Cortex” (I. Chester Jones and I. W. Henderson, eds.), Vol. 3, pp. 349-393. Academic Press, London. O’Grady. M. P., and Hall, N. R. (1991). Long-term effects of neuroendocrine-immune interactions during early development. In “Psychoneuroimmunology” (R. Ader, D. L. Felten. and N. Cohen, eds.), pp. 561-572. Academic Press, San Diego, CA. Olsen, N. J., and Kovacs, W. J. (1996). Gonadal steroids and immunity. Endocr. Rev. 17, 369-384.

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

125

Orr, T. E., and Mann, D. R. (1992). Role of glucocorticoids in the stress-induced suppression of testicular steroidogenesis in adult male rats. Horm. Behav. 26, 350-363. Ottenweller, J. E., Tapp, W. N., Burke, J. M.. and Natelson, B. H. (1985). Plasma cortisol and corticosterone concentrations in the golden hamster (Mesocricefusaurafus).Life Sci. 37, 1551-1558. Paavonen, T. (1994). Hormonal regulation of immune responses. Annals of Medicine 26, 255-258. Packer, C., Collins, D. A,, Sindimwo, A,, and Goodall, J. (1995). Reproductive constraints on aggressive competition in female baboons. Nature (London) 373, 60-63. Pedersen, C. A,, and Prange, A. J., Jr. (1985). Oxytocin and mothering behavior in the rat. Pharmacol. Ther. 28, 287-302. Perrson, L., Gullberg, B., Hanson, B. S., Moestrup, T., and Ostergren, P. 0. (1994). HIV infection: social network, social support, and CD4 lymphocyte values in infected homosexual men in Malmo, Sweden. J. Epidemiol. Cornmunify Health 48, 580-585. Phoenix, C. H., Goy, R. W., Gerall, A. A,, and Young, W. C. (1959). Organizing action of prenatally administered testosterone propionate on the tissues mediating mating behavior in the female guinea pig. Endocrinology (Baltimore) 65, 369-382. Plaut, M., and Friedman. S. B. (1981). Psychosocial factors in infectious disease. In “Psychoneuroimmunology” (R. Ader, ed.), pp. 3-30. Academic Press, New York. Plaut, S. M., Ader, R., Friedman, S. B., and Ritterson, A. L. (1969). Social factors and resistance to malaria in the mouse: Effects of group vs. individual housing on resistance to Plasmodium berghei infection. Psychosom. Med. 31, 536-552. Politch, J. A., and Leshner, A. I. (1977). Relationship between plasma corticosterone levels and levels of aggressiveness in mice. Physiol. Behav. 19, 775-780. Popova, N. K., and Naumenko, E. V. (1972). Dominance relations and the pituitary-adrenal system rats. Anim. Behav. 20, 108-111. Price, V.A. (1982). “Type A Behavior Pattern. A Model for Research and Practice.” Academic Press, New York. Raab, A,, Dantzer, R., Mormede, M. P.. Taghzouti, K., Simon, H.,and Le Moal. M. (1986). Behavioural, physiological and immunological consequences of social status and aggression in chronically coexistingresident-intruderdyads of male rats. Physiol. Behav. 36, 223-228. Rabin. B. S.. Kusnecov, A,, Shurin, M., Zhou, D., and Rasnick, S. (1994). Mechanistic aspects of stressor-induced immune alteration. I n “Handbook of Human Stress and Immunity” (R. Glaser and J. K. Kiecolt-Glaser, eds.), pp. 23-51. Academic Press, San Diego, CA. Rabin, D., Gold, P. W.. Margioris, A. N., and Chrousos, G. P. (1988).Stress and reproduction: Physiologic and pathophysiologic interactions between the stress and the reproductive axes. Adv. Exp. Med. Biol. 245,377-387. Ratcliffe. H. L., Luginbuhl, H., Schnarr, W. R., and Chacko, K. (1969). Coronary arteriosclerosis in swine: Evidence of a relation to behavior. J. Comp. Physiol. Psychol. 68,385-392. Rau, M. E. (1983). Establishment and maintenance of behavioural dominance in male mice infected with Trichinella spiralis. Parasitology 86, 311-318. Read, A. F. (1990). Parasites and the evolution of host sexual behaviour. I n “Parasitism and Host Behaviour” (C. J. Barnard and J. M. Benke, eds.), pp. 117-157. Taylor & Francis, London. Reite, M. L., Harbeck, R. J., and Hoffman, A. (1981). Altered cellular immune response following peer separation. Life Sci. 29, 1133-1136. Riad-Fahmy, D.. Read, G. F., Walker, R. F., and Griffiths, K. (1982). Steroids in saliva for assessing endocrine function. Endocr. Rev. 3, 367-395. Riechert, S. E. (1988). The energetic costs of fighting. Am. 2001.28,877-884.

126

DIETRICH VON HOLST

Riley, V. (1981). Psychoneuroendocrine influences on immunocompetence and neoplasia. Science 212, 1100-1109. Rivier, C. (1991). Neuroendocrine mechanisms of anterior pituitary regulation in the rat exposed to stress. In “Stress, Neurobiology and Neuroendocrinology” (M. R. Brown. G. F. Koob. and C. Rivier, eds.), pp. 119-136. Dekker, New York. Rose, R. (1985). Psychoendorcinology. In “Williams Textbook of Endocrinology” (J. Wilson and D. Foster, eds.), pp. 653-681. Saunders, Philadelphia. Rose, R., Holaday, J., and Bernstein, I. (1971). Plasma testosterone, dominance rank and aggressive behaviour in male rhesus monkeys. Nature (London) 231, 366-368. Rose, R. M.. Gordon. T. P., and Bernstein. 1. S. (1972). Plasma testosterone levels in the male rhesus: Influences of sexual and social stimuli. Science 178, 643-645. Rose, R. M., Bernstein, 1. S., Gordon. T. P., and Catlin. S. F. (1974). Androgen and aggression: A review and recent findings in primates. In “Primate Aggression, Territoriality and Xenophobia” (R. L. Halloway. ed.). pp. 275-304. Academic Press, New York. Rose. R. M., Bernstein. I. S., and Gordon, T. P. (1975). Consequences of social conflict on plasma testosterone levels in rhesus monkeys. Psychosom. Med. 37, 50-61. Rosenblum. L. A., and Plimpton, E. H. (1981). The infant’s effort to cope with separation. In “The Uncommon Child” (M. Lewis and L. A. Rosenblum, eds.), pp. 225-257. Plenum. New York. Rosenfeld, P., Wetmore, J. B., and Levine, S. (1992). Effects of repeated maternal separations on the adrenocortical response to stress of preweanling rats. Physiul. Behav. 52,787-791. Rosenman, R. H. (1986). Current and past history of Type A behavior pattern. In “Biological and Psychological Factors in Cardiovascular Disease” (T. H. Schmidt, T. M. Dembroski, and G. Blumchen. eds.), pp. 15-40. Springer, Berlin. Saad, M. B., and Bayle, J. D. (1985). Seasonal changes in plasma testosterone, thyroxine, and cortisol levels in wild rabbits (Oryctolagus citniculus algirus) of Zembra island. Gen. Comp. Endocrinol. 57,383-388. Saboureau. M . , Laurent, G., and Boissin, J. (1977). Daily and seasonal rhythms of locomotory activity and adrenal function in male hedghog (Erinaceus europaeus L.). J. Inrerdiscip. Cycle Res. 10, 245-266. Sachser, N. (1986). Different forms of social organization at high and low population densities in guinea pigs. Behaviour 97, 253-272. Sachser, N. (1987). Short-term responses of plasma norepinephrine, epinephrine. glucocorticoid and testosterone titers to social and non-social stressors in male guinea pigs of different social status. Physiol. Behav. 39, 11-20. Sachser, N. (1993). The ability to arrange with conspecifics depends on social experiences around puberty. Physiol. Behav. 53,539-544. Sachser, N. (1994a). “Sozialphysiologische Untersuchungen an Hausmeerschweinchen. Gruppenstrukturen, soziale Situation und Endokrinium, Wohlergehen.” Parey, Berlin. Sachser, N. (1994b). Social dominance and health in nonhuman mammals: A case study in guinea pigs. In “Social Stratification and Socioeconomic Inequality” (L. Ellis. ed.). Vol. 2, pp. 113-121. Praeger. Westport, CT. Sachser, N., and Kaiser, S. (1996). Prenatal social stress masculinizes the females’ behaviour in guinea pigs. Physiol. Behav. 60, 589-594. Sachser, N., and Lick, C. (1989). Social stress in guinea pigs. Physiol. Behav. 46, 137-144. Sachser. N., and Lick, C. (1991). Social experience, behavior, and stress in guinea pigs. Physiol. Behav. 50, 83-90. Sachser, N., and Prove, E. (12186). Social status and plasma-testosterone titers in male guinea pigs (Cavia aperea f: porcellits). Ethology 71, 103-1 14.

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

127

Sachser, N., and Renninger, S. V. (1993). Coping with social conflict: The role of social rearing in guinea pigs. Ethol. EcoL Evol. 5, 65-74. Sachser, N., Lick, C., and Stanzel, K. (1994). The environment, hormones, and aggressive behaviour: A 5-year-study in guinea pigs. Psychoneuroendocrinology 19, 697-707. Sachser, N., Diirschlag. M., and Hirzel, D. (1998). Social relationships and the management of stress. Psychoneuroenr~ocrinology(in press). Sade, D. S., Cushing, K., Cushing, P., Dunaif, A., Figueroa, J., Kaplan, R., Laver, C., Rhodes. D., and Schneider. J. (1976). Population dynamics in relation to social structure on Cay0 Santiago. Yearb. Physiol. Anthropol. 20, 253-262. Sapolsky, R. M. (1982). The endocrine stress-response and social status in the wild baboon. Horm. Behav. 20,279-292. Sapolsky. R.M. (1983). Individual differences in cortisol secretory patterns in the wild baboon: Role of negative feedback sensitivity. Endocrinology (Baltimore) 113, 2263-2267. Sapolsky, R. M. (1985a). Endocrine aspects of social instability in the olive baboon (Papio anubis). A m . J. Primntol. 5, 365-379. Sapolsky, R. M. (198Sb). Stress-induced suppression of testicular function in the wild baboon: Role of glucocorticoids. Endocrinology (Baltimore) 116, 2273-2278. Sapolsky, R. M. (1988). Individual differences and the stress response: Studies of a wild primate. A d v . Exp. Med. B i d . 245, 399-411. Sapolsky, R. M. (1990). Adrenocortical function, social rank, and personality among wild baboons. Biol. Psychiatry 28, 862-878. Sapolsky, R. M. (1991). Effects of stress and glucocorticoids on hippocampal neuronal survival. I n “Stress, Neurobiology and Neuroendocrinology” (M. R. Brown, G. F. Koob, and C. Rivier. eds.), pp. 293-322. Dekker, New York. Sapolsky, R. M. (1992). “Stress, the Aging Brain, and the Mechanisms of Neuron Death.” MIT Press, Cambridge, MA. Sapolsky, R. M., and Mott, G. E. (1987). Social subordination in wild baboons is associated with suppressed high density lipoprotein-cholesterol concentrations: The possible role of chronic stress. Endocrinology (Baltimore) 121, 1605-1610. Sassenrath, E. N. (1970). Increased adrenal responsiveness related to social stress in rhesus monkeys. Horm. Behav. 1,283-298. Schaftenaar, W., Buiter, R. M., and Dieleman, S. J., eds. 1992. Proceedings from the First International Symposium on Faecal Steroid Monitoring in Zoo Animals. 1992 Feb 28-29, Rotterdam, The Netherlands. Schiml, P. A,, Mendoza, S. P., Saltzman, W., Lyons, D. M., and Mason, W. A. (1996). Seasonality in squirrel monkeys (Saimiri sciurew): Social facilitation by females. Physiol. Behav. 60, 1105-1113. Schoenbach, V. C., Kaplan, B. H.. Fredman, L.. and Kleinbaum, D. G . (1986). Social ties and mortality in Evans County, Georgia. Am. J. Epidemiol. 123, 577-591. Schonheiter, R. (1992). Der Einfluss von Jahreszeit, Geschlecht und sozialen Faktoren auf Verhalten und physiologische Parameter bei adulten Europaischen Wildkaninchen (Oryctolagus cuniculus L.). Unpublished Doctoral Dissertation, University of Bayreuth, Bayreuth, Germany. Schuhr, B. (1987). Social structure and plasma corticosterone level in female albino ice. Physiol. Behav. 40,689-693. Schuurman, T. (1981). Endocrine processes underlying victory and defeat in the male rat. Unpublished Doctoral Dissertation, Rijksuniversiteit, Groningen, Netherland. Schuurs, A. H. W. M.. and Verheul, H. A. M. (1990). Effects of gender and sex steroids on the immune response. J. Steroid. Biochem. 35, 157-172.

128

DIETRICH VON HOLS?

Scott, J. P., and Fredericson. E. (1951). The causes of fighting in mice and rats. Physiol. Zool. 24,273-309. Seligman, M. E. (1975). “Helplessness: O n Depression, Development, and Death.” Freeman, San Francisco. Selye, H. (1936). A syndrome produced by diverse nocuous agents. Nature (London) 138, 32-34. Selye. H. (1950). “Stress.” Acta, Montreal. Selye, H. (1952). “The Story of the Adaptation Syndrome.” Acta, Montreal. Selye, H. (1976). “Stress in Health and Disease.” Butterworth, Boston. Selye, H. (1981). The stress concept today. In “Handbook on Stress and Anxiety. Contemporary Knowledge, Theory, and Treatment” ( I . L. Kutash and L. B. Schlesinger. eds.). pp. 127-143. Jossey-Bass, San Francisco. Shapiro, L. E., and Dewsbury, D. A. (1986). Male dominance, female choice and male copulatory behavior in two species of voles (Microtus ochrogasrer and Microtus montanus). Behav. Ecol. Sociohiol. 18, 267-274. Sherman, P. W., Jarvis, J. U. M., and Alexander, R. A,, eds. (1991). “The Biology of the Naked Mole-rat.” Princeton University Press, Princeton. NJ. Shively, C.. and Kaplan. J. (1984). Effects of social factors on adrenal weight and related physiology of Macaca fascicularis. Physiol. Behav. 33, 777-782. Silk, J. B., Clark-Wheatley, C. B., Rodman, P. S.. and Samuels. A. (1981). Differential reproductive success and facultative adjustment of sex ratios among captive female bonnet macaques (Macaca radiata). Anim. Behav. 29, 106-120. Smelik, P. G., and Vermes, I. (1980). The regulation of the pituitary-adrenal system in mammals. I n “General, Comparative and Clinical Endocrinology of the Adrenal Cortex” ( I . Chester Jones and 1. W. Henderson, eds.), pp. 1-55. Academic Press, London. Snyder, R. L. (1968). Reproduction and population pressures. Prog. Physiol. Psychol. 2, 119-160. Solomon, G. F., and Amkraut, A. A. (1981). Psychoneuroendocrinological effects on the immune system. Annu. Rev. Microhiol. 35, 155-184. Southern. H. N. (1940). The ecology and population dynamics of the wild rabbit (Oryctolagus cuniculus). Ann. Appl. B i d . 27, 509-526. Stalker, A,, Hermo, L., and Antakly, T. (1989). Covalent affinity labeling, radioautography, and immunocytochemistry localize the glucocorticoid receptors in rat testicular Leydig cells. Am. J. Anar. 186, 36Y-377. Stefanski, V., and Ben-Eliyahu, S. (1996). Social confrontation and tumor metastasis in rats: Defeat and R-adrenergic mechanisms. Physiol. Behav. 60, 277-282. Stafanski, V., and Hendrichs, H. (1996). Social confrontation in male guinea pigs: Behavior, experience, and complement activity. Physiol. Behav. 60, 235-241. Stefanski, V., Solomon, G. F.. Kling, A. S . , Thomas, J., and Plaeger, S. (1996). Impact of social confrontation on rat CD4 T cells bearing different CD45R isoforms. Brain, Behavior, and Immunity 10, 364-379. Sternberg, E. M. (1988). Monokines, lymphokines, and the brain. In “The Year in Immunology 1988” (J. M. Cruse and R. E. Lewis, eds.). Vol. 5, pp. 205-217. Karger, Basel. Stoddard, S. L. (1991). Hypothalamic control and peripheral concomitants of the autonomic defense response. In “Stress, Neurobiology and Neuroendocrinology” (M. R. Brown, G. F. Koob, and C. Rivier, eds.), pp. 231-253. Dekker, New York. Stohr, W. (1986). Heart rate in tree shrews and its persistent modification by social contact. In “Biological and Psychological Factors in Cardiovascular Disease’’ (T. H. Schmidt, T. M. Dembroski, and G. Bliimchen, eds.), pp. 508-516. Springer, Berlin.

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

129

Stohr, W. (1988). Longterm heart rate telemetry in small mammals: A comprehensive approach as a prerequisite for valid results. Physiol. Behav. 43, 567-576. Stumpfe, K. (1973). “Der psychogene Tod.” Hippokrates, Stuttgart. Suomi, S. J. (1976). Factors affecting responses to social separation in rhesus monkeys. In “Animal Models in Human Psychobiology” ( G . Serban and A. Kling, eds.), pp. 9-26. Plenum, New York. Takahashi, L. K. (1990). Hormonal regulation of sociosexual behavior in female mammals. Neurosci. Biohehav. Rev. 14, 403-413. Takahashi, L. K. (1991). Ontogeny of stress-induced ultrasonic vocalization and pituitaryadrenal hormone secretion in preweanling Norway rats. Psychol. Rec. 41, 159-174. Tecoma, E. S., and Huey, L. Y. (1985). Psychic stress and the immune system. Life Sci. 36, 1799-1812. Temoshok, L., and Peeke, H. V. S. (1988). Individual behavior differences related to induced tumor growth in the female Syrian hamster: Two studies. Inr. J. Neurosci. 38, 199-209. Thierry, B.. Steru, L., Chermat, R., and Simon, P. (1984). Searching-waiting strategy: A candidate for an evolutionary model of depression? Behav. Neirral B i d . 41, 180-189. Thiessen. D. D., and Rice, M. (1976). Mammalian scent gland marking and social behavior. Psychol. Bull. 83, 505-539. Thoenen, H., Mueller. R. A,, and Axelrod, J. (1969). Trans-synaptic induction of adrenal tyrosine hydroxylase. J. Pharmacol. Exp. Ther. 169,249-254. Toates, F. (1987). The relevance of models of motivation and learning to animal welfare. In “Biology of Stress in Farm Animals: An Integrative Approach” (P. R. Wiepkema and P. W. M. van Adrichem, eds.), pp. 151-186. Martinus Nijhoff, Dordrecht, The Netherlands. Tobach, E., and Bloch, H. (1955). A study of the relationship between behavior and susceptibility to tuberculosis in rats and mice. Adv. Tuberc. Res. 6, 62-89. Tobach, E., and Bloch, H. (1958). Effects of stress by crowding prior to and following tuberculous infection. A m . J. Physiol. 187, 399-402. Toft, C. A., and Karter, A. J. (1990). Parasite-host coevolution. Trends Ecol. Evol. 5,326-329. Turner, J. W. (1984). Sex differences in the response to neonatal steroid treatment in the mongolian gerbil. Physiol. Behav. 33, 173-176. Unden, A. L., Orth-Gomer, K., and Elofsson, D. (1991). Cardiovascular effects of social support in the working place: Twenty-four hour ECG monitoring of men and women. Psychosom. Med. 53, 50-60. Ungar, A,. and Phillips, J. H. (1983). Regulation of the adrenal medulla. Physiol. Rev. 63, 787-843. Uno, H., Eisele, S., Sakai, A., Shelton, S., Baker, E., DeJesus, O., and Holden, F. (1994). Neurotoxicity of glucocorticoids in the primate brain. Horm. Behav. 28, 336-348. Ursin, H., and Olff. M. (1993). The stress response. In “Stress. From the Synapse to the Syndrome” (S. C. Stanford and P. Salmon, eds.), pp. 3-22. Academic Press, London. Ursin, H., Baade. E., and Levine, S. (1978). “Psychobiology of Stress.” New York: Academic Press, New York. Uvnas-Moberg, K. (1994). Role of efferent and afferent vagal nerve activity during reproduction: Integrating function of oxytocin on metabolism and behavior. Psychoneuroendocrinology 19, 687-695. Vander, A. J., Henry, J. P., Stephens, P. M., Kay, L. L., and Mouw, D. R. (1978). Plasma renin activity in psychosocial hypertension of CBA mice. Circ. Res. 42, 496-502. van Eck, M., Berkhof, H., Nicolson. N.. and Sulon, J. (1996). The effects of perceived stress, traits, mood states, and stressful daily events on salivary cortisol. Psychosom. Med. 58, 447-458.

130

DIETRICH VON HOLST

van Oortmerssen, G. A., Benus. I., and Dijk, D. J. (1985). Studies in the wild house mice: Genotype-environment interactions for attack latency. Neth. J. Zool. 35, 155-169. van Oortmerssen, G. A., Dijk, D. J., and Schuurman, T. (1987). Studies in wild house mice 11. Testosterone and aggression. Horm. Behav. 21, 139-152. Vellucci, S. (1990). Primate social behavior-anxiety or depression. fharmacol. Ther. 47, 167-180. Vessey, S. H. (1964). Effects of grouping on levels of circulating antibody levels in mice. Proc. Soc. Exp. Biol. Med. 115, 252-255. Vining, R. F., and McGinley, R. A. (1986). Hormones in the saliva. Crit. Rev. Clin. Lab. Sci. 23, 95-146. von Holst, D. (1969). Sozialer Stress bei Tupajas (Tupai belangeri). Die Aktivierung des sympathischen Nervensystems und ihre Bezihung zu hormone11ausgelosten ethologischen und physiologischen Veranderungen. Z . Vergl. Physiol. 63, 1-58. von Holst, D. (1972a). Renal failure as the cause of death in Tupaia belangeri exposed to persistent social stress. J. Con7p. Physiol. 78, 236-273. van Holst, D. (1972b). Die Funktion der Nebennieren mannlicher Tupaia belangeri. J. Cornp. Physiol. 78, 289-306. von Holst, D. (1985a). Coping behaviour and stress physiology in male tree shrews (Tupaia belangeri). Fortschr. Zool. 31, 461-470. von Holst, D. (1985b). The primitive eutherians 11: A case study of the tree shrew, Tupaia belangeri. In “Social Odours in Mammals” (R. E. Brown and D. M. McDonald. eds.), Vol. 1, pp. 155-216. Clarendon Press, Oxford. von Holst, D. (1986a). Vegetative and somatic components of tree shrew’s behavior. J. Auton. Nerv. Syst., Suppl., pp. 657-670. von Holst, D. (1986b). Psychosocial stress and its pathophysiological effects in tree shrews (Tccpaia belangeri). In “Biological and Psychological Factors in Cardiovascular Disease” (T. H. Schmidt, T. M. Dembroski, and G. Bliimchen. eds.). pp. 508-516. Springer, Berlin. von Holst, D. (1987). Physiologie sozialer Interaktionen-Sozialkontakte und ihre Auswirkungen auf Verhalten sowie Fertilitat von Tupajas. Physiol. Aktuel. 3, 189-208. von Holst, D. (1994). Auswirkungen sozialer Kontakte bei Saugetieren. Biol. Unserer Zeit 24, 164-174. Wade, S. E. (1991). An optimized method for measurement of salivary corticosteroids. In “Assessment of Hormones and Drugs in Saliva in Biobehavioral Research” (C. Kirschbaum, C., F. R. Read, and D. H. Hellhammer, eds.), pp. 3-17. Hogrefe & Huber, Seattle, WA. Walker. M. L., Gordon, T. P., and Wilson, M. E. (1983). Menstrual cycle characteristics of sexually breeding rhesus monkeys. Biol. Reprod. 29, 841-848. Ward, I. L. (1984). The prenatal stress syndrome: Current status. fsychoneicroendocrinology 9, 3-11. Wasser, S. K. (1996). Reproductive control in wild baboons measured by fecal steroids. Biol. Reprod. 55, 393-399. Watson, A., and Moss, R. (1970). Dominance, spacing behaviour and aggression in relation to population limitation in vertebrates. In “Animal Populations in Relation to their Food Resources” (A. Watson, ed.), pp. 167-218. Blackwell, Oxford. Wedekind, C. (1994). Mate choice’ and maternal selection for specific parasite resistance before, during and after fertilization. fhilos. Trans. R. Soc. London, Ser. B. 346,303-31 1. Weiner. H. (1977). “Psychobiology and Human Diseases.” Elsevier, New York. Weiner, H. (1991). Behavioral biology of stress and psychosomatic medicine. In “Stress, Neurobiology and Neuroendocrinology” (M. R. Brown, G. F. Koob, and C. Rivier, eds.), pp. 23-51. Dekker, New York.

STRESS AND ITS RELEVANCE FOR ANIMAL BEHAVIOR

131

Weinmann. C. J., and Rothman. A. H. (1067). Effects of stress upon acquired immunity to the dwarf tapeworm, Hymenolepis nana. Exp. Parasitol. 21, 61-67. Weiss, J. M. (1971). Effects of punishing the coping response (conflict) on stress pathology in rats. 1. Conzp. Physiol. Psychol. 77, 1-13. Weiss, J. M. (1972). Influence of pyschological variables on stress-induced pathology. In “Physiology, Emotion and Psychosomatic Illness” (R. Porter and J. Knight, eds.), Ciba Found. Symp. 8, pp. 253-279. Elsevier, Amsterdam. Weiss. J. M. (1984). Behavioral and psychological influences on gastrointestinal pathology: Experimental techniques and findings. In “Handbook of Behavioral Medicine” ( W. D. Gentry, ed.), pp. 174-221. Guilford. New York. Wenar, C. (1983). “Psychopathology from Infancy through Adolescence. A Developmental Approach.” Random House, New York. White, P. J. (1986). Female discrimination of male dominance by urine cues in hamsters. Physiol. Behav. 37, 273-277. Whitman, D. C.. and Albers. H. E. (1995). Role of oxytocin in the hypothalamic regulation of sexual receptivity in hamsters. Brain Res. 680, 73-79. Wiener, S. G., Lowe, E. L., and Levine. S. (1992). Pituitary-adrenal response to weaning in infant squirrel monkeys. Psychobiology 20, 65-70. Williams, J. R., Carter, C. S., and Insel, T. (1992). Partner preference in the development in female prairie voles is facilitated by mating and central infusion of oxytocin. Ann. N. Y. Acad. Sci. 652, 487-489. Williams, J. R., Insel, T. R., Harbaugh, C. R.. and Carter, C. S. (1994). Oxytocin administered centrally facilitates formation of a partner preference in female prairie voles (Mirrorus ochrogaster). J . Neuroendocrinol. 6, 247-250. Wilson. M. E., Gordon, T. P., and Bernstein, 1. S. (1978). Timing of births and reproductive success in rhesus monkey social groups. J . Med. Primatol. 7, 202-212. Winslow, J. T., Shapiro, L., Carter, C. S., and Insel, T. R. (1993). Oxytocin and complex social behavior: Species comparison. Psychopharmacol. Bull. 29,409-414. Wise, D. A,, Eldred. N. L., McAfee, J., and Lauber, A. (1985). Litter deficits of socially stressed and low status hamster dams. Physiol. Behav. 35, 775-777. Woolley, P. (1966). Reproduction in Antechinus spp. and other dasyurid marsupials. Symp. Zool. Soc. London 15,281-294. Wurtman. R. J., and Axelrod, J. (1966). Control of enzymatic synthesis of adrenaline in the adrenal medulla by adrenal corticoid steroids. J. Clin. Chem. 241, 2301-2305. Yu, G. Z., Okutani. F., Takahashi, S., and Higuchi, T. (1996). The olfactory bulb: A critical site for oxytocin in the induction of maternal behaviour in the rat. Neurosciences 72,10831988. Zalcman. S., Kerr, L., and Anisman, H. (1991). Immunosuppression elicited by stressors and stress-related odors. Brain, Behav., Immrcnol. 5, 262-274. Zobelein. H. (1996). Sozialer Rang, Reproduktionserfolg und Fitness mannlicher Europaischer Wildkaninchen Oryctolagtcs cuniculus L. Unpublished Doctoral Dissertation, University of Bayreuth, Bayreuth, Germany. Zuckerman, S. (1932). “The Social Life of Monkeys and Apes.” Harcourt, New York. Zuk, M. (1994). Immunology and the evolution of behavior. I n “Behavioral mechanisms in Evolutionary Ecology” (L. A. Real, ed.), pp. 354-368. University of Chicago Press, Chicago. Zwilling. B. S. (1994). Neuroimmunomodulation of macrophage function. I n “Handbook of Human Stress and Immunity” (R. Glaser and J. K. Kiecolt-Glaser, eds.), pp. 53-76. Academic Press, San Diego, CA.

This Page Intentionally Left Blank

ADVANCES IN THE STUDY OF BEHAVIOR, VOL. ?I

Stress and Immune Defense VICTORAPANIUS

DEPARTMENT OF BIOLOGICAL SCIENCES FLORIDA INTERNATIONAL UNIVERSITY 33199 UNIVERSITY PARK,MIAMI,FLORIDA

I. INTRODUCTION It is a clichC to state that stress reduces immunocompetence. Although this phrase is widely believed and supported by incontrovertible evidence, the biological reality is much more complex and the topic remains an active area of research by ethologists, endocrinologists, neurobiologists, immunologists, pathologists, and parasitologists. Interest in the relationship between behavior and immune function has spawned the field of psychoneuroimmunology, with several new journals reporting recent research. A resurgence of interest in the role of parasites in the evolution of host life histories has provided an impetus for integrating this biomedical information into a Darwinian framework. The purpose of this review is to demonstrate the simplistic nature of the phrase “stress suppresses immunity” and to discuss how stress alters immunocompetence. A consideration of the words stress and immunocompetence will show that these terms can be twisted to serve any purpose. Progress in this field requires precise functional definitions of stress and immunocompetence with an appreciation of the multifactorial and nonlinear nature of these subjects. For example, the stress response suppresses particular immunological mechanisms while enhancing others. Stress, which can be immunosuppressive in the short term, can also enhance immunological reactivity in the long term. It is hypothesized that these immunological mechanisms have differential costs and benefits and that these shifts enhance survival based on energetic considerations. There is no shortage of reviews concerning the endocrinological aspects of the stress response (Chrousos and Gold, 1992; Sapolsky, 1992), the endocrine regulation of immunity (Blalock, 1989, 1994; Reichlin, 1993; Leonard and Song, 1996), and the relationship between stress and immunity (Khansari et al., 1990; Stein and Miller, 1993; Daynes et al., 1995; Besedovsky, 1996; Friedman et al., 1996; Ottaviani and Franceschi, 1996). The 133

Copyright 0 1998 by Academic Press All rights of reproduction in any form reserved 0065-3454/’)8 $25 00

134

VICTOR APANIUS

purpose of this review is to highlight certain aspects of the relationship between stress and the vertebrate immune system that cause generalizations to be biologically simplistic. An additional goal is a synthesis based on energetic considerations that may provide a framework for understanding the adaptive significance of this phenomenon. It is hoped that the reader will appreciate the daunting complexity of the neuroendocrine-immune axis but will also be stimulated to think about the evolution of this vital system from a functional rather than purely mechanistic viewpoint. 11. THENATURE OF STRESS Almost by definition, living organisms respond homeostatically to environmental variation, whether it is the subtle variations in salinity experienced by an estuarine crustacean or the transhemispheric migration of shorebirds. The delineation between a normal homeostatic response and a stress response is problematic. Typically, physiological parameters, for example, plasma glucocorticoid concentration, are used to delineate the range of environmental or social conditions that are stressful to an animal. There are a number of problems with this approach, including the fact that it is tautological. That is because we define stress as the state where stress hormones are elevated above some arbitrarily defined threshold. A more quantitative biochemical approach uses the ratio of energycharged adenine nucleotides to all such adenine nucleotides to give a measure called the adenine energy charge (AEC): AEC

=

(ATP) + 1/2(ADP) (ATP) + (ADP) + (AMP)

(Hochachka and Somero, 1984). It has been argued that this energetically based measurement reflects the physiological performance of the organism in particular environments and provides a universal metric for comparisons across taxa. Individuals living in stressful environments or experiencing stressful events would have a lower AEC, refleding a lower biosynthetic capacity for growth, reproduction, or storage. Although A E C can be conveniently measured, it is more problematic to relate it to survival rates or lifetime reproductive success. For the evolutionary ecologist or geneticist, a definition of stress should logically incorporate the individual’s (inclusive) fitness. A stressful environment can be defined as one in which conditions do not allow population persistence through local reproduction. Species ranges are limited because of physical or biotic factors that would intrinsically be stressful to individuals. Selection would favor phenotypic plasticity as a response to the unpre-

STRESS A N D IMMUNE DEFENSE

135

dictable biophysical and biotic regimes that occur not only at the species margin but throughout its range (Hoffmann and Parsons, 1991). Phenotypic plasticity is one way that the stress can be detected but phenotypic plasticity will also depend on the extent of gene flow. It can be argued that all of these approaches simply provide operational definitions, but the root of the problem is that stress is often considered to be a dichotomous state, which is defined on the basis of continuously varying physiological parameters. This allows proliferation of the situations to which the word stress is applied to the point where it is meaningless. Because glucocorticoid levels are generally higher during the active phase than during sleep, one can talk about the stress of daily life. Glucocorticoid levels are elevated during reproduction, allowing one to speak of the stress of reproduction. Glucocorticoid levels are elevated in migrant songbirds, suggesting that there is a stress of migration. Parasitic infections are associated with increased glucocorticoid levels, implying a stress of infection. Glucocorticoid levels are elevated in subdominant individuals following the initial agonistic encounters that establish dominance hierarchies, an example of social stress. Yet all of these conditions are typical events for these animals. Therefore, a useful definition of stress needs to capture the intensity and transient nature of these conditions. Operational definitions of stress are further complicated by the neuroendocrine adaptation to stressful stimuli, which is manifested by the “coping” strategies of individual animals to escapable versus nonescapable stress. Nonetheless, laboratory and field experiments can informatively test hypotheses about the adaptive significance of the stress response when stressful stimuli are clearly defined and consistently applied to the subjects. The animal’s capacity to respond to stressful situations has a genetic component (Hoffmann and Parsons, 1991). The magnitude of the generalized stress response has been altered through artificial selection in poultry. Turkeys have been selected for high and low responsiveness to cold stress (Brown and Nestor, 1973). Chickens have been selectively bred for high and low corticosterone production in response to social stress (Gross and Siegel, 1985). As expected, the birds that responded to stress with higher circulating levels of corticosterone showed a greater susceptibility to virusinduced tumors and coccidiosis and increased ectoparasite populations. This increased susceptibility to malignant and infectious disease in the birds responding with high corticosterone levels was associated with lowered immunocompetence, as measured by antibody and cell-mediated responses (defined later). This research on domestic birds not only demonstrates a genetic basis for the magnitude of the stress responses but also demonstrates its lability in response to selection.

136

VICTOR APANIUS

Stress-induced immunosuppression is a paradoxical phenomenon. The unfettered operation of the immune system is crucial to an individual’s health, vigor, and, ultimately, survival. Yet, at the time when an animal’s survival is challenged by extreme environmental change or by tremendous physical exertion, the immune system appears to be downregulated. Prolonged stress can cause the premature regression of the primary lymphoid organs and effectively truncate the ontogeny of immunity. Prolonged stress can also induce atrophy of the secondary lymphoid organs, rendering the animal more susceptible to infectious and malignant disease. Even acute episodes of stress can reduce the effectiveness of the immune system to control opportunistic parasites. It is likely that immunity is temporarily downregulated to make nutrients available for other organismal processes that have a higher priority, that is, the nervous system and the musculature. The mechanistic basis of stress-induced immunosuppression is reasonably well understood. The physiological basis for re-allocating nutrients from lymphoid tissue to other somatic compartments is poorly understood. The evolutionary basis for stress-induced immunosuppression is only dimly understood and will require a multidisciplinary approach to understanding this paradox. In conclusion, stress refers to a perturbation of the organism’s homeostatic mechanisms that entails a definable suite of physiological responses. For any particular organism, the stressful stimuli and the physiological responses need to be operationally defined, especially in terms of the frequency and duration of the perturbation. Stress leads to a reduction in the energy and materials available for maximizing lifetime reproductive success. Because the organism expends nutrients to surmount proximate physiological and ecological threats to survival, less nutrients are available to fuel reproductive functions. The measurement and interpretation of stressful conditions remains a rigorous challenge to researchers investigating the relationship between stress and immunocompetence.

111. THENATURE OF IMMUNOCOMPETENCE

Resistance to disease and immunocompetence are often conflated and used interchangeably in nonmedical writing. Properly, resistance is defined as the ability of the host to prevent disease arising from endogenous (e.g., tumors) or exogenous (e.g., infectious agents) sources. Resistance is the most general term and includes genetic, behavioral, and environmental resistance modes. Allelic variants coding for alternative biochemical products can prevent infection or reduce the disease severity because the parasite is not capable of efficient utilization of the variant protein of the host.

STRESS AND IMMUNE DEFENSE

137

The sickle cell allele of hemoglobin is the most widely cited example of genetically based resistance. Preening or grooming to remove ectoparasites exemplifies behavioral resistance. Commensal microbial populations in the gut provide protection from more invasive enteric bacteria and provide an example of environmental influences on disease resistance. Notice that all of these examples of resistance lack an immunological component. Immune-mediated resistance to infectious disease relies on two important arms of the vertebrate immune system-natural (or innate) immunity and acquired immunity. There is an unfortunate tendency for immunologists to use the term “adaptive” in place of “acquired” immunity. Natural immunity involves protective mechanisms that are constituitively expressed, such as macrophages and natural killer cells, and does not require prior exposure to the infectious agent. In contrast, acquired immunity is based on induced responses that are very specific for particular parasite-derived antigens and that retain memory of them. Acquired immune responses are generated by lymphocytes derived from the thymus (T lymphocytes) and the bursa of Fabricius (B lymphocytes), or the nonavian equivalent of the bursa-bone marrow. Historically, acquired immunity has been divided into humoral (antibody-mediated) and cellular (cell-mediated cytotoxic) responses, depending on whether immunity can be transferred by soluble or cellular elements, respectively. Currently, lymphocyte-mediated responses are classified as TH1or TH2,based on the pattern of cytokines (intercellular regulatory molecules) produced during the response. Generally speaking, TH1 responses involve cell-mediated cytotoxic responses against intracellular pathogens, whereas TH2responses involve antibody responses that are most effective against bacteria, extracellular protozoa, and most helminths. The regulatory switch between TH1 and TH2immune responses is actively being investigated and it appears that elevated steroid hormones are one factor that shifts the balance from T H 1 to TH2 (Mason, 1991; Rook et af., 1994). This simple overview not only provides the necessary background for the ensuing discussion but also demonstrates that steroid hormones do not have a simple, single effect on immune function. Immune responses arise from the interaction of numerous cells types and are regulated by endocrine and cytokine networks with complex feedback loops. As an example, bacteria entering through broken skin attract granulocytes, such as neutrophils that emigrate through the endothelial wall to the site of infection. These inflammatory cells degranulate and release toxic compounds into the tissue, which results in killing of bacteria and localized necrosis. Bacteria and antigens released from the lysed bacteria encounter macrophages in the draining lymph node. These macrophages ingest, proteolytically process, and then present antigen-derived peptides on the cell surface to receptors on T helpers (TH2) and B lymphocytes.

138

VICTOR APANIUS

When activated, these lymphocytes proliferate rapidly, doubling every 14 hr. T helper lymphocytes secrete cytokines that drive the clonal expansion of B lymphocytes bearing receptors with the highest affinity for antigen. These cells secrete antibodies, which are used by granulocytes and macrophages to clear the remaining bacteria. The essential features of a lymphocyte-mediated immune response is the interaction with macrophages and the intense proliferation of lymphocytes at the peak of the response. The combination of natural immune mechanisms and the induction and regulation of antigen-specific immune responses leads to the phenotypic trait called immunocompetence. This can be thought of as the homeostatic regulation of host-parasite interactions, as animals are continually challenged by invasive microbes in the gut as well as opportunistic parasite infections. Research utilizing transgenic “knock-out’’ animal models has revealed a surprising robustness of the vertebrate immune system. For example, disrupting the antigen presentation pathway used for the detection of intracellular parasites had a surprisingly modest effect when the deficient animals were challenged with experimental infections using a diversity of intracellular infectious agents (reviewed in Apanius et al., 1997). From these studies, one can infer that there is a remarkable redundancy in immune functions. Deletion of antigen-specific responses in the mice mentioned here was compensated by increased activity of natural killer cells, an arm of the natural immune system. This underscores one of the paradoxes of the relationship between stress and immunocompetence. For example, elevated corticosteroids are related to increased traffic of granulocytes in the vasculature and sequestration of lymphocytes in lymphoid tissue (Dhabhar et al., 1995). This leads to an increased leukocyte count, which has been misinterpreted by some to signal increased immunocompetence (Dufva and Allander, 1995; Gustafsson et al., 1994). This redistribution of leukocytes permits an immediate mobilization of granulocytes, which are preformed and are able to enforce the first line of defense at lesions. Antigens released from the site of infection drain to the lymph nodes where lymphocytes have collected to cooperate with antigen-presenting cells in antigen-driven proliferation. If the stressful condition continues and chronically elevates glucocorticoid levels, then nonlymphocyte mediated immune mechanisms, for example, natural killer cells, are upregulated to compensate (Fowles et al., 1993). Thus, stress does not necessarily reduce immunocompetence because, in some cases, it actually enhances some components of immunocompetence as part of a compensatory response. This requires investigators to be very specific about the form, intensity, and duration of stress and the particular measure of immunocompetence that is taken.

STRESS AND IMMUNE DEFENSE

139

Assessing immunocompetence is made difficult because of (1) the complexity of a seemingly simple process, such as an antibody response against a single antigen; (2) the overlapping and redundant nature of immune mechanisms; and ( 3 ) the ability of one component of immunity to be upregulated to compensate for ineffective responses by other components. In practical terms, a battery of immunological tests are performed to assess elements of the natural, humoral, and cellular immunity. These tests can be performed only on captive animals, except under exceptional circumstances such as in nestling birds, and usually entail euthanasia to collect spleen or lymph node tissue. Under these circumstances, it is not difficult to understand that immunocompetence of wild animals is often discussed only in theoretical terms due to the paucity of empirical data. The situation is made more difficult because of phylogenetic variation in immune mechanisms and because immune function of wild animals may not correspond to that measured in domestic animals living in the sanitary conditions of captivity. A surprising gap in our understanding of immunocompetence is the physiological cost of immune function. A number of studies have clearly shown that nutrient limitation, especially protein deficiency, can reduce immunocompetence. Yet, some immune mechanisms, principally elements of natural immunity, are apparently enhanced by moderate nutrient limitation, possibly as a compensation for reduced acquired immune responsiveness. In general, the nutritional studies show that immune function can be altered by trace nutrient, calorie, and protein deficiency and that lymphocyte-mediated cytotoxic responses are most sensitive, followed by antibody responses and then natural immunity (Chandra and Newberne, 1977;Gershwin et al., 1985; Klasing et al., 1991; Cunningham-Ruddles, 1993). Nonetheless, the metabolic cost of specific immune mechanisms is unknown and the total fraction of basal metabolic rate or daily energy expenditure that is allocated to immune function remains conjectural. This is unfortunate because the immunoregulatory strategies under stressful conditions may involve re-allocation of nutritional resources in the light of the cost of particular immune mechanisms. Without a better understanding of the physiological cost of immune function these hypotheses cannot be tested. With these caveats in mind, this review will discuss the current understanding of how stress alters immunocompetence. Generalizations about the relationship of stress and host susceptibility to parasitism will not be directly addressed because so many additional factors influence hostparasite interactions. This review will necessarily rely on detailed endocrinological and immunological information gathered using laboratory animal models. To the extent possible, the relationship of the neuroendocrineimmune axis to other physiological processes will be considered with the

140

VICTOR APANIUS

hope of reaching a better understanding of the strategy for stress-induced immunomodulation.

IV. NEUROLOGICAL LINKAGE OF STRESS AND IMMUNOCOMPETENCE The response to stress is a neurological cascade of events that begins at the electrophysiological level in specific sites in the brain and continues with sustained secretion of adrenal hormones that entrain long-term physiological processes leading eventually to habituation or emelioration of the stressful condition. This chain of events is mediated by two pathways: the sympathoadrenal (SA) and the hypothalamo-pituitary-adrenocorticalaxis (HPA). In this section, events within the central nervous system (CNS) will be reviewed. Also considered will be immunological feedback to the CNS, that is, the influence of immunological processes on neurological activity and behavior of the host. This communication between the immune and nervous systems is significant because most disease states can be considered as a source of stress and the response involves immune mechanisms that are assisted by particular host behavior modifications, such as sleep. Thus, the effect of stress on immune function is not a simple causal chain but involves feedback loops that make the relationship more complex. The immediate neurological response to stressful stimuli is activation of cerebral catecholamine circuits, neurons which show an elevated secretion and catabolism of norepinephrine, dopamine, and epinephrine. Cerebral serotonin circuits also appear to play a role in mediating the stress response in the central nervous system. Although these neurons project to many regions of the brain, the turnover of norepinephrine in the hypothalamus is notable (reviewed in Dunn, 1996). Corticotropin-releasing factor (CRF) appears to be the central regulator of the stress response at the neural level. Injection of CRF directly into the brain mimics many of the endocrine and physiological responses to stress, including (1) activation of the HPA and SA axes; (2) inhibition of growth hormone and gonadotropins; and (3) behavioral responses such as anorexia and reduced physical activity (Owens and Nemeroff, 1991). These responses result from the pituitary release of adrenocorticotrophic hormone, which leads to the production of glucocorticoids by the adrenal cortex, which then induce these systemic effects. These will be covered in more detail in the next section. An important finding that has stimulated psychoneuroimmunological research has been the discovery of autonomic nervous system innervation of the peripheral lymphoid tissue. This provides an empirical basis for a direct linkage of neural activity with immune function. To date, it has

STRESS AND IMMUNE DEFENSE

141

been established that noradrenergic sympathetic nerve fibers innervate the primary and secondary lymphoid organs. The primary lymphoid organs, for example, thymus and bone marrow, are the sites of lymphopoiesis during ontogeny. These organs are invested with noradrenergic fibers that appear to influence cellular proliferation, differentiation, and emigration rates. In the thymus, noradrenergic nerve fibers enter the tissue with the vasculature and are found in highest density at the corticomedullary junction, which is a location that can potentially control the flux of differentiated cells through the tissue. In secondary lymphoid tissues, for example, spleen, lymph nodes, and gut, noradrenergic fibers radiate throughout the parenchmya and contact lymphocytes as well as macrophages and other antigenpresenting cells. The precise role for these nerve tracts is not clear, but the rich assortment of cognate receptors for neurotransmitters on immunocytes suggests that the neurons are signaling to resident and circulating cells of the immune system (reviewed in Ader et al., 1990). It is these histological and neurochemical observations that lend credence to a remarkable body of experiments on physiological and behavioral conditioning of immune function. These experiments date back to the 1920s in the Soviet Union and follow the classical conditioning paradigm of coupling conditioned and unconditioned stimuli. In recent experiments conducted by Ader and colleagues, the novel taste of saccharin (conditioned stimulus) was paired with an immunosuppressive agent (unconditioned stimulus). Upon reexposure to saccharin, depressed immune function was observed. Because these experiments were greeted with great skepticism, they have been repeated in a number of laboratories with additional immunological assays and experimental psychological techniques. These experiments have demonstrated that behavioral conditioning affects many components of immune function, such as antibody production, lymphocyte proliferation, foreign tissue rejection, as well as delayed-type allergic reactions. Most importantly, enhancement of immune responses was also observed when immunostimulatory drugs were paired with the conditioned stimulus. This and other sources of evidence argue against conditioning of immune function being simply a stress response (reviewed in Husband, 1993). The sympathetic innervation of the lymphoid compartments not only complements the hypothalamo-pituitary-adrenal axis for modulating immune function during stress but also forms another regulatory network with feedback. Cytokines released in localized immune reactions can enter the systemic circulation and bind to cognate receptors throughout the brain. Interleukin 1 (IL-1) is secreted by macrophages at the initial stages of a specific immune response and has profound effects at a number of physiological levels. In the brain, elevated IL-1 levels are associated with an increased time spent in slow-wave (non-REM) sleep and an increased

142

VICTOR APANIUS

hypothalamic set point for body temperature. IL-1 is also associated with anorexia, lethargy, and inhibition of locomotory activity, which characterizes the malaise of various disease states. Interestingly, a number of bacterial cell-wall products, most notably muramyl peptides, have similar biological properties. The neural and behavioral effects induced by IL-1 span a wide spectrum. There is a circadian rhythmicity in physiological concentrations of IL-1 that induces mild but measurable changes, in body temperature, for example, and at the other extreme chronic malignant disease is often associated with a chronic wasting syndrome that can be re-created in experimental animals with exogenous IL-1 administration. The effects of IL-1 do not occur in isolation. Not only is this cytokine part of a central immunoregulatory network, which includes glucocorticosteroids, but it also mediates numerous other metabolic processes in the liver, skeletal muscle, vascular system, and hematopoietic system (reviewed in Dinarello, 1992), all of which ultimately affect the physiological condition and performance of the individual. The linkage of the nervous and immune systems through neurotransmitters and cytokines appears to be deeply integrated with organismal processes such as foraging behavior, digestion, energy expenditure, as well as reproduction. The relationship between stress and immunocompetence can be better rationalized if these additional behavioral and physiological processes are also considered.

v.

ENDOCRINE LINKAGE OF STRESS A N D

IMMUNOCOMPETENCE

Glucocorticoids have had a long history in therapeutic medicine, yet current research continues to enlarge our understanding of these central mediators of stress and immunocompetence. The conventional wisdom is that glucocorticoids suppress immune responses and reduce inflammation. These hormones link stress, as perceived in the brain, with: (1) impaired immunocompetence; (2) increased susceptibility to infectious and malignant disease; and (3) decreased susceptibility to autoimmune diseases. Longterm administration of high doses of glucocorticoids leads to a reduced mass of primary and secondary lymphoid organs due to a depletion of lymphocyte cellularity and regression to a condensed epithelial reticulum (reviewed in Cupps and Fauci, 1982). These statements are supported by a formidable body of experiments and underlie many modern medical therapies. What is less widely appreciated is that glucocorticoid levels vary on a daily basis and are increased during the course of an immune response. Thus, glucocorticoids are an essential component of the endogenous immunoregulatory network.

STRESS AND IMMUNE DEFENSE

143

Circulating glucocorticoid levels show a circadian oscillation. Levels increase during an animal’s active phase and decline during the inactive phase. IL-1 shows an inverse circadian pattern, with higher levels in the animal’s inactive period (Zabel et af., 1990). These circadian oscillations are mirrored by leukocyte traffic patterns (Abo et al., 1981; Kawate et af., 1981) and antibody responsiveness (Fernandes et af., 1976). Moreover, sleep deprivation impairs antibody responsiveness and the immunological deficit can be restored by administration of IL-1 (Brown et al., 1989). It is now known that IL-1 regulates the secretion of adrenocorticotrophic hormone and other pituitary hormones, including growth hormone, thyroidstimulating hormone, and prolactin (Bernton et al.,1987). The combination of pituitary hormones and IL-1 is associated with: (1) increased body temperature; (2) increased catabolism of glycogen; (3) increased catabolism of skeletal muscle; and (4) increased turnover of circulating amino acids (Klasing, 1988). These metabolic changes can be viewed as a mobilization of nutrient stores to provide substrates for activation and proliferation of cells involved in inflammation and immunity. Thus, the flow of nutrients into or away from the lymphoid compartment is principally controlled by the levels of IL-1 and corticosterone, respectively. When a specific immune response is induced by infection or immunization, circulating glucocorticoid levels transiently rise in the first 24 hr of the response. There is increasing evidence that a modest increase of corticosteroid levels does not always depress immunocompetence and in some cases may actually be associated with enhanced immune responsiveness. For example, prior handling or saline injections of mice to condition them to the immunization procedure led to reduced antibody responses compared to controls that were handled only at immunization. Pre- and postimmunization glucocorticoid levels were comparable and stable in the mice conditioned to injections or handling but hormone levels were elevated postimmunization in the mice handled only once. The unconditioned animals had higher antibody titers despite having higher corticosteroid levels following immunization (Moynihan er af., 1989). This would not be expected from the traditional view that glucocorticoids are immunosuppressive. There are several plausible interpretations of this outcome. One line of reasoning regards the elevated corticosterone levels in animals with elevated antibody responses as evidence that glucocorticoids released during an induced immune response are part of a negative feedback loop, whereby the enhanced response is reduced by the concomitant increase in immunosuppressive hormone later in the response. This is supported by a great deal of empirical evidence whereby hypophysectomized animals suffer increased immunopathological disease from overreactive immune responses. Another interpretation is that elevated glucocorticoid levels were responsible for

144

VICTOR APANIUS

increasing the flux of nutrients in the immunized animal and the dual signals increased cell proliferation and antibody production in lymphoid tissues. Because we currently lack the relevant information to address this hypothesis it cannot be rejected outright. The important point is that physiological and pharmacological levels of glucocorticoids are often associated with reduced immunocompetence, but not unequivocally. The additional dimension of behavioral conditioning, through daily handling or repeated injections, shows that the effects of glucocorticoids on immunity are not simply linear and additive but interact with other endocrine factors and the neural processes. The relationship between energy metabolism and immunocompetence is suggested by additional endocrine factors that regulate both of these processes. Growth hormone, prolactin, and thyroid hormone generally have stimulatory effects on lymphocytes. Congenital deficiency of growth hormone is associated with fewer cells in the primary lymphoid organs as well as reduced natural and acquired immunity, which can be reversed, in part, by exogenous administration of the hormone in birds and mammals. Experimental administration of growth hormone enhances immunocompetence and even reverses some of the age-related decline in immune function (Gelato, 1996). There is evidence that circulating growth hormone levels increase (1) during infectious disease; (2) in response to elevated IL-1; and (3) in response to bacterial cell-wall products called endotoxins (Gala, 1991). Furthermore, lymphoid cells of the thymus and spleen secrete growth hormone as well as express growth-hormone-releasing hormone receptors (Guarcello et al., 1991). Thus, growth hormone acts within the immune system, probably in concert with insulinlike growth factor I, to positively stimulate lymphocyte-mediated responses. Systemic effects of growth hormone on nutrient mobilization would coincide with increased activity of the lymphoid compartment. Thyroid hormones, notably thyroxine, appear to play an immunostimulatory role within the endocrinekytokine network. Endogenous administration of thyroxine produces variable results in terms of altering the immunocompetence of normal animals. However, animals with congenitally or pharmacologically impaired thyroid function have reduced thymus and spleen mass and decreased cell-mediated responses, which can be, at least partially, restored with exogenous administration of thyroid hormones (reviewed in Marsh, 1992). More recently, thyroid-stimulating hormone has been shown to be secreted by activated lymphocytes, which in turn promotes the ability of helper T lymphocytes to increase antibody production by B lymphocytes in vitro (Kruger et al., 1989). This provides another example that hormones that regulate metabolism also control lymphocytemediated immunity.

STRESS AND IMMUNE DEFENSE

145

In the generalized stress response, prolactin secretion by the anterior pituitary is increased. Growth hormone secretion may transiently increase early in the stress response but is typically decreased in some species or later in the response. Despite this difference in hormone secretion pattern, prolactin has many of the same effects on immune function as growth hormone. Markedly increased or decreased circulating levels of prolactin are immunosuppressive, suggesting that normal physiological concentrations of this hormone are necessary for immunological homeostasis. Immunogenic stimulation leads to increased production of prolactin by the pituitary as well as lymphocytes. Lymphocytes have prolactin receptors indicating autocrine and paracrine functions for this hormone within the immune system (Matera, 1996). Prolactin levels are dramatically elevated in reproducing animals and these same individuals are known to be immunosuppressed. It has been suggested that the elevated prolactin levels in lactating sheep is the basis for the well-documented phenomenon of periparturient increase in fecal shedding of intestinal nematode eggs. However, pharmacological manipulation of ewes has been successful in both increasing prolactin in nonlactating females and decreasing prolactin in lactating females, while no effect has been found on measures of immunity or on intestinal parasite egg shedding (reviewed in Barger, 1993). Although additional endocrine factors have been proposed to account for immunosuppression associated with parturition and lactation, it is equally likely that the high level of energy expenditure associated with lactation (Thompson, 1992) may be the ultimate cause. This hypothesis is supported by manipulation of parental effort in birds. The proportion of individuals infected with blood parasites has been positively associated with artificially increased parental effort in a number of studies (reviewed in Sheldon and Verhulst, 1996). It is also known that daily energy expenditure (Deerenberg et al., 1995) and corticosterone levels (Silverin, 1982; Hegner and Wingfield, 1987) are positively related to parental effort in birds. A recent study provides evidence that increased blood parasitism in birds with artificially increased brood sizes is associated with a hematological measure of stress, increased granulocyte : lymphocyte ratio (Ots and Hbrak, 1996). In sum, these studies indicate that increased reproductive effort (1) increases nutrient demand; (2) increases energy turnover; (3) increases circulating glucocorticoid levels; and (4) increases the prevalence and intensity of parasitism. These studies link high levels of physical activity and energy metabolism with increased susceptibility to parasitism. VI.

ALTERS IMMUNOCOMPETENCE WHYSTRESS

It is tempting to finish with a diagram or table to summarize the bidirectional communication between the neuroendocrine and immune systems,

146

VICTOR APANIUS

with pluses and minuses to show positive and negative effects. Readers seeking this information should consult the reviews listed in Section I. These summary figures fail to show the complex feedback networks, compensatory cross-regulation, nonlinear dose dependence, and temporal dependence that characterize these phenomena. As such, the depictions suggest that the relationships are generalizable and consistent across species. The generalizations that can be offered have already been made in this review and any detailed listing of hormones and cytokines will soon be out of date. In its place, I offer a synthesis that moves the question away from mechanisms and pathways, which are only partially understood to date, to focus on the adaptive significance of stress-induced alteration in immunocompetence. What is the strategy underlying endocrine regulation of immunity and how has this strategy been shaped by evolution through natural selection? Stress requires that animals make an immediate response that can be physiologically costly. Typically, nutritional reserves are mobilized and nutrient acquisition is suppressed. This physiological regime entails a reduction of physiological processes that are not immediately vital. Immunity is one of the processes that is reduced and this may be due to two factors. First, immune function appears to be costly, although direct evidence is lacking. It is possible that immune function requires the same quantity of nutritional resources as the nervous system. Yet the cost of immune function can be ameliorated by selectively downregulating the most costly components without complete loss of immunity. The other factor that makes immune function susceptible to downregulation during stress is that induction of immune mechanisms in the poststress period permits compensation for the period of suppressed immunity. Thus, the increased level of microparasites that have replicated in the host and the increased number of infective stages of parasites acquired during the stress period may be cleared afterward. At least, that would be the teleological expectation from the host’s point of view. This explanation is supported by empirical observations on the effect of long-term stress on immunity in mice. Mice were subjected to daily auditory stress. From the initial day and continuing for 20 days afterward, the mice showed depressed lymphocyte function and elevated plasma cortisol levels. After that point, and continuing for approximately 20 days, they showed enhanced lymphocyte function and relatively low levels of plasma cortisol (Monjan and Collector, 1977). This indicates that immune function does not simply adapt to chronic stress by returning to the basal level of activity but appears to be increased in a compensatory manner. This may have occurred in these laboratory experiments because food was provided ad libitum.

STRESS AND IMMUNE DEFENSE

147

The relationship between imposed stress and the animal’s nutrient budget is of critical importance in discussing pathogenesis of infections during the stress period. If the stressful event is associated with depressed body temperature or nutrient limitation, then replication of microparasites may be inhibited because of their dependence on the host’s nutritional resources. Viral replication and pathogenicity is often linked to the flux of nutrients through host tissues, especially nitrogen-rich substrates for nucleic acid synthesis. In these circumstances, stress-induced immunosuppression entails less risk to the host than might be generally appreciated. However, stress generally induces a short-term increase in the flux of nutrients through host tissues, to support thermogenesis or muscular activity, and this can promote microparasite replication and macroparasite reproduction. The fact that the host is often immunosuppressed as well makes it difficult to disentangle the two effects. Nevertheless, from first principles one can infer that the stress response in a well-nourished animal is associated with high nutrient turnover, which entails a greater risk of fulminating infectious disease because of the additive and possibly synergistic operation of these two factors. Moderate levels of exercise in humans and captive animals also increases nutrient turnover. This increased nutrient flux is associated with increased circulating levels of epinephrine, adrenocorticotrophin, glucocorticoids, pendorphin, metenkephalin, prolactin, growth hormone, and thyroid hormone (Smith and Weidemann, 1990). Since these hormones also regulate immune function, then physical activity also affects immune function. Generally, moderate levels of activity enhance immunocompetence and parasite resistance (Cannon and Kluger, 1984), possibly through elevated nutrient flux. It is also possible that the increased levels of endotoxin associated with exercise enhance immunocompetence (Cannon and Kluger, 1983). Endotoxin is released in the digestive tract from the breakdown of gramnegative bacteria and increased dietary throughput would thus result in increased endotoxin levels. At higher levels of physical exercise, especially among “elite athletes,” upper respiratory tract infections are more common and a hormonal regime associated with immunosuppression is observed (Hoffmann-Goetz and Pederson, 1994). These studies suggest that an increased flow of nutrients in conjunction with normal immune function can control opportunistic infections in all but the most demanding circumstances. These considerations apply to parasites where there is a direct relationship between host nutrient acquisition and pathogenesis. These parasites are typically associated with economic disease observed in animal production settings, for example, intestinal nematode infections of ruminants. Wellnourished animals harbor low-grade, chronic infections and it appears that

148

VICTOR APANIUS

parasite populations are at equilibrium with immune-mediated expulsion. This type of parasitism is probably widespread in wild animals and it has been extremely difficult to measure the fitness consequences in nature. At the same time, it is also widely recognized that these chronic infections show an opportunistic increase in intensity in the stressed host. A good example is the relapse or recrudescence of blood parasite infections observed in wild bird populations. In avian malaria, there are seasonal increases in prevalence and parasitemia associated with host reproduction. Relapse of avian hematozoa infections can be induced by corticosterone injections (Applegate, 1970). It is informative that although hematozoan parasitemia or nematode egg-shedding increases during the reproductive cycle, these infections are seldom associated with morbidity or mortality in nature. It is tempting to suggest that the increased intensity of parasitism that accompanies reproduction is an important component of the cost of reproduction (Sheldon and Verhulst, 1996), but, to date, there is only one study that shows that the increased parasitism reduces future reproductive success (Mgller, 1993). However, these data are difficult to collect, so the question remains an open one. The available data show that the intensity of parasitism increases during the reproductive cycle and that the infections are eventually controlled, presumably by host immune mechanisms. This may be an example of adaptive modulation of immunity mediated by endocrine factors associated with stress. The host strives to maximize its reproductive output in a seasonal reproductive episode. It re-allocates nutrients from immune function to reproductive effort during peak periods of physical performance and incurs a greater parasite burden. This burden is reduced by immunological mechanisms following the reproductive bout. This hypothesis permits predictions that can be tested through comparative and experimental approaches and provides a framework for viewing the stress response as an adaptive strategy. Numerous studies have documented the causes of death in wildlife populations and have attempted to identify the factors involved. A never-ending controversy centers around the role of infectious agents in natural mortality. Are the individuals that expire over the winter and that are found to harbor larger parasite loads killed by environmental circumstances or because of these parasitic infections? It is often stated that the higher parasite load predisposes these individuals to stress-induced mortality. It is equally probable that individuals stressed by adverse weather may have been unable to control their parasite load. Although it would appear to be simple to tease these factors apart, in practice there are few studies that adequately address the contributions of environmentally induced stress and parasitism to mortality in wild animals.

STRESS AND IMMUNE DEFENSE

149

It is this intertwined process, that is, the progression of infectious disease in a stressed animal, that is crucial to understanding the adaptive nature of stress-induced immunosuppression. Two polarized views can be advanced. One is that infectious disease in a stressed animal is biologically self-limiting. As the host deteriorates in condition during a severely stressful event, the nutrients to support parasite replication or reproduction diminish and the infection becomes less pathogenic. Indeed, there are examples of parasites that are highly pathogenic during the refeeding stage of hosts that have been nutritionally deprived, for example, malaria in undernourished human populations. In this view, stress-induced immunosuppression is clearly advantageous due to the nutrient savings during the stress period and because mortality solely due to parasitism is unlikely. The opposite view is that stress-induced immunosuppression is advantageous only in the short term where the nutrient savings may be critical during the stress period. But as the period of immunosuppression lengthens, then the severity of infectious disease increases in a runaway process in which disease progression contributes to mortality. Despite the intuitive appeal of this view, the supportive evidence is anecdotal and seldom experimental. This stems, in part, from the methodological difficulty of identifying multicausal factors of mortality. Ultimately, selection must favor stress-induced immunosuppression because of the short-term advantages. Whether immunosuppression exacerbates mortality during prolonged stressful episodes remains to be demonstrated.

VII. SUMMARY Alterations in immune function by stress are a graded response that is tightly regulated by neuroendocrine pathways. In mild stress responses, the principal immunological mechanisms shift from lymphocyte-mediated responses, which have high metabolic demands due to rapid cell proliferation, to natural immunity that utilizes the current store of cells and proteins. This provides the individual with a saving in nutrients. Because the less expensive immune mechanisms do not retain memory and are not as specific as those mediated by lymphocytes, there is a modest decrease in immunocompetence over the remaining lifetime of the individual. More intense stress leads to the suppression of additional mechanisms with presumed additional savings and costs. As the stress period lengthens, either the immune mechanisms are restored to their previous levels or they continue to decline, depending on the intensity of the stressful stimuli. In the latter case, immunological control of preexisting infections or new infections breaks down and an increased

150

VICTOR APANIUS

severity of infectious disease is observed. It is at this point that it is no longer possible to interpret the outcome in adaptive terms. Up to this point it can be consistently argued that stress-induced immunoregulation is part of the generalized metabolic response to provide the resources for a shortterm solution to an immediate threat to survival. The stress response incurs a fitness cost in decreased parasite resistance, which may be measured proximally as increased susceptibility or ultimately through decreased immunological memory. The assertion that long-term social stress induces immunosuppression, which in turn allows malignant and infectious diseases t o progress to the point of significant morbidity or mortality, is intriguing but remains conjectural (Sapolsky, 1992). It is biologically plausible, but its relevance, except in highly contrived circumstances, for example, in captive wildlife, is questionable. This assessment is based in part on the compensatory nature of the immune system. Low-grade, chronic stress does alter immune function but does not necessarily reduce disease resistance. More likely, chronic social stress affects access to food and the likelihood of parasite transmission, and these may directly affect the animal’s health in nature. References Abo, T., Kawate, T., Itoh, K., and Kumagai, K. (1981). Studies on the bioperiodicity of the immune response. I. Circadian rhythms of human T. B and K cell traffic in the peripheral blood. J. Immunol. 126, 1360-1363. Ader, R., Felten, D.. and Cohen, N. (1990). Interactions between the brain and immune system. Annu. Rev. Pharmacol. Toxicol. 30, 561-602. Apanius, V., Penn. D., Slev, P. R.. Ruff, L. R., and Potts, W. K. (1997). The nature of selection on the major histocompatibility complex. Crii. Rev. Inmunol. 17, 179-224. Applegate. J. E. (1970). Population changes in latent avian malaria infections associated with season and corticosterone treatment. J. Parasitol. 56, 439-443. Barger, 1. A. (1993). Influence o f sex and reproductive status on susceptibility of ruminants to nematode parasitism. Inl. J. Parasiiol. 23, 463-469. Bernton. E. W.. Beach, J. E., Holiday, J. W.. Smallridge. R. C., and Fein. H. G. (1987). Release of multiple hormones by a direct action of interleukin-I on pituitary cells. Science 238, 519-521. Besedovsky, H. 0. (1996). Immune-neuro-endocrine interactions: Facts and hypotheses. Endoer. Rev. 17, 64-102. Blalock, J. E. (1989). A molecular basis for bidirectional communication between the immune and neuroendocrine systems. Physiol. Rev. 69, 1-32. Blalock, J. E. (1 994). The syntax of immune-endocrine communication. Imrnunol. Today 15, 505-510. Brown, K. I., and Nestor, K. E. (1973). Some physiological responses of turkeys selected for high and low adrenal responses to cold stress. Poiilr. Sci. 52, 1948-1954. Brown, R., Price. R. J., King, M. G.. and Husband, A. J. (1989). Interleukin-1 beta and muramyl dipeptide can prevent decreased antibody response associated with sleep deprivation. Brain, Behav., Immunol. 3, 320-321.

STRESS AND IMMUNE DEFENSE

151

Cannon, J. G.. and Kluger, M. J. (1983). Endogenous pyrogen activity in human plasma after exercise. Science 220, 617-619. Cannon, J. G., and Kluger, M. J. (1984). Exercise enhances survival rate in mice infected with Salmonella typhimurium. Proc. Soc. Exp. Biol. Med. 175, 518-521. Chandra, R. K.. and Newberne, P. M. (1977). “Nutrition, Immunity, and Infection.” Plenum, New York. Chrousos, G. P., and Gold, P. W. (1992). The concepts of stress and stress system disorders. JAMA, J. Am. Med. Assoc. 267, 124441252, Cunningham-Ruddles, S., ed. (1993). “Nutrient Modulation of the Immune Response.” Dekker, New York. Cupps, T. R., and Fauci, A. S. (1982). Corticosteroid-mediated immunoregulation in man. Immunol. Rev. 65, 133-155. Daynes, R. A,, Araneo, B. A,, Hennebold, J., Enioutina, E., and Mu, H. H. (1995). Steroids as regulators of the mammalian immune response. J. Invest. Dermatol. 105, 14s-19s. Deerenberg, C., Pen, I., Dijkstra, C., Arkies, B.-J., Visser, G. H., and Daan, S. (1995). Parental energy expenditure in relation to manipulated brood size in the European kestrel. Zoology, Analysis of Complex Systems 99,38-47. Dhabhar, F. S.. Miller, A. H., McEwen, B. S., and Spencer, R. L. (1995). Effects of stress on immune cell distribution: Dynamics and hormonal mechanisms. J. Immunol. 154,551 15527. Dinarello. C. A. (1992). Role of interleukin-1 in infectious diseases. Immunol. Rev. 127, 119-146. Dufva, R.. and Allander, K. (1995). Intraspecific variation in plumage coloration reflects immune response in Great tit (Parus major) males. Funct. Ecol. 9, 785-789. Dunn, A. J . (1996). Psychoneuroimmunology, stress and infection. In “Psychoneuroimmunology, Stress, and Infection” (H. Friedman, T. W. Klien, and A. L. Friedman, eds.). pp. 25-46. CRC Press, Boca Raton, FL. Fernandes, G., Halberg, F., Yunis, E. J., and Good, R. A. (1976). Circadian rhythmic plaqueforming cell response of spleens from mice immunized with SRBC. J. Immunol. 116, 962-966. Fowles, J. R., Fairbrother, A,. Fix, M., Schiller, S.. and Kerkvliet, N. I. (1993). Glucocorticoid effects on natural and humoral immunity in mallards. Dev. Comp. Irnmunol. 17,165-177. Friedman. H.. Klein, T. W., and Friedman, A. L., eds. (1996). “Psychoneuroimmunology, Stress and Infection.” CRC Press, Boca Raton, FL. Gala, R. R. (1991). Prolactin and growth hormone in the regulation of immunity. Proc. Soc. Exp. Biol. Med. 198,513-519. Gelato, M. C. (1996). Aging and immune function: A possible role for growth hormone. Horm. Res. 45,46-49. Gershwin, M. E., Beach, R. S., and Hurley. L. S. (1985). “Nutrition and Immunity.” Academic Press, Orlando, FL. Gross, W. B., and Siegel, P. B. (1985). Selective breeding of chickens for corticosterone response to social stress. Poult. Sci. 64, 2230-2233. Guarcello, V., Weigent, D. A,, and Blalock, J. E. (1991). Growth hormone releasing hormone receptors on thymocytes and splenocytes from rats. Cell. Imrnunol. 136,291-302. Gustafsson, L., Nordling, D., Anderson, M. S.. Sheldon, B. C., and Qvarnstrom, A. (1994). Infectious diseases, reproductive effort, and the cost of reproduction in birds. Philos. Trans. Soc. London, Ser. B 346,321-331. Hegner, R. E., and Wingfield. J. C. (1987). Effects of brood size manipulations on parental investment, breeding success, and reproductive endocrinology of house sparrows. Auk 104,470-480.

152

VICTOR APANIUS

Hochachka. P. W., and Somero, G. N. (1984). “Biochemical Adaptation.” Princeton University Press, Princeton, NJ. Hoffman-Goetz, and Pederson, B. K. (1994). Exercise and the immune system: A model of the stress response. Irnmunul. Today 15, 382-387. Hoffmann, A. A,. and Parsons, P. A. (1991). “Evolutionary Genetics and Environmental Stress.” Oxford University Press, Oxford. Husband, A. J. (1993). Role of central nervous system and behavior in the immune response. Vaccine 11, 805-816. Kawate, T., Abo, T., Hinuma. S., and Kumagai, K. (1981). Studies on the bioperiodicity of the immune response. 11. Co-variations of muringT and B cells and a role of corticosteroid. J. Immunol. 126, 1364-1367. Khansari. D. N., Murgo. A. J.. and Faith. R. E. (1990). Effects of stress on the immune system. Immunol. Today 11, 170-175. Klasing, K. C. (1988). Nutritional aspects of leukocyte cytokines. J . Nurr. 118, 1436-1446. Klasing, K. C., Johnstone, B. J., and Benson, B. N. (199 1). Implications of an immune response on growth and nutrient requirements of chicks. In “Recent Advances in Animal Nutrition” (W. Haresign and D. J. A. Cole, eds.). pp. 135-146. Butterworth Heinemann, Stoneham, MA. Kruger, T. E.. Smith, L. R., Harbour, D. V.. and Blalock, J. E. (1989). Thyrotropin: An endogenous regulator of the in v i m immune response. J. Immunol. 142, 744-747. Leonard. B. E., and Song, C. (1996). Stress and the immune system in the etiology of anxiety and depression. Pharmacol., Biochern. Behav. 54,299-303. Marsh, J. A. (1992). Neuroendocrine-immune interactions in the avian species- a review. Poulr. Sci. Rev. 4, 129-167. Mason, D. (1991). Genetic variation in the stress response: susceptibility to experimental allergic encephalomyelitis and implications for human inflammatory disease. Immunol. Today 12, 57-60. Matera, L. (1996). Endocrine, paracrine and autocrine actions of prolactin on immune cells. Life Sci. 59, 599-614. Moiler, A. P. (1993). Ectoparasites increase the cost of reproduction in their hosts. J. Anim. Ecol. 62, 309-322. Monjan, A. A,, and Collector, M. I. (1977). Stress-induced modulation of the immune response. Science 196, 307-308. Moynihan. J., Koota, D., Brenner, G., Cohen. N., and Ader. R. (1989). Repeated intraperitoneal injections of saline attentuate the antibody response to a subsequent intraperitoneal injection of antigen. Brain, Behav., Immunol. 3, 90-96. Ots. I., and Horak. P. (1996). Great tits Parus major trade health for reproduction. Proc. R. Soc. London, Ser. 263, 1443-1447. Ottaviani. E.. and Franceschi. C. (1996). The neuroimmunology of stress from invertebrates to man. Prog. Neurohiol. 48, 421-440. Owens, M. J., and Nemeroff, C. B. (1991). Physiology and pharmacology of corticotropinreleasing factor. Pharmacol. Rev. 43, 425-449. Reichlin, S. (1993). Neuroendocrine-immune interactions. N. Engl. J . Med. 329, 1246-1253. Rook, G. A. W., Hernandez-Pando, R.. and Lightman, S. L. (1994). Hormones, peripherally activated prohormones and regulation of the Thl/Th2 balance. Immunol. Today 15, 301-303. Sapolsky, R. M. (1992). Neuroendocrinology of the stress response. In “Behavioral Endocrinology” (J. B. Becker. S. M. Breedlove. and D. Crews, eds.). pp. 287-324. MIT Press, Cambridge MA.

STRESS A N D IMMUNE DEFENSE

153

Sheldon, B. C., and Verhulst, S. (1996). Ecological immunology, costly parasite defenses and trade-offs in evolutionary ecology. Trends Ecvl. Evol. 11, 3 17-321. Silverin, B. (1982). Endocrine correlates of brood size in adult pied flycatchers Ficediiln hypoleiicn. Geti. Cvmp. Etidvcrinvl. 47, 18-23. Smith, J. A.. and Weidemann, M. J . (1990). The exercise and immunity paradox: A neuroendocrinelcytokine hypothesis. Med. Sci. Res. 18, 751 -755. Stein, M., and Miller. A. H. (1993). Stress, the hypothalamic-pituitary-adrenalaxis, and immune function. Adv. Exp. Med. B i d . 335, 1-5. Thompson, S. D. (1992). Gestation and lactation in small mammals: Basal metabolic rate and limits of energy use. 1n “Mammalian Energetics: Interdisciplinary Views on Metabolism and Reproduction” (T. E. Tomasi and T. H. Horton. eds.). pp. 213-259. Comstock, Ithaca. NY. Zabel. P., Horst. H.-J.. Kreiker. C.. and Schlaak. M. (1990). Circadian rhythm of interleukin1 production by monocytes and the influence of endogenous and exogenous glucocorticoids in man. Klin. Wochmschr. 68, 1217-1221.

This Page Intentionally Left Blank

ADVANCES IN THE STUDY OF BEHAVIOK. VOL. 27

Behavioral Variability and Limits to Evolutionary Adaptation under Stress P. A. PARSONS SCHOOL OF GENETICS A N D HUMAN VARIATION LA TROBE UNIVERSITY

BUNDOORA, VICTORIA 3083 AUSTRALIA

I. INTRODUCTION A. STRESSA N D ENERGY BALANCES

Interactions between organisms and environment are central for understanding evolution. For some evolutionary biologists, the occurrence of environmental perturbations of an unpredictable nature emphasizes physical factors as the major determinants of the distribution and abundance of organisms. Even so, the effects of abiotic factors can be modulated by interactions with biotic factors (Dunson and Travis, 1991). However, the abiotic environment should be tracked more predictably than the biotic as the time scale lengthens. From a diffuse literature Hoffmann and Parsons (1991) concluded that abiotic stresses mainly of climatic origin are important in many evolutionary and ecological processes. Furthermore, inadequate nutrition is usual in free-living populations, so that animals normally struggle to exist in hostile environments. White (1993) amassed much evidence, especially in herbivores, indicating that the abundance of organisms is often determined by a shortage of protein, especially for the young. For instance, pollen digestion is important for early breeding of Darwin’s finches of the Galapagos Islands (Grant, 1996). Consequently, many organisms are born but few are expected to survive due to a combination of climatic stress interacting with and causing nutritional stress. Validity of this environmental model is suggested by the rarity of creatures that commonly die of old age in free-living populations. In any case, a reference point is provided as a boundary for comparisons with more benign environments, especially certain human populations of recent times. 155

Copyright 0 1998 hy Academic Press All rights of reproduction in any form reserved. O(~S-34.541YX$25 00

156

P. A. PARSONS

The assumption of substantial stress contrasts with approaches to the environment by many evolutionary biologists, whose ideas appear to be influenced by the apparent existence of adequate nutrition in many human populations today. But these populations may represent a benign environmental outlier when considered in a historical context, both past and future. One direct effect of stress is an increase in the expenditure of metabolic energy, implying a cost (Odum et al., 1979). As exposure to stress is the norm, there is a need for some energy to exist in any habitat. The habitats of organisms can then be expressed by an interaction between stress intensity, magnitude of environmental fluctuations, and energy from resources as a first approximation. The interaction of stress of various types causing energy costs and energy gains (provided from resources) is central in relating the distribution and abundance of organisms to energy balances (Hall et al., 1992). As energy costs increase as conditions deviate from optimal (Porter and Gates, 1969), physical conditions can limit the occurrence of organisms in particular habitats. Biotic variables, such as competition, can be incorporated into this model via an increase in energy costs, but these effects are usually second-order compared with abiotic stresses (Parsons, 1996a,b). Stress reduces the fitness of organisms by deflecting energy from processes such as maintenance and survival, reproduction, growth, and genetic adaptation (De Kruipf, 1991). Fitness therefore is inversely related to the stress level as a first approximation. Furthermore, the impact of environmental perturbations can be expressed as a stress gradient (Odum et af., 1979), on which the potential for genetic adaptation falls as stress increases to an extreme where survival is threatened. The net energy balance of a species should be relatively high in central regions of its distribution. However, the margins of distributions of at least some animal species are limited by physiological constraints. Genes allowing further adaptation may not arise or, if they do, the animals carrying them may not survive (Parsons, 1991; Hoffmann and Blows, 1994). Physiological constraints can therefore determine the location of species borders, for example, in many North American bird species and in some small mammals (Bozinovic and Rosenmann, 1989; Root, 1993). In any case, organisms living at the very extremes of a species range are rarely the healthiest and most vigorous members of that species. B. HABITATS PREFERRED Habitats in which minimum energy is expended should be preferentially occupied. Intermediate temperatures between limiting extremes should therefore be preferred in an environment where temperature gradients exist. In these regions, maximum energy should be available for behavior,

BEHAVIORAL VARIABILITY AND EVOLUTIONARY ADAPTATION

157

growth, and reproduction (Huey, 1991), and resistance to stress should be higher than elsewhere (Klieber, 1961; Arking et al., 1988; Zotin, 1990). Of course, these are the circumstances in which high population sizes may lead to maximum competition, but in all but the most abiotically stable of habitats such abiotic effects should be transient. Insects living in habitats with steep abiotic gradients can be useful for habitat preference studies. For example, in temperate zone rain forests, adults of an Australian Drosophila species, D . inornata, tend to rest on the fronds of tree ferns in the 15-20" range, with a mean of 17.7 ?I 2.0"C, following behavioral responses to the microenvironment. Consequently, flies attempt to select microhabitats as close as possible to optimal for temperature/humidity conditions, where the energy cost from the physical environment would be minimized (Parsons, 1993a). Similarly, Jones et al. (1987) found behavioral flexibility for thermal niche preference in D . melanoguster, whereby the effects of temperature extremes were ameliorated by habitat selection. Survival is thereby enhanced, since animals in early developmental stages are intolerant of energetically costly extreme conditions. In this context, the microenvironment (soil moisture and air temperature) experienced by a larva during wandering and pupation is important for pupal survivorship (Rodriguez et al., 1992). Furthermore, larvae from populations from dry habitats in Tunisia pupate closer to food than those from wet habitats (Rodriguez et al., 1991). Adult food-searching range is dependent on temperature. In D . melanogaster, the range searched is substantially smaller at the stressful temperature of 30°C than at 25°C (Good, 1993), presumably because of the high energetic cost of surviving at 30°C. Genetic shifts in preferred temperatures occurred in a gradient, when flies were reared at 25, 27, and 30°C for 15 generations. In addition, Yamamoto (1994) found temperature preferences in natural populations of D. immigruns and D . virilis to be heritable. Ye et al. (1994) found substantial genetic differences between strains of D. melanogaster from six local populations for adult starvation resistance, presumably as adaptations to cope with specific microhabitats. The six strains ranked for search behavior in parallel with starvation resistance, such that resistant strains searched over more extended ranges than did sensitive strains. In the two-spotted spider mite, Tetranychus urticae, there is genetic variation in aerial dispersal behavior, associated with resistances to the environmental stresses of desiccation and starvation (Li and Margolies, 1994). These few examples suggest that behavioral flexibility can reduce the energy costs of physical stresses. Furthermore, genetic changes can occur, enabling adaptation to varying habitats, assessed abiotically. Hence, ecobehavioral genetic adaptation can evolve under varying stress levels in free-

158

P. A. PARSONS

living populations, which should ameliorate the direct effects of stress in an evolutionary sense. Turning now to the physiological state of insects, in D. melanogaster, starved flies are less discriminatory in responding to alternative resources in orchards than when unstressed (Hoffmann and Turelli, 1985). Resource selection is therefore most efficient when organisms do not simultaneously need to cope with extreme stress. However, nutritional stress appears to be the norm in natural populations of Drosophifa. For example, within a French population of D. melanogaster, high reproductive potential is not normally expressed under natural conditions, because of substantial and variable fluctuations in food availability (BoulCtreau-Merle et af., 1987). Furthermore, in a widespread montane North American butterfly, Speyeria mormonia, experiments with varying feeding regimes show that survival claims resources as a priority over reproduction (Boggs, 1994). In summary, more emphasis is needed on the study of habitat preferences under a range of realistic abiotic environmental conditions. In the remainder of this chapter, some background is provided for the consideration of limits to adaptation under predominantly rigorous conditions.

11. ENERGY LIMITSTO ADAPTATION

A. NONSEXUAL BEHAVIOR

Energetically expensive behaviors are common, for example, web construction in spiders, and insect and avian flight. However, the amount of energy that can be assimilated from food is finite (Weiner, 1992). Oxygen consumption can increase to meet a higher demand for ATP production, but the maximum possible oxygen consumption (Bennett, 1991) sets a limit to total behavioral activity. Consequently, any superimposed environmental stress would be deleterious by increasing respiration and hence stress sensitivity. For example, metabolic rate and whole-body thermal conductance increased in polar bears exposed to oil pollution, which increased mortality during the stress of a hard winter (Hurst et af., 1991). In D. melanogaster, high-metabolicrate “shaker” mutants show high levels of behavioral activity, and are very sensitive to environmental stresses, including high temperature, desiccation, and exposure to an unsaturated aldehyde acrolein (Barros et af., 1991; Parsons, 1992). In fasting rats, minimum heat production occurs in the thermoneutral zone and increases at higher and especially lower temperatures in association with reduced stress resistance (Klieber, 1961; Blaxter, 1989). In the social Damara mole rat, Cryptomys damarensis, body tempera-

BEHAVIORAL VARIABILITY AND EVOLUTIONARY ADAPTATION

159

ture remains stable at ambient temperatures from 7 to 30°C, so that there is a substantial metabolic cost at extremes; at 7°C the metabolic rate is more than four times higher than in the thermoneutral zone (Lovegrove, 1986). In C. damarensis, which is from arid regions in southern Africa, the rate of metabolism is much lower than for comparable species from wetter regions. This is an energy-saving device enabling adaptation to aridity stress. In addition, cooperative searching and food sharing can reduce energy demands further (Lovegrove, 1986). Such adaptations can occur seasonally, or transiently during periods of bad weather. For instance, in the house martin, Delichon urbica, energy is saved during transient problems in finding food in the breeding season, by a complex of physiological and behavioral adaptations, including low basal metabolic rate, low thermal conductance, clustering behavior, high tolerance of the young to periods of low food supply, and the ability to become torpid (Prinzinger and Siedle, 1988). In subterranean blind mole rats of the Spalax ehrenbergi superspecies complex, aggression tendency and basal metabolic rate decrease geographically across Israel as the climatic stresses of temperature, and especially aridity, increase (Nevo, 1991). These responses would minimize water and energy expenditure, and are adaptations to counter extreme stress. Furthermore, in isolates from the environmentally harsh Sahara Desert of northern Egypt, mole rats were totally pacifist, presumably as an adaptation to an environment that is even more extreme than that in Israel (Nevo et al., 1992). Consequently, a behavioral-ecophysiological response has evolved based on selection against aggression. This response has enabled the spread of S. ehrenbergi into extremely arid environments (Ganem and Nevo, 1996). In summary, energetically costly behaviors occur frequently, and can determine limits of adaptation of organisms. Under these circumstances, any additional stress would be rapidly restrictive. Adaptations to high stress levels include reductions in resting metabolic rate, social behavior patterns that conserve energy, and pacifist behavior.

B. SEXUAL BEHAVIOR AND SEXUAL SELECTION The energy used in calling can exceed resting levels by up to 20 times in frogs (Ryan, 1988), suggesting that the mating process can be energetically expensive. In some frog species, mating success of males increases with increasing chorus tenure (Pough, 1989). As the time a male spends in the breeding chorus is important in determining mating success, fitness assessed in this way is likely to be related to available metabolic energy. In damsel flies, Calopteryx maculata, territorial contests favor males with the greatest energy reserves, measured by fat content (Marden and Waage, 1990). In great tits, Parus major, and male pied flycatchers, Ficedula hypoleuca, in

160

P. A. PARSONS

breeding condition, resting metabolic rate is positively correlated with dominance rank (Roskaft et al., 1986). In the fish Betta splendens, winners and dominant individuals in a hierarchy consume more energy per unit time than losers and submissives (Haller and Wittenberger, 1988). In pupfish, Cyprinodon pecosensis, critical swimming speed is higher in territorial than in nonterritorial males, indicating a positive correlation of vigor with social status (Kodric-Brown and Nicoletto, 1993). In summary, these and other examples imply that fitness in mating is normally correlated with energy consumption. Similarly, daily energy expenditure increased significantly with increasing display rate and time spent in the lek in the male sage grouse, Centrocercus urophasianus (Vehrencamp et al., 1989). Daily energy expenditure for the most vigorously displaying males was two times higher than for a nondisplaying male, and four times higher than the basal metabolic rate. The increased levels of lek attendance and display levels appear fueled by increased quantity or quality of food, since the more actively displaying males can forage further from the lek. Furthermore, the abiotic environment is relevant, since metabolic expenditure increases as temperature falls; this would be ultimately restrictive. The inadequate resources available for free-living organisms should be used efficiently. Accordingly, resources are normally channeled to only some of those seeking to use them, so that a dominant few survive; the remainder in a population are vulnerable. This can be achieved by territorial and social behaviors, largely restricting resources to the dominant few, as demonstrated in passerine bird species by Moller (1991). Analogously, polygyny can replace monogamy in traditional human societies, when there are substantial fitness differences among men, following pathogen stress. The minority of resistant men are dominant in mating, because they are more skilled in promoting polygyny; these skills include hunting, winning disputes, and resource acquisition (Low, 1990). Considering stress from parasites, in the lizard, Sceloporus occidentalis, Schall and Sarni (1987) found that the time males spend in social behaviors was reduced when infected with the malarial parasite, Plasmodium mexicanum, and furthermore, infected males perch more often in shade. Hence, the energy cost from parasites reduces social behaviors, and stressful microhabitats are avoided. In feral rock doves, Columbia liviu, lice reduced feather and host body mass, and increased thermal conductance and metabolic rate, indicating an energy cost. This is exacerbated in a deteriorating abiotic environment during winter (Booth et ul., 1993). Finally, in the colonially nesting cliff swallow, Hirundo pyrrhonutu, mark-recapture experiments over an 8-year period showed that the annual survivorship of birds

BEHAVIORAL V A R I A B I L I T YAND EVOLUTIONARY ADAPTATION

161

parasitized with cimicid bugs, fleas, and chewing lice was .38, compared with .57 for fumigated, nonparasitized birds (Brown et al., 1995). Turning to reproduction, de Lope et al. (1993) found that the ectoparasitic house martin bug, Oeciacius hirundinis, had larger negative effects on the reproduction of its host, Delichon urbica, when nutritional conditions were poor during the second compared with the first clutch in the season. In red jungle fowl, Gallus gallus, chicks infected with parasites grew more slowly than uninfected controls (Zuk et al., 1990). Since this effect was most pronounced for secondary sexual traits, there is a channeling of resources into the normal growth of nonornamental traits under parasite stress. Generally, birds and fish with high parasite loads engage in less courtship display and obtain fewer mates than those with lower loads (Hamilton and Zuk, 1982; Kennedy et al., 1987; Clayton, 1990). Therefore, the energy cost of parasites in combination with abiotic stresses can preclude the full development of ornamental traits, and reduce mating and fitness generally. Furthermore, if there are energy restrictions from nutritional stress, an ornament can rapidly regress, as found for the nuptial crest of male newts of the genus Triturus (see Halliday, 1978). In addition, sexual ornamentation in some birds is restricted to the breeding season, indicating an excessive cost in less favorable abiotic environments. Sexual selection can be constrained by costs associated with mate choice, when interacting with unfavorable abiotic circumstances. The same should apply to biotic effects, although less obviously. Especially for predation, theoretical models (e.g., Pomiankowski, 1987) predict that female preference should decrease with increasing costs of mate choice. Accordingly, individual females should modify their choice behavior to minimize this risk. In terms of energy costs, this means that females should become less discriminatory when given a choice among potential mates at times when the predation risk is increased. For example, male pipefish, Syngnathus typhle, exposed to the cod, Gadus morhua, as a predator, copulated infrequently and indiscriminately, whereas control males copulated more often with large than with small females (Berglund, 1993); and predation from the cichlid fish, Crenicichla alta, reduced female preference in the guppy, Poecilia reticulata (Godin and Briggs, 1996). Even so, Magnhagen (1991) has cautioned that only in a very few cases has it been shown that individuals actively change their mating behavior according to predation risk. Additional studies would be useful. Overall, the stressful scenario in nature should reduce the tendency for traits involved in the sexual selection process to become progressively more extreme, thereby limiting the runaway process in which the sexual ornament is continuously exaggerated (Fisher, 1930; Lande, 1980). Therefore, a tradeoff occurs, since the energy cost of the development and maintenance of

162

P. A . PARSONS

ornaments of increasing size is countered by the cost of stress (Parsons, 1995a). The most extreme ornaments should therefore occur when the stress level is relatively low in the background environment, and the size of ornaments should fall with increasing stress. Species with morphologically complex sexual ornaments should be vulnerable during periods of environmental stress (McLain, 1993), such as extinction events. For instance, McLain et al. (1995) record that sexually dimorphic bird species are more vulnerable to extinction than are monomorphic species, following their introductions into the oceanic islands of Oahu and Tahiti. Certainly, in the initial occupation of new adaptive zones, sexually dimorphic species would appear to be at a disadvantage compared with more generalist species, because of the energy costs in developing and maintaining secondary sexual characters and in sexual display, which in total may approach the maximum limit of available energy (Moller, 1994a). In many bird species the cost of secondary sexual ornaments can be reduced by an investment in physiological and anatomical adaptations. These adaptations coevolve with the secondary sexual characters, thereby permitting levels of sexual display considerably higher than those observed in their absence (Moller, 1996). This indicates strong selection at energetic limits, implying extreme vulnerability of birds to any increase in stress. Finally, and in accord with the previous considerations, a recent theoretical analysis concludes that sexual selection in a changing environment enhances population extinction by increasing selection intensities on a male trait (Tanaka, 1996).

BOUNDARIES C. SPECIES In general, the available data on species boundaries are restricted to successful species, and represent end points of adaptive change during the speciation process. Accordingly, it is appropriate to consider briefly the boundaries between closely related species, especially those that are mainly sympatric. The predominantly sympatric sibling species, D. rnelanogaster and D. sirnulans, are distinguishable physiologically, based on differing resistances to environmental extremes, especially high and low temperatures, and toxic levels of ethanol and acetic acid. Furthermore, within these extremes, ecobehavioral differences indicate very different microhabitats for the species in nature, especially for larvae. Even so, these species of subgenus Sophophora are ecobehaviorally and physiologically very similar, and contrast substantially with another widespread species, D. irnrnigrans of subgenus Drosophila (Ehrman and Parsons, 1981).

BEHAVIORAL VARIABILITY AND EVOLUTIONARY ADAPTATION

163

Similarly, marine sympatric invertebrate species show distinct habitat preferences defined by depth, salinity, exposure, or preference for host substrates (Knowlton, 1993). In carnivorous stoneflies (Plecoptera), eggs are laid at similar times, but newly hatched larvae rarely occur together in the same habitat, defined by hatching temperature, incubation temperature, and time of hatching, thereby largely eliminating competition btween the species (Elliott, 1995). In two closely related hymenopteran parasitoid species of Drosophila, there is odor-mediated avoidance of competition, as Leptopilina heterotoma can recognize patches on stinkhorns where groups of L. clavipes females occur (Janssen et al., 1995). These examples indicate that in responding to abiotic stresses, physiological and ecobehavioral traits are important in adapting to habitats, and are likely to be important targets of selection in evolutionary shifts underlying the speciation process. Consequently, the resources of the environment become utilized efficiently, as the sibling species do not directly compete with each other. Divergence at the ecobehavioral and physiological levels is therefore primary to morphological divergence. Only to the extent that it has a functional role can morphology be regarded as a direct target of selection (Bonner, 1988). Thermal constraints on the time and energy budgets of lizards have been investigated extensively, and upper and lower critical thermal limits can be determined (Adolph and Porter, 1993). Canyon lizards, Sceloporus merriami, have a characteristic body temperature of 32.2”C, which is lower than that of other North American desert iguanids. Under this thermal environment, individual activities (movement rate, feeding strikes, and social displays) are restricted to a 2-hr period beginning around local sunrise and to a brief period in the late afternoon. When the average temperature was around 32.2”C, maximum activity and maximum use of microenvironments occurred. However, as the temperature deviated from 32.2”C, the use of microenvironments became more constrained (Grant and Dunham, 1988; Adolph and Porter, 1993). Presumably, energy costs would increase in parallel with divergence from 32.2”C. As thermal regimes deviate from optimal, lizard activity becomes more restrictive, which is a behavior that influences home range size, population density, fecundity, social history, and ultimately survival. From populations across a range of elevations, complex relationships have been established between biophysical constraints and fitness mediated through daily time budgets and seasonal energy-mass budgets. These relationships underlie life-history variation and the adaptation of organisms to specific environments (Dunham et al., 1989). Adolph and Porter (1993) argue for the importance of activity as a connecting link between thermal environment and lizard life histories. The

164

P. A. PARSONS

implication is that lizard activity is a target of selection, which is reasonable because of its energy cost. When the energy available for activity becomes so restrictive that there is no discretionary energy for reproduction, growth, or storage, species boundaries are likely to be located. This is in accord with the physiological constraints that appear important at the margins of distributions of an increasing number of animal taxa (see Section 1,A).

A N D THE SURVIVAL OF VARIANTS 111. VARIABILITY

Under environments that are demonstrably extreme, heterozygotes tend to be favored, especially for polymorphisms in natural populations. Even under less extreme conditions characteristic of the laboratory, the level of heterozygosity of individual organisms in populations tends in some cases to correlate with measures of performance or fitness, in particular growth rate and developmental stability. Enzyme loci influencing metabolism and contributing to the amount of energy available for development and growth show the most significant positive associations with heterozygosity (Mitton, 1993). Consequently, heterozygotes should be differentially favored in growth and reproduction as stress increases, and when resources become limited (Parsons, 1996b). In the white-tailed deer, Odocoileus virginianus, antler size, body mass, fat levels, and other dimensions were found to be correlated with heterozygosity, dominance status, and reproductive success by Scribner et al. (1989), who emphasized the importance of metabolic efficiency of the heterozygotes. In bighorn sheep, Ovis canadensis, horn size largely determines breeding superiority, as large horns give access to estrous ewes. In addition, such rams have superior foraging ability, energy efficiency, and disease resistance (Hogg, 1987). In the seventh year of life, which is around the time of onset of breeding, 21% of variation in horn volume can be explained by an association with heterozygosity. In contrast, the horns of young rams show little variation in size that is attributable to genetic factors (Fitzsimmons et al., 1995). Therefore, when energy demands from the development and maintenance of horns and from the mating process itself are high, heterozygote advantage is maximal. In any case, mating is an expensive process energetically, and fitness assessed by mating success can often be related to available metabolic energy (see Section 11,B). Consequently, heterozygotes should be favored during mating, as documented in a number of species, especially insects and fish (Thornhill and Gangestad, 1993). Furthermore, Rolan-Alvarez et al. (1995) found a positive association between heterozygosity and sexual fitness for males in populations of the marine snail, Littorina mariae.

B E H A V I O R A L V A R I A B I L I T Y AND EVOLUTIONARY ADAPTATION

165

In the context of this discussion, competition can be used as an example of stress under laboratory conditions. In offspring from a diallel cross involving three inbred strains of mice, several traits were studied in a normal cage, and a smaller cage with enhanced crowding. In the normal cages, 14% of inseminated females did not produce offspring; 29.4% did not in the smaller cages, suggesting that crowding reduced reproductive fitness. Additive genetic variability increased under crowding stress, especially for preimplantation mortality, litter size, and relative adrenal weight. For preimplantation mortality and litter size, nonadditive effects and heterozygote advantage increased under stress (Belyaev and Borodin, 1982). Genetic variability for behavioral traits can therefore be high under stress (Parsons, 1988). These and many other examples show that under highly stressed situations, especially in free-living populations, genetic variability is not normally expected to be restrictive. In addition to heterozygote advantage under stress, there is a substantial body of data indicating increased mutation, recombination, developmental variability, and phenotypic variability as stresses approach levels where extinctions become a real possibility (Parsons, 1987; Hoffmann and Parsons, 1991). For novel variants, the issue then turns to the conditions under which their survival and reproduction is likely. Following Fisher (1930), the chances of survival of a novel variant should be inversely related to the magnitude of its phenotypic change, or in the context of this discussion, the energy cost of the change. In addition, survival should be inversely related to the magnitude of energy cost of existing in a variable environment, so that the more extreme the environment, the smaller the change that can be accommodated by organisms for their survival (Parsons, 1996b). In summary, the level of genetic variability is unlikely to be restrictive for adaptation, but the ecological conditions determining the survival of the variants may be. It is in this light that the issue of extending the limits of adaptation in populations will now be considered, for a continuum of environments, from extreme at species borders to benign, where learning is possible. The survival of variants should increase as conditions become less severe, even though variants can apparently appear under all conditions.

I v . EXTENDING THE LIMITS OF ADAPTATION A. ABIOTIC STRESSES A N D RESOURCES Speciation necessarily involves a shift in the limits of adaptation of established species. Therefore, the process of speciation, approached ecologi-

166

P. A. PARSONS

cally, implies energy costs (Van Valen, 1976a,b). Shifts in limits can in principle involve changes in resistance to abiotic stresses, changes in resource availability and usage, or a combination of these variables. Commencing with abiotic stress, thermophilia occurs in some desert ant genera, enabling successful foraging for arthropods that have succumbed to extreme heat. For instance, the Saharan silver ant, Cataglyphis bombycina, scavenges for the corpses of insects and other arthropods that have SUCcumbed to the heat stress of their desert environment in a small thermal window with a maximum width of 7°C (46.5-53.6"C). The boundaries of the window are underlain by predatory pressure exerted by a desert lizard at the lower limit and heat stress at the upper limit (Wehner et al., 1992). Parallel situations occur in the Australian ant, Melophorus bagoti (Christian and Morton, 1992), and in the burrowing spiders Seothyra in Namib desert dunes (Lubin and Henschel, 1990). Finally, the diamond above, Geopelia cuneara, an inhabitant of the arid savannahs and semideserts of Australia, is extremely heat tolerant, and consequently activity occurs throughout the day under dry and hot conditions when potential predators and food competitors are reduced (Schleucher, 1993). These are examples of extreme abiotic stress, where predators and competitors are likely to be absent, enabling the occupation of extreme habitats. Ultimately, as found in the lizard, Scleroporus merriami, at extreme temperatures, the energy for activity becomes so restricted that there is no discretionary energy for reproduction, growth, or storage, and species boundaries occur (Adolph and Porter, 1993). Heat shock protein synthesis may occur in association with thermotolerance, as found in the ant Cataglyphis (Gehring and Wehner, 1995). However, the formation of heat shock proteins is likely to have a metabolic cost, thereby reducing fitness (Krebs and Loeschke, 1994). Consequently, assuming that species boundaries are regions of energy restriction, it seems difficult to envisage much widening of windows of opportunity for direct abiotic extremes of climatic origin, especially as the survival of novel variants would be unlikely. Turning to resources, innovation can involve ecobehavioral traits in shifting to alternatives at stressful times (Parsons, 1993b). Examples include: (1) the evolution of specialization of D. sechellia onto a single resource, Morinda citrifolia, from a stressful window of opportunity from D. simulans, which finds the resource toxic (R'Kha et al., 1991); (2) the evolution of host races of a stem-galling tephritid, Eurosta solidaginis, on two goldenrod host species assisted by a 10- to 14-day difference in emergence times on the two hosts (Craig et al., 1993); and (3) the evolution of races of the tephritid fly, Rhagoletis pomonella, from its native host hawthorn to introduced fruits maturing at different times (Feder et al., 1988).

BEHAVIORAL VARIABILITY AND EVOLUTIONARY ADAPTATION

167

Therefore, resource heterogeneity can underlie divergence, especially if associated with the simultaneous need to adapt to some abiotic stress. The summed environmental change must be intense enough to cause disruptive selection for sufficiently long that any incipient divergence can become established, and consequently to have the potential to lead to isolation. A possible example comes from the intertidal snail, Littorina saxatilis, where there is assortative mating leading to incipient reproductive isolation associated with habitat selection by two morphs; one of these occurs in the uppershore barnacle belt and the other in the lower-shore mussel belt, indicating physiological and behavioral adaptations by the morphs to the two differing environments (Johannesson etal., 1995). In contrast, in the Galapagos finch, Geospiza conirostris, partial isolation of a population based on resource heterogeneity occurred following a drought, but prolonged divergence was prevented by extreme fluctuations in the abiotic environment (Grant and Grant, 1989).

B. RESOURCE POLYMORPHISMS Genetically based resource and habitat polymorphisms permit the occupation of more than one niche within a species, and can underlie divergence ( West-Eberhard, 1986; Stanhope et al., 1992). For instance, sympatric populations of the tropical sponge-dwelling coral-reef shrimp, Synalpheus brooksi, occupy two alternative host species of sponge, and in laboratory situations tend to choose native sponge species. This promotes assortative mating and hence divergence, as shown by significant host-associated genetic divergence of shrimp in two of three reefs based on proteinelectrophoretic variation (Duffy, 1996). In the Arctic charr, Salvelinus alpinus, there are benthivorous, planktivorous, and/or piscivorous forms in lakes in Iceland, which show substantial morphological, developmental, and behavioral specialization for discrete resource categories. The behavioral differences break down when food is artificially superabundant, occurring only in the nutritionally restricted environments of free-living populations (Skulason et al., 1993). Under these latter conditions, energy returns to charr appear to be maximized by genetic divergence among morphs, enabling the efficient exploitation of differing resource categories. Novel variants promoting such divergence would be favored on grounds of energetic efficiency. Schluter and McPhail (1993) record multiple examples of fish in lowdiversity postglacial lakes, where there are sympatric species involving limnetic and benthic forms. The limnetic forms, which exploit plankton in open water, are typically smaller, with a narrower mouth and longer, more numerous gill rakers than the benthic forms, which consume larger prey;

168

P. A. PARSONS

the morphological differences are adaptations to differing food requirements. The benthic-limrtetic split could therefore be a predictable first step in the diversification of many fish taxa (Schluter and McPhail, 1993), and this split would be favored by energy efficiency in resource utilization. While Skulason and Smith (1995) argue that resource polymorphism has been underestimated as an evolutionary force leading to divergence, this could be precluded by abiotic instability, as noted in the finch G. conirosfris. More generally, during a 2-year California drought in 1976-1977, pressures on development time intensified in colonies of the specialist insect herbivore, the butterfly Euphdryas editha, because host plants senesced rapidly (Ehrlich et al., 1980). Conversely, continuous rainfall can retard postdiapause larval development so that adult flight is delayed beyond plant senescence (Dobkin et al., 1987). Such climatic perturbations would appear sufficient to swamp the selection for energy-use efficiency based on resource polymorphisms that can occur under more stable and less stressful abiotic conditions. The utilization of heterogeneous resources therefore is likely to be the most efficient when organisms d o not simultaneously need to cope with the energy costs of extreme stresses (see also Section 1,B). C. LEARNING The finch Pinaroloxias inornata, of Cocos Island, Costa Rica, has extremely generalist feeding habits, spanning those of several families of birds on the mainland, encompassing insects, Crustacea, seeds, fruits, nectar from many flower species, and perhaps lizards. In contrast, individuals feed as specialists year-round, often using just one of the many feeding techniques and resources observed at the population level. Apparently, these specializations are transmitted at least partly culturally, from the observation of other individuals. Hence, these tropical birds have developed learning ability, permitting the exploitation of heterogeneous resources (Werner and Sherry, 1987), which implies high energy-use efficiency. The ecological situation on Cocos Island appears permissive of this situation, as it is an aseasonal environment with very few competing species, and with high availability, variety, and predictability of food resources. Under these circumstances of relatively low energy constraints from the environment, specialization has apparently occurred. Ultimately, such behavioral specialization could be assimilated genetically, perhaps following varying biochemical demands made on finches from differing food categories. For instance, differences in feeding behavior in the crustacean Gammarus palustris are associated with genetic variation in the properties of a digestive enzyme (Guarna and Borowsky, 1993).

B E H A V I O R A L VARIABILITY AND EVOLUTIONARY ADAPTATION

169

Another example of learning comes from bluegill sunfish, Lepomis macrochirus, in North American freshwater lakes, where learning assisted the behavioral modification needed to search efficiently in vegetated and open-water habitats (Ehlinger, 1990). Habitat-specific foraging efficiency occurred, thereby increasing the energetic effectiveness of resource exploitation. Therefore, behaviors governing resource use may be influenced by the previous experience of individuals. In insects, prior exposure to a particular resource can enhance a female’s tendency to oviposit on that type of resource. For instance, in the true fruit fly, Rhagoletis pomonella, the propensity to accept a particular fruit prior to the deposition of an egg can be modified by previous ovipositional experience (Prokopy and Papaj, 1988). Learning should assist in exploiting windows of opportunity presented by introduced fruit species, which would be enhanced if mating occurred at the resource. Learning may therefore assist in adaptation to new hosts from the original host. Hence, following Baldwin (1896) learning may be a factor in switches into novel habitats, thereby assisting in the integration of genetic components of behavior into the gene pool. As learning eases the process of genetic change (Anderson, 1995), the energy costs in the occupation of novel habitats would be reduced. For instance, fifteen-spined sticklebacks, Spinachia spinachia, attack Gamrnarus and Artemia more efficiently as a result of experience. By decreasing handling time, learning increased the profitability of specific prey, expressed in terms of energy expended per given time period (Croy and Hughes, 1991).

V. FROM STRESS-RESISTANCE GENOTYPES TO A CONNECTED METABOLISM

A. STRESS-RESISTANCE GENOTYPES Koehn and Bayne (1989) argue that high stress resistance is associated with the efficient use of metabolic resources for growth and reproduction, especially when resources are limited. Since stress-resistance phenotypes tend to have a low metabolic rate (Hoffmann and Parsons, 1991), a low maintenance requirement is implied. Consequently, growth should be supportable over a wide range of conditions. In particular, the association between metabolic efficiency and stress resistance suggests that genes for stress resistance should be favored during the metabolically costly process of the development and maintenance of sexual ornaments and mating itself (Parsons, 1995a). During mating, the preferred male trait may reflect the underlying genetic quality of the male, so that females mating with these males gain additional

170

P. A. PARSONS

advantages for themselves or their offspring outside of mating (Thornhill and Alcock, 1983). Such advantages are conventionally regarded as being under the control of “good genes,” which enhance fitness both during the mating process, conferring direct benefits to females, or by producing offspring with superior fitness (Moore, 1994). Wedekind (1994) argued that sexual selection for stress-resistance genes is important in improving the survival chances of offspring. In other words, mate preferences would be most efficient if coupled with resistance genes in parents and offspring. Following Hamilton and Zuk (1982), this conclusion comes from an assessment of parasite-driven sexual selection. In the pheasant, Phasianus colchicus, male spur length correlates with male viability, female mate choice, and offspring survival (von Schantz et al., 1996). Genetic analyses show that the major histocompatibility complex genotype is associated with variation in male spur length and male viability. Von Schantz et al. (1996) conclude that these data directly support the “good genes” hypothesis (Hamilton and Zuk, 1982) that females discriminate among males based on secondary sexual characters, and so pass on genes for disease resistance that improve offspring fitness. In any case, a premium on stress resistance and hence metabolic efficiency conferring overall fitness is expected, assuming that populations are normally exposed to high levels of stress (Parsons, 1996b, 1997a). Furthermore, because “good genes” reflect fitness under these environmental conditions, it should be possible to incorporate other fitness traits into this scheme. For instance, in an African cockroach, Nauphoeta cinerea, females have offspring that develop relatively quickly following mating with the most attractive males (Moore, 1994). This suggests that the choosing female prefers individuals carrying “good genes,” which also underlie rapid development. Additional examples cited by Moore (1994) include heritable variation in plumage as an indicator of viability in male great tits, Parus major (Norris, 1993), and improved growth and survival of offspring of peacocks, Pavo cristatus, with more elaborate trains (Petrie, 1994). In the damselfly, Zschnura graellsii, Corder0 (1995) found that the best predictor of male lifetime mating success was mature life-span. In barn swallows, Hirundo rustica, Mdler (1994b) found that offspring longevity is positively related to that of their fathers, and to the ornament size of the male parent. Considering aging, assuming that a long life-span and rapid development depend on metabolically efficient stress-resistance genes, individuals having high inherited stress resistance should develop fastest and live longest (Parsons, 1996b,c). Accordingly, ornament size, mating success, longevity, and development time can perhaps be viewed as a coordinated suite of characters assuming the stressful environments of free-living populations. If a major target of selection of stress is at the level of energy carriers, “good

BEHAVIORAL VARIABILITY AND EVOLUTIONARY ADAPTATION

171

genes” therefore should be stress-resistance genes, and these should be increasingly important for ensuring fitness as limits to adaptation are approached. O n a cautionary note, the paucity of empirical observations for this predicted relationship presumably relates to the point that studies carried out under relatively benign laboratory conditions are unlikely to be efficient in revealing such associations, because selection for stress resistance is necessarily less intense than in free-living populations. On the other hand, irrespective of the background environment, associations of development time and life-span with mating success and the size of sexual ornaments should be the most readily detectable, because mating and the development and maintenance of sexual ornaments are normally energetically expensive processes. Lifetime reproductive success has not been considered to any extent in this article, and in any case it is usually strongly correlated with longevity. However, direct extrapolations from laboratory to natural populations cannot be assumed. For instance, in D. melanogaster, substantial and variable deficiencies in food availability under natural conditions preclude the expression of reproductive potential (BoulCtreau-Merle et af., 1987). In any case, adults of English Drosophila populations had a mean life expectancy of 1.3 to 6.2 days, which is at least an order of magnitude less than survival under equable laboratory conditions (Rosewell and Shorrocks, 1987). In a recent review of genetic variation and aging, Curtsinger et al. (1995) argued for a model where old and young fitness components are correlated, which is in accord with a prediction from the stress theory of aging (Parsons, 1993b). Accordingly, survival at any age should be a predictor of lifetime reproductive success in free-living populations (see Parsons, 1997b, where this conclusion is considered in the light of various evolutionary theories of aging).

B. FITNESS A N D METABOLIC EFFICIENCY In Section 111, it was noted that heterozygosity levels tend to be correlated with fitness during the mating process, especially for enzyme loci controlling metabolism and hence energy availability. For instance, in bighorn sheep, Hogg (1987) argued that this association reflects a “good genes” strategy favoring heterozygotes at an energetically demanding time. Extending to other fitness traits, in particular development rate but also life-span, substantial evidence suggests that heterozygosity tends to be associated with high fitness in a wide range of taxa, especially under stressful circumstances (Mitton, 1993; Parsons, 1996b,c, 1997a).

172

P. A . PARSONS

Consequently, two approaches imply parallel associations for a range of fitness traits. The first approach commences at the whole organism level and leads to genes for stress resistance for promoting fitness, while the second approach commences at the gene level using electrophoretic variants, and leads to generalized heterozygous advantage for promoting fitness under stress. Although these approaches have developed largely independently, they can be linked by a requirement for metabolic efficiency in the face of the stress to which free-living populations are normally exposed. Detailed gene location studies based on natural populations appear necessary for additional elaborations. The generalized advantage of heterozygotes under stress does, however, suggest that many interacting loci may be involved in promoting metabolic efficiency, so that it appears more appropriate to talk of “good genotypes” than good genes. This does not, of course, preclude the involvement of some major genes, such as the more anodal allozymehozyme at the phosphoglucose isomerase locus, which is favored in a range of stressful situations, including high temperature, high salinity, anoxia, and desiccation in natural populations of a wide range of taxa (Riddoch, 1993). In summary, a wide-ranging literature suggests that stress resistance and metabolic efficiency are associated for a range of fitness measures (Table I). The ranking 1 to 10 in Table I represents a continuum of organizational levels ranging from the essentially molecular (1) to the organismic. The items of main concern in this paper are categories 7 and 8, and correlations with other life-history characteristics, especially 6 and 9, are noted under 10.

TABLE 1 ASSOCIATIONS PREDICTED I N STRESSED FREE-LIVING POPULATIONS“ ~~~~~~~

1. Stress-resistance genes

2. High (electrophoretic) heterozygosity 3. High vitality, vigor. and resilience 4. High homeostasis in response to external stresses 5. Low fluctuating asymmetry 6. Rapid development 7. High male mating success 8. Extremes of sexual ornaments 9. Long life span 10. Positive correlations among fitness traits a See Parsons (199%. 1966b.c. 1997a) for detailed discussions from which this table was derived.

BEHAVIORAL V A R I A B I L I T YAND EVOLUTIONARY ADAPTATION

173

Items 3-5 are various measures of homeostasis, from the morphological to physiological levels. For instance, survival to an old age is associated with high vitality, vigor, and resilience (3), and high homeostasis in response to external stresses (4). Fluctuating asymmetry (FA) measures the degree to which an individual can control development under given environmental and genetic conditions ( 5 ) , and is a measure of individual phenotypic quality or fitness (Zakharov, 1989; Parsons, 1990; Markow, 1994; Polak, 1997; Moller, this volume). One manifestation of energy dissipation is increased FA (Mitton, 1993), for instance the FA of antlers of reindeer is positively correlated with parasite intensity (Folstad ef al., 1996). Accordingly, low F A should occur in organisms in which metabolic efficiency or fitness is highest (5). Similarly, low FA should be associated with genes for stress resistance, which implies that FA should be heritable to some extent, as noted from associations between development rate and life-span (Parsons, 1996d), but more generally from a meta-analysis of 29 studies of 13 species, which revealed a mean heritability of FA of 0.27 (Moller and Thornhill, 1997). Furthermore, because the level of heterozygosity of organisms tends to correlate with performance or fitness, correlations with high FA should be maximal in heterozygotes. There are now sufficient data sets to infer that rapid development, a long life-span, success in mating, and extremes of sexual ornaments tend to be associated with low FA, and this tends to be clearest in heterozygotes. However, there is a need to devise laboratorybased experiments to model the environments of free-living populations to obtain additional empirical data to explore these apparent and rather tentative generalizations more directly. Finally, the extreme stress scenario, which is the basic assumption underlying this paper, gains support from Kauffman (1993) who argues that the normal situation faced by organisms is an extremely perturbed world. Under these circumstances, he argues that a connected metabolism is important for the facilitation of adaptive change in response to environmental challenges. There is a convergence with the model in this paper, as the associations in Table I are underlain by selection by environmental stress, which targets energy carriers in free-living populations. In any case, an energetic approach to fitness has appeared previously. For instance, Van Valen (1976b) argued that energy underlies fitness, which can then be viewed as the rate at which resources, exceeding those needed for growth and maintenance, are available for reproduction in the broadest sense (Brown et al., 1993). In the context of this paper, mating, the development and maintenance of sexual ornaments, and various nonsexual behaviors can extract substantial energy from resources, and so may be critical in determining limits to adaptation in extreme environments.

174

P. A. PARSONS

V1. SUMMARY Energy expenditure is a prerequisite for organisms to exist in any habitat, as exposure to biotic and especially abiotic stress is the norm in free-living populations. Therefore, the distribution and abundance of organisms can be related to energy balances, derived from the costs of various stresses interacting with gains from resources. Consequently, the behavioral selection of preferred habitats imposing low energy costs is adaptive. On the other hand, limits to adaptation occur when available energy becomes totally restrictive. Therefore, energetically costly behaviors, especially those involving sexual selection, are important in determining limits. Assuming that species borders are regions of energy restriction, it is difficult to envisage much widening of windows of opportunity for direct stresses of climatic origin. However, when combined with resource heterogeneity, evolutionary divergence led by behavioral shifts appears more likely, provided that abiotic perturbations are not extreme. In abiotically benign environments, implying minimal energy constraints, resource use specialization by learning appears possible. Heterozygotes tend to be favored in extreme environments because of their energy and metabolic efficiency. Therefore, genetic variability is unlikely to be restrictive in stressed free-living outlier populations; however, ecological circumstances can preclude the survival of novel variants. Consequently, the primary key to understanding limits to adaptation for behavioral traits is likely to be ecological. Under stressed free-living conditions, favored “good genotypes” are likely to be stress resistant and heterozygous. An association between success in mating, the development of extreme sexual ornaments, rapid development, and a long life can be postulated based on the metabolic efficiency of stress-resistance genotypes. While these postulated associations among fitness traits are supported by only limited empirical evidence, they may be important in any habitat where organisms are close to their limits of survival. If many organisms are born but few survive to reproduce because of climatic stress interacting with and causing nutritional stress, this situation may be quite normal. Although I contend that this model of the environment is generally valid, a reference point for comparisons with more benign environments is certainly provided.

References Adolph. S. C.. and Porter, W. P. (1993). Temperature, activity and lizard life histories. Am Nnt. 142, 213-295.

BEHAVIORAL VARIABILITY AND EVOLUTIONARY ADAPTATION

175

Anderson, R. W. (1995). Learning and evolution: A quantitative genetics approach. J. Theor. Biol. 175, 89-101. Arking, R.. Buck, S., Wells, R. A,. and Pretzalff. R. (1988). Metabolic rates in genetically based long lived strains of Drosophilu. Exp. Gerontol. 23, 59-76. Baldwin, J. M. (1896). A new factor in evolution. Am. Nut. 30, 441-451. Barros, A. R., Sierra, L. M.. and Comendador, M. A. (1991). Decreased metabolic rate as an acrolein resistance mechanism in Drosophilu melunogaster. Behav. Genet. 21,321-328. Belyaev, D. K., and Borodin. P. M. (1982). The influence of stress on variation and its role in evolution. Biol. Zentralbl. 100, 705-714. Bennett, A. F. (1991). The evolution of activity capacity. . I . Exp. B i d . 160, 1-23. Berglund, A. (1993). Risky sex: Male pipefishes mate at random in the presence of a predator. Anim. Behav. 46, 169-175. Blaxter, K. (1989). “Energy Metabolism in Animals.” Cambridge University Press, New York. Boggs, C. L. (1994). The role of resource allocation in understanding reproductive patterns. In ”Individuals, Populations and Patterns in Ecology” (S. R. Leather, A. D. Walt, and N. J. Mills, eds.), pp. 25-33. Intercept Ltd., Andover, Hampshire. Bonner, J. T. (1988). “The Evolution of Complexity by Means of Natural Selection.” Princeton University Press, Princeton, NJ. Booth, D. T., Clayton, D. H., and Block, B. A. (1993). Experimental demonstration of the energetic cost of parasitism in free-ranging hosts. Proc. R. Soc. London Ser. B 223, 123-129. Boulttreau-Merle, J., Fouillet, P., and Terrier, 0. (1987). Seasonal variations and balanced polymorphisms in the reproductive potential of temperate D. melanoguster populations. Entomol. Exp. Appl. 43, 39-48. Bozinovic, F., and Rosenmann, F. (1989). Maximum metabolic rate of rodents: Physiological and ecological consequences in distributional limits. Acnct. Ecol. 3, 173-181. Brown, C. R., Brown, M. B., and Rannala, B. (1995). Ectoparasites reduce long-term survival of their avian host. Proc. R. Soc. London, Ser. B 262, 313-319. Brown. J. H.. Marquet. P. A,, and Taper, M. L. (1993). Evolution of body size: Consequences of an energetic definition of fitness. Am. Nut. 142, 373-384. Christian, K. A., and Morton, S. R. (1992). Extreme thermophilia in a Central Australian ant, Melophoriis bugoti. Physiol. Zool. 65, 885-905. Clayton. D. H. (1990). Mate choice i n experimentally parasitized rock doves: Lousy males lose. Am. Zool. 30, 251-261. Cordero. A. (1995). Correlates of male mating success in two natural populations of the damselfly Ischnuru gruellsii (Odonata: Coenagrionidae). Ecol. Entomol. 20, 213-222. Craig, T. P.. Itami. J. K., Abrahamson. W. G., and Horner, J. D. (1993). Behavioral evidence for host-race formation in Eitrostasolidaginis. Evolution (Lawrence, Kuns.)47,1696-1710. Croy, M. I., and Hughes. R. N. (1991). The role of learning and memory in the feeding behavior of the fifteen-spined stickleback. Spinachiu spinachia L. Anim. Behav. 41, 149-159. Curtsinger, J. W., Fukui, H. H., Khazaeli, A. A,, Kirscher. A,, Pletcher, S. D., Promislow. D. E. L., and Tartar, J. (1995). Genetic variation and aging. Annu. Rev. Genet. 29,553-575. De Kruipf, H. A. M. (1991). Extrapolation through hierarchical levels. Comp. Biochem. Physiol. IOOC, 291-299. de Lope, F., Gonzalez. G., Perez, J. J.. and Moiler. A. P. (1993). Increased detrimental effects of ectoparasites on their bird hosts during adverse environmental conditions. Oecologiu 95, 234-240. Dobkin, D. S., Olivieri, I.. and Ehrlich. P. R. (1987).Rainfall and the interaction of microclimate with larval resources in the population dynamics of checkerspot butterflies (Euphydryus editha) inhabiting serpentine grassland. Oecologica 71, 161-166.

176

P. A. PARSONS

Duffy, J . E. (1996). Resource-associated population subdivision in a symbiotic coral-reef shrimp. Evolution (Lawrence, Kans.) 50, 360-373. Dunham, A. E., Grant, B. W., and Overall, K. L. (1989). Interfaces between biophysical and physiological ecology and the population ecology of terrestrial vertebrate ectotherms. Physiol. 2001.62, 335-355. Dunson, W. A,, and Travis, J . (1991). The role of abiotic factors in community organization. Am. Nat. 138, 1067-1081. Ehlinger. T. J. (1990). Habitat choice and phenotype-limited feeding efficiency in bluegill: Individual differences and tropic polymorphism. Ecology 71, 886-896. Ehrlich. P. R., Murphy, D. D.. Singer, M. C., Sherwood. C. B., White, R. R., and Brown, I. L. (1980). Extinction, reduction, stability and increase: The responses of checkerspot butterfly (Euphydryas) populations to the California drought. Oecologia 46, 101-105. Ehrman, L., and Parsons, P. A. (1981). “Behavior Genetics and Evolution.” McGraw-Hill, New York. Elliott, J . M. (1995). Egg hatching and ecological partitioning in carnivorous stoneflies (Plecoptera). C. R. Acad. Sci., Life Sci. 318, 237-243. Feder, J . L., Chilcote, C. A,, and Bush, G. L. (1988). Genetic differentiation between sympatric host races of the apple maggot fly Rhagoletis pomonella. Nature (London) 336,61-64. Fisher, R. A. (1930). “The Genetical Theory of Natural Selection.” Clarendon Press. Oxford. Fitzsimmons, N. N., Buskirk, S. W., and Smith, M. H. (1995). Population history, genetic variability, and horn growth in bighorn sheep. Conserv. Biol. 9, 314-323. Folstad, I., Arneberg, P., and Kanter. A. J. (1996). Antlers and parasites. Oecologia 105, 556-558. Ganem, G., and Nevo, E. (1996). Ecophysiological constraints associated with aggression. and evolution towards pacifism in Spalax ehrenbergi. Behav. Ecol. Sociobiol. 38,245-252. Gehring, W. J., and Wehner. R. (1995). Heat shock protein synthesis and thermotolerance in Cataglyphis, an ant from the Saharan desert. Proc. Natl. Acad. Sci. U.S.A.92,2994-2998. Godin, J.-G. J., and Briggs, S. E. (1996). Female mate choice under predation risk in the guppy. Anim. Behav. 51, 117-130. Good, D. S. (1993). Evolution of behaviours in Drosophila melanogaster in high temperatures: Genetic and environmental effects. J. Insect Physiol. 39, 537-544. Grant, B. R. (1996). Pollen digestion by Darwin’s finches and its importance for early breeding. Ecology 77, 489-499. Grant, B. W., and Dunham, A. E. (1988). Thermally imposed time constraints on the activity of a desert lizard Scleroporus merriami. Ecology 69, 167-176. Grant, P. R., and Grant,B. R. (1989). Sympatricspeciation and Darwin’s finches. In “Speciation and its Consequences” (D. Otte and J. A. Endler, eds.), pp. 433-457. Sinauer, Sunderland, MA. Guarna, M. M., and Borowsky. R. L. (1993). Genetically controlled food preference: Biochemical mechanisms. Proc. Natl. Acad. Sci. U.S.A. 90, 5257-5261. Hall, C. A. S., Stanford, J. A,, and Hauer. F. R. (1992). The distribution and abundance of organisms as a consequence of energy balances along multiple environmental gradients. Oikos 65,377-390. Haller, J.. and Wittenberger, C. (1988). Biochemical energetics of hierarchy formation in Betta splendens. Physiol. Behav. 43, 447-450. Halliday, T. R. (1978). Sexual selection and mate choice. In “Behavioural Ecology: An Evolutionary Approach” ( J . R. Krebs and N. B. Davies. eds.), pp. 180-213. Blackwell, Oxford. Hamilton, W. D., and Zuk. M. (1982). Heritable true fitness and bright birds: A role for parasites? Science 218,384-387.

BEHAVIORAL VARIABILITY AND EVOLUTIONARY ADAPTATION

177

Hoffmann, A. A.. and Blows. M. W. (1994). Species borders: Ecological and evolutionary perspectives. Trends Ecol. Evol. 9, 223-227. Hoffmann, A. A., and Parsons, P. A. (1991). “Evolutionary Genetics and Environmental Stress.” Oxford University Press, Oxford. Hoffmann, A. A.. and Turelli, M. (1985). Distribution of Drosophila melanogasrer on alternative resources: Effects of experience and starvation. A m . Nai. 126, 662-679. Hogg, J. T. (1987). Intrasexual competition and mate choice in Rocky Mountain bighorn sheep. Ethology 75, 119-144. Huey, R. B. (1991). Physiological consequences of habitat selection. Am. Nat. 137, S91-Sl15. Hurst, R. J., Watts. P. D., and Oritsland, N. A. (1991). Metabolic compensation in oil-expressed polar bears. J. Therm. Biol. 16, 53-57. Janssen, A., Van Alphen, J. J. M.. Sabelis, M. W., and Bakker, K. (1995). Odour-mediated avoidance of competition in Drosophila parasitoids: The ghost of competition. Oikos 73,356-366. Johannesson. K., Rolan-Alvarez, E.. and Ekendahl. A. (1995). Incipient reproductive isolation between two sympatric morphs of the intertidal snail. Lirrorina saxafilis. Evolution (Lawrence, Kans.) 49, 1180-1 190. Jones, J. S., Coyne, J. A., and Partridge, L. (1987). Estimation of thermal niche of Drosophila melanogasrer using a temperature-sensitive mutation. Am. Nar. 130, 83-90. Kauffman. S. A. (1993). “The Origins of Order: Self-organization and Selection in Evolution.” Oxford University Press, New York. Kennedy. C. E. J., Endler, J. A,, Poynton, S. L., and McMinn, H. (1987). Parasite load predicts mate choice in guppies. Behav. Ecol. Sociobiol. 21, 291-295. Klieber, M. (1961). “The Fire of Life: An Introduction to Animal Energetics.” Wiley, New York. Knowlton, N. (1993). Sibling species in the sea. Annic. Rev. Ecol. Syst. 24, 189-216. Kodric-Brown, A., and Nicoletto, P. F. (1993). The relationship between physical condition and social status in pupfish, Cyprinodon pecosensis. Anim. Behav. 46, 1234-1236. Koehn. R. K., and Bayne, B. L. (1989). Towards a physiological and genetical understanding of the energetics of the stress response. B i d . J. Linn. Soc. 37, 157-171. Krebs, R. A., and Loeschcke, V. (1994). Costs and benefits of activation of the heat-shock response in Drosophila melanogasier. Fiinci. Ecol. 8, 730-737. Lande, R.(1980). Sexual dimorphism. sexual selection, and adaptation in polygenic characters. Evolution (Lawrence, Kans.) 34, 292-305. Li, J., and Margolies. D. C. (1994). Responses to direct and indirect selection on aerial dispersal behaviour in Tetranychirs urticae. Heredity 72, 10-22. Lovegrove. B. G . (1986). The metabolism of social subterranean rodents: Adaptation to aridity. Oecologia 69, 551-55.5. Low, B. S. (1990). Marriage systems and pathogen stress in human societies. Am. Zool. 30, 325-339. Lubin, Y. D., and Henschel, J. R. (1990). Foraging at the thermal limit: Borrowing spiders (Seothyru, Eresidae) in the Namib desert dunes. Oecologia 84,461-467. Magnhagen, C. (1991). Predation risk as a cost of reproduction. Trends Ecol. Evol. 6,183-186. Marden, J. H., and Waage, J. K. (1990). Escalated damselfly territorial contests are energetic wars of attrition. Anim. Behav. 39, 954-959. Markow, T. (1994). “Developmental Instability: Its Origins and Evolutionary Implications.” Kluwer, Dordrecht, The Netherlands. McLain, D. K. (1993). Cope’s rules, sexual selection, and the loss of ecological plasticity. Oikos 68,490-500.

178

P. A. PARSONS

McLain, D. K., Moulton, M. P., and Redfearn, T. P. (1995). Sexual selection and the risk of extinction of introduced birds on oceanic islands. Oikos 74, 27-34. Mitton, J. B. (1993). Enzyme heterozygosity, metabolism, and developmental variability. Genetica 89, 47-63. Mdler, A. P. (1991). Clutch size, nest predation, and distribution of avian unequal competitors in a patchy environment. Ecology 72, 1336-1349. Mdler, A. P. (1994a). “Sexual Selection and the Barn Swallow.” Oxford University Press, Oxford. Meller, A. P. (1994b). Male ornament size as a reliable cue to enhanced offspring viability in the barn swallow. Proc. NUB Acad. Sci. U.S.A. 91, 6929-6932. Mdler, A. P. (1996). The cost of secondary sexual characters and the evolution of costreducing traits. Ibis 138, 112-1 19. M~iller,A. P., and Thornhill, R. (1997). A meta-analysis of the heritability of developmental stability. J. Evol. B i d . 10, 1-16. Moore, A. J. (1994). Genetic evidence for the “good genes” process of sexual selection. Behav. Ecol. Sociobiol. 35, 235-241. Nevo, E. (1991). Evolutionary theory and processes of active speciation and adaptive radiation in subterranean mole rats, Spalax ehrenbergi superspecies, in Israel. Evol. B i d . 25,l-125. Nevo, E., Simson, S., Heth, G., and Beiles, A. (1992). Adaptive pacifistic behaviour in subterranean mole rats in the Sahara desert, contrasting to and originating from polymorphic aggression in Israeli species. Behaviour 123, 70-76. Norris, K. (1993). Heritable variation in a plumage indicator of viability in male great tits Parus major. Nature (London) 326,537-539. Odum, E. P., Finn, J. T., and Franz, E. H. (1979). Perturbation theory and the subsidy stress gradient. BioScience 29, 349-352. Parsons, P. A. (1987). Evolutionary rates under environmental stress. Evol. Biol. 21,311-347. Parsons, P. A. (1988). Behavior, stress and variability. Behav. Genet. 18, 293-308. Parsons, P. A. (1990). Fluctuating asymmetry: An epigenetic measure of stress. B i d . Rev. Cambridge Philos. Soc. 65, 131-145. Parsons, P. A. (1991). Evolutionary rates: Stress and species boundaries. Annu. Rev. Ecol. Syst. 22, 1-16. Parsons, P. A. (1992). Evolutionary adaptation and stress: The fitness gradient. E d . B i d . 26, 191-223. Parsons, P. A. (1993a). Habitat preference: An interaction between genetic variability and the costs of stress. Etologia 3, 1-9. Parsons, P. A. (1993b). Evolutionary adaptation and stress: Energy budgets and habitats preferred. Behav. Genet. 23, 231-238. Parsons, P. A. (1995a). Stress and limits to adaptation: Sexual ornaments. J . Evol. B i d . 8,455-461. Parsons, P. A. (1995b). Inherited stress resistance and longevity: A stress theory of ageing. Heredity 75, 216-221. Parsons, P. A. (1996a). Competition versus abiotic factors in variably stressful environments: Evolutionary implications. Oikos 75, 129-132. Parsons, P. A. (1996b). Stress, resources, energy balances, and evolutionary change. Evol. Biol. 29, 39-72. Parsons, P. A. (1996~).Rapid development and a long life: An association expected under a stress theory of ageing. Experientia 52, 643-646. Parsons, P. A. (1997a). Success in mating: A co-ordinated approach to fitness through genotypes incorporating genes for stress resistance and heterozygous advantage under stress. Behav. Genet. 27, 75-81.

BEHAVIORAL VARIABILITY A N D EVOLUTIONARY ADAPTATION

179

Parsons, P. A. (1997b). Stress-resistance genotypes, metabolic efficiency and interpreting evolutionary change. In “Stress, Adaptation and Evolution” ( V . Loeschcke and R. Bijlsma, eds.), pages 291-305. Birkhaueser, Basel. Petrie, M. (1994). Improved growth and survival of offspring of peacocks with more elaborate trains. Nature (London) 371, 598-599. Polak, M. (1997). Parasites, fluctuating asymmetry and sexual selection. In “Parasites: Effects on Host Hormones and Behavior” (N. E. Beckage, ed.). Chapman & Hall, New York. In press. Pomiankowski. A. (1987). The costs of choice in sexual selection. J. Theor. Biol. 128,195-218. Porter, W. P., and Gates, D. M. (1969). Theormodynamic equilibia of animals with environment. Ecol. Monogr. 39,227-244. Pough. F. H. (1989). Organismal performance and Darwinian fitness: Approaches and interpretations. Physiol. Zool. 62, 199-236. Prinzinger. R., and Siedle, K. (1988). Ontogeny of metabolism, thermoregulation and torpor in the house martin, Delichon urbica (L.) and its ecological significance. Oecologia 76, 307-312. Prokopy, R. J.. and Papaj, D. R. (1988). Learning of adult fruit biotypes by apple maggot flies. J. Insect. Behav. 1, 67-74. Riddoch, B. J. (1993). The adaptive significance of electrophoretic mobility in phosphoglucose isomerase (PGI). Biol. J. Linn. Soc. 50, 1-17. R’Kha, N., Capy. P., and David, J. R. (1991). Host-plant specialization in the Drosophila melanogaster species complex: A physiological, behavioral, and genetical analysis. Proc. Natl. Acad. Sci. U.S.A. 88, 1835-1839. Rodriguez, L., Sokolowski, M. B., and Carton, Y. (1991). Intra- and inter-specific variation in population behaviors of Drosophila from different habitats. Can. J. Zool. 69,2616-2619. Rodriguez, L.. Sokolowski. M. B., and Shore, J. S. (1992). Habitat selection by Drosophila mrlanogaster larvae. J. Evol. Biol. 5, 61-70. Rolan-Alvarez, E., Zapata, C., and Alvarez, G. (1995). Multilocus heterozygosity and sexual selection in a natural population of the marine snail Littorina mariae (Gastropoda: Prosobranchia). Heredity 75, 17-25. Root, T. ( I 993). Effects of global climate change on North American birds. In “Biotic Interactions and Global Change” (P. M. Kareiva, J. G. Kingsolver, and R. B. Huey, eds.), pp. 280-292. Sinauer, Sunderland, MA. Rosewell, J . , and Shorrocks, B. (1987). The implication of survival rates in natural populations of Drosophila; Capture-recapture experiments in domestic species. Biol. J. Linn. SOC. 32,373-384. Reskaft. E., Jarvi. T.. Bakken, M., Bech, C., and Reinertsen, R. E. (1986). The relationship between social status and metabolic rate in great tits (Parus major) and pied flycatchers (Ficedula hypolercca) Anim. Behav. 34, 838-842. Ryan, M. J. (1988). Energy, calling, and selection. Am. Zool. 28, 885-898. Schall, J. J., and Sarni, G. A. (1987). Malarial parasitism and the behavior of the lizard, Sceloporus occidentalis. Copeia 1, 84-93. Schleucher, E. (1993). Life in extreme dryness and heat: A telemetric study of the behaviour of the diamond dove Geopelia cuneata in its natural habitat. Emu 93, 251-258. Schluter, D., and McPhail, J. D. (1993). Character displacement and replicate adaptive radiation. Trends Ecol. Evol. 8, 197-200. Scribner, K. T . ,Smith, M. H.. and Johns, P. E. (1989). Environmental and genetic components of antler growth in white-tailed deer. J . Mammal. 70, 284-291. Skulason, S., and Smith, T. B. (1995). Resource polymorphism in vertebrates. Trends E d . Evol. 10. 366-370.

180

P. A. PARSONS

Skulason. S.. Snorrason. S. S., Ota. D.. and Noakes. D. L. G. (1993). Genetically based differences in foraging behaviour among sympatric morphs of arctic charr (Pisces: Salmonidae). Anirn. Behav. 45, 1179-1 192. Stanhope, M. J., Leighton, B. J., and Hartwick. B. (1992). Polygenic control of habitat preference and its possible role in sympatric population subdivision in an estuarine crustacean. Heredity 69, 279-288. Tanaka, Y. (1996). Sexual selection enhances population extinction in a changing environment. J. Theor. Biol. 180, 197-206. Thornhill, R., and Alcock. J. (1983). “The Evolution of Insect Mating Systems.” Harvard University Press, Cambridge, MA. Thornhill. R., and Gangestad. S. W. (1993). Human facial beauty: Averageness, symmetry and parasite resistance. Hum. Net. 4, 237-269. Van Valen, L. (1976a). Ecological species, multispecies, and oaks. Taxon 25, 233-239. Van Valen. L. (1976b). Energy and evolution. Evol. Theory 1, 179-229. Vehrencamp, S. L., Bradbury. J. W.. and Gibson, R. M. (1989). The energetic cost of display in male sage grouse. Anim. Behav. 38,885-896. von Schantz,T.. Wittzell, H.. Goransson, G.. Grahn, M.. and Persson, K. (1996). MHC genotype and male ornamentation: Genetic evidence for the Hamilton-Zuk model. Proc. R. Soc. London, Ser. B 263, 265-271. Wedekind, C. (1994). Handicaps not obligatory i n sexual selection for resistance genes. J. Theor. Biol. 170, 57-62. Wehner. R.. Marsh, A. C., and Wehner, S. (1992). Desert ants on a thermal tightrope. Nature (London) 357,585-587. Weiner, J. (1992). Physiological limits to sustainable energy budgets in birds and mammals: Ecological implications. Trends Ecol. Evol. 7 , 384-389. Werner, T. K., and Sherry, T. W. (1987). Behavioral specialization in Pinaroloxias inornara, the “Darwin’s finch” of COCOS Island, Costa Rica. Proc. Natl. Acad. Sci. U.S.A.84,5506-5510. West-Eberland, M. J. (1986). Alternative adaptations, speciation, and phylogeny (A review). Proc. Natl. Acad. Sci. U.S.A. 83, 1388-1392. White, T. C. R. (1993). “The Inadequate Environment: Nitrogen and the Abundance of Animals.” Springer-Verlag. Berlin. Yamamoto, A. H. (1994). Temperature preferences of Drosophila immigrans and D. virilis: Intra- and interpopulation genetic variation. Jpn. J. Genet. 69, 67-76. Ye, S., Aracena, J., Good. D. S., and Bell, W. J. (1994). Correlation between survival during food deprivation and search behaviour in populations of Drosophilu melanogasfer. J . Insect Physiol. 40, 137-142. Zakharov. V. M. (1989). Future prospects for population phenogenetics. Sov. Sci. Rev.,Sect. F: Physiol. Gen. Biol. 4, 1-79. Zotin, A. I . (1990). “Thermodynamic Bases of Biological Processes: Physiological Reactions and Adaptations.“ de Gruyter, Berlin. Zuk, M., Thornhill, R., Ligon, J. D., and Johnson, K. (1990). Parasites and mate choice in red jungle fowl. Am. Zool. 30, 235-244.

ADVANCES IN THE STUDY OF BEHAVIOR, VOL 27

Developmental Instability as a General Measure of Stress ANDERS PAPE MBLLER LABORATOIRE D'ECOLOGIE U N I V E R S I T ~PIERRE ET MARIE CURIE PARIS CEDEX

5,

FRANCE

I. INTRODUCTION Any organism, animal, plant, or fungus, is characterized by regularity of its phenotype. That is why we often recognize species and sometimes even sexes and age classes of particular species. A number of different mechanisms ensure that development does not go wrong, and that developmental processes are kept within certain limits. Disruptions of developmental trajectories are caused by developmental noise from the environment, but also inferior developmental processes caused by the genetic setup of the individual. This ability to control development under given environmental conditions is called developmental stability. Developmental stability cannot be measured directly, but deviations from a regular phenotype provide information on developmental instability. A number of such measures have been proposed including fluctuating asymmetry, the frequency of phenodeviants, and deviations from modal behavior, physiology, and immunology. Most organisms display bilateral or radial symmetry, and random deviations from such symmetry is termed fluctuating asymmetry (Ludwig, 1932; Van Valen, 1962; Palmer and Strobeck, 1986; Parsons, 1990; Moller and Swaddle, 1997). In other words, characters displaying fluctuating asymmetry have signed right-minus-left trait values and normal frequency distributions with a mean value of zero (Fig. 1). The level of asymmetry of an individual belonging to a population demonstrating fluctuating asymmetry is simply called its asymmetry or individual fluctuating asymmetry; the term fluctuating asymmetry, however, is a population parameter. It was only recently that asymmetry at the level of the individual was used as a measure of phenotypic quality (Moller, 1990). This individual approach is a most powerful tool in many different kinds of studies, as we shall see later. Two other kinds of morphological asymmetry are common: Antisymmetry is 181

Cvpyriphf 0 I9YX hy Academic Press All rights ol reproduction i n any iorm rcwrvcd. IK)hS-34S4/98 $2S.M1

182

ANDERS PAPE MOLLER

A

Directional asymmetry

-10 -8 -6 -4 -2

B

0

2

4

6

8 10

6

8 10

Antisymmetry

2ol 16

-10 -8 -6 -4 -2

c

0 2

4

Fluctuating asymmetry

-10 -8 -6 -4 -2

0

2

4

6

8 10

Right-minus-left character value FIG. 1. Frequency distributions of signed right-minus-left character values for traits demonstrating ( A ) directional asymmetry, (B) antisymmetry, and (C) fluctuating asymmetry. Adapted from Moller and Swaddle (1997).

DEVELOPMENTAL INSTABILITY AND STRESS

183

characterized by individuals being asymmetric in a random direction, and the frequency distribution of signed left-minus-right character values having a deficiency of symmetric phenotypes (Fig. 1). Typical examples of antisymmetry are the signaling claw of male fiddler crabs Uca spp. and the beak of crossbills Loxia spp. The third kind of asymmetry is directional asymmetry, which is usually displayed in a particular direction (Fig. 1). The mean value of the frequency distribution of signed left-minus-right character values therefore deviates significantly from zero. Examples of directional asymmetry are the size of testes in mammals and the structure of ears in certain species of owls. There is currently some controversy over whether only fluctuating asymmetry, or also the two other kinds of asymmetry, reflects developmental instability, although a number of cases clearly suggest that antisymmetry and directional asymmetry may reflect poor developmental conditions (review in Moller and Swaddle, 1997). Developmental instability can also be estimated from other measures of deviant phenotypes. Gross abnormalities such as a position of the heart in the right side of the body cavity in some humans and four or six rather than five fingers on each hand are termed phenodeviants. Their frequency is actually positively correlated with fluctuating asymmetry and phenodeviants therefore reflect developmental instability (e.g., Rasmuson, 1960). A number of other valid and useful measures have been proposed for specific kinds of organisms, and their common feature is the morphological invariance. For example, snails grow their shells at a specific, constant angle, which results in more and more narrow whorls, but when exposed to environmental perturbations such as those caused by acid rain, the angle of growth changes (Graham et al., 1993). The deviation from the normal angle of growth is therefore a measure of developmental instability. Many kinds of plants have composite leaves consisting of leaflets that are typically exactly juxtaposed to one another. Deviations from commonly encountered environmental conditions result in the stalks of the leaflets being displaced, and the average deviation from perfectly juxtaposed leaflets is therefore a measure of developmental instability (Freeman et al., 1993). A number of other measures of developmental instability are listed by Graham ef al. (1993) and Mgller and Swaddle (1997). It is important to emphasize that phenomena other than developmental problems may give rise to normal frequency distributions of signed leftminus-right character values. Since individual fluctuating asymmetry is often very small, with the majority of all individuals having asymmetries less than 1% of the size of the character, measurement errors may contribute significantly to the overall estimate of fluctuating asymmetry. In fact, measurement errors also have normal frequency distributions with a mean value of zero! Scientists working on fluctuating asymmetry traditionally test for

184

ANDERS PAPE M0LLER

the magnitude of measurement errors in a sample of individuals or in the entire sample. The importance of measurement error can thereby be evaluated or directly partialled out (see Palmer, 1994, and Moller and Swaddle, 1997, for methods). Organisms demonstrate regularity of their phenotypes because such regularity promotes superior performance. Just think of the wealth of asymmetry within the human body and contrast that with the exterior phenotype. Race horses with symmetric skeletons win more races (Manning and Ockenden, 1994). Symmetry is the best solution to the engineering problem of constructing phenotypes that are well designed for locomotion. Many organisms are sessile and do not need streamlined bodies for efficient locomotion. Even though selection for symmetry may be most severe in mobile organisms, there is still selection, albeit weaker, for symmetry in fungi, plants, and sessile animals such as corals and sponges. The reason is that symmetry also gives rise to more efficient resource use (such as light and nutrients) and dispersal of propagules (such as spores, gametes, and seeds), but also results in less severe effects of the abiotic environment such as wave action and wind. Any deviation from symmetry is likely to impose performance costs, and a number of such costs have been proposed or directly demonstrated experimentally (see review in Moller and Swaddle, 1997). The optimal solution therefore appears to be given a priori; it is a symmetric phenotype. Bilateral or radial symmetry therefore differs from any other phenotypic measure because we know the optimum in advance. We might be able to identify the optimum for other traits, but not without extensive research and then only for an environment with particular characteristics. If developmental control is a costly process, then perfect symmetry might not be the optimal solution, since the incremental decrease in asymmetry achieved by further investment is a function with diminishing returns. We might be able to cope with average asymmetries of .1 mm in the length of our fingers, while 1 mm or 10 mm might pose problems in certain situations. Developmental instability is affected by a wide range of environmental (external) and genetic (internal) factors that contribute to disruption of the stable development of the phenotype. These factors are briefly reviewed below. The control of developmental processes proceeds most efficiently under the commonly encountered environmental conditions, and increasing deviations from such conditions result in the use of energy for stress tolerance that could otherwise be used for developmental control, growth, reproduction, and survival (Hoffmann and Parsons, 1989; Parsons, 1990; Alekseeva et al., 1992; Ozernyuk et al., 1992). Development, like any other process, is an energy-dissipating activity, and energy used for developmental control has to be diverted away from other vital activities. As the future performance of any organism depends on the developmental stability of

DEVELOPMENTAL INSTABILITY A N D STRESS

185

its phenotype, deviations from this goal can be used as a reliable measure of the challenge experienced by an individual in its natural environment. It is inherently impossible to generate a perfectly symmetrical phenotype; the level of fluctuating asymmetry will provide extremely important information about the developmental performance of an individual. This argument is of utmost importance because it allows us to obtain reliable information about the state of individual organisms in their natural environment as perceived by the organisms themselves. We cannot readily ask organisms how they perceive their environment, but we can use their developmental instabilities as an indirect answer to this question. If fluctuating asymmetry provides reliable information on the well-being of populations, and individual asymmetry does the same for individuals, then it should suffice to measure a single character and extract the information. However, it is a common finding that asymmetries in different characters often are not significantly positively correlated (review in Mdler and Swaddle, 1997). Why should that be the case? As already stated, developmental instability measured as individual fluctuating asymmetry integrates the effects of a number of different environmental and genetic factors, and if we were to rerun the development of an individual once more, we might not end up with exactly the same level of asymmetry. The estimate of developmental instability is basically an estimate of a variance based on a single measurement of two different morphological characters, and such an estimate is bound to have a high degree of uncertainty. The ability of an individual to develop the same phenotype repeatedly is called the repeatability of its individual fluctuating asymmetry. This quantity can be estimated readily when we make some simplifying assumptions and know the coefficient of variation of the asymmetry measure, the magnitude of our measurement errors, and the phenotypic variance of the character in question (Whitlock, 1996). The correlation in symmetry among a number of different characters is likely to provide a serious underestimate of the true correlation because of the lack of repeatability of developmental instabilities. An unbiased estimate of the true correlation turns out simply to be the correlation between the asymmetries divided by the square root of the product of the repeatabilities (Whitlock, 1996). There are other explanations for this general absence of a correlation between asymmetries of different characters (review in Mgller and Swaddle, 1997), but the explanation presented above is perhaps the most likely. A way of resolving this problem when choosing characters for measurement is to choose a number of different characters and use a composite measure of asymmetry as an indicator of the overall level of developmental instability. The ability to control developmental processes and generate a stable phenotype differs among individuals, and both genetic and environmental

186

ANDERS PAPE MOLLER

components affect this ability, as for any other character. A large number of studies of different organisms have shown that there is indeed a statistically significant heritability of measures of developmental instability (Mdler and Thornhill, 1997a). Although the quality of these studies differs in a number of ways that may affect the estimates of the genetic and environmental components, independently of how the data are selected the conclusion remains stable: developmental instability has a statistically significant heritability (Houle, 1997; Leamy, 1997; Markow and Clarke, 1997; Palmer and Strobeck, 1997; Pomiankowski, 1997; Swaddle, 1997; Whitlock and Fowler, 1997; Moller and Thornhill, 1997b). The signficance of this finding is that relatives will resemble each other with respect to developmental instability, and that developmental stability may evolve. In the following three sections, I briefly review (1) the genetic and environmental determinants of developmental instability, (2) the relationship between developmental instability and mode of selection, and (3) the relationship between developmental instability and fitness.

A N D ENVIRONMENTAL DETERMINANTS OF 11. GENETIC

DEVELOPMENTAL INSTABILITY Developmental instability as estimated from fluctuating asymmetry and the frequency of phenodeviants has been investigated in many hundreds of studies, and some general patterns have emerged concerning the factors that contribute to increased developmental problems. These can basically be divided into internal genetic and external environmental causes, which are briefly reviewed here. An extensive review is provided by Mdler and Swaddle (1997). A.

INSTABILITY GENETIC CAUSESOF DEVELOPMENTAL

The genetic factors that increase developmental instability include inbreeding, homozygosity, hybridization, and mutation. Inbreeding results in a reduction in additive genetic variation, but also in the exposure of deleterious recessive alleles that become fully expressed in recessive homozygotes. Either of these effects may disrupt the stable development of a phenotype. A recent review of the literature has shown that a large majority of the studies indeed found developmental instability to be negatively associated with inbreeding (Moller and Swaddle, 1997). Exceptions to this finding may be explained in a number of different ways of which the selective loss of asymmetric homozygotes at early embryonic stages is a plausible one.

DEVELOPMENTAL INSTABILITY A N D STRESS

187

Heterozygosity at protein encoding loci is an indicator of genetic variability, and individuals that are able to produce a more diverse array of biochemical products should be capable of coping with a wide range of environmental conditions. There is considerable evidence in agreement with this suggestion, and metabolism and therefore also growth have repeatedly been shown to be most efficient among heterozygous individuals (review in Mitton and Grant, 1984). If stable development is associated with efficient metabolism and controlled growth processes, one might hypothesize a positive relationship between developmental homeostasis and heterozygosity (as first suggested by Lerner, 1954). A large number of studies have addressed this question, but with very mixed results. A meta-analysis (a statistically based review of overall effects and heterogeneity in effects among studies) revealed that there was no consistent association between heterozygosity and measures of developmental instability such as fluctuating asymmetry (Vdlestad et al., 1997). However, there was statistically highly significant heterogeneity in the effects among studies. Some of this could be explained by whether the study organisms were heterothermic or homeothermic (the latter providing a more stable and protected developmental environment). Therefore, this review provided little evidence for a general association. Vdlestad et al. (1997) discussed several explanations for this lack of an association. A particularly likely explanation is that many studies have been performed under relatively benign laboratory conditions that do not result in severe environmental stress. Perhaps a high degree of heterozygosity is beneficial only under stressful conditions. This hypothesis has been tested explicitly by Mulvey et al. (1994) on the fish Gambusia holbrooki. Fluctuating asymmetry and heterozygosity were negatively related at high temperatures, but completely unrelated at an optimal water temperature. Another possibility is that the relationship holds only for specific enzymes that are of particular functional importance in a specific context (Mitton, 1995). This could certainly account for some of the results for specific enzymes, but not for relationships between asymmetry and general heterozygosity as found in a number of different studies. Hybridization results in the genomes of two species being mixed, and this may have severe fitness consequences because the functioning of biological systems depends on the cooperation of the different components of the genome. Of course, hybridization also results in the generation of novel genetic variation and may therefore in some cases of closely related species result in hybrid vigor. Divergence and reproductive isolation are outcomes of the process of speciation, which may result in the acquisition of different coadapted gene complexes. If different genomes are combined, this may result in disruption of development because the gene combinations of the hybrids have not been subject to natural selection. This will obviously

188

ANDERS PAPE M0LLER

particularly be the case when different species have been isolated from each other for a long time rather than recently (in evolutionary terms). Experiments with eggs of the Florida large mouth bass, Micropterus salmonides floridanus, fertilized by the sperm of ten different species resulted in an increase in developmental deviants as the divergence measured by the genetic distance increased (Parker et al., 1985). A review of the literature demonstrated a clear general pattern of increased developmental instability in hybrids as compared to the parental species, particularly if the species in question had diverged considerably (Moller and Swaddle, 1997). Mutations usually result in a deterioration of the phenotype, but every now and then a slight improvement may arise. It has been known for a long time that mutations usually result in deviant phenotypes with properties similar to those of asymmetric and phenodeviant individuals (GoldSchmidt, 1940, 1955). Particularly mutations with a low penetrance give rise to deviant phenotypes, while highly penetrant mutations show less phenotypic effects with respect to developmental regularity (Goldschmidt, 1940, 1955). Again, the explanation for this effect appears to be the lack of genetic coadaptation. If genes are manipulated by use of recent molecular techniques, the resultant phenotype becomes asymmetric for ETS2-alleles, which have effects similar to those of Down’s syndrome (Sumarsono et al., 1996). the vascular endothelial growth factor gene (Carmeliet et af., 1996), and homeobox genes (Davis et al., 1995). In conclusion, a number of different genetic factors contribute to the development of a stable phenotype, although no consensus exists concerning the relative roles of the different factors. The genetic mechanisms involved in generating asymmetric phenotypes appear to be gene coadaptation, but potentially also other mechanisms. B. ENVIRONMENTAL CAUSES OF DEVELOPMENTAL INSTABILITY Environmental factors include temperature, food, pollutants, population density, sound, light, and parasites. The diversity of environmental stresses that have been shown to cause an increase in asymmetry is probably not exclusive; many other kinds of stress might provide similar effects. Temperatures that deviate from optimal conditions result in increased energy expenditure for stress resistance. Increased temperature differences from the normal range encountered have been shown to result in increased asymmetry in Drosophila (Beardmore, 1960), rats Rattus norvegicus (Gest et al., 1983, 1986), and a number of other organisms (review in Moller and Swaddle, 1997). Nutritional stress has been shown to increase asymmetry in a number of different organisms under experimental conditions. For example, European

DEVELOPMENTAL INSTABILITY AND STRESS

189

nuthatches, Sitta europaea, that had a couple of feathers removed during winter, regrew these feathers more symmetrically when provided with extra food than did control birds (Nilsson, 1994). Similar results have been obtained for a wide variety of organisms (Mdler and Swaddle, 1997). Free-living organisms encounter a diverse chemical environment during development, particularly in species with external fertilization. Deviations from commonly encountered conditions result in increased developmental instability. This is the case for several different kinds of pollutants, but also for chemicals found in the food consumed by animals. For example, gray seals, Hulichoerus grypus, from the Baltic had increased asymmetry in their skulls during the 1950s and 1960s (Zakharov and Yablokov, 1990), but experienced a decrease in asymmetry as the concentrations of pollutants decreased in recent decades (Zakharov ef al., 1989). In a similar vein, alcohol consumption in pregnant women resulted in increased asymmetry in their children (Kieser, 1992). Further examples are discussed in Section V,A. Increasing population density results in a reduction in the amount of nutrients available per individual, but also in energy spent on stress resistance that could otherwise be used for control of developmental processes. Several studies have shown that asymmetry and phenodeviants increase as a consequence of increased density. For example, elevated larval density of Australian sheep blowflies, Lucilia cuprina, resulted in increased asymmetry in the adult flies (Clarke and McKenzie, 1992). Similarly, although under more natural conditions, skeletal asymmetry followed population density in the small mammal cycles of the common shrew, Sorex araneus, in Siberia (Zakharov et al., 1991). A final example derives from a study of similarly aged clones of poplars, Popitlus arnericanus, planted at three different densities (Rettig et aZ., 1997). The effect of density on asymmetry increased linearly from a density of .167 to 2.0 plants per square meter. Audiogenic stress may also increase asymmetry of the phenotype. Early experiments by Siege1 and Smookler (1973) demonstrated that pregnant rats that were exposed to noise had pups with increased dental asymmetry. This effect has been repeated in a number of subsequent experiments on rats and other rodents (review in Mdler and Swaddle, 1997). A final example of an environmental component that can increase asymmetry is exposure to predators (Witter and Lee, 1995). Molting starlings, Sturnus vulgaris, were kept in aviaries with food provided near or away from shelter. Hence, there was no differential exposure to predators per se, but just a perceived difference in exposure. Starlings that developed their feathers while feeding at an exposed food source developed significantly greater feather asymmetry than did controls. As asymmetric individuals are more likely t o fall prey to predators (Section IV), increased perceived

190

ANDERS PAPE M 0 L L E R

risks of predation may actually result in increased predation, if increased morphological asymmetries give rise to reduced performance. In conclusion, the development of an individual integrates the effects of a wide range of environmental and genetic factors that affect the stability of developmental processes. This is an advantage for the scientists because the overall effect of many different factors is added up into the phenotype, but a disadvantage because we will not obtain information on the particular factor that is causing an increase in asymmetry.

111.

DIRECTIONAL SELECTION A N D DEVELOPMENTAL INSTABILITY

The previous section dealt with a range of different genetic and environmental factors that tend to increase the level of fluctuating asymmetry. However, this is not the entire story, as characters of individuals even when conditions are kept constant still may differ in their measures of developmental instability. The recent (in evolutionary terms) history of selection affecting a character may strongly influence the potential for development of asymmetry. Characters that have been subject to a history of directional or disruptive selection are generally less stable than characters subject to a history of stabilizing selection (Moller and Pomiankowski, 1993a,b; Moller and Swaddle, 1997). The reason for this phenomenon appears to be that intense directional selection selects against any mechanisms that control the full expression of a character. These control mechanisms are also involved in the stable expression of the phenotype. Stabilizing selection has the opposite effect of incorporating developmental mechanisms that prevent the expression of extreme phenotypes, but also avoid the expression of asymmetric phenotypes. The evidence for this scenario comes from a range of different sources (sexual selection, life-history traits, plant-animal interactions, the paleontological record, domestication, and laboratory experiments) of which three are mentioned below. Sexual selection results in the evolution of extravagant characters that are costly to produce and maintain, and therefore do not ameliorate the effects of natural selection (Darwin, 1871). The divergence in secondary sexual characters is generally much larger than in ordinary morphological traits, apparently leading to pre- and postcopulatory species isolation mechanisms. This extreme divergence is evidence of a recent history of evolutionary change. We should therefore expect secondary sexual characters in general, but particularly those currently subject to intense directional selection, to have elevated levels of asymmetry. This appears to be the case in some comparative studies (Moller 1992b; Moller and Hoglund, 1991), but not in others (Balmford et al., 1993; Tomkins and Simmons, 1995). This

DEVELOPMENTAL INSTABILITY AND STRESS

191

apparent discrepancy may be explained by differences in the patterns of asymmetry between species in which there is currently a female mate preference for the most ornamented traits and those in which there is no such preference (Moller, 1993b). A number of different studies have investigated in specific studies or experiments whether secondary sexual characters are more susceptible to the negative effects of stress, and this appears generally to be the case (review in Moller and Swaddle, 1997). Animal and plant breeding results in intense directional selection during a large number of generations to achieve preferred phenotypes. We know from the appearance of plants and animals that there have been dramatic responses to selection as determined from the extreme variance in phenotypes (Darwin, 1868). Just attend an exhibition of cats, dogs, or poultry and the diversity of morphology will become apparent. We should therefore expect to find considerably more asymmetry in these domesticated forms than in their wild ancestors. This appears to be the case. It is well known to pet breeders that abnormal numbers of digits appear at a frequency much higher than in free-living populations. Similarly, domestic strains of chickens are on average considerably more asymmetric than their wild ancestors in the jungles in Southeast Asia (Moller et al., 1995a). The null expectation might reasonably be the opposite, as free-living jungle fowl, Callus gallus, necessarily must be more severely restricted by limited access to food and more frequent exposure to debilitating parasites. A confounding factor is that many domesticated animals and plants may have suffered from the negative effects of bottlenecks and inbreeding, factors that are known to result in increased morphological asymmetry. A more reliable source of information is the large number of laboratory selection experiments performed over the years. This literature is reviewed in Moller and Swaddle (1997). Directional and disruptive selection experiments generally result in an increase in asymmetry, while stabilizing selection has the opposite effect of reducing asymmetry. Since most of these experiments are performed in ways that avoid inbreeding, and as they usually only last relatively few generations, before any depletion of additive genetic variance has taken place, this provides the most firm evidence for the relationship between mode of selection and developmental instability. In conclusion, there is evidence from a number of different sources suggesting that characters that have diverged because of a recent history of directional selection are developmentally less stable than characters that have been subject to a history of stabilizing selection. This observation implies that the former kinds of traits may be more suitable for determining the effects of adverse genetic and environmental conditions on developmental instability.

192

ANDERS PAPE M 0 L L E R

OF DEVELOPMENTAL INSTABILITY IV. FITNESSCORRELATES

If performance generally depends on developmental stability, then one should predict individual fluctuating asymmetry to be a reliable predictor of fitness. Indeed, this appears to be the case. First, asymmetrical individuals suffer more from intra- and interspecific competition and have an elevated probability of becoming parasitized and falling prey to a predator. Second, sexual selection and mating success in particular have been shown persistently to depend on morphological asymmetry. Finally, other fitness components such as growth performance, clutch size, offspring survival, and adult survival have been shown often to be inversely related to asymmetry. Ecological and behavioral studies of interactions usually depend on the ability to identify individuals differing in their ability to perform well in an interaction. Asymmetric individuals indeed appear to perform less well than symmetric ones in a number of different kinds of interactions. Parasitism differentially affects asymmetric hosts, and this is caused by greater susceptibility in at least some cases (Moller, 1996~). Predation also affects asymmetric indivduals differentially in organisms as diverse as domestic flies, Musca domestica, preyed upon by dung flies, Scatophaga stercoraria, and by barn swallows, Hirundo rustica, and barn swallows being preyed upon by European sparrowhawks, Accipiter nisus (Moller, 1996d; review in Moller and Swaddle, 1997). Finally, intraspecific and interspecific competition has been shown in a few cases to affect asymmetry. An experimental study of foliar asymmetry in clones of poplars showed that both intraspecific competition as determined from three levels of population density, but also interspecific competition as determined from absence or presence of an herb layer increased leaf asymmetry (Rettig et al., 1997). Given these negative relationships between asymmetry and performance in interactions, it is perhaps not surprising that the fitness of asymmetric individuals generally appears to be reduced. Sexual selection results from nonrandom variance in mating success being associated with particular phenotypes. A general finding is that females of a wide variety of species prefer males with more extreme phenotypes. Male secondary sexual characters that are larger and brighter, vocalizations that are louder, and pheromones that are more powerful result in higher mating success (Andersson, 1994). Because secondary sexual characters have been subject to a recent history of intense directional selection (as stated earlier), they often have elevated levels of fluctuating asymmetry. These asymmetries are often the direct target of sexual selection, as demonstrated by a large number of observational and experimental studies. A recent metaanalysis of developmental instability and sexual selection based on 61 studies of 36 species of animals revealed compelling evidence for an intermedi-

DEVELOPMENTAL INSTABILITY AND STRESS

193

ately sized effect of asymmetry on mating success (Moller and Thornhill, 1997~).An extremely large number of unpublished, negative results would have to exist in order to nullify this effect. There was considerable statistical heterogeneity in the data, and some of this could be explained by (1) lack of tests for fluctuating asymmetry and measurement errors, (2) poor experimental design that resulted in unwanted side effects of the manipulation, (3) experimental studies demonstrating stronger effects, apparently because experiments control potentially confounding variables, and (4) asymmetry in secondary sexual characters generally showing stronger effects than asymmetry in ordinary morphological traits (Moller and Thornhill, 1997~).There is also evidence suggesting that asymmetry plays an important role in pollinator preferences and sexual selection in plants (Moller and Swaddle, 1997). A number of other fitness components have also been shown to be associated with morphological asymmetry (Moller, 1997). A review showed that 10 of 12 studies found increased growth performance, 16 of 17 studies found increased fecundity, and 19 of 21 studies found increased survival rates of the more symmetric individuals (Mdler, 1997). These findings are particularly remarkable, given the different methods of study and the observational nature of the approach in most studies. As developmental instability appears to have a heritable component (Mdler and Thornhill, 1997a,b), asymmetric parents should also on average produce relatively asymmetric offspring. A review has demonstrated that viability selection often acts against asymmetric gametes, embryos, and juvenile individuals, and that parents may use developmental selection against offspring with deviant phenotypes as a way of allocating resources to viable offspring (Mgller, 1997b). Such developmental selection against asymmetric phenotypes is widespread in both animals and plants. The final part of this chapter considers the potential uses of developmental instability as a tool in a number of different contexts. Behavioral biologists will find information of interest for their own field of research.

V. PRACTICAL USESOF DEVELOPMENTAL INSTABILITY

With the information on the causes and consequences of developmental instability provided in the previous paragraphs in mind, we can start considering how this kind of information about the adaptation of individuals to their natural environment can be used. A few examples are provided here and more can be found in Mdler and Swaddle (1997). The topics covered include (1) environmental monitoring, (2) conservation biology, (3) animal welfare, (4) human and veterinary medicine, and ( 5 ) behavioral studies.

194

ANDERS PAPE M0LLER

A. ENVIRONMENTAL MONITORING Large numbers of biologists are employed to monitor the environment for regional, national, and international bodies. Vast sums of money are spent on monitoring environmental quality, which is usually done by simply determining measures of quality such as the presence or abundance of socalled indicator species as well as the pollution levels of air and water. When indicator species disappear, this is often a sign of severe, irreversible damage. We would often like to interfere before this state has been reached. Sublethal effects of environmental deterioration would therefore often be preferred. How should we proceed if we want to know how animals and plants perceive their environment? Measures of developmental instability become elevated well before severe effects on fitness components appear (e.g., Graham et al., 1993c), and asymmetry and similar measures may therefore be useful early indicators of the level of stress in the environment as experienced by free-living animals. Alternatively, the same strain of laboratory animals such as fruitflies or mosquitoes could be used to assess the environmental conditions across a range of sites. This would allow testing of environmental conditions with homogeneous strains, but also allow the use of particularly susceptible strains for monitoring. For example, the shaker mutant of Drosophila melanogaster is particularly susceptible to environmental stress because of the high activity level of flies with this mutation. They may for this reason be particularly useful for assessment of the effects of pollutants on developmental instability (Parsons, 1991). The sublethal effects of pollutants on phenotypic expression is an aspect of utmost importance. Because there is usually strong natural and sexual selection against individuals with asymmetric phenotypes (see Section IV), severe negative effects of pollution may appear to become hidden or even absent if asymmetric individuals are selectively removed from the population. There is a relatively large number of studies of the effects of a range of different kinds of pollutants on the asymmetry of plants and animals. The pollutants range from heavy metals and organic compounds to electromagnetic radiation and radioactivity (review in Maller and Swaddle, 1997). Early work on Arabidopsis thaliana revealed that radiation caused an increase in developmental instability (Bagchi and Iyama, 1983). Further examples are now available from Chernobyl in Ukraine. Maller (1993a) investigated the relationship between morphological asymmetry and exposure to radiation in the barn swallow. Individuals were captured and measured in a contaminated area near the nuclear power plant and in an uncontaminated control area southeast of Chernobyl. Samples predating the contamination accident were obtained from museum collections for both areas. There was

DEVELOPMENTAL INSTABILITY A N D STRESS

195

a statistically significant increase in asymmetry in tail length (a secondary sexual character) in the Chernobyl area after the contamination, but not in other characters in males or in any characters in female barn swallows. This result shows that there was a differential effect on the secondary sexual character. A subsequent study five years later confirmed these results and demonstrated increased mutation rates in swallows from Chernobyl (Ellegren et al., 1997). Asymmetry in three species of plants (black locust tree, Robinia pseudoacacia, rowan, Sorbus aucuparia, camomile, Matricaria perforata) sampled along a gradient from Chernobyl toward uncontaminated areas in Southeast Ukraine revealed elevated asymmetries near Chernobyl (Moller, 1997a). The asymmetries of the three species were concordant across the gradient, indicating that the asymmetry was responding to the same environmental conditions, and asymmetries were positively correlated with radiation from a radioactive isotope of cesium ('37Cs) in soil samples. A number of studies have determined the effects of less severe pollutants on the stable expression of phenotypes. For example, heavy metal pollution is common around large melter factories in various parts of the world. Leaf asymmetry in two species of birch Betufa was severely elevated near such sources of pollution in Finland and Russia, and the degree of foliar asymmetry was directly related to the concentration of the pollutant (Kozlov et al., 1996). Additional studies around a range of different kinds of chemical factories in Russia have documented similar effects on different measures of developmental instability in plants (Freeman et al., 1993). A number of studies concerning animals are mentioned in Mraller and Swaddle (1997). In conclusion, a diversity of organisms respond to sublethal exposure to pollutants by developing increased asymmetries. The effects of pollutants may thereby be assessed directly without using the traditional LD50 criterion (the dose at which 50% of a population has died), and the sublethal effects may also be more consistent with the generally accceptable levels of exposure to pollutants. B.

CONSERVATION BIOLOGY

Conservation biology is concerned with the factors that determine the sustainability of viable populations of animals in their natural habitats (e.g., Meffe and Carroll, 1994). Species are threatened for a number of different reasons including habitat destruction and other kinds of human activity as well as reductions in genetic variability of populations. These factors can be considered to result in a deterioration of the environment as perceived by the organisms in question. Although some species are better able to cope with environmental deterioration than others (Parsons, 1994), a number of different environmental and genetic factors are directly associated with

196

ANDERS PAPE M0LLER

increased stress. Again, the argument is the same as that already developed at the start of this chapter: Different kinds of stress give rise to a deterioration of the energy balance of an individual, and may result in poor developmental control independent of whether this is due to a reduction in the availability of food, shelter, or other kinds of vitally important resources. A reliable and objective insight into the ways that the environment is perceived by organisms can be gained directly from their individual fluctuating asymmetries or other measures of developmental instability. One of the first studies to adopt this approach was based on dental asymmetry in gorillas Gorilla gorilla (Manning and Chamberlain, 1993). The habitats of gorillas have suffered from continuous destruction with severe reductions in the living conditions of large parts of the populations. If secondary sexual characters are particularly susceptible to stress, for the reasons stated previously, we should especially expect to see an increase in the degree of asymmetry of gorilla canine teeth, but less so for other kinds of teeth. This was exactly the pattern that was found. Asymmetry in the sexually size-dimorphic canines has increased considerably since the beginning of the last century; this is not the cause for sexually sizemonomorphic teeth. This study suggests that the living conditions of gorillas indeed have deteriorated during the last 150 years. Large proportions of threatened animals are currently found in national parks throughout the world, and they are superficially safe from threats that otherwise may cause declines and extinction of less well protected populations. Park populations are often sold or culled because of rapid increases in numbers, and decisions have to be made concerning which animals to remove. These decisions are not easily made on reasonable scientific grounds. One possibility that has not been considered is that individuals with symmetric phenotypes may have characteristics that allow them to cope better with stressful conditions. Selective culling of asymmetric individuals would actually be indirect selection for increased stress resistance. A study addressing this question concerns the gemsbok, Oryx gazella, a large antelope confined to very dry habitats such as dry savannahs and deserts in southern Africa (Moller et al., 1996). Gemsboks have long, lancelike horns that are used for interactions and antipredator defense. Adults of both sexes were in better condition if they had symmetrical horns. Fights between individuals of the same sex with similarly sized horns were also most often won by individuals with symmetric horns, which therefore had differential access to limiting resources such as drinking water. Adult females more often had a calf if their horns were symmetric, and males with symmetric horns more often had access to females than did the flock-living males with asymmetric horns. A couple of gemsbok eaten by predators all had asymmetric horns. These observations suggest that symmetrical

DEVELOPMENTAL INSTABILITY AND STRESS

197

individuals indeed were more fit than asymmetrical individuals, and that symmetrical individuals would contribute more disproportionately to the maintenance of a growing population than would asymmetric ones. A final example concerns a number of different species of butterflies that have become threatened and in many cases have disappeared from large areas during recent decades. Poulsen (1997) investigated whether butterfly species that were currently threatened in Denmark differed in their wing asymmetry from closely related, common species. This pairwise comparative approach helped control for a number of different factors that potentially could affect developmental instability. Butterflies that were currently threatened had considerably higher degrees of asymmetry than their sister species (Poulsen, 1997). A second comparison determined whether there had been a temporal increase in asymmetry in the threatened species. Again, there was evidence for a significant increase in wing asymmetry in the threatened species, but not in the common sister species. Although the direct cause of the asymmetry cannot be pinpointed, this study provided evidence for fluctuating asymmetry being a reliable predictor of future conservation status. In conclusion, conservation biological studies may benefit from the use of measures of developmental instability for assessment of how organisms perceive their environment. A number of other ways in which measures of developmental instability can be of use in the context of conservation biology are discussed by Clarke (1995), Sarre et al. (1994), and Moller and Swaddle (1997).

C. ANIMAL WELFARE Scientific questions of animal welfare consider ways in which to decide objectively about the state of animals and ways in which to ameliorate poor conditions. A number of different solutions to these problems of welfare have been suggested (e.g., Broom and Johnson, 1993; Toates, 1995), but none of these approaches has satisfied the farmer community and the decision makers. The reason is that there are no objective, a priori ways of determining whether a specific criterion for rearing animals will improve conditions in any appreciable way. Common measures of stress such as behavioral or physiological variables are themselves subject to selection during the domestication process, and they may not provide reliable information about the welfare state of animals. Developmental instability can be used as a direct measure of how animals and plants perceive their environment, since we know a priori the optimal phenotypic solution to the engineering problem of constructing a well-functioning organism; this is a symmetric phenotype. Deviations from perfect symmetry can therefore

198

ANDERS PAPE MOLLER

be used to objectively assess environmental conditions in whichever way is of current interest. The optimal solution is a very low level of asymmetry, and if a particular environmental gradient is suspected to cause increased levels of stress, this can be assessed directly from the relationship between asymmetry and increased values of the environmental variable. A concrete example of this approach can be found in a study of chicken asymmetry in relation to rearing density (MGller et al., 1995a). Chickens of two different breeds were reared at three different densities of 20, 24, and 28 chickens per square meter (the latter being the normal density in commercial chicken farms), and their level of skeletal asymmetry was assessed when slaughtered. There was a considerable increase of on average 30% in the degree of asymmetry across this range of densities (Fig. 2 ) . The next step in this line of research would obviously be to extend the relationship to even lower densities until a minimum has been found for the curve relating asymmetry to density. Of course, this approach is not restricted to density alone, and other conditions of rearing can be investigated as well. In a second study, Mdler et al. (1998) investigated the relationship between light regime and skeletal asymmetry in chickens. Commercial chicken farmers expose their chickens to continuous light, but we managed to manipulate the light to some extent: (1) continuous light, ( 2 ) a changing light regime, (3) a 16 :8 hours light cycle with the exception of the first and last days of life of the chickens; the lack of a strict difference in light regime was caused by constraints imposed by the chicken farmers (MGller et al., 1998). Chickens

20

24

28

Density (inds. per square meter) FIG. 2. Skeletal fluctuating asymmetry in relation to rearing density of chickens of the breeds ScanBrid and Ross 208. Values are means (SE). Adapted from Moller et al. (1995).

DEVELOPMENTAL INSTABILITY AND STRESS

199

reared in continuous light had an average 40% larger asymmetry than chickens reared in the two other treatments, again suggesting that continuous light imposed significant amounts of stress on the chickens. A final example concerns cow asymmetry and milk production (J. T. Manning, pers. comm.). Cows were measured on a number of skeletal traits for asymmetry, and both the single asymmetry values and a composite index of overall asymmetry were found to be negatively related to milk production and the quality of milk as determined by the dairy companies. This result differs to some extent from the studies of chickens because the conditions that will improve animal welfare (a reduction in the stress factors causing fluctuating asymmetry), will also improve productivity. Hence, in this case the interests of people studying animal welfare and farmers are congruent. In conclusion, problems of animal welfare arise from the fact that we cannot make the theoretical inferences about the optimal conditions under which animals are reared. This problem can be resolved if measures of fluctuating asymmetry are used as a way of determining the conditions under which morphological asymmetry reaches a minimum level. These conditions will reflect the environment in which a specific animal is reared with a minimum amount of stress.

D. HUMAN AND VETERINARY MEDICINE The practice of medicine has for a long time been isolated from evolutionary theory, and several evolutionary biologists believe that progress has been prevented by this lack of scientific knowledge about the interactions between pathogens and human hosts. Darwinian medicine based on evolutionary theory may be a way of resolving these problems (Nesse and Williams, 1995). Similar arguments can be raised for veterinary medicine. Since measures of developmental instability provide information about the developmental state of individuals, this information may be useful for understanding interactions between parasites and their human hosts, but also for making inferences about the current health status of individuals (Thornhill and Mgller, 1997). A number of examples of this approach are listed in Thornhill and Mgller (1997) and Mgller (1996~).Here I will provide only a couple of examples. A number of different kinds of cancer associated with the use of contraceptives have increased dramatically in frequency during recent decades, and they are presumably more abundant as a consequence of dramatic hormonal fluctuations experienced by regularly menstruating women (Eaton et al., 1994). Breast cancer is one of these now common kinds of reproductive cancers of women, and large sums of money have been used

200

ANDERS PAPE MOLLER

to identify predictors of susceptibility to cancer. Preliminary studies showed that breast asymmetry was a relatively important predictor of low lifetime fecundity and hence potentially of exposure to repeated surges of estrogen and other kinds of reproductive hormones (Moller et al., 199%; Manning et al., 1996). A large data set of mammograms from a hospital was used to determine whether breast asymmetry was a reliable predictor of breast cancer (Scutt et al., 1997). Previous work had indicated that a number of factors associated with increased exposure to estrogens such as high body mass, large breasts, and repeated menstrual cycles were correlated with elevated risks of acquiring cancer. However, breast asymmetry proved to be an even better predictor of breast cancer than any of the previous variables (Scutt et al., 1997). This has important implications for prevention and treatment of breast cancer. It is possible that a number of cases of cancer can be prevented simply by asking women with elevated breast asymmetry to report for more regular checks than other women. Another possibility is to attempt to reduce exposure to estrogens and thereby reduce the risks of breast cancer. Cancers resemble to some extent developmental instability in the sense that both phenomena are the result of uncontrolled growth processes. If the lack of control of the two types of growth processes has a similar cause, measures of developmental instability may be useful for predicting other kinds of cancer. A large number of studies have investigated the relationship between developmental instability and parasitism in a wide variety of plants and animals (review in Moller, 1996~).Asymmetric individuals throughout this range of organisms usually have higher parasite loads than do symmetric individuals, although the causal relationship between symmetry and parasitism is not known in most of these cases. Parasitism was the cause of increased asymmetry in studies of elms Ulmus glabra, fruitflies Drosophila nigrospiracula, barn swallows, and reindeer Rangifer tarandus (Moller, 1992a, 1997b; Polak, 1993, 1997; Folstad et al., 1996). Studies of elms, fruitflies D. nigrospiracula, domestic flies, barn swallows, and humans have found increased susceptibility of asymmetric individuals (Fig. 3; Moller, 1992a, 1996d, 1997b; Polak, 1993, 1996; Shapiro, 1992). This might have important implications for prevention of disease, but also for treatment, and selection for disease resistance in domesticated organisms. If the reduced resistance of certain individuals is caused by a low level of stress resistance in general, and thereby loss of energy that could otherwise be used for coping with parasite attacks, then selection for increased stress resistance might be an important management tool. In conclusion, a number of areas in human and veterinary medicine are likely to benefit from the use of information on developmental instability

DEVELOPMENTAL INSTABILITY A N D STRESS

Uninfected

201

Infected

Infection status FIG. 3. Wing and tibia asymmetry in male domestic flies Musca domestica that acquired or did not acquire a fungus infection after exposure to fungal spores. Values are means (SE). Adapted from Moiler (1W6d).

at the level of individuals. Future research will decide the extent to which this approach will be of general use. MEASURES OF DEVELOPMENTAL INSTABILITY E. BEHAVIORAL In this section on behavioral measures of developmental instability I adopt two approaches. First, I briefly review behavior as affected by deviant morphology. Second, I suggest that phenodeviant behavior also can be considered a measure of developmental instability. 1. Behavioral Invariance

The approach adopted for morphological developmental instability is to consider deviations from the morphological invariant to reflect instability. In a similar way, behavior also goes through an ontogenetic phase after which a behavioral phenotype is developed. If this phase of ontogeny is disturbed by mutations, an inability to learn, or by exposure to deviant models from which a behavioral pattern can be learned, then this will result in a deviant behavioral phenotype. The main problem for behavioral studies, as for morphological ones, is to identify the invariance against which phenotypes can be compared. For morphology this is bilateral or radial symmetry against which deviations can be determined. Irregular behavior can be considered phenodeviant with respect to modal behavioral

202

ANDERS PAPE M0LLER

patterns. An explicit model for this approach was developed by Markow and Gottesman (1993) for schizophrenia, a mental disease with a presumed polygenic threshold inheritance. Many behavioral patterns have a quantitative genetic inheritance with heterozygotes displaying modal phenotypes and homozygotes exhibiting extreme behavior. High levels of homozygosity were presumed in the model to result in poor developmental stability, which would be reflected not only in elevated asymmetry of the morphological phenotype, but also in deviant behavior and phenodeviance with respect to the anatomy and biochemistry of the central nervous system. Some support for this model was obtained for schizophrenia, but also for abnormal behavior in Drosophila. Many behavioral patterns are incredibly repetitive, particularly when it comes to display and signals. Typical examples of such repetitiveness is the same song being sung over and over again by a bird during an entire breeding season and the same courtship display being performed repeatedly by a male during a reproductive season. The function of this repetitiveness could be that it reveals behavioral phenodeviance; only individuals that are able to repeatedly perform the same behavior over and over again can be considered to possess a developmentally stable phenotype. A measure of the ability to perform such repetitiveness can be considered a behavioral invariant (MGller and Swaddle, 1997). One measure of the ability to perform a behavior in a similar way is the repeatability of the activity. Repeatability is a measure from quantitative genetics of the variance among individuals in relation to a measure of the variance within individuals (Falconer, 1989). Since the repeatability is a measure of the genotypic variance and the general environmental variance relative to the entire phenotypic variance, it also has the property of being an upper estimate of the heritability of a trait (Falconer, 1989). Repeatabilities vary from 0 to 1 with low values reflecting extreme variation within as compared to among individuals, and values of 1 reflecting a complete ability to repeat the same behavior again and again. Repeatabilities have been used in studies of behavior for some time in various contexts (Boake, 1989), but never as a measure of developmental instability. A high value of repeatability can be considered the developmentally stable state of a behavioral parameter, while lower values are unstable states. The use of repeatability of behavioral traits as a measure of developmental instability also has the advantage of different traits being directly comparable because of the range of repeatabilities from 0 to 1. I would predict that repeatability would be particularly high for behavioral traits that are closely related to fitness, while less important behavior will have lower repeatabilities. For example, alarm calls will undoubtedly have high repeatabilities, while contact calls will have lower repeatabilities.

DEVELOPMENTAL INSTABILITY AND STRESS

203

This argument parallels that for morphological developmental instability, which appears to be inversely related to the functional importance of a morphological character. An important exception to this prediction will be behavioral traits that have been subject to intense directional selection. These are predicted to be developmentally relatively unstable (see Section 111). Of course, it is important when estimating repeatabilities, for example, in the behavior of an individual on different days, that comparisons among individuals are made while controlling for the time scale of sampling, the reproductive state of the individual, and other potentially confounding factors. The kinds of behavioral traits that can be evaluated with respect to repeatability are limited only by the imagination of the scientist; examples include duration, interval, frequency, volume, and complexity. A possible behavioral example of this approach, although not couched in terms of repeatability per se, is a study of drift in the song of the great tit, Parus major (Lambrecht and Dhondt, 1988a,b). Great tit males repeat a song many times during a song bout, but if the frequency of the song (measured in kHz) is estimated, males differ considerably in their ability to maintain a consistently high frequency of their songs. The reduction during a song bout in the frequency of the song, which is termed drift, can be considered an indirect measure of repeatability. Interestingly, male great tits that are able to sing without experiencing considerable drift have higher reproductive success than other males. Obviously, the underlying mechanism that generates differences in drift may well be morphological asymmetries that result in muscle fatigue being reached at an early stage. Another suggestion for a behavioral invariance is the fractal dimension of behavioral sequences (Esc6s et al., 1995). Fractals reflect a measure of the scaling constant that describes the relationship between size of a character and the scale of measurement (Hastings and Sugihara, 1993). For example, the length of a coastline depends on the ruler that is used for measuring the coast; as the ruler decreases in size, the length of the coastline increases, although the pattern of the coastline is always the same. Anybody can convince themselves of this fact by taking a map of an arbitrary archipelago like the British Isles. Start out by placing a transparent grid with side length of say 10 cm (this choice is not really important) on the map and count the number of squares that include the coastline. Then redo this using increasingly smaller squares with a side length of 5 cm, 2.5 cm, 1.25 cm, and .625 cm. The regression coefficient relating the number of squares necessary to cover the entire coastline to the size of the squares is an unbiased estimate of the fractal dimension. Esc6s et al. (1995) used this approach t o determine the fractal dimension of foraging and scanning bouts in Spanish ibex, Capra pyrenaica. As is typical for such sequences of behavior, periods of foraging are interspersed randomly with periods of

204

ANDERS PAPE MBLLER

scanning for predators or competitors. Any bout of foraging and scanning measured at a particular time scale resembles a bout at another time. This constancy is described by the fractal dimension, which was recorded to be 1.15 in the present case. Two groups of Spanish ibex were compared: Individuals affected by scabies (an ectoparasitic disease) and healthy individuals. Disease status changed the fractal dimension from 1.15 to .94. The difference in fractal dimension between the modal phenotype and the alternative phenotype can be considered a measure of developmental instability. This conclusion obviously depends on the certainty with which the baseline fractal dimension has been determined. While symmetry a priori can be considered the optimal phenotype, it is not straightforward to make the same claim for the fractal dimension of a particular behavioral pattern. In fact, the baseline fractal dimension will depend enormously on the spatial and temporal homogeneity of the sample obtained.

2. Behavioral and Morphological Developmental Instability All behavior has at least two morphological bases; the morphology that is used directly in the production of the behavior and the neural tissue used for performing the behavior. Developmental instability in both these morphological bases may have important consequences for the regularity of the behavioral output. Even small morphological differences in any of these morphological bases may result in large behavioral differences. Just a small asymmetry in the size of a paired, bilaterally symmetrical muscle may after considerable repetition of a behavior translate into large differences in behavioral performance. The study of the morphological basis of phenodeviant behavior is still in its infancy. Early studies of morphological asymmetries demonstrated experimentally that there were direct effects on locomotion. This was the case for asymmetry of tail and wing feathers of birds (Mflller, 1991; Evans et al., 1994; Swaddle et al., 1996). These studies demonstrated what one might predict from aerodynamic theory. Perhaps the question of whether morphological asymmetries affect other kinds of behavior differs only in degree. The most well studied example of behavioral asymmetry is vocalizations. Vocalizations play an important role in animal communication and there is ample evidence for considerable intraspecific variation in such calls. The first study addressing the question of quality of calls and morphological asymmetry dealt with mating calls of the field cricket, Cryflus campesfris (Simmons and Ritchie, 1996).Males produce a mating call with an unpaired structure called the harp, and the level of asymmetry in this morphological structure translates directly into perceivable differences in calls. Males with more asymmetrical harps produce calls that are more attractive to females

DEVELOPMENTAL INSTABILITY AND STRESS

205

in phonotaxis experiments. Given the inherent asymmetry of the harp, and the potential disadvantages in terms of natural selection of such asymmetry, asymmetry can be maintained only by an oppositely directed selection force such as sexual selection. A similar coupling between morphological asymmetry and call characteristics was found in the oilbird, Steatornis caripensis, although the functional importance of these differences in vocalizations was not studied (Suthers, 1994). Although these few studies all concern calls, there is no a priori reason why similar principles may not apply equally well to other modes of communication. The second kind of morphological basis of behavior is neural anatomy. Nerve cells and brains do not differ quantitatively from any other structure with respect to their development, and deviations from regular neural phenotypes will invariably affect the ways and the efficiency with which various kinds of behavior can be performed. Stresses of various kinds are predicted to severely affect the regulated development of the nerve system, and similar kinds of environmental and genetic stresses that affect the stable development of the ordinary phenotype will also affect the stable development of the neural system. Deviant neural systems will arise from insults to the stable development of the phenotype during ontogeny. There is already evidence for this prediction from the relationship between genetic deviants and the functioning of the brain in humans affected by chromosomal abnormalities and other genetically based diseases (Thornhill and Mdler, 1997). I would also predict that deviant neural systems will give rise to behavioral phenodeviance; a prediction that still needs to be rigorously tested. Recent progress in the study of developmental instability at the neural level has important implications for these kinds of predictions. Morphological asymmetries in the human brain were determined from magnetic emission scanning (MES) and magnetic resonance imaging (MRI) studies in a sample of subjects from a university in the United States (Thoma, 1996). Asymmetries in various components of the brain were subsequently correlated with body asymmetries that were recorded blindly (i.e., without knowing the identity of the subjects with respect to the brain scans). There were relatively strong correlations between brain asymmetry and body asymmetry, indicating that the same factors responsible for the development of fluctuating asymmetry in the skeleton also caused asymmetry in brain structures (Thoma, 1996). This result can be interpreted in a number of different ways. One possibility is that nerve cells play an important role in stable morphogenesis of the phenotype, and that this causes covariation in asymmetry in brains and the skeleton. A second study investigated this relationship further by testing whether the performance of the brain as estimated from a so-called “culture free”

206

ANDERS PAPE M0LLER

version of a standard IQ test was directly related t o body asymmetry (Furlow et al., 1997). IQ tests have been severely criticized for not providing reliable estimates of the intelligence of individual humans. However, a recent study has demonstrated that there is a statistically significant positive correlation between IQ and pH in the brain, suggesting that the IQ score provides an estimate of a trait that has a physiological basis (Rae etaf.,1996). Again, the data were collected blindly, so there was no prior knowledge of the IQ of subjects that were assessed for fluctuating asymmetry, and vice versa. There was a significant negative relationship between IQ and body asymmetry, suggesting that asymmetry on the outside of people was directly correlated with asymmetry on the inside. This result is based on a correlation and the causation can go either way; or, perhaps most parsimoniously, the stable development of the phenotype may be controlled by growth processes controlled by neural signals. Any deviant neural anatomy may subsequently give rise to a deviant body morphology and as well a deviant neural system. A number of potentially confounding factors that are known from previous studies to be correlated with IQ were controlled statistically, but did not affect the relationship. These results are potentially of general biological interest and should be investigated experimentally in other organisms. The potential for a better understanding of the functioning of the brain and the neural underpinnings of behavior is certainly present. In conclusion, behavioral phenodeviance can be estimated when the invariance of a behavioral pattern has been determined. Such measures of behavioral invariance are the repeatability of a behavioral pattern and the fractal dimension of a behavior. Behavior has morphological bases in the structures that are involved in production of the behavior, but also in the neural underpinnings of behavior. Developmental instability in behavior may simply arise from developmental instability in either of these morphological bases. Several important advances have been made recently in attempts to investigate the relationship between behavior and developmental instability. There are many possibilities for further developments, as suggested in the following section.

AND PROSPECTS FOR FUTURE STUDIES VI. CONCLUSIONS

Measures of developmental instability are fascinating in many different ways. A particularly important point is that they are easy to obtain and do not require any especially sophisticated and expensive equipment. The generality of the concept also implies that there are no obvious limits to the applications. Any aspect of biological phenomena can in one way or another be linked either directly or indirectly to the degree of stability in

DEVELOPMENTAL INSTABILITY AND STRESS

207

the development of the phenotype. Measures of developmental instability such as fluctuating asymmetry are unique in the sense that the optimal phenotype is known a priori; it is the symmetric phenotype. This makes developmental instability different from any other phenotypic character. Finally, measures of developmental instability, because of their strong correlations with fitness components such as growth, mating success, fecundity, and survival, provide the most readily accessible measure of fitness that we usually can achieve under natural conditions for free-living organisms. Behavioral biologists could contribute t o further development of the developmental instability approach to biology in a number of different ways. An obvious starting point would be to consider Tinbergen’s (1951) four classical approaches to the study of ethology: ontogeny, mechanism, function, and evolution. I will not provide extensive lists of studies that could be done, but only suggest a few areas that might prove particularly fruitful. There is a deficiency of studies, particularly experimental ones, investigating the relationship between morphology and behavior. I have reviewed three studies considering the effects of morphological asymmetries on signals in the auditory domain. When measures of developmental invariance regularly become adopted in studies of behavior, we may start investigating the relationship between morphology and behavior. This approach will also open up for studies of the link between behavioral developmental instability and the consequences of behavior in terms of fitness. Developmental instability of behavior at different levels that subsequently result in mating success and reproductive success (from display and handling of mates to copulation, fertilization, and parental care) can potentially be investigated as a series of behavioral events that lead to the fitness of the individual. The previous section dealt with the association between brain asymmetry and body asymmetry in humans. Future studies of neuroethology may benefit considerably from considering measures of developmental instability of the neural system, but also of the phenotype in general, as a means of exploring the association between the environment and the way in which the brain functions. Learning is a central theme in ethology, and the mechanisms resulting in the learning of particular tasks, and the variability among individuals in this ability, are of great theoretical and practical importance. Learning inability among humans has in several studies been associated with developmental instability (review in Thornhill and Mflller, 1997). It is possible that fluctuating asymmetry may provide a phenotypic marker in future studies of learning. Recent studies of humans have also indicated that various measures of psychological stability and mood are directly associated with body fluctuating asymmetry (Shackelford and Larsen, 1997). It remains to be determined whether similar relationships exist among other animals.

208

ANDERS PAPE MBLLER

VII. SUMMARY Developmental stability reflects the ability of organisms to buffer their developmental trajectories against disturbance. An inability to fulfill this predetermined goal can be assessed at the phenotypic level in terms of fluctuating asymmetry, the frequency of phenodeviants, o r other measures of developmental instability. Such random deviations from perfect asymmetry are related to a number of different kinds of deviant environmental and genetic factors. Organisms are generally believed to be adapted to the most commonly encountered environmental conditions, and deviations from such optimal conditions result in energy being spent on maintenance. The control of growth processes is energetically costly, and since the total amount of energy available has to be allocated to either maintenance, growth, storage, or reproduction, a larger fraction of the total energy budget allocated to maintenance reduces the amount available for developmental control. A range of environmental stresses such as food deficiency, pollutants, parasites, and deviant temperatures gives rise to elevated developmental instability. Similarly, a range of genetic factors such as mutations, inbreeding, and hybridization increases developmental instability. In other words, measures of developmental instability provide an integrated estimate of the quality of the environment of an individual with a given genetic background, as experienced by the individual itself. Developmental instability may be particularly useful for studies of environmental monitoring, conservation biology, animal welfare, human and veterinary medicine, and behavioral studies, as shown by a large number of examples.

Acknowledgments I am grateful for constructive criticism provided by J. T. Manning, M. Milinski. and P. J. B. Slater. This paper was written while I was supported by the Danish Natural Science Research Council.

References Alekseeva, T. A,, Zinichev, V. V., and 2 0 t h . A. I. (1992). Energy criteria of reliability and stability of development. Acta Zool. Fenn. 191, 159-165. Andersson, M. (1994). “Sexual Selection.” Princeton University Press, Princeton, NJ. Bagchi, S., and Iyama, S. (1983). Radiation induced developmental instability in Arabidopsis tlialiana. Theor. Appl. Genet. 65, 85-92. Balmford, A., Jones, I. L., and Thomas, A. L. R. (1993). On avian asymmetry: Evidence of natural selection for symmetrical tails and wings in birds. Proc. R. Soc. London, Ser. B 252,245-251.

DEVELOPMENTAL INSTABILITY AND STRESS

209

Beardmore, J. A. (1960). Developmental stability in constant and fluctuating temperatures. Heredity 14, 41 1-422. Boake. C. R. B. (1989). Repeatability: Its role in evolutionary studies of mating behavior. Evol. Ecol. 3, 173-182. Broom, D. M., and Johnson, K. G. (1993). “Stress and Animal Welfare.” Chapman & Hall, London. Carmeliet, P., Ferrteira, V., Breier, G., Pollefeyt, S., Kieckens, L.. Gertsenstein, M., Fahrig, M.. Vandenhoeck, A., Harpal, K., Eberhardt, C., Declercq, C., Pawling, J., Moons, L., Collen, D., Risau, W.. and Nagy, A. (1996). Abnormal blood vessel development and lethality in embryos lacking single VEGT allele. Nature (London) 380,435-439. Clarke, G. M. (1995). Relationships between developmental stability and fitness: Application for conservation biology. Conserv. Biol. 9, 18-24. Clarke, G. M., and McKenzie, J. A. (1992). Fluctuatingasymmetry as a quality control indicator for insect mass rearing projects. J. Econ. Entomol. 85, 2045-2050. Darwin, C. (1868). “The Variation of Animals and Plants under Domestication.” John Murray. London. Darwin, C. (1871). “The Descent of Man, and Selection in Relation to Sex.” John Murray, London. Davis, A. P., Witte. D. P.. Hsieh-Li, H., Potter, S. S.. and Capecchi, M. R. (1995). Absence of radius and ulna in mice lacking hoxa-11 and hoxd-21. Nature (London) 375,791-795. Eaton, S. B., Pike, M. C . , Short, R. V.. Lee, N. C.. Trussell, J.. Hatcher, R. A,, Wood. J. S., Worthman. C. M., Blurton Jones. N. G.. Konner. M. J., Hill, K. R., Bailey, R., and Hurtado. A. M. (1994). Women’s reproductive cancers in evolutionary perspective. Q. Rev. Biol. 69, 353-367. Ellegren, H., Lindgren. G., Primmer, C. R., and Mgiller, A. P. (1997). Fitness loss and germ line mutations in barn swallows breeding in Chernobyl. Nafure (London) 389, 593-596. Escos. J. M., Alados, C. L., and Emlen, J. M. (1995). Fractal structures and fractal functions as disease indicators. Oikos 74, 310-314. Evans, M. R., Martins, T. L. F., and Haley. M. (1994). The asymmetrical cost of tail elongation in red-billed streamertails. Proc. R. Soc. London, Ser. B 256, 97-103. Falconer, D. S. (1989). “Introduction to Quantitative Genetics,” 3rd ed. Longman, New York. Folstad, I., Arneberg, P., and Karter, A. J. (1996). Antlers and parasites. Oecologia 105, 556-558.

Freeman, D. C.. Graham, J. H., and Emlen, J. M. (1993). Developmental stability in plants: Symmetries. stress and epigenesis. Generica 89, 97-1 1Y. Furlow, F. B.. Armijo-Prewitt, T.. Gangestad. S. W., and Thornhill, R. (1997). Fluctuating asymmetry and psychometric intelligence. Proc. R. Soc. London, Ser. B 264, 823-829. Gest, T. R.,Siegel, M. I., and Anistranski, J. (1983). Increased fluctuating asymmetry in the long bones of neonatal rats stressed in cold, heat, and noise. Am. J. Phys. Anthropol. 60, 196-197. Gest, T. R.. Siegel, M. I., and Anistranski, J. (1986). The long bones of neonatal rats stressed by cold, heat, and noise exhibit increased fluctuating asymmetry. Growth 50, 385-389. Goldschmidt, R. B. (1940). “The Material Basis of Evolution.” Yale University Press, New Haven, CT. Goldschmidt, R. B. (1 955). “Theoretical Genetics.” Cambridge University Press, Berkeley. CA. Graham, J. H., Freeman, D. C.. and Emlen, J. M. (1993). Developmental stability: A sensitive indicator of populations under stress. ASTM Spec. Tech. Publ. STP 1179, 136-158. Graham, J. H.. Roe, K. E., and West, T. B. (1993~).Effects of lead and benzene on the developmental stability of Drosophilu melunogaster. Ecotoxicology 2, 185-195.

210

ANDERS PAPE M0LLER

Hastings, H. M.. and Sugihara, G. (1993). “Fractals: A User’s Guide for the Natural Sciences.” Oxford University Press. Oxford. Hoffmann. A. A., and Parsons, P. A. (1989). “Evolutionary Genetics and Evironmental Stress.” Oxford University Press, Oxford. Houle. D. (1997). Comment on ‘A meta-analysis of the heritability of developmental stability’ by Msller and Thornhill. J. Evol. Biof. 10, 17-20. Kieser, J. A. (1992). Fluctuating odontometric asymmetry and maternal alcohol consumption. Ann. Hiim. Biol. 19, 513-520. Kozlov, M. V., Wilsey, B. J., Koricheva, J.. and Haukioja. E. (1996). Fluctuating asymmetry of birch leaves increases under pollution impact. J. Appl. Ecol. 33, 1489-1495. Lamhrecht. M., and Dhondt, A. A. (1988a). Male quality, reproduction, and survival in the great tit (Punts major). Behav. Ecol. Sociobiol. 19, 57-64. Lambrecht, M., and Dhondt, A. A. (1988b). The anti-exhaustion hypothesis: A new hypothesis to explain song performance and song switching in the great tit. Anim. Behav. 36,327-334. Leamy, L. (1997). Is developmental stability heritable? J. Evol. B i d . 10, 21-29. Lerner, 1. M. “Genetic Homeostasis.” Oliver & Boyd, London. Ludwig, W. (1932). “Das Rechts-Links Problem in Tierreich und beim Menschen.” SpringerVerlag. Berlin. Manning, J. T., and Chamberlain, A. T. (1993). Fluctuating asymmetry in gorilla canines: A sensitive indicator of environmental stress. Proc. R. Soc. London, Ser. B 255, 189-193. Manning, J. T., and Ockenden, L. (1994). Fluctuating asymmetry in racehorses. Nutiire (London) 370, 185-186. Manning, J. T., Scutt, D., Whitehouse, G. H.. and Leinsteb. S. J. (1996). Breast asymmetry and phenotypic quality in women. Efhol. Sociobiol. (in press). Markow, T. A,. and Clarke, G. M. (1997). Meta-analysis of the heritability of developmental stability: A giant step backward. J . Evol. Biol. 10, 31-37. Markow, T. A,, and Gottesman, 1. 1. (1993). Behavioral phenodeviance: A Lerneresque approach. Genetica 89,297-305. Meffe. G. K., and Carroll. C. R.. eds (1994). “Principles of Conservation Biology.” Sinauer, Sunderland, MA. Mitton. J. B. (199.5). Enzyme heterozygosity and developmental stability. Acfa Theriol., Siippl. 3, 33-54. Mitton. J. B., and Grant. M. C. (1984). Associations among protein heterozygosity, growth rate, and developmental homeostasis. Annu. Rev. Ecol. Syst. 15, 479-499. Msller, A. P. (1990). Fluctuating asymmetry in male sexual ornaments may reliably reveal male quality. Anim. Behav. 40, 1185-1187. Msller, A. P. (1991). Sexual ornament size and the cost of fluctuating asymmetry. Proc. R. Soc. London, Ser. B 243, 50-62. Moller, A . P. (1992a). Parasites differentially increase fluctuating asymmetry in secondary sexual characters. J. Eva/. B i d . 5, 691-699. Msller, A. P. (1992b). Patterns of fluctuating asymmetry in weapons: Evidence for reliable signalling of quality in beetle horns and bird spurs. Proc. R. Soc. London, Ser. B 248, 199-206. Msller. A. P. (1993a). Morphology and sexual selection in the barn swallow Hirundo rusficu in Chernobyl. Ukraine. Proc. R. Soc. London. Srr. B 252, 51-57. Moiler, A. P. (l993b). Patterns of fluctuating asymmetry in sexual ornaments predict female choice. J. Evol. Biol. 6, 481-491. Moller. A. P. (1997). Developmental stability and fitness: A review. Am. Nut. 149, 916-932. M@ller,A. P. (1997b). Developmental selection against developmentally unstable offspring and sexual selection. 1. Theor. B i d . 185, 415-422.

DEVELOPMENTAL INSTABILITY A N D STRESS

21 1

Moller, A. P. (1996~).Parasitism and developmental instability of hosts: A review. Oikos 77, 189- 196. Mprller, A. P. (lYy6d). Sexual selection. viability selection, and developmental stability in the domestic fly Musca domesticn. Evolution (Lawrence, Kans.) 50, 746-752. Moller, A. P. (1997a). Developmental instability of plants and radiation from Chernobyl. Oikos (in press). Moller, A. P. (1997b). Elm Ulmus gfabra leaf asymmetry and Dutch elm disease. Oikos (submitted for publication). Moller, A. P., and Hoglund, J. (1991). Patterns of fluctuating asymmetry in avian feather ornaments: Implications for models of sexual selection. Proc. R. SOC. London, Ser. B 245, 1-5. Moller. A. P.. and Pomiankowski, A. (1993a). Fluctuating asymmetry and sexual selection. Genefica 89, 267-279. M0ller. A. P., and Pomiankowski, A. (1993b). Punctuated equilibria or gradual evolution: Fluctuating asymmetry and variation in the rate of evolution. J. Theor.Biof. 161,359-367. Moller, A. P., and Swaddle, J. P. (1997). “Asymmetry, Developmental Stability and Evolution.” Oxford University Press, Oxford. Moller. A. P.. and Thornhill. R. (1997a). A meta-analysis of the heritability of developmental stability. J . Evol. Biof. 10, 1-16. Moller. A. P., and Thornhill, R. (1997b). Developmental instability is heritable. J . Evol. Biol. 10,69-76 Moller. A. P.. and Thornhill, R. (1997~).Bilateral symmetry and sexual selection: A metaanalysis. Am. Nut. (in press). Moiler. A. P., Sanotra, G. S., and Vestergaard, K. S. (1995a). Developmental stability in relation to population density and breed of chickens Gulfus galfus. Poult. Sci. 74, 1761-1771. Mdler, A. P., Soler, M.. and Thornhill, R. (199%). Breast asymmetry, sexual selection and human reproductive success. Erhof. Sociobiol. 16, 207-219. Moller, A. P., Cuervo, J. J., Soler, J. J., and Zamora-Muiioz, C . (1996). Horn asymmetry and fitness in gemsbok Oryx g. gazeffa. Behav. Ecol. 7,247-253. Moller. A. P., Sanotra, G. S., and Vestergaard, K. S. (1998). Developmental instability and light regime in chickens. Appl. Anim. Behav. Sci. (in press). Mulvey. M., Keller, G. P.. and Meffe, G. K. (1994). Single- and multiple-locus genotypes and life-history responses of Gamhusia holbrooki at two temperatures. Evolution (Lawrence, Knns.) 48, 1810-1819. Nesse, R. M., and Williams, G. C. (1995). “Evolution and Healing.” Weidenfeld & Nicolson, London. Nilsson, J.-A. (1994). Energetic stress and the degree of fluctuating asymmetry: Implications for a long-lasting, honest signal. Evol. Ecol. 8, 248-255. Ozernyuk. N. D., Dyomin, V. I., Prokofyev, E. A,, and Androsova, I. M. (1992). Energy homeostasis and developmental stability. A c f a Zoof. Fenn. 191, 167-175. Palmer, A. R. (1994). Fluctuating asymmetry analyses: A primer. In “Developmental Instability: Its Origins and Evolutionary Implications” (T. A. Markow, ed.), pp. 335-364. Kluwer, Dordrecht. The Netherlands. Palmer, A. R., and Strobeck, C. (1986). Fluctuating asymmetry: Measurement, analysis, patterns. Annri. Rev. Ecol. Sysr. 17, 391-421. Palmer, A. R., and Strobeck, C. (1997). Fluctuating asymmetry and developmental stability: Heritability of observable variation vs. heritability of inferred cause. J. Evol. Biol. 10, 39-49. Parker, H. R., Philipp. D. P., and Whitt, G. S. (1985). Gene regulatory divergence among species estimated by altered developmental patterns in interspecific hybrids. Mol. Biol. Evol. 2, 217-250.

212

ANDERS PAPE M0LLER

Parsons, P. A. (1990). Fluctuating asymmetry: An epigenetic measure of stress. Biol. Rev. Cambridge Philos. Soc. 65, 131-145. Parsons, P. A. (1991). Can atmospheric pollution be monitored from the longevity of stress sensitive behavioural mutants in Drosophila? Funct. Ecol. 5, 713-715. Parsons, P. A. (1994). The energetic cost of stress: Can biodiversity be preserved? Biodiv. Lett. 2, 11-15. Polak, M. (1993). Parasites increase fluctuating asymmetry of male Drosophila nigrospiracu/at Implications for sexual selection. Genetica 89, 255-265. Polak, M. (1997). Ectoparasitism in mothers causes higher fluctuating asymmetry in their sons: Implications for sexual selection. Am. Naf. 149, 955-974. Pomiankowski, A. (1997). Genetic variation in fluctuating asymmetry. J. Evol. Biol. 10,51-55. Poulsen, M. G. (1997). Fluctuating asymmetry and conservation status of butterflies. MSc thesis, Univ. of Copenhagen, Denmark. Rae, C . , Scott, R. B., Thompson, C. H., Kemp, G. K., Dumughin, I., Styles. P., Tracey. I., and Radda. G. K. (1996). Is pH a biochemical marker of IQ? Proc. R. Soc. London, Ser. B 263, 1061-1064. Rasmuson, M. (1960). Frequency of morphological deviations as a criterion of developmental stability. Heredifas 46, 51 1-536. Rettig, J. E., Fuller, R. C., Corbett, A. L., and Getty, T. (1997). Indicates levels of competition in an even-aged poplar clone. Oikos 80, 123-127. Sarre, S., Dearn, J. M., and Georges, A. (1994). The application of fluctuating asymmetry in the monitoring of animal populations. Pac. Conserv. B i d . 1, 118-122. Scutt, D.. Manning, J. T.. Whitehouse, G. H., Leinster, S. J., and Massey, C. P. (1997). The relationship between breast asymmetry, breast size and the occurrence of breast cancer. Br. J. Radio/. (in press). Shackelford, T. K., and Larsen, R. J. (1997). Facial asymmetry indicates psychological, emotional, and physiological illness. J. Personal. Sue. Psycho/. (in press). Shapiro, B. L. (1992). Development of human autosomal aneuploid phenotypes (with an emphasis on Down syndrome). Acta Zool. Fenn. 191, 97-105. Siegel, M. I., and Smookler, H. H. (1973). Fluctuating dental asymmetry and audiogenic stress. Growth 37, 35-39. Simmons, L. W., and Ritchie, M. G. (1996). Symmetry in the songs of crickets. Proc. R. Soc. London, Ser. B 263,305-3 11. Sumarsono, S. H., Wilson, T. J.. Tymms, M. J.. Venter, D. J.. Corrick. C. M., Kola, R., Lahoud, M. H., Papas, T. S., Seth, A,, and Kola, I . (1996). Down’s syndrome-like skeletal abnormalities in ETS2 transgenic mice. Nature (London) 379, 534-537. Suthers, R. A. (1994). Variable asymmetry and resonance in the avian vocal tract: A structural basis for individually distinct vocalizations. J. Comp. Physiol. A 175, 457-466. Swaddle, J. P. (1997). On the heritability of developmental stability. J. Evol. B i d . 10,57-61. Swaddle, J. P., Witter, M. S., Cuthill. I. C., Budden. A,, and McCowen, P. (1996). Plumage condition affects flight performance in starlings: Implications for developmental homeostasis, abrasion and moult. J. Avian Biol. 27, 103-111. Thoma, R. (1996). Developmental instability. handedness, and brain lateralization: MES and MRI correlates. MSc Thesis, University of New Mexico, Department of Psychology, Albuquerque. Thornhill. R.. and Mgiller, A. P. (1997). Disease and developmental stability. B i d . Rev. Cambridge Philos. Soc. (in press). Tinbergen, N. (1951). “The Study of Instinct.” Oxford University Press, Oxford. Toates, F. (1 995). “Stress: Conceptual and Biological Aspects.” Wiley, Chichester.

DEVELOPMENTAL INSTABILITY AND STRESS

213

Tomkins. J. L., and Simmons, L. W. (199.5). Patterns of fluctuating asymmetry in earwig forceps: No evidence for reliable signalling. Proc. R. Soc. London, Ser. B 259, 89-96. Van Valen, L. (1962). A study of fluctuating asymmetry. Evolufion (Lawrence, Kans.) 16, 125-142. V~llestad,L. A., Hindar, K., and M ~ l l e r .A. P. (1997). Patterns of correlation between heterozygosity and fluctuating asymmetry. J. Evol. Biol. (submitted for publication). Whitlock, M. (1996). The heritability of fluctuating asymmetry and the genetic control of developmental stability. Proc. R. Soc. London, Ser. B 263, 849-854. Whitlock, M. C., and Fowler, K. (1997). The instability of studies of instability. J. Evol. Biol. 10,6347. Witter. M. S., and Lee, S. J. (199.5).Habitat structure, stress and plumage development. Proc. R. Soc. London, Ser. B 261,303-308. Zakharov, V. M.. and Yablokov, A. V. (1990). Skull asymmetry in the Baltic grey seal: Effects of environmental pollution. Amhio 19, 266-269. Zakharov, V. M., Olsson. M., Yablokov, A. V.. and Esipenko. A. G . (1989). Does environmental pollution affect the developmental stability of the Baltic grey seal (Halichoerus grypus)? In “Influence of Human Activities on the Baltic. Proceedings of a Soviet-Swedish Symposium, Moscow, 14-18 April 1986” (A. V. Yablokov and M. Olsson. eds.). pp. 96-108. Gidrometeorizdat, Leningrad. Zakharov, V. M.. Pankakoski, E., Sheftel, B. I., Peltonen, A., and Hanski, I. (1991). Developmental stability and population dynamics in the common shrew, Sorex araneus. Am. Nut. 138, 797-810.

This Page Intentionally Left Blank

ADVANCES IN THE STUDY OF BEHAVIOR. VOL. 27

Stress and Decision Making under the Risk of Predation: Recent Developments from Behavioral, Reproductive, and Ecological Perspectives STEVEN L. LIMA DEPARTMENT OF LIFE SCIENCES INDIANA STATE UNIVERSITY TERRE HAUTE. INDIANA 47809

I.

INTRODUCTION

My objective here is to provide a comprehensive review of recent empirical and theoretical work on antipredator decision making. The ways in which predators influence the behavioral decisions made by their prey is now the subject of a large and growing literature. This sustained interest in the behavioral aspects of predator-prey interactions is readily traced to the fact that virtually all animals are subjected to some form of predation, and many biological and ecological insights can be gained from an understanding of the ways in which predators influence their prey’s behavior. Prey decision making under the risk of predation essentially allows an animal to manage predator-induced stress. Stress is not a term commonly associated with the study of antipredator &cision making, but this is largely a matter of semantics, and one can relate stress to such decision making in several contexts. If one defines stress as an environmental condition that diminishes Darwinian fitness through either reproduction or survival (e.g., Sibly and Calow, 1989), then few aspects of an environment would lead to more stress than predators. Note that death due to predation is not the sort of stress that I consider here: observable predator-induced stress in animals is (in part) a result of prey decision making itself, such as the energetic stress caused by choosing to feed less in the presence of predators. One might thus consider the adaptive management of this sort of predatorinduced stress as a main function of antipredator decision making. Forms of non-predator-induced stress, such as energetic stress caused by food shortages, will also influence such decision making. This view of stress is 215

Copyright C IYYK by Academic Press All rights of reproduction in any form reserved 0065-3454198 $25 00

216

STEVEN L. LIMA

ecological or evolutionary in perspective, and most of the existing literature deals with stress in this context. A more classical definition of stress concerns the rapid increase in certain hormones (e.g., glucocorticoids) in response to some threatening situation (Weiner, 1992). This hormonal response is considered to be a biological marker of fear (Boissy, 1995), and there exists a substantial literature on fear and the physiological (neuroendocrine) stress response (for a review, see Boissy, 1995). However, relatively few studies on physiological stress have worked with predators, and research relating such stress to antipredator behavior is still in its infancy (Bercovitch ef al., 1995; Boissy, 1995). Further research into these physiological aspects of stress may ultimately have several important implications for how we view antipredator decision making, several of which I summarize in a closing section. Regardless of the topic being addressed, all of the published works included in this review share certain characteristics. First, the behaviors/ decisions in question respond in ecological time to changes in some component of the risk of predation (sensu Lima and Dill, 1990). That is, this review concerns plastic behavioral traits that respond to short-term perceived changes in the risk of predation. Thus, I do not consider in detail those aspects of behavior that respond to predation over evolutionary time (see Edmunds, 1974; Endler, 1991). Second, figuring prominently in most studies included herein is the inevitable trade-off between the benefits of avoiding possible predation and the costs of doing so, in terms of feeding, survival, or reproduction (i.e., stresses as defined earlier). Third, for reasons of manageability, I include work published primarily during the last 7-8 years. This covers roughly the time period since the publication of several relevant reviews that were written in the late 1980s (Dill, 1987; Sih, 1987; Lima and Dill, 1990). The present review nevertheless encompasses about twice the number of papers covered by Lima and Dill (1990), whose comprehensive coverage extended over almost a 15-year period! I have strived to provide perspectives on antipredator decision making that encompass several levels of biological organization. I thus cover the spectrum from short-term decision making by individuals to the consequences of such decision making for long-term fitness, population dynamics, and species interactions. Work on short-term decision making under the risk of predation has a relatively long history of study (Milinski, 1986; Sih, 1987; Lima and Dill, 1990), whereas most of the work on its consequences has appeared in recent years. My choices for the topics organizing this review reflect an attempt to provide a representative perspective on the current state of the field-I hope they succeed. In the hope of synthesizing available studies as much as possible, I have also classified studies across several topics to the extent warranted.

PREDATOR-INDUCED STRESS AND BEHAVIOR

217

11. BEHAVIOR OF FEEDING ANIMALS: CLASSICAL MOTIVATIONS

It is appropriate to begin with an examination of recent empirical and theoretical work on foraging behavior. By “classical,” I refer to studies motivated directly or historically by optimal foraging theory (Stephens and Krebs, 1986). Work in this area still forms the main empirical, theoretical, and philosophical basis for the study of decision making under the risk of predation. Note that while this section focuses mainly on classical issues, subsequent sections often deal with the behavior of feeding animals to one degree or another. STRESS AND STATE-DEPENDENT RISKTAKING A. ENERGETIC 1. Empirical Studies

One of the best ways to demonstrate that animals trade off safety against feeding is to manipulate their internal (energetic) state (Milinski, 1993). Such a manipulation is usually accomplished via a period of food deprivation. Provided that riskier behavioral options are also those that allow for a higher rate of energy intake, then an energetically stressed animal should accept a relatively high risk of predation while feeding. This idea goes back to the very earliest of studies on anti predator trade-offs (Milinski and Heller, 1978; Dill and Fraser, 1984). Work on state-dependent risk taking began in earnest during the late 1980s, and the pace of research has accelerated in recent years (Table I). Recent demonstrations of state-dependent risk taking make clear that such behavior is widespread. Almost without exception, over a wide range of decision making and taxa (Table I), energetically stressed animals will accept relatively great risk to obtain food. Most studies manipulated an animal’s energetic state (hunger), with a few exceptions addressing issues such as the effects of reproductive or migratory state. Moore’s (1994) study is particularly interesting in this regard; warblers in a migratory state (and thus in need of large energetic reserves for long-distance flight) took greater risks than control birds even though the former had higher energetic reserves than the latter. Another unusual result concerns the demonstration that bumblebee workers accept greater risks for increased food intake when their colony is experiencing energetic stress (Cartar, 1991). In related work, Weary et al. (1996) found that slower growing piglets accept a higher risk of maternal crushing to secure increased milk intake.

2.

Theory: The Rise of Stochastic Dynamic Programming

State-dependent decision making under the risk of predation is at the heart of stochastic dynamic programming (SDP). The introduction of SDP

TABLE I RECENT STUDIES EXAMINING STATE-DEPENDENT RISK-TAKING IN ANIMALS Animal

t ! DO

Statelstress

Invertebrates Bumble bee (Bombus occidentalis)

Energetic

Barnacle (Balanus glandula)

Energetic

Stonefly larvae (Paragnerina media) Mayfly larvae (Baetis tricaudatus)

Energetic Energetic

Stonefly larvae (Acroneuria and Paragnetina, 2 spp.) Whirligig beetle (Dineutes assimilis)

Energetic

Whirligig beetle (D. assindis)

Energetic

Backswimmer (Notonecta hoffmnni)

Energetic

Dogwhelk (Nucella lapillus)

Energetic

Vertebrates Ground squirrel (Spermophihrs beldingi)

Body mass

Energetic

Context and result Foraging workers are reluctant to flee from predator when colony’s reserves are low Poorly fed barnacles resume feeding faster following encounter with predator No apparent effect of hunger on use of space Hungry larvae increase feeding by spending less time in refuges No effect of hunger on tendency to enter drift

Source Cartar (1991) Dill and Gillett (1991) Feltmate and Williams (1989a) Kohler and McPeek (1989) Rader and McArthur (1995)

Hungry beetles occupy profitable but risky outer portion of group Hungry beetles adopt solitary foraging to increase energetic gain Hungry individuals resume feeding faster following encounter with predator Hungry individuals move more and spend less time in aerial refuges

Romey (1995)

Individuals with low body mass show reduced vigilance following alarm calls

Bachman (1993)

Romey and Rossman (1995) Sih (1992a) Vadas et al. (1994)

t!

Stickleback (Spinachia spinachia) Stickleback (Gasterosteus aculeatus)

Energetic Energetic

Stickleback (C. aculeatus)

Energetic

Atlantic salmon (Salmo salar)

Energetic

Pika (Ochotona collaris)

Reproductive

Frog larvae (2 Rana spp.) Stickleback (C. aculeatus)

Energetic Energetic

Willow tit (Parus montanus)

Energetic

Roach (Rutihts rutilus)

Energetic

Dark-eyed junco (Junco h y e m a h )

Energetic

Coho salmon (Oncorhynchus kisurch)

Energetic

Yellow-rumped warbler (Dendroica coronata) Crucian carp (Carassius carassius) Porcupine (Erethizon dorsatum)

Migratory Energetic Body mass

Hungry fish choose riskier but more profitable patches Croy and Hughes (1991) Hungry fish choose increasingly safer but less profitable prey Godin (1990) as they satiate Hungry fish increase predator inspection. reflecting a greater Godin and Crossman (1994) need for information (?) Hungry fish resume feeding faster following encounter Gotceitas and Godin (1991) with predator Lactating females feed in riskier but more profitable Holmes (1991) microhabitats Hungry tadpoles increase activity under all levels of risk Horat and Semlitsch (1994) Hungry fish feed on dense but dangerous portions of prey Jakobsen era/. (1994) swarms Hungry birds resume feeding faster following encounter Koivula ef 01. (1995) with predator Hungry fish occupy profitable but risky periphery of the Krause et al. (1992); Krause (1993a) group Hungry birds increase rate of energy intake by reducing Lima (1995) vigilance Hungry fish are more willing to attack distant prey following Martel and Dill (1993) recent exposure to predator Birds in migratory state resume feeding faster following Moore (1994) encounter with predator Hungry fish feed in riskier but more profitable microhabitats Petterson and Bronmark (1993) Individuals with low body mass feed in risky but Sweitzer and Berger (1992) profitable microhabitats

220

STEVEN L. LIMA

to behavioral ecology was spurred by the need to combine disparate quantities like predator avoidance and food intake into a common framework for making predictions about state-dependent behavior (McNamara and Houston, 1986; Mangel and Clark, 1988). SDP models use the numerical technique of backward induction to develop an optimal behavioral “program” in which optimal behavior is specified for all possible internal states and environmental contingencies. The intuitive and conceptual appeal of such behavioral programs, and the relative accessibility of SDP modeling to biologists (via Mangel and Clark, 1988), have led to much recent interest in SDP. Published SDP models cover a wide range of behavioral issues in decision making under the risk of predation. The more “classically” oriented models examine issues of state dependence (typically energetic stress) in diet selection (Godin, 1990; Burrows and Hughes, 1991) or patch use (Newman, 1991; see also Houston et al., 1993; McNamara and Houston, 1994), while others have explored state-dependent foraging activity/effort (Werner and Anholt, 1993; Crowley and Hopper, 1994). Rosland and Giske (1994) and Fiksen and Giske (1995) have developed models of optimal die1 vertical migration in aquatic animals. Houston and McNamara (1989) have used SDP to explore the issue of foraging effort in closed versus open experimental systems. The use of rules of thumb regarding uncertainty about predation risk has also been addressed with SDP (Bouskila and Blumstein, 1992). In addition, a series of SDP models incorporating body-mass-dependent predation in birds addresses issues of optimal body mass (McNamara and Houston, 1990; Houston and McNamara, 1993; Bednekoff and Houston, 1994; see also Bull et al., 1996, for related work with fish), the decision to hoard food (Lucas and Walter, 1991), and the temporal patterning of daily foraging behavior (Bednekoff and Houston, 1994; McNamara et al., 1994). McNamara and Houston (1992), Houston et al. (1993), and Bednekoff (1997) have used SDP to explore several issues surrounding the tradeoff between feeding and antipredatory vigilance. SDP models have also addressed the influence of predation risk on optimal sociality (Szekely et al., 1991; Paveri-Fontana and Focardi, 1994), parental behavior (Clark and Ydenberg, 1990a,b), and various aspects of mating behavior (Sargent, 1990; Crowley et al., 1991; KAlAs et al., 1995; Lucas and Howard, 1995; Lucas et al., 1996). This interest in SDP modeling has not yet produced a corresponding increase in empirical tests of such models. Few SDP models are even accompanied by much empirical information (but see Godin, 1990; Burrows and Hughes, 1991; Lucas and Walter, 1991; Rosland and Giske, 1994; Bull et al., 1996). Furthermore, most studies demonstrating state-dependent antipredatory decision making (Table I) do not directly address SDP theory.

PREDATOR-INDUCED STRESS AND BEHAVIOR

221

The problem here may lie in (1) the sometimes extremely complex nature of SDP models (e.g.. Burrows and Hughes, 1991; Crowley and Hopper, 1994; Rosland and Giske, 1994; Fiksen and Giske, 199.5; Lucas and Howard, 199.5), which may have outstripped the empiricist’s ability to provide even qualitative tests of theory, and (2) the fact that qualitative predictions regarding state-dependent behavior often do not require SDP modeling. The value of SDP models is nonetheless clear and important, especially with regard to the link between short-term decision making and life-history phenomena (Clark, 1994; McNamara et al., 1995). I return to the issue of testability and the importance of models later in this section. BEHAVIOR B. THEp/g RULEFOR OPTIMAL The p / g rule specifies that an animal can maximize its fitness, or optimally manage its predator-induced stress, by choosing the behavioral option that minimizes the rate of mortality ( p ) per unit increase in growth rate (8). Gilliam (1982; see also Werner and Gilliam, 1984) derived this rule for animals that experience continuous growth up to some reproductive size, but it has since been broadened to other animals in the form the p/f rule, where f represents feeding rate (Gilliam and Fraser, 1988; Gilliam, 1990). In all of its guises, the p/g rule has undeniable appeal. It has been applied to the question of patch choice (Gilliam and Fraser, 1987, 1988; Moody et al., 1996; Sih, 1998), diet choice (Gilliam, 1990), foraging effort (Werner and Anholt, 1993), avian migration (Lindstrom, 1990), and life-history evolution (Werner, 1986; Aksnes and Giske, 1990). Furthermore, the p/g rule has been derived in several contexts (Clark and Levy, 1988; McNamara and Houston, 1992,1994; Houston ef a/., 1993) and without the dynamical theory used in its original formulation (Aksnes and Giske, 1990; Leonardsson, 1991; Brown, 1992; Clark, 1994; Clark and Dukas, 1994; Hugie and Dill, 1994; Dukas and Clark, 199.5). Several recent papers caution that the p/g rule has its limitations (many of which were noted in Gilliam, 1982). Ludwig and Rowe (1990) and Rowe and Ludwig (1991) show that time constraints in reaching reproductive size can negate the simple p/g rule. Clark (1994) adds that the p/g rule implies an unlikely scenario in which reproductive value does not change over time. McNamara and Houston (1994; see also Houston et al., 1993) show further that the p/g rule requires no stochasticity or state dependency in p or g (or f ) . Most importantly, McNamara and Houston show that the p/ g rule applies only when long-term foraging options are not subject to change. Such an environment is unlike that in which the p/g rule might be tested experimentally.

222

STEVEN L. LIMA

Despite these apparent limitations, the p/g rule can perform well even when some of the above conditions are clearly violated (Werner and Anholt, 1993; Crowley and Hopper, 1994). This suggests that animals might actually use some p/g-like rule in their decision making. However, only Gilliam and Fraser (1987) provide quantitative empirical support for such a rule. Gotceitas (1990) claims empirical support for the p/g rule, but his results suggest a more simple alternative explanation (see McNamara and Houston, 1994) in which the fish studied simply acted t o minimize p . Other tests (Bowers, 1990; Turner and Mittelbach, 1990) provide only qualitative support that appears consistent with the general expectations of several different models. In any case, the p/g rule remains a powerful heuristic tool in the study of decision making under the risk of predation.

C. FORAGING I N A PATCHY ENVIRONMENT Here, I address primarily the relatively abstract theoretical and empirical aspects of foraging in a patchy environment, typically a laboratory or “mathematical” environment. I address the more ecologically motivated studies of habitat use in a later section. This distinction is not always easily made, but it is a useful one.

1. Patch Choice Recent theoretical developments in this area concern the Ideal Free Distribution (IFD) model of patch choice. This model posits that animals with perfect (ideal) information are free to choose patches such that they maximize their fitness, subject to the choices made by other animals. A common prediction is that the distribution of animals among patches will eventually stabilize (at the IFD) and match the distribution of food resources among patches (Milinski and Parker, 1991). Predators can certainly disrupt the IFD, and Moody et al. (1996) provide a much-needed theoretical perspective on this phenomenon. They show that (1) undermatching of food resources is a universal expectation when resource-rich patches are also the riskier patches (e.g., Abrahams and Dill, 1989), and that (2) multiple stable distributions are possible under some circumstances. Other recent IFD-based models examine situations in which predators respond to the distribution of prey, and prey, in turn, respond to the distribution of both their food resources and predators (Schwinning and Rosenzweig, 1990; Hugie and Dill, 1994; Sih, 1998; see also van Baalen and Sabelis, 1992). The overall results indicate that stable distributions across patches of both predator and prey are possible outcomes in many situations (but see Schwinning and Rosenzweig, 1990). These multi-trophic-level models also make the counterintuitive predictions that (1) the distribution of predators should

PREDATOR-INDUCED STRESS AND BEHAVIOR

223

tend to match their prey’s resource distribution (Hugie and Dill, 1994; Sih, 1998), and (2) the prey distribution may not closely match the distribution of prey resources (Hugie and Dill, 1994; but see Sih, 1998). Empirical tests of these predictions ought to be feasible, but none has been reported (but for related empirical studies, see Sih, 1984; Formanowicz and Bobka, 1989). Empirically, predator-induced deviations from the Ideal Free Distribution have been used to assess the “energetic equivalence” of predator avoidance in predation-risk-dependent patch choice (Abrahams and Dill, 1989; Todd and Cowie, 1990; Utne et af., 1993); the ultimate goal here is to express food intake and predator avoidance in the common currency of energy (see also Kotler and Blaustein, 1995, for a different perspective on this matter). Kennedy et al. (1994) criticized such IFD-based studies for assuming an IFD rather than assessing the possibility of systematic deviations from the IFD. Kennedy et al. also present a non-IFD-based alternative to assessing the energetic equivalence of predator avoidance, but Moody et al. (1996) warn that this alternative has no functional basis. Moody et al. caution further that the entire enterprise of determining such energetic equivalencies may rest on shaky conceptual ground. There have been relatively few non-IFD-related developments regarding patch choice under the risk of predation. Theoretically, Gilliam and Fraser (1988) extend the pulg rule to patch choice with depleting resources. Houston et al. (1993) provide a cogent discussion and review of the relationships among models of optimal patch choice under the risk of predation. Empirically, there have been several recent demonstrations that patch choice represents an energy-predation trade-off when dangerous patches are also energetically profitable (e.g., Gotceitas, 1990; Gotceitas and Colgan, 1990a,b; Brown et al., 1992a,b; Pettersson and Bronmark, 1993; Scrimgeour and Culp, 1994a; Scrimgeour et al., 1994). These studies complement many similar studies reviewed in Lima and Dill (1990). Nonacs and Dill (1990) provide the unique result that a worker ant’s decision to feed in a risky patch reflects the contribution that its efforts make to colony growth.

2. Time in Patches Recent theoretical treatments of patch use differ considerably in their predictions. Newman (1991) indicates that optimal patch residence time may be influenced little by the risk of predation. In contrast, Brown (1992) develops several models in which optimal patch residence times are highly predation-risk dependent. This discrepancy may reflect disparate assumptions about whether patches vary in predation risk or energetic quality. Empirically, there is much evidence that the degree to which small mammals exploit patches is predation-risk dependent (see Section VII1,A).

224

STEVEN L. LIMA

3. Choice of Foraging Location Several recent (and somewhat difficult to categorize) papers come under this general heading, which addresses within-patch decisions about where to feed. For instance, Jakobsen et al. (1994) found that sticklebacks forage on denser portions of zooplanktonic swarms only when energetically stressed or safe from attack; this reflects a trade-off between predator detection and feeding rate (see also Milinski and Heller, 1978; Godin and Smith, 1988). Vasquez (1994) found that a small cricetid rodent becomes a refuge-seeking, central-place consumer of food when feeding under a threat of predation. Peterson and Skilleter (1994) found that clams shift feeding location from substrate (risky but profitable option) to water column (safe but less profitable option) after suffering partial siphon loss to foraging fish. This shift is consistent with an energy-predation trade-off, but it is not clear whether clams can assess the risk of (partial) predation independent of the act itself, or whether they could effectively employ both foraging options after partial siphon loss (see also Lindsay and Woodin, 1995). D. DIETSELECTION Recent work provides much-needed theoretical perspectives on diet selection under the risk of predation. Gilliam (1990) describes a particularly insightful extension of the p / g rule to the question of diet selection. This model exhibits quasi-classical behavior (see Stephens and Krebs, 1986) in which prey-specific predation risks are a determinant of prey ranking. Godin (1990) developed an SDP model that predicted that profitable but risky prey (large prey whose consumption interferes with predator detection) should be consumed preferentially only by energetically stressed animals. Burrows and Hughes (1991) presented an ambitious SDP model in which mortality and digestive constraints combine to cause a general contraction of the diet with increasing risk of predation. Empirical work on diet selection has been limited. Godin (1990) provided empirical evidence that diet selection in guppies is predationrisk and state dependent as predicted (qualitatively) by his SDP model; further support for this model lies in the observation that fish may prefer large, profitable items only under low predation risk (Ibrahim and Huntingford, 1989). Phelan and Baker (1992) suggested that predationrisk-related travel costs influence diet selection in mice, but their test suffered from a lack of any manipulation of risk. Brown and Morgan (1995) showed that a squirrel’s apparent preference for certain food types can be predation-risk dependent, even though one food type may be inherently preferred over others.

PREDATOR-INDUCED STRESS AND BEHAVIOR

225

E. TESTA~ILITY AND THE ROLEOF THEORETICAL MODELS It is appropriate at this point to address some important issues regarding the role of modeling in the study of predator-induced stress and antipredator decision making, as “classically motivated” work is the most theory-rich area that I consider in this review. The following discussion, however, applies generally to subsequent sections. These issues regarding the role of modeling concern the virtual absence of quantitative tests of theory. Besides the efforts of Gilliam and Fraser (1987) and Gotceitas (1990), few attempts at quantitative tests have been reported. There are probably several reasons for this phenomenon. First, many simple models are obviously caricatures of reality that demand no quantitative test. At the opposite extreme, some ambitious SDP models may outstrip the ability of empiricists to provide even qualitative tests of predictions. A more fundamental problem concerns our inability to measure the risk of predation itself (or its various components). Only a few field studies have much quantitative information on the risk of predation (e.g., Watts, 1990;Harfenist and Ydenberg, 1995), and none provides information that relates an animal’s conceivable behavioral options to particular risks of predation. This sort of information is critical to making quantitative behavioral predictions about the adaptive management of predatorinduced stress. To what extent are we limited by our inability to provide quantitative tests of theory? Two lines of argument suggest that this limitation is not too severe. First, qualitative tests of carefully reasoned predictions should prove enlightening in most situations. Second, even without quantitative tests, there has been an invaluable interplay between theory and empiricism in the study of decision making under the risk of predation, and I see no reason why this will not continue. On the other hand, Brown (1992) argues that models with rather disparate fitness formulations can yield similar qualitative predictions. Quantitative tests may ultimately be needed to determine which fitness formulation is superior. Given our ongoing inability to provide quantitative tests of theory, modelers have little choice but to strive for qualitative predictions that distinguish among various hypotheses. I personally prefer relatively simple models with broad heuristic value, but Abrams (1993a) argues that simple models can also be misleading. In any case, a pluralism of modeling approaches should continue to provide a strong conceptual basis for further empirical and theoretical progress. 111. PAITERNS OF ACTIVITY

“Activity” studies examine the influence of predators on both the level and the temporal patterning of prey activity. I consider each of these areas

226

STEVEN L. LIMA

in turn. These studies on prey activity provide some of the best documented behavioral responses by prey to the presence of predators, and form the foundation for much behaviorally explicit ecological research on predatorprey interactions (e.g., Werner, 1992; Wooster and Sih, 1995). OF ACTIVITY A. LEVELS

I distinguish between two types of activity, movement and refuging. An animal can in principle vary these two types of activity independently in response to the risk of predation (Sih and Kats, 1991; Werner and Anholt, 1993). By movement, I refer to things like speed of movement, length of moves, frequency of movement, and so on. Refuging refers to a situation in which an animal retreats to a refuge and emerges infrequently and for only brief periods; a refuge is, for example, a burrow or rock crevice (as opposed to a safe habitat) in which an animal cannot readily feed, locate mates, and so on (e.g., Sih et al., 1988). Categorizing a given activity as either movement or refuging is usually straightforward, although many studies do not define in detail the behaviors under examination. A decrease in prey activity following a heightened threat of predation has been a reasonably well established result for some time (Sih, 1987; Lima and Dill, 1990; Kolar and Rahel, 1993; Wooster and Sih, 1995), one that also figures prominently in studies related to physiological (neuroendocrine) stress (Boissy, 1995). Work in recent years indicates that such a response is indeed ubiquitous across diverse taxa (Table 11). Almost all species studied exhibit decreased movement, increased refuging, or both (if both types of behavior were examined) in response to an increase in the risk of predation. Several studies indicate that many aquatic (and even some terrestrial) animals respond to the chemical evidence of predators as well as the actual presence of predators (for an extensive review, see Kats and Dill, 1998). There were exceptions to the general result of decreased activity with increasing risk (Table 11). Some cases with no response to predator manipulation may have involved prey large enough to be invulnerable to predators (e.g., Willman et al., 1994), while in others a nonsignificant effect was in the typical direction (e.g., Walls, 1995). Houtman and Dill (1994) found a decrease in movement by marine sculpins only if the background provided some degree of crypticity. Larval Ambystoma salamanders decreased movement only in the absence of a refuge; otherwise, movement increased in an effort to reach a refuge (Sih and Kats, 1991). The case of increased movement in toad larvae in response to an alarm substance (Hews, 1988) may also represent refuge-seeking behavior. Sih and Krupa (1992) argued that female water striders take advantage of a predator-induced decrease

TABLE I1 RECENT STUDIES EXAMINING CHANCES I N PREY ACTIVITY I N RESPONSE TO PREDATOR PRESENCE

OR

PERCEPTION THEREOF

Change in activity' with Species Invertebrates Aquatic snail (Physella and Planorbella, 2 spp.) Crayfish (Pacifastacus leniusciilus)

Predatof

Activityh

Risk

t

Food

t

Hunger

Source

Alexander and Covich (1991a.b) Blake and Hart (1993)

Alarm substance

R

Incd

NP

Fish C

M R M

Dee Inc Dec (1 sp.). NR (1 sp.) Dee Inc Inc

-

-

Blois-Heulin el al. (1990)

NP NP NP

Crow1 and Covich (1994) Culp et al. (1991)

-

Damselfly larvae (2 Enallagrna SPP.) Shrimp (Atya lanipes)

Fish P. larval odonate P Large shrimp P

Larval mayfly (Paraleptophlebia heteronea) Grass shrimp (Palaernonetes pugio) Stonefly larvae (Paragnetina media) Crayfish (Orconecfesvirilis) Crayfish (3 Orconectes spp.) Isopod (Lirceus fontinah)

Fish P

M R R

Fish P Fish P

R M

Inc Dee

NP NP

Everett and Ruiz (1993) Feltmate er al. (1992)

Alarm substance Fish P Fish P

M

NP

M

Dec Inc Dec

Amphipod ( Garnrnarus minus)

Fish C

M

Dec

NP

lsopod ( L . fontinalis) Damselfly larvae (2 Enallagnia SPP.1 Damselfly larvae (Coenagrion hastidaturn) Mayfly larvae (Baetis fricaudatus)

Fish P Larval dragonfly P

M M

Dec Dee

-

Inc

Hazlett (1994) Hill and Lodge (1994) Holomuzki and Short (1990) Holomuzki and Hoyle (1990) Huang and Sih (1990, 1991) Jeffries (1990)

Larval dragonfly P

M

Dec

NR

Johansson (1993)

Fish P

M R

Dec Inc

NR Dec

Kohler and McPeek (1989)

R

-

-

(continued)

TABLE I1 (Continued) Change in activity’ with Species

Predatof

Risk t

Activity6

Food

t

Hunger f

Source

Chironomid larvae (Chironomus tentans) Chironomid larvae (C. tentans) Damselfly larvae (4 Enallagma SPP.1 Worker ants (Lasius pallitarsis) Caddis larvae (Rhyacophila nubila) Water flea (2 Daphnia spp.)

Fish P

R

Incd

NP

Macchiusi and Baker (1991)

Fish P Fish P, larval dragonfly P Large ant P None Copepod P

R M

Inc Dec

Dec -

Macchiusi and Baker (1992) McPeek (1990)

M R M

-

Dec

Marine snail (Stramonita haemastoma) Ostracod (Cypridopsis vidua) Mayily larvae (Baetis, Ephemerella, Claassenia spp.) Isopod (L.fontinalis)

Crab P

R

Dee (1 sp.), inc (1 SP.1 Inc

Inc NP -

Fish C Fish P, stonefly P

R R

Inc Inc

Fish C

M

Dec

Water strider (Aquarius remigis)

Fish P

M

Isopod (Saduria entomon)

Larger isopod P

R

Dec (male), inc (female) Inc

Mayfly larvae (2 Baetis spp.)

Fish P

Aquatic snail (Physella gyrina) Dogwhelk (Nucella lapillus)

Alarm substance Crab C, alarm substance Fish P Fish P

M R R M R R R

Lobster (Homarus americanus) Crayfish (3 Orconectes spp.)

NR Incd Inc Dec Inc Inc Dec, 1sp., NR, 2 spp!

-

Dee

-

-

NP

Inc NP

Nonacs (1990) Otto (1993) Ramcharan and Sprules (1991) Richardson and Brown (1992) Roca et al. (1993) Scrimgeour et al. (1994) Short and Holomuzki (1992) Sih and Krupa (1992,1995) Sparrevik and Leonardsson (1995) Tikkanen et al. (1994)

Inc Dec

Turner (1996) Vadas et al. (1994) Wahle (1992) Willman et al. (1994)

Fish Fathead minnow (Pimephales promelas) Fathead minnow (P.promelas)

R M R

Inc Dec Inc

M

Dec

Chivers and Smith (1994. 1995) Gelowitz et al. (1993)

M

Dec'

Houtman and Dill (1994)

Duck C

M

Dec

Martel and Dill (1993)

Duck P Alarm substance, fish P Alarm substance Fish P Fish P

M M R M M M

Dec Dec Inc Dec Dec' Dee

Martel and Dill (1995) Mathis and Smith (1993a): Mathis er al. (1993) Mathis and Smith (1993b) Radabaugh (1989) Williams and Brown (1991)

Larval dragonfly P Adult newt C Fish P

M R R

Inc Inc

R

Inc

Toad (Bufo americanus) Toad (B. americanus) Toad larvae (B. americanus) Frog larvae ( 2 Rana spp.)

Fish and salamander C Snake P Snake P Alarm substance Fish C

Salamander larvae (A. texanum) Frog, toad larvae (Hyla and Bufo, 4 SPP.)

Fish P Fish, newt, dragonfly P

M M M M R R M

Dec Dec Inc Dec NR Inc Dec

Brook stickleback (Culaea inconstans) Marine sculpin (Oligocottus maculosus) Coho salmon (Oncorhynchus kisutch) Coho salmon (0.kisutch) Fathead minnow (P.promelas)

h) W

h)

Brook stickleback (C. inconstans) Darter (3 Etheostoma spp.) Lumpfish larvae (Cyclopterus lumpus) Amphibians Frog larvae (Rana catesbeiana) Newt larvae (Taricha torosa) Salamander larvae (Ambystoma maculatum) Frog larvae (Ascaphus truei)

Alarm substance Alarm substance, fish P Alarm substance, fish C Alarm substance

Brown et al. (1995)

Dee NP

NP NP Dec or NR NR -

-

-

Inc Inc NR -

Anholt and Werner (1995) Elliott et al. (1993) Figiel and Semlitsch (1990) Feminella and Hawkins (1 994) Hayes (1989) Heinen (1994a,b) Hews (1988) Horat and Semlitsch (1994) Huang and Sih (1990,1991) Lawler (1989) -

(continued)

TABLE I1 (Continued) Change in activity‘with Species

3

Activity’

Frog larvae (R. temporarin) Salamander larvae ( A . babouri) Salamander larvae ( A . babotcri) Salamander larvae ( A . baboirri) Toad larvae (B. americanus) Frog larvae ( H y l a versicolor) Frog larvae (2 Pseudncris spp.) Frog larvae (2 R a m spp.)

Fish, crayfish C Fish P Fish C Fish P Larval dragonfly Salamander P Dragonfly larva P Fish P,C

M R M R M M M M

Dec Inc Dec Inc Dec Dec Decd Dec

Salamander larvae (2 Ambysroma SPP.) Frog larvae (R. aurora) Other Vertebrates Gerbils (2 Gerbilhus spp.) Rat (Ratt~issp.)

Large salamander P Alarm substance

R M

Inc (1 sp.), NR ( 1 SP.) Dec

Bank voles (Clethrionornys

Mammals C

M M R M

Dec Dec Inc Dec

Falcon P Owls P

M M

Dec Dec

Snake C

M

Dec

Owl P Cat P

Risk

7

Predatof’

Food

7

Hunger 7

NP

-

Manteifel (1995) Sih et al. (1988) Sih and Kats (1991) Sih et al. (1992) Skelly and Werner (1990) Skelly (1992) Skelly (1995) Stauffer and Semlitsch (1993) Walls (1995)

NP

-

Wilson and Lefcort (1993) Abramsky et al. (1996) Blanchard and Blanchard (1989) Jedrzejewska and Jedrzejewski (1990); Jedrzejewski et al. (1993) Korpimaki et al. (1996) Longland and Price (1991)

glareolus)

Field vole (Microtus agrestis) Desert rodents (4 heteromyids, 1 cricetid) Lizard (Lacerta viviparn)

Source

NP

-

Van Damme et al. (1990)

P, Predator present; C. chemical scent of predator: “alarm substance” usually refers to a chemical emanating from a killed or injured conspecific. R, Refuging; M. movement. NR, No response: Dec and Inc, decrease or increase, respectively, in the activity in question; -, no manipulation; NP, food not present. Response varied according to body size; some size classes may have been invulnerable to predators. ‘May see only under cryptic conditions.

‘

PREDATOR-INDUCED STRESS AND BEHAVIOR

231

in male activity to pursue their own activities free from male harassment, hence their atypical response to predator presence. Surprisingly few studies have examined an animal’s level of activity in the context of managing stress caused by a lower rate of feeding. In fact, food (or an identifiable impetus for nonzero activity) was not present in approximately 40% of the studies in Table 11. Food was present but unmanipulated in an additional 40% of studies; presumably, under these circumstances, a reduction in activity led to a decreased feeding rate. Studies manipulating food levels show mainly a decrease in activity with increasing food availability. Such a decrease is consistent with theoretical expectations (Abrams, 1991; Werner and Anholt, 1993), provided that risk increases with activity. The few studies manipulating an animal’s state show a consistent increase in activity (increased movement, decreased refuging) in energetically stressed animals. Such a state-dependent response is indicative of a trade-off between activity and the risk of predation (see Section 11,A). Underlying any functional explanation for a predator-induced decrease in activity is the assumption that increased activity raises the risk of predation. Presumably, increased activity raises the probability of being detected or encountered by a predator (but see also Houtman and Dill, 1994). This assumption receives support from several recent studies involving diverse predator-prey systems (e.g., Vaughn and Fisher, 1988; Daly et al., 1990; FitzGibbon, 1990; Rahel and Kolar, 1990; Everett and Ruiz, 1993; Otto, 1993; Heinen, 1994a; Anholt and Werner, 1995; Martel and Dill, 1995). Skelly (1994) provides a particularly nice demonstration of this effect by comparing predation on active and partially anesthetized tadpoles. Furthermore, a predator-induced increase in activity in Daphnia oregonensis (Ramcharan and Sprules, 1991) actually led to greater mortality. Interspecific patterns in predation linked to differing levels of activity (Hershey, 1987; Lawler, 1989; Chovanec, 1992; Azevedo-Ramos et al., 1992; Juliano et al., 1993; Grill and Juliano, 1996) provide further support for this important assumption.

B. TEMPORAL PATTERNS I N ACTIVITY

1. Die1 Vertical Migration by Zooplankton Zooplankton undertaking die1 vertical migration (DVM) descend to the depths during the day, and ascend to the surface at night; cases of reverse DVM (the opposite activity pattern) are also known (Ohman, 1990). Gliwicz and Pijanowska (1988) and Lampert (1989) note that, by the mid-l980s, many studies suggested that DVM is an adaptation against visually feeding predators rather than one related to the reduction of energetic stress, as

232

STEVEN L. LIMA

once thought. Recent work on DVM collectively provides an unusually comprehensive view of predator-induced stress and decision making. The antipredator hypothesis posits that animals engaged in DVM trade off the energetic benefits of remaining in the warm and food-rich surface layers against the safety of the dark, but relatively cold and food-poor deeper water (Lampert, 1989; Fiksen and Giske, 1995). Accordingly, the addition of a predatory stimulus to experimental test chambers induces or enhances DVM in many cladocerans (primarily Daphnia spp.; Dodson, 1988; Leibold, 1990; Dawidowicz and Loose, 1992; Dini and Carpenter, 1992; Young and Watt, 1993; Loose and Dawidowicz, 1994), copepods (Bollens and Frost, 1989b;Neill, 1992),and Chaoborus midges (Dawidowicz er al., 1990; Leibold, 1990;Tjossem, 1990). Similar effects occur upon wholelake additions or removals of planktivorous fish (Dini et al., 1993). These experimental results have been corroborated by field work showing that changes in DVM correspond closely to behavioral and distributional changes in planktivorous fish (Dini and Carpenter, 1988; Bollens and Frost, 1989a, 1991; Dodson, 1990; Levy, 1990a; Ohman, 1990; Ringelberg er al., 1991; Frost and Bollens, 1992). The way in which predation risk interacts with nonpredatory factors (e.g., food abundance, water temperature) to influence DVM is relatively unexplored territory. However, recent work suggests that DVM can be enhanced with the addition of food near the water’s surface (Leibold, 1990) or can be diminished with food addition to deeper water (Dini and Carpenter, 1992); observational evidence also suggests a strong effect of resource depth distribution on DVM (Gliwicz and Pijanowska, 1988). Fiksen and Giske (1995) suggest further that the effects of food abundance on optimal DVM may be markedly nonlinear and circumstance dependent. Theory also suggests that factors such as light transmission and water temperature may be important determinants of the optimal depth of DVM (Aksnes and Giske, 1990; Levy, 1990b; Fiksen and Giske, 1995), but there appears to be relatively little experimental work in this area. Gabriel and Thomas (1988) present a game-theoretical model of DVM suggesting that at evolutionary stability some members of a population may not engage in DVM. There is no strong evidence for such an effect (but see Guisande et al., 1991), although clonal (genetic) differences in DVM are known to occur (De Meester, 1993; De Meester et al., 1995). It is also known that species or size classes most vulnerable to fish predation tend to be those whose migratory behavior is most affected by changes in the predatory regime (Dodson, 1988; Ohman, 1990; Leibold, 1991; Neill, 1992;Watt and Young, 1994;see also Fiksen and Giske, 1995). Conspicuous, egg-carrying females may also be reluctant to ascend to the surface even under relatively dark conditions (Bollens and Frost, 1991).

PREDATOR-INDUCED STRESS AND BEHAVIOR

233

The proximate factors influencingDVM have also been examined. Chemicals emitted by predators are sufficient (and perhaps necessary) to induce DVM in most species studied (see Larsson and Dodson, 1993,for a review). Some progress has been made in characterizing the chemical(s) that signal the presence of predators (Parejko and Dodson, 1990; Loose et al., 1993). Rapidly changing light levels may also induce DVM (Ringelberg, 1991a,b; see also Clark and Levy, 1988), but zooplankton may initiate migration well in advance of changing light levels (Young and Watt, 1993). Studies examining the long-term stress induced by DVM associate slower growth (Dawidowicz and Loose, 1992; Gliwicz, 1994; Loose and Dawidowicz, 1994) and delayed reproduction (Vuorinen, 1987) with descending into the depths during the day. Loose and Dawidowicz (1994) argue that these costs of DVM are due mainly to the colder temperatures of deep water (see also Aksnes and Giske, 1990). Despite these costs, demographic analyses (Ohman, 1990;Bollens and Frost, 1991) suggest that DVM confers a net advantage if it results in even a modest lowering of the risk of predation.

2. Die1 Migration in Fish Fish may also engage in diel migrations, both vertical and horizontal (Helfman, 1986; Clark and Levy, 1988; Levy, 1990a,b;Gliwicz and Jachner, 1992). Clark and Levy (1988) outline several hypotheses for such migratory behavior, which parallel those proposed for zooplankton (see earlier discussion). One of these hypotheses suggests that DVM in planktivorous fish reflects little more than the DVM of their prey, but this alone cannot explain DVM in such fish (Clark and Levy, 1988; Levy, 1990b; Rosland and Giske, 1994). Furthermore, these fish may undergo DVM even in the absence of DVM in their prey (Gliwicz and Jachner, 1992; Rosland and Giske, 1994). Theoretical and empirical evidence suggests that DVM in planktivores reflects in part the risk imposed upon them by piscivores (Clark and Levy, 1988; Gliwicz and Jachner, 1992; Rosland and Giske, 1994). In any case, there appears to be little definitive experimental work on diel migration in fish.

3. Nocturnal versus Diurnal Activity Several recent studies show that animals will switch between nocturnal and diurnal activity, depending on the activity patterns of predators. Fenn and Macdonald (1995) showed that normally nocturnal rats may shift to diurnal activity in response to nocturnal activity by foxes. Such flexibility in rat activity was anticipated in recent psychological work on the patterning of rat behavior in response to threatening stimuli (Lester and Fanselow, 1992; Helmstetter and Fanselow, 1993). A literature review by McNeil et

234

STEVEN L. LIMA

al. (1992) suggests further that some birds may shift to nocturnal activity to avoid a strong diurnal risk of predation. Similarly, tiger moth (Spibsoma congrua) larvae become more nocturnal after diurnal encounters with wasps (Stamp and Bowers, 1993). On the more aquatic side of things, Culp and Scrimgeour (1993) and Cowan and Peckarsky (1994) showed that mayflies (Baetis spp.) switch from largely aperiodic to nocturnal feeding in the presence of visually hunting fish. Juvenile crayfish become more nocturnal in the presence of fish, but more diurnal in the presence of larger (and nocturnal) adult crayfish (Blake et al., 1994). 4.

Die1 Drift Periodicity in Stream Insects The tendency for large benthic stream insects to enter the nocturnal drift (to move via the current to a downstream site) has long been interpreted as an antipredator response, as these insects would be at risk to sizeselective fish predators in the diurnal drift (Allan, 1978). Flecker (1992) found support for this idea in a comparative study of streams with and without fish, and suggested that such nocturnal drift periodicity was a fixed (evolutionary) response to predation (see also Anderson et al., 1986; Malmqvist, 1988). However, much recent work shows clearly that at least some stream insects actively decide to enter the nocturnal drift in response to an increased local risk of predation (Williams, 1990; Poff et al., 1991; Andersen et al., 1993; Douglas et al., 1994; Forrester, 1994a,b; McIntosh and Townsend, 1994; Tikkanen et al., 1994). Rader and McArthur (1995) show further that the tendency of stoneflies to enter the nocturnal drift is reduced in habitats with abundant refuges. 5. Daily Activity Patterns and Body Mass in Birds

Bednekoff and Houston (1994) and McNamara et al. (1994) argue theoretically that patterns in the daily feeding activity of birds should reflect a trade-off between the costs (reduced speed or maneuverability) and benefits (reduced energetic stress) of carrying high fat reserves. These models suggest that such a trade-off can produce the bimodal daily feeding pattern commonly seen in birds (McNamara et al., (1994) even in the absence of die1 cycles in temperature, food availability, and so on. However, the behavioral consequences of such trade-offs have received little experimental attention (but see Witter et al., 1994). Observational evidence nevertheless suggests an important role for fat-reserve-related predatory effects in avian biology (Witter and Cuthill, 1993).

6. Nondiel Temporal Patterns in Activity a. Activity and the Lunar Cycle. The brighter portion of a lunar cycle represents a period of elevated risk for animals hunted by predators like

PREDATOR-INDUCED STRESS AND BEHAVIOR

235

owls. Accordingly, recent studies have demonstrated repeatedly that small, nocturnal mammals are relatively inactive under bright moonlight. This is the case in gerbils (Kotler et al., 1991, 1993a,b; Hughes and Ward, 1993; Hughes et af., 1994), for whom Kotler ei al. (1991) verify an elevated risk of owl predation under bright conditions. Kotler et al. (1994a) also found that gerbils reduce activity in anticipation of moonrise, indicating that the simple avoidance of light is not necessarily the proximate factor controlling lunar-based activity cycles. Recent work on heteromyid rodents also shows strong moonlight avoidance (Bowers, 1990; Daly et al., 1992; Bouskila, 1995; see also Lockard and Owings, 1974; but see Longland and Price, 1991). Daly et al. (1992) found that heteromyid kangaroo rats compensate for the lack of activity during periods of full moon by increased crepuscular activity, which actually makes them more vulnerable to diurnal predators. Work on murid rodents (in addition to gerbils; Wolfe and Summerlin, 1989; Simonetti, 1989; Dickman, 1992; Vasquez, 1994) and Old World porcupines (Brown and Alkon, 1990) indicates the same general trends in moonlight avoidance. The generality of moonlight avoidance in small nocturnal mammals is clear, but there appears to have been little recent work on nonmammalian species. However, Gliwicz (1986) and Dodson (1990) suggest that the lunar cycle can also affect the strength of die1 vertical migration in zooplankton. b. Activity on Other Time Scales. Nondiel patterns in activity have received relatively little attention outside of the context of the lunar cycle. However, tidal cycles may influence risk taking by refuging barnacles (Dill and Gillett, 1991) and migrating intertidal-feeding fish (Burrows and Gibson, 1995). On a shorter time scale, Speakman er al. (1995) suggest that temporal clumping in the nightly emergence of bats from maternity colonies represents an attempt by individuals to dilute the risk of owl predation. Kalcounis and Brigham (1994) nevertheless found that the presence of a vocal owl model had no impact on any aspect of bat emergence patterns. Activity cycles expressed over an entire season have received almost no attention. Lucas et al. (1996) provide an interesting exception in their dynamic game analysis of chorusing behavior in male frogs. Their analysis suggests that an interaction between predation risk, energetic stress, male density, and female behavior may produce pulses (or waves) of chorusing activity over the breeding season.

AN ENCOUNTER WITH A PREDATOR IV. AFTER

Recent work on postencounter decision making covers a variety of topics, such as the resumption of activity, the choice of escape behavior, and flight

236

STEVEN L. LIMA

initiation distance. In covering these topics, I focus on behavior that is flexible with respect to changes in the predatory environment; papers describing simple evasive behaviors in response to attack are outside the scope of this review. A. POSTENCOUNTER RESUMPTION OF ACTIVITY Prey typically reduce activity via reduced movement, increased refuging, or both, upon an encounter with a predator (Table 11). An animal must, of course, resume its normal activity at some point. The period of reduced activity may range from a few seconds in hermit crabs (Scarratt and Godin, 1992) to several days in small mammals (Jedrzejewski and Jedrzejewska, 1990; Kotler, 1992; Saarikko, 1992). However, despite the many activityrelated studies in Table 11, there is relatively little work on the factors affecting an animal’s decision to resume activity. One factor influencing the decision to resume activity is the nature of the predatory threat, with animals remaining inactive for longer periods in riskier situations (Scarratt and Godin, 1992; Sih, 1992a; Gotceitas and Godin, 1993; Johansson and Englund, 1995). Several recent studies also demonstrate that energetically stressed animals resume activity sooner than those well fed (Dill and Gillett, 1991;Gotceitas and Godin, 1991;Sih, 1992a; Koivula et al., 1995). Moore (1994) found that birds in a migratory state (with large energy reserves and a need to acquire even more) were more eager to resume feeding than nonmigratory birds after exposure to a hawk. Theoretical studies on the resumption of activity are few. However, Sih (1992a) provides a good theoretical discussion of the ways in which energetic stress and information combine to influence the postencounter resumption of activity in refuging prey. Stochastic dynamic programming could also be usefully applied to this temporal phenomenon, but apparently only one such model has been presented (KBlBs et al., 1995, dealing with the resumption of lekking following a predatory encounter). Johansson and Englund (1995) present a much-needed (but brief) game-theoretical perspective on the resumption of activity, which suggests that prey will generally outwait all but the most persistent predators. B. PURSUIT-DETERRENCE SIGNALS

Upon detecting a predator, an animal may signal that (1) the predator has been detected, and (2) it is able to escape; such signals should deter further pursuit. This mutually beneficial form of communication between prey and predator (Hasson, 1991) should be subject to some form of costbenefit analysis on the part of prey (Caro, 1995),but few studies have taken

PREDATOR-INDUCED STRESS AND BEHAVIOR

237

such a perspective. Car0 (1994) and Car0 et ul. (1995) provide exceptions in their thoughtful consideration of antipredator signaling in ungulates. Car0 (1994) argues convincingly that much antipredator behavior in African ungulates is pursuit-deterrence signaling. Similarly, Car0 et ul. (1995) conclude that tail flagging in white-tailed deer (Odocoileus virginiunus) functions as a pursuit-deterrence signal (see also Smith, 1991). The tail-flicking response of rails (Aves) to various aspects of predation risk also suggests that such behavior functions as a pursuit-deterrence signal (Alvarez, 1993). Furthermore, predator inspection behavior has been implicated as a form of pursuit-deterrence signaling in fish and mammals (see later discussion). On a theoretical note, Vega-Redondo and Hasson (1993) suggest that “honest” antipredator signaling can be evolutionarily stable depending on the processes by which predators and prey encounter each other.

C. FLIGHT INITIATION Prey often allow a predator to approach up to a certain point (the flight initiation distance, FID) before initiating escape behavior. Several recent studies complement earlier work (see Ydenberg and Dill, 1986; Lima and Dill, 1990), suggesting that FIDs increase in riskier situations, and are thus the outcome of a cost-benefit analysis by prey. A good example of such decision making occurs in woodchucks (Murmotu monax), which increase FIDs with an increase in the distance to the nearest refuge burrow (Bonenfant and Kramer, 1996) and when the predator approaches from the side opposite such a refuge (Kramer and Bonenfant, 1997);these studies complement similar work on tree squirrels (Dill and Houtman, 1989). Fish may increase their FID when far from a refuge (Dill, 1990) or when in smaller groups (Abrahams, 1995). Bulova (1994) also found a positive relationship between distance to refuge and FID in two iguanid lizards, and (surprisingly) a tendency toward shorter FIDs when approached directly by a predator (as opposed to a more tangential approach). There are still few studies examining nonpredatory influences on decisions regarding flight initiation. However, Scrimgeour and Culp (1994a) and Scrimgeour et al. (1994) found that FIDs in mayflies were shorter in patches with a better food supply. Gravid female lizards may have lower FIDs than nongravid females, perhaps reflecting the former’s relative inability to flee from predators (Braiia, 1993). D. CHOICE OF ESCAPE BEHAVIOR Animals generally have several escape options and may perform various escape maneuvers at differing intensities. Legault and Himmelman (1993)

238

STEVEN L. LIMA

showed that the intensity of evasive behavior in several molluscs and echinoderms varied positively with the danger posed by an encounter with a starfish; these results imply a cost to escalated escape behavior, but the nature of this cost was not clear. Dill etal. (1990) found that alarmed aphids were less likely to drop off high-quality plants than poor-quality plants, and suggested that aphid escape behavior is a function of both lost feeding opportunities (post-escape) and mortality associated with the extreme escape option of dropping off a plant. However, Stadler el al. (1994) found that aphids drop off plants more readily under better feeding conditions; this contradiction may be related to reproductive considerations. In related work, Cartar (1991) found that threatened worker bumblebees were relatively unlikely to initiate escape maneuvers (i.e., cease feeding) when their colony was under energetic stress. Finally, badgers faced with a dangerous predatory encounter choose the nearest available burrow for escape; they may seek a more distant but safer burrow with a lesser threat (Butler and Roper, 1994). A N D INSPECTING PREDATORS E. APPROACHING

There are many possible benefits and costs associated with the odd behavior of approaching predators, many of which are discussed by Dugatkin and Godin (1992a) in a wide-ranging review. Here, I focus my attention on the phenomenon of “predator inspection” by fish, which has received much attention in recent years. Predator inspection by fish usually involves one or more fish breaking away from a larger group to approach a predator (Dugatkin and Godin, 1992a). Such inspections may serve to gain information about the type of predator encountered (Magurran and Girling, 1986) or the predator’s readiness to attack (Licht, 1989). Dugatkin (1992) demonstrates a mortality cost to such behavior (but see Godin and Davis, 1995), and evidence suggests that inspectors assess these costs when approaching a predator. For instance, inspectors approach more closely when in larger groups, avoid a moving predator, and approach preferentially the tail end of the predator (Pitcher el al., 1986; Magurran and Seghers, 1990a; Dugatkin and Godin, 1992b). Larger fish, with presumably better escape abilities, may inspect more closely (Kiilling and Milinski, 1992) than smaller individuals. Energetically stressed fish may also inspect more than others, presumably because such fish must feed and thus have a greater need for information on predation risk (Godin and Crossman, 1994; McLeod and Huntingford, 1994). Predator inspection may also serve as a form of pursuit-deterrence signaling (Magurran, 1990; Godin and Davis, 1995; see also FitzGibbon, 1994, for a possible mammalian example).

PREDATOR-INDUCED STRESS A N D BEHAVIOR

239

A great deal of controversy surrounds the related claims that (1) pairs of inspecting fish are caught in the “prisoner’s dilemma,” and (2) such fish engage in a strategy of conditional cooperation resembling the tit-for-tat (TFT) strategy of Axelrod and Hamilton (1981). Evidence in favor of TFT cooperation suggests that inspecting fish exhibit the sort of reciprocation, retaliation, and forgivingness that one might expect in a TFT-like strategy (Milinski, 1987; Dugatkin, 1988; Milinski ef al., 1990a,b; Dugatkin and Alfieri, 1991a,b; Huntingford et al., 1994; see also Pitcher, 1992; Chivers et al., 1995b). Evidence against such a strategy suggests that inspectors may not be caught in the prisoner’s dilemma in the first place (and thus the TFT strategy would not apply; Magurran and Nowak, 1991; Murphy and Pitcher, 1991; Magurran and Seghers, 1994; Godin and Davis, 1995; Stephens et al., 1997). I cannot resolve this controversy, but much work clearly remains to be done regarding the nature of predator inspection.

V. SOCIAL SITUATIONS A. ADAPTIVE SOCIALITY

Decision making by individuals ought to influence the nature of sociality under the risk of predation (e.g., Pulliam and Caraco, 1984). The last few years have seen considerable progress in the study of such decision making, but there are still surprisingly few studies in this area (see also Lima and Dill, 1990; Krause, 199413). Recent years have also seen advances in the comparative study of predation and sociality (notably in primates; e.g., Boesch, 1991; Cowlishaw, 1994; van Schaik and Horstermann, 1994; Stanford, 1995), but such work is outside the scope of this review.

1. Spatial Position in Groups Fish may seek out the innermost (safest) area in a group when threatened by predators (Krause, 1993b). However, energetically stressed fish (Krause et al., 1992; Krause, 1993a) and aquatic beetles (Romey, 1995) may seek better feeding opportunities at their group’s (risky) periphery. A similar “spatial conflict” between feeding and safety may influence the location of web-building spiders within the larger colony (Rayor and Uetz, 1990, 1993). Krause (1994b) provides a cogent review of these and related studies on spatial positioning in social animals.

2. Choice of Group Larger groups should provide greater safety from predators than smaller ones, all else being equal. Accordingly, fish given a choice prefer larger

240

STEVEN L. LIMA

groups, especially when under a heightened risk of predation (Hager and Helfman, 1991; Ashley et al., 1993; Krause and Godin, 1994). Startled fish may also join the largest available group, unless this group is much farther away than a nearby, smaller group (Tegeder and Krause, 1995). A larger group in a risky area may also be avoided (Ashley et al., 1993), and a preference for larger groups may be overridden by a preference for individuals of a similar size (Krause and Godin, 1994; see also later discussion). Krause and Godin (1995) found that large groups of fish may suffer more attacks, but argue that prey are still better off in large groups (see also Wrona and Dixon, 1991; Uetz and Hieber, 1994). Poysa (1991) suggests that a duck’s choice of group may not be influenced by the risk of predation, although these ducks may have realized that the predator in question was not much of a threat. An SDP model by Szekely et al. (1991) suggests that energetically stressed birds should be less social (to avoid competitors) than those better fed. I know of no studies testing this prediction in birds, but Romey and Rossman (1995) describe such an effect in aquatic beetles. Paveri-Fontana and Focardi (1994) developed a model of optimal herd size selection in ungulates; they related the results to various ecological processes, but the model’s predictions for sociality per se were unclear.

3. Size-Assortative Grouping A small individual in a group of large individuals (or vice versa) may be conspicuous to predators and thus suffer a greater risk of attack (Wolf, 1985; Theodorakis, 1989). Such an effect may explain why fish in a group associate preferentially with others of their size under a heightened risk of predation (Theodorakis, 1989; Ranta et al., 1992a,b; Krause, 1994a; Krause and Godin, 1994). However, under such conditions larger fish may aggressively occupy the group’s central position, and thus preclude the intermingling of size classes irrespective of any effect of conspicuousness per se (Theodorakis, 1989; Krause, 1994a). B. VIGILANCE Many animals face a constant conflict between the need to be alert for attack and the need to feed. A ubiquitous observation is that individuals become progressively less vigilant (alert) as group size increases (see Elgar, 1989, for a brenchmark review). This “group size effect” is seen as an outcome of the fact that individual group members can devote less time to vigilance (i.e., more time to feeding) with increasing group size without detracting from the group’s collective ability to detect attack (Elgar, 1989).

PREDATOR-INDUCED STRESS AND BEHAVIOR

241

Interest in antipredatory vigilance has remained high in recent years, and has entered a period of retrospection and reassessment of old ideas. Insightful theory (Packer and Abrams, 1990; McNamara and Houston, 1992) and empirical studies (Krause and Godin, 1996) have elucidated some key issues in the maintenance of social vigilance in selfish animals. Refinements and challenges to the basic concept of collective detection have appeared (Lima and Zollner, 1996; Roberts, 1996). The group size effect itself has received better documentation (e.g., Roberts, 1995). Behavioral sequences involving vigilance have received much needed attention (Desportes et af., 1989; Roberts, 1994). Some exceptional observational studies of predator-prey interactions shed further light on social vigilance (e.g., Cresswell, 1994a). These studies and other developments have been reviewed by Roberts (1996). VI. REPRODUCTION Sih (1994) summarizes the current state of affairs with regard to reproductive decision making under the risk of predation: “Although predation risk is often viewed as an important component . . . of the evolution of mating behavior, . . . little effort has gone into gaining a deep, ecologically-rooted understanding of how predation risk influences reproductive behavior.” A similar sentiment is expressed in Lima and Dill (1990), Magnhagen (1991, 1993), and Reynolds (1993). Recent years have nonetheless seen considerable progress in understanding such reproductive behavior in many contexts. I review this work below, and in keeping with my overall theme, I focus on the management of predator-induced stress in ecological time. Sih (1994) provides an excellent discussion of the more general evolutionary and ecological aspects of reproductive behavior. I should note that “stress” in this section refers ultimately to a loss of reproductive output, which may or may not reflect a more standard form of stress (e.g., energetic) on the animal in question.

A. MATECHOICE Crowley et af.’s (1991) ground-breaking model of mate choice suggests that females should become less choosy with an increase in the risk of predation associated with locating mates. In other words, a given class of males will enjoy a diminished mating advantage under a high risk of predation. This prediction is supported by observations of predator-induced random mate choice in fish (Forsgren, 1992; Berglund, 1993). Godin and Briggs (1996) also report a predator-induced lowering of female choosiness in

242

STEVEN L. LIMA

guppies, but only in females from high-risk streams (but mate choice copying by such guppies may not be influenced by predation risk, Briggs et al., 1996). Similarly, the mating advantage enjoyed by longer-calling male crickets may be overridden if females can approach short-calling males in relative safety (Hedrick and Dill, 1993). On the other hand, large male water striders enjoy an increased mating advantage under a high risk of predation (Sih et al., 1990; Sih and Krupa, 1992, 1995,1996). This unusual result may reflect the female-harassment-based mating system in water striders (Krupa and Sih, 1993). Under a heightened risk of fish predation, males harass females less (i.e., become less active), which may then allow females to be more selective in their choice of mates or avoid mating altogether.

TACTICS B. ALTERNATIVE MALEMATING Male guppies may court females via conspicuous visual displays, or attempt “sneaky” forced copulations. Endler (1987) found that male guppies attempted more sneaky copulations in the presence of predators. Similar results have been reported in captive (Magurran and Seghers, 1990b) and free-living guppies (Godin, 1995). It is perhaps intuitive that male guppies would adopt the less conspicuous “sneaker” strategy in risky situations (see also Lucas and Howard, 1995; Lucas et al., 1996), but sneaky males may also be taking advantage of a female’s preoccupation with predator inspection in the presence of predators (Magurran and Nowak, 1991; Godin, 1995). One might envision other scenarios of predator-induced flexibility in alternative male mating tactics, but there appear to be no other reported cases. However, Magnhagen (1995) found that the riskier tactics used by sneaker and territorial common gobies (Pornatoschistus rnicrops) are used less frequently in the presence of predatory fish.

C. MATING DYNAMICS The act of mating itself may be influenced by the risk of predation. For instance, Travers and Sih (1991) found that male semiaquatic hemipteran insects accept lowered mating success under a high risk of predation by spending less time in tandem (copulating) with a female; tandem pairs make tempting targets for predators (Sih, 1988). Sih and Krupa (1995,1996) also found a decrease in mating duration and frequency in water striders in the presence of fish, presumably at some reproductive cost to males; tandem pairs once again are at greater risk than singletons (Fairbairn, 1993; Rowe, 1994). Razorfish (Xyrichtys splendens) spawn closer to the (safe) sea floor in high-risk situations, which may limit the dispersal success of

PREDATOR-INDUCED STRESS AND BEHAVIOR

243

resulting zygotes (Nemtzov, 1994). Finally, copulation frequency and number in pipefish (Syngnathus typhle) may decrease in the presence of predators (Berglund, 1993; Fuller and Berglund, 1996), but copulation time may increase to compensate (Berglund, 1993). D. COURTSHIP Conspicuous activities associated with courtship can lead to a higher risk of predation for males (Lima and Dill, 1990; Magnhagen, 1991). Hence, one might expect lowered courtship activity in the presence of predators. This has been observed in several fish species (Endler, 1987; Berglund, 1993; Forsgren and Magnhagen, 1993; Nemtzov, 1994; Chivers et af., 1995~). Area-specific differences in courtship activity by male fish may also be determined by the local abundance of predators (Hastings, 1991); Lister and Aguayo (1992) suggest that similar effects occur in lizards. Predators may also inhibit courtship and spermatophore deposition by male salamanders (Uzendoski et al., 1993). Following a predatory disturbance, the resumption of courtship chorusing by male frogs is quicker in larger groups, perhaps reflecting a greater dilution of risk in such groups (Jennions and Blackwell, 1992). E. OVIPOSITIONAL BEHAVIOR Mating dragonflies are sensitive to the presence of frogs in their choice of oviposition sites (e.g., Michiels and Dhondt, 1990). However, dragonflies appear unable to detect frogs lying in ambush (Rehfeldt, 1992). This inability may explain why dragonflies are attracted to groups of ovipositing pairs, as such groups form only in the absence of frog attacks (Rehfeldt, 1990, 1992). Regarding theory, Mange1 (1989) and Weisser et af. (1994) developed models of optimal ovipositional behavior by parasitoids searching in dangerous, patchy environments (see also Iwasa et af., 1984). The results suggest that optimal patch residence times should be sensitive to the risk of mortality experienced by ovipositing females. These models challenge the standard view that parasitoids should act only to maximize their rate of oviposition, but I know of no explicit tests of their predictions.

F. PREGNANCY AND PARENTING Observational evidence suggests that pregnant or lactating ground squirrels (MacWhirter, 1991) and bighorn sheep (Berger, 1991) take greater risks in order to meet the energetic stresses of mammalian reproduction.

244

STEVEN L. LIMA

In contrast, egg-carrying marine copepods avoid food-rich surface waters (Bollens and Frost, 1991); their opaque eggs make them vulnerable to detection by surface-feeding fish, even at night. Nest building and defense by male fish may also be predation-risk dependent. Magnhagen (1990) showed that nest building by male black gobies (Gobius niger) diminished in the presence of predators. The lack of such an effect in sand gobies (Pomatoschistus minutus) was attributed to their brief life-span (Magnhagen, 1990), which puts a premium on reproducing as soon as possible. Magnhagen and Vestergaard (1991) found that male common gobies took greater risks to defend their broods as their young matured (and presumably became more vulnerable); Magnhagen (1992) provides a general review of brood defense and parental risk taking in fish. Surprisingly few studies on nestling provisioning in birds consider risk to the parent to be an important determination of parental behavior (Ydenberg, 1994). However, Harfenist and Ydenberg (1995) suggest that rhinoceros auklet (Cerorhinca monocerata) chicks fledge younger and at lower body mass in areas frequented by eagles because parents terminate feeding earlier in high-risk areas. Such a decision is in accord with the predictions of Clark and Ydenberg (1990a,b). SUPPRESSION G. BREEDING

A growing body of work, focused almost exclusively on small boreal mammals (but see Fraser and Gilliam, 1992), addresses the issue of predation risk and the decision to engage in reproduction. Ylonen (1989) first reported that bank voles (Clethrionomys glareolus) strongly suppress reproduction upon exposure to mustelid predators. Similar degrees of breeding suppression have been observed in several other laboratory experiments on bank voles (Ylonen et al., 1992; Ronkainen and Ylonen, 1994; Ylonen and Ronkainen, 1994), other Clethrionomys voles (Ylonen et al., 1992; Heikkila et al., 1993), and Microtus voles (Koskela and Ylonen, 1995). Korpimaki et al. (1994) also demonstrate long-term breeding suppression in bank voles under field conditions. The mechanism behind this breeding suppression is not well understood. However, female Clethrionomys voles aggressively avoid male advances upon exposure to the scent of mustelid predators (Ylonen and Ronkainen, 1994; Ylonen, 1994). Male Microtus voles may themselves show less sexual activity in high-risk situations (Koskela and Ylonen, 1995). Energetic stress resulting from reduced feeding under high-risk conditions may also be involved (Heikkila et al., 1993) in suppressing breeding. Research into the estrous cycle of voles suggests that the mechanism behind breeding suppression has a strong physiological component (Koskela et al., 1996).

PREDATOR-INDUCED STRESS AND BEHAVIOR

245

This physiological link might conceivably relate to the negative effects of physiologicalheuroendocrine stress (caused by exposure to predators) on reproduction (Weiner 1992). Hansson (1995) suggests that reproduction in some boreal voles may be sensitive to physiological stress of any sort, not just that induced by predators. VII. LONG-TERM CONSEQUENCES OF DECISION MAKING Most studies on antipredatory decision making accept the idea that any decision has associated with it both a fitness cost (some form of predatorinduced stress) and benefit (avoiding an early death). How much do we really know about these issues? There are now several studies demonstrating that antipredator decision making does indeed lower an animal’s risk of predation (as per examples mentioned throughout this review). However, such benefits of antipredator decision making remain a presumption in many research programs, especially those involving terrestrial vertebrates. It is thus perhaps disturbing that a few studies have found antipredator responses to be inadequate in some way. For instance, strong refuging behavior in larval salamanders can be inadequate as a defense against fish predators (Sih et al., 1988; Sih, 1992b);a similar scenario is apparent in an amphipod predator-prey system (Sparrevik and Leonardsson, 1995). McPeek (1990), Werner and McPeek (1994), and Skelly (1995) report cases in which reduced activity in the presence of predators failed to prevent predation; however, these cases involved a lack of coevolutionary history between predator and prey. Demonstrations of the long-term costs of antipredator behavior are relatively uncommon. Recent years have nevertheless witnessed considerable progress in identifying and quantifying these costs (Table 111). A common theme in this work is that antipredator decisions that lower risk (usually habitat shifts or decreased activity) also lead to some form of energetic stress, typically manifest in lower growth rates. Slower growth may lead to a smaller size at maturity (Skelly and Werner, 1990) or prolonged development (Skelly, 1992). Exceptionally complete analyses of such predatorinduced stress, covering growth, development, and fecundity, have been possible in mayflies (Peckarsky et al., 1993; Scrimgeour and Culp, 1994b) and chironomids (Ball and Baker, 1995,1996). These insects have nonfeeding adult life stages, and thus reduced larval growth translates directly into reduced adult fitness (see also Feltmate and Williams, 1991; but see Duvall and Williams, 1995, for a more complicated situation in stoneflies). It is possible that a smaller size at maturity may reflect not only predatorinduced stress but also a predator-induced change in life history. There is,

TABLE 111 RECENT EXPERIMENTAL DEMONSTRATIONS O F A LONG-TERM COSTOF ANTIPREDATOR DECISION MAKING Prey

Invertebrates Chironomid larvae (Chironomus tentans)

2

Predatof

Prey response

Conditions

Fish P

Reduced activity

Laboratory

Cladoceran spp. Water flea (Daphnia magna)

Copepod P,C Fish C

Reduced activity (?) Vertical migration

Laboratory Laboratory

Damselfly larvae (Zschnura verticalis)

None

Reduced activity

Laboratory

Ant (Lusius pallitarsis) Dogwhelk (Nucellu lapillus) Mayfly larvae (Baetis bicaudatus)

Large ant P Crab C

Reduced activity Reduced activity

Laboratory Laboratory

Stonefly P

Escape-induced loss of feeding

Semifield

Marine snail (Strarnonita haemastoma) Mafly larvae ( B . tricaudatus)

Crab P

Reduced activity

Laboratory

Model fish P

Reduced activity

Laboratory

Cost

Source

Slower growth and development, lower adult mass at emergence, fewer eggs Slower growth Deeper migrators experience slower growth

Ball and Baker (1995, 1996)

Simulated predator-induced reduction in feeding slows growth and development Slower colony growth Slower (or zero) growth Adults emerge at lower mass, with fewer eggs (no effect on development time) Slower growth Slower growth, lower adult mass, longer development, fewer and smaller eggs

Gliwicz (1994) Dawidowicz and Loose (1992); Loose and Dawidowicz (1994) DixonandBaker (1988)

Nonacs and Dill (1990) Palmer (1990) Peckarsky et al. (1993)

Richardson and Brown ( 1992) Scrimgeour and Culp (1994b)

Buckmoth larvae (Hemileuca lucina) Copepod (Eurytemora hirundoides) Vertebrates Juvenile perch (Perca flu viatilis) Salamander larvae (Ambystoma maculatum) Guppy (Poecilia reticulata)

N

P 4

Juvenile roach (Rutilzu rutilus) Toad larvae (Bufo americanus) Tree frog larvae (Hyla versicolor) Tree frog larvae (2 Pseudacris spp.) Crucian carp (Carassius carassius)

Wasp P

Microhabitat shift

Semifield

Slower growth

None

Vertical migration

Laboratory

Simulated vertical migration leads to longer development

Fish P

Habitat shift

Semifield

Diehl and Eklov (1995)

Fish P

Reduced activity

Laboratory

Slower growth (due to increased competition) Slower growth

Fish P

Reduced activity, habitat shift Habitat shift

Field SemifieId

Reduced egg production and growth Slower growth

Reduced activity

Laboratory

Metamorphose at smaller size

Reduced activity

Semifield

Slower growth and development

Fraser and Gilliam (1992) Persson and Eklov (1995) Skelly and Werner (1990) Skelly (1992)

Reduced activity

Laboratory

Slower growth

Skelly (1995)

Habitat shift

Field

Slower growth (due to increased competition)

Tonn et al. (1992)

Fish P Larval odonate P Salamander P Larval odonate P Fish P

P, predator present; C, chemical scent of predator.

Stamp and Bowers (1991) Vourinen (1987)

Figiel and Semlitsch (1990)

248

STEVEN L. LIMA

however, no clear evidence for such adaptive life-history changes (Skelly and Werner, 1990; Ball and Baker, 1996). The current emphasis on predator-induced reductions in growth rates is entirely appropriate as most animals develop as free-living organisms for whom successful reproduction means reaching adult size (cf. Werner and Gilliam, 1984). However, for many birds and mammals, growth is often largely complete before they strike out on their own. For such creatures, the stress resulting from antipredator behavior is probably manifest in decreased body condition (e.g., Hik, 1995; Sinclair and Arcese, 1995), leading ultimately to lower female fecundity or male competitive ability. Such costs may also be manifest in energetic stress experienced by young being provisioned by parents attempting to avoid predation (Harfenist and Ydenberg, 1995). The survival-growthireproduction trade-offs apparent in Table 111 seem adaptive, given that an early death is the likely alternative to incurring some form of predator-induced stress. However, the degree to which “adaptive” approximates “optimal” is unknown. This should come as no surprise, given our inability to quantify many aspects of predation risk (see Section 11,E). Perhaps Nonacs and Dill (1990) come closest to making this distinction. They estimated the benefits to an ant colony from extra foraging and the cost of losing workers to predators, and found that the risks taken by workers reflected the potential increase in colony growth as a result of extra foraging.

VIII. ECOLOGICAL INFLUENCES AND IMPLICATIONS Decision making under the risk of predation can influence the nature of ecological systems. Understanding these influences has long been a major driving force in the study of antipredator decision making (Sih, 1980; Werner et al., 1983). Here, I discuss recent work in this area within three main contexts: the use of space by individuals, population-level consequences, and species interactions. This work involves mostly field or semifield experimentation. Although often not achieving the controlled rigor of laboratory experimentation, this work nevertheless illustrates the potential ecological effects of predator-induced stress and antipredatory decision making. A. USEOF SPACE

Table IV summarizes studies indicating that predators have a pervasive effect on the use of space by a variety of animals. This work adds to the many (but far fewer) studies on the use of space discussed in Lima and

TABLE IV RECENT STUDIES EXAMINING THE USEOF SPACE UNDER THE RISKOF PREDATION

Prey

Predatof

Scaleb

Results

Source

Invertebrates

Chironomid larvae (Chironomus tentans) Juvenile crayfish (Pacifmtacus leniusculus) Mayfly larvae (Baetis bicaudatus) Stonefly larvae (Paragnetina media) Epibenthic invertebrates (several spp.) Mayfly larvae (B. tricaudatus) Damsellly larvae (Ischnura venicalis) Whelk (Baccinum undarum) Isopod (Saduria entomon) Juvenile lobster (Homarus americanus) Hermit crabs (Clibanarius, Pagurus spp.)

Baker and Ball (1995)

m

No preference for predator-free areas (main response involved lower activity) No consistent preference for safer microhabitats

m

Avoid profitable but risky feeding locations

Cowan and Peckarsky (1994)

Fish

m

Feltmate and Williams (1989b)

Fish

M

Strong preference for color-matching substrate, (undiminished in absence of predator) Choose areas rich in refuges (woody debris)

Fish

m

Avoid profitable but risky feeding locations

Kohler and McPeek (1989)

Fish

m

Moum and Baker (1990)

Starfish

m, M

Large isopod Fish. crab

m m, M

Crab C, alarm substance

m

Strong preference for dark (safe) substrates, which may be enhanced in the presence of predators m: attracted to feeding starfish M: avoid areas with abundant starfish Avoid areas with abundant predators Predator-induced preference for safe, cobble substrate Crabs with ill-fittingshells seek areas with recently killedshell occupants;other crabsflee such areas

Fish

M

Shift from pelagic to littoral zone after

Brabrand and Faafeng (1993)

Fish

m

Fish, adult crayfish Fish C

Blake et al. (1994)

Everett and Ruiz (1993)

Rochette et al. (1995) Sparrevik and Leonardsson (1995) Wahle and Steneck (1992) Rittschoff et al. (1992)

Flsb

Roach (Rutilus rutilus) Stream fish (4 spp.)

Fish

m, M

Bluegill and shad (Lepomis and Dorosoma spp.)

Fish

m

predator introduction Juveniles and smaller species shifi to shallow water at both micro- and macroscales Only bluegill shift to shallow water in presence of predator

Brown and Moyle (1991) De Vries (1990) (continued)

TABLE IV (Continued) Prey

Predator"

Scaleh

Results

Source

Juvenile perch (Perca fluviatilis)

Fish

m, M

m: remain close to refuge in presence of predator M: avoid profitable but risky refuge-poor habitats

Perch and rudd (Perca and Scardinus spp.) Perch and roach (Perca and Rutilris spp.)

Fish

M

Large perch

m, M

Small stream fish (Riuuhcs and Poecilia spp.) Killifish (Rivulus hartii) Juvenile salmon (Oncorhynchus tshawytscha) Juvenile cod (Gadus morhua) Sculpin (Cottus bairdi) Small stream fish (mainly juv. Lepomis) Small, soft-rayed fish (4 SPP.) Stickleback (Gasterosteus aculeatus) Bleak (Alburnus alburnus)

Fish

M

Species segregate into pelagic vs littoral habitats based in part on vulnerability to predator m: remain close to refuge in presence of predator M: prefer refuge-rich habitat in presence of predator Avoid stream pools with predators; move to riffles

Eklov and Persson (1995); Christensen and Persson 1993); Persson (1991, 1993) Fraser and Gilliam (1992)

Fish Bird, fish

m, M m

Fish avoid streams populated by predators Prefer deeper water under nonturbid conditions

Fraser et al. (1995) Gregory (1993)

Large cod

m

Gotceitas and Brown (1993)

Fish Fish

m m, M

Fish

M

Fish

m

Predator-induced preference for safe, cobble substrate Microhabitat use unaffected by predator presence m: shift to shallow water in presence of predator M: avoid pools with predators Much emigration from lake (into outlet stream) following predator introduction Stay close to bottom in presence of predator

Fish, alarm substance Fish

m

Alarm substance

m

Arctic charr (Salvelinus alpinus) Fathead minnow (Pimephales promelas)

M

Preference for vegetated habitats is enhanced by predators and diminished by food in open water Ontogenetic shift to pelagic habitat is delayed under risky conditions Avoid areas marked with alarm substance

Diehl and Eklov (1995); Eklov and Diehl (1994); Persson and Eklov (1995 Eklov and Hamrin (1989)

Grossman et al. (1995) Harvey (1991) He and Kitchell (1990) Ibrahim and Huntingford (1989) Jachner (1995a,b) L'AbCe-Lund et al. (1993) Mathis and Smith (1992); Chivers et al. (1995a)

Perch (P.puviatilis)

Fish

M

Small fish (several spp.)

Fish, crab

M

Large pollock

m

Fish

m

Alarm substance

m

Fish

m, M

Large salamander

m

Eagle

m, M

Escape tactic may constrain birds to steep terrain

Bland and Temple (1990)

Raptors

m, M

m: juveniles feed in risky, profitable microhabitats

Cresswell (1994b)

Juvenile pollock (Theragra chalcogramma) Mosquitofish (Gambusia holbrooki) Brook stickleback (Culaea inconstans) Amphibians and Reptiles Salamander larvae (Ambystoma barbouri) Salamander larvae (2 Ambystoma spp.) fj Birds 3 Himalayan snowcock ( Tetraogallus himalay ensis) Redshank (Tringa totanus)

Titmice (2 Parus spp.)

Raptor

m

Willow tit (P. montanus)

Raptors

m

Sparrows (2 emberizid spp.) Anna’s hummingbird (Calypte anna) Small granivores (7 spp.. mostly emberizids)

Raptors Terrestrial birds

m, M m

Raptors

M

Choice of littoral (safe) or pelagic zone of lake determined by presence of non-gape-limited predator Preference for shallow water reflects risk in deep water Predator-induced preference for vegetated habitats Larvae may avoid adult cannibals by associating with predators that are avoided by adults Avoid areas marked with alarm substance

m: predator-induced preference for shallow water M: avoidance of pools with predators Shift to deeper water in presence of predator (one species only)

M: prefer less profitable but safe habitat (mussel beds) Feed in open (away from vegetation) only when forced to do so by aggression Feed in open only when forced to do so by aggression Willingness to feed in open related to escape tactic Avoid profitable feeding opportunities close to ground Large-scale habitat choice influenced by escape tactics

Persson et al. (1996)

Ruiz e f al. (1993) Sogard and Olla (1993) Winkleman and Aho (1993) Wisenden ef al. (1994)

Sih et al. (1992) Walls (1995)

Hinsley et al. (1995) Koivula et al. (1994) Lima (1990a) Lima (1991) Lima and Valone (1991) (continued)

TABLE IV (Continued) Prey Downy woodpecker (Picoides pubescem) Brambling (Fringilla montifringilla) Duck (Anas penelope) Sparrows (3 emberizid spp.)

Predatof

Scaleb

Raptors

M

Raptors, humans Raptors

m m

Small granivores (several spp.. old-world granivores) White-crowned sparrow (Zonotrichia leucophrys) Titmice (2 Parus spp.)

Raptors

m

Raptors

m

Raptors

m

Small birds (several spp.. mostly passerines)

Raptors

M

Blue tit (P. cueruleus) Sparrows ( 2 emberizid spp.)

Raptors Raptors

m m, M

Savannah sparrow (Passerculus sandwichensis) Mnmmals: Rodents Gerbils (2 Gerbillus spp.)

Raptors

m

Owls

m

Kangaroo rat (Dipodomys merriami)

Owls (?)

m

Results Choice of feeding site reflects vigilance-escape trade-off Prefer forest habitat over profitable but risky open habitat Reluctant to feed far from water (refuge) Avoid open areas, even those with high food density (except one sp.) General avoidance of relatively profitable but open areas Feed in open only when forced to do so by aggression Feed in open only when forced to do so by aggression Small (vulnerable) species avoid nesting in vicinity of (up to 1 km or more from) falcon nests Avoid profitable but open (risky) feeding sites Choice of feeding location influenced by escape tactics Reluctant to feed far from vegetated refuge

Avoid open (nonvegetated) areas when risk is increased under field conditions Avoid profitable but risky open microhabitats

Source Lima (1992) Lindstriim (1990) Mayhew and Houston (1989) Repasky and Schluter (1994) Schluter (1988) Slotow and Rothstein (1995) Suhonen (1993a,b; Suhonen et al. (1993) Suhonen et al. (1994) Todd and Cowie (1990) Watts (1990) Watts (1991)

Abramsky et al. (1996) Bowers (1990)

W

White-footed mouse (Peromyscus leucopus) Squirrels (Sciurus and Tamias, 2 spp.) Kangaroo rats ( 2 Dipodomys spp.) Desert rodents (2 heteromyid, 1 sciurid) Crested porcupine (Hystrix indica) Fox squirrel (S. niger)

Mammals, raptors

m, M

Avoid feeding opportunities in open habitats

Bowers and Dooley (1993)

Raptors

m

Avoid profitable but risky open microhabitats

Bowers et al. (1993)

Snakes

m

Bouskila (1995)

Raptors (mainly)

m

Large mammals

m, M

Raptors/mammals

m

Avoid feeding opportunities in vegetation that might be occupied by active snakes Avoid open areas, but kangaroo rats are more likely to be in open habitat than other species Avoid profitable feeding opportunities in open habitats Avoid profitable but risky open microhabitats

Gerbil (G. allenbyi)

Owls (?)

m (M?)

Guinea pig (Cavia aperea)

Raptordmammals

m

Prairie dog (Cynomys ludovicianus) House mouse (Mus domesticus) Gerbil ( G . tytonis)

Raptors

m

Mammals

m, M

Perceive increased risk in rocky habitats, which are usually avoided Appear to perceive higher risk when away from vegetation Avoid feeding far from refuge (burrow) unless feeding in groups Seek out vegetated habitats under increased risk

Raptors/mammals

m

Avoid profitable but risky open microhabitats

Field vole (Microtus agrestis) Gerbils (2 GerbiZZus spp.)

Raptors/weasels

m

Owls

m

Gerbils (2 Gerbillus spp.)

Snakes

m

Degu (Octodon degus)

Raptors

m

Desert rodents (4 heteromyids, 1 cricetid)

Owls

m

Avoid open areas in presence of kestrel, may avoid cover when in presence of weasels Avoid profitable but risky open microhabitats (can distinguish risk posed by different owl species) Avoid feeding opportunities in vegetation that might be occupied by active snakes Appear to perceive higher risk when away from vegetation Avoid open areas, but kangaroo rats are more likely to be in open habitat than other species

Brown (1989) Brown and Alkon (1990) Brown et al. (1992a); Brown and Morgan (1995) Brown et al. (1992b) Cassini (1991); Cassini and Galante (1992) Devenport (1989) Dickman (1992) Hughes and Ward (1993); Hughes et al. (1994) Korpimaki et al. (1996) Kotler (1992); Kotler et al. (1991, 1994a); Kotler and Blaustein (1995) Kotler et al. (1992, 1993a,b) Lagos et al. (1995a,b) Longland and Price (1991)

(continued)

TABLE IV (Continued) Prey

g *

Townsend’s vole ( M . townsendit) Desert rodents (2 heteromyids, 1 cricetid) Small rodents (5 spp.. mostly cricetids) Ground squirrels (Spermophilus and Tamias, 2 spp.) Porcupine (Enthizon dorsafum) Mammals: Nonrodent Bighorn sheep (Ovis canadensis) Hedgehog (Erinaceus europaeus) Pika (Ochotona cdlaris) Ibex (Capra ibex)

Jackrabbit (Lepus californicus) Buffalo (Syncerus cafer)

Predator“

Scale”

Results

Source

Mammal C

m

Avoid feeding opportunities in open habitats

Merkens et al. (1991)

Snake

m

Pierce e f al. (1992)

Raptors

m

Raptors, mammals

m

No consistent effect of snakes on use of space (on very constrained spatial scale) Avoid open areas (which may not be very profitable) Avoid profitable but risky open microhabitats: faster species feeds farther from cover

Mammals

m, M

Avoid feeding in open but more profitable habitats

Sweitzer and Berger (1992)

Large mammals

m, M

Berger (1991)

Badger

M

Pregnant sheep leave relative safety of steep terrain for better foraging May choose habitats in which predators are absent

Raptors, mammals Large mammals

m

Holmes (1991)

m (M?)

Raptorsimammals

m

Avoid profitable but risky microhabitats away from refuge Preference for cliffs over flat terrain may be due to increased perceived predation risk in latter habitat Perceive higher risk when away from vegetation

Lions

m. M

No clear indication that lions influence use of space, despite spatial variation in predation risk

Prins and Iason (1989)

C, chemical scent of predator only; otherwise predators were present in environment.

” m. microscale: M. macroscale.

Simonetti (1989) Smith (1995)

Doncaster (1993, 1994)

Kotler e f al. (1994b)

Longland (1991)

PREDATOR-INDUCED STRESS A N D BEHAVIOR

255

Dill (1990). In Table IV, the “microscale” category refers to an animal’s use of its immediate surroundings, very often in the vicinity of a refuge from attack. The “macroscale” category is more difficult to specify, but refers to a scale at which changes in the use of space require a significant investment in movement. The absolute spatial scale of macro- and microhabitat use is, of course, species-specific.

1. Invertebrates Studies on invertebrates (Table IV) show a tendency for individuals to avoid risky micro- or macrohabitats, even if such habitats offer good feeding opportunities. Rochette et al. (1995) describe an unusual case in which whelks avoid predatory starfish on a macroscale, but feed close to preyconsuming starfish on a microscale; starfish occupied by prey consumption are not dangerous, and produce “scraps” on which whelks can feed. Similarly, hermit crabs with ill-fitting shells may be attracted to areas of recent predation on gastropods in an attempt to obtain a better fitting shell; individuals with proper-fitting shells often flee from such area (Rittschoff et al., 1992). Note that I have already reviewed the use of space by certain invertebrates in other contexts. For instance, die1 vertical migration in zooplankton (see Section II1,B) involves a macroscale change in the use of the water column. Nocturnal drift in stream-dwelling arthropods (Section II1,B) also involves a macroscale change in location within a stream. Sih and Wooster (1994) and Wooster and Sih (1995) provide excellent reviews of drift behavior in stream animals and its consequences for local prey population regulation; subsequent work by Crow1 and Covich (1994), Forrester (1994a,b), Rader and McArthur (1995), and Kratz (1996) will also interest anyone working in this general area. Taking a different perspective, inadequate antipredator behavior may be a major determinant of the large-scale distribution of certain invertebrates. For instance, Daphnia aregonensis is largely absent from lakes occupied by a predator toward which its antipredator behavior is ineffective (Ramcharan and Sprules, 1991). Larval damselflies typical of fish-free ponds exhibit antipredator responses that are inadequate against the fish in permanent ponds, and vice versa (Blois-Heulin et al., 1990; McPeek, 1990;McPeek et al., 1996. Henrikson (1988) suggests similarly that inappropriate escape responses toward fish limit a libellulid dragonfly larva to fish-free lakes. Note, however, that these odonate larvae do not directly make decisions regarding their distribution among ponds or lakes; such decisions are made by ovipositing adults. 2. Fish Many recent studies demonstrate that predators are a major determinant of the use of space by fish. At the microscale, fish tend to remain in or

256

STEVEN L. LIMA

near safe habitats (e.g., shallow water, vegetation, safe substrates), while at the macroscale they avoid predator-rich or refuge-poor habitats (Table IV; see also Sih, 1987; Lima and Dill, 1990; Milinski, 1993). Eklov, Persson, and colleagues provide an unusually complete look at the use of space by small fish, which covers the spectrum from mechanistic studies of prey behavior (Eklov and Persson, 1996) to field studies examining whole-lake phenomena (Persson et al., 1996). Most fish-related studies in Table IV deal with lake systems, but the distribution of fish within and among stream pools is also influenced by predators (Brown and Moyle, 1991; Harvey, 1991; Fraser and Gilliam, 1992; see also Power et al., 1985; Schlosser, 1987; but see Grossman et al., 1995). Along these lines, Fraser et al. (1995) link small-scale decisions regarding the use of space to whole-drainage patterns in the distribution of killifish.

3. Amphibians and Reptiles Work on the use of space by these animals has been limited (Table IV), and there is a clear need for work on reptiles. The few existing studies suggest that predator-induced effects in larval amphibians are similar to those seen in fish (see also Lima and Dill, 1990). Some studies also show that larval amphibians stay as far from predators as possible in small laboratory containers (e.g., Hews, 1988; Skelly and Werner, 1990), suggesting that their microhabitat use might be predation-risk dependent. Morey (1990) and Heinen (1993, 1994b) also found that frogs and toads, respectively, choose substrates against which they are most cryptic; this has obvious implications for the use of space under the risk of predation. On a large scale, some studies link inadequate antipredator behavior to the distribution of larval amphibians within streams (Sih, 1992b; Feminella and Hawkins, 1994) or among temporary versus permanent ponds (Kats et al., 1988; Werner and McPeek, 1994). As in similar cases with invertebrates, however, the choice of temporary versus permanent ponds is made not by these larvae but by adults (Resetarits and Wilbur, 1989). 4.

Birds

An emerging avian theme is that the use of space relative to vegetative cover is determined to a large extent by escape tactics (Lima, 1993). Birds with vegetation-dependent escape tactics are reluctant to feed far from vegetative cover (Table IV). Observations of raptor predation on birds confirm the adaptive nature of this reluctance to feed in the open (Watts, 1990; Suhonen, 1993a,b; Hinsley et al., 1995). Although less well studied, birds with vegetation-independent tactics may avoid vegetative cover altogether (Lima, 1993).

PREDATOR-INDUCED STRESS AND BEHAVIOR

257

Most avian studies take a microscale perspective (Table IV), but patterns at this scale may also translate to larger spatial scales (Lima and Valone, 1991; Watts, 1991). Bland and Temple (1990) describe a situation in which a bird’s gravity-assisted, downhill escape tactic may explain its geographic restriction to mountainous terrain. Birds may enter macrohabitats not well suited to their escape tactics, or relatively risky macrohabitats, only (1) if forced to do so by aggression (e.g., Cresswell, 1994b), or (2) if such habitats offer exceptional foraging opportunities (e.g., Lindstrom, 1990). On a different note, the location of falcon nests may also influence the large-scale distribution of breeding passerines (Suhonen et al., 1994).

5. Mammals Recent work shows convincingly that small mammals (mostly rodents) avoid feeding far from protective cover (e.g., vegetation), even at the cost of forgoing high-quality feeding opportunities (Table IV). Thermophysiological stress in the open cannot account for the avoidance of open areas (Bozinovic and Simonetti, 1992; Sweitzer and Berger, 1992; Kotler et al., 1993d; Bowers et al., 1993; Lagos et al., 1995a), but such effects deserve more attention. The strong attraction of woody vegetative cover for desert rodents can be reduced or reversed when such vegetation harbors predatory snakes (Table IV). In this regard, Kotler et af. (1992) and Korpimaki et al. (1996) note the possibility of “predator facilitation” in which the avoidance of vegetative cover makes prey more available to open-hunting predators (or vice versa; see also Daly et al., 1992). Schooley et al. (1996) note furthermore that vegetation may present obstacles to escape and predator detection for some diurnal rodents, hence their preference for open areas. Work on large mammals is sparse and mixed (Table IV). Predation risk may be a factor in the use of space by bighorn sheep (Berger, 1991) and ibex (Kotler et al., 1994b), but perhaps not by African buffalo (Prins and Iason, 1989). Work on the use of space by mammals usually focuses on small spatial scales (Table IV). Doncaster’s (1993, 1994) work on hedgehogs provides a notable exception. It nevertheless seems likely that the ubiquitous microscale avoidance of open areas by small mammals (Table IV) will translate to larger spatial scales. In other words, habitats with little vegetative cover will probably be avoided by animals reluctant to forage away from such cover (see also Price ef al., 1994). B. POPULATION-LEVEL CONSEQUENCES Antipredatory decision making could in principle influence many aspects of prey population dynamics and regulation (e.g., Desy et al., 1990; Chesson

258

STEVEN L. LIMA

and Rosenzweig, 1991; Schluter and Repasky, 1991; Sinclair and Arcese, 1995). This possibility is readily apparent given the long-term negative effects of predator-induced stress (Table 111). However, translating behavioral decisions to their population-level consequences has proven difficult. Actually, the extent to which this is true depends on the scale of analysis. The influence of predators on local population dynamics can often be understood in terms of decisions affecting the large-scale distribution of animals (see previous section). Nevertheless, studies covering whole populations are unusual. The “whole population barrier” has been broken by some experimental studies focusing on small lakes in which entire populations of predators and prey can be manipulated and monitored (although often with limited replication). He and Kitchell (1990) provide a particularly good case in point. They showed that the “crash” in the prey population following the introduction of pike into a lake was caused by a large-scale movement of prey fish out of the lake and into the outflow stream (see also H e and Wright, 1992). Tonn et al. (1992) also performed a whole-lake manipulation of predators. In this case, predatory perch induced an almost exclusive use of the shallow (safe) littoral zone by young crucian carp. This led to a competitive bottleneck that ultimately limited recruitment to adult life stages relative to a control population (see also Diehl and Eklov, 1995). Individuals surviving this bottleneck grew much larger than control fish after shifting to the competition-free pelagic zone. This scenario parallels that in Werner el al.’s (1983) landmark study in a bass-sunfish system. Recent work in similar systems suggests that such bottlenecks can alter the competitive relationship among prey species (Brabrand and Faafeng, 1993; see also next section). Furthermore, an understanding of these predatorinduced bottlenecks can provide insight into the nature of stock-recruitment relationships of importance to fisheries management (Walters and Juanes, 1993). Models of predator-prey population dynamics abound (Crawley, 1992), but very few incorporate adaptive antipredator behavior. Abrams (1993b) argues that most predator-prey models actually suffer from assumptions not easily supported by adaptive antipredator behavior. Ruxton (1995) found that adaptive antipredator behavior acts to stabilize otherwise oscillatory predator-prey population dynamics, complementing results from earlier modeling (Ives and Dobson, 1987). Crowley and Hopper (1994) present an extraordinary modeling attempt linking a stochastic-dynamic game between predator and prey to stock-recruitment curves and resulting population dynamics. Predator-prey population cycling might also be influenced by antipredator decision making by prey. Hik (1995) presents evidence that energetic

PREDATOR-INDUCED STRESS AND BEHAVIOR

259

stress following a predator-induced microhabitat shift by snowshoe hare (Lepus arnericanus) causes a lowering of hare reproductive output, which then hastens the decline and lengthens the recovery phase in the cyclic population dynamics of hare and their mammalian predators. Similarly, Ylonen (1994) and Oksanen and Lundberg (1995) suggest that predatorinduced breeding suppression (see Section VI,G) hastens the crash phase in the cyclic population dynamics of boreal voles and their mustelid predators. Ylonen (1994) outlines the specific idea that breeding suppression represents an attempt by female voles to ride out (in a high-survival, nonreproductive state) the high-predation part of a population cycle, after which they and their offspring would have a better probability of survival. Lambin et al. (1995) leveled some harsh criticism against this idea regarding breeding suppression and vole population dynamics, claiming that many of its key assumptions are unsupported (especially the assumption of enhanced survivorship in nonreproductive females). Ylonen’s idea still has considerable merit, but there is clearly a need for critical experimentation and quantitative modeling regarding the role of breeding suppression in predator-prey population dynamics.

C. SPECIES INTERACTIONS Recent studies illustrate how antipredator decision making might influence species interactions. These studies emphasize the role of indirect interactions between predators and other species mediated by the predators’ effect on the behavior of a third (transmitter) species (Abrams, 1995). Such indirect interactions have been termed higher order interactions (Werner, 1992) or trait-transmitted indirect effects (Abrams, 1995), but for clarity I will use the term behaviorally transmitted indirect effects. Behaviorally transmitted indirect effects may act in a variety of ways to alter the outcome of interspecific competition (Werner, 1992). For example, similar refuging behavior under a high risk of predation may lead to one (transmitter) species excluding another from the refuges. This has the effect of leaving the lesser competitor exposed to greater predation, which may ultimately tip the competitive balance in favor of the transmitter species. Such a scenario may apply in fish-crayfish systems (Hill and Lodge, 1994; Soderback, 1994) and a fish-salamander-isopod system (in which fish consume both salamanders and isopods; Huang and Sih, 1990). Werner (1991) argues that greater larval bullfrog activity (movement) in the presence of predators gives them a competitive advantage over larval green frogs; these two species are evenly matched competitors in the absence of predators. This effect of differential activity ultimately interacts with direct predatory effects in determining the distribution of these two species among perma-

260

STEVEN L. LIMA

nent versus temporary ponds (Werner, 1994; Werner and McPeek, 1994). Similar movement-related effects may influence competition between larval mosquitos ( Juliano et al., 1993; Grill and Juliano, 1996). On the other hand, Tayasu et af. (1996) argue on empirical and theoretical grounds that similar levels of predator-induced inactivity in two shrimp species may allow for coexistence that would not otherwise be possible. Here, lowered activity in the superior competitor favors coexistence via a reduction in the overall level of interference competition. Behaviorally transmitted indirect effects may also be evident when predators influence a particular species’ use of space (Werner, 1992). Leibold (1991) describes a case in which competitive exclusion between two zooplankton species may be prevented by a predator-induced habitat shift in the superior competitor (the transmitter species). Cases have also been reported in which the similar use of space in the presence of predators intensifies interspecific competition among fish (Person, 1991, 1993; Brabrand and Faafeng, 1993) and desert rodents (Hughes et al., 1994). Finally, recent work on gerbils provides a cautionary tale regarding the use of space and its ultimate effects on species interactions. Despite the fact that two competing gerbil species may use space differently in the presence of predators (Kotler et al., 1991), the temporal partitioning of activity appears to form the basis for their coexistence (Kotler et al., 1993c; Ziv et al., 1993; Brown et al., 1994). Behaviorally transmitted indirect effects have also been implicated in cases of strong “top-down’’ ecosystem regulation; such regulation dictates that a change in the abundance of top predators causes indirect ecological effects, which are transmitted all the way down to the lowest trophic levels of a food web (Power, 2992). For instance, Turner and Mittelbach (1990) found that the strong indirect effect of piscivorous bass on zooplankton communities is transmitted by predator-induced changes in the use of space by planktivorous sunfish. Diehl and Eklov (1995) and Person ef al. (1996) describe a very similar situation in a piscivore-+perch-+invertebrate trophic system (arrows indicate predator-prey relationships). In a sunfish-salamander+isopod system simulated by Huang and Sih (1991). a positive effect of fish on isopods is transmitted primarily via a strong refuging response by salamanders to the presence of fish. Turner (1997) provides an extreme case of behaviorally transmitted top-down effects in a simulated predator-+snail+algae system in which the mere chemical scent of predation drives the system. Finally, Hill and Lodge (1995) describe a case in which the (nonlethal) presence of predators mediates top-down effects system via both behavioral in a fish+crayfish-+macroinvertebrate+plant changes and increased mortality in crayfish (the latter being caused by increased fighting for refuges).

PREDATOR-INDUCED STRESS A N D BEHAVIOR

261

The importance of behaviorally transmitted indirect effects in ecological systems has also been explored theoretically in recent years. Abrams (1992,1995) and Abrams and Matsuda (1993) make a convincing case that (1) community-level models ignoring such indirect effects may be misleading, and (2) a variety of indirect effects may be expected if both predator and prey can change their behavior adaptively (see also Kotler and Holt, 1989). Abrams (1995) notes also that such adaptive behavioral traits may make it difficult to even distinguish and classify direct versus indirect effects. Indirect effects also figure prominently in models suggesting that ecological communities will be more speciose if prey exhibit predator-specific rather than generalist antipredator behavior (Matsuda el al., 1993,1994,1996; see also Brown and Vincent, 1992, for a different perspective on this issue).

IX. ADDITIONAL CONSIDERATIONS In this section I group four disparate topics about which relatively little is known. These topics nonetheless address several important issues in the study of decision making under the risk of predation. AND DECISION MAKING STRESSRESPONSE A. PHYSIOLOGICAL

A threatening situation often induces the classic “fight or flight” physiological (neuroendocrine) stress response, which involves (among other things) the immediate production of hormones like cortisol, epinephrine, and norepinephrine (Weiner, 1992); recent work suggests that this response is even greater than previously thought (Le Maho et al., 1992). One of the short-term physiological effects of the basic stress response is to make more energy available for immediate action like escape (Weiner, 1992). Many stimuli will produce this stress response, such as aggressive conspecifics, unfamiliar terrain, novel objects, and so on (Boissy, 1995). Of course, predators may also induce such a response, but relatively little work addresses the effects of predators per se (but see Levine et al., 1993; Boissy, 1995). However, work on stress caused by being approached or handled by humans (Le Maho etaf.,1992; Boissy, 1995) has an obvious relationship to physiological stress caused by predators. The physiological stress response is well known, but its relationship to antipredator decision making represents unexplored territory. Indeed, the relationship between the basic stress response and subsequent behavior is not always clear (Boissy, 1995). Experimental work in which the stress response is chemically blocked does suggest, however, that elevated levels of stress hormones affect (in part) various antipredator behaviors (Berco-

262

STEVEN L. LIMA

vitch et al., 1995). Boissy (1995) argues further that individual differences in “fearfulness” among animals are related causally to such differences in the stress response. It thus seems likely that the physiological stress response is mechanistically linked in some way to antipredator decision making. It is, in fact, conceivable that the stress response is to a significant extent a target of selection in the evolution of antipredator behavior in general, especially as it relates to short-term changes in responsiveness to predators. It is also conceivable that an unusually extreme stress response may actually impair decision making in some way; Mesa et al, (1994) suggest such a possibility with regard to non-predator-induced physiological stress, but the same might well hold for stress caused by chronic exposure to unusually high predation risk (see also following discussion). It is also tempting to speculate further that certain aspects of antipredator decision making are designed to avoid the long-term effects of a chronic physiological stress, such as stress-induced diseases and suppression of the immune system (Ader et al., 1991); such a realization may have important implications for the design of experiments on antipredator behavior (see following discussion). As mentioned earlier, the reproductive effects of such physiological stress may also impinge on our interpretation of predator-induced breeding suppression (see Section V1,G). All of the forgoing discussion on physiological stress pertains to vertebrates. In fact, most research has been conducted on only a small number of mammals, birds, and fish of economic or medical importance (Schreck, 1990; Mesa etaf., 1994; Boissy, 1995). The results obtained thus far probably apply to most vertebrates, but their relevance (if any) to physiological stress and the antipredator behavior of invertebrates seems largely unexplored. B. ASSESSING THE RISKOF PREDATION

An assessment of the risk of predation must in some way form the basis for antipredator decision making (Blumstein and Bouskila, 1996), but little is known about the way in which such assessments are made. A great deal is known about the sorts of predatory stimuli that animals interpret with alarm (see Curio, 1993, for an excellent discussion), but the way in which animals integrate information on predator abundance, the likelihood of escape, and so on, into some sort of assessment of predation risk is unknown. Following Lima and Dill (1990), it seems likely that animals use “rules of thumb” in assessing the prevailing risk of predation. It also seems likely that any such assessment will be fraught with uncertainty. In this regard, Bouskila and Blumstein (1992) argue that animals might adaptively overestimate the risk of predation to avoid the relatively high costs of underesti-

PREDATOR-INDUCED STRESS A N D BEHAVIOR

263

mating risk. Abrams (1994) cautions, however, that underestimating the risk of predation can be favored under certain circumstances. Work on the chemical detection of predators might shed light on this issue of assessing risk. Scores of studies show that a variety of animals can detect a threat of predation via chemicals emitted by predators (for recent reviews, see Weldon, 1990; Smith, 1992; Larsson and Dodson, 1993; Dodson et al., 1994; Kats and Dill, 1998). As argued by Kats and Dill (1998), the concentration of such chemicals might provide an accurate estimate of predation risk. This might explain why the strength of antipredator behavior in zooplankton (Ramcharan et al., 1992; Loose and Dawidowicz, 1994) and tadpoles (Horat and Semlitsch, 1994) increases with the concentration of fish-emitted chemicals. However, very few studies examine behavioral responses to varying chemical concentrations, nor have such concentrations been related to mortality, predator abundance, and so on. Future work in this area might well demonstrate that predator-emitted chemicals provide many types of animals with an accurate estimate of the risk of predation (Kats and Dill, 1998).

C. PREYACTION A N D PREDATOR REACTION The study of antipredatory decision making is hindered by a lack of information on the way in which predators respond (in ecological time) to the antipredatory actions of their prey. In fact, a tacit assumption in the vast majority of studies reviewed herein is that factors like attack rate are fixed entities to which prey determine their optimal response. There are nonetheless many scenarios in which prey behavior might influence predator behavior (and thus the components of risk controlled by predators, e.g., Lima, 1990b). The smattering of studies addressing this issue of “action and reaction” cover a wide range of phenomena. Johansson and Englund (1995) consider explicitly the behavioral interaction between a refuging prey and a persistent predator. Piscivorous perch change from an active to a sit-and-wait foraging mode when their prey shift from an open to a refuge-rich habitat (Eklov and Diehl, 1994). Of conceptual importance in the study of vigilance are observations that predators avoid attacking relatively vigilant prey (FitzGibbon, 1989; Krause and Godin, 1995). O n a different note, piscivorous pike may defecate away from their feeding areas so as to avoid being detected chemically by prey (Brown et al., 1995). Recent attempts to model multi-trophic-level games of habitat selection (Schwinning and Rosenzweig, 1990; Hugie and Dill, 1994; Sih, 1998) provide notable instances in which the crux of the matter is the real-time interaction between prey response and predator reaction.

264

STEVEN L. LIMA

D. SCALING TO THE REALWORLD To what extent d o small-scale laboratory microcosms simulate the situation faced by animals avoiding predators in their natural environment? Lima and Dill (1990) raised this question with regard to the common experimental situation in which predator and prey are maintained in very close proximity. Under such situations, the prey’s response to predators may be so strong as to be potentially misleading. Richardson and Brown (1992) report just such a situation in which a strong response by snails to nearby crabs in the laboratory could not be replicated under field conditions. Similarly, the relatively brief reduction in gerbil activity following an encounter with an owl under semifield conditions (Abramsky et al., 1996) did not reflect the marked reduction in gerbil activity following an exposure to captive owls at close quarters (Kotler, 1992). Perhaps the application of most laboratory studies to the real world would not be so problematic, but I nevertheless urge caution in the use of experimental protocols in which prey and predator are in close proximity. Such caution may also be warranted in light of the possibility that decision making may be impaired by an abnormally intense physiological stress response under these circumstances (see Mesa et al., 1994). The general issue of “scaling to the real world” concerns not just the spatial proximity of predator and prey, but also the temporal scale of the interaction. Many studies demonstrate that animals respond markedly to a brief but acute exposure to predators, perhaps with a complete cessation of feeding. In effect, these animals are able to “ride out” a short period of high risk. However, such strong responses may not be indicative of those to a chronic exposure to high risk; animals must eventually eat.

X. CONCLUSIONS A N D SUMMARY Recent years have witnessed increasing interest in the study of antipredatory decision making and its consequences. This recent work is much too vast to summarize in detail, but some notable recent advances include clear demonstrations that antipredatory decision making (1) may influence many aspects of reproductive behavior, (2) has demonstrable long-term consequences for individual fitness, and (3) may influence the nature of ecological systems themselves. There have also been many advances in the theory of antipredator behavior, which should provide a sound conceptual basis for further progress. Overall, combined with earlier work (Sih, 1987; Lima and Dill, 1990), these recent advances lead to the inescapable conclusion that the risk of predation may influence any aspect of animal decision making. Just about all of the areas covered in this review deserve more attention. This is particularly true of areas that have emerged most recently. In this

PREDATOR-INDUCED STRESS AND BEHAVIOR

265

regard, of great value would be further work on the effects that predator and prey have on the other’s behavioral decisions. The range of reproductive behaviors influenced by the risk of predation also requires much more investigation. Work on the long-term costs of antipredator decision making needs more empirical documentation and greater taxonomic diversity. Work on the ecological implications of antipredatory decision making has only “scratched the surface,” especially with regard to population-level effects and species interactions. Theoretical investigations should also play a prominent role in future work. While I am not sanguine about the possibilities that such theoretical models can be tested quantitatively, theory is nevertheless essential to the continued conceptual development of the field. Finally, I suspect that research exploring the link between antipredator decision making and the physiological stress response will prove rewarding. What are the next “big steps” in the study of decision making under the risk of predation? Two areas seem to have particularly good prospects. The first concerns the aforementioned application of antipredator decision making to the understanding of ecological systems. Such work will be particularly interesting given that the early development of behavioral ecology was spurred (in part) by the prospect that behavioral studies might provide key insights into the workings of ecological systems; this prospect may well be realized in the study of predator-prey interactions. The second area concerns the development of a view of antipredator decision making that encompasses phenomena expressed over both ecological and evolutionary time. Work in this area promises to integrate the study of antipredator decision making with recent advances in the larger field of evolutionary biology. I have not been able to cover this emerging area to any great extent, but Sih (1992b) and McPeek etal. (1996) provide thoughtful discussions and examples of how such an integration might proceed.

Acknowledgments

I thank Peter Slater. Manfred Milinski, and Anders M d l e r for their comments on the manuscript, and their efforts regarding this volume on stress and behavior. Peter Bednekoff and Patrick Zollner also commented on the manuscript. Chris Mathews provided competent assistance with the literature search. Hilary Philpot helped in the preparation of the References section. Finally, much of the background work in preparing this review was made possible by a sabbatical leave granted by Indiana State University, for which I am most grateful.

References Abrahams, M. V. (1995). The interaction between antipredator behaviour and antipredator morphology: Experiments with fathead minnows and brook sticklebacks. Can. J. Zool. 73,2209-2215.

266

STEVEN L. LIMA

Abraham, M. V., and Dill, L. M. (1989). A determination of the energetic equivalence of the risk of predation. Ecology 70, 999-1007. Abrams, P. A. (1991). Life history and the relationship between food availability and foraging effort. Ecology 72, 1242-1252. Abrams, P. A. (1992). Predators that benefit prey and prey that harm predators: Unusual effects of interacting foraging adaptations. Am. Nat 140,573-600. Abrams, P. A. (1993a). Optimal traits when there are several costs: The interaction of mortality and energy costs in determining foraging behavior. Behav. Ecol. 4, 246-253. Abrams, P. A. (1993b). Why predation rate should not be proportional to predator density. Ecology 73,726-733. Abrams, P. A. (1994). Should prey overestimate the risk of predation? Am. Nat. 144,317-328. Abrams, P. A. (1995). Implications of dynamically variable traits for identifying, classifying. and measuring direct and indirect effects in ecological communities. A m . Nat. 146, 112-134. Abrams, P., and Matsuda, H. (1993). Effects of adaptive predatory and anti-predator behaviour in a two-prey-one-predator system. Evol. Ecol. 7, 312-326. Abramsky. Z . , Strauss, E., Subach, A,, Kotler, B. P., and Riechman, A. (1996). The effect of barn owls (Tyto alba) on the activity and microhabitat selection of Gerbillus allenbyi and G. pyramidum. Oecologia 105, 313-319. Ader, R., Felten, D., and Cohen, N., Eds. (1991). “Psychoneuroimmunology.” Academic Press, San Diego, CA. Aksnes, D. L., and Giske, J. (1990). Habitat profitability in pelagic environments. Mar. Ecol.: Prog. Ser. 64,209-215. Alexander, J . E., Jr., and Covich, A. P. (1991a). Predator avoidance by the freshwater snail Physella virgata in response to the crayfish Procambarus simulans. Oecologia 87,435-442. Alexander, J. E., Jr., and Covich, A. P. (1991b). Predation risk and avoidance behavior in two freshwater snails. Biol. Bull. (Woods Hole, Mass.) 180, 387-393. Allan, J . D. (1978). Trout predation and the size composition of stream drift. Limnol. Oceanogr. 23, 1231-1237. Alvarez, F. (1993). Alertness signalling in two rail species. Anim. Behav. 46, 1229-1231. Andersen, T. H., Friberg, N., Hansen, H. 0.. Iverson, T. M., Jacobsen, D., and Krojgaard, L. (1993). The effects of introduction of brown trout (Salmo trutta L.) on Gammarus pulex L. drift and density in two fishless Danish streams. Arch. Hydrobiol. 126,361-371. Anderson, K. G., Bronmark. C.. Herrmann, J., Malmqvist, B.. Otto, C., and Sjorstrom, P. (1986). Presence of sculpins (Conus gobio) reduces drift and activity of Gammarus pulex (Amphipoda). Hydrobiologia 133, 209-215. Anholt. B. R., and Werner, E. E. (1995). Interaction between food availability and predation mortality mediated by adaptive behavior. Ecology 76, 2230-2234. Ashley, E. J.. Kats. L. B., and Wolfe. J. W. (1993). Balancing trade-offs between risk and changing shoal size in northern red-belly dace (Phoxinus eos). Copeia, pp. 540-542. Axelrod, R., and Hamilton, W. D. (1981). The evolution of cooperation. Science 211, 13901396. Azevedo-Ramos, C., Van Sluys, M., Hero, J.-M., and Magnusson, W. E. (1992). Influence of tadpole movement on predation by Odonate naiads. J. Herpetol. 26, 335-338. Bachman, G. C. (1993). The effect of body condition of the trade-off between vigilance and foraging in Belding’s ground squirrels. Anim. Behav. 46, 233-244. Baker, R. L., and Ball, S. L. (1995). Microhabitat selection by larval Chironomus tentans (Diptera: Chironomidae): Effects of predators, food, cover, and light. Freshwater Biol. 34, 101-106. Ball, S. L., and Baker, R. L. (1995). The non-lethal effects of predators and the influence of food availability on life history of adult Chironomus tentans (Diptera: Chironomidae). Freshwater Biol. 34, 1-12.

PREDATOR-INDUCED STRESS AND BEHAVIOR

267

Ball. S. L.. and Baker, R. L. (1996). Predator-induced life history changes: Antipredator behavior costs or facultative life history shifts? Ecology 77, 1116-1124. Bednekoff, P. A. (1997). Mutualism among safe, selfish sentinels: A dynamic game. Am. Nut., (in press). Bednekoff, P. A., and Houston, A. 1. (1994). Avian daily foraging patterns: Effects of digestive constraints and variability. Evol. Ecol. 8, 36-52. Bercovitch, F. B., Hauser. M. D.. and Jones, J. H. (1995). The endocrine stress response and alarm vocalizations in rhesus macaques. Anim. Behav. 49, 1703-1706. Berger. J. (1991). Pregnancy incentives, predation constraints and habitat shifts: Experimental and field evidence for wild bighorn sheep. Anim. Behuv. 41, 61-77. Berglund, A. (1993). Risky sex: Male pipefishes mate at random in the presence of a predator. Anim. Behuv. 46, 169-175. Blake, M. A,, and Hart, P. J. B. (1993). The behavioural responses of juvenile signal crayfish Pacifasfacrrs lenircsciilus to stimuli from perch and eels. Freshwater Biol. 29, 89-97. Blake, M.. Nystrom, P., and Hart, P. J. B. (1994). The effect of weed cover on juvenile signal crayfish (Pucifustucus leniusculus Dana) exposed to adult crayfish and non-predatory fish. Ann. 2001.Fenn. 31,297-306. Blanchard. R. J., and Blanchard. D. C. (1989). Antipredator defensive behaviors in a visible burrow system. J. Comp. Psychol. 103, 70-82. Bland, J. D., and Temple, S. A. (1990). Effects of predation-risk on habitat use by Himalayan snowcocks. Oecologiu 82, 187-191. Blois-Heulin, C., Crowley, P. H., Arrington, M.. and Johnson, D. M. (1990). Direct and indirect effects of predators on the dominant invertebrates of two freshwater littoral communities. Oecologiu 84, 295-306. Blumstein, D., and Bouskila, A. (1996). Assessment and decision-making in animals: A mechanistic model underlying behavioral flexibility can prevent ambiguity. Oikos 77,569-576. Boesch, C. (1991). The effects of leopard predation on grouping patterns in forest chimpanzees. Behuviour 117, 220-241. Boissy. A. (1995). Fear and fearfulness in animals. Q. Rev. Biol. 70, 165-191. Bollens, S. M., and Frost, B. W. (1989a). Zooplanktivorous fish and variable die1 vertical migration in the marine planktonic copepod Culanus pacificus. Limnol. Oceunogr. 34, 1072-1083. Bollens, S. M., and Frost, B. W. (1989b). Predator-induced die1 vertical migration in a planktonic copepod. J . Plankton Res. 11, 1047-1065. Bollens, S. M., and Frost, B. W. (1991). Ovigerity, selective predation, and variable die1 vertical migration in Euchaefa elongura (Copepoda: Calanoida). Oecologiu 87, 155-161. Bonenfant, M., and Kramer, D. L. (1996). The influence of distance to burrow on flight initiation distance in the woodchuck, Marmotu monux. Behuv. Ecol. 7, 299-303. Bouskila, A. (1995). Interactions between predation risk and competition: A field study of kangaroo rats and snakes. Ecology 76, 165-178. Bouskila. A,, and Blumstein, D. T. (1992). Rules of thumb for predation hazard assessment: Predictions from a dynamic model. Am. Nut. 139, 161-176. Bowers, M. A. (1990). Exploitation of seed aggregates by merriam’s kangaroo rat: Harvesting rates and predatory risk. Ecology 71, 2334-2344. Bowers. M. A., and Dooley, J. L., Jr. (1993). Predation hazard and seed removal by small mammals: Microhabitat versus patch scale effects. Oecologiu 94, 247-254. Bowers, M. A,. Jefferson, J . L.. and Kuebler, M. G. (1993). Variation in giving-up densities of foraging chipmunks (Tumias sfriatus) and squirrels (Sciurus carolinensis). Oikos 66, 229-236.

268

STEVEN L. LIMA

Bozinovic, F., and Simonetti, J . A. (1992). Thermoregulatory constraints on the microhabitat use by cricetid rodents in central Chile. Mammalia 56, 363-369. Brabrand, A,, and Faafeng, B. (1993). Habitat shift in roach (Rutilus rutihts) induced by pikeperch (Stizostedion htcioperca) introduction: Predation risk versus pelagic behaviour. Oecologia 95, 38-46. Braiia, F. (1993). Shifts in body temperature and escape behaviour of female Podarcis muralis during pregnancy. Oikos 66, 216-222. Briggs, S. E., Godin, J.-G., J., and Dugatkin, L. A. (1996). Mate-choice copying under predation risk in the Trinidadian guppy (Poecilia reiiculaia). Behav. Ecol. 7, 151-157. Brown. G. E., Chivers, D. P., and Smith, R. J. F. (1995). Localized defection by pike: A response to labelling by cyprinid alarm pheromone? Behav. Ecol. Sociohiol. 36, 105-1 10. Brown, J. S. (1989). Desert rodent community structure: A test of four mechanisms of coexistence. Ecol. Monogr. 59, 1-20. Brown, J. S. (1992). Patch use under predation risk I. Models and predictions. Ann. 2001. Fenn. 29, 301-309. Brown, J. S., and Alkon. P. U. (1990). Testing values of crested porcupine habitats by experimental food patches. Oecologia 83, 512-518. Brown, J. S., and Morgan, R. A. (1995). Effects of foraging behavior and spatial scale on diet selectivity: A test with fox squirrels. Oikos 74, 122-136. Brown, J. S., and Vincent. T. L. (1992). Organization of predator-prey communities as an evolutionary game. Evoluiion (Lawrence, Kans.) 46, 1269-1283. Brown, J. S., Morgan, R. A,, Dow, B. D. (1992a). Patch use under predation risk: 11. A test with fox squirrels, Sciurus niger. Ann. Zoo/. Fenn. 29, 311-318. Brown, J. S., Arel, Y.,Abramsky, 2..and Kotler, B. P. (1992b). Patch use by gerbils (Gerhillus allenhyi) in sandy and rocky habitats. J. Mammal. 73, 821-829. Brown. J. S., Kotler, B. P., and Mitchell, W. A. (1994). Foraging theory. patch use, and the structure of a Negev Desert granivore community. Ecology 75, 2286-2300. Brown, L. R.. and Moyle, P. B. (1991). Changes in habitat and microhabitat partitioning within an assemblage of stream fishes in response to predation by Sacramento squawfish (Ptychocheihcs grandis). Can. J. Fish. Aquat. Sci. 48, 849-856. Bull, C. D., Metcalfe, N. B., and Mangel, M. (1996). Seasonal matching of foraging to anticipated energy requirements in anorexic juvenile salmon. Proc. R. Soc. London, Ser. B 236, 13-18. Bulova, S. J. (1 994). Ecological correlates of population and individual variation in antipredator behavior of two species of desert lizards. Copeia, pp. 980-992. Burrows. M. T., and Gibson, R. N. (1995). The effects of food, predation risk and endogenous rhythmicity on the behaviour of juvenile plaice, Pleuronectes plaiessa L. Anim. Behav. 50,41-52. Burrows, M. T.. and Hughes. R. N. (1991). Optimal foraging decisions by dogwhelks, Nucella lapillits (L.): Influences of mortality risk and rate-constrained digestion. Funci. Ecol. 5,461-475. Butler, J. M.. and Roper, T. J. (1994). Escape tactics and alarm responses in badgers Meles melest A field experiment. Ethology 99, 313-322. Caro. T. M. (1994). Ungulate antipredator behaviour: Preliminary and comparative data from African bovids. Behaviour 128, 189-228. Caro, T. M. (1995). Pursuit-deterrence revisited. Trends Ecol. Evol. 10, 500-503. Caro. T. M.. Lombardo, L., Goldizen, A. W., and Kelly, M. (1995). Tail-flagging and other antipredator signals in white-tailed deer: New data and synthesis. Behav. Ecol. 6,442-450. Cartar, R. V. (1991). Colony energy requirements affect response to predation risk in foraging bumble bees. Eihology 87, 90-96.

PREDATOR-INDUCED STRESS AND BEHAVIOR

269

Cassini. M. H. (1991). Foraging under predation risk in the wild guinea pig Cavia aperea. Oikos 62, 20-24. Cassini, M. H.. and Galante. M . L. (1992). Foraging under predation risk in the wild guinea pig: The effect of vegetation height on habitat utilization. Ann. Zool. Fenn. 29,285-290. Chesson, P., and Rosenweig. M. (1991). Behavior, heterogeneity, and the dynamics of interacting species. Ecology 72, 1187-1195. Chivers, D. P., and Smith, R. J. F. (1994). Fathead minnows, Pimephales promelas, acquire predator recognition when alarm substance is associated with the sight of unfamiliar fish. Anim. Behav. 48, 597-605. Chivers, D. P., and Smith, R. J. F. (1995). Free-living fathead minnows rapidly learn to recognize pike as predators. J . Fish Biol. 46, 949-954. Chivers, D. P., Wisendon, B. D., and Smith. R. J. F. (1995a). The role of experience in the response of fathead minnows (Pimephales promelas) to skin extract of Iowa darters (E/heos/oma exile). Behavioirr 132, 665-674. Chivers. D. P., Brown. G. E.. and Smith, R. J. (1995b). Familiarity and shoal cohesion in fathead minnows (Pimephales promelas): Implications for antipredator behaviour. Can. J. Zoo/. 73, 955-960. Chivers, D. P.. Wisenden, B. D.. and Smith. R. J. F. (1995~).Predation risk influences reproductive behaviour of Iowa darters, E/heos/oma exile (Osteichthyes, Percidae). E/hology 99,278-285. Chovanec, A. (1992). The influence of tadpole swimming behaviour on predation by dragonfly nymphs. Amphibin-Rrplilia 13, 341-349. Christensen. B., and Persson. L. (1993). Species-specific antipredatory behaviours: Effects on prey choice in different habitats. Behav. Ecol. Sociobiol. 32, 1-9. Clark, C. W. (1994). Antipredator behavior and the asset-protection principle. Behav. Ecol. 5, 159-170. Clark, C. W.. and Dukas. R. (1994). Balancing foraging and antipredator demands: An advantage of sociality. Am. Nat. 144, 542-548. Clark, C. W., and Levy, D. A. (1988). Die1 vertical migrations by juvenile sockeye salmon and the antipredation window. A m . Nut. 131, 271-290. Clark, C. W., and Ydenberg, R. C. (1990a). The risks of parenthood. I. General theory and applications. Evol. Ecol. 4, 21-34. Clark, C. W., and Ydenberg. R. C. (1990b). The risks of parenthood 11. Parent-offspring conflict. Evol. Ecol. 4, 312-325. Cowan, C. A., and Peckarsky, B. L. (1994). Die1 feeding and positioning periodicity of a grazing mayfly in a trout stream and a fishless stream. Can. J. Fish. Aqua/. Sci. 51,450-459. Cowlishaw, G. (1994). Vulnerability to predation in baboon populations. Behaviour 131, 293-304. Crawley. M. J. (1992). Population dynamics of natural enemies and their prey. In “Natural Enemies: The Population Biology of Predators. Parasites. and Diseases” (M. J. Crawley. ed.), pp. 40-89. Blackwell. Oxford. Cresswell. W. (1994a). Flocking is an effective anti-predation strategy in redshanks, Tringa /o/aiii~s. Anim. Behav. 47, 433-442. feeding location: Cresswell. W. ( 1994b). Age-dependent choice of redshank (Tringa /o/ani~s) Profitability or risk’?J . Anim. Ecol. 63, 589-600. Crowl. T. A., and Covich. A. P. (1994). Response of a freshwater shrimp to chemical and tactile stimuli from a large decapod predator. J. North Am. Benthol. Soc. 13,291-298. Crowley, P. H., and Hopper, K. R. (1994). How to behave around cannibals: A densitydependant dynamic game. A m . Na/. 143, 117-154.

270

STEVEN L. LIMA

Crowley. P. H.. Travers, S. E.. Linton, M. C., Cohn, S. L., Sih, A., and Sargent. R. C. (1991). Mate density, predation risk, and the seasonal sequence of mate choices: A dynamic game. Am. Nut. 137,567-596. Croy. M. I., and Hughes, R. N. (1991). Effects of food supply, hunger, danger and competition on choice of foraging location by the fifteen-spined stickleback. Spinachia spinachia L. Anim. Behav. 42, 131-139. Culp, J. M., and Scrimgeour, G. J. (1993). Size-dependent die1 foraging periodicity of a mayfly grazer in streams with and without fish. Oikos 68,242-250. Culp, J. M., Glozier, N. E., and Scrimgeour, G. J. (1991). Reduction of predation risk under the cover of darkness: Avoidance responses of mayfly larvae to benthic fish. Oecologia 86, 163-169. Curio, E. (1993). Proximate and developmental aspects of antipredator behavior. Adv. Study Behav. 22, 135-238. Daly. M., Wilson, M., Behrends. P. R., and Jacobs, L. F. (1990). Characteristics of kangaroo rats, Dipodomys merriami, associated with differential predation risk. Anim. Behav. 40, 380-389. Daly, M., Behrends, P. R., Wilson, M. I., and Jacobs, L. F. (1992). Behavioural modulations of predation risk: Moonlight avoidance and crepuscular compensation in a nocturnal desert rodent. Anim. Behav. 44, 1-9. Dawidowicz, P., and Loose, C. J. (1992). Metabolic costs during predator-induced die1 vertical migration of Daphnia. Limnol. Oceanogr. 37, 1589-1595. Dawidowicz, P., Pijanowska, J., and Ciechomski, K. (1990). Vertical migration of Chaohorus larvae is induced by the presence of fish. Limnol. Oceanogr. 35, 1631-1637. De Meester, L. (1993). Genotype, fish-mediated chemicals, and phototactic behavior in Daphnia magna. Ecology 74, 1467-1474. De Meester, L., Weider, L. J., and Tollrian, R. (1995). Alternative antipredator defences and genetic polymorphism in a pelagic predator-prey system. Nature (London) 378,483-485. Desportes, J.-P., Metcalfe, N. B., Cezilly, F.. Lauvergeon, G., and Kervella. C. (1989). Tests of the sequential randomness of vigilance behavior using spectral analysis. Anim. Behav. 38,771-777. Desy, E. A,. Batzli, G. 0..and Liu, J. (1990). Effects of food and predation on behaviour of prairie voles: A field experiment. Oikos 58, 159-168. Devenport, J. A. (1989). Social influences on foraging in black-tailed prairie dogs. J. Mammal. 70, 166-168. DeVries, D. R. (1990). Habitat use by bluegill in laboratory pools: Where is the refuge when macrophytes are sparse and alternative prey are present? Environ. B i d . Fishes 29,27-34. Dickman. C. R. (1992). Predation and habitat shift in the house mouse, Miis domesticus. Ecology 73, 313-322. Diehl, S., and Eklov, P. (1995). Effects of piscivore-mediated habitat use on resources, diet, and growth of perch. Ecology 76, 1712-1726. Dill, L. M. (1987). Animal decision making and its ecological consequences: The future of aquatic ecology and behaviour. Can. J. Zool. 65, 803-811. Dill, L. M. (1990). Distance-to-cover and the escape decisions of an African cichlid fish, Melanochromis chipokae. Environ. Biol. Fishes 27, 147-152. Dill. L. M., and Fraser. A. H. G. (1984). Risk of predation and the feeding behavior of juvenile coho salmon (Oncorhynchus kisutch). Behav. Ecol. Sociohiol. 16,65-71. Dill. L. M.. and Gillett, J. F. (1991). The economic logic of barnacle Balanus glandula (Darwin) hiding behavior. J. Exp. Mar. B i d . Ecol. 153, 115-127. Dill, L. M.. and Houtman. R. (1989). The influence of distance to refuge on flight-initiation distance in the gray squirrel (Sciurus carolinensis). Can. J. Zool. 67, 232-235.

PREDATOR-INDUCED STRESS AND BEHAVIOR

27 1

Dill, L. M.. Fraser, A. H. G., and Roitberg, B. D. (1990). The economics of escape behaviour in the pea aphid, Acyrthosiphon pisum. Oecologia 83, 473-478. Dini, M. L., and Carpenter, S. R. (1988). Variability in Daphnia behavior following fish community manipulations. J. Plankton Res. 10, 621-635. Dini, M. L., and Carpenter, S . R. (1992). Fish predators, food availability and diel vertical migration in Daphnia. J. Plankton Res. 14, 359-377. Dini, M. L., Soranno, P. A,, Scheuerell, M.. and Carpenter, S . R. (1993). Effects of predators and food supply on diel vertical migration of Daphnia. In “The Trophic Cascade in Lakes” ( S . R. Carpenter and J. F. Kitchell, eds.), pp. 153-171. Cambridge University Press, Cambridge, UK. Dixon, S . M., and Baker. R. L. (1988). Effect of size on predation risk, behavioral response to fish, and cost of reduced feeding in larval Ischnura verticalis (Coenagrionidae: Odonata). Oecologia (Berlin) 76, 200-205. Dodson, S. (1988). The ecological role of chemical stimuli for the zooplankton: Predator avoidance behavior in Daphnia. Limnol. Oceanogr. 33, 1431-1439. Dodson, S . (1990). Predicting diel vertical migration of zooplankton. Limnol. Oceanogr. 35,1195-1200. Dodson, S. I., Crowl, T. A,, Peckarsky, B. L., Kats, L. B., Covich, A. P., and Culp, J. M. (1994). Non-visual communication in freshwater benthos: On overview. J. North Am. Benthol. SOC. 13, 268-282. Doncaster, C. P. (1993). Influence of predation threat on foraging pattern: The hedgehog’s gambit. Rev. Ecol. 48, 207-213. Doncaster. C. P. (1994). Factors regulating local variations in abundance: Field tests on hedgehogs, Erinaceus europaeus. Oikos 69, 182-192. Douglas, P. L., Forrester. G. E., and Cooper, S. D. (1994). Effects of trout on the diel periodicity of drifting in baetid mayflies. Oecologia 98, 48-56. Dugatkin, L. A. (1988). Do guppies play tit for tat during predator inspection visits? Behav. Ecol. Sociobiol. 25, 395-399. Dugatkin, L. A. (1992). Tendency to inspect predators predicts mortality risk in the guppy (Poecilia reticulata). Behav. Ecol. 3, 124-127. Dugatkin, L. A,, and Alfieri, M. (1991a). Guppies and the tit-for-tat strategy: Preference based on past interaction. Behav. Ecol. Sociobiol. 28, 243-246. Dugatkin, L. A,, and Alfieri, M. (1991b). Tit-for-tat in guppies (Poecilia reticulata): The relative nature of cooperation and defection during predator inspection. Evol. Ecol. 5,300-309. Dugatkin, L. A,, and Godin, J.-G. J. (1992a). Prey approaching predators: A cost-benefit perspective. Ann. Zool. Fenn. 29, 233-252. Dugatkin, L. A., and Godin, J . 4 . J. (1992b). Predator inspection, shoaling, and foraging under predation hazard in the Trinidadian guppy, Poecilia reticulata. Environ. Biol. Fishes 34,265-276. Dukas, R., and Clark, C. W. (1995). Sustained vigilance and animal performance. Anim. Behav. 49, 1259-1267. Duvall, C. J., and Williams, D. D. (1995). Individuality in the growth of stonefly nymphs in response to stress from a predator. Arch. Hydrobiol. 133, 273-286. Edmunds, M. (1974). “Defence in Animals.” Longman, New York. Eklov, P., and Diehl, S. (1994). Piscivore efficiency and refuging prey: The importance of predator search mode. Oecologia 98, 344-353. Eklov, P.. and Hamrin, S. F. (1989). Predatory efficiency and prey selection: Interactions between pike Esox lucius, perch Perca puviatilis and rudd Scardinus erythrophthalmus. Oikos 56, 149-156.

272

STEVEN L. LIMA

Eklov, P.. and Persson, L. (1995). Species-specific antipredator capacities and prey refuges: Interactions between piscivorous perch (Percu ,puviatilis) and juvenile perch and roach (Rutilics rittilus). Behav. Ecol. Sociobiol. 37, 169- 178. EklBv. P., and Persson. L. (1996). The response of prey to the risk of predation: Proximate cues for refuging juvenile fish. Anim. Behav. 51, 105-115. Elgar, M. A. (1989). Predator vigilance and group size in mammals and birds: A critical review of the empirical evidence. B i d . Rev. Cambridge Philos. Soc. 64, 13-33. Elliott, S. A,, Kats, L. B., and Breeding. J. A. (1993). The use of conspecific chemical cues for cannibal avoidance in California newts (Taricha torosa). Ethology 95, 186-192. Endler. J. A. (1987). Predation. light intensity, and courtship behaviour in Poecilio retictilafa (Pisces: Poeciliidae). Anim. Behav. 35, 1376- 1385. Endler. J. A. (1991). Interactions between predators and prey. In ”Behavioral Ecology: An Evolutionary Approach” ( J . R. Krebs and N. B. Davies, eds.). 3rd ed., pp. 169-201. Blackwell, Oxford. Everett. R. A.. and Ruiz. G. M. (1993). Coarse woody debris as a refuge from predation in aquatic communities: An experimental test. Oecologia 93, 475-486. Fairbairn, D.J. (1993). Costs of loading associated with mate-carrying in the waterstrider, Aquarius reniigis. Behav. Ecol. 4, 224-231. Feltmate, B. W.. and Williams. D. D. (1989a). Influence of rainbow trout (Oncorhynchiis mykiss) on density and feeding behaviour of a perlid stonefly. Can. J. Fish. Aquat. Sci. 46, 1575-1580. Feltmate, B. W., and Williams, D. D. (1989b). A test of crypsis and predator avoidance in the stonefly Paragnetitin media (Plecoptera: Perlidae). Anirn. Behav. 37, 992-999. Feltmate, B. W.. and Williams. D. D. (1991). Evaluation of predator-induced stress on field population of stoneflies (Plecoptera). Ecology 72, 1800-1806. Feltmate, B. W., Williams. D. D.. and Montgomerie. A. (1992). Relationship between diurnal activity patterns, cryptic coloration. and subsequent avoidance of predaceous fish by perlid stoneflies. Can. J. Fish. Aquat. Sci. 49, 2630-2634. Feminella, J. W., and Hawkins, C. P. (1994). Tailed frog tadpoles differentially alter their feeding behavior in response to non-visual cues from four predators. J. North Am. Benthol. Soc. 13, 310-320. Fenn, M. G. P., and Macdonald, D. W. (1995). Use of middens by red foxes: Risk reverses rhythms of rats. J. Mammal. 76, 130-136. Figiel. C.R.. Jr.. and Semlitsch, R. D. (1990). Population variation in survial and metamorphosis of larval salamanders (Ambysroma nzaculaturn) in the presence and absence of fish predation. Copeia, pp. 818-826. Fiksen, 0.. and Giske, J. (1995). Vertical distribution and population dynamics of copepods by dynamic optimization. ICES J. Mur. Sci. 52, 483-503. FitzCibbon, C. D. (1989). A cost to individuals with reduced vigilance in groups of Thomson’s gazelles hunted by cheetahs. Anim. Behav. 37, 508-510. FitzGibbon. C.D. (1990). Anti-predator strategies of immature Thomson’s gazelles: Hiding and the prone response. Anim. Behuv. 40, 846-855. FitzGibbon, C. D. (1994).The costs and benefits of predator inspection behaviour in Thomson’s gazelles. Behrtv. Ecol. Sociobiol. 34, 139-148. Flecker, A. S. (1992). Fish predation and the evolution of invertebrate drift periodicity: Evidence from neotropical streams. Ecology 73, 438-448. Forrnanowicz, D.R., Jr.. and Bobka, M. S. (1989). Predation risk and microhabitat preference: An experimental study of the behavioral responses of prey and predator. Am. Mid/. Nut. 121,379-386.

PREDATOR-INDUCED STRESS AND BEHAVIOR

273

Forrester, G. E. (1994a). Influences of predatory fish on the drift dispersal and local density of stream insects. Ecology 75, 1208-1218. Forrester, G. E. (1994b). Die1 patterns of drift by five species of mayfly at different levels of fish predation. Can. J . Fish. Aqrmf. Sci. 51, 2549-2557. Forsgren. E. (1992). Predation risk affects mate choice in a gobiid fish. A m . Nat. 140,1041-1049. Forsgren. E.. and Magnhagen, C. (1993). Conflicting demands in sand gobies: Predators influence reproductive behaviour. Behaviour 126, 125-135. Fraser, D. F., and Gilliam, J. F. (1992). Nonlethal impacts of predator invasion: Facultative suppression of growth and reproduction. Ecology 73, 959-970. Fraser. D. F.. Gilliam. J. F., and Yip-Hoi. T. (1995). Predation as an agent of population fragmentation in a tropical watershed. Ecology 76, 1461-1472. Frost, B. W., and Bollens. S. M. (1992). Variability of die1 vertical migration in the marine plankton copepod Pseitdocalonus newmani in relation to its predators. Can. J. Fish. Aquat. Sci. 49, 1137-1 141. Fuller, R., and Berglund. A. (1996). Behavioral responses of a sex-role reversed pipefish to a gradient of perceived predation risk. Behav. Ecol. 7, 69-75. Gabriel, W., and Thomas, B. (1988). Vertical migration of zooplankton as an evolutionarily stable strategy. A m . Nut. 132, 199-216. Gelowitz. C. M., Mathis, A,. and Smith, R. J . F. (1993). Chemosensory recognition of northern pike (Esox lucius) by brook stickleback (Culaea inconstans): Population differences and the influence of predator diet. Behaviour 127, 105-1 18. Gilliam, J. F. (1982). Habitat use and competitive bottlenecks in size-structured fish populations. Ph.D. Dissertation, Michigan State University, East Lansing. Gilliam. J. F. (1990). Hunting by the hunted: optimal prey selection by foragers under predation hazard. N A T O AS1 Ser., Ser. G 20, 797-818. Gilliam, J. F.. and Fraser, D. F. (1987). Habitat selection when foraging under predation hazard: A model and a test with stream-dwelling minnows. Ecology 68, 1856-1862. Gilliam, J. F.. and Fraser, D. F. (1988). Resource depletion and habitat segregation by competitors under predation hazard. In “Size-Structured Populations” (B. Ebenman and L. Persson. eds.), pp. 173-184. Springer-Verlag. Berlin. Gliwicz, Z. M. (1986). A lunar cycle in zooplankton. Ecology 67, 883-897. Gliwicz, 2. M. (1994). Retarded growth of cladoceran zooplankton in the presence of a copepod predator. Oecologia 97,458-461. Gliwicz. Z . M., and Jachner. A. (1992). Die1 migrations of juvenile fish: A ghost of predation past or present? Arch. Hydrobiol. 124, 385-410. Gliwicz, Z . M.. and Pijanowska, J. (1988). Effect of predation and resource depth distribution on vertical migration of zooplankton. Bull. Mar. Sci. 43, 695-709. Godin. J.-G. J. (1990). Diet selection under the risk of predation. N A T O AS1 Ser., Ser. C 20,739-769. Godin. J.-G. J. (1995). Predation risk and alternative mating tactics in male Trinidadian guppies (Poecilia reticulafa).Oecologia 103, 224-229. Godin, J.-G. J., and Briggs, S. E. (1996). Female mate choice under predation risk in the guppy. Anim. Behav. 51, 117-130. Godin. J.-G. J., and Crossman. S. L. (1994). Hunger-dependent predator inspection and foraging behaviours in the threespine stickleback ( Gasterosfeus aculeatus) under predation risk. Behav. Ecol. Sociobiol. 34, 359-366. Godin, J . 4 . J., and Davis, S. A. (1995). Who dares, benefits: Predator approach behaviour in the guppy (Poecilia reticulata) deters predator pursuit. Proc. R. Soc. London, Ser. B 259, 193-200.

274

STEVEN L. LIMA

Godin, J.-G. J., and Smith, S. A. (1988). A fitness cost of foraging in the guppy. Nature (London) 333,69-71. Gotceitas, V. (1990). Foraging and predator avoidance: A test of a patch choice model with juvenile bluegill sunfish. Oecologia 83, 346-351. Gotceitas, V., and Brown, J. A. (1993). Substrate selection by juvenile Atlantic cod (Gadus morhua): Effects of predation risk. Oecologia 93, 31-37. Gotceitas. V., and Colgan, P. (1990a). Behavioural response of juvenile bluegill sunfish to variation in predation risk and food level. Ethology 85, 247-255. Gotceitas, V., and Colgan, P. (1990b). The effects of prey availability and predation risk on habitat selection by juvenile bluegill sunfish. Copeia, pp. 409-417. Gotceitas, V., and Godin. J.-G. J. (1991). Foraging under the risk of predation in juvenile Atlantic salmon (Salmo salar L.): Effects of social status and hunger. Behav. Ecol. Sociob i d . 29, 255-261. Gotceitas, V. and Godin, J . 4 . J. (1993). Effects of aerial and in-stream threat of predation on foraging by juvenile Atlantic salmon (Salmo sahr). Can. Spec. Publ. Fish. Aquar. Sci. 118, 35-41. Gregory, R. S. (1993). Effects of turbidity on the predator avoidance behaviour of juvenile chinook salmon (Oncorhynchus rshawytschu). Can. J. Fish. Aquat. Sci. 50, 241-246. Grill, C. P., and Juliano, S . A. (1996). Predicting species interactions based on behaviour: Predation and competition in container-dwelling mosquitoes. J. Anim. Ecol. 65, 63-76. Grossman, G. D., Ratajczak, R. R., Jr.. and Crawford. M. K. (1995). Do rock bass (Ambloplites rupestris) induce microhabitat shifts in mottled sculpin (Cottus bairdi)? Copeia, pp. 343-353. Guisande. C., Duncan, A,, and Lampert, W. (1991). Trade-offs in Daphnia vertical migration strategies. Oecologia 87, 357-359. Hager, M. C., and Helfman, G. S. (1991). Safety in numbers: Shoal size choice by minnows under predatory threat. Behav. Ecol. Sociobiol. 29,271-276. Hansson, L. (1995). Is the indirect predator effect a special case of generalized reactions to density-related disturbances in cyclic rodent populations? Ann. Zool. Fenn. 32, 159-162. Harfenist, A,, and Ydenberg, R. C. (1995). Parental provisioning and predation risk in rhinoceros auklets (Cerorhinca monocerata); Effects on nestling growth and fledging. Behav. E d 6, 82-86. Harvey, B. C. (1991). Interactions among stream fishes: Predator-induced habitat shifts and larval survival. Oecologia 87, 29-36. Hasson. 0. (1991). Pursuit-deterrant signals: Communication between prey and predator. Trends Ecol. Evol. 6, 325-329. Hastings, P. A. (1991). Flexible responses to predators in a marine fish. Ethol. Ecol. Evol. 3, 177-184. Hayes. F. E. (1989). Antipredator behavior of recently metamorphosed toads (Bufo a. americanus) during encounters with garter snakes (Thamnophis s. sirtulis). Copeia, pp. 10111015. Hazlett, B. A. (1994). Alarm responses in the crayfish Orconectes virilis and Orconectes propinquus. J. Chem. Ecol. 20,1525-1535. He, X., and Kitchell. J. F. (1990). Direct and indirect effects of predation on a fish community: A whole-lake experiment. Trans. Am. Fish. Soc. 119, 825-835. He, X., and Wright, R. A. (1992). An experimental study of piscivore-planktivore interactions: Population and community responses to predation. Can. J . Fish. Aquat. Sci. 49,1176-1 183. Hedrick, A. V., and Dill, L. M. (1993). Mate choice by female crickets is influenced by predation risk. Anim. Behav. 46, 193-196.

PREDATOR-INDUCED STRESS A N D BEHAVIOR

275

Heikkila, J., Kaarsalo, K., Mustonen, O., and Pekkarinen, P. (1993). Influence of predation risk on early development and maturation in three species of Clethrionomys voles. Ann. Zool. Fenn. 30, 153-161. Heinen, J. T. (1993). Substrate choice and predation risk in newly metamorphosed American toads Bufo americanus: An experimental analysis. A m . Midl. Nat. 130, 184-192. Heinen, J. T. (1994a). Antipredator behavior of newly metamorphosed American toads (Bufo a. americanus). and mechanisms of hunting by eastern garter snakes (Thamnophis s. sirtalis). Herpetologica 50, 137-14.5. Heinen, J. T. (1994b). The significance of color change in newly metamorphosed American toads (Bufo a. americanus). J. Herpetol. 28, 87-93. Helfman. G. S. (1986). Behavioral responses of prey fishes during predator-prey interactions. In “Predator-Prey Relationships: Perspectives and Approaches from the Study of Lower Vertebrates” (M. E. Feder and G. V. Lauder, eds.), pp. 135-156. University of Chicago Press, Chicago. Helmstetter, F. J., and Fanselow. M. S. (1993). Aversively motivated changes in meal patterns of rats in a closed economy: The effects of shock density. Anim. Learn. Behav. 21,168-175. Henrikson, B.-I. (1988). The absence of antipredator behaviour in the larvae of Leucorrhinia dubia (Odonata) and the consequences for their distribution. Oikos 51, 179-183. Hershey, A. E. (1987). Tubes and foraging behavior in larval Chironomidae: Implications for predator avoidance. Oecologia 73, 236-241. 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. Anim. Behav. 36, 125-133. Hik, D. S. (1995). Does risk of predation influence population dymamics? Evidence from the cyclic decline of snowshoe hares. Wild. Res. 22, 115-129. Hill, A. M., and Lodge, D. M. (1994). Die1 changes in resource demand: Competition and predation in species replacement among crayfishes. Ecology 75, 21 18-2126. Hill, A. M., and Lodge, D. M. (1995). Multi-trophic-level impact of sublethal interactions between bass and omnivorous crayfish. J. North A m . Benthol. SOC. 14,306-314. Hinsley, S. A,, Bellamy, P. E., and Moss, D. (1995). Sparrowhawk Acciprer nisus predation and feeding site selection by tits. Ibis 137, 418-428. Holmes, W. G. (1991). Predator risk affects foraging behaviour of pikas: Observational and experimental evidence. Anim. Behav. 42, 111-119. Holomuzki, J. R.. and Hoyle, J. D. (1990). Effect of predatory fish presence on habitat use and die1 movement of the stream amphipod, Cammarus minus. Freshwater Biol. 24,509-517. Holomuzki, J. R., and Short, T. M. (1990). Ontogenic shifts in habitat use and activity in a stream-dwelling isopod. Holarctic Ecol. 13, 300-307. Horat, P., and Semlitsch, R. D. (1994). Effects of predation risk and hunger on the behaviour of two species of tadpoles. Behav. Ecol. Sociobiol. 34, 393-401. Houston, A. I., and McNamara. J. M. (1989). The value of food: Effects of open and closed economies. Anim. Behav. 37,546-562. Houston, A. 1.. and McNamara, J. M. (1993). A theoretical investigation of the fat reserves and mortality levels of small birds in winter. Ornis Scand. 24,205-219. Houston, A. I., McNamara, J. M., and Hutchinson, J. M. C. (1993). General results concerning the trade-off between gaining energy and avoiding predation. Philos. Trans. R. SOC. London, Ser. B 341, 375-397. Houtman, R., and Dill, L. M. (1994). The influence of substrate color on the alarm response of tidepool sculpins (Oligocotrus maculosiis; Pisces, Cottidae). Erhology %, 147-154.

276

STEVEN L. LIMA

Huang. C.. and Sih, A. (1990). Experimental studies on behaviorally meditated, indirect interactions through a shared predator. Ecology 71, 15 15-1522. Huang. C.. and Sih, A. (1991). Experimental studies on direct and indirect interactions in a three trophic-level stream system. Oecologia 85, 530-536. Hughes, J. J., and Ward, D. (1993). Predation risk and distance to cover affect foraging behaviour in Namib desert gerbils. Anim. Behav. 46, 1243-1245. Hughes, J . J.. Ward. D., and Perrin, M. R. (1994). Predation risk and competition affect habitat selection and activity of Namib Desert gerbils. Ecology 75, 1397-1405. Hugie. D. M., and Dill, L. M. (1Y94). Fish and game: A game theoretic approach to habitat selection by predators and prey. J. Fish B i d . 45, 151-169. Huntingford. F. A.. Lazarus, J.. Barrie, B. D.. and Webb. S. (1994). A dynamic analysis of cooperative predator inspection in sticklebacks. Anim. Behav. 47, 413-423. Ibrahim. A. A., and Huntingford. F. A. (1989). Laboratory and field studies of the effect of predation risk on foraging in three-spined sticklebacks (Gasterosfeusaculeatus).Behavioitr 109,46-57. lves, A. R., and Dobson, A. P. (1987). Antipredator behavior and the population dynamics of simple predator-prey systems. Am. Nat. 130, 431-447. Iwasa, Y., Suzuki. Y.. and Matsuda, H. (1984) Theory of oviposition strategy of parasitoids. I. Effect of mortality and limited egg number. Theor. Popicl. Biol. 26, 205-227. Jachner, A. (1995a). Chemically-induced habitat shifts in bleak (Alburnus alburnus L.) Arch. Hydrobiol. 133, 7 1-79. Jachner, A. (199%). Changes in feeding behavior of bleak (Alburnus alburnus L.) in response to visual and chemical stimuli from predators. Arch. Hydrobiol. 133, 305-314. Jakobsen, P. J., Birkeland, K.. and Johnsen. G. H. (1994). Swarm location in zooplankton as an anti-predator defence mechanism. Anim. Behav. 47, 175-178. Jedrzejewska, B., and Jedrzejewski. W. (1990). Antipredatory behaviour of bank voles and prey choice of weasels: Enclosure experiments. Ann. Zool. Fenn. 27, 321-328. Jedrzejewski. W.. and Jedrzejewska. B. (1990). Effect of a predator’s visit on the spatial distribution of bank voles: Experiments with weasels. Can. J . Zool. 68, 660-666. Jedrzejewski, W., Rychlik. L., and Jedrzejewska, 5. (1993). Responses of bank voles to odours of seven species of predators: Experimental data and their relevance to natural predatorvole relationships. Oikos 68, 251-257. Jeffries, M. (1990). Interspecific differences in movement and hunting success in damselfly larvae (Zygoptera: Insecta): Responses to prey availability and predation threat. Freshwater Biol. 23, 191-196. Jennions, M. D., and Blackwell, P. R. Y. (1992). Chorus size influences on the anti-predator response of a neotropical frog. Anim. Behav. 44,990-992. Johansson. A., and Englund. G. (1995). A predator-prey game between bullheads and casemaking caddis larvae. Anim. Behav. 50, 785-792. Johansson, F. (1993). Effects of prey type, prey density and predator presence on behaviour and predation risk in a larval damselfly. Oikos 68, 481-489. Juliano, S. A,. Hechtel, L. J., and Waters, J. R. (1993). Behavior and risk of predation in larval tree hole mosquitoes: Effects of hunger and population history of predation. Oikos 68,229-241. Kilis, J . A., Fiske. P.. and Szther, S. A. (1995). The effect of mating probability on risk taking: An experimental study in lekking great snipe. Am. Nut. 146, 59-71. Kalcounis, M. C., and Brigham. R. M. (1994). Impact of predation risk on emergence by little brown bats, Myotic lucifzrgiis (Chiroptera: Vespertilionidae), from a maternity colony. Ethology 98, 201 -209.

PREDATOR-INDUCED STRESS A N D BEHAVIOR

277

Kats, L. B., and Dill, L. M. (1998). The scent of death: Chemosensory assessment of predation risk by prey animals. Ecoscience (in press). Kats. L. B.. Petranka, J. W.. and Sih. A. (1988). Antipredator defenses and the persistence of amphibian larvae with fishes. Ecology 69, 1865-1870. Kennedy, M.. Shave, C. R., Spencer, H. G.. and Gray. R. D. (1994). Quantifying the effect of predation risk on foraging bullies: No need to assume an IFD. Ecology 75,2220-2226. Kohler, S. L.. and McPeek, M. A. (1989). Predation risk and the foraging behavior of competing stream insects. Ecology 70, 1181-1825. Koivula. K., Lahti. K., Rytkonen. S., and Orell. M. (1994). Do subordinates expose themselves to predation? Field experiments on feeding site selection by willow tits. J. Avian Biol. 25, 178-183. Koivula, K., Rytkonen, S.. and Orell, M. (1995). Hunger-dependency of hiding behaviour after a predator attack in dominant and subordinate willow tits. Ardea 83, 397-404. Kolar, C. S., and Rahel, F. J. (1993). Interaction of biotic factor (predator presence) and an abiotic factor (low oxygen) as an influence on benthic invertebrate communities. Oecologia 95,210-219. Korpimaki, E.. Norrdahl, K., and Valkama, J. (1994). Reproductive investment under fluctuating predation risk: Microtine rodents and small mustelids. Evol. Ecol. 8, 357-368. Korpimaki, E.. Koivunen, V., and Hakkarainen. H. (1996). Microhabitat use and behavior of voles under weasel and raptor predation risk: Predator facilitation? Behav. Ecol. 7, 30-34. Koskela. E.. and Ylonen, H. (1995). Suppressed breeding in the field vole (Microtus agrestis): An adaptation to cyclically fluctuating predation risk. Behav. Ecol. 6, 31 1-315. Koskela, E., Horne, T. J., Mappes. T., and Ylonen, H. (1996). Does risk of small mustelid predation effect the oestrous cycle in the bank vole, Clethrionomys glareolus? Anim. Behav. 51, 1159-1163. Kotler. B. P. (1992). Behavioral resource depression and decaying perceived risk of predation in two species of coexisting gerbils. Behav. Ecol. Sociobiol. 30,239-244. Kotler, B. P., and Blaustein, L. (1995). Titrating food and safety in a heterogeneous environment: When are risky and safe patches of equal value? Oikos 74, 251-258. Kotler, B. P., and Holt. R. D. (1989). Predation and competition: The interaction of two types of species interactions. Oikos 54, 256-260. Kotler, B. P., Brown, J. S., and Hasson, 0. (1991). Factors affecting gerbil foraging behavior and rates of owl predation. Ecology 72,2249-2260. Kotler, B. P., Blaustein, L.. and Brown, J. S. (1992). Predator facilitation: The combined effect of snakes and owls on the foraging behavior of gerbils. Ann. Zool. Fenn. 29, 199-206. Kotler, B. P.. Blaustein, L., and Dednam, H. (1993a). The specter of predation: The effects of vipers on the foraging behavior of two gerbilline rodents. Isr. J. Zool. 39, 11-21. Kotler, B. P., Brown, J. S., Slotow. R. H., Goodfriend, W. L.. and Strauss. M. (1993b). The influence of snakes on the foraging behavior of gerbils. Oikos 67, 309-316. Kotler, B. P., Brown, J. S.. and Subach, A. (1993~).Mechanisms of species coexistence of optimal foragers: Temporal partitioning by two species of sand dune gerbils. Oikos 67,548-556. Kotler, B. P., Brown, J. S., and Mitchell, W. A. (1993d). Environmental factors affecting patch use in two species of gerbilline rodents. J. Mammal. 74, 614-620. Kotler, B. P.. Ayal, Y., and Subach, A. (1994a). Effects of predatory risk and resource renewal on the timing of foraging activity in a gerbil community. Oecologia 100, 391-396. Kotler, B. P., Gross, J. E., and Mitchell, W. A. (1994b). Applying patch use to assess aspects of foraging behavior in Nubian ibex. J . Wild. Manage. 58,299-307.

278

STEVEN L. LIMA

Kramer, D. L., and Bonenfant, M. (1997). Direction of predator approach and the decision to flee to a refuge. Anim. Behav. (in press). Kratz, K. W. (1996). Effects of stoneflies on local prey populations: Mechanisms of impact across prey density. Ecology 77, 157351585, Krause, J. (1993a). The relationship between foraging and shoal position in a mixed shoal of roach (Rutilus rutilus) and chub (Leuciscus cephaius): A field study. Oecologia93,356-359. Krause, J. (1993b). The effect of ‘Shreckstoff’ on the shoaling behaviour of the minnow: A test of Hamilton’s selfish herd theory. Anim. Behav. 45, 1019-1024. Krause, J. (1994a). The influence of food competition and predation risk on size-assortative shoaling in juvenile chub (Leuciscus cephalus). Ethology 96, 105- 116. Krause, J. (1994b). Differential fitness returns in relation to spatial position in groups. Biol. Rev. Cumbridge Philos. SOC.69, 187-206. Krause, J., and Godin, J.-G. J. (1994). Shoal choice in the banded killifish (Fundulusdiaphanus, Teleostei, Cyprinodontidae): Effects of predation risk, fish size, species composition and size of shoals. Ethology 98, 128-136. Krause, J., and Godin. J.-G. J. (1995). Predator preferences for attacking particular prey group sizes: Consequences for predator hunting success and prey predation risk. Anim. Behuv. 50, 465-473. Krause, J., and Godin, J.-G. J. (1996). Influence of prey foraging posture on flight behavior and predation risk: Predators take advantage of unwary prey. Behav. Ecol. 7, 264-271. Krause, J., Bumann, D., and Todt, D. (1992). Relationship between the position preference and nutritional state of individuals in schools of juvenile roach (Rutilus rutihs). Behav. Ecol. Sociobiol. 30, 177-180. Krupa, J. J., and Sih, A. (1993). Experimental studies on water strider mating dynamics: Spatial variation in density and sex ratio. Behav. Ecol. Sociobiol. 33, 107-120. Kiilling, D., and Milinski, M. (1992). Size-dependent predation risk and partner quality in predator inspection of sticklebacks. Anim. Behav. 44,949-955. L’Abte-Lund, J. H., Langeland. A,, Jonsson. E., and Ugedal. 0. (1993). Spatial segregation by age and size in Arctic charr: A trade-off between feeding possibility and risk of predation. J. Anim. Ecol. 62, 160-168. Lagos, V. 0.. Contreras, L. C., Meserve, P. L., GutiCrrez, J. R., and Jaksic, F. M. (1995a). Effects of predation risk on space use by small mammals: A field experiment with a neotropical rodent. Oikos 74, 259-264. Lagos, V. O., Bozinovic, F., and Contreras, L. C. (1995b). Microhabitat use by a small diurnal rodent (Octoden degus) in a semiarid environment: Thermoregulatory constraints or predation risk? J. Mammal. 76,900-905. Lambin, X., Ims. R. A,, Steen, H., and Yoccoz, N . G. (1995). Vole cycles. Trends Ecol. Evol. 5, 204. Lampert, W. (1989). The adaptive significance of die1 vertical migration of zooplankton. Funct. Ecol. 3, 21-27. Larsson, P., and Dodson, S. (1993). Chemical communication in planktonic animals. Arch. Hydrobiol. 129, 129-155. Lawler, S. P. (1989). Behavioural responses to predators and predation risk in four species of larval anurans. Anim. Behav. 38, 1039-1047. Legault, C.. and Himmelman, J. H. (1993). Relation between escape behaviour of benthic marine invertebrates and the risk of predation. J. Exp. Mar. Biol. Ecol. 170,55-74. Leibold, M. A. (1990). Resources and predators can affect the vertical distributions of zooplankton. Limnol. Oceanogr. 35, 938-944. Leibold, M. A. (1991). Trophic interactions and habitat segregation between competing Daphnia species. Oecologia 86, 510-520.

PREDATOR-INDUCED STRESS AND BEHAVIOR

279

Le Maho, Y . , Karmann. H., Briot, D., Handrich, Y . , Robin, J.-P., Mioskowski, E., Cherel, Y . .and Farni, J. (1992). Stress in birds due to routine handling and a technique to avoid it. Am. J. Physiol. 263, R7755R781. Leonardsson, K. (1991). Predicting risk-taking behaviours from life-history theory using static optimization technique. Oikos 60, 149-154. Lester, L. S.. and Fanselow, M. S. (1992). Nocturnality as a defensive behavior in the rat: An analysis in terms of selective association between light and aversive stimulation. Psychological Record 42,221 -253. Levine, S.. Atha, K., and Wiener, S. G. (1993). Early experience effects on the development of fear in the squirrel monkey. Behav. Neural B i d . 60, 225-233. Levy. D. A. (1990a). Reciprocal die1 vertical migration behaviour in planktivores and zooplankton in British Columbia lakes. Can. J. Fish. Aqccaf. Sci. 47, 1755-1764. Levy, D. A. (1990b). Sensory mechanism and selective advantage for die1 vertical migration in juvenile sockeye salmon, Oncorhynchus nerka. Can. J. Fish. Aquat. Sci. 47,1796-1802. Licht, T. (1989). Discriminating between hungry and satiated predators: The response of guppies (Poecilia reticulafa) from high and low predation sites. Efhology 82, 238-243. Lima, S. L. (1990a). Protective cover and the use of space: Different strategies in finches. Oikos 58, 151-158. Lima, S. L. (1 990b). Evolutionarily stable antipredator behavior among isolated foragers: On the consequences of successful escape. J . Theor. Biol. 143, 77-89. Lima, S. L. (1991). Energy, predators and the behaviour of feeding hummingbirds. Evol. Ecol. 5,220-230. Lima, S. L. (1992). Vigilance and foraging substrate: Anti-predatory considerations in a nonstandard environment. Behav. Ecol. Sociobiol. 30, 283-289. Lima. S. L. (1993). Ecological and evolutionary perspectives on escape from predatory attack: A survey of North American birds. Wilson Bull. 105, 1-47. Lima, S. L. (1995). Back to the basics of anti-predatory vigilance: The group size effect. Anim. Behav. 49, 11-20. Lima, S. L.. and Dill, L. M. (1990). Behavioral decisions made under the risk of predation: A review and prospectus. Can. J . Zool. 68, 619-640. Lima, S. L., and Valone. T. J. (1991). Predators and avian community organization: An experiment in a semi-desert grassland. Oecologia 86, 105-112. Lima, S. L., and Zollner, P. A. (1996). Anti-predatory vigilance and the limits to collective detection: Visual and spatial separation between foragers. Brhav. Ecol. Sociobiol. 38, 355-363. Lindsay, S. M., and Woodin, S. A. (1995). Tissue loss induces switching of feeding mode in spionid polychaetes. Mar. Ecol: Prog. Ser. 125, 159-169. Lindstrom, A. (1990). The role of predation risk in stopover habitat selection in migrating bramblings, Fringilla montifringilln. Behav. Ecol. 1, 102-106. Lister, B. C., and Aguayo, A. G . (1992). Seasonality, predation, and the behaviour of a tropical mainland anole. J . Anim. Ecol. 61, 717-733. Lockard, R. B., and Owings, D. H. (1974). Moon-related surface activity of bannertail (Dipodomys spectabilis) and Fresno (D. nitratoides) kangaroo rats. Anim. Behav. 22, 262-273. Longland, W. S. (1991). Risk of predation and food consumption by black-tailed jackrabbits. J. Range Manage. 44,447-450. Longland, W. S., and Price, M. V. (1991). Direct observation of owls and heteromyid rodents: Can predation risk explain microhabitat use? Ecology 72, 2261-2273. Loose, C. J., and Dawidowicz, P. (1994). Trade-offs in die1 vertical migration by zooplankton: The cost of predator avoidance. Ecology 75,2255-2263.

280

STEVEN L. LIMA

Loose. C. J., von Elert. E.. and Dawidowicz. P. (1993). Chemically-induced diet vertical migration in Daphnia: A new bioassay for kairomones exuded by fish. Arch. Hydrohid. 126, 329-337. Lucas, J. R., and Howard, R. D. (1995). O n alternative reproductive tactics in anurans: Dynamic games with density and frequency dependance. Am. Nor. 146,365-397. Lucas, J. R., and Walter, L. R. (1991). When should chickadees hoard food‘? Theory and experimental results. Anim. Behav. 41, 579-601. Lucas, J. R.. Howard, R. D.. and Palmer. J . G. (1996). Callers and satellites: Chorus behaviour in anurans as a stochastic dynamic game. Anim. Behav. 51, 501-518. Ludwig, D., and Rowe, L. (1990). Life history strategies for energy gain and predator avoidance under time constraints. Am. Nar. 135, 686-707. Macchiusi, F., and Baker, R. L. (1991). Prey behaviour and size-selective predation by fish. Freshwater Biol. 25, 533-538. Macchiusi. F.. and Baker, R. L. (1992). Effects of predators and food availability on activity and growth of Chironomus tenfans (Chironomidae: Diptera). Freshwater Biol. 28,207-216. MacWhirter, R. B. (1991). Effects of reproduction on activity and foraging behaviour of adult female Colurnbian ground squirrels. Can. J. Zool. 69, 2209-2216. Magnhagen. C. (1990). Reproduction under predation risk in the sand goby, Pomaroschistus minutus, and the black goby, Gohirrs niger: The effect of age and longevity. Behav. Ecol. Sociobiol. 26, 331 -335. Magnhagen. C . (1991). Predation risk as a cost of reproduction. Trends Ecof. Evol. 6,183-186. Magnhagen. C. (1992). Parental care and predation risk in fish. Ann. Zool. Fenn. 29,227-232. Magnhagen, C. (1993). Conflicting demands in gobies: When to eat, reproduce. and avoid predators. Mar. Behav. Physiol. 23, 79-90. Magnhagen. C. (1995). Sneaking behaviour and nest defence are affected by predation risk in the common goby. Anim. Behnv. 50, 1123-1128. Magnhagen, C.. and Vestergaard, K. (1991). Risk taking in relation to reproductive investments and future reproductive opportunities: Field experiments on nest-guarding common gobies, Pomatoschistus microps. Behav. Ecol. 2, 351-359. Magurran. A . E. (1990). The adaptive significance of schooling as an anti-predator defence in fish. Ann. Zool. Fenn. 27, 51-66. Magurran, A . E., and Girling, S. L. (1986). Predator recognition and response habituation in shoaling minnows. Anim. Behav. 34, 510-518. Magurran. A. E., and Nowak. M. A. (1991). Another battle of the sexes: The consequences of sexual asymmetry in mating costs and predation risk in the guppy. Poecilia reticulata. Proc. R. Sue. London. Ser. B 246, 31-38. Magurran. A. E., and Seghers. B. H. (199th). Population differences in predator recognition and attack cone avoidance in the guppy Poecilia reticrtlata. Anim. Behav. 40, 443-452. Magurran. A. E., and Seghers, B. H. (1990b). Risk sensitive courtship in the guppy (Poecilia reticrtlata). Behaviour 112, 194-201. Magurran. A . E., and Seghers. B. H. (1994). Predator inspection behaviour covaries with schooling tendency amongst wild guppy. Poecilia reticulata. populations in Trinidad. Behaviour 128, 121-134. Malmqvist. B. (1988). Downstream drift in Madeiran levadas: Tests of hypotheses relating to the influence of predators o n the drift of insects. Aqiiot. Insects 10, 141-152. Mangel, M. (1989). Evolution of host selection in parasitoids: Does the state of the parasitoid matter? Am. Nut. 133, 688-705. Mangel. M.. and Clark, C . W. (1988). “Dynamic Modeling in Behavioral Ecology.” Princeton University Press, Princeton, NJ.

PREDATOR-INDUCED STRESS AND BEHAVIOR

281

Manteifel. Y. (1995). Chemically-mediated avoidance of predators by Rana temporaria tadpoles. J . Herpetol. 29, 461-463. Martel, G., and Dill, L. M. (1993). Feeding and aggressive behaviours in juvenile coho salmon (Oncorhynchus kisutch) under chemically-mediated risk of predation. Behav. Ecol. Soczob i d . 32, 365-370. Martel, G., and Dill, L. M. (1995). Influence of movement by coho salmon (Oncorhynchus kisutch) parr on their detection by common mergansers (Mergus merganser). Ethology 99, 139-149. Mathis, A., and Smith, R. J. F. (1992). Avoidance of areas marked with a chemical alarm substance by fathead minnows (Pimephales promelas) in a natural habitat. Can. J . Zool. 70, 1473-1476. Mathis, A., and Smith, R. J. F. (1993a). Fathead minnows, Pimephales promelas, learn to recognize northern pike, Esox lucius, as predators on the basis of chemical stimuli from minnows in the pike's diet. Anim. Behav. 46,645-656. Mathis, A,, and Smith, R. J. F. (1993b). Intraspecific and cross-superorder responses to chemical alarm signals by brook stickleback. Ecology 74, 2395-2404. Mathis, A,, Chivers, D. P., and Smith, R. J. F. (1993). Population differences in responses of fathead minnows (Pimephales promelas) to visual and chemical stimuli from predators. Ethology 93, 31-40. Matsuda, H., Abrams. P. A,, and Hori, M. (1993). The effect of adaptive anti-predator behavior on exploitative competition and mutualism between predators. Oikos 68,549-559. Matsuda. H.. Hori, M.. and Abrams, P. A. (1994). Effects of predator-specific defence on community complexity. Evol. Ecol. 8, 628-638. Matsuda, H., Hori, M., and Abrams, P. A. (1996). Effects of predator-specific defence on biodiversity and community complexity in two-trophic-level communities. Evol. Ecol. 10, 13-28. Mayhew, P., and Houston, D. (1989). Feeding site selection by widgeon Anus penelope in relation to water. Ibis 131, 1-8. McIntosh. A. R., and Townsend, C. R. (1994). Interpopulation variation in mayfly antipredator tactics: Differential effects of contrasting predatory fish. Ecology 75, 2078-2090. McLeod. P. G . ,and Huntingford, F. A. (1994). Social rank and predator inspection in sticklebacks. Anim. Behav. 47, 1238-1240. McNamara, J . M., and Houston, A. I. (1986). The common currency for behavioral decisions. Am. Nat. 127, 358-378. McNamara, J. M., and Houston, A. I. (1990). The value of fat reserves in terms of avoiding starvation. Acta Biotheor. 38, 37-61. McNamara. J . M.. and Houston, A. 1. (1992). Evolutionarily stable levels of vigilance as a function of group size. Anim. Behav. 43, 641-658. McNamara, J. M., and Houston. A. I. (1994). The effect of a change in foraging options on intake rate and predation rate. Am. Nut. 144, 978-1000. McNamara, J. M., Houston, A. I.. and Lima, S. L. (1994). Foraging routines of small birds in winter: A theoretical investigation. J. Avian Biol. 25, 287-302. McNamara, J. M., Webb, J. N., and Collins, E. J . (1995). Dynamic optimization in fluctuating environments. Proc. R. Soc. London, Ser. B 261,279-284. McNeil, R.. Drapeau, P., and Goss-Custard, J . D. (1992). The occurrence and adaptive significance of nocturnal habits in waterfowl. Biol. Rev. Cambridge Philos. Soc. 67, 381-419. McPeek. M. A. (1990). Behavioral differences between Enalfagmaspecies (Odonata) influencing differential vulnerability to predators. Ecology 71, 1714-1726. McPeek, M. A,. Schrot, A. K.. and Brown, J. M. (1996). Adaptation to predators in a new community: Swimming performance and predator avoidance in damselflies. Ecology 77, 617-629.

282

STEVEN L. LIMA

Merkens, M., Harestad, A. S., and Sullivan, T. P. (1991). Cover and efficacy of predatorbased repellents for Townsend’s vole, Microtus townsendii. J. Chem. Ecol. 17, 401-41 2. Mesa, M. G., Poe, T. P., Gadomski, D. M., and Petersen, J. H. (1994). Are all prey created equal? A review and synthesis of diffrential predation on prey in substandard condition. J . Fish Biol. 45, Suppl. A, 81-96. Michiels, N. K., and Dhondt, A. A. (1990). Costs and benefits associated with oviposition site selection in the dragonfly Sympetrum danae (Odonata: Libellulidae). Anim. Behav. 40,668-678. Milinski, M. (1986). Constraints placed by predators on feeding behaviour. In “The Behaviour of Teleost Fishes” (T. Pitcher, ed.). pp. 236-252. Croom Helm, London. Milinski, M. (1987). Tit for tat in sticklebacks and the evolution of cooperation. Nature (London) 325,433-435. Milinski, M. (1993). Predation risk and feeding behaviour. In “Behaviour of Teleost Fishes” (T. Pitcher, ed.), 2nd ed., pp. 285-305. Chapman & Hall, London. Milinski, M., and Heller, R. (1978). Influence of a predator on the optimal foraging behaviour of stickelbacks (Gasterosteus aculeatus L.). Nature (London) 275, 642-644. Milinski. M., and Parker, G. A. (1991). Competition for resources. In “Behavioral Ecology: An Evolutionary Approach” ( J . R. Krebs and N. B. Davies. eds.). 3rd ed., pp. 137-168. Blackwell. Oxford. Milinski, M., Kiilling, D., and Kettler, R. (1990a). Tit for tat: Sticklebacks (Gasterosteus aculeatus) “trusting” a cooperating partner. Behav. Ecol. 1, 7-1 1. Milinski. M.. Pfluger, D.. Kulling, D., and Kettler, R. (1990b). Do sticklebacks cooperate repeatedly in reciprocal pairs? Behuv. Ecol. Sociobiol. 27, 17-21. Moody, A. L., Houston, A. I., and McNamara, J. M. (1996). Ideal free distributions under predation risk. Behav. Ecol. Sociobiol. 38, 131-143. Moore, F. R. (1994). Resumption of feeding under risk of predation: Effect of migratory condition. Anim. Behav. 48, 975-977. Morey, S. R. (1990). Microhabitat selection and predation in the pacific treefrog, Pseudacris regilla. J. Herpetol. 24, 292-296. Mourn. S. E., and Baker, R. L. (1990). Colour change and substrate selection in larval Ischnuru verticalis (Coenagrionidae: Odonata). Can. J. Zool. 68, 221-224. Murphy, K. E., Pitcher, T. J . (1991). Individual behavioural strategies associated with predator inspection in minnow shoals. Ethology 88, 307-319. Neill, W. E. (1992). Population variation in the ontogeny of predator-induced vertical migration of copepods. Nuture (London) 356,54-57. Nemtzov, S. C. (1994). Intraspecific variation in sand-diving and predator avoidance behavior of green razorfish, Xyrichtys splendens (Pisces, Labridae): Effect on courtship and mating success. Environ. Biol. Fishes 41, 403-414. Newman, J. A. (1991). Patch use under predation hazard: Foraging behavior in a simple stochastic environment. Oikos 61, 29-44. Nonacs, P. (1990). Death in the distance: Mortality risk as information for foraging ants. Behaviour 112,23-35. Nonacs, P., and Dill, L. M. (1990). Mortality risk vs. food quality in a common currency: Ant patch preferences. Ecology 71, 1886-1892. Ohman, M. D. (1990). The demographic benefits of die1 vertical migration by zooplankton. Ecol. Monogr. 60, 257-281. Oksanen, L., and Lundberg. P. (1995). Optimization of reproductive effort and foraging time in mammals: The influence of resource level and predation risk. Evol. Ecol. 9,45-56. Otto, C. (1993). Long-term risk sensitive foraging in Rhyacophila nubila (Trichoptera) larvae from two streams. Oikos 68, 67-74.

PREDATOR-INDUCED STRESS AND BEHAVIOR

283

Packer, C., and Abrams, P. (1990). Should co-operative groups be more vigilant than selfish groups? J. Theor. Biol. 142, 341-357. Palmer, A. R. (1990). Effect of crab effluent and scent of damaged conspecifics on feeding, growth, and shell morphology of the Atlantic dogwhelk Nucella lapillits (L.). Hydrobiologia 193, 155-182. Parejko. K.. and Dodson, S. (1990). Progress towards characterization of a predatodprey kairomone: Daphnia pirlex and Chaoborus americanus. Hydrobiologia 198, 51-59. Paveri-Fontana, S. L., and Focardi, S. (1994). A theoretical study of the socioecology of ungulates. 11. A dynamic programming study of the stochastic formulation. Theor. Popul. Biol. 46,279-299. Peckarsky, B. L., Cowan, C. A,. Penton, M. A.. and Anderson, C. (1993). Sublethal consequences of stream-dwelling predatory stoneflies on mayfly growth and fecundity. Ecology 74, 1836-1846. Persson, L. (1991). Behavioral response to predators reverses the outcome of competition between prey species. Behav. Ecol. Sociobiol. 28, 101-105. Persson, L. (1993). Predator-mediated competition in prey refuges: The importance of habitat dependent prey resources. Oikos 68, 12-22. Persson. L., and Eklov, P. (1995). Prey refuges affecting interactions between piscivorous perch and juvenile perch and roach. Ecology 76,70-81. Persson, L., Anderson, J., Wahlstrom, E., and Eklov, P. (1996). Size-specific interactions in lake systems: Predator gape limitation and prey growth rate and mortality. Ecology 77,900-91 1. Peterson, C. H., and Skilleter, G. A. (1994). Control of foraging behavior of individuals within an ecosystem context: The clam Macoma balthica, flow environment, and siphon-cropping fishes. Oecologia 100,256-267. Pettersson, L. B., and Bronmark, C. (1993). Trading off safety against food: State dependent habitat choice and foraging in crucian carp. Oecologia 95, 353-357. Phelan, J. P., and Baker, R. H. (1992). Optimal foraging in Peromyscus polionotus: The influence of item-size and predation risk. Behaviour 121, 95-109. Pierce, B. M., Longland, W. S., and Jenkins, S. H. (1992). Rattlesnake predation on desert rodents: Microhabitat and species-specific effects on risk. J. Mammal. 73, 859-865. Pitcher, T. J . (1992). Who dares, wins: The function and evolution of predator inspection behaviour in shoaling fish. Neth. J. Zool. 42, 371-391. Pitcher, T. J., Green, D. A,, and Magurran, A. E. (1986). Dicing with death: Predator inspection behaviour in minnow shoals. J. Fish Biol. 28,439-448. Poff, N. L., DeCino, R. D., and Ward, J. V. (1991). Size-dependent drift responses of mayflies to experimental hydrolic variation: Active predator avoidance or passive hydrodynamic displacement? Oecologia 88, 577-586. Power, M. E. (1992). Top-down and bottom-up forces in food webs: Do plants have primacy? Ecology 73, 733-746. Power, M. E., Matthews, W. J., and Stewart, A. J. (1985). Grazing minnows, piscivorous bass, and stream algae: Dynamics of a strong interaction. Ecology 66, 1448-1456. Poysa, H. (1991). Effects of predation risk and patch quality on the formation and attractiveness of foraging groups of teal, Anas crecca. Anirn. Behav. 41,285-294. Price, M. V.. Goldingay, R. L., Szychowski, L. S.. and Waser, N. M. (1994). Managing habitat for the endangered stephen’s kangaroo rat (Dipodomys stephensi): Effects of shrub removal. Am. Midl. Nat. 131, 9-16. Prins, H. H. T., and Iason. G. R. (1989). Dangerous lions and nonchalant buffalo. Behaviour 108, 262-296.

284

STEVEN L. LIMA

Pulliam, H. R., and Caraco. T. (1984). Living in groups: Is there an optimal group size? In “Behavioural Ecology: an Evolutionary Approach” ( J . R. Krebs and N. B. Davies. eds.). 2nd ed., pp. 127-147. Blackwell, Oxford. Radabaugh, D. C. (1989). Seasonal colour changes and shifting antipredator tactics in darters. J. Fish Biol. 34, 679-685. Rader, R. B., and McArthur, J. V. (1995). The relative importance of refugia in determining the drift and habitat selection of predaceous stoneflies in a sandy-bottomed stream. Oecologia 103, 1-9. Rahel. F. J., and Kolar, C. S. (1990). Trade-offs in the response of mayflies to low oxygen and fish predation. Oecologia 84, 39-44. Ramcharan, C . W., and Sprules, W. G. (1991). Predator-induced behavioral defense and its ecological consequences for two calanoid copepods. Oecologia 86,276-286. Ramcharan, C. W., Dodson, S. I., and Lee, J. (1992). Predation risk, prey behavior, and feeding rate in Daphnia pulex. Can. J . Fish. Aquut. Sci. 49, 159-165. Ranta, E., Juvonen. S.-K., and Peuhkuri, N. (1992a). Further evidence for size-assertive schooling in sticklebacks. J. Fish Biol. 41, 627-630. Ranta, E., Lindstrom, K.. and Peuhkuri, N. (1992b). Size matters when three-spined sticklebacks go to school. Anim. Behav. 43, 160-162. Rayor, L. S., and Uetz, G. W. (1990). Trade-offs in foraging success and predation risk with spatial position in colonial spiders. Behav. Ecol. Sociobiol. 27, 77-85. Rayor, L. S., and Uetz, G. W. (1993). Ontogenetic shifts within the selfish herd: Predation risk and foraging trade-offs change with age in colonial web-building spiders. Oecologia 95, 1-8. Rehfeldt, G. E. (1990). Anti-predator strategies in oviposition site selection of Pyrrhosoma nyrnphula (Zygoptera: Odonata). Oecologia 85, 233-237. Rehfeldt, G . E. (1992). Aggregation during oviposition and predation risk in Syrnpetrum vulgatum L. (Odonata: Libellulidae). Behav. Ecol. Sociohiol. 30, 317-322. Repasky, R. R., and Schluter, D. (1994). Habitat distribution of wintering sparrows along an elevational gradient: Tests of the food. predation and microhabitat structure hypotheses. J . Anim. Ecol. 63,569-582. Resetarits, W. J., Jr.. and Wilbur, H. M. (1989). Oviposition site choice in Hyla chrysoscelis: Role of predators and competitors. Ecology 70, 220-228. Reynolds, J. D. (1993). Should attractive individuals court more? Theory and a test. Am. Nar. 141,914-927. Richardson, T. D., and Brown. K. M. (1992). Predation risk and feeding in an intertidal predatory snail. J . Exp. Mar. Biol. Ecoi. 163, 169-182. Ringelberg, J. (1991a). Enhancement of the phototactic reaction in Dnphnia hyalina by a chemical mediated by juvenile perch (Perca puviatilis). J. Plankton Res. 13, 17-25. Ringelberg, J. (1991b). A mechanism of predator-mediated induction of die1 vertical migration in Daphnia hyalina. J. Plankton Res. 13, 83-89. Ringelberg, J., Flik, B. J. G., Lindenaar, D., and Royackers, K. (1991). Die1 vertical migration of Daphnia hyalina (sensu latiori) in Lake Maarsseveen: Part 1. Aspects of seasonal and daily timing. Arch. Hydrobiol. 121, 129-145. Rittschoff, D., Tsai, D. W., Massey. P. G., Blanco, L., Kueber. G . L., Jr., and Haas, R. J., Jr. (1992). Chemical mediation of behavior in hermit crabs: Alarm and aggregation cues. J. Chem. Ecol. 18,959-984. Roberts. G . (1994). When to scan: An analysis of predictability in vigilance sequences using autoregression models. Anim. Behav. 48, 579-585. Roberts, G. (1995). A real-time response of vigilance behaviour to changes in group size. Anim. Behav. 50, 1371-1374.

PREDATOR-INDUCED STRESS A N D BEHAVIOR

285

Roberts. G. (1996). Why individual vigilance declines as group size increases. Anim. Behav. 51,1077-1086. Roca, J . R., Baltanas, A., and Uiblein, F. (1993). Adaptive responses in Cypridopsis vidua (Crustacea: Ostracoda) to food and shelter offered by a macrophyte (Chara frugilis). Hydrobiologia 262, 127- 13 I. Rochette, R., Morissette. S.. and Himmelman. J. H. (1995). A flexible response to a major predator provides a whelk Buccinum undatuni L. with nutritional gains. J. Exp. Mar. B i d . Ecol. 185, 167-180. Romey, W. L. (1995). Position preferences within groups: Do whirligigs select positions which balance feeding opportunities with predator avoidance? Behav. E d . Sociobiol. 37, 195-200. Romey, W. L., and Rossman, D. S. (1995). Temperature and hunger alter grouping tradeoffs in whirligig beetles. A m . Midl. Nor. 134, 51-62. Ronkainen, H., and Ylonen, H. (1994). Behaviour of cyclic bank voles under risk of mustelid predation: Do females avoid copulations? Oecologia 97, 377-381. Rosland, R.. and Giske, J. (1994).A dynamic optimization model of the die1 vertical distribution of pelagic planktivorous fish. Prog. Oceanogr. 34, 1-43. Rowe, L. (1994). The costs of mating and mate choice in water striders. Anim. Behav. 48,10491056. Rowe, L., and Ludwig, D. (1991). Size and timing of metamorphosis in complex life cycles: Time constraints and variation. Ecology 72,413-427. Ruiz, G . M.. Hines, A. H., and Posey, M. H. (1993). Shallow water as a refuge habitat for fish and crustaceans in non-vegetated estuaries: An example from Chesapeake Bay. Mar. Ecol.: Prog. Ser. 99, 1-16. Ruxton, G. D. (1995). Short term refuge use and stability of predator-prey models. Theor. Popul. B i d . 47, 1-17. Saarikko, J. (1992). Risk of predation and foraging activity in shrews. Ann. Zool. Fenn. 29,291-299. Sargent. R. C. (1990). Behavioural and evolutionary ecology of fishes: Conflicting demands during the breeding season. Ann. Zool. Fenn. 27, 101-118. Scarratt, A. M., and Godin, J.-G.J. (1992). Foraging and antipredator decisions in the hermit crab Pagtcrus acadianus (Benedict). J. Exp. Mar. Bid. Ecol. 156, 225-238. Schlosser, I . J. (1987). The role of predation in age- and size-related habitat use by stream fishes. Ecology 68, 651 -659. Schluter, D. (1988). The evolution of finch communities on islands and continents, Kenya vs. Galapagos. Ecol. Monogr. 58,229-249. Schluter. D., and Repasky. R. R. (1991). Worldwide limitation of finch densities by food and other factors. Ecology 72, 1763-1 774. Schooley. R. L., Sharpe, P. B., and Van Horne, B. (1996). Can shrub cover increase predation risk for a desert rodent? Can. J. Zool. 74, 157-163. Schreck, C. B. ( 1990). Physiological, behavioral, and performance indicators of stress. A m . Fish. SOC.Symp. 8, 29-37. Schwinning, S., and Rosenzweig, M. L. (1990). Periodic oscillations in an ideal-free predatorprey distribution. Oikos 59, 85-91. Scrimgeour, G. J.. and Culp, J. M. (1994a). Foraging and evading predators: The effect of predator species on a behavioural trade-off by a lotic mayfly. Oikos 69, 71-79. Scrimgeour, G. J.. and Culp, J. M. (1994b). Feeding while evading predators by a lotic mayfly: Linking short-term foraging behaviours to long-term fitness consequences. Oecologia 100, 128-134.

286

STEVEN L. LIMA

Scrimgeour, G. J., Culp, J. M., and Wrona, F. J. (1994). Feeding while avoiding predators: Evidence for a size-specific trade-off by a lotic mayfly. J. Norrh Am. Benfhol. Soc. 13, 368-378. Short, T. M., and Holomuzki, J. R. (1992). Indirect effects of fish on foraging behaviour and leaf processing by the isopod Lirceus fonfinaiis. Freshwater Biol. 27, 91 -97. Sibly, R. M.. and Calow, P. (1989). A life-cycle theory of responses to stress. B i d . J. Linn. SOC. 37, 101-116. Sih, A. (1980). Optimal behavior: can foragers balance two conflicting demands? Science 210, 1041-1043. Sih, A. (1984). The behavioral response race between predator and prey. Am. Nut. 123, 143-150. Sih, A. (1987). Predators and prey lifestyles: An evolutionary and ecological overview. In “Predation: Direct and Indirect Impacts on Aquatic Communities” (W. C. Kerfoot and A. Sih, eds.), pp. 203-224. University Press of New England, Hanover, NH. Sih, A. (1988). The effects of predators on habitat use, activity and mating behaviour of a semi-aquatic bug. Anim. Behav. 36, 1846-1848. Sih, A. (1992a). Prey uncertainty and the balancing of antipredator and feeding needs. Am. Nut. 139, 10.52-1069. Sih, A. (1992b). Integrative approaches to the study of predation: General thoughts and a case study on sunfish and salamander larvae. Ann. Zool. Fenn. 29, 183-198. Sih, A. (1994). Predation risk and the evolutionary ecology of reproductive behaviour. J. Fish Biol. 45, 111-130. Sih, A. (1998). Three trophic level ideal free distributions: A game-theory approach to understanding the predator-prey behavioral response race. In “Advances in Game Theory and the Study of Animal Behavior” (L. A. Dugatkin and H. K. Reeve, eds.), pp. 221-238. Oxford University Press, Oxford. Sih, A,, and Kats, L. B. (1991). Effects of refuge availability on the responses of salamander larvae to chemical cues from predatory green sunfish. Anim. Behav. 42, 330-332. Sih. A,, and Krupa, J. J. (1992). Predation risk, food deprivation and non-random mating by size in the stream water strider, Aquarius remigis. Behuv. Ecol. Sociobiol. 31, 51-56. Sih, A,, and Krupa, J. J. (1995). Interacting effects of predation risk and male and female density on malelfemale conflicts and mating dynamics of stream water striders. Behav. E d . 6,316-325. Sih, A., and Krupa, J. J. (1996). Direct and indirect effects of multiple enemies on water strider mating dynamics. Oecologiu 105, 179-188. Sih, A., and Wooster, D. E. (1994). Prey behavior, prey dispersal, and predator impacts on stream prey. Ecology 75, 1199-1207. Sih, A., Petranka, J. W., and Kats, L. B. (1988). The dynamics of prey refuge use: A model and tests with sunfish and salamander larvae. Am. Nat. 132,463-483. Sih, A,. Krupa, J. J., and Travers, S. (1990). An experimental study on the effects of predation risk and feedingregime on the mating behavior of the water strider. Am. Nut. 135,284-290. Sih, A., Kats, L. B., and Moore, R. D. (1992). Effects of predatory sunfish on the density, drift. and refuge use of stream salamander larvae. Ecology 73,1418-1430. Sirnonetti, J. A. (1989). Microhabitat use by small mammalsincentral Chile. Oikos56,309-318. Sinclair, A. R. E.. and Arcese, P. (1995). Population consequences of predation-sensitive foraging: The Serengeti wildebeest. Ecology 76,882-891. Skelly, D. K. (1992). Field evidence for a cost of behavioral antipredator response in a larval amphibian. Ecology 73, 704-708. Skelly, D. K. (1994). Activity level and the susceptibility of anuran larvae to predation. Anim. Behav. 47, 465-468.

PREDATOR-INDUCED STRESS AND BEHAVIOR

287

Skelly, D. K. (1995). A behavioral trade-off and its consequences for the distribution of Pseudacris treefrog larvae. Ecology 76, 150-164. Skelly, D. K., and Werner, E. E. (1990). Behavioral and life-historical responses of larval American toads to an odonate predator. Ecology 71, 2313-2322. Slotow, R., and Rothstein, S. I. (1995). Importance of dominance status and distance from cover to foraging white-crowned sparrows: An experimental analysis. Auk 112,107-117. Smith, R. J. (1995). Harvest rates and escape speeds in two co-existing species of montane ground squirrels. J. Mammal. 76, 189-195. Smith, R. J. F. (1992). Alarm signals in fishes. Rev. Fish Biol. Fish. 2, 33-63. Smith, W. P. (1991). Ontogeny and adaptiveness of tail flagging behavior in white-tailed deer. Am. Nat. 138, 190-200. Soderback, B. (1994). Interactions among juveniles of two freshwater crayfish species and a predatory fish. Oecologia 100, 229-235. Sogard, S. M., and Olla, B. L. (1993). The influence of predator presence on utilization of artificial seagrass habitats by juvenile walleye pollock, Theragra chalcogramma. Environ. Biol. Fishes 37, 57-65. Sparrevik, E., and Leonardsson, K. (1995). Effects of large Saduria entomon (Isopoda) on spatial distribution of their small S. enfornon and Monoporeia afinis (Amphipoda) prey. Oecologia 101, 177-184. Speakman, J. R., Stone, R. E., and Kerslake, J. E. (1995). Temporal patterns in the emergence behaviour of pipistrelle bats, Pipistrellus pipisfrellus, from maternity colonies are consistent with the anti-predator response. Anim. Behav. 50, 1147-1156. Stadler, B., Weisser, W. W., and Houston, A. I. (1994). Defence reactions in aphids: The influence of state and future reproductive success. J. Anim. Ecol. 63,419-430. Stamp, N. E., and Bowers, M. D. (1991). Indirect effect on survivorship of caterpillars due to presence of invertebrate predators. Oecologia 88,325-330. Stamp, N. E., and Bowers, M. D. (1993). Presence of predatory wasps and stinkbugs alters foraging behavior of cryptic and non-cryptic caterpillars on plantain (Plantago lanceolafa). Oecologia 95, 376-384. Stanford, C. B. (1995). The influence of chimpanzee predation on group size and anti-predator behaviour in red colobus monkeys. Anim. Behav. 49,577-587. Stauffer, H.-P., and Semlitsch, R. D. (1993). Effects of visual, chemical and tactile cues of fish on the behavioural responses of tadpoles. Anim. Behav. 46, 355-364. Stephens, D. W., and Krebs, J. R. (1986). “Foraging Theory.” Princeton University Press, Princeton, NJ. Stephens, D. W., Anderson, J. P., and Benson, K. E. (1997). On the spurious occurrence of Tit for Tat in pairs of predator-approaching fish. Anim. Behav. 53, 113-131. Suhonen, J. (1993a). Predation risk influences the use of foraging sites by tits. Ecology 74,11971203. Suhonen. J. (1993b). Risk of predation and foraging sites of individuals in mixed-species tit flocks. Anim. Behav. 45, 1193-1198. Suhonen, J., Halonen, M., and Mappes, T. (1993). Predation risk and the organization of the Pnrus guild. Oikos 66, 94-100. Suhonen, J., Norrdahl, K., and Korpimaki, E. (1994). Avian predation risk modifies breeding bird community on a farmland area. Ecology 75, 1626-1634. Sweitzer, R. A,, and Berger, J. (1992). Size-related effects of predation on habitat use and behavior of porcupines (Erethizon dorsatum). Ecology 73, 867-875. Szekely, T., Sozou, P. D., and Houston, A. I. (1991). Flocking behaviour of passerines: A dynamic model for the non-reproductive season. Behav. Ecol. Sociobiol. 28, 203-213.

288

STEVEN L. LIMA

Tayasu, I., Shigesada. N.. Mukai, H., and Caswell, H. (1996). Predator-mediated coexistence of epiphytic grass shrimps that compete for refuges. Ecol. Morlell. 84, 1-10. Tegeder, R. W.. and Krause, J. (199.5). Density dependence and numerosity in fright stimulated aggregation behaviour of shoaling fish. Philos. Trans. R. Soc. London. Ser. B 350,381-390. Theodorakis, C. W. (1989). Size segregation and the effects of oddity on predation risk in minnow schools. Anim. Behav. 38, 496-502. Tikkanen, P., Muotka, T., and Huhta. A. (1994). Predator detection and avoidance by lotic mayfly nymphs of different size. Oecologia 99, 252-259. Tjossem, S. F. (1990). Effects of fish chemical cues on vertical migration behavior of Chaoborus. Limnol. Oceanogr. 35, 1456-1468. Todd, 1. A., and Cowie, R. J. (1990). Measuring the risk of predation in an energy currency: Field experiments with foraging blue tits, Parirs caemleus. Anim. Behav. 40, 112-1 17. Tonn, W. M., Paszkowski, C. A,, and Holopainen, I. J. (1992). Piscivory and recruitment: Mechanisms structuring prey populations in small lakes. Ecology 73, 951-958. Turner, A. M. (1996). Freshwater snails alter habitat use in response to predation. Anim. Behav. 51,747-756. Turner, A. M. (1997). Contrasting short-term and long-term effects of predation risk on consumer habitat use and resource dynamics. Behav. Ecol. 8, 120-125. Turner. A. M., and Mittelbach. C. G. (1990). Predator avoidance and community structure: Interactions among piscivores, planktivores, and plankton. Ecology 71, 2241-22.54. Travers, S. E.. and Sih, A. (1991). The influence of starvation and predators on the mating behavior of a semiaquatic insect. Ecology 72, 2123-2136. Uetz, G. W.. and Hieber, C. S. (1994). Group size and predation risk in colonial web-building spiders: Analysis of attack abatement mechanisms. Behav. Ecol. 5, 326-333. Utne, A. C. W., Aksnes, D. L.. and Giske. J. (1993). Food, predation risk and shelter: An experimental study on the distribution of adult two-spotted goby Gobiuscultrspavescens (Fabricius). J. Exp. Mar. Biol. Ecol. 166, 203-216. Uzendoski, K., Maksymovitch, E.. and Verell. P. (1993). D o the risks of predation and intermale competition affect courtship behavior in the salamander Desmognathus ochrophaeus? Behav. Ecol. Sociohiol. 32, 42 1-427. Vadas, R. L., Sr.. Burrows, M. T., and Hughes, R. N. (1994). Foraging strategies of dogwhelks. Nucella lapillus (L.): Interacting effects of age, diet and chemical cues to the threat of predation. Oecologia 100, 439-450. van Baalen, M., and Sabelis. M. W. (1992). Coevolution of patch selection strategies of predator and prey and the consequences for ecological stability. Am. Not. 142, 646-670. Van Damme, R., Bauwens. D., Vanderstighelen, D., and Verheyen, R. F. (1990). Responses of the lizard Lacerto vivipara to predator chemical cues: The effects of temperature. Anim. Behav. 40,298-305. van Schaik, C. P., and Horstermann. M. (1994). Predation risk and the number of adult males in a primate group: A comparative test. Behav. Ecol. Sociohiol. 35, 261-272. Visquez. R. A. (1994). Assessment of predation risk via illumination level: Facultative central place foraging in the cricetid rodent Phyllotis darwini. Behav. Ecol. Sociobiol. 34,375-381. Vaughn, C. C.. and Fisher, F. M. (1988). Vertical migration as a refuge from predation in intertidal marsh snails: A field test. J . Exp. Mar. B i d . Eco/. 123, 163-176. Vega-Redondo. F., and Hasson, 0.(1993). A game-theoretic model of predator-prey signaling. J . Theor. B i d . 162, 309-319. Vuorinen. I . (1987). Vertical migration of Eiirytemora (Crustacea, Copepoda): A compromise between the risks of predation and decreased fecundity. J . Hankton Res. 9, 1037-1046. Wahle. R. A. (1992). Body-size dependent anti-predator mechanisms of the American lobster. Oikos 65,.52-60.

PREDATOR-INDUCED STRESS AND BEHAVIOR

289

Wahle, R. A., and Steneck, R. S. (1Y92). Habitat restrictions in early benthic life: Experiments on habitat selection and in situ predation with the American lobster. J . Exp. Mar. Biol. Ecol. 157, 91-114. Walls, S. C. (199.5). Differential vulnerability to predation and refuge use in competing larval salamanders. Oecologia 101, 86-93. Walters. C. J., and Juanes, F. (1993). Recruitment limitation as a consequence of natural selection for use of restricted feeding habitats and predation risk taking by juvenile fishes. Ciin. J . Fish. Ayuuf. Sci. 50, 2058-2070. Watt, P. J.. and Young, S. (1YY4). Effect of predator chemical cues on Duphniu behaviour in both horizontal and vertical planes. Anim. Behrrv. 48, 861-889. Watts, B. D. (1990). Cover use and predator-related mortality in song and savannah sparrows. Auk 107, 775-778. Watts, B. D. (1991). Effects of predation risk on distribution within and between habitats in savannah sparrows. Ecology 72, 1515-1.519. Weary, D. M., Pajor. E. A,. Thompson, B. K., and Fraser, D. (1996). Risky behaviour by piglets: A trade off between feeding and risk of mortality by maternal crushing? Anim. Behav. 51,619-624. Weiner. H. (1992). “Perturbing the Organism: The Biology of Stressful Experience.” University of Chicago Press, Chicago. Weisser, W. W., Houston. A. I., and Volkl, W. (1994). Foragingstrategies in solitary parasitoids: The trade-off between female and offspring mortality rates. Evol. Ecol. 8, 5877597. Weldon. P. J. (1990). Responses by vertebrates to chemicals from predators. In “Chemical Signals in Vertebrates 5” (D. W. Macdonald, D. Miiller-Schwarze. and S. E. Natynczuk, eds.), pp. 500-521. Oxford University Press, Oxford. Werner. E. E. (1986). Amphibian metamorphosis: Growth rate, predation risk, and the optimal size at transformation. Am. Nut. 128, 319-341. Werner, E. E. (1991). Nonlethal effects of a predator on competitive interactions between two anuran larvae. Ecology 72, 1709-1720. Werner, E. E. (1992). Individual behavior and higher-order species interactions. Am. Nat. 140, S5-S32. Werner, E. E. (1994). Ontogenetic scaling of competitive relations: Size-dependent effects and responses in two anuran larvae. Ecology 75, 197-213. Werner, E. E., and Anholt, B. R. (1993). Ecological consequences of the trade-off between growth and mortality rates mediated by foraging activity. Am. Nat. 142, 242-272. Werner, E. E., and Gilliam, J. F. (1984). The ontogenetic niche and species interactions in size-structured populations. Annu. Rev. Ecol. Sysf. 15, 393-425. Werner, E. E., and McPeek. M. A. (1994). Direct and indirect effects of predators on two anuran species along an environmental gradient. Ecology 75, 1368-1382. Werner, E. E., Gilliam, J. F., Hall, D. J., and Mittelbach, G. G. (1983). An experimental test of the effects of predation risk on habitat use in fish. Ecology 64, 1540-1548. Williams, D. D. (1990). A field study of the effects of water temperature. discharge and trout odour on the drift of stream invertebrates. Arch. Hydrohiol. 119, 167-181. Williams, P. J., and Brown, J. A. (1991). Developmental changes in foraging-predator avoidance trade-offs in larval lumpfish Cyclopterus lumpus. Mar. Ecol.: Prog. Ser. 76,.53-60. Willman, E. J., Hill, A. M., and Lodge, D. M. (1994). Response of three crayfish congeners (Orconectes spp.) to odors of fish carrion and live predatory fish. Am. Midl. Naf. 132,44451. Wilson. D. J., and Lefcort. H. (1993). The effect of predator diet on the alarm response of red-legged frog, Rana aurora, tadpoles. Anim. Behav. 46, 1017-1019. Winkleman, D. L., and Aho, J. M. (1993). Direct and indirect effects of predation on mosquitofish behavior and survival. Oecologia 96, 300-303.

290

STEVEN L. LIMA

Wisenden, B. D., Chivers, D. P., and Smith, R. J. F. (1994). Risk-sensitive habitat use by brook stickleback (Culaea inconstans) in areas associated with minnow alarm pheromone. J. Chem. Ecol. 20,2975-2983. Witter, M. S., and Cuthill, I. C. (1993). The ecological costs of avian fat storage. Philos. Trans. R. SOC. London, Ser. B 340,73-92. Witter, M. S., Cuthill, I. C., and Bonser, R. H. C. (1994). Experimental investigations of mass-dependent predation risk in the European starling, Sturnus vulgaris. Anim. Behav. 48,201-222. Wolf, N. G. (1985). Odd fish abandon mixed-species groups when threatened. Behav. Ecol. Sociohiol. 17, 47-52. Wolfe, J. L., and Summerlin, C. T. (1989). The influence of lunar light on nocturnal activity of the old-field mouse. Anim. Behav. 37,410-414. Wooster, D., and Sih, A. (1995). A review of the drift and activity responses of stream prey to predator presence. Oikos 73, 3-8. Wrona, F. J., and Dixon, R. W. J. (1991). Group size and predation risk: A field analysis of encounter and dilution effects. Am. Nut. 137, 186-201. Ydenberg, R. C. (1994). The behavioral ecology of provisioning in birds. b s c i e n c e 1,1-14. Ydenberg, R. C., and Dill, L. M. (1986). The economics of fleeing from predators. Adv. Study Behav. 16,229-249. Ylonen, H. (1989). Weasels Mustela nivalis suppress reproduction in the cyclic bank voles Clethrionomys glareolus. Oikos 55, 138-140. Ylonen, H. (1994). Vole cycles and antipredatory behaviour. Trends Ecol. Evol. 9,426-430. Ylonen, H., and Ronkainen, H. (1994). Breeding suppression in the bank vole as antipredatory adaptation in a predictable environment. Evol. Ecol. 8, 658-666. Ylonen, H., Jedrzejewska, B., Jedrzejewski, W., and Heikkila, J. (1992). Antipredatory behaviour of Clethrionomys voles: ‘David and Goliath’ arms race. Ann. 2001.Fenn. 29,207-216. Young, S., and Watt, P. (1993). Behavioral mechanisms controlling vertical migration in Daphnia. Limnol. Oceanogr. 38, 70-79. Ziv, Y., Abramsky, Z . , Kotler, B. P., and Subach, A. (1993). Interference competition and temporal and habitat partitioning in two gerbil species. Oikos 66, 237-246.

ADVANCES IN THE STUDY OF BEHAVIOR, VOL. 27

Parasitic Stress and Self-Medication in Wild Animals GEORGE A. LOZANO* DEPARTMENT OF BIOLOGICAL SCIENCES UNIVERSITY OF CALIFORNIA

RIVERSIDE, CALIFORNIA

92521

I. INTRODUCTION In the physical sciences, “stress” is defined as the force per unit area, or pressure, acting upon a solid body, resulting in the deformation (strain) of the solid. At low stresses the strain is said to be elastic, directly proportional to the stress and reversible; the solid returns to its original shape after the stress is removed. As the stress increases, the elastic limit is reached, after which the strain is said to be plastic, increasing exponentially with increasing stress and nonreversible. Plastic deformation continues until the rupture strength is reached, at which point the material breaks. The term stress was adopted by biologists to refer to factors that interfere with the maintenance of homeostasis, the effects of which range from the minor, temporary, and easily reversible, to the complete breakdown of homeostatic mechanisms (Cannon, 1935). As applied to vertebrates, the term “stress” is generally used to denote stimuli that elicit a specific set of physiological responses, particularly the release of corticosteroids (Vander, 1981; Kopin, 1995; Mims et al., 1995). However, these responses are not characteristic of all taxa, so this definition is not inclusive. Stresses can also be defined more broadly as aversive stimuli (McGrath, 1970; Selye, 1976), regarded as selective forces, and studied along with the adaptations that have evolved to reduce their negative effects (see Thornhill and Furlow, this volume). Under this view, stresses can take a myriad of forms, as indicated by the wide range of topics included in this volume. Along with competition and predation, parasitism is one of the main sources of biotic stress facing all organisms. For the purposes of this discussion, parasites will be functionally defined as organisms that live in or on a heterospecific * Present address: Behavioral Ecology Research Group, Department of Biological Sciences, Simon Frazer University, Burnaby, British Columbia, V5A 156 Canada 291

Copyright 0 1YYX by Academic Prew All rights of reproduction in any form reserved 0065 3454/YX $25 00

292

GEORGE A. LOZANO

animal (the host), draw their nutrients primarily from the host, and have the potential to reduce its fitness. Therefore, this definition includes both endoparasites and ectoparasites, but excludes micropredators or animals that use their hosts solely for shelter. Second, both macroparasites ( h e h n t h s , arthropods) and microparasites (viruses, bacteria, protozoa, fungi) are included. Finally, parasites need not be harmful all the time, or even most of the time. Parasites can often coexist with their hosts without causing any measurable deleterious effects, but parasites are also opportunistic, and can quickly increase in numbers and overwhelm a host weakened by other forms of stress, such as malnutrition or reproduction (Walzer and Genta, 1989). To counteract actual or potential fitness losses due to parasitism, animals have evolved a variety of anatomical, physiological, and behavioral adaptations, and parasites have developed an equally impressive array of countermeasures to bypass these defenses (Behnke and Barnard, 1990). In some cases parasites have even evolved ways to manipulate their hosts’ behavior for their own interests (e.g., Bethel and Holmes, 1973; Brassard el af.,1982; Maitland, 1994). The effects of parasites on host behavior include the manipulation of host behavior by parasites (reviewed by Moore and Gotelli, 1990), and host behavioral adaptations for protection against parasitism (reviewed by Hart, 1990; Mdler et al., 1993). Recently, it has become recognized that animal diets may also be shaped by the need for protection from parasites. Foraging behavior evolves primarily to meet the need of a nutritionally adequate diet. However, just as foraging behavior can be affected by predators (e.g., Milinski and Heller, 1978; Krebs, 1980; Sih, 1980; Edwards, 1983; Abrahams and Dill, 1989; Lima and Dill, 1990) and competitors (e.g., Baker et al., 1951; Milinski, 1982; Millikan ef al., 1985), some features of diet selection seem to have evolved to stave off, or reduce parasitism. These adaptations can include the avoidance of foods that are also potential sources of parasitic infection, the use of prophylactic substances, and the consumption of therapeutic substances (Phillips-Conroy, 1986; Lozano, 1991). Self-medication includes the latter two types of responses. Although in this chapter I deal largely with self-medication in the context of feeding, it may also occur under other circumstances, including the use of plants with potentially antibacterial chemicals for nest material (Wimberger, 1984; Clark and Mason, 1985), and the topical application of antifungal and antibacterial compounds (Ehrlich ef al., 1986; Baker, 1996; Gompper and Hoylman, 1993). In this chapter I first incorporate selfmedication into the broader phenomenon, namely, the effects of plant chemicals across several trophic levels, and categorize self-medicating behavior into two basic forms: prophylactic and therapeutic. In the body of the chapter I review in detail current evidence in the published literature

SELF-MEDICATION IN WILD ANIMALS

293

for the occurrence of self-medication in nonhuman vertebrates. Finally, I discuss behavioral mechanisms that may play a role in self-medication, and highlight potential implications for other areas of research.

11. SELF-MEDICATION

The effects of secondary plant metabolites are not always limited to the herbivores that consume them, but may also affect the herbivores’ predators, parasites, and parasitoids. For example, in several herbivorous insects susceptibility to pathogens differs depending on the plant on which the hosts feed (e.g., Hare and Andreadis, 1983; Krischik et al., 1988). Such interactions have long been studied in the general framework of chemical ecology, mostly in insects (reviewed by Duffey, 1980; Price et al., 1980). Nevertheless, animals in other taxa are also able to ingest secondary plant metabolites and accumulate them in their tissues. These compounds can make prey unpalatable to predators (e.g., Brower, 1958; Rothschild, 1972; Hay et al., 1990; Pennings, 1994; Daly e f al., 1994), or less susceptible to parasitoids (e.g., Campbell and Duffey, 1979). Sequestered compounds, specifically carotenoids can also play a role in sexual selection by altering the showiness of secondary sexual ornaments in males (e.g., Kodric-Brown, 1989; Zuk, 1992; Milinski and Bakker, 1990; Hill, 1994), although it is unknown whether these traits are important in sexual selection because they indicate foraging ability or immunocompetence (Endler, 1980; Lozano, 1994). It is therefore clear that plant chemicals can have effects across several trophic levels. The use of secondary plant metabolites by vertebrates for the purpose of self-medication can be viewed as a special case of this broader phenomenon. Janzen (1978) was probably responsible for bringing to the forefront of western scientific inquiry the idea of self-medication in nonhumans. He compiled anecdotal accounts of unusual feeding habits by several species of mammals. For example, just before starting long trips, Indian elephants (Elephas muximus) reportedly feed on Entuda schefferi (Leguminosae). Indian wild boars (Sus scrofu) consume the roots of Boerhavia diffusa (Nyctaginaceae), a plant used in traditional medicine as an anthelminthic. Sumatran rhinoceroses (Didermocerus sumatrensis) have been observed eating copious quantities of the tannin-laden bark of mangroves (Ceriops candoleana, Rhizophoraceae). Janzen pointed out that energy requirements and chemical avoidance were probably not adequate to explain these observations, and raised the possibility that animals use plant secondary metabolites as stimulants, antihelminthics, laxatives, antibiotics, or even as antidotes for previously consumed toxins.

294

GEORGE A. LOZANO

Despite this apparent taxonomic and behavioral diversity, selfmedication can take only two functionally distinct forms, preventive (prophylactic) and therapeutic (Phillips-Conroy, 1986; Lozano, 1991). The two processes yield different predictions and require distinct behavioral mechanisms. By viewing self-medication under a more general framework, these behaviors need not be studied as a series of isolated cases, but rather can be considered in terms of common elements. For example, the consumption of food items for preventative purposes would be related to the risk of parasitism, but not necessarily to the presence of parasites. The biological effects of these medicines may be aimed solely at the infectious stage of the parasite, and could have no effect at all on established parasites. Furthermore, the consumption of medicinal substances may not vary substantially among individuals within a population, but could differ considerably between populations. Lastly, if the risk of parasitism is predictable, seasonally, for example, dietary shifts may be largely genetically determined, and not depend on individual or social learning. This also means that the consumption of prophylactic food items will probably be difficult to demonstrate conclusively, even for a single parasite-host-medicine system, because the consumption of these food items would likely be integrated with the regular diet. On the other hand, in cases of therapeutic self-medication, only sick individuals would be expected to consume medicinal substances. These food items would not be expected to be in the animal’s regular diet, and would be consumed only upon infection. Therapeutic medications would probably be more potent than preventative ones, and consequently would carry a greater risk of negative side effects. Medicinal substances could be aimed at the infection, in which case their biological effect would be directed at parasites already established within the host. Alternatively, the purpose of medicinal substances could be to alleviate discomfort, similar to the use of medicines for the common cold by humans, and have no effect at all on the parasites. In either situation, the ability to self-diagnose, prescribe, seek, and consume the appropriate medicine requires a fairly complex mechanism of individual and/or social learning.

111. PROPHYLACTIC SELF-MEDICATION Studies have not always made clear the distinction between preventative and curative self-medication. As previously indicated, the difference is that therapeutic self-medication is a specific response to a particular situation; that is, the deliberate consumption of medicinal substances by ill individuals. In this section I discuss instances in which secondary plant metabolites seem

SELF-MEDICATION IN WILD ANIMALS

295

to affect parasites or disease, but, so far, there is no evidence suggesting intentionality. The classification of the following behaviors as preventative self-medication is therefore not definitive, as further work may show that these behaviors are also examples of therapeutic self-medication. A. GEOPHACY IN PRIMATES

Geophagy, the deliberate consumption of soil, dirt, or rock, has been observed in several herbivorous and omnivorous mammals (reviewed by Kreulen, 1985). Geophagy may be used to control gut p H (Oates, 1978; Davies and Baille, 1988), to meet nutritional requirements of trace minerals (Davies and Baille, 1988; Johns and Duquette, 1991), to satisfy a specific hunger for sodium (Mahaney et al., 1990), andlor to detoxify secondary plant metabolites (Johns and Duquette, 1991). Recently, it has also been suggested that some primates may use geophagy to combat intestinal problems, particularly diarrhea (Mahaney et al., 1995a,b). Geophagy has been studied in the context of self-medication in Japanese macaques (Macaca fuscata) (Mahaney et al., 1993), rhesus macaques (Macaca rnularta) (Mahaney et al., 1995a), mountain gorillas (Gorilla gorilla) (Mahaney 1993; Mahaney et al., 1995b), and chimpanzees (Pan troglodytes) (Mahaney et al., 1996,1997). Analyses of the soils consumed by these four species have detected at least one of three mineralogically similar clays: halloysite, metahalloysite, and kaolinite, the last of which is the principal ingredient of the commercial antidiarrheal Kaopectate TM (Vermeer and Ferrell, 1985). So far, support for the idea of geophagy as selfmedication is limited to these mineralogical analyses. There have been no studies relating geophagy to the incidence or risk of diarrhea, nor have there been studies on the physiological effects of these clays in nonhumans.

B. STIMULANT USE IN BABOONS Hamilton et al. (1978) classified food items consumed by chacma baboons (Papio ursinus), into four categories: (1) animals, (2) fruits and seeds, ( 3 ) leaves, and (4) “euphorics.” The fourth group consisted of plants that were widely available and consumed consistently, but only in minute quantities. Furthermore, these plants were known to be hallucinogenic and highly toxic to humans, and presumably also to other mammals (Hamilton et al., 1978). These “euphorics” included Croton megalobotrys (Euphorbiaceae), Euphorbia avasmontana (Euphorbiaceae), Datura innoxia (Solanaceae), and D. stramonium. Subsequent authors (Huffman and Seifu, 1989; Wrangham and Goodall, 1989) have cited this study as an example of self-medication; however, aside from labeling these plants as “euphorics,”

296

GEORGE A. LOZANO

Hamilton et al. (1978) did not speculate on their possible role(s). There has been no further work with this system. C. ANTISCHISTOSOMAL DRUGUSEBY BABOONS Phillips-Conroy (1986) examined the diet of baboons along the Awash River Valley, Ethiopia, which is divided by waterfalls into two distinct habitats, with water flow being faster upstream, but slower after the falls. The valley was populated by anubis baboons (Papio anubis) above the falls, and hamadryas baboons (Papio hamadryas) and anubis-hamadryas hybrids below the falls. The risk of schistosomiasis infection varied for these populations because snails (Biomphalaria sp.), the intermediate hosts of Schistosoma spp., were absent upstream from the waterfalls, but were abundant downstream. Finally, although the shrub Balanites aegyptica (Balanitaceae) was common throughout the valley, only downstream from the falls did baboons consume its leaves and fruits. Balanites fruits contain diosgenin, a hormone precursor. Phillips-Conroy (1986) suggested that Balanites is consumed because it hinders the development of schistosomes, but experimental work in schistosome-infected mice showed that ingestion of diosgenin actually increases the number of schistosome eggs in the liver; it enhances the disease (Phillips-Conroy and Knopf, 1986).

FOLIAGE AS NESTMATERIAL D. ANTIBACTERIAL Several bird species place in their nests fresh vegetation that does not constitute part of the nests’ structure. Wimberger (1984) noted that fresh plants probably contain more volatile secondary compounds than does dried vegetation, and he hypothesized that birds use these plants t o repel or even kill ectoparasites. Using data from egg collections of North American and European Falconiformes, and based on the premise that nest reuse leads to increased parasite transmission, Wimberger (1984) showed that Falconiformes that reused their nests in successive years were more likely to use green foliage in their nests, and those that did not were less likely to do so. Clark and Mason (1985) conducted a similar comparison using selected North American passerines and found that cavity nesters were more likely to use green foliage than were open cup nesters (Table I). Clark and Mason (1985) also demonstrated that plant use by starlings (Sturnus vulgaris) was not random, as the plants selected did not simply reflect the availability in the surrounding areas. Furthermore, preferred plants were more effective at reducing the hatching success of lice (Menacanthus sp.) eggs and inhibiting bacterial growth than a random subset of the available vegetation. Subsequently, they showed experimentally that leaves of wild carrot (Daucus carom, Umbelliferae), one of the preferred

297

SELF-MEDICATION IN WILD ANIMALS

TABLE I USEOF GREENPLANTS AS NESTMATERIAL I N RELATION TO NESTREUSE A N D TYPEOF NEST AMONG FALCONIFORMES A N D NORTH AMERICAN PASSERINES. RESPECTIVELY (EXPECTED FREQUENCIES I N PARENTHESES: FROM WIMBERCER. 1984; CLARK A N D MASON.1985.) Use of green vegetation

a) Falconiformes ( n = 48) Reuse nests Build new nests b) North American Passerines (PZ Enclosed nests Open nests

=

Present

Absent

x?

P

22(17.5) 6(10.5)

8(12.5)

8,28


12(4.5)

lg(9.1) 28(36.9)

82(73.1)

137) 9(17.9)

16.4

<0.001

species, significantly reduced the number of fowl mites (Ornithonysus sylviarum) in starling nests (Clark and Mason, 1988). The decrease in mite abundance had no effect on nestling growth, but nestlings from nests with carrot leaves had higher hemoglobin levels than chicks from control nests. Therefore, it seems fairly clear that starlings select nest material with insecticidal and antibacterial properties. However, contrary to what would be expected according to the chemically mediated parasite-protection hypothesis, starlings add green vegetation to their nests only during nest building, and, unlike Clark and Mason (1988), not while eggs or young are in the nest. Also, males that reuse a nest box during one breeding season, whether because the first brood fledged or was lost, collect less foliage than males nesting concurrently but for the first time (Gwinner, 1997). Finally, the hypothesis does not explain why only males add green vegetation to their nests, and first-year males use less fresh vegetation than older males (Clark and Mason, 1985). Several other hypothesis, not necessarily alternative, have been proposed to explain the use of green vegetation in nests. Green foliage may serve to attract females (Fauth et al., 1991; Gwinner, 1997) and actually be a rudimentary bower; it may be used to cover debris and keep the nest clean; it may advertise nest occupancy, or it may prevent egg desiccation. It would be interesting to know whether other species behave similarly, and whether starlings use more green foliage in response to higher levels of parasitism.

E. ANTING A N D FURRUBBING Anting refers to a behavior in which birds rub crushed ants throughout their plumage. Birds also ant by lying on ant nests and letting ants crawl

298

GEORGE A. LOZANO

over their plumage. This behavior occurs in a variety of birds (Potter, 1970) and it has been suggested that anting is used to soothe irritated skin, help with feather maintenance, and prevent or reduce the abundance of skin parasites (Potter, 1970; Clunie, 1976; Ehrlich et al., 1986). Birds also “ant” with other invertebrates, plants, and inanimate objects, such as millipedes (Clunie, 1976), lime fruit (Clayton and Vernon, 1993), and mothballs (Clark et al., 1990), all of which have some antiparasitic properties. Anting has also been observed in mammals (Bagg, 1952; Hauser, 1964; Longino, 1984). An analogous behavior, fur rubbing, occurs in some mammals. Baker (1996) observed capuchin monkeys (Cebus capucinus) in Costa Rica rubbing their fur with the fruits of several species of Citrus (Rutadeae), and with the leaves or stems of the vines Piper rnarginaturn (Piperaceae) and Clematis dioica (Ranunculaceae). These plants have a wide range of bioactive compounds and are used in traditional medicine to treat a variety of ailments. White-nosed coatis (Nasua narica) have been observed coating their bodies with Trattinnickia aspera (Burseraceae) resin. Although information on the medicinal uses of T. aspera is limited, Gompper and Hoylman (1993) suggested this behavior serves a medicinal function. In conclusion, support for the idea that anting and fur rubbing are primarily antiparasitic behaviors is still largely anecdotal; more detailed and experimental studies are presumably forthcoming.

SELF-MEDICATION IV. THERAPEUTIC In contrast to prophylactic self-medication, evidence for therapeutic selfmedication is more compelling, and has attracted considerably more attention. This evidence comes from a single source: chimpanzees at Gombe National Park and the Mahale Mountains, Tanzania, but it is very diverse in nature. Conclusions are based on direct observations of chimpanzees in the wild, fecal analyses, traditional medicine, and biochemistry. Several other plants may be involved, but most work has concentrated on the possible therapeutic use by chimpanzees of three specific plants. A. ASPILIA

The first report of a possible case of therapeutic self-medication was based on several peculiarities of the consumption of leaves of Aspilia pluriseta (Compositae), A . rudis, and A. rnossarnbicencis (Wrangham and Nishida, 1983). Field observations of chimpanzees and subsequent fecal analyses revealed that entire leaves were swallowed without being chewed. Instead, these leaves were taken singly and rolled around the mouth before being

299

SELF-MEDICATION IN WILD ANIMALS

swallowed. It was later suggested that this feeding technique may facilitate the intake of any existing medicinal substances via the buccal mucosa (Newton and Nishida, 1990). At Gombe consumption occurred only during the morning, but at Mahale it occurred all day. Finally, there was no between-individual variation in the tendency to consume Aspilia leaves. Based on these observations and the widespread use of Aspilia in traditional medicine, Wrangham and Nishida (1983) suggested that these leaves are consumed because of their pharmacological effects. However, because of the lack of individual variation and because, at Gombe, consumption occurred only during the morning, Wrangham and Nishida (1983) concluded Aspilia was probably used as a stimulant, rather than as a medicine. It is difficult to draw conclusions about the seasonal variation of Aspilia consumption. At Mahale, the percentage of chimpanzee feces containing Aspilia leaves was highest in January and February (Wrangham and Nishida, 1983). In contrast, at Gombe, the presence of Aspilia leaves in fecal samples was highest during June and July, but behavioral observations indicated that consumption peaked in January, November, and May (Wrangham and Goodall, 1989) (Fig. 1). Further work has shown that

--*Gombe feces I

0

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

FIG. 1 . Seasonal variation of Aspilia consumption by chimpanzees at Mahale and Gombe, based on fecal samples and behavioral observations. (From Wrangham and Nishida, 1983 and Wrangham and Goodall. 1989).

300

GEORGE A. LOZANO

the prevalence of infection by the intestinal nematode Oesophagostomurn stephanostornurn is highest during the rainy season (November to March), but there are no seasonal patterns in the prevalence of two other intestinal nematodes: Trichuris trichiura and Stongyloides fuelleborni (Huffman et al., 1997). Several secondary metabolites have been obtained from other Aspilia species (Mabry et al., 1977; Bohlmann et al., 1984; Ganzer et al., 1992). Methanol extracts of A . rnossarnbicensisleaves have limited biological activity against a variety of insects, herbs, and fungi (Ohigashi et al., 1991a). In contrast, chloroform extracts of dried leaves yielded thiarubrine-A, a naturally occurring phototoxic compound also found in other species of Compositae (Rodriguez et al., 1985). In the presence of UV-A light, thiarubrine-A is toxic to several bacteria and viruses, and at least one freeliving nematode, but its toxicity decreases in the absence of light. Under acidic or alkaline conditions thiarubrine-A readily changes into thiopheneA, which is toxic only in the presence of UV-A light (Towers et al., 1985; Hudson et al., 1986, Table 11). Page et al. (1992) found thiarubrine-A in the roots of A . mossarnbicensis, but were not able to isolate it from samples of either fresh or dried leaves. They did, however, isolate two diterpenes, kaurenoic acid and grandiflorenic acid from dried leaves, and showed that these compounds stimulate contracT A B L E I1

IN

VITRO TOXICITY OF THIARUBRINE-A AND

Organism

Caenorhabditis riegans Saccharomyces cerevisinc, Candida albicans Staphylococcrrs nlbrrs Bacillus subtilis Streptococcus fecalis Escherichia coli Pseudotriotias ,floirrescerzs Mycobacrrririm phlei Murine cytomegalovirus Sindbis virus T4 bacteriophage M 13 bacteriophage

Thiarubrine-A Light Dark

++ ++ ++ ++ ++ ++ ++ -

++ ++ ++

+

-

1 = Towers et al. (1985) 2 = Hudson er a/. (1986) " ++ = highly toxic. + = weakly toxic,

THIOPHENE-A" Thiophene-A Light Dark

+ ++ ++ -

++ -

+ -

+ ~

~

- =

no effect. nt

nt nt

++ ++

nt nt ~

-

nt nt

nt nt

nt nt nt nt nt nt

nt nt nt nt nt nt

+

=

~

not tested.

Ref. 1 1 1 1 1 1 1 1 I 2 2 2 2

SELF-MEDICATION IN WILD ANIMALS

301

tions of a guinea pig uterus in vitro. Observations of chimpanzees revealed that at Gombe more females than males selected Aspilia leaves, and the number of days in which Aspilia leaves were consumed was also significantly higher for females (Wrangham and Goodall, 1989), which led Page et af. (1992) to suggest that pregnant chimpanzees consume the leaves of A . mossambicensis to induce labor. This hypothesis would predict increases in Aspilia consumption by pregnant females as they approach their due dates, and perhaps even by females ill suited to carry their fetuses to term. However, there is no information on Aspilia use by pregnant females, nor any evidence to indicate that Aspilia induces labor in vivo. This idea has not received further consideration. Conclusions based on the chemical analyses must be considered tentative for at least two reasons. First, it is difficult to build a case for self-medication based on the ingestion of thiarubrine-A, the biological activity of which is markedly decreased, or completely absent, without light. Given its properties, thiarubrine-A seems an unlikely medicinal substance, except if used as an external antibiotic (see Ehrlich et al., 1986; Gompper and Hoylman, 1993; Baker, 1996). Second, two subsequent studies (Page et af., 1992; Huffman et al., 1996) have failed to detect thiarubrine-A in leaf samples of A . mossambicensis, as first reported by Rodriguez et al. (1985). If indeed only the roots of A . mossambicensis contain thiarubrine-A, then its biological activity is irrelevant to leaf-eating chimpanzees. However, it has recently been suggested that the leaves of Aspilia sp. and other suspected medicinal plants may be consumed not because of their chemical properties, but rather because of their characteristically rough surfaces, which may aid in the mechanical removal of intestinal parasites (Messner and Wrangham, 1996; Huffman et al., 1996).

B.

VERNONIA

The recognition of Vernonia amygdafina (Compositae) as a possible chimpanzee medicinal plant was also the result of detailed field observations (Huffman and Seifu, 1989). An adult female, dubbed CH, was observed during 2 consecutive days, for a total of about 11 hr. For 35 min during the afternoon of the first day CH foraged almost exclusively on the branches of V. amygdafina, a plant that was not consumed by other members of her group. When feeding on Vernonia, she chewed the young branches, sucked and swallowed the pith juice, and discarded the remaining fibers. During the afternoon of the first day and the morning of the second day, CH spent an unusually long time lying down and very little time foraging; she seemed to have trouble defecting, and her urine seemed unusually dark. Her behavior and urine color returned to normal in the afternoon of the second

302

GEORGE A. LOZANO

day. Interestingly, CH had previously been observed swallowing leaves of another plant, Lippia plicata (Verbenaceae), presumably also for medicinal purposes (Takasaki and Hunt, 1987). Huffman et al. (1993) observed another adult female, dubbed FT, for a total of about 5 hr, over a period of 2 days. During this time she fed on clay from a termite mound and on at least four species of plants, among them V.amygdahza. Like CH, FT did not consume the leaves of V. amygdalina, but instead chewed, sucked, and discarded young branches. A fecal sample obtained during the afternoon of the first day was yellowish and liquid, and contained 130 eggs of the intestinal nematode Ternidens sp. per gram of feces. A second stool sample, obtained in the morning of the next day, was solid, and contained only 15 eggs per gram of feces. Huffman et al. (1993) also presented data on normal infestation levels, based on repeated fecal sampling of seven other individuals. However, these data were not detailed enough to determine whether the decrease in Ternidens eggs in FT was within the normal range of daily variation, in the absence of Vernonia or clay consumption. Several factors, including herbivory, can affect the production and distribution of secondary metabolites within individual plants (Rhoades, 1979; Karban and Myers, 1989), so care must be taken to ensure that the leaves used for analysis are a suitable representation of those consumed (e.g., Huffman et al., 1996). Representative samples collected from the actual Vernonia plants consumed by FT showed that, contrary to expectations, the concentration of two biologically active compounds, vernonioside B1 and vernodalin, was higher in young leaves than in young stems (Huffman et al., 1993; Ohigashi et al., 1994). Several other secondary metabolites have been extracted from V. amygdalina (Kupchan et al., 1969; Ganjian et al., 1983; Ohigashi et al., 1991a,b; Jisaka et al., 1992a, 1993a), and, as expected, the biological activity of these compounds is diverse. Extracts from V. amygdalina deter insect herbivory (Ganjian et al., 1983), are toxic to schistosomes (Jisaka et al., 1992b; Ohigashi et al., 1994), and have antitumoral (Kupchan et al., 1969), antibacterial (Jisaka et al., 1993b), and antioxidant (Igile et al., 1994) properties. Like Aspilia, V. amygdalina is used widely in Africa by humans as a medicinal plant for a variety of ailments.

C. RUBIA Wrangham (1995) examined the relationship between a parasitic tapeworm infection and the peculiar habit of leaf swallowing by chimpanzees. Fecal droppings containing whole leaves of Aneilema aequinoctiafe (Commelinaceae) and Rubia cordifolia (Rubiaceae) were found sporadically throughout 6 years. Tapeworm fragments were detected in these droppings

SELF-MEDICATION IN WILD ANIMALS

303

on 14 occasions, spanning 7 months. During this time, the frequency of droppings containing tapeworm fragments and whole leaves was significantly greater than expected. Wrangham (1995) concluded that heavily infected chimpanzees purposely swallow whole leaves, which cause the shedding of tapeworm proglottides. However, Wrangham conceded that, because proglottid shedding is part of tapeworm’s normal life cycle, leaf swallowing is not necessarily an effective method of tapeworm control. Another study (Messner and Wrangham, 1996), also involving R. cordifolia, is the only one so far in which the biological activity of a presumed medicinal plant has been compared with that of other plants comprising the regular diet of chimpanzees. Messner and Wrangham compared R. cordifolia to six other plants, but found no differences in their toxicity to free-living adults or larvae of Strongyloides spp.

V. SKEPTICISM

Although the idea of therapeutic self-medication in animals has been discussed for over a decade, only in the semipopular literature have we seen a healthy dose of skepticism (Sapolsky 1994, his pun). Sapolsky raised three main concerns: the absence of controls with which to compare the biological effects of these presumed medicinal plants, the lack of and need for studies in vivo, and the absence of clearly demonstrated behavioral mechanisms by which therapeutic self-medication can arise and be maintained in a population. I deal with the first two points here, and with behavioral mechanisms in the subsequent section. Currently, there is adequate evidence that some plants are consumed under unusual circumstances, and that the leaves or roots of these plants have secondary compounds with uterotonic, antiviral, antibacterial, andlor anthelminthic properties. Little else is required if the goal is merely to identify bioactive compounds present in chimpanzees’ diets. However, if we consider that all plants have secondary metabolites, and that the main role of these chemicals is protection from herbivores, fungi, and bacteria, it is not particularly surprising to find that, for any given plant, even if selected at random, some of these secondary metabolites will be biologically active. The presence of bioactive secondary metabolites in suspected medicinal plants is therefore not conclusive evidence. A t best, it can be concluded only that, when ill, some chimpanzees deviate from their normal diet. To demonstrate that plants are consumed to deal with specific diseases, we need to know the ailment affecting an individual, and show not only that the plant parts consumed alleviate this condition, but that their effect is greater than that of plants that make up a healthy individual’s regular diet.

304

GEORGE A. LOZANO

Alternatively, it is probably difficult to determine the exact relationships between specific diseases and their corresponding medicines, so, when ill, individuals may simply choose plants containing a wide spectrum of biological effects, akin to taking a general antibiotic. In these cases, these plants would not be necessarily effective against the disease, but they would be expected to contain more and/or stronger biologically active compounds than plants normally consumed. Third, therapeutic medicines may be consumed to alleviate the symptoms of the disease, and have no effects at all on the pathogen itself. In such cases medicinal plants would be expected to contain analgesics and other compounds that affect only the host, and not the parasite. Whether an animal consumes medicinal plants as general antibiotics, or to deal with specific diseases, or merely the symptoms of disease, comparisons with other plants comprising the animal’s regular diet are needed before firm conclusions can be drawn. There has been only one study in which such controls have been used. Messner and Wrangham (1996) found no differences in the biological activity against Strongyloides spp. between methanol extracts of R. cordifolia and six other plants regularly eaten by chimpanzees. Messner and Wrangham pointed out that these results do not necessarily mean that R. cordifolia does not affect intestinal nematodes because (1) the extraction method may have failed to obtain all bioactive compounds, (2) the nematodes used in this test were not the parasitic stage, but rather free-living adults and larvae, and (3) in vitro tests are a poor replacement for the complex interactions that may occur in vivo. Unfortunately, these same caveats would have also been valid had Messner and Wrangham (1996) found significant differences between the biological activity of R. cordifolia and the six other plants. However, Messner and Wrangham (1996) did raise an important point: experimental tests in vivo are needed. Understandably, in vivo trials may not be practical or ethical in wild chimpanzees, so they must take a back seat to observational and to phytochemical studies. Furthermore, chimpanzees at Gombe and Mahale have been the focus of ongoing research for several decades, so it would be undesirable to carry out invasive experiments with these populations. Nevertheless, such studies are necessary if we wish to understand the effects of these plants, and could perhaps be conducted with chimpanzee populations elsewhere, captive chimpanzees, or other primates. VI. BEHAVIORAL MECHANISMS

Therapeutic self-medication requires fairly complex and interesting behavioral mechanisms of food selection. However, as Sapolsky (1994)

SELF-MEDICATION I N WILD ANIMALS

305

pointed out, this aspect has received little attention in the self-medication literature. In this section I discuss food selection mechanisms that may be involved in therapeutic self-medication, dealing first with individual learning and then with social learning. The following discussion is not meant to be a critical review of the literature on the mechanisms of food selection, and is based largely on several comprehensive reviews (Rozin and Kalat, 1971: Galef, 1976, 1996; Rozin, 1976; Bandura, 1977). A. INDIVIDUAL LEARNING If therapeutic self-medication is learned individually, a series of steps must take place. First, upon infection by a parasite, or when the infection reaches a particularly uncomfortable level, the host must begin sampling unfamiliar food items, and in some cases overcome their natural aversion to new foods and bitter-tasting plants (Garcia and Hankins, 1975). The infected animal must then chance upon a medicinal plant and fortuitously consume it in sufficient quantities for the plant to be effective against the offending parasite. Upon recovery, which may occur many hours after the medicinal plant was consumed, the animal must return to its regular diet. Several relevant mechanisms have been demonstrated experimentally in rats (Rafrus norvegicus), apparently the preferred species for experimental work on the mechanisms of food selection. Richter (1943) showed that rats faced with a limited number of single-nutrient food items were able to select a nutritionally adequate diet. Furthermore, rats with deficiencies of specific nutrients were able to obtain these nutrients by altering their diets. In theory, the ability to obtain a balanced diet may be the result of specific hungers, under strict genetic control, without the flexibility of learned behavior. For every single nutrient it requires, an animal could have the ability to sense physiological deficiencies, and recognize its presence in food. The animal would need the physiological mechanisms to identify each nutrient individually, presumably by taste or smell, and to monitor constantly for deficiencies. This would mean a separate monitoring and identification system for each essential amino acid, vitamin, and mineral. Clearly, such a system would seldom be necessary or particularly useful. Specific hungers do exist, but they are limited to extremely important nutrients. For example, carnivores need not be concerned with individual nutrients, as each prey item provides them with a full range of essential nutrients in an adequate balance. Domestic chickens (Gullus gallus) have a specific hunger for water that includes the ability to taste it, but not identify it visually (Hunt and Smith, 1967). It is therefore possible to have dehydrated newly hatched chicks walking through water and being completely unaware of the obvious solution to their problem. They soon learn

306

GEORGE A. LOZANO

to identify it visually, but only after having pecked at it and tasted it. The adult blowfly (Phormia regina) has specific hungers for sugar, water, and salt, and its feeding response is under direct control of separate internal and external chemoreceptors for each of these nutrients (Dethier, 1969). Most mammals have a specific hunger for water (Rozin, 1976). Among primates, sodium hunger has been shown in humans (Beauchamp et al., 1990) and baboons (Denton et al., 1993). Given the large number and unusual nature of chemicals involved, it is doubtful that specific hungers play a role in therapeutic self-medication. Whereas the rule of thumb “when suffering from dehydration, drink plenty of water” could be solely under genetic control, the directive “when suffering from malaria, drink water from under a cinchona tree, or better yet, chew on the tree’s bark” is far more complex and more likely to be a learned response. In the absence of specific hungers, diet selection must be the result of learned preferences for suitable diets, or learned aversions for inadequate diets. Rozin (1967) observed that the behavior of rats toward their regular, palatable, but nutritionally deficient diet was similar to their behavior toward highly unpalatable diets. In both cases rats approached the food tray tentatively, spilled some food, and then moved away and chewed on some inedible object. These rats were quick to consume any new diet, regardless of whether it was nutritionally adequate. These observations showed that diet changes in rats are not the result of learned preferences for new or nutritionally adequate diets, but rather the result of a learned aversion for the initial, nutritionally deficient diet. Whether diet changes are the result of aversions or preferences, several problems arise when attempting to apply diet selection mechanisms to therapeutic self-medication. Self-medication requires that animals consume unusual food items temporarily and maybe exclusively, and then revert to their regular diets. This process does not entail a permanent preference for the alternative diet, or a permanent aversion to the regular diet. It may be possible to explain self-medication in terms of a dual aversion, first to the regular diet, and then to the medicine. However, this would be possible only if the initial aversion to the regular diet is strong enough to cause the initial shift, yet mild enough to be subsequently forgotten. Another problem for a self-medicating animal is learning to associate its recovery with its diet over the past several hours, and not with other events that may have occurred concurrently. Experiments in rats have shown that aversions do not develop to the location or the type of food container, but are limited to the nutrient-deficient food itself. Garcia and Koelling (1966) exposed rats to taste, sound, and light stimuli, paired with either electrical shocks or poisoning, via injection or radiation. Poisoned rats learned to

SELF-MEDICATION I N WILD ANIMALS

307

avoid the taste, but not the sound or light, whereas shocked rats developed an aversion to the light and sound, but not the taste. This experiment showed that certain associations between stimuli are learned more easily than others. Specifically, visceral responses are more likely to be associated with food consumption, which suggests that intestinal ailments would be more likely associated with medicines consumed orally, and external ailments with topical medicines. A self-medicating animal must learn not only to associate its recovery with food, but also to determine which food, out of the several items that may have been consumed, is responsible for its recovery. While trying to find an adequate diet, rats do not sample alternative foods randomly. Instead, their sampling pattern facilitates the possibility of associating recovery with a specific item. Feeding bouts are temporally separated and include only one new food source, and only a few new foods are sampled each day (Rozin, 1969). So far, no studies have dealt specifically with the food sampling behavior of sick chimpanzees.

B. SOCIAL LEARNING Social interactions play an important role in every aspect of chimpanzee behavior; hence much of their knowledge concerning ways to interact with their environment does not necessarily come from individual experience. Food preferences may be influenced by the food choices of conspecifics, so self-medication may not be learned de n o w by every individual. Although the effects of social learning on self-medicating chimpanzees have not been studied yet, several potentially relevant mechanisms have been demonstrated experimentally in other species. For example, in rats, protein deficiency increases the effect of social learning on diet preferences (Galef et uf., 1991). These results suggest that sick animals in poor condition may be more likely to alter their diets. It has also been demonstrated that rats are more likely to learn the unfamiliar, rather than the familiar or usual diet of their demonstrators (Galef, 1993). In spotted hyenas (Crocutu crocutu), individually learned food aversions can be attenuated and even eliminated by the observation of conspecifics feeding on the avoided food (Yoerg, 1991). In red-winged blackbirds (Agefuiusphoeniceus), aversions can develop from observing conspecifics becoming ill after consuming a food item (Mason and Reidinger, 1982), which shows that blackbirds learn to associate visual cues of illness in conspecifics with particular foods. However, there have not been any studies demonstrating the reverse: the ability to associate the recovery of a sick conspecific with its consumption of a specific food item.

308

GEORGE A. LOZANO

VII. IMPLICATIONS A. CONSERVATION ECOLOGY The many ways in which animals interact with their environment are seldom readily apparent. Self-medication in wild animals may be one such relationship that we are only now beginning to recognize, much less understand. This lack of knowledge further demonstrates that we do not have the ability to reconstruct natural ecosystems; therefore, conservation ecology requires the protection of entire communities, with all their species and interrelationships intact (Clayton and Wolfe, 1993). Captive breeding programs can be successful at preserving individual species, but do not preserve the relationships of an animal with its natural environment. Hence, the preservation of species should be considered as an important fail-safe option, but only part of more holistic conservation ecology strategies. The existence of self-medication may also affect the ease with which animals can be reintroduced to the wild, especially in cases for which knowledge about self-medication is culturally transmitted. Depending on the extent to which self-medication and other parasite avoidance behaviors are culturally transmitted, naive animals being returned to their natural environment may be subjected to unusually high parasite loads. The negative effects of parasites may be further exacerbated in host populations with heavily fragmented habitats, a factor that should be considered in designing biological reserves (Loye and Carroll, 1995).

THEORY B. FORAGING Optimal foraging models were initially based on the assumption that the primary goal of foragers was to maximize net energy or protein intake (Stephens and Krebs, 1986). Other factors, such as the risk of predation (e.g., Milinski and Heller, 1Y78; Sih, 1980; Edwards, 1983; Abrahams and Dill, 1989; Lima and Dill, 1990), and the effects of intraspecific (e.g., Baker et al., 1981; Milinski, 1982) and interspecific (e.g., Millikan el al., 1985) competition, have been subsequently incorporated into diet choice models, and increased their predictive powers. The effects of parasitic infections on foraging behavior have also been examined (e.g., Crowden and Broom, 1980; Milinski, 1984; Giles, 1987). So far, however, diet choice has been largely ignored as a way in which potential hosts could actively reduce parasitism. Diet choice has also evolved under the selection pressures brought about by parasites. It is therefore reasonable to expect that optimal diets are not only nutritionally and energetically adequate, but also take into account

SELF-MEDICATION IN WILD ANIMALS

309

the potentially detrimental effects of parasites. Hosts could alter their diets to counteract the risks of parasitism by (1) avoiding food items that are common sources of parasites, (2) selectively consuming certain food items to decrease their susceptibility to parasites, and (3) acitvely consuming foods with antiparasitic properties upon infection (Lozano, 1991). As evidence of self-medication continues to accumulate, future diet choice models must consider the effects of parasitism.

c.

BEHAVIORAL MECHANISMS OF FOOD SELECTION

Diet preferences evolve under many constraints, including parasitism, so it is easy to envision that plants with antiparasitic properties may become part of an animal’s regular diet. In contrast, therapeutic self-mediation requires a sick animal to deviate away from its regular diet, and seek and consume medicinal substances. It requires intentionality, and is, by necessity, a learned behavior. It is sometimes difficult for even welldocumented phenomena to be generally accepted without clear mechanisms by which they can occur. So far, self-medication cannot .be explained in terms of experimentally demonstrated food selection mechanisms, so it may prove to be an interesting challenge, if it is to be demonstrated conclusively. D. HUMAN MEDICINE Although estimates vary, it is generally agreed that a large proportion of our current medicinal drugs are derived from plants (Fansworth ef al., 1985;Balandrin et al., 1985;Caldecott, 1987; McKenna, 1996). The resources are not available to sample every species, so the identification of plants with potential medicinal uses is a major impediment in the discovery and development of new medicines. It has been repeatedly stated that the study of self-medication in nonhuman animals may lead to the discovery of new medicinal compounds (Cowen, 1990; Newton, 1991; Clayton and Wolfe, 1993; Rodriguez and Wrangham, 1993; Sapolsky, 1994). However, this is not supported by the cases studied so far. As noted earlier, one important reason to suspect that chimpanzees consume certain plants for medicinal purposes is that these plants are also used in human traditional medicine. So, in essence, it was already known that these plants may have medicinal properties. It is possible that further research will indeed yield new medicines. However, the search for new pharmaceuticals cannot be considered the primary goal of this line of research, for we would probably fare much better by exploring plants used in traditional medicine (e.g., Johns, 1990; Johns and Chapman, 1995; Wagner, 1995) instead of plants used by self-medicating chimpanzees.

310

GEORGE A. LOZANO

VIII. SUMMARY A N D CONCLUSIONS Parasites present a ubiquitous selective force that has led to the evolution of a vast array of behavioral adaptations. The need to avoid and reduce parasites can affect foraging patterns and diet choice, and conceivably lead to self-medicating behavior. Self-medication can be viewed as a specific case of the more widespread phenomenon of chemical interactions across trophic levels. Despite the many apparently disparate examples suggestive of self-medication, it can take only two functionally distinct forms: preventative and therapeutic. These two processes require separate mechanisms, and yield different and explicit predictions. By viewing it in a more general framework, self-medication can be studied in terms of common elements, instead of isolated examples. Rodriguez and Wrangham (1993) proposed the term “zoopharmacognosy” to describe the scientific study of the use of plants by wild animals for the prevention and treatment of disease. Current research on therapeutic self-medication is still solely limited to chimpanzees at Mahale and Gombe, but work could be carried out with other populations or other taxa. The multidisciplinary nature of this relatively new field means that problems can be tackled from many different angles, and many avenues of research are still open. Contributions are possible from ethologists, biochemists, parasitologists, pharmacologists, behavioral ecologists, immunologists, psychologists, and statisticians, whether working directly in the field, or simply being aware of the possibility of self-medication in wild animals when conducting other lines of research. Finding out that baboons indulge in the recreational use of pharmaceuticals or that chimpanzees practice a primitive form of medicine may challenge some individuals’ convictions regarding the uniqueness of humans. Understandably, public interest is high, and discussions on self-medication have not been limited to academic media (Bower, 1986; Cowen, 1990; Sears, 1990; Gibbons, 1992; Strier, 1993; Sapolsky, 1994). O n the other hand, most scientists would probably consider such findings extraordinary, but not necessarily disturbing. Scientific interest, therefore, results from more than a mere fascination with newly discovered behaviors; as noted earlier, the study of self-medication in wild animals may have implications for a variety of related fields. Although several recent synopses (Wrangham and Goodall, 1989; Newton, 1991; Huffman, 1993; Rodriguez and Wrangham, 1993; Huffman and Wrangham, 1994) have presented the evidence for selfmedication as being fairly conclusive, I must conclude that the evidence for therapeutic self-medication in nonhumans is still only suggestive. Nevertheless, the possibility of prophylactic or therapeutic self-medication in nonhumans remains a fascinating prospect, and is certainly a fertile ground for further innovative research.

SELF-MEDICATION IN WILD ANIMALS

311

Acknowledgments I thank C-L. Adams. D. Kristan, R. E. Lemon, L. Lefebvre, M. Milinski, A. P. Moller, J. Mountjoy, M. Sclafani. P. J. B. Slater, and M. Zuk for taking the time to discuss ideas with me, guiding me toward relevant literature, and forcing me to find better ways to express myself. Financial support was provided by FCAR.

References Abrahams, M. V., and Dill, L. (1989). A determination of the energetic equivalence of the risk of predation. Ecology 70, 999-1007. Bagg. A. M. (1952). Anting not exclusively an avian trait. J. Mammal. 33, 243. Baker, M. (1996). Fur rubbing: Use of medicinal plants by capuchin monkeys (Cebus capucinus). Am. J. Primatol. 38, 362-370. Baker, M. C., Belcher, C. S., Deutsh, L. C., Sherman, G. L., and Thompson, D. B. (1981). Foraging success in junco flocks and the effect of social hierarchy. Anim. Behav. 29, 137- 142. Balandrin, M. F., Klocke. J. A,, Wurtele, E. S., and Bollinger, W. H. (1985). Natural plant chemicals: Sources of industrial and medicinal materials. Science 228, 1154-1 160. Bandura. A. (1977). “Social Learning Theory.” Prentice-Hall, Englewood Cliffs, NJ. Beauchamp. G. K., Bertino, M., Burke, D., and Engelman, K. (1990). Experimental sodium depletion and salt taste in normal human volunteers. Am. J. Clin. Nutr. 51, 881-889. Behnke, J. M., and Barnard, C. J. (1990). Coevolution of parasites and hosts: Host-parasite arms races and their consequences. In “Parasites: Immunity and Pathology” (J. M. Behnke, ed.), pp. 1-22. Taylor & Francis, London. Bethel, W. M.. and Holmes. J. C. (1973). Altered evasive behavior and responses to light in amphipods harboring acanthocephalan cystacanths. J. Parasitol. 59, 945-956. Bohlmann. F., Gerke, T.. Jakupovic, J., Borthakur, N., King, R. M., and Robinson, H. (1984). Diterpene lactones and other constituents from Wedelia and Aspilia species. Phytochemistry 23, 1673-1676. Bower, B. (1986). Herbal medicine: R, for chimps? Sci. News 129,38-39. Brassard, P., Rau, M. E.. and Curtis, M. A. (1982). Parasite-induced susceptibility to predation in diplostomiasis. Parasitology 85, 495-501. Brower. J. V. (1958). Experimental studies of mimicry in some North American butterflies. I: The Monarch, Danausplexippus, and Viceroy, Limenitis archippus archippus. Evolution (Lawrence, Kans.) 12, 32-47. Caldecott, J. (1987). Medicine and the fate of tropical forests. Br. Med. J. 295, 229-230. Campbell, B. C.. and Duffey, S. S. (1979). Tomatine and parasitic wasps: Potential incompatibility of plant antibiosis with biological control. Science 205, 700-702. Cannon, W. B. (1935). Stresses and strains of homeostasis. Am. J. Med. Sci. 89, 1-14. Clark, C. C., Clark. L., and Clark, L. (1990). “Anting” behavior by common grackles and European starlings. Wilson Bull. 102, 167-169. Clark, L., and Mason, J. R. (1985). Use of nest material as insecticidal and anti-pathogenic agents by the European starling. Uecologia 67, 169-176. Clark. L., and Mason, J. R. (1988). Effect of biologically active plants used as nest material and the derived benefit to starling nestlings. Uecologia 77, 174-180. Clayton, D. H.. and Vernon, J. G. (1993). Common grackle anting with lime fruit and its effect on ectoparasites. Auk 110, 951-952.

312

GEORGE A. LOZANO

Clayton, D. H., and Wolfe, N. D. (1993). The adaptive significance of self-medication. Trends Ecol. Evol. 8, 60-63. Clunie, F. (1976). Jungle mynah “anting” with a millipede. Notornis 23, 77. Cowen. R. (1990). Medicine on the wild side: Animals may rely on a natural pharmacy. Sci. News 138,280-282. Crowden. A. E., and Broom, D. M. (1980). Effects of the eyefluke. Diplostomum spafhaceum, on the behaviour of dace (Letrciscits leuciscus). A n i m Behav. 28, 287-294. Daly. J. W.. Garraffo, H. M.. Spande, T. F., Jaramillo. C., and Rand, A. S. (1994). Dietary source for skin alkaloids of poison frogs (Dendrobatidae)? J . Cheni. Ecol. 20, 943-955. Davies, A. G.. and Baillie, I. C. (1988). Soil-eating by red leaf monkeys (PrcJsbytisruhicundu) in Sabah, Northern Borneo. BioTropica 20,252-258. Denton. D. A,. Eichberg, J. W.. Shade. R.. and Weisinger, R. S. (1993). Sodium appetite in response to sodium deficiency in babbons. Am. J . Physiol. 264, R539-RS43. Dethier, V. G. (1969). Feeding behavior of the bowfly. A d v . Study Behav. 2, 111-266. Duffey, S. S. (1980). Sequestration of plant natural products by insects. Annu. Rev. Entoviol. 25,447-477. Edwards, J. (1983). Diet shifts in moose due to predator avoidance. Oecologia 60, 185-189. Ehrlich, P. R.. Dobkin. D. S.. and Wheye, D. (1986). The adaptive significance of anting. Aitk 103, 835. Endler, J. A. (1980). Natural selection on color patterns in Poecilia reticulafa. Evoliifion (Lawrence, Kans.) 34, 76-91. Fansworth, N. R.. Akerele, O., Bingel, A. S.. Soejarto, D. D., and Guo, Z . G. (1985). Medicinal plants in therapy. Bull. W.H.O.63, 83-97. Fauth, P. T., Krementz. D. G., and Hines, J. E. (1991). Ectoparasitism and the role of green nesting material in the European starling. Oecologia 88,22-29. Galef, B. G., Jr. (1976). Social transmission of acquired behavior: A discussion of tradition and social learning in vertebrates. A h . Sritdy Behav. 6, 77-100. Galef. B. G., Jr. (1093). Functions of social learning about food: A causal analysis of effects of diet novelty on diet preference. Anini. Behav. 46, 257-265. Galef, B. G.. Jr. (1996). Social enhancement of food preferences in Norway rats: A Brief review. In “Social Learning in Animals: The Roots of Culture” (C. M. Heyes and B. G. Galef. Jr., eds.), pp. 49-64. Academic Press. Toronto. Galef. B. G.. Jr.. Beck, M., and Whiskin. E. E. (IYYl). Protein deficiency magnifies social influence on the food choices of Norway rats (Rartits norvegicus). J . Conip. Psycho/. 105,ss-59. Ganjian, 1.. Kubo. I., and Fludzinski, P. (1983). Insect antifeedant elemanolide lactones from Vemonia amygtlalinu. Phytocheinistry 22, 252552526, Ganzer. U., Jakupovic. J., Bohlmann, F., and King. R. M. (1992). A highly oxygenated germacrane from Aspilia eenii and constituents from other Aspilia species. Phytochemistry 31, 209-212. Garcia, J., and Hankins. W. G. (1975). The evolution of bitter and the acquisition of toxiphobia. In “Olfaction and Taste: V ” (D. A. Denton and J. P. Coghlan. eds.). pp. 39-45. Academic Press. New York. Garcia, J.. and Koelling. R. A. (1966). Relation of cue to consequence in avoidance learning. Psychon. Sci. 4, 123-124. Gibbons, A. (1992). Plants of the apes. Science 255,921. Giles, N. (1987). Predation risk and reduced foraging activity in fish: Experiments with parasitized and non-parasitized three-spined sticklebacks, Gasterosteus aculeafirs L. J. Fish B i d . 31,37-44.

SELF-MEDICATION IN WILD ANIMALS

313

Gompper, M. E., and Holyman, A. M. (1993). Grooming with Trattinnickia resin: Possible pharmaceutical plant use by coatis in Panama. J. Trop. Ecol. 9, 533-540. Gwinner. H. (1997). The function of green plants in nests of European starlings (Sturnus vulgaris). Behaviour 134, 337-351. Hamilton, W. J.. 111. Buskirk, R. E.. and Buskirk. W. H. (1978). Omnivory and utilization of food resources by chacma baboons, Pupio ursinus. Am. Nut. 1l2, 911-924. Hare, J. D.. and Andreadis, T. G. (1983). Variation in the susceptibility of Leptinotarsa decemlineata (Coleoptera: Chrysomelidae) when reared on different host plants to the fungal pathogen, Beauveria bassiana in the field and laboratory. Environ. Entomol. 12, 1891-1896. Hart, B. L. (1990). Behavioral adaptations to pathogens and parasites: Five strategies. Neitrosci. Biobehav. Rev. 14, 273-294. Hauser, D. C. (1964). Anting by gray squirrels. J. Mammal. 45, 136-138. Hay, M. E., Duffy, J. E.. Paul, V. J., Renaud, P. E., and Fenical, W. (1990). Specialist herbivores reduce their susceptibility to predation by feeding on the chemically defended seaweed A vrainvillea longicaulis. Limnol. Oceanogr. 35, 1734-1743. Hill. G. E. (1994). Geographic variation in male ornamentation and female mate preference in the house finch: A comparative test of models of sexual selection. Behav. Ecol. 5,64-73. Hudson, J. B., Graham, E. A,, Fong, R.. Finlayson, A. J., and Towers, G. H. N. (1986). Antiviral properties of thiarubrine-A, a naturally occurring polyine. Planta Med., 51-54. Huffman, M. A. (1993). An investigation of the use of medicinal plants by chimpanzees: Current status and future prospects. Primate Res. 9, 179-187. Huffman. M. A,, and Seifu, M. (1989). Observations on the illness and consumption of a possibly medicinal plant Vernonia amygdalina (Del.), by a wild chimpanzee in the Mahale Mountains National Park. Tanzania. Primates 30, 51-63. Huffman. M. A,. and Wrangham, R. W. (1994). Diversity of medicinal plant use by chimpanzees in the wild. I n “Chimpanzee Cultures” (R. W. Wrangham. W. C. McGrew, F. B. De Waal. and P. G. Heltne. eds.), pp. 129-148. Harvard University Press, Cambridge, MA. Huffman, M. A,, Gotoh. S., Izutsu, D., Koshimizu, K., and Kalunde, M. S. (1993). Further observations on the use of the medicinal plant Vernoniu umygdalina (Del.), by a wild chimpanzee, its possible effect on parasite load, and its phytochemistry. Afr. Stud. Monogr. 14,227-240. Huffman, M. A,, Gotoh, S., Turner, L. A,, Hamai, M., and Yoshida. K. (1997). Seasonal trends in intestinal nematode infection and medicinal plant use among chimpanzees in the Mahale Mountains, Tanzania. Primates 38, 111-125. Huffman, M. A,, Page, J. E., Sukhdeo, M. V. K., Gotoh, S., Kalunde, M. S., Chandrasiri, T., and Towers, G. H. N. (1996). Leaf-swallowing by chimpanzees: A behavioral adaptation for the control of strongyle nematode infections. In/. J . Primatol. 17, 475-503. Hunt, G. L., and Smith, W. J. (1967). Pecking and initial drinking responses in young domestic fowl. J. Comp. Physiol. Psychol. 64, 230-236. Igile, G. 0..Oleszek, W., Jurzysta, M., Burda, S., Fafunso, M.. and Fasanmade, A. A. (1994). Flavonoids from Vernnnia amygdalinu and their antioxidant activities. J. Agric. Food Chem. 42,2445-2448. Janzen. D. H. (1978). Complications in interpreting the chemical defenses of trees against tropical arboreal plant-eating vertebrates. I n “The Ecology of Arboreal Foliovores” (G. G. Montgomerie, ed.), pp. 73-84. Smithsonian Institution Press, Washington, DC. Jisaka, M., Ohigashi, H., and Takagaki, T., Nozaki, H., Tada, T. Hirota, M., hie, R., Huffman, M. A., Nishida, T.. Kaji. M., and Koshimizu, K. (1992a). Bitter steroid glycosides, vernoniosides A,, A?, A3. and related B, from a possible medicinal plant, Vernonia amygdalinu, used by wild chimpanzees. Tetrahedron 48,625-632.

314

GEORGE A. LOZANO

Jisaka, M., Kawanaka, M., Sugiyama, H., Takegawa, K.. Huffman, M., Ohigashi. H., and Koshimizu, K. (1992b). Antischistosomal activities of sesquiterpene lactones and steroid glucosides from Vernonia amygdalina, possibly used by wild chimpanzees against parasiterelated diseases. Biosci. Biotechnol. Biochem. 56, 845-846. Jisaka, M., Ohigashi. H., and Takegawa. K.. Hirota, M., hie, R., Huffman, M. A,, and Koshimizu. K. (1993a). Steroid glucosides from Vernonia amygdalina, a possible chimpanzee medicinal plant. Phytochemistry 34, 409-413. Jisaka, M., Ohigashi, H.. and Takegawa, K., Huffman, M. A,, and Koshimizu, K. (1993b). Antitumoral and antimicrobial activities of bitter sesquiterpene lactones of Vernonia amygdalina, a possible medicinal plant used by wild chimpanzees. Biosci. Biotechnol. Biochem. 57, 833-834. Johns, T. (1990). “With Bitter Herbs they Shall Eat It: Chemical Ecology and the Origins of Human Diet and Medicine.” University of Arizona Press, Tucson. Johns, T., and Chapman, L. (1995). Phytochemicals ingested in traditional diets and medicines as modulators of energy metabolism. Recent Adv. Phytochem. 29, 161-188. Johns. T., and Duquette, M. (1991). Detoxification and mineral supplementation as functions of geophagy. Am. J. Clin. Nutr. 53, 448-456. Karban, R., and Myers, J. H. (1989). Induced plant responses to herbivory. Annu. Rev. Ecol. Syst. 20, 331-348. Kodric-Brown. A. (1989). Dietary carotenoids and male mating success in the guppy: An environmental component to female choice. Behav. Ecol. Sociobiol. 25, 393-401. Kopin, I. J. (1995). Definitions of stress and sympathetic neuronal responses. Ann. N. Y. Acad. Sci. 771, 19-40. Krebs, J. R. (1980). Optimal foraging, predation risk and territory defense. Ardea 68,83-90. Kreulen, D. A. (1985). Lick use by large herbivores: A review of benefits and banes of soil consumption. Mamm. Rev. 15, 107-123. Krischik, V. A., Barhosa, P., and Reichelderfer. C. F. (1988). Three trophic level interactions: Allelochemicals. Manduca sexta (L.), and Bacillus thrtringiensis var. kurstaki Berliner. Environ. Entomol. 17, 476-482. Kupchan, S. M., Hemingway, R. J., Karim, A,, and Werner, D. (1969). Tumor inhibitors. XLVII. Vernodalin and vernomygdin, two new cytotoxic sesquiterpene lactones from Vernonia amygdalina Del. J. Org. Chem. 34, 3908-391 1. Lima, S. L., and Dill, L. M. (1990). Behavioral decisions made under the risk of predation: A review and prospectus. Can. J. Zool. 68,619-640. Longino, J. T. (1984). True anting by the capuchin (Cebus capucinus). Primates 25,243-245. Loye, J., and Carroll. S. (1995). Birds, bugs, and blood: Avian parasitism and conservation. Trends Ecol. Evol. 10,232-235. Lozano, G. A. (1991). Optimal foraging theory: A possible role for parasites. Oikos63,391-395. Lozano, G. A. (1994). Carotenoids, parasites, and sexual selection. Oikos 70, 309-31 1. Mahry, T. J., Gill, J. E., Burnett, W. C., Jr., and Jones, S. B., Jr. (1977). Antifeedant sesquiterpene lactones in the Compositae. ACS Symp. Ser. 62, 165-178. Mahaney, W. C. (1993). Scanning electron microscopy of earth mined and eaten by mountain gorillas in the Virunga mountains, Rwanda. Primates 34, 31 1-319. Mahaney, W. C., Aufreiter, S., and Hancock, R. G. V. (199%). Mountain gorilla geophagy: A possible seasonal behavior for dealing with the effects of dietary changes. Int. J. Primatol. 16, 475-487. Mahaney, W. C., Hancock, R. G. V.. Aufreiter, S., and Huffman, M. A. (1996). Geochemistry and clay mineralogy of termite mound soil and the role of geophagy in chimpanzees of the Mahale mountains, Tanzania. Primates 37, 121-134.

SELF-MEDICATION I N WILD ANIMALS

315

Mahaney. W. C.. Hancock, R. G. V., and Inoue, M. (1993). Geochemistry and clay mineralogy of soils eaten by Japanese macaques. Primates 34, 85-91. Mahaney, W. C., Milner, W. M.,Sanmugadas, K.,Hancock, R. G. V.. Aufreiter.S., Wrangham, R., and Pier. H. W. (1997). Analysis of geophagy soils in Kinale Forest, Uganda. Primates 38, 159-176. Mahaney, W. C.. Stambolic, A,, Knezevich, M.. Hancock. R. G. V.. Aufreiter. S., Sanmugadas. K., Kessler. M. J., and Grynpas, M. D. (1995a). Geophagia amongst rhesus macaques on Cay0 Santiago, Puerto Rico. Primates 36, 323-333. Mahaney. W. C., Watts, D. P., and Hancock, R. G. V. (1990). Geophagia by mountain gorillas (Gorilla gorillu heringei) in the Virunga mountains. Rwanda. Primates 31, 113-120. Maitland, D. P. (1994). A parasitic fungus infecting yellow dungflies manipulates host perching behaviour. Proc. R. Soc. London Ser. B 258, 187-193. Mason, J. R., and Reidinger, R. F., Jr. (1982). Observational learning of food aversions in red-winged blackbirds (Agelaius phoeniceus). Auk 99, 548-554. McGrath, J. E. (1970). A conceptual formulation for research on stress. In “Social and Psychological Factors in Stress’’ (J. E. McGrath, ed.), pp. 10-21. Holt, Rinehart & Winston, New York. McKenna, D. J . (1996). Plant hallucinogens: Springboards for psychotherapeutic drug discovery. Behav. Brain Res. 73, 109-116. Messner, E. J., and Wrangham. R. W. (1996). In vitro testing of the biological activity of Rubia cordifolia leaves on primate Strongyloides species. Primates 37, 105-108. Milinski, M. (1982). Optimal foraging: The influence of intraspecific competition on diet selection. Behav. Ecol. Sociohiol. 11, 109-1 15. Milinski, M. (1984). Parasites determine a predator’s optimal feeding strategy. Behav. Ecol. Sociohiol. 15, 35-37. Milinski, M.. and Bakker. T. C. M. (1990). Female sticklebacks use male coloration in mate choice and hence avoid parasitized males. Nature (London) 344,330-333. Milinski, M.. and Heller, R. (1978). Influence of a predator on the optimal foraging behaviour of sticklebacks (Gasterosteus acitleatus L.). Nature (London) 275, 642-644. Millikan, G. C.. Caddis, P.. and Pulliam, H. R. (1985). Interspecific dominance and the foraging behavior of juncos. Anim. Behav. 33,428-435. Mims, C. A., Dimmock, N., Nash. A., and Stephen. J. (1995). “Mims’ Pathogenesis of Infectious Disease.”- 4th ed. Academic Press, London. Maller, A. P., Dufva, R., and Allander, K. (1993). Parasites and the evolution of host social behavior. Adv. Study Behav. 22, 65-102. Moore, J., and Gotelli, N. J. (1990). A phylogenetic perspective on the evolution of altered host behaviours: A critical look at the manipulation hypothesis. In “Parasitism and Host Behaviour” (C. J. Barnard and J. M. Behnke, eds.). pp. 193-233. Taylor & Francis, London. Newton, P. (1991). The use of medicinal plants by primates: A missing link: Trends Ecol. Evol. 6, 297-299. Newton, P. N.. and Nishida, T. (1990). Possible buccal administration of herbal drugs by wild chimpanzees. Pan troglodytes. Anim. Behav. 39, 798-801. Oates, J. F. (1978). Water-plant and soil consumption by Guereza monkeys (Colobusguereza): A relationship with minerals and toxins in the diet? BioTropica 10, 241-253. Ohigashi, H., Takagaki, T., Koshimizu, K., Watanabe, K.. Kaji, M., Hoshino, J.. Nishida, T., Huffman, M. A.. Takasaki. H., Jato. J.. and Muanza. D. N. (1991a). Biological activities of plant extracts from tropical Africa. Afr. Stud. Monogr. U,201-210. Ohigashi, H., Jisaka. M., Takagaki, T., Nozaki, H., Tada, T., Huffman, M. A,, Nishida. T., Kaji. M.. and Koshimizu. K. (1991b). Bitter principle and a related steroid glucoside from

316

GEORGE A. LOZANO

Vernonia arnygdalina, a possible medicinal plant for wild chimpanzees. Agric. B i d . Chem. 55, 1201-1203. Ohigashi, H., Huffman. M. A,, Izutsu, D., Koshimizu, K., Kawanaka, M.. Sugiyama. H., Kirby, G . C., Warhurst, D. C., Allen, D., Wright, C. W., Phillipson, J. D., Timon-David, P., Delmas. F., Elias, R., and Balansard. G . (1994). Toward the chemical ecology of medicinal plant use in chimpanzees: The case of Vernonia arnygdalina, a plant used by wild chimpanzees possibly for parasite-related diseases. J . Chern. Ecol. 20, 541-553. Page. J. E., Balza. F.. Nishida, T., and Towers, G . H. N. (1992). Biologically active diterpenes from Aspilia rnossarnbicensis, a chimpanzee medicinal plant. Phytochemistry 31, 34373439. Pennings, S. C. (1994). Interspecific variation in chemical defenses in the sea hares (Opisthobranchia: Anaspidea). J. Exp. Mar. B i d . Ecol. 180, 203-219. Phillips-Conroy, J. E. (1986). Baboons, diet. and disease: Food plant selection and schistosomiasis. In “Current Perspectives in Primate Social Dynamics” (D. M. Taub and F. A. King, eds.), pp. 287-304. Van Nostrand-Reinhold, New York. Phillips-Conroy, J. E., Knopf, P. M. (1986). The effects of ingesting plant hormones on schistosomiasis in mice: an experimental study. Biochemical Systematics and Ecology 14, 637-645. Potter, E. F. (1970). Anting in wild birds. its frequency and probable purpose. Auk 87,692-713. Price, P. W., Bouton, C. E., Gross, P., McPheron. B. A., Thompson, J. N., and Weis, A. E. (1980). Interactions among three trophic levels: Influence of plants on interactions between insect herbivores and natural enemies. Annu. Rev. Ecol. Syst. ll, 41-65. Rhoades, D. F. (1979). Evolution of plant chemical defenses against herbivores. In “Herbivores: Their Interaction with Secondary Plant Metabolites” (G. A. Rosenthal and D. H. Janzen, eds.), Chapter 1, pp. 3-54. Academic Press, New York. Richter, C. P. (1943). Total self-regulatory functions in animals and human beings. Harvey Lect. 38, 63-103. Rodriguez, E., and Wrangham, R. (1993). Zoopharmacognosy: The use of medicinal plants by animals. Recent Adv. fhytochern. 27, 89-105. Rodriguez, E., Aregullin, M., Nishida, T., Uehara, S., Wrangham, R.. Abramowski, Z., Finlayson, A., and Towers, G . H. N. (1985). Thiarubrine-A, a bioactive constituent of Aspilia (Asteraceae) consumed by wild chimpanzees. Experientia 41, 419-420. Rothschild, M. (1972). Some observations on the relationship between pants, toxic insects and birds. In “Phytochemical Ecology” (J. B. Halborne, ed.), pp. 1-12. Academic Press, London. Rozin, P. (1967). Specific aversions as a component of specific hungers. J. Comp. fhysiol. Psychol. 64, 237-242. Rozin, P. (1969). Adaptive food sampling patterns in vitamin deficient rats. J. Comp. Physiol. fsychol. 69, 129-132. Rozin. P. (1976). The selection of food by humans, rats, and other animals. Adv. Study Behav. 6, 21-76. Rozin, P., and Kalat, J. W. (1971). Specific hungers and poison avoidance as adaptive specializations of learning. Psychol. Rev. 78, 459-486. Sapolsky, R. M. (1994). Fallible instinct: A dose of skepticism about the medicinal “knowledge” of animals. Sciences 34, 13-15. Sears, C. (1990). The Chimpanzee’s medicine chest. New Sci. 1728, 42-44. Selye, H. (1976). “Stress in Health and Disease.” Butterworth, Boston. Sih, A. (1980). Optimal behavior: Can foragers balance two conflicting demands? Science 210, 104-1043.

SELE-MEDICATION IN WILD ANIMALS

317

Stephens, D. W., and Krebs. J . R. (1986). “Foraging Theory.” Princeton University Press. Princeton, NJ. Strier. K. B. (1993). Menu for a monkey. Nat. Hist. 102, 34-43. Takasaki, H.. and Hunt, K. (1987). Further medicinal plant consumption in wild chimpanzees? Afr. Stiid. Monogr. 8, 125-128. Towers, G. H. N.. Abramowski. Z., Finlayson, A., and Zucconi. A. (1985). Antibiotic properties of thiarubrine A,. a naturally occurring dithiacyclohexadiene polyine. Planfa Med. 225-229. Vander, A. J . (1981). “Nutrition, Stress, and Toxic Chemicals.” University of Michigan Press. Ann Arbor. Vermeer, D. E.. and Ferrell, R. E.,Jr. (1985). Nigerian geophagical clay: A traditional antidiarrheal pharmaceutical. Sciznce 227, 634-663. Wagner. H. K. M. (1995). Immunostimulants and adaptogens from plants. Recent Adv. Phytochem. 29, 1-18. Walzer, P. D., and Genta, R. M. (1989). “Parasitic Infections in the Compromised Host.” Dekker, New York. Wimberger. P. H. (1984). The use of green plant material in bird nests to avoid ectoparasites. Auk 101, 615-618. Wrangham. R. W. (1995). Relationship of chimpanzee leaf-swallowing to a tapeworm infection. Am. J . Primatol. 37, 297-303. Wrangham. R. W., and Goodall. J . (1989). Chimpanzee use of medicinal leaves. In “Understanding Chimpanzees” (P. G. Heltne, and L. A. Marquardt. eds.), pp. 22-37. Harvard University Press. Cambridge, MA. Wrangham, R. W., and Nishida, T. (1983). Aspilia spp. leaves: A puzzle in the feeding behavior of wild chimpanzees. Primates 24,276-282. Yoerg. S . 1. (1991). Social feeding reverses learned flavor aversions in spotted hyenas (Crocufa crocufa).J . Comp. Psychol. 105, 185-189. Zuk, M. (1992). The role of parasites in sexual selection: Current evidence and future directions. Adv. Study Behav. 21, 39-68.

This Page Intentionally Left Blank

ADVANCES IN THE STUDY OF BEHAVIOR. VOL. 21

Stress and Human Reproductive Behavior: Attractiveness, Women’s Sexual Development, Postpartum Depression, and Baby’s Cry RANDY THORNHILL AND BRYANT FURLOW DEPARTMENT OF BIOLOGY

THE UNIVERSITY OF NEW MEXICO ALBUQUERQUE, NEW MEXICO

87131

I. INTRODUCTION Stress resistance is highly variable among individuals within populations, and this variation seems to be an important mediator of the association between phenotypic variation and individual variation in reproductive success (i.e., an important mediator of fitness variance) (Parsons, 1996). Stress tolerance and resistance requires energy that must be diverted from other functional channels of the organism, which negatively affects the organism’s energy allocations to the other functions. Parsons (1996) has emphasized that, although biotic stressors have been central in analyses by evolutionary biologists, abiotic stressors are often important sources of selection. He emphasizes also that an interaction between biotic and abiotic stressors is often seen, and that this interaction can be subtle and easily overlooked in favor of interpretations based on centrality of biotic stressors as selection agents. Because of the often intimate association between abiotic and biotic stressors, we choose here to use a unitary concept and not distinguish between abiotic and biotic stressors. Stress can be considered in ecological and evolutionary terms. An ecological stress becomes significant in microevolutionary terms when it generates selection for coping with the stress. An ecological stress is significant in long-term evolution when an adaptation for coping with the stress arises. Stress-related selection generates specieswide adaptation over long periods of evolutionary time when the selection is consistently directional and is not completely opposed by selection in other contexts. Thus, a species’ adaptation that functions against a stress is unequivocal evidence that the stress was important to fitness over the evolutionary history of the species. 31 9

Copyright 0 1998 by Academic Press All rights of reproduction in any form reserved. 0065-3454/98 $25.00

320

RANDY THORNHILL A N D BRYANT FURLOW

Sweating in humans is an adaptation for prevention of overheating of the other adaptations of the body beyond their optimal performance temperatures. Human hunger is an adaptation to deal with stressfully low levels of nutrients. Women’s preference for socially successful men as mates is an adaptation for obtaining mates who provide resource stability and thus ameliorate resource-related stresses. In an important sense, all adaptations are evolved stress reducers, because adaptations are caused by selection, and all selection-each of Darwin’s hostile forces-involves stress. Said differently, an adaptation is an evolved phenotypic solution to an environmental (abiotic or biotic) problem or stressor. Accordingly, all human behavior is stress related because all human behavior is the output of information-processing psychological adaptation. From the vast domain of human behavior and stress, we have selected a number of topics that we feel are advancing in terms of empirical or theoretical progress. Human attraction and attractiveness is one such topic. Attractiveness may be largely a certification of the ability to cope with environmental and genetic stresses acting during development. Attraction appears to arise from adaptations designed to secure benefits for self and offspring that offset stress. We examine the role of stress-related attraction and attractiveness in the three major areas of human social behavior: mating, nepotism, and reciprocity. Another topic that we address centers around the stresses of the home environment during girls’ upbringing and how this may functionally mold the development of women’s sexuality, particularly the variation among women in sexual arousal. There is considerable evidence that women raised in low-resource settings exhibit different sexual tactics than women raised in settings with more resources. The former women show earlier maturity and age of copulation, more sex partners, less stable pair-bonds, and less emotional warmth toward mates. It has been proposed that these tactics serve women’s interest in low-resource settings by promoting mating with multiple males, each of whom can provide limited resources at best. We argue that because coital female orgasm has a bonding effect, limited and highly selective coital orgasm is expected in low-resource settings, because it will promote relatively weak bonds with multiple males. There is evidence that low-orgasmic women are concentrated in low-resource settings and that reduced parental investment is associated with the ontogeny of loworgasmic response in women. We also examine the stresses on human parents surrounding parental investment in babies. Our focus is on postpartum depression of mothers and on infants’ crying. We propose and test the hypothesis that postpartum depression is an evolved manifestation of discriminative parental solicitude that motivates mothers to eliminate newborn babies under ecological situa-

STRESS AND HUMAN BEHAVIOR

321

tions that were not conducive to infant survival in human evolutionary history. The baby’s cry is hypothesized to be an evolved, honest signal of offspring reproductive value. Data bearing on this hypothesis are reviewed. Both the postpartum depression hypothesis and the baby cry hypothesis receive support.

11. HUMAN ATTRACTION A N D ATTRACTIVENESS

We discuss and test the hypothesis that human sexual competition revolves around assessment of and displays of stress resistance (Thornhill and Gangestad, 1993; Grammer and Thornhill, 1994). The stress resistance is seen in developmental stability and in development of secondary sexual traits. We treat developmental stability briefly, as it is covered in detail by Moller (this volume). We then discuss the connection between secondary sexual traits and stress. After treating the role of stress in sexual attraction and attractiveness, we turn to the importance that attractiveness based on stress resistance may play in human nepotism and reciprocal altruism. STABILITY A. TYPESOF DEVELOPMENTAL Developmental stability occurs when the evolved or adaptive developmental trajectory is achieved despite environmental and genetic perturbations during development. There are a number of kinds of developmental instability (Mdler, this volume). Two important kinds in our discussion are phenodeviance and fluctuating asymmetry. Phenodeviants reflect conspicuous deviations from the adaptive developmental trajectory. Human birth defects and so-called minor physical anomalies (Waldrop et al., 1968) are examples of human phenodeviants. Birth defects may arise from environmental perturbations (e.g., maternally ingested toxins during pregnancy; Profet, 1995) or genetic perturbations. Minor physical anomalies arise in the first three months of pregnancy. Phenodeviance refers to any relatively gross deviation from the adaptive phenotypic target of development, whether it be morphological, behavioral, physiological, immunological, or of other origin (Zakharov, 1992; Mdler, this volume). Phenodeviants are relatively rare in populations of organisms because they reflect only the most radical disruptions of development. Developmental instability is most often measured as fluctuating asymmetry because fluctuating asymmetry is a more sensitive measure of developmental disturbance than is phenodeviance. Fluctuating asymmetry is a deviation from bilateral symmetry in normally bilaterally symmetrical

322

R A N D Y THORNHILL A N D BRYANT FURLOW

morphological traits for which signed differences between the left and right sides have a mean of zero and are normally distributed within a population (Van Valen, 1962). A long-standing interest in fluctuating asymmetry (Ludwig, 1932) derives from the notion that perfect bilateral symmetry is the optimum and fluctuating asymmetry results from imprecise expression of developmental design. The corresponding two sides of a bilaterally symmetrical trait are encoded by the same genes; therefore, fluctuating asymmetry arises from environmental stressors or stressors from a hostile genetic environment within the genome that lower developmental homeostasis. Accordingly, fluctuating asymmetry increases with exposure to a wide range of environmental insults during ontogeny such as parasites (Bailit et al., 1970; Moller, 1992; Moller and Saino, 1994; Polak, 1997), pollutants (Parsons, 1990, 1992), radiation (Moller, 1993a), extreme temperatures, and other adverse physical conditions (Parsons, 1990,1992), including marginal habitats (Moller, 1995). Fluctuating asymmetry also increases with exposure to genetic perturbations such as inbreeding (Lerner, 1954), deleterious recessives (Lerner, 1954; Parsons, 1990), homozygosity (Lerner, 1954; Mitton and Grant, 1984; Mitton, 1993), directional selection (Moller and Pomiankowski, 1993), hybridization of genetically distinct populations (Moller and Swaddle, 1997), and chromosomal abnormalities (Shapiro, 1992; Fraser, 1994). Fluctuating asymmetry can be used to compare populations or individuals within populations. Within populations, asymmetry can vary considerably across individual organisms. Individual fluctuating asymmetry refers to an individual’s deviation in either direction from perfect bilateral symmetry. This variation, in part, is due to genetic differences. A meta-analytic review of 34 studies of 17 species revealed that fluctuating asymmetry has significant heritability in general, and multiple studies of human fluctuating asymmetry show significant additive genetic variance (Moller and Thornhill, 1997a,b). Fluctuating asymmetry is importantly related to the action of Darwinian natural and sexual selection. In a wide range of animal species, individuals’ asymmetry is negatively correlated with fitness components pertinent to natural selection, specifically fecundity, growth rate, and survival (Watson and Thornhill, 1994; reviews in Mdler, 1997a; Moller and Swaddle, 1997). In the last several years, evolutionary biologists have studied the role of fluctuating asymmetry in sexual selection. Sexual selection is the individual variation in offspring production that is the consequence of individual differences in traits that affect access to mates. In a diversity of animal species low-asymmetry males tend to obtain more mates than highasymmetry males, through either advantage in male-male competition for females or female mate choice (Moller and Pomiankowski, 1993; Watson and Thornhill, 1994; Polak, 1997;review in Moller & Swaddle, 1997). Recent

STRESS AND HUMAN BEHAVIOR

323

evidence reveals that fluctuating asymmetry plays a role in human sexual selection as well, as we discuss later.

SEXUAL CHARACTERS AND STRESS B. SECONDARY Secondary sexual traits evolve by sexual selection and are fundamentally involved in courtship and other intrasexual contests for mates. Thus, secondary sexual traits are signals. An important discovery in modern studies of secondary sexual traits is that the expression of such traits often signals reliable information about the phenotypic condition of the signaler. Such traits function as reliable signals for two related reasons. First, their expression is facultatively plastic and thus is importantly dependent on the environment in which they develop, including the bearer’s condition during development of the trait. Second, many secondary sexual traits are survival handicaps (Zahavi, 1975, 1977a). They are costly to produce, but more so to individuals of low rather than high phenotypic quality. A high-quality individual is therefore able to develop the most extravagant secondary sexual character, but at a relatively low cost. The differential cost of the secondary sexual trait is the mechanism that ensures reliable signaling of quality, because only high-quality individuals with superior viability genes will be able to survive with the most extreme levels of sexual display (Haywood, 1989; Iwasa et al., 1991). A special version of the handicap mechanism is the immunocompetence handicap hypothesis (Folstad and Karter, 1992; Wedekind and Folstad, 1994), which suggests that secondary sexual characters develop in response to circulating androgens (or other self-regulating biochemicals) that increase the expression of secondary sexual characters, but reduce the functioning of the immune system, and thus suppress the ability of individuals to raise an immune defense against parasites. In other words, there may be an intricate, negative feedback mechanism among host secondary sexual characters, host hormones, host immune defense, and parasites. Highquality males will be able to develop large sexual traits, cope with high levels of androgens, and to only a relatively small extent compromise their immune defense. Sex differences in the course of parasite infections (Bundy, 1989; Zuk, 1990) and relationships between sex hormones and parasitism (Alexander and Stimson, 1989) are consistent with the immunocompetence handicap hypothesis. Secondary sexual traits typically show much more fluctuating asymmetry than other traits (Mdler and Pomiankowski, 1993). This is expected because such traits are often condition dependent, elaborate, and energetically demanding. This combination presents a developmental challenge for an individual to make secondary sexual traits perfectly symmetrical. This is not

324

RANDY THORNHILL AND BRYANT FURLOW

to say that secondary sexual traits are the only traits in which fluctuating asymmetry will correlate with phenotypic quality. Indeed, fluctuating asymmetry in ordinary traits may often be a good predictor of Darwinian fitness and genetic quality. It is to say that secondary sexual traits will often be the most revealing of phenotypic and genetic quality differences among individuals. Furthermore, it is likely that many bilateral secondary sexual traits are designed to signal symmetry per se and thereby phenotypic quality; this is supported by experimental evidence from studies of male secondary sexual traits in birds (Mgller, 1993b; Swaddle and Cuthill, 1994a,b). Secondary sexual traits are exaggerated by a directional sexual selection process. Directional selection appears to increase fluctuating asymmetry and developmental instability in general regardless of whether the selected trait is sexual or ordinary (Mgller and Pomiankowski, 1993; Pomiankowski and Mgller, 1995;review in M d l e r and Swaddle, 1997). Directional selection is thought to increase fluctuating asymmetry by incorporation of new alleles with exaggerated trait value and by disruption of the mechanisms that stabilize development. The new alleles may cause genetic stresses during development. The disruption of developmental mechanisms associated with the evolutionary response to directional selection makes the developing organism susceptible to both environmental and genetic perturbations. Thus, human secondary sexual traits may importantly display stress resistance because they may be immunocompetence handicaps and because they may manifest relatively high developmental instability. C. STRESS RESISTANCE A N D HUMAN SEXUAL SELECTION

Current knowledge of human attractiveness indicates that it may be a certification of stress resistance and thereby a health certification. Human attractiveness research is thus potentially important to the health professions. Human attractiveness relates to health through phenotypic quality in general and immunocompetence and developmental stability, in particular, and through immunocompetence handicapping sex hormones (Thornhill and Mgller, 1997). Cross-cultural research has shown that, although men place more value on physical attractiveness of a mate than do women, both sexes highly value it (Buss, 1989). The aesthetic judgments of faces made by individuals of different cultures tend to agree, and children as young as 6 months show preferences similar to those of adults (see review in Cunningham et al., 1995; Jones, 1995). Certain human facial features are secondary sexual traits, arising or increasing in size at sexual maturity under the proximate influence of androgens and estrogens. Both sexes have both hormones, but the ratio at puberty

STRESS AND HUMAN BEHAVIOR

325

is sex specific. Relatively high testosterone levels lead to growth of lower face and jaw, cheekbones and brow ridges, and projection of the central face between the brow and bottom of the nose. Relatively high estrogen at puberty prevents this growth, but yields increased lip size. These features distinguish young men’s and women’s faces (see reviews in Enlow, 1990; Symons, 1995). As estrogen declines with female age, testosterone masculinizes the female face. Thus, a highly estrogenized female face marks youth and thus high fertility (Symons, 1995; but see Jones, 1995). Abundant evidence indicates that markers of high estrogen such as small lower jaw and lower face, in general, and large lips are attractive in women’s faces (Johnston and Franklin, 1993; Perrett et af., 1994; Cunningham et af., 1995; Jones, 1995). By contrast, a large lower jaw is rated as dominant and attractive in the male face (reviewed by Symons, 1995). The dominance rating of adult men’s faces correlates positively with their amount of previous sexual experience and with testosterone level during puberty (Mazur et af., 1994). Facial secondary sex traits, in addition to conveying sex and sexual maturity, appear to be designed to advertise phenotypic and genetic quality, because of their connection to sex hormones. As mentioned earlier, testosterone appears to be an immunosuppressor. Estrogen may also negatively impact immunocompetence when at high titers (Grossman, 1985; Ahmed and Talal, 1990). Sex-hormone-facilitated markers honestly advertise immunocompetence because the high hormone titer needed to produce attractive features simultaneously handicaps disease resistance accordingly, and thus can be afforded only by individuals of extraordinary immunocompetence. This model applied to human facial secondary sexual traits suggests that the facial hormone markers are conditionally expressed advertisements of phenotypic quality or overall health, and simultaneously supports the view that attractiveness judgments based on these features arise from adaptation designed to detect mate quality. Crammer and Thornhill (1994) found that opposite-sex health ratings of facial photos positively correlate with the degree of development of the sex-specific facial secondary sexual traits. Attractive expressions of human facial secondary traits actually covary with health (T. Shackleford and R. Larsen, personal communication, January 1996). Subjects completed daily records of psychological, emotional, and health symptom status over a 2month period. Shackleford and Larsen found evidence that women with highly estrogenized facial structure-prominent cheekbones, short chin, and narrow lower jaw-and men with highly testosteronized facial structure-prominent cheekbones, large chin, and wide lower jaw-are indeed emotionally, psychologically, and physiologically healthier. For example, compared to women with wide lower jaws, women with small lower jaws

326

RANDY THORNHILL AND BRYANT FURLOW

are less depressed, less emotionally labile, more excited and enthusiastic, and experience fewer gastrointestinal problems and less muscle soreness. Men with wide jaws are less neurotic, less emotionally labile, feel less sluggish, and experience fewer physical symptoms of ill health daily than men with narrow jaws. It should be noted that prominent cheekbones are attractive in both sexes, but reflect sex-specific hormone effects (see Symons, 1995). They reflect high midfacial flatness in women due to high estrogen. In men, they reflect lateral bone growth due to androgens. There is evidence that parasites play a critical role in human attractiveness. Disease organisms causing skin infections and rashes influence human attractiveness (Symons, 1995). Also, the value that people of either sex place on physical attractiveness in choice of a mate correlates positively with the prevalence of parasites across 29 human societies (Gangestad and Buss, 1993). Moreover, maxillary sinusitis of probable bacterial etiology appears to cause irregular growth of most facial bones of the lower face. Children with these facial irregularities are recognizable on the basis of their facial appearance (Yates, 1928). It is unknown, however, how sinusitis may influence the structure of facial secondary sexual traits at puberty. Both size and symmetry of facial structures may be influenced by sinus infections. Other evidence that physical attractiveness relates to phenotypic quality has come from research on facial symmetry. In three studies, in both sexes, faces with high bilateral symmetry are rated more attractive than less symmetrical faces (Grammer and Thornhill, 1994; D. Perrett and M. Burt, personal communication, July 1996; L. Mealey, R. Bridgstock, and G. Townsend, personal communication, November 1996), and facially symmetric individuals exhibit more emotional, psychological, and physiological health (Shackleford and Larsen, 1997). Faces show no evidence of systematic directionality in size of hard or soft tissue, according to Sackheim’s (1985) review of 50 years of research on facial asymmetry. Thus, most facial asymmetry appears to be fluctuating (see also Hershkovitz et al., 1992). Facial asymmetry is the rule. Probably all faces show some asymmetry of both hard and soft tissue and the degree varies considerably among healthy subjects (Farkas and Cheung, 1981; Sutton, 1969; Sackheim, 1985; Peck et al., 1991; Hershkovitz et al., 1992). Two studies involving making perfectly symmetrical faces from hemifaces using computer techniques have shown that such faces are rated lower in attractiveness than natural faces, which exhibit some asymmetry (Langlois et al., 1994; Swaddle and Cuthill, 1995). This effect apparently is due largely to the unnatural facial feature sizes and textures created by the hemiface method (D. Perrett and M. Burt, personal communication, July 1996; L. Mealey, R. Bridgstock and G. Townsend, personal communication, November 1996). Also, as Swaddle and Cuthill point out, the low attractiveness

STRESS AND HUMAN BEHAVIOR

327

ratings of hemifaces may reflect, in part, the fact that perfectly symmetrical faces look unnatural, perhaps even dead, given that emotional expression in the face is asymmetric (e.g., see Sackheim, 1985). There is evidence that facial asymmetries correlate within individuals. Peck et al. (1991) examined facial skeletal asymmetry in 52 adult subjects. Eye orbit asymmetry was positively correlated with zygion (cheekbone) asymmetry and zygion with gonium (mandible) asymmetry. Farkas and Cheung (1981), in a large study of soft-tissue asymmetries of children and young adults, also found significant correlations within individuals in certain facial asymmetries. It is conceivable that these within-individual correlations may be generated by variation in sinus infections prior to puberty or in early adulthood among individuals, given the strong effect of sinusitis on facial bone growth (Yates, 1928). Also indicating that facial attractiveness marks phenotypic quality is the relationship between facial asymmetry and facial sex hormone markers. Facial symmetry in each sex correlates with the sex-specific attractive expression of facial secondary sexual traits (e.g., symmetry positively correlates with chin size in men and negatively with chin size in women) (Gangestad and Thornhill, 1997). Thus, people with symmetric faces tend to have attractive sex hormone markers in their faces. Moreover, composite faces made from many individual photos are rated more attractive than the majority of the individual photos used to make the composites (Langlois and Roggman, 1990; Langlois et al., 1994). This means that, at least for certain facial features (not the secondary sexual traits), being near the mean is associated with greater attractiveness (also see Jones and Hill, 1993). It has been suggested that averageness in certain facial features may mark phenotypic and genetic quality because averageness in traits under stabilizing selection often positively covaries with heterozygosity (Thornhill and Gangestad, 1993). Heterozygosity is sometimes associated with increased developmental stability and may be associated with reduced susceptibility to parasites (see Thornhill and Gangestad, 1993; Wedekind, 1994; Maller and Swaddle, 1997). The immunocompetence model may also hold for nonfacial secondary sexual traits of men. For example, male body mass may signal immunocompetence. Adult body mass is sexually differentiated, with males being larger. This is due, in part, to men’s greater muscle mass, which arises at puberty as a result of testosterone. Male body mass in the Ache Indians (who do not use contraception) positively correlates with number of offspring produced, apparently because heavier men are more attractive to women (Hill and Hurtado, 1996). In the West, women rate male athletic builds more attractive than other male builds (Gangestad and Thornhill, 1997), and the mates of larger men tend to orgasm during copulation more frequently

328

RANDY THORNHILL A N D BRYANT FURLOW

(Thornhill et al., 1995). Copulatory orgasm may be a mechanism of female choice (see later discussion). Men’s mass and athleticism appear to signal men’s developmental health, that is, low-fluctuating asymmetry, which may reflect immunocompetence. Male body mass is positively correlated with male body symmetry (Manning, 1995; Thornhill et al., 1995), and, as already mentioned, men of high body symmetry are more muscular and vigorous. Such men also show lower resting metabolic rates than men of low body symmetry (Manning et al., 1997). Women’s body mass negatively correlates with their body symmetry. Thus, large men and small women exhibit greater developmental stability (Manning, 1995). The relatively low metabolic rate of symmetric men implies that such men have more energy available for other body functions, which may account for their greater vigor. If so, vigor itself in men is an honest signal of developmental health. Men exhibiting developmental stability are sexually attractive to women. Compared to men with high asymmetry, men with low asymmetry have greater facial attractiveness, greater numbers of lifetime sex partners, more extrapair copulations, and quicker access to copulation in romantic relationships. Low-asymmetry men also are chosen more often as extrapair copulatory partners and begin sexual intercourse earlier in their life history (Gangestad et al., 1994; Thornhill and Gangestad, 1994; Baker, 1997; Gangestad and Thornhill, 1997). Also, the mates of symmetrical men show the most reported copulatory orgasms (Thornhill et al., 1995; but see Baker, 1997). Female copulatory orgasm may function in selective bonding with males and in selective ejaculate retention. Thus, female orgasm may be a subtle form of female choice involving choice of sire (Smith, 1984; Birkhead and Moller, 1993; Baker and Bellis, 1995; Baker, 1997). Nonfacial human secondary sexual features in women appear to signal phenotypic quality, and higher quality expressions are judged more attractive. The sexual dimorphism in waist-to-hip ratio in humans arises at puberty and is facilitated by sex hormones. Waist-to-hip ratio in women correlates negatively with estrogen, fertility, and health, and positively with age, and low waist-to-hip ratios (.6-.7) are maximally attractive in women (Singh, 1993, 1995). Adult female breasts develop at puberty under the influence of estrogen. Breast-size symmetry was positively correlated with reported fertility (number of children during their lifetime) in two separate samples of women (United States and Spain) (Mdler et al., 1995). Breast symmetry positively affects attractiveness judgments of men, and men’s interest in both short-term and long-term relationships (Singh, 1995). In theory, if symmetry marks high phenotypic and genetic quality, and bilateral secondary sexual traits honestly signal quality as reduced asymme-

STRESS AND HUMAN BEHAVIOR

329

try, it is predicted that bilateral secondary sexual traits will exhibit high levels of fluctuating asymmetry, specifically higher levels than in ordinary traits not influenced by sexual selection. This has been found in numerous nonhuman species (Mdler and Swaddle, 1997). This same pattern is found in Homo sapiens. Absolute breast size asymmetry in a sample of 172 women in Spain showed a mean of 1.23 cm (Moller et af.,1995). Similar high breast size asymmetry values were obtained from samples in New Mexico (Mdler et al., 1995) and in the United Kingdom (Manning et al., 1996). Average relative breast size asymmetry was about 5% in these samples (i.e., absolute breast size asymrnetry/breast size). Other morphological characters in humans usually show 2% or less relative fluctuating asymmetry (Livshits and Kobyliansky, 1991; Thornhill and Gangestad, 1994: Manning, 1995). Human facial asymmetry also may show the pattern of more fluctuating asymmetry with facial secondary sexual traits (Peck et af.,1991). Eye socket asymmetry showed relatively low asymmetry, whereas more asymmetry was found in the two sexual traits zygion (cheekbone at most outward lateral point) and lower jaw. Farkas and Cheung (1981) measured lateral facial soft-tissue asymmetry of children and 18-year-olds of both sexes. Only in the 18-year-olds was there a significant sex difference (69% of males and 47% of females) in asymmetry in the facial trait nasion to tragion distance. This is the distance on the lateral aspect of the face between a point on the upper end of the nose (nasion) to the temple (tragion). This distance would include the outward pubertal growth of the central facial region (eyebrows to bottom of the nose) seen in males (Enlow, 1990; Symons, 1995), which is facilitated by testosterone. Current knowledge of human physical attractiveness leads to the conclusion that the sexual selection responsible for designing many of the secondary sexual traits was not Fisherian, that is, male and female winners of sexual competition in human evolutionary history did not sexually signal with arbitrarily attractive traits. Human secondary sexual traits reveal developmental stress resistance, and their attractive expressions signal the ability to resist developmental stress. Attractive adult faces of both sexes reflect secondary sexual traits requiring high titers of sex-specific hormones that may connote immunocompetence, and attractive facial secondary sexual traits may connote greater emotional and psychological health. Also, attractive faces are developmentally healthy, as seen in their high developmental stability. The adult male body is attractive when reflecting testosterone effects and athleticism, which covary with physical and developmental health. The adult female body is attractive when it shows good health by developmental stability in breasts and high estrogen in low waist-tohip ratio.

330

RANDY THORNHILL A N D BRYANT FURLOW

The use of simple neural network models has led to the hypothesis that preferences for exaggerated or symmetrical stimuli that arise from these models are inevitable by-products of visual information processing and not the result of selection for detection of signaler quality (e.g., Johnstone, 1994). However, Dawkins, and Guilford (1995) have pointed out that these simple models do not behave like the visual systems of animals, and the preferences they generate may be incidental to the inadequacies in the models themselves. Furthermore, the by-product hypothesis of preferences for exaggeration or symmetry is unlikely to apply to humans, because of the association between preference and tangible benefits, specifically stress resistance in a mate and the genetic and material benefits a stress-resistant mate can provide.

D. STRESS RESISTANCE AND NONSEXUAL SOCIAL BEHAVIOR The above discussion of human attractiveness in relation to stress focuses on sexual selection, that is, on issues of competition within each sex by display to the opposite sex and between sex mate preference. Human attractiveness judgments occur in social behavioral domains other than sexual competition. Humans everywhere are immersed in a complex set of interactions with relatives and nonrelatives other than mates. Under the hypothesis that attractiveness reflects phenotypic and genetic quality, specifically the ability to withstand developmental stresses, and the complimentary hypothesis that attraction reflects a preference for stress resistance during development, the general prediction is that human reciprocity and nepotism will be importantly based on physical attractiveness (Thornhill and Gangestad, 1993). If, as proposed here, physical attractiveness is a stress-resistance certification and thereby a marker of phenotypic quality, people are expected to have adaptations that use attractiveness information (1) in choosing social allies for reciprocal alliances: (2) for dispensing benefits to these social allies: and (3) for dispensing nepotism to relatives. The reproductive benefits of aid in the form of reciprocity or nepotism depend on survival of the beneficiary of that aid. According to our view, physical attractiveness was a reliable and consistent predictor of individual survival in human evolutionary history because attractiveness reflects phenotypic quality. Thus, we anticipate that people will prefer to interact socially with individuals exhibiting bodily health certificates (e.g., attractive secondary sexual traits and low asymmetry). Freeland (1976) suggests that, in nonhuman primates, health of a potential social ally should be an important factor influencing the willingness of individuals to engage in social interactions, and that healthy individuals should be preferred social companions. We agree with Freeland but add that physical beauty may be an important

STRESS AND HUMAN BEHAVIOR

331

means of assessing the health-related reproductive value of social allies to self. Abundant evidence indicates that people show greater interest in being associated socially with physically attractive people than with unattractive people and that people treat attractive people with favoritism (reviewed in Alley and Hildebrandt, 1988; Eagly et al., 1991; Jackson, 1992). Physically attractive people are viewed as having greater social skills and competence, as having higher status, and as being more successful, interesting, educated, and intelligent than unattractive people. That is, they are viewed as having more social assets and as better social allies. Not only are attractive people perceived as having these social assets, they are treated as if they have them. Moreover, numerous studies reveal that the physical attractiveness of a person has a major positive influence on the number and stability of friendships with same-sex individuals. No study has tested for positive effects of facial symmetry or beauty of secondary sexual traits on people’s interest in and selection of potential allies for reciprocity. The hypothesis that people base social interactions on physical attractiveness because it is related to phenotypic and genotypic stress resistance ancestrally may apply to discriminative nepotism patterns in people. It is likely that stress tolerance and resistance was a major determinant of survival and well-being of one’s genetic relatives during human evolutionary history. Thus, everything else being equal, we would expect people to direct more nepotistic benefits to attractive than to unattractive relatives. Studies have provided positive results bearing on the predicted relation between attractiveness and nepotism (see review in Alley and Hildebrandt, 1988; also Langlois et al., 1995). Mothers of children with facial anomalies rate their children more negatively than do mothers with facially normal children. Parents have higher expectations for attractive than for unattractive children. Also, mothers of more attractive newborns are more affectionate toward their babies than are the mothers of unattractive newborns (Langlois et af.,1995). All of these patterns suggest differential investment by parents in offspring, with attractive offspring receiving more. We know of no study that has directly examined the effect of body symmetry or other markers of attractiveness on wealth inheritance or any other form of nepotism.

E. CONCLUSIONS No finding in social psychology is more robust and replicable than the finding that individuals respond more positively to physically attractive than to unattractive persons. Indeed, attractiveness may be the single best predictor of human social preference. It clearly affects how one is treated by others. Attractiveness also appears to affect how one treats others.

332

RANDY THORNHILL AND BRYANT FURLOW

Attractive individuals seem to be very discriminating socially-this is seen in the positive assortment on the basis of looks both in mateships and friendships (Jackson, 1992); it is also seen in the reduced investment of attractive men in romantic relationships (Gangestad and Thornhill, 1997). Only relatively recently has research focused on the most fundamental and intriguing question about the power of looks in human everyday life. What is the evolutionary basis of the priority placed on looks in the human mind? There is increasing evidence that human physical attractiveness and attraction owe their ultimate existence to Darwinian selection for displaying health and for assessing health in others. Health or phenotypic quality can be appropriately viewed as synonymous with stress tolerance and resistance. Accordingly, physical attractiveness reflects (1) developmental stability, or the ability to cope with environmental andfor genetic perturbations during development; and (2) expressions of secondary sexual traits that appear to mark immunocompetence and the ability to cope with sex hormones. There is need for more experimental research on human attractiveness in relation to symmetry and secondary sexual traits. Singh (1993,1995), D. Perrett and M. Burt (personal communication, July 1996), and L. Mealey, R. Bridgstock and G. Townsend (personal communication, November 1996) have made pioneering studies in this regard, but most studies are observational and correlational with only statistical control of potential confounds.

111. PARENT-DAUGHTER RELATIONS A N D WOMEN’S SEXUAL BEHAVIOR

There is a voluminous literature dealing with the effects of parental relationship behavior on women’s romantic relationship attachment. There is also a large literature dealing with parental behavior in relation to the development of women’s sexuality (Belsky et af., 1991). Evolutionary psychologists have reviewed this literature in the light of their evolutionary developmental theory of women’s behavior pertaining to relationships and sexuality (Draper and Harpending, 1982; MacDonald, 1985, 1988, 1992; Surbey, 1990; Belsky, 1990; Belsky et af., 1991; Cashdan, 1993). According to this theory of women’s development, which is based on life-history theory (Roff, 1992), the parents’ ability and willingness to invest in each other and their children affects the sexual developmental pathway followed by a daughter. Parents and daughters are viewed as strategists. Parents’ patterns of child rearing prepare daughters for the adult social environment they will be likely to face. Daughters internalize the rearing setting and learn the emotional and relationship skills that are suitable for the probable environment they will face in adulthood. Factors such as marital discord, high stress in

STRESS AND HlJMAN BEHAVIOR

333

the home, inadequate resources, and father absence or emotional distance are developmental stresses that cue in daughters early puberty and sexual activity, short-term and unstable mateships characterized by exploitation, emotional coolness and insensitivity, and limited parental care. On the other hand, interparental harmony, adequate resources, father presence and involvement in child rearing cue later puberty and restricted sexual activity, stable relationships, greater parental investment, and emotional warmth and positive affection. Daughters with the former set of traits are viewed as best suited in terms of reproductive success for an environment in which resources will be relatively limited. The latter set of traits equips daughters reproductively for a social environment in which resources will be abundant and consistently available. Women in environments of reliable resources are expected to perform best in terms of reproductive self-interest by positive and enduring interpersonal, social relations (including romantic ones) based on trust. Sexual restraint may allow women in resource-rich environments to secure a mate with resources because of the positive influence of women’s restricted sexual history on men’s perception of high paternity reliability (Low, 1989). Women in resource-poor environments are served by a mistrustful and opportunistic relationship style. Women use their sexuality to access resources from men (e.g., Symons, 1979; Buss, 1994). In stressful environments, early maturity would lead to quicker access to resources that are missing as a result of limited parental investment. Given that individual men will be less able to provide sufficient investment in resource-poor situations, women’s weak attachment with a mate would adaptively promote multiple sex partners and thus a greater total amount of male-provided resources than if they were faithful to a single mate. The massive literature on female child development shows a remarkable fit to this evolutionary view of sexual and adult relationship behaviors of daughters being dependent on upbringing. This literature comprises many detailed studies in the West spanning several decades, as well as crosscultural studies, including hunter-gatherer societies (see previous references: also Low, 1989; Hurtado and Hill, 1992; other papers in Hewlett, 1992). This view sees female development as a conditional strategy with tactics being selected by an adaptation that reads current rearing conditions pertaining to resources and reliability of parental investment and forecasts the future social environment based on rearing environment. The different tactics involve different schedules of reproductive and somatic effort and different allocations of reproductive effort to mating and parenting. Because an unrestricted sexual history is a detriment for women in securing a mate with abundant resources-such males are choosy about females in whom they invest (e.g., Symons, 1979; Buss, 1994)-it is expected that early sexual activity and a sexual history of many partners will often track

334

RANDY THORNHILL AND BRYANT FURLOW

a young woman into a promiscuous future. Thus, there may be considerable continuity in the benefits and costs of unrestricted sexuality throughout the life history of a woman in a low-resource setting. Although there is reason to believe that a girl’s experience in a rearing environment with a low- versus high-resource base may have long-term consequences on her sexuality, it does not follow that women’s sexuality will show no flexibility in terms of resource variation in the adult environment. The adaptive hallmark of condition dependence is the ability to shift tactics to increase reproductive success if circumstances encountered are promising in this regard. Thus, it is anticipated that women will have the ability to assess current circumstances of male investment potential and make adjustments in sexuality that will best capitalize on current conditions. Research by Cashdan (1993) suggests women may have this lability. She has shown that the pursuit of short-term versus long-term romantic relationships by women (Buss and Schmidt, 1993) depends on whether a woman perceives that men in the environment will invest and commit themselves in romantic relationships. Compared with women who perceive that they are in an environment in which men will invest, women who perceive that they are in an environment where men are not likely to invest act and dress in a more sexually provocative manner, and use copulation to attract desirable men. This variation can be understood as a facultative response of an evolved female sexual psychology designed for an output of sexual restraint (thus giving cues of paternity reliability) when investing males may be accessible, and less sexual restraint (to access material benefits) when each male can or will invest little (Cashdan, 1993). However, it is unclear from Cashdan’s research how much development versus current environment affected the sexual differences among the women. Father-daughter relations have figured importantly in the evolutionary theory of women’s sexual development. Women developing in settings of reduced paternal investment associated with divorce, marital strife, or father’s death show earlier onset of menarche and sexual behavior and greater numbers of sexual partners. Reduced paternal investment may have impacted on daughters severely throughout human evolution. Ache Indian girls without fathers are more likely to have infectious disease, suffer mortality, and be captured by men in raids (Hurtado and Hill, 1992). In the Ache context, the advantage to girls without fathers of speeding up sexual maturity and beginning sexual activity early is crystal clear. This literature’s focus on the fathers’ effects on daughters’ sexual development does not imply that mothers don’t influence daughters (see Patterson, 1986). Nor does it mean that maternal investment is insignificant in influencing daughters’ life-history decisions about life pursuits. However, there is good reason to believe that adult males provided certain essential parental

STRESS A N D HUMAN BEHAVIOR

335

resources more consistently than did females in human evolutionary history. Presumably, this is the reason that male parental investment evolved in humans in the first place. Based on human sex differences in parenting behavior, the resources provided by ancestral males were meat from hunting, teaching hunting skills to sons, and physical protection (e.g., Hewlett, 1992). Even in cases in which mothers reject, neglect, or desert offspring, an underlying cause may involve absent or unreliable paternal investment. It is clear that social support from fathers is important in decisions by mothers to kill offspring rather than invest in them (Daly and Wilson, 1988). We hypothesize here that there is an additional component of women’s sexuality that may be influenced significantly by rearing environment: variation among women in sexual arousal. Among women, sexual response is highly variable (Fisher, 1973). For example, some women are totally anorgasmic, others are anorgasmic during copulation, and others show frequent copulatory orgasm. The evolutionary theory of women’s sexuality discussed earlier includes the following tactical components: time of sexual maturation/puberty, onset of sexual activity, stability of the pair-bond and sex partner number, and emotional coldness/warmth toward the mate. Sexual arousal is likely to be a tactical component of female strategy because of its connection to pair-bonding and sperm use. Specifically, we suggest that limited and highly selective orgasmic response is an evolved facultative expression of female development in the stressful environments of limited male spousal and parental investment that would complement the other sexual tactics shown by women in such environments. Below we examine first the hypothesis that female sexual arousal is a female choice adaptation. We then more fully discuss variation in female sexual strategy and the relationship between female sexual response and rearing environment, and then we specifically treat the daughter-father developmental environment. A. FEMALE SEXUAL AROUSAL AS FEMALE CHOICE

Female sexual arousal can be viewed as a female choice adaptation with multiple outputs that range from no arousal to orgasm. There is considerable evidence that the range of outputs reflects female choice. Many have commented on the connection between women’s sexual arousal and female choice. When female choice is circumvented, as in rape, women do not show sexual arousal. This is a reason that women often label rape as a violent act rather than a sexual act, whereas men view sexually coercive sex as a sexual experience (Thornhill, 1994). Also, women’s sexual interest and arousal are importantly tied to a man’s investment in the relationship: female mate preference is also importantly predicted by the ability and willingness of a man to invest (Symons, 1979; Grammer, 1993; BUSS,1994).

336

RANDY THORNHILL A N D BRYANT FURLOW

Moreover, women’s copulatory orgasmic frequency is significantly predicted by marital happiness and husband’s income and status, and by the woman’s own socioeconomic status with low frequencies in women of lower socioeconomic status (Fisher, 1973). Finally, as mentioned earlier, female copulatory orgasm is positively correlated with the developmental stability of the mate, a pattern predicted by the hypothesis that women are selectively bonding with and retaining preferentially the sperm of men of high phenotypic and genetic quality (Thornhill et al., 1995; but see Baker, 1997). Women are more sexually interested in and aroused by men capable and willing to invest and by men of high phenotypic and genetic quality. Thus, all components of female sexual arousal from absence of arousal to copulatory orgasm may be strategically related to female choice of mate and sire. Female orgasm could interfere with women’s sexual strategy when women are in settings in which limited sexual and emotional commitment to each man is in their evolved interests, that is, when men have reduced resources. Said differently, like all traits, female orgasm has costs in addition to benefits. The costs of orgasm may include strong pair bonding with a mate in environments in which short-term mateships are optimal. Oxytocin, a hormone known to have effects on social bonding, is released in large quantities during female orgasm (see Thornhill et al., 1995, for evidence and discussion). If oxytocin plays an important role in differential bonding of women to men, it may lead to reducing female interest in male nonmates and thereby limit resources females can obtain by exchanging sex for resources from multiple mates. Females with multiple mating partners could still benefit from orgasm’s sperm retention function by limited and highly selective orgasm in relation to male genetic quality. Also, women appear to control the fertilizing potential of ejaculates in ways other than by copulatory orgasmic upward suction. First, masturbatory, that is, cryptic, orgasm may be important because it can result in acidic cervical conditions that are hostile to subsequently placed ejaculates (Baker and Bellis, 1995). Second, the timing of copulation in relation to ovulation may allow women some control over which male sires offspring. Strategic faked orgasm may be especially common in women pursuing resources from multiple sex partners. Romantically involved women who are flirtatious with men other than their usual partner seem to fake more orgasms (Thornhill et al., 1995). The main idea here is that limited and highly selective female coital orgasm is predicted in women who live in environments in which resources are limited or unreliable. This form of female sexual response would compliment the female’s other traits that have been viewed by others as strategic in such settings. In a resource-rich, high-male investment setting, female

STRESS AND HUMAN BEHAVIOR

337

orgasm will bond the woman with her providing man, limit infidelity, and result in his sperm being used. This is not to say that female infidelity is uncommon by women paired with providing males. During infidelity, these women may use orgasm for selective sperm retention. There is evidence that women’s extrapair copulations result in higher copulatory orgasm frequency than their inpair copulations (Baker and Bellis, 1995). SEXUAL AROUSAL B. FATHER-DAUGHTER RELATIONS A N D FEMALE There has been some in-depth research on the relationship between fathers’ and mothers’ behaviors and presence and daughters’ sexual arousal. Most of this was inspired by Freudian psychoanalytic theory, which sees daughters desiring sexual experience with father and thus competing sexually with mother. Particularly salient in Freudian theory is the need for women to resolve incestuous desires for father in order to be orgasmic during copulation. Copulatory orgasm in Freudian terms is the marker of an emotionally mature and stable woman, and the more orgasmic a women is during mating, the more mental health she possesses. We will not provide a full critique of Freud’s views of women’s sexuality. We do point out that evolutionary knowledge of mind design implies that Freud’s view is false: women will not have evolved to desire father as a mate (Thornhill and Thornhill, 1984). At any rate, the literature conains data bearing on the connection between women’s sexual arousal and the parent-daughter relationship even though the data were not collected to examine Darwinian hypotheses. Fisher (1973) studied a sample of about 300 married, middle-class women = 25). The women completed ranging in age from 21 to 45 years questionnaires about the parent-child relation and parental behavior during childhood. Separate questionnaires were used for mothers and fathers. The questionnaire by Roe and Siegelman (1963) examined parental behaviors such as protective, demanding, rejecting, neglecting, casual, loving, and punishment. The women rated each parent on a five-point scale for each variable. The questionnaire also allowed women to judge their relationship with each parent. The women separately provided tape-recorded oral narratives of their impressions of the parent-child relationship, and included any separation between parent and child (for each parent). The women self-rated their consistency in attaining orgasm during sexual intercourse on a six-category scale (1 = always, 6 = never). Retesting of the orgasm rating over intervals of 3 to 7 days was highly significant. There are several findings from Fisher’s research that are relevant to the prediction that resource-poor childhood environment will cue less sexual arousal in women. Parental loss or absence was examined in relation to

(x

338

RANDY THORNHILL AND BRYANT FURLOW

low- and high-orgasmic women. Childhood separations from mother were essentially absent in the sample and too infrequent for analysis. References to separation from father were more numerous. The low-orgasm consistency women exceeded those with high consistency in the frequency with which they described their fathers as having been dead, separated, or absent during their childhood. Although not an ideal measure of resource limitation during childhood, it is likely that, all else being equal, women reporting father separation grew up in homes with fewer resources than women reporting no father separation. In the questionnaire data that systematically evaluated the women’s attitudes toward their parents, there were differences dpending on sex of parent. There was no statistically significant relationship between a woman’s recall of how her mother had treated her and that woman’s ability to attain orgasm or her orgasmic consistency. Some interesting patterns emerged, however, in the women’s recall of how father had behaved. A statistically significant negative correlation was observed between orgasm consistency and casualness of paternal behavior. Orgasm consistency was lower in those women who perceived the father as conforming to the following paradigm of a casual parent from Roe and Siegelman’s questionnaire (1963, p. 357): They [high-casual parents] will be responsive to him [the child] if they are not busy about something else. They d o not think about him or plan for him very much, but take h i d h e r as a part of the general situation. They don’t worry much about him and make little definite effort to train him. They are easygoing, have few rules, & do not make much effort to enforce those they have.

Demanding fathers are at the opposite end of the continuum of casualdemanding fathering behavior. Demanding refers to the father’s imposition of strict rules and demands of obedience, whereas casualness refers to the father’s easygoing manner and the few rules that were rarely enforced. There was a significant positive correlation between demanding paternal behavior and women’s orgasm consistency. Overall, then, the greater a woman’s coital orgasm consistency, the less permissive and the more controlling she perceives her father to have been. The casualness-demanding dimension may tap importantly into paternal investment. The value of the daughter to the casual father is probably less than the value of the daughter to demanding fathers. It would appear that demanding fathers are investing more time and other parental effort into daughters. That casual fathers are less investing in daughters is supported by Fisher’s own conclusion (p. 263): “. . . these behaviors on a [casual] father’s part could be viewed as being thinly separated from indifference and lack of genuine concern about his daughter’s welfare.”

STRESS AND HUMAN BEHAVIOR

339

The oral narratives of high- and low-orgasmic women depicted their fathers significantly differently. Low-orgasm consistency women, more frequently than high-orgasmic women, described their fathers as unavailable for a substantial or consistent relationship with them. The closenessdistance factor did not show this pattern in the recollections of the women about their mothers’ behavior. The narrative aspect of Fisher’s study and the results on casualnessdemanding behaviors of fathers converge in support of the conclusion that fathers of low-orgasmic women invest less in their daughters than fathers of high-orgasmic women. Mothers’ behavior, however, was unrelated to daughters’ sexual responsiveness during copulation. Also consistent is the finding in many studies in diverse cultures (see Fisher, 1973, for review) that women of low socioeconomic level consistently report significantly lower copulatory orgasm frequency than women of the middle class. Paternal investment is reduced and often absent in lower socioeconomic settings. Another component of Fisher’s research is also consistent with our view here. We have suggested that low-orgasm women are those who strategically have less emotional commitment and sensitivity in mateships and less stable pair-bonds. Fisher’s results on women’s fantasies and concerns in relation to orgasm consistency show that low-orgasm women, in comparison to high-orgasm women, are especially concerned about how transitory relationships in general are and how easily loved ones can be lost. Such concerns in low-orgasm women may reflect a perception of greater likelihood of and preparation for mate rejection or desertion. Although Fisher’s various results discussed are supportive, they are only moderately so. More research is needed to examine the relationship between women’s sexual arousal and resource level and parental involvement in the home of origin. Measurement of resource availability and how parental behaviors actually relate to parental investment are critical. Women’s recall of conditions in the home could be cross-checked by other observers. Also, longitudinal studies of girls who subsequently become sexually active women and who grow up in homes differing in parental investment could address causal factors. The hypothesis proposed is one based on a conditional expression of a sex-specific female sexuality adaptation. However, genetic differences between families producing low- versus high-orgasmic women could explain Fisher’s results as well as the socioeconomic pattern in orgasm. Studies of female twins who are separated at birth and adopted by families differing in the variables of interest (resource level, stability of marriage, father presence/absence) could provide useful data to separate the heritability of sexual arousal from condition-dependent effects on sexual arousal.

340 C.

RANDY THORNHILL A N D BRYANT FURLOW

CONCLUSIONS

There is little doubt that certain important aspects of women’s sexuality and related heterosexual relationship behavior-age of puberty and first sex, number of sex partners, stability of consortships, interpersonal attitude toward mate (trust, sensitivity, interest, support)-are correlated with factors that would have reflected resource availability and reliability of parental investment, especially father’s investment, in the rearing environments of daughters in human evolutionary history. Women’s sexuality seems to show two distinct modes, each composed of tactics that combine to form a sexuality that is consistent with functioning sexually in either low- or high-resource environments. Women raised under conditions of resource stress, that is, limited resource base and parental investment, show early age of maturity and sexual intercourse, more sexual partners, more infidelity, less enduring pair-bonds, and a more opportunistic, insensitive, and exploitative orientation toward a mate. These tactics serve women’s interests in an environment in which it was historically adaptive to pursue resources from multiple mates. To the suite of tactics that are functional for women raised in limited resource situations we have added the tactic of reduced and highly selective sexual arousal, which we have suggested would promote pursuit of multiple partners by reducing the emotional commitment shown by a woman toward each partner. The suite of sexual tactics shown by women reared in conditions of resource availability and consistent parental investment also would fit them to that adult environment. Accessing investment from single males would be promoted by restricted sexual history, reduced infidelity, and greater sensitivity and trust toward the mate. Orgasmic consistency may be an important component of sexuality of women in environments in which single males can provide sufficient resources for successful female reproduction. The data bearing on the connection between orgasm consistency in women and limited resources during development are far from convincing, but are all consistent, from the father absence or disinterest data to the concerns about loss of relationships shown by these women. Also, the effect of socioeconomic level on women’s arousal corroborates the same trend. As Fisher emphasized, more research needs to be done here, but there is consistency in the data collected by different methods and investigators. Fisher and others before him looked at many variables in attempts to understand female copulatory orgasm variation. In all this literature, the associations between orgasm frequency and the father-child relationship and other measures of resource stress during upbringing emerge as the most robust and consistent pattern. Note that the argument we have provided makes no value judgments about orgasmic capacity of women. Even a superficial survey of the female

STRESS AND HUMAN BEHAVIOR

341

sexual arousal literature reveals that many workers assume that female sexuality is incomplete without copulatory orgasm, and women’s sexual competence and general well-being are correlated with frequency of coital orgasm (e.g., Fisher, 1973). This value system is based on a male sexuality model: copulation equals orgasm. As Symons (1979) pointed out, women’s more conditional sexual response often is seen as repressed from the ideal of male sexual response. There is no way that our approach could inform a value system based on orgasm as better among those who are naive about the naturalist fallacy, because we are saying that sexual arousal patterns of women from no arousal to orgasm are serving women’s evolved interests. Of course, what has evolved or is otherwise natural can never be used logically as a source of moral guidance. Those who define right and wrong in terms of what is natural commit the naturalist fallacy.

Iv. POSTPARTUM DEPRESSION Upon reaching reproductive maturity, human females become significantly more likely than males to suffer serious psychological depressions (Nolen-Hoeksema, 1990). This is not the case before puberty, and some evidence even suggests that male children experience a greater incidence of depressive symptomology than do girls (Nolen-Hoeksema, 1990). Traditional feminist interpretations of this increased vulnerability to depression among reproductively mature women generally posit that psychological stress due to social gender inequities are to blame, while biomedical researchers have typically conducted studies based on the assumption that sex hormones associated with reproductive status are at the root of women’s greater incidence of psychological suffering. Such proximate-causation approaches do not address the evolutionary significance of psychological pain itself, however, and thus lack an important component of any coherent conceptual framework from which predictions about the nature of psychological pain can be generated and tested. In this section, we propose and preliminarily test the hypothesis that one type of female depressionpostpartum depression (PPD)-may be an evolutionary psychological adaptation for discriminative maternal solicitude, encouraging mothers to cease investment in newborn offspring under circumstances in which the offspring would have had low reproductive value in ancestral environments. A. PSYCHOLOGICAL PAIN

AS A N

EVOLUTIONARY ADAPTATION

Throughout much of this century, researchers treated emotions as though they were little more than disruptions of normal psychological functioning

342

RANDY THORNHILL A N D BRYANT FURLOW

(see Nesse, 1990). Many behavioral biologists and psychologists, especially those informed by a neo-Darwinian perspective, have since rejected this assumed meaninglessness of emotions and have begun in recent years to search for the functional significance of emotional states (e.g., Daly et al., 1985; Leeper, 1948; Nesse, 1990; Thornhill and Thornhill, 1989; also see Hamida, 1996). Yet, many biomedical researchers have continued to treat depression as if it were always a “disorder,” a pathological malfunctioning of neurological hardware rather than a fitness-buffering reaction to stress (such researchers tend to emphasize the proximate causation of depressions only, typically focusing on depression’s hormonal or other physiological correlates; e.g., Goodwin et al., 1976; Livingston et al., 1978; Handley et al., 1977; Ballinger et af., 1979; Okano and Nomura, 1992). This assumption about the nature of depression has shaped doctors’ attitudes toward proper treatment (usually, prescribing antidepressive drugs; Nesse, 1990 or manipulating hormone levels; Gregoire et al., 1996), and has led to an inaccurate presentation in the medical literature of “biological” and “environmental” (or “psychosocial”) theories of psychological pain as alternative, mutually exclusive explanations. Evolutionary psychologists, on the other hand, have proposed that psychological pain reflects evolved responses to environmental cues related to impacts to a person’s inclusive fitness-that is, to evolutionarily relevant sources of stress (Thornhill and Thornhill, 1989; Nesse, 1990). Thornhill and Thornhill (1989) argue that psychological pain was selectively favored through evolutionary time because it increased victims’ lifetime reproductive success and inclusive fitness by forcing “an assessment of the circumstances surrounding social problems in the lives of individual humans” (Thornhill and Thornhill, 1989), and that depression is proximately caused by dramatically fitness-reducing social events. Hence, psychological pain is evolutionarily analogous to physical pain; pain receptors are found only where they are useful in terms of correcting evolutionarily recurrent fitness impacts (Alexander, 1986; Thornhill and Thornhill, 1989), and are designed to demand immediate attention. Evolutionary predictions about the proximal causes of different types of psychological pain vary, reflecting the particular challenges to fitness presented by different social circumstances during human evolution. If a given form of depression is an evolved strategy for dealing with ancestral fitness impacts, and occurs when current environments contain cues that trigger those evolved states of psychological pain, then it should be possible to make predictions about the social environment of the depressed person. For instance, in a review of the largest investigation of victims’ anguish following rape (McCahill et al., 1979), Thornhill and Thornhill (1989) predicted and confirmed that women of reproductive age suffer greater psycho-

STRESS AND HUMAN BEHAVIOR

343

logical distress following rape than do pre- or postreproductive-age victims, reflecting the increased risk of pregnancy by an unwanted and uninvesting mate. Another evolutionarily relevant cost of rape is the disruption of support from an investing mate. Victims’ mates often view claims of rape as denials of what was actually infidelity, and men frequently abandon women after a sexual assault (McCahill et al., 1979). Hence, whereas McCahill et al. expected that increased physical trauma during rape would correlate positively with subsequent psychological pain, the evolutionary psychological hypothesis of postrape anguish generated the opposite prediction: that physical injury or signs of brutalization would constitute strong evidence of actual rape victimization rather than consensual sexual infidelity to mates, and would therefore correlate with lower levels of psychological pain suffered by rape victims. A single episode of rape constitutes a less serious paternity threat than multiple copulations during sexual affairs, and is less likely to result in questions about fidelity in the minds of mates. The counterintuitive prediction from the evolutionary hypothesis was supported by the data presented by McCahill et al. (1979; Thornhill and Thornhill, 1989): physical trauma during rape negatively correlates with the amount of psychological pain suffered by rape victims. (See Thornhill, 1997, for a recent discussion of the evolutionary psychology of rape victim trauma.) Postrape anguish is but one of the many forms of psychological pain suffered by women. From an evolutionary perspective, it should be expected that women would suffer a greater incidence of psychological pain, especially in response to events affecting their reproductive status. Women’s lifetime offspring number is physiologically constrained such that each reproductive event constitutes a larger proportion of a female’s lifetime reproductive potential than a male’s.

B. POSTPARTUM PSYCHOLOGICAL PAIN Maternal postpartum psychological pain was first described in the West by Hippocrates (Jones, 1923). Many researchers believe that western women are more likely to suffer serious depression soon after childbirth than at any other time in life (but see O’Hara er al., 1990). In contrast with pregnancy, which is associated with very few psychiatric conditions (Paullekhoff, 1992), depression during the postpartum period is the most common serious psychiatric syndrome among western women (Hamilton, 1988; Inwood, 1985), and appears to be more common than depression at other times in life (Inwood, 1985; Paffenbarger, 1961; Whiffen, 1992). The prevalence of PPD in Arab and Ugandan populations match that of western nations (Cox, 1983; Ghubash and Abou-Saleh, 1997). Women with previous histories of psychiatric disorders are at significantly more risk of postpartum

344

R A N D Y THORNHILL A N D BRYANT FURLOW

psychiatric hospitalization (Inwood, 1985). Because postpartum psychological pain in these women cannot be disentangled from their preexisting psychiatric illnesses, we will not consider this population in our review of PPD. Further, because male postpartum depression does not share the physiological etiology of maternal PPD (see Perry, 1992), we consider it to be a distinct psychological phenomenon (although perhaps a convergent one, shaped by related selective pressures during human evolution). Since very little is known about paternal PPD, we deal only with maternal PPD in this chapter. Feminist perspectives on PPD have emphasized the distress inflicted on women during child birth in male-dominated medical settings in which “birth is manipulated to suit institutional requirements and women are pressured to accept passive and dependent roles” (Oakley, 1980). However, empirical evidence fails to support this view’s prediction that a greater proportion of hospital births than home births in the West leads to PPD (Pop et al., 1995). Maternal postpartum dysphoria varies along a continuum of symptomology from ubiquitous but mild and transient “baby blues” to “postpartum psychosis.” Baby blues occur from a few hours to a few days after giving birth, and involve bouts of tearfulness and relatively mild levels of anxiety (Hamilton et al., 1992). More serious is PPD, a clinical depression involving longer lasting feelings of grief, guilt, irritability, tearfulness, feelings of hopelessness and failure, and apathy or hostility toward the newborn (Hamilton et al., 1992). Typically, PPD is mild compared to nonpostpartum clinical depressions and involves a markedly lower incidence of suicidal ideation (Whiffen, 1992). Serious, clinical PPD symptomology often begins after a few weeks postpartum and, unlike the baby blues, lasts up to several months (Hamilton et al., 1992). Unusually serious cases of PPD are sometimes called “postpartum psychotic depression” if onset is acute immediately after childbirth, but in diagnostic practice there is considerable overlap and confusion about such differentiations (Hamilton et al., 1992; Paullekhoff, 1992). Postpartum psychosis (PPP) is typified by delusional and obsessive thoughts about failure, an inability to love or care for the newborn, and guilt resulting from these feelings (Herzog and Detre, 1976; Paullekhoff, 1992). As Daly and Wilson (1995) point out, however, the content of these “delusions” is not always particularly illogical (see also Herzog and Detre, 1976). Beliefs in many cultures around the world that deformed newborns are demons or demonic progeny are irrational and may be considered psychotic by western psychiatrists, but the resulting infanticidal behavior-while morally reprehensible to us-is nevertheless adaptive, evolutionarily speaking (Daly and Wilson, 1995). We will treat PPD and PPP as different points along the same

STRESS AND HUMAN BEHAVIOR

345

continuum of psychological distress (collectively referred to by us simply as “PPD”), because even if real psychiatric differences exist between the two diagnoses, they give rise to similar fitness-affecting behaviors (most notably, neglect or abuse of newborns). Immediately after birth, a mother is apparently at special risk of depressive symptomology because of reduced levels of the stress-mediating corticotropin-releasing hormone (CRH; Magiakou et al., 1996), rendering her vulnerable to sources of stress. Deficits in corticotropin-releasing hormone are not a sufficient explanation of PPD’s origins until we answer the question of why this postpartum vulnerability to stress (with its often dramatically fitness-affecting results) was not expunged from human psychological design by natural selection during human evolution. One intuitive and potentially popular answer to this question is that the female human body plan is somehow physiologically constrained, such that no escape from vulnerability to PPD has ever emerged throughout the course of human evolution. For instance, CRH levels may regulate prenatal and delivery events (Karalis et al., 1996), and until postpartum maternal hormone levels are differentiated from a prenancy-typical profile, CRH may confound postpartum behavioral adaptations. A more readily tested alternative hypothesis is that natural selection acting in ancestral populations favored a maternal vulnerability to stresses after giving birth. INVESTMENT BEHAVIORS I N EVOLUTIONARY CONTEXT C. PARENTAL As stated earlier (Section II,C), parents do not invest maximally in all of their offspring. To understand the evolved psychology of variable parental investment, we must clarify the selective pressures responsible for shaping human parental psychology. Parent-Offspring Conflict Theory was the first systematic attempt to delineate the selective forces that shape interactions between parents and their offspring in modern, neo-Darwinian terms (Trivers, 1974). In this section, we introduce Trivers’s theory, and then describe the more recent, complimentary model of differential parental solicitude. Finally, we place the study of PPD in this evolutionary framework and present predictions derived from our hypothesis about PPD’s evolved function. These predictions are tested against the medical literature in the following section. While extensive and representative, our review of the literature is almost certainly not exhaustive. Trivers argued that the nonidentical genetic self-interests of diploid parents and offspring leads to natural selection favoring behavioral strategies of mutual exploitation (Trivers, 1974). To some degree, for instance, offspring who exaggerate their nutritional needs will be favored over siblings who do not, even if their exaggeration reduces their parents’ long-term

346

RANDY THORNHILL A N D BRYANT FURLOW

fitness. Parental psychology and reproductive physiology, on the other hand, are expected to be designed to counter exploitation by offspring (Haig, 1993; Peacock, 1991) and to favor investment in those offspring that have higher probable eventual fitness (an offspring’s “reproductive value”). Hence, contrary to popular notions of automatic and unconditional motherinfant bonding, evolutionary theory leads to the prediction that parental investment in offspring will be preceded by an evaluation of the newborn’s reproductive value (Daly and Wilson, 1995). This line of thinking-termed “differential parental solicitude” by Daly and Wilson (e.g., 1988)-has been expanded in recent years, and now includes an explicit model of maternal bonding psychology to replace the simpler, naive version of unconditional, “automatic” bonding, which still enjoys wide popularity (Daly and Wilson, 1995; Eyer, 1994). The evolutionary bonding model posits three stages of maternal attachment to a newborn (Daly and Wilson, 1995):initial assessment of the baby’s reproductive value, followed by discriminative emotional attachment to the baby, followed in turn by a gradual deepening of appreciation and attachment to that individual as his or her reproductive value increases with age. Infanticidal behaviors frequently correlate with low infant reproductive value (Daly and Wilson, 1988), and involve a failure of the second step of maternal attachment, as described above. In many cultures, including those of the industrialized West, mothers react to cues of low infant reproductive value with lethal neglect or active infanticide; in other words, mothers make a negative assessment of offspring reproductive value and act on evolved motivational states that encourage cessation of investment in the offspring (Daly and Wilson, 1995). Infants’ deformities are grounds for socially accepted infanticide in many cultures, for instance (Daly and Wilson, 1988). This infanticide is an example of what Moiler (1997b) calls developmental selection against developmentally unstable offspring. This postnatal assessment of offspring fitness may allow a more subtle evaluation than prenatal maternal assessments, and hence, postnatal infanticide may be a more fitness-enhancing strategy of investment withdrawal than adaptive (spontaneous) abortion. Human mothers represent their offspring’s main or only source of nutrition for the first 6 months to 2 years (Wood, 1994). Since lactation is significantly more metabolically costly to the mother than gestation (Miller and Huss-Ashmore, 1989; WorthingtonRoberts et al., 1985), infanticide may constitute a strategic maternal investment withdrawal at the beginning of the truly costly stage of reproduction. It is important to note that the first stage of maternal attachment (assessment of reproductive value) includes factors in a mother’s life that are extrinsic to the child’s inherent reproductive value, because there are circumstances under which even the healthiest newborn is unlikely to survive

STRESS AND HUMAN BEHAVIOR

347

to reproductive maturity. Hence, if PPD has been designed by natural selection to motivate strategic investment withdrawal from infants of low reproductive value, predictions other than higher incidence of PPD in mothers of ill or deformed infants can be generated and tested. We propose the hypothesis that PPD is a form of evolved psychological pain, designed to encouarge maternal withdrawal of investment from offspring when doing so would have increased long-term maternal fitness in ancestral environments. We therefore expect that stress factors that would have related to the intrinsic and extrinsic reproductive value of a newborn during human evolution will be highly correlated with the incidence of maternal PPD. We derive six predictions from our hypothesis that we believe must be true if our hypothesis is correct:

1. Cues of compromised health and/or developmental integrity of the newborn will correlate positively with the incidence of maternal PPD. 2. Low levels of paternal investment will correlate positively with the incidence of PPD. 3. Low levels of social support from kin and social allies other than an investing mate will correlate postively with the incidence of PPD. 4. Cues of a compromised resource base (famine or poverty) will correlate with the incidence of PPD. 5. Increases in the incidence of infanticidal ideation by mothers will accompany PPD. 6. Cultures in which ritual displays of social support and paternal investment do not accompany childbirth will have an increased incidence of PPD. In the following sections, these predictions are tested against the existing English-language medical literature on PPD. 1. Infant Health and the Incidence of PPD

If postpartum depression is a psychological adaptation encouraging the cessation of investment in offspring of low reproductive value, then the most obvious correlate of PPD should be unhealthy infants. Indeed, mothers of infants at higher medical risk report higher levels of emotional distress and PPD symptomology, difficulty expressing affection toward their baby, and greater dissatisfaction with the levels of social support they receive (Bennett and Slade, 1991). Kumar and Robson (1984) found a significantly increased incidence of PPD in mothers of premature babies compared to mothers of full-term newborns. Premature infants would have had slim survival prospects in ancestral environments, so evolved discriminative parental investment mechanisms may assess these infants as possessing low intrinsic reproduc-

348

R A N D Y THORNHILL A N D B R Y A N T FURLOW

tive value despite recent medical advances that enhance the odds of survival for such newborns. Condition of the newborn negatively correlates with PPD (Hopkins etal., 1987). However, Davidson (1972) found no correlation between infant state and PPD. Medical complications during delivery (e.g., excessive bleeding or fever during delivery, emergency cesarean section) correlate significantly with PPD in some studies (Campbell and Cohn, 1991; Morgenshy, 1982; O’Hara et al., 1984; Paykel et al., 1980) but not others (O’Hara et af., 1982). Burger et al. (1993) found that women with severe complications during pregnancy are significantly more prone to suffer PPD, even after variables such as premature birth and neonatal hospitalization were controlled for. It appears possible that such pregnancy or delivery-related “emergency” cues can negatively affect maternal assessment of infant viability.

2. Social Support and P P D Low levels of social support are markedly related to a higher risk of PPD (Cutrona and Troutman, 1986; Nilsson and Almgren, 1970; O’Hara et al., 1983; Pitt, 1968; Richman et al., 1991; Spangenberg and Pieters, 1991). Even the additional social support afforded by volunteer “labor companions” from the local community reduces the risk of PPD in middleclass South African women (Wolman et al., 1993). Among Arab women, presence of a housemaid may reduce the risk of PPD (Ghubash and AbouSaleh, 1997). A review of mild PPD (cases not characterized as psychotic) reveals that a lack of social support is significantly correlated with higher rates of PPD (Cutrona, 1982). Workplace social support may be relevant to postpartum adjustment as well (Leathers et al., 1997). However, Hopkins et af. (1987) report no significant correlation between PPD and low levels of social support in the sample they studied. A recent review by Wilson et al. (1996) reports “fair” evidence for a PPD/social support relationship. Poor relationships of women with their mothers is related to their risk of experiencing PPD (Douglas, 1963; Kumar and Robson, 1984; Richman et af., 1991). These relationships may be viewed as indications of low levels of kin support, an important component of social support networks for new mothers. In adolescents, the correlation between poor relationship with own mother and PPD may be due to the fact that the emotionally or materially uninvesting mother is the new mother’s primary or only source of support (see Quijano and Cobliner, 1983).

3. Paternal Investment and Maternal PPD A perceived lack of support from a husband is associated with maternal PPD (Buchwald and Unterman, 1982; Unterman el af., 1990). Help with parenting duties, displays of interest in the newborn, and support for the

STRESS A N D HUMAN BEHAVIOR

349

mother correlate with a significantly reduced incidence of maternal PPD, whereas marital discord or lack of investment behaviors by the father correlate with a higher incidence of PPD (Campbell and Cohn, 1991;Collins et al., 1993; Gordon and Gordon, 1967; Kumar and Robson, 1984; Logsdon et al., 1994; Paykel et al., 1980; Richman et al., 1991; Sosa et al., 1980; Spangenberg and Pieters, 1991). Arab women with PPD are significantly more likely to report marital probkms than controh (Ghubash and AbouSaleh, 1997). Indeed, Close (1980) argues that the best treatment for PPD is extra affection, understanding behavior, and help with parenting from the husband (see also Hickman 1992). Barnett et al. (1996) report that among adolescent mothers, social support from their mothers or from the infants’ fathers was significantly associated with lower rates of PPD. Wilson et al. (1996) report “good evidence for association” between poor marital relations and PPD, in their recent review. In this section, we have emphasized the hypothesis that PPD constitutes a mechanism whereby maternal investment in a newborn is suspended because of low newborn reproductive value. Another purpose for PPD may be the demonstration of need to people likely to respond by changing the environmental or social conditions that are extrinsic to the offspring but nevertheless lower its reproductive value. In the only other evolutionary psychological analysis of PPD that we are aware of, it has been proposed that in addition to investment withdrawal from infants, defecting from univesting or coercive mateships is a primary function of PPD (Hagen, 1996). This display-component hypothesis posits that PPD forces interested parties to invest more in the mother and her offspring lest she abandon the infant. Hence, PPD may serve as an unconscious strategy for displaying need to a social support network via a threat of child neglect or infanticide, and the vulnerability to PPD may have been evolutionarily retained not just as an effective mechanism for strategic resource withdrawal from offspring, but also may have enhanced maternal fitness by forcing others to correct circumstances that lowered offspring reproductive value in ancestral environments. Symptoms such as weeping and irritability, if this is the case, are best viewed as social signaling (Hagen, 1996), while apathy or hostility toward the newborn reflect a resource-withdrawal function of PPD. It is interesting to note that while adaptationist explanations have been proposed for human grief displays in other contexts (such as the loss of a loved one), the honesty of grief displays is suspect because of the potential fitness benefits of exaggeration (Thornhill and Thornhill, 1989). Whether a display is a reflection of real need or whether it is a counterfeit performance is not easily assessed by onlookers. For displays to constitute an evolutionarily stable signaling strategy, they either must make assertions that are readily verified by signal receivers, who in turn reliably punish

350

RANDY THORNHILL A N D BRYANT FURLOW

exaggerations, or they must carry a physiological or fitness cost to the signal sender (a “handicap”; Grafen, 1990; Zahavi, 1977b). Weeping is a speciestypical correlate of human grief, and is an obvious component of PPD. Tearfulness in adult women is immunosuppressive (Labott et al., 1990; Martin et al., 1993); sad mood alone does not affect immune system functioning, but shedding tears results in depressed salivary immunoglobulin-A counts. Since immunological handicapping serves as a physiological enforcer of signal reliability in other signaling contexts (Folstad and Karter, 1992), we speculate that weeping may be a Zahavian social signal of need. Assuming that maternal tearfulness during PPD is physiologically similar to experimental subjects’ evoked tearfulness, immunosuppression may have limited the exaggeration of need during PPD weeping displays to social allies among our ancestors, leaving tears, on average, an honest indicator of a woman’s psychological pain and real need. Even small degrees of immunosuppression during the postpartum period may have had important consequences in human evolutionary history because of the greater impact of parasites than that existing in contemporary western societies. 4. Resource Base and PPD Women living in lower socioeconomic strata (Davidson, 1972), recently immigrated to a new nation (Zelkowitz and Milet, 1995), or expressing perceptions of poor housing quality (Paykel et al., 1980) experience a higher incidence of PPD than other women, but not a higher incidence of less serious forms of postpartum dysphoria (“blues”). Socioeconomic level involvement in PPD incidence may be confounded by some unmeasured variable such as a higher incidence of poor marital relations in lower socioeconomic stratum marriages. Women who express concern over finances or their husband’s employment reliability exhibit an increased incidence of PPD (Heitler and McCrensky, 1976), while women who are themselves employed experience lower rates of PPD than do women without jobs (Richman et al., 1991; Zelkowitz and Milet, 1995). Western housewives may perceive poor social support because they experience relative social isolation, caring for infants in the home (E. H. Hagen, personal communication). Unterman et al. (1990) state that “economic problems” constitute a risk factor for PPD. During the Great Depression, rates of PPD in the United States increased significantly, perhaps partly because of an increased incidence of unemployment or general employment insecurity (Wick, 1941). 5. Infanticidal Ideation and PPD

As stated earlier, women suffering PPD exhibit decreased levels of suicidal ideation compared to those suffering nonpostpartum clinical depres-

STRESS AND HUMAN BEHAVIOR

351

sions. However, a higher incidence of homicidal (infanticidal) ideation associated with PPD is reported in the literature (Beck, 1992; Wisner et al., 1994), as is an increased incidence of actual infanticides (Herz, 1992; Kumar and Marks, 1992). Infanticidal thoughts appear to be a major source of the guilt feelings suffered by women during PPD (Beck, 1992; Wisner et al., 1994), and may be more common than is reported. A number of authors report that women suffering from PPD are unable to feel love or maternal concern for their newborns (e.g., Affonso and Arizmendi, 1986; Kumar and Robson, 1984). 6. Cultural Traditions and PPD PPD appears to be very rare in traditional, nonwestern cultures (Harkness, 1987; Kelly, 1967; Kruckman, 1992; Stephenson et al., 1978). The traditionally recognized “amakiro” postpartum illness in Uganda, typified by mental confusion and a desire to kill the newborn, may represent an exception (Cox, 1978,1979). Domination by colonial cultures and resulting disruption of traditional native culture appears to result in the incidence of PPD similar to that of Westerners (e.g., East Africans and New Zealand’s aboriginal Maori; Harris, 1981; Webster et al., 1994). Apparently, people from traditional cultures are no less susceptible to PPD, given certain stresses, than Westerners are. Those cultures that observe traditional, ritual displays of social support to the new mother seem largely “immune” from PPD (Kelly, 1967; Kruckman et al., 1983; Upreti, 1979). Among the Nigerian Ibibio, for instance, women are placed in a “fattening room” after giving birth, while others take over their daily duties (Kelly, 1967). In China and Taiwan, a traditional postpartum custom of “doing the month” (to yueh; 30 days of rest) similarly demonstrates social support to the new mother, who is viewed as unusually vulnerable and in need of great physical and emotional support (Fried and Fried, 1980; Pillsbury, 1978). The Chinese and the Ibibio both apparently enjoy an absence of PPD, according to ethnologists (although it should be noted that ethnologists often have low sample sizes). It appears that displays of social support from a woman’s kin and allies can mitigate the stress leading to PPD, and that people around the world have noted a mother’s special postpartum vulnerability and responded with cultural practices that mitigate relevant sources of stress. D. CONCLUSIONS Our review of the medical literature on postpartum depression provides preliminary, correlational support for our hypothesis that PPD constitutes an evolved mechanism of differential maternal solicitude, discouraging in-

352

RANDY THORNHILL AND BRYANT FURLOW

vestment in offspring under circumstances in which the offspring are unlikely to survive to reproductive maturity. Two of the eighteen studies reviewed with regard to our first two predictions failed to support our hypothesis; Davidson (1972) found no clear correlation between infant condition and incidence of PPD, and Hopkins et al. (1987) found social support to play an unimportant role in subjects’ depressions. The variance of infant conditions and the range of social support represented in a given study is likely to affect detection of correlations between social stress and PPD. Thus, these two studies may not be exceptions to our hypothesis; however, even if they are treated as true exceptions, the overall pattern is highly supportive of our hypothesis (sign test for predictions 1-6,35 positive excluding the Wilson etaf., 1996 review [also positive], 2 negative,p < .OOl). Available cross-cultural evidence suggests that many cultures observe traditional displays of social support that reduce the incidence of PPD. If similar preventative measures are adopted in western nations (for instance, doctors could discuss with a woman’s kin and mate the importance of displays of a willingness to help raise the coming child), we predict that the incidence of PPD would drop dramatically. Little research has been done to evaluate whether identifying and correcting the relevant social variables after PPD has been diagnosed can improve subsequent motheroffspring interactions, but this is a reasonable prediction of both our and Hagen’s hypotheses. Both our and Hagen’s hypotheses identify selective pressures, which, through evolutionary time, could have maintained maternal vulnerability to PPD. Predictions from the two models overlap, and both receive support in the medical literature. Future research should emphasize distinctions between evolved social need signal components (e.g., weeping) and resource withdrawal components (e.g., infanticidal urges) of PPD, so that the relative importance of each of these apparent evolved functions of PPD may be approximated.

V. INFANT CRYING AS

A

SIGNAL OF PHENOTYPIC QUALITY

Care-soliciting vocalizations are not limited to human neonates. They are common in bird and in other mammal species, and occur throughout the primate order (Newman, 1985). Human neonatal crying generally enhances caregiver proximity and investment (Ainsworth, 1969), but also carries serious risks as an apparently causal focal point of child abuse and even infanticide (Frodi, 1981; Steele and Pollack, 1968; Weston. 1968). This has been called the “crying paradox” (Barr, 1990), because crying is apparently necessary for the neonate to elicit care but may result in fitness-impacting reactions by caregivers. A resolution of this apparent paradox must explain

STRESS AND HUMAN BEHAVIOR

353

both the beneficial and the fitness-reducing (for the baby) aspects of caregivers’ reactions. It has been proposed that the resolution of the crying paradox rests with an understanding of evolved patterns of differential parental solicitude (Furlow, 1997), discussed earlier in this chapter (Section IV,C). Zahavi (1975, 1977a) and others (e.g., Grafen, 1990) argue that because deceived signal receivers regularly suffer significant fitness impacts (unnecessarily abandoning resources to a bluffing competitor, for instance, or mating with a genetically substandard suitor), signal receiving mechanisms should evolve that favor signals that are costly to send. Signal costs (fitness impacts) limit exaggeration during communication between organisms during which “claims” are being made about a signal sender’s phenotypic quality, because only high-quality signal senders can afford to suffer the associated fitness reductions. Just as it is important for courted female birds or competing red deer males to accurately evaluate the phenotypic quality of the individuals before them, parents who discriminatively evaluate the phenotypic quality of their offspring before responding to offspring solicitation can avoid the fitness costs of investing in nonviable babies. While attention has been paid to the potential for deceptive solicitation by primate infants (e.g., Hauser, 1986), few attempts have been made to analyze human crying in the context of the evolutionary theory of honest, signaling. Godfray (1991) addresses the evolutionarily stable signaling of offspring need in vertebrates, and presents a model that presupposes that some sort of criteria exists for parental assessment of offspring reproductive value, as the intrinsic reproductive value of a newborn is as important from the parental perspective as its asserted level of need. Some species appear to have such fitness-assessment components in their neonatal solicitation displays (Lyon et al., 1994; Bustamante et al., 1992). In this section, we evaluate predictions derived from the hypothesis that the human neonatal cry itself contains criteria that allow parental assessment of offspring reproductive value. Hence, whereas previous authors have studied the motivational (offspring need) component of crying, we focus instead on the question of whether the acoustic structure of cry vocalizations accurately indicates the condition of a crying neonate, and hence, correlates with its intrinsic reproductive value. Parental bonding and investment patterns hinge on an early critical assessment of offspring reproductive value (Daly and Wilson, 1995), and we propose that the earliest example of human vocal communication-the neonatal cry-constitutes an important source of offspring fitness information for parents during this assessment period. The previous section of this chapter, on postpartum depression, included “infant condition” as one variable of importance to maternal assessments of an offspring’s reproductive value. Many physical and behavioral cues of offspring health are available to parents (e.g., rash, irregular breathing),

354

R A N D Y THORNHILL A N D B R Y A N T FURLOW

but because the cry involves neurological, cardiorespiratory, and vocal phenotype, which are not as readily assessed as a rash on the skin, crying may reveal information about the internal structural and functional integrity of the offspring. Selection on parents to evaluate offspring reproductive value should have favored discriminative responses to these cues. (Discriminative investment is not limited to intrinsic offspring fitness, of course.) An honest signal model of infant crying predicts that crying will be more expensive to infants with relatively low phenotypic quality than to those with relatively high intrinsic fitness. Indeed, the energetic expense of crying is 11% higher than basal infant metabolism (Brignol et al., 1993). This expense is not likely to impact on healthy, well-nourished infants, but probably constitutes a serious threat to the fitness of ill and/or malnourished infants who must partition energy to mounting immunological defenses or who simply lack sufficient energy to afford the expense of crying. This is consistent with the expectations of the Zahavian paradigm of signal evolution. If infant crying contains acoustic cues of reproductive value, and is a costly honest signal in the Zahavian sense, then two testable predictions become immediately apparent: acoustic cry parameters will be correlated with infant phenotypic condition, and the relevant cry parameters will correlate with parental investment behaviors (or with emotions, like hostility or concern, which probably inspire given investment behaviors, like infanticidal abuse, neglect, or immediate investment). CONDITION A N D CRYACOUSTICS A. OFFSPRING Healthy human infants’ cries have an average fundamental frequency (pitch) of approximately 300-600 Hz (Furlow, 1997; Zeskind, 1983). From brain-damaged infants to malnourished, premature, asphyxiated, and parasitized infants, low infant fitness is correlated with significantly increased cry pitch (Frodi, 1985; Morley et al., 1991; Wasz-Hockert et al., 1985). Cry pitch is more exclusively indicative of serious illness than symptoms typically noted by pediatricians, including a change in respiration or pulse rate, rash, nasal discharge, stridor, pale skin, wheezing, temperature, sweating, dehydration, or behavioral symptoms (Morley et af., 1991). The magnitude of pitch abnormality correlates with the magnitude of the infant’s fitness impacts. Diabetic mothers’ infants have a mean maximum pitch of 1480 Hz, whereas infants of diabetic mothers who are also hypoglycemic produce cries with a mean maximum pitch of 1520 Hz; hypoglycemic infants of diabetic mothers who have hyperbilirubinemia (jaundice) have a mean maximum pitch of 1980 Hz (Wasz-Hockert et al., 1985). Likewise, the severity of asphyxiation can be predicted by the magnitude of cry pitch

STRESS AND H U M A N BEHAVIOR

355

abnormality (Michelsson, 1971). Infants with only peripheral asphyxia have a mean maximum pitch of 1000 Hz, but those with central asphyxia produce cries with a mean maximum pitch of 1460 Hz. Central asphyxia causes neurological damage, and constitutes a more serious impact to offspring reproductive value. The severity of pitch abnormality in infants with bacterial meningitis is more pronounced (higher pitched) in babies later diagnosed with neurological sequelae (Michelsson et af., 1977). Cry pitch is also associated with subsequent cognitive development (Donzelli et af., 1995; Lester, 1987). At 18 months and 5 years of age, children’s scores on cognitive tests were predicted by analysis of cry pitch during early infancy-babies with higher and more variably pitched cries score significantly lower on cognition tests than other children (Lester, 1987). Neurodevelopmental integrity is evidenced by normal cry pitch (Donzelli et al., 1995). B. CRYING A N D PARENTAL REACTIONS Crying is recognized in many cultures as an important sign of a healthy infant. Among the African Igbo people, for instance, babies who do not cry vigorously are abandoned in the forest (Basden, 1966). Laboratory studies of adult reactions to recorded cries report a negative emotional reaction to high-pitched cries in western cultures (e.g., Crowe and Zeskind, 1992; Zeskind and Lester, 1978). Adult reactions to cry pitch are similar, despite differences in caregiving experience, subject gender, and age (Furlow, 1997). Cardiac response and skin conductance levels support selfreported emotional states in study participants (Frodi et af., 1978; Weisenfeld et af., 1981). Heart rate increases are greater when participants are played high-pitched cries than when they are played normal cries, and skin conductance diminishes with time to normal cries, but not to high-pitched cries (Frodi et af., 1978; Weisenfeld et af., 1981). Pitch appears to be unrelated to aversiveness ratings under 610 Hz, roughly the upper limit of the range of normal (healthy) cries (Bisping et af., 1990). Hence, adults’ ratings of cry aversiveness correlate with infant health. It should be noted that cry pitch may be linked with another, unmeasured, acoustic parameter, and may not itself be the exact parameter adults respond to in reaction-to-cry studies. This seems unlikely, however, as pitch is obviously a salient acoustic parameter of cry vocalizations. The role of cry pitch in the incidence of PPD has yet to be studied. C. CONCLUSIONS Crying conveys information about infants’ health and may serve as an important cue of offspring reproductive value for parents. Parents, in turn,

356

R A N D Y THORNHILL A N D BRYANT FURLOW

react negatively to at least one acoustic correlate of low infant reproductive value (pitch). If aversiveness ratings (and their physiological correlates) in the lab are indicative of parental emotions to crying in more natural settings, it would appear that unhealthy children are likely to suffer reduced investment or even parental hostility. Zahavian selection for a signal that accurately communicates signal senders’ phenotypic quality seems to have played a potentially important role in the evolution of the human neonatal cry. A limitation to this review is that it is based on correlational data, and with the exception of adult emotional reactions to variation in cry acoustics, causality is therefore assumed rather than empirically established.

VI. SUMMARY We addressed four major topics under the heading of stress and human behavior that are currently advancing areas of research in human behavior. The first topic is human attraction and attractiveness. It is well established that looks or physical attractiveness matter a great deal in everyday human life and that attractive juveniles and adults of both sexes have social advantages. Only relatively recently have evolutionary psychologists and behaviorists explored in any detail the evolutionary basis of attraction and attractiveness in humans using hypotheses based on modern sexual selection theory. The hypothesis that attractiveness is a phenotypic marker of stress resistance is discussed, as is the complimentary hypothesis that attraction involves assessment of stress resistance. These hypotheses entail selection favoring individuals who viewed physical beauty as phenotypic stress resistance, and selection favoring individuals who displayed their stress resistance in physical features. Stress resistance is equated with low developmental instability, particularly low fluctuating asymmetry, and well-developed body and facial secondary sexual traits. Fluctuating asymmetry is known to be caused by various environmental and genetic perturbations acting during development. The development of secondary sexual traits in humans is mediated by sex hormones, which handicapkompromise the immune system. Thus, low asymmetry and highly developed secondary sexual traits signal an ability to cope with stresses during development. Evidence is reviewed showing that body and facial symmetry, as well as secondary sexual traits, have positive effects on sexual attractiveness. Attractive faces of both sexes exhibit relatively low asymmetry and relatively high development of sex-specific facial sex hormone markers. Facially symmetrical individuals and individuals with attractive expressions of facial secondary sexual traits exhibit greater emotional and physiological health,

STRESS AND HUMAN BEHAVIOR

357

a further indicator of their ability to resist stress. Also, nonfacial body symmetry correlates with attractiveness in men when attractiveness is measured as mating success (sex partner number, age of first sex, and other related factors). Moreover, nonfacial body attractiveness in both sexes is explicable by sex-specific sex hormone effects. It is concluded that human attraction and attractiveness are importantly related to stress resistance or general health and phenotypic and genetic quality. Most of the research on sexual attractiveness has involved correlational analyses with statistical control of confounding variables. However, some of the recent work on facial attractiveness has been experimental. There is need for further, careful experimentation with symmetry and secondary sexual traits. We argue that humans have been selected to make physical attractiveness judgments in the context of nepotism and reciprocal altruism based on the two markers of stress resistance: low asymmetry and secondary sexual traits. There is vast evidence that attractiveness affects nepotism as well as reciprocity such as friendships. However, the role of the two markers of stress resistance per se in this effect has not been studied. This may be a fruitful research domain in human behavior. The second area of behavioral research explored in this paper is the relationship between women’s sexual arousal and the stressful condition during women’s upbringing of limited or absent paternal investment. It has been hypothesized by developmental evolutionary psychologists that paternal divestment during development has important consequences for women’s sexuality. Empirically, the divestment appears to relate to earlier sexual maturity and sexual intercourse, reduced stability of romantic pairbonds, more sex partners, and emotional coldness toward mates. This suite of sexual traits has been interpreted by evolution-minded psychologists as evolved condition-dependent tactics for securing material benefits from many males in a social environment in which paternal investment from individual males is not reliable. In an ecological setting of paternal divestment, the suite of female sexual traits mentioned is viewed as maximizing the amount of male-provided benefits by consortships with many men, each of whom is capable or willing to invest to a small degree. We hypothesize that reduced frequency and highly selective female orgasm would be functional for women to couple with the other traits mentioned for exploiting resources held by men in a social environment of male paternal divestment. Female copulatory orgasm apparently has a selective bonding function. Thus, limited orgasm would promote women’s strategic pursuit of multiple partners by reducing the emotional commitment shown by a woman toward individual partners. The literature of female orgasm frequency is examined in the light of this hypothesis. As predicted, there is a positive relationship

358

RANDY THORNHILL A N D BRYANT FURLOW

between orgasm frequency in women and paternal investment during a girl’s upbringing. We emphasize that more research is needed to clarify the role of a father’s involvement on a developing female’s orgasm behavior. Next, we turn to the fascinating behaviors and associated mental states of women’s postpartum depression (PPD). Data derived from the medical literature on PPD is consistent with the predictions of our evolutionary psychological “resource withdrawal function” hypothesis for PPD. Through human evolutionary history, females who withdrew resources from infants exhibiting cues of low phenotypic quality (and therefore, low reproductive value to parents), or who withdrew resources from offspring during times of resource limitation would have improved their own long-term reproductive success. Mothers who invest indiscriminately in unhealthy offspring or who invested maximally during times of resource shortages would have been selected against, and should not represent our ancestors. As we predicted, cues of infant health, social support, mate support, and compromised resource base all correlate positively, in general, with the incidence of PPD. Cultures that observe ritualized, exaggerated displays of social and material support to new mothers seem to have a remarkably low incidence of PPD, but in those cultures in which European imperialism has disrupted traditional cultural customs, the incidence of PPD appears to be markedly higher. In addition, women suffering from PPD more often report thinking about killing their babies, and the incidence of actual infanticides is higher among PPD sufferers than among other mothers, as should be expected if PPD is an evolved baby killing (resource withdrawal) adaptation. In this way, vulnerability t o PPD could have been shaped and retained for its fitness-enhancing benefits for sufferers living in ancestral environments. We d o no argue or assume that PPD is currently adaptive. Our final topic is the baby’s cry. To our knowledge, no researcher has yet studied the role of abnormal acoustic quality in babies’ cries in the etiology of PPD. Based on our review of the relevant literature, we predict that abnormal cry quality contributes to maternal PPD, because the cry represents an assay of internal phenotypic quality (and hence, reproductive value) of offspring, and abnormal cry quality is a potent indicator of compromised infant phenotypic quality. Human neonatal crying appears to have been shaped by natural (parental) selection to reveal the soliciting offspring’s reproductive value, in addition to any need assertion or need-type communication (pain, hunger, fear, etc.) conveyed by cries. As is the case with our review of PPD, all published support for the Zahavian or “honest” signal hypothesis of cry quality is correlational rather than experimental. A further limitation is that studies of adult responses to infant cry quality are almost entirely limited to the western, industrialized nations. One of

STRESS AND HUMAN BEHAVIOR

359

us (Furlow) is initiating cross-cultural studies of adult reactions to cry quality in the neotropics.

Acknowledgments R T and BF worked together on all topics in the manuscript. RT was invited to write the manuscript and he is. therefore, first author. M. Milinski, A. Moiler, and P. J. B. Slater provided provided helpful suggestions on the manuscript. J. Belsky, E. Cashdan, K. MacDonald, and M. Surbey provided useful discussion about fathers and women’s sexuality. E. Hagen’s comments on postpartum depression were helpful, as were references on lactation that he brought to our attention. A. Rice’s help in preparation of the manuscript is greatly appreciated.

References Affonso, D., and Arizmendi, T. (1986). Disturbances in post-partum adaptation and depressive symptomatology. J . Psychosom. Obstet. Gynaecol. 5, 15-32, Ahmed, S. A.. and Tala], N. (1990). Sex hormones and the immune system. Part 2. Animal data. Bailliere’s Clin. Rheumatol. 4, 13-31. Ainsworth, M. (1969). Object relations, dependency, and attachment: A theoretical review of the mother-infant relationship. Child Dev. 40, 969-1025. Alexander, J., and Stimson, W. H. (1989). Sex hormones and the course of parasitic infection. Parasitol. Today 4, 1891-1 893. Alexander. R. D. (1986). Ostracism and indirect reciprocity: The reproductive significance of humor. Ethol. Sociobiol. 7 , 105-122. Alley, T. R., and Hildebrandt, K. A. (1988). Determinants and consequences of facial aesthetics. In “Social and Applied Aspects of Perceiving Faces” (T. R. Alley, ed.), pp. 101-140. Erlbaum, Hillsdale, NJ. Bailit. H. L., Workman, P. L., Niswander. J. D., and MacLean, C. J. (1970). Dental asymmetry as an indicator of genetic and environmental conditions in human populations. Hum. Biol. 42, 626-638. Baker, R. R. (1997). Copulation, masturbation and infidelity: State-of-the-art. Proc. Int. SOC. Hum. Ethol. (in press). Baker, R. R., and Bellis. M. A. (1995). “Human Sperm Competition: Copulation, Masturbation and Infidelity.” Chapman & Hall, London. Ballinger. B. C., Buckley, D.. Naylor, G. J., and Stransfield, D. (1979). Emotional disturbance following childbirth: Clinical findings and urinary excretion of cyclic AMP. Psychal. Med. 9(2), 293-300. Barnett. B.. Joffe, A., Duggan, A,, Wilson. M.. and Repke, J. (1996). Depressive symptoms, stress, and social support in pregnant and postpartum adolescents. Arch. Pediafr.Adolesc. Med. 150, 64-69. Barr. R. G. (1990). The early crying paradox. Hum. Nat. 1(4), 355-389. Basden, G. T. (1966). “Among the Ibos of Nigeria.” Frank Cass & Co., London. Beck, C. T. (1992). The lived experience of postpartum depression: A phenomenological study. Niirs. Res. 41(3), 166-170. Belsky. J. (1990). Parental and non-parental care and children’s socioemotional development: A decade in review. .I. Marriage Family 52, 885-903.

360

RANDY THORNHILL A N D BRYANT FURLOW

Belsky, J., Steinberg. L., and Draper, P. (1991).Childhood experience. interpersonal development and reproductive strategy: An evolutionary theory of socialization. Child Dev. 62,647-670. Bennett, D. E., and Slade, P. (1991). Infants born at risk: Consequences for maternal postpartum adjustment. Br. J. Med. Psychol. 64, 159-172. Birkhead. T.,and Moller. A. P. (1993). Female control of paternity. Trends Ecol. Evol. 8, 100-104. Bisping. R., Steingrueber, H.. Oltmann, M.. and Wenk. C. (1990).Adults’ tolerance of cries: An experimental investigation of acoustic features. Child Dev. 61, 1218-1229. Brignol, M., Rao, M., Blass, E., Marino, L.. and Glass. L. (1993).Effect of crying on energymetabolism in human neonates. Clin. Res. 41, A597. Buchwald, J.. and Unterman, R. (1982).Precursors and predictors of postpartum depression: A retrospective study. J. Prev. Psychiatry l(3). 293-308. Bundy. D. A. P. (1989).Gender-dependent patterns of infection and disease. Parusitol. Todiry 4, 186-189. Burger. J.. Horwitz, S., Forsyth, B., Leventhal. J.. and Leaf, P. (1993).Psychological sequelae of medical complications during pregnancy. Pediatrics 91(3).566-571. Buss, D. M. (1989).Sex differences in human mate preferences: Evolutionary hypotheses tested in 37 cultures. Behav. Brain Sci. 12, 1-49. Buss, D. (1994). “The Evolution of Desire: Strategies of Human Mating.” Basic Books, New York. Buss, D., and Schmidt, D. P. (1993).Sexual strategies theory: An evolutionary perspective on human mating. Psycho/. Rev. 100,204-232. Bustamante, J., Cuervo, J, and Moreno, J. (1992).The function of feeding chases in the chin strap penguin, Pygoscelis antartica. Anim. Behav. 44,753-759. Campbell, S.. and Cohn. J. (1991).Prevalence and correlates of postpartum depression in first-time mothers. J . Abnornz. Psychol. 100, 594-599. Cashdan, E. (1993).Attracting mates: Effects of parental investment on mate attraction. Ethol. Sociobiol. 14, 1-24. Close, S.(1980).“Birth Report: Extracts from Over 4000 personal Experiences. Nfer Publishing Company, Windsor, Berks. Collins, N.. Dunkel-Schetter. C.. Lobel, M.. and Scrimshaw. S. C. (1993).Social support in pregnancy: Psychosocial correlates of birth outcomes and postpartum depression. J . Pres. Soc. Psychol. 65(6), 1243-1258. Cox, J. L. (1978).Some socio-cultural determinants of psychiatric morbidity associated with childbearing. In “Mental Illness in Pregnancy and the Puerperium” (M. Sandler, ed.), pp. 91-98. Oxford University Press, London. Cox, J. L. (1979).An Ugandan puerperal psychosis. SOC. Psychiatry 14,49-52. Cox, J. L. (1983).Postnatal depression: A comparison of African and Scottish women. Soc. Psychint. IS,25-28. Crowe, H., and Zeskind, P. (1992).Psychophysiological and perceptual responses to infant cries varying in pitch: Comparison of adults with low and high scores on the Child Abuse Potential Inventory. Child Ahiise Neglect 16, 19-29. Cunningham, M. R.,Roberts, A. R., Barbee, A. P., Druen, P. B., and Wu, C.-H. (1995).Their ideas of beauty are, on the whole, the same as ours: Consistency and variability in the cross-cultural perception of female physical attractiveness. J . Pers. Soc. Psychol. 68,261-279. Cutrona. C. ( 1982).Nonpsychotic postpartum depression: A review of recent research. Clin. Psychol. Rev. 2(4),487-503.

STRESS AND HUMAN BEHAVIOR

361

Cutrona. C.. and Troutman, B. (1986). Social support, infant temperament. and parenting self-efficacy: A mediational model of postpartum depression. Child Dev. 57, 1507-1518. Daly, M., and Wilson, M. (1988). “Homicide.” de Gruyter. New York. Daly, M., and Wilson, M. (1995). Discriminative parental solicitude and the relevance of evolutionary models to the analysis of motivational systems. In “The Cognitive Neurosciences” (M. S . Gazzaniga, ed.), pp. 1269-1286. MIT Press, Cambridge, MA. Daly, M., Wilson, M., and Weghorst, S . J. (1985). Male sexual jealousy. Ethol. Sociobiol. 3, 11-27. Davidson, J. R. (1972). Post-partum mood change in Jamaican women: A description and discussion on its significance. Br. J . Psychiatry 121, 659-663. Dawkins, M. S.. and Guilford. T. (1995). An exaggerated preference for simple neural network models of signal evolution. Proc. R. Soc. London, Ser. B 261, 357-360. Donzelli. G., Rapisardi, G., Moroni, M., Scarano, E.. Ismaelli. A,. and Bruscaglioni, P. (1995). Neurodevelopmental prognostic significance of early cry analysis in preterm infants. Pediutr. Res. 38, 432. Douglas. G. (1963). Puerperal depression and excessive compliance with the mother. Br. J. Med. Psychol. 36, 271-278. Draper, P., and Harpending, H. (1982). Father absence and reproductive strategy: An evolutionary perspective. J . Anthropol. Res. 38, 255-273. Eagly, A. H., Ashmore, R. D., Makhijani, M. G., and Longo, L. C. (1991). What is beautiful is good, but . . . : A meta-analysis review of research on the physical attractiveness stereotype. Psychol. Bull. 110, 109-128. Enlow, D. H. (1990). “Facial Growth,” 3rd ed. Harcourt Brace Jovanovich. Philadelphia. Eyer, D. E. (1994). Mother-infant bonding: A scientific fiction. Hum. Not. 5(1), 69-94. Farkas, L. G., and Cheung, G. (1981). Facial asymmetry in healthy North American Caucasians: An anthropometrical study. Angle Orthodontist 51, 70-77. Fisher, S. (1973). “The Female Orgasm: Psychology, Physiology, Fantasy.” Basic Books, New York. Folstad. I . , and Karter, A. J. (1992). Parasites, bright males and the immunocompetence handicap. A m . Nut. 139, 603-622. Fraser, F. C. (1994). Developmental instability and fluctuating asymmetry in man. In “Developmental Instability: Its Origins and Evolutionary Implications” (T. A. Markow. ed.), pp. 319-334. Kluwer, Dordrecht. The Netherlands. Freeland, W. J. (1976). Pathogens in the evolution of primate sociality. BioTropicu 8, 12-24. Fried, M. N., and Fried. M. H. (1980). “Transitions: Four Rituals in Eight Cultures.” Norton. New York. Frodi, A. (1981). Contribution of infant characteristics to child abuse. A m . J . Ment. Defici. 85,341-349. Frodi, A. (1985). When empathy fails: Aversive infant crying and child abuse. In “Infant Crying” (B. Lester and C. Boukydis, eds.). pp. 107-123. Plenum, New York. Frodi. A.. Lamb, M., Leavitt. L., Donovan, W.. Neff, C.. and Sherry, D. (1978). Fathers’ and mothers’ responses to the appearance and cries of premature and normal infants. Dev. Psychol. 14, 490-498. Furlow. F. B. (1997). Human neonatal cry quality as an honest signal of fitness. Evol. Hum. Behuv. 18, 175-193. Gangestad, S. W., and Buss, D. M. (1993). Pathogen prevalence and human mate prefence. Ethol. Sociobiol. 14, 89-96. Gangestad. S. W.. and Thornhill, R. (1997). Human sexual selection and developmental stability. In “Evolutionary Social Psychology” ( J . A. Simpson and D. T. Kenrick, eds.), pp. 169-195. Erlbaum, Hillsdale, NJ.

362

RANDY THORNHILL AND BRYANT FURLOW

Gangestad, S. W., Thornhill. R.. and Yeo, R. (1994). Facial attractiveness, developmental stability and fluctuating asymmetry. Ethol. Sociobiol. 15, 73-85. Ghubash, R., and Abou-Saleh, M. T. (1997). Postpartum psychiatric illness in Arab culture: Prevalence and psychosocial correlates. Brit. J . Psychiat 171, 65-68. Godfray, H. (1991). Signaling of need by offspring to their parents. Nature (London) 352, 328-330. Goodwin. J . W., Godden. O., and Chance, G. W. (1976). “Perinatal Medicine: The Basic Science Underlying Clinical Practice.” Williams & Wilkins, Baltimore, MD. Gordon, R. E., and Gordon, K. K. (1967). Factors in postpartum emotional adjustment. Am. J. Orthopsychiatry 37, 359-360. Grafen, A. (1990). Biological signals as handicaps. J. Theor. Biol. 144, 517-546. Crammer, K. (1993). “Signale de liebe: Die biologischen gesetze der partnerschaft. Hoffman and Campe, Hamburg. Grammer, K., and Thornhill, R. (1994). Human (Homo sapiens) facial attractiveness and sexual selection: The role of symmetry and averageness. J . Comp. Psychol. 108,233-242. Gregoire, A. J., Kumas, R., Everitt. B.. Henderson, A. F., and Studd, J . W. (1996). Transdermal estrogen for treatment of severe postnatal depression. Lancet 347, 930-933. Grossman, C. J. (1985). Interactions between the gonadal steroids and the immune system. Science 227, 257-261. Hagen, E. H. (1996). Postpartum depression: A model for depression as a strategy to defect from coercive and exploitative relationships. Paper presented at the Human Behavior and Evolution Society Meeting, 1996. Haig, D. (1993). Genetic conflicts in human pregnancy. Q. Rev. Biol. 68, 495-532. Hamida, S. B. (1996). Mate preferences: Implications for the gender difference in unipolar depression. ASCAP Newsl. 9(5), 4-29. Hamilton, J . A. (1988). “Rick Factors for Depression in Women and Diagnostic Subtypes.” National Institute on Drug Abuse, Washington, DC. Hamilton, J. A,, Harberger. P. N., and Parry, B. L. (1992). The problem of terminology. In “Postpartum Psychiatric Illness” ( J . A. Hamilton and P. Harberger. eds.), pp. 33-40. University of Pennsylvania, Philadelphia. Handley, S. L., Dunn, T. L., Baker, J., Cockshott, C.. and Gould, S. (1977). Mood changes in puerperium, and plasma tryptophan and cortisol concentrations. Br. Med. J . 78,18-20. Harkness, S. (1987). The cultural mediation of postpartum depression. Med. Anrhropol. Q. 6(2), 194-209. Harris, B. (1981). Maternity blues in East African clinic attenders. Arch. Gen. Psychiatry 38( 11). 1293-1295. Hauser, M. (1986). Parent-offspring conflict: Care elicitation behavior and the “cry wolf” syndrome. In “Primate Ontogeny, Cognition, and Social Behavior” ( J . Else and P. Less, eds.), pp. 293-303. Cambridge University Press, Cambridge, UK. Haywood, J. S. (1989). Sexual selection by the handicap principle. Evolurion (Lawrence, Kans.) 43, 1387-1397. Heitler, S., and McCrensky, K. (1976). Postpartum depression: A multidimensional study. Diss. Absrr. Inf. 36(11B), 5792-5793. Hershkovitz, I., Ring, B., and Kobyliansky. E. (1992). Craniofacial asymmetry in Bedouin adults. Am. J. Hum. Biol. 4, 83-92. Herz, E. K. (1992). Prediction, recognition, and prevention. In “Postpartum Psychiatric Illness” ( J . A. Hamilton and P. Harberger, eds.), pp. 65-77. University of Pennsylvania Press, Philadelphia. Herzog, A,, and Detre, T. (1976). Psychotic reactions associated with childbirth. Dis.Nerv. Sys. 31, 229-235.

STRESS AND HUMAN BEHAVIOR

363

Hewlett, 9. S., ed. (1992). “Father-child Relations: Cultural and Biological Contexts.” de Gruyter, New York. Hickman, R. (1992). Husband support: A neglected aspect of postpartum psychiatric illness. In “Postpartum Psychiatric Illness” ( J . A. Hamilton and P. Harberger, eds.), pp. 305-312. University of Pennsylvania Press, Philadelphia. Hill, K.. and Hurtado, A. M. (1996). “Ache Life History: The Ecology and Demography of a Forest People.” de Gruyter, New York. Hopkins, J., Campbell, S. B., and Marcus, M. (1987). Role of infant-related stressors in postpartum depression. J . Abnorm. Psychol. 96, 237-241. Hurtado. A. M., and Hill. K. R. (1992). Parental effect on offspring survivorship among Ache and Hiwi hunter-gathers: Implications for modeling pair-bound stability. In “Father-child Relations: Cultural and Biosocial Contexts” (B. S. Hewlett, ed.), pp. 31-56. de Gruyter, New York. Inwood, D. G., ed. (1985). “Recent Advances Postpartum Psychiatric Disorders.” American Psychiatric Association Press, Washington, D.C. Iwasa, Y., Pomiankowski, A,. and Nee, S. (1991). The evolution of costly mate preferences. 11. The “handicap” principle. Evolution (Lawrence, Kans.) 45, 1431-1442. Jackson, L. A. (1992). “Physical Appearance and Gender: Sociobiological and Sociocultural perspectives.” State University of New York Press, Albany. Johnston, V. S.,and Franklin, M. (1993). Is beauty in the eye of the beholder? Ethol. Sociobiol. 14, 183-199. Johnstone, R. A. (1994). Female preference for symmetrical males as a by-product of selection for mate recognition. Nature (London) 372, 172-175. Jones, D. (1995). Sexual selection, physical attractiveness and facial neotony. Curr. Anthropol. 36,723-748. Jones, D., and Hill, K. (1993). Criteria of facial attractiveness in five populations. Hum. Nat. 4, 271 -296. Jones. W. H. S. (1923). “Hippocrates with an English Translation.” Heinemann, London. Kelly, J. V. (1967). The influence of native customs on obstetrics in Nigeria. Obster. Gynecol. 30, 608-612. Karalis, K., Goodwin, G., and Majzoub, J. A. (1996). Cortisol blockade of progesterone: A possible mechanism involved in the initiation of human labor. Nat. Med. 2, 556-560. Kruckman, L. D. (1992). Rituals and support: An anthropological view of postpartum depression. In “Postpartum Psychiatric Illness” (J. A. Hamilton and P. Harberger, eds.). pp. 137-148. University of Pennsylvania Press, Philadelphia. Kruckman, S., Kruckman, G.. and Kruckman, L. D. (1983). Multi-disciplinary perspectives on postpartum depression: An anthropological critique. Soc. Sci. Med. 17, 1027-1041. Kumar. R., and Marks, M. (1992). Infanticide and the law in England and Wales. In “Postpartum Psychiatric Illness” ( J . A. Hamilton and P. Harberger, eds.), pp. 257-274. University of Pennsylvania Press, Philadelphia. Kumar, R., and Robson, K. (1984). A prospective study of emotional disorders in childbearing women. Br. J. Psychiatry 144, 35-47. Labott, S., Ahleman, S., Wolever, M., and Martin, R. (1990). The physiological and psychological effects of the expression and inhibition of emotion. Behav. Med. 16, 182-189. Langlois. J. H.. and Roggman. L. A. (1990). Attractive faces are only average. Psychol. Sci. 1, 115-121. Langlois, J. H., Roggman, L. A., and Musselman, L. (1994). What is average and what is not average about attractive faces? Psychol. Sci. 5, 214-220. Langlois, J. H., Ritter. J. M. Casey, R. J.. and Sawin, D. B. (1995). Infant attractiveness predicts maternal behaviors and attitudes. Dev. Psychol. 31,464-472.

364

RANDY THORNHILL AND BRYANT FURLOW

Leathers, S. J., Kelley, M. A,, and Richman, J. A. (1997). Postpartum depressive symptomatology in new mothers and fathers. J. Nerv. Ment. Dis. 185, 129-139. Leeper. R. W. (1948). A motivational theory of emotions to replace “emotions as a disorganized response.” Psychol. Rev. 55, 5-21. Lerner, I. M. (1954) “Genetic Homeostasis.” Wiley, New York. Lester, B. (1987). Developmental outcome prediction from acoustic cry analysis in term and preterm infants. Pediatrics 80, 529-534. Livingston, J. E., MacLeod, P. M., and Applegarth. D. A. (1978). Vitamin B6 status in women with postpartum depression. Am. J . Clin. Nutr. 31(5), 886-891. Livshits, G., and Kobyliansky, E. (1991). Fluctuating asymmetry as a possible measure of developmental homeostasis in humans: A review. Hum. B i d . 63,441-466. Logsdon, M.. McBride, A,, and Birkimer. J. C. (1994). Social support and postpartum depression. Res. Nurs. Health 11,449-457. Low, B. S. (1989). Cross-cultural patterns in the training of children: An evolutionary perspective. J . Cornp. Psychol. 103, 311-319. Ludwig. W. (1932). “Das rechts-links Problem in Tierreich und beim Menschen.” Springer. Berlin. Lyon, B., Eadle. J., and Hamilton, L. (1994). Parental choice selects for ornamental plumage in American coot chicks. Noture (London) 371, 240-243. MacDonald, K. (1985). Early experience, relative placisticity and social development. Dev. Rev. 5, 99-121. MacDonald, K.. ed. (1988). “Sociobiological Perspectives on Human Development.” SpringerVerlag, New York. MacDonald. K. (1992). Warmth as a developmental construct: An evolutionary analysis. Child Dev. 63,753-774. Magiakou. M. A.. Mastorakos, G., Rabin. D., Dubbert, D., Gold, P. W., and Chrousos. G. P. (1 996). Hypothalamic corticotropin-releasing hormone suppression during the postpartum period: Implications for the increase in psychiatric manifestations at this time. J. Clin.Endocrinol. Matahol. 8, 1912-1917. Manning, J. T. (1995). Fluctuatingasymmetry and body weight in men and women. Implications for sexual selection. Ethol. Sociobiol. 16, 145-153. Manning, J. T.. Koukourakis, K.. and Brodie, D. A. (1997). Fluctuating asymmetry. metabolic rate and sexual selection in human males. Ethol. Sociobiol. 18, 15-21. Manning, J. T., Scutt, D., Whitehouse, G. H., Leinster, S. J., and Walton, J. M. (1996). Asymmetry and the menstrual cycle in women. Ethol. Sociobiol. 17, 129-143. Martin, R., Guthrie, C., and Pitts, C. (1993). Emotional crying. depressed mood, and secretory immunoglobulin-A. Behav. Med. 19, 111-1 14. Mazur, A,, Halper, C., and Udry, J. R. (1994). Dominant-looking male teenagers copulate earlier. Ethol. Sociohiol. 15, 87-94. McCahill. T. W., Meyer, L. C.. and Fischmann. A. M. (1979). “The Aftermath of Rape.” Heath and Co., Lexington. MA. Michelsson, K. (1971). Cry analysis of symptomless low birth weight neonates and of asphyxiated newborn infants. Acra Paediatr. Suppl. 216. Michelsson. K.. Sirvio. P., and Wasz-Hiickert, 0. (1977). Sound spectrographic cry analysis of infants with bacterial meningitis. Dev. Med. Child Neurol. 19,309-315. Miller, J. E., and Huss-Ashmore. R. (1989). Do reproductive patterns affect maternal nutritional status‘?An analysis of maternal depletion in lesotho. Am. J. Hum.Biol. 1,409-419. Mitton. J. B. (1993). Enzyme heterozygosity, metabolism and developmental stability. Cenetica 89,47-65.

STRESS AND HUMAN BEHAVIOR

365

Mitton, J. B.. and Grant, M. C. (1984). Associations among protein heterozygosity, growth rate and developmental homeostasis. Annu. Rev. Ecol. Syst. 15, 479-499. Moller, A. P. (1992). Parasites differentially increase the degree of fluctuating asymmetry in secondary sexual characters. J. Evol. Biol. 5, 691-699. Moller. A. P. (1993a). Morphology and sexual selection in the barn swallow. Hirttndo rusticu, in Chernobyl, Ukraine. Proc. R. Soc. London, Ser. B 252, 51-57. Moller, A. P. (1993b). Female preference for apparently symmetrical male sexual ornaments in the barn swallow. Hiricndo rustico. Behav. Ecol. Sociobiol. 32, 371-376. Moller, A. P. (1995). Patterns of fluctuating asymmetry in sexual ornaments of birds from marginal and central populations. A m . Nut. 145, 316-327. Moller. A. P. (1997a). Developmental stability and fitness: A review. A m . Nut. 149, 916-932. Moller. A. P. (1997b). Developmental selection against developmentally unstable offspring and sexual selection. J . Theor. B i d . 185, 415-422. Moller. A. P.. and Pomiankowski, A. (1993). Fluctuating asymmetry and sexual selection. Genetico 89, 267-279. Moller, A. P.. and Saino. N. (1994). Parasites, immunology of hosts. and host sexual selection. J. Pnrasitol. 80, 850-858. Moller. A. P., and Swaddle, J. P. (1997). “Asymmetry, Developmental Stability, and Evolution.” Oxford University Press, Oxford. Moller, A. P., and Thornhill, R. (1997a). A meta-analysis of the heritability of developmental stability. J . Evol. Biol. 10, 1-16. Moller, A. P.. and Thornhill. R. (1997b). Developmental stability is heritable. J . Evol. B i d . 10,69-76. Moller, A. P., Soler, M.. and Thornhill, R. (1995). Breast asymmetry, sexual selection and human reproductive success. Ethol. Sociobiol. 16, 207-219. Morgenshy, C. (1982). Psychological and attitudinal reaction to childbirth of recently parturient women. Dissertation, Fordham University. Dissertation Abstracts, Ann Arbor, MI. Morley, C., Thornton, A,, Cole, T., Fowler, M., and Hewson, P. (1991). Symptoms and signs in infants younger than 6 months of age correlated with severity of their illnesses. Pediatrics 88,1119-1124. Nesse. R. M. (1990). Evolutionary explanations of emotions. Hum. Not. l(3). 261-289. Newman, J. (1985). The infant cry in primates: An evolutionary perspective. In “Infant Crying” (B. Lester and C. Boukydis, eds.). pp. 307-323. Plenum, New York. Nilsson. A,. and Almgren, P. (1970). Para-natal emotional adjustment: A prospective investigation of 165 women. Acto Pscyhintri. Scund., Siippl. 220, 9-61. Nolen-Hoeksema. S. (1990). “Sex Differences in Depression.” Stanford University Press, Stanford, CA. Oakley. A. (1980). “Women Confined.” Schocken Books, New York. O’Hara, M. W., Rehm, L., and Campbell, S. (1982). Predicting depressive syrnptomatology: Cognitive-behavioral models and postpartum depression. J. Abnorm. Psychol. 91, 457-461. O’Hara. M. W.. Rehm, L., and Campbell, S. (1983). Postpartum depression: A role for social network and life stress variables. J. Nerv. Ment. Dis. 171(6). 336-341. O’Hara. M. W.. Neunaber, D.. and Zekosko, E. (1984). Prospective study of postpartum depression: Prevalence, course, and predictive factors. J . Abnorm. Psychol. 93, 158-171. O’Hara, M. W.. Zekoski, E.. Philipps. L., and Wright, E. (1990). Controlled prospective study of postpartum mood disorders: Comparison of childbearing and nonchildbearing women. J . Ahtiortn. Pscyhol. 99, 3-15. Okano, T., and Nomura, J. (1992). Endocrine study of the maternity blues. Prog. Neuropsychophurmacol. B i d . Psychiatry 16, 921-932.

366

RANDY THORNHILL AND BRYANT FURLOW

Paffenbarger, R. S. (1961). The picture puzzle of the post-partum psychosis. J. Chronic Dis. 13, 161-173. Parry, B. L. (1992). Reproductive-related depressions in women: Phenomena of hormonal kindling? In “Postpartum Psychiatric Illness’’ ( J . A. Hamilton and P. Harberger, eds.), pp. 200-218. University of Pennsylvania Press, Philadelphia. Parsons, P. A. (1990). Fluctuating asymmetry: An epigenetic measure of stress. Biol. Rev. Cambridge Philos. Soc. 65, 131-145. Parsons, P. A. (1992). Fluctuating asymmetry: A biological monitor of environmental and genomic stress. Heredity 68, 361-364. Parsons, P. A. (1996). Stress. resources, energy balances and evolutionary change. Evol. Bid. 29, 39-72. Patterson. G. R. (1986). Maternal rejection: Determinants or product for deviant child behavior? In “Relationships and Development” (W. W. Hartup and Z . Rublin, eds.), pp. 73-94. Erlbaum, Hillsdale, NJ. Paullekhoff, B. (1 992). Toward the diagnosis of postpartum psychotic depression. In “Postpartum Psychiatric Illness” ( J . A. Hamilton and P. Harberger, eds.), pp. 239-249. University of Pennsylvania Press. Philadelphia. Paykel, E. S., Emms, E.. Fletcher, J., and Rassaby, E. (1980). Life events and social support in puerperal depression. Br. J. Psychiatry 136, 339-346. Peacock, N. (1991). An evolutionary perspective on the patterning of maternal investment in pregnancy. Hum. Nut. 2(4), 351-385. Peck, S.. Peck, L., and Kataja. M. (1991). Skeletal asymmetry in esthetically pleasing faces. Angle Orthodontist 61, 43-48. Perrett, D. I., May, K. A,, and Yoshikawa, S. (1994). Facial shape and judgements of female attractiveness. Nature (London) 368, 239-242. Pillsbury, B. (1978). Doing the month, confinement and convalescence of Chinese women after childbirth. Soc. Sci. Med. 12, 11-22. Pitt, B. (1968). Atypical depression following childbirth. Br. J . Psychiatry 114, 13251335. Polak, M. (1997). Parasites, fluctuating asymmetry and sexual selection. In “Parasites: Effects on Host Hormones and Behavior” (N. E. Beckage, ed.), Chapman & Hall, New York (in press). Pomiankowski, A,, and Mbller, A. P. (1995). A resolution of the lek paradox. Proc. R. SOC. London, Ser. B 260,21-29. Pop, V. J., Wijnen, H. A,, Vanmontfort, M., Essed, G., Degeus, C. A., Watson, M., and Komproe, I. H. (1995). Blues and depression during early puerperium: Home versus hospital deliveries. Br. J. Obstet. Gynaecol. 102(9), 701-706. Profet, M. (1995). “Protecting your Baby to be: Preventing Birth Defects in the First Trimester.” Addison-Wiley. Reading, MA. Quijano. E. C., and Cobliner. G. (1983). Postpartum depression in adolescents. J . Adolesc. Health Care 4(3), 213. Richman, J.. Raskin, V., and Gaines, C. (1991). Gender roles, social support, and postpartum depressive symptomology. J. Nerv. Ment. Dis. 179, 139-147. Roe, A.. and Siegelman. L. (1963). A parent-child relations questionnaire. Child Dev. 34, 355-369. Roff, D. A. (1992). “The Evolution of Life Histories.” Chapman & Hall, London. Sackheim, H. A. (1985). Morphologic asymmetries of the face: A review. Brain Cognition 4, 296-312. Shackelford, T. K., and Larsen, R. J. (1997). Facial asymmetry as an indicator of psychological, emotional and physiological distress. J. Pers. Soc. Psychol. 72, 456-466.

STRESS AND HUMAN BEHAVIOR

367

Shapiro, B. L. (1992). Development of human autosomal aneuploid phenotypes (with an emphasis on Down syndrome). Acfa Zool. Fenn. 191,97-105. Singh, D. (1993). Adaptive significance of female physical attractiveness: Role of waist-tohip ratio. J . Pers. SOC.Psychol. 59, 1192-1201. Singh, D. (1995). Female health, attractiveness and desirability for relationships: Role of breast asymmetry and waist-to-hip ratio. Erhol. Sociobiol. 16, 465-481. Smith, R. L. (1984). Human sperm competition. I n “Sperm Competition and the Evolution of Animal Mating Systems” (R. L. Smith, ed.). pp. 601-660. Academic Press, London. Sosa, R., Kennell, J., Klaus, M., Robertson, S., and Urrutia, J. (1980). The effect of a supportive companion on perinatal problems, length of labor, and mother-infant interaction. N. Engl. J. Med. 303(1l), 597-599. Spangenberg, J., and Pieters, H. (1991). Factors related to postpartum depression. S.Afr. J. Psychof. 21, 159-165. Steele. B., and Pollack, C. (1968). A psychiatric study of parents who abuse infants and small children. I n “The Battered Child” (R. Helfer and C. Kempe, eds.), pp. 103-145. University of Chicago Press, Chicago, IL. Stephenson, R., Huxel, L., and Harui-Walsh, E. (1978). From wife to mother: An exploratory study of Micronesian postpartum practices. Preliminary Report, unpublished manuscript, University of Guam. Surbey, M. K. (1990). Family composition, stress and human menarche. I n “The Socioendocrinology of Primate Reproduction” (T. E. Ziegler and F. B. Bercovitch, eds.), pp. 11-32. Wiley-Liss, New York. Sutton, P. R. N. (1969). Bizygomatic diameter: The thickness of the soft tissues over the zygions. Am. J. Phys. Anthropol. 30,303-310. Swaddle, J. P., and Cuthill. I. C. (1994a). Female zebra finches prefer symmetric males. Narure (London) 367, 165-166. Swaddle, J. P., and Cuthill, I. C. (1994b). Female zebra finches prefer males with symmetric chest plumage. Proc. R. Soc. London, Ser. B 258, 267-271. Swaddle, J . P., and Cuthill, 1. C. (1995). Asymmetry and human facial attractiveness: Symmetry may not always be beautiful. Proc. R. Soc. London, Ser. B 261, 111-116. Symons, D. (1979). “The Evolution of Human Sexuality.” Oxford University Press, Oxford. Symons, D. (1995). Beauty is in the adaptations of the beholder: The evolutionary psychology of human female sexual attractiveness. In “Sexual NaturelSexual Culture” (P. R. Abramson and S. D. Pinkerton, eds.), pp. 80-118. University of Chicago Press, Chicago. Thornhill, N. W., and Thornhill, R. (1984). Book review: Incest: A biosocial view. Ethol. Sociobiol. 5, 211-214. Thornhill, R. (1994). Is there psychological adaptation to rape? Anal. Kritic 16, 68-85. Thornhill, R. (1997). Rape-victim psychological pain revisited. In “Human Nature: A Critical Reader” (L. Betzig, ed.), pp. 239-240. Oxford University Press. Oxford. Thornhill, R., and Gangestad, S. W. (1993). Human facial beauty: Averageness, symmetry and parasite resistance. Hum. Nat. 4, 237-269. Thornhill, R., and Gangestad, S. W. (1994). Human fluctuating asymmetry and sexual behavior. Psychol. Sci. 5,297-302. Thornhill, R., and Moller, A. P. (1997). Developmental stability, disease and medicine. Biol. Reviews 72, 497-528. Thornhill, R., and Thornhill, N. W. (1989). The evolution of psychological pain. I n “Sociobiology and the Social Sciences” (R. Bell, ed.), pp. 73-103. Texas Tech University Press, Lubbock. Thornhill, R., Gangestad. S. W., and Comer, R. (1995). Human female orgasm and mate fluctuating asymmetry. Anim. Behav. 50, 1601-1615.

368

RANDY THORNHILL A N D BRYANT FURLOW

Trivers, R. (1974). Parent-offspring conflict. Am. Zool. 14, 249-264. Unterman, R.. Posner, N., and Williams. K . (1990). Postpartum depressive disorders: Changing trends. Birh 17, 131-136. Upreti, N. S. (1997). A study of the family support system: Child bearing and child rearing rituals in Kathmandu, Nepal. Unpublished doctoral dissertation, University of Wisconsin. Madison. Van Valen, L. (1962). A study of fluctuating asymmetry. Evolution (Lawrence, Kans.) 16, 125-142. Waldrop. M. F., Pedersen, F. A.. and Bell, R. Q. (1968). Minor physical anomalies and behavior in preschool children. Child Dev. 39,391-400. Wasz-Hockert, O., Koivisto. M., Vuorenkoski. V., Partanen. T.. and Lind. J . (1985). Twentyfive years of Scandinavian cry research. In “Infant Crying” (B. Lester and C. Boukydis, eds.), pp. 187-215. Plenum. New York. Watson. P. J., and Thornhill, R. (1994). Fluctuating asymmetry and sexual selection. Trends Ecol. Evol. 9, 21-25. Webster, M.. Thompson, J., Mitchell, E., and Werry. J. (1994). Postnatal depression in a community cohort. Aitsf. N . Z. J. Psychiatry 28, 42-49. Wedekind, C. (1994). Mate choice and maternal selection for specific parasite resistance before, during and after fertilization. Philos. Trans. R. Soc. London, Ser. B 346,303-31 1. Wedekind. C.. and Folstad, 1. (1994). Adaptive or non-adaptive immunosuppression by sex hormones? Am. Nat. 143, 93-98. Weisenfeld, A,. Malatesta, C. Z., and DeLoach, L. (1981). Differential parental responses to familiar and unfamiliar infant distress signals. Infant Behav. Dev. 4(3), 281-290. Weston. J. (1968). The pathology of child abuse. I n “The Battered Child” (R. Helfer and C. Kempe, eds.). pp. 77-100. University of Chicago Press, Chicago. Whiffen, V. E. (1992). Is postpartum depression a distinct diagnosis‘? Clin. P.Tycho1. Rev. 12, 485-508. Wick. S. (1941). Puerperal psychoses. Wis.Med. J. 40, 299-302. Wisner, K. L., Peindl. K., and Hanusa, B. H. (1994). Symptomatology of affective and psychotic illnesses related t o childbearing. J. Affeciive Disord. 30(2), 77-87. Wolman. W. L.. Chalmers, B., Hofmeyr, G . .and Nikodem. V . C. (1993). Postpartum depression and companionship in the clinical birth environment: A randomized. controlled study. Am. J. Ohster. Gynecol. 168(5). 1388-1393. Wood. J. (1994). “Dynamics of Human Reproduction: Biology, Biometry, Demography.” de Gruyter, New York. Worthington-Roberts. B., Vermeersch, J., and Williams, S. (1985). “Nutrition in Pregnancy and Lactation.” MirrorlMosby College Publishing, St. Louis. MO. Yates. A . L. (1928). The role of nose and throat disease in the production of deformities of the jaw. Dent. Rec. 48, 609-623. Zahavi. A. (1975). Mate selection-A selection for a handicap. J. Theor. Biol. 53, 205-214. Zahavi. A . (1977a). The cost of honesty (further remarks on the handicap principle). J . Theor. Biol. 67, 603-605. Zahavi, A. (l977b). Reliability in communication and the evolution of altruism. In “Evolutionary Ecology” (B. Stonehouse and C. Perrins, eds.), pp. 253-260. Macmillan. London. Zakharov, V. M. (1992). Population phenogenetics: Analysis of developmental stability in natural populations. Acia Zool. Fenn. 191, 7-30. Zelkowitz, P., and Milet. T. H. (1995). Screening for postpartum depression in a community sample. Con. J. Psychiatry 40(2), 80-86. Zeskind, P. (1983). Production and spectral analysis of neonatal crying and its relation to other biobehavioral systems in the infant at risk. In “Infants Born at Risk: Physiological,

STRESS AND HUMAN BEHAVIOR

369

Perceptual. and Cognitive Processes” (T. Field and A. Sostek. eds.), pp. 23-43. Grune & Stratton, New York. Zeskind, P.. and Lester, M. (1978). Acoustic features and auditory perceptions of cries of newborns with prenatal and natal complications. Child Dev. 49, 580-589. Zuk. M. (1990). Reproductive strategies and sex differences in disease susceptibility: An evolutionary viewpoint. Pornsirol. Today 6, 231-233.

This Page Intentionally Left Blank

ADVANCES I N THE STUDY OF BEHAVIOR. VOL. 27

Welfare, Stress, and the Evolution of Feelings DONALD M. BROOM DEPARTMENT OF CLINICAL VETERINARY MEDICINE UNIVERSITY OF CAMBRIDGE MADINGLEY ROAD

CAMBRIDGE, CB3 OES UNITED

KINGDOM

I. FEELINGS, THEIR ROLEA N D THEIR EVOLUTION

A. FEELINGS, EMOTIONS, AND CONSCIOUSNESS Three of the ways in which feelings can arise in an individual are as follows. First, inputs to sensory systems may result in changes in the brain, which we refer to as sensations or perceptions. Some of these have wideranging effects within the brain in addition to information processing, storage as memory, or initiation of activity modification. They lead to feelings in the individual, for example, pain or sexual gratification (Ottoson, 1983; Swenson and Reece, 1993). Second, various neural and hormonal changes result in physiologically describable conditions in individuals, which we refer to as emotions. The emotional state may involve electrical and neurochemical activity in well-defined parts of the brain, hormone release, and peripheral consequences. These various changes may also result in feelings, such as lust or anxiety, although emotional states may exist without any accompanying feeling, for example, as active regions of the amygdala with no cortical activity or during sleep (Guyton, 1982). Third, even in the absence of sensory input, or hormonal change, or activity in emotional centers in the brain, complex or simple brain processing can lead to the existence of feelings, for example, pleasure of achievement, guilt, or boredom. Each feeling is an internal brain construct, which the individual concerned may be able to describe but which will not necessarily have any external manifestations. As Dawkins (1993, p. 142) has pointed out, feelings may have observable signs, for example, happiness can lead to laughter, but happiness is not laughter and cannot be defined in terms of laughter. It is part of the state of the individual at that time. Some of the characteristics of feelings and examples of feelings are detailed in Table I. 37 1

Copyright 0 1998 hy Academic Press All rights of reproduction in any form reserved 0065-3454/98 $25 00

372

DONALD M. BROOM

TABLE I CHARACTERISTICS A N D EXAMPLES OF FEELINGS 1. A feeling is a brain construct within an individual, sometimes with peripheral links. 2. The brain construct includes something additional to that required for other functioning. 3. A feeling is recognizable by the individual when it recurs. 4. Feelings may change behavior immediately or eventually hut need not do so. 5. Feelings often act as reinforcers when learning. 6 . Feelings can he positive or negative in that they promote approach or avoidance. 7. Examples of feelings are: pain, malaise. tiredness, hunger, thirst, thermal discomfort, fear, anxiety. grief, frustration, guilt, depression, boredom. loneliness, general suffering, lust, jealousy, anger, sexual pleasure. eating pleasure, exhilaration. other sensory pleasure. achievement pleasure, general happiness.

Each kind of feeling can vary greatly in strength according to the magnitude and duration of the eliciting input. The mechanism for initiating the feeling in the individual in response to an input can also vary. Indeed, individuals will vary in sensory functioning, other physiological processing, and analytical ability according to their genotype and environment during life. Hence feelings will vary from one individual to another. As indicated above, some feelings are largely elicited by low-level neural processing, while others depend on very complex processing. Pain depends on inputs from nociceptive pathways, usually originating in nociceptive receptor cells, and does not require high-level processing in the brain. Similarly, thirst is principally dependent on inputs from body fluid monitors and mouth receptors, thermal discomfort results from local or general peripheral input, and pleasure associated with food or sexual intercourse is principally due to sensory input. Fear, in contrast to pain, requires high-level processing, usually involving the comparison of sensory inputs with established models in the brain of what constitutes a familiar or a dangerous stimulus. Likewise, frustration is complex because it necessitates precise expectations to compare with actual inputs. The most complex processes may be involved in deriving pleasure from the solving of a difficult problem, or in some situations that lead to anxiety. Simpler, perhaps more primitive, processes leading to feelings are likely to be more widespread in the animal kingdom than the most complex process. However, the existence of a range of levels of complexity in the origins of different feelings does not mean that the feelings themselves, or any behavioral consequences that they may have, also differ in complexity. All may be equally simple. It is also likely that most feelings, whatever the complexity of their initiation, can be amplified by complex brain processes when much attention is devoted to the source of the feeling, or diminished by active reduction in the brain resources employed in that area.

WELFARE. STRESS. AND THE EVOLUTION OF FEELINGS

373

The variation in the extent of high-level processing, which is likely to be involved in the initiation of feelings, is clearly relevant to the relationship between having feelings and being conscious or aware. Emotion, in the sense of neural activity in the emotion centers of the brain or specific hormonal changes, can occur without any feelings being reported, but how aware does an individual have to be in order to feel something such as pain or the pleasure associated with eating a favorite food? Pain is universally referred to as a feeling, indeed the definition of pain given in a veterinary dictionary by Blood and Studdert (1988) starts with the words “a feeling.” The feeling is what distinguishes pain from nociception. Most medical and veterinary usage of the words “conscious” or “aware” would imply that an individual cannot have a feeling without being conscious and aware. Blood and Studdert (1988) define conscious as “capable of responding to sensory stimuli; awake; aware,” a definition that refers to the lower threshold of what people might call conscious. The Oxford English Dictionary defines a feeling as “pleasurable or painful consciousness, emotional appreciation or sense.” Others wish to elevate the terms conscious and aware to mean something more complex, in some cases defining them so that they are exclusively human qualities. Griffin (1981) says that “awareness involves the experiencing of interrelated mental images” and (1984) defines conscious as “aware of what one is doing or intending to do, having a purpose and intention in one’s actions.” It would seem likely that many feelings involve mental images, and hence are associated with being aware and conscious as defined by Griffin but we cannot know this for certain. Statements by others, on the other hand, as Rodd (1990, pp. 51-54) has pointed out, seem designed to specifically exclude feelings from consciousness. Gallup (1983) refers to consciousness as the demonstrable capacity to reflect about the self, specifically the ability to recognize oneself in mirrors, and Humphrey (1986, pp. 93-94) states that there is no consciousness in the development of a human baby until it recognizes itself in a mirror or makes guesses about other people’s feelings. Dennett (1991) uses the term consciousness in a much wider and vaguer way but tends (pp. 171-226) ot equate the complexity of the individual’s world with the existence of and necessity for consciousness. The definition of awareness that will be used here is: “awareness is a state in which complex brain analysis is used to process sensory stimuli or constructs based on memory.” It is clear that there is a range of degrees of awareness (Sommerville and Broom, 1998), the highest of which might require a greater extent of intellectual processing than is needed for most feelings; but it is also clear that, according to the commoner meanings of the concept, feelings involve awareness. The term consciousness, however, is so widely used in the Blood and Studdert

374

DONALD M. BROOM

sense (discussed earlier) that it would be simplest to limit its meaning to this and use awareness when referring to different degrees. When the word “feelings” is used, some authors qualify it with “subjective.” This usage seems redundant, as every feeling is necessarily limited to one individual or subject. The word subjective can also be confusing because it is sometimes used to refer to a conclusion that is based not on the observations, but on self-examination by the observer. The various feelings are discussed later with reference to their possible evolutionary origins, but it is useful at this stage to consider the widely used categorization of feelings into pleasant and unpleasant. Unpleasant feelings are those that the individual concerned would avoid if possible, whereas efforts are generally made to experience pleasant feelings. Single unpleasant feelings or combinations of unpleasant feelings may sometimes be referred to as suffering. Dawkins (1990) stated that “suffering occurs when unpleasant subjective feelings are acute or continue for a long time because the animal is unable to carry out the actions that would normally reduce risks to life and reproduction in those circumstances.” However, as Broom and Johnson (1993, p. 81) have pointed out, most people would include all but the milder, briefer kinds of pain and malaise within the term suffering and such problems do not necessarily involve inability to reduce risks to life and reproduction. Hence, a better definition might be “suffering is an unpleasant feeling, which is prolonged or severe.” An individual in distress may well have unpleasant feelings, but could also be showing physiological consequences of adverse conditions without having such feelings. Hence “distress” is not listed as a feeling. B. EVIDENCE FOR FEELINGS The existence of a feeling in an individual may result in a change in its physiology or behavior that can be recognized by other individuals, including a human scientist. If the change were unique to individuals with such a feeling, the possibility would arise for the existence of the feeling to be recognized whenever the change occurred. For example, persons who feel grief exhibit particular facial expressions and may produce tears. Individual actions are often not unique to particular feelings, but combinations of actions can be good indicators. However, we know that such expressions and tears can be faked, and although the observer might be quite efficient in ability to identify the signs of grief and to discriminate between fake and real signs, the achievable precision in the identification of grief will be no more than a high probability. A person can also use words to express aspects of his or her grief, but words are more easily faked than the behavior changes, and so are generally less reliable as indicators of feelings. Despite

WELFARE. STRESS. A N D THE EVOLUTION OF FEELINGS

375

the possibilities of deliberate deceit or other simulation, words and behavior observation can often provide evidence for reasonably confident recognition of feelings in other people. For most observers, the recognition of grief is initiated after personal experience of such a feeling and facilitated by previous experience of what was assumed, from observing context and consequences, to be grief in others. Examples of other feelings that can substantially affect behavior are pain and guilt and those feelings easily related to particular needs. Some needs are associated with feelings and these feelings are likely to change when the need is satisfied (Broom, 1996). Examples of such feelings include hunger, eating pleasure, lust, and sexual pleasure. As needs are part of motivational state and motivational state alters the likelihood of occurrence of behaviors, the existence of needs can be deduced from the frequency and pattern of behavior (Fraser and Broom, 1990, pp. 31-38, 263-264). Hence, reasonable predictions about various feelings can be obtained from observations of behavior. The major problem with the recognition of feelings from observations of behavior is that feelings may often exist without any behavioral or physiological change to indicate them. Abdominal pain can lead to particular postures and movements and leg pain can lead to limping (for other examples, see Fraser and Broom, pp. 296-304), but severe pain can exist without any detectable sign. In some cases an observer or experimenter can contrive situations so as to maximize the likelihood that a measurable behavioral or physiological change will indicate the existence and extent of a feeling. For example, suspected localized pain may be identified by palpation of the area and measurement of behavioral, heart rate, and adrenal cortex responses. Experimental studies of animal preferences may also be carried out in order to obtain some information about a feeling such as hunger, frustration, or an aspect of sensory pleasure. The existence of a strong preference for some resource or possibility to carry out a behavior gives some indirect information about what an animal is likely to be feeling, but may not discriminate between working hard because of the existence of a negative feeling and working hard in order to obtain a positive feeling. Some feelings are not easy to investigate experimentally in this way, for example, malaise or boredom, because, for different reasons, the feeling may affect the likelihood of carrying out the preference test. Some information that can help in the recognition of feelings can be obtained from measurement of physiological changes including brain state. Heart rate and adrenal cortex activity changes, as already mentioned, and measurements of hormonal or neural activity changes may coincide with or precede changes in feelings. Brain scanning techniques can indicate sites and pathways in the brain where activity is occurring. Such activity in the

376

DONALD M. BROOM

brain may be related to reports or behavior changes that give other evidence of the existence of feelings and hence may themselves come to be used as evidence for particular feelings. The general conclusion about evidence for the existence and extent of feelings is that, even with sophisticated techniques, it is not possible to know exactly the feelings of any other individual, whatever the species, but reasonable predictions may be made using evidence, most reliably that from carefully studied behavior. The argument that no feelings can be recognized in others unless the individual can describe the feeling in words is wrong.

C. AREFEELINGS FUNCTIONAL OR JUST EPIPHENOMENA? I.

Feelings in General

It is thought by some scientists that all feelings are merely trivial byproducts of processes within the body. Skinner (1974, p. 17) said “what is felt or introspectively observed is not some non-physical world of consciousness, mind or mental life but the observer’s own body. This does not mean . . . that what are felt or introspectively observed are the causes of behavior.” Further, he said (1978, p. 124), “One feels various states and processes within one’s body, but these are collateral products of one’s genetic and personal history. No creative or initiating function is to be assigned to them.” This view of feelings as solely an accident of individual development with no function or relevance to any other individuals is certainly not held by many people. As Dawkins (1993, p. 5 ) points out, the actions of people are much affected by a belief that these might cause pain, happiness, or sorrow in others. Once it is accepted, as it is by most people, especially pet owners, that feelings exist in other individuals of our own and other species, the idea that they must have some function often follows. Y.-K. Ng (personal communication) says that consciousness is a major mechanism in individuals and hence it “must contribute to fitness to survive natural selection.” Although adaptive characteristics will survive in populations because selection will favor their survival, if feelings were just an accidental, nonadaptive consequence of other adaptive mechanisms, would they persist in a population? Provided that a nonadaptive characteristic has some genetic component (as almost everything has) and some cost, it is likely that it would disappear from the population. Because various feelings have persisted, they are probably adaptive. The nature of any advantage to animals of having feelings is explored here in some detail, but first it is necessary to consider possible origins of feelings and the possibility that they are still accidental and nonfunctional.

WELFARE, STRESS. A N D THE EVOLUTION OF FEELINGS

377

An assumption outlined in Section I,A is that feelings involve brain activity additional to the minimum required for information processing, storage as memory, and motor output. During the early stages of development of systems such as those for recognizing and responding to predators or for recognizing and ingesting food, it seems very likely that there would have been some accidental activation of parts of the brain that were not essential for the neural function. Also, as the system became more efficient, it is likely that pathways that were at first necessary, perhaps to ensure effective communications between receptor and effector, became redundant but were not immediately eliminated from the functioning system. Both of these are examples of nonfunctional epiphenomena of the system, which might have the kinds of effects in the brain that became feelings. Some of these epiphenomena might continue to be nonfunctional but inevitable side effects of an essential system. Others might have had effects that eventually became functional, as discussed later in relation to the various feelings. It is also possible that there are feelings that were once functional in the ancestors of present-day animals but that now have little or no function. Finally, it is important to consider the extent to which feelings might sometimes be harmful (Broom and Johnson, 1993, p. 80), perhaps by making it more difficult for individuals to show the most appropriate responses. The idea that pain has the function of preventing body damage has been espoused on many occasions, but the more general concept of all feelings being functional is relatively recent. Dawkins (1977) stated that “It is reasonable to assume that subjective feelings (like other characteristics) evolved because animals which possessed them were fitter than those which did not” and “feelings must be a product of natural selection. They are part of biology.” In a much more extensive exposition, Cabanac (1979) said, We experience feelings of hunger because that is part of our mechanism for rectifying a food deficit and getting something to eat. We experience fear and pain because they are part of our body’s way of removing us from situations that are life-threatening. Conscious experiences are there as survival aids.

A similar argument was presented by Wiepkema (1985) who asserted that feelings are involved in monitoring the effectiveness of regulatory actions, being positive when the regulation is successful and negative when it is not. A further statement that the evolutionary advantage of having feelings is considerable was made by Dawkins (1990), and Broom and Johnson (1993, p. 334) said, A final point about the evolution of adaptation to the vicissitudes of the physical and

social environment is that a very important part of that evolution has been the develop-

378

DONALD M. BROOM

ment of the complex appreciation of the interactions of an individual with the world in which it lives, which we call feelings. Complex brains, like those of vertebrates, have complex systems for regulating those interactions which are not just the product of automatic responses to stimuli. If an individual has a system of feelings which involves changes in its mental, and perhaps in its hormonal, functioning because a certain kind of body regulation or because an anticipated event has not occurred, such an individual will have increased fitness in comparison with a genetically different individual which has no such system.

Similarly, Y.-K. Ng (personal communication) argues that awareness contributes to fitness but is limited to species in which there is plasticity in brain and behavior, and that “affective feelings must have evolved fairly quickly after the evolution of awareness, if not concomitantly.” In order for feelings to confer an adaptive advantage it is essential that they should have an effect (Dawkins, 1993, p. 169); the exploration of such effects and how they might be recognized is a major part of this paper. The effect of the feelings might be “that the individual is more likely to carry out some adaptive action and hence more likely to survive” (Broom, 1996). The most likely way in which this would occur is that the feeling acts as a reinforcer, which makes it more likely that the individual will learn to carry out the adaptive action. Indeed “if the state of the individual in certain conditions is desirable from an evolutionary viewpoint, there should be a propensity for that individual to have good feelings. On the other hand, if a state is one which should be quickly altered, it should be associated with unpleasant feelings which prompt avoidance or some other action” (Broom, 1996). If a feeling does have an effect such that it can act as a reinforcer to promote adaptive behavior, the effect might be coincident with the occurrence of the feeling, or it could be that the feeling is remembered so that its beneficial effect occurs long after the feeling itself has finished. Such effects would be difficult to distinguish from the effects of other events in the life of the individual and hence to attribute to the feeling. The general argument that consciousness in general and feelings in particular make a difference in the way in which organisms possessing it function is presented with some force by Dawkins (1993, pp. 9, 143,167-181). One straightforward but strong argument is that most people report that many of the feelings listed in Table I can have a considerable effect on the way in which they organize their lives. Observation of human behavior supports this. Proposals about the effects and functions of consciousness have suggested that it allows more effective analysis of the environmental and prediction of the future (Crook, 1980, p. 31) or that it improves the efficiency of information processing in general (Weiskrantz, 1987). Dawkins (1993) suggests that “Consciousness might . . . make actions more decisive or give better anticipation of the future” and “have an evolutionary effect

WELFARE. STRESS, AND THE EVOLUTION OF FEELINGS

379

and an effect which we could detect.” Dawkins does not claim that every action will be more efficient if the individual carries it out consciously, and refers to the observation by Baars (1988) that certain actions progress more efficiently if the individual is not actively thinking about what is being done. However, this argument about the function of consciousness in the sense used by those authors is more relevant to the control of actions than to areas where feeling plays an important role. 2. Pain

Although the word pain is used colloquially to refer to a wide range of unpleasant experiences, its scientific and medical meaning is limited to refer to a sensation, that is, to the immediate consequences of a particular sensory input. The sensation elicits immediate avoidance or subsequent modification of behavior whose effect is to reduce the likelihood of recurrence of the sensation. Hence, a definition of pain is “A sensation which, without involving higher level brain processing, such as that associated with fear, is very aversive” (Broom and Johnson, 1993, p. 27). Pain usually involves specialized nociceptive receptor cells and some degree of injury. Even in the case of phantom limb pain, the specialized nociceptive neural pathways are involved. Pain normally elicits protective reactions, causes emotional responses, and results in learned avoidance behavior. The pain system includes: specialized receptor cells, nerves in which evoked electrical responses to mechanical or thermal damage can be detected, neural pathways involving characteristic transmitters such as substance P, brain mechanisms, which include endogenous analgesic opioids, and the propensity to initiate avoidance behavior. This system is present in all vertebrates that have been studied, including fish, and most aspects of it are also present in some invertebrates, for example, cephalopod molluscs. As mentioned already, all pain is regarded as being a feeling and if there is activity in the nociceptive system with no feeling, perhaps because of naturally occurring opioid-induced inhibition or the use of an externally applied analgesic, then it is not pain. We cannot know whether pain in another individual is the same as that which we feel ourselves, but observation of behavior in all species with complex nociceptive systems suggests that there is likely to be much similarity among the different kinds of animals in the feeling of pain. The importance of feeling pain in promoting individual survival is considerable (Melzack, 1973). When pain is felt, the individual can take action to minimize tissue damage being caused, and the greater the feeling of pain, the faster the initial action is likely to be. Once pain has been felt in a recognizable situation, the possibility of learning to avoid any future pain, and hence damage, of the same kind is increased. Again, more intense

380

DONALD M. BROOM

pain is likely to have a greater effect. After physical trauma, or during a pathological condition, recovery will often be facilitated by modification of behavior so as to avoid further damage to the affected parts of the body (Wall, 1979). Chronic pain can therefore be functional in that it increases the chances that activity level and type will be modified in an adaptive way. Hence, as explained by Broom and Johnson (1993, p. 29), to suppress pain would in many situations be disastrous. The existence of a feeling of pain in an individual may not be obvious and it seems very likely that in various species extreme pain can occur without any recognizable modification of behavior (Morton and Griffiths, 1985). Indeed, as pointed out by Fraser and Broom (1990 pp. 269-273), pain is not necessary unless some action has to be taken, and animals of different species will vary in the kind of behavioral response to pain that is most adaptive. For example, vocalization when in pain may be advantageous for a young animal or a social animal that might be helped by its mother or by members of its own social group, but disadvantageous when a dangerous predator is close and no effective help from any other individual is likely. Hence, it is not surprising that young pigs, dogs, and humans, which could be helped, make a lot of noise when in pain, but young sheep, which are less likely to be helped against most predators, do not. Indeed, the sheep response to tissue damage is considerable in terms of physiological change (Shutt et al., 1987) and subsequent avoidance of the situation where the damage occurred (Rushen, 1990), so it is extremely likely that they are feeling pain, but they show little behavioral response at the time of the damage. Although it is often possible to hypothesize that in many situations pain has a function that would result in natural selection favoring genotypes in which the individuals were able to feel pain, in other situations it is not easy to ascribe any function to the pain. Bateson (1991) refers to the extreme pain associated with a kidney stone stuck in a ureter and other forms of extreme pain that result in the individual writhing in agony. The writhing could conceivably have a function, but this seems unlikely when recovery from surgery seems to be facilitated by analgesia. Perhaps the various gradations of the lower levels of pain are adaptive but some extreme pain is an inevitable but nonfunctional consequence of the system. The possibility that extreme pain interferes with various aspects of normal functioning, and hence potentially impairs the efficacy of adaptive responses, is often quoted as a reason why endogenous opioid analgesic mechanisms evolved. An individual might be in such pain that it could not show an escape response unless the analgesic opioid inhibited or ameliorated the pain.

WELFARE. STRESS. A N D THE EVOLUTION OF FEELINGS

381

Another critique of the idea that pain is adaptive is that it should not be necessary for there to be such an elaborate within-individual communication system when all of the cells of the body have the same genotype and hence do not need any more than a very simple message. However, there are other elaborate communications systems within the body and a fast, effective system of messages about actual or potential body damage is important. The conclusion of these arguments about the function of pain is that many forms of pain are important for survival but some are likely to be accidental and nonfunctional. Functional pain seems to occur in all vertebrates that have been studied and in some invertebrates. Nonfunctional pain probably also occurs in all animals. There is no reason why there should be any differences between humans and other vertebrates in the proportions of functional and nonfunctional pain.

3. Malaise Malaise is a feeling of illness or discomfort associated with some pathology or inadequacy of bodily function. It is more general in its effects than pain, which is localized in a particular part of the body. There could be wide-ranging effects on the body when there is a reduction in the availability of energy providing resources because these are being used to fight pathogens. Similarly, an accumulation of toxins could have consequences in various parts of the body. In both of these examples, the effects on the brain, perhaps mediated via the peripheral nervous system, could lead to the feeling of malaise. The exact details of the feeling are likely to vary according to the kinds of effects of the toxin or pathogen, but a general feeling of lethargy is common to malaise with various causes. Most people think of malaise, or of feeling ill, as an unfortunate byproduct of infection, but it often affects behavior as well as physiology and its major effect may well be adaptive. Although some effects of parasites or pathogens on host behavior are induced by the parasite or pathogen for its own benefit, many of such effects help the host. As Hart (1988, 1990) points out, animals that are sick are often depressed, lethargic, show no interest in eating, and fail to exhibit body care, but such behaviors “appear to comprise an adaptive behavioral mode that facilitates recovery from illness.” When the immunological and other defenses of the body are having to work hard and consume a lot of energy, it is advantageous for the individual to rest and to be able to concentrate available energy on fighting the cause of the malaise. Even high fever would be adaptive if the net benefit from killing pathogens and suppressing activity outweighed the net cost of tissue damage (Kluger, 1979; Ewald, 1980, 1983; Hart, 1988). The

382

DONALD M. BROOM

general feeling of malaise can be thought of as part of the means by which active defense against pathogens, with its multitude of consequences including increased levels of interleukin 1, can act so that behavior is adaptively modified. There may also be specific aspects of malaise that are adaptive. For example, disorders of the gut may be remedied fastest if no further food is taken in; therefore, a feeling of nausea is advantageous in this circumstance. Too much consumption of a poison, such as alcohol, may overload liver resources and hence also lead to nausea and reduction in further intake of food or poison. Infection or poison-induced nausea is likely to be adaptive, but some nausea, like that induced by certain forms of motion, might be accidental. The nausea, vomiting, and food aversion that occurs in early human pregnancy could be a mechanism for minimizing the risk that toxins from various normally edible foods might adversely affect the developing fetus (Profet, 1992). These feelings are also likely to suppress activity in the mother and could be a consequence of a risk of mechanical damage to the fetus.

4.

Tiredness

Muscular fatigue is a sometimes unwanted consequence of muscular activity, but it may have neural effects that are felt as tiredness. Other forms of exertion and prolonged periods of being awake rather than asleep can also lead to a feeling of tiredness. As sleep tends to occur with a fairly regular rhythm, tiredness could sometimes indicate that the normal time for sleeping has arrived rather than that the individual has been involved in much exertion. Tiredness as an indicator of high levels of exertion can be an adaptive feeling in that it tends to prevent levels of exertion that might be damaging. Tiredness as a prompter of sleeping can be adaptive, just as sleep can, in that it ensures that the individual is inactive and inconspicuous at a time of day when accidents are more likely and predators abound. There may also be some recuperative function.

5. Hunger The feeling of hunger is generally associated with emptiness at some level in the gut, or low levels of nutrients within some organ of the body, or with input from an internal clock that indicates that the time of day when some food intake normally occurs has arrived (Grossman, 1973). As with other feelings, hunger varies from slight to very severe. The stronger the feeling, the more likely it is that the motivational state will be greatly influenced by causal factors relating to hunger. Hunger is generally rather unspecific, but there can be specific hungers that are satisfied only by the consumption of a particular nutrient. Although individuals can overeat and

WELFARE, STRESS, AND THE EVOLUTION OF FEELINGS

383

can select the wrong nutrients, in most circumstances the feeling of hunger would seem to be adaptive. 6.

Thirst

The sensory inputs that lead to feeling thirst come from body fluid concentration monitors, including those in the mouth (Toates, 1986). Thirst varies from a minor feeling, which has little effect on behavior, to an all-pervading feeling, which dominates all behavior in that the individual devotes all possible resources to finding water while conserving it. In most circumstances, the feeling of thirst would seem to be adaptive, although it is possible that some very thirsty individuals behave in a rash way in order to obtain water. 7. Thermal Discomfort The physiological responses to, and consequences of, exposure to very high or very low temperatures have been described in detail (Milner, 1970). The feeling of discomfort when too hot or too cold results in changes in behavior of various kinds (Broom, 1981, pp. 108-113). Most of the consequences of the feelings that have been described have the effect of increasing survival chances. 8. Fear

Fear is a feeling that occurs when there is perceived to be actual danger or a high risk of danger. Although a fast looming visual stimulus, a sudden sound, or an acrid smell might elicit a startle response, the feeling of fear depends on some more complex analysis in which current sensory input is compared with memories of previously experienced events. Blood and Studdert (1988) define fear as “a normal emotional response to consciously recognized external sources of danger.” Hence, the recognition precedes the feeling rather than being a part of it. There are two very different kinds of response to fear (Broom, 1981, pp. 162-175). One is to actively escape or defend, while the other is to rapidly and substantially reduce behavioral and physiological activity so as to render the individual inconspicuous to an attacker. The propensity of predators to notice fleeing prey and ignore immobile objects is utilized in this response. The physiological changes associated with feeling fear are also of two kinds (Broom and Johnson, 1993, pp. 92-107). If active responses are a possibility, first the adrenal medulla response and then the adrenal cortex response prepare the individual for precipitant activity. If suppression of movement is important, then bradycardia, which tends to suppress movement, occurs. In many cases, however, either the active or the passive kind of response could be shown and animals may show first one and then the other. For

384

DONALD M. BROOM

example, when a young domestic chick is startled, it usually freezes for a while and then shows escape activity, the proportions of each response depending on the conditions and the previous experience of the bird (Broom, 1969a,b). The freezing response can be very effective as a means to survive predator attack, but it would be extremely inappropriate as a response to the immediate risk of a large quantity of rocks falling onto the individual. Hence, the efficiency of analysis of the situation and appropriate selection of response is of great importance. If a predator was detected that would not be deceived by a freezing response but that could be outrun, then it would again be important to choose the response correctly. The response to feeling fear could also be maladaptive if carried to an extreme. For example, an extreme form of the freezing response is catalepsy. This can be an adaptive response to the presence of a predator but if it leads to hypovolemic shock, the individual can be damaged by the response itself. Shock reactions are on the borderline between the life saving and the clearly disadvantageous, being the former when predator attack is a major factor in the life of the individual, but being the latter if predation risk is insignificant.

9. Anxiety Anxiety is “a feeling of uneasiness, apprehension or dread” (Blood and Studdert, 1988). It depends on an ability t o predict future risk, usually based on recent stimuli and always on previous experience. The horse that is reluctant to pass a gateway where it once had a frightening experience is showing the characteristic signs of an individual feeling anxiety. Whenever information about disturbing events is stored, there must be a potential that there will be recall of this information and consequences of the recall that activate emotional systems in the individual. The recall can occur at an entirely cerebral level, that is, with no current stimulus involved, but with consequences of the feeling that may be physiological as well as conceptual. The feelings of anxiety probably potentiate the response to the risk. In some cases, showing an appropriately high level of response to a risk can be life saving, while in other cases anxiety can be damaging to the individual (Nesse and Williams, 1995). A genotype that promoted the ability to explore previous experiences in the brain and to feel anxiety when high levels of risk were deduced would be strongly favored in natural selection, given the very considerable advantages of anxiety in some circumstances. The disadvantages of unnecessary anxiety would not be sufficient to counterbalance the advantages of having the feeling. It could be that anxiety was much more important to our ancestors than to present-day humans so that many individuals, and some in particular, feel much anxiety that confers

WELFARE. STRESS. AND THE EVOLUTION OF FEELINGS

385

no benefit to them. Indeed, the propensity to feel much anxiety might be expected to be greater in females because of the difficulties of maternal care when offspring are young and very vulnerable. The anxiety is of such importance at that time that its selective advantage then outweighs any disadvantages associated with its persistence throughout life during times when it is unnecessary or harmful to the individual. Anxiety may indeed be greater at times of parental responsiblity but may be difficult to switch off at other phases in life.

10. Grief Grief and sorrow are unpleasant feelings associated with unwanted life events, particularly those connected with social relationships. Grief is a stronger version of sorrow. For the most part they are private feelings with no effect on others, although their outward manifestations may have an effect of maintaining social position, preserving good relationships, or informing others, honestly or dishonestly, of the properness of the individual’s feelings. When a significant unpleasant event occurs during life, for example, an important failure to succeed in some endeavor, or the loss of a social relationship, increased brain processing might occur. However, that increase in brain processing and activation of emotional centers that result in grief could then increase even more by positive feedback in a way that may help the individual to respond adequately to the event, both immediately and in terms of later learning. Hence, the grief could be functional in that, by amplifying the effects of the event, it ensures that something of importance is not processed too rapidly and then stored without sufficient likelihood of affecting future decision making. Grief is only one example of a feeling as an amplifier of the significance of important life events. Simple brain processing of perceived events may not allow sufficient allocation of attentional and analytical resources to those events that demand more consideration if the individual is to survive and reproduce effectively. However, mechanisms whose general effect is amplification can operate too strongly in a particular instance or may frequently exaggerate the significance of real situations in certain individuals. Grief, like other feelings, seems to have adaptive advantages on some occasions but not on others. The most obvious situation in which grief is felt by humans is when a relative or close friend departs or dies, a circumstance in which the amplifying effect of grief helps in the evaluation of the importance of relationships. There are many reports of pets, especially dogs, and monkeys (de Waal, 1996, pp. 54-56) showing the same sort of behavior that humans show in such circumstances. This behavior is sometimes described as mourning and has also been described in horses, pigs, and elephants. As the arguments for the selective advantages of grief would

386

DONALD M. BROOM

apply to other species, especially those that have a long-lasting and elaborate social structure, it would seem likely that, as de Waal proposes, some degree of grief occurs in other species. 11. Frustration

The existence of elaborate motivational mechanisms in a wide variety of animals has already been mentioned. Whenever a well-organized decisionmaking process exists, the individuals will sometimes be unable to do what they most want to do. Broom (1985) said that an animal is frustrated “if the levels of most of the causal factors which promote a behavior are high enough for the occurrence of the behavior to be very likely but, because of the absence of a key stimulus or the presence of some physical or social barrier, the behavior cannot occur.” The feeling of frustration could originally have arisen as an accident because of some alternative channeling of output from the decision-making system. This feeling could be damaging to the individual in that it might tend to make it carry out behavior such as self-mutilation or to activate physiological systems that either use up energy or promote pathological effects. Alternatively, the feeling of frustration could lead to behavioral and physiological changes that help the individual to cope with the frustrating situation. Some behavior changes resulting from frustration, such as stereotypies, might be adaptive in some situations but harmful in others (Mason, 1991; Broom and Johnson, 1993, pp. 139141). Frustration is likely to be an important feeling in most complex animals. Because there will have been strong selection for systems that allocate resources effectively (Broom, 1981, pp. 79-96), as Houston (1997) points out, animals will be strongly motivated to perform certain important activities, and suffering is likely if they are prevented from performing them. The feeling of frustration, which at a high level might be referred to as suffering, is therefore likely to occur in a wide range of animals.

12. Guilt The feeling of individuals who have behaved in a way that they or their social group members would normally condemn or punish is referred to as guilt. People report that they feel guilt in a variety of situations. It is clear that the brain processing underlying the feeling must often involve sophisticated analysis of actions in relation to experience of the consequences of such actions for the individual carrying them out and for others. This brain activity could easily have been the origin of certain wide-ranging effects that are now grouped together as the feeling of guilt. The consequences of feeling guilt are often changes in behavior described in some detail for dogs and primates by de Waal (1996). One of the human

WELFARE. STRESS, AND THE EVOLUTION OF FEELINGS

387

responses is blushing, dogs may hang their heads, and monkeys may behave in an unusually submissive way and show “grin” expressions on their faces. The feeling of guilt could be advantageous to the individual in that it forces attention channels and processing capacity to be allocated to consideration of a situation of importance in relation to future decision making. However, behavior generally associated with guilt is shown much more often in social situations than by individuals that are by themselves. The information conveyed to others in the social group could be of value to them and might therefore increase the spread of a gene promoting the feeling and expression of guilt because of close relatedness or the potential for reciprocal altruism. However, the guilt expressor might also derive a direct social advantage in that he or she would be perceived as continuing to be an honest and constructive member of the group. Perhaps without the expressions of guilt he or she would be expelled from the group. Another way of putting this is that the blushing, o r other behavioral manifestation of guilt, is an honest signal. Blushing can also be a contrived signal with the intent of making the blushing individual more sexually attractive to the observer. 13. Depression

Depression is a clinical condition associated with certain neurological disorders and with extreme malaise, fear, anxiety, grief, or frustration. It can be described as a feeling separate from all of these, as it is reported by people as being different. Depressed people describe how nothing matters to them and most of their normal activities are reduced or absent. These behavioral signs can exist in individuals of other species, at least in mammals and birds. An experimentally induced parallel is the “learned helplessness” described in laboratory animals by Seligman (1975). Some of the possible advantages of feeling depressed have already been described as extreme forms of fear, grief, and so on. In social situations in which the individual is performing very badly, behavior and physiology indicative of depression may occur. Extreme examples are the tree shrews and rodents that have been defeated in fights with rival conspecifics but that remain caged with them (Koolhaas et al., 1983; von Holst, 1986). These individuals show very depressed behavior and often die quickly. However, passive responses in such a situation, and perhaps the feeling of depression itself, may be an effective strategy for avoiding future attack (Mend1 et al., 1992). Even the tree shrews might benefit from such behavior if they had the opportunity to escape completely from the rival. Similarly, some human depression could be specific to perceived failure in a specific situation, such as inability to compete adequately in a particular segment of society, and could reduce the chances of attack by dominant individuals (Nesse and Williams, 1995). Depression, linked with the possibility to move to a differ-

388

DONALD M. BROOM

ent segment of society in which coping is easier, could be an effective response. The depressed period could also allow time to work out a better life strategy. However, it would appear that in some cases, feeling depressed is maladaptive, perhaps because there is no viable alternative lifestyle. 14. Boredom

Boredom is a feeling associated with a lack of novel input, perhaps with a lack of input in total. Its occurrence in various animals has been discussed by Wemelsfelder (1993). She refers to animals in impoverished conditions increasingly directing their behavior toward inadequate stimuli, exaggerating normal behavior, and establishing stereotypies. A threshold in the increasing series, which she defines as boredom, is “that state of behavioral fixation in which the animal’s orientation towards a novel stimulus loses its inquisitive and manipulative character.” The feeling of boredom may arise because of severe sensory deprivation or because those potentially interesting stimuli that are detectable are repeated exactly. In the first case, the individual might be in a plain, empty cage; in the second case, there might be a machine present that continually undergoes the same movements with the same periodicity. In either case a sequence of behavior whose function is to carry out useful exploration is continued until lack of useful consequences tend to result in its inhibition. At the same time, or even in the absence of the attempts at exploration, a feeling of boredom arises and increases, sometimes associated with the abnormalities of behavior described here. It may be that brain function can be impaired when there is too little input, or too little new input, for a prolonged period. Studies of sensory deprivation suggest that this is so. It may be that the existence of the feeling of boredom, with its concomitant behavior abnormalities, provides sufficient neural activity to maintain the brain in a low input situation. Boredom that continues to the point where inadequate input results in greater and greater risk of neural system damage may have stronger and stronger effects on the individual, which, although more and more unpleasant, are more likely to prevent the damage. 15. Loneliness

In social animals, social contact is often important for survival of the individual as well as for reproduction. Hence, it is not surprising that a specific feeling can occur when inadequate social contact occurs. In our human ancestors, predator avoidance, food finding, and control of the physical environment were all very much facilitated by, or perhaps possible only in, a group-living situation. Hence, an individual separated from all

WELFARE, STRESS, A N D THE EVOLUTION OF FEELINGS

389

conspecifics might feel lonely and make all efforts to restore adequate social contact. 16. General Suffering All of the negative feelings mentioned so far could constitute suffering if extreme enough. Combinations of feelings might also be referred to as suffering. Individuals with a range of negative feelings might be able to respond by moving to a new living place, so suffering could lead to adaptive responses. In most cases, however, the most extreme forms of suffering will not have beneficial effects. The various forms of suffering may well be adaptive when moderate but neutral or harmful in relation to fitness when severe. The form of suffering that occurs when there is real or perceived failure to cope with the environment is sometimes called despair, which in its extreme form will probably not help the individual to survive. However, a genotype that promoted the existence of such feelings, and their consequent behavioral manifestations, might spread because close relatives observing the behavior might respond in a way that preserved their lives. 17. Jealousy

Jealousy is a feeling that occurs in social situations when another individual is perceived to have achieved something that the subject would like to have achieved. The term envy may be used when possession of an object or a position, rather than a social relationship, has been obtained. In both of these cases there may be an element of frustration. Some of the most extreme examples of jealousy are those feelings of male humans when they perceive that the female whom they regard as their mate is courted by or attracted to another male. A large proportion of murders of women arise from male jealously and they refer to the relative uncertainty about paternity as compared with the certainty about maternity (Nesse and Williams, 1995). Hence, it would seen that the feeling of jealousy can be adaptive in that it reduces the chances that males will lose their mates or invest resources in offspring that are not their own. The biological situation can exist in various species in addition to humans; therefore, the feeling may also exist in them. On some occasions, jealousy may involve a complex analysis of what might have been achieved as compared with what has been achieved. Such comparison can be useful to the individual and the feeling of jealousy may draw attention to the situation in such a way that future performance is improved. Excessive feelings of jealously or envy can be damaging. 18. Lust

There are times when eagerness to behave in a certain way or to obtain a certain objective is so great that a generally pervading feeling of lust

390

DONALD M. BROOM

exists in the individual. A clear-cut example of this is the lust of a male mammal to mate with a female. The sequence of events following exposure to an actually or apparently receptive female is an increase in plasma testosterone, which results in approach and display and, by means of a positive feedback loop, more testosterone production, more approach and display, and so on. The feeling of the individual is an increasing lust for mating with that female. Other resources can elicit lust to various degrees. The biological value of this feeling of lust is to increase the chances that the objective, for example, successful mating, will be achieved. A hazard of lust is that antisocial or directly damaging results may occur.

19. Anger Anger is a feeling of intense displeasure in situations in which the individual is not able to control events adequately. The lack of control may be because of the actions of other individuals, or changes in the physical world, or inadequacies in the abilities of the individual who becomes angry. The feeling of anger includes emotional components and may involve aggressive and violent behavior. Anger may be preceded by other feelings, especially frustration or pain but also fear, anxiety, grief, or lust. An individual who is angry can sometimes achieve objectives that would not otherwise be achieved. Some of these objectives are of considerable importance in terms of individual survival or of reproduction. On other occasions, however, anger can have damaging effects on the individual. However, as these damaging effects may be relatively slight, even if anger results in fitness increment only occasionally, those genotypes that facilitate expression of the feeling could spread in the population. The effects of anger are often more extreme where substantial increases in plasma testosterone can occur, that is, in males. The ideal balance, in terms of fitness, between ready anger and retention of individual control over social relationships and other environmental variables will vary according to the effectiveness of the anger in achieving objectives and the kind of environment in which the individuals live. Highly organized societies with much reciprocal altruism leave less room for anger. 20. Sexual Pleasure The feeling at orgasm is much more wide-ranging in its effects than might be expected from the localized nature of the receptors involved. There can be little doubt that genotypes that led to expression of substantial sexual pleasure survived better in the population than those that led to little sexual pleasure because their bearers would have made more efforts to achieve mating. Indeed, it would be predicted that sexual pleasure would be one of the most substantial forms of pleasure because otherwise individuals would not make effective mating a sufficiently high priority objective. This

WELFARE. STRESS, AND THE EVOLUTION OF FEELINGS

391

argument seems to be widely accepted in relation to males. However, Baker and Bellis (1995) seem reluctant to use the same argument for females, as they present (p. 49) two “hypotheses concerning the functioning of the female copulatory orgasm,” which are (1) that it increases the chances that the female will lie down after copulation in order to reduce sperm loss, and (2) that the orgasmic movements tend to suck up sperm during copulation. Hypothesis (1) receives little support from Baker and Bellis, but the more important adaptive advantage of sexual pleasure in general and the orgasm component in particular must surely be to increase the likelihood that copulation will be repeated. All species would be expected to have some feeling of sexual pleasure, but in those species in which mating would reduce fitness in certain seasons, this feeling might be restricted to the appropriate periods during the year. Some difference in the mechanism that leads to sexual pleasure would also be expected according to whether the ideal was one single mating or multiple matings. A further possible advantage of sexual pleasure that would be more likely to occur in species in which multiple matings and shared parental care occur is that the feeling might serve to encourage bonding between the partners.

21. Eating Pleasure When animals eat, a variety of sensory systems are in operation, and it is not surprising that a generalized feeling of pleasure resulting from the procedure of eating and the taste sensations has arisen. Anticipation of the pleasure of eating would result in one of the causal factors promoting eating. Hence, this feeling would tend to make individuals work harder for food than they would if there was no such pleasure. As some of the pleasure is likely to emanate from sensations obtained during feeding on appropriate foods, the feeling would also have the function of encouraging the individual to eat appropriate rather than inappropriate foods. Some eating pleasure occurs when specific foods that are damaging to the individual are consumed or when overeating occurs, so the effects of this feeling are not always adaptive. It would seem likely that most species with complex nervous systems have some eating pleasure. The particular actions or substances that elicit this pleasure will vary from species to species. However, Cabanac (1979) described how the relative proportions of different strengths of sugar solutions drunk by rats and humans were very similar. Hence, it would appear that this aspect of eating pleasure is similar in rats and humans. 22. Exhilaration Pleasure is often reported by people as being felt when they walk in the countryside, take various forms of exercise, reach a mountaintop, or stand

392

DONALD M. BROOM

by the sea or on a boat. Some of this feeling may be associated with inhaling a high oxygen concentration or a certain small quantity of ozone. Other aspects may reflect a high level of control over muscular activity. Yet another aspect of the feeling may indicate an action of endogenous opioids in the brain. Each of these is likely to be promoting the fitness of the individual; therefore, a feeling that increases the chances of such effects is of value to the animal.

23. Other Sensory Pleasure A wide variety of events in life can have effects that are detectable by sense organs and that result in pleasant feelings. Particular odors, sounds, sights, or mechanoreceptor inputs can evoke pleasure. In some cases these events are brief, or are subtle and difficult to appreciate. The feeling of pleasure can be an indicator of good sensory and brain function and thus help with the monitoring of body functioning. Other sensations are associated with important life events, so their occurrence should be promoted. Just as with eating pleasure, some of the events may not be beneficial for that individual at that time; not all of such feelings are adaptive.

24. Achievement Pleasure A variety of intellectual tasks, as well as some that combine physical effort and much brain activity, can result in feelings of pleasure. The feeling could have been a by-product of the processing of information originally but its continuation would have been promoted if the feeling itself conferred an advantage on the individuals. Perhaps the pleasure resulting from effective high-level brain processing helps to make energy available for brain functioning or changes the biochemical or electrical characteristics of a region of the brain so that further efficient brain functioning can occur. In the competitive and difficult social world in which many animals live, those animals that process information in the brain efficiently are at a considerable advantage over those that do not; feelings that promote this should become widespread.

25. General Happiness Intense pleasure of any kind can lead to people declaring themselves to be happy. If several different kinds of pleasure are combined, this is more likely. However, most people who are asked to say what happiness is will refer to the absence of problems as an important part of happiness. When Cabanac (1979) refers to a widespread control of behavior through pleasure seeking, a major part of this must be the diminution of any bad feelings. It is difficult to identify a general feeling of happiness, but the seeking of this feeling is clearly a major factor in the life of complex animals.

WELFARE. STRESS, AND THE EVOLUTION OF FEELINGS

393

AS PART OF COPING SYSTEMS D. FEELINGS

Animals have a wide range of systems for trying to cope with their environments. These coping systems include positive and negative aspects: useful actions should be carried out and resources obtained but damaging events should be avoided. Coping is having control of mental and bodily stability (Fraser and Broom, 1990). Coping and control systems are described at length by Broom and Johnson (1993, Chapters 2 and 3). A substantial part of physiology and behavior plays a role in attempts to cope with the environment. This varies from straightforward, low-energy homeostatic mechanisms to high-urgency emergency responses. The general idea that feelings often have a function and help individuals to cope with their environment has been advanced (see Section CJ), especially by Dawkins (1990,1993), Broom and Johnson (1993), and Broom (1996). The central thrust of the arguments is that all feelings can be functional to a greater or lesser extent. Feelings are a part of the biology of the individual that has evolved. They are used in order to maximize its fitness, often by helping it to cope with its environment. The coping systems used by animals operate on different time scales. Some must operate during a few seconds in order to be effectual, others take hours or months. Optimal decision making depends not only on an evaluation of energetic costs and benefits but on the urgency of action: in other words, the costs associated with injury, death, or failure to find a mate (Broom, 1981, p. 80). In the fastest acting, urgent coping responses, such as avoidance of predator attack or risk of immediate injury, fear and pain play an important role. In longer time scale coping procedures, where various risks to the fitness of the individual are involved, feelings rather than just intellectual calculations are among the causal factors affecting what decisions are taken. In attempts to deal with long-term problems that may harm the individual, aspects of suffering contribute significantly to how the individual tries to cope. In the organisation of behavior so as to achieve important objectives, pleasurable feelings and the expectation that these will occur have a substantial influence. The general hypothesis advanced is that whenever a situation exists in which decisions are made that have a big effect on the survival or potential reproductive output of the individual, it is likely that feelings will be involved. This argument applies to all animals with complex nervous systems, such as vertebrates and cephalopods, and not just to humans. Feelings are not a minor influence on coping systems, they are an important part of them. In circumstances in which individuals are starting to lose control and fail to cope, feelings may exist. These feelings might have a role in damage limitation, which is functional. However, they might also occur when the

394

DONALD M. BROOM

individual is not coping at all, and hence the feelings have no survival function. Extreme suffering or despair are probably not adaptive feelings, but an observer of the same species might benefit, and a scientist might use indications of such feelings to deduce that the animal is not coping.

11. WELFARE, STRESS,A N D FEELINGS A. DEFINING WELFARE, STRESS,AND HEALTH: THERELEVANCE OF FEELINGS

I.

Welfare

Welfare is a term restricted to animals, including humans, and hence not used for other organisms or inanimate objects. It is used in science and in legislation and therefore must have a meaning precise enough for such use. Welfare refers to a characteristic of an individual rather than to something given to it, it must be measurable in a scientific way, and it must vary over a scale from very good to very poor (Curtis, 1986; Duncan, 1987; Broom, 1988, 1991, 1996; Broom and Johnson, 1993, pp. 74-76). The original use of the word welfare, meaning how well an individual fares or goes through life, is followed in the definition proposed by Broom (1986): the welfare of an individual is its state as regards its attempts to cope with its environment. Its state includes how well or how badly it is coping and how much difficulty it is having in coping. As emphasized by Broom (1991, 1996) and Broom and Johnson (1993, pp. 80-82) the feelings of the individual are an important part of that state. The assessment of how good or how poor the welfare is depends on a wide range of measures of behavior, physiology, brain functioning, immune system functioning, pathology, injury, and life expectancy . This definition of welfare was not found satisfactory by Duncan (1993, 1996), who considered that it understated the importance of feelings. Indeed, this was a valid criticism of the original paper (Broom, 1986), which did not refer to feelings. An aim of the present chapter is to explain, first, that feelings are important biologically, second, that they are very significant contributors to coping systems, and hence third, that they are encompassed properly in this definition of welfare. The key point, which has been made in Section I, is that feelings have extremely important biological functions, especially in coping systems. A further point is that although feelings are part of coping systems, they do not make up all of them, so if the concept of welfare is to be usable it must refer to all aspects of coping systems and not just to feelings (Broom, 1993). The separation of feelings from all other

WELFARE. STRESS. A N D THE EVOLUTION O F FEELINGS

395

aspects of coping systems in biologically unsound and impractical when welfare assessment is attempted.

2. Stress Two other important terms in relation to welfare are stress and health. As argued in detail by Broom and Johnson (1993, pp. 57-73), the term stress should be limited to refer to something that is bad for the individual. It should not be used to mean any kind of disturbance of homeostasis, and should not be merely equated with adrenal responses. The use of stress to refer to environmental impacts on the individual, whatever their consequences, and hence the idea that stress can be good for an individual is confusing and renders the term virtually useless. Similarly, the fact that adrenal cortex activity can occur during beneficial activities such as mating and prey catching and may not occur during hemorrhage, dehydration, or harmful increases in body temperature means that such activity cannot be used in defining stress. A measure of what constitutes adverse or harmful effects is needed in the definition of stress, and hence, following earlier versions of the definition by Broom (1983) and Fraser and Broom (1990), Broom and Johnson (1993, p. 72) conclude that: “stress is an environmental effect on an individual which overtaxes its control systems and reduces its fitness or appears likely to do so.” The “environment” here is that which is outside the brain. Hence, when an individual attempts to cope with its environment but fails to do so, welfare will be poor and stress will occur. The failure to cope would usually be associated with one or more of the unpleasant feelings previously described and, although many aspects of stress are psychological, stress is not itself a feeling. It is possible for stress to occur without any associated feelings. An individual knocked unconscious and otherwise injured so that, despite the operation of various coping systems, death occurs without consciousness being regained is stressed but does not suffer. The relationship between stress and welfare is straightforward. Welfare refers to a wide range of states from very good to very poor, but the welfare of any stressed individual is poor. However, there can be poor welfare without stress when the individual is coping but only with considerable difficulty. For example, on some occasions when a person is succeeding in coping with a minor injury, he or she may be in some pain and may show some physiological and behavioral responses; the welfare is poor, but there is no likelihood of any reduction of fitness and therefore no stress. 3. Health The word “health,” like “welfare,” can be qualified by “good” or “poor” and varies over a range. However, health refers to the state of body systems,

396

DONALD M. BROOM

including those in the brain, which combat pathogens, tissue damage, or physiological disorder. Welfare is a broader term covering all aspects of coping with the environment and taking account of a wider range of feelings and other coping mechanisms than those that affect health, especially at the positive end of the scale. Although people regularly refer to poor health, they sometimes use the word to mean absence of illness or injury in the same way that people refer to welfare when they mean good welfare. However, the precise and scientific use of health and welfare must refer to states varying from very good to very poor. “Health” is encompassed within the term “welfare,” and indeed is a very important part of welfare. B. ASSESSING WELFARE, STRESS, A N D FEELINGS

f. What To Measure? The assessment of welfare is discussed at length by Dantzer and Mormkde (1979), Smidt (1983), and Broom and Johnson (1993), and is not described in detail here. As Mason and Mend1 (1993) and Fraser (1995) have pointed out, no single, all-embracing measure of welfare is available, and indeed, all of the authors mentioned above refer to the necessity to use a range of measurements in studies where welfare following different short-term or long-term treatments is assessed. It is occasionally possible to recognize very poor welfare in a set of individuals using a single measurement, for example, when all of the animals show early death or severe abnormalities of behavior, physiology, or immune system function. However, the range of methods used to try and cope, the range of effects of adversity, and the range of feelings that should be considered means that good studies should use an array of measures. Fraser (1995) describes welfare as a “type 3 concept,” like “safety,” which has multiple attributes so that the results of different kinds of assessment must be weighed one against the other. However, in comparing the welfare of animals in two conditions, each scientific measurement and its interpretation must be objective and unaffected by prejudice about which condition is better. Occasionally the data available may not allow a conclusion to be reached, but eventually the scientific approach will provide data that give information about how good or how poor the welfare is in each case. Once this has been done, a moral judgment about what is acceptable and what is not can be made (Broom, 1996). Part of the assessment of welfare will involve the various indicators of stress, that is, reduced fitness or potential reduction of fitness as deduced from impaired growth, disease risk, serious injury, and so on. Another part of the assessment will concern effects on health ascertained especially from evaluating any pathological effect. A further part will concern observations that indicate the extent to which the individual has good or bad feelings,

WELFARE, STRESS. AND THE EVOLUTION OF FEELINGS

397

for example, the extents to which strongly preferred behaviors can be shown and the magnitude of physiological and behavioral changes usually associated with pain. Many measures of behavior and physiology reveal the extent of difficulty in coping without necessarily giving direct information about unpleasant feelings, and some measures, such as those of failure in immune system function, may not be related to feelings at all.

2.

The Role of Preference Tests

Animals normally prefer what is good for them, and studies that provide information about the strengths of individual preferences are of considerable value in designing conditions for animals and deciding what should or should not be done in procedures involving animals (Dawkins, 1980,1990; Duncan, 1993, 1996; Broom and Johnson, 1993, pp. 145-157). However, preferences are affected by previous experience (Mendl, 1990), and there are various examples of animals choosing something that is not good for them (Duncan, 1978). Dawkins (1990) concludes that several studies with different individuals will often be necessary as a consequence when trying to find out which treatments or conditions are likely to result in good welfare. The preference studies do not provide comprehensive information about the welfare of the individuals, but should be followed up by studies using other welfare indicators to make sure that what is preferred does not lead to poor health or other problems. Carefully designed preference studies can give good indications about some feelings such as pain, thirst, or thermal discomfort. However, they may not indicate whether the behavior is affected most by a positive or by a negative feeling, they d o not always allow discrimination of what kind of feeling is being avoided or sought, and they may not make it possible to recognize some feelings such as guilt and jeolousy. Like measures of abnormalities of behavior or physiology, preference tests cannot tell us exactly what the individual studied is feeling. Preference tests are valuable in some situations but not appropriate in others. Studies of foods and physical and social conditions are readily carried out by measuring how hard animals will work to obtain each resource, and measures of strength of aversion help in determining how painful or frightening some events or stimuli are. However, it would be difficult to assess stunning procedures or disease effects using preference tests, and there are problems in interpreting strong preferences for some harmful foods or drugs (Broom, 1996).

3. Welfare Assessment and Time Sometimes poor welfare is very brief but very severe; at other times, poor welfare is prolonged but slight. Similarly, pleasant experiences may

398

DONALD M. BROOM

be intense or mild, and momentary or extended in effect. Time should be considered as a variable in decisions about welfare. In general (Broom and Johnson, 1993,pp. 109-110), it would seem biologically meaningful to make some assessment of intensity. Then, if PW is intensity of poor welfare, and t is time, the multiplicative effect (PW X t ) can be calculated (Fig. 1). We cannot be sure that a simple multiplicative relationship is correct, and this requires experimental investigation, but the problem should not be ignored.

very good

Poor welfare x time = X Very poor Time very good

(b) Welfare

5

Neutral Poor welfare x time = 3X very Time very good

........................ ........................ ........................ ........................ ........................ Poor welfare x time = 3 X Very poor Time

FIG. I . Relationships between poor welfare and time. The effects of two levels of poor welfare (PW) and two durations of environmental conditions on an individual are shown. From Broom and Johnson (1993).

WELFARE. STRESS, AND THE EVOLUTION OF FEELINGS

399

C. FEELINGS AS PART OF WELFARE We are so aware of our own feelings that it is not surprising that they are at the forefront of our minds when we think of our welfare; indeed, some people refer to feelings as being the sum total of welfare. As discussed at length earlier, the propensity to have feelings has evolved because some of the feelings in each of the general categories considered have some function, and feelings are an important part of mechanisms used in coping with the environment. Hence, it is impossible to consider a concept of welfare without feelings being included in that concept (Dawkins, 1990; Broom and Johnson, 1993, pp. 33-34,8042; Broom, 1996). However, some feelings may be epiphenomena of neural activity (Broom and Johnson, 1993, p. 80) and hence need not affect coping with the environment. If the definition of welfare were limited to the feelings of the individual it would not be possible to refer to the welfare of any individual that was asleep, or anesthetized, or drugged, or suffering from a disease that affects awareness. Neither would any disease be considered to affect welfare unless it altered the feelings of the individual. The welfare of an individual who was dying but was briefly euphoric because of drug administration would be described as very good. Most people would say that welfare can be poor in circumstances in which there are no bad feelings. A further problem results if only feelings are considered in assessing welfare: a great deal of important evidence in assessing welfare in practical studies would not be used. Animals may be studied and found to have neuromas, or extreme physiological responses, or abnormalities of behavior, or immunosuppression, or disease, or inability to grow and reproduce, or reduced life expectancy, but this evidence would not be used to indicate poor welfare unless bad feelings could be demonstrated to be associated with them in such individuals. This argument is already being used by some people to say that systems for housing farm animals are not proved to result in poor welfare, because the abnormalities of behavior and physiology that occur in these conditions have not been linked with certainty to unpleasant feelings. It is often difficult to convince scientists about evidence concerning feelings, and some people are never convinced about feelings in any other individual, even of their own species. Hence, it is dangerous to let decisions about welfare depend entirely on evidence about feelings. We should carry out research in which we try to find out as much as possible about the feelings of individuals whose welfare we are trying to assess. However, we should incorporate such studies in investigations of the whole range of coping methods by studying coping systems in their entirety. Individuals cope by using their immune systems, adrenal responses,

400

DONALD M. BROOM

and behavioral regulatory methods as well as by responding to their feelings. Welfare assessment must take account of all of these. As Broom (1993) and Broom and Johnson (1993, pp. 82-84) point out, welfare is worse when there is an injury than when there is no injury, even if the individual is narcotized or asleep, but the welfare is worse still if the individual is awake and suffering. Similarly, an individual whose immune system functioning is suppressed is coping with its environment less well than an individual with normal immune system functioning, so its welfare is worse even if it is unaware of the immunosuppression. However, the welfare of the immunosuppressed individual would be worse still if it were diseased, especially if the effects of the disease caused suffering. Welfare assessment should involve a combination of studies of feelings and of other factors providing information about coping.

111. SUMMARY

Animals have systems for recognizing harmful or favorable stimuli and for monitoring the effects of environmental conditions on themselves. They also have a range of methods for attempting to cope with perceived or actual adversity. Mechanisms involving learning tend to maximize the chances that things that enhance fitness will happen to them, and tend to minimize the chances that things that reduce fitness will happen. As part of these various methods and mechanisms, positive and negative feelings have evolved. These feelings have a biological role that complements various other anatomical, physiological, and behavioral mechanisms. Each of 24 different kinds of feelings is discussed and the origin and possible function of each is considered. All have some potential for improving fitness and most are likely to have been the subject of considerable selection pressure, but some aspects of feelings are likely to be just epiphenomena of neural mechanisms. With this view that most aspects of feelings have evolved like other biological mechanisms and that they help significantly in coping and responding, a single view of welfare as the state of an individual as regards its attempts to cope with its environment becomes clearer. Feelings are an important part of the welfare of an individual and should be assessed as well as possible. Other coping procedures and effects of the environment on the individual should also be assessed. An effect on an individual that is adverse in the long term is categorized as stress. Programs for trying to evaluate and improve welfare should combine the use of experiments to assess what is important to the individual by measuring the strengths of preferences, with monitoring studies in which feelings and other aspects of welfare are assessed more directly.

WELFARE, STRESS. AND THE EVOLUTION OF FEELINGS

401

References

Baars. J. (1988). “A Cognitive Theory of Consciousness.“ Cambridge University Press. Cambridge. UK. Baker, R. and Bellis. M. A . (1995). “Human Sperm Competition.” Chapman and Hall. London. Bateson. P. (1991). Assessment of pain in animals. Anim. Behov. 42, 827-839. Blood, D. C., and Studdert, V. P., eds. (1988). “Baillitre‘s Comprehensive Veterinary Dictionary.” Baillitre Tindall. London. Broom, D. M. (1969a). Reactions of chicks to visual changes during the first ten days after hatching. Anim. Behav. 17, 307-315. Broom, D. M. (1969b). Effects of visual complexity during rearing on chicks’ reactions to environmental change. Anim. Behav. 17,773-780. Broom. D. M. (1981). “Biology of Behaviour.” Cambridge University Press, Cambridge, UK. Broom, D. M. (1983). The stress concept and ways of assessing the effects of stress in farm animals. Appl. Anim. Ethol. 11, 79. Broom, D. M. (1985). Stress, welfare and the state of equilibrium. In ”Proceedings of the Second European Symposium on Poultry Welfare” (R. M. Wegner, ed.), pp. 72-81. World Poultry Science Association. Celle, Germany. Broom, D. M. (1986). Indicators of poor welfare. Br. Vet. J. 142, 524-526. Broom. D. M. (1988). The scientific assessment of animal welfare. Appl. Anim. Behav. Sci. 20, 5-19. Broom. D. M. (1991). Assessing welfare and suffering. Behav. Process. 25, 117-123. Broom, D. M. (1993). A usable definition of animal welfare. J . Agric. Environ. Ethics 6, Suppl. 2. 15-25. Broom. D. M. (1996). Animal welfare defined in terms of attempts to cope with the environment. Acto Agric. Scarid.. Anini. Sci. Siippl. 27, 22-28. Broom, D. M.. and Johnson, K. G . (1993). “Stress and Animal Welfare.” Chapman & Hall, London. Cabanac. M. (1979). Sensory pleasure. Q. Rev. B i d 54, 1-29. Crook, J. H. (1980). “The Evolution of Human Consciousness.” Oxford University Press, Oxford. Curtis, S. E. ( 1 983). Perception of thermal comfort by farm animals. In “Farm Animal Housing and Welfare” (S. H. Baxter, M. R. Baxter, and J. A. C. MacCormack, eds.), Curr. Top. Vete. Med. Anim. Sci, pp. 59-66. Martinus Nijhoff, The Hague. Curtis, S. E. (1986). The case for intensive forming of food animals. I n “Advances in Animal Welfare Science” (M. W. Fox and L. D. Mickey, eds.), pp. 245-255. The Humane Society of the United States, Washington, D.C. Dantzer, R.. and Mormkde, P. (1979). “Le stress en 6lCvage intensii. Masson. Paris. Dawkins, M. (1977). D o hens suffer in battery cages? environmental preferences and welfare. Anim. Behav. 25, 1034-1046. Dawkins, M. (1980). “Animal Suffering: The Science of Animal Welfare.” Chapman & Hall, London. Dawkins. M. (1990). From an animal’s point of view: Motivation, fitness and animal welfare. Behav. Brain Sci. 13, 1-61. Dawkins. M. (1993). “Through our Eyes Only.” Freeman, Oxford. Dennett, D. C. (lY91). “Consciousness Explained.” Penguin, London. de Waal. F. (1996). “Good Natured: The Origins of Right and Wrong in Humans and Other Animals.” Harvard University Press, Cambridge, MA. Duncan, 1. J. H. (1978). The interpretation of preference tests in animal behaviour. Appl. Anim. Ethol. 4, 197-200.

402

DONALD M. BROOM

Duncan, I. J. H. (1987). The welfare of farm animals: An ethological approach. Sci. Prog. 71, 317. Duncan, I. J. H. (1993). Welfare is to do with what animals feel. J. Agric. Environ. Ethics. 6, Suppl. 2, 8-14. Duncan, I. J. H. (1996). Animal welfare definition in terms of feelings. Actu Agric. Scand., Sect. A, Anim. Sci. Suppl. 27, 29-35. Ewald, P. W. (1980). Evolutionary biology and the treatment of signs and symptoms of infectious disease. J . Theor. Bioi. 86, 169-176. Ewald, P. W. (1983). Host-parasite relations vectors and the evolution of disease severity. Annic. Rev. Ecol. Syst. 14, 465-485. Fraser. A. F., and Broom, D. A. (1990). “Farm Animal Behaviour and Welfare” 3rd ed. C.A.B. International, Wallingford. Fraser, D. (1995). Science, values and animal welfare: Explaining the inextricable convention. Anim. Welfare 4, 103-117. Gallup, G. G. (1983). Towards a comparative psychology of mind. In “Animal Cognition and Behaviour (R. L. Mellgren, ed.). pp. 502-503. North-Holland Publ., New York. Griffin, D. R. (1981). “The Question of Animal Awareness. 2nd ed. Rockefeller University Press. New York. Griffin, D. R. (1984). “Animal Thinking.” Harvard University Press, Cambridge, MA. Grossman, S. P. (1973). “Essentials of Physiological Psychology. Wiley, New York. Guyton, A. C. (1982). “Human Physiology and Mechanisms of Disease.” Saunders, Philadelphia. Hart, B. L. (1988). Biological bases of the behavior of sick animals. Neurosci. Biobehav. Rev. 12, 123-137. Hart. B. L. (1990). Behavioural adaptations to pathogens and parasites: Five strategies. Neurosci. Biobehav. Rev. 14, 273-294. Houston, A. I. (1997). Demand curves and welfare. Anim. Behav. 53,983-990. Humphrey, N. K. (1986). “The Inner Eye.” Faber and Faber. London. Kluger, M. J. (1979). Fever in eitotherms: Evolutionary implications. A m . Zool. 19,295-304. Koolhaas. J . M., Schuurmann, T., and Fokkema, D. S. (1983). Social behaviour of rats as a model for the psychophysiology of hypertension. In “Biobehavioural Bases of Coronary Heart Disease” (T. M. Dembrowski. T. H. Schmidt, and G. Blumchen, eds.), pp. 391-400. Karger, Basel. Mason, G. J. (1991). Stereotypies and suffering. Behav. Process. 25, 103-115. Mason, G. J., and Mend!, M. ( 1 9 9 3 Why is there no simple way of measuring animal welfare? Anim. Werfare 2,301-319. Melzack. R. (1973). “The Puzzle of Pain.” Basic Books, New York. Mendl, M. (1990). Developmental experience and the potential for suffering: Does “out of experience” mean “out of mind?” Behav. Brain Sci. 13, 28-29. Mendl, M.. Zanella, A. J., and Broom, D. M. (1992). Physiological and reproductive correlates of behavioural strategies in female domestic pigs. Anim. Behav. 44, 1107-1121. Milner, P. M. (1970). “Physiological Psychology.” Holt, Rinehart and Winston, New York. Morton, D. B.. and Griffiths, P. H. B. (1985). Guidelines on the recognition of pain, distress and discomfort in experimental animals and an hypothesis for assessment. Vet. Rec. 116,431-436. Nesse, R. M., and Williams, G. C. (1995). “Evolution and Healing.” Weidenfeld & Nicolson, London. Ottoson, D. (1983). “Physiology of the Nervous System.” Macmillan. London. Profet. M. (1992). The antiteratogen theory of morning sickness. In .‘The Adapted Mind” ( J . H. Barkow, ed.), pp. 327-365. Oxford University Press, New York.

WELFARE. STRESS, A N D THE EVOLUTION O F FEELINGS

403

Rodd, R. (1990). “Biology. Ethics and Animals.” Oxford University Press, Oxford. Rushen. J. (1990). Use of aversion learning techniques to measure distress in sheep. Appl. Anim. Behav. Sci. 28, 3-14. Seligman, M. E. P. (1975). “Helplessness: On Depression, Development and Death.” Freeman, San Francisco. Shutt, D. A,, Fell, L. R., Cornell, R., Bell, A. K., Wallace, C. A,, and Smith, A. I. (1987). Stress-induced changes in plasma concentrations of immunoreactive p endorphin and cortisol in response to routine surgical procedures in lambs. Ausr. J. B i d Sci. 40,97-103. Skinner, B. F. (1974). “About Behaviorism.” Jonathan Cape, London. Skinner, B. F. (1978). “Reflections on Behaviorism and Society.” Prentice Hall, Englewood Cliffs, NJ. Smidt. D., ed. (1983). “Indicators Relevant to Farm Animal Welfare,“ Curr. Top. Vet. Med. Anim. Sci.. Vol. 23. Martinus Nijhoff. Dordrecht. The Netherlands. Sommerville, B. A.. and Broom, D. M. (1998). Olfactory awareness in domesticated animals. Appl. Anim. Behav. Sci. (in press). Swenson, M. J., and Reece, W. 0. (1993). “Duke’s Physiology of Domestic Animals,” 11th ed. Comstock, Ithaca, NY. Toates, F. (1986). “Motivational Systems.” Cambridge University Press. Cambridge, UK. von Holst, D. (1986). Vegetative and somatic components of tree shrews’ behaviour. J. Aurvn. Nerv. Syst., Sicppl., pp. 657-670. Wall, P. (1979). On the relation of injury to pain. Pain 6, 253-264. Weiskrantz, L. (1 987). Neuropsychology and the nature of consciousness. I n “Mindwaves” (C. Blakemore and S. Greenfield, eds.). pp. 307-320. Blackwell, Oxford. Wemelsfelder, F. (1993). The concept of animal boredom and its relationship to stereotyped behaviour. I n “Stereotypic Animal Behaviour” (A. B. Lawrence and J. Rushen, eds.), pp. 65-95. C.A.B. International, Wallingford. Wiepkema. P. R. (1985). Abnormal behaviour in farm animals: Ethological implications. Neth. J 2001.35,279-289.

This Page Intentionally Left Blank

ADVANCES IN THE STUDY OF BEHAVIOR, VOL. 21

Biological Conservation and Stress HERIBERT HOFER AND MARION L. EAST MAX-PLANCK-INSTITUT FUR VERHALTENSPHYSIOLOGIE

D-82319 SEEWIESEN POST

STARNBERG

GERMANY

I. INTRODUCTION In this contribution we review why stress has important implications for biological conservation and consider practical ways in which conservationists can identify and tackle problems caused by stress. We take an evolutionary approach to stress, its consequences, and the adaptations that permit animals to cope with stress. Using information from several scientific disciplines we outline the factors known to cause stress and the Darwinian fitness consequences of stress. Increasingly anthropogenic factors such as environmental pollution, tourism and leisure activities, hunting, noise, and global warming are thought to cause stress. We investigate when anthropogenic factors are likely to generate stress, what conditions favor detrimental fitness consequences, and which stress responses provide a reliable indicator of a potentially harmful situation. There is extensive literature on both stress and biological conservation, but there has been little integration of the two disciplines (for an admirable exception, see Hoffmann and Parsons, 1 9 9 1 ) , and there has been no comprehensive review of the importance of stress to conservation. In part it may be because biological conservation studies and stress studies tend to focus on different parameters. Biological conservation generally concentrates on populations and species communities and rarely considers mechanisms operating at the level of the individual that may be ultimately responsible for population and community changes (e.g., Frankel and SoulC, 1981; SoulC, 1986; Primack, 1993). Research on stress is normally concerned with individual organisms and their developmental, behavioral, metabolic, endocrinological, and immunological responses to stress, but it rarely (for exceptions, see Fraser and Broom, 1990; Hoffmann and Parsons, 1991) considers the consequences of stress for the reproductive success of individuals and the persistence of populations. It is the purpose of this review to show that (1) stress is a valuable concept in biological conservation; 405

Copyright 0 I Y Y X by Academic Press All rights of reproduction In any form reserved IW65-1454/YX $25 00

406

HERlBERT HOFER A N D MARION L. EAST

(2) conservation efforts often generate stress and unless this is appreciated no effort will be made to minimize stress: and (3) the success of conservation efforts could be improved if detrimental Darwinian fitness consequences of stress were minimized. Ideally, a theory of stress and biological conservation should guide conservation actions by accurately predicting the likely response to both natural and anthropogenic forms of stress for individuals, populations. species, and communities. Currently no such theory is available. However, useful elements for such a theory can be drawn from many disciplines such as population biology, evolutionary genetics, evolutionary ecology, physiological ecology, ethology, animal welfare, immunology, and endocrinology. We suggest that a Darwinian approach is essential for any theory of stress in biological conservation. This approach considers the stress response of organisms as an evolved trait with an adaptive value and provides a yardstick (Darwinian fitness) to measure the consequences of anthropogenic and natural sources of stress that is of direct relevance to conservation. In this review we first consider the problem of defining stress and its relevance to conservation, and how an evolutionary framework can be incorporated into studies of stress that are relevant for conservation. We then review how stress has been studied and develop a set of criteria to evaluate published work. We review the natural history of stress, in particular factors influencing inter- and intrapopulation variation in the stress response, and outline elements of a theory of stress in biological conservation. This is followed by a survey of indicators of stressed states. We then discuss the potential €or anthropogenic environmental factors such as pollution, disturbance, hunting, noise, and climatic (global) warming to cause stress and evaluate whether stress responses to these anthropogenic stimuli can be equated to natural stress stimuli. We also consider whether research and conservation activities should be considered potential stressors. We finally discuss actions that minimize the occurrence and impact of stress in conservation research and management. We show that such actions readily fall into two categories: minimizing the occurrence of stress and maximizing the ability of individuals to cope successfully with stress. We have attempted to use examples from a diverse range of taxa, including invertebrates, although as most studies have been done on birds and mammals, this review is necessarily biased toward them. We use four case studies to illustrate the difficulties and progress that stress studies in biological conservation have made: (1) the mountain pygmy possum and development of tourist facilities in Australia; (2) the mass die-off of seals in the North Sea in 1988; (3) whale watching; and (4) disturbance of Antarctic penguin breeding colonies by tourism.

BIOLOGICAL CONSERVATION AND STRESS

IN 11. STRESS

A

407

CONSERVATION BIOLOGY CONTEXT

In this section we develop a concept of stress for biological conservation and a framework for studies of stress that is relevant to conservation. A N D STRESS RESPONSE A. STRESS, STRESSORS,

Currently there is no generally accepted definition of stress, or agreement on how to use the word. Ambiguity arises because the word has been used to describe changes in or states of the environment that may cause a change of the organism’s internal state (e.g., “heat stress”), the internal state of the organism, and the actions of the organism in response to environmental stimuli (Toates, 1995). Changes in or states of the environment are often called stressors, the internal state of the organism is called a stressed state, and the actions of the organism, the stress response. The inability to agree on one rigorously defined and universally accepted concept of stress has led some authors to believe that stress as a scientific concept should be abandoned (Rushen, 1986). If a similar rigor were applied to other biological disciplines, then ethologists would abandon concepts like territoriality and dominance (see Kaufmann, 1983), and population ecologists would dispense with density dependence or population regulation (e.g., den Boer, 1990, 1991; but see Dennis and Taper, 1994; Sinclair and Pech, 1996). All these terms continue to be used because they can be rigorously defined in some way. They are also useful because they are parsimonious in the sense that they summarize in an economic manner a host of phenomena that would otherwise be cumbersome to describe. For the purpose of biological conservation, a definition of stress should be theoretically well founded, useful in practice, and help predict the consequences of conservation actions. Before we explain which concept of stress we consider useful for this purpose, it is necessary to look at the criteria that have been used to define stress. A systems view of stress (Moberg, 1985; Toates, 1995) is helpful in this respect. It acknowledges that organisms need to evaluate potential or actual challenges to internal homeostasis (mental and bodily stability) and then act in such a way that their state is restored to the one prior to that of a challenge (Fig. 1). The criteria that distinguish concepts of stress include the following: 1. An organism’s experience of an environmental stimulus. A general concept of a stress or stressor is anything that threatens (Chrousos et al., 1988) or changes internal homeostasis (Sapolsky, 1994). This concept does not distinguish between pleasant or aversive stimuli, and thus includes food or sexual encounters as stressors. More restricted definitions of stressful

408

HERIBERT HOFER AND MARION L. EAST

INPUT

OUTPUT

I LateOi’ environ ment

+

\

00 lenviron mental st mrlus

internal

states

/I external actions

-

l

”

J

l

FITNESS CONSEQUENCES

FIG. 1. A systems view of the links between environmental stimulus, organismal state, and organismal response in the context of stress. Changes in or states of the environment may be perceived by the organism as a potential challenge to internal homeostasis. The organism then either finds itself in an actual state of deviation from homeostasis or anticipates such a deviation. An evaluation of the likely success of potential internal andlor external actions to restore homeostasis may preceed such actions. If the actions fail to restore homeostasis, then the cycle starts again and eventually may cause pathological changes to the organism. Partially derived from diagrams and discussions in Moberg (1985). Broom and Johnson (1991), and Toates (1995). The numbers refer to the criteria that distinguish different concepts of stress (see text).

situations stipulate that the organism’s experience of the stimulus is aversive rather than pleasant, a view widely adopted in animal welfare and psychology (see Broom and Johnson, 1991).

2. The anticipation of environmental stimuli. The absence of a stimulus may create a physiological and behavioral response (“frustration”) reminiscent of a response that would in other circumstances be considered a stress response (Toates, 1995). Anticipation of aversive stimuli under certain conditions may trigger a more substantial internal response than the actual stimulus itself (Arthur, 1987). Thus, the state of the environment in relation to the animal’s control, familiarity, and expectation may be important (Toates, 1995).

3. The temporal pattern of stimuli. In evolutionary ecology, stimuli are distinguished by their predictability and their frequency of occurrence relative to the generation time of an organism. Single, unpredictable events (e.g., an attempt to capture an animal) are contrasted with continuous stimuli (e.g., pollution of the environment with heavy metals). Hoffman

BIOLOGICAL CONSERVATION AND STRESS

409

and Parsons (1991) consider the first type to be stress, whereas Rollo (1995) labels these as “(strong) disturbances” and reserves the term stress for the second class of stimuli.

4. Kinds of internal changes. Selye’s (1946) original concept viewed the stress response as composed of a specific response and a nonspecific response common to a wide variety of environmental stimuli and determined only by the intensity of the stimulus. He termed this the “general adaptation syndrome.” The responses involved include cardiovascular effects and a hormonal response associated with (1) the sympathetic-adrenal medullary system (catecholamines such as epinephrine and norepinephrine, also known as adrenaline and noradrenaline); and (2) the hypothalamicpituitary-adrenocortical (HPA) axis (also known as the adrenocortical response: adrenocorticotropic hormone (ACTH), cortisol, corticosterone, other corticosteroids). Thus, he tied the definition of stress to an explicit response mechanism. Such a criterion may be too narrow and too unspecific (Toates, 1995). Too narrow, because some life-threatening situations fail to evoke a corticosteroid response (Freeman, 1985) and invoke physiological responses other than these hormonal responses (Fraser et al., 1975;Moberg, 1987), and because it would exclude taxa such as invertebrates that do not have these systems. Too unspecific, because similar behaviors or hormonal changes may be observed under both pleasant and aversive stimuli. For instance, toxins, infections, pain, sleep, and exercise all stimulate the hypothalamus and lead to an increase in corticosteroid secretion (Rivier, 1989), and field studies have sometimes found evidence for a positive, rather than a negative, association between corticosteroid secretion and reproductive activity (Wilson and Wingfield, 1992, 1994; Saltzman et al., 1994). Hence, a corticosteroid response could be considered a condition that is neither necessary nor sufficient, although it might be a helpful indicator in many situations.

5 . Presence or absence of external actions (behavior). Some concepts require behavioral changes as a necessary condition (McCarty, 1983). Such a restriction would exclude most cases of “prenatal stress” in mammals and many interesting cases in evolutionary ecology and environmental toxicology where the response involves changes in energy and resource allocation within the body. 6. Kinds of behavior. Some concepts expect animals to show certain kinds of behavior when in a stressful situation, for example, changes in alertness and attention span, decreases in reflex time, and suppression of feeding and sexual behavior (Chrousos et al., 1988). Toates (1995) argued that this is not a very useful criterion, as any behavior needs to be evaluated

410

HERIBERT HOFER A N D MARION L. EAST

in terms of what the animal achieves by it and what would happen if it was somehow prevented from executing it. For instance, both extreme agitation and apathetic withdrawal might be considered responses to stressful situations (Mason, 1991).

7. The results of the organism's response. The most general concept in this respect is that the results of the organism's action, its success in restoring the state of homeostasis or stability prior to a stimulus, does not matter. Thus, a temporary heat wave that the animal successfully escapes from by seeking a cool refuge could be considered a stressor (cell groups A and B in Fig. 2). More restrictive definitions consider situations stressful only if the action of the organism fails to restore homeostasis (cell groups E and F in Fig. 2), or if response mechanisms are being chronically stretched or are failing (Toates, 1995), that is, the cell groups labeled 3 in Fig. 2. If this

4L stimulus

~timulus

stimulus

I

P E

time

time

time

RESTORATION OF HOMEOSTASIS within a

I-1-I I

~

EFFECT ON

I

I I

~~

unchanged modified stretched/ failing

1

'EFFECTON

, NO 1 RESPONSE

modified

i I

short period of time ~

1

;;01 ~

~

/B1I&-m

over a yr;d

never restored; organism remains in a modified state ~-

~

F1

1

,

I

FIG.2. Separating the time course of restoration of homeostasis from fitness consequences and the impact of the environmental stimulus on the organism's response system helps to clarify and classify different concepts of stress. Training of the response system by successive stimuli is indicated by the links between cells C1 and C2 and A1 and A2, and cells D1 and D2 and B1 and B2. Definitions of stress commonly used in psychology and ethology (Toates. 1995) restrict the concept of stress to cells E3/F3, those used in animal welfare studies to the shaded block of cells A3/C3/E3 and B3/D3/F3 (Broom and Johnson, 1991). The evolutionary concept of stress advocated in this chapter includes all cases of cell groups A. C. and E.

BIOLOGICAL CONSERVATION AND STRESS

41 1

happens, then the organism’s state may change in a more permanent fashion, and go through a prepathological state (Moberg, 1985) before processes such as the development of an enlarged adrenal cortex, gastric ulcers, or shrinkage of lymphatic tissues (Selye’s (1973) “stress syndrome”) create pathological states. Some definitions demand evidence of prepathological or pathological states before a situation can be termed stressful (Moberg, 1985). None of the restrictive definitions copes well with training effects (an example are changes in the organismal response in successive challenges, as indicated by arrows in Fig. 2).

8. Fitness consequences. Fraser and Broom (1990) suggested that the Darwinian fitness of an organism may be used to evaluate objectively whether an animal is coping, that is, regaining control of mental and bodily stability. Following their lead, a restrictive definition of stress could be constructed by arguing that stress encompasses only situations that, all else being equal, are likely to or do lead to a reduction in fitness. In this form the concept avoids arguments such as: capturing a wild animal and placing it with conspecifics in captivity might result in an increase in fitness because the threat of predation is removed (e.g., Toates, 1995). In fact, confinement in captivity leads to a decrease in reproductive success. In evolutionary ecology stressors are usually subsumed under the term environmental stress if they exert selection pressure (Hoffmann and Parsons, 1991; Rollo, 1995). 9. Immediately lethal versus sublethal effects. Immediately lethal “stimuli” (e.g., successful predation) may be contrasted with effects that are at least initially sublethal. Conventionally, only initially sublethal effects are considered stressors, but this does not exclude sublethal effects that may eventually cause the death of an organism. However, the distinction between immediately lethal and nonlethal stimuli is not as clear-cut. Suppose there is variation between individuals in the tolerance to the concentration of heavy metals in the soil. A high concentration causes outright death in some but not all individuals in a population. Did only the survivors experience stress but not the nonsurvivors? Environmental stress as defined in evolutionary ecology ignores this restriction. Here stressful effects are defined as exerting selection pressure and thus effects that may be immediately lethal are included (see Hoffmann and Parsons, 1991; Rollo, 1995).

CONCER OF STRESS B. A N EVOLUTIONARY Biological conservation strives to enhance the persistence of populations and ecosystems in their natural setting or through ex situ measures. Population persistence depends on individual attributes such as fecundity, survival, and lifetime reproductive success as well as genetic and spatial population

412

HERIBERT HOFER AND MARION L. EAST

structure and interactions between individuals (e.g., Huffaker, 1958; Ginzburg et a!., 1990; Karr, 1990; Walter, 1990; Renshaw, 1991; Boyce, 1992). In this context, a concept of stress that considers the fitness consequences of the organism’s response to stimuli appears eminently useful. Broom and Johnson (1991) advance a similar argument in the context of animal welfare. Thus, we define stress as an environmental effect that is likely to or does reduce the Darwinian fitness of the organism. The Darwinian fitness of a genotype, a particular assemblage of genes, is the per capita rate of increase of its corresponding phenotype, the traits of the individual that carries the genes (Sibly and Calow, 1983). In practice, fitness can be approximated by the long-term reproductive capabilty and expected reproductive success of individuals. These are linked to their fecundity, survival, and age at first breeding (Sibly and Calow, 1986). Reductions in fitness can therefore be measured as changes in the age at first breeding, survival to adulthood, adult survival between breeding attempts, or fecundity (female offspring) per female breeding attempt (Charlesworth, 1994). For practical conservation activities it makes sense to distinguish sublethal from outright lethal effects (criterion 9) in many cases, so our preference is to usually limit the application of the concept of stress to cases where sublethal effects are observed. Thus, we would exclude deadly road traffic accidents as incidences of stress but include the effect of noise due to road traffic on the breeding performance of woodland birds. By requiring a reduction in fitness we apply a restrictive version of criterion 1, as pleasant stimuli would not be considered stressful (unless they have or are likely to have detrimental fitness consequences). Rollo (1995) used the temporal course and predictability of environmental stimuli to distinguish “stress” from “disturbance” (criterion 3). For our purposes, we do not consider this distinction useful. Thus, states of environment (criterion 2) are included provided they have or are likely to have detrimental fitness consequences. Also, the time course over which restoration of homeostasis may or may not be completed (criterion 4) becomes irrelevant for our concept. Consider Fig. 2: It is possible that a response may achieve restoration of homeostasis quickly, slowly, or may fail, in which case the organism continues in a modified state. In each case fitness may be reduced or may remain unchanged. We include all cases where detrimental fitness consequences are observed or are likely to happen. We are therefore not restricted to those cases where the response mechanism is stretched or actually failing, as others demand (Moberg, 1985; Toates, 1995). Using a reduction in Darwinian fitness to define stress has a number of advantages. The first is universality. It does not require specific mechanisms (criteria 4, 5, and 6) that restrict the concept to certain taxa, or classes of stimuli. The evolutionary concept also recognizes that organisms may em-

BIOLOGICAL CONSERVATION AND STRESS

413

ploy a variety of mechanisms to cope with different stressful stimuli. Using fitness consequences as the decisive factor provides a comparable measure of the impact of stress across a wide variety of organisms, environmental stimuli, and organismal response systems. The second advantage is that it allows us to incorporate cases where the restoration of homeostasis happens quickly but where a reduction of fitness nevertheless occurs (criterion 7, cell group A in Fig. 2). Consider disturbance by boats to great crested grebes, Podiceps cristatus, incubating eggs in vegetation on the edges of lakes. Grebes respond to approaching boats by covering up the eggs, flying away, and returning to the nest after boats have moved on. It is unlikely that coping mechanisms are being stretched or are failing to restore internal homeostasis to the adult. Boating, however, increases exposure of eggs because adults stay away longer after boating disturbances than after other kinds of disturbance and egg losses due to predation are likely if the eggs remain uncovered (Keller, 1989a,b). We think it is useful to call this kind of disturbance stressful. The third advantage is that it emphasizes that for the purpose of biological conservation, studies of the effects of anthropogenic disturbance and environmental change ought to measure fitness consequences. This is important from both a theoretical and a practical point of view. The practical point of view is that without measuring fitness consequences the interpretation of changes in stress response may be erroneous. In her study of grebes, Keller (1989a) showed that the distance between nests and boats at which incubating grebes fled depended on the amount of boating. Grebes exposed to a high amount of boating fled at shorter distances. This may be interpreted as a positive sign that the grebes had successfully habituated to the disturbance by boaters, and indeed grebes with shorter flight distances had a higher reproductive success than those with longer flight distances. However, grebes that had shorter flight distances and fled from approaching boats were also more likely to cover up their nests incompletely than those with longer flight distances (Keller, 1989a), and this may have increased the likelihood of egg losses (Ingold et al., 1992), as even the most successful grebes on sites with intensive recreational activity still had a lower reproductive success than on those without (Keller, 1989a). Note that defining stress in conservation as a condition that reduces or is likely to reduce fitness does not affect the study of potential stressors, particularly anthropogenic ones, nor does it invalidate research that investigates the organism’s stress response without looking at fitness consequences. It merely emphasizes that for conservation purposes such studies remain incomplete unless it can be convincingly argued that the kind and intensity of the organism’s stress response predict detrimental fitness consequences. Nor does such a concept prescribe whether the effects of a potential stressor

414

HERIBERT HOFER AND MARION L. EAST

should be considered acceptable or desirable if no detrimental fitness consequences are found. For instance, if plumage fouling of a species of penguin by low amounts of crude oil has no detrimental fitness consequences, then crude oil pollution of the investigated amount, consistency, and application pattern would not be considered a stressor for this species. However, this does not make oil pollution acceptable. The same application pattern of crude oil pollution may have detrimental fitness consequences for other species. For example, Fowler et al. (1995) demonstrated that even low levels of oil fouling in Magellanic penguins, Spheniscus magellanicus, significantly elevated plasma corticosterone concentrations in females and reduced their fitness. Also, other organisms that were not investigated may suffer detrimental fitness consequences from pollution, habitat quality may decline, or the composition of ecological communities may change. Note that all concepts assume that there is an anticipated o r actual deviation from homeostasis; the implication is that this can be recognized easily. In practice, however, this may be a point of contention. For instance, body mass losses of birds during the breeding season have been considered an evolutionary adaptation designed to reduce increased energetic costs of flight of foraging parents during the breeding season, and a stressful consequence of increased energetic demands by the young that leads to parents obtaining insufficient food so that they use up their own body reserves (Moreno, 1989). The controversy hinges on the notion whether the loss of body mass is an unexpected deviation from homeostasis or whether it is an adaptive change in anticipation of the demands of the breeding season. Theoretical analyses suggest that adaptive body mass changes are expected as a consequence of trade-offs between costs and benefits of fat storage (Lima, 1986; Houston and McNamara, 1993). The evidence mostly favors the adaptationist view (Korpimaki, 1990; Merkle and Barclay, 1996), although there have been cases of mass losses consistent only with a stress explanation (Korpimaki, 1990). We return to the relationship between stress and evolutionary adaptations in Section IV.

C. A FRAMEWORK FOR STUDYING STRESS I N BIOLOGICAL CONSERVATION We now introduce a framework for studying stress in biological conservation. It concentrates on the natural history of stress and emphasizes the evolutionary importance of stress (Fig. 3). This permits us to focus on aspects that have been neglected in conceptual debates about stress and that direct us to some interesting questions for biological conservation. A natural history of stress explores the link between the variation in environmental stimuli, the evaluation of these stimuli by organisms, and the range of responses to and consequences of stressful stimuli. Current

BIOLOGICAL CONSERVATION AND STRESS

415

knowledge of potential stressors (Fig. 3) is probably incomplete, but each listed factor has been hypothesized or shown to be involved in stressful situations. Natural stressors include physical and chemical characteristics of the environment (radiation, temperature, pH, water/drought, salinity, minerals, environmental catastrophes such as storms), and factors emanating from within a species (population density, social instability, breeding) and other species (resource availability, pathogens, predators). Anthropogenic factors include global warming, pollution, noise (boats, powerful acoustic communication and listening devices in the marine environment, road and air traffic), disturbance by visitors and leisure activities (tourist game viewing, hiking, paragliding, mountain biking, skiing, canoeing), seismic exploration, development (often for tourist game viewing or leisure activities), and sports hunting. Conservation-activity-related stressors include intervention and handling, confinement and captivity, transportation, and translocation. If stress is an evolutionary force, then an organism’s evaluation of an environmental stimulus and its response may be an evolved trait. Variation in evaluation and response may thus not be random but adaptive, and hence predictable, in at least those cases where a particular type of stimulus occurred sufficiently frequently during the evolutionary history of a population. Section IV reviews evidence that variation in stress response is predictable and an evolved trait, and summarizes current knowledge about the sources of variation in stress response relevant to biological conservation and their consequences. The natural history of anthropogenic stress is considered in Sections V (general anthropogenic stressors) and VI (conservation-activity-related stressors). For many species we know little about the kinds of stressors they face, what factors influence the organism’s evaluation of the stimulus, what kinds of response are available, and what consequences result from the response. However, even if our knowledge of the natural history of the stress response of a species to natural stimuli were well known, could we predict how they would respond to anthropogenic stimuli? This is an important question that has not been reviewed before (Sections V and VII).

D. QUESTIONS ABOUT STRESS IN THE CONTEXT OF BIOLOGICAL CONSERVATION We introduce our discussion of questions about stress in the context of biological conservation by a case study on the development of tourist facilities and its impact on the mountain pygmy-possum, Burramys parvus. This case study is interesting because it illustrates many questions that can be asked about stress-related issues and because a scientific study

416

HERIBERT HOFER AND MARION L. EAST

(POTENTIAL) STRESSOR

MODULATING FACTORS

STRESS RESPONSE

CONSEQUENCI

transportation confinement / captivity handling / intervention

behavioral response (habitat choice, movements, social inter-

sports hunting ieveloprnent noise visitor disturbance

coids, catecholarnines, androgens)

immunocompetence

pollution

energy requirements

global foraging efficiency radiation

survival

temperature

breeding success

PH water I drought

breeding suppression

physiological

I

I

salinity catastrophes

allocation)

resource availability population density breeding social instability pathogens predators

I

I developmental response

BIOLOGICAL CONSERVATION A N D STRESS

417

accompanied the development of a ski resort from the design stage onward so that the setup (one developed and disturbed versus one undeveloped and undisturbed habitat) approximated a scientific experiment. 1. Case Study I : The Mountain Pygmy-Possum and

Tourist Development The mountain pygmy-possum is one of the most threatened mammalian species. Its distribution is restricted to less than 10 km2 total world habitat. Its world population comprises roughly 2600 individuals in several disjointed populations in the Australian Alps. Mountain pygmy-possums are the longest lived small terrestrial mammal, they are female dominated, feed on plants and insects, and hibernate during the winter (Mansergh and Broome, 1994). The development of a ski resort, accompanied by the building of roads and the construction of ski slopes, had far-reaching consequences for the pygmy-possum populations. Development involved clearing vegetation, habitat destruction, and habitat fragmentation. Habitat destruction and fragmentation in a key high-quality site reduced the proportion of adult females and increased the proportion of adult males in the period prior to hibernation. Female survival during hibernation was only half of the survival in the undisturbed area and female fecundity was reduced (Mansergh and Broome, 1994). Why? Hibernating animals rely on fat rather than food caches to survive the winter (Geiser and Broome, 1991). Males and juveniles are usually ejected by females from high-quality sites and this ensures that the females accumulate sufficient fat in autumn. Prehibernation fattening occurs at a time of reduced activity and energy expenditure (Kortner and Geiser, 1995). It is thus possible that overcrowding the habitat with juveniles and males disturbed the females and raised their activity levels so that

FIG. 3. A framework for the study of stress in biological conservation that emphasizes the role of stress as an evoultionary force. Stressors are environmental stimuli that include both natural (bottom box, left side of figure) and anthropogenic stimuli (top two boxes, left side of figure), separated into stimuli arising from conservation activities (top box) and other anthropogenic stimuli. If stress is an evolutionary force. then we would expect that the evaluation of a stressor by the organism as reflected in the organism’s subsequent response is shaped by natural selection. Thus, we expect the evaluation to vary, and, in the case of natural environmental stimuli at least, to vary in an adaptive way. The suite of factors known to modify response systems (modulating factors) are listed in the shaded block. The subsequent response of the organism may activate a behavioral. hormonal, immune, developmental, and/ or physiological response. Stimulus, evaluation. and response may modify the organism’s energy requirement and foraging efficiency, and its survival and reproductive activity and success.

418

HERIBERT HOFER AND MARION L. EAST

accumulation of fat reserves was insufficient (Mansergh and Broome, 1994). Snow-grooming of ski slopes by graders, ski-mobiles, and skiers may also have contributed because snow-grooming compacts snow and creates noise and vibrations. This may affect microclimate and incite arousal from torpor more often than usual. Excessive arousal would deplete energy reserves and hence diminish survival chances. Several conservation activities were considered or implemented. The creation of a habitat corridor (a tunnel of rock boulders under the road) reconnected key habitats on both sides of the main alpine highway, and within a year the population structure in the key areas for females reverted to the pattern known from undisturbed sites (Mansergh and Scotts, 1989). Habitat restoration, however, is considered to be of limited use, as growth of a key food plant, the mountain plum-pine, Podocarpus lawrencei, is slow (.25 mm increase in diameter per year). The chance of successful restoration is further reduced by siltation from areas where plum-pines were removed by ski slope development, inadvertently lost due to grazing earlier this century, or lost by fire (Mansergh and Broome, 1994). Ex situ conservation efforts at first faced considerable difficulties. Captive-bred animals exposed to autumn and wintering conditions in the laboratory approximating conditions in the wild did not put on weight and did not hibernate at first (Geiser e f a[., 1990), but did so under modified conditions (Broome and Geiser, 1995; Kortner and Geiser, 1995). Captive breeding and release efforts may still require fine-tuning before they can be considered an “insurance” against extinction (Mansergh and Broome, 1994). Captive breeding might become valuable if the predicted winter temperature rise caused by global warming becomes reality. Such a rise may increase winter mortality because the energetic costs of hibernation critically depend on ambient winter temperature (Geiser and Broome, 1993). Increased summer temperatures may also limit dispersal. Animals quickly become hyperthermic at ambient temperatures above 30°C (Fleming, 1985). In order to disperse during summer, animals would have to cross valleys and hence descend to lower altitudes where temperatures are higher. Models show that an increase of average ambient temperature by 1°C might lead to species extinction because no area would retain the climate pattern to which the mountain pygmy-possum is assumed to be currently adapted (Mansergh and Broome, 1994; Brereton et af., 1995). In addition, climatic change might also affect the migratory behavior of the Bogong moth, Agrotis infusa, a major food source of the pygmy-possum during the breeding season. This possibility has not yet been investigated.

2. Questions about Stress This case study illustrates that answers to basic questions about the natural history of stress responses to natural and anthropogenic stressors

BIOLOGICAL CONSERVATION AND STRESS

419

are crucial if conservation activities are to be effective. What are the fitness consequences of a stressor and by what mechanism does the stressor reduce fitness? Do animals react to stressors with developmental, behavioral, energetic, hormonal, or immunological adjustments, and are these aspects of the stress response linked? At what frequency of occurrence and/or magnitude does a stressor cause a reduction in fitness? Do characteristics of a potential stressor (frequency of occurrence, magnitude) reliably predict the magnitude of fitness consequences and the kind of stress response an animal will exhibit? Is the pattern of the stress response a reliable predictor of fitness consequences? Under what conditions do potential stressors cause pathological states? The case study also illustrates that answers to questions concerning stress may determine the success or failure of conservation activities. To answer these questions a comparison of managed (experimental) and unmanaged (control) segments of the population may be required. The case study shows that application of experimental principles are essential if the impact of management is to be understood. Theories of conservation actions emphasize the preservation of genetic diversity (SoulC, 1987). Genetic studies have not been undertaken on the mountain pygmy-possum, but the pattern of several small populations with potentially restricted gene flow between them (Mansergh and Broome, 1994) suggests that it would be valuable to look at genetic variation between populations. Conservation actions often aim for maximizing “general” genetic diversity, as measured by mean population heterozygosity across many alleles, or the proportion of alleles that are heterozygous (see Nevo et al., 1984). How compatible is such a goal with the goal that population persistence may be improved by maximizing stress tolerance or stress resistance? If past selection favored individuals that are stress resistant or stress tolerant, and if such selection contributed to genetic variation between populations, then the two goals may not be compatible. Futuyma (1983) argued that if biotic stressors (pathogens, parasites, competitors, predators) are an important threat to a population, then it may be important to preserve rare alleles, rather than overall genetic diversity. This may be important for conservation programs that involve the translocation of populations. Translocated populations and refuge populations may show reduced genetic diversity because alleles rare in the parental population got lost (Stockwell et al., 1996). Hoffmann and Parsons (1991) and Parsons (1989a, 1995) argued that small marginal populations with restricted gene flow to large central populations are more likely to experience strong selection for stress resistance or stress tolerance. They expect marginal populations to contain increased genetic diversity as well as a preponderance of stress-tolerant or stress-resistant individuals. Their line of argument predicts that, if management actions increased the genetic flow between central and marginal popu-

420

HERIBERT HOFER A N D M A R I O N L. EAST

lations, for example, by creating habitat corridors, then population persistence may be reduced. For successful conservation efforts, detailed knowledge about population structure, gene flow between populations, and genetic variation for stress tolerance between marginal and central populations might prove to be essential. Are individuals with higher stress tolerance at a fitness disadvantage compared with normal individuals if stressful conditions are rare (e.g., Krebs and Loeschke, 1994)? Because inbreeding depression may be more severe under conditions of environmental stress (Miller, 1994), reduced genetic variability through inbreeding may be a much greater problem for recently reintroduced free-ranging populations than for populations in a relatively benign captive environment. Does captive housing of animals in temperature-regulated facilities over several generations reduce the capacity to adapt to extreme temperatures (Kohane and Parsons, 1988)? This type of knowledge may help guide the design of captive facilities, the selection of individuals for breeding and reintroduction programs, or the matching of reintroduction sites with individuals best able to cope with the prevailing forms of stress. These sets of questions summarize an outline of stress research in biological conservation. This research focuses on the natural history of the stress response and links developmental, behavioral, energetic, endocrinological, and immunological aspects of the stress response with fitness consequences. It identifies aspects of the natural history of a species that predict the consequences of particular stressors. It assesses genetic variation in stress tolerance or stress resistance between populations. It emphasizes minimalinvasive or noninvasive monitoring and research techniques that minimize stress caused by conservation activities. Our survey indicates that there are many promising beginnings but that answers to many questions are not yet available.

A CONSERVATION STUDY T o MEASURE STRESS 111. DESIGNING AND ITS IMPACT

Careful design of research studies and conservation activities can increase the value of the results and foster more effective conservation. The value of conservation management increases substantially if it conforms to standard scientific protocol, as this improves the assessment of its impact and efficacy. Although the recommendations in this section are not new, the violation of scientific protocols in conservation-oriented studies suggests that a brief summary of some important issues would be useful. Elaborations can be found in Cochran (1977), Green (1979), Mead (1991), or Manly (1992).

BIOLOGICAL CONSERVATION AND STRESS

A.

421

MEASURING SEVERAL COMPONENTS OF THE STRESS RESPONSE Is DESIRABLE

Stressed states and stress responses may be measured in several ways. A comprehensive assessment of the impact of stress should include more than one component of the response, as it is increasingly accepted that there is activation of more than one physiological system (Fig. 3). A virus infection, for instance, may trigger immunosuppression as well as neurochemical and endocrine responses (Dunn et al., 1987, 1989), although the physiological relevance and precise mechanisms that link different response systems after an infection are only beginning to be explored (Olsen et al., 1992; Stefan0 and Smith, 1996). Hormonal responses have also been shown to interact with foraging behavior, energy storage, and release systems (Astheimer et al., 1992; Rogers et al., 1993). If individuals vary in the extent to which hormonal, resource allocation, immunological, and behavioral systems are triggered (e.g., Hurst et al., 1996), and if these systems follow different time courses (e.g., Sachser and Lick, 1989), then measuring only part of the response may provide incomplete information about the magnitude and consequences of the full response. Thus, it is valuable to monitor hormonal, behavioral, and immunological consequences of stress and consider their links (Coe e? al., 1994; Toates, 1995; Hurst et al., 1996; Sapolsky, 1997). DETERMINES WHICHEFFECTS CAN B. THETIMESCALEOF MEASUREMENT BE STUDIED What time scales are required to detect fitness consequences and the full range of the stress response? What time scales are required to assess the efficacy of conservation efforts? Answers to such questions are vital for the experimental design of a study. Short time scales in terms of minutes, hours, or days usually cover most components of the immediate stress response. The immediate hormonal consequences of ACTH stimulation can be monitored over hours (Sapolsky, 1982) and even severe stresses may cause changes in behavior that persist for only a few days (Astheimer et al., 1995). However, some consequences of the immediate response, such as a disturbance of ultradian and circadian rhythms, may take more than one week before they are rectified, long after a challenge has terminated (Harper et al., 1996). We know little about the time course of energy reallocation processes. Short-term monitoring is unlikely to identify pathological consequences. The development of pathological states may require days (von Holst, 1986; Sachser and Lick, 1989) or weeks (dasyurid marsupials: Lee and Cockburn,

422

HERIBERT HOFER AND MARION L. EAST

1985; salmonids: Hare and Robertson, 1959; Robertson and Wexler, 1957, 1960). An exception to this rule is the phenomenon of capture myopathy, a debilitating and often fatal syndrome associated with extreme muscular exertion during pursuit, restraint, or handling (see Section V1,C). Death due to capture myopathy may occur within hours (dugongs, Diigong dugon: Anderson, 1981; African lion, Panthera Leo: Joubert and Stander, 1990) or days (white-tailed deer, Odocoileus virginianus: Beringer et al., 1996). The relationship between short-term and long-term measures of fitness consequences of stress is complex. Short-term monitoring may sometimes suffice to detect significant effects of handling or intervention in field studies (see later discussion). However, long-term monitoring of the fate of individuals subjected to intervention in field studies may often be useful because interventions can result in delayed mortality, as in the case of capture myopathy (Beringer et al., 1996), or other forms of detrimental fitness consequences (Putman, 1995). In such cases, short-term measures would conclude that there are no detrimental fitness consequences when in fact they do occur. Short-term assessment of fitness consequences may also misjudge long-term effects on population persistence. For instance, pollution may produce pronounced short-term effects on behavior of adults that may be associated with a reduction of reproductive success during the current breeding season. However, some studies found no significant longterm effects on population persistence by the specific stress and application method (insecticides and several species of passerines: Busby et al., 1990; Millikin and Smith, 1990; crude oil and Leach’s storm petrel, Oceanodroma leucorhoa: Butler et al., 1988). We emphasize that long-term monitoring is particularly crucial for the assessment of the efficacy of conservation activities. Such long-term monitoring is not standard practice. For instance, one set of guidelines that eloquently summarizes protocols for vaccinating wildlife (Hall and Harwood, 1990) does not mention long-term monitoring of vaccinated individuals to judge vaccination success. Long-term monitoring is always advisable, as it is unwise to assume that conservation actions cannot make things worse (Thorne and Williams, 1988; Hall and Hanvood, 1990). Short-term measures of significant seroconversion of a high proportion of vaccinated individuals may not be a reliable guide to the long-term persistence of protective levels of antibodies (rabies and domestic dogs: Sage et al., 1993). Vaccination may also have detrimental effects on individual life expectancy (canine distemper and black-footed ferret, Mustela nigripes: Carpenter et al., 1976).These effects of vaccines may become apparent only after monitoring individuals over several months, and the effects of other forms of interventions may be apparent only after years of monitoring (rabies and African wild dogs, Lycaon pictus: Burrows et al., 1994, 1995).

BIOLOGICAL CONSERVATION AND STRESS

423

Temporal characteristics of the occurrence of stressors have not been well studied but are of major importance to evolutionary models that relate spatial and temporal characteristics of habitats to patterns of selection pressure (Southwood, 1988; Wingfield et al., 1992a; Parsons, 1994; Rollo, 1995). Such temporal characteristics include the degree to which environmental factors remain constant or change from season to season, or how frequently these factors change in relation to the generation time or lifetime of individuals. Unpredictable, severe events are considered to be rare but may have pronounced effects on the pattern of hormonal and behavioral response (Wingfield, 1988; Smith et al., 1994; Astheimer et al., 1995).

C. MEASURING FITNESS CONSEQUENCES Is IMPORTANT Why is it important to measure fitness consequences? Consider a study that demonstrates that disturbance by people approaching an incubating bird increases the animal’s heart rate by a certain percentage compared to some baseline value. Does this result indicate a conservation problem? Internal changes following the occurrence of a potential stressor cannot, for the purpose of biological conservation, be properly interpreted without measuring fitness consequences, or at least establishing a calibration curve that relates the stress response to fitness consequences. As argued above, population persistence depends on components of Darwinian fitness. Fitness consequences of stressors that have been measured within the context of intervention include failure to breed (Sapolsky, 1985), abandonment of offspring (Colwell et al., 1988), a decline in the number of surviving independent young (Rodway et al., 1996), an increase in mortality (Cotter and Gratto, 1995), and a decline in longevity (Burrows et al., 1994). D. GOODSTUDIES AND MANAGEMENT ACTIVITIES PAYATTENTION TO PRINCIPLES OF EXPERIMENTAL DESIGN Good studies select subjects carefully, chose appropriate controls, avoid confounding factors, optimize the number of subjects required, avoid pseudoreplication, and design the study in such a way that the power of statistical tests is sufficiently high to recognize effects if there are in fact any (Green, 1979; Hurlbert, 1984; Machlis et al., 1985; Manly, 1992). 1. Selecting Individuals

Most statistical tests, including those commonly used in ecology, animal behavior, or physiology, require a random sample of individuals. This requirement is sometimes not easy to fulfill and often ignored. An example will illustrate this. A study aims to assess the influence of vessel-based

424

HERIBERT HOFER AND MARION L. EAST

nearshore whale watching (e.g., Stone et al., 1992) on the behavior and fitness of migrating whales. Principles of experimental design specify that a random sample of individuals should be constructed by randomly selecting individuals prior to migration at, say, their overwintering site and assigning them to either the control group (no whale watching) or the treatment group (whale watching). The effect of whale watching could then be assessed by applying a statistical test to selected measurements. It is unlikely that all or even most whales assigned to one group will actually end up in that group because the researcher is unlikely to influence the route of migration of individual whales, so the actual composition of control and treatment groups may be very different from the one decided by the observer prior to migration. The measurements could still be evaluated as planned, but now there is a caveat. Imagine that animals in poor condition stay closer to the shore than animals in good condition and are therefore more likely to experience whale watching. Any differences between treatment and control groups now may be either due to body condition, or due to whale watching, or due to the combined effect of both. In order to avoid such problems, periods of whale watching might be alternated with periods without, or the experimental manipulation could consist of “moving” whale watching vessels rather than whales. A common problem with observational studies or “natural” experiments is that individuals within experimental and control groups have often entered these groups by their own choice because of some unmeasured trait. Such studies provide correlational evidence only if individuals are not randomly assigned to experimental and control group. This does not detract from the value of observational studies that often summarize empirical evidence difficult to obtain otherwise. It just means that we ought to be careful about how far the results of such studies can be interpreted. For instance, in an observational study of the great skua, Catharacta s k u , Thompson et al. (1991) found no relationship between mercury concentrations of individuals and their breeding performance or survival. They suggested that mercury concentrations were not linked to fitness because there was substantial individual variation in choosing feeding areas used before the breeding season. Can we conclude from this study that mercury pollution does not have detrimental fitness consequences in great skuas? No, because animals in the group with high concentrations of mercury may have operated a different foraging tactic or used some other means to reduce reproductive effort compared with animals in the group with low concentrations of mercury, and the fitness costs of mercury pollution may become apparent only at high levels of reproductive effort. Another example is that if, understandably, researchers tend to radiocollar those animals they consider particularly healthy in order to minimize

BIOLOGICAL CONSERVATION AND STRESS

425

potentially detrimental effects of collaring, then a comparison of collared and uncollared individuals becomes biased and cannot be done using standard statistical tests because of the requirement of random samples. For instance, Creel (1996) and Creel et al. (1997) explicitly selected those individual African wild dogs for radio-collaring that they considered were least likely to suffer from any potential effects of collaring, and then went on to compare this nonrandom sample with unhandled African wild dogs using standard statistical procedures to argue that collaring did not cause “stress.” Pilot studies may help to decide whether nonrandom sampling is a problem. If the observer cannot be sure that individuals in control and treatment groups are random samples from a larger population, then extrapolations of the results of a study to other segments of the population may not be warranted.

2. Controls Are Important Having control groups, and selecting them wisely, is important. In the absence of a control, it is often impossible to assess the effect of a stressor on the success or otherwise of a conservation action. It is important to observe control and experimental animals prior to the conservation action as well as after the action (Green, 1979). Without knowing whether differences between control and experimental groups were already present prior to the conservation action it is impossible to know whether differences between experimental and control groups after the action are due to it or some other factor. With a wisely chosen set of control animals or conditions, the only difference between experimental and control groups is the stress factor of interest, and thus any observed difference can be attributed to the experimental factor. This is important for conservation activities that are perceived to be beneficial, and thus applied to entire populations. An example is the use of wildlife vaccinations where there is a tendency to think that saving the entire population from the perceived threat of a viral or bacterial infection is more important than considering such an intervention as an experiment whose efficacy ought to be properly assessed. When the Serengeti population of African wild dogs was vaccinated against rabies, there were no unvaccinated and monitored control groups (Heinsohn, 1992). This is part of the reason why it is unlikely that the cause of the extinction of the study population, which occurred shortly after the vaccination, will be conclusively identified (Burrows et al., 1994).

3. Careful Experiments Attempt to Avoid Confounding Factors To obtain unambiguous results, clearly only one factor should be changed between experimental and control groups. Changing two or more factors simultaneously in a conventional experimental design will ensure that any

426

HERIBERT HOFER AND MARION L. EAST

observed differences between experimental and control groups cannot conclusively be assigned to one particular factor. With special designs it is possible to study the effects of several factors simultaneously (Cochran, 1977; Green, 1979; Mead, 1991; Manly, 1992).

4. Optimizing the Number of Individuals Required Optimizing the number of individuals required for a study is of both practical and ethical importance, particularly if the experiments are contentious, painful, or stressful. It is often possible to reduce the number of individuals needed if the same individual is used both in control and in experimental groups, where experimental conditions permit this. Such experimental data can be analyzed with paired tests (e.g., Wilcoxon signedrank test), which are particularly powerful. More specialist test statistics designed for sequential samples can also be used (Gottman and Roy, 1990). Still (1982) and McConway (1992) discuss these and other recommendations. A second problem is that sample sizes may be too small if the power of a test is small, or if pseudoreplication occurs (see the next two sections). The power of a test can be calculated using procedures described by Cohen (1988). It is usually better to have less detailed data on more individuals than to have very detailed data on only a few individuals. 5.

The Power of a Test Decides Whether a Study Is Likely to Find Significant Effects

The potential impact of anthropogenic factors is usually tested by formulating a null and an alternative hypothesis, conducting observations or experiments, and then applying a statistical test (e.g., de Swart et al., 1995a; Ross et al., 1995). A significant effect emerges if the null hypothesis is rejected. Such a test is associated with two types of error: rejecting the null hypothesis when in fact it is true (Type I error with probability a ) , and accepting the null hypothesis when it is in fact false (Type I1 error with probability p). The power of a test is 1 - p, that is, the probability of rejecting the null hypothesis when it is false, and can be calculated by following the procedures outlined by Cohen (1988). If the power is high, then the test is highly likely to find a significant difference between treatment and control if a difference really exists. If the power is small, however, then this is unlikely. In randomized trials, the power should be approximately .8-.9 (e.g., Parmar and Machin, 1995). Ignoring the power of a test may lead to false conclusions: A lack of a significant difference may be due to a lack of power to distinguish between treatment and control, not because there is no real difference (Thomas and Juanes, 1996). This is of practical importance. For instance, if the purpose is to show that handling does not stress the animal, then a statistical

BIOLOGICAL CONSERVATION AND STRESS

427

test with low power is unlikely to find an impact of handling even if there was one. Thus, reporting the power of tests used to examine intervention effects should become routine. However, such reports are an exception and when values of power are reported they are usually low (.l-S), so there is room for improvement (e.g., Houston and Greenwood, 1993; Gammonley and Kelley, 1995; Thirgood et al., 1995). Without information on power, previously published conclusions that handling has no effect should be considered tentative. An example is the debate about the potential impact of researcher presence on the likelihood of predation on cheetah, Acinonyx jubatus, cubs (Laurenson and Caro, 1994; O’Brien, 1994; May, 1995). What does the power of a test depend on? Apart from the choice of test statistic it depends on the sample size, the magnitude of the expected effect, and the sample variance. Small sample sizes imply low power, large sample sizes increase power. White and Garrott (1990) reanalyzed data from a previous publication (Garrott e f al., 1985) on the effect of radio-collaring of mule deer, Odocoileus hemionus, fawns on the chance of being preyed upon by coyotes, Canis latrans. Garrott et al. (1985) concluded that 45 radio-collared fawns did not suffer ( p = .67) increased predation mortality compared to 46 ear-tagged controls. White and Garrott (1990) then showed that if collars were supposed to have doubled predation mortality, the minimum sample size to detect a significant (at p = .05) difference between collared animals and controls would have been 48 animals in each group, so their sample size was insufficient to show this. To demonstrate that an increase of merely 20% in predation mortality due to collaring was significant would require a sample of 408 animals in each group! Because the calculation of adequate minimum sample sizes requires an idea of the magnitude of “treatment” effects, pilot studies may be useful to assess the magnitude of effects likely to be encountered (Underwood and Kennelly, 1990).

6. Pseudoreplication Inflates the Number of Independent Sampling Events Pseudoreplication occurs when individuals contribute more than one data point to a data set and each data point is treated as if it were independent (Hurlbert, 1984; Machlis et al., 1985). For instance, an experiment with six fish provides six independent data points, one for each individual, even if each individual was subjected to ten trials (for a discussion of a recent case, see Lamprecht and Hofer, 1994; Lombardi and Hurlbert, 1996). The frequent mistake is to assume that the sample size n required in standard statistical tests is the number of experiments (60) rather than the number of independent sampling events, the individuals (6). Following such a mistake

428

HERIBERT HOFER AND MARION L. EAST

significant results are more likely because p-values depend on a. If the number of data points varies between individuals, a second problem arises: The number of data points per individual operates as a weighting factor. In extreme cases, where one individual contributes a substantial proportion of data points, the results of the statistical test depend on the behavior of that individual. There are several solutions to the issue of pseudoreplication. If there is more than one data point per individual, data may be averaged, yielding one composite value per individual, or they may be analyzed with special tests for repeated sampling (Hurlbert, 1984; Machlis et al., 1985). Because of the problem of pseudoreplication it is generally better to collect less detailed data on more individuals than detailed data on very few. This also increases the power of a test (see previous discussion).

HISTORY OF STRESS IV. THENATURAL In this section we focus our discussion of the natural history of stress on issues that are relevant to biological conservation. Our discussion will center on factors that modulate an individual’s stress response. Figure 3 emphasizes that there is a wide range of such factors, an aspect neglected in many conceptual discussions of stress but of immediate and wide-ranging practical importance for biological conservation. Such a range of modulating factors may be expected because different populations and species have different evolutionary histories, and optimal levels of environmental conditions, where individuals experience no stress or a minimum of stress (Hoffmann and Parsons, 1991), may vary between populations. Different species may therefore respond in different ways to the same stimulus, there may be variation in stress response between populations within a species, and there are a host of attributes that cause variation in stress response between individuals within a population (Fig. 3). We review this variation in three contexts before we outline elements of a theory for such variation and consider whether the stress response can have detrimental fitness consequences and yet be considered an adaptive trait. We then review issues surrounding the impact of stress on organisms relevant to biological conservation and evaluate measures that may help to identify stressed states of organisms and populations and thereby direct conservation work. For a comprehensive discussion of the organismal stress response we ask the reader to consult the chapters by Apanius, von Holst, Lima, Mlbller, and Parsons in this volume. Before we look at the factors that modulate the organismal stress response we briefly introduce the “cellular stress response” as a reminder of the evolutionary continuum that links stress

BIOLOGICAL CONSERVATION AND STRESS

429

responses of all organisms. We should point out that for many of the studies we discuss in Section IV it is not clear whether the environmental stimuli examined had detrimental fitness consequences and thus would be considered stressors according to our definition. A. THECELLULAR STRESSRESPONSE

The “cellular stress response” is the activation of a class of proteins known as “heat-shock’’ proteins (HSP) when a cell is submitted to a transient rise in temperature, low pH, low oxygen level, or other physical or chemical treatments (Morimoto et al., 1990; Nover, 1991; van Eden and Young, 1996). These treatments lead to the accumulation of denatured and/or aggregated proteins inside the cell. HSPs repair the damage resulting from these conditions in a variety of ways (Watson, 1990; Burel et al., 1992; Parsell et al., 1994). Hence, HSPs are often called stress proteins and an increase in their production is described as a “cellular stress response.” HSPs are present in all cells of all prokaryotic and eukaryotic organisms and the amino acid sequences of these proteins are highly conserved in all groups of organisms (Morimoto et al., 1990; Nover, 1991; van Eden and Young, 1996). These characteristics indicate an evolutionary continuity of selection pressures exerted by stress since the earliest life forms and suggest that the cellular response may provide the molecular basis of universal components of the organismal stress response. Links between molecular processes on a cellular level and the organismal response are increasingly recognized (van Eden and Young, 1996). For instance, the capacity to cope with infections can be linked to the state of repair of cells of the immune system (Macario, 1995). Thus, for an evolutionary understanding of stress, the cellular stress response is of considerable interest. STRESSRESPONSE CANBE HIGHLY VARIABLE B. THEORGANISMAL It is important to appreciate that individuals vary in their response to the same stressor. This variation may be random, but more often it can be predicted from modulating factors (Fig. 3). The following sections review empirical evidence on factors that modulate the stress response and discuss hypotheses that predict variation in response in three contexts: the ecology of reproduction, metabolic rate and energetics, and social factors. It concludes with an outline of elements that a theory would require to predict individual variation in stress response relevant for biological conservation. Predictions about intraspecific and interspecific variation in the stress response would be valuable for anticipating the consequences of conservationrelated activities.

430

H E R B E R T HOFER AND MARION L. EAST

Experiments that involve situations in which the stressor has been standardized have been conducted in evolutionary studies with heat or desiccation as a stressor (Hoffmann and Parsons, 1991), and in field endocrinological studies with capture of immobilization as a stressor (Wingfield et al., 1996). This involves first immobilizing or capturing and handling an individual and then measuring a hormonal response such as corticosteroid production by blood sampling at regular intervals over a standard time period of one or a few hours (Wingfield et al., 1996). An extension of this protocol involves the initial injection of a standardized amount of ACTH (Sapolsky, 1982). In the following sections we concentrate on studies with this approach.

1. Factors Influencing the Stress Response Differences between species in their corticosteroid stress response have been documented in birds (e.g., Wingfield et al., 1992b) and mammals (e.g., Widmaier et al., 1994). In some cases, species differences were experimentally shown to be a consequence of the differences in social organization (new-world monkeys: Mendoza and Mason, 1986), or were presumed to be related to differences in life-history parameters (Wingfield et al., 1995b; Cockburn, 1997). Within species, there may be significant differences in stress responses (of any kind) between populations (Krebs and Loeschcke, 1994; Wingfield et al., 1994a, 1995a; Dunlap and Wingfield, 1995). Evolutionary studies have demonstrated significant genetic between-population variation in stress response and stress tolerance to physical factors such as heat or desiccation, for example, in Drosophila (see Hoffmann and Parsons, 1991; Krebs and Loeschcke, 1994) and intertidal organisms (Etter, 1988). The successful breeding of strains of laboratory animals for a specific stress response demonstrates that the stress response may in part be genetically determined (Satterlee et al., 1993; Backstrom and Kauffman, 1995; Lemaire and Mormkde, 1995). Field studies of between-population differences in the adrenocortical response have only recently begun (Dunlap and Wingfield, 1995; Wingfield et al., 1994a, 1995a). None of these studies provides direct evidence that between-population variation may be genetic. Their results are nevertheless instructive. For instance, in the western fence lizard, Sceloporus occidentalis, the adrenocortical response to acute stress was higher in populations at the margin of the species range than in central populations. This study controlled for possible seasonal changes, differences in individual physiological condition, seasonal changes in physiological condition, and population differences in physiological condition. Population differences therefore may have been generated by genetic variation, or possibly were a conse-

BIOLOGICAL CONSERVATION AND STRESS

431

quence of differential ontogenetic sensitization, but this has not been investigated (Dunlap and Wingfield, 1995). Significant differences between individuals of a population may be due to sex (Matt et al., 1983; Wingfield, 1985; Astheimer et al., 1994; Logan and Wingfield, 1995) and season (Bradley et al., 1980; Gustafson and Belt, 1981; Wingfield et al., 1992b; Astheimer et al., 1994; O’Reilly and Wingfield, 1995). Individual attributes that may change within the lifetime of an individual and significantly affect the stress response include age (Kock et al., 1995), body condition (Smith et al., 1994; Wingfield et al., 1994a,b), molt (Astheimer et al., 1995), lactation status in females (Higuchi et al., 1989), stage of territoriality in males (Hannon and Wingfield, 1990), whether females have arrived on a male territory or not (Beletsky et al., 1990), social status (Sapolsky, 1982,1997; Sachser, 1987; Sapolsky and Ray, 1989), social experience (Sapolsky and Ray, 1989; Sachser and Lick, 1989,1991), social support (Sachser and Beer, 1995), personality (Sapolsky, 1997), and individual differences in agonistic behavior (Line et al., 1996). 2.

The Adrenocortical Response and the Ecology of Reproduction

The adrenocortical response (activation and tuning of the HPA axis) is an important component of the hormonal stress response because it has at least two potential fitness effects. First, glucocorticoids mobilize energy and tissue nitrogen as a source of protein through gluconeogenesis, which may improve fitness because it can improve reproductive success (Lee and Cockburn, 1985). Second, it may also be responsible for immunosuppression and the creation of pathological states (Lee and McDonald, 1985;Wingfield, 1988). If the adrenocortical response is an evolved adaptation, then we would expect that different species vary in the way in which this response is fine-tuned. Figure 4 illustrates four options for tuning the adrenocortical response known to vary between species or populations. The anterior pituitary may or may not change its sensitivity to environmental stimuli (option 1). Adrenal activity and gonadal activity may or may not affect each other in different species (option 2). Gonadal activity may increase or decrease the rate of production of corticosteroid binding proteins (CBG), which determines the concentration of biologically active glucocorticoids (option 3). Plasma concentration of glucocorticoids may or may not control adrenal activity via negative feedback through the anterior pituitary (option 4). a. Changing Sensitivity To Environmental Stimuli (Option 1). During a severe snowstorm, Lapland longspurs, Calcarius lapponicus, had a higher adrenocortical response to capture compared to that in other weather conditions, suggesting increased sensitivity of the HPA axis in response to severe conditions (Asheimer et al., 1995). In the common diving petrel, Pelecanoides urinatrix, corticosterone levels usually decline with improved

432

HERIBERT HOFER AND MARION L. EAST

I environment 1 .L@ 1

I 1 I

roid binding

gluco-

I

J

I gluconeogenesis I

I pathological s t a t e s I

mating s u c c e s s

jimmunosupressionl

I

I

FIG. 4. The basic components of the adrenocortical response system, the hypothalamicpituitary-adrenocortical (HPA) axis. The anterior pituitary secretes adrenocorticotropic hormone (ACTH), which in turn stimulates the production of glucocorticoids in the adrenals. The biologically active fraction of glucocorticoids is the fraction of total plasma glucocorticoids that remains after most glucocorticoids are bound to corticosteroid binding proteins (CBG). Negative feedback of glucocorticoids on the anterior pituitary regulates the rate of production of ACTH. The actions of glucocorticoids may contribute both fitness benefits (gluconeogenesis. sometimes associated with increased mating success) and fitness costs (pathological changes in organs and immunosuppression). Comparisons between species suggest that natural selection has tuned the system in different ways in at least four compartments. (1) The anterior pituitary may change sensitivity t o environmental stimuli. (2) Adrenal activity and gonadal activity may be negatively coupled o r uncoupled. (3) Gonadal activity may either increase or decrease rate of production of CBGs. (4) Negative feedback is either present or switched off. Continuous lines: positive effects: dashed lines: negative effects: dotted lines: effects can be positive o r negative o r the two compartments can be uncoupled.

body condition. This effect vanished during stormy weather even though body condition differences persisted, suggesting a change in the sensitivity of the HPA axis in some animals (Smith et al., 1994). 6. Negative Coupling of Adrenal and Gonadal Activity (Option 2). Increased plasma levels of glucocorticoids typically impair fertility and reproductive function, and reproductive function typically reduces the capacity to increase glucocorticoid production (Greenberg and Wingfield, 1987). For instance, in wild baboons, Papio anubis, stress-induced suppression of

BIOLOGICAL CONSERVATION A N D STRESS

433

testicular function and testosterone production is due to cortisol (Sapolsky, 1985). Prolonged captivity changed adrenal gland morphology and led to reproductive failure in female nine-banded armadillos, Dasypus novemcinctus, in a manner similar to free-ranging individuals experiencing a harsh winter (Rideout et al., 1985). The link between gonadal and adrenal activity may pose a problem during periods of reproduction if predictable (seasonal changes) or unpredictable environmental events (inclement weather) activated the HPA axis, perhaps maximizing survival but interfering with gonadal function. We would therefore expect different species exposed to different kinds of environmental stimuli and under different breeding regimes to vary in the way they weigh the importance of responding to inclement conditions versus the continuation of reproduction. Wingfield et al. (1995b) considered the following hypotheses: (1) In the Arctic, a very short breeding season and the potential of summer snowstorms can restrict food supplies to passerine birds. It should be adaptive for arctic passerines to show less sensitivity to environmental stressors than midlatitudinal species to avoid delays in breeding that might prevent a breeding attempt in that season. This reasoning also applies to other habitats where the severity of environmental conditions is similar. (2) If adverse environmental conditions persist, then the HPA axis should become sensitized to prevent debilitating consequences. (3) Larger species, or individuals, should have relatively larger energy reserves and thus should show a muted response to acute stressors compared to smaller ones, and individuals with a high fat score should respond less than individuals with a low fat score. (4) Short-lived species should be more resistant to acute stressors than long-lived species that have several opportunities to breed. Hence, short-lived species should show a muted response compared to long-lived species. ( 5 ) Species, or individuals, with a high level of parental care should show a muted response compared to species, or individuals, with little parental care. Evidence is equivocal, both supporting and failing to support some of these predictions. The adrenocortical response to acute stress was muted in many bird species in the Arctic (see Wingfield et al., 1995b) and the Sonoran Desert (Wingfield et al., 1992b), sensitized during Arctic summer snowstorms (Astheimer et al., 1995), reduced in large individuals or those with a high fat score (Wingfield et al., 1994a,b), and muted in species and sexes that display high levels of parental care (Wingfield, 1986; Wingfield et al., 1995b). In contrast to the predictions, the adrenocortical response to acute stress was the same, or even higher and more variable, in some populations in the more severe habitat and with the shorter breeding season (Astheimer et al., 1994; Wingfield et al., 1994b, 1995a), and there appeared

434

HERIBERT HOFER A N D MARION L. EAST

to be no effect of body mass in some species, or of longevity in general (Wingfield et al., 1995b). c. Uncoupling Adrenal and Gonadal Activity (Option 2). In all observed cases, an uncoupling of adrenal and gonadal activity was associated with changes to CBG production (option 3) and the negative feedback loop (option 4). The adaptedness of these changes has not yet been fully explored and there are currently no hypotheses that convincingly explain why such changes occurred in some species but not in others. The principal benefit appears to be the mobilization of tissue nitrogen as a source of protein and of energy reserves through gluconeogenesis at a time of peak mating activity. This requires an increase in the concentration of biologically active glucocorticoids either by manipulating the amount of CBG (option 3) or by increasing the total amount of glucocorticoids by suppressing negative feedback (option 4), which would otherwise reduce adrenal activity. We later describe studies of a passerine and dasyurid marsupials where these effects have been clarified. Similar systems operate in salmonids (Hare and Robertson, 1959; Robertson and Wexler, 1957, 1960) and the plaice, Pleuronectes pfatessa (Wingfield and Grimm, 1977). In Gambel’s white-crowned sparrow, Zonotrichia leucophrys gambelii, adrenal activity during the breeding season is increased by uncoupling (option 2) and lack of negative feedback (option 4; Astheimer et af., 1994). Because testosterone simultaneously increased corticosterone-binding capacity (option 3) in males (Wingfield and Farner, 1980), corticosteronebinding capacity did not decrease below maximum corticosterone plasma levels. Concentration of biologically active corticosterone was still effectively controlled, preventing the potentially pathological effects of high concentrations of active glucocorticoids (Fig. 4). In several species of dasyurid marsupials, males increase the rate of gluconeogenesis during the breeding season by uncoupling, eliminating negative feedback (McDonald et al., 1986; Bradley, 1990) and decreasing corticosteroid-binding capacity through androgen stimulation, thereby losing control over the concentration of active glucocorticoids (Bradley et al., 1980;Bradley, 1987). Within a few days or weeks, the soaring concentrations of active glucocorticoids cause immune suppression (Barker et af., 1981), a reduction in imrnunocompetence, gastric ulcers, diseases, and lethal hemorrhage (Bradley et al., 1980; Bradley, 1987), leading to total male mortality within two weeks of mating (Lee et al., 1977; Lee and Cockburn, 1985; Bradley, 1987). The evolutionary origins of this kind of mass male mortality at the end of the mating season are unclear. In life-history terms its ultimate cause is a question of the evolution of semelparity, which has been extensively studied in other taxa (Roff, 1992; Stearns, 1992). Its proximate cause was

BIOLOGICAL CONSERVATION AND STRESS

435

cited by Lee et al. (1977) as an example for the stress hypothesis advanced by Christian (1971, 1978, 1980). This hypothesis states that an increase in population density raises the level of agonistic behavior between individuals provoking a stress response that ultimately impairs survival and reproduction. It viewed such stress responses as nonadaptive. Lee and McDonald (1985) have reviewed the evidence and concluded that there is scant empirical evidence in support of this hypothesis. An alternative view (Lee and Cockburn, 1985) suggested that such a stress response is adaptive because increased gluconeogenesis during times of intense competition for mates when food is in short supply reduces the need to spend time on foraging. However, this and other adaptive hypotheses cannot accommodate all cases of dasyurid species with total male mortality (Cockburn, 1997). This debate suggests that even spectacular cases of mass mortality involving pathological states may not necessarily be a consequence of nonadaptive, aberrant pathological events. They may instead be part of a suite of characters that maximize Darwinian fitness within the constraints set by past phylogeny and current ecology.

3. Metabolic Rate and Energetics Breeding experiments demonstrate that stress resistance is often related to a genetically reduced metabolic rate (Parsons, 1993b). Cattle bred for high growth rate under tropical conditions and increased resistance to high parasite loads had lower mass-specific metabolic rates under fasting than did controls (Frisch, 1981). When metabolic rate is genetically reduced, genetic correlations may increase resistance to several environmental stresses simultaneously (Hoffmann and Parsons, 1989). Stress-tolerant and stress-intolerant individuals within the same population vary in their metabolic efficiency and capacity for growth and reproduction, suggesting that stress-resisting processes are energetically expensive (Sibly and Calow, 1989; Calow and Sibly, 1990; Parsons, 1990c; Tranvik et al., 1993; Wynn et al., 1995). Experimental investigations of the link between metabolic efficiency and the adrenocortical response have shown that activation of the HPA axis reduces metabolic efficiency (e.g., Wingfield, 1988). Domestic lambs that were immunized against ACTH and thus could not activate the HPA axis increased metabolic efficiency by up to 20% during exercise, as assessed by a reduction in oxygen consumption (Wynn et al., 1995). In birds, corticosterone reduces responsiveness to external stimuli and thus may promote nocturnal restfulness (Buttemer et al., 1991). Thus, an energetic approach that considers the metabolic cost accompanying major genetic changes in response to stressors of various types may be useful (Parsons, 1990~). This may also be helpful in predicting geographical distribution and physiological attributes of species. For instance, the range

436

HERIEERT HOFER AND MARION L. EAST

of wintering bird species in North America may be limited by the amount of energy needed to keep warm, regardless of body size or habitat; genotypes to function beyond these stressful limits probably do not exist because of physiological limitations (Root, 1988). Lower metabolic rates of Galapagos fur seal, Arctocephalus galapagoensis, females in comparison with Antarctic fur seal, Arctocephalus gazella, females are probably an adaptation to reduce thermal stress on land in the warm equatorial habitat (Costa and Trillmich, 1988). Careful experimental studies have shown that parent birds work well below the maximum effort they can sustain when rearing young (Masman et al., 1989), although increased daily work may have detrimental fitness consequences (Daan et al., 1996) or may increase the incidence of malaria, suggesting that the extra effort of tending a brood can reduce immunocompetence (Oppliger et al., 1996a,b). Measurements of corticosterone concentrations in parent birds that increased their effort during the breeding season also provide an equivocal picture. In some studies adults that tended larger broods were not unduly stressed by extra efforts of feeding more nestlings (Hegner and Wingfield, 1987), whereas in others successful territory owners had the highest circulating levels of corticosterone during most of the breeding season, implying greater energetic demands and stress (Beletsky et al., 1989). Trade-offs in energy allocation have been repeatedly studied in ecotoxicology (Walker er al., 1996). Introducing animals from toxin-free environments to a toxic environment is likely to change the optimal allocation of resources compared with a toxin-free environment (Holloway er al., 1990). 4. Social Factors

The importance of social factors has been effectively summarized by the hypothesis that social skills, the control and predictability over the outcome of social interactions, and the available social support, determine the level of stress. Thus, it is the extent to which an organism feels challenged by a social context, rather than the context as such, that determines its physiological attributes (e.g.. cortisol or catecholamine concentrations) and its stress response (change in these concentrations; Henry and Stephens, 1977). This hypothesis predicts that individuals (1) with appropriate social skills and an experience of a high degree of control of social interactions, and (2) that have high social support and may reliably predict the outcome of interactions, should show low levels of glucocorticoids and a small hormonal and cardiovascular stress response. Individuals with a low degree of control over social interactions, that cannot predict the course and outcome of interactions, and that lack social support or appropriate skills should show elevated levels of glucocorticoids and a strong response to socially stressful

BIOLOGICAL CONSERVATION AND STRESS

437

situations. Supporting evidence has come from many studies of mammals in both captive and wild settings (Sapolsky and Ray, 1989; Sachser and Lick, 1991; Sachser et al., 1994; Sachser and Beer, 1995; Sapolsky, 1997). Individuals that are predicted to show a strong response often, but not invariably, include individuals of low social status (Sapolsky, 1997). However, because predictability is an important component, stability or instability of social situations may be more important than social status in some cases. Several studies have demonstrated that in stable contexts low-ranking individuals may respond similarly to high-ranking individuals, whereas differences in baseline levels and in stress response are pronounced in both high- and low-ranking individuals if situations change from social stability to social instability (e.g., Sachser, 1994). Thus, the same context may present itself as a different challenge to different individuals of the same population, or to individuals from different species. Holding captive primates in large groups may be very stressful for socially monogamous species but may cause little stress for species living naturally in large groups (Mendoza and Mason, 1986).

5. Elements of a Theory of the Stress Response How can the existence and effects of modulating factors be explained? The appropriate evolutionary theory for such phenomena would consider the stress response as a trait within the context of an individual’s life history and use an optimality approach to identify traits favored by natural selection (Sibly and Calow, 1986, 1989; Roff, 1992; Stearns, 1992; McNamara and Houston, 1996). Throughout its life, an organism allocates resources to production (growth and future reproduction) and to defense mechanisms that reduce the chances of mortality. It is often the case that improving defense mechanisms (e.g., by building detoxification procedures against environmental pollutants) results in a reduced allocation of resources to production. Trade-offs between life-history traits, particularly mortality and production, are central to life-history theory, which aims to explain how selection acts on traits when such trade-offs are involved (Sibly and Calow, 1986,1989; Roff, 1992; Stearns, 1992). How animals respond to such tradeoffs may depend on their state, for example, the amount of fat stored in body tissue. Many modulating factors in Fig. 3 characterize variation in the state of an individual. State-dependent models of life histories (McNamara and Houston, 1996) will therefore be an important tool for future theoretical conservation-oriented studies. An example is the verbal model of Wingfield et al. (1997, in press) that describes a general state-dependent “emergency life history stage”. The development of life-history models for stress response traits is uneven. Within the context of environmental pollution (ecotoxicology), there

438

HERIBERT HOFER A N D MARION L. EAS?

are testable models that examine the impact of stress on resource allocation within an individual and its consequences for population dynamics and population persistence (Kooijman et al., 1989; Calow and Sibly, 1990; Holloway et al., 1990; Baveco and De Roos, 1996; Walker et al., 1996). Other contexts currently lack such models. However, it should be possible to expand on current hypotheses, for example, on variation in the adrenocortical response in relation to the ecology of reproduction (see previous discussion), by placing them in a framework of state-dependent life-history evolution. Models developed for other traits, for example, state-dependent lifehistory models for clutch size evolution (McNamara and Houston, 1992), could be modified to predict how individuals should respond to stress as a function of their body condition, the amount of parental work they do, and so on. For instance, a trade-off well known in behavioral ecology is that between risk of predation and/or starvation and foraging. The optimal behavioral response to such a trade-off depends on the amount of fat stored in body tissue and other factors (see, e.g., Lima, this volume; Krebs and Davies, 1997). As human disturbance is often considered equivalent to a form of predation risk (Sections IV,D,2, and VII, Sutherland, 1996), this approach may be fruitful in conservation-oriented studies to explain variation in alarm reactions, for example, flight distances or aggregation behavior, as a function of body condition, breeding status, and other factors (Section V,C,2). 6.

The Stress Response as an Adaptation

Can the stress response be considered an adaptively tuned trait if stress is supposed to have detrimental fitness consequences? A stress response would be considered an evolved adaptation if there was genetic variation in the stress response (see Parsons, 1988b, 1993c) and natural selection tuned the stress response to match the challenge posed by the environmental stimulus (see Sober, 1993, for a more extensive discussion). As Lee and McDonald (1985) show, Selye (1946) in his original concept of a “general adaptation syndrome” viewed stages one, alarm, and two, “adaptation” (the organism adjusts to the presence of a stressor) of the organism’s response as adaptive in the evolutionary sense. That is, organisms showing such responses would be favored by natural selection over those that do not. Stage three, exhaustion and collapse of the system, was viewed by Selye as maladaptive or nonadaptive in the evolutionary sense. Do such events imply a nonadaptive situation? Or can we argue that a stress response is an evolutionary adaptation even though pathological states occur in many or even all cases? The answers to these two questions are no and yes. Whether or not adaptive tuning eliminates detrimental fitness consequences depends on a number of factors: (1) If current antropogenic condi-

BIOLOGICAL CONSERVATION AND STRESS

439

tions are different from those under which the stress response evolved, then the stress response may not work well because, for instance, the magnitude and/or frequency of stress due to human actions is greater than under natural conditions. However, the stress response would still be considered an adaptation for the conditions under which it evolved. (2) It is possible that natural selection favored a stress response that increased the Darwinian fitness of the genotype but included pathological states. Such a condition could evolve if fitness benefits outweigh detrimental consequences (costs); focusing only on costs would be inappropriate. A likely example is the stress-induced mortality of male dasyurid marsupials discussed earlier. (3) The evolved response may be the optimal option among available ones in the sense that it minimizes detrimental fitness consequences but does not eliminate them. Selection would favor this response and the evolved trait would be viewed as adaptive (Williams, 1966; Sober, 1993). Thus, we have to be careful: It is unwise to assume that a response could not have been adaptively tuned it if fails to protect the organism from all detrimental fitness consequences. Natural selection maximizes fitness but not necessarily the well-being of organisms.

C. THEIMPACTOF STRESSON

AN

ORGANISM

A detailed review of the impact of stress on the organism and the development and interactions of the energetic, hormonal, and immunological subsystems of a stress response is given by Toates (1995) and several contributions to this volume, particularly those by von Holst and Apanius. Energetic costs of avoiding predators (including human disturbance) through increased vigilance or the necessity to shift into safer but poorer microhabitats and perhaps increased competition for refuges in prey are reviewed by Lima (this volume). The purpose of this section is therefore t o point to some impacts that are not yet sufficiently appreciated but relevant for biological conservation purposes. 1. Stress May Modify Resource Allocation and Individual Energy Budgets

Coping with stress is usually energetically expensive, favoring genotypes that have a reduced metabolic rate (see previous discussion). Whereas numerous measurements of the energetics of foraging or breeding under natural conditions are available, direct energetic measures of the consequences of anthropogenic disturbances, including handling, intervention by researchers, and pollution are rare. One such study calculated that each human approach to an incubating great cormorant, Phalacrocorax carbo,

440

HERrBERT HOFER AND MARION L. EAST

colony that induces incubating birds to fly off requires an additional consumption of 23 g fish per bird or 23 kg per disturbance event for a typical colony (GrCmillet et a/., 1995). Speakman et a/. (1991) demonstrated that each event of a tactile stimulus significantly decreased fat stores of several species of hibernating bats by .05 g, whereas each event of a nontactile stimulus (head torch, photographic flash, sound, speech, temperature increase) decreased fat stores by merely .001 g. Contamination of sea otters, Enhydra lutris, with crude oil increased the thermal conductance of the fur by a factor of 1.8. This caused the otters to increase their average metabolic rate by a factor of 1.9 through voluntary activity, and shivering, time spent grooming, and swimming by a factor of 1.7. Metabolism and thermal conductance returned to baseline levels 3-6 days after cleaning (Davis e f al., 1988). In the mountain crab, Pseudothelphusa garmani, handling doubled the average increase in aerial oxygen uptake (Innes et a/., 1986). The impact of social and environmental stress on allocation of resources during juvenile growth has been experimentally studied. In general, stressors slow down growth. For instance, polychlorinated biphenyl (PCB) intake slowed growth of captive-bred mallards, Anus platyrhynchos, but not that of wild wood duck, Aix sponsa (Brisbin et al., 1986), suggesting that there are species-specific responses to environmental contaminants. Scott and Koehn (1990) demonstrated that heterozygous coot clams, Mulinia lateralis, grew faster than homozygous individuals and that heterozygote advantage was significantly more important when individuals grew up under conditions of environmental stress, suggesting that environmental stress accentuates genetic variation in stress resistance. Correlational evidence suggests that subordinate juvenile steelhead trout, Salmo gairdneri, experienced higher rates of aggression and grew more slowly than dominant juveniles even though both categories of individuals received equal food rations (Abbott and Dill, 1989). Additional experiments would be required to check that there were no genetically based differences in growth rate that determined social status in the first place. Several studies have assessed the increase in energy expenditure caused by back-mounted data loggers in swimming or flying animals (Obrecht et a/., 1988; Culik and Wilson, 1991a, 1992; Bannasch et al., 1994; Culik et al., 1994). Additional energy expenditure because of data loggers may be as high as 42% (Culik and Wilson, 1991a), but can be substantially reduced by optimizing size, shape, and location of the device (Bannasch et al., 1994). Many studies have recorded changes in heart rate in response to anthropogenic disturbance. Their value is limited because an increase in heart rate does not necessarily constitute evidence of stress. Evidence for stress would require comprehensive records that compare intensity and frequency of occurrence of heart rate changes due to anthropogenic stressors with

BIOLOGICAL CONSERVATION AND STRESS

44 1

changes in natural contexts, and a demonstration that the observed changes in heart rate have detrimental fitness consequences through an increase in energy expenditure or other mechanisms. 2. Stress May Impair Immunological Competence

Apanius (this volume) has reviewed in detail the relationship between stress and the immune system. We therefore restrict our comments to conceptual and factual issues that are still insufficiently appreciated by many practitioners of biological conservation. Incorrect but widely held beliefs include: (1) The immune system may be impacted only if there is chronic stress. (2) Stress does not influence an organism’s capacity to cope with an infection by a pathogen. (3) Latent viruses (viruses that are inactive and whose presence inside the host cannot be demonstrated yet after the appropriate stimulus become active and pathogenic) are rare, are benign, and are not activated by stress. (4) Conservation activities do not need to consider the current physical condition of an organism, that is, stress caused by handling cannot impair the organism’s capacity to cope with current infections. a. Stress Impairs Immunological Functions. It is increasingly recognized that stressors modify an organism’s immunological competence. For instance, glucocorticoids inhibit immune function by inhibiting interleukin 1 synthesis (Haour et al., 1995). Careful, detailed measurements of the impact of stressors relevant in a biological conservation context have been undertaken only recently. In harbor seals, Phoca vitulina, the consumption of organochlorine-contaminated prey reduced the cytotoxic activity of peripheral blood mononuclear cells (natural killer cells) to a level approximately 25% lower than that in controls (Ross et al., 1996b). Several studies have recently considered stress associated with social instability and aspects of housing in captivity, including “enrichment” of housing conditions in primates .and rodents. In rhesus monkeys, Macaca mulatta, perturbation of early rearing environment created numerous behavioral abnormalities and stereotypies, but also reduced the proportion of CD8 cells, natural killer cell activity, and increased lymphocyte proliferation response to mitogen stimulation (Lubach et al., 1995). In the same species, “enriched” housing enhanced phytohemagglutinin-stimulated production of interleukin 4 and gamma interferon but reduced natural killer cell activity (Schapiro et al., 1996). In pigtail macaques, Macaca nemestrina, and bonnet macaques, Macaca radiata, separation of infants from their mothers for 2 weeks increased infant vocalization and caused them to withdraw into a huddle. Changes in indicators of innate immunity (e.g., natural killer cell activity) paralleled behavioral changes but were transient (Laudenslager et al., 1996). In longtailed macaques, Macaca fascicularis, a change of membership in captive

442

HERIBERT HOFER AND MARION L. EAST

social groups produced aggression and fear, increased lymphocyte counts, reduced natural killer cell activity, and decreased lymphocyte proliferation particularly among individuals showing high levels of fear (Line et al., 1996). None of these studies has, however, considered or demonstrated detrimental fitness consequences of the experimental stressor. b. Stress Impairs the Organism’s Response to Pathogen Infection. There are some experimental studies that assess the potentially detrimental impact of stress on an organism’s capacity to respond to pathogen infections. An experimental study in domestic pigs, Sus scrofa, demonstrated that fluctuating temperatures enhanced the susceptibility to Aujeszky’s disease virus (ADV) infection because fewer immunoglobulin-containing cells responded to ADV infection (Narita et al., 1992). Hybrids of wild boar Sus scrofa and domestic pigs that were transported with unfamiliar conspecifics for 5 h in confinement showed a reduction in interferon-a production after exposure to a standardized monolayer of porcine cells infected with ADV (Wattrang et al., 1994). Barnard et af. (1996) studied laboratory mice that were placed in “enriched” housing facilities. In support of previous studies, enrichment of cages with objects that are defensible increased rates of aggression, which in turn reduced immunocompetence, as measured by immunoglobulin G concentrations and resistance to experimental infections with a protozoan parasite. However, the provision of refuges as part of the enriched environment helped to alleviate some of the detrimental effects. Thus, environmental enrichment may start a complex chain of behavioral and immunological changes. Experimental injection of glucocorticoids in two species of dasyurid marsupials increased the prevalence of lesions in the prostate (cytomegalic disease) caused by a herpes virus (Barker et al., 1981). Several experimental studies of laboratory mice, Mus musculus, have been able to clarify the precise mechanism by which immobilization and manual restraint impair the organism’s response to a variety of viruses (Ozherelkov et al., 1990; Bonneau et al., 1991; Kramskaya et al., 1991; BenNathan, 1994; Pokhil’ko et al., 1995). Many observational studies suggest a detrimental impact of stress on coping with an infection. For instance, synergistic effects between pollution and pathogen infections that may lead to mass mortalities have been considered in the context of catastrophic population declines in marine mammals (Section V,A). A problem with observational studies can be that stress may be used as an explanation of “last resort.” However, there are some observational studies in which the evidence is suggestive. Moving a captive herd of Dall’s sheep, Ovis dafli, to a new exhibit in a zoo may have caused an outbreak of pneumonia caused by Mycoplasma ovipneumoniae (Black et al., 1988). Transportation and confinement may have caused immunosuppression in yellow baboons, Papio cynocephalus, exacerbating the conse-

BIOLOGICAL CONSERVATION AND STRESS

443

quences of mixed helminth infections and causing substantial mortality of 30% (Farah, 1996). Harsh environmental conditions in a marginal population of apparently healthy white-tailed deer were held responsible for a high (53%) seroprevalence of a herpes virus (Lamontagne et al., 1989). Low water levels and high population density substantially increased mortality of white sturgeon, Acipenser transrnontanus, infected with white sturgeon iridovirus (LaPatra et al., 1993). Osmotic stress exacerbated the consequences of infection with the iridovirus-like VEN virus and led to mass mortality in Pacific herring, Clupea harengus pallasi (Meyers et al., 1986). An epidemic outbreak of bluetongue virus disease and high mortality was recorded in a herd of indigenous domestic sheep that walked in strong sunlight for 5 days. This outbreak was surprising given the relatively mild nature of bluetongue in indigenous sheep breeds (Eisa et al., 1980). c. Reactivation of Latent Viruses. The most detailed studies on latency and reactivation of viruses have been done on herpes viruses, where it is now possible to experimentally establish latent infections in the gut of mice (Gesser et al., 1994). Jenkins and Baum (1995) review what is currently known about herpes simplex virus (HSV) latency and reactivation in humans and consider mechanisms by which stress-induced changes in the host immune and nervous system might assist in establishing or reactivating latent viral infections. Because the existence of latent viruses is hard to prove, identification of their presence in wildlife is difficult. They may therefore be much more common than current records suggest. Nevertheless, reactivation of latent viruses by stressful conditions and glucocorticoids has been observed or experimentally demonstrated across a wide range of vertebrate hosts and viruses (Table I). Reactivation of latent rabies virus has also been implicated in the extinction of a study population of African wild dogs in the Serengeti, Tanzania (Burrows, 1992; Burrows et al., 1994).

3. Prenatal and Perinatal Stress Modifies Behavior and Irnrnunocompetence Another aspect of stress that is little understood is the impact of environmental stressors on fetuses or infants shortly after birth. Because neurotransmitter systems also continue to develop after birth, interactions between the individual and its environment after birth may affect the development of these systems and have long-term effects on behavior (Rogeness and McClure, 1996). Experimental studies on captive rhesus monkeys, rats, mice, cats, and guinea pigs demonstrated that the impact of environmental stimuli during these periods may alter many aspects of neurological structures, physical and behavioral maturation, growth, locomotory competence, cellular immune response, the incidence of exploratory play, aggressive, abnormal, and disturbance behavior, and gonadal activity

TABLE I EXPERIMENTAL A N D OBSERVATIONAL STUDIES THAT SUGGEST OR DEMONSTRATE REACTIVATION OF LATENT VIRUSES BY POTENTIAL STRESSORS Virus

Study host

Herpes virus

American plaice. Hippoglossoides platessoides Several parrot species

Adenovirus

Common murre

Infectious laryngotracheitis (ILT) virus Herpes virus

Domestic chicken

Lymphocystis disease virus

B virus (Herpesvirus simiae) Infectious bovine rhinotracheitis virus (IBR) Rabies virus Rabies virus Pseudorabies virus (PRV)

Brush-tailed phascogale, brown antechinus Several species of macaques

Stress

Results

Reference

Reduction of salinity

Reactivation of latent iridovirus

Berthiaume et (1993)

Treatment of psittacosis

Reactivation of latent herpes virus caused massive mortality Activation of a latent viral infection in the kidney Onset of egg laying reactivated shedding of large amounts of virus; other treatments reactivated sometimes or not at all Reactivation of latent infection in kidneys

Eskens et nl. (1994)

The most effective known inducer of latent B virus, in adults with low antibody titers against the virus Reactivation of virus, in a study that attempted to investigate reactivation of latent herpes virus Reactivation of latent virus and subsequent mortality Reactivation of latent virus and subsequent mortality Reactivation of latent pseudorabies virus in 8 of 9 piglets with maternal antibodies born to 2 vaccinated sows at postinoculation months 3-4

Zwartouw er a/. (1984)

Oil pollution, handling, confinement Unfamiliar birds, onset of egg lay. corticosteroid treatment Corticosteroid injection

Calf of domestic cattle

Assemblage of groups of adult strangers to form breeding colonies Dexamethasone injection

Raccoon. Procyon loror

Cortisone treatment, pregnancy

Guinea pig. Cuvia aperea Domestic pig

ACTH injection. crowding Injection of corticosteroids

a/.

Lowenstine and Fry (1985) Hughes ef al. (1989)

Barker et al. (1981)

Castrucci er a/. (1980) McLean (1975) Soave et ul. (1961): Soave (1964) van Oirschot and Gielkens (1984)

BIOLOGICAL CONSERVATION AND STRESS

445

during subsequent adulthood. These alterations persist for months or years and may be permanent (Insel et al., 1990; Maestripieri et af., 1991; Meaney et al., 1991; Schneider, 1992c; Clarke and Schneider, 1993; Schneider and Coe, 1993; Clarke et al., 1994; Gonzalez et al., 1994; McCune, 1995; Lubach et al., 1996; Sachser and Kaiser, 1996; Coe et al., 1997). Stressful stimuli examined included cold temperature, handling of the pregnant mother or the neonatal offspring, injection of ACTH, repeated removal from familiar housing and exposure to short, unpredictable noises, and unstable social environments (single or chronic disruption of social relationships) during pregnancy or lactation. However, the effect of stimuli need not always be negative (Section VIII). None of these studies has been undertaken on free-ranging animals, or has investigated fitness consequences of prenatal or perinatal stress.

D. INDICATORS OF STRESSED STATES Many studies have attempted to develop reliable criteria that indicate whether an individual or a population is in a stressed state, driven by Selye’s (1946) recognition of the “general adaptation syndrome” (Section 11,A). Selye’s insight stimulated hope that general indicators might be found that reliably signal a stressed state. However, the search for such indicators is often driven by the implicit assumption that all individuals of a population or a species will respond to a stressor in the same manner. This is unlikely to be so because it ignores the possibility of a conditional response that depended on the individual’s state (Section IV,B, Fig. 3). The impact of a stressor may also be contingent on the individual’s state in that the same environmental stressor may have detrimental fitness consequences if the individual is in bad shape but no impact if it is in good shape. Indicators should therefore take the individual’s state into consideration. The search for reliable indicators is also difficult because physiological processes that are part of a stress response are frequently activated by the organism in other contexts (Section 11,A). Indeed, it may be parsimonious to assume that natural selection used physiological systems that already operated in other contexts and modified them in such a way that they may also be used as part of a stress response, rather than assume that entirely new systems are specifically dedicated to dealing with stress (Sibly and Calow, 1986). Particular physiological processes might therefore be necessary components of a stress response but on their own may sometimes be insufficient to demonstrate that an organism is stressed. Indicators may be valuable to answer some questions but not others. One indicator may be useful to answer the question “How much does anthropogenic disturbance reduce the effective population size in a given

446

HERIBERT HOFER AND MARION L. EAST

habitat?,” whereas others may be more useful to answer the question “Does pollution cause stress?” Other problems may occur: The time course of a stress response may be captured incompletely by the indicator chosen; or the stress response may include hormonal, energetic, and immunological components, that may be invoked to different degrees by different individuals, and measuring only one component may therefore provide a misleading picture (Section 111,A). Thus, the conclusion from this survey is again a note of caution. Selye’s (1946) “general adaptation syndrome” and more recent compilations of critical signs that assist in evaluations of stressed states (e.g., Wiepkema and Kolhaas, 1993) are helpful, but they should not be applied in an uncritical fashion. In particular, numerous studies demonstrate that the most common prejudice-lack of “stressful behavior” indicates lack of stress-is ill founded because behavior is often an unreliable guide to the extent to which an organism experiences stress. 1. Indicators of an Organism’s Stressed State

Most experimental studies on this have been on vertebrates, particularly mammals and birds, and asked the question how reliably the organism’s behavior indicates a stressed state. These studies have been conducted under both field and captive conditions. Before we turn to behavior, we briefly mention some other indicators. a. Molecular Markers. Examples of molecular markers as indicators of stressed states come from studies of invertebrates. Fang et al. (1991) demonstrated experimentally that the ATP content of the corals Acropora hyacinthus and A. formosa is a reliable indicator of the degree of “stress” caused by desiccation and bleaching. The ATP content changed rapidly after exposure and increased during the recovery phase in subsequent weeks. Cochrane et al. (1994) used nucleic acid probes based on chaperonin and 70-heat-shock protein (see Section IV,A) to monitor stress-related changes in mRNA abundance in cells of the rotifer Brachionus plicatilis after exposure to heat. They suggested that this method may be feasible for other marine invertebrates. Because stress increases energy expenditure (Sections IV,B,3 and IV,C,Z), a measure of the metabolic energy available to an organism may indicate how stressed the organism is. High levels of available energy would suggest stress-free conditions, low levels stressful conditions. Ivanovici and Wiebe (1981) examined the usefulness of a measure of available metabolic energy, the adenylate energy charge (AEC), expressed as AEC

=

[ATP] + ;[ADPI [ATP] + [ADP] + [AMP],

where [ATP], [ADP] and [AMP] are the amounts of adenosine triphos-

BIOLOGICAL CONSERVATION AND STRESS

447

phate, diphosphate, and monophosphate, respectively. Values of around .8-.9 were typical for stress-free optimal conditions, whereas stressful conditions with losses of viability usually led to values of the AE C of around .5 (Ivanovici and Wiebe, 1981). b. Torpor. Many mammals and birds use torpor to minimize energy expenditure during periods of low temperature (Lyman et al., 1982). This has led some authors to assume that the occurrence of torpor is a sign of energy reserve depletion, often called “nutritional stress” (Dawson and Hudson, 1970). Carpenter and Hixon (1988) used field observations of the rufous hummingbird, Selasphorus rufus, to argue that torpor in small endotherms does not necessarily indicate that animals are energetically stressed. Individuals in good condition may also use torpor to conserve fat reserves that, in the case of the rufous hummingbird, may be later used on migration. Because experimental field studies of torpor are rare, this issue deserves further investigation. c. Body Mass Changes May Be an Unreliable Indicator of a Stressed State. Body mass changes during the breeding season have been frequently debated as an indicator that breeding is “stressful.” We argued in Section II,B that this controversy really depends on whether body mass losses are an unexpected deviation from homeostasis or whether they are an adaptive change in anticipation of the demands of the breeding season. Thus, although body mass changes might suggest a stressed state, such a claim must be verified. d. Reproductive Suppression. The suppression of gonadal activity in socially subordinate mammals and birds has been frequently linked to “psychological stress” caused by agonistic interactions (harassment) with socially superior conspecifics (Bowman et ul., 1978; Keverne et ul., 1982; Wasser and Barash, 1983; Abbott, 1989; Bronson, 1989). If agonistic interactions increased because of environmental changes, for example, decreased food availability and reduced foraging efficiency caused by anthropogenic disturbance, could cessation of gonadal activity be a reliable indicator of environmental stress? Cessation could arise from general effects of raised glucocorticoid levels that interfere with gonadal activity (see Section IV,B) or from specific neuroendocrine mechanisms independent of the general adrenocortical stress response. In the former case, reproductive suppression would be an indicator of stress relevant to conservation, whereas in the latter it might not. Intensive work in recent years suggests that the social organization of a species probably provides the decisive clue. In cooperatively breeding birds and mammals a single dominant pair breeds and other adult group members assist in the rearing of the offspring. Reproductive suppression is a natural condition experienced by most members of a population at some stage of

448

HERIBERT HOFER AND MARION L. EAST

their life cycle even in free-ranging populations. In such species, specific neuroendocrine mechanisms are the prime cause of reproductive suppression, whereas the general adrenocortical response plays no role or only a minor one (Mays et al., 1991; Schoech et al., 1991; Wingfield and Lewis, 1993; Faulkes and Abbott, 1997; French, 1997; reviewed by Abbott ef al., 1997; Sapolsky, 1997). Because little is known about how such specific neuroendocrine mechanisms may be triggered by environmentally dependent changes in agonistic behavior, it is unclear to what extent the activation of such mechanisms may be used as an indicator of stressed states. In contrast, in many other group-living species, most or all females and/ or males attempt to reproduce. Socially subordinate individuals in such species may experience reproductive failure as a consequence of an increased production of glucocorticoids caused by harassment by dominant conspecifics (Wasser and Starling 1986; Sapolsky, 1987, 1997; Altmann el al., 1988; Dunbar, 1989). Note that such harassment has sometimes been interpreted as an adaptive reproductive strategy of social superiors (e.g., Wasser and Starling, 1986). Thus, reproductive suppression may occur because stress triggers a significant adrenocortical response, but further information is required to assess whether the stress derives from environmental conservation-relevant sources, or whether it is the consequence of adaptive behavior of conspecifics. e. The Adrenocortical Response. Among several components of the hormonal response system (Section II,A), the adrenocortical response is frequently cited as a key component of a generalized hormonal stress response that can be conveniently measured even under field conditions. This is because activation requires several minutes before a rise in plasma glucocorticoid concentration is detectable, and field conditions often require such time periods before a blood sample can be taken (Wingfield el al., 1996). It is not, however, appropriate to consider it uncritically as an all-embracing panacea. Stress responses need not involve the adrenocortical axis (Section II,A), and glucocorticoid production may serve other functions in the context of reproductive activity, basal metabolism, or mobilization of reserves (Buttemer et al., 1991; Wilson and Wingfield, 1992, 1994; Saltzman et al., 1994). However, the adrenocortical response is a reliable indicator in many cases, and there is evidence that raised levels of glucocorticoids induce profound changes in immunological competence (see previous discussion) and are associated with an increase in heart rate and energy expenditure (see following discussion). Such changes result in detrimental fitness consequences in dasyurid marsupials (see previous discussion) and alpine marmots, Marmota marmota (Arnold and Dittami, 1997). J The Sympathetic-Adrenal Medrtllary Response. In contrast to the adrenocortical response, the sympathetic-adrenal medullary response (produc-

BIOLOGICAL CONSERVATION A N D STRESS

449

tion of catecholamines such as epinephrine and norepinephrine) has been neglected in measurements of the hormonal response to stress in wildlife. This is possibly because its measurement poses greater difficulties than the measurement of the adrenocortical response. Because the sympatheticadrenal medullary response can operate very quickly, handling procedures for blood sampling may increase plasma concentrations of catecholamines severalfold within only 2 min (Le Maho et al., 1992). Le Maho et al. (1992) used remote-controlled blood sampling on completely undisturbed captive domestic geese, Anser anser to demonstrate that mean basal values for epinephrine and norepinephrine recorded by their procedure were 90fold and 5-fold, respectively, below the lowest values previously recorded, suggesting that significant handling effects on measurements must have been widespread. However, because of its potentially devastating effects on the cardiovascular system, hyperactivity of the sympathetic-adrenal medullary system in response to conservation-relevant stressors may be arguably as important as the adrenocortical response (Henry and Stephens, 1977). For instance, catecholamines were shown to be more important than cortisol or testosterone in the response by guinea pig males to agonistic encounters with other males (Sachser, 1987). The measurement of catecholamines therefore deserves closer attention by future studies. Instead of direct measurements of catecholamine concentrations, measurements of indicators may become feasible. A recent study on pigs demonstrated that the role of certain vocalizations may be a reliable indicator of increasing epinephrine concentrations caused by isolation of the individual from group members (Schrader and Todt, in press). g. Heart Rate. Heart rate has been repeatedly shown to be a reliable predictor of oxygen consumption, and thus of metabolic rate and energy expenditure, under field conditions (e.g., Culik, 1992). At the same time, heart rate may be closely related to plasma cortisol concentration (Harlow et al., 1987a,b). Whereas changes in heart rate may reliably track changes in metabolic rate, they do not necessarily indicate stress without knowing the behavioral or physiological context of changes in metabolic requirements. h. Behavior Is Often an Unreliable Indicator of a Stressed State. Maestripieri et al. (1992) reviewed numerous primate studies (mostly from captive settings) and came to the conclusion that the measurement of displacement activities might be a useful indicator of “psychosocial stress.” Escos et al. (1995) and Alados et al. (1996) introduced two parameters as indicators of stressful behavioral changes that characterize the temporal pattern of activity and feeding behavior. These are the fractal dimension of head-lift frequency and the power spectrum of the time distribution of feeding. Under “stressful conditions” (parasitic infection, pregnancy) the

450

HERIBERT HOFER A N D MARION L. EAST

complexity of temporal behavior patterning (fractal dimension) declines. Other indices have been suggested that attempt to capture the degree of disturbance experienced by animals in terms of their activity levels, amount of exploratory behavior, and other parameters, particularly in the animal welfare literature (see Jordan and Burghardt, 1986; Broom and Johnson, 1991). However, the relationship of such parameters to cardiovascular, adrenocortical, energetic, or immunological aspects of the stress response is frequently unclear, and their value for predicting fitness consequences is unknown. Some experimental studies in both captive and field settings have attempted to use physiological measurements to verify behavioral measures as reliable indicators of a stressed state. The most frequent measures were heart rate, often recorded by telemetry devices, and plasma cortisol concentrations. Behavior was not a reliable guide to elevation in heart rates of AdClie penguins, Pygoscelis adeliae, as a response to approaches to the breeding colony by people or aircraft (Culik et af.,1990; Wilson et at., 1991). Heart rates increased before any change in behavior became apparent and substantial increases in heart rate occurred even when behavioral changes were absent. Heart rate was also a better indicator than behavior of the stress response of free-ranging domestic sheep, Ovis aries, and bighorn sheep, Ovis canadensis, exposed to a variety of experimental stressors, including human disturbance (MacArthur et al., 1982; Baldock and Sibly, 1990), and correlated closely with blood cortisol levels, which in turn were an indicator of immunological competence (Harlow et al., 1987a,b). Similar results were obtained in captivity studies. In domestic geese, behavior was an unreliable guide to the physiological response to handling (Le Maho et al., 1992). Significant cardiovascular and hormonal response with simultaneous lack of a behavioral response were also recorded in baboons (Bentson et al., 1996), Goeldi’s monkeys, Callimico goeldii (Dettling and Pryce, 1996), rhesus monkeys (Preston et al., 1996), and squirrel monkeys, Saimiri sciureus (Martel et al., 1996). Experiments that measured the behavioral response of domestic pigs to an approaching human demonstrated that the pig’s behavioral response was not correlated with its response in terms of plasma cortisol concentrations or subsequent growth performance (Paterson and Pearce, 1992). Hurst et al. (1996) demonstrated that in captive rats Rattus norvegicus subjected to intense social pressure, activity increased, the sleep phase was reduced, and pathophysiological states were pronounced. 2. Indicators of a Population’s Stressed State

In population management the question frequently arises whether food shortage, pollution, or human disturbance reduce foraging success, thereby

BIOLOGICAL CONSERVATION AND STRESS

45 1

decreasing the animal’s nutritional status, reproductive success or survival, or the effective population size in a given habitat. Several measures have been proposed and tested that might be used as an early warning system. The analysis of urine samples collected from snow has been successfully used to develop measures of the physiological condition of populations of terrestrial mammals (wolf, Canis lupus: Mech et af.,1987; white-tailed deer: Delgiudice et uf., 1989; red deer, Cervus elaphus: Delgiudice et al., 1991). Williams and Rothery (1990) argued that in gentoo penguins, Pygoscefis papua, the duration of foraging trips was a useful measure of food shortage, provided the breeding stage, timing of trips, and the possibility of overnight trips were considered. They suggested that recording coarse patterns of the duration of foraging trips for many individuals was more valuable than detailed records for a few. It is difficult to assess population states in many seabird species because individuals are long lived, forage far at sea, and breed in colonies. Foraging trips are difficult to record and changes in population status may not be easily or quickly detected. Successful alternatives t o traditional measures of individual foraging effort or mortality are the measurement of the number of nonbreeders at breeding colonies and a change in age at recruitment into the breeding population, as demonstrated for the kittiwake, Rissa tridactylu, by Porter and Coulson (1987), and for the great skua, Catharacta skua, by Klomp and Furness (1992). In many ungulates nutritional stress may decrease growth rate, delay puberty, and decrease fertility of young females. Williamson (1991) showed that the number, fertility, and condition of young females worked well as an indicator of population performance in red lechwe, Kobus leche leche. Gill et al. (1996) investigated if anthropogenic disturbance of foraging pink-footed geese, Anser brachyrhynchus, reduced their opportunity to exploit food resources. They developed a general model that described the trade-off between resource use and risk of disturbance as conceptually similar to the trade-off between resource use and risk of predation. The model requires the measurement of the amount of resources in a number of patches, the proportion of these resources exploited by the animals, the total number of individuals supported by the resource, and a measure of disturbance on each patch. In the case of the geese this information was successfully applied to derive an estimate of how many animals a disturbed habitat could support if disturbance was reduced. Population density itself is not necessarily a good indicator of the stressed state of a population. If it were, density would be positively linked to habitat quality and thus provide a composite measure of the impact of environmental stressors. Van Horne (1983) and Reijnen and Foppen (1995) demonstrated empirically that density is a reasonable measure of habitat quality in years with low density. However, in years with high density,

452

HERIBERT HOFER AND MARION L. EAST

a decline in habitat quality as measured by density in that habitat was substantially underestimated because animals continued to crowd in habitats that had lost quality. In theories of habitat use, density also does not reflect habitat quality when overall density is high, whereas it might when overall density is low (Fretwell, 1972; Bernstein et al., 1991; Sutherland, 1996). Studies that attempt to assess the impact of anthropogenic stressors using population density therefore need to assess whether the population is at a high or low density. If an organism experiences environmental stress during its development it may have inadequate resources to buffer developmental processes against distortions introduced by environmental stimuli. Measures of developmental stability, most importantly fluctuating asymmetry, could therefore be used as an early warning system to indicate that a population is in a stressed state. This topic has been reviewed by Parsons (1990a,b, 1992), Clarke (1995, 1996), and in detail by Moller (this volume). Warwick (1986) developed a model of how pollution or other forms of environmental stress may modify the composition of freshwater and estuarine macroinvertebrate or macrozoobenthos communities. It suggests that stressors would remove large competitive dominant species from the community. Changes in community composition would be indicated by a comparison of abundance (numerical species diversity) with biomass (biomass species diversity). This approach works well for freshwater systems but not for estuarine communities (Meire and Dereu, 1990). A computer software package (RIVPACS) is now available that compares the actual faunal composition of a river with a predicted composition based on a small number of environmental features that characterize a site (Wright et al., 1993; Wright, 1995).

V. EFFECTS OF ANTHROPOGENIC STRESSORS In previous sections we used mainly experimental evidence from studies using natural or anthropogenic stress to answer questions about the measurement of the impact of stress and evaluate possible indicators of stressed states. We now summarize the effects of several anthropogenic stressors, again emphasizing experimental studies. We first introduce pollution and disturbance by tourism with case studies 2 through 4 (see next section) to illustrate the complexity of some of the issues involved. We then discuss the impact of hunting, noise, and climatic warming. Section VI deals with the impact of conservation management or research-related activities. We focus on the effect of anthropogenic stressors on individuals and populations because there is very little information on the impact on community struc-

BIOLOGICAL CONSERVATION AND STRESS

453

ture or composition. This is an area that will be of major importance in the future (Bodini et al., 1994; Knight and Gutzwiller, 1995). A. POLLUTION Environmental pollutants or contaminants have long been held responsible for an impairment of reproductive activity in natural populations (e.g., Krylov, 1990). However, pollutants rarely occur in isolation from other potential stressors. For instance, a Dutch harbor seal, Phoca vitulina, population declined from 350 to 17 individuals; suspected causes included high hunting pressure, environmental pollution, loss of habitat, and disturbance at resting places (Mees and Reijnders, 1994). Experimental evidence to separate the contributions of various factors is therefore vital. Research on environmental pollutants has concentrated on laboratory tests to identify lethal doses (risk assessment), the physiological effects of pollutants, and the geographical distribution and taxonomic occurrence of pollutants. There have been comparatively few experiments that assess the mechanisms by which pollutants reduce the Darwinian fitness of organisms in their natural environment. An exception to this was the detailed studies on the effects of pollutants on eggshell thinning in birds (reviewed by Peakall, 1993) and lead poisoning in swans (e.g., Sears et al., 1989); other examples are reviewed by Walker et al. (1996). The main reason seems to be that studies of individuals in free-ranging populations rarely look at pollution, whereas studies that look at pollution do not follow individual life histories in the field. The 1988 mass die-off of North Sea harbor seals is a well-documented environmental catastrophe that illustrates this and other points. 1.

Case Study 2: The Mass Die-Off of North Sea Harbor Seals in 1988

The mass die-off of North Sea harbor seals in 1988 is an interesting case that demonstrates the difficulties of identifying factors underlying population catastrophes and separating cause and effect. A massive research effort that continues to this day has led to the publication of many papers. Despite this effort, the cause(s) of the 1988 mass die-off have still not yet been conclusively identified, although a number of contributing factors have emerged. Progress was directly related to the introduction of experimental techniques, but historical reviews and comparisons with other populations and die-offs were also important. In 1988, around 18,000 harbor seals died in European waters. The first casualties were noted in Danish and Swedish waters in April 1988, and by autumn 1988 seals in Norway, Germany, the Netherlands, and the United Kingdom were affected (Dietz et al., 1989). The immediate cause of death was identified as phocine distemper virus (PDV) (Cosby et al., 1988). How-

454

HERIBERT HOFER AND MARION L. EAST

ever, this does not necessarily provide an explanation of the spatial and temporal pattern of mass mortalities. Where did PDV come from and why did it cause mass mortalities? Were other environmental factors involved? Several hypotheses were proposed (Simmonds, 1991). These were not mutually exclusive and focused on different aspects of the mass die-off: 1. Higher than average monthly temperatures may cause seals to leave the water and aggregate on land in unusually high densities. Large aggregations and high densities may have facilitated widespread and rapid transmission of pathogens (Lavigne and Schmitz, 1990). There was indeed a correlation between temperatures and several seal mass die-offs known to have occurred in the twentieth century, including the 1988 die-off (Lavigne and Schmitz, 1990). 2. Individuals that carried the virus may have originated from infected seal populations that “invaded” a PDV-naive population, thereby triggering mass mortality (Heide-Jorgensen el al., 1992). These invaders may have been either harp seals, Phoca groenlandica, from the Barents Sea that were known to have “invaded” Scandinavian waters in 1987 (Harwood and Grenfell, 1990), harbor porpoises, Phocoena phocoena (Kennedy et al., 1988), or a variety of seal species from Greenland, Canada, or the United States (see Simmonds, 1991). Analysis of serological samples demonstrated that the North Sea harbor seal population was PDV-naive and that a variety of other marine mammals carried PDV (Simmonds, 1991; Heide-Jorgensen etal., 1992). There are problems assigning the origin of PDV to a particular marine mammal population because the data on the geographic progress of the seal die-off and possible contact times of putative carrier candidates do not match well, and there are no data on actual contact rates between various species (Simmonds, 1991). It was doubted that the virus could have been effectively transmitted under field conditions because it requires rather specific conditions to survive transmission, but the virus must have spread well because Danish farmed mink, Mustela vison, became infected with PDV in 1989 (Heide-Jorgensen et al., 1992). Thus, the origin and mode of transmission of PDV in harbor seals is currently not known. 3. An exceptional bloom of the alga Chrysochromulina polylepsis in early 1988 may have produced high concentrations of toxic substances that may have stressed seals and suppressed their immune systems, which in turn may have reduced resistance to pathogen infection and caused quick mass mortality (Lavigne and Schmitz, 1990). However, there are discrepancies between the timing and spatial spread of the algal bloom and the initiation of the seal die-off, the potential toxicity of the algal bloom is unknown, and evidence on links between toxic algal blooms and previous

BIOLOGICAL CONSERVATION AND STRESS

455

marine mammal die-offs is contentious (Lavigne and Schmitz, 1990; Simmonds, 1991). 4. Marine pollution, particularly organochlorines such as PCBs, may have exacerbated the consequences of pathogen infection and thereby initiated mass mortality (Simmonds and Johnston, 1989). When Lavigne and Schmitz (1990), Simmonds (1991), and Heide-Jgrgensen et al. (1992) published their papers, evidence of the possible impairment of immunocompetence by contaminants such as organochlorines and the fitness consequences of contaminants was largely circumstantial and based on observational data in other marine mammals. Lavigne and Schmitz (1990) therefore proposed that this explanation contributed little explanatory power to other factors such as elevated temperatures and very high population densities. Similarly, Skaare et al. (1990) concluded that observed organochlorine and heavy metal concentrations from harbor seals in Norway gave no support to suggestions that organochlorines and heavy metal pollution may have been directly involved in the observed seal deaths. However, impairment of fecundity by organochlorines had been demonstrated by observational studies in Dutch harbor seals (Reijnders, 1980) and feeding experiments with Dutch harbor seals fed on PCB-contaminated fish (Reijnders, 1986). It was already known that exogenous toxins such as PCB and other organochlorines seemed to interfere primarily with the endocrine system causing changes similar to those present in hyperadrenocorticism in gray seals, Halichoerus grypus, and ringed seals, Phoca hispida (Bergman and Olsson, 1986). St. Aubin and Geraci (1986) identified the zona glomerulosa as an organ through which a stress response to contaminants or other environmental stimuli might exhaust adrenal hormone reserves or desensitize the adrenal cortex to other physiological stimuli. Data on blood biochemistry and thyroid hormone levels from relatively uncontaminated individuals became available in 1995 (Schumacher et al., 1995), studies on the historical progress of stress-related morphological changes were conducted in 1994 (Olsson et al., 1994), and the first experiments to investigate potential links between contaminants and impairment of immune function were published in 1994 (de Swart et al., 1994). More is now also known about historical changes in levels of contaminants in seal populations and what proportion of contaminants are transferred between mothers and pups. Basic blood biochemcial parameters of seals changed between areas containing polluted and unpolluted prey and were not affected by captivity conditions if the animals were held in captivity for a long time (9 months, Schumacher et al., 1995). Baltic gray seals and ringed seals suffered from a disease complex described as a primary lesion in the adrenals causing secondary reactions in various other organs, including

456

HERIBERT HOFER AND MARION L. EAST

skull bone lesions. The incidence of skull bone lesions has increased in Baltic seals since World War 11, indicating the presence of unnatural stress factors. Analytical results and pathological findings suggested a particular class of contaminants that include PCBs to be the most likely candidate to instigate the disease complex (Olsson et al., 1994). Finally, long-term experiments have demonstrated that immune function in harbor seals was impaired if they were fed fish from polluted waters (de Swart et al., 1994). Contaminated food not only changed cellular immunity, it also impaired virus-specific immune responses and thus caused immunosuppression (de Swart et al., 1993, 1995a; Ross et al., 1995). Exposure to contaminants may therefore have had an adverse effect on the defense against virus infections, affecting the severity of viral infections, survival rates, and the spread of infections during recent epizootics (Ross et al., 1996b). Further studies produced interesting evidence on vertical transmission of contamination between mothers and pups. Female Baikal seals, Phoca sibirica, transferred about 20% of their total DDTs and 14% of their total PCBs to the pup during lactation (Nakata et al., 1995). Short-term fasting typical for lactating mother harbor seals did not aggravate immunosuppression in animals with high burdens of organochlorines (de Swart et al., 1995b), but lymphocyte functionality and total immunoglobulin G levels were reduced in mothers at the end of lactation (Ross et al., 1993). Pups at birth and females late in lactation may therefore be more susceptible to infection by viral and bacterial agents than other population segments. Perinatal exposure to environmental contaminants represented a greater immunotoxic threat than exposure as juvenile or adult (Ross et al., 1996a). This spate of recent experimental and historical studies suggests that environmental pollutants may have contributed to the severity and extent of distemperlike infections in seals and dolphins in recent years (Osterhaus et al., 1995).

2. Experimental Studies of Physiological Responses to and Fitness Consequences of Pollution Detailed measurements under well-defined conditions of the impact of pollutants have not been undertaken until recently (cf. Dunnet, 1982). Experimental studies both in captivity and the field now provide evidence that the physiological response to and the fitness consequences of experimental pollution may be substantial, vary between conditions of application, species, and type of pollutant, and that short-term measures may misjudge long-term impacts of pollution on fitness (see following discussion; Sections III,B and IV,C,I; Table I; Walker et al., 1996). Fowler et al. (1995) demonstrated that even low levels of oil fouling in Magellanic penguins significantly elevated plasma corticosterone concen-

BIOLOGICAL CONSERVATION A N D STRESS

457

trations in females and reduced their fitness because few of the pairs with an oiled partner later established nests with eggs. PCB-contaminated food fed to captive mink before and during pregnancy increased urinary cortisol concentrations during pregnancy and were associated with an increase in the resorption of fetal tissue, and a reduction in the chance of successful parturition (Kihlstrom et al., 1992; Madej el al., 1992). In some cases, pollution may produce pronounced short-term effects on behavior of adults that may be associated with a reduction of reproductive success during the current breeding season. However, some studies found no significant long-term effects on population persistence by the specific pollutant and application method. Internal and external crude oil contamination of one adult in breeding pairs in Leach’s storm petrel reduced hatching success and fledging success in a dose-dependent manner, as adults temporarily abandoned their breeding burrows, but there were no longterm effects on reproductive success (Butler er al., 1988). In a 5-year study, the experimental application of the insecticide fenitrothion on Canadian forests caused temporary changes in habitat utilization in the chestnutsided warbler, Dendroica pensylvanica, the magnolia warbler, D. magnolia, and the white-throated sparrow, Zonorrichia albicollis, but no long-term changes in time budgets, suggesting that behavior was affected in the shortterm but not in the long-term (Millikin and Smith, 1990). Another study that also experimentally applied fenitrothion on forests found that whitethroated sparrows abandoned territories, or failed to defend them, that the application disrupted incubation, and may have been responsible for some clutch desertion (Busby et al., 1990). However, clutch size and hatching success were not affected. BY TOURISM AND LEISURE ACTIVITIES B. DISTURBANCE

Recreational activities and other forms of nonconsumptive utilization of wildlife are being increasingly recommended as a method to conserve populations, and have become very important for conservation because tourism is now the largest industry in the world and thereby a major economic force in many countries (Goodwin, 1996). Whether tourism can protect both indigenous people and places is an unresolved debate (King and Stewart, 1996). Understanding of the mechanisms of the impact of tourism on protected areas, their ecological significance, and the capacity to manage tourism in protected areas lags behind the growth of tourism to protected areas (Goodwin, 1996), because recreational, nonconsumptive use of wildlife does not fit well into the existing paradigm of wildlife management (Duffus and Dearden, 1990). Thus, the consequences of tourismrelated or recreation-related disturbance deserve increasing attention.

458

HERIBERT HOFER AND MARION L. EAST

Some examples illustrate this. Large caves used by bats for roosting in Mexico are preferentially targeted for tourism development but have the highest species richness and are the only places where rare, threatened, or endangered bat species occur (Arita, 1996). Uncontrolled tourist feeding is a health and welfare threat to the barbary macaques, Macaca sylvanus, of Gibraltar because they become overweight and their health deteriorates (O’Leary, 1996). Increased rates of tourists visiting and interacting with howler monkey, Alouatta pigra, troops led to acclimation to tourists and an 8%population increase over 10 years, although guides regularly touched, fed, and howled at one troop (Lash and Horwich, 1996). However, because ecotourism started only after the area received sanctuary status in order to protect the population from hunting, these results merely indicate that interactions with and disturbance by tourism had less impact than hunting, not that disturbance was not harmful. In one area in Morocco, high tourist numbers and general human activity were responsible for clumping of food resources in a low-density population of barbary macaques. Resource clumping in this population increased the frequency of aggressive behavior, caused fatalities from intraspecific fights, and had a destructive effect on the socioecological organization compared to a high-density population elsewhere where resource clumping did not occur (A.S. Camperio Ciani, personal communication). Studies of the effect of tourism disturbance have typically concentrated on recording behavioral responses of disturbed animals in observational studies, but have rarely attempted to use experiments to investigate the behavioral or physiological response, or the fitness consequences of disturbance. In this section we look at two case studies and summarize experimental studies of the behavioral and physiological responses to and the fitness effects of disturbance by tourism. A key problem is the design of appropriate control groups. Case study 3 on whale watching introduces these difficulties. Case study 4 on the potential disturbance of Antarctic penguins by tourism illustrates some of the most advanced techniques available to modern investigations, yet the question of how much penguins are stressed by tourism has not been fully resolved.

1. Case Study 3: Whale Watching Whale watching is one of the fastest growing forms of ecotourism, worth annually more than $35 million in direct and indirect benefits in eight countries, excluding two of the biggest whale-watching nations, the United States and Australia (International Fund for Animal Welfare [IFAW], 1996). Potential disturbance effects of watching cetaceans (whales and dolphins) include the presence of and the noise (see Section V,D) generated by the engines of whale-watching vessels. These may cause whales to change

BIOLOGICAL CONSERVATION AND STRESS

459

durations of dives, respiration patterns, and surface resting periods, reduce feeding or nursing, destroy herd integrity or disrupt social groupings, and displace them from preferred feeding habitats (reviewed by Evans, 1987, 1996; Richardson et al., 1995). Reaction to disturbance could thus adversely affect individual energy budgets, foraging tactics, distribution and habitat use, predation risk, and the buildup of reserves prior to migration. Data on the potential impact of whale watching are almost entirely confined to observations of short-term effects (minutes or hours, sometimes days) on behavior and have been primarily carried out on baleen whales, whereas toothed whales have been studied less often. A large number of behavioral studies, typically unpublished reports and dissertations, have been summarized by Richardson et al. (1995) and Evans (1996), from which the following examples are taken. In Hawaii, humpback whale, Megaptera novaeangliae, presence was inversely related to the amount of daily boat traffic. Short-term reactions to small boats at distances of 500-1000 m that suggest avoidance reactions include: (1) increased frequencies of dives started without raised flukes and surfacings without blows; (2) reduced time spent surfacing; and (3) movements out of favored areas. Humpbacks have also been known to sometimes “threaten” boats by “charging” toward them and “screaming” at them. In Alaska, humpback whales responded to vessel traffic and approaching boats at distances of several kilometers by increasing the duration of their dives, reducing the duration of surfacing, moving away from vessels, and temporarily abandoning preferred feeding areas. Off Cape Cod, humpbacks responded little to boats that followed established guidelines for whale watching by approaching slowly and steadily and keeping a minimum distance, but rapid and close (within 30 m) approaches induced avoidance behavior. Bowhead whales, Balaena mysticetus, and right whales Eubalaena glacialis and E. australis fled from vessels that were noisy or approached rapidly, but could often be approached by quiet and slowly moving boats. The same reaction was noted in blue whales, Balaenoptera musculus, and fin whales, B. physalus, in the St. Lawrence estuary. In the Gulf of Maine, fin whales shortened dives and reduced the number of surface blows in the presence of boats. Off New Zealand, some sperm whale, Physeter catodon, individuals avoided outboard-powered whale-watching boats when approached to within 2 km by reducing the number of blows per surfacing, the intervals between blows, and the duration of surfacing. They also changed dive patterns and surface movements. These studies identified many factors that influence the response of cetaceans to whale watching. They include season, differences in responses due to status (residents versus transient animals), age and sex, activity prior to approach, group size, group composition, population, species, the number

460

HERlBERT HOFER AND MARION L. EAST

of vessels present, the speed, distance, and angle of approach of vessels, the type of vessel, the general amount of vessel traffic, and the history of whale watching in a particular area (Richardson et af., 1995; Evans, 1996; Janik and Thompson, 1996). Other factors include the changing habits of whale watchers, changes in regulations of whale watching, and a diversity of regulations across the globe (IFAW et al., 1996). A majority of studies have found some sort of short-term effect of boats or whale watching on the behavior of whales, whereas a minority of studies started that they did not detect any significant effects. Studies of whale watching are typically shore based (Stone et af., 1992; Janik and Thompson, 1996) or conducted from aircraft. Typically, these studies suffer from methodological problems that need to be overcome to make substantial progress in the future. Shore-based observations cover only a limited amount of space; the approach and/or presence of aircrafts as such may have significant effects on cetacean behavior (Richardson et al., 1995). A key problem is the lack of individual identification of study animals. Without individual identification, “control” observations of unknown cetaceans are less valuable than comparisons of known individuals in the presence and absence of whale-watching vessels. Experiments such as playbacks of recorded noise (Section V,D) are of limited usefulness if carried out in captivity (e.g., Thomas ef al., 1990), yet field experiments (e.g., Evans et al., 1994) are difficult to implement unless individual identification ensures adequate “control” and “treatment” observations. If “control” and “treatment” groups also have small sample sizes, lack of power (Section 111,DJ) will ensure that differences are unlikely to be recognized as significant and that conclusions from studies that find no significant effects should be considered tentative. Lack of power may also prevent a meaningful comparison of the effect of different classes of boats that may frequent a whale-watching area, or differences in vessel behavior, because sample sizes are likely to be even smaller if split between vessel categories. However, such studies are essential for understanding the impact of variations in vessel behavior, and hence for a scientific foundation of guidelines for whale watching. Moreover, no study has attempted as yet to quantify potential fitness consequences of whale watching and little is known about the physiological consequences of anthropogenic disturbance of whales or the issue of vessel noise masking acoustic channels important for intraspecific communication or antipredator vigilance (Richardson et al., 1995). Reviewers of the impact of anthropogenic disturbance on cetaceans frequently suggest that apparent stability or increases in population size in a particular geographical area indicate that this population must be tolerating the disturbance well and hence individuals are unlikely to experience detrimental fitness conse-

BIOLOGICAL CONSERVATION AND STRESS

46 1

quences. In the absence of data on individual fitness, this conclusion is usually not justified for several reasons: 1. Because of their slow reproductive rates, many whale populations are still believed to be recovering from excessive hunting earlier this century (Evans, 1987). In an expanding population, a significant but moderate negative effect from anthropogenic disturbance may manifest itself as a reduction in population growth, rather than a population decline. Observed population stability or increases as such are therefore unreliable indicators of population effects of whale watching. 2. Significant anthropogenic disturbance might reduce habitat quality. At overall high population densities low-quality habitats would contain larger populations than expected, masking any fitness effects of a decline in habitat quality (Section IV,D,2). Hence, population density or size can be a misleading indicator of habitat quality, particularly of low-quality habitats. 3. Availability, size, and distance of alternative high-quality habitats and their degree of occupancy are usually unknown but strongly influence habitat choice and duration of stay (Fretwell, 1972; Bernstein et af., 1991; Sutherland, 1996).

4. Population estimates based on censuses often have large confidence limits, making it unlikely that changes in population size can be identified unless they are substantial. Caveats also apply to the issues of reactions interpreted as benign and to behavioral changes after repeated exposure to whale watching. Approaches of boats by cetaceans are often viewed as a benign reaction, yet it might be an aggressive response to defend resources (prey or potential mates), which might also be disturbed by vessel presence, and be associated with an interruption of essential activities and a physiological stress response. Secondary detrimental effects of approaches (e.g., incidental pollution) may also be possible ( Janik and Thompson, 1996). Behavioral changes might include habituation, a gradual decline in response to continuous or discrete repeated stimuli (yet still be associated with detrimental fitness consequences; see Section II,B and following discussion), and sensitization, an increase in responsiveness under the same conditions (see any textbook on animal behavior). Clear evidence of habituation or sensitization is scarce because of the lack of observations of individually identified animals in the presence of repeated or continuous stimuli (Richardson et al., 1995; Evans, 1996). Another problem could be the continued illegal offtake of protected species such as humpbacks as recently as 1993 (Baker and Palumbi, 1994).

462

HERIBERT HOFER AND MARION L. EAST

Populations that continue to experience hunting are likely to react differently to approaching whale-watching vessels than nonhunted populations (Richardson et al., 1995; see Section V,C). It will be important to establish the history of hunting of a population to ensure that adverse reactions t o the presence of boats are attributed to the appropriate cause. Some of these problems were recognized by a recent workshop on scientific aspects of managing whale watching, which emphasized the importance of (1) conducting carefully controlled long-term studies; (2) using experiments to assess the impact of whale watching on the behavioral response of cetaceans; (3) improving methods to assess the hormonal, immunological, and physiological response to vessel presence; and (4) assessing the fitness consequences of whale watching (IFAW et al., 1996). Whale watching is a case where intensive efforts are under way to establish internationally standardized guidelines and regulations even though scientific information on the potential fitness consequences of whale watching is not yet available (IFAW, 1996; IFAW et af., 1996).

2. Case Study 4: Antarctic Penguins and Tourism Increasing numbers of tourists visit breeding colonies of several species of Antarctic penguins every year ever since shipborne tourism to the Antarctic began in 1958 (Enzenbacher, 1993). Concern about the potential impact of human presence on breeding performance of penguins increased when it became clear that colonies in the vicinity of Antarctic research stations declined during the operation of these stations and recovered when operations were terminated (Culik and Wilson, 1991b). Wilson et al. (1991) used external electrocardiogram recording devices or surgically implanted heart rate transmitters to assess the impact of disturbance on heart rate in AdClie penguins living in a colony close to a research station. Significant heart rate increases (50% or more) were recorded from breeding penguins exposed to a solitary human approaching nests or a major penguin commuter pathway, and to low-flying aircraft approaching the colony, even when the penguins showed no obvious behavioral signs of excitement. However, penguins also responded behaviorally by deviating 70 m from the commuting path for several hours after the human had left the area, fleeing from the nest, or refraining from returning to their nest (Wilson et al., 1991). Three days of helicopter exposure caused 80% of penguins to abandon their nests (Wilson et af., 1991). Culik and Wilson (1991b) concluded that “tourism does adversely affect breeding penguins, almost irrespective of how ‘well-behaved’ the tourists are.” Nimon et al. (1994,1995) pointed out that the handling procedures used by Wilson et af. (1991) may have caused the study animals to be negatively predisposed toward humans and that the records of heart rate by Wilson

BIOLOGICAL CONSERVATION AND STRESS

463

et al. (1991) reflected the effect of handling rather than the effect of disturbance by the approaching human or aircraft. They placed an artificial egg with an infrared heartbeat sensor in nests incubated by gentoo penguins in order to record changes in heart rate caused by disturbance. They expressed the hope that their minimal handling procedure would not influence records. Their preliminary results showed that heart rates were not significantly increased when a single experimental visitor approached slowly and in a nonobtrusive fashion, and they concluded that tourism did not necessarily cause a disturbance. In reply, Culik and Wilson (1995) pointed out that (1) the heartbeat records of Nimon et al. (1994,1995) were taken during incubation, whereas those of Wilson et ul. (1991) were obtained from penguins guarding crkches; and (2) the reaction of incubating birds are minimal compared with reactions during other stages of the reproductive period and thus cannot be extrapolated to these other stages. Also, as tourists come in batches rather than as single visitors, the experimental setup by Nimon et al. (1994, 1995) was unlikely to represent the situation of tourist visitors (but may well represent the situation of researchers studying penguin colonies). They further pointed out that there was considerable intra- and interspecific variation in responses to disturbance and that the strength of response (e.g., in terms of flight distance) may depend on the general level of disturbance experienced by a colony. Thus, disturbance of penguin colonies sometimes had significant detrimental fitness consequences, behavioral responses varied between reproductive stages and depended on the general level of disturbance, and responses varied between species. These results suggest that (1) more refined methods and long-term records are required before interspecific generalizations can be safely drawn; and (2) nondisruptive behavior of visitors is a minimal requirement to protect penguins from significant disturbance. 3. Experimental Studies of the Effects of Tourism and Leisure Activities Several experimental studies have demonstrated that disturbance associated with tourism may change heart rate, increase energy budgets, and decrease reproductive performance (case studies 1, 3, and 4; Sections IV,B-D; see following discussion). However, the potential disturbance caused by several important types of tourism, notably wildlife viewing on safaris in eastern and southern Africa, has not been studied in this way. Disturbance of oystercatchers, Huematopus ostrulegus, by people, artificial kites, and single-engine planes increased heart rate in incubating individuals (Huppop and Hagen, 1990). Disturbance of ptarmigans, Lagopus mutus, by hikers near incubating birds decreased heart rate associated with freezing behavior (Ingold et al., 1992). Low-level disturbance during the

464

HERIBERT HOFER A N D MARION L. EAST

incubation and nestling phases increased (1) time and energy expenditure of marsh harrier, Circus aeruginosus, parents on reproductive and nonreproductive activities, and (2) plasma urea concentrations in nestlings, suggesting a detrimental effect on lifetime reproductive success (Fernandez and Azkona, 1993). Madsen (1995) reported on the results of a “natural” experiment, in which farmers either disturbed or did not disturb grazing pinkfooted geese, on their subsequent body condition and breeding success. Disturbed geese had a poorer body condition and were significantly less likely to breed successfully. The presence and behavior of tourists at nesting beaches of green turtles, Chelonia rnydas, reduced the arrival of females by a third but did not influence the proportion of successful nesting, incomplete nesting, or no nesting attempts (Jackson and Lopez, 1994). Recreational disturbance of winter denning sites of American black bears, Ursus arnericanus, resulted in the abandonment of dens and cubs and a delay in entry to hibernation (Goodrich and Berger, 1994). Experimental harassment of mule deer by all-terrain vehicles induced reproductive pauses in the subsequent breeding season (Yarmoloy et ul., 1988). Does wildlife habituate to disturbance? In many cases it does, as evidenced by a gradual decline in the behavioral response of disturbed wildlife to groups of tourists (e.g., Van Heezik and Seddon, 1990). However, as discussed in Section II,B, disturbance may continue to have detrimental fitness consequences even when animals are habituated. Habituation cannot be expected to occur under all circumstances, and is less likely to occur if the disturbance involves close approaches or unusual or unpredictable events (Huppop and Hagen, 1990). For instance, bighorn sheep, red deer, and marmots, Marrnotu rnurrnotu, were less responsive to people on major roads or hiking trails than to people encountered off-trial or with a domestic dog as companion (Schultz and Bailey, 1978;MacArthur et al., 1982;Mainini et al., 1993). However, the experiments reported in these studies in some cases cannot exclude the possibility that individuals less sensitive to disturbance, for example, individuals prepared to take higher risks because they are of below-average quality, settle closer to major roads or hiking trails. Some studies demonstrate that animals did not habituate, even after years of exposure, o r may even become sensitized by repeated disturbance events. For instance, paragliding in the Swiss Alps initiated flight responses in chamois, Rupicupru rupicupru, and Capricorn, Capra ibex, and led to changes in habitat utilization by these species. Experimentally disturbed animals did not show any evidence of habituation (Schnidrig et al., 1992; Ingold et al., 1996), although differences in the routine of experimental and usual paragliders may have prevented habituation to occur. Mountain goats, Orearnnos urnericus, showed substantial individual differences in response

BIOLOGICAL CONSERVATION A N D STRESS

465

to, but did not habituate to, intense industrial exploration activity (Foster and Rahs, 1983). D o successful education and awareness campaigns always reduce disturbance of wildlife? A recent study has suggested that such campaigns may also initiate new forms of disturbance. Initially, water-based tourism posed the greatest threat to the survival of manatees, Trichechus manatus, because of mortality and injuries from boat propellers (Shackley, 1992). After a highly successful public-awareness campaign, new forms of tourism developed, including helicopter flights, canoeing, and SCUBA diving, creating new sources of disturbance. The long-term consequences of these forms of disturbance are not yet known. C. HUNTING Hunting by people is a form of human disturbance that is thought to closely resemble predation and may sometimes even operate synergistically with natural predation. For instance, Schauer and Murphy (1996) observed that the rate of egg loss from nests of colonial cliff-nesting common murres was high on days when human hunters discharged firearms and shot adults on or near cliffs. Flushed adults often accidentally dislodged eggs, and abandoned eggs were taken by natural egg predators such as glaucous gulls, Larus hyperboreus. Potentially stressful impacts of hunting may be relevant to conservation efforts for two reasons. Hunting is often considered a recreational activity, and thus should be considered in the context of disturbance of wildlife by recreational activities. A second aspect is that one school of thought proposes that conservation will work only if potential conflicts between local communities and conservation activities can be minimized. According to this idea, the minimization of such conflicts requires the economic exploitation of wildlife because consumptive exploitation, including sports hunting, pays local communities, and hence ensures the conservation of wildlife populations (see Taylor and Dunstone, 1996a). This approach encourages hunting as a method to advance conservation. In both contexts it might be useful to know how stressful hunting is. Here we consider three aspects of hunting that may have detrimental fitness consequences: the physiological consequences of pursuit, modification of behavior between hunting and nonhunting seasons, or hunting and nonhunting areas, and the consequences of crippling. 1.

Physiological and Fitness Consequences of Pursuit

Some forms of hunting, for instance riding with hounds to hunt red foxes, Vulpes vulpes, red deer, or hares, Lepus europaeus, consist of driving the target animals over considerable distances. We call this pursuit hunting.

466

HERIBERT HOFER AND MARION L. EAST

Proponents of pursuit sports hunting frequently argue that there is little difference between the pursuit of a red deer by natural predators or the pursuit of a red deer by a staghunt. However, there has been no quantitative investigation that compares the “stress” response of the target animal to pursuits by predators with the response to pursuits by hunts, and there are comparatively few descriptive or experimental studies that have explored the physiological consequences of pursuit hunting (see following discussion). Such measurements should include the hormonal, immunological, and anatomical response to hunts as a function of the duration of a pursuit, the distance covered and average and maximum travel speeds, and subsequent fitness consequences (e.g., delayed mortality). The hypothesized parallels between pursuits by natural predators and hunters fail for such practices as badger, Meles meles, “digging” using terriers in the United Kingdom. Again, data are few on details of such hunting practices (for an exception, see Griffiths, 1994) and none are available to assess the physiological response or its fitness consequences for the hunted animals. In an experimental study, Harlow et al. (1992) assessed the consequences of five to six pursuit chases of cougar, Felis concofor, over the course of one hunting season by comparing the adrenocortical response of individuals to ACTH stimulation before and after chases at the beginning and at the end of the hunting season. These pursuit chases were sufficient to permanently alter the physiological response of the adrenals. It is unclear, however, whether such changes enhanced or reduced the ability to cope with stress and whether they would have detrimental fitness consequences. Measurements of the adrenocortical, immunological, or anatomical response to pursuits have not been undertaken until recently, although they would have contributed to a more scientifically informed debate on the welfare implications of sports hunts in the United Kingdom (Taylor and Dunstone, 1996b). A recent 2-year study compared the adrenocortical and anatomical response of red deer that were shot, involved in car traffic accidents, pursued by sports hunts. Red deer that were pursued during the course of normal staghunts were chased over average distances of 19 km and at the end of the chase had no blood sugar left, substantial damage to blood cells and muscles (reminiscent of capture myopathy), and significantly higher levels of cortisol and beta endorphins than deer that were killed in other ways (Bateson, 1997). This suggests that pursuits may cause substantial damage and that deer that escaped at the end of the chase may die from capture myopathy or other debilitating consequences of the pursuit. Observations on several species indicate that hunts of whales are usually associated with prolonged struggles of the pursued individual and wounded animals may take up to 1 hour to finally succumb (Kestin, 1995). A descriptive study of freshly killed fin whales and sei whales, Baluenoptera borealis,

BIOLOGICAL CONSERVATION A N D STRESS

467

tried to assess whether the pursuit of whales during whale-catching operations may cause “heat stress” in terms of raised core body temperatures. It found no evidence for raised core temperatures and concluded that either high propulsive efficiency or high thermoregulatory capacity prevented heat stress in whales after an intensive pursuit (Brodie and Paasche, 1985). However, as details of pursuit intensity were not available, and temperature measurements took place some time after death and were restricted to a few sites, generalizations from this study are probably unwarranted without further evidence. Pursuits of individuals for the purpose of capture may result in capture myopathy (Section V1,C). Species probably vary substantially in their tolerance of pursuits before capture myopathy occurs. In the dugong, Dugong dugon, even short pursuit distances by traditional hunting methods are likely to lead to severe physiological problems including mortality from capture myopathy (Anderson, 1981). Continuous pursuit of African lions, Panthera leo, by cattle farmers with rifles, dogs, and gin traps over 2 days led to severe pathological changes (as subsequently revealed by autopsy), and capture myopathy with subsequent mortality when the individual was immobilized on the third day of the hunt (Joubert and Stander, 1990). 2. Effects on Behavior and Habitat Choice

It has been repeatedly argued that hunting has a pronounced influence on activity budgets and the reaction of wildlife to people. One hypothesis considers hunting as just one of many forms of human disturbance and predicts that animals should react to hunter presence and activity in a manner similar to that for other forms of human disturbance. A second hypothesis, however, explicitly distinguishes hunting from other forms of human disturbance and argues that in areas with hunting or during periods of hunting wildlife tends to be shy and is more likely to flee from any kind of approaching human being, particularly during daytime. By contrast, in areas without hunting or during periods of no hunting wildlife habituates quickly to human presence and can habituate even to high densities of visitors involved in recreational activities. Some observational, experimental, and theoretical studies have attempted to quantify the impact of disturbance caused by hunters and obtained variable results, but so far no rigorous experimental tests have been carried out that were designed to distinguish between these hypotheses. There are several reports of marine mammals that alter their response to other forms of anthropogenic disturbance if the population is subjected to hunting (Richardson et al., 1995). Grizzly bear, Ursus arctos, females avoided private lands on which they would be hunted (Mace et aL, 1996). Madsen (1995) reported on the progress of a large-scale experiment that

468

HERIBERT HOFER A N D MARION L. EAST

aims to assess whether hunting is responsible for habitat shifts and other behavioral changes in waterfowl. Every year experimental refuge areas (areas where hunting was prohibited for that year) were set up in a new location and the distribution of individuals over hunting and nonhunting areas was subsequently observed. Both species that were hunted and species that were legally protected significantly preferred the refuge areas. In an observational study, Jeppesen (1987) compared the behavior of a population of red deer between seasons of no hunting, moderate hunting, and intensive hunting. During hunting periods red deer moved greater distances and enlarged their home ranges. During an intensive hunting period high tourist numbers led to deer completely abandoning one type of plantation, a major habitat shift by the deer. In previous years high tourist numbers were associated with only moderate hunting pressure and habitat shifts were not observed. Jeppesen (1987) argued that this singled out hunting as a particularly severe form of disturbance and that it was the increase in hunting pressure that had led t o habitat shifts by the deer, but he was unable to conclusively identify cause and effect in this case. Skogland and Grovan (1988) demonstrated that foraging and aggregation behaviors of reindeer, Rangifer tarandus, were significantly affected by hunting and asked whether such behavioral changes depended on the initial body condition and the foraging needs of individual reindeer (Section IV,B,S). They found that during the hunting season, well-fed animals aggregated into larger groups, spent more time alert, and foraged less than during the nonhunting period. That is, they pursued a risk-minimization tactic. Poorly fed animals, however, pursued a risky nutrient maximization strategy by moving more and losing more body mass during the hunting season than well-fed animals. Frederick et al. (1987) used a stochastic simulation model to analyze the effect of increasing hunting pressure on population size and emigration rates of the lesser snow goose, Chen caerulescens. They found that direct mortality from hunting had a smaller effect on population size than the reduced energy gains and increased emigration rates of geese caused by disturbance during feeding by hunters. Several studies have found no effect of hunting on behavior. In an experimental study, Olsson et al. (1996) compared the movements of radiocollared willow grouse, Lagopus lagopus, in heavily hunted areas with those in areas in which hunting was prohibited. No significant differences in movement distances or rates, or in the likelihood of emigration from the study area were detected. Olsson et al. (1996) interpreted this as evidence for a predator-avoidance strategy that relies on utilizing a familiar area with known escape sites. Kernohan et al. (1996) asked whether 24-hour habitat use can be predicted from diurnal habitat use in white-tailed deer during seasons of hunting and no hunting, and found that this was the case

BIOLOGICAL CONSERVATION AND STRESS

469

for both types of seasons. This does not, however, preclude the possibility that changes in activity patterns did occur between the hunting and no hunting seasons, but if so, they did not affect the diurnal behavior of deer.

3.

The Consequences of Crippling

Hunting by shooting or snaring may also cause mutilations that permanently cripple the victim and generally make life more difficult. There are few data on the proportion of animals crippled as a result of hunting in wildlife populations, or the fitness consequences of crippling. Crippling such as amputations and persistent injuries might increase energetic expenditure during foraging trips or decrease foraging skills, increasing the chance of mortality or decreasing the ability to successfully raise young. Wheeler et ul. (1984) estimated that during periods of high-intensity shooting of ducks in autumn in an American marsh 24-32% of ducks were crippled, but they did not record fitness consequences of crippling. In spotted hyenas, Crocuta crocuta, the chance of escaping from snares set by poachers to capture herbivores was 25-64% (Hofer et al., 1993). Individuals that successfully escaped from snare locations by biting through the tethering wire retained snares on their bodies for variable periods. Snare wounds may become infected and persist until they eventually kill individuals, skin may grow over the snare and the snare may become a permanent feature of the animal, or snare wounds lead to an amputation of parts of an extremity. In several lactating females, snare wounds and amputations led to an increase in suckling intervals and the loss of the litter (H. Hofer and M. L. East, unpublished data). Clearly, such impairments can have detrimental fitness consequences (see Hofer et al., 1996). In forest chimpanzees, Pun troglodytes, 11 out of 34 individuals (32%)suffered injuries from snares set for bushbuck, Trugeluphus scriptus crippling or amputating one or both hands a n d o r feet. Observations of foraging groups showed that disabled individuals were well integrated into the groups (Quiatt et ul., 1994). Chimpanzees suffering from snaring injuries were typically lower ranking and fed at greater heights but achieved the same feeding rate in fig trees (Smith, undated). Fitness consequences of these injuries have not yet been measured in chimpanzees.

D. Noise There is an increasing recognition that environmental noise, particularly noise generated by human activities, is harmful not only to humans (Clark, 1992) but also to wildlife (Fletcher and Busnel, 1978). Sources of environmental anthropogenic noise are flying devices, including hot-air balloons, low-flying military and civilian aircraft and helicopters, road, boat, and vessel traffic, terrestrial and marine seismic exploration and geophysical

470

HERIBERT HOFER AND MARION L. EAST

surveys, explosions, marine sonars, terrestrial and marine resource extraction including mining, oil and gas drilling and production, dredging and construction, and recreational activities including wildlife viewing and the chatter emanating from safari vehicles. Noise can influence animals over distances of up to 30 km (Cosens and Dueck, 1993) and impact the entire animal community. Ideally, carefully designed and executed experimental studies on wildlife assess the response to and identify the fitness consequences of anthropogenic noise, and separate the effect of noise as such from other aspects (e.g., visual cues) of disturbance. However, comparatively few experimental studies have considered these issues and most information comes from observational studies. These either provide evidence of the detrimental effects of noise or demonstrate that a reduction of noise levels had a positive effect, but often cannot separate the influence of noise from that of other factors. For instance, the decrease in breeding density of hazel hen, Bonasia bonasia, in partly urbanized areas was suspected to be due to a permanent increase in background noise level, which may have reduced the chance of individuals to hear alarm calls, thereby increasing mortality from predation (Scherzinger, 1979), but appropriate experimental tests were not carried out. Examples of experimental evidence for the beneficial effects of noise reduction come from studies in captivity. Quiet handling, based on an understanding of behavior, reduced excitement and the incidence of injuries in domestic cattle (Grandin, 1987). Fear-related behavior and heart rate of red deer housed indoors were reduced by sound-proofing housing facilities (Price et al., 1993). Examples of experimental evidence for the detrimental effects of noise include short startle noises used to induce “prenatal stress” in captive, nonhuman, pregnant, female primates after removal from their home cage. These studies produced significant effects of prenatal stress on the behavior and immunocompetence of infants (Lubach et af., 1996; Schneider 1992a,b,c; Clarke & Schneider, 1993; Clarke et al., 1994), but did not separate the effect of removing the pregnant female from her home cage from that of noise itself. The hormonal or physiological response of wildlife to noise is not well researched. The effect of noise on blood pressure and sleep (known to be detrimentally affected in humans, Berglund et al., 1990) has to our knowledge not yet been studied, nor has the effect of human disturbance and noise on heart rate of marine mammals (Richardson et al., 1995). Noise from car traffic increased heart rate and thus may affect birds breeding close to highways (Helb and Huppop, 1991). In free-ranging mountain sheep, low-flying aircraft and vehicle traffic produced no overt behavioral response and a modest increase in heart rate (modest compared to the increase observed after the animal was handled for instrumenation) (Mac-

BIOLOGICAL CONSERVATION A N D STRESS

47 1

Arthur et al., 1982, 1986). Harrington and Veitch (1991) noted that the sound rather than the visual appearance of low-flyingjet aircraft was responsible for startling Alaskan caribou, Rangifer tarandus. In a partly experimental study, Harrington and Veitch (1992) demonstrated a significant negative correlation between the frequency of exposure of Alaskan caribou to low-level jet overflights during the calving and postcalving season and subsequent calf survival. In an observational study of 43 species of woodland-breeding birds, Reijnen et al. (1995) identified noise emanating from car traffic along main roads as the key variable that depressed breeding densities over distances of 100-1500 m away from main roads, when the results were controlled for other potentially confounding factors. This confirmed a single-species study of the willow warbler, Phylloscopus trochilus, that suggested that the noise emanating from car traffic along main roads prevented males close to a highway from attracting or keeping females (Reijnen and Foppen, 1994); as a consequence these males moved away from the road during the following breeding season. Short-term seismic exploration and timber harvest in an area with few permanent developments and low levels of vehicular traffic caused little overt response and minimal changes in habitat utilization by grizzly bears (McLellan and Shackleton (1988, 1989). However, in an area with a higher intensity of vehicle traffic, Mace et al. (1996) found that grizzly bears avoided buffer zones of 500-m width surrounding roads with a traffic intensity of more than 10 vehicles per day, whereas their response was neutral to roads with a traffic intensity below that. The literature on the effect of anthropogenic noise on marine animals has been reviewed by Myrberg (1990), Richardson et al. (1995), and Evans (1996). Most studies recorded the behavioral response of mammals and fish to offshore petroleum exploration and production, using observations and playback experiments. Intensity of response depended on noise level, activity at the time of exposure to human-made noise, and prior experience. Where such studies combined playbacks with measurements of physiological variables, sample sizes are so low that the power (Section II1,DS) to identify significant effects is very low (e.g., Thomas et al., 1990). However, few quantitative data are available on the hearing ability of most marine animals, their physiological response to sounds of different intensities, frequencies, and durations, or the fitness consequences of such noise (Richardson et al., 1995). Some evidence on detrimental fitness consequences comes from several species of finfish and shellfish where high intensity of anthropogenic noise caused abnormal growth and reproductive processes (Myrberg, 1990). We conclude that noise is currently suspected to be a potential stressor, but there is little factual evidence on its impact on animals, or on how noise management can improve the success of conservation activities.

472

HERIBERT HOFER AND MARION L. EAST

E. CLIMATIC WARMING Climatic warming is likely to have a pronounced effect on many populations and species (Peters and Lovejoy, 1992; Kareiva et al., 1993) because (1) temperature fluctuations as well as the mean ambient temperature are key factors that determine the survival of individuals (Hoffmann and Parsons, 1991), and (2) climatic warming is also associated with other forms of environmental stress, particularly atmospheric pollutants, that are likely A rigorous experimental to operate in a synergistic fashion (Parsons, 1990~). test of the impact of global warming on fitness is not feasible, so that predictions of the impact of global warming frequently rely on measurements of temperature and pollutant tolerance in single species or the modeling of extinction probabilities of particular populations in response to environmental change. Of particular value will be observations of species that are at the top of the food chain in an ecosystem (Stirling and Derocher, 1993), and of the changes in the structure of interactions in community food webs (Bodini et af., 1994). Harsh temperature conditions may have a profound influence on the survival of populations or whole species (case study 1) because temperatures may simply be too hot or cold to ensure individual survival. On a more moderate level, harsh temperatures may be responsible for a decline in immunocompetence (e.g., Lamontagne et af.,1989) or fertility (Krebs and Loeschcke, 1994). Experimental evidence is accumulating that changes in average temperatures of as little as 1-2°C may be sufficient to cause population or species extinction (Parsons 1989b; Baur and Baur, 1993), and observational evidence indicates that short-term regional climatic changes such as the El Niiio of the southern Pacific Ocean initiate profound changes in biological communities (Arntz and Fahrbach, 1991; Trillmich and Ono, 1992). For conservation purposes, the key question will be whether populations can tolerate climatic changes or whether adaptive evolution in response to climatic change can be quick enough. There is some evidence from genetic studies suggesting that fast evolution of stress tolerance is possible (Hoffmann and Parsons, 1991), but the available evidence is insufficient to predict the outcome of climatic change on specific populations or species (Hoffmann and Blows, 1993). Genetic studies have demonstrated that there is a genetic basis for heat resistance, that some populations improve stress resistance by increasing the plasticity of phenotypes, and that there is variation in the genetic basis of heat resistance between populations of a species (Hoffmann and Blows, 1993). The study of the evolutionary genetics of stress resistance in marginal populations will help answer questions about the impact of climatic warming (Parsons, 1990~).

BIOLOGICAL CONSERVATION AND STRESS

473

VI. CONSERVATION RESEARCH A N D MANAGEMENT AS STRESSORS ACTIVITIES Both research and management activities may be a source of stress. Failure to recognize this problem often reduces the value of a study or a conservation activity, makes the subjects experience more hardship than necessary, and prevents the design and implementation of procedures that minimize the occurrence, magnitude, and consequences of stress. Even if ecological field work is not explicitly designed as a manipulative experiment, field studies may routinely cause disturbance because of the presence of an observer or specific interventions involving capture, handling, and marking. Whether an intervention causes discomfort, distress, or a reduction in fitness ought to be the subject of careful scientific evaluation. The widespread tacit assumption that interventions are always benign, or least do not influence the results of a study, is not justified, as careful experiments have demonstrated (see following discussion). However, interventions do not necessarily have detrimental consequences, and as there is currently no general theory available that predicts the circumstances under which interventions may be harmful, the impact of interventions ought to be evaluated on a case-by-case basis. In most studies reviewed later in this chapter that report the absence of an intervention effect, the power of tests (Section II,D,5) was not evaluated, and if it was evaluated it was low, so the failure to reject the null hypothesis of intervention having no detrimental effects must be considered tentative in each case. A.

LITTLEATTENTION HASBEENPAIDTO CONSERVATION ACTIVITIES

THE

IMPACT OF INTERVENTION

AND

Many field and laboratory studies and conservation activities use various forms of intervention in field experiments (capture, handling, blood samping, radio-collaring, vaccination, and manipulation of populations or aspects of their environment). Compared with the number of cases where such procedures are employed, it is surprising how little attention has been paid to the ethical and conservation implications of interventions and field experiments (Cuthill, 1991; Putman, 1995). Handling techniques are known to provoke a strong physiological response, particularly in the case of nonacclimated animals (Gartner et al., 1980; Pottinger and Calder, 1995), that might bias or invalidate physiological results. Despite this, little attention has been paid to such biases in several disciplines, including physiology and toxicology (Rowan, 1990; Pottinger and Calder, 1995). Stress-sensitive physiological data from blood samples were analyzed in 58 out of the 397 publications on macaques surveyed by Reinhardt (1991a), yet 81% of the

474

HERIBERT HOFER AND MARION L. EAST

studies did not provide any information as to how subjects were caught or immobilized for blood sampling. Common practice suggests that the animals were physically restrained with squeeze-backs or forced with fear-inducing techniques to leave their home areas and enter a transport cage. Such methods of enforced restraint result in significantly increased adrenal activity and significant changes in other physiological parameters (Reinhardt et al., 1995), suggesting that the effects of handling may have biased the results of many of these studies. THE TARGET POPULATION AND BIAS B. CENSUSING MAYDISTURB CENSUSRESULTS

Most studies that estimate population size or habitat use assume that the censusing process does not disturb populations and thereby influence censusing results. Several studies have tried to experimentally test this assumption. Dufour et al. (1993) tested the idea that baited traps preferentially attract mallard individuals in bad condition and that such an effect may change with sample size. They found that there was some trap selectivity but that it was unlikely to vary with sample size. Bleich et al. (1994) investigated the effect of helicopter surveys on movements of mountain sheep, Ovis canadensis, across habitats and sampling blocks. Surveys increased movements between sampling blocks and altered habitat use, potentially increasing “nutritional stress” or susceptibility to predation. Bleich et al. (1994) suggest that with significant disturbance effects censuses are likely to violate fundamental assumptions of population estimators. Mallet et al. (1987) assessed disturbance effects from netting or handling of butterflies Heliconius sp. in mark-recapture studies. They concluded that disturbance caused mark-recapture estimates to be so inaccurate that the capture of almost all individuals of a population is required, making the use of the Lincoln Index unnecessary. Similarly, cane toads, Bufo marinus, were highly sensitive to disturbance by trapping or handling, reducing the chance of recapture, and making mark-recapture estimates unreliable (Lampo and Bayliss, 1996). Other forms of population monitoring require visits to breeding or nesting sites. Such visits may or may not influence breeding success. Checking of nesting burrows of a Canadian population of Atlantic puffins, Fratercula arctica, by researchers caused birds to abandon nests and reduced breeding success by more than a third (Rodway et al., 1996). Visits at different rates by researchers to nesting sites of rock ptarmigans, Lagopus mucus, however, did not influence predation rate, clutch size, nesting, fledging, or hatching success (Cotter and Gratto, 1995). These experimental studies suggest that population censusing may sometimes stress target populations and that studies should assess the potential

BIOLOGICAL CONSERVATION A N D STRESS

475

influence of the censusing process on results. As population size estimates are a vital ingredient of conservation activities, potential biases in these estimates caused by the censusing procedure may have profound consequences. A N D CONSERVATION ACTIVITIES ARE C. INTERVENTIONS SOMETIMES STRESSORS

The possibility that interventions and conservation activities may themselves be the source of stress is frequently ignored. Retrospective, descriptive comparisons of handled versus unhandled individuals are often the only way to analyze data in the observational studies when there is a suspicion that handling might have a significant impact on fitness. Such studies may provide important hints and suggest factors responsible for handling effects. However, descriptive retrospective studies sometimes use small sample sizes that imply low statistical power and make it unlikely that a significant difference between handled and unhandled animals could be identified if there really was one; other caveats may apply (Section 111,D). To overcome these problems, experimental studies of intervention are required in which subjects are assigned randomly to experimental and control groups (Section 111,DJ). Observational studies may sometimes meet these criteria if it can be argued that the selection of individuals for interventions had proceeded in a random way. The following sections emphasize results from careful experiments that in many cases measured fitness consequences of interventions. When considering the results of these studies it is important to remember that some measures of fitness are shortterm ones that may not necessarily predict long-term effects.

Capture, Measurements and Palpation, and Blood Sampling Because a typical handling event usually consists of capture, restraint, examination, measurements, and blood sampling of an animal, these components of handling will be considered together. Capture, measurements, and blood sampling may sometimes cause a substantial physiological response. An experimental study of captive domestic geese demonstrated that weighing, injecting, and blood sampling disrupted the acid-base balance and caused a dramatic increase in the level of humoral indexes of stress (catecholamines, corticosterone, and lactate) within 2 min (Le Maho et al., 1992). Routine restraint for 1 h, handling, and examination in captivity (including measuring and weighing) provoked substantial increases in the level of glucocorticoids in several species of bats (Widmaier et al., 1994). “Unpleasant” handling of captive domestic pigs alone or in groups reduced growth rate and feed conversion efficiency and led to higher glucocorticoid levels I.

476

HERIBERT HOFER AND MARION L. EAST

in individually handled animals (Hemsworth and Barnett, 1991). Handling captive rats and domestic cattle for blood sampling when animals were not acclimated to the procedure raised plasma concentrations of cortisol and other hormones, substantially changed heart rate, and depressed tonic luteinizing hormone secretion (Gartner et al., 1980; Echternkamp, 1984). The potential fitness consequences of capture, basic body measurements, and blood sampling has been repeatedly examined by retrospective descriptive studies in many species. Research activities that included handling and ringing are now considered a major factor responsible for high mortality and the decline in population size in many European species of cave bats in the 1970s (Gaisler et al., 1981). In birds, where handling effects have been most intensively studied, detrimental fitness consequences of interventions were sometimes recorded in retrospective studies; these, and other studies that found no detrimental consequences, were reviewed by the American Ornithologists’ Union (1988) and Kania (1992). Several experimental studies show that capture, measurements, and blood sampling had no fitness consequences in the wild. In the red-winged blackbird, Agelaius phoeniceus, capture and blood sampling did not affect migratory behavior, annual return rates, chance of territory loss, or reproductive success in the field and did not affect mass changes in captive birds (Hoysak and Weatherhead, 1991). In the endangered Chatham Island black robin, Petroica traversi, capture in traps, measuring, and blood sampling had no detrimental effect on behavior and adult survival for one year (Ardern et al., 1994). In the red-cockaded woodpecker, Picoides borealis, sampling blood and feather pulp from nestlings had no effect on nestling survival (Stangel and Lennartz, 1988). Capture, handling, and blood sampling of white-crowned sparrows, Zonotrichia leucophrys, during the breeding season did not affect survival, migration, annual return rates, chance of territory loss, or reproductive success (Wingfield and Farner, 1976). The behavioral ecology literature contains many examples in which manipulations of study animals were carried out, and controls that assessed the effects of handling as such found no effects on fitness. Some experimental studies have also found significant effects of handling on fitness. In the semipalmated sandpiper, Calidris pusilla, the spotted sandpiper, Actitis macularia, the red-necked phalarope, Phalaropus lobatus, and Wilson’s phalarope, Phalaropus tricolor, capture and blood sampling during different stages of the breeding season caused little mortality, but behavioral responses with fitness consequences varied across reproductive stages and among species with different mating systems (Colwell et al., 1988). Parental desertion was least common in uniparental species. Wilson’s phalarope was more likely to desert when captured during the laying stage. Incubating semipalmated sandpipers were more likely to desert if both

BIOLOGICAL CONSERVATION AND STRESS

477

parents were bled rather than if one or neither of the adults was bled. Of the previously mentioned experimental studies, only Colwell et al. (1988), Hoysak and Weatherhead (1991), and Ardern et al. (1994) used experiments to study both changes in behavior and potential fitness consequences of capture, handling, and blood sampling in birds. As previously mentioned, a well-known detrimental consequence of capture and handling of animals is the phenomenon of capture myopathy (Chalmers and Barrett, 1982). Death due to capture myopathy has been described for birds (Spraker et al., 1987; Dabbert and Powell, 1993), marine mammals (Anderson, 198l), and several orders of terrestrial mammals including marsupials (Shepherd, 1986), ungulates (Kock et al., 1987a; Beringer et al., 1996), primates (Harthoorn, 1976), and carnivores (Joubert and Stander, 1990). Sometimes the animal may survive but suffer from chronic pathological consequences for months to years (Kock et al., 1987b). It is difficult to predict which species are likely to suffer from capture myopathy. Australian dugongs, Dugong dugon, are large, marine mammals that are considered to be susceptible to capture myopathy and die within hours of a pursuit (Anderson, 1981), whereas in a large sample of captured American manatees, a closely related and ecologically very similar species, there was no evidence of capture myopathy (Oshea et al., 1985). Capture myopathy can sometimes be cured or avoided (Harthoorn et al., 1974). 2.

Manual Restraint

Data on the impact of manual restraint are largely restricted to physiological responses. A comparison of the adrenocortical response to capture by physical restraint with that to chemical restraint (chemical immobilization) in 18 mammalian species suggested that manual restraint was less stressful than chemical immobilization because cortisol levels rose more substantially after chemical immobilization (Morton et al., 1995). Traditional involuntary restraint techniques in studies of nonhuman primates are an intrinsic source of distress causing fear, resulting in significantly increased adrenal activity and significant changes in a variety of other physiological parameters (Reinhardt et al., 1995). In mice, short-term restraint reduces immunocompetence in response to a herpes simplex virus infection (Bonneau et al., 1991). Short-term restraint of domestic sheep led to high increases in epinephrine, norepinephrine, cortisol, glucose, and free fatty acids (Niezgoda et al., 1993). In Wied’s black-tufted ear marmoset, Callithrix kuhli, isolation in captivity followed by short mammal restraint provoked a significant adrenocortical response, as measured by urinary cortisol (Smith and McGreer-Whitworth, 1996). The response was sensitive to subtle changes in stressor severity in a dose-dependent manner. “Gentle” handling and manual restraint for brief periods did not appear to cause chronic stress, as measured by plasma

478

HERIBERT HOFER AND MARION L. EAST

corticosterone levels, in the ball python, Python regius, or the blue-tongued skink, Tifiqua scincoides. Restraint of pythons in a container, however, resulted in short-term elevation of corticosterone but not in behavioral changes (Kreger and Mench, 1993). Manual restraint thus may produce a significant physiological response, but as yet little is known about its fitness consequences.

3. Chemical Immobilization and Anesthesia Chemical immobilization has often been mistakenly thought to cause little or no stress if an animal remains quiet and shows no behavioral signs of excitement (e.g., Creel, 1992). We have already pointed out that behavior is often unreliable as an indicator of a stressed state (Section IV,DJ). “Stress” in the sense of a significant physiological, hormonal, or immunological response to chemical immobilization results from disorientation before unconsciousness and does not require a display of behavioral excitement (Sapolsky, 1982). In the gray wolf immobilization and anesthesia caused significant reduction in heart rate and hypertension (Kreeger et al., 1987). In captive rhesus monkeys capture, injection, and disorientation prior to anesthesia were responsible for increased ACTH and cortisol concentrations (Clarke et al., 1994). Tethering, sedation, surgery, and chronic catheterization increased cortisol levels in long-tailed macaques (Crockett et al., 1993). Handling, ether vapor anesthesia, and blood sampling consistently increased serum luteinizing hormone and prolactin concentrations in male rats (Euker et al., 1975). Handling, ether anesthesia, and cardiac puncture induced significant but variable elevations of serum prolactin in female golden hamsters but not in males (Matt et al., 1983), suggesting that the response to chemical immobilization may be sex specific. Because chemical immobilization can be presented as a standardized stimulus, it has been extensively used as an experimental paradigm to understand the factors that mold the adrenocortical response (Sapolsky, 1997). There is now also considerable evidence that chemical immobilization and anesthesia may compromise the immune system, causing a deficient cellular response and a delay and depression in antibody production in response to an infection (Ozherelkov et af., 1990 Kramskaya et af., 1991; Pokhil’ko et al., 1995) or vaccination (Mayr et al., 1990), with detrimental fitness consequences in terms of increased mortality (e.g., Hansbrough et al., 1985). Even single events of anesthesia may depress cellular and antibody immunity, increasing susceptibility to infection (Thomas et al., 1982; Hansbrough et al., 1985; Felsburg et al., 1986). Such effects may be stronger and more important in wild than in captive individuals, which suggests that testing a procedure with captive animals may be of limited value. For instance, in an experimental study with coyotes, Canis Iatrans, Smith and

BIOLOGICAL CONSERVATION AND STRESS

479

Rongstad (1980) showed that glucose levels and leukocyte counts after capture, handling, immobilization, and blood sampling were significantly higher in wild compared to pen-raised or captive animals. The effects of chemical immobilization plus radio-tagging may depend on life-history stage or social competition. African wild dogs that were immobilized and radio-tagged prior to emigration survived significantly longer than individuals immobilized and radio-tagged after they had recently immigrated into a new group (Burrows et al., 1994). Dominance struggles of immigrant African wild dogs with residents or among themselves may entail injuries, while they are not overt among animals prior to dispersal. Immobilization during immigration may have exacerbated the decrease in immunological competence associated with immigration (Alberts et al., 1992; Sapolsky, 1992). 4.

Measurement of Body Temperature

Even such a simple procedure as measuring body temperature by a rectal probe may have a measurable impact on the organism. A comparison of body temperatures of laboratory rats measured by rectal probe (requiring handling of the animal) and those measured by implanted telemetric devices (recorded by remote control) usually reveal average discrepancies of approximately 1°C. A careful experimental study by Dilsaver et al. (1992) measured rectal temperature and telemetered core body temperature in the same individuals. They demonstrated that the higher body temperatures recorded by rectal probes were due to the handling necessary for the rectal probe. Other factors that influence temperature measurements in grouphoused laboratory mice include the sequence in which individuals are measured and the interval between measurements. Mice measured later in the sequence had higher temperatures and the percentage of mice showing hyperthermia increased with increasing interval between subsequent measurements (Zethof et al., 1994). 5. Surgery

Surgery, often accompanied by anesthesia, changes the metabolism of animals in that after the operation urinary and plasma cortisol levels, blood pH, partial oxygen pressure, glucose conservation, and lipolysis may be detrimentally affected (e.g., laboratory rat: Schofield et al., 1986; red fox: Kreeger et al., 1990; Pacific oyster, Crassostrea gigas: Jones et al,, 1993; long-tailed macaque: Crockett et al., 1993). Anesthesia and surgical trauma may also be responsible for immunosuppression with its potentially negative fitness consequences (e.g., Medleau et al., 1983). The fitness consequences of surgery, including muscle biopsies, for wild animals have been rarely considered or subjected to experimental investiga-

480

HERIBERT HOFER A N D MARION L. EAST

tion. In part, this is because surgery is often an emergency procedure, as in the case when animals become entangled in objects, or when it is likely that the animal will die or suffer if nothing is done. However, some procedures require routine rather than emergency surgery. These include muscle biopsies for sampling genetic material or the implantation of radio transmitters and other devices that in the course of research projects collect data from the animal’s body. In some species, surgery and radio implantation may cause substantial postsurgery mortality. Trials with implanted American river otters, Lutra canadensis, illustrated the importance of diet and extended periods of postsurgery recovery time in appropriate holding pens (Woolf et al., 1984; Hoover et al., 1985). Even under improved conditions, 30% of otters died during the postsurgery holding period. Mortality within 12 months of radio implantation of African wild dogs in Kruger National Park was more than twice as high as mortality of unhandled African wild dogs (East, 1996). These results contrast with numerous studies that found no effects of surgery and implantation of radio transmitters on subsequent survival or behavior (white-footed mouse, Peromyscus leucopus: Smith, 1980; canids: Green et al., 1985; yellow-bellied marmot, Marmota flaviventris: Van Vuren, 1989; armadillo, Dasypus novemcinctus: Herbst, 1991;ring-necked pheasant, Phasianus colchicus: Ewing et al., 1994). The effect of muscle biopsies on fitness in birds varies between species. In white-throated sparrows and Indigo buntings, Passerina cyanea, muscle biopsies did not affect body condition or survival of either overwintering or breeding individuals (Westneat, 1986; Westneat et al., 1986). In contrast, muscle biopsies of white ibises, Eudocimus albus, led to complete nest desertion (Frederick, 1986). Frederick (1986) could show that this effect was not due to the procedures of capture or handling as such, as ibises that were captured, handled, and blood sampled, but not biopsied, were not affected.

6. Tagging and Radio- Tagging For individual identification purposes, individuals are commonly tagged with plastic ear tags, neck or foot rings, freeze-branded, or hot-branded. The fitness consequences of such tags and improvements that may minimize such consequences have been reviewed repeatedly (see Putman, 1995). One key issue that is often neglected is that social relationships and social success may be permanently affected by tagging (because of the associated disturbance, human smell, etc.) or by the quality of the tags themselves. For instance, radio-tagging mule deer fawns may result in females rejecting their fawns, which in turn reduces the fawn’s chance of survival (Goldberg and Haas, 1978). Individually distinct color-banding may permanently affect

BIOLOGICAL CONSERVATION A N D STRESS

481

attractivity and reproductive success of individuals and shift female mate preferences. Males ringed with attractive color-bands were preferred over males without any color-bands or males with unattractive color-bands, as experiments on finches have shown (Burley et al., 1982; Burley, 1986,1988; Johnson et al., 1993). Providing birds with one ring or band or mammals with one ear tag may render their appearance asymmetrical. Mdler (1992) and others have shown that asymmetrical traits can reduce individuals’ attractiveness to mates. The short-term physiological response to tagging an animal with a radio transmitter may be pronounced. In bighorn sheep, trapping, handling, and radio instrumentation caused a sustained increase in heart rate for 2 h after release and a cardiac recovery time ten times longer than the maximum recovery time recorded for any anthropogenic disturbance to which individuals were subsequently exposed (MacArthur et af., 1986).The fitness consequences of radio transmitters attached by some form of harness or collar have been more frequently studied by rigorous experiments than any other aspect of intervention. Numerous studies have demonstrated that transmitters may or may not change behavior or foraging success, or reduce the chance of breeding, survival, nesting success, or the chance of predation (reviewed by Kenward, 1987; White and Garrott, 1990. Are there any rules that predict under what circumstances a significant detrimental effect of attached transmitters is likely to occur? Most studies have been concerned with the impact of radio transmitters on birds, even though there are many more wildlife radio-tracking studies of mammals than of birds (White and Garrott, 1990). White and Garrott (1990) argued that studies on the consequences of attaching transmitters were more likely to be undertaken if the species was small and depended on flight because of the frequently voiced concern that relative transmitter weight (RTW) may determine whether or not attachment of a transmitter has detrimental fitness consequences. Consequently, little attention has been paid to the effect of transmitters on large carnivores or ungulates where RTW is well below 1% of body weight (BW) and where it is conventionally assumed that transmitters have no detrimental effects (White and Garrott, 1990). There has been no literature survey that looked at the relationship between RTW and of the likelihood of detrimental fitness consequences. Such a survey would have to tackle the issue that if studies were more likely to be undertaken if a problem was suspected in the first place, then the results of the survey might be biased. We therefore prefer to review experimental studies that looked at the effects of RTW within a species. Greenwood and Sargeant (1973) used three different weight classes of transmitters and showed that captive blue-winged teals, Anus discors, lost

482

HERIBERT HOFER A N D MARION L. EAST

more body weight as RTW increased, whereas there was no relationship between RTW and body weight loss in captive mallards; all three weight classes had a significant impact. Houston and Greenwood (1993) also looked at three weight classes in captive mallards and could discern no effect of transmitters in general, and no effect of RTW on a variety of fitness measures including clutch size and nesting interval, but the power (Section 111,DS)of their comparison was low. However, Pietz et af. (1993) and Rotella et af. (1993) found a significant effect of transmitters on fitness measures in wild mallards, and Rotella et af. (1993) found that this effect varied significantly between three different types of transmitters. Amlaner et af. (1979) also used three weight classes to demonstrate that the survivorship of clutches of herring gulls, Larus argentatus, declined as RTW increased. Warner and Etter (1983) demonstrated that female longevity of ring-necked pheasants declined with increasing RTW. In AdClie penguins the length of foraging trips and the incidence of nest desertion increased with increasing volume of fitted measurement devices (Wilson et af., 1989). Cotter and Gratto (1995) showed that male rock ptarmigans, Lagopus mutus, with light transmitters (2.3% BW) had the same survival rate as unmarked males, whereas males with heavy (3.6% BW) transmitters had significantly lower survival than unmarked males. Can these results suggest a rule of thumb for an acceptable RTW? Small animals, such as the greater horseshoe bat, Rhinolophus ferrumequinum, at 15-30 g body weight, can carry transmitters weighing 12% BW without experiencing detrimental fitness consequences (Stebbings, 1982). Larger animals (above 50 g body weight) should not carry transmitters with an RTW above 4-6%. A sensible limit for the largest birds and mammals is probably 1-2% or less (see Kenward, 1987). Even if transmitter weights stay below these limits, it is unwise to assume that just because a species may be large transmitters are unlikely to reduce fitness. For instance, in chinstrap penguins, Pygoscefis antarctica, radio-tagged adults were more likely to abandon nesting attempts than were controls (Croll et af., 1996), and in mule deer, instrumentation was observed to reduce fitness (Goldberg and Haas, 1978; Garrott et af., 1985). Other aspects of radio-tagging may decide whether detrimental fitness consequences are likely to occur. An example is the timing of radio-tagging relative to the reproductive phase of individuals. The general recommendation is to avoid tagging animals during their reproductive period when they appear to be sensitive to disturbance (White and Garrott, 1990). However, wood ducks that were radio-tagged before nesting started were less likely to start incubation than were controls, whereas wood ducks radio-tagged during incubation did not differ in fitness measures from controls (although the power of these comparisons was low; Gammonley and Kelley, 1995). Radio-tagging can also cause sex-specific fitness effects. Female kangaroo

BIOLOGICAL CONSERVATION AND STRESS

483

rats, Dipodomys merriami, reduced excursions from their burrow for the first few nights after radio-tagging and suffered only 47% of predationrelated mortality compared to radio-tagged males, which did not reduce their movements (Daly et al., 1992). 7. Vaccination Several observational and experimental studies in humans and domestic animals suggest that stress may modify vaccination success and be responsible for vaccination failure. Anesthesia and surgery after exposure to rabies virus has been held responsible for vaccination failure (death of the patient) in humans (Fescharek et aL, 1994). A study of elderly humans demonstrated that a chronic stressor such as the duty of caregiving for a spouse with a progressive dementia was responsible for a downregulation of antibody response to vaccination against influenza virus (Kiecolt-Glaser et al., 1996). Cattle that were vaccinated with a dead vaccine against the IBWIPV virus excreted virus after they were subjected to confinement and transport (Frerking et a/., 1995). An experimental study showed that injection of chickens with corticosterone after vaccination against Marek’s disease virus and a new challenge by the virus significantly increased the incidence of Marek’s disease in the vaccinated chickens and downregulated their immune response (Powell and Davison, 1986). After vaccination against caprivox virus, first-calf lactating cows suffered a decrease in milk production and severe generalized skin lesions from which the virus could be isolated, whereas nonlactating cattle did not develop any reactions (Yeruham et al., 1994). There have been no experimental studies of the impact of vaccination on wildlife populations. However, observational studies of both captive and free-ranging wildlife populations demonstrate that it is unwise to assume that vaccinations cannot make things worse. Examples are vaccination failures resulting in the vaccine-induced death of captive, endangered blackfooted ferrets (Carpenter et al., 1976), captive African wild dogs (Durchfeld et al., 1990), and lesser pandas, Ailurus fulgens (Bush et a/., 1976), all vaccinated against canine distemper virus. A retrospective observational study of the life expectancy of African wild dogs in both the Serengeti and Mara ecosystems, after vaccination against rabies with vaccine-filled dart syringes, demonstrated that vaccinated individuals survived for a significantly shorter period than radio-collared ones, and animals either vaccinated or radio-collared were less likely to survive for 12 months than unhandled animals (Burrows et al., 1994, 1995).

8. Captivity, Housing, and Enrichment For the purpose of biological conservation, several aspects of captive housing are important. Housing ought to provide adequate standards of

484

HERIBERT HOFER A N D MARION L. EAST

welfare (Fraser and Broom, 1990; Broom, this volume), and provide an appropriate environment for successful ex situ conservation activities, principally breeding. Poor housing and a boring environment are responsible for permanent changes in behavior (Mason, 1991) and physiology. For instance, solitary housing of vervet monkeys, Cercopithecus aethiops, caused a withdrawal response and was associated with permanent adrenal changes and an increase in the incidence of gastric ulcers with subsequent mortality (Tarara et al., 1995). Stress from crowded captive conditions was a possible primary cause of high chick mortalities in Cape francolins, Francolinus capensis (Hey1 et al., 1988). In pinnipeds, captivity or environmental stressors may cause a potentially fatal sodium imbalance caused by exhaustion of adrenal hormone reserves or desensitization of the cortex to other physiological stimuli (St. Aubin and Geraci, 1986). In zoos, one neglected aspect is the effect of the behavior of visitors on zoo animals when visitors interacted with zoo animals, which could be modified by appropriate housing facilities (Nimon and Dalziel, 1992). There have been strong efforts to improve captive housing by methods summarized as “environmental enrichment.” Environmental enrichment is the provision of objects in captive housing that increases spatial heterogeneity, structural complexity, facilitates an increased variety of activities, and creates different sites of shelter (Markovitz, 1982; Chamove, 1989). A key assumption of environmental enrichment is that an improvement of housing conditions automatically leads to a reduction of stress and an improvement in terms of animal welfare. The measurement of the hormonal and immunological response of captive organisms to enriched housing facilities and their fitness consequences is still in its infancy but has already yielded some surprises (Section IV,C,2). Some studies have demonstrated that enrichment causes a reduction in cortisol levels (e.g., Chamove, 1988). However, results of studies of species held in groups suggest that the impact of enrichment may depend on the social organization of the species, for example, the presence or absence of territoriality (Section IV,C,2). This suggests that enriched environments cannot automatically be considered less “stressful” than conventional environments and that enrichment is unlikely to achieve its aims unless it takes the social organization of a species into account. Good, appropriate housing, however, often ends up as being a relatively benign environment where access by competitors and predators is usually prevented. The problem for biological conservation is that such benign environments (1) may be of limited use to prepare captive animals for the challenges that await them once they are released into the wild (see the next section); (2) stifle the development and fine-tuning of complex behaviors, for example, antipredator behavior (see Curio, 1996); and (3) may not

BIOLOGICAL CONSERVATION A N D STRESS

485

reveal whether the captive stock is of poor genetic quality. Animals that grow up in benign captive environments may lack appropriate antipredator (Jarvi and Uglem, 1993) or foraging and hunting behavior (Scheepers and Venzke, 1995), show a reduced ability to adapt to environmental fluctuations (Kohane and Parsons, 1988), or lack exposure to the level of intraand interspecific competition necessary to reveal the consequences of inbreeding depression (Miller, 1994). Finding the optimal trade-off between welfare concerns and the requirements of biological conservation will be a major task for the future (Wuichet and Norton, 1995).

9.

Transportation and Translocation

Transportation and translocation may be an important stressor, but studies of the physiological response to and fitness consequences of such activities have only recently begun (Woodford and Kock, 1991). Transportation of frequently handled sheep caused a substantial increase in heart rate (Baldock and Sibly, 1990). In bighorn sheep released after capture and instrumentation, heart rate was increased for 2 h after release and the cardiac recovery time was 10 times longer than the maximum recovery time for any disturbance to which individuals were subsequently exposed (MacArthur et al., 1986). The physiological response to transportation includes a reduction in immunocompetence as demonstrated by a reduction in the production of interferon and other components of the immune response (Wattrang et al., 1994; Section IV,C,2; subsection on Vaccination in this section). Fitness consequences of transportation in terms of high mortality may be substantial and are often associated with a decline in immunocompetence (Section IV,C,2). Although there have been many translocations or reintroductions, few studies have reviewed or experimentally explored factors that determine the success or failure of such projects, or monitored the fate of released animals in their new environment. In a review of translocation of mussels, Cope and Waller (1995) found that only 16% of translocated populations were monitored for five or more consecutive years, mortality in translocated populations was unreported in 27% of projects, and there was little guidance on the methods for translocation or for monitoring the subsequent longterm status of translocated mussels. Wolf er al. (1996) reviewed a large number of translocations and concluded that they are more likely to succeed if animals were released into the core of the historical range of a species, into habitat of good to excellent quality, if the population released was large, and if the species had an omnivorous diet. Genetic factors may also play a role because translocated populations and refuge populations show reduced allozyme diversity, as allozymes rare in the parental population are lost (Stockwell et al., 1996). Two behavioral hypotheses that underpin

486

HERIBERT HOFER AND MARION L. EAST

many translocation efforts are that animals acclimated to the novel site before release do better than those who are not and that wild-caught animals do better than captive-bred ones (Bright and Morris, 1994). In other words, there is a suspicion that released animals are “stressed” and fail to cope because they are not sufficiently acclimated or lack the training to face the challenges posed by a new unknown environment. There is observational and experimental evidence that support both hypotheses. An experimental study by Bright and Morris (1994) was set up to test both hypotheses and found support for both of them. Stussy et al. (1994) estimated that translocated female red deer had a lower annual survival rate than resident females; they suggested that translocation may result in higher survival if conducted outside winter months when conditions are less severe. A case study of a released Iberian lynx, Lynx pardinus, by Rodriguez et al. (1995) suggested that the successful release was based on careful feeding-training and avoidance of human contact during the captive phase, as well as selecting a site with good habitat quality.

VII. THEEQUIVALENCE OF NATURAL A N D ANTHROPOGENIC STRESSORS An anthropogenic stressor would be considered equivalent to a natural stressor if an organism’s response to the two was similar or identical. Figure 5 summarizes some hypotheses about such equivalence relationships between natural and anthropogenic stressors. Because equivalence relationships may vary between species, Figure 5 illustrates potential equivalence relationships that may apply to some but not all species. This is not an exhaustive list of all documented links and the links discussed later in this chapter are supported by evidence to a varying degree. We believe that this exercise is useful because there is a well-developed body of theory (behavioral and evolutionary ecology) that predicts how environmental and social factors, resource availability, population density, pathogens, and predators mold life-history tactics, behavior, foraging, survival, and reproductive success of wildlife (e.g., Roff, 1992; Steams, 1992; McNamara and Houston, 1996; Krebs and Davies, 1997). Because there are natural equivalents for most anthropogenic factors, there may be evolved abilities of coping with anthropogenic factors (Sections IV and V). The predictive power of stress studies for biological conservation purposes would therefore be greatly enhanced if this body of theory could be utilized (Section IV,B,S). The best documented link is that between predation and several anthropogenic stressors. These include visitor disturbance (Section V,B), sports hunting (Section V,C), and handling (Section V1,C). It could be argued

BIOLOGICAL CONSERVATION AND STRESS

487

FIG. 5. Hypothesized equivalence relationships between natural and anthropogenic stresson. Evidence from stress response studies suggests that the anthropogenic stressors in the central unshaded group evoke stress responses that are borrowed from evolved responses to natural stressors arranged in the outer shaded group.

that housing in captivity may be related to predation risk because housing in captivity is inevitably associated with handling events. Anthropogenic stressors that have a similar effect to temperature are climatic (global) warming (Section V,E), and urbanization or development. An experimental study demonstrated that local extinction of a land snail was most likely caused by a small increase in ambient temperature caused by thermal radiation from an urban area (Baur and Baur, 1993). Climatic warming may also have effects similar to those of pathogens, because it may favor pathogens and/or because host energy budgets are often related to ambient temperature. Pathogens can modify the energy budgets of hosts, with detrimental fitness consequences for the host (Munger and Karasov, 1989, 1991; Holmes and Zohar, 1990; Forstad et al., 1991). Pollution and handling are linked to pathogens because both reactivate latent viruses (Table I) and impair immunocompetence (Sections IV,C, V,A, V1,C). Natural catastrophes could sometimes be considered the natural equivalent of pollution events when they are associated with significant changes in atmospheric, terrestrial, freshwater, or marine contaminants, and thereby may also contribute to global warming. Visitor disturbance and development may be functionally equivalent to a change in patch richness and other aspects of natural resource availability that influence animals’ decisions on time budgeting and foraging tactics (Gill et al., 1996; Sutherland, 1996). This may also include disturbance caused by sports hunting (Section V,C). Housing in captivity often implies

488

HERIBERT HOFER AND MARION L. EAST

food availability ad libitum, or practices equivalent to clumping of resources (e.g., Boccia et al., 1988). We suspect that anthropogenic factors that are equivalent to social instability may include translocation, which often is functionally equivalent to migration or dispersal (Bright and Morris, 1994) because it removes individuals or populations from one site and introduces them at another site where conspecifics may already exist. The links between high population density and confinement, transportation, translocation, and captivity follow the hypothesis of Christian (1971, 1978, 1980).

OCCURRENCE AND IMPACT OF STRESS IN CONSERVATION VIII. MINIMIZING RESEARCH A N D MANAGEMENT Conservation actions are often initiated to reduce the impact of anthropogenic stressors. For instance, guidelines for tourists visiting Antarctic penguin colonies (Wilson et al., 1991) or internationally standardized rules on whale watching (IFAW, 1996; IFAW et al., 1096) aim to minimize the potential disturbance effects of tourism. Probably no disturbance is better than little disturbance, and little disturbance is better than a lot of disturbance. Beyond such simple rules, however, it is currently impossible to make general recommendations based on a sound factual and theoretical basis because there are few rigorous studies that predict the fitness consequences of stressors in a quantitative, dose-dependent fashion. Predictions would be further complicated by the diversity of organisms, stress response systems, and anthropogenic stressors that may occur (Sections V and VI). As a result, many currently practiced recommendations are essentially hunches based on rules of thumb. In many cases (e.g., wildlife viewing by tourists in East African savannas), very different rules are practiced in different countries even if they share the same ecosystem (e.g., the Serengeti shared by Tanzania and Kenya) and current information is insufficient to evaluate which rules are most appropriate. Even an international comparison of the implementation and consequences of different guidelines of how to approach wildlife would be a major step forward. In this section we concentrate on the issue of minimizing the stressful consequences of conservation management and research activity itself. Here, the data are better, alternatives have been frequently developed, and sometimes their consequences have been rigorously tested. The first step to minimize the occurrence and impact of stress in conservation research and management is to recognize that conservation activities may be potentially stressful. The second is to pay attention to principles of study design (Section 111,D). The third step is to recognize that actions that minimize

BIOLOGICAL CONSERVATION AND STRESS

489

the consequences of potentially stressful conservation activities readily fall into two categories. The first category includes options for minimizing the occurrence of stressors. The occurrence of stressors is sometimes controlled by individual conservation or research programs (e.g., the frequency of handling) and sometimes it is not (e.g., global warming), although concerted efforts might be able to eventually reduce the impact of large-scale anthropogenic stressors such as global warming. The second category includes cases in which it is accepted that under certain conditions the occurrence of a potential stressor is unavoidable and actions aim to maximize an individual’s ability to cope with a stressor. There are several general reviews useful for the practical design of conservation research activities. Recommendations for the use of wild birds in research can be found in the publications by the American Ornithologists’ Union (1988), Hoysak and Weatherhead (1991), and Le Maho et al. (1992). Putman (1995) developed a framework to assess the costs and benefits of capture, handling, and marking mammals in ecological field studies. Other reviews and recommendations are provided by Cuthill (1991) and the Association for the Study of Animal Behaviourkhe Animal Behavior Society published at regular intervals in the journal Animal Behaviour. Procedures for the identification of subjects, capture, telemetry, and sampling of urine and feces in socially living primates have been reviewed, among others, by Rasmussen (1991), for capture, medical management, and anesthesia of free-ranging wildlife by Jessup (1992), and for vaccinating wildlife by Hall and Harwood (1990). THE OCCURRENCE OF STRESS A. MINIMIZING

Because interventions can be a major stressor, reducing the incidence of interventions by replacing standard techniques with noninvasive ones or at least improving standard research equipment to minimize the impact of interventions would reduce the occurrence of stress in some instances. 1. Minimal-Invasive and Noninvasive Alternatives to

Standard Procedures The development of minimal-invasive and noninvasive procedures has been greatly facilitated by recent advances in molecular techniques. These techniques can sometimes replace interventions that include blood sampling as the conventional method of choice to answer many research questions. As an example of the scope of the new techniques, fecal samples may now be used to sex animals by analyzing fecal sex steroids (giant panda, Ailuropoda melanoleuca: Kubokawa, 1993), or prove infection by canine

490

HERIBERT HOFER AND MARION L. EAST

parvovirus using negative contrast electron microscopy (gray wolf Muneer et al., 1988). a. Adrenocortical Response. Blood sampling to determine the magnitude of the adrenocortical response (secretion of corticosteroids) has on occasion been successfully replaced by the analysis of urine samples (domestic dog: Jones et al., 1990; bighorn sheep: Miller et al., 1991; mule deer: Saltz and White, 1991; mink: Madej et al., 1992; rhesus monkey: Crockett et al., 1993; domestic cat: Graham and Brown, 1996), saliva samples (humans: Kirschbaum and Hellhammer, 1994; white rhinoceros, Ceratotherium simum: Schmidt and Sachser, 1996), or fecal samples (bighorn sheep: Miller et a/., 1991; domestic cat: Graham and Brown, 1996). 6. Reproductive State. Daily monitoring of female reproductive state in breeding programs may be possible by analyzing urine (Goeldi’s monkey, Callimico goefdii: Jurke et a/., 1994) or fecal samples (Most1 et al., 1984). Scrota1 dimensions have been determined noninvasively by conditioning captive subjects to hold themselves in a standard vertical position on the mesh walls of their cages so that the maximum width of the scrotum may be compared to a square paper card (cotton-top tamarin, Saguinus oedipus: Ginther and Washabaugh, 1996). Vibrostimulation is a much gentler, yet more reliable method than conventional electro-ejaculation to assess sperm quality and quantity (squirrel monkey: Yeoman et af., 1996). c. Conservation Genetics. Protein diversity may be assessed from feather pulp, obviating the need to blood-sample study animals (Marsden and May, 1984). Hair samples and fecal samples can be used t o identify mitochondrial and nuclear gene sequences in conservation genetics studies (brown bear, Ursus arctos: Taberlet and Bouvet, 1992; Kohn et al., 1995). A single plucked feather may be sufficient for genetic studies that require only small amounts of DNA, for example, when analyzing mitochondria1 DNA or microsatellites with the help of the polymerase chain reaction (blue tit, Purus caeruleus: Taberlet and Bouvet, 1991). d. Monitoring Environmental Contaminants. Instead of killing study animals, levels and effects of organochlorines (PCB and DDT) may be monitored by drawing blood samples (gray seal: Jenssen et af., 1994, 1995). Monitoring mixed function oxidase activity and organochlorine content of body tissue can now be accomplished by remote skin and hypodermic biopsy (fin whale, Balaenoptera physalus, striped dolphin, Stenelfa coeruleoalba: Fossi et a/., 1992). Measurements of contamination with copper, zinc, mercury, cadmium, or lead has been carried out on hair samples collected from resting sites of endangered species in the absence of individuals (Mediterranean monk seal, Monachus monahus: Yediler et al., 1993).

2. Improving Standard Research Equipment Sometimes, noninvasive alternatives are not yet available. In such cases, improving standard research equipment to minimize stress is an option.

BIOLOGICAL CONSERVATION AND STRESS

491

Designs that minimize stress may reduce the number of handling events by avoiding the need to recapture animals. An example is the attachment of a buoyant pack containing a VHF transmitter and a data recorder to the pelt of harbor seals, which drops off after a predefined duration (Ellis and Trites, 1992). Data collection by implantable transmitters can now be remote-controlled by a computer. These devices are sophisticated enough to detect acute changes in heart rate and core body temperature and discriminate among several different types of behavior without restraining or handling animals (Diamant et al., 1993; Kramer et al., 1993). The stressfulness of the handling event may also be reduced. Delgiudice et al. (1990) reported on the development of a capture collar that includes a syringe with immobilization drugs. Once the capture collar is attached to an animal, chemical immobilization can be initiated by a remote-controlled radio signal, obviating the need for darting the animal. Under some experimental conditions, regular disturbance or handling of captive animals for blood sampling can be avoided with remote-controlled blood sampling techniques (Le Maho et al., 1992; Alexander et al., 1996). The shape and point of attachment of data loggers can be designed so that they match the body contour of the animal, resulting in substantial energy savings during foraging and locomotion (penguins: Bannasch et al., 1994; Culik et al., 1994). THE ABILITY To COPESUCCESSFULLY WITH STRESS B. MAXIMIZING

In some cases, interventions or other forms of potential stress may be unavoidable. How can an organism be prepared to maximize its ability to cope with stress? 1. Early Experience and Behavioral Training In captivity, rat pups exposed to brief periods of innocuous handling early in life showed a reduced adrenocortical response to a wide variety of potential stressors, and this effect persisted throughout the life of the animal. This has been interpreted to mean that early experience of handling can reduce the stressfulness of handling later in life (Meaney et al., 1991; Gonzalez et al., 1994; Rostene et al., 1995). Recent experiments with cats and rhesus monkeys confirmed these results (Meaney et al., 1993; McCune, 1995). Apparently this effect occurs because handling modifies the transcription activity of glucocorticoid receptor genes early in life, permanently altering neuroendocrine responsitivity to stress (Meaney et al., 1993). Conservation programs often consider the release of captively bred individuals into the wild where they might encounter environmental stressors and predators, and the success of translocation or release programs may be heavily affected by this (e.g., Scheepers and Venzke, 1995). Jarvi (1990) and Jarvi and Uglem (1993) used a series of elegant experiments to investi-

492

HERIBERT HOFER A N D MARION L. EAST

gate the impact of encounters of salmon smolts that had been hatchery reared in freshwater with predators under conditions of osmotic stress (in seawater). Smolts were either naive or previously trained by exposing them to predators in a contact (freely hunting predator) or noncontact fashion (predator behind a glass screen). Trained smolts were more likely to respond with appropriate behavior and show a reduction in the physiological stress response than naive smolts, and contact-exposed smolts did better than noncontact smolts. Active antipredator training might therefore improve the performance of released animals and enhance the success of release programs. A soft release (in which animals are acclimated to a release site and/or decide on their own when to leave a holding area) may be viewed as a self-training exercise and thus be predicted to achieve a similar positive effect. Bright and Morris (1994) experimentally demonstrated that soft releases were indeed more successful than hard releases (no acclimation at the release site).

2.

Training to Cooperate

Numerous reports demonstrate that nonhuman primates can be trained to cooperate with rather than resist common handling procedures such as capture, injection, blood sampling, and veterinary examination (Reinhardt et al., 1995). The same applies to captive cetaceans trained to drape their flukes over the edge of the holding pool for blood samples (Thomas et al., 1990). Animals can also be trained to donate urine or fecal samples in appropriate containers (Kelley and Bramblett, 1981; Phillippi-Falkenstein and Clarke, 1992; Anzenberger and Gossweiler, 1993). Cooperative animals showed reduced or absent behavioral and physiological signs of distress (Reinhardt et al., 1995). Some of these training procedures invoke aversive stimulation, at least in primates (Rasmussen, 1991). Several studies demonstrate that individuals or groups can be trained using positive reinforcement procedures without aversive stimulation and that the effort required is much less than skeptics might expect. Voluntary presentation by rhesus monkeys of their leg for blood collection without mechanical restraint was easily achieved (Reinhardt, 1991b). Only a minimal time investment was needed to train a large troop of laboratory rhesus monkeys to cooperate in a capture procedure, minimizing risk to personnel and distress to the animals (Luttrell et al., 1994). Positive reinforcement was used to rapidly train chimpanzees to move to indoor quarters on a verbal cue, allowing personnel safe access to enclosures for maintenance, to move animals to other quarters, and facilitate veterinary and research procedures (Stone et al., 1996). The success of habituation and positive reinforcement procedures is not restricted to nonhuman primates, as training of nyala, Tragelaphus

BIOLOGICAL CONSERVATION AND STRESS

493

angasi, to enter a wooden crate for veterinary examination and blood sampling has demonstrated (Grandin et al., 1995).

3. Social Support Social stability and the presence of preferred social partners may assist to significantly reduce the glucocorticoid stress response in an acute challenging situation (Sachser and Beer, 1995; Sachser et al., 1997; Sapolsky, 1997). This may be because the presence of preferred social partners improves the confidence of an individual that it will be successful in meeting the challenge, or because a potentially stressful situation is experienced as less challenging if a preferred social partner is present. 4.

Tranquilizers and Sedatives

Tranquilizers and sedatives may help to tone down the apparent behavioral response of an animal to a handling event. For instance, placing a sedative injector in the nest of breeding seabirds and administering an intramuscular sedative by remote-control to facilitate capture avoided fear and stress-related behavior during capture of birds (Wilson and Wilson, 1989). However, this does not necessarily imply that the physiological response to intervention is muted in a similar way to the behavioral response. Tranquilizers did not suppress the physiological stress response in impala, Aepyceros melampus, to repeated capture, handling, and blood sampling, as measured by increases in lactate, glucose, cortisol, total catecholamines, osmolality, and hematocrit (Knox et al., 1990).

5. Optimizing Intervention Conditions Several experimental studies confirm that optimizing intervention conditions can result in substantial downregulation of the physiological response to intervention. Predictable handling and husbandry routines and the provisioning of appropriate places for concealment reduce the distress displayed by domestic cats responding to handling by familiar or unfamiliar humans (Carlstead et al., 1993). Quiet, efficient handling based on an understanding of behavior reduces apparent levels of excitement and the incidence of injuries when domestic cattle are tagged with ear implants and injected (Grandin, 1987). The impact of handling in terms of the adrenocortical response may also be reduced by handling captive individuals in a familiar environment rather than taking them to a restraint apparatus (Reinhardt et al., 1991). Enclosures may be optimized for efficient handling and data collection, simultaneously reducing the potential of injury to the animal. An example is the portable, semirigid enclosure with opaque, vertical sides of vinyl-coated nylon developed by Davis and Allen (1989) to handle captured waterfowl. Fear-related behavior and heart rate of red deer, either

494

HERIBERT HOFER AND MARION L. EAST

confined in an unfamiliar holding pen with a stationary human present or housed indoors, was reduced by darkening the holding area (Pollard and Littlejohn, 1999, or by sound-proofing housing facilities to minimize noise (Price et al., 1993).

IX.

CONCLUSIONS: How IMPORTANT Is STRESSIN BIOLOGICAL CONSERVATION?

From our review a few interesting rules and trends emerge. It is unwise to assume that: (1)issues related to stress can be safely ignored in conservation or research; (2) short-term observations of behavior provide reliable evidence about whether an organism is stressed; (3) a single-factor explanation of stress-related individual and population responses is sufficient; and (4) interventions in the course of research or conservation actions are always benign. There are, however, good reasons to believe that: (1) the awareness of the importance of stress in biological conservation has increased steadily and will continue to do so; (2) in some cases individuals and populations are adequately equipped to deal with anthropogenic stress of moderate intensity because such stress mimics other factors to which these animals have evolved an adaptive response; ( 3 ) it is possible to measure the extent to which resource use declines with disturbance and quantify how much anthropogenic disturbance reduces the effective population size of a habitat; (4) there are now many noninvasive alternatives to conventional research procedures; and (5) stress associated with conservation and research activities can be substantially reduced. Conservation efforts attempt to improve population size, population persistence, habitat distribution and quality, minimize the chance of genetic depauperation within species, and preserve community structure and interactions as intact as possible. How do stress studies help in this respect? We believe that stress studies contribute in three major ways. The first refers to the practicalities of conservation work: Should minimization of stress be considered an important element of the implementation of conservation activities? And would the success of conservation efforts be improved by minimizing stress-related fitness consequences? Our review suggests that the answer to both questions is yes. Recognizing that stress is an important factor paves the way for organizing conservation activities around two paradigms: Minimizing the occurrence of stress, and maximizing an individual’s or population’s ability to cope successfully with stress (Section VIII). In this context it would be helpful if conservation activists and researchers routinely reported the precise circumstances of interventions and explain how they attempted to control for the effects of intervention.

BIOLOGICAL CONSERVATION AND STRESS

495

The second refers to the recognition that stress must be considered an important factor in evolution (Section IV and Hoffmann and Parsons, 1991; Parsons, 1988a, 1991, 1993a,b, 1994). Anthropogenic changes of the environment are likely to create, or have already created, substantial selection pressures on populations, increasing additive genetic variance and recombination rates (Parsons, 1988a). If we want to predict the future course of populations, then understanding the magnitude and effects of such selection pressures is important. The third is that recent work on the genetics of stress resistance suggest that conservation geneticists need to address several points of concern neglected by standard treatments of evolutionary genetics. Current evolutionary theory is not well geared to understand the effects of fluctuating and stressful environments, as most work in evolutionary genetics focuses on equilibrium populations in constant and stable environments (see Hoffmann and Parsons, 1991). Maximizing stress resistance or stress tolerance may require: (1) the preservation of rare alleles rather than that of overall genetic diversity (Futuyma, 1983); (2) an emphasis on the preservation of marginal rather than central populations (Parsons, 1989a, 1995); (3) the prevention of “genetic flooding” of marginal populations caused by the creation of habitat corridors (Hoffmann and Parsons, 1991); (4) the recognition that the consequences of potential inbreeding depression in a newly released captive population may not be predicted from a relatively benign captive environment (Miller, 1994); ( 5 ) the recognition that heterozygotes may express superior stress resistance only under conditions of environmental stress and not in benign environments (e.g., Scott and Koehn, 1990); (6) the acknowledgment that long-term breeding in benign captive environments (e.g., in temperature-regulated facilities) may reduce the capacity of the captive population to adapt to conditions of natural habitats (Kohane and Parsons, 1988). What are the consequences of ignoring the possible effects of stress in conservation work? They may be profound if stress is part of the reason why populations or species decline, if appropriate activities to stop the decline are available and yet stress is not identified as a contributing factor. This problem has been recently highlighted by the “cheetah controversy” (May, 1995). O’Brien (1994) argued that in cheetahs small litter sizes and difficulties encountered by captive breeding programs point to inbreeding depression, which is thought to be caused by abnormally low levels of genomic diversity. Caughley (1994) and Merola (1994) suggested instead that low reproductive performance was not an adequate indicator of inbreeding depression because breeding performance in captivity may have been impaired by the stress caused by inadequate conditions. With im-

496

HERIBERT HOFER AND MARION L. EAST

proved conditions, captive breeding programs should, and do, increase reproductive success (May, 1995). We conclude that a biologically appropriate conservation effort recognizes potential stress situations and attempts to minimize the stressful consequences of conservation activities.

X. SUMMARY This chapter reviews why stress has important implications for biological conservation and considers practical ways in which conservationists can identify and tackle problems caused by stress. We take an evolutionary approach that emphasizes links between stress and its consequences and possible adaptations that permit animals to cope with stress. Using information from several scientific disciplines we outline current knowledge of the kinds of factors known to generate stress and the Darwinian fitness consequences of stress. The magnitude and nature of the organism’s stress response can be highly variable between species, populations of the same species, individuals, and even change with the reproductive state or body condition of an individual. Detailed knowledge on the sources of this variation and the rules that govern it is urgently required to predict the likely impact of anthropogenic stressors. Such knowledge is lacking because the plethora of retrospective observational studies of natural or anthropogenic stressors contrasts with a paucity of rigorously designed experimental studies. Experimental studies have demonstrated that anthropogenic factors such as environmental pollution, tourism and leisure activities, hunting, noise, and global warming d o sometimes-but not invariably-cause stress, with detrimental fitness consequences, in a wide variety of species, contexts, and dosages. There are currently no general rules available to predict when anthropogenic factors are likely to generate stress that will result in detrimental fitness consequences, or which factors are likely to have the most severe impact. Conservationists are not sufficiently aware of the impact of natural stress, anthropogenic stress, or stress created by conservationrelated activities and scientific research, which are usually considered benign. We outline a research program for stress in conservation biology and provide recommendations that can improve the success of conservation efforts by minimizing the frequency of occurrence of conditions that generate stress and maximizing the chance of organisms to successfully cope with stress. Acknowledgments

We are grateful to numerous colleagues who have discussed some of these issues with us over the years or provided information. reprints. o r preprints. In particular we would like to

BIOLOGICAL CONSERVATION AND STRESS

497

thank D. H. Abbott. G. Anzenberger. F. Aureli, N. Bahr. M. Bekoff. R. Burrows. A. Camperio Ciani, A. Cockburn. E. Curio, P. Deimer. N. Hillgarth. L. Haas. J. Lamprecht. G. Lubach, C. Richard-Hansen, R. M. Sapolsky. D. Quiatt, V. Reynolds, N. Sachser. J. Silk, T. Smith, W. J. Sutherland, and J. Wingfield. We are grateful to R. Klein, B. Knauer, K. Schulz, A. Turk, W. Wickler, and the Max-Planck-Gesellschaft for assistance and financial support and to the referees for constructive comments.

References Abbott, D. H. (1989).Social suppression of reproduction in primates. In “Comparative Socioecology” (V. Standen and R. A. Foley, eds.). pp. 285-304. Blackwell, Oxford. Abbott, D. H.. Saltzman, W., Schultz-Darken. N. J.. and Smith, T. E. (1997).Specific neuroendocrine mechanisms not involving generalized stress mediate social regulation of female reproduction in cooperatively breeding marmoset monkeys. Ann. N. Y. Acad. Sci. 807, 219-239. Abbott, J. C..and Dill, L. M. (1989).The relative growth of dominant and subordinate juvenile steelhead trout (Salrno gairdneri) fed equal rations. Behavioitr 108, 104-1 13. Alados, C. L.. Escos. J. M., and Emlen. J. M. (1996).Fractal structure of sequential behavior patterns. Anim. Behav. 51, 437-443. Alberts. S. C.. Sapolsky. R. M.. and Altmann, J. (1992).Behavioral, endocrine, and immunological correlates of immigration by an aggressive male into a natural primate group. Horrn. Behnv. 26, 167-178. Alexander, S. L., Irvine, C. H. G., and Donald, R. A. (1996). Dynamics of the regulation of the hypothalamo-pituitary-adrenal (HPA) axis determined using a nonsurgical method for collecting pituitary venous blood from horses. Front. Neuroendocrinol. 17, 1-50. Altmann. J., Hausfater, G., and Altmann. S. A. (1988).Determinants of reproductive success in savannah baboons, Papio cynocephaliis. In “Reproductive Success” (T. H. CluttonBrock, ed.). pp. 403-418. University of Chicago Press, Chicago. American Ornithologists’ Union. (1988). Report of committee on the use of wild birds in research. Auk 105, Suppl., la-41a. Amlaner. C. J., Sibly, R. M., and McCleery. R. (1979). Effects of telemetry transmitter weight on breeding success in herring gulls. In “Proceedings of the Second International Conference on Wildlife Biotelemetry” (F. M. Long, ed.), pp. 254-259. University of Wyoming, Laramie. Anderson. P. K. (1981).The behavior of the dugong (Dugong dicgon) in relation to conservation and management. Bull. Mar. Sci. 31, 640-647. Anzenberger, G., and Gossweiler, H. (1993).How to obtain individual urine samples from undisturbed marmoset families. Am. J. Prirnatol. 31, 223-230. Ardern, S. L.. McLean, 1. G . , Anderson, S.. Maloney, R., and Lambert, D. M. (1994).The effects of blood sampling on the behavior and survival of the endangered Chatham Island black robin (Petroicn traversi). Conserv. B i d . 8, 857-862. Arita, H. T. (1996).The conservation of cave-roosting bats in Yucatan. Mexico. Biol. Conserv. 76, 177-185. Arnold, W., and Dittami, J. (1997).Reproductive suppression in male alpine marmots. Anirn. Behav. 53, 53-66. Arntz, W. E., and Fahrbach, E. (1991). “El NiAo. Klimaexperiment der Natur.” Birkhauser, Basel. Arthur, A. Z. (1987).Stress as a state of anticipatory vigilance. Percept. Mot. Skills 64,7545. Astheimer, L. B., Buttemer, W. A,, and Wingfield, J. C. (1992).Interactions of corticosterone with feeding activity and metabolism in passerine birds. Ornis Scand. 23, 355-365.

498

HERIBERT HOFER AND MARION L. EAST

Astheimer, L. B., Buttemer, W. A,, and Wingfield. J. C. (1994). Gender and seasonal differences in the adrenocortical response to ACTH challenge in an arctic passerine, Zonotrichia leucophrys gambelii. Gen. Comp. Endocrinol. 94, 33-43. Astheimer, L. B., Buttemer, W. A,, and Wingfield, J. C. (1995). Seasonal and acute changes in adrenocortical responsiveness in an arctic-breeding bird. Horm. Behav. 29, 442-457. Backstrom, L., and Kauffman, R. (1995). The porcine stress syndrome: A review of genetics, environmental factors, and animal well-being implications. Agri-Practice 16(8), 24-30. Baker, C. S.. and Palumbi, S. R. (1994). Which whales are hunted? A molecular genetic approach to monitoring whaling. Science 265, 1538-1539. Baldock. N. M., and Sibly, R. M. (1990). Effects of handling and transportation on the heart rate and behavior of sheep. Appl. Anim. Behav. Sci. 28, 15-39. Bannasch, R., Wilson, R. P., and Culik, B. M. (1994). Hydrodynamic aspects of design and attachment of a back-mounted device in penguins. J. Exp. Biol. 194,83-96. Barker, I. K., Carbonell, P. L., and Bradley, A. J. (1981). Cytomegalovirus infection of the prostate in the dasyurid marsupials. Phascogale tapoatafa and Antechinus stuartii. J. Wildl. Dis. 17, 433-441. Barnard, C. J., Behnke, J . M., and Sewell, J. (1996). Environmental enrichment, immunocompetence and resistance to Babesia microri in male laboratory mice. Physiol. Behav. 60,12231231. Bateson, P. P. G. (1997). “The behavioural and physiological effects of culling red deer.” Report to the Council of the National Trust, 77p. Baur, B., and Baur, A. (1993). Climatic warming due to thermal radiation from an urban area as possible cause for the local extinction of a land snail. J. Appl. Ecol. 30,333-340. Baveco, J. M., and De Roos. A. M. (1996). Assessing the impact of pesticides on lumbricid populations: An individual-based modelling approach. J. Appl. Ecol. 33, 1451-1468. Beletsky, L. D., Orians, G. H., and Wingfield, J. C. (1989). Relationships of steroid hormones and polygyny to territorial status, breeding experience, and reproductive success in male red-winged blackbirds. Auk 106, 107-1 17. Beletsky, L. D., Orians, G. H., and Wingfield. J. C. (1990). Steroid hormones in relation to territoriality, breeding density, and parental behavior in male yellow-headed blackbirds. Auk 107, 60-68. Ben-Nathan, D. (1994). Stress and infectious disease. Isr. J . Vet Med. 49, 105-112. Bentson, K. L., Astley, C. A,. Miles, F. P., Goldstein, D. S., Holmes, C., and Smith, 0. A. (1996). Behavioral and physiological aspects of exposure of a group of baboons to an intruder. Ahstr., Congr. Int. Primatol. Soc., 16th, 92. Berglund, B., Lindvall, T., and Nordin, S. (1990). Adverse effects of aircraft noise. Environ. Int. 16,315-338. Bergman, A,, and Olsson, M. (1986). Pathology of baltic gray seal (Halichoerus grypus) and ringed seal (Phoca hispida botnica) females with special reference to adrenocortical hyperplasia: Is environmental pollution the cause of a widely distributed disease syndrome. Riistatieteellisia Julkaisuja 44,47-62. Beringer, J., Hansen, L. P., Wilding, W., Fischer, J., and Sheriff, S. L. (1996). Factors affecting capture rnyopathy in white-tailed deer. J . Wildl. Manage. 60, 373-380. Bernstein, C., Krebs, J. R., and Kacelnik, A. (1991). Distribution of birds amongst habitats: Theory and relevance to conservation. In “Bird Population Studies” (C. M. Perrins, J . D. Lebreton, and G. M. Hirons, eds.), pp. 317-345. Oxford University Press, Oxford. Berthiaume. L., Heppell. J., Desy, M., Leblanc, L., Lallier, R., Bailey, R., and Dutil, J. D. (1993). Manifestation of lymphocystis disease in American plaice (Hippoglossoides platessoides) exposed to low salinities. Can. J. Fish. Aquat. Sci. 50, 430-434.

BIOLOGICAL CONSERVATION AND STRESS

499

Black, S. R., Barker, I. K., Mehren, K. G., Crawshaw, G. J., Rosendal, S.. Ruhnke. L., Thorsen. J., and Carman, P. S. (1988). An epizootic of mycoplasma ovipneumoniae infection in captive Dall’s sheep (Ovis dalli dalli). J. Wildl. Dis. 24, 627-635. Bleich. V. C., Bowyer, R. T.. Pauli, A. M., Nicholson, M. C., and Anthes, R. W. (1994). Mountain sheep Ovis canadensis and helicopter surveys: ramifications for the conservation of large mammals. Biol. Conserv. 70, 1-7. Boccia, M. L., Laudenslager, M., and Reite, M. (1988). Food distribution, dominance and aggressive behaviors in bonnet macaques. Am. J . Primatol. 16, 123-130. Bodini, A,, Giavelli, G., and Rossi, 0.(1994). The qualitative analysis of community food webs: Implications for wildlife management and conservation. J . Environ. Manage. 41,49-65. Bonneau, R. H., Sheridan, J. F., Feng, N., and Glaser, R. (1991). Stress-induced effects on cell-mediated innate and adaptive memory components of the murine immune response to herpes simplex virus infection. Brain, Behav., Immun. 5, 274-295. Bowman, L. A,, Dilley, S. R., and Keverne, E. B. (1978). Suppression of oestrogen-induced LH surges by social subordination in talapoin monkeys. Nature (London) 275, 56-58. Boyce, M. S. (1992). Population viability analysis. Annu. Rev. Ecol. Syst. 23, 481-506. Bradley, A. J. (1987). Stress and mortality in the red-tailed phascogale, Phascogale calura (Marsupialia: Dasyuridae). Gen. Comp. Endocrinol. 67, 85-100. Bradley, A. J. (1990). Failure of glucocorticoid feedback during breeding in the male redtailed phascogale Phascogale calura (Marsupialia: Dasyuridae). J . Steroid Biochem. Mol. Biol. 37, 155-163. Bradley, A. J., McDonald, I. R., and Lee, A. K. (1980). Stress and mortality in a small marsupial (Antechinus sfuartii).Gen. Comp. Endocrinol. 40, 188-200. Brereton, R., Bennett, S., and Mansergh, I. (1995). Enhanced greenhouse climate change and its potential effect on selected fauna of south-eastern Australia: A trend analysis. Biol. Conserv. 72,339-354. Bright, P. W.. and Morris, P. A. (1994). Animal translocation for conservation: Performance of dormice in relation to release methods, origin and season. J. Appl. Ecol. 31,699-708. Brisbin, I. L., White, G. C., and Bush, P. B. (1986). Polychlorinated biphenyl intake and the growth of waterfowl: Multivariate analyses based on a reparameterized Richards sigmoid model. Growrh 50, 1- 11. Brodie. P.. and Paasche, A. (1985). Thermoregulation and energetics of fin and sei whales based on postmortem stratified temperature measurements. Can. J. 2001.63,2267-2269. Bronson, F. H. (1989). “Mammalian Reproductive Biology.” University of Chicago Press, Chicago. Broom, D. M., and Johnson, K. G. (1991). “Animal Welfare.” Chapman & Hall, London. Broome, L. A., and Geiser, F. (1 995). Hibernation in free-living mountain pygmy-possums, Burramys parvrcs (Marsupialia: Burramyidae). Aust. J. Zool. 43, 373-379. Burel, C., Mezger, V., Pinto, M.. Rallu, M., Trigon, S., and Morange, M. (1992). Mammalian heat shock protein families. Expression and functions. Experientia 48, 629-634. Burley, N. (1986). Comparison of the band-color preferences of two species of estrildid finches. Anim. Behav. 34, 1732-1741. Burley, N. (1988). Wild zebra finches have band-color preferences. Anim. Behav. 36, 12351237. Burley. N., Krantzberg, G., and Radman, P. (1982). Influence of colour-banding on the conspecific preferences of zebra finches. Anim. Behav. 27,686-698. Burrows, R. (1992). Rabies in wild dogs. Nature (London) 359, 277-277. Burrows, R., Hofer, H., and East, M. L. (1994). Demography, extinction and intervention in a small population: The case of the Serengeti wild dogs. Proc. R. Soc. London, Ser. B 256, 281 -292.

500

HERIBERT HOFER AND MARION L. EAST

Burrows, R.. Hofer. H.. and East, M. L. (1995). Population dynamics, intervention and survival in African wild dogs. (Lycaon pictus). Proc. R. Soc. London, Serv. B 262, 235-245. Busby, D. G.. White, L. M., and Pearce, P. A. (1990). Effects of aerial spraying of fenitrothion on breeding white-throated sparrows. J . Appl. E d . 27,743-755. Bush, M.,Montali, R. J., Brownstein, D., James, A. E., and Appel. M. J. G. (1976). Vaccineinduced canine distemper in a lesser panda. J. Am. Vet. Med. Assoc. 169, 959-960. Butler, R. G.. Harfenist. A., Leighton, F. A., and Peakall, D. B. (1988). Impact of sublethal oil and emulsion exposure on the reproductive success of Leach’s storm petrels: Short and long-term effects. J. Appl. Ecol. 25, 125-143. Buttemer, W. A., Astheimer. L. B., and Wingfield, J. C. (1991). The effect of corticosterone on standard metabolic rates of small passerine birds. J . Comp. Physiol. B 161, 427-431. Calow. P.. and Sibly, R. M. (1990). A physiological basis of population processes: ecotoxicological implications. Funct. Ecol. 4, 283-288. Carlstead. K., Brown. J. L., and Strawn. W. (1993). Behavioral and physiological correlates of stress in laboratory cats. Appl. Anim. Behau. Sci. 38, 143-158. Carpenter. F. L., and Hixon, M. A. (1988). A new function for torpor fat conservation in a wild migrant hummingbird. Condor 90, 373-378. Carpenter, J. W., Appel, M. J. G., Erickson, R. C.. and Novilla, M. N. (1976). Fatal vaccineinduced canine distemper virus infection in black-footed ferrets. J. Am. Vet. Merf.Assoc. 169, 961-964. Castrucci. G., Frigeri, F.. Cilli. V., Tesei, B., Arush, A. M.. Pedini, B., Ranucci. S., and Rampichini. L. (1980). Attempts to reactivate bovid herpesvirus-2 in experimentally infected calves. Am. J . Vet. Res. 41, 1890-1893. Caughley. G. (1994). Directions in conservation biology. J. Anim. Ecol. 63, 215-244. Chalmers. G. A., and Barrett. M. W. (1982). Capture myopathy. In “Non-infectious Diseases in Wildlife” (G. L. Hoff and J. W. Davies, eds.), pp. 84-94. Iowa State University Press, Ames. Chamove, A. S. (1988). Cage design reduces emotionality in mice. Lab. AnOn. 23, 215-219. Chamove, A. S. (1989). Environmental enrichment: A review. Anim. Technol. 40, 155-178. Charlesworth, B. (1994). “Evolution in Age-structured Populations.” Cambridge University Press, Cambridge, UK. Christian, J. J. (1971). Population density and reproductive efficiency. Biol. Reprod. 4,248-294. Christian. J. J. (1978). Neurobehavioural endocrine regulation in small mammal populations. In “Populations of Small Mammals under Natural Conditions” (D. P. Snyder, ed.), pp. 143-158. University of Pittsburgh Press, Pittsburgh. Christian, J. J. (1980). Endocrine factors in population regulation. In “Biosocial Mechanisms in Population Regulation” (M. N. Cohen, R. S. Malpass, and H. Klein, eds.), pp. 55-115. Yale University Press, New Haven, CT. Chrousos. G. P., Loriaux, D. L., and Gold. P. W. (1988). The concept of stress and its historical development. In “Mechanisms of Physical and Emotional Stress” (G. P. Chrousos, D. L. Loriaux, and P. W. Gold. eds.). pp. 3-7. Plenum. New York. Clark, W. W. (1992). Hearing: The effects of noise. Otolaryngol. Head Neck Surg. 106,669-676. Clarke, A. S., and Schneider. M. L. (1993). Prenatal stress has long-term effects on behavioral responses to stress in juvenile rhesus monkeys. Dev. Psychobiol. 26,293-304. Clarke, A. S., Wittwer, D. J., Abbott. D. H.. and Schneider, M. L. (1994). Long-term effects of prenatal stress on HPA axis activity in juvenile rhesus monkeys. Deu. Psychobiol. 27,257-269. Clarke, G. M. (1995). Relationships between developmental stability and fitness: Application for conservation biology. Conserv. Bid. 9, 18-24.

BIOLOGICAL CONSERVATION AND STRESS

501

Clarke, G. M. (1996). Relationships between fluctuating asymmetry and fitness: How good is the evidence? Pac. Conserv. Biol. 2, 146-149. Cochran, W. G. (1977). “Sampling Techniques.” Wiley, New York. Cochrane, B. J., Mattley, Y. D., and Snell, T. W. (1994). Polymerase chain reaction as a tool for developing stress protein probes. Environ. To.ricol. Chem. 13, 1221-1229. Cockburn. A. (1997). Living slow and dying young: senescence in marsupials. I n “Recent Advances in Marsupial Biology” (N. Saunders and L. Hinds, eds.). University of New South Wales Press, New Kensington, Australia (in press). Coe, C. L., Lubach, G. R., and Ershler, W. B. (1994). Maternal influences on the development of immune competence in infancy. Psychol. Beitr. 36, 15-21. Coe, C. L., Lubach, G. R., Karaszewski, J. W.. and Ershler, W. B. (1996). Prenatal endocrine activation alters postnatal cellular immunity in infant monkeys. Brain, Behav.. Immun. 10, 221-234. Cohen, J. (1988). “Power Analysis for the Behavioral Sciences.” 2nd ed., Erlbaum. Philadelphia. Colwell, M. A., Gratto, C. L., Oring, L. W., and Fivizzani, A. J. (1988). Effects of blood sampling on shorebirds: Injuries, return rates and clutch desrtions. Condor 90, 942-945. Cope, W. G., and Waller, D. L. (1995). Evaluation of freshwater mussel relocation as a conservation and management strategy. Regul. Rivers: Res. Manage. 11, 147-155. Cosby, S. L., McQuaid, S., Duffy, N., Lyons, C., Rima, B. K., Allan, G. M., McCullough. S. J., Kennedy, S., Smyth, J. A., McNeilly, F., and Craig, C. (1988). Characterization of a seal morbillivirus. Nature (London) 336, 115-115. Cosens, S. E., and Dueck, L. P. (1993). Icebreaker noise in Lancaster Sound. Northwest Territories. Canada: Implications for marine mammal behavior. Mar. Mammal. Sci. 9, 285-300. Costa, D. P., and Trillmich, F. (1988). Mass changes and metabolism during the perinatal fast: A comparison between Antarctic (Arcrocephalus gazella) and Galapagos fur seals (Arctocephalus galapagoensis). Physiol. Zool. 61, 160-169. Cotter, R. C., and Gratto, C. J. (1995). Effects of nest and brood visits and radio transmitters on rock ptarmigan. J. Wildl. Manage. 59, 93-98. Creel, S. (1992). Causes of wild dog deaths. Nature (London) 360, 633-633. Creel, S. (1996). Conserving wild dogs. Trends Ecol. Evol. 11, 337-337. Creel, S. R., Creel, N. M., and Monfort, S. (1997). Radiocollaring and stress hormones in African wild dogs. Conserv. Biol. 11,544548. Crockett, C. M., Bowers, C. L., Sackett, G. P., and Bowden, D. M. (1993). Urinary cortisol responses of longtailed macaques to five cage sizes, tethering, sedation, and room change. Am. J. Primatol. 30,55-14. Croll, D. A., Jansen, J. K., Goebel, M. F., Boveng, P. L., and Bengtson, J. L. (1996). Foraging behavior and reproductive success in chinstrap penguins: The effects of transmitter attachment. J. Field Omithol. 67, 1-9. Culik, B. M. (1992). Diving heart rates in Adtlie penguins (Pygoscelisadeliae).Comp. Biochem. Physiol. A 102,487-490. Culik, B. M., and Wilson, R. P. (1991a). Swimming energetics and performance of instrumented AdClie penguins (Pygoscelis adeliae). J. Exp. Biol. 158,355-368. Culik, B. M., and Wilson, R. P. (1991b). Penguins crowded out? Nature (London) 351,340-340. Culik, B. M., and Wilson, R. P. (1992). Field metabolic rates of penguins instrumented penguins using doubly-labelled water. J. Comp. Physiol. B 162, 567-573. Culik, B. M., and Wilson, R. P. (1995). Penguins disturbed by tourists. Nature (London) 376,301-302.

502

HERIBERT HOFER A N D MARION L. EAST

Culik, B. M., Adelung, D.. and Woakes, A. J. (1990). The effect of disturbance on the heart rate and behaviour of Adelie penguins (Pygoscelis adeliae) during the breeding season. In “Antarctic Ecosystems. Ecological Change and Conservation” (K. R. Kerry, and G. Hempel, eds.), pp. 171-182. Springer, Berlin. Culik. B. M., Bannasch, R.. and Wilson, R. P. (1994). External devices on penguins: How important is shape? Mar. Biol. 118,353-357. Curio, E.(1996). Conservation needs ethology. Trends Ecol. Evol. 11, 260-263. Cuthill, I. (1991). Field experiments in animal behaviour: methods and ethics. Anim. Behav. 42, 1007-1014. Daan. S., Deerenberg. C., and Dijkstra, C. (1996). Increased daily work precipitates natural death in the kestrel. J. Anim. E d . 65, 539-544. Dabbert, C. B., and Powell, K. C. (1993). Serum enzymes as indicators of capture myopathy in mallards Anus platyrhynchos. J. Wildl. Dis. 29, 304-309. Daly. M., Wilson, M. I., Behrends, P. R.. and Jacobs. L. F. (1992). Sexually differentiated effects of radio transmitters on predation risk and behaviour in kangaroo rats Dipodoniys merriami. Can. J. Zool. 70, 1851-1855. Davis. D. S., and Allen, H. A. (1989). An improved waterfowl enclosure: Considering animal welfare as a research priority. J. Field Ornifhol. 60, 162-167. Davis, R. W., Williams, T. M., Thomas, J. A,, Kastelein, R. A,, and Cornell, L. H. (1988). The effects of oil contamination and cleaning on sea otters (Enhydra lutris). 11. Metabolism, thermoregulation and behavior. Can. J. Zool. 66, 2782-2790. Dawson, W. R., and Hudson, J. W. (1970). Birds. In “Comparative Physiology of Thermoregulation I ” (G. C. Whittow, ed.), pp. 223-310. Academic Press, New York. Delgiudice, G. D., Mech, L. D., and Seal, U. S. (1989). Physiological assessment of deer populations by analysis of urine in snow. J. Wildl. Manage. 53, 284-291. Delgiudice, G. D., Kunkel, K. E., Mech, L. D., and Seal, U. S. (1990). Minimizing capturerelated stress on white-tailed deer with a capture collar. J. Wildl. Manage. 54, 299-303. Delgiudice, G. D., Seal, U. S., and Mech. L. D. (1991). Indicators of severe undernutrition in urine of free-ranging elk during winter. Wildl. Soc. Bull. 19, 106-110. den Boer, P. J. (1990). On the stabilization of animal numbers. Problems of testing. 3. What do we conclude from significant test results? Oecologia 83, 38-46. den Boer, P. J. (1991). Seeing the trees for the wood: Random walks or bounded fluctuations of population size? Oecologia 86, 484-491. Dennis, B., and Taper, M. L. (1994). Density dependence in time series observations of natural populations: Estimation and testing. Ecol. Monogr. 64, 205-224. de Swart, R. L., Kluten, R. M. G., Huizing, C. J., Vedder. L. J., Reijnders, P. J. H., Visser. 1. K. G., UytdeHaag, F. G. C. M., and Osterhaus, A. D. M. E. (1993). Mitogen and antigen induced B and T cell responses of peripheral blood mononuclear cells from the harbour seal (Phoca vitulina). Vet. Inmunol. Immunopathol. 37, 217-230. de Swart, R. L.. Ross, P. S., Vedder, L. J., Timmerman. H. H., Heisterkamp, S. H.. van Loveren, H., Vos, J. G.. Reijnders, P. J. H., and Osterhaus, A. D. M. E. (1994). Impairment of immune function in harbour seals (Phoca vitulinu) feeding on fish from polluted waters. Ambio 23, 155-159. de Swart, R. L., Ross, P. S., Timmermann, H. H., Vos. H. W.. Reijnders, P. J. H., Vos. J . G., and Osterhaus. A. D. M. E. (1995a). Impaired cellular immune response in harbour seals (Phoca vitulina) feeding on environmentally contaminated herring. Clin. Exp. Immunol. 101,480-486. de Swart, R. L., Ross. P. S., Timmerman, H. H.. Hijman. W. C., de Ruiter. E. M., Liem, A. K. D., Brouwer, A,, van Loveren, H., Reijnders, P. J. H., Vos. J. G.. and Osterhaus, A. D. M. E. (199%). Short term fasting does not aggravate immunosuppression in harbour

BIOLOGICAL CONSERVATION AND STRESS

503

seals (Phoca vitulina) with high body burdens of organochlorines. Chemosphere 31,42894306. Dettling, A., and Pryce, C. R. (1996). Social and emotional development in captive Goeldi’s monkeys. Abstr. Congr. Int. Primatol. Soc., 16th, 088. Diamant, M., Van Wolfswinkel, L., Altorffer, B., and De Wied, D. (1993). Biotelemetry: Adjustment of a telemetry system for simultaneous measurements of acute heart rate changes and behavioral events in unrestrained rats. Physiol. Behav. 53, 1121-1 126. Dietz, R., Ansen. C. T., Have, P., and Heide-Jflrgensen, M. P. (1989). Mass deaths of harbor seals (Phoca vitulina) in Europe. Ambio 18, 258-264. Dilsaver, S. C., Overstreet, D. H., and Peck, J. A. (1992). Measurement of temperature in the rat by rectal probe and telemetry yields compatible results. Pharmacol., Biochem. Behav. 42,549-552. Duffus. D. A., and Dearden, P. (1990). Non-consumptive wildlife-oriented recreation: A conceptual framework. Biol. Conserv. 53, 21 3-232. Dufour. K. W., Ankney, C. D., and Weatherhead, P. J. (1993). Nonrepresentative sampling during waterfowl banding: Emphasis on body condition. J. Wildl. Manage. 57, 741-751. Dunbar, R. I. M. (1989). Reproductive strategies in female gelada baboons. In “The Sociobiology of Sexual and Reproductive Strategies” (0.A. E. Rasa. C. Vogel, and E. Voland, eds.), pp. 74-92, Chapman & Hall, London. Dunlap, K. D., and Wingfield, J. C. (1995). External and internal influences on indices of physiological stress: I. seasonal and population variation in adrenocortical secretion of free-living lizards, Sceloporus occidentalis. J. Exp. Zool. 271, 36-46. Dunn. A. J., Powell, M. L.. Moreshead, W. V., Gaskin, J. M., and Hall, N. R. (1987). Effects of Newcastle disease virus administration to mice on the metabolism of cerebral biogenic amines, plasma corticosterone. and lymphocyte proliferation. Brain, Behav., Immunol. 1, 216-230. Dunn, A. J., Powell, M. L., Meitin, C., and Small, P. A. (1989). Virus infection as a stressor: Influenza virus elevates plasma concentrations of corticosterone, and brain concentrations of MHPG and tryptophan. Physiol. Behav. 45,591-594. Dunnet, G. M. (1982). Oil pollution and seabird populations. Philos. Trans. R. Soc. London, Ser B 297,413-427. Durchfeld, B., Baumgartner, W., Herbst, W., and Brahm, R. (1990). Vaccine-associated canine distemper infection in a litter of African hunting dogs (Lycaon pictus). Zentralbl. Veterinarmed., Reihe, B 37, 203-212. East, M. L. (1996). Survivorship in African wild dogs. Conserv. Biol. 10, 313-313. Echternkamp, S. E. (1984). Relationship between luteinizing hormone and cortisol in acutely stressed beef cows. Theriogenology 22, 305-312. Eisa, M., Osman, 0.M., Karrar, A. E., Abdel Rahim, A. H. (1980). An outbreak of bluetongue in sheep in the Sudan. Vet. Rec. 106,481-482. Ellis, G. M., and Trites, A. W. (1992). The RAM-packs came back: A method for attaching and recovering pinniped data recorders. A q u a Mamm. 18, 61-64. Enzenbacher, D. J. (1993). Tourists in Antarctica: Numbers and trends. Tourism Manage. 14, 142-146. Escbs, J . M.. Alados, C. L., and Emlen, J . M. (1995). Fractal structures and fractal functions as disease indicators. Oikos 74, 310-314. Eskens, U., Kaleta. E. F., and Unger, G. (1994). Eine Herpesvirus-bedingte EnzootiePachecosche Papageienkrankheit-in einem Psittazidenbestand. Tieriirztl. Praxis 22, 542-553. Etter, R. J. (1988). Physiological stress and color polymorphism in the intertidal snail Nucella lapillus. Evolution (Lawrence, Kans.) 42, 660-680.

504

HERIBERT HOFER AND MARION L. EAST

Euker, J. S., Meites, J.. and Riegle, G. D. (1975). Effects of acute stress on serum LH and prolactin in intact. castrate and dexamethasone-treated male rats. Endocrinology (Baltimore) 96, 85-92. Evans, P. G. H. (1987). “The Natural History of Whales and Dolphins.” Christopher Helm. London. Evans, P. G. H. (1996). Human disturbance of cetaceans. I n “The Exploitation of Mammal Populations” (V. J. Taylor and N. Dunstone, eds.), pp. 376-394. Chapman & Hall. London. Evans, P. G. H., Carson, Q., and Fisher, P. (1994). A study of the reactions of harbour porpoises to various boats in the coastal waters of S. E. Shetland. I n ”European Research on Cetaceans 8” (P. G. H. Evans, ed.), pp. 60-64. European Cetacean Society, Cambridge. UK. Ewing. D. E.. Clark, W. R.. and Vohs, P. A. (1994). Evaluation of implanted radio transmitters in pheasant chicks. J. Iowa Acad. Sci. lOl(3-4), 86-90. Fang, L. S., Chen, Y.W., and Chen, C. S. (1991). Feasibility of using ATP as an index for environmental stress on hermatypic coral. Mar. E d . : Progr. Ser. 70, 257-262. Farah. I. 0. (1996). Stress related enterotyphylocolitis in baboons. Abstr., Congr. In?. Primatol. Soc.. 16th p. 642. Faulkes, C. G., and Abbott. D. H. (1997). The physiology of a reproductive dictatorship: Regulation of male and female reproduction by a single breeding female in colonies of naked mole rats. In “Cooperative Breeding in Mammals” (N. J. Solomon and J. A. French, eds.), pp. 302-334. Cambridge University Press, Cambridge, UK. Felsburg, P. J., Keyes, L. L., Krawiec, D. R., and Rubin, S. I. (1986). The effect of general anesthesia on canine lymphocyte function. Vet. Immitnol. Immunopnthol. 13, 63-70. Fernandez. C., and Azkona, P. (1993). Human disturbance affects parental care of marsh harriers and nutritional status of nestlings. J. Wildl. Manage. 57, 602-608. Fescharek, R., Franke. V., and Samuel, M. R. (1994). Do anaesthetics and surgical stress increase the risk of post-exposure rabies treatment failure? Vaccine 12, 12-13. Fleming, M. R. (1985). The thermal physiology of the mountain pygmy-possum. Burramys parvics (Marsupialia: Burramyidae). Aust. Mammal. 8, 79-90. Fletcher, J . L., and Busnel, R. G. (1978). “Effects of Noise on Wildlife.” Academic Press, New York. Forstad, I . , Nilssen. A. C., Halvorsen, O., and Andersen, J. (1991). Parasite avoidance: The cause of post-calving migrations in Rangifer? Can. J. 2001.69,2423-2429. Fossi, M. C., Marsili, L., Leonzio, C., Notarbartolo, D. I., Sciara, G . , Zanardeli, M., and Focardi, S. (1992). The use of non-destructive biomarker in rnediterranean cetaceans: Preliminary data on MFO activity in skin biopsy. Mar. Pollut. Bull. 24, 459-461. Foster, B. R., and Rahs, E. Y. (1983). Mountain goat response to hydroelectric exploration in northwestern British Columbia. Environ. Manage. 7, 189-197. Fowler, G . S.. Wingtield, J. C., and Boersma, P. D. (1995). Hormonal and reproductive effects of low levels of petroleum fouling in Magellanic penguins (Spheniscus magellanicus). Auk 112,382-389. Frankel. 0. H., and SoulC, M. E. (1981). “Conservation and Evolution.” Cambridge University Press, Cambridge, UK. Fraser. A. F., and Broom, D. M. (1990). “Farm Animal Behaviour and Welfare.” Baillikre Tindall, London. Fraser, D., Ritchie, J. S. D., and Fraser, A. F. (1975). The term ‘stress’ in a veterinary context. Br. Vet. J. 131, 653-662. Frederick, P. C. (1986). Parental desertion of nestlings by white ibis (Eudocirnus albus) in response to muscle biopsy. J. Field Omithol. 57, 168-173.

BIOLOGICAL CONSERVATION AND STRESS

505

Frederick, R. B., Clark, W. R., and Klaas, E. E. (1987). Behavior energetics and management of refuging waterfowl: A simulation model. Wildl. Monogr. 96, 1-35. Freeman, B. M. (1985). Stress and the domestic fowl: Physiological fact or fantasy? World’s Poult. Sci. J. 41, 45-51. French. J. A. (1997). Proximate regulation of singular breeding in callitrichid primates. In “Cooperative Breeding in Mammals” (N. S. Solomon and J. A. French, eds.), pp. 34-75. Cambridge University Press, Cambridge, UK. Frerking, H.. Kramer, R., Schiele, R., and Roder, B. (1995). Klinische Erfahrungen mit Paramunitats-inducern beim Rind anlablich von Tierschauen. Dsch. Tierarztl. Wochenschr. 102, 188-189. Fretwell. S. D. (1 972). “Populations in a Seasonal Environment.” Princeton University Press, Princeton. NJ. Frisch. J. E. (1981). Changes occurring in cattle as a consequence of selection for growth rate in a stressful environment. J. Agric. Sci. 96, 23-38. Futuyma, D. J. (1983). Interspecific interactions and the maintenance of genetic diversity. In “Genetics and Conservation” (C. M. Schonewald-Cox, S. M. Chambers, B. MacBryde, and L. Thomas, eds.), pp. 364-373. BenjaminKummings, London. Gaisler. H., Hanak, V., and Horocek, I. (1981). Remarks on the current status of bat populations in Czechoslovakia. Myotis 18-19, 68-75. Gammonley, J. H., and Kelley, J. R. (1995). Effects of back-mounted ratio packages on breeding wood ducks. J . Field Ornithol. 65, 530-533. Garrott. R. A., Bartmann, R. M.. and White, G. C. (1985). Comparison of radio-transmitter packages relative to deer fawn mortality. J. Wildl. Manage. 49, 758-759. Gartner, K., Biittner, D.. Dohler, K., Friedel, R., Lindena, J., and Trautschold, I. (1980). Stress response of rats to handling and experimental procedures. Lab. Anim. 14,267-274. Geiser. F., and Broome, L. S. (1991). Hibernation in the mountain pygmy-possum Burramys parvus (Marsupialia). J . Zool. 223, 593-602. Geiser. F., and Broome, L. S. (1993). The effect of temperature on the pattern of torpor in a marsupial hibernator. J. Comp. Physiol. B 163, 133-137. Geiser, F.. Sink. H. S., Stahl, B., Mansergh, I., and Broome, L. S. (1990). Differences in the physiological response to cold in wild and laboratory-bred mountain pygmy-possums, Burramys parvus (Marsupialia). Aust. Wildl. Res. 17, 535-539. Gesser, R. M., Valyi-Nagy, T.. Altschuler, S. M., and Fraser, N. W. (1994). Oral-oesophageal inoculation of mice with herpes simplex virus type 1 causes latent infection of the vagal sensory ganglia (nodose ganglia). J. Gen. Virol. 75, 2379-2386. Gill, J. A., Sutherland, W. J., and Watkinson, A. R. (1996). A method to quantify the effects of human disturbance on animal populations. J. Appl. Ecol. 33, 786-792. Ginther, A. J.. and Washabaugh, K. F. (1996). Non-invasive measurement of scrota1 width in captive cotton-top tamarins (Suguinus o. oedipus): A tool for evaluating development. Abstr., Congr. Int. Primatol. Soc. 16th, 114. Ginzburg, L. R.. Ferson, S., and Akqakaya, H. R. (1990). Reconstructibility of density dependence and the conservative assessment of extinction risks. Conserv. Biol. 4, 63-70. Goldberg, J. S., and Haas, W. (1978). Interactions between mule deer dams and their radiocollared and unmarked fawns. J. Wildl. Manage. 42,422-425. Gonzalez, A. S., Echandia, E. L. R.,Cabrera, R., and Foscolo. M. R. (1994). Neonatal chronic stress induces subsensitivity to chronic stress in adult rats: 11. Effects on estrous cycle in females. Physiol. Behav. 56, 591-595. Goodrich, J. M., and Berger, J. (1994). Winter recreation and hibernating black bears llrsus umericunus. Biol. Conserv. 67, 105-110. Goodwin, H. (1996). In pursuit of ecotourism. Biodiversity Conserv. 5, 277-291.

506

HERIBERT HOFER AND MARION L. EAST

Gottman, J. M.. and Roy. A. K. (1990). “Sequential Analysis. A Guide for Behavioral Researchers.” Cambridge University Press, Cambridge, UK. Graham, L. H.. and Brown, J. L. (1996). Cortisol metabolism in the domestic cat and implications for non-invasive monitoring of adrenocortical function in endangered felids. Zoo Biol. 15, 71-82. Grandin, T. (1987). Animal handling. Vet. Clin. North Am. Food Anim. Pract. 3, 323-338. Grandin. T., Rooney, M. B., Phillips, M., Cambre, R. C., Irlbeck. N. A,, and Graffam, W. (1995). Conditioning of nyala (Tragelaphiis nngasi) to blood sampling in a crate with positive reinforcement. Zoo B i d . 14, 261-273. Green, J . S., Golightly, R. T., Lindsey, S. L., and Leamaster, B. R. (1985). Use of radio transmitter implants in wild canids. Great Basin Nat. 45, 567-570. Green, R. H. (1979). “Sampling Design and Statistical Methods for Environmental Biologists.” Wiley, New York. Greenberg, N.. and Wingfield. J. C. (1987). Stress and reproduction: Reciprocal relationships. In “Reproductive Endocrinology of Fishes, Amphibians and Reptiles” (D. 0.Norris and R. E. Jones. eds.), pp. 461-503. Wiley, New York. Greenwood, R. B., and Sargeant, A. B. (1973). Influence of radio packs on captive mallards and blue-winged teal. 1. Wildl. Manage. 47, 369-375. Gremillet, D., Schmid. D.. and Culik, B. M. (1995). Energy requirements of breeding great cormorants. Phalacrocorax carbo sinensis. Mar. Ecol.: Prog. Ser. 121, 1-9. Griffiths, H. 1. (1994). The effects upon badgers (Meles meles) of the activities of a single, persistent poacher. Anim. Welfare 3, 219-225. Gustafson. A. W.. and Belt. W. D. (1981). The adrenal cortex during activity and hibernation in the male little brown bat, Myoris hccifugiis htcifilgus: Annual rhythm of plasma cortisol levels. Gen. Comp. Endocrinol. 44, 269-278. Hall, A., and Harwood, J. (1990). “The Intervet Guidelines to Vaccinating Wildlife.” Intervet, Cambridge, UK. Hannon, S. J., and Wingfield. J. C. (1990). Endocrine correlates of territoriality, breeding stage, and body molt in free-living willow ptarmigan of both sexes. Can. J. 2001.68,2130-2134. Hansbrough, J. F., Zapata-Sirvent. R. L., Bartle, E. J., Anderson, J. K., Elliott, L., Mansour, M. A., and Carter, W. H. (1985). Alterations in spleenic lymphocyte subpopulations and increased mortality from sepsis following anesthesia in mice. Anesthesiology 63,267-273. Haour. F., Marquette. C., Ban, E., Crumeyrolle-Arias, M., Rostene, W.. Tsiang, H., and Fillion, G. (1995). Receptors for interleukin-I in the central nervous and neuroendocrine systems: Role in infection and stress. Ann. Endocrinol. 56, 173-179. Hare, S.. and Robertson, 0.H. (1959). Changes in plasma 17-hydroxycorticosteroids accompanying sexual maturation and spawning of the Pacific salmon (Oncorhynchirs tschawyrscha) and rainbow trout (Salmo gairdneri). Proc. Natl. Acad. Sci. U.S.A. 45, 886-893. Harlow. H. J., Thorne. E. T., Williams, E. S., Belden, E. L., and Gem, W. A. (1987a). Adrenal responsiveness in domestic sheep (Ovis aries) to acute and chronic stressors as predicted by remote monitoring of cardiac frequency. Can. J. Zool. 65,2021-2027. Harlow, H. J., Thorne, E. T., Williams, E. S., Belden, E. L., and Gern, W. A. (1987b). Cardiac frequency: A potential predictor of blood cortisol levels during acute and chronic stress exposure in Rocky Mountain bighorn sheep (Ovis canadensis). Can. J. Zool. 65, 20282034. Harlow, H. J.. Lindzey, F. G., Van Sickle, W D., and Gem, W. A. (1992). Stress response of cougars to nonlethal pursuit by hunters. Can. J. 2001.70, 136-139. Harper, D. G., Tornatzky. W.. and Miczek, K. A. (1996). Stress induced disorganization of circadian and ultradian rhythms: Comparisons of effects of surgery and social stress. Physiol. Behav. 59,409-41 9.

BIOLOGICAL CONSERVATION AND STRESS

507

Harrington, F. H., and Veitch, A. M. (1991). Short-term impacts of low-level jet fighter training on caribou in Labrador. Arctic 44,318-327. Harrington. F. H.. and Veitch, A. M. (1992). Calving success of woodland caribou exposed to low-level jet fighter overflights. Arctic 45, 213-218. Harthoorn, A. M. (1976). “The Chemical Capture of Animals.” Bailliere Tindall. London. Harthoon, A. M., van der Walt, K.. and Young, E. (1974). Possible therapy for capture myopathy in captured wild animals. Nature (London) 274,577-577. Harwood, J., and Grenfell, B. (1990). Long term risks of recurrent seal plagues. Mar. Polltit. Bitll. 21, 284-287. Hegner, R. E., and Wingfield. J. C. (1987). Effects of brood-size manipulations on parental investment. breeding success and reproductive endocrinology of house sparrows. Auk 104,470-480. Heide-Jorgensen, M. P.. Harkonen. T., Dietz, R., and Thompson, P. M. (1992). Retrospective of the 1988 European seal epizootic. Dis. A q u a . Org. 13,37-62. Heinsohn, R. (1992). When conservation goes to the dogs. Trends Ecol. Evol. 7, 214-215. Helb, H. W., and Huppop, 0. (1991). Herzschlagrate als Map zur Beurteilung des Einflusses von Storungen vei Vogeln. In “Ornithologenkalender 1992” (E. Bezzel, P. H. Barthel. H. H. Bergmann. H. W. Helb. and K. Witt, eds.), pp. 217-230. A d a , Wiesbaden. Hemsworth, P. H., and Barnett, J. L. (1991). The effects of aversively handling pigs either individually or in groups on their behavior, growth and corticosteroids. Appl. Anim. Behav. Sci. 30, 61-72. Henry, J. P.. and Stephens, P. M. (1977). “Stress, Health, and the Social Environment.” Springer, New York. Herbst, L. (1991). Pathological and reproductive effects of intraperitoneal telemetry devices on female armadillos. J. Wildl. Manage. 55, 628-631. Heyl, C. W., Bigalke, R. C., and Pepler, D. (1988). Captive rearing of the Cape francolin and prospects for stocking. S. Afr. J . Wildl. Res. 18, 22-29. Higuchi, T., Negoro, H., and Arita, J. (1989). Reduced responses of prolactin and catecholamine to stress in the lactating rat. J. Endocrinol. 122, 495-498. Hofer, H., East. M. L., and Campbell, K. L. I. (1993). Snares. commuting hyaenas, and migratory herbivores: Humans as predators in the Serengeti. Symp. Zool. Soc. London 65,347-366. Hofer, H., Campbell, K. L. I.. East, M. L., and Huish. S. A. (1996). The impact of game meat hunting on target and non-target species in the Serengeti. In “The Exploitation of Mammal Populations” (V. Taylor and N. Dunstone, eds.). pp. 117-146. Chapman & Hall, London. Hoffmann, A. A., and Blows, M. W. (1993). Evolutionary genetics and climate change: Will animals adapt to global warming? In “Biotic Interactions and Global Change” (P. M. Kareiva, J. G. Kingsolver, and R. B. Huey, eds.), pp. 165-178. Sinauer, Sunderland, MA. Hoffmann. A. A., and Parsons, P. A. (1989). Selection for increased desiccation resistance in Drosophila melanogaster: additive genetic control and correlated responses for other stresses. Generics 122, 837-846. Hoffmann, A. A.. and Parsons, P. A. (1991). “Evolutionary Genetics and Environmental Stress.” Oxford University Press, Oxford. Holloway, G. J., Sibly, R. M., and Povey, S. R. (1990). Evolution in toxin-stressed environments. Funct. Ecol. 4, 289-294. Holmes, J. C.. and Zohar, S. (1990). Pathology and host behaviour. In “Parasitism and Host Behaviour” (C. J. Barnard and J. M. Behnke, eds.). pp. 34-63. Taylor & Francis, London. Hoover, J. P., Bahr. R. J., Nieves, M. A., Doyle, R. T., Zimmer, M. A,, and Lauzon, S. E. (1985). Clinical evaluation and prerelease management of American river otters in the second year of a reintroduction study. J. Am. Vet. Med. Assoc. 187, 1154-1161.

508

HERIBERT HQFER A N D MARION L. EAST

Houston, A. I.. and McNamara. J. M. (1993). A theoretical investigation of fat reserves and mortality levels in small birds in winter. Ornis Scand. 24, 205-219. Houston, R. A., and Greenwood, R. J. (1993). Effects of radio transmiters on nesting captive mallards. J. Wildl. Manage. 57, 703-709. Hoysak. D. J. and Weatherland, P. J. (1991). Sampling blood from birds: A technique and assessment of its effect. Condor 93, 746-752. Huffaker, C. B. (1958). Experimental studies of predation: Dispersion factors and predatorprey interactions. Hilgardia 27, 343-383. Hughes, C. S., Gaskell, R. M., Jones, R. C., Bradbury, J. M., and Jordan. F. T. (1989). Effects of certain stress factors on the re-excretion of infectious laryngotracheitis virus from latently infected carrier birds. Res. Vet. Sci. 46, 274-276. Huppop. O., and Hagen, K. (1990). Der Einfluss von Storungen auf Wildtiere am Beispiel der Herzschlagrate brutender Austernfischer Haematopus ostralegus. Vogelwarfe35,301-310. Hurlbert, S. H. (1984). Pseudoreplication and the design of ecological field experiments. Ecol. Monogr. 54,187-21 1. Hurst, J. L., Barnard, C. J.. Hare. R., Wheeldon, E. B., and West, C. D. (1996). Housing and welfare in laboratory rats: Time-budgeting and pathophysiology in single-sex groups. Anim. Behav. 52,335-360. Ingold, P., Huber, B., Mainini, B., Marbacher, H., Neuhaus, P., Rawyler, A., Roth, M., Schnidrig, R., and Zeller, R. (1992). Freizeitaktivitaten-ein gravierendes Problem fur Tiere? Ornithol. Beob. 89, 205-216. Ingold, P., Schnidrig-Petrig, R., Marbacher, H., Pfister, U., and Zeller, R.(1996). “Tourismusl Freizeitsport und Wildtiere im Schweizer Alpenraum.” Schriftenr. Umwelt 262, Bundesamt fur Umwelt, Wald iind Landschaft, Bern, Switzerland. Innes, A. J., El Haj, A. J., and Gobin, J. F. (1986). Scaling of the respiratory cardiovascular and skeletal muscle systems of the freshwater-terrestrial mountain crab Pseudorhelphusa garmani garmani. J. Zool., Ser. A 209,595-606. Insel, T. R., Kinsley, C. H., Mann, P. E., and Bridges, R. S. (1990). Prenatal stress has longterm effects on brain opiate receptors. Brain Res. 511, 93-97. International Fund for Animal Welfare (IFAW). (1996). “Report of the International Workshop on the Special Aspects of Watching Sperm Whales, Roseau, Commonwealth of Dominica, East Caribbean 8th January-11th January 1996.” Int. Whal. Comm., Sci. Comm. Doc. SC/48/0 26. IFAW, Crowborough, East Sussex, UK. International Fund for Animal Welfare (IFAW), Tethys Research Institute and Europe Conservation. (1996). “Report of the Workshop on the Scientific Aspects of Managing Whale Watching, Montecastello di Vibio, Italy.” IFAW, Crowborough, East Sussex, UK. Ivanovici, A. M., and Wiebe, W. J. (1981). Towards a working “definition” of ‘stress’: A review and critique. I n “Stress Effects on Natural Ecosystems” (G. W. Barrett and R. Rosenberg, eds.), pp. 13-27. Wiley, New York. Jacobson, S. K., and Lopez, A. F. (1994). Biological impacts of ecotourism: Tourists and nesting turtles in Tortuguero National Park, Costa Rica. Wildl. SOC. Bull. 22, 414-419. Janik, V. M., and Thompson, P. M. (1996). Changes in surfacingpatterns of bottlenose dolphins in response to boat traffic. Mar. Mammal. Sci. 12, 597-602. Jarvi, T. (1990). Cumulative acute physiological stress in Atlantic salmon smolts: The effect of osmotic imbalance and the presence of predators. Aquaculfure 89, 337-350. Jarvi, T., and Uglem, I. (1993). Predator training improves the anti-predator behaviour of hatchery reared Atlantic salmon (Salmo salar) smolt. Nord. J. Freshwater Res. 68,63-71. Jenkins. F. J., and Baum, A. (1995). Stress and reactivation of latent herpes simplex virus-a fusion of behavioral medicine and molecular biology. Ann. Behav. Med. 17, 116-123.

BIOLOGICAL CONSERVATION AND STRESS

509

Jenssen, B. M., Skaare, J. U., Ekker. M., von Graven, D., and Silverstone, M. (1994). Blood sampling as a non-destructive method for monitoring levels and effects of organochlorines (PCB and DDT) in seals. Chemosphere 28, 3-10. Jenssen, B. M., Skaare, J. U., Woldstad, S., Nastad, A. T.. Haugen, 0..Kloven, B., and Sormo, E. G. (1995). Biomarkers in blood to assess effects of polychlorinated biphenyls in freeliving grey seal pups. Dev. Mar. Bid. 4, 607-615. Jeppesen, J. L. (1987). Impact of human disturbance on home range, movements and activity of red deer Cervus eluphus in a Danish environment. Dan. Rev. Game Biol. 13(2), 1-38. Jessup, D. A. (1992). Veterinary contributions toward improving capture, medical management, and anesthesia of free-ranging wildlife. J. Am. Vet. Med. Assoc. 200, 653-658. Johnson, K., Dalton. R., and Burley, N. (1993). Preferences of female American goldfinches Carduelis tristis for natural and artificial male traits. Behuv. Ecol. 4, 138-143. Jones, C. A,, Refsal, K. R., Lippert, A. C., Nachreiner, R. F., and Schwacha, M. M. (1990). Changes in the adrenal cortisol secretion as reflected in the urinary cortisollcreatinine ratio in dogs. Domest. Anim. Endocrinol. 7, 559-572. Jones, T. O.,Bourne, N. F., Bower, S. M., and Iwama, G. K. (1993). Effect of repeated sampling on haemolymph pH, partial pressure of oxygen and haemocyte activity in the pacific oyster Crussostreu gigus Thunberg. J. Exp. Mar. Biol. Ecol. 167, 1-10, Jordan, R. H., and Burghardt, G. M. (1986). Employing an ethogram to detect reactivity of black bears (Ursus americanus) to the presence of humans. Ethology 73,89-115. Joubert. F. G., and Stander, P. E. (1990). Capture myopathy in an African lion. Madoquu 17,51-52. Jurke, M. H., Pryce, C. R., Doebeli, M., and Martin, R. D. (1994). Non-invasive detection and monitoring of pregnancy and the postpartum period in Goeldi’s monkey (Cullimico Am. J. Primutol. 34, 319-331. goeldii) using urinary pregnanediol-3-alpha-glucuronide. Kania, W.(1992). Safety of catching adult European birds at the nest. The Ring 14,5-50. Kareiva, P. M., Kingsolver, J. G., and Huey, R. B. (1993). “Biotic Interactions and Global Change.” Sinauer, Sunderland, MA. Karr, J. R. (1990). Avian survival rates and the extinction process on Barro Colorado Island. Panama. Conserv. Biol. 4, 391-397. Kaufmann, J. H. (1983). On the definitions and functions of dominance and territoriality. Biol. Rev. Cambridge Philos. Soc. 58, 1-20. Keller, V. (1989a). Variations in the response of great crested grebes Podiceps cristutus to human disturbance: A sign of adaptation? Biol. Conserv. 49, 31-45. Keller, V. (1989b). Egg-covering behaviour by great crested grebes Podiceps cristutus. Ornis Scand. 20, 129-131. Kelley, T. M., and Bramblett, C. A. (1981). Urine collection from vervet monkeys by instrumental conditioning. Am. J. Primutol. 1, 95-97. Kennedy. S., Smyth, J. A,. Cush. P. F., McCullough, S. J., Allan, G. M., and McQuaid. S. (1988). Viral distemper now found in porpoises. Nature (London) 336, 21-21. Kenward, R. (1987). “Wildlife Radio Tagging.” Academic Press, London. Kernohan. B. J., Jenks, J. A,, Naugle. D. E.. and Millspaugh. J. J. (1996). Estimating 24-h habitat use patterns of white-tailed deer from diurnal use. J. Environ. Munuge. 48,299-303. Kestin, S. C. (1995). Welfare aspects of the commercial slaughter of whales. Anim. Welfare 4, 11-27. Keverne, E. B., Meller, R. E., and Eberhardt. J. A. (1982). Dominance and subordination: Concepts or physiological states? I n “Advanced Views on Primate Biology” (0.Chiarelli, ed.), pp. 81-94. Springer, New York. Kiecolt-Glaser. J. K., Glaser, R., Gravenstein, S., Malarkey, W. B., and Sheridan, J. (1996). Chronic stress alters the immune response to influenza virus vaccine in older adults. Proc. Nutl. Acud. Sci. U.S.A.93, 3043-3047.

510

HERIBERT HOFER AND MARION L. EAST

Kihlstrom, J . E.. Olsson, M., Jensen, S., Johansson, A,. Ahlborn, J., and Bergman, A. (1992). Effect of PCB and different fractions of PCB on the reproduction of the mink (Mustela vison). Ambio 21, 563-569. King, D. A.. and Stewart, W. P. (1996). Ecotourism and comodification: Protecting people and places. Biodiversity Conserv. 5, 293-305. Kirschbaum, C., and Hellhammer, D. H. (1994). Salivary cortisol in psychoneuroendocrine research: Recent developments and applications. Psychoneuroendocrinolo~y19,313-333. Klomp, N. I., and Furness, R. W. (1992). Non-breeders as a buffer against environmental stress: Declines in number of great skuas on Foula, Shetland, and prediction of future recruitment. J. Appl. Ecol. 29, 341-348. Knight, R. L., and Gutzwiller. K. J. (1995). “Wildlife and Recreationists.” Island Press, Washington, DC. Knox, C. M., Hattingh, J.. and Raath, J. P. (1990). The effect of tranquilizers on the immediate responses to repeated capture and handling of boma-kept impala. Comp. Biochem. Physiol. C. 95C, 247-251. Kock, M. D.. Jessup, D. A., Clark, R. K., Franti, C. E., Weaver, R. A. (1987a). Capture methods in five subspecies of free-ranging bighorn sheep: An evaluation of drop-net, drive-net, chemical immobilization and the net-gun. J. Wildl. Dis. 23, 634-640. Kock. M. D., Clark, R. K., Franti, C. E., Jessup, D. A,, and Wehausen, J. D. (1987b). Effects of capture on biological parameters in free-ranging bighorn sheep Ovis canadensis: Evaluation of normal. stressed and mortality outcomes and documentation of postcapture survival. J. Wildl. Dis. 23, 652-662. Kock, M. D., Morkel, P.. Atkinson. M., and Foggin, C. (1995). Chemical immobilization of free-ranging white rhinoceros (Ceratotherium simum simum) in Hwange and Matobo National Parks. Zimbabwe, using combinations of etorphine (M99), fentanyl, xylazine, and detomidine. J. Zoo Wildl. Med. 26, 207-219. Kohane, M. J.. and Parsons. P. A. (1988). Domestication. Evolutionary change under stress. Evol. Biol. 23, 31-48. Kohn, M., Knauer, F., Stoffella, A,, Schroder, W., and Paabo, S. (1995). Conservation genetics of the European brown bear: A study using excremental PCR of nuclear and mitochondria1 sequences. Mol. Ecol. 4, 95-103. Kooijman, S. A. L. M., van der Hoeven. N., and van der Werf, D. C. (1989). Population consequences of a physiological model for individuals. Funct. Ecol. 3, 325-336. Korpimaki, E. (1990). Body mass of breeding Tengmalm’s owls Aegolius funereus: Seasonal, between-year, site and age-related variation. Ornis Scand. 21, 169-178. Kortner, G., and Geiser, F. (1995). Effect of photoperiod and ambient temperature on activity patterns and body weight cycles of mountain pygmy-possums, Burramysparvus (Marsupiaha). J. Zool. 235, 311-322. Kramer, K., van Acker, S. A,, Voss, H. P., Grimbergen, J. A,, and van der Vijgh, W. J., and Bast. A. (1993). Use of telemetry to record electrocardiogram and heart rate in freely moving mice. J. Pharmacol. Toxicol. Methods 30, 209-215. Kramskaya,T. A,, Chetverikova, L. K., Frolov, B. A,, and Polyak, R. Y. A. (1991). Disturbance of resistance to a repeated infection by homological influenza virus in mice after stress. Vestn. Akad. Med. Nauk SSSR 4, 47-50. Krebs, J. R., and Davies. N. B. (1997). “Behavioural Ecology.” Blackwell, Oxford. Krebs, R. A.. and Loeschcke, V. (1994). Response to environmental change: Genetic variation and fitness in Drosophila buzzarii following temperature stress. In “Conservation Genetics” (V. Loeschcke, J. Tomiuk, and s. K. Jain, eds.). pp. 309-321. Birkhauser, Basel and New York.

BIOLOGICAL CONSERVATION AND STRESS

511

Kreeger, T. J., Faggella, A. M.. Seal, U. S., Mech, L. D., Callahan, M., and Hall, B. (1987). Cardiovascular and behavioral responses of gray wolves to ketamine-xylazine immobilization and antagonism by yohimbine. J. Wildl. Dis. 23, 463-470. Kreeger, T. J., Seal, U. S., Tester, J. R., Callahan, M., and Beckel, M. (1990). Physiological responses of red foxes Vulpes viclpes to surgery. J. Wildl. Dis. 26, 219-224. Kreger, M. D., and Mench, J. A. (1993). Physiological and behavioral effects of handling and restraint in the ball python (Pyrhon regius) and the blue-tongued skink (Tiliqua scincoides). Appl. Anim. Behav. Sci. 38, 323-336. Krylov, V. 1. (1990). Ecology of the Caspian seal. Finn. Game Res. 47, 32-36. Kubokawa, K. (1993). Field endocrinology in Japan. J. Reprod. Dev. 39, j117-jl25 (in Japanese). Lamontagne, L., Sadi, L., and Joyal, R. (1989). Serological evidence of bovine herpesvirus 1-related virus infection in the white-tailed deer population on Anticosti Island Quebec, Canada. J. Wildl. Dis. 25, 202-205. Lampo, M., and Bayliss, P. (1996). Density estimates of cane toads from native populations based on mark-recapture data. Wildl. Res. 23, 305-315. Lamprecht, J., and Hofer, H. (1994). Cooperation among sunfish: Do they have the cognitive abilities? Anim. Behav. 47, 1457-1458. LaPatra. S. E., Groff, J. M., Jones, G. R., Munn, B., Patterson. T. L., Holt, R. A,, Hauck, A. K., and Hedrick, R. P. (1993). Occurrence of white sturgeon iridovirus infections among cultured white sturgeon in the pacific northwest. Aquaculture 126. Lash, G. Y. B., and Horwich, R. H. (1996). Successes and problems of community-managed ecotourism in protecting primates: A case study in Belize. Abstr., Congr. Int. Primatol. Soc.. 16th, 470. Laudenslager, M., Berger, C., Bocchia, M., Worlein, J., and Reite. M. (1996). Innate immunity in young macaques during brief social separation experiences. Absrr., Congr. Inf.Primatol. SOC., 16th, 757. Laurenson, M. K., and Caro, T. M. (1994). Monitoring the effects of non-trivial handling in free-living cheetahs. Anim. Behav. 47, 547-557. Lavigne, D. M., and Schmitz, 0.J. (1990). Global warming and increasing population densities: A prescription for seal plagues. Mar. Pollur. Bull. 21, 280-284. Lee, A. K., and Cockburn, A. (1985). “Evolutionary Ecology of Marsupials.” Cambridge University Press, Cambridge, UK. Lee, A. K., and McDonald, I. R. (1985). Stress and population regulation in small animals. Oxford Rev. Reprod. Biol., pp. 261-304. Lee, A. K., Bradley, A. J., and Braithwaite, R. W. (1977). Corticosteroid levels and male mortality in Antechinus stuarfii. In “The Biology of Marsupials” (B. Stonehouse and D. Gilmore, eds.), pp. 209-220. Macmillan, London. Le Maho, Y., Karmann, H., Briot, D., Handrich, Y., Robin, J. P., Mioskowski, E., Cherel, Y., and Farni, J. (1992). Stress in birds due to routine handling and a technique to avoid it. Am. J. Physiol. 263, R775-R781. Lemaire, V., and Mormkde, P. (1995). Telemetered recording of blood pressure and heart rate in different strains of rats during chronic social stress. Physiol. Behav. 58,1181-1 188. Lima, S. L. (1986). Predation risk and unpredictable feeding conditions: Determinants of body mass in birds. Ecology 67, 377-385. Line, S. W., Kaplan, J. R., Heise, E. R., Hilliard, J. K.. Cohen, S., Rabin, B. S.. and Manuck, S. B. (1996). Effects of social reorganization on cellular immunity in male cynomolgus monkeys. Am. J. Primatol. 39,235-249. Logan, C . A., and Wingfield, J. C. (1995). Hormonal correlates of breeding status, nest construction, and parental care in multiple-brooded northern mockingbirds, Mimum polyglotfos. Horm. Behav. 29, 12-30.

512

HERIBERT HOFER AND MARION L. EAST

Lombardi, C. M.. and Hurlbert. S. H. (1996). Sunfish cognition and pseudoreplication. Anini. Behav. 52,419-422. Lowenstine, L. J.. and Fry, D. M. (1985). Adenovirus-like particles associated with intranuclear inclusion bodies in the kidney of a common murre (Uria aalge). Avian Dis. 29, 208-213. Lubach. G. R., Coe. C. L., and Ershler, W. B. (1995). Effects of early rearing environment on immune responses of infant rhesus monkeys. Brain. Behav., imnirtn. 9, 31-46. Lubach, G. R., Karaszewski, J. W., and Schneider. M. L. (1996). Pregnancy conditions influence immune and behavioral development of rhesus monkey infants. Abstr., Congr. Int. Prinzato/. Soc., 16th. 103. Luttrell. L., Acker, L., Urben. M., and Reinhardt, V. (1994). Training a large troop of rhesus macaques to co-operate during catching: Analysis of the time investment. Anim. Welfare 3, 135-140. Lyman, C. P., Willis, J. S., Malan, A., and Wang, L. C. H. (1982). “Hibernation and Torpor in Mammals and Birds.” Academic Press, London. Macario, A. J. L. (1995). Heat-shock proteins and molecular chaperones: Implications for pathogenesis, diagnostics, and therapeutics. Itit. J . Clin. Lab. Res. 25, 59-70. MacArthur, R. A., Geist. V.. and Johnson, R. H. (1982). Cardiac and behavioural responses of mountain sheep to human disturbance. J . W i l d . Manage. 46,351-358. MacArthur, R. A., Geist, V., and Johnson, R. H. (1986). Cardiac responses of bighorn sheep to trapping and radio instrumentation. Can. J. 2001.64, 1197-1200. Mace, R. D., Waller. J. S., Manley, T. L.. Lyon, L. J., and Zuuring, H. (1996). Relationships among grizzly bears, roads and habitat in the Swan Mountains, Montana. J . Appl. Ecol. 33, 1395-1404. Machlis, L., Dodd, P. W. D., and Fentress. J. C. (1985). The pooling fallacy: Problems arising when individuals contribute more than one observation to the data set. Z . Tierpsychol. 68,201-214. Madej, A,, Forsberg. M., and Edqvist, L. E. (1992). Urinary excretion of cortisol and oestrone sulfate in pregnant mink fed PCB and fractions of PCB. Anzbio 21, 582-585. Madsen, J. (1995). Impacts of disturbance on migratory waterfowl. ibis 137, Suppl, S67-S74. Maestripieri D.. Badiani, A,, and Puglisiallegra S. (1991). Prepartal chronic stress increases anxiety and decreases aggression in lactating female mice. Behav. Neurosci. 105,663-668. Maestripieri, D., Schino. G., Aureli. F., and Troisi, A. (1992). A modest proposal: Displacement activities as an indicator of emotions in primates. Anim. Behav. 44,967-979. Mainini, B., Neuhaus, P., and Ingold, P. (1993). Behavior of marmots Marmota marmota under the influence of different hiking activities. B i d . Conserv. 64, 161-164. Mallet, J., Longino, J . T., Murawski, D., Murawski. A,. Simpson de Gamboa. A. (1987). Handling effects in Heliconiiis: Where do all the butterflies go? J. Aninz. Ecol. 56,377-386. Manly, B. F. J. (1992). “The Design and Analysis of Research Studies.” Cambridge University Press, Cambridge, UK. Mansergh, I. M., and Broome, L. (1994). “The Mountain Pygmy-possum of the Australian Alps.” New South Wales University Press, New Kensington. Australia. Mansergh, I. M.. and Scotts, D. J . (1989). Habitat continuity and social organization of the mountain pygmy-possum restored by tunnel. J . Wildl. Manage. 53,701-707. Markovitz, H. (1982). “Behavioral Enrichment in Zoos.” Van Nostrand-Reinhold, New York. Marsden. J. E., and May, B. (1984). Feather pulp: A non-destructive sampling technique for electrophoretic studies of birds. Artk 101, 173-175. Martel, F. L., Lyons, D. M., and Levine, S. (1996). Long-term consequences of postnatal stress: Behavioral and pituitary-adrenal responses in squirrel monkeys. Abslr., Congr. Int. Primarol. Soc., 16th, 128.

BIOLOGICAL CONSERVATION A N D STRESS

513

Masman. D.. Dijkstra. C., Daan, S.. and Bult, A. (1989). Energetic limitation of avian parental effort: Field experiments in the kestrel Falco rinnuncubcs. J. Evol. Biol. 2,435-455. Mason, G. J. (1991). Stereotypes: A critical review. Anim. Behav. 41, 1015-1037. Matt, K. S.. Soares. M. J., Talamantes. F., and Bartke, A. (1983). Effects of handling and ether anesthesia on serum prolactin levels in the golden hamster. Proc. Soc. Exp. B i d . Med. 173,463-466. May. R. M. (1995). The cheetah controversy. Nariire (London) 374, 309-310. Mayr. B.. Honig, A.. and Gutbrod, F. (1990). Untersuchungen iiber die Wirksamkeit und Unschadlichkeit einer Schutzimpfung gegen Parvovirus bzw. Tollwut bei narkotisierten Hundewelpen. Tieriirztl. Praxis 18, 165-169. Mays, N. A., Vleck. C. M.. and Dawson. J. W. (1991). Plasma luteinizing hormone, steroid hormones, behavioral role, and nest stage in the cooperatively breeding Harris’ hawk (Pambuteo itnicincrus). Auk 108, 619-637. McCarty. R. (1983). Stress, behaviour and the sympathetic-adrenal medullary system. In “Stress and Alcohol Use” (L. A. Pohorecky and J. Brick, eds.), pp. 7-22. Elsevier, New York. McConway. K. (1992). The number of subjects in animal behavior experiments: is Still right? In “Ethics in Research on Animal Behaviour” (M. S. Dawkins and L. M. Gosling, eds.), pp. 35-58. Academic Press, London. McCune. S. (1995). The impact of paternity and early socialization on the development of cats’ behavior to people and novel objects. Appl. Anim. Behav. Sci. 45, 109-124. McDonald. I. R.. Lee. A. K.. Than, K. A,. and Martin, R. W. (1986). Failure of glucocorticoid feedback in males of a population of small marsupials (Antechinus swainsonii) during the period of mating. J. Endocrinol. 108, 63-68. McLean, R. G. (1975). Racoon rabies. I n “The Natural History of Rabies 2” (G. M. Baer. ed.). pp. 53-77. Academic Press, New York. McLellan, B. N.. and Shackleton, D. M. (1988). Grizzly bears and resource-extraction industries: Effects of roads on behaviour. habitat use and demography. J. Appl. Ecol. 25, 45 1-460. McLellan, B. N., and Shackleton, D. M. (1989). Grizzly bears and resource-extraction industries: Habitat displacement in response t o seismic exploration, timber harvesting and road maintenance. J . Appl. Ecol. 26,371-380. McNamara, J. M., and Houston, A. I. (1992). State-dependent life history theory and its implications for optimal clutch size. Evol. Ecol. 6, 170-185. McNamara, J . M., and Houston, A. I. (1996). State-dependent life-histories. Nature (London) 380, 215-221. Mead, R. (1991). “The Design of Experiments. Statistical Principles for Practical Application.” Cambridge University Press, Cambridge, UK. Meaney. M. J., Aitken, D. H., Bhatnagar, S., and Sapolsky, R. M. (1991). Postnatal handling attenuates certain neuroendocrine, anatomical. and cognitive dysfunctions associated with aging in female rats. Neurohiol. Aging 12, 31-38. Meaney, M. J., Bhatnagar. S., Diorio, J.. Larocque, S.. Francis. D., O’Donnell. D.. Shanks, N.. Sharma. S., Smythe, J., and Viau. V. (1993). Molecular basis for the development of individual differences in the hypothalamic-pituitary-adrenal stress response. Cell. Mol. Neurobiol. 13, 321 -347. Mech, L. D.. Seal. U. S.. and Delgiudice. G. D. (1987). Use of urine in snow to indicate condition of wolves. J . Wildl. Manage. 51, 10-13. Medleau. L., Crowe, D. T., and Dawe, D. L. (1983). Effect of surgery on the in vitro response of canine peripheral blood lymphocytes t o phytohemagglutinin. Am. J. Vet. Res. 44,859-860.

514

HERIBERT HOFER AND MARION L. EAST

Mees, J., and Reijnders. P. J. H. (1994). The harbour seal, Phoca vitulina, in the Oosterschelde: Decline and possibilities for recovery. Hydrohiolagy 282-283, 547-555. Meire, P. M., and Dereu, J. (1990). Use of the abundancelbiomass comparison method for detecting environmental stress: Some considerations based on intertidal macrozoobenthos and bird communities. J. Appl. Ecol. 27, 210-223. Mendoza, S. P., and Mason, W. A. (1986). Contrasting responses t o intruders and to involuntary separation by monogamous and polygynous New World monkeys. Physiol. Behav. 38, 795-801. Merkle, M. S., and Barclay, R. M. R. (1996). Body mass variation in breeding mountain bluebirds. Sialia currucoides: Evidence of stress or adaptation for flight? J . Anim. Ecol. 65,401-413. Merola, M. (1994). A reassessment of homozygosity and the case for inbreeding depression in the cheetah. Acinonyxjrbarus: Implications for conservation. Conserv. Biol. 8,961-971. Meyers, T. R.. Hauck, A. K., Blankenbeckler, W. D., and Minicucci, T. (1986). First report of viral erythrocytic necrosis in Alaska, USA, associated with epizootic mortality in pacific herring Chipea harengus pallasi Valenciennes. J. Fish Dis. 9, 479-491. Miller, M. W., Thompson, Hobbs. N., and Sousa, M. C. (1991). Detecting stress responses in Rocky Mountain bighorn sheep (Ovis canadensis canadensis): Reliability of cortisol concentrations in urine and feces. Can. J. 2001.69, 15-24. Miller, P. S. (1994). Is inbreeding depression more severe in a stressful environment? Zoo Biol. 13, 195-208. Millikin, R. L., and Smith, J. N. M. (1990). Sublethal effects of fenitrothion on forest passerines. J . Appl. Ecol. 27,983-1000. Moberg, G. P. (1985). Biological response to stress: Key to assessment of animal well-being? In “Animal Stress” (G. P. Moberg, ed.), pp. 27-49. Am. Physiol. SOC.. Bethesda, MD. Moberg, G. P. (1987). Problems in defining stress and distress in animals. J . Am. Vet. Med. Assoc. 191, 1207-1211. Moller, A. P. (1992). Female choice selects for male sexual tail ornaments in the monogamous swallow. Nature (London) 357,238-240. Moreno, J . (1989). Strategies of mass changes in breeding birds. B i d . J . Linn. Soc. 37,297-310. Morimoto, R. I., Tissitres, A,, and Georgopoulos. C. (1990). “Stress Proteins in Biology and Medicine.” Cold Spring Harbor Lab. Press, Cold Spring Harbor, NY. Morton, D. J.. Anderson, E., Foggin, C. M., Kock, M. D., and Tiran, E. P. (1995). Plasma cortisol as an indicator of stress due to capture and translocation in wildlife species. Vet. Rec. 136, 60-63. Mostl, E., Choi. H. S., Wurm, W.. Ismail, N.. and Bamberg, E. (1984). Pregnancy diagnosis in cows and heifers by determination of oestradiolL17cy in faeces. Br. Vet. J . 140,287-291. Muneer. M. A.. Farah, I. O., Pomeroy, K. A., Goyal, S. M., and Mech, L. D. (1988). Detection of parvoviruses in wolf feces by electron microscopy. J. Wildl. Dis. 24, 170-172. Munger, J . C., and Karasov, W. H. (1989). Sublethal parasites and host energy budgets: Tapeworm infection in white-footed mice. Ecology 70, 904-917. Munger, J . C., and Karasov, W. H. (1991). Sublethal parasites in white-footed mice: Impact on survival and reproduction. Can. J. Zoo1 69, 398-404. Myrberg, A. A. (1990). The effects of man-made noise o n the behavior of marine animals. Environ. Int. 16, 575-586. Nakata, H.. Tanabe, S., Tatsukawa, R.. Amano, M., Miyazaki, N., and Petrov, E. A. (1995). Persistent organochlorine residues and their accumulation kinetics in Baikal seal (Phoca sibirica) from Lake Baikal. Russia. Environ. Sci. Technol. 29, 2877-2885. Narita, M., Nanba, K., Haritani, M., and Kawashima, K. (1992). lmmunopathology in Aujeszky’s disease: Virus-infected pigs exposed to fluctuating temperatures. J. Comp. Pathol. 107, 221-229.

BIOLOGICAL CONSERVATION AND STRESS

515

Nevo, E., Beiles, A,. and Ben-Shlomo. R. (1984). The evolutionary significance of genetic diversity: Ecological, demographic and life history correlates. In “Evolutionary Dynamics of Genetic Diversity” (G. S. Mani, ed.), pp. 13-213. Springer, Berlin. Niezgoda, J.. Bobek, S.. Wronska-Fortuna, D., and Wierzchos. E. (1993). Response of sympatho-adrenal axis and adrenal cortex to short-term restraint stress in sheep. J. Vet. Med. AM, 63 1-638. Nimon. A. J., and Dalziel, F. R. (1992). Cross-species interaction and communication: A study method applied to captive siamang Hylobates syndactylus and long-billed corella Cacatua tenuirostris contacts with humans. Appl. Anim. Behav. Sci. 33, 261-272. Ninion, A. J., Oxenham, R. K. C., Schroter, R. C.. and Stonehouse, B. (1994). Measurement of resting heart rate and respiration in undisturbed and unrestrained, incubating gentoo penguins (Pygoscelzs papua). J. fhysiol. (Cambridge) 481,57P-58P. Nimon, A. J., Schroter, R. C., and Stonehouse, B. (1995). Heart rate of disturbed penguins. Nature (London) 374, 415-415. Nover, L. (1991). “The Heat Shock Response.” CRC Press, Boca Raton, FL. Obrecht. H. H., Pennycuick, C. J., and Fuller, M. R. (1988). Wind tunnel experiments to assess the effect of back-mounted radio transmitters on bird body drag. J. Exp. B i d . 135,256-273. O’Brien, S. J. (1994). The cheetah’s conservation controversy. Conserv. Biol. 8, 1153-1 155. O’Leary, H. (1996). Contrasts in diet amongst barbary macaques on Gibraltar: Human influences. Anim. Welfare 5, 177-188. Olsen, N. J., Nicholson, W. E., Debold, C. R., and Orth, D. N. (1992). Lymphocyte-derived adrenocorticotropin is insufficient to stimulate adrenal steroidogenesis in hypophysectomized rats. Endocrinology (Bakimore) 130, 2113-2119. Olsson, G. E., Willebrand, T., and Smith, A. A. (1996). The effects of hunting on willow grouse Lagopus lagopus movements. Wildl. B i d . 2, 11-15. Olsson, M., Karlsson, B.. and Ahnland, E. (1994). Diseases and environmental contaminants in seals from the Baltic and the Swedish west coast. Sci. Total Environ. 154, 217-227. Oppliger, A,. Christe, P., and Richner, H. (1996a). Clutch size and malaria resistance. Nature (London) 381,565. Oppliger, A.. Christe. P., and Richner, H. (1996b). Nature (London) 382, 502. O’Reilly, K. M.. and Wingfield, J. C. (1995). Spring and autumn migration in arctic shorebirds: Same distance, different strategies. Am. Zool. 35, 222-233. Oshea, T. J.. Rathbum, G. B., Asper, E. D., and Searles, S. W. (1985). Tolerance of WestIndian manatees Trichechus manatus to capture and handling. Biol. Conserv. 33,335-350. Osterhaus. A. D. M. E., de Swart, R. L.. Vos. H. W., Ross, P. S., Kenter, M. J., and Barrett, T. (1995). Morbillivirus infections of aquatic mammals: Newly identified members of the genus. Vet. Microbial. 44,219-227. Ozherelkov, S. V., Khozinsky. V. V.. and Semenov, B. F. (1990). Replication of Langat virus in immuno-competent cells of mice subjected to immobilization stress. Acfa Virol. 34,291 -294. Parmar. M. K. B.. and Machin, D. (1995). “Survival Analysis. A Practical Approach.” Wiley. Chichester. Parsell. D. A., Kowal, A . S.. Singer. M. A,, and Lindquist, S. (1994). Protein disaggregation mediated by heat-shock protein hsp 104. Nature (London) 372,475-478. Parsons. P. A. (1988a). Evolutionary rates: Effects of stress upon recombination. B i d . J. Linn. Soc. 35,49-68. Parsons. P. A. (1988b). Behavior, stress and variability. Behav. Genet. 18, 293-308. Parsons. P. A. (1989a). Environmental stresses and conservation of natural populations. Annu. Rev. Ecol. Syst. 20, 29-50.

516

HERIBERT HOFER AND MARION L. EAST

Parsons, P. A. (1989b). Conservation and global warming: A problem in biological adaptation to stress. Ambio 18, 322-325. Parsons, P. A. (1990a). Fluctuating asymmetry and stress intensity. Trends Ecol. Evol. 5,97-98. Parsons, P. A. (1990b). Fluctuating asymmetry: An epigenetic measure of stress. B i d . Rev. Cambridge Philos. SOC.65, 131-145. Parsons, P. A. (1990~).The metabolic cost of multiple environmental stresses: Implications for climatic change and conservation. Trends Ecol. Evol. 5, 315-317. Parsons, P. A. (1991). Evolutionary rates, stress and species boundaries. Annu. Rev. Ecol. Syst. 22, 1-18. Parsons, P. A. (1992). Fluctuating asymmetry: A biological monitor of environmental and genomic stress. Heredity 68, 361-364. Parsons, P. A. (1993a). Stress, extinctions and evolutionary change: From living organisms to fossils. B i d . Rev. Cambridge Philos. Soc. 68, 313-333. Parsons, P. A. (1993b). The importance and consequences of stress in living and fossil populations from life-history variation to evolutionary change. Am. Nut. 142, S5-S20. Parsons, P. A. (1993~).Evolutionary adaptation and stress: Energy budgets and habitats preferred. Brhav. Genet. 23, 231-238. Parsons, P. A. (1994). Habitats, stress, and evolutionary rates. J. Evol. B i d . 7, 387-397. Parsons, P. A. (1995). Evolutionary response to drought stress: Conservation implications. B i d . Conserv. 74, 21-27. Paterson, A. M., and Pearce, G. P. (1992). Growth response to humans and corticosteroids in male pigs housed individually and subjected to pleasant. unpleasant or minimal handling during rearing. Appl. Anim. Behav. Sci. 34, 315-328. Peakall, D. B. (1993). DDE-induced eggshell thinning: An environmental detective story. Environ. Rev. 1, 13-20. Peters, R. L., and Lovejoy. T. E. (1992). “Global Warming and Biological Diversity.” Yale University Press, New Haven, CT. Phillippi-Falkenstein, K., and Clarke, M. R. (1992). Procedure for training corral-living rhesus monkeys for fecal and blood sample collection. Lab. Anim. Sci. 42, 83-85. Pietz, P. J., Krapu, G. L., Greenwood, R. J.. and Lokemoen, J. T. (1993). Effects of harness transmitters on behavior and reproduction of wild mallards. J. Wildl. Manage. 57,696-703. Pokhil’ko, A. V., Kramskaia, T. A,. and Poliak, R. I. A. (1995). Stress and viral infection: Dynamics of expression of influenza A viral antigens during immobilization stress. Vopr. Virusol. 40, 76-79. Pollard, J. C., and Littlejohn, R. P. (1995). Effects of lighting on heart rate and positional preferences during confinement in farmed red deer. Anim. Welfare 4,329-337. Porter, J. M., and Coulson, J. C. (1987). Long-term changes in recruitment to the breeding group, and the quality of recruits at a kittiwake ROsu tridactyla colony. J. Anim. Ecol. 56, 675-689. Pottinger, T. G., and Calder, G. M. (1995). Physiological stress in fish during toxicological procedures: A potentially confounding factor. Environ. Toxicol. Water Qual. 10,135-146. Powell, P. C., and Davison, T. F. (1986). Induction of Marek’s disease in vaccinated chickens by treatment with betamethasone or corticosterone. Isr. J. Vet. Med. 42, 73-78. Preston. S. D., Aureli, F., and Insel, T. R. (1996). Heart rate telemetry as a window to the emotionality of freely-moving rhesus macaques: a pilot study. Abstr., Congr. Int. Primatol. Soc., 16th, 350. Price. S., Sibly, R. M., and Davies, M. H. (1993). Effects of behaviour and handling on heart rate in farmed red deer. Appl. Anim. Behav. Sci. 37, 111-123. Primack, R. B. (1993). “Essentials of Conservation Biology.” Sinauer, Sunderland. MA.

BIOLOGICAL CONSERVATION AND STRESS

5 17

Putman, R. J. (1995). Ethical considerations and animal welfare in ecological field studies. Eiodiversify Conserv. 4, 903-915. Quiatt, D., Rutan, B., Stone, T., and Reynolds, V. (1994). “Bundongo Forest Chimpanzees: Composition of Feeding Groups During the Rainy Season, with Attention t o the Spatial Integration of Disabled Individuals.” Budongo For. Proj. Rep. 21. Institute of Biological Anthropology, Oxford. Rasmussen, K. L. R. (1991). Identification, capture and biotelemetryof socially living monkeys. Lab. Anim. Sci. 41,350-354. Reijnders, P. J. H. (1980). Organochlorine and heavy metal residues in harbour seals from the Wadden Sea and their possible effects on reproduction. Nefh. J. Sea Res. 14, 30-65. Reijnders, P. J . H. (1986). Reproductive failure in common seals feeding on fish from polluted coastal waters. Nature (London) 324,456-457. Reijnen, R., and Foppen, R. (1994). The effects of car traffic on breeding bird populations in woodland. I. Evidence of reduced habitat quality for willow warblers (Phylloscopw trochilus) breeding close to a highway. J . Appl. Ecol. 31,235-94. Reijnen, R., and Foppen, R. (1995). The effects of car traffic on breeding bird populations in woodland. 1V. Influence of population size on the reduction of density close to a highway. J . Appl. Ecol. 32, 481-491. Reijnen, R., Foppen, R., ter Braak, C., and Thissen, J. (1995). The effects of car traffic on breeding bird populations in woodland. 111. Reduction of density in relation to the proximity of main roads. J. Appl. Ecol. 32, 187-202. Reinhardt, V. (1991a). Impact of venipuncture on physiological research conducted in conscious macaques. J. Exp. Anim. Sci. 34,212-217. Reinhardt, V. (1991b). Training adult male rhesus monkeys to actively cooperate during inhomecage venipuncture. Anim. Technol. 42, 11-17. Reinhardt, V., Cowley, D., Eisele, S., and Scheffler, J. (1991). Avoiding undue cortisol responses to venipuncture in adult male rhesus macaques. Anim. Technol. 42,8346. Reinhardt, V., Liss, C., and Stevens, C. (1995). Restraint methods of laboratory non-human primates: A critical review. Anim. Welfare 4, 221-238. Renshaw, E. (1991). “Modelling Biological Populations in Space and Time.” Cambridge University Press, Cambridge, UK. Richardson, W. J., Greene, C. R., Malme, C. I., and Thomson, D. H. (1995). “Marine Mammals and Noise.” Academic Press, San Diego, CA. Rideout, B. A., Gause, G. E., Benirschke, K., and Lasley, B. L. (1985). Stress-induced adrenal changes and their relation to reproductive failure in captive nine-banded armadillos (Dasypus novemcinctus). Zoo Biol. 4, 129-137. Rivier, C. (1989). Involvement of endogenous corticotropin-releasing factor (CRF) in modulating ACTH and LH secretion function during exposure to stress, alcohol or cocaine in the rat. In “Molecular Biology of Stress” (S. Breznitz and 0. Zinder, eds.), pp. 31-47. Alan R. Liss, New York. Robertson, 0.H., and Wexler, B. C. (1957). Pituitary degeneration and adrenal tissue hyperplasia in spawning Pacific salmon. Science 125, 1295-1296. Robertson, 0. H., and Wexler. B. C. (1960). Histological changes in the organs and tissues of migrating and spawning Pacific salmon (genus Oncorhynchus). Endocrinology (Ealrimore) 66, 222-239. Rodriguez, A., Barrios, L., and Delibes, M. (1995). Experimental release of an Iberian lynx (Lynx pardina). Eiodiv. Conserv. 4, 382-394. Rodway, M. S., Montevecchi, W. A,, and Chardine, J. W. (1996). Effects of investigator disturbance on breeding success of Atlantic puffins. Fratercula arctica. Biol. Conserv. 76,311-319.

518

H E R B E R T HOFER AND MARION L. EAST

Roff. D. A. (1992). “The Evolution of Life Histories.” Chapman & Hall, London. Rogeness, G. A,. and McClure, E. B. (1996). Development and neurotransmitter-environmental interactions. Dev. fsychoparhol. 8, 183-199. Rogers, C. M.. Ramenofsky, M., Ketterson, E. D.. Nolan, V., and Wingfield, J. C. (1993). Plasma corticosterone. adrenal mass. winter weather, and season in nonbreeding populations of dark-eyed juncos (Junco hyernalis hyernalis). Auk 110, 279-285. Rollo, C. D. (1995). “Phenotypes. Their Epigenetics, Ecology and Evolution.” Chapman & Hall, London. Root, T. ( 1988). Energetic constraints on avian distributions and abundances. Evolufion (Lawrence, Kuns.) 69, 330-330. Ross, P. S., Pohajdak, B., Bowen, W. D., and Addison, R. F. (1903). Immune function in free-ranging harbor seal (Phoca virulina) mothers and their pups during lactation. J . Wild. Dis. 29, 21-29. Ross, P. S.. de Swart. R. L., Reijnders, P. J . H., van Loveren, H.. Vos, J. G., and Osterhaus. A. D. M. E. (1995). Contaminant-related suppression of delayed-type hypersensitivity and antibody responses in harbor seals fed herring from the Baltic Sea. Environ. Health Perspect. 103, 162-167. Ross. P. S.. de Swart, R. L., Addison, R., van Loveren, H.. Vos, J . G., and Osterhaus, A. D. M. E. (1996a). Contaminant-induced immunotoxicity in harbour seals-wildlife at risk. Toxicology 112, 157-169. Ross, P. S., de Swart, R. L., Timmerman. H. H., Vedder. L. J., van Loveren, H., Vos, J . G., Reijnders, P. J. H., and Osterhaus, A. D. M. E. (1996b). Suppression of natural killer cell activity in harbour seals (Phoca virulina) fed Baltic Sea herring. Aquat. Toxicol. 34, 71-84. Rostene, W., Sarrieau. A., Nicot. A,, Scarceriaux. V.. Betancur, C., Gully, D., Meaney, M., Rowe, W.. de Kloet. R.. Pelaprat, D., and Berod. A. (1995). Steroid effects on brain functions: An example of the action of glucocorticoids on central dopaminergic and neurotensinergic systems. J. Psychiafr. Neurosci. 20, 349-356. Rotella, J . J., Howerter, D. W.. Sankowski, T. P.. and Devries, J. H. (1993). Nesting effort by wild mallards with three types of radio transmitters. J. Wildl. Manage. 57, 690-695. Rowan, A. N. (1990). Refinement of animal research technique and validity of research data. Fundam. Appl. Toxicol. 15, 25-32. Rushen, J. (1986). Some problems with the physiological concept of ‘stress.” Ausf. Ver. J . 63, 359-361. Sachser, N. (1987). Short-term responses of plasma norepinephrine. epinephrine, glucocorticoid and testosterone titers to social and non-social stressors in male guinea pigs of different social status. Physiol. Behav. 39, 11-20. Sachser. N. (1994). Social stratification and health in non-human mammals-a case study in guinea pigs. In “Social Stratification and Socioeconomic Inequality” (L. Ellis, ed.), Vol. 2, pp. 113-121. Praeger, Westport, CT. Sachser. N.. and Beer. R. (1995). Long-term influences of social situation and socialization on adaptability in behaviour. In “Research and Captive Propagation” (U. Ganslosser. J. K. Hodges, and W. Kaumanns, eds.). pp. 207-214. Filander, Furth. Germany. Sachser. N., and Kaiser. S. (1 996). Prenatal social stress masculinizes the females’ behaviour in guinea pigs. Physiol. Behav. 60, 589-594. Sachser, N., and Lick, C. (1989). Social stress in guinea pigs. Physiol. Behav. 46, 137-144. Sachser. N.. and Lick, C. (1991). Social experience, behavior, and stress in guinea pigs. Physiol. Behuv. 50, 83-90. Sachser. N., Lick, C., and Stanzel, K. (1994). The environment, hormones, and aggressive behavior: A 5-year-study in guinea pigs. Psychoneuroendocrinvlo~y19, 697-707.

BIOLOGICAL CONSERVATION AND STRESS

519

Sachser, N., Diirschlag. M., and Hirzel, D. (1997). Social relationships and the management of stress. Psychoneuroendocrinology (in press). Sage, G., Khawplod, P., Wilde, H., Lobaugh, C., Hemachudha, T., Tepsumethanon. W., and Lumlertdaecha, B. (1993). Immune response to rabies vaccine in Alaskan dogs: Failure to achieve a consistently protective antibody response. Trans. R. SOC. Trop. Med. Hyg. 87, 593-595. Saltz, D., and White, G. C. (1991). Urinary cortisol and urea nitrogen responses to winter stress in mule deer. J. Wildl. Manage. 55, 1-16. Saltzman. W., Schultz-Darken, N. J., Scheffler. G., Wegner, F. H.. and Abbott, D. H. (1994). Social and reproductive influences on plasma cortisol in female marmoset monkeys. Physiol. Behav. 56, 801-810. Sapolsky, R. M. (1982). The endocrine stress-response and social status in the wild baboon. Horm. Behav. 16,279-292. Sapolsky. R. M. (1985). Stress-induced suppression of testicular function in the wild baboon: Role of glucocorticoids. Endocrinology 116, 2273-2278. Sapolsky, R. M. (1987). Stress, social status, and reproductive physiology in free-living baboons. In “Psychobiology of Reproductive Behavior” (D. Crews, ed.), pp. 291-322. Prentice Hall, Englewood Cliffs. NJ. Sapolsky. R. M. (1992). Neuroendocrinology of the stress response. In “Behavioral Endocrinology” ( J . B. Becker, S. M. Breedlove. and D. Crews, eds.). pp. 287-324. MIT Press, Cambridge, MA. Sapolsky, R. M. (1994). “Why Zebras Don’t Get Ulcers.” Freeman, New York. Sapolsky, R. M. (1997). The physiological and pathophysiological implications of social stress in mammals. In “Handbook of Physiology” (in press). Sapolsky, R. M., and Ray, J . C. (1989). Styles of dominance and their endocrine correlates among wild olive baboons (Papio anribis). Am. J. Primntol. 18, 1-13. Satterlee, D. G., Jones, R. 9.. and Ryder, F. H. (1993). Effects of vitamin C supplementation on t h e adrenocortical and tonic immobility fear reactions of japanese quail genetically selected for high corticosterone response to stress. Appl. Anim. Behav. Sci. 35,347-357. Schapiro, S. J., Nehete, P. N., Perlman, J. E., and Sastry, K. J. (1996). Effects of environmental enrichment on cellular immune function of rhesus macaques. Ahstr., Congr. Int. Primatol. Soc.. 16th, 353. Schauer, J. H. S., and Murphy, E. C. (1996). Predation on eggs and nestlings of common murres (Uria cralge) at Bluff. Alaska. Colon. Waterbirds 19, 186-198. Scheepers. J. L.. and Venzke, K. A. E. (1995). Attempts to reintroduce African wild dogs Lycaon pictics into Etosha National Park, Namibia. S. Afr. J. Wildl. Res. 25, 138-140. Scherzinger. W. (1979). Zum Feindverhalten des Haselhuhnes (Bonasa bonasia). Vogelwelt 100, 205-217. Schmidt. C.. and Sachser, N. (1996). Auswirkungen unterschiedlicher Futterverteilungen auf Verhalten und Streflhormonkonzentrationene von Breitmaulnashornern (Ceratotherium siniirm simum) im Allwetterzoo Munster. Verhl. Dsch. Zoo/. Ges. 89(1), 294-294. Schneider, M. L. (1992a). The effect of mild stress during pregnancy on birthweight and neuromotor maturation in rhesus monkey infants (Macaca mulafta). Infant Behav. Dev. 15,389-403. Schneider. M. L. (1992b). Delayed object permanence development in prenatally stressed rhesus monkey infants (Macaca m d a t m ) . Occup. Ther. J. Res. 12,96-110. Schneider, M. L. (1992~).Prenatal stress exposure alters postnatal behavioral expression under conditions of novelty challenge in rhesus monkey infants. Dev. Psychohiol. 25,529-540. Schneider. M. L., and Coe, C. L. (1993). Repeated social stress during pregnancy impairs neuromotor development of the primate infant. J. Dev. Behav. Pediatr. 14, 81-87.

520

HERIBERT HOFER AND MARION L. EAST

Schnidrig, R., Marbacher. H., and Ingold. P. (1992). How is paragliding affecting the behaviour and general condition of chamois Rupicapra rupicapra: A study in the Swiss alps. In “Ongules/Ungulates 91” (F. Spitz, ed.). pp. 609-610. Inst. Rech. Grands Mammifkres, Toulouse, France. Schoech. S. J., Mumme, R. L., and Moore, M. C. (1991). Reproductive endocrinology and mechanisms of breeding inhibition in cooperatively breeding Florida scrub jays. (Aphelocoma c. coerulescens). Condor 93, 354-364. Schofield, P. S., McLees, D. J., Kerbey, A. L. and Sugden, M. C. (1986). Activities of cardiac and hepatic pyruvate dehydrogenase complex are decreased after surgical stress. Biochem. Int. 12, 189-197. Schrader. L. and Todt, D. (in press). Vocalisations can indicate an endocrine stress response in domestic pigs (Sus scrofa domesticn). Ethology, in press. Schultz. R. D.. and Bailey, J. A. (1978). Response of white-tailed deer to snowmobiles and snowmobile trails in Maine. Can. Field Nat. 92, 334-344. Schumacher, U., Heidemann, G . ,Skirnisson, K., Schumacher, W., and Pickering, R. M. (1995). Impact of captivity and contamination level on blood parameters of harbour seals (Phoca vitulina). Comp. Biochem. Physiol. A 112A, 455-462. Scott, T. M.. and Koehn, R. K. (1990). The effect of environmental stress on the relationship of heterozygosity to growth in the coot clam Mulinia lateralis (Say). J. Exp. Mar. B i d . Ecol. 135, 109-116. Sears, J., Cooke, S. W., Cooke, Z . R., and Heron, T. J. (1989). A method for the treatment of lead poisoning in the mute swan (Cygnus olor) and its long-term success. Brit. vet. J. 145, 586-595. Selye, H. (1946). The general-adaptation syndrome and disease of adaptation. J. Clin. Endocrinol. Metab. 6, 117-230. Selye, H. (1973). The evolution of the stress concept. Am. Sci. 61, 692-699. Shackley, M. (1992). Manatees and tourism in southern Florida: opportunity or threat? J. Environ. Manage. 34, 257-265. Shepherd, N. (1986). Capture myopathy. I n “Care and Handling of Australian Native Mammals” (S. J. Hand, ed.), pp. 143-148. Surrey Beatty, Chipping Norton, Australia. Sibly. R. M., and Calow, P. (1983). An integrated approach to life-cycle evolution using selective landscapes. J. Theor. Biol. 102, 527-547. Sibly, R. M., and Calow, P. (1986). “Physiological Ecology of Animals.” Blackwell, Oxford. Sibly, R. M., and Calow, P. (1989). A life-cycle theory of responses to stress. Biol. J . Linn. Soc. 37, 101-116. Simmonds, M. (1991). The involvement of environmental factors in the 1988 european seal epidemic. Rev. Int. Oceanogr. Med. 101-104, 109-1 14. Simmonds, M., and Johnston, P. A. (1989). Seals, sense and science. Mar. Pollut. Bull. 20, 580-583. Sinclair. A. R. E., and Pech, R. P. (1996). Density dependence, stochasticity, compensation and predator regulation. Oikos 75, 164-173. Skaare. J. U., Markussen, N. H., Norheim, G.. Haugen, S., and Holt, G. (1990). Levels of polychorinated biphenyls, organochlorine pesticides, mercury, cadmium, copper, selenium, arsenic and zinc in the harbor seal (Phoca vitulina) in Norwegian waters. Environ. Poll~it.66, 309-324. Skogland, T., and Grovan, B. (1988). The effects of human disturbance on the activity of wild reindeer in different physical condition. Rangifer 8, 11-19, Smith, G. J., and Rongstad, 0. J. (1980). Serologic and hematologic values of wild coyotes in Wisconsin. J . Wildl. Dis. 16, 491-497.

BIOLOGICAL CONSERVATION AND STRESS

521

Smith. G. T., Wingfield, J. C.. and Veit. R. R. (1994). Adrenocortical response to stress in the common diving petrel. Pelernnoides urinafrix. Physiol. Zoo/. 67, 526-537. Smith, H. R. (1980). Growth. reproduction and survival in Peromyscus leucopus carrying intraperitoneally implanted transmitters. In “ A Handbook of Biotelemetry and Radiotracking’’ (C. A. Amlaner and D. W. Macdonald. eds.), pp. 367-374. Pergamon. Oxford. Smith. R. J . (undated). Some effects of limb injuries on the chimpanzees of the Budongo Forest. Ms, Institute of Biological Anthropology, Oxford. Smith. T. E., and McGreer-Whitworth, B. (1996). Psychosocial stress and urinary cortisol excretion in marmoset monkeys (Callirhrix kuhli). Abstr. Congr. Int. Primatol. Soc. I6rh, 086. Soave, 0.A . (1964). Reactivation of rabies virus in a guinea pig due to the stress of crowding. Am. J . Vet. Rex 25, 268-269. Soave, 0. A,, Johnson, H. N.. and Nakamura, K. (1961). Reactivation of rabies virus infection with adrenocorticotropic hormones. Science 133, 1360-1360. Sober, E. (1 993). “Philosophy of Biology.” Oxford University Press. Oxford. Soule. M. E. (1986). “Conservation Biology.” Sinauer, Sunderland, MA. SoulC, M. E. (1987). “Viable Populations for Conservation.” Cambridge University Press. Cambridge, UK. Southwood, T. R. E. (1988). Tactics, strategies and templets. Oikos 52, 3-18. Speakman. J. R., Webb. P. I.. and Racey. P. A. (1991). Effects of disturbance on the energy expenditure of hibernating bats. J. Appl. Ecol. 28, 1087-1095. Spraker. T. R., Adrian, W. J . , and Lance, W. R. (1987). Capture myopathy in wild turkeys Meleagri.s gnllopavo following trapping, handling and transportation in Colorado. USA. J . Wildl. Dis. 23, 447-453. Stangel, P. W.. and Lennartz. M. R . (1988). Survival of red-cockaded woodpecker nestlings unaffected by sampling blood and feather pulp for genetic studies. J. Field Ortiithol. 59,389-394. St. Aubin. D. J., and Geraci, J . R. (1986). Adrenocortical function in pinniped hyponatremia. Mur. Mammol. Sci. 2, 243-250. Steams. S. C. (1992). “The Evolution of Life Histories.” Oxford University Press. Oxford. Stebbings, R. E. (1982). Radio tracking greater horseshoe bats with preliminary observations on flight patterns. Symp. Zoo/. Soc. London 49, 161-173. Stefano. G. B., and Smith, E. M. (1996). Adrenocorticotropin-a central trigger in immune responsiveness: Tonal inhibition ‘of immune activation. Med. Hypotheses 46, 471-478. Still, A. W. (1982). O n the number of subjects used in animal behaviour experiments. Anim. Behav. 30, 873-880. Stirling, I.. and Derocher. A. E. (1993). Possible impacts of climatic warming on polar bears. Arctic 46, 240-245. Stockwell, C. A.. Mulvey. M.. and Vinyard, G. L. (1996). Translocations and the preservation of allelic diversity. Conserv. B i d . 10, 1133-1 141. Stone. A. M., Bloomsmith, M. A,, Laule. G. E.. and Alford. P. L. (1996). Positive reinforcement training for voluntary movement of group-housed chimpanzees. Abstr., Cungr. fnt. Primcrtol. Soc.. 16th, 679. Stone, G. S.. Katona, S. L., Mainwaring, A,. Allen. J. M., and Corbett, H. D. (1992). Respiration and surfacing rates of fin whales (Balaenopferu physalus) observed from a lighthouse tower. Rep. Inr. Whal. Cornm. 42, 739-745. Stussy. R. J., Edge. W. D., and O’Neil, T. A. (1994). Survival of resident and translocated female elk in the Cascade Mountains of Oregon. Wildl. Soc. Birll. 22, 242-247. Sutherland. W. J. (1996). “From Individual Behaviour to Population Ecology.” Oxford University Press. Oxford.

522

HERIBERT HOFER A N D MARION L. EAST

Taberlet, P.. and Bouvet. J . (1991). A single plucked feather as a source of DNA for bird genetic studies. Auk 108, 959-960. Taberlet. P., and Bouvet, J. (1992). Bear conservation genetics. Nature (London)358,197-197. Tarara, E. B., Tarara, R. P.. and Suleman, M. A. (1995). Stress-induced gastric ulcers in vervet monkeys (Cercopitlzecus aethiops): The influence of life history factors. Part 11. J . Zoo Wildl. Med. 26, 72-75. Taylor, V. J.. and Dunstone, N., eds. (1996a). “The Exploitation of Mammal Populations.” Chapman & Hall. London. Taylor, V. J., and Dunstone. N. (1996b). The exploitation, sustainable use and welfare of wild mammals. In “The Exploitation of Mammal Populations” (V. J. Taylor and N. Dunstone, eds.), pp 3-15. Chapman & Hall, London. Thirgood. S. J.. Redpath, S. M.. Hudson, P. J.. Hurley, M. M.. and Aebischer, N. J. (1995). Effects of necklace radio transmitters on survival and breeding success of red grouse Lagopus lagopi~sscoficus. Wildl. B i d . 1, 121-126. Thomas, J., Carver, M., Haische, C., Thomas, F.. Welch, J., and Carchman, R. (1982). Differential effects of intravenous anaesthetic agents on cell-mediated immunity in the rhesus monkey. Clin. Exp. lmmrinol. 47, 457-466. Thomas, J. A,, Kastelein. R. A,, and Awbrey. F. T. (1990). Behavior and blood catecholamines of captive belugas during playbacks of noise from an oil drilling platform. Zoo Biol. 9, 393-402. Thomas, L.. and Juanes. F. (1996). The importance of statistical power analysis: An example from Animal Behaviour. Anim. Behav. 52,856-859. Thompson, D. R.. Hamer, K. C.. and Furness, R. W. (1991). Mercury accumulation in great skuas Catharacta skuci of known age and sex and its effects upon breeding and survival. J . Appl. Ecol. 28,672-684. Thorne, E. T.. and Williams, E. S. (1988). Disease and endangered species: The black-footed ferret as a recent example. Conserv. B i d . 2, 66-74. Toates, F. (1995). “Stress. Conceptual and Biological Aspects.’’ Wiley, Chichester. Tranvik. L., Bengtsson. G.. and Rundgren, S. (1993). Relative abundance and resistance traits of two Collemhola species under metal stress. J. Appl. Ecol. 30, 43-52. Trillmich, F., and Ono. K. A. (1992). “Pinnipeds and El Niiio. Responses to Environmental Stress.” Springer. Berlin. Underwood, A. J., and Kennelly. S. J. (1990). Pilot studies for designs of surveys of human disturbance of intertidal habitats in New South Wales. Australia. Auut. J . Mar. Freshwater Res. 41, 165-174. van Eden, W., and Young. D. B. (1996). “Stress Proteins in Medicine.” Dekker. New York. Van Heezik, Y.. and Seddon. P. J . (1990). Effect of human disturbance on beach groups of jackass penguins. S. A,fr. J . Wildl. Rrs. 20, 89-93, Van Horne, B. (1983). Density as a misleading indicator of habitat quality. J. W i l d Manage. 47, 893-90 1. van Oirschot. J. T.. and Gielkens, A. L. J. (1984). In vivo and in vitro reactivation of latent pseudorabies virus in pigs born to vaccinated sows. Am. J . Vet. Res. 45, 567-571. Van Vuren, D. (1989). Effects of intraperitoneal transmitter implants on yellow-bellied marmots. J. Wildl. Manage. 53, 320-323. von Holst, D. (1986). Psychosocial stress and its pathophysiological effects in tree shrew (Tupaia belangeri). In “Biological and Psychological Factors in Cardiovascular Disease” (T. H. Schmidt, T. M. Dembroski. and G. Blumchen. eds.), pp. 476-489. Springer. Heidelberg. Walker. C. H.. Hopkin. S. P.. Sibly, R. M.. and Peakall, D. B. (1996). “Principles of Ecotoxicology.” Taylor & Francis, London.

BIOLOGICAL CONSERVATION AND STRESS

523

Walter, H. S. (1990). Small viable population: The red-tailed hawk of Socorro Island. Conserv. B i d . 4, 441-443. Warner. R. E., and Etter, S. L. (1983). Reproduction and survival of radio-marked hen ringnecked pheasants in Illinois. J. Wildl. Mange. 47, 369-375. Warwick. R. M. (1986). A new method for detecting pollution effects on marine macrobenthic communities. Mar. Biol. 92, 557-562. Wasser. S. K., and Barash, D. P. (1983). Reproductive suppression among female mammals: Implications for biomedicine and sexual selection theory. Q. Rev. Biol. 58, 513-538. Wasser. S. K.. and Starling, A. K. (1986). Reproductive competition among female yellow baboons. In “Primate Ontogeny. Cognition and Social Behaviour” ( J . G. Else and P. C. Lee. eds.). pp. 343-354. Cambridge University Press, Cambridge, UK. Watson, K. (1990). Microbial stress proteins. A h . Microb. Physiol. 31, 183-223. Wattrang, E.. Edfors-Lilja, I., Anderson. L.. and Fossum, C. (1994). Genetic and stressmediated influence on Aujeszky’s disease virus induced interferon-alpha production in porcine leukocytes. Acfa Vet. Hung. 42, 331-336. Westneat. D. F. (1986). The effects of muscle biopsy on survival and condition in white throated sparrows. Wilson Bull. 98, 281-285. Westneat, D. F., Payne. R. B.. and Doehlert, S. M. (1986). Effects of muscle biopsy on survival and breeding success in Indigo buntings. Condor 88, 220-227. Wheeler. W. E., Gatti, R. C.. and Bartelt. G. A. (1984). Duck breeding ecology and harvest characteristics on Grand River Marsh Wildlife Area, Wisconsin, USA. Wis. Dep. Nar. Res. Tech. Bull. 145, 1-49, White, G. C., and Garrott. R. A. (1990). “Analysis of Wildlife Radio-tracking Data.” Academic Press, San Diego, CA. Widmaier, E. P.. Harmer, T. L.. Sulak, A. M., and Kunz, T. H. (1994). Further characterization of the pituitary-adrenocortical responses to stress in chiroptera. J . Exp. Zool.269,442-449. Wiepkema, P. R., and Kolhaas, J. M. (1993). Stress and animal welfare. Anim. Welfare 2, 195-218. Williams, G. C . (1966). “Adaptation and Natural Selection.” Princeton University Press. Princeton, NJ. Williams. T. D., and Rothery. P. (1990). Factors affecting variation in foarging and activity patterns of gentoo penguins (Pygoscelis papitu) during the breeding season at Bird Island, South Georgia. J. Appl. Ecol. 27, 1042-1054. Williamson. D. T. (1991). Condition, growth and reproduction in female red lechwe (Kobiis leche leche Gray 1850). Afr. J. Ecol. 29, 105-117. Wilson, B. S.. and Wingfield, J. C. (1992). Correlation between female reproductive condition and plasma corticosterone in the lizard Ura stansburiana. Copeia, pp. 691-697. Wilson, B. S.. and Wingfield, J. C. (1994). Seasonal and interpopulational variation in plasma levels of corticosterone in the side-blotched lizard (Ura sransburiana). Physiol. 2001. 67, 1025-1049. Wilson. R. P., and Wilson, M. P. T. J. (1989). A minimal-stress bird-capture technique. J. Wildl. Manage. 53, 77-80. Wilson. R. P., Coria. N. R.. Spairani, H. J.. Adelung. D., and Culik, B. M. (1989). Humaninduced behavior in AdClie penguins (Pygoscelis adeliae). Polar Biol. 10, 77-80. Wilson, R. P., Culik, B. M., Danfeld. R.. and Adelung, D. (1991). People in Antarctica: How much do AdClie penguins (Pygoscrlis adeliae) care? Polar B i d . 11, 363-370. Wingfield. J. C. (1985). Influences of weather on reproductive function in female song sparrows (Melospiza melodia). J. Zool. 205, 545-558. Wingfield. J. C. (1988). Changes in reproductive function of free-living birds in direct response to environmental perturbations. In “Processing of Environmental Information in Vertebrates” (M. H. Stetson. ed.), pp. 121-148, Springer, New York.

524

HERIBERT HOFER AND MARION L. EAST

Wingfield. J. C.. and Farner. D. S. (1976). Avian endocrinology-field investigations and methods. Condor 78, 570-573. Wingfield, J. C., and Farner, D. S. (1980). “Endocrinologic and Reproductive States of Bird Populations under Environmental Stress,” U S . EPA Report Contract No. 699095. U S . Environ. Prot. Agency, Washington, DC. Wingfield, J. C.. and Grimm. A. S. (1977). Seasonal changes in plasma cortisol, testosterone and oestradiol-17-beta in the plaice, Pleuronectes platessa L. Gen. Comp. Endocrinol. 31, 1-11. Wingfield, J. C.. and Lewis. D. M. (1993). Hormonal and behavioural responses to simulated territorial intrusion in the cooperatively breeding white-browed sparrow weaver. Plocepasser mahali. Anim. Behav. 45, 1-1 1. Wingfield, J. C., Hahn. T. P.. Levin, R.. and Honey, P. (1992a). Environmental predictability and control of gonadal cycles in birds. J. Exp. Zoo/. 261,214-231. Wingfield. J. C., Vleck. C. M., and Moore. M. C. (1992b). Seasonal changes of the adrenocortical response to stress in birds of the Sonoran desert. J. Exp. Zoo/. 264, 419-428. Wingfield, J. C., Deviche, P.. Sharbaugh, S.. Astheimer. L. B.. Holberton, R.. Suydam, R., and Hunt, K. (1994a). Seasonal changes of the adrenocortical responses to stress in redpolls, Acanthis fiammea, in Alaska. J. Exp. Zoo/. 270, 372-380. Wingfield. J. C.. Suydam, R.. and Hunt, K. (1994b). The adrenocortical responses to stress in snow buntings (Plectrophenux nivn/is) and Lapland longspurs (Calcnriirs iapponicus) at Barrow, Alaska. Comp. Biochem. Physiol. C 108,299-306. Wingfield, J. C., Kubokawa, K., Ishida, K., Ishii, S., and Wada. M. (1995a). The adrenocortical response to stress in male bush warblers, Cettia diphone: A comparison of breeding populations in Honshu and Hokkaido, Japan. Zoo/. Sci. 12, 615-621. Wingfield, J. C.. O’Reilly, K. M., and Astheimer, L. B. (1995b). Modulation of the adrenocortical responses to acute stress in Arctic birds: a possible ecological basis. Am. Zoo/. 35, 285-294. Wingfield, J. C.. Hunt, K.. Breuner. C., Dunlap, K., Fowler. G. S., Freed, L.. and Lepson, J. (1 996). Environmental stress, field endocrinology, and conservation biology. I n “Behavioral Approaches to Conservation in the Wild” ( J . R. Clemmons and R. Buchholz, eds.), pp. 95-131. Cambridge University Press, Cambridge, UK. Wingfield. J. C., Breuner, C., and Jacobs. J. (1997). Corticosterone and behavioral responses to unpredictable events. I n “Perspectives in Avian Endocrinology” (S. Harvey and R. J. Etches. eds.), pp. 1-12. Journal of Endocrinology Ltd, Bristol. Wingfield. J. C., Maney, D. L., Breuner, C. W.. Honey, P. K., Jacobs, J. D., Lynn, S.. Ramenofsky, M., and Richardson, R. D. (in press). Ecological bases of hormone-behavior interactions: The “emergency life history stage”. Am. Zool., (in press). Wolf, C. M., Griffith, B., Reed, C.. and Temple, S. A. (1996). Avian and mammalian translocations: Update and reanalysis of 1987 survey data. Consrrv. B i d . 10, 1142-1 154. Woodford. M. H., and Kock, R. A. (1991). Veterinary considerations in re-introduction and translocation projects. Symp. Zool. Soc. London 62, 101-1 10. Woolf, A,, Curl. J. L., and Anderson, E. (1984). Inanitation following implantation of a radiotelemetry device in a river otter. J. Am. Vet. Med. Assoc. 185, 1415-1415. Wright, J . F. (1995). Development and use of a system for predicting the macroinvertebrate fauna in Howing waters. Atrst. J. Ecol. 20, 181-197. Wright, J. F., Furse. M. T., Armitage, P. D., and Moss, D. (1993). New procedures for identifying running-water sites subject to environmental stress and for evaluating sites for conservation based on the macroinvertebrate fauna. Arch. Hvdrobiol. 127, 3 19-326. Wuichet. J.. and Norton, B. (1995). Differing conceptions of animal welfare. In “Ethics on the Ark: Zoos, Animal Welfare. and Wildlife Conservation” (B. G . Norton. M. Hutchins,

BIOLOGICAL CONSERVATION A N D STRESS

525

E. F. Stevens. and T. L. Maple. eds.). pp. 235-250. Smithsonian Institution Press. Washington, D C and London. Wynn. P. C., Shahneh. A. Z.. Ribgy. R. D. G . . Behrendt, R.. Giles, L. R., Gooden, J. M., and Jones. M. R. (1995). Physiological consequences of the induction of auto-immunity to adrenocorticotropin (ACTH). Livest. Prod. Sci. 42, 247-254. Yarmoloy, C.. Bayer, M., and Geist, V. (1988). Behavior responses and reproduction of mule deer. Odocoileus hentionus, does following experimental harassment with an all-terrain vehicle. Can. Field Nat. 102, 425-429. Yediler. A.. Panou, A., and Schramel, P. (lY93). Heavy metals in hair samples of the mediterranean monk seal (Monachrrs monc7hus). Mar. Polhir. B d l . 26, 156-159. Yeoman. R. R.. Ricker. R. B.. Williams. L. E.. Sonksen. J.. and Abee. C. R. (1996). Vibrostimulation (VS) is superior to electro-ejaculation (EE) in squirrel monkeys (Suirniri). Ahstr.. Congr. Int. Primatol. Soc.. 16th, 222. Yeruham, I.. Perl. S., Nyska. A,, Abraham. A., Davidson, M.. Haymovitch, M., Zamir, 0.. and Grinstein, H. (1994). Adverse reactions in cattle to a capripox vaccine. Vet. Rec. 135,330-332. Zethof, T. J. J . , van der Heyden. J . A. M., Tolboom, J. T. B. M., and Olivier. B. (1994). Stress-induced hyperthermia in mice: a methodological study. Physiol. Behav. 55,109-1 15. Zwartouw. H. T., MacArthur. J. A,. Boulter, E. A,, Seamer. J. H.. Marston, J . H.. and Chamove, A . S. (1984). Transmission of B virus infection between monkeys especially in relation to breeding colonies. Lob. Anirri. 18, 125-130.

This Page Intentionally Left Blank

Index

A Abiotic stress adaptation. 165-166 energy balance, 155-156 Achievement pleasure, 392 Acoustics, cry, offspring condition, 354-355 ACTH (adrenocorticotrophic hormone) central effects, 20-21 glucocorticosteroid release. 19-20 Active stress response, 107-108 Cannon’s, 3-5. 62 chronic stress, 10 dominance relationships, 62, 63, 88-89 gonadal activity, 14 tree shrews, 54-58 Activity, as indicator of stressed state. 449-450 Adaptation abiotic stress, 165-166 chronic challenges, 5-10 critical situations. 3 energy limits. 158-164 evolutionary, relationship to stress, 414 extending limits. 165-169 learning, 168-169 resource heterogeneity, 166-167 resource polymorphisms, 167-168 as solution to environmental stress, 320 stress response as. 438-439 Adaptation diseases, 6 Adaptation syndrome, general. 5-6. 62 Adaptive sociality. predation. 239-241 Adenine energy charge, 134, 446-447 Adrenal cortex hormonal activation, 19-20 sympathetic activation, 20 zones, 19 Adrenal glands enlargement. long-term stress. 22 hormone measurements, usefulness. 22-29 527

weight measurements psychosocial stress, 48-49 usefulness, 22-23 Adrenal medulla, see Sympatheticoadrenomedullary system Adrenocortical hormones, 19-22 Adrenocortical response, see also Pituitary-adrenocortical system; Stress response aggressive behavior, 45-47, 59 assessment, 22-29, 490 bidirectional nature, 63 challenge test, 25-29 chemical immobilization and anesthesia, 478 dominance relationships rabbits, 64-65 tree shrews, 55-57 emotionally induced, 6-7 energetics, 435-436 fitness effects, 431 gonadal activity negative coupling, 432-434 uncoupling, 434-435 manual restraint, 477 metabolic rate, 435-436 passive chronic stress, 10-11 predictability and control, 7. 8-10. 63. 88-89, 437 psychosocial stress. 48-49 reproduction, 431 -436 Selyean concept, 5-6. 62 sensitivity, environmental stress, 431-432. 433-434 social bonds, disruption. 90, 93 social factors, 436-437 tuning options, 431-435 variability, 430-431 Adrenocorticotrophic hormone central effects, 20-21 glucocorticosteroid release, 19-20

528

INDEX

AEC, see Adenine energy charge African lions (fandirra leo), pursuit hunting, 467 African talapoin (Miopifhwrrs frilqmin), stress response. dominance relationships, 71 African wild dogs (Lvcaon picfirs). stress response, dominance relationships. 72 Agdaiiis phoeniceiis. food aversions. 307 Aggressive behavior adrenocortical response, 45-47. 59 cardiovascular response. 67-69 mortality. 1-2, 434-435 social stress. 45-47 testosterone, 75. 86-87 Alarm reaction. general adaptation syndrome. 5 Aldosterone, stress response. 22 Algal bloom, 454-455 American black bears (Ursrrs rinirriconirs), tourism-related disturbance. 464 Amphibians. use of space, antipredator decision making, 256 Androgens adrenal. stress response, 22 physiological effects. 3.5 production and release. 36 Anesthesia. as stressor. 478-479. 479 Anger. function. 390 Animal breeding, developmental instability, 191

Animal handling, see Handling Animal housing equivalence to natural stressors. 487-488 as stressor. 16. 483-485 Animal models. “knock out,” immune system. 138 Animal welfare assessment, 396-397 coping systems. 400 preference tests. 397-398 time. 398 definition. 394-395 devclopmental instability, 197-199 relationship to health. 396. 397 relationship to stress. 396. 397 relevance of feclings. 371, 395. 399-400 stress concept, 407-408 Antarctic penguins, tourism-related disturbance, case study. 462-463

Anfechinus, stress reactions, males, 1-2. 434-435 A nthropogenic st ressors effects. 452-472 energetic measures, 439-441 equivalence to natural strcssors, 486-488 examples. 415 fitness consequences. 413 Antibacterial foliage. 296-297 Antibodies. 38. 41 Anticipation. stress concept. 408 Anting, birds, 297-298 Antipredator decision making. see NISO Predator-induced stress developmental instability, 189-190. 192 ecological effects, 248-261 foraging behavior. 217-225 future research. 264-265 long-term consequences. 245-248 overview, 215-216 patterns of activity. 225-235 levels, 226-23 1 temporal. 231-235 types, 226 population dynamics. 257-259 postencounter, 235-239 approaching and inspecting predators. 238-239 escape behavior. 237-238 flight initiation. 237 pursuit-deterrence signals, 236-237 resumption of activity. 236 predator reaction, 263 reproductive, 241-245, 259 risk assessment. 262-263 role of modeling. 225 scaling to real world. 264 social situations. 239-241 species interactions. 259-261 stress response. 261 -262 use of space. 248-257 Antipredator training. 492 Antischistosomal drug use. baboons. 296 Antisymmetry, 181, 183 Ants, adaptation. abiotic stresses. 166 Anxiety. function, 385 Aphids, escape behavior, 238 Arctic charr (Srrlvelinrrs alpinus). resource polymorphisms. 167

INDEX

Aspilia. self-medication. chimpanzees. 298-301 Asymmetry behavioral. 204-205 correlations. 185 directional. 183 facial, 326-327 fitness correlates. 192-193 fluctuating. see Fluctuating asymmetry morphological behavioral effects, 204 types, 181, 183 performance costs, 184 secondary sexual characters. 190-191, 323-324 selection pressure. 184 types. 181-183 Athleticism. immunocompetence. 327-328 ATP. as indicator of stressed state. 446-447 Attraction and attractiveness. humans, 321-332,356-357 developmental stability, 321-323 secondary sexual characters. 323-324 stress resistance nonsexual social behavior, 330-331 sexual selection, 324-330 Audiogenic stress, developmental instability, 189 Autonomic nervous system. 29 innervation of lymphoid organs, 140-141 Aves, pursuit-deterrence signals. 237 Awareness, definition, 374

B Baboons antischistosomal drug use. 296 coronary artery disease, chronic social stress, 70 dominance relationships female reproductive success. 81 immunological consequences, 83. 85 stress reactions, 69, 71 stimulant use. 295-296 Baby blues, 344 Bacterial infection. immune response, 137-138

529

Badgers escape behavior. 238 hunting effects. 466 Baltic gray seals, skull bone lesions. 455-456 Barnacles. tidal activity patterns. 235 Behavior asymmetrical, 204-205 as indicator of stressed state, 44Y-450 invariance. 201-204 observation. recognition of feelings, 375-376 sequences. fractal dimension, 203-204 stress concept. 409-410 Behavioral conditioning. immune function, 141, 144 Behaviorally transmitted indirect effects activity level, 259-260 top-down ecosystem regulation. 260 use of space. 260 Behavioral research, stress concept. 14-15. 44-45 Behavioral traits. repeatability. 202-203 Betfo splendms. sexual behavior, energy costs, 160 Bighorn sheep (Ovis canadensis), heterozygosity and fitness, 164 Biological conservation. see Conservation biology Biomedicine, behavioral stress research. 44-45 Biopsies, muscle. as stressor, 479-480 Birds activity patterns, 234 anting, 297-298 body mass loss. breeding season. 414 conservation research recommendations. 489 handling effects, 476-477 parental effort. immune effects. 145 plant use, 296-297 radio transmitter effects, 481-482 use of space, antipredator decision making, 256-257 Blood pressure aggressive behavior, 67-69 measurement techniques. 34-35 psychosocial stress, 49 stress response, 14, 33. 34-35

530

INDEX

Blood sampling challenge test, 26-29, 42 chemical contamination monitoring, 490 immunological parameters. 40-42 methodological problems, 16-18, 23-24 remote-controlled, 491 as stressor, 476-477 Blowfly (Phormia regina). specific hungers, 306 Bluegill sunfish (Lepomis macrochinis). learning, 169 Blushing, function, 387 B lymphocytes, 38, 40, 41 Body mass birds, breeding season, 414 as indicator of stressed state, 447 male, immunocompetence, 327-328 Body temperature, measurement, as stressor, 479 Bonnet macaques (Macnca radiata), social stress, immunological consequences, 441-442 Boredom, function, 388-389 Brain asymmetry, 205-206 Breast asymmetry, 199-200, 328-329 Breeding suppression. predation, 244-245, 259 Bumblebees, escape behavior, 238 Burramys parvrts, tourism-related stress, 415, 417-420 Butterflies, wing asymmetry, 197

C Calopteryx tnaciilara, sexual behavior. energy costs, 159 Cannon's fight or flight response, 3-5. 10, 62 Capricorn (Capra ibex), tourism-related disturbance, 464 Captivity equivalence to natural stressors. 487-488 as stressor. 483-485 Capture, as stressor, 476-477 Capture collar, 491 Capture myopathy. 422, 467, 477

Cardiovascular system, see also Heart rate chronic stress effects, 34, 70 stress response, dominance relationships, 55, 66. 67-69 Cotaglyphis bombycina, adaptation, abiotic stresses. 166 Catastrophes. equivalence to anthropogenic stressors. 487 Catecholamines biosynthesis, 31 challenge tests. 32-33 circuits, activation, 31 immunosuppressive effect, 39 as indicator of stressed state, 449 Cave bats. handling effects, 476 Cavia crperea f. poreellus, see Guinea pigs Cellular factors. immune system. 38, 137 Cellular stress response. 429 Censuring disturbance, 474-475 Central nervous system ACTH effects, 20-21 immune system connections, 38-39 Centrocercus iirophasianus, sexual behavior, energy costs, 160 Challenge test catecholamines, 32-33 glucocorticosteroids, 26-29 Chamois (Riipicapra rupicapra), tourism-related disturbance, 464 Cheetah, inbreeding depression. 495-496 Chelonia mynas, tourism-related disturbance, 464 Chemical contamination, see Pollution Chemical detection, predators, 263 Chemical immobilization remote-controlled, 491 as stressor. 477, 478-479 Chen caerrrlescens, hunting effects, 468 Chickens skeletal asymmetry, 197-198 specific hungers, 305-306 Chimpanzees self-medication, 298-303, 304 snare wounds, 469 Chronic stress active vs. passive coping, 10-11 adaptation, 5-10 blood pressure, 67-69 cardiovascular effects, 34, 70 glucocorticosteroid-induced effects, 19. 22

INDEX

immune response, 138 mortality. 70-71 Selye's general adaptation, 5-6, 62 social experience. 64 Circus aencginosics, tourism-related disturbance, 464 Cliff swallow (Hiruntlo pyrrhonala), parasite stress, energy costs, 160-161 Climatic warming. 472, 487 Colicmba livia, parasite stress, 160 Community composition, as indicator of stressed state, 452 Complement system, assessment, 41 Confinement, equivalence to natural stressors, 488 Confounding factors, experimental design, 425-426 Consciousness definition. 373. 374 evolution, 377-379 Conservation biology, 405-491 activity-related stressors attention given. 473-474 examples, 415. 474-486 minimizing consequences. 488-494 case studies Antarctic penguin, 462-463 mountain pygmy-possum. 417-418 North Sea harbor seal, 453-456 whale watching, 458-462 developmental instability, 195-197 equipment, improvements, 490-491 experimental design issues, 423-428. 489 importance of stress, 494-496 stress concept, 407-414 stress research contributions, 494-495 experimental design issues, 423-428 framework, 414-415 measurement issues, 421-423 questions about stress, 41.5, 417-420 Conservation ecology, self-medication, implications, 308 Conservation genetics, molecular techniques. 490 Control, see also Coping strategies definition, 8 developmental. energy costs, 184-185

531

relationship to physiological stress response, 7, 8-10, 63, 88-89, 437 testosterone level, 75-76 Control groups, 42.5 Cooperation training, 492-493 Coping strategies active. 10 appraisal, 12-13 passive, 3, 5 , 10-1 1 relationship to social status and physiological stress response, 88-89 stress concept, 411, 412-413 Coping systems relevance of feelings, 393-394, 395 welfare assessment, 400 Copulatory orgasm, females female choice, 335-337 with symmetrical men, 328 Coronary artery disease, chronic social stress, 70 Corticosteroid binding proteins. 432. 434 Corticosterone dominance relationships, 21-22, 67 species variation, 19 Corticotrophin-releasing hormone glucocorticosteroid release, 19 physiological effects, 140 postpartum depression, 345 Cortisol, species variation, 19 Cougar (Fehs concolor), hunting effects. 466 Courtship, predation, 243 Cows, asymmetry, milk production, 199 Crayfish, nocturnal activity pattern, 234 CRH, see Corticotrophin-releasing hormone Crickets, mate choice, predation, 242 Crippling, consequences, 469 Crocuta crocuta food aversions, 307 snare wounds, 469 Crowding, experimental immunological consequences, 81-82 population regulation, 43-44 Crying adult women, fitness benefits, 350 infant. see Infant crying Cry pitch, offspring condition, 354-355 Cryptomys damarensis, energy limits, nonsexual behavior, 158-159

532

INDEX

Culture. postpartum depression. 351 Cyprinodon pecosensis. sexual behavior, energy costs, 160 Cytokines, 38. 41. 137. 141-142

D Daily activity patterns, birds, 234 Damara mole rats (Cryptornys rlamnrmsis), energy limits, nonsexual behavior, 158-159 Damsel flies (Calopteryx maculntn). sexual behavior, energy costs, 159 Dasyurid marsupials (Antechinus). stress-induced male mortality, 1-2. 434-43s Data loggers, 491 Death, see Mortality Delichon urbicn, energy limits. nonsexual behavior. 159 Dendritic atrophy, glucocorticosteroidinduced, 19 Depression females vs. males, 341 function, 388 postpartum, 341-352 Despair, function. 389. 394 Development, environmental. equivalence to natural stressors, 487 Developmental control. energy costs. 184- 185 Developmental stabilityhstability animal welfare. 197-199 attractiveness, 328 behavioral, 201-206 conservation biology. 195-197 definition, 181 directional selection. 190-191 environmental causes, 184. 188-190 fitness correlates. 192-193 future studies. 206-207 genetic causes, 185-188 human and veterinary medicine, 199-201 as indicator of stressed state, 452 measurement, 181-1 84, 201-206 monitoring, 194- 195 morphological, 204-206 practical uses, 193-206 types. 321-323

Development rate, mate choice, 170 Die1 drift periodicity, stream insects. 234 Die1 migration fish, 233 zooplankton, 23 1-233 Diet selection learned aversions, 306-307 parasitism. 292, 308-309 predation. 224 relationship to stress, 147 self-medication, 309 social learning, 307 specific hungers, 305-306 Differential parental solicitude infant crying, 353 postpartum depression, 345-346 Directional asymmetry. 183 Directional selection. 190-191, 324 Disease adaptation, 6 equivalence to anthropogenic stressors. 4x7 induction, 40 relationship to stress. 39-40 resistance to behavioral, 137 definition, 136 genetically based. 136-137 immune-mediated, 136-137 secondary sexual characters. 323 social stress, 81-8.5 stress-induced impairment, 442-443 Displacement activities, as indicator of stressed state, 449 Diurnal activity patterns. 233 Diurnal rhythms. physiological parameters. 17 Divergence resource heterogeneity. 167 resource polymorphisms. 167-168 Dogs, stress reactions, predictability. 7 Dominance relationships, see also Social stress female. reproductive success, 76-81 male female choice. 85-86 physiological costs, 85-88 role of ACTH. 20-21 role of corticosterone, 21-22. 67

INDEX

stress responses, 45-89 active. 62. 63. 88-89 assessment of rival, 58-61 cardiovascular, 55. 66, 67-69 establishment of dominance, 51-58 female, 72 gonadal, 73-81 guinea pigs, 58-59 immune, 57. 60-61, 81-85 mice, 62-64 passive. 62, 63, 88-89 pituitary-adrenocortical, 61 -73 primates. 69-73, 83 rabbits, 64-66 rats, 66-69 relationship to control and predictability, 88-89 sympathetico-adrenomedullary, 61 -73 tree shrews, 51-58, 59-61. 76 testosterone, 63, 67. 73-76 Dopamine, stress response, 140 Dragonflies, ovipositional behavior, predation, 243 Drosophila energy limits. nonsexual behavior, 158 habitat preference, 157, 158 species boundaries, 162, 163 Ducks, crippling consequences, 469 Dugongs (Dngong dugon) handling effects, 477 pursuit hunting, 467 Dwarf mongooses (Hrlogule purvula), dominance relationships reproductive success, 78-79 stress responses, 72

E Eating pleasure, function, 391 Echinoderms, escape behavior, 238 Ecological stress, 31 9-320 Ecosystem regulation, top-down, indirect effects, 260 El Niiio, 472 Emotions functional significance, 342 initiation of feelings. 372 stress response, 6-7, 10, 50-51 Endotoxin. 147

533

Energetics adrenocortical response. 435-436 anthropogenic stressors, 439-441 predator avoidance, 223 Energetic stress postencounter resumption of activity, 236 state-dependent risk taking, 217-221 Energy balances, 155-156,439-441 Energy charge, adenine. 134, 446-447 Energy costs developmental control, 184-185 learning, 168-169 nonsexual behavior, 158-159 resource polymorphisms, 167-168 sexual behavior, 159-162 speciation, 165-166 species boundaries. 162-164 Energy metabolism. immunocompetence. I44 Environmental contamination, sre Pollution Environmental enrichment immunological consequences, 442 as stressor, 484 Environmental factors, developmental instability, 184, 188-190, 194-195 Environmental stress, 41 1 adaptation, 320 adrenocortical response sensitivity, 431-432, 433-434 prenatal and perinatal impact, 443, 445 resistance, genetic variation, 440, 495 resource allocation, 440 Envy, function, 390 Enzyme activities, sympatheticoadrenomedullary activation, 33 Epidemiological studies, 15, 39-40 Epinephrine biosynthesis. 31 physiological effects, 30-31 secretion, 29 stress response, 10. 140 Error measurement. fluctuating asymmetry, 183-184 statistical tests, 426-427 Escape behavior, postencounter. 237-238 Estrogen facial secondary sexual traits. 325 physiological effects, 36

534

INDEX

European wild rabbits (Orvcrolagus cuniculus) female, social rank, 78 physiological stress responses. 46-47 social bonds. stress-reducing effects, 94-98 Exhaustion stage, general adaptation syndrome. 6 Exhilaration, function. 391-392 Experience early, handling. 491 social role in stress responses, 12-13. 64 self-medication, 307 stress concept, 407-408 Experimental design. conservation biology. 423-428.489

F Face secondary sexual traits, 324-326 symmetry, 326-327 Falconiformes. plant use. 296 Father-daughter relations postpartum depression, 348-350 women’s sexual behavior, 334, 337-339 Fear definition, 384 function, 384-385 initiation, 373 responses, 384 Feather pulp sample. protein assessment, 490 Fecal samples, 25, 489-490 Fecundity. developmental instability. 193 Feeding behavior, see Foraging behavior Feelings adaptive effects. 378-379 characteristics. 372 consciousness, 373-374 definition, 374 evidence, 375-376 evolution, 376-379 examples, 372 function. 393-394 initiation, 372-373 as part of coping systems, 393-394. 395

pleasant vs. unpleasant. 374-375 relevance to welfare, 395. 397. 398-400 Female choice costs. 161-162 dominance relationships, 85-86 orgasm, 328, 335-337 sexual arousal, 335-337 Females dominance relationships, 72. 76-81. 85-86 mate choice, see Female choice secondary sexual characters, immunocompetence. 328 sexual behavior, humans, 332-341, 357-358 Fenitrothion contamination, 457 Ficediila hypoleuca, 159-160 Fight or flight response. 3-5, 10, 62 Fights, see Aggressive behavior Finch (Pinaroloxias inornoto), learning. 168 Fish courtship. predation, 243 die1 migration, 233 flight initiation distance. 237 predator inspection. 238-239 predator-prey population dynamics, 258 resource polymorphisms, 167-1 68 use of space, antipredator decision making, 255-256 Fitness consequences adrenocortical response, 431 examples, 423 importance of measuring, 423 stress response, 438-439 timescale. 421-422 developmental instability, 192-193 long-term monitoring. 422 mating. relationship to energy consumption, 159-160 measurement, 172-173,412. 423 metabolic efficiency, 171-173 offspring, assessment, 346-347. 353-354, 354-355 psychological pain. 342 reduction. stress concept, 411. 412-413 relationship to stress level, 156 secondary sexual characters, 325-326 stress-resistance. 170- 171. 319 Flight initiation distance, 237

INDEX

Fluctuating asymmetry, 321-323 definition, 181 etiology. 322 heterozygosity, 173 as indicator of stressed state, 452 individual, 181, 185. 192-193, 322 measurement, 173, 183-184 secondary sexual characters, 328-329 sexual selection, 322. 323-324 Follicle-stimulating hormone, 36 Foraging behavior energetic stress and state-dependent risk taking, 217-221 food selection, see Diet selection as indicator of stressed slate, 449-451 location. 224 mortality/growth rate rule. 221-222 patch choice, 222-223 role of modeling, 225 time in patches, 223 Foraging theory. self-medication, 308-309 Fractal dimension, as measure of developmental instability, 203-204 Freezing response. function, 384 Freudian theory, female sexual arousal, 337 Frogs courtship, predation. 243 seasonal activity patterns, 235 sexual selection. energy costs, 159 Frustration function. 386-387 initiation, 373 jealousy, 390 Fur rubbing, mammals. 298

G Gastric ulceration. rats, physiological stress response, 8-9 Gemsbok (Oryx gnzella), horn asymmetry, 196-1 97 Gene coadaptation, 188 General adaptation syndrome, 5-6, 62 Genetic diversity measurement, 419 preservation. 419, 495 under stress, 164-165, 419-420

535

Genetic factors developmental instability, 185-188 resistance to disease, 136-137 stress response, magnitude, 135 Genetic variability, 164-165, 202 Genotypes. stress-resistant, 169-171. 440, 495 Geophagy, primates, 295 Gerbils, moonlight avoidance, 235 Global warming, 472, 487 Glucocorticosteroids, 19-22 acute effects, 19 challenge tests, 26-29 circadian oscillation, 143 immune effects, 142-144 long-term effects, 19, 22 measurement. 23-29 ratio of free to bound, 24 synthesis and release. 22 transport proteins, 22 Gobies, nest building and defense. predation, 244 Gonadal activity, see also Pituitary-gonadal system adrenocortical activity negative coupling, 432-434 uncoupling, 434-435 dominance relationships, 73-81 as indicator of stressed state, 447-448 psychosocial stress, reproductive success, 80-81 stress response active vs. passive, 14 assessment, 35-37 Gonadotropin. suppression, psychosocial stress, 80 Gonadotropin-releasing hormone, 36 Good genes hypothesis good genotype vs., 172 stress-resistance, 170-171 Gorillas (Gorilla gorilla), dental asymmetry, 196 Great crested grebes (Podiceps crisratus), boating-induced stress, 413 Great tits (Parus major) sexual behavior, energy costs, 159-160 song drift, 203 Green turtles (Chelonia mydas), tourism-related disturbance, 464

536

INDEX

Grief displays, fitness benefits. 349-350 function. 385-386 recognition. 375 Grizzly bear (Ursus arcfos), hunting effects, 467 Group choice, predation, 239 Growth hormone, immune effects. 144 Growth performance, developmental instability, 193 Guilt, function, 387 Guinea pigs maternal-infant separation. 92 social bonds, stress-reducing effects. 94 stress responses dominance relationships, 58-59 role of social experiences, 12- 13 Guppies. mate choice, predation, 242

H Habitat preference energy balance, 156-158 hunting effects. 467-469 Habitat quality, 451-452 Haemutopits ostralegus. tourism-related disturbance, 463 Hair samples, 490 Hamadryas baboons (Papio hrmiotiryus). cardiovascular stress responses. dominance relationships. 69 Handling early experience, 491 equivalence to natural stressors, 486. 487 fitness consequences, 476-477 physiological response. 16- 17. 473-474. 475-476 predictable. 493 Happiness. function, 392 Harbor seals North Sea. case study. 453-456 stress-induced immunosuppression, 44 1 Health developmental, metabolic rate, 328 infant, postpartum depression. 347-348 relationship to animal welfare. 396, 397

Heart rate. see rrlso Cardiovascular system anthropogenic stress. 440-441 dominance relationships rabbits, 66 tree shrews. 55 as indicator of stressed state. 449 as indicator of sympatheticoadrenomedullary activity. 33-34 measurement techniques, 34 Heat shock proteins. 166. 429 Helogale parvulm. 72, 78-79 Helplessness learned. 388 passive chronic stress, 10-1 1 Hemifaces. attractiveness, 326-327 Hemipteran insects. mating dynamics. predation. 242 Herpes viruses. stress-induced reactivation, 443 Heterocephahis glcrber, reproductive success. social rank, 80 He terozygosi ty developmental instability. 173. 187 facial features, averageness. 327 relationship to fitness, 164-165. 171-173 Hippocampus, glucocorticosteroid effects. 19 Homeostasis measurement. 173 stress concept, 407. 410-41 1 Homozygosity. developmental instability, 186 Hormone concentrations catecholamines, 31-33 fecal, 25 glucocorticosteroids, 25-29 salivary, 25-26 serum o r plasma. 23-24 sex hormones, 37 urinary. 24-25 Hormones chemical immobilization and anesthesia effects, 478 initiation of feelings. 372 linkage to stress and immunocompetence, 142-145 psychological influences, 7 rhythms, 17 House martin (Delichon iirhica), energy limits, nonsexual behavior, 15Y

INDEX

House mice (Mus rniucuhcs), stress responses. 12 Housing, animal equivalence to natural stressors. 487-488 as stressor. 16, 483-485 Humans attraction and attractiveness, 321-332. 356-357 developmental stability. 321-323 secondary sexual characters, 323-324 stress resistance, 324-331 infant crying, 352-356, 358-359 postpartum depression, 341-352.358 sexual behavior. women, 332-341. 357-358 social bonds, stress-reducing effects, 98 stress response adrenocortical, 63, 64 sympathetico-adrenomedullary. 64 Humoral factors, immune system, 38, 40-41. 137 Humpback whale (Megoprero novueungliue), response to tourism, 459 Hunger, function, 383 Hunting behavioral effects, 467-469 crippling. consequences, 469 physiological and fitness consequences, 465-467 sports, 486. 487 Hybridization, developmental instability, 187-188 Hypothalamo-pituitary-adrenocorticalaxis, see Adrenocortical response; Pituitaryadrenocortical system Hypothalamo-pituitary-gonadal axis, see Pituitary-gonadal system

I Ideal free distribution model, patch choice. 222-223 Immobilization, chemical remote-controlled, 491 as stressor, 477, 478-479 Immune system adaptivelacquired immunity, 38. 137 anesthesia effects, 478-479 assessment problems, 139

537

behavioral conditioning, 141, 144 cellular factors, 38. 137 central connections, 38-39 chemical immobilization effects. 478-479 compensation, 146 components, 38 costs, 139. 146 energy metabolism, 144 enhancement socially induced. 100 stress-induced, 138, 143-144 environmental enrichment effects, 442 handicap hypothesis, 323, 350 humoral factors, 38, 40-41. 137 innatelnatural immunity, 37-38. 137 prenatal and perinatal stress, 443, 445 regulation. 137- 138 resistance to disease. 136-137 robustness, 138 secondary sexual characters, 86, 323, 327-328 stress response, 14, 38-39 assessment. 37-42 compensatory, 138 dominance relationships, 57, 60-61, 81-85 hormonal linkage, 142-145 neurological linkage, 140-142 relationship to disease, 39-40 testosterone, 87-88 suppression catecholamines, 39 sex hormones, 87-88, 325 socially induced, 90-91, 92-93, 96-98 stress-induced, 136, 145-149, 441-443 tearfulness, 350 surgery effects. 479 transportation effects, 485 Inbreeding cheetahs, 495-496 developmental instability, 186 Inclusive fitness, definition of stress, 134-135 Infant health, postpartum depression, 347-348 maternal bonds, 90-92, 346 lnfant crying acoustics, offspring condition, 354-355 honest signal model, 354 offspring need model, 353-354

538

INDEX

parental reactions, 355 phenotypic quality, 352-356. 358-359 postpartum depression, 358-359 Infanticide. 346. 350-351 Infection, see also Disease assessment, fecal samples, 489 susceptibility. stress-induced, 42, 147- I48 Intelligence test, body asymmetry, 206 Interleukin 1 circadian oscillation, 143 linkage to stress and immunocompetence. 141-142, 143 Internal changes, stress concept, 409 Interspecific competition behaviorally transmitted indirect effects, 259-261 developmental instability, 192 Interventions optimizing. 493-494 stressful, see Conservation biology, activity-related stressors Intraspecific competition, developmental instability, 192 Invasive procedures. alternatives, 489-490 Invertebrates, use of space, antipredator decision making, 255

energy expenditure. 168-169 self-medication. 305-307 Lepomis macrochinis, learning, 169 Lepus americanics, predator-prey population dynamics, 258-259 Lesser snow geese (Chen caerulescens), hunting effects, 468 Lethality of effects, stress concept, 41 1 Leukocytes, assessment, 40-41 Leydig cells, 36 Life history models, stress response traits, 437-438 Life-span. reproductive success, 170 Lizards courtship, predation, 243 flight initiation distance, 237 parasite stress, energy costs, 160 species boundaries, thermal environment. 163-164 Loneliness, function, 389 Lunar activity patterns, 234-235 Lust, function. 390 Luteinizing hormone, 36 Lycaon pictus, dominance relationships, stress responses. 72 Lymphocytes. 38. 40, 41, 138 Lymphoid organs, sympathetic innervation. 140-141 M

J Macaca fasciciilrzris. 70, 71, 93

Java monkeys (Maraca fascicularis) coronary artery disease, chronic social stress, 70 dominance relationships, stress reactions, 71 social bonds, disruption, 93 Jealousy, function, 390

L Laboratory selection experiments, 191 Lagopiis lagoptis. hunting effects, 468 Lagopus muhis, tourism-related disturbance, 463 Learning, see also Social experience developmental instability, 207

Macaques maternal-infant separation, 90-91 social stress. immunological consequences, 441-442 Malaise, function, 381-382 Males dominance relationships female choice, 85-86 physiological costs, 85-88 father-daughter relations, 334, 337-339. 348-350 mating tactics, predation. 242 secondary sexual characters, immunocompetence, 327-328 stress-induced mortality, 1-2, 434-435 Mammals active vs. passive stress responses. 62. 63

539

INDEX

conservation research recommendations, 489 female, social rank. reproductive success, 76-81 fur rubbing, 298 marine. hunting effects. 467 self-medication, 293. 298 social stress, 42-106 specific hungers. 306 use of space, antipredator decision making, 256-257 Manatees (Trichechus manatus). tourism-related disturbance. 465 Manual restraint. as stressor, 477-478 Marine animals hunting effects, 467 noise disturbance, 471 Marmoset monkeys, social rank, reproductive success, 79-80 Marrnota monax, flight initiation distance, 237 Marsh harrier (Circus aeruginosirs), tourism-related disturbance, 464 Mate choice costs, female preference, 161-162 development rate, 170 female, dominance relationships, 85-86 predation, 241-242 Mating predation, 242-243 stress-resistance, 169-170 Mayflies nocturnal activity pattern, 234 predator-induced changes, 245, 248 Measurement issues conservation biology research, 421-423 stress induction. 476-477, 479 Medicine, see also Self-medication Darwinian. 199-201 developmental instability, 199-201 Megapiera novaeangliae, response to tourism, 459 Metabolic energy. as indicator of stressed state. 446-447 Metabolism adrenocortical response, 435-436 developmental health. 328 heterozygosity, I71 -173 stress response, 14 Metal pollution, developmental instability. 19s

Mice crowding stress. 165 dominance relationships gonadal endocrine activity, 74-75 stress responses, 62-64 stress-induced immunosuppression, 81-82. 84-85,442 stress responses. 12. 62-64 Microtiis ochrogaster, social bonds. stress-reducing effects, 100, 103 Mineralocorticosteroids. 22 Molecular markers, stressed state, 446-447 Mole rat (Spalax ehrenhergi), energy limits, nonsexual behavior, 159 Molluscs, escape behavior. 238 Monkeys, maternal-infant separation, 90-91 Moonlight avoidance, 235 Mortality aggressive behavior, 1-2, 434-435 chronic social stress. 70-71 parasitism. 147-149 psychosocial stress, 49-51 stress-induced emotional arousal, 50-51 Mortality/feeding rate (pff) rule. 221-222 Mortality/growth rate ( d g ) rule diet selection, 224 optimal behavior, 221 -222 patch choice, 223 Mother-daughter relations postpartum depression. 338 women's sexual behavior, 334-335 Mother-infant bonds. 90-92, 346 Mountain goats (Oreamnos americus), tourism-related disturbance. 464-465 Mountain pygmy-possum (Burramys parvus), tourism-related stress, case study, 415,417-420 Movement, prey activity, 226 MRNA, as indicator of stressed state, 446 Mule deer, tourism-related disturbance, 464 Muscle biopsies, as stressor, 479-480 MUSmusculus, stress responses, 12 Mutations, developmental instability, 188 Myopathy, capture, 422, 467, 477

N Naked mole rats (Heterocephalus glaher), social rank, reproductive success, 80 Natural catastrophes, equivalence to anthropogenic stressors, 487

540

INDEX

Natural experiments, design issues, 424 Natural killer cells, 38, 40, 41 Natural stressors equivalence to anthropogenic stressors. 486-488 examples, 415 Nausea, function. 382 Nepotism, physical attractiveness, 331 Nervous system asymmetry, 205-206 autonomic, 29, 140-141 central, 20-21, 38-39 initiation of feelings, 372 sympathetic, see Sympatheticoadrenomedullary system Nest material, antibacterial, 296-297 Neuroendocrine stress response, see Stress response Nociceptive systems, 380 Nocturnal activity patterns, 234 Noise stress, 189, 469-471 Noninvasive procedures, 489-490 Norepinephrine biosynthesis, 31 secretion, 29 stress response. 10. 140 North Sea harbor seals, case study, 453-456 Nutrition, see Diet selection Nutritional stress, 447, 451 developmental instability, 188-189 immune effects. 139

0 Observational studies, design issues, 424 Odocoileris virginianus, 164, 237, 468-469 Offspring fitness, parental assessment, 346-347. 353-354 Oil fouling. consequences, 456-457 Olive baboons (Papio anubis) coronary artery disease, chronic social stress, 70 dominance relationships female reproductive success, 81 immunological consequences, 83 Omegas, 69 Orerimnos umericus, tourism-related disturbance, 464-465 Organochlorines, 455. 490

Orgasm, female costs, 336 female choice, 328. 335-337 frequency father-daughter relations, 338-339 female choice, 335-337 value judgments, 340-341 Ornaments, see also Secondary sexual characters energy costs, 161-162 Oryctolagus cuniculus, 46-47, 78, 94-98 Oryx gazella, horn asymmetry, 196-197 Osmoregulation, stress response, 14 Ovaries, stress response, 36, 37 Ovipositional behavior, predation, 243 Ovis canadensis, heterozygosity and fitness. 164 Oxytocin female orgasm, 336 social bonds, 103 stress-reducing effects, 103 Oystercatchers (Haematopus ostralegus), tourism-related disturbance, 463

P Pain definition. 373, 379-380 extreme, 381 function, 381 initiation, 373 responses, 380-381 system, 380 Pair bonding costs of orgasm, 336 stress-reducing effects, 98-100. 103 Pantherrr leo. pursuit hunting, 467 Papio anribis chronic social stress, coronary artery disease, 70 dominance relationships female reproductive success, 81 immunological consequences, 83 Papio hamadryas, dominance relationships, cardiovascular stress responses, 69 Paragliding disturbance, 464 Parasitism contribution to mortality, 147-149 definition, 291-292

INDEX

developmental instability, 192, 200 diet selection, 292, 308-309 effects on host behavior, 292 energy costs, 161 human attractiveness, 326 sexual behavior. 160-161 susceptibility, stress-induced. 42, 147-148 Parental investment physical attractiveness. 331 postpartum depression, 345-351 Parental solicitude infant crying, 353 postpartum depression, 345-346 Parent-daughter relations, women’s sexual behavior, 332-341, 337-339 Parenting, predation, 243-244 Parent-offspring conflict theory, 345-346 Pariis miijor sexual behavior, energy costs, 159-160 song drift, 203 Passerines, plant use, 296 Passive stress response, 108 chronic stress, 10-11 dominance relationships, 62. 63, 88-89 gonadal activity, 14 maternal separation, 90, 91 Selyean. 5-6. 62 tree shrews, 54-58 Patch choice. 222-223 Patch location, 224 Patch residence times, 223 Paternal investment postpartum depression, 348-350 women’s sexual behavior, 334. 337-339 Pathogens. see Disease PCB contamination, 455. 457, 490 Penguins. tourism-related disturbance, case study. 462-463 Perinatal stress, 443, 445 Petaurus breviceps, dominance relationships, testosterone, 73-74 Phagocytes. 38, 41 Phenodeviant behavior morphological basis, 204-206 repeatability. 202-203 Phenotype deviance, 183.321, 352-356 Pheny lethanolamine-N-methyltransfcrase epinephrine conversion, 31 regulation, variability, 31

541

sympathe tico-adrenomedullary activity. 33 Phoca vitulina, stress-induced immunosuppression, 441 Phocine distemper virus, 453-454 Physical activity, immune effects, 147 Physical attractiveness, see Attraction and attractiveness Physiological parameters assessment of stress, 15-16 evidence of feelings, 376 rhythms, 17 Pied flycatchers (Ficedula hypoleiica). sexual behavior, energy costs. 159-160 Pigs. stress-induced immunosuppression, 82-83, 442 Pigtail macaques (Macaca nemestrina). social stress, immunological consequences, 83,441-442 Pinaroloxias inornata, learning, 168 Pink-footed geese, tourism-related disturbance, 464 Pipefish (Syngnathus typhle). mating dynamics. predation, 243 Pituitary-adrenocortical system assessment, 19-29, 448 dominance relationships, 61-73 immune system connections, 39 regulation, 140 stress response, see Adrenocortical response Pituitary-gonadal system. see also Gonadal activity psychosocial stress. reproductive success, 80-81 stress response, assessment, 35-37, 447-448 Plant breeding, developmental instability, I91 Plant use. see also Self-medication birds, 296-297 secondary metabolites, 293 Pleasure. 374-375 achievement, 393 eating, 391 sensory, 392-393 sexual, 391-392 Plecoptera, species boundaries, 163 PNMT, see Phenylethanolamine-Nmethyltransferase

542

INDEX

Podiceps crisrarus. boating-induced stress. 413 Pollution case study, 453-456 developmental instability, 189, 194-195 equivalence to natural stressors. 487 experimental studies, 456-457 monitoring techniques, 490 Population censuring, disturbance, 474-475 Population density developmental instability, 189 equivalence to anthropogenic stressors. 488 as indicator of stressed state, 451-452 social stress, 42-44 Population dynamics predator-prey interactions, 257-259 stress response, 430-431 Population persistence, 41 1-412, 420 Postpartum depression, 341-352, 358 continuum, 344-345 cultural traditions, 351 infant crying, 358-359 infant health, 347-348 infanticidal ideation, 350-351 parental investment behaviors, 345-351 paternal investment, 348-350 prevalence, 343-344 resource base, 350 social support, 347-348, 350 tearfulness, fitness benefits, 350 Postpartum psychosis. 344 Power, statistical tests, 426-427 PPD, see Postpartum depression Prairie vole (Microrus ochrogaster), social bonds, stress-reducing effects, 100, 103 Predator-induced stress, see also Antipredator decision making definition. 215-216 equivalence to anthropogenic stress, 486-487 long-term costs, 245, 248 physiological response, 261-262 role of modeling, 225 Predictability handling effects, 493 relationship to stress response. 7. 8-10, 63.88-89,437 testosterone level, 75-76 Preference tests, role, 397-398

Prenatal stress impact, 443.445 primates, noise disturbance. 470 Preputial glands, social rank, 75 Primates conservation research recommendations, 48’) dominance relationships immunological consequences, 83 reproductive success, 77-78 stress responses, 69-73 geophagy. 295 noise disturbance, 470 Prisoner’s dilemma, fish, 239 Progesterone, physiological effects, 36 Prolactin, immune effects, 145 Proteins assessment, feather pulp sample, 490 corticosteroid binding, 432, 434 heat shock, 166, 429 transport, 22. 24 Pseudoreplication, 427-428 Psychical stress, 6-7 Psychological pain as evolutionary adaptation, 341-343 postpartum, 343-345 Psychology, stress concept, 407-408 Psychoneuroimmunological research, 45 Psychosocial stress, 48-50 mortality, 49-51 reproductive success, 80-81 Ptarmigan (Lagopus nzutus), tourism-related disturbance, 463 Pupfish (Cyprinodon pecosensis). sexual behavior, energy costs, 160 Pursuit-deterrence signals, 236-237 Pursuit hunting, 465-466

R Rabbits, see European wild rabbits (Oryctolagus cuniculus) Radiation, developmental instability, 194-1 95 Radio implantation, as stressor. 480 Radio-tagging. as stressor, 480-483 Radiotelemetry blood pressure, 35 heart rate, 34 improvements, 49 1

543

INDEX

Rails (Aves). pursuit-deterrence signals, 237 Random selection. 423-425 Rangifer tarandus, 468 Rape, victims’ anguish, 342-343 Rats diurnal activity pattern. 233 food selection, 305. 306-307 maternal-infant separation. 92 social stress. immunological consequences, 82, 84 stress response dominance relationships, 66-69 predictability and control, 8-9 Razorfish (Xvrichtys splendens), mating dynamics, predation, 242-243 Reciprocity. physical attractiveness, 330-331 Recreation disturbance, see Tourism/ recreation-related disturbance Red deer hunting effects, 466, 468 noise disturbance, 470 Red-winged blackbirds (Agelaius phoeniceus), food aversions. 307 Refuging, prey activity, 226 Reindeer (Rongifer tarandus), hunting effects, 468 Release, soft, 492 Repeatability individual fluctuating asymmetry, 185 as measure of developmental stability, 202-203 Reproduction. see also Gonadal activity adrenocortical response, 431-436 costs, parasitism, 148 decision making, predation, 241-245 dominance relationships, females, 76-81 parental effort, immune effects, 145 phenotypic variation, stress resistance, 319, 324-330 prenatal stress, 43-44, 73 social rank, female mammals, 76-81 social stress, 43-44, 73 Reproductive system assessment, molecular techniques, 490 state. as indicator of stressed state. 447-448 stress response, 14

Reptiles, use of space, antipredator decision making, 256 Resistance stage, general adaptation syndrome, 5-6 Resources heterogeneity, adaptation, 166-167 polymorphisms, energy expenditure, 167-168 Resource stress allocation differences, 439-441 equivalence to anthropogenic stressors, 487-488 postpartum depression, 350 sexual behavior, 160,333-334.338-339, 340 Restraint, as stressor. 477-479 Rhesus monkeys (Macaca mularta) dominance relationships, 70-71 immunological consequences, 83 stress responses, 69. 72 maternal-infant separation, 90 social bonds, disruption, 92-93 stress-induced immunosuppression, 441 Rhythms die1 drift periodicity, 234 disturbance, timescafe, 421 physiological parameters, 17 Risk assessment antipredator decision making, 262-263 coping style, 12-13 dominance relationships, stress responses, 58-61 Rock doves (Columbia livia), parasite stress. energy costs, 160 Rodents, moonlight avoidance, 235 Rubia, self-medication, chimpanzees, 302-303,304

5

Sage grouse (Cenfrocercus urophasianus), sexual behavior. energy costs, 160 Saharan silver ant (Cataglyphis bombycina), adaptation, abiotic stresses, 166 Saimiri sciureus, see Squirrel monkeys Saliva samples, 25-26, 490 Salt, specific hunger, 306 Sample size. 426

544 Sceloporirs occidentalis. stress response,

variability. 430-431 Schizophrenia, phenodeviance, 203 SDP, see Stochastic dynamic programming Seasonal activity patterns, 235 Secondary sexual characters, see also Sexual selection asymmetry, 190-191,323-324,328-329 developmental instability. 192-193, 323-324 facial, 324-326, 327 immunocompetence. 86, 323, 327-328 nonfacial, 327-330 stress, 191. 323-324 Sedatives, 493 Selection directional. 324 sexual, see Sexual selection stress-related, 31Y-320 Self-medication behavioral mechanisms, 304-307 chimpanzees, 298-303 contexts. 292 forms. 294 future research, 310 implications. 308-310 mammals, 293. 298 prophylactic, 294-298 baboons, 295-296 birds. 296-298 primates, 295 therapeutic vs.. 294 secondary plant metabolites, 293 sexual selection, 293 skepticism, 303-304 therapeutic, 294, 298-303 individual learning, 305-307 social learning. 307 Selye’s general adaptation syndrome, 5-6, 62 Sensation initiation of feelings, 372 pleasurable. function. 392-393 Separation. mother-infant, 90-92 Serotonin, stress response, 140 Sex determination, fecal samples, 489 Sex hormones measurement. 37 stress response, assessment, 35-37

INDEX

Sex steroids stress response, 22 transport proteins, 22 Sexual arousal, father-daughter relations, 337-339 Sexual behavior, women arousal father-daughter relations, 337-339 female choice, 335-337 flexibility. 334 parent-daughter relations, 332-341, 357-358 resource level. 333-334. 338-339 Sexual pleasure, function, 391-392 Sexual selection, see ulso Secondary sexual characters developmental instability. 190-191, 192-193 energy costs. 159-162 fluctuating asymmetry, 322, 323-324 parasite stress, 86, 161 self-medication, 293 stress resistance. 170-171 humans, 324-330 Sheep, parental effort, immune effects, 145 Signal evolution fitness impacts, 353 infant crying. 350, 354 Sinusitis. facial secondary sexual characters. 326 Size-assortative grouping, predation, 240 Sleep, function, 383 Sneaky copulations, predation, 242 Snowshoe hare (Lepus americnni~s), predator-prey population dynamics, 258-259 Social behavior nonsexual. stress resistance, 330-331 sexual, see Sexual behavior Social b o n d s h p p o r t costs of orgasm, 336 disruption, 89-93 oxytocin. 103 postpartum depression, 347-348 stress-reducing effects. 93-103, 493 Social density. 44 Social experience. see also Learning role in stress responses, 12-13, 64 self-medication, 307

INDEX

Social factors adrenocortical response. 436-437 antipredator decision making. 239-241 stress response, 89-103, 436-437 Sociality, adaptive. 239-241 Social rank, see Dominance relationships Social stress, 42-105 aggressive behavior. 45-47 behavioral research. 44-45 disruption of social bonds, 89-93 dominance relationships. see Dominance relationships population regulation, 42-44, 73 psychological vs. physical processes. 48-50 reproductive success. 80-81 resistance to disease. 81-85 resource allocation, 440 social conflict. 45-61 social support effects, 93-103 tree shrews, 50-58 Social systems, stability, 88-89. 488 Social vigilance, 240-241 Socioeconomic level postpartum depression. 350 women's sexual behavior. 340 Sodium, specific hunger, 306 Soft release. 492 Space, use. predator-induced stress. 248-257 Spnlm ehrenhergi. energy limits. nonsexual behavior, 159 Spatial positioning, predation, 239 Speciation. energy costs, 165-166 Species boundaries, energy limits, 162-164 Species interactions. antipredator decision making. 259-261 Species variation. stress response, 430 Spider mite (Tetraqvchus urticrie). habitat preference, 1.57 Sports hunting, equivalence to natural stressors. 486. 487 Spotted hyenas (Crocirtn crociirri) food aversions. 307 snare wounds. 469 Squirrel monkeys (Sairniri sciureus) dominance relationships stress reactions. 71 testosterone. 74 maternal-infant separation. YO

545

social bonds disruption, 90, 92 stress-reducing effects, 93-94 Stability, see also Control developmental, see Developmental stability/instability relationship to stress response. 7. 8-10, 63, 88-89,437 social systems, 88-89 Starlings (Stirrnus vulgaris). plant use. 296-297 State-dependent life history models, stress response traits, 437-438 State-dependent risk taking empirical studies, 217. 218-219 stochastic model. 217, 220-221. 224, 236 Statistical tests, power, 426-427 Stimulant use, baboons, 295-296 Stochastic dynamic programming. 217, 220-221.224. 236 Stoneflies (Plecoptera). species boundaries, 163 Stream insects, die1 drift periodicity, 234 Stress, see also specific type definition, 2. 14-15, 395-396, 407 biological, 291 classical. 216 evolutionary. 215-216. 412-414 operational. 134-135 physical, 291 Selyean, 5 natural history. 414-415, 428-452 relationship to disease. 39-40 relationship to welfare, 396. 397 systems view. 407 types. 62 Stress concepts development. 3-15 distinguishing criteria, 407-41 1 evolutionary. 41 1-414.437-438 Stress-odor exposure. mice, immunological consequences, 84 Stressors anthropogenic, see Anthropogenic stressors conservation activity-related attention given, 473-474 examples. 41.5. 474-486 minimizing consequences, 488-494 definition, 407

546 natural, 415, 486-488 physical vs. psychical, 6-7 temporal characteristics, 423 Stress proteins, 166, 429 Stress resistance genotypes, 169-171,440,495 metabolic rate, 435-436 Stress response active, see Active stress response as adaptation, 438-439 adrenocortical, see Adrenocortical response adrenomedullary, see Sympatheticoadrenomedullary system, stress response antipredator decision making. 261-262 assessment, 445-452 epidemiological approach, 15, 39-40 methodological issues. 16-18. 420428 molecular markers. 446-447 organisrnal markers, 446-450 physiological approach, 15-16 physiological markers, 18-42 Cannon’s, 3-5, 62 cellular, 429 conceptual issues. see Stress concepts definition, 407 emotions, 6-7, 10, 50-51 magnitude, genetic component, 135 modulating factors, 416, 429. 430-431 multiple components. 421 nonspecific. 7-8, 14 passive. see Passive stress response physiological pattern, 13-14 predictability and control, 7, 8-10, 63, 88-89.437 Selye’s, 5-6, 62 social factors, 89-103. 436-437 sympathetic, see Sympatheticoadrenomedullary system. stress response theories. see Stress concepts timescale, 421-422 variability, 429-439. 438 Stress state definition. 407 indicators, 445-452 organismal, 446-450 population. 450-452

INDEX

Strrrnrrs vulgoris. plant use. 296-297 Submission, see Dominance relationships Suffering definition, 374-375 function, 389, 394 Sugar. specific hunger, 306 Sugar gliders (Petourus hreviceps), dominance relationships, testosterone. 73-74 Surgery, as stressor, 479-480 Survival aids, 378 Survival handicaps. 323. 350 Survival rates, developmental instability, 193 Survival variants, 165 Swine, social bonds, disruption, 93 Symmetry, see also Asymmetry facial, 326-327 selection pressure, 184 Sympathetico-adrenomedullary system, 20 active chronic stress, 10 fight or flight response. 3 immune system connections, 39 physiological effects, 29-31 regulation, 140 in Selyean concept, 5 stress response assessment, 29-35, 31-35, 448-449 dominance relationships, 54-55, 56-57. 61-73 rabbits, 46 tree shrews, 54-55, 56-57 Syngnrrthits typhle, mating dynamics, predation, 243 Syrian hamsters, dominance relationships, immunological consequences, 82

T Tagging, as stressor, 480-483 Tamarins, social rank, reproductive success, 79-80 Tearfulness. fitness benetits, 350 Temperature, measurement. as stressor, 479 Temperature deviations climatic, 472. 487 developmental instability. 188 equivalence to anthropogenic stressors, 487

547

INDEX

Temporal patterns as indicator of stressed state. 449-450 prey activity, 231-235 Testes. stress response, 35-36, 37 Testosterone aggressive behavior, 75. 86-87 dominance relationships, 63, 67, 73-76 facial secondary sexual traits, 325 immunological consequences, 87-88, 325 physiological effects, 35 predictability and control, 75-76 stress effects, 36-37 Tefranychus urficae, habitat preference. 157 TH, see Tyrosine hydroxylase Thermal discomfort, function. 383-384 Thiarubrine-A, 300, 301 Thirst function. 383 initiation, 373 Thymus, noradrenergic innervation, 140-141 Thyroid hormone, immune effects, 144 Tidal activity patterns, 235 Time patch residence, 223 stress concept, 408-409 Timescale animal welfare assessment, 398 conservation biology research, 421 -422 Tiredness, function, 383 Tit-for-tat strategy, fish, 239 Titi monkeys (Cdlicehus moloch), social bonds, stress-reducing effects, 94 T lymphocytes, 38, 40. 41 Top-down ecosystem regulation, indirect effects. 260 Torpor, as indicator of stressed state, 447 Tourism/recreation-related disturbance, 457-465 Antarctic penguins. case study, 462-463 development-related stress, case study, 415. 417-420 experimental studies. 463-465 hunting, 465-471 whale watching, case study, 458-462 Training antipredator, 492 cooperation, 492-493 Tranquilizers. 493

Translocation equivalence to natural stressors. 488 soft release, 492 as stressor, 485-486 Transmitters improvements, 491 radiotelemetry, 34, 35 weight, 481-482 Transportation equivalence to natural stressors. 488 as stressor, 485-486 Transport proteins, 22, 24 Tree shrews challenge test, 27-29 social bonds, stress-reducing effects, 98- 100 stress response active and passive, 54-58 dominance relationships, 51-58, 59-61, 76 new housing, 16 social stress, 50-58 Trichechus manatus, tourism-related disturbance, 465 Tyrosine hydroxylase adrenal concentration, 31 catecholamine biosynthesis, 31 as measure of sympatheticoadrenomedullary activity, 33

U Ulcers, rats, predictability and control, 8-9 Ungulates, pursuit-deterrence signals, 237 Urbanization, equivalence to natural stressors, 487 Urine samples, 24-25, 451, 490 Urscts americanus, tourism-related disturbance, 464 Ursus arctos, hunting effects, 467

V Vaccination long-term monitoring, 422 as stressor, 483 Variability, genetic, 164-165. 202 Variants, survival, 165

548

INDEX

Vernonia, self-medication, chimpanzees, 301-302 Vervet monkeys (Cercopifhecus nethiops sohoeus), dominance relationships, stress reactions, 71 Veterinary medicine, developmental instability, 199-201 Vigilance. social, 240-241 Viruses. stress-induced reactivation, 443 Visitor disturbance, see also Tourismi recreation-related disturbance equivalence to natural stressors, 486. 487 Vocalizations, behavioral asymmetry, 204-205 Voles, breeding suppression, predation. 244-245. 259

Wildlife conservation research recommendations, 489 hunting effects, 465-469 noise disturbance, 469-471 tourism-related disturbance, see Tourism/ recreation-related disturbance vaccination effects, 483 Willow grouse (Lagopus Iugopirs). hunting effects, 468 Woodchucks (Morninfa monux), night initiation distance. 237

X Xvrichfys splendens, mating dynamics. predation. 242-243

W Waist-to-hip ratio, phenotypic quality, 328-329 Water, specific hunger, 305-306 Waterfowl, hunting effects, 467-468 Water striders mate choice, predation, 242 mating dynamics, predation, 242 Welfare, see Animal welfare Western fence lizard (Sceloporcrs occidentrilis),stress response, variability, 430-431 Whales hunting effects, 466-467 tourism, case study, 458-462 White-tailed deer (Odocoileus virginionus) heterozygosity and fitness. 164 hunting effects. 468-469 pursuit-deterrence signals, 237

Y Yellow baboons, dominance relationships, immunological consequences, 85

2

Zahavian paradigm, signal evolution, 350, 353,354 Zona fasciculata enlarged, long-term stress. 19 glucocorticosteroid production, 19 Zona glomerulosa. mineralocorticosteroid production, 22 Zooplankton. die1 vertical migration, 23 1-233 Zoos, housing stress, 484

~~

~~

~~

~

Contents of Previous Volumes Volume 16 Sensory Organization of Alimentary Behavior in the Kitten K. V. SHULEIKINA-TURPAEVA Individual Odors among Mammals: Origins and Functions ZULEYMA TANG HALPIN The Physiology and Ecology of Puberty Modulation by Primer Pheromones JOHN G. VANDENBERGH AND DAVID M. COPPOLA Relationships between Social Organization and Behavioral Endocrinology in a Monogamous Mammal C. SUE CARTER, LOWELL L. GETZ, AND MARTHA COHEN-PARSONS Lateralization of Learning Chicks L. J. ROGERS Circannual Rhythms in the Control of Avian Migrations EBERHARD GWINNER

Behavioral Ecology: Theory into Practice NEIL B. METCALFE AND PAT MONAGHAN The Dwarf Mongoose: A Study of Behavior and Social Structure in Relation to Ecology in a Small, Social Carnivore 0. ANNE E. RASA Ontogenetic Development of Behavior: The Cricket Visual World RAYJOND CAMPAN, GUY BEUGNON. AND MICHEL LAMBIN

Volume 18 Song Learning in Zebra Finches (Taeniopygia guttata): Progress and Prospects PETER J. B. SLATER, LUCY A. EALES, AND N. S. CLAYTON Behavioral Aspects of Sperm Competition in Birds T. R. BIRDHEAD

The Economics of Fleeing from Predators R. C. YDENBERG AND L. M. DILL

Neural Mechanisms of Perception and Motor Control in a Weakly Electric Fish WALTER HEILIGENBERG

Social Ecology and Behavior of Coyotes MARC BEKOFF AND MICHAEL C. WELLS

Behavioral Adaptations of Aquatic Life in Insects: An Example ANN CLOAREC

Volume 17

The Cicadian Organization of Behavior: Timekeeping in the Tsetse Fly, A Model System JOHN BRADY

Receptive Competencies of LanguageTrained Animals LOUIS M. HERMAN Self-Generated Experience and the Development of Lateralized Neurobehavioral Organization in Infants GEORGE F. MICHEL

Volume 19 Polyterritorial Polygyny in the Pied Flycatcher P. V. ALATALO AND A. LUNDBERG 549

550

CONTENTS OF PREVIOUS VOLUMES

Kin Recognition: Problems, Prospects. and the Evolution of Discrimination Systems C. J. BARNARD Maternal Responsiveness in Humans: Emotional, Cognitive. and Biological Factors CARL M. CORTER AND ALISON S. FLEMING The Evolution of Courtship Behavior in Newts and Salamanders T. R. HALLIDAY Ethopharmacology: A Biological Approach to the Study of Drug-Induced Changes in Behavior A. K. DIXON, H. U. FISCH, AND K. H. MCALLISTER Additive and Interactive Effects of Genotype and Maternal Environment PIERRE L. ROUBERTOUX, MARIKA NOSTEN-BERTRAND, AND MICHELE CARLIER Mode Selection and Mode Switching in Foraging Animals GENE S. HELFMAN Cricket Neuroethology: Neuronal Basis of Intraspecific Acoustic Communication FRANZ HUBER Some Cognitive Capacities of an African Grey Parrot (Psirrnclts erirhncits) IRENE MAXINE PEPPERBERG Volume 20 Social Behavior and Organization in the Macropodoidea PETER J. JARMAN The 1 Complex: A Story of Genes. Behavior. and Population SARAH LENINGTON *

Lekking in Birds and Mammals: Behavioral and Evolutionary Issues R. HAVEN WILEY Volume 21 Primate Social Relationships: Their Determinants and Consequences ERIC B. KEVERNE The Role of Parasites in Sexual Selection: Current Evidence and Future Directions MARLENE ZUK Conceptual Issues in Cognitive Ethology COLIN BEER Responses in Warning Coloration in Avian Predators W. SCHULER AND T. J. ROPER Analysis and Interpretation of Orb Spider Exploration and Web-Building Behavior FRITZ VOLLRATH Motor Aspects of Masculine Sexual Behavior in Rats and Rabbits GABRIELA MORAL^ AND CARLOS BEYER On the Nature and Evolution of Imitation in the Animal Kingdom: Reappraisal of a Century of Research A. WHITEN AND R. HAM Volume 22 Male Aggression and Sexual Coercion of Females in Nonhuman Primates and Other Mammals: Evidence and Theoretical Implications BARBARA B. SMUTS AND ROBERT W. SMUTS

The Ergonomics of Worker Behavior in Social Hymenoptera PAUL SCHMID-HEMPEL

Parasites and the Evolution of Host Social Behavior ANDERS PAPE M0LLER, REIJA DUFVA, AND KLAS ALLANDER

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

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

CONTENTS OF PREVIOUS VOLUMES

Proximate and Developmental Aspects of Antipredator Behavior E. CURIO Newborn Lambs and Their Dams: The Interaction That Leads to Sucking MARGARET A. VINCE The Ontogeny of Social Displays: Form Development. Form Fixation, and Change in Context T. G. GROOTHUIS

551

The Behavioral Diversity and Evolution of Guppy, Poecilin rericrtlata, Populations in Trinidad A. E. MAGURRAN. B. H. SEGHERS. P. W. SHAW, AND G. R. CARVALHO Sociality, Group Size, and Reproductive Suppression among Carnivores SCOTT CREEL AND DAVID MACDONALD

Volume 23

Development and Relationships: A Dynamic Model of Communication ALAN FOGEL

Sneakers. Satellites. and Helpers: Parasitic and Cooperative Behavior in Fish Reproduction MICHAEL TABORSKY

Why Do Females Mate with Multiple Males? The Sexually Selected Sperm Hypothesis LAURENT KELLER AND HUDSON K. REEVE

Behavioral Ecology and Levels of Selection: Dissolving the Group Selection Controversy LEE ALAN DUGATKIN AND HUDSON KERN REEVE Genetic Correlations and the Control of Be. havior. Exemplified by Aggressiveness in Sticklebacks T H E 0 C. M. BAKKER Territorial Behavior: Testing the Assumptions JUDY STAMPS Communication Behavior and Sensory Mechanisms in Weakly Electric Fishes BERND KRAMER Volume 24

Is the Information Center Hypothesis a Flop? HEINZ RICHNER AND PHILIPP HEEB

Cognition in Cephalopods JENNIFER A. MATHER

Volume 25

Parental Care in Invertebrates .STEPHEN T. TRUMBO Cause and Effect of Parental Care in Fishes: An Epigenetic Perspective STEPHEN S. CRAWFORD AND EUGENE K. BALON Parental Care among the Amphibia MARTHA L. CRUMP An Overview of Parental Care among the Reptilia CARL CANS Neural and Hormonal Control of Parental Behavior in Birds JOHN D. BUNTIN

Maternal Contributions to Mammalian Reproductive Development and the Divergence of Males and Females CELlA L. MOORE

Biochemical Basis of Parental Behavior in the Rat ROBERT S. BRIDGES

Cultural Transmission in the Black Rat: Pine Cone Feeding JOSEPH TERKEL

Somatosensation and Maternal Care in Norway Rats JUDITH M. STERN

552

CONTENTS OF PREVIOUS VOLUMES

Experiential Factors in Postpartum Regulation of Maternal Care ALISON S. FLEMING. HYWEL D. MORGAN, A N D CAROLYN WALSH Maternal Behavior in Rabbits: A Historical and Multidisciplinary Perspective GABRIELA GONZALEZ-MARISCAL A N D J A Y S . ROSENBLATT Parental Behavior in Voles ZUOXIN WANG A N D THOMAS R. INSEL Physiological, Sensory, and Experiential Factors of Parental Care in Sheep F. LEVY, K. M. KENDRICK, E. B. KEVERNE, R. H. PORTER, A N D A. ROMEYER Socialization, Hormones, and the Regulation of Maternal Behavior in Nonhuman Simian Primates CHRISTOPHER R. PRYCE Field Studies of Parental Care in Birds: New Data Focus Questions on Variation among Females PATRICIA A D A I R GOWATY

Volume 26 Sexual Selection in Seawood Flies THOMAS H. D A Y A N D A N D R E S. G ILB U RN Vocal Learning in Mammals VINCENT M. JANIK A N D PETER J. B. SLATER Behavioral Ecology and Conservation Biology of Primates and Other Animals KAREN B. STRIER How to Avoid Seven Deadly Sins in the Study of Behavior MANFRED MILINSKI Sexually Dimorphic Dispersal in Mammals: Patterns, Causes, and Consequences L A U R A SMALE. SCOTT NUNES, A N D K A Y E. HOLEKAMP Infantile Amnesia: Using Animal Models to Understand Forgetting H. M O O R E A R N O L D A N D NORMAN E. SPEAR

Parental Investment in Pinnipeds FRITZ TRILLMICH

Regulation of Age Polyethism in Bees and Wasps by Juvenile Hormone SUSAN E. F A H R B A C H

Individual Differences in Maternal Style: Causes and Consequences of Mothers and Offspring LYNN A. FAIRBANKS

Acoustic Signals and Speciation: The Roles of Natural and Sexual Selection in the Evolution of Cryptic Species G A R E T H JONES

Mother-Infant Communication in Primates D A R I O MAESTRIPIERI A N D JOSEP CALL

Understanding the Complex Song of the European Starling: An Integrated Ethiological Approach M A R C E L EENS

Infant Care in Cooperatively Breeding Species CHARLES T. SNOWDEN

ISBN 0-12-004527-3

Representation of Quantities by Apes S A R A H T. BOYSEN