the behavior of animals

MECHANISMS, FUNCTION, AND EVOLUTION

SECOND EDITION

EDITED BY JOHAN J. BOLHUIS, LUC-ALAIN GIRALDEAU

AND JERRY A. HOGAN
This edition first published 2022
©2022 John Wiley & Sons, Inc.
Edition History
John Wiley & Sons, Inc. (1e, 2004)
All rights reserved. No part of this publication may be reproduced, stored in a retrieval
system, or transmitted, in any form or by any means, electronic, mechanical, photocopying,
recording or otherwise, except as permitted by law. Advice on how to obtain permission to
reuse material from this title is available at http://www.wiley.com/go/permissions.
The right of Johan J. Bolhuis, Luc-Alain Giraldeau, and Jerry A. Hogan to be identified as
the author(s) of the editorial material in this work has been asserted in accordance with law.
Registered Office(s)
John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA
Editorial Office
111 River Street, Hoboken, NJ 07030, USA
For details of our global editorial offices, customer services, and more information about
Wiley products visit us at www.wiley.com.
Wiley also publishes its books in a variety of electronic formats and by print-on-demand.
Some content that appears in standard print versions of this book may not be available in
other formats.
Limit of Liability/Disclaimer of Warranty
The contents of this work are intended to further general scientific research, understanding,
and discussion only and are not intended and should not be relied upon as recommending
or promoting scientific method, diagnosis, or treatment by physicians for any particular
patient. In view of ongoing research, equipment modifications, changes in governmental
regulations, and the constant flow of information relating to the use of medicines,
equipment, and devices, the reader is urged to review and evaluate the information provided
in the package insert or instructions for each medicine, equipment, or device for, among
other things, any changes in the instructions or indication of usage and for added warnings
and precautions. While the publisher and authors have used their best efforts in preparing
this work, they make no representations or warranties with respect to the accuracy or
completeness of the contents of this work and specifically disclaim all warranties, including
without limitation any implied warranties of merchantability or fitness for a particular
purpose. No warranty may be created or extended by sales representatives, written sales
materials, or promotional statements for this work. The fact that an organization, website,
or product is referred to in this work as a citation and/or potential source of further
information does not mean that the publisher and authors endorse the information or
services the organization, website, or product may provide or recommendations it may
make. This work is sold with the understanding that the publisher is not engaged in
rendering professional services. The advice and strategies contained herein may not be
suitable for your situation. You should consult with a specialist where appropriate. Further,
readers should be aware that websites listed in this work may have changed or disappeared
between when this work was written and when it is read. Neither the publisher nor authors
shall be liable for any loss of profit or any other commercial damages, including but not
limited to special, incidental, consequential, or other damages.
A catalogue record for this book is available from the Library of Congress
Paperback ISBN: 9781119109501; ePub ISBN: 9781119109525; ePDF ISBN: 9781119109518.
Cover image: © Mike Truchon/Shutterstock
Cover design by Wiley
Set in 10/12.5pt Rotis Serif Std by Integra Software Services, Pondicherry, India
Contents
Cover
Title page
Copyright
List of Contributors
Foreword by Robert A. Hinde
Preface
1. The Study of Animal Behavior
2. Stimulus Perception
3. Motivation and Emotion
4. Biological Rhythms and Behavior
5. Brain and Behavior
6. Hormones and Behavior
7. Development of Behavior
8. Learning and Memory
9. Animal Cognition
10. Applied Animal Behavior and Animal Welfare
11. The Function of Behavior
12. Mate Choice, Mating Systems, and Sexual Selection
13. Animal Personality, the Study of Individual Behavioral
Differences
14. Animal Communication
15. Evolution of Behavior
16. The Evolution of Hominin Behavior
17. Evolutionary Approaches to Human Behavior
Name Index
Subject Index
End User License Agreement
List of Illustrations
Chapter 2
Figure 2.1 (A) The “reflection experiment.”...
Figure 2.2 Looking into a “black-box”:...
Figure 2.3 Configurational features in sign-stimuli....
Figure 2.4 The prey (a, c) vs. threat (b, d) configuration...
Figure 2.5 (a) Common toad displaying anti-predator
behavior...
Figure 2.6 Male silk moth in alerted position, with
combed...
Figure 2.7 Principles of scent detection in moths by...
Figure 2.8 Prey-feature analysis in common toads....
Figure 2.9 Neuroimaging toad’s visual system...
Figure 2.10 “Window hypothesis” of...
Figure 2.11 Visual processing streams in the...
Chapter 3
Figure 3.1 Conception of behavior systems. Stimuli from
the...
Figure 3.2 Lorenz’ model of motivation. The tap (T)...
Figure 3.3 Results of an experiment on guppy courtship.,...
Figure 3.4 Courtship and mating behavior of the...
Figure 3.5 Genetically featherless chicks...
Figure 3.6 Upright postures of the herring gull:...
Figure 3.7 “Waltzing” in a male...
Figure 3.8 Facial expressions of fear and aggression...
Chapter 4
Figure 4.1 Rest-activity records from...
 Figure 4.2 (A) Daily rhythm of core body...
Figure 4.3 Plasticity of rest–activity...
Figure 4.4 A phase-response curve (PRC)...
Figure 4.5 The rat brain and some areas...
Figure 4.6 The molecular basis of...
Chapter 5
Figure 5.1 The facial ruff of the barn owl...
Figure 5.2 Acoustic measurements taken from
microphones...
Figure 5.3 (a) Auditory neurons and neurons in the...
Figure 5.4 Electron micrograph of the head...
Figure 5.5 (a) The brain of the honeybee...
Figure 5.6 (a) Schematic sagittal view of the...
Figure 5.7 Juvenile zebra finch males were injected at...
Chapter 6
Figure 6.1 Simplified metabolic pathway...
Figure 6.2 Schematic view of the endocrine...
Figure 6.3 Schematic model illustrating how a...
Figure 6.4 Reciprocal interactions between...
Figure 6.5 Schematic representation of the...
Chapter 7
Figure 7.1 “Epigenetic landscape”...
Figure 7.2 Schematic illustration of two...
Figure 7.3 Zebra finch (left) and...
Figure 7.4 Stages in song development in...
Figure 7.5 Three ways in which experience,...
Figure 7.6 Mean preference scores...
 Figure 7.7 Schematic drawings of the...
Chapter 8
Figure 8.1 (a) Associative connections in...
Figure 8.2 (a) A typical color-based...
Figure 8.3 A serial position curve...
Figure 8.4 An experiment on chunking in...
Figure 8.5 A schematic of the experiment...
Figure 8.6 The effect of electroconvulsive...
Chapter 9
Figure 9.1 (a) Ammophiia sabulosa,...
Figure 9.2 Greylag goose retrieving an egg. (a)...
Figure 9.3 Supernormal stimulus. Herring gull...
Figure 9.4 (a) Computational framework for the...
Figure 9.5 The lifetime of an engram. The formation...
Figure 9.6 Observational learning in doves...
Figure 9.7 Bumblebee using a string to...
Figure 9.8 Young capuchin monkey using a stick to...
Figure 9.9 (a) Female Portia fimbriata...
Figure 9.10 Neural encoding of regret in...
Chapter 10
Figure 10.1 A dog in the Australian...
Figure 10.2 A moose (Alces alces)...
Figure 10.3 A serval (Felis serval)...
Figure 10.4 The price that caged American mink...
Figure 10.5 Calf demand for two types of...
Figure 10.6 Facial expressions of a newborn...
Figure 10.7 Cognitive bias in dairy calves...
Chapter 11
Figure 11.1 The power expressed as multiples...
Figure 11.2 The results of Krebs et...
Figure 11.3 Results of Giraldeau and Kramer...
Figure 11.4 Red knots (Calidris canutus)...
Figure 11.5 The payoff functions of the...
Figure 11.6 The value of two habitats...
Figure 11.7 The university ducks experiment...
Chapter 12
Figure 12.1 Mating system described as a...
Figure 12.2 Polygyny threshold model with...
Figure 12.3 Male mating success in...
Figure 12.4 Egg survival (%) in relation to...
Figure 12.5 Survival rate of male barn...
Figure 12.6 Sex ratio of zebra...
Chapter 13
Figure 13.1 Representation of personality...
Figure 13.2 Illustration of the distribution...
Figure 13.3 The broccoli model of...
Figure 13.4 A) Fitness increases with...
Figure 13.5 The relative importance...
Chapter 14
Figure 14.1 Non-informational and...
Figure 14.2 A taxonomy in which signals...
Figure 14.3 Representative flight paths...
Figure 14.4 A graphical version of the handicap
mechanism...
 Figure 14.5 A graphical version of a differential benefits
model...
Figure 14.6 Spectrograms of three...
Chapter 15
Figure 15.1 (left) Darwin predicted that...
Figure 15.2 In the hypothetical phylogeny...
Figure 15.3 Mean body hue of males and...
Figure 15.4 (top) The dose-dependent...
Figure 15.5 (left) (a) Several ecological (beak depth...
Figure 15.6 (A) The relationship between two...
Figure 15.7 (top) Weber’s Law predicts...
Chapter 16
Figure 16.1 Graphic representation of encephalization...
Figure 16.2 Reconstruction by the paleoartist Jay...
Figure 16.3 Side views of the toothrows of (from top to
bottom):...
Figure 16.4 Highly approximate tentative genealogical...
Figure 16.5 Diorama at the American Museum of...
Figure 16.6 The “Nariokotome Boy“...
Figure 16.7 Two Acheulean implements from the type...
Figure 16.8 Three-quarter view of the most...
Figure 16.9 Reconstruction of the largest of the...
Figure 16.10 Comparison between a reconstructed...
Figure 16.11 The best preserved of the small...
Figure 16.12 Monochrome rendering of a...
Chapter 17
Figure 17.1 Charles Darwin (from Wikimedia Commons)...
 Figure 17.2 Frontispiece of Thomas...
Figure 17.3 John B. Watson (from Wikimedia Commons)...
Figure 17.4 Field studies have shown that...
Figure 17.5 Sarah Hrdy (courtesy of Dan Hrdy)...
Figure 17.6 The human archaeological record...
Figure 17.7 A) In most mammals, the individuals...
Figure 17.8 In an experimental study, the...
Figure 17.9 Capuchin monkeys...

List of Tables
Chapter 2
Table 2.1 Invariance in Gestalt perception...
Table 2.2 From heterogeneous summation...
Table 2.3 Feature discrimination causes dishabituation.
Table 2.4 Command neuron hypothesis.
Table 2.5 Constants α and β...
Table 2.6 Minimum number of cell types...
Chapter 8
Table 8.1 Experimental designs in classical conditioning
Chapter 10
Table 10.1 Three Conceptions of Animal...
Chapter 11
Table 11.1 Payoffs to the row player...
contributors
Professor Gregory F. Ball
Department of Psychology
University of Maryland
2141 Tydings Hall
7343 Preinkert Drive
College Park, MD 20742
USA
Professor Jacques Balthazart
University of Liege
GIGA Neurosciences
Quartier Hôpital
15 Avenue Hippocrate
Tour Pharmacie (Bat. B36, 1er étage)
B-4000 Liège
Belgium
Professor Johan J. Bolhuis
Cognitive Neurobiology
Department of Psychology
Utrecht University
PO Box 80.086
Yalelaan 2
3584 CM Utrecht
The Netherlands
Professor Gillian R. Brown
School of Psychology & Neuroscience
University of St. Andrews
South Street
St. Andrews, Fife, KY16 9JP
United Kingdom
Dr Harald Burghagen†
Universität Kassel
Fachbereich Naturwissenschaften
Abteilung Zoologie/Physiologie, Neurobiologie
Heinrich-Plett-Str. 40
D-34132 Kassel
Germany
Dr Catharine P. Cross
School of Psychology & Neuroscience
University of St. Andrews
South Street
St. Andrews, Fife, KY16 9JP
United Kingdom
Professor Dr Jörg-Peter Ewert
Universität Kassel
Fachbereich Naturwissenschaften
Abteilung Zoologie/Physiologie, Neurobiologie
Heinrich-Plett-Str. 40
D-34132 Kassel
Germany
Professor David Fraser
Animal Welfare Program
Faculty of Land and Food Systems
University of British Columbia
2357 Main Mall
Vancouver, British Columbia
Canada V6T 1Z4
Professor Luc-Alain Giraldeau
Institut national de la recherche scientifique
490, rue de la Couronne
Québec QC
Canada G1K 9A9
Professor Geoffrey Hall
Department of Psychology
University of York
York YO10 5DD
United Kingdom
Professor Robert A. Hinde, FRS†
St. John’s College
Cambridge CB2 1TP
United Kingdom
Professor Jerry A. Hogan
Department of Psychology
University of Toronto
Toronto, Ontario
Canada M5S 3G3
Professor Kimberly Kirkpatrick
Director, Cognitive and Neurobiological Approaches to Plasticity
Center
Department of Psychological Sciences
Kansas State University
496 Bluemont Hall
Manhattan, KS 66506
USA
Professor Kevin N. Laland
Centre for Biological Diversity
School of Biology
University of St. Andrews
Sir Harold Mitchell Building
St. Andrews
Fife KY16 9TF
United Kingdom
Professor Ralph E. Mistlberger
Department of Psychology
Simon Fraser University
8888 University Drive
Burnaby, British Columbia
Canada V5A 1S6
Dr Anders Pape Møller
Directeur de Recherche
Ecologie Systématique Evolution
Université Paris-Saclay
CNRS, AgroParisTech
F-91405 Orsay Cedex
France
Professor Pierre-Olivier Montiglio
Groupe de Recherche en Ecologie Comportementale et Animale
Département des Sciences Biologiques
Université du Québec à Montréal
CP 8888, succursale centre-ville
Montréal, Québec,
Canada H3C 3P8
Professor Stephen Nowicki
Department of Biology
Duke University
130 Science Drive
Durham, NC 27708
USA
Professor Denis Réale
Groupe de Recherche en Ecologie Comportementale et Animale
Département des Sciences Biologiques
Université du Québec à Montréal
CP 8888, succursale centre-ville
Montréal, Québec,
Canada H3C 3P8
Professor Benjamin Rusak
Departments of Psychiatry and Psychology & Neuroscience
Dalhousie University
5909 Veterans Memorial Lane
Halifax, Nova Scotia
Canada B3H 2E2
Professor Michael J. Ryan
Department of Integrative Biology
University of Texas
Austin, TX 78712
USA
Professor William A. Searcy
Department of Biology
University of Miami
Coral Gables, FL 33124
USA
Professor David F. Sherry
Advanced Facility for Avian Research
Departments of Psychology and Biology
Western University
1393 Western Road
London, ON
Canada N6G 1G9
Professor Ian Tattersall
Division of Anthropology
American Museum of Natural History
New York, NY 10024
USA
Professor Daniel M. Weary
Animal Welfare Program
Faculty of Land and Food Systems
University of British Columbia
2357 Main Mall
Vancouver, British Columbia, Canada V6T 1Z4
foreword
ROBERT A. HINDE
Writing a foreword for such a stimulating series of chapters, which
represent the state of animal behavior studies at this time, is a
considerable responsibility. Perhaps I can do best by looking not
forward, as might seem appropriate, but backward, and thus attempt
to provide a context for the chapters that follow. Of course it cannot
be a fully objective backward view, because I am looking from where
I am now, and what I see is biased by my own experience. It is bound
also to involve simplification. But I hope that it will provide a useful
perspective.
In the early decades of the twentieth century, most studies of animal
behavior fell into two groups. In one were the naturalists, mostly
amateurs, without scientific pretensions but with a long tradition
stretching back beyond the nineteenth century. In the other were the
psychologists, producing an increasing body of data and theory
mostly concerned with learning processes. Of course this dichotomy
is already unjust and simplistic. Darwin himself could be called a
naturalist; and an originator of learning theory (J.B. Watson) started
from naturalistic observation. However, the work of the learning
theorists, impressive in its own right, was not to have much impact
on the traditions that led to the chapters in this book until much
later.
Those traditions can be said to stem from the emergence of ethology
in the 1930s. This was due to Lorenz, an Austrian MD with a PhD in
comparative anatomy, and Tinbergen, a Dutch zoologist who moved
to England a few years after the end of World War 2. Both men had a
passionate interest in animals, but this was expressed in very
different ways. Lorenz kept a menagerie of diverse animals in his
home, though also studying the local jackdaws and the semi-tame
geese that he reared. Tinbergen, by contrast, was a dedicated field
naturalist. Although he later worked with captive animals, it was
always with problems that he had brought in from the field, and he
liked best to be in the field himself. Tinbergen’s first pupil, Baerends,
suggested that the contrast lay in their attitudes to their subjects:
Tinbergen saw himself as a nonparticipant hidden observer of
animals, Lorenz as an adopted alien member and protector. Lorenz
was a thinker who tried to relate or contrast his observations with
current biological and philosophical views, while Tinbergen was
much more empirical, an experimenter as well as an observer.
But both rejected the vitalist view that the phenomena of “instinct”
were unanalysable and the misuse of the Gestalt concept to imply
that analysis is unnecessary because the whole is always more than
the sum of its parts. They also rejected the focus of most learning
theorists on the input/output relations of the whole organism, with
neglect of the “physiological machinery,” and the sterility of the
artificial environments used to study animals in many psychological
laboratories.
The term “ethology” has been applied primarily to the work of
students who, though differing widely in the problems they tackled,
the methods they used, the level of analysis at which they worked,
and the theoretical interpretations (if any) that they adopted, shared
certain orienting attitudes. They insisted that the proper description
of behavior is a necessary preliminary to its analysis; and that the
behavior of an animal must be studied in relation to the environment
to which it has become adapted in evolution. In addition they held
that full understanding of behavior required knowledge not only of
its development and causation but also of its biological function and
its evolution. The result was a vast amount of data on the behavior of
animals and a certain amount of model-building to elucidate the
mechanisms underlying behavior. In 1973 Lorenz and Tinbergen
(together with von Frisch) were awarded the Nobel Prize in
Physiology and Medicine.
Although ethology was primarily a European phenomenon in the
early postwar years, two research workers in the USA were to have a
considerable influence on its development, though in very different
ways. Beach, a behavioral endocrinologist interested primarily in the
hormonal control of sexual behavior, met Tinbergen in the USA and
became a powerful supporter of ethology. Schneirla, a comparative
psychologist working at the American Museum of Natural History,
was intensely critical. One of his students, Lehrman, published a
hard-hitting critique of ethology in 1953. There were three main
issues. Lehrman and Schneirla took exception to Lorenz’s distinction
between “innate” and “acquired” behavior as neither empirically
valid nor heuristically valuable. They objected to the energy model of
motivation that Lorenz used, though this also came in for criticism
from within ethology. And both were unhappy about the ethologists’
tendency to apply concepts across a wide range of species differing in
their levels of cognitive capacity. On their side, the ethologists felt
that the adjective in “comparative psychology” was a sham, for
contrasting distantly related species did not constitute comparison in
a biological sense. They were also unhappy about the manner in
which many comparative psychologists (though not so much those
influenced by Schneirla) generalized on the basis of studies of a few
mammalian species, predominantly the laboratory rat. For a while
the differences between the two groups seemed irreconcilable.
However, soon after his critique was published, Lehrman came to
Europe and met a number of European ethologists. Tinbergen,
Lorenz, and Lehrman were all bird watchers, with a passionate
enthusiasm for natural history. Lehrman had an infectious geniality,
and friendships were soon made. This brought about a
rapprochement between ethology and many of the members of
Schneirla’s group, a rapprochement that came not so much from
academic discussion, but from the compatibility of personalities. On
the issue of ontogeny, both sides changed their emphasis, the
comparative psychologists withdrawing from their extreme emphasis
on experience, and the differences in approach to the “comparative”
issue were recognized.
It seems to happen not infrequently in the history of science that
those regarded as originating a branch of science are subsequently
seen to have been wrong in many of their generalizations. For
instance, Freud (psychoanalysis) used a misleading model of
motivation, Piaget (developmental psychology) based generalizations
on a tiny sample of subjects, and Jeffreys (geophysics) refused to
accept the evidence for continental drift. This was also the case with
ethology. Many of the concepts that had been invaluable tools in the
early days of ethology—the “innate releasing mechanism” and “fixed
action pattern” for instance—were subsequently seen to involve
oversimplification, and now seldom figure in the literature. But not
surprisingly the change in outlook was not adopted simultaneously
by all ethologists, and this led to some divisions within ethology.
Lorenz, whose influence was particularly strong in Germany and the
USA (through two research workers who had worked with him, Hess
and Barlow), was slower to relinquish the innate/acquired
dichotomy and energy model of motivation than Tinbergen and
workers in the Netherlands and the UK.
An issue important for the nature/nurture debate became prominent
in the 1960s. Both Tinbergen and Lorenz had long argued on the
basis of empirical evidence that species were specially equipped for
particular learning tasks that were biologically important for them.
Thorpe’s book on birdsong, published in 1961, showed that the
chaffinch was predisposed to learn the species-characteristic song
pattern. A few years later, Rozin, Garcia, and others demonstrated a
predisposition to avoid toxins in mammals. Such findings were
directly contrary to the orientation of the learning theorists, who
were searching for laws of learning valid for all species and all
situations. It thus became apparent that, in many cases, what was
“innate” was a predisposition to learn some things in particular
contexts. This was to be of special importance for the study of human
behavior.
Lorenz, originally a comparative anatomist, had used species
differences and similarities as a taxonomic tool, and Tinbergen had
always had an interest in the function of behavior. But, of the four
problems of causation, development, function, and evolution, the
main (though by no means the only) emphasis in ethology had been
on the first two. In the 1970s this changed with the publication of
Wilson’s Sociobiology. The orienting attitudes of ethology continued
but the motivational models disappeared and many of the old
concepts fell into disuse. Behavioral ecology came to the fore, and
new theoretical approaches made possible the study of function in a
quantitative fashion. Data on foraging behavior were compared with
the behavior to be expected (on certain assumptions that were not
always made fully explicit) from an organism foraging with maximal
efficiency. Later, attention turned to such issues as sperm
competition and the role of fluctuating asymmetry. Hamilton’s work
on kin selection, first published in 1964 but neglected for much of the
next decade, made possible a new approach to social behavior. Game
theory was recruited, and mathematical modeling came to be a much
used tool in studies of behavior.
At the same time, the influence of ethology started to penetrate into a
number of other disciplines. Lehrman and Rosenblatt, as well as
Beach and his many students, adopted the orienting attitudes of
ethology in their work on behavioral endocrinology. Von Holst had
already studied the elicitation of fixed action patterns by brain
stimulation through implanted electrodes, and the importance of
using unconfined animals where possible was recognized by
neurophysiologists. Bowlby, a psychoanalyst concerned with the
effects of maternal deprivation in children, realized that what had
been called the “irrational fears of childhood” (fear of falling, being
alone, etc.) would have been highly adaptive in the environments in
which early hominids lived, and an ethological element was
incorporated in the “attachment theory” which he elaborated, an
approach that was to become central in studies of child development.
The study of human nonverbal communication profited from the
input of ethologists, such as Eibl Eibesfeldt. The techniques of the
behavioral ecologists were applied in studies of preindustrial human
groups. An ethological influence is to be seen in studies of human
personal relationships, and even in studies of religion and morality.
Thus, while ethology as a set of concepts or as a theory of animal
behavior has been largely superseded, the influence of its orienting
attitudes has increased and is potent in other disciplines.
While behavioral ecology took center stage in the study of animal
behavior, many felt it to be impoverished by the neglect of problems
of development and causation. This book will go a long way toward
setting the balance straight. Each of the four problems is covered,
and the chapters introduce the growing points in the study of animal
behavior at the start of the twenty-first century.
preface
The idea for this book arose out of a need that we (and many of our
colleagues) felt for a comprehensive textbook on animal behavior.
There is no shortage of animal behavior textbooks, so why did we
want to produce a new one? First, animal behavior is a dynamic field
of research, and we believe that a modern textbook should
incorporate all the contemporary subdisciplines of behavioral
biology, such as animal welfare, evolutionary psychology, animal
cognition, and behavioral neuroscience. In some ways, the science of
animal behavior has become a victim of its own success, as it covers a
much wider field than it did initially. Gone are the days when one
author could write a textbook both comprehensive and authoritative:
Robert Hinde’s classic Animal Behaviour: A Synthesis of Ethology
and Comparative Psychology (1970) is an outstanding example of
such a book, and it continues to inspire many of us. Given the
breadth of contemporary animal behavior research, we felt that it
was important to invite experts in the respective subdisciplines to
write a chapter about their specialist topic.
Second, a large proportion of extant textbooks are single-author
volumes that approach animal behavior from a particular
perspective, for example from an evolutionary point of view or with
the emphasis on mechanisms. We believe that a modern science of
animal behavior should encompass both functional and causal
approaches. For such a comprehensive approach, we found the
classic formulation of the aims and methods of ethology (the study of
animal behavior) by Niko Tinbergen, one of its founding fathers,
most useful. Tinbergen suggested that there are four basic questions
in animal behavior, namely about causation, development, survival
value (function), and evolution. We agree with Tinbergen that all
these four questions are equally important. Hence all of them are
represented in this book. Like Tinbergen, we also find it important to
distinguish among the four questions. In particular, it is important to
realize which of the four questions is addressed, and to use a
research approach appropriate for that question. Most chapters in
the book focus on one of the four questions, with cause and
development being the main subjects of the first ten chapters, and
survival value (function) and evolution being the main subjects of the
last seven chapters. But most chapters are also concerned with more
than one question, noted separately of course, supporting
Tinbergen’s claim that all questions should be answered.
We are very pleased with the enthusiastic response we received from
the authors invited to contribute to this book. They are all leaders in
their respective fields, and we feel privileged that they participated in
this project. Robert Hinde has passed away since his foreword was
written, but his words are just as relevant to the second edition of
this book as they were to the first. His influence remains.
1
the study of animal behavior
JOHAN J. BOLHUIS, LUC-ALAIN GIRALDEAU AND JERRY A.
HOGAN

INTRODUCTION
The scientific study of animal behavior is often called ethology, a
term used first by the nineteenth century French zoologist Isidore
Geoffroy Saint Hilaire, but then used with its modern meaning by
the American zoologist Wheeler (1902). Ethology is derived from
the Greek ethos, meaning “character.” The word “ethics” is also
derived from the same Greek word, which makes sense, because
ethics is basically about how humans ought to behave.
Unfortunately, the word “ethology” is also often confused with
the word “ethnology” (the study of human peoples), with which it
has nothing in common. In fact, the very word processor with
which we are writing this chapter keeps prompting us to replace
“ethology” by “ethnology”! For whatever the reason, the word
“ethology” is not used as much as it used to be, although there is
still an active animal behavior journal bearing this name. Instead
of “ethology,” many authors now use the words “animal behavior”
or “behavioral biology” when they refer to the scientific study of
animal behavior.

A Brief History of Behavioral Biology

Early days
Scientists (and amateurs) have studied animal behavior long before
the word “ethology” was introduced. For instance, Aristotle had
many interesting observations concerning animal behavior. The
study of animal behavior was taken up more systematically, however,
mainly by German and British zoologists around the turn of the
nineteenth century. The great British naturalist Charles Darwin
(1809–1882), in his classic book on the theory of evolution by
natural selection (1859), devoted a whole chapter to what he called
“instinct.” As early as 1873, a British amateur investigator, Douglas
Spalding, recorded some very interesting observations on the
instinctive behavior of young domestic chicks, including a
phenomenon that was later called “imprinting” (Chapter 7). At the
beginning of the twentieth century, the behavior of animals was also
studied in the context of learning by the Russian physiologist Ivan P.
Pavlov (1927) and the American psychologist Edward L. Thorndike
(1911; Chapter 8).

Lorenz and Tinbergen

In the middle of the twentieth century, the study of animal behavior
became an independent scientific discipline, called ethology, mainly
through the efforts of two biologists, the Austrian Konrad Lorenz
(1903–1989) and the Dutchman Niko Tinbergen (1907–1988). It can
be said that Lorenz was the more philosophical and theoretical of the
two. He put forward a number of theoretical models on different
aspects of animal behavior such as evolution and motivation. He was
also the more outspoken of the two men, and some of his
publications met with considerable controversy. Tinbergen was very
much an experimentalist, who together with his students and
collaborators conducted an extensive series of field and laboratory
experiments on the behavior of animals of many different species. In
1973, Lorenz and Tinbergen were awarded the Nobel Prize for
Physiology and Medicine. They shared their prize with Karl von
Frisch (1886–1982), an Austrian comparative physiologist and
ethologist who had pioneered research into the dance “language” of
bees (Chapter 14).

Ethology and comparative psychology

During the early days of ethology there was a certain amount of
scientific rivalry between mainly European ethologists and North
American experimental psychologists, who also studied animal
behavior in what was usually called comparative psychology. The
European ethologists emphasized that animal behavior is a biological
phenomenon, and as such a product of evolution. This is exemplified
by the use of the word “instinct” (e.g., in the title of Tinbergen’s 1951
classic book The Study of Instinct), which referred to the “innate”
components of behavior that are subject to natural selection. A
prominent critique of this way of thinking came from the American
psychologist Daniel Lehrman, in his 1953 paper “A critique of
Konrad Lorenz’ theory of instinctive behavior.” In this paper he
argued against Lorenz’ theory in which behavior can be dissected
into “innate” and “acquired” (learned) components (see Chapter 7 for
a more detailed discussion of these issues). In general, American
psychologists emphasized the effects of learning on behavior. Pavlov
had already demonstrated the importance of what we now call
Pavlovian (or classical) conditioning, and Thorndike studied learning
processes that are now known as instrumental conditioning (Chapter
8). Another difference between the ethologists and experimental
psychologists was that the former often observed animals of many
different species in their natural environment, while the latter,
despite the name comparative psychology, often concentrated on one
species, such as the rat or the pigeon, that was studied exclusively in
the laboratory.

Behaviorism
The emphasis of the North American psychologists on learning was
epitomized by the rise of behaviorism in the 1930s. Behaviorism was
a very influential school of thought initiated by the American
psychologist John B. Watson (1878–1958), with his book
Behaviorism (1924). Essentially, Watson considered psychological
phenomena to be physical activity rather than some kind of mental
event. He proposed that we cannot make any scientific statements
about what might be going on in our minds, and that introspection
was unreliable. Rather, for behaviorists, psychology is the study of
observable behavior and of the external physical factors that
influence it. At the time, behaviorism was extremely influential in
science and beyond. Within North American psychology it was the
dominant school of thought for several decades. Behaviorist theory
also affected education practice, particularly with Watson’s book
Psychological Care of Infant and Child (1928). Watson once made
the famous statement:
Give me a dozen healthy infants, well-formed, and my own specified
world to bring them up in and I’ll guarantee to take any one at
random and train him to become any type of specialist I might select
—doctor, lawyer, artist, merchant-chief, and, yes, even beggarman
and thief, regardless of his talents, penchants, tendencies, abilities,
vocations, and race of his ancestors.
This epitomizes behaviorist ideas about child rearing. Watson
considered the upbringing of children to be an objective, almost
scientific exercise, without the need for affection or sentimentality.
Watson’s most famous student was Burrhus Frederic Skinner (1904–
1990), who applied behaviorist ideas to the study of learning. For
Skinner (1938) and his behaviorist colleagues, learning had to do
with changing relationships between visible entities, not with what
might be going on inside the animal’s head. In particular, behaviorist
learning theorists suggest that learning refers to changes in the
frequency of responding due to its consequences. Most of their
experiments involve operant conditioning (see Chapter 8), in which a
certain response by the animal (e.g., pressing a lever) is rewarded
(“reinforced”) with food.

Cognitive psychology
Within experimental psychology there came a reaction to
behaviorism in what we now call cognitive psychology. In contrast to
behaviorism, cognitive psychologists start with the assumption that
individuals (humans and other animals) have a mental life that can
be investigated (Chapter 9). For instance, Skinner (1957) maintained
that language development in children was a learning process, in
which responses (i.e., uttering certain sounds) were reinforced. The
American linguist Noam Chomsky (1959) wrote a highly critical
review of Skinner’s book on language development, suggesting that
language acquisition is not a case of instrumental conditioning, but a
much more complex interaction between experience and internal
cognitive mechanisms (Chapter 7). At birth the child is already
endowed with essential knowledge of language, the theory of which
is known as Universal Grammar. Clearly, learning is involved in the
development of language, but it is not the only factor. Chomsky and
colleagues have recently described the growth of language in the
child as “the interplay of three factors: domain-specific principles of
language (Universal Grammar), external experience, and properties
of non-linguistic domains of cognition including general learning
mechanisms and principles of efficient computation” (Yang et al.
2017). Another important publication that signaled the beginning of
the cognitive revolution is a book by the British psychologist Donald
Broadbent (1958) who, in contrast to Skinner, analyzed learning and
memory in terms of cognitive mechanisms rather than stimulus–
response relations. Hogan (1988, 2017) has noted that what cognitive
psychologists call “cognitive structures” are in fact the same as the
causal mechanisms that were proposed by ethologists such as Lorenz
and Tinbergen (Chapters 3 and 9).

Four Questions in the Study of Animal

Behavior

Niko Tinbergen published a very important paper in 1963, in which

he outlined four major questions in the study of animal behavior,
namely causation, development, survival value, and
evolution. As he readily admitted, Tinbergen wasn’t original,
because three of these questions (causation, survival value, and
evolution) had already been put forward by the British biologist
Julian Huxley (1887–1975) as the major questions in biology, but
Tinbergen added a fourth question, development. Many authors,
including ourselves, use the word function as a substitute word for
survival value, but the term function is used in many different ways
in biology (Wouters 2003), and care is necessary when using it. It
should also be mentioned that the evolutionary biologist Ernst Mayr
(1904–2005) in 1961 popularized a different categorization of
problems in biology: proximate and ultimate causation. Proximate
causation is similar to Tinbergen’s causal question, but ultimate
causation is a controversial term that deals with evolutionary issues
somewhat differently from Tinbergen’s questions of survival value
and evolution. However, no matter how these questions are broken
up, it is crucially important that students of animal behavior be quite
clear about the type of question they are addressing when they study
or speak of animal behavior.
We find Tinbergen’s analysis so important that we would say you
cannot really understand animal behavior if you do not also
understand the meaning of his four questions. Some of the more
heated contemporary debates in the field of animal behavior can
often be traced to misunderstandings about the meaning of these
questions (e.g., Hogan 1994, 2017; Bolhuis & Macphail 2002). It is
essential, therefore, that any productive discussion about animal
behavior involves participants that are capable of clearly stating
which of the four questions they are addressing. This view of animal
behavior has also served as the framework for the organization of the
present book, with the first half covering mostly causal and
developmental topics while the second half deals with questions of
survival value (or function) and evolution.
Tinbergen’s four questions are sometimes also called the four whys,
because they represent four ways of asking “why does this animal
behave in this way?” Let’s consider a bird singing at dawn, say a male
song sparrow (Melospiza melodia). The question is: why is this bird
singing? This seems a perfectly straightforward question, but in fact
it is not, because answers can take any of four different forms. These
different forms—you’ve guessed it—have to do with Tinbergen’s four
questions. The first question concerns causation: what causes the
bird to sing? The answers include the stimuli or triggers of behavior
whether they be internal or external, the way in which behavioral
output is guided, factors that stop behavior, and the like. These are
questions concerning the causation of behavior. Sometimes this is
called motivation, a topic discussed at length in Chapter 3.
Tinbergen’s question of causation also concerns the mechanisms or
structure of behavior. These mechanisms involve the “machinery”
that operates within the animal and which are responsible for the
production of behavioral output (Chapters 5 and 9).
The second question is about development: How did the singing
behavior of the bird come about in the lifetime of an animal? A male
song sparrow does not sing immediately after it has hatched from the
egg, and it takes quite some time before it has developed a song, a
process that involves learning. Such questions that concern
development of behavior, sometimes also called ontogeny, will be
discussed explicitly in Chapter 7. The third question has to do with
the consequences of singing for the singer’s fitness: What is the
function of the bird singing; what is it singing for? Does singing help
the bird keep intruding males away from his nest? Or does it serve to
attract females? Or does it do both? The topic of function (survival
value), its methods of enquiry, and main findings will be discussed
explicitly in Chapter 11. The fourth question concerns evolution: how
did this behavior come about in the course of evolution? Behavior
does not leave fossils behind and so the study of its evolutionary
history requires the development of special methods. These methods,
based on taxonomy and comparisons among species, will be
discussed in detail in Chapter 15.
The previous paragraph illustrates that the question “why does this
bird sing?” is not very useful, as it can have four different meanings.
It can be very confusing if a biologist studying birdsong does not
make it clear which of the four “why questions” he/she is asking, and
it could lead to futile arguments concerning whether the bird is
singing to attract mates or because it learned its song. The same
problem arises in all other areas of animal behavior, so it is very
important to make clear which question is being addressed in any
study. Of course, it is possible that a particular investigator wants to
address more than one question at a time. This is perfectly
legitimate, as long as it is made explicit which of the questions are
addressed at what time. A famous example of this is an experimental
paper by Tinbergen and his associates (Tinbergen et al. 1962) on the
behavior of blackheaded gulls (Larus ridibundus). After the chicks
have hatched, the adult birds remove the empty eggshells from the
nest. Tinbergen and his students investigated both the causation and
the function (survival value) of this behavior using elegantly
designed simple field experiments. They discovered the stimulus
characteristics of items removed from the nest and, in the same
paper, also reported results relevant to nest predation.
There is also considerable overlap among the four questions. For
instance, the development of behavior is essentially a causal problem
but may also involve functional aspects (Chapter 7). The evolution of
behavior often depends on mechanism. For instance, emergent
properties of an animal’s sensory and perceptual capabilities
(mechanisms) may create opportunities for sexual selection to
operate in the evolution of extravagant traits (Chapters 12 and 14).
Finally, questions in one domain (e.g., function) can provide clues
for questions in another domain (e.g., causation). For instance, a
number of bird species cache food, some for a few hours, others for
months (Vander Wall 1990). It is plausible that the ecological
circumstances that have given rise to these different forms of food
caching may have also influenced the birds’ ability to memorize
spatial locations. In fact, a large number of studies are concerned
with investigating the spatial memory of food caching versus
nonfood-caching birds (Chapter 8).

Trends in the Study of Animal Behavior

Behavioral ecology: from mechanism to function

Much of the research and theorizing of early ethologists such as
Lorenz and Tinbergen was concerned with the causation of
behavior. When Tinbergen was invited to move from the Dutch
University of Leiden to the University of Oxford, he established the
Animal Behaviour Research Group, while at the same time the
ecological ornithologist David Lack was taking over the newly
founded Edward Grey Institute of Ornithology. The coincidence of
having both these scientists and their followers in the same
department in Oxford sowed the seeds of a discipline that was to
blossom much later in the mid-1970s under the name of behavioral
ecology.
Behavioral ecology arose out of the fusion of evolutionary ecology,
population ecology, and ethology. A number of conditions were ripe
in the mid-seventies for such an event. In 1975 the Harvard
entomologist Edward O. Wilson published Sociobiology, The New
Synthesis (1975). Wilson’s book was firmly grounded in population
genetics and evolutionary biology. Its clear presentation of William
D. Hamilton’s concepts of inclusive fitness, kin selection, the
evolution of altruism and social groups among others, provided the
essential foundations for a successful evolutionary approach to social
behavior. Not long after that, in 1978, John R. Krebs at Oxford
University and Nicholas B. Davies at Cambridge co-edited a book
they called Behavioural Ecology: An Evolutionary Approach, which
applies a similar evolutionary approach but this time to all, not just
social, behavior. The publication of that book marks the official birth
of behavioral ecology, which now includes sociobiology (see Chapter
17).
Behavioral ecology today is more of an approach than a body of
accumulated fact. Its initial success grew out of a combination of
optimality theory and evolutionary thinking that pictures the
expression of behavioral traits as constrained trade-offs between
their evolutionary benefits and costs (Chapter 11). The development
of the concept of the evolutionarily stable strategy (ESS) by the
British evolutionary biologist John Maynard Smith (1982) allowed
this cost–benefit approach to be applied to a wide range of
behavioral interactions. Evolutionary thinking and the cost–benefit
approach cast a new light on behavioral systems such as foraging,
fighting, and habitat selection (Chapter 11). When applied to
communication it raised an important number of questions
concerning the design of signals and their functions (Chapter 14).
While early ethologists tended to picture sexual reproduction as a
cooperative venture between males and females, the evolutionary
approach has somewhat subverted this idyllic view. Mating systems
and mate choice (Chapter 12) as well as conflicts of interests between
mates (Chapter 14) have become exciting and rapidly developing
areas of the discipline. Darwin himself pictured behavior as a
character that was modified over generations by selection. Behavior,
hence, has a history that can be and is studied with contemporary
organisms (Chapters 15 and 16).

Neuroethology and cognitive neuroscience

The mechanisms underlying behavior are also represented in the
workings of the central nervous system. In fact, Tinbergen often used
neural analogies and metaphors in his models of behavior. We shall
see in Chapters 2 and 5 how the central nervous system obtains and
processes information about its external world. As knowledge about
the brain, both its gross and fine-level morphology as well as the way
its neurons are connected, led to increased interface between brain
and behavior a new subdiscipline arose that is called neuroethology
(Ewert 1980; Chapter 2). In the early days of this new discipline,
researchers concentrated on the study of the neural mechanisms of
perception and movement, often in insects or simple vertebrates.
More recently the study of the brain mechanisms of behavior is also
directed at higher cognitive processes such as learning and memory
or spatial orientation. Often, the terms behavioral neuroscience or
cognitive neuroscience are used to describe these disciplines. Now,
the combination of an extraordinary array of powerful techniques
from electrophysiological recording to molecular analyses of RNA
sequences allows researchers to delve deeper into the connection
between behavior and its neural substrate (Chapters 5 and 9).

Cognitive ecology and neuroecology

Perhaps as a result of the success of behavioral ecology, the
mechanisms of brain and cognition have also been studied more
recently from a functional and/or an evolutionary perspective, in
new fields known as cognitive ecology (Healy & Braithwaite 2000;
Macphail & Bolhuis 2001) and neuroecology (Bolhuis & Macphail
2001). According to the former, an animal’s ability to collect and
process information should be heavily influenced by its ecology.
Neuroecology refers to the study of the neural mechanisms of
behavior guided by functional and evolutionary principles. How do
the evolutionary pressures for complex birdsong affect the evolution
of the underlying neural substrate? How does having a large home
range affect one’s ability to navigate? Does having to store food place
selective pressure on spatial memory and its underlying brain
regions? These functional and evolutionary approaches to the study
of brain and cognition have come under considerable criticism from
authors who claim that they are flawed, because Tinbergen’s four
whys are being confounded (Bolhuis & Macphail 2001; Macphail &
Bolhuis 2001; Bolhuis & Wynne 2009; Bolhuis 2015). For example,
food storing in certain species of titmice and corvids has been
interpreted as a case of adaptive specialization (Healy & Braithwaite
2000; Shettleworth 2010). According to this hypothesis, food-storing
birds would have evolved superior spatial memory as well as a larger
hippocampus (thought to be its neural substrate) compared to their
nonstoring counterparts. Both parts of this hypothesis have not been
confirmed by the data (see Chapter 8 for further discussion).

Animal welfare and human nature

Many people are interested in animal behavior out of mere curiosity,
the need to know more about something. This is all very fine but
there always comes a time when someone will ask “what is the
purpose of studying animal behavior?” This question, whether from
a research colleague, a friend, or a granting agency requires an
answer expressed in terms of benefits to society. We see two areas in
which animal behavior research can contribute to human society:
animal welfare and understanding human nature.
Animals are important contributors to wealth and quality of life.
They provide us with nourishment, the means to find cures and
treatments for our illnesses, as well as invaluable companionship.
Almost all of the information contained in this book relies on
experiments and research conducted with animals. There is a
growing concern that animals used for human benefit, however, be
exposed to as little unpleasantness as necessary. Are housing cages
too small, or densities of individuals too high? Is the knowledge
acquired from experiments sufficiently important to authorize
animal experimentation? The answer to such, often difficult,
questions depends in many ways on knowing something about an
animal’s behavior (Chapter 10).
People are endlessly curious about people and the sheer number of
disciplines devoted uniquely to the study of human beings is
eloquent testimony to this fact (e.g., medicine, anthropology,
psychology, sociology, criminology). Animal behavior can provide
insight into human behavior in two ways. More conventionally,
phenomena observed in animals can be generalized, although often
in some modified way, to humans. For example, just as a new
antibiotic drug that cures an infection in some nonhuman primate
can also be used, perhaps in a slightly modified way, to cure
infections in humans, so can knowledge about how an animal learns
be extended and applied to human learning. The second way,
however, involves generalizing an approach rather than a result. For
instance, can we learn anything new about human behavior by
applying an evolutionary cost–benefit analysis to the things we do?
This is what an area known as evolutionary psychology does
(Chapter 17).

SUMMARY AND CONCLUSIONS

The study of the behavior of animals has grown into a highly
diverse set of approaches and disciplines. Its subject area ranges
from molecules and neurons to individuals and populations. One
of Tinbergen’s major contributions to the study of animal
behavior has been to make its goals explicit and clarify four types
of questions that can be asked of behavior: causation,
development, survival value, and evolution. In this book we
strongly advocate Tinbergen’s position that behavior can only be
understood through research on all four questions. In addition,
we suggest that it should be made clear which of Tinbergen’s
questions is addressed when a behavioral problem is
investigated: a problem in one domain should not be investigated
with concepts from another. The early chapters in this book
examine primarily causal, mechanistic, and developmental
questions, while the latter chapters examine survival value and
evolution issues. But examples in many chapters also illustrate
that multiple approaches are necessary for understanding a
problem, affirming Tinbergen’s view.

FURTHER READING
Tinbergen’s (1963) paper on the four whys is essential reading for
any serious student of animal behavior. It was reprinted in Houck
and Drickamer (1996), which is a collection of classic papers on all
aspects of animal behavior, and in Tinbergen’s Legacy (Bolhuis &
Verhulst 2009), which is a collection of contemporary essays
reflecting on Tinbergen’s classic paper. The four-part reader by
Bolhuis and Giraldeau (2010) is organized on the basis of
Tinbergen’s four whys and has a collection of classic and
contemporary papers on the evolution, function, development, and
causation of animal behavior. Tinbergen’s (1951) classic book is still
very much worthwhile. It was reprinted in 1992 and is still available.
The British ethologist William Thorpe (1979) has written a brief
history of ethology, viewed from the inside, while Dewsbury (1989)
provides a more recent account from the North American
perspective. Boakes (1984) is an excellent review of the history of the
study of animal behavior by psychologists, while Laland and Brown
(2011) provide a very clear account of the different ways in which the
behavior of animals (including humans) can be studied from an
evolutionary perspective. Functional and evolutionary aspects of
behavior are also discussed in Davies et al. (2012). Hogan’s (2017)
book provides an exposition and critique of behavioral concepts and
of historical and contemporary studies of behavior.

REFERENCES
Boakes, R.A. 1984. From Darwin to behaviorism. Psychology and
the minds of animals. Cambridge, UK: Cambridge University
Press.
Bolhuis, J.J. 2015. Evolution cannot explain how minds work.
Behavioural Processes, 117, 82–91.
Bolhuis, J.J. & Giraldeau, L.-A. (eds.). 2010. Animal behaviour, 4
volumes. London: Sage Publications.
Bolhuis, J.J. & Macphail, E. M. 2001. A critique of the neuroecology
of learning and memory. Trends in Cognitive Sciences, 5, 426–
433.
Bolhuis, J.J. & Macphail, E.M. 2002. Everything in neuroecology
makes sense in the light of evolution. Trends in Cognitive
Sciences, 6, 7–8.
Bolhuis, J.J. & Verhulst, S. (Eds.) 2009. Tinbergen’s legacy:
Function and mechanism in behavioral biology. Cambridge:
Cambridge University Press.
Bolhuis, J.J. & Wynne, C.D.L. 2009. Can evolution explain how
minds work? Nature, 458, 832–833.
Broadbent, D.E. 1958. Perception and communication. London:
Pergamon.
Chomsky, N. 1959. A review of B.F. Skinner’s ‘Verbal Behavior’.
Language, 35, 26–58.
Darwin, C. 1859. On the Origin of Species by Means of Natural
Selection, or the Preservation of Favoured Races in the Struggle
for Life. London: John Murray.
Davies, N.B., Krebs, J.R. & West, S.A. 2012. An introduction to
behavioural ecology. Chichester: Wiley-Blackwell.
Dewsbury, D. 1989. A brief history of the study of animal behavior in
North America. In: P.P.G. Bateson & P.H. Klopfer (eds.),
Perspectives in ethology, pp. 85–122. London: Plenum.
Ewert, J.P. 1980. Neuroethology. An introduction to the
neurophysiological fundamentals of behavior. Berlin: Springer.
Healy, S. & Braithwaite, V. 2000. Cognitive ecology: a field of
substance? Trends in Ecology and Evolution, 15, 22–26.
Hogan, J.A. 1988. Cause and function in the development of
behavior systems. In E. M. Blass (ed.), Handbook of behavioral
neurobiology, vol. 9, pp. 63–106. New York: Plenum.
Hogan, J.A. 1994. The concept of cause in the study of behavior. In J.
A. Hogan & J. J. Bolhuis (eds.), Causal mechanisms of
behavioural development, pp. 3–15. Cambridge, UK: Cambridge
University Press.
Hogan, J.A. 2017. The study of behavior: Organization, methods,
and principles. Cambridge, UK: Cambridge University Press.
Houck, L.D. & Drickamer, L.C. (eds.). 1996. Foundations of animal
behavior. Classic papers with commentaries. Chicago: University
of Chicago Press.
Krebs, J.R. & Davies, N.B. (eds.). 1978. Behavioural ecology: An
evolutionary approach. Sunderland, MA: Sinauer.
Laland, K.N. & Brown, G.R. 2011. Sense and nonsense. Evolutionary
perspectives on human behaviour, 2nd ed. Oxford: Oxford
University Press.
Lehrman, D.S. 1953. A critique of Konrad Lorenz’ theory of
instinctive behavior. Quarterly Review of Biology, 28, 337–363.
Macphail, E.M. & Bolhuis, J.J. 2001. The evolution of intelligence:
adaptive specialisations versus general process. Biological
Reviews, 76, 341–364.
Maynard Smith, J. 1982. Evolution and the Theory of Games.
Cambridge: Cambridge University Press.
Mayr, E. 1961. Cause and effect in biology. Science, 134, 1501–1506.
Pavlov, I.P. 1927. Conditioned reflexes. Oxford: Oxford University
Press.
Shettleworth, A.J. 2010. Cognition, evolution, and behavior. New
York: Oxford University Press.
Skinner, B.F. 1938. The behavior of organisms. New York: Appleton-
Century-Crofts.
Skinner, B.F. 1957. Verbal behavior. Englewood Cliffs, N.J.: Prentice
Hall.
Spalding, D.A. 1873. Instinct, with original observations on young
animals. Macmillan’s Magazine, 27, 282–293. Reprinted in 1954
in: British Journal of Animal Behaviour, 2, 2–11.
Thorndike, E.L. 1911. Animal Intelligence. New York: Macmillan.
Thorpe, W.H. 1979. The origins and rise of ethology. London:
Heinemann/Praeger.
Tinbergen, N. 1951. The study of instinct. Oxford: Oxford University
Press.
Tinbergen, N. 1963. On aims and methods in ethology. Zietschrift für
Tierpsychologie, 20, 410–433.
Tinbergen, N., Broekhuysen, G.J., Feekes, F. et al. 1962. Egg shell
removal by the black headed gull, Larus ridibundus: a behaviour
component of camouflage. Behaviour, 19, 74–117.
Vander Wall, S.B. 1990. Food hoarding in animals. Chicago:
University of Chicago Press.
Watson, J.B. 1924. Behaviorism. New York: W.W. Norton.
Wheeler, W.M. 1902. ‘Natural history’, ‘ecology’ or ‘ethology’?
Science, 15, 971–976.
Wilson, E.O. 1975. Sociobiology, the new synthesis. Cambridge, MA:
Belknap/Harvard University Press.
Wouters, A.G. 2003. Four notions of biological function. Studies in
History and Philosophy of Biological and Biomedical Sciences,
34, 633–668.
Yang, C., Crain, S., Berwick, R.C., Chomsky, N. & Bolhuis, J.J. 2017.
The growth of language: Universal Grammar, experience, and
principles of computation. Neuroscience & Biobehavioral
Reviews, 81, 103–119.
2
stimulus perception
H. BURGHAGEN AND J.-P. EWERT
INTRODUCTION
Driving a car during rush-hour, we are exposed to a flood of
information that bombards our sensory systems through various
channels: visual, auditory, vibratory, somatosensory, etc. If the
central nervous system (CNS) were to respond to all this
information simultaneously, chaos would develop. Thus, on the
one hand, the CNS must be ready to collect information from
different sensory channels and to process these in parallel and
concurrently; on the other hand, it must be selective: perceiving
the right thing in the right place at the right time—say, a traffic
sign—and responding to it appropriately, for example, by
stepping on the brakes. This involves localization, identification,
and decision-making. In general, all animals employ their
sensory instruments for the translation of perception into action
in order to select a specific goal-oriented skill.
In this chapter, we start with a survey of sensory modalities and
show that sense organs and corresponding neural networks
(sensory maps) provide animals with their own sensory worlds.
In the light of current investigations in different animal species,
including humans, we select examples showing that Niko
Tinbergen’s ideas and concepts have paved the way for
ethological and neuroethological studies over the last six decades.
We go on to describe quantitative relationships between stimulus
and behavioral response, including discussion of the concepts of
sign-stimulus, innate releasing mechanism (IRM), heterogeneous
summation, supernormal releaser, and the influences of attention
and motivation. Our intention is to show that various classical
ethological concepts can be redefined, filled with physiological
content, and thus integrated into our current knowledge.
Moving from the behavioral to the neurophysiological level of
analysis, we explore stimulus perception and the behavior that
ensues, from which some general principles across species
emerge. In the CNS, there are stimulus-response mediating
pathways and neural loops that modulate, modify, or even specify
that mediation. Using neuronal correlates of releasing
 mechanisms as well as neural network modeling, which operate
as sensori-motor interfaces, we discuss sensory structures
involved in feature detection including olfaction in insects,
configurational visual object perception in toads and monkeys,
and visual perception in primates.

Stimulus Reception

Sensory information to be processed comes from outside an

organism’s CNS and must get first in contact with the nervous
system via its receptors. This information concerns basic sensory
modalities:

Photoreception: response to radiant energy in the visible

wavelength range of the electromagnetic spectrum (photons).
Thermoreception: response to radiant thermal energy in the
nonvisible wavelength range of the electromagnetic spectrum.
Mechanoreception: response to kinetic energy, including
hearing, vibration, touch, balance, etc.
Chemoreception: response to chemical energy, including smell
and taste.

Particular perceptual capabilities include electroreception (response

to electrical energy) and magnetoreception (response to energy of a
magnetic field). Nociception, the reception of pain, involves specific
cell physiological responses to severe tissue damage caused by
thermal, kinetic, and/or chemical energy. The form of energy to
which the receptor cell responds determines the sensory modality.
Within a sensory modality (e.g., vision), different stimulus qualities
(color) and stimulus quantities (brightness) can be distinguished.
A stimulus, sensed by a receptor cell, is transduced by intracellular
chemical processes (see Chapter 5). These lead to a change in the
(receptor) membrane potential, which—depending on cell type—can
generate nerve impulses. To respond to weak stimuli, receptors may
have their own amplifying system. In rod photoreceptor cells of the
vertebrate retina, for example, one photon absorbed by one molecule
of rhodopsin gives rise to a signaling cascade of intracellular
biochemical events that activate 6×106 molecules of cyclic guanosine
monophosphate, cGMP. This intracellular messenger influences the
ion channels of the cell membrane leading to the receptor potential.
Scent receptors in mammals have similar properties.

Receptor cells provide organisms with information on

their own sensory worlds
Each organism is equipped with sets of sensory receptors that open
the gates toward the world in which the organism lives. Jacob von
Uexküll (1921) pointed out that each animal species lives in, and
communicates with, its own sensory world, the “Umwelt.” Different
species perceive their Umwelt differently, and quite differently from
the way we humans perceive our environment. Knowledge about the
capabilities of sense organs indicates the kinds of stimuli perceived
by organisms and suggests what their perceptual worlds look like
(Prete 2004).
Certain species have evolved perceptual talents. In birds, visual
acuity and accommodation speed of the eyes are maximized at 20/4
vision, compared to 20/20 in humans. From high altitudes raptors
can magnify and detect small prey on the ground while monitoring a
wide field of vision. Many insects—like some birds—are sensitive to
ultraviolet light and may use polarized skylight for navigation
(Labhart & Meyer 2002). Barn owls, Tyto alba, hunting in darkness,
localize sound sources by tiny interaural time differences (Konishi
2003) (Chapter 5). Dogs and mice live mainly in a world of smell,
spiders in a world of vibration. Rattlesnakes, Crotalus viridis, have
infrared-sensitive pit organs for object detection in the dark
(Newman & Hartline 1982). Dolphins and nocturnal bats use
biosonar (Suga 1990), and weakly electric fish produce electric fields
for object detection and communication (Heiligenberg & Rose 1985).
Sharks utilize electroreceptors to detect electric fields of about 0.005
μV/cm, e.g., from breathing-muscle potentials by a flatfish hidden in
the sand. Like migratory birds, loggerhead turtles, Caretta, detect
geomagnetic cues that vary across the earth’s surface and employ
magnetic-compass orientation for long-distance navigation
(Wiltschko & Wiltschko 2002).
Male silk moths, Bombyx mori, are “smell champions.” One
molecule of the female’s sex attractant pheromone bombykol is
sufficient to elicit a nerve impulse in a receptor cell specialized for
bombykol (Kaissling 2014). The female’s pheromone glands contain
less than 1 µg bombykol, which, theoretically, is sufficient to lure 1013
Bombyx males. In males the behavioral olfactory threshold is 103
bombykol molecules/cm3 in an air stream at a velocity of about 50
cm/s. Actually, 200 simultaneous molecule hits at 200 of the more
than 25,000 bombykol receptors tell the Bombyx male: a female is
present in the direction from which the wind blows. For comparison,
in German shepherd dogs the behavioral threshold for butyric acid, a
component of sweat, is 5.9×103 molecules/cm3.
Whichever type of energy the receptor cells respond to—odor,
infrared, sonar, etc.—the corresponding sensory systems have a
comparable basic structure: receptor cells and nerve cells connected
to them. Perception involves the neurosensory as well as the
motivation and attention related processes, by which an organism
becomes aware of and localizes and recognizes external stimuli.
A few comments on stimulus recognition are appropriate.
Recognition (identification, interpretation, detection) in the present
context means that stimuli from the environment are classified into
meaningful units, i.e., categories of functional (behavioral)
significance. A category can be a class of behaviorally significant
objects that approximately share a set of defining features.
Approximate, rather than specific, is the suitable term, since it
enables “openness” in category formation. How organisms sort
objects into different categories according to features is one of the
basic questions in the behavioral and cognitive sciences. The
problem is general, as an “object” can be any recurring signal
(associated with experience or species-specific knowledge) and
“sorting” is related to any differential response to it. Categorization—
the ways that categorize information—plays a critical role in
perception, thinking, and language (see Harnad 1987).
 Sign Stimuli

Stimulus perception in male sticklebacks

To understand how an animal interprets a stimulus means
investigating how the animal responds to it. In his famous textbook
The Study of Instinct, Tinbergen (1951) introduced the topic of
stimulus perception with an experiment (Figure 2.1A):
Figure 2.1 (A) The “reflection experiment.” Toward its mirror-
image a male stickleback assumes a vertical threat posture.
(Abstracted after a photo in Tinbergen 1951). (B) Great blue heron,
Ardea herodias, in threat-display: bill/neck vertically straightened.
(Abstracted after a photo in:
http://redandthepeanut.blogspot.de/2010/06/two-great-blue-
herons-face-off-in.html [accessed: 08/11/20]).
A male three-spined stickleback, Gasterosteus aculeatus, seeing its
reflection in a mirror, assumes a vertically oriented body posture
with the head pointing downward.
Tinbergen listed a set of issues that must be addressed in order to
answer why, in a causal sense (Chapter 1), the animal does this.
These questions concern the processes hidden in a “black-box” that,
so to speak, translates the stimulus into the behavioral reaction
(Figure 2.2).
Figure 2.2 Looking into a “black-box”: principles of brain
function mediating between sensory input and motor output.
(Modified after Ewert 1976).
First, we must classify the reaction in a behavioral context (e.g.,
reproduction, male–male aggression). The fact that a male
stickleback in reproductive state shows this behavior suggests the
involvement of hormones, which requires neurochemical
investigations. Then we should examine the releasing features of the
visual stimulus and analyze the neuronal instruments that extract
the features, exploring the processes responsible for stimulus
recognition and localization. Furthermore, the releasing value of a
stimulus may depend on motivation and experience. The
information regarding stimulus features, locus, memory, attention,
and motivation yields a releasing mechanism that activates motor
pattern generation responsible for eliciting the adequate behavior.
What Tinbergen is suggesting is that “neural orchestration” in the
whole brain participates in what appears, initially at least, as a
relatively simple stimulus–response.
Back to the “reflection experiment” (Figure 2.1A): what is the
ethological background? A male stickleback encountering a
conspecific male in its territory changes its longitudinal body axis
into a vertical head-down position, thus signaling threat. In fact,
when placing a male into a narrow glass tube (Figure 2.3a), the
responses of territorial males were stronger when the tube was
oriented vertically compared with when the tube was oriented
horizontally (Ter Pelkwijk & Tinbergen 1937; cited in Tinbergen
1951). Furthermore, the red coloration of the male’s belly contributed
to the release of aggressive attacks. Strongest responses were
obtained when both features—vertical posture and red color—
occurred together. Later, Heiligenberg et al. (1972) demonstrated
quantitatively in the male perch, Haplochromis burtoni, that its
black eye-bar delivered a threat-signal to conspecific males that, too,
was enhanced if the fish assumed a vertical head-down posture.
Figure 2.3 Configurational features in sign-stimuli. (a) Three-
spined stickleback, in a glass tube, in threat posture and prevented
from assuming a threat posture. (Ter Pelkwijk & Tinbergen 1937.) (b)
Parent thrushes simulated by a head(h)/rump(r) model. Nestling’s
gaping (arrow) is directed toward parent’s head. (Tinbergen &
Kuenen 1939.) (c) A bird model moving overhead turkeys resembles
a goose-like bird or a hawk, depending on the movement direction.
(Tinbergen 1948; cit. 1951) (d) A moving small stripe signals either
prey or threat to a toad (cf. Figure 2.4), depending on its orientation
relative to the direction of movement (arrow). (Ewert 1968; cit
1984.).
Figure 2.4 The prey (a, c) vs. threat (b, d) configuration of a stripe
traversing a common toad’s visual field in different directions
(arrows). (Ewert et al. 1979; cit. Ewert 1984; cf. Suggested Reading,
Movie A2.).
Tinbergen referred to such a stimulus, composed of different
features, as a sign-stimulus. More generally, he pointed out that a
feature A combined with a feature B may provide a certain sign-
stimulus, but that feature A in combination with a feature C may
provide a different sign-stimulus. For example, in male sticklebacks:

red belly and head-down posture addresses a threat signal to

conspecific males, but not to females;
 red belly and zigzag dance (Chapter 3) addresses a courtship
signal to conspecific females, but not to males.

It is the combination—configuration or “Gestalt” (e.g., see Koffka

1922)—of behaviorally relevant features that determines the
releasing value of a sign-stimulus in the sense of a “stimulus-
pattern.” Its perception requires pattern recognition—a process, in
which genetic and/or learning factors can be involved.
A configuration is perceived as the whole. This means that the sum of
the responses to the features, when each feature is presented alone,
is significantly less than the response to the complete array.
Furthermore, recognition is independent of certain changes in other
stimulus parameters as long as these do not affect the configuration.
This phenomenon is called invariance.
Sign-stimuli provide parsimonious ways of encoding information to
release adequately motivated behaviors. They also have survival
value, since they are recognized quickly and responded to rapidly
and unambigiously. With these attributes in mind, we continue to
use the term sign-stimulus. Its efficacy can be analyzed in
experiments using dummies by changing, omitting, adding, or
exaggerating certain features.
Sign-stimuli allow humans to communicate with other animals. The
wildlife biologist Kent Clegg used his ultralight aircraft as a sign-
stimulus for captive-bred endangered whooping cranes, Grus
americanus. Simulating their parents, he painted the wings of the
plane white with black tips and thus instructed the young cranes to
fly and follow the small aircraft. After leaving Idaho and making
three overnight stops, he succeeded in having the young follow the
plane on their first migratory trip at 35 mph—matching crane’s
flight-speed—for 800 miles to their winter residence in New Mexico
(see also https://friendsofthewildwhoopers.org/whooping-cranes-
facts-management [accessed: 08/11/20]).

Principle of configurational sign-stimuli: picking out

visual key features
Static and dynamic configurations
Sign-stimuli can be simple, such as vibration or bodily touch,
suitable to elicit escape in a crayfish, or more complex. For example,
configurational stimuli are determined by several features.
Depending on the way features are related to each other, Tinbergen
(1951) distinguished a static configuration (involving spatial
relationships) from a dynamic configuration (involving
spatiotemporal relationships, such as motion).
An example of static configuration is the sign-stimulus eliciting
gaping in nestling thrushes, Turdus merula, toward the parent. In
experiments using dummies (Figure 2.3b), the parent was modeled
by two solid disks of different diameters, the large one simulating its
rump r, the smaller one its head h. Nestlings aimed at the head, if the
head-to-rump ratio was 1:3. This was examined with a “two-head-
rump” model giving nestlings a choice between a head in 1:2 ratio
and an adjacent one in 1:3 (Tinbergen & Kuenen 1939; cit. Tinbergen
1951). Hence h and r are features; their spatial arrangement yields
the configuration. Its recognition is invariant to a change in other
stimulus parameters, e.g., in size—within limits—provided the 1:3
ratio is preserved. Although movement improved the efficacy, the
spatial relationship between head and rump turned out to be the
prominent sign.
An example of dynamic configuration concerns the goose/hawk
discrimination. Tinbergen (1951) showed that a bird model (Figure
2.3c) elicited escape in young turkeys, Gallopavo meleagris, when
the model was flown overhead with the short end and the wings
leading, simulating the silhouette of an airborne bird of prey. But the
same model flown with the long end leading, resembling a harmless
goose-like bird, was ignored. Subsequent research has shown that
this configurational discrimination resulted from stimulus-specific
habituation discussed later in this chapter.

Are there comparable signs with threatening stimuli across

species?
Common toads, Bufo bufo, interpret small elongated objects—a
worm or millipede—as prey. Experimentally, it can be shown that a
2.5 × 40 mm stripe oriented parallel to the direction of its movement
releases eager prey-catching (see Further Reading, Movie A2). The
same stripe suddenly oriented crosswise to the direction of
movement—in a split-second—leads the toad to “freeze.” That
configuration signals threat. Prey-like again, the stripe immediately
elicits strong prey-catching activity, etc. The discrimination of these
dynamic configurational “opposite stimuli” is invariant to changes in
movement direction (Figures 2.3d, 2.4a–d) and velocity. This
phenomenon was also observed in terrestrial urodeles (Finkenstädt
& Ewert 1983).
Interestingly, the mudskipper Periophthalmus barbarus, an
amphibious fish (Burghagen in Kutschera et al. 2008), and the
preying insect Sphodromantis lineola (Kral & Prete 2004; cit., 2004)
respond to such test-stripes in the same way as toads. Toward the
threat-configuration, mudskippers may even raise their dorsal fins,
thus threatening back.
Generalizing, a configurational feature perceived as threat here, is a
contrast-border aligned crosswise to the direction of its movement.
Unlike earthworms, snakes—the “archenemies” of toads—reveal such
visual cues during locomotion: elevating head, undulating sidling
body. Faced with a snake (Figure 2.5a) or a dummy snake (Figure
2.5b) a toad keeps eye contact, stops locomotion, discharges skin
poison glands, swells up, assumes a stiff-legged defensive stance,
and, eventually, presents its head and back like a shield difficult for
the snake to handle (Ewert & Traud 1979, cit. Ewert 1984).

Figure 2.5 (a) Common toad displaying anti-predator behavior

toward a ring snake or (b) to a snake dummy. (Courtesy of R. Traud.)
.
Threatening postures, as components of agonistic behavior, shown
for stickleback (Figure 2.1A), perch (Heiligenberg et al. 1972), or
great blue heron (Figure 2.1B), are widespread in the animal
kingdom. The caterpillar of Hemeroplanes triptolemus, if attacked,
displays a scaring snake-like posture that is disregarded as prey. If
human divers encounter a shark, experts recommend assuming an
erect posture that will not fit shark’s prey schema.
We are tempted to speculate that during human evolution, the
upright gait had—inter alia—a bearing on emitting a threatening
signal. Wagging the erected forefinger against somebody as threat
gesture, we interpret as a ritualized behavior. In fact, the significance
of signs of that kind was known phylogenetically for a long time, such
as with the amphibian genera Bombina and Bufo that emerged in the
Jurassic and Paleocene, respectively (cf. Ewert & Burghagen 1979;
cit. Ewert 1984).

Configurational Stimuli in Human Perception

Configurational perception allows one to extract a figure from
background. This is not feasible if both figure and ground are
composed of similar items: the figure is masked like a needle in a
haystack. However, if items of the figure differ from the background
in one aspect, say if they move coherently, the figure isolates. When
an observer moves in front of a patterned stationary three-
dimensional landscape, nearby objects move faster than those
farther away and isolate from each other as flowers, bushes, trees,
hills, etc.
In social communication, the recognition of faces plays a prominent
role. Human neonates track a face model markedly further than they
will follow scrambled face components (Valenza et al. 1996). Babies
are “face-recognition experts.” At two months of age, the learned
recognition of individual faces proceeds, while generating different
face categories. At six months, a pattern of perceptual narrowing
emerges, whereby a category of familiar faces is favored (Pascalis et
al. 2002; cf. also Bower 1966 and Table 2.2). A comparable
ontogenetic phenomenon of narrowing the perceptual window is also
known from auditory perception. Young infants distinguish the
phonemes of different languages much better than do infants from 8
months of age, who focus on the native language.

Table 2.2 From heterogeneous summation to Gestalt-perception

in 24 babies at different ages (n = 6/age). (Modified after Bower
1966).
In human Gestalt perception invariance plays a role. In e-mails, the
configurations (:-) and (:-/ are correctly interpreted despite their
sideways orientation. We recognize the letter B independently of
boldface, contrast, italics, fonts, or size (Table 2.1). However,
invariances are not unlimited: depending on orientation and context,
it can be interpreted differently.

Table 2.1 Invariance in Gestalt perception and its limitation.

An example of invariance in auditory perception concerns melodies.
We recognize Mozart’s “Eine kleine Nachtmusik” independently of
the key in which it is played or its instrumentation or whether it is
whistled or sounded on a comb.
The Relational/Combinatorial Principle of Sign-stimuli from
Other Sensory Modalities
The notion that the releasing values of sign-stimuli often depend on
characteristic relationships between features accords with examples
from other sensory modalities, such as chemical or auditory signals
used in communication.
Female moths use sex-attractant pheromones to invite their males.
In many moth species these consist of blends of two chemical
compounds (Kaissling 2014). Whereas the same two components can
be emitted by phylogenetically related species, a species-specific
signal is produced by a characteristic mixture: “principle of
parsimony.” Among species of the North American female leaf-roller
moth the proportion of the pheromone compounds [Z]-11-
tetradecenyl-acetate: [E]-11-tetradecenyl-acetate is:

90:10 in Archips mortuanus

60:40 in Archips argyrospilus
17:83 in Archips cerasivoranus

In various species of frogs, to choose another example, the females

are attracted by the advertisement calls of the conspecific males if
certain low-frequency and high-frequency components of sufficient
energy in the call power-spectrum coincide (Capranica 1976):
200 Hz and 1400 Hz, bullfrog, Rana catesbeiana
500 Hz and 1500 Hz, leopard frog, Rana pipiens
900 Hz and 3000 Hz, green tree frog, Hyla cinerea.
Male adult bullfrogs among themselves evoke advertisement calls in
a chorus. However, if young males start to take part, the adults
become mute. Although immature, the young display sort of
advertisement call, albeit a bit unsounded: high-frequency peak at
1400 Hz, but the low-frequency peak not at 200 Hz rather around
600 Hz. This 600-Hz component inhibits adult male’s mating-calling
and female`s interest to follow. With age the vocal cavities of young
males increase, so that the low-frequency peak shifts perfectly to 200
Hz (Capranica 1976). Hence, females are protected twofold from
meeting an immature partner.
Unlike Rana or Hyla, the advertisement call of the Puerto Rican
male treefrog, Eleuterodactylus coqui, is made up of a sequence of
two notes with different conspecific addressees: the “Co”-note of
1100 Hz is addressed to males in male–male territorial interactions;
it is followed by the “Qui”-note, upward sweeping from 1800 to 2100
Hz, to attract females (Narins 1981). The advertisement call of male
tungara frogs, Engystomops pustulosus, too, consists of a sequence
of different notes: a whine followed by several lower-pitched chucks
(Ryan & Rand 1995). Whereas the whine serves species recognition,
the chucks enhance the call’s attractiveness to females. Males of the
related species E. coloradorum whine but do not chuck. If given a
choice, their females prefer E. pustulosus calls (Chapter 12).
Enhancing effects may also result from combining features of
different sensory channels. When prey is difficult to carry for
Aphenogaster ants, they recruit help from workers by emitting a
pheromone. If food competitors are present and time is short, ants
also deliver a vibrational stimulus. Vibration alone has no effect on
worker recruitment, but in combination it enhances the pheromone’s
efficacy (Markl & Hölldobler 1978). Workers of leaf-cutting ants,
Atta cephalotes, respond to a pheromone or vibration with
recruitment behavior. If these ants are given a choice, they choose
the combination, suggesting a kind of summation effect.

From Stimulus Summation to Supernormal

Stimuli

Heterogeneous summation
The concept of Gestalt implies that the efficacy of a configuration is
greater than the sum of the efficacies of its components (features).
However, there are examples showing that independent different
stimulus features are additive in their efficacy, whereby the whole is
equal to the sum of its parts (Seitz 1940: “Reizsummenphänomen”).
For instance, herring gulls, Larus argentatus, recognize their eggs by
different features, such as size, shape, color. These features are
additive in their influence upon retrieval of an egg having rolled out
of the nest. Heiligenberg and coworkers (see Leong 1969)
quantitatively demonstrated a comparable phenomenon in male
perch, Haplochromis burtoni. The level of aggression A in these fish
averaged at x bites/min. In a model fish (a), a black eye-bar
increased the attack rate of conspecific males at
Aa = x + 2.8 bites/min,
while a model fish (b) with orange spots on skin, but no eye-bar,
lowered the attack rate:
Ab = x–1.7 bites/min.
A model fish (c) containing eye-bar and orange spots caused an
attack rate of
Ac = x + 1.1 bites/min,
which correlates well with the algebraic sum of the rates (1.0)
obtained in (a) and (b). Hence, the opposite effects of both features
summed algebraically: “heterogeneous summation.”
It is suggested that Gestalt perception of human faces is derived from
heterogeneous summation both in phylogenetic and ontogenetic
histories. For example, Bower (1966) showed in 2-month-old human
babies that the number of conditioned orienting responses toward a
face-model—consisting of head-outline, eye-dots, and cross—equaled
the algebraic sum of the orienting activities measured to each
component of the model (Table 2.2). In 5-month-old babies the
responses to the face-model were about twice as high as the sum of
the responses to each face component: Gestalt perception.

Supernormal stimulus
Stimulus summation introduces the phenomenon of supernormal
stimulus, i.e., a stimulus that is more effective at eliciting a response
than the stimulus for which it evolved. In studies using dummies,
various examples show that exaggeration of a sign-stimulus leads to
an extraordinary increase in its efficacy. The courting behavior of a
male stickleback depends on the swollen abdomen of the pregnant
female. When presented with two model females, one showing a
normal swollen and the other a hyper-swollen abdomen,
respectively, the male will choose the latter (Rowland 1989).
Another example concerns the egg-retrieval behavior of brooding
greylag geese, Anser. After an egg rolls out of its nest, the goose
extends its neck toward the egg and with its flat mandible slowly rolls
it back toward the nest (Lorenz & Tinbergen 1938; cit. Tinbergen
1951). If the goose is offered a choice of eggs of different sizes placed
outside the nest, the largest one will be preferred, even if it is twice
the size of its own egg and thus difficult to handle. Egg retrieval by
the brooding herring gull, Larus argentatus, has been studied in
even more detail by Baerends and his colleagues (Baerends & Drent
1982). In hundreds of experiments carried out over many years, they
showed that a green, speckled egg the size of a football was preferred
over the gull’s own brown egg, which is the size of a large hen’s egg.
Humans, too, take advantage of exaggerated stimuli for various
communicative purposes. Messages can be made appealing by
stressing features. In cartoons, caricatures, and graphic icons certain
cues are accentuated in order to abstract and emphasize the
expression of a subject or an object. Women’s eyes and lips are
underlined cosmetically in order to make the face attractive and
distinctive from other faces.

Behavioral Ways of Stimulus Selection

Stimulus-specific habituation implies stimulus

discrimination
Hinde (1954) showed that, in many cases, the responsiveness of
animals to the same repeatedly presented stimulus passes three
stages: an initial increase (warming-up), a plateau of maximal
activity, and a subsequent decrease until the stimulus is neglected
(habituation) (see Chapter 8). There are instances showing that
habituation can be stimulus-specific, such that a small change in the
stimulus to which an animal has become habituated can produce
dishabituation: a sudden increase in response compared to
prehabituation levels. The phenomena of stimulus-specific
habituation and dishabituation provide a method to investigate
distinctive features of stimuli (Table 2.3).

Table 2.3 Feature discrimination causes dishabituation.

For example, the prey-catching activity in the common toad
habituates when the animal is repeatedly offered a small, orthogonal,
triangular piece of black cardboard moved with its small side leading
and the tip trailing (a):
If immediately after habituation, the toad is offered the triangle’s
mirror image (b), the prey-capture responses return immediately
(Ewert & Kehl 1978; cit. Ewert 1984; see also Further Reading, Movie
A3). Another example: when young gallinaceous birds are exposed to
any medium-sized silhouette from a bird flying above, they exhibit
escape behavior. However, over time, young turkeys, Gallopavo
meleagris, become habituated to goose-like birds (long neck, short
tail) flying overhead recurrently, whereas the less frequently seen
birds of prey (short neck, long tail) continue to be avoided (Schleidt
1961). This explains the goose/hawk discrimination (Figure 2.3c).
The phenomenon of stimulus-specific habituation also occurs in
human perception. During exercise in a fitness center we habituate
rapidly to the smell of our own sweat but readily detect the smell of
another person. We habituate to the ticking of an old-fashioned
clock. If the clock is replaced by another one, ticking somewhat
differently, we will notice this sound—until habituation.

Search images facilitate stimulus recognition

When birds discover a tasty cryptic prey in their environment, for
example, a type of insect difficult to detect because it is embedded in
masking distractors, they employ a search image. A predator using a
search image takes one type of prey and neglects others, even if the
types—e.g., investigated in a choice procedure—appear equally
attractive (Langley et al. 1996). A search image neglects certain cues
from the complete image of the object being sought, but rather
focuses attention on particular cues of the search object. The
discrimination principle is in some ways opposite to stimulus-
specific habituation, since the searcher tends to see what it expects to
see.
There are parallels in human perception. Suppose we want to pick
blueberries in the forest. At first glance the bushes seem to be empty
since the dark-green leaves distract from the berries. By
concentrating on the dark-blue coloration of the berries—and
“printing” a search image—suddenly it seems quite easy to collect
them. Expecting a visitor at an airport, not seen for a long time, we
have a search image in mind of what the visitor will look like, based
upon experience or a photograph—and this works. But if the visitor
has changed his image, e.g., wearing a beard, we might have been
more successful in identifying him without search image.

The behavioral meaning of stimuli can depend on motivation

Responses to particular stimuli are also subject to internal (e.g.,
hormonal) and external (e.g., photoperiod) factors (Figure 2.2; see
also Chapters 3–8). In spring, when nesting motivation in birds is
high, pieces of branch are attractive for nest building. After hatching,
a juicy piece of branch may be chosen as food for the nestlings. The
sharpness of stimulus identification, too, may depend on the level of
motivation. During the mating season, if in a pond there is no female
toad rapidly available to a highly motivated conspecific male, the
male may clasp even a piece of bark it encounters—somewhat
reminiscent of Goethe’s Mephisto who promises Faust “with this
drink in your body, soon you’ll greet a Helena in every girl you meet.”

Analyzing Neural Processes that Underlie

Perception of Sign-stimuli
 The conclusion that a sign-stimulus is configurational is merely a
provisional way of describing the complexity of the sensory
stimulating process. It is thus a challenge rather than a solution, a
challenge to analyse the complex system of processes denoted by
the convenient collective name ‘stimulus’. […] Accepting a mere
descriptive term as a causal explanation […] causes a false
satisfaction which is a hindrance to further research.(Tinbergen
1951; reissued 1989, p. 79 top)
Here, Tinbergen is stressing the need to analyze the processes
underlying ethological concepts—down to the neuronal level—by
means of a broad spectrum of physiological/anatomical methods. He
introduced the discipline of this causal-analytical research as
“ethophysiology,” today called neuroethology or behavioral
neurobiology (Ewert 1976; Carew 2004; Zupanc 2019).

The classical concept of innate releasing mechanism

What mechanism “translates” a sign-stimulus into the adequate
behavior (Figure 2.2)? The concept of innate releasing mechanism
(IRM), introduced by Konrad Lorenz and Niko Tinbergen, concerns
the observation that some organisms are apparently able to
recognize behaviorally meaningful stimuli never before experienced
in their environment and to respond to them in a predictable manner
(Tinbergen 1951). A sign-stimulus—also called “key-stimulus” was
thought to activate the IRM much like a safe is unlocked by a key.
The comparison with a key-lock principle may be misleading,
however, since it suggests that the IRM is just responsive to one
specific stimulus. This is not the case. More likely a sort of “group
key” works, since the mechanism responds to a category of stimuli
that share a set of defining features. Controversy also surrounds the
term “innate” (Chapter 7). A revised concept considers that both
experience and genetic factors contribute to behavior, and most
authors now refer to the concept as a “releasing mechanism” (RM),
or a “releasing system” (Ewert 1987, 1997).

Toward neuronal correlates of releasing systems

The concept of “releasing system” suggests a neuronal sensorimotor
interface translating perception into action (Figure 2.2). At its
afferent (input) side this interface has stimulus recognition and
localization properties; its efferent (output) side runs (“commands”)
the corresponding motor pattern generating system (Table 2.4).
Theoretically, the simplest structure of a ballistic releasing system
would entail a “command neuron,” CN, operating in a processing
stream that—once triggered—proceeds to completion.

Table 2.4 Command neuron hypothesis.

At first glance the giant axon of giant lateral interneuron in crayfish,
controlling the escape tail-flip in response to mechanosensory
stimulation, serves as a CN (Wiersma & Ikeda 1964). The idea that a
neuron triggers a behavior was challenging and sparked intense
debate among neuroethologists. Kupfermann and Weiss (1978)
pointed out that a CN must fulfill two conditions: its excitation is not
only necessary but also sufficient to activate the behavior. This
strong definition can be examined: electrically exciting that neuron
should be sufficient to elicit the corresponding behavior; removing it
should abolish the behavioral response to the peripheral stimulus.
For experimental studies, the best candidate of a CN is the
reticulospinal Mauthner cell in teleost fish that, in response to
vibratory stimulation, triggers the fast-body-bend escape reaction.
However, quantitative investigations showed that a Mauthner cell
did not fulfill the “double-condition” (Eaton 2001); a command
neuron that does it, convincingly, remains to be discovered. If several
neurons are involved in a command function, we use the term
“command system.”
In fact, a releasing system—relying on attention and motivation—
may take advantage of adequate receptor cells and assemblies of
feature-sensitive/selective interneurons. This system translates the
information of a sign-stimulus into a command, which is appropriate
to activate the corresponding motor system (Figure 2.2).
Scent-coding by specialized receptor cells in insects
Scent perception in many insects involves receptor cells best tuned to
chemical sign-stimuli. In the silk moth, Bombyx mori, scent hairs on
the male’s antennae are equipped with “olfactory specialist receptor
cells.” These respond maximally to a female’s key pheromone
bombykol, as mentioned above (Figures 2.6 and 2.7a). A bombykol
receptor is also somewhat sensitive to chemically related
compounds, however, at 10- to 1000-fold higher concentrations
(Kaissling 2014).
Figure 2.6 Male silk moth in alerted position, with combed
antennae elevated (top). (Courtesy of R.A. Steinbrecht.) Bottom:
schematic section through a scent hair with pore tubuli and two
scent receptor cells. Arrangement for recording impulses from
bombykol receptor. (Modified after R.A. Steinbrecht).
Figure 2.7 Principles of scent detection in moths by specialist
receptor cells and interneurons. (a) In male silk moths a receptor
channel is specialized for female’s sex pheromone bombykol. (b) In
male nun moths a receptor channel is specialized for the sex
pheromone (+)-disparlure. (c) In male gypsy moths two types of
receptor channel are specialized for two pheromone compounds:
(+)-disparlure stimulates interneurons, whereas (−)-disparlure
inhibits their response to (+)-disparlure. (d) In male leaf-roller
moths the concurrent excitatory influences of the two receptor
channels specialized for the two sex pheromone stereoisomers
(Z)-11-tetradecenyl-acetate and (E)-11-tetradecenyl-acetate are
essential to activate interneurons. (Compiled from data in Hansen
1984; Kaissling 2014.) Note that a behavioral response requires
activation of many specialist receptor cells and corresponding
interneurons.
Other examples of such narrow-band olfactory specialists are the
meat receptor in Necrophorus beetles, the rotten-meat receptor in
blow-flies Calliphora erythrocephala, and the grass receptor in the
locust Locusta migratoria. Grass receptors respond to chemically
related components of fresh grass, such as hexenol, hexenal, and
hexenic acid (Kafka 1970).
In leaf-roller moth species, as mentioned above, the males have two
types of specialized receptor cells, each one tuned to a different
pheromone, both emitted by the conspecific female in a
characteristic proportion (e.g., Figure 2.7d), which minimizes the
risk of mating with males of inadequate species.
Another way of species separation are interspecific inhibitors.
Females of the nun moth, Lymantria monacha, and the gypsy moth,
L. dispar, produce the male-attracting compound (+)-disparlure
(Figure 2.7b, c). However, the female nun moth also produces (−)-
disparlure, which stimulates a specialist receptor cell in the male
gypsy moth to inhibit its behavioral response to (+)-disparlure
(Figure 2.7c) (Hansen 1984). This kind of species separation would
function in one direction, since a male nun moth could be attracted
by (+)-disparlure of the female nun or gypsy moth. Presumably,
female gypsy moths are less attractive to male nun moths because of
the high emission concentrations of (+)-disparlure (Figure 2.7c ).
In addition to specialist receptors there are “generalist receptor
cells.” These—showing partly overlapping response spectra—respond
differently to various odor compounds. Such cells may be suitable to
distinguish odorous substances by individual experience.

Visual feature detection in amphibians: a

multimethodological analysis
Summertime in a forest, evening is falling, twilight—a toad Bufo bufo
sits in front of its shelter. Suddenly, a moving milliped shows up. The
toad orients its head and body toward it, watchfully creeps up, fixates
it binocularly, and snaps. In the catching sequence, the releasing
system of each behavioral component decides “what is it?” The
releasing systems differ in determining “where is it?” and, in case of
prey, “how can it be caught?”, e.g., by stalking or pouncing, tongue-
flipping, or jaw-grasping. Key factors of these releasing systems
involve prey-selective neurons in the midbrain optic tectum. How
was this discovered?
Toward a features-relating-algorithm as a principle in
configurational perception
How does prey look to toads? Using dummies, an experimental
procedure enables one to measure a toad’s prey-catching activity in
response to the visual feature in question and to change it, while
other stimulus parameters are held constant. This allows one to
evaluate the effect of this change on prey-category formation (Ewert
1974, 1984; see also Further Reading, Movie A2).
Since toads hunt moving objects, it makes sense to change a visual
variable related to movement. Figure 2.8A shows that stepwise
extending a moving rectangular black stripe against white
background parallel to the movement direction raises the prey value
within limits, e.g., resembling woodlice, millipedes, or worms (chart
p). Extending a stripe crosswise to the movement direction
progressively reduces the prey value and signals threat (chart c).
Toward squares of different sizes, the influences of both features p
and c interact, thus yielding a preferred square of approximately 10
mm edge length, e.g., simulating bugs (chart s). A large moving
square of about 100 mm edge length, like a shadow from an airborne
predator, elicits escape.
Figure 2.8 Prey-feature analysis in common toads. The
configurational paradigm involves three types of moving rectangular
black stimuli: stripes extended parallel (p) or crosswise (c) to the
movement direction, and squares (s) of different edge lengths. A)
Prey-catching activity, RB (mean values, n = 15 toads (Ewert 1984).
B) Activity, RN, of prey-selective tectal T5.2 neurons (mean values, n
= 18 neurons) in awake toads being pharmacologically immobilized
(required for stable neuron recordings through the entire stimulus
program within a perimeter device); deg: degrees of visual angle
(Wietersheim v. & Ewert 1978; cit. Ewert 1984.).
In reminiscence of the term key-stimulus, the “key” is inherent in an
algorithm—calculation method—that enables the toad’s visual
system to get access to the configurational prey-category by
weighting p and c (Figure 2.8A) according to the equations shown in
Table 2.5.
Table 2.5 Constants α and β determining essential traits of the
features (p,c)-relating-algorithm; k1 and k2 depending on other
stimulus parameters and the toad’s prey-catching motivation. RB =
prey-catching orienting activity [responses/30s]. (Ewert 1984).
Variations in other stimulus parameters, e.g., movement direction
(Figure 2.4), velocity, motion pattern, or background pattern (cf.
Movie A2) influence a toad’s general prey-catching activity, but the
basic effects of the configurational features p and c (Figure 2.8A,
charts p,c) are invariant to those changes (Burghagen & Ewert 1982,
1983).
Developmental studies showed that the features-relating-algorithm
emerges—without prey-catching experience—after metamorphosis
with transition from aquatic to terrestrial life (Traud 1983; cit. Ewert
1984). The principle is common to terrestrial anurans but shows
species specificities (Burghagen 1979; Ewert & Burghagen 1969; cit.
Ewert 1984).

What does the eye tell its brain?

In leopard frogs, Rana pipiens, Barlow (1953) and Lettvin et al.
(1959) recorded spike activities of retinal ganglion cells—from axon
terminals in the optic tectum—toward objects traversing the center
of a cell’s visual receptive field, RF (Figure 2.9A). The four classes of
retinal ganglion cells differ in sensitivities to dimming/brightening,
and to an object’s contrast, velocity, and size. The latter is correlated
with the diameter of the excitatory RF ranging among different cell
classes from 2 to 16 degrees visual angle.
Figure 2.9 Neuroimaging toad’s visual system. (A) Dorsal view of
toad’s brain. A stimulus (S) traverses the receptive field (RF) of a
retinal ganglion cell (G) of the right eye (E), whose optic nerve (ON)
projects to left optic tectum (T) and pretectal thalamus (TH). R,
receptor cells; DT, dorsal tectal lobe; VT, ventral tectal lobe; MP,
telencephalic ventro-medial pallium; M, medulla oblongata. (B)
Functional neuroimaging: 14C-2DG-uptake in brain transverse
sections at levels a-d. (see also Suggested Reading, Movie A1). (a)
After hand-conditioning of a right-eyed toad, left MP showed 14C-
2DG-uptake toward the conditioned stimulus (Finkenstädt & Ewert
1988). (b) Toad escaping a predator stimulus showed strong 14C-
2DG-uptake in DT and TH. (c) Toad stiffening toward a threat-like
moving stripe presented to the right eye showed moderate 14C-2DG-
uptake in left TH and less so in DT. (d) Toad binocularly snapping
toward a prey-like stripe showed strong 14C-2DG-uptake bilaterally
in VT (Finkenstädt et al. 1985.).
The pioneering work by Barlow and Lettvin and coworkers made us
realize that retinal ganglion cells—the output neurons of the retinal
network—perform a first-stage analysis of visual input. Subsequent
quantitative investigations, however, showed that retinal processing
is not sufficient to recognize prey, as suggested formerly. Applying
the configurational stimulus paradigm, no correlation was found
between retinal neuronal and prey-catching activities in common
toads (Ewert & Hock 1972; cit. Ewert 1984). Consequently,
configurational feature analysis requires further processing by
retina-recipient neurons in the brain.

In search of brain structures involved in feature detection

The visual field of toad’s retina is mapped—via retinal ganglion cell
axons along the optic nerve—inter alia mainly in the contralateral
optic tectum and pretectal thalamus (Figure 2.9A ). If, in the absence
of retinal input, a locus in the tectum of a free-moving toad was
excited by trains of electrical impulses delivered by an implanted
electrode, the toad responded with orienting or snapping. Probably,
the electrical stimulus excited neurons mediating information on
prey recognition (Ewert 1974, 1984). In the pretectal thalamus, focal
electrical stimulation elicited avoidance, such as ducking or jumping
or freezing associated with secretion of skin poison glands—all-up
behaviors known to be released by airborne or ground predators.
A neuroimaging technique allows one to check the regional neural
activities in response to prey or predator stimuli (Finkenstädt et al.
1985). If 14C-labeled 2-deoxy-D-glucose, 14C-2DG, was administered
systemically to the toad, active neurons were confusing the 2-deoxy-
D-glucose with glucose, hence taking it up, but failing in
decomposing it like glucose. The more active neurons were, the
greater the storage of 14C-2DG and thus the radioactivity measured
in brain sections later on (Figure 2.9B; see also Suggested Reading,
Movie A1).
Figure 2.9 Bd shows a color-coded autoradiographic image of a
transverse section through the midbrain of a toad snapping toward a
prey-like stripe moving in the binocular field. Strong radioactivity
was focused bilaterally on the ventrolateral tectum: “snapping-
evoking areas” (Figure 2.9 Bd, VT). In a toad escaping from a moving
large square, the overall radioactivity was high, strongest in tectal
and pretectal/thalamic structures (Figure 2.9 Bb, DT and TH). This
substantiates Tinbergen’s prediction that “neural orchestration” of
the whole brain may participate in a stimulus-response. In a toad
becoming stiffened toward a threat-like stripe moving in the right
visual field, moderate radioactivity in the corresponding left pretectal
thalamus was stronger than in the optic tectum (Figure 2.9 Bc).

Configurational object perception involves parallel processing

streams and their interaction
Optic tectum and pretectal thalamus are involved in prey catching
and predator avoidance. At the neuronal level of analysis,
extracellular recordings from toad’s optic tectum reveal monocular
T5-type neurons. One type T5.1 is sensitive to extension of feature p.
A second type T5.2 is selective in that its activity—according to prey-
catching activity (Table 2.5)—increases with extension of p (Figure
2.8 Bp), within limits, but progressively decreases with extension of c
(Figure 2.8 Bc) The activity of type T5.2 in response to different
configurational objects reflects the probability that an object fits the
prey category (Figure 2.8, cf. A, B).
Recordings from pretectal thalamus reveal various types of TH-type
neurons, among them monocular neurons TH3 responsive to
extension of c or p and c. All these neurons are integrated in a
feature-analyzing network.
The “window hypothesis” (Figure 2.10A–C) suggests that feature p is
analyzed in a retinotectal processing stream originating in certain
classes of retinal ganglion cells (classes R2 and R3) and continuing in
tectal prey-selective T5.2-neurons. In parallel, feature c is analyzed
in a retinopretectal/thalamic processing stream originating in partly
other retinal ganglion cells (classes R3 and R4) and continuing in
TH4-neurons that are selective to predatory objects (see also
Suggested Reading, Movie A1).

Figure 2.10 “Window hypothesis” of configurational feature

analysis in toads. Illustrative schemes of feature-sensitive/selective
neurons (symbolized by circles) integrated in a neuronal network
(Ewert 1974, 2004). For explanations see text.
In detail, the evaluation of an object as belonging to the prey
category results from convergence of both processing streams on
tectal T5.2-neurons that weight feature p by excitatory tectal input
and feature c by inhibitory pretectal/thalamic input (Figure 2.10A).
Predator evaluation results from convergence of both processing
streams on pretectal/thalamic TH4-neurons that are weighing p and
c by excitatory pretectal/thalamic and excitatory tectal inputs
(Figure 2.10C). We speak of parallel distributed interactive
processing of information.

The size constancy phenomenon

Common toads evaluate the absolute size of prey binocularly and/or
monocularly (Ewert & Gebauer 1973; cit. Ewert 1984). Among
configurationally “neutral” square black prey dummies (Figure
2.8As), they prefer an object—at different distances—of about s = 10
mm edge length, generally:
s = 0.37 w for 5 ≤ w ≤ 26 [mm]; mouth-width w
including, e.g., the anuran species Bufo, Rana, Hyla, Bombina,
Alytes. The size constancy phenomenon performs with stereoscopic
vision after metamorphosis (Ewert & Burghagen 1969; cit. Ewert
1984). Monocularly, it requires an estimation of object’s distance
calculated in connection with toad’s movement relative to the object:
e.g., by triangulation (brief head shifting) and/or accommodation
(lens shifting) (Collett 1977). Depth information, for example,
reaches T5-type neurons.
Evidence: In pharmacologically immobilized awake toads, T5-type
neurons measure objects in degrees of visual angle (Figure 2.8B,s).
However, in free-moving toads, T5-type neurons are sensitive to
object’s absolute size (Spreckelsen et al. 1995). So, in accordance
with that, the features(p,c)-relating-algorithm is traceable in scales
of absolute or visual angular size (Figure 2.8 cf. A,B).

Visuomotor access
Prey-selective T5.2-neurons send their axons from the optic tectum
to the medullary premotor/motor cells that innervate jaw and tongue
muscles (Weerasuriya & Ewert 1981, Satou & Ewert 1983; cit. Ewert
1984; see also Table 2.6).
Table 2.6 Minimum number of cell types linking photoreception
and snapping reaction. Additional interacting cells (e.g., retinal
horizontal and amacrine cells; pretectal/thalamic neurons) have
tuning and specifying functions at different processing levels.
Such a neuron can be examined by recording its activity with an
electrode chronically implanted in the tectum of a freely moving toad
in response to a prey-like stripe: a strong burst of impulses preceded
toad’s tongue-projection at prey (Schürg-Pfeiffer et al. 1993). The
same stripe presented in threat configuration elicited little neuronal
activity and no prey-catching, probably due to pretectal/thalamic
inhibition (see Figure 2.10B).
Test: If a second electrode—fastened at the skull—delivered an
electrolytic lesion to the ipsilateral pretectal thalamus, a moving
stripe strongly activated the prey-selective T5.2 neuron and elicited
prey-capture regardless of whether the stripe was presented in prey
or threat configuration. Pretectal/thalamic lesion (Figure 2.10D)
impaired the discrimination between prey and threat both
neuronally and behaviorally—hence evidencing linkage between
prey-selective neuronal activity and prey-catching behavior (Schürg-
Pfeiffer et al. 1993).
No motor command can be issued when motivation and attention
are not appropriate: if a toad was satiated after feeding on
mealworms or frightened by some noise by the experimenter, a prey
object neither activated T5.2 neurons nor any prey capture.

Modification of species-specific feature detection by learning

Snapping a hive-bee, the painful/distasteful incident—conditioned
with bee’s appearance—prevents a toad from catching such bees
again (Cott 1936).
Can a toad be trained to catch a threatening stimulus? Usually, toads
are frightened by a moving hand. Feeding a toad daily with a
mealworm presented in the experimenter’s hand, the toad associated
the hand with food and became tame (Brzoska & Schneider 1978).
Approximately after a fortnight of hand-feeding—once a day—the
moving hand alone released snapping and, generalizing, a moving
large square or threat-like stripe was included to the toad’s prey
category. The species-specific prey recognition was extended by
individual experience—in terms of classic ethology, a “modified IRM”
developed (Ewert 1997; see also Further Reading, Movie A3).
How can this modification be explained? Recalling the “window
hypothesis,” we suggest in the telencephalon a neural structure that
—during hand-feeding training—becomes sensitized to the
contiguous presentation of prey and threat signals. In the learning
phase, the sensitized neurons inhibit threat detection, which leads to
a sort of “disinhibition” in prey detection.
How was this checked out? The study passed three steps: Step 1: a
toad—left eye covered—was trained by hand-feeding. Step 2: during
prey-catching of the trained toad—eyes uncovered—toward a moving
large square, the 14C-2DG-uptake rose in the posterior ventromedial
pallium of the left telencephalon (Figure 2.9 Ba, MP). This structure
obtains visual information from the right eye (uncovered during
training) and projects to thalamic/pretectal regions. This
telencephalic pallial structure is homologous to the mammalian
hippocampus known to be involved in learning. Step 3: after lesions
to that pallial structure, the training effects disappeared and the
property of species-specific prey recognition—classic ethologically
speaking, the “IRM”—reappeared (Ewert et al. 1994).
Summarizing, prey-catching releasing systems in toads contain—
inter alia—midbrain neurons with species-specific prey-selective
characteristics. Ontogenetically speaking, configurational prey
selection is present after toad’s metamorphosis in the context of
aquatic-to-terrestrial transition. It can be modified by individual
experience, e.g., via a forebrain-loop involving ventromedial pallium,
the “primordium hippocampi.”
There are promising studies concerning modulatory functions of
diencephalic pretectal/thalamic and hypothalamic nuclei on the
stimulus-response pathways that mediate prey-catching and threat-
avoiding (e.g., see Ewert & Schwippert 2006; Islam et al. 2019;
Prater et al. 2020).

Sensorimotor codes
The concept of command releasing system CRS interprets
Tinbergen’s concept of (innate) releasing mechanism in a
neurophysiological context.
A CRS considers combinatorial aspects of stimulus perception as a
sensorimotor code in a sensorimotor interface. A coded command
involves different types of neurons, each type monitoring or
analyzing a certain stimulus aspect, e.g., prey-selective T5.2 neurons.
The idea is that a certain combination of such command elements
cooperatively activates a certain motor pattern–generating system in
the presence of adequate motivational and attentional inputs. It is
suggested that certain command elements can be shared by different
sensorimotor codes.

Modeling toad’s visual pattern recognition

Building on the neuroethological results of the toad’s visual system
(e.g., Figure 2.10), artificial pattern recognizers were developed—
using systems theoretical approaches (Ewert & v. Seelen 1974; cit.
Ewert 1984)—computer models taking advantage of the relevant
cytological brain structures (Lara et al. 1982), and artificial neuronal
nets, ANNs, trained by backpropagation algorithms (Ewert 2004).
Different ANNs applying algorithms for reinforcement learning,
classical contitioning, and genetic operations are described by
Reddipogu et al. (2002) and Yoshida (2016). Hence, there are
various ways of modeling brain/behavior functions: global models
are heuristic; ANNs subserve approximation and optimization, e.g.,
by implementation of an algorithm.
Why modeling? 1) A model offers a representation of the processes
within the modeled system. Hence, models have explanatory
function. 2) Models are predictive. Predictions can be tested by
adequate experiments. The results, in turn, may improve the model.
3) Models are sort of creative since they may exhibit unexpected
properties. 4) Models provide tools toward artificial intelligence,
such as in the growing field of neuroengineering.
For example, the German Federal Ministry for Research and
Technology (BMFT) supported a joint project called “Sensori-Motor
Coordination of Robotic Movements with Neuronal Nets” SEKON
established in 1991–1994 by scholars from neurobiology,
neuroinformatics, and robotics. To study interfaces between
perception and action, in one experimental platform a modular
structured ANN simulating toad vision (Fingerling et al. 1993)—in
connection with a CCD-camera—instructed a robot to select and pick
out differently shaped work pieces from a conveyor belt (see also
Further Reading, Movie A1).

Visual Perception in Primate Cortex:

Dedicated, Modifiable, Crossmodal, and
Multifunctional Properties in Concert

Roughly comparable to toads and other vertebrates, sensory

information processing in primates proceeds in a parallel-distributed
and interactive fashion. Tremendous complexity arises from cortical
neuronal circuits with regard to feature analysis, plasticity,
multisensory integration, and sensory substitution (Kaas 1991).
Combining and binding of features is a common task (Singer 1995).
There is no one single place for perception at the “top” of a sensory
system. All processing levels contribute to the resulting picture
(Damasio 1990).
In primate vision, Ungerleider and Mishkin (1982) showed that two
neural processing streams are involved to answer the questions
“what kind of object?” and “where is the object?” (see also Hubel &
Livingstone 1987).

Ventral processing stream answering “what”

This processing stream originates in the small-celled system of the
retina, passes the related structures of the diencephalic lateral
geniculate nucleus, LGN, and reaches—via corresponding cortical
areas V1-4—the integration and association fields of the inferior
temporal cortex, ITC (Figure 2.11).

Figure 2.11 Visual processing streams in the primate cortex.

Simplified diagram; bidirectional pathways not shown. LGN, lateral
geniculate nucleus; V1-V5, visual cortical areas; ITC, inferior
temporal cortex; PPC, posterior parietal cortex. (Compiled after data
by Ungerleider & Mishkin 1982; Hubel & Livingstone 1987).
In area V1 the different orientations of contrast borders in terms of
lines (|/\—) are determined by certain neurons arranged in columns
as prerequisite for the analysis of combinations of lines and angles
between them (L V ∧ T) explored in areas V2 and V3 (Hubel &
Wiesel 1977). Shape and color are analyzed separately in layers of
area V2 and are combined in V4, thus allowing assignments like
“yellow banana.” Associations depend on connections with the
hippocampus.
The ITC is involved in the recognition of gestures and postures, e.g.,
suitable for social communication. Neuronal responses selective to
faces were discovered by Perrett and Rolls (1983) (Chapter 5).
Comparable face-selective neurons were recorded in the temporal
cortex of Dalesbred sheep: some neuron types preferring a
conspecific’s face, others responding selectively to a German
shepherd dog’s face (Kendrick 1994). It is suggested that an assembly
of differently face-tuned neurons code for the recognition of an
individual face (cf. Cohen & Tong 2001).

Dorsal processing stream answering “where”

“Where is the object?” deals with “how should it be responded to?”
This requires spatial vision in connection with analyses of object
motion and depth: starting in the large-celled retina, continuing in
related structures of LGN and processing—via corresponding areas
V1-3, V5—in the posterior parietal cortex, PPC (Figure 2.11).
The PPC contains neurons responsible for target-oriented reaching
or grasping involving arm, hand, and fingers. Such neurons fulfill
integrative tasks. Motivation plays an essential role. If a satiated
monkey was offered a banana, its visual fixation neurons failed to
respond or discharged sluggishly and the animal ignored the banana
(Mountcastle et al. 1975), in reminiscence of a comparable situation
observed in toads.
The “what” and “where/how” processing streams are not completely
segregated. A patient with damage to the “what” stream was able to
reach for an object; however, if the object’s shape required an
appropriate grasping pattern, the patient failed to grasp it.

Selective attention: what an individual does not like to

see, it may not see
Animals, including humans, may guide their perception toward
interesting parts of a scene and suppress uninteresting ones. In a
behavioral experiment monkeys were trained to draw their attention
either to a red or a green stripe (Barinaga 1997). In the
neurophysiological experiment both stripes were presented in the
excitatory visual receptive field of a red-sensitive neuron of area V4.
The neuron responded to the red stimulus. If both stimuli were
presented and the monkey was prompted to focus on the red one, the
neuron fired as expected. However, requested to focus on the green
stripe, the red-sensitive neuron was silent albeit the red stripe, too,
was present in its excitatory receptive field.
Studies applying functional neuroimaging technologies in humans
offer a look at the activity pattern in cortical visual areas. The
regional neural activity depends on which property of an object the
test-person shows interest. If a person is asked to focus on an
object’s motion, an area corresponding to V5 is mainly activated.
Paying attention to the color, an area corresponding to V4 is
principally responsive. When shape is the focus, greatest activation is
elsewhere along the “what” processing stream. In another task,
photos of faces in frames were shown. Asking whether the faces
looked different, ITC was strongly activated. Requesting whether a
face was positioned symmetrically in the frame, activation shifted to
the PPC.

Does imagination reactivate processes involved in visual

perception?
Studies from neuroimaging in humans suggest that during visual
imagination of an object, a “protocol” of the neurons operating
during visual perception of that object is reactivated. Ishai and Sagi
(1995) discuss common mechanisms of visual perception and visual
imagination. This explains, for example, why humans with lesions in
an area corresponding to V4 are unable both to recognize and
imagine colors.
A different phenomenon of imagination is sensing a stimulus that
actually affects another person. You see a child touching a hot stove
and you will feel pain, caused by activity in appropriate structures of
your brain. This kind of compassion is mediated by so-called mirror
neurons (for details see Chapter 5).

Sensory maps shrink or expand depending on supply

and demand
Sensory space is represented in the brain by topographic maps. From
lesion and regeneration studies in amphibians it is known that the
visual map of the retina in the optic tectum can be either compressed
or expanded depending on the available tectal room and on the
disposable retinal input. For example, if in a frog half of the retina of
one eye is destroyed, a blown-up map of the remaining retina will
regenerate in the entire contralateral tectum (Udin 1977).
Although such regeneration capability is not conferable to mammals,
sensory cortical maps, too, may expand or shrink depending on the
available cortical space and the need for perceptual skills. The
phenomenon is called functional remodeling. Studies applying
neuroimaging showed in the somatosensory cortex of string
instrument players an expansion of the representation of the active
digits to the detriment of the less active thumb (Elbert et al. 1995).

Universal potential of neural networks allows sensory

substitution
The fact that underemployed cortical regions take over functions of
overemployed regions is documented by neuroimaging in people
blind from early age. Their visual cortex is activated by tactually
reading Braille or embossed Roman letters or by other tactile
discrimination tasks. Evidence of this sensory substitution was
provided by transient disruption of the visual cortex by means of
transcranial magnetic stimulation TMS. This induced errors in
tactual discrimination tasks and distorted tactile perception in blind,
but not in sighted test subjects (Cohen et al. 1997). Blindness causes
the visual cortex to be recruited to a role in somatosensory
processing, which contributes to the superior tactile perceptual
talents of blind people.
Convergence of different sensory channels implies crossmodal
interactions. In humans and most animals, a sudden touch to the
body can enhance vision near that body part. Actually, cutaneous
stimulation facilitates visual responses in the visual cortex (Cohen et
al. 1997; Macaluso et al. 2000).
The notion that neural structures may have a “universal potential” is
also in line with the fact that a perceptual principle may be
implemented in different neural structures. “Lateral inhibition”—a
principle by which contrast-borders are highlighted—discovered in
the compound eye of the horseshoe crab, Limulus (Hartline 1949), is
such a principle that works also in visual and tactile perception in
vertebrates. The features-relating-algorithm—a principle of prey-
selection—realized in the brains of toad (amphibian), mudskipper
(fish), and mantis (insect) provides another example.
SUMMARY AND CONCLUSIONS
The examples and comparisons across various sense modalities
and species we have reviewed show that perceptual worlds are
shaped up to the nature of the sensory systems which in turn are
adapted and adaptable to behaviorally relevant stimuli. No
matter what the sensory world looks like, the orientation and
communication in that world requires basic strategies of
stimulus perception involving recognition and localization. The
employed tactics take advantage of individual experience as well
as peculiarities that emerged during evolution of the species in
dependence on the ecological benefits and constraints.
Animals—including humans—tend to abstract objects in terms of
configurational features. The resulting sign-stimulus elicits an
assigned behavior that depends on motivation and attention.
Sign-stimuli are as simple as possible and resemble the originals
as closely as necessary. They facilitate both perception and
communication, thereby minimizing misinterpretations between
“sender” and “receiver.” Neurobiological instruments underlying
stimulus perception range from specialized receptor cells in sense
organs to feature-analyzing assemblies of cells and feature-
detecting cells in the CNS. Certain assemblies may function in a
manner of a sensorimotor-coded releasing system, that
depending on motivation and attention, selects the appropriate
behavior.
For most sensory modalities there are sensory brain maps
containing populations of neurons selectively tuned to different
stimulus features, feature combinations or configurations.
Neurosensory networks may be omnipotent in that they display
various degrees of plasticity involving remodeling, sensory
substitution, crossmodal interaction, and learning.

FURTHER READING
Textbooks
The Study of Instinct by Tinbergen (1951) is a classic textbook of
ethology—especially impressive in view of Tinbergen’s foresight, e.g.,
in terms of the neuroethological fundamentals of behavior.
Hogan (2017) provides new insights in the study of animal behavior
including behavioral ecology, neuroscience, cognitive psychology,
and evolutionary developmental biology.
Prete (2004) presents a multi-author textbook describing in depth
what the perceptual worlds of animals of various species might be.
Readers interested in the research of olfactory perception in insects
will enjoy the ambitious review by Kaissling (2014).
Carew (2004) and Zupanc (2019) provide the best up-to-date
treatments of neuroethology.

Movies
Ewert, J.-P. & IWF (Institut für den Wissenschaftlichen Film,
Göttingen). Voice-Over: English.
If you scan this QR code with the QR app of your smartphone, or
click the URL, you’re directed to an internet TIB|AV-Portal, which
allows you to watch three English versions of movies about the
visually guided prey-catching and threat-avoidance behaviors in
toads and the underlying neurophysiological processes.
A1: Image Processing in the Visual System of the Common Toad:
Behavior, Brain Function, Artificial Neuronal Net (No.: C1805).
https://av.tib.eu/media/15148
This weblink refers to the movie dealing with: Image processing in
the toad’s visual system from behavior to brain function, which is
explained by a global model (“window hypothesis”) and simulated
by an artificial neuronal net that—in an experimental platform of
neuroengineering—advises a robot to select and pick out different
objects moving on a conveyor belt.
A2: Gestalt Perception in the Common Toad-1: Innate Prey
Recognition (No.: C1430). https://av.tib.eu/media/15241
This weblink refers to the movie dealing with: Species-specific prey-
selection in the common toad with reference to the “worm” vs. “anti-
worm” discrimination and its invariance under changes of other
stimulus parameters, such as object motion, movement pattern,
shift of retinal image (induced movement), direction of movement,
background contrast, and background texture.
A3: Gestalt Perception in the Common Toad-2: Modification of Prey
Recognition by Learning (No.: C1431).
https://av.tib.eu/media/15242
The weblink refers to the movie dealing with: Modification of
species-specific prey selection in common toads by learning, such as
visual/olfactory associative learning, visual/visual (hand-feeding)
associative learning and non-associative learning (stimulus-specific
habituation).

REFERENCES
Baerends, G.P. & Drent, R.H. (eds). 1982. The herring gull and its
eggs. Behaviour, 82, 1–416.
Barinaga, M. 1997. Visual system provides clues to how the brain
perceives. Science, 275, 1583–1585.
Barlow, H.B. 1953. Summation and inhibition in the frog’s retina.
Journal of Physiology London, 173, 377–407.
Bower, T.G.R. 1966. Heterogeneous summation in human infants.
Animal Behaviour, 14(4), 395–398.
Brzoska, J. & Schneider, H. 1978. Modification of prey-catching
behaviour by learning in the common toad (Bufo bufo L., Anura,
Amphibia): Changes in response to visual objects and effects of
auditory stimuli. Behavioural Processes, 3, 125–136.
Burghagen, H. 1979. Der Einfluss von figuralen, visuellen Mustern
auf das Beutefangverhalten verschiedener Anuren. Math.-Nat.
Dissertation, University of Kassel.
Burghagen, H. & Ewert, J.-P. 1982. Question of “head preference” in
response to worm-like dummies during prey-capture of toads.
Bufo bufo. Behavioural Processes, 7, 295–306.
Burghagen, H. & Ewert, J.-P. 1983. Influence of the background for
discriminating object motion from self-induced motion in toads
(Bufo bufo L.). Journal Comparative Physiology, 152, 241–249.
Capranica, R.R. 1976. Morphology and physiology of the auditory
system. In R. Llinas & W. Precht (eds.), Frog Neurobiology, pp.
551–575. Berlin: Springer.
Carew, T.J. 2004. Behavioral Neurobiology, 2nd ed. Sunderland,
MA: Sinauer.
Cohen, J.D. & Tong, F. 2001. The face of controversy. Science, 293,
2405–2407.
Cohen, L.G., Celnik, P., Pascual-Leone, A. et al. 1997. Functional
relevance of cross-modal plasticity in blind humans. Nature, 389,
180–183.
Collett, T. 1977. Stereopsis in toads. Nature, 267, 349–351.
Cott, H.B. 1936. The effectiveness of protective adaptations in the
hive-bee, illustrated by experiments on the feeding reactions,
 habit formation and memory of the common toad (Bufo bufo).
Proceedings of the Zoological Society London, 1, 113–133.
Damasio, A.R. 1990. Category-related recognition defects as clue to
the neural substrates of knowledge. Trends in Neurosciences, 13,
95–98.
Eaton, R.C. 2001. The Mauthner cell and other identified neurons of
the brainstem escape network of fish. Progress in Neurobiology,
63, 467–485.
Elbert, T., Pantev, C., Wienbruch, C., Rockstroh, B. & Taub, E. 1995.
Increased cortical representation of the fingers of the left hand in
string players. Science, 270, 305–307.
Ewert, J.-P. 1974. The neural basis of visually guided behavior.
Scientific American, 230(3), 34–42.
Ewert, J.-P. 1976. Neuro-Ethologie. Berlin: Springer, [Engl. edn.
1980, Neuroethology. Introduction to the Neurophysiological
Fundamentals of Behavior. Berlin: Springer. Japanese edn. 1982,
Tokyo: Baifukan. – Chinese edn. 1986, Beijing: Science Press,
Acad. Sinica].
Ewert, J.-P. 1984. Tectal mechanisms that underlie prey-catching
and avoidance behaviors in toads. In: H. Vanegas (ed.),
Comparative Neurology of the Optic Tectum, pp. 247–416.
[Review 1965-1984]. London: Plenum Press.
Ewert, J.-P. 1987. Neuroethology of releasing mechanisms. Behavior
and Brain Sciences, 10, 337–405.
Ewert, J.-P. 1997. Neural correlates of key stimulus and releasing
mechanism: a case study and two concepts. Trends in
Neurosciences, 20, 332–339.
Ewert, J.-P. 2004. Motion perception shapes the visual world of
amphibians. In F.R. Prete (ed.), Complex Worlds from Simpler
Nervous Systems, pp. 117–160. Cambridge, MA: MIT Press.
Ewert, J.-P., Dinges, A.W. & Finkenstädt, T. 1994. Species-universal
stimulus responses, modified through conditioning, reappear
 after telencephalic lesions in toads. Naturwissenschaften, 81,
317–320.
Ewert, J.-P. & Schwippert, W.W. 2006. Modulation of visual
perception and action by forebrain structures and their
interactions in amphibians. In E.D. Levin (ed.), Neurotransmitter
Interactions and Cognitive Function, pp. 99–136. Birkhäuser
Verlag.
Fingerling, S., Ewert, J.-P., Menzel, R. & Pfeiffer, F. 1993. From the
toad to a robot—implementation of neurobiological principles of
object discrimination in neural engineering.
Naturwissenschaften, 80, 321–324.
Finkenstädt, T., Adler, N.T., Allen, T.O., Ebbesson, S.O.E. & Ewert,
J.-P. 1985. Mapping of brain activity in mesencephalic and
diencephalic structures of toads during presentation of visual key
stimuli—a computer assisted analysis of 14C-2DG
autoradiographs. Journal Comparative Physiology A, 156, 433–
445.
Finkenstädt, T. & Ewert, J.-P. 1983. Processing of area dimensions of
visual key stimuli by tectal neurons. Salamandra salamandra.
Journal Comparative Physiology A, 153, 85–98.
Finkenstädt, T. & Ewert, J.-P. 1988. Effects of visual associative
conditioning on behavior and cerebral metabolic activity in toads.
Naturwissenschaften, 75, 95–97.
Hansen, K. 1984. Discrimination and production of disparlure
enantiomers by the gypsy moth and the nun moth. Physiological
Entomology, 9, 9–18.
Harnad, S.E. (ed.). 1987. Categorial perception: The groundwork of
cognition. Cambridge University Press.
Hartline, H.K. 1949. Inhibition of activity of visual receptors by
illuminating nearby retinal areas in the Limulus eye. Federation
Proceedings, 8(1), 69.
Heiligenberg, W. & Rose, G. 1985. Neural correlates of the jamming
avoidance response (JAR) in the weakly electric fish.
 Eigenmannia. Trends in Neurosciences, 8, 442–449.
Heiligenberg, W.F., Kramer, U. & Schulz, V. 1972. The angular
orientation of the black eye-bar in Haplochromis burtoni
(Cichlidae, Pisces) and its relevance to aggressivity. Zeitschrift
vergleichende Physiologie, 76, 168–176.
Hinde, R.A. 1954. Changes in responsiveness to a constant stimulus.
Behaviour, 2, 41–54.
Hogan, J.A. 2017. The study of behavior. Cambridge, UK: Cambridge
University Press.
Hubel, D.H. & Livingstone, M.S. 1987. Segregation of form, colour,
and stereopsis in primate area 18. Journal of Neuroscience, 7,
378–415.
Hubel, D.H. & Wiesel, T.N. 1977. Functional architecture of macaque
monkey visual cortex. Proceedings of the Royal Society of London
B, 198, 1–59.
Ishai, A. & Sagi, D. 1995. Common mechanisms of visual imagery
and perception. Science, 268, 1719–1720.
Islam, R., Prater, C.M., Harris, B.N. & Carr, J.A. 2019.
Neuroendocrine modulation of predator avoidance/prey capture
tradeoffs: Role of tectal NPY2R receptors. Comparative
Endocrinology, 282, 113214. doi: 10.1016/j.ygcen.2019.113214.
Kaas, J.H. 1991. Plasticity of sensory and motor maps in adult
mammals. Annual Review of Neuroscience, 14, 137–167.
Kafka, W.A. 1970. Molekulare Wechselwirkungen bei der Erregung
einzelner Riechzellen. Zeitschrift Vergleichende Physiologie, 70,
105–143.
Kaissling, K.-E. 2014. Pheromone reception in insects. The example
of silk moths. In C. Mucignat-Caretta (ed.), Neurobiology of
chemical communication, pp. 99–146. Taylor & Francis: CRC
Press.
Kendrick, K.M. 1994. Neurobiological correlates of visual and
olfactory recognition in sheep. Behavioural Processes, 33, 89–
112.
Koffka, K. 1922. Perception: An introduction to the Gestalt-theorie.
Psychological Bulletin, 19, 531–585.
Konishi, M. 2003. Coding of auditory space. Annual Review of
Neuroscience, 26, 31–55.
Kupfermann, I. & Weiss, K.R. 1978. The command neuron concept.
Behavior and Brain Sciences, 1, 3–39.
Kutschera, U., Burghagen, H. & Ewert, J.-P. 2008. Prey-catching
behaviour in mudskippers and toads: A comparative analysis.
OnLine Journal Biological Sciences, 8(2), 41–43.
Labhart, T. & Meyer, E.P. 2002. Neural mechanisms in insect
navigation: polarization compass and odometer. Current Opinion
in Neurobiology, 12, 707-714.
Langley, C.M., Riley, D.A., Bond, A.B. & Goel, N. 1996. Visual search
and natural grains in pigeons (Columba livia): search images and
selective attention. Journal of Experimental Psychology: Animal
Behavior Processes, 22, 139–151.
Lara, R., Cervantes, F. & Arbib, M. 1982. Two-dimensional model of
retinal-tectal-pretectal interactions for the control of prey-
predator recognition and size preference in amphibia. In: S.
Amari & M.A. Arbib (eds.), Lecture notes in biomathematics 45.
Competition and cooperation in neural nets, pp. 371–393. Berlin:
Springer-Verlag.
Leong, C.Y. 1969. The quantitative effect of releasers on the attack
readiness of the fish Haplochromis burtoni (Cichlidae: Pisces).
Zeitschrift vergleichende Physiologie, 65, 29–50.
Lettvin, J.Y., Maturana, H.R., McCulloch, W.S. & Pitts, W.H. 1959.
What the frog’s eye tells the frog’s brain. Proceedings of the
Institute for Radio Engineering, 47, 1940–1951.
Macaluso, E., Frith, C.D. & Driver, J. 2000. Modulation of human
visual cortex by crossmodal spatial attention. Science, 289, 1206–
1208.
Markl, H. & Hölldobler, B. 1978. Recruitment and food-retrieving
behavior in Novomessor (Formicidae, Hymenoptera). Behavioral
Ecology and Sociobiology, 4, 183–216.
Mountcastle, V.B., Lynch, J.C., Georgopoulos, A., Sakata, H. &
Acuna, C. 1975. Posterior parietal association cortex of the
monkey: command functions operations within extrapersonal
space. Journal of Neurophysiology, 38, 871–908.
Narins, P.M. 1981. Responses of torus semicircularis cells of the
coqui treefrog to fm sinusoids. In J.-P. Ewert, R. R. Capranica &
D. J. Ingle (eds.), Advances in vertebrate neuroethology, pp.
889–894. New York: Plenum Press.
Newman, E.A. & Hartline, P.H. 1982. The infrared “vision” of snakes.
Scientific American, 246(3), 116–127.
Pascalis, O., De Haan, M. & Nelson, C. A. 2002. Is face processing
species-specific during the first year of life? Science, 296, 1321–
1323.
Perret, D.I. & Rolls, E.T. 1983. Neural mechanisms underlying the
visual analysis of faces. In J.-P. Ewert, R. R. Capranica, & D. J.
Ingle (eds.), Advances in vertebrate neuroethology, pp. 543–566.
New York: Plenum Press.
Prater, C.M., Harris, B.N., & Carr, J.A. 2020. Tectal CRFR1 receptor
involvement in avoidance and approach behaviors in the South
African clawed frog, Xenopus laevis. Hormones and Behavior,
120, 104707. doi:10.1016/j.yhbeh.2020.104707.
Prete, F.R. (ed.). 2004. Complex worlds from simpler nervous
systems. Cambridge, MA: MTT Press.
Reddipogu, A., Maxwell, G. & MacLeod, C. 2002. An innovative
neural network based on the toad’s visual system. Proceedings of
ACIVS, Advanced Concepts for Intelligent Vision Systems, Ghent,
Belgium.
Rowland, W.J. 1989. Mate choice and the supemormality effect in
female sticklebacks (Gaster-osteus aculeatus). Behavioral
Ecology and Sociobiology, 24, 433–438.
Ryan, M.J. & Rand, A.S. 1995. Female responses to ancestral
advertisement calls in the tungara frog. Science, 269, 390–392.
Schleidt, W. 1961. Über die Auslösung der Flucht vor Raubvögeln bei
Truthühnern. Naturwissenschaften, 48, 141–142.
Schürg-Pfeiffer, E., Spreckelsen, C. & Ewert, J.-P. 1993. Temporal
discharge patterns of tectal and medullary neurons chronically
recorded during snapping toward prey in toads Bufo spinosus.
Journal Comparative Physiology A, 173, 363–376.
Seitz, A. 1940. Die paarbildung bei einigen cichliden. Zeitschrift
Tierpsychologie, 4, 40–84.
Singer, W. 1995. Development and plasticity of cortical processing
architectures. Science, 270, 758–764.
Spreckelsen, C., Schürg-Pfeiffer, E. & Ewert, J.-P. 1995. Responses of
retinal and tectal neurons in non-paralyzed toads Bufo and B.
marinus to the real size versus angular size of objects moved at
variable distance. Neuroscience Letters, 184, 105–108.
Suga, N. 1990. Biosonar and neural computation in bats. Scientific
American, 262, 60–68.
Ter Polkwijk, J.J. & Tinbergen, N. 1937. Eine reizbiologische Analyse
einigen Verhaltensweisen von Gasterosteus aculeatus. Zeitschrift
für Tierpsychologie, 1, 194–200.
Tinbergen, N. 1951. The Study of Instinct. Oxford: Clarendon Press.
Reissued in 1989 by Oxford University Press, New York.
Tinbergen, N. & Kuenen, D.J. 1939. Über die auslösenden und die
richtunggebenden Reizsituationen der Sperrbewegung von jungen
Drosseln (Turdus m. merula L. und T. e. ericetorum Turton).
Zeitschrift für Tierpsychologie, 3, 37-60.
Udin, S. 1977. Rearrangements of the retinotectal projection in Rana
pipiens after unilateral caudal half-tectum ablation. Journal
Comparative Neurology, 173(3), 561–582.
Ungerleider, L.G. & Mishkin, M. 1982. Two cortical visual systems.
In D.J. Ingle, M.A. Goodale & R.J.W. Mansfield (eds.), Analysis of
visual behavior, pp. 549–586. Cambridge, MA: MIT Press.
Valenza, E., Simion, F., Cassia, V.M. & Umilta, C. 1996. Face
preference at birth. Journal of Experimental Psychology: Human
Perception and Performance, 22, 892–903.
Von Uexkiill, J. 1921. Umwelt und Innenwelt der Tiere. Berlin:
Springer.
Wiersma, C.A.G. & Ikeda, K. 1964. Interneurons commanding
swimmeret movements in the crayfish, Procambarus clarkii
(Girard). Comparative Biochemistry and Physiology, 12, 509–
525.
Wiltschko, W. & Wiltschko, R. 2002. Magnetic compass orientation
in birds and its physiological basis. Naturwissenschaften, 89,
445–452.
Yoshida, N. 2016. From retina to behavior: prey-predator
recognition by convolutional neural networks and their
modulation by classical conditioning. Adaptive Behavior, 1–23.
doi:10.1177/1059712316650265.
Zupanc, G.H. 2019. Behavioral neurobiology. An integrative
approach, 3rd ed. Oxford: Oxford University Press.
3
motivation and emotion
JERRY A. HOGAN

INTRODUCTION
The word motivate means “to cause to move,” and I will use the
concept of motivation to refer to the study of the immediate
causes of behavior: those factors responsible for the initiation,
maintenance, and termination of behavior. Thus, motivation is
another word for aspects of Tinbergen’s causal question (see
Chapter 1). Causal factors for behavior include stimuli, hormones,
and the intrinsic activity of the nervous system. How do these
factors cause a female rat to behave maternally to her pups? Or a
chicken to bathe in dust in the middle of the day? Or a male
stickleback fish to stop responding sexually to receptive females?
These are the types of questions asked in the first part of this
chapter.
Motivated behavior often produces emotion, but the concept of
emotion is problematic because there is no consensus about its
definition. In the second part of this chapter I will analyze the
concept of emotion as applied primarily to humans and conclude
with a section on nonhuman emotion and its relation to animal
welfare.

Behavior Systems

A major problem in the study of both motivation and emotion is that

different authors use these concepts in different ways. The concept of
a behavior system is useful in understanding many of these
differences. I have proposed perceptual, central, and motor
mechanisms as the basic structural units of behavior. These entities
are viewed as corresponding to structures within the central nervous
system. They consist of an arrangement of neurons (not necessarily
localized) that acts independently of other such mechanisms.
Perceptual mechanisms analyze incoming sensory information and
solve the problem of stimulus recognition. An example is the
releasing mechanism discussed in Chapter 2. The motor mechanisms
are responsible for coordinating the neural output to the muscles,
which results in recognizable patterns of movement. The central
mechanisms coordinate the perceptual and motor mechanisms and
also provide the basis for an animal’s mood or internal state. These
units are called behavior mechanisms because their activation
results in an event of behavioral interest: a particular perception, a
specific motor pattern, or an identifiable internal state.
Behavior mechanisms can be connected with one another to form
larger units called behavior systems, which correspond to the level of
complexity indicated by feeding, sexual, and aggressive behavior
(Baerends 1976; Hogan 2001). The organization of the connections
among behavior mechanisms determines the nature of the behavior
system. Thus, a behavior system can be considered a description
of the structure of behavior. A pictorial representation of this concept
is shown in Figure 3.1.
Figure 3.1 Conception of behavior systems. Stimuli from the
external world are analyzed by perceptual mechanisms. Output from
the perceptual mechanisms can be integrated by central mechanisms
and/or channeled directly to motor mechanisms. The output of the
motor mechanisms results in behavior. In this diagram, central
mechanism I, perceptual mechanisms 1, 2, and 3, and motor
mechanisms A, B, and C form one behavior system; central
mechanism II, perceptual mechanisms 3, 4, and 5, and motor
mechanisms C, D,and E form a second behavior system; 1-A, 2-B,
and so on can also be considered less complex behavior systems.
(From Hogan 1988).
Causal factors not only motivate behavior, they can also change the
structure of behavior; that is, they have developmental effects. The
formation of associations and the effects of reinforcement are
developmental processes, and developmental processes have played
an important role in many theories of motivation, especially in
experimental psychology (see Hogan 1998). In this chapter, I will
restrict the term motivation to the modulating effects causal factors
have on the activation of behavior mechanisms. Development refers
to the permanent effects causal factors have on the structure of the
behavior mechanisms and on the connections among the behavior
mechanisms and is discussed in Chapters 7 and 8. Emotion is
considered to be one of the consequences of activating behavior
mechanisms and will be discussed later.

Some Motivational Issues

In addition to problems concerning the definition of the concept of

motivation, there have been four issues that have dominated
discussions of motivation. 1. What role should a concept of energy
play in motivational theories? 2. What is the relative role of internal
versus external causal factors? 3. Do causal factors have specific or
general effects? 4. Is the locus of action of causal factors peripheral
or central? I will briefly discuss each of these issues in this section.

The concept of motivational energy

One attribute of living matter is its activity, the continuous
transformation of energy from one form to another. It was natural,
therefore, when people began to seek explanations of their own
activity, to invoke some concept of energy. And indeed, the earliest
scientific theories of motivation invoked concepts such as instinctual
impulses (James 1890), libido (Freud 1905, 1915), and psycho-
physical energy (McDougall 1923). Within American psychology
these concepts became replaced by the concept of drive, but as
Lashley (1938) pointed out, drives continued to have all the dynamic
properties of the old instinctual urges. A particularly influential
theory of motivation was proposed by Lorenz (1937). The core of his
theory was an energy variable, action-specific energy, discussed
below. For a variety of reasons all these theories were strongly
criticized (Hinde 1960), and energy concepts quickly disappeared
from most accounts of behavior. However, some authors have
pointed out that many of the phenomena that used to be explained
using energy concepts are still not accounted for by other concepts:
they have suggested that an energy concept may still play a useful
theoretical role (e.g., Toates & Jensen 1991; Hogan 1997). We will see
some examples later in the chapter.

External versus internal causal factors

In popular usage, the word motivation often refers only to internal
causes of behavior. We speak of an animal’s search for food as
motivated by hunger, but of chewing and swallowing as reflex actions
to stimuli in the mouth. On close inspection, however, it turns out
that a thoroughly sated animal will often spit out the same food it
would have chewed if it were hungry; and hungry animals are clearly
guided by environmental cues as they search for food. In fact, any
behavior must be caused by some combination of both internal and
external factors.
Many years ago, Lorenz proposed a motivational model of behavior
that illustrates the interdependence of internal and external factors.
This model is shown in Figure 3.2. According to Lorenz, each
behavior pattern (motor mechanism) is associated with a reservoir
that can hold a certain amount of energy. Whenever the behavior
pattern occurs, energy is used up; but when the behavior pattern
does not occur, energy can build up in the reservoir. The higher the
level of energy, the more pressure it exerts on the valve. When the
value opens, as a result of external stimulation, energy is released
and the behavior occurs. Thus, in this model, a particular behavior
pattern cannot occur without at least some internal causal factors as
well as some external ones. Further, the model makes it clear that
internal and external factors can substitute for each other in
determining the intensity of a behavior pattern: a strong stimulus
can compensate for weak internal factors and vice versa.
Figure 3.2 Lorenz’ model of motivation. The tap (T) supplies a
constant flow of endogenous energy to the reservoir (R). The valve
(V) represents the releasing mechanism and the spring (S) the
inhibitory functions of the higher coordinating mechanisms. The
scale pan (Sp) represents the perceptual part of the releasing
mechanism, and the weight applied corresponds to the impinging
stimulation. When the valve is open, energy flows out into the trough
(Tr), which coordinates the pattern of muscle contractions. The
intensity of the response can be read on the gauge (G). (From Lorenz
1950).
The fact that both internal and external factors are essential for any
behavior to occur does not imply, of course, that one cannot study
the effects of internal and external factors separately. The effects of
varying various internal factors can be determined if the external
situation is kept relatively constant, as can the effects of external
factors if the internal state of the animal is held constant. A classic
example of such a study is the one by Baerends et al. (1955) on the
courtship behavior of the male guppy (Lebistes reticulatus).
Courtship by the male comprises a number of behavior patterns,
including posturing in front of the female, a special sigmoid posture,
and copulation attempts. These authors were able to derive a scale of
internal motivation using the relation of marking patterns on the
body of the male to the number of copulation attempts; external
stimulation was considered to be proportional to the size of a female.
Figure 3.3 shows the results of an experiment in which females of
different sizes were presented to males at different levels of internal
motivation. The points plotted on the graph represent the
relationship between the measures of internal and external
stimulation at which particular patterns of behavior were observed.
If it is assumed that the total motivation necessary for a specific
behavior pattern to occur is always the same, the patterns
“posturing,” “sigmoid intention,” and “sigmoid” can be seen to
represent increasing values of courtship strength. The lines
connecting these points of equal motivation have been called
motivational isoclines by McFarland and Houston (1981).
Figure 3.3 Results of an experiment on guppy courtship. (a)
Relationship between the intensity of the external stimulation, the
intensity of the internal stimulation, and the kind and degree of
development of the resulting activity. (b) “Calibration curve” for
determining the place of the different marking patterns on the
abscissa of (a). CA, copulation attempt; S, sigmoid posture; Si,
sigmoid intention; Pf, posturing in front of the female. (From
Baerends et al. 1955).

Specific versus general effects of causal factors

Ever since the psychologist Hull (1943) postulated a general drive, a
recurring question in the study of motivation has been whether
causal factors have general or specific effects. Does a hungry dog
merely eat its food more quickly and accept less preferred foods
more readily, or does it also attack a stranger more fiercely and
copulate more vigorously? There is evidence to support either point
of view, but common sense suggests that some causal factors are
likely to have broad effects whereas others will have only limited
effects. A man who is worried about difficulties at work may show
exaggerated or even inappropriate responses in feeding, aggressive,
and sexual situations. On the other hand, the same man will
probably only drink an extra glass of water if he has lost more body
fluid than usual on a warm dry day.
In general, any particular causal factor will most likely have both
specific and general effects; which effects are more important will
depend on the question of interest. For example, specific effects of
causal factors are implied in Lorenz’s model of motivation. The
model posits that the fluid in the reservoir is specific to the particular
behavior pattern with which it is associated: Lorenz spoke about
action-specific energy . On the other hand, the circadian clock
will be seen to have an important influence on many behavior
systems. I will examine specific and general effects of causal factors
in some detail in the section on displacement activities.

Central versus peripheral locus of action

A fourth pervasive issue in motivation concerns the locus of action of
causal factors. Do causal factors operate within the central nervous
system (CNS) or at a more peripheral level? Once again, common
sense suggests that they must act in both places; nonetheless, this
has also been a controversial issue. Historically, the controversy
arose as a reaction by the early behaviorist school in psychology to
the views of the introspectionists, who thought one could understand
behavior by reflecting on one’s own experiences (see Chapter 1). The
behaviorists were skeptical of internal causes that could not be
investigated directly, and they attempted to explain as much
behavior as possible in terms of stimuli and responses that could be
measured physically. However, as it has become more and more
possible to measure and manipulate events that occur within the
CNS, one major objection to the postulation of central factors has
been removed. Nonetheless, some researchers continue to emphasize
central or peripheral factors.

Causal Factors

Motivation is concerned with the factors that control the activity of

the behavior mechanisms of the individual. These factors are
generally considered to be stimuli, hormones and other substances in
the blood, and the intrinsic activity of the nervous system. Each of
these factors will be briefly discussed.

Stimuli
Stimuli can control behavior in many ways: they can release, direct,
inhibit, and prime behavior. Chapter 2 discussed many examples of
stimuli that release and direct various behavior patterns. Some
stimuli can have exactly the opposite effect: rather than facilitate
behavior, they inhibit it. A good example is provided by the nest-
building behavior of many species of birds. Birds typically build their
nests using specific behavior patterns. The stimuli that release and
direct their behavior have been studied in several cases and conform
to the general principles already discussed. However, at a certain
point the birds stop building and no longer react to the twigs,
lichens, or feathers with which they construct their nest. There are
several possible reasons why they stop, but one reason is that the
stimuli provided by the completed nest inhibit further nest building.
This can be seen when a bird takes over a complete nest from the
previous season and shows very little nest-building behavior. Other
birds, in the same internal state, that have not found an old nest
show a great deal of nest-building behavior (Thorpe 1956).
Another example of the inhibitory effects of stimuli is seen in the
courtship behavior of the three-spined stickleback (Gasterosteus
aculeatus), a small fish. Male sticklebacks set up territories in small
streams early in the spring, build a nest of bits of plant material, and
will generally court any female that may pass through their territory.
Courtship includes a zigzag dance by the male, appropriate posturing
by the female, leading to and showing of the nest entrance by the
male, following and entering the nest by the female, laying eggs, and
finally fertilization (see Figure 3.4). The female swims away and the
male then courts another female. The male could continue courting
egg-laden females for many days, but usually he does not.
Experiments in which eggs were removed from or added to the nest
have shown that visual stimuli from the eggs inhibit sexual activity: if
eggs are removed from the nest, the male will continue courting
females, but if eggs are added to the nest he will cease courting,
regardless of the number of eggs he has fertilized (Sevenster-Bol
1962). This is an especially interesting example because the same
visual stimulus that inhibits sexual activity has an activating effect on
the parental behavior (fanning the eggs) of the same male.
Figure 3.4 Courtship and mating behavior of the three-spined
stickleback. The male is on the left and the female, with a swollen
belly, is on the right. A typical courtship sequence is indicated below
the diagram. (From Tinbergen 1951).
Stimuli not only control behavior by their presence, but in many
cases continue to affect behavior even after they have physically
disappeared. When a stimulus has arousing effects on behavior that
outlast its presence, priming is said to occur. Aggressive behavior
in the male Siamese fighting fish (Betta splendens) provides a good
example (Hogan & Bols 1980). This fish shows vigorous aggressive
display and fighting toward other males of its species (including its
own mirror image). If a fish is allowed to fight with its mirror image
for a few seconds and the mirror is then removed, it is very likely to
attack a thermometer introduced into the aquarium. If the
thermometer had been introduced before the mirror was presented,
the fish very likely would have ignored it. Thus, the sight of a
conspecific not only releases aggressive behavior, it must also change
the internal state of the fish for some time after the conspecific
disappears. We can say that the stimulus primes the mechanism that
coordinates aggressive behavior or, more simply, that it primes
aggression. Similar priming effects have been demonstrated with
food and water in rats and hamsters, and with brain stimulation in
several species (see Hogan & Roper 1978). An especially elegant
mathematical analysis of priming in cichlid fish and crickets is
presented by Heiligenberg (1974).
These examples of priming all occur during the time span of a few
minutes. Some stimuli prime behavior over a much longer period.
Stimuli from the eggs of the stickleback inhibit sexual behavior, as
we have just seen, but they also prime parental behavior. Male
sticklebacks fan the eggs in their nest by moving their fins in a
characteristic manner, which directs a current of water into the nest
and serves to remove debris and provide oxygen to the developing
embryos. The amount of fanning increases over the 7 days it takes for
the eggs to hatch. It has been shown that CO2, which is produced by
the eggs, is one of the stimuli releasing fanning, and the amount of
CO2 produced is greater from older eggs. Thus, one might expect that
the increased fanning is a direct effect of CO2 concentration. This
supposition was tested in an experiment by Van Iersel (1953). He
replaced the old eggs on day 4 with newly laid eggs from another
nest. There was a slight drop in fanning with the new eggs, but
fanning remained much higher than the original day-1 level. Further,
the peak of fanning activity was reached the day the original eggs
would have hatched. This means that the stimuli from the eggs must
prime a coordinating mechanism and that the state of the
coordinating mechanism is no longer completely dependent on
stimulation from the eggs after 3 or 4 days.
A similar example is provided by the development of ovulation in
doves. A female ring dove (Streptopelia risoria) will normally lay an
egg if she is paired with an acceptable male for about seven days. If
the male is removed after 2 or 3 days, the developing egg regresses
and is not laid. However, if the male is allowed to remain with the
female for 5 days before he is removed, the majority of females will
lay an egg 2 days later. Experiments by Lehrman (1965) and his
colleagues demonstrated that it is the stimuli from the courting male
that prime the mechanism responsible for ovulation.
Longer-term effects of stimuli can be seen in the yearly cycle of
gonad growth and regression in some birds and fish as a result of
changes in day length. And changes in day length can also stimulate
a host of other physiological changes including those that prepare
migratory birds for their long-distance flight (e.g., Piersma & Van
Gils 2011) or various mammals for hibernation in the winter (Nelson
2016).

Hormones and other substances

Hormones are substances released by endocrine glands into the
bloodstream; many of them are known to have behavioral effects.
Lashley (1938) suggested that hormones could affect behavior in at
least four different ways: during the development of the nervous
system, by effects on peripheral structures through alteration of their
sensitivity to stimuli, by effects on specific parts of the central
nervous system (central behavior mechanisms), and by nonspecific
central effects. Abundant evidence for all these modes of action has
accumulated since Lashley’s time, although the mechanisms by
which hormones influence behavior have turned out to be more
complex and diverse than early investigators realized (Beach 1948).
Both peripheral and central effects of the hormone prolactin are
seen, for example, in the parental feeding behavior of the ring dove.
Prolactin is responsible for the production of crop “milk” sloughed-
off cells from the lining of the crop that are regurgitated to feed
young squabs. Lehrman (1955) hypothesized that sensory stimuli
from the enlarged crop might induce the parent dove to approach the
squab and regurgitate. His experiments showed that local anesthesia
of the crop region, which removes the sensory input, reduced the
probability that the parents will feed their young. More recent
experiments have confirmed that prolactin has both peripheral and
central effects on the dove’s parental behavior (Buntin 1996).
The maternal behavior of the rat provides an example that illustrates
the variety of hormonal effects. The hormones released at parturition
change the dam’s olfactory sensitivity to pup odors, reduce her fear
of the pups, and facilitate learning about pup characteristics; they
also activate a part of the brain essential for the full expression of
maternal behavior (see Fleming & Blass 1994). More extensive
coverage of the relation between hormones and behavior can be
found in Chapter 6.
Substances released from the neuron terminals into the synapse are
known as transmitters; many of these are known to be involved in
activating specific behavior systems such as feeding and drinking
(see Nelson 2016). Transmitters such as dopamine are thought to
mediate the motivational effects of stimuli for a wide range of
behavior systems, especially their reinforcing effects (Glimcher
2011). Examples of these effects are given in Chapter 6. Psychoactive
drugs, which are thought to exert their effects by altering
neurotransmitter functioning in the brain, are also causal factors for
behavior, but will not be considered further here.

Intrinsic neural factors

In living organisms, the nervous system is continuously active, and
this has many consequences for the occurrence of behavior. Adrian
et al. (1931) was the first to demonstrate spontaneous firing of an
isolated neuron, and Von Holst (1935) showed that such nervous
activity underlay the endogenous patterning of neural
impulses responsible for swimming movements in fish. That
behavior can occur spontaneously, i.e., without any apparent
external cause, was an idea that was long resisted by many
behavioral scientists. As we have seen (Chapter 1), there is a long
history of behavioral scientists fighting against mentalistic concepts
such as consciousness or intentions as causes of behavior. However,
it has gradually become clear that intrinsic causes can be studied
scientifically and that any explanation of behavior that only takes the
effects of external stimuli into account will be incomplete.
One of the earliest attempts to incorporate intrinsic causes into the
scientific study of behavior was made by Skinner (1938). His concept
of the operant is a motor unit of behavior that occurs originally due
to unspecified, intrinsic causes. It is only as a result of conditioning
that the operant comes to be controlled by specific stimuli. The
motivational model of Lorenz was another attempt (Figure 3.2).
Lorenz postulated that motivational energy builds up as a function of
time. His model predicts that the probability that a particular
behavior pattern will occur increases with the time since its last
occurrence. One might imagine that as the pressure in the reservoir
increases, it becomes more and more difficult to prevent the energy
from escaping through the valve. In fact, behavior does sometimes
occur in the absence of any apparent external stimulus. Such
behavior has been called a vacuum activity . Lorenz described the
behavior of a captive starling that performed vacuum insect hunting.
This bird would repeatedly watch, catch, kill, and swallow an
imaginary insect.
Lorenz’s model implies a continuously active nervous system kept in
check by various kinds of inhibition. A particularly striking example
concerns the copulatory behavior of the male praying mantis (Mantis
religiosa). Mantids are solitary insects that sit motionless most of the
time waiting in ambush for passing insects. Movement of an object at
the correct distance and up to the mantis’s own size releases a rapid
strike. Any insect caught will be eaten, even if it is a member of the
same species. This cannibalistic behavior might be expected to
interfere with successful sex, because the male mantis must
necessarily approach the female if copulation is to occur. Sometimes
a female apparently fails to detect an approaching male and he is
able to mount and copulate without mishap, but very often the male
is caught and the female then begins to eat him. Now an amazing
thing happens. While the female is devouring the male’s head, the
rest of his body manages to move round and mount the female, and
successful copulation occurs.
In a series of behavioral and neurophysiological experiments, Roeder
(1967) showed that surgical decapitation of a male, even before
sexual maturity, releases intense sexual behavior patterns. He was
then able to demonstrate that a particular part of the mantis’s brain,
the subesophageal ganglion, normally sends inhibitory impulses to
the neurons responsible for sexual behavior. By surgically isolating
these neurons from all neural input, he showed that the neural
activity responsible for sexual activity is truly endogenous.
A more recent example of endogenous control is dustbathing
behavior in fowl. Most animals possess behavior patterns that can be
used for cleaning themselves or for keeping their muscles, skin, or
feathers in good condition. These patterns range from simply
stretching or rubbing up against some object to complex integrated
sequences of behavior used for grooming in many species, such as
dustbathing in fowl. This behavior comprises a sequence of
coordinated movements of the wings, feet, head, and body of the bird
that serve to spread dust through the feathers. One might suppose
that this behavior is primarily a reaction to dirt or parasites in the
feathers. However, a series of experiments, testing young chicks after
periods of dust deprivation, has provided strong evidence that
dustbathing is primarily endogenously controlled. In one experiment
it was possible to test genetically featherless chicks (Figure 3.5);
these chicks also showed a strong correlation between length of dust
deprivation and amount of dustbathing (Vestergaard et al. 1999).
Figure 3.5 Genetically featherless chicks dustbathing in sand.
Courtesy of Klaus Vestergaard.
The intrinsic factors just discussed are all related to the motivation of
specific behavior patterns. One additional intrinsic factor is the
pacemaker or oscillator cells that are thought to be responsible for
biological clocks (see Chapter 4). These clocks do not control any
specific behavior pattern, but rather modulate the behavior
mechanisms that control many different types of behavior.
Dustbathing in chickens provides an example. A bout of dustbathing
can last for half an hour and usually occurs in the middle of the day
(Vestergaard 1982). Experiments have confirmed that an internal
clock is an important causal factor in the timing of dustbathing
(Hogan & Van Boxel 1993). The timing of human sleep is also an
important example of oscillator control of behavior (Borbély et al.
2001).
There are many other examples of oscillator control of behavior, but
most of the experimental work has investigated the oscillators
responsible for daily (circadian) rhythms, often at a
neurophysiological or genetic level. There has also been considerable
work on the oscillators controlling interval and hourglass timers
(Buhusi & Meck 2005). Timing mechanisms and biological rhythms
are discussed further in Chapter 4.

Interactions among Behavior Systems

Causal factors for many behavior systems are present at the same
time, yet an animal can generally only do one thing at a time. This is
a situation of motivational conflict. In this section I consider the
kinds of behavior that occur in conflict situations, as well as some
mechanisms that have been proposed for switching from one
behavior to another. There have been two major ways of studying
conflict behavior. One way to classify conflicts is in terms of the
direction an organism takes from a goal object: either toward or
away. Many psychologists have distinguished three basic kinds of
motivational conflict, each designated according to the direction
associated with the specific tendencies aroused: approach–approach,
avoidance–avoidance, and approach–avoidance (e.g., Miller 1959).
Other psychologists have used approach/withdrawal concepts as the
basis for a theory of behavioral development (Schneirla 1965). An
alternative way to classify conflict situations, used by ethologists, is
to look at the specific behavior systems that are activated and analyze
the behavior that is actually seen. Four major types of outcome have
been studied: inhibition, ambivalence, redirection, and
displacement. I will briefly discuss each.

Inhibition and Intention Movements

The most common outcome in a conflict situation is that the
behavior system with the highest level of causal factors will be
expressed and all the other systems will be suppressed. A male
stickleback that is foraging in its territory will stop foraging when a
female enters and will begin courting. The male’s hunger has not
changed, nor has the availability of food. It follows that the activation
of the systems responsible for courtship must have inhibited the
feeding system. In general, behavior system inhibition can be said
to occur when causal factors are present that are normally sufficient
to elicit a certain kind of behavior, but the behavior does not appear
as a result of the presence of causal factors for another kind of
behavior.
Sometimes, inhibition of a behavior system is not complete, and
incipient movements belonging to the suppressed behavior systems
are seen. These provide an indication of the relative strength of the
causal factors for other behaviors that are activated in the situation.
They have been called intention movements because they suggest
to an observer, human or conspecific, what behavior might occur
next. Intention movements have played an important role in theories
of the evolution of motor mechanisms (Tinbergen 1952).

Ambivalence
When a female stickleback enters the territory of a male, she is both
an intruder and a potential sex partner. The appropriate response to
an intruding conspecific is to attack it; the appropriate response to a
sex partner is to lead it to the nest. The male essentially does both; he
performs a zigzag dance (see Figure 3.4). He makes a sideways leap
followed by a jump in the direction of the female, and this sequence
may be repeated many times. Sometimes the sideways leap continues
into leading to the nest, and sometimes the jump toward the female
ends in attack and biting. Thus, the zigzag dance can be considered a
case of successive ambivalence. Ambivalent behavior is behavior that
includes motor components belonging to two different behavior
systems; in successive ambivalence , these components occur in
rapid succession.
A somewhat similar case is provided by the “upright” posture of the
herring gull (Larus argentatus; Figure 3.6). This display often occurs
during boundary disputes when two neighboring gulls meet at their
mutual territory boundary. The bird’s neck is stretched and its bill
points down; the carpal joints (wrists) of the wings are raised out of
the supporting feathers; the plumage is sleeked. The position of the
bill and wings are characteristic of a bird that is about to attack
(fighting in this species includes pecking and wing beating the
opponent), and the stretched neck and sleeked plumage are
characteristic of a frightened bird that is about to flee. Further, actual
fighting or fleeing often follows the upright posture. Thus, the
upright posture is a behavior pattern that includes motor
components belonging to two different behavior systems. Unlike the
zigzag dance of the stickleback, however, these components occur
simultaneously. The upright posture can be considered a case of
simultaneous ambivalence . Figure 3.6 also shows that the
upright posture can occur in varying forms. In the “aggressive
upright,” components of attack predominate, whereas in the “anxiety
upright,” components of fleeing predominate.

Figure 3.6 Upright postures of the herring gull: (a) “aggressive”

upright; (b) “intimidated” upright; (c) “anxiety” or “escape” upright.
(From Tinbergen 1959).
The simultaneous occurrence of components belonging to different
behavior systems greatly increases the number and variety of
behavior patterns in a species’ repertoire. A technique called
motivation analysis can be used to explore such ambivalent
behavior patterns, which include many of the bizarre displays
exhibited by many species. In a motivation analysis, one looks at the
form of the behavior, the situation in which it occurs, and other
behaviors that occur in association with it (Tinbergen 1959). An
example is provided by Kruijt’s (1964, p. 61) analysis of “waltzing” by
the male junglefowl (Gallus gallus spadiceus), the wild ancestor of
the domestic chicken (Figure 3.7).

Figure 3.7 “Waltzing” in a male junglefowl. (From Kruijt 1964).

 It is a lateral display: the waltzing bird walks sideways around or
toward the opponent. Back and shoulders are held oblique, the
inner side (the side nearest the opponent) lower than the outer
side. Both wings are lifted out of the supporting feathers; the
upper and lower arms are slightly lowered so that the rump
becomes visible. Otherwise, the inner wing and upper and lower
arm of the outer wing remain folded. The hand of the outer wing
is lowered perpendicularly to the ground and pulled forward, its
plane near the body. The primaries touch the ground and the
outer foot makes scratching or stepping movements through the
primaries. Head and neck are held at the level of the back and
either in the medial plane or slightly turned toward the opponent.
The tail spreads and is turned toward the opponent; breast and
belly feathers are often ruffled, especially those of the other side.
Kruijt noted that the side of the bird’s body near the hen expressed
many components of escape behavior, whereas the side further from
the hen expressed many components of attack behavior. It was “as if
the part of the animal which is nearest to the opponent tries to
withdraw, whereas the other half, which is further away, tries to
approach” (Kruijt 1964, p. 65). He also noted that waltzing was
always directed toward a conspecific. Somewhat surprisingly, young
males directed waltzing equally to males and females, even though
adult males almost always direct it toward females. In about two-
thirds of the cases it was performed immediately before, during, or
immediately after fighting, and in some of these cases behavior
associated with escape was also seen. Thus, on the basis of form,
situation, and associated behavior, Kruijt could conclude that
waltzing is indeed an ambivalent behavior pattern expressing both
attack and escape, with attack predominating. Sexual motivation
appears to be unnecessary.
Motivation analysis of many complex courtship displays in both
birds and mammals has revealed that they are ambivalent activities
very frequently involving primarily the attack and escape systems.
Such activities are usually essential for successful courtship and
reproduction. This means, as mentioned above, that the sex system
by itself is often insufficient for achieving these ends, and illustrates
clearly why causal and functional questions need to be kept separate.
Redirection
When two herring gulls meet at their mutual boundary, causal
factors for both attack and escape behavior are present. As we have
just seen, the birds usually adopt the ambivalent upright posture in
this situation. A common occurrence during this mutual display is
that one of the birds viciously pecks a nearby clump of grass and
then vigorously pulls at it. In form, “grass pulling” resembles the
feather pulling seen during a heated fight between two gulls. This
behavior can be considered a case of redirected behavior because the
motor components all belong to one of the behavior systems for
which causal factors are present (i.e., aggression), but it is directed
toward an inappropriate object. The causal factors for the other
behavior (in this case, escape or fear) must be responsible for the
shift in object. Redirection of aggressive behavior seems to be
especially common in many species including humans (Lorenz
1966).

Displacement
Ambivalent behavior and redirected behavior are appropriate
responses to causal factors that are obviously present in the situation
in which the animal finds itself. Sometimes, however, an animal
shows behavior that is not expected, in that appropriate causal
factors are not apparent. A male stickleback meets its neighbor at the
territory boundary and shows intention movements of attack and
escape; then it suddenly swims to the bottom and takes a mouthful of
sand (which is a component of nest-building behavior). A young
chick encounters a wriggling mealworm and shows intention
movements of approach to peck and eat the mealworm and of
retreating from the novel object; then, while watching the mealworm
the chick falls asleep. A pigeon, actively engaged in courtship,
suddenly stops and preens itself. A student studying hard for an
exam, puts down her book, walks to the kitchen, and makes herself a
sandwich. These behaviors are all examples of displacement
activities that are controlled by a behavior system different from
the behavior systems one might expect to be activated in a particular
situation.
In the case of the stickleback, it is reasonable to show components of
attack and escape behavior at the boundary of its territory because
the neighboring fish is an intruder when it crosses into our subject’s
territory, and our subject loses the security of home when it ventures
into its neighbor’s territory. But why should it engage in nest-
building behavior? The stickleback has probably already built its nest
elsewhere and, in any case, would not normally build it at the edge of
its territory. What are the causal factors for nest building in this
situation? Similar considerations apply to the other examples as well.
In all cases, causal factors for the displacement activity appear to be
missing. It is this apparent inexplicableness of displacement
activities that has caused so much attention to be focused on them.
Why does this unexpected behavior occur?
There have been two main theories put forward to account for
displacement activities: the overflow theory and the disinhibition
theory. The original theory was proposed independently by Kortlandt
(1940) and by Tinbergen (1940) and is usually called the overflow
theory. They proposed that when causal factors for a particular
behavior system (e.g., aggression) were strong, but appropriate
behavior was prevented from occurring, the energy from the
activated system would “spark” or flow over to a behavior system
that was not blocked (e.g., nest building) and a displacement activity
would be seen. The appropriate behavior might be prevented from
occurring because of interference from an antagonistic behavior
system (e.g., fear or escape) or the absence of a suitable object or
thwarting of any sort.
This theory was formulated in the framework of Lorenz’s model of
motivation, which accounts for the graphic metaphor of energy
sparking over or overflowing. In more prosaic terms, this is actually a
theory in which causal factors have general as well as specific effects.
Many examples of displacement activities are described as being
incomplete or hurried—the stickleback does not calmly proceed to
build a nest during a boundary conflict—and such observations give
support to a theory that posits general effects of causal factors. It can
be noted that Freud’s (1940/1949) theory of displacement and
sublimation of sexual energy (libido) is basically the same as the
overflow theory: sexual energy is expressed in nonsexual activities
such as creating works of art.
The alternative theory is called the disinhibition theory. In
essence, it states that a strongly activated behavior system normally
inhibits weakly activated systems. If, however, two behavior systems
are strongly activated (e.g., sex and aggression), the inhibition they
exert on each other will result in a release of inhibition on other
behavior systems (e.g., parental) and a displacement activity will
occur. The general idea was proposed by several scientists, but the
most detailed exploration of the theory was made by Sevenster
(1961). He studied displacement fanning in the male stickleback,
which often occurs during courtship before there are any eggs in the
nest. The sex and aggression behavior systems are known to be
strongly activated during courtship. By careful measurements, it was
possible to show that fanning occurred at a particular level of sex and
aggression when their mutual inhibition was the strongest. Of special
importance for the disinhibition theory, the amount of displacement
fanning that occurred depended on the strength of causal factors for
the parental behavior system. When extra CO2 was introduced into
the water, there was an increase in fanning.
The primary difference between the two theories is that according to
the disinhibition theory the displacement activity is motivated by it
own normal causal factors and the conflict between systems merely
serves a permissive role; whereas according to the overflow theory
the displacement activity is motivated by causal factors for one or
both of the conflicting systems. In the disinhibition theory, causal
factors always have specific effects; in the overflow theory, they have
general effects. Which theory is correct? As is so often the case,
neither theory, by itself, is able to account for all the phenomena
associated with displacement activities. The disinhibition theory is in
many ways more satisfying because it only requires that causal
factors have their normal and expected effects on behavior.
Nonetheless, more general effects of causal factors must be invoked
to account for the frantic or excited aspects of displacement activities
seen in many situations.
It is frequently true that the causation of a behavior pattern is even
more complicated. For example, ground pecking occurs as a
displacement activity during aggressive encounters between two
male junglefowl. Arguments for considering this activity as a
displaced feeding movement include the fact that it is often directed
to food pieces on the ground and the fact that it occurs more
frequently when the animals are hungry. This same activity can also
be considered redirected aggression, and experimental evidence also
supports this interpretation. Thus, one behavior pattern can be both
a displacement activity and a redirected activity at the same time
(Feekes 1972). Each contribution to the causation of a behavior
pattern can be analyzed separately, but the list of causal factors
affecting the behavior pattern can be very long. Indeed, multiple
causation of behavior is the rule rather than the exception. The
causation of behavior is a very complex question, and it is
unreasonable to expect a simple answer.

Mechanisms of Behavioral Change

What determines when a particular behavior will occur, how long it

will continue, and what behavior will follow it? One can imagine that
all an animal’s behavior systems are competing with each other for
expression, perhaps in a kind of free-for-all. For example, if the level
of causal factors for eating is very high, the hunger system will
inhibit other systems and the animal will eat. As it eats, the causal
factors for eating will decline while the causal factors for other
behaviors, say drinking, will become higher than those for eating and
the animal will change its behavior. If a predator approaches, the
escape system will be strongly activated, which will inhibit eating and
drinking, and the animal will run away. And so on.
Unfortunately, as attractive as this account appears, it is clearly an
oversimplification of reality. Perhaps its most serious shortcoming is
that if there were a real free-for-all and only the most dominant
behavior system could be expressed, many essential but generally
low-priority activities might never occur. If a hungry animal never
stopped to look around for danger before the predator was upon it, it
would not long survive. Since most animals do survive, this must
imply that the rules for behavioral change are more complex than the
“winner take all” model. Lorenz (1966) has compared the
interactions among behavior systems to the working of a parliament
that, though generally democratic, has evolved special rules and
procedures to produce at least tolerable and practicable
compromises between different interests. The special rules that
apply to interactions among behavior systems have only begun to be
studied, but a few principles are beginning to emerge.
One important mechanism for behavioral change arises from the fact
that most behavior systems are organized in such a way that “pauses”
occur after the animal has engaged in a particular activity for a
certain time. The level of causal factors for the activity may remain
very high, but during the pause other activities can occur. For
example, in many species, feeding occurs in discrete bouts; between
bouts there is an opportunity for the animal to groom, look around,
drink, and so on. It appears that the dominant behavior system (in
this case, the hunger system) releases its inhibition on other systems
for a certain length of time. During the period of disinhibition, other
behavior systems may compete for dominance according to their
level of causal factors or each system may, so to speak, be given a
turn to express itself. McFarland (1974) has compared these kinds of
interactions among behavior systems with the “time-sharing” that
occurs when multiple users share the same computer system.
A striking example of this sort of behavioral organization is the
incubation system of certain species of birds. Broody hens sit on
their eggs for about 3 weeks. Once or twice a day, the hen gets off the
eggs for about 10 minutes. During this interval she eats, drinks,
grooms, and defecates. The proportion of the 10 minutes spent
eating will vary depending on the state of her hunger system, but
even 24 hours of food deprivation does not change the pattern of
leaving the eggs (Sherry et al. 1980).
Another type of mechanism for behavior change depends upon the
reaction of an animal to discrepant feedback. A male Siamese
fighting fish, for example, will not display as long to its mirror image
as to another displaying male. This is because the behavior of the
mirror image is always identical to the behavior of the subject, but
identical responses are not part of the “species expectation” of
responses to aggressive display (Bols 1977). These mechanisms, and
undoubtedly many others, all interact to produce the infinite variety
of sequences of behavior characteristic of the animal in its natural
environment.
 Human Emotion

Emotion is one of the consequences of activating various behavior

mechanisms. But specifying the concept in more detail is
problematic because there is no consensus about its definition.
Books have been written on the subject (e.g., Ekman & Davidson
1994; Barrett & Russell 2015), and two major journals, Emotion and
Emotion Review, publish studies on emotion. The term usually
refers to certain subjective experiences called feelings, but
observable features often accompany them. I have defined a feeling
as the activation of a specific central behavior mechanism (Hogan
2015), but characterizing those behavior mechanisms that subserve
emotion is as intractable as the concept of emotion itself. I think
some insight into the problem can be gained by considering some of
the views of William James (1890, vol. 2). Note that in reading these
quotes from James, when he uses the term instinct, I would replace
it with the term “behavior system.”
In speaking of the instincts it has been impossible to keep them
separate from the emotional excitements which go with them.
Objects of rage, love, fear, etc., not only prompt a man to outward
deeds, but provoke characteristic alterations in his attitude and
visage, and affect his breathing, circulation, and other organic
functions in specific ways. When the outward deeds are inhibited,
these latter emotional expressions still remain, and we read the
anger in the face, though the blow may not be struck, and the fear
betrays itself in voice and color, though one may suppress all
other sign. Instinctive reactions and emotional expressions thus
shade imperceptibly into each other. Every object that excites an
instinct excites an emotion as well. Emotions, however, fall short
of instincts, in that the emotional reaction usually terminates in
the subject’s own body, whilst the instinctive reaction is apt to go
farther and enter into practical relations with the exciting object.
(p. 442)
He goes on to state his famous theory of emotion.
 Our natural way of thinking about these coarser emotions [grief,
fear, rage, love] is that the mental perception of some fact excites
the mental affection called the emotion, and that this latter state
of mind gives rise to the bodily expression. My theory, on the
contrary, is that the bodily changes follow directly the perception
of the exciting fact, and that our feeling of the same changes as
they occur is the emotion.(p. 449)
In other words, when we meet a bear and run away, we are not
running away because we are frightened; we are frightened because
we are running away—fright is our perception of all the bodily
changes that occur when we run away. This theory has been subject
to much criticism to which I will return shortly. But first I will quote
some more of James’ views.
Now the moment the genesis of an emotion is accounted for, as
the arousal by an object of a lot of reflex acts which are forthwith
felt, we immediately see why there is no limit to the number of
possible different emotions which may exist, and why the
emotions of different individuals may vary indefinitely, both as
to their constitution and as to objects which call them forth….
such a question as “What is the ‘real’ or ‘typical’ expression of
anger, or fear?” is seen to have no objective meaning at all.
Instead of it we now have the question as to how any given
‘expression’ of anger or fear may have come to exist; and that is a
real question of physiological mechanics on the one hand, and of
history [evolution] on the other…(p. 454)
I have quoted from James so extensively because the ideas he
expressed have continued to dominate studies of emotion from his
time to the present. Historically, the idea that has been most
discussed is his idea that emotion is the perception of the feedback
one gets from bodily changes in response to some arousing situation.
Everyone agrees that most emotional situations cause a multitude of
visceral changes such as increases in heart rate, vasoconstriction,
sweating, etc., all of which are caused by sympathetic nervous system
action. However, Cannon (1927) challenged the notion that
perception of these changes is the emotion for several reasons. He
pointed out that the viscera are relatively insensitive structures and
that visceral changes are too slow to account for emotional feelings
that occur demonstrably quicker. He also cited results of Marañon,
who injected epinephrine, a sympathetic nervous system stimulant,
into human subjects and asked them to describe their feelings. The
results showed a clear distinction “between the perception of the
peripheral phenomena of vegetative emotion (i.e., the bodily
changes) and the psychical emotion proper, which does not exist and
which permits the subjects to report on the vegetative syndrome with
serenity, without true feeling” (tr. by Cannon 1927, p. 113).
Cannon also pointed out that the same visceral changes occur in very
different emotional states as well as in nonemotional states. What
distinguishes the different emotional states from each other? What is
responsible for the “feeling” of each emotional state? James was
aware of this problem, as we have seen, because he showed that there
is no typical expression of each emotion. But James does list four
“coarser” emotions and discusses a large number of “subtler”
emotions, so presumably he thought that we can introspectively
distinguish among them, perhaps on the basis of the nonorganic
emotional expressions such as facial features or postures. Schachter
and Singer (1962) proposed that the quality of an emotion is arrived
at by a process of cognitive appraisal. In one of their experiments
they injected epinephrine into human subjects. Some of the subjects
were told what effects they might expect and others were not told
anything about the effects. All subjects were then observed in the
presence of a confederate of the experimenter who acted in either an
elated or angry manner. The subjects were later asked about their
emotional reactions. The informed subjects reported very little
emotion (as also the subjects of Marañon); but the uninformed
subjects did report emotional feelings, and the kind of emotion they
felt tended to mimic that of the confederate. In other words, in
precisely the same state of physiological arousal, emotional labels
depended on the cognitive aspects of the situation.
Since 1962, there have been hundreds of studies on emotion.
Gendron and Barrett (2009) review the history of scientific ideas
about emotion and posit three major approaches to its study: basic
emotion, appraisal, and psychological construction. The basic
emotion approach has been a major focus of studies in “affective
science,” as studies of emotion have come to be called. The goal of
researchers in this area is to discover and characterize the basic
emotions, which are considered to be inherent in our biological
endowment. However, in spite of years of research, there is still no
consensus about the identity of the basic emotions. Tracy and
Randles (2011) recently reviewed four models of basic emotions
proposed by four prominent researchers in the area (Ekman, Izard,
Levenson, Panksepp). Their lists of basic emotions are somewhat
similar (fear is included in all four lists, and sadness, anger, and
disgust are included in three of the four lists), but there are still
many differences and many problems of definition of terms remain.
Panksepp’s (2005) list is the most divergent from the other three,
which probably reflects the fact that he bases his list on his analysis
of “the neurodynamics of brain systems that generate instinctual
emotional behaviors” in various mammalian species, whereas the
others base their lists on experimental studies of human subjects.
The appraisal approach assumes “that emotions are not merely
triggered by objects in a reflexive or habitual way, but arise from a
meaningful interpretation of an object by an individual” (p. 317). It
considers the identification of emotional quality (its meaning) to
require appraisal of the object and situation; it is the meaning that
then leads to internal state changes: we are afraid of the bear when
we see it (we appraise the situation) and then we become aroused
and flee. Although Schachter & Singer use the word “appraisal” in
their theory, Gendron & Barrett consider their theory closer in
content to the psychological construction approach.
The psychological construction approach posits that emotions
are constructed out of more basic psychological ingredients that are
not themselves specific to emotion. Two such basic components were
proposed by Russell (2003): core affect and affective quality. “Core
affect is that neurophysiological state consciously accessible as the
simplest raw (nonreflective) feelings evident in moods and
emotions” (p. 148). It is a single integral blend of two dimensions:
pleasure-displeasure (which can range from elation to agony) and
activation-deactivation (which can range from frenetic excitement to
sleep). The feeling is an assessment of one’s current condition.
Affective quality is a property of the stimulus: its capacity to change
core affect. Perception of affective quality together with core affect
allows a person to construct the emotion.
Barrett (2013) believes that psychological construction constitutes a
paradigm for the scientific study of emotion that is different from the
“faculty” psychology paradigm of the basic emotion and appraisal
approaches. She points to three principles of psychological
construction that define this difference: the principles of variation, of
core systems, and of emergentism and holism. I will not discuss
these principles here, but I will say that in many respects the details
of these principles bear a striking resemblance to many of the ideas
originally expressed by William James (and acknowledged by Barrett
and Russell). Whether these ideas will actually change the way
researchers on emotional issues behave, remains to be seen. Barrett
also calls psychological construction the Darwinian approach to the
science of emotion, primarily because pre-Darwin, species were
considered fixed whereas post-Darwin, the variety within species
could be exploited by natural selection and lead to new kinds.
Although there are some similarities between the two approaches, I
doubt that most biologists would be impressed with the analogy.
In another recent development, LeDoux (2012) has proposed
rethinking the emotional brain in terms of survival circuits. The
survival circuits proposed by LeDoux correspond almost exactly to
behavior systems as I have defined them, although he places more
constraints on which brain circuits would be considered survival
circuits. His survival circuits are considered to be “innate” and to
have functional [survival] significance for the organism. There are
many problems with the concept of “innate” (see Chapter 7) and
functional explanations cannot solve causal problems (Hogan 2015).
LeDoux also notes that his list of survival circuits does not align well
with human basic emotions.
In considering these various approaches to the study of emotion, I
would propose that it is the activated behavior system that
determines the quality of the emotion. The study of emotion then
becomes the study of what behavior systems exist in any organism,
what motivational factors activate them, and how they are expressed.
Emotions are the subjective aspect of strongly activated behavior
systems. A corollary of this conceptualization is that the felt emotion
becomes an epiphenomenon: like the whistle of the steam engine, it
has no causal significance—which is, of course, consonant with
James’ viewpoint. Much of the research on emotion in the past 50
years can be understood in these terms.

Nonhuman Emotion

I have defined emotion as the subjective aspect (feeling) of strongly

activated behavior systems. Since we have no access to the subjective
experience of any animal (except ourselves), any discussion of
nonhuman emotion must rely on investigation of the expression of
such strongly activated behavior systems. One of the first systematic
studies of the expression of the emotions in man and (other) animals
was that of Darwin (1872). Darwin was primarily interested in
similarities between animal expression of presumed emotional states
such as anger, terror, and joy and human expression of these and
other emotions. Darwin assumed that animals such as dogs, cats,
horses, and monkeys had such emotional states and tried to show
that the expression of these emotions in humans could be traced to
their expression in various animals as support for his theory of
evolution. Since we know that the nervous systems of all animals
have similar components, it should be possible to infer the emotional
state of an animal from observations of its behavior. We would be
inferring the state of activation of an animal’s various behavior
systems, irrespective of whatever subjective experience the animal
might be having. In effect, we would be performing a motivation
analysis (p. 62). We have already seen examples of this with respect
to the upright posture of the herring gull, the zig-zag dance of the
stickleback, and waltzing in junglefowl. A similar example is Lorenz’s
(1966) analysis of the facial expressions of fear and aggression in
dogs (Figure 3.8). In this figure, increasing aggression goes from left
to right and increasing fear goes from top to bottom. In (a) the dog is
calm and unemotional; in (b) and (c) it is becoming more afraid; in
(d) and (g) it is becoming more aggressive. The other figures depict
ambivalent expressions. It can be seen that as fear is increasing, the
ears and the corners of the mouth are drawn backward and
downward; as aggression is increasing, the upper lip is raised and the
mouth opened. These examples show that it is possible to ascertain
which behavior systems in an animal are activated. But how strongly
does the system have to be activated in order to be considered an
emotion? And how do we measure strength?

Figure 3.8 Facial expressions of fear and aggression in dogs.

Explanation in text. (From Lorenz 1966).
Strength has been measured both behaviorally and physiologically,
and recently, “cognitively” as well. An early investigator of
“emotionality” in animals (the rat in this case) was Hall (1934). He
showed that defecation and urination in a standard situation were
valid measures of individual differences in emotionality. Hall
considered emotionality a trait, characteristic of an individual. He
felt that attempts to differentiate specific emotions were extremely
speculative, a view still held by many investigators of human
emotion, as we have seen above. Hall, as also most prior and
subsequent investigators of animal emotion, was really interested in
using animal studies as a model for understanding human emotion.
And soon thereafter, many other measures of bodily changes in
animals, both behavioral and physiological, began to be used in
investigations of various aspects of emotion. Paul et al. (2005) review
the various approaches to measuring emotional processes in
animals, past and present, including new non-linguistic cognitive
measures.
Most studies of animal emotion are directed to understanding
human emotion, but the rise of interest in animal welfare has led
many investigators to study animal emotion per se. In the context of
welfare, it is crucial to discover what makes an animal ‘feel good’ (or,
at least, not suffer). However, feelings are subjective and we can
never know what an animal feels (see Panksepp 2010 and Dawkins
2015 for recent discussions of animal consciousness). M. Dawkins
(2008) suggests that a scientific study of animal suffering and
welfare can be based on answers to two questions: Will the situation
improve animal health? And, will it give animals something they
want? The answer to the second question can be determined by
discovering what the animal finds positively and negatively
reinforcing (what they want and do not want) in a learning situation.
Even here, however, a difference between “wanting” and “liking”
(Berridge 2004) makes interpretation of the results not
straightforward (an animal may like something, but not want it at
this moment). Nonetheless, Dawkins’ approach seems the most
reasonable proposal to date. Theoretically, Mendl et al. (2010) have
proposed a framework that integrates the discrete emotion approach
(i.e., the basic emotion approach above) with the dimensional
approach (i.e., the psychological construction approach above) for
the study of animal emotion and mood. The cognitive aspects of the
dimensional approach allow one to experimentally dissociate “liking”
something from currently “wanting” it, which solves some problems.
In all cases, however, the feelings of the animal remain a conjecture.
The topic of animal welfare is considered again in Chapter 10.
 SUMMARY AND CONCLUSIONS
Motivation refers to the immediate causes of behavior, that is, to
Tinbergen’s causal question. All behavior is caused by the action
of a combination of internal and external causal factors, some of
which have very specific effects on behavior and others more
general effects. Stimuli can release, direct, inhibit, and prime
behavior. These effects all depend on the internal state of the
animal, which is controlled by hormones and other substances
and by the intrinsic activity of the nervous system. The
“psychohydraulic” model of behavior proposed by Lorenz
provides a useful analogy for understanding how all these factors
interact with each other. In general, causal factors for more than
one behavior system are present at the same time. Sometimes the
system with the strongest causal factors inhibits all the other
systems, but most of the time animals engage in some type of
ambivalent, redirected, or displacement behavior.
Emotions are the subjective aspect of strongly activated behavior
systems. They can be studied in animals using Tinbergen’s
motivation analysis (p. 62). Operant training techniques can be
used to assess animal welfare.

FURTHER READING
Amplification of many of the ideas in this chapter can be found in
The Study of Behavior (Hogan 2017). An Introduction to Behavioral
Endocrinology (Nelson 2016) provides an excellent source for details
on the physiological control of most of the behavior systems
mentioned in this chapter. Toates’ (1986) book Motivational
Systems presents a review of behavior systems with an emphasis on
control theory variables, and Enquist and Ghirlanda (2005) discuss
behavior systems from a neural network point of view. Lorenz’
(1966) controversial, but entertaining, book On Aggression gives his
views on how ethological concepts can be applied to human
behavior. James’ (1890) chapter on emotion should be read by
anyone interested in the subject.

REFERENCES
Adrian, E.D., Cattell, M. & Hoagland, H. 1931. Sensory discharges in
single cutaneous nerve fibres. Journal of Physiology, 72, 377–391.
Baerends, G.P. 1976. The functional organization of behaviour.
Animal Behaviour, 24, 726–738.
Baerends, G.P., Brouwer, R. & Waterbolk, H.T. 1955. Ethological
studies on Lebistes reticulatus (Peters): 1. An analysis of the male
courtship pattern. Behaviour, 8, 249–334.
Barrett, L.F. 2013. Psychological construction: The Darwinian
approach to the science of emotion. Emotion Review, 5, 379–389.
Barrett, L.F. & Russell, J.A. (eds.). 2015. The Psychological
Construction of Emotion. Guilford Press.
Beach, F.A. 1948. Hormones and Behavior. New York: Hoeber.
Berridge, K.C. 2004. Motivation concepts in behavioral
neuroscience. Physiology and Behavior, 81, 179–209.
Bols, R.J. 1977. Display reinforcement in the Siamese fighting fish,
Betta splendens: Aggression motivation or curiosity? Journal of
Comparative and Physiological Psychology, 91, 233–244.
Borbély, A.A., Dijk, D.-J., Achermann, P.& Tobler, I. 2001. Processes
underlying the regulation of the sleep-wake cycle. In: J.S.
Takahashi, F.W. Turek & R.Y. Moore (eds.), Circadian Clocks:
Handbook of Behavioral Neurobiology, vol. 12, pp. 458–479. New
York: Kluwer Academic/Plenum.
Buhusi, C.V. & Meck, W.H. 2005. What makes us tick? Functional
and neural mechanisms of interval timing. Nature Reviews
Neuroscience, 6, 755–765.
Buntin, J.D. 1996. Neural and hormonal control of parental behavior
in birds. Advances in the Study of Behavior, 25, 161–213.
Cannon, W.B. 1927. The James-Lange theory of emotions: A critical
examination and an alternative theory. American Journal of
Psychology, 39, 106–124.
Darwin, C. 1872. The Expression of the Emotions in Man and
Animals. London.
Dawkins, M.S. 2008. The science of animal suffering. Ethology, 114,
937–945.
Dawkins, M.S. 2015. Animal welfare and the paradox of animal
consciousness. Advances in the Study of Behavior, 47, 5–37.
Ekman, P. & Davidson, R.J. (eds.). 1994. The Nature of Emotion:
Fundamental questions. New York: Oxford University Press.
Enquist, M. & Ghirlanda, S. 2005. Neural Networks and Animal
Behavior. Princeton, NJ: Princeton University Press.
Feekes, F. 1972. “Irrelevant” ground pecking in agonistic situations
in Burmese red junglefowl (Gallus gallus spadiceus). Behaviour,
43, 186–326.
Fleming, A.S. & Blass, E.M. 1994. Psychobiology of the early mother-
young relationship. In: J.A. Hogan & J.J. Bolhuis (eds.), Causal
Mechanisms of Behavioural Development, pp. 212–241.
Cambridge: Cambridge University Press.
Freud, S. 1905. Three Contributions to the Theory of Sexuality.
London: Hogarth Press.
Freud, S. 1915. Instincts and Their Vicissitudes. London: Hogarth
Press.
Freud, S. 1940/1949. An Outline of Psycho-Analysis. New York:
Norton.
Gendron, M. & Barrett, L.F. 2009. Reconstructing the past: a century
of ideas about emotion in psychology. Emotion Review, 1, 316–
339.
Glimcher, P.W. 2011. Understanding dopamine and reinforcement
learning: the dopamine reward prediction error hypothesis.
Proceedings of the National Academy of Sciences, 108, 15647–
15654.
Hall, C.S. 1934. Emotional behavior in the rat: 1. Defecation and
urination as measures of individual differences in emotionality.
Journal of Comparative Psychology, 18, 385–403.
Heiligenberg, W. 1974. Processes governing behavioral states of
readiness. Advances in the Study of Behavior, 5, 173–200.
Hinde, R.A. 1960. Energy models of motivation. Symposia of the
Society for Experimental Biology, 14, 199–213.
Hogan, J.A. 1988. Cause and function in the development of
behavior systems. In: E.M. Blass (ed.), Handbook of Behavioral
Neurobiology, vol. 9, pp. 63–106. New York: Plenum Press.
Hogan, J.A. 1997. Energy models of motivation: A reconsideration.
Applied Animal Behaviour Science, 53, 89–105.
Hogan, J.A. 1998. Motivation. In: G. Greenberg & M.M. Haraway
(eds.), Comparative Psychology: A Handbook, pp. 164–175. New
York: Garland.
Hogan, J.A. 2001. Development of behavior systems. In: E.M. Blass
(ed.), Handbook of Behavioral Neurobiology, vol. 13, pp. 229–
279. New York: Kluwer Academic/Plenum.
Hogan, J.A. 2015. A framework for the study of behavior.
Behavioural Processes, 117, 105–113.
Hogan, J.A. 2017. The Study of Behavior: Organization, Methods,
and Principles. Cambridge, UK: Cambridge University Press.
Hogan, J.A. & Bols, R.J. 1980. Priming of aggressive motivation in
Betta splendens. Animal Behaviour, 28, 135–142.
Hogan, J.A. & Roper, T.J. 1978. A comparison of the properties of
different reinforcers. Advances in the Study of Behavior, 8, 155–
255.
Hogan, J.A. & Van Boxel, F. 1993. Causal factors controlling
dustbathing in Burmese red junglefowl: some results and a model.
Animal Behaviour, 46, 627–635.
Hull, C.L. 1943. Principles of Behavior. New York: Appleton-
Century-Crofts.
James, W. 1890. Principles of Psychology, vol. 2. New York: Holt.
Kortlandt, A. 1940. Wechselwirkung zwischen Instinkten. Archives
Néerlandaises de Zoologie, 4, 443–520.
Kruijt, J.P. 1964. Ontogeny of social behaviour in Burmese red
junglefowl (Gallus gallus spadiceus). Behaviour, Suppl. 9, 1–201.
Lashley, K.S. 1938. Experimental analysis of instinctive behavior.
Psychological Review, 45, 445–471.
LeDoux, J. 2012. Rethinking the emotional brain. Neuron, 73, 653–
676.
Lehrman, D.S. 1955. The physiological basis of parental feeding
behaviour in the ring dove (Streptopelia risoria). Behaviour, 7,
241–286.
Lehrman, D.S. 1965. Interaction between internal and external
environments in the regulation of the reproductive cycle of the
ring dove. In: F.A. Beach (ed.), Sex and Behavior, pp. 355–380.
New York: Wiley.
Lorenz, K. 1937. Über die Bildung des Instinktbegriffes.
Naturwissenschaften, 25, 289-300, 307-318, 324-331.
Lorenz, K. 1950. The comparative method in studying innate
behaviour patterns. Symposia of the Society for Experimental
Biology, 4, 221–268.
Lorenz, K. 1966. On Aggression. London: Methuen.
McDougall, W. 1923. An Outline of Psychology. London: Methuen.
McFarland, D.J. 1974. Time-sharing as a behavioral phenomenon.
Advances in the Study of Behavior, 5, 201–225.
McFarland, D.J. & Houston, A. 1981. Quantitative Ethology: The
State Space Approach. London: Pitman.
Mendl, M., Burman, O.H.P. & Paul, E.S. 2010. An integrative and
functional framework for the study of animal emotion and mood.
Proceedings of the Royal Society B, 277, 2895–2904.
Miller, N.E. 1959. Liberalization of basic S-R concepts: extensions to
conflict behavior, motivation, and social learning. In: S. Koch
(ed.), Psychology: A Study of a Science, vol. 2, pp. 196–292. New
York: McGraw-Hill.
Nelson, R.J. 2016. An Introduction to Behavioral Endocrinology, 4th
ed. Sunderland, MA: Sinauer.
Panksepp, J. 2005. Affective consciousness: Core emotional feelings
in animals and humans. Consciousness and Cognition, 14, 30–80.
Panksepp, J. 2010. Affective consciousness in animals: Perspectives
on dimensional and primary process emotion approaches.
Proceedings of the Royal Society B, 277, 2905–2907.
Paul, E.S., Harding, E.J. & Mendl, M. 2005. Measuring emotional
processes in animals: the utility of a cognitive approach.
Neuroscience and Biobehavioral Reviews, 29, 469–491.
Piersma, T. & Van Gils, J.A. 2011. The Flexible Phenotype. Oxford:
Oxford University Press.
Roeder, K.D. 1967. Nerve Cells and Insect Behavior, Revised ed.
Cambridge, MA: Harvard University Press.
Russell, J.A. 2003. Core affect and the psychological construction of
emotion. Psychological Review, 110, 145–172.
Schachter, S. & Singer, J.E. 1962. Cognitive, social, and physiological
determinants of emotional state. Psychological Review, 69, 379–
399.
Schneirla, T.C. 1965. Aspects of stimulation and organization in
approach/withdrawal processes underlying vertebrate behavioral
development. Advances in the Study of Behavior, 1, 1–74.
Sevenster, P. 1961. A causal analysis of a displacement activity
(fanning in Gasterosteus aculeatus L.). Behaviour, Suppl. 9, 1–
170.
Sevenster-Bol, A.C.A. 1962. On the causation of drive reduction after
a consummatory act. Archive Néerlandaises de Zoologie, 15, 175–
236.
Sherry, D.F., Mrosovsky, N. & Hogan, J.A. 1980. Weight loss and
anorexia during incubation in birds. Journal of Comparative and
Physiological Psychology, 94, 89–98.
Skinner, B.F. 1938. The Behavior of Organisms. New York: Appleton-
Century-Crofts.
Thorpe, W.H. 1956. Learning and Instinct in Animals. London:
Methuen.
Tinbergen, N. 1940. Die Übersprungbewegung. Zeitschrift für
Tierpsychologie, 4, 1–40.
Tinbergen, N. 1951. The Study of Instinct. Oxford: Oxford University
Press.
Tinbergen, N. 1952. Derived activities: Their causation, biological
significance, origin and emancipation during evolution. Quarterly
Review of Biology, 27, 1–32.
Tinbergen, N. 1959. Comparative studies of the behaviour of gulls
(Laridae): a progress report. Behaviour, 15, 1–70.
Toates, F.M. 1986. Motivational Systems. Cambridge, UK:
Cambridge University Press.
Toates, F.M. & Jensen, P. 1991. Ethological and psychological models
of motivation: towards a synthesis. In: J.-A. Meyer & S. Wilson
(eds.), From Animals to Animats, pp. 194–205. Cambridge, MA:
MIT Press.
Tracy, J.L. & Randles, D. 2011. Four models of basic emotions: A
review of Ekman and Cordaro, Izard, Levenson, and Panksepp
and Watt. Emotion Review, 3, 397–405.
Van Iersel, J.J.A. 1953. An analysis of the parental behaviour of the
three-spined stickleback (Gasterosteus aculeatus L.). Behaviour,
Suppl. 3, 1–159.
Vestergaard, K.S. 1982. Dust-bathing in the domestic fowl: diurnal
rhythm and dust deprivation. Applied Animal Ethology, 8, 487–
495.
Vestergaard, K.S. et al. 1999. Regulation of dustbathing in feathered
and featherless domestic chicks: The Lorenzian model revisited.
Animal Behaviour, 58, 1017–1025.
Von Holst, E. 1935. Über den Prozess der zentralnervösen
Koordination. Pflügers Archiv für die gesamte Physiologie, 236,
149–158.
4
biological rhythms and behavior
RALPH E. MISTLBERGER AND BENJAMIN RUSAK

INTRODUCTION

A clockwork chipmunk
The eastern chipmunk Tamias striatus is a solitary terrestrial
squirrel that inhabits forests of eastern North America. It sleeps
at night in complete darkness in an underground burrow. During
the summer months, it emerges daily to collect food and eat. In
the autumn it switches to hoarding prodigious amounts of acorns,
beechnuts, and maple samaras in its burrow to eat during the
winter months, which it spends in its burrow under layers of
snow. If food runs short, the chipmunk becomes torpid to save
energy. By early spring the chipmunk emerges from its den to
seek a mate. The chipmunk’s behavior thus varies predictably
with time of day and season, expressing highly regular cycles in
synchrony with its environment. How are these rhythms
controlled? How does the chipmunk know the correct time to
emerge from its den each day in the summer or at the end of the
winter, given that it rests underground in constant dark and near
constant temperature? These questions of timing are the topic of
this chapter. As we shall see, biological organisms have evolved
endogenous (internal) timing devices that can be used like clocks
to enable them to do the right thing at the right time.
 The Spectrum and Discovery of Biological
Rhythms

Life on earth evolved in environments characterized by regular and

highly predictable changes in levels of light, temperature, humidity,
tidal activity, and many secondary consequences of these cycles.
These changes are caused by the unique configuration and
movements of earth, moon, and sun relative to each other. The spin
of the earth on its own axis gives rise to the solar day (a 24 hour
period), while the rotation of the moon around the earth creates the
lunar day and tidal cycles (24.8 hours). The relative movements of
earth and moon around the sun give rise to the lunar month (~29.5
days) and solar year (~365.25 days). The many changes in local
conditions that result from these physical cycles have affected the
course of biological evolution and provided a temporal framework
for both the physiology and behavior of organisms.
Rhythmicity of form and function that mirrors the rhythmicity of the
external world is evident in most organisms, from cyanobacteria to
humans, and at all levels of biological organization, from the
transcription of genes to the movements of animal populations. Daily
(24 hours) rhythms are especially prominent in virtually all
eukaryotic organisms (those with nucleated cells), but cycles of
behavior and physiology with periodicities of hours (“ultradian”
rhythms, «24 hours), months, or years (“infradian” rhythms, »
24 hours) are also evident in many species. In this chapter,
discussion will be limited primarily to a subset of periodicities known
collectively as the “circa” rhythms. These rhythms share two
characteristics:

1. under natural conditions, they are synchronized to major

geophysical cycles in the environment; and
2. under artificially constant conditions, they persist with a period
that approximates that of the corresponding geophysical cycle.
Four “circa” rhythms are recognized; circatidal (~12.4 hours
ultradian rhythms synchronized to the semidiurnal tidal rhythm);
circadian (from the Latin circa [about] and dies [a day], i.e., ~24
hour rhythms synchronized to the solar day); circalunar (~29.5 day
infradian rhythms synchronized to the lunar month); and
circannual (~365 day infradian rhythms synchronized to the solar
year). Circadian rhythms are by far the best known and most
ubiquitous of the circa rhythms and will therefore be the focus of this
chapter, but the mechanisms and functions of some ultradian and
infradian behavioral rhythms will also be discussed.

Circadian rhythms persist in temporal isolation

The rest-activity cycle is among the most conspicuous of daily
rhythms, and most animals can be classified as predominantly day
active (diurnal), night active (nocturnal), or dawn-and-dusk
active (crepuscular). These temporal patterns could be the result of
a direct effect on behavior of daily variations in light, temperature,
humidity, or some other stimulus. Direct effects of light and dark on
behavior are in fact well known; the activity of diurnal animals is
typically inhibited by darkness while that of nocturnal animals is
inhibited by light (Mrosovsky 1999). However, direct effects of light
and dark on behavior reinforce or modify daily rhythms; they are not
necessary for their expression. Plants and animals maintained in
carefully controlled conditions of constant light, temperature, and
humidity (i.e., in temporal isolation) continue to exhibit daily
rhythms in their biochemistry, physiology, and behavior. In many
organisms these rhythms can persist (free run) virtually indefinitely
in the complete absence of environmental time cues, indicating that
the rhythms are both endogenously generated and self-sustaining
(Figure 4.1). In such constant conditions, free-running rhythms
express a period (denoted by the Greek letter τ, the reciprocal of
which is frequency, e.g., one cycle per day) that usually differs
slightly from 24 hours (which is why they are called “circadian”
rather than daily rhythms). This deviation of τ from 24 hours causes
circadian rhythms in temporal isolation to gradually drift out of
synchrony with local (external) time. Since free-running τ can vary
both within and between individuals (Figure 4.1), the rhythms of
organisms recorded in the same environment may drift out of
synchrony with each other. Circadian rhythms in plants and animals
have been shown to persist even in orbit, completely removed from
the natural cycles on the earth’s surface. These observations
constitute powerful evidence that daily rhythms are the product of an
endogenous, self-sustaining, clock-like mechanism.
Figure 4.1 Rest-activity records from representatives of four
mammalian species recorded continuously for several months or
more. In each record, time within a day is plotted from left to right,
and successive days are aligned top to bottom. The records are
double-plotted, so that each line represents two consecutive days (48
hours). Heavy bars indicate time bins when the animals are active
(Panels A, B, D) or when the subject is asleep (C). (A) Wheel-running
activity of a male mouse (Mus musculus, C57BL/6 strain) in a light–
dark (LD) cycle (lights on is denoted by a heavy yellow bar above day
1 of recording) and then in constant dark (DD). Running is almost
exclusively nocturnal in LD and exhibits a precise 24 hour cycle. In
DD, the daily rhythm of running persists with a circadian periodicity
of ~23.7 hours. This demonstrates that the rest–activity cycle is
endogenously generated and that LD cycles control both the phase
and period of the rhythm (a process defined as entrainment)
(Mistlberger lab archives). (B) Activity rhythm of a female rat
(Rattus norvegicus; Wistar albino strain) in DD, recorded using a
motion sensor, and then with a running wheel. The periodicity of the
activity rhythm is >24 hours without a running wheel (rhythm drifts
to the right, i.e., later each day) and <24 hours with a running wheel
(rhythm drifts to the left, i.e., earlier each day). This illustrates how
the animal’s own behavior (and the method of measuring activity)
can modify a parameter of the circadian clock (its periodicity). The
wheel running activity also shows an “infradian” cycle of 4–5 days in
the amount of running, due to variations in estrogen and
progesterone across the estrous cycle. (Adapted from Yamada et al.
1988). (C) Sleep (heavy bars) and core body temperature (grey bars;
representing lower temperatures that normally coincide with sleep)
rhythms of a male human subject maintained in a laboratory setting
first under a LD cycle and then in temporal isolation (no time-of-day
cues, self-selected sleep, and lighting). This example illustrates how
the sleep–wake cycle in humans can shift relative to the body
temperature cycle and even become completely dissociated from it
during occasional very long spontaneous sleep and wake bouts.
(Adapted from Czeisler et al. 1980). (D) Activity rhythm of a
reindeer (Rangifer tarandus platyrhynchus) recorded continuously
for one year in its natural habitat in polar Norway (70o N). Orange
lines denote civil twilight at dawn and dusk (~10 lux light). Yellow
lines denote sunrise and sunset. This record illustrates how circadian
organization of rest–activity states may be abandoned in some
species during the continuous light of arctic summer (Adapted from
Van Oort et al. 2005).

Circadian rhythms are temperature compensated

Persistence in constant conditions with a τ near 24 hours is one
defining property of circadian rhythms; another is stability of τ
despite large changes in the temperature of organisms.
Homeothermic animals employ physiological and behavioral
mechanisms to minimize the variability of their internal
temperature, but the environment largely determines the
temperature of plants and poikilothermic animals. Biochemical
reactions normally vary with temperature, doubling or tripling in
rate with every 10o C increment. Early studies demonstrating relative
invariance of τ across a wide temperature range in poikilothermic
organisms appeared to violate this physical “law” and led to
considerable debate over whether daily rhythms might be driven
exogenously, by an unknown environmental factor, rather than by an
endogenous clock based on biochemical processes.
The exogenous hypothesis failed, however, to parsimoniously
account for the variation in τ among individuals housed in the same
environment, or within individuals over time or in different
physiological conditions, and no covert environmental “Factor X”
was ever identified. It is now accepted that an internal clock system
drives these rhythms, and temperature compensation is viewed as a
logical necessity for circadian timers. If the frequency of a
mechanism that drives circadian rhythms varied with temperature,
its output would reflect the organism’s history of temperature
changes rather than the passage of time. It would, in other words,
make a fine thermometer but a poor clock. Recent advances in our
understanding of the molecular basis of circadian timekeeping,
discussed further below, suggest mechanisms by which temperature
compensation is achieved physiologically.
 Rhythm Parameters and Terminology

Chronobiologists use a specialized terminology as convenient

shorthand for communicating with each other. The terms or their
usage are, however, often unfamiliar to nonspecialists, so it is
worthwhile to explicitly define a few important terms in any
introduction to this field. A rhythm may be formally defined as any
process that repeats itself at regular intervals. A device that produces
a rhythm is an oscillator. If more than one oscillator is involved in
regulation of a rhythm, the one that ultimately sets the rhythm’s
long-term periodicity is the pacemaker for that rhythm. An
oscillator can be used in several different ways. By counting cycles or
portions of an oscillator’s cycle, a measure of duration or elapsed
time can be obtained, and events can be triggered at specified
intervals. By synchronizing an oscillator to an external cycle (e.g., the
solar day), it can then be consulted as a clock to recognize local time
(e.g., any arbitrary time of day). The latter is a critical timekeeping
function, since only a few times of day are sharply marked in the
environment (e.g., dusk and dawn), but behavioral or physiological
processes might be profitably linked to any number of other times.
Circadian rhythms of behavior and physiology, whether they are
continuously variable (e.g., body temperature, blood pressure, heart
rate) or discrete events (e.g., onset of locomotor activity, mealtime,
sleep onset) can be conceptualized as hands of the circadian clock.
By monitoring these observable hands, we can identify the position
(phase, or instantaneous state) of the clock within its cycle (Figure
4.2). These processes are distinct from the underlying clock
mechanism (metaphorically, the gears of the clock), which may not
be directly observable. The distinction is important because altering
overt behavior or physiology (e.g., preventing sleep onset at its
predicted phase) may have no effect on the underlying clock
mechanism driving the observed rhythm, even if overt rhythmicity is
altered temporarily.
Figure 4.2 (A) Daily rhythm of core body temperature recorded
by radiofrequency transmitters implanted in the peritoneal cavity of
adult male Sprague Dawley rats. This illustrates how continuous
variables begin to change in advance of changing LD conditions
(temperature rises in anticipation of the dark period and falls before
light onset). (B) Circadian rhythms can be modeled mathematically
by sine waves and other functions. These waves can be characterized
by their wavelength (duration or period, denoted by the Greek letter
τ) and by their amplitude (peak level relative to mean level). The
time of the peak level defines the acrophase. Zeitgebers, such as the
LD cycle, also have a cycle duration (T) and may have an amplitude.
Under constant conditions, τ of an overt rhythm is an accurate
reflection of the periodicity of the underlying pacemaker driving that
rhythm. When a free-running pacemaker is exposed to a regularly
recurring external cycle that affects its periodicity sufficiently, it
becomes entrained to that cycle, by adopting a stable phase relation
to it and expressing the same periodicity as the external cycle. Any
recurring stimulus that can entrain a circadian rhythm is referred to
as a Zeitgeber (German “time giver”).
The timing of an entrained rhythm can be expressed relative to other
rhythms by comparing particular phases of each cycle. For example,
“wake-up” time, a phase of the daily sleep–wake cycle, can be
expressed in minutes or in radian degrees of a full 360o cycle relative
to a defined phase of the solar day, such as sunrise. Thus, wake-up
time may lead sunrise by 1 hour or lag (follow) it by 2 hours.
Different overt rhythms have different phase relations to the solar
day.
Some rhythms have proven to be particularly useful markers of
circadian pacemaker phase. The most commonly used laboratory
animals in circadian biology are mice, rats, and hamsters, in large
part because they exhibit robust, easily measured daily rhythms of
wheel-running activity. Wheel-running rhythms can be remarkably
precise (e.g., Figure 4.1). The standard deviation of successive daily
activity onsets by flying squirrels (Glaucomys volans) free-running
for weeks in constant dark may be as low as 6 minutes, or 0.4% of a
24 hours (1,440 minutes) cycle. Thus, given the time of the last
activity onset in a constant environment and knowledge of the
average τ, the next activity onset can be predicted to within a few
minutes. The underlying circadian pacemaker must be at least as
stable as any of the measurable rhythms it drives, and its cycle-to-
cycle precision has, in fact, been estimated to be twice that of the
overt activity rhythm (Pittendrigh & Daan 1976).
The phase and period of circadian rhythms reflect the timing of the
underlying circadian clock. Another feature of a rhythm is its
waveform, which describes the variation in level (e.g., body
temperature) across successive phases of the cycle (e.g., Figure
4.2A). Circadian waveforms come in a variety of shapes (e.g.,
sinuosoidal, sawtooth) and may have one or more daily peaks.
Sinusoidal rhythms can be described mathematically by a cosine
function, the peak of which is designated the acrophase (Figure
4.2B). The difference between acrophase and mean value can be used
as a measure of rhythm amplitude. However, unlike the steady-
state rhythm period, rhythm amplitude cannot be assumed to reflect
a property of the underlying clock. A stopped clock has a zero
amplitude oscillation, as will all of the rhythms that it drives, but a
stopped rhythm (e.g., if an animal is kept awake for 24 hours,
thereby eliminating the sleep–wake cycle) cannot be taken to imply a
stopped clock. Many factors downstream from the clock can affect
the overt expression, and therefore the waveform, of specific
rhythmic variables.

Evolution and Adaptive Uses of Circadian

Clocks

Circadian clocks are phylogenetically ancient

True circadian rhythms (i.e., those that persist in temporal isolation)
are found in at least some prokaryotic organisms (those without
nucleated cells; e.g., cyanobacteria Synechococcus, and extremophile
archaeon Halobacterium salinarum NRC-1) and in the simplest
unicellular eukaryotes (e.g., green algae Chlamydomonas). Circadian
clocks must therefore have evolved very early in the history of life on
earth, at least 3.5 billion years ago. Genetic analysis (reviewed below)
suggests that circadian clocks may have evolved several times in
different evolutionary lines (convergent evolution).
The early appearance of circadian clocks was likely driven by the
need to shield certain biochemical processes from ionizing solar
radiation (the “escape from light” hypothesis; Pittendrigh 1993).
Ultraviolet light damages DNA, which is particularly vulnerable
during replication. Circadian clocks can minimize photooxidative
damage by programming cell division to occur at night. In mobile
organisms, the circadian clock can further contribute by stimulating
movement away from sources of brighter light (e.g., the surface of
the ocean) in anticipation of sunrise.
Another hypothesis for the evolution of circadian clocks is suggested
by evidence for circadian cycling in the oxidation state of
peroxiredoxin (antioxidant) molecules in model organisms from all
three domains of life (Bacteria, Archaea, and Eukaryota; Edgar et al.
2012). The metabolic oscillator driving these rhythms may have
arisen at the time of the Great Oxygenation Event (~2.5 billion years
ago), to coordinate the activity of cellular antioxidant mechanisms
with daily metabolic cycles in early aerobic organisms.
The evolution of circadian clocks may also have been favored by the
need to separate incompatible biochemical processes. Cyanobacteria,
for example, exhibit daily rhythms of nitrogen fixation and of
photosynthesis that are in “antiphase”: photosynthesis occurs in the
day, nitrogen fixation at night. Photosynthesis releases oxygen,
which inactivates the nitrogen-fixing enzyme nitrogenase; thus,
these processes must be separated, either spatially or temporally. In
simpler organisms lacking intracellular compartments, circadian
programming can achieve the necessary temporal segregation.
These hypotheses about the origins of endogenous circadian
timekeeping emphasize two primary functions of circadian clocks:
optimal coordination of the organism with the outside world, and
optimal coordination of internal biochemical and physiological
processes. There are numerous examples that illustrate the adaptive
value of temporal coordination of behavior and physiology with the
environment, but the importance of circadian timing for internal
physiological coordination is less well documented.
Some animals can apparently function without circadian control at
the cellular and systems physiology levels. These species have
adapted to life in more or less constant environments, such as
underground burrows or deep caves, or at high latitudes where light
is continuously present or absent for months at a time. Caverniculus
(cave dwelling) organisms typically lack circadian rhythms of
behavior and at least some lack circadian rhythms of gene expression
(Idda et al. 2012). Similar absence of rhythmicity has been
documented in insect, avian, and mammalian species in some
environments or life stages (Van Oort et al. 2005; Lu et al. 2010;
Bloch et al. 2013) (e.g., Figure 4.1D). Clearly, 24 hour rhythmicity is
not an obligatory temporal framework for physiology in all species.
This conclusion is reinforced by many observations that animals with
severely disrupted or absent circadian organization, resulting from
molecular, neural, or environmental manipulations, survive in the
lab. Nonetheless, there is increasing evidence that loss of circadian
rhythmicity in normally rhythmic species has a range of subtle
metabolic effects that compromise health and longevity. In one
striking demonstration, disruption of circadian rhythms, by repeated
exposure to LD shifts (simulating chronic jetlag), was shown to
significantly increase mortality in mice (Davidson et al. 2006). These
results confirm the importance of circadian regulation for optimal
functioning of physiological systems in species that are adapted to
life in a 24 hour world.

Circadian clocks coordinate behavior with the external

world
Complex, multicellular organisms no longer need a circadian
mechanism to “escape from light” because of their numerous
adaptations to protect against solar radiation. Nevertheless, they use
their circadian systems and specialized sensory and motor capacities
to adopt diurnal or nocturnal lifestyles. For these species, foraging,
reproductive, and other behaviors are most efficient—and safest—if
restricted to appropriate times of day. Diurnal animals (e.g.,
flycatchers) have high-acuity, low-sensitivity visual systems optimal
for daytime light levels. Foraging at night would be inefficient in
competition with species with excellent night vision or enhanced
auditory capabilities (e.g., bats). It would also be potentially
dangerous, if nighttime activity increased exposure to daily extremes
of local climate (e.g., cold) or to nocturnally adapted predators.
Circadian rhythmicity is important for timing not only daily activities
but also developmental events that occur only once in a lifetime. The
prototypical example is the process of eclosion (emergence of an
adult from the pupal case) in the fruit fly, Drosophila melanogaster.
Air that is too hot or dry may damage the wings of newly eclosed flies
by drying them before they can flatten into an appropriate shape.
Eclosion is therefore restricted to the early morning hours, when air
temperature tends to be lowest and humidity highest. Although each
fly emerges only once, a daily rhythm of eclosion is evident at the
population level, with groups of new adults emerging only at a time
corresponding to subjective morning, even under constant
environmental conditions. Circadian control of eclosion gains
additional significance in insects that mate only for a brief time
immediately after emergence; if members of the population do not
emerge in synchrony, there will be fewer successful matings and
fewer offspring. The necessity of synchronizing behavior to facilitate
mating is likely to exert a significant selection pressure for accurate
circadian timekeeping in many species.
In birds, fledging is a developmental event that may entail significant
risk of predation. One solution to minimize mortality within a
population is to restrict fledging to a particular time of day. If all
fledglings leave the nest at the same time, this will reduce the total
number lost to predation by overwhelming the maximum killing
capacity of predators. Circadian control of fledging behavior, and its
advantage for survival, have been documented in guillemots (Uria
lomvia), cliff-breeding arctic sea birds that suffer heavy predation
from glaucous gulls (L. hyperboreus) (Daan 1981).
Circadian control of developmental events illustrates an advantage of
circadian timing that may also be important at other life stages. At
the time of a critical developmental event, an organism may lack the
sensory capacity or access to information that would allow detection
of optimal environmental conditions (e.g., detecting external
humidity from inside a sealed pupal case). Without direct access to
relevant external conditions, the organism can time critical events to
occur at the appropriate time by entraining their circadian clocks to a
highly correlated and more readily detectable cue (e.g., the lighting
cycle). Mature organisms can exploit the same strategy in other
situations, such as in the selection of nesting sites. Many species of
bats roost in caves sufficiently deep to provide constant darkness and
temperature. Other animals, like the clockwork eastern chipmunk
introduced at the beginning of this chapter, nest in burrows in the
ground or hollows in trees, which also obscure daily variations in
light and temperature. These habitats can be utilized because an
appropriately synchronized circadian clock can function as an alarm
clock to ensure emergence from the nest site at an appropriate time
of day, without the need for continuous access to information about
current external conditions.
The general rule that there is an optimal time of day for sleep and
wake extends to specific portions of the active phase. Different
sources of food may be available at different times of day, so effective
foraging strategies may require attention to both spatial and
temporal cues that define where and when food may be obtained. For
example, flowers produce nectar and open their petals at particular
times of day, thereby limiting the resources they must expend to
attract potential pollinators. The same approach also enhances the
probability that pollen will be transported within a species and
minimizes the spread of pollen to unrelated species. Honeybees can
learn and remember both the time and the place at which a nectar
source is available and can communicate this information to other
foragers in the hive via the well-known waggle dance (see Chapter
14). Laboratory studies indicate that bees can use their circadian
clocks to associate food locations with as many as nine times of day.
This ability to “sense” time of day and organize behaviors
appropriately for each phase (Zeitgedächtnis ; German, “time
memory”) demonstrates that the bee possesses a true clock by which
it can recognize local time, analogous to our use of a wristwatch to
schedule activities precisely within our own time-ruled society. The
circadian clock is thus more than an on–off switch for behavioral
states; rather, it can be “continuously consulted” for use as a
chronometer to permit a highly differentiated temporal program of
daily foraging and other activities.
A similar ability to recognize time of day and generate phase-
appropriate learned behaviors using internal circadian cues has now
been demonstrated in several other species, including birds, rodents,
fish, and insects. Rats and mice can learn that food is available at one
or two times of day and exhibit so-called “anticipatory activity” prior
to the daily mealtimes (Mistlberger 2009). The clock mechanism
that controls daily food anticipation is separate from the clock that
synchronizes rest–activity rhythms to daily LD cycles and will be
discussed further below. Lab rodents can also discriminate between
at least two times of day, enabling them to associate different times
of day with different feeding locations (time–place learning; Mulder
et al. 2013). Some bird species can learn at least four time–place
associations for food access.
Anticipation of mealtimes highlights a primary adaptive feature of
endogenous circadian timing; namely, the ability it confers to
prepare in advance for the major environmental changes associated
with the day–night cycle. In mammals and birds, body temperature
is lowered during the daily rest period, which minimizes energy
expenditure. To prepare for the active period, the circadian clock
must initiate autonomic and hormonal events (e.g., cortisol
secretion) to raise body temperature and mobilize stored energy
reserves prior to dawn or dusk. Thus, when daylight brings the
opportunity to forage, mate, or defend a territory, the organism is
ready, rather than beginning its preparations only in response to the
signal provided by dawn.
Organisms that forage over a wide area may have to begin the
migration back to their resting sites well in advance of day or night
onset. The need for anticipatory behavior may be particularly crucial
for slow-moving organisms; for example, worms vulnerable to
desiccation during the day may need to migrate down in the soil
before there are obvious changes in moisture or temperature, lest
they be caught high and dry in the morning sun.
These examples illustrate how circadian timing can contribute to
homeostasis. Homeostasis (L. “similar standing”) refers to the
regulation of vital physiological functions within certain optimal
limits. It has traditionally been conceptualized as a reactive process,
based on a negative feedback mechanism. Thus, changes in a
regulated variable, such as body temperature, result in compensatory
responses, such as shivering, sweating, or seeking sun or shade, until
the variable is restored to the desired value. The circadian clock can
minimize disruptions of homeostasis by initiating responses in
anticipation of a predictable daily challenge. Vertical migration of a
worm into the soil anticipates, and thereby prevents, a severe
challenge to hydration. This adaptive feature of circadian timing has
been referred to as “predictive” homeostasis (Moore-Ede 1986).
Despite the many apparent advantages of circadian timekeeping for
anticipating environmental change and coordinating behavior with
daily cycles, few studies have attempted to test directly whether
circadian organization of behavior is necessary for survival in natural
habitats. One means of conducting such a test is to remove the
circadian clock by a surgical or genetic procedure and to evaluate
survival after release into the natural habitat. This type of
experiment has been attempted twice using wild-caught diurnal
rodents. In both cases, the circadian clock was removed by a
localized brain lesion (of the suprachiasmatic nucleus, SCN;
discussed below). In one study, SCN-ablated and intact white-tailed
antelope ground squirrels (Ammospermophilus leucurus) were
maintained in a large outdoor enclosure in their habitat of origin
(DeCoursey 2014). Squirrels with SCN ablations were active at food
sites during both day and night, while control squirrels were almost
exclusively day-active. The unplanned entry of a feral cat resulted in
overnight loss of 60% of the squirrels with SCN ablations, compared
to 29% of the intact squirrels. In a second study, wild-caught eastern
chipmunks (Tamias striatus) received SCN ablations and were
returned to their home range after behavioral tests had confirmed
that their circadian rhythms were either eliminated or severely
disrupted (DeCoursey 2014). Although telemetric recordings
indicated that SCN-ablated and control animals remained in their
burrows at night (perhaps as a direct response to the absence of
light), the SCN-ablated group suffered significantly greater predation
from nocturnally active weasels. It was hypothesized that nocturnal
restlessness of SCN-ablated chipmunks in their burrows may have
alerted weasels to their location. These results are consistent with the
prediction that circadian rhythms of rest and activity states have
adaptive significance and suggest that any genetic mutation that
greatly attenuates or alters the timing of circadian activity rhythms
would be rapidly eliminated from the gene pool.

The circadian clock is also a compass and a calendar

Circadian clocks have the capacity to be synchronized precisely to
periodic environmental stimuli, such as daily light–dark (LD) cycles.
A clock synchronized to the solar day provides not only knowledge of
local time but also the basis for an internal compass (knowledge of
direction) and a calendar (knowledge of season). Many species have
exploited the circadian clock in one or both of these ways. The
honeybee uses its circadian clock to recognize and remember the
time at which good nectar sources are available, but it also uses the
clock to navigate to and from the hive. The bee takes its flight
heading relative to the sun’s azimuth, the point on the horizon below
the apparent position of the sun, which changes throughout the day
as the earth rotates. The circadian clock is used to compensate for
this continuous motion, so that the flight heading relative to the
azimuth is continuously adjusted by reference to circadian clock
phase. If the circadian clock is shifted artificially, the direction of
flight also shifts in parallel. Circadian clock–compensated sun-
compass orientation is also used by birds, mammals, reptiles, and
fish that migrate over great distances, typically on a seasonal basis
(Wallraff et al 1981).
Annual rhythms of activity, metabolism, reproduction, and other
functions are widespread in animals and plants. In some cases,
annual rhythms are controlled by an internal, circannual clock,
which in mammals appears to be physically distinct from the
circadian clock. More commonly, annual rhythms are timed in part
by measuring seasonal changes in daylength (photoperiod), using a
mechanism involving the circadian clock, sometimes in combination
with endogenously timed components (see below for further
discussion of seasonality).
 Environmental Synchronization of
Circadian Rhythms

Entrainment by light–dark cycles

The circadian clock coordinates behavior and physiology with
external cycles and serves the ancillary functions of compass and
calendar. All of these adaptive functions depend on maintaining
stable synchronization of the circadian clock to the day–night cycle.
In the absence of stable synchronization, behavior would not be
timed appropriately to match daily rhythms in the environment, and
both compass and calendar functions would be compromised. The
process of matching the phase and period of an endogenously
rhythmic circadian clock to an environmental cycle is known as
entrainment.
How is entrainment achieved? In temporal isolation, circadian
rhythms typically deviate systematically from 24 hours. For a LD
cycle with a given periodicity “T” (e.g., 24 hours) to entrain a free-
running rhythm with periodicity τ (e.g., 23.5 hours), it must be
capable of altering the motion of the circadian clock to correct for the
difference τ - T (in this case 30 minutes). To maintain synchrony
with the LD cycle, the clock must either be permanently slowed by 30
minutes/cycle (a “continuous” effect on its rate of oscillation), or it
must be reset (shifted) each day (a “discrete” effect on its phase,
analogous to setting a fast wristwatch or alarm clock back by 30
minutes each day). An effect of LD on τ is also known as a
“parametric” effect of light, because period, along with amplitude, is
a parameter of a rhythm. A discrete effect of LD on phase is known
as a nonparametric effect of light, because phase is a point on a cycle,
and not a parameter needed to describe that cycle.
Light has both continuous and discrete effects on circadian timing,
and both likely contribute to entrainment. In constant light, τ
lengthens (the clock runs slower) in proportion to light intensity in
most nocturnal animals and shortens (the clock runs faster) in most
(but not all) diurnal animals. (This parametric effect is one of
“Aschoff’s Rules,” named after one of the pioneers of circadian
biology, the German scientist Jürgen Aschoff.) In constant dark, a
single “pulse” of light (from seconds to an hour or more in duration)
can rapidly shift the phase of circadian rhythms in either direction,
depending on when within the circadian cycle light exposure occurs.
Phase shifts may also be accompanied by small changes in τ, again
depending on the timing of the light exposure. In nocturnal species,
shifts of phase can account for the principal features of entrainment
to LD cycles under laboratory conditions. In some, if not all diurnal
species, changes in τ may also be necessary.

LD entrainment in nocturnal rodents: The nonparametric

model
Figure 4.3 illustrates a series of experiments that map out the phase-
shifting effects of light exposure at different phases of the circadian
cycle. Phase shifts can be readily quantified by comparing a standard
phase of the circadian rhythm in the cycles before and after the light
stimulus. For many organisms, the beginning of the daily active
period is a convenient phase marker for measuring shifts. Regression
lines can be fit to successive daily activity onsets before and after the
light stimulus. The temporal displacement of the two regression lines
on the intercept day (typically the day after the light pulse) provides
the magnitude and direction of phase shift. The data can then be
summarized by plotting a phase-response curve (PRC; see
Figure 4.3F) illustrating how the size and direction of phase shift
relate to the circadian phase of stimulus presentation.
Figure 4.3 Plasticity of rest–activity rhythms in response to
varying external conditions. (A) Wheel-running activity of a rat fed
ad libitum and then restricted to a 2 hour daily meal provided once
every 24 hours (denoted by white vertical bar) and then once every
25 hours (oblique white bar). The rat exhibits a daily rhythm of food
anticipatory running that tracks mealtime when it is delayed by 1
hour each day. During the first 2–3 weeks of food restriction, the rat
also shows a gradual shift of nocturnal activity into the light period,
but throughout restricted feeding, the rat continues to exhibit small
bouts of nocturnal activity. Multiple lines of evidence indicate that
rat behavior under these conditions is controlled by two types of
circadian clocks, one that remains entrained to the LD cycle and
another that is entrained by scheduled feeding (adapted from
Mistlberger & Marchant 1995). (B) Activity rhythm of a mouse
maintained in LD and then required to work for food by running in
its wheel. As the work requirement per food reward increases, the
mouse gradually shifts its predominantly nocturnal activity into the
light period. When food is then available freely during the last week,
in constant dark, the mouse reverts to a nocturnal pattern, indicating
that the LD-entrained clock has not been shifted by this procedure.
This illustrates that under certain conditions the timing of activity
can shift relative to the timing of the circadian clock, i.e., the mouse
may switch is temporal niche from nocturnality to diurnality
(adapted from Hut et al. 2011). (C) Activity rhythms of two castrated
montane voles (Microtus montanus) maintained continuously in an
LD cycle with 8 hours light per day. In the upper chart, the vole
received a subcutaneous implant of slow release testosterone (T),
while in the vole in the lower panel received a control implant (C)
with no hormone. Activity in both voles exhibits a 3–4 hours
ultradian rhythm, with most activity occurring during the light
period prior to the implants. After the implants, the vole receiving
testosterone continues to exhibit an ultradian rhythm, but now most
of the activity occurs in the dark period. The castration procedure
simulates gonadal regression that occurs seasonally in the winter,
while the testosterone treatment simulates gonadal regrowth that
occurs in the summer. These records demonstrate how overt activity
patterns can simultaneously express ultradian (3–4 hours), circadian
(24 hours), and infradian (annual) rhythmicity (Adapted from
Rowsmitt 1986).
In an LD cycle, a nocturnal animal rests during the day and is active
most of the night. In constant dark, the rest phase of the circadian
cycle is designated the subjective day and the active phase the
subjective night (the opposite is true for diurnal animals). When
light exposure occurs during the subjective day, there is little or no
effect on the circadian clock. However, light exposure in the
subjective night induces an abrupt phase shift. The direction and
magnitude of the phase shift depend on the precise time of light
exposure. Early in the subjective night, light induces a phase delay
shift, i.e., the next cycle occurs later than predicted based on the
preceding cycles, and the rhythm resumes its free-run from this new
phase. Later in the subjective night, light induces a phase advance
shift, i.e., the next cycle occurs earlier than predicted. The resulting
PRC therefore has three regions: an unresponsive zone in the
subjective day, a phase delay zone in the early part of the subjective
night, and a phase advance zone late in the subjective night.
The bidirectional shape of the PRC is characteristic of all organisms
that can be entrained by light, regardless of whether they are
nocturnal, diurnal, or crepuscular, or whether the circadian clock has
a period longer or shorter than a solar day. A slow circadian clock
(e.g., τ = 24.5 hours) will shift later each day relative to a 24 hours
LD cycle until lights-on (“sunrise”) begins late in the animal’s
subjective night or early in the subjective day, when the circadian
clock is phase advanced by light. The circadian clock will stop
drifting and appear to express a precise 24 hours τ when sunrise
occurs at that point in the clock cycle when the net phase advance is
equal to τ - T (in this case, 30 minutes). A faster circadian clock (e.g.,
τ = 23.5) will shift earlier relative to a 24 hours LD cycle until the end
of the light period (“sunset”) reaches the beginning of the subjective
night, sufficient to induce a net phase delay shift of 30 minutes. Once
entrained, a clock that would naturally delay by 30 minutes each
cycle cannot because it is advanced 30 minutes by morning light,
while a clock that would naturally advance by 30 minutes each cycle
cannot because it is delayed 30 minutes by evening light.
The PRC and τ together can account for individual differences in the
precise phase relation between rhythm and LD cycle that we
recognize in the concept of chronotypes. For a given PRC shape, a
shorter τ will result in a more advanced phase of entrainment. In
humans, this would result in early rising (the “early-bird” or
“morning” chronotype). A longer τ would result in late rising (the
“night-owl” or “evening” chronotype). This predicted relation
between endogenous τ and phase of entrainment has been confirmed
empirically in humans and other species. For a given τ, differences in
PRC shape will also have predictable consequences for entrainment.
Thus, there are two circadian clock properties—its rate of cycling and
its circadian rhythm of sensitivity to light—that can be modified to
adjust or preserve the phase of entrainment as environmental
conditions change seasonally or over generations.
It should be noted that while we typically refer to sharply defined
markers of the circadian cycle such as awakening or sleep onset,
what is actually entrained is a complex circadian program that affects
functions throughout the circadian cycle. Thus, most humans show a
daily pattern of increasing alertness and improved performance
throughout the day, until an evening hour when both alertness and
performance begin to deteriorate. The timing of this pattern is
predictably related to individuals’ preferences for early or late
awakening, with morning-type individuals showing earlier peaks and
troughs of performance relative to external clock time (Dijk & Von
Schantz 2005).
The PRC and τ can also account for some empirically established
constraints on LD entrainment. In some work settings, humans may
lack exposure to natural time-of-day cues; examples include nuclear
submarines that submerge for many days and workstations in orbit
or near the poles. In these settings, to meet requirements for
continuous operations, it may seem desirable to employ work
schedules that differ from 24 hours. However, humans (and other
animals) are unable to entrain to such schedules if they differ by
more than 1 or 2 hours from 24 hours. This observation is
predictable from the PRC to light; maximum phase shifts to single
pulses of light are on the order of 1–3 hours, suggesting a maximum
range of entrainment of ~24 hours ± 3 hours.
The same phase-shifting constraints underlie the phenomena of jet
lag and shift-work malaise. These conditions are characterized by
transient sleep disturbances, daytime fatigue, and gastrointestinal
complaints, associated with rapid travel to new time zones, or
rotations between day, evening, and night shifts. The cause is a
temporary mismatch between circadian clock time and the new
sleep–wake schedule. Travel between North America and Europe
results in a rapid 4–9 hour shift of the LD cycle (an advance if
traveling east, a delay if traveling west). The circadian clock cannot
shift by this amount in one cycle but will re-entrain to the new local
time cues in jumps of ~1–2 hours per day.
Complicating this analysis is evidence that these observations apply
only to the central neural pacemaker. Other parts of the body
(muscles, liver, etc.) have their own, “local” circadian oscillators,
which may respond only sluggishly to the light-induced shifts of the
pacemaker (Yamazaki et al. 2000; Balsolabre 2001). Thus, full
physiological adaptation to a major shift in local time cues may take
many more days than suggested solely by analysis of the pacemaker
itself or of rhythms closely regulated by it. In fact, the resulting
physiological discordance among peripheral organs may be a major
contributor to the malaise resulting from such abrupt shifts in
schedules.
There is evidence that τ, and therefore the phase of entrainment,
changes over the lifetime of individuals. Some studies of rodents
have demonstrated a shortening of τ associated with aging (e.g.,
Pittendrigh & Daan 1974). In humans, older people typically wake up
earlier, which would be consistent with a shortened circadian τ,
although a difference in τ between young and old has not been
confirmed empirically (Czeisler & Dijk 2001). The delayed phase of
sleep onset and awakening preferred by adolescents (Roenneberg et
al. 2004) would be consistent with a lengthening of τ during and
after puberty, for which there is some evidence (Carskadon et al.
2004; Micic et al. 2016). Such physiological changes would be
potentiated by social factors promoting late-night activities in older
adolescents, such that the phase of circadian clock–controlled
awakening would be misaligned with societal demands for early
morning attendance at school, for example. Misalignment between
“body time” and “social time” has been conceptualized as social jet
lag, and, like repeated jet lag caused by travel across time zones, is
thought to be a risk factor for impaired health and performance
(Roenneberg et al. 2013)
LD entrainment in diurnal species
The model of entrainment by daily phase adjustments in response to
morning or evening light works well to explain entrainment in semi-
fossorial, nocturnal animals such as Syrian hamsters, flying
squirrels, mice, and rats studied under controlled laboratory
conditions (DeCoursey 1986). For some diurnal species, such as the
European ground squirrel, the model cannot explain activity
rhythms observed in the field (Hut et al. 1999). These animals sleep
in dark burrows underground at night, emerge to forage each day
several hours after sunrise, and return to their burrows several hours
before sunset. Consequently, light exposure is entirely self-selected.
According to the “nonparametric” entrainment model, the circadian
clock in these animals should free-run until the active phase either
begins at dawn (if τ < 24) or ends at dusk (if τ > 24 hours). Instead,
European ground squirrels exhibit precise LD entrainment without
ever seeing dawn or dusk. Entrainment in this case appears to
depend on an effect of light on τ, as initially proposed by Aschoff.
In constant dark or dim light, a single light “pulse” (e.g., 30 minutes
duration) that induces a phase delay shift on average lengthens τ,
while a pulse that induces a phase advance shift shortens τ. Diurnal
species tend to exhibit a lengthening of τ in response to light
exposure during the first half of the subjective day (their active
period) and a shortening of τ in response to light during the second
half of the subjective day. The daily timing of activity exhibited by
European ground squirrels in their natural environment could only
be simulated mathematically when a τ response curve (τRC) was
combined with the PRC in the model (Hut 1999). Applying the model
more generally suggests that τ responses enhance the precision of LD
entrainment (i.e., reduced day to day variability of activity onset time
relative to sunset) (Beersma et al. 1999).
Although the phase of entrainment of individual animals can be
successfully modeled using parametric and nonparametric responses
to light, the opposite phasing of rest–activity states that define
nocturnal and diurnal animals is not explained by LD entrainment
models. In fact, nocturnality and diurnality do not appear to reflect
properties of the master circadian clock that mediates LD
entrainment. These differences instead reflect how the circadian
clock is coupled to brain mechanisms that are downstream in the
flow of temporal information from the environment to animal
behavior and physiology. Many of the downstream targets of master
clock output also contain local circadian clocks, and many of these
are entrainable by other time cues, collectively known as “nonphotic”
zeitgebers.

Nonphotic stimuli and plasticity of circadian phenotypes

Most of what we know about the formal properties of circadian
rhythms and entrainment comes from studies of just a few model
species. Under laboratory conditions, these species are day or night
active, defining nocturnal and diurnal circadian phenotypes,
respectively. There is substantial evidence, however, that phenotypes
are plastic and not rigid, and that much of this plasticity reflects
sensitivity of circadian clocks to nonphotic stimuli. The most
important of these, depending on the species, are food, temperature,
behavioral arousal, and social cues (Mistlberger & Skene 2004).
The impact of food availability on the timing of animal behavior is
readily demonstrated by restricting food access to a particular time
of day. In laboratory mice and rats (and many other species of
mammals, birds, and fish), restricted feeding schedules induce a
daily rhythm of food-anticipatory activity, with formal properties
indicating circadian clock control (i.e., persistence for several cycles
in constant conditions with no food; circadian limits to entrainment;
gradual rather than immediate resetting to shifts in meal timing;
Mistlberger 2009) (Figure 4.4A). Importantly, food anticipatory
activity rhythms in mammals appear to be controlled by a circadian
clock separate from the clock that mediates entrainment of activity
rhythms to LD cycles (the SCN, discussed further below). This
enables animals to anticipate a meal at any time of day, without
shifting the phase of the LD-entrained clock, which in some species
is critical for annual rhythms (see below) or for clock-based
navigation. An interesting outcome of this food-entrainment process
is that nocturnal species may express most of their foraging behavior
during the daytime, exhibiting an apparent switch of their temporal
niche. The ability to exploit food sources that are available at unusual
times of day is of obvious adaptive value.
Figure 4.4 A phase-response curve (PRC) to light and its
derivation. The PRC is a convenient plotting convention to
summarize the empirical observation that light can phase shift
circadian rhythms and that the direction and magnitude of the shift
depends on the circadian phase of the organism’s clock at which the
light exposure occurs. Panels A–E represent simulated free-running
activity rhythms of nocturnal animals kept in constant dark, with the
heavy bars denoting the active period of each circadian cycle. Free-
running period is slightly >24 hours (the rhythms are drifting later
each day). In each panel, the square indicates the timing of a brief
exposure to light (e.g., 15 minutes). In Panel A, light falls during the
animal’s rest period (in this case, the subjective (internal) day, when
the sun would normally be up). No phase shift occurs. In Panels B–E,
light falls early, mid, or late in the active period (the nocturnal
animal’s subjective night), and in each case this produces a rapid and
permanent shift of the animal’s rhythm. However, phase delays
result only when light falls in the first half of the night (with larger
delays the later the exposure), and phase advances only when light
falls in the latter half of the night (with large advances near the
middle of the night). The clock thus has a circadian rhythm of
sensitivity to light, which ensures entrainment. If the clock tends to
drift later, “morning” light will advance it, whereas if it drifts earlier,
“evening” light will delay it. (Adapted from Moore-Ede et al. 1982).
Temporal niche switching in nocturnal rodents can be induced
without limiting the time of food access, by increasing the work
required to obtain food (Hut et al. 2012). At some level of effort, mice
choose to eat less rather than work more. At this point, the active
(food seeking) period begins to gradually shift from the dark to the
light phase (Figure 4.4B). The mice then spend most of the night
inactive with a reduced body temperature approaching torpor. This
is conceptualized as a strategy to limit total daily energy expenditure
when food is difficult to acquire, by reducing waking activity at night
when the costs of maintaining a high body temperature are expected
to be greatest (nights in the real world are cooler than days). Like
food anticipation, this effect does not involve shifting of the light-
entrained circadian clock in the SCN.
Entrainment to social cues may be important for some species,
particularly those that live in complex social groups (Castillo-Ruiz et
al. 2012). Social entrainment may be mediated by effects of social
stimuli on behavioral state. This is suggested by evidence that in
some mammalian species simply keeping an animal awake and
active (‘exercising’) during its usual rest phase can shift or entrain
circadian rhythms in constant light or dark (Webb et al. 2014). In
humans, exercise very early in the morning (around the minimum of
the daily cycle of body temperature, prior to usual wake-up time) can
induce phase shifts, but any effects of sleep–wake schedules (by
contrast with scheduled exercise) appear to be very weak. For most
species, in natural environments, social stimuli may be important
primarily to the extent that they control exposure to LD cycles.
Other environmental factors affect expression of daily rhythms.
Thus, the presence of other species, such as aggressive competitors,
shifted the daily timing of activity of desert hamsters (Phodopus
roborovskii) (Scheibler et al. 2013). Similarly, the timing of daily
sleep episodes differed between related sloth species (Bradypus sp.)
as a function of differences in exposure to predation pressure (Voirin
et al. 2014).
Daily temperature cycles may be important Zeitgebers for some
species (particularly plants and poikilothermic organisms) but are
thought to play little if any direct role in entraining circadian rest–
activity cycles in mammals. Entrainment to temperature cycles
observed in some laboratory animals could be secondary to effects of
temperature on overt behavior.
The number of field observations of animals being active at the
presumed “wrong” time of day is accumulating, suggesting that
circadian timing is more flexible than expected based on evidence
from the laboratory (Hut et al. 2012). A striking example is the
contrast between the strongly bimodal (dawn and dusk) activity
patterns of Drosophila under controlled laboratory conditions and
their activity patterns under natural daylight and temperature
conditions outdoors. The sharp light–dark transitions and flattened
temperature cycles in the laboratory contrasted to the gradual dawn
and dusk transitions and high-amplitude temperature cycles of a
summer day in Italy.
The impact of these combined light and temperature differences
included replacing the characteristic laboratory pattern of a midday
period of inactivity (“siesta”) by the most sustained activity bout
shown daily, along with attenuated activity peaks at dawn and dusk
(Vanin et al. 2012). Much remains to be learned about how photic
and nonphotic stimuli combine to regulate circadian timing of
organisms in complex, natural habitats.

Neural Mechanisms of Circadian Rhythms

The physiological system that is responsible for circadian

rhythmicity must include three parts: a self-sustaining circadian
clock, one or more input pathways by which LD cycles (and other
cues) can entrain the clock, and one or more output pathways by
which the clock can control the timing of behavior and physiology.
The available evidence suggests that in most, if not all, species, there
are multiple entraining inputs, clocks, and outputs. There appear to
be many common themes across phylogeny in the organization of
circadian systems, but there is also substantial species diversity in
the details. The same is true at the molecular level of analysis.

Mammals
Localization of a circadian pacemaker in mammals
Localization of the circadian clock in mammals was guided by the
reasonable assumption that it must receive information about light
and dark, possibly by a direct pathway from the retina. Cutting the
optic nerves in mammals eliminates entrainment to LD cycles but
does not eliminate circadian rhythms of behavior and physiology;
therefore, the eyes must contain the photoreceptors necessary for
entrainment but not the circadian pacemaker itself. The axons of
retinal ganglion cells form the optic nerves, which communicate
photic information to the brain. The two nerves cross, at least
partially, forming the optic chiasm at the base of the brain just below
the hypothalamus (Figure 4.5). The nerves, now renamed the optic
tracts, exit the chiasm and project to various areas of the brain
responsible for vision, oculomotor reflexes, and other functions. If all
optic tracts are cut posterior to the chiasm, the animal is rendered
visually blind; however, it remains entrained to LD cycles. Therefore,
the retinal pathway mediating photic entrainment must enter the
brain at the level of the chiasm. Sensitive tract-tracing techniques
revealed that some retinal fibers do innervate the hypothalamus
above the chiasm, particularly the SCN, which, as its name implies,
lies atop the chiasm. Surgical ablation of this small, bilateral
structure was then shown to completely eliminate circadian rhythms
of locomotor activity, body temperature, and hormones in rats, mice,
hamsters, and other species in constant conditions (Weaver 1998).
Figure 5.5 The rat brain and some areas that exhibit circadian
rhythmicity in vivo or in vitro. The SCN (suprachiasmatic nucleus) in
the hypothalamus functions as a master circadian pacemaker that is
entrained to daily LD cycles by a direct input from the retina. The
SCN exhibits circadian oscillations of neural activity and expression
of so-called clock genes, which can persist indefinitely in explants
maintained in culture (see the panel plotting bioluminescence driven
by the clock gene Per2 fused with the firefly gene encoding
luciferase). The retina itself harbors a circadian oscillator, which
controls among other things retinal production of melatonin and
light sensitivity. The SCN coordinates oscillators elsewhere in the
brain (a subset of regions that oscillate are labeled) and in the body
(peripheral organs and tissues), both directly (solid arrows) and
indirectly (dashed arrows), by its control over autonomic efferents,
hormones, body temperature, and behaviors such as eating.
(Adapted from Guilding & Piggins 2007).
Additional experiments were needed to clarify the role of the SCN: Is
it the site of the clock, or is it merely permissive for the expression of
circadian rhythms (e.g., by conveying output from the clock to the
rest of the brain)? To address its role, studies demonstrated that
stimulation of the SCN, electrically or with drugs, caused phase shifts
of circadian rhythms, mimicking those produced by light or by other
stimuli. The capacity of the SCN to generate a circadian rhythm of
metabolic and electrical activity was demonstrated both in vivo and
when isolated from the rest of the brain in a constant environment in
vitro (the SCN are dissected out and maintained in a perfusion
chamber providing oxygen and nutrients). This demonstration of
endogenous capacity to generate circadian rhythms was followed by
evidence that transplanting an SCN from a donor animal to one in
which the SCN had been ablated surgically could restore free-
running circadian rhythms of behavior. Notably, the restored
rhythms displayed a τ characteristic of the donor animal’s genotype,
not that of the host animal (Ralph & Lehmann 1991). This
remarkable series of lesion, stimulation, recording, and transplant
experiments, conducted over approximately 20 years in numerous
laboratories, established that the SCN are an autonomous circadian
oscillator responsible for setting the phase and period of circadian
rhythms (Klein et al. 1991). The SCN have thus been accorded the
status of master oscillator, or pacemaker, within the mammalian
circadian system.

The cellular and molecular clockworks

How does the SCN circadian clock tick? One possibility is that SCN
neurons are wired in such a way that signals passed from one neuron
to the next take ~24 hours to complete a single loop through the
ensemble. An alternative to this “reverberating circuit” model is that
many or all neurons in the SCN are themselves circadian clocks,
mutually coupled so as to produce a coherent circadian output
signal. Electrical recordings from SCN neurons grown in culture
support the latter hypothesis; single cells recorded from the same
culture exhibited free-running circadian rhythms with different τ’s
(Welsh et al. 1995). This result is consistent with evidence from
unicellular organisms that circadian clocks are cell autonomous and
confirms that the intracellular nature of circadian timekeeping has
been conserved across phylogeny.
If the circadian clock is intracellular, there are likely to be genes and
proteins that cycle with a 24 hour period, and these cycles must be
self-perpetuating. The first putative clock genes were discovered in
the 1970s in the bread mold Neurospora and the fruit fly Drosophila.
These genes, frequency (frq) in neurospora and period (per) in
Drosophila, exhibit circadian rhythms of expression driven by
autoregulatory feedback and feedforward loops (Figure 4.6; Patke et
al. 2020). Both genes code for proteins (FRQ and PER, respectively)
that accumulate in the cytoplasm, translocate back to the nucleus,
and inhibit their own expression. This causes protein production to
decline and protein levels in the cytoplasm and nucleus to fall until
the genes are no longer inhibited. Other genes encode proteins that
drive a subsequent increase in expression, and the cycle is repeated.
Mutations of these (and related) genes are associated with loss of
circadian rhythmicity or changes in τ, establishing their essential role
for circadian cycling (Buhr & Takahashi 2013).
Figure 4.6 The molecular basis of circadian clocks. Inset:
Circadian clocks based on transcription-translation feedback loops
employ negative feedback loops by which gene expression is
regulated by the protein it encodes. The gene transcribes mRNA,
which is translated into a protein, which then feeds back to turn off
transcription. The transcription-translation feedback loop (TTFL)
design is conserved across phylogeny. Main figure: Details of the
mammalian TTFL circadian clock. The clock proteins BMAL1 (blue)
and CLOCK (green) dimerize and drive expression of multiple clock
genes (per 1, 2 and 3, red; cry 1 and 2, yellow; Rev-erbα, Rorα) and
other clock-controlled genes that regulate myriad cellular functions
and outputs. The PER and CRY proteins form homo- and hetero-
dimer complexes and translocate and suppress transcriptional drive
from the BMAL1:CLOCK dimers. PER and CRY proteins are
gradually removed by degradation, relieving the autoinhibition of
per and cry gene expression, and initiating the next cycle of
transcription. BMAL1 exhibits a circadian rhythm of abundance due
to negative feedback from REVERBα and positive drive from RORα.
The specific genes and proteins differ across species, but the
principle of positive and negative elements connected by feedback
and feedforward loops are a conserved feature. Recent work suggests
that circadian rhythms can also be generated by nontranscriptional
processes that regulate protein oxidation (several sources).
The principle of circadian cycling by autoregulatory transcription–
translation feedback loops (TTFLs) is evident across phylogeny, and
appears to be an example of convergent evolution, because the
specific genes in these loops are unrelated among bacteria, fungi,
plants, and animals (Figure 4.6). Within kingdoms of life, species
differences are also apparent in the number of genes (e.g., mice have
three per genes, compared to only one in fruit flies) and in their
specific roles. Stimuli that entrain the circadian clock may do so by
acutely altering the level of clock gene proteins, either delaying their
fall or advancing their rise, thereby resetting the timing cycle.
Following the discovery of circadian clock genes in mammals, these
genes were found to exhibit circadian cycling in many brain regions
outside of the SCN, including the retina, olfactory bulb, and lateral
habenula, and in all of the major organs and most tissues (Stratmann
& Schibler 2006; Mohawk et al. 2012) (Figure 4.5B-D). Cells from
many of these tissues can oscillate for a number of cycles in vitro,
indicating that each harbors a functional circadian clock. Non-SCN
clock cells can be shifted and entrained by a variety of stimuli,
including daily cycles of feeding, temperature, and systemic
hormonal signals. Circadian rhythms of physiology and behavior are
therefore the result of a distributed system of local circadian
oscillators that are synchronized with each other and with the solar
day by a variety of internal and external time cues. Within this
system, the SCN pacemaker plays a special role by controlling the
daily rhythms of food intake, body temperature, some hormones,
and autonomic neural outputs to peripheral organs, all of which
impact on the timing of peripheral clocks. In mammals, the SCN is
also unique in mediating light–dark entrainment.
Although SCN ablation eliminates circadian behavioral rhythms in
animals provided free access to food, it does not eliminate circadian
rhythms of food-anticipatory activity that emerge when food is
restricted to one or two fixed times of day (Mistlberger 2011).
Therefore, there must be circadian clocks in other brain regions (or
body tissues) that entrain to feeding related cues and impose a daily
rhythm on the expression of food-seeking activity. These clocks have
not yet been localized, but there are many candidates, as most
circadian oscillators in the brain and body are entrainable by feeding
cycles, in intact and SCN-ablated mice and rats. Coordinating
behavior and physiology with daily feeding opportunities is critical
for survival in many species, so it is not surprising that feeding cues
can dominate LD cues when these are in conflict.
Non-transcriptional circadian clocks
Although circadian clocks based on transcription-translation
feedback loops appear to operate in rhythmic species across
phylogeny, accumulating evidence for circadian cycling without
transcription indicates that circadian clocks can operate on other
principles. Circadian rhythms of cellular redox state have now been
described in model organisms from all of the kingdoms of life,
including in cells lacking DNA (e.g., red blood cells in humans). The
mechanism by which these rhythms are driven is currently unknown
and may represent a proto-clock common to most life forms (O’Neil
et al. 2013).
Nonmammalian vertebrates
The circadian systems of birds and reptiles (sauropsid vertebrates)
are less centralized than those of mammals (Gwinner & Brandstatter
2001; Cassone 2014). In adult mammals, the photoreceptors for
entrainment are found only in the retina. In sauropsid vertebrates,
photoreceptors important for entrainment are found in the retina,
pineal gland, and in one or more parts of the brain (encephalic
photoreceptors). In mammals, the SCN is a master pacemaker that
coordinates secondary oscillators in the retina, periphery, and
possibly other brain areas. In sauropsids, a homolog of the SCN is
localized to one or more structures in the suprachiasmatic area, but
ablation of this area in birds disrupts or eliminates circadian
rhythms only in constant light or dark. In some avian species,
circadian rhythms are eliminated or greatly disrupted by ablation of
the pineal gland, while in others combined removal of the retina and
the pineal is necessary to eliminate circadian rhythmicity. All
rhythms may not be equally affected, at least in starlings (Sturnus
vulgaris), leading to some ambiguity as to the interpretation of these
lesion effects. In Japanese quail (Coturnix coturnix japonica), the
circadian oscillator in the eyes is necessary for normal rhythmicity.
In most sauropsid species for which data are available, including the
house sparrow (Passer domesticus), chicken (Gallus gallus), and
anolis lizard (Anolis carolinensis), the isolated pineal maintains a
circadian oscillation in vitro, whereas in mammals, the pineal cannot
oscillate without rhythmic input provided by the SCN. In the desert
iguana (Dipsosaurus dorsalisa), as in mammals, only SCN ablations
disrupt circadian rhythms, and the pineal does not oscillate in vitro.
However, in another species, the green iguana (Iguana iguana), the
pineal, retinas, and parietal eye all oscillate robustly in vitro, and an
intact pineal is essential for circadian thermoregulatory rhythms, but
not for activity or ERG rhythms (Tosini & Menaker 1998). The
sauropsid circadian system is thus characterized by a more
distributed circadian regulatory system and by marked species
variability in the role of its oscillatory and photoreceptive elements.
Mutual coupling of the primary oscillators in the retina,
hypothalamus, and pineal, possibly by an endocrine factor such as
melatonin, may serve to amplify and increase the precision of the
aggregate circadian signal. The ecological significance of species
variability in the details of sauropsid circadian organization remains
to be explored.
Less is known about the neural and endocrine bases of circadian
timekeeping in amphibians and fish, but the available evidence
suggests an important role for the pineal as photoreceptor and
pacemaker in the few species studied. In the clawed frog, Xenopus,
retinal photoreceptors maintain circadian oscillations in vitro. In the
hagfish (Eptatretus stouti), ablation of the hypothalamus eliminates
circadian rhythms, but the site of the circadian clock within the
hypothalamus is uncertain. The zebrafish has become a valuable
model for molecular analysis of the circadian clock in fish (Idda et al.
2012). A unique feature of this species is that self-sustaining
circadian oscillators are found in every organ, and each of these is
directly entrainable by light, in vivo and in tissue cultures. Among
vertebrates, the zebrafish may represent an extreme of decentralized
circadian control.

Seasonal Rhythmicity

In many habitats, day length, temperature, and precipitation change

markedly with season, altering the availability of food, cover, and
shelter, as well as the activity of resident predators, competitors, and
parasites. These annual cycles present formidable challenges for
survival and reproduction. One solution is seasonal migration to a
more hospitable environment during the winter or dry season and
return migration at an appropriate time.
Another solution is to remain in place but undergo seasonal changes
in behavior (e.g., reproduction, ingestion, nesting, activity; Figure
4.4C), physiology (e.g., metabolism, reproduction), and morphology
(e.g., fur or feather density and color) that facilitate survival and that
restrict the birth of offspring to seasons optimal for survival. Both
solutions require that behavioral and physiological adjustments
begin well in advance of the target season. To anticipate the change
of season, most animals measure daylength and in some cases use
daylength to entrain an endogenous circannual clock.
Migration often entails dramatic changes in daily behavioral
patterns. Birds that fly continuously over water for days during their
seasonal migrations reduce their sleep amounts dramatically (e.g.,
the great frigate bird, Fregata minor; Rattenborg et al. 2016). When
normally migratory species are studied in captive conditions and
exposed to photoperiod changes during the migration season, they
show characteristic increases in activity (Zugunruhe or migratory
restlessness) oriented toward the direction of that season’s usual
migration (Kramer 1950). This activity is often strongly concentrated
in the night, replacing the sleep shown at that phase in other seasons
by species such as the black-headed bunting (Emberiza
melanocephala) and white-crowned sparrow (Zonotrichia
leucophrys gambelii) (Rattenborg et al. 2004; Rastogi et al. 2011).
Daily sleep and activity patterns undergo remarkable modifications
to meet the demands of seasonal migrations in many bird species.
But limited seasonal breeding opportunities that form part of this
annual cycle also modify daily rhythms. Male pectoral sandpipers
(Calidris melanotos) breed in the Arctic. During their short breeding
season, they strongly curtail sleep in favor of courtship and mating
opportunities. Those with the longest active periods and largest
reductions in sleep sire the most offspring (Lesku et al. 2012).

Circannual clocks
Seasonal rhythms driven by a circannual clock are those that persist
for one or more years in constant environments. Circannual clocks
have been demonstrated in a variety of longer-lived species,
including ground squirrels, marmots, deer, sheep, and some species
of bats and primates. Although generated endogenously, circannual
rhythms, like circadian rhythms, are entrained by environmental
changes in light exposure. In all species studied, annual changes in
daylength are the primary, if not exclusive, circannual Zeitgeber. In
some species, entrainment may require a gradual change in
photoperiod, as occurs in the natural habitat.
The location of the circannual clock is not known for any species
(Paul et al. 2008). Ablation studies have ruled out a necessary role
for the SCN circadian clock in most species, although in experiments
with ground squirrels, a minority of animals with SCN ablations
failed to exhibit circannual rhythms in body mass. Under some
conditions, the circannual cycle of hibernation is also disrupted by
SCN damage.
Although the pineal gland is also not necessary for the generation of
circannual rhythms, it is necessary for their entrainment by
photoperiod. The pineal gland secretes the hormone melatonin at
night, under the control of the SCN circadian pacemaker. Neural
activity in the SCN increases during the daytime (in both diurnal and
nocturnal animals) and the duration of this increased activity is
proportional to daylength. Pineal melatonin synthesis is also
inhibited by light (in mammals via the SCN). The duration of
melatonin secretion each night thus varies with the length of the day
and serves as an endogenous signal of season, when combined with
information about the direction of change in duration. Surgical
removal of the pineal abolishes entrainment of circannual rhythms.

Photoperiodism and interval timers

Annual rhythms in most species studied to date do not involve a fully
autonomous circannual clock but depend on seasonal changes in
photoperiod for one of more of the annual physiological transitions
(Paul et al. 2008). Shortening daylengths in the fall induces
behavioral and physiological responses appropriate to autumn and
winter, whereas lengthening days induces responses appropriate to
spring and summer. However, the critical daylength for eliciting a
physiological change of seasons varies by species, by the particular
variable being measured, and by the animals’ recent photoperiod
history. History dependence is important because, apart from the
annual extremes, most daylengths are ambiguous as to season, given
that they occur twice a year, once before and once after the summer
solstice. Consequently, the response to an intermediate photoperiod
often depends on whether it is longer or shorter than the prior
photoperiod. In rare cases, winter responses may occur only within a
narrow range of photoperiods, with summer responses stimulated by
longer or shorter days.
In some species, the winter state reverses spontaneously after several
months of continuous exposure to short days. In a well-studied
model system, the Syrian hamster, the gonads regress and
reproductive behaviors cease in response to decreasing daylength,
but within 20–25 weeks, the gonads spontaneously recrudesce
(regrow) even if they are maintained in short days or continuous
darkness, presumably in preparation for the end of winter and the
onset of the breeding season. For several months after
recrudescence, hamsters are unresponsive to short days.
Photosensitivity is only restored by several weeks of exposure to long
days. The annual cycle therefore appears to be regulated by an
endogenous interval timer that is triggered by short days and reset
by long days.
The circadian clock plays an essential role in photoperiodic time
measurement, due to its intrinsic circadian rhythm of sensitivity to
light. The daily pattern of light exposure differs in summer and
winter, stimulating different phases of the circadian clock. Even very
brief pulses of light, timed to occur at certain circadian phases, can
trigger long-day physiological responses. Thus, season is sensed by
the timing of light relative to a circadian template, but there are
alternative hypotheses about the exact mechanisms involved in
daylength measurement. A consensus view is that daylength affects
circadian cycling of the clock and that changes in clock output signals
affect neuroendocrine regulators of seasonal responses, such as the
duration of nocturnal melatonin secretion.
The issue of human photoperiodic responses has attracted
considerable attention, in part because of the psychiatric syndrome
seasonal affective disorder (SAD), which is a regularly recurring
seasonal (usually winter) depression (Danilenko & Levitan 2012).
The fact that in many patients this condition can be treated
successfully with timed brief light exposure that effectively extends
the light phase of a short winter day has prompted an exploration of
whether it is based on photoperiod. There is evidence that supports
the idea that the circadian profile of melatonin secretion differs
between SAD patients and normal controls in the winter, but not
when patients are asymptomatic in the summer. This evidence
reinforces the conclusion from a review of many years of birth
records that humans express at least a rudimentary seasonal
rhythmicity. The degree to which this rhythmicity has been modified
in industrial societies by artificial regulation of light and
temperature, and the proximal cues that regulate seasonality, remain
open questions.

Ultradian Rhythmicity

Ultradian rhythms can be defined as those that complete two or

more cycles per circadian day. The definition is broad and the cutoffs
arbitrary because, unlike circadian and circannual rhythms,
ultradian rhythms do not correspond to any geophysical cycle in the
environment. Ultradian rhythms can be further classified as those
that are controlled by interval timers and those that are controlled by
self-sustaining oscillators.
Interval timers can express a rhythm only if they are reset
periodically. An example is a sand-filled hourglass timer. The
duration of a single cycle is set by the length of time it takes for sand
to run out, and the cycle will not be repeated unless the hourglass is
reset. Ultradian rhythms in feeding and locomotor activity evident in
many organisms likely reflect the rate at which ingested nutrients are
absorbed and metabolized, an hourglass process reset by each major
feeding event.
In some rodent species, such as voles (Figure 4.4C), ultradian cycles
of activity have been shown to persist even if food is withheld,
indicating that these rhythms are generated by a true ultradian
oscillator (Gerkema et al. 1993). The physical location and cellular
mechanisms of such ultradian oscillators remain to be established.
From a functional perspective, it has been hypothesized that
synchrony of ultradian rhythms of foraging within a local population
of animals vulnerable to predation may reduce both the probability
of being killed (safety in numbers) and the total number of kills
(swamping the predator’s capacity; Daan 1981).
Ultradian rhythms are also evident in sleeping organisms. In normal
adult humans, sleep progresses through three nonrapid eye
movement (NREM) sleep stages, followed by a rapid eye movement
(REM) sleep episode. This cycle is repeated at ~ 90–120 minutes
intervals through a typical 7–8 hour sleep period. The cycle is
modulated by circadian phase, such that the duration of the REM
sleep episode is greatest near the nocturnal body temperature
minimum (late in a typical night’s sleep). The latency to the first
REM bout is also shorter when sleep is initiated closer to the
temperature minimum, but sleep onset REM episodes in adults are
rare and are normally symptomatic of a sleep disorder. Thus, the
NREM–REM cycle reflects an ultradian process that is reset by sleep
onset but modulated by circadian phase.
Discovery of the ~90 minute sleep cycle stimulated the search for a
comparable periodicity in waking activities. Numerous behavioral
and physiological variables have since been shown to exhibit
rhythmicity in the 90 minute range during waking, leading to the
hypothesis of a “basic rest–activity cycle” manifest in
neurobehavioral functions during waking and in the NREM–REM
cycle during sleep (Kleitman 1963, 1993). However, the evidence for
continuity between waking 90 minute cycles and the sleep cycle is
weak, possibly because the sleep stage cycle is clearly reset at sleep
onset.

SUMMARY AND CONCLUSIONS

Under usual environmental conditions, a large proportion of the
variance in animal behavior can be accounted for by a stable
behavioral program regulated by ultradian, circadian, and
circannual oscillators and timers. The precision of circadian
timekeeping is such that the activities of bees and flying squirrels
can be predicted to within minutes from knowledge of the timing
of their rhythms during the previous circadian cycle. Human
behavior is similarly predictable; the probability of sleep or wake
onset is closely related to circadian phase, as are appetite and
cognitive and physical performance. Neurobehavioral analysis of
endogenous rhythmicity is central to understanding the temporal
dimension of animal behavior and the human condition.
FURTHER READING
As of January 29, 2021, the USA National Library of Medicine search
engine retrieves 93,504 scholarly publications from the search term
“circadian rhythm.” For help navigating the literature on circadian
and other biological rhythms in humans and other animals, a
number of edited trade books and textbooks are available.
Chronobiology as an identifiable field of scientific inquiry can
arguably be traced to the landmark Cold Spring Harbor Symposium
on Quantitative Biology (vol. 25) held in 1960, which summarized
empirical findings and general principles of biological rhythms to
that date. Comprehensive updates include two volumes of the
Handbook of Behavioral Neurobiology (vol 4, 1981, J. Aschoff, ed.;
vol. 12, 2002, J.S. Takahashi et al., eds.) and the Cold Spring Harbor
Symposium on Quantitative Biology (vol. 72) published in 2007.
Textbooks of note include Chronobiology: Biological Timekeeping
(J. Dunlap et al. 2004), Circadian Medicine (C.S. Colwell, 2015) and
Circadian Physiology, 3rd edition (R. Refinetti 2016). Excellent
introductions to the field suitable for general readers include
Rhythms of Life (R.G. Foster & L. Kreitzman, 2005), Circadian
Rhythms: A Very Short Introduction (R. Foster & L. Kreitzman,
2017) and Internal Time (T. Roenneberg 2012). For an introduction
to annual rhythms, see Seasons of Life (R.G. Foster & L. Kreitzman
2009).

REFERENCES
Balsalobre, A. 2001. Clock genes in mammalian peripheral tissues.
Cell and Tissue Research, 309, 193–9.
Beersma, D.G., Daan, S. & Hut, R.A. 1999. Accuracy of circadian
entrainment under fluctuating light conditions: contributions of
phase and period responses. Journal of Biological Rhythms,
14(4), 320–9. doi:10.1177/074873099129000740.
Bloch, G., Barnes, B.M., Gerkema, M.P. & Helm, B. 2013. Animal
activity around the clock with no overt circadian rhythms:
patterns, mechanisms and adaptive value. Proceedings of the
Royal Society B. Biological Sciences, 280(1765), 20130019.
Buhr, E.D. & Takahashi, J.S. 2013. Molecular components of the
Mammalian circadian clock. Handbook of Experimental
Pharmacology, 217, 3–27.
Carskadon, M.A., Acebo, C. & Jenni, O.G. 2004. Regulation of
adolescent sleep: implications for behavior. E.P. Annals of the
New York Academy of Science, 1021, 276–91.
Cassone, V.M. 2014. Avian circadian organization: a chorus of clocks.
Frontiers in Neuroendocrinology, 25(1), 76–88.
Castillo-Ruiz, A., Paul, M.J. & Schwartz, W.J. 2012. In search of a
temporal niche: social interactions. Ch. 16. In: A. Kalsbeek, M.
Merrow, T. Roenneberg & R.G. Foster (eds.), Progress in Brain
Research, vol. 199, pp. 267–80. Elsevier.
Czeisler, C.A. & Dijk, D.-J. 2001. Human circadian physiuology and
sleep-wake regulation. In: J.S. Takahaski, F.W. Turek & R.Y.
Moore (eds.), Circadian Clocks. Handbook of Behavioral
Neurobiology, vol. 12, pp. 531–69. New York: Kluwer
Academic/Plenum.
Czeisler, C.A., Duffy, J.F., Shanahan, T.L. et al. 1999. Stability,
precision, and near-24-hour period of the human circadian
pacemaker. Science, 284(5423), 2177–81.
Czeisler, C.A., Weitzman, E.D., Moore-Ede, M.C., Zimmerman, J.C.
& Knauer, R.S. 1980. Human sleep: its duration and organization
depend on its circadian phase. Science, 210(4475), 1264–7.
Daan, S. 1981. Adaptive daily strategies in behavior. In: J. Aschoff
(ed.), Handbook of Behavioral Neurobiology 4: Biological
Rhythms, pp. 275–98. New York: Plenum.
Danilenko, K.V. & Levitan, R.D. 2012. Seasonal affective disorder.
Handbook of Clinical Neurology, 106, 279–89.
Davidson, A.J., Sellix, M.T., Daniel, J., Yamazaki, S., Menaker, M. &
Block, G.D. 2006. Chronic jet-lag increases mortality in aged
mice. Current Biology, 16, R914–6.
DeCoursey, P.J. 1986. Light-sampling behavior in photoentrainment
of a rodent circadian rhythm. Journal of Comparative Physiology
A, 159(2), 161–9. doi:10.1007/BF00612299.
DeCoursey, P.J. 2014. Survival value of suprachiasmatic nuclei
(SCN) in four wild sciurid rodents. Behavioral Neuroscience,
128(3), 240–9.
Dijk, D.-J. & Von Schantz, M. 2005. Timing and consolidation of
human sleep, wakefulness, and performance by a symphony of
oscillators. Journal of Biological Rhythms, 20, 279–90.
Edgar, R.S., Green, E.W., Zhao, Y. et al. 2012. Peroxiredoxins are
conserved markers of circadian rhythms. Nature, 485(7399),
459–64.
Gerkema, M.P., Daan, S., Wilbrink, M., Hop, M.W. & Van Der Leest,
F. 1993. Phase control of ultradian feeding rhythms in the
common vole (Microtus arvalis): the roles of light and the
circadian system. Journal of Biological Rhythms, 8, 151–71.
Guilding, C. & Piggins, H.D. 2007. Challenging the omnipotence of
the suprachiasmatic timekeeper: are circadian oscillators present
throughout the mammalian brain? European Journal of
Neuroscience, 25, 3195–3216.
Gwinner, E. & Brandstatter, R. 2001. Complex bird clocks.
Philosophical Transactions of the Royal Society of London B,
356, 1801–10.
Hut, R.A. 1999. Natural entrainment of circadian systems: A study of
the diurnal ground squirrel (Spermophilus citellus). Doctoral
dissertation, University of Groningen.
Hut, R.A., Kronfeld-Schor, N., Van Der Vinne, V. & De La Iglesia, H.
2012. In search of a temporal niche: Environmental factors. In: A.
Kalsbeek, M. Merrow, T. Roenneberg & R.G. Foster (eds.),
Progress in Brain Research, vol. 199, Ch. 17, pp. 281–304.
Elsevier.
Hut, R.A., Pilorz, V., Boerema, A.S., Strijkstra, A.M. & Daan, S. 2011.
Working for food shifts nocturnal mouse activity into the day.
 PLoS One, 6(3), e17527.
Hut, R.A., Van Oort, B.E. & Daan, S. 1999. Natural entrainment
without dawn and dusk: the case of the European ground squirrel
(Spermophilus citellus). Journal of Biological Rhythms, 14(4),
290–9. doi:10.1177/074873099129000704.
Idda, M.L., Bertolucci, C., Vallone, D. et al. 2012. Circadian clocks:
lessons from fish. Progress in Brain Research, 199, 41–57.
Klein, D.C., Moore, R.Y. & Reppert, S.M. 1991. Suprachiasmatic
Nucleus: The Mind’s Clock. New York: Oxford University Press.
Kleitman, N. 1963. Sleep and wakefulness. Chicago: University of
Chicago Press.
Kleitman, N. 1993. Basic rest-activity cycle. In: M.A. Carskadon (ed.),
Encyclopedia of Sleep and Dreaming, pp. 65–66. New York:
Macmillan.
Kramer, G. 1950. Orientierte Zugaktivität gekäfigter Singvögel. Die
Naturwissenschaften, 37, 188.
Lesku, J.A., Rattenborg, N.C., Valcu, M. et al. 2012. Adaptive sleep
loss in polygynous pectoral sandpipers. Science, 337, 1654–8.
Lu, W., Meng, Q.J., Tyler, N.J.C., Stokkan, K.A. & Loudon, A.S.I.
2010. A circadian clock is not required in an arctic mammal.
Current Biology, 20, 533–7.
Micic, G., Lovato, N., Gradisar, M. et al. 2016. Circadian melatonin
and temperature taus in delayed sleep-wake phase disorder and
non-24-hour sleep-wake rhythm disorder patients: an ultradian
constant routine study. Journal of Biological Rhythms, 31, 387–
405.
Mistlberger, R.E. 2009. Entrainment of circadian rhythms by food:
concepts and methods. European Journal of Neuroscience, 30(9),
1718–29.
Mistlberger, R.E. 2011. Neurobiology of food anticipatory circadian
rhythms. Physiology and Behavior, 104, 535–45.
Mistlberger, R.E. & Marchant, E.G. 1995. Computational and
entrainment models of circadian food-anticipatory activity:
Evidence from non-24 h feeding schedules. Behavioral
Neuroscience, 109, 790–8.
Mistlberger, R.E. & Skene, D.J. 2004. Social influences on circadian
rhythms in mammals: animal and human studies. Biological
Reviews, 79(3), 1–23.
Mohawk, J., Green, C.B. & Takahashi, J.S. 2012. Central and
peripheral circadian clocks in mammals. Annual Review of
Neuroscience, 35, 445–62.
Moore-Ede, C.A., Sulzman, F.M. & Fuller, C.A. 1982. The Clocks That
Time Us: Physiology of the Circadian Timing System.
Cambridge, MA: Harvard University Press.
Moore-Ede, M.C. 1986. Physiology of the circadian timing system:
predictive versus reactive homeostasis. American Journal of
Physiology, 250, 737–52.
Mrosovsky, N. 1999. Masking: history, definitions, and
measurement. Chronobiology International, 16, 415–29.
Mulder, C.K., Gerkema, M.P. & Van Der Zee, E.A. 2013. Circadian
clocks and memory: time-place learning. Frontiers in Molecular
Neuroscience, 6, article 8.
O’Neill, J.S., Maywood, E.S. & Hastings, M.H. 2013. Cellular
mechanisms of circadian pacemaking: beyond transcriptional
loops. Handbook of Experimental Pharmacology, 217, 67–103.
Patke, A., Young, M.W. & Axelrod, S. 2020. Molecular mechanisms
and physiological importance of circadian rhythms. Nature
Reviews Molecular and Cell Biology, 21(2), 67–84.
doi:10.1038/s41580-019-0179-2.
Paul, M.J., Zucker, I. & Schwartz, W.J. 2008. Tracking the seasons:
the internal calendars of vertebrates. Philosophical Transactions
of the Royal Society of London B, 363(1490), 341–61.
Pittendrigh, C.S. 1993. Temporal organization: reflections of a
Darwinian clock-watcher. Annual Review of Physiology, 55, 16–
54.
Pittendrigh, C.S. & Daan, S. 1974. Circadian oscillations in rodents: a
systematic increase of their frequency with age. Science, 186,
548–50.
Pittendrigh, C.S. & Daan, S. 1976. A functional analysis of circadian
pacemakers in nocturnal rodents: IV Entrainment: Pacemaker as
clock. Journal of Comparative Physiology A, 106, 291–331.
Ralph, M.R. & Lehman, M.N. 1991. Transplantation: a new tool in
the analysis of the mammalian hypothalamic circadian
pacemaker. Trends in Neuroscience, 14, 362–66.
Rastogi, A., Kumari, Y., Rani, S. & Kumar, V. 2011. Phase inversion
of neural activity in the olfactory and visual systems of a night-
migratory bird during migration. European Journal of
Neuroscience, 34, 99–109.
Rattenborg, N.C., Mandt, B.H., Obermeyer, W.H. et al. 2004.
Migratory sleeplessness in the white-crowned sparrow
(Zonotrichia leucophrys gambelii). PLoS Biology, 2, 924–36.
Rattenborg, N.C., Voirin, B., Cruz, S.M. et al. 2016. Evidence that
birds sleep in mid-flight. Nature Communications, 7, 12468.
Roenneberg, T., Kantermann, T., Juda, M., Vetter, C. & Allebrandt,
K.V. 2013. Light and the human circadian clock. Handbook of
Experimental Pharmacology, 217, 311–31. doi:10.1007/978-3-
642-25950-0_13.
Roenneberg, T., Kuehnle, T., Pramstaller, P.P. et al. 2004. A marker
for the end of adolescence. Current Biology, 14(24), R1038–9.
Rowsmitt, C.N. 1986. Seasonal variations in activity rhythms of male
voles: mediation by gonadal hormones. Physiology and Behavior,
37, 797–803.
Scheibler, E., Wollnik, F., Brodbeck, D. et al. 2013. Species
composition and interspecific behavior affects activity pattern of
 free-living desert hamsters in the Alashan Desert. Journal of
Mammalogy, 94, 448–58.
Stratmann, M. & Schibler, U. 2006. Properties, entrainment, and
physiological functions of mammalian peripheral oscillators.
Journal of Biological Rhythms, 21(6), 494–506.
Tosini, G. & Menaker, M. 1996. Circadian rhythms in cultured
mammalian retina. Science, 272, 419–21.
Van Oort, B.E., Tyler, N.J., Gerkema, M.P. et al. 2005. Circadian
organization in reindeer. Nature, 438(7071), 1095–6.
Vanin, S., Bhutani, S., Montelli, S. et al. 2012. Unexpected features of
Drosophila circadian behavioural rhythms under natural
conditions. Nature, 484, 371–5.
Voirin, B., Scriba, M.F., Martinez-Gonzalez, D. et al. 2014. Ecology
and neurophysiology of sleep in two wild sloth species. Sleep, 37,
753–61.
Wallraff, H.G. 1981. Clock-controlled orientation in space. In: J.
Aschoff (ed.), Handbook of Behavioral Neurobiology, vol. 4, pp.
275–98. New York: Plenum.
Weaver, D.R. 1998. The suprachiasmatic nucleus: a 25-year
retrospective. Journal of Biological Rhythms, 13, 100–12.
Webb, I.C., Antle, M.C. & Mistlberger, R.E. 2014. Regulation of
circadian rhythms in mammals by behavioral arousal. Behavioral
Neuroscience, 128(3), 304–25.
Welsh, D.K., Logothetis, D.E., Meister, M. & Reppert, S.M. 1995.
Individual neurons dissociated from rat suprachiasmatic nucleus
express independently phased circadian firing rhythms. Neuron,
14, 697–706.
Yamada, N., Shimoda, K., Ohi, K., Takahashi, S. & Takahashi, K.
1988. Free-access to a running wheel shortens the period of free-
running rhythm in blinded rats. Physiology and Behavior, 42,
87–91.
Yamazaki, S., Numano, R., Abe, M. et al. 2000. Resetting central and
peripheral circadian oscillators in transgenic rats. Science,
288(5466), 682–5.
5
brain and behavior
DAVID F. SHERRY

INTRODUCTION
The brains of animals do many things. For the purposes of this
chapter the things brains do are organized into three broad
categories. Brains obtain information from the environment, they
carry out cognitive operations, and they control movement:
input, central processing, and output. These categories are broad
because although they are certainly the conventional way of
thinking about the brain and behavior, from the beginning of
systematic work on the topic (Lashley 1929) to present-day
research (Poeppel et al. 2020), it is often not possible to find a
clear demarcation between, say, receiving sensory input from the
environment and performing cognitive operations on that input.
The “cognitive” category itself is a very broad one, including
everything from memory to spatial orientation, timing, decision-
making, and social interactions. These categories are therefore
more an organizational convenience that corresponds to how we
think about the things a brain does than how the nervous system
actually divides up the tasks it performs. With this in mind, let us
begin with the first category, how animals obtain information
about their environment.

Sensing and Perceiving the Environment

The auditory world of the barn owl (Tyto alba)

Just below the eaves of an abandoned farmhouse, a barn owl pokes
its full-moon face through a gap in weathered grey boards. It
hunches forward and launches itself into the darkness. It will spend
the night flying in and out of a nearby orchard, tracing long loops
over meadows and pastures in search of prey. Its flight is almost
silent. It can hunt successfully on the darkest nights because it finds
its prey, mostly mice, voles, and shrews, by sound. The sounds the
barn owl detects are faint, quiet as a mouse in fact, but its sensitive
hearing pinpoints sound sources with such accuracy that it can
swoop even in complete darkness and fly off with prey in its talons.
How the barn owl brain localizes environmental sounds provides our
first illustration of neural processing of the sensory world (Figure
5.1).
Figure 5.1 The facial ruff of the barn owl. Auricular feathers in the
middle of the ruff (white circle) are transparent to sound. Reflector
feathers at the border of the ruff (white triangle) direct sound toward
the ears. From Wagner et al. (2013).

The barn owl’s auditory map of space

All sound, like the rustling of a mouse in dry grass, comes from a
location in space. The question is: how does the auditory system
determine where the sound is coming from? This is not a new
question. Lord Rayleigh addressed it and in general terms solved it in
the early 1900s (Rayleigh 1907). The key is that the two ears,
whether owl ears or human ears, placed on opposite sides of the
head, hear the same sound slightly differently. The auditory system
combines the signals received by the two ears to produce the
perception of a single sound source at a fixed location, a process
called binaural fusion. Stereo recordings exploit binaural fusion to
create the illusion of different instruments and voices occupying
different places in the space between two stereo speakers.
When sound from a single source in the environment reaches the
ears of the barn owl, the sounds received by the two ears are not the
same. There is a difference in the timing of the sound at the two ears,
and there is a difference in the intensity of the sound. These are,
respectively, the interaural time difference (ITD) and the interaural
level difference (ILD). The time difference is a consequence of the
spacing between the ears: an expanding wave of sound will reach one
side of the head before it reaches the other. The intensity difference
occurs because diffraction of sound by the head reduces the
intensity at one ear compared to the other. The structure of barn owl
ears also exaggerates intensity differences because the ears are
asymmetric. The right ear is located below eye level and is directed
upward, whereas the left ear is located above eye level and is directed
downward. A sound coming from above eye level, for example, will
be perceived as louder in the right ear than in the left. Timing
differences and intensity differences between the two ears are
handled by two separate processing streams in the barn owl auditory
system. Input from each ear is processed first as separate time and
intensity signals, then combined to detect timing and intensity
differences between the two ears, before being finally assembled into
an auditory map of space.

Two auditory processing streams

In the abstract, the problem of locating a sound in space is the
problem of specifying the coordinates of a point on an imaginary
sphere centered on the head of the listener. One of these coordinates
is the azimuth, the horizontal position on the equator of the sphere,
and the other is the elevation, the height above the equator. In the
barn owl, one processing stream is dedicated to each coordinate.
ITDs are almost perfectly correlated with the azimuth of a sound
source, whereas ILDs are correlated with its elevation (Figure 5.2).
Konishi and his colleagues confirmed this by recording sound from
tiny microphones placed in the ears of barn owls (Moiseff & Konishi
1981). As they moved a speaker in a predetermined pattern around
the barn owl’s head, ITDs varied systematically with the azimuth of
the speaker, whereas ILDs varied systematically with its vertical
position. By attaching two coils of copper wire to the barn owl’s head
at right angles and placing the barn owl in a larger array of coils that
produced a magnetic field, they were able to record the orientation of
the barn owl’s head as the location of the speaker changed (Moiseff
1989). The eyes of barn owls are tubular and move very little in the
head. The owls instead move their entire head to direct attention to a
location in space. Barn owls turned their heads very accurately to
face the speaker when it produced a sound. Furthermore, an owl will
do the same to a phantom sound source created by delivering sounds
to tiny earphones placed in the owl’s ears (Moiseff & Konishi 1981).
Varying interaural time and intensity differences between the
earphones caused the owl to turn and face the location in space that
would correspond to a sound that produced the same ITD and ILD.
Figure 5.2 Acoustic measurements taken from microphones in a
barn owl’s ears show how ITD (measured in μs) and ILD (measured
in dB) vary with azimuth and elevation (a and b). Azimuth of 0° is
directly in front of the owl. Mean values for nine barn owls taken at
0° elevation are shown in c and d. Shading in c and d shows standard
deviations; circles and triangles in d show two individual owls. From
Kettler et al. (2017).
Sounds are converted to neural signals by auditory receptors in the
basilar membrane of the cochlea, a small spiral chamber inside the
ear. The site of stimulation on the basilar membrane of the cochlea
corresponds to the frequency of the sound, high-frequency sounds
stimulating receptors at the proximal end, nearest the tympanum or
eardrum, and low-frequency sounds stimulating receptors at the
distal end. Movement of the basilar membrane causes
neurotransmitter release in hair cells, which causes excitation of the
primary auditory neurons. The firing patterns of primary auditory
fibers in response to movement of the basilar membrane encode both
sound intensity and frequency. Sound is a waveform and auditory
neurons convey frequency information by firing at a particular
phase angle of the sound wave (Figure 5.3a). Such cells are said to
be phase-locked. Auditory neurons responsive to the same
frequency are locked to the same phase angle of the sound wave. The
frequencies of most sounds are much higher than the frequency at
which neurons can fire. A 10-kHz pure tone oscillates 10,000 times
per second whereas few neurons can fire faster than several hundred
times per second. So not all phase-locked neurons fire at every cycle
of the sound, but when they do fire, they fire at the same phase angle
of the sound wave.
Figure 5.3 (a) Auditory neurons and neurons in the nucleus
magnocellularis of the barn owl brain code sound frequency by firing
at a particular phase angle of the sound wave. The phase angle of a
periodic wave is an angular value ranging from 0 to 360°. The
illustration shows a neuron that fires only at the 90° phase angle of a
sound wave. Because of the neuron’s refractory period, it does not
fire at every 90° phase angle. The neuron is phase-locked and
neurons sensitive to the same frequency are locked to the same phase
angle. (b) Schematic diagram of coincidence detecting neurons in the
left and right nucleus laminaris (NL). Sounds at different locations
on the azimuth (coloured speakers) arrive at the left and right ears
with different ITDs. Phase-locked neurons in the nucleus
magnocellularis (NM) relay the interaural phase difference to NL
where the characteristic delay between the ears represents a location
in space. From Seidl et al. (2010). (c) A coronal section through the
dorsal brainstem shows axonal projections from the right nucleus
magnocellularis to the ipsilateral and contralateral nucleus laminaris
(NL). The left nucleus magnocellularis (NM) sends similar axonal
projections to its ipsilateral and contralateral nucleus laminaris.
Phase-locked signals from the two ears converge in this way on
neurons in NL. The path from the ipsilateral nucleus magnocellularis
acts as a delay line and neurons in the nucleus laminaris act as
coincidence detectors. Scale bar = 1 mm. From Carr & Konishi
(1990). Copyright 1990 Society for Neuroscience. (d) Timing and
intensity differences between the ears are relayed by different
neuroanatomical paths to the inferior colliculus. This schematic
drawing omits reciprocal connections between the two sides of the
brain as well as other nuclei involved in auditory processing. VLVp,
nucleus ventralis lemnisci lateralis, pars posterior.
The primary auditory fibers project to the first sound-processing site
in the barn owl brain, the cochlear nucleus magnocellularis, located
in the brainstem. It is in this nucleus that the temporal properties of
the monaural signal from each ear are extracted. Recordings from
single neurons in this nucleus show that these cells fire with the
same phase-locked pattern found in the auditory fibers but do not
respond differentially to the intensity of the auditory signal.
Signals from the ipsilateral and contralateral nucleus
magnocellularis converge at the next step in the time-processing
pathway, the nucleus laminaris, also located in the brainstem.
Because the signals from the two ears are out of phase unless the
sound is directly in front of or behind the head (Figure 5.3b), the
interaural phase difference (IPD) indicates where the sound source is
located on the azimuth (Moiseff & Konishi 1983; Carr & Konishi
1990). Neurons in the nucleus laminaris are maximally responsive to
phase-locked signals from the two ears that arrive in synchrony. That
is, these neurons act as coincidence detectors, firing when stimulated
simultaneously by signals from the left and right nucleus
magnocellularis that are in phase, that is, with IPD = 0. Some
mechanism is therefore required to bring the two phase-locked
signals into register. Jeffress (1948) proposed a simple model of how
this might occur, and in the barn owl brain Jeffress’s model is
implemented by a remarkably simple arrangement: the conduction
velocities of axons from the contralateral and ipsilateral nucleus
magnocellularis are different (Figure 5.3c). The contralateral path
crosses the midline between the hemispheres and goes directly to the
ventral portion of the nucleus laminaris. The ipsilateral path
terminates in the dorsal portion of the nucleus laminaris, but
conduction velocity is slowed by smaller axon diameters and more
closely spaced nodes of Ranvier at various points in this path (Seidl
et al. 2010). The path from the ipsilateral nucleus magnocellularis
acts as a delay line, slowing the time of arrival of its phase-locked
signal. The contralateral and ipsilateral axons travel through the
nucleus laminaris, intermingling and terminating on large laminaris
cells. At some laminaris cells, the phase-locked signals from the two
ears arrive in register, and the identity of the cells that fire indicates
the phase difference between the auditory signals received by the left
and right ears. Because of the delay in projections from the ear on
the same side of the brain, laminaris cells triggered by signals that
are in phase are in fact responding to a particular IPD between the
left and right ears. The nucleus laminaris is organized
tonotopically; different parts of the nucleus receive input encoding
different frequencies of sound. In this way, the neuroanatomical
address of laminaris cells stimulated by signals arriving from the two
ears in phase encodes IPD over the range of frequencies that make
up the sound (Peña et al. 2001).
Intensity differences between the two ears are handled by a different
processing stream that determines the elevation of the sound source.
The cochlear nucleus angularis contains neurons sensitive to
intensity signals carried by the auditory fibers (see Figure 5.3d). A
nucleus in the pons, the nucleus ventralis lemnisci lateralis pars
posterior (VLVp), receives excitatory input from the contralateral
nucleus angularis and inhibitory input from the contralateral VLVp.
VLVp neurons thus receive an excitatory signal conveying loudness
from the opposite ear and an inhibitory signal conveying loudness
from the ear on the same side of the brain that has been relayed
through the contralateral VLVp. VLVp cells respond selectively to
particular ILDs, with a topographical representation of loudness
differences in the VLVp similar to the topographical representation
of time differences in the nucleus laminaris. Neurons in the ventral
part of the VLVp respond more strongly when the sound is louder in
the ear on the same side of the brain. Neurons in the dorsal part
respond more strongly when the sound is loudest in the opposite ear
(Konishi 1995).

The auditory map

The time- and intensity-processing streams eventually converge in
the midbrain in the external nucleus of the inferior colliculus (Figure
5.3d). Cells in this nucleus are selectively sensitive to different
combinations of ITDs and ILDs that specify the azimuth and
elevation of the sound source. As described earlier, barn owls will
turn their head toward a sound in their environment and will also
turn their head toward phantom sound sources created by presenting
combinations of ITDs and ILDs to the two ears using tiny earphones.
In this way, Konishi and his colleagues tested hypotheses about how
time and intensity differences between the two ears are interpreted
by the owl. Using the same technique for sound presentation, but
with anesthetized owls, they were able to find neurons in the inferior
colliculus sensitive to particular combinations of time and intensity
differences. These cells have as their receptive fields combinations
of azimuth and elevation that represent locations on the imaginary
sphere surrounding the owl’s head: neurons that fire only when a
sound comes from a particular point in space. Selective inactivation
of time and intensity inputs to the external nucleus of the inferior
colliculus by injection of minute amounts of local anesthetic into the
nucleus magnocellularis and nucleus angularis, respectively,
confirmed that these two cochlear nuclei were the origin of azimuth
and elevation inputs (Takahashi et al. 1984). The space-specific
neurons of the inferior colliculus in turn project to motor areas
controlling the owl’s head-turning and to the optic tectum, where a
further auditory and visual representation of space is assembled.
ITD alone, however, cannot fully account for the barn owl’s ability to
localize a sound source on the azimuth. This is because sounds at
different points on the azimuth can produce the same ITD, causing
the owl to perceive both true and phantom sound sources (Kettler et
al. 2017). On each side of the head there is a “cone of confusion” with
its apex at the midpoint of the head and its long axis extending along
a line passing through both ears: all points on the surface of this cone
have the same ITD (Wagner et al. 2013). For barn owls, ILD resolves
most of the ambiguity on the cone of confusion by specifying the
elevation of the sound source, but there are still two points where the
cone intersects the azimuth that produce the same ITD. Kettler et al.
(2017) were able to show that small variations in ILD in the
horizontal plane are enough to resolve the ambiguity.
To summarize, barn owls localize sounds in space by first
decomposing sounds into a phase-locked time component and a
loudness component for each ear. Time differences between the two
ears are computed by coincidence detectors in the nucleus laminaris
that respond maximally when signals from the two ears arrive in
phase, corresponding to a characteristic delay between the signals
from the two ears. Intensity differences between the two ears are
computed by neurons in VLVp that receive an excitatory signal from
one ear and an inhibitory signal from the other. ITD and ILD are
relayed to the inferior colliculus where a map of space is assembled
using ITD to specify azimuth and ILD to specify elevation of the
sound source.
There are many additional questions about sound localization in
barn owls that we have not discussed. The barn owl map of auditory
space that has been described is two-dimensional; it localizes sounds
on an imaginary sphere surrounding the owl’s head. Barn owls are
also sensitive to the third dimension, depth, but it is not clear how
this is computed by the auditory system. Barn owls are also known to
anticipate the location of moving prey and intercept it. Barn owls
move their heads parallel to the direction in which the prey is
travelling, moving ahead of the prey to the point of interception and
then returning to the prey. They also make large horizontal head
movements regardless of the prey’s direction of travel. It is not
known how auditory perception contributes to or is affected by these
head movements (Fux & Eliam 2009). Furthermore, Konishi found
that the barn owl brain can detect ITDs as short as 10 μs, even
though a single neural impulse is a hundred times longer than this.
Some additional mechanism must exist that is able to extract these
tiny timing differences from neural signals that are much less precise
in their temporal firing properties (Konishi 1995).
Research on sound localization by barn owls illustrates two general
principles about the brain and behavior. The first is that the brain
interprets environmental events like sounds by parsing them into
their component parts in parallel processing streams and then
reassembling the components into a neural representation of the
external world. Vision in vertebrates and invertebrates, bat sonar,
electroreception by fish, and most other sensory systems operate in
this way. The second general principle shown by this research is that
understanding the function of a complex system such as auditory
localization in the barn owl makes it easier to search for and analyze
its causal structure. As Konishi (1995, p. 270) puts it, “had the
researchers not known the perceptual problems the animals must
solve, they would not have looked for neurons selective for these
natural stimuli.”

Perception and sexual selection

Perceiving the environment is a major part of foraging, predator
avoidance, movement and migration, social behavior, parental care,
and indeed practically all kinds of behavior. Perception can also have
unexpectedly subtle effects on mate choice and sexual selection. The
courtship displays of birds, usually performed by males, can be
complex multimodal events. The spectacular displays of male
manakins (Pipridae) are familiar to many from natural history
documentaries and online videos, but even familiar birds like the
North American brown-headed cowbird (Molothrus ater) perform
eye-catching displays accompanied by distinctive vocalizations. Male
brown-headed cowbirds puff their feathers, spread their wings, and
bow forward while producing a few low frequency notes called
“glugs” followed by a high frequency 5–12 kHz phrase called the P2.
Females respond to preferred displays with a copulatory solicitation
display (CSD). The prevailing view is that in most species, females
select mates on the basis of male display because the quality of the
display is correlated with direct benefits the male can provide to the
female and her offspring or with indirect genetic benefits the male
can pass on to offspring (Andersson 1994; Searcy & Nowicki 2005).
This supposes, however, that displays honestly advertise male quality
and that there is consensus among females about what is a good
display. Ronald et al. (2018) examined how female auditory and
visual perception affects female cowbirds’ preference for male
displays. They measured the duration and latency of female CSDs to
video recordings of male displays that varied in intensity, paired with
male songs that varied in the likelihood of eliciting a CSD. Females’
auditory sensitivity and ability to resolve rapid temporal change in
sound were determined from auditory brainstem responses.
Females’ visual temporal resolution was determined from visual
evoked potentials and visual spatial resolution from the density of
retinal cone receptors.
Auditory and visual temporal resolution were both found to affect
the duration of females’ CSDs to male displays and songs. Females
with poor auditory temporal resolution preferred males whose P2
phrases were longer, lower in frequency, and more resembled pure
tones. Females with good auditory temporal resolution had the
opposite preference. Females with poor visual temporal resolution
preferred males who gave high intensity displays with more feather
puffing and greater wing extension. Females with good visual
temporal resolution had the opposite preference (Ronald et al. 2018).
The finding that variation within a population in female auditory and
visual perception affects mate choice raises challenging questions for
sexual selection based on female choice. Does the pattern discovered
in brown-headed cowbirds lead to assortative mating by intensity of
male display and female perceptual resolution? Is the observed
variation in female perception condition dependent or perhaps the
outcome of developmental processes? Have tradeoffs in the cost of
female auditory and visual resolution led to variation in the intensity
of male displays? If females with different auditory and visual
resolution prefer different intensities of display, what direct or
indirect benefits are these females receiving from the males they
prefer, and what exactly are male displays advertising? Whatever the
answers to these questions, Ronald, Fernandez-Juricic & Lucas’s
(2018) results clearly show that females are not all resolving or
responding to male signals in the same way.

Cognition
We turn next to the cognitive functions of the brain. “Cognition” in
animals means many things to many people (see also Chapter 9). For
some it implies that animals think, possess consciousness, and
experience the world as we do (Griffin 2001). We will not be using
the term “cognition” in this sense. Apart from the obvious problem of
verifying what another organism, even another human, thinks or
experiences, using our own thoughts or conscious experience to
understand cognition can be misleading. Neuropsychological
research with humans shows that what we experience is rarely a
reliable guide to how our cognitive processes actually work (Schacter
1996). In addition, using human experience as a model for cognition
in animals seems likely to seriously underestimate the diversity of
animal experience, whatever it might be.
The term “cognition” is used by many researchers, in this chapter
and in Chapter 9, in quite a different way, to refer to information
processing in a general sense. The term is broad enough to embrace
widespread forms of learning, like Pavlovian conditioning, as well as
kinds of learning that seem to follow rules of their own, like song
learning, imprinting, and some learned components of navigation.
So at the cost of fuzziness about exactly what cognition is and what it
is not, our working definition of cognition will be the processing of
information about the animal’s environment.

Honeybee learning
Honeybees (Apis mellifera) have a characteristic response to sugars
that does not require learning. When sensory cells on their legs,
antennae, or mouthparts detect sucrose, usually in the nectar of a
flower, bees extend their proboscis and siphon up the nectar
(Figure 5.4). Honeybees do not respond, however, to the colors,
shapes, or odors of flowers in this way. They have preferences for
approaching certain colors, shapes, and odors more than others, but
identifying which flowers within flying distance of the hive are
producing nectar and which are not requires learning the visual and
olfactory characteristics of flowers and associating them with the
presence of nectar.
Figure 5.4 Electron micrograph of the head of a honeybee
showing the antennae and proboscis. (From the Centre for Electron
Optical Studies, University of Bath, Bath, UK; Copyright University
of Bath).
The neural pathways involved in learning the olfactory traits of
flowers and associating them with the presence of nectar have been
investigated extensively by Menzel and his colleagues (Menzel &
Müller 1996; Menzel et al. 2001). The neuroanatomical circuitry, and
especially the intracellular processes involved in the formation of
memory, are the topics of the next section. Learning in the honeybee
provides an illustration of the neural basis of what is, conceptually at
least, a relatively simple cognitive event: forming a Pavlovian
association between two stimuli (see Chapter 8).
Pavlovian conditioning involves an unconditioned stimulus (US)
that elicits an unlearned response. For the honeybee, the US is
sucrose and the response is extension of the proboscis. The second
stimulus is the conditioned stimulus (CS), in this case floral odor,
to which the bee does not initially respond with proboscis extension.
As a result of experiencing a contingent relation between detecting
the odor and encountering sucrose, the odor eventually comes to
elicit the conditioned response of proboscis extension. In the
terminology of contemporary Pavlovian learning theory, an
association is formed between odor and sucrose because the odor is a
good predictor of the presence of sucrose. Conditioning of the
proboscis extension response (PER) in the honeybee provides an
opportunity to examine the cellular and molecular processes that
take place in the honeybee brain during the formation of this learned
association.
The idea that a record of experience—whether we call it an
association, a memory trace, or an engram—involves a structural
change at the cellular level was most clearly articulated by Donald
Hebb in 1949, and this idea has come to be called Hebb’s rule:
When an axon of cell A is near enough to excite a cell B and
repeatedly or persistently takes part in firing it, some growth
process or metabolic change takes place in one or both cells such
that A’s efficiency, as one of the cells firing B, is increased.(Hebb
1949, p. 62)
A “Hebb synapse” is thus a connection between neurons that
becomes more effective at stimulating its target neuron as a
consequence of correlated activity in the two cells. A great deal of
contemporary research in the neurosciences has been directed at
finding cells and synapses that conform to Hebb’s rule. There are
some limitations to Hebb’s rule. In its simplest form it allows the
probability that one cell will fire another to increase indefinitely,
even when cells occasionally fire together by coincidence, and so
modifications of Hebb’s rule are used in modelling neural networks.
But Hebb’s rule remains a useful heuristic and neurons in the
mushroom bodies of the honeybee brain that are responsible for
Pavlovian conditioning are one example of a system that has been
discovered to follow the rule Hebb had in mind.

Mushroom bodies
The mushroom bodies are a paired structure, one on the left and one
on the right, in the most anterior part of the honeybee brain, the
protocerebral lobe. The mushroom bodies vary enormously among
insects and can be identified in many other arthropods, including
spiders, scorpions, horseshoe crabs, and the pycnogonids or sea
spiders (Strausfeld et al. 1998). The honeybee mushroom body
consists of two cup-shaped structures, the calyces, joined by a stem
called the peduncle (Figure 5.5a,b). Surrounding the calyces are the
cell bodies of Kenyon cells, which receive converging input from
olfactory receptors in the proboscis and antennae and sucrose
receptors in the tarsae. Much of the structure of the mushroom body
consists of densely packed fiber projections of the Kenyon cells and
axons conveying input from, or sending output to, other parts of the
brain.
Figure 5.5 (a) The brain of the honeybee (Apis mellifera) shown
as a three-dimensional reconstruction in its approximate location in
the head. Red: mushroom bodies; blue: central body; yellow: optic
lobes; green: antennal lobes. (b) Structure of the mushroom body. l,
lateral; m, medial; col, collar; Pe, peduncle; br, basal ring; vl, vertical
lobe. a & b from Avargues-Weber & Giurfa. (2013). (c) Acetylcholine
(ACh) released at the conditioned stimulus (CS) synapse causes an
influx of calcium ions (Ca2+) in the Kenyon cell. Octopamine (Oc)
released at the unconditioned stimulus (US) synapse causes an influx
of Ca2+ and activation of adenylate cyclase (AC). Adenylate cyclase
converts ATP to cyclic AMP (cAMP), which activates protein kinase A
(PKA). PKA phosphorylates the cAMP-binding protein (CREB).
Signals from olfactory receptors in the honeybee antennae, the
pathway conveying information about the odor CS, are relayed first
to the antennal lobes and then to the mushroom body calyces. Input
from receptors in the proboscis, the pathway conveying the
information about the sucrose US, is relayed first to the
suboesophageal ganglion and then along a branch of the ventral
unpaired median neuron (VUM) to the mushroom body calyces.
Electrical stimulation of this branch of the VUM or injection of
octopamine, the neurotransmitter used by this neuron, can
substitute for the sucrose US and produce conditioning of the PER to
an odor.

Converging input
Encountering a floral odor followed by detection of sucrose should
thus cause firing of Kenyon cells by input from the antennae and,
immediately following or concurrently, further firing caused by input
from the proboscis. If Pavlovian conditioning of the PER follows
Hebb’s rule we would expect to find mushroom-body Kenyon cells
that, as a result of this concurrent neural activity, are more easily
fired by odor input alone. Such change in the ability of an odor to
activate a Kenyon cell would be the fundamental basis of the
conditioned PER to odor. A great deal of progress has been made in
identifying the molecular events within Kenyon cells that follow CS
and US stimulation. Interestingly, these molecular events are broadly
similar to the intracellular events that underlie conditioning in the
sea slug Aplysia and are also involved in the phenomenon of long-
term potentiation (LTP) in vertebrates (Bliss & Lømo 1973). LTP
is an experimental procedure that follows the Hebb rule par
excellence. Repeated stimulation of neurons in the rat hippocampus,
for example, results in a greater response by these cells to
subsequent stimulation. This increased responsiveness can persist
for days or weeks (hence “long-term” potentiation) and is thought to
resemble processes involved in neural plasticity and learning.
Induction of LTP depends on one of the several different receptors
for the neurotransmitter glutamate, the N-methyl-D-aspartate
(NMDA) receptor. These receptors only function when there is both
depolarization of the postsynaptic neuron and activation of the
NMDA receptor by glutamate released from the axon terminals of
the presynaptic neuron. Induction of LTP thus depends on
simultaneous activation of the presynaptic and postsynaptic neuron,
the same mechanism that Hebb proposed.
Activation of mushroom body Kenyon cells by a CS signal from the
antennae and a US signal from the proboscis initiates a cascade of
molecular events inside the cell that leads ultimately to gene
transcription, protein synthesis, and permanent structural or
metabolic changes in neurons. The discussion that follows describes
only a few of these molecular events, but enough to give an idea of
the molecular processes known to characterize learning and the
formation of memory across a wide variety of animals.
Release of acetylcholine, the neurotransmitter in the CS pathway
from the antennal lobe to the mushroom bodies, causes an increase
in the concentration of calcium ions (Ca2+) inside the Kenyon cell
(Figure 5.5c). Release of octopamine, the neurotransmitter in the US
pathway from the proboscis, increases activity of the enzyme
adenylate cyclase as well as increasing the intracellular concentration
of Ca2+. Adenylate cyclase converts adenosine 5´-triphosphate (ATP)
to cyclic adenosine monophosphate (cAMP). cAMP is one of the best
known and the first discovered second messengers, agents that
respond to signals originating outside the cell and which relay these
signals inside cells. One of the effects of cAMP is to activate protein
kinase A (PKA). The protein kinases are a family of catalysts that
facilitate phosphorylation, the transfer of phosphate from ATP to a
wide variety of proteins, regulating the activity of these proteins.
If the two events of CS activation and US activation occur
successively, their effects not only summate but interact. This is
because one consequence of elevated Ca2+ concentration inside the
cell is greater adenylate cyclase activity, the signal that ultimately
leads to elevated levels of PKA. Thus, if Ca2+ levels produced by the
CS odor signal are still high when the US sucrose signal arrives, there
will be greater PKA activity than produced by the sucrose signal
alone.
PKA activity, and the effect it has on phosphorylation of proteins
inside the Kenyon cell, is known to be a necessary step in
conditioning of the PER in the honeybee. Interfering with PKA
synthesis impairs memory measured 1 day after training (Fiala et al.
1999). Similar memory impairments occur in the fruit fly Drosophila
with mutations affecting PKA activity.
How does phosphorylation of a protein by PKA result in memory
formation? The activity of some genes is regulated by a region of
DNA called the cAMP-response element (CRE). Activation of the
CRE sequence is in turn controlled by a protein, the CRE-binding
protein (CREB). CREB is one of the proteins under the control of
phosphorylation mediated by PKA.
This series of interacting switches and controls, initiated by CS and
US stimulation, leads finally to gene transcription and the synthesis
of proteins inside the Kenyon cell. In the honeybee, disruption of
protein synthesis has little effect on learning in the first 24 hours
after odor conditioning, but 3 days later there are significant
impairments (Wüstenberg et al. 1998). The activity of genes coding
for many proteins are regulated by CREB. One such gene codes for
synapsin I, a protein that releases vesicles containing
neurotransmitter and allows them to move to the axon terminal
where they can be released into the synapse (Montminy & Bilezikjian
1987). Another gene regulated by CREB codes for ubiquitin, a
protein that forms part of a feedback loop that causes successive
bouts of PKA activity to be increasingly long-lasting, producing long-
term changes in cell chemistry, gene transcription, and protein
synthesis (Chain et al. 2000).
The second messenger system and its target, gene expression, are
tightly regulated to produce functional behavioral consequences.
Molecular signals of the kind involved in Pavlovian conditioning of
proboscis extension in the honeybee are known to underlie learning
and memory, as well as other forms of neural plasticity, in a wide
variety of animals and appear to be relatively conserved in evolution
despite enormous change in behavior. Molecular signaling systems
and the regulation of gene expression are essential cell functions
throughout the body, not just in the nervous system, and even in the
nervous system serve many functions beside forming permanent
records of experience. These cell signaling systems have been
recruited to produce the neural plasticity involved in the formation
of associations and give us a glimpse of the events inside neurons
that make cognition possible.

Control of Behavior

We come finally to the output side of the brain’s control of behavior.

There are many illustrations of the neural control of behavior that we
could examine: the control of flight in insects, the sequencing of
complex behavior in rodents, eye movements, locomotion, reaching,
grasping, orientation, and many others. However, one of the classic
models for research on the control of behavior is the neural control
of birdsong, a topic of fascination for students of animal behavior for
many years and one that continues to yield new insights about the
brain and behavior (Bolhuis and Moorman 2015).

Vocal motor control

There are many vocalizations produced by both male and female
songbirds year round, but the term “song” is usually reserved for
“long complex vocalizations produced by males in the breeding
season” (Catchpole & Slater 1995, p. 10). Although song is sometimes
heard outside the breeding season and is sometimes produced by
females, this broad definition applies for most species of songbirds
(or oscines). Song has a variety of functions. It can announce
territory occupancy, resolve contests over territory ownership, and,
as we saw earlier, provide cues used by females to choose mates.
Song production is controlled by a series of nuclei in the avian
forebrain (Figure 5.6a). One of these nuclei, HVC, sends projections
to another, the robust nucleus of the archistriatum (RA), which in
turn sends projections to the tracheosyringeal component of the
hypoglossal nucleus controlling motor output to the trachea and
syrinx. Song is learned in every species of oscine that has been
studied. HVC is also part of the neural circuits involved in song
learning and auditory feedback during song production, but the
focus in this section is on the motor side of song production (see
Chapter 7 for discussion of song development and learning).
Figure 5.6 (a) Schematic sagittal view of the songbird brain
showing the motor pathways for song with approximate locations of
nuclei and brain regions. The motor pathway descends from HVC to
RA, then to the tracheosyringeal area of the hypoglossal nucleus
(nXIIts), and then on to the trachea and syrinx. In male canaries,
neural activation occurs in the nuclei shown in orange: HVC, RA, Av,
LMAN, and Area X. Av, nucleus avalanche; DLM, medial subdivision
of the dorsolateral nucleus of the anterior thalamus; DM,
dorsomedial subdivision of nucleus intercollicularis of the
mesencephalon; HVC, at one time the “high vocal center” but now a
non-acronymic label; LMAN, lateral magnocellular nucleus of the
anterior nidopallium; LMO, lateral oval nucleus of the mesopallium;
Nlf, interfacial nucleus of the nidopallium; nXIIts, tracheosyringeal
portion of the nucleus hypoglossus (nucleus XII); Pam, nucleus
retroambiguus medullaris; RA, robust nucleus of the arcopallium;
RAm, nucleus paraambiguus medullaris; Uva, nucleus uvaeformis;
VTA, ventral tegmental area. From Moorman et al. (2011). (b) ZENK
gene expression in male canaries, shown as a multiple of the control
level, in HVC, RA, and Area X as a function of the number of songs
sung in the 30-min period preceding sacrifice of the bird. Each dot
represents one bird. Values describe the linear correlations between
the number of songs sung and the amount of ZENK expression.
From (Jarvis & Nottebohm 1997). Copyright (1997) National
Academy of Sciences USA.
Song is produced at a very high rate during the breeding season and
is heard less often at other times of year. There is variation among
species in the exact seasonal timing of song production, but in
general song is produced only when it can serve the functions of
territory defense and mate attraction. Nottebohm (1981) discovered
that in male canaries (Serinus canarius) not only does singing
behavior wax and wane seasonally, but so does the size of HVC and
RA. The seasonal change in size of the song control nuclei can be
substantial. In the wild, HVC and RA can be 50–100% larger during
breeding in sparrows of the family Emberizidae (Tramontin &
Brenowitz 2000).
Increase in the size of HVC and RA occurs in different ways. In RA,
the size of neurons and the spacing between them both increase
(Tramontin 1998). In HVC, in contrast, the number of neurons
increases (Goldman & Nottebohm 1983). Like the increase in total
size of HVC, the increase in neuron number can be dramatic. In song
sparrows (Melospiza melodia), the number of neurons in HVC
increases from about 100,000 outside the breeding season to over
200,000 in early spring (Tramontin & Brenowitz 1999). Although
there had been previous reports of neurogenesis in the adult brain
(Altman 1962), the discovery that new neurons were formed in the
canary song control nuclei, and in such large numbers (Goldman &
Nottebohm 1983), overturned the prevailing view in neurobiology
that neurons can only be lost from, not added to, the adult brain.
New neurons in the avian brain originate along the ventricle and
migrate into the brain from these proliferation zones following the
long radial processes of glial cells. New neurons can be identified by
injecting cell division markers, such as [3H]thymidine or
bromodeoxyuridine (BrdU). Both are incorporated into the DNA of
dividing cells during the S-phase—the DNA synthesis phase—of
mitotic cell division. Because [3H]thymidine is radioactive, cells that
have divided and incorporated [3H]thymidine can be identified in
brain sections autoradiographically. BrdU can be identified
immunocytochemically with antibodies to BrdU that attach a
fluorescent chemical or other visible marker to cells that have
incorporated BrdU into DNA in their nucleus (Balthazart et al.
2008). Not all cells labeled by these methods are necessarily
neurons, so additional tests must be used to confirm the neural
identity of these newborn cells, such as a neuronal phenotype seen in
a light or electron microscope, or the presence of neuron-specific
proteins.
Newly divided cells identified in this way are found mostly along the
ventricle in the first few days following injection, concentrated in
proliferation “hotspots” (Alvarez-Buylla et al. 1990). At this early
stage, the new cells do not have a neuronal phenotype. In the
following days and weeks these cells move into the forebrain,
differentiating into neurons as they travel and eventually take up
positions in HVC, the hippocampus, and elsewhere. They establish
functional connections with other neurons in the brain, confirmed by
injecting an anterograde chemical tracer into the cells that travels
along the axon to target cells, or by injecting a retrograde tracer that
travels backward along the axon from projection areas to the new
neuron.
Only a small proportion of the large number of new neurons born
along the ventricles survive for long in the avian brain. Most die
within a few weeks. Those that do survive and establish connections
with other neurons survive for months or longer. Neurogenesis in the
avian brain therefore seems to involve populations of cells that are
regularly replaced.
Why are new neurons incorporated annually into the song control
nuclei of birds? There have been many proposals. Because some
species learn new songs each year, or modify their existing songs,
new neurons may be required in the areas that control song output to
store instructions for producing new songs. Memory for new motor
instructions that are sent to the syrinx may require new neurons.
Like adult neurogenesis itself, this idea overturns some conventional
beliefs about the brain. Neural plasticity is known to involve changes
in the effectiveness with which one neuron activates another. This
can be achieved through changes at the level of the synapse or by
growth of new synaptic connections, without the replacement of
neurons. Because neurons can be replaced in adulthood, however,
neurogenesis may play a role in neural plasticity and the formation
of memory. The difficulty with this idea as a general account of the
function of neurogenesis is that some species of songbirds that do
not modify their songs from year to year still incorporate new cells
into HVC (Tramontin & Brenowitz 1999).
An alternative possibility is that neurogenesis in the song control
nuclei is not directly involved in storing new motor programs for
song but rather is involved in seasonal stereotypy of song (Tramontin
& Brenowitz 2000). When song is produced during the breeding
season, the acoustic characteristics of song are highly stereotyped.
The same identifiable song syllables (in some species very large
numbers of different syllables) are produced over and over with
almost exactly the same frequency and time properties. When song is
occasionally produced outside the breeding season, it is much less
stereotyped. Perhaps new cells in the song control nuclei play a part
in song stereotypy, ensuring that songs do not deviate from a
narrowly prescribed structure, analogous to recruiting additional
workers to ensure quality control in a difficult manufacturing
process.
Whether the function of new neurons involves memory or not, other
questions remain about adult neurogenesis. New neurons obviously
have the useful property that they can form new synaptic
connections, and the processes that guide neuronal migration may
be able to influence where new cells take up residence and send their
projections. New neurons also differ physiologically and
neurochemically from mature neurons (Gould et al. 1999; Doetsch &
Hen 2005). New neurons can have very large action potentials
compared with mature cells, have longer-lasting LTP, and respond
differently to some neurotransmitters. They are preferentially
incorporated, compared to mature neurons, into circuits involved in
spatial memory in mice (Kee et al. 2007). Adult neurogenesis may
maintain a pool of new neurons because some brain functions exploit
the specific properties of immature neurons.

Gene expression and song production

In the earlier discussion of honeybee learning we saw that structural
change in neurons involves the transcription of genes coding for
proteins that alter the properties of neurons. The control of song
production by HVC, RA, and other song nuclei provides an
opportunity to see gene transcription in action. Immediate early
genes are genes that are transcribed very rapidly, usually within
minutes following depolarization of a neuron. Most immediate early
genes code for transcription factors, proteins that influence the
expression of other genes. Jarvis and Nottebohm (1997) determined
the amount of expression of the immediate early gene ZENK in the
song nuclei of male canaries that either sang, heard canary song, or
both. The amount of ZENK activity was determined with a
radioactive riboprobe, a segment of RNA complementary to the
mRNA produced during ZENK transcription.
The birds were exposed to different conditions and experimental
treatments in order to create groups of birds that differed in singing
behavior and in what they heard. Some birds heard playback of taped
canary song, which induced them to sing themselves. These birds
showed striking increases in ZENK expression, up to 60 times
greater than control levels in the song control nuclei HVC and RA
(Figure 5.6b). Other birds could sing but not hear song. ZENK
expression in this group was found in the song control nuclei HVC
and RA but not in auditory areas. Unexpectedly, high levels of ZENK
expression were also found in a forebrain nucleus called Area X that
was not thought to be part of the circuit controlling song motor
output. Area X can be removed in adults without affecting song
production in the short term, and electrophysiological recording
from cells in Area X indicates no increased activation during song.
Nevertheless, strikingly elevated ZENK expression also occurred in
this area (see Figure 5.6b). Another group of birds was made unable
to sing by sectioning the hypoglossal nerve that innervates the
syrinx. These birds adopted a singing posture and opened their bill
but produced no song. ZENK expression in these birds was like that
of canaries that sang but did not hear the song of other birds. This
last result is important because it shows that it is the execution of
motor routines for singing, whether or not song is produced as a
result, that causes immediate early gene activity in the song control
nuclei.
What does ZENK expression reveal about how the brain organizes
motor output? Because immediate early genes are the first step in
transcription of other genes that code for proteins, it may reveal a
number of things. It may be that proteins are rapidly depleted in
neurons involved in song production and so must be synthesized and
replenished. ZENK activation is also thought to indicate long-term
changes in neurons. If this is the case, these results may indicate that
motor performance like singing can lead to further consolidation
of the motor routine each time it is performed (Jarvis & Nottebohm
1997).

Vocal motor memory

What, exactly, are the motor routines that drive song production in
the songbird brain? We know that these motor routines are learned
and we know that HVC neurons are at the top of the chain of
command during song production (Figure 5.6a). Neurons in another
nucleus, however, Nlf, send input to HVC and may control the
duration of phrases of a song. Using optogenetic methods, Zhao et al.
(2019) were able to control the firing of neurons in Nlf and observe
the effects on song production and song learning in zebra finches
(Taeniopygia guttata). Optogenetics combines genetic methods for
inserting light-sensitive protein channels into the cell membranes of
neurons and optical methods for activating these channels with light
(Deisseroth 2011; Shaaya et al. 2021). A viral vector is used to deliver
genes for the light sensitive protein channel—Zhao et al. (2019) used
Channelrhodopsin2—to a targeted population of neurons.
Channelrhodopsins are photochemicals found in algae and are
activated by blue light (Hegemann & Nagel 2013). The protein
channel is synthesized by the neuron and incorporated into the
neuronal cell membrane. An optic fiber (100–400 μm in diameter) is
implanted over the target population of neurons to illuminate the
light-sensitive channels resulting, in the case of channelrhodopsins,
in depolarization of the membrane and firing of the neuron.
Zhao et al. (2019) first verified that optogenetic activation of neurons
in Nlf selectively caused neurons in HVC to fire. They next raised
juvenile male zebra finches without adult song tutors and instead
“tutored” them with optogenetic activation of Nlf. Zebra finch
juveniles that are exposed to adult male tutors later sing as adults a
number of distinct phrases each around 100 ms in duration.
Juveniles raised without tutors sing longer phrases, each around 170
ms in duration. Juvenile males optogenetically tutored with repeated
50 ms pulses of light to Nlf later sang as adults phrases around 60
ms in duration. Juveniles tutored with repeated 300 ms pulses of
light sang as adults phrases that were much more variable but
centered on a median duration of 320 ms (Figure 5.7). Song
properties other than duration did not differ systematically between
optogenetically tutored males and either normally tutored males or
males with no tutors. Further experiments showed that
optogenetically tutored males used their songs in directed singing to
females in the normal way. Finally, and perhaps most strikingly,
when juvenile male zebra finches were raised with adult male tutors
and optogenetic activation of Nlf was made contingent on the tutor’s
song, birds did not copy the phrase duration of their live tutor but
the duration of their optogenetic “tutor.” When live tutoring and
optogenetic driving of Nlf were placed in competition, birds adopted
the durations that had been implanted optogenetically.
Figure 5.7 Juvenile zebra finch males were injected at 45 days
post hatch (dph) with a viral vector that delivered genes for
Channelrhodopsin2 (ChR2) to the nucleus Nlf. Two weeks later birds
began receiving optical stimulation of Nlf with either 50 ms (a) or
300 ms (b) pulses of light. As adults, these males sang phrases that
differed significantly in duration (c). Males tutored with 50 ms
pulses to Nlf sang phrases 60 ms in median duration; males tutored
with 300 ms pulses sang phrases 320 ms in median duration. From
Zhao et al. (2019).
So what are the motor routines that drive song production? At least
in the case of phrase duration, they are memories that can be
implanted in the vocal motor control system by activation of neurons
in the forebrain nucleus Nlf. Nlf provides auditory input to HVC and
in the course of song acquisition provides information about phrase
durations of tutor song to HVC. Lesioning Nlf following song
acquisition does not disrupt adult song performance (Zhao et al.
2019), so memory for phrase duration is not stored in Nlf. Instead
Nlf provides input to, or is responsible for, the sensorimotor
transformation of phrase durations that are heard into phrase
durations that are sung.
SUMMARY AND CONCLUSIONS
We began by looking at neural mechanisms for sensing and
perceiving the environment and found that there can be
remarkable specialization in the processing of information by the
brain. Complex auditory signals are decomposed by the barn owl
auditory system, then reassembled to construct a map of auditory
space. Sensation and perception shape how animals interact with
the world around them. As we saw, variation in female auditory
and visual temporal resolution can affect their preference for
male displays and ultimately which males they mate with. We
looked next at the formation of associations in the honeybee
brain. Cognitive processes in animal behavior can, of course, be
more complex than the formation of Pavlovian associations and
can involve more complex neural processes than the ones we
looked at in the honeybee brain. Cognitive processes of all kinds,
however, ultimately depend on relatively permanent changes that
take place inside neurons in response to the signals they receive
from other parts of the brain. Finally, we saw that what may seem
to be the rather straightforward matter of motor output can
involve phenomena such as the incorporation of new neurons
into the adult brain, the rapid transcription of genes, and the
formation of motor memories derived ultimately from sensory
input.
This survey of a few areas of research on the neural basis of
animal behavior has also given a glimpse of the many methods
available for answering questions about the brain and behavior,
from electrical recording of single neurons to molecular methods
for identifying intracellular signals that initiate gene
transcription, and optogenetic methods for activating select
populations of neurons. Techniques are techniques, however, not
ideas, and powerful techniques offer no substitute for careful
observation of behavior and curiosity about its cause, function,
development, and evolution. There has been explosive recent
growth in the neurosciences and rapid progress in many areas. It
is important to remember as research probes deeper into the
 brain that there is only one reason to care about the brain at all—
discovering more about the brain will help us understand the
causes of behavior.

FURTHER READING
Sound localization in birds, mammals, and reptiles is discussed by
Ashida and Carr (2011) and the evolution of sound localization by
Carr & Christensn-Dalsgaard (2016). Multimodal courtship
displays like those of cowbirds are examined by Mitoyen et al.
(2019). Leonard and Masek (2014) take an ethological approach
to the neural and cognitive mechanisms of odor and color learning
in bees, and the neural basis of olfaction and olfactory learning in
honeybees is reviewed by Paoli and Galizia (2021). The
neurobiology of vocalization in birds and mammals is examined
by Mooney (2020) and contrasting views on adult neurogenesis
are presented in a collection of papers in the Journal of
Neuroscience (vol. 22, no. 3, February 2002). A wealth of
information on animal communication can be found in Bradbury
and Vehrencamp (2011) and Searcy and Nowicki (2005). Poeppel
et al. (2020) contains many chapters by researchers in the
neurosciences on auditory and visual perception, cognition, and
motor control.

ACKNOWLEDGMENTS
I would to thank Scott MacDougall-Shackleton, Jeff Martin, Maddie
Brodbeck, and the late Peter Cain for their many helpful comments
and suggestions on this chapter and Carrie Branch for discussion of
perception and female choice.

REFERENCES
Altman, J. 1962. Are neurons formed in the brains of adult
mammals? Science, 135, 1127–8.
Alvarez-Buylla, A., Theelen, M. & Nottebohm, F. 1990. Proliferation
“hotspots” in adult avian ventricular zone reveal radial cell
division. Neuron, 5, 101–9.
Andersson, M. 1994. Sexual Selection. Princeton, NJ: Princeton
University Press.
Ashida, G. & Carr, C.E. 2011. Sound localization: Jeffress and
beyond. Current Opinion in Neurobiology, 21, 745–51.
Avargues-Weber, A. & Giurfa, M. 2013. Conceptual learning by
miniature brains. Proceedings of the Royal Society B-Biological
Sciences, 280, 20131907.
Balthazart, J., Boseret, G., Konkle, A.T.M., Hurley, L.L. & Ball, G.F.
2008. Doublecortin as a marker of adult neuroplasticity in the
canary song control nucleus HVC. European Journal of
Neuroscience, 27, 801–17.
Bliss, T.V.P. & Lomo, T. 1973. Long-lasting potentiation of synaptic
transmission in the dentate area of the unanaesthetized rabbit
following stimulation of the perforant path. Journal of Physiology
London, 232, 331–56.
Bolhuis, J.J. & Moorman, S. 2015. Birdsong memory and the brain:
In search of the template. Neuroscience and Biobehavioral
Reviews, 50, 41–55.
Bradbury, J.W. & Vehrencamp, S.L. 2011.Principles of animal
communication, 2nd ed. Sunderland MA: Sinauer.
Carr, C.E. & Christensen-Dalsgaard, J. 2016. Evolutionary trends in
directional hearing. Current Opinion in Neurobiology, 40, 111–7.
Carr, C.E. & Konishi, M. 1990. A circuit for detection of interaural
time differences in the brain stem of the barn owl. Journal of
Neuroscience, 10, 3227–46.
Catchpole, C.K. & Slater, P.J.B. 1995. Bird Song: Biological Themes
and Variations. Cambridge: Cambridge University Press.
Chain, D.G., Schwartz, J.H. & Hegde, A.N. 2000. Ubiquitin-
mediated proteolysis in learning and memory. Molecular
Neurobiology, 20, 125–42.
Deisseroth, K. 2011. Optogenetics. Nature Methods, 8, 26–9.
Doetsch, F. & Hen, R. 2005. Young and excitable: the function of
new neurons in the adult mammalian hippocampus. Current
Opinion in Neurobiology, 15, 121–8.
Fiala, A., Müller, U. & Menzel, R. 1999. Reversible downregulation of
protein kinase a during olfactory learning using antisense
technique impairs long-term memoiy formation in the honeybee,
Apis mellifera. Journal of Neuroscience, 19, 10125–34.
Fux, M. & Eilam, D. 2009. How barn owls (Tyto alba) visually follow
moving voles (Microtus socialis) before attacking them.
Physiology & Behavior, 98, 359–66.
Goldman, S.A. & Nottebohm, F. 1983. Neuronal production,
migration, and differentiation in a vocal control nucleus of the
adult female canary brain. Proceedings of the National Academy
of Sciences USA, 80, 2390–4.
Gould, E., Tanapat, P., Hastings, N.B. & Shors, T.J. 1999.
Neurogenesis in adulthood: a possible role in learning. Trends in
Cognitive Sciences, 3, 186–92.
Griffin, D.R. 2001. Animal Minds: Beyond Cognition to
Consciousness, 2nd ed. Chicago: Chicago University Press.
Hebb, D.O. 1949. The Organization of Behavior: &
Neuropsychological Theory. New York: John Wiley & Sons.
Hegemann, P. & Nagel, G. 2013. From channelrhodopsins to
optogenetics. Embo Molecular Medicine, 5, 173–6.
Jarvis, E.D. & Nottebohm, F. 1997. Motor-driven gene expression.
Proceedings of the National Academy of Sciences USA, 94, 4097–
102.
Jeffress, L.A. 1948. A place theory of sound localization. Journal of
Comparative and Physiological Psychology, 41, 35–9.
Kee, N., Teixeira, C.M., Wang, A.H. & Frankland, P.W. 2007.
Preferential incorporation of adult-generated granule cells into
spatial memory networks in the dentate gyrus. Nature
Neuroscience, 10, 355–62.
Kettler, L., Griebel, H., Ferger, R. & Wagner, H. 2017. Combination
of interaural level and time difference in azimuthal sound
localization in owls. eNeuro, 4, e0238-0217.2017.
Konishi, M. 1995. Neural mechanisms of auditory image formation.
In: M.S. Gazzaniga (ed.), The Cognitive Neurosciences, pp. 269–
77. Cambridge, MA: MTT Press.
Lashley, K.S. 1929. Brain Mechanisms and Intelligence. Chicago:
University of Chicago Press.
Leonard, A.S. & Masek, P. 2014. Multisensory integration of colors
and scents: insights from bees and flowers. Journal of
Comparative Physiology a-Neuroethology Sensory Neural and
Behavioral Physiology, 200, 463–74.
Menzel, R., Giurfa, M., Gerber, B. & Hellstem, F. 2001. Cognition in
insects: the honeybee as a study case. In: G. Roth & M.F.
Wullimann (eds.), Brain Evolution and Cognition, pp. 333–66.
New York: John Wiley & Sons.
Menzel, R. & Müller, U. 1996. Learning and memory in honeybees:
from behavior to neural substrates. Annual Review of
Neuroscience, 19, 379–404.
Mitoyen, C., Quigley, C. & Fusani, L. 2019. Evolution and function of
multimodal courtship displays. Ethology, 125, 503–15.
Moiseff, A. 1989. Bi-coordinate sound localization by the barn owl.
Journal of Comparative Physiology A, 164, 637–44.
Moiseff, A. & Konishi, M. 1981. Neuronal and behavioral sensitivity
to binaural time differences in the owl. Journal of Neuroscience,
1, 40–8.
Moiseff, A. & Konishi, M. 1983. Binaural characteristics of units in
the owl’s brainstem auditory pathway: precursors of restricted
spatial receptive fields. Journal of Neuroscience, 3, 2553–62.
Montminy, M.R. & Bilezikjian, L.M. 1987. Binding of a nuclear
protein to the cyclic-AMP response element of the somatostatin
gene. Nature, 328, 175–8.
Mooney, R. 2020. The neurobiology of innate and learned
vocalizations in rodents and songbirds. Current Opinion in
Neurobiology, 64, 24–31.
Moorman, S., Mello, C.V. & Bolhuis, J.J. 2011. From songs to
synapses: Molecular mechanisms of birdsong memory. BioEssays,
33, 377–85.
Nottebohm, F. 1981. A brain for all seasons: cyclical anatomical
changes in song control nuclei of the canary brain. Science, 214,
1368–70.
Paoli, M. & Galizia, G.C. 2021. Olfactory coding in honeybees. Cell
and Tissue Research, 383, 35–58.
Peña, J.L., Viete, S., Funabiki, K., Saberi, K. & Konishi, M. 2001.
Cochlear and neural delays for coincidence detection in owls.
Journal of Neuroscience, 21, 9455–9.
Poeppel, D., Manguin, G.R. & Gazzaniga, M.S. (eds.). 2020. The
cognitive neurosciences, 6th ed. Cambridge MA: MIT Press.
Rayleigh, L. 1907. XII. On our perception of sound direction.
Philosophical Magazine Series 6, 13, 214–32.
Ronald, K.L., Fernandez-Juricic, E. & Lucas, J.R. 2018. Mate choice
in the eye and ear of the beholder? Female multimodal sensory
configuration influences her preferences. Proceedings of the
Royal Society B, 285, 20180713.
Schacter, D.L. 1996. Searching for Memory. New York: Basic Books.
Searcy, W.A. & Nowicki, S. 2005. The evolution of animal
communication: Reliability and deception in signaling systems.
 Princeton NJ: Princeton University Press.
Seidl, A.H., Rubel, E.W. & Harris, D.M. 2010. Mechanisms for
adjusting interaural time differences to achieve binaural
coincidence detection. Journal of Neuroscience, 30, 70–80.
Shaaya, M., Fauser, J. & Karginov, A.V. 2021. Optogenetics: The art
of illuminating complex signaling pathways. Physiology, 36, 52–
60.
Strausfeld, N.J., Hansen, L., Yongsheng, L., Gomez, R.S. & Ito, K.
1998. Evolution, discovery, and interpretations of arthropod
mushroom bodies. Learning and Memory, 5, 11–37.
Takahashi, T., Moiseff, A. & Konishi, M. 1984. Time and intensity
cues are processed independently in the auditory system of the
owl. Journal of Neuroscience, 4, 1781–6.
Tramontin, A.D. 1998. Seasonal plasticity and sexual dimorphism in
the avian song control system: stereological measurement of
neuron density and number. Journal of Comparative Neurology,
396, 186–92.
Tramontin, A.D. & Brenowitz, E.A. 1999. A field study of seasonal
neuronal incorporation into the song control system of a songbird
that lacks adult song learning. Journal of Neurobiology, 40, 316–
26.
Tramontin, A.D. & Brenowitz, E.A. 2000. Seasonal plasticity in the
adult brain. Trends in Neurosciences, 23, 251–8.
Wagner, H., Kettler, L., Orlowski, J. & Tellers, P. 2013.
Neuroethology of prey capture in the barn owl (Tyto alba L.).
Journal of Physiology-Paris, 107, 51–61.
Wüstenberg, D., Gerber, B. & Menzel, R. 1998. Long- but not
medium-term retention of olfactory memories in honeybees is
impaired by actinomycin D and anisomycin. European Journal of
Neuroscience, 10, 2742–5.
Zhao, W.C., Garcia-Oscos, F., Dinh, D. & Roberts, T.F. 2019.
Inception of memories that guide vocal learning in the songbird.
Science, 366, 83–9.
6
hormones and behavior
JACQUES BALTHAZART AND GREGORY F. BALL
INTRODUCTION
In order to form pluricellular organisms, it is necessary for cells
to be able to communicate with each other in an adaptive manner
that allows them to synchronize metabolism, rhythmic activity,
and eventually behavior between the parts forming a given
individual. Two significant solutions emerged over the course of
the evolution of life to address this problem, especially in
vertebrates: a centralized nervous system and hormones. Within
the nervous system, cells communicate via electrochemical
means or direct electrical communication (see Chapter 5).
Hormones are, in contrast, chemical messengers secreted by
specialized endocrine cells, organized into ductless glands, that
transport information via body fluids (mainly blood in
vertebrates) to specialized cells with specific receptors that allow
them to bind the hormone, which then results in a change in
signaling or metabolic activity.
In this chapter we present, in a synthetic manner, the wealth of
information that has accumulated primarly in the past 50–60
years concerning the mechanisms through which hormones
modulate behaviors. In the first section, we provide a brief
introduction to how hormones are able to affect brain activity in
relation to the control of behavior. The second section then
describes the major concepts that have been established to
account for the reciprocal interactions between hormones and
behavior. A third and final section will illustrate these concepts
with a few selected examples. We will concentrate on examples of
hormone–behavior interactions that we think are among the best
understood, namely the roles of sex steroid hormones in the
control of reproductive behaviors in birds and mammals. A few
other behavioral systems modulated by hormones will, however,
also be briefly discussed.
 A Primer on Hormone Action in the Brain

From a reductionist point of view, most behaviors can be viewed as

series of muscular contractions triggered by an organized series of
nerve impulses. To affect the expression of a behavior, a hormone
thus has to modify electrical activity in specific groups of neurons.
Note, however, that hormones do not induce behaviors. Hormones
only change the probability of the occurrence or intensity of a
behavior in a particular situation; they modify the activity of neural
circuits so that an individual is more likely to react to relevant
stimuli with the expression of a given behavior. According to a widely
accepted terminology, it is said that hormones activate behaviors
when their presence greatly increases the probability that a behavior
will by produced in response to a given stimulus situation.
It is generally agreed that the first true experiment in behavioral
endocrinology was performed by Arnold Adolph Berthold who
identified, in 1849, the critical role of the testes in the expression of
courtship and copulatory behaviors of domestic fowl (Gallus
domesticus). If the testes of young male chicks were removed by
castration, they developed as adult males whose appearance (e.g.,
secondary sexual characters such as the comb and wattles) and
behavior (i.e., crowing and mounting females) did not develop as in
gonadally intact males. In contrast, if testes were reimplanted in
these castrated chicks, they developed the rooster-typical
morphology, plumage, and behavior. Because the testes could be
grafted anywhere in the abdominal cavity, these experiments
indicated that a blood-borne substance was responsible for induction
of behavioral changes observed in adults. This substance was
identified as the sex steroid hormone testosterone at the beginning
of the twentieth century and pure testosterone, first purified from
animal testes and soon thereafter chemically synthesized, became
broadly available for experimentation. Experiments have now
accumulated clearly demonstrating that, in a broad range of
vertebrate species belonging to all classes from fishes to mammals,
castration eliminates or at least markedly reduces the expression of
male copulatory behaviors, which can subsequently be restored by a
treatment with exogenous testosterone.
Testosterone is, like all other steroids, derived from cholesterol
through a series of enzymatic conversions that take place almost
exclusively in specialized endocrine structures (testes, ovaries,
adrenal glands, and also the brain itself; see Figure 6.1). These
enzymatic reactions produce hundreds if not thousands of different
compounds, many of which are able to modulate behaviors. This
chapter will only consider a few of them besides testosterone,
including the two sex steroids that play a critical role in the
activation of female sexual behavior, estradiol-17β and progesterone,
and also the steroids produced by the cortex of adrenal glands,
appropriately called corticosteroids (cortisol or corticosterone,
depending on the species) that play a key role in the stress response.
Figure 6.1 Simplified metabolic pathway illustrating the synthesis
and metabolism of sex steroid hormones that are critically involved
in the control of reproductive behaviors. Through a multi-step suite
of enzymatic reactions, cholesterol is transformed in endocrine
glands (testes, ovaries, adrenal gland) and probably in the brain into
pregnenolone, the precursor of all sex steroids. An isomerization
produces progesterone, which can then be transformed into the
androgens testosterone and androstenedione. Aromatizaton of these
compounds leads to synthesis of the estrogens estradiol-17β and
estrone, respectively, that are responsible at the cellular level for
many behavioral effects of androgens. The 5α-reduction of
testosterone into 5α-dihydrotestosterone (5α-DHT) is also
contributing to many actions of testosterone on behavior and is
critical during ontogeny for the formation of external genitalia in
mammals. Finally, 5β-reduction of testosterone produces 5β-DHT
that is largely devoid of behavioral action but plays a critical role in
hematopoiesis. Androgens are highlighted in blue, estrogens in pink,
and progestagens in brown. 17β-HSDH = 17β-hydroxysteroid
dehydrogenase.
Together, these four classes of steroids, namely androgens,
estrogens, progestogens, and corticosteroids (respectively,
testosterone, estradiol-17β, progesterone, cortisol/corticosterone and
related compounds), are involved in the modulation of most if not all
aspects of social life including sexual, aggressive, and parental
behavior as well as memory and various aspects of cognition.
Interestingly, some of them (androgens and estrogens mostly) are
also implicated in the control of numerous aspects of brain and
general physiology including angiogenesis, tumor growth, brain
plasticity including neurogenesis, and neurodegenerative processes.
Studies of steroids reveal most general principles of hormonal
regulation of behavior and constitute the primary focus of the
present chapter. However, several other hormones implicated in
behavioral regulation will also be discussed briefly (see Figure 6.2).
Figure 6.2 Schematic view of the endocrine glands and hormones
they produce that are implicated in the control of social behaviors.
Approximate location of glands is presented in a chicken, but the
same organization is essentially present in all tetrapods.
Abbreviations: ACTH, adrenocorticotropic hormone; CCK,
cholecystokinin; FSH, follicle stimulating hormones; GnRH,
gonadotropin releasing hormone; LH, luteinizing hormone; VIP,
vasoactive intestinal polypeptide.
Hormones secreted by the anterior pituitary gland are peptides or
proteins, i.e., amino acid chains of various lengths. They also modify
a variety of physiological and behavioral processes. For example,
prolactin seems to control the initiation or maintenance of behaviors
such as egg incubation in birds, parental care in mammals, or food
consumption. Another anterior pituitary hormone, the
AdrenoCorticoTropic Hormone (ACTH), is part of the
neuroendocrine circuit involved in stress adaptation. During
stressful circumstances (e.g., chronic subordination in a dominance
hierarchy), ACTH is released from the pituitary gland and stimulates
corticosteroid release from the adrenals, which normally feeds back
to inhibit further release of anterior pituitary hormones. Besides its
role as a trophic hormone, ACTH also acts on the brain to modulate
behaviors related to stress and escape.
The posterior pituitary peptides, oxytocin and vasopressin, were
initially associated with their role in mammalian milk ejection and
water retention, respectively. More recently, however, a large body of
evidence has implicated these hormones in maternal and sexual
behavior, affiliation in general, and in the control of social
organization including features such as gregariousness vs.
territoriality or monogamy vs. polygamy (Goodson & Kabelik 2009;
Balthazart & Young 2014).
Many other hormones are also implicated more or less directly in
behavior control. This includes thyroid hormones produced from the
amino acid tyrosine by the thyroid gland that regulates brain
development, sexual maturation, and seasonal cycles and also
several gut peptide hormones (e.g., cholecystokinin, CCK, and
vasoactive intestinal polypeptide, VIP) that are released from the
gastrointestinal tract and control digestion but are also found in the
brain where they regulate, as neurotransmitters and/or
neuromodulators, feeding or forms of learning and memory.

Cellular mechanisms of hormone action on behavior

Sex steroids activate behaviors by molecular mechanisms that are for
the most part similar to those used to produce their morphological
effects (e.g., growth of the oviduct for estradiol, of the prostate for
testosterone): they bind to intracellular receptors and regulate the
transcription of specific genes coding for proteins that will later
affect neurotransmission (Etgen & Pfaff 2009). A major goal of
current research in molecular neuroendocrinology is to identify
genes regulated by steroids and their function in relation to
physiology and behavior. More recent research, however, shows that,
in addition, the behavioral effects of steroids are also the result of
other brain-specific mechanisms involving namely direct effects on
neuronal membranes.
Intracellular receptors and the control of genomic
transcription
Steroid hormones, and other lipophilic hormones such as the
thyroxine, can more or less freely enter all neurons and they produce
biological effects by binding to specific intracellular receptors in cells
that express them. The complex formed by the hormone and its
receptor then acts, in the cell nucleus, as a transcription factor
leading to changes in the transcription of new messenger RNA
(mRNA) and production of new proteins that ultimately alter cell
function (see Figure 6.3). These effects are usually fairly slow and
take hours to days to even weeks to develop.
Figure 6.3 Schematic representation of the neural circuits
controlling vocalizations in songbirds. Key parts of the vocal as well
as auditory pathways are illustrated. For the auditory pathways, only
the last relays that interface with vocal structures are indicated and
are labeled “auditory” in the list of abbreviations. Aspects of the vocal
pathways are steroid sensitive (presence of estrogen and/or
androgen receptors) and are represented in the figure by red areas.
The tentative connections between the medial preoptic nucleus and
the song system via the PAG are indicated in green. The anterior
forebrain pathway mediating song learning is indicated by dotted
arrows. Abbreviations: DLM, nucleus dorsolateralis antherior
thalami, pars lateralis; HVC, used as proper name (formerly high
vocal center); ICo, nucleus intercollicularis; lMAN: lateral nucleus of
the anterior nidopallium; MLd, nucleus mesencephalicus lateralis,
pars dorsalis (auditory); NCM, caudal medial nidopallium
(auditory); NIf, nucleus interfacialis; nXIIts, motornucleus of the
hypoglossal nerve, tracheosyringeal part; Ov, nucleus ovoidalis
(auditory); PAG, periaqueductal gray (also called substantia grisea
centralis in birds); POM, nucleus preopticus medialis; RA: nucleus
robustus of the arcopallium; Ram: nucleus retroambigualis.
The neuroanatomical distribution of binding sites for androgens,
estrogens, and progestagens is remarkably consistent among
vertebrate species. This distribution was originally mapped in the
1970s using in vivo autoradiographic techniques. It was confirmed
more recently by immunocytochemistry and by in situ hybridization
histochemistry that allows one to localize the corresponding mRNA
in intact tissue slices.
A dense expression of sex steroid receptors is found in the medial
preoptic area (mPOA), several nuclei of the hypothalamus (anterior
hypothalamic area, ventromedial nucleus [VMN], tuberal
hypothalamus), telencephalic structures that are part of the limbic
system (amygdala, lateral septum, bed nucleus of stria terminalis),
and in specific parts of the mesencephalon (optic tectum).
Importantly, this distribution is strikingly similar across all
vertebrates from fishes to mammals. This evolutionary stability
testifies toward the functional importance of these receptors.
Furthermore, the anatomical distribution of androgen-concentrating
cells is in general similar to that of the estrogen-concentrating cells.
However, differences have been observed in the intensity and the
number of labeled cells, as well as in their precise distribution within
a given nucleus.
Only a fraction of these sex steroids receptors are directly implicated
in the control of behavior. These critical sites have been identified by
a combination of lesion experiments and of stereotaxic implantation
of steroids directly in the relevant brain areas of gonadectomized
subjects. Quite generally the mPOA is a critical site for the activation
male sexual behaviors by androgens (and derived estrogens; see
below), whereas estrogens and progestagens play a similar role in the
VMN for activating female behavior. However, additional sites are
also implicated. In males, for example, androgen action in the
septum, bed nucleus of the stria terminalis, and amygdala modulate
the expression of male sexual behavior even if the action of
androgens in the mPOA alone is sufficient to activate this behavior in
many cases.

Steroid metabolism
When entering the target tissues, testosterone, progesterone, and to
some extent estradiol undergo metabolic transformations. Two
enzymes, aromatase and 5α-reductase, catalyze the transformation
of testosterone into behaviorally relevant metabolites estradiol and
5α-dihydrotestosterone (5α-DHT), respectively. Additional enzymes
inactivate the steroid (e.g., 5β-reductase producing the behaviorally
inactive compound 5β-DHT). Progesterone is similarly metabolized
by the 5α- and 5β-reductase into 5α- and 5β-progesterone that have
a very different behavioral impact compared to progesterone; 5α-
progesterone namely is a high affinity ligand for the GABA (gamma-
aminobutyric acid) receptors. Estradiol is also subjected to
hydroxylations in brain cells to form compounds known as
catecholestrogens but their functional significance remains poorly
understood.
Because the concentration and activity of all these enzymes is
regulated, the ratio of active versus inactive metabolites that are
produced in the brain that can interact with the corresponding
receptors is affected by factors such as the sex, age, season, or
hormonal condition of the subjects. This provides a mechanism,
potentially fine-tuning the expression of hormone-dependent
behavior.
In many species, testosterone acts as a pro-hormone, i.e., it must
first be metabolized into another steroid before acting at the cellular
level. In a large number of species of birds and mammals, including
rats and Japanese quail, it is estradiol, sometimes erroneously
considered as a “female hormone,” that is locally produced in specific
regions of the male brain by aromatization of testosterone that
induces, at the cellular level, the neurochemical changes resulting in
the activation of male sexual behavior. In other species, such as
rabbits or guinea pigs, it is the 5α-DHT obtained by local 5α-
reduction of testosterone that activates copulatory behavior. In most
species, however, the full activation of male sexual behavior results
from a synergistic action of both testosterone metabolites, estradiol
and 5α-DHT. The relative role of these two steroids in this process
varies between species, but it seems that both are usually involved.
In many fish species, the main androgenic steroid found in the blood
that plays a major role in the activation of male sexual and aggressive
behaviors is not testosterone itself but a related compound derived
from testosterone called 11-ketotestosterone (11KT). In this case,
however, the critical transformation of testosterone into 11KT
already takes place in the testis, and the active steroid is also the
steroid found in the blood, whereas in the avian and mammalian
examples just discussed, testosterone is the circulating pro-hormone
and the transformation into the active metabolite (estradiol and/or
5α-DHT) takes place in the target organ itself, in specific neurons
that express the relevant enzymes. Testosterone therefore often acts
as a pro-hormone in the control of behavior, and the chemical signal
that acts at the cellular level to affect behavior is in fact a metabolite
of this prohormone. Interestingly, thyroxine is also a pro-hormone
released into the blood by the thyroid gland and it is locally
transformed via enyzmatic action to an active metabolite
triiodothyronine in target tissues. Thus, an important concept in
behavioral endocrinology is that hormonal effects can be amplified
via their transformation at specific times and and in specific
locations to more active forms.
Membrane receptors and the modulation of intracellular
signaling cascades
Steroid-induced behavioral changes described so far usually appear
after an exposure to the steroid, ranging from a few hours to a few
days. Such a time course can explain changes in reproductive
behavior that are observed over months during the annual cycle in
seasonally breeding animals. However, much faster actions of
steroids have also been identified, suggesting that these hormones
may also act via fundamentally different mechanisms. This is
particularly the case for estradiol, a steroid produced in the brain by
aromatization of testosterone that plays a key role in the activation of
male sexual behavior.
It was observed already in the 1970s that estradiol is able to
modulate the electrical activity of hypothalamic neurons within
minutes if not seconds. More recently, evidence has accumulated
indicating that estrogens acutely influence behavioral processes such
as pain perception, memory, and aggressive and sexual behaviors. It
was, for example, demonstrated that a subcutaneous injection of
estradiol stimulates mounts and anogenital investigations within 35
minutes in castrated rats (Rattus norvegicus), a latency too short to
be compatible with an activation by genomic mechanisms.
Subsequent studies demonstrated that a single injection of estradiol
facilitates the expression of most aspects of male sexual behavior
within 10–15 minutes in quail (Coturnix japonica) and mice (Mus
musculus). The existence of such rapid behavioral effects of estradiol
seems to be an ancient feature in vertebrates since these effects are
also observed broadly across taxonomic groups including fishes. For
example, the injection of estradiol modulates within minutes the
production of courtship vocalization in the plainfin midshipman fish
(Porichthys notatus).
Although purely cytoplasmic effects have also been described, these
rapid effects are generally initiated by steroids acting at the plasma
membrane, resulting in the activation of a wide variety of
intracellular signaling pathways. Even if many questions remain
open, it is now clear that multiple estrogen receptors are present at
the neuronal membrane including the two classical nuclear estrogen
receptors (ERα and ERβ, which in addition to being present in the
cytoplasm and the nucleus can fuse to the membrane via
palymitolyation processes and association with caveolin-1), a G-
protein-coupled receptor (GPR30) and two other membrane
receptors that have been postulated based on functional or
pharmacological evidence but have not been formally identified at
this time. The binding of estradiol to these receptors activates a
variety of intracellular signaling cascades that result in the
phosphorylation of various proteins such as enzymes or receptors,
changes in intracellular calcium concentration or even activation of
genomic transcription (indirect genomic effects) that ultimately
result in behavioral effects.
It is currently difficult to assess how general these membrane-
initiated steroid effects on behavior are, but research in this field is
very active at present and regularly leads to new discoveries. It has
been argued that in this short-term temporal context, estradiol
displays most, if not all, functional characteristics of a
neurotransmitter or at least a neuromodulator and thus can regulate
short-term changes in behavior that are usually considered to be
controlled by classical transmitters such as dopamine, noradrenalin,
and serotonin. Estradiol and other steroids could thus control both
the long-term and short-term changes in behavior allowing the
animal to adapt to lasting changes in the environment (e.g., seasons)
but also to more discrete disruption events (presence of a receptive
female, of a predator, snow storm, etc.). These steroids thus act in in
two different time domains to control long- and short-term changes
in behavior (Cornil et al. 2015).
It should also be noted that in addition to sex steroids, other larger
or more polar molecules such as the peptidergic or protein hormones
(e.g., vasopressin/vasotocin, oxytocin, gonadotropin-releasing
hormone, prolactin, etc.) that also play a role in the control of social
behaviors similarly produce their effects by binding to receptors
located at the cell membrane that are coupled to the activation of a
second intracellular messenger system (e.g., activation of adenylate
cyclase leading to the synthesis of cyclic AMP). It is, however, beyond
the scope of this chapter to review all these cellular mechanisms that
differ among hormones, brain regions, and species.
 Basic Physiological Mechanisms
Mediating Hormone Effects on Behavior

Central versus peripheral actions

The brain is the major site of hormone action for the expression of
sexual and related social behaviors and most of this chapter
therefore concerns the ways in which steroids act at this level to
regulate behavioral expression. As will be discussed later, steroid
hormones can act on multiple neural sites with diverse cellular
consequences to modulate the activation of complex behaviors.
However, hormones have widespread effects throughout the entire
organism and, besides actions on the brain, there are at least three
other ways that hormones can act on peripheral tissues that are
relevant to behavior control (Hinde 1970).
Steroids can first affect sensory inputs to the brain. It has also been
shown that sex steroids increase the sensitivity of the inner ear in the
female plainfin midshipman fish (P. notatus) to the male calls
(hums) during the breeding season and in this way modulate their
reproductive behavior (Forlano et al. 2015). In rodents there are
well-known effects of hormonal state on olfactory processing such
that females in estrus are more responsive to conspecific olfactory
signals as compared to ovariectomized females or females in an
anestrus state (Pfaff & Pfaffmann 1969; Baum & Bakker 2013).
Second, steroids such as androgens are known to have anabolic
(trophic) effects on muscles that are ultimately the effector organs of
behaviors. For example, androgens affect the morphology of the
penis in mammals and the equivalent intromittent organ that is
present in some avian species (such as Ratites and select waterfowl
species) as well as related organs in species that lack an intromitten
organ such as the cloacal gland in Japanese quail. The effector
organs for more complex social behaviors used in the context of
reproduction are also often the target of steroid hormone action. The
mass of muscles controlling the syrinx (the primary vocal production
organ in birds) is increased by androgens. Testosterone also
modulates the cholinergic activity in the songbird syrinx. These
morphological and biochemical changes in the syrinx represent one
way that androgens can modify singing behavior in addition of
course to their central effects on the neural song control system.
Similarly, the androgen-dependent development of the thumb pad in
male frogs during the reproductive season is clearly needed to
facilitate egg fertilization by males. During mating, the male grasps a
female with his front legs (amplexus) and then releases his sperm in
synchrony with female oviposition. Appropriate amplexus is only
possible in these aquatic animals following development of the
thumb pad.
Finally, steroids also alter social signals that will indirectly affect
behavioral expression. Many birds exhibit a marked steroid-
dependent sex dimorphism in their plumage and in a number of
integumentary derivatives and skin appendages such as beak, comb,
wattles, and cloacal gland. These structures play an important role as
social signals during sexual interactions and steroid-induced changes
in these signals will affect the behavior of congeners and thus have
profound effects on the expression of sexual behaviors. Steroid-
dependent olfactory signals are also produced in males and females
of many vertebrate species from fishes to mammals and directly
influence the expression of sexual behaviors.

Steroids regulate social behaviors at multiple sites in a

non-redundant fashion
The neuroanatomical distribution of sex steroid receptors initially
guided investigations on the neural circuits controlling male and
female sexual behaviors. Stereotaxic lesion and hormone implant
studies all clearly demonstrated the importance of the mPOA in the
activation of male copulatory behavior. The importance of this region
for male behavior has been conserved through evolution in that
lesions to this area impairs copulation in male rats and in a large
number of other mammalian species as well as in all species of birds,
reptiles, amphibians, and fishes that have been investigated.
The mPOA is clearly a steroid-sensitive integration site
bidirectionally connected to a large number of brain regions. In
particular, the mPOA receives, directly or indirectly, inputs from
most, if not all, sensory modalities and is therefore ideally positioned
to integrate information from the environment and adjust responses
made by the organism to environmental and endocrine inputs. These
connections have been investigated in a few species (e.g., rat, quail)
by a variety of techniques including anatomical tract-tracing
(stereotaxic injection of compounds that will be transported from the
neuronal cell body to the axon terminal [anterograde tracing] or
conversely from the terminals to the perikarya [retrograde tracing])
or the quantification of the expression of immediate early genes such
as c-fos or egr-1 or of the increased glucose accumulation following
expression of the behavior. These studies identified inputs coming
from multiple steroid-sensitive regions such as the ventromedial
hypothalamus, the amygadala, the septum, and the bed nucleus of
the stria terminalis. These nuclei and a few others form a circuitry
that has been called the social behavior network that is implicated,
apparently in most vertebrate species, in the control of various
aspects of social behaviors (sex, but also aggression, gregariousness,
territoriality, etc.).
In addition, the mPOA sends a prominent projection to the
periaqueductal gray (PAG, also named by its Latin name, substantia
grisea centralis). This projection to the PAG represents a key link
between the diencephalic center where endocrine stimuli and
sensory inputs originating from the female are integrated with the
spinal module controlling motor outputs reflected in sexual behavior
itself. The specific function of each node in this network has not
always been identified, but information is available in some specific
cases. For example, the steroid-sensitive amygdala in hamsters
(Mesocricetus auratus) clearly serves as a filtering and integration
site for the olfactory stimuli originating from the female and the
endocrine stimuli produced by the gonads.
A similar anatomical organization of the circuits controlling sexual
behavior is present mutatis mutandis in females. However the
critical site of steroid action is the ventromedial nucleus of the
hypothalamus where stereotaxic implantation of estradiol and/or
progesterone is usually sufficient to activate sexual receptivity and
proceptivity. However, detailed studies of the circuit controlling
sexual receptivity have revealed that steroids have general effects on
arousal in addition to specific effects in the ventromedial nucleus in
promoting particular motor behaviors associated with receptivity
and proceptivity (Pfaff 1980).

Sex differences in steroid action

A variety of behaviors in animals, including humans, are
preferentially or in some cases even exclusively exhibited in one sex
(Goy & McEwen 1980; Becker et al. 2008). It was initially thought
that these sex differences resulted from the presence of a different
hormonal milieu in adult males and females, such as high plasma
testosterone in males and high plasma estrogen and progesterone in
females. This is true in some cases: female quail who never produce
the crowing vocalization will crow like males do if treated with
testosterone, but this is not the rule and, by the middle of the
twentieth century, this interpretation was found to be generally
incorrect. Estrogens often cannot activate female-typical behaviors in
adult males, and conversely testosterone often fails to activate male-
typical behaviors in adult females.
An early and irreversible sexual differentiation of the brain response
to steroids was suspected to take place based on a variety of
experimental data. In 1959 this work was conceptualized in a seminal
paper by Phoenix and collaborators demonstrating, in guinea pigs,
that besides their effects on the activation of behavior in adulthood,
steroids produce an irreversible sexual differentiation of brain and
behavior during early life (Phoenix et al. 1959). This process of sexual
differentiation has been identified in all classes of tetrapods, but
although the general principle remains the same in all species (i.e.,
early action of steroids organize in an irreversible manner the
responsiveness of the adult brain to steroid hormones), there are
many variations on this general theme.

Sexual differentiation of behavior: the concept of

organizational effect of steroid hormones
Mammals
In mammals, early exposure to testosterone produces a masculine
phenotype. Behavioral characteristics of the male are enhanced
(masculinization), and more or less independent of this process, the
capacity of males to display female-typical behavior is diminished or
lost (defeminization). The female phenotype in contrast develops in
the absence of (high) concentrations of sex steroids. Recent evidence,
however, suggests that small amounts of estrogens are required
during ontogeny, at least in rats and mice, to allow the development
of a female brain that will be able to support in adulthood the
expression of the full complement of female behaviors.
These organizing effects of sex steroids that take place during normal
ontogeny can be mimicked by neonatal gonadectomy or injection of
testosterone. Thus the development of the masculine and feminine
phenotypes of responsiveness to sex steroids in adulthood can be
obtained irrespective of the genetic sex of the subjects by neonatal
endocrine manipulations. It is also important to note that these
organizational actions of steroids only take place during a limited
period of development (the embryonic or early postnatal life,
depending on the species) that is usually referred to as the critical
period.
Although the masculinization and defeminization of male mammals
is triggered by testicular androgens, it is often not testosterone per
se, but its aromatized metabolite estradiol that is responsible for the
differentiation of behavior at the brain level. Prenatal ovaries in rats
and mice are not active during the embryonic period (they secrete no
or very small amounts of estrogens), but the ovaries of pregnant
mothers secrete high levels of estradiol that should easily permeate
the placental barrier and should masculinize/defeminize female
embryos. However, the plasma of perinatal rats contains high
concentrations of α-fetoprotein (AFP), a protein that binds and
sequesters estrogens in the blood so that they are unavailable for
brain differentiation. Testicular testosterone, which does not bind to
APF can in contrast reach the brain of males where it exerts its
masculinizing effects after being locally aromatized to estradiol.
Some species including primates do not appear to have detectable
levels of circulating AFP or an analogous protein that would bind
estrogens. The relative importance of brain aromatization of
testosterone in the process of brain sexual differentiation in these
species is therefore questionable and could also be quite variable
from species to species. In monkeys for example, androgens
themselves may be more important than estrogens for sexual
differentiation of brain and behavior.
Note also that the relative importance of the masculinization and
defeminization processes vary from one species to another. For
example, the sexual differentiation of male rats consists mostly of a
defeminization associated with a limited masculinization of behavior,
while in rhesus monkeys a marked masculinization of behavior takes
place that is not associated with defeminization.
More subtle differences in hormone exposure during development
can also influence the degree to which an animal will display male or
female-typical behavior. A female rat positioned between two males
in utero, and thereby exposed to slightly elevated concentrations of
steroid hormones, will display higher levels of aggression and other
male typical behaviors as an adult compared to sisters who were not
positioned between two males.

Birds
In avian species that have been studied in detail, mainly quail
(Coturnix japonica) and to some extent chickens (Gallus
domesticus) and ducks (Anas platyrhynchos), exposure to steroids
early in ontogeny also affects the responsiveness to steroids in
adulthood, but this sexual differentiation process seems to affect
exclusively male-typical behaviors. A variety of male-typical sexual
behaviors have been described that cannot be activated in females
even after treatment with high doses of testosterone. Female
reproductive behaviors (e.g., sexual receptivity postures such as
squatting) are in contrast not sexually differentiated and can be
activated in both sexes provided an adequate treatment with
estrogens is administered. In these avian species, the sex difference
in responsiveness to adult testosterone is the result of the early
exposure of female embryos to ovarian estrogens (as opposed to
exposure of males to testicular androgen as is the case in mammals).
In the absence of embryonic steroids, the male phenotype of
responsiveness to testosterone develops. In the presence of
estrogens, behavior is demasculinized: birds lose the capacity of
express, as adults, male-typical copulatory behavior in response to
testosterone.
This physiological process can be experimentally manipulated so that
the male or female behavioral phenotypes are induced in either
genetic sex by adequate embryonic endocrine manipulations. In
Japanese quail, treatment of a male embryo with estrogens on day 9
of incubation produces an adult that will never exhibit the male-
typical copulatory patterns even after treatment with behaviorally
active doses of testosterone. Conversely, early treatment of female
embryos with an aromatase inhibitor that blocks estrogen synthesis
leads to the development of a fully masculine behavioral phenotype.
Contrary to what is observed in mammals, in the absence of gonadal
steroids, these avian species thus develop a male phenotype while the
female phenotype develops in the presence of estrogens.
These contrasting patterns of differentiation in birds and mammals
(i.e., relative absence of endocrine stimulation during early life
results in male behavioral phenotype in birds and in female
behavioral phenotype in mammals) might be related to the fact that
females are the homogametic sex in mammals (female XX and male
XY) while males are homogametic in birds (male ZZ and female ZW).
This observation could then highlight a more general rule according
to which the phenotype of the heterogametic sex (male mammals
[XY] and female birds [ZW]) would always develop in response to
hormonal stimulation during ontogeny while the behavioral
phenotype of the homogametic sex would be the “default” sex
observed in the absence of early endocrine stimulation (also referred
to as the “neutral” sex). This principle must be accepted cautiously at
present because the proximate mechanisms that might explain this
connection between the nature of the sex chromosomes in males and
females and the process of behavioral differentiation have still not
been identified. In particular, the process of sexual differentiation of
the gonads is reasonably well established in mammals (the sry gene
of the Y chromosome induces formation of the testes) but not in
birds where the differentiation of the gonad seems to be the result of
the presence on one or two Z chromosomes of the gene DMRT1 (gene
dosage effect).

Other vertebrates
In other vertebrate classes including fishes, amphibians, and reptiles,
many species do not have sex chromosomes (Bachtrog et al. 2014).
Their brain and behavior differentiate by endocrine mechanisms that
are driven by the physical (e.g., temperature) or social environment
(presence of congeners of the other sex, position in a dominance-
subordinate hierarchy). For example, in reptiles, the first identified
event signaling sexual differentiation is an increase in aromatase
expression in the female gonad leading to an increase in estrogen
concentration that will by itself transform the undifferentiated gonad
into an ovary. Whether gonadal steroids have an early and
irreversible effect on the brain and its responsiveness to steroids in
adulthood is not broadly established in fish, amphibians, and
reptiles. There are, however, indications that such a phenomenon
may take place at least in some species. For example, in species with
alternative mating strategies, early events in development seem to
have long-lasting consequences. In tree lizards (Urosaurus ornatus),
concentrations of progesterone and testosterone during development
determine, in an apparently irreversible manner, whether the adult
males will be territorial during their entire life or will switch from
nomadic to satellite male as a function of environmental conditions.
In the midshipman fish (P. notatus), the physiological “decision” to
become a territorial male who builds a nest to attract females or a
sneaker (satellite male) is made early in development and is
apparently irreversible, but the endocrine bases of this “decision” are
unclear at present. However, given the plasticity in the expression of
sex-typical behaviors that can be observed by adult fish species, it
seems likely that, in many species, the brain and the behavioral
phenotype are not determined in an irreversible manner by the early
endocrine environment.

Brain correlates of sex differences in behavior

Large numbers of sex differences in brain structures implicated in
the control of sexual behavior have been observed. These differences
can be morphological in nature (e.g., volume of groups of neurons in
the POA or ventromedial hypothalamus, size of neurons or of their
dendritic arborization more developed in females), but also
biochemical. For example, the concentration and turnover of various
neurotransmitters or neuropeptides as well as the density of their
receptors can be sex-specific (McEwen et al. 1988). Some of these
anatomical or biochemical differences persist in adult
gonadectomized animals placed in similar hormonal environments.
They are thus presumably caused by the early organizing effects of
sex steroids and potentially linked in a causal manner to behavioral
differences between sexes. Overall, the organizational effects of
steroids on behavior are the consequence of changes in gene
expression that result in the multiplication, migration, or death of
neurons, as well as in their functional differentiation (e.g.,
expression of neurotransmitters, neuropeptides, and their
receptors). The causal links between sex differences in brain and
behavior remain, however, poorly understood except in a few cases
(e.g., for some aspects of brain function in rats) and much research
on this topic is still fully warranted (Ball et al. 2014).

Genetic effects on sex differences in brain and behavior

that do not occur via hormonal action
Based on the previous section, it could seem that, in birds and
mammals, all stable (i.e., still present when adults of both sexes are
placed in identical endocrine conditions) sex differences in behavior
are the result of early organizational effects of sex steroids and do not
relate in a direct manner to the genetic make-up of the individuals.
Sex chromosomes would only determine the sex of the gonads and
the sex steroids secreted by these gonads would during early life
determine all other phenotypic aspects of the sex difference. Some
suspicion that this mechanism was not universal had been raised by
the discovery of phenotypic sex differences and sex differences in
gene expression that develop before the differentiation of the gonads
(and thus cannot be induced by exposure to a differential hormonal
milieu).
A clear challenge to this broad principle was then identified in zebra
finches (Taeniopygia guttata ), a species of songbirds that conforms
to the avian pattern as far as the distribution of sex chromosomes is
concerned (ZZ males and ZW females) but whose singing behavior
differentiates by mechanisms that do not seem to follow the pattern
identified in quail and chicken. In zebra finches, estrogens
administered to males early in life demasculinize male-typical
copulatory behavior as observed in quail but the same early estrogen
treatment masculinizes singing behavior in females as well as
morphological aspects of the brain nuclei that control this behavior.
It is therefore difficult to understand how the full behavioral
repertoire of the male differentiates under physiological endocrine
conditions in this species given that normal untreated males both
copulate and sing in adulthood while females never exhibit these two
behaviors. All experimental manipulations of the early endocrine
milieu of these birds made in an attempt to resolve this paradox were
unsuccessful but the discovery of a single individual who was a
mosaic of male cells on the right side and female cells on the left size
gave a clear clue to the underlying mechanism. This bird known as
the gynandromorphic zebra finch demonstrated the well-established
sex difference in the volume of the song control nucleus HVC
between the left and right side (HVC larger on the male right side
than in the female left side) despite the fact that both brain sides had
obviously been exposed to a similar action of peripheral sex steroids
(Agate et al. 2003). This pointed to the existence of a more direct
effect of genes present on the sex chromosomes on the adult neural
phenotype.
To further test this assumption, researchers turned to genetically
modified mouse models in which either the sry gene located on the Y
chromosome had been mutated so that it could no longer induce the
formation of testes (the XYSRY- mouse) or the sry gene had been
transfected to an autosome so that XX females would develop testes
during the early embryonic life (XXSRY mouse). Togeher with XX and
XY wild type subjects, this provided four separate genotypes in which
the presence of testes or ovaries could be disentangled from the
presence of an XY or XX genotype, the four core genotype (FCG)
model.The comparative study of these mice revealed that in general,
behavioral and neuroanatomical traits directly related to
reproduction become sexually differentiated mostly under the
organizing influence of gonadal steroids as explained before (XY and
XXSRY males are different from the XX and XYSRY- females).
However the sex difference in a number of other neurobehavioral
responses such as the susceptibility to specific neural diseases or
some aspect of pain sensitivity rather seem to depend on the
complement of sex chromosomes rather than on the sex of the
gonads (XY and XYSRY- subjects are different from the XX and XXSRY
subjects) (see (Arnold & Chen 2009) for review). Future research will
have to establish whether and to what extent this mode of sexual
differentiation also applies to some of the behaviors discussed in this
chapter

Epigenetic mechanisms mediating sex differences in the

brain
Finally, we cannot conclude this section without mentioning that
steroids and related compounds can, in addition to their effects
mediated via binding to intracellular receptors that then act as
transcription factors, affect in a long-lasting, eventually
transgenerational manner the expression of specific genes and in this
way affect behavior. The expression of genes is deeply affected by
specific biochemical modifications targeting either the DNA itself
(e.g., methylation of cytosine residues) or the surrounding histone
proteins (acetylations, methylations, phosphorylations…). These
modifications of the DNA or the associated histones can be induced
by sex steroids and by a variety of compounds that mimic or inhibit
steroid action, globally grouped under the name of endocrine
disruptors. Importantly, some of these chemical modifications are
long lasting and have even been shown to be transmitted from one
generation to the next (Rissman & Adli 2014). The process of sexual
differentiation seems actually to result in part from the permanent
repression of the expression of specific genes by the early exposure to
sex steroids. The study of these mechanisms called epigenetic is
however in its early days and it is difficult at this stage to determine
exactly how important they will be for behavior control. Based on
current knowledge, they should, however, play a crucial role.

Hormone Secretion Modulated by

Behavioral Interactions

The relationship between hormones and behavior is not

unidirectional. Hormones activate behavior, but the behavior of a
congener (as well as other environmental stimuli) can also alter the
endocrine state of an animal. For example, in seasonal breeders,
hormones secreted by the pituitary gland and gonads in many
species respond to environmental stimuli such as the change in
photoperiod that synchronize the onset of reproductive behavior
with the time of year most conducive for high reproductive success.
In these species, the secretion of gonadal hormones is also
influenced by social stimuli such as the behavior of conspecifics that
can enhance or retard reproductive development (see Figure 6.4).
For example, the view of and interaction with a female will in many
species acutely and transiently increase plasma testosterone
concentration in the male, and conversely the sexual displays of the
male will promote steroid hormone release and maturation of
oocytes in females. These effects of behavior on hormone secretion
will ensure an optimal synchronization between partners of a pair, as
elegantly demonstrated as early as the 1960s by a suite of
experiments on two avian species, the ring dove (Streptopelia
risoria) studied by Daniel Lehrman and his group at Rutgers
University (Newark, NJ, USA) and the canary (Serinus canaria) that
was studied by Robert Hinde and his collaborators at Cambridge
University, UK.
Figure 6.4 Schematic model illustrating how a single endocrine
signal such as an increase in circulating testosterone concentration
can conceptually lead to the activation of a complex behavior such as
copulation in males. Testosterone can be metabolized to androgenic
(5α-dihydrotestoserone or 5α-DHT) and estrogenic (estradiol)
compounds that bind to their respective receptors. This binding will
activate a cascade of reactions ultimately leading to the association of
the occupied receptors with specific responses elements on the DNA.
These events then alter transcription and translation of specific
genes coding for a variety of proteins (enzymes, receptors, co-
regulators) that will ultimately affect neurotransmission and
behavior.
Stimulatory as well as inhibitory effects of the socio-sexual
interactions on hormone secretion have been demonstrated in both
sexes. In laboratory mice (Mus musculus) for example, the
occurrence of puberty is accelerated in the presence of an adult male
(Vandenbergh effect). In the anestrus adult female, the contact with
the male can induce the resumption of ovulation and estrus cyclicity.
This is the case in the laboratory mouse when anestrus is induced by
a social stress (Whitten effect), as well as in the ewe or the goat
during the seasonal or post-partum anestrus. During the estrus cycle
itself, the sensory stimulation among females can influence the
length of the cycle as demonstrated in the rat, or even result in
synchronized ovulations (wild boar, human). In mice, keeping
females in large groups can induce abnormalities in the estrus cycles
(Lee-Boot effect). In colonies of African naked mole rat
(Heterocephalus glaber), the presence of a single sexually active
female prevents sexual cyclicity in many other females. A similar
phenomenon is observed in family groups of small rodents: the
presence of the mother inhibits puberty in young females. Many
effects of social interactions on the endocrine system are mediated by
pheromones perceived by the accessory olfactory (vomeronasal)
system.

Examples of Species in Which the

Hormonal Regulation of Behavior Has
Been Studied

We will now end this overview by briefly describing how the

mechanisms described before integrate to coordinate specific social
behaviors, mostly in the context of reproduction, in a few well-
studied model systems.

Lordosis in rats
In rats, like in many other species, sexual activity of the female is
restricted to a limited period of time around ovulation when the
chances of fertilization are maximal and rapid changes in endocrine
conditions take place. In the female rat, increased levels of estrogens
are observed for a couple of days before a peak of progesterone that
takes place immediately before the onset of sexual receptivity.
Ovariectomy suppresses all aspects of female sexual behavior that
can be restored by a sequential treatment with estrogens and
progesterone that mimics these physiological endocrine changes
normally observed during the estrous cycle. In other species (e.g.,
prairie vole, ferret, Japanese quail), estrogens alone are fully
effective in restoring sexual receptivity in the ovariectomized female.
A substantial amount of research has been carried out to dissect the
neural mechanisms that mediate these behavioral effects of
estrogens and progesterone. Estrogens and progesterone receptors
(ERα and PR) are densely expressed in the ventromedial
hypothalamus (VMH) and detailed stereotaxic implantation studies
by Barfield and colleagues showed that their activation is usually
sufficient to stimulate lordosis in ovariectomized females. One
critical aspect of estrogen action during the cycle is in fact the
induction of PR expression in the hypothalamus. The increased
ovarian progesterone secretion will then act on a pre-sensitized brain
to activate behavior.
The activation of the lordosis response is, however, not only a
hypothalamic phenomenon and involves a complex neural circuitry.
The male mounting behavior stimulates pressure receptors in the
female’s flanks, rump, and perineum. Axons of these receptors then
project to the spinal cord, and these inputs will be relayed to the
medullary reticular formation, the hindbrain, and the midbrain
central gray area. In response to these inputs and to
estrogen/progesterone action, several brain regions, including the
VMH and mPOA, then activate spinal motoneurons innervating
deepback muscles whose contraction will result in the characteristic
lordosis position. Estrogens and progesterone act at several of these
levels and in particular estrogens increase the size of the receptive
field of flank sensory neurons, so that tactile stimuli from the male
become more prominent. In addition to these well-documented
effects on sensory inputs and motor outputs there is an effect of
steroids that induces a generalized arousal that is essential for
successful reproduction to occur.
Estrogens alone or in association with progesterone also modulate
the concentration and activity of numerous neurotransmitters and
neuropeptides including dopamine, noradrenaline, serotonin,
acethylcholine, oxytocin, substance P, β-endorphin, and the
gonadotropin releasing hormone (GnRH). These changes are often
mediated by a regulation of the transcription of corresponding
enzymes such as acetylcholine esterase (acetylcholine catabolism),
tyroxine hydroxylase and dopamine β-hydroxylase (synthesis of
dopamine and noradrenaline), catechol-o-methyl-transferase and
monoamine-oxidases (dopamine and noradrenaline catabolism), or
of the neurotransmitters/neuropeptides receptors. Through these
neurochemical changes, steroids directly and indirectly modify
synaptic transmission and thus behavior.

Male sexual behavior: the case of Japanese quail

Appetitive vs. consummatory sexual behavior
To reproduce successfully, males must produce gametes but also
detect and attract a sexually receptive female and then copulate with
her (or release gametes in a coordinated manner into the external
environment for species with external fertilization such as most
fishes). The different aspects of this interaction with the female
correspond to different phases of sexual behavior that have been
called “appetitive sexual behavior” (ASB) and “consummatory sexual
behavior” (CSB). CSB is usually followed by a period of variable
duration during which the male’s interest in the female is greatly
reduced or nonexistent (refractory period).
The distinction between ASB and CSB was originally made by
Charles Sherrington and the European ethologists of the first
generation (Konrad Lorenz and Niko Tinbergen) and was introduced
to the field of behavioral endocrinology by Frank Beach in the 1950s
(see Ball & Balthazart 2008 for more information). However, until
recently, most research has been devoted to the neuroendocrine
controls of CSB and much less is known concerning the controls of
ASB. Studies on male ASB in a diverse set of species are particularly
important from a clinical perspective. Patterns of male sexual
performance (CSB) are often stereotypic and can be species-specific.
Generalizations from non-human animals to humans are thus
sometimes difficult. In contrast, mechanisms underlying sexual
motivation and ASB seem to be more widespread among vertebrates
and could potentially be more easily transposed to humans.
Sophisticated behavioral tests have now been designed to quantify
ASB separately from CSB in quail. In the “learned social proximity
response,” a male quail will stand for extended periods of time in
front of a window that provides him with visual access to a female
after he has been copulating with that female in the same arena. This
robust, easily quantifiable response provides a useful way to
investigate the mechanisms regulating male ASB. The response is,
however, learned only after the male performs copulatory behavior
in the testing chamber and, as a consequence, cannot be studied
completely independent of the occurrence of copulation. In contrast,
the “rhythmic cloacal sphincter movements” (RCSM) are produced
in anticipation of copulation but do not require copulatory behavior
to occur in order to be produced. RCSM are greatly increased (10- to
50-fold) in males, including sexually naive males, by the simple view
of a female. These movements produce a meringue-like foam that is
transferred to females during copulation and enhances the
probability of fertilization.

Endocrine controls
In male quail, the expression of both ASB and CSB responses are
androgen-dependent. These behaviors are markedly inhibited if not
completely suppressed in castrated males and are restored to rates
observed in intact sexually mature males by treatments with
exogenous testosterone either injected systemically or implanted
directly in the mPOA. As observed for copulatory behavior, effects of
testosterone on the two measures of ASB seem to be mediated by its
aromatization into an estrogen in the mPOA. The activating effects of
testosterone on these two measures are indeed blocked by the
concurrent injection of an aromatase inhibitor or an estrogen-
receptor blocker (antiestrogen) and they can be mimicked to a large
extent by a treatment with exogenous estrogens.
Thus, both ASB and CSB (copulation) are activated by similar if not
identical endocrine stimuli. This is understandable from an ultimate
causation point of view since natural selection should favor the
control by similar hormones of behaviors that must by nature be
expressed in sequence to ensure successful reproduction. This
similarity in the endocrine controls of ASB and CSB has also been
observed in rats

Neural circuits
Despite this similarity in endocrine regulation, neural controls of
ASB and CSB should by necessity be distinct to some degree. Barry
Everitt and his colleagues at Cambridge University, UK, initially
showed that lesions to the mPOA in rats eliminate male-typical
copulatory behavior but have more limited or no effects on some
measures of sexual motivation. Rats with such lesions still pursue
and attempt to mount females. They also perform learned
instrumental responses (in operant conditioning paradigms) to gain
access to females. In contrast, lesions to the basolateral amygdala
inhibited the ability of males to acquire learned responses that are
rewarded with access to females. These observations suggested the
existence of a double dissociation between brain areas mediating
CSB (mPOA) on the one hand and ASB/arousal/motivation
(amygdala, bed nucleus striae terminalis) on the other hand.
Experimental analysis of the neuroanatomical bases of these
behaviors in a variety of species, however, also indicates a clear role
for the mPOA in the control of ASB. This was confirmed in quail
where data also suggested the existence of anatomical specializations
within mPOA areas controlling ASB vs. CSB.
In one set of experiments, discrete electrolytic lesions aimed at the
mPOA strongly inhibited, as expected, copulatory behavior and also
decreased to various degrees the expression of the learned social
proximity response. Closer inspection of the data revealed that
lesions in the caudal mPOA were associated with decreased
expression of CSB, while slightly more rostral lesions were
specifically associated with inhibition of the measure of ASB. In
another experiment, mPOA lesions also completely abolished the
expression of RCSM induced by the view of a female, but the
anatomical specificity of this latter effect could not be established
due to the limited number of available subjects.
A second set of studies correlatively indicated a localized induction of
the expression of the immediate early gene c-fos in birds that had
expressed ASB vs. CSB just before brain collection. Testosterone-
treated castrated males who had been able to interact freely with a
female and express the full sequence of copulatory behavior
displayed 90 minutes later an increased c-fos expression throughout
the rostrocaudal extent of the mPOA. In contrast, birds that were
only allowed to view the females and had therefore only performed
RCSM, one measure of ASB, showed an increased c-fos expression in
the rostral mPOA only. These data thus support the idea that the
mPOA is implicated in the expression of both ASB and CSB but the
rostral part would be more specifically implicated in the control of
ASB while copulatory behavior sensu stricto would rather be
controlled by the posterior part of this brain region
This anatomical specificity within the mPOA might not be restricted
to quail. In rats also, lesions of the caudal POA and anterior
hypothalamus block the expression of copulatory behavior, but more
rostral lesions in the POA had little or no effect in at least one
experiment. In hamsters (Mesocricetus auratus), pheromones alone
are able to activate c-fos expression in neurons of the mPOA in the
absence of copulatory interaction with females, indicating that
stimuli encountered during the appetitive phase of male sexual
behavior are processed, at least in part, in the mPOA. Similarly, in a
songbird, the house sparrow (Passer domesticus), female-directed
song, an appetitive behavior that precedes copulation, relates
positively to the induction of the immediate early genes c-fos and
zenk in the rostral POM but not the caudal part of this nucleus.

Role of dopamine
Testosterone and the derived estrogens activate behaviors by
modulating the activity of neurotransmitters. Extensive evidence
indicates that in quail, like in rats, dopamine (DA) release and
activity critically control male sexual motivation and copulatory
performance. A suite of pharmacological experiments tested the
effects on the performance of ASB and CSB of agonists and
antagonists and the D1-like and D2-like DA receptors. A consistent
effect of all drugs was observed on CSB that was stimulated by the D1
agonist but inhibited by the D2 agonists. Antagonists consistently
displayed the opposite effects (inhibition of behavior by D1
antagonist and stimulation by D2 antagonists). Far fewer effects of
the treatments were detected on measures of ASB: they were
decreased by treatment with D2 agonists but not affected by the
other treatments.
An in vivo microdialysis technique was also established and allowed
to measure changes in DA concentrations in the mPOA during social
and sexual interactions. This demonstrated that preoptic DA
concentrations increase as soon as a male detects the presence of a
female and starts interacting with her. This rise in DA actually
predicts whether copulation will or will not follow the initial
interaction and therefore represents an accurate measure of the
male’s motivation to engage in sexual behavior.

Rapid effects of estrogens

In male quail, testosterone acting via its aromatization in the mPOA
activates sexual motivation (ASB) and copulatory behavior (CSB) in
castrated males. Besides these slow, presumably genomic, effects
that develop after latencies of several days, it has also been
demonstrated that injections in the third ventricle of estrogens on
the one hand and aromatase inhibitors or antiestrogens on the other
hand, respectively, activate or inhibit within minutes ASB, as
measured by both the RCSM frequency and the learned social
proximity response, without affecting the bird’s ability to copulate. It
was correlatively shown that local aromatase activity in the mPOA
and in other hypothalamic or telencephalic nuclei acutely changes
within minutes in response to sexual stimuli (view of or sexual
interaction with a female) or to an acute stress. These in vivo changes
in aromatase activity are presumably mediated by post-translation
modifications of the enzyme, such as phosphoryations, that have
been shown during in vitro studies to be controlled namely by
changes in afferent glutamatergic and possibly dopaminergic inputs,
presumably resulting in fluctuations of the intracellular calcium
concentrations.
Together these data suggest that localized acute changes in estrogen
bioavailability have rapid behavioral consequences and therefore
that estrogens acting via membrane-initiated, presumably non-
genomic, mechanisms control sexual motivation in a manner similar
to the action of a neurotransmitter or at least a neuromodulator
(Balthazart & Ball 2006; Saldanha et al. 2011). A same steroid,
estradiol, has thus evolved complementary mechanisms to regulate
different behavioral components (motivation vs. performance) in
distinct temporal domains (long- vs. short-term) so that diverse
reproductive activities can be properly coordinated to improve
reproductive fitness.

Singing in songbirds: multiple brain sites of steroid

action
Song production in songbirds (oscines) is an active topic of research
for multiple reasons including the fact that (i) song is a learned
vocalization, like human speech, (ii) song production is sexually
differentiated in many temperate zone species, and (iii) singing is
controlled by a discrete network of brain nuclei (the song control
system) in which a high degree of neurogenesis in the adult brain has
been identified. Several aspects of song control are reviewed in the
chapters by Bolhuis (song learning) and Sherry (neural control and
neurogenesis) in this volume and we shall focus here on endocrine
controls.
Males of many songbird species display seasonal fluctuations in both
singing activity and testosterone plasma concentration that both
peak during the breeding season. Interestingly, androgen-
accumulating cells have been identified in several telencephalic
nuclei of the song control system, namely HVC (now used as a proper
name) and the robust nucleus of the arcopallium, RA (see Figure
6.5). These cells were later confirmed to express androgen receptors
as identified by immunohistochemistry and detection by in situ
hybridization of the corresponding mRNA.
Figure 6.5 Reciprocal interactions between hormones and
behavior in the control of reproduction illustrated based on the
canary literature. In the spring, the increasing daylength activates
the hypothalamo-pituitary-gonadal axis of both males and females.
The resulting increase in circulating concentrations of testosterone in
the male activates courtship and singing behavior. Perception of
these songs will in turn affect the female endocrine physiology and
behavior, including increased expression of immediate early genes
(IEG) such as Fos or Zenk (also known as egr-1) in the secondary
auditory area NCM (nidopallium medialis pars caudalis) and
increased concentration of estradiol-17β (E2) locally in the brain (in
NCM) and in the periphery. These endocrine changes will result in
the expression of the female-typical copulation solicitation display
that will invite males to mount and also feedback in a positive
manner on their plasma testosterone concentrations.
The presence of high densities of steroid receptors in these
telencephalic areas represents a remarkable specialization among
vertebrates; a dense expression of sex steroid receptors is only found
from fishes to mammals in diencephalic and limbic brain regions. A
similar exception to the general pattern has, however, also been
uncovered in other vertebrate species that produce vocalizations in
the context of reproduction, namely the midshipman fish (P.
notatus) and several species of amphibians such as the African
clawed frog (Xenopus laevis). In all these species, vocalizations are
produced by a complex neuronal circuitry that includes a number of
androgen-sensitive nuclei in telencephalic brain regions that do not
normally contain androgen receptors in other vertebrates.
Because the volume of these androgen-sensitive nuclei undergoes a
marked seasonal plasticity, it was initially suggested that they are the
main/sole site of steroid action for the activation of singing.
However, things are not so simple; steroids may act at multiple levels
of the songbird brain to regulate specific aspects of song. Blocking
testosterone action in HVC reduces song quality but does not affect
song rate and conversely implanting testosterone in HVC or RA of
castrated white-crowned sparrows does not activate singing.
Moreover, lesions of HVC in canaries creates birds that produce
“silent song,” i.e., they assume all of the postural components
associated with song production but do not produce audible
vocalizations. It was also shown that lesions of the mPOA decrease
the motivation to sing in starlings (Sturnus vulgaris). Since
testosterone action in this brain region activates both the appetitive
and consummatory aspects of male sexual behavior and bird song
can be considered as an appetitive response by which the male tries
to attract a female, recent work tested the effects on signing of
testosterone implants in the mPOA of castrated male canaries. These
implants activated an intense singing activity quite comparable to
the activity seen in birds systemically treated with testosterone but
song quality was poor and could only be restored to nearly intact
levels by an additional testosterone implant in HVC. It seems,
therefore, that testosterone action is required at multiple levels
(HVC, POM, and also probably at the syrinx) to produce normal song
with occurrence frequencies typical of the reproductive condition
and songbirds represent an excellent model system to analyze the
interaction of these multiple actions of the steroid.

Alternative Strategies of Reproduction

Besides these general mechanisms of behavior control that apply to

the majority of vertebrate species, a few species or families of
vertebrates have adopted unusual modes of reproduction usually in
reponse to unusual environmental conditions. Selected examples are
gathered in this last section.

Associated vs. dissociated reproductive cycles

In some species, the period of reproduction during the annual cycle
does not correspond to the maximal endocrine activity of the gonads.
One of the best-documented examples concerns the red-sided garter
snakes (Thamnophis sirtalis) that copulate in the early spring almost
immediately after they emerge from the den where they hibernated
(Crews 2005). No gonadal endocrine signal is apparently required to
activate copulation. Indeed, castration in the fall before hibernation
does not prevent spring copulation, and treatment with exogenous
testosterone does not affect the rate of behavior expression.
Spermatogenesis and steroid secretion take place during the summer
months and precede by at least three-quarters of a year the period
when copulation actually takes place. Recent research suggests that
during hibernation, brain aromatase activity continues at a very low
pace and progressively accumulates estradiol in the brain so that this
steroid would be responsible for the behavioral activation in the early
spring. This would explain why a period of “vernalization” (exposure
to cold) of at least a few weeks must take place before the behavior is
expressed but additional work is needed to confirm this idea. A
similar dissociated pattern of reproduction has also been identified
in some bat species.

Sex change and successive hermaphroditism

A large diversity of reproductive patterns is found in fish species that
have no sex chromosomes (Bachtrog et al. 2014) and change sex
during their life either as a function of age, of environmental factors,
or of the social situation. This “sequential hermaphroditism” can
involve a change from male to female (protandry) or female to male
(protogyny) and is associated with huge variations in plasma
concentrations of sex steroids and with a substantial remodeling of
brain neurochemistry affecting variables such as the expression of
various neuropeptides (vasotocin, gonadotropin-releasing hormone)
or the intracellular metabolism of steroids. In all these species,
testosterone or 11KT remains, however, the steroid responsible for
the activation of male sexual behavior, and the POA is a key area
controlling reproduction. A few piscine species (deep-sea, some
serraninae, or sea basses) have also been identified that seem, based
on examination of gonadal tissue, to exhibit simultaneous
hermaphroditism (simultaneous presence of testicular and ovarian
tissue in the same subjects). Little or no information is, however,
available on the endocrine control of reproductive behavior in such
species.

Parthenogenesis
Parthenogenetic reproduction, that is, asexual reproduction in which
females can reproduce without fertilization by a male, is sometimes
observed in invertebrates (aphids, some bees, and parasitic wasps)
but is very rare in vertebrates (a few sharks and reptiles species). It
has, however, been extensively documented in one species of lizard.
Cnemidophorus uniparens females are able to lay eggs that will
produce female offspring in the absence of sperm. They also display,
over time, cycles of sexual activity during which they successively
assume the role played by males in closely related lizard species and
they mount a female that is about to lay eggs or they will be mounted
and will lay eggs. The female-like receptive behavior is displayed just
before ovulation when circulating estrogen concentrations are high,
while the male-typical mounting behavior is displayed when plasma
concentrations of progesterone are elevated. Accordingly, these two
types of behaviors can be induced in the laboratory by injecting
ovariectomized females either with estrogens or with progesterone.
Progesterone rather than testosterone is thus responsible for the
activation of male-typical copulatory behavior in this species.
Progesterone, however, still acts in the POA that expresses a high
density of progesterone receptors. These pseudocopulations play an
important functional role in reproduction: females that undergo
mounting release more eggs and produce more offspring than
females that do not engage in this behavioral interaction.

Alternative reproductive phenotypes

Some selected species also adopt multiple reproductive phenotypes
within the same sex, including large males that defend territories and
smaller (“satellite”) males that often display morphological and
behavioral features of females (smaller size, absence of colorful
displays) and steal copulations from the larger dominant male. Such
social systems have been observed in most, if not all, vertebrate
classes. However, the hormonal bases of these alternative
reproductive tactics have been studied in detail in only a few cases
(the plainfin midshipman fish, P. notatus, or the tree lizard, U.
ornatus). In these examples, sneaker males were shown to display
reduced levels of androgenic steroids (testosterone or 11KT in fishes)
as compared to dominant males. A somewhat similar behavioral and
morphological polymorphism is seen in a shore bird, the ruff
(Philomachus pugnax). Ruffs have two types of males: territorial
males with dark (brown or black) long fluffy feathers in their neck
(“the ruff”) and occipital “head tufts” that defend small territories on
leks that will be visited by females during the breeding season and
satellite males (approximately 16% of the population) in which these
long feathers have a light white or creamy color that do not defend
such territories but stay in the vicinity and try to sneak copulations
with females visiting the lek. Morph type inheritance is consistent
with a single-locus two-allele autosomal genetic model and
differences between morphs do not seem to reflect differences in
testosterone, athough the hormone is needed to activate the
phenotype. If females that do not show these male morphological
attributes are treated with a standardized dose of testosterone, they
will display the phenotype corresponding to their genotype.
SUMMARY AND CONCLUSIONS
Hormones are chemical messengers secreted by endocrine glands
that change the probability and intensity of a behavioral response
to a particular stimulus. There are several types of hormones
known to influence behavior including steroid hormones,
peptide/protein hormones, and thyroid hormones. Steroids are
among the best studied hormone systems in relation to
behavioral activation, and we illustrate general principles about
how steroid hormones can organize a suite of traits into an
adaptive response by focusing on testosterone and reproduction.
Steroids have two different types of effects in adulthood: one is
relatively fast (seconds to minutes) and involves actions at the
cell membrane; and in the other type, action is relatively slow
(hours to days to weeks) and involves the initiation of gene
transcription. These adult effects are known as activational
effects. Steroid hormones can also act during ontogeny to
produce enduring sex-specific effects on the response of an
individual to hormones. So for example, female rodents will not
respond to male concentrations of testosterone in adulthood with
male-typical behaviors because their brain has been organized in
a female-typical fashion and does not have the capacity to exhibit
male-like traits. These early ontogenetic effects are known as
organizing effects of steroid hormones.
Hormone effects on behavior are often multi-dimensional and
complex. So for example, in the case of birdsong, testosterone can
act in several brain areas simultaneously as well as in the
periphery to independently regulate different aspects of song
behavior such as the motivation to sing, aspects of the quality of
song, and the loudness of song. In the case of testosterone, the
hormone secreted by the gonad is often metabolized locally at the
target brain area or in peripheral structure into an androgenic or
estrogenic metabolite that are the effectors of the behavioral
change at the cellular level.
FURTHER READING
An excellent in-depth scholarly summary of the ideas and data
reviewed here can be found in Hormones and Animal Social
Behavior (Adkins-Regan 2005). A good recent textbook that covers
the field is An Introduction to Behavioral Endocrinology (Nelson
and Kriegsfeld 2016). A scholarly book on testosterone’s many
behavioral effects that occur via regulation by estrogenic metabolites
can be found in Brain Aromatase, Estrogens and Behavior
(Balthazart & Ball 2012). An excellent summary of how
glucocorticoid hormones regulate the stress response in many
species can be found in Romero and Wingfield (2016). Young and
Alexander (2014) is a text on neuropeptides and affiliation designed
for a popular audience. Donald Pfaff has written several books in
recent years that provide insightful reviews into hormones and many
aspects of behavior.

ACKNOWLEDGMENTS
Our research reviewed in this chapter has been supported by grants
from the NIMH (MH 50388) and NINDS (NS104008) as well as by
support from the Belgian FNRS and the IAP program. We thank our
students and post-docs for their hard work and good discussions
about the data we have reviewed here.

REFERENCES
Adkins-Regan, E. 2005. Hormones and Animal Social Behavior.
Princeton, NJ: Princeton University Press.
Agate, R.J., Grisham, W., Wade, J. et al. 2003. Neural, not gonadal,
origin of brain sex differences in a gynandromorphic finch.
Proceedings of the National Academy of Sciences USA, 100,
4873–8.
Arnold, A.P. & Chen, X. 2009. What does the “four core genotypes”
mouse model tell us about sex differences in the brain and other
tissues? Frontiers in Neuroendocrinology, 30, 1–9.
Bachtrog, D., Mank, J.E., Peichel, C.L. et al. 2014. Sex
determination: why so many ways of doing it? PLoS Biology, 12,
e1001899.
Ball, G.F. & Balthazart, J. 2008. How useful is the appetitive and
consummatory distinction for our understanding of the
neuroendocrine control of sexual behavior? Hormones and
Behavior, 53, 307–11.
Ball, G.F., Balthazart, J. & McCarthy, M.M. 2014. Is it useful to view
the brain as a secondary sexual characteristic? Neuroscience and
Biobehavioral Reviews, 46, 628–638.
Balthazart, J. & Ball, G.F. 2006. Is brain estradiol a hormone or a
neurotransmitter? Trends in Neurosciences, 29, 241–9.
Balthazart, J. & Ball, G.F. (eds.). 2012. Brain Aromatase, Estrogens
and Behavior. Oxford: Oxford University Press.
Balthazart, J. & Young, L.J. 2014. Mate selection, sexual orientation
and pair bonding. In: T.M. Plant & A.J. Zeleznik (eds.), Knobil
and Neill’s Physiology of Reproduction, pp. 2157–2210.
Amsterdam: Elsevier.
Baum, M.J. & Bakker, J. 2013. Roles of sex and gonadal steroids in
mammalian pheromonal communication. Frontiers in
Neuroendocrinology, 34, 268–84.
Becker, J.B., Berkley, K.J., Geary, N., Hampson, E., Herman, J.P. &
Young, E.A. 2008. Sex Differences in the Brain. From Genes to
Behavior. Oxford: Oxford University Press.
Cornil, C.A., Ball, G.F. & Balthazart, J. 2015. The dual action of
estrogen hypothesis. Trends in Neurosciences, 38, 408–16.
Crews, D. 2005. Evolution of neuroendocrine mechanisms that
regulate sexual behavior. Trends in Endocrinology and
Metabolism, 16, 354–61.
Etgen, A.M. & Pfaff, D.W. 2009. Molecular Mechanisms of Hormone
Actions on Behavior. Amserdam: Elsevier.
Forlano, P.M., Sisneros, J.A., Rohmann, K.N. & Bass, A.H. 2015.
Neuroendocrine control of seasonal plasticity in the auditory and
vocal systems of fish. Frontiers in Neuroendocrinology, 37, 129–
45.
Goodson, J.L. & Kabelik, D. 2009. Dynamic limbic networks and
social diversity in vertebrates: from neural context to
neuromodulatory patterning. Frontiers in Neuroendocrinology,
30, 429–41.
Goy, R.W. & McEwen, B.S. 1980. Sexual Differentiation of the Brain.
Cambridge, MA: MIT Press.
Hinde, R.A. 1970. Animal Behaviour, 2nd ed. New York: McGraw-
Hill.
McEwen, B.S., Luine, V.N. & Fischette, C.T. 1988. Developmental
actions of hormones: from receptors to function. In: S.S.J. Easter,
K.F. Barald & B.M. Carlson (eds.), From Message To Mind, pp.
272–87. Sunderland, MA: Sinauer.
Nelson, R.A. and Kriegsfeld, L.J. 2016. An Introduction to
Behavioral Endocrinology, 5th ed. Sunderland, MA: Sinauer.
Pfaff, D.W. 1980. Estrogen and brain function. New York: Springer.
Pfaff, D.W. & Pfaffmann, C. 1969. Olfactory and hormonal influences
on the basal forebrain of the male rat. Brain Research, 15, 137–56.
Phoenix, C.H., Goy, R.W., Gerall, A.A. & Young, W.C. 1959.
Organizational action of prenatally administered testosterone
propionate on the tissues mediating behavior in the female guinea
pig. Endocrinology, 65, 369–82.
Rissman, E.F. & Adli, M. 2014. Minireview: transgenerational
epigenetic inheritance: focus on endocrine disrupting compounds.
Endocrinology, 155, 2770–80.
Romero, L.M. & Wingfield, J.C. 2016. Tempests, Poxes, Predators
and People: Stress in Wild Animals and How They Cope. Oxford:
Oxford University Press.
Saldanha, C.J., Remage-Healey, L. & Schlinger, B.A. 2011.
Synaptocrine signaling: steroid synthesis and action at the
synapse. Endocrine reviews, 32, 532–49.
Young, L. & Alexander, B. 2014. The Chemistry Between Us; Love,
Sex and the Science of Attraction. Penguin.
7
development of behavior
JOHAN J. BOLHUIS
INTRODUCTION
What is behavioral development? Roughly, it is about the changes
in behavior and particularly its underlying mechanisms in
individuals from conception to death. This definition includes the
behavior of individuals before they are born, which may come as
a surprise. In fact, the embryo has a rich behavioral repertoire,
with which it communicates with its parents or siblings. Also,
embryos can learn and remember. Later in this chapter we shall
discuss some examples of behaviors studied in what is known as
behavioral embryology. Is behavioral development important?
Classical ethological theory had surprisingly little to say about the
development of behavior, in spite of the fact that early in the last
century Konrad Lorenz (1935) had published a landmark paper
on the phenomenon of imprinting, a key concept in behavioral
development that is discussed in detail later in this chapter. For
instance, Niko Tinbergen’s (1951) famous book, The Study of
Instinct, has only one short chapter on development, and only
one paragraph on imprinting. Lehrman (1953), in his influential
critique of ethological theory, pointed out this neglect of
developmental questions, which subsequently led many
behavioral biologists to consider problems of development (Kruijt
1964; Bateson 1966). It also led Tinbergen (1963), some 10 years
later, to reformulate his views on the aims of ethology. In the first
chapter of this book, we described how in this seminal paper
Tinbergen considered development to be so important that he
added it to Huxley’s three questions in biology and made it one of
the four main problems in the study of animal behavior. So it is
important, then. It is certainly the case that some of the major
scientific discussions in animal behavior involve different
concepts of development, and we will discuss these debates. In
addition, we shall see that many concepts and findings
concerning behavioral development in animals have had
important consequences for the study of human development.
 Basic Developmental Issues

Learning and development

Learning is often interpreted as being part of behavioral
development; this is because learning, like other developmental
processes, involves changes in the mechanisms underlying behavior
over time. For example, during associative learning (see Chapter 8),
representations are formed of the stimuli or events that are
associated, and these representations are somehow linked to each
other. Representations are cognitive structures (see Chapter 9) in
which the external or internal environment of an individual is
somehow coded or represented. Thus, after successful learning there
will be behavior patterns and psychological and neural structures in
the individual that were not there before the learning episode, which
according to our broad definition above is a form of development.
Many developmental processes, such as filial and sexual imprinting
and birdsong learning, explicitly involve learning, but they also
involve other mechanisms. This raises the difficult question of the
difference between learning and other forms of development. The
difference is gradual and has to do with the specificity of the
representations. In learning, representations are formed (of the
stimuli or events and the relationship between them) that are
specific to those external events. That is, the representation of the
sound of the bell in Pavlov’s dogs (see Chapter 8) is only addressed
when the dog hears the bell (or perhaps something that sounds very
much like it, in the case of generalization). In other forms of
development the influence of external experience may be
nonspecific, or internal causal factors may be involved. For instance,
an increase in sex hormones around the time of birth may lead to the
development of sexual behavior in a certain direction; one would not
call this learning. These differences are not substantial, however.
One could say that learning is a subset of behavioral development in
general. Learning as such is discussed in Chapter 8, while imprinting
and song learning are dealt with separately later in this chapter.
Embryology and behavioral development
The Dutch ethologist Jaap Kruijt (1964) argued that the study of
behavioral development could benefit greatly from concepts
developed in embryology. Indeed, many concepts in the
contemporary study of behavioral development originated in
embryology, particularly the work of the Scottish embryologist
Conrad Waddington (1966). Waddington (1966) was far ahead of his
time and wrote very clearly about the basic concepts in embryology,
which are still important and which have proved to be crucial also for
the understanding of the development of behavior after birth.
Waddington pointed out that a major dichotomy in developmental
thinking ever since Aristotle has been the distinction between
preformation and epigenesis. Preformation refers to the idea that at
the time of fertilization, the egg or the sperm already contains the
basic features of the adult individual. Shortly after the invention of
the microscope, some eighteenth-century anatomists believed they
could see the shape of an adult individual in an egg, or even a little
man coiled up in a sperm! This is of course, as Waddington put it,
“mere imagination, dependent on the inadequacy of their
instruments.” However, later versions of preformationism held that
certain aspects of the adult are somehow represented in the embryo.
Epigenesis, in contrast, believes that none of the adult features of
individuals are represented in the embryo but rather that they
develop. As Waddington suggests, embryological development is in
fact a combination of preformation (such as the expression of genes)
and epigenesis, and the same can be said for the development of
behavior.
An important concept introduced by Waddington is that of
canalization. This basically means that under certain circumstances a
particular developmental process will follow a fairly fixed and
predictable path. This is reflected in the fact that many
developmental processes are quite consistently similar between
individuals of the same species. This is certainly the case in
embryonic development. However, certain events can disturb this
stereotyped (“canalized”) developmental process, and development
can proceed in a different direction. We will later see examples where
in the absence of normal conditions, perceptual preferences in chicks
or ducklings may develop in quite a different way. To illustrate the
principle of canalization and its possible disturbance, Waddington
(1966) introduced the concept of the epigenetic landscape (Figure
7.1). This is a metaphor for developmental processes, with balls
rolling down a hill representing the development of different
morphological traits. A similar metaphor could be used to illustrate
the development of different behaviors. Waddington stressed the
correcting effect of the walls of the valley on the path of the rolling
ball. When a ball moves to the left, the slope of the wall pushes it
back on to the right course. Such regulatory mechanisms lead to
balanced development overall. A particular ball can follow a fixed
path down the slope in one of the valleys, but it can hit a ridge and be
diverted in another direction and follow quite a different path. A
major question in embryology is that of differentiation: how is it that
various parts of the embryo develop into different organs, or, using
the metaphor of the epigenetic landscape, how do different balls end
up in different valleys? Essentially, differentiation comes about
through the “switching on” or expression of different genes in
different cells of the embryo or at different times during
development.
Figure 7.1 “Epigenetic landscape” suggested by Waddington
(1966). See text for further explanation.
A related concept in embryology is that of induction. Waddington
defines embryonic induction as the process by which one part of an
embryo influences a neighboring part, thereby making that part
develop into an organ when it would not otherwise have done. For
example, cells taken from an egg of a newt (a small amphibious
animal) that would normally develop into skin cells can be induced
to develop into neurons when placed in a salt solution together with
cells taken from a different part of the embryo known as the crescent
region (because it is shaped like a crescent in many species; this
region normally develops into the structures of the backbone).
Alternatively, one can transplant crescent region cells into another
embryo, such that these cells are positioned next to a part of the host
embryo that would normally develop into the skin. In this case also
the presence of the crescent region cells leads to induction, such that
the host cells now develop into the brain. Gottlieb (1980, 2002a)
later adopted the term “induction” to denote the effects of
nonspecific external experience on perceptual development, as we
shall discuss later. Related to induction is a phenomenon called
competence. This refers to the fact that certain parts of the embryo
can cause induction of different other parts at different times.
Waddington (1966) gives the example of a piece of tissue of a newt
embryo called ectoderm. When this ectoderm tissue is removed and
immediately wrapped around the brain–eye region taken from an
older embryo, the host tissue will develop into a brain and eye.
However, when the ectoderm is kept in a saline solution for 36 hours
before it is wrapped around the host tissue, the latter will develop
into a lens rather than brain and eye. This is known as competence:
the host tissue is first competent to develop into a brain and eye but
at a later stage is competent to develop into a lens. In terms of gene
expression this means that certain genes can only be switched on at
particular times during development. This is an example of sensitive
periods, which can also occur in behavioral development and which
we will discuss in detail later in this chapter.

Genes and behavior

Especially in the popular literature one often comes across the word
“innate” to characterize certain behaviors. For instance, it is
sometimes suggested that aggression is “innate.” This is unfortunate,
because it is not clear what is meant by “innate,” which does not
make it a very useful term. In addition, we shall see that “innate” is
actually meaningless as a developmental process. Bateson (1999)
listed seven possible different meanings of the word “innate”: (i)
present at birth; (ii) a behavioral difference caused by a genetic
difference; (iii) adapted over the course of evolution; (iv) unchanging
throughout development; (v) shared by all members of a species; (vi)
present before the behavior serves any function; and (vii) not
learned. Discussions of the concept of “innate” are not made any
easier by the fact that authors rarely state which of these possible
meanings they are using. In popular discussions, “innate” usually
implies that a behavior is somehow “genetically determined” or “in
our genes.” One should keep in mind that there is no direct, one-to-
one effect of genes on behavior. Genes are sections of DNA that code
for proteins, not for behavior patterns. This is not to say that
behavior has nothing to do with the expression of genes, but the
relationship between the two is far more complex than the naive idea
that there are genes for certain behaviors, as one often reads in the
popular press. For a start, it is likely that several genes are somehow
involved in a certain behavior and that they interact with each other
as well as with the animal’s internal and external environment. Also,
as we have seen already in the previous section, during
embryological development, gene expression is very much dependent
on where in the embryo the cells are, and on the particular time
during development. This is likely to be the case in behavioral
development as well.
What one can say is that differences in behavior may vary with
genetic differences. For instance, it is possible to select animals
artificially on certain behavioral traits, such as aggression.
Lagerspetz (1964) selected mice on certain characteristics of
aggressive behavior, in particular the time it took for an individual to
attack another individual in a standard situation. Fast attackers were
mated with each other, and so were slow attackers. After selecting in
this way over a number of generations, two subpopulations
developed, with one subpopulation showing a high aggression score
and the other showing a much lower aggression score. Clearly, the
differences between the two behavioral extremes are related to
genetic differences. However, we do not know much about the
relationship between genes and behavior in this case, which is likely
to be complex. For instance, it is not clear which aspect of aggressive
behavior, or its underlying mechanisms, was actually selected for.
Work by van Oortmerssen and collaborators (Benus et al. 1990)
shows that the two extremes of mice selected for attack latency also
differ in a number of behavioral and physiological measures other
than aggressive behavior. For instance, mice that have a high
“aggression” score also show more exploratory activity when
confronted with a novel environment. Thus it is not clear what was
actually selected for: a whole complex of genes, or some factor that
varies with genetic differences and which is somehow involved in a
host of different behaviors. Also, we do not know the exact nature of
the genetic difference between the two extremes. Suppose the
difference is in only one gene (or one allele), we do not know for
which protein (or proteins) this gene codes (and it may do so in
collaboration with other genes). This protein (or proteins) could have
all kinds of functions; for example, it could be a receptor for
testosterone, or an enzyme or receptor in the cell membrane. These
cellular effects need not have an effect on aggressive behaviors but
could also somehow affect any of a whole host of other behavioral
factors involved in aggression or, alternatively, in various other
behaviors. All these considerations make it very difficult to say
anything meaningful about a direct role of genes in behavior.
Importantly, they show that statements about behavior being
“genetically determined” or “innate” are not particularly helpful. For
this reason, whenever the word “innate” is used in this chapter
(because it is still widely used in the literature) it will be placed in
quotation marks.

The “nature/nurture” debate

The early days of ethology saw a battle of ideas between those who
thought that behavioral development could be analyzed in terms of
“innate” and acquired components (Lorenz 1935, 1937b) and those
who thought that development was more complex (Lehrman 1953,
1970). Other terms used in this dichotomy are “innate” versus
“learned,” “genes” or “genetically determined” versus “environment,”
and “instinct” versus “learning.” A general expression for this
dichotomy is “nature versus nurture.” Konrad Lorenz (1935, 1937b)
postulated that behavior could be considered a mixture of “innate”
and acquired elements (Instinkt-Dressur-Verschränkung,
intercalation of fixed pattern and learning) and that analysis of the
development of the “innate” elements (fixed patterns) was a matter
for embryologists. Lorenz suggested that the influence of the
environment could be controlled for by rearing animals in isolation.
This is known as a Kaspar Hauser experiment, after the boy who was
found abandoned at the market in Nürnberg (Germany) in 1828. No
one seemed to know the boy, who was very confused and could not
tell where he came from. Kaspar Hauser died at the age of 20. Up to
this day his story is shrouded in mystery, with numerous conflicting
publications and an extensive folklore as to who he was and where he
came from. The myth arose that he was reared in social isolation;
hence the use of his name for experiments involving isolated
animals. Lehrman (1953), in an important critique of ethological
theory, argued that it is impossible to rear an individual in complete
isolation from the environment: it simply would not survive, because
“structures and activity patterns” (Lehrman 1953) at any time during
development need to interact with the internal and external
environment of the organism. Lehrman argued that terms such as
“innate” or “genetically fixed” obscure the importance of
investigating processes in order to understand mechanisms of
behavioral development. He suggested a much more complex
process of continual interaction between the organism and its
internal and external environment at every stage of development.
Lehrman stressed that the interaction during behavioral
development is not between what he called “heredity” and
environment, but between the organism and its (internal and
external) environment. Furthermore, he noted that an organism is
different at each different stage of development. Lehrman’s view
evokes a much more complex and dynamic picture of behavioral
development than Lorenz’s suggestions (Figure 7.2).
Figure 7.2 Schematic illustration of two different views of
behavioral development. (a) Simple view that the individual and its
behavior are influenced by both genes and environment. (b) More
complex view of development such as suggested by Lehrman (1953,
1970). In this view the individual organism (represented by the blue
forms) continually interacts (represented by bidirectional black
arrows) with its internal and external environment (represented by
lighter, larger two-dimensional yellow-orange shapes). The
individual (and its environment) are different at different times
during development.
In reaction to Lehrman’s (1953) critique, Lorenz (1965) changed his
formulation somewhat. He argued that the information necessary for
a behavior element to be adapted to its species’ environment can
only come from two sources: from information stored in the genes or
from an interaction between the individual and its environment.
Lorenz insisted that it is possible to isolate an individual from
relevant environmental information; one could then determine
whether the behavior element still developed in its normal adaptive
form. He also argued that learning itself is a mechanism that adapted
in the course of evolution. He used the expression “innate
schoolmarm” to indicate that the organism in some sense “knows
what is good for it.” The “schoolmarm” was thought to determine
what an animal does or does not learn.
Lehrman (1970) published a reaction to Lorenz’s book, his last major
paper before his untimely death. In this important paper, Lehrman
notes that both he and Lorenz agree on the facts, which means that
their disagreements must depend on their differing interpretations of
those facts and on the different meanings they give to the same
words. That is, Lehrman thought that Lorenz and he were really
interested in two different problems. Lehrman was interested in
studying the effects of all types of experience on all types of behavior
at all stages of development, very much from a causal perspective,
whereas Lorenz was interested only in studying the effects of
experience on behavior at the stage of development at which
behavior patterns begin to function as modes of adaptation to the
environment (see Hogan 1988, 2001 for further discussion).
Beyond nature/nurture
It is important to realize that the debates between the Lorenz and the
Lehrman camps were not about nature versus nurture, although it is
the case that European ethologists tended to emphasize “instinct”
whereas North American comparative psychologists focused on
learning (see Chapter 1). Rather, the debates were about viewing
development as having both “nature” and “nurture” elements as
opposed to regarding development as a much more complex and
dynamic process. Despite these long-standing debates, and modern
developments in behavioral biology, Lorenz’s view corresponds to
the way many people, laymen as well as scientists, still think about
development. In the popular press one can still regularly read stories
about how a certain behavior is thought to be “innate,” thereby
completely ignoring development. In behavioral biology, on the
whole authors are much more sophisticated when it comes to these
issues, but many of them still find it difficult to abandon Lorenzian
terminology (and thus presumably Lorenzian concepts), even when
they want to be “interactionists.” One often finds that the old
Lorenzian dichotomy is merely rephrased in interactionist terms, for
instance when development is seen as a continuous interaction
between “innate predispositions” and “environmental factors.” We
have already seen that use of the term “innate” is problematic.
Furthermore, Lehrman (1953) has shown that during development
the “interaction” is much more complex than that. The problem is
that Lehrman’s views are not easily expressed in a catchy sound bite,
so this issue is likely to remain contentious.

Developmental discontinuities and ontogenetic

adaptations
Is behavioral development a continuous process, with different
phases following each other in a predictable fashion, and behaviors
at a certain stage emerging out of preceding behaviors? In this view,
behaviors at one stage of development are a kind of preparation for a
subsequent stage, the former being essential for the emergence of the
latter. Alternatively, are there discontinuities in the development of
behavior, with abrupt changes, where there is no causal relationship
between two subsequent behaviors (and between their underlying
mechanisms)? Both these views have been proposed, and it is likely
that behavioral development is a mixture of both. Hinde and Bateson
(1984) have discussed this issue in some detail. They suggest that
overt changes in behavior during development do not necessarily
imply changes in the underlying mechanisms. This makes the
investigation of developmental (dis)continuities not as
straightforward as it may seem at first sight.
In the first chapter of this book, we argued that Tinbergen’s four
main questions in the study of animal behavior are all important, but
that they should also be separated conceptually. It has been argued
that behavioral development is essentially a causal problem, i.e., it
has to do with mechanisms, not with function or evolution (Hogan
1988, 1994). Nevertheless, it is important to realize that different
phenotypes at different developmental stages can be subject to
different selection pressures. Consequently, behavior at a particular
developmental stage may have a different adaptive value compared
with earlier or later stages. This was recognized by the embryologist
Oppenheim (1981). He suggested that stages in development are
often not merely a kind of immature preparation for the adult state,
although they certainly can be, but that each developmental phase
may involve unique adaptations to the environment of the
developing animal, i.e., ontogenetic adaptations. As a consequence,
certain early behavior patterns may disappear in the course of
development, a phenomenon that Oppenheim termed a
“retrogressive process.”
A clear example of ontogenetic adaptation is metamorphosis such as
occurs in insects and amphibians. For instance, a butterfly emerges
from a pupa, which is itself preceded by the caterpillar stage. In each
of these stages the animal looks completely different and also
behaves in a completely different way. For instance, a butterfly flies
but a caterpillar clearly does not, whereas a pupa does not seem to
show much behavior at all. It is unlikely that the movements of the
caterpillar are a necessary prerequisite for the flight of the butterfly.
Rather, their respective behavior patterns are adaptations for those
particular stages in development. However straightforward the
example of metamorphosis may seem, Hinde and Bateson (1984) cite
examples of underlying continuities even here, where experience in
the larval phase affects adult behavior. For instance, adult female
moths of certain species lay their eggs on substrates on which they
were reared as larvae. Oppenheim (1981) also suggested that
imprinting and birdsong learning (both discussed later in this
chapter) may be other examples of ontogenetic adaptations because
they occur only early during ontogeny.
A good behavioral example of ontogenetic adaptation is provided by
Hall and Williams’ (1983) review demonstrating that in rat pups
“suckling isn’t feeding.” Suckling behavior is qualitatively different
from later feeding behavior and is the sole means of food intake until
weaning, which according to Oppenheim (1981) qualifies it as an
ontogenetic adaptation. Hall and Williams (1983) discuss two
possible developmental scenarios for suckling and feeding in rats.
One possibility is that suckling merges into feeding at weaning, and
the two behaviors share internal and external causal factors.
Alternatively, the two behaviors are relatively separate and they
share only some internal and external causal factors. Hall and
coworkers found that adult ingestion is not a continuation of
suckling, as the two behaviors have different internal and external
causal factors. Furthermore, if pups were deprived of suckling by
feeding them with a cannula, later feeding behavior emerged
normally, suggesting that suckling is not a necessary antecedent for
adult feeding (Hall 1979).

Imprinting

Imprinting has often been regarded as a showcase for behavioral

development in general. The phenomenon of filial imprinting has
been known for a long time and was described as early as 1518 by Sir
Thomas More in Utopia. However, imprinting was investigated
experimentally much later by the British amateur biologist Douglas
Spalding in 1873 and by the German naturalist Oskar Heinroth in
1911. Konrad Lorenz, who gave the phenomenon its name,
subsequently provided a detailed description of imprinting in a
number of bird species in an influential work published in 1935; a
shorter, English version was published in 1937 (Lorenz 1937a).
Contemporary researchers define imprinting as the process through
which the social preferences of young animals become restricted to a
particular stimulus or class of stimuli (Bateson 1966; Bolhuis 1991).
Images of Konrad Lorenz being followed around by a group of
goslings are well known. These images illustrate some of the
characteristics of imprinting, but they have also led to some
confusion. Most of the research into imprinting has indeed been
conducted with young individuals of precocial bird species, such as
geese, ducks, or chickens. These are species where the young can
move around and are relatively independent soon after hatching. The
example of Lorenz also shows that, apparently, young birds can form
a social bond with something that does not resemble their natural
mother at all. However, these images also suggest that the goslings
have acquired an irreversible social bond with Lorenz, that they treat
him as their mother or even as a potential mating partner, and that
they ignore members of their own species. Indeed, Lorenz thought
that imprinting was a unique process, for at least three reasons.
First, he thought that imprinting was unlike any other form of
learning, such as classical conditioning (see Chapter 8), as it did not
require any kind of reward or reinforcement. Second, imprinting was
thought to occur only during a particular period in development,
what Lorenz called a critical period. Third, Lorenz thought that the
process of imprinting was irreversible: once the animal had formed a
social bond with a particular individual, it could not form a bond
with another individual. Lorenz also thought that the main outcome
of imprinting was that the animal learned the characteristics of its
future mating partner. Nowadays a distinction is made between filial
imprinting, where a social bond is formed between the young animal
and its parent, and sexual imprinting, which involves the formation
of sexual preferences that are expressed later in life.

Filial imprinting
Although filial imprinting may occur in mammals (Sluckin 1972), it
has been studied mostly in precocial birds. Soon after hatching, these
birds will approach and follow an object to which they are exposed.
In a natural situation the first object the young bird encounters is
usually its mother. In the absence of the mother, other animals or
even Lorenz can be adequate mother-surrogates. Amazingly,
inanimate mother-surrogates such as colored balls or illuminated
boxes are also effective in eliciting approach and following behavior
and the animals will readily form a social preference for these
unnatural stimuli (Bateson 1966; Sluckin 1972; Horn 1985, 2004;
Bolhuis 1991; McCabe 2019). When the chick or duckling is close to
an appropriate object (see below), it will attempt to snuggle up to it,
frequently emitting soft twitters. Initially the young bird approaches
a wide range of objects. After the bird has been exposed to one object
long enough, it remains close to this object and may run away from
novel ones. If the familiar object is removed, the bird becomes
restless and emits shrill calls. When given a choice between the
familiar stimulus and a novel one, the bird preferentially approaches
the familiar stimulus. It is important to realize that filial imprinting
refers to the acquisition of a social preference and not just an
increase in following (Sluckin 1972; Bolhuis 1991).

Conditions for imprinting

To study visual imprinting in the laboratory, chicks or ducklings may
be hatched in darkness and exposed for a period of 1–2 hours to a
conspicuous object when they are about 24-h old. The animals are
then returned to a dark incubator and kept there until their
preferences are tested by exposing them to the familiar object and a
novel object. A widely used measure of filial preference is approach
to the familiar object relative to approach to a novel object. The
effectiveness of imprinting stimuli varies. For example, young
ducklings approach and follow objects larger than a matchbox, but
peck at smaller objects. For chicks, red and blue objects are more
effective imprinting stimuli than yellow and green objects.
Movement, brightness, contrast, and sound all enhance the
attractiveness of an imprinting stimulus.

Imprinting and learning

Lorenz (1935) thought that imprinting had “nothing to do with
learning.” Indeed, filial imprinting proceeds without any obvious
reinforcement (see Chapter 8) such as food or warmth (see Bolhuis
et al. 1990 for discussion). However, an imprinting object may itself
be a reinforcer, i.e., a stimulus that an animal finds rewarding. Just
as a rat can learn to press a lever to receive a food reward, so a
visually naive chick is able to learn to press a pedal to see an
imprinting object. As we will see in Chapter 8, this is a case of
instrumental or operant conditioning, and in the case of the chick the
mere sight of an imprinting object is reinforcing the response.
Similarly, when chicks are exposed to two imprinting stimuli (e.g., a
visual object and a sound) simultaneously, they learn more about the
individual stimuli than when they are exposed to the stimuli
sequentially or to only one stimulus. This so-called within-event
learning has also been found in conditioning paradigms in rats and
humans (Bolhuis & Honey 1998). So although imprinting looks quite
different from classical or operant conditioning, nevertheless it
shares many characteristics with these forms of associative learning.
Thus, it may be that only the characteristics and the circumstances in
which imprinting occurs differ from other forms of learning, but that
the underlying mechanisms are similar if not the same.

Sexual imprinting
Lorenz suggested that the main consequence of imprinting is the
determination of adult sexual preferences. Other research suggests
that filial imprinting and sexual imprinting are two separate
(although perhaps partially overlapping) processes. First, the time of
expression of the preference differs, with filial preferences being
expressed in very young birds, whereas sexual preferences are
expressed during courtship, when the animals are sexually mature.
Second, the period of time during which experience affects
preferences also differs between filial and sexual imprinting. Sexual
preferences continue to be affected by experience up to the time of
mating. Furthermore, filial preferences may be formed after a
relatively short period of exposure to an object. In contrast, sexual
preferences develop as the result of a long period of exposure to, and
social interaction with, the parents as well as the siblings.
Normally, sexual imprinting ensures that the bird will mate with a
member of its own strain or species. However, when the young bird
is cross-fostered (i.e., reared with adults of a different species) it may
develop a sexual preference for the foster species. For instance, when
young zebra finch (Taeniopygia guttata) males are reared with
Bengalese finch (Lonchura striata, also known as the society finch)
parents (Figure 7.3), they will later show courtship behavior mostly
to Bengalese finch females. Interestingly, when young zebra finch
males are reared with mixed parents (one Bengalese finch and one
zebra finch), they will later show a sexual preference for a female of
the species with which they had interacted most. More specifically, in
Japanese quail (Coturnix coturnix japonica) and domestic chickens,
mating preferences are for individual members of the opposite sex
that are different, but not too different, from individuals with which
the young bird was reared (Bateson 1978).

Figure 7.3 Zebra finch (left) and Bengalese finch (right).

(Photograph courtesy of Hans-Joachim Bischof).

Is imprinting really irreversible?

Lorenz’s assertion, often cited in animal behavior textbooks, that
imprinting is irreversible has strengthened the myth that the main
developmental processes cannot be reversed. It is important to
realize that there are two ways in which Lorenz’s claim of
irreversibility of imprinting can be interpreted. On the one hand, it
could mean that once the young bird has formed an attachment with
a particular object, it will never direct its social behavior toward a
different, novel object. A weaker form of the claim is that although
the animal may show social behavior toward novel objects, it will not
forget what it has learned about the object to which it was exposed
originally. The latter form of the claim of irreversibility is in fact
more in line with Lorenz’s (1937a) account, for instance when he says
that “the recognition response cannot be ‘forgotten’!”
The strong form of the claim of irreversibility has been refuted, both
for filial imprinting (Salzen & Meyer 1967; see Bolhuis 1991) and for
sexual imprinting (Kruijt & Meeuwissen 1991; see Bolhuis 1991). For
instance, Salzen and Meyer (1967) reared chicks with a colored ball
that was suspended in their home cage for 3 days. When given a
simultaneous preference test, the animals preferred the familiar
stimulus to a novel stimulus of a different color. When the familiar
stimulus was then removed and replaced with the alternative colored
ball for a further 3 days, the chicks had reversed their preference to a
preference for the novel stimulus. Other research has addressed the
weak version of the claim of irreversibility, i.e., that information
about the first stimulus is not forgotten. Specifically, it was found
that although filial preferences are reversible, under certain
circumstances the original preference may return (Bolhuis & Bateson
1990).
In sexual imprinting, as in filial imprinting, it has been found that
the strong claim of irreversibility cannot be maintained, whereas
there is evidence that supports the weak version of the claim (Bischof
1994). When zebra finch males were reared with Bengalese finch
foster parents for the first 40 days of life they showed a strong
courtship preference for Bengalese finch females in preference tests
when adult. When such males were exposed to a zebra finch female
for several months they would show courtship behavior toward that
female, but after separation from the zebra finch female the males
once again preferred a female of the foster species in preference
tests. However, if the males did not receive a preference test before
exposure to the zebra finch female, most of them altered the
direction of their courtship behavior toward a stable preference for
zebra finch females. Direct physical interaction with the females was
not necessary, either during the brief preference tests or during the
longer exposure periods when the animals were sexually mature.
Thus, sexual preferences could be altered by later exposure to a
female of the other species and, furthermore, brief exposure to a
female of the rearing species was sufficient to “consolidate” or
“stabilize” the animal’s sexual preference. When zebra finch males
reared by Bengalese finch parents were exposed to a Bengalese
female first they subsequently courted Bengalese finch females;
however, when they were exposed to a zebra finch female first, a
large proportion of the males changed their preference toward a
zebra finch female. On the basis of these and other results, Bischof
(1994) suggested that sexual imprinting is a two-stage process, with
an acquisition phase during which the animal learns the
characteristics of its parents and siblings, and a consolidation phase
during which the sexual preference is stabilized or modified
according to the species of the individual to which the male is
exposed.

Development of attachments in humans and other

primates

Theories and methods employed in animal behavior research have

played an important role in the formulation and development of
British psychiatrist John Bowlby’s (1969) theory of attachment in
humans. This theory was originally developed to explain the
behavior of children who had been separated from their mothers and
raised in a wartime nursery during the Second World War, and was
greatly influenced by Lorenz’s ideas about imprinting. In many ways,
the attachment system postulated by Bowlby is analogous to the
systems involved in filial behavior in young birds. In both cases, the
newborn infant or chick possesses a number of behavior patterns
that keep it in contact with the parent (or other caregiver) and which
attract the attention of the parent in the parent’s absence. Further,
both infant and chick must learn the characteristics of the parent,
which is important for the formation of a social bond between the
two. Factors influencing the formation of the bond are also similar,
including all the factors discussed above such as length of exposure,
sensitive periods, and predispositions. Studying the importance of
these factors in the human situation has resulted in a large body of
literature, some of which has supported the theory and some of
which has not (see Rutter 2002). The theory itself has been modified
to take these results into account and has also been expanded to
include development of attachments throughout life.
Ironically, the study of attachment behavior involved experiments
with monkeys that nowadays would be considered too cruel to be
allowed. To study the effects of maternal separation on infant
behavior, Harlow (1958), for example, raised infant rhesus monkeys
in complete social isolation, which had severe effects on the infant’s
subsequent behavior. Less intrusive methods, such as raising infants
with other infants (Harlow & Harlow 1962) or separating infants
from their mothers for brief periods of time (Hinde 1977), led to less
dramatic results, although these methods are still unacceptable for
human research. Harlow and his collaborators conducted an
extensive series of experiments that involved rearing young rhesus
monkeys with surrogate mothers. In many of these experiments the
infant monkey could choose between a cloth mother surrogate and a
wire mother surrogate, with only the latter providing food to the
infant. Whenever there was some stressful situation, the infant
invariably preferred to stay close to the cloth mother. Again, these
experiments appear to us nowadays as cruel. However, they must be
seen in the light of prevailing ideas at the time concerning child
rearing. These were strongly influenced by behaviorist ideas (see
Chapter 1) suggesting that reinforcement (e.g., food) was the main
factor in the development of children as well as animals. Harlow’s
work has been extremely important in showing rather poignantly
that, unlike behaviorist dogma, infants need the type of contact and
assurance provided by parents in order to develop as normal social
beings.
Bowlby felt that the best method for studying human development
was to observe infants in real-life situations, in much the same way
as many ethologists study the behavior of other animals in natural or
seminatural settings. Much of his theorizing about human
attachment was based upon such research carried out by his
colleague Mary Ainsworth. She and her colleagues (1978) developed
a standardized strange situation test in which a stranger approaches
an infant with and without the parent being present and various
aspects of the infant’s behavior are measured. This method is now
widely used and has allowed researchers to characterize specific
patterns of attachment and their determinants.
 Birdsong Learning

Approximately half of all bird species belong to the order

Passeriformes, the songbirds. A subgroup of this order known as the
oscines (real songbirds) need to learn their song from a “tutor,”
usually an adult male of their own species, often the bird’s father
(Catchpole & Slater 1995; Bolhuis et al. 2010; Bolhuis & Everaert
2013; Bolhuis & Moorman 2015). The developmental time at which
this learning takes place varies widely between species: in some
species learning occurs early in development, whereas in others
learning takes place when the birds have dispersed to an area other
than where they were reared, usually when they are in their first
breeding season (see Snowdon & Hausberger 1997 for review). In
addition, a distinction is made between “open-ended learners” (e.g.,
the canary Serinus canarius) that can learn new songs throughout
life and species (e.g., the zebra finch) where one song is learned early
in life and remains unaltered (Marler 1987). We will discuss only
these “age-limited learners” (Marler & Peters 1987; for reviews see
DeVoogd 1994; Marler 1976).
Pioneering work on birdsong was done by William Thorpe at the
University of Cambridge. The classic work of one of Thorpe’s
students, British-born ethologist Peter Marler, and his coworkers has
led to the distinction of two main phases in the song copying process:
a memorization phase and a sensorimotor phase (Figure 7.4; Bolhuis
& Moorman 2015). During the memorization phase (when the bird
does not yet sing itself) the young male is thought to form a
“template” or memory of the tutor song. Later, during the
sensorimotor phase, the young will start to vocalize, gradually
matching its own vocalizations with the information stored in the
template. Song learning in age-limited learners thus goes through
various stages, called subsong, plastic song, and crystallized or full
song, respectively (Figure 7.4). During subsong, the animal produces
highly variable and unstructured song patterns. During the period of
plastic song, the male produces song patterns that it has learned
earlier, but there is still a degree of plasticity in song output. This is
the phase when the animal is thought to “match” its own song output
with the stored template. Eventually, there is crystallization of vocal
output into full song, which is fixed for life (Marler 1976, 1987).
Figure 7.4 Stages in song development in some songbird species
and their presumed underlying mechanisms. Figures are sonograms
(frequency over time) of the song of a zebra finch (Taeniopygia
guttata) at different stages of development. See text for further
explanation. (After Bolhuis & Moorman 2015, adapted from; Slater
(1983); Bolhuis and Gahr (2006), with permission.).

Birdsong and the development of speech and language

in human infants
There are some interesting parallels between birdsong learning and
the development of speech and language in human infants (Doupe &
Kuhl 1999; Bolhuis et al. 2010; Berwick et al. 2011; Bolhuis &
Everaert 2013). First, as in songbirds, human infants go through a
phase in development where they do not yet vocalize themselves but
during which they hear others vocalize, the characteristics of which
they may learn. Later, the infant starts to vocalize itself; this cannot
yet be termed proper speech but is called “babbling.” The infant
produces sounds that resemble real words but which are not
intelligible as such. This is very much like subsong in songbirds, as
described above (Snowdon & Hausberger 1997; Doupe & Kuhl 1999;
Bolhuis et al. 2010). In both cases it looks like the young individual is
“practicing” its vocalizations, comparing them to information stored
in memory. Second, in both cases there is a sensitive period (see
below) early in life, during which the acquisition of new vocalizations
proceeds more easily than before or after.
In humans, speech is a possible externalization of language, which is
seen as a mind-internal cognitive system (Bolhuis & Everaert 2013;
Everaert et al. 2015). It is important to realize that there is a
fundamental distinction between language (a cognitive system) and
speech (a possible outward manifestation, with, e.g., signing being
another). Given the parallels between birdsong learning and infant
speech acquisition that we have discussed, the question arises as to
whether birdsong has a form of syntax or grammar in the way that
human language has (Bolhuis et al. 2010, 2014; Berwick et al. 2011;
Bolhuis & Everaert 2013). Many songbird species do have a certain
structure in the order of elements in their songs, and the sequence of
these elements is not random, but there is no evidence to suggest
that this “phonological syntax” (Marler 1977) is used in a
“generative,” creative way and conveys meaning in the way that
human language syntax does (Bolhuis et al. 2018). In human
language, meaning is associated with structure through the principle
of compositionality, whereby the meaning of a complex expression is
a function of the meaning of its constituent parts and the mode of
composition (Bolhuis et al. 2018). Crucially, human language has a
hierarchical structure, where the position of a word in the mind-
internal hierarchy is important, rather than its serial position in a
string of spoken words, hence Everaert et al. (2015)’s observation
that human language involves “structures, not strings.” Despite
claims to the contrary (see Chapter 14 for details), a recent review
found no evidence for compositionality in bird vocalizations (Bolhuis
et al. 2018). Thus, there are interesting similarities in the way in
which auditory-vocal learning proceeds in both birdsong and human
speech. However, these similarities do not extend to an underlying
cognitive mechanism that builds hierarchical structures, which
seems to be the exclusive property of human language. This has
important consequences for the study of the evolution of human
language, which should be distinguished clearly from the evolution
of auditory-vocal learning (Bolhuis & Everaert 2013; Bolhuis et al.
2014, 2018).

Predispositions and Sensitive Periods

As we have discussed, the difference between learning and other

forms of behavioral development is gradual and has to do with the
specificity of external experience. There are many examples of
perceptual preferences that may develop without any experience
with the particular stimuli involved. In this case, researchers often
use the term “predispositions” to denote the behavioral tendency or
indeed the underlying mechanism. Predispositions have been found
to play a role in birdsong learning (Marler 1991) and auditory
preferences in ducklings (Gottlieb 1980). Further, there is an
important influence of predispositions in the perception of faces in
neonatal human infants (Johnson & Morton 1991) and in the
development of filial preferences in chicks (Johnson et al. 1985;
Bolhuis 1996). In the study of the development of filial preferences in
chicks, the term “filial predispositions” has been used (Bolhuis
1996). These were defined as perceptual preferences that develop in
young animals without experience of the particular stimuli involved
(Bolhuis & Honey 1998).
A predisposition for species-specific sounds has been demonstrated
in song learning in certain avian species (Marler 1987, 1991). Under
certain circumstances, young males of some songbird species can
learn their songs, or at least part of their songs, from tape recordings
of tutor songs. When fledgling male song sparrows (Melospiza
melodia) and swamp sparrows (Melospiza georgiana) were exposed
to taped songs that consisted of equal numbers of songs of both
species, they preferentially learnt the songs of their own species.
Males of both species are able to sing the songs of the other species,
and it appears therefore that perceptual predispositions are involved
in what Marler (1991, p. 200) called the “sensitization of young
sparrows to conspecific song.”
In an extensive series of elegant experiments, the American biologist
Gilbert Gottlieb (1976, 1980) investigated the mechanisms
underlying the preferences that young ducklings of a number of
species show for the maternal call of their own species over that of
other species. This might suggest that these preferences are “innate,”
in the sense of “present at birth.” However, it turns out that reality is
much more complex. Gottlieb (1980, 2002a) found that differential
behavior toward the species-specific call could already be observed at
an early embryonic stage, before the animal started to vocalize itself.
However, a posthatching preference for the conspecific maternal call
was only found when the animals received exposure to embryonic
contact-contentment calls, played back at the right speed (Gottlieb
1980) and with a natural variation, within a certain period in
development. Thus, the expression of the species-specific
predisposition in ducklings is dependent on particular experience
earlier in development. In fact, Gottlieb found that these preferences
can be induced, maintained, and facilitated by external experience.
An illustration of these principles of perceptual development is
shown in Figure 7.5.
Figure 7.5 Three ways in which experience can influence the
development of perceptual preferences: (a) maintenance, (b)
facilitation, and (c) induction. The behavior of interest is shown on
the y-axis as a measure of preference, e.g., for the maternal call of
the bird’s own species in ducklings (Gottlieb 1980) or for a stimulus
with a head and neck in domestic chicks (Bolhuis 1996). The time of
onset of divergence of the two lines can be before hatching, as in the
case of auditory predisposition in ducklings (Gottlieb 1980), or
shortly after hatching, as in the case of visual predisposition in chicks
(Bolhuis 1996). Maintenance and facilitation were demonstrated in
some of Gottlieb’s own work on the development of auditory
preferences in ducklings (see Gottlieb 1980) and in other paradigms
(see text). Induction was demonstrated, for example, in the
development of filial predisposition in domestic chicks (Bolhuis
1996). Solid line, experience; dashed line, no experience. (After
Gottlieb 1980, with permission.).

Predispositions in chicks and human infants

The development of filial behavior in the chick involves two systems
that are neurally and behaviorally dissociable (Horn 1985, 1998,
2004; Johnson et al. 1985; Bolhuis 1996; Bolhuis & Honey 1998).
Behavioral evidence for the existence of a predisposition was
provided by a study in which day-old dark-reared chicks received
imprinting training by exposing them to either a rotating red box or a
rotating stuffed junglefowl hen (Johnson et al. 1985). Chicks in a
control group were exposed to white overhead light for the same
amount of time. The approach preferences of the chicks were
measured in a subsequent test where the two training stimuli were
presented simultaneously. Preferences were tested at either 2 hours
(Test 1) or 24 hours (Test 2) after the end of training (Figure 7.6). At
Test 1 the chicks preferred the object to which they had been exposed
previously. At Test 2 there was a significantly greater preference for
the fowl in both experimental groups as well as in the light-exposed
control group. Thus, the preference for the junglefowl increased from
the 2-h to the 24-h test and did so regardless of the stimulus with
which the chicks had been trained.
Figure 7.6 Mean preference scores (expressed as a preference for
the stuffed fowl) of chicks previously trained by exposure to a
rotating stuffed junglefowl (brown bars), a rotating red box (red
bars), or white light (white bars). Preference scores are defined as
activity when attempting to approach the stuffed junglefowl divided
by total approach activity during the test. Preferences were measured
in a simultaneous test either 2 hours (Test 1) or 24 hours (Test 2)
after the end of training. K1–K4, differences between the preferences
of the trained chicks and the controls; Δy, difference in preference
between the control chicks at Test 2 and at Test 1. See text for further
explanation. (Adapted from Horn 1985, by permission of the Oxford
University Press, after Johnson et al. 1985).
These results suggest that the preferences of trained chicks are
influenced by at least two different systems. On the one hand, there
is an effect of experience with particular stimuli (reflected in the
differences K1–K4 in Figure 7.6), i.e., filial imprinting. On the other
hand, there is an emerging predisposition to approach stimuli
resembling conspecifics (reflected as ΔY for the control group in
Figure 7.6). Training with a particular stimulus is not necessary for
the predisposition to emerge. In fact, visual experience is not
necessary to “trigger” or induce the predisposition (Gottlieb 1976;
Bolhuis 1996). The predisposition can emerge in dark-reared chicks,
provided that they receive a certain amount of nonspecific
stimulation within a certain period in development (Johnson et al.
1989).
Stimulus characteristics that are important for the filial
predisposition to be expressed were investigated in chicks that had
developed the predisposition. The tests involved an intact stuffed
junglefowl versus a series of increasingly degraded versions of a
stuffed junglefowl (Johnson & Horn 1988). The degraded versions
ranged from one where different parts of the model (wings, head,
torso, legs) were reassembled in an unnatural way to one in which
the pelt of a junglefowl had been cut into small pieces and stuck onto
a rotating box. The intact model was preferred only when the
degraded object possessed no distinguishable junglefowl features. In
addition, chicks did not prefer an intact junglefowl model over an
alternative object that contained only the head and neck of a stuffed
fowl. Thus, the head and neck region contains stimuli that are
relevant for the predisposition. In subsequent experiments it was
found that the chicks did not prefer a stuffed junglefowl hen over a
stuffed Gadwall duck (Anas strepera) or even a stuffed polecat
(Mustela putorius). Thus, the predisposition is not species or even
class specific. Subsequent studies showed that eyes are an important
stimulus but that other aspects of the stimulus are also sufficient for
the expression of a predisposition (Bolhuis 1996).
There are interesting similarities between the development of filial
preferences in chicks and the development of face recognition in
human infants (Johnson & Morton 1991). British psychologist Mark
Johnson and his colleagues tested visual preferences in human
babies as young as 30 minutes old. The newborn babies were shown
various slowly moving objects that resembled table tennis rackets.
On these rackets were painted either four dots in the configuration of
two eyes, a nose, and a mouth, or the dots were jumbled up; the
control was a racket with no dots. Newborn infants followed a
moving face-like stimulus with their eyes significantly more than the
other two stimuli (Johnson & Morton 1991). It is not known whether
the development of face preferences in infants is dependent on
previous experience, but the parallel with the emergence of
perceptual predispositions in birds is striking. Similarly, in both
human infants and young precocial birds, the features of individual
stimuli need to be learned.

Sensitive periods
Sensitive periods are an important characteristic of developing
behavior. At the same time, the concept of “sensitive period” is often
misunderstood. It needs to be separated from the idea of “critical
period,” which implies a much stricter separation of sensitivity and
insensitivity to external influences. The idea of a critical period
simply does not correspond with biological reality. However, there is
considerable evidence to suggest that during development there are
periods of increased sensitivity to external experience, preceded and
followed by periods of less sensitivity. The transition between these
developmental states may be gradual.
Bateson (1979) likened some of the different interpretations of
sensitive periods to a train traveling through a landscape. The train is
a metaphor for the developmental process. At the beginning of the
journey all the windows of the train, which are opaque, are closed so
that passengers cannot see the landscape outside the train. In the
simplest interpretation, the windows open at some stage and close at
a later stage. This is very much like the old idea of a critical period.
Another interpretation is that some windows open at some stage and
others open at other stages, i.e., different onset of sensitivity at
various stages. The different windows may or may not close again.
The most realistic interpretation, inspired by studies of filial
imprinting, is that windows may open at different stages and not
close again. The passengers can obtain information from outside the
train and may decide on the basis of that information that they need
to get off the train at the next station. In other words, the end of an
apparent sensitive period is not a result of an end to external
experience influencing the organism (i.e., windows being closed
again) but the result of the effect that external experience has on the
organism. In the case of imprinting, when the young bird has been
exposed to an imprinting stimulus and learned its characteristics, it
will form a social preference for this stimulus. As long as that
stimulus is present, the animal will follow it around and spend most
of its time close to it, so the animal will not have much opportunity to
imprint on other stimuli. As we saw earlier, if the original imprinting
stimulus is removed and replaced with a novel stimulus, the animal
can imprint on the novel stimulus, suggesting that sensitivity for
experience relevant for imprinting has not waned.
Various models have been put forward to explain sensitive periods,
ranging from endogenous factors (e.g., “clock models”) to self-
terminating mechanisms, where an apparent sensitive period comes
to an end solely as a result of external experience (Bateson 1979; Ten
Cate 1989). It has become clear that sensitive periods should not be
seen as rigid mechanisms where a window to the external
environment opens briefly, never to be opened again. Rather,
sensitive periods are regarded as flexible mechanisms, the timing of
which can be modified, and that depend for a large part on external
influences on the organism (Bolhuis 1999).

Development of Brain and Behavior

Many developmental processes (including behavioral processes) are

constrained by the developing nervous system. An illustrative
example is the development of face recognition in human infants, as
discussed above (Johnson & Morton 1991). For some time there were
conflicting reports about whether newborn infants showed a
preference for stimuli that look like faces over stimuli that do not.
Some researchers claimed that it was not until infants were 2–3
months old that they could recognize faces, whereas others reported
face recognition in 10-minute-old babies. Johnson and his
collaborators discovered that these conflicting findings have
something to do with the method of testing. When newborn infants
are tested by showing them static figures, they do not prefer to look
at faces compared with nonfaces. However, when the stimuli are
moved around, the infants follow faces with their eyes and head
significantly more than nonfaces. How can we explain this?
Johnson and his colleagues suggested that it has something to do
with the differential development of the human neocortex and
subcortical structures. Certain subcortical visual structures such as
the superior colliculus are fully developed at birth, whereas the
development of the cortex continues for a considerable time after
birth. The subcortical structures, which are sufficient for face
recognition, receive projections particularly from the peripheral part
of the retina. When the infant has to follow a moving stimulus with
its eyes, there is a greater chance of light reflected by this stimulus
falling on the periphery of the retina, and thus of the subcortical
structures being activated. Thus, as the subcortical structures are
fully developed at birth, and the visual cortex is not, only when
stimuli are moving are the fully developed neural structures involved
and faces recognized.
Experience during development can also affect the structure of the
nervous system. Rats reared in “enriched” environments (with a
running wheel and toys, and sometimes with other rats) have a
different cortex from control rats reared in standard laboratory cages
(Greenough et al. 1987). In particular, the overall weight of the cortex
is greater in the “enriched” rats and they have more elaborate
dendritic branching (see Chapter 5). Arguably more interesting are
direct relations between cognitive experience and brain structure, as
in the case of learning. Filial imprinting has become a prominent
model for the study of the neural mechanisms of learning and
memory (Horn 1985, 1998, 2004; McCabe 2019). A big advantage of
imprinting is that domestic chicks do not actually eat food until they
are about 3 days old. Thus, they can be kept in the dark until the start
of an imprinting experiment (just like in the natural situation where
they spend most of their first days in darkness, under the mother
hen). Exposure to an imprinting stimulus will then be their first
visual experience and, as we have seen, relatively limited experience
can have a big impact on the animal’s behavior. Cambridge
neurobiologist Gabriel Horn and his collaborators have studied the
neural mechanisms of filial imprinting in great detail (Horn 1985,
2004; McCabe 2019). They have found that there are a number of
plastic changes in a restricted region of the chick forebrain, the
intermediate and medial mesopallium (IMM) (Figure 7.7) during
imprinting. For instance, there is increased protein synthesis in the
IMM and also increased expression of immediate early genes (see
Chapter 5), suggesting increased neuronal activation. The level of
this neuronal activation correlates with the strength of imprinting, as
measured in preference tests. In other words, the more the animals
have learned, the more neuronal activity in the IMM and not in other
parts of the brain. Thus, there appears to be localization of function,
as only a very restricted part of the brain seems to be involved in
memory storage. Analysis of the brain at the subcellular level, using
an electron microscope, revealed that imprinting led to a significant
increase in synaptic contact area between neurons. This suggests that
learning may lead to increased synaptic efficiency, which is
consistent with proposals put forward by the Canadian psychologist
Donald Hebb as long ago as 1949 (see Chapter 5).
Figure 7.7 Schematic drawings of the brain of the domestic chick.
Top: Lateral aspect of the brain. The line indicates the approximate
plane of the coronal section of the brain below. Scale bar, 5 mm.
Bottom: Simplified diagram of a coronal section of the chick brain at
the level of the intermediate and medial mesopallium (IMM). The
extent in the coronal plane of the IMM, as removed in biochemical
studies (see Horn 1985, 2004), is indicated by the red area.
Abbreviations: M, mesopallium; VL, lateral ventricle; LaM,
mesopallial lamina. (Adapted from Moorman & Nicol 2015, with
permission, courtesy of Sanne Moorman).
In Chapter 5, the neural correlates of birdsong are discussed,
particularly those in the “song system” that are involved in vocal
production (see also DeVoogd 1994; Nottebohm 2000; Bolhuis &
Gahr 2006; Bolhuis et al. 2010; Bolhuis & Moorman 2015; Prather et
al. 2017). Mello and Clayton (1994) found that in zebra finches,
hearing a song led to neuronal activation (measured as expression of
immediate early genes) in brain regions outside the conventional
song system, in particular, the caudomedial mesopallium (CMM)
and caudomedial nidopallium (NCM) (see Figure 5.6). Interestingly,
the level of neuronal activation in one of these regions (NCM) is
correlated with the strength of song learning, measured as the
number of song elements that a bird has copied from its tutor
(Bolhuis et al. 2000; for reviews see Bolhuis & Gahr 2006; Bolhuis et
al. 2010; Bolhuis & Moorman 2015). Thus, analogous to the
imprinting case, there appears to be learning-related neuronal
activation in a restricted brain region. This strongly suggests that the
NCM is (part of) the neural substrate for tutor song memory, that is
distinct from the neural substrate of the memory of the bird’s own
song, which, as we saw in Chapter 5, is located in the song system
(Bolhuis et al. 2012; Bolhuis & Moorman 2015).
In addition to the behavioral parallels that we have discussed, there
are also similarities between birdsong learning and human speech in
terms of the underlying neural organisation (Bolhuis et al. 2010;
Bolhuis & Moorman 2015). The neural “division of labor” in
songbirds between production-related regions (the song system) and
memory- and perception-related regions (including the NCM) is
reminiscent of the functional distinction between Broca’s area and
Wernicke’s area in the human brain (Bolhuis et al. 2010; Bolhuis &
Moorman 2015; Prather et al. 2017; Friederici et al. 2017). In
addition, memory-related neuronal activation is lateralized to the left
hemisphere in both human infants and songbirds (Moorman et al.
2012; Moorman & Nicol 2015).
It is interesting that both imprinting and song learning, two
prominent models for the neural mechanisms of learning and
memory, are occurring early in development. It may be that such
developmental events are so prominent in the life of an individual
that it is easier to detect neural correlates of learning in such
behaviors.
 Development of Motivational Systems

Kruijt (1964) was one of the first to describe the development of

behavior systems in detail. He provided a detailed description of the
development of behavior in the Burmese junglefowl (Gallus gallus
spadiceus), the ancestor of the domestic chicken. He suggested that,
initially, many of the motor components of behavior appear as
independent units prior to any opportunity for practice, and that
only later, often after specific experience, do these motor
components become integrated into more complex systems such as
hunger, aggression, or sex. An example is aggressive behavior in
young male junglefowl chicks: behavioral components such as
hopping, pecking, and kicking occur quite independently at first, but
later become integrated into behavioral sequences seen during
agonistic encounters with other males.
Hogan (1988) has generalized these proposals and suggested a
general framework for the analysis of behavioral development using
the concept of behavior systems, as described in Chapter 3. A
behavior system consists of various elements: a central mechanism,
perceptual mechanisms, and motor mechanisms. These mechanisms
correspond to structures in the brain, and one could also call them
cognitive structures. The structural definition of a behavior system is
“any organization of perceptual, central and motor mechanisms that
acts as a unit in some situations” (see Chapter 3 and Figure 3.1).
According to Hogan (1988), behavioral development is essentially
the development of these mechanisms and the changes in the
connections among them. As we shall see, many of these
mechanisms and their connections only develop after functional
experience, i.e., experience with the particular stimuli involved, or
with the consequences of performing specific motor patterns.

Hunger
An example of a developing behavior system is the hunger system in
the junglefowl chick (Hogan 1988). This system involves perceptual
mechanisms for the recognition of features (color, shape, etc.),
objects (grain, worms, etc.), and functions (food versus nonfood).
Then there are motor mechanisms underlying behavior patterns
such as ground scratching and pecking, and there is a central hunger
mechanism. Importantly, several of these mechanisms and the
connections between them (dashed lines in Figure 3.5) develop as a
result of specific functional experience. For instance, only after a
substantial meal will the chick differentiate between food items and
nonfood items to eat.
On the motor side of the system, the mechanisms underlying
pecking, scratching, and walking are present as soon as the chick
hatches, as is the effective coordination of these behaviors into
foraging sequences. On the other hand, for at least 3 days, feeding-
related behavior (in this case pecking) is not dependent on the level
of food deprivation. Thus, the central hunger mechanism and the
pecking motor mechanism are not initially connected. Only after the
experience of pecking and swallowing (and not necessarily of food:
pecking and swallowing sand is equally effective) do the two
mechanisms become connected, and only then is the level of pecking
dependent on the level of food deprivation (Hogan 1984). A similar
phenomenon occurs in the case of suckling in rat pups, as described
earlier (Hall & Williams 1983). Suckling decreases as weaning
approaches but, importantly, suckling behavior does not become
deprivation dependent until about 2 weeks after birth.
Unfortunately, we still do not know whether functional experience is
required, and if so what type of experience is needed, to connect the
suckling motor mechanism with the central hunger mechanism in
the rat pup.

Dustbathing
The development of behavioral structure is not uniform but may
proceed along different pathways for different behavior systems. An
example of this is the development of dustbathing in junglefowl
chicks, as studied by the Danish ethologist Klaus Vestergaard and
coworkers (Vestergaard et al. 1990, 1993; see also Chapter 3).
Dustbathing is a behavior that adult birds of many species frequently
engage in. It consists of a sequence of coordinated movements of the
wings, feet, head, and body that serve to spread dust through the
feathers (see Figure 3.6). The function of this behavior is to remove
excess lipids from the feathers and to maintain good feather
condition. Unlike the development of feeding behavior in rats or
chicks, dustbathing is deprivation dependent as soon as it appears in
the animal’s behavioral repertoire (Hogan 2001). Thus, in this case
chicks do not require functional experience to connect the motor
mechanisms with the central dustbathing mechanism.
On the perceptual side, other experiments have shown that initially
the chick will perform dustbathing on virtually any kind of surface,
including wire mesh, suggesting that the perceptual mechanism and
the central mechanism are not yet connected. The perceptual
mechanism itself develops more quickly with some substrates than
with others, which is similar to the development of perceptual
mechanisms in song learning and filial predispositions discussed
earlier. Furthermore, it turns out that preferences for functionally
unlikely surfaces (e.g., a skin of junglefowl feathers) can be acquired
as a result of experience with them. This is another example of the
development of a perceptual mechanism, and one that is not
dissimilar to filial imprinting.
 SUMMARY AND CONCLUSIONS
According to Niko Tinbergen, development is one of the four
main questions that should be asked about the behavior of
animals. The study of the development of animal behavior has
often been hampered by misrepresentation, mainly in the
popular literature, of early theoretical proposals. In particular,
interpretations of the concept of “instinct” have led to a stubborn
belief in the existence of “innate” behaviors, and the mistaken
idea that genes “determine” behavior. Also, some early intuitions
about imprinting have led to a rather rigid view on behavioral
development, where events occurring during “critical periods”
early in life are crucial for the development of behavior, an idea
that could be expressed metaphorically as once one has missed
the developmental bus there is no way back. It turns out that
there is considerable plasticity in development, extending into
adulthood. Nevertheless, events during development have a great
influence on adult behavior, for instance in imprinting and the
development of birdsong. In social behavior, even brief
separation from the mother can have profound effects on the
development of attachment. At the same time, there is
considerable plasticity and flexibility in the way in which these
events affect the development of animal behavior.

FURTHER READING
A number of classic papers on behavioral development,
supplemented with some important recent ones, have been collected
in the book by Bolhuis and Hogan (1999), which has crucial
publications on all aspects of development. Volume 2 of the 4-part
reader by Bolhuis and Giraldeau (2010) is devoted to behavioral
development. A collection of contemporary essays on the
development of behavior can be found in Hogan and Bolhuis (1994).
Gottlieb (2002b) provides a monograph on the relation between
development and evolution. Classic and contemporary papers on the
development of brain and cognition have been collected in a
comprehensive reader (Johnson et al. 2008), while a concise
introduction to the field of developmental cognitive neuroscience is
provided by Johnson and De Haan (2015). Review papers on the
development of brain and behavior include Bolhuis (1999), Hogan
(2001), Prather et al. (2017) and Yang et al. (2017).

ACKNOWLEDGMENTS
I am grateful to Luc-Alain Giraldeau, Jerry Hogan and the late
Gabriel Horn, for their comments on earlier versions of this chapter.

REFERENCES
Ainsworth, M.D.S., Blehar, M.C., Waters, E. & Wall, S. 1978.
Patterns of Attachment: Assessed in the Strange Situation and at
Home. Hillsdale, NJ: Erlbaum.
Bateson, P. 1979. How do sensitive periods arise and what are they
for? Animal Behaviour, 27, 470–86.
Bateson, P. 1978. Imprinting as a process of competitive exclusion.
In: J.P. Rauschecker & P. Marier (eds.), Imprinting and Cortical
Plasticity: Comparative Aspects of Sensitive Periods, pp. 151–68.
New York: John Wiley & Sons.
Bateson, P. 1999. Foreword. In: J.J. Bolhuis & J.A. Hogan (eds.), The
Development of Animal Behavior: A Reader, Oxford: Blackwell
Publishers.
Bateson, P.P.G. 1966. The characteristics and context of imprinting.
Biological Reviews, 41, 177–220.
Bateson, P.P.G. 1978. Sexual imprinting and optimal outbreeding.
Nature, 273, 659–60.
Benus, R.F., Den Daas, S., Koolhaas, J.M. & Van Oortmerssen, G.A.
1990. Routine formation and flexibility in social and nonsocial
behavior of aggressive and nonaggressive male mice. Behaviour,
112, 176–93.
Berwick, R.C., Okanoya, K., Beckers, G.J.L. & Bolhuis, J.J. 2011.
Songs to syntax: The linguistics of birdsong. Trends in Cognitive
Sciences, 15, 113–21.
Bischof, H.-J. 1994. Sexual imprinting as a two-stage process. In:
J.A. Hogan & J.J. Bolhuis (eds.), Causal Mechanisms of
Behavioural Development, pp. 82–97. Cambridge: Cambridge
University Press.
Bolhuis, J.J. 1991. Mechanisms of avian imprinting: a review.
Biological Reviews, 66, 303–45.
Bolhuis, J.J. 1996. Development of perceptual mechanisms in birds:
predispositions and imprinting. In: C.F. Moss & S.J. Shettleworth
(eds.), Neuroethological Studies of Cognitive and Perceptual
Processes, pp. 158–84. Boulder, CO: Westview Press.
Bolhuis, J.J. 1999. The development of animal behavior. From
Lorenz to neural nets. Naturwissenschaften, 86, 101–1.
Bolhuis, J.J. & Bateson, P.P.G. 1990. The importance of being first: a
primacy effect in filial imprinting. Animal Behaviour, 40, 472–83.
Bolhuis, J.J., Beckers, G.J.L., Huybregts, M.A.C., Berwick, R.C. &
Everaert, M.B.H. 2018. Meaningful syntactic structure in songbird
vocalizations? PloS Biology, 16(6), e2005157.
Bolhuis, J.J., De Vos, G.J. & Kruijt, J.P. 1990. Filial imprinting and
associative learning. Quarterly Journal of Experimental
Psychology, 42B, 313–29.
Bolhuis, J.J. & Everaert, M. (eds.). 2013. Birdsong, Speech &
Language. Exploring the Evolution of Mind and Brain.
Cambridge, MA: MIT Press.
Bolhuis, J.J. & Gahr, M. 2006. Neural mechanisms of birdsong
memory. Nature Reviews Neuroscience, 7, 347–57.
Bolhuis, J.J. & Giraldeau, L.-A. (eds.). 2010. Animal Behaviour, vol.
2. Development of Animal Behaviour. London: Sage Publications.
Bolhuis, J.J., Gobes, S.M.H., Terpstra, N.J., Den Boer-Visser, A.M. &
Zandbergen, M.A. 2012. Learning-related neuronal activation in
the zebra finch song system nucleus HVC in response to the bird’s
own song. PLoS One, 7(7), e41556.
Bolhuis, J.J. & Hogan, J.A. (eds.). 1999. The Development of Animal
Behavior: a Reader. Oxford: Blackwell Publishers.
Bolhuis, J.J. & Honey, R.C. 1998. Imprinting, learning and
development: from behaviour to brain and back. Trends in
Neurosciences, 21, 306–11.
Bolhuis, J.J. & Moorman, S. 2015. Birdsong memory and the brain:
In search of the template. Neuroscience & Biobehavioral Reviews,
50, 41–55.
Bolhuis, J.J., Okanoya, K. & Scharff, C. 2010. Twitter evolution:
Converging mechanisms in birdsong and human speech. Nature
Reviews Neuroscience, 11, 747–59.
Bolhuis, J.J., Tattersall, I., Chomsky, N. & Berwick, R.C. 2014. How
could language have evolved? PLoS Biology, 12(8), e1001934.
Bolhuis, J.J., Zijlstra, G.G.O., Den Boer-Visser, A.M. & Van Der Zee,
E.A. 2000. Localized neuronal activation in the zebra finch brain
is related to the strength of song learning. Proceedings of the
National Academy of Sciences USA, 97, 2282–5.
Bowlby, J. 1969. Attachment and Loss, vol. 1. Attachment. London:
Hogarth Press.
Catchpole, C.K. & Slater, P.J.B. 1995. Bird Song: Biological Themes
and Variations. Cambridge: Cambridge University Press.
DeVoogd, T.J. 1994. The neural basis for the acquisition and
production of bird song. In: J.A. Hogan & J.J. Bolhuis (eds.),
Causal Mechanisms of Behavioural Development, pp. 49–81.
Cambridge: Cambridge University Press.
Doupe, A.J. & Kuhl, P.K. 1999. Birdsong and human speech:
common themes and mechanisms. Annual Review of
Neuroscience, 22, 567–631.
Everaert, M.B.H., Huybregts, M.A.C., Chomsky, N., Berwick, R.C. &
Bolhuis, J.J.2015. Structures, not strings: Linguistics as part of
the cognitive sciences. Trends in Cognitive Sciences, 19, 729–43.
Friederici, A.D., Chomsky, N., Berwick, R.C., Moro, A. & Bolhuis, J.J.
2017. Language, mind and brain. Nature Human Behaviour, 1,
713–22.
Gottlieb, G. 1976. The roles of experience in the development of
behavior and the nervous system. In: G. Gottlieb (ed.), Neural
and Behavioral Specificity: Studies in the Development of
Behavior and the Nervous System, pp. 237–80. New York:
Academic Press.
Gottlieb, G. 1980. Development of species identification in
ducklings: VI. Specific embryonic experience required to maintain
species-typical perception in Peking ducklings. Journal of
Comparative and Physiological Psychology, 94, 579–87.
Gottlieb, G. 2002a. On the epigenetic evolution of species specific
perception: the developmental manifold concept. Cognitive
Development, 17, 1287–1300.
Gottlieb, G. 2002b. Individual Development and Evolution: The
Genesis of Novel Behavior. Mahwah, NJ: Erlbaum.
Greenough, W.T., Black, J.E. & Wallace, C.S. 1987. Experience and
brain development. Child Development, 58, 539–59.
Hall, W.G. 1979. Feeding and behavioral activation in infant rats.
Science, 190, 1313–5.
Hall, W.G. & Williams, C.L. 1983. Suckling isn’t feeding, or is it? a
search for developmental continuities. Advances in the Study of
Behavior, 13, 219–54.
Harlow, H.F. 1958. The nature of love. American Psychologist, 13,
573–685.
Harlow, H.F. & Harlow, M.K. 1962. Social deprivation in monkeys.
Scientific American, 207, 136–46.
Hebb, D.O. 1949. The Organization of Behavior: &
Neuropsychological Theory. New York: John Wiley & Sons.
Heinroth, O. 1911. Beiträge zur Biologie, nahmentlich Ethologie und
Psychologie de Anatiden. Verhandlungen des 5. Internationaler
Ornithologischer Kongress Berlin, 589–702.
Hinde, R.A. 1977. Mother-infant separation and the nature of inter-
individual relationships: experiments with rhesus monkeys.
Proceedings of the Royal Society of London Series B, 196, 29–50.
Hinde, R.A. & Bateson, P. 1984. Discontinuities versus continuities
in behavioural development and the neglect of process.
International Journal of Behavioral Development, 7, 129–43.
Hogan, J.A. 1984. Pecking and feeding in chicks. Learning and
Motivation, 15, 360–76.
Hogan, J.A. 1988. Cause and function in the development of
behavior systems. In: E.M. Blass (ed.), Handbook of Behavioral
Neurobiology, vol. 9, pp. 63–106. New York: Plenum Press.
Hogan, J.A. 1994. The concept of cause in the study of behaviour. In:
J.A. Hogan & J.J. Bolhuis (eds.), Causal Mechanisms of
Behavioural Development, pp. 3–15. Cambridge: Cambridge
University Press.
Hogan, J.A. 2001. Development of behavior systems. In: E.M. Blass
(ed.), Handbook of Behavioral Neurobiology, vol. 13, pp. 229–79.
New York: Kluwer Academic/Plenum.
Hogan, J.A. & Bolhuis, J.J. (eds.). 1994. Causal Mechanisms of
Behavioural Development. Cambridge: Cambridge University
Press.
Horn, G. 1985. Memory, Imprinting, and the Brain. Oxford:
Clarendon Press.
Horn, G. 1998. Visual imprinting and the neural mechanisms of
recognition memory. Trends in Neurosciences, 21, 300–5.
Horn, G. 2004. Pathways of the past: the imprint of memory. Nature
Reviews Neuroscience, 5, 108–120.
Johnson, M.H., Bolhuis, J.J. & Horn, G. 1985. Interaction between
acquired preferences and developing predispositions during
imprinting. Animal Behaviour, 33, 1000–6.
Johnson, M.H., Davies, D.C. & Horn, G. 1989. A sensitive period for
the development of a predisposition in dark-reared chicks. Animal
Behaviour, 37, 1044–6.
Johnson, M.H. & De Haan, M. 2015. Developmental Cognitive
Neuroscience: An Introduction, 4th ed. Chichester: Wiley-
Blackwell.
Johnson, M.H. & Horn, G. 1988. Development of filial preferences in
dark-reared chicks. Animal Behaviour, 36, 675–83.
Johnson, M.H. & Morton, J. 1991. Biology and Cognitive
Development: The Case of Face Recognition. Oxford: Wiley-
Blackwell.
Johnson, M.H., Munakata, Y. & Gilmore, R.O. 2008. Brain
Development and Cognition: a Reader, 2nd ed. Oxford: Wiley-
Blackwell.
Kruijt, J.P. 1964. Ontogeny of social behaviour in Burmese red
junglefowl (Gallus gallus spadiceus). Behaviour, (Suppl. 9), 1–
201.
Kruijt, J.P. & Meeuwissen, G.B. 1991. Sexual preferences of male
zebra finches: effects of early and adult experience. Animal
Behaviour, 42, 91–102.
Lagerspetz, K. 1964. Studies on the aggressive behaviour of mice.
Annales Academiae Scientiarum Fennicae, Series B, 131, 1–131.
Lehrman, D.S. 1953. A critique of Konrad Lorenz’s theory of
instinctive behavior. Quarterly Review of Biology, 28, 337–63.
Lehrman, D.S. 1970. Semantic and conceptual issues in the nature-
nurture problem. In: L.R. Aronson, E. Tobach, D.S. Lehrman &
 J.S. Rosenblatt (eds.), Development and Evolution of Behavior,
pp. 17–52. San Francisco: Freeman.
Lorenz, K. 1935. Der Kumpan in der Umwelt des Vogels. Journal fur
Ornithologie, 83, 137–213, 289–413.
Lorenz, K. 1937a. The companion in the bird’s world. Auk, 54, 245–
73.
Lorenz, K. 1937b. Uber die Bildung des Instinkbegriffes.
Naturwissenschaften, 25, 289–300, 307–18, 324–31.
Lorenz, K. 1965. Evolution and Modification of Behavior. Chicago:
University of Chicago Press.
Marler, P. 1976. Sensoiy templates in species-specific behavior. In: J.
Fentress (ed.), Simpler Networks and Behavior, pp. 314–29.
Sunderland, MA: Sinauer.
Marler, P. 1977. The structure of animal communication sounds. In:
T.H. Bullock (ed.), Recognition of Complex Acoustic Signals:
Report of the Dahlem Workshop on Recognition of Complex
Acoustic Signals, pp. 17–35. Berlin: Abakon-Verlagsgesellschaft.
Marler, P. 1987. Sensitive periods and the roles of specific and
general sensory stimulation in birdsong learning. In: J.P.
Rauschecker & P. Marler (eds.), Imprinting and Cortical
Plasticity: Comparative Aspects of Sensitive Periods, pp. 99–135.
New York: John Wiley.
Marler, P. 1991. Song-learning behavior: the interface with
neuroethology. Trends in Neurosciences, 14, 199–206.
Marler, P. & Peters, S. 1987. A sensitive period for song acquisition in
the song sparrow, Melospiza melodia: a case of age-limited
learning. Ethology, 76, 89–100.
McCabe, B.J. 2019. Visual imprinting in birds: Behavior, models, and
neural mechanisms. Frontiers in Physiology, 10, 658.
Mello, C.V. & Clayton, D.F. 1994. Song-induced ZENK gene
expression in auditory pathways of songbird brain and its relation
 to the song control system. The Journal of Neuroscience, 14,
6652–66.
Moorman, S., Gobes, S.M.H., Kuijpers, M., Kerkhofs, A.,
Zandbergen, M.A. & Bolhuis, J.J. 2012. Human-like brain
hemispheric dominance in birdsong learning. Proceedings of the
National Academy of Sciences USA, 109, 12782–7.
Moorman, S. & Nicol, A.U. 2015. Memory-related brain
lateralisationin birds and humans. Neuroscience & Biobehavioral
Reviews, 50, 86–102.
Nottebohm, F. 2000. The anatomy and timing of vocal learning in
birds. In: M.D. Hauser & M. Konishi (eds.), The Design of Animal
Communication, pp. 63–110. Cambridge, MA: MIT Press.
Oppenheim, R.W. 1981. Ontogenetic adaptations and retrogressive
processes in the development of the nervous system and
behaviour: a neuroembryological perspective. In: K.J. Connolly &
H.F.R. Prechtl (eds.), Maturation and Development: Biological
and Psychological Perspective, pp. 73–109. Philadelphia:
Lippincott.
Prather, J., Okanoya, K. & Bolhuis, J.J. 2017. Brains for birds and
babies: Neural parallels between birdsong learning and speech
acquisition. Neuroscience & Biobehavioral Reviews, 81, 225–37.
Rutter, M. 2002. Nature, nurture, and development: from
evangelism through science toward policy and practice. Child
Development, 73, 1–21.
Salzen, E.A. & Meyer, C.C. 1967. Imprinting: reversal of a preference
established during the critical period. Nature, 215, 785–6.
Slater, P.J.B. 1983. The development of individual behaviour. In:
T.R. Halliday & P.J.B. Slater (eds.), Genes, Development and
Learning. Animal Behaviour, vol. 3. New York: W.H. Freeman.
Sluckin, W. 1972. Imprinting and Early Learning. London:
Methuen.
Snowdon, C.T. & Hausberger, M. (eds.). 1997. Social Influences on
Vocal Development. Cambridge: Cambridge University Press.
Spalding, D.A. 1873. Instinct, with original observations on young
animals. Macmillan’s Magazine, 27, 282–293. Reprinted in 1954
in British Journal of Animal Behaviour, 2, 2–11.
Ten Cate, C. 1989. Behavioral development: toward understanding
processes. In: P.P.G. Bateson & P.H. Klopfer (eds.), Perspectives
in Ethology, vol. 8, pp. 243–69. New York: Plenum Press.
Tinbergen, N. 1951. The Study of Instinct. Oxford: Clarendon Press.
Tinbergen, N. 1963. On aims and methods of ethology. Zeitschria für
Tierpsychologie, 20, 410–33.
Vestergaard, K.S., Hogan, J.A. & Kruijt, J.P. 1990. The development
of a behavior system: dustbathing in the Burmese red junglefowl:
I. The influence of the rearing environment on the organization of
dustbathing. Behaviour, 112, 99–116.
Vestergaard, K.S., Kruijt, J.P. & Hogan, J.A. 1993. Feather pecking
and chronic fear in groups of red junglefowl: its relations to
dustbathing, rearing environment and social status. Animal
Behaviour, 45, 1117–26.
Waddington, C.H. 1966. Principles of Development and
Differentiation. New York: Macmillan.
Yang, C., Crain, S., Berwick, R.C., Chomsky, N. & Bolhuis, J.J. 2017.
The growth of language: Universal Grammar, experience, and
principles of computation. Neuroscience & Biobehavioral
Reviews, 81, 103–19.
8
learning and memory
KIMBERLY KIRKPATRICK AND GEOFFREY HALL

INTRODUCTION
When first faced with a particular set of circumstances, an animal
will behave in a certain way; but when these same circumstances
occur again, its behavior may be different. The animal, having
interacted with its environment, is changed by the experience and
thus becomes capable of behaving differently in the future. The
process of interaction that produces the change in the animal is
called learning, and the mechanisms involved in this process
form the subject matter of the first two main sections of this
chapter; the change itself is often referred to as the formation of a
memory, and the properties of animal memory will be
considered in the third and fourth sections of the chapter. Most of
this chapter will be taken up with discussing not the functional
implications of learning, but the mechanisms by which it is
achieved. Although field studies have supplied some important
information, our knowledge of these mechanisms comes mainly
from experimental work conducted with laboratory animals;
discussion of such experiments forms the bulk of the chapter.
Those who have done this work (often experimental psychologists
rather than ethologists) claim to have detected learning
mechanisms of general relevance—mechanisms that form the
basis of a whole range of seemingly different types of learning,
and that operate in most or all species. Others have doubted this
claim, suggesting that the narrow focus of laboratory studies of
learning has led researchers to overlook specialized mechanisms
of learning and memory that have evolved. This issue will be
discussed in the final section of the chapter.
 Procedures for the Study of Learning and
their Results

Exposure to a single event

Given the definition of learning offered above, an obvious way to
study the phenomenon is to present the animal with a well-defined
stimulus, observe the behavior that results, and then determine how
behavior changes with repeated presentations of the stimulus. The
usual outcome is that the likelihood or vigor of the response declines.
For example, Humphrey (1933) demonstrated that if a sharp tap is
applied to the substrate on which a snail (Helix) is moving, the
animal will respond by retracting it horns. The same may happen the
second time this stimulus is applied, but after a series of applications
of the stimulus, the animal will continue on its way with horns
extended. This phenomenon, the waning of a response with repeated
presentation of its eliciting stimulus, is known as habituation.
Examples are found throughout the animal kingdom.
Consideration of habituation allows us to refine our definition of
learning. Although all learning involves a change in behavior as a
result of experience, it is possible that such a change may sometimes
occur because of the operation of processes that do not deserve this
label—the change in the snail’s behavior might be a consequence
simply of fatigue in the muscular system responsible for the
response, or of a loss of sensitivity in the sensory system that detects
the stimulus. In other cases a change in responsiveness may be
attributed to motivational variables, as when an animal initially
responds with vigor when offered food (by eating it) but responds
less eagerly as its hunger is satisfied. Learning, it is usually supposed,
involves longer-term changes; to be sure that learning has occurred
it is necessary to demonstrate that the change observed is not a
consequence of these other, short-term, processes. For the case of
habituation relevant evidence comes from an effect called
dishabituation. If a habituated snail is subjected to a novel and
intense stimulus (Humphrey dropped a weight onto the board on
which the snail was moving) it is found that the response to the
original stimulus is partially restored, at least for a short time. This
reappearance of the response shows that its previous non-occurrence
cannot have been a consequence just of muscular fatigue or its
sensory equivalent.
In some cases, exposure to a stimulus produces not merely a waning
of the original response but establishes a new one. A classic example
is imprinting. The initial response of a newly hatched domestic chick
to a salient object (e.g., a brightly colored, moving box) may be to
show signs of fear, but this will soon be replaced by a characteristic
set of filial responses that include approaching the object and
following it when it moves. To show the behavior it does, the chick
needs to learn the characteristics of the object and discriminate these
from those possessed by other similar objects. Examples of this
perceptual learning effect (an enhanced ability to discriminate after
mere exposure to stimuli) may be found in other species and are not
confined to immature animals. For example, Hall (1979) found that
adult laboratory rats, exposed in their home cages to simple
geometrical shapes, showed an enhanced ability, subsequently, to
learn a discrimination task in which they had to choose one shape
over the other to obtain a food reward.

Pairing of events
Classical conditioning
Although the phenomenon of imprinting had been demonstrated
some years earlier, the experimental study of animal learning began
in earnest only about 1900 when the Russian physiologist Ivan
Pavlov turned his attention to the topic. Pavlov’s experimental
procedure involved explicit pairing of stimuli. From his earlier work
on the processes controlling digestive secretions, Pavlov knew that a
dog would salivate in response to a wide range of stimuli—not just to
the presentation of food but, for instance, to the appearance of the
laboratory attendant who supplied the food. This latter response
clearly depended on experience, and by taking a version of it into the
laboratory, Pavlov hoped that he would be able to use it to elucidate
the brain mechanisms responsible for learning.
In his standard experimental situation (Pavlov 1927) a lightly
restrained dog, isolated in a quiet room, received a series of training
trials in which presentations of food were paired with (usually
slightly preceded by) presentations of a neutral event, such as the
flashing of a light or the clicking of a metronome. From the outset,
the presentation of food evoked salivation—this response did not
require special training and was therefore described as an
unconditional (or unconditioned) response. The event that elicited it
was called an unconditioned stimulus (US). Over the course of
training, the light, which had originally been ineffective in this
respect, acquired the power to evoke salivation; this response, since
it was conditional on the animal having received training, was called
a conditioned response and the light itself a conditioned stimulus
(CS). The whole procedure thus became known as conditioning, and
since this term has also been applied to other examples of learning,
Pavlov’s original version is distinguished by the qualifier classical.
The central feature of the classical conditioning procedure is that the
animal is subjected to the paired presentation of two stimuli. The
result is a change in behavior (the dog comes to salivate to a stimulus
that did not originally elicit this response), but it should be noted
that there is nothing in the procedure that requires the animal to do
anything—pairings of the stimuli are scheduled irrespective of what
the animal does. Although Pavlov’s salivary conditioning procedure
has been little used in recent years, its essence is to be found in
other, widely used, training paradigms. For example:

1. The autoshaping procedure, often used with pigeons, is similar

to Pavlov’s original in that the bird is given trials in which the
presentation of a light (the CS), projected onto a movement-
sensitive response key, precedes the delivery of food (the US).
With training, the bird comes to peck at the lit key prior to the
arrival of food, a behavior known as known as sign-tracking.
2. In the conditioned emotional response procedure as commonly
used with rats, the CS is the presentation, for a period of a
minute or so, of an initially neutral stimulus (a light or the
sounding of a tone). A brief electric shock (the US) follows the
CS. After just a few trials the rat starts to show signs of fear (e.g.,
freezing) when the CS occurs.
 3. In flavor aversion learning (also called conditioned taste
aversion) the paired events are a distinctive taste as the CS (e.g.,
a saccharin solution) and a state of gastric distress as the US
(usually produced by injection of a mild toxin). One such trial
can be enough to produce a change in the rat’s behavior, so that
it now refuses the sweet solution that it consumed readily when
first presented.

In their details, the procedures used in these examples move

progressively further away from that used by Pavlov for salivary
conditioning. But all share these features: The experimenter arranges
paired presentations of CS and US and the CS comes to evoke
behavior that it did not evoke originally.

Instrumental (operant) conditioning

As a physiologist, Pavlov took to the study of learning as a way to
discover the workings of the brain. In the West a different
perspective prevailed. During the latter half of the nineteenth
century, biologists had become interested in the nature and extent of
animal intelligence—Darwinian theory required “continuity” among
species and thus it might be expected that intelligence, of the sort
shown by humans, should also be detectable in nonhuman species.
Experimental attempts to demonstrate this involved setting animals
puzzles of various sorts and showing that they could learn to solve
them. Perhaps the best known example comes from the work of
Thorndike, a psychologist from the United States. He confined cats
in an apparatus (appropriately called a puzzle box) from which they
could escape and gain access to a food reward, by manipulating an
arrangement of latches and levers. Initially the cat took a long time
to escape but, with repeated trials in the apparatus, performance
improved to the extent that the cat became able to open the box in
just a few seconds. Whether this behavior is evidence of real
intelligence is open to debate. Thorndike’s own interpretation—that
reward simply “stamps in” or reinforces the pattern of behavior that
happens to precede it—implies that it is not. However this may be, a
version of this basic procedure has been extensively used in
subsequent laboratory studies of animal learning.
The apparatus most commonly used has not been Thorndike’s puzzle
box but that devised by another psychologist from the United States,
B. F. Skinner (and thus known as the Skinner box). The experimental
subject (usually a rat) remains within the box for the duration of a
training session and presses on a lever, resulting in the delivery of a
reward (also referred to as a reinforcer) to an adjacent food cup (see
Skinner 1938). The measure taken is rate of responding, which is low
initially but increases with successive reinforced lever presses. This is
an example of appetitive conditioning (so called because the
procedure is one in which the animal comes to satisfy an appetite, in
this case for food). The apparatus can also be used to study the
aversive conditioning paradigm known as punishment by arranging
that the response produces an aversive event, such as an electric
shock to the rat’s feet. With this procedure the rate of a previously
reinforced response declines.
Much research has been done, particularly for the appetitive case, on
the effects of scheduling reinforcement for only some lever presses.
Response rate shows great sensitivity to such schedules. For
example, the rat will show bursts of rapid responding when it is
arranged that reward will be delivered only after a certain number of
responses has been made (the fixed ratio schedule, so called because
a fixed number of responses is required to earn each reward). On the
variable interval schedule, by contrast, only one response is
required but the experimenter arranges that reward will be made
available only when a certain length of time has passed since the last
rewarded response. The interval between consecutive rewards varies
around some mean value (e.g., from a few seconds to several
minutes, around a mean of 1 minute). In these circumstances the rat
comes to respond at a fairly steady rate. Other commonly used
schedules include the variable ratio schedule, in which the number
of responses required to earn the reward changes unpredictably, and
the fixed interval schedule, in which reward is presented for the first
response that occurs after a fixed amount of time since the previous
reward was earned.
At the time of its introduction, the distinction between the
instrumental training procedure and that used by Pavlov was not
fully appreciated, and the term conditioning was applied to both.
But, as Skinner himself was at pains to make clear, there is an
important difference. Both involve the presentation of a biologically
important event (such as food), but in the Pavlovian procedure this is
independent of what the animal does, whereas in instrumental
conditioning it is a consequence of the animal’s response (the
response is instrumental in producing the outcome). Skinner
preferred to use the term operant conditioning to emphasize that in
this form of learning the animal operates upon its environment.

Complex conditioning procedures

In the simple conditioning procedures just described, the
experimenter arranges a contingency between two events, the CS
being followed by the US in the classical procedure, the response by
the reinforcer in the instrumental case. More recent research has
continued to use these basic procedures but has introduced a variety
of elaborations.

Conditional control
The experimenter can arrange, for a rat lever-pressing in the Skinner
box, that the response will produce food only when certain other
conditions are met (e.g., only when a tone is sounding). The
response-reinforcer contingency holds, therefore, only when the tone
is on. The rat will come to respond only in the presence of the tone
and will withhold responding in its absence. This phenomenon,
known as stimulus control, reveals an ability to learn about higher-
order relationships (food will be presented only if a response is
made, and even then, only when the tone is present). Such
conditional learning is not confined to the instrumental procedure.
In the Pavlovian analogue the animal receives trials on which CSA is
followed by the US only if it is accompanied or preceded by some
other stimulus B; no US occurs when A is presented on its own. The
animal shows its sensitivity to this conditional relationship as it
comes to produce the CR to A only on the B-A trials.

Discrimination learning
In the procedure just described the animal must discriminate
between two sets of circumstances in which a given CS occurs. The
standard discrimination training procedure is similar in principle,
but two different stimuli are used and they are given different
consequences. For example, a pigeon may be given autoshaping
training consisting of trials in which presentations of a red light (the
positive stimulus) are followed by food, intermixed with trials on
which a green light (the negative stimulus) is presented, but no food
follows. Not only will the bird come to respond to the red light, as
might be expected, it may also initially show some responding to
green. This phenomenon is known as generalization (the CR
established to the CS generalizes to another, similar stimulus). With
extended training, however, the tendency to respond to green will
decline—a discrimination between red and green will be established.
The instrumental version of discrimination training often involves a
choice procedure. For example, the rat at the choice point of a maze
can be confronted with two arms, one black and one white. If a move
into the white arm gives access to food whereas choice of the black
does not, the rats will, after a few training trials, come to choose the
white arm on each occasion, even when the left-right position of the
arms is swapped at random from trial to trial. This same apparatus
can be used to study spatial discrimination learning. In this case no
extra visual cues are presented and reward is available after choice of
one arm, and never after choice of the other. Rats learn spatial
discriminations readily, even when the task cannot be solved by
learning to perform a specific response. In one widely used
procedure (Morris 1981) the rat is placed in a pool of water from
which it can escape by climbing on to an invisible platform, the top of
which lies just below the surface. After a few trials the rat learns to
swim directly to the platform from whatever point it is placed in the
pool. To do this, the rat must have learned something about the
spatial relationships between the platform and the various cues
present in the room in which the pool is located.

Mechanisms of Learning

Associative analysis of conditioning

Classical conditioning
In classical conditioning the experimenter arranges that two events,
the CS and the US, occur together. What could be more natural than
to assume that what the animal learns directly reflects this? Pavlov’s
own interpretation was that activation of the brain center
responsible for perception of the CS, at the same time as the center
responsive to the US was also activated, resulted in the strengthening
of a link or association between them. The existence of this
stimulus–stimulus (S–S) association means that presentation of the
CS will be able to activate the US center (by way of the link shown as
link 1 in Figure 8.1(a)) and thus produce the behavior normally
evoked by the US itself (i.e., produce a CR). This analysis implies that
classical conditioning should be possible even when the event used
as the US elicits no response. Sensory preconditioning supplies
an example of such an effect. In this procedure (see Table 8.1) the
animal receives initial conditioning trials with two stimuli that elicit
no obvious response, for example a light is paired with a tone. After a
second stage of training, in which the tone is paired with a US, it is
found that the light is now able to elicit the CR that has been
conditioned to the tone. This result seems to indicate that the light
has acquired the power to activate the representation of the
subsequently conditioned tone—that the initial stage of training
established an (S–S) association between the light and tone.
Figure 8.1 (a) Associative connections in classical conditioning.
Rectangles represent events in the environment and circles represent
the brain centers activated by these events. Solid lines represent
inbuilt links; dotted lines indicate links that might form as a result of
conditioning: (1) an S–S link; (2) an S–R link. CS: conditioned
stimulus; US: unconditioned stimulus: R: response. (b) An
associative mechanism for conditional control. CS A has a
connection to the US, but this will only operate fully when CS B has
been presented.
In classical conditioning the experimenter arranges that two events,
the CS and the US, occur together. What could be more natural than
to assume that what the animal learns directly reflects this? Pavlov’s
own interpretation was that activation of the brain center
responsible for perception of the CS, at the same time as the center
responsive to the US was also activated, resulted in the strengthening
of a link or association between them. The existence of this
stimulus–stimulus (S–S) association means that presentation of the
CS will be able to activate the US center (by way of the link shown as
link 1 in Figure 8.1(a)) and thus produce the behavior normally
evoked by the US itself (i.e., produce a CR). This analysis implies that
classical conditioning should be possible even when the event used
as the US elicits no response. Sensory preconditioning supplies
an example of such an effect. In this procedure (see Table 8.1) the
animal receives initial conditioning trials with two stimuli that elicit
no obvious response, for example a light is paired with a tone. After a
second stage of training, in which the tone is paired with a US, it is
found that the light is now able to elicit the CR that has been
conditioned to the tone. This result seems to indicate that the light
has acquired the power to activate the representation of the
subsequently conditioned tone—that the initial stage of training
established an (S–S) association between the light and tone.
Table 8.1 Experimental designs in classical conditioning
Phase 1 Phase 2 Test
Sensory Preconditioning A->B B - > US A
Blocking A - > US AB - > US B
Note. A and B represent conditioned stimuli; US, unconditioned stimulus. In each phase of
training several presentations of each set of events will occur.

It has been suggested that another type of association is important in

classical conditioning. Although the experimenter arranges a pairing
of stimuli, the animal will necessarily experience a pairing of a
stimulus (the CS) and a response (the UR). A stimulus–response (S–
R) association could thus form between the center that is activated
by the CS and the center that is responsible for organizing and
emitting a response (shown as link 2 in Figure 8.1(a))—indeed, for
many years this S–R account was the dominant interpretation of
classical conditioning, being championed by adherents of a version
of behaviorism that tried to explain all behavior in terms of the S–R
formula. But experimental analysis has failed to support this
doctrine. In one study, Holland and Straub (1979) gave rats pairings
of a noise (CS) and a food pellet (US) so that they acquired a goal-
tracking CR of approaching the food cup, when the noise came on.
In a second stage of training the rats were given free access to food
pellets followed by a nausea-inducing injection so that flavor
aversion learning occurred and the rats were now unwilling to eat
pellets of this type. A final test in the original conditioning apparatus
revealed that the noise was now less effective in evoking the CR. This
result is not predicted by the S–R analysis—there is no reason why
the link between CS and R (link 2 of Figure 8.1(a)), once formed,
should be affected by a subsequent change in the value of the US; but
if the S–S theory is correct, and the CR depends on the ability of the
CS center to activate the US center, then sensitivity to change in the
value of the US is just what would be expected.
Although it is now widely accepted that S–S associations play a
dominant role in classical conditioning, it would be wrong to jump to
the conclusion that animals cannot form S–R associations—indeed,
the experiment just described provides some evidence that they can.
In addition to measuring food cup approach, Holland and Straub
(1979) measured another CR: the increase in general activity that
occurs in the presence of an auditory CS signaling food. They found
that the vigor of this CR was not influenced by devaluation of the US,
suggesting that the behavior depends on the rats having formed an
S–R link. This observation points to a more general conclusion. For
some psychologists the debate about the S–R and S–S mechanisms
for classical conditioning has been seen as a battle over rival
approaches to the discipline as a whole, some equating behaviorism
with the S–R account; others, persuaded of what is sometimes
referred to as a “cognitive” perspective, arguing in favor of the S–S
account. But, as the results just described show, the choice between
these alternative approaches cannot be made in terms of the nature
of the association in conditioning. Animals will form both types of
association, according to the details of the training procedure.

Instrumental conditioning
The associative analysis has also been applied to instrumental
conditioning, and the debate over the nature of the association
involved has much in common with classical conditioning. Again, the
S–R account, which dominated for many years, has been challenged
by later research showing the importance of another association (in
this case between the response and its outcome); and again the
resolution turns out to be that both forms of association may be
established, given the appropriate conditions.
Thorndike’s interpretation of instrumental learning was that the
relevant association was between S and R. This was couched in terms
of his law of effect. In its most general form, this law is simply that
the effect produced by an action will change the likelihood of that
action for the future (e.g., the frequency of lever-pressing will
increase when that behavior produces a food pellet). In the specific
form proposed by Thorndike it was suggested that the mechanism
responsible for the increase in response frequency was the
strengthening (reinforcement) of a connection between the
current stimulus situation (the sight of the lever in this example) and
the response (depressing the lever).
Experiments investigating this interpretation have made use of a
reinforcer-devaluation procedure, analogous to the US-devaluation
experiment of Holland and Straub (1979), described earlier. In one of
these, Adams and Dickinson (1981) trained rats to press the lever in a
Skinner box for food, training that should, according to the theory,
simply establish a link between the S of the lever and the R of
pressing. In a further stage of training the rats ate food pellets of the
type they had previously earned by lever pressing and this was
followed by an injection known to induce nausea. The fact that the
rats now showed an aversion to this form of food should, according
to the S–R theory, be irrelevant to the behavior shown in the Skinner
box—the reward would have done its job of reinforcing the S–R
connection in the first stage of training and subsequent changes in
value should be of no consequence. It was found, however, that when
the rats were returned to the Skinner box, those that had been
subjected to the reinforcer-devaluation procedure showed a reduced
willingness to press the lever. We may conclude that the role of the
reinforcer in this procedure is not simply to stamp in some
connection between a stimulus and a response—rather, the animal
clearly knows something about what the outcome of its response will
be. Put in associative terms, this means that the critical link is
between R (in this case the lever-press) and S (the outcome of the
response). When the outcome is of value, the R–S connection results
in a high rate of response; when it is not, the rate of response will be
low.
It remains to point out that some conditions of instrumental training
will generate responding that is independent of the current value of
the reinforcer; in these cases the rat will continue to perform a
response that was initially established using a reinforcer that is no
longer of value. There are several possible reasons why this might
arise, but among them is the possibility that what the animal has
learned in initial training is an S–R rather than an R–S association.
Particularly intriguing in this context is the observation that animals,
given extensive initial training, can become insensitive to the value of
the reinforcer, an outcome consistent with the fact of everyday
experience that an act acquired initially with reference to the effect it
produced can, if repeated often enough, turn into an automatic habit.

Principles of association formation

Experimental study has confirmed that contiguity is important for
association formation—conditioning occurs best when the CS and the
US are presented close together in time. If a delay is imposed
between the offset of the CS and the presentation of the US,
conditioning is impaired. This is not to say that conditioning is
impossible in these conditions (in the case of flavor aversion learning
some effect is found even when several hours separate CS and US),
but this observation may simply indicate that the central activity
induced by the CS may persist for a time and thus still be present
when the US occurs. A more fundamental issue is raised by
experiments showing that sometimes conditioning can fail to occur
even when the animal receives contiguous presentations of a CS and
US.
The clearest demonstration that contiguity does not necessarily
generate conditioning is supplied by the phenomenon known as
blocking. This phenomenon was discovered by Kamin (1968) in
experiments using the conditioned emotional response procedure.
The procedure involved two stages of training and two CSs (see Table
8.1). In the first stage rats received trials in which one CS (a light)
was followed by shock US so that a CR was established. CS–US
pairings continued in the second stage but now another CS (a noise)
was presented along with the light. Although the animal experienced
a series of contiguous presentations of the noise and the shock, in a
final test phase in which the noise was presented on its own no CR
was observed. Pretraining with the light had blocked learning about
the noise that was added in the second stage (control subjects that
received just the stage of training with the noise–light compound
paired with shock showed adequate conditioning to the noise).
This result shows that contiguous presentation of the CS and US is
not enough to establish a link between them; rather it seems that the
CS must supply information about the upcoming US. In the blocking
procedure the first stage of training ensures that the occurrence of
the US is fully predicted by the light—the noise is redundant
(supplies no new information) and is thus not learned about. To
some extent, this interpretation is just a re-description of the result
obtained, and the theorist will want to specify the mechanism by
which it occurs. This was addressed in the influential theory first
proposed by Rescorla and Wagner (1972; see Box 8.1) and
subsequently developed by Wagner (1981). These theorists took as
their starting point the notion that a CS–US link would not form if
the US was already expected. This means that the US center is
already activated, by way of a CS–US link, before the US actually
arrives. If we assume that a US center that is already activated
cannot react in the normal way when the US is presented then the
experimental results may be explained. Although, in the blocking
experiment, the experimenter arranges contiguous presentations of a
noise CS and the US, this is not what the animal experiences—the CS
will activate its center but (given that the relevant center is already
activated by the light) the US will not. In a sense, therefore, the
central role of contiguity is maintained in this account. What
matters, however, is not the contiguous occurrence of events in the
animal’s environment, but the concurrent presence of appropriate
activity in the relevant brain centers.
 Box 8.1 Rescorla-Wagner Model

The account of conditioning proposed by Rescorla and Wagner

(1972) was expressed in terms of a simple equation that specifies
by how much the strength of the associative connection between
a CS and US will change as the result of a conditioning trial. A
simplified version is given below. The change in the associative
strength (∆V) that occurs on a conditioning trial is determined
by:

ΔV = α(λ−∑V)
where α is parameter that varies with the salience of the CS, λ
represents the maximum associative strength that the US can
support, and ∑V is the summed associative strength of all CSs
present on the trial. It will be noted that when ∑V equals λ there
can be no further change in associative strength, as the value
inside the brackets will be zero. Conditioning with A as the CS (as
in the first stage of the blocking experiment, see Table 8.1) will
thus proceed until the associative strength of A (VA) has risen to
the value of λ. Further trials in the second stage, in which A is
presented in compound with B will produce no further increase
in VA. What is more, they will not produce any increase in the
associative strength of the newly introduced B, as the value of
λ−∑V will be zero by virtue of the associative strength of A. This
outcome is the result obtained experimentally (the blocking
effect).
This simple equation neatly captures the notion that an expected
US (one preceded by a CS that has a high level of associative
strength) will be poor at supporting further learning. It has
proved to have wide explanatory power and has been very
influential not only in the study of conditioning, but also in the
psychology of cognition more generally.

General applicability of associative theory

The associative theory of conditioning is important not simply
because it can explain effects observed in conditioning experiments.
Its wider importance comes from the suggestion (put forward
originally by Pavlov himself but taken up enthusiastically in recent
years by adherents to the theory known as connectionism) that
associative mechanisms may underlie all forms of learning. To
establish this point it would be a useful first step to show that the
associative analysis can apply to the various forms of learning
described earlier in this chapter. For the case of discrimination
learning and generalization, this is a fairly simple matter.
Generalization will occur because two similar stimuli hold features in
common; pairing one of these stimuli with a US will thus mean that
features of the other will also be associated with the US, and thus this
untrained stimulus will tend to evoke the CR to some extent. But
presentation of a CS in the absence of the US will reduce the effective
strength of any association that may exist between them (a
phenomenon referred to as extinction). What follows is that the
discrimination training procedure will reduce the strength of these
common features. As a result the animal will cease to respond to the
negative stimulus but will continue to respond to the positive
stimulus, the unique features of which will still be strongly associated
with the US. These principles also successfully predict the behavior
shown when the animal is required to choose between the two
stimuli.
The explanation of conditional control is still a matter for debate.
One possibility is that it reflects the formation of an association
between a stimulus and another association. When an animal comes
to respond to CSA only when B has been presented this may be
because a link has been formed between B and the A–US association
that allows B to activate the association. (Stimulus control is
explained in much the same way by the suggestion that the stimulus
is able to activate the R-S association.) An alternative possibility is
that the animal is able to put A and B together as a configural cue
and that conditioning occurs to this configuration rather than to any
of the separable elements. An advantage of this interpretation is that
it suggests a possible explanation of the spatial discrimination
learning phenomena described above. The ability of the rat to find its
way to the platform in a swimming pool is not easily explained in
terms of conditioning to any single event, but it can be analyzed in
terms of the acquisition of behavior that is guided by a configuration
of cues.
More problematic for the associative analysis are the seemingly
simple learning effects produced by exposure to a single stimulus. By
their very nature (since only one stimulus is presented) it is difficult
to explain these effects in terms of the formation of an association
between two events. Rather it seems that mere exposure to a
stimulus can engage a learning process that changes the way in
which that event is perceived. One popular account of habituation is
that repeated exposure to a stimulus allows the animal to build up a
“model” or central representation of that event. Habituation occurs
because the incoming stimulus is found to match an existing central
representation; a novel stimulus, one that finds no match, evokes a
response. And animals that possess well-formed representations of
stimuli might be expected to be good at discrimination between
them; that is, they should show the perceptual learning effect. It
remains to be seen whether associative mechanisms are involved in
the learning process by which such representations are formed.

Procedures for the Study of Memory and

their Results

The preceding sections considered the means by which behavior can

be modified by experience. For this process to occur, there must be a
representation of past experience. In simple conditioning, for
example, a conditioned response could not emerge during the
conditioned stimulus unless there was some memory of the previous
pairings with the unconditioned stimulus. Thus, memory can be
thought of as the effect of past experience on behavior in the present.
There may be many kinds of memory, but the two major divisions
are between shorter term, working memory, and longer term,
reference memory. The memory of previous pairings in a
conditioning task would be an example of reference memory. The
reference memory that is formed is the content of learning. Working
memory, on the other hand, is a shorter-term form of storage. The
ability to remember where you parked your car while you are in the
market is an example of working memory.

Working or short-term memory

The study of working memory was first undertaken by Walter Hunter
(1913) when he developed the delayed-response paradigm. An
animal was placed in an apparatus containing several boxes, one of
which was baited with food. The baited box was signaled by the
presentation of a light stimulus. Then, after a delay (with the light
off) the animal was released and allowed to select a box. If the animal
chose the correct site, then it received the food. A correct choice
indicates that the animal remembered the baited location over the
delay period. One of the problems with Hunter’s procedure was that
some animals learned to orient themselves toward the baited box
during the delay period, thereby avoiding having to retain any
memory of which box was baited.
The next major procedural development occurred when Blough
(1959) introduced the delayed matching-to-sample (DMTS)
paradigm for studying memory in pigeons. A trial shown in Figure
8.2(a), typically involves the presentation of a sample stimulus (a red
keylight) followed by two choice stimuli (red and green keylights).
The pigeon has to choose the same item as the sample to receive a
reward, thereby matching the sample. By inserting a delay or
retention interval between the sample and choice, it is possible to
determine how well the pigeon can remember the sample over time.
Figure 8.2(b) displays some data from DMTS (Grant 1976). Accuracy
was nearly perfect at the 0-s retention interval but decreased as the
interval between sample and choice was increased, providing
evidence that the sample is forgotten over time. Thus, it can be seen
that a key feature of working memory is its time-limited nature.
Additional aspects of forgetting will be considered in the following
section.
Figure 8.2 (a) A typical color-based delayed matching to sample
task where the sample is presented on the center key followed by a
retention interval and then the presentation of the choices on the
side keys. Choosing the item that matches the previous sample
results in reward, in this case the red key. (b) The effect of increasing
retention interval on performance in a choice delayed matching-to-
sample task with pigeons. As the time between the sample and choice
is increased, performance progressively deteriorates demonstrating
the time-limited nature of working memory. Adapted from Grant
(1976).
A third method for studying working memory is list learning, which
has been popularly used with humans and more recently adapted for
animals. In one experiment, Harper et al. (1993) demonstrated list
learning in a 12-arm radial maze. The rats were taught a list of
seven items by allowing access to only one baited arm at a time. The
rat was placed in the center of the radial maze and one arm was
opened. The rat traveled down the arm and retrieved the reward and
then returned to the center. The first arm would then close and a
second arm would open, and so on until seven arms had been visited.
After a 5-s delay, the rats were then given tests with pairs of arms.
One arm was an item from the previous list, whereas the second arm
was novel. The rat had to visit the novel arm to receive food. By
testing different pairs of arms, Harper et al. determined how well the
rats remembered the individual items in the list. Shown in Figure 8.3
are results that demonstrate that (like humans) the rats best
remembered the first and last items in the list and performed more
poorly on intermediate items, a serial position curve. The
superior memory for early items in a list is a primacy effect,
whereas the superior memory for items near the end of a list is a
recency effect. Recency effects are often easier to demonstrate in
animals than primacy effects, but research with dogs in the radial
maze has shown evidence of a primacy effect but no recency effect
(Craig et al. 2012). It is possible that recency and primacy vary by
species and by the type of task. The possible source of primacy and
recency effects will be discussed in the following section on
forgetting.
Figure 8.3 A serial position curve obtained from rats in a radial-
arm maze. Memory is superior for items at the beginning (primacy
effect) and end (recency effect) of the list, but poorer for items in the
middle of the list. Adapted from Harper et al. (1993).

Forgetting of items in working memory

Because working memory is usually time-limited, forgetting occurs.
One source of forgetting is passive decay (see Figure 8.2(b)). Roberts
and Grant (1976) proposed that items in working memory consist of
traces that decay over time. The strength and durability of the trace
are determined by factors such as the intensity and duration of the
stimulus. In DMTS, there may be multiple items in memory (the
current sample, previous samples, and previous choices). The item
with the strongest trace value will determine the choice on the
current trial. Errors occur when the traces of items become so weak
that there is no clear winner, such as when the delay interval is long.
This account correctly predicts two common effects in DMTS—
increasing either sample duration or the time between trials
increases performance (Roberts & Grant 1976; Nelson & Wasserman
1978). Increasing sample duration would enhance the strength of the
trace of the current sample, whereas increasing the time between
trials would allow previous items from earlier trials to decay further
and thereby produce less interference with the current trace.
A second source of forgetting is thought to be due to the limited
capacity of working memory. The classic 7 ± 2 rule in humans
exemplifies the capacity issues of working memory (Miller 1956). It
has been consistently reported that humans can only maintain about
seven (give or take a couple) items in short-term memory before
interference effects begin to occur. The same also appears true in
other animals, although the limits may vary across species. Wagner
(1976) proposed that there is a limit on the number of traces that can
be maintained at once. The introduction of a new item in working
memory will weaken or displace previous traces, and previous items
in working memory can interfere with new information.
One factor that can alleviate the problem of limited capacity is the
ability to chunk items together into meaningful clusters. An example
of chunking was provided by Terrace (1991) in a list-learning
experiment with pigeons. Pigeons learned lists of five items, which
were composed of colors and/or shapes. The pigeon had to peck five
lighted keys in a specified order to receive a reward. The five lit keys
appeared at random positions in an array of eight possible keys
(Figure 8.4(a)). Different groups received different lists (Figure
8.4(b)). The list for Group II contained a natural chunk, with three
colors followed by two shapes. Group IV received a larger initial
chunk of four colors followed by one shape. The remaining groups
experienced sequences that did not contain meaningful chunks of
items. It was discovered that the pigeons in Groups II and IV learned
more quickly and they were faster at completing the sequences at the
end of training. The results indicate that the presence of chunks in a
list aids learning, perhaps by decreasing the load on working
memory.
Figure 8.4 An experiment on chunking in a list-learning task. (a)
Pigeons were given an array of eight keys, of which a random five
were lit with different colors or shapes. The pigeons had to peck the
lit keys in a previously specified order. (b) Different groups were
trained with different sequences of key lights that contained either
two chunks (Groups II and IV) or no chunks (Groups I, III, and V).
The lists that contained chunks were learned faster than lists that
didn’t contain chunks. Adapted from Terrace (1991).
One facet of limited capacity is that interference effects may occur.
There are two main types of interference: proactive and retroactive.
Proactive interference is the effect of past information on current
retention ability. The interfering effect of short intertrial intervals in
the DMTS task is one example. Retroactive interference is the
effect of information that intervenes between exposure to an item
and the later recall of the item. This can be produced in DMTS by
adding a distracter cue between the sample and choice. One situation
where proactive and retroactive interference may operate together is
the serial position curve (Figure 8.3). Performance in the middle of
the list may be impaired because both proactive (earlier items) and
retroactive (later items) interference operate together (Baddeley
1976). Items near the start of the list would be remembered more
easily because they suffer solely from retroactive interference, and
items at the end of the list would be remembered more easily
because they suffer solely from proactive interference.
A final feature of working memory is that rehearsal aids the
maintenance of items for longer periods of time. One way that
rehearsal has been studied in animals is through the use of the
directed-forgetting paradigm (Maki & Hegvik 1980). The
paradigm is an extension of the DMTS procedure in which a cue is
added after the sample to indicate whether the animal will be tested
for sample memory or not. Occasionally, the animal receives a forget
cue but is given the memory test anyhow; on these trials,
performance is usually impaired relative to remember-cued trials.
The fact that a forget cue results in poorer memory indicates that
animals may engage in maintenance rehearsal when they have to
remember the sample.
Reference or long-term memory
All learning involves the formation of reference memory and vice
versa, so in many ways the concept of reference memory is difficult
to distinguish from the concept of learning. Most studies of learning
assess the formation of reference memories, whereas most studies of
reference memory assess the capacity and durability of information
that is learned. Reference memory can be assessed in the context of
almost any experiment in which new information is learned.
Reference memory, unlike working memory, appears to have a
relatively unlimited capacity and long duration. One of the most
striking examples in animals is found in the New World corvid,
Clark’s nutcracker (Nucifraga columbiana), which stores about
30,000 seeds in thousands of locations. The nutcrackers store the
seeds in the fall and retrieve them throughout the winter and spring,
with a high success rate. Laboratory studies have indicated that this
ability is due to memory, rather than other factors such as storing or
retrieving strategies. Thus, reference memory can contain a large
number of distinctive items over a period of at least months.
This ability has been demonstrated in other species as well. Vaughan
and Greene (1984) trained pigeons to discriminate between pairs of
photographs shown together on a viewing screen. For each pair, one
of the photos was designated consistently as the reinforced stimulus
and the other as the non-reinforced stimulus. If the pigeon correctly
pecked the reinforced stimulus, then it would receive food. If, on the
other hand, it erroneously responded to the non-reinforced stimulus,
then food would be withheld. Initially, the birds were trained to
discriminate 40 pairs of photographs. Once they had successfully
learned these, then they were given further sets of photos, until they
had learned 160 different pairs of slides. Although learning of 320
slides over hundreds of training sessions may not be as impressive as
the cache-recovery behavior of the Clark’s nutcracker, the study by
Vaughan and Green indicates that pigeons can maintain a large
number of items in reference memory.
Vaughan and Greene (1984) went on to test the durability of the
pigeons’ memory for the photos that they had received over the
course of the study. The pigeons were tested for memory of the 320
photos after either 237 or 629 days and were found to have retained
the discrimination well above chance. These results and other studies
indicate that animals can retain information over periods of weeks to
years much like humans. However, long-term retention is not
perfect; there is some degree of forgetting and there may be capacity
limits (Cook et al. 2005). Capacity limits may be more likely when
animals are solving problems that are not tied to specialized
adaptations such as food storage in the Clark’s nutcracker (Qadri et
al. 2018).
A special type of reference memory that has received interest in
animals is episodic-like memory. Tulving (1972) defined episodic
memory as memory for a personal experience and this type of
memory is the basis for autobiographical memories in humans.
Episodic-like memories contain the combination of three elements—
what, where, and when. The first study to demonstrate episodic-like
memory in animals was by Clayton and Dickinson (1998) in Western
scrub jays. The jays cache and later retrieve food in their natural
environment, much like the Clark’s nutcrackers. In their study, the
jays were allowed to cache peanuts and waxworms within sand-filled
ice cube trays (there were lego blocks positioned around the trays to
provide landmarks for orientation). There were two caching periods
that were 120 hours apart followed by a retrieval period 4 hours after
the second caching period (see Figure 8.5). In cache 1, the jays were
allowed to cache either peanuts or waxworms on one side of the tray
and in cache 2 the jays cached the other food type. The jays usually
prefer waxworms, but on trials when the waxworms were cached first
they would be rotten by the time of retrieval 124 hours later. The
peanuts were always good to eat. They found that the jays altered
their preferences depending on what they had cached at different
times, preferring peanuts when the worms were cached first (and
would be rotten at the time of retrieval) and preferring worms when
the peanuts were cached first. This shows memory for what (peanuts
versus waxworms), when (4 versus 124 hours ago), and where (the
position in the tray where the items were cached) memory, meeting
the definition for episodic-like memory. A control group that
received a condition where the waxworms never decayed showed a
general preference for the worms regardless of when they were
cached, demonstrating that the experimental group moderated their
retrieval preferences based on their experiences with the worms
decaying. Subsequent to these findings, the effects have been
demonstrated in rats in the radial arm maze as well, suggesting that
multiple species may have this ability (Babb & Crystal 2006).

Figure 8.5 A schematic of the experiment by Clayton & Dickenson

(1998) demonstrating episodic-like memory in scrub jays. The jays
could cache peanuts or worms at two different times, separated by
120 hours, and could retrieve items 4 hours after the second cache
time. The jay’s preferences varied depending on what they cached
first.

Forgetting of items in reference memory

Information in reference memory may simply be lost over time due
to passive decay. Experimentally, passive decay is difficult to
disentangle from other sources of forgetting. In working memory,
passive decay occurs quite rapidly so it is easy to observe in the
absence of other factors that may cause forgetting. In reference
memory, passive decay may occur over the course of months to
years.
There are two main factors that have been demonstrated to affect
storage and/or recall of information in reference memory—
consolidation and retrieval. Consolidation is a hypothetical
process that has been proposed to occur when new memories are
formed. In the minutes to hours following a learning experience, it is
believed that there is an ongoing process in the nervous system that
results in new memory formation. If this process is disrupted then
memory formation may be incomplete, resulting in poor recall of
information. For example, retrograde amnesia is commonly observed
in people following head trauma where patients suffer a loss of
memory for recent events that occurred prior to the injury.
Retrograde amnesia has been studied in psychiatric patients
undergoing electroconvulsive shock (ECS) treatment for severe
depression (Squire & Slater 1975; Squire & Fox 1980). The patients
showed deficits in recall of information that had been learned up to
three years prior to ECS treatment, but information acquired earlier
was not affected. Retrograde amnesia is usually temporally graded;
that is, the severity of disruption in memory is greatest for most
recent events and becomes progressively less severe as a function of
time prior to the onset of amnesia.
Evidence of memory consolidation has also been obtained in a
variety of species. For example, Duncan (1949) examined the effect
of ECS on reference memory in rats. The rats were trained with a
light that was followed by a shock to the feet. To prevent the shock
from occurring, the rat had to move to the other side of the box
before the light went off. Following each training trial, the rats
received ECS administration, with the time of ECS occurrence
varying from 20 seconds to 14 hours after the training trial between
different groups. As seen in Figure 8.6, when ECS occurred soon
after the trial, avoidance learning was greatly disrupted, but delayed
ECS had no effect on avoidance behavior. These results indicate that
the trial information was still being consolidated during the first few
minutes after a trial and was vulnerable.
Figure 8.6 The effect of electroconvulsive shock (ECS) treatment
on reference memory. When ECS was given soon after avoidance
training, performance suffered, suggesting that electro-convulsive
shock had disrupted memory consolidation. Adapted from Duncan
(1949).
Although interruptions in consolidation may result in poor memory
formation, once a memory is formed then any memory failures
appear to be due primarily to retrieval failures. Retrieval accounts
(Spear 1973; Lewis 1979) maintain that memories are not lost, but
instead cannot be retrieved from reference memory. Retrieval
theories propose that reference memories exist in two categories:
active and inactive. When an item is in active memory it is highly
likely to be retrieved, but when in item is in inactive memory it is
much less likely to be retrieved.
The primary evidence in support of retrieval theories comes from
studies that have used reactivation treatments. Deweer et al.
(1980) trained rats on a six-arm radial maze. One group of rats was
tested immediately after training (group Immediate), a second group
was tested in the maze after 25 days (group Delay), and a third group
was tested after 25 days (group Reactivate), but with a brief
reactivation treatment prior to testing. The reactivation treatment
involved placing the rat in a wire cage next to the maze for 90
seconds. Group Immediate performed at a high level of accuracy, but
group Delay demonstrated decrements in performance indicating
that they had forgotten the task. However, group Reactivate
performed at similar levels of accuracy to group Immediate. Thus,
despite the 25-day delay, the reactivation treatment resulted in a
recovery of performance on the task. These results suggest that
group Delay had not lost information from reference memory, but
rather failed to retrieve the relevant memories for the task. By
reactivating those memories, retrieval was boosted to normal levels.

Working versus reference memory

There are a number of differences between working and reference
memory that have been described thus far. For example, working
memory operates over a shorter time span, has a more limited
capacity, and is more prone to interference than reference memory.
Working and reference memory also interact. In terms of initial
storage, information must pass through working memory (perhaps
with rehearsal and/or repeated use) before that information is stored
in reference memory. Moreover, when an item is recalled from
reference memory, it is held in working memory during the time that
the item is in use. Items that are recalled more frequently are also
recalled more easily, through the facilitation of retrieval processes.

General and Special Processes

The experimental studies of learning and memory described above
were conducted on species chosen largely for their ease in adapting
to the special conditions of the laboratory. The stimuli and
reinforcers manipulated in the experiments and the patterns of
behavior chosen for observation were selected as much for the
convenience of the experimenter as for any other reason. The
experimenters who conducted this work were sustained by the hope
that these factors might be irrelevant—that their laboratory studies
might reveal principles of learning and memory that apply to most or
all species and to most or all stimuli and responses. To some extent
their hope appears justified. For example, Pavlov’s original results
were obtained from studies of the salivary reflex in dogs, but, as we
have seen, the basic associative principles he discovered have been
applied successfully to a range of other procedures and species.
Studies of the mechanisms of memory have also revealed a high
degree of similarity across a range of species. In the DMTS, task
performance deteriorates as a function of retention interval,
regardless of the species of animal that is being tested; serial position
curves of a similar shape have been obtained with several species,
although there are some differences in the time courses of these
curves; and there is considerable evidence for a capacious and
durable long-term memory in a variety of species. An exhaustive
survey of the relevant evidence conducted by Macphail (1982) led
him to conclude that the same basic learning mechanisms operate in
all nonhuman vertebrate species. (Humans were viewed as
exceptions by Macphail on the grounds of their unique ability to use
language to organize and encode information.)
This “general-process” view of learning has its merits but also its
challenges. There are a number of observations of belongingness
effects in which the ease of learning is affected by the nature of the
stimuli involved. Garcia and Koelling (1966) examined associations
of taste as the CS and nausea as the US (as is usual) and also used
exteroceptive cues as a CS and foot shock as a US. During training,
the rats drank a saccharin-flavored solution and at the same time a
flash of light and a burst of noise was presented. Half the animals
then experienced a nausea-inducing treatment as the US; the others
were given electric shock. In a subsequent test, rats in the first
condition showed an aversion to saccharin but were willing to drink
plain water even when this was accompanied by the light and noise.
Rats in the other condition drank saccharin readily but refused the
“bright noisy water.” The important feature of this pattern of results
is that it contains evidence that all the events used as CSs and USs
were effective and fully perceived by the rats. Clearly the taste was
effective as a CS since it readily became associated with nausea;
equally clearly the shock was effective as a US: it supported
conditioning perfectly well with the noise-light CS. Standard learning
theory thus has no reason to expect that an association would fail to
form between the taste CS and the shock US or between the noise-
light CS and the nausea-inducing US.
What the Garcia and Koelling results seem to show is that
association formation can occur selectively; that gastric upset is
especially readily associated with taste, whereas an event that
impinges of the body surface becomes associated more readily with
exteroceptive cues. It is not hard to see this as being a specialization
generated by the evolutionary history of the species—given that an
attack on the body surface is likely to be preceded by exteroceptive
cues (as when a predator strikes), and that gastric illness is likely to
be preceded by the characteristic taste of a particular food.
One reaction to results of this sort was to reject general-process
theory entirely. According to one influential learning theorist, such
results show that “our principles of learning no longer have a claim
to universality … that learning depends in very important ways on
the kind of animal being considered, the kind of behavior that is
required of it, and the kind of situation in which the behavior occurs”
(Bolles 1979, p. 165). Pursuing this view leads to the suggestion that
general-process learning theory has quite overlooked the significance
of the phenomenon—that the mechanisms responsible for learning
have evolved like all the other attributes an animal possesses, and
that they should be viewed as specializations that help the individual
to survive in the particular habitat in which its ancestors have
evolved. Rather than looking for general processes the investigator
should accept that each species may possess a niche-specific learning
mechanism, different from that possessed by other species. The
remarkable cache-recovering abilities of food-storing birds are often
cited in this context. Birds that are heavily dependent on food storing
have a larger hippocampus (a brain structure known to be involved
in spatial memory) than related species that engage in very little
storing behavior (See Chapter 1 for further discussion and Bolhuis &
Macphail 2001; Bolhuis 2015 for critical reviews). The implication is
that the special selection pressures operating on food-storers might
have induced structural adaptations that allow them to exhibit
learning processes substantially different from those found in other
species.
Having sketched out the arguments, it is time now to attempt an
assessment. The first thing to say is that, on closer examination,
much of the evidence cited against the general-process position is
not as damaging as it may seem. It consists largely of showing that
there are quantitative differences between the learning that occurs in
different training situations, not that the mechanisms involved are
different. Most species can be trained to acquire spatial
discriminations (and in these the hippocampus has often been found
to play an important role); food-storing birds may be better at it, but
may not be doing anything radically different from the rat that learns
to find the platform in a water tank. It is possible that such
quantitative differences reflect not so much differences in the basic
learning mechanisms involved as differences between animals in
their sensory and perceptual systems (see Macphail 1982).
The dispute between theorists who emphasize the generality of
learning mechanisms and those who are impressed by the special
abilities shown by certain species (e.g., Macphail & Bolhuis 2001;
Bolhuis & Macphail 2001; Shettleworth 2010; Bolhuis 2015) will not
be easy to resolve. A perspective may be obtained by considering
learning as a biological process like many others—such as, for
example, respiration. At one level of analysis, all animals that respire
aerobically share certain common processes—the demands of a
biochemical system that depends on oxidation for the generation of
energy require that this be so. But the precise nature of the
environment in which these processes operate in a given species may
require specializations (gills are good for gas exchange when the
oxygen is dissolved in water; lungs are more effective for air-
breathers). Learning and memory can be viewed in the same light. At
one level all animals must live in a world in which events have
consequences and accordingly it would be no surprise to find that all
are equipped with mechanisms that allow them to internalize such
relationships. But the special features of the niche occupied by any
given species make it possible that special mechanisms may evolve to
supplement the general processes shared by all.

SUMMARY AND CONCLUSIONS

Learning has been studied in the laboratory using procedures
such as classical and instrumental conditioning. In classical
conditioning, pairings of a previously neutral CS followed by a US
result in the emergence of conditioned responding during the CS.
In instrumental conditioning, a particular target behavior such as
pressing a lever is reinforced through the presentation of an
appetitive event such as food or punished through the
presentation of an aversive event such as shock. Over trials, the
target behavior will increase (with food delivery) or decrease
(with shock delivery) in probability of occurrence. These changes
in behavior can be explained by presuming that a link or
association occurs between the events that the experimenter
arranges. Memory may be closely related to learning; it is
operationally defined as a delayed effect of an experience on
behavior. Working memory is studied with a delayed-response
paradigm and is usually time-limited, capacity-limited, and
susceptible to interference. Reference memory is the source of
storage of learned information and has no clear capacity or
duration limits, but interference can occur in the form of
consolidation and/or retrieval failures. Although the basic
mechanisms of learning and memory vary somewhat across
species, a general-process view can incorporate most instances of
learning and memory across the animal kingdom.

FURTHER READING
For a review of earlier work on animal learning and memory see
Spear et al. (1990), who cover not only the topics dealt with in the
present chapter but also includes information on the development of
learning and memory in young animals. An approachable account
that deals with later work is provided by Pearce (2013). The book
edited by Mackintosh (1994) is an advanced text that will make
considerable demands on the reader but which provides state-of-the-
art expositions by specialist contributors on many of the topics
covered in the present chapter. The contributors to this book are
mainly experimental psychologists. For a different perspective, see
the book by Shettleworth (2010); this book offers an emphasis on the
evolutionary approach to the understanding of learning and
memory.

REFERENCES
Adams, C.D. & Dickinson, A. 1981. Instrumental responding
following reinforcer devaluation. Quarterly Journal of
Experimental Psychology, 33B, 109–21.
Babb, S.J. & Crystal, J.D. 2006. Episodic-like memory in the rat.
Current Biology, 16, 1317–21.
Baddeley, A.D. 1976. The psychology of memory. New York: Basic
Books.
Blough, D.S. 1959. Delayed matching in the pigeon. Journal of the
Experimental Analysis of Behavior, 2, 151–60.
Bolhuis, J.J. 2015. Evolution cannot explain how minds work.
Behavioural Processes, 117, 82–91.
Bolhuis, J.J. & Macphail, E.M. 2001. A critique of the neuroecology
of learning and memory. Trends in Cognitive Sciences, 5, 426–33.
Bolles, R.C. 1979. Learning Theory, 2nd ed. New York: Holt,
Rinehart and Winston.
Clayton, N.S. & Dickinson, A. 1998. Episodic-like memory during
cache recovery by scrub jays. Nature, 395, 272–4.
Cook, R.G., Levinson, D.G., Gillett, S.R. & Blaisdell, A.P. 2005.
Capacity and limits of associative memory in pigeons.
Psychonomic Bulletin and Review, 12, 350–8.
Craig, M., Rand, J., Mesch, R., Shyan-Norwalt, M., Morton, J. &
Flickinger, E.2012. Domestic dogs (Canis familiaris) and the
radial arm maze: Spatial memory and serial position effects.
Journal of Comparative Psychology, 126, 233–42.
Deweer, B., Sara, S.J. & Hars, B. 1980. Contextual cues and memory
retrieval in rats: alleviation of forgetting by a pretest exposure to
background stimuli. Animal Learning & Behavior, 8, 265–72.
Duncan, C.P. 1949. The retroactive effect of electroshock on learning.
Journal of Comparative and Physiological Psychology, 42, 32–
44.
Garcia, J. & Koelling, R.A. 1966. The relation of cue to consequence
in avoidance learning. Psychonomic Science, 5, 123–4.
Grant, D.S. 1976. Effect of sample presentation time on long-delay
matching in the pigeon. Learning and Motivation, 7, 580–90.
Hall, G. 1979. Exposure learning in young and adult laboratory rats.
Animal Behaviour, 27, 586–91.
Harper, D.N., McLean, A.P. & Dalrymple-Alford, J.C. 1993. List item
memory in rats: Effects of delay and task. Journal of
Experimental Psychology: Animal Behavior Processes, 19, 307–
16.
Holland, P.C. & Straub, J.J. 1979. Differential effects of two ways of
devaluing the unconditioned stimulus after Pavlovian appetitive
training. Journal of Experimental Psychology: Animal Behavior
Processes, 5, 65–78.
Humphrey, G. 1933. The nature of learning. London: Kegan Paul.
Hunter, W.S. 1913. The delayed reaction in animals and children.
Behavior Monographs, 2, 1–86.
Kamin, L.J. 1968. “Attention-like” processes in classical
conditioning. In: M.R. Jones (ed.), Miami symposium on the
prediction of behavior: Aversive stimulation, pp. 9–33. Miami:
University of Miami Press.
Lewis, D.J. 1979. Psychobiology of active and inactive memory.
Psychological Bulletin, 86, 1054–83.
Mackintosh, N.J. (ed.). 1994. Handbook of perception and
cognition. Vol. 9: Animal learning and cognition. San Diego:
Academic Press.
Macphail, E.M. 1982. Brain and intelligence in vertebrates. Oxford:
Clarendon Press.
Macphail, E.M. & Bolhuis, J.J. 2001. The evolution of intelligence:
adaptive specialisations versus general process. Biological
Reviews, 76, 341–64.
Maki, W.S. & Hegvik, D.K. 1980. Directed forgetting in pigeons.
Animal Learning & Behavior, 8, 567–74.
Miller, G.A. 1956. The magical number seven plus or minus two:
Some limits on our capacity for processing information.
Psychological Review, 63, 81–97.
Morris, R.G.M. 1981. Spatial localization does not require the
presence of local cues. Learning and Motivation, 12, 239–60.
Nelson, K.R. & Wasserman, E.A. 1978. Temporal factors influencing
the pigeon’s successive matching-to-sample performance: Sample
duration, intertrial interval, and retention interval. Journal of the
Experimental Analysis of Behavior, 30, 153–62.
Pavlov, I.P. 1927. Conditioned Reflexes. New York: Dover. (reprinted
1960).
Pearce, J. 2013. Animal learning and cognition, 3rd ed. New York:
Psychology Press.
Qadri, M.A.J., Leonard, K., Cook, R.G. & Kelly, D.M. 2018.
Examination of long-term visual memorization capacity in the
Clark’s nutcracker (Nucifraga columbiana). Psychonomic
Bulletin and Review, 25, 2274–80.
Rescorla, R.A. & Wagner, A.R. 1972. A theory of Pavlovian
conditioning: Variations in the effectiveness of reinforcement and
 nonreinforcement. In: A.H. Black & W.F. Prokasy (eds.), Classical
conditioning II: Current research and theory, pp. 64–99. New
York: Appleton-Century-Crofts.
Roberts, W.A. & Grant, D.S. 1976. Studies of short-term memory in
the pigeon using the delayed matching to sample procedure. In:
D.L. Medin, W.A. Roberts & R.T. Davis (eds.), Processes of animal
memory, pp. 79–112. Hillsdale, NJ: Erlbaum.
Shettleworth, S.J. 2010. Cognition, evolution, and behavior, 2nd ed.
New York: Oxford University Press.
Skinner, B.F. 1938. The behavior of organisms. New York: Appleton-
Century-Crofts.
Spear, N.E. 1973. Retrieval of memory in animals. Psychological
Review, 80, 163–75.
Spear, N.E., Miller, J.S. & Jagielo, J.A. 1990. Animal memory and
learning. Annual Review of Psychology, 41, 169–211.
Squire, L.R. & Fox, M.M. 1980. Assessment of remote memory:
Validation of the television test by repeated testing during a
seven-day period. Behavioral Research Methods, Instruments
and Computers, 12, 583–6.
Squire, L.R. & Slater, P.C. 1975. Forgetting in very long-term
memory as assessed by an improved questionnaire technique.
Journal of Experimental Psychology: Human Learning and
Performance, 104, 50–4.
Terrace, H.S. 1991. Chunking during serial learning by a pigeon: I.
Basic evidence. Journal of Experimental Psychology: Animal
Behavior Processes, 17, 81–93.
Tulving, E. 1972. Episodic and semantic memory. In: E. Tulving & W.
Donaldson (eds.), Organization of Memory, pp. 381–403. New
York: Academic Press.
Vaughan, W., Jr. & Greene, S.L. 1984. Pigeon visual memory
capacity. Journal of Experimental Psychology: Animal Behavior
Processes, 10, 256–71.
Wagner, A.R. 1976. Priming in STM: An information-processing
mechanism for self-generated or retrieval-generated depression in
performance. In: T.J. Tighe & R.N. Leaton (eds.), Habituation:
Perspectives from child development, animal behavior, and
neurophysiology, pp. 95–128. Hillsdale, NJ: Erlbaum.
Wagner, A.R. 1981. SOP: A model of automatic memory processing
in animal behavior. In:N.E. Spear & R.R. Miller (eds.),
Information processing in animals: Memory mechanisms, pp. 5–
47. Hillsdale, NJ: Erlbaum.
9
animal cognition
JERRY A. HOGAN

INTRODUCTION
According to the American Heritage Dictionary, cognition is “the
mental process or faculty by which knowledge is acquired.” What
does it mean to study the mental processes in animals? For some
social scientists, “animal cognition” is an oxymoron: they believe
that nonhuman animals do not possess anything like what they
consider mental processes. For others, like the early behaviorists
(see Chapter 1), mental processes are too amorphous and difficult
to study. But animals, from insects, to fish, birds, rodents, and
apes, behave in ways that go well beyond reacting in an inflexible
manner to the situations in which they find themselves. They
perceive things, they learn, they find their way around in their
environment, they remember, they plan, they organize into
groups, and do many other things that imply mental processes
are occurring. During the past hundred years, psychologists and
behavioral biologists have developed many ways to study these
topics, some of which have been discussed in other chapters in
this book. In this chapter, I will be concerned primarily with
evidence that focuses on the nature and usage of knowledge. At
the end I will return to the question of what a mental process is.

A Wasp’s Story
In early summer, in certain regions of central Netherlands, digger
wasps of the species Ammophila campestris emerge from the
underground chambers in which they have developed from an egg
deposited the previous summer. The following description of the
wasps’ behavior is derived from the work of the Dutch ethologist
Gerard Baerends (1941) who studied these animals for his PhD
dissertation. Males emerge first, and females a few days later.
Courtship and mating occur as soon as the females appear. Each
female then goes immediately about the business of digging nests,
laying eggs, and providing her offspring with food. Nest building is a
complicated, arduous task. The female first searches for an
appropriate nesting site in the diluvial sand soil amid the heath
vegetation and pine trees. She may scratch and bite the ground in
several locations before choosing a site for the nest. When a suitable
site is found, she begins digging with her mandibles and forelegs.
With the sandy soil wedged between her mandibles and mouth, she
flies about 20 cm from the nest, sprays the sand on the ground, and
then returns to digging. This is repeated many times until she has
completed digging a tunnel about 2 cm deep with an elliptical, 2½
cm chamber at the bottom. She then closes the nest, which is a feat in
itself. First, she must find a clump of sand of the right size and shape
to fit in the tunnel but not fall into the chamber, and then other bits
of wood and stone to fill the tunnel. Finally, she covers it with sand
so that it is not visible. Then she searches for a caterpillar that she
stings to paralyze and carries back to her nest, either flying with the
caterpillar suspended under her body like a missile or in some cases
dragging it on the ground. She pulls the paralyzed caterpillar down
the tunnel into the chamber and deposits an egg on it (see Figure
9.1). After one or a few days, she returns to the nest, opens it, and
determines whether the egg has hatched. Whether it has or not, she
closes the nest. If the egg has not hatched, she returns again a few
days later and inspects again. If the egg has hatched and the larva
has devoured the caterpillar, she searches for a new caterpillar,
reopens the nest, replenishes the food supply, and closes the nest.
She may do this several times, depending on the amount of food in
the nest. Each time, she opens and closes the nest. In the course of
the summer, the female wasp may dig and provision five to ten nests
often with two or sometimes even three nests active at the same time,
each at a different stage of development. The nests of an individual
female may be a meter or more distant from each other, but the nests
of other females are located in between.
Figure 9.1 (a) Ammophiia sabulosa, a close relative of A.
campestris. Copyright Beentree, Wikimedia creative commons. (b)
Steps in provisioning and egg-laying the Ammophila nest. 1.
Bringing the prey. 2. Opening the nest. 3-6. Dragging the prey into
the nest. 7. Egg laying. 8. Leaving the nest. From Baerends (1941).
There are many more fascinating aspects of the behavior of these
wasps, but the details presented here already illustrate the
considerable cognitive abilities required of these tiny (about 1½ cm)
creatures. The female must first discover a suitable location for the
nest and perform the activities necessary to build the nest and to
close it. She must be able to find and recognize a caterpillar; she
must be able to find her way back to the nest after a foraging
expedition and then discriminate among her nests and the nests of
other females. She must inspect each nest to plan how much food
each nest will need. She must remember the state of development of
each nest even when multiple nests are at different stages at the
same time. How does she do these things? These are the types of
questions that are discussed in this chapter.
In the previous paragraph, I have used the words discover, build,
forage, recognize, discriminate, remember, inspect, and plan. I will
use these and similar words throughout this chapter to describe the
activities of many animals, including humans. These are all words
naming the function of different activities of an animal, but it is
important to realize that these words do not imply that the cognitive
mechanisms underlying these activities are the same in different
species. In particular, it is important not to assume that these
mechanisms are the same as they are in people. Doing so is called by
the name anthropomorphism: attributing human characteristics
to other animals. I chose the behavior of digger wasps to illustrate
the problems that mental processes are used to solve because digger
wasps are so different from humans that attributing human abilities
and motives to them is not likely. Such attributions, however,
become much more likely with animals we feel are more similar to
us, as we will see later in the chapter.

Basic Mechanisms: Knowledge

The definition of cognition given above uses the word “knowledge”:

acquiring knowledge is the purpose of mental processes. But what is
knowledge? From the point of view of this chapter, knowledge refers
to structures residing in the mind (or brain) of an animal that,
when activated, produce a result that is behaviorally interesting: a
perception, a memory, an idea, an intention, a thought, a feeling, or a
motor response. Various words are used to refer to these structures
including representation, behavior mechanism (cf. Chapter 3),
and mental structure. As well, many researchers use the word
“brain” as a synonym for “mind” and neural ensemble or engram
as synonyms for mental structures. These words are not pure
synonyms, but they all refer to the concept of knowledge. It should
also be noted that we humans are often aware of seeing something,
remembering something, or thinking of something, but there is no
way of knowing what other animals might be aware of. The
experimental work of animal behavior researchers is devoted to
determining the properties of the mental structures of other animals
and discovering how they function, regardless of whether or not the
animal is aware.
How is knowledge acquired? From the point of view of this chapter,
knowledge is acquired through general developmental principles as
described in Chapter 7, prefunctionally before birth and elaborated
through functional individual experience after birth. The processes
after birth are often subsumed under the term learning, and these
processes will be considered in more detail throughout the chapter.
The cognitive mechanisms, or knowledge, possessed by the digger
wasp are clearly present when it emerges from the nest, though there
is also some elaboration and formation of new cognitive mechanisms
afterward. In other animals, too, many cognitive mechanisms are
present at birth or shortly thereafter, and new ones continue to
develop throughout the life of the individual.

Motor and Perceptual Mechanisms

Ethologists have been studying motor and perceptual mechanisms
for many years, though without calling them cognitive. The classical
study was carried out by Konrad Lorenz and Niko Tinbergen (1939)
on the egg-retrieval response of the greylag goose, Anser anser (see
Figure 9.2). When a brooding goose sees an egg just outside the nest,
she is likely to rise, orient toward the egg, and then stretch her neck
toward it. If she can reach the egg, she places her bill behind it and
pulls it back to the nest in a very characteristic and stereotyped
manner. An observation of special significance in these experiments
was that the egg-retrieval movement itself continued quite normally
to completion even if the egg rolled away. The incongruous sight of a
goose retrieving an imaginary egg led Lorenz to suggest that the form
of the movement was not determined by stimuli from the
environment, but instead was programmed in the central nervous
system. He called it an Erbkoordination (inherited coordination), a
motor mechanism. Much of the behavior of the digger wasp is clearly
controlled by such mechanisms. There was considerable controversy
at the time about the properties of an Erbkoordination, but today it
is generally agreed that there are motor programs (mechanisms) in
the central nervous system and that they can be either inherited or
learned (e.g., piano playing: motor skill learning—Diedrichsen &
Kornysheva 2015).
Figure 9.2 Greylag goose retrieving an egg. (a) The goose sees the
egg outside the nest. (b) She gets up off the nest and approaches the
egg with outstretched neck. (c) She places the underside of her bill
over the egg, at which point the behavior pattern begins and
continues until (d), when the egg is on the nest rim. After Lorenz &
Tinbergen 1938.
Lorenz and Tinbergen also did experiments to determine what
stimulus characteristics of the egg were necessary to elicit the egg-
retrieval response and proposed the concept of an innate releasing
mechanism (a perceptual mechanism) on the basis of their results.
This idea was examined further in experiments by Baerends and his
colleagues on the egg-retrieval behavior of the herring gull (Larus
argentatus) and by Jörg-Peter Ewert and his colleagues on the prey-
catching behavior of the common toad (Bufo bufo). As we have seen
in Chapter 2, energy in the environment comes in many forms, but
an animal can only detect those forms of energy that it uses in its
everyday life. Sensory systems can differ greatly among species, both
with respect to the kinds of energy that can be detected and the
sensitivity of the receptor. But once detected, the basic organization
of sensory systems is generally very similar. The receptor cell
transforms the incoming energy into a neural impulse, which can
then activate a perceptual mechanism, either directly or through a
series of synapses. Unlike behavior patterns that can be directly
observed, however, perceptions occur within the brain and can only
be studied indirectly.
One way of studying the perceptual world of animals has been very
successful. The first step is to observe the objects or events that are
associated with particular behavior patterns in a relatively natural
environment. For instance, the natural prey of the common toad is a
worm or millipede. The second step in the analysis is to determine
what it is about the stimulus object that makes it effective. What
makes an object a prey item for a toad? Its size, color, shape,
movement, orientation? Ewert (1997) has done many experiments
investigating this question and they are described in detail in
Chapter 2. Another example is the work of Baerends (1982) and his
colleagues on egg retrieval. The natural object for egg retrieval is the
gull’s egg. What makes an object an egg for the gull? Herring gulls
nest in colonies and normally brood a clutch of three eggs. If the
gulls are alarmed while brooding, they fly off the nest but return as
soon as the potential danger has disappeared. In these experiments,
two experimenters walked into the colony, which caused the birds to
leave their nests. One nest was chosen for study and two of its three
eggs were removed. These eggs were replaced by two wooden egg
models placed on the rim of the nest. One of the experimenters then
hid in a small portable tent that had been set up near the chosen
nest, and the other experimenter left the colony. The nest owner
soon returned, and the experimenter in the tent could easily observe
and record the behavior of the returning bird. Typically, the gull
would enter the nest and sit on the one egg remaining. It would then
rise, look at the model eggs on the nest rim, retrieve one egg and
then the other. (The retrieval movement itself is similar to the egg-
retrieval movement of the greylag goose described above.). The egg
that was retrieved first was considered to be a better releasing
stimulus than the other egg. Thousands of such experiments were
performed using pairs of models that differed from each other in
shape, size, color, pattern of speckles, and various other attributes;
possible side preferences and other experimental details were also
controlled for.
One of the first insights gained from such experiments is that most
animals pay very little attention to much of the information available
to them, and what they respond to often depends on their
motivational state. This selectivity by animals led to the concept of a
sign stimulus. A sign stimulus is that part of the total stimulus
situation that is relatively most effective for releasing a response. In
the case of the gull, shape had relatively little effect on the releasing
value of the egg: round, square, oblong, and egg-shaped models were
retrieved with about equal frequency. Color, size, and speckle
pattern, on the other hand, were all very important. Green was the
most highly preferred color, larger models up to five times normal
size were always preferred over smaller ones, and speckled models
were always preferred over nonspeckled ones. Thus, green, a large
size, and speckles are three of the sign stimuli for egg retrieval in a
brooding herring gull. By varying several stimulus attributes of the
model eggs at the same time, it was possible to show that each of the
sign stimuli increased the releasing value of the model
independently. The case of the herring gull is especially interesting in
this respect. The normal egg is dark beige, speckled, and slightly
larger than a large chicken egg, whereas the most effective egg is
green, speckled, and almost as large as a football. This egg has been
called a supernormal stimulus because it is invariably chosen
over the gull’s own egg in a choice test (see Figure 9.3).
Figure 9.3 Supernormal stimulus. Herring gull choosing to
retrieve a large, green, speckled egg rather than its own smaller,
brown egg. Courtesy of Gerard Baerends.
The concept of innate releasing mechanism, as proposed by Lorenz
and Tinbergen, was soon found to be problematic in several respects.
One is the difficulty in defining the concept of innate (see Chapters 2
and 7), which refers to the development of the mechanism. Another
is the motivational implications of the word release. Herring gulls
not only retrieve eggs, but they also eat them, and Baerends found
that gulls prefer small, red eggs to eat. Thus, the motivational state of
the gull is one factor determining how it sees the world. Since the
releasing mechanism is considered to be a mental structure, it is
preferable to use structural terminology to refer to it: an object
recognition mechanism, a basic perceptual cognitive mechanism
(Hogan 2017). The idea of stimulus summation from independent
stimulus dimensions is still applicable, however, and will be seen
again in the discussion of human semantic representations below.

Central Mechanisms
Central cognitive mechanisms come in several forms, and I find it
convenient to distinguish among comparator mechanisms,
oscillator mechanisms, and representations. Comparator
mechanisms involve feedback and are basically expectations. An
extremely important exemplar is the reafference system and
efference copy. The principle has been explained by von Holst
(1954). He considers all stimulus input to an organism from all its
receptors to be afference, and information sent from the central
nervous system (a “command”) to motor mechanisms to be
efference. He then distinguishes between exafferent stimulation that
is produced by movement in the external world and reafferent
stimulation that is produced by an animal’s own movement. For
example, “if I shake the branch of a tree, various receptors of my skin
and joints produce a re-afference, but if I place my hand on a branch
shaken by the wind, [stimulation] of the same receptors produce[s]
an ex-afferance…. The same receptor can serve both the ex- and re-
afferance. The CNS [central nervous system], must, however, possess
the ability to distinguish one from the other” (p. 89). His solution
was the efference copy. The “command” leaves a copy of itself in the
motor mechanism, which is compared with reafference produced by
the animal’s movement. If the feedback from the movements of the
animal (reafference) matches the expectation of what the feedback
should be (the efference copy) the animal perceives the world as
stable and perceives itself as having moved. If there is no efference
copy (i.e., there is no expectation of movement) stimulus input
(afference) is perceived as the world moving. The concept of
efference copy has been experimentally shown to be applicable to
animals from insects to humans and the comparison principle, or
expectation, is important in most fields of behavior studies.
Oscillators or central pattern generators are another class of central
behavior mechanisms. At a neural level, oscillators are neurons or
groups of neurons (neural systems) that fire in a rhythmic pattern in
the absence of any rhythmic input. Oscillators are also called
pacemakers. Oscillators can be found throughout the brain and are
generally thought to be responsible for rhythmic activity in many
systems. Their rhythms range from milliseconds to days and perhaps
even years. It was von Holst (1937) who showed that fin movements
in fish are controlled by central pattern generators and D. Wilson
(1966) postulated oscillators as controlling factors in insect
locomotion. Lorenz’ (1937) concept of Erbkoordination also invokes
the operation of central pattern generators in many types of actions.
Gallistel (1980) provides a general discussion of the role of
oscillators in the organization of action. Beyond their role in
determining the form of behavior patterns, oscillators are also
involved in the timing of most behavioral events. They are important
for our perception of time and play an essential role in biological
rhythms as discussed in Chapter 4. As with most basic comparator
mechanisms, so also with oscillators, both are central behavior
mechanisms that generally operate in the background of
consciousness or awareness.
Representations are the mental (neural) correlates of objects, ideas,
concepts, memories, feelings, and the like. Because these entities are
difficult to specify empirically, there have been many different
approaches to studying them. We have seen one method of studying
the representation of an object (an egg for a gull). But there are many
approaches, often specific to the kind of representation being studied
and the kind of question being asked. We will look at some studies of
memory and concepts.

Memory
Memory is the representation of past experience and can be
considered a process leading to the formation of behavior
mechanisms (representations) as well as the outcome of the process.
Cognitive psychologists who study memory are interested in the
relations among the acquisition, retention, and retrieval mechanisms
of memory. These are actually three different kinds of questions
about memory: acquisition is a problem of development or change in
structure; retention is a problem of structure—how (in what form)
and where are memories stored; and retrieval is a problem of
motivation. Psychologists have also noted that there are different
kinds of memory, though there is no consensus as to what those
kinds are. That is because distinctions among kinds of memory
depend on which aspect of memory a researcher is considering: Are
there different modes of acquisition? Are there different forms of
memory, and different locations in which memories are stored? Are
there different kinds of retrieval? Do memories serve different
functions? One distinction, discussed in detail in Chapter 8, is
between working memory (short-term) and reference memory (long-
term). Another distinction, made by Endel Tulving (1972), is between
semantic memory and episodic memory, two forms of reference
memory. Semantic memory, sometimes called conceptual
knowledge, is the aspect of memory that corresponds to general
knowledge of objects, word meaning, and facts without connection to
any particular time or place. Episodic memory, as the name suggests,
is memory for specific events in time and place. Work on these forms
of memory has concentrated on humans, but the results are
applicable to much work on memory in other animals.
In 2007, Karalyn Patterson, Peter Nestor, and Timothy Rogers
published a paper entitled “Where do you know what you know? The
representation of semantic knowledge in the human brain.”
Conceiving of how an object, an egg for example, might be
represented in the brain is a problem because any object has many
attributes, and representations of those attributes (color, shape, size,
etc.) must be located in different appropriate locations. For any
particular object (or idea or concept or word) it is not too difficult to
imagine how it might be represented, but all objects share some
attributes, and representations interact with each other. Patterson et
al. proposed the scheme depicted in Figure 9.4 as a solution to the
problem: the hub-and-spoke model. The model is based on decades
of research with brain-damaged patients suffering from semantic
dementia, neuroimaging studies on normal individuals, and
computational modeling. The details of this model are beyond the
purview of this chapter, but one can see that the representation of an
object is widely distributed throughout the brain (the spokes), but
that the representation comes together in the anterior temporal lobe
(the hub). Representations of different objects, ideas, concepts, etc.
“mingle” in the hub and in that way can interact with each other. The
scheme of visual processing presented in Figure 2.11 is, in principle,
a simplified version of the hub-and-spoke model. I later discuss how
this model relates to representations in other animals.
Figure 9.4 (a) Computational framework for the hub-and-spoke
model. (b) A neuroanatomical sketch of the location of the hub and
spokes is presented. The hub is located within the anterior temporal
lobe (ATL) region, whereas the modality-specific spokes are
distributed across different neocortical regions (the same color
coding is used as for the computational model). Each spoke
communicates bidirectionally with the ATL hub through short- and
long-range white-matter connections (arrows). From Ralph et al.
(2017).
The concept of episodic memory has been a source of controversy
among many cognitive scientists. Tulving (1983) believed that
episodic memory was characteristic of humans and was not present
in other species. As we have seen in Chapter 8 (Figure 8.5), however,
scrub jays (Aphelocoma californica) also exhibit episodic-like
memory (Clayton & Dickinson 1998) as do other species. The digger
wasp female inspects one of her three active nests, spends the night
resting in the fallen needles of a pine tree some distance from the
nest, and returns the next morning with a caterpillar to provision the
nest she inspected the night before. Unsurprisingly, there is still
debate about whether similar behavior implies equivalent cognitive
mechanisms (Clayton 2015). This question is explored further later
in this chapter.
Neuroscientists are also actively studying memory: “Memories are
thought to be encoded as enduring physical changes in the brain, or
engrams” (Josselyn et al. 2015, p. 521). A cartoon of the development
of a memory (engram) is shown in Figure 9.5. There has been
considerable progress in recent years in unravelling the neural
mechanisms involved in these structural changes, and Josselyn et al.
(2015, 2017) review the neurophysiological evidence for these
mechanisms. Of special note is the fact that the neural ensembles
that encode the memory (engram) are widely distributed in the
brain, which is in accord with the hub-and-spoke model of memory
above. A behavioral example of the consolidation and
reconsolidation processes in engram development is the
development of the partner recognition mechanism (sexual
imprinting) in zebra finches discussed in Chapter 7.
Figure 9.5 The lifetime of an engram. The formation of an engram
(encoding) involves strengthening of connections between
collections of neurons (neuronal ensembles) that are active (red)
during an event. Consolidation further strengthens the connections
between these neurons, which increases the likelihood that the same
activity pattern can be recreated at a later time, allowing for
successful memory retrieval. During consolidation, the engram
enters a mainly dormant state. Memory retrieval returns the engram
back to an active state and transiently destabilizes this pattern of
connections. The engram may be restabilized through a process of
reconsolidation and re-enter a more dormant state. Therefore, an
engram may exist in a dormant state between the active process of
encoding and retrieval required to form and recover the memory. In
this way, an engram is not yet a memory, but provides the necessary
conditions for a memory to emerge. From Josselyn et al. (2015)

Concepts
The American Heritage Dictionary defines a concept as “a general
idea or understanding, especially one derived from specific instances
or occurrences.” This is a rather vague definition, but in most cases,
it refers to an abstract feature common to a group of objects that can
differ from each other in many or all of their other features. “Red” is
a concept. The sky can be red, a flower can be red, a fire engine can
be red, the belly of a male stickleback can be red. The hub-and-spoke
model of semantic knowledge discussed above can provide a
framework for understanding a concept as a mental structure:
overlapping spokes. Concepts and concept formation in humans
have been studied in experimental psychology for many years (e.g.,
Bruner et al. 1956), but Chittka & Niven (2009) have argued that
insects can also form concepts such as same or different.
Recently, Avargues-Weber et al. (2012) have provided evidence that
honeybees are even able to learn two abstract concepts
simultaneously: a spatial relationship (above/below vs right/left)
and a same/different relationship. The bees were trained in a Y-maze
with stimuli presented vertically on the back wall of each arm. A
small nipple in the center of each wall contained either a drop of
sucrose solution (a reward) or a drop of quinine (an aversive
substance). The correct stimulus always had two components that
could be either the same or different (in color and pattern) and could
be arranged either one above the other or next to each other. Bees
were trained on the spatial arrangement and reached a level of 80%
correct in 30 trials. During training, all the stimulus components
were different. On transfer tests after training, the bees chose the
stimulus with different components when the spatial arrangement
no longer provided discriminant information. The authors
concluded: “…bees learned that they had to choose stimuli arranged
in a specific spatial relationship and that all stimuli were composed
of different visual elements” (p. 7484).

Using Mental Structures: Using

Knowledge

Orientation
Orientation is generally considered to be a change of an organism’s
position in space as a response to an external stimulus. It is therefore
one category of the consequences of activating behavior mechanisms.
Orientation reactions range from simple reflexes to highly integrated
sequences of behavior. Three types of orientation reactions are
considered here: taxes, navigation, and migration.

Taxes
Taxes are movements defined by the fact that their form is
continuously determined by stimuli from the external environment.
When the greylag goose is retrieving her egg, she continuously makes
sideways movements of her bill that keep the egg from rolling away.
However, if the egg does roll away, the sideways movements stop
even though the retrieval movement itself continues (Figure 9.2).
Most behavior patterns have a taxis component, and there is also a
large older literature on taxes in various invertebrates. One of the
most influential figures in this field was the German/American
physiologist Jacques Loeb (1918), who was active at the end of the
nineteenth and the beginning of the twentieth centuries. He led a
crusade against the anthropomorphic and teleological explanations
of invertebrate behavior then current in the writings of men such as
Romanes (1883) and attempted to understand all behavior in
physical and chemical terms. Although many of his ideas were
oversimplified or incorrect, his work stimulated a great deal of
experimental work on the behavior of invertebrates, much of it
summarized in the book by Fraenkel & Gunn (1940/1961).
Movement toward, away from, or with a fixed relation to a stimulus
is probably the simplest form of orientation. The behavior
mechanisms responsible for the movement, however, can range from
simple to highly complex. The negative klinotaxis to light in the
blowfly maggot (Calliphora erythrocephala), for example, depends
on a comparison of the light intensities on the two sides of its
foreregion. As it moves, it slowly waves its foreregion from side to
side. If the light intensity on one side is higher than on the other, it
turns away from the light (see Tinbergen 1951, pp. 92–93). Much
more complex is the rheotaxis of the larval zebrafish (Danio rerio).
Most aquatic species orient toward a current and swim to hold their
position. This rheotaxic behavior is known to involve the lateral line
and visual systems. Olive et al. (2016) developed an assay to study
freely swimming larvae and were able to show that there is a clear
transition from exploratory to counterflow swimming. By changing
the sensory modalities accessible to the fish (visual only, lateral line
only, or both) and comparing the swim patterns at different ages,
they were able to detect and characterize two different mechanisms
for position holding, one mediated by the lateral line and one
mediated by the visual system. When both sensory modalities were
accessible, the visual system overshadowed the lateral line. Oteiza et
al. (2017), in other experiments, showed that, in the absence of visual
cues, larval zebrafish perform rheotaxis by using flow velocity
gradients as navigational cues. The fish use their mechanosensory
lateral line to first sense the curl (or vorticity) of the local velocity
vector field to detect the presence of flow and, second, to measure its
temporal change after swim bouts to deduce flow direction. Their
results reveal an elegant navigational strategy that generalizes to a
wide range of animal behaviors in moving fluids. Florian Engert and
his group are currently investigating the neural circuits involved in
this behavior.

Navigation
Our ability to easily find our way about in space raises many
additional questions about the mechanisms of orientation. In
everyday life, we subjectively take a stable representation of space for
granted. But how the nervous system might accomplish such a
representation poses huge problems. Further, other species from
insects to birds and other mammals also behave in ways that support
the idea that they too have a stable representation of space. Three
lines of evidence speak to this issue: spatial memory, path
integration, and cognitive maps.

Spatial Memory
Place memory, that is, remembering where something is located, is a
basic requirement for view-based navigation. There is abundant
evidence that species from insects to humans use view-based
navigation, but what is actually being remembered? In early studies
of digger wasps, Tinbergen (1932/1972) and Baerends (1941) saw
that the wasps made an orientation flight above the nest before
departing on a foraging trip. Experiments using displacement of
landmark arrays near the nest (see Figure 17.4) showed that the
wasps relied on visual local landmarks to relocate their nests. Black-
capped chickadees (Parus atricapillus) store seeds and invertebrate
prey in concealed locations scattered throughout their home range.
Experiments, similar to the wasp studies, showed that the birds rely
on visual information from nearby landmarks to locate concealed
caches. The appearance of the cache sites themselves seemed to be
relatively unimportant in cache retrieval (Sherry & Duff 1996).
Sherry and Vaccarino (1989), in a laboratory study, removed
(aspirated) the hippocampus of black-capped chickadees after the
birds had had an opportunity to cache sunflower seeds and found
that recovery of stored seeds was reduced to the chance rate, even
though the rate of searching behavior was not affected. The results of
a second experiment showed that both memory for places and
working memory were disrupted by hippocampal damage, and that
both these memory capacities were essential for cache recovery.
Spatial memory has been studied in a wide variety of species, using
many different tasks. Richard Morris (1981) observed rats swimming
in a circular pool of water. A platform, on which the animals could
remain dry, was submerged just below the surface. The platform was
not visible due to milk being added to the water. The rats quickly
learned the location of the platform. By altering the location of the
platform, it was determined that the rats were using distal cues in the
room as landmarks. Interestingly, rats that were tested when the
platform was not submerged and was visible did not learn
substantially faster. Ken Cheng (1986) observed rats in a large
rectangular sandbox. The rats learned that food was buried in a
particular location and then were tested when no food was present.
Analysis of the errors made by the rats while learning and when
being tested with no food present showed that the animals were
encoding the location of the food in relation to the geometry of the
enclosure, local cues. What cues are used when both distal and local
cues are available varies with the species and situation.
How the landmark memories are used is another question. In a study
of landmark learning in bees, Cartwright & Collett (1982) suggested
that bees trained to forage at a place specified by landmarks do not
construct a Cartesian map of the arrangement of landmarks at the
food source. Instead, they proposed that the bees store something
like a two-dimensional snapshot of their surroundings at the food
source. To return there, the bees move so as to reduce discrepancies
between the snapshot and their current retinal image. Some
variation of image matching is likely to be a general mechanism used
in spatial navigation. Gallistel (1990), Shettleworth (2010), and
Collett et al. (2013) discuss these issues in detail.

Path Integration
“It is often assumed that navigation implies the use, by animals, of
landmarks indicating the location of the goal. However, many
animals (including humans), are able to return to the starting point
of a journey, or to other goal sites, by relying on self-motion cues
only. This process is known as path integration, and it allows an
agent to calculate a route without making use of landmarks” (Etienne
& Jeffery 2004, p. 180). The term “path integration” was first used by
Mittelstaedt & Mittelstaedt (1980) with respect to pup retrieval in
the gerbil (Meriones unguiculatus), but the principle was already
suggested by Darwin (1873) with respect to humans and derived
from “dead reckoning” as used by mariners in navigating across open
seas with no visible landmarks. The path integration mechanism is
often thought to be a neural accumulator that continuously monitors
directional cues and distances traveled and that integrates them so
that the animal always “knows” the direction from its current
position to its home or other starting position. Such a mechanism
implies the existence of some sort of neural representation of space
(Gallistel 1990; Gallistel & Matzel 2013).
There have been many experiments studying navigation of insects,
especially ants and honeybees, and many of these experiments have
demonstrated the use of path integration. Karl Von Frisch (1955), for
example, showed that honeybees (Apis mellifera) communicate the
distance and direction of a food source to other bees by means of the
“wagging dance” (see Chapter 14). In one experiment, he
demonstrated that the subject bee communicated the straight-line
distance from the hive to the food source even though the subject bee
had only visited the food source by detouring around a large rock and
never by flying directly herself; this information could only have
been acquired through path integration. Other experiments by
Wehner and his colleagues (e.g., Wehner & Menzel 1990) on the
desert ant (Cataglyphis fortis) also demonstrate the use of path
integration in navigation. Collett and Collett (2000) reviewed the
results from desert ants and honeybees and discuss three models of
how an accumulator might work. They also provide new
experimental results that show that insects can use both path
integration and landmark navigation, and which is employed
depends on conditions in the specific case.
As mentioned above, path integration is also used by birds and
mammals. In a series of very careful experiments, Ariane Etienne
and her colleagues (1996) demonstrated that foraging golden
hamsters (Mesocricetus auratus) can return to their home using
only information derived from path integration. Her results also
showed that the return home need not even be direct. Nonetheless,
under normal environmental conditions, path integration
information is combined with landmark information in navigation.

Cognitive Maps
Tolman (1948) introduced the term “cognitive map” into psychology
to interpret the results of his experiments on maze learning in rats.
He suggested “that in the course of learning, something like a field
map of the environment gets established in the rat’s brain” (p. 192).
Such a map would be an internal representation of the geometric
relations among noticeable points in the animal’s environment.
Evidence for such a map came from a number of sources, but
especially from observations that, when physical barriers were
removed, a rat would run directly to the source of food. Thirty years
later, O’Keefe and Nadel (1978) presented neurophysiological
evidence for such a view in their book, The Hippocampus as a
Cognitive Map.
Since then, there has been an explosion of experimental work, both
behavioral and neurophysiological, exploring the characteristics of
such maps. It soon became apparent, however, that investigators had
differing views of what was meant by the concept of a cognitive map
and whether insects, for example, did or did not have such a map.
Sara Shettleworth (2010, pp. 296–310) provides a thoughtful and
comprehensive review of the behavioral evidence for and against the
idea of a cognitive map. She concludes that “there is little if any
unambiguous evidence that any creature gets around using a
representation that corresponds to an overall metric survey map of
its environment” and that mapping-like behavior is “better explained
by reference to what cues the animals are actually using, how they
are using them, and how they come to do so than to the ill-defined
notion of a cognitive map.”
The reality of the hypothesized existence of a cognitive map has been
supported by the recent discovery of spatial grid cells in the
entorhinal cortex (a neighboring structure of, and with direct
connections to, the hippocampus) of the rat that encode both
distance and direction information independent of the animal’s
immediate location (Moser & Moser 2016). Together with
information from “place cells” in the hippocampus, this system could
be used by the animal to find its way about in space. (John O’Keefe,
May-Britt Moser, and Edvard Moser were awarded the Nobel Prize
for physiology in 2014 for this work.) However, as Redish (2001)
pointed out, there are different kinds of maps: a local map of one’s
surroundings; a map of the city in which one resides; a map of the
world. How are these maps related? And how is space represented in
animals such as birds, fish, insects, and cephalopods that do not have
the same brain structures as mammals? Humans have an internal
model of the outside world, but whether “cognitive map” is the best
metaphor for how the outside world is represented is moot. It is
perhaps better to follow Shettleworth’s advice and look for the
mechanisms each species is actually using to find its way about in
space.

Migration
Migration refers to movement from one place to another, usually in
groups, and often seasonally. The cognitive mechanisms used by
human migrants from Afghanistan to Europe are clearly different
from those used by wildebeest moving across the African plains or
from birds migrating from the Arctic to the Antarctic. In this chapter,
I will only be considering some aspects of bird migration.
The phenomenon of bird migration has been known since ancient
times, although explanations of it have radically changed since then
(Wiltschko & Wiltschko 2003). Bird migration can vary in distance
from a few hundred to several thousand kilometers. Many species
interrupt their migratory flights in places where they can rest and
refuel, but at least one species, the bar-tailed godwit (Limosa
lapponica), has been tracked making an 11,000 km-long nonstop
flight from Alaska to New Zealand and eastern Australia (Battley et
al. 2012). However, regardless of distance, there are three questions
all migrants face with respect to orientation: In which direction do I
go? How far do I go? How do I know I have arrived in the right
place? Another question, “When do I go?”, is a question of
motivation and is not considered in detail here. Most of these
questions have been investigated using birds kept in aviaries,
although some studies have observed free-flying birds.
“In which direction do I go?” is the aspect of migration that has been
most studied. During the migratory season, most aviary birds show
an increase in activity (Zugunruhe or migratory restlessness). This
Zugunruhe has been shown to be due to physiological changes
controlled by the photoperiod and a circannual rhythm (Gwinner
1986, 1996). Early experiments by Kramer (1952) on starlings
(Sturnus vulgarus) showed that the birds were able to orient their
flights using a sun-compass (see Chapter 4), but since most birds,
including starlings, migrate at night, later experiments have
concentrated on other means of orientation. Emlen (1967) tested
captive indigo buntings (Passerina cyanea) for celestial orientation
under the natural night sky and inside a planetarium and showed
that the birds displayed a consistent tendency to orient in the
direction appropriate for the migration season in question (i.e., fall
or spring migration). Outside periods of Zugunruhe, the birds
oriented in random directions. Other experiments by the Wiltschkos
and their colleagues (review in 2003) provide convincing evidence
that most birds can also use the earth’s magnetic fields for
orientation.
The sun-compass, the starry sky, and magnetic fields are all tools
that birds can use for orientation, but they do not actually answer the
question of which direction is correct. Emlen (1969) provided
evidence that changes in the internal physiological state of the bird
are responsible for the seasonal reversal of preferred migration
direction, and experiments with birds of various species that were
migrating for the first time, demonstrate that they set off in the
correct direction. These results all suggest that migratory direction
can develop prefunctionally, but social interaction of some sort with
experienced conspecifics is probably the determining factor in most
cases.
“How far should I fly?” is a question that has not been much studied,
but there are some data that suggest possible answers. Berthold
(1973) found, in various species of migratory warblers that he and his
colleagues were studying, that the amount of Zugunruhe shown by
aviary birds during the migratory period correlated highly with the
distance that each species flew on its migratory trips. Gwinner (1986,
1996) proposed that a circannual rhythm produced timing programs
that could play a major role in determining migratory distance.
Another possible mechanism is related to the magnitude of the
energy stores built up prior to migration (see Piersma & Van Gils
2011). The bird would fly until its store was used up. In experienced
birds, memory of previous flights would be sufficient to determine
distance.
“How do I know I have arrived?” has also not been extensively
addressed experimentally. One study of habitat preference in the
dark-eyed junco (Junco hyemalis) suggests a possible general
mechanism. Roberts and Weigl (1984) recorded the time juncos
spent in front of pictures of their summer and winter habitats.
Juncos kept on winter or summer photoperiods preferred winter or
summer habitat pictures, respectively. Insofar as habitat preference
is based on visual features of the environment, birds would know
that they have arrived when the actual habitat matches their mental
picture of the appropriate habitat. An experiment with red knots
(Calidris canutus), a long-distance migratory shorebird, has
provided suggestive evidence: the birds’ visual preference for
projected habitat pictures of wintering mudflats and breeding tundra
did change as the birds went from winter to summer physiological
condition in the lab (Kok et al. 2020).

Social Learning
Social learning refers to any learning that is influenced by interaction
with, or observation of, another animal or its products (Heyes 1994).
What makes it different from the types of learning discussed in
Chapter 8? With respect to the mechanisms of learning, probably
nothing! Nonetheless, it has been studied as a separate field for
many years for at least two major reasons. The first is that many
investigators have been interested in comparing human and
nonhuman intelligence: the presence or absence of social learning in
various species might make it possible to differentiate humans from
nonhumans. The second reason is that social learning provides a
mechanism for transmission of behavior between generations and
could thus provide insight into the evolution of behavior. I will
present some examples of social learning and then discuss possible
mechanisms.

Observational Learning
When a broody hen, in the presence of young chicks, finds a morsel
of food such as a mealworm, she picks it up and makes a “food call.”
This attracts the chicks. She then drops the food and one of them
picks it up and swallows it. Subsequently, all the chicks will show a
preference for similar foods (Kruijt 1964). Young kittens provide a
somewhat similar example. Young kittens are often seen “playing”
with small moveable objects, which is also seen with a live mouse.
When the mother cat joins the kittens and catches the mouse, opens
it up, and eats it, the kittens subsequently treat mice as food (Berry
1908). Another more complex example concerns foraging behavior in
pigeons (Columbia livia) and doves (Zenaida macroura). There have
been several reports that pigeons and doves can perform relatively
complex novel food-finding behavior after observing the actions of
an experienced conspecific. The most convincing evidence comes
from a well-designed series of experiments by Palameta and Lefebvre
(1985). These investigators were able to show that a pigeon, merely
by observing another pigeon performing a learned response for food,
can learn both where to direct its feeding behavior and what motor
act to use. The experimental set-up used in their experiments is
shown in Figure 9.6.
Figure 9.6 Observational learning in doves. The dove on the left is
watching the dove on the right lift a cover off a source of food.
Courtesy of Louis Lefebvre.
In a review of studies of social learning in insects, Elli Leadbeater &
Lars Chittka (2007) concluded that there are many well-
substantiated cases of observational learning. In a recent example,
Leadbeater (2015) studied observational learning in bumblebees
(Bombus terrestris). In the experiment, “subject” bumblebees were
initially trained to associate conspecific presence with either a
sucrose reward or an aversive substance. They were then permitted
to watch conspecific “demonstrators” choosing a particular color of
an artificial flower through a screen. The set-up was analogous to
that in Figure 9.5, adapted for bees, of course. When later allowed to
forage alone, bees that had learned to associate conspecifics with
sucrose preferred the flower color the demonstrators chose, while the
aversive group actively avoided the same colors. A control group of
naïve bees was not influenced by the demonstrators’ choices. In
another study (Alem et al. 2016), investigators trained bumblebees to
obtain a sucrose reward from an artificial flower lying under a
transparent cover using a string-pulling task (see Figure 9.7).
Observer bees were first allowed to obtain the sucrose reward from
exposed flowers. They then were allowed, individually in a small
transparent chamber nearby, to watch a demonstrator bee retrieve
the flower and obtain the reward. After the demonstrator bee had
been removed and the flower replenished, the observer bee was
allowed to enter the flight arena. Sixty percent (15 of 25) of the
observer bees tested managed to pull the string (a tool) and obtain
the reward on the first trial after having observed the demonstration.
Control bees, without a demonstration, never solved the problem.

Figure 9.7 Bumblebee using a string to gain access to an artificial

flower. From Alem et al. 2017.
Monkey see, monkey do. Even folk wisdom attributes observational
learning to primates, although it was not expected that the primates
would be using tools and passing on traditions to subsequent
generations. Use of tools by nonhuman primates became a focus of
scientific interest after Jane Goodall published her observations of
wild chimpanzees pushing sticks into the nests of ants and termites
and eating the insects that crawl onto the stick. Observations of
infants with their mothers suggested that the young chimps imitate
the older animals in this behavior (Van Lawick-goodall 1970).
Another example of social influences on tool use is provided by
studies of capuchin monkeys. Of special interest are two groups of
Sapajus libidinosus that have been observed in the wild in Serra da
Capivara National Park in northeastern Brazil. These two groups
manufacture tools to probe for lizards by trimming branches and
thinning their tips and use pounding stones for breaking and/or
enlarging holes in tree trunks or rocks, as well as for cracking nuts
(Falótico & Ottoni 2014, 2016; see Figure 9.8); the use of pounding
stones to open cashew nuts goes back at least 600 years (Haslam et
al. 2016). These behaviors have not been observed in other groups of
capuchins. The young monkeys have been seen observing older
monkeys performing these behaviors, but it takes several years of
practice before the young monkeys become proficient. Resende et al.
(2014) have studied how the young monkeys learn to crack nuts. An
example of observational learning in humans (Figure 17.8), as well as
cultural transmission, is discussed in Chapter 17.

Figure 9.8 Young capuchin monkey using a stick to probe for

lizards in Serra da Capivara National Park in northeasten Brazil.
Courtesy of Tiago Falótico.
All the examples of observational learning discussed above might be
described by the word imitation. “If one can from an act witnessed
learn to do the act…he imitates” (Thorndike 1898, p. 50). Thorndike,
for his PhD thesis, was studying the “associative processes in
animals” and all the examples above can be understood using the
concepts of associative learning as presented in Chapter 8. Some of
the motor patterns used by the bees and monkeys need to be
practiced before the animals become proficient, but no new cognitive
mechanisms need to be invoked. Nonetheless, scientists have been
arguing for more than a century about what “imitation” means and
whether particular animals can imitate. Imitation is a functional
concept and can include different mechanisms. Associative learning
mechanisms were sufficient to explain the examples here, but there
may be cases where other mechanisms need to be invoked. Excellent
discussion of this issue can be found in articles by Jeff Galef (1988,
2013) and in the book by Sara Shettleworth (2010).

Perceptual Learning
As we have seen in Chapter 7, social interaction is essential for the
development of species’ recognition mechanisms through filial and
sexual imprinting. It is also essential for the development of a
template (perceptual mechanism) in bird song learning: the bird
stores a memory of the song it heard while very young that it uses
when learning to sing the song itself later. Social influences are in
fact ubiquitous in many aspects of vocal development (Snowdon &
Hausberger 1997). An important example is the development of
alarm calling in vervet monkeys (Cercopithecus aethiops, now
Chlorocebus aethiops) living in Amboseli National Park, Kenya.
Struhsaker (1967) noted that these monkeys gave acoustically
different alarm calls in response to different types of predators and
that each call evoked a different and seemingly adaptive response.
His suggestions were followed up in a 14-month field study by
Seyfarth et al. (1980) who observed the responses of vervet moneys
to predators in natural encounters and to playbacks of recorded
alarm calls. Their results confirmed the observations of Struhsaker:
“animals on the ground respond to leopard alarms by running into
trees, to eagle alarms by looking up, and to snake alarms by looking
down” (p. 1070).
These findings started a conversation about the problem of meaning
in alarm calls (Macedonia & Evans 1993). Do nonhuman animal
vocalizations have “functional reference”: do the callers only give the
leopard alarm when they spot a leopard, or do they give the leopard
alarm when they spot any ground predator or nonpredator; do the
receivers always behave appropriately when they hear the alarm for a
predator they cannot see? Do the monkeys have representations of
the predator analogous to human representations? The question of
meaning is logically an aspect of the search for the adequate (sign)
stimulus. Subsequent studies of the development of alarm calling
and responses to alarm calls in young vervet monkeys from 3 to 7
months of age found that both the motor mechanism of the sender
and the perceptual mechanism of the receiver require functional
experience (i.e., learning) in order to attain adult form (Seyfarth &
Cheney 1997). Their results suggest that the alarm calls of these
monkeys do indeed have functional reference.

“Higher” Mental Processes

Portia fimbriata is a species of jumping spider that eats other
spiders. Robert Jackson has studied its hunting tactics for many
years (review in Jackson & Cross 2011). When hunting other spiders,
Portia uses mimicry, detours, or deception depending on the prey.
These tactics are often modified by trial-and-error learning, but
perhaps Portia’s most striking talent is its apparent ability to plan an
attack. An example is its pursuit of Argiope appensa, a spider that
builds orb webs on tree trunks. “Portia often walks up the tree trunk
toward A. appensa and then stops, looks around, goes off in a
different direction and reappears above the web. If there is a vine
over the web, for example, Portia seems to look at the web, the vine
and the neighboring vegetation before moving away, perhaps going
to where the web is completely out of view, crossing the vegetation
and coming out on the vine above the web. From above the web,
Portia drops on a silk line alongside but not touching the intended
victim’s web. Then, when parallel with the spider in the web, Portia
swings in to make a kill” (Jackson & Wilcox 1998, p. 393).
A recent experiment takes the behavior of Portia from anecdote to
established fact. The apparatus and procedure are described in
Figure 9.9. Jumping spiders have excellent eyesight, and each trial of
the experiment began with the test spider on the top of the tower
from which it could view the two goalboxes with lures and the two
pathways that led to the goalboxes. The test spiders had all been
raised individually and had had no experience with any situation
similar to the experimental set-up. Further details are in the legend
to Figure 9.9. Of the spiders that completed the trial, 251 chose the
correct pathway and only 15 chose the incorrect pathway (Cross &
Jackson 2016). Cross (2020) points out that no associative
mechanisms are needed to explain the spiders’ behavior; she also
discusses ideas of animal intelligence.

Figure 9.9 (a) Female Portia fimbriata. This is the posture Portia
adopts when stalking other jumping spiders. Courtesy of Robert
Jackson. (b) Example of apparatus used in detour-choice
experiments. Trial began with a test spider walking out of the pit and
on top of the tower where it could view two boxes. One box contained
four lures made from Oecobius amboseli (another species of spider)
and the other box contained lures made from four green-leaf pieces.
Which box contained prey was determined at random. After the test
spider left the pit and walked down from the top of the tower, all of
the lures were removed from the apparatus. To complete a successful
trial, the test spider chose a walkway after it left the tower and then
walked across the platform. The thick arrows indicate the path the
test spider took from the tower to the beginning of the correct
walkway and then to the end of that walkway. The apparatus sat in a
shallow pan filled with water (not shown). Drawing modified from
Cross and Jackson (2016)
Planning is sometimes characterized as “thinking before acting,” and
Portia’s hunting behavior and Ammophila’s provisioning behavior fit
that definition, as does going to the grocery store with a shopping
list. But thinking ranges from simple to complex, and many scientists
are interested in studying more complex thinking in animals.
Güntürkün and Bugynar (2016) reviewed studies on bird and
primate cognition looking at these cognitive “skills”: object
permanence, delay of gratification, mental time travel, reasoning,
meta-cognition, mirror self-recognition, theory of mind, and vocal
learning. The purpose of the review was to show that many birds,
primarily corvids and parrots, have these skills at the same level as
many primates, especially chimpanzees and apes. Such a result
supports the hypothesis that mammal and bird cognition (brain
structures) evolved independently and that birds and mammals have
evolved similar cognitive abilities convergently. Implications of this
conclusion are considered below. But first I want to consider a study
investigating learning and cognitive processes in rats using advanced
neurophysiological methods as well as behavioral observation.
Steiner and Redish (2014) studied “regret” in rats using the
Restaurant Row apparatus seen in Figure 9.10. The economic
foraging task required a rat to run counterclockwise around the loop,
making stay or skip decisions as it passed each spoke (restaurant).
Upon entering the zone around each spoke, the rat encountered
different offers of delays ranging from 1 to 30 seconds, indicated by
the frequency of the tone. If the rat stayed the allotted time, it
received a flavored pellet from a feeder at the end of the spoke. If it
left the zone early, the offer was rescinded, and the rat moved on to
the next zone. Each restaurant offered a different flavor, and the rat
had a total of 60 minutes in each session. Because the session was
time-limited, the decision to stay or skip a zone was not independent
of the other zones: waiting at one zone was time that could have been
spent at another zone. An economically maximizing rat should
distribute its time among the offers, waiting for valuable offers but
skipping expensive offers. Assuming that an animal likes some
flavors more than others, the economic value of an offer should
depend on the delay offered and the animal’s preferences.
Figure 9.10 Neural encoding of regret in orbitofrontal cortex
(OFC) and ventral striatum in rats performing the Restaurant Row
task. (a) Rats run counterclockwise in an economic foraging task in
which there were four reward zones (restaurants). If rats waited long
enough in each zone, a flavored pellet was delivered at the end of the
zone’s arm. The length of the wait time was signaled by the frequency
of a tone presented on entering that zone. Two zones, banana and
cherry, are shown in this example. The rat likes both of these flavors.
The plots at the bottom represent single cell activity seen in OFC
during performance of this task. In this example, activity is high
when the rat considers cherry and low when it considers banana.
Thus, this neuron is selective for cherry. Other neurons are selective
for other flavors. (b) The rat skips the cherry zone because the delay
to reward exceeds the rat’s threshold of willingness to wait for
cherry. However, once the rat gets to the banana zone, the new delay
for banana far exceeds the rat’s willingness to wait for banana. At
that instant the rat looks back with regret. During this look back,
activity in OFC and ventral striatum is similar to that observed
previously when the rat considered the cherry zone. From Bissonette
et al. (2014) (after Steiner & Redish 2014).
There are clearly many decisions the rat must make in order to
maximize receiving the most preferred pellets, but I will only
consider one here. “Regret” entails recognition that an alternative
action would have produced a more valued outcome. In humans, the
orbitofrontal cortex (OFC) is active during expressions of regret, and
humans with damage to the OFC do not express regret. In rats and
nonhuman primates, both the OFC and the ventral striatum have
been implicated in reward computations. In this study, neural
ensembles from OFC and ventral striatum were recorded as the rat
traversed the Restaurant Row. Some results are shown in Figure
9.10. The results described give strong support to the supposition
that the rat is engaged in deliberative thinking and is perhaps feeling
“regret”; but whether the rat “feels” anything and how such a feeling
compares to a human feeling remain matters of conjecture.

Mental Structure and Mental Processes

Cognition is knowledge, and knowledge resides in the mental

structures of the mind (brain). These structures include motor
programs; sensory systems; representations of objects, concepts
(including words), thoughts, and memories; comparator
mechanisms; path integration mechanisms; and the like. These are
the basic cognitive structures common to all animals and there are
various methods for studying them. But the structures themselves
differ among species in complexity and scope. Consider an object
recognition mechanism. We have discovered most of the features
(sign stimuli) that define the representation of an egg for a herring
gull. What are the features that define the representation of an egg
for a human? A moment’s thought reveals that a human’s
representation of an egg has many more features, but which are
important depends on the context: size, shape, color; a chicken egg, a
duck egg, an ostrich egg; a human egg, a frog’s egg, a digger wasp’s
egg; a fresh egg, an old egg, a rotten egg. The herring gull has at least
two conceptions of an egg—to be retrieved, to be eaten—but we
humans have many, many more. In the context of semantic
cognition, the hub-and-spoke model allows for complexity in the
number and range of connections among different regions of the
brain (the spokes) and the integration of these representations in the
hub. A rotten egg can refer to a person, as can a bad apple.
What accounts for cognitive differences among species? One
suggestion is that cognitive abilities and intelligence are positively
correlated with brain size. Within mammalian species, this may be
true (Sol et al. 2008). However, other dimensions of brains are as
important, if not more so. I noted above that birds, until fairly
recently, had been considered cognitively inferior to mammals
because of their small brains, “birdbrains,” but Güntürkün (2020)
has discussed how brain architecture in birds and mammals is being
realized as more similar than previously thought and that bird
neurons are more densely packed, so that gross size becomes less
relevant. And even neuron number is limited in explanations of
cognition. Chittka and Niven (2009) point out that the brains of
insects, with relatively very few neurons, support highly
differentiated motor repertoires, extensive social structures, as well
as a surprising variety of cognitive processes, many of which have
been described above. Using neural network analyses, they showed
that cognitive features found in insects may require only very limited
neuron numbers: Similar cognitive achievements do not imply
similar neural architecture.
Knowledge is acquired through general developmental processes,
including learning, as described in Chapters 7 and 8. Developmental
processes are complicated, however, and learning is actually a rather
nebulous concept. I would suggest there are three basic mental
processes: association, thinking, and reasoning. Association is the
attachment of one structure (idea, thought, or representation) to
another. Thinking (sometimes called “computation”) is the
integration of two or more sources or information (structures) into a
new structure. Reasoning is deconstructing a structure into its parts
and reconstructing a new structure. Flexibility in reasoning then
becomes the definition of intelligence. Unfortunately, thinking and
reasoning are two processes that are not yet fully understood; we
know when they are occurring, but we are not able to prescribe their
occurrence. Flexibility is possible to investigate, but only from a
human point of view. I will end with some food for thought. The
philosopher William James, in his chapter on reasoning, commented
(1890, vol. 2, pp. 332–333):
 …it is truly said that to know one thing thoroughly would be to
know the whole universe… one thing is related to everything else;
and to know all about it, all its relations need be known…. All
ways of conceiving a concrete fact, if they are true ways at all, are
equally true ways. There is no property ABSOLUTELY essential
to any one thing.

SUMMARY AND CONCLUSIONS

Animal cognition refers to the mental structures (knowledge)
possessed by an animal and the uses it makes of them (it). The
structures include perceptual and motor mechanisms and
central mechanisms (representations, memories, comparator
mechanisms, path integration mechanisms, and the like).
These mechanisms can vary in complexity and scope in
different species, from insects to primates. Many insects, in
fact, possess very sophisticated cognitive structures. These
structures are conceived of as components of the mind, and in
many cases have recognized neural correlates. Animals use
these structures to orient in space, perform everyday
activities, plan, and solve problems. Each species has evolved
the specific mechanisms it needed to survive and reproduce in
its ecological niche, and in many cases these mechanisms
differ from those found in humans.

FURTHER READING
Cognition, Evolution, and Behavior by Sara Shettleworth (2010) is
the authoritative source for details on most of the topics discussed in
this chapter. The Study of Behavior (Hogan 2017) puts cognitive
issues in a broader perspective. Wynne and Udell (2020), Animal
Cognition: Evolution, Behavior, and Cognition, present a wide
range of facts and examples of animal cognition suitable for the
general reader.
REFERENCES
Alem, S., Perry, F.J., Zhu, X. et al. 2016. Associative mechanisms
allow for social learning and cultural transmission of string
pulling in an insect. PLoS Biology, 14, 10.
Avargues-Weber, A., Dyer, A.G., Combe, M. & Giufa, M. 2012.
Simultaneous mastering of two abstract concepts by the miniature
brain of bees. Proceedings of the National Academy of Sciences
USA, 109, 7481–6.
Baerends, G.P. 1941. Fortpflanzungsverhalten und Orientierung der
Grabwespe Ammophila campestris. Jur. Tijdschrift voor
Entomologie, 84, 71–275.
Baerends, G.P. & Drent, R.H. (eds.). 1982. The herring gull and its
eggs. Behaviour, 82, 1–416.
Battley, P.F. et al. 2012. Contrasting extreme long-distance migration
patterns in bar-tailed godwits. Limosa lapponica. Journal of
Avian Biology, 43, 1–12.
Berry, C.S. 1908. An experimental study of imitation in cats. Journal
of Comparative Neurology and Physiology, 18, 1–12.
Berthold, P. 1973. Relationships between migratory restlessness and
migration distance in six Sylvia species. Ibis, 115, 594–9.
Bissonette, G.B., Bryden, D.W. & Roesch, M.R. 2014. You won’t
regret reading this. Nature Neuroscience, 17, 892–3.
Bruner, J.S., Goodnow, J.J. & Austin, G.A. 1956. A Study of
Thinking. New York: Wiley.
Cartwright, B.A. & Collett, T.S. 1982. How honeybees use landmarks
to guide their return to a food source. Nature, 295, 56–564.
Cheng, K. 1986. A purely geometric module in the rat’s spatial
representation. Cognition, 23, 149–78.
Chittka, L. & Niven, J. 2009. Are bigger brains better? Current
Biology, 19, R995-R1008.
Clayton, N.S. 2015. Ways of thinking: from crows to children and
back again. Quarterly Journal of Experimental Psychology, 28,
209–41.
Clayton, N.S. & Dickinson, A. 1998. Episodic-like memory during
cache recovery by scrub jays. Nature, 395, 272–4.
Collett, M., Chittka, L. & Collett, T.S. 2013. Spatial memory in insect
navigation. Current Biology, 23, R789-R800.
Collett, M. & Collett, T.S. 2000. How do insects use path integration
for their navigation? Biological Cybernetics, 83, 245–59.
Cross, E.R. 2020. Arthropod intelligence? The case for Portia.
Frontiers in Psychology, 11, 568049.
Cross, E.R. & Jackson, R.R. 2016. The execution of planned detours
by spider-eating predators. Journal of the Experimental Analysis
of Behavior, 105, 194–210.
Darwin, C. 1973. Origin of certain instincts. Nature, 7, 417–8.
Diedrichsen, J. & Kornysheva, K. 2015. Motor skill learning between
selection and execution. Trends in Cognitive Sciences, 19, 227–33.
Emlen, S.T. 1967. Migratory orientation in the indigo bunting.
Passerina cyanea. Auk, 84(309-342), 463–89.
Emlen, S.T. 1969. Bird migration: influence of physiological state
upon celestial orientation. Science, 165, 716–8.
Etienne, A.S. & Jeffery, K.J. 2004. Path integration in mammals.
Hippocampus, 14, 180–92.
Etienne, A.S., Maurer, R. & Séguinot, V. 1996. Path integration in
mammals and its integration with visual landmarks. Journal of
Experimental Biology, 199, 201–9.
Ewert, J.-P. 1997. Neural correlates of key stimulus and releasing
mechanism. Trends in Neurosciences, 20, 332–39.
Falótico, T. & Ottoni, E.B. 2014. Sexual bias in probe tool
manufacture and use by wild bearded capuchin monkeys.
 Behavioural Processes, 108, 117–22.
Falótico, T. & Ottoni, E.B. 2016. The manifold use of pounding tools
by wild capuchin monkeys of Serra da Capivara National Park,
Brazil. Behaviour, 153, 421–42.
Fraenkel, G.S. & Gunn, D.L. 1940/1961. The Orientation of Animals.
New York: Dover.
Galef, B.G., Jr. 1988. Imitation in animals: history, definition and
interpretation of data from the psychological laboratory. In T.R.
Zentall & B.G. Galef Jr. (eds.), Social Learning: Psychological
and Biological Perspectives, pp. 3–28. Hillsdale, NJ: Erlbaum.
Galef, B.G., Jr. 2013. Imitation and local enhancement: detrimental
effects of consensus definitions on analyses of social learning in
animals. Behavioural Processes, 100, 123–30.
Gallistel, C.R. 1980. The Organization of Action. Hillsdale, NJ:
Erlbaum.
Gallistel, C.R. 1990. The Organization of Learning. Cambridge, MA:
MIT Press.
Gallistel, C.R. & Matzel, L.D. 2013. The neuroscience of learning:
beyond the Hebbian synapse. Annual Review of Psychology, 63,
169–200.
Güntürkün, O. 2020. The surprising power of the avian mind.
Scientific American, 322, 48–55.
Güntürkün, O. & Bugynar, T. 2016. Cognition without cortex. Trends
in Cognitive Sciences, 20, 291–303.
Gwinner, E. 1986. Circannual rhythms in the control of avian
migration. Advances in the Study of Behavior, 16, 191–228.
Gwinner, E. 1996. Circadian and circannual programmes in avian
migration. Journal of Experimental Biology, 199, 39–48.
Haslam, M., Luncz, L.V. & Falótico, T. 2016. Pre-Columbian monkey
tools. Current Biology, 26, R515-R522.
Heyes, C.M. 1994. Social learning in animals: categories and
mechanisms. Biological Reviews, 69, 207–31.
Hogan, J.A. 2017. The Study of Behavior. Cambridge, UK:
Cambridge University Press.
Jackson, R.R. & Cross, E.R. 2011. Spider cognition. Advances in
Insect Physiology, 41, 115–74.
Jackson, R.R. & Wilcox, R.S. 1998. Spider-eating spiders. American
Scientist, 86, 350–7.
James, W. 1890. The Principles of Psychology. New York: Henry
Holt. [Reprinted 1950 by Dover Publications.].
Josselyn, S.A., Köhler, S. & Frankland, P.W. 2015. Finding the
engram. Nature Reviews Neuroscience, 16, 521–34.
Josselyn, S.A., Köhler, S. & Frankland, P.W. 2017. Heroes of the
engram. Journal of Neuroscience, 17, 4647–57.
Kok, E.M.A., Hogan, J.A. & Piersma, T. 2020. Experimental tests of a
seasonally changing visual preference for habitat in a long-
distance migratory shorebird. Ethology, 126, 681–93.
Kramer, G. 1952. Experiments on bird orientation. Ibis, 94, 265–85.
Kruijt, J.P. 1964. Ontogeny of social behaviour in Burmese red
junglefowl (Gallus gallus spadiceus). Behaviour, Suppl. 9.
Leadbeater, E. 2015. What evolves in the evolution of social learning.
Journal of Zoology, 295, 4–11.
Leadbeater, E. & Chittka, L. 2007. Social learning in insects—From
miniature brains to consensus building. Current Biology, 17,
R703–R713.
Loeb, J. 1918. Forced Movements, Tropisms and Animal Conduct.
Philadephia: Lippincott.
Lorenz, K. 1937. Über die Bildung des Instinktbegriffes.
Naturwissenschaften, vol. 25, pp. 289–300, 307–318, 324–331.
[Tr. as: The establishment of the instinct concept. In: Lorenz
 1970, Studies in Animal and Human Behaviour, vol. 1, pp. 259–
315. London: Methuen].
Lorenz, K. & Tinbergen, N. 1939. Taxis und Instinkthandlung in der
Eirollbewegung der Graugans. Zeitschrift für Tierpsychologie,
vol. 2, pp. 1–29. [Tr. as: Taxis and instinctive behaviour pattern in
egg-rolling by the greylag goose. In: Lorenz 1970, pp. 316–50].
Macedonia, J.M. & Evans, C.S. 1993. Variation among mammalian
alarm call systems and the problem of meaning in animal signals.
Ethology, 93, 177–97.
Mittelstaedt, M.-L. & Mittelstaedt, H. 1980. Homing by path
integration in a mammal. Naturwissenschaften, 67, 566–7.
Morris, R.G.M. 1981. Spatial localization does not require the
presence of local cues. Learning and Motivation, 12, 239–60.
Moser, M.-B. & Moser, E.J. 2016. Where am I? Where am I going?
Scientific American, 314(1), 26–33.
O’Keefe, J. & Nadel, L. 1978. The Hippocampus as a Cognitive Map.
Oxford: Clarendon.
Olive, R., Wolf, S., Dubreuil, A. et al. 2016. Rheotaxis of larval
zebrafish: Behavioral study of a multi-sensory process. Frontiers
in Systems Neuroscience, 10, 14.
Oteiza, P., Odstrcil, I., Lauder, G., Portugues, R. & Engert, F. 2017. A
novel mechanism for mechanosensory-based rheotaxis in larval
zebrafish. Nature, 547, 445–8.
Palameta, B. & Lefebvre, L. 1985. The social transmission of a food-
finding technique in pigeons: what is learned? Animal Behaviour,
33, 892–6.
Patterson, K., Nestor, P. & Rogers, T.T. 2007. Where do you know
what you know? The representation of semantic knowledge in the
human brain. Nature Reviews Neuroscience, 8, 976–87.
Piersma, T. & Van Gils, J. 2011. The Flexible Phenotype. Oxford:
Oxford University Press.
Ralph, M.A.L., Jefferies, E., Patterson, K. & Rogers, T.T. 2017. The
neural and computational bases of semantic cognition. Nature
Reviews Neuroscience, 18, 42–55.
Redish, A.D. 2001. The hippocampal debate: are we asking the right
questions? Behavioral Brain Research, 127, 81–98.
Resende, B.D., Nagy-Reis, M.B., Lacerda, F.N., Pagnotta, M. &
Savalli, C. 2014. Tufted capuchin monkeys (Sapajus sp.) learning
how to crack nuts: does variability decline throughout
development? Behavioural Processes, 109, 89–104.
Roberts, E.P. & Weigl, P.D. 1984. Habitat preference in the dark-
eyed junco (Junco hyemalis): the role of photoperiod and
dominance. Animal Behaviour, 32, 709–14.
Romanes, C.J.R. 1883. Mental Evolution in Animals. London.
Seyfarth, R.M. & Cheney, D.L. 1997. Some general features of vocal
development in nonhuman primates. In: Snowdon & Hausberger,
1997, pp. 249–73.
Seyfarth, R.M., Cheney, D.L. & Marler, P. 1980. Vervet monkey
alarm calls: semantic communication in a free-ranging primate.
Animal Behaviour, 28, 1070–94.
Sherry, D.F. & Duff, S.J. 1996. Behavioural and neural bases of
orientation in food-storing birds. Journal of Experimental
Biology, 199, 165–71.
Sherry, D.F. & Vaccarino, A.L. 1989. Hippocampus and memory for
food caches in black-capped chickadees. Behavioral
Neuroscience, 103, 308–18.
Shettleworth, S.J. 2010. Cognition, Evolution, and Behavior. New
York: Oxford University Press.
Snowdon, C.T. & Hausberger, M. 1997. Social Influences on Vocal
Development. Cambridge, UK: Cambridge University Press.
Sol, D., Bacher, S., Reader, S.M. & Lefebvre, L. 2008. Brain size
predicts the success of mammal species introduced into novel
 environments. American Naturalist, 172, S63–S71.
Steiner, A.P. & Redish, A. 2014. Behavioral and neurophysiological
correlates of rat decision-making on a neuroeconomic task.
Nature Neuroscience, 17, 995–1002.
Struhsaker, T.T. 1967. Social structure among vervet monkeys
(Cercopithecus aethiops). Behaviour, 29, 6–121.
Thorndike, E.L. 1898. Animal intelligence. Psychological Review,
Monograph suppl. 2, 1–113.
Tinbergen, N. 1932/1972. On the orientation of the digger wasp
Philanthus triangulum Fabr. In: N. Tinbergen (ed.), The Animal
in Its World, vol. 1, pp. 103–27. London: George Allen & Unwin.
Tinbergen, N. 1951. The Study of Instinct. Oxford: Oxford University
Press.
Tolman, E.C. 1948. Cognitive maps in rats and men. Psychological
Review, 55, 189–208.
Tulving, E. 1972. Episodic and semantic memory. In: E. Tulving & W.
Donaldson (eds.), Organization of Memory, pp. 381–403. New
York: Academic Press.
Tulving, E. 1983. Elements of Episodic Memory. New York: Oxford
University Press.
Van Lawick-goodall, J. 1970. Tool using in primates and other
vertebrates. Advances in the Study of Behavior, 3, 195–249.
Von Frisch, K. 1955. The Dancing Bees. New York: Harcourt, Brace.
Von Holst, E. 1937. Von Wesen der Ordnung im
Zentralnervensystem. Naturwissenschaften, vol. 25, pp. 625-631,
641–47. [Tr. as: On the nature of order in the central nervous
system. In: von Holst 1973, The Selected Papers of Erich von
Holst, vol. 1, pp. 3–32. London: Methuen].
Von Holst, E. 1954. Relations between the central nervous system
and the peripheral organs. British Journal of Animal Behaviour,
2, 89–94.
Wehner, R. & Menzel, R. 1990. Do insects have cognitive maps?
Annual Review of Neuroscience, 13, 403–14.
Wilson, D.M. 1966. Central nervous mechanisms for the generation
of rhythmic behaviour in arthropods. Symposia of the Society for
Experimental Biology, 20, 199–228.
Wiltschko, R. & Wiltschko, W. 2003. Avian navigation: from
historical to modern concepts. Animal Behaviour, 65, 257–72.
Wynne, C.D.L. & Udell, M.A.R. 2020. Animal Cognition, 3rd ed.
London: Red Grobe Press.
10
applied animal behavior and animal welfare
DAVID FRASER AND DANIEL M. WEARY

INTRODUCTION
Why do captive tigers pace for hours, tracing and retracing the
same path in their pens? Does the presence of tourists drive
grizzly bears away from their spring feeding areas? Why do some
pigs chew the tails off their pen-mates? Why is it so hard to get
wild rats to eat poisoned bait? Is it frustrating for a hen to live in
a cage with no secluded place to lay her eggs? Why do certain
dogs never become fully socialized with their human families?
Since animals can’t talk, how do we know when they are in pain?
These are a small sample of the questions that the field of applied
animal behavior has tried to answer as it grapples with practical
problems and animal welfare issues arising in the management of
wild, farmed, companion, and laboratory animals. These are the
issues discussed in this chapter.

An Old or New Field?

In one sense, applied animal behavior is one of the oldest fields of

science. In the foothills of the Rocky Mountains is a historic site
named “Head-Smashed-In Buffalo Jump” where the native people of
the plains hunted American buffalo (Bison bison) on foot for more
than 5,000 years. In the autumn the people would locate a group of
buffalo grazing in the vicinity of a particular cliff top. They would
then begin the slow process, perhaps extending over several days, of
moving the buffalo gradually closer to the cliff. Some of the hunters
dressed themselves in the skins of buffalo calves and positioned
themselves between the herd and the cliff, taking advantage of the
buffaloes’ poor eyesight and natural tendency to maintain protective
contact with calves. Others would dress in wolf skins and menace the
buffalo from the other side to assist the gradual movement of the
herd. Eventually this would bring the buffalo into the final valley
where they entered a configuration of rock cairns, each large enough
to conceal a person, arranged in a funnel-shaped pattern leading to
the cliff top. As the buffalo passed each cairn, a person would leap
into view, frightening the animals to create panic and keep them
bunched together until the herd was thundering at full gallop toward
the cliff. The site had been carefully chosen: it had a slight rise before
the precipice, so that the lead animals would not see the drop until it
was too late, and many would fall to their death. In this way, people
were able to move and kill herds of wild animals much larger and
stronger than themselves using little more than a highly developed
understanding of animal behavior.
In modern science, however, applied animal behavior is a
remarkably new field (see Box 10.1). Before 1900, practical problems
of hunting and game management stimulated some early
observational work on animal behavior (Thorpe 1979), and animals
of social and economic importance played a significant role in basic
behavioral research. Most notably, Pavlov (1927) used domestic dogs
in his research on classical conditioning; Von Frisch (1967) studied
domestic honeybees in his pioneering work on insect
communication; and domestic chickens served as a model in early
research on social organization (Allee c.1938). Nonetheless, it was
not until the 1950s and 1960s that scientific books on the application
of animal behavior to practical problems of animal management
appeared; the first Textbook of Domestic Animal Behaviour was
published in 1962 (Hafez 1962), followed by specialist books on the
reproductive and abnormal behavior of animals (Fox 1968; A.F.
Fraser 1968). The first scientific society in the field was founded in
1966, and the first journal in 1974. After this late beginning,
however, the field has grown rapidly, with an extensive scientific
literature, numerous textbooks, and growing educational
opportunities for biologists, animal scientists, veterinarians, and
other animal care professionals to study animal behavior and its
applications.
Box 10.1 The development of applied animal
behavior and animal welfare science
Period Applied animal Animal welfare Animal welfare
behavior concerns science
1900– • Domestic
1950 animals used in
basic behavioral
research
• Gradual
increase in
applied
research

1950s • Many papers • Animal welfare

on domestic & issues re-gain
applied animal prominence
behavior
• Special focus on
• Wild Animals laboratory animals
in Captivity and slaughter

1960s • Textbook of • Animal Machines • Thorpe (1965)

Domestic and the Brambell outlines how
Animal Committee (1965) scientific research
Behaviour focus attention on could assess and
the welfare of farm improve animal
• Early
specialist books animals welfare

• Society for
Veterinary
Ethology (SVE)
founded
1970s • Journal • Occasional books • Early research
Applied Animal on animal papers
Ethology ethics/advocacy
• First textbook
founded 1974 • Animal welfare chapter on farm
codes and guidelines animal welfare
 • Books on written
behavior of
farm animals
and poultry

1980s • Rapid • Many books on • Animal

increase in animal Suffering: The
research ethics/advocacy Science of Animal
literature • Many new animal Welfare
• Many books advocacy • First university
on domestic organizations chair in animal
and applied created welfare science
animal
behavior

1990s • Applied • Widespread media • Two scientific

animal attention to animal journals founded
behavior taught welfare • Numerous books
at many • Farm animal
universities • Clear influence
welfare standards in
of science on
legislation and animal welfare
international
standards
agreements
• Some food
corporations adopt
animal welfare
standards

2000s • 50th • Animal welfare • Several new

anniversary of promoted by major books per year
International international • Major multi-
Society for agencies (OIE, FAO)
institution
Applied • Widespread research projects
Ethology
attention to animal
(formerly SVE) welfare by
corporations
Column 1: The 1950s and 1960s saw a steady increase in scientific
papers on applied animal behavior together with books such as
Wild Animals in Captivity (Hediger 1950), the first Textbook of
Domestic Animal Behavior (Hafez 1962), and specialist books on
reproductive behavior (A.F. Fraser 1968) and abnormal behavior
(Fox 1968). The Society for Veterinary Ethology (now called
International Society for Applied Ethology) was founded in 1966
and the journal Applied Animal Ethology (now called Applied
Animal Behaviour Science) in 1974. The 1970s saw additional
books on the behavior of farm animals (A.F. Fraser 1974) and
chickens (Wood-Gush 1971), followed in the 1980s by a rapid
increase in the research literature and many new books (e.g.,
Craig 1981; Houpt & Wolski 1982; Waring 1983; Kilgour & Dalton
1984). By the 1990s the field was well established and taught at
many universities. Growth of the field continued in the 2000s
when the International Society for Applied Ethology celebrated
its 50th anniversary with over 500 members in more than 30
countries.
Column 2: Concern over animal welfare dates to ancient times
and regained prominence in Western countries after the Second
World War with particular emphasis on the use of laboratory
animals and humane slaughter. In the 1960s public concern
about living conditions for farm animals was catalyzed in the
United Kingdom by Animal Machines (Harrison 1964) followed
by the Brambell Committee (1965), which called for evaluation
and improvements in the welfare of farm animals kept in
intensive production systems. The 1970s saw other books on
animal ethics and advocacy such as Animal Liberation (Singer
1975) and Slaughter of the Innocent (Ruesch 1978), and welfare
codes and guidelines for farm animals were written in several
countries. The 1980s saw many additional books (e.g., Rollin
1981, Midgley 1983, Regan 1983) and the formation of many new
animal advocacy organizations. By the 1990s animal welfare and
ethics were receiving widespread media attention, and farm
animal welfare standards were being adopted in legislation,
international agreements, and by some corporations. The 2000s
saw a continuation of major media attention and development of
standards, and animal welfare received explicit attention by
 major international agencies including the World Organisation
for Animal Health (OIE) and the Food and Agriculture
Organization of the United Nations (FAO).
Column 3: Beginning in the 1960s, concerns about animal welfare
led to calls for scientific research to assess and improve animal
welfare with W.H. Thorpe’s (1965) essay playing a key role. The
1970s saw early research papers with titles like “Frustration in
the fowl” (Duncan 1970) and “Do hens suffer in battery cages?”
(Dawkins 1977), together with the first textbook chapter on farm
animal welfare (Wood-Gush et al. 1975). The 1980s saw the first
book on animal welfare science (Animal Suffering: The Science
of Animal Welfare by Dawkins 1980) and the creation of the first
university chair in animal welfare science at Cambridge in 1986.
In the 1990s, two scientific journals were founded (Animal
Welfare in 1992 and Journal of Applied Animal Welfare Science
in 1998), numerous books were published (e.g., Webster 1994;
Rollin 1995, Appleby & Hughes 1997, Fraser & Broom 1997), and
the role of scientific research in the emerging animal welfare
standards was clear. In the 2000s research continued to expand,
several new books were published every year, and two major
multi-institution studies (“Welfare Quality” and “Animal Welfare
Indicators”) were funded by the European Commission.

Just when applied animal behavior was developing as a field of

study, an emerging social issue catapulted it into prominence. In
1964 an English animal welfare advocate published Animal
Machines, a book describing confinement systems of animal
production such as cages for hens and crates for veal calves, which
were then becoming common (Harrison 1964). The book argued that
these systems are so restricting and unnatural for animals that they
cause widespread suffering. Animal Machines created such a strong
public reaction that the British government set up a technical
committee to investigate the welfare of farm animals. Included on
the committee was the eminent ethologist William Thorpe, known
for his research on bird song, who recognized that scientific research
on animal behavior and related fields could do much to help answer
questions about the welfare of animals in captivity. In an essay that
formed part of the committee report, Thorpe (1965) proposed how
behavioral research could be used to detect pain and discomfort, to
understand the cognitive powers of animals, and to identify
motivations that are thwarted in captivity; he also noted that
physiological studies can indicate stress in animals, and that choice
tests could be used to identify environments that animals prefer. The
committee, clearly impressed by these ideas, called for research into
the welfare of animals, especially through studies of animal behavior.
Such research began in the 1960s and 1970s initially in the United
Kingdom, and spread rapidly to other countries, mainly in Europe
and the English-speaking world. Thus, barely 3 years after its first
textbook had been published, the field of applied animal behavior
was cited as a primary source of guidance on the highly contentious
issue of animal welfare.
As a result of these developments, the field of applied animal
behavior has acquired two somewhat distinct but overlapping
mandates: more conventionally, to understand the behavior of
animals in order to solve practical problems of animal housing,
management, and health; and more controversially, to contribute to
“animal welfare science” that seeks to improve our understanding of
the welfare of animals, partly as a basis for humanitarian action and
social policy.

Practical Problems, Practical Solutions

Using the Abilities of Animals

For thousands of years, people have used certain abilities of animals,
such as the horse’s strength and the dog’s powerful sense of smell, to
complement human talents and to accomplish tasks that people
alone could not do. For such use of animals to be successful, handlers
need to understand and shape the behavior of the animals. For
example, “detector dogs” can find narcotics concealed in luggage and
other parcels (Figure 10.1). In the Australian Customs Service, dogs
are trained for this task in a series of stages that make use of the
animals’ fondness for playing tug-of-war. Initially the dogs are
trained to retrieve a stick or ball, with praise and a game of tug-of-
war as the reward. The stick is then replaced by a cloth tug-of-war
toy that is subsequently hidden in open cardboard boxes; at this
stage, the dog must search for the toy and bring it to the trainer for
the reward. Next, the cloth is impregnated with the scents of various
narcotics, and the toy is concealed in closed boxes. The dogs thus
learn to use the scent to find the toy, and subsequently they search
eagerly for narcotic scents in luggage and other containers. While the
dog is working in this way, the trainer keeps the dog’s toy close by
and continues to play tug-of-war as a reward for successful detection
(Adams & Johnson 1994).
Figure 10.1 A dog in the Australian Customs Service trained to
search for narcotics. Photo (by Peter Sanderson) courtesy of Dr. G.J.
Adams.
Scent detection is one of the simpler animal talents that humans
exploit, but animals—dogs in particular—are also used for
complicated tasks such as herding sheep, apprehending criminals
and, mostly during the past century, a wide range of “service”
activities to assist the blind, the deaf, and people with a range of
physical or mental disabilities. These tasks require suitable
temperament and cognitive skills plus extensive training. However,
many dog trainers report a high failure rate among dogs that begin
the long and costly training process. After years of experience with
dogs, Emily Weiss (2002) developed a series of tests that can be used
to screen adult dogs from animal shelters for suitability as service
animals. The tests involve training dogs to follow simple instructions
and exposing them to a wide range of situations including other
dogs, strange people, and fear-producing events. Dogs fail the test if
they are fearful, unruly, or slow to learn, if they overreact to other
dogs or to a sudden noise or visual stimulus (an opened umbrella), or
if they react aggressively to mild pain. By identifying better and
worse candidates, the test procedure can allow animal trainers to
focus on animals with a high probability of success.
Recent years have seen an explosion of new uses of dogs for medical
purposes. Dogs have been trained to identify people with lung cancer
from breath samples and people with bladder cancer from urine
samples, presumably by detecting traces of chemicals produced by
the tumors. Dogs have also been used to predict seizures in epileptics
and hypoglycemia in diabetics, perhaps from odor cues or from
subtle changes in behavior. The next decades are likely to see major
advances in such use of animals and in understanding of the
mechanisms that allow dogs to perform these useful tasks (Wells
2009).

Preventing Undesirable Behavior

As well as beneficial behavior that people try to exploit, animals often
perform harmful behavior that people try to prevent. Rodents can be
serious crop pests; deer can destroy orchards; introduced species can
cause significant damage to ecological systems. Billions of animals
have been killed deliberately, often by trapping and poisoning, but
such measures are often futile because the offending animals are
simply replaced through reproduction or movement.
However, we can sometimes find more effective, nonlethal solutions
through knowledge of animal behavior. In North America, deer
(Odocoileus hemionus) and elk (Cervus elephas) sometimes cause
severe damage to plantations of seedling trees. In a search for an
effective repellent, researchers assembled many candidate chemicals
and screened them for repellency by spraying them on samples of
feed which were then presented to captive deer. They found that
decomposed protein-rich material was highly repugnant to deer. On
this basis they developed a fermented egg product treated so as to
give off a rotting odor, which repelled deer even in minute
concentrations, perhaps because these ruminant herbivores may
have evolved a capacity to detect and avoid rotting material that their
digestive systems are not equipped to tolerate. The research formed
the basis of various commercial repellents that protect vegetation by
giving off sulfurous odors (Nolte 1998).
As a more sophisticated approach, scientists have used “conditioned
taste aversion” (a form of classical conditioning) to prevent certain
undesirable behavior. On the island ecosystems of New Zealand,
introduced predators have brought certain species of birds to the
brink of extinction. A famous example is the Kakapo (Strigops
habroptilus), a large flightless parrot, whose decline has occurred
partly because its eggs are eaten by introduced rats. An ingenious
approach to controlling such predation has been to add nonlethal
toxins to some eggs so that predators learn that this type of food
makes them ill. In an experimental demonstration of this effect,
researchers exposed crows (Corvus brachyrhynchos) to chicken eggs
that had been inoculated with toxins and painted green. The crows
quickly learned to avoid these eggs and subsequently avoided green
eggs without the toxin (Nicolaus et al. 1983).
In an interesting twist on this approach, Australian researchers have
used conditioned aversion to protect predators from dangerous prey.
Cane toads (Bufo marinus) were introduced to Australia in the 1930s
and now number in the hundreds of millions. They carry a toxin that
is harmful to humans and fatal to many small predators that attack
them. The northern quoll (Dasyurus hallucatus) is a cat-sized
marsupial predator whose numbers have plummeted in areas where
cane toads have invaded. In an effort to protect reintroduced quoll,
scientists tested whether the animals could learn to avoid cane toads
through conditioned taste aversion. They created a group of “toad-
smart” quoll by feeding them a dead, nontoxic toad to which they
had added the nausea-inducing chemical thiabendazole. After one
such exposure, the toad-smart animals were more reluctant to attack
a toad than a control group, and they appeared to show longer
survival when released into an area where toads were present
(O’Donnell et al. 2010).
Some captive animals direct harmful behavior to themselves or their
pen-mates. Some pigs chew each other’s tails to the point of causing
serious injury. Some chickens peck their cage-mates to the point of
denuding, injuring, or even killing them. Certain zoo or laboratory
animals mutilate themselves or attack their newborn young. By
understanding the causes of such behavior, applied ethologists can
often find ways to prevent or mitigate the problem. One of the most
intractable examples is “cross sucking” by young calves. Dairy calves
are normally removed from the mother within the first day after
birth, at an age when they still have high motivation to suck, and
they are generally fed a milk-based or milk-like diet from a bucket. If
young calves are housed in groups, they will often suck avidly on the
navel, ears, or prepuce of other calves, sometimes to the point of
causing damage. The most common solution has been to house
calves in individual cubicles, but the practice deprives the young
animals of social contact. However, if the calves obtain their milk by
sucking from artificial teats, this appears to satisfy their motivation
both for milk for the act of sucking and allows the animals to be
raised in groups. Even for calves in stalls, dummy teats attached to
the wall allow the animals to suck in a harmless way and with the
additional advantage of causing a release of digestive enzymes that
are not released to the same extent when animals drink from a
bucket (De Passillé & Rushen 1997).
Domestic cats, both household cats and unowned “feral” cats, are
important predators of small birds and other prey. The problem is
considered especially severe in New Zealand where the native birds
evolved in the absence of mammalian predators and seem to have
very poor defenses against cats. A study in the city of Dunedin
recorded the known kills of 144 household cats. By applying the
average kill rate to the total number of cats in the city, the
researchers hypothesized that the cats likely exterminate most of the
city’s population of certain birds each year and that the persistence
of those species in the city depends on an annual influx of birds from
outlying areas with fewer cats. However, the research also showed
that most of the deaths were due to about a quarter of the cats
whereas many cats did little or no hunting (Van Heezik et al. 2010.).
The death rate could be reduced by identifying the persistent
predators and keeping them indoors during daytime or fitting them
with a collar and bell, which significantly reduces the effectiveness of
their hunting (Gordon et al. 2010).

Improving Animal Handling

A better understanding of animal behavior often allows people to be
more effective in handling animals. Early in his career, Australian
researcher Paul Hemsworth studied a group of 12 small, one-
operator pig farms in the Netherlands. The farms used the same
genetic line of pigs, the same diet, and the same building design, but
the number of piglets born per sow each year differed considerably
from farm to farm. Hemsworth found a striking behavioral
difference between animals on farms with the greatest and least
number of piglets born. On the more productive farms, the breeding
sows would more readily approach a human visitor, while on the less
productive farms, the animals tended to shy away. This observation
and a large body of subsequent research (Hemsworth & Coleman
2011) led to the theory that rough, unskillful, or inconsistent
handling by people can create learned fear responses that interfere
with the basic endocrine processes underlying reproduction and
growth. In one striking demonstration, inappropriate handling of
dairy cows interfered with milk yield. In this experiment, cows
received rough handling, including slaps and shouts, from one
handler, and more gentle handling from another. When later tested
in the milking parlor, cows in the presence of the rough handler
“retained” more milk (i.e., they failed to release the milk during
milking), likely because the neuro-endocrine responses triggered by
fear can counteract the milk-releasing action of the hormone
oxytocin (Rushen et al. 1999).

Mitigating Harm to Animals

Animal behavior has also been applied in efforts to reduce harm to
animals caused by human actions. In forested areas of Canada, many
moose (Alces alces) are killed on highways every year. Moose show a
strong attraction to natural mineral-rich springs at roughly the times
of year when highway collisions are common. “Cafeteria”
experiments, using pails of pure salt solutions presented at mineral
springs (Figure 10.2), showed that of the various minerals available
in the springs, it is sodium that attracts the animals (Fraser &
Reardon 1980). Subsequent research on highway accidents showed
that many accident “hot-spots” involve poorly drained areas of the
roadside where highway deicing salt, applied during the winter,
remains dissolved in stagnant pools of sodium-rich water.
Management of the problem involves eliminating some of the
stagnant pools by better roadside drainage and by placing warning
signs in the immediate vicinity of wet, salty areas that cannot be
drained successfully (Rea et al. 2014).
Figure 10.2 A moose (Alces alces) drinking from one of several
plastic pails sunk into the ground at a natural mineral spring in
Canada, as part of a preference experiment to determine which
chemicals attract moose to the springs. Photo by Hank Hristienko.
During the past century, the world has seen an explosion of tall
communication towers for radio, television, and mobile telephones.
From the beginning, concerned citizens recognized that towers were
killing large numbers of birds, especially on overcast nights.
Research in North America has shown that most of the collisions
involve tropical migrants flying at night during the spring and
autumn migration. Nighttime behavioral observations made by a
portable radar device showed birds flying in circles around a lit tower
under foggy conditions. The behavior suggests that birds become
disoriented or blinded by the lights on the towers under some
weather conditions and may fly nearby until they collide with the
tower or the network of supporting cables (Larkin & Frase 1988).
Further research compared towers with different lighting systems
and found many fewer collisions at towers with flashing rather than
steady-burning lights. The researchers suggested that the number of
collisions could be cut possibly in half simply by the use of flashing
lights, although very tall towers (>305 m in height) remain serious
killers regardless of the lighting system (Gehring et al. 2009).
Windows are another modern hazard for birds. In the past, windows
were typically composed of small panes surrounded by enough
opaque material that birds could recognize the window as a solid
structure. However, with the large sheets of glass now being used in
buildings, windows seem to be invisible to birds, and a vast number
of birds collide with windows every year. In seeking possible
solutions, ornithologist Daniel Klem studied birds in flight tunnels
and field experiments where birds would fly near different types of
glass and plastic. His results indicated that birds detect certain
wavelengths of ultraviolet light that are invisible to people. He
therefore tested glass surfaces that had been covered with a film that
created patterns of stripes or grids that would reflect or absorb
ultraviolet light. Windows treated in such ways had far fewer
collisions than plain glass. Klem estimated that such films, if
retrofitted into buildings or incorporated into new window designs,
could save hundreds of millions of birds each year (Klem 2009).

Designing Better Environments for Animals

One of the major applications of animal behavior is the design of

better environments for captive, farmed, and laboratory animals,
partly for the practical goal of making the environments function
better, and partly to improve the welfare of the animals that live in
them.

Accommodating Animals’ Natural Behavior

One strategy for improving animal housing has been to devise
environments where animals can carry out elements of their natural
behavior. As primatologist Jane Goodall (1971) discovered during her
field studies, wild chimpanzees (Pan troglodytes) will use long twigs
as “fishing rods” for termites. A chimpanzee pokes the twig into a
hole in a termite mound, and when the termites react by biting the
twig, the chimpanzee withdraws the twig and eats the termites. Real
termite fishing would be difficult to re-create in most zoos, but if
captive chimpanzees are provided with a large container with holes,
and some long sticks, they will occupy a good deal of time fishing for
a taste-treat such as honey or mustard, with a substantial decrease in
the time they spend inactive (Celli et al. 2003).
Servals (Felis serval) are long-legged cats of west central Africa that
hunt by leaping on their prey, sometimes flushing birds from low
vegetation and catching them by spectacular jumps into the air.
When kept and fed in a standard zoo cage, servals spend much of
their time inactive or pacing repetitively in the cage. In his classic
account of behavioral enrichment for zoo animals, Hal Markowitz
(1982) described how the servals became much more animated—and
more interesting to zoo visitors—when fed “flying meatballs”
attached to a rope or rod and swung over the heads of the cats
(Figure 10.3). Servals fed in this way would leap vertically twice their
body length to capture the prey.
Figure 10.3 A serval (Felis serval) leaping into the air to capture a
“flying meatball” in the San Diego Zoo. Photo by Dr. Hal Markowitz.

Testing Environmental Preferences

Another common strategy in improving animal environments is to
“ask” animals what environments they themselves prefer. William
Thorpe, in his initial proposal for using science to improve animal
welfare, cited the example of a group of African buffalo (Syncerus
caffer) that had to be relocated to Nairobi National Park in Kenya.
While being prepared for release, they were enclosed in paddocks
near the park office. After they were released, they kept returning
toward nightfall and tried to reenter the paddocks. Thorpe concluded
that at nighttime the buffalo preferred the cramped but protected
environment of the paddocks ahead of open parkland with its
abundance of predators (Thorpe 1965).
Since that unplanned experiment, scientists have used
environmental preference research with many animal species to
identify preferred levels of temperature, illumination, social contact
and space allowance, preferred flooring and bedding materials, and
preferred design features for pens, cages, and animal handling
equipment (Fraser & Nicol 2017). One example involved using
preference research to improve the handling of pigs. Ramps are often
used in loading pigs into vehicles, but pigs refuse to walk up certain
ramps for reasons that were not well understood. Engineer Peter
Phillips tested the animals’ preferences as a basis for improving
designs. He placed pigs in a small pen with four ramps that the pigs
could climb at will. One experiment offered ramps with four different
slopes, another compared four levels of illumination, and so on. The
pigs clearly preferred to walk on ramps with shallow slopes (up to
about 25 degrees) and ramps with closely spaced horizontal foot-
holds that provided secure footing. In contrast, they showed no
preference related to the width of the ramp, the level of illumination,
or the openness of the side walls. Using this information, Phillips
was able to recommend design features that would allow pigs to be
loaded easily and with little stress to the animals (Phillips et al.
1988).
Testing Motivation Strength
By themselves, simple preferences do not indicate the degree of
importance that the animal attaches to the preferred option. This
weakness led to attempts to measure the strength of animals’
motivation to obtain preferred options or to avoid unpreferred ones.
In an influential paper called “Battery hens name their price,” British
ethologist Marian Dawkins (1983) proposed a way to “titrate” an
unknown motivation against a known one. “Dustbathing” is a natural
behavior whereby chickens work dust, sand, or other loose material
into their feathers, thus keeping the feathers in good condition by
absorbing excess oil. Because the behavior is impossible to perform
in standard commercial cages, a debate arose over how strongly
birds are motivated to perform it. In one experiment Dawkins
trained hens to enter two cages from a common choice point; one
cage contained dust-bathing material and the other contained food.
The hens were then required to choose between the two cages after
different periods of food deprivation. The experiment showed that
the hens’ motivation to dustbathe (under the conditions tested)
appeared to be about as strong as their motivation to eat when food-
deprived for several hours.
As a further refinement, animals can be trained to perform an
operant task, such as pressing a lever or pecking a key, in order to
obtain a reward, and researchers can then compare how much work
animals are willing to perform in order to obtain access to various
environmental features. American mink (Mustela vison) are active,
partially aquatic carnivores that are raised commercially for fur. In
the wild, mink perform a wide range of behavior that is impossible in
captivity; for example, they swim, rest in several nest sites, survey
the environment from raised perching places, and explore the
burrows of potential prey animals. But how important is it to mink to
be able to carry out these types of behavior? In one study, mink in
standard cages were trained to push against weighted doors for
access to various rewards including a tunnel, a raised platform, an
alternative nest box, and a small pool of water where they could
swim. The experimenters then varied the amount of weight that the
animals had to lift to open the different doors. The amount of work
the animals performed, expressed as the total amount of weight they
lifted per unit of time, was much greater for access to the pool of
water than for the raised platform, the tunnel, and other rewards
(Figure 10.4; Mason et al. 2001).

Figure 10.4 The price that caged American mink (Mustela vison)
will pay for access to various environmental features. Price is
expressed in the total weight (kg) lifted by mink over a 6-week test.
Values are means (+ S.E.) from 16 animals tested. Data are from
Mason et al. (2001)
An especially sophisticated example of this approach comes from a
study by Laura Webb and colleagues who used “cross-point” analysis
of double demand functions to test which types of forage are most
valued by calves. In one experiment Webb and colleagues trained
calves to push their muzzle against a panel and thus trigger the
release of roughage (chopped hay), and they varied the “price” using
a ratio of 7, 14, 21, 28, or 35 presses per reward (called the Fixed
Ratio or FR). This generated a classic “demand curve” with the calves
receiving fewer rewards as the price increased. However, the calves
could also switch to a second panel that released long hay at a price
of 42 minus the FR for chopped hay. Thus, when chopped hay was
available at 7 presses, long hay was available at 35 presses, and vice
versa. If the calves found the two forages equally attractive, we would
expect that the two demand curves should intersect exactly in the
middle; i.e., calves would show equal demand for the two options
when the price was the same (Fixed Ratios of 21 and 21). Instead,
however, the calves appeared to value long hay more because they
received equal proportion of long and chopped hay when the price of
long hay was higher (Figure 10.5; Webb et al. 2014).

Figure 10.5 Calf demand for two types of forage, illustrating

cross-point analysis. The figure shows the cross point (cp) for one
calf, using the demand curves for chopped hay (circles) and long hay
(squares). The y-axis shows the number of chopped hay rewards or
long hay rewards as a proportion (p) of the total number of rewards
earned. The x-axis shows the number of muzzle presses required to
earn one reward of chopped hay (the “fixed ratio” or FR), with the
price increasing from 7 to 35 presses. The number of presses
required for long hay was 42 - FR. Figure is reprinted with
permission from Webb et al. (2014)
 Behavioral Genetics and Animal Welfare

Instead of designing environments that suit the animals’ behavior,

could we solve animal welfare problems by the reverse strategy? That
is, could we use genetic selection to create animals that are so well
adapted to restrictive, artificial environments that their welfare is not
compromised? In fact, has the artificial selection of domestic animals
already achieved this result?
In some cases, the answer is no. A wild sow (Sus scrofa), in the days
before she gives birth, seeks out a suitable nesting site, digs a shallow
depression, and builds a nest of branches and grass that create a soft,
protected area for the newborn piglets. Presumably the behavior was
important for the survival of the young under natural conditions, and
evolution appears to have equipped sows with strong motivation,
probably under the control of the hormonal changes leading to
parturition, to find and prepare a nest. Domestic sows, however,
have been bred for many decades in protected, indoor environments
where nesting behavior is impossible to perform and is no longer
needed for piglet survival. Nonetheless, when researchers have
released domestic sows in forested areas, the animals behaved much
like their wild cousins, suddenly performing elaborate nesting
behavior that had not been seen for generations. Moreover, when
kept in indoor stalls, sows show intense restlessness during the day
before parturition, changing posture repeatedly, performing digging
movements against the hard floor, biting the bars of the stall, and
showing an intense interest in straw or other loose material.
Evidently, the motivation to find and prepare a nest site is still strong
in the domestic sow, and an inability to perform the behavior
remains an animal welfare concern despite many generations of
artificial selection (Wischner et al. 2009).
In other cases, especially involving general temperament rather than
specific behavior patterns, fairly rapid genetic change may be
possible. One of the welfare concerns for captive fur-bearing
carnivores is that these animals remain fearful of people and may
experience substantial stress from close human proximity. In
Denmark, where there is extensive commercial breeding of mink,
applied ethologists have devised a rapid test for temperament
differences: they put a stick into the front of the mink’s cage and note
whether the animal approaches and investigates the stick, attacks it,
or retreats from it. By breeding selectively from mink that approach
and investigate the stick, scientists have produced, within just a few
generations, animals that appear to remain calm in the presence of
humans. The animals’ welfare is presumably improved because they
show little fear of people and a less pronounced physiological stress
response to handling. There are also commercial advantages because
the animals, being calmer, can be mated at a younger age than
fearful ones (Malmkvist & Hansen 2001).
In yet other cases, selective breeding may help reduce welfare
problems that likely arose through previous genetic selection. On
commercial egg farms, where laying hens are often housed in cages
of 3–10 birds, the hens sometimes peck each other to the point of
causing feather loss and occasionally injuries or even death. The
problem has likely been accentuated because geneticists, in breeding
selectively from birds with the highest egg production in a group,
may inadvertently have favored highly competitive birds whose
aggressiveness reduces the production of their cage-mates. As an
alternative breeding strategy, some poultry geneticists have housed
hens in cages of closely related birds, and then bred selectively from
those groups where the birds achieved a high level of production on
average. Applying such group-based selection for several generations
produced a line of chickens (which the breeders called KGB for “kind
gentle birds”) whose egg production was substantially better, in part
because of reduced aggression, fewer deaths, and significant changes
in physiological stress responses (Cheng 2010).

Animal Welfare and Affective States

Many concerns about animal welfare are primarily concerns about

the affective states of animals—their “emotions,” “feelings,” and
other pleasant or unpleasant experiences. Are the animals in our care
experiencing fear, pain, hunger, and other negative states, or
alternatively are they comfortable, contented, happy? If scientists are
to provide guidance on animal welfare, they must do their best to
confront these difficult issues (Duncan 1993).

Abnormal Behaviors
Abnormal animal behavior has provided a strong stimulus to
understand affective states of animals. In a classic article on
behavioral disturbances, psychiatrist David Levy (1944) described
the repeated rocking movements and other stereotyped behavior
shown by some emotionally disturbed children. In the same article
Levy described similar behavior in farm animals—including
repetitive weaving by horses in stables and stereotyped head
movements by caged chickens—and Levy proposed that these
abnormalities reflect the same kinds of problems seen with children
in severely deprived environments. The idea generated enormous
interest in abnormal animal behavior and its implications for animal
welfare.
A striking example is seen with pregnant sows. During pregnancy
most sows have to be limited in their food intake in order to prevent
excessive weight gain and later health problems, but the restricted
diet can lead to serious aggression if the animals are fed in groups. A
common solution is to house sows individually, often in narrow stalls
where they are sometimes tethered by a collar around the neck.
Some such sows develop stereotyped movement patterns; for
example, a sow may make three rooting movements to the left, swing
her head to the right, and bite the bar of the stall, and then repeat
that same sequence of movements for several hours every day. But
what is the significance of these bizarre movements? Various
scientists have proposed that the behavior develops from exploratory
motivation, or from attempts to escape from a confined space, or that
the behavior helps animals to “cope” with an aversive environment,
for example by causing a release of endogenous opioids (see
Lawrence & Rushen 1993).
However, research in Scotland showed that food restriction plays a
large role in the development of stereotyped behavior of sows. One
hypothesis is that food restriction causes a motivation to forage for
food, but in a barren environment where normal foraging is
impossible, elements of foraging behavior become fused together
into stereotyped sequences of behavior, and because the behavior
never leads to actual eating, it does not turn off in the normal
manner. In a critical experiment, Claudia Terlouw and coworkers
(1991) tested the food-restriction theory. They housed some sows in
narrow stalls where they were tethered by a chain while others were
loose-housed in larger pens equipped with chains hanging from the
walls. In each housing treatment, half the animals were fed a
restricted diet typical of commercial practice while the others
received nearly twice as much. In both pens and stalls, the restricted-
fed animals spent considerable time in chewing, biting, and rooting
the chain and in other seemingly functionless and stereotyped
behavior, whereas these activities were much less common among
the geneyrously fed animals. Terlouw et al. (1991) concluded that
chain-manipulation and other stereotyped behavior of sows are more
related to hunger than to restraint itself, but they noted that none of
the available theories adequately explains this bizarre behavior.

Assessing Affective States

Of the various affective states, fear is one of the most widely studied.
An intriguing example arose over the mechanical “harvesting” of
chickens that are raised for meat. These birds, unlike caged laying
hens, are normally housed in large, open buildings; when they reach
market weight, crews of people are employed to catch the birds,
usually grabbing them by the legs and carrying them upside-down to
the shipping crates. A newer alternative is a mechanical “chicken
harvester”—a large machine that moves through the pens, gathers
the birds in counter-rotating rubber fingers, and transfers them to
the shipping crates by a conveyor belt.
When chicken harvesters first appeared, there was concern that they
would cause unnecessary fear in the birds. To investigate this
concern, Ian Duncan—one of the pioneers of animal welfare research
—monitored the heart rate of birds when they were captured by hand
or by machine and found (perhaps surprisingly) that the rapid heart
rate of newly caught birds returned to normal much more quickly if
they had been caught by machine rather than by hand. A widely used
behavioral test for fear in chickens involves flipping a bird suddenly
onto its back, whereupon the bird will often stay totally immobile for
many minutes in a reaction called tonic immobility. It can be shown
experimentally that chickens tend to remain in tonic immobility
longer if they have been frightened before the test. Duncan and co-
workers found that chickens that had been loaded by hand
maintained tonic immobility for over 10 minutes on average,
whereas those that had been loaded by machine righted themselves
much sooner. The evidence thus suggested that machine catching
actually caused less fear than manual catching, probably because the
mechanical device, being so foreign to the birds, did not trigger any
form of predator recognition (Duncan et al. 1986). On this basis,
many animal welfare advocates came to support the use of
mechanical catchers.
Pain is another state that animal welfare scientists have often tried to
quantify. Many routine procedures performed on animals—including
ear-cropping of puppies, castration of piglets, and dehorning of
calves—are likely painful, and much recent research has been
directed at developing measures of this pain so that less painful
procedures can be identified.
An interesting example comes from pediatric medicine. Circumcision
of babies, like many of the procedures done to farm animals, has
typically been performed without any intervention to control pain. In
one influential study, infant boys were observed during circumcision
with and without a local anesthetic applied to the foreskin. The
researchers evaluated pain responses using facial actions scored
from videotape. These included bulging of the brow, squeezing the
eyes closed, opening the mouth, pursing the lips, holding the tongue
taut, quivering of the chin, and protrusion of the tongue (Figure
10.6). Babies that received the local anesthetic showed 12–49% less
facial activity than those that received a placebo. Measures of vocal
activity and heart rate showed similar effects: the anesthetic-treated
infants cried for half as long, and their heart rate increased less than
the placebo-treated controls (Taddio et al. 1997). These results
helped lead to the recommendation that infants should receive a
local anesthetic before circumcision.
Figure 10.6 Facial expressions of a newborn human infant before
and immediately after heel-lancing for a blood sample.
Characteristics of the facial response to pain include bulging of the
brow, squeezing the eyes closed, opening the mouth, pursing the lips,
holding the tongue taut, quivering of the chin, and protrusion of the
tongue. Photo from Grunau and Craig (1987)
Facial expressions can also be used to assess pain in some animals.
For example, mice are often used by laboratory researchers to study
pain and pain treatments. Work by Jeffrey Mogil and colleagues has
shown that some of the same facial features of human infants are
also seen in mice that are in pain, along with other behavior (changes
in ear and whisker position) not shown by the human infants
(Langford et al. 2010). Similar research has led to “grimace scales”
for various species such as horses (Dalla Costa et al. 2014
But could such behavioral signs be merely reflexive responses and
hence not indicative of consciously felt pain or emotion? Conditioned
place preference experiments allow us to infer that emotions are
consciously experienced, by “asking” animals about their memory of
hedonic experiences when the stimuli responsible are no longer
present. In conditioned place preference experiments, animals
experience a treatment (e.g., morphine) in a distinctive test chamber,
for example with brightly colored walls. Animals are then re-tested
without the treatment and allowed to move between the distinctive
test chamber and a control chamber. If the animal preferentially
visits (or avoids) the distinctive test chamber, we can infer that the
animal had a positive (or negative) hedonic experience in the test
chamber and that their memory of this experience is driving the
preference. In one such study, researchers used this technique to
assess chemicals used to euthanize laboratory fish (Wong et al.
2014). Zebrafish (Danio rerio) are commonly used in laboratory
research and are typically euthanized at the end of the study with an
overdose of tricaine methanesulfonate (TMS). In this study fish were
allowed to swim between two tanks attached by a narrow swim tube.
One tank was brightly lit and the other was kept dark. The fish
preferred the bright tank and would consistently swim to the bright
side when first placed into the dark tank. On one day TMS was added
to the water in the bright tank, before the fish were allowed to enter.
The fish then entered the bright tank and became unconscious. The
fish were then allowed to recover in another tank and then retested.
Even though there was now no chemical on either side of the tank,
most of the fish that had been exposed to TMS spent little or no time
in the bright tank. When fish were exposed to other anesthetics, like
clove oil, all fish continued to prefer the bright side. These results
indicate that the zebrafish found exposure to the TMS unpleasant
and remembered this unpleasant experience (even after just one
exposure), and that the memory was enough to overcome the
previous preference for the bright tank. More practically, these
results identified more humane alternatives for euthanasia of fish.
Studies in humans show that repeatedly experiencing pain, fear, or
other unpleasant states can lead to low mood, including depression,
which in turn causes people to interpret information more
pessimistically. Research in animal welfare has now used these
cognitive changes to assess mood states in animals. Emma Harding,
Elizabeth Paul, and Michael Mendl pioneered this method in animals
(Harding et al. 2004). They trained rats to discriminate between two
sounds that differed in pitch. One pitch (the positive tone) signaled
that pressing a lever would lead to a food reward; the other (the
negative tone) signaled that the same response would lead to an
unpleasant noise. Once rats had learned the discrimination they
were tested with unrewarded “probes” of intermediate pitch. As
expected, rats responded more to the test probes closer in pitch to
the positive tone. However, when a group of rats was kept in
unpleasant and unpredictable housing (thought to induce low mood
from other studies), the rats reacted pessimistically by treating the
test probes more like the negative tone (i.e., by not responding).
Such “cognitive-bias” testing has now been used to assess mood
states in various animals. One study used this approach to assess
how two routine farm practices affect mood states in dairy calves.
Like the rats described above, calves were trained to discriminate
between two cues—in this case different colors on a video monitor.
The animals were then re-tested in the hours after they had
undergone “hot-iron disbudding” (a painful procedure that prevents
the horn-buds from developing), and they showed a pessimistic bias
in responding to intermediate test colors. A few weeks later the same
calves were separated from their mothers, and the calves showed a
similar cognitive bias in the days that followed. These results indicate
that routine procedures like disbudding and separation can affect
longer term mood states in animals. The results provide a further
scientific basis for the recommendation that calves receive treatment
for pain after disbudding (which is often not provided on farms), and
indicate that efforts are required to minimize the social distress
associated with separating calves from their mothers (Figure 10.7;
Daros et al. 2014).
Figure 10.7 Cognitive bias in dairy calves after routine farm
procedures. The figure shows the mean ± SE percent responses to
the video screen during test sessions (a) before and after hot-iron
disbudding (dehorning), and (b) before and after separation from the
mother. Responses are shown separately for the two training colors
(positive and negative) and for the three ambiguous probe colors
(near-positive, intermediate, and near-negative). In the graphs, both
groups responded at 100% to the positive screen and at 0% to the
negative, but they differed in their responses to one or more of the
ambiguous screens. Figure is reprinted with permission from Daros
et al. (2014)
Fear, pain, separation distress, avoidance learning, and stereotyped
behavior are some of the simpler issues being studied by animal
welfare scientists. At the frontier of the field, scientists are asking
more unconventional questions. Can we operationalize “boredom” in
animals (Meagher & Mason 2012)? Can we find indices of positive
states (“well-being”) that will apply to both humans and nonhuman
species (Boissy et al. 2007)? What is suffering in animals and do
negative experiences such as pain and fear interact to result in
suffering (Weary)?

Probing the Limits of Science

Science and the mental experience of animals

But can science really help us understand the affective states of other
species? A century ago biological scientists had little doubt on this
issue. The last great work of Charles Darwin (1872) was called The
Expression of the Emotions in Man and Animals. Darwin’s friend
and contemporary George Romanes, in his book Animal Intelligence,
assembled narrative accounts of animal behavior from what he
regarded as reliable sources. From these he tried to infer the
emotional capacity of many species; he argued, for example, that
elephants are capable of vindictiveness toward humans or elephants
that have harmed them, and sympathy toward the injured (Romanes
1904). The pioneering ethologist Julian Huxley (1914), quoted by
Burkhardt 1997, p. 8) proposed that through the study of behavior,
“we can deduce the bird’s emotions with much more probability of
accuracy than we can possibly have about their nervous processes.”
However, at the time when Darwin was pondering the evolution of
emotions, another influential scientist was proposing a view that
excluded such interests from science altogether. Auguste Comte was
the 19th century French thinker whose “Philosophie Positive” gave us
the school of thought that we call “Positivism.” Comte was
attempting to draw a clear distinction between science—interpreted
as the study of the material world—and other branches of thought
such as theology and metaphysics. Science, as seen by the Positivists,
is concerned only with what we can observe. With this emphasis on
the tangible, Positivist thinkers held that we should not postulate
unobservable processes to explain observable ones and that
processes that cannot be observed fall outside the realm of scientific
enquiry. On this basis, some early Positivists believed that the origin
of the universe, the atomic theory of matter, and the theory of
evolution are not part of science. These ideas have now faded from
memory, but Positivism had a more lasting influence on the study of
animal behavior, with both behaviorists such as John Watson and
ethologists such as Niko Tinbergen claiming that the mental states of
animals should play no role in the scientific study of behavior (Rollin
1989; Burkhardt 1997). In the words of Tinbergen (1951, p. 4),
“Because subjective phenomena cannot be observed objectively in
animals, it is idle either to claim or to deny their existence.”
Arguably, this attitude imposed a valuable discipline on behavioral
science. It helped prevent scientists from unrestrained speculation
about the mental lives of animals, from assuming that the mental
experiences of nonhumans are the same as those of humans, and
from believing that behavior is explained when it is merely rephrased
in mentalistic terms. These errors are all too easily made by writers
purporting to describe the subjective lives of animals. For example,
one popular writer, after seeing a televised sequence of a puma that
had just killed a bighorn sheep, described the puma as gazing
“fondly” into the sheep’s eyes and “tenderly” patting the sheep’s face
(Thomas 1994, p. 25). This was subsequently presented by other
popular writers as evidence that predators feel “gratitude” toward
their prey (Masson & McCarthy 1995, p. 174). The moral is that when
we use scientific methods to probe the affective states of animals, we
will need rigorous thinking and clear means of testing our ideas, or
else we will see all manner of gratuitous speculation jumbled
together with science.
Nonetheless, Marian Dawkins, while fully recognizing the need for
intellectual discipline, proposed that we not shrink from the
challenge of understanding the mental states of animals:
Consciousness is the greatest remaining mystery in biology…
Ethical concerns about the welfare of animals come from many
people’s deeply held conviction that many nonhuman animals
consciously experience emotions such as fear, anxiety, and
boredom… So the study of animal welfare, by its very nature … is
forced to confront the greatest remaining mystery in biology.
(Dawkins 2001, p. S19)
The philosopher Daniel Dennett has given us a vocabulary for
understanding the different approaches to the issue of animal
consciousness. Dennett (1987) refers to a “stance” as a conceptual
framework or set of working presuppositions that scientists adopt to
interpret observations and guide further empirical study. The view of
the animal welfare scientist—that higher vertebrates can consciously
experience states such as pain, fear, and hunger—is not a hypothesis
that we can expect to either prove or disprove; rather, it is a stance
adopted by animal welfare scientists (the “affective stance,” in
contrast to the Positivist stance adopted by Watson and Tinbergen)
that will ultimately be judged by its usefulness in suggesting
interesting questions, in leading to correct predictions, and in
providing satisfying explanations.
Apart from its role in the study of animal welfare, could efforts to
understand the affective states of animals also give us better
explanatory models and predictive theories of behavior? Jane
Goodall (1971) gave many narrative descriptions of the behavior of
free-living chimpanzees. She described, for example, how “Mike,” an
adult male, repeatedly collected empty paraffin cans and banged
them together while charging toward other, more dominant
chimpanzees; this caused the others to scatter and was followed by a
distinct rise in Mike’s dominance status within the group. Can we
provide a satisfactory explanation of this behavior without
postulating certain mental processes such as a desire for dominance
and planning of the display?
At the time, many animal behavior scientists dismissed Goodall’s
descriptions as mere “anecdotes” that fell outside the realm of
science. But were these descriptions, in fact, a useful expansion of
the paradigm of animal behavior research? The standard paradigm
at the time involved a search for some form of average or central
tendency, with little attention to the details of behavior or to
differences between individuals. Such research created little
apparent need for explanations involving mental and emotional
states. For example, as long as scientists confined themselves to
measuring the average speed of rats running toward a food source
after different periods of deprivation, the results could be expressed
mathematically, and nothing would be gained by postulating that the
rats “felt” hungry and “expected” to find food. But if we expand our
notion of “data” to include such unique behavior as Mike charging
his group-mates while banging noisy objects together, then theories
involving the emotional and cognitive capacities of species may play
a key role in understanding behavior (Fraser 2009). In other fields,
such as physics and biology, scientists reached a point where they
needed to postulate the existence and properties of phenomena such
as electron spin and speciation by natural selection that could not be
observed directly. Has animal behavior reached a similar stage, or
are there reasons why science cannot, or should not try to,
understand the mental states of animals?

Animal Welfare, Science, and Ethics

The application of science to animal welfare also raises important
questions about the relationship between science and ethics.
Philosopher David Hume (1711–1776) famously pointed out that we
cannot derive an “ought” from an “is”—in other words, that we
cannot use facts to answer ethical questions. Questions about animal
welfare are ultimately motivated by concerns over how we ought to
treat animals, whereas science deals with questions of fact. How,
then, can there be a science of animal welfare?
One response has been to draw a clear line between research on
variables associated with animal welfare—such as signs of fear, pain,
and disease—and ethical decisions about what constitutes an
acceptable level of these variables. The distinction is important, but
the interplay between science and values is more subtle than this
simple compartmentalization would suggest.
The variables that scientists choose to study in assessing animal
welfare, and the interpretation they attach to these variables, are
underlain by value-laden ideas about what constitutes a good life for
animals. Some scientists, including some veterinarians and
agriculturalists, tend to emphasize the biological functioning of
animals (health, growth, etc.) as fundamental to animal welfare; in
assessing animal welfare, these scientists are likely to use traditional
measures such as disease incidence, reproductive success, and stress
physiology (Table 10.1). Other scientists tend to emphasize affective
states as the basis for animal welfare, and their scientific measures of
welfare include indicators of pain, distress, and related states. Yet
others see satisfactory animal welfare as requiring that animals can
live relatively natural lives, in accordance with their evolved
adaptations; in studying animal welfare these scientists tend to use
the occurrence of natural behavior (body care, normal social
interaction) as indicators of good animal welfare, and abnormal
behavior (stereotyped behavior, self-mutilation, excessive
aggression) as denoting impaired welfare.

Table 10.1 Three Conceptions of Animal Welfare, and Typical

Measures Used to Provide Positive (+) or Negative (-) Evidence of
Animal Welfare
Conception of Typical measures
animal welfare
Biological functioning -increase in stress hormones
-reduction in immune competence
-incidence of disease and injury
+longevity
+growth rate
+reproductive success
Conception of Typical measures
animal welfare

Affective states -behavioral signs of fear, pain, frustration,

etc.
-physiological changes thought to reflect
fear, pain, etc.
-behavioral signs of aversion or learned
avoidance
+behavioral indicators of
comfort/contentment
+performance of behavior (e.g., play)
thought to be pleasurable
+behavioral signs of approach/preference
Natural living -behavioral/physiological indicators of
thwarting of natural behavior
-performance of abnormal behavior
+performance of natural behavior
These different views of animal welfare sometimes lead to similar
conclusions. For example, allowing a sow to wallow in cool mud
when the weather is hot should be good for animal welfare because
the sow should avoid stress and disease problems caused by heat (a
biological functioning criterion), should not suffer from the heat (an
affective state criterion), and because wallowing is the animal’s
natural way of cooling off (a natural living criterion).
Nonetheless, the three views involve quite different areas of
emphasis and sometimes lead to conflicting conclusions. For
example, hens in a confined, high-health flock may have little disease
and high rates of growth and production, whereas hens kept
outdoors are freer to perform natural behavior but they may also
have more parasites, lay fewer eggs, and be more prone to fear of
predators. In many cases there is no logically or empirically correct
way of weighing these different elements. Hence, in the scientific
assessment of animal welfare our empirical studies are underlain by
a conceptual framework where values play a key and sometimes
subtle role in the selection and interpretation of variables. From its
beginnings, therefore, animal welfare research has included debate
about the values underlying the science, the interplay between values
and science, and the ethical implications of the results (Fraser 2008).
In writing this chapter, we found a pleasant irony in these last few
pages. Applied animal behavior began as an attempt to solve real-life
problems through an understanding of animal behavior. The field
thus tended to be strong on practical ingenuity but light on theory—a
kind of rough-and-ready country cousin of theoretical science. As it
has developed, however, applied animal behavior has found itself
probing the conceptual frontiers of animal behavior research, re-
thinking whether and how to study affective experience in animals,
confronting fundamental questions about animal consciousness, and
articulating the often subtle interplay between empirical and value
issues in the conduct and interpretation of research.

SUMMARY AND CONCLUSIONS

Applied animal behavior science began through the application of
animal behavior to practical problems: improved handling and
raising of animals, designing better housing for animals in
captivity, prevention of abnormal or undesirable behavior,
genetic manipulation of behavioral traits, and more effective use
and training of animals. Since the 1960s the field has also
provided means of assessing and improving animal welfare
through studying the preferences and motivations of animals,
abnormal behavior, and affective states such as fear, pain, and
distress. Both goals have given rise to a suite of research
approaches that tend to combine behavioral observations and
experiments with elements of stress physiology, veterinary
medicine, animal production, and environmental design. In
dealing with animal welfare issues, the field has been stretched
beyond the traditional boundaries of animal behavior research—
into probing the affective states of animals and understanding the
interplay of empirical and ethical elements that occurs when we
apply science to an area of social action.
FURTHER READING
The major scientific journal dealing with applied animal behavior is
Applied Animal Behaviour Science, published since 1974 initially
under the name Applied Animal Ethology; however, papers on
animal behavior applied to practical problems appear in many
journals covering animal, poultry and dairy science, animal behavior,
veterinary medicine, zoo biology, laboratory animal science, and
other fields. Journals specializing in the scientific study of animal
welfare are Animal Welfare and Journal of Applied Animal Welfare
Science; journals dealing with human-animal interaction and
animals in society include Animals and Society and Anthrozoös.
Recent books providing a general introduction to farm or domestic
animal behavior include Broom and Fraser (2015), Jensen (2009),
and Houpt (2011). Recent books on animal behavior have been
written about many individual species including cats (Bradshaw et
al. 2012; A.F. Fraser 2012; Turner & Bateson 2014), dogs (Horowitz
2014), and horses (Waring 2003; Hausberger et al. 2007; A.F. Fraser
2010). Specialized topics that have received book-length treatment
include stereotyped behavior (Mason & Rushen 2006),
transportation and handling of animals (Grandin 2014), and
human–animal interaction (Hemsworth & Coleman 2011). Recent
works on animal welfare science include Appleby et al. (2014, 2018);
Fraser (2008); and Mellor et al. (2009) as well as numerous books
on individual species.

REFERENCES
Adams, G.J. & Johnson, K.G. 1994. Sleep, work, and the effects of
shift work in drug detector dogs, Canis familiaris. Applied Animal
Behaviour Sciemce, 41, 115–26.
Allee, W.C. c.1938. The Social Life of Animals. New York: W.W.
Norton & Co.
Appleby, M.C. & Hughes, B.O. (eds.). 1997. Animal Welfare.
Wallingford: CABI.
Appleby, M.C., Olsson, I.A.S. & Galindo, F. (eds.). 2018. Animal
Welfare, 3rd ed. Wallingford: CABI.
Appleby, M.C., Weary, D.M. & Sandøe, P. (eds.). 2014. Dilemmas in
Animal Welfare. Wallingford: CABI.
Boissy, A., Manteuffel, G., Jensen, M.B. et al. 2007. Assessment of
positive emotions in animals to improve their welfare. Physiology
and Behavior, 92, 375–97.
Bradshaw, J.W.S., Casey, R.A. & Brown, S.L. 2012. The Behaviour of
the Domestic Cat, 2nd ed. Wallingford: CABI.
Brambell, F.W.R. (chairman). 1965. Report of the Technical
Committee in Enquire into the Welfare of Animals kept under
Intensive Livestock Husbandry Systems. London: Her Majesty’s
Stationery Office.
Broom, D.M. & Fraser, A.F. 2015. Domestic Animal Behaviour and
Welfare, 5th ed. Wallingford: CABI.
Burkhardt, R.W., Jr. 1997. The founders of ethology and the problem
of animal subjective experience. In: M. Dol, S. Kasanmoentalib, S.
Lijmbach, E. Rivas & R. Van Den Bos (eds.), Animal
Consciousness and Animal Ethics, pp. 1–13. Assen: Van Gorcum.
Celli, M.L., Tomonaga, M., Udono, T., Teramoto, M. & Nagano, K.
2003. Tool use task as environmental enrichment for captive
chimpanzees. Applied Animal Behaviour Science, 81, 171–82.
Cheng, H.-W. 2010. Breeding of tomorrow’s chickens to improve
well-being. Poultry Science, 89, 805–13.
Craig, J.V. 1981. Domestic Animal Behavior: Causes and
Implications for Animal Care and Management. Englewood
Cliffs: Prentice-Hall.
Dalla Costa, E., Minero, M., Lebelt, D. et al. 2014. Development of
the Horse Grimace Scale (HGS) as a pain assessment tool in
horses undergoing routine castration. PLoS One, 9, e92281.
Daros, R.R., Costa, J.H.C., Von Keyserlingk, M.A.G., Hötzel, M.J. &
Weary, D.M. 2014. Separation from the dam causes negative
judgment bias in dairy calves. PLoS One, 9, e98429.
Darwin, C. 1872. The Expression of the Emotions in Man and
Animals. Reprinted 1965. Chicago: University of Chicago Press.
Dawkins, M.S. 1977. Do hens suffer in battery cages? Environmental
preferences and welfare. Animal Behaviour, 25, 1034–46.
Dawkins, M.S. 1980. Animal Suffering: The Science of Animal
Welfare. London: Chapman & Hall.
Dawkins, M.S. 1983. Battery hens name their price: Consumer
demand theory and the measurement of ethological “needs.”
Animal Behaviour, 31, 1195–1205.
Dawkins, M.S. 2001. Who needs consciousness? Animal Welfare,
10(Suppl.), S19–S29.
De Passillé, A.M. & Rushen, J. 1997. Motivational and physiological
analysis of the causes and consequences of non-nutritive sucking
by calves. Applied Animal Behaviour Science, 53, 15–31.
Dennett, D.C. 1987. The Intentional Stance. Cambridge, MA: MIT
Press.
Duncan, I.J.H. 1970. Frustration in the fowl. In: B.M. Freeman &
R.F. Gordon (eds.), Aspects of Poultry Behaviour, pp. 15–31.
Edinburgh: British Poultry Science Ltd.
Duncan, I.J.H. 1993. Welfare is to do with what animals feel. Journal
of Agricultural and Environmental Ethics, 6(Suppl. 2), 8–14.
Duncan, I.J.H., Slee, G., Kettlewell, P., Berry, P. & Carlisle, A.J. 1986.
Comparison of the stressfulness of harvesting broiler chickens by
machine and by hand. British Poultry Science, 27, 109–14.
Fox, M.W. (ed.). 1968. Abnormal Behavior in Animals. Philadelphia:
Saunders.
Fraser, A.F. 1968. Reproductive Behaviour in Ungulates. London:
Academic Press.
Fraser, A.F. 1974. Farm Animal Behaviour. London: Baillière
Tindall.
Fraser, A.F. 2010. The Behaviour and Welfare of the Horse, 2nd ed.
Wallingford: CABI.
Fraser, A.F. 2012. Feline Behaviour and Welfare. Wallingford: CABI.
Fraser, D. 2008. Understanding Animal Welfare: The Science in its
Cultural Context. Oxford: Wiley-Blackwell.
Fraser, D. 2009. Animal behaviour, animal welfare and the scientific
study of affect. Applied Animal Behaviour Science, 118, 108–17.
Fraser, A.F. & Broom, D.M. 1997. Farm Animal Behaviour and
Welfare, 3rd ed. Wallingford: CABI.
Fraser, D. & Nicol, C.J. 2017. Preference and motivation research. In:
M.C. Appleby, I.A.S. Olsson & F. Galindo (eds.), Animal Welfare,
3rd ed., pp. 213–231. Wallingford: CABI.
Fraser, D. & Reardon, E. 1980. Attraction of wild ungulates to
mineral-rich springs in central Canada. Holarctic Ecology, 3, 36–
40.
Gehring, J., Kerlinger, P. & Manville, A.M. 2009. Communication
towers, lights, and birds: successful methods of reducing the
frequency of avian collisions. Ecological Applications, 19, 505–14.
Goodall, J. 1971. In the Shadow of Man. London: Wm Collins.
Gordon, J.K., Matthaei, C. & Van Heezik, Y. 2010. Belled collars
reduce catch of domestic cats in New Zealand by half. Wildlife
Research, 37, 372–8.

Grandin, T. (ed.). 2014. Livestock Handling and Transport, 4th ed.

Wallingford: CABI.
Grunau, R.V.E. & Craig, K.D. 1987. Pain expression in neonates:
facial action and cry. Pain, 28, 395–410.

Hafez, E.S.E. (ed.). 1962. The Behaviour of Domestic Animals. (3rd

ed., 1975) Baltimore: Williams & Wilkins.
Harding, E.J., Paul, E.S. & Mendl, M. 2004. Cognitive bias and
affective state. Nature, 427, 312.
Harrison, R. 1964. Animal Machines. London: Vincent Stuart.
Hausberger, M., Sondergaard, E. & Martin-Rosset, W. (eds.) 2007.
Horse Behaviour and Welfare. EAAP publication No. 122.
Wageningen: Wageningen Academic Publishers.
Hediger, H. 1950. Wild Animals in Captivity (tr. G. Sircom).
London: Butterworth.
Hemsworth, P.H. & Coleman, G.J. 2011. Human-livestock
Interactions: The Stockperson and the Productivity and Welfare
of Intensively Farmed Animals, Revised. Wallingford: CABI.
Horowitz, A. (ed.). 2014. Domestic Dog Cognition and Behavior.
Heidelberg: Springer.
Houpt, K.A. 2011. Domestic Animal Behavior for Veterinarians and
Animal Scientists, 5th ed. Ames: Wiley-Blackwell.
Houpt, K.A. & Wolski, T.R. 1982. Domestic Animal Behavior for
Veterinarians and Animal Scientists. Ames: Iowa State University
Press.
Jensen, P. (ed.). 2009. The Ethology of Domestic Animals: An
Introductory Text, 2nd ed. Wallingford: CABI.
Kilgour, R. & Dalton, C. 1984. Livestock Behaviour: A Practical
Guide. Boulder: Westview Press.
Klem, D., Jr. 2009. Preventing bird–window collisions. Wilson
Journal of Ornithology, 121, 314–21.
Langford, D.J., Bailey, A.L., Chanda, M.L. et al. 2010. Coding of
facial expressions of pain in the laboratory mouse. Nature
Methods, 7, 447–9.
Larkin, R.P. & Frase, B.A. 1988. Circular paths of birds flying near a
broadcasting tower in cloud. Journal of Comparative Psychology,
102, 90–3.
Lawrence, A.B. & Rushen, J. (eds.). 1993. Stereotypic Animal
Behaviour: Fundamentals and Applications to Welfare.
Wallingford: CABI.
Levy, D.M. 1944. On the problem of movement restraint: tics,
stereotyped movements, hyperactivity. American Journal of
Orthopsychiatry, 14, 644–71.
Malmkvist, J. & Hansen, S.W. 2001. The welfare of farmed mink
(Mustela vison) in relation to behavioural selection: a review.
Animal Welfare, 10, 41–52.
Markowitz, H. 1982. Behavioral Enrichment in the Zoo. New York:
Van Nostrand Reinhold.
Mason, G.J., Cooper, J. & Clarebrough, C. 2001. Frustrations of fur-
farmed mink. Nature, 410, 35–6.
Mason, G.J. & Rushen, J. (eds.). 2006. Stereotypic Animal
Behaviour: Fundamentals and Applications to Welfare, 2nd ed.
Wallingford: CABI.
Masson, J.M. & McCarthy, S. 1995. When Elephants Weep: The
Emotional Lives of Animals. New York: Delacorte Press.
Meagher, R.K. & Mason, G.J. 2012. Environmental enrichment
reduces signs of boredom in caged mink. PLoS One, 7, e49180.
Mellor, D.J., Patterson-Kane, E. & Stafford, K.J. 2009. The Sciences
of Animal Welfare. Oxford: Wiley-Blackwell.
Midgley, M. 1983. Animals and Why They Matter. Athens, GA:
University of Georgia Press.
Nicolaus, L.K., Cassel, J.F., Carlson, R.B. & Gustavson, C.R. 1983.
Taste-aversion conditioning of crows to control predation on eggs.
Science, 22, 212–4.
Nolte, D.L. 1998. Efficacy of select repellents to deter deer browsing
on conifer seedlings. International Biodeterioration and
Biodegradation, 42, 101–7.
O’Donnell, S., Webb, J.K. & Shine, R. 2010. Conditioned taste
aversion enhances the survival of an endangered predator
imperilled by a toxic invader. Journal of Applied Ecology, 47,
558–65.
Pavlov, I.P. 1927. Conditioned Reflexes: An Investigation of the
Physiological Activity of the Cerebral Cortex (G.V. Anrep. tr.,
1960 Reprint). New York: Dover Publications.
Phillips, P.A., Thompson, B.K. & Fraser, D. 1988. Preference tests of
ramp designs for young pigs. Canadian Journal of Animal
Science, 68, 41–8.
Rea, R.V., Johnson, C.J. & Emmons, S. 2014. Characterizing moose–
vehicle collision hotspots in northern British Columbia. Journal
of Fish and Wildlife Management, 5, 46–58.
Regan, T. 1983. The Case for Animal Rights. Berkeley: University of
California Press.
Rollin, B.E. 1989. The Unheeded Cry: Animal Consciousness,
Animal Pain and Science. Oxford: Oxford University Press.
Rollin, B.E. 1992. Animal Rights and Human Morality, Revised ed.
(1st edition, 1981). Buffalo: Prometheus Books.
Rollin, B.E. 1995. Farm Animal Welfare: Social, Bioethical, and
Research Issues. Ames: Iowa State University Press.
Romanes, G.J. 1904. Animal Intelligence, 8th ed. London: Kegan
Paul, Trench, Trübner & Co.
Ruesch, H. 1978. Slaughter of the Innocent. New York: Bantam
Books.
Rushen, J., De Passillé, A.M. & Munksgaard, L. 1999. Fear of people
by cows and effects on milk yield, behavior and heart rate at
milking. Journal of Dairy Science, 82, 720–7.
Singer, P. 1990. Animal Liberation, 2nd ed. (1st edition, 1975). New
York: Avon Books.
Taddio, A., Stevens, B., Craig, K. et al. 1997. Efficacy and safety of
lidocaine-prilocaine cream for pain during neonatal circumcision.
New England Journal of Medicine, 336, 1197–1201.
Terlouw, E.M.C., Lawrence, A.B. & Illius, A.W. 1991. Influences of
feeding level and physical restriction on development of
stereotypies in sows. Animal Behaviour, 42, 981–91.
Thomas, E.M. 1994. The Tribe of the Tiger. New York: Simon &
Schuster.
Thorpe, W.H. 1965. The assessment of pain and distress in animals.
In: F.W.R. Brambell chairman, Report of the Technical
Committee in Enquire into the Welfare of Animals kept under
Intensive Livestock Husbandry Systems, pp. 71–79 (Appendix
III). London: Her Majesty’s Stationery Office.
Thorpe, W.H. 1979. The Origins and Rise of Ethology: The Science
of the Natural Behaviour of Animals. London: Heinemann
Educational Books.
Tinbergen, N. 1951. The Study of Instinct. Oxford: Clarendon Press.
Turner, D.C. & Bateson, P.P.G. (eds.). 2014. The Domestic Cat: The
Biology of its Behaviour, 3rd ed. Cambridge: Cambridge
University Press.
Van Heezik, Y., Smyth, A., Adams, A. & Gordon, J. 2010. Do
domestic cats impose an unsustainable harvest on urban bird
populations? Biological Conservation, 143, 121–30.
Von Frisch, K. 1967. Dance Language and Orientation of Bees (L.E.
Chadwick. tr.). Cambridge, MA: Belknap Press of Harvard
University Press.

Waring, G.H. 2003. Horse Behavior, 2nd ed. (1st edition, 1983).
Norwich (USA): Noyes.
Weary, D.M. 2014. What is suffering in animals? In: M.C. Appleby,
D.M. Weary & P. Sandøe (eds.), Dilemmas in Animal Welfare, pp.
188–202. Wallingford: CABI.
Webb, L.E., Bak Jensen, M., Engel, B. et al. 2014. Chopped or long
roughage: what do calves prefer? Using cross point analysis of
double demand functions. PLoS One, 9(2), e88778.
Webster, J. 1994. Animal Welfare: A Cool Eye towards Eden.
Oxford: Blackwell Science.
Weiss, E. 2002. Selecting shelter dogs for service dog training.
Journal of Applied Animal Welfare Science, 5, 43–62.
Wells, D.L. 2009. The effects of animals on human health and well-
being. Journal of Social Issues, 65, 523–43.
Wischner, D., Kemper, N. & Krieter, J. 2009. Nest-building
behaviour in sows and consequences for pig husbandry. Livestock
Science, 124, 1–8.
Wong, D., Von Keyserlingk, M.A.G., Richards, J.G. & Weary, D.M.
2014. Conditioned place avoidance of zebrafish (Danio rerio) to
three chemicals used for euthanasia and anaesthesia. PLoS One,
9, e88030.
Wood-Gush, D.G.M. 1971. The Behaviour of the Domestic Fowl.
London: Heinemann Educational.
Wood-Gush, D.G.M., Duncan, I.J.H. & Fraser, D. 1975. Social stress
and welfare problems in agricultural animals. In: E.S.E. Hafez
(ed.), The Behaviour of Domestic Animals, 3rd ed., pp. 182–200.
London: Baillière Tindall.
11
the function of behavior
LUC-ALAIN GIRALDEAU AND JERRY A. HOGAN
INTRODUCTION
Avian migration, the seasonal displacement of millions of birds
from one hemisphere to another is, as all mass migrations, an
astonishing phenomenon: tons of biomass moving from one
continent to another, spending enormous amounts of energy in
the process and all to what end? This remarkable sequence of
events can be approached from a purely mechanistic point of
view. Birds, for instance, respond to changing photoperiod,
which, added to signals from an internal clock, trigger a
hormonal cascade that leads to accumulation of fat reserves and
makes the bird inclined to fly off at night with a strong directional
preference based on stellar position. However, as we pointed out
in Chapter 1, no matter how hard you study this chain of events
and learn how birds migrate, no matter how detailed the
knowledge you get of the physiological processes involved in
hormone production or the way the birds use the stars as a
compass, you will be no closer to knowing why birds migrate in
the first place. Why is it that the internal machinery of some bird
species, but not others, reacts to changing photoperiod in a long
chain of events that lead to long distance migration? To answer
this type of question we must leave the realm of cause and
mechanism and enter the sphere of function and adaptation. This
chapter reviews some of the approaches and methods used to
study the adaptive function of behavior (see also Chapter 15). It
first looks at how adaptation can be studied by formulating
quantitative hypotheses about function using backward
engineering and simple optimality models. The approach will be
illustrated with two classic examples taken from foraging theory:
prey choice and patch residence times, including some results
from experimental tests of the models. Then in the next section it
introduces another version of the optimality approach,
evolutionary game theory and its solution, the Evolutionarily
Stable Strategy, using three examples: the Hawk-Dove and
Producer-Scrounger games and the Ideal Free Distribution.
 Introducing Optimality

That enhanced organ and behavior design have evolved due to

natural selection can be inferred through examples of convergent
evolution. Compare, for instance, the shapes of appendages in a
number of different aquatic organisms such as fish, sea turtles,
dolphins, and penguins. Despite the diverse evolutionary origins of
these fin-like appendages (some were wings, others legs or arms)
their forms now resemble each other. This resemblance no doubt
follows because they serve a similar purpose; a means to power
movement through a dense medium. The similarity in shape can be
an adaptation if we can ascribe it to the action of natural selection
that favored certain variants from originally quite different-looking
appendages that made them more efficient at accomplishing their
function of propelling individuals through water. The argument
seems reasonable. However, in order to go beyond such apparently
sensible comments about adaptation and turn them into science, we
need to become a bit more quantitative and explicit about what we
expect natural selection to have produced.

Backward Engineering
Engineers are asked to design machines that will accomplish some
function with maximal efficiency. They build the machine knowing
its purpose. Biologists, however, have the complete machine before
them and puzzle about its purpose. To figure out the purpose they
follow the reverse route, they postulate a function, and then ask
whether the details of the machine are consistent with maximal
efficiency of this function. That, in essence, is what we mean by
backward engineering. The method of backward engineering
assumes that selection has shaped traits, organs, limbs, behaviors
etc, so that after many generations of being submitted to the same
selective pressures they offer the greatest possible efficiency in the
accomplishment of their purpose. We say greatest possible because
evolution (much like the engineer) is often constrained. The design is
assumed to be maximally efficient given physical, genetic,
morphological, historical, and other constraints. So, if a trait is
thought to be adaptive, we hypothesize its function and ask: what
would be the characteristics of a trait that is optimally designed to
accomplish this hypothetical function? We use mathematical,
economic, and engineering techniques to predict these
characteristics of optimal design and compare them to observations.
If we find a quantitative fit between some feature of an organ, say its
length or width, and the precise length or width expected of an organ
optimally designed to accomplish that same function, then we would
feel more confident that selection honed the trait we are dealing with
to accomplish that purpose and can with more certainty conclude
that it is an adaptation.

Optimal Flight Speeds: An Example of the Logicof

Backward Engineering
Each bird species flies at some typical speed when engaged in its
everyday activities. Is this speed chosen randomly or is it an
adaptation; has it been selected to accomplish some specific adaptive
function? One possibility is that selection favored birds that
minimize energy use per unit time spent flying: the minimum power
speed. Alternatively, selection might have favored birds that
maximize the distance covered per unit energy used: the maximum
range speed. These are two hypotheses about the adaptive function
of flight speed, hypotheses concerning the selective pressures that
have given rise to the flight speeds we observe in birds today. Let’s
see how a behavioral ecologist would set out to test these hypotheses.
The physics of flight are well known and so we can use backward
engineering to construct an optimality model that generates
quantitative expected flight speeds for each hypothetical adaptive
function. From the shape of a bird’s wing and its body mass it is
possible to construct a power function that describes the relationship
between the power required to sustain level flight and flight speed
for that wing shape. This power function is generally a parabola that
first declines as speed increases, reaches a minimum, and then
increases again (Figure 11.1). The minimum point of the function
gives the minimum power flight speed. That is the speed that
requires the least power and hence provides the greatest energetic
savings. The maximum range speed can be obtained by drawing a
line from the origin and tangent to the power curve. The point of
tangency gives the maximum range speed; the speed that allows the
animal to cover the greatest distance per unit power used. If flight
speed is an adaptation to minimize power, then birds should fly at
precisely the minimum power speed. Alternatively, if flight speed is
designed to maximize the range, the birds should fly at the maximum
range speed. We can test these hypotheses by comparing the
predicted flight speeds with the observed flight speeds.
Figure 11.1 The power expressed as multiples of basal metabolic
rate (BMR) required to sustain level flight of different speeds by a
bird. Very slow and very fast flight speeds are very costly. Vmin
corresponds to the speed that requires the least power. Vmax
corresponds to the flight speed that covers the greatest range per
unit energy spent. Both are potential optimal flight speeds that are
subject to empirical testing.
If we found that dark-eyed juncos flew at the minimum power speed
calculated from its mass and wing shape we could legitimately
conclude that the junco’s flight speed is an adaptation to reduce its
flight costs. But it is more important to agree about what it means if
the bird does not fly at either of the predicted optimal flight speeds,
or more formally when the hypothesis is rejected.
What does it mean when a hypothesis about adaptation
is rejected?
Some may argue that the negative results should be taken as
evidence that flight speed simply has no function; it is not an
adaptation so that we should give up the quest and move on to
something else. The problem with that view is that not only does it
simply end the enquiry, but it is difficult to be certain that the speed
does not serve any function after only one or two rejected
hypotheses. After all, flight speed could serve some other adaptive
function that remains to be investigated. Because of that, the absence
of adaptation can only be invoked as a last resort, once most
reasonable alternatives have been found to be wrong. When a
model’s predictions are wrong we must first be certain that the
assumptions upon which the predictions were based are correct.
Only if all the assumptions happen to be correct can we clearly reject
the hypothetical adaptive function we were testing. If we do reject it,
then the next step is to come up with an alternative functional
hypothesis that will also be subject to testing.
This position of constantly hypothesizing adaptation for traits has
given evolutionary and behavioral ecologists a bad name (Gould &
Lewontin 1979). Because of it, behavioral ecologists are often
depicted as naive optimists, arguing that everything must and does
have a purpose. However, it is important to understand that to study
adaptation, and hence function, it is necessary to hypothesize
adaptation as a starting proposition. This is not decreeing that
adaptation exists but rather stating that it is a hypothesis open to
scientific scrutiny. Behavioral ecologists do not believe that all traits
are adaptive; in fact, it is likely that many traits are neutral or have
no current function. However, we all agree that any claim for
adaptation must be subjected to empirical testing (Williams 1966).
Concluding that a trait has no adaptive value is simply an absolute
last resort because it ends any further functional investigation.

Optimality Models In Foraging

One of the most successful applications of optimality theory to the
study of behavior lies in what is known as Optimal Foraging
Theory (Stephens & Krebs 1986; Giraldeau & Caraco 2000). Indeed
the very birth of this theory can be traced to the pioneering work of a
few ecologists such as Robert MacArthur and Eric Pianka (1966) who
realized that the answers to a number of questions in population
ecology required some knowledge of animal behavior. That thought
was a radical departure from the way ecologists then considered
foraging behavior: they assumed that animals eat what they find and
stop when they are sated or when the food is gone. But is it
reasonable to expect that natural selection would have optimized the
shape of wings and flight speed, a bumblebee’s tongue length, the
architecture of a spider’s orb web, and the morphology of a fish’s
mouth while leaving untouched the very behavior that allows the
collection of the energy required to live and reproduce? MacArthur
and Pianka thought it was not and proposed an optimality analysis of
diet choice for predators that encounter their prey sequentially. Since
then the optimality approach they pioneered has been applied to the
diet selection of large herbivores that must deal with toxins,
predators that encounter prey simultaneously, or even that ambush
prey rather than search for them. Reviewing all these different
optimal diet models would require a book in itself and goes beyond
the objective of introducing the optimality approach (but see Sih &
Christensen 2001). So, we deal here only with the simplest sequential
prey encounter optimal diet model to illustrate the approach.

To Choose or Not to Choose: Optimal Diet Models

A black-capped chickadee (Poecile atricapillus), a territorial small
passerine bird that inhabits the forests of North America, is busy
collecting insects to feed itself and its noisy and growing nestlings
during the late spring. Let’s imagine that there are only two species
of insects at the moment in the forest: large, rare, and nutritious
insects and small, common, but less nutritious ones. An animal that
takes what it finds will eat each type of insect in proportion to their
relative abundances in the environment. The less nutritious one
being more common, it should be more common in the parent’s and
the nestlings’ diets. But could the chickadee do better by being more
selective? Could it nourish itself and its young more effectively by
exercising some choice over the items it eats and brings back to the
nest? If it could do better by choosing then should we not expect that
selection would have favored choosy individuals over those that were
not? And if this were the case, then how should the frequency of prey
captured by a forager be affected by the preys’ relative abundances?
To answer these questions we must use backward engineering and
build an optimality model.
Three steps to building the optimal prey model
Step one
The first thing we need to do is make explicit the alternative courses
of action that we wish to analyze. In the case of sequential prey
choice, it is: given an encounter with prey item x, should the
predator attack, capture, and eat prey x, OR should it ignore the
prey and continue searching?
Step two
The second thing we need to do is come up with a way to compare
the fitness consequences of the two alternative courses of action. To
do this we formulate a hypothesis concerning the adaptive function
or survival value that we can assign to each alternative. One
possibility is that the faster the rate at which a bird harvests food the
greater its fitness. This makes sense given that greater harvest rates
would allow nestlings to grow faster, fledge sooner, and be healthier
when they fledge. So, let’s hypothesize that the function of prey
choice is to maximize the rate of prey harvesting. Surely fitness is not
just correlated with the number of prey harvested but must also be
affected by their quality. Assigning units to prey quality is not simple.
Nutritive value may be important but it is difficult to quantify in a
single axis as it involves proteins and lipids, carbohydrates, minerals,
and a number of other nutrients. However, nutrition above all fuels
the animal’s metabolism and the offsprings’ growth. So let’s assume
that energy is a good approximation of nutritive value and
hypothesize a function: prey choice has been designed to maximize
the rate of energy intake (E/T) with Joules per time spent foraging as
units. In the behavioral ecologist’s jargon, rate of energy intake is the
currency of fitness or adaptive function that is under
experimental scrutiny. This is the currency with which we will be
able to compare the adaptive value of alternative courses of action
such as being choosy or not.
Step three
The third and last thing we need to do is specify the assumptions and
constraints under which we assume the animal makes its decision
(choice). These constraints often take the form of assumptions
required to keep the model realistic, simple, and general. The
sequential prey encounter model makes a number of assumptions
about the world in which the animal lives. It assumes that energy (E)
can only be acquired by consuming a prey item and that consuming
an item requires some handling time (H). The profitability of a
prey type is set by the ratio E/H and all individuals of a prey type
offer the same exact profitability. E and H are characteristics of the
prey and the forager cannot do anything to change these prey
characteristics (e.g., handle them faster or get more energy from one
by eating it differently). The model assumes that each prey type is
encountered randomly during search (S) at some rate λ that depends
on its abundance in the environment. The model also makes
assumptions about the foraging predator. It assumes that it knows
what encounter rates to expect from each prey type, that it is capable
of recognizing each prey type instantly without error upon
encounter, that it cannot search while it is eating a prey item, and its
search efficiency and speed remain constant and unalterable.
These assumptions may seem peculiar, possibly even unrealistic.
However, it is impossible to capture all the elements of the real world
in a single quantitative model. It would make the mathematics very
complicated such that it would require a model to study the model!
Modelers secretly hope that not every element present while the
animal is behaving is important. So, in trying to capture the essential
features of the problem, they discard some elements, simplify others,
and ignore many. To do this successfully, however, it is essential to
have a good idea of the biology behind the problem being addressed.
Not too surprisingly, models often oversimplify and this is why,
when a model fails, instead of flatly rejecting the adaptive hypothesis
behavioral ecologists first turn to scrutinizing the validity of the
assumptions.
Two steps to testing the optimal prey model
Step one: Formulating the predictions
Now that the decision, the currency, and the main constraints are
known we can formulate the predictions that will be subjected to
testing. We predict that if E/T is the real currency of fitness then the
animal should adopt the diet that provides the highest value of E/T.
So we need to calculate the E/T outcomes of all possible diets. For
simplicity we assume only two prey types exist (the same argument
and conclusions would apply to any number of prey types). In such a
two-prey system there are just two reasonable outcomes: take all
prey as encountered (no choice) or always attack the most profitable
item (prey type 1) and never attack the less profitable item (prey type
2). An animal that does not choose behaves as a generalist and the
one that chooses only the best item is a dietary specialist.
To compare the value of each option we must calculate their
respective adaptive value or function using the currency of fitness.
We calculate the E/T that corresponds to a generalist diet by
calculating first the numerator E. After S time spent in search a
predator will have encountered Sλ1E1 prey of type 1 and Sλ2E2 prey of
type 2. Its overall energy harvest will be:

To estimate T, the denominator, we must account for the time spent

in each activity during foraging. Each time the predator eats prey
type 1 it spends H 1 time handling it (we assume attack time is zero).
So, in total it will have spent Sλ1H1 time handling prey type 1 and
Sλ2H2 handling prey type 2. To this handling time we must add the
time spent searching such that the total amount of time spent is:

Placing numerator over denominator we get:

which corresponds to the adaptive value of a generalist diet. Now we

can do the same for the specialist diet, which gives:
When does being a specialist provide a higher currency of fitness that
being a generalist? When:

which, after some algebra simplifies to:

When this inequality holds, the forager should be a generalist and

accept the less profitable prey when it encounters it. When it does
not hold, the forager should exclude the less profitable prey and
specialize only on the most profitable. This inequality makes a
number of rather interesting predictions that follow from the
hypothesis that choice is designed to maximize E/T.
The predictions

1. The relative abundances of each prey type should not affect the
choice. Only the encounter rate with the most profitable prey
(prey type 1) remains in the final equation. That means that
whether prey type 2 gets included in the diet does not depend on
its own abundance but on the abundance of the other, more
profitable item. The more prey type 1 is abundant, the more
likely prey type 2 will drop out of the diet.
2. The optimal diet policy does not depend on the duration of the
search time, because the variable S falls out during
simplification. That means the same policy holds whether an
animal searches just briefly or for long foraging periods.
3. The optimal diet does not allow partial preferences; a prey type
is either always accepted or always rejected.

Step two: Experimental testing

The main difficulty with testing this model is obtaining an accurate
estimate of the forager’s encounter rates with the prey types. John
Krebs and his colleagues at Oxford University devised an ingenious
experimental apparatus to overcome this difficulty (Krebs et al.
1977). They placed small captive wild birds (Great tits, Parus major)
in an experimental apparatus where they could obtain mealworms by
perching above a small window giving access to a conveyor belt upon
which prey of each type appeared at experimentally set rates. Hence,
the experimenters could set the encounter rates for each prey type in
this microcosm. Once a prey item was attacked and held in the bill,
the bird had to fly away from the window to a perch in the rear of the
apparatus in order to eat it. In this way the bird could not handle and
search at the same time. The birds encountered two types of prey
sequentially (only one prey type appeared in the window at a time)
on the conveyor belt, short four segment and long eight segment
mealworms. The handling times were similar for both types of prey
and so the short mealworms were less profitable than the long ones.
The results of the conveyor belt experiment were remarkable (Figure
11.2). They provided the first experimental evidence that relative
prey abundances were irrelevant to the animal’s choice; choice was
independent of the abundance of the least profitable prey type. No
matter how many short mealworms were available, if the long
mealworms were sufficiently common the shorter ones would simply
be rejected… almost always. Although the birds followed the
qualitative predictions of the model they nonetheless exhibited
partial preference rather than the all or nothing choice predicted by
the model. This deviation from expectations is not really surprising
and several hypotheses, based on reappraisal of the initial constraint
assumptions have since been formulated to account for the result.
For instance, it may not be correct to expect that foragers
instantaneously and without error always recognize each prey type
upon encounter. Also, maybe predators expect prey types to change
in quality over time and so sample them once in a while to update
their information (Getty & Krebs 1985).
Figure 11.2 The results of Krebs et al’s (1977) conveyor belt
experiment. Great tits (Parus major) could pick pieces of mealworms
from a moving conveyor belt. The pieces were either long or short,
corresponding to most and least profitable items, respectively. The
birds were exposed to four prey densities corresponding to the four
columns of the histograms. At low prey densities they encountered
both prey types equally. The diet model predicts unselective foraging
and that is what the birds did. When the density of the profitable
mealworms was increased, the model predicts they should now
specialize on only the longest more profitable ones. The results show
that they almost always chose to ignore the least profitable items.
When density of long items was kept constant but the density of less
profitable short items was increased and made equal to the density of
good items, the model still predicts specialization and the birds
continued to take almost exclusively only the most profitable longest
items. Finally, keeping the density of good items constant but
making the density of the least profitable items double the density of
good items still predicts that animals should specialize. The birds
again conform closely to this prediction. From Krebs (1978)
Sih and Christensen (2001) provide a review and appraisal of the
optimal diet approach. Their analysis includes 134 published studies
that were either experimental or observational, conducted in the
laboratory or the field. They conclude that the optimality models
have worked well for systems in which predators exploit immobile
prey. The models, however, do poorly when prey are mobile because
the current models do not take into account the effectiveness of prey
evasion strategies on the value of pursuing alternative prey. The
current trend at the moment is to include these factors in optimal
prey models. For instance, we will see in Chapter 12 that prey that
have detected their predator may be more difficult to capture. So, it
pays these prey to communicate to predators that they have been
detected, indicating that their profitability as prey has just declined
compared to other prey that have not yet detected the predator (see
Getty 2002).

Foraging Theory: Patch Models

Organisms are rarely uniformly distributed in space. As a
consequence an animal’s food often occurs in patches: clumps of
eggs, fruit in a bush, flowers on a plant, etc. When resources are in
clumps, what should the optimal patch exploitation strategy be?
Should foragers consistently deplete clumps, or exploit them only
partially? What is the function of patch exploitation decisions? What
ecological factors affect this decision? Consider the following
example: a bumblebee takes nectar from a flower. Each time its
tongue is inserted in the nectary (the structure that holds nectar) it
draws a bit less nectar: the nectary is depleting. At some point the
bumblebee must give up the current flower and search for another,
hopefully fuller one. At what point during flower exploitation should
this occur, or in terms of the currency of fitness we used for prey
choice, at what point during the exploitation of a patch would
giving it up maximize long-term rate of energy intake? Let’s use
backward engineering once again and construct an optimality model
to analyze the decision and predict the optimal patch departure time.
This model is known as the patch model and as all optimality models
it has three parts.
Three steps to building an optimal patch residence
model
Step one: The decision
The decision in this model is: whether to continue searching for
prey in the current patch OR cease exploitation and begin searching
for the next patch. What is being traded off here is current intake
rate against a potential intake rate achieved by foraging at a better
fuller patch, taking the time required to travel this patch into
account.
Step two: The currency of fitness
As for the optimal diet model above, we must choose a currency of
fitness with which to compare the adaptive values of all possible
patch exploitation times. For the same reasons as for the diet model
let’s adopt E/T as a reasonable currency of fitness.
Step three: The constraints
The model assumes that all prey items are randomly distributed
within identifiable patches; no food can be found between patches.
The time required to uncover patches is set by patch density in the
habitat, so that as patch density declines the time required to find a
patch (T) increases. A forager can only gain energy while spending
time (P) in a patch. The density of items in the patch sets the
forager’s encounter rate with prey and, because the density declines
as exploitation progresses, so does the forager’s encounter rate with
prey.
The model also makes assumptions about the forager. It assumes
that the forager has no say in the way its rate of encounter with
patches and its encounter rate with prey within a patch progresses.
The model assumes that a forager does not know beforehand the
precise time it will take to reach the next patch nor does it know
whether the patch it will find will be rich or poor. The decision must
therefore be based on the average time to reach a patch in a given
environment and the average patch quality expected in that
environment. The model also assumes that a forager cannot estimate
the actual quality of the patch it is currently exploiting and behaves
toward it as if it was a patch of average quality for that habitat.
Two steps to testing the optimal patch residence model
Step one: Drawing the predictions
The patch problem consists of looking for an optimal patch residence
time that corresponds to the maximum rate of energy intake for a
habitat characterized by a given density of patches of known average
quality. The predictions, therefore, will concern how a change in
patch density or a change in average patch quality in a habitat will
affect the extent to which resource clumps of that habitat will be
depleted by foragers. These predictions are best obtained graphically,
and the method is illustrated in
The predictions

1. When the patch density declines and hence the average travel
time between patches increases, animals should spend longer
foraging within patches and deplete them to a greater extent.
2. When average patch quality in a habitat increases, foragers
should spend less time exploiting each patch and deplete each to
a lesser extent.

Step two: Experimental test of the prediction

Like the diet model presented above, there have been numerous tests
of the patch model's predictions (see Stephens & Krebs 1986). In
most cases experimenters have found a qualitative fit for both
predictions listed above (Nonacs 2001). A typical example of such
studies involves the collecting of seed loads by eastern chipmunks
(Tamias striatus), a ground-living squirrel indigenous to the east
coast of the northern United States and Canada. Chipmunks are
central place foragers because rather than eating food where they
find it, they collect seeds (maple samaras, acorns, beech nuts, etc) in
extensible cheek pouches and carry them back to their underground
burrow to be eaten later, often during winter. Seed loading by
chipmunks while in a clump occurs at a decelerated rate as the cheek
pouches fill and adding prey becomes more difficult. If the
chipmunk’s patch exploitation decision is an evolutionary adaptation
to maximize its rate of seed delivery to the burrow, then as the
distance between burrow and patch increases the chipmunk should
spend more time in the patch to collect larger seed loads (Box 11.1).
Box 11.1 Graphical representation of the optimal solution
to the patch model

In this graph the x-axis runs in two directions from the y-axis. To
the right, patch time increases. To the left, travel time increases.
The curve gives the cumulative increase in energy intake as the
animal spends time in the average patch. Note that as the animal
spends longer and longer in the patch its instantaneous rate
declines. This is the result of patch depletion. There are two
habitats represented here, one with short travel times and the
other with long travel times, each one represented on the x-axis.
When the animal forages in the habitat with the short travel time
it can leave at any point on the exploitation curve. The problem
consists in finding the point that provides the maximal long-term
intake rate. The intake rate will have units E divided by the sum
of travel time and the chosen patch time. The slope of the line
starting at the short travel time and touching the exploitation
function will have a slope that gives the rate of intake. The slope
of any line linking a travel time to a point on the exploitation
function will give the rate of intake achieved by that patch time.
We are looking for the patch time that corresponds to a line with
the maximum slope. That line has to be the point of tangency to
the curve. Note that the point of tangency for short travel time
corresponds to a short optimal patch time, whereas the longer
travel time leads to a point of tangency at a longer optimal patch
 time. This means that the optimal patch times are predicted to
increase when the average travel time between patches in an
environment increases.

In a field experiment, chipmunks were given the opportunity to

collect loads of sunflower seeds from artificial patches placed at
known distances from their burrow. They were observed to collect
larger loads when the distance to the food patch increased (Figure
11.3; Giraldeau & Kramer 1982; Giraldeau et al. 1994) as predicted by
the patch model. However, the model was also somewhat of a
quantitative disaster given that the load sizes observed to be
collected by the chipmunks were much smaller than those predicted
by the model. This is a clear case of qualitative support and
quantitative mismatch, which required additional constraints to be
added to the model (see McAleer & Giraldeau 2006).
Figure 11.3 Results of Giraldeau and Kramer’s (1982) eastern
chipmunk (Tamias striatus) central place foraging experiment.
Chipmunks were offered trays of sunflower seeds at various
distances from their burrow. They made repeated trips to the same
tray, collecting seeds in their cheek pouches and carrying them to
their burrow where they were stored. The top graph shows the mean
observed time spent in the patch collecting seeds, the middle graph
the weight of sunflower seeds collected, and the bottom graph the
rate of food delivery achieved at various distances and hence travel
times between tray and burrow. The lines on the graphs are the
quantitative predictions of changes in patch time, load size, and the
corresponding rates of food delivery from the optimal patch model
presented in Box 10.1 using three different estimated exploitation
functions. The chipmunks did collect larger loads and spent more
time doing this as distance and hence travel time to the burrow
increased. The rates of food delivery achieved were generally lower
than those they would have achieved had they taken the larger loads.
From Giraldeau and Kramer (1982)

Expanding Foraging Models

A qualitative fit but quantitative failure is a common result in
optimal foraging tests, which has led to closer scrutiny of the various
assumptions of the model. For example, honeybees have a nectar
crop that can be filled while the bee is foraging. Schmid-Hemple et
al. (1985) found that the bees did not fill up their crops as full as
would be expected if they were maximizing rate of energy intake.
Instead, it turned out the bees were maximizing energy efficiency. On
their return to the hive, the weight of the nectar affects energy
consumption much more in the bee than the weight of the insects or
grain does for birds or rodents returning to their nests or burrows.
Sometimes simple consideration of a species’ biology can suggest
different assumptions or additional constraints for the model, but
sometimes the constraints are very unexpected.
Theunis Piersma and his colleagues have carried out a series of field
and laboratory studies on the foraging of red knots, Calidris canutus,
over a period of more than 20 years (see Piersma & Van Gils 2011,
for a review of these studies). Knots are medium-size shorebirds that
breed in the Arctic tundra and over-winter in various places further
south, including the Dutch Wadden Sea. There, they forage during
low tide on bivalve molluscs and various crustraceans that live
buried in mudflats along the coast (see Figure 11.4). They swallow
their prey whole and crush the hard shells in their gizzard, a strong
muscular stomach. Early studies on knot foraging found that the
birds were feeding at a slower pace than expected and were not
maximizing rate of energy intake as predicted by the patch residence
model. But foraging knots face a special problem due to the
composition of their prey, hard-shelled molluscs. Their gizzard,
necessary for crushing the shell and extracting the energy-rich meat,
can only crush shells at a certain rate, which is slower than the
foraging rate. So, when the stomach is full the birds need to wait for
some digestion to take place before they can recommence eating.
When the constraint of digestion time is added to the model, the
birds are found to be maximizing rate of energy intake.

Figure 11.4 Red knots (Calidris canutus) foraging in the Wadden

Sea. Courtesy of Jan van de Kam.
Foraging knots also posed a problem for the patch choice model. As
well as eating hard-shelled cockles (Cerastoderma edute), they also
eat soft-shelled shrimp (Crangon crangon). The cockles and shrimp
generally occur in separate patches, and energy calculations show
that cockles are more profitable than shrimp with respect to rate of
energy intake and inter-capture interval. But some birds choose
shrimp patches over cockle patches, and captive birds in the
laboratory always prefer shrimp. It turns out that the gizzard is the
problem again: it changes size over the course of the year, and a
small gizzard digests shrimp more efficiently than cockles. One more
constraint then, gizzard size, needed to be added to the model.
Red knots are actually a group of at least six subspecies, one of which
over-winters in the Netherlands. Another subspecies over-winters
along the west coast of Africa, and its foraging behavior has been
studied at the Banc d’Arguin in Mauritania. The prey species in
Mauritania are different from those in the Netherlands, and the two
main prey types are the molluscs Loripes lucinalis and Dosinia
isocardia. Based on energy measurements of the prey species and
results from the Wadden Sea in the Netherlands, the optimal
foraging model predicted the birds should choose only Loripes. In
fact, the knots actually chose a mixed diet of the two mollusc species.
In this case, the confounding factor turned out to be a consequence
of the peculiar metabolism of Loripes: sulfide-oxidizing bacteria that
cause a toxic effect when ingested excessively. When the constraint
of toxicity was added to the model, the knots were found to be
maximizing energy intake rate (Oudman et al. 2014.

Game Theory and Evolutionarily Stable

Strategies

The optimality models presented up to now were simple in the sense

that it was possible to compute the outcome of all strategies available
to the animal (the various diets in the prey model, or all possible
patch residence times for the patch model) without reference to the
behavior used by other individuals. Sometimes, however, the
benefits derived from a given course of action cannot be specified
without also specifying the frequency with which the other animals
use that alternative. This section considers three behavioral systems
in which the payoffs are characterized by this sort of frequency
dependence: choosing (i) fighting displays; (ii) resource harvesting
strategies; and (iii) where to settle or forage.

The Function of Fighting Displays

In a contest over resources individuals often engage in aggressive
encounters. Bald eagles fight over access to the carcasses of dead
salmon; male dragonflies chase male intruders from their territory
and maintain a monopoly over females that come to lay their eggs
there. What is the function of behavior expressed during fighting? Is
the behavior meant to inflict injury upon opponents or does it serve
some other purpose? Except for some of the fighting scenes in Walt
Disney’s classic Bambi, male deer rarely use their antlers to inflict
injury by ramming each other by surprise, or in any other place but
the other male’s antlers. Male deer (and other ungulates) that fight
with their antlers use them in rather stereotyped fighting displays:
they slowly and purposefully interlock them before they engage in an
all out pushing match at the end of which one male leaves promptly
as the other chases it away. Similarly, when cichlid fish are engaged
in a conflict over a territory, or access to a female, two highly
valuable resources, they interlock jaws and push and pull, until one
gives up, conceding victory to the other without inflicting any injury
despite their razor-sharp teeth. When early ethologists such as
Konrad Lorenz and Niko Tinbergen described fighting displays they
were struck by the fact that very little blood if any was drawn during
fights, a stark contrast with the outcome of aggression among people.
They assumed, as many did in those days, that noninjurious fighting
displays evolved by group selection and that their adaptive function
was to avoid the useless waste of individuals during animal fights.
We now know that it is unlikely that group selection could have been
strong enough to favor such widespread occurrence of noninjurious
fighting displays (Williams 1966). The challenge that is put to
behavioral ecologists therefore is what is the adaptive function of
noninjurious displays given that they must have arisen by selection
acting at the level of individuals?
John Maynard Smith and G.R. Price (1973) met this challenge by
applying a mathematical technique known as game theory to the
evolutionary problem of aggression. Game theory existed in
economics and was used to analyze problems in trade, auctions, and
even war. However, to apply this mathematical tool to evolutionary
problems, it was important to realize that solutions to the game
depend on the evolutionary process: these solutions need to be the
most likely evolutionary outcome of the game. The stage was set for
one of the first formal analyses of an evolutionary game.
Can noninjurious fighting displays exist as the most plausible
evolutionary outcome when selection acts to maximize the fitness of
individuals and not species? Shouldn’t an individual designed to
maximize its fitness go all out and risk injury by attempting to injure
its opponent? To answer the question we must first specify the payoff
derived from using either strategy. The payoff provided by each
strategy depends on whether the opponent is likely to use injurious
or noninjurious levels of fighting itself. Hence simple optimality
modeling is of little utility here and we must use game theory to
analyze the problem and predict the most likely evolutionary
outcome.

Three Steps to Analyzing the Fighting Game

Step one: Specifying the alternative strategies
To analyze a game we must first specify the alternatives open to the
players. We can simplify the options and place them in two discrete
categories: a noninjurious option called Dove and an all-out
escalating option called Hawk. A Dove displays until it or the
opponent gives up or until the opponent escalates to injurious levels
of fighting. A Hawk always uses escalated fighting and leaves only if
injured.
Step two: Specifying the payoffs to each alternative
We will not specify whether the animals are fighting for food, mates,
shelter, or whatever. Let’s keep our analysis general and so, instead
of a currency of fitness, we will use fitness directly to measure the
payoffs of both strategy. To calculate the payoffs of each alternative
we must first specify the value (V) of the resource that is sought in
terms of its replacement cost expressed in fitness units. If the
animals were fighting over food, then the same food resource would
have greater value for a starving individual compared to a well-fed
individual because the starving individual would incur a greater risk
of death by having to replace the lost food than the well-fed
individual. Individuals that get injured in a fight suffer a cost (C),
also in fitness units. To keep the analysis simple, we assume that the
players are of equal strength and each value the resource equally.
Technically this means the players are “symmetric” so that we can
represent the payoffs of individuals strictly by what they and their
opponent play.
No matter what strategy one plays, the payoffs from fighting against
Dove will be different from the payoffs of fighting against Hawk.
There are four possible combinations of strategies. The payoffs of a
Dove that encounters a Hawk will be written as E(D,H), so that the
other three remaining combinations are: E(D,D), E(H,D), and
E(H,H). We can analyze what happens in each one of these
encounters.
When a Dove meets a Dove. When two Dove individuals meet they
both use a noninjurious display. Because the players are symmetric
let’s assume that they each win encounters half of the time. So,
E(D,D) = 0.5 V.
When a Dove meets a Hawk. Dove always runs away when the Hawk
escalates. The Dove avoids injury but never obtains the resource. So,
E(D,H) = 0.
When a Hawk meets a Dove. Hawks always win against Doves
because Dove always avoids escalated fighting by running away when
the opponent escalates to injurious levels. So E(H,D) = V.
When a Hawk meets a Hawk. When two Hawks meet we assume
that one wins half the time and the other wins the other half because
the players are symmetric. Because a Hawk only leaves if it gets
injured the loser suffers a fitness costs (C) and the winner gets the
resource (V). So E(H,H) = 0.5 V + 0.5 C = (V + C)/2. The four
potential payoffs can be expressed in a matrix (Table 11.1 )
Table 11.1 Payoffs to the row player engaged in a hawk–dove game
Hawk Dove
Hawk E(H,H) = (V + C)/2 E(H,D) = V
Dove E(D,H) = 0 E(D,D) = V/2
Step three: Finding the expected evolutionary solution
The solution to an evolutionary game is known as the Evolutionarily
Stable Strategy, or ESS (Maynard Smith & Price 1973; Maynard
Smith 1982). An ESS is a strategy such that, when all members of a
population adopt it, no mutant strategy can invade. An evolutionary
thought experiment comes in handy to find the ESS in the Hawk–
Dove game. Imagine an ancestral population of individuals that all
play Dove. Can a rare mutant that plays Hawk invade this
population? Because the population is all Dove, the Hawk mutant
meets only Dove players and so will do well because E(H,D) >
E(D,D) (or V > V/2 from the payoff matrix). Hawk, therefore, can
invade a population of Doves. But the argument does not stop there.
It is not enough to know that Hawk can invade a population of
Doves, we also need to know whether a population of Hawks can
resist invasion from rare Dove mutants. Or, more formally, E(H,H) >
E(D, H)? Using the values in the payoff matrix, the Dove will be
unable to invade Hawks when (V-C)/2 > 0. As long as C < V, the
costs of injury are smaller than the replacement value of the
resource, Hawk will remain the ESS to the game. So we conclude
from the analysis that a population of individuals engaged in
noninjurious fighting is simply not an evolutionarily stable solution.
What happens when C > V?
When C > V, Hawk is no longer the ESS. Under that condition rare
Dove mutants do better in a population of Hawk than do the Hawks
(E(D,H) > E(H,H)). That happens because the Hawks are inflicting
costly injuries to each other while Dove never gets injured (even
though it never gets the resource). Even though rare Dove mutants
do well, they cannot form an ESS because as we saw above there is
no condition under which Dove can ever be an ESS; E(H,D) >
E(D,D). In this case, neither Hawk nor Dove can be evolutionary
solutions to the game. Each time the population evolves toward all of
one strategy, the alternative strategy can invade. In this case there
must be a combination of Hawk and Dove that is an ESS. Going back
to our evolutionary thought experiment we start once again with a
population of all Dove in which a mutant Hawk arises. This mutant
does much better than Dove and so it spreads. However, as it spreads
in the population, it starts encountering more and more of its own
kind and so suffers an increasingly important injury cost. At some
point the average payoff of Hawk declines until it reaches the payoff
to Dove. When the payoff to Hawk and Dove are the same, selection
can no longer favor one over the other and no further change in the
frequency of alternative fighting strategies is possible: evolution is
stuck! This happens whenever the population reaches an ESS
mixture of the two strategies. We can use the fact that at the ESS the
fitness payoffs to both strategies must be equal to calculate the ESS
frequency of Hawk (p*). The proportion of Hawk strategists in a
population is p and Dove is 1-p. At ESS:

placing the payoffs from the matrix in their appropriate places gives:

which after some algebra and simplification gives:

This means that the ESS proportion of Hawk in the population

depends on the ratio of the value of the resource to the costs of
injury; the greater the ratio, the more Hawks are expected at
equilibrium.

Lessons from the Hawk–Dove Game

The results of the analysis are a little disappointing. The Hawk–Dove
game predicts that Dove, the noninjurious display we are trying to
account for, is never an ESS. Moreover for Hawk, injurious fighting
is the ESS when the costs of injury are less than the value of the
resource. If Dove exists at all, it can only be in a mixed ESS where the
frequency of Hawk is set by the ratio of the value of the resource to
the cost of injury. Noninjurious displays are much more common
than this and so there must be some important element here that
was ignored by the Hawk–Dove model.
In the real world, players are unlikely to be as equal in fighting ability
as those imagined in the simple Hawk–Dove game. It turns out that
when players are asymmetric, the predicted ESS is quite different. In
Box 11.2 we modified the Hawk–Dove game to allow for a third
strategy called Assessor that assesses the fighting ability of its
opponent relative to itself. Assessor plays Hawk when it assesses to
be stronger and Dove when it assesses to be weaker. The payoff
matrix of the Hawk, Dove, Assessor game in Box 11.2 shows that so
long as there is a cost to being injured (C > 0) Assessor is the only
ESS. When asymmetries occur, the ESS fighting strategy is to assess
asymmetries reliably. Now it so happens that a population composed
of Assessors would almost never use injurious fighting. All the
individuals would be engaged in assessing each other’s strength until
one contestant, having realized it was the weakest and hence most
likely to be injured, played Dove and abandoned the contest.
Injurious level fighting would occur only in those unlikely cases
where opponents are so closely matched in fighting ability that they
would be incapable of assessing which is the weaker individual.
Box 11.2 Hawk–Dove–Assessor game:when players are
different
Say a contest involves asymmetric players, either because the
resource represents a different value for the players or because
players differ in fighting ability. A strategy called Assessor
compares the values of the resource for both contestants or their
fighting abilities and plays according to the asymmetry: it plays
Hawk if it values the resource more or has higher fighting ability
and Dove if the opponent values the resource more or has higher
fighting ability.
Defining Payoffs
Since we have already gone through the explanations of the
payoffs to the Hawk–Dove game we only consider here the
interactions that involve the Assessor strategy.
When Hawk meets Assessor, E(H, A)
For simplicity let’s assume that half the time the individual
playing Hawk has greater fighting ability (or values the resource
more) than the individual playing Assessor. When the individual
playing Hawk is stronger, its Assessor opponent assesses this and
plays Dove. The individual playing Hawk then gets all the
resource. When the individual playing Hawk is weaker, however,
the Assessor knows this and now plays Hawk. Because the
Assessor individual is stronger, it always wins and inflicts an
injury on Hawk. The average payoff of Hawk against Assessor is
therefore / V + / C = (V + C)/2.
1
2
1
2

When Dove meets Assessor, E(D, A)

Let’s assume that half the time Dove has greater fighting ability
(or values the resource more) than Assessor. When this is the
case Assessor plays Dove but, unlike with Hawk strategies,
because there is no actual fighting, the asymmetry has no
consequence on the outcome. So, half the time Dove wins and
half the time the Dove-playing Assessor wins, so the payoff to
Dove when it is the stronger of the two is ½(½ V).
In the other situation Dove has less fighting ability (or values the
resource less) than Assessor. Assessor then plays Hawk and Dove
loses the contest and gets 0. So the average payoff is
1
/2(1/2V ) + 1/20 = V /4.

When Assessor meets Assessor, E(A, A)

In this case half the time one individual, call it ego, has greater
fighting ability (or values the resource more) than the other, call
it alter. Ego plays Hawk, alter plays Dove so that ego gets V. The
other half of the time, ego has lower fighting ability (or values the
resource less) than alter so that ego plays Dove and alter plays
Hawk. In this case ego gets 0 (and hence avoids the costs of
injury). The average payoff in this case is / V + / 0 = V /2.
1
2
1
2

When Assessor meets Hawk, E(A, H)

When Assessor has greater fighting ability (or V) than Hawk,
both play Hawk but Assessor always wins. When Assessor has
lower fighting ability than Hawk, Assessor plays Dove and gets 0.
The average payoff is then / V + / 0 = V /2.
1
2
1
2

When Assessor meets Dove, E(A,D)

When Assessor is the stronger of the two, it plays Hawk and gets
V against Dove. When Dove is the strongest, Assessor then also
plays Dove but because they do not really fight the outcome is
independent of the asymmetry. So each wins the resource half the
time. The average payoff to Assessor against Dove is then ½V +
½ ½V = 3/4V.
The payoff matrix of a game between Hawk, Dove, and Assessor,
in a situation where Assessor has perfect information about any
asymmetry, is shown below.
Analysis of the Game
We know from the Hawk–Dove game that Dove can never be an
ESS. The question that needs to be answered is whether Hawk
can still be an ESS against Assessor, i.e., is E(A, H) > E(H, H)?
Using entries in the payoff matrix we see that V/2 > (V − C)/2
and so Assessor can invade Hawk. Clearly, Hawk is not an ESS in
 an asymmetric game so long as the cost of injury is nonzero. So
yes, Assessor can invade Hawk.
We know that Hawk is not a pure ESS against Assessor but this
does not mean that Assessor is an ESS. Can Assessor resist
invasion from Hawk or is E(H, A) > E(A, A)? Using entries from
the payoff matrix we find that this requires that ¼(V − C) > ½V.
Clearly this can never be true given that C > 0, and thus Assessor
cannot be invaded by Hawk.
Can it be invaded by Dove? For this to happen E(D, A) > E(A, A)
or, using entries from the matrix ¼V > ½V, which clearly can
never be true. Hence we can conclude that so long as C > 0, i.e.,
there is a cost of being injured, Assessor is the pure ESS.

In the real world, players are unlikely to be as equal in fighting ability

as those imagined in the simple Hawk–Dove game. It turns out that
when players are asymmetric, the predicted ESS is quite different. In
Box 11.2 we modified the Hawk–Dove game to allow for a third
strategy called Assessor that assesses the fighting ability of its
opponent relative to itself. Assessor plays Hawk when it assesses to
be stronger and Dove when it assesses to be weaker. The payoff
matrix of the Hawk, Dove, Assessor game in Box 11.2 shows that so
long as there is a cost to being injured (C > 0) Assessor is the only
ESS. When asymmetries occur, the ESS fighting strategy is to assess
asymmetries reliably. Now it so happens that a population composed
of Assessors would almost never use injurious fighting. All the
individuals would be engaged in assessing each other’s strength until
one contestant, having realized it was the weakest and hence most
likely to be injured, played Dove and abandoned the contest.
Injurious level fighting would occur only in those unlikely cases
where opponents are so closely matched in fighting ability that they
would be incapable of assessing which is the weaker individual.
This means that noninjurious fighting can be the most likely
evolutionary outcome of fighting games so long as injuries impose a
fitness cost. When deer interlock antlers and push as hard as they
can, they are likely assessing their relative strengths. The
information obtained by using this display is highly reliable because
individuals cannot pretend to be stronger than they really are in a
pushing contest. The same holds for many other displays such as
mouth wrestling in fish. The displays are forceful but noninjurious.
Do you know of other examples that seem to fit this conclusion?

Alternative Resource Harvesting Strategies

There are often many alternative ways to harvest resources and the
game theoretical framework we just saw for fighting can help us
understand how polymorphisms and alternative strategies can be
maintained within populations. For instance, when male natterjack
toads Bufo calamita sing (if we can call it that) to attract females it is
possible for some individuals to acquire females without singing at
all, intercepting them as they travel toward the signing individuals
(Arak 1988). These so-called satellite males (satellite because they
remain on the periphery of singing male territories) obtain females
at little cost compared to the energy invested in calling by the males
he exploits. The same has been noted for singing crickets (Gryllus
integer; Cade 1978). To reproduce, female golden digger wasps
Sphex ichneumoneus dig a burrow, provision it with katydids, and
then lay an egg (Brockmann et al. 1979). Sometimes, instead of
digging her own burrow, she enters and provisions a burrow that has
already been dug up by another female. Sometimes she digs, other
times she enters. Coho salmon males Oncorhynchus kisutch that
spawn in streams on the west coast of North America come in two
forms: large hooknose males that mature after many months at sea
and jacks that mature more quickly but reach a smaller adult size.
Large males use their hooknoses to fight among each other for
opportunities to spawn with females. Small jacks don’t stand a
chance in a fight with a hooknose. They get to spawn nonetheless by
exploiting the hooknoses’ courtship behavior. Once a hooknose male
has courted a female that is just about to shed her eggs, a jack that
has been hiding in the neighboring rocks sneaks into position
between the male and female and sheds his own sperm, some of
which will get to fertilize the female’s eggs (Gross 1985). The list of
examples of such cases of alternative resource harvesting strategies
can be quite long (see Barnard 1984). What is common with all these
instances is that one of the alternatives (silent male toads, female
enterers, jacks) exploits the behavioral investment of the alternative
strategy (calling males, digger females, hooknose males). The
exploiting strategy could not exist without the presence of the
investing strategy. It was Christopher Barnard and his colleague
Richard Sibly who first recognized the generality of this phenomenon
and proposed a general game they called producer–scrounger game
to help us analyze the conditions under which to expect the co-
occurrence of alternative resource harvesting strategies (Barnard &
Sibly 1981).
Payoff to Hawk Dove Assessor
Hawk E(H,H) = ½ (V − C) E(H,D) = V E(H,A) = ½(V − C)
Dove E(D,H) = 0 E(D,D) = ½V E(D,A) = ¼V
Assessor E(A,H) = ½V E(A,D) = ¾V E(A,A) = ½V

The producer-scrounger game

The producer–scrounger game assumes that resources can be
harvested in two mutually exclusive ways: producer, which consists
of investing effort in making the resource available (calling mates,
finding food, building nests, provisioning young etc), and
scrounger, which consists of seeking opportunities to exploit the
resources that have become available as a result of the producer’s
efforts. The game is characterized by strong negative frequency
dependent payoffs: scroungers do very well when they are rare (and
hence producers common) but very poorly when they are common
(and hence producers rare) (Figure 11.5). The scroungers do well
when they are rare because they have the opportunity of exploiting a
large number of producers while suffering little competition from
other animals. However, when they become common the numerous
scroungers must compete for the resources made available by a
dwindling number of producers. Given this negative
frequencydependence what is the ESS solution to the producer–
scrounger game? Scrounger cannot be the ESS because it cannot gain
any resource on its own and would thus go extinct. Producer is not
an ESS because scroungers who take advantage of the many
opportunities for exploitation can invade it. Once again, as was seen
earlier in the hawk–dove game, the solution to the game is a mixture
of strategies that occur at a frequency for which the payoffs to each
strategy are equal.
Figure 11.5 The payoff functions of the producer–scrounger game
expressed as a function of the frequency of the scrounger strategists
in the group. When there are no scroungers in the group the payoff of
producers (black line) is maximal. The payoff to scrounger (dashed
line). indicates that the first few scroungers do much better than the
producers. The evolutionary consequence of this is that scroungers
will increase in frequency and replace producers in the population.
As the scrounger frequency declines, the payoff to scroungers also
declines. At some point the payoffs to scrounger and producer are
equal and the combination of the two strategies qualifies as an
evolutionarily stable strategy (ESS).
Lessons from the producer–scrounger game
The producer–scrounger game is a general game. To make
predictions for specific cases of alternative resource harvesting
behavior it would be necessary to go through the three steps of
building optimality models: stating the decision to be analyzed,
formulating a hypothetical currency of fitness, and listing the
constraint assumptions. The game has rarely been taken to this level
of detail except perhaps for its application to the study of alternative
foraging tactics within flocks of seed-eating birds (Giraldeau &
Beauchamp 1999). The producer–scrounger game has been
qualitatively successful at predicting the change in the stable
frequency of the scrounger tactic within bird flocks (Giraldeau &
Caraco 2000). However, beyond its specific applications the game
provides a general lesson for all those interested in the study of
evolution and adaptation. Whatever the system explored, scroungers
often impose a cost on group living; they take resources away from
producers and do not contribute to the group’s corporate effort at
uncovering resources. The producer–scrounger game helps us
understand that scroungers exist, despite the cost they impose upon
other group members, because that is the only stable evolutionary
outcome. A group of all producers would no doubt do much better;
however, scroungers inevitably invade because a mixture that
includes scroungers is the ESS.
Earlier we saw that one of the critiques aimed at behavioral
ecologists is that they live in a rosy world where every trait seems to
exist for the benefit of individuals. Our analysis of the producer–
scrounger game makes a serious dent in this shiny view of selection
by showing that evolutionary stability is not always synonymous with
maximization of benefits. The solution is expected because it is
evolutionarily stable, not because it is better. Scroungers exist
despite making it worse for all, including themselves, because that is
the only outcome that can be expected from natural selection. So
much for the rosy world of adaptationists!

Predictions for games with behavioral adjustment

The producer–scrounger game, as are most evolutionary games, is
formulated in terms of evolutionary stability. This means that the
alternative strategies equilibrate over evolutionary time as selection
adjusts the frequency toward the equilibrium. In many situations
animals have the behavioral plasticity to adjust rapidly to local
conditions, perhaps using learning (Beauchamp 2000). Behavioral
adjustment is expected to lead to the same solution to a game as
would evolution. In fact, the birds tested in the foraging producer–
scrounger game can adjust to a new equilibrium scrounger frequency
within a few hours (Giraldeau & Livoreil 1998). When animals can
adjust so rapidly the stability of the frequency is not really
evolutionary. This is why behavioral stable strategy may be a
more appropriate expression for solutions to evolutionary games that
are solved by behavioral assessment (Morand-Ferron & Giraldeau
2010).
The simple optimality diet and patch models we presented earlier
predict the behavior expected of each individual in a population. If
they all forage under the same conditions, all the chickadees should
exhibit the same prey choice. This is no longer the case for
predictions that follow from a game theoretical analysis. The game
can predict the co-existence of two strategies within a population but
it doesn’t predict the behavior of a given individual in the population.
For example, a game analysis of producing and scrounging could
predict that the stable frequency of scrounger is 75%. That level of
scrounging can be achieved by having 75% of the population always
playing scrounger and 25% always playing producer: a
polymorphism. Or, it could occur by having all individuals playing
scrounger 75% of the time and producer 25% of the time. Or, it could
have a range of individuals each playing different combinations of
producer and scrounger that happen to average out over the
population to 75% scrounger and 25% producer. All of these
solutions are equally stable. But only one of them (can you figure out
which it is?) actually predicts the behavior expected of each group
member; all others merely specify the group mean.

Choosing Where to Live or Forage: The

Ideal Free Game
Above, in the section on patch models we mentioned that organisms
were rarely uniformly distributed. They occur in aggregations.
Sometimes these aggregations result from the benefits derived from
being with others (see Chapter 14). Sometimes, however, there are
no advantages and perhaps even competitive costs of being close to
others, a situation known as a dispersion economy. We can ask in
these conditions what selective pressures have led animals to adopt
the distribution they currently have. The answer requires some
evolutionary game theoretic analysis because if all animals go to the
best place, then competition can become so intense for resources
there that it is better to go elsewhere, somewhere less suitable but
with less competition. This distribution decision has been analyzed
by a game known as the Ideal Free Distribution (IFD), by far the
most successful theory produced by behavioral ecology.

The ideal free distribution game

Stephen Fretwell (1972) proposed a simple model that can allow us
to analyze an animal’s decision to establish in one habitat or another.
Inspired by the theory of perfect gases he simply assumed that
animals:

1. know the value of all alternative habitats (i.e., they are ideal).
2. are free to go in the best habitats (i.e., they are free).

If animals are ideal and free they will always go instantaneously to

the habitat with the highest current value. If all animals do this then
all movement will cease once the experienced value in each habitat
becomes equal; no improvement is possible when this happens.
Knowing this, in a world of k individuals having to colonize two
habitats, one that is initially of greater value than the other as
illustrated in Figure 11.6, Fretwell predicts that:

1. A stable distribution will be attained when the value experienced

in each habitat is equal.
2. If habitat 1 had higher initial value (B1) than habitat 2 (B2) the
density of animals at stability will be greater in habitat 1 than in
habitat 2 (d1 > d2, d1 + d2 = k).
 3. At stability, the ratio of the populations in each habitat will be
the same as the ratio of resources available in the habitats, that
is:

Figure 11.6 The value of two habitats (B1 and B2) declines as the
density of competitors in them increases. Habitat B1 has a higher
initial value than habitat B2. The first ideal free individuals should go
to habitat B1. As the density of individuals in habitat B1 increases, its
value declines until the value in B1 becomes equal to the value in B2.
When this happens, all new arrivals should distribute equally to
habitats B1 and B2 until the population has been completed
distributed. The stable distribution will have individuals that
experience equal values in both habitats but there will be a higher
density of competitors in habitat B1 than in B2 (d1 > d2).
This form of the IFD has been called habitat matching (e.g.,
Pulliam & Caraco 1984). Another way to express this prediction is to
say that the proportion of individuals in a habitat should match the
proportion of resources provided by that habitat, or

The general empirical success of the model is impressive, but a

number of tests of the predicted distribution indicate consistent
deviations from proportional habitat matching (Kennedy & Grey
1993; Tregenza 1995).

Testing the Ideal Free Distribution

Testing the IFD requires that we apply it to the specific decision of

interest. For instance, in a foraging problem we must replace the
“value” of a habitat by a more specific statement or currency of
fitness. We can then use the IFD to generate a prediction based on
the hypothesis that animals distribute over concurrently available
foraging patches in a way that is consistent with maximizing their
rate of energy intake. This is exactly what Harper (1982) did when he
and his undergraduate students measured how a population of 33
adult, free-living mallard ducks (Anas platyrhynchos) residents of a
pond in the University Botanical Gardens of Cambridge University
(England) distributed themselves at two feeding stations located 20
m apart at two points on the edge of the pond. The resources
available in the habitats consisted of pre-cut, pre-weighed pieces of
white bread that were thrown at constant rates in the water by two
student observers. The initial value of the two habitats could be
changed in two ways: (i) by changing the rate at which the bread was
thrown, and (ii) by changing the size of the pieces of bread thrown in
the pond. During the experiments, there was a student observer at
each habitat who counted the number of mallards in each habitat at
regular intervals. Harper tested two predictions based on
maximization of food intake: that the proportion of birds at a site
would match the proportion of resources there and that at stability
all individuals in the habitats would obtain similar feeding rates
(Figure 11.7).

Figure 11.7 The university ducks experiment where students

threw bread at two stations on a pond in University Parks Cambridge
(Harper 1982). In the left panel the students in one station threw
bread at twice the rate as the other. The points show the number of
ducks that were present at the rich patch from the time the
experiment began. The horizontal line gives the predicted number
using the ideal free distribution game assuming the ducks are
designed to maximize their food intake. In the right panel the
students threw bread at the same rate, but in one station the size of
the bread was double that in the other. The ducks initially used the
rate of throwing to distribute but then slowly approached the
number predicted by the ideal free distribution based on the quantity
of food offered. Modified from Shettleworth (1998).
He found that when the two sites provided equal food input rates,
the birds divided themselves equally between the two places. When
he doubled the food input rate of one of the two sites, the birds re-
assorted themselves 2:1, as predicted by an IFD based on food intake
maximization (Figure 10.7A). He then kept the throwing rates
constant at both sites but doubled the size of the pieces of bread at
one site making it double the suitability of the other. The birds were
tricked at first by the throwing rates and distributed equally between
both sites. However, after a few seconds the birds re-distributed
according to the amount of bread being thrown (Figure 10.7B). Even
though the individuals assorted themselves in an ideal free way,
however, a more detailed analysis revealed that not all ducks were
deriving similar feeding rates from their foraging, a clear violation of
the second prediction of the game. His results would turn out to be
typical of most other studies of ideal free distribution (Kennedy &
Gray 1993). Harper’s experiment is one you could easily replicate
yourselves using a flock of city pigeons.
Another test of the model looked at the spatial distribution of the
foraging red knots discussed above. The main prey of the knots,
cockles, occur in patches that are spread out over a large area of
more than 1000 km2 in the mudflats of the Wadden Sea. The birds
feed in aggregations and Van Gils et al. (2006) found that they
distributed themselves in patches that conformed to an ideal, but not
free, model. The “not-free” aspect was due to the fact that the birds
can only feed during low tides, twice a day, which means they
necessarily must return to their roosts during high tides. Thus, the
second assumption of the IDF, that they are free to go to the best
habitats, was not met. But their spatial distribution during low tides
implies that the first assumption of the IFD, that they know the value
of all alternative habitats, was met. This raised the causal question:
how do they know? A likely hypothesis was that the birds use “public
information,” that is, observing the behavior of other birds. While
foraging, knots could observe the behavior of other knots and
perhaps thereby discover the location and quality of the best
available foraging grounds.
Bijleveld et al. (2015) devised two laboratory experiments to test this
idea. In the first, a focal bird was allowed to observe “demonstrator”
birds foraging on two separated patches, two demonstrators on each
patch. One patch had blue mussels hidden in the sand, the other
patch had none. The demonstrators were all hungry and did not
know which patch they occupied, so they all foraged. The only
difference between the demonstrators was that one set found and
could eat mussels, while the other could only forage. Focal birds
joined the patch with food more than 76% of the time. In the second
experiment, knots were allowed to forage in groups of one, two,
three, or four in an arena with 48 small patches, only one of which
was baited with mussels. As the number of foragers increased, the
time to find the baited patch by all birds decreased proportionally.
Neither dominance status nor sex appeared to be important for rate
of energy intake in either experiment. In this case, a test of the IFD
led to testable hypotheses about the model’s assumptions.
The IFD has been documented in an extraordinary diverse set of taxa
and circumstances (Tregenza 1995; see also Giraldeau & Caraco
2000). In most cases, animals adopt near ideal free distributions
despite important deviations from the game’s assumptions. The most
important and consistent violation of the ideal-free assumptions has
been the existence of unequal competitive abilities among animals.
Some individuals are simply capable of out-exploiting others or
keeping them away from good resource patches (Milinski & Parker
1991). It turns out that even when dealing with many of these
changes in assumptions, whether unequal competitors, perceptual
limitations, costs of moving among habitats etc., the outcome has
always been similar, a predicted deficit of individuals in the best
habitats and slight overcrowding in the poorer ones. Plots of the log
of the ratio of animals in each habitat against the log ratio of
resources in those habitats should produce a slope of 1.0 when
animals follow the IFD. In their reanalysis of 24 experimental studies
of IFD in a range of organisms, Kennedy and Gray (1993) report that
in most cases the density of animals in good habitats increases more
slowly than the quantity of resources in that habitat. Nonetheless,
given the general robustness of this theory to violations of its
assumptions, it is no surprise that it has become one of the first
important contributions of behavioral ecology to applied issues such
as conservation ecology (Sutherland 1996).
 SUMMARY AND CONCLUSIONS
The existence of natural selection makes it possible to study the
adaptive function of an organism’s traits. Adaptations are
characters that exist as a result of the action of natural selection
because they accomplish a function that contributes to an
individual’s fitness. To study function it is important that the
adaptive hypothesis be submitted to scientific testing. This
requires generating predictions by using backward engineering to
build an optimality model. Simple optimality and backward
engineering have been used with some success to predict prey
selection behavior as well as optimal patch exploitation decisions.
When the payoff associated with a behavioral option cannot be
defined without reference to the behavior adopted by other
members of the population, simple optimality must be replaced
by game theory and the evolutionarily stable strategy. We used a
game theoretical analysis of the Hawk–Dove game to show that
selection operating at the level of individuals could lead to
evolutionarily stable noninjurious fighting strategies. However,
we pointed out using the producer–scrounger game that selection
does not always lead to maximization of individual fitness.
Scroungers occur despite the costs they impose because the only
probable evolutionary outcome is a mixture of producers and
scroungers. Finally, we considered choice of habitats and the
ideal free distribution, which predicts that animals should settle
in the most profitable habitats. The outcome is that at
equilibrium individuals in all occupied habitats should
experience the same fitness.

FURTHER READING
Davies et al. (2012) provides an excellent and easy-to-read
introduction to behavioral ecology. More about foraging theory can
be obtained by reading Dave Stephens and John Krebs’ (1986) now
classic book and its more recent multi-authored follow-up by
Stephens et al. (2007). Those who wish to know more about games of
foraging should read Giraldeau and Dubois’ (2008) more recent
chapter on social foraging. For those interested in more general
game theory, Geof Parker’s (1984) excellent chapter as well as John
Maynard Smith’s (1982) book are two important sources while Lee
Dugatkin and Kern Reeve (1998) provide an up-to-date view of the
many applications of game theory in animal behavior. To explore the
relevance of producer–scrounger games to natural resource
havesting and public health see Giraldeau et al. (2017). For exploring
the interaction between functional and mechanistic approaches
students should consider Sara Shettleworth’s (2010) excellent book
and Reuven Dukas’ (1998) multi-authored book on cognitive ecology.

REFERENCES
Arak, A. 1988. Callers and satellites in the natterjack toad:
evolutionarily stable decision rules. Animal Behaviour, 36, 416–
32.
Barnard, C.J. 1984. Producers and Scroungers: Strategies of
Exploitation and Parasitism. London: Croom Helm.
Barnard, C.J. & Sibly, R.M. 1981. Producers and scroungers: a
general model and its application to captive flocks of house
sparrows. Animal Behaviour, 29, 543–50.
Beauchamp, G. 2000. Learning rules for social foragers: Implications
for the producer-scrounger game and ideal free distribution
theory. Journal of Theoretical Biology, 207, 21–35.
Bijleveld, A.I., Van Gils, J.A., Jouta, J.A. & Piersma, T. 2015. Benefits
of foraging in small groups: An experimental study on public
information use in red knots Calidris canutus. Behavioural
Processes, 117, 74–81.
Brockmann, H.J., Grafen, A. & Dawkins, R. 1979. Evolutionarily
stable nesting strategy in a digger wasp. Journal of Theoretical
Biology, 77, 473–96.
Cade, W. 1978. Of cricket song and sex. Natural History, 87, 64–72.
Davies, N.B., Krebs, J.R. & West, S.A. 2012. An Introduction to
Behavioural Ecology. Chichester: Wiley-Blackwell.
Dugatkin, L.A. & Reeve, H.K. 1998. Game Theory and Animal
Behavior. New York: Oxford University Press.
Dukas, R. 1998. Cognitive Ecology: The Evolutionary Ecology of
Information Processing and Decision Making. Chicago:
University of Chicago Press.
Fretwell, S.D. 1972. Populations in a Seasonal Environment.
Princeton, NJ: Princeton University Press.
Getty, T. 2002. The discriminating babbler meets the optimal diet
hawk. Animal Behaviour, 63, 397–402.
Getty, T. & Krebs, J.R. 1985. Lagging partial preferences for cryptic
prey: a signal detection analysis of great tit foraging. American
Naturalist, 125, 39–60.
Giraldeau, L.-A. & Beauchamp, G. 1999. Food exploitation: searching
for the optimal joining policy. Trends in Ecolology and Evolution,
14, 102–6.
Giraldeau, L.-A. & Caraco, T. 2000. Social Foraging Theory.
Princeton, NJ: Princeton University Press.
Giraldeau, L.-A. & Dubois, F. 2008. Social foraging and the study of
exploitive behaviour. Advances in the Study of Behavior, 38, 59–
104.
Giraldeau, L.-A., Heeb, P. & Kosfeld, M. (eds). 2017. Investors and
Exploiters in Ecology and Economics: Principles and
Applications. Strüngmann Forum Reports Cambridge, MA: MIT
Press.
Giraldeau, L.-A. & Kramer, D.L. 1982. The marginal value theorem: a
quantitative test using load size variation in a central place forager
the eastern chipmunk, Tamias striatus. Animal Behaviour, 30,
1036–42.
Giraldeau, L.-A., Kramer, D.L., Deslandes, I. & Lair, H. 1994. The
effect of competitors and distance on central place foraging in
eastern chipmunks, Tamias striatus. Animal Behaviour, 47, 621–
32.
Giraldeau, L.-A. & Livoreil, B. 1998. Game theory and social foraging.
In: L.A. Dugatkin & H.K. Reeve. (eds.), Game Theory and Animal
Behavior, pp. 16–37. New York: Oxford University Press.
Gould, S.J.&. & Lewontin, R.C. 1979. The spandrels of San Marco
and the Panglossian paradigm: a critique of the adaptationist
programme. Proceeding of the Royal Society London B, 205,
581–98.
Gross, M.R. 1985. Disruptive selection for alternative life histories in
salmon. Nature, 313, 47–48.
Harper, D.G.C. 1982. Competitive foraging in mallards: “ideal free”
ducks. Animal Behaviour, 30, 575–84.
Kennedy, M. & Gray, R.D. 1993. Can ecological theory predict the
distribution of foraging animals? A critical evaluation of
experiments on the ideal free distribution. Oikos, 68, 158–66.
Krebs, J.R. 1978. Optimal foraging: Decision rules for predators. In:
J.R. Krebs & N.B. Davies. (eds.), Behavioural Ecology: An
Evolutionary Approach, pp. 23–63. Sunderland, MA: Sinauer.
Krebs, J.R., Erichsen, J.T., Webber, M.I. & Charnov, E.L. 1977.
Optimal prey-selection by the great tit (Parus major). Animal
Behaviour, 25, 30–8.
MacArthur, R.H. & Pianka, E.R. 1966. On optimal use of a patchy
environment. American Naturalist, 100, 603–9.
Maynard Smith, J. 1982. Evolution and the Theory of Games.
Cambridge, UK: Cambridge University Press.
Maynard Smith, J. & Price, G.R. 1973. The logic of animal conflict.
Nature, 246, 15–18.
McAleer, K. & Giraldeau, L.-A. 2006. Testing central place foraging
in eastern chimunks, Tamias striatus, by altering loading
functions. Animal Behaviour, 71, 1447–53.
Milinski, M. & Parker, G.A. 1991. Competition for resources. In: J.R.
Krebs & N.B. Davies. (eds.), Behavioural Ecology: An
Evolutionary Approach, pp. 137–68. Sunderland, MA: Sinauer.
Morand-Ferron, J. & Giraldeau, L.-A. 2010. Learning behaviorally
stable solutions to producer-scrounger games. Behavioral
Ecology, 21, 343–8.
Nonacs, P. 2001. State dependent behavior and the Marginal Value
Theorem. Behavioral Ecology, 12, 71–83.
Oudman, T., Onrust, J., De Fouw, J., Spaans, B., Piersma, T. & Van
Gils, J.A. 2014. Digestive capacity and toxicity cause mixed diets
in red knots that maximize intake rate. American Naturalist, 183,
650–9.
Parker, G.A. 1984. Evolutionary stable strategies. In: J.R. Krebs &
N.B. Davies. (eds.), Behavioural Ecology: An Evolutionary
Approach, pp. 30–61. Sunderland, MA: Sinauer.
Piersma, T. & Van Gils, J.A. 2011. The Flexible Phenotype. New York:
Oxford University Press.
Pulliam, R.H. & Caraco, T. 1984. Living in groups: Is there an
optimal group size? In: J.R. Krebs & N.B. Davies. (eds.),
Behavioural Ecology: An Evolutionary Approach, pp. 122–147.
Sunderland, MA: Sinauer.
Schmid-Hampel, P., Kacelnik, A. & Houston, A.I. 1985. Honeybees
maximise efficiency by not filling their crop. Behavioral Ecology
and Sociobiology, 17, 61–6.
Shettleworth, S.J. 1998. Cognition, Evolution, and Behavior. New
York: Oxford University Press.
Shettleworth, S.J. 2010. Cognition, Evolution, and Behavior, 2nd ed.
New York: Oxford University Press.
Sih, A. & Christensen, B. 2001. Optimal diet theory: when does it
work, and when and why does it fail? Animal Behaviour, 61, 379–
90.
Stephens, D.W., Brown, J.S. & Ydenberg, R.C. 2007. Foraging:
Behavior and Ecology. Chicago: University of Chicago Press.
Stephens, D.W. & Krebs, J.R. 1986. Foraging Theory. Princeton, NJ:
Princeton University Press.
Sutherland, W.J. 1996. Predicting the consequences of habitat loss
for migratory populations. Preceedings of the Royal Society of
London, Series B, 263, 1325–7.
Tregenza, T. 1995. Building on the ideal free distribution. Advances
in Ecological Research, 26, 253–307.
Van Gils, J.A., Spaans, B., Dekinga, A. & Piersma, T. 2006. Foraging
in a tidally structured environment by red knots (Calidris
canutus): ideal, but not free. Ecology, 87, 1189–1202.
Williams, G.C. 1966. Adaptation and Natural Selection. Princeton,
NJ: Princeton University Press.
12
mate choice, mating systems, and sexual
selection
ANDERS PAPE MØLLER
INTRODUCTION
Male crickets Gryllus integer sing to attract females but thereby
run the risk of being parasitized by tachinid flies that are also
attracted to the calls (Cade 1975). Similarly, foxes Vulpes vulpes
prey on peacocks Pavo cristatus that are likely to be easier to
catch because of their cumbersome ornamentation, with
exaggerated feathers that weigh more than 1 kg and, when wet
after a shower, more than 3 kg (Petrie 1992). Why do females not
pay a similar cost? How can such apparently maladaptive traits
that reduce the probability of survival evolve and be maintained?
Charles Darwin developed the theory of sexual selection to
account specifically for the evolution and the maintenance of
exaggerated traits that do not benefit the individual in terms of
survival prospects. Many characters, such as the antlers of deer,
the train of the peacock, the calls of frogs and crickets, and the
exaggerated body size of one sex but not the other in the gorilla
Gorilla gorilla, can be attributed to the effects of sexual selection.
This chapter starts with a discussion of sexual selection followed
by a definition of mating systems and their determinants. Next, I
describe male–male competition and its consequences, analyze
female choice and the adaptive and nonadaptive bases for mate
choice, and briefly discuss evolutionary conflicts between the
sexes and their consequences. I then investigate the cost of
secondary sexual characters and its ecological and evolutionary
consequences and discuss the reasons for the presence of
multiple secondary sexual characters and their significance. In
the final part, I analyze sex ratio theory and how it relates to
sexual selection.

Sexual Selection
Sexual selection is a result of nonrandom variance in mating success
in one or both sexes. It is important to note that males (or females)
may vary in mating success for simple stochastic reasons: for
example, because of random rates of encounter with individuals of
the other sex, and because of a finite number of individuals in the
local neighborhood. Nonrandom variation in mating success
(hereafter called variation in mating success) may result in the
evolution of exaggerated traits when there is a quantitative genetic
basis (several genes with additive effects on the expression of a trait)
for such traits and when individuals with more extreme phenotypes
enjoy increased mating success.
Sexual selection differs from natural selection in terms of the kinds
of characters that can be attributed to the two processes of sexual
selection (intrasexual selection, usually male–male competition,
and intersexual selection, usually female mate choice) and,
therefore, in terms of the effects of the two kinds of processes on the
relative importance of survival and mate selection. While natural
selection results in the evolution of traits that are beneficial in terms
of survival, sexual selection is generally detrimental to the survival of
individuals. The reason for this effect is that traits that become
exaggerated by sexual selection may progress beyond the optimum
under natural selection, because of the mating advantages that they
confer. However, phenotypic traits can become immensely
exaggerated and cumbersome to the extent that males suffer
dramatic reductions in survival prospects; this process of
exaggeration will continue as long as some males are more than
compensated for the costs by advantages in terms of mating success.
Why is sexual selection important? Sexual selection accounts for
most differences in phenotype between males and females, and an
understanding of any phenomenon that can be attributed to
differences between the sexes among reproducing adults can only be
achieved by considering sexual selection. Since sexual selection
usually affects males and females differentially, ecological questions
such as the distribution of males and females can also only be
understood by considering sexual selection. Sexual selection also has
profound effects on the mortality rates of the two sexes, the
stochastic component of population fluctuations in the numbers of
individuals of the two sexes, the probability of and variance in
successful reproduction, and hence the likelihood of survival of a
population. Finally, most aspects of life among adult humans differ
between men and women, and we can only understand such
differences and the social, psychological, and medical problems that
they cause by considering sexual selection. Thus, we cannot
understand ecology, life-history theory, or human existence without
considering sex and sexual selection.

Mating Systems

Mating system is the label used for describing the way in which
males and females are distributed in reproductive units, and the
consequences of this distribution for reproductive behavior and
parental care. Why are some species polygynous (one male, several
females), whereas others are polyandrous (one female, several
males)? Why do females in most mating systems provide parental
care whereas males less frequently do so? Before answering these
questions we have to remember that a mating system is more than
meets the eye. Although a monogamous mating system is
described as a reproductive relationship between a single male and a
single female, sexual relationships between a “monogamous” female
and other males may render the term “monogamy” anomalous. Great
advances during the last 15 years in the study of parentage as a result
of the introduction of molecular genetic analyses have clearly
demonstrated that social and genetic mating systems are sometimes
not at all congruent. This superimposition of social and genetic
mating systems can most readily be resolved by considering that
sexual selection is a sequence of selection events starting from the
acquisition of a social mate, time required for acquiring and
processing these mates, followed by copulation, fertilization,
abortion, infanticide, parental care, and differential parental
investment (Figure 12.1). Intraspecific variation in each of the
variables in Figure 12.1 may be associated with the expression of
male secondary sexual characters. For example, attractive peacocks
with extravagant trains have shorter search time, take shorter time to
handle a female, and thus have a shorter mate processing time. This
results in such males having more copulation partners and hence
higher fecundity. Such attractive males cause females to invest
differentially in offspring, which increases fecundity per mate, again
increasing total fecundity. Since attractive peacocks also have lower
mortality, total fitness is presumably greater for the most attractive
males. Thus a thorough description of the mating system requires
quantification of components of sexual selection at these different
sequential stages and their interactions.

Figure 12.1 Mating system described as a sequence of sexual

selection events. (Adapted from Møller 1994)
Behavioral biologists have described a wide variety of mating
systems in different taxa. How do mating systems arise? What
determines whether males, females, or both are the choosy sex?
Originally, Bateman (1948) showed that although female fecundity in
Drosophila fruit flies did not increase after mating with a couple of
males, male fecundity continued to increase with number of mates.
Thus, the two sexes should differ in their willingness to mate since
only males would benefit from continued mating with additional
females. Subsequently random distributions of individuals
demonstrated that the observed distribution of individuals of the two
sexes often did not differ from the null distribution of a random
distribution. Hence male fruit flies should seek additional females,
whereas females should be reluctant to mate and therefore become
choosy. Later, Emlen and Oring (1977) suggested that the
operational sex ratio, defined as the ratio of fertilizable females to
sexually active males at any given time, is the main determinant of
the opportunity for sexual selection. For instance, when many
females are available per male we expect that the mating system is
biased toward polyandry (many males, one female), whereas a male-
biased operational sex ratio should result in a polygynous or lekking
mating system in which several males aggregate on small display
territories.
Several studies, however, have indicated that there is no clear
relationship between operational sex ratio and stronger sexual
selection for male traits favored in competition over mates and
thereby mating systems. For example, female water striders Gerris
odtogaster are generally reluctant to mate, but this reluctance
decreases when the sex ratio becomes male biased (Arnqvist 1989,
1992). Males of this species use an abdominal grasping organ for
holding on to reluctant females, and the success of males with large
abdominal organs decreases with increasing male-biased sex ratios.
Hence, female behavior may clearly influence the strength of sexual
selection. Clutton-Brock and Vincent (1991) suggested that relative
reproductive rates, measured as the potential rate of offspring
production of the two sexes, should reflect their relative level of
sexual competition. In an extensive review of animal species with
exclusively male parental care, they revealed that almost all cases
could be predicted from the rate of offspring production of males
relative to females. For example, in some species of pipefishes in
which males carry embryos in a placenta-like structure, males take
longer to produce a brood than do females. Males therefore have
slower reproductive rates than females, and they are predicted to be
less intensely sexually selected than females. Indeed, females are
larger and more colorful than males in these species, in accordance
with the prediction (Berglund et al. 1989). Such male limitation of
female reproductive success is rare, and polyandry is therefore much
less prevalent in the animal kingdom than is polygyny.
Although mating systems may be explained by a number of
ecological or evolutionary factors, as described above, we should not
forget that particular mating associations may arise for entirely
random reasons. For example, Sutherland (1985) showed that
chance alone can generate a sex difference in variance in mating
success, and he was able to explain Bateman’s famous result in fruit
flies simply on the basis of randomness. Surprisingly, a simple null
model based on chance alone has only recently been developed to
predict variation in mating systems. Bessa-Gomes et al. (2003)
created artificial populations of males and females in a computer
program, and by using finite populations of males and females
randomly skewed sex ratios were demonstrated to cause clear
patterns in mating systems. This was the case even without any
variation in parental care or the strength of mate preferences by
individuals of the two sexes, implying that the operational sex ratio
was not the important factor resulting in a particular mating system.
For example, lek breeding systems may arise due to randomly
skewed sex ratios biased toward females, whereas polyandry may
arise when the reverse sex ratio is found. Thus, random factors in
small populations may cause a skew in sex ratio, giving rise to
particular mating systems, even without any effects of operational
sex ratio.
Modeling of mating systems may thus provide important
information concerning evolution. A good example of a heuristic
model is the polygyny threshold model, first proposed by Verner
(1964) and Orians (1969). If females and males are distributed
according to the distribution of resources, and if individuals of the
choosy sex behave as if individuals of the chosen sex are yet another
resource, we can predict the distribution of females and hence the
mating success of males. The polygyny threshold model suggests that
sometimes a female may be better off by becoming a secondary
female of a polygynously mated male rather than the mate of a
monogamous male on a poor quality territory (Figure 12.2). The
polygyny threshold is thus the difference in quality of the breeding
situation for a monogamous and a secondary female, whereas the
cost incurred by a female for settling with an already-mated male is
described by the difference in reproductive success between a
simultaneously mated monogamous and secondary female. Pribil
and Searcy (2001) have provided experimental evidence for red-
winged blackbirds Agelaius phoeniceus consistent with these
predictions.
Figure 12.2 Polygyny threshold model with curves showing the
fitness of females in terms of reproductive success of monogamously
mated and secondary polygynously mated females. C, cost in terms
of fitness for a female mated to an already-mated male; PT, polygyny
threshold, i.e., gain in breeding situation necessary for a female
settling on the territory of an already-mated male compared with an
unmated male. (Adapted from Orians 1969).
Waders (shorebirds) belonging to the order of wading birds
Charadrii vary in social mating systems from polyandry to
monogamy to polygyny, with different mating systems having
evolved multiple times. The components of their mating systems are
complex, and the role of genes and environment are difficult to
assess. One example of such a complex mating system is that of the
ruff (Philomachus pugnax), a medium-sized wader that breeds in
grassy marshes and moist meadows across northern Eurasia.
Courthip and mating in the ruff occur on a lek, which can range in
size from 3 to 30 or more residences. Each residence is about 30 cm
in diameter and about 1 to 1½ meters from other residences. Each is
“owned” by a single male but may be visited by one to four satellite
males. Females also visit residences, which is where copulation takes
place; both resident and satellite males copulate (Hogan-Warburg
1966; Van Rhijn 1991). There are striking differences between the
two types of males, in appearance and behavior: dark, aggressive
owners and light, submissive satellites. These differences are a
genetically based behavioral polymorphism (Lank et al. 1995). A
third morph, called a “faeder,” has also been discovered that is a
female mimic and may also copulate (Jukema & Piersma 2006).
Recent studies of the genomics of ruffs have revealed that structural
genomic changes underlie the alternative reproductive strategies
(Lamichhaney et al. 2015), and that all three male morphs are
controlled by a single “super-gene” (Küpper et al. 2016). These
recent studies may help to understand the organization and
evolution of these complex mating systems.

Intrasexual Selection

Intrasexual selection is a very important force resulting in the

evolution of armament and greater body size in one sex compared
with the other. This component of sexual selection is usually male–
male competition because males gain higher reproductive success by
acquisition of additional mates, although female–female competition
is intense in the relatively few cases when polyandry has evolved.
Therefore, males will typically compete with each other for access to
more or higher quality females, and the outcome of such competition
is decided by differences in brute force or armament between the
contestants. Males and females often differ in body size, and such
differences may be consequences of natural or sexual selection.
Either sex may enjoy an advantage of small or large body size, and
sex differences in such advantages may account for the evolution of
sexual size dimorphism. For example, in some species females with
larger body size may have higher fecundity and therefore enjoy an
advantage over smaller conspecifics of the same sex. In other species
small females may enjoy a selective advantage because they mature
earlier and have shorter generation times. Similarly, in males large
body size may be advantageous because larger size provides an
advantage in competition over females or resources. However, in
other species, such as some spiders, small body size may be
advantageous because small males are more maneuverable, mature
earlier, and have higher success in certain kinds of competition.
Interspecific patterns of sexual size dimorphism have received
considerable attention. For example, Arak (1988b) investigated
sexual size dimorphism in nine species of anurans in relation to the
difference in selection gradients between individuals of the two
sexes. The selection gradient is estimated as the partial regression
coefficient of relative fitness on the trait (in this case, body size). A
positive difference in selection gradient implies that males benefit
more from large size than do females, whereas a negative difference
implies the reverse situation. Indeed, species of anurans with
relatively large males had a greater selective advantage of large body
size for males than for females. Similarly, Alexander et al. (1979)
investigated sexual size dimorphism in pinnipeds, ungulates, and
primates including humans and found consistent evidence across
taxa of greater dimorphism in species in which reproducing males
were able to monopolize a larger number of females. Perhaps
surprisingly, Alexander et al. (1979) even found evidence of greater
sexual size dimorphism in human societies with a higher frequency
of polygyny, suggesting that sex differences in body size within
species evolve readily in response to differences in selection
pressures on the two sexes. Arak (1988a) found evidence of a similar
intraspecific phenomenon among populations of anurans.
Extravagant weaponry such as spurs, horns, antlers, and exaggerated
canines in males are commonly found across many taxa. The
evolutionarily stable level of weaponry is predicted to reflect the size
of contested groups of females. Males have been shown to use their
weapons in fights with other males, providing a direct link between
morphological structure and context of use. Although large
morphological characters may appear to have a function as weapons,
alternative explanations should also be considered. Male characters
may also play a role in defense against predators, although this
hypothesis has received little support. Packer (1983) in his analysis
of horns in African antelopes showed convincingly that males have
evolved large horns that are thicker at the base compared with horns
in females. Female horns are mainly present in antelopes with large
body size, and their straight and thin structure make them eminent
stabbing devices when encountering or defending offspring against a
predator. The greater thickness of male horns seems to have evolved
directly in response to selection for greater strength, since males
have a higher frequency of broken horns compared with females.
Male characters have also been suggested to function as signals of
strength to other males or as indicators of sexual vigor to choosy
females. There is very little evidence for these alternative
explanations, although the few and generally small experimental
tests may caution against any general conclusions. Bro-Jørgensen
(2007) showed for bovids a similar pattern in sexual size
dimorphism, providing a striking similarity in pattern of body size
and horn size.
Several studies have suggested that males with large degrees of
weaponry indeed enjoy a mating success through male–male
competition. The forked fungus beetle Bolitotherus cornutus is a
small beetle that lives on or near fungi growing on the bark of
deciduous trees in North America. Conner (1988) has extensively
analyzed the relationship between length of male horns and sexual
selection, after controlling statistically for potentially confounding
factors such as body size. Selection analyses revealed a positive
association between horn length and male success (Figure 12.3),
supporting the prediction that males benefit in terms of sexual
selection by having long horns. In this species females do not appear
to choose mates based on horn length.
Figure 12.3 Male mating success in relation to horn length in the
forked fungus beetle Bolitotherus cornutus. (Adapted from Conner
1989).
The role of male–male competition in the evolution and
maintenance of weaponry is further supported by the evolution of
defensive properties in males of the same species. Any male able to
sustain a blow from the weapon of a competitor would potentially
enjoy an advantage. Hence, Jarman (1989) argued that there should
be selection for thick skin in males of species that have evolved
weaponry. Not only is the skin of males thicker than that of females
in large herbivores, it is also relatively thicker in the particular parts
of the body that commonly receive blows during fights. Kangaroos
use feet for fighting and therefore the skin has become thicker on
upper parts of the body that are subject to scratching. In antelopes
that use horns for fighting, the skin has become thicker on flanks and
other places susceptible to injury by horns. Such differences in skin
thickness between the sexes and among species would be difficult to
explain except when viewed in the light of sexual selection and the
evolutionary arms race of armament and counter-armament.
Males differ considerably in the architecture of their weaponry, from
pointed stabbing devices in some antelopes, through broad shovel-
like structures in some deer, to intricate branched structures in
beetles and deer and elaborate lifting or grasping devices in beetles.
The variety of use of weapons fully matches the variety of
morphological structures that have evolved. Hence, interpretations
of function of weapons will rely on observation of the actual behavior
during male-male interactions in fights.

Intersexual Selection: Female Preferences

Intersexual selection occurs because individuals of the two sexes

differ in their reproductive rates and in their willingness to mate.
Usually females do not benefit from mating with additional males
because it does not increase their fecundity, whereas males benefit
from mating with additional females. Females are therefore choosy
in most mating systems, whereas male choosiness occurs in
polyandrous and monogamous mating systems. Females often show
considerable concordance in mate preferences, with most females
agreeing upon which potential partner is most attractive. This is
epitomized in lekking species where one or a few males may account
for most matings, and the remaining males have few or no matings at
all. This is all the more surprising given that any genetic benefit
should long have been depleted due to persistent directional
selection. Why females show great consistency in mate preferences
despite the genetic benefits being small constitutes the so-called lek
paradox.
Quantitative genetic studies (which compare the phenotype of
offspring with that of their parents) or selection experiments (in
which individuals with extreme phenotypes are selected each
generation and used as breeders to produce the subsequent
generation) of more than 25 species have shown that mate
preferences have a significant heritability (Bakker & Pomiankowski
1995). This implies that mate preferences can readily evolve and that
any changes in the costs and benefits of such preferences may result
in microevolutionary change. The environmental determinants of
mate preferences are linked to previous or current experience, own
phenotype, and own attractiveness. For example, studies of the
stickleback Gasterosteus aculeatus have shown that females in
prime condition show stronger preferences for males with sexual
coloration than do females in poor condition (Bakker et al. 1999).
Similarly, studies of preferences for facial characteristics in terms of
asymmetry in humans show that women who themselves are
attractive also have a strong preference for more attractive men
(Little et al. 2001). Such a preference can be beneficial if women
mated to a relatively attractive partner invest relatively more in
reproduction. Differential investment in reproduction will be costly,
and the ability of women to make such an investment may depend on
their own condition.
Mate preferences can be considered to have two components: the
preference curve and the behavioral decision rule that a female uses
to choose a mate. Female mate preferences can be described by the
mathematical function that accounts for the relationship between
intensity of preference and expression of the male trait. Several
theoretical models have assumed that preference functions have an
intermediate maximum, reach an asymptote, or are “open-ended,”
i.e., “more is better.” Empirical studies of a few invertebrates have
suggested that the shape of preference curves may often be open-
ended. Decision rules have been hypothesized to be based on random
choice, sequential assessment, or a best-of-n-males rule (Janetos
1980). Jennions and Petrie (1997) reviewed the literature and found
considerable differences in how females sample males and discuss
the number of cues females use. They showed that sexual‐selection
studies have paid much less attention to variation among females
than to variation among males.
Empirical studies of individual females and their mate-searching
behavior under field conditions suggest that females often visit a
number of males, albeit a relatively small number (usually less than
10), before making a mate choice. Females will often return to a
previously visited male and mate with that particular individual. For
example, a study of great reed warblers Acrocephalus arundinaceus
using females provided with transmitters that allowed radio-tracking
showed that each female on average only visited six males (range 3–
11 males) before making a choice, and that females followed a best-
of-n-males rule of mate choice (Bensch & Hasselquist 1992). Thus,
there is much evidence that females do not choose the first best male
they encounter but spend considerable amounts of time and energy
searching for mates, thereby running risks of being eaten by a
predator, infected by parasites, or being too late to find an
appropriate mate and hence losing the possibility of any
reproduction at all.
The evolutionary origin of female mate preferences has received
considerable attention. Some theories suggest that preferences
evolved in conjunction with male traits, whereas other theories
propose that the female preference in fact predates the evolution of
the mate preferences. For example, Ryan et al. (1990) suggested that
sensory exploitation is the basis for evolution of male traits. Female
mate preferences for particular traits in frogs and fish appear to be
present in a group of species that evolved earlier than the group with
the male trait. Such analyses are inherently only as reliable as the
phylogenetic hypothesis on which they are based (see Chapter 16).
They also make assumptions about the origin of male traits, and the
absence of as yet undescribed or extinct species renders this
hypothesis difficult to test.

Benefits of Mate Choice

Why should females go to great lengths in finding a particular

partner rather than mating with any randomly chosen individual?
Given that females spend considerable lengths of time and amounts
of energy on mate-searching behavior, we can infer that there are
considerable benefits accruing to choosy females, otherwise natural
selection would have eliminated choosy females from the population.
Female mate preferences have long been suggested to arise and be
maintained in order to allow mating with individuals of the right
species. There is very little empirical evidence supporting this
hypothesis, and it is only presented here as a matter of completeness.
Fitness benefits accruing to choosy females are traditionally
categorized as direct or indirect benefits. Direct fitness benefits
are material benefits that females obtain in the current generation,
whereas indirect benefits are genetic benefits that females gain
through their offspring in the subsequent generation. Given this time
delay in acquisition of indirect benefits, we can predict that direct
benefits will be more important than indirect ones. Theoretical
models have generally not considered direct material benefits
because the mechanism and the advantages are so intuitively simple
that no underlying formal genetic model is needed and no problem
of maintenance of genetic variation arises. Direct fitness benefits
accrue to a choosy female if the expression of a secondary sexual
character directly reflects the quality of a territory, the quality or the
quantity of male parental care, the quality of sperm provided by a
male, or simply the absence of any directly transmitted disease from
a male to his partner (and then on to the offspring). An example of a
direct fitness benefit is the study of parental care in the bicolor
damselfish Stegastes partitus (Knapp & Kovach 1991). Males of this
fish court females vigorously, but males differ considerably in the
intensity of their courtship display. Females prefer males that court
more intensely. Males that courted females more intensely were also
more efficient at providing paternal care, measured as the survival
rate of eggs (Figure 12.4), and females therefore benefit directly from
their mate choice.
Figure 12.4 Egg survival (%) in relation to male courtship rate in
the bicolor damselfish Stegastes parrtitus. (Adapted from Knapp &
Kovach 1991).
Surprisingly, many species show exactly the opposite pattern, with
attractive males providing less care than unattractive males (see
Differential allocation below), and general patterns of sexual
selection for direct fitness benefits therefore need to be considered
carefully. Numerous studies have investigated how the expression of
male traits reflects the magnitude of direct fitness benefits. Møller
and Jennions (2001) quantitatively assessed the literature and found
that male traits or attractiveness on average accounted for 6.3% of
the variance in direct benefits in terms of fertility of eggs, for 2.3% of
the variance in fecundity in females through courtship, food, or
ejaculate components, for 1.3% of the variance in male parental care
in terms of feeding of offspring, and for 23.6% of male parental care
in terms of hatching rate in male-guarding ectotherms. Thus,
although some systems, such as egg-guarding fish, provide
considerable amounts of direct fitness benefits that far exceed the
genetic benefits in terms of good genes (see below), the direct
benefits for species of birds with male provisioning of offspring are
on average of a magnitude similar to the indirect benefits of “good
genes.” Why there are such large differences among ectotherms and
endotherms remains to be determined.
Indirect fitness benefits are genetic benefits that accrue to
choosy females through offspring, i.e., through delayed effects in the
subsequent generation. The initial hypothesis for such indirect
fitness benefits is due to Ronald Fisher (1930), who suggested that
mate preferences and male secondary sexual characters would co-
evolve to ever more exaggerated versions, provided that both were
partly genetically determined. Imagine that initially males differed
slightly in morphology and females differed slightly in their ability to
perceive such differences. Females that were better able to
discriminate among potential partners would tend to mate with
males that were first detected. Such pairs would produce sons that
had slightly enlarged characters that could more readily be detected
and daughters that also had a strong preference for such males.
Genes for male trait and female preference would tend to co-occur in
the same individuals and hence become linked. During subsequent
generations both male trait and female preference would co-evolve in
a runaway fashion to ever more extreme versions, until one or both
traits had gone to fixation (i.e., no more genetic variability present
because all individuals only had the dominant alleles, since these
alleles would confer the greater selective advantage) or until the
benefits for males in terms of mating success were balancing the
costs in terms of reduced viability due to the presence of the
exaggerated male trait. A pure Fisherian character will not show any
signs of condition dependence but entirely rely on attractiveness to
females for its maintenance. This is the reason why pure Fisherian
traits can be distinguished from condition-dependent traits
(e.g., those associated with the “good genes” hypothesis; see below)
by differences in the relationship between character size and
viability; males with the most exaggerated traits are likely to survive
better if the traits are condition-dependent (Jennions et al. 2001). In
practice, it is difficult to distinguish between this hypothesis and the
hypothesis of “good genes” (see below) because sample sizes need to
be sufficiently large to conclude that the generally small effects of
“good genes” are absent in purely Fisherian traits. However, the
Fisher hypothesis will always be at work if the assumptions of the
model are met. Many traits in species with little or no direct fitness
benefits of sexual selection are therefore likely to show a component
of Fisherian attractiveness.
“Good genes” provide genetic benefits that accrue to choosy females
as a consequence of mating with an attractive partner. If the
secondary sexual character reflects the genetic constitution of a male,
a female mating with an attractive male may produce offspring that
inherit the genetic qualities of the father. For example, the
expression of a secondary sexual character may directly reflect the
ability of an individual to resist parasites (Hamilton & Zuk 1982),
and attractive males will thus tend to sire resistant offspring.
Alternatively, attractive males may sire offspring that have general
high viability independent of disease resistance. An example of the
effect of “good genes” is offspring viability in the gray tree frog Hyla
versicolor. “Good genes” effects are difficult to estimate because
breeding designs must take maternal and common environment
effects into account. Welch et al. (1998) used a clever breeding
design controlling for such effects by randomly allocating males to
groups of females. This allowed estimation of the offspring viability
effect to be 10.4% of the variance in viability among sires. Although
this breeding design forcefully controlled for any obvious effects of
maternal or common environment, it did not control for differential
investment effects by females due to differences in perceived quality
among sires.
The social and sexual environments can also have significant impacts
on sexual behavior. For example, females may reduce the number of
males sampled when they are in an urgent stage of mate sampling
(Bastien et al. 2019). Such condition-dependent mate sampling has
now been reported for several species. In males, Gil et al. (1999)
showed that manipulation of testosterone in male eggs had different
effects on their song features. Manipulation of male tail length had
similar effects on song features as compared to males that had the
length of their tails shortened. Analyses of testosterone contents in
eggs among species also suggested that there were effects of social
and sexual environments on sexual signals (Gil et al. 2007).
How large are the so-called “good genes” benefits in general? Møller
and Alatalo (1999) quantified “good genes” benefits in terms of
increased viability of offspring sired by attractive males and found
that viability increased by about 1.5% due to “good genes” effects
across 22 species. There were significant differences in the
magnitude of the viability effect, with studies of birds showing
stronger effects than studies of other kinds of organisms. In addition,
species with a stronger skew in mating success among males indeed
had greater “good genes” benefits, as predicted if the basis for the
generally strong consistency in mate preferences among females in
lekking species is offspring viability benefits. Although this effect
may seem extremely small, a viability difference of this magnitude is
very large on an evolutionary time scale, and costly female mate
preferences can readily be maintained by the presence of such a
“good genes” effect.
Hamilton and Zuk (1982) suggested that male sexual displays may
reliably reveal the health status of a signaler because only high-
quality individuals would be able to sustain attacks from parasites.
Choosy individuals could thus obtain reliable but indirect
information on the level of resistance of a male to parasites by simply
inspecting costly secondary sexual characters. Numerous studies
have investigated this hypothesis and found general evidence for
males with larger degrees of ornamentation indeed having fewer
parasites than the average male in the population (review in Møller
et al. 1999). The original hypothesis of Hamilton and Zuk was based
on an argument about resistance. Hence, this hypothesis makes
predictions about the level of resistance rather than the level of
parasitism (the latter might reflect exposure as well as resistance). A
quantitative analysis of the literature revealed that measures of
immune response were better predicted by the size of secondary
sexual characters than were parasite loads (Møller et al. 1999).
Hence, females seem to be better able to obtain reliable information
on immune responses than on parasite loads by assessing male
secondary sexual characters. Saino et al. (1997) showed that male
barn swallows Hirundo rustica with long outermost tail feathers,
which are the subject of a directional female mate preference, have
better immune responses than short-tailed males, and that an
experimental elongation of tail length causes a reduction in antibody
production to challenge with a novel antigen. Antibody titers of
males were a very good predictor of male survival during the annual
migration to sub-Saharan Africa and back, independent of tail
manipulation, with naturally long-tailed males surviving the best.
This suggests that the male secondary sexual character subject to a
female preference provides direct information about the ability of
males to produce an immune response, and that both male trait and
its information on immune status are reliable predictors of survival
prospects.
The immune system is a very costly physiological defense system that
is closely integrated with the endocrine system, sexual selection, and
life history. The link between sexual display, immune function, and
endocrinology was formalized by the immunocompetence handicap
hypothesis. Folstad and Karter (1992) suggested that the level of
reliability of sexual signals may come about because the ontogeny of
such displays is under the influence of the endocrine system, which
in turn may have detrimental effects on the immune system. Hence,
individual signalers might be forced to adjust their level of signaling
to their body condition because otherwise they would risk sacrificing
their own survival. The original hypothesis was based on the
enhancing effects of testosterone on the level of sexual display but
antagonistic effects on immunity; effects of other hormones or
indirect effects of hormones on sexual display are also possible. Yet
another twist to this story was raised by von Schantz et al. (1999),
who proposed that the currency of extravagant sexual display and
efficient immune defense is free radical scavengers such as
carotenoids and vitamins A and E. Since many sexual displays are
based on carotenoids and since carotenoids cannot be synthesized
but only obtained through ingestion, individual signalers would have
to allocate limiting amounts of carotenoids to sexual signals and
competing physiological functions such as immunity and free radical
scavenging (Møller et al. 2000). In this scenario, individuals may
have to balance the use of such antioxidants to sexual signals and to
physiological functions such as immunity and detoxification.
Recently, more than 100 studies have investigated the relationship
between bilateral asymmetry and sexual selection. The basis for
this interest is that small differences in morphology between the two
sides of the body may reveal developmental instability, which is
defined as the inability of individuals to develop a stable phenotype
under given environmental conditions. Small asymmetries are
generally randomly directed, with most individuals in the population
being symmetrical. Choosy individuals may benefit from having a
symmetric partner because such a partner is better at providing
parental care, since an asymmetric phenotype interferes with
efficient locomotion. However, since the same patterns are found
across taxa independent of parental care, it is possible that the
advantage accruing to choosy individuals is in terms of the ability to
produce offspring with a symmetric and regular phenotype. Many
experimental studies and observations have shown that the degree of
fluctuating asymmetry of an individual is negatively associated with
mating success or attractiveness (reviews in Møller & Thornhill 1998;
Møller & Cuervo 2003). This holds for visual, auditory, and olfactory
signals and for organisms as diverse as insects, fish, birds, and
mammals including humans. These patterns are independent of
publication bias and other sources of error (Møller & Cuervo 2003),
and the relationships are as strong as for character size in the species
where both relationships have been investigated (Thornhill & Møller
1998).
A third genetic hypothesis concerning mate choice is based on
genetic complementarity. Individuals that differ in alleles at
particular loci may enjoy an advantage in terms of fecundity. For
example, individuals with different alleles at loci of the major
histocompatibility complex (MHC, an assemblage of loci involved in
production of immune defenses) are able to produce more diverse
defenses against parasites than individuals with similar alleles.
Hence, a female able to recognize the genotype of a potential mate
through its phenotypic characters and preferentially mate
disassortatively with males with a complementary genotype would
produce more viable offspring. This mechanism has been suggested
to account for mate preferences for males that differ in body odor in
mice and humans, since MHC genotype can be directly assessed
from odor (Wedekind et al. 1995). The generality of preferences for
genetic complementarity in animals remains to be determined.
The distinction between genetic and nongenetic components of
sexual selection is far from trivial, and empirical tests addressing
these issues have been marred by the difficulties of partitioning the
variance into its various component parts without confounding
genetic effects with maternal effects or common environment
effects. A particularly interesting phenomenon that seems to bridge
the “good genes” effects and maternal effects is differential parental
investment. Females of several species have been shown to invest
differentially in reproduction when mated to an attractive partner.
For example, Nancy Burley (1986) showed for the zebra finch
Taeniopygia guttata that females mated to males with attractive red
leg rings worked harder, laid more eggs, and reproduced more
frequently than females mated to males with unattractive green
rings. This maternal effect of differential investment is costly to
females since the increased parental activity and the elevated rate of
reproduction are costly in terms of time and energy use and in other
respects. In a field experiment, de Lope and Møller (1993)
manipulated the length of the outermost tail feathers of male barn
swallows after males had acquired a mate. Subsequently, females
mated to males with elongated tails worked harder in terms of
feeding of offspring than did females mated to males with shortened
tails. This feat was repeated in the second brood, since females
mated to tail-elongated males more often laid a second brood and
also worked harder than their mates when feeding the offspring.
Females mated to presumably attractive, long-tailed males suffered a
reduction in survival of 20% as a consequence of these kinds of
maternal effects. Subsequent studies have shown similar effects of
differential parental investment in many other species, including
differential allocation of testosterone, antibodies, and antioxidants
by females to their eggs depending on the phenotype of the male
partner (Gil et al. 1999; Petrie et al. 2001). Costly maternal effects
can only be maintained in the population if females benefit in terms
of increased viability of offspring and/or increased mating success of
their sons. Thus, the mere presence of maternal effects underlying
differential parental investment suggests strong indirect fitness
benefits of mate choice.
 Sexual Conflict

Male and female evolutionary interests are rarely if ever congruent,

and sexual conflicts of interest are a direct consequence of sexual
reproduction and differences in selection pressures on individuals of
the two sexes. Such conflict can be a great force in evolutionary
divergence. For example, Eberhard (1996) suggested that female
control of fertilization has resulted in very strong selection pressures
on males, leading to divergent evolution in reproductively related
phenotypic traits. A case in point is the evolution of genitalia, which
generally show dramatic divergence compared with other
morphological traits, the one and only other exception being
secondary sexual characters. This divergence in genital morphology
is so dramatic that species identification in invertebrates is often
based primarily on such characters. Arnqvist (1998) demonstrated in
a comparative analysis of genital morphology in insects that
divergence among species is indeed much greater than for ordinary
morphological characters. This finding suggests that ongoing sexual
selection is likely to exert intense pressure on any male trait
associated with reproduction and traits associated with control of
reproduction in females.
Sexual conflict may in fact be running sexual selection. The chase-
away model of sexual selection suggests that females evolve
resistance to male traits that increase male fitness at the expense of
female fitness (Holland & Rice 1998). For example, seminal fluid
proteins of male Drosophila fruit flies that reduce the remating
propensity of females are deleterious to female fitness. In an elegant
experiment, female evolution was arrested to investigate the effects
of continued male evolution on female fitness. As predicted, there
was a significant reduction in female fitness caused by ongoing
male–male competition resulting in male evolution (Rice 1996).
Conversely, when male evolution was arrested, the reverse effect was
found (Holland & Rice 1999). Although these effects were
demonstrated for seminal fluid proteins, similar mechanisms based
on other phenotypic traits are certainly imaginable. Further
experiments revealed that male fitness increased when females were
eliminated from the gene pool with implications for the Y gene pool
(Rice 1998).

Costs of Sexual Ornamentation

Costs of exaggerated phenotypes are an inherent aspect of

evolutionary models of sexual selection. Such costs arise because
characters become exaggerated beyond the optimum under natural
selection. Even in the case of condition-dependent expression of
secondary sexual characters, a cost is an inevitable consequence of
sexual selection. Condition-dependent characters are more
exaggerated among individuals in prime condition, and a positive
phenotypic correlation between ornamentation and condition should
therefore arise. What this means is that those individuals with the
most exaggerated secondary sexual characters tend to be in better
condition both before and after development of the character.
Therefore, such individuals have higher survival prospects than the
average individual in the population. For example, a study of the wolf
spider Hygrolycosa rubrofasciata demonstrates this point. In early
spring (April–May) males of this spider use old leaves on the ground
to produce a drumming sound, and females are attracted to this
acoustic display. Males with higher drumming rates are more
successful in terms of mating success but are also of superior
viability under both laboratory and field conditions compared with
less vigorously drumming males (Kotiaho et al. 1999). Male
drumming explains 6.8% of the variance in survival rate among
males.
This relationship between male sexual display and viability is a
general one since males with the most exaggerated secondary sexual
characters also tend to be those that survive the best (Jennions et al.
2001). On average, male ornamentation or attractiveness accounted
for 1.4% of the variance in male survival rate across studies.
Surprisingly, the amount of variance explained was of very similar
magnitude to that in studies of viability effects in “good genes” sexual
selection (Møller & Alatalo 1999). The positive relationship between
ornamentation and survival prospects does not imply that secondary
sexual characters are cheap, only that the tradeoff between survival
and mating success as determined by sexual display is masked by
individual differences in condition. Genetic and environmental
determinants of male differences in viability are possible, and studies
of “good genes” effects suggest that part of this difference in viability
is caused by genetic effects. Only an experiment can reveal the
underlying costs of ornamentation by displacing individuals
randomly from their once chosen optimum. An experiment
manipulating the length of the outermost tail feathers of male barn
swallows reversed the positive relationship between survival and
ornamentation among unmanipulated males to the predicted
negative association after experimentally increasing or decreasing
the tail length of males (Figure 12.5). This study demonstrates that
the underlying costs of ornamentation may be revealed by
phenotypic manipulations.
Figure 12.5 Survival rate of male barn swallows Hirundo rustica
in relation to the length of their tails. Males had the length of their
tails experimentally reduced, increased, or kept as controls (I,
treated control group; II, untreated control group). Values are mean
(SE) deviations for each of four experiments from the mean survival
rate in the populations, which was set at zero. (Adapted from Møller
1994; Møller & de Lope 1994).
The two sexes often differ markedly in mortality rate, and these
differences appear to be related to sexual selection. Larger male
mortality rates are associated with more intense competition among
males for access to females. In comparative studies of sex-specific
patterns of mortality in Promislow et al. (1992, 1994) have shown
that males have greater mortality in species where males have a
greater degree of exaggeration of plumage color than females, and
that the difference increases proportionally with the increase in
sexual dichromatism. Likewise, sex differences in longevity are
common in all human societies, with men living 5–10 years less than
women in Western societies. This difference is caused by dramatic
differential mortality between ages 16 and 28 years, when men suffer
a nearly 200% higher mortality than women (reviews in Trivers
1985; Wilson & Daly 1985). Causes of death that partly account for
this difference include a threefold greater risk of being murdered,
and higher risks of mortality from accidents, including car accidents,
even when taking the number of kilometers driven into account. Men
generally take much greater risks than women and seek medical
treatment much more rarely than women, even for the same
ailments. Men also suffer from a number of different diseases to a
greater extent than women; some of these effects are caused by
recessive alleles on the male’s unguarded X chromosome being
expressed in men but not women (Trivers 1985). The overall sex
difference in mortality can best be understood as arising from
intense male–male competition and sexual differences in sexual
psychology of men and women.
The actual mechanisms giving rise to sex differences in mortality in
animals arises from predation and/or parasitism. Many studies have
investigated predation risk in relation to sexual display or the
expression of secondary sexual characters. There is ample evidence
that predators use sexual signals to target males, and that predation
can be considered a general cause of sex differences in mortality
(review in Zuk & Kolluru 1998). However, several studies have
shown that males with the most extravagant secondary sexual
characters in fact run lower risks of predation than the average male.
In such cases, male condition seems to be a confounding factor
because males with large secondary sex traits not only pay the costs
of sexual display but do so based on an initially superior condition.
Hence, we can only expect to reveal the tradeoff between display and
predation risk when males are displaced experimentally from the
chosen level of sexual signaling.
Males often show a greater prevalence of parasites than females, and
many diseases also predominate in males. Parasitism can be
considered the second major cause of sex differences in mortality
(Alexander & Stimson 1989; Zuk 1990). Such sex differences in level
of parasitism may be partly mediated by endocrine effects on
susceptibility and parasite resistance, as suggested by Alexander and
Stimson (1989) and Folstad and Karter (1992). A comparative study
of birds showed that males generally had smaller immune defense
organs (bursa of Fabricius and spleen, a primary and a secondary
lymphoid organ) than females (Møller et al. 1998). Furthermore, the
sex difference in the size of immune defense organs was negatively
related to the frequency of extrapair paternity, which is a measure of
the intensity of sexual selection. Thus, in species with frequent
extrapair paternity and hence intense competition among males for
access to fertile females, males had a much larger reduction in spleen
size compared with females than in species with little or no extrapair
paternity.
Reductions in the viability of individuals of an intensely sexually
selected population may result in a population decrease in viability,
particularly during periods of adverse environmental conditions.
Similarly, a high variance in reproductive success, the associated
strong mate preferences for individuals with particular phenotypes,
and differential investment by females into reproduction with
attractive males may all be detrimental to the survival prospects of
small populations of threatened species. There are numerous
examples of small populations of threatened species in nature
reserves or in captivity with little or no reproductive success, and
where numbers remain small or decrease because of little or no
reproduction. For example, newspapers and magazines often feature
stories of giant pandas Ailuropoda melanoleuca or other enigmatic
animals being transported from one zoo to another to make yet
another attempt at successful reproduction.
A hypothesis that accounts for such problems is that sexual selection
causes a reduction in population viability (Møller & Legendre 2001).
Costs of sexual selection may reduce viability of small populations
because there is a biased number of individuals of the two sexes for
completely stochastic reasons. This effect may become exaggerated
by strong female mate preferences for particular males, since all such
males may be absent for random reasons. Similarly, females may in
such situations reduce investment in reproduction or not reproduce
at all when mated to unattractive males. Thus, small populations or
populations in captivity, such as those of species threatened by
extinction, may run increased risks of extinction simply because they
have an evolutionary history of intense sexual selection.
A common objection to the idea of costs of sexual selection is that it
may be balanced or even exceeded by benefits in terms of the
positive effects of “good genes.” This is not necessarily so. Depending
on the nature of the “good genes,” there may be little or no positive
effect on individuals in the population. For example, differences in
genetic constitution among males in the population may arise as a
consequence of more individuals of poor genetic quality being
present in a highly sexually selected species.

Multiple Ornaments

The previous sections have generally dealt with sexual signaling for a
single phenotypic character. However, in real cases of signaling,
individuals often simultaneously exhibit many different kinds of
secondary sexual characters, such as exaggerated morphological
traits, colors, and vocalizations. The evolution and maintenance of
such multiple signals need to be explained. Candolin (2005)
reviewed the literature on multiple ornamentation focusing on the
benefits of multiple choice, but also on the information content of
the mate choice of the cues. The evolution of multiple characters can
derive from the following.

1. Each character provides information about different aspects of

the phenotype of a signaler.
2. Each secondary sexual character provides redundant
information about the overall condition of an individual. A
single signal may not reveal complete information about a
particular quality of an individual, and multiple signals may
provide independent information that in combination reflect its
overall quality.
3. Some characters do not currently reflect properties of the
signaler but did so in the past. If the traits are not particularly
 costly, they may be maintained for a long time after they cease
providing information on signalers (Møller & Pomiankowski
1993).

The junglefowl (Gallus gallus spadiceus) provides a well-known

example of multiple signals. Cockerels have elongated tail and hackle
feathers, bright coloration of different parts of the body, a red wattle
and comb, and a crowing call. Studies of sexual selection in this
species have revealed that only a single trait, the size of the comb,
provides reliable information about infection status with a nematode
parasite. During mate-choice sessions, females only paid attention to
this single trait revealing parasite status (Zuk et al. 1990), suggesting
that the other traits were not currently reflecting information about
the condition of males. Other studies of other organisms have
revealed evidence for the first and second hypotheses of multiple
ornaments. A general understanding of the function and importance
of multiple signals is still far from being achieved.

Sex Ratios

Why are equally many males and females born in most species?
Ronald Fisher (1930) suggested that an equal sex ratio would be
advantageous to an individual because any deviation from this ratio
on average would result in the production of fewer grandparental
offspring. This view of Fisher’s sex ratio theory is based on the
premise that sons and daughters cost the same, whereas the sex ratio
in fact should be based on the combined cost of sons versus
daughters. Deviations from these predicted patterns of equal sex
ratios occur when siblings mate with each other, in which case sex
ratios are generally biased strongly toward daughters.
Adaptive variation in sex ratio has attracted considerable attention
recently, particularly because of its relationship with sexual
selection. If males are especially sexually attractive for genetic
reasons, male offspring of such sires should enjoy a disproportionate
mating success compared with the average male in the population.
Females mated to such attractive males should therefore produce
relatively more sons than daughters. Indeed, attractive males of
several species seem to have mates that produce more sons than
daughters. Nancy Burley (1981), in a classic study of the zebra finch,
showed that when males became more attractive by simple
experimental manipulation of the color of their plastic leg rings, this
affected the sex ratio of their offspring. Female zebra finches
preferred males with red rings and avoided males with green rings.
Females mated to red-ringed males produced significantly more sons
than did females mated to green-ringed males, as predicted by
theory (Figure 12.6). Subsequent studies of several other species
have provided similar evidence for adaptive adjustment of sex ratio
in response to the phenotype of the sire, although experimental
studies testing this prediction are still the exception.
Figure 12.6 Sex ratio of zebra finches Taeniopygia guttata in
relation to the color of plastic leg rings of their mates, where red
rings are attractive and green rings are unattractive. (Adapted from
Burley 1981).
Numerous studies have demonstrated evidence of female condition
being related to sex ratio of their offspring. If sons are more costly to
produce than daughters and if female condition changes consistently
with age, we should expect sex ratios to change similarly. Many
studies have shown a preponderance of sons being produced by
dominant females or by females in prime condition. Likewise,
senescence seems to be associated with a sex ratio bias toward
daughters.
Studies of sex ratios in humans provide equally striking effects of
environmental conditions. For example, studies at the US West Point
military academy of the sex ratio of cadets in relation to military rank
show a preponderance of sons produced by women married to high-
ranking officers. A similar result has been reported for the sex ratio
of the offspring of US presidents, vice presidents, and cabinet
secretaries. Among the US executive branch, a staggering 70% of
sons have been recorded during the first 20 presidents, falling to
53% during the following 20 presidents, compared with the null
expectation of 50% recorded in human populations in general
(Betzig & Weber 1995). Similarly, many different studies have
recorded changes in sex ratios associated with resource abundance
and other indicators of environmental conditions. For example, in
rural Portugal during the period 1671–1720 the sex ratio was 112.1
sons per 100 daughters in good harvest years but only 90.7 sons per
100 daughters in poor harvest years (Trivers 1985). Twins are more
costly to produce than two singletons, and we should thus expect the
sex ratio of twins to be biased toward daughters. In fact, the sex ratio
of human twins is 3% lower than that of singletons (Trivers 1985). If
this deviation from an equal sex ratio is adaptive, we should expect
that male twins have disproportionately lower reproductive success
than female twins. Indeed, both male and female twins have lower
reproductive success than singletons, based on data from eighteenth-
and nineteenth-century Mormons in the USA. However, the
reduction in reproductive success among male twins is twice as large
as that for female twins, consistent with the prediction (Trivers
1985). More recently, sex ratio theory has been invoked to explain
variation in sex ratios in societies with enforced family planning such
as China. Families generally favor sons over daughters when a one-
child policy is enforced, and this happens as a consequence of
selective abortion, infanticide, and other mechanisms. Surprisingly,
in families with more than a single child there are dramatic changes
in sex ratios of the second and later children, depending on the sex
ratio of the first child (Low 2000). If the first child is a son, the sex
ratio of any subsequent children is biased toward daughters; the
reverse situation occurs when the first child is a daughter. These
studies, and many others reviewed by Trivers (1985), suggest that
human sex ratios are closely linked to sexual selection and the
advantages that accrue to women in terms of future reproductive
potential through the sex of their offspring.

SUMMARY AND CONCLUSIONS

The pervasive effects of reproduction on parents in sexually
reproducing species have caused sexual selection to affect most
aspects of the lives of individuals. Sexual selection and mate
choice does not only affect the appearance of individuals and how
they behave toward each other during the breeding season. The
effects of sexual selection appear before this, when parents make
decisions about the sex ratio of their offspring and how these
offspring are to be provisioned. This is followed by decisions
about developmental rates for secondary sexual characters, when
these first appear, and the degree to which these characters are
exaggerated. Individuals of the choosy sex are equally affected by
sexual selection, namely their sales resistance when approached
by potential mates, their choice of partner, and their subsequent
reproductive decisions. Sexual selection also affects the way in
which individuals are spatially distributed and therefore has
important consequences for social organization. The costs of
secondary sexual characters can be dramatic and may affect
populations to the extent that they may constitute an important
risk of extinction for some species. These effects of sexual
selection on most aspects of the life of different organisms are
equally prominent in humans, where everything from sex ratio,
sex differences in mortality, and sex differences in sexual
behavior and risk-taking can be considered to have arisen as a
consequence of this evolutionary force.

FURTHER READING
Andersson (1994) provides a general overview of sexual selection
theory and summarizes the empirical evidence. Monographs on
sexual selection in particular species include Houde (1997) on the
guppy, Møller (1994) on the barn swallow, and Ryan (1985) on the
túngara frog. Sex ratio theory is reviewed extensively by Trivers
(1985). Low (2000) provides an extensive treatise on the
consequences of sexual selection for humans. The consequences of
sexual selection for conservation are described by Møller (2000) and
Wilson et al. (1998).
Methodological addendum: Many model systems have been
used repeatedly in order to investigate different predictors or other
factors that may reveal novel mechanisms or functions of behavioral
traits. Such repeated investigations can be evaluated using meta-
analyses, which basically represents the quantitative assessment of
multiple studies and their synthesis. Parker (2013) investigated over
1200 relationships between the effects of coloration on signalling
and other phenotypic traits in the blue tit Cyanistes caeruleus with
only 52 studies being suited to be included in a meta-analysis.
Mainly small effects with weak relationships were the outcome of
this meta-analysis. Romano et al. (2017a) conducted a similar
analysis of barn swallow behavior that showed many more
intermediate to large effects, demonstrating differences in intensity
of sexual selection among populations, consistent with a role of
sexual selection in divergence and speciation. Romano et al. (2017b)
showed consistent and significant effects of the expression of
viability being linked to small to intermediate effects. These findings
open up novel ways to investigate behavior and to quantitatively
assess such patterns of behavior. The links between behaviour,
ecology, and evolution are all open for these novel ways of
quantitative analysis.

REFERENCES
Andersson, M. 1994. Sexual Selection. Princeton, NJ: Princeton
University Press.
Arak, A. 1988a. Callers and satellites in the natterjack toad:
evolutionarily stable decision rules. Animal Behaviour, 36, 416–
32.
Arak, A. 1988b. Sexual dimorphism in body size: & model and & test.
Evolution, 42, 820–5.
Arnqvist, G. 1989. Sexual selection in the water strider: the function,
mechanism of selection and heritability of & male grasping organ.
Oikos, 56, 344–50.
Arnqvist, G. 1992. Spatial variation in selective regimes: sexual
selection in the water strider. Gerris odontogaster. Evolution, 46,
914–29.
Arnqvist, G. 1998. Comparative evidence for the evolution of
genitalia by sexual selection. Nature, 393, 784–8.
Bakker, T.C.M., Kunzler, R. & Mazzi, K. 1999. Condition-related
mate choice in sticklebacks. Nature, 401, 234.
Bakker, T.C.M. & Pomiankowski, A. 1995. The genetic basis of female
mate preferences. Journal of Evolutionary Biology, 8, 129–81.
Bastien, B., Farley, G., Ga, F., Malin, J.S., et al. 2019. The waiting
and mating game. Condition-dependent mate sampling in female
gray tree frogs (Hyla versicolor). Frontiers in Ecology, 2018,
00140.
Bateman, A.J. 1948. Intra-sexual selection in Drosophila. Heredity,
2, 349–68.
Bensch, S. & Hasselquist, D. 1992. Evidence for active female choice
in & polygynous warbler. Animal Behaviour, 44, 301–12.
Berglund, A., Rosenqvist, G. & Svensson, I. 1989. Reproductive
success of females limited by males in two pipefish species.
American Naturalist, 133, 506–16.
Bessa-Gomes, C., Legendre, S., Clobert, J. & Møller, A.P. 2003.
Modeling mating patterns given mutual mate choice: the
importance of individual mating preferences and mating system.
Journal of Biological Systematics, 11, 205–19.
Betzig, L.L. & Weber, S. 1995. Presidents preferred sons. Politics and
the Life Sciences, 14, 61–4.
Bro-Jørgensen, J. 2007. The intensity of sexual selection predicts
weapon size in bovids. Evolution, 61, 1316–26.
Burley, N. 1981. Sex ratio manipulation and selection for
attractiveness. Science, 211, 721–2.
Burley, N. 1986. Sexual selection for aesthetic traits in species with
biparental care. American Naturalist, 127, 415–45.
Cade, W.H. 1975. Acoustically orienting parasitoids: fly phonotaxis to
cricket song. Science, 190, 1312–3.
Candolin, U. 2005. The use of multiple cues in mate choice.
Biological Reviews, 78, 575–95.
Clutton-Brock, T.H. & Vincent, A.C.J. 1991. Sexual selection and the
potential reproductive rates of males and females. Nature, 351,
58–60.
Conner, J. 1988. Field measurements of natural and sexual selection
in the fungus beetle. Bolitotherus cornutus. Evolution, 42, 735–
49.
Conner, J. 1989. Density-dependent sexual selection in the fungus
beetle. Bolitotherus cornutus. Evolution, 43, 1378–86.
De Lope, F. & Møller, A.P. 1993. Female reproductive effort depends
on the degree of ornamentation of their mates. Evolution, 47,
1152–60.
Eberhard, W.G. 1996. Female Control: Sexual Selection by Cryptic
Female Choice. Princeton, NJ: Princeton University Press.
Emlen, S.T. & Oring, L.W. 1977. Ecology, sexual selection, and the
evolution of mating systems. Science, 197, 215–23.
Fisher, R.A. 1930. The Genetical Theory of Natural Selection.
Oxford: Clarendon Press.
Folstad, I. & Karter, A.J. 1992. Parasites, bright males, and the
immunocompetence handicap. American Naturalist, 139, 603–
22.
Gil, D., Biard, C., Lacroix, A., Spottiswoode, C., et al. 2007. Yolk
androgens in birds: Development, coloniality and sexual
dichromatism. American Naturalist, 160, 802–19.
Gil, D., Graves, J.A., Hazon, N. & Wells, A. 1999. Male attractiveness
and differential testosterone investment in zebra finch eggs.
Science, 286, 126–8.
Hamilton, W.D. & Zuk, M. 1982. Heritable true fitness and bright
birds: a role for parasites? Science, 218, 384–87.
Hogan-Warburg, A.J. 1966. Social behavior of the ruff, Philomachus
pugnax (L.). Ardea, 54, 109–229.
Holland, B. & Rice, W.R. 1998. Chase-away sexual selection:
antagonistic seduction versus resistance. Evolution, 52, 1–7.
Holland, B. & Rice, W.R. 1999. Experimental removal of sexual
selection reverses intersexual antagonistic coevolution and
removes a reproductive load. Proceedings of the National
Academy of Sciences USA, 96, 5083–88.
Houde, A.E. 1997. Sex, Color and Mate Choice in Guppies.
Princeton, NJ: Princeton University Press.
Janetos, A.C. 1980. Strategies for female mate choice: a theoretical
analysis. Behavioral Ecology and Sociobiology, 1, 107–12.
Jarman, P.J. 1989. On being thick-skinned: dermal shields in large
mammalian herbivores. Biological Journal of the Linnean
Society, 36, 169–91.
Jennions, M.D., Møller, A.P. & Petrie, M. 2001. Sexually selected
traits and adult survival: a meta-analysis of the phenotypic
relationship. Quarterly Review of Biology, 76, 3–36.
Jennions, M.D. & Petrie, M. 1997. Variation in mate choice and
mating preferences: A review of causes and consequences.
Biological Reviews, 72, 283–327.
Jennions, M.D. & Petrie, M. 2000. Why do females mate multiply? A
review of the genetic benefits. Biological Reviews, 75, 21–64.
Jukema, J. & Piersma, T. 2006. Permanent female mimics in a
lekking shorebird. Biology Letters, 2, 161–4.
Knapp, R.A. & Kovach, J.T. 1991. Courtship as an honea indicator of
male parental quality in the bicolor damselfish. Stegastes
partitus. Behavioral Ecology, 2, 295–300.
Kotiaho, J.S., Alatalo, R.V., Mappes, J. & Parri, S. 1999. Sexual
signalling and viability in & wolf spider (Hygrolycosa
rubrofasciata): measurements under laboratory and field
conditions. Behavioral Ecology and Sociobiology, 46, 123–8.
Küpper, C., Stocks, M., Risse, J.E., et al. 2016. A supergene
determines highly divergent male reproductive morphs in the ruff.
Nature Genetics, 48, 79–83.
Lamichhaney, L., Fan, G., Widemo, F., et al. 2015. Structural
genomic changes underlie alternative reproductive strategies in
the ruff (Philomachus pugnax). Nature Genetics, 48, 84–8.
Lank, D.B., Smith, C.M., Hanotte, O., Burke, T. & Cooke, F. 1995.
Genetic polymorphism for alternative mating behaviour in lekking
male ruff. Philomachus pugnax. Nature, 378, 59–62.
Little, A.C., Burt, D.M., Penton-Voak, I.S. & Perrett, D.I. 2001. Self-
perceived attractiveness influences human female preferences for
sexual dimorphism and symmetry in male faces. Proceedings of
the Royal Society of London Series B, 268, 39–44.
Low, B.S. 2000. Why Sex Matters. Princeton, NJ: Princeton
University Press.
Møller, A.P. 1994. Sexual Selection and the Barn Swallow. Oxford:
Oxford University Press.
Møller, A.P. 2000. Sexual selection and conservation. In: M. Gosling
& W.J. Sutherland (eds.), Behaviour and Conservation, pp. 172–
197. Cambridge: Cambridge University Press.
Møller, A.P. & Alatalo, R.V. 1999. Good genes effects in sexual
selection. Proceedings of the Royal Society of London Series B,
266, 85–91.
Møller, A.P., Biard, C., Blount, J.D., et al. 2000. Carotenoid-
dependent signals: indicators of foraging efficiency,
 immunocompetence or detoxification ability? Poultry and Avian
Biology Reviews, 11, 137–59.
Møller, A.P., Christe, P. & Lux, E. 1999. Parasite-mediated sexual
selection: effects of parasites and hoa immune function. Quarterly
Review of Biology, 74, 3–20.
Møller, A.P. & Cuervo, J.J. 2003. Asymmetry, size and sexual
selection: factors affecting heterogeneity in relationships between
asymmetry and sexual selection. In: M. Polak (ed.),
Developmental Instability, pp. 262–275. New York: Oxford
University Press.
Møller, A.P. & De Lope, F. 1994. Differential costs of a secondary
sexual character: an experimental tea of the handicap principle.
Evolution, 48, 1676–83.
Møller, A.P. & Jennions, M.D. 2001. How important are direct
fitness benefits of sexual selection? Naturwissenschaften, 88,
401–15.
Møller, A.P. & Legendre, S. 2001. Allee effect, sexual selection and
demographic stochasticity. Oikos, 92, 27–34.
Møller, A.P. & Pomiankowski, A. 1993. Why have birds got multiple
sexual ornaments? Behavioral Ecology and Sociobiology, 32,
167–76.
Møller, A.P., Sorci, G. & Erritzoe, J. 1998. Sexual dimorphism in
immune defense. American Naturalist, 152, 605–19.
Møller, A.P. & Thornhill, R. 1998. Bilateral symmetry and sexual
selection: a meta-analysis. American Naturalist, 151, 174–92.
Orians, G.H. 1969. On the evolution of mating systems in birds and
mammals. American Naturalist, 103, 589–603.
Packer, C. 1983. Sexual dimorphism: the horns of African antelopes.
Science, 221, 1191–3.
Parker, T.H. 2013. What do we really know about the signalling role
of plumage colour in blue tits: A case study of impediments to
 progress in evolutionary biology. Biological Reviews, 88, 511–36.
Petrie, M. 1992. Peacocks with low mating success are more likely to
suffer predation. Animal Behaviour, 44, 485–6.
Petrie, M., Schwabl, H., Brande-Lavridsen, N. & Burke, T. 2001. Sex
differences in avian yolk hormone levels. Nature, 412, 498–99.
Pribil, S. & Searcy, W.A. 2001. Experimental confirmation of the
polygyny threshold model for red-winged blackbirds. Proceedings
of the Royal Society of London Series B, 268, 1643–6.
Promislow, D.E.L., Montgomerie, R. & Martin, T.E. 1992. Mortality
costs of sexual dimorphism in birds. Proceedings of the Royal
Society of London Series B, 250, 143–50.
Promislow, D.E.L., Montgomerie, R. & Martin, T.E. 1994. Sexual
selection and survival in North American waterfowl. Evolution,
48, 2045–50.
Rice, W.R. 1996. Sexually antagonistic male adaptation triggered by
experimental arrest of female evolution. Nature, 361, 232–4.
Rice, W.R. 1998. Male fitness increases when females are eliminated
from the gene pool: Implications for the Y chromosome.
Proceedings of the National Academy of Sciences USA, 95, 6217–
21.
Romano, A., Costanzo, A., Rubolini, D., Saino, N. & Møller, A.P.
2017a. Geographical and seasonal variation in the intensity of
sexual selection in the barn swallow Hirundo rustica: A meta-
analysis. Biological Reviews, 92, 1582–1600.
Romano, A., Saino, N. & Møller, A.P. 2017b. Viability and expression
of sexual ornaments in the barn swallow Hirundo rustica: A meta-
analysis. Journal of Evolutionary Biology, 30, 1929–35.
Ryan, M.J. 1985. The Túngara Frog. Chicago: University of Chicago
Press.
Ryan, M.J., Fox, J.H., Wilczynski, W. & Rand, A.S. 1990. Sexual
selection for sensory exploitation in the frog. Physalaemus
 pustulosus. Nature, 343, 66–7.
Ryan, M.J. & Rand, A.S. 1993. Sexual selection and signal evolution:
the ghost of biases past. Philosophical Transactions of the Royal
Society of London Series B, 340, 187–95.
Saino, N., Bolzern, A.M. & Møller, A.P. 1997. Immunocompetence,
ornamentation and viability of male barn swallows (Hirundo
rustica). Proceedings of the National Academy of Sciences USA,
94, 549–52.
Sutherland, W.J. 1985. Chance can produce a sex difference in
variance in mating success and explain Bateman’s data. Animal
Behaviour, 33, 1349–52.
Thornhill, R. & Møller, A.P. 1998. The relative importance of size and
asymmetry in sexual selection. Behavioral Ecology, 9, 546–51.
Trivers, R.L. 1985. Social Evolution. Menlo Park, CA:
Benjamin/Cummings.
Van Rhijn, J.G. 1991. The Ruff. London: Poyser.
Verner, J. 1964. The evolution of polygamy in the long-billed marsh
wren. Evolution, 18, 252–61.
von Schants, T., Bensch, H., Grahn, M. et al.1999. Good genes,
oxidative stress and condition-dependent sexual signals.
Proceedings of the Royal Society of London B, 266, 1–12.
Wedekind, C., Seebeck, T., Bettens, F. & Paepke, A.J. 1995. MHC-
dependent mate preferences in humans. Proceedings of the Royal
Society of London Series B, 260, 245–9.
Welch, A.M., Semlitsch, R.D. & Gerhardt, H.C. 1998. Call duration as
an indicator of genetic quality in male gray tree frogs. Science,
280, 1928–30.
Wilson, M. & Daly, M. 1985. Competitiveness, risk taking, and
violence: the young male syndrome. Ethology and Sociobiology,
6, 59–73.
Wilson, M., Daly, M. & Gordon, S. 1998. The evolved psychological
apparatus of human decision-making is one source of
environmental problems. In: T. Caro (ed.), Behavioral Ecology
and Conservation Biology, pp. 501–23. Oxford: Oxford University
Press.
Zuk, M. 1990. Reproductive strategies and sex differences in disease
susceptibility: an evolutionary viewpoint. Parasitology Today, 6,
231–3.
Zuk, M. & Kolluru, G.R. 1998. Exploitation of sexual signals by
predators and parasitoids. Quarterly Review of Biology, 73, 415–
438.
Zuk, M., Thornhill, R., Ligon, J.D., et al. 1990. The role of male
ornaments and courtship behavior in female mate choice of red
jungle fowl. American Naturalist, 136, 459–73.
13
Animal Personality, the Study of Individual
Behavioral Differences
DENIS RÉALE AND PIERRE-OLIVIER MONTIGLIO
INTRODUCTION
At the turn of the twenty-first century, the idea that social
networks comprise individuals has become one of the new pillars
of our contemporary societies. Every day, millions of people
commute singly listening to personalized radio programs or
engage in all-inclusive communication with their personal
electronic devices. Jogging has become one of the most popular
activities of urban populations. Some people spend hours
broadcasting their private lives on social networks, and several of
us are attracted by electronic devices whose names start with the
prefix “i.” The questions asked by scientists both reflect and
impact the trends of a society. It is thus not surprising to see the
interest that biological sciences have recently developed around
questions related to individuality. Indeed, the individual
phenomenon has started invading several fields of biology: we
speak of personalized medicine, we recognize the importance of
individuals for epidemiology, energetics, endocrinology, ecology,
and life history. Behavioral ecology has not been spared by the
trend: since the late 1990s studies on personality or temperament
have become a major topic in the field. It is now recognized that
animal personality research has built up bridges between
different fields of biology, ecology, and psychology, and is part of
the current trend in the development of integrative biology.
In this chapter we will review the main aspects related to the
recent trends in animal personality. Newcomers in animal
personality raise some questions about the field and its concepts.
Thus, we will first discuss terminological and other issues related
to personality. We then present a brief history, with an emphasis
on behavioral ecology. We next review the different sources of
consistent individual behavioral variation and describe the main
methodological advances in the field over the last few decades.
Finally, we will provide an overview of the role that animal
personality can play in several aspects of animals’ ecology.
 Definition(s) of Personality

After two decades of debate, most biologists now agree with the
definition of personality as consistent individual differences in
behavioral traits over time and across situations (Dall et al. 2004;
Réale et al. 2007). In other words, it represents the repeatable
dimension of the variation of a behavioral trait in a population.
What does this mean exactly? Most traits vary, and this variation can
be explained by three main sources: differences among individuals,
differences within individuals, and measurement errors. If we
consider measurement errors to be negligible and randomly
distributed across all the measures, we are left with differences in the
measures among and within individuals. The first one constitutes
personality and can itself be decomposed into other components as
we will see below. Variation within individuals can mostly be
explained by plastic changes in the behavior resulting from changes
in internal or external conditions experienced by individuals during
the interval of our measures (Figure 13.1). Following that definition,
personality for an individual refers to how much his/her average
behavioral phenotype (or behavioral type) deviates from the
average value of that behavior in a population. For instance, some
individuals may be more aggressive, less shy, less active, more
voracious, or more impulsive than the average individual in the
population. (Figure 13.2).
Figure 13.1 Representation of personality differences or
individual behavioral consistency. Left hand side: there are
personality differences among individuals; the behavior of three
genotypes or individuals is measured in three environments (E1, E2,
E3; here environment can be replaced by time). Although each of
these genotypes exhibits some phenotypic plasticity or some
differences in the phenotype of the behavior with the environment,
the three genotypes differ consistently in the level of their
phenotypes across environments. They are said to show similar
reaction norms, and the relative value of an individual in one
environment can be predicted by its relative value in another
environment. Furthermore, each individual only shows one fraction
of the total phenotypic variation observed. Right hand side: absence
of personality differences among individuals. In this case the
phenotypic value of an individual in one environment is
unpredictable and all individuals show the whole range of variation
observed in the population for that behavior. These two examples
represent the two extreme cases and in personality differences are
observed somewhere along the gradient between these two extremes
cases. From Réale & Dingemanse 2010.
Figure 13.2 Illustration of the distribution (a) and of the range of
variation (b–f) in a trait among individuals depending on its
repeatability. Here we simulated data for 12 individuals and 50
measurements per individual for a trait with repeatability values of
0, 0.1, 0.25, 0.5, and 0.75 (b–f, respectively). When repeatability is
very low (e.g., 0.1) phenotypic values largely overlap among
individuals. In contrast, very high repeatability corresponds to no
overlap of values among individuals (high among-individual
consistent differences). Most repeatability estimates for behavioral
traits range between 0.2 and 0.5 (Bell et al. 2009), which means that
a lot of them will show significant overlap in the data points among
individuals, showing that the consistency of ranks is not as obvious
as one would expect.
The definition of personality provided above solves several questions
about the phenomenon. First, it shows that it is important to
distinguish the emergence of consistent differences in the behavior of
groups, populations, or species, from the idea that individuals are
consistent (i.e., stable) in their behavior over time, which is clearly
the property of a given individual. It is central to understanding that
these two concepts are not equivalent. Individual consistency in
behavior is not a prerequisite for the existence of consistent
differences in behavior among individuals. For example, all
individuals could plastically change their behavior in response to
changes in their environment, while still preserving their behavioral
differences (Figure 13.1). We will detail how these concepts are
linked later in this chapter.
Considering personality as the repeatable dimension of a behavioral
trait, has also the benefit of separating the notion of personality from
the notion of behavior. Hence, under this definition, personality is a
feature of a behavioral trait in a population, and there is no such
thing as a “personality trait.” Furthermore, although students of
personality have restricted their investigation to five main traits,
personality can be studied on any type of behavior (Réale et al.
2007). Also, using the term personality without reference to a
particular behavioral trait is probably too vague to be informative.
For instance, asking whether personality and cognition should be
linked does not make much sense because there could be a
personality dimension in cognitive traits. We should rather ask
whether and why some specific behavioral traits should be linked to
specific cognitive traits.
The debate over the terminology of personality has mainly been
caused by the plethora of terms that exist around the notion of
personality, such as temperament, behavioral types, coping styles, or
behavioral syndromes (Gosling 2001; Réale et al. 2007). There are a
few nuances among these terms. For example, psychologists
differentiate temperament—the inherited and biological basis for
behavioral tendencies appearing early in life and serving as a
foundation for personality—from personality itself, which is built on
the experiences (physical, cultural…) accumulated by the individual
during its whole life (Gosling 2008). However, from a biological
standpoint there is no need to separate temperament from
personality, particularly if one is interested in the evolutionary and
ecological consequences of this phenomenon. People working on the
behavioral and neurophysiological reaction of animals in response to
a stressful situation talk about coping styles (Koolhaas et al. 1999).
People interested in examining the associations among behavioral
traits at the phenotypic or genetic level mostly refer to behavioral
syndromes (Sih et al. 2004).
The definition of personality given above also shows that the
correlation among traits into a behavioral syndrome is not a
necessary condition to define personality. The notion of behavioral
syndrome is a key biological concept and is important for the study
of animal personality. However, including the notion of syndrome in
the definition of personality generates a problem if people take that
definition too strictly and reject the existence of personality based on
this criterion. Researchers should focus on whether behavioral traits
are correlated or not, and why. It is, thus, more profitable for the
study of personality to consider that personality could be found in a
set of behavioral traits, these traits being correlated to a greater or
lesser extent.

A Brief History of the Study of Personality

A very thorough description of the long history of the study of
personality can be found in Dumont’s book (2010). In a nutshell, the
interest in individualism and individual behavioral differences that
started to develop with Ancient Greeks almost disappeared for more
than a millennium, in a Europe dominated by Christianity. The end
of the nineteenth century and the twentieth century saw a new rise of
interest in personality, with the development of modern psychology.
While most researchers were interested in human personality
differences, a few other scientists, including Pavlov, Yerkes, and
Hebb, started analyzing behavioral differences in animals (Gosling
2008), and it is possible to find a surge of interest in animal
personality in the psychological literature since the 1970s (Gosling
2001).
Surprisingly, the topic of individual behavioral differences was not
popular among biologists until the late twentieth century. Darwin’s
book On the Origin of Species by Means of Natural Selection (1859)
is a central component of any curriculum in biology, and it
demonstrates that phenotypic variation within a population is the
material on which selection operates; without phenotypic variation,
there is no possible evolution.
Traces of the idea of individual behavioral differences can be found
within the research of some pioneers in behavioral ecology.
Huntingford (1976), for example, showed a positive link between
aggression and risk taking in three-spined sticklebacks (Gasterosteus
aculeatus). Later Clark & Elhinger (1987) wrote a fundamental
review of the role of individual behavioral differences in biology and
discussed the existence and the role of behavioral syndromes.
Armitage (1986) studied individual variation in social behavior in
yellow-bellied marmots (Marmota flaviventris). Others have used
individual behavioral variation to generate hypotheses about the
ecology of species. In his review in 1977, Bekoff provided evidence of
individual behavioral variation in mammals and hypothesized that
dispersal in mammals results from the ontogeny of individual
behavioral phenotypes, which was in turn affected by social
(agonistic) interactions within a litter. In his 1967 paper, Chitty
argued that population self-regulation may be affected by individual
(genetic) differences in social behavior (aggression and spacing)
rapidly changing in frequency under natural selection pressures.
More important, different behavioral phenotypes were assumed to
show different fitnesses depending on the population density.
Following Chitty’s hypothesis, Myers and Krebs (1971) tested
whether individual variation in aggressive behavior and exploration
were linked to dispersal in two species of voles (Microtus
ochrogaster and M. pennsylvanicus) and showed some evidence for
a dispersal syndrome in these species. Clark and her colleagues
examined the ecological consequences of individual differences in
behavior in pumpkinseed sunfish (Wilson et al. 1993). This study led
Wilson et al. (1994) to publish a review that was probably one of the
papers that triggered the booming interest for personality in the
early twenty-first century. In the last 20 years, there has been an
explosion of papers on personality, both from a mechanistic and
from the evolutionary and ecological point of view. The study of
personality in ecology started at the turn of the new millennium.
Among the first to start, Drent and his colleagues initiated a major
research project on the great tits combining both work in the
laboratory and in the field, which to date, is still one of the most
complete cases of the ecological study of personality in wild animals
(Drent et al. 2003; Dingemanse et al. 2003; Van Oers et al. 2008).
The lack of popularity of animal personality among behavioral
ecologists and more broadly ecologists during the late twentieth
century reflected the fact that the focus of research on the evolution
of behavior at that time was on the concept of optimality (see
Chapter 11). Behavioral ecologists investigated the agents of selection
acting on behavior and described how long-term, invariant selection
pressures could explain the behavioral strategies currently expressed
by animals. This shaped, in important ways, the way we see the
evolution of behavior as an ecological adaptation. By focusing on this
question, however, behavioral ecologists neglected unexplained
individual differences (Dall et al. 2004). More recently, with an
increase of interest in personality research, evolutionary behavioral
ecologists have recognized the multivariate dimensions of an
organism; researchers in behavior are now equipped with both the
optimality approach and a series of tools and concepts derived from
other fields of evolutionary biology with which they can analyze the
selection pressures shaping behavioral traits as adaptations (Réale &
Dingemanse 2010).
 What Is Personality Made Of?

Personality ecologists have two main objectives: to study how eco-

evolutionary mechanisms have maintained the variation in
behavioral traits, and to investigate the implications of that variation
for the evolutionary ecology of species (Dall et al. 2004; Réale et al.
2007; Dingemanse & Réale 2013). To fulfill their first objective, they
test hypotheses from the theory of adaptive consistent differences in
behavior (Dall et al. 2004). For instance, they look at how potentially
diversifying selection pressures such as heterogeneous, frequency-
dependent, sexual antagonistic, or correlational selection can act on
behavioral traits (Dingemanse & Réale 2013). This is done by
estimating the strength and shape of selection acting on those traits
(Réale et al. 2007; Dingemanse & Réale 2013; Moiron et al. 2019). To
achieve their second objective, they analyze how individuals with
different behavioral types cope differently with specific situations
and the ecological consequences of such variation. We will provide
illustrations of these two objectives in the final section of the chapter,
but one important step in this endeavor is first to estimate the
sources of variation in a behavior in a population.
Many different sources of variation can be the origin of consistent
behavioral differences among individuals in a population (Figure
13.3). Each of them is interesting from a biological standpoint. First,
consistent individual differences at the phenotypic level can be
caused by heritable genetic differences (i.e., dominance effects or
additive genetic effects or the effects of many genes each with small
individual effects). Dominance genetic effects and inbreeding can
affect behavioral traits (Meffert et al. 2002) and thus generate
consistent individual differences in behavior. Through their
behavioral interactions with their offspring, the level of care they
provide, the concentration of hormones they transmit to offspring
though the egg or the placenta, or the environmental conditions they
provide to their offspring, parents can affect the behavior of their
offspring permanently (Groothuis & Carere 2005). These parental
effects can either be adaptive or not, and can themselves have a
genetic and an environmental basis (Räsänen & Kruuk 2007).
Environmental conditions early in life can also have persistent
effects on the behavior of an individual. For example, the sex of
neighboring sibs in utero in rodents can have persistent effects on
the ontogeny of reproductive behavior (Vom Saal et al. 1983) and
competition with offspring in a nest can modulate the aggressive
behavior of individuals later in life (Carere et al. 2005). During
ontogeny, each individual experiences particular environmental
conditions, which affect its behavior and bias the probability that this
individual encounters the same experience later in life. Thereby, a
series of feedback loops between an individual’s behavioral type and
experience or state across developmental stages can modulate the
development of behavior throughout time (Stamps & Groothuis
2010; Sih et al. 2015). Furthermore, an individual can learn to select
the most appropriate conditions for its behavioral type, a process
called niche-picking (Stamps & Groothuis 2010). In many species,
repeated social interactions with particular conspecifics have an
important effect on the development of personality and can lead to
the development of particular specialized social roles, a phenomenon
called social niche specialization (Bergmüller & Taborsky 2010;
Montiglio et al. 2013).
Figure 13.3 The broccoli model of variance decomposition of a
behavioral trait. Using a mixed-effect model approach, one can
decompose the variation measured in the trait into multiple
components. At the base of the stem, the phenotypic variance
(variance of all the measures for a trait) can be decomposed into the
among- and the within-individual variance. To do so we only need
repeated measures of the trait for a group of individuals. With
additional information on the pedigree, the among-individual
component can itself be split into additive genetic, permanent
environmental, and maternal-effect variance. Maternal-effect
variance can originate either from effects of the environment that are
affecting maternal performances and thus the focal individuals’
phenotype (i.e., maternal environmental effects) and effects of
maternal performances on the focal individuals’ phenotype caused
by genetic differences among the mothers (i.e., maternal genetic
effects). The text in the inflorescences shows some examples of
causes that are involved in the different variance components. The
percentages of variance are used as an illustration of the relative
importance of the different components of the phenotypic variance.
These percentages, however, change from one trait to another or one
population to another.
Physiological processes can translate these diverse genetic or
environmental effects into behavioral differences. For example,
coping styles are linked to variation in the activity of the
hypothalamus–pituitary–adrenocortical (HPA) axis, the main
nervous system in charge of dealing with stress (Koolhaas et al.
1999). These differences translate into differences in the production
of stress hormones (Koolhaas et al. 1999; Øverli et al. 2007; Baugh et
al. 2012) or testosterone (While et al. 2010; Baugh et al. 2012).
Furthermore, individuals with different behavioral types differ in the
activity and reactivity of the sympathetic and parasympathetic
nervous systems (Koolhaas et al. 1999, 2007). These differences are
measured by how much heart rate increases and how much it varies
in a stressful situation. Individual differences in the turnover or the
expression of neurotransmitters, such as dopamine and serotonin, or
polymorphism of these neurotransmitters are linked behavioral
differences among individuals (Van Oers & Mueller 2010; Coppens et
al. 2010).
 How to Study Personality

Repeatedly measuring a behavior trait in a group of individuals is

essential to the study of animal personality (Réale et al. 2007;
Dingemanse & Dochtermann 2012). It is the only way to separate
among- from within-individual variation and to estimate
repeatability. Repeatability can be calculated as r = VA /(VA + VW
), from the estimates of the variance among individuals (VA ) and the
variance within individuals (VW ), which can themselves be obtained
from a mixed-effect model or an analysis of variance (Dingemanse &
Dochtermann 2012). Remember that selection acts on the among-
individual portion of the phenotypic variance and, thus, phenotypic
variance (VP = VA + VW ) does not tell us much about the extent to
which a trait could be subject to selection. As we can see in Figure
13.4, selection on a trait increases with its repeatability.
Furthermore, under some conditions, repeatability provides an
estimate of the upper limit of heritability, or what evolutionary
biologists call the evolutionary potential, of a trait (Boake 1989). The
evolutionary response of a trait to selection (i.e., change in the
average trait value from one generation to the next) increases with its
heritability.
Figure 13.4 A) Fitness increases with the phenotypic value of a
behavior. Here behavioral phenotypes and fitness have been
measured X times on X individuals. However, the selection acting on
behavior [measured as the covariance between individuals’
behavioral type (i.e., its average value for the behavior) and average
individual fitness] will be null if individuals fail to express any
consistent differences in the behavior (b). Ultimately, the strength of
selection will increase as a function of the amount of variation
observed among individuals in behavioral type (panels c, d, e, and f).
Note that although the effect of behavior on fitness is identical in all
panels, selection still varies.
Knowing how many repeated measures on how many individuals
represents a challenge for personality ecologists, as these may affect
their ability to detect “significant” repeatability. Several papers have
dealt with these sampling issues (e.g., Martin et al. 2011;
Dingemanse & Dochtermann 2012), and we will not go into too much
detail here. Remember that our ability to detect significant
repeatability depends on the real repeatability of the trait analyzed
and sample size used to estimate it.
Another challenge is the interval at which successive measures on the
same individual should be taken. Extending that interval provides
estimates of repeatability that are easier to interpret but doing that
may not always be possible. We can classify the sources affecting the
phenotype of a trait in an individual at any time into two intrinsic
and the extrinsic sources (Figure 13.3). Whether a source is classified
as intrinsic or extrinsic depends on whether its effect on the
phenotype is fixed or variable during the time frame of study. For
example, genetic effects, parental, or early environmental effects,
and long-term learning/experience effects on the trait are intrinsic
sources responsible for the persistence of among-individual
phenotypic differences over the time frame of the study, and thus
belong to the individual. Intrinsic sources are also responsible for the
association (i.e., covariation) between different traits among
individuals (Dingemanse et al. 2012). In contrast, all the immediate
environmental effects affecting the phenotype of an individual
through short-term phenotypic plasticity are extrinsic sources
because they are changing the value of one trait or several traits for
that individual, within the time frame of the study. In other words,
extrinsic sources create variation in one trait or covariation between
two traits that happens within an individual and is associated with
plastic changes (Réale & Dingemanse 2010; Dingemanse et al. 2010,
2012). This dichotomy is somewhat arbitrary and depends on the
temporal scale at which the study is conducted: extrinsic sources
may affect the phenotype of one or more traits in a way that looks
persistent over a relatively short time scale, and thus become
intrinsic. For example, an animal attacked by a predator may
decrease its activity within the next few days, or temporary changes
in an individual’s body condition or state may temporarily affect one
or more of its behaviors. As a consequence repeatability declines
with the interval between the measures (Bell et al. 2009). When the
same traits or two traits are measured over a short-time interval
within an individual, a significant part of the individual consistency
estimated could be caused by such extrinsic sources of variation,
creating pseudo-personality or pseudo-syndrome.
Personality researchers also used to say that personality is illustrated
by the fact that rank order of individual phenotypes is maintained
over time (i.e., the most aggressive individual in a given context is
the most aggressive individual in another context). However,
applying this definition too strictly may lead to the flawed notion
that personality represents no overlap in the data between
individuals (see Figure 13.1); the degree to which rank order is
maintained, however, depends on the degree of repeatability of the
trait. Figure 13.3 shows that most behavioral traits with repeatability
ranging between 0.2 and 0.5 show some overlapping in the data
points among individuals. Traits are repeatable when within-
individual variation is small relative to among-individual variation.
Behavioral reaction norms provide an interesting framework to study
both personality and plasticity within the same context(Figure 13.5;
Dingemanse et al. 2010). With repeated measures of a behavior on
several individuals varying along an environment, one can use a
mixed-effect model approach to estimate personality as the variance
among individuals (at the intercept), and phenotypic plasticity as the
variance in slopes of the trait among individuals because of changes
in the environment (Dingemanse et al. 2010). This approach has the
advantage of showimg that personality can be found for plastic traits.
Figure 13.5 The relative importance of animal personality and
individual behavioral plasticity in generating consistent individual
differences. (a) individuals could theoretically all exhibit the same
range of behavioral values and fail to express any behavioral
plasticity in response to their environment. (b) Alternatively,
individuals can all exhibit a similar level of behavioral plasticity in
response to changes in their environment without showing any
variation in their behavioral type. (c) Note that even in absence of
variation in behavioral type, or plasticity, one can still detect
consistent individual differences in behavior when individuals are
measured over a different range of environmental conditions.
Personality and plasticity can be observed independently of each
other: (d) individual variation in behavioral type can be observed
although individuals express behavioral plasticity, and (e)
individuals can differ in their behavioral plasticity in absence of
variation in behavioral type. Finally, individuals can simultaneously
differ in their behavioral type and in their plasticity (f).
Personality psychologists working mostly on humans have developed
sophisticated tools to profit from the multivariate information
gathered in questionnaires. Persons filling out a personality
questionnaire such as the classical NEO-PI-R are asked to score their
attitudes, emotions, and behavior in a series of questions
representing many situations or contexts (Costa & McRae 1992).
Multivariate analyses are then run on the questionnaire responses of
many persons and the general dimensions of personality estimated
from these analyses (Nunnaly & Bernstein 1994). Psychologists use
the same general approach when working on animals and thus are
often restricted to captive or domestic animals, where keepers can
rate different individual animals (Uher 2008; Weiss & Adams 2013).
Although this approach, called the rating approach, has proven to be
highly reliable in these conditions (Weiss & Adams 2013), it is
seldom applicable to wild animals that are not as easily observable
for long periods (Réale & Dingemanse 2010). Some researchers have
used a classical ethological approach based on observation of
individuals in a natural setting, an approach called scoring (Uher
2008; Weiss & Adams 2013).
Ecologists studying animals in the wild often use an experimental
approach or what we could call the etching method. This approach is
built on the assumption that, if the personality of an individual is
mainly illustrated by its reaction to a stressful (or novel and thus
stressful) situation, then measuring the behavior of that individual in
a stressful situation will inform us about its behavioral type. This
approach has also been proposed by comparative psychologists with
a series of tests such as the open-field test, the novel-object test, the
startle test, or the mirror or the dyadic encounter test. With each of
these tests, although we can only capture a facet of the behavior, it
remains a very robust way of measuring personality in animals. Its
undeniable advantage is that contrary to the scoring approach,
experimental tests are done in standard conditions, and thus
uncontrolled extrinsic effects cannot bias estimates of differences
among individuals for the trait. In contrast, with the scoring method
it is impossible to control for every potential effect responsible for
the consistency of individuals. For example, individuals may behave
more boldly than others because they live in a safer environment. It
is thus impossible to separate the cause from the consequence do
bolder individuals live in a safer environment or do individuals that
live in a safer environment behave more boldly?

Consequences of Individual Personality

Differences for Ecology and Evolution

More and more work has analyzed the consequences of an animal’s

behavioral type for its ecology within many species. We now know
that differences in the behavioral type of organisms within
populations are associated with differences in their morphology, life
history, and performance at every life stage. Differences in
behavioral types are linked with important differences in the ecology
and fitness of organisms within populations. This is perhaps where
work on animal personality has made its biggest progress. An
example of such studies linking personality and ecology can be found
in Box 13.1. Below we describe some of the most striking links
between personality differences and some facets of an organism’s
life, suggesting that personality affects an organism’s life. The early
stages of life for most animals entails growing to reach an adult size,
while surviving predation, cannibalism by larger conspecifics, or
starvation. The way individuals negotiate this trade-off between
growth and survival can vary within populations as a function of
individuals behavioral type (Stamps 2007). Bolder, more aggressive,
or active individuals for example, could be regarded as weighing
growth more heavily over survival because these individuals would
forage more, encounter more food patches, or secure more resources
than shyer, less aggressive, or less active individuals. Boldness,
aggressiveness, and activity should also lead to higher mortality rates
(Wolf et al. 2007; Réale et al. 2010b).
Box 13.1

(a) The eastern chipmunk (Tamia striatus) is a small Sciurid

rodent living in the forests of eastern North America. Eastern
chipmunks are solitary and defend an individual burrow. They
spend the winter in their burrow system subsisting on stored food
and using torpor to save energy. (b) Chipmunks differ in how fast
they explore a novel environment (i.e., open-field arena)
throughout their life. Fast explorers are also less docile, and they
are more likely to be trapped. Faster explorers also show a higher
heart rate under restraint and a lower daily variability in cortisol
production over the summer (Montiglio et al. 2012; 2015). (c) In
some forests in Southern Quebec chipmunks rely mostly on seeds
of the American beech tree (Fagus grandifolia) for their winter
 survival and reproduction (Bergeron et al. 2011). Beech trees
show fluctuations in seed production, with masts occurring in the
fall (“F”) every 2–3 years. As a result of these fluctuations
chipmunks reproduce either in the summer preceding a mast
(dark diamonds, i.e., cohort B & D) or in the following spring
(i.e., cohorts A & C). Thus, chipmunks from the summer cohorts
are able to reproduce at 7 months of age, whereas those from the
spring cohorts are constrained to reproduce at 15 months of age.
(d) Irrespective of their birth cohort, faster explorers reach their
maximum reproductive output earlier in life than slower
explorers. Fast explorers have a higher lifetime reproductive
success than slow explorers among summer-born individuals,
whereas slower explorers have a higher lifetime reproductive
success than faster explorers among spring-born individuals
(Montiglio et al. 2014). Selection acting on behavioral type and
life history oscillates between birth cohorts. Variation in
exploration is influenced by genetic and cohort effects, suggesting
that exploration can evolve in response to selection but also that
exploration is determined by maternal effects related to
photoperiod.
Charline Couchoux

In species providing parental care, offspring have to go through a

stage during which they are dependent on the care provided by the
parents (Royle et al. 2011). That situation has led to the evolution of
conflicts between parents and their offspring, as well as among
siblings, over the care that should be provided to each offspring
(Trivers 1974). It is hard to imagine that personality differences
among offspring do not play a role in parent–offspring dynamics and
sibling conflicts. However, few studies have been published on the
question. For instance, the level of aggressiveness of siblings can
shape the intensity of begging and sibling competition in altricial
birds (Roulin et al. 2010), and broods with more aggressive offspring
may show a lower average growth rate because of an increased
sibling competition, or exhibit stronger variation in growth if more
aggressive individuals interfere with the feeding of less aggressive
ones. For precocial species, such as many ungulate species, where the
offspring is mobile a few hours after birth, differences in personality
may play an important role in the dynamics of emancipation and the
socialization of the offspring. Explorative offspring may go too far
from the herd and get caught by predators, which means that
environmental conditions may lead to selection for specific
behavioral types in offspring. While several hypotheses have been
suggested on the effects of personality during early life stages, fewer
studies have tested these hypotheses so far. Thus, the studying links
between personality and early life stages is a great subject to conduct
empirical work (Carere et al. 2005).
Dispersal, habitat selection, and space use. There comes a
point when offspring have to leave their parental environment to
establish themselves in a new place. At this point individuals differ in
how far they disperse from their natal habitat. Individual differences
in exploration, aggressiveness, or activity are often linked to
differences in whether individuals disperse. Highly aggressive
(Duckworth & Badyaev 2007), lowly neophobic (Dingemanse et al.
2003), or lowly sociable individuals (Cote & Clobert 2007) are more
likely to disperse or to disperse over longer distances. For example,
juveniles of the common lizard (Lacerta vivipara) differ in how
attracted they are to the odor of other members of their population.
Some juveniles are sociable and are consistently attracted to this
odor, while others are less sociable and are repulsed by it (Cote &
Clobert 2007). At low lizard density, the sociable juveniles had a
higher probability of dispersing than less sociable ones. In contrast,
at high density, less sociable juveniles had a higher probability of
dispersing than sociable ones.
During dispersal, animals have to select the habitat in which they
will live and reproduce. The habitats available may differ in their
level of food abundance or quality, in the level of predation risk, or in
the density of refuges available. Habitats may also differ in the social
environment, the density, the sex ratio, or even the average
behavioral type of conspecifics, and individuals with different
behavioral types can prefer to settle in different habitats. For
example, sociable dispersing lizards settle in densely populated
habitats, and less sociable individuals settle in sparsely populated
habitats (Cote & Clobert 2007).
Once established, individuals use their home range in different ways
according to their behavioral type. For example, fast-exploring North
American red squirrels (Tamiasciurus hudsonicus) are caught more
often outside their territory and at longer distances from their
territory than slow explorers (Boon et al. 2008). Fast-exploring
Siberian chipmunks (Tamias sibiricus) also show larger home range
use and higher rate of captures (Boyer et al. 2010). Roe deer
(Capreolus capreolus) use a variety of habitats from forested areas to
open agricultural fields. Using GPS data from individuals with
known personality, Bonnot et al. (2015) have shown that bolder
individuals used open habitats more frequently. Such differences in
habitat use can have consequences for survival when the risk of
mortality varies with the habitat. Elks (Cervus elaphus) with a higher
rate of movements and using open terrains close to roads more
frequently had a higher probability of being killed during the hunting
season (Ciuti et al. 2012).
Mating. An animal’s mating success depends on its behavioral type
for some behavioral traits, and thus these traits can be the target of
sexual selection (Schuett et al. 2010a). Being more active, or more
aggressive, allows males to acquire more mates in some systems
(Patterson & Schulte-Hostedde 2011; Sih et al. 2014). However,
animal populations often show alternative mating tactics (Oliveira et
al. 2008) and the coexistence of these tactics can be associated with
among-individual variation in behavioral traits (Reaney & Backwell
2007; Sih & Bell 2008). For example, in the great tit (Parus major),
males that are faster explorers tend to sire more offspring out of their
nest, while slower ones sire instead a greater proportion of the young
present in their nest (Van Oers et al. 2008; Patrick et al. 2012).
Aggressive male junglefowl secure more matings and more partners,
but sire fewer offspring per partner than less aggressive ones
(McDonald et al. 2017). Furthermore, different behavioral types may
be differentially favored by selected, depending on environmental
conditions. Fluctuations in the density of potential mates over space
and time favor differentially less active or highly active male water
striders (Aquarius remigis) (Sih et al. 2002). The evolutionary
mechanisms invoked to explain the maintenance of mating tactics
(Oliviera et al. 2008) could thus also explain the maintenance of
personality differences throughout evolution.
Personality can also act as a sexual signal. Behavior can thus be
involved in mate choice if they predictably inform potential partners
of the cost of mating (e.g., in cannibalistic spiders, Johnson & Sih
2005; Rabaneda-Bueno et al. 2014), future fecundity (Kralj-Fiser et
al. 2013), cooperation over offspring care (Schuett et al. 2011a), or
where individual behavioral types determine the probability of
encountering mates (Sih et al. 2002). Such behavior-dependent mate
choice may lead to assortative or dissasortative mating (Kralj-Fišer et
al. 2013; Montiglio et al. 2016).
An individual’s behavioral type may also affect its mate preferences
or level of choosiness. In Trinidadian guppies, fast-exploring females
strongly prefer unfamiliar males (Lucon-Xiccato et al., 2019). In
zebra finches (Taeniopygia guttata), fast-exploring females prefer to
mate with fast-exploring males, while less exploratory females do not
exhibit any preference (Schuett et al. 2011b). Interestingly, in this
species, pairs of more exploratory individuals produce heavier
offspring that reach independence in better condition (Schuett et al.
2011a, 2011b). In great tits (Parus major), fast explorers exhibited a
preference for fast-exploring females, while slow explorers did not
express any preference (Groothuis & Carere 2005). These studies
hint that the necessity for parents to coordinate their behavior
during territory defense and parental care might both favor
individual behavioral consistency and allow different behavioral
types to have equal mating success.
Social interactions and social roles. Many animals repeatedly
interact with their conspecifics, either because they live in social
groups or because they establish territories and defend it against
their neighbors. Individuals of many species exhibit consistent
differences in their tendency to seek out and tolerate conspecifics
(e.g., Trinidadian guppy, Poecilia reticulata: Budaev 1999; spiders:
Kralj-Fišer et al. 2013; yellow-bellied marmot: Armitage 1986) or in
their preferred group size (Cote et al. 2012). Differences in social
tendencies can be linked to other behavioral traits. For example,
shyness can increase the tendency to seek out conspecifics, prefer
large groups, or proximity to neighbors, which reduces predation
risk (Best et al. 2015).
Individuals may develop social preferences based on their
personality. Trinidadian guppies interact preferentially with
conspecifics of similar sociability, while bolder individuals had fewer
social ties with their shoal members (Croft et al. 2009). Bolder three-
spined sticklebacks (Gasterosteus aculeatus) maintain more social
connections, more evenly distributed across group members, than
shyer individuals (Pike et al. 2008). Fast-exploring great tits (Parus
major) form a greater number of less stable social bonds over the
winter than slow explorers, and these differences are likely to
originate from different social preferences among individuals (Aplin
et al. 2013). Because of their position within the social structure of
foraging flocks, fast explorers are more likely to discover novel food
patches than slow ones (Aplin et al. 2013). Males great tits also nest
near neighbors with similar levels of boldness (Johnson et al. 2017).
Positive assortment based on behavior can generate different social
conditions for individuals with different behavioral types
(Bergmüller & Taborsky 2010; Montiglio et al. 2013). Difference in
social conditions among groups can in turn affect selection pressures
acting on the trait. In contrast, negative assortment can homogenize
social conditions (Montiglio et al. 2018b), and different compositions
of behavioral types in a group shape the group social dynamics and
affect natural selection acting on behavior (Krause et al. 2010). In
water striders Aquarius remigis, rare but unusually aggressive males
affect the general mating dynamics of a whole group by harassing
both males and females and driving them into hiding (Sih & Watters
2005). The presence of such hyperaggressive males decreases the
intensity of sperm competition experienced by individuals in the
population (Wey et al. 2014). In social foraging species, shyer
individuals increase the group cohesion through their higher
propensity to aggregate (Pike et al. 2008; Michelena et al. 2009).
Parental care. Parental behavior can typically include caring for
the eggs (e.g., incubation, oxygenation, or insulation) or the offspring
(e.g., feeding, grooming, warming up), but also defending the
offspring directly against conspecific aggression, cannibalism, or
predators. Some parental behaviors show consistent individual
differences, while offspring that receive such reduced parental care
are smaller when hatching and are less active under predation threat
than offspring with normal parental care. For example, house
sparrows Passer domesticus show individual and sex differences in
chick provisioning rate, even after accounting for differences in the
environment during parental care (Westneat et al. 2011). In contrast,
the time spent incubating the eggs did not vary consistently among
parents (Nakagawa et al. 2007). In monogamous species, where both
parents contribute to parental care, males and females negotiate
their parental effort either in real time or over evolutionary time
(Nakagawa et al. 2007). In three-spined sticklebacks, Gasterosteus
aculeatus, males differ in the time they spend fanning their clutch
and in their willingness to defend their offspring (Stein & Bell 2014).
These two behaviors are correlated such that males spending more
time fanning their eggs are also more willing to defend them. This
suggests that males exhibit a clear behavioral syndrome linking
multiple aspects of parental behavior. Such syndrome can also be
found between the level of parental care and other, nonparentally
related behaviors. For example, brood feeding rate decreases with
aggressiveness in male blue tits Cyanistes caeruleus (Mutzel et al.
2013).
Differences in parental care among individuals can also generate
some differences in behavioral types among offspring through
programming. Male sticklebacks decrease the time they spend
fanning the eggs when exposed to predation threat; offspring that
experienced such a reduced parental care hatch at smaller sizes and
express a decrease their activity under predation threat to a greater
extent than offspring with normal parental care (Stein & Bell 2014).
Such findings parallel extensive laboratory studies in rodents. In
these animals, mothers differ in their grooming and licking behavior
they provide to the offspring during the first few days after birth
(Champagne et al. 2006). Pups that were less groomed or licked
grow to differ in many ways from pups that received more maternal
care.
Getting old. There are two ways to see potential links between
behavioral traits and age. The first one is purely developmental: at
each stage of their life, the costs and benefits associated with
expressing a particular personality may change and thus the average
personality phenotype may change with age (Niemelä et al. 2012).
The second potential link occurs through differential survival of
individuals with different personality types, which leads to changes
in the abundance of different personalities among different age
classes. In both cases, differences in personality types with age may
inform biologists about the functional role of behavior on life history
decisions.
Differential survival among individuals with different personalities
has been the topic of recent developments following the observation
that personality was linked to life-history strategies (Réale et al.
2000, 2010b; Stamps 2007; Wolf et al. 2007; Biro & Stamps 2008).
According to their theoretical model, Wolf et al. (2007) predicted
that if behavioral decisions affected an individual’s reproductive
assets and thus if behavior played a functional role in both survival
and reproduction, then we should expect a coevolution between
some behavioral traits and life-history strategies, with fast
personality phenotypes (i.e., high aggression, high boldness, fast
exploration) being associated with a fast life history (early life
reproduction and short life), and slow personality phenotypes (low
aggression, shyness and slow exploration) being associated with a
slow life history (late reproduction and long-lived type). Following
the seminal paper by Ricklefs and Wikelski (2002) on the pace of life
and the link between physiology and life history, Réale et al. (2010b)
proposed a general conceptual framework, according to which a
whole suite of personality, physiological, or immunological traits
were expected to be associated with a position along the slow-fast life
history continuum, and these links were observed for individuals
within a population, for populations within a species, or for different
species.
An increasing number of studies have tested that framework, either
completely or partially, with mixed results (Royauté et al. 2018;
Montiglio et al. 2018a). One reason for mixed results may be that
studies have not verified the two basic assumptions of the models.
First, a trade-off should be found between early reproductive effort
and survival or late reproduction. Second, behavior or physiological
traits should have a functional role either in terms of survival or
reproduction. For example, bold or aggressive types should grow
faster and mature earlier in life but die younger than shy, lowly
aggressive types. Unfortunately, studies have rarely tested for the
presence of these two assumptions. More work is needed on the link
between personality and life history decision, growth rate, age at
maturity, or senescence.

SUMMARY AND CONCLUSIONS

Over the last two decades, there has been an unprecedented
increase in studies of animal personality that is not restricted to
psychology. Recent studies have demonstrated how considering
individual consistent behavioral differences has changed the field
of animal behavior by linking several disciplines such as
psychology, physiology, development, ecology, and evolution
together to provide answers about this phenomenon. We have
listed multiple cases where the study of personality highlighted
how individuals within populations or groups are not ecologically
equivalent and have seen how investigating animal personality
pushed behavioral ecologists to establish links with other fields in
evolutionary biology.

REFERENCES
Aplin, L.M., Farine, D., Morand-Ferron, J., Cole, E.F., Cockburn, A.
& Sheldon, B.C. 2013. Individual personalities predict social
behaviour in wild networks of great tits (Parus major). Ecology
Letters, 16, 1365–1372.
Armitage, K.B. 1986. Individuality, sociability, and reproductive
success in yellow-bellied marmots. Ecology, 67, 1186–1193.
Baugh, A.T., Schaper, S.V., Hau, M., Cockrem, J.F., De Goede, P. &
Van Oers, K. 2012. Corticosterone responses differ between lines
of great tits (Parus major) selected for divergent personalities.
General and Comparative Endocrinology, 175, 488–494.
Bekoff, M. 1977. Mammalian dispersal and the ontogeny of
individual behavioral phenotypes. American Naturalist, 111, 715–
732.
Bell, A.M., Hankison, S.J. & Laskowski, K.L. 2009. The repeatability
of behaviour: a meta-analysis. Animal Behaviour, 77, 771–783.
Bergmüller, R. & Taborsky, M. 2010. Animal personality due to
social niche specialisation. Trends in Ecology and Evolution, 25,
504–511.
Best, E.C., Blomberg, S.P. & Goldizen, A.W. 2015. Shy female
kangaroos seek safety in numbers and have fewer preferred
friendships. Behavioral Ecology, 26, 639–646.
Biro, P.A. & Stamps, J.A. 2008. Are animal personality traits linked
to life-history productivity? Trends in Ecology and Evolution, 23,
361–368.
Boake, C.R. 1989. Repeatability: its role in evolutionary studies of
mating behavior. Evolutionary Ecology, 3, 173–182.
Bonnot, N., Verheyden, H., Blanchard, P., et al. 2015. Interindividual
variability in habitat use: evidence for a risk management
syndrome in roe deer? Behavioral Ecology, 26, 105–114.
Boon, A.K., Réale, D. & Boutin, S. 2008. Personality, habitat use, and
their consequences for survival in North American red squirrels.
Tamiasciurus hudsonicus. Oikos, 117, 1321–1328.
Boyer, N., Réale, D., Marmet, J., Pisanu, B. & Chapuis, J.-L. 2010.
Personality, space use and tick load in an introduced population
of Siberian chipmunks. Tamias sibiricus. Journal of Animal
Ecology, 79, 538–547.
Budaev, S.V. 1999. “Personality” in the guppy (Poecilia reticulata): A
correlational study of exploratory behavior and social tendency.
Journal of Comparative Psychology, 111, 399–411.
Carere, C., Drent, P.J., Koolhaas, J.M. & Groothuis, T.G.G. 2005.
Epigenetic effects on personality traits: early food provisioning
and sibling competition. Behaviour, 132, 1329–1355.
Champagne, F.A., Weaver, I.C.G., Diorio, J., Dymov, S., Szyf, M. &
Meaney, M.J. 2006. Maternal care associated with methylation of
the estrogen receptor-α1b promoter and estrogen receptor-α
 expression in the medial preoptic area of female offspring.
Endocrinology, 147, 2909–2915.
Chitty, D. 1967. The natural selection of self-regulatory behaviour in
animal populations. Proceeding of the Ecological Society of
Australia, 2, 51–78.
Ciuti, S., Muhly, T.B., Paton, D.G., et al. 2012. Human selection of elk
behavioural traits in a landscape of fear Proceedings of the Royal
Society of London B, 279, 4407–4416.
Clark, A.B. & Ehlinger, T.J. 1987. Pattern and adaptation in
individual behavioral differences. Perspectives in Ethology, 7, 1–
47.
Coppens, C.M., De Boer, S.F. & Koolhaas, J.M. 2010. Coping styles
and behavioural flexibility: towards underlying mechanisms.
Philosophical transactions of the Royal Society of London B, 365,
4021–4028.
Costa, P.T., Jr. & McCrae, R.R. 1992. NEO PI-R professional
manual. Odessa, FL: Psychological Assessment Resources, Inc.
Cote, J. & Clobert, J. 2007. Social personalities influence natal
dispersal in a lizard. Proceedings of the Royal Society of London
B, 274, 383–390.
Cote, J., Clobert, J., Brodin, T., Fogarty, S. & Sih, A. 2010.
Personality-dependent dispersal: characterization, ontogeny and
consequences for spatially structured populations. Philosophical
Transactions of the Royal Society of London B, 365, 4065–4076.
Cote, J., Fogarty, S. & Sih, A. 2012. Individual sociability and
choosiness between shoal types. Animal Behaviour, 83, 1469–
1476.
Croft, D.P., Krause, J., Darden, S.K., et al. 2009. Behavioural trait
assortment in a social network: patterns and implications.
Behavioral Ecology and Sociobiology, 63(10), 1495–1503.
Dall, S.R.X., Houston, A.I. & McNamara, J.M. 2004. The behavioural
ecology of personality: consistent individual differences from an
 adaptive perspective. Ecology Letters, 7, 734–739.
Darwin, C. 1859. On the origin of species by means of natural
selection, or the preservation of favoured races in the struggle
for life, First ed. London: John Murray.
Dingemanse, N.J., Both, C., Van Noordwijk, A.J., Rutten, A.L. &
Drent, P.J. 2003. Natal dispersal and personalities in great tits
(Parus major). Proceedings of the Royal Society of London B,
270, 741–747.
Dingemanse, N.J. & Dochtermann, N.A. 2012. Quantifying
individual variation in behaviour: mixed-effect modelling
approaches. Journal of Animal Ecology, 82, 39–54.
Dingemanse, N.J., Dochtermann, N.A. & Nakagawa, S. 2012.
Defining behavioural syndromes and the role of “syndrome
deviation” in understanding their evolution. Behavioral Ecology
and Sociobiology, 66, 1543–1548.
Dingemanse, N.J., Kazem, A.J.N., Réale, D. & Wright, J. 2010.
Behavioural reaction norms: animal personality meets individual
plasticity. Trends in Ecology and Evolution, 25, 81–89.
Dingemanse, N.J. & Réale, D. 2013. What is the evidence for natural
selection maintaining animal personality variation? In: C. Carere
& D. Maestripieri (eds.), Animal Personalities: Behavior,
Physiology and Evolution. Chicago: Univesity of Chicago Press.
Drent, P.J., Oers, K.V. & Noordwijk, A.J.V. 2003. Realized
heritability of personalities in the great tit (Parus major).
Proceedings of the Royal Society of London B, 270, 45–51.
Duckworth, R.A. & Badyaev, A.V. 2007. Coupling of dispersal and
aggression facilitates the rapid range expansion of a passerine
bird. Proceedings of the National Academy of Science, USA, 104,
15017–15022.
Dumont, F. 2010. A History of Personality Psychology: Theory,
Science, and Research from Hellenism to the Twenty-first
Century. Cambridge, UK: Cambridge University Press.
Gosling, S.D. 2001. From mice to men: what can we learn about
personality from animal research? Psychological Bulletin, 127,
45–86.
Gosling, S.D. 2008. Personality in non-human animals. Social and
Personality Psychology Compass, 2, 985–1001.
Groothuis, T.G.G. & Carere, C. 2005. Avian personalities:
characterization and epigenesis. Neuroscience and Biobehavioral
Reviews, 29, 137–150.
Huntingford, F.A. 1976. A comparison of the reaction of sticklebacks
in different reproductive conditions towards conspecifics and
predators. Animal Behaviour, 24, 694–697.
Johnson, K.V.A., Aplin, L., Cole, E.F., Farine, D.R., Firth, J., Patrick,
S.M. & Sheldon, B.C. 2017. Male great tits assort by personality
during the breeding season. Animal Behaviour, 128, 21–32.
Johnson, J. C. & Sih, A. 2005. Precopulatory sexual cannibalism in
fishing spiders (Dolomedes triton): a role for behavioral
syndromes. Behavior Ecology and Sociobiology, 58, 390–396.
Koolhaas, J.M., De Boer, S.F., Buwalda, B. & Van Reenen, K. 2007.
Individual variation in coping with stress: A multidimensional
approach of ultimate and proximate mechanisms. Brain,
Behavior and Evolution, 70(4), 218–226.
Koolhaas, J.M., Korte, S.M., De Boer, S.F., et al. 1999. Coping styles
in animals: current status in behavior and stress-physiology
Neuroscience and Biobehavioral Reviews, 23, 925–935.
Kralj-Fišer, S., Sanguino Mostajo, G.A., Preik, O., Pekár, S. &
Schneider, J.M. 2013. Assortative mating by aggressiveness type
in orb weaving spiders. Behavioral Ecology, 24, 824–831.
Krause, J., James, R. & Croft, D.P. 2010. Personality in the context of
social networks. Philosophical Transactions of the Royal Society
of London B, 365, 4099–4106.
Krebs, J.R., MacRoberts, M.H. & Cullen, M. 1972. Flocking and
feeding in the great tit Parus major – an experimental study. Ibis,
 114, 507–530.
Lucon-Xiccatoa, T., Bisazz, A. & Pilastro, A. 2019. Exploratory
behaviour covaries with preference for unfamiliar males in female
guppies. Animal Behaviour, 155, 217–224.
Martin, J.G.A., Nussey, D.H., Wilson, A.J. & Réale, D. 2011.
Measuring individual differences in reaction norms in field and
experimental studies: a power analysis of random regression
models. Methods in Ecology and Evolution, 2, 362–374.
McDonald, G.C., Spurgin, L.G., Fairfield, E.A., Richardson, D.S. &
Pizzari, T. 2017. Pre‐and postcopulatory sexual selection favor
aggressive, young males in polyandrous groups of red junglefowl.
Evolution, 71, 1653–1669.
Meffert, L.M., Hicks, S.K. & Regan, J.L. 2002. Nonadditive genetic
effects in animal behavior. American Naturalist, 160, S198 – 213.
Michelena, P., Sibbald, A.M., Erhard, H.W. & McLeod, J.E. 2009.
Effects of group size and personality on social foraging: the
distribution of sheep across patches. Behavioral Ecology, 20,
145–152.
Moiron, M., Laskowski, K.L. & Niemelä, P.T. 2019. Individual
differences in behaviour explain variation in survival: a meta-
analysis. Ecology Letters, 23, 399–408.
Montiglio, P.-O., Dammhahn, M., Messier, G.D. & Réale, D. 2018a.
The pace-of-life syndrome revisited: the role of ecological
conditions and natural history on the slow-fast continuum.
Behavioral Ecologyt Sociobiology, 72, 116.
Montiglio, P.O., Ferrari, C. & Réale, D. 2013. Social niche
specialization under constraints: personality, social interactions
and environmental heterogeneity. Philosophical Transactions of
the Royal Society of London B, 368, 20120343–20120343.
Montiglio, P.O., Garant, D., Pelletier, F. & Réale, D. 2014. Intra-
individual variability in fecal cortisol metabolites varies with
lifetime exploration and reproductive life history in eastern
 chipmunks (Tamias striatus). Behavioral Ecology and
Sociobiology, 69, 1–11.
Montiglio, P.O., McGlothlin, J.W. & Farine, D.R. 2018b. Social
structure modulates the evolutionary consequences of social
plasticity: a social network perspective on interacting phenotypes.
Ecology and evolution, 8, 1451–1464.
Montiglio, P.-O., Wey, T., Chang, A.T., Fogarty, S. & Sih, A. 2016.
Multiple mating reveals complex patterns of assortative mating by
personality and body size. Journal of Animal Ecology, 85, 125–
135.
Mutzel, A., Dingemanse, N.J., Araya-Ajoy, Y.G. & Kempenaers, B.
2013. Parental provisioning behaviour plays a key role in linking
personality with reproductive success. Proceedings of the Royal
Society of London B, 280, 20131019.
Myers, J.H. & Krebs, C.J. 1971. Genetic, behavioral, and reproductive
attributes of dispersing field voles Microtus pennsylvanicus and.
Microtus ochrogaster. Ecological Monographs, 41, 53–78.
Nakagawa, S., Gillespie, D.O.S., Hatchwell, B.J. & Burke, T. 2007.
Predictable males and unpredictable females: sex difference in
repeatability of parental care in a wild bird population. Journal of
Evolutionary Biology, 20, 1674–1681.
Niemelä, P. T., Vainikka, A., Hedrick, A. V. & Kortet, R. 2012.
Integrating behaviour with life history: boldness of the field
cricket, Gryllus integer, during ontogeny. Functional Ecology, 26,
450–456.
Nunnally, J.C. & Bernstein, I.H. 1994. Psychometric Theory, 3rd ed.
New York: McGraw-Hill.
Oliveira, R.F., Taborsky, M. & Brockmann, H.J. 2008. Alternative
Reproductive Strategies: An Integrative Approach. Cambridge,
UK: Cambridge University Press.
Øverli, Ø., Sørensen, C., Pulman, K.G.T., et al. 2007. Evolutionary
background for stress-coping styles: relationships between
physiological, behavioral, and cognitive traits in non-mammalian
 vertebrates Neuroscience and Biobehavioral Reviews, 31, 396–
412.
Patrick, S.C., Chapman, J.R., Dugdale, H.L., Quinn, J.L. & Sheldon,
B.C. 2012. Promiscuity, paternity and personality in the great tit.
Proceedings of the Royal Society of London B, 279, 1724–1730.
Patterson, L.D. & Schulte-Hostedde, A. 2011. Behavioural correlates
of parasitism and reproductive success in male eastern chipmunk.
Tamias striatus. Animal Behaviour, 81, 1129–1137.
Pike, T.W., Samanta, M., Lindström, J. & Royle, N. 2008.
Behavioural phenotype affects social interactions in an animal
network. Proceedings of the Royal Society of London B, 275,
2515–2520.
Rabaneda‐Bueno, R., Aguado, S., Fernández‐Montraveta, C. & Moya‐
Laraño, J. 2014. Does female personality determine mate choice
through sexual cannibalism? Ethology, 120, 238–248.
Räsänen, K. & Kruuk, L.E.B. 2007. Maternal effects and evolution at
ecological time-scales. Functional Ecology, 21, 408–421.
While, G.M., Isaksson, C., McEvoy, J., et al. 2010. Repeatable intra-
individual variation in plasma testosterone concentration and its
sex-specific link to aggression in a social lizard Hormones and
Behavior, 58, 208–213.
Wilson, D.S., Clark, A.B., Coleman, K. & Dearstyne, T. 1994. Shyness
and boldness in humans and other animals. Trends in Ecology
and Evolution, 9, 442–446.
Wilson, D.S., Coleman, K., Clark, A.B. & Biederman, L. 1993. Shy-
bold continuum in pumpkinseed sunfish (Lepomis gibbosus): An
ecological study of a psychological trait. Journal of Comparative
Psychology, 107, 250.
Wolf, M., Van Doorn, G.S., Leimar, O. & Weissing, F.J. 2007. Life-
history trade-offs favour the evolution of animal personalities.
Nature, 447, 581–584.
14
animal communication
WILLIAM A. SEARCY AND STEPHEN NOWICKI
INTRODUCTION
Many of the most spectacular and beautiful traits of animals
function in communication. Especially striking to humans are
certain visual signals, such as the vivid color patterns of coral reef
fish and arrow poison frogs and the exaggerated feather
ornaments of birds of paradise, egrets, and peafowl. Also obvious
and beautiful to humans are the auditory signals of some
animals, such as the songs of humpback whales, nightingales, and
wood thrushes. Other animal signals tend to pass us by because
of our sensory limitations; we miss out, for example, on most of
the olfactory signals of insects such as butterflies and ants, and
we are entirely oblivious to the electrical signals produced by
some groups of fishes. Whether we are personally aware of them
or not, animal signals pose interesting evolutionary questions.
Natural selection should favor the evolution of signaling
behaviors that benefit the signaler rather than the receiver, but
then why do receivers pay attention to signals that have evolved
to benefit others? Is it because signals contain information
valuable to receivers? But why should such information be
reliable (honest) enough to be worth attending to? Such
evolutionary questions will be one focus of this chapter. A second
focus will be the relationship between animal communication and
human language. Human language is far more complex and
cognitively sophisticated than the communication system of any
non-human animal. Nonetheless we still can ask about specific
ways in which human language is more advanced than animal
communication, and about the nature of commonalities between
the two. Before we get to any of these questions, however, we
define communication and its relationship to information.

Communication and Information

Communication can be defined as the production of acts or
structures that affect the behavior of other individuals and that have
evolved because of those effects (Wheeler et al. 2011). Such acts or
structures are termed signals. The first part of this definition
specifies that communication is considered to occur only if the
signals produced by one individual influence the behavior of another.
This influence can be overt and immediate or subtle and delayed.
The second part of the definition specifies that communication
involves signals that have evolved because of their effects on others.
Acts or structures that affect other individuals but have not evolved
because of those effects are termed cues rather than signals, and
their production is not considered communication. An example of a
cue is the compound 4-methylphenol, which is found in human
sweat, and which has the effect of attracting Anopheles mosquitos
(Hallem 2004. Attraction of Anopheles mosquitos (the vector for
malaria) cannot be selectively advantageous for humans, so we
cannot have evolved production of 4-methylphenol for that purpose.
4-methylphenol is therefore a cue and not a signal, and humans are
not considered to be communicating to mosquitos when they
produce this chemical.
Some definitions of communication add the requirement that
information must be transferred for communication to occur. In this
context information means a reduction in a receiver’s uncertainty
about the state of a signaler or its environment (Wheeler et al. 2011).
Thus a begging call given by a nestling bird might reduce the
uncertainty of the parents as to the nestling’s state of need, and an
alarm call given by one monkey might reduce the uncertainty of its
fellows as to whether a predator is nearby. We have not included a
stipulation of information transfer in our basic definition of
communication because of the possibility that signalers in some
cases have evolved to manipulate the behavior of receivers without
providing any information. One way for such manipulation to occur
is through sensory bias (Ryan et al. 1990): receivers have pre-
existing response biases because of selection on their sensory
systems in other contexts, and signals evolve to exploit these biases.
An example that illustrates the idea involves the leg-trembling
display used to court females by male water mites of the species
Neumania papillator. The trembling display produces vibrations in
the surrounding water at frequencies within the range produced by
the copepod prey of the mites, and female mites react to both sets of
vibrations by orienting towards and attempting to clutch the source
(Proctor 1991). A male responds to being clutched by depositing
spermatophores, which may then be taken up by the female to
fertilize her eggs. The hypothesis that females respond to the male
display as if it represents stimuli produced by prey is supported by
evidence that hungry females are more likely to respond to male
displays than are well fed females (Proctor 1991). Furthermore,
phylogenetic evidence is consistent with the female foraging behavior
having evolved first and the male display afterwards (Proctor 1992).
The sensory bias interpretation is that female response to vibrations
evolved to improve foraging, and that males evolved their trembling
signal to exploit this response. Under this interpretation, female
response is explained without reference to any information provided
by the male display.
Although non-informational explanations for signals can be
plausible, as illustrated by the water mite example, such hypotheses
make assumptions that are unlikely to hold in general. Figure 14.1
illustrates the assumptions made by informational and non-
informational interpretations of communication. Both diagrams
illustrate the evolution of the simplest kind of signaling system, one
in which a single signaler communicates with a single receiver. The
two interpretations make identical assumptions about the evolution
of the signaler side of the system: in both, signaling is assumed to
evolve to change the receiver’s behavior in the signaler’s interests.
Where the two interpretations differ is in how the receiver side of the
system is assumed to evolve. The non-informational interpretation
assumes that there is no evolutionary response on the part of the
receiver to the actions of the signaler; receivers continue to be
manipulated in the interests of the other party, and do not respond
to that manipulation evolutionarily. In contrast, the informational
interpretation assumes that receiver response does evolve, and of
course the evolution of receiver behavior, if it occurs, should be in a
direction that increases the receiver’s fitness, not the signaler’s. If
receiver response does evolve, then to maintain receiver response
there has to be some benefit to the receiver in attending to the signal,
and the only plausible benefit is an informational one. Receivers thus
are expected to maintain response only to signals that on average
provide sufficiently valuable information to make their response
adaptive.

Figure 14.1 Non-informational and informational interpretations

of animal communication. Under both interpretations, signaling
behavior evolves to influence receiver behavior in the signaler’s
interest. Under the non-informational interpretation, receiver
response does not evolve in response to the signaler’s actions, so
signalers are able to continue to manipulate receivers in the
signalers’ interests without providing information useful to the
receivers. Under the informational interpretation, receiver response
evolves to benefit the receivers, so response is only maintained if the
signals provide information of benefit to the receivers.
Under the informational interpretation, signaling systems are
expected to be at an evolutionary equilibrium, in which neither party
can do better given what the other party is currently doing. Under
the non-informational interpretation, signaling systems are viewed
as being out of equilibrium, in that superior response behaviors are
possible for receivers but for some reason have not evolved. The
broad explanation for why superior response behaviors have not
evolved is the existence of evolutionary constraints; in the water mite
case, for example, females are constrained from evolving more
advantageous responses to male signals by the necessity of
maintaining proper responses to prey. Many cases exist, however, in
which receiver responses are known to have evolved, or in which
evolutionary constraints on receiver response seem unlikely. All such
systems should be at equilibrium, and therefore should be
informational.

Signal Reliability

Signal reliability is an important issue for any communication

system that is informational. Reliability in this context basically
means honesty; signals are reliable if they are honest. A more formal
definition is that signals are reliable if some characteristic of the
signal is consistently correlated with an attribute of the signaler or
the signaler’s environment, and receivers benefit from knowing
about that attribute (Searcy & Nowicki 2005).
Signal reliability presents a puzzle whenever the interests of signaler
and receiver are not identical. The puzzle is most easily understood
with respect to signaling during aggression, as this is the context in
which interests are most clearly opposed. Assume that two
individuals are contesting for a resource that is not shareable, such
that if one wins the other loses. Assume also that the two contestants
are evenly matched in fighting ability, but that one is willing to fight
harder than the other to claim the resource; if the contest comes
down to a physical fight, this more aggressive individual will win. An
efficient solution would be for both individuals to signal their
respective levels of aggressiveness honestly, with the individual
producing the less aggressive signal giving way to the other. The
resource would then be allocated in the same way as if there had
been a fight, with both individuals benefiting from avoiding the costs
associated with fighting. The problem with such a system is that it is
vulnerable to cheating: animals that exaggerate their aggressiveness
would benefit by winning additional contests and garnering more
resources. Selection should favor exaggeration, and as the tendency
to exaggerate spreads through the population, reliability of the signal
will be undermined. Once reliability is sufficiently low, receivers
should evolve to cease responding to the signal, and once that has
happened, signalers should evolve to cease giving the signal. The
result of unchecked exaggeration should thus be the disappearance
of the signaling system.
Problems with signal reliability are also possible in a variety of other
signaling contexts. In mate choice, for example, males might benefit
from exaggerating signals used to communicate their quality to
females, in that exaggeration might induce additional females to
choose them for mating. In systems in which offspring beg for food
from their parents, individual offspring might benefit from
exaggerating their neediness, either from securing a larger
proportion of the food the parents have gathered or by inducing the
parents to work harder. The fact that these signaling systems persist
implies that some mechanism or mechanisms are at work to
maintain signal reliability. Figure 2 shows a classification of the
possible mechanisms for ensuring reliability and the kinds of signals
they lead to (Hurd & Enquist 2005). In the rest of this section, we
explain each of these mechanisms in turn.

Signaling When Interests are Not Opposed

For equilibrium signaling, the first division in the classification in
Figure 14.2 is between cases where the interests of signalers and
receivers do or do not oppose. What we have said above about the
temptations of unreliability applies only to cases in which interests
do oppose. If interests are congruent, then unreliability has no
benefits for the signaler, and selection can act to maximize
information transfer between signaler and receiver, as used to be
imagined for animal communication as a whole. Interests are most
likely to be congruent in systems in which signaling occurs between
close genetic relatives. The best known such systems are found in
eusocial insects of the order Hymenoptera.
Figure 14.2 A taxonomy in which signals are classified according
to the mechanism that maintain their reliability. See text for
explanations of the mechanisms. Based in large part on Hurd and
Enquist (2005).
Eusocial (“truly social”) species are ones in which castes of
individuals exist that for the most part do not themselves reproduce,
and instead work to aid the reproduction of others (Crespi & Yanega
1995). In Hymenopteran colonies, the worker caste consists only of
females, who are usually the daughters of a single queen and are thus
each other’s sisters or half-sisters. Because the workers do not
normally reproduce directly, they experience fitness gains only by
helping the queen and each other to raise the queen’s offspring. In
the absence of worker reproduction (Ratnieks 1988), conflicts of
interest between workers are minimal, and under these conditions
communication systems especially rich in information have evolved.
The most remarkable of such communication systems is the dance
language of honeybees.
The dance language was first described by Karl von Frisch, who was
awarded a Nobel Prize in 1973 largely for this work. The dancers are
worker bees who have gone out from the hive and found food, and
who use the dance to signal information about their discoveries to
other workers. The dance is performed on the vertical surface of the
honeycomb. The dancer moves in a figure eight pattern, the center
part of which is called the “waggle run” because the bee waggles her
abdomen during this segment. The dance contains three kinds of
information about the food that the bee has visited. First, the quality
of the food source, specifically in terms of its sugar concentration, is
positively correlated with both the rate of waggle run production and
the total time spent dancing (Seeley et al. 2000). Second, the
distance to the food source is positively correlated with the duration
of each waggle run (Seeley 1997). Third, and most remarkably, the
direction of the food is signaled by the direction of the waggle run
relative to the vertical, with the convention that the vertical
represents the direction of the sun (Von Frisch 1967). If, for example,
the food is located 60 degrees to the right of the sun, the waggle run
points 60 degrees to the right of vertical. The precision of the dance
is such that 50% of the bees recruited by the dance arrive within 7
degrees of the food source (Gould 1976).
In the context of signaling between fellow workers from the same
colony, both signalers and receivers benefit if receivers assess the
food source accurately and locate it efficiently. Thus interests are not
opposed, and the reliability of the signal is not surprising. What is
surprising is the volume and sophistication of the information that
we infer is communicated by an animal as small and as distantly
related to humans as a honeybee. Critics have questioned this
inference and have given an alternative explanation for receiver
behavior that does not depend on communication via the dance: that
recruits find the food using odor cues alone (Wenner 2002). Von
Frisch (1967) himself had shown earlier that bees both pick up scents
at food sources and release pheromones there, and later experiments
showed that bees rarely arrive at food sources that are unscented
even if they are indicated by dancing (Wenner et al. 1969). The
hypothesis that recruits follow odors rather than the dance is
attractive in providing a parsimonious explanation for the behavior
of recruits but leaves unexplained why bees would perform elaborate
dances if no one acts upon them.
A variety of evidence has since upheld the dance hypothesis. In an
especially ingenious test, harmonic radar was used to track the flight
paths of individual bees as they responded to dancing (Riley et al.
2005). The dance was performed for a food source that was placed
200 m east of the hive and that none of the tracked bees had
previously visited. Radar tracking showed that almost all the bees
responding to dancing flew quite close to the correct easterly
direction and flew out for distances close to the correct 200 meters
(Figure 14.3). Although the tracked bees arrived in the vicinity of the
unscented food source, none actually found it. Recruits that were
displaced by the researchers to new release points again flew the
direction and distance indicated by the dance, though (because of the
displacement) their flights brought them nowhere near the food
source (Figure 15.3). Thus, receivers do get fairly precise direction
and distance information from observing dances, but the final
approach to the food seems to require odor cues.
Figure 14.3 Representative flight paths taken by honey bees after
observing a waggle dance at the hive indicating the location of a
feeder (red square). Some bees departed directly from the hive (black
lines) whereas others were released at displaced points (blue lines).
The fact that the displaced bees flew in approximately the right
direction for approximately the right distance is particularly strong
evidence that the bees followed the information given in the dance
rather than following cues emanating from the food. Based on the
results of Riley et al. (2005).

Constrained Signals
Within the category of signals given between individuals with
opposing interests, a distinction is made between signals that can be
made by any individual and those whose production is somehow
constrained (Figure 14.2). Two types of constraints are considered,
physical and informational. A physical constraint means that some
individuals are unable to produce certain types of signals or certain
signal features because of the physical limitations of their signal
production mechanisms, whereas an informational constraint means
that some individuals cannot produce certain signals because they
lack the necessary information.
Signals subject to a physical constraint are termed index signals
and are often considered to be inherently honest because of an
inescapable relationship between a signal feature and the physical
characteristics of the signaler (Maynard Smith & Harper 2003). A
now-classic example involves formant frequencies in red deer
(Cervus elaphus). Male red deer produce loud roaring vocalizations
during the fall breeding season. As in other mammals, the sound of
any vocalization is initially produced by vibrations of vocal folds in
the larynx, and then passes to the outside through the vocal tract,
which in red deer is essentially a tube formed by the throat and oral
cavity. Passage through this tube emphasizes certain frequencies,
those whose wave lengths correspond to the acoustic resonances of
the tube. These emphasized frequencies are termed formants, with
the frequency of successive formants and the spacing between them
being inversely related to the tube’s length. Because the length of the
vocal tract is likely to be greater in larger animals, one would expect
these formant characteristics to be inversely correlated with body
size (Fitch & Reby 2001). In red deer, as expected, body mass is
strongly correlated with certain formant traits, especially with the
minimum spacing between formants. These correlations are not
perfect, but are nevertheless informative; variation in formant
spacing, for example, predicts about 40% of the variation among
males in body mass (Reby & McComb 2003). Both sexes of red deer
respond to such information: females prefer to approach playback of
roars with low formant spacing, while males vocalize more
themselves in reply to such roars (Reby et al. 2005; Charlton et al.
2007).
The responses of other red deer suggest that it would be
advantageous for a male to produce roars with lower formant
spacing in order to exaggerate his apparent size. In theory, cheating
on the roar in this way should not be possible, because only a truly
large individual can have the long vocal tract needed to produce the
formant attributes indicative of large size. Nevertheless, evidence
exists of both past and present exaggeration of this signal (Fitch &
Reby 2001). The larynx of most mammals is situated high in the
throat, but has descended deeper in the throat in the evolution of
humans. The “descended larynx” was long thought to be a uniquely
human characteristic, until it was shown that the larynx of red deer is
also found in a similar descended position. One explanation for the
descended larynx of red deer is that natural selection has favored
lowering the larynx because doing so elongates the vocal tract, thus
changing formant characteristics in a way that exaggerates apparent
size. Natural selection for size exaggeration thus would have led to
the gradual evolution of the larynx’s descended position. In addition
to this evolutionary exaggeration of apparent size, red deer
manipulate apparent size behaviorally by using muscles to pull the
larynx even lower when they roar. Despite the evidence for cheating
on the signal on both evolutionary and behavioral time scales, it can
still be argued that if all males have the descended laryngeal position
and all pull down on the larynx when roaring, then any remaining
variance in vocal tract length and thus in formant characteristics is
still constrained to reflect body size.
An example of an informational constraint is provided by song type
matching in songbirds. In many species of songbirds, each male
sings multiple versions of the species song; the different versions are
termed “song types.” Song type matching is a behavior in which one
individual replies to another with the same song type that the other
has just sung. Matching has been suggested to be a signal of
attention, demonstrating that the matcher is paying attention to the
bird that it matches. The signal is constrained to be honest, because
matching at above chance levels is only possible if the matcher is
indeed paying attention to the other, so that it knows what the other
has just sung. In this sense, having information about what song type
was sung and the ability to match it functions as a kind of “password”
in the signaling interaction. The constraint is an informational one
because it is information on what the other male has sung that limits
ability to match, rather than physical ability to produce the signal.

Handicaps
Signals that are not subject to constraints are sometimes said to
represent a free strategic choice (Hurd & Enquist 2005) (Figure
14.2): “free” because all individuals are able to produce such signals,
and a “strategic choice” because each individual chooses whether or
not to produce the signals based on the costs and benefits of doing
so. It is then the relationships between costs and benefits together
with individual attributes or environmental circumstances that can
combine to make such signals reliable.
The best known hypothesis for how signal costs can produce signal
reliability is the handicap principle of Amotz Zahavi. Zahavi
(1975) proposed that display characters used in mate choice will be
honest about signaler quality if they lower the signaler’s survival.
Those individuals that survive despite the costly display have passed
a test that individuals of lower quality cannot pass. Possession of the
trait thus conclusively demonstrates an individual’s quality. Zahavi
(1977) later amended the idea to allow the development of the
display to be adjusted to individual quality within the lifetime of a
signaler, removing the assumption that death must act to cull
through the signalers in order to produce reliability. A costly signal
whose development depends on individual quality is referred to as a
condition dependent handicap. Such traits are handicaps not in
the sense of a physical disability but rather in the sense of the extra
weight that an especially fast racehorse is made to carry.
Particularly important in convincing scientists of the validity of the
handicap principle were mathematical models by Grafen (1990),
which demonstrated that the assumptions of the handicap principle
can lead to signals that are both evolutionarily stable and reliable.
The models are complex, but their essentials are easily grasped from
a graphical version proposed by Johnstone (1997) and illustrated in
Figure 14.4. The model graphs the fitness costs and benefits of a
signal against a measure of its intensity, which might be size or color
for a visual display, or amplitude or production rate for an auditory
display. The model shows the fitness benefit of the display increasing
with signal intensity. If, for example, the display is a male courtship
signal, this relationship would arise because the higher a male’s
signal intensity, the more females he attracts, leading to more
offspring and thus higher fitness. The benefit curve reaches an
asymptote, based on the reasonable assumption in this mating signal
example that there is some upper limit to the number of females a
male can mate with. Two costs curves are shown, both having fitness
costs that increase linearly with increasing signal intensity. The
reason there are two cost curves is that the model assumes that
fitness costs increase more steeply for low quality signalers than for
high quality ones, another reasonable assumption. Suppose that the
signal is a call used in mate attraction, such as a frog call, and that
calling takes considerable energy, as is indeed true for calling in
frogs. In this system, a low quality signaler would be one with poor
energy reserves, and a high quality signaler ones with good energy
reserves. The fitness costs of expending energy by calling would then
increase more rapidly with calling rate for a low quality signaler than
for a high quality one. Only a single benefit curve is drawn because
we assume that receivers cannot judge signaler quality independent
of the signal, making the fitness benefit of a signal solely dependent
on its intensity.
Figure 14.4 A graphical version of the handicap mechanism. The
fitness costs of giving the signal increase linearly with signal
intensity, with a greater slope for low quality than for high quality
signalers. Fitness benefits increase in the same curve for both
categories of signalers. The optimum signaling level is found where
the difference between benefit and cost is maximized; that level is
greater for a high quality signaler than for a low quality one. Model
based on Johnstone (1997) and Grafen (1990).
The optimal signaling level is that value of signal intensity at which
the difference between signal benefit and signal costs is greatest.
Because of the difference in cost curves, the optimum level is higher
for signalers of good quality than for signalers of poor quality (Figure
14.4). Therefore, if all individuals signal at their optimum levels,
signal intensity reveals signaler quality. Cheating is not favored,
because signalers of poor quality that raise their signal intensity
beyond their equilibrium level experience a greater increase in signal
costs than in signal benefits. The signal is reliable, however, only
along the dimension in which it is costly. A fitness cost that arises
from an energy cost, for example, makes the signal reliable about
energy balance, but not about other attributes such as age or agility.
An example of a signal that illustrates the assumptions of the
handicap model is courtship drumming in the wolf spider
Hygrolycosa rubrofasciata. In this species, males court females by
drumming their abdomens against dry leaves, producing both
substrate vibrations and an air-borne sound. Females respond
preferentially to higher drumming rates both when choosing among
live males (Kotiaho et al. 1996) and when responding to playback of
drumming presented in the absence of males (Parri et al. 1997).
Captive males provisioned at higher levels drum more than males
provisioned at lower levels, demonstrating that drumming is
condition-dependent (Kotiaho 2000). As assumed in the handicap
model, the signal has a fitness cost: males induced to drum at a high
rate (by proximity to females) lose weight more rapidly and suffer
higher mortality than males drumming less (Mappes et al. 1996). In
an experiment in which both food level and drumming rate were
manipulated simultaneously, the two factors interacted in their
effects on survival (Kotiaho 2000), supporting the assumption that
the fitness costs of signaling are greater for males in poor condition
than for males in good condition.
In the wolf spider case, it is the energy cost of producing the signal
that leads to a fitness cost that maintains signal reliability, as we
suggested in laying out the general logic of the handicap theory.
Other types of costs are also possible. In some species of birds, such
as the barn swallow, females prefer males with longer tails (Møller
1988a; Vortman et al. 2011), and tail length is a predictor of aspects
of male quality (Møller 1994). Producing longer tail feathers may
impose a trivial energy cost in many cases; the major cost of this
signal instead appears to be that tail elongation reduces flight
performance (Rowe et al. 2001). In certain species of fish, such as
three-spined sticklebacks (Gasterosteus aculeatus) (Millinksi &
Bakker 1990), and birds, such as house finches (Haemorhous
mexicanus) (Hill 1991), females prefer males that are redder in color.
The red color is typically produced by carotenoid pigments, which
vertebrates are unable to synthesize, and which consequently must
be obtained in food. One cost to red coloration is that carotenoids
allocated to producing color are taken away from their alternative
functions in promoting health (Lozano 1994). Finally, some signals
have costs that are not experienced when the signals are used in
communication, but instead are experienced earlier in life, when the
signals are developing. For example, learned aspects of bird song
may have a developmental cost because song learning requires
considerable investment in the brain regions that support song
during a period of early life when resources are limited (Nowicki et
al. 1998, 2002).

Differential Benefits
In many species of birds, offspring remain in the nest and are fed by
their parents for some days or weeks after hatching. When a parent
visits the nest with food, the young produce a mix of signals:
stretching their necks upwards, opening their mouths widely, and
producing shrill calls. The common sense interpretation of these
signals, that they function to beg food from the parents, turns out to
be well justified. Young are fed more the more intensely they beg
(Krauss & Yasukawa 2013) and experimental enhancement of
begging calls via playback results in an increase in feeding by the
parents (Burford et al. 1998). Begging intensity increases with
increasing time since the last feeding (Kilner et al. 1999), so the
signal contains reliable information about the hunger (or need) of
the nestlings. Begging signals that communicate need from offspring
to parents are also found in mammals such as seals (Smiseth &
Lorentsen 2001) and meerkats (Manser et al. 2008).
Begging has some costs, as would be expected under a handicap
interpretation. Begging requires energy expenditure (McCarty 1996;
Leech & Leonard 1996), which, though slight, appears to be enough
to lower growth (Kilner 2001). Lower growth should have negative
effects on fitness. Begging also may attract predators (Haff &
Magrath 2011), which again has negative fitness consequences.
Furthermore, the energy cost of begging seems likely to act
differentially according to signaler quality, as assumed in the
handicap model. Nestlings that have been fed recently, and thus are
not very hungry, should pay a lower fitness cost for expending a
given amount of energy by begging than would a nestling that has
not been fed recently, and is accordingly nearer starvation. Given
that signal costs exist and that they increase with signal intensity at
different rates for high and low quality signalers, reliability of
begging seems well explained by our handicap model – except that
the predicted relationship between hunger and begging intensity is
directly opposite to that observed! Note with reference to Figure 14.4
that because the very-hungry chicks are the group for which fitness
costs increase more steeply with increasing signal intensity, and the
not-very-hungry chicks are the group for which fitness costs increase
less steeply, the handicap model predicts lower begging from the
very-hungry chicks than from the not-very-hungry ones, which is the
reverse of what is observed.
Although the reliability of begging cannot be explained by our
handicap model (Figure 14.4), it can be explained by an alternative
version of Grafen’s general model (Grafen 1990). In this second
version (Johnstone 1997), illustrated in Figure 14.5, a single cost line
is assumed to apply to all signalers. Two different benefit curves are
assumed, one for signalers of high need and the other for signalers of
low need. The benefit curve rises more steeply for signalers of high
need than for those of low need because the delivery of a given
amount of food has a greater impact on the fitness of a starving chick
than on the fitness of a well-fed chick. The equilibrium signaling
level is again the signal intensity at which the difference between
signal benefit and signal cost is greatest. This model predicts a higher
signaling level for signalers of high need than for signalers of low
need, which is in accord with what is observed.
Figure 14.5 A graphical version of a differential benefits model.
The benefits of signaling are assumed to increase with increasing
signal intensity more rapidly for signalers of high need than for those
of low need, while costs follow the same line for both categories of
signaler. The optimum signaling level is found where the difference
between benefit and cost is maximized; that level is greater to
signalers of high need. Model based on Johnstone (1997) and Grafen
(1990)
This alternative model may or may not be considered a handicap
model, depending on one’s perspective. Signal costs are needed to
stabilize the signaling system, but it is really the differential benefits
experienced by different categories of signalers that create the
predicted relationship between signal intensity and signaler
attributes. We therefore refer to this as a differential benefits model
(Figure 14.2).
Conventional Signals
The final category of explanations for signal reliability invokes
receiver dependent costs (Figure 14.2). These are costs that are
not intrinsic to the production or development of the signal but are
instead generated by the response of receivers. When signals are
subject to a physical constraint (like red deer roars) or a production
cost (like wolf spider drumming), the meaning of the signal, in the
sense of its information content, is determined by the signal’s
physical makeup. Signals that are not subject to physical constraints
or production costs, but only to receiver-dependent costs, have
meanings that are arbitrary with respect to their physical makeup,
and that are instead determined only by convention. For this reason,
signals that are of this type are often referred to as conventional
signals (Guilford & Dawkins 1995).
One possible example of a conventional signal is soft song in song
sparrows (Melospiza melodia) and other songbirds. Songs are long,
elaborate vocalizations produced in the context of mate attraction
and territory defense. Soft songs are simply low amplitude versions
of these vocalizations. Although in some species soft songs are
produced mainly during courtship (Dabelsteen et al. 1998), in song
sparrows they are produced only during aggression. In fact, of the
array of signals that song sparrows use during aggressive
interactions, soft song is the one best predictor of aggressive
escalation, more specifically of an actual attack (Searcy et al. 2006;
Akçay et al. 2013). The physical attribute that separates soft song
from normal, broadcast song is its low amplitude, a physical feature
that is neither costly to produce nor constrained to be associated
with aggression. At the same time, as a highly aggressive signal soft
song does appear to provoke aggressive retaliation from rivals and
thus to be subject to a receiver dependent cost (Anderson et al.
2012).
As emphasized earlier, aggression is a context in which conflicts of
interest are most extreme, and in which unreliable signaling seems
most likely to be advantageous. Can receiver-dependent costs
actually produce reliable signaling in such cases? An answer to this
question was first provided by Enquist’s (1985) game theoretical
model. The model assumes there are two possible signals: a signal of
strength that we will call A and a signal of weakness that we will call
B. An honest signaling strategy is to give A if strong and B if weak.
Individuals giving B concede defeat if their opponent gives A and
fight if their opponent gives B. Individuals giving A wait for their
opponent to concede if it gives B and fight if it gives A. Given these
assumptions, there is a temptation for weak individuals to cheat by
giving the signal of strength: a weak individual giving A will cause
honest weak individuals to concede, thus winning contests it might
otherwise lose. This benefit of cheating is balanced by a cost: weak
individuals that give the signal of strength will be attacked by strong
opponents, thus getting into fights that they will necessarily lose, and
which honest weak individuals can avoid. Game theory analysis
shows that if the cost of fighting a stronger opponent is high relative
to the value of winning contests, then cheating will not have a net
advantage, and reliability will prevail. Thus, reliable signaling is not
inevitable in such a system, but can be a stable outcome with the
right parameter values. Note that in such a conventional signaling
system, honest individuals do not actually pay a signal cost, so
honesty can be thought of as being maintained by the potential costs
of cheating rather than by realized costs (Számadó 2011).

Deception
Thus far we have laid out a number of hypotheses to explain why
animal signals are often reliable, but we have not considered the
opposite possibility: that animal signals are sometimes deceptive.
When applied to humans, the term deception implies that one
individual has an intention to cause another to form a false belief.
Intentions and beliefs are mental states that are difficult, if not
impossible, to assess in nonhuman animals. Accordingly, scientists
have adopted a definition of deception in animal communication that
does not stipulate such mental states. Here deception is defined as
occurring when a signaler produces a signal Y that is usually
associated with condition X; a receiver gives a response to Y that is
appropriate under condition X and that benefits the signaler; and
condition X does not actually hold (Searcy & Nowicki 2005). An
example helps to clarify this definition. Suppose that signal Y is an
alarm call that is usually given when a predator is present (condition
X), that receivers usually respond to the alarm by freezing, and that
freezing is appropriate in that it makes the freezing individual less
likely to be observed and attacked by the predator. Alarming would
be considered to be deceptive when the alarm is given in a context in
which no predator is actually present and the signaler benefits from
the receiver’s freezing response.
Instances that meet this definition of deception have long been
known from interspecific communication. Batesian mimicry, in
which a harmless prey species evolves a resemblance to a dangerous
one, provides many examples. Such examples typically involve visual
resemblance, as when a non-poisonous king snake evolves a pattern
of red, yellow, and black rings similar to that of a highly poisonous
coral snake, and visual predators such as birds are thereby deterred
from attacking the king snake (Greene & McDiarmid 1981; Brodie
1993). Because the visual signals provided by the king snake have
evolved in order to influence the behavior of the predators, such a
case meets our definition of communication. Mimicry can also occur
with respect to other sensory modalities, as when palatable moths
mimic the ultrasonic clicks that noxious moths use to warn off
predatory bats (Barber & Conner 2007).
Evidence has also been found for intraspecific deception. A
particularly clear example is provided by topi (Damaliscus lunatus),
a species of savannah antelope (Bro-Jørgensen & Pringle 2010). In
topi, both males and females give alarm snorts when they detect a
predator such as a lion or cheetah, and then stand staring at the
predator with ears pricked. Male topi give false alarms, that is alarms
in the absence of any predator, in one particular circumstance: when
a sexually receptive female is on the male’s territory and is starting to
leave it. A female is sexually receptive for one day, and during that
day she typically visits about 10 male territories, mating with about
four of the owners. Territory owners are much more likely to give
false alarms when a receptive female is on their territory than when
no females are present, and they are especially likely to give false
alarms when the receptive female attempts to leave. False alarms are
acoustically indistinguishable from true alarms, and males giving
them prick their ears and stare into the distance, just as they do
when giving true alarms. Females respond to playback of both true
and false alarms by first standing still briefly and then walking away
from the source of the sound. Because males give false alarms when
positioned between the female and the nearest boundary, the
walking away response usually brings the female back toward the
center of the male’s territory, thus delaying her departure. Males on
average achieve about three extra matings by using this false alarm
tactic. Males thus give a false signal and benefit from doing so, as
required by our definition of deception. Such a system can remain at
evolutionary equilibrium despite the occurrence of deception if a
sufficient proportion of the signals are reliable to make response
advantageous on average. Thus, female topi may be selected to
continue to respond to alarm calls despite the occurrence of false
alarms because the fitness consequences of failing to respond to an
honest alarm are potentially disastrous.
False alarms are used for other purposes in other species, for
example to draw competitors away from food in birds (Møller 1988b)
and monkeys (Wheeler 2009) and to cause sexual rivals to cease
moving in squirrels (Tamura 1995). Deception is also known in non-
alarm systems, for example in threat displays in stomatopod
crustaceans (Steger & Caldwell 1983) and in food calls in domestic
chickens (Gyger & Marler 1988).

Eavesdropping
Eavesdropping refers to the use of signals by unintended receivers
and represents another way that signaling systems can be diverted
from the functions for which they originally evolved. That receivers
are actually unintended is clearest in cases in which the receivers are
predators or parasites that use the signals of their victims to locate
individuals to attack. A case in point is provided by the túngara frog
(Engystomops pustulosus), whose mate attraction calls are exploited
by both predators and parasites. Male túngara frogs can produce
either simple calls, consisting of a frequency-modulated whine, or
complex calls, consisting of a whine plus one or more broad-band
chucks. The calls function in mate attraction; female túngara frogs
approach calling males or playback of calls and prefer to approach
complex calls over simple ones (Rand & Ryan 1981). A frog-eating
bat, Trachops cirrhosus, is also attracted to the male calls (Tuttle &
Ryan 1981), and like the female frogs the predatory bat is more likely
to approach complex calls than simple ones (Ryan et al. 1982).
Several species of blood-sucking flies of the genus Corethrella are
also attracted to the calls, and again respond more to complex calls
to simple ones (Bernal et al. 2006). Complex calls appear to be easier
for the bats to localize (Page & Ryan 2008), but this does not seem to
be true for the flies (Bernal et al. 2006). Although it is not always
clear why these natural enemies prefer complex calls, the existence of
this preference helps explain why the male frogs sometimes produce
simple calls even though complex calls are more effective mating
signals.
Eavesdropping has also been suggested to occur within species.
When a territorial male songbird countersings with an intruder on
his territory, or with song playback simulating an intruder,
information derived from the exchange of signals may be acted on by
other, neighboring territory owners (Peake et al. 2005) or by nearby
females (Otter et al. 1999). Such third party receivers probably
should not be classified as “unintended” in the sense that predators
and parasites are, in that influencing these other conspecifics may
have been one of the selective advantages that originally led to the
evolution of the signaling behavior and that still favors its
maintenance. Nevertheless, the perspective that communication
often occurs in extended “networks” (McGregor & Dabelsteen 1996),
rather than exclusively in signaler/receiver dyads, is an important
one and should be kept in mind when analyzing signaling systems.

Animal Communication and Human

Language

The animal signals discussed thus far have for the most part been
capable of communicating only rather simple information: the body
size of the signaler, for example, or its aggressive intentions or level
of hunger. The simplicity of these signaling systems contrasts greatly
with human language, as well as with the way that animals are
imagined to communicate in works of literature such as Watership
Down and The Jungle Book. In this section we consider whether
non-human animals are capable of greater sophistication in
communication than we have thus far seen, using human language as
a point of comparison.

Animal Signals as Symbols

In 1960, the philosopher Charles Hockett published a now-classic
paper on the “design features” of human language, features that he
thought all human languages possess and that are essential to their
functioning. Hockett started his paper by asserting that “man is the
only animal that can communicate by means of abstract symbols.”
Philosophers debate how to define symbols, but Hockett’s list of
design features included attributes that he thought necessary,
including 1) “semanticity”, the property of being able to convey
meaning, especially with reference to objects and events external to
the signaler; 2) “arbitrariness,” in the sense that that the meaning of
the signal is arbitrary with respect to the signal’s physical features,
and 3) “displacement,” meaning that the signal could be used to refer
to things that are remote in space or time from the signaling event.
Hockett suggested that some of the animal signals known to him had
one or two of these properties, but that none combined all three.
Much has been discovered about animal communication since
Hockett’s landmark paper. One important set of animal signals
discovered in the intervening years are predator-specific alarm
signals, as exemplified by the alarm calls of vervet monkeys
(Chlorocebus pygerythrus) (Seyfarth et al. 1980a, 1980b). Vervets
possess three distinct alarm calls (Figure 14.6), each given when a
different type of predator has been encountered: 1) leopard alarms –
a series of short tonal calls given for large mammalian carnivores
such as leopards, lions, and cheetahs; 2) eagle alarms – low
frequency, staccato grunts given for large raptors such as martial
eagles; and 3) snake alarms – high pitched “chutters” given for large
snakes such as pythons and cobras (Seyfarth et al. 1980a, 1980b).
Vervets hearing these alarms react differently to each: for leopard
alarms they run up into trees, for eagle alarms they look up into the
sky or run into dense bushes, and for snake alarms they look down at
the ground (Struhsaker 1967; Seyfarth et al. 1980a). In each case, the
response ought to make the monkey safer from the specific type of
predator associated with that alarm. Although vervet alarm calls
clearly contain information on predator identity, it can be claimed
that this information is not communicated to listeners, and that
instead an alarm simply alerts the monkeys to the presence of some
danger, whereupon they look around, see and identify the predator,
and then take the proper action. This alternative interpretation was
disproven in experiments in which vervets were played alarms from
a loudspeaker in the absence of any predator; under these
conditions, the monkeys still made the correct response for each of
the three alarm call types (Seyfarth et al. 1980a, 1980b).
Figure 14.6 Spectrograms of three alarm calls given by vervet
monkeys in response to three specific types of predators: leopards,
eagles, and snakes. Note that the acoustic structure of each alarm call
is distinctive, allowing receivers to discriminate between them.
Vervets respond differently to hearing each of the call types. Calls
kindly provided by R. M. Seyfarth and D. L. Cheney.
Although vervet alarm calls appear to refer to different classes of
predators, we cannot say whether these signals have meaning to
vervets in the same way that words do to humans. Does a snake
alarm produce in an animal’s mind a mental image of a snake, the
way the word “snake” does for us? To address this question, Suzuki
(2020) presented to coal tits (Periparus ater) a quasi-snake-like
object, a wooden stick made to move in a snake-like manner, paired
with playback of one of three call types from the vocal repertoire of
the Japanese tit (Parus minor), a related species that occurs in the
same habitat. Coal tits were significantly more likely to approach the
stick when paired with the Japanese tit’s snake alarm than when
paired with its general alarm call or recruitment call. Coal tits failed
to respond to the stick paired with snake alarms if the stick moved in
a non-snake-like manner. The results suggest that the snake alarm
evokes a mental “search image” for snakes in coal tits. Although
suggestive, these results do not conclusively demonstrate that
predator-specific alarms have meaning to non-human animals
equivalent to the meaning of words and other symbols to humans.
Consequently, these and similar signals in other animals are said to
be functionally referential (Macedonia & Evans 1993), implying
that they function as if they refer to things external to the animals,
without committing to whether they are fully referential.
Functionally referential alarms arguably have the property of
semanticity. These signals also have the property of arbitrariness, in
that there is typically no apparent tie between the acoustic structure
of the calls and the predator type they are associated with. These
alarms, however, lack the third property specified for symbols,
namely displacement: the calls are used only in the immediate
presence of a predator, and not when the predator is distant in either
space or time. Functionally referential signals with a similar mix of
properties are found in other animals in food calls (Slocombe &
Zuberbühler 2006) as well as in alarms (Manser 2001).
Hockett (1960) credited one animal signal with the property of
displacement: the honeybee waggle dance described earlier. Recall
that the dance conveys the quality of a food source by the rate of
waggle run production and the time spent dancing, the distance to
the food by the duration of a waggle run, and the direction to the
food by the direction of the waggle run relative to vertical. Because
the waggle dance can be used to describe a food source that is
kilometers distant, it clearly exhibits displacement. Hockett also
credited the waggle dance with semanticity, as it is used to refer to
things external to the dancer. Hockett denied, however, that the
waggle dance has the property of arbitrariness. Although one can
agree with Hockett that some of the conventions in the dance are not
arbitrary, such as the use of a longer waggle run to indicate that the
food is farther away, the convention for indicating direction could be
considered arbitrary. At any rate, the honeybee dance language
probably comes closest to meeting the criteria for symbolic
communication of any of the animal communication systems that
have thus far been studied, a remarkable feat for a tiny invertebrate.

Vocal Learning
Another design feature that Hockett ascribed to all human languages
is “traditional transmission,” by which he meant that the “detailed
conventions” of any particular language are acquired through
learning. Because human languages are primarily vocal, language
transmission is typically through “vocal learning,” though humans
also have the capacity to acquire fully functional sign languages
through “gestural learning.” Most animals, including many that rely
heavily on vocal communication, are incapable of vocal learning; in
particular, primates other than humans have relatively little capacity
in this regard (Egnor & Hauser 2004). This generalization applies
specifically to vocal production learning, in which animals learn how
to form their vocalizations by imitating others. Other aspects of vocal
learning, such as learning how to respond to vocalizations, are more
widespread (Janik & Slater 2000).
Among animals other than humans, vocal production learning has
been most thoroughly studied in songbirds. All songbirds appear to
learn their songs, but as there are over 4000 songbird species, there
is a great deal of scope for variation in patterns of learning (Beecher
& Brenowitz 2005). One common pattern is illustrated by song
sparrows. Male song sparrows prevented in early life from hearing
the songs of adults grow up to produce songs that are obviously
abnormal (Kroodsma 1977), but which nevertheless preserve some
species-typical features (Marler & Sherman 1985). Isolated males
tutored with recorded songs learn the details of those songs and
show a strong preference for learning own-species songs rather than
songs of a closely related species (Marler & Peters 1988). Young male
song sparrows are particularly likely to learn from recorded models
that they hear during a “critical learning period” that spans
approximately 10 to 100 days post-hatching (Marler & Peters 1987).
The existence of a critical learning period provides one striking
parallel with human language development; a second parallel is that
song sparrows and other songbirds pass through a subsong phase in
which they sing relatively unformed versions of their song, similar to
the babbling stage shown in human infants (Marler & Peters 1982).
In the wild, young male song sparrows do not learn songs from their
own fathers (Cassidy 1993), and instead learn from males they
encounter after dispersing from their natal territories and that are
likely to become their territorial neighbors (Nordby et al. 1999).
Vocal production learning occurs in two groups of birds other than
songbirds, hummingbirds and parrots, and in a few groups of
mammals other than humans, such as whales, dolphins, and bats
(Searcy & Nowicki 2019). Why vocal learning has evolved in these
various groups and not others is not well established, but one idea is
that the selective advantage of vocal learning lies in allowing the
expansion of the repertoire of vocal signals (Nowicki & Searcy 2014).
In humans, expansion of the vocal repertoire may have originally
been advantageous mainly in allowing information sharing among
kin (Fitch 2010) through increases in numbers of referential signals.
In the other vocal learners, however, signals subject to learning are
not referential, but instead are most often sexual signals used in
mate choice and aggressive competition, as is true of songbird song.
Increases in the vocal repertoire in such cases do not lead to greater
information sharing in the same way as seen in humans. Because
vocal production learning involves learning new referential signals in
humans but not in other animals, non-human animals can be
considered at most to approximate this design feature of human
language.

Syntax
Syntax refers to rules governing how smaller signal elements, such
as words, are assembled into longer strings, such as sentences.
Syntax in this simple sense is fairly common among animals. Many
species of songbirds, for example, sing multiple song types, each of
which conveys the same two messages: an aggressive, keep-away
message directed at same sex conspecifics, and a courtship, mate-
attraction message directed at opposite-sex conspecifics. Some
species sing such song types with “eventual variety,” meaning that a
singer produces a series of renditions of one song type before
switching to a bout of a second song type, while other species sing
with “immediate variety,” continually switching song types after a
single rendition of each. Although a species may follow one of these
syntactical rules faithfully, syntax in these songbirds is
fundamentally different than in human language in that how signal
elements are combined has little or no effect on meaning (Berwick et
al. 2011).
Some nonhuman primates show a more complex level of syntax, in
which signal elements have meanings that change when those
elements are combined. In Campbell’s monkey (Cercophithecus
campbelli), for example, two boom calls mean that a male is
separated from his group, two booms followed by a series of krak-oo
calls mean that a tree is falling, and a pair of booms plus krak-oo
calls interspersed with hok-oo calls mean that another group of
monkeys is approaching (Ouattara et al. 2009). Note that in this
example, meaning is affected by what calls are combined, and there
seems to be regularity in the order in which call types are given, but
it is not explicitly shown that order affects meaning. Syntax of this
order of complexity is known for several nonhuman primates
(Zuberbühler 2019).
The most complex level of syntax known for a nonhuman
communication system occurs not in a primate but in a bird, the
Japanese tit (Suzuki et al. 2019). Japanese tits produce A, B, and C
notes as alert calls when they perceive a nearby predator and
produce D notes to recruit others to social contexts that are not
threatening. Playback of ABC notes causes listening birds to scan
their surroundings, while playback of D notes causes listening birds
to approach. Playback of ABC-D combinations causes listeners to
approach while scanning. Crucially, ABC-D sequences are much
more effective in causing the approach while scanning response than
are D-ABC sequences (Suzuki et al. 2016), demonstrating that
element order affects response and suggesting that order affects
meaning. Japanese tits also respond to tää recruitment calls
produced by willow tits (Poecile montanus), a species that they often
flock with (Suzuki et al. 2017). Japanese tits again respond with both
approach and scanning to playback of completely novel ABC-tää
combinations, while showing very little response to equally novel
tää-ABC combinations (Suzuki et al. 2017). Thus, novel sequences
are also interpreted with respect to their element order.
Although Japanese tit syntax is surprisingly complex, it is still far
simpler than the widely varying syntactical rules of human
languages. What the essential differences are between human and
nonhuman syntax has been much debated. One proposal is that the
one truly distinguishing feature of the human “faculty of language” is
that only humans have a “capacity of recursion” (Hauser et al. 2002),
where recursion refers to the sequential placement of components
inside other components of the same type. An example is provided
by the sentence “the car the doctor drove broke down,” in which the
phrase that describes what the car did (the car broke down) has
embedded within it a specification of what car we are talking about
(the one the doctor drove). Humans are sometimes said to have an
infinite capacity of recursion, but in fact adding one more level of
recursion is more than we typically attempt in speaking and is almost
more than we can comprehend, as in “the car the doctor Sally knew
drove broke down.”
Although there is currently no clear evidence that any non-human
animal uses recursion in its natural signaling system, studies have
asked whether animals can learn to recognize recursive structures in
human-imposed systems. In one such study, Gentner et al. (2006)
trained European starlings (Sturnus vulgaris) to discriminate
between two categories of sequences made up of two types of starling
phrases, rattles (R) and warbles (W). One category included strings
such as RRWW and RRRWWW, with a general form that can be
written as RnWn. These strings have a recursive structure, with RW
pairs embedded within other RW pairs. The second category
consisted of sequences such as RWRW and RWRWRW; these can be
written as (RW)n and do not have a recursive structure. Starlings
proved capable of discriminating between the two categories of
training sequences and of generalizing this discrimination to new
examples of the sequence types. Probe stimuli were used to show
that the starlings were not using simple rules-of-thumb, such as
classifying based on whether the first two phrases were the same (RR
= recursive) or different (RW = non-recursive). One classification
strategy that was not eliminated was counting: starlings might have
counted the number of R’s given first and then the number of W’s
given subsequently, and accepted the sequence as recursive when
those numbers were equal (Corballis 2007). Ironically, counting is a
strategy that humans often follow to solve this discrimination
problem. Another even simpler possibility is that the starlings used
overall acoustic similarity (i.e. which strings sounded similar) to
classify sequences, without doing any kind of syntactic analysis at all
(Van Heijningen et al. 2009). Acoustic similarity may also explain
the results of other studies in which non-human animals appear to
discriminate stimuli based on artificial syntactical rules (Beckers et
al. 2017).
Although it is still debated whether birds and other non-human
animals can be taught to judge whether strings of signals have a
proper recursive structure, no one has suggested that non-human
animals can recover meaning from such structures. It is after all both
much easier (and less useful) to discern that the sentence “the car the
doctor Sally knew drove broke down” has the correct number of
verbs relative to subjects than to figure out that the car broke down,
the doctor drove, and Sally knew. Once again, the communication
capabilities of other animals fall short of those shown in human
language.

Pragmatics
In the study of human language, pragmatics concerns how context
influences communication in general and meaning in particular.
Context definitely has important effects on meaning in human
speech, effects that are traditionally illustrated using statements that
are ambiguous without knowledge of the context in which they are
given. As one example, the statement “the missionaries are ready to
eat” (Mey 2001) has a different meaning if given when a group of
religious workers are sitting down to dinner than if given after the
same individuals have been captured by cannibals. Context also
affects communication acts in humans, with speakers altering their
speech with respect to the composition of the audience they are
addressing, the presumed state of knowledge of the individuals in
that audience, and so forth.
Context has been demonstrated to affect the interpretation of signals
in non-human animals just as in humans. In song sparrows, for
example, territory owners are normally less aggressive towards
playback of a neighbor’s song than towards playback of a stranger’s
song, as long as the neighbor’s song is played from the correct
boundary, that is the one shared with the neighbor in question
(Stoddard et al. 1991). If, however, playback is first used to simulate
the intrusion of a neighbor onto a male’s territory, that male is
subsequently much more aggressive towards boundary playback of
the song of this “bad neighbor” than towards song of an unoffending
“good neighbor” (Akçay et al. 2009). Moreover, if playback is first
used to simulate the intrusion of one neighbor (the “defector”) onto
the territory of another neighbor (the “victim”), subjects are
subsequently more aggressive towards boundary playback of the
defector’s songs than of the victim’s songs (Akçay et al. 2010). Song
sparrows thus modulate their aggressive response to song not just
with respect to the presumed signaler’s past behavior towards
themselves, but also with respect to the presumed signaler’s past
behavior towards others.
Context in non-human animals can affect signaler behavior as well as
receiver response. In audience effects, for example, the presence
and identity of potential receivers affects patterns of signaling. One
of the first demonstrations of such effects was in chickens: roosters
that encountered food were found to be very likely to give a food call
when a female was present, less likely to call when no audience was
present, and not at all likely to call when only a rival male was
present (Marler et al. 1986). Audience effects have also been found in
non-human primates: for example, the latency with which capuchin
monkeys (Cebus apella) give food calls in response to the discovery
of bananas was found to decrease as the number of nearby monkeys
increased (Di Bitetti 2005). More subtle audience effects have been
demonstrated in chimpanzees: in experiments in which playback was
used to simulate the approach of another individual to a silently
feeding chimpanzee, subjects were more likely to give food calls for
the approach of a closely associated individual (a “friend”) than for
less closely associated individuals (Schel et al. 2013).
Altogether, non-human animals have been shown to exhibit
considerable abilities with respect to pragmatics aspects of
communication. Humans undoubtedly do far more, in particular in
terms of adjusting signaling with respect to the state of knowledge of
their receivers. Nevertheless, pragmatics may be the aspect of
communication in which animal signaling systems most closely
resemble human language.
SUMMARY AND CONCLUSIONS
Communication is defined as the production of acts or structures
that affect the behavior of others and that have evolved because
of such effects. Signals often contain information of value to
receivers, and this value explains the maintenance of receiver
response over evolutionary time. The vulnerability of such
systems is that signalers may be selected to manipulate receiver
response by providing false information; most communication
systems can be evolutionarily stable only if such deception is
somehow held in check. One way that deception can be limited is
for signaling to be constrained by inescapable relationships
between signal features and physical characteristics of the
signaler, producing what are termed index signals. Another
possibility, which applies especially to signals of quality, is for
signals to have a fitness cost that is higher for low quality than for
high quality signalers, so that optimal signaling levels are higher
for those of high quality. A third possibility, which applies
especially to signals of need, is for the fitness benefits of signaling
to be higher for individuals of high need than for those of low
need. Finally, conventional signals are signals that are not
physically constrained and that have negligible intrinsic costs and
are instead stabilized by costs imposed by receiver responses.
Although some animal signaling systems are impressively
complex, none approaches the complexity of human language. A
subset of animal signals have some of the properties of symbols,
such as the ability to refer to things external to the signaler, but
none are accepted as being fully symbolic. A few animal groups
learn to produce their signals, but none are known to learn to
produce new referential signals with new meanings, as do
humans. Some animals follow rules about signal order, but it
seems to be very rare for signal order to affect meaning, as occurs
pervasively in human language. Animal signals approach human
language most closely with respect to pragmatics, the effects of
context on meaning and on signaling behavior, but even here
 non-human animals seem not to adjust their signals with respect
to the state of knowledge of their audience.

FURTHER READING
Principles of Animal Communication by Bradbury and Vehrencamp
(2011) provides comprehensive coverage of all aspects of animal
communication. Those interested specifically in issues relating to
signal reliability and deception might consult the books by Maynard
Smith and Harper (2003) and Searcy and Nowicki (2005), while
those interested in the relationship of animal communication to
human language will find much of value in The Evolution of
Language by Fitch (2010). Monographs on specific systems of
animal communication that are both authoritative and entertaining
include Von Frisch (1967) on the dance language of honeybees, the
two books by Cheney and Seyfarth (1990, 2007) on communication
in non-human primates, and Catchpole and Slater (2008) on
birdsong.

REFERENCES
Akçay, C., Reed, V.A., Campbell, S.E., Templeton, C.N. & Beecher,
M.D. 2010. Indirect reciprocity: song sparrows distrust aggressive
neighbours based on eavesdropping. Animal Behaviour, 80,
1041–7.
Akçay, C., Tom, M.E., Campbell, S.E. & Beecher, M.D. 2013. Song
type matching is an honest early threat signal in a hierarchical
animal communication system. Proceedings of the Royal Society
of London B, 280, 20122517.
Akçay, C., Wood, W.E., Searcy, W.A., et al. 2009. Good neighbour,
bad neighbour: song sparrows retaliate against aggressive rivals.
Animal Behaviour, 78, 97–102.
Anderson, R.C., Searcy, W.A., Hughes, M. & Nowicki, S. 2012. The
receiver-dependent cost of soft song: a signal of aggressive intent
in songbirds. Animal Behaviour, 83, 1443–8.
Barber, J.R. & Conner, W.E. 2007. Acoustic mimicry in a predator-
prey interaction. Proceedings of the National Academy of
Sciences USA, 104, 9331–4.
Beckers, G.J.L., Berwick, R.C., Okanoya, K. & Bolhuis, J.J. 2017.
What do animals learn in artificial grammar studies?
Neurosicence and Biobehavioral Reviews, 81, 238–46.
Beecher, M.D. & Brenowitz, E.A. 2005. Functional aspects of song
learning in songbirds. Trends in Ecology and Evolution, 20, 143–
9.
Bernal, X.E., Rand, A.S. & Ryan, M.J. 2006. Acoustic preferences
and localization performance of blood-sucking flies (Corethrella
Coquillett) to túngara frog calls. Behavioral Ecology, 17, 709–15.
Berwick, R.C., Okanoya, K., Beckers, G.J.L. & Bolhuis, J.J. 2011.
Songs to syntax: the linguistics of birdsong. Trends in Cognitive
Science, 15, 113–21.
Bradbury, J.W. & Vehrencamp, S.L. 2011. Principles of Animal
Communication, 2nd ed. Sunderland, MA: Sinauer.
Brodie, E.D. 1993. Differential avoidance of coral snake banded
patterns by free-ranging avian predators in Costa Rica. Evolution,
47, 227–35.
Bro-Jørgensen, J. & Pangle, W.M. 2010. Male topi antelopes snort
deceptively to retain females for mating. American Naturalist,
176, e33-e39.
Burford, J.E., Friedrich, T.J. & Yasukawa, K. 1998. Response to
playback of nestling begging in the red-winged blackbird,
Agelaius phoeniceus. Animal Behaviour, 56, 555–61.
Cassidy, A.L.E.V. 1993. Song variation and learning in island
populations of song sparrows. PhD Dissertation, University of
British Columbia, Vancouver.
Catchpole, C.K. & Slater, P.J.B. 2008. Bird Song: Biological Themes
and Variations, 2nd ed. Cambridge: Cambridge University Press.
Charlton, B.D., Reby, D. & McComb, K. 2007. Female red deer prefer
the roars of larger males. Biology Letters, 3, 382–5.
Cheney, D.L. & Seyfarth, R.M. 1990. How Monkeys See the World:
Inside the Mind of Another Species. Chicago: University of
Chicago Press.
Cheney, D.L. & Seyfarth, R.M. 2007. Baboon Metaphysics: The
Evolution of a Social Mind. Chicago: University of Chicago Press.
Corballis, M.C. 2007. Recursion, language, and starlings. Cognitive
Science, 31, 697–704.
Crespi, B.J. & Yanega, D. 1995. The definition of eusociality.
Behavioral Ecology, 6, 109–15.
Dabelsteen, T., McGregor, P.K., Lampe, H.M., Langmore, N.E. &
Holland, J. 1998. Quiet song in song birds: an overlooked
phenomenon, Bioacoustics, 9, 89–105.
Di Bitetti, M.S. 2005. Food-associated calls and audience effects in
tufted capuchin monkeys. Cebus apella nigritus. Animal
Behaviour, 69, 911–9.
Egnor, S.E.R. & Hauser, M.D. 2004. A paradox in the evolution of
primate vocal learning. Trends in Neuroscience, 27, 649–54.
Enquist, M. 1985. Communication during aggressive interactions
with particular reference to variation in choice of behaviour.
Animal Behaviour, 33, 1152–61.
Fitch, W.T. 2010. The Evolution of Language. Cambridge:
Cambridge University Press.
Fitch, W.T. & Reby, D. 2001. The descended larynx is not uniquely
human. Proceedings of the Royal Society of London B, 268,
1669–75.
Gentner, T.Q., Fenn, K.M., Margoliash, D. & Nusbaum, H.C. 2006.
Recursive syntactic pattern learning by songbirds. Nature, 440,
1204–12.
Gould, J.L. 1976. The dance-language controversy. Quarterly Review
of Biology, 51, 211–44.
Grafen, A. 1990. Biological signals as handicaps. Journal of
Theoretical Biology, 144, 517–46.
Greene, H.W. & McDiarmid, R.W. 1981. Coral snake mimicry: does it
occur? Science, 213, 1207–12.
Guilford, T. & Dawkins, M.S. 1995. What are conventional signals?
Animal Behaviour, 49, 1689–95.
Gyger, M. & Marler, P. 1988. Food calling in the domestic fowl,
Gallus gallus: the role of external referents and deception. Animal
Behaviour, 36, 358–65.
Haff, T.M. & Magrath, R.D. 2011. Calling at a cost: elevated nestling
calling attracts predators to active nests. Biology Letters, 7, 493–
5.
Hallem, E.A., Fox, A.N., Zwiebel, L.J. & Carlson, J.R. 2004.
Mosquito receptor for human-sweat odorant. Nature, 427, 212–3.
Hauser, M.D., Chomsky, N. & Fitch, W.T. 2002. The faculty of
language: what is it, who has it, and how did it evolve? Science,
298, 1569–79.
Hill, G.E. 1991. Plumage coloration is a sexually selected indicator of
male quality. Nature, 350, 337–9.
Hockett, C.F. 1960. The origin of speech. Scientific American, 203,
88–96.
Hurd, P.L. & Enquist, M. 2005. A strategic taxonomy of biological
communication. Animal Behaviour, 70, 1155–70.
Janik, V.M. & Slater, P.J.B. 2000. The different roles of social
learning in vocal communication. Animal Behaviour, 60, 1–11.
Johnstone, R.A. 1997. The evolution of animal signals. In: J.R. Krebs
& N.B. Davies (eds.), Behavioural Ecology, pp. 155–178. Oxford:
Blackwell.
Kilner, R.M. 2001. A growth cost of begging in captive canary chicks.
Proceedings of the National Acadamy of Sciences USA, 98,
11394–8.
Kilner, R.M., Noble, D.G. & Davies, N.B. 1999. Signals of need in
parent-offspring communication and their exploitation by the
common cuckoo. Nature, 397, 667–72.
Kotiaho, J.S. 2000. Testing the assumptions of conditional handicap
theory: costs and condition dependence of a sexually selected
trait. Behavioral Ecology and Sociobiology, 48, 188–94.
Kotiaho, J.S., Alatalo, R.V., Mappes, J. & Parri, S. 1996. Sexual
selection in a wolf spider: male drumming activity, body size and
viability. Evolution, 50, 1977–81.
Krauss, N. & Yasukawa, K. 2013. How do female red-winged
blackbirds allocate food within broods? Condor, 115, 198–208.
Kroodsma, D.E. 1977. A re-evaluation of song development in the
song sparrow. Animal Behaviour, 25, 390–9.
Leech, S.M. & Leonard, M.L. 1996. Is there an energetic cost to
begging in nestling tree swallows (Tachycineta bicolor)?
Proceedings of the Royal Society of London B, 263, 983–7.
Lozano, G.A. 1994. Carotenoids, parasites, and sexual selection.
Oikos, 70, 309–11.
Macedonia, J.M. & Evans, C.S. 1993. Variation among mammalian
alarm call systems and the problem of meaning in animal signals.
Ethology, 93, 177–97.
Manser, M.B. 2001. The acoustic structure of suricates’ alarm calls
varies with predator type and the level of response urgency.
Proceedings of the Royal Society London B, 268, 2315–24.
Manser, M.B., Madden, J.R., Kunc, H.P., English, S. & Clutton-
Brock, T. 2008. Signals of need in a cooperatively breeding
mammal with mobile offspring. Animal Behaviour, 76, 1805–13.
Mappes, J., Alatalo, R.V., Kotiaho, J. & Parri, S. 1996. Viability costs
of condition-dependent sexual male display in a drumming wolf
spider. Proceedings of the Royal Society of London B, 263, 785–
9.
Marler, P., Dufty, A. & Pickert, R. 1986. Vocal communication in the
domestic chicken: II. Is a sender sensitive to the presence and
nature of a receiver? Animal Behaviour, 34, 194–8.
Marler, P. & Peters, S. 1982. Subsong and plastic song: their role in
the vocal learning process. In: D.E. Kroodsma & E.H. Miller
(eds.), Acoustic Communication in Birds, Vol. 2, pp. 25–50. New
York: Academic Press.
Marler, P. & Peters, S. 1987. A sensitive period for song acquisition in
the song sparrow, Melsopiza melodia: A case of age-limited
learning. Ethology, 76, 89–100.
Marler, P. & Peters, S. 1988. The role of song phonology and syntax
in vocal learning preferences in the song sparrow. Melospiza
melodia. Ethology, 77, 125–49.
Marler, P. & Sherman, V. 1985. Innate differences in singing
behaviour of sparrows reared in isolation from adult conspecific
song. Animal Behaviour, 33, 57–71.
Maynard Smith, J. & Harper, D. 2003. Animal Signals. Oxford:
Oxford University Press.
McCarty, J.P. 1996. The energetic cost of begging in nestling
passerines. Auk, 113, 178–188.
McGregor, P.K. & Dabelsteen, T. 1996. Communication networks. In:
D.E. Kroodsma & E.H. Miller (eds.), Ecology and Evolution of
Acoustic Communication in Birds, pp. 409–25. Ithaca, New York:
Cornell University Press.
Mey, J.L. 2001. Pragmatics: An Introduction. Oxford: Blackwell.
Milinski, M. & Bakker, T.C.M. 1990. Female sticklebacks use male
coloration in mate choice and hence avoid parasitized males.
Nature, 344, 330–3.
Møller, A.P. 1988a. Female choice selects for male sexual tail
ornaments in the monogamous swallow. Nature, 332, 640–2.
Møller, A.P. 1988b. False alarm calls as a means of resource
usurpation in the great tit. Parus major. Ethology, 79, 25–30.
Møller, A.P. 1994. Male ornament size as a reliable cue to enhanced
offspring viability in the barn swallow. Proceedings of the
National Academy of Sciences USA, 91, 6929–32.
Nordby, J.C., Campbell, S.E. & Beecher, M.D. 1999. Ecological
correlates of song learning in song sparrows. Behavioral Ecology,
10, 287–97.
Nowicki, S., Peters, S. & Podos, J. 1998. Song learning, early
nutrition and sexual selection in songbirds. American Zoologist,
38, 179–90.
Nowicki, S. & Searcy, W.A. 2014. The evolution of vocal learning.
Current Opinion in Neurobiology, 28, 48–53.
Nowicki, S., Searcy, W.A. & Peters, S. 2002. Brain development, song
learning and mate choice in birds: a review and experimental test
of the “nutritional stress hypothesis”. Journal of Comparative
Physiology A, 188, 1003–14.
Otter, K., McGregor, P.K., Terry, A.M.R., et al. 1999. Do female great
tits (Parus major) assess males by eavesdropping? A field study
using interactive song playback. Proceedings of the Royal Society
of London B, 266, 1305–9.
Ouattara, K., Lemasson, A. & Zuberbühler, K. 2009. Campbell’s
monkeys concatenate vocalizations into context-specific call
sequences. Proceedings of the National Academy of Sciences
USA, 106, 22026–31.
Page, R.A. & Ryan, M.J. 2008. The effect of signal complexity on
localization performance in bats that localize frog calls. Animal
Behaviour, 76, 761–9.
Parri, S., Alatalo, R.V., Kotiaho, J. & Mappes, J. 1997. Female choice
for male drumming in the wolf spider. Hygrolycosa
 rubrofasciata. Animal Behaviour, 53, 305–12.
Peake, T.M., Matessi, G., McGregor, P.K. & Dabelsteen, T. 2005.
Song type matching, song type switching and eavesdropping in
male great tits. Animal Behaviour, 69, 1063–8.
Proctor, H.C. 1991. Courtship in the water mite Neumania
papillator: males capitalize on female adaptations for predation.
Animal Behaviour, 42, 589–98.
Proctor, H.C. 1992. Sensory exploitation and the evolution of male
mating behaviour: a cladistic test using water mites (Acari:
Parasitengona). Animal Behaviour, 44, 745–52.
Rand, A.S. & Ryan, M.J. 1981. The adaptive significance of a complex
vocal repertoire in a neotropical frog. Zeitschrift für
Tierpsychologie, 57, 209–14.
Ratnieks, F.L.W. 1988. Reproductive harmony via mutual policing by
workers in eusocial Hymenoptera. American Naturalist, 132,
217–36.
Reby, D. & McComb, K. 2003. Anatomical constraints generate
honesty: acoustic cues to age and weight in the roars of red deer
stags. Animal Behaviour, 65, 519–30.
Reby, D., McComb, K., Cargnelutti, B., et al. 2005. Red deer stags
use formants as assessment cues during intrasexual agonistic
interactions. Proceedings of the Royal Society of London B, 272,
941–7.
Riley, J.R., Greggers, U., Smith, A.D., Reynolds, D.R. & Menzel, R.
2005. The flight paths of honeybees recruited by the waggle
dance. Nature, 435, 205–7.
Rowe, L.V., Evans, M.R. & Buchanan, K.L. 2001. The function and
evolution of the tail streamer in hirundines. Behavioral Ecology,
12, 157–63.
Ryan, M.J., Fox, J.H., Wilczynski, W. & Rand, A.S. 1990. Sexual
selection for sensory exploitation in the frog. Physalaemus
pustulosus. Nature, 343, 66–7.
Ryan, M.J., Tuttle, M.D. & Rand, A.S. 1982. Bat predation and sexual
advertisement in a neotropical anuran. American Naturalist, 119,
136–9.
Schel, A.M., Machanda, Z., Townsend, S.W., Zuberbühler, K. &
Slocombe, K.E. 2013. Chimpanzee food calls are directed at
specific individuals. Animal Behaviour, 86, 955–65.
Searcy, W.A., Anderson, R.C. & Nowicki, S. 2006. Bird song as a
signal of aggressive intent. Behavioral Ecology and Sociobiology,
60, 234–241.
Searcy, W.A. & Nowicki, S. 2005. The Evolution of Animal
Communication: Reliability and Deception in Signaling Systems.
Princeton, New Jersey: Princeton University Press.
Searcy, W.A. & Nowicki, S. 2019. Birdsong learning, avian cogntion
and the evolution of language. Animal Behaviour, 151, 217–27.
Seeley, T.D. 1997. Honey bee colonies are group-level adaptive units.
American Naturalist, 150, S22-S41.
Seeley, T.D., Mikheyev, A.S. & Pagano, G.J. 2000. Dancing bees tune
both duration and rate of waggle-run production in relation to
nectar-source profitability. Journal of Compative Physiology A,
186, 813–819.
Seyfarth, R.M., Cheney, D.L. & Marler, P. 1980a. Monkey responses
to three different alarm calls: evidence of predator classification
and semantic communication. Science, 210, 801–803.
Seyfarth, R.M., Cheney, D.L. & Marler, P. 1980b. Vervet monkey
alarm calls: Semantic communication in a free-ranging primate.
Animal Behaviour, 28, 1070–1094.
Slocombe, K.E. & Zuberbühler, K. 2006. Food-associated calls in
chimpanzees: responses to food types or food preferences?
Animal Behaviour, 72, 989–999.
Smiseth, P.T. & Lorentsen, S.H. 2001. Begging and parent-offspring
conflict in grey seals. Animal Behaviour, 62, 273–279.
Steger, R. & Caldwell, R.L. 1983. Intraspecific deception by bluffing:
a defense strategy of newly molted stomatopods (Arthropoda:
Crustacea). Science, 221, 558–560.
Stoddard, P.K., Beecher, M.D., Horning, C.L. & Campbell, S.E. 1991.
Recognition of individual neighbors by song in the song sparrow,
a species with song repertoires. Behavioral Ecology and
Sociobiology, 29, 211–215.
Struhsaker, T.T. 1967. Auditory communication among vervet
monkeys (Cercopithecus aethiops). In: S.A. Altman (ed.), Social
Communication Among Primates, pp. 281–324. Chicago: Univ. of
Chicago Press.
Suzuki, T.N. 2020. Other species’ alarm calls evoke a predator-
specific search image in birds. Current Biology, 30, 1–5.
Suzuki, T.N., Griesser, M. & Wheatcroft, D. 2019. Syntactic rules in
avian vocal sequences as a window into the evolution of
compositionality. Animal Behaviour, 151, 267–274.
Suzuki, T.N., Wheatcroft, D. & Griesser, M. 2016. Experimental
evidence for compositional syntax in bird calls. Nature
Communications, 7, 10986.
Suzuki, T.N., Wheatcroft, D. & Griesser, M. 2017. Wild birds use an
ordering rule to decode novel call sequences. Current Biology, 27,
2331–2336.
Számadó, S. 2011. The cost of honesty and the fallacy of the handicap
principle. Animal Behaviour, 81, 3–10.
Tamura, M. 1995. Postcopulatory mate guarding by vocalization in
the Formosan squirrel. Behavioral Ecology and Sociobiology, 36,
377–386.
Tuttle, M.D. & Ryan, M.J. 1981. Bat predation and the evolution of
frog vocalizations in the neotropics. Science, 214, 677–678.
Van Heijningen, C.A.A., De Visser, J., Zuidema, W. & Ten Cate, C.
2009. Simple rules can explain discrimination of putative
 recursive syntactic structures by a songbird species. Proceedings
of the National Academy of Sciences USA, 106, 20538–20543.
Von Frisch, K. 1967. The Dance Language and Orientation of Bees.
Cambridge, MA: Harvard University Press.
Vortman, Y., Lotem, A., Dor, R., Lovette, I.J. & Safran, R.J. 2011. The
sexual signals of the East-Mediterranean barn swallow: a different
swallow tale. Behavioral Ecology, 22, 1344–1352.
Wenner, A.M. 2002. The elusive honey bee dance “language”
hypothesis. Journal of Insect Behavior, 15, 859–878.
Wenner, A.M., Wells, P.H. & Johnson, D.L. 1969. Honey bee
recruitment to food sources: olfaction or language? Science, 164,
84–86.
Wheeler, B.C. 2009. Monkeys crying wolf? Tufted capuchin monkeys
use anti-predator calls to usurp resources from conspecifics.
Proceedings of the Royal Society of London B, 276, 3013–3018.
Wheeler, B.C., Searcy, W.A., Christiansen, M.H., et al. 2011.
Communication. In: R. Menzel & J. Fischer (eds.), Animal
Thinking: Contemporary Issues in Comparative Cognition, pp.
187–205. Cambridge, MA: MIT Press.
Zahavi, A. 1975. Mate selection - a selection for a handicap. Journal
of Theoretical Biology, 53, 205–214.
Zahavi, A. 1977. The cost of honesty (further remarks on the
handicap principle). Journal of Theoretical Biology, 67, 603–605.
Zuberbühler, K. 2019. Evolutionary roads to syntax. Animal
Behaviour, 151, 259–265.
15
evolution of behavior
MICHAEL J. RYAN

INTRODUCTION
Darwin often referred to his theory of evolution by natural
selection as descent with modification. Descent refers to the
existence of traits because they were passed down from a
previous ancestor; modification refers to variation in such traits
shaped by selection. As we have seen in Chapter 1, Tinbergen
(1963) summarized the four main “questions” of ethology as
causation, ontogeny, evolution, and survival value. The latter two
map quite well onto “descent with modification.” In this chapter
we focus on the historical components of behavior, which is what
Tinbergen meant by “evolution” and what Darwin meant by
“descent”: How many of the details of an animal’s behavior were
inherited from its previous ancestor, and what are the general
historical patterns by which behaviors emerge? But we also
explore the inextricable links between history and selection. How
can we parse the dual influences of history and selection on
behaviors of extant species? Toward this end we will discuss
various approaches that are used to infer historical patterns of
behavioral evolution and apply some of these approaches to aid
us in understanding adaptations in a historical context.

What Is the Evolution of Behavior?

Evolutionary thinking has been applied to behavior for several

reasons. Humans are obsessed with historical roots, whether it be
their own genealogies, the derivation of Roman mythology from that
of the Greeks, or the possibility that the American game of baseball
might be a descendent of the European sport of cricket. It is this
same interest that motivates some behavioral biologists’ pursuit of
evolution. Charles Darwin (1872) analyzed the evolution of emotions
in animals, including humans. He noted that there are striking
similarities in the facial expression of humans and some of their
closest relatives and posed the question: Did such expressions come
about in each species separately, or can their similarities be best
explained by being inherited from a common ancestor of humans
and other primates? In this way he specifically applied his notion of
descent with modification to behavior.
Some of the early ethologists were trained in the traditions of
comparative morphology and taxonomy. Characters that are shared
through common descent give some indication of how species are
related to one another. During the development of ethology,
morphology provided the primary data used to derive phylogenetic
relationships much as genetic data today provide the bulk of data
used in phylogenetics. Lorenz (1941, 1971), Lorenz (1958, 1967)
thought that the analysis of differences and similarities in behavior
might also provide useful information for taxonomy and phylogeny,
as well as being used to understand the historical patterns of
behavioral evolution. In fact, Lorenz’ emphasis on comparing
homologous behaviors among species to understand patterns of
behavioral evolution presages the current trend in studies of
behavioral evolution.
Besides describing historical patterns and using behavior for
taxonomy, behavioral biologists have more recently been using
historical information to test hypotheses about the adaptive
functions of behavior (reviewed in Martins 1996; Ryan 1996; Pagel
1999; Paradis 2011). For example, consider how variation in a
species’ mating system might drive the evolution of a species’
morphology (as discussed in Chapter 12). One might hypothesize
that large testis size is an adaptation for sperm competition in
promiscuous mating systems, as opposed to having small testes in
monogamous species. This hypothesis can be tested by
reconstructing the evolution of these two traits, testis size and
mating system, determining how often the predicted correlation
between the traits evolves, and controlling for phylogenetic effects.
When we ask questions about history we are asking about the past.
Unlike the three other areas of Tinbergen’s ethology,
experimentation and observations in the wild are usually not
sufficient tools for glimpsing what has already happened. In this
chapter we will be exposed to a different logic and a different set of
tools for exploring behavior, those of historical analysis.

Patterns of behavioral evolution deduced from strong

inference
In many examples of behavioral evolution, especially older ones, a
formal phylogenetic analysis was not used or even needed to deduce
the past history of a behavior. Instead, a detailed understanding of
an animal’s behavior in the wild combined with a basic
understanding of evolution is sufficient to allow one to propose likely
scenarios for how behaviors have evolved.
A simple and stunning example concerns the star orchid of
Madagascar. In his book on how orchids are fertilized, Darwin (1862)
described this flower as having its nectar in the bottom of a flower
spur almost a foot in depth. Such nectar could only be accessed by a
very special pollinator—one with a tongue at least as long. No such
beast was known from Madagascar at the time, but Darwin predicted
one would be found. And it was—a hawkmoth with the longest
known tongue (relative to body size) in the Old World (Xanthopan
morganni praedicta; Figure 15.1). In this case, a simple observation
about a plant yielded a specific and testable prediction about a
behavioral adaptation of an animal that was not even known to exist
at the time.
Figure 15.1 (left) Darwin predicted that the Madagascar star
orchid, Angraecum sesquipedale, would require pollination by an
insect with an exceptionally long tongue. Later the hawkmoth,
Xanthopan morganni praedicta, was discovered. (Drawing
originally from Wallace, A. R. 1867, Creation by Law, Q. J. Sci
4:470–488, and reprinted in Trends Ecol. Evol. 1998,13, 259).
(right) An illustration of the Andean bat Anoura fistulata extending
its elongated tongue (from
http://www.oloommagazine.com/Articles/ArticleDetails.aspx?
ID=2463).
Another beast with a lengthy tongue was more recently discovered in
the cloud forests of Ecuador. This tongue is the longest of any
mammal and among all vertebrates, second to only the chameleon.
The tongue of the Neotropical bat Anoura fistulata is 150% of its
body length and protrudes 8.5 cm (Figure 15.1). It is so long is needs
to be stored in a shaft that goes all the way back to its heart
(Muchhala 2006). And for what purpose, we might ask, has this
tongue and its unusual sheath evolved? Harking back to Darwin, the
researcher found a type of bell flower, Centropogon nigricans, which
is bat pollinated and also has an unusually deep corolla: 8–9 cms.
The nectar is at the bottom, and only a pollinator with a very long
tongue could reach it. Thus, the bat with its tongue and the flower
with its corolla all seem to fit quite well together.
Understanding the evolution of behavior is usually not as simple as
the two examples of pollinators and flowers, where two traits fit so
well there seems to be only one logical scenario. For one, the traits of
interest often are more complex than the length of an organ.
Understanding the evolution of complex traits, such as the eye, has
always been especially challenging (Dawkins 1996).
Instead of looking for fits between traits of two disparate taxa, many
studies of behavioral evolution compare the descent and
modification of traits within a lineage. In many cases behaviors with
one function can evolve to serve a different function, just as a
reptilian jawbone evolved into a mammalian ear bone.
Julian Huxley (1914) pondered the complicated courtship patterns of
the grebe (Porpodiceps cristatus). In one part of the grebe’s
courtship, the penguin dance, the birds exhibit stereotyped behavior
patterns that bear a striking resemblance to other behaviors
associated with nest building. It seemed to Huxley unlikely that such
complex motor patterns so gracefully coordinated between mates
arose de novo and just happened to resemble behaviors used in other
contexts. Instead, Huxley suggested such display behaviors were
derived or “ritualized” from more simple motor patterns that served
other functions. How does ritualization take place?
Here is a simple example. When a dog is about to attack an
opponent, it must open its mouth; thus, the open mouth becomes a
cue of an impending attack. The open mouth may then evolve into an
exaggerated and stereotyped behavior that evolves into a threat
signal. Tinbergen (1952) summarized the ways in which an original
behavior can become ritualized into a display: the original behavior
becoming more intense, is performed more slowly, repeated
rhythmically, combined with other behaviors, and no longer directed
toward the stimulus that originally elicited it. A bird squatting and
pointing its beak skyward prior to flight evolved from behaviors
associated with preparation for flight and is one example of a
ritualized behavior. Another example is the rhythmic courtship
drumming that woodpeckers evolved from their behavior of hunting
insects. Thus, similarities in behavior within a species that serve
different functions might be indicative of how these behaviors arose.
This type of logic has also been applied to patterns by which similar
behaviors have arisen within a group of closely related species. A
stellar example involves the evolution of courtship behavior of a
group of empidid flies known as balloonflies. In some species males
form leks, areas where males aggregate solely for the purpose of
displaying to females (see Chapter 12), and present to females large
empty balloons of silk. Females choose mates as if they are
evaluating the size of the balloons. Can we possibly understand how
something like this might evolve?
Kessel studied other species of empidids and found an interesting
range of courtship behaviors. Some species hunt small insects that
they use as nuptial gifts, and some but not all of these species with
nuptial gifts gather in leks. Furthermore, the presentation of these
gifts to females varies among species: some partially wrap the nuptial
gifts in a single strand of silk, others totally wrap the gift in silk,
while the males of other species suck the juices out of the gift before
wrapping it. Even empidids that feed on nectar instead of insects buy
into this gift-giving scheme, but instead of hunting down a prey item,
they wrap the balloon around dead insects they find. And finally,
back to where we started, some species present the female with a
large empty balloon.
Given the behavior of the various species of empidids, Kessel (1955)
proposed a logical series of behavioral transitions. Among insect-
hunting empidids there is first the evolution of nuptial gift giving,
and then the adornment or “packaging” of the gift. Once packaged,
males either cheat by sucking the juices out of the prey or by not even
putting a gift in the package. At one level, Kessel’s suggested scenario
seem logically compelling. This was another example where the logic
seems tight, and the scenario seems right. As Cumming (1994) points
out, however, these scenarios are drawn from a few species in
different genera in the balloon flies’ subfamily Empidinae and are
described in a scenario from seemingly simple to complicated. The
subfamily contains more than 1500 species and this popular scenario
still needs to be verified using some of the more rigorous
phylogenetic tools we will discuss below.
There is also an alternative to Kessel’s explanation. Perhaps the
females’ yearning for a nuptial gift is exploited by males who
substitute a worthless token for a nutritious gift. Although this might
seem far-fetched, a study of another empidid fly, the dancing fly
Rhamphomyia sulcata, supports this assumption. In an experiment
the nuptial gift of a courting male was replaced by a cotton ball.
Females accepted the cotton ball as a gift, copulated with the males
for the same duration as if the males presented a small nuptial gift
(although not a large one), and males with tokens instead of real gifts
were as likely to remate (LeBas & Hockham 2005). Sakuluk (2000)
reviews similar examples in other insects with nuptial gifts.
One sort of behavioral evolution that is a bit more complicated
occurs when the behavior of one species influences the behavior of
another and vice versa. Fireflies have species-specific patterns of
flashing that are used in conspecific mate choice (summarized in
Lloyd 1984). Male fireflies fly about to search for a mate, the
stationary female responds with her own flash, and the male
approaches the female and mates with her. Or at least that is usually
how it happens. Fireflies in the genus Photuris prey on other fireflies,
including those in the genus Photinus. Female Photuris sometimes
mimic the flashing signal of female Photinus to lure unsuspecting
male Photinus. When the male Photinus approaches the mimicked
flash, he is confronted not with a conspecific female but with a
predatory heterospecific. Instead of becoming a mate to a conspecific
he becomes a meal to a heterospecific.
This game of deceit is not over yet. Male Photuris will sometimes
mimic the flash of the male Photinus to find their own conspecific
female (Photuris). When the male Photuris, looking like a male
Photinus, does so the female Photuris responds with the mimicked
flash of her prey (female Photinus). Now the tables are turned; when
the male Photuris approaches the female (Photuris) she finds that he
is a potential mate rather than a potential meal.
How did all this come about, by what pattern did this sequence of
deceitful behaviors evolve? Strong logic suggests what seems to be
the most likely hypothesis. First, each species has its own species-
specific flash pattern, then female Photuris evolves to exploit male
Photinus, then male Photuris evolve to exploit female Photuris.
Although we cannot prove that this scenario is the correct one, none
of the alternatives seem to be as feasible. If there were no species-
specific signals there would be nothing for the female Photuris to
exploit, and if female Photuris did not try to lure their prey by
mimicking the flashes of Photinus females, they could not be
exploited by Photuris males. The logic seems tight, and the scenario
seems right. Of course, the scenario is just a hypothesis that could be
wrong.

Behavior, taxonomy, and phylogenetics

As mentioned above, Lorenz and others used behavioral characters
in both taxonomy and in deducing phylogenetic relationships. Until
recently there were substantial arguments as to whether behavior
was a reliable taxonomic character (e.g., Atz 1970; Wenzel 1992;
Wimberger and de Quiroz 1996). Nowadays, however, phylogenetic
reconstruction relies almost exclusively on DNA sequence data
(Wiley & Lieberman 2011) and behavioral and morphological data
play a diminished role in constructing phylogenetic hypotheses. The
main interaction between phylogenetics and behavior usually is in
the other direction; phylogenetics provides a historical background
against which behaviorists interpret patterns of evolution and test
hypotheses of adaptation. This pursuit has become so popular that it
is sometimes forgotten that there are other approaches to behavioral
evolution, which we discussed above, and that phylogenetic methods
are not even appropriate in some circumstances (Price 1997;
discussed below).
Although behavioral characters are not often used in phylogenetic
reconstruction, they can be critical in species level taxonomy. Gilbert
White’s The Natural History and Antiquities of Selbourne,
published in 1789, for example, was one of the first treatises to
explore bird song as a means of differentiating species (Drickamer
2010). Differences between species in courtship displays continue to
often give the first hint that a currently recognized taxon might be
more than one species (Blair 1962). This is true for an obvious
reason. Species are often operationally defined as a group of
potentially interbreeding individuals, and animals often use
courtship traits to determine if a potential mate is a conspecific.
Thus, courtship displays such as plumage pattern and color, calls and
songs, flashes and pheromones can all provide the choosing sex, the
discerning taxonomist, and even the avid birder with information
about species status.

Phylogenetics and the Evolution of

Behavior

In most of the examples discussed above, variation in behavior

among species was examined and a logical scenario of how such
behaviors might have evolved was proposed. A more rigorous
approach is to examine such logical scenarios and tenable alternative
ones in a phylogenetic context. Knowing the phylogeny can eliminate
certain explanations of patterns of evolution as being unlikely and
offer further support for other scenarios. The interpretations of
behavioral phylogenetics are only as strong as the phylogenetic
hypothesis used to interpret behavioral evolution which, in turn, is
only as strong as the data and the methods used to analyze it (Ryan
1996). Phylogenetics has made an important contribution to studies
of behavioral evolution, but the caveats are serious ones, and we
should never assume that congruence between our hypothesis of
behavioral evolution and a phylogenetic hypothesis ever “proves”
anything. That is not the way science works (Platt 1964).
One of the strengths of behavioral phylogenetics is that it offers a
formal and repeatable method for considering patterns of behavioral
evolution. Phylogenetics is an active field in which much of the
research is devoted to developing techniques for reconstructing past
history. All of the techniques used are based on Darwin’s notion of
descent with modification; that is, the traits exhibited by a taxa result
from an interaction of inheritance of the trait from ancestors and
modification of the trait through time.
Most phylogenetic methods rely on extracting characters, such as
DNA sequence, morphology, or behavior, from extant species and
then hypothesizing how evolution from a common ancestor of these
taxa could have produced the observed distribution of traits we see
today. The hypothesis of phylogenetic relationships is usually
presented in the form of a tree, or some other genealogical nexus. In
phylogenetics the term “tree” is usually synonymous with
phylogenetic hypothesis. Traits or characters that are similar among
species because they were present in the common ancestor are called
homologous with the understanding that homologous traits can
vary among taxa; they can exhibit different character states such as
brown eyes and blue eyes. Traits that independently evolve to be
similar, traits that converge on one another, are homoplasious.
The pattern of shared homologous traits is the footprint of evolution;
if this pattern can be distinguished correctly from independent
evolution of traits as a response to selection, or convergence, then we
should be able to reconstruct the branching pattern or phylogenetic
tree by which taxa diverged from one another. The challenge often is
distinguishing between homologous and homoplasious traits
(Wenzel 1992).
All phylogenetic methods examine sets of homologous traits (Box
15.1). Under the assumption that evolution is a conservative process,
many scientists rely on the principle of parsimony or Occam’s
Razor (not necessarily the method of maximum parsimony—see Box
15.1), which posits that that the simplest explanation is more likely to
be correct than a more complicated one. For an example of this logic,
consider that there are approximately 5500 species of mammals and
they all have four-chambered hearts. We can ask, did this trait evolve
5500 times independently or is it present in all mammals because it
is shared through a common ancestor? Parsimony favors the latter
explanation, and if true this trait contributes support to the
hypothesis that mammals are a monophyletic group. This exercise in
logic is applied only to the evolutionary origin of the trait. It does not
make any judgment as to whether the trait has current survival
value, that is, whether it is maintained by selection.
 Box 15.1 Phylogenetic Techniques
There are several techniques commonly used in phylogenetic
reconstruction (Archibald et al. 2003; Thornton 2004).
Maximum parsimony asserts that the phylogenetic hypothesis
most likely to be correct is that which invokes the fewest number
of phylogenetic changes, summing the number of changes of each
character at each node of the phylogeny. This is a nonparametric
method that does not take into account branch lengths (which
depict the amount of character evolution since divergence from a
common ancestor). This method also does not allow for varying
the model of evolution, for example the fact that in DNA
substitutions, transitions (changes in nucleotides A↔G, or
C↔T) are more common than transversions (A↔C, A↔T,
G↔T, G↔C) despite the fact that there are more possible types of
transversions. Maximum Parsimony performs badly when there
are large differences in rates of evolution along different branches
of the tree, a phenomenon called long-branch attraction. On the
other hand, maximum parsimony is computationally simple and
was especially valuable when it was first introduced by Fitch
(1971) in the early 1970s, before the explosion in computational
power.

A more popular phylogenetic method is maximum likelihood. It

calculates the likelihood function, which determines the likelihood of
a particular tree being true given the observed data. The single tree
with the greatest likelihood value is usually considered the tree most
likely to be true. This is a parametric method that invokes specific
models of evolution. Although the ability to specify models of
evolution makes maximum likelihood a statistically more robust
approach than parsimony, it also includes a burden. Incorrect
models, meaning wrong assumptions about how characters evolve,
are more likely to produce incorrect trees.
Another model gaining more traction is Bayesian inference. This
method also uses the likelihood function but then invokes Bayes’s
theorem by taking into account the prior probability of a tree being
correct. It then calculates the posterior probability distribution based
on the prior probabilities, the specific model of evolution, and the
data being analyzed. Instead of proposing a single tree, Bayesian
inference usually proposes a set of trees with various degrees of
statistical support.
There are several methods that are used to reconstruct phylogenetic
relationships, the most common being maximum parsimony,
maximum likelihood, and Bayesian inference (Box 15.1). All three of
these methods can also be used to estimate the value of characters at
ancestral nodes, often referred to as character reconstruction. This is
a type of data of interest to many behaviorists. The same advantages
and disadvantages that characterize each of the methods when used
for phylogenetic reconstruction (Box 15.1) also apply to ancestral
character reconstruction. Not only can each technique generate
different trees and different character reconstructions on those trees,
but different assumptions incorporated into the models can result in
different outcomes. Let us consider the simple phylogenetic pattern
depicted in Figure 15.2. In this hypothetical example the form of
parental care is described for each species in a species group as well
as a less closely related species, which is referred to as the outgroup.
The species differ in having paternal care (P), maternal care (M), or
biparental care (B). How did this distribution of care arise in these
extant species? What type of care was exhibited in the ancestral
species? Or more specifically, what is the likely mode of care at Node
1 on the phylogeny? Maximum parsimony suggests that this ancestor
exhibited maternal care. Maximum likelihood agrees as it assigns a
high probability to M being the state at Node 1.
Figure 15.2 In the hypothetical phylogeny branch length is
proportional to time. Letters at the tips note the type of parental care
exhibited. M, maternal; B, biparental; P, paternal. Nodes 1 and 2 are
indicated.
The two methods differ, however, in their hypothesis as to the
character state at Node 2.
Maximum parsimony hypothesizes that all three character states are
equally likely to characterize this ancestor. Any of the character
states at Node 2 would require three evolutionary changes to result
in the distribution of character states among the extant species.
Maximum likelihood, however, assigns a high probability to
character state M because there is little opportunity for evolution
between Node 1 and Node 2. The difference between the methods in
this case is because maximum likelihood but not maximum
parsimony takes into account branch lengths.

Describing patterns of evolution

Many groups of closely related animals tend to share similar
behaviors, but there are often exceptions as some taxa lack these
behaviors. Examples include voiceless frogs, flightless birds, and
day-flying moths. A fundamental question about behavioral
evolution asks how often do similar behaviors evolve within a group.
Alternative male reproductive tactics are common in a number of
species (Oliveira et al. 2008). Damselflies are unusual in that females
exhibit alternative reproductive tactics; some species have two
morphs, one of which is nearly identical in color pattern to the males.
This female polymorphism is especially common on the Hawaiian
Islands where at least 7 of the 25 species investigated show this
polymorphism (Cooper et al. 2016). Plotting the female color
patterns on a phylogenetic tree shows that female polymorphism
seems to have evolved independently at least four times and thus is
more likely to be subject to recent and perhaps ongoing selection
than if there were a single origin of the trait (Figure 15.3). This type
of analysis is critical in beginning to explore hypotheses of
adaptation and identify the target species for such an investigation.
Figure 15.3 Mean body hue of males and females, including
second female morphs of Hawaiian damselflies in the genus
Megalagrion. Hue is shown in colored blocks. The island locations
are shown for each species (Kauai, Hawaii, Maui, Molokai, or Oahu;
the relative location of the islands are shown on the map). The
question marks denote unknown values. Most species show sexual
dimorphism in hue, and four species contain a female-limited
dimorphism on at least one island (M. calliphya, M. blackburni, M.
hawaiiense, and M. paludicola; from Cooper et al. 2016).
A similar example involves morphology related to display behavior in
the Anolis lizards of the Caribbean (Nicholson et al. 2007). Males of
these species perform a head bob display as they extend the skin
under the chin, the dewlap. The action pattern of the display tends to
be species specific as does the color and pattern of the dewlap. There
is astounding variation in this color and pattern of the dewlap in this
group of species, which clearly results from numerous independent
evolutionary events. These data eventually provided the evidence
that dewlap pattern and color are more likely adapted to allow
discrimination from sympatric species rather than being adapted to
the local habitat.

Uncovering ancestral states

A more specific question about behavioral evolution concerns their
origin. Almost all birds fly, for example. But how did flight evolve,
from the ground up or from the trees down? The former suggests
that flapping wings were first used to increase running speed and
leaping distance, and the other suggests it was used to enhance
gliding. There still seems to be no consensus although strong
opinions abound (Padian 2001).
Phylogenetics offers a formal strategy to explore the origin and
transitions of behavioral traits. A naïve error, often made, called the
consensus method, is that the most common and most apparent
behaviors are often the ancestral state and that exceptions are
specializations that are secondarily derived from the group’s
consensus trait. This has been the train of thought in songbirds,
passerines in the subfamily of oscines. Males often have complex
song repertoires that they sing to defend their territories and attract
mates. Unlike vocalizations of most other animals (some exceptions
are hummingbirds, parrots, some cetaceans, and humans), males
learn their song. But females do sing in some species and the
assumption seems to be that female song is something that has been
derived in certain types of mating systems, such as long-term pair
bonds that tend to be more common in tropical than temperate
environs. Female song is often treated as an afterthought, and far
less attention has been given to male song—at least until recently.
Odom et al. (2014) surveyed 323 species from 34 of the 44 songbird
families where they had confidence in determining the
presence/absence of female song in a species. Of course, if female
song was documented it exists; the challenge was in determining that
the lack of evidence was compelling. Female song was present in
more species than most would have guessed, 71%. The researchers
conducted ancestral state reconstruction using both parsimony and
maximum likelihood analysis combined with different parameters,
such as varying how missing data were treated and including or
excluding a group of Old World sparrows (Paridae) whose
phylogenetic placement is troublesome. In most of the analyses by
these researchers there is strong statistical support for the
interpretation that female song is not only common but that it was
the original condition for songbirds. This suggests that sexual
selection is not the only force that might have promoted song
evolution and leads to a variety of questions about the evolution of
song learning and dimorphisms in song learning centers in the
songbird’s brain.

Testing functionality of ancestral traits

Behaviors do not fossilize but they often leave footprints, sometimes
literally. The close proximity of nests of dinosaur eggs (Horner &
Makela 1979) and intermingled footprints of adults and juveniles
(Day et al. 2002) suggest a certain degree of sociality. In other cases,
morphologies associated with behaviors are sometimes unearthed,
and biomechanical analyses are performed to determine if the
proposed function–structure correlation is tenable. For example,
biomechanical modelling shows that the wings of the earliest
Archeopteryx could support self-powered flight (Houck et al. 1990)
and that the crests of hadrosaur dinosaurs could act as resonators for
the animal’s vocalizations (Weishampel 1981).
Not all mechanisms underlying behavior are morphological.
Neurons, hormones, and genes can all have prominent influences on
behavior. An especially important set are hormones and steroids.
They play prominent roles in development, sexual differentiation,
and activation of numerous behaviors critical for survival and
reproduction. But hormones need receptors, and usually each
hormone has a receptor that binds most effectively to it. Receptors,
in turn, have their own set of genes that code for their structure. If
we had a fossil of an ancient receptor gene, we could then put it into
a living cell, synthesize the receptor, see how it works, and identify
what hormones can bind to it. Unfortunately, genes are not much
better at fossilizing than behaviors. But recent advances in molecular
evolution now allow researchers to resurrect ancient genes, describe
their DNA sequence, and determine their functionality (Thornton
2004; Dean & Thornton 2007). One of the most compelling gene
resurrection studies is of steroid receptors.
Humans have six different steroid receptors. Each binds optimally to
a different steroid hormone and then eventually activates
transcription of hormone-responsive genes. What was the ancestral
steroid receptor? Would it have been able to bind equally well to all
of the extant steroid hormones, and if not, to which hormones would
it exhibit the highest binding affinity? This is quite an old ancestor as
the ancestral steroid receptor gene existed before the protostomes
and deuterostomes split some 600 to 1,000 million years ago. This
single gene gave rise to all other genes that code for different steroid
receptors through a process of sequence divergence and gene
duplication. But what did it look like and how did it function? The
only way this question could be answered would be by resurrecting
the ancient gene.
Thornton and his colleagues (2003) utilized the following gene
resurrection protocol: they analyzed 73 steroid receptor sequences of
the sea slug Aplysia and humans and then used these data to infer
the amino acid sequence of the steroid receptor shared by the most
recent common ancestor of these two species—which of course was
not all that recent. Not all of the sites within steroid receptor genes
are critical for receptor function. Interestingly, the amino acid
sequence of the critical sites was identical between the ancestral
reconstructed gene and the modern day human estrogen receptor,
and these amino acid sequences differed from human steroid
receptors that bind to other steroid hormones.
The researchers then used PCR to recover the DNA sequence of the
ancestral gene from its amino acid sequence. Up to this point, the
study of the ancient gene is descriptive. But the researchers were
concerned with not only what the gene looked like (its DNA
sequence) but how it functioned. In addition, verifying that this
receptor actually functions at all is an experimental test of the
validity of the phylogenetic reconstruction.
The functionality of the receptor gene was tested by cloning the DNA
into vectors that were transfected into cultured mammalian cells.
The response of the receptors in these cells to estrogen was then
determined. In response to low doses of estrogen these receptors
effectively activated gene transcription as efficiently as did extant
human estrogen receptors. The same was not true in response to
other steroid hormones. These experiments show that the
reconstruction of the ancient gene really works, that the ancient gene
does make a receptor that binds to hormones, and it initiates a
cascade of responses that lead to gene activation. In addition, the
ancient receptor is an estrogen receptor.
The evolutionary interpretation of these results is intriguing.
Estrogen is the endpoint of a steroid synthesis pathway in which
testosterone and progesterone are intermediates, and in vertebrates
these products are also hormones and have their own specific
receptors. But in the “ancient” condition there were only receptors
for estrogen; the synthesis of progesterone and testosterone were
merely intermediate steps in a biosynthesis pathway. Without
receptors, progesterone and testosterone could not act as hormones
and were merely biochemical byproducts. Thus, the original receptor
was for the endpoint of this pathway and the biochemical
intermediates only became hormones as this ancestral receptor
duplicated and diverged to specifically bind the intermediate
products in the synthesis pathway.

Revealing behavioral genetic innovation

Often, we are concerned with not only the ancestral state of a
behavior but with the specific changes that have taken place in
modification of the behavior. At this point, it is worth noting once
again that behavior has underlying mechanisms. Preferences, for
example, arise from the interaction of sensory, perceptual, and
cognitive biases and stimulus variation. Our attraction to various
foods, for example, can be traced in part to our taste buds. To
understand the evolution of a behavior such as a taste preference it is
necessary to understand how the mechanisms modulating taste
evolve.
Imagine losing your sense of taste. Now imagine losing your sense of
taste if your job was to taste things, or at least make things taste good
for others. This was the fate that befell the chef Grant Achatz, one of
the leaders in the field of molecular gastronomy whose goal is to take
familiar foods and use scientific techniques to give them new tastes
and textures. Achatz’s restaurant, the Alinea in Chicago, is one of the
best at molecular gastronomy. In an almost unimaginable tragedy
Achatz, who at 33 never smoked a cigarette in his life, was diagnosed
with tongue cancer. Doctors wanted to remove his tongue and
replace it with his arm, or at least tissues from his arm. He said no,
but he did allow them to remove his lymph nodes, and he underwent
chemo- and radiation therapy. The therapy caused Achatz to lose his
hair—and for him, even worse, his sense of taste. He continued to
concoct new dishes based on his knowledge of molecular gastronomy
and his memory of taste. Of course, it was not the same as doing it all
himself. Remarkably, however, 3 years later his sense of taste slowly
returned (Max 2008).
Taste can be lost in the near term when taste buds are damaged, or in
evolutionary time when genes coding for taste receptors are lost. A
class of G-protein coupled receptors, called T1Rs, are responsible for
tastes that we experience as savory (or umami) and sweet; the T1R1-
T1R3 diamer is responsible for savory taste and the T1R2-T1R3 for
sweet taste. Not all vertebrates have both sweet and umami
receptors, however, and the lack of certain taste receptors is related
to diet. T1R2 (codes for sweet taste) is lacking in cats as well as
chickens, turkeys, and zebra finches. This suggests, and behavioral
tests verify, that these animals lack the ability to taste sugars.
Chickens and zebra finches are birds that have no need to taste
things sweet, but what about hummingbirds, all of whom feed on the
sugar-rich nectar of flowers? Surely, they can taste sugar? If so, how
do they do it without the sweet taste receptor?
Baldwin et al. (2014) addressed this question. They first asked if the
lack of T1R2 (sweet taste) in chickens and zebra finches was
indicative of birds in general, and especially hummingbirds. This
appears to be the case. Their genome analysis showed that T1R2 was
lacking in all 13 bird genomes examined, including hummingbirds
and their close relative the insectivorous chimney swift. When did
T1R2 disappear? Not long ago it seems. It is present in alligators,
which suggests that T1R2 occurred within Dinosauria, but that it was
lost when birds diverged from other reptiles.
Hummingbirds do have T1R1-T1R3 receptors, which are usually
associated with the umami or savory taste but is “blind” to sugar.
Could it be that T1R1-T1R3 was retooled in hummingbirds to detect
sugars? The researchers tested this hypothesis by cloning the
receptor and transfecting it into in vitro heterologous cells where the
response of the receptors to a variety of potential ligands was
quantified. The hummingbird T1R1-T1R3 receptors were responsive
to a variety of sugar alcohols but not to some other compounds that
humans find sweet, such as aspartame (Figure 15.4). Thus, the
hummingbird T1R1-T1R3 receptor acquired a new function in the
42–72 million years since it diverged from swifts and was retooled to
detect carbohydrates, as does the T1R2-T1R3 receptor in many other
vertebrates.
Figure 15.4 (top) The dose-dependent responses of the T1R1-
T1R3 or “savory” receptor of different species in response to amino
acids (blue) and sugars (red). All species but the hummingbird show
a strong response to most of the amino acids, as is expected for the
savory receptor. Only the hummingbird’s T1R1-T1R3 receptor shows
a strong response to the sugars, which is expected for a sweet
receptor but not a savory receptor. (bottom left) A homology model
of the Venus flytrap domain of T1R3 shows the putative ligand
binding site (yellow), and mutations that confer sugar binding to the
hummingbird’s “savory” receptor. These mutations cluster in three
distinct locations (red, green, and blue). (bottom right) An
illustration of a Ruby Throated hummingbird feeding, and the
results of taste preference tests in the field. The normalized time
drinking is on the Y axis and the X axis shows the different
treatments that varied in sucrose concentrations (white, 500 mm;
gray, 1 M; black, indicated). Significant differences in time spent
feeding on the two solutions within a treatment are indicated by ***
(P ≤ 0.001; from Baldwin et al. 2014).
“Retooled” does not really tell us what happened to the receptor, only
that something happened. How was the receptor retooled? The
researchers produced chimeric receptors by combining portions of
the sequence of chicken and hummingbird T1R1-T1R3 receptors.
They concentrated on the “venus flytrap” domain of T1R1, which is
known to modulate ligand binding. When this domain was
transferred from chicken to hummingbird, the taste receptors no
longer responded to sucrose; when the transfer was reversed, the
chicken umami receptor was now transformed into a sweet receptor.
So now we know not only that the hummingbird’s umami receptor
was modified to make it a sweet receptor, but we have a good idea of
how this took place (Figure 15.4).
An important assumption underlying this entire study is that the
source of the hummingbird’s affinity for sucrose-rich nectar is in its
mouth. It might seem that this must be the case, but the bird’s
preference for sweet could be driven by post-ingestion effects rather
than taste. So, which is it, the mouth or the gut, or more precisely, do
the cellular responses of hummingbird T1R1-T1R3 predict the
behavior of the whole organism? The researchers determined the
preference of hummingbirds for several compounds: water; sucrose;
erythritol, a sucrose agonist (mimic); and aspartame, a human
sweetener that did not activate the hummingbird’s T1R1-T1R3
receptor. Results were as expected. The birds preferred sucrose to
both water and to aspartame (Figure 15.4).
This study of hummingbird sweet preferences is an ideal example of
the power of integrating levels of analysis in addressing questions in
behavioral evolution. Without the proper phylogenetic framework,
the genomic analysis, and the cellular and behavioral experiments,
this entire story could not have emerged, or at least not in so
convincing a manner.

Assessing divergence time in behavioral evolution

In addition to elucidating the relationships among extant taxa and
reconstructing ancestral states, phylogenetic analysis can estimate
the time over which these taxa evolved. The basic idea is that the rate
of mutation is fairly constant, or at least random. Comparing
nucleotide substitutions among taxa should give some idea as to the
relative time since divergence from a common ancestor. If one can
calibrate this molecular clock, by coincidence with geological events
for example, then one can estimate the time of divergence for all
extant taxa. Tobias et al. (2013) provide a compelling example of how
this information can be used to test alternative hypotheses about
species divergence.
It is evolutionary dogma that interactions among sympatric species
can promote ecological divergence in order to avoid competition; the
phenomenon is often referred to as character displacement. The
spectacular variation in the beaks of finches on the Galapagos
Islands, which allow them to exploit different feeding niches, is the
canonical example of this process. The general prediction is that
sympatric species should differ more in homologous traits than
allopatric species, and this prediction has been borne out in
numerous studies (Schluter & McPhail 1992).
It is also possible, however, that differences in traits among species
that are now sympatric might have arisen in allopatry before species
came into contact with current neighbors. To further complicate
matters, sympatric interactions could promote convergence in some
traits rather than divergence. The convergence of territorial calls, for
example, might be advantageous because it would allow the calls to
be understood by heterospecifics as well as conspecifics and thus
enhance an individual’s ability to defend its territory. Past history,
ecological and social functions of traits, and current species
distributions can all become intertwined in asking an otherwise
simple question: Do sympatric species differ more than allopatric
species? It would be helpful to have some window into the past to
understand why traits might be different or similar in groups of
currently sympatric species.
Tobias et al. (2014) analyzed three ecological and social traits in 330
species of ovenbirds (Figure 15.5). Beak dimensions are linked to
feeding, tarsus length to locomotion, and song to social interactions.
For each species they compared these traits to its closest relative with
which it is sympatric and its closest relative with which it is
allopatric. All traits were more different between the sympatric pairs
than the allopatric ones, as would be expected from the character
displacement hypothesis. But these results are confounded by the
fact that the time since divergence is older for the sympatric
comparisons than for the allopatric ones. They then repeated their
comparisons while they controlled for time since divergence and
found that there was no significant difference in beak or tarsus
measures in the sympatric–allopatric comparisons. Their
interpretation is that when currently sympatric species were
allopatric in the past, differences in traits continued to accumulate
with time since divergence from a common ancestor. When these
species became sympatric, they were already different enough to
obviate the need for character displacement in sympatry since they
were already different enough to avoid much ecological competition.
Figure 15.5 (left) (a) Several ecological (beak depth (a) and length
(b); tarsus length (c to d)) and (b) social traits (spectrograms
(frequency/time) and waveforms (amplitude/time) of typical songs
show duration (e to f), pace (e to f/number of notes) and peak
frequency (g)) of ovenbirds. Each species was compared to its closest
relative in sympatry and its closest relative in allopatry. Sympatric
closest relatives were more divergent than were allopatric closest
relatives in the following traits: (c) beak (P < 0.001), (d) tarsus (P <
0.001), (e) song (P < 0.009). The time since divergence from their
most recent common ancestor was also greater between closest
relatives in sympatry than in allopatry (f, P < 0.001). The units of
measures used are: c, PC1 scores derived from a phylogenetic
principal components analysis on three beak variables; d, mm; e,
distance between centroids derived from a phylogenetic principal
components analysis of all acoustic traits for song; f, millions of years
(Myr), calculated from mitochondrial DNA sequence divergence. The
height of the bars represents the mean level of divergence plus one
standard deviation; the sample sizes are given in c. (Photographs
(Syndactyla striata) and spectrograms (upper, Synallaxis
erythrothorax; lower, Cinclodes aricomae). g, Variation in beak
morphology and plumage across 350 lineages in 12 major clades of
ovenbirds. Species richness in each clade is represented by distance
on the circumference of the phylogram; clades are coloured to
facilitate interpretation. For a list of species names see Supplemental
Figure 1 in Tobias et al. (2014).
The pattern is different for social traits. Song was actually more
similar in sympatry than allopatry after controlling for time since
divergence, and the longer the time since divergence the more
similar the song (Figure 15.5). These results support the assumption
that although there should be selection to favor sympatric species to
be different in ecological traits, there should be selection for them to
be similar in some social traits, such as those involved in territorial
defense. This study also highlights the importance of considering
past history when asking why species are different.

Testing Process with Pattern

The section above discusses how phylogenetic approaches can be

used to address several fundamental questions about behavioral
evolution: What is the historical pattern by which behavioral traits
arose? What is the ancestral condition of a suite of homologous
behaviors? How did ancestral traits function relative to extant
homologous traits? And, what are the details of innovation in the
genetics underlying behavior? To what degree do ecological
conditions predict variation in behavior? The final question begins to
bridge the gap between phylogenetics and studies of adaptation.

Correlations of variables with independent contrasts

We usually think of an adaptation as a phenotypic trait that evolved
for some function that enhances the individual’s Darwinian fitness
(Williams 1966). The elongated tongues of the hawkmoth and bat are
good examples—the tongue enhances the ability to reach rich food
sources that, we assume, influences the individual’s survival and
reproductive success. This fit of the animal to its environment is a
key prediction tested in studies of adaptation. The environment to
which it adapts can be an abiotic (climate) or biotic (a flower, a
parasite, a host) factor, or even the environment within itself or other
conspecifics (communication signals and receivers). Such hypotheses
of adaptation are best tested by comparing a number of closely
related species in which the independent (the agent of selection) and
the dependent variables (the target of selection) vary among taxa. A
critical caveat is the unit of comparison. It is not always the species.
Instead, it is the lineage that independently evolved the trait under
study. The unit of comparison for hair, for example, would be the
lineage of all mammals, while for language the lineage of comparison
would be Homo sapiens or perhaps the hominid lineage (Hauser et
al. 2014).
We can illustrate the dilemma at hand by considering the prediction
that males should have larger testes in more promiscuous mating
systems—a prediction from sperm competition theory. For example,
assume that we have data on the testis size and mating system for a
group of primates. The variation in the mating system is quantified
as the number of males per group. The more males in a group, the
greater the possibility of sperm competition due to females mating
with multiple males. Assume that when we examine these two
variables there is a strong correlation—males of species with more
promiscuous mating systems have larger testes. These data would
seem to support the sperm competition hypothesis. But such an
interpretation depends not on the number of species that have the
predicted relationship between testis size and mating system, but the
number of times this relationship evolved.
Assume that the mating system of the common ancestor was
monogamous (one male per group) and the males had small testes (a
general approach is outlined in Figure 15.6). This common ancestor
then gave rise to a large number of descendent species, and in each
species independently there was an independent increase in both
testis size and the number of males per group. This pattern would
support the sperm competition hypothesis because the two variables
have changed in concert a large number of times. Alternatively,
assume that the same common ancestor, monogamous with small
testes, gave rise to two descendents, one with testes and a mating
system like itself, and the other with larger testes and a more
promiscuous mating system. Each of these two species then gave rise
to many descendent species but there was no further evolution of
testes size or mating system. Although in this scenario there are
many species with the predicted relationship between testes size and
mating system, there is only one independent evolutionary event, the
initial evolution of larger testes and more promiscuous mating
system. All of the other species exhibit their values of these two
variables because they inherited them from a common ancestor
rather than by independent evolution.

Figure 15.6 (A) The relationship between two arbitrary variables,

X and Y, for each of the extant taxa, A–E, shows a strong correlation
between these two traits, which would support the hypothesis of an
evolutionary interaction between them (e.g., testes size and mating
system). The table on the right lists measures of two traits for each
taxon. (B) The hypothetical phylogenetic relationships between the
six extant taxa A–E. (C) The arrows denote where on the
phylogenetic tree independent contrasts are made for X and Y by
comparing the differences in each trait between closest relatives of
extant taxa (comparisons 1–3) and for estimates of X and Y in
hypothetical ancestors (comparisons 4–5). The table on the lower
right shows the values of X and Y for each of those independent
contrasts. Estimates of X and Y for any hypothetical ancestor was
made to minimize the total amount of evolutionary change relative to
the two descendent taxa. (D) Plotting the independent contrasts of X
and Y, which corrects for phylogenetic signal, shows that there is
only a weak correlation between X and Y, which rejects the
hypothesis for an evolutionary interaction between them.
How can we estimate the degree to which associations between traits
within species have evolved independently? A popular method is
called independent contrasts (Felsenstein 1985, Figure 15.6).
Again, one must begin with a hypothesis of phylogenetic
relationships. Here one is not interested in the actual values of the
variables of interest for each species, but in how much these
variables have changed between species since they have diverged
from a common ancestor. Thus, the first step is to estimate the
variables of interest for the ancestors or the nodes on the
phylogenetic tree. The second step is to determine the degree to
which the two variables differ between sister taxa, be they on the tips
or the nodes of the tree. It is these differences or independent
contrasts that are tested for the predicted relationships. As Figure
15.6 indicates, the relationships between the variables of the extant
species and their contrasts can be quite different.
There are now a number of more recent techniques to control for
phylogenetic effects in testing hypotheses of adaptation. The most
important step, however, is to realize that phylogenetic relatedness
can be a confounding factor and then to deal with it in an
appropriate manner.
There is somewhat of a consensus among behavioral biologists that
the comparative approach needs to control for history. Many studies
that predated this consensus have been reevaluated. For example,
the data first presented by Harcourt et al. (1981) showing the
predicted relationship in primates between testis size and mating
system that we discussed above did not control for phylogenetic
relationship. When this hypothesis was tested with a larger
taxonomic range of species with appropriate phylogenetic controls
the data still supported the sperm competition hypothesis (Harvey &
May 1989). Alternatively, Hamilton and Zuk (1982) predicted that
species of birds that were more exposed to parasites would have
brighter plumage, and a comparison of species without considering
phylogeny supported this hypothesis. This support, however, was
lessened when the phylogenetic corrections were included in the
analyses (Read & Harvey 1989).

Coevolution
When selection favors the evolution of traits needed to face
environmental challenges, such as survival in harsh temperatures,
there could be the evolution of an optimal solution that would be
stable over time. For example, animals in Arctic climes might evolve
an optimal degree of fat and fur given the various tradeoffs of such
an adaptation. The fact that the animal evolves an adaption to be
warmer does not feedback on the environmental temperature. In this
example there is one independent variable (the agent of selection,
environmental temperature) and one dependent variable (the target
of selection, the animal’s phenotype).
There are situations in which an individual can be both an agent and
a target of selection. This occurs when one individual influences the
evolution of a second and this then causes feedback on the evolution
of the first individual, a process referred to as coevolution.
Coevolution can occur between species, such as when a predator
evolves more efficient hunting tactics and a prey evolves a greater
ability to evade those tactics, or when bacteria evolve drug resistance
to the development of new antibiotics by pharmaceutical companies
(Bull & Wichman 2001). These situations can also occur within
species, such as when male courtship traits drive preferences for
those traits and vice versa (Kirkpatrick & Ryan 1991). We often refer
to these cycles as arms races (Dawkins & Krebs 1978).
A particular type of coevolution that has seen widespread interest is
that of “cospeciation.” A well-considered example is provided by
studies of figs and the wasps that pollinate them. In general, each
species of fig is pollinated by a single species of wasp, and each fig-
pollinating wasp will pollinate only that species of fig. This is one of
the most extreme cases of obligate pollination known. A female wasp
will enter a fig, which has its flowers enclosed within the fruit. The
female will pollinate some of the flowers, lay eggs in others, and then
die inside the fig. After her offspring hatch, the male and female
offspring mate, the females gather pollen from some of the flowers,
and leave to find another fig in which they can oviposit, die, and thus
continue the life cycle (Machado et al. 2001). Previous studies have
argued for strong patterns of cospeciation (e.g., Machado et al.
2001), but a more recent analysis suggests that groups of genetically
well-defined wasp species coevolve along with frequently hybridizing
groups of figs (Machado et al. 2005).
Sexual selection and sensory exploitation
Coevolution can also occur within a species, especially in sexual
communication systems.
In many sexually reproducing species, males produce advertisement
signals that are specific to the species, and females are attracted
preferentially to males producing the conspecific signal in contrast to
males that would produce signals of other species. The evolution of
such mate recognition systems is a critical part of the speciation
process. In one simple scenario of how speciation comes about,
imagine that the range of an ancestral species becomes split by a
geographic barrier resulting in two isolated populations.
Reproductive interactions are constrained to individuals on either
side of the barrier. These populations become different in various
aspects of their phenotype, including the mate recognition system,
due to random genetic drift or local adaptation. Eventually the
populations differ to a degree that they no longer recognize their
former conspecifics as appropriate mates. Speciation has occurred
(Mayr 1942).
During the process of speciation there is often evolution of a new
communication system that recognizes mates. For this to happen, it
is thought there must be a change not only in the signal used by
males but also in the females’ perception of that signal. Many studies
have shown that various aspects of the receivers’ neural systems are
tuned or biased to properties of the species-specific signal, whether it
be in the auditory, visual, chemosensory or electrical-reception
modality (Ryan & Wilczynski 2011).
Sexual selection is responsible for the evolution of exaggerated male,
and sometimes female, traits that enhance an individual’s ability to
acquire mates even if the exaggerated traits reduce survivorship.
Sometimes the exaggerated traits give the bearer armaments that are
used in combat, but in many cases the elaboration involves
ornaments that males use to attract females. Although sexual
selection can be important in driving evolution of traits used in
species recognition (Lande 1982; West Eberhard 1983), much of the
interest in sexual selection is in trying to explain the evolution of
exaggerated traits within a lineage.
A central focus in the study of sexual selection is understanding why
females would prefer males with traits that reduce survivorship,
especially in mating systems in which males offer no resources to
females but their sperm (see also Chapter 12). Two hypotheses have
received most of the attention: the good genes theory and
Fisher’s theory of runaway sexual selection. Both hypotheses
posit that the variation in the genes underlying the male trait and the
female preference become correlated, and that evolution of the male
trait in response to female preference generates correlated evolution
of the preference itself. Thus, tight coevolution of the trait and
preference should be apparent. A third hypothesis is sensory
exploitation. This hypothesis states that females will have general
sensory or perceptual biases, as detailed by the more general theory
of sensory drive, and that males who evolve traits that exploit these
biases will be favored by sexual selection.
The two hypotheses of coevolution can be distinguished from that of
sensory exploitation if one can reconstruct the evolution of sexual
selected traits and female preferences for those traits. If the trait is
restricted to one lineage but the preference for that trait
encompasses not only the lineage with the trait but others without it,
then the most logical interpretation is that the preference existed
prior to the trait. If the preference, however, is restricted to the
lineage in which the trait is present, then the coevolution hypotheses
are more tenable (Ryan 1990, 1998).
This approach to sexual selection was initially taken in two groups of
animals, swordtail fishes (Xiphophorus helleri, Basolo 1990) and
túngara frogs (Physalaemus pustulosus, Ryan et al. 1990). Female
swordtails prefer males with longer swords, an elaboration of the
bottom rays of the caudal fin. At the time of the experiments,
swordtails were thought to be a monophyletic group consisting of
two smaller groups, northern swordtails and southern swordtails. A
third group, the platyfish, is a monophyletic group belonging to the
same genus. Male platys lack swords, but Basolo showed that if she
appended a plastic sword to a male platyfish his own females found
him more attractive. Experiments with another fish, Priapella
olmacae, added further support to this hypothesis. This fish is in a
genus closely related to Xiphophorus but which, like platyfish, has
swordless males. And like platyfish, they have females that prefer
swords (Basolo 1995).
A similar result was found in a very different mating system that
relies on acoustic rather than visual cues. Male túngara frogs
produce a call consisting of a whine and from 0–6 chucks. Females
prefer males with chucks and the larger males that make lower
frequency chucks. Except for its sister species P. petersi, all other
Physalaemus species (> 30 known species) lack chucks, so the chuck
seems to have been derived from the ancestor of these two species.
But females of a closely related species, P. coloradorum, prefer the
whines of their own species when a three-chuck component of the
túngara frog is appended to the normal chuckless call more so than
their own normal, chuckless whines. But the preference for chucks is
not apparent in response to a whine with one chuck (Ron 2008).
Thus, it appears that among some Physalaemus there is a
preexisting bias for chucks and male túngara frogs evolved chucks to
exploit that bias (Ryan & Rand 1993). There is also a more subtle
exploitation going on. The relationship between the tuning of one of
the frog’s two inner ear organs and the frequencies in the chuck
results in female túngara frogs preferring the lower-frequency chucks
of larger males. Other species of Physalaemus in the same species
group, however, all have similar tuning properties (Ryan et al. 1990;
Wilczynski et al. 2001). Thus, it seems that the properties of the
chuck evolved to match what is a very conservative feature of this
animal’s neurobiology. Again, this is an example in which the
importance of phylogenetic inference in providing a deep
understanding of behavioral evolution should be clear. But it also
highlights the integrative nature of animal behavior in which
proximate and historical analyses are cross-illuminating (Ryan &
Wilczynski 2011).

Brain, Behavior, and Evolution

Not only does the brain modulate the expression of behavior, it can
also drive its evolution. Furthermore, the brain evolves under a
variety of constraints and under selection in a variety of domains.
Sensory, perceptual, and cognitive biases in the brain should not be
expected to be optimal for all tasks in which it is involved (Ryan
2011). For example, the wavelengths of light to which an animal is
most sensitive is influenced to a great degree by the sensitivity of its
photopigments. Light sensitivity influences performance in many
tasks but optimal performance in each task might require different
wavelength sensitivity. It is not possible to know a priori if
photopigment sensitivities are optimized for one of several tasks or a
compromise among competing demands. Cummings (2007) showed,
for example, that photopigment sensitivity in surfperch is optimal for
discriminating foraging targets in the ambient light environment,
and this sensitivity, in turn, has driven the evolution of male
courtship colors. Numerous other examples of similar phenomena
are reviewed in Ryan and Cummings (2013).
It has long been known that the neural and cognitive mechanisms by
which signals are processed by a receiver can lead to the evolution of
elaborate or exaggerated signals without concomitant change in the
receiver. As noted above, this can happen to some extent in sexual
selection by sensory exploitation. This is best illustrated by
Tinbergen’s notion of the supernormal stimulus, which occurs when
a stimulus has some properties exaggerated relative to the normal
stimulus and because of this elicits a greater response from the
receiver (see also Chapter 2). Two examples from the early
ethological literature include male sticklebacks and their bright red
nuptial coloration rushing toward a large red postal van driving past
their aquarium (Tinbergen 1952), and oystercatchers preferentially
retrieving a large model of an egg back to their nests in preference to
the smaller, real egg that has been removed (Tinbergen 1951). There
are other types of stimulus–response patterns that suggest that
internal biases of animals can drive evolution in certain directions.
In a more artificial setting, pigeons can exhibit a well-known
psychological phenomenon called peak shift displacement
(Hanson 1959; Hogan et al. 1975; Staddon 1975). In these situations,
an animal is trained to a positive stimulus and a negative stimulus,
such as wavelength. Peak shift displacement occurs when in
subsequent trials the subject shows a greater positive response to a
novel stimulus that is less similar to the original negative stimulus
than one that is more similar to the positive stimulus. Analogous
situations occur in nature, as in some birds in which individuals
prefer to mate with those conspecifics who appear most different
from heterospecifics in order to avoid the costs of mating with
heterospecifics (Grant & Grant 2010). In another example, zebra
finches learn to discriminate sexes from experience with the different
beak colors of their mother and father, and later when mature males
prefer females as mates that are more extreme in their beak color
than their own mothers (Ten Cate et al. 2006). Ten Cate and Rowe
(2007) review many other natural tasks in which animal responses
involve peak shift displacement.
There are also biases in how animals perceive stimuli in relation to
that of others, such as in comparisons of quantity. For example,
Weber’s Law predicts that stimuli are compared based on
proportional rather than absolute differences, such that: ∆I/I = K,
where ∆I is the minimum difference required to discriminate two
stimuli of magnitude I, and K is a constant (Stevens 1975). Weber’s
Law could impose important limitations on the evolution of signal
elaboration as an ever-increasing magnitude of traits must evolve for
them to be perceived as different from the norm. This seems to be
true in preferences for call complexity in túngara frogs. As noted
above, females prefer males who produce chucks and the more
chucks the better. But female preference does not increase linearly
with chuck number. The differences in the strength of preference for
two chucks versus one, for example, is much stronger than the
preference for six chucks over five chucks. This pattern of preference
strength scales with chuck number as predicted by Weber’s Law
(Akre et al. 2011). Furthermore, this mode of preference does not
appear to be an adaptation for mate choice per se in túngara frogs.
Call complexity is under counterselection by predation by the frog-
eating bat, Trachops cirrhosus. The bats are able to locate calling
frogs passively by listening to the frog’s call, and the bats like the
female frogs also prefer calls with more chucks to fewer chucks. The
pattern of the bat’s preference for chuck number also follows a
Weber-like pattern that is indistinguishable from that of the female
frogs. Weber’s law has not been investigated extensively in
naturalistic settings, but Akre & Johansen (2014; Figure 15.7) argue
how it could play an important role in other domains besides sexual
selection, such as species recognition and host–parasite evolution in
cowbirds.
Figure 15.7 (top) Weber’s Law predicts that comparisons are
based on proportional rather than absolute differences. When this is
the case, the magnitude of the difference between two stimuli needed
for them to be perceived as different (just noticeable difference,
JND) is greater when the overall stimulus magnitude is greater, and
the JND is smaller when the overall magnitude of the stimuli are
smaller. (bottom) Weber’s Law could have important effects on the
evolution of sexually selected traits. In this hypothetical example,
female peahens prefer male peacocks with longer tails. But the same
increase in tail length that made a male more attractive when tails
were short does not endow a male with an attractiveness advantage
when tails are long. If genetic variation for tail length is exhausted or
the costs of increasing tail length are not sustainable, selection might
then favor other innovations, such as eye spots, that increase the
male’s attractiveness in another dimension (from Akre & Johnsen
2014).
Another biased transformation of nature into our perceptions occurs
with perceptual illusions. Kelley and Kelley (2014) make a strong
case for the importance of visual illusions driving the evolution of a
variety of animal phenotypes. The most convincing example deals
with how male bowerbirds arrange decorations around their bower.
Using stones, snails and a variety of other objects, males create an
avenue leading up to the bower, the male’s lair where he courts a
willing female. The objects are not arranged randomly, but the larger
objects are at the front of the avenue and the smaller ones closer to
the bower. The males are very particular about this; when
researchers rearrange the objects the males put them back where
they belong. This particular arrangement creates what is called
forced perspective. Any object at the far end of the bower will be
perceived as relatively larger than it is due to the smaller size of the
decorations at the avenue’s end near the bower.
Humans readily exploit forced perspective to manipulate what we
think we see. Hobbits and dwarves in a movie seem to be standing
with other characters but are actually some distance away to make
them appear smaller. The Cinderella Castle at Disney’s Magic
Kingdom looks larger because the proportion of structures get
smaller as the Castle becomes taller, thus they appear farther away
than they really are, the building taller than it really is. It seems
bowerbirds have cashed in on this same trick.

SUMMARY AND CONCLUSIONS

Tinbergen considered understanding the past evolutionary
history of behavior as one of the four major aims of ethology. We
have seen here that in some cases sufficient knowledge of the
animal’s natural history can suggest how behavioral adaptations
have come about. Arguments based only on strong logical
inference, however, especially when making a priori assumptions
about how behaviors should evolve (e.g., from simple to
complex), might be more prone to error than arguments that are
framed in the context of the animal’s phylogenetic relationships.
Furthermore, it is becoming clear that knowledge of the
mechanisms underlying behavior, what Tinbergen called
causation, is also critical since such mechanisms can bias the
direction of evolution. The brain is an important landscape over
which many social traits evolve, and understanding the sensory,
perceptual, and cognitive biases of the brain are critical in
understanding how it influences the tempo and mode of
behavioral evolution. Although we have concentrated on the
evolution of behavior, a deep understanding of behavior must
involve all of Tinbergen’s four questions.

FURTHER READING
Lorenz (1958 (1967)) provides some instructive examples of how the
early ethologist explored patterns of evolution of homologous
behavior, while Greene (1994) offers a more recent as well as
insightful synopsis of the general issue of establishing behavioral
homology. Felsenstein’s (1985) independent contrast method was
critical for promoting the use of phylogenetic data to test hypotheses
of adaptation, while Thornton et al. (2003) reviews the use of
phylogenetic information to reconstruct ancestral characters of
hormone receptors and test the functionality of the ancestral
receptor. Ryan and Cummings (2013) review evidence for the role of
perceptual biases in driving signal evolution, and the textbook by
Ryan and Wilczynski (2011) emphasizes the importance of
integrative analyses in studies of animal behavior.

ACKNOWLEDGMENTS
I thank Dave Hillis for discussion and Sofia Rodriguez and Alex
Jordan for comments on the manuscript.

REFERENCES
Akre, K.L., Farris, H.E., Lea, A.M., Page, R.A. & Ryan, M.J. 2011.
Signal perception in frogs and bats and the evolution of mating
signals. Science, 333, 752–3.
Akre, K.L. & Johnsen, S. 2014. Psychophysics and the evolution of
behavior. Trends in Ecology and Evolution, 29, 291–300.
Archibald, J.K., Mort, M.E. & Crawford, D.J. 2003. Bayesian
inference of phylogeny: a non-technical primer. Taxon, 52, 87–
191.
Atz, J.W. 1970. The application of the idea of homology to behavior.
In: L.R. Aronson, E. Tobach, D.S. Lehrman & J.S. Rosenblatt
(eds.), Development and Evolution of Behavior, pp. 53–74. San
Francisco: Freeman.
Autumn, K., Ryan, M.J. & Wake, D.B. 2002. Integrating historical
and organismal biology enhances the study of adaptation.
Quarterly Review of Biology, 77, 383–408.
Baldwin, M.W., Toda, Y., Nakagita, T., et al. 2014. Evolution of sweet
taste perception in hummingbirds by transformation of the
ancestral umami receptor. Science, 345, 929–33.
Basolo, A.L. 1990. Female preference predates the evolution of the
sword in swordtail fish. Science, 250, 808–10.
Basolo, A.L. 1995. A further examination of a pre-existing bias
favouring a sword in the genus Xiphophorus. Animal Behaviour,
50, 365–75.
Blair, W.F. 1962. Non-morphological data in anuran classification.
Systematic Zoology, 11, 72–84.
Brennan, S.E. 1985. The caricature generator. Leonardo, 18, 170–8.
Bull, J.J. & Wichman, H.A. 2001. Applied evolution. Annual Review
of Ecology and Systematics, 32, 183–217.
Cooper, I.A., Brown, J.M. & Getty, T. 2016. A role for ecology in the
evolution of colour variation and sexual dimorphism in Hawaiian
damselflies. Journal of Evolutionary Biology, 29, 418–27.
Cumming, J.M. 1994. Sexual selection and the evolution of dance fly
mating systems (Diptera: Empdidae: Empidinae). Canadian
Entomologist, 126, 907–20.
Cummings, M.E. 2007. Sensory trade‐offs predict signal divergence
in surfperch. Evolution, 61, 530–45.
Darwin, C. 1862. On the Various Contrivances by which Orchids are
Fertilized by Insects. London: John Murray.
Darwin, C. 1872. The Expression of the Emotions in Man and the
Animals. London: John Murray.
Dawkins, R. 1996. Climbing Mount Improbable. New York: Norton.
Dawkins, R. & Krebs, J.R. 1978. Animal signals: Information or
manipulation? In: J.R. Krebs & N.B. Davies (eds.), Behavioural
Ecology, An Evolutionary Approach, pp. 282–309. Oxford:
Blackwell.
Day, J.J., Upchurch, P., Norman, D.B., Gale, A.S. & Powell, H.P.
2002. Sauropod trackways, evolution, and behavior. Science, 296,
1659.
Dean, A.M. & Thornton, J.W. 2007. Mechanistic approaches to the
study of evolution: the functional synthesis. Nature Reviews
Genetics, 8, 675–88.
Drickamer, L.C. 2010. Animal behavior: The seventeenth to the
twentieth centuries. In: M.D. Breed & J. Moore (eds.),
Encyclopedia of Animal Behavior, volume 3, pp. 453–461.
Oxford: Academic Press.
Enquist, M. & Arak, A. 1993. Selection of exaggerated male traits by
female aesthetic senses. Nature, 361, 446–8.
Enquist, M. & Arak, A. 1998. Neural representation and the evolution
of signal form. In: R. Dukas (ed.), Cognitive Ecology, pp. 21–87.
Chicago: University of Chicago Press.
Felsenstein, J. 1985. Phylogenies and the comparative method.
American Naturalist, 125, 1–15.
Fitch, W.M. 1971. Toward defining the course of evolution: minimum
change for a specific tree topology. Systematic Biology, 20, 406–
416.
Grant, B.R. & Grant, P.R. 2010. Songs of Darwin’s finches diverge
when a new species enters the community. Proceedings of the
National Academy of Sciences USA, 107, 20156–63.
Greene, H.W. 1994. Homology and behavioral repertoires. In: B.K.
Hall (ed.), Homology: The Hierarchical Basis of Comparative
Biology, pp. 369–392. San Diego: Academic Press.
Hamilton, W.D. & Zuk, M. 1982. Heritable true fitness and bright
birds: a role for parasites? Science, 218, 384–7.
Hanson, H.M. 1959. Effects of discrimination training on stimulus
generalization. Journal of Experimental Psychology, 58, 321–34.
Harcourt, A.H., Harvey, P.H., Larson, S.G. & Short, R.V. 1981. Testis
weight, body weight and breeding system in primates. Nature,
293, 55–7.
Harvey, P.H. & May, R.M. 1989. Out for the sperm count. Nature,
337, 508–9.
Hauser, M.D., Yang, C., Berwick, R., et al. 2014. The mystery of
language evolution. Frontiers in Psychology, 5, Article 401.
Hogan, J.A., Kruijt, J.P. & Frijlink, J.A. 1975. Supernormality in a
learning situation. Zeitschrift für Tierpsychologie, 38, 212–8.
Horner, J.R. & Makela, R. 1979. Nest of juveniles provides evidence
of family structure among dinosaurs. Nature, 282, 296–8.
Houck, M., Gauthier, J.A. & Strauss, R.E. 1990. Allometric scaling in
the earliest. Archaeopteryx lithographica. Science, 247, 195–8.
Huxley, J. 1914. The courtship habits of the great crested grebe.
Porpodiceps cristatus. Proceeding of the Zoological Society,
London, 2, 491–562.
Johnson, K.P., McKinney, F., Wilson, R. & Sorenson, M.D. 2000.
The evolution of postcopulatory displays in dabbling ducks
(Anatini): a phylogenetic perspective. Animal Behaviour, 59,
953–63.
Kelley, L.A. & Kelley, J.L. 2014. Animal visual illusion and confusion:
the importance of a perceptual perspective. Behavioral Ecology,
25, 450–63.
Kessel, E.L. 1955. The mating activities of balloon flies. Systematic
Zoology, 4, 97–104.
Kirkpatrick, M. & Ryan, M.J. 1991. The evolution of mating
preferences and the paradox of the lek. Nature, 350, 33–8.
Lande, R. 1982. Rapid origin of sexual isolation and character
divergence in a cline. Evolution, 36, 213–23.
LeBas, N.R. & Hockham, L.R. 2005. An invasion of cheats: the
evolution of worthless nuptial gifts. Current Biology, 15, 64–7.
Lloyd, J.E. 1984. On deception, a way of all flesh, and firefly
signalling and systematics. Oxford Surveys in Evolutionary
Biology, 1, 48–84.
Lorenz, K. 1941 (1971). Comparative studies of the motor patterns of
Anatinae. In: R. Martin (ed.), Studies on Animal and Human
Behavior, vol. 2, pp. 14–114. London: Methuen.
Lorenz, K. 1958 (1967). The evolution of behavior. In: J.L. McGaugh,
N.M. Weinberger & R.E. Whalen (eds.), Psychobiology, The
Biological Bases of Behavior, Readings from Scientific American,
pp. 33–42. San Francisco: Freeman.
Machado, C., Robbins, N., Gilbert, T.P. & Herre, E.A. 2005. Critical
review of host specificity and its coevolutionary implications in
the fig/fig-wasp mutualism. Proceedings of the National
Academy of Sciences USA, 102, 6558–65.
Machado, C.A., Jousselin, E., Kjellberg, F., Comptom, S.G. & Herre,
E.A. 2001. Phylogenetic relationships, historical biogeography
and character evolution of fig-pollinating wasps. Proceedings of
the Royal Society, London B, 268, 685–94.
Martins, E. (ed.) 1996. Phylogenies and the Comparative Method in
Animal Behavior. Oxford: Oxford University Press.
Max, D.T. 2008. A man of taste. The New Yorker, May 12.
Mayr, E. 1942. Systematics and the Origin of Species. New York:
Columbia University Press.
Muchhala, N. 2006. Nectar bat stows huge tongue in its rib cage.
Nature, 444, 701–2.
Nicholson, K.E., Harmon, L.J. & Losos, J.B. 2007. Evolution of
Anolis lizard dewlap diversity. PLoS One, 2, e274.
Odom, K.J., Hall, M.L., Riebel, K., Omland, K.E. & Langmore, N.E.
2014. Female song is widespread and ancestral in songbirds.
Nature Communications, 5, Article 3379.
Oliveira, R.F., Taborsky, M. & Brockmann, H.J. (eds.) 2008.
Alternative Reproductive Tactics: An Integrative Approach.
Cambridge, UK: Cambridge University Press.
Padian, K. 2001. Cross-testing adaptive hypotheses: phylogenetic
analysis and the origin of bird flight. American Zoologist, 41,
598–607.
Pagel, M. 1999. The maximum likelihood approach to reconstructing
ancestral character states of discrete characters on phylogenies.
Systematic Biology, 48, 612–22.
Paradis, E. 2011. Analysis of Phylogenetics and Evolution with R.
New York: Springer.
Phelps, S.M. & Ryan, M.J. 2000. History influences signal
recognition: Neural network models of túngara frogs. Proceedings
of the Royal Society, London, B, 267, 1633–9.
Platt, J.R. 1964. Strong inference. Science, 46, 347–53.
Price, T. 1997. Correlated evolution and independent contrasts.
Philosophical Transactions of the Royal Society, B, 352, 519–29.
Read, A.F. & Harvey, P.H. 1989. Reassessment of comparative
evidence for Hamilton and Zuk theory on the evolution of
secondary sexual characters. Nature, 339, 618–20.
Ron, S.R. 2008. The evolution of female mate choice for complex
calls in túngara frogs. Animal Behaviour, 76, 1783–94.
Ryan, M.J. 1990. Sensory systems, sexual selection, and sensory
exploitation. Oxford Surveys in Evolutionary Biology, 7, 157–95.
Ryan, M.J. 1996. Phylogenetics and behavior: some cautions and
expectations. In: E. Martins (ed.), Phylogenies and the
Comparative Method in Animal Behavior, pp. 1–21. Oxford:
Oxford University Press.
Ryan, M.J. 1998. Receiver biases, sexual selection and the evolution
of sex differences. Science, 281, 1999–2003.
Ryan, M.J. 2011. The brain as a source of selection on the social
niche: Examples from the psychophysics of mate choice in
túngara frogs. Integrative and Comparative Biology.
doi:10.1093/icb/icr065.
Ryan, M.J. & Cummings, M.E. 2013. Perceptual biases and mate
choice. Annual Review of Ecology, Evolution and Systematics,
44, 437–59.
Ryan, M.J., Fox, J.H., Wilczynski, W. & Rand, A.S. 1990. Sexual
selection for sensory exploitation in the frog. Physalaemus
pustulosus. Nature, 343, 66–7.
Ryan, M.J. & Rand, A.S. 1993. Sexual selection and signal evolution:
the ghost of biases past. Philosophical Transactions of the Royal
Society, B, 340, 187–95.
Ryan, M.J. & Rand, A.S. 1995. Female responses to ancestral
advertisement calls in the túngara frog. Science, 269, 390–2.
Ryan, M.J. & Rand, A.S. 1999. Phylogenetic influences on mating call
preferences in female túngara frogs (Physalaemus pustulosus).
Animal Behaviour, 57, 945–56.
Ryan, M.J. & Wilczynski, W. 2011. An Introduction to Animal
Behavior, An Integrative Approach. Cold Spring Harbor, New
York: Cold Spring Harbor Laboratory Press.
Sakaluk, S.K. 2000. Sensory exploitation as an evolutionary origin to
nuptial food gifts in insects. Proceedings of the Royal Society of
London, B, 267, 339–43.
Schluter, D. & McPhail, J.D. 1992. Ecological character displacement
and speciation in sticklebacks. American Naturalist, 140, 85–108.
Staddon, J.E.R. 1975. A note on the evolutionary significance of
‘supernormal’ stimuli. The American Naturalist, 109, 541–5.
Stevens, S.S. 1975. Psychophysics. Transaction Publishers.
Ten Cate, C. & Rowe, C. 2007. Biases in signal evolution: learning
makes a difference. Trends in Ecology and Evolution, 22, 380–7.
Ten Cate, C., Verzijden, M.N. & Etman, E. 2006. Sexual imprinting
can induce sexual preferences for exaggerated parental traits.
Current Biology, 16, 1128–32.
Thornton, J.W. 2004. Resurrecting ancient genes: experimental
analysis of extinct molecules. Nature Reviews Genetics, 5, 366–
75.
Thornton, J.W., Need, E. & Crews, D. 2003. Resurrecting the
ancestral steroid receptor: ancient origin of estrogen signaling.
Science, 301, 1714–7.
Tinbergen, N. 1951. The Study of Instinct. Oxford: Oxford University
Press.
Tinbergen, N. 1952. The curious behavior of the stickleback.
Scientific American, 187, 2–6.
Tinbergen, N. 1963. On aims and methods in ethology. Zeitschrift für
Tierpsychologie, 20, 410–33.
Tobias, J.A., Cornwallis, C.K., Derryberry, E.P., et al. 2014. Species
coexistence and the dynamics of phenotypic evolution in adaptive
radiation. Nature, 506, 359–63.
Weishampel, D.B. 1981. Acoustic analyses of potential vocalization in
lambeosaurine dinosaurs (Reptilia: Ornithischia). Paleobiology,
7, 252–61.
Wenzel, J.W. 1992. Behavioral homology and phylogeny. Annual
Review of Ecology and Systematics, 7, 361–81.
West Eberhard, M.J. 1983. Sexual selection, social competition and
speciation. Quarterly Review of Biology, 58, 155–83.
Wilczysnki, W., Rand, A.S. & Ryan, M.J. 2001. Evolution of calls and
auditory tuning in the Physalaemus pustulosus species group.
Brain, Behavior and Evolution, 58, 137–51.
Wiley, E.O. & Lieberman, B.S. 2011. Phylogenetics: Theory and
Practice of Phylogenetic Systematics. New York: Wiley.
Williams, G.C. 1966. Adaptation and Natural Selection: A Critique
of Some Current Evolutionary Thoughts. Princeton, NJ:
Princeton University Press.
Wimberger, P.H. & De Queiroz, A. 1996. Comparing behavioral and
morphological characters as indicators of phylogeny. In: E.
Martins (ed.), Phylogenies and the Comparative Method in
Animal Behavior, pp. 206–33. Oxford: Oxford University Press.
16
the evolution of hominin behavior
IAN TATTERSALL

INTRODUCTION
Human beings are very unusual creatures. Not only is their
upright bipedality a very unusual form of locomotion, with
enormous structural ramifications throughout the body, but they
process information about their exterior and interior
environments in an utterly unique way. Among all of the many
evolutionary alterations that occurred along the way to Homo
sapiens it is this latter acquisition that is the most astonishing,
and that was least predicted by anything that had gone before. In
this chapter, I look at what can be said about how a specialized
but otherwise unremarkable lineage of higher primates became
transformed into the altogether unprecedented cognitive entity
that our species is today. Using the modern African great apes as
a very approximate yardstick for gauging where the story begins,
the chapter surveys what can be gleaned of behavioral and
cognitive change over the seven- to eight-million-year history of
the subfamily Homininae to which our species Homo sapiens
belongs.

Primate Phylogeny

We humans are primates, members of a distinctive order of

mammals with roots stretching back to the earliest part of the
Cenozoic Era that followed the demise of the dinosaurs at around 65
million years (myr) ago. Opinions continue to differ on whether the
somewhat squirrel-like plesiadapiforms that abounded during the
Paleocene, the first epoch of the Cenozoic, should properly be
classified as primates; but by the beginning of the succeeding Eocene
epoch, a little over 50 myr ago, “lower” primates that generally
resembled the lemurs that still flourish in Madagascar today were
already established in Eurasia, Africa, and North America (see
Fleagle 2013).
The “higher” primate group to which humans belong (along with the
monkeys of the New and Old Worlds, the gibbons, and the African
and Asian great apes) emerged initially during the Oligocene epoch,
around 30 myr ago. Within this larger group, the superfamily
Hominoidea that embraces apes and humans began diversifying
importantly during the early part of the Miocene epoch that began 23
myr ago. Today’s surviving great apes, the chimpanzees, bonobos,
gorillas, and orangutans, have virtually no fossil record. But on
molecular evidence the group that contains just them and us is
believed to have begun differentiating around 14 myr ago (see
discussion in Fleagle 2013). Among the great apes, it is the
chimpanzees and bonobos that are generally believed to share the
most recent common ancestry with humans, various different lines
of evidence converging to suggest that our last common ancestor
with them lived toward the end of the Miocene, around 7 or 8 myr
ago.
Modern humans are classified in the zoological family Hominidae (or
the subfamily Homininae; for present purposes the difference is
merely notional), which also contains all extinct forms closer to us
than to the living apes. As we will see, there are good reasons for
believing that the earliest hominins may not have been organized
socially like any of the extant apes. Nonetheless, it is generally
assumed that chimpanzees and bonobos provide a pretty good
approximation to the general cognitive prowess of those ancestors,
having encephalization quotients (EQ: the ratio between actual brain
mass and that predicted for a particular body size, see Figure 16.1)
that are more or less in the same range. To put this observation in
context, as well as to show how dramatic a feature encephalization
has been in the evolution of the human lineage, by one estimate
(Roth & Dicke 2005) modern humans have EQs of around 7.5, and
chimpanzees of up to 2.5, while a typical monkey is at 2.1 and dogs
are at about 1.2.

Figure 16.1 Graphic representation of encephalization in various

mammals. Brain and body weights are plotted against each other (log
scales), showing that humans are far from alone in having higher
than expected brain mass. Modified after Roth and Dicke (2005).

Our Closest Living Relatives

Chimpanzees and bonobos are clearly highly intelligent in the greater

mammalian scheme of things (see review by Cohen 2010). They are
ingenious and often remarkably efficient problem solvers. They react
in complex and subtle ways to social and other stimuli. They form
strategic alliances, and even engage in “politics” to reinforce or
improve their social statuses. They recognize themselves in mirrors.
They spontaneously make and employ rudimentary tools, for
example using wooden “spears” to impale sleeping galagos, prepared
twigs to “fish” for termites, and anvils and hammer-stones to crack
hard nuts. They pass such technologies down through the
generations as local “cultural” traditions, albeit apparently acquired
by imitation rather than deliberately “taught.” Under appropriate
controlled conditions, apes have even been shown to understand the
combined meanings of short strings of verbal and other symbols, for
example the simple command to “take … green … ball … outside.”
And recent studies have shown that apes have theory of mind (ToM)
to the (quite complex) extent that they can recognize that other
individuals might harbor a false belief (Krupenye et al. 2016; Kano et
al. 2019).
The living apes thus show many of the basic abilities that underpin
our own modern human cognitive processes, something that it is
hardly surprising to encounter in forms that share a huge
preponderance of their evolutionary histories with us. But despite
their ability to recognize symbols, and even to use them additively,
they do not manipulate information in their minds in the same way
as us. What we modern Homo sapiens do, apparently uniquely, is to
deconstruct both our surroundings and our internal experiences into
a vocabulary of intellectual symbols that we can then rearrange,
according to rules, to make statements about the world not simply as
it is, but as it might be. As a result, we human beings are able to live
much of the time in the worlds we reconstruct in our heads, rather
than directly in the world that nature presents to us.
Such symbolic reordering of reality exceeds the ape cognitive range,
largely because apes appear unable to reduce their observations to
abstractions. As the cognitive psychologist Daniel Povinelli (2004)
put it: “in contrast to modern humans, chimpanzees rely strictly
upon observable features of others to forge their social concepts …
[they] do not realize that there is more to others than their
movements, facial expressions, and habits of behavior … they [do]
not understand that other beings are repositories of private, internal
experience … [they] do not reason about what others think, precisely
because they do not form such concepts in the first place” (Povinelli
2004, p. 33). To put this in slightly different terms, as noted by the
ToM researchers (Krupenye et al. 2016, p. 113), apes appear to show
“implicit understanding of belief,” (italics added) rather than the
explicit understanding that symbolic ordering permits.
Of course, there is no way for us to know the subjective quality of
apes’ experience of the world. But both the potentials and the
limitations of chimpanzee cognition are almost certainly relevant to
any attempt to define the starting point from which our hominid
ancestors set out on their seven-million-year trajectory. As a result,
tracing the evolution of human behavior is in significant part a
matter of trying to reconstruct how the cognitive gulf that separates
contemporary human beings from chimpanzees was bridged.
However, unlike the hard tissues of the body, behavior itself does not
preserve in the material record. This means that we have to seek
preservable proxies for any ancient behaviors we are interested in. At
the beginning of the human story those proxies are very indirect
indeed, since all we have to go on is a fossil record that is limited to
hard-tissue anatomies and what can be inferred from them. But the
behavioral archive improves dramatically once our precursors began,
at some point over about 2.5 myr ago, to leave behind not only fossils
but an archaeological record. Consisting initially of no more than the
durable stone tools our forerunners made, and of the physical
consequences of their use, this behavioral record bears witness to
stepwise increases over time in hominid behavioral sophistication
(see review in Tattersall 2012).

The Earliest Hominins

The earliest putative hominins consist of a rather miscellaneous

assortment of fossil species, all African and mostly pretty poorly
known, that date to between about 7 and 4 myr ago (Figure 16.2; see
also Gibbons 2006). What they all have most strikingly in common—
and the principal reason for including them all in Homininae—is that
they all seem to have shared the founding adaptation of walking
upright when on the ground. The times in which they lived saw an
irregular but progressive drying-out of the African climate that,
together with a net increase in seasonality, was leading to the local
replacement of forest by woodlands and bushlands. These
environmental changes made available new terrestrial opportunities
for exploitation by previously arboreal primates (see review by Van
Covering 2014).

Figure 16.2 Reconstruction by the paleoartist Jay Matternes of

the skeleton of the 4.4 myr-old putative hominid Ardipithecus
ramidus, front and side views. Note the long arms and the divergent
(grasping) first digit of the foot. Courtesy of and © Jay Matternes.
Modern apes spend a lot of time on the ground; but when there, they
travel on all fours. This is clearly the posture in which they feel most
comfortable, even though as “facultative bipeds” they are perfectly
capable of standing and moving short distances on two feet. This
distinguishes them from our own ancient precursors who adopted
bipedality when on terra firma, suggesting that these latter were so
much in the habit of holding their bodies erect in the trees that
upright posture naturally felt most comfortable when they descended
to the ground. This is a more plausible explanation for terrestrial
bipedality than the traditional scenarios suggesting that, once on the
ground, a quadrupedal ancestor stood awkwardly upright to free its
hands, to see farther, to shed heat more effectively, or whatever (see
discussion by Harcourt-Smith 2014). After all, once you have stood
upright you have all the advantages of this posture—and all of the
disadvantages, too, which importantly include slowness of
movement and a consequently increased vulnerability to predation.
One other feature shared—where known—by all the very early
hominins is the reduction of the canine teeth, and of the associated
sharpening surface on the front lower premolar (Figure 16.3). Apes
have a formidable natural weaponry in the form of a sharp, slashing
upper canine, and a compensating advantage of losing this feature
may have come in the form of an ability to chew more effectively by
moving the lower jaw from side to side (which can’t happen if large
interlocking canines restrict it to an up-and-down motion). This in
turn suggests that a descent to the ground may have opened up new
dietary possibilities to the early hominins.
Figure 16.3 Side views of the toothrows of (from top to bottom): a
modern ape (chimpanzee); a very early hominid (Ardipithecus
ramidus); and a modern Homo sapiens. The ape has a large, pointed
upper canine tooth that locks and hones against the lower premolar
behind it; the human has a barely projecting upper canine (plus an
overbite and absent wisdom teeth); and the early hominid shows an
intermediate condition in which the canine crowns are reduced but
remain somewhat projecting. Drawing by Jennifer Steffey.
That this was the case is supported by the generally large size of the
early hominids’ molar teeth, and by stable-isotope analyses
suggesting that their diets were much more generalized than those of
apes. Chimpanzees living in relatively open habitats comparable to
those frequented by early hominins tend to stick to the same
foodstuffs—mainly fruit and leaves—that also sustain them in the
forests (see Sponheimer & Lee-Thorp 2014). Hominins, on the other
hand, seem from the beginning to have consumed a much wider
range of foods that probably included both tubers and animal
proteins.
Nonetheless, the early hominins remained in many ways “bipedal
apes.” Notably, their brains and EQs seem to have remained on a par
with those of chimpanzees for several million years; and at very early
stages there is nothing to suggest that their cognitive prowess had
moved significantly beyond the ape zone. Povinelli’s speculation
seems reasonable that they were “intelligent … creatures who deftly
attend[ed] and learn[ed] about the regularities that unfold[ed] in the
world around them,” but who did “not reason about unobservable
things; they [had] no ideas about the ‘mind,’ no notion of
‘causation.’” (Povinelli 2004, p. 34).

Australopithecus afarensis

The first hominin we know well is a form known as Australopithecus

afarensis. Its fossils are well known from Ethiopia and elsewhere in
eastern Africa in the period between about 4 and 3 myr ago.
Australopithecus afarensis is a possible descendant of the earlier
and more sparsely represented—but definitely bipedal—
Australopithecus anamensis, known as far back as 4.2 myr in Kenya;
and its most famous representative is the 3.2 myr-old skeleton
“Lucy,” discovered in Ethiopia in 1974. We can consider A. afarensis
emblematic of the diverse “australopith” hominin group that
flourished in Africa between about 4 and 2 million years ago (see
Figure 16.4), until our own genus Homo came on the scene.
Figure 16.4 Highly approximate tentative genealogical tree of the
hominins, showing its densely branching nature and the consistent
co-occurrence of several different hominin lineages prior to the
emergence of behaviorally modern Homo sapiens. Drawn by Kayla
Younkin.
Australopithecus afarensis was a small-bodied but quite strongly
sexually dimorphic hominin, with females standing about 3´6” (107
cm) tall and larger males maybe a foot (30 cm) more. Its
commitment to terrestrial bipedality is clearly shown in its broad
pelvis; in its foot with a great toe in line with the others; and in its
rather short legs that nonetheless show an inward angle from hip to
knee, as ours do. As a result, some characterize A. afarensis as a
“habitual” biped. Above the waist, though, the picture is rather
different. In its shoulder girdle, arms, and hands, A. afarensis shows
numerous adaptations to climbing and suspending its weight from
above (though its hands were somewhat shorter and broader than
those of modern apes). Indeed, we seem to be looking here at a
creature that was still very much at home in the trees, even as it
moved bipedally when on the ground (see Harcourt-Smith 2014).
With this mixture of adaptations, there can be little doubt that
australopiths both foraged extensively in the trees and sought
arboreal shelter at night. Their fossils are most commonly found
along with those of animals indicating that riverine or lakeside
forests, bushlands, and woodlands were nearby; and almost certainly
they would have exploited resources available in all of these
environments, many of which teemed with large predators. And it is
unlikely that the early hominins did not adjust lifestyles that were
initially developed in less dangerous arboreal settings to match this
pressing reality. As a result, while paleoanthropologists have often
been tempted to believe that early hominins were pair-bonding like
their modern descendants, or to extrapolate their social
organizations from those of the living apes, it has recently been
suggested that a better living model is presented by generalist
primates like macaques and baboons that live in comparable
environmental settings, and that are similarly menaced by the
predators that patrol them (Hart & Sussman 2009).
In contrast to the relatively small social groups of the great apes,
macaques and baboons live in very large and spatially organized
social units that sometimes number in the hundreds. The
reproductive core of females and young offspring is protected by
dominant males, and the vulnerable periphery is made up of more
reproductively expendable members such as subordinate males and
subadults of both sexes. This key spatial arrangement is achieved by
maintaining rigid social hierarchies that are reinforced by constant
social interaction, much of which involves not only affiliative
behaviors such as grooming, but also vigorous competition among
adult males for access to sexually receptive females.
Suggesting that australopiths had lived in large, structured groups of
this kind is a far cry from earlier interpretations that saw them
essentially as junior-league humans, with pair-bonding and other
modern human social features. Indeed, this alternative scenario may
well imply that the conspicuous “prosociality,” or extreme
cooperativeness, of modern human beings actually began to develop
only after this stage in human evolution had been passed. But in
many ways this newer picture is a lot more convincing than the one it
replaces. What is more, there is—at least to begin with—no strong
evidence that the early australopiths displayed any significant
cognitive or behavioral advances relative either to the very earliest
hominids or to today’s apes—which, as we have seen, are cognitively
very sophisticated. The evidence of brain size (admittedly, only a
very approximate measure of “intelligence”) is not entirely clear. The
very early hominins seem to have had brains that, relative to body
size, were about on a par with those of today’s apes. And while
australopith brains were, on average, perhaps a quarter larger than
this, they rarely achieved a volume much in excess of about 500
milliliters (ml), not much more than a third of the size of an average
modern human brain (Holloway et al. 2004). Accordingly, their skull
proportions remained ape-like, with substantial faces
(accommodating large chewing teeth) jutting out in front of notably
small braincases. This is the very reverse of what we see in Homo
sapiens, where a tiny face is tucked beneath the front of a huge and
balloon-like braincase.
But then something happened.
 The First Stone Tool Makers

The “have your cake and eat it too” forest-to-woodland-to-bushland

strategy turned out to be a very successful way of life, supporting a
morphology that remained unchanged in its essentials for millions of
years, even as the australopiths diversified and new species came and
went. But any significant biological or behavioral innovation has
necessarily to be made inside an existing group; and the most fateful
behavioral innovation ever made among hominins was no exception.
At a place in northern Ethiopia called Dikika, researchers discovered
some fossil mammal bones bearing grooves of exactly the kind that
are made by sharp stone flakes during the butchery of a carcass
(McPherron et al. 2010). A hominin, presumably belonging to
Australopithecus afarensis whose fossils are found in the same
sediments, had apparently been using stone tools to dismember
mammal carcasses there some 3.4 myr ago. Unfortunately, no
identifiable stone tools are found in the Dikika deposits, although
some oddly large apparent artifacts of similar age have been reported
from Kenya (Harmand et al. 2015). The clear-cut archaeological
record thus begins only with the “Oldowan” stone tool industry
(named for Tanzania’s Olduvai Gorge), which appears in both Kenya
and Ethiopia at around 2.6 myr ago (see Semaw 2000). This has led
to the speculation that the earlier Dikika hominins had used
naturally occurring sharp flakes of stone, fractured in riverbeds, to
do the earliest butchery, and that only later did their descendants
figure out how to produce such cutting tools at will. Still, whatever
the exact case, there is fair presumptive evidence at Dikika that
australopiths had begun to depart from their essentially vegetarian
ancestral lifeway and to incorporate a significant component of
animal proteins and fats into their diets.
It is well known that chimpanzees at some localities regularly hunt
small mammals such as colobus monkeys and young bushpigs (see
review by Newton-Fisher 2014). But the meat they thus obtain
appears, via sharing among group members, to have a greater social
significance than a dietary one: raw meat is often excreted largely
undigested (Wrangham 2009). Something else was clearly
happening, or was about to happen, among the australopiths.
Whatever the exact case, the deliberate manufacture and
employment of stone tools following around 2.5 myr ago
demonstrates that by that time hominins had moved cognitively, as
well as in their subsistence and presumed social organization, well
beyond the envelope described by the living great apes. Experiments
aimed at teaching apes—most notably the bonobo Kanzi, a star in
“ape language” experiments—to make simple stone tools were not a
notable success (Toth et al. 1993). After intensive coaching Kanzi
managed to produce the occasional sharp flake, but he never got the
idea of striking one cobble with another at exactly the right angle and
force necessary to reliably produce a cutting implement. This
apparent failure to understand the principles of flaking rock (rather
than just the notion of breaking it) was clearly due to much more
than the limited dexterity conferred by Kanzi’s long, slender hands,
adapted to powerful grasping rather than to fine manipulation. The
cognitive equipment just wasn’t there. What is more, the early stone
tool makers showed considerable foresight and a planning capacity
that probably lay well beyond the ape range. That is because good
raw materials for stone tool making are not found everywhere, and
the refitting of flakes found at early butchery sites into entire cobbles
shows that the ancient tool makers had selected suitable stones in
advance and then carried them around for many kilometers in
anticipation of needing them later (see Plummer 2004).
By at least 2.5 myr ago, then, a creature of unprecedented cognitive
capacities was around in Africa; and it is fair to surmise that tool
making brought with it modified behaviors of many kinds, though
any attempt to specify them in detail would be pure speculation. Still,
it is probably fair to say that the new cognitive style was an
incremental—albeit huge—improvement on what had existed before,
rather than expressing a qualitative shift in cognitive style. The
techniques of flaking rock can be passed along by imitation, rather
than having to be consciously taught; and certainly, the introduction
of the new behavior, radical as it might have been, does not appear to
have heralded a new dynamic of change. For once the new behavior
had been introduced, things stayed remarkably stable for the best
part of a million years, both in terms of the material products
themselves and of whatever behaviors were associated with them.

The Striding Biped

Compared to us, the australopiths were both small-brained and

small-bodied; and even though they walked upright when on the
ground, they did not walk around exactly as we do. They were indeed
harbingers of what was to come in some respects; but they were
certainly not simply junior versions of what we are today, even long
after they had begun the fateful transition from being prey species to
being hunters. Our modern technologies have long insulated us from
the dangers posed by large predators with skills and anatomical
weaponry honed by many millions of years of evolution. But with
only very rudimentary implements at their disposal, early hominins
emerging onto the expanding woodlands and grassland of Africa in
the early Pleistocene would have been hugely vulnerable to
predation, even while they themselves were almost certainly
incorporating increasing quantities of animal protein into their diets.
This would have been particularly true for the australopiths, with
their small bodies and archaic limb proportions; but following about
1.9 myr ago these hominins were joined by a radically different kind
of relative, the success of which presumably lay in a different set of
responses to the new ecological pressures that came with life away
from the shelter of the trees.
This first hominin known to have had body proportions close to our
own is Homo ergaster, a species best known from fossils found in
northern Kenya (Figure 16.5). Various earlier fossils from eastern
Africa are often referred to as “Homo habilis,” or as “early Homo,”
but there is little to suggest that these forms had anything closely
resembling modern human body structure. Exactly when that
structure emerged is a little difficult to say. The best-preserved
Homo ergaster cranium, with a reduced face and a substantial brain
volume of 848 ml (Holloway et al. 2004), was found to the east of
Kenya’s Lake Turkana in the 1970s. It is now dated to about 1.7 myr,
but we know nothing about the structure of the individual’s body.
Much more informative was the discovery in 1984, on the other side
of Lake Turkana, of the amazingly well-preserved “Nariokotome
Boy” skeleton (see Walker & Leakey 1993, also Figure 16.6), now
dated to about 1.6 myr ago.

Figure 16.5 Diorama at the American Museum of Natural History

showing two Homo ergaster individuals in northern Kenya at about
1.7 million years ago. The practice of animal butchery was well
established by then; but note that, although they are advanced in
body build, these hominids are still using simple flake tools of the
kind that their archaic predecessors had already been using almost a
million years earlier (possibly more). Photo: Denis Finnin/AMNH.
Figure 16.6 The “Nariokotome Boy“ Homo ergaster skeleton
from West Turkana, Kenya. Approximately 1.6 million years old.
Drawn by Don McGranaghan.
Belonging to an adolescent who would have been as tall as an
average modern person at maturity (Graves et al. 2010), the
Nariokotome skeleton possesses the long legs and short arms that
are the hallmark of modern humans, despite many detail differences
from us (Walker & Leakey 1993). The skeleton is unquestionably that
of an “obligate biped,” an upright strider who would have been at
home in the open savanna, and who had lost the ancestral climbing
adaptations of the upper body. To go with this new body form, the
Boy boasted a reduced face and an impressive brain volume of 900
ml (Holloway et al. 2004). On the other hand, developmental studies
of his teeth indicate that he had matured a lot faster than humans do
today, though not as fast as a young ape does (Dean & Smith 2009).
This is significant, since the extended maturation of modern
juveniles provides a long period of dependence during which a vast
library of learned behaviors is acquired. The accelerated
developmental period of apes severely reduces this learning interval.
Significantly, the first appearance of Homo ergaster is not associated
with any new technology, presaging a consistent future disconnect in
the hominid record between anatomical and behavioral innovations.
Most very early representatives of the new hominin species are found
in sediments that yield only flakes of the early kind, the “cores” from
which they were struck, and the “hammers” that were used to strike
them and for other pounding tasks. The first intimations of a
radically new tool type come at 1.78 million years ago, when crude
examples of the teardrop-shaped “Acheulean handaxe” (Figure 16.7)
are found at a site in Kenya (Lepre et al. 2011). Interestingly,
however, this first appearance of the handaxe seems to be an outlier,
implements of this kind becoming common only significantly later,
after about 1.5 million years ago.
Figure 16.7 Two Acheulean implements from the type locality of
St. Acheul in France. On the left is a typical teardrop-shaped
“handaxe,” while the truncated variant on the right is known as a
“cleaver.” Photo by Willard Whitson.
Handaxes represent a revolutionary concept in stone tool making,
since they are relatively large implements deliberately fashioned to a
particular shape—typically, that bilaterally-symmetrical “teardrop,”
though there are variants. For the first time, stone tool makers were
producing implements of a predetermined shape that must have
been held in a “mental template” in the toolmaker’s mind before
knapping began. This approach contrasted with that of the earliest
toolmakers, who were simply after an attribute: a sharp cutting edge.
It didn’t matter to them what the tool looked like. Clearly, in the
ability to imagine form within a nucleus of stone we are seeing
evidence of a cognitive advance among the handaxe makers,
although it is anyone’s guess what other behavioral innovations
might have been associated with it.
Still, the anatomical form of Homo ergaster by itself shows that
hominins at this stage had definitively quit the shelter of the trees,
committing themselves to life on the open savanna (Walker & Leakey
1993). And although the archaeological record itself is silent on this
point, there are some inferences that can be made about the possible
changes in hominid lifestyle that might have accompanied this new
ecological commitment. Homo ergaster had a brain almost double
the size of its counterparts among the australopiths; and even though
some of this enlargement may have been accounted for by increased
body size, there were also energetic consequences. For brain tissue is
very energy-hungry: while your brain only accounts for about 2
percent of your body size, it uses upwards of 20 percent of the energy
you consume.
That extra energy has to come from somewhere, even though the
demands of the enlarging brain may have been partially
compensated for by the reduction of the hominid gut, also a very
metabolically “expensive” tissue (Aiello & Wheeler 1995). Apes need
very large intestines to process their rather low-quality diet of fruits
and leaves, and the broad pelvises and wide lower rib cages of the
australopiths certainly also supported voluminous guts. In contrast,
the slenderer build of the Nariokotome Boy not only argues for
significant reduction in his intestinal volume, but for a higher-quality
diet that his smaller abdominal organs needed to process more
effectively in order to feed his largish brain. The only evident source
of that higher-quality diet was animal fats and proteins, implying
that such foods were becoming ever-more-significant components of
the hominid diet. So, while it is entirely possible that the earliest
hominin butchers back in Dikika times were scavenging the
carcasses of animals that had been killed by predators or had died of
natural causes, or that later tool-using australopiths may have
“power-scavenged” by driving competitors away from carcasses by
hurling rocks or wielding heavy tree branches, by the Boy’s time it
had become absolutely imperative for hominids to have access to
animal fats and proteins, most likely by becoming active hunters of
small or medium-sized animals.
Some authorities would go even farther than this. The primatologist
Richard Wrangham (2009) has argued that, for an animal lacking a
specialized carnivore’s gut, consuming flesh raw would have yielded
insufficient energetic returns to underwrite the expansion of the
brain that became such a consistent theme in hominid evolution
from the Nariokotome Boy’s time onwards. He believes that the
requisite energy could have been obtained only if such foods had
been cooked, which of course implies the mastery of fire—a huge
behavioral leap. Wrangham points out that cooking not only
increases the energy available to the gut from both plant and animal
materials, even as it makes them easier to masticate, but that it also
detoxifies them, an important consideration in tropical
environments in which pathogens multiply with extreme rapidity.
What’s more, Wrangham suggests that, for a slow and biologically
defenseless biped out on the savanna, fire might even have been
crucial as a means of discouraging predators, especially at night.
Wrangham’s circumstantial case is a strong one, so it is unfortunate
that there is no clear evidence for the use of fire at living sites until a
million years ago at most (Berna et al. 2012), and that fire only
becomes a routine feature in the archaeological record following
about 400 thousand years (kyr) ago (Roebroeks & Villa 2011). Still, it
remains true that physical evidence for fire is typically fragile, so that
its apparent absence from the early record might be an artifact of
poor preservation. It is unfortunate that our perceptions may be
skewed in this way, for the domestication of fire was without any
question an epic innovation for hominids, and not only energetically.
For example, by providing a focus for social interaction at living sites
fire almost certainly had profound social as well as economic
consequences. Still, Zink and Lieberman (2016) have suggested that
there is a potential alternative to fire as an agent for making food
more digestible and its energy more available. Namely, its reduction
to small pieces prior to mastication, by the use of stone tools. These
authors have, moreover, linked this form of exogenous food
processing to the reduction of the chewing teeth that is such a salient
feature of the genus Homo. Whatever the case, the move to the
savanna is almost certainly indicative of increasing hominin
behavioral complexity, and it is unfortunate that we have no way of
gauging exactly how the individuals involved perceived the world.
 Out of Africa

The new body form was almost certainly implicated in the first
movement by hominins beyond the African continent. During the
australopith period hominins, while diversifying vigorously, had
remained confined to the continent of their birth. But by 1.8 myr ago
they had already penetrated as far as Dmanisi, a site in Georgia, just
to the east of the Black Sea (Lordkipanidze et al. 2013). And by not
long after that, they had reached all the way to eastern Asia, where
the iconic form Homo erectus is documented on the island of Java as
far back as 1.3 myr ago. Exactly what made hominins so
unprecedentedly mobile at this point is not known for sure, although
all hominins known from outside Africa have been attributed (rightly
or wrongly) to the genus Homo. This attribution is certainly not the
result of notably enlarged brains: the Dmanisi hominins, now known
from five skulls and two partial skeletons, all have brains in the
australopith-like 600 ml range (Figure 16.8). What’s more, brain size
is the main thing all the Dmanisi specimens have in common,
because in other respects quite diverse morphologies are
represented. Nonetheless, the entire assemblage has most recently
been attributed to a single “paleodeme” of the ubiquitous Homo
erectus (Lordkipanidze et al. 2013). Alternatively, the name Homo
georgicus is also available for them, in the unlikely event that all
belong to the same species.
Figure 16.8 Three-quarter view of the most distinctive hominid
cranium from Dmanisi, in the Republic of Georgia, holotype of
Homo georgicus. Approximately 1.8 million years old. Photo by
Jeffrey H. Schwartz.
Numerous stone artifacts are known from Dmanisi, and all are of the
ancient Oldowan kind. Improved technology thus does not seem to
be implicated in these hominins’ ability to leave the tropics and
penetrate the cooler temperate zone. And, to judge by the small brain
sizes of the hominins involved, neither does higher intelligence—if
indeed this quality maps efficiently onto brain size. Physically, this
leaves us only with the structure of the locomotor apparatus as a
potential explanation of the Dmanisi hominins’ remarkable mobility
—which will, it seems, remain the default explanation for the time
being. For, while the Dmanisi hominins all appear to have been of
rather short stature, their limb skeletons have been described as
more modern- than australopith-like. It is worth noting that an
external event such as a climatic amelioration might also have
permitted the ancestors of the Dmanisi hominins to expand beyond
Africa for the first time; but, however this colonization of new
environments was achieved, it is difficult based on the little we know
to find a specific behavioral explanation—either for the colonization
itself, or for the remarkable hominin adaptability it clearly implies.
The fossil and archaeological records both now suggest that
hominins had penetrated Europe, notably the Italian and Iberian
peninsulas, by about 1.4 myr ago (Bermudez de Castro et al. 2011)—
still in an Oldowan technological context. The Acheulean came late
to Europe, and barely managed to penetrate eastern Asia at all. As
might have been expected, both Europe and eastern Asia hosted
indigenous developments on the biological front, with Homo erectus
documented early on as a regional phenomenon in eastern China
and on Java, while the lineage leading to Homo neanderthalensis
was established in Europe well over 400 kyr ago.

The First Cosmopolitan Hominid

While handaxes did become on balance thinner and more elegant

over time, no radically new tools appeared on the scene for more
than a million years after their invention. The implication seems to
be that, whatever the exact nature of the ecological and behavioral
equilibria achieved by the early hominin occupants of the savanna,
they were successful and enduring even in the face of climates that
oscillated ever more severely. Clearly, when presented with
substantial local environmental challenges, the hominins of this
period responded by adapting old stone tools to new purposes, rather
than by inventing new technologies to cope, as modern Homo
sapiens typically does. As a result, it was stasis rather than change
that dominated the material record of the time. On the biological
front, in contrast, local diversification seems to have been the rule in
Africa, with a range of (as yet inadequately characterized)
morphologies documented between about 1.5 myr and 200 kyr.
Still, by about 600 kyr ago a new form, Homo heidelbergensis, had
appeared in both Africa and Europe. Possible representatives show
up in China somewhat later than this, and the species persisted in
places until possibly as little as 200 kyr ago. With Homo
heidelbergensis, then, we find the first hominin species that was
ubiquitous throughout the Old World. Individuals of this species
were strongly built and had massive faces, but they also possessed
brains in the 1200 ml range (compared to our 1330 ml: Holloway et
al. 2004). What is more, during the tenure of this species (although
rarely with a direct association between fossils and artifacts), several
very significant cultural innovations were made.
One such innovation is the deliberate construction of freestanding
shelters. During the 1960s, archaeologists excavating a 380 kyr-old
site on the Mediterranean coast of France uncovered evidence for
some hut-like structures (De Lumley & Boone 1976). The largest of
these (Figure 16.9) consisted of an oval ring of saplings stuck into the
ground and brought together at the top, its periphery reinforced by a
ring of stones. A gap in the stones indicated the hut’s entrance, just
inside which fire had burned in a hearth. This is the first good
evidence we have for the construction of artificial shelters that—in
various forms—rapidly became an increasingly common feature in
the archaeological record, along with fireplaces and evidence for
cooking. Interestingly, the associated stone artifacts found at the site
were relatively crude, although this may in part have been due to a
dearth of good raw materials in the vicinity.
Figure 16.9 Reconstruction of the largest of the 380 kyr-old hut-
like structures excavated at Terra Amata in southern France.
Through the cutaway can be seen, just inside the doorway, a hearth
in which a controlled fire had burned. Concept by Henry de Lumley;
drawing by Diana Salles.
Another startling finding from over 300 kyr ago was made during the
1990s in a brown-coal bog at Schoeningen in northern Germany
(Thieme, 1997). The anoxic conditions of the bog made possible the
preservation of a series of very unusual artifacts, most strikingly
wooden spears, a couple of meters long, that were carefully shaped to
a form very much like today’s javelins, with the center of balance
lying far forward. The spears were sharpened to a point, although in
the absence of a hard stone tip their penetrating power has been
called into question. Nevertheless, these are very clearly throwing
implements, suggesting that ambush-hunting was already practiced.
Using thrusting spears for hunting can be dangerous, because to use
them you have to be uncomfortably close to your quarry; but
throwing spears can be hurled safely from a distance. Their discovery
at Schoeningen implies that hunting techniques had become much
more sophisticated by this time than one might have guessed from
the stone tool kit alone. And the ability to kill from a distance
ushered in an important and rather sinister strand in subsequent
human history.
Enhancements to the stone tool kit are also evident at Schoeningen,
in the form of several pieces of wood that had been grooved at one
end, presumably to accept stone flakes. This is the first evidence we
have from anywhere for compound tools, which represent a huge
improvement in efficiency compared to flakes held just in the fingers.
What is more, the next advance in the manufacture of the stone tools
themselves also occurred during the tenure of Homo
heidelbergensis. This was the invention of the “prepared-core” tool,
whereby a high-quality stone core was carefully shaped on both sides
until a final blow could detach a more or less finished implement,
with a continuous cutting edge all around it (see Klein 2009). This
radically new technological innovation presumably reflects yet
further cognitive advance, though once more we find ourselves with
a picture of a generally increasing behavioral complexity without
knowing precisely how the hominins involved were perceiving the
world.
One thing does seem clear, though. Because, as resourceful and
intelligent as Homo heidelbergensis clearly was, its members do not
seem to have been processing information symbolically, as we do.
Cognitively, they were not simply more primitive versions of
ourselves. How do we know this? Well, since intellectual styles leave
no direct trace in the material record, we have to seek proxies for
them. And, to cut a long story short, the only sure proxies for
symbolic thought are explicitly symbolic objects. Just being complex
isn’t enough; you can evidently be very complex behaviorally without
being symbolic. And, significantly, Homo heidelbergensis
bequeathed us nothing that is unequivocally symbolic. A pebble
vaguely in the shape of a female torso, found at the 230 kyr-old site
of Berekhat Ram in the Golan Heights, was supposedly modified to
enhance its human resemblances; but it remains hotly disputed (see
d’Errico & Nowak 2000) and did not evidently form part of any
larger symbolic corpus developed by the hominin that made it.
 The Neanderthals

The best-documented of all extinct hominin species is Homo

neanderthalensis. Neanderthals flourished in Europe and western
Asia following about 200 kyr years ago, though their roots in the
region go back beyond a half-million years. Its members possessed
brains as large as ours, or even larger. They averaged about 1500 ml
in volume, the same as for Ice Age Homo sapiens (though for reasons
we can only speculate about, but that probably have to do with the
greater efficiency of the symbolic information processing algorithm,
our brains seem to have shrunk noticeably since: see discussion by
Tattersall 2018). However, those large Neanderthal brains were
enclosed in skulls of very different proportions. A large face jutted
out in front of a long, low braincase, rather than being retracted
beneath a rounded braincase as ours is. Brain casts reveal relatively
small prefrontal and large visual areas compared to those of Homo
sapiens. Below the neck, Neanderthal body skeletons betray a very
robust build, with numerous differences from Homo sapiens in
structural detail (Figure 16.10).
Figure 16.10 Comparison between a reconstructed Neanderthal
skeleton (left) and that of a modern human of similar stature. Note
the considerable differences between the two, most notably in
general robusticity and in cranial and thoracic/pelvic structure.
Homo neanderthalensis reconstruction by Gary Sawyer and Blaine
Maley; photo by Ken Mowbray.
Physically, then, Homo neanderthalensis presents us with a picture
of a hominid very different from modern Homo sapiens, despite
recent genomic indications that some interbreeding occurred soon
after the first Homo sapiens entered the Neanderthals’ European
redoubt some 40 thousand years ago (Green et al. 2010).
Significantly, within a few millennia following the initial Homo
sapiens invasion of their territory the Neanderthals were gone,
clearly as a result of competition rather than of wholesale
assimilation. The same may be said of the ghostly but widely
distributed Denisovans (Krause et al. 2010), known almost
exclusively from their DNA, who evidently interbred in Asia with
modern humans who were on their way to populate the Far East.
Adding to this picture of biological replacement is the contrast
between the archaeological records left behind by the two species.
Whereas the invading Homo sapiens entered Europe showing all the
signs of a fully-fledged modern sensibility, the Neanderthals left us
little if anything to indicate that they reasoned symbolically. They
were excellent craftsmen, who wielded the prepared-core stone-
working technology with great deftness, and who created compound
tools using resin mastic to haft them. But there is a strong argument
to be made that even sophisticated practical technologies of this sort
do not depend on a symbolic sensibility; they can be passed along by
imitation and can be understood at a purely intuitive level (Ohnuma
et al. 1997; Tattersall 1998). Saying this is not to belittle the
Neanderthals in any way; in their day they were as clever and
ingenious as any hominins that had ever lived. Their intuitive
capacities were evidently formidable; in some places, at least, they
habitually hunted fearsomely large game such as woolly mammoth
and rhino; they successfully accommodated to many different
geographic milieus; and they survived extreme climatic swings.
But, as clever as they undoubtedly were, the Neanderthals did not
show the spark of inventiveness that typifies modern humans; and
there are only sparse indications that they ever produced overtly
symbolic objects. So, while there are now hints that Neanderthals
may sometimes have decorated their bodies (e.g., Radovčić et al.
2015), and they have even been implicated in the coloring of cave
walls (Hoffmann et al. 2018; though see Slimak et al. 2018), there is
little reason to suppose that, on the whole, they had departed
substantially from long-established ways of manipulating
information. This is not to deny that the Neanderthals were complex
creatures who led complex lives, but simply to observe that the vast
preponderance of the evidence suggests that they interacted with the
world, and with each other, very differently from the ways in which
behaviorally modern humans do. To take just one minor example,
they carried out acts of cannibalism (a practice with a tradition in
Europe that long predated them) in an alien and rather chillingly
prosaic manner (see discussion in Tattersall 2012). In sum, while the
archaeological record of the Neanderthals furnishes us with a
portrait of a hardy and sophisticated species, that portrait is of a
hominin that did not do business in the world in the peculiar way
that distinguishes modern Homo sapiens .

The Evolution of Homo Sapiens

The earliest fossils of hominins that looked to all intents and

purposes like we do today, and that can with reasonable confidence
be assigned to our species Homo sapiens, come from sites in
Ethiopia dating between about 200 and 160 kyr ago (Whiteet al.
2003; McDougall et al. 2005). Homo sapiens is morphologically very
distinctive, and while there was clearly a lot of evolutionary
experimentation going on among African hominins in the period
before about 200 kyr ago, it is impossible currently to identify the
biological lineage from which H. sapiens sprang (as it is possible to
do with the Neanderthals). Significantly, though, the rather sparse
archaeological associations of the first H. sapiens are less than
impressive; and indeed, in one case they are strikingly archaic,
including, for instance, the very latest handaxes ever found on the
African continent (Klein 2009). At the same locality, an infant skull
had apparently been defleshed (a common practice going back to the
earliest Homo heidelbergensis in Africa), and carried around in a bag
for long enough to acquire a curious polish; what this implies about
the mental processes of the adult who carried it is anyone’s guess.
But there is at present no firm basis for concluding that the earliest
biologically recognizable Homo sapiens were behaving significantly
different from their hominin contemporaries.
The earliest archaeological intimations of a qualitative behavioral
change in the modern human lineage come only appreciably later
than the biological appearance of our species. At around 100 kyr ago
small marine shells, apparently pierced for stringing and sometimes
bearing traces of ochre (possibly derived from contact with painted
skin), begin to be found at coastal sites in northern and southern
Africa and in the Levant (Bouzouggar et al. 2007; d’Errico et al.
2010). The significance of such objects in isolation is arguable: one
swallow doesn’t make a summer, and complex behaviors by
individuals don’t necessarily map on to the behaviors of entire
societies. But the routine appearance of such “beads” at
archaeological sites might well herald a qualitative cognitive change,
since in all historically and ethnologically documented societies
bodily painting and ornamentation carry strongly symbolic
overtones of identity and status.
But we do not have to wait long for more compelling evidence of a
cognitive shift. At South Africa’s Blombos Cave, archaeological strata
about 77 kyr old have produced a series of smoothed ochre plaques
and plaque fragments (Figure 16.11) engraved with a geometric motif
(Henshilwood et al. 2002). There is some stratigraphic separation
among the plaques, suggesting that the geometric symbol concerned
had maintained its significance over time and was not just the result
of doodling by a single individual. Additionally, a similar motif has
been found at a slightly younger site some hundreds of kilometers
away (Texier et al. 2010), reinforcing this suggestion. What is more,
not far away along the southern coast of Africa, the caves of Pinnacle
Point have yielded 72 kyr-old (and likely much older) evidence of a
multi-stage fire-hardening technology that (unlike almost any other
Old Stone Age procedure) is so complex that it can be taken as prima
facie evidence of symbolic mental processes on the part of those who
developed and practiced it (Brown et al. 2009).

Figure 16.11 The best preserved of the small, geometrically

engraved ochre plaques from Blombos Cave on the southern coast of
South Africa. Drawn by Kayla Younkin.
So the conclusion has to be, at least for now, that Homo sapiens as
we know it today emerged in two distinct stages, significantly if not
greatly separated in time. First, the anatomically distinctive species
appeared, probably as the result of a change in gene regulation that
had cascading developmental consequences throughout the body.
This change would have arisen, and have become fixed, in one of the
tiny isolates of hominid hunters and gatherers that were thinly
scattered across the continent in the period around 100 thousand
years ago. A recent study suggests that this isolate might have been
located in northern Botswana (Chan et al. 2019), though the
mitochondrial record on which this conclusion largely depends may
not provide ideal resolution. A surmise about which we can be more
confident is that the many hard-tissue modifications that
accompanied the emergence of Homo sapiens were accompanied by
increased interconnectivities within the brain. It was those crucial
neural pathways that would underwrite the cognitive shift that was
to come.
If this scenario holds, the potential for symbolic cognition was
present from the very anatomical origins of our species. But clearly,
since the biology was necessarily already in place (just as the
ancestors of birds had possessed feathers long before using them to
fly), its possessor had to “discover” this potential through the action
of what was necessarily a cultural or behavioral stimulus (Tattersall
2012).
Most attempts to reconstruct the evolution of modern human
cognition have depended on a potential feedback role of increasing
social complexity, often seen as reflected in ToM—broadly, the ability
to know what another individual of your species is thinking. At first
glance, the undeniable trend toward hominid brain size increase over
the last two million years seems to support the search for a selective
agent that would have acted consistently over this long period. And
in an era of wildly oscillating climates and environments that clearly
provided the reverse of consistency, social imperatives are obviously
an attractive place to look for the causes of what looks like a long-
term trend. However, this option is undermined by the fact that
while brain expansion occurred independently in multiple lineages of
the genus Homo (Tattersall 2018), it led to symbolic cognition in
only one of them. And recall also that modern human brain sizes
have actually diminished since the acquisition of symbolic cognition.
It is naturally impossible for us to know for sure whether, given
enough time, the Neanderthals in Europe or Homo erectus in eastern
Asia (the other lineages of Homo in which brain size increase is
documented) might or might not have spontaneously attained
symbolic reasoning, or whether a Neanderthal infant raised as a
modern human might have been able to acquire it. But while they
make for nice thought experiments neither of these options seem at
all plausible, given both the radical nature of the developmental
reorganization we know Homo sapiens underwent just prior to
switching to the symbolic algorithm, and the alacrity with which the
new capacity was subsequently expressed and elaborated. All in all, it
seems most reasonable to regard the highly complex ToMs of which
human beings are unquestionably capable as expressions of symbolic
cognition, rather than as drivers of it.
A much more plausible stimulus for the arrival of symbolic cognition
is the spontaneous invention of externalized language. Like the
cognitive capacity that underlies it, language depends on the
formation of a vocabulary of mental symbols that can then be
recombined to permit an unlimited variety of statements from a
finite number of elements. Given the improbability that a hominid
could be linguistic without being symbolic (and almost certainly vice
versa), it seems entirely reasonable to associate the two, and to
surmise that spoken language was spontaneously invented, in some
small late Pleistocene African isolate of the already biologically
preadapted Homo sapiens, by the simple expedient of attaching
specific meaning to specific sounds. Possibly this was done initially
by children at play, for we know that children can spontaneously
invent and subsequently elaborate grammatical sign languages
(Senghas & Coppola 2001). Necessarily superimposed on earlier
modes of ordering and communicating information, this new mode
of expression could have co-opted newly existing pathways between
sensory, association, and motor modalities in the brain into a
feedback that inaugurated the unprecedented symbolic cognitive
mode (Tattersall 2012). And since (unlike ToM) spoken language is
externalized, once established it would have been poised to spread
rapidly within a species that consisted of populations that were all
biologically pre-enabled for it.
Of course, a virtually instantaneous shift from a nonsymbolic,
nonlinguistic ancestor to a symbolic, linguistic descendant may seem
intuitively implausible, given that this is likely the most profound
cognitive transition that ever happened. But we do know that the
transition has to have occurred; we know that it is yet more
improbable that it could slowly have developed over the eons; and,
most importantly of all, we know that the archaeological record
shows a fundamental inflection at the point at which we see the first
stirrings of symbolic activities.
Over the first 2.5 myr of the archaeological record, material change
had been both highly sporadic and extremely rare. The pace began to
pick up a bit once Homo heidelbergensis was on the scene; but it was
only after our own species Homo sapiens had made its debut that
cultural change became the norm rather than the exception. At about
70 kyr ago (new evidence from China suggests possibly a bit earlier:
Wu et al. 2015), the newly symbolic Homo sapiens moved
definitively out of Africa. Earlier forays by nonsymbolic Homo
sapiens had evidently been unsuccessful (see Tattersall 2012); but,
once it had adopted symbolic thinking, our species rapidly took over
the world, eliminating all hominid competition in the process. What
is more, once symbolic behaviors had begun to be expressed, change
became not simply routine in human experience but exponential. It
took at most 50 kyr after those first stirrings in Africa before the
great age of cave and portable art in Europe (and islands in eastern
Asia: Aubert et al. 2014, 2018) began, furnishing us with some of the
most powerful expressions of the human symbolic spirit ever
documented (Figure 16.12). As soon as that extraordinary tradition
had begun to wane with the environmental changes of the end of the
last Ice Age, the most fateful behavioral revolution of all occurred:
the adoption of sedentary existence, and humanity’s consequent
declaration of independence from the ecosphere (Eldredge 1995).
And then, within a mere 10 thousand years, cities had blossomed, the
Industrial Revolution had begun to irrevocably change the Earth’s
surface and atmosphere, and people had walked on the moon.

Figure 16.12 Monochrome rendering of a now badly faded

polychrome wall painting, probably around 14 thousand years old, at
the southwestern French cave of Font de Gaume. A female reindeer
kneels before a male, who leans forward and delicately licks (or
sniffs) her forehead. Drawing by Diana Salles, after a rendering by
the Abbé Breuil.

SUMMARY AND CONCLUSIONS

Today we accept as normal both that our species is the lone hominid
on the planet, and that through our own activities the world will
change almost out of recognition in our lifetimes. But it was not ever
thus. For almost all of its existence, members of the hominid family
had simply responded as best they could to the climatic and
environmental vicissitudes thrown at them by nature, mainly by
adapting technologically fairly meager old ways and means to new
purposes. Now the dynamic is entirely different and, for the first
time in hominin history, we are finding it necessary to deal with the
ramifying unintended consequences of our heedless activities. Or
somehow to justify not dealing with them.

FURTHER READING
Recent books on the subject by Ian Tattersall are Masters of the
Planet: The Search for Our Human Origins (2012), The Strange
Case of the Rickety Cossack and Other Cautionary Tales from
Human Evolution (2015), and (with Rob DeSalle) The Accidental
Homo sapiens: Genetics, Behavior, and Free Will (2019).

REFERENCES
Aiello, L. & Wheeler, P. 1995. The expensive-tissue hypothesis: The
brain and the digestive system in human and primate evolution.
Current Anthropology, 36, 199–221.
Aubert, M., Brumm, A., Ramli, M., et al. 2014. Pleistocene cave art
from Sulawesi, Indonesia. Nature, 514, 223–7.
Aubert, M., Setiawan, P., Oktaviana, A.A., et al. 2018. Palaeolithic
cave art in Borneo. Nature, 564, 254–7.
Bermúdez De Castro, J.M., Martinon-Torres, M., Gómez-Robles, A.,
et al. 2011. Early Pleistocene human mandible from Sima del
Elefante (TE) cave site in Sierra de Atapuerca (Spain): A
comparative morphological study. Journal of Human Evolution,
61, 12–25.
Berna, F., Goldberg, P., Horwitz, L.K., et al. 2012. Microstratigraphic
evidence of in situ fire in the Acheulean strata of Wonderwerk
Cave, Northern Cape province, South Africa. Proceedings of the
National Academy of Sciences, USA, 109(20), 1215–20.
Bouzouggar, A., Barton, N., Vanhaeren, M., et al. 2007. 82,000-year-
old shell beads from North Africa and implications for the origins
of modern human behavior. Proceedings of the National
Academy of Sciences, USA, 104, 9964–9.
Brown, K.S., Marean, C.W., Herries, A.I.R., et al 2009. Fire as an
engineering tool of early modern humans. Science, 325, 859–62.
Chan, E.K.F., Timmermann, A., Baldi, B.F., Moore, A.E., et al. 2019.
Human origins in a southern African palaeo-wetland and first
migrations. Nature, 575, 185–189. 10.1038/s41586-019-1714-1.
Cohen, J. 2010. Almost Chimpanzee: Searching for What Makes Us
Human in Rainforests, Labs, Sanctuaries, and Zoos. New York:
Times Books/Henry Holt.
d’Errico, F. & Nowel, A. 2000. A new look at the Berekhat Ram
figurine: implications for the origins of symbolism. Cambridge
Archaeological Journal, 10, 123–67.
d’Errico, F., Salomon, H., Vignaud, C. & Stringer, C. 2010. Pigments
from Middle Paleolithic levels of es-Skhūl (Mount Carmel, Israel).
Journal of Archaeological Science, 37, 3099–3110.
De Lumley, H. & Boone, Y. 1976. Les structures d’habitat au
Paléolithique inférieur. In: H. De Lumley (ed.), La Préhistoire
française, vol. 1, pp. 635–43. Paris: CNRS.
Dean, M.C. & Smith, H. 2009. Growth and development of the
Nariokotome Youth, KNM-WT 15000. In: F.E. Grine (ed.), Origin
and Early Evolution of the Genus Homo, pp. 101–20. Heidelberg:
Springer.
Eldredge, N. 1995. Dominion. Berkeley: University of California
Press.
Fleagle, J.G. 2013. Primate Adaptation and Evolution, 3rd ed. San
Diego, CA: Academic Press.
Gibbons, A. 2006. The First Human: The Race to Discover Our
Earliest Ancestors. New York: Doubleday.
Graves, R.R., Lupo, A.C., McCarthy, R.C., Wescott, D.J. &
Cunningham, D.L. 2010. Just how strapping was KNM-
WT15000? Journal of Human Evolution, 59, 542–54.
Green, R.E., Krause, J., Briggs, A.W., et al. 2010. A draft sequence of
the Neandertal genome. Science, 328, 710–22.
Harcourt-Smith, W.E.H. 2014. Origin of bipedal locomotion. In: W.
Henke & I. Tattersall (eds.), Handbook of Paleoanthropology, vol.
3, pp. 1919–60. Heidelberg and New York: Springer.
Harmand, S., Lewis, J.E., Feibel, C.S., Lepre, C.J., et al. 2015. 3.3-
million-year-old stone tools from Lomekwi 3, West Turkana,
Kenya. Nature, 521, 310–6.
Hart, D. & Sussman, R.W. 2009. Man the Hunted: Primates,
Predators, and Human Evolution, Expanded Edition ed. Boulder,
CO: Westview Press.
Henshilwood, C.S., d’Errico, F., Yates, R., et al. 2002. Emergence of
modern human behavior: Middle Stone Age engravings from
South Africa. Science, 295, 1278–80.
Hoffmann, D.L., Standish, C.D., Pettitt, P.B., et al. 2018. U-Th dating
of carbonate crusts reveals Neandertal origin of Iberian cave art.
Science, 359, 912–5.
Holloway, R.L., Broadfield, D.C. & Yuan, M.S. 2004. The Human
Fossil Record, Vol. 3: Hominid Endocasts, The Paleoneurological
Evidence. New York: Wiley-Liss.
Johanson, D.J. & Edgar, B. 2006. From Lucy to Language, 2nd ed.
New York: Simon and Schuster.
Kano, F., Krupenye, C., Satoshi Hirata, S., Tomonaga, N. & Call, J.
2019. Great apes use self-experience to anticipate an agent’s
action in a false-belief test. Proceedings of the National Academy
of Sciences USA, 116, 20904–9.
Klein, R. 2009. The Human Career: Human Biological and Cultural
Origins, 3rd ed. Chicago: University of Chicago Press.
Krause, J., Fu, Q., Good, J.M., Viola, B., Shunkov, M.V., Derevianko,
A.P. & Pääbo, S. 2010. The complete mitochondrial DNA genome
of an unknown hominin from southern Siberia. Nature,
464(7290), 894–7.
Krupenye, C., Kano, F., Hirata, S., Call, J. & Tomasello, C. 2016.
Great apes anticipate that other individuals will act according to
false beliefs. Science, 354, 110–4.
Lepre, C.J., Roche, H., Kent, D.V., et al. 2011. An earlier origin for
the Acheulian. Nature, 477, 82–5.
Lordkipanidze, D., Ponce De León, M.S., Margvelashvili, A., et al.
2013. A complete skull from Dmanisi, Georgia, and the
evolutionary biology of early Homo. Science, 342, 326–31.
McDougall, I., Brown, F.H. & Fleagle, J.G. 2005. Stratigraphic
placement and age of modern humans from Kibish, Ethiopia.
Nature, 433, 733–6.
McPherron, S., Alemseged, Z., Marean, C.W., et al. 2010. Evidence
for stone-tool-assisted consumption of animal tissues before 3.39
million years ago at Dikika, Ethiopia. Nature, 466, 857–60.
Newton-Fisher, N.E. 2014. The hunting behavior and carnivory of
wild chimpanzees. In: W. Henke & I. Tattersall (eds.), Handbook
of Paleoanthropology, 2nd ed., vol. 2, pp. 1661–92. Heidelberg
and New York: Springer.
Ohnuma, K., Aoki, K. & Akazawa, T. 1997. Transmission of tool-
making through verbal and non-verbal communication:
Preliminary experiments in Levallois flake production.
Anthropological Science, 105(3), 159–68.
Plummer, T. 2004. Flaked stones and old bones: Biological and
cultural evolution at the dawn of technology. Yearbook of
Physical Anthropology, 47, 118–64.
Povinelli, D.J. 2004. Behind the ape’s appearance: Escaping
anthropocentrism in the study of other minds. Daedalus, 133(1),
29–41.
Radovčić, D., Sršen, A.O., Radovčić, J. & Frayer, D.W. 2015.
Evidence for Neandertal jewelry: Modified White-tailed Eagle
claws at Krapina. PLoS One, 10(3), e0119802.
Roebroeks, W. & Villa, P. 2011. On the earliest evidence for habitual
use of fire in Europe. Proceedings of the National Academy of
Sciences USA, 108(13), 5209–14.
Roth, G. & Dicke, U. 2005. Evolution of the brain and intelligence.
Trends in Cognitive Science, 9(5), 250–7.
Semaw, S. 2000. The world’s earliest stone artifacts from Gona,
Ethiopia: Their implications for understanding stone technology
and patterns of human evolution between 2.6-1.5 million years
ago. Journal of Archaeological Science, 27, 1197–1214.
Senghas, A. & Coppola, M. 2001. Children creating sign language.
Psychological Science, 12(4), 323–8.
Slimak, L., Fietzke, J., Geneste, J.-M. & Ontanon, R. 2018. Comment
on “U-Th dating of carbonate crusts reveals Neandertal origin of
Iberian cave art.” Science, 361, 912–5.
Sponheimer, M. & Lee-Thorp, J. 2014. Hominin paleodiets: The
contribution of stable isotopes. In: W. Henke & I. Tattersall (eds.),
Handbook of Paleoanthropology, 2nd ed., vol. 1, pp. 671–702.
Heidelberg: Springer.
Tattersall, I. 1998. The origin of the human capacity. New York:
American Museum of Natural History (James Arthur Lecture
series).
Tattersall, I. 2012. Masters of the Planet: In Search of Our Human
Origins. New York: Palgrave Macmillan.
Tattersall, I. 2018. Brain size and the emergence of modern human
cognition. In: J.H. Schwartz (ed.), Rethinking Human Evolution,
pp. 319–34. Cambridge, MA: MIT Press.
Texier, P.J., Porraz, G., Parkington, J., et al. 2010. A Howiesons
Poort tradition of engraving ostrich eggshell containers dated to
60,000 years ago at Diepkloof Rock Shelter, South Africa.
 Proceedings of the National Academy of Sciences USA, 107,
6180–5.
Thieme, H. 1997. Lower Palaeolithic hunting spears from Germany.
Nature, 385, 807–10.
Toth, N., Schick, K.D., Savage-Rumbaugh, E.S., Sevcik, R.A. &
Rumbaugh, D.M. 1993. Pan the toolmaker: Investigations into the
stone tool-making and tool-using capabilities of a bonobo (Pan
paniscus). Journal of Archaeological Science, 20, 81–91.
Van Couvering, J.A. 2014. Quaternary geology and
paleoenvironment. In: W. Henke & I. Tattersall (eds.), Handbook
of Paleoanthropology, 2nd ed., vol. 1, pp. 537–56. Heidelberg and
New York: Springer.
Vanhaeren, M., d’Errico, F., Stringer, C., et al. 2006. Middle
Paleolithic shell beads in Israel and Algeria. Science, 312, 1785–8.
Walker, A.C. & Leakey, R.E.F. 1993. The Nariokotome Homo erectus
skeleton. Cambridge, MA: Harvard University Press.
White, T.D., Asfaw, B., DeGusta, D., et al. 2003. Pleistocene Homo
sapiens from Middle Awash, Ethiopia. Nature, 423, 742–7.
Wrangham, R. 2009. Catching Fire: How Cooking Made Us
Human. New York: Basic Books.
Wu, L., Martinon-Torres, M., Cai, Y.-J. & Xing, S. 10 others. 2015.
The earliest unequivocally modern humans in southern China.
Nature, 526, 596–600.
Zink, K.D. & Lieberman, D.E. 2016. Impact of meat and Lower
Paleolithic food processing techniques on chewing in humans.
Nature, 531, 139–44.
17
evolutionary approaches to human behavior
GILLIAN R. BROWN, CATHARINE P. CROSS, AND KEVIN N.
LALAND
INTRODUCTION
In this chapter, we overview how evolutionary theory has been
applied to investigate human behavior and cognition, examining
whether human beings can be studied in the same way as any
other species of animal. Charles Darwin proposed that natural
selection and sexual selection could act on our behavioral traits,
just as on our physical traits, and argued that human beings
should not be considered as separate from the rest of the animal
kingdom. In the first half of this chapter, we briefly review the
history of applying evolutionary theory to human behavior from
Darwin to the present day, highlighting recurrent controversies
such as the “nature versus nurture” debate. During the 1950s,
ethologists applied their methodologies to human behavioral
traits, and the advent of sociobiology in the 1970s prompted a
further rush of evolutionary hypotheses for human behavior,
leading to vocal criticism from some opponents. Subsequently, a
number of modern sub-fields have emerged, including Human
Behavioral Ecology, Evolutionary Psychology, and Cultural
Evolution. We provide examples of research within these sub-
fields and evaluate the strengths and weaknesses of each
approach, before examining whether the core assumptions of
these sub-fields are complementary or contradictory. In the
second half of this chapter, we critically evaluate the assumption
that human beings can be studied as if they were any other
animal species. Human beings are potentially different from
other animals in terms of (i) their reliance on culture and the
extent to which human beings modify their selective
environments, and (ii) the fidelity, efficiency, and breadth of
information transfer, which is enhanced by language and
teaching. Other animal species modify their environments and
exhibit social learning, but not to the same extent as human
beings, and we provide examples of how our cultural traits have
altered selection pressures on the human genome. Language and
teaching capacities of human beings are clearly more potent than
related systems in other species, and human beings exhibit high
levels of cooperation, with potential implications for the
 evolutionary trajectory of our species. We conclude that human
behavior is often on a continuum with other animals but is also
distinctive in some important respects, and that certain aspects of
our cognition have dramatically affected the selection pressures
acting on the human mind. Finally, we argue that the resolution
of current debates surrounding the evolution of human behavior
will enhance our understanding of the behavior of nonhuman
animals.

Applying Evolutionary Theory to Human

Behavior

Evolutionary explanations for human behavior have entered into a

broad range of academic disciplines, including psychology, medicine,
economics, and criminology (e.g., Wilson & Gowdy 2013; Buss 2015;
Durrant & Ward 2015; Brüne & Schiefenhövel 2019). In addition,
evolutionary theory has been used to describe changes in the
frequencies of entities other than genes (Plotkin 1994; Hull et al.
2001), such as words in languages and connections in social
networks (e.g., Kossinets & Watts 2006; Pagel et al. 2007).
Evolutionary theory has thus had a tremendous impact on our
understanding of the human mind and behavior. However, some
examples of research on evolution and human behavior that have
entered the mainstream media have rightly provoked strong
criticism, given that any pattern of behavior or physical trait could be
described as being adaptive in our ancestral environment
(Monaghan & Birkhead 2001). More generally, some critics have
argued that evolutionary accounts bolster the erroneous idea that
genes “control” our behavior and that inequalities between the sexes,
races, or social classes are therefore inevitable (e.g., Rose et al. 1984;
Buller 2006), and many social scientists remain suspicious of, even
hostile to, evolutionary approaches (Ingold 2007; Toren 2018).
Nonspecialists are, understandably, left wondering whether to accept
or reject the idea that evolutionary theory can be applied to human
behavior.
In the following section, we provide an overview of the debates
surrounding the application of evolutionary theory to human
behavior, with the intention of allowing readers to make evaluations
for themselves. We show that some questions, such as the roles of
“nature” and “nurture,” have been debated for centuries, whereas
new questions about the evolution of our species have been raised by
more recent sources of evidence, such as the Human Genome
Project. We start by briefly summarizing the history of applying
evolutionary theory to human behavior from Darwin to the present
day, in order to provide a clearer understanding of how the current
debates and controversies in this field arose. We provide a short
summary of how Darwin applied his theories of natural selection and
sexual selection to human beings, how the idea of “instinctive
behavior” was reinvigorated by the ethologists, and how the advent
of sociobiology, which revolutionized our understanding of animal
behavior, was met with strong criticism when applied to human
beings. We then introduce three modern sub-fields within the
human evolutionary behavioral sciences, namely Human Behavioral
Ecology, Evolutionary Psychology, and Cultural Evolution. We
evaluate the key assumptions and methodologies that underpin these
sub-fields and finish with a brief discussion of some specific points of
contention, such as the extent of “adaptive lag” in human behavior
and role of culture in the evolutionary process.

From Darwin to behaviorism

In The Origin of Species (1859), Darwin laid out his theory of
natural selection, which postulated that individuals in a population
whose anatomical, physiological, and behavioral traits best fitted the
environment would have the greatest chances of surviving and
reproducing (Figure 17.1). If these traits were inherited by offspring,
then the next generation would contain a higher frequency of
individuals with these fitter traits, and the characteristics of the
population would change over time. The key components required
for natural selection to operate, therefore, are (i) variation between
individuals in a particular trait, (ii) competition between trait-
bearers for survival and reproduction, and (iii) heritability of the
trait across generations. While much of the controversy that
surrounded publication of The Origin of Species was directed at the
proposition that human beings evolved from an ape-like ancestor,
Darwin (1859) did not specifically discuss human evolution in The
Origin of Species, except in the final pages, where he stated:
Figure 17.1 Charles Darwin (from Wikimedia Commons).
 In the distant future I see open fields for far more important
researches. Psychology will be based on a new foundation, that
of the necessary acquirement of each mental capacity by
gradation. Light will be thrown on the origin of man and his
history.
Subsequently, Darwin (1871) argued that the mental lives of human
beings have been subject to natural selection and merely differ from
those of other animals in terms of the degree of intellectual and
cognitive capacities, rather than in the types of mental processes
possessed by ourselves and other animals. For Darwin, human
mental capacities were on a continuum with all other species and not
set apart in a pre-ordained hierarchy. Darwin’s supporter, Thomas
Huxley (1863), made the parallel argument that human anatomy,
physiology, and life history are continuous with those of other
animals (Figure 17.2). The argument for continuity is supported by
compelling evidence for impressive cognitive abilities in nonhuman
animals (e.g., Tomasello & Call 1997; Güntürkün & Bugnyar 2016;
Seed & Mayer 2017), alongside genetic evidence confirming that
great apes are our closest living relatives (Goodman 1999; Ellegren
2005).
Figure 17.2 Frontispiece of Thomas Huxley’s (1863) Evidence as
to Man’s Place in Nature (from Wikimedia Commons).
More than a decade after The Origin of Species, Darwin published
his views on human behavior in The Descent of Man and Selection in
Relation to Sex (1871) and in The Expression of Emotions in Man
and Animals (1872). In both books, Darwin again sought to reduce
the presumed distance between the cognitive and emotional lives of
human beings and other animals, arguing that human beings have
been influenced by the same evolutionary processes as the rest of the
natural world. For example, in The Descent of Man, Darwin
proposed that sex differences in mental capacities could relate to
selection that reflects male competition for mates and female
nurturance of offspring, and that the diversity in human secondary
sexual traits could relate to cultural differences in aesthetic tastes
that lead to selection for specific physical attributes. Some of
Darwin’s views on the sexes now appear dated, and contemporary
writers, such as Antoinette Brown Blackwell (1875), were already
arguing that evolutionary theory did not require females to be
considered as inferior to males. Blackwell pointed out that the sexual
divisions of labor that characterized the late nineteenth century
societies of Europe and North American did not reflect the patterns
typical in other animals—these controversies have continued to the
present day (see Box 17.1).
Box 17.1 Darwinism and sex differences in human behavior

While Darwin (1859) did not write explicitly about human

evolution in The Origin of Species, he did hint a few times in the
final pages that human beings were not excluded from his theory
of evolution by natural selection. Darwin noted the commonality
between the patterning of bones in the limbs of bats, porpoises,
horses, and humans, and argued that these commonalities “at
once explain themselves on the theory of descent with slow and
slight successive modifications” (1859, p. 424). He also suggested
that, because of his theory, “Light will be thrown on the origin of
man and his history” (1859, p. 432). Later, Darwin (1871)
surmised that sexual selection had left its mark on human beings
in the form of differences in temperament and intelligence
between men and women. He described men as more
pugnacious, competitive, ambitious, and selfish than women,
resulting from the necessity of competition between males of
many species, including most mammals, for securing access to
females and hence reproducing. Darwin described women as
more tender and selfless; he believed these characteristics were
the result of needing to have these qualities to raise offspring (for
more on Darwin’s writings about women and contemporary
women’s responses, see Hamlin 2014). With regard to
intelligence, Darwin argued (citing Galton) that “if men are
capable of decided eminence over women in many subjects, the
average standard of mental power in man must be above that of
woman” (1871, p. 327). He further argued that sexual selection
had probably been an important driver of this sex difference. On
the other hand, he suggested that the difference was not
immutable: “In order that woman should reach the same
standard as man, she ought, when nearly adult, to be trained to
energy and perseverance, and to have her reason and imagination
exercised to the highest point” (1871, p. 329).
While modern research has thoroughly dispelled the notion that
men and women differ in levels of intelligence, the idea that sex
differences in behavioral tendencies and personality traits have
resulted from differential selection pressures acting on males and
females during our ancestral past remains prevalent. For
 example, a large body of literature is devoted to the argument
that sex differences in human aggression are the legacy of sexual
selection for competition in males (Archer 2009), and women’s
greater role in childcare has been suggested to underlie women’s
greater fearfulness (Campbell 1999), behavioral control
(Bjorklund & Kipp 1996), and risk-aversion (Cross et al. 2011).
Men are assumed to be more competitive and willing to take
risks, and less choosy in their mating decisions than women, as a
result of the greater variation in male, compared to female,
reproductive success (Bateman 1948; Trivers 1972); in many
mammal species, a subset of males will be very successful in
competing for, and attracting, mating partners and will
consequently have much higher reproductive success than other
males, while all females will tend to have similar reproductive
success to each other due to the costs imposed by raising
offspring, particularly gestation and lactation. However, modern
sexual selection theory has shown that levels of competitiveness
and choosiness will vary according to factors other than the
relative amount of investment by males and females in offspring,
including the probability of encountering potential mating
partners and rates of mortality (e.g., Kokko & Monaghan 2001).
Therefore, current views on the evolution of sex differences in
behavior focus on the impact of cultural processes on the
development of sex differences and the diversity of sex roles
across human populations (e.g., Brown et al. 2009) and time
periods (Cross et al. 2013).

In The Expression of Emotions, Darwin documented the emotional

lives of nonhuman animals and proposed that emotions such as fear,
pleasure, and shyness, which were assumed to be specific to human
beings, are found in a wide range of species. For instance, Darwin
described how monkeys can be fearful, ants playful, and dogs
courageous. Furthermore, Darwin argued that similar facial
expressions, such as the fearful grimaces of monkeys and appeasing
smiles in human beings, are associated with the same specific
emotions across species. While Darwin’s use of anecdotal examples
lacks the rigor of modern science, subsequent research has
supported the idea that nonhuman animals share some emotional
states with human beings (Panksepp 2011; Damasio & Carvalho
2013), and human emotions and expressions have been found to
exhibit both universal characteristics and cross-cultural variation
(Ekman 1994; Elfenbein & Ambady 2002). More broadly, the idea
that evolutionary theory can be usefully applied to the study of our
own species had an immediate and long-lasting impact on the
development of numerous scientific disciplines (Boakes 1984).
Within psychology, for example, the notion that our mental
capacities have evolved was further developed by Herbert Spencer,
who wrote the highly influential Principles of Psychology (1855,
1870) and coined the phrase “survival of the fittest.” Spencer argued
that different species exhibit heritable differences in behavioral
tendencies and that the mental abilities of human beings can be
considered as more complex forms of similar processes found in
“lower” organisms. However, Spencer also argued that individuals in
Westernized societies were inherently more intelligent than
individuals in other societies and that competition between
institutions and business within societies was “natural.” The term
“Social Darwinism” is commonly applied to the social theories
propagated by Spencer and others that advocated self-interest and
competition at the expense of weaker individuals or groups. Darwin
was, in fact, a highly vocal opponent of the idea that different groups
or “races” of people should be considered as superior to others and
spent much effort on fighting for anti-slavery laws (Desmond &
Moore 2009). Regardless of Darwin’s protestations, however, by the
end of the nineteenth century, the idea that human behavior is
underpinned by “instincts” and unlearned abilities that vary between
populations, sexes, and social classes had become accepted within
the field of psychology.
In the early 1900s, the new field of behaviorism gained momentum
within psychology. This field emphasized the importance of learning
and counteracted the notion that “instincts” could explain the
majority of differences between individuals. In a key paper, John B.
Watson (1913) outlined the behaviorist’s view of human psychology
and argued that studies on learning in nonhuman animals should be
more appreciated as being relevant to our understanding of human
learning (Figure 17.3). Behaviorists favored direct observation of
behavior under experimental conditions and devised a set of learning
rules that could be demonstrated to fit with observable behavioral
data from rodents and pigeons. Later experimental studies
confirmed that similar learning rules are deployed by human beings.
Watson’s (1924) research on child development further showed his
emphasis on the role of conditioning and other forms of learning on
behavior. Researchers studying learning processes in nonhuman
animals during the mid-1900s, such as B. F. Skinner, were generally
referred to as “comparative psychologists,” even though most studies
were conducted on a limited range of laboratory species and rarely
made explicit between-species comparisons. In recent decades,
researchers have substantially built upon the early work of the
comparative psychologists by investigating cognitive abilities in a
broader range of animal species (e.g., Emery & Clayton 2004; Bshary
et al. 2014) and by examining how cognitive abilities vary in relation
to life-history traits (e.g., Boogert et al. 2018).
Figure 17.3 John B. Watson (from Wikimedia Commons).

Human ethology and sociobiology

In the mid-twentieth century, the emerging field of ethology
reignited interest in the idea of “instinctive” behavior in animals,
with a focus on relatively “fixed” motor patterns and species-typical
behavior. In contrast to the early comparative psychologists, who
generally studied behavior under laboratory conditions, the
ethologists conducted experiments on the behavior of nonhuman
animals in natural environments. For example, in order to examine
how digger wasps locate their nest entrances, the ethologist Niko
Tinbergen set up rings of pine cones around a nest entrance, then
moved the pine cones a short distance away while the wasp was
foraging, which resulted in the wasp returning to the incorrect
entrance location (Tinbergen & Kruyt 1932; Figure 17.4). Tinbergen
concluded that wasps were using the pine cones as visual cues in
memory formation. Another influential ethologist, Konrad Lorenz
(1935), described how young birds imprint onto adults during early
stages of life and subsequently follow them around. The key concepts
of ethology were outlined by Tinbergen in his influential book The
Study of Instinct (1951). Behavior was characterized as species-
specific and shaped by evolved tendencies, while learning was
perceived to be tightly constrained toward adaptive goals. Key
concepts, such as “instincts,” were soon replaced by a more balanced
approach to the study of animal behavior that acknowledged the key
role of behavioral flexibility and individual learning, as well as the
constant interplay between internal and external causal factors
(Hinde 1966). However, the strong tradition of field experimentation
of early ethology, and emphasis on species-specific adaptations, has
been maintained within animal behavior research to the present day.
Figure 17.4 Field studies have shown that female digger wasps
(Philanthus triangulum) learn the location of their nests using local
landmarks (Tinbergen & Kruyt 1932). a) When a female wasp was in
a nest, the experimenters placed a ring of twenty pine cones around
the entrance; upon emerging, the wasp spent several seconds flying
over the site, before leaving the area. b) The experimenters then
moved the pine cones approximately 30 cm away from the nest
entrance, and the wasp was observed to return to the incorrect nest
location marked by the pine cones (from Tinbergen 1951).
Some ethologists were also willing to extend their concepts and
methodologies to the study of human behavior. For instance, Robert
Hinde (1974), who studied the development of behavior in both
children and nonhuman primates, provided novel insights into the
dynamics of human social relationships and attachment processes.
The research of Hinde and similarly minded child psychologists,
such as John Bowlby (1969), was highly influential in the
development of attachment theory, which still shapes our
understanding of the link between early childhood experiences and
later behavioral traits, such as personality. However, other
applications of ethological concepts to human behavior were more
controversial. For example, in On Aggression, Lorenz (1966) argued
that fighting and war were natural expressions of human aggressive
“instincts,” and that aggressive impulses are inevitably manifested
and must be released, either through fighting or competitive sports.
Lorenz’s views on aggression received immediate criticism, including
from other ethologists. Similarly, The Naked Ape, written by
ethologist Desmond Morris (1967), provoked a hostile response from
academics, despite popular appeal and impressive sales figures.
Morris described modern human behavior as inevitably reflecting an
ancestral stage in human evolution and presented scientifically
unsupported statements about how human behavior can be
interpreted, based on the assumption that we are “naked apes” living
in a modern world to which we are ill-suited.
Another major controversy in the application of evolutionary theory
to human behavior surrounded the publication of Sociobiology: The
New Synthesis by Edward Wilson (1975). Wilson was an
entomologist who synthesized recent advances in evolutionary
approaches to animal behavior that had emerged from the research
of George Williams, William Hamilton, Robert Trivers, and others.
This new perspective stressed the importance of considering how
selection acts at the level of the gene, rather than the species, and
replaced previous evolutionary hypotheses, known as “group
selection” hypotheses, that had relied on explanations for how the
group or species might benefit from an individual’s behavior. For
example, a group-selectionist interpretation of the roaring and
posturing behavior shown by pairs of male stags during the rutting
season would be that, when a subordinate male withdraws from the
confrontation, this male is acting for the good of the species by
allowing the more dominant, healthier male to sire offspring. Naïve
group-selectionist arguments were shown to be problematic by
Williams (1966) and were succeeded by a new perspective that
focused on selection at the level of the gene. For instance, in the
above example, the new perspective proposed that both males are
attempting to gain mating opportunities and that selection had
favored genes that underpinned the expression of optimal behavioral
strategies. Individual stags that engaged in roaring and posturing,
rather than immediately resorting to aggressive interactions, would
potentially avoid the high costs of losing fights and therefore have
higher lifetime reproductive success, assuming that roaring and
posturing provide at least partially reliable information about the
fighting abilities of opponents. This “gene’s-eye perspective”
revolutionized how the behavior of nonhuman animals was
interpreted and still forms the core of animal behavior research.
The major controversy resulted from the application of the new
gene-selectionist ideas, which Wilson labeled as “sociobiology,” to
human behavior. In the last chapter of his book, Wilson (1975)
provided numerous examples of how sociobiology could be applied
to a broad range of questions about human behavior, including the
sexual division of labor, homosexuality, morality, and social
stratification. For instance, Wilson suggested that genes underlying
homosexuality could have been selected if homosexual individuals
enhanced the reproductive success of close relatives by investing
time and energy in raising their relatives’ offspring, based on the
concept of “kin selection” that had proven successful in explaining
the presence of worker castes in social insects. Perhaps
unsurprisingly, Wilson’s somewhat naïve and simplistic ideas about
human behavior were received with vocal criticism, including from
some prominent biologists, and Wilson’s ambitious claim that all of
the social sciences would eventually be subsumed within
sociobiology further alienated social scientists who were already
hostile to evolutionary approaches to human behavior (see
Segerstråle, 2000). In contrast, other examples of human
sociobiology research have arguably enhanced our understanding of
human behavior. For example, Martin Daly and Margo Wilson
(1985) provided a plausible account of why the rates of child abuse
among families with step-parents were higher than among other
families, given that selection is likely to favor higher investment in
genetically related children (but see Buller 2006; Daly & Wilson,
2005), and Sarah Hrdy (1981, 1999; Figure 17.5) convincingly argued
that female primates, including human mothers, could be selected to
adjust levels of investment in offspring depending upon current
social relationships and the extent of available social support.

Figure 17.5 Sarah Hrdy (courtesy of Dan Hrdy)

Recent evolutionary perspectives on human behavior

Following the hostile reception to human sociobiology, a number of
sub-fields have emerged within the human behavioral evolutionary
sciences that build upon the work of Williams, Hamilton, Trivers and
others, yet typically reject the name “sociobiology” (Laland & Brown
2011). The term sociobiology has also fallen out of favor with
researchers studying animal behavior, and the seminal research of
Hamilton, Trivers, and others, such as David Lack, has instead given
rise to the field of animal behavioral ecology (Owens 2006), which
initially focused on the function of behavior in specific ecological
settings or life-history stages and has continued the ethological
tradition of studying animal behavior in the wild. During the 1970s,
anthropologists, such as Irven DeVore and William Irons, began to
apply the idea that selection should be viewed as mainly acting on
genes, rather than on groups or species, to human behavior and
emphasized that human behavior is highly flexible and variable
(Cronk 1991). These researchers started with the assumption that
human behavior varies adaptively in response to specific
environmental and social parameters in a manner that optimizes
lifetime reproductive success, where reproductive success is
estimated by recording the number of offspring or a proxy measure
such as rate of food acquisition. The predictions of these optimality
models were then tested against observable, quantitative data. While
much of the early research in this sub-field, which is now termed
Human Behavioural Ecology, was conducted on foraging behavior in
small-scale, hunter-gatherer societies, the topics and populations
under investigation have greatly expanded in recent years (Nettle et
al. 2013).
During the 1980s, the Human Behavioral Ecology approach received
stern criticism from psychologists who argued that researchers
should focus on the psychological mechanisms, rather than the
current function of behavioral responses, because they reasoned that
the brain is where adaptations will be found (Cosmides & Tooby
1987; Symons 1989). The emerging sub-field of Evolutionary
Psychology proposed that the human brain consists of numerous
evolved psychological mechanisms that have been selected for their
role in solving recurrent problems encountered by human ancestors,
with selection particularly favoring modules in the brain that solve
problems within specific domains, such as finding a mate, tracking
prey, and cooperating with others (Tooby & Cosmides 2005).
Evolutionary psychologists have argued that measuring the current
fitness of behavior is unlikely to reveal adaptive outcomes, as the
modern world differs so greatly from the environment that was
encountered by our human ancestors and in which our psychological
adaptations were thought to have evolved (Symons 1989). The
relevant “environment of evolutionary adaptedness” (EEA) for
understanding human psychological mechanisms was proposed to be
the African savanna environment that was inhabited by human
ancestors during the Pleistocene epoch (Tooby & Cosmides 1990). In
modern environments, human beings are therefore said to
experience an “adaptive lag,” i.e., a mismatch between current
selection pressures and a character, as a result of a mismatch
between our stone-aged mind and the current physical and social
environments (Li et al. 2018). The concept of the EEA received
strong criticism from archaeologists (e.g., Foley 1995), who pointed
out that a variety of hominins existed during the Pleistocene and
inhabited a broad range of physical environments across the globe,
and other researchers have emphasized that behavioral flexibility
potentially allows for optimal behavioral responses even in modern,
Westernized populations (Laland & Brown 2006; Brown et al. 2011).
The conflicts between the human behavioral ecologists and
evolutionary psychologists appeared to have been somewhat
reconciled by the idea that universal, evolved psychological
mechanisms, which are shared by all human beings, could produce
locally specific adaptive responses in some circumstances (Nettle
2009). The concept of behavioral flexibility was further incorporated
into the evolutionary psychology framework by considering
“condition-dependent strategies,” such that the behavioral responses
of individuals vary predictably in relation to the specific situation.
For example, evolutionary psychologists studying human mate
choice proposed that women prefer masculine men when choosing
short-term mates and nurturing men when choosing long-term
partners, as a result of context-specific strategies that are pre-
programmed into our evolved psychological mechanisms (e.g.,
Gangestad & Simpson 2000). Alternatively, individuals are proposed
to develop particular behavioral strategies as a result of
environmental inputs that act on evolved developmental systems
during early life (Gluckman et al. 2005). However, as discussed later,
other researchers have questioned whether the human mind consists
of pre-specified programs and have argued that brain development is
strongly influenced by socially transmitted cultural information
(Brown et al. 2009). From this perspective, development during
early life involves a bi-directional interaction between the developing
nervous system and the behavioral outputs of the brain, expressed
within human-made physical and social environments (Flynn et al.
2013).
Some evolutionary psychologists continued to stress the idea that
evolved psychological mechanisms have remained relatively
unchanged since the end of the Pleistocene, based on the assumption
that human minds are built from co-adapted gene complexes that are
unable to respond rapidly to selection (Tooby & Cosmides 2005).
The opportunity to test whether this assumption is reasonable was
provided by data from the Human Genome Project, which
successfully sequenced and mapped the genes of individual members
of our species. In contrast to the assumption made by evolutionary
psychologists, the genetic evidence shows that the human genome
has undergone substantial selection since the end of the Pleistocene
and that as much as 10% of our genes may have been subject to
strong, positive selection over the past 30 thousand years, with much
of the selection occurring since the start of the agricultural revolution
(approximately 12,000 years ago). The genes that have undergone
selection include those expressed in the nervous system, such as
genes involved in neurotransmission, brain energy production,
neuronal plasticity, and brain development (Laland et al. 2010;
Somel et al. 2013; Figure 17.6) which suggests that culturally induced
changes in our environment have influenced how our brains are
constructed and how our minds work (Richerson et al. 2010; Laland
2017). Selection experiments and observations of natural selection in
nonhuman animals have also confirmed that evolution can be
extremely fast, with significant genetic and phenotypic change
sometimes observed in a handful of generations (Kingsolver et al.
2001; Siepielski et al. 2011). Therefore, genetic evidence from both
human beings and other species is consistent with the notion that
human evolution has continued since the end of the Pleistocene.
Figure 17.6 The human archaeological record documents the
emergence of tools and cultural artifacts, and increasing cranial
volumes, during the evolution of the Homo genus. Enhanced
nutrition that resulted from tool use for hunting, scavenging, and
digging tubers, and the subsequent control of fire and cooking, is
thought to have promoted the increase in brain size. Comparative
genomics has identified a number of genes in the modern human
genome that could underlie these changes in brain volume and
behavioral traits; for instance, a deletion associated with the growth
arrest and DNA-damage-inducible gamma (GADD45G) gene
potentially underpinned brain size expansion, while changes in other
genes are postulated to have increased dendritic spine density in the
cortex (SLIT-ROBO Rho GTPase activating protein 2, SPGAP2, gene)
or have been linked to the emergence of language (forkhead box P2,
FOXP2, gene) (from Somel et al. 2013).
Many of the changes in our environment that have occurred since the
end of the Pleistocene have resulted from human cultural practices,
such as agriculture, urbanization, and medical advances, and these
cultural practices have, themselves, changed over time. In The Selfish
Gene, Richard Dawkins (1976) introduced the concept of the
“meme,” which he defined as a cultural replicator that could be
thought of as infecting human minds, or “hosts,” and as being
susceptible to similar evolutionary processes as genes. While the
“meme” meme has waned in popularity (see Box 17.2), the idea that
changes in the frequencies of cultural traits over time can be
modeled using the principles of evolutionary theory developed into a
mathematically sophisticated field of research (Richerson & Boyd
2005; Henrich 2016; Boyd 2017). This sub-field, known as cultural
evolution, developed the idea that, because cultural traits exhibit
“variation” and “competition,” and are “inherited” both between and
within generations, evolutionary modeling can be applied to cultural
traits, even though the exact mechanisms of selection or inheritance
may vary to some degree from those involved in genetic evolution
(Mesoudi 2011). Cultural evolution models have been used to predict
the spread of cultural traits, including maladaptive traditions, and
have also examined the interactions between cultural and genetic
inheritance systems, known as gene-culture coevolutionary
processes. In the following section, we examine whether the extent of
human culture, and the cognitive processes that underpin
information transmission in our species, requires a new framework
for applying evolutionary theory to human behavior.
Box 17.2 The rise and fall of memetics

The idea that units of cultural information might compete with

each other to be transmitted between minds and hence flourish is
not new. Darwin (1871) himself wrote, with respect to language
evolution, that a “struggle for life is constantly going on amongst
the words and grammatical forms in each language. The better,
the shorter, the easier forms are constantly gaining the upper
hand, and they owe their success to their own inherent virtue.”
Dawkins, in his book The Selfish Gene (1976), brought Darwin’s
idea into the mainstream by coining the term “meme.” A
successful meme—something that can be imitated, such as an
idea, or a word, or a dance move—has the essential properties of a
target of natural selection: it has longevity (we can retain it in our
minds for long periods), it has fecundity (it is easy to copy, and
spreads rapidly), and it has fidelity (transmission, when it occurs,
is accurate). Dawkins argued that the imitative ability of humans
made us a fertile breeding ground for “memes”: they began to
hijack our “lumbering robot” minds, invading us like a mind-
virus. The Selfish Gene remains one of the most popular science
books of the twentieth century, and the idea of “memetics”—the
systematic study of memes and their transmission—spawned a
number of highly successful books (e.g., Dennett 1995; Blackmore
1999) and even a Journal of Memetics. However, the endeavor
stalled; the journal closed down in 2005 and, in academic circles,
interest in memes appeared to wane, most likely because of a
failure to translate the early meme concept into an empirical
science that provided novel predictions and tested hypotheses
(although the concept remains highly prevalent in popular
culture, where the term designates an idea or image that has
“gone viral”). It is also clear that not all socially transmitted
knowledge operates like a selfish element, or virus, probably
because human minds have been subject to selection to filter out
maladaptive knowledge. Nonetheless, the study of cultural
evolution has flourished, with a new Cultural Evolution Society
founded in 2015.
Are Human Beings Different from Other Animals?
One major theme in the criticism leveled at human sociobiology and
evolutionary psychology is the claim that different rules apply to
human beings compared to other animals. While there are many
variants on this premise, which typically place emphasis on one or
more uniquely human capability, such as our language, culture, or
intelligence, the basic argument is that humans are different or
special in some important regard. Of course, all species are unique to
some degree, but the claim here is stronger than simply stating that
human beings are unique: it is that we are so different that our
behavior must be understood using quite distinct principles. In this
section, we consider this possibility more closely, focusing on (i) the
reliance of human beings on culture and the manner in which this
reliance allows us to modify our selective environments, and (ii) the
fidelity, efficiency, and breadth of human information transfer,
which is uniquely enhanced by language and teaching.

Human reliance on culture and its expression in niche

construction
While other animals are capable of learning from others and
exhibiting localized behavioral traditions, human beings engage in
much higher levels of social learning than any other species and
exhibit cumulative culture, defined as cultural information that is
accumulated over time (Dean et al., 2014; Mesoudi & Thornton
2018). Cumulative culture impacts upon every aspect of our lives, as
exemplified by technologies and institutions that have followed an
iterative process of change, with each version being built upon
previous knowledge and traditions (Basalla 1988). This accumulated
knowledge is expressed in our tools, technology, and engineering,
which allow human beings to transform the world in which they live,
for example by building houses and roads, domesticating plants and
animals, and developing medicines and weapons. Some aspects of
cumulative culture, such as contraceptives and medical advances,
appear to have dampened out the effects of natural selection on our
species. Other aspects, such as agricultural practices, appear to have
created new demands on our physiology, for example on our
digestive and immune systems. For instance, culturally induced
changes in agricultural practices have selected for genes involved in
digestion and disease resistance (Laland et al. 2010; Voight et al.
2006). A well-documented example is the link between dairy farming
and the prevalence of genes that allow adults to digest lactose, a
sugar found in milk and milk products (Gerbault et al. 2011, 2013;
Figure 17.7). The physical environment and the social niche that
human beings have inhabited throughout their evolutionary history
have, therefore, largely been constructed by human activities.
Figure 17.7 A) In most mammals, the individuals lose the ability
to digest milk beyond weaning age, due to a reduction in the enzyme
lactase, which is responsible for converting lactose into glucose and
galactose. In human beings, around 35% of adults continue to
produce lactase and can therefore digest fresh milk, a trait known as
lactase persistence (LP). a) The distribution of LP varies across
human populations, showing highest prevalence in parts of Europe,
Africa, the Middle East, and south-west Asia (color = proportion of
population, black dot = data collection sites); the global distribution
of LP closely matches local cultural traditions of dairy farming and
pastoralism. b) Genetic studies have found a number of single
nucleotide polymorphisms (SNPs) that are associated with LP; for
instance, a nucleotide change in the –13910*T allele, which affects
activity of the lactase gene promotor, is likely to underpin LP in
European populations (and other SNPs have been associated with LP
in other areas of the world); the date estimate for the emergence of
this SNP matches with the archaeological evidence of the spread of
animal domestication and dairy farming into northwest Europe,
supporting the hypothesis that cultural practices selected for genetic
variants that allowed individuals to access this source of energy into
adulthood (color = proportion of population, black dot = data
collection sites) (from Gerbault et al. 2011).
While all living organisms engage in niche construction, for instance,
by building nests, storing food and modifying soil chemistry, human
niche construction is more potent than that of any other species on
the planet (Odling-Smee et al. 2003; Smith 2007). Cultural niche
construction, in particular, has driven dramatic changes in our
environment and has modified subsequent selection pressures,
thereby impacting on the direction and rate of human evolution
(Laland et al. 2010). In contrast to the evolutionary psychology
perspective that focuses on the Pleistocene environment and
assumes large-scale adaptive lag, this niche construction perspective
highlights the fact that human beings are flexible, active constructors
of their niches. In most instances, niche-constructing activities are
expected to result in an adaptive fit between the organism and the
environment, as selection will have favored organisms that perturb
their environments in a manner than enhances fitness (Odling-Smee
et al. 2003). For example, termites build complex nest structures
that dampen environmental fluctuations in temperature and
humidity, thus constructing an environment that reduces the
variability in these parameters and preserves the adaptiveness of
previously selected traits that are temperature- and humidity-
sensitive. Similarly, during human evolution, activities such as
manufacturing clothes, building shelters, and storing food have
allowed individuals to reduce variability in their environment and
survive in conditions that would otherwise have been outside of their
physiological capability, and modern technologies, such as
refrigeration and air conditioning, have further extended the ability
to construct environments that suit our existing adaptations. Thus,
while modern environments are certainly different from ancestral
conditions, the extent of adaptive lag is likely to have been
significantly reduced by niche construction.
Human beings also exhibit an ability to deplete resources, degrade
environments, and devastate biodiversity. In the short term, the
human populations that create these effects might benefit from such
activities, in terms of having new sources of food or fuel. However,
unlike most animal species, human beings can respond to earlier
niche construction activities through further cultural developments.
For instance, pollution of the environment or overuse of water
resources can stimulate the invention and spread of new
technologies that reduce contamination or provide access to new
water sources, or the emergence of epidemics can lead to the
development of new vaccines. One consequence of cultural niche
construction is that genetic variants that have negative effects on
human fitness have potentially increased in frequency, as cultural
activities can cancel out the negative effects. For example, genetic
variation that underpins short sightedness might have been strongly
selected against in ancestral environments but is potentially
unrelated to fitness in populations with access to prescription
eyeglasses. Theoretical analyses suggest that cultural responses to
modified selection pressures (e.g., glasses, corrective eye surgery)
typically occur more rapidly than genetic responses do and will often
render genetic responses unnecessary (Laland et al. 2001). While the
capacity of human populations to construct solutions to self-imposed
problems is not always successful, this capability leaves human niche
construction uniquely potent and fast-acting.
A niche construction perspective demonstrates that the adaptive lag
in human behavior might not be as large as suggested by
evolutionary psychologists (Laland & Brown 2006). Evolution is a
response to changed selection pressures, and that response cannot
be instantaneous. Consequently, it is a truism that all organisms
must experience some adaptive lag, and, in this respect, humans are
not unique. Even the most adaptable of creatures will experience
limits to its tolerance space, outside of which it is unable to behave
adaptively. However, the idea that modern humans experience an
atypically large adaptive lag does not appear to be upheld; to the
contrary, the plasticity in behavioral responses that is made possible
by cultural niche construction may have left human beings with
unusually small levels of adaptive lag. This conclusion does not
preclude the possibility that certain human characteristics, such as
our taste for salt and sugar-rich foods, might be maladaptive in the
modern world. Instead, acknowledging the role of culture in human
evolution should enhance our understanding of how interactions
between genes and culture have impacted, and continue to alter, the
direction of human evolution. The ability of human beings to
construct their environments can influence the trajectory and rate of
evolution, meaning that, while evolutionary theory is still highly
relevant to understanding human behavior, the complexity of human
niche construction activities leads to feedback loops and interactions
that must themselves be incorporated into evolutionary theory.

Cognitive processes underlying information

transmission
The fact that human beings exhibit specialized cognitive processes
that underpin the effective transmission of cultural knowledge, such
as language and teaching, could also be said to have changed the
dynamics of human evolution (Brown & Richerson 2014; Laland
2017). Many cognitive traits that were once considered unique to
human beings, such as tool use, have fallen by the wayside with
reports, for example, of tool use in chimpanzees, birds, and dolphins
(Seed & Byrne 2010). Still, a large body of evidence shows that
human cognition differs from that of other animals in important
respects. For example, a study that compared the behavior of human
children, chimpanzees, and orangutans on a battery of cognitive tests
revealed that human toddlers outperformed adults of the other ape
species (Herrmann et al. 2007). Even at such a young age, human
children already had comparable cognitive skills to adult
chimpanzees and orangutans for dealing with the physical world
(e.g., spatial memory, object rotation) and had far more
sophisticated cognitive skills than both adult chimpanzees and
orangutans within the social realm (e.g., learning from others,
understanding intentions) (Herrmann et al. 2007). The differences
in the mental capabilities of humans and other animals could change
how researchers approach the study of human behavior. Here, we
focus on language and teaching, as these cognitive processes
potentially enhance the fidelity (accuracy) with which information is
transferred between individuals, increase the efficiency of knowledge
acquisition, in terms of the time taken for the naïve individual to gain
the information, and expand the breadth of information that can be
transferred, allowing for knowledge about past time periods and
distant places to be communicated between individuals.
Communication is perhaps the most obvious respect in which human
abilities appear to be qualitatively different from those of nonhuman
animals. Human languages differ from other animal communication
systems in that languages employ grammatical and semantic
categories, such as nouns and verbs in present, past, and future
tenses, to express exceedingly complex meanings. Even among
nonhuman apes that have been trained to comprehend the meaning
of a large numbers of words and symbols, none of these apes
acquired anything resembling the complex grammar that is
characteristic of human language (Fitch 2010). While some studies
have provided evidence that primate alarm calls refer to specific
categories of predators (e.g., Seyfarth et al. 1980; Zuberbühler
2000), vocal and visual signals in animal communication systems
generally relate to the immediate circumstances, and compelling
evidence that primate vocalizations are composed into complex,
semantic structures remains controversial (Wheeler & Fischer 2012).
In contrast, human language is entirely open-ended, allowing
humans to produce an infinite set of utterances and to create entirely
new sentences through the use of symbols. Language has been
argued to structure our thought processes, support complex social
interactions, and enhance the transmission of knowledge between
individuals (Bickerton 2010). By organizing our thoughts, human
language allows for rich forms of cognition that could not otherwise
occur, increasing by orders of magnitude the scale and accuracy of
information exchange relative to other animals, and perhaps
allowing for forms of cooperation that would be infeasible in other
species.
All human societies are dependent on an extensive series of
cooperative acts, often among unrelated individuals. In recent years,
human cooperation has been subject to intense investigation using
economic games (Fehr & Rockenbach 2004). One such game is
called the “ultimatum game,” where two players must decide how to
split a sum of money. The first player proposes how to divide the
sum among them, and the second player can either accept or reject
this proposal. If the second player accepts, the money is split
according to the proposal, but if the second player rejects, neither
player receives anything. Although the first player should be
expected to make a low offer, and the second player should rationally
accept any offer above zero, the empirical evidence suggests that
players frequently make generous offers and that players often reject
small offers (e.g., Roth et al. 1991). Moreover, the magnitude of offers
and rejection rates vary from one human society to the next in a
manner consistent with local cultural norms. For instance,
particularly generous offers are observed in societies with a culture
of gift giving (Henrich et al. 2001). Humans seem predisposed
toward cooperative interactions, often motivated by a sense of
fairness and consideration of other’s perspective, and frequently
adhere to the conventions in their society. In contrast, chimpanzees
and other primates respond to similar experimental tasks by taking
the highest reward, regardless of whether other individuals could
also have received a reward (Jensen et al. 2007), which suggests that
hominins, in particular, have been subject to selection promoting
other-regarding preferences and sensitivity to local norms of fairness
(Richerson & Boyd 2005).
Box 17.3 4Animal traditions

Many animals, including mammals, birds, fishes, and even

insects, acquire knowledge and skills from others (Hoppitt &
Laland 2013). Animals learn valuable skills and knowledge via
social learning, like what to eat or what a predator looks like, by
copying other animals. Social learning also is thought to lead to
variation in behavioral repertoires between and within regions, in
a manner that is not easily explained by ecological or genetic
variation and is often described as “animal culture” (e.g., Whiten
et al. 1999; Rendell & Whitehead 2001). Some animals, including
chimpanzees, orangutans, capuchins, whales, and dolphins,
appear to have an unusually broad cultural repertoire, with
multiple diverse traditions and distinctive behavioral profiles in
each community (Fragaszy & Perry 2003; Hoppitt & Laland 2013;
Whiten et al. 2017). Experimental studies of chimpanzees provide
strong evidence for imitation and other social learning
mechanisms that could underpin local behavioral traditions
(Whiten 2017; Figure 17.8). However, evidence for teaching in
chimpanzees is very limited (Rapaport & Brown 2008). Young
chimpanzees appear to learn skills such as termite fishing and
nut-cracking from observing others, but knowledgeable
individuals do not appear to aid their learning, for example by
sharing tools or by slowing down their foraging efforts in the
presence of naïve individuals.
Figure 17.8 In an experimental study, the efficacy of
transmission of stone tool-making skills along a chain of adult
participants was shown to be enhanced by teaching and,
particularly, language-based interactions (Morgan et al. 2015). a)
A skilled stone-knapper created flakes from a stone core using a
hammer stone, and the transmission of this skill to the first
participant, second participant, and so on (up to a maximum of
10 participants) took place using one of five different
transmission mechanisms: b) reverse engineering, in which the
pupil only saw the flakes produced by the tutor; c)
imitation/emulation, in which the pupil observed the tutor
making a flake; d) basic teaching, in which the tutor could
manually shape the pupil’s grasp or change their own actions to
improve the pupil’s view; e) gestural teaching, in which the pupil
and tutor could interact using gestures; or f) verbal teaching, in
which the pupil and tutor were allowed to speak. Flake quality,
and the number of viable flakes produced, showed greater
improvement across chains with gestural or verbal teaching
compared to the other chains (from Morgan et al. 2015).
Teaching in animals requires that: (i) an individual (X) with some
specific knowledge or skill modifies their behavior selectively in
the presence of an individual or individuals (Y) without that
knowledge or skill; (ii) this modification of behavior is costly to
individual X (or, at least, has no immediate benefit); and (iii)
individual Y acquires the knowledge or skill more rapidly as a
result of X’s behavior change than Y otherwise would have done
(Caro & Hauser 1992). In fact, the best evidence from teaching in
nonhuman animals comes from more distantly related species,
such as meerkats, pied babblers, and ants (Hoppitt et al. 2008).
For example, when young meerkats are learning to catch
scorpions, adults first provide youngsters with scorpions from
which they have removed the sting (Thornton & McAuliffe 2006).
Given that the potential costs of individual, trial-and-error
learning are high in this example, teaching potentially offers the
advantage of transmission of knowledge that is costly to learn
otherwise. Humans are unusual in being able to teach a broad
range of knowledge, and theoretical and empirical studies imply
that this is because our capacity for teaching and cumulative
 culture coevolved (Fogarty et al. 2011; Dean et al. 2012). For
example, experimental evidence suggests that teaching enhances
the transmission of skills required for knapping stone tools, such
as those made by our hominin ancestors (Morgan et al. 2015).

Human beings are also highly proficient at passing on skills to other

individuals through teaching, broadly defined as behavior that
functions to enhance the fidelity of information transmission. While
nonhuman animals are able to learn from others, exhibit local
behavioral traditions, and show evidence of teaching (Hoppitt et al.
2008; see Box 17.3), human beings engage in extensive social
learning and complex, proactive teaching, both through
demonstration and through scaffolding of the learner’s behavior.
Experimental studies have supported the hypothesis that, when the
fidelity of information transfer is high, culture is more likely to
exhibit cumulative properties (Tomasello 1994; Lewis & Laland
2012). For example, Dean and colleagues (2012) presented small
groups of nursery-school children, chimpanzees, and capuchin
monkeys with a “cumulative culture” task, comprising a puzzle box
that could be solved at three different levels to yield ever more
desirable rewards. As expected, the children, but not the other two
species, showed clear evidence of cumulative cultural learning, with
multiple children in multiple groups solving the task at the highest
level. The success of the children in reaching higher-level solutions
was strongly associated with a package of sociocognitive processes,
including teaching through verbal instruction, imitation, and
prosociality, that were observed only in the children. Cumulative
culture, teaching, and language are likely to coevolve, since each
reinforces the other (Fogarty et al. 2011; Morgan et al. 2015; Laland
2017; Figure 17.9). Such studies suggest that, while evolutionary
theory can be applied to human behavior, the complex cognitive
processes underlying information transmission in human beings
alter the dynamics of the evolutionary process in our species.
Figure 17.9 Capuchin monkeys (Cebus apella) can learn how to
open an artificial foraging device by observing the actions of a social
group member (Dindo et al. 2009). In each of two social groups, an
adult male capuchin monkey was trained to open the foraging device
using either a) a lift technique or b) a slide technique, and, when all
members of the social groups were allowed to interact with the
apparatus, c) all of the other capuchins in the two social groups were
found to preferentially adopt the technique used by the trained
subject, despite the apparatus functioning both ways during this
phase (each box represents the proportion of lift (black) and slide
(hashed) methods used by each subject during the 7 testing days,
with subjects added to the chart on successfully gaining food from
the apparatus) (from Whiten & Van De Waal 2017).

Implications for the Study of Behavior

So far, we have argued that evolutionary principles can be used to

investigate human behavior, but that different or refined
evolutionary methodologies are required to understand humans,
compared to all, or most, other animal species. However, humans are
not the only species to “break the rules.” To the contrary, it is now
well recognized that other species, such as chimpanzees, also possess
a second general inheritance system comprised of socially
transmitted knowledge and skills (Whiten et al. 1999; Whiten 2005),
and that gene-culture coevolution may be a process that has operated
in several animal taxa, albeit probably to a lesser degree than in
hominins (Avital & Jablonka 2000; Whitehead & Rendell 2015). We
ask next whether there are reasons to think that cognitive differences
between humans and other animals should influence how
researchers approach the study of human behavior. We specify four
ways in which the distinctive aspects of human cognition make the
study of human behavior more challenging than the study of other
animals: (i) accurate predictions of human behavior require
knowledge of human culture and history, (ii) reproductive success
will not always be a useful measure of evolutionary fitness in
humans, (iii) advanced human cognitive capacities, such as language
and theory of mind, allow for extensive deception and dishonest
signaling, and (iv) novel forms of cooperation can result from our
ability to engage in cultural learning, teaching, language, and
reasoning.
Accurate prediction of human behavior requires knowledge of
human culture and history. The breadth and depth of human
teaching and social transmission, largely mediated by the uniquely
cheap, flexible, and effective communication system that human
beings possess, leaves the behavior of humans far more dependent
on social learning, and social information more generally, than all
other animals. Language allows precise specification of laws, rules,
codes, conventions, institutions, and morals, which hugely shape
human behavior in a manner that diverges from all other animals.
This observation does not mean that evolved predispositions exert
no influence on human behavior, but it almost certainly means that it
will be more difficult to predict human behavior accurately solely
using knowledge of such predispositions. Human culture may be
broadly adaptive, but compliance with local laws, traditions, and
conventions will lead human behavior to approach local, rather than
global, optima. Accurate prediction of which equilibria will be
reached can only be achieved with knowledge of the history of
behavior in the local environment (Richerson & Boyd 2005; Mathew
& Perreault 2016). Culture, combined with potent niche
construction, partly detaches human behavior from the ecological
environment. As discussed above, the complexity of human niche
construction, and the ensuring interactions between genetic and
cultural inheritance systems, therefore must be incorporated into
evolutionary accounts of human behavior.
Reproductive success will not always be a useful fitness currency in
humans. While niche construction likely reduces adaptive lag in
humans, maladaptive traditions can spread in human populations
(Durham 1991; Richerson & Boyd 2005). Such maladaptive
traditions, which impact negatively on reproductive success, can
spread in circumstances where the cultural trait has a high
transmission rate and is spread effectively, despite negative impacts
on individual health and long-term survival (e.g., the spread of
tobacco smoking), or where language and culture allows human
populations to misattribute the negative consequences of behavior;
for instance, the deadly disease Kuru (a neurological illness that is
transmitted through consumption of infected tissue) was attributed
to witchcraft rather than cannibalism by New Guinea tribes (Durham
1991). In these instances of maladaptive behavior, an understanding
of the spread of the behavioral tradition and the impact on
reproductive success requires gene-culture coevolutionary analyses
that track how gene-culture combinations change frequency over
time, rather than analyses that focus on either genetic or cultural
inheritance of traits separately (Laland & Brown 2011). In addition,
the invention and use of contraception has partly detached sexual
behavior from reproduction in most human societies, requiring
human behavioral ecologists to rely on alternative fitness currencies.
The complexities of gene-culture coevolution thus again require
complex feedback loops and interactions to be incorporated into
evolutionary models of human behavior.
Language and theory of mind allow extensive human deception and
dishonest signaling. In other biological systems, most forms of
communication within species are thought to involve honest
signaling of information between individuals. The costs associated
with producing signals are considered sufficient to ensure that
signaling remains honest, and receivers are expected to respond to
signals as long as the information being transmitted is reliable on
average (Searcy & Nowicki 2005; see Chapter 14 ). Honest signaling
theory can also be used to make sense of aspects of human behavior:
for instance, human beings are thought to signal status through
conspicuous shows of wealth or risky activities such as hunting (e.g.,
Smith et al. 2003). However, human language constitutes a uniquely
cheap and flexible signaling device, allowing humans to engage in
“cheap talk” in an unprecedented range of circumstances (Dunbar
1997). Moreover, our capacity for theory of mind, and mental state
attribution, allows humans to anticipate the beliefs and intentions of
others and to exploit this knowledge to deceive others more
effectively (Tomasello 1999), leading to a well-established disconnect
between what people say and what they do. Our dependence on
social information may even have fashioned a highly gullible mindset
in humans (Richerson & Boyd 2005; Henrich 2016), which,
combined with mass media facilitated by technology and
hierarchically organized societies in which wealth and power are
unevenly distributed, leaves humans vulnerable to manipulation.
Levels of deception and dishonesty probably reach unparalleled
levels in our species and may frequently depart from equilibria
predicted by biological evolutionary models.
Novel forms of cooperation can result from human cognitive
capacities, including cultural learning, teaching, language, and
reasoning. Numerous researchers have argued that human culture
massively increases the scope of cooperation in our species (e.g.,
Richerson & Boyd 2005; Nowak & Highfield 2011; Pagel 2012;
Henrich 2016). Human language allows for trade, a form of
mutualistic exchange that outside of humans is virtually absent in
the animal kingdom (Ridley 2010). Norms, institutions, and laws
stabilize cooperative interactions between nonrelatives on huge
scales (Richerson & Boyd 2005), while language ensures that these
rules are specified in detail and widely known. In addition, imitation,
teaching, and institutionalized punishment ensure that differences in
behavior between groups are maintained despite dispersal of
individuals between groups, while reasoning and mental simulation
allow humans to devise complex sequences of goal-directed behavior
that involve future planning, often over periods of years. These
capacities allow an unusually stable form of group selection to arise,
known as “cultural group selection,” whereby culturally successful
groups can out-compete other cultural groups and proliferate by
eliminating other groups or absorbing them into their own traditions
and customs. An extensive body of formal models shows that cultural
group selection can generate large-scale cooperation (e.g., Boyd et al.
2003; Nowak et al. 2010). Our capacity for culture has therefore
apparently taken human populations down evolutionary pathways
not available to noncultural species, generating an evolved
psychology (including tribal “instincts,” docility, shame applied to
norms) that is entirely different from what can evolve through genes
alone (Richerson & Boyd 2005; Chudek & Henrich 2011; Henrich
2016). While evolutionary theory can be applied to human behavior,
provided that the models take cultural processes into account,
cultural group selection may be one respect in which humans are
distinct from other species.
Although human beings differ from other animals in terms of their
cultural niche construction activities and advanced cognitive
processes that enhance information transmission, evolutionary
theory can generally accommodate the complex interactions between
genetic and cultural inheritance systems that characterize human
evolutionary history. A fuller account of human behavior must take
into consideration the feedback loops that occur between human
niche construction activities and selection pressures acting on our
bodies and minds, and the fact that cultural traits, and cultural
groups, exhibit the basic properties required for natural selection to
ensue, namely variation, competition, and inheritance. The cognitive
processes underlying information transmission have had a dramatic
impact on the rate and direction of human evolution, including
altering the selection pressures feeding back on the human mind.
The emerging view of human evolution therefore involves complex,
reciprocal interactions between individuals and their environments.
Gaining a more complete understanding of human evolution has
implications for our understanding of animal behavior. More
specifically, the complex, reciprocal interactions between the
developing organism and the external, constructed environment
require researchers to reevaluate whether questions about behavior
should be separated into those that investigate “proximate”
mechanisms and those that investigate “ultimate” evolutionary
processes (as defined by Mayr 1961). The proximate mechanisms
themselves become part of the selective environment, and the
continuous selective feedback generated by the proximate
mechanisms therefore form a key component of the “ultimate”
causes of behavior (Laland et al. 2011).
 SUMMARY AND CONCLUSIONS
Darwin’s view that human behavior and cognition can be
understood through the application of evolutionary principles
holds true, as long as researchers take into account some
uniquely human attributes, such as our culture and the impact of
our niche construction. Traits that enhance the transmission of
social information in humans, including language and teaching,
have played an important role in the feedback loops between
culturally constructed environments and subsequent bouts of
selection on human beings. Genetic evidence confirms that our
species has continued to evolve since the end of the Pleistocene,
and the idea that evolutionary processes are too slow to have
causally influenced recent evolution of the human brain has been
refuted. Human cognitive capacities have themselves almost
certainly altered our species’ evolutionary trajectory, including
through modifying selection on the human brain. Evolutionary
theory can be broadened to incorporate the impact of human
culture, including the presence of maladaptive behavior, and
human cognition on the evolution of our species. The broader
evolutionary perspective needed to understand human behavior
will also provide a richer understanding of the behavior of
nonhuman animals, as behavior in all species involves many
reciprocal causal interactions between the developing organism
and the constructed environment.

FURTHER READING
Laland, K. N. & Brown, G. R. 2011. Sense and Nonsense:
Evolutionary Perspectives on Human Behaviour, 2nd ed.

REFERENCES
Archer, J. 2009. Does sexual selection explain human sex differences
in aggression? Behavioral and Brain Sciences, 32, 249–66.
Avital, E. & Jablonka, E. 2000. Animal Traditions: Behavioural
Inheritance in Evolution. Cambridge, UK: Cambridge University
Press.
Basalla, G. 1988. The Evolution of Technology. Cambridge, UK:
Cambridge University Press.
Bateman, A.J. 1948. Intra-sexual selection in Drosophila. Heredity,
2, 349–68.
Bickerton, D. 2010. Adam’s Tongue: How Humans Made Language,
How Language Made Humans. New York: Hill and Wang.
Bjorklund, D.F. & Kipp, K. 1996. Parental investment theory and
gender differences in the evolution of inhibition mechanisms.
Psychological Bulletin, 120, 163–88.
Blackmore, S. 1999. The Meme Machine. Oxford, UK: Oxford
University Press.
Blackwell, E.B. 1875. The Sexes throughout Nature. New York: G.P.
Putnam’s Sons.
Boakes, R. 1984. From Darwin to Behaviourism: Psychology and
the Minds of Animals. Cambridge, UK: Cambridge University
Press.
Boogert, N.J., Madden, J.R., Morand-Ferron, J. & Thornton, A.
2018. Measuring and understanding individual differences in
cognition. Philosophical Transactions of the Royal Society B, 373,
20170280.
Bowlby, J. 1969. Attachment and Loss: Volume 1 Attachment.
London: Hogarth Press.
Boyd, R. 2017. A Different Kind of Animal. Princeton, NJ: Princeton
University Press.
Boyd, R., Gintis, H., Bowles, S. & Richerson, P.J. 2003. The evolution
of altruistic punishment. Proceedings of the National Academy of
Sciences USA, 100, 3531–5.
Brown, G.R., Dickins, T., Sear, R. & Laland, K.N. 2011. Evolutionary
accounts of human behavioural diversity. Philosophical
Transactions of the Royal Society B, 366, 313–24.
Brown, G.R., Laland, K.N. & Borgerhoff Mulder, M. 2009. Bateman’s
principles and human sex roles. Trends in Ecology and Evolution,
24, 297–304.
Brown, G.R. & Richerson, P. 2014. Applying evolutionary theory to
human behaviour: past differences and current debates. Journal
of Bioeconomics, 16, 105–28.
Brüne, M. & Schiefenhövel, W. (eds.) 2019. The Oxford Handbook of
Evolutionary Medicine. Oxford: Oxford University Press.
Bshary, R., Gingins, S. & Vail, A.L. 2014. Social cognition in fishes.
Trends in Cognitive Sciences, 18, 465–71.
Buller, D.J. 2006. Adapting Minds: Evolutionary Psychology and
the Persistent Quest for Human Nature. Cambridge, MA: MIT
Press.
Buss, D.M. (ed.) 2015. The Handbook of Evolutionary Psychology,
2nd ed., 2 vols. New York: Wiley.
Campbell, A. 1999. Staying alive: Evolution, culture, and women’s
intrasexual aggression. Behavioral and Brain Sciences, 22, 203–
52.
Caro, T. M. & Hauser, M. D. 1992. Is there teaching in nonhuman
animals? Quarterly Review of Biology, 67, 151–174.
Chudek, M. & Henrich, J. 2011. Culture-gene coevolution, norm-
psychology and the emergence of human prosociality. Trends in
Cognitive Sciences, 15, 218–26.
Cosmides, L. & Tooby, J. 1987. From evolution to behavior:
evolutionary psychology as the missing link. In: J. Dupré (ed.),
The Latest on the Best: Essays on Evolution and Optimality.
Cambridge, MA: MIT Press.
Cronk, L. 1991. Human behavioral ecology. Annual Review of
Anthropology, 20, 25–53.
Cross, C.P., Copping, L.T. & Campbell, A. 2011. Sex differences in
impulsivity: a meta-analysis. Psychological Bulletin, 137, 97–130.
Cross, C.P., Cyrenne, D.L.M. & Brown, G.R. 2013. Sex differences in
sensation-seeking: a meta-analysis. Scientific Reports, 3, 2486.
Daly, M. & Wilson, M. 1985. Child abuse and other risks of not living
with both parents. Ethology and Sociobiology, 6, 197–210.
Daly, M. & Wilson, M. 2005. The ‘Cinderella effect' is no fairy tale.
Trends in Cognitive Sciences, 9, 507–508.
Damasio, A. & Carvalho, G.B. 2013. The nature of feelings:
evolutionary and neurobiological origins. Nature Reviews
Neuroscience, 14, 143–52.
Darwin, C. 1859. The Origin of Species by Means of Natural
Selection, or the Preservation of Favoured Races in the Struggle
for Life. London: John Murray. (1st edition reprint, London:
Penguin Books, 1968).
Darwin, C. 1871. The Descent of Man and Selection in Relation to
Sex. John Murray; London (1st edition reprint, Princeton NJ:
Princeton University Press, 1981).
Darwin, C. 1872. The Expression of the Emotions in Man and
Animals. London: John Murray. (3rd edition reprint, London:
Harper-Collins, 1998).
Dawkins, R. 1976. The Selfish Gene. Oxford: Oxford University Press.
Dean, L., Vale, G., Laland, K., Flynn, E. & Kendal, R. 2014. Human
cumulative culture: a comparative perspective. Biological
Reviews, 89, 284–301.
Dean, L.G., Kendal, R.L., Schapiro, S.J., Thierry, B. & Laland, K.N.
2012. Identification of the social and cognitive processes
underlying human cumulative culture. Science, 335, 1114–8.
Dennett, D. 1995. Darwin’s Dangerous Idea: Evolution and the
Meanings of Life. London: Penguin Books.
Desmond, A. & Moore, J. 2009. Darwin’s Sacred Cause: Race,
Slavery and the Quest for Human Origins. London: Penguin
Books.
Dindo, M., Whiten, A. & De Waal, F.B.M. 2009. In-group conformity
sustains different foraging traditions in capuchin monkeys (Cebus
apella). PLoS ONE, 4, e7858.
Dunbar, 1997. Grooming, Gossip, and the Evolution of Language.
Cambridge, MA: Harvard University Press.
Durham, W.H. 1991. Coevolution: Genes, Culture and Human
Diversity. Stanford, CA: Stanford University Press.
Durrant, R. & Ward, T. 2015. Evolutionary Criminology: Towards a
Comprehensive Explanation of Crime. London: Academic Press.
Ekman, P. 1994. Strong evidence for universals in facial expressions:
a reply to Russell’s mistaken critique. Psychological Bulletin, 115,
268–87.
Elfenbein, H.A. & Ambady, N. 2002. On the universality and cultural
specificity of emotion recognition: a meta-analysis. Psychological
Bulletin, 128, 203–35.
Ellegren, H. 2005. Natural selection in the evolution of humans and
chimps. Current Biology, 15, R919–922.
Emery, N.J. & Clayton, N.S. 2004. The mentality of crows:
convergent evolution of intelligence in corvids and apes. Science,
306, 1903–7.
Fehr, E. & Rockenbach, B. 2004. Human altruism: economic, neural,
and evolutionary perspectives. Current Opinion in Neurobiology,
14, 784–90.
Fitch, W.T. 2010. The Evolution of Language. Cambridge, UK:
Cambridge University Press.
Flynn, E.G., Laland, K.N., Kendal, R.L. & Kendal, J.R. 2013.
Developmental niche construction. Developmental Science, 16,
296–313.
Fogarty, L., Strimling, P. & Laland, K.N. 2011. The evolution of
teaching. Evolution, 65, 2760–70.
Foley, R. 1995. The adaptive legacy of human evolution: a search for
the environment of evolutionary adaptedness. Evolutionary
Anthropology, 6, 194–203.
Fragaszy, D.M. & Perry, S. (eds.) 2003. The Biology of Traditions:
Models and Evidence. Cambridge, UK: Cambridge University
Press.
Gangestad, S.W. & Simpson, J.A. 2000. The evolution of human
mating: trade-offs and strategic pluralism. Behavioral and Brain
Sciences, 23, 573–644.
Gerbault, P., Liebert, A., Itan, Y. et al. 2011. Evolution of lactase
persistence: an example of human niche construction.
hilosophical Transactions of the Royal Society B, 366, 863–77.
Gerbault, P., Roffet-Salque, M., Evershed, R.P. & Thomas, M.G.
2013. How long have adult humans been consuming milk?
IUBMB Life, 65, 983–90.
Gluckman, P.D., Hanson, M.A. & Spencer, H.G. 2005. Predictive
adaptive responses and human evolution. Trends in Ecology and
Evolution, 20, 527–33.
Goodman, M. 1999. The genomic record of humankind’s
evolutionary roots. American Journal of Human Genetics, 64, 31–
9.
Güntürkün, O. & Bugnyar, T. 2016. Cognition without cortex. Trends
in Cognitive Sciences, 20, 291–303.
amlin, K.A. 2014. From Eve to Evolution: Darwin, Science, and
Women’s Rights in Gilded Age America. Chicago: University of
Chicago Press.
Henrich, J. 2016. The Secret of Our Success: How Culture is Driving
Human Evolution, Domesticating our Species, and Making us
Smarter. Princeton, NJ: Princeton University Press.
Henrich, J., Boyd, R., Bowles, S. et al. 2001. In search of Homo
economicus: behavioral experiments in 15 small-scale societies.
American Economic Review, 91, 73–7.
Herrmann, E., Call, J., Hernàndez-Lloreda, M.V., Hare, B. &
Tomasello, M. 2007. Humans have evolved specialized skills of
social cognition: the cultural intelligence hypothesis. Science, 317,
1360–6.
Hinde, R.A. 1966. Animal Behaviour: a Synthesis of Ethology and
Comparative Psychology. New York: McGraw-Hill.
Hinde, R.A. 1974. Biological Bases of Human Social Behaviour. New
York: McGraw-Hill.
Hoppitt, W.J.E., Brown, G.R., Kendal, R. et al. 2008. Lessons from
animal teaching. Trends in Ecology and Evolution, 23, 486–93.
Hoppitt, W.J.E. & Laland, K.N. 2013. Social Learning: Mechanisms,
Methods and Models. Princeton, NJ: Princeton University Press.
Hrdy, S.B. 1981. The Woman that Never Evolved. Cambridge, MA:
Harvard University Press.
Hrdy, S.B. 1999. Mother Nature: Natural Selection and the Female
of the Species. London: Chatto and Windus.
Hull, D.L., Langman, R.E. & Glenn, S.S. 2001. A general account of
selection: biology, immunology, and behavior. Behavioral and
Brain Sciences, 24, 511–73.
Huxley, T.H. 1863. Evidence as to Man’s Place in Nature. London:
Williams and Norgate.
Ingold, T. 2007. The trouble with ‘evolutionary biology’.
Anthropology Today, 23, 13–7.
Jensen, K., Call, J. & Tomasello, M. 2007. Chimapnzees are rational
maximizers in an ultimatum game. Science, 318, 107–9.
Kingsolver, J.G., Hoekstra, H.E., Hoekstra, J.M. et al. 2001. The
strength of phenotypic selection in natural populations. American
Naturalist, 157, 245–61.
Kokko, H. & Monaghan, P. 2001. Predicting the direction of sexual
selection. Ecology Letters, 4, 159–65.
Kossinets, G. & Watts, D.J. 2006. Empirical analysis of an evolving
social network. Science, 311, 88–90.
Laland, K.N. 2017. Darwin’s Unfinished Symphony: How Culture
Made the Human Mind. Princeton, NJ: Princeton University
Press.
Laland, K.N. & Brown, G.R. 2006. Niche construction, human
behaviour and the adaptive-lag hypothesis. Evolutionary
Anthropology, 15, 95–104.
Laland, K.N., Odling-Smee, F.J. & Feldman, M.W. 2001. Cultural
niche construction and human evolution. Journal of Evolutionary
Biology, 14, 22–33.
Laland, K.N., Odling-Smee, F.J. & Myles, S. 2010. How culture has
shaped the human genome: bringing genetics and the human
sciences together. Nature Reviews: Genetics, 11, 137–48.
Laland, K.N., Sterelny, K., Odling-Smee, J., Hoppitt, W. & Uller, T.
2011. Cause and effect in biology revisited: Is Mayr’s proximate-
ultimate dichotomy still useful? Science, 334, 1512–6.
Lewis, H.M. & Laland, K.N. 2012. Transmission fidelity is the key to
the build-up of complex culture. Philosophical Transaction of the
Royal Society B, 367, 2171–80.
Li, N.P., Van Vugt, M. & Colarelli, S.M. 2018. The evolutionary
mismatch hypothesis: implications for psychological science.
Current Directions in Psychological Science, 27, 38–44.
Lorenz, K. 1935. Der Kumpan in der umwelt des vogels: der
artgenosse als auslösendes moment sozialer verhaltungsweisen.
[The companion in the bird’s world: fellow members of the
 species as releasers of social behaviour]. Journal für Ornithologie,
83, 137–213.
Lorenz, K. 1966. On Aggression. London: Methuen (reprinted in
1996 by Routledge).
Mathew, S. & Perreault, C. 2016. Cultural history, not ecological
environment, is the main determinant of human behavior.
Proceedings of the Royal Society B, 283, 1826.
Mayr, E. 1961. Cause and effect in biology. Science, 134, 1501–6.
Mesoudi, A. 2011. Cultural Evolution: How Darwinian Theory Can
Explain Human Culture and Synthesize the Social Sciences.
Chicago: Chicago University Press.
Mesoudi, A. & Thornton, A. 2018. What is cumulative cultural
evolution? Proceedings of the Royal Society B, 285, 20180712.
Monaghan, P. & Birkhead, T. 2001. Spot the difference. Trends in
Ecology and Evolution, 16, 527.
Morgan, T.J.H., Uomini, N.T., Rendell, L.E. et al. 2015. Experimental
evidence for the co-evolution of hominin tool-making, teaching
and language. Nature Communications, 6, 6029.
Morris, D. 1967. The Naked Ape. London: Vintage.
Nettle, D. 2009. Ecological influences on human behavioural
diversity: a review of recent findings. Trends in Ecology and
Evolution, 24, 618–24.
Nettle, D., Gibson, M.A., Lawson, D.W. & Sear, R. 2013. Human
behavioral ecology: current research and future prospects.
Behavioral Ecology, 24, 1031–1340.
Nowak, M. & Highfield, R. 2011. Supercooperators: Evolution,
Altruism and Human Behaviour or, Why We Need Each Other to
Succeed. New York: Canongate Books.
Nowak, M.A., Tarnita, C.E. & Wilson, E.O. 2010. The evolution of
eusociality. Nature, 466, 1057–62.
Odling-Smee, F.J., Laland, K.N. & Feldman, M.W. 2003. Niche
Construction: The Neglected Process in Evolution. Princeton, NJ:
Princeton University Press.
Owens, I.P.F. 2006. Where is behavioural ecology going? Trends in
Ecology and Evolution, 21, 356–61.
Pagel, M. 2012. Wired for Culture: The Natural History of Human
Cooperation. London: Allen Lang.
Pagel, M., Atkinson, W.D. & Meade, A. 2007. Frequency of word-use
predicts rates of lexical evolution throughout Indo-European
history. Nature, 449, 717–20.
Panksepp, J. 2011. The basic emotional circuits of mammalian
brains: do animals have affective lives? Neuroscience and
Biobehavioral Reviews, 35, 1791–1804.
Plotkin, H. 1994. Darwin Machines and the Nature of Knowledge.
Cambridge, MA: Harvard University Press.
Rapaport, L.M. & Brown, G.R. 2008. Social influences on foraging
behaviour in young non-human primates: learning what, where,
and how to eat. Evolutionary Anthropology, 17, 189–201.
Rendell, L. & Whitehead, H. 2001. Culture in whales and dolphins.
Behavioral and Brain Sciences, 24, 309–24.
Richerson, P.J. & Boyd, R. 2005. Not by Genes Alone: How Culture
Transformed Human Evolution. Chicago: Chicago University
Press.
Richerson, P.J., Boyd, R. & Henrich, J. 2010. Gene-culture
coevolution in the age of genomics. Proceedings of the National
Academy of Sciences USA, 107, 8985–92.
Ridley, M. 2010. The Rational Optimist: How Prosperity Evolves.
New York: Harper.
Rose, S., Lewontin, R.C. & Kamin, L.J. 1984. Not in Our Genes:
Biology, Ideology, and Human Nature. London: Penguin Books.
Roth, A.E., Prasnikar, V., Okuno-Fujiwara, M. & Zamir, S. 1991.
Bargaining and market behavior in Jerusalem, Ljubljana,
Pittsburgh, and Tokyo: an experimental study. American
Economic Review, 81, 1068–95.
Searcy, W.A. & Nowicki, S. 2005. The Evolution of Animal
Communication. Princeton, NJ: Princeton University Press.
Seed, A. & Byrne, R. 2010. Animal tool-use. Current Biology, 20,
R1032–9.
Seed, A.M. & Mayer, C. 2017 Problem-solving in animals. In J. Call,
G.M. Burghardt, I.M. Pepperberg, C.T. Snowdon & T.R. Zentall
(eds.), APA Handbook of Comparative Psychology, pp. 601–25.
Washington: American Psychological Association.
Segerstråle, U.S. 2000. Defenders of the Truth: The Sociobiology
Debate. Oxford: Oxford University Press
Seyfarth, R.M., Cheney, D.L. & Marler, P. 1980. Vervet monkey
alarm calls: semantic communication in a free-ranging primate.
Animal Behaviour, 28, 1070–94.
Siepielski, A.M., DiBattista, J.D., Evans, J.A. & Carlson, S.M. 2011.
Differences in the temporal dynamics of phenotypic selection
among fitness components in the wild. Proceedings of the Royal
Society of London B, 278, 1572–80.
Smith, B. 2007. Human niche construction and the behavioural
context of plant and animal domestication. Evolutionary
Anthropology, 16, 188–99.
Smith, E.A., Bliege Bird, R. & Bird, D.W. 2003. The benefits of costly
signaling: Meriam turtle hunters. Behavioral Ecology, 14, 116–26.
Somel, M., Liu, X. & Khaitovich, P. 2013. Human brain evolution:
transcripts, metabolites and their regulators. Nature Reviews
Neuroscience, 14, 112–27.
Spencer, H. 1855, 1870. Principles of Psychology. (1st edn., 2nd
edn.) London: Longman.
Symons, D. 1989. A critique of Darwinian anthropology. Ethology
and Sociobiology, 10, 131–44.
Thornton, A. & McAuliffe, K. 2006. Teaching in wild meerkats.
Nature, 313, 227–9.
Tinbergen, N. 1951. The Study of Instinct. Oxford: Oxford University
Press (reprinted 1989).
Tinbergen, N. & Kruyt, W. 1932. Über die orientierung des
bienenwolfes (Philanthus triangulum Fabr.). [On the orientation
of the bee wolf] Zeitschrift für vergleichende Physiologie, 25, 292–
344.
Tomasello, M. 1994. The question of chimpanzee culture. In: R.
Wrangham, W. McGrew, F. De Waal & P. Heltne (eds.),
Chimpanzee Cultures, pp. 301–17. Cambridge, MA: Harvard
University Press.
Tomasello, M. 1999. The Cultural Origins of Human Cognitions.
Cambridge, MA: Harvard University Press.
Tomasello, M. & Call, J. 1997. Primate Cognition. Oxford: Oxford
University Press.
Tooby, J. & Cosmides, L. 1990. The past explains the present:
emotional adaptations and the structure of ancestral
environments. Ethology and Sociobiology, 11, 375–424.
Tooby, J. & Cosmides, L. 2005. Conceptual foundations of
evolutionary psychology. In: D. Buss. (ed.), The Handbook of
Evolutionary Psychology, pp. 5–67. Hoboken, NJ: Wiley.
Toren, C. 2018. Human ontogenies as historical processes: lessons
from ethnography. In: E. Hannon & T. Lewens (eds.), Why we
Disagree about Human Nature, pp. 170–85. Oxford: Oxford
University Press.
Trivers, R.L. 1972. Parental investment and sexual selection. In: B.
Campbell (ed.), Sexual Selection and the Descent of Man, 1871–
1971, pp. 136–70. Chicago: Aldine.
Voight, B. F., Kudaravalli, S., Wen, X. & Pritchard, J. K. 2006. A map
of recent positive selection in the human genome. PLoS Biology,
4, e72.
Watson, J.B. 1913. Psychology as the behaviorist views it.
Psychological Review, 20, 158–77.
Watson, J.B. 1924. Behaviorism. New York: Norton.
heeler, B.C. & Fischer, J. 2012. Functionally referential signals: a
promising paradigm whose time has passed. Evolutionary
Anthropology, 21, 195–205.
Whitehead, H. & Rendell, L.R. 2015. The Cultural Lives of Whales
and Dolphins. Chicago: University of Chicago Press.
Whiten, A. 2005. The second inheritance system of chimpanzees and
humans. Nature, 437, 52–5.
Whiten, A. 2017. Culture extends the scope of evolutionary biology in
the great apes. Proceedings of the National Academy of Science
USA, 114, 7790–7.
Whiten, A., Ayala, F., Feldman, M.W. & Laland, K.N. 2017. The
extension of biology through culture. Proceedings of the National
Academy of Science USA, 114, 7775–81.
Whiten, A., Goodall, J., McGrew, W.C. et al. 1999. Culture in
chimpanzees. Nature, 399, 682–5.
Whiten, A. & Van De Waal, E. 2017. Social learning, culture and the
‘socio-cultural brain’ of human and non-human primates.
Neuroscience and Biobehavioral Reviews, 82, 58–75.
Williams, G.C. 1966. Adaptation and Natural Selection. Princeton,
NJ: Princeton University Press.
Wilson, D.S. & Gowdy, J.M. 2013. Evolution as a general theoretical
framework for economics and public policy. Journal of Economic
Behavior and Organization 90S, S3–S10.
Wilson, E.O. 1975. Sociobiology: the New Synthesis. Cambridge,
MA: Harvard University Press.
Zuberbühler, K. 2000. Referential labelling in Diana monkeys.
Animal Behaviour, 59, 917–27.
name index
Adams, C.D., 205
Adams, G.J., 255
Adams, M.J., 352
Adli, M., 152
Adrian, E.D., 58
Agate, R.J., 151
Aiello, L., 440
Ainsworth, M.D.S., 179
Akçay, C., 380, 389
Akre, K.L., 420–421
Alatalo, R.V., 327, 331
Alem, S., 240
Alexander, B., 163
Alexander, J., 332
Alexander, R.D., 321
Allee, W.C., 254
Altman, J., 126
Alvarez-Buylla, A., 128
Ambady, N., 462
Anderson, R.C., 380
Andersson, M., 119
Aplin, L.M., 358
Appleby, M.C., 254, 276
Arak, A., 303, 321
Archer, J., 461
Archibald, J.K., 403
Armitage, K.B., 346, 358
Arnold, A.P., 151
Arnqvist, G., 318, 330
Ashida, G., 132
Atz, J.W., 402
Aubert, M., 451
Avargues-Weber, A., 123F, 233
Avital, E., 479

Babb, S.J., 215

Bachtrog, D., 149, 161
Backwell, P.R.Y., 357
Baddeley, A.D., 212
Badyaev, A.V., 356
Baerends, G.P., 24, 50, 51, 53, 224, 224F, 227, 228, 229F, 235
Bakker, J., 145
Bakker, T.C.M., 323, 378
Baldwin, M.W., 410, 411F
Ball, G.F., 150, 155, 158, 163
Balsalobre, A., 93
Balthazart, J., 128, 140, 155, 158, 163
Barber, J.R., 381
Barinaga, M., 39
Barlow, H.B., 31
Barnard, C.J., 303, 304
Barrett, L.F., 68–70
Basalla, G., 471
Basolo, A.L., 418
Bastien, B., 327
Bateman, A.J., 318, 461
Bateson, P., 173, 174
Bateson, P.P.G., 166, 170, 175, 177, 178, 186, 187
Battley, P.F., 237
Baugh, A.T., 349
Baum, M.J., 145
Beach, F.A., 57
Beauchamp, G., 305, 306
Becker, J.B., 146
Beckers, G.J.L., 389
Beecher, M.D., 386
Beersma, D.G., 94
Bekoff, M., 346
Bell, A.M., 345F, 351, 352, 357, 359
Bensch, S., 324
Benus, R.F., 170
Berglund, A., 318
Bergmuller, R., 349, 358
Bermúdez De Castro, J.M., 443
Berna, F., 441
Bernal, X.E., 383
Berridge, K.C., 73
Berry, C.S., 239
Berthold, P., 238
Berwick, R.C., 180, 182, 387
Bessa-Gomes, C., 318
Betzig, L.L., 336
Bickerton, D., 475
Bijleveld, A.I., 310
Bilezikjian, L.M., 125
Birkhead, T., 457
Biro, P.A., 360
Bischof, H.-J., 178
Bissonette, G.B., 245
Bjorklund, D.F., 461
Blackmore, S., 470
Blackwell, E.B., 460
Blair, W.F., 402
Blass, E.M., 58
Bliss, T.V.P., 124
Bloch, G., 85
Blough, D.S., 209
Boake, C.R., 350
Boakes, R., 462
Boakes, R.A., 9
Boissy, A., 271
Bolhuis, J.J., 1, 2, 4–6, 8, 9, 126, 175–176, 178, 180–183, 184F,
185–187, 189, 219
Bolles, R.C., 219
Bols, R.J., 55, 67
Bonnot, N., 357
Boogert, N.J., 463F
Boon, A.K., 357
Boone, Y., 444
Borbély, A.A., 60
Bouzouggar, A., 448
Bower, T.G.R., 21, 23
Bowlby, J., 178, 179, 465
Boyd, R., 470, 476, 480–481
Boyer, N., 357
Bradbury, J.W., 133
Bradshaw, J.W.S., 276
Braithwaite, V., 8
Brambell, F.W.R., 252, 253
Brandstatter, R., 101
Brenowitz, E.A., 126, 128, 386
Broadbent, D.E., 4
Brockmann, H.J., 303
Brodie, E.D., 381
Bro-Jørgensen, J., 382
Broom, D.M., 254, 276
Brown, G.R., 9, 461, 467–468, 474, 476, 480
Brown, K.S., 449
Brüne, M., 457
Bruner, J.S., 233
Brzoska, J., 36
Bshary, R., 463
Budaev, S.V., 358
Bugnyar, T., 459
Bugynar, T., 244
Buhr, E.D., 99
Buhusi, C.V., 60
Bull, J.J., 416
Buller, D.J., 457, 467
Buntin, J.D., 58
Burford, J.E., 378
Burghagen, H., 19, 31
Burkhardt, R.W., 271, 272
Burley, N., 329, 335, 335F
Buss, D.M., 457
Byrne, R., 474

Cade, W.H., 303, 315

Caldwell, R.L., 382
Call, J., 459
Campbell, A., 461
Candolin, U., 334
Cannon, W.B., 69
Capranica, R.R., 22
Caraco, T., 285, 305, 308, 310
Carere, C., 349, 356, 358
Carew, T.J., 26
Caro, T. M., 476
Carr, C.E., 115, 117, 132
Carskadon, M.A., 93
Cartwright, B.A., 235
Carvalho, G.B., 462
Cassidy, A.L.E.V., 386
Cassone, V.M., 101
Castillo-Ruiz, A., 96
Catchpole, C.K., 126, 180
Celli, M.L., 261
Chain, D.G., 125
Champagne, F.A., 359
Chan, E.K.F., 449
Charlton, B.D., 374
Chen, X., 151
Cheney, D.L., 242
Cheng, H.-W., 266
Cheng, K., 235
Chittka, L., 233, 239, 246
Chitty, D., 346
Chomsky, N., 4
Christensen, B., 285, 290
Chudek, M., 481
Ciuti, S., 357
Clark, A.B., 346
Clayton, D.F., 189
Clayton, N.S., 215F, 232, 463
Clobert, J., 356, 357
Clutton-Brock, T.H., 318
Cohen, J.D., 38, 429
Cohen, L.G., 40
Coleman, G.J., 259
Collett, M., 235, 236
Collett, T.S., 35, 235, 236
Conner, J., 321, 322F
Conner, W.E., 381
Cook, R.G., 214
Cooper, I.A., 405, 406F
Coppens, C.M., 349
Coppola, M., 450
Corballis, M.C., 388
Cornil, C.A., 144
Cosmides, L., 467, 468
Costa, P.T. Jr., 352
Cote, J., 356–358
Cott, H.B., 36
Craig, J.V., 253
Craig, K.D., 268F
Craig, M., 211
Crespi, B.J., 371
Crews, D., 161
Croft, D.P., 358
Cronk, L., 467
Cross, C.P., 461
Cross, E.R., 242, 243F, 244
Crystal, J.D., 215
Cuervo, J.J., 328
Cumming, J.M., 401
Cummings, M.E., 419
Czeisler, C.A., 80, 93

Daan, S., 84, 86, 93, 105

Dabelsteen, T., 381, 383
Dall, S.R.X., 343, 347
Dalla Costa, E., 268
Dalton, C., 253
Daly, M., 331, 467
Damasio, A.R., 37, 462
Danilenko, K.V., 104
Daros, R.R., 270F, 271
Darwin, C., 2, 7, 71, 236, 271, 346, 397–399, 399F, 456, 458–
462, 470, 482
Davidson, A.J., 85
Davidson, R.J., 68
Davies, N.B., 7, 9
Dawkins, M.S., 73, 254, 263, 272, 380
Dawkins, R., 400, 416, 470
Day, J.J., 407
Dean, A.M., 408
Dean, L., 471, 477, 479
Dean, M.C., 438
DeCoursey, P.J., 88, 93
Deisseroth, K., 130
De Lope, F., 329, 332F
De Lumley, H., 444
Dennett, D.C., 272, 470
De Passillé, A.M., 258
De Queiroz, A., 402
d’Errico, F., 445, 448
Desmond, A., 462
DeVoogd, T.J., 180, 189
Deweer, B., 217
Dewsbury, D., 9
Di Bitetti, M.S., 390
Dicke, U., 428F, 429
Dickinson, A., 205, 215F, 232
Diedrichsen, J., 227
Dijk, D. J., 92, 93
Dindo, M., 478F
Dingemanse, N.J., 344F, 347–348, 350, 352, 356
Dochtermann, N.A., 350
Doetsch, F., 129
Doupe, A.J., 180, 181
Drent, P.J., 347
Drent, R.H., 24
Drickamer, L.C., 9, 402
Duckworth, R.A., 356
Duff, S.J., 235
Dumont, F., 346
Dunbar, 481
Duncan, C.P., 216, 216F
Duncan, I.J.H., 254, 266–267
Durham, W.H., 480
Durrant, R., 457

Eaton, R.C., 27
Eberhard, W.G., 330
Edgar, R.S., 85
Egnor, S.E.R., 386
Ehlinger, T.J., 346
Eilam, D., 118
Ekman, P., 68, 462
Elbert, T., 40
Eldredge, N., 451
Elfenbein, H.A., 462
Ellegren, H., 459
Emery, N.J., 463
Emlen, S.T., 238, 318
Enquist, M., 74, 371, 372F, 375, 380
Etgen, A.M., 140
Etienne, A.S., 236
Evans, C.S., 242, 385
Everaert, M.B.H., 180, 182
Ewert, J.-P., 7, 16, 18–20, 25, 26, 30–37, 227

Falótico, T., 241

Feekes, F., 66
Fehr, E., 475
Felsenstein, J., 414
Fiala, A., 125
Fingerling, S., 37
Finkenstädt, T., 19, 32, 33
Fischer, J., 475
Fisher, R.A., 326, 334
Fitch, W.M.404
Fitch, W.T., 374, 387, 475
Fleagle, J.G., 428
Fleming, A.S., 58
Flynn, E.G., 468
Fogarty, L., 477, 479
Foley, R., 468
Folstad, I., 328, 332, 333
Forlano, P.M., 145
Fox, M.M., 216
Fox, M.W., 253, 254
Fraenkel, G.S., 234
Fragaszy, D.M., 476
Frase, B.A., 260
Fraser, A.F., 253–254, 276
Fraser, D., 260, 262, 273, 275
Fretwell, S.D., 307
Freud, S., 50, 65
Friederici, A.D., 189
Fux, M., 118

Gahr, M., 189

Galef, B.G., 242
Gallistel, C.R., 230, 235–236
Gangestad, S.W., 468
Garcia, J., 218
Gehring, J., 260
Gendron, M., 69, 70
Gentner, T.Q., 388
Gerbault, P., 472, 473F
Gerkema, M.P., 105
Getty, T., 290
Ghirlanda, S., 74
Gibbons, A., 430
Gil, D., 327, 329
Gingins, S., 463
Giraldeau, L.A., 285, 293, 294, 294F, 305, 306, 310
Giurfa, M., 123F
Glimcher, P.W., 58
Gluckman, P.D., 468
Goldman, S.A., 126
Goodall, J., 261, 272, 273
Goodman, M., 459
Goodson, J.L., 140
Gordon, J.K., 258
Gosling, S.D., 344, 346
Gottlieb, G., 169, 182, 183, 184F, 185
Gould, E., 129
Gould, J.L., 372
Gould, S.J., 284
Gowdy, J.M., 457
Goy, R.W., 146
Grafen, A., 376, 377F, 379, 379F
Grant, B.R., 419
Grant, D.S., 210–212
Grant, P.R., 419
Graves, R.R., 438
Gray, R.D., 308–310
Green, R.E., 447
Greene, H.W., 381
Greene, S.L., 214
Greenough, W.T., 188
Griffin, D.R., 120
Groothuis, T.G.G., 349, 358
Gross, M.R., 303
Grunau, R.V.E., 268F
Guilding, C., 98
Guilford, T., 380
Gunn, D.L., 234
Güntürkün, O., 244, 246, 459
Gwinner, E., 101, 238
Gyger, M., 382
Hafez, E.S.E., 253, 254
Haff, T.M., 382
Hall, C.S., 72
Hall, G., 198
Hall, W.G., 174, 191
Hallem, E.A., 368
Hamilton, W.D., .326, 327, 416
Hamlin, K.A., 461
Hansen, K., 28, 29
Hansen, S.W., 265
Hanson, H.M., 419
Harcourt, A.H., 415
Harcourt-Smith, W.E.H., 431, 433
Harding, E.J., 269
Harlow, H.F., 179
Harlow, M.K., 179
Harmand, S., 436
Harnad, S.E., 15
Harper, D.G.C., 308, 309, 309F, 374
Harper, D.N., 210, 211F
Harrison, R., 253, 254
Hart, D., 435
Hartline, H.K., 41
Hartline, P.H., 14
Harvey, P.H., 416
Haslam, M., 241
Hasselquist, D., 324
Hausberger, M., 242
Hauser, M.D., 386, 388, 414, 476
Healy, S., 8
Hebb, D.O., 122, 124, 189
Hediger, H., 253
Hegemann, P., 130
Hegvik, D.K., 213
Heiligenberg, W.F., 14, 16, 19, 56
Heinroth, O., 174
Hemsworth, P.H., 259
Hen, R., 129
Henrich, J., 470, 476, 481
Henshilwood, C.S., 448
Herrmann, E., 474
Heyes, C.M., 239
Highfield, R., 481
Hill, G.E., 378
Hinde, R.A., 24, 51, 145, 153, 173, 174, 179, 465
Hockett, C.F., 383–386
Hockham, L.R., 401
Hoffmann, D.L., 447
Hogan, J.A., 1, 4, 5, 9, 41, 49F, 50, 51, 55, 56, 60, 68, 71, 74, 172,
173, 190, 191, 228, 419
Hogan-Warburg, A.J., 320
Holland, B., 330
Holland, P.C., 204, 205
Hölldobler, B., 22
Holloway, R.L., 435, 437, 438, 444
Honey, R.C., 176, 183
Hoppitt, W.J.E., 476
Horn, G., 175, 183, 185, 185F, 188, 188F
Horner, J.R., 407
Houck, L.D., 9
Houck, M., 408
Houpt, K.A., 253
Houston, A., 52
Hrdy, S.B., 466F, 467
Hubel, D.H., 37–38
Hughes, B.O., 254
Hull, C.L., 53
Hull, D.L., 457
Humphrey, G., 198
Hunter, W.S., 209
Huntingford, F.A., 346
Hurd, P.L., 371, 372F, 375
Hut, R.A., 91, 94, 96
Huxley, J., 400
Huxley, T.H., 458, 460

Idda, M.L., 85, 102

Ikeda, K., 27
Ingold, T., 457
Ishai, A., 39
Islam, R., 36
Jablonka, E., 479
Jackson, R.R., 242, 243, 243F, 244
James, W., 50, 68–71, 246
Janetos, A.C., 323
Janik, V.M., 386
Jarman, P.J., 322
Jarvis, E.D., 127F, 129, 130
Jeffery, K.J., 236
Jeffress, L.A., 115
Jennions, M.D., 323, 325, 326, 331
Jensen, K., 476
Jensen, P., 51
Johnsen, S., 420, 421
Johnson, J. C., 357
Johnson, K.G., 255
Johnson, K.V.A., 358
Johnson, M.H., 182, 183, 185–187
Johnstone, R.A., 376, 377F, 379, 379F
Josselyn, S.A., 232, 232F
Jukema, J., 320

Kaas, J.H., 37
Kabelik, D., 140
Kafka, W.A., 28
Kaissling, K.-E., 14, 21, 27, 29
Kamin, L.J., 206
Karter, A.J., 328, 332, 333
Kee, N., 129
Kelley, J.L., 421
Kelley, L.A., 421
Kendrick, K.M., 38
Kennedy, M., 308–310
Kessel, E.L., 400, 401
Kettler, L., 114F, 118
Kilgour, R., 253
Kilner, R.M., 378
Kingsolver, J.G., 469
Kipp, K., 461
Kirkpatrick, M., 416
Klein, D.C., 99
Klein, R., 445, 448
Kleitman, N., 106
Klem, D., 260
Knapp, R.A., 325
Koelling, R.A., 218
Koffka, K., 17
Kok, E.M.A., 239
Kokko, H., 461
Kolluru, G.R., 332
Konishi, M., 14, 113–115, 117–119
Koolhaas, J.M., 345, 349
Kornysheva, K., 227
Kortlandt, A., 65
Kossinets, G., 457
Kotiaho, J.S., 331, 377
Kovach, J.T., 325
Kralj-Fišer, S., 357
Kramer, D.L., 293, 294F
Kramer, G., 103, 238
Krause, J., 358, 447
Krauss, N., 378
Krebs, C.J., 347
Krebs, J.R., 7, 285, 289, 290, 290F, 292, 416
Kriegsfeld, L.J., 163
Kroodsma, D.E., 386
Kruijt, J.P., 63, 63F, 64, 166, 167, 178, 190, 239
Krupenye, C., 429
Kruuk, L.E.B., 349
Kruyt, W., 464, 464F
Kuenen, D.J., 18
Kuhl, P.K., 180, 181
Kupfermann, I., 27
Kupper, C., 320
Kutschera, U., 19

Labhart, T., 14
Lagerspetz, K., 170
Laland, K.N., 9, 467–469, 472, 474, 476, 479, 480, 482
Lamichhaney, L., 320
Lande, R., 417
Langford, D.J., 268
Langley, C.M., 25
Lank, D.B., 320
Lara, R., 37
Larkin, R.P., 260
Lashley, K.S., 50, 57, 111
later, P.J.B., 126, 180
Lawrence, A.B., 266
Leadbeater, E., 239
Leakey, R.E.F., 437, 438, 440
LeBas, N.R., 401
LeDoux, J., 70, 71
Leech, S.M., 378
Lee-Thorp, J., 433
Lefebvre, L., 239
Legendre, S., 333
Lehman, M.N., 98
Lehrman, D.S., 2, 57, 58, 166, 171, 172, 172F, 173
Leonard, M.L., 378
Leong, C.Y., 23
Lepre, C.J., 439
Lesku, J.A., 103
Lettvin, J.Y., 31
Levitan, R.D., 104
Levy, D.M., 266
Lewis, D.J., 217
Lewis, H.M., 479
Lewontin, R.C., 284
Li, N.P., 468
Lieberman, D.E., 441
Little, A.C., 323
Livingstone, M.S., 37, 38
Livoreil, B., 306
Lloyd, J.E., 401
Loeb, J., 234
Lomo, T., 124
Lordkipanidze, D., 442, 443
Lorentsen, S.H., 378
Lorenz, K., 51, 52F, 54, 58, 59, 64, 65, 67, 71, 72F, 73, 74, 166,
171–179, 226, 226F, 227, 228, 230, 398, 402, 464, 465
Low, B.S., 326
Lozano, G.A., 378
Lu, W., 85

Macaluso, E., 40
MacArthur, R.H., 285
Macedonia, J.M., 242, 385
Machado, C.A., 416
Macphail, E.M., 5, 8, 218–219
Magrath, R.D., 382
Makela, R., 407
Maki, W.S., 213
Malmkvist, J., 265
Manser, M.B., 378, 385
Mappes, J., 377
Marchant, E.G., 91
Markl, H., 22
Markowitz, H., 261
Marler, P., 180, 182, 183, 382, 386, 390
Martin, J.G.A., 350
Martins, E., 398
Mason, G.J., 263F, 264, 271
Masson, J.M., 272
Mathew, S., 480
Matzel, L.D., 236
Max, D.T., 409
May, R.M., 416
Mayer, C., 459
Maynard Smith, J., 7, 297, 299, 374
Mayr, E., 4, 417, 482
McAleer, K., 294
McAuliffe, K., 476
McCabe, B.J., 175, 188
McCarthy, S., 272
McCarty, J.P., 378
McComb, K., 374
McDiarmid, R.W., 381
McDonald, G.C., 357
McDougall, I., 448
McDougall, W., 50
McEwen, B.S., 146, 150
McFarland, D.J., 52, 67
McGregor, P.K., 383
McPhail, J.D., 412
McPherron, S., 436
McRae, R.R., 352
Meagher, R.K., 271
Meck, W.H., 60
Meeuwissen, G.B., 178
Meffert, L.M., 349
Mello, C.V., 189
Menaker, M., 102
Mendl, M., 73, 269
Menzel, R., 121, 236
Mesoudi, A., 470, 471
Mey, J.L., 389
Meyer, C.C., 178
Meyer, E.P., 14
Michelena, P., 359
Micic, G., 93
Midgley, M., 253
Milinski, M., 310, 378
Miller, G.A., 212
Miller, N.E., 60
Mishkin, M., 37, 38
Mistlberger, R.E., 87, 91, 94, 95, 101
Mittelstaedt, H., 236
Mittelstaedt, M. L., 236
Mohawk, J., 99
Moiron, M., 348
Moiseff, A., 113–115
Møller, A.P., 325–329, 331, 332F, 333, 334, 377, 382
Monaghan, P., 457, 461
Montiglio, P.O., 349, 355, 356, 358, 360
Montminy, M.R., 125
Moore, J., 462
Moore-Ede, C.A., 95
Moore-Ede, M.C., 88
Moorman, S., 126, 127F, 188F, 189
Morand-Ferron, J., 306
Morgan, T.J.H., 477, 477F, 479
Morris, D., 465
Morris, R.G.M., 202, 235
Morton, J., 182, 186, 187
Moser, E.J., 237
Moser, M.-B., 237
Mrosovsky, N., 81
Mueller, J.C., 349
Mulder, C.K., 87
Muller, U., 121
Mutzel, A., 359
Myers, J.H., 347

Nadel, L., 237

Nagel, G., 130
Nakagawa, S., 359
Narins, P.M., 22
Nelson, K.R., 212
Nelson, R.A., 163
Nelson, R.J., 57–59
Nestor, P., 231
Nettle, D., 467, 468
Newman, E.A., 14
Newton-Fisher, N.E., 436
Nicholson, K.E., 406
Nicol, A.U., 188F, 189
Nicol, C.J., 262
Nicolaus, L.K., 258
Niemela, P. T., 359
Niven, J., 233, 246
Nolte, D.L., 257
Nonacs, P., 292
Nordby, J.C., 386
Nottebohm, F., 126, 127F, 129, 130, 189
Nowak, M.A., 481
Nowel, A., 445
Nowicki, S., 119, 370, 378, 381, 386, 387, 480

Odling-Smee, F.J., 472

Odom, K.J., 407
O’Donnell, S., 258
Ohnuma, K., 447
O’Keefe, J., 237
Olive, R., 234
Oliveira, R.F., 357, 405
Oppenheim, R.W., 174
Orians, G.H., 319, 319F
Oring, L.W., 318
Oteiza, P., 234
Otter, K., 383
Ottoni, E.B., 241
Ouattara, K., 387
Oudman, T., 296
Owens, I.P.F., 467

Packer, C., 321

Padian, K., 407
Page, R.A., 383
Pagel, M., 398, 457, 481
Palameta, B., 239
Pangle, W.M., 382
Panksepp, J., 70, 73, 462
Paradis, E., 398
Parker, G.A., 310, 311
Parri, S., 377
Pascalis, O., 21
Patke, A., 99
Patrick, S.C., 357
Patterson, K., 231
Patterson, L.D., 357
Paul, E.S., 72, 269
Paul, M.J., 103, 104
Pavlov, I.P., 2, 199–201, 203, 208, 218, 254
Peake, T.M., 383
Pena, J.L., 117
Perreault, C., 480
Perret, D.I., 38
Perry, S., 476
Peters, S., 180, 386
Petrie, M., 315, 323, 326, 329, 331
Pfaff, D.W., 140, 145, 146
Pfaffmann, C., 145
Phillips, P.A., 262
Phoenix, C.H., 147
Pianka, E.R., 285
Piersma, T., 57, 238, 295, 320
Piggins, H.D., 98
Pike, T.W., 358, 359
Pittendrigh, C.S., 84, 93
Platt, J.R., 402
Plotkin, H., 457
Plummer, T., 436
Poeppel, D., 111
Pomiankowski, A., 323, 334
Povinelli, D.J., 429, 433
Prater, C.M., 36
Prather, J., 189
Prete, F.R., 14
Pribil, S., 319
Price, G.R, 297, 299
Price, T., 402
Proctor, H.C., 368
Promislow, D.E.L., 331
Pulliam, R.H., 308

Qadri, M.A.J., 214

Rabaneda-Bueno, R., 357

Radovčić, D., 447
Ralph, M.A.L., 231F
Ralph, M.R., 98
Rand, A.S., 22, 383, 418
Randles, D., 69
Rapaport, L.M., 476
Rasanen, K., 349
Rastogi, A., 103
Ratnieks, F.L.W., 371
Rattenborg, N.C., 102, 103
Rayleigh, L., 113
Read, A.F., 416
Reale, D., 344, 344F, 347–348, 350, 352, 354, 360
Reaney, L.T., 357
Reardon, E., 260
Reby, D., 374
Reddipogu, A., 37
Redish, A.D., 237, 244, 245F
Regan, T., 253
Rendell, L.R., 476, 479
Rescorla, R.A., 207
Resende, B.D., 241
Rice, W.R., 330
Richerson, P.J., 469, 470, 474, 476, 480, 481
Ricklefs, R.E., 360
Ridley, M., 481
Riley, J.R., 373, 373F
Rissman, E.F., 152
Roberts, E.P., 238
Roberts, W.A., 211, 212
Rockenbach, B., 475
Roebroeks, W., 441
Roeder, K.D., 59
Roenneberg, T., 93, 106
Rogers, T.T., 231
Rollin, B.E., 253, 254, 272
Rolls, E.T., 38
Romanes, C.J.R., 234
Romanes, G.J., 271
Romero, L.M., 163
Ron, S.R., 418
Ronald, K.L., 119, 120
Roper, T.J., 56
Rose, G., 14
Rose, S., 457
Roth, A.E., 475
Roth, G., 428F, 429
Roulin, A., 356
Rowe, L.V., 378
Rowland, W.J., 24
Rowsmitt, C.N., 91
Royaute, R., 360
Royle, N.J., 354
Ruesch, H., 253
Rushen, J., 258, 259, 266
Russell, J.A., 68, 70
Rutter, M., 179
Ryan, M.J., 22, 324, 368, 383, 398, 402, 416–419

Sagi, D., 39
Saino, N., 327
Saldanha, C.J., 158
Salzen, E.A., 178
Schachter, S., 69, 70
Schacter, D.L., 120
Scheibler, E., 96
Schel, A.M., 390
Schibler, U., 99
Schiefenhövel, W., 457
Schleidt, W., 25
Schluter, D., 412
Schmid-Hampel, P., 295
Schneider, H., 36
Schneirla, T.C., 61
Schuett, W., 357, 358
Schulte-Hostedde, A., 357
Schürg-Pfeiffer, E., 35
Schwippert, W.W., 36
Searcy, W.A., 119, 319, 370, 380–381, 386, 387, 480
Seed, A.M., 459, 474
Seeley, T.D., 372
Segerstråle, U.S., 466
Seidl, A.H., 115, 117F
Seitz, A., 23
Semaw, S., 436
Senghas, A., 450
Sevenster, P., 65
Sevenster-Bol, A.C.A., 55
Seyfarth, R.M., 242, 384, 385, 475
Shaaya, M., 130
Sherman, V., 386
Sherry, D.F., 67, 235
Shettleworth, A.J., 8
Shettleworth, S.J., 219, 235, 237, 242, 309F
Sibly, R.M., 304
Siepielski, A.M., 469
Sih, A., 285, 290, 345, 349, 357, 359
Simpson, J.A., 468
Singer, J.E., 69, 70
Singer, P., 253
Singer, W., 37
Skene, D.J., 94
Skinner, B.F., 3, 4, 58, 200, 201, 205
Slater, P.C., 216
Slater, P.J.B., 180, 181F, 386
Slimak, L., 447
Slocombe, K.E., 385
Sluckin, W., 175
Smiseth, P.T., 378
Smith, E.A., 472, 481
Smith, H., 438
Snowdon, C.T., 242
Sol, D., 246
Somel, M., 469, 469F
Spalding, D.A., 2, 174
Spear, N.E., 217
Spencer, H., 462
Sponheimer, M., 433
Spreckelsen, C., 35
Squire, L.R., 216
Staddon, J.E.R., 419
Stamps, J.A., 349, 354, 360
Steger, R., 382
Stein, L.R., 359
Steiner, A.P., 244, 245F
Stephens, D.W., 285, 292
Stevens, S.S., 419
Stimson, W.H., 332
Stoddard, P.K., 389
Stratmann, M., 99
Straub, J.J., 204, 205
Strausfeld, N.J., 122
Struhsaker, T.T., 242, 384
Suga, N., 14
Sussman, R.W., 435
Sutherland, W.J., 310, 318
Suzuki, T.N., 385, 387
Symons, D., 467, 468
Számadó, S., 381

Taborsky, M., 349, 358

Taddio, A., 268
Takahashi, J.S., 99
Takahashi, T., 118
Tamura, M., 382
Tattersall, I., 430, 446, 447, 449–451
Ten Cate, C., 419
Terlouw, E.M.C., 267
Ter Polkwijk, J.J., 16, 18
Terrace, H.S., 212, 213F
Texier, P.J., 448
Thieme, H., 444
Thomas, E.M., 272
Thorndike, E.L., 2, 241
Thornhill, R., 328
Thornton, A., 471, 476
Thornton, J.W., 403, 408
Thorpe, W.H., 9, 55, 252, 254, 255, 262
Tinbergen, N., 2, 4–6, 8, 9, 12, 15–19, 24, 26, 33, 36, 41, 48, 56F,
61, 62F, 63, 65, 73, 166, 173, 191, 226, 226F, 227, 228, 234, 235,
272, 397, 398, 400, 419, 421, 422, 463, 464, 464F
Toates, F.M., 51, 74
Tobias, J.A., 411, 412F, 413F
Tolman, E.C., 236
Tomasello, M., 459, 479, 481
Tong, F., 38
Tooby, J., 467, 468
Toren, C., 457
Tosini, G., 102
Tracy, J.L., 69
Tramontin, A.D., 126, 128
Tregenza, T., 308, 310
Trivers, R.L., 331, 332, 336, 354, 461, 465, 467
Tulving, E., 214, 230, 231
Tuttle, M.D., 383
Udin, S., 40
Uher, J., 352
Ungerleider, L.G., 37, 38

Vaccarino, A.L., 235

Vail, A.L., 463
Valenza, E., 21
Van Boxel, F., 60
Van Couvering, J.A., 430
Vander Wall, S.B., 6
Van De Waal, E., 478F
Van Gils, J.A., 57, 238, 295, 309
Van Heijningen, C.A.A., 389
Van Iersel, J.J.A., 57
Vanin, S., 97
Van Lawick-goodall, J., 241
Van Oers, K., 347, 349, 357
Van Oort, B.E., 80, 85
Van Rhijn, J.G., 320
Vaughan, W. Jr., 214
Vehrencamp, S.L., 133
Verhulst, S., 9
Verner, J., 319
Vestergaard, K.S., 59, 60F, 191
Villa, P., 441
Vincent, A.C.J., 318
Voight, B. F., 472
Voirin, B., 96
Vom Saal, F.S., 349
Von Frisch, K., 236, 254, 371–373
Von Holst, E., 58, 229, 230
Von Schants, T., 328
Von Schantz, M., 92
Von Uexkiill, J., 14
Vortman, Y., 377

Waddington, C.H., 167, 168, 169F

Wagner, A.R., 207, 212
Wagner, H., 112F, 118
Walker, A.C., 437, 438, 440
Wallraff, H.G., 89
Ward, T., 457
Waring, G.H., 253
Wasserman, E.A., 212
Watson, J.B., 3, 462, 463F
Watters, J.V., 359
Watts, D.J., 457
Weary, D.M., 271
Weaver, D.R., 97
Webb, I.C., 96
Webb, L.E., 264, 264F
Weber, S., 336
Wedekind, C., 329
Wehner, R., 236
Weigl, P.D., 238
Weishampel, D.B., 408
Weiss, A., 352
Weiss, E., 256
Weiss, K.R., 27
Welch, A.M., 327
Wells, D.L., 257
Welsh, D.K., 99
Wenner, A.M., 373
Wenzel, J.W., 402, 403
West Eberhard, M.J., 417
Westneat, D.F., 359
Wey, T.W., 359
Wheeler, B.C., 368, 382, 475
Wheeler, P., 440
Wheeler, W.M., 1
While, G.M., 349
White, T.D., 448
Whitehead, H., 476, 479
Whiten, A., 476, 478F, 479
Wichman, H.A., 416
Wiersma, C.A.G., 27
Wiesel, T.N., 38
Wikelski, M., 360
Wilcox, R.S., 243
Wilczynski, W., 417
Williams, C.L., 174, 191
Williams, G.C., 285, 297, 413, 465
Wilson, D.M., 230
Wilson, D.S., 347, 457
Wilson, E.O., 6, 465, 466
Wilson, M., 331, 467
Wiltschko, R., 14, 237, 238
Wiltschko, W., 14, 237, 238
Wimberger, P.H., 402
Wingfield, J.C., 163
Wischner, D., 265
Wolf, M., 354, 360
Wolski, T.R., 253
Wong, D., 269
Wood-Gush, D.G.M., 253, 254
Wouters, A.G., 4
Wrangham, R., 436, 441
Wu, L., 451
Wustenberg, D., 125
Wynne, C.D.L., 8

Yamada, N., 80
Yamazaki, S., 93
Yanega, D., 371
Yang, C., 4
Yasukawa, K., 378
Yoshida, N., 37
Young, L.J., 140, 163
Zahavi, A., 375
Zhao, W.C., 130, 131F
Zink, K.D., 441
Zuberbühler, K., 385, 387, 475
Zuk, M., 326, 327, 332, 334, 416
Zupanc, G.H., 26
subject index
Acheulean implements, 440f
acoustic display, 331
acrophase, 83–84
action-specific energy, 51, 54
activating effects, 156
adaptations, 172–174, 214, 219, 281–285, 398, 399, 402, 413–
417, 433, 434, 438, 465, 467, 468, 472
affective quality, 70
affective states, 274
abnormal behaviors, 266–267
and animal welfare science, 266–271
assessing, 267–271

African buffalo (Syncerus caffer), 262

African clawed frog (Xenopus laevis), 159
African naked mole rat (Heterocephalus glaber), 154
aggression, 64–66, 71, 72, 169, 170, 266, 297, 346, 347, 359,
360, 380, 465
alarm calls, 242, 382, 384, 385, 475
Alces alces, 259f, 260
algorithm, 30, 31, 35, 37, 41, 446, 450
alternative resource harvesting strategies, 303–304
alternative strategies, 160, 298, 303, 305
altruism, 7
ambivalence, 61–64
American buffalo (Bison bison), 251
American Heritage Dictionary, 233
American mink (Mustela vison), 263
Ammophiia sabulosa, 224f
Ammophila campestris, 223
amphibians, 29, 40, 102, 146, 149, 159, 174
visual feature detection in, 29–37

amplitude, 83, 84, 90, 97, 116, 376, 380, 412

analogous, 87, 90, 129, 148, 179, 189, 205, 239, 242, 419
Anas platyrhynchos, 308
Anas strepera, 186
ancestral traits functionality, 407–409
animal behavior, 1, 4–6
animal welfare, 8–9
behavioral ecology, 6–7
cognitive ecology, 8
cognitive neuroscience, 7
human nature, 8–9
neuroecology, 8
neuroethology, 7

animal experimentation, 8
animals
better environments, designing, 260–264
environmental preferences, testing, 262
handling, improving, 259
mitigating harm to, 259–260
 motivation strength, testing, 262–264
natural behavior, accommodating, 261–262
undesirable behavior, preventing, 257–258
using the abilities, 255–257
verses human beings, 471

animal signals, as symbols, 383–386

animal traditions, 476–477
animal welfare, 8–9, 73, 251–255, 265–267, 269, 271–276
and affective states, 266–271
behavioral genetics and, 265–266

Anolis carolinensis, 101

anolis lizard (Anolis carolinensis), 101
Anopheles mosquitos, 368
anthropomorphism, 225
anticipatory activity, 87
Aphelocoma californica, 232
Apis mellifera, 120
appetitive conditioning, 200
“appetitive sexual behavior” (ASB), 155–158
applied animal behavior, 252–254
appraisal approach, 70
Aquarius remigis, 357
arms races, 322, 416
assessor, 300–302
association formation conditioning, 206–208
associative analysis, conditioning, 203–206
associative theory, conditioning, 208–209
attachment theory, 465
audience effects, 390
auditory map, 117–119
space, barn owl, 113–125

auditory processing
barn owl (Tyto alba), 112
cognition, 120
honeybee learning, 120–125
perception and sexual selection, 119–120
streams, 113–117

Australopithecus afarensis, 433–435

Australopithecus anamensis, 433
autoshaping, 199, 202
procedure, 199

aversive conditioning, 200

azimuth, 89, 113–118

backward engineering, 282–283

logic, example, 283–284

barn owl (Tyto alba), 112

bar-tailed godwit (Limosa lapponica), 237
basic emotions, 69
Batesian mimicry, 381
battery cages, 254
Bayesian inference, 404
behavioral adjustment, 305, 306
predictions for games, 305–306

behavioral biology, 1–4

behavioral change, 66, 67, 163, 448
mechanisms of, 66–67

behavioral development, 166–192

basic developmental issues, 167–179
beyond nature/nurture, 173
developmental discontinuities, 173–174
embryology and, 167–169
filial imprinting, 175–176
genes and behavior, 169–171
humans and primates, attachments development, 178–179
imprinting, 174–175
learning and, 167
“nature/nurture” debate, 171–172
ontogenetic adaptations, 173–174
physiological mechanisms, hormone effects, 144–152
sexual imprinting, 176–177

behavioral ecology, 6–8, 306, 310, 342, 343, 346, 456, 458, 467
mechanism to function, 6–7

behavioral enrichment, 262

behavioral flexibility, 465, 468
behavioral genetic innovation, 409–411
behavioral genetics, 265–266
behavioral interactions, hormone secretion, 152–154
behavioral neuroscience, 7
behavioral stable strategy, 306
behavioral syndromes, 345
behavioral ways, stimulus selection, 24–26
behaviorism, 3, 204, 205, 458, 462
behavior mechanisms, 50, 54, 57, 59, 67–68, 225, 229, 230, 233,
234
behavior systems, 48–50, 58, 60–62, 64–68, 71–73, 190
ambivalence, 61–64
displacement, 64–66
inhibition and intention movements, 61
interactions, 60–66
redirection, 64

Bengalese finch (Lonchura striata), 177, 177f

bilateral asymmetry, 328
binaural fusion, 113
binding of features, 37
biological rhythms, 78, 79, 81, 83, 85, 87, 89, 91, 93, 95, 97, 99,
101, 103, 105
spectrum and discovery of, 79–82

birdsong
learning, 180–182
speech and language, human infants, 180–182

Bison bison, 251

black-capped chickadees (Parus atricapillus), 235
black-capped chickadees (Poecile atricapillus), 285
blackheaded gulls (Larus ridibundus), 6
blocking, 159, 206, 207
blowfly maggot (Calliphora erythrocephala), 234
bonobos, 429
brain
control of behavior, 125–132
primer on hormone action, 137–144

brainstem, 115, 117, 119

Calidris canutus, 238, 296f

Calliphora erythrocephala, 234
canalization, 168
Capreolus capreolus, 357
Capuchin monkeys (Cebus apella), 478f
castrated rats (Rattus norvegicus), 143
Cataglyphis fortis, 236
categorization, 4, 15
causal factors, 54–60
external versus internal, 51–53
hormones and substances, 57–58
intrinsic neural factors, 58–60
specific versus general effects, 53–54
stimuli, 55–57

causation, 4–6, 9, 66, 156, 397, 422, 433

Cebus apella, 478f
Cerastoderma edute, 295
Cercopithecus aethiops, 242
Cervus elaphus, 257, 357
character displacement, 411
chase-away model, 330
chicken (Gallus gallus), 101
chimpanzees, 429, 433
chipmunks (Tamias sibiricus), 357
chunking, 212
circadian, 79, 81, 84–90, 92–102, 104
circadian clocks
compass and calendar, 89
coordinate behavior with external world, 86–88
evolution and adaptive uses, 84–89
phylogenetically ancient, 84–85

circadian phenotypes, nonphotic stimuli and plasticity, 94–97

circadian rhythms
environmental synchronization, 89–97
mammals, 97–101
neural mechanisms, 97–102
nonmammalian vertebrates, 101–102
temperature compensated, 81–82
in temporal isolation, 79–81

circalunar, 79
circannual, 79
circatidal, 79
classical conditioning, 175, 199, 203–205, 220, 254, 257
clock genes, 98–100
closest living relatives, 429–430
cochlea, 114, 116
coevolution, 360, 416–418, 479, 480
cognition, 223–247
central mechanisms, 228–233
concepts, 233
knowledge, basic mechanisms, 225–233
memory, 230–232
mental structure and processes, 245–246
mental structures, knowledge, 233–245
motor and perceptual mechanisms, 226–228
wasp, 223–225

cognitive bias, 270f

cognitive ecology, 8
cognitive neuroscience, 7
cognitive processes, 474–479
cognitive psychology, 3–4
cognitive structures, 4
Columbia livia, 239
command releasing system (CRS), 36
common environment effects, 327, 329
communication, 367–391, 475
constrained signals, 374–375
conventional signals, 380–381
deception, 381–382
differential benefits, 378–380
 handicaps, 375–378
and information, 368–370
non-informational and informational interpretations, 369f
signaling, 371–374
signal reliability, 370–371

comparative psychology, 2, 3
comparator mechanisms, 228
competence, 169, 274, 328
complex conditioning procedures, 201–202
conditional control, 201–202
discrimination learning, 202

condition dependent handicap, 376

condition-dependent traits, 326
conditioned emotional response, 199, 206
conditioned response, 121, 199, 209
conditioned stimulus (CS), 32, 121, 123, 199, 203, 209
conditioned taste aversion, 200, 257, 258
conditioning, 2, 122, 124, 125, 176, 199–201, 203–209, 220
associative analysis of, 203–206
classical conditioning, 203–205
instrumental conditioning, 205–206

configurational sign-stimuli, 17–22

comparable signs with stimuli across species, 19
dynamic configurations, 17–19
in human perception, 21
static configurations, 17–19
configurational stimuli, 17, 21
conflict, motivational, 60
connectionism, 208
consciousness, 58, 73, 120, 230, 272, 275
consolidation, 130, 178, 216, 217, 220, 232
“consummatory sexual behavior” (CSB), 155–158
contiguity, 206, 208
contralateral, 32, 40, 115, 117
cooperation, 357, 457, 475, 479, 481
coping styles, 345, 349
core affect, 70
Corvus brachyrhynchos, 257
cospeciation, 416
Coturnix coturnix japonica, 101, 177
Coturnix japonica, 143
Crangon crangon, 295
crepuscular, 79, 92
critical period, 147, 175, 186
crows (Corvus brachyrhynchos), 257
cues, 368
cultural evolution, 470
cultural information, 471
culture
human reliance, 471–474

cumulative culture, 471

currency of fitness, 286–288, 291, 298, 305, 308
Damaliscus lunatus, 382
Danio rerio, 234, 269
dark-eyed junco (Junco hyemalis), 238
Darwinism, 460–461
Dasyurus hallucatus, 257
dead reckoning, 236
deception, 242, 381, 382, 390, 479–481
decision, 12, 111, 150, 287, 291, 292, 305–308, 323
deer (Odocoileus hemionus), 257
delayed matching-to-sample (DMTS), 209, 210, 212, 213, 218
depolarization, 124, 129, 130
desert ant (Cataglyphis fortis), 236
desert iguana (Dipsosaurus dorsalisa), 102
development, 4, 5, 147, 166–168, 170–174, 179, 180, 183, 186,
187, 190–192, 242
of brain and behavior, 187–189
of motivational systems, 190–191

developmental instability, 328

differentiation, 147–152, 168, 408
diffraction, 113
Dipsosaurus dorsalisa, 102
direct benefits, 119, 324–326
directed-forgetting paradigm, 213
direct fitness, 324–326
benefits, 324–326

discrete emotion approach, 73

discrimination training, 202, 208
dishabituation, 24, 25, 198
disinhibition theory, 65, 66
dispersal, 346, 347, 356, 357, 481
displacement activities, 65
displacement activity, 65, 66
display, 63, 64, 67, 119, 120, 148, 162, 328, 331, 332, 368, 375,
376, 400, 406
diurnal, 79, 81, 86, 88, 90, 92–94, 103
domestic fowl (Gallus domesticus), 137
dorsal processing stream, 39
Dove, 57, 58, 152, 240, 298–302, 305, 311
doves (Zenaida macroura), 239
dustbathing, 263

earliest putative hominins, 430–433

eastern chipmunks (Tamias striatus), 292
eavesdropping, 382–383
efference copy, 229
elevation, 113, 114, 117, 118
elks (Cervus elaphus), 257, 357
emotion
human, 67–71
nonhuman, 71–73

Engystomops pustulosus, 382

entrained, 82, 91, 92, 96–99, 103
entrainment, 89, 90, 92–97, 100, 101, 103
environmental preference research, 262
environment of evolutionary adaptedness (EEA), 468
epigenesis, 168
epigenetic landscape, 168
episodic-like memory, 214
Erbkoordination, 230
ethology, 1, 2, 6, 9, 166, 171, 252, 253, 397, 398, 463–465
and comparative psychology, 2–3

events pairing, 199–201

classical conditioning, 199–200
instrumental (operant) conditioning, 200–201

evolution, 4, 397–402
ancestral states, uncovering, 407
brain and, 418–421
divergence time, assessing, 411–413
patterns of, 405–406
phylogenetics and, 402–405
strong inference, 398–399

evolutionarily stable strategy (ESS), 7, 296–306, 349

evolutionary arms race, 322
evolutionary psychology, 9, 456, 458, 467, 468, 471, 472
evolutionary theory, 456–457
Darwin to behaviorism, 458–463
to human behavior, 457–479
implications, behavior study, 479–482
exafferent stimulation, 229
extinction, 208, 257, 333, 337

facial expressions, 268

facilitation, 184, 217
Felis serval, 261f, 262
fighting displays, 297
fighting game
alternative strategies, specifying, 298
expected evolutionary solution, 299–300
payoffs to each alternative, specifying, 298

filial imprinting, 174–176, 178, 185, 186, 188, 191

first stone tool makers, 435–437
Fisherian character, 326
Fisherian traits, 326
Fisher’s theory of runaway sexual selection, 417
fitness, 5, 286, 288, 291, 297, 298, 311, 319, 324–326, 330, 351,
376–379, 390, 480
flavor aversion learning, 200, 204, 206
free-run, 81, 90, 94
frequency, 22, 81, 114–116, 119, 205, 296, 300, 304–306, 321,
384, 412, 418
function, 4–6, 140, 150, 281–286, 291, 293, 297, 311, 321, 400,
404, 408, 413

Gadwall duck (Anas strepera), 186

Gallus domesticus, 137
Gallus gallus, 101
Gallus gallus spadiceus, 63
game theory, 282, 296–306, 311, 381
Gasterosteus aculeatus, 358
gene expression, 129–130
generalist, 29, 288, 435
generalization, 167, 202, 208, 386
general-process learning theory, 219
genetic complementarity, 329
gene transcription, 124, 125, 129, 132, 163, 408
gerbil (Meriones unguiculatus), 236
glaucous gulls (L. hyperboreus), 86
glial cells, 126
golden hamsters (Mesocricetus auratus), 236
good genes, 326, 327, 329, 331, 333, 417
theory, 417

great tits (Parus major), 357, 358

grebe (Porpodiceps cristatus), 400
green iguana (Iguana iguana), 102
group selection, 297, 465, 481
Gryllus integer, 303
guillemots (Uria lomvia), 86
guppy courtship, 53f

habitat matching, 308

habitat selection, 7, 356
habituation, 19, 24, 25, 198, 209
handicap, 328, 375–380
hypothesis, 328

handling time, 287, 288

hands, 82, 431, 433, 436
hard-shelled cockles (Cerastoderma edute), 295
Hawk, 18, 19, 25, 282, 298–302, 305, 311
Hawk–Dove game, 299t, 300–303, 305, 311
hawkmoth (Xanthopan morganni praedicta), 399
Hebb’s rule, 122
heritability, 323, 350, 458
herring gull (Larus argentatus), 61
Heterocephalus glaber, 154
heterogeneous summation, 12, 23–24
“higher” mental processes, 242–245
high vocal center (HVC), 126–131, 151, 159–160
homeostasis, 88
home range, 8, 88, 235, 357
hominin behavior, 427–452
first cosmopolitan hominid, 443–445
Neanderthals, 445–447
out of Africa, 442–443

Homo ergaster, 437, 438f, 440

homologous behaviors, 398, 403, 413
homoplasious, 403
Homo sapiens, 429, 435
 evolution of, 448–451

honest signaling, 380, 480, 481

honeybee learning, 120–125
converging input, 124–125
mushroom bodies, 122–123

honeybees (Apis mellifera), 120, 236

hormonal regulation, 154–160
hormone action, 137–144
cellular mechanisms, 140–144
central versus peripheral actions, 144–145
genomic transcription control, 140–142
intracellular receptors, 140–142
intracellular signaling cascades modulation, 143–144
membrane receptors, 143–144
multiple sites, non-redundant fashion, 145–146
sex differences, steroid action, 146–147
steroid metabolism, 142–143
steroids, social behaviors, 145–146

hormone secretion, behavioral interactions, 152–154

house sparrow (Passer domesticus), 101
Human Behavioral Ecology approach, 467
human emotion, 67–71
human ethology, 463–467
human language, 383–390
human nature, 8–9
ideal free distribution (IFD), 282, 306–311
game, 307–308
testing, 308–310

ideal free game, 306–308

Iguana iguana, 102
imitation, 241, 242, 429, 437, 447, 476, 477, 479, 481
immediate early genes, 129, 130, 146, 153, 157, 189
imprinting, 166, 167, 174–179, 185–189, 191, 192, 198
conditions for, 176
irreversible, 177–178
and learning, 176

independent contrasts, 413–415

index signals, 374
indigo buntings (Passerina cyanea), 238
indirect benefits, 120, 324, 326
indirect fitness, 326, 329
benefits, 326

individual personality differences, 354–360

dispersal, habitat selection, and space use, 356–357
getting old, 359–360
mating, 357
parental care, 359
social interactions and social roles, 358

induction, 124, 137, 154, 157, 168, 169, 184

information, 8, 12, 13, 87, 171, 212, 214, 216, 217, 236, 237, 327,
334, 367–369, 371, 374, 475, 480
 transmission, 474–479

infradian rhythms, 79
inhibition, 35, 40, 59, 61, 65–67, 157, 158
innate, 2, 26, 71, 169–173, 183, 227, 228
innate releasing mechanism (IRM), 12, 26–29, 36, 227, 228
neuronal correlates, releasing systems, 27
scent-coding, 27–29
specialized receptor cells, insects, 27–29

insight, 9, 68, 239

instinct, 68
instrumental conditioning, 2, 4, 201, 205, 220
intensity of sperm competition, 359
intention movements, 61
intersexual selection, 316, 323
female preferences, 323–324

interval timer, 104

intrasexual selection, 316, 320–322
intrinsic neural factors, 58–60
invariance, 17, 21, 81
ipsilateral, 35, 115, 117
irreversibility, 177, 178

Japanese quail (Coturnix coturnix japonica), 101, 155–158, 177

appetitive vs. consummatory sexual behavior, 155–156
endocrine controls, 156
neural circuits, 156–157
 rapid effects of estrogens, 158
role of dopamine, 158

jet lag, 93
Junco hyemalis, 238

Kakapo (Strigops habroptilus), 257

kin selection, 7, 466

laboratory mice (Mus musculus), 153

Larus argentatus, 61
Larus ridibundus, 6
larval zebrafish (Danio rerio), 234
law of effect, 205
learning, 3, 4, 120, 167, 176, 180, 182, 189, 197–199, 208, 209,
218, 219, 239, 242, 386, 462, 476
complex conditioning procedures, 201–202
events pairing, 199–201
general and special processes, 217–220
mechanisms of, 203–209
single event, exposure, 198

Lee-Boot effect, 154

lek, 162, 319, 320, 323
lekking mating system, 318
L. hyperboreus, 86
light–dark (LD) cycles entrainment, 89–90
in diurnal species, 93–94
 in nocturnal rodents, 90–93
nonparametric model, 90–93

limits of science
Animal Welfare, Science, and Ethics, 273–275
mental experience of animals, 271–273
probing, 271–275

Limosa lapponica, 237

Lonchura striata, 177
long-term memory, 214–215
long-term potentiation (LTP), 124, 129
lordosis, rats, 154–155

maintenance, 48, 184, 213, 315, 322, 325, 326, 334, 357, 383,
390
male junglefowl (Gallus gallus spadiceus), 63
male praying mantis (Mantis religiosa), 59
male sexual behavior
appetitive vs. consummatory sexual behavior, 155–156
endocrine controls, 156
Japanese quail, 155–158
neural circuits, 156–157
rapid effects of estrogens, 158
role of dopamine, 158

male song sparrows (Melospiza melodia), 5, 183

mallard ducks (Anas platyrhynchos), 308
mammals, 97–101
 cellular and molecular clockworks, 99–101
circadian pacemaker, localization, 97–99
non-transcriptional circadian clocks, 101

Mantis religiosa, 59
Marmota flaviventris, 346
marmots (Marmota flaviventris), 346
masculinization, 147, 148
mate choice benefits, 324–329
maternal effects, 329, 356
mating success, 316, 318, 319, 321, 322, 326–329, 331, 335, 357,
358
mating systems, 316–320, 398, 414–416, 418
maximum parsimony, 403
Melospiza georgiana, 183
Melospiza melodia, 5
memetics, 470–471
memory, 197, 209–217
general and special processes, 217–220

Meriones unguiculatus, 236

Mesocricetus auratus, 236
mice (Mus musculus), 143
Microtus ochrogaster, 347
midbrain, 30, 33, 36, 117, 155
midshipman fish (Porichthys notatus), 143, 159
migration, 281
mimicry, 242, 381
monogamous mating system, 317
monogamy, 140, 317, 320
monophyletic, 404, 418
moose (Alces alces), 259f, 260
motivation, 16, 26, 41, 48, 50–54, 59, 63–65, 71, 73, 156, 158,
262, 263, 265
motivational conflict, 60
motivational energy, 50–51
motivational issues, 50–54
central versus peripheral locus of action, 54
external versus internal causal factors, 51–53
motivational energy, 50–51
specific versus general effects, causal factors, 53–54

motivational systems, 190–191

dustbathing, 191
hunger system, 190–191

motivation analysis, 62–64, 71, 73

multifunctional properties, concert, 37–41
multimodal, 119
multiple ornaments, 333–334
mushroom body, 122–124
Mus musculus, 143, 153
Mustela putorius, 186
Mustela vison, 263

Nariokotome Boy, 437, 439f, 441

natural selection, 2, 282, 285, 311, 316, 346, 347, 375, 456, 458,
460, 469, 470, 472
neural ensemble, 225
neural networks
sensory substitution, 40–41
universal potential, 40–41

neural processes, 26–37

neuroecology, 8
neuroethology, 7, 26
neuron, 27, 30, 35, 38, 39, 58, 116, 122, 124, 126, 128–130, 245,
246
neurotransmitter, 58, 114, 122, 124, 125, 144, 158
niche construction, 471–474
nocturnal, 79–81, 86, 88, 90–96, 103–105
nonhuman emotion, 71–73
nonmammalian vertebrates, 101–102
Nucifraga columbiana, 214
nuptial gifts, 400, 401
nutcracker (Nucifraga columbiana), 214

object recognition mechanism, 228

Odocoileus hemionus, 257
ontogenetic adaptations, 173, 174
ontogeny, 5, 147–149, 163, 174, 328, 346, 349, 397
operant, 3, 58, 73, 157, 176, 200, 201, 263
operant conditioning, 3, 157, 176, 201
operational sex ratio, 318, 319
optimal diet models, 285–286
optimal foraging theory, 285
optimality, 282–285
backward engineering, 282–283
hypothesis, adaptation, 284–285
optimal flight speeds, 283–284

optimality models, 283, 286, 291, 311

in foraging, 285–296
foraging models, expanding, 295–296
patch models, foraging theory, 291

optimality theory, 7, 285

optimal patch residence model, 291
constraints, 291–292
currency of fitness, 291
decision, 291
drawing the predictions, 292
experimental test of prediction, 292–294

optimal prey model, 286–287

experimental testing, 289–290
formulating predictions, 287–289
testing, 287–290

organizing effects, 147, 150, 163

orientation reactions, 233
navigation, 234
taxes, 233–234

oscillator, 59, 60, 82, 85, 98, 99, 101, 105, 228
mechanisms, 228
overflow theory, 65, 66
oviposition, 145

pacemaker, 59, 82, 84, 93, 97–103

Pan troglodytes, 261
parse, 397
partial preference, 289
Parus atricapillus, 235
Parus major, 357, 358
Passer domesticus, 101
Passerina cyanea, 238
patch, 282, 291–296, 306, 309–311
pattern, 413–416
Pavlovian conditioning, 120–122, 124, 125
peak shift displacement, 419
perceptual learning, 198, 209, 242
period, 79, 81, 83, 84, 87, 90–92, 95, 99, 147, 155, 161, 175–177,
186, 209, 215, 438, 449
personality, 350–354
brief history of, 346–347
definition(s) of, 343–346
main objectives, 347–349

phase, 82, 84, 87, 89, 90, 92–96, 105, 115–118, 128, 174, 178,
180
advance shift, 90
angle, 115
delay shift, 90
 shift, 90, 95

phase-locked, 115–118
phase-locked cells, 115
phase-response curve (PRC), 90, 92–95
phenotypic variance, 350
pheromone, 14, 22, 27–29
Philomachus pugnax, 320
phylogenetic reconstruction, 402–404, 408
phylogenetics, 398, 402–405
phylogenetic tree, 403, 405, 415
phylogeny, 97, 99–101, 398, 402–405, 416, 427
Physalaemus pustulosus, 418
pigeons (Columbia livia), 239
Poecile atricapillus, 285
Poecile montanus, 388
polecat (Mustela putorius), 186
polyandrous species, 317
polyandry, 318–320
polygynous species, 317
polygyny, 318–321
threshold model, 319

pons, 117
Porichthys notatus, 143, 159
Porpodiceps cristatus, 400
pragmatics, 389–390
predispositions, 173, 179, 182–187, 191, 480
 in chicks and human infants, 183–186

preformation, 168
primacy effect, 211
primate phylogeny, 427–429
priming, 55–57
principle of parsimony, 22, 403
proactive interference, 212, 213
proboscis, 120
producer-scrounger game, 304–305
psychological construction, 70
punishment, 201

quail (Coturnix japonica), 143

quoll (Dasyurus hallucatus), 257

Rattus norvegicus, 143

reactivation treatments, 217
reafferent stimulation, 229
recency effect, 211
receptive fields, 118
red knots (Calidris canutus), 238, 296f
red-sided garter snakes (Thamnophis sirtalis), 161
red squirrels (Tamiasciurus hudsonicus), 357
reference memory, 209, 214–215
forgetting of items in, 215–217

rehearsal aids, 213

reinforcement, 205
reinforcer, 200
releasing mechanisms, 13
reliability, 370–371
repeatability, 350
representations, 228
reproduction
alternative reproductive phenotypes, 162
alternative strategies of, 160–162
associated vs. dissociated reproductive cycles, 161–162
parthenogenesis, 161–162
sex change and successive hermaphroditism, 161

Rescorla-Wagner Model, 207

retention interval, 210
retrieval, 216
retroactive interference, 212, 213
retrograde amnesia, 216
rhythm, 82
amplitude, 84
parameters, 82–84

ring dove (Streptopelia risoria), 152

ritualization, 400
roe deer (Capreolus capreolus), 357
ruff (Philomachus pugnax), 320

scrub jays (Aphelocoma californica), 232

search image, 25, 385
seasonal affective disorder (SAD), 104
seasonal rhythmicity, 102–105
circannual clocks, 103
photoperiodism and interval timers, 104–105

selective attention, 39
sensitive periods, 186–187
sensory bias, 368
sensory exploitation, 417
sensory maps, 40
sensory modalities, 21–22
sensory preconditioning, 204
sensory receptors, 14
sensory substitution, 37, 40, 41
sensory systems, 12, 14, 41, 119, 227, 245, 368
sequential prey encounter model, 287
serial position curve, 211
servals (Felis serval), 261f, 262
sex ratios, 334–336
sexual conflict, 330
sexual differentiation
birds, 148–149
brain correlates, behavior, 150
epigenetic mechanisms, brain, 152
genetic effects, 150–151
in human behavior, 460–461
mammals, 147–148
 other vertebrate classes, 149–150
steroid action, 146–147

sexual ornamentation, costs, 330–333

sexual selection, 6, 316
chase-away model, 330
and sensory exploitation, 417–418

short-term memory, 209–211

signal reliability, 370–371
signals, 368
sign-stimuli, 15–22
configurational sign-stimuli, 17–22
innate releasing mechanism, 26–29
neural processes, 26–37
relational/combinatorial principle of, 21–22
sensory modalities, 21–22
visual feature detection, 29–37

sign-tracking, 199
simultaneous ambivalence, 62
singing crickets (Gryllus integer), 303
Social Darwinism, 462
social learning, 239–242
observational learning, 239–242

sociobiology, 463–467
soft-shelled shrimp (Crangon crangon), 295
songbirds (oscines)
multiple brain sites, steroid action, 159–160
 singing in, 159–160

song production, 129–130

song types, 375
spatial discrimination learning, 202
spatial memory, 235–239
cognitive maps, 236–237
migration, 237–239
path integration, 236

starlings (Sturnus vulgaris), 101, 159, 238

stimulus control, 202, 208
stimulus perception, male sticklebacks, 15–17
stimulus reception, 13
receptor cells, 14–15
sensory worlds, 14–15

stimulus recognition, 25
stimulus selection
behavioral meaning on motivation, 25–26
behavioral ways of, 24–26

stimulus summation, 23–24

Streptopelia risoria, 152
striding biped, 437–442
Strigops habroptilus, 257
Sturnus vulgaris, 101, 159, 238
subjective day, 90
subjective night, 90
successive ambivalence, 61
supernormal stimulus, 24, 228
suprachiasmatic nucleus, 88
survival circuits, 70
survival value, 4
swamp sparrows (Melospiza georgiana), 183
swordtail fishes (Xiphophorus helleri), 418
Syncerus caffer, 262
syntax, 387–389

Taeniopygia guttata, 130, 151, 177, 358

Tamiasciurus hudsonicus, 357
Tamias sibiricus, 357
Tamias striatus, 292
taxonomy, 402
testosterone, 137
Thamnophis sirtalis, 161
theory of mind, 244, 429, 479–481
thermoreception, 13
three-spined sticklebacks (Gasterosteus aculeatus), 358
tonic immobility, 267
tonotopical, 117
topi (Damaliscus lunatus), 382
topographic maps, 40
túngara frogs (Engystomops pustulosus), 382
túngara frogs (Physalaemus pustulosus), 418
Tyto alba, 112
ubiquitin, 125
ultradian rhythmicity, 105–106
“ultradian” rhythms, 79
unconditional (or unconditioned) response, 199
unconditioned stimulus (US), 121, 199
Uria lomvia, 86

vacuum activity, 58
Vandenbergh effect, 153
variable interval schedule, 201
variable ratio schedule, 201
variables correlations, 413–416
ventral processing stream, 38
vervet monkeys (Cercopithecus aethiops), 242
visual cortex, 40, 187
visual feature detection, 29–37
in amphibians, 29–37
brain structures, feature detection, 32–33
configurational object perception, 33–34
eye and brain, 31–32
features-relating-algorithm, configurational perception, 30–
31
multimethodological analysis, 29–37
sensorimotor codes, 36–37
size constancy phenomenon, 34–35
species-specific feature detection, learning, 36
 toad’s visual system, 37
visuomotor access, 35

visual imagination, 39, 40

visual perception
imagination, 39–40
primate cortex, 37–41

visual signals, 367, 381, 475

vocal learning, 386–387
vocal motor control, 126–129
vocal motor memory, 130–132
voles (Microtus ochrogaster), 347

waggle dance, 87, 373, 385

wasp, 223–225
water striders (Aquarius remigis), 357
waveform, 84, 115
Weber’s Law, 419, 420f
well-being, 271
Whitten effect, 153
wild chimpanzees (Pan troglodytes), 261
willow tits (Poecile montanus), 388
working memory, 209–211
forgetting of items in, 211–213
versus reference memory, 217

Xanthopan morganni praedicta, 399

Xenopus laevis, 159
Xiphophorus helleri, 418

zebra finches (Taeniopygia guttata), 130, 151, 177, 177f, 358

zebrafish (Danio rerio), 269
Zeitgeber, 82
Zeitgedächtnis, 87
Zenaida macroura, 239
Zugunruhe, 238
WILEY END USER LICENSE AGREEMENT
Go to www.wiley.com/go/eula to access Wiley’s ebook EULA.