Perspectives in Ethology Volume - VVAA PDF

PERSPECTIVES
IN ETHOLOGY
Volume 2
CONTRIBUTORS
John Archer
The University of Sussex
Ethology Group, School of Biological Sciences
Brighton, Sussex, U.K.
Marc Bekoff
Department of Environmental, Population, and Organismic Biology
Ethology Group
University of Colorado, Boulder, Colorado
Paul Byers
Teachers College
Columbia University, New York, N. Y.
Robert M. Fagen
Provisional Department of Ecology, Ethology, and Evolution
The University of Illinois, Urbana, Illinois
Han Golani
Department of Zoology
Tel-Aviv University, TelAviv, Israel
J. Hanby
MRC Unit on the Development and Integration of Behavior
Madingley, Cambridgeshire, U.K.
H. Ronald Pulliam
Department of Ecology and Evolutionary Biology
University of Arizona, Tucson, Arizona
Wolfgang M. Schleidt
University of Maryland, College Park, Maryland
Nicholas S. Thompson
Department of Psychology
Qark University, Worcester, Massachusetts
A Continuation Order Plan is available for this series. A continuation order will bring
delivery of each new volume immediately upon publication. Volumes are billed only upon
actual shipment. For further information please contact the publisher.
PERSPECTIVES
IN ETHOLOGY
Edited by
P. P. G. Bateson
Sub-Department of Animal Behaviour
University of Cambridge
Cambridge, England
and
Peter H. Klopfer
Duke University
Durham, North Carolina
Volume 2
PLENUM PRESS • NEW YORK AND LONDON

Library of Congress Cataloging in Publication Data
Bateson, Paul Patrick Gordon, 1938-
Perspectives in ethology.
Includes bibliographies and indexes.
1. Animals, Habits and behavior of. I. Klopfer, Peter H., joint author. II. Title.
[DNLM: 1. Ethology - Essays. QL751 P467'j
QL751.B188 591.5 73-79427
ISBN-13: 978-1-4615-7574-0 e-ISBN-13: 978-1-4615-7572-6
DOl: 10.1007/978-1-4615-7572-6
©1976 Plenum Press, New York

Softcover reprint of the hardcover 1st edition 1976
A Division of Plenum Publishing Corporation
227 West 17th Street, New York, N.Y. 1O011
United Kingdom edition published by Plenum Press, London
A Division of Plenum Publishing Company, Ltd.
Davis House (4th Floor), 8 Scrubs Lane, Harlesden, London, NWI0 6SE, England
All rights reserved
No part of this book may be reproduced, stored in a retrieval system, or transmitted,
in any form or by any means, electronic, mechanical, photocopying, microfilming,
recording, or otherwise, without written permission from the Publisher
PREFACE
In the preface to the first volume of this series we set out our aims, which
were to encourage fresh perspectives in ethology and provide a forum for
new ideas. We still feel that in the perfectly proper search for high stan-
dards of evidence, methodology has tended to remain the master rather than
the servant of most aspects of ethological work. It is easy for us all to forget
that the kinds of data we collect are largely determined by the kinds of
questions we ask. Even an ethologist with the professed goal of providing a
straightforward account of behavior must incorporate into his or her
descriptions a great many assumptions about the organization of that be-
havior. Inevitably some facets of what went on will have been selected at the
expense of others. This is sometimes done, for example, in the service of a
theory that the fundamental unit for description is the fixed action pattern.
Our point is not that constraints on the collection of data are bad but that
the theory which gives rise to the selection of evidence should not be
neglected.
In the first volume, the choice of topics and authors was based upon
our views about the exciting or developing issues in ethology. This volume
represents a more opportunistic approach: the articles were selected from
among the many offered to us as best conforming to our aims. Neverthe-
less, certain themes do emerge. It is particularly appropriate, we believe,
that fresh thought should now be brought to bear on what methods of
description are appropriate for ethology.
In the third volume, which is now in preparation, we intend to take
social behavior as our theme, for there has been a remarkable resurgence of
thinking about this subject. We have already commissioned some articles,
but once again we invite the submission of manuscripts, as developed in
subsequent volumes.
P. P. G. Bateson
P. K. Klopfer
v
CONTENTS
Chapter 1
SOCIOSEXUAL DEVELOPMENT IN PRIMATES
J. Hanby
I. Abstract .............................................. .
II. Introduction .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
A. Definitions.......................................... 3
B. Age Categories ...........,........................... 7
III. Infancy ................................................ 9
A. Thrusting and Mounting .............................. 9
B. Intromission ........................................ 12
C. Orgasm ............................................ 13
D. Partner Selection .................................... 13
E. Presentation Patterns and Genital Touching. . . . . . . . . . . . .. 15
F. Restricted Rearing ................................... 21
G. Hormones.......................................... 24
H. Conclusion.......................................... 25
IV. Juvenile Period. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 25
A. Sequential Mountings ................................ 25
B. Seasonal Patterns .................................. .. 26
C. Partner Selection .................................... 26
D. Posturing ............................ :.............. 27
E. Contexts............................................ 29
F. Interference and Imitation. . . . . . . . . . . . . . . . . . . . . . . . . . . .. 29
G. Restricted Rearing .................. . . . . . . . . . . . . . . . .. 30
H. Conclusion.......................................... 31
V. Adolescence and Adulthood ............................. " 32
A. Male Group Change, Status, and Hormones ............. 32
B. Dominance and Mating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 34
C. Mature Copulatory Patterns in Males ................... 36
D. Other Adult Sociosexual Patterns ...................... 37
E. Male-Male Sexual Behavior. . . . . . . . . . . . . . . . . . . . . . . . . .. 40
vii
viii Contents
F. Incest and Cross-Species Copulation ........ . . . . . . . . . . .. 41

G. Old Age .................. '" ... . .... .... ... . . ...... 42
H. Restricted Rearing: Males. . . . . . . . . . . . . . . . . . . . . . . . . . . .. 43
I. Female Behavior: General ..... . . . . . . . . . . . . . . . . . . . . . . .. 45
J. Female Mounting.................................... 46
K. Female Presenting ................................... 48
L. Female Orgasm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 48
M. Restricted Rearing: Females. . . . . . . . . . . . . . . . . . . . . . . . ... 49
VI. General Discussion ...................................... 50
A. Comparisons between Nonhuman Primates and Humans. .. 50
B. Research Suggestions ............................... " 55
C. Conclusion.......................................... 59
VII. Acknowledgments ....................................... 60
VIII. References ............................................. 61
Chapter 2
HOMEOSTATIC MOTOR PROCESSES IN MAMMALIAN
INTERACTIONS: A CHOREOGRAPHY OF DISPLAY
lIan Golani
I. Abstract ............................................... 69
II. Introduction .... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 70
III. The Dimensions .................... ~ . . . . . . . . . . . . . . . . . . .. 72
IV. The Splicing of the Flow ................................ " 76
A. Definition of an Actual Elementary Movement ........... 76
B. The Necessity for At Least Two Descriptions. . . . . . . . . . . .. 76
C. The Possibility of Several Descriptions .................. 79
D. Anatomy and Movement.... . . . . . . . . . .. ... . . . . . . ... ... 81
V. The Connectedness of the Flow. . . . . . . . . . . . . . . . . . . . . . . . . . .. 84
A. Temporal Connectedness. . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 85
B. Spaciotemporal Connectedness ....................... " 89
C. Anatomical Connectedness ............................ 114
VI. Further Hypotheses and Future Fields of Study .............. 119
VII. Togetherness within Shared Routes of Convergence ........... 127
VIII. Summary .............................................. 128
IX. Acknowledgments ....................................... 129
X. References ............................................. 130
Chapter 3
BIOLOGICAL RHYTHMS AS INFORMATION CHANNELS
IN INTERPERSONAL COMMUNICATION BEHAVIOR
Paul Byers
I. Introduction ............................................ 135
II. Brain Rhythms .......................................... 136
Contents ix
III. Other Rhythms .......................................... 138

IV. Information Processing ................................... 142
V. Research Procedures and Data ............................ 144
VI. Summary and Discussion ................................. 159
VII. References ............................................. 162
Chapter 4
ANIMAL PLAY: PROBLEMS AND PERSPECTIVES
Marc Bekoff
I. Abstract ............................................... 165
II. Introduction ............................................ 165
III. Theories of Play ......................................... 167
A. The Surplus Energy Theory ............................ 167
B. The Practice Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 168
C. The "Arousal" Theory of Play ......................... 172
IV. The Characterization of Social Play ........................ 173
A. How Has Play Been Characterized? . . . . . . . . . . . . . . . . . . . .. 173
B. The Analysis of Temporal Sequences. . . . . . . . . . . . . . . . . . .. 173
C. Are Actions Exaggerated When Performed During Play? . .. 175
D. Play and Anthropomorphism: Can Play Be Characterized as
Being Pleasurable? ............................... 176
V. The Communication of Play Intention ...................... 177
VI. Conclusion ............................................. 182
VII. Acknowledgments ....................................... 182
VIII. References ............................................. 183
Chapter 5
EXERCISE, PLAY, AND PHYSICAL TRAINING IN
ANIMALS
Robert M. Fagen
I. Abstract ............................................... 189
II. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 189
III. Physical Exercise in Animal Play .......................... 190
IV. Exercise Physiology and Training Responses . . . . . . . . . . . . . . . .. 191
V. Consequences of the Physical Training Hypothesis ............ 197
A. Why Playful Rather than Serious Forms of Activity? ...... 197
B. Age Dependence ..................................... 199
C. Play Deprivation and Exercise Deprivation. . . . . . . . . . . . . .. 205
D. Warmup ............................................ 207
E. Play with Familiar Objects ............................ 207
F. Why Only Homoiotherms? ............................ 208
x Contents
VI. Aspects of Play Which the Physical Training Hypothesis Alone

Does Not Explain ................................... 209
A. ManipUlative Play ................................... 209
B. Social Play ......................................... 210
C. Diversive Play . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 210
VII. Summary of Predictions .................................. 212
VIII. Acknowledgments ....................................... 213
IX. References ............................................. 213
Chapter 6
MY DESCENT FROM THE MONKEy ....................... 221
Chapter 7
THE ORGANIZATION OF AGGRESSION AND FEAR IN
VERTEBRATES
John Archer
I. Abstract ............................................... 231
II. Introduction ............................................ 232
III. Conditioned Attack and Fear Behavior .................... " 236
IV. Situations Evoking Unconditioned Attack or Fear Behavior .. " 238
A. Pain ................ " ...... '" ................... " 238
B. Individual Distance Intrusion .......................... 239
C. "Territory": The Presence of a Novel Object in a Familiar
Situation ....................................... 240
D. Presence of an Unfamiliar Environment ................. 244
E. Familiar Object in an Unfamiliar Place .................. 245
F. Frustrative Nonreward ................................ 246
G. Thwarting .......................................... 246
H. Low Reinforcement Schedules ......................... 247
I. Other Situations ..................................... 248
1. Other Types of Responses Evoked by these Situations ..... 249
V. Interaction of Situations Evoking Attack and Fear Behavior ... 250
VI. Common Properties of Situations Evoking Attack and Fear
Behavior ........................................... 251
VII. Internal Factors Affecting Decision Process 1 ................ 258
A. Degree of Discrepancy between Observed and Expected
Events ..... , .................. , .............. '" 258
B. Strength of Ongoing Motivation ........................ 261
C. Hormonal Factors ................................... 262
D. Experiential Factors .................................. 271
E. Interaction of Variables Affecting First Stage of Decision
Process ......................................... 274
Contents xi
VIII. External Factors Affecting Decision Process 2 ............... 275

A. Attack and Target Properties .......................... 275
B. Two Types of Fear Behavior ........................... 278
C. Blocked Escape Behavior. . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 279
D. Other Situational Factors ............................. 279
IX. Discussion.............................................. 280
X. Acknowledgments ....................................... 284
XI. References ............................................. 284
Chapter 8
ON INDIVIDUALITY: THE CONSTITUENTS OF
DISTINCTIVENESS
I. Abstract ............................................... 299
II. Definitions ............................................. 300
III. The Geometric Model of Distinctiveness .................... 303
A. Uniform Distribution ................................. 304
B. Normal Distribution .................................. 304
C. General Considerations ............................... 306
IV. Discussion.............................................. 307
V. Summary .............................................. 309
VI. Acknowledgments ....................................... 309
VII. References ............................................. 309
Chapter 9
THE PRINCIPLES OF OPTIMAL BEHAVIOR AND THE
THEORY OF COMMUNITIES
H. Ronald Pulliam
I. Abstract ............................................... 311
II. Introduction ............................................ 312
III. Models of Optimal Diet .................................. 314
IV. Optimal Models of Habitat Selection . . . . . . . . . . . . . . . . . . . . . .. 318
V. Models of Optimal Sociality .............................. 321
VI. Speculations and Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 328
VII. References ............................................. 331
INDEX ...................................................... 333

Chapter 1
SOCIOSEXUAL DEVELOPMENT IN
PRIMATES
J. Hanby
MRC Unit on the Development and Integration oj Behavior
Madingley. Cambridgeshire. u.K.
I. ABSTRACT
The development of sociosexual patterns in nonhuman primates is traced

from infancy through adulthood. Recent published and unpublished studies
on a number of group-living captive and free-ranging Old World monkeys
and apes show that most species develop two distinguishable sociosexual
systems: copulatory and contacting. The two systems overlap in form and in
their apparent motivational and emotional bases. The early development of
distinct patterns seems to be facilitated by group companions of both sexes
and all ages that interact with the infant, by dorsal riding and terrestrial
habits.
Restrictions on social experience and space retard the development and
integration of elements such as mounting, thrusting, intromission, and
presenting into two distinct systems. Throughout development the young
nonhuman primate learns to associate certain behaviors with certain
partners and situations. During infancy, if mothers and others are absent,
opportunities for learning are especially limited, with little improvement in
adulthood even under the best conditions.
The implications for our understanding of human sociosexual behavior
are discussed and some suggestions are made for the improvement of re-
search into this extremely important area of overlap in sexual and social
patterns.
II. INTRODUCTION
Gestures of love and hate, or affectional and agonistic patterns, have

recently received attention in a number of popular books (e.g., Eibl-
2 J. Hanby
Eibesfeldt, 1970; Lorenz, 1966; Morris, 1967; Wickler, 1972). These works
and many others (e.g., Kinsey et al., 1948, 1953) have repeatedly made use
of information on our fellow primates to support arguments and generaliza-
tions about human behavior. Sexual behavior as part of love has had a large
share of the attention, but since the sexual behavior of humans is so little
understood and prejudices so great, the literature on nonhuman primates
has been widely utilized. Unfortunately, the literature has not been espe-
cially plentiful, and human biases and limited focus make much of the work
difficult to interpret. Most knowledge about the sexual behavior of free-
ranging primates comes from studies on a few species of macaques, on ba-
boons, and-to a lesser extent-on chimpanzees. Many studies are
conducted in the laboratory and under restricted conditions, factors which
impede comparisons and generalizations owing to differences in species,
testing and rearing conditions, observers, etc.
Since generalizations are constantly being made despite the inadequacy
of our knowledge, it seems imperative that more effort be made to review
and reassess what is known about primate sexual behavior. This review is
one such attempt. Its main aim is to draw together much of the recent
published and unpublished work on the behavior of nonhuman primates that
have been reared in groups containing both sexes and all ages, and to thus
provide some standard of comparison with primates that have been sub-
jected to a variety of restricted conditions. Although comparisons with hu-
mans must always be made with caution, an underlying assumption of this
review is that our understanding of our own behavior can be greatly
increased by considering what characteristics we do or do not share with
other primates.
In all primates there appears to be a significant overlap between sexual
patterns and behaviors that are common in a wide variety of other situa-
tions. For instance, an embrace can occur between a male and female in a
copulatory context, but it can also occur between any two monkeys, apes,
or humans of whatever age or sex, and between kin, friends, or strangers, in
a wide range of contexts: reunion, greeting, play, or even as part of a strug-
gle. While a given form of behavior may be similar in different species or
situations, the interpretation of the behavior varies from sexual to affilia-
tive. In this review I am especially concerned with this difficult area of
overlap between behaviors that are part of the copulatory pattern and ap-
pear in other patterns as well. Such behaviors are termed "sociosexual" and
occur in all group-living primates studied to date. 1
1 Only the Old World monkeys and apes are considered in this review, as they have received the
most attention and the data on them are more adequate by far than those on New World
monkeys and prosimians. The work of Ploog (e.g., Ploog and Maurus, 1973) on squirrel
monkey penile displays deserves special mention though it is not reviewed here.
Sociosexual Development in Primates 3
Sociosexual behaviors such as mounting, presenting, genital touching,

embracing, etc., are not only widespread but quite variable in appearance
and situation. The outline drawings which illustrate this review give the
reader a general idea of the variety of postures and social contexts in which
these behaviors appear. All the drawings have been made from
photographs, motion picture film, or life.
Posturing is only one attribute of sociosexual behaviors that I shall
consider. This review is structured such that form, onset, frequency, pat-
terning, social context, and partner selection are all given some attention in
each of the three sections that trace sociosexual development (infancy, the
juvenile period, adolescence, and adulthood). Also, at each level of develop-
ment, the effects of restricted rearing conditions are evaluated. The main
thesis argued throughout is that nonhuman primates develop two distin-
guishable patterns, the copulatory and the affinitive or contacting. The two
patterns often share common elements (such as mounting) but these develop
along different lines at different stages. Both patterns are necessary for the
individual to cope successfully with the requirements of group living, mat-
ing, and child rearing.
Following the review of sociosexual development in nonhuman pri-
mates, some extrapolations to and comparisons with humans are discussed.
The last section consists of research suggestions. At the outset, some terms
must be defined and the age categories delineated.
A. Definitions
Sociosexual: Behavior elements or patterns that are usually present in

the heterosexual copulatory context and appear in other social situations as
The references cited here are not comprehensive because reports are always increasing.
Rhesus monkeys Macaca mulatta: Maslow, 1936; Carpenter, 1942; Chance, 1963; Altmann,
1962; Hinde and Rowell, 1962; Loy, 1969; Alexander and Perachio, 1970; stumptail monkeys
M. speciosa: Bertrand, 1969; Chevalier-Skolnikoff, 1971a, 1974a; crab-eating monkeys M.
irus: Schloeth, 1956-7; Japanese monkeys M. fuscata: Yamada, 1966; Hanby and Brown,
1974; bonnet monkeys M. radiata: Simonds, 1965; Rahaman and Pathasarathy, 1968; pigtail
monkeys M. nemestrina: Bernstein, 1967; Tokuda et al.. 1968; Barbary "apes" M. sylvana:
Deag, 1974; Burton, 1972; hamadryas baboons Papio hamadryas: Kummer, 1968; savanna,
yellow, etc. baboons P. ursinus. anubis. papio. or cynecephalus: Bolwig, 1959; Maxim and
Buettner-Janusch, 1963; Hall and DeVore, 1965; Rowell, 1967b; Owens, 1973; Vervet monkeys
Cercopithecus aethiops: Struhsaker, 1967; Gartian, 1969; langurs Presby tis entellus: Jay, 1965;
Sugiyama et al.. 1965; chimpanzees Pan troglodytes: Bingham, 1928; Crawford, 1940; Galt,
1947; Reynolds and Reynolds, 1965; Goodall, 1968a,b; Kellar et al.• 1968; McGinnis, 1973; By-
gott, 1974. Further references can be found in Wickler (1967), who has also reviewed the ap-
pearance of sociosexual signals in humans. (Throughout this review I use common names
instead of species names, but the foregoing list has been presented in part for reference
purposes.)
4 J. Hanby
well. Only a few of the possible behaviors that could be included in this
broad category are discussed in any detail in this review.
Presentation: Standing still or moving backward with the buttocks
oriented to the recipient, usually with the head turned toward the partner
and, if there is one, the tail turned away from the genital area. There are
many variations of the presentation posture, some of which can be seen in
Figs. 1,3,4,5.
Mounting: Any position or posture which involves one individual
climbing (wholly or partly) onto another. Some variations can be seen in
Figs. 2, 6d, 7, 9, 10. The particular posture should be made explicit because
mount postures are not necessarily equivalent to copulatory postures. This
is most easily seen in chimpanzees; the copulatory posture typically does
not include the male resting his weight on the female (see Figure 2c).
In some species, such as Japanese, rhesus, and Barbary macaques and
hamadryas and chacma baboons, copulation typically involves a series of
(a) (b)
Fig. 1. Some presentation patterns. (a) An adult female anubis baboon presents to another
adult female and her recent infant (from a photograph in Hall and DeVore, \965). (b) A male
hamadryas baboon "notifies" his group leader of the direction of his departure (from a
photograph in Kummer, 1968). (c) A chimpanzee female approaches and presents her bottom
to an adult male who gently extends a hand to touch it (from a photograph by D. Bygott). (d)
A hamadryas female presents to her harem leader during an altercation with another female not
drawn here (from a photograph in Kummer, \968).
Fig. 2. Some copulatory postures. (a) Adult male and female langurs (from a photograph in
Jay, 1965). (b) Adult male and female anubis baboons (from a photograph in McGinnis, 1973).
(c) Adult male and female chimpanzees. (d) Adult male and female gorillas (from a
photograph by D. Sorby).
mounts. In other species there may be only one mount with varying num-
bers of thrusts. Comparisons between species and also between copulatory
and other types of mounting situations are facilitated by use of the term
mount event, which means an interaction between specified individuals (age,
sex, etc.) involving one or more mounts and terminating when mounting
stops or the partners separate. Contingent behaviors (e.g., thrusting, in-
tromission), postures, the situation, etc., are specified.
Copulation: Essentially penile intromission, either anally or vaginally,
with or without ejaculation, and in any position (Fig. 2). The term is
synonymous with coitus (or coition) and sexual intercourse.
Ejaculation: Usually recognized by semen or a "vaginal plug" (coagu-
lated semen) and various behavioral elements, such as an increased rate in
thrusting or manipulation of the penis, followed by a pause and body rigor.
Orgasm: Refers to behavior that resembles the male ejaculatory pat-
6 J. Hanby
tern (pausing after a period of thrusting or genital stimulation) even though

there may not be any reliable external or known internal signs.
Mating: Includes copulation and the pre- and postcopulatory behavior
as well.
Consorting: Refers to a persistence in mating by a particular pair.
Estrus: Refers to behavior and does not presume hormonal state. The
term's meaning has been confused in the literature (Rowell, 1972) and
seems little more relevant to nonhuman primates than it would be to the
human female, though it is useful when there are limited periods of seeking,
giving, or receiving sexual attention.
Beach (1965) has discussed the varied meanings attributed to the term
sexual behavior; it is used here in the narrow sense of involving prolonged
genital contact (autoerotic ally or with a partner). In as many cases as possi-
ble, the behavior elements themselves are referred to.
In this review the various terms are used as defined above, since all too
often it is not clear what they mean in the literature. For example, it is
seldom made explicit whether copulation means mating, mounting, consort-
ing, or intercourse, or whether ejaculation is a necessary condition.
Most sociosexual patterns are seen between males and females, though
homosexual interactions are widespread in mammals. Unfortunately, since
the emphasis and usage has been on the sexual part of the word, I shall use
precise if boring terms such as male-male mounting, female-female
presenting, heterosexual copulation, etc.
Although the bulk of this review concerns actual behavior, there is no
way to avoid problems with terms that imply motivational state simply be-
cause of the structure of language. "Sexual" refers both to behavior and
presumed causal factors such as hormones and to the motivations and emo-
tions of those who behave. As humans we invest the terms and the behavior
with many meanings according to our biases. A bias that is apparent in the
present review is that sociosexual gestures have a variety of motivations and
therefore the causes for any act can only be understood in terms of the im-
mediate social context and early rearing, which are at least as important as
hormonal state.
Two especially bothersome and emotionally-loaded terms are domi-
nance and submission. Mounting is usually taken to be an expression of
dominance (see, for example, Chevalier-Skolnikoff, 1971 b) and presenting
or accepting mounts to be an expression of submission. In most marriage
manuals, the male atop is called superior and the female's position inferior,
or dominant and subordinate terms are used. When referring strictly to
position these terms may be understandable, but most often they connote a
more general value judgment. I use dominance here to denote one aspect of
a relationship between individuals: their pattern of agonistic interactions. If
A hits Band B never or seldom reciprocates, then A is dominant to B or B

is subordinate. Submission refers to active response to an aggressive action
by another individual. If A approaches B with arm raised and B cowers, B
is acting submissively.
This review shows that few sociosexual gestures are obviously
expressions of dominance or submission, for they rarely occur in agonistic
settings. More often they precede peaceful or friendly interactions, and
might be termed gestures of deference or invitations to interact. The moti-
vation for the gesture will always remain obscure though we can gain clues
by examining the contexts as well as our own experience.
A term that seems nearly impossible to define is that of "normal" be-
havior, and yet it is constantly used. "Normal" does not usually mean the
most prevalent behavior (e.g., masturbation is rare but is considered
normal). Nor is the term equivalent to "natural" although it is sometimes
used in that sense. The term can only refer to behavior in a specific situa-
tion: what is "normal" in a prison, zoo, or laboratory may not be "normal"
under free-ranging conditions. Most laboratory studies tend to use the be-
havior of wild caught or feral animals as the standard of normality in the
laboratory situation. Many of these animals are caught as juveniles or even
with their mothers while very young and are raised or kept in the labora-
tory. Since environmental conditions and early experience can greatly affect
behavior, it seems that the assumption that these captive, feral-born animals
exhibit "normal" behavior must be questioned.
Another term with assorted usages and vague meanings is "adequate"
behavior. Many reports on the effects of early rearing use this term, and it
even seems that usage has shifted subtly from "normal" to "adequate."
Perhaps this reflects an understanding that while certain behaviors may be
adequate for living in a captive environment that allows handlers and ex-
perimenters to deal with the creatures, these behaviors may not be the norm
in the wild. Likewise, behavior adequate to allow two individuals to copu-
late in some manner may well have nothing to do with the behavior
necessary for rearing a child that will become and remain an integral
member of a social group. As people, we are especially concerned with how
primates learn to relate, not just procreate.
B. Age Categories
Harlow has described stages in heterosexual sexual development and

insists that ages are irrelevant (e.g., Harlow, 1965). It is increasingly ap-
parent that ages are relevant, and certain kinds of experience at certain ages
are absolutely necessary for mature sociosexual expression (e.g., Harlow
8 J. Hanby
and Harlow, 1969). In the following review reference is made to infancy, the
juvenile period, puberty, adolescence, and adulthood. The age ranges to
which these terms refer are difficult to delineate precisely, partly due to dif-
ferent authors' usages, but mainly because age is a continuum on which
periods or stages are imposed. Hinde (1971) has reviewed reports that refer
to age periods and has presented tables that summarize mother-infant and
behavioral developments. The reader is referred to Hinde's article and to
Chalmers (1972) for the changes in the mother-infant relationship over time
for several species.
In this paper I shall use the following terms: Infancy for macaques and
baboons refers to the period from birth to approximately 1-1 Y2 years. This
period can be subdivided into two stages, the differentiation occurring at
about 3-5 months, or waning of coat color change (see Deag, 1974; Ber-
trand, 1969). Infancy in chimpanzees and humans covers up to about 3
years. The juvenile period covers the period between infancy and the onset
of puberty2 (e.g., approximately 1Y2-2 or 3 years for macaques and baboons,
3-6 or 7 years for chimpanzees, and 3-12 years for humans). Adolescence or
subadulthood (a more neutral term) means the period between puberty onset
and full adulthood (= maturity) or full body size, weight, and coloration,
stability of sexual cycles and behavior (approximately 3-5 years for macaque
and baboon females; 4-7 years for males; 7-11 or 14 years for chimpanzees;
and 12-18 or 21 years for humans; sex differences on reaching maturity in the
great apes are not yet clear enough to state here). The term immature refers
2 Puberty means onset of "sexual maturity," i.e., capability to reproduce. It seems increasingly
clear, by the contrast between estimates of puberty onset in captivity (e.g., van Wagenen,
1972) and in the wild, that either primates reach puberty earlier in captivity or the criteria for
puberty are different in the different situations. Probably both factors operate to produce dif-
ferent estimates. Weight gain and nutritional differences between the two situations seem to
be basically responsible for differences in puberty onset.
To determine actual onset of puberty requires, of course, actual hormonal measurement. This
is not easy to do, and for the sake of more reliability across laboratory and field studies it
may be better to use regular cycle and consistent ejaculation as the criteria for puberty or
adolescence rather than the first appearance of menses or ejaculation. Hopefully, stable
external signs correspond with stable hormonal levels. (For instance, Dierschke et al.. 1974,
found that although rhesus females showed menses, they did not respond to injections of
estrogen with a surge or show spontaneous surges of gonadotropin until 4-8 months after
menarche.)
Onset of puberty also varies with species: seasonal breeders such as the rhesus, Japanese, and
Barbary macaques may reach puberty in the Y2 year (e.g., 3V2 female; 4Y2 males) whereas
nonseasonal species may show more variable onset. Still another problem is "adolescent
sterility," or the fact that in macaques, chimpanzees, and human societies, adolescents may
copulate but not reproduce; hence the criteria of capability to reproduce can be questioned.
In any case, external signs that are related to stable hormonal levels and can be agreed upon
are probably better criteria than potentialities or rare occurrences.
to any age less than full adulthood. I must stress that these ages and periods
are approximate and are impositions on a continuum.
III. INFANCY
This period is characterized. by initially complete dependence on a

mother or caretaker, developing to relative independence and increasing
interaction with others. During infancy the behavioral elements of thrusting,
mounting, presenting, and, for males, intromission seem to be integrated
into definite patterns, which depend on the availability of partners. I
consider in this section the integration and onset of the elements and their
patterning.
A. Thrusting and Mounting
The newborn monkey or ape is in almost constant contact with the

mother's body. From a very early age (around I month old) penile erections
in the male (stumptail macaques, Chevalier-Skolnikoff, 1971b; Japanese,
rhesus, and Celebes macaques, personal observation) and thrusting in both
sexes are directed to the mother's body, especially when the baby is awake
and lively and the mother is relaxed. This also seems to be true of humans;
Lewis (1965) describes pelvic thrusting that occurs under conditions of
maximum security in moments of delight and decreases with the decrease of
intimate holding contact.
When young infants become mobile (within the first few days or weeks
in monkeys) and thereafter, the infant frequently makes "cuddle" move-
ments upon re-establishing contact with the mother's ventrum. Often these
movements seem to involve thrusting. In macaques, baboons, and chimps,
thrusting appears independently of mount attempts, but as the infant starts
to leave its mother's ventral surface, thrusts become directed to other parts
of the body and the infant begins to mount.
Mounting seems to involve boarding the mother or another individual
by whatever route is available. Mounting can be initiated by the baby or
elicited by a variety of usually subtle mount gestures made by the mother or
others (Fig. 3). Babies in many species are carried dorsally, and mO)lnt at-
tempts with thrusting are seen as the infant boards his partner, who may not
necessarily be the mother. (Japanese macaques, Hanby and Brown, 1974;
stumptail macaques, Bertrand, 1969; Barbary macaques, Burton, 1972;
olive or anubis baboons, Owens, 1973; chimpanzees, Goodall, 1968a).
10 J. Hanby
(a)
'-----""~~~ (b)
Fig. 3. Presenting by mothers and boarding mounts by infants. (a) Chimpanzees (from
Goodall, 1968b). (b) Japanese monkeys (from life).
Dorsal riding (Fig. 4) in nonhuman primates may facilitate the

development of the correct orientation of the mount to the rear. Correct
positioning helps bring the genital area of the infant into contact with the
partner's genitals, and promotes linking of the mount with thrusting. Also,
the infant learns to respond to various mount and dismount gestures. Dorsal
riding is not absolutely necessary for correct mount posturing or for the
linking of thrusting to mounting, since free-living adults of other species
exhibit rear-oriented mounts with thrusts even though the infants are car-
ried ventrally or in other positions until independence. A few species have
not been observed to thrust on mothers or others at all frequently or to
"practice" mounting (e.g., gorillas: Fossey, personal communication;
orangutans: MacKinnon, personal communication; gibbons: Chivers, per-
sonal communication).
Nevertheless, it may be significant that dorsal-riding species seem to
show the earliest onset of properly-oriented mounting and various other so-
ciosexual patterns. These same species also tend to have the most elaborate
aggressive displays and the most varied greeting gestures, in addition to (or
as a result of) highly complex social organizations and perhaps as a result of
their terrestrial habits. Dorsal riding may simply facilitate proper orienta-
tion and early sociosexual development.
Rear-oriented mounting appears in male stumptail macaques reared in
mixed age/sex groups somewhere between the fourth and ninth week (Chev-
alier-Skolnikoff, 1971 b; Bertrand, 1969). Rhesus males have been seen to
execute rear-oriented mounts at 12 weeks of age (Hinde and Spencer-Booth,
1967), Japanese macaques at about 9 weeks (Hanby and Brown, 1974),
olive baboons by the third month (Owens, 1973), vervets at about 12 weeks
(Chalmers, cited in Hinde, 1971), talapoins at 9 weeks (Hill, 1966), langurs
at 10 months (Sugiyama, 1965), and chimpanzees at 9 months (Goodall,
1968a) and 17 months (F. Plooij, 1976). Less well oriented mounting has
been observed at younger ages in most species. Not all males start mounting
at these earliest-cited times, and the range of onset, the partners, and the
context (e.g., play or agonistic) all need to be more carefully described.
It is clear that there are differences depending on the ongoing group
setting, in onset and continuation of sociosexual behaviors. An example that
Fig. 4. Dorsal riding. (a) Hamadryas baboons (drawn from a photograph in Kum-
mer, 1968). (b) Chimpanzees (from life). (c) Japanese monkeys (from life). (d)
Anubis baboons (from life).
12 J. Hanby
needs to be followed by more data is that of the onset difference seen

between studies of chimpanzees by Goodall and Plooij. In the former study,
the chimpanzees were being provisioned and were together frequently,
which probably facilitated or even necessitated early development of so-
ciosexual mounting (and presenting as well, see below). Plooij, with a more
intensive study of infants, shows later development at a time when the chim-
panzees no longer gathered in large groups at the feeding area. Thus the dif-
ference between these two studies in age of onset may have been due to
differences between situations rather than between individual chimps or
observers.
Although females seem to thrust as often as males in the early weeks,
by the end of the first year males thrust more often and mount much more
frequently than females. When females do mount, however, they orient to
the rear of their partners and often even footclasp (rhesus, personal observa-
tion; Japanese macaques, Hanby and Brown, 1974; olive baboons, Owens,
1973). Both males and females seem to be recipients of mounts around the
time they themselves begin mounting. Besides being mounted by other in-
fants, they also have their rumps held or clasped. How this (and genital
manipulation, see below) affects their own development of like patterns is
not clear, but some imitative behavior does seem to occur.
B. Intromission
Once the correct orientation and posture have been attained, mounts
then become linked to thrusting and intromission. Intromission is difficult
to determine in field studies and even in captive conditions, and the relation
between thrusting and intromission is tenuous. For example, Hanby and
Brown (1974) found in Japanese macaques that most mounts between
partners of any age or sex were accompanied by thrusts, though in-
tromission seemed to occur only when there were deep pelvic thrusts; most
nonadult male~female thrusting was perfunctory, as though a token gesture.
Nevertheless, since thrusting is much more reliably observable than in-
tromission, one must often use thrusts as an indication that intromission has
been achieved. Once achieved, one might expect the basic copulatory pat-
tern to become more stable and frequent.
Owens (1973) has described anubis baboons achieving intromission
with adult females at around 10 months of age, followed by a sudden
increase in mounts with thrusts and an increase in total mount frequency.
Although intromission could not be observed reliably, we found that infant
Japanese macaque males thrusted on females of any age more often than
they did on males once they reached one year of age (Hanby and Brown,
1974).
Infant males thrusting and occasionally achieving intromission (espe-
cially with adult females) has been noted in other baboons (Anthoney, 1968;
Kummer, 1968); this is also true of chimpanzees (Bingham, 1928; Goodall,
1968a,b; Tutin, in preparation; McGinnis, 1973). An infant stumptail male
managed to obtain heterosexual and homosexual intromission (into the male
partner's anus) in his sixth month (Chevalier-Skolnikoff, 1971b), but since
later observations were not made, the increase or stability of these patterns
cannot be assessed.
The fact that young males are more likely to mount and thrust on fe-
males, and especially adult females, is probably due to a variety of factors:
females may be and usually do seem to be more willing to hold still; females
present and gesture for mounting more often; the female's bright pink bot-
tom in some species may be especially attractive; cycling females pre-
sumably exude pheromones (Michael and Keverne, 1970); females are often
in proximity to the mother of the infant; and most important, intromission
is more likely into the vagina of a willing adult female than that of a young
virgin female or into a male's anus. The availability of these partners un-
doubtedly makes a difference to infant males.
C. Orgasm
While thrusting, mounting, and intromission become linked for most

infant males, the element of orgasm has been described only for infant
anubis baboon males (Owens, 1973). Orgasm has also been described for
both human sexes (Kinsey et al., 1948, 1953; Bakwin, 1952, 1971, 1972).
The rarity of descriptions for nonhuman primates may be due to true
species differences, difficulty in observation, or observer bias. All infant pri-
mates thrust and receive genital stimulation, but whether there is an orgasm
or "peak experience" and the nature of its onset, form, frequency, etc.,
remain open and intriguing questions.
D. Partner Selection
The mounting of females more often than other m~Ues seems to have
long been taken as a sign of sexual maturity in males. Even so, Zuckerman
(1932), after long experience with baboons, macaques, and apes concluded:
One has to examine the sexual responses of immature monkeys and apes to discover
whether or not either sex shows a preference for heterosexual or homosexual contact.
14 J. Hanby
My own opinion, based on prolonged observation, is that neither sex, especially among
the younger immature animals, shows any preference.... (p. 288)
More recent observations do not contradict Zuckerman's conclusion

but do show that partner selection varies to a considerable extent with fac-
tors such as species and rearing conditions. Throughout the literature the
term "partner preference" is used with the criteria for "preference" left
absent or vague. It should be kept in mind that "preference" most often
means simply that an animal does something to one individual more often
than to another. This is not necessarily a preference on the actor's part; his
behavior could be provoked, for example, by the recipient of the behavior.
Infant male Japanese monkeys mounted both sexes equally until al-
most 1Y2 years of age, when females were more often the recipients of
mounts. In their study of mixed age/sex groups of rhesus, Hinde and
Spencer-Booth (1967) note that " ... although male young showed signifi-
cantly more mounting than females, there was no clear difference between
the sexes in the amount they were mounted" (p. 185). Presumably this state-
ment refers to males up to 2Y2 years of age.
The data collected by Goy and his co-workers (Goy and Goldfoot,
1973) on rhesus raised in peer groups and in groups in which the mothers
were present show that there are clear differences in partner selection de-
pending on rearing conditions. The majority of peer-raised males showed no
"preference" for female partners even at 3 years of age. On the other hand,
6 out of 8 males raised with both mothers and peers mounted females more
often than they did other males before they were 1 year of age. Preference
for play partners remained the same, namely, other males, for both groups.
Data on males of other species showing the development of selection or
acceptance of partners of like or opposite sex are very limited, and em-
phasize the variability seen in groups of rhesus and Japanese monkeys.
Evans (1967) showed that in a colony of pigtail macaques males at 6
months mounted females more often than males, but infants raised in indi-
vidual cages did not seem to show differential frequency according to sex.
An infant male stumptail reared in a small social group, studied by Chev-
alier-Skolnikoff (1974), appeared to prefer a juvenile male as a mounting,
thrusting, oral-anal-genital contact partner, though he also mounted the
other two infants present-both females-and his adoptive mother. It is im-
plied in Anthoney's (1968) descriptions of zoo colony reared infant baboons
that females with bright pink and swollen bottoms were most frequently
mounted. In chimpanzees, infant males mount swollen females more often
than they do males of any age, and also thrust on stones and tree trunks (F.
Plooij, personal communication). In a nonswelling species, the langur,
studied by Jay (1965), infant and juvenile males apparently mounted adult
males before they did females, and the mount-embrace-touching relation-

ship between older and younger males was quite striking.
The rather scanty and imperfect data make it obvious that several fac-
tors must be considered in the ontogeny of partner preference or selection
by infant males. The size and sex ratios of groups (number of "available"
partners) are critical, and the number and sex ratios of peers are also im-
portant. The species-specific patterns of greeting (e.g., langurs, Jay 1965),
notification (hamadryas baboons, Kummer, 1968, see Fig. Ib), or "tension"
reduction (e.g., macaques, Hanby, 1975) seem to be very important. Each
species-and especially those comprising terrestrial, relatively aggressive,
highly complex societies-has particular sociosexual patterns that appear to
function as contact-promoting signals. Infant males are learning two pat-
terns during this early period: (1) response to adult or other female
presentations by mounting, thrusting, and intromission; and (2) the response
to any age or sex-depending on the species and situation-with mounting,
dorsal embracing, riding, genital touching, etc. It can be concluded,
however, that a definition of the male preference for male partners as an
"immature" pattern and of the early and frequent mounting of females as a
"mature" pattern is misleading. Nevertheless, the presence of females in the
early environment seems to allow an infant a better opportunity to integrate
his copulatory pattern, while the presence of other males may facilitate the
more general sociosexual patterns.
E. Presentation Patterns and Genital Touching
Most infants are of course first presented to by their mothers as a

boarding invitation or what could also be interpreted as a "lure" to get the
infant to walk (Fig. 3). Lilyan White has reminded me that in many species
the females' bottoms are usually bright pink after birth (as they are during
pregnancy) and that the toddler will lurch toward this vivid area and reach
for and attempt to climb on if the mother remains still. Other species (e.g.,
chimps) do not have pink bottoms after giving birth and attract infants by
facing them (Plooij, 1976). I have already mentioned that the swollen, pink
bottoms of other adult females are especially attractive to infant males
(e.g., Goodall, 1968, and Anthoney, 1968, both specifically mention this for
chimpanzees and baboons respectively). Also, the bright red anal field of
the male hamadryas baboon may well help to keep his young females and
other followers oriented to him. If very young infants are not able to recog-
nize their mothers for a while, it may be very important in the wild for an
infant to readily respond to a brightly colored object (like the rear) by
16 J. Hanby
contacting and climbing attempts. This hypothesis could be tested with

models in the laboratory.
Being presented to and the appearance of the present posture by the in-
fant do not seem to be correlated. On the other hand the link between
seeing a present and responding by mounting and attempting contact is
early and close. Presenting and mounting, though obviously related in
context, appear to have different developmental sequences. While mounting
is primarily a response to a presenting gesture, presenting is interlaced with
being restrained, having the hips held, and the genitals and perineal area in-
spected (some examples are shown in Fig. 5).
Genital inspection and grooming are part of the experience of all infant
primates (humans included, see Bakwin, 1971; Ford and Beach, 1951;
Marshall and Suggs, 1971). Although some may struggle, most infants do
not resist such touches and even seem to relax and "enjoy" it. If enjoyable,
it is more understandable that infants soon present their genitals to others,
and once mobile often orient their rears to adults and playmates whether or
not their genitals are touched.
Infants of many species not only orient their buttocks to some group
members but actually walk backward into the recipient, pushing their hips
into the face. Ransom and Rowell (1972) illustrate this behavior of young
anubis baboons, and in many species (e.g., stumptails: Bertrand, 1969;
Japanese, Celebes, rhesus: personal observation) it seems more directed to
adults, especially males, than to other infants, juveniles, or mothers. The in-
fant may approach another animal in a variety of postures (e.g., buttocks
first, sideways, with footback "presents") and often the recipient will inves-
tigate the infant's genitals, touch the rump, back, or tail, or groom the in-
fant briefly.
An especially interrsting variant of infant genital investigation is that
of the Barbary macaques described by Deag (1974). In this species, the
adults use babies (Deag distinguishes babies-less than 6 months-from
older infants) as part of greeting one another. This is particularly true of
male-male interactions, and Deag shows clearly that most adult male
interactions involve babies. These interactions consist of the male not in
possession of the baby performing "rump-lifting," or picking up the hips of
the baby and sniffing or touching its rear (Fig. Sf). Females also use the
baby rump-lifting gesture; it seems to allow adults peaceful approaches to
one another, but the baby is obviously also influenced.
As well as receiving inspections and touches, especially on their bot-
toms and genitals, infants are often allowed to explore the same regions of
others. An unusual example of genital contact with others is described in
the infant stumptail male in a small captive group observed by Chevalier-
(e)
Fig. 5. Genital touching and presenting in infants. (a) A rhesus mother holds the tail of her
infant male to restrain him from leaving (drawn from a photograph in Southwick et al., 1965).
(b) An adult baboon inspects the genitals of an infant (from a photograph in Ransom and
Rowen, 1972). (c) An infant stumptail monkey puts feet on an adult (from Bertrand, 1969). (d)
An infant anubis baboon presents bottom to an adult male, putting feet and tail into contact
(from a photograph in Ransom and Rowen, 1972). (e) An infant stumptail monkey presents
with head down and looking backwards (from Bertrand, 1969). (f) "Social buffering" in Bar-
bary macaques. Two males hold an infant which one has brought over to the other; the
recipient raises the infant's bottom and inspects it (from a photograph in Deag, 1974).
18 J. Hanby
Skolnikoff (l97Ia, 1974a,b), (Fig. 6). This is nowhere so clear as in chim-

panzees. H. and F. Plooij (personal communication) vividly describe how in-
fants avidly inspect the genitals of adult males and females, especially the
pink bottoms of the females and the pink penes of the males. The infants
sniff, touch, and lick as early as 4 months, and as soon as they can walk they
become persistently curious about and attracted to copulating pairs.
Fondling of their own genitals has been described in young human in-
fants (e.g., Moll, 1912; Kinsey et aI., 1948); and Bakwin (e.g., 1952, 1972)
recounts cases of "masturbation" and erotic elements in self-directed be-
havior with indications of orgasm. Such behavior exists in other primates
but seems less directed by the infant to himself than by others to the infant.
Another difference seems to be that in humans, infant females may be the
more frequent masturbators than infant males, while in nonhuman primates
the infant males seem to give and receive more genital attention.
Genital attention to and from very young infants gradually involves less
direct contact and exploration, being replaced by more independent (no
contact) and clearer present posturing. Presenting of the bottom to another
Fig. 6. Unusual infant behavior. The infant stumptail

monkey male mounting the juvenile has an erection and
mouths the penis of this partner (drawing from Chevalier-
Skolnikoff, 1974, by permission of the author and Plenum
Publishing Corp.).
is a gesture reported to appear quite early in the terrestrial, highly social

primates even though posturing is variable and less obvious to an observer
than a mount.
The earliest full presenting in stumptails has been observed on day 22
(Chevalier-Skolnikoff, 1971b) and in the seventh week (Bertrand, 1969).
Rhesus males usually present somewhere between 25-30 weeks (Hinde and
Spencer-Booth, 1967), though L. White and I have recently observed males
clearly presenting (tail raised, looking back over the shoulder, buttocks ele-
vated) at 6 and 7 weeks. Infant rhesus females show the present posture at
13-18 weeks of age (Hinde and Spencer-Booth, 1967). This earlier onset of
rhesus male presenting in captive groups needs to be confirmed by field
studies before any conclusions can be drawn about sex differences. Infant
male and female Japanese monkeys do not so clearly present as do other
species-they more often simply hold still-yet before one year of age fe-
males present earlier and more often than males. Chimpanzee males and fe-
males begin presenting in their eleventh month (Goodall, 1968a), but a later
appearance is described by F. and H. Plooij.
The development of the present posture is given less attention than
mounting in almost all studies, yet it is at least as important to both sexual
and general social relations. Both males and females of many species
present and hold still for a mount more than they actually mount. Present-
ing occurs frequently without being followed by a mount. For example, in
wild rhesus, 41 % of presents were not followed by mounts (Lindberg, 1971),
and in captive rhesus sometimes 60% were not followed by mounts
(unpublished data). Presenting is seen in more social situations and has even
more postural variations than mounting. Possibly due to the variety of pos-
tures and situations for presenting, there has been great difficulty in inter-
preting this social signal and describing its development.
As an example of the confusion, Harlow and various colleagues (e.g.,
Harlow et al., 1966b) have described infant "feminine behavior patterns" as
"withdrawal, passivity, and rigidity" in the social situation. This same pat-
tern has also been described as presenting though it certainly appears more
like holding still at best or avoiding interactions altogether. The presenting
pattern is very different from the crouching pattern indicative of extreme
submission or the rigidity that is all too commonly seen in laboratory-raised
females.
Laboratory-reared infants of either sex do not seem to show the present
pattern until relatively late, if at all, for it is seldom described. Harlow
(1965, p. 243) has stated that "There is no semblance of normal sex postur-
ing in either the male or female infant monkey." "Normal" presumably
refers to footclasp mounting and presenting; as shown here, males and fe-
males of many species do show such normal behavior in infancy. Harlow's
20 J. Hanby
reference is made to the "normal" development of monkeys reared with

surrogates and allowed interaction with peers only, not in groups with
mixed ages and sexes. In another study, monkeys reared with just their
mothers and then introduced to a group at 13 months all showed clear
present patterns to others (three of the four within just 2 months of group
experience) (Spencer-Booth, 1969).
A distinction must be made between holding still with rump oriented to
the other, the more distinctive approach with rump or genital area
"presented" to the other (with all its variants), and crouching when ap-
proached. Concomitant gestures (such as lip smacking, teeth chattering,
grimacing, or screaming) should also be noted. The pattern called present-
ing develops early and appears frequently in animals raised in mixed groups,
but appears late or never in restricted animals. Holding still and crouching
are basic to any primate regardless of rearing conditions, but may be
frequent only in confined conditions.
Besides the need for a clearer description of posturing and other ges-
tures, the context of presenting should be made explicit. For example, in the
case of one infant rhesus male observed by the author, the presenting pat-
tern appeared early (7 weeks) in its complete form and was soon repeated
frequently in certain contexts: the infant male would run back and forth
along the cage fencing, threatening at the adjacent group and squeaking,
screeching, and presenting to the adult male of his group. Such "protected
threats" have been vividly described by Kummer (1967, 1968) for ham-
adryas baboons and they appear in many other group-living species (see
Wickler, 1967). The development of this sociosexual pattern-the present-
is especially intriguing since it is so important in agonistic interactions. Its
early appearance in the infant male described above may reflect the fact
that his mother had a special relationship with the adult male and was the
highest ranking female at the time; the group was also undergoing a period
of general heightened agonistic encounters. The age of the onset of present-
ing certainly varies with the social situation as well as with species and
group composition, as is especially clear in infant chimpanzees under high
density (Goodall, 1968b) and more dispersed conditions (F. Plooij, 1976).
Chevalier-Skolnikoff (1971a, 1974) described first presentings in stump-
tails as being adult in form, all three infants (1 male, 2 females) presenting by
the fourth month (one to the adult male, one to a sllbadult female, and the
other to adult females), context unspecified. Nevertheless, since all infants
presented first to adults, the situation could be interpreted as an instance of
the "polite" use of presenting, as in greeting or agonistic settings.
Chevalier-Skolnikoff did distinguish the onset of the above behavior
from groom presenting, which also appeared in the fourth month. Goodall
(l968a) likewise distinguished between the initial onset of presenting as a
greeting (11 months) and the later onset of presenting as appeasement (14-
23 months). She also refers to presenting as submissive (1968b), presumably
because infant chimps present more often when they come into proximity
with a high-ranking individual (90% of 61 presents) than when they have
been attacked or threatened. Since almost all chimps are higher ranking
than an infant, it is not surprising that infants present to "dominants." It is
more interesting that in all cases in the literature, presenting by infants is
more often directed to older animals (especially adults) in nonaggressive
situations as a gesture of "respect" or deference, perhaps best termed a
polite greeting. Presenting by infants- is seldom seen as a "submissive"
response to an aggressive or sexual invitation.
F. Restricted Rearing
Primates reared in restricted conditions show delayed and often dis-

torted development of sociosexual behaviors. The onset of mounting,
presenting and other behaviors, and their orientation, manner (with or
without footclasps, thrusting, bites, intromission, lip smacking, etc.), and
partner "preference" varies with the rearing conditions. Many studies of
sexual and general social development have been done on animals reared in
environments that not only restrict them physically, but also, more im-
portantly, limit the numbers and kinds of companions. The results of these
studies need re-examination in the light of our increased knowledge of pri-
mates reared under less impoverished conditions; they are also extremely
important to our understanding of human behavior since our various
societies enforce certain rearing experiences with kinship structures and
other restrictions due to social institutions. Many of the behaviors seen (or
not seen) in nonhuman primates that have grown up in certain restricted en-
vironments are reminiscent of various human activities. I now consider dif-
ferent environments, from the most to the least restricted.
Isolate and surrogate reared males have no early social opportunities
and are totally inadequate socially and sexually as adults (Harlow et al..
1966a). Infants reared only with mothers show variable sociosexual develop-
ment. Originally Harlow and Harlow (1962) reported on one male and one
female, each reared only with its mother and in cages with no visual access
until 7 months. When the infants were allowed to be together, the female
seemed to try to establish contact but the male never even left the living
cage in the two months of exposure. After separation from the mothers at
13 months they still showed no interactive play in two more months of ex-
posure. On the basis of this pilot study the researchers concluded that "even
22 J. Hanby
normal mothering is not enough to produce socially adequate offspring"

(Harlow and Harlow, 1962 p. 145).
Another male and female reared alone with their mothers were ob-
served by Joslyn (reported in Harlow and Harlow, 1969) and they showed a
reverse pattern: the male was active and the female withdrew; but still
neither associated. A study by Alexander (1966) showed that two groups of
infants reared with just the mother for the first 4-8 months of life were wary
of body contact, had fewer affectional interchanges with peers, and were
more aggressive than infants reared with mothers but with access to peers.
The longer the infant was alone with just the mother, the more exaggerated
the traits. Additional tests made at 13 months when the infants were
removed from their mothers showed that these characteristics (more
aggression, less affection) were retained. Harlow and Harlow (1969)
suggested:
Such animals placed in a heterogeneous social group including animals of all ages might
well become loners or outcasts by virtue of their lower affiliative tendencies and
hyperaggressive tendencies. In a group restricted to age-mates, however, they could con-
ceivably establish themselves as dominant members because of their readiness to ag-
gress. Nevertheless, in a group similarly reared they do show adequate play and normal
sex behavior suggesting that to a large extent mothers may substitute fairly effectively
for age-mates in at least the first 8 months of the rhesus monkey's life. (p. 21)
The statement that mothers can "substitute" for peers reflects

Harlow's apparent belief that monkeys raised with peers only show
"normal" behavior development. I examine this idea shortly; here, it is im-
portant to determine whether infants raised with just the mother and then
introduced to a heterogeneous group do indeed become outcasts because of
lower affiliative and higher aggressive tendencies.
A study by Hinde and Spencer-Booth (1967) is relevant here. They ob-
served four mother-infant pairs (three females, one male), each pair living
alone in a large cage with all but tactile contact with other animals in the
colony for the first year of the infant's life. They reported that on the whole
the behavior of the mother-only babies differed little from the group-living
ones, though the groupless mothers were less restrictive and the infants
ranged to a distance more freely. Hinde and Spencer-Booth specifically
mention that the only male in the "isolate" group masturbated (by holding
his penis) quite frequently from week 22 onward; they never observed this in
a group-living baby.
When introduced at 13 months into groups containing various ages and
both sexes, all the infants "avoided or showed smaller readiness to respond
socially to other individuals. However, the interactions were qualitatively
normal" (Spencer-Booth, 1969, p. 85). Specifically, the infant male and one
female mounted and were mounted by other infants. Spencer-Booth cau-
tioned that the introduction to a new group, rather than the rearing condi-
tions, may have led to the reluctance to interact. Only one of the four in-
fants survived to breed and bear young, and indeed the female seemed to
have lower affiliative and higher aggressive tendencies though she was by no
means an outcast (personal observation).
Infant rhesus reared together without mothers or mother substitutes
seem to form strong physical and presumably emotional attachments
("together-together infants, "Chamove, 1966, 1973; Harlow and Harlow,
1962, 1965, 1969). Social experience with more than one peer either in a
group or consecutively from 10 weeks of age to 22 weeks resulted in ani-
mals that "were seen to adopt normal sexual posturing and it was felt that
when mature their skills were adequate for successful copulation" (Cha-
move, 1973, p. 13).
Once again, "adequacy" and "normalcy" need further thought, but
another study showed that male rhesus raised with peers only (or with peers
after a 3 month period with just the mother) show striking deficiencies in
the mounting pattern. Goy and Goldfoot (1973) have carefully compared
the mounting pattern shown by peer-raised males with that shown by infants
reared with mother and peers together (instead of in succession). The group-
reared males showed a preponderance of "mature" (rear-oriented) single or
double footclasp mounts (with or without thrusts or intromission) in the
first 100 days after observations begun at 3 months (making the infants
about 6 months old). During the same age period, of the 25 peer only reared
males, only 13 showed mounting at all and the ratio of mature to immature
mounts was low. Harlow, Joslyn, Senko, and Dopp (1966) describe another
group of peer only raised males as showing single or double footclasps on
65% of their mounts by 1Y2 years. By 1Y2 years Japanese macaques reared in
a corralled troop used footclasps (almost always double) on 90% of their
mounts (Hanby and Brown, 1974).
These differences in the development of mounting in various rearing
conditions seem largely due to differences in the availability of partners.
Correct orientation to the rear is crucial in the sequence linking erection to
thrusting and ultimately to intromission. For macaques and baboons, a dou-
ble footclasp must normally be achieved if intromission is to occur (foot-
clasping is definitely not so important for the arboreal or larger apes, e.g.,
gibbons, orangs; chimpanzees and gorillas keep their feet on a limb or the
ground as well). When macaques and baboons are prevented from riding
dorsally by spatial restrictions or the absence of a mother, footclasping is
delayed. Restrictions on possible partners that willingly hold still or present
would further retard successful mounting attempts.
The "availability" of partners is a matter that deserves special
consideration. Briefly, an available partner can mean any animal who offers
some portion of the anatomy on which a mount can be executed. However,
24 J. Hanby
there is obviously a great difference between a partner who cringes or

crouches upon being approached and one who actively solicits. From their
work on rhesus reared in the laboratory, Harlow and Harlow (1965)
speculated:
In all probability, the withdrawal and passivity behavior of the female and the forceful
behavior of the male gradually lead to the development of primary sexual behaviors.
The tendency of the female to orient away from the male and of the male to clasp and
tussle at the female's buttocks predisposes the consorts to assume the proper sexual
positions. (p. 325)
We might expect this to be true of same-age infants in a small space

where females cannot run away from a lunging, playful male, but in a less
rigid situation the female is less rigid: she can actively present, adopt a
holding-still posture, or simply run away leaving the male behind. Thus a fe-
male has the opportunity of having the male thrust on her buttocks instead
of on her back or not at all. Likewise, in the mixed group, the male has a
variety of willing partners of both sexes and of all ages and stages of recep-
tivity. The importance of willing partners cannot be overemphasized.
I have primarily discussed the effects of restricted rearing on rhesus
development because the picture is more complete. There is similar evi-
dence on chimpanzees and gorillas, however, which confirms that it is crucial
for infants to develop a relationship with a mother (sometimes humans can
substitute) and also with others. (See articles by Rogers and Davenport,
1969; Lemmon, 1970; and Nadler, 1975).
G. Hormones
The influence of hormones during the period from birth to puberty is

presumably very small. Goy (1964; Goy and Goldfoot, 1973) has de-
scribed the behavior of castrate males and pseudohermaphrodite females
that were reared with mothers for 3 months and then with peers. The males
were castrated at birth or at weaning and in addition to peer experience
some were provided with a female a few times for pair-performance tests.
Although the animals showed much play and agonistic behavior they did
not groom, and no sexual behavior was seen until 2 months after observa-
tions began, at which time (around 9-10 months) erections and thrusts
increased sharply. Interestingly, the frequencies of attempted mounting,
thrusting, and erection were higher for castrates than intact males, which
was due at least partly to the greater social experience of the castrates.
Neither group showed footclasp mounting at this time, a reflection of the
retardation caused by peer group rearing and confinement. Hormones do
Sociosexual Development in Primates 2S
not seem to be necessary for the expression of sociosexual patterns in

infancy.
H. Conclusion
During the first year of life infant male nonhuman primates integrate
the elements of erection, thrusting, mounting, and, to some degree, in-
tromission. In the olive baboon it appears that orgasm (Owens, 1973) is in-
tegrated into the pattern by one year (see also, Zuckerman, 1932). Also, by
one year, both males and females hold still for mounts and show the present
pattern in some social contexts. It must be emphasized that these state-
ments refer only to the various macaque and baboon species that have been
studied and to chimpanzees. Other species may not show these patterns and
different age changes can definitely be expected. Early sociosexual behavior
is completely tied to the specific rearing environment, and is delayed or dis-
torted when mothers and other group members are unavailable.
IV. JUVENILE PERIOD
Most of the elements of the copulatory pattern have been integrated by

the time a macaque, baboon, or chimpanzee enters the juvenile period.
Other sociosexual patterns have appeared during infancy but seem to be less
clearly tied to situations or to particular partners. They may also be more
variable in form. During the juvenile period the copulatory pattern in males
and variations in presenting and mounting in both sexes become more
clearly tied to specific kinds of partners and social situations. The influence
of individual characteristics, group social organization, and species-typical
patterns also become more apparent. The behavior of juvenile primates is
much less well documented than that of infants, and the following trends
must be taken as mere sketchwork.
A. Sequential Mountings
Male rhesus and Japanese macaques and hamadryas and chacma ba-
boons show sequential mounting before ejaculation as adults and it is during
the juvenile period that mounting takes on this pattern. Kummer (1968)
describes juvenile male hamadryas "copulating" with swollen females,
26 J. Hanby
though it is not clear whether this means mounting in series. Male Japanese
monkeys definitely exhibit series mounting during the juvenile period, their
sequences being especially long since they do not terminate in ejaculation.
Studies on laboratory-reared rhesus and other species seldom differentiate
series from single mountings, and the large increase in mount frequency
usually reported for this period may be the result of attempts by males to
mount sequentially. It is undoubtedly very difficult to distinguish a series
pattern from repeated mountings, though certain criteria could be applied,
e.g., so many minutes between mounts, no change in partners, and a
minimum of two or three mounts in sequence.
Another species-typical pattern for which there seems to be little
developmental information is the number of thrusts per mount. Michael,
Wilson, and Plant (1973) have summarized the apparent diversity of thrusts
per mount in the same and in different species. In chimpanzees and olive
baboons, which are both single mount to ejaculation species, the number of
thrusts per intromission is highest in juvenile males (C. Tutin, in preparation;
Owens, 1973), again because intromission is not followed by ejaculation. It is
likely that there are large individual differences in the number of thrusts per
mount and the number of mounts necessary for ejaculation. While such pat-
terns may occur in the juvenile period, they may only be observed after pu-
berty and stabilization in full adulthood. Much more information is needed.
B. Seasonal Patterns
Rhesus and Japanese macaques have relatively strict breeding seasons:

series mountings with ejaculation and conceptions occur only at certain
times of the year. The development of the seasonal changes associated with
series mountings was a striking feature of juvenile Japanese macaque males
(Hanby and Brown, 1974). The 2Y2 and 3Y2 year old males mounted adult fe-
males in series during the fall-winter breeding season. Fewer of the younger
juveniles participated in breeding season activity, but all were observed to
mount in series at some time.
C. Partner Selection
Not only did series mountings of males on adult females become tied to
a seasonal patterning, but the absolute number of encounters that involved
mounting (numbers of mount events) also increased. At the same time,
these juvenile males developed a clear seasonal patterning in the choice of
their mount partners. Adult females were mounted in breeding season, while
males of any age were mounted at any time of year but mainly outside the
breeding season. Instead of series mounts, as were performed on adult fe-
males, juvenile and all older males usually mounted other males and young
females with only a single mount.
Juvenile females do not generally participate in copulatory activity, or
at least it is seldom commented upon. Nevertheless, Kummer (1968) has
described copulation between an adult male (harem leader) and his juvenile,
recently "kidnapped," females. These females have sexual skin swellings
during their juvenile period though they apparently do not conceive and it
must be assumed that they have not yet reached puberty.
Since most species do not have strict breeding seasons or require a
series of mounts to ejaculation, developmental changes during the juvenile
period may be harder to note. Also, there are social factors that may
prevent the juvenile males from exhibiting copulatory behavior. Thus,
Hinde and Spencer-Booth (1967) comment that the decline in juvenile male
mounting in their rhesus groups may have been due to the inhibitory in-
fluence of the single dominant adult male. The participation of juvenile
Japanese macaque males in breeding season was dependent on the number
of females who were in estrus; for during later breeding seasons, when there
were many females with recent babies-these females are less active
sexually-they mounted adult females much less frequently. Whether this
was due to unsuccessful competition with older and more dominant males
or to the disinterest of the females is unclear; both factors probably affected
the juvenile males' behavior.
D. Posturing
Posturing becomes more stereotyped in juvenile primates, and most

studies note an increase in mature mounting patterns and increased present-
ing whether in the wild or in captivity, in mixed groups or with peers only
(Harlow et al .. 1966b; Goy and Goldfoot, 1973; Anthoney, 1969; Owens,
1973; Goodall, 1968a). Some studies have noted that young monkeys and
apes assume a ventro-ventral posture while thrusting. For example, Bingham
(1928) and Goodall (1968a) describe this for chimpanzees and we saw it in
Japanese macaques (Hanby and Brown, 1974). This particular pattern often
disappears with age or experience, but is reminiscent of the relation between
close maternal contact, ventro-ventral embraces, and genital stimulation.
Young males often mount in play and other situations with a wide va-
riety of postures. Play mounts by rhesus or Japanese macaque males on
males seem to be "sloppy" in that they sometimes lack double footclasps
and thrusts, the partner may not hold still, and many mounts are accom-
28 J. Hanby
panied by bites on the shoulders or neck. While mounts on females tend to

have footclasps and longer or more thrusting, thrusting on females can oc-
cur in any position.
Both juvenile males and females use a wide variety of posturing in their
interactions with other group members in greeting, agonistic situations,
play, etc. (Fig. 7). A variety of facial gestures and vocalizations also accom-
panies sociosexual patterns during the juvenile period; these gestures and
vocalizations are absent or infrequent during infancy. For instance, juvenile
male Japanese monkeys utter a low "girrning" sound when mounting in
what seem to be tense peer group encounters, and lip smacking more
(d) (e)
Fig. 7. Some juvenile sociosexual patterns. (a) A rhesus juvenile female presents with foot
back to a young infant before a bout of chasing play (drawn from life). (b) Presenting for
grooming between young stumptail monkeys (drawn from a photograph in Bertrand, 1969). (c)
A juvenile male anubis baboon presents with foot back to an adult male (from a photograph in
Ransom and Rowell, 1972). (d) Mounting with a neck bite during play between two male ju-
venile Japanese monkeys (from Hanby and Brown, 1974). (e) Mutual genital holding during
excitement between two young stumptail monkeys (from a photograph in Bertrand, 1969).
regularly accompanies mounts by juveniles. Bertrand (1969) mentions teeth

chattering as accompanying sociosexual gestures in stumptail monkeys, and
Chevalier-Skolnikoff (1974a,b) has described other facial gestures for the
same species.
The connection of the "appropriate" sociosexual gesture with a certain
age/sex of partner (in different social situations and during certain times of
year) more characterizes the juvenile period than do changes in the connec-
tions between elements or their frequency.
E. Contexts
The most frequent context for juvenile sociosexual patterns seems to be

play, as it is in the later stages of infancy as well. Mounting, thrusting, em-
bracing, and presenting all frequently occur in play, and it seems especially
important that juveniles perform certain patterns with particular partners.
While males tend to mount and play with other males during play, they tend
to mount females in isolated instances (e.g., greeting), and often mount
either sex during tense or disturbing situations (e.g., while a human stares at
them).
Juvenile females often present to, pick up, carry, and allow an infant
male to thrust on them (e.g., Fig. 7a). This type of activity in the juvenile fe-
male may contribute to both her future sexual and maternal behavior
(Lancaster, 1972). "Play presents" by both males and females during the
juvenile period become clearly distinguished from presenting before groom-
ing or mounting (Celebes, Japanese and rhesus macaques, personal
observation).
Kummer (1960) describes the appearance of "notification" behavior in
young hamadryas baboons, a special case of presenting between males
before the group moves off (Fig. 1b). Kummer (1968) also notes that ju-
venile males mount swollen females but then are quiescent until they form
harem units of their own.
F. Interference and Imitation
During the juvenile period (and, less commonly, in infancy) some indi-
viduals show frequent "interference" behavior, and all primate young seem
very interested in copulating pairs. Young rhesus, Japanese monkeys (per-
sonal observation), irus macaques (deBenedictis, 1973), stumptail macaques
(Chevalier-Skolnikoff, 1971b), anubis baboons (Owens, 1973), chimpanzees
(McGinnis and H. and F. Plooij, personal communication), and, to a lesser
30 J. Hanby
extent, gorillas (Fossey, personal communication) seem attracted to and

intently watch mating pairs. Japanese macaque males very occasionally
harass, follow closely, and peer at mating pairs. Rowell (1967a, b) describes
the behavior of young wild baboon males which "often danced round a
copulating pair making repeated coughs and touching them." Owens (1973)
reports similar behavior and notes that the young baboon males sometimes
mounted a female after watching a mating pair. Chimpanzee infants and ju-
veniles, especially males, have been seen to copulate after watching an adult
pair mate (F. Plooij, 1976). Also, chimpanzees are much more active and
consistent in their interference, climbing on the female (often the mother)
and poking at the male (Goodall, 1968a). Both male and female youngsters
do this (McGinnis, personal communication), although females usually in-
terfere only with copulations involving the mother whereas males interfere
with any pair (Tutin, personal communication).
The importance of interference patterns and the imitation of adults is
not easy to assess. Such behavior does not seem necessary for later so-
ciosexual expression, and may simply reflect the excitability or arousal that
group-living primates experience when able to view sexual behavior. In ju-
veniles, such arousal may be diffuse and may engender behavior that is full
of fearful, agonistic, protective, and exploratory as well as sexual
components. Perhaps close observation or imitation helps the juvenile male
(and female?) to associate such arousal to copulatory behavior. Such ideas
are speculative, but interference and imitation in primate juveniles are
probably more important than has been realized, and both need more
attention.
G. Restricted Rearing
If rhesus monkeys are not allowed to interact with any other monkeys
for the first two years of life, but have some heterosexual peer experience
during the late juvenile period, both males and females remain utterly
unable to display any sociosexual patterns that could be called "normal"
(Harlow et af.. 1966b). If reared with a mother for 7 months and then
allowed to live with one peer of the same sex, these juveniles develop so-
ciosexual patterns that at least allow them to copulate in some manner with
a member of the opposite sex (Erwin, Mitchell, and Brandt, 1973). Thus it
appears that some experience with a mother facilitates later learning with a
peer or some other partner (see Lemmon, 1970, for examples from chim-
panzees). That the mother is not necessary has been stated by Harlow and
Harlow (1962), and Chamove (1973) found that peer only raised monkeys
adopted normal sexual posturing; he also felt that when mature their skills
were "adequate" for successful copulation.
Peer only raised juvenile males, however, show inadequacies when

compared with those raised with mother and peers (Goy and Goldfoot, 1973).
Only some of the peer only raised males developed high rates of mature to
immature mounts, and some even stopped mounting altogether during the
juvenile period. Harlow et al. (1966b) reported on other juvenile males
reared with mothers and with varying amounts of peer experience, and these
showed rather high levels of mounting, about 65% of which included at least
a single footclasp. However, one male mounted his mother as frequently as
he mounted all females combined; another also started mounting his
mother, and even one female mounted her mother. The high levels of
mounting and mother-mounting may be at least partly explained by the fact
that these monkeys were separated from their mothers at 13-18 months of
age, and were then reunited and allowed 2 hours of peer contact per day.
Much of the mounting was probably "reunion" mounting (i.e., mounting as
a form of contact greeting, especially in exciting or mildly stressful situa-
tions). This has also been termed "reassurance" mounting (e.g., Goodall,
1968a).
The limited laboratory situation may delay or simply never allow the
young primate to connect some sociosexual behaviors with certain postures
and social situations. Goy and Goldfoot have also described the lack of fe-
male over male mounting preference in peer-raised male monkeys when
compared with mother and peer raised males. They attributed the lack of
preference for females to the fact that the mothers handled agonistic (or, in
their terms, dominant-subordinate) encounters between the young in the
group, whereas the peer-raised males had to settle their disputes themselves;
this explanation implies that mounting is a form of dominance behavior.
Since mounting often appears in excited or tense situations, males might be
expected to mount males during the brief daily peer-group encounters (this
is essentially an exciting and reunion situation). Also, the females in such a
situation may even avoid all interactions, in which case the males interact
with one another. All of these factors (lack of willing female partners, high
agonistic behavior, and unstable social situation) probably contribute to the
lack of male-female mounting in peer groups. It would be interesting to
compare the daily peer-interaction groups of Harlow and Harlow and Goy
and Goldfoot with the constant peer groups of Chamove.
H. Conclusion
The juvenile period seems to be a time for further integration of the

copulatory pattern involving partner selection and seasonal, sequential pat-
terning, at least in some species. It is also an important period for the
elaboration of sociosexual components that become integrated into patterns
32 J. Hanby
specific to the species and to certain partners and social situations. Both
sexes seem to acquire experience in sociosexual patterning (including
agonistic, maternal, and sexual components) in the general content of play,
a name given to rather diverse combinations of social behaviors. Early rear-
ing restrictions result in retarded sociosexual development.
v. ADOLESCENCE AND ADULTHOOD
The onset of puberty implies a surge of pituitary and gonadal hormones

along with an increase in size and weight. Behavior changes contingent upon
puberty and their relation to hormones are not clearly documented in any
species. Behavior changes themselves, however, are widely known.
Human societies have well-known problems with adolescents, and
social regulation of the young during this period is almost universal.
Adolescence also seems to be a trying time for other primates. Aldrich-
Blake and Chivers (1973) describe how adolescent male and female
siamangs have to leave their families and establish their own. Chimpanzee
males and females seem to be especially upset, unsure, and changeable as
females first begin to show sexual swellings and often leave their home
"community"; males engage in long consorts with females and struggle
wii.l:l their male~male relationships (Thorndahl, Pusey, Bygott; personal
communication). Male macaques and baboons usually leave their natal
groups and have to establish themselves in a new group or even create their
own, as in hamadryas baboons (Kummer, 1968). Sociosexual patterns seem
to be utilized especially often during this extended period of change.
A. Male Group Change, Status, and Hormones
Adolescent male macaques and baboons seem to be particularly mo-

bile, often transferring from their own group to another. Observations on
anubis baboon groups (Packer, 1975) have revealed that if a male stays in his
natal group his opportunities for mating are limited (though he might achieve
high rank). Young males must seek sexual partners in other groups. Lindberg
(1969) noted that rhesus males who transferred during the breeding season
copulated with females of the groups newly joined. Those males who shifted
"permanently" (for the duration of Lindberg's study, or somewhere between
1 and 9 months) "copulated at a significantly higher frequency per observa-
tion hour" than did males who stayed for shorter periods. Lindberg com-
mented that these transferring males also copulated with females in their
original groups. In his study, "subadult" males were not seen to transfer, but
Boelkins and Wilson (1972) found that 3 and 4 year old rhesus males were the
most common transfers. Transfers were most frequent during breeding
season.
Drickamer and Vessey (1973), after 6 years of study of rhesus male
transfers, found that the average age of first leaving was about 4 years, all
males leaving their natal group by 7 years. They concluded that age, mating
season, and adult sex ratio all influenced when males would shift, though
whether shifts increased copulatory opportunity was not stated. Opportu-
nities for incest were at least greatly reduced. This point was also em-
phasized by Norikoshi and Koyama (1975), who studied group shifting in
Japanese macaques for 8 years. As in other macaques, shifts were most
frequent in breeding season, and almost all males over 4Y2 years shifted.
Langur males also move from their predominantly female or male
groups, and the turnover seems to be related somewhat to population
density (Rudran, 1973). The ages of the transferring males are often not
known, but the changes are associated with aggression and the death or
disappearance of young animals and with copulation of females with the
new male (Rudran, 1973; Parthasarathy and Rahaman, 1974).
Obviously, captive animals cannot change groups, and the above in-
formation suggests that sexual patterns may decline in frequency as the cap-
tive male matures. This seems to be the case for males who remain in their
natal groups at Madingley (as was mentioned for juveniles by Hinde and
Spencer-Booth, 1967), although quantitative information is lacking due to
the fact that most subadult males are removed. We were able to observe one
male who was allowed to remain until 5 years of age. He was furtive, paced,
and seldom interacted with others. His testicles were withdrawn, his face
and scrotum were pale, and he was never seen to mount except in rare play
bouts. We removed the only mature male for a short while, and the young
male immediately began to interact with others more frequently, though he
was still timid. He was then removed and placed with regularly cycling fe-
males. The change in his demeanor was striking: he was at first very ag-
gressive, and then resumed his docility; his face and scrotum reddened and
his testicles filled the scrotal sac. He also copulated and inseminated several
females. Koford (1963) has commented upon the fact that many young free-
ranging rhesus males who theoretically should mate do not do so because of
social inhibition.
In the example mentioned above of the captive male, and in other cases
of wild young males, the lack of sexual activity in the home group could be
regarded as being due to inhibition in behavior (due to the presence of one,
34 J. Hanby
or a number of, older, more dominant males), to the disinterest or avoi-

dance of the females, or to the inability of the young male (due, it is
presumed, to low levels of testosterone).
In captivity, male rhesus have testosterone increases at around 3 years
of age which stabilize by 3lj2 years (Resko, 1967). Spermatozoa are present
(van Wagenen and Simpson, 1954) and the testes are adultlike (Conaway
and Sade, 1969). If this is so, 3 year old males could be expected to ejacu-
late and even inseminate a female; this has been shown in captives by
Erwin, Mitchell, and Brandt (1973). Nevertheless, social factors do affect
testosterone levels and mating opportunities. This has been clearly shown by
an interdisciplinary team working with rhesus macaques at Yerkes (Rose et
at., 1971; Rose et at., 1972; Gordon and Bernstein, 1973). Testosterone
levels in males were related to their status in the group, to their aggressive-
ness, and to the availability of females.
Since most adolescent males are subordinate to the older males in their
groups, it is not unlikely that social inhibition actually affects their
hormonal levels. However, as I have observed in Japanese macaques, a
young male's opportunities to mate depend at least partly on the availability
of receptive females. In one breeding season, when many females in the
troop were mating, most of the males of whatever dominance or age-even
the juveniles-copulated. During the next breeding season, when most of the
females had suckling infants, few males mated; subordinates, adolescents,
and juveniles showed the greatest reduction in copulations. Mature females
tend to mate with mature males (not necessarily the highest ranking males),
and since most dominant males are mature it is predictable that these males
will be readily accepted by females. Also, dominant males may actively
prevent a female from mating with other males by directing aggression
toward her or toward her partner.
It is quite clear from this information that there is an interaction
between age, partner availability, position in a dominance hierarchy,
general social situation (breeding season, sex-age ratios), and hormonal
state. The emphasizing of one factor at the expense of the others does not
clarify anything at this stage of our knowledge, yet field and laboratory
studies continue to fixate on male dominance as the primary factor affect-
ing mating. Let me Qriefly evaluate some of the evidence.
B. Dominance and Mating
The evidence for a positive relationship between dominance rank and

breeding is largely confined to studies of mixed groups. Family units and
harems (one-male units) are given much less attention, even though females
in such groups may obviously mate not only with the single or dominant
male of their group, but also with fringe, peripheral, and young males (e.g.,
Kummer, 1968; Boese, 1973).
A positive correlation between rank and sexual activity was found for
free-ranging Cayo Santiago rhesus by Carpenter (1942), Altmann (1962),
Conaway and Koford (1965), and Kaufmann (1965), though in no case was
the coefficient very high. Loy (1969) found no "striking correlation between
dominance rank and frequency of mating." We found a positive correlation
between rank and mount events in Japanese monkeys (Hanby et al .• 1971)
but it was not high. In baboon groups, both DeVore (l965) and Packer (in
preparation) found strong evidence for a positive relationship between copu-
lation and dominance rank. The number of completed copulations for male
langurs, however, did not always correspond to their rank (Jay, 1965). Simi-
larly, in chimpanzees, a male's rank does not predict how sexually active he
will be in terms of completed copulations or the ability to take a female
away on "safari" (McGinnis, 1973; Bygott, 1974), though this may change
with the particular males and the state of the male community (Tutin,
1974).
The relatively low and changing correlations between rank and sexual
behavior reflect the confusion of many different factors in studies of these
species. The trouble lies mainly in (a) the multitude of factors that de-
termine a male's dominance (age, size, aggressiveness, background, group
setting, etc.), and in (b) the many different ways sexual behavior is defined
and measured. The nature of such correlations is also a part of the ambi-
guity, since rank order correlation coefficients ignore the fact that an alpha
male may copulate many times more than any other male. Such a statistic
only asks whether A is higher than B is higher than C, and so on. More im-
portant than the correlation coefficient is the fact that sexual activity is
variously defined and measured. Ideally, we would like to know about the
many characteristics of a male and about insemination, conception, etc.
Copulation frequency-or even ejaculation frequency--is not enough to es-
tablish reproductive success.
Relatively exclusive copulations for certain males at the middle of a fe-
male's cycle (presumably the time of conception) have been reported in a
number of species: rhesus (Carpenter, 1942; Conaway and Koford, 1965;
Altmann, 1962; Kaufmann, 1965; Southwick, Beg, and Siddiqi, 1965),
Japanese monkeys (Tokuda, 1961), baboons (Hall and DeVore, 1965;
Packer, 1975; Kummer, 1968), langurs (Jay, 1965), and chimpanzees
(McGinnis, 1973; Tutin, 1974). Even in these studies, however, the relatively
exclusive consortships at midcycle are not necessarily with the dominant
males, and certainly do not occur in an order correlating predictably with
order of dominance in males. Also, females cycle after becoming pregnant,
36 J. Hanby
so it is not always clear whether exclusive access to a female midcycle has

anything to do with conception.
Correlations between dominance rank and mating do not seem to be a
particularly fruitful approach to the problem of male sexual expression.
Other studies have concentrated and are beginning to concentrate more on
the complex of factors that affects the behavior of particular males. The
social situation of primates is so changeable and so variable from species to
species and group to group that our ideas are constantly in need of
reevaluation as more information accumulates. I would like to emphasize
that agonistic dominance is only one factor in a multiplex of relationships
that may affect a male's copulatory success.
C. Mature Copulatory Patterns in Males
Maturity is hard to define in terms of age for most primates. Mature

sociosexual patterns in males may be present in juveniles and may never ap-
pear in some adults. A mature pattern cannot be fully described for any
species, since patterns vary over time and from group to group. Neverthe-
less, some studies show that the important factors in a fully-developed so-
ciosexual pattern include stability and stereotypy in a male's copulatory be-
havior, and an adaptability to situations and partners in general.
In Japanese macaques, mount postures, series patterning, and seasonal
changes in mount event frequency and partner selection were stable by early
adulthood (Hanby and Brown, 1974). Copulatory sequences of older males
were more predictable in terms of the number of mounts to ejaculation, the
interval between sequences, the proportion of events that led to ejaculation,
and the age and identity of partners.
Increased stability and stereotyping in the copulatory repertoire does
not mean that the males ceased to change or show variation. The partners
of a crippled female, for example, managed to devise ways to copulate with
her. One male stood bipedally, held her onto him ventrally, and thrusted;
another continually lifted up her bottom so he could gain intromission. Ac-
commodation to partners may be characteristic of mature mating patterns.
The effects of age on the actual copulatory pattern may be more ap-
parent in males of other species, though there are few data. For instance, it
is not clear when the genital lock, or tie, of the stumptail macaque (Che-
valier-Skolnik off, 1974) enters the male's repertoire. Although form and
frequency of copulation may change slightly at the onset of puberty,
adolescence seems to be primarily a time for sorting out partners and
practicing courtship procedures. For example, an adolescent captive male
chimpanzee described by Tutin and McGrew (l973b) showed a charac-
teristic "lip flip" in his courtship which disappeared when he became

dominant to the females. An adolescent and very high ranking Japanese
macaque male acquired an increasing array of "courtship" behaviors
characteristic of mature, usually dominant males (e.g., sliding over the fe-
male's back, stalking approaches). Courtship displays of males also differ
from group to group (Stephenson, 1973) and may be learned partly by
example.
Wild adolescent male chimpanzees can copulate but do not seem to
consort with females; they may need a few years to acquire the ability to
"persuade" a female to go off with them alone, thereby increasing their op-
portunities to copulate and to sire offspring (McGinnis, 1973; Tutin, 1974).
Likewise, young male hamadryas baboons must "kidnap" and learn to herd
their acquired females (Kummer, 1968). Such a process also seems to take
place in Guinea baboons (Boese, 1973). The maturing male in all these
groups seems to be concerned more with status, control, and courtship than
with copulation, an activity he would have perfected in early adolescence.
As males grow older, their copulatory practices and partnerships be-
come more specialized. For example, mature males may only copulate with
females in the middle of their "estrous" period. This is especially obvious
when females have full sexual swellings at midcycle and only consort at that
time. If mature males "monopolize" fully estrous females, presumably
adolescents can then copulate only at other stages. Mature males seem to
elicit more indications of "satisfaction" from the female (e.g., copulatory
grunts in baboons, Paterson, 1973), although the precise reason for this is
not clear. Other behavior patterns and familiarity may lead to increasingly
stable copulatory preferences between mature males and certain mature fe-
males. Such preferences may in some cases even lessen a male's chances of
siring offspring. For example, the leader rhesus male described by Conaway
and Koford (1965), and the alpha Japanese macaque male in a confined
troop both mated with older females that were often already pregnant (but
see Stephenson, 1973).
D. Other Adult Sociosexual Patterns
In addition to the increasingly stable and individually characteristic

patterns of courtship, copulation, and partner selection, other sociosexual
patterns "mature." Some typical adult sociosexual mounting is shown in
Fig. 8. Male-male mounting is a common adult pattern in group-living non-
human primates, and most often occurs in situations that do not involve
copulatory or agonistic elements. Male-male mounting is not usually
sexual, and does not relate to dominance of either partner; rather, it often
38 J. Hanby
(a) (b) (c)

Fig. 8. Adult sociosexual mounting. (a) A mount between bonnet monkey males-the mount
recipient is reaching back to touch the testicles of the mounter-in a greeting situation (drawn
from a photograph in Simonds, 1965). (b) Mounting between two Guinea baboon females after
an agonistic interaction (from a photograph by A. Mertl). (c) A Japanese monkey adult male
pulls on hips of a young adult female during a bout of play. She reaches back and pushes under
his chin before dashing away (from a 16 mm film frame by J. Hanby).
Fig. 9. Some contact gestures between adult chimpanzees (all drawings from photographs by
D. Bygott). (a) Female touches rump of male who has just given her a beating. (b) One male
mounts another during excitement over being fed bananas. (c) Genital holding during period of
calling by a distant group of chimpanzees. (d) Embrace with neck bite between two males dur-
ing food excitement. (e) Mutual presenting with bottom touching between females during food
excitement.
occurs as a prelude to other interactions in greeting, reunion, and "tense"

situations (Hanby, 1974). Other male-male patterns, such as testicle touch-
ing and holding, may increase as greeting patterns stabilize (e.g., langurs,
Weber, 1973; hamadryas baboons, Kummer, 1968; chimpanzees, Bygott,
1974). Hip touching, back embracing, and other greeting-reassurance pat-
terns may also develop and stabilize between certain males or between
males and females. Some sociosexual patterns may be typical of only
certain groups (e.g., Japanese macaques, Yamada, 1966; and Stephenson,
1973; chimpanzees, Goodall, 1973).
Adult chimpanzees show an especially wide variety of contact gestures,
some of which are shown in Fig. 9. Most of the sociosexual gestures
illustrated occur in situations of excitement or tension. This seems to be
true for adult sociosexual patterns in most species, and there is a conspic-
uous component of tactile and bodily contact, especially on sensitive areas
such as the mouth, rump, and genitals. The distribution of interactive
touching among chimpanzees, according to the sexes of the partners, is
shown in Fig. 10. The large percentage of touches on the genitals, except for
Male on Male (N = 194)
10
"25
5_
Male on Female (N = 85) Female on Female (N = 56)
Fig. 10. Distribution of brief contact with the hand between adult chimpanzees. Contact oc-
curred in interactions that could not be classified as grooming, play or attack. Arrows and
numbers indicate the percentage of occasions on which the face, head, back, genitals, feet, and
hands were touched (from Bygott, 1974).
40 J. Hanby
females touching males, is noteworthy. An analysis of the distribution of

touching in other primates (especially in unclothed humans) would be
interesting.
E. Male-Male Sexual Behavior
Homosexual behavior may occur in nonhuman males under special

conditions in captive groups; it still has not been reported in the wild. Che-
valier-Skolnikoff (1974b) has described and illustrated the occurrence of
much male-male oral-genital behavior (but no male-male ejaculation) in
her small group of captive stumptail macaques (Fig. 11). Bertrand (personal
communication) has not seen such behavior in her several captive groups.
In corralled Japanese macaques, males would on rare occasions (about 1%
of the 1269 male-male mount events observed) mount others in the series
pattern so typical of heterosexual copulation. Only one male was ever
observed to ejaculate with a male partner(probably with anal intromission).
All the males who mounted one another sequentially were heterosexually
active and were scattered throughout the dominance hierarchy (Hanby,
1974).
Gordon and Bernstein (1973) have also reported homosexual behavior
in an all-male group of rhesus monkeys. This was especially interesting be-
c"use these males were caged within sight, sound, and smell of a mixed
group of males and females, and the all-male group showed the same
seasonal changes in testosterone and copulation as did the mixed group.
Another group of males caged nearby, without sight of but presumably
Fig. 11. Male-male genital holding. An unusual pattern illustrated by

Chevalier-Skolnikoff (1974b) for stumptail monkeys (by permission of the
author and Plenum Publishing Corp.).
within olfactory and auditory range of both the other groups, showed only
the typical male-male sociosexual patterns, involving a single mount and no
seasonal hormonal fluctuation. It is presumed that if males are gonadally
active and do not have access to receptive females, they may turn to other
males for sexual interaction (e.g., humans in prison).
Homosexual behavior may also be due in part to conditions of rearing,
as is discussed below. One might expect male homosexual behavior to be less
frequent in those species in which there is only one adult male in the group,
and less frequent in family and solitary livers. A study of the behavior of
males in wild all-male groups could be informative (e.g., as in langurs, Ru-
dran, 1973; or patas monkeys, Gartlan, 1975).
F. Incest and Cross-Species Copulation
Much has been made' of some very infrequent examples of mother-son

mating in some monkeys (e.g., rhesus, Sade, 1968). This is definitely not a
common phenomenon in free-ranging or even in mixed captive groups.
Brother-sister mating is more common (chimpanzees, McGinnis, Bygott,
personal communication; Japanese and rhesus macaques, personal observa-
tion) but still is rare. The more common appearance of "incest" when males
are adolescents is more interesting, despite its overall low frequency.
Sade ~ 1968) attributed one occurrence of mother-son mating to the
young male's rise in dominance over the mother, hypothesizing that nor-
mally the mother's dominance inhibits mother-son mating. Another
hypothesis, supported by observations on Japanese and rhesus macaques
and by an observation on humans in kibbutzim (Shepher, 1971), is that early
intimate contact normally inhibits later sexual attraction to a partner. Since
sexual behavior overlaps considerably with gestures that may be said to
reassure or comfort, an individual can get "confused" in certain situations.
For example, a young Japanese monkey male in his first breeding year
(when he could ejaculate) would occasionally run from his copulatory
partner back to his mother, mount and thrust on her once or several times,
and then run back to his other partner. It is easy to see that, in a more
agonistic setting, elements of aggression, submission, fear, and tension, as
well as sex, can combine to produce a mixture of behaviors or a sociosexual
gesture that expresses the conflict.
Males in many nonhuman primate species mount their mothers on oc-
casion throughout development, but there is still no clear evidence of
ejaculation with a mother in free-living species. Ejaculation does occur in
brother-sister pairs.
Another unusual behavior is cross-species copulation. There is much
42 J. Hanby
old lore about apes and humans copulating, and "bestiality" is still
considered a crime, or at least a perversion, in most societies. Copulation
between closely related species is considered as hybridization and is not
uncommon (e.g., Nagel, 1973). Most reports of cross-species copulation (or
attempts at copulation) are for captive animals. In the wild, young male
chimpanzees responded to presentings by adolescent female baboons (show-
ing their first swellings) with attempts at or even achievement of double
foot-clasp mounts and intromission (H. and F. Plooij, personal communica-
tion and photos) (Fig. 12); this also illustrates how primates adapt posturing
to a given situation.
G. Old Age
Information on the relationship between old age and sociosexual be-

havior in males is lacking since such work must be very long term.
Nevertheless, it does seem that very old males and females remain so-
Fig. 12. Cross-species copulation. A young male chimpanzee mounts a young female anubis
baboon undergoing her first sexual cycle. This drawing (from a photograph by H. and F.
Plooij) shows how the chimpanzee male, whose usual copulatory posture involves squatting
(Fig. 2c), is attempting to mount with footclasping, which is the usual baboon pattern (Fig.
2b). The male actually gained intromission on several different occasions.
ciosexually active. I have observed a rhesus male older than 20 years mount
his very old female cagemate during excitement (e.g., strange humans star-
ing at them), or when the female "incited" the male (presenting frequently)
while she was threatening people or other monkeys. This pair even managed
to produce an offspring, although it was deformed. In the wild, the ability of
old Japanese macaque and hamadryas baboon males to remain integrated
with or in control of their groups may decline. Opportunities to mate may
thus decline with social changes, as well as with loss of health and physical
prowess, in groups in which the males' breeding opportunities are related to
competition or status. In family-living species, in those in which social orga-
nization is not strict, and in captive groups, old age may not decrease copu-
lation activity as much, especially if preferred or familiar partners remain
or new ones are acceptable or tempting.
H. Restricted Rearing: Males
The effect of early rearing on rhesus adolescent and adult male sexual
behavior has also been shown in a number of studies by Harlow and
colleagues (Harlow and Harlow, 1969, provide a review and references) and
by Goy and Goldfoot (1973); the same has been shown in chimpanzees by
Davenport and Rogers (1970) and Lemmon (1970). Males very clearly need
both maternal and peer experience during infancy, and possibly during the
juvenile period, if they are to display sociosexual patterns like those of
group-living wild animals. Each type of restricted rearing produces many
interesting patterns, such as homosexual behavior, self-stimulation, ag-
gression, and a wide variety of heterosexual practices, as well as a complete
lack of sociosexual behavior in some individuals.
Males reared in isolation for the first year or more seem to be totally
inadequate at mounting; they are probably also inadequate at presenting
and other sociosexual gestures necessary for group living, though these are
seldom mentioned. Males reared with surrogate mothers did not mount nor-
mally or achieve intromission (Mason, 1960; Harlow and Harlow, 1969;
Harlow et al., 1966b).
Hand rearing by humans and subsequent placement with conspecifics
seems to have various effects in males. C. Wilson (personal communication)
described one male rhesus mounting his pigtail cagemate frequently and
subsequently managing to inseminate females. T. Rowell (personal com-
munication) has also found that her hand-reared baboon males would copu-
late as adults, as did a hand-reared patas monkey. However, she also
found that neither a hand-reared male talapoin nor one reared with parents
showed any signs of mating. Lemmon (1970) describes successful copula-
44 J. Hanby
tion in some hand-reared chimpanzees. Hand rearing by humans may be

better for later copulatory success than isolation or surrogate "mothers," as
long as the conspecific partners are willing and capable.
Experience with other monkeys in adulthood does not seem to mitigate
the effects of early social isolation. A report by Meier (1965) implied that
several laboratory-reared male rhesus monkeys had managed to inseminate
some females in the breeding colony at Puerto Rico, but Missakian (1969)
tested two of Meier's males and found both incapable of normal sexual be-
havior. Even if the laboratory-reared males were able to inseminate females,
that alone is not a basis for assuming that their sexual behavior was adequate
for anything more than insemination. Most isolate or surrogate-reared ani-
mals were tested either with other like-reared monkeys or with experienced,
usually feral, animals. Mason (1960) found that, when tested with other
restricted animals, young rhesus males showed low frequencies of sexual be-
havior, but with experienced females the incidence of mounting per session
rose from 0.2 to 2.4 and the incidence of thrusting from 0.13 to 1.8, although
most mounts were still inappropriately oriented. Mason attributed the
increased performance to either the experience gained in the social tests or
the presence of an experienced partner. However, Mason's males were also
older by the second round of testing, and the cycle state of the females was
not considered: they may have been more "attractive" (see Herbert, 1970)
and more experienced, and presumably were more receptive.
Experience as adults per se does not seem to help macaques, as shown
by Missakian (1969, 1972), who compared (as adults) laboratory-reared
males with laboratory-reared females: "At no time was a cage-reared male,
either with or without adult social experience, observed to execute a normal
mount" (Missakian, 1969, p. 405).
Species differences in the degree of "recovery" from restricted rearing
have been shown by Davenport and Rogers (1970) in chimpanzees. They
pointed out that restricted chimpanzees, unlike monkeys, could gain some
degree of proficiency as adults after practice. The greater flexibility of
chimpanzee sexual posturing could account for this, but adequacy in copu-
lation is not equivalent to adequacy in this species' other complex so-
ciosexual gestures. Adult gorillas reared in captivity-mostly captured after
the mothers were shot-show just as many reproductive deficiencies as do
other apes and monkeys, and these have also been attributed to early rearing
(Nadler, 1975).
As adults, some rhesus males reared with mothers or peers in varying
combinations and at different times can copulate, but other males seem
inadequate. Chamove (1973) reported on different varieties of rearing con-
ditions, and concluded that all mUltiple peer raised males were "adequate"
sexually. Likewise, single peer mother only raised and mother peer-raised
males showed "no deficiencies" in sexual behavior. Goy and Goldfoot,

however, found that of 23 mother, then peer, raised males, only 9 displayed
intromissions and only 3 ejaculated when they were put with receptive fe-
males at 4-6 years of age.
Adolescent male and female rhesus that had been reared with first only
the mother then only one like-sexed partner were able to copulate and even
reproduce. The behavior of these young ejaculators in the laboratory,
described by Erwin, Mitchell, and Brandt (1973), is especially interesting.
When paired with members of the opposite sex, the monkeys exhibited so-
ciosexual patterns that seem distorted compared with patterns in mixed
group-reared monkeys. Aggression, females "collapsing," crouching,
disoriented thrusting, and ejaculation upon mounting with no intromission
were seen, in addition to double footclasp series mountings and presenting.
One female showed fellatio with her partner during bouts of grooming and
also thrusted on his ventrum. Another female was very solicitous but her
male often ignored her; still another mounted her partner regularly. After
this heterosexual experience, it was determined that one pair of males
preferred each other as sexual partners; they had lived together as late in-
fants and juveniles. Such idiosyncratic patterns of behavior are most in-
triguing, and are reminiscent of human practices.
Unfortunately, the emphasis in almost all studies of restricted rearing
has been on sexual performance. Other sociosexual patterns have been dis-
regarded arui some, like male-male mounting, have been assumed to be ab-
normal or solely dominance related. Patterns of integration into a group
and of mounting and presenting outside the heterosexual context need
further examination. A recent study by Sackett (unpublished manuscript)
showed that out of six males with various backgrounds that were released
on an island, one of the two survivors was feral and the other was a
mother peer-reared male. More work along these lines would be greatly
illuminating.
I. Female Behavior: General
In females, there is a clear positive relationship between the onset of

regular menstrual cycles and the appearance of mounting, reception of
mounts by males, presenting to males, and certain vocalizations and ges-
tures. Other sociosexual patterns, such as thrust-cuddling with youngsters
and more maternal behaviors, may even decline as the females experience
regular estrus cycles, get pregnant, and bear young. Adolescent female
chimpanzees (Tutin and McGrew, 1973a) and Japanese monkeys (Hanby
46 J. Hanby
and Brown, 1974) were very active sexually, and their participation in mat-
ing activity seemed to increase with experience.
Wild adolescent chimpanzee females may leave their original com-
munity (Nishida and Kawanaka, 1972; Pusey, 1974) and they seem to cycle
regularly for several years without getting pregnant. (This may be another
case of "adolescent sterility.") Since they copulate with many males, often in
groups, such extended sexual intercourse may help to establish the female in
her new group and provide her familiarity with all the males (a two-way
benefit). Although there are not yet data enough to support the idea, females
who have given birth to one infant may be more prone to consort only with
certain males. In any case, experience certainly contributes to the female's
partner choices and acceptances, as it does to the male's.
There seem to be wide differences between species, groups, and indi-
viduals in the way in which age is related to female copulatory behavior. Fe-
male anubis baboons, for example, may predictably jump away from the
males after copulation in one group but not in another (Paterson, 1973).
Kaufmann (1965) found that mature (> 7 years) rhesus females were the
most sexually active, whereas in Japanese macaques 5% year old females
were most active (Hanby et al.. 1971). This is probably less a difference in
species than in the reproductive state of the females (e.g., recency of off-
spring), a reminder that female sexual behavior is probably much more inti-
mately related to hormonal condition than to age or even social position.
I cannot presume in this paper to fully evaluate the intertwining of
cycle state, reproductive condition in general, age, and social status. There
are so many factors, such as recency of offspring, parity, dominance, etc.,
that any review of them would require many pages. The reader is referred to
an excellent assessment of this problem by Rowell (1972); I shall try to briefly
cover some other sociosexual patterns of females.
J. Female Mounting
Although adult female mounting is a widespread mammalian

phenomenon (Beach, 1968), the data on primates are sparse. Hamilton
(1914) noted adult female monkeys occasionally mounting one another and
displaying lip smacking and pelvic movements which, if the context is
ignored (e.g. a mother returning to her group and mounting her daughter),
might be interpreted as "sexual" behavior. On the other hand, Carpenter
(1942) described females mounting other females in a free-ranging and
newly formed group of rhesus monkeys. These females formed "consort"
pairs, and most were in some stage of "estrus." These cases of female
mounting were probably hormone dependent. Other examples of female
mounting in chimpanzees (Bingham, 1928; Yerkes, 1939) are ambiguous.

Altman (1962) lists 9 (out of 121) cases of mounting between rhesus fe-
males, and 4 cases of females mounting males, with no further comment
about the type of mounting, the situation, the season of year, etc. Adult
Japanese macaque females were seen to mount other females on 34 occa-
sions, and mounted males on 226 occasions. Most of these events were dur-
ing breeding season and had sexual elements (Hanby and Brown, 1974).
Adult female-female mounting in many different situations occurred in an
all-female captive baboon group observed by A. Mertl at the Chicago
Zoological Park (1967; personal communication) (Fig. 7b). These female-
female mountings were probably both sexual and nonsexual. I have ob-
served reciprocal female-female mountings of a series nature in rhesus
monkeys. All involved one dominant female with her two adult female cage-
mates; the male of the group mounted all of these females as well. Bertrand
(1969) and Chevalier-Skolnikoff (1974b) both described female mounting in
stumptail macaques.
Japanese macaque females in a corral occasionally mounted one
another reciprocally and in series during the breeding season, but they much
more frequently mounted or rode on their male consorts (Fig. 13). Half of
the 22 females in one breeding season mounted males with either the double
footc1asp posture, or more commonly sat or lay on the male, and often
(a) (b)
Fig. 13. Female mounting. (a) An adult female mounts a young adult female, rubbing her
genitals back and forth. These are Japanese monkeys, but the posture is seen in other ma-
caques as well (e.g., stumptail monkeys, Chevalier-Skolnikoff,1974b)(from a photograph by J.
Hanby). (b) Female dorsal riding on male in Japanese monkeys during their consortship (from
a photograph by J. Hanby).
48 J. Hanby
thrusted on him (Hanby et al .. 1971). It was interesting to learn from later

observers (Eaton, personal communication) that the females used primarily
the double footclasp posture in later seasons. Stephenson (1973) has shown
that, in Japan, some groups of Japanese macaques show female-male
mounting as part of the consort relationship, and others do not. The female
lying on the male's back and being carried (Fig. lOb) was very reminiscent
of early dorsal riding, and was part of the consort behavior of certain pairs.
Thus female-male and female-female mounting is a product of group con-
ditions, traditions, sex ratios, the hormonal state of the female, and the
idiosyncratic patterns of particular females.
K. Female Presenting
Presenting between females may be related to dominance (Richards,

1972; Wickler, 1967); females often present to new mothers (e.g., DeVore,
1963; Fig. la), and in situations of excitement may touch rumps with a
hand or even rump-to-rump (see Fig. ge). Rowell (1972) calls presenting a
universal gesture of "politeness" and mentions that she has never seen it
used to divert attack, as has been the common explanation of this gesture
for years (e.g., Zuckerman, 1932; Wickler, 1967). Rowell also notes that
presenting does not seem to relate to the state of swelling in those females
that do swell.
As we have seen, presenting has an early and diverse appearance in
both sexes and in various social situations. In general, females present to in-
dicate deference and respect or to request recognition, contact, grooming,
or "permission" to remain near; female presenting also invites infants to
board and males to mount (Figs. la, c, 3a, b). Despite its meaning so many
different things, many workers still label the posture a "sex present." Our
potential for understanding would be greatly improved if researchers would
indicate partners (age, sex, dominance, etc.) and social situation, and would
avoid labelling a present as sexual. The causal nature of the present posture
seems to be too diverse to assume a sexual basis. In fact, Phoenix et al.,
(1967) commented that "Not only does the response appear to be inde-
pendent of estrogen and progesterone but independent of any gonadal
hormone ... "
L. Female Orgasm
The question whether nonhuman primate females experience orgasm is

currently a matter of some debate. Burton (1971) found no evidence for or-
gasm in artificially stimulated rhesus monkey females, but Michael (1971)

showed that uterine contractions increased when male rhesus thrusted on
their ejaculatory mounts. Chevalier-Skolnikoff (1971a) has described a pat-
tern of female behavior in stumptail macaques that clearly resembles the
male's ejaculatory pattern. Rowell and Packer (personal communication by
each) both agree that olive, like chacma, baboon females (Saayman, 1970)
probably experience orgasm. These baboons emit special grunts on the last
thrust by the male and leap away after copulation with a series of grunts
(H. Plooij, unpublished sonograms). McGinnis (1973; personal communica-
tion) has described the behavior of chimpanzee females that emit special vo-
calizations during copulation which may indicate at least extreme sexual ex-
citement if not actual orgasm.
Zumpe and Michael (1968) associated reaching back and certain other
gestures with rhesus female "orgasm," but we found in the Japanese
monkey that females of any age, males, and copulating or playing pairs
would often reach back more as a mount or dismount gesture than as a ges-
ture that could be reliably associated with the end of a series of mounts or
male ejaculation. Nevertheless, orgasm could certainly occur in some
females, even though they do not show a pattern resembling male ejacula-
tions. If orgasm is as variable and subtle-and often absent-in the non-
human female as it is in the human female (e.g., Singer and Singer, 1972a,b),
it would be very difficult to observe at all, let alone reliably.
M. Restricted Rearing: Females
Restrictions of social opportunity during early rearing have been said

to be less drastic for females than for males, since some females at least
present properly when put with males (e.g., Harlow and Harlow, 1965).
Nevertheless, isolate and surrogate-reared females were still deficient in
"normal" sexual behavior, showed self-biting and clutching, were aggressive
to males, did little or no grooming, and their maternal behavior was
pathetically distorted (Harlow et 01., 1966b; Harlow and Harlow, 1965).
Heterosexual experience as juveniles and adolescents did not improve adult
females' heterosexual or maternal behavior (Harlow et 01., 1966b).
However, isolate females seem to be better mothers with their second babies
than with their first, and surrogate-reared mothers might not be as ag-
gressive to their babies as the total isolates (Harlow and Harlow, 1969;
Harlow et 01., 1966a). Rearing with a mother alone (Alexander, 1966), or
with a mother and then peers, or with one other peer female (Chamove,
1973; Erwin et 01., 1973), seems to improve sexual behavior to "adequacy,"
though descriptions of other sociosexual gestures and other behavior
50 J. Hanby
systems are meager. Even though a female can be more readily inseminated
than a male can inseminate, this does not necessarily mean that a female
could behave appropriately enough to ensure that she and her offspring
could remain members of a group, and thereby presumably profit from its
protection and its social advantages for the children.
Observations on the mothering behavior of an 8-year-old hand-reared
chimpanzee female showed that she exhibited many sociosexual and ma-
ternal behaviors, especially after menarche. She mothered various objects
and a kitten, clasped, thrusted, and occasionally showed behavior similar to
that often shown by chimpanzee females after copulation (Savage et al ..
1974; Lemmon and Temerlin, 1974). Her care giving, greeting, and
courtship sequences appeared to be less well integrated and, of course, mis-
directed when compared with sequences in wild chimpanzees.
The general presence of maternal "adequacy," and the fact that some
females can present properly when deprived of early social experience, both
suggest that these behavior patterns are very basic elements in a female's
repertoire. The importance of the present is also indicated by its universality
across species and the fact that it is common in males as well. Undoubtedly,
behaviors that promote the well being of a female and her offspring would
be ensured by natural selection, which explains the presence of presenting
and other contact-requesting, cohesion-promoting, and agression-reducing
behaviors in group-living primates. While experience in handling young as a
juvenile may lessen the awkwardness of a female with her first baby
(Lancaster, 1972), such early experience may not be necessary to her adult
behavior because maternal behavior is probably rather strongly selected for.
Nevertheless, adult behavior in both sexes in all group-living primates must
depend on a complex intertwining of experiential, genetic, and hormonal
factors.
VI. GENERAL DISCUSSION
A. Comparisons between Nonhuman Primates and Humans
This section involves extrapolations from the preceding review and is

necessarily speculative and limited. Ford and Beach (1951) and Kinsey et al.
(1953) compared human with nonhuman primate sexual behavior in their
books; in this paper I can only consider some of the newer items of in-
formation and their implications.
The sociosexual behavior of group-living nonhuman primates can be
divided into distinct "contacting" and "copulatory" systems, both arising
from infantile clinging, riding, thrusting, and genital contact. The copula-
tory system develops early into a recognizable, stereotyped sequence. The

contacting system, however, is more variable and situationally dependent.
Its components (mounting, presenting, genital touching, embracing, etc.),
though superficially resembling copulation, are in many ways similar to in-
fantile and maternal behaviors, and contextual evidence suggests that they
reflect a desire for reassuring contact rather than sexual gratification.
Both systems develop in parallel, individuals learning to perform the
appropriate behavior with an appropriate partner in any given situation.
The process depends heavily on the availability of familiar, varied, and
experienced partners. In most species, all the elements of the male copula-
tory pattern appear and become fully integrated during infancy. Infant fe-
males, however, show only some components, such as thrusting, presenting,
embracing. During adolescence the full female copulatory repertoire ap-
pears, and both sexes develop their precopulatory and courtship patterns.
They also perfect other sociosexual patterns concerned mainly with status
and with relationships to other group members. Adolescence is a time when
many male primates (but females in chimpanzees) transfer from one group
to another. Whether an adolescent transfers or not, it still has to build its
own relationships with adults outside its maternal family, and is thus largely
concerned with attempting to gain adult acceptance and recognition. This
process may be facilitated by sexual activity with other members of the
group, and by the exercise of sociosexual patterns which were traditionally
assumed to be submissive but perhaps more often reflect a request for non-
aggressive contact. Thus presenting and mounting are employed in es-
tablishing affinitive as well as sexual relationships.
Observers often interpret all sociosexual gestures of nonhuman pri-
mates as being sexually motivated, though data on the development and
hormonal status of monkeys and apes imply that they are better able to dis-
tinguish copulatory from contacting patterns than are humans. To make
matters worse, many observers have labelled the more obviously nonsexual
interactions as "dominance" behavior (e.g., Wickler, 1967; Chevalier-
Skolnik off, 1971 b). There is much evidence to suggest that sociosexual
mounting and presenting are related more to reassurance than to domi-
nance (using "dominance" to refer to an agonistic relationship between two
individuals). Rather than being "submissive," presenting seems in most
cases to be a "polite" or relatively stereotyped deference gesture, often
preceding a positive gesture or preventing an agonistic one. Likewise, a
mount does not usually mean that the mounter is expressing dominance.
The mount is a basic part of a primate male's repertoire that gains him
nonaggressive contact with other males and females. Although mounting
and presenting occasionally occur in agonistic settings, this is by no means
their most frequent context.
Any interpretation of these and other sociosexual behaviors can only be
52 J. Hanby
made by knowing the age, sex, and other attributes of the partners and the
immediate and general social context (e.g., "an adult male mounting a ju-
venile female during a play bout outside the breeding season"). It is particu-
larly important to interpret sociosexual gestures with reference to the rela-
tionship between the partners. Observers are being simplistic when they
judge social structure and behavior in terms of dominance hierarchies based
on aggression, since kinship structures and the patterns associated with
general filial behavior (e.g., grooming, protection, play, huddling,
proximity) are at least as important as aggression to the structure and func-
tion of a nonhuman primate society.
The ease with which humans misinterpret the sociosexual behavior of
nonhuman primates underlines the extent to which copulatory and contact-
ing systems are confused in human behavior. Although some human cul-
tures have been able to keep these two systems separate (e.g., Gorer, 1938),
they overlap more in humans-particularly Western cultures-than in any
other primate species. This confusion is a natural consequence of human at-
tributes such as bipedalism, the face-to-face orientation of all social interac-
tions, the long rearing period with restrictions on the availability of so-
ciosexual partners, and the lack of clear estrous periodicity or hormonal
seasonality. These factors combine to delay the development of the copula-
tory pattern and restrict the expression of contacting patterns. The
adolescent human, therefore, may experience a strong yet unsatisfied need
for contact-reassurance in addition to his or her awakening sexual desire,
and will eventually seek a partner of the opposite sex to fulfill most so-
ciosexual needs (which, for a nonhuman primate, may be satisfied by a wide
variety of other group members). Most humans learn to attach contacting
and copulatory patterns to one or possibly several partners, but have diffi-
culty in expressing to other adults nonsexual contact gestures which provide
reassurance and comfort. This tendency probably helps to perpetuate mo-
nogamous pairing and the nuclear family; but difficulties in sociosexual
expression between adult humans may also lead to difficulties in many other
social spheres, to a lack of cohesion, to disturbances in relationships, and to
a reliance on nonsocial comforts and verbal reassurance.
This admittedly superficial analysis of the problem only emphasizes the
importance of carefully evaluating our rearing methods and social stan-
dards. If one is restricted from sexual experience and even from comforting
contact with other humans-toys and material possessions are poor sub-
stitutes-a child may grow up to behave in a manner reminiscent of
restricted monkeys and apes. To extrapolate from nonhuman primates, the
restrictions usually imposed upon children imply that "normal" adult
sexual behavior will probably be highly idiosyncratic, and will be especially
variable if both partners are inexperienced. Often partners can adjust and
find mutual satisfaction; but often they cannot, and consequently they
sometimes seek gratification elsewhere or risk having their relationship
suffer.
The critical point in development for determining adult sociosexual
expression appears to be infancy, both in humans and in nonhuman pri-
mates. Recent research supports this view, held for years by psychoanalysts,
though it does not support some details of psychoanalytic theory and
practice (see Bowlby, 1969); e.g., infant primates do not seem to go through
oral, anal, and genital stages, nor to have oedipal conflicts, castration con-
flicts, or a period of juvenile latency in sexual activity. Of course, the young
primate suckles and is thereby oral, excretes and is thereby anal, and fon-
dles its own genitals or lets others do so; but the total mother-infant rela-
tionship has more overall psychological import. Species differ in their type
of mother-infant relationship (Hinde, 1971; Kaufman and Rosenblum,
1969), and this may influence developing and mature sociosexual patterns;
the question needs much more exploration. The mother seems to be espe-
cially necessary for acquiring emotional control and affinitive expression;
peers seem to be necessary for learning how to behave appropriately in
many social situations and how to form and maintain relationships. The
presence of other older and younger companions allows the young primate to
learn and practice copulatory, maternal, and other sociosexual patterns.
Maternal care, other adult companions, and peers are all important,
and no one of these factors can replace another. This point cannot be over-
emphasized. Harlow, in particular, has claimed that infant monkeys
develop more "adequate" or "normal" sexual behavior when raised in a
peer group than they would if raised with just the mother. I have tried to
make it abundantly clear that this conclusion was based on inadequate data,
and further research has revealed the many inadequacies of peer-raised
monkeys in their development of both sexual and affinitive patterns.
Mothers and peers are both crucial for sociosexual development, and no
human mother can justify putting her infant into a creche or other peer
group on the basis that "studies with monkeys prove that peers are more
important for development than mothers."
Clearly, humans limit the various possible kinds of early sociosexual
expression and learning. In particular, the viewing of copulation is usually
forbidden, older experienced sex partners are not available, and sex play
with peers or younger children is restricted or punished. What is the
developing human to do with his or her inherently sensitive body, growing
sexuality, and increasingly early puberty, in a society that demands rapid
and clear skills in meeting and greeting? Difficulties are obviously inevita-
ble. The romantic concept of "true love" seems to require human primates
to completely overlap copulatory and contacting patterns, both emotionally
54 J. Hanby
and behaviorally. Such a degree of overlap exists only in the infant stage for
other primates. It is understandable that the human goal of "true love" re-
quires great mutual dependency and the ability of partners to satisfy each
other's desires on a great many levels; but this expectation of "true love," the
differences in experience between marital partners, and the isolation of many
couples in highly mobile cultures, all combine to place a great strain on
marital relationships, as is reflected in divorce, broken homes, and mental
illness.
Early rearing restrictions-in the form of a close relationship with the
mother, combined with a lack of opportunity for experience with other
adults and children, and followed by boarding schools where the usual ju-
venile social adjustments involve aggression, play, and strange or demand-
ing situations-must create other problems. A child seeking the reassurance
and comfort previously provided by the mother may turn to peers or to
adults of either sex. Intense relationships often result, and the confusion of
sex with comfort seeking is part of this intensity. Temporary or more
permanent homosexual relationships may develop. This is not necessarily a
reason for censure, but it is an illustration of the way in which rearing
practices may affect sociosexual development.
Early close physical contact with a specific individual-usually the
mother-during the infant period seems to result in the later inability of the
partners to form a sexual relationship with one another, although in
monkeys and apes offspring may occasionally attempt copulation with the
mother before adolescence. If sexual inhibition results from early intimate
contact, then instinctive "incest taboos" in humans are understandable.
Normally, only mothers and siblings come into close and frequent bodily
contact with the newborn. An increasing participation by fathers in early in-
fant handling might reduce father-daughter incest and might provide more
contact and comfort for both fathers and children, whereas multiple mar-
riages, absent fathers, and rare or restricted contact of fathers with their in-
fants may be associated with the incidence of incest and other problems.
All of the above points are open to argument, and hopefully further
discussion and re-evaluation of primate studies will broaden our under-
standing and perspective. The last point I wish to make is that a human is
not a chimpanzee or monkey. Humans, even when they have experienced
severely restricted early rearing, can and do adjust, and even appear to
behave "adequately" in social settings. Our ability to learn and our strong
motivations to fit into the social system and to be "happy" help in this ad-
justment; but many people are unhappy, do not fit into society, and have
learned sociosexual patterns unacceptable to society. An increase in human
contact, especially with infants, and a broader acceptance of the primate
need for reassuring contact at all ages, may alleviate some of the strains in-
duced by modern society.
B. Research Suggestions
Throughout the foregoing review, I have tried to indicate weaknesses in

the literature that suggest new directions for research. The following sum-
mary of the main problem areas in the study of sociosexual development
should help to clarify the issues.
1. Species Differences
Most of the work on species differences has been done on a few ter-
restrial multi male species: macaques, baboons, and chimpanzees. The indi-
vidual and group variations in these species indicate that species-specific so-
ciosexual patterns can only be determined after a number of groups have
been studied. Age changes need much more documentation for each species
because response to early rearing probably differs between species. Work on
humans is urgently needed.
2. Rearing Environment: Physical
Every stage of development seems to be important. Although there has

been a remarkable clarity in descriptions of the physical environment (cage
dimensions, etc.), adequate descriptions of actual amounts and types of ac-
cess to other animals (conspecifics, humans, etc.) is needed. Such descrip-
tions can be complicated but must be recorded, especially when animals are
removed and returned to various cages, pens, compounds, or groups. The
"richness" of the habitat is a factor in both captive and free-ranging condi-
tions, but is seldom described. Each study is unique, and no one should ever
presume to have described the "typical" or "normal" for any species or
group.
3. Rearing Environment: Social
The numbers, sexes, ages, and background experience of all animals

under study must be considered. The length of time with partners and the
type of access need to be specified and evaluated, e.g., daily interaction,
continuous housing, multiple separations, etc. More importantly, the
quality of the interactions needs to be described in much more detail, espe-
56 J. Hanby
cially as regards (a) mother-infant relationships, (b) adult or juvenile male-

infant relations, and (c) female-other relationships at all stages. The.social
setting generally needs more description and evaluation. For example, the
sex ratio and the age ratio both affect sociosexual behaviors, but additional
measures of relationships, such as proximity, agonistic encounters, and
grooming, would be most valuable in understanding how a particular indi-
vidual fits into the group as a unit.
4. Hormones and Genetics
Of course the "ideal study" would have information on the genetic

background of the animals involved, and would give prenatal, postnatal, pu-
bertal-adolescent, and adult hormone measures, as well as behavioral data.
There is increased hope that gonadal hormones can be measured by sam-
pling fetal and prenatal fluids (Resko et al., 1973), and techniques such as
radioimmunoassay can help to detect hormones such as testosterone in the
developing and adult female as well as in the male.
The increasing use of swelling species (e.g., pigtail macaques) may
facilitate the linking of hormones to behavior and vice versa. More studies
on the interactions between social, seasonal, and hormonal factors in all
species must be done (e.g., Gordon and Bernstein, 1973; Rose, et al., 1971).
5. Observations: Sociosexual Behavior as a Dependent Variable
The measures of sociosexual behavior need to be defined with care.

Many laboratory studies classify all female presents as "sexual," even
though it is perfectly clear by situation and hormonal state that they most
likely are not sexual. The avoidance of labelling sociosexual behaviors with
motivational terms seems desirable (Hanby, 1975).
A wide variety of presenting and mounting patterns can be precisely
described and defined in terms of:
l. Postures (e.g., "full" presents with tail averted, head over shoulder
vs. "partial presents" with simple rear oriented to recipient and
steady stance; double footclasp mounts vs. mounts with weight not
on partner).
2. Associated elements (e.g., mounts with or without thrusts, in-
tromission, or ejaCUlation; reaching back with foot towards partner
being presented to).
3. Partners (e.g., female-male, male-female; old-young; dominant-
subordinate).
4. Immediate social situation (e.g., reunion, reintroduction, meeting, in
association with a disturbance, in play).
5. General social situation (e.g., breeding season, changes in group

composition, leadership changes).
Laboratory work and field studies will obviously focus on different be-
haviors but in all cases such imprecise terms as "normal," or even "ade-
quate," should be used with caution and delineation or not at all. Actual be-
havior elements or units should be used in preference to general terms such
as "estrous" or "receptive" behavior. If they are used, all such general
terms should be precisely defined. Likewise, specific terms, such as copula-
tion, ejaculation, or orgasm, cannot be assumed to mean the same thing in
different studies; definitions must be made. Appendices of definitions would
be a valuable addition to reports.
6. Observations: Testing or Study Conditions
Developmental studies are necessarily long term. The time invested can
be put to good use by collecting data on behavior patterns and interactions
throughout the early growth period until full adulthood and, even better,
through old age. Many studies end after 6 months of observation, when
follow-up studies could be done at 6-month intervals thereafter, or at salient
ages such as pUberty. Juvenile and adolescent behavior remains far more
obscure than infant and adult behavior. Longitudinal studies are of course
preferabl;::, but many more cross-sectional studies could be made.
Studies that completely ignore early rearing, and test only after pu-
berty or in adulthood, are not very useful for inferring developmental
processes, especially when few, or possibly incorrect, records have been
made of the subjects' experiences after puberty. (This is altogether too com-
mon in human studies.) Studies involving any laboratory-reared animal
should include good records of living conditions and exposure to other ani-
mals, even when spot observations cannot be made. Any systematic records
would be of great help, not only for publications, but also presumably for
the investigators and colony managers. Primates are not as freely available
as they were formerly, and all possible efforts should be directed toward
their conservation and reproductive success in captivity. Since a primate's
early socialization is very important to its ability to breed and bear
young-at minimum-all primate workers would stand to gain by an at-
tempt to understand the complex effects of early rearing. Better records are
necessary to help further this understanding (e.g., records of the effects of
hand rearing and multiple species rearing, as well as of the varieties of
mothers and others). Veterinarians and colony managers continue to
concentrate on cleanliness and convenience, and are surprised at the poor
breeding success of monkeys reared alone in colony cages (e.g., Honjo et
58 J. Hanby
al., 1974). Efforts should be made to disseminate knowledge of the effects

of early rearing.
Testing conditions are very important, since they are also experience
periods for the subjects. In addition to careful records of age of exposure,
length, behavior, etc., much greater attention should be paid to the
experience of the partner or partners and to female cycle state and be-
havior. Much less is known about the backgrounds of "feral" animals than
about those of laboratory-reared animals, and they should be used as test
partners with caution. Individual differences between animals are liable to
be great in adulthood as a result of special rearing and subsequent
experience, whether animals are wild born or captive. As adults, animals
with ambiguous backgrounds should be tested with a variety of partners in
order to establish some "baseline" behaviors. If animals are kept in a
group, observations should be made in some systematic way. Removals and
returns may affect both the individual and the group companions.
Testing conditions, such as cage size, home cage of the male or female,
etc., affect behavior, as do the special characteristics and number of indi-
viduals tested in the situation. Pairs, triads, and groups of all kinds each
constitute a special condition. Increases in complexity and group composi-
tion require an increased battery of measures and time spent in observa-
tions; if the cost is too great, simpler tests should be done. The aims of
investigations will continue to differ, but whether one is interested in the va-
riety and adaptability of sociosexual behavior in the group situation, or
solely interested in rearing animals that can breed in captivity, the attain-
ment of the desired aims will rest on an understanding of the ways in which
rearing, environment, experience, hormones, and individual/species charac-
teristics interact. Field and laboratory work will profit by better records and
a broader outlook.
7. Observations: Observers
A broader outlook and better observations require more staff, more

opinions, and teamwork. Field workers are notoriously independent, and
scientists in general are not noted for sociability and cooperativeness. The
fact that most of the work is done by male humans is reflected in the
concentration of research on male sexual behavior, male dominance, etc.
This is not to say that female humans are less biased, but that both sexes
should be part of a research tea,m.
Maturity on the part of observers would also help. Many observers are
very young or are themselves inexperienced sociosexually, and thus miss
many subtleties, or misread them, when watching primates interact. No ob-
server or researcher should presume that his or her biases, early rearing,
and sociosexual experience do not affect his or her observations and in-
terpretations.
8. Interpretation and Generalization
Needless to say, the most important part of research is interpretation

of the results. Great caution should be used in generalizing from very few
animals or situations (e.g., the conclusion that peers were more important
than mothers to rhesus monkey social behavior was based on the study of
very few animals in a very limited situation). Likewise, the utmost caution
is needed in comparisons of nonhumans with humans; but since such com-
parisons and extrapolations are constantly being made-and may be
necessary if we wish to understand our own behavior-workers must also
make more frequent evaluations of the information. In particular, this is a
recommendation that investigators, and especially those employed in
managing laboratories, delegate or encourage' periodic reassessment of the
work. Open and critical minds are necessary on the part of all concerned
persons, and more time and trouble should be taken in thinking out the way
the many experiential, hormonal, and social factors intertwine.
Social and sexual intercourse is of great importance to all primate
societies, and with appropriate care we should be able to increase our
understanding of our own sociosexual patterns by attending to the early
experiences of other primate species. It is to be hoped that we will also
someday attend more directly to human sociosexual development.
C. Conclusion
Nonhuman primates develop two distinguishable sociosexual systems

which I have termed the copulatory and contacting. The separation of these
two systems and varieties of the behavior is most apparent and well known
in free-ranging, terrestrial, Old World monkeys and apes that live in groups
containing both sexes and all ages. There is some degree of overlap in these
systems even in the wild. The more restricted the rearing of a primate in
terms of social experience with the mother and like and opposite sexes of all
ages, the less distinguishable, the more variable, and sometimes the more
distorted are its behavior patterns.
Human primates appear to exhibit perhaps the greatest degree of
overlap and the most variability in these sociosexual systems, from infancy
through adulthood. It is completely understandable that the word "love" in
many languages has two connotations: affectional and sexual. For a variety
60 J. Hanby
of reasons, some of which have been discussed, there is a delayed and all too
often distorted development of sociosexual patterns in some humans, to the
detriment of individual relationships and the entire society. Not all
societies, and certainly not all humans, restrict sociosexual expression in in-
fancy or adulthood, and free expression is not synonymous with stability
and mental and physical health. The point is that our behavior can be better
understood (and thereby perhaps be accepted or changed) by considering
carefully the accumulating evidence on the behavior of other primates in
both free-ranging and a variety of restricted conditions.
The importance of sociosexual patterns is obvious: they include greet-
ing, deferential, and affectional gestures, as well as sexual and maternal ges-
tures, all of which serve to keep individuals peacefully together. An area of
such importance deserves more research attention, discussion, revision, and
a broader perspective.
VII. ACKNOWLEDGMENTS
A pitcher of appreciation for David Bygott, who has portrayed the so-
ciosexual behaviors in this paper and has helped at all stages. A libation
al"o to Lilyan White, for her patience, tolerance, and physical-
psychological help; to Robert Hinde for inspiration and help; to those two
and to Anne and Michael Simpson, who together are still clearing a path to
the understanding of socialization processes; and to Carolyn Tutin, Hettie
and Franz Plooij, Mirreille Bertrand, Suzanne Chevalier-Skolnikoff, Anne
Mertl, Thelma Rowell, Tim Ransom, Carolyn Wilson, Anne Pusey, Mitzi
Thorndahl, Hans Kummer, and John Deag, all of whom contributed data
and/or photos, drawings, or discussions.
Thanks to the staff of the Oregon Primate Center, especially to
Charles Phoenix, Gray Eaton, Joel Ito, Margaret Barss; and to escapee
Bruce Alexander, who got me started.
Acknowledgment also to the Wisconsin Primate Center emigrants:
Jane and Conrad Brown, Robert Goy, Dave Goldfoot, and Kim Wallen.
A present to Lee Robertson, for his early help and his supportive
interest, and to John Fentress; and of course to Mum, for her tolerance of
our scientific endeavors. Grateful acknowledgments to the National
Science Foundation (U.S.A.) and to the Medical Research Council (Eng-
land) for funds and employment.
And finally, many, many thanks to the wonderful Phyllis Osbourn,
who typed the manuscript.
VIII. REFERENCES
Aldrich-Blake, F. P. G., and Chivers, D. J. (1973). On the genesis of a group of siamang. Am.
J. Phys. Anthro. 38:631-636.
Alexander, B. K. (1966). The effects of early peer deprivation on juvenile behavior of rhesus
monkeys. Ph.D. Univ. Wisconsin.
Alexander, M., and Perachio, A. (1970). An influence of social dominance on sexual behavior
in rhesus monkeys. Am. Zoo. 10(3).
Altmann, S. A. (1962). A field study of the sociobiology of rhesus monkeys, Macaca mulatta.
Ann. N.Y. Acad. Sci. 102:338-435.
Anthoney, T. R. (1968). The ontogeny of greeting, grooming, and sexual motor patterns in
captive baboons (Superspecies Papio cynocephalus). Behavior 31 :358-372.
Bakwin, H. (1952). Masturbation in infants. J. Pediat. 40:675.
Bakwin, H. (1971). Erotic feelings in children. Indian J. Pediat. 38:135.
Bakwin, H. (1972). Erotic feelings in infants and young children. Am. J. Dis. Chi/d. 126:52-54.
Beach, F. A. (1965). Retrospect and Prospect. In Beach, F.A. (ed.), Sex and Behavior, Wiley,
New York, pp. 535-569.
Beach, F. A. (1968). Factors involved in the control of mounting behavior by female mam-
mals. In Diamond, M. (ed.), Reproduction and Sexual Behavior, Indiana Univ. Press,
Bloomington, pp. 83-131.
Bernstein, I. S. (1967). A field study of the pigtail monkey (Macaca nemestrina). Primates
8:217-228.
Bertrand, M. (1969). The behavioral repertoire of the stumptail macaque. Bibliotheca
Primatologica No. 11. S. Karger, New York, pp. 1-273.
Bingham, H. C. (1928). Sex development in apes. Compo Psych. Mono. 5:1-161.
Boelkins, R. C. and Wilson, A. P. (1972). Intergroup social dynamics of the Cayo Santiago
rhesu6 (Macaca mulatta) with special reference to changes in group membership by males.
Primates 13:125-140.
Boese, G. K. (1973). Social behavior and ecological considerations of West African baboons
(Papio papio). Proc. IXth Int'l. Congress Anthropological and Ethnological Science,
Chicago.
Boiwig, N. (1959). A study of the behaviour of the Chacma baboon, Papio ursinus. Behaviour
14:136-163.
Bowlby, J. (1969). Attachment and Loss. Vol. I, Aitachment. Basic Books Inc., New York,
pp.I-428.
Burton, F. (1971). Sexual climax in Macaca mula/la. Proc. 3rd Int. Congr. Primatol., Zurich,
3:181-191.
Burton, F. D. (1972). The integration of biology and behavior in the socialization of Macaca
sylvana of Gilbraltar. In Poirier, F. E. (ed.), Primate Socialization, Random House, New
York, pp. 29-62.
Bygott, D. (1974). Agonistic behaviour and dominance in wild chimpanzees. Ph.D. thesis,
Cambridge University.
Carpenter, C. R. (1942). Sexual behavior of free ranging rhesus monkeys (Macaca mulatta). I:
Specimens, procedures and behavioral characteristics of estrus; II: Periodicity of estrous,
homosexual, autoerotic, and nonconformist behavior. J. Compo Psychol. 33:113-142, 142-
162.
Chalmers, N. R. (1972). Comparative aspects of early infant development in some captive Cer-
copithecines. In Poirier, F. E. (ed.), Primate SOcialization, Random House, New York,
pp.63-82.
62 J. Hanby
Chamove, A. S. (1966). The effects of varying infant peer experience on social behavior of the
rhesus monkey. M.A. Thesis, Univ. Wisconsin.
Chamove, A. S. (1973). Varying infant rhesus social housing. J. Institute An. Technicians
24:(1) 5-15.
Chance, M. R. A. (1963). The social bond of primates. Primates 4:1-22.
Chevalier-Skolnikoff, S. (197Ia). The ontogeny of communication in Macaca speciosa. Ph.D.
thesis, Dept. of Anthro., Univ. of Calif., Berkeley.
Chevalier-Skolnikoff, S. (197Ib). Homo- and heterosexual behaviour in stumptail monkeys
Macaca speciosa. Paper presented to the symposium on Biological and Cultural Bases of
Sex Role Differentiation, American Association for the Advancement of Science.
Philadelphia.
Chevalier-Skolnikoff, S. (1974a). The ontogeny of communication in the stumptail, Macaca
arctoides. Contributions to Primatology, Vol. I, Karger, Basel.
Chevalier-Skolnikoff, S. (1974b). Male-female, female-female, and male-male sexual behavior
in the stumptail monkey, with special attention to the female orgasm. Archives oj Sexual
Behavior 3(2):95-115.
Conaway, C. H., and Koford, C. B. (1965). Estrous cycles and mating behavior in a free-ranging
band of rhesus monkeys.J. Mammal. 45:577-588.
Conaway, C. H., and Sade, D. S. (1969). Annual testis cycle of the green monkey (Cercopithecus
aethiops) on St. Kitts, West Indies. J. Mammal. 50:833-835.
Crawford, M. P. (1940). The relation between social dominance and the menstrual cycle in the
female chimpanzee.J. Compo Psychol. 3:483-513.
Davenport, R. K., and Rogers, C. M. (1970). Differential rearing of the chimpanzee: A project
survey. The Chimpanzee 3:337-360.
Deag, J. (1974). A study of the social behavior and ecology of the wild Barbary macaque, Ma-
caca sylvanus. Ph.D. thesis, Bristol Univ.
de Benedictis, T. (1973). The behavior of young primates during adult copulation: Observa-
tions of a Macaca irus colony. Amer. Anthropologist 75:1469-1484.
DeVore, I. (1965). Baboon sexual behavior. In F. A. Beach (ed.), Sex and Behavior, Wiley and
Sons, New York.
Dierschke, D. J., Weiss, G., and Knobil, E. (1974). Sexual maturation in the female rhesus
monkey and the development of estrogen-induced gonadotropic hormone release. Endocr.
94:198-206.
Drickamer, L. C., and Vessey, S. H. (1973). Group changing in free-ranging male rhesus
monkeys. Primates 14(4):359-368.
Eibl-Eibesfeldt, I. (1970). Love and Hate, Methuen, London.
Erwin, J., Mitchell, G., and Brandt, E. (1973). Initial heterosexual experiences of three-year-
old rhesus macaques. Paper presented to annual meeting of International Society for
Developmental Psychobiology, San Diego, California.
Evans, C. S. (1967)., Methods of rearing and social interaction in Macaca nemestrina. Anim.
Behav. 15:263-266.
Ford, C. S., and Beach, F. A. (1951). Patterns oj Sexual Behavior, Harper and Brothers, New
York, 307 pp.
Freud, S. (1957). Three Essays on the Theory of Sexuality. In: The Standard Edition oj the
Complete Psychological Works oj Sigmund Freud, Vol. VII, Hogarth Press and the In-
stitute of Psycho-Analysis, London.
Galt, W. E. (1947). Sex behavior in primates. Ann. N.Y. Acad. Sci. 47:617-630.
Gartlan, J. S. (1969). Sexual and maternal behavior of the vervet monkey, Cercopithecus ae-
thiops. J. Reprod. Ferr. Suppl. 6:137-150.
Gartlan, J. S. (1975). Adaptive social structure of Erythrocebus patas. Symp. 5th Congr. Int.
Primatoi. Soc., Japan Science Press, Tokyo.
Goodall, J. van Lawick (1968a). The behaviour of free-living chimpanzees in the Gombe Stream
Reserve. Anim. Behaviour Mono. 1,165-311.
Goodall, J. van Lawick (1968b). A preliminary report on expressive movements and communi-
cation in the Gombe Stream chimpanzees. In Jay, P. C. (ed.), Primates-Studies in
Adaptation and Variability, Holt, Rinehart and Winston, New York and London, pp. 313-
374.
Goodall, J. (1973). Cultural elements in a chimpanzee community. In Montagna, W. (ed.),
Symp. 4th Int. Congr. Primat. Vol. I, Karger, Basel.
Gordon, T. P., and Bernstein, I. S. (1973). Seasonal variation in sexual behavior of all-male
rhesus troops. Am. J. Phys. Anthro. 38:221-226.
Gorer, G. (1938). Himalayan Vii/age: An Account oj the Lepchas oj Sikkim, M. Joseph,
London, pp. 7-519.
Goy, R. W. (1964). Reproductive behavior in mammals. In Lloyd, C. W. (ed.), Human Re-
production and Sexual Behavior, Lea and Febiger, Philadelphia, pp. 409-411.
Goy, R. W., and Goldfoot, D. A. (1973). Experiential and hormonal factors influencing
development of sexual behaviour in the male rhesus monkey. In The Neurosciences, Third
Study Program, M.I.T. Press, Cambridge, Mass.
Hall, K. R. L., and DeVore, I. (1965). Baboon social behavior. In I. DeVore (ed.), Primate Be-
havior, pp. 53-110, Holt, Rinehart, and Winston, New York.
Hamilton, G. V. (1914). A study of sexual tendencies in monkeys and baboons. H. Anim. Behav.
4:295-318.
Hanby, J. (1974). Male-male mounting in Japanese macaques. An. Behav., 22:836-849.
Hanby, J. (1975). The social nexus: problems and solutions in the portrayal of primate social
structures. In Symp. 5th Congr. Int. Primatol. Soc., Japan Science Press, Tokyo.
Hanby, J., and Brown, C. E. (1974). The development of sociosexual behaviours in Japanese ma-
caques M acaca juscata. Behaviour 49: 152-196.
Hanby, J., Robertson, L., and Phoenix, C. (1971). Sexual behavior in a confined troop of
Japanese macaques. Folia Primatol. 16:123-143.
Harlow, H. F. (1965). Sexual behavior in the rhesus monkey. In Beach, F. A. (ed.), Sex and
Behavior, Wiley and Sons, New York.
Harlow, H. F., and Harlow, M. K. (1962). Social deprivation in monkeys. Sci. Am.
207(5): 136-146.
Harlow, H. F., and Harlow, M. K. (1965). The affectional systems. In Schrier, Harlow, and
Stollnitz (eds.), Behaviour oj Nonhuman Primates, Vol. 1. Academic Press, N.Y. and
London,pp.287-334.
Harlow, H. F., and Harlow, M. K. (1969). Effects of various mother-infant relationships on
rhesus monkey behaviors. In Foss, B. M. (ed.), Determinants oj Injant Behavior, Vol. 4,
Methuen, London.
Harlow, H. F., Harlow. M. K., Dodsworth, R. D., and Arling, G. L. (1966a). Maternal be-
havior of rhesus monkeys deprived of mothering and peer associations in infancy. Proc.
Amer. Phi/os. Soc. 100(1):58-66.
Harlow, H., Joslyn, D., Senko, M., and Dopp A. (1966b). Behavioral aspects of reproduction
in primates. J. Anim. Sci. 25:49-67.
Herbert, J. (1970). Hormones and reproductive behaviour in rhesus and talapoin monkeys. J.
Reprod. Fert. Suppl. 11:119-140.
Hill, W. C. O. (1966). Laboratory breeding, behavioural development and relations of the
talapoin (Miopithecus talapoin). Extrait de Mammalia, 30(3):353-370.
64 J. Hanby
Hinde, R. A. (1971). Development of Social Behaviour. In Schrier, A. M., and Stollnitz, F.,
(eds.), Behaviour oj Nonhuman Primates, Vol. III, Academic Press, New York and
London, pp. 1-68.
Hinde, R. A., and Rowell, T. E. (1962). Communication by postures and facial expressions in
the rhesus monkey Macaca mulalla. Proc. Zool. Soc. Lond. 138:1-21.
Hinde, R. A., and Spencer-Booth, Y. (1967). The effect of social companions on mother-infant
relations in rhesus monkeys. In Morris, D. (ed.), Primate Ethology, Weidenfeld and
Nicolson, London.
Honjo, S., Fujiwara, T., and Cho, R. (1974). A comparison of breeding performance of indi-
vidual cage and indoor gang cage systems in cynomolgus monkeys. Paper presented at the
5th International Primatological Congress, Nagoya, Japan.
Jay, Phyllis (\965). The common langurs of North India. In Primate Behaviour: Field Studies
oj Monkeys and Apes. Holt, Rinehart and Winston, New York.
Kaufman, L. C. and Rosenblum, L. A. (1969). Effect of separation from mother on the emo-
tional behavior of infant monkeys. Ann. N.Y. Acad. Sci. 159:681-695.
Kaufmann, J. H. (1965). A three-year study of mating behavior in a free ranging band of
rhesus monkeys. Ecology 46(4):500-512.
Kellar, C. J., Beckwith, W. c., and Edgerton, R. B. (1968). Sexual behavior in the ARL
Colony chimpanzees.J. Nerv. Ment. Dis. 147:444-459.
Kinsey, A. c., Pomeroy, W. B., and Martin, C. E. (1948). Sexual Behavior in the Human
Male, Saunders, Philadelphia.
Kinsey, A. C., Pomeroy, W. 8., Martin, C. E., and Gebhard, P. H. (\953). Sexual Behavior in
the Human Female, Saunders, Philadelphia.
Koford, C. 8. (1963). Rank of mothers and sons in bands of rhesus monkeys. Science 141:356-
357.
Kummer, H. (1967). Tripartite relations in hamadryas baboons. In Altmann, S. A. (ed.),
Social Communication among Primates, Univ. of Chicago Press, Chicago.
Kummer, H. (\968). Social organization of hamadryas baboons. Bibliotheca Primatologica
No.6, S. Karger, New York.
Lancaster, J. B. (\972). Play-mothering: the relations between juvenile females and young in-
fants among free-ranging vervet monkeys. In Poirier, F. E. (ed.), Primate Socialization,
Random House, New York.
Lemmon, W. 8. (\970). Experiential factors and sexual behavior in male chimpanzees.
Medical Primatology, Proc. 2nd Conj. on Exp. Med. Surg. Primates, New York 1969,
Karger, Basel, pp. 432-440.
Lemmon, W. B., and Temerlin, J. W. (\974). The development of human-oriented courtship be-
havior in a human-reared chimpanzee female. Paper presented at the 5th International
Congress of Primatology, Nagoya, Japan.
Lewis, W. C. (1965). Coital movements in the first year of life. Int. J. Psycho-Anal. 46:372-
374.
Lindberg, D. G. (\969). Rhesus monkeys: mating season mobility of adult males. Science
16: 1176-1178.
Lorenz, K. (1966). On Aggression, Harcourt, Brace and World Inc., New York.
Loy, J. (\969). Estrous behavior in free-ranging rhesus monkeys Macaca multalla, a study of
continuity and variability. Ph.D. thesis, Northwestern Univ.
Marshall, D. S., and Suggs, R. C. (eds.) (1971). Human Sexual Behavior, Basic Books Inc.,
New York.
Maslow, A. H. (1936). The role of dominance in the social and sexual behavior of infra-human
primates. III. A theory of sexual behavior of intra-human primates. J. Genet. Psychol.
48:310-338.
Mason, W. A. (1960). The effects of social restriction on the behavior of rhesus monkeys. I.
Free social behavior.l. Compo Physiol. Psychol. 53:582-589.
Masters, W. H., and Johnson, V. E. (1966). Human Sexual Response. Little, Brown and Co.,
Boston.
Masters, W. H., and Johnson, V. E. (1970). Human Sexual Inadequacy. Little, Brown and
Co., Boston.
Maxim, P. E., and Buettner-Janusch, J. (1963). A field study of the Kenya baboon. Amer. J.
Phys. Anthrop. 21:165-180.
McGinnis, P. (1973). Sexual behaviour of chimpanzees. Ph.D. thesis, Cambridge.
Meier, G. W. (1965). Other data on the effects of social isolation during rearing upon adult re-
productive behaviour in the rhesus monkey Macaca mulatta. Anim. Behav., 13:228-231.
Mertl, A. (1963). Baboon social behaviour. M. A. thesis, Univ. of Chicago.
Michael, R. P. (1971). Neuroendocrine factors regulating primate behavior. In Martini, L.,
and Ganong, F. (eds.), Neuroendocrinology. Oxford Univ. Press, New York, pp. 359-398.
Michael, R. P., and Keverne, E. B. (1970). Primate sex pheromones of vaginal origin. Nature
225:84-85.
Michael, R. P., Wilson, M. and Plant, T. M. (1973). Sexual behaviour of male primates and
the role of testosterone. In Michael, R. P., and Crook, J. H. (eds.), Comparative Ecology
and Behaviour of Primates. Academic Press, London, pp. 236-313.
Missakian, E. A. (\969). Reproductive behavior of socially deprived male rhesus monkeys Ma-
caca mulatta. J. Compo Physiol. Psychol. 69:403-407.
Missakian, E. A. (1972). Genealogical and Cross-genealogical dominance relations in a group
of free-ranging rhesus monkeys M acaca mulatta on Cayo Santiago. Primates 13(2): 169-
180.
Moll, A. (\ 912). The Sexual Life of the Child. translated by Paul, E., George Allen, London.
Morris, D. (1967). The Naked Ape. Cape, London, 252 pp.
Nadler, R. D. (1975). Determinants of variability in reproductive behaviour of captive gorillas.
Symp. 5th Congr. Int. Primatol. Soc .. Japan Science Press, Tokyo.
Nagel, V. (1973). A comparison of Anubis baboons, hamadryas baboons and their hybrids at a
species border in Ethiopia. Folia Primatol. 19:104-165.
Nishida, I. and Kawanaka, K. (1972). Inter-United-Group relationships among wild chim-
panzees of the Mahali Mountains. In Umesao, T. (ed.) Kyoto Univ. African Studies. 7:1-
169.
Norikoshi, K., and Koyama, N. (1975). Group shifting and social organization among Japanese
monkeys. Symp. 5th Congr.Int. Primatol. Soc .. Japan Science Press, Tokyo.
Owens, N. (1973). The development of behaviour in free-living baboons Papio anubis. Ph.D.
thesis, Univ. of Cambridge.
Packer, C. (\ 975). Male transfer in olive baboons. Nature 255:219-220.
Parthasarathy, M. D., and Rahaman, H. (1974). Infant killing and dominance assertion among
the hanuman langur. Proc. 5th Int. Congr. Primatol .. Nagoya.
Paterson, J. D. (1973). Ecologically differentiated patterns of aggressive and sexual behavior in
two troops of Ugandan baboons Papio anubis. Am. J. Phys. Anthro. 38: 641-648.
Phoenix, C., Goy, R., and Young, C. 1967. Sexual Behavior: General aspects. In Martini, L.,
and Ganong, F. (eds.), Neuroendocrinology II. Academic Press, New York, pp. 163-196.
Ploog, D. and Maurus, M. (1973). Social communication among squirrel monkeys: Analysis by
sociometry, bioacoustics and cerebral radio-stimulation. In Michael, R. P., and Crook, J.
H. (eds.), Comparative Ecology and Behaviour of Primates. Academic Press, London and
New York, pp. 211-233.
Plooij, F. (1976). Behavioral development of infant chimpanzees. Ph.D. thesis.
Pusey, A. (1974). Intercommunity transfer of chimpanzees in Gombe National Park. Wenner-
66 J. Hanby
Gren Conjerence on the Great Apes (to be published by Holt, Rinehart, and Winston, New
York). _
Rahaman, H., and Parthasarathy, M. D. (1968). The expressive movements of the bonnet ma-
caques. Primates 9:259-272.
Ransom, T. W., and Rowell, T. E. (1972). Early social development of feral baboons. In
Poirier, F. E. (ed.), Primate Socialization, Random House, New York, pp. 105-144.
Resko, J. (1967). Plasma androgen levels of the rhesus monkey: Effects of age and season. En-
docr.81:1203-1225.
Resko, J. A., Malley, A., Begley, D., and Hess, D. (1973). Radioimmunoassay of testosterone
during fetal development of the rhesus monkey. Endocrinology 93:156-161.
Reynolds, V., and Reynolds, F. (1965). Chimpanzees of the Budongo Forest. In DeVore, I.
(ed.), Primate Behavior: Field Studies oj Monkeys and Apes, Holt, Rinehart and Winston,
New York.
Richards, S. (1972). Tests for behavioural characteristics in rhesus monkeys. Ph.D. thesis,
University of Cambridge.
Rogers, C. M., and Davenport, R. K. (1969). Sexual behavior of differentially-reared chim-
panzees. Proc. 2nd Int. Congr. Primatol.. Vol. I, S. Karger, Basel and New York.
Rose, R. M., Gordon, T., and Bernstein, I. S. (1972). Plasma testosterone levels in the male
rhesus: Influences of sexual and social stimuli. Science 178:693-695.
Rose, R. M., Holladay, J. W., and Bernstein, I. S. (1971). Plasma testosterone, dominance rank
and aggressive behavior in male rhesus monkeys. Nature 231 :366-368.
Rowell, T. E. (1967a). Female reproductive cycles and the behavior of baboon and rhesus ma-
caques. In Altman, S. A. (ed.), Social Communication in Primates. Univ. of Chicago Press,
Chicago, pp. 15-32.
Rowell, T. E. (1967b). A quantitative comparison of the behaviour of a wild and a caged ba-
boon troop. Anim. Behav. 15:499-509.
Rowell, T. E. (1972). Female reproductive cycles and social behavior in primates. Adv. Study
!Jehav. 4:69-105.
Rudran, R. (1973). Adult male replacement in one-male groups of purple-faced langurs
(Presbutes senex senex) and its effect on population structure. Folia Primatol. 19: 166-192.
Saayman, G. S. (1970). The menstrual cycle and sexual behavior in a troop of free-ranging
chacma baboons, Papio ursinus. Folia Primatol. 12:81-111.
Sade, D. S. (1968). Inhibition of son-mother mating among free-ranging rhesus monkeys. Sci.
and Psychoanal. 12:18-38.
Savage, E. S., Temerlin, J., and Lemmon, W. B. (1974). The appearance of mothering be-
havior toward a kitten in a human reared chimpanzee. Paper presented at the 5th Interna-
tional Congress of Primatology, Nagoya, Japan.
Schloeth, R. (1956-7). Zur psychologie der begegnung-zwischen tieren. Behavior 10: 1-80.
Shepher, Joseph (1971). Self-imposed incest avoidance and exogamy in second generation kib-
butz adults. Ph.D. thesis, Rutgers Univ., The State Univ. of New Jersey.
Simonds, S. E. (1965). The bonnet macaque in South India. In DeVore, I. (ed.), Primate Be-
havior: Field Studies oj Monkeys and Apes, Holt, Rinehart and Winston, New York, pp.
175-196.
Singer, I., and Singer, J. (1972a). Periodicity of sexual desire in relation to time of ovulation in
women.J. Biosoc. Sci. 4:471-481.
Singer, I., and Singer, J. (1972b). Types of female orgasm. The Journal oj Sex Research
8:(4):255-267.
Southwick, C. H., Beg, M. A. and Siddiqi, M. R. (1965). Rhesus monkeys in N. India. In
DeVore, I. (ed.), Primate Behavior, Holt, Rinehart and Winston, New York.
Spencer-Booth, Y. (1969). The effects of rearing rhesus infants in isolation with their mothers
on their subsequent behavior in a group situation. Mammalia 33:80-86.
Stephenson, G. R. (1973). Testing for group specific communication patterns in Japanese ma-
caques. Symp. 4th Int. Congr. Primatol., Vol. I, Karger, Basel, pp. 51-75.
Struhsaker, T. (1967). Behavior of vervet monkeys Cercopithecus aethiops U. Calif. Publ.
Zool.82:1-74.
Sugiyama, Y. (1965). On the social change of hanuman langurs Presby tis entellus in their
natural condition. Primates 6,381-418.
Sugiyama, Y., Yoshiba, K., and Pathasarathy, M. D. (1965). Home range, mating season,
male group and inter-group relations in hanuman langurs Presby tis entellus. Primates
6(1):73-106. .
Tokuda, K. (1961). A study on the sexual behavior in the Japanese monkey troop. Primates 3:1-
40.
Tokuda, K., Simons, R., and Jensen, G. (1968). Sexual behavior in a captive group of pigtailed
monkeys Macaca nemestrina. Primates 9:283-294.
Tutin, C. (in preparation). Mating patterns in a community of wild chimpanzees. Ph.D. thesis,
Univ. of Edinburgh.
Tutin, C. (1974). Exceptions to promiscuity in a feral chimpanzee community. Paper read at the
5th International Congress of Primatology, Nagoya, Japan.
Tutin, C. E. G., and McGrew, W. C. (l973a). Sexual behaviour of group-living adolescent chim-
panzees. (Proc. IVth Int. Congr. Primat.) Am. J. Phys. Anthro. 38:195-200.
Tutin, C. E. G., and McGrew, W. C. (1973b). Chimpanzee copulatory behaviour. Folia
Primatol. 19:237-256.
van Wagenen, G. (1972). Vital statistics from a breeding colony: Reproduction and pregnancy
outcome in Macaca mulatta. J. Med. Prim. 1:3-28.
van Wagenen, G., and Simpson, M. E. (1954). Testicular development in the rhesus monkey.
Anat. Rec. 118(2):231-251.
Weber, I. (1973). Tactile communication among free-ranging langurs. Am. J. Phys. Anthro.
38(2):481-486.
Whiting, B. (ed.) (1966). Six Cultures: Studies of Child Development, Wiley, New York.
Wickler, W. (1967). Socio-sexual signals and their intra-specific imitation among primates. In
Morris, D. (ed.), Primate Ethology, Weidenfeld and Nicolson, London.
Wickler, W. (1972). The Sexual Code, Doubleday, New York.
Yamada, M. (1966). Five natural troops of Japanese monkeys in Shodoshira Island. I. Dis-
tribution and social organization. Primates 7(3):315-362.
Yerkes, R. M. (1939). Social dominance and sexual status in the chimpanzee. Q. Rev. Bioi.
14:115-136.
Zuckerman, S. (1932). The Social Life of Monkeys and Apes, Kegan Paul, London.
Zumpe, D., and Michael, R. P. (1968). The clutching reaction and orgasm in the female rhesus
monkey Macaca mulatta. J. Endocr. 40:117-123.
Chapter 2
HOMEOSTATIC MOTOR PROCESSES IN

MAMMALIAN INTERACTIONS: A
CHOREOGRAPHY OF DISPLAY l
Han Golani
Tel-Aviv University
Tel A viv, Israel
"All advances and flows, trembles and aspires"

-R. A. Y. H. Kook, Orot Hakodesh
I. ABSTRACT
The Eshkol-Wachmann movement notation is used to describe motor se-

quences in the interactions of both golden jackals (Canis aureus) and
Tasmanian devils (Sarcophilus harrisii). Motor behavior is rigorously
described in terms of the elementary movements of limb segments. The
same movements are described in four coordinate systems: in relation to the
animal's own body, in relation to the environment, in relation to a partner,
and in relation to the topography of the contact point with the partner on
the animal's own body.
This analysis reveals several directive processes, termed "routes of con-
vergence," which appear in the behavior of individuals and in interactions.
Some routes of convergence are homeostatic. In interactions, they often ap-
pear as a process of steady maintenance of a fixed contact point between
the two animals. This fixed contact point is described as a "joint" that
seemingly forces the two interacting animals to move in unison as one
1 This paper is dedicated to my parents and to Michal. This study was supported by a Ford Foun-
dation Grant B-9. The paper was written under the auspices of the Smithsonian Institution,
Washington, D.C., U.S.A., and the Dept. of Psychology, Dalhousie University, Halifax, N.S.,
Canada.
69
70 lIan Golani
"superorganism." It is shown that a large number of previously described

fragmentary postures in mammalian display are nothing but "twists" within
a continuous process of joint maintenance. The options available for the
animals in terms of the sequencing of joints are conveyed in "kinetic field"
diagrams, which describe the location of joints and the trajectories between
them on the animals' own bodies. As is shown, these diagrams describe the
structuring of individual space during interactions.
In this approach the emphasis is on: (1) describing the moment to mo-
ment interaction in a language which is appropriate for the description of
the rich world of movement; (2) making the observer's contribution to the
description as explicit as possible; (3) focusing on the motor strategies that
lead to common final positions, rather than on the final positions to which
the movements converge; (4) suspending judgment as to the possible
motivational, ontogenetic, phylogenetic, or physiological mechanisms of
movement.
The rigorous description of display, and the demonstration of its
directedness, can serve as a beginning of a zoosemiotic of mammalian dis- .
play, as a description of the demand made upon the central nervous system,
and as a baseline for quadrupedal mammalian display.
II. INTRODUCTION
The study of movement involves the use of a particular language. Lan-

guage creates a disposition of thought by ordering attention and directing
action in a specific mode (e.g., Whitehead, 1929; Bohm, 1971; and Whorf,
1956). Distinctions made by language can prevail all the way through to the
terminal results of a specific study. These distinctions tacitly imply the
route and the end results of the study, since a distinction does not stand
only for itself, but also for all that it was distinguished from. As Whitehead
puts it:
... every proposition [e.g., series of distinctions 1 refers to a universe exhibiting some
general systematic ... character ... Nothing has been defined, because every definite
entity requires a systematic universe to supply its requisite status. Thus, every proposi-
t~on proposing a fact must, in its complete analysis, propose the general character of the
universe required for that fact. There are no self-sustained facts floating in nonentity
(Whitehead, 1929, p. 17; see also Spencer Brown, 1969).
Thus, to make several distinctions is tantamount to implicitly creating

a whole universe of discourse. To use a specific language as a tool in a study
means to use a series of distinctions that splice reality and create "facts"
which are later used to illustrate some initial insights.
Homeostatic Motor Processes in Mammalian Interactions 71
For example, the use of human emotional and functional language for
the description of animal movements already implies the result that animals
have drives, and that these drives sometimes conflict (e.g., Fisher, 1965).
Baerends' choice to describe guppy motor display in terms of a language
that employs intensities of behaviors, e.g., lower and higher intensities of the
Sigmoid posture, etc., already implied the result that guppies have underly-
ing continuous drives of varying intensities (Baerend, et al., 1955). Some
students of behavior, following Tinbergen's (1951) dictum, attempted to
avoid such "biased" distinctions that were made in the first steps of
ethological descriptive studies. They felt that behavioral events should be la-
beled by "neutral" descriptive terms, and they used discrete, fragmented la-
bels, such as "zigzagging" and "ovaIling" in Nelson (1965), and "bobs
head" and "holds tail erect" in Altman (1965). They started with a frag-
mented, alienated language, and ended up with stochastic models. Their
pledge for initial ignorance as to the connectedness of motor behavior
shaped itself initially in a language of separate unrelated labels, and ended
up with models which described this basic ignorance in scientific terms (see
Poincare, 1952, p. 64, on the probabilistic approach as an expression of ig-
norance). This should not be interpreted as a value judgment, but rather as
an effort to indicate how initial distinctions later emerge as scientific
results. It is our initial insight that, unless modified by a new insight during
the course of a study, frames both our initial percepts and language in such a
way that we can later support and establish this initial insight (see also Fried
and Agassi, 1974). In this respect, science may be viewed as a logically valid,
rigorously verified illustration for an entertained metaphysics.
If, indeed, terminal results are extensions of initial distinctions, then
the choice of a particular descriptive language is crucial.
This paper is based on the premise that an holistic theory of motor be-
havior must be founded on more consistent and coherent descriptive lan-
guages. One approach to this ideal is to develop and employ notation
systems with a richness similar to that found in the system widely and suc-
cessfully used for musical notation. To the same extent that it is inconceiva-
ble to describe and analyze music either through the use of everyday verbal
language or through the use of fragmentary "neutral" terms, it is incon-
ceivable to study the specific field of motor behavior through a language
that is incapable of conveying the richness of movement. I believe that the
Eshkol-Wachmann (E-W) movement notation, which I have been using in
my studies of the golden jackal (Canis aureus) and the Tasmanian devil
(Sarcophilus harrisii), is capable of conveying this richness.
Developed in Israel (Eshkol and Wachmann, 1958), the notation has
already been successfully applied to the description of a variety of subjects,
such as classical ballet (Eshkol et al., 1968), physical education (Eshkol et
72 lIan Golani
al., 1969), folk dances (Eshkol et al., 1970b, 1972a), sign language ofthe deaf
(Eshkol et al., 1971), the Feldenkrais physiotherapeutic method (Eshkol et
al., 1972b), the precopulatory behavior of the golden jackal (Golani et al.,
1969), the generation and movement of plane form in graphic kinetic art
forms (Harries, 1969), and the study of the nature of simultaneous move-
ment of adjacent limb segments, as plotted in Cartesian space (Eshkol et
al., 1970a).
Since the notation describes the angles obtaining between all the limb
segments of the vertebrate body, motor behavior can be fully reconstructed
from it without any prior knowledge of the behavior. This is not the case with
some other notations (e.g., Laban, 1956; Birdwhistell, 1970). The motor be-
havior of an animal is described as an "orchestration" of movements of limb
segments which extend temporally over each other. Both the discrete and the
continuous aspects of movement are conveyed, so that movement is no longer
conceived as a sequence of one event at a time or as a mosaic of successive
instantaneous configurations of discrete events (e.g., Golani, 1973). The
contribution of the observer to the description becomes explicit in the specifi-
cation of the coordinate system in which the movements are described-
either bodywise, or environmentwise, or partnerwise-and the observer has
the freedom to shift from one coordinate system to another. Unlike other
globographic techniques (e.g., Albert, 1876; Strasser and Gassman, 1893;
Dempster, 1955; Prost, 1974a), which describe the behavior in only one coor-
diTlate system in the form of continuous graphs, the E-W notation provides
an articulated conceptual framework, i.e., a language which is a disposition
for thought. Finally, the E-W movement notation provides a language with
which one can relate to motor behavior as a process, without reverting to a
cybernetic language that uses a distinction between input and output. Such a
distinction is speculative as long as the study is limited to the behavioral level.
In the following, I will outline in some detail the above principles, as
well as several others that I entertain; demonstrate how these principles find
their expression in the E-W movement notation; and show how I am apply-
ing this notation.
In some cases I found it necessary to provide a somewhat detailed ac-
count of the behavior. Readers who find this exposition difficult may skip it
and follow the argument where it is resumed.
III. THE DIMENSIONS
Movement is a multidimensional process, the dimensions being body,

time, and space. One of the determinants of order in movement is anatomy.
Therefore, in the E-W movement notation, the partitioning of the body di-
mension is determined by the vertebrate joints, i.e., each limb segment is a
separate part along the body dimension. Each animal is described in terms
of a series of serially connected limb segments (Fig. 1).
When a limb segment moves about a "fixed" joint, the surface, or
"paths," produced by its movement are confined within an abstracted
sphere, which is carried along with the joint. The extremity of the limb seg-
ment moves on the surface of the sphere (Fig. 2). A coordinate network on
the surface of the sphere describes the position of the far end of the limb
segment in terms of its displacement on both the horizontal and the vertical
planes (Fig. 3). Thus, within a given degree of resolution, every position
may be specified by two numerals.
A movement is described in terms of the initial position, the final posi-
tion, and the trajectory between the two positions. The trajectory can be
described by specifying the sense of the movement (clockwise or anti-
clockwise), and the angle obtaining between the axis of the limb segment
and the axis of movement throughout the movement. The axis of movement
is the axis of symmetry of the movement (Fig. 4). As a matter of conven-
tion, three kinds of movement are described: rotatory movements, in which
the axis of the limb and the axis of movement coincide; conical movements,
in which the angle of movement is less than 90°; and plane movements, in
which the angle of movement is 90° (Fig. 4). This partitioning may be
avoided by specifying in one continuous system the angle between the mov-
ing limb and the axis of movement. For example, a neck movement, from a
forward horizontal position to an oblique horizontal position 45° clockwise,
Fig. I. A pair of golden jackals during precopulatory behavior. The

superimposed bars indicate the body parts that were considered as
separate limb segments during the present study.
74 lIan Golani
o~
, --- ---
" '-
---
----,---
"-
..........-,-- - /
-------~-
I
/
/'
Fig. 2. The extremity of a limb segment moves on the surface of

an imaginary sphere. (By permission of the Movement Notation So-
ciety, Israel.)
through a trajectory in which the jackal maintains his neck on the same
level with regard to the joint with the torso, is specified as an injunction for
movement,
Iml~R I----;Lm I
which reads: go from mto [n, for six time units, in a plane movement. The
trajectory is specified as a plane movement, since the angle obtaining
between th~ axis of the neck and the axis of movement (which is, in this
case, perpendicular to the axis of the neck) is 90°. In the above example,
the chosen degree of resolution is 45°; therefore, the number above the ar-
row specifies one amount of movement. The temporal and spatial degrees of
resolution depend, of course, both on the interest of the observer and on the
limitations of his recording equipment.
Note that the spherical coordinate system in which the movement is
described is determined, or rather induced, by the vertebrate skeletal
anatomy. Instead of movement being described in a homogeneous, diffuse,
and highly abstracted Cartesian space, which is external to and independent
@r--r---~
~:--+--....l@
rig. 3. The E-W coordinate system. For each pair of numerals, the
lower indicates the horizontal, and the upper the vertical coordinate.
In the present study one unit of displacement equals 45°. (By per-
mission of the Movement Notation Society, Israel.)
\\ \, \\
\-1\
I \
\-1-
I \ \~
.\
.~. \ ...
\
i \ \
~.-.
"\... ...
\I
L-
'-..
\. \I
!::::=-
..\
Fig. 4. One of the criteria that define an elementary movement is the angle
obtaining between the axis of movement and the axis of the moving limb seg-
ment. (By permission of the Movement Notation Society, Israel.)
76 Han Golani
of the animal, it is described in a structured space, the structure stemming

from the animal's skeletal anatomy.
The fact that the vertebrate body is composed of several limb segments
is represented on the score page by a series of horizontal lines, one line for
each segment. The temporal dimension is represented on the score page by
columns representing successive time units (Fig. 5).
IV. THE SPLICING OF THE FLOW
A. Definition of an Actual Elementary Movement
An actual elementary movement is deJined as a movement perJormed

by a limb segment, and is specified by Jour variables: the initial position; the
angle between the moving limb and the axis oj movement; the sense oj the
movement (clockwise or anticlockwise); and the speed of the movement. A
constant rate of change in the first or the last of the above variables still
defines one elementary movement. A discrete change in any of these varia-
bles terminates one movement and starts another. As a matter of conven-
tion, actual movements with zero speed are termed actual positions.
The term "actual" is used in this paper in Whitehead's sense of a
concrete event, in order to maintain a distinction between various levels of
abstraction. For example, the abstracted term of a plane movement of the
head from [~] to m may represent a set of actual, concrete movements that
were indeed performed by one or more animals within specific contexts.
Actual movements and positions are the most concrete elements of motor
behavior defined within this system. Further reduction would "shift" us
from the level of the movement of limb segments to the level of muscle
contractions.
B. The Necessity for At Least Two Descriptions
At least two descriptions are required Jor an actual movement as a

basis Jor Jurther analysis. The fact that actual movements are atomistic
does not in any way mean that they are simple. In the E- W movement
notation, an actual movement is described in terms of (I) an injunction for
a movement, and (2) the end position of this movement. For example, the
following description:
d
6 +--1 second---..;. 7
Eyes F F
Ears [iJ" 'J
I.!J [~JA m
Head [6J H [8]
Neck [;J [6]
Torso [8) "" >- " [~
Pelvis (t,
Tail [~l -vl,-- [8] [tl l..- I-- [.:]
Front legs
Back legs
Weight
Front
Miscellaneous
Relative body
positions
Eyes I
Ears [~] til ~r
Head [2] ~ [gl "- [~] [6] ~ [:;] :.l [g]
Neck [b] '::J [0;] (<I- [S] /I Jg: ~ [,,] iL. [0';] K Pt]
i::>rso [8]
Pelvis 6) 61=
Tail [~ 31 [~]
I1IIIIIIIffilllillf II H!HHIfP
Front legs
Back legs
1(')IIIIH~(")lllltHfMIIII
Weight
Front
Miscellaneous
Fig. 5. A motor score of four seconds of jackal precopulatory behavior. The numerals on top
indicate time in seconds. The first column on the left describes the initial position of the two
animals. Arcs indicate the duration of movements. ~ and ...... , T and L stand for positive
(clockwise), and negative horizontal, and positive and negative vertical plane movements
respectively. Ii and U stand for positive and negative shifts of front and rotations. Oblique ar-
rows stand for movements with both horizontal and vertical components. Numerals in brackets
indicate initial and final positions, and the different brackets indicate the coordinate system in
which the behavior was described (see Section C). Other numerals indicate amount of move-
ment. 1 indicates walking.
78 lIan Golani
includes an injunction for a plane movement of two amounts of movement, or

90°. The result of thIs injunction is that the limb will end up in position m
whether specified or not, so that specification of the end position seems re-
dundant. However, this apparent redundancy makes explicit these two funda-
mental properties of the actual movement: its actual genesis'l which
essentially identifies it; and its end point, or consequence, which is the start-
ing point for the following actual movement (description by consequence).
The coherence of an actual movement, i.e., its significance, or function, in
the restricted sense of its effect on other actual movements and positions of
the same or another animal, can potentially be derived from either aspect of
the movement-the constrained flow (trajectory) or the fixed (final
abstracted position) aspect-depending on the coordinate system in which
the movement was initially described and later examined. For example, if a
certain movement involves the steady maintenance of a particular abstracted
position, then one way to look at it is to take the movement as given, so that
its significance is disclosed in its consequence, i.e., the steady maintenance of
the abstracted position. Another way is to assume the final abstracted posi-
tion as given, and then the trajectory of the movement itself is viewed as a
consequence, i.e., as a movement of the limb segment around the given,
constant position. I will show later that localization is an abstraction based
on abstracted final positions, whereas coordination and integration are con-
cepts based on the movements to these positions.
The score of a sequence of actual movements and actual positions can
also be conceived of as an occurrence graph, in Holt's sense (Holt, in
Bateson, 1972, Chapter 4), where the arcs describe stable intervals of condi-
tion holding, and the "points"-the abstracted positions-describe the
instant of change, after which new conditions hold.
Actual positions are concrete behavioral events with temporal thickness.
Abstracted positions are not concrete; they are abstracted instantaneous
boundaries of junctions between actual movements and actual positions.
These abstracted positions disclose the fixed and discrete aspects of motor
behavior, but they are not real biological events since they have no temporal
extension. For example, by tracing a continuous figure 8 with my stretched
arm, starting from the center of the 8, I perform two actual movements with
my upper arm, namely, two complete conical movements. The two actual
movements are bridged by an imaginary abstracted position between the two
full cones. By starting from the top of the figure 8, three actual movements
are performed, because there are three instantaneous points where new condi-
tions hold. In the same way, every actual position is bounded by two
abstracted (identical) positions. The first abstracted position indicates the
instant of change after which the new conditions of the actual position hold,
and the second indicates a further change to still new conditions.
Imaginary contact positions between actual continuous movements are

the basis for the abstracted concept of instantaneousness.
C. The Possibility of Several Descriptions
An actual movement can have several descriptions, each consisting oj a

trajectory and an abstracted jinal position, depending on the number oj
coordinate systems in which it is described. The E-W sphere, which serves
as the coordinate system, is attached to the joint of the examined limb at
the root of the movement, and is carried along with the next proximal limb.
The coordinates which are projected on the sphere can be induced by the
next proximal limb, so that the description of the limb's movement is
"bodywise"; it can be induced or projected by the environment so that the
description is "environmentwise," or "absolute"; and it can also be
projected by an interacting partner, so that the description is "partnerwise,"
etc. (Fig. 6).
A fourth frame of reference, devised by Eshkol specifically for the
study of relative body positions in jackal interaction sequences, describes
Bodywise Absolute Partnerwise
Vertic~1
en
Q)
~
<tI
iLateral view)
c:
...
"0
~ ~-€B
0
0 Horizontal
()
(Dorsal view) 3 7 3 2 I 3 -,0 3 2 2 1
, '0
2 '
c:
0
.~
Head [~J.>-------[~j (12.)~(~) (~.};:::----(~}
<tI
[U~[~
~
0 Neck (~)~~) (~V------(~+}
z
Fig. 6. The same head and neck movements are schematically illustrated by. stick figures, and
notated in three different coordinate systems. The horizontal bar on the left of each illustration
stands for a stationary torso. The neck bar and the head bar are serially connected to it. For
each coordinate system, initial (left), and final positions are illustrated. Lateral view shows
vertical, and dorsal view horizontal coordinates. [ 1 stands for a "bodywise," ( ) for an
"absolute," and I I for a "partnerwise" coordinate system. The movements are notated
under each set of illustrations. The lower numerals within the brackets indicate horizontal
coordinates, and the upper numerals vertical coordinates. Obviously, in the case of a part-
nerwise description, the partern induces the coordinates, and once he moves the coordinates
move with him. In the above diagram, however, the partner was considered to be stationary for
the sake of simplicity.
80 lIan Golani
the movements of an animal in terms of their topographical consequences

on the animal's own body. The grid of the E-W sphere is superimposed
schematically on the animal's own body, as if the animal carried it along.
Figure 7 describes the horizontal values attached to the animal's body.
Any change in the relationship between the environment and the
animal is described in terms of its topographical consequence on the
animal's own body. A female may be shifted from one imaginary or real
contact point to another on the male's body through any combination of the
male moving around the female, the male freezing and waiting for the fe-
male to move around, or both animals moving simultaneously or suc-
cessively. In relation to the shift on the male's body, however, all of the
above movements may in principle have identical consequences in terms of
topographical end points.
The biological sense of such a description is that it is a description by
consequence, as if the animal "is at the center of the universe, coordinating
it from an unchanging position" (quoted by Whitehead, 1923, p. 40, as
"Wildon Carr's excellent phrase"). The animal is continuously manipulat-
ing the environment around its own body, either by freezing and letting the
environment move around, or by moving' together with the environment
thereby keeping it constant, or by moving alone, etc. Description in this
frame of reference shifts us from a Newtonian homogeneous conception of
individual space to a highly structured epistemological space. (For further
discussion of the two types of space-the a priori Newtonian or Kantian
space, which exists as if it were independent of the animal's movements and
Fig. 7. The horizontal aspect of the E-W coordinate system

3 superimposed schematically on the surface of the body of a male
Tasmanian devil. In defining this system, the longitudinal axis of
the animal's body is imagined as coinciding with the axis of the
coordinate system. Clearly, the animal is not in fact spherical, but
by regarding it as a modification of a sphere the positions of the
coordinative grid of the sphere can be mapped into the surfaces of
the animal. In the above modification, only horizontal coordinates
were attached. 0 stands for front, 4 for back, 2 for right, and 6 for
left. H,N,C, and P stand for head, neck, torso, and pelvis respec-
tively. The arrows indicate that the male shifted the contact point
with the female from II to the right flank of his head, then to the
right flank of his neck, and then to midway between neck and
torso where the contact point shifted minimally back and forth.
"outside" of the animal; and. the a posteriori, epistemological, relativistic,

Poincare type space, which is constructed in ontogeny by the animal's own
movements-see below).
By describing the same behavior within several coordinate systems,
various aspects of the movement are brought into attention: regularity in
respect to the animal's own body emerges in a bodywise description; regu~
larity in respect to the environment emerges in an "absolute" coordinate
system; and regularity in respect to a partner emerges in a partnerwise
description. Altman's (1962) suggestion that the social context should never
be a part of the definition of a pattern is therefore replaced by the opposite
suggestion, making the social context part of the definition; unless a be-
havior is described in the appropriate context or coordinate system its regu-
larity might not emerge.
The observer's methodological and technical freedom to examine be-
havior from several points of view, by shifting from one coordinate system
to another, is one of the most powerful properties of the E-W movement
notation, in that it explicates the nature of the interaction between the ob-
server and that which is observed. Such interaction is immanent to the
process of description (see also Maturana, 1970).
The fact that the various limbs of an animal, or even one and the same
limb, move in a patterned manner with respect to several coordinate
systems, whether simultaneously or alternatingly, is in part the source of the
ambiguity :)f motor behavior.
D. Anatomy and Movement
The skeletal anatomy oj a vertebrate partly shapes its actual move-

ments and actual positions. The anatomical bounds oj joints can provide a
theoretical jrame oj rejerence within which these actual movements can be
examined. The insight that anatomy and behavior are extensions of each
other was one of the foundations of classical ethology. Heinroth (1911) and
Whitman (1899) related initially to animal behavior as a domain that
further extended their zoological insight, which was in turn originally based
on a profound knowledge of anatomy.
Later workers, like Lorenz, in using a motivational language for the
description of behavior, redirected their attention from their earlier interest
in the interface between anatomy and behavior to the interface between be-
havior and the central nervous system. Their followers, who still showed
vivid interest in jorms of behavior, diverged even further from the insight
that anatomical structures give shape to forms of behavior, and centered
their attention almost exclusively on the action of hypothetical brain
82 lIan Golani
mechanisms as formative forces on behavior. Once relatively stereotyped

forms of movement, termed "displays," were discerned, they were auto-
matically related to centrally programmed nervous command. The concept
of species-specific stereotyped behavior presupposed-whether implicitly or
explicitly-the central programming of a chain of a highly complex nervous
output, mainly in the form of a command (as distinct from a chain of control
whereby the animal is continuously modifying its behavior) which ·flows all
the way through after its triggering to its full expression as a gestalt. (See
McCulloch, in Bateson, 1972, p. 204, for the distinction between control and
command; and Barlow, 1968, for the concept of the triggering of a fixed ac-
tion pattern.)
Notwithstanding the nature of the contribution of the central nervous
system, it is an almost trivial fact that an animal's display is qualified by its
anatomy. Anatomy shapes actual movements and actual positions both by
giving them rough form (in the sense that a bird's display must differ from a
mammal's) and by forming their bounds (in the sense that freedom can be
experienced only within the limits of a given skeletal anatomy).
On the level of interspecific comparisons, anatomical constraints must
be sifted out before central programming can be assumed as a formative
force. For example, during precopulatory motor interaction sequences,
golden jackals (Canis aureus) may rotate either their necks or their heads or
both. 2 Horizontal flexions of the torso and neck are abundant. During
sili.ilar sequences, Tasmanian devils (Sarcophilus harrisii) mainly rotate
their necks, hardly flexing their necks in the horizontal plane. The dif-
ferences between these formal aspects of species-specific behavior are due to
the fact that a Tasmanian devil can hardly rotate its head on its neck be-
cause of the bulky head musculature and the structure of the relevant joints,
and its bulky and short neck can hardly flex horizontally. On the other
hand, due to their ability to shift easily from a digitigrade to a plantigrade
stance, Tasmanian devils perform many pelvis and torso movements in the
vertical plane, by generating a variety of combinations, such as lowered
hindquarters and raised forepart, raised hindquarters and lowered forepart,
lowered hindquarters and lowered forepart, etc., which are seldom observed
injackals.
On the level of intraspecific study, anatomically determined conditions
provide the range or bound within which actual movements and actual posi-
tions are carried out. The specification of what part within this bound is
indeed exploited (or, occasionally, not exploited) for specific limbs in
specific contexts might prove instructive. Such bounds can be described
2 Because of the elongated skull, head rotations where the axis of rotation lies between the
condylles and the forehead look like, and were described in the present paper as, plane move-
ments. Jackals can rotate their heads up to 45° to each side.
within the E-W theoretical sphere, and provide a frame of reference in

which observed actual movements can be examined in relation to all ana-
tomically possible movements for both individuals and species. A considera-
ble advance along these lines has already been carried out by Dempster
(1955), who described the range of motion for some human cadaver joints,
and by Prost (1967, 1974a, and in press), who described ranges of motion for
humans, gibbons, and baboons. Both authors describe ranges of movement
within a globographic coordinate system.
Figures 8 and 9 present summaries of actual head and neck movements
performed during ten sequences of jackal precopulatory behavior (T-se-
quences), in a context where the two animals circle each other.
During equivalent interaction sequences, head movements exploit all
the available anatomical space, whereas neck movements exploit only part
of it. Horizontally, neck movements never cross planes [2] and [6] for both
partners, although such movements are anatomically feasible. Vertically,
they very seldom cross plane [0] downward. When, however, neck move-
ments do cross this plane, the movements involve a large horizontal
component, starting for example at plane [6], the furthest sideways position
observed in this context, and proceeding for at least 90° horizontally. In
comparison with head movements, neck movements in intermediate planes
are much more abundant and have larger amplitude. Movements in the
horizontal planes are comparatively scarce; i.e., the movements usually in-
volve either purely vertical components or simultaneous vertical and hori-
zontal components. As is shown in the next section, this verticality relates
to specific patterns of locomotion that are employed in the context of
mutual circling.
Fig. 8. Wide lines represent the trajectories of actual head movements. Each arrow represents
one actual movement. The coordinates of the spheres in Figs. 8 and 9 wiJI be used in the follow-
ing descriptions.
84 lIan Golani
Fig. 9. As in Fig. 8 for neck movements.
The vertical upper bound for the male's neck movements in this
context is plane [1}, whereas for the female it is plane [1+]. Examination in
more than a hundred observed sequences of the context in which the female
raises her neck to (1+] shows that a movement of the female's neck to this
position always precedes the male's licking of her vulva. Unlike other ob-
served neck movements, these movements are restricted to interaction se-
quences, and never occur when the female locomotes on her own. In those
cases in which the male proceeds to lick the female's vulva, and a move-
ment on the female's part to this position does not precede his move, the fe-
male moves away before he starts to lick. Once the female's neck moves to
[1+], she will not move away if the male proceeds to lick her vulva. A possi-
ble interpretation of this particular neck movement is given below.
V. THE CONNECTEDNESS OF THE FLOW
This section addresses the following questions: How do actual move-

ments and actual positions become organized into definite, coherent, recog-
nizable motor behaviors, such as displays? How do successive atomistic
events become welded into a continuum of precopulatory behavior?
A. Temporal Connectedness
1. The Structure oj Organismic Simultaneity
An organism may be conceived of as an "orchestra," or rather, as a

"jam session," where the interacting "players" are the various limb seg-
ments. The actual movements and actual positions may be conceived of as
the "playing" behavior of the various limb segments, or "players."
Within a specific limb segment, actual movements and actual positions
are of course described as following each other continuously, so that they
form a partitioning of the flow. 3 Since all limbs perform simultaneously in
this way, whether or not movement occurs, actual movements and actual
positions extend over each other so that they form a multidimensional con-
catenate. This property of temporal extensive connectedness (compare
Whitehead, 1929) is an almost trivial but at the same time a fundamental
property of motor behavior. In the E-W notation manuscript page, which is
a spatial representation of the temporal relations between events (Fig. 5),
behavior is represented by layers which are comprised of actual movement.s
and actual positions that extend over other actual movements and actual
positions like shingles over shingles. Just how these movements become
connected to each other, i.e., the description of the structuring of motor be-
havior, is the subject of the following sections. In Holt's language (Holt, in
Bateson, 1972, Chapter 10), the question is to what extent the system
guarantees the priority of some events vis-a-vis others, and to what extent
certain events are concurrent, in the sense that they show an independence
that develops out of complexity within a general context of detailed inter-
connection.
Simultaneity is a diffuse concept which incorporates many possible
temporal relations between events. Two events are usually considered si-
multaneous either when they start together and end together, or when one
starts first and they both end together, or when one starts earlier and ends
later, etc. The variety of these and other temporal relations between events
is usually described using the concept of simultaneity. Common language is
too poorly equipped, in its use of unitary terms, to deal with the subtleties of
temporal topological relations between events. Because of the crucial role of
language in 'ordering attention and directing action in scientific enquiry, an
expanded and more precise terminology for the various temporal relations
between events is needed if one is at all interested in temporal patterning.
Since the decision whether two events start or end together is dependent on
3 Sometimes, however, a limb may perform two simultaneous movements, such as a rotation
and a plane movement.
86 lIan Golani
the smallest time unit used during observation, the use of such a termi-
nology to describe temporal relations between events has to be accompanied
by a specified time unit on which the observed temporal relationships are to
be based. If the smallest time unit were the life span of an animal, for
example, then all of its movements would be described as starting together
and ending together. The finer the degree of resolution, the more intricate
the description of temporal relations between events becomes. The "appro-
priate" degree of resolution, or the "natural" time unit for a specified level
of description, is partly dependent on the quality of the results obtained
through the use of such a specified time unit, as well as on the intuition of
the observer, his interests, and the limitations of his recording equipment.
It would be of some help if the proposed terminology could be opera-
tional, in the sense that the terms were grouped in various ways by the
various temporal topological properties that they convey. Since it is im-
possible to convey all the topological properties of such temporal rela-
tionships within one term, only two aspects of the relationship were chosen
in the following proposed terminology; these are expressed in bisyllabic Latin
terms.
Definitions: Let p and q stand for two actual movements performed by
two limb segments.
1. If p and q are temporally contiguous, i.e., follow each other im-
mediately, the relationships will be designated by the suffix "vene."
2. If during every "instant" of the occurrence of p, q occurs, i.e., p
either starts together with or later than the start of q, and p ends
together with or earlier than the end of q, the relationship will be
designated by the suffix "dure."
3. If only during the later part of the occurrence of p, q or part of it also
occurs, i.e., p starts before q and ends after the start of q but before or
together with the end of q, the relationship will be designated by the
suffix "vade."
4. If p starts after or together with the start of q, but before the end of
q, and ends after the end of q, the relationship will be designated by
"cede."
5. If q occurs only during the middle part of p, i.e., p starts before the
start of q and ends after the end of q, the relationship will be
designated by "case."
Variants within the five groups are distinguished by the prefixes pre-
(before) and super-(after); in-(going in) and ex-(going out); pri-(prior) and
post-(later); con-(together); ent-(within); and en-(around).
Logical combinations are coined in Fig. lO.
Definition: If p relates to q in any of the above relationships, then p con-
catenates q.
:c
0
,.a
0
rIO
Pendsjust P ends after q P ends together P ends after g

;:;.
before q starts st Jrted but be- withq q ended
(contiguous) fore q ended
::
S
..,
..,"tI
P starts before ,...,0
rIO
q started rIO
----
prevene invade convade encase ,.
rIO
or
::
10
a
a
P starts together a
;.
withq pridure condure concede
-i=
10
::;.
P starts after q o·
rIO
=
started but before
it ended entdure postdure excede
P starts immediately
after q ended
(contiguous) supervene
Fig. 10. Temporal topological relations between "simultaneous" actual movements. The bars represent the temporal extension of the movements.
For each pair of bars the upper bar stands for p. the lower for q. The time dimension proceeds from left to right.
QO
-J
88 I1an Golani
Definition of transitivity: If p concatenates q and q concatenates r, then

p concatenates r.
In this way, we shall be able to form and describe multidimensional
concatenates, in which the body dimension includes a number of examined
limb segments of one or more animals, and in which the events are actual
movements and positions.
2. Concatenates, or "Chords," oj Movement
The following are one-animal concatenates of actual movements and

actual positions. Concatenates between actual positions only are not
examined in this paper.
A one-animal concatenate is a particulate process of movement, a
"chord" of behavior. Such concatenates are isolated from each other by
chords or concatenates of the actual positions of the examined limb seg-
ments. One way to visualize a concatenate of actual movements is to imagine
an animal moving in a cloud chamber and leaving behind itself traces of its
movements. Because of the simultaneous movement of the various limb seg-
ments, the result is a highly complex morphological structure, in which time
is seemingly transcended and objectified in spatial form. Figure 11 shows
such a form reproduced from the work of Bernshtein (1940), a brilliant
Russian student of movement whose work has only been starting to have its
appropriate impact on the Western neurophysiology of movement. The stick
figure describes a slab of a side view of a human locomotion concatenate.
Like a snail's shell, this three dimensional structure transcends time as a
Fig. II. A stick figure of a side view of human locomotion. Reproduced from Bern-
shtein (1940).
fossil of what was fundamentally a process of movement (see also Holt, in

Bateson, 1972, Chapter 10).
Figure 12 presents 16 concatenates that were extracted from jackal
interaction sequences. For each animal, the examined parameters are head
movement, neck movement, shift of the animal's weight as a whole, and
shift of the front of the animal's torso. Other body parameters were
eliminated for the purpose of the present analysis.
Two separate lines are allocated on the movement notation manuscript
page for shift of weight and shift of front. In the E-W movement notation,
the direction of shift of weight is usually described in relation to the
animal's environment, i.e., in "absolute" space. However, in the present
paper it is described in relation to the animal's own longitudinal axis. In the
above concatenates, shift of weight was always in the direction of the
animal's front, or "zero," so that whenever shift of weight occurred without
shift of front the animal proceeded forward, and whenever both occurred
together the animal proceeded in a circle. Shift of front describes the
amount of horizontal displacement of the longitudinal axis of the animal's
torso within the E-W sphere, which is attached in this case to the ground.
One amount of movement equals 45°. Clockwise displacement is
represented by n; anticlockwise displacement is represented by U.
B. Spatiotemporal Connectedness
The directedness oj the jlow is immanent to motor behavior at all

levels. In this section I will try to show how a few principles of directed ness
account for a large component of jackal concatenate structuring. Principles
of directedness on this level are defined and described as "routes of con-
vergence" and "fixations."
The directedness of the flow of motor behavior is already implied in the
E- W notation at the elementary level of actual movements and actual posi-
tions. The description of an actual movement using a twofold description of
a trajectory and a final abstracted position already implies that the trajec-
tory leads to a fixed final end point. Only after the position is manifested
does the movement acquire its coherence as a particulate process of be-
havior. Thus, by definition, an actual movement may be conceived of as
the simplest route of convergence of a limb segment to an abstracted final
position.
In the same manner, actual movements of serially connected limb seg-
ments may be conceived of as routes of convergence to particular positional
values in the environment or on the animal's own body. In that case, it is
the actual movements of several limb segments that converge to a positional
90 lIan Golani
Head [6'] ~ [6]j [6] ~ [&)

Neck [~] 1- llll
Weight [0) I [0]
Front &
- - I (j\
[3
[b '"
k
[0]
[8)
[9.1 7 f-- ~
[k ..
[8 1-
[01
--- [~l
J.:;]
'"
III IV
[~) ~
- [6] "- [3'] [~. ~[8i t[l;
.el-
[1)1 '"
[0)
- ['li ~- [',']
[,,'
[0]1
[8
~ 1
VI
d ~ -[.;] ;..,
~
f·[?] -7 [oj
--
Gil
[3J L?- I~jt>- I~ " ['i'] <- [S] "' ["J
~ --.... !;:;2-
--
[0] [oJ [0
0l--- M)
VII
d [.\] 1'-
[~] "
[0]
[s] ..
[I] !- [~] "-
[3]
[:] .
1- [6]1
Jqj
I
Ii;\
I
VIII
[B]1- [6 i,.J::. [3)

[1] I< [?j [g] [6]
<'-
[0] [oj
'"
me-
IX
Fig. 12. Concatenates of jackal precopulatory behavior. Extracts from the above will be used
in the following text to demonstrate various motor processes. Time unit is one sixth of a
second. -> and.-, T and L stand for positive and negative horizontal, and positive and negative
vertical plane movements, respectively. nand U stand for positive and negative rotations and
shifts of front. Numerals within the rotation and shift of front sign stand for amount of move-
fo"'" [6 -1' [8 ~~
[i] +
[0)
[bJ [~ '" (oj
[~ " I-- [g ~
(oj
[9' ,. [8
[0]
"' ...... [~
\&J
XI
d !;i] '[8] .[6] ~ -[1 » ~[o

[I] t[~ ;:... [8 f[~J f[ZJ ;Z-[? .or I'?
[0] [oj
t1;\
XII
d
~J .1- [g ..:t- [g <-- [;:].[7; ~ [~T ;J.- [;1;
1[9' .[~ <- [~ 7_ [9 .;r [SI: "'7 ['f;j <-I-- [91
(0) [oj (0]
'"' ~
XIII
[~!-J- [iii • [~ll

1'1-- r- W~~
['liJ~ [8J
ri-
[0] I
-
J
XIV
[~I +[81 j:... [8] J I I'" I [b)

[b] -[g] -t [8] !.- [?5l il'''' I
[1,]
[0) [0) I I I I
@;
i I I I 1
XV
--,-
[t] -" -~r~-~· ~
~,-
[bl ~ [~l -" I- lbi

(0)
- I- ~.
,~
XVI
ment where one amount equals 45°. m stands for "minimal" amount of movement (less than
22° 30'). [ 1 stands for a bodywise, and ( ) for an absolute coordinate system. The lower
numerals in the brackets stand for the horizontal coordinates. A plus sign on the right of a
numeral indicates an amount of 22° 30' more than the numeral stands for. Arcs represent du-
ration of movement.
92 lIan Golani
value. When such a given positional value is further maintained, due to a

series of self-compensatory antagonistic movements of the various limb seg-
ments, the route of convergence is defined as a "fixation." A fixation can be
directed at a positional value on the environment, on a partner, or on the
animal's own body. For example, as a result of raising the neck in relation
to the torso, and of lowering the head in relation to the neck for the same
amount of movement, the head may be "fixated" in the absolute coordinate
system on the same vertical positional value, i.e., it will steadily maintain its
initial angle with respect to the ground. This is a vertical fixation of the
head in the absolute frame of reference. By the same token, when an animal
changes direction, the neck first shifts to a particular horizontal value, and
then fixates on this value through two antagonistic movements of the torso
and the neck; the movement of the torso to the particular absolute hori-
zontal value to which the neck formerly moved is accompanied by a si-
multaneous bodywise antagonistic neck movement, so that the neck is
steadily maintained in the absolute horizontal particular value to which it
originally moved (see Fig. 14). This constitutes a fixation of the neck in an
absolute horizontal positional value. In other cases, particular positional
values of the head or other parts of the body are steadily maintained with
respect to the head or other parts of the body of another animal. In'these
cases, the fixation is often a result of the simultaneous coordinated move-
ments of both animals. For instance, one animal may fixate its snout on the
other animal's cheek, and compensate for movements performed by its
partner with its own movements, thereby forming a "snout to cheek joint"
with the partner. In other cases, a fixation may be performed in relation to
a point on the animal's own body; e.g., an animal keeping its own head in
contact with a particular point on its own hindquarters, and then steadily
maintaining this contact by simultaneous "twisting" of the head, the neck,
and the torso.
Thus, motor behavior can be conceived of as a directed process, and
the convergence of movement can be conceived of on the level of single
actual movements, on the level of actual movements of serially connected
limb segments that converge to particular positional values, and on the level
of fixations. At a still higher level of motor behavior, Craig (1918) described
directedness through the concepts of appetitive and consumatory behavior.
Several routes of convergence that show regularity in terms of their
form and their final positional values will be described and discussed in
following sections. In order to unfold this regularity I will have to use
several coordinate systems for the description of movement; so far, I have
described behavior only bodywise, i.e., the E- W sphere was attached to the
joint of the examined limb at the root of the movement, and its coordinates
were projected by the next serially connected-proximal limb. What counted
was the angle obtaining between the distal limb examined and its adjacent
carrying proximal limb. In what follows we will have to continually change
from one coordinate system to another, in order to be able to comprehend
movement in terms both of its mechanism, or actual genesis, and of its final
abstracted result, or consequence. This can be done by transforming the
bodywise positional values to "absolute" positional values of a sphere whose
coordinates are projected by the environment, or to "partnerwise" posi-
tional values of a sphere whose coordinates are projected by the partner,
etc.
Some regular routes of convergence occur both during locomotion of a
single individual and during interaction sequences. These are termed loco-
motory routes oj convergence. Others are exclusive for motor interaction
sequences. In the rest of this paper, I will try to demonstrate that a large
component of jackal motor display in particular, and mammalian motor
display in general, is comprised of successive or simultaneous regularly re-
curring routes of convergence on the level of both a single animal and a pair
of animals.
1. Locomotory Routes oj Convergence
Locomotory route oj convergence No.1: jixation oj the head in a given

vertical absolute positional value. When the neck is above the bodywise
horizontal, lowering or raising of the neck is condured Q :) by a move-
ment of the head in the opposite vertical direction for the same amount of
movement.
As a result of the two opposing equal vertical displacements, the angle
obtaining between the head and the neck changes, but the angle obtaining
between the head and the ground is maintained throughout movement
(Figure 13).
Transformation of the bodywise coordinates (Fig. 13, row 1) to absolute
coordinates (Fig. 13, row 3) makes it evident that the absolute vertical head
position value is maintained steady through movement. The eight examples
of pairs of movements are motor transformations that map the set of vertical
values of the head positions onto themselves. The set of concatenates
presented in Fig. 13 may be viewed as exemplifying a mapping process, or a
function that maps a set of positional values onto themselves. Essentially, it is
a set of steady-state processes achieved through continual spatiotemporal
convergence. The vertical absolute values of the head may be viewed as
fixated soIl werts, or must values.
Locomotory route oj convergence No.2: movement oj the neck to an
absolute horizontal value which specifies the direction oj shift oj weight, and
jixation oj both the neck and the head on this value. The neck moves to a
94 lIan Golani
~ll:1 IIII~I
Head
Bodywise
Neck
Absolute Head
('1 I I I I 1('+)1
III IV
~ ~ ~.;
[~ll~1 I I H~ll [6]tS1~1

f-il" 8]
[gl '" [),1
[~l;.. ~~ ~~I~ III]~~
(1)1
I I I I 1(1)I ('llj (')EE (')1 I II 11(')1
IX XI XIII XIV XVI
Fig. 13. Locomotory route of convergence No. I. Examples were extracted from the con-
catanates in Fig. 12. Each example includes a bodywise notation, an absolute notation of the
vertical positional values of the head, and a stick figure that represents vertical displacements
(lateral view) of the neck (left) and the head. Note that the absolute notation provides a
de~cription by consequence, in this case. Arrows in notation stand for plane movements, lower
numerals in brackets stand for horizontal and upper for vertical coordinates. [ 1for body wise,
( ) for absolute coordinate system. I = 45°.
specific horizontal absolute "direction" and fixates on it. The longitudinal

axis of the torso converges to the same horizontal absolute value, so that
both the neck and the torso align in the direction of shift of weight (Fig. 14).
The first movement of the neck to the specific absolute horizontal
value, in which it later fixates while shifting weight, or weight and front in
the same direction, either invades (~) (III, Fig. 14) or pridures (~)
(VI, VII, X, andXIb, Fig. 14) shift of weight, i.e., it is always "attached" to
the beginning of shift of front and weight. However, when large shifts of front
and weight are involved, several neck movements in the direction of the shift
might entdure (I I I I) it; it is as if the animal sets a horizontal value for the
shift, fixates on it, then sets a new value, then fixates on the new value, etc.
This is characteristic only of slow shifts of front.
The second bodywise neck movement to [0], which fixates the neck in
the absolute frame of reference, is, of course, in the opposite direction of
shift of front and weight, and usually ends together with the shift of front
(III, VII, IX, and Xlb in Fig. 14), Le., postdures (~), convades
(1---;:=1), or condures (~). During large shifts of front, however, it
might entdure (~) (Xla in Fig. 14).
If the longitudinal axis of the torso is already in the absolute horizontal
value which specifies the direction of shift of weight, the concatenate in-
cludes only a movement of the neck to [0], which again either prevenes, or
invades, or pridures shift of weight forward (IX, VI, and X in Fig. 14).
The set of concatenates comprising this locomotory route of con-
vergence may be described as a mapping process that maps both the torso
and the neck to the same absolute horizontal value: a set of any given
absolute horizontal neck values are mapped onto another set of given
absolute horizontal neck values, which are further mapped onto themselves.
Bodywise, a set of given horizontal neck values are mapped onto another set
of given horizontal values, that are further mapped into [0].
Locomotory route oj convergence No.3: lowering oj neck at the begin-
ning oj shift oj weight. Lowering of the neck sometimes prevenes (~),
but usually invades e-----il---\) or pridures (~) shift of weight (Figure
15).
Lowering usually converges to the horizontal, so that the angle between
Neck [0J I! ~J ~ [0][

/
J '" [,] «-
r [0]
/
Weight [0] [oj
Front \!J ~
III VI VII
[,]It [,] ;<:'- 101 [6J ,. [0] J~ [7J

:---
'" [01
f91
'"
[oj
IV ...... .
19]
IX XI • XI b
Fig. 14. Locomotory route of convergence No.2. Examples were extracted from the con-
catenates in Fig. 12. Stick figures represent successive changes in absolute shift of front and
bodywise neck positions. The lower edge of each stick figure is aligned with the onset and
termination of notated movements. Arrows on sticks indicate direction of following movement.
The lower bar stands for the torso, the upper for the neck. Shift of weight is not represented.
Numerals in the "front" space represent amounts of movement, where 1 = 45°.
96 lIan Golani
Neck
Weight ["lffiffNI I I I I I I l:m
III
VI VII VIII
['1H£]
IX XI XII
XIII XIV
['l~IIII:rlllllfl
XV XVI
Fig. 15. Locomotory route of convergence No.3. Symbols and legends as in previous
notation figures.
the neck and the torso is 1800 • When a shift of front of more than 90 0 is in-
volved, lowering of neck may go below the absolute horizontal to form an
angle of up to 225 0 with the torso. If the neck is already horizontal at the
start of shift of weight, no lowering occurs. Unlike the two previous routes,
which include the mapping of a set of values onto themselves, this route
maps a set of vertical positional values onto a set of lower vertical posi-
tional values. Since the neck is carried in this context on a torso which is
constantly horizontal, there is no differentiation between a bodywise and an
absolute description of vertical neck movements. The same applies for the
next route.
Locomotory route oj convergence No.4: Raising oj neck at the end oj
shift oj weight. Raising of the neck postdures (~), and sometimes ex-
cedes (~), shift of weight (Figure 16).
The neck is usually raised to 45 0 or less above the horizontal. Raising
it higher is exclusive for the female, and results from the simultaneous oc-
currence of route of convergence No.6 (see below).
Note that the present route is symmetrical to the previous locomotory
route of convergence: lowering of neck invades (~) or pridures
(~) shift of weight, and raising of neck postdures (~) or excedes
(t---i~) shift of weight.
Locomotory route oj convergence No.5 oj head to [0] at the beginning
oj a concatenate. If at the beginning of a concatenate the horizontal

bodywise position of the head is not [0], a movement to [0] occurs at the
start of the first neck movement. Head movement to [0] prevenes, vades,
pridures, or condures neck movement. That is, a jackal does not start mov-
ing its neck at the beginning of slow concatenates without first or at the
same time moving its head to [0]. This route may be described as a mapping
of a set of positional values onto a [0] bodywise positional value.
Since during motor interaction sequences both mates locomote, i.e.,
shift front and weight around each other, a large component of the structur-
ing of their behavior is explicable in terms of the above five locomotory
routes of convergence, whether in the presence of a partner or not. When a
jackal shifts front slowly, he usually moves his neck to a particular hori-
zontal position and then maintains the neck in this absolute position, while
shifting weight forward and aligning the torso in the same direction as the
neck (route No.2). Lowering of the neck accompanies the beginning of
shift of weight forward or forward and sideways, and raising of the lowered
neck accompanies the end of shift of weight (routes No.3 and No.4). The
above routes involve neck movements which are usually accompanied by
appropriate head movements (summarized under routes No. 1 and No.5).
These locomotory regularities are only a part of the repertory of particulate
motor processes used during motor interaction sequences. Other such
processes, not dealt with in the present paper, are the route of convergence
VII VIII
IX XI XII
XIII XIV
['ll~[] I liEf III I IfI

XV XVI
Fig. 16. Locomotory route of convergence No.4. Symbols and legends as in previous
notation figures.
98 lIan Golani
of the eyes and the head to particular positional values, and the several
routes of convergence of the neck, the torso, and the pelvis to particular
absolute vertical values.
The route of convergence of the eyes and the head to particular hori-
zontal and vertical values is described, and its underlying neurophysiology is
analyzed in some detail, in cats by Robinson (1974): "The vestibulo-ocular
reflex maintains clear vision during head movements by quickly and exactly
changing the position of the eye in the head to compensate for changes in
the position of the head" (see Robinson, 1974, Fig. 2, p. 199). It might be of
interest to neurophysiologists who study motor behavior that basic struc-
tures similar to those described for eye and head movements, can be
demonstrated in the coordination of most limb segments during motor
interactions.
The routes of convergence of neck, torso, and pelvis to particular
vertical values do not occur in the context of slow circling, and are prev-
alent in jackals in the context of mutual upright during fighting and play
behavior. The neck or the torso goes either up or down to a particular
absolute vertical value, and the other limb segments (neck, torso, and
pelvis) follow it and align in the same absolute vertical value. These routes
are mostly prevalent in the behavior of Tasmanian devils, where, by shifting
from a plantigrade to a digitigrade stance, and vice versa, the animals can
raise and lower their foreparts and hindparts while either extending or flex-
ing the forelimbs or the hindlimbs. Whereas a jackal usually aligns his neck,
torso, and pelvis in the same oblique vertical line by jumping with his fore-
legs high in the air, Tasmanian devils often obtain the same alignment by
extending their forelegs, and flexing, and assuming a plantigrade stance
with their hindlegs. Another typical route of convergence employed by
Tasmanian devils, and sometimes by jackals, during motor interaction se-
quences, is performed by lowering the neck to the ground, flexing the fore-
legs, and raising the hindquarters by extending the hindlegs, so that again
the neck, torso, and pelvis align in the same absolute vertical line. The
details of the vertical routes of convergence will be given elsewhere (Golani,
in preparation).
These basic particulate processes, which are performed on the level of
one animal, together with the routes of convergence, which are performed
on the level of an interaction (see next section), can serve as a baseline
repertory for the study of motor interaction sequences. When superimposed
on each other, they create highly patterned behaviors. Regular deviations
from these routes may be interpreted as modulations, and the specific
properties of the modulation should be looked for in the recurring specific
circumstances of the interaction. In such a framework, the fact that a
certain movement does not occur in the characteristic temporal ordering, or
that it is missing completely, acquires significance. A discussion of the way

in which various concurrent routes modify each other is beyond the scope of
the present paper.
2. Interactional Routes oj Convergence
If the movements of the male around the female are examined within
the female's frame of reference, it soon becomes evident that the male often
moves during an interaction so that his head "traces" a horizontal circle
around the female's head. Through a series of highly coordinated shifts of
weight, front, neck, and head movements, the male fixates his head at a
specific distance from the female's head, and keeps this distance steady
while continuously changing his position in absolute and relative space. The
only common denominator of positional values obtained through an
interaction of this kind is the specific distance from the female's head. The
concatenates involved in the fixation of distance belong to one out of
several possible interactional routes of convergence that can be
demonstrated most explicitly in a "partnerwise" frame of reference. In
other interactional routes of convergence, specific points on the two
partners' bodies converge to each other: The male's head and the female's
vulva converge, the female's head· and the male's inguinal region converge,
both pairmates' heads converge, one partner's shoulder and one partner's
head converge, and, of course, the genitals of the two partners converge.
Still other routes involve both convergence to a specific point on the mate's
body and to a specific distance from this point. Each such route consists of
a set of concatenates in which a single concatenate describes a specific
strategy of movement toward a common final end position. Such positions,
when attained, are often further maintained either through "freezing" or
through a steady-state process of fixation.
Interactional route oj convergence to head-to-head distance: con-
vergence to a given vertical value on the animal's own body. This route is
possibly comprised of several more specific routes of convergence.
However, at this stage of my study, I cannot tell these routes apart.
Once a given head-to-head distance is attained, it is maintained steady
through a series of self-compensatory movements of either one or both ani-
mals, until the system "shifts" into a new distance, which is again steadily
maintained. Sometimes the movements only partly compensate for the
change in distance, so that, for instance, a larger change in distance due to
shift of front and weight toward the mate's head is accompanied by a
smaller antagonistic neck movement away from the mate's head. Such a be-
havior of the system is reminiscent of what Waddington (1966) describes as
100 lIan Golani
self righting; it is as if the system "has an inbuilt tendency to stick to a

path, and it is quite difficult to divert from it by any influence-external or
internal." A change of distance due to shift of weight may be partly or com-
pletely buffered out by an antagonistic shift of front, neck, or head move-
ments, or by horizontal arching of torso, neck, etc.
In the context of "shifting" from a relatively short distance to an even
shorter distance, the approaching animal may often lower its neck and
rotate its head or neck, thereby "presenting" the ventral portion of its head
and neck to the approached animal. The occurrence of these movements in
the context of multiple self-righting movements that converge to a
particular head-to-head distance raises the question of the relevance of this
particular movement. Since the present study is mainly concerned with a
context poor in head and neck rotations, I am not yet in a position to assert
the relevance of rotations in general. A thorough study of contexts rich in
rotations, such as play and fight behaviors, is needed before their signifi-
cance can be fully assessed. Thus, the following relates to only one specific
context in which rotations occur.
In order to demonstrate that rotations of the head and the neck are in
some cases part of a route of convergence to distance, I must resort to a
description that contemplates movement in terms of its consequences on the
animal's own body. If a jackal moves in such a way as to keep certain
aspects of the environment (e.g., the female's head) fixated on its own body,
then the assessment of a particular relationship between that aspect and its
own body can be described in terms of imaginary "contact points" between
the two animals; as previously described, the E- W sphere can be
schematically superimposed on the animal's body so that every part of the
body is labeled by specific horizontal and vertical coordinate values. The
point on the body with which one animal relates to the other animal's
closest point is termed a "contact point." Once the animals change their
relative body positions, the ensuing shift of the "contact point" can be
described in terms of both horizontal and vertical displacements on the
examined animal's own body. When, for example, the male's neck is
broadside to the female's face, or parallel to the female's head facing the
same direction, a lowering of the male's neck results in a vertical upward
shift of the contact point on the male's own body. This upward vertical shift
can in principle be buffered out by an appropriate rotation of the head or
the neck of the male so as to keep the contact point steady on his own body.
In a similar way, an approach to the female's head with lowered head
and neck also results in a vertical upward shift of the contact point. Such a
shift could again, in principle, be buffered out by appropriate head and neck
rotations. Such rotations do indeed occur in the context of lowering the
head or decreasing head-to-head distance. These rotations could therefore
be interpreted as self-righting movements in a route oj convergence to a
given vertical value on the animal's own body. The performance of this
rotation of either the head or the neck, in the context of the shifting of
weight and front toward another animal's head, was formerly interpreted by
Lorenz (1949, 1953) as the exposure of the most vulnerable parts of the
body to attack by the opponent. Schenkel (1967) later pointed out that the
rotation was performed by the attacking (i.e., approaching) animal, and he
argued that it was a signal of "dominance" rather than of "submission."
Some of his illustrations demonstrate very well the contingency of lowered
neck, head, and neck rotation, and shifting of weight and front toward the
opponent (Fig. 17).
a. The Extensive Connectedness oj Several Routes oj Convergence.
Just how several routes of convergence can find expression in one and the
same unitary process is exemplified in the analysis of the following con-
catenate (Fig. 18), which is accompanied by a series of drawings of the
interaction (Fig. 19).
Readers who are not proficient in the notation and find the analysis of
the concatenate difficult to follow may skip it. I nevertheless present it here
in some detail because, except for its specific content, it can serve as a para-
digm for a zoosemiotic of motor behavior.
The analysis demonstrates how simultaneous and successive routes of
convergence can concatenate to form a slab of behavior. In a way, this is
part of the deep structure underlying the behavior. It shows how one func-
tional group (i.e., route) can extend over or be embedded within another
group, which is again embedded within another group. In Fig. 19, for
example, group 5 contains group 4, which contains group 3. This is a
hierarchy of embedding, not an ecclesiastical hierarchy in Tinbergen's
(1951) or Miller et al.'s (1960) sense (compare Nelson, 1973, p. 372). Within
Fig. 17. Illustrations from Schenkel (1967) demonstrate the occurrence of head and neck
rotations in the context of approaching and lowering of head. This contingency could be in-
terpreted as serving a homeostatic maintenance of contact between the two interacting
animals.
102 lIan Golani
001 016 024
~
""_, ;, f' ,
032 056 064
068 076 080
096 112 116

18
Figs, 18 and 19. A five-second concatenate of a precopulatory motor interaction in golden
jackals is used for a demonstration of the process of notation and further interpretation of
such a hierarchy, one and the same actual movement can be a part of
several routes, thereby serving several mutually compatible functions.
The phenomenon of one and the same behavioral event serving several
simultaneous functions is described in the psychoanalytical literature as the
property of overdetermination of symptoms (Wilden 1972, p. 34), and its
dH Ibl
00'
O~ ." .56
I~ 1-16
06' ...
~
076
Hj] ~
OBO ... 112
[0]
116
N ji;] I?- [~ll> [.l.j

,. "- ['i'I~
~
[g '" [7]
W [oj (oj {oi IPJ [01
F 01 h ~ .
00 w 1
9~li 1111111111111111
~
~lsecond~
• ii' _
, ,_• , m
iii. _i i i ...
i~ •
Wjl f ~ n
ii'
_~ i'
II
ii'
.~ ,
I
d [;]
Head
[:j
Neck
Weight
Front
(b)
9 Head
[:1
Neck
~]
Weight
Front
19
motor display. Both drawings and notation were made of the same 24 fps 16 mm film. Numerals
on both drawings and notation indicate frame number and correspond to each other. Location of
numerals in drawings is kept constant on the ground, so that displacement of animals in absolute
space can be inferred. Drawings were made of frames at the beginning or end of actual move-
ments. In Fig. 18 the male is- on the right. Figure 19a provides the actual notation of the con-
catenate. Figure 19b replicates the same concatenate with a superimposed interpretation in
terms of routes of convergence. Each route is enclosed in a box which is numerically labeled and
discussed in the text. Other symbols are as in Figs. 5 and 12.
demonstration on this level of motor behavior is of some interest. This

overdetermination of actual movements is one source of the ambiguity of
motor behavior. Another source of ambiguity, mentioned earlier, is the fact
that actual movements are performed with simultaneous regard to several
coordinate systems, i.e., bodywise, absolute, and partnerwise.
104 Han Golani
The spatiotemporal extensive connectedness of the various abstracted

directive groups of actual movements is represented by the overlapping
boxes in Fig. 19. Each box describes a functional self-compensatory group
of movements, unified by one and the same positional value to which they
converge.
Analysis oj the concatenate presented in Figs. 18 and 19.
1. Shift of weight toward the female's rump is postdured (~) by
neck movement [g]::::tln in the direction of the female's head, which at the
same time sets the absolute direction for the following shift of front and
weight (route No.2), and keeps the male's head fixated on the circumfer-
ence of an imaginary circle centered around the female's head; the shifting
of weight away from the female's head is buffered out by a neck movement
toward the female's head (Figs. 18 and 19,001-024).
2. Shift of front toward the female's head coincides with a fixation of
the neck in the absolute frame of reference (route No.2), and with raising
of neck at the end of shift of front and weight (route No.4). However, route
No. 2 is modulated to the extent that the torso and the neck do not
converge to the same absolute horizontal value. The final horizontal neck
position is the basis for the following shift of front and weight in the same
direction. (Figs. 18 and 19,016-040).
3. The concatenation of shift of front and weight, and the two hori-
zontally opposing neck movements, conform to route No.2, i.e., the neck
shifts to a given horizontal value and then fixates on it. Lowering of the
neck pridures (~) shift of weight, which conforms to route No.3. I do
not know why a raising component is missing at the end of the shift (route
No.4). This is possibly a preparation for the next supervening (~) shift
of front and weight. The concatenation of these movements shifts the male's
head to a shorter distance from the female's head (Figs. 18 and 19,040-076).
This shift is compensated in two ways:
4. First, the male rotates his head clockwise, thereby buffering out at
least part of the vertical shift of the "contact" point on his own body (Figs.
18 and 19,040-076).
5. Second, when the head-to-head distance is still decreasing, due to
the male's continuation of the shift toward the female, and at the same time
the male moves his head in the female's head direction ([~]~[m, the self-
righting response of the system shifts from the male to the female, who
moves her neck away, thus keeping the head-to-head distance constant (Figs.
18 and 19,040-076).
6. Male head movement to m prevenes (t----it--t), a male head and
neck concatenation that conforms with route No.5., i.e., head movement to
[0] condures (I II) neck movement to [ ] (Figs. 18 and 19,068-112).
7. Instead of shifting the neck toa given horizontal value and then
fixating the neck by shifting the torso to the same absolute value, the neck
indeed moves in the direction of shift of front and weight, but does not fix-
ate. Possibly, the male's neck follows the female's head, which is moving
away, thereby modulating route No.2. The raising of the neck at the end of
shift of front conforms to route No.4 (Figs. 18 and 19,076-112).
8. The male's head movement and shift of front and weight in the direc-
tion of the female's head is again buffered out by an exceding horizontal fe-
male head movement away from the male's head (Figs. 18 and 19,068-116).
9. The female's raising of her head to [1+] (Figs. 18 and 19,096-116),
which excedes the male's concatenate, bHongs to the following route of con-
vergence: ,
Route of convergence No.6 of female's neck and head to head [6+] and
neck [1+]. Raising of the neck to P+] is dured (~J====1; ~) by
raising of the head to [6+] (Fig. 20, V and Xlb).
The performance of this route is exclusive for the female during
interaction sequences. It consists of both the neck and the head ending up in
the same final specified bodywise and absolute vertical positions. Either the
head or the neck is kept out of horizontal plane [0]. Once the female attains
these neck and head positions, she will not walk away if the male proceeds
to lick her vulva. The fact that this behavior is associated with vulva licking
hints at the possible significance of this final posture: her raising of the neck
to this extraordinarily high position, and at the same time keeping either the
head or the neck out of horizontal plane [0], gives the female some visual
access to her own hindquarters, the next possible contact point with the
male.
The performance of neck and head movements to these positions
usually occurs at the end of female concatenates. Emergence from these
positions occurs at the onset of concatenates (Fig. 20, VII and Xla). As a
result, these behaviors often occur in the same temporal context as route No.
1 (fixation of the head in the vertical plane), and the fact that the head move-
ment does not fully compensate for the vertical shift of the neck makes this
Head
Neck
XI b VII XI a
Fig. 20. Locomotory route of convergence No.6 symbols and legend as in previous notation
figures. Stick figures represent successive changes in the vertical coordinates of the head and
the neck. The lower bar stands for the neck. Arrows on sticks indicate direction of following
movement. Examples were extracted from the concatenates in Fig. 12.
106 Han Golani
route of convergence often modulate route No. 1. Since, during the per-
formance of this route, the torso is always horizontal to the ground, and the
convergence is possibly to eye contact with the animal's own hindquarters,
the result is a reiterating general posture in which the female freezes prior to
possible vulva licking. Both a bodywise and an absolute description convey
the same information regarding this posture.
b. Routes oj Convergence to Horizontal Positional Values on the
Pairmates' Own Bodies. In order to illustrate these interactional routes of
convergence, I will resort to the description of relative body positions dur-
ing motor interaction sequences of Tasmanian devils (Sarcophilus harrisii)
(see also Eisenberg and Golani, 1976). These animals were originally selected
for study at the suggestion of J. Eisenberg of the Smithsonian Institution,
who at the time believed that a fine-grained comparison of the behavior of
the jackal (a Eutherian carnivore) with the behavior of the anatomically re-
mote devil (a Metatherian carnivore) might provide us with what he called
"a paseline" for quadrupedal mammalian display.
The analysis of relative body position is carried out by schematically
superimposing the grid of the E-W sphere on the animals' bodies and
describing for each animal's own body the shift of the "contact point" with
the other animal. A description of the procedure was given above.
The splicing of the flow is decided by two criteria: (1) the onset and
termination of either a continuous shift or of a constant maintenance of the
"contact point" on each animal's own body, and (2) the nature of the
contribution of the animal to the shift, either actively through the animal's
own movements, or passively through the movements of the partner. For
example, a continuous shift of the female on the male's body from his front
(U) to his hindquarters (4) may be spliced by the fact that at first the shift is
carried out by the male, who actively shifts the female to his left pelvis (gp)
in an anticlockwise shift, at which point he freezes, while the female
continuously proceeds with the shift on his body by moving to his
hindquarters (4). The continuous shift will then be spliced into two discrete
events, namely,
lol~1 FJillI±£1
which reads: shift the female for 5 time units, continuously and anti-
clockwise, from your '0' to your 6P, and then, continuously and passively
(i.e., by freezing and letting the female move), shift her anticlockwise to 4'
(see explanation for signs in the legend for Fig. 7). The continuous
maintenance of a specific contact point might also be spliced by whether it
is maintained actively or passively. Thus,
t2HIQF£I
reads: shift partner from the right flank of your head for 3 time units in an an-
ticlockwise manner to your front; then, by freezing, maintain the female on
your front for two time units, and finally, actively maintain the female on
your front. Obviously, during the last phase, the female is moving, and the
male has to actively compensate for her movements in order to keep the
contact point steady. Thus, the particularity of a shift as an elementary
atomistic unit is decided both by the continuity of the shift and by the form of
the animal's participation in it. The behavioral flow of a single animal is
therefore comprised of a continuous succession of active shifts, passive shifts,
passively maintained contact points, and actively maintained contact points,
in all of which a specific contact point is steadily maintained in spite of, or
rather as a result of, the movement of both animals. The shift of the contact
point is described simultaneously for both animals (see Fig. 5).
Figures 21, a-e were abstracted from interactions in which the male
circled the female's head, which protruded out of a box to which she had
retreated. The shift of the contact point on the male's body is therefore exclu-
sively due to his own shifts of front and weight.
The most evident characteristic of Fig. 21 is the continuous shift of the
contact point from one flank of the male to the other via the hindquarters,
and never via the front. The male relates to the female with one flank, and
then, usually in one continuous move, shifts her to his other flank. This is
accomplished in the following typical manner: the male "releases" the
contact with the female by walking away in a circle, and once the other
flank of his body is exposed to the female's head, he stops shifting weight
forward, and starts to shift weight sideways in the female's direction, so that
his pelvis is approaching the female's head at a higher speed than his fore-
legs and head. He thus approaches the female, while at the same time shift-
ing front away from her, and may ultimately touch her or bump into her
with his pelvis or his torso. Alternatively, he may freeze in front of her in a
broadside position. As a result of this strategy, the female is always shifted
around his hindquarters. While in a broadside position, the male often
makes a series of "minimal" movements: (1) minimal shifts of front and
weight which result in small fluctuations of the contact point; (2) small head
and neck movements in the direction of the female's head and away; these
movements are performed with antagonistic muscular tension and seem
very strained; and (3) rhythmic release and establishment of leg contact
with the ground. Whereas the contact point on his flanks is maintained for
relatively long periods of time, the shift to the opposite flank is swift and
short. The grouping of all these behaviors creates the impression that the
male is moving in a highly structured field, as if certain alignments of his
body in this "kinetic field" were more stable than others; and once he extri-
cates himself from a stable alignment, he is seemingly "sucked in" into the
symmetrically opposite stable alignment. The small movements which are
108 Han Golani
a c
d e
Fig. 21. "Contact" management in a male Tasmanian devil. All five diagrams were ex-
tracted from notated motor scores of precopulatory interactions with a stationary female. Only
the horizontal domain is represented. Arrows stand for shifting of female around the male's
body through his own movements. Short arrows stand for minimal (less than the degree of
resolution) shifts. Numerals on arrows indicate serial order of shifts, H,N,C, and P stand for
head, neck, torso, and pelvis, 0 for front, 4 for hindquarters, 2 for the right, and 6 for the left
side of the animal's own body. In all the sequences, "contact" is maintained from a distance.
performed while in the stable alignment create the impression of fluctua-

tions around a specific spatial sollwert, or must value.
Since the above structure is due exclusively to the male's behavior, it
might be of interest to see how this structure changes as a result of the ac-
tive participation of the female. Figure 22 describes the kinetic fields of
both animals during an interaction in which the female was partly lame and
thus somewhat slow to move. During this interaction, the female either
moved a few steps forward or backward in response to the male's approach,
or turned her back to the male and walked away. Figure 23 describes an en-
counter in which both animals were in top physical condition.
The partial freedom of movement of the female eliminates the contact
points on the pelvis and the torso of the male's kinetic field (Fig. 22). Once
the female participates actively in the structuring of this field, the
continuous shifts via the male's hindquarters converge to the male's head
Fig. 22. Kinetic fields (contact management) of male and female Tasmanian devils during a
sequence of precopulatory behavior in which the female was partly lame. Arrows that start and
end at the same topographical position on the animals' bodies indicate fixations. Uninterrupted
lines indicate an active contribution to the shift by the described animal. Interrupted lines
stand for shift due to the movement of the partner only. Lines I, 2, and 3 on the male are
drawn as interrupted lines, in spite of the male's contribution, and describe shifts on the ventral
side of the male's neck. Numerals on arrows indicate serial order of shift for each animal
separately and do not correspond to each other. Diagrams do not make a distinction between
physical contact and contact from a distance.
no lIan Golani
and neck. This is partly due to the fact that once the male shifts the female
to one of his flanks, the female actively "slides" the contact point toward
his head. Her movement postdures the male's movement, and since the
male has an active contribution throughout, this property is not represented
on the male in Fig. 22.
Once a contact point is established on the male's head or neck, it is ac-
tively maintained through a fixation (see Figs. 22 and 23). As in previously
described routes of convergence, the movement flows to a particular posi-
tional value and is then steadily maintained. In the present case, the con-
vergence is to a particular horizontal value on the male's own body.
The shifts of the contact point to the male's front (Fig. 22) occur when
the female walks away, thereby shifting the male to her hindquarters. Only
then does the male shift the contact point to his front by following her.
The most outstanding feature of the male's kinetic field in Fig. 22 is the
continuous shift of the female from his lli, anticlockwise for 540 0 , to his '0'
(kinetic line No. 12). This shift occurred when the male started to shift the
female from the right flank of his head via the hindquarters to his left
flank. When the female was behind his hindquarters, she shifted her front
away from him and walked away, so that when the male had ended one
complete shift of front of 180 0 , the female was not there, and he went on
shifting front for another complete circle of 3600 , where, on confronting her
hindquarters with his front, he followed her and mounted her from her rear.
In another encounter, unfortunately not filmed, when the female responded
in this same way and walked away in the middle of the male's circling, the
male went on rotating for five complete circles like a spinning top, while at
the same time shifting weight in her direction, until he ultimately started to
follow her without shifting front.
This phenomenon again creates the impression of a directive system
that converges to a particular value under "ordinary" circumstances. Ordi-
narily, the female would be in the "appropriate" position, and her head
would converge to the appropriate contact point on the flank of the male's
head or neck, thereby "satisfying" the circling and terminating it. Once the
female is not there, and the crucial contact to which the system converges is
eliminated, "satisfaction" or "punctuation" of the process is impossible,
and the system reiterates the process successively for four more times in the
search for the appropriate final sollwert before a new parameter is in-
troduced and the behavior shifts to a more adaptive pattern of following the
female. The lag in the response of the system to an abrupt and unexpected
change in the environment is reminiscent of the egg-retrieving behavior of
the Grey-lag goose (Lorenz and Tinbergen, 1939). There, too, the behavior
may be interpreted as a route of convergence to a particular positional
value on the surface of the egg which is being retrieved into the nest. First,
Homeostatic Motor Processes in Mammalian Interactions III
the goose establishes contact with a position on the egg's equator. The
contact is then steadily maintained through a process of fixation of the beak
to the equator, which is achieved by head and neck movements that are
compensatory and antagonistic to the movements of the egg. If, during the
retrieving-which is essentially a process of management of contact-the
contact is unexpectedly and abruptly eliminated due to the sliding of the
egg, the goose lags in responding and continues the retrieving movement.
Thus, the behavior may be interpreted as a failure in control rather than the
execution of a centrally-programmed command.
Most aspects of the male devil's kinetic field can also be demonstrated
in jackals. However, the reiteration of circling in the manner described
above is not found in jackals, and is possibly a characteristic of the rela-
tively low degree of plasticity and control in Tasmanian devils' motor
interaction sequences. When a jackal misses a bump during a similar cir-
cling process, he will immediately stop the circling and will respond in a
more adaptive way to the new situation.
The kinetic lines on the handicapped female's body converge to the
sides of her head, to her front, and to her hindquarters (Fig. 22). The occa-
sional convergence to her torso occurs when the male is further away. She
either confronts the male or goes away. The confrontation is comprised of
sliding the male's neck or the side of his head toward the side of her own
head, so that in the final configuration, which is later maintained steady
through fixation, the male's head is always in front of her snout and never
behind it. This is achieved by freezing while the male is circling, until the
flank of his torso is in front of her, and then actively shifting along his body
with her snout and stopping when his neck or the side of his head is in front
of her snout. From the male's point of view, once his flanks are actively
presented to the female, the female actively slides on his body toward his
head. In the final position, the male may continue to shift weight in her di-
rection with his pelvis, thereby pushing her, and the female continues to
shift weight forward, thereby pushing the male's neck or head. The con-
vergence to the female's hindquarters is sometimes accomplished with the
male's front when the male approaches her from the rear. More often,
however, the male establishes mouth contact with the dorsal aspect of the
female's neck, and then pivots on this contact point to a mounting position,
which may ultimately terminate in copulation. Note that no kinetic lines go
continuously from one flank of the female to the other via her hindquarters;
this property of the female's kinetic field was consistent throughout the
observations.
The vertical kinetic lines that are drawn on the male's body in Fig. 22
go from one side of his head to the other, and back via the ventral side of
his head. The male shifted the female's head from the right side of his head
112 lIan Golani
to the left side by first rotating his head and neck and then rotating his
torso, and eventually rolling on his back. In terms of the shift of the contact
point on his body, the male accomplishes the same final end position of
contact by rolling on his back as he would by shifting the female around his
hindquarters through circling. This interpretation of rolling as equivalent in
some respects to circling is supported by other observations in which the
significance of the contact point as a formative factor of behavior is
evident: during some motor sequences, a particular contact point is steadily
maintained through a series of shifts of weight and front, rolling on the
back, and other movements. The result is a highly variable motor interac-
tion sequence, in which the only invariable constant aspect is the
maintenance of steady, fixed contact points on the two animals' own bodies.
The various movements form a commutative group whereby a particular
contact point is steadily maintained throughout the sequence. Figure 24
presents such an interaction. The whole motor interaction sequence be-
comes intelligible once it is realized that the set point which is steadily
maintained throughout the interaction is the common contact point between
the two animals. This contact point is described in the next section as a
"joint" between the two animals. Similar sequences are abundant in mam-
malian courtship, agonistic, and play behavior.
Figure 23 presents the kinetic fields of both animals during an interac-
Fig. 23. Kinetic fields of same animals as before in top physical

condition. Legend for symbols as in Figs. 21, 22.
tion in which both were· in top physical condition. It differs from the pre-
viously described fields in that the kinetic lines often converge to the pos-
terior parts of the male's body. The active female approaches the male by
running toward him, and the male responds by running away, i.e., shifting
the female to his hindquarters. The "release of contact" between the two
animals, designated by the sign (=), occurs when the female's front ('0) is
directed toward the withdrawing male's pelvis (2P). More successful en-
counters would, of course, converge to the opposite configuration, in which
the male's front would converge to the stationary female's pelvis or
hindquarters, and whereby the male would eventually mount the female.
The concept of a kinetic field is in some respects analogous to the
physical concepts of gravitational and electromagnetic fields, in that it is an
abstracted concept based on the condensation of changing spatiotemporal
relationships between bodies into an instantaneous visual representation.
Like its physical analogues, the kinetic field concept replaces the
Newtonian or Kantian concept of a homogeneous given space within which
movement occurs, with a biological relativistic space, constructed by the
animals in their mutual movements. As was consistently pointed out to me
by my friend and colleague, Yehuda Fried, M.D.,4 himself a student of
Piaget, the development of such a concept, stemming from the actual
process of behavior, was indispensable for the study of animal behavior. It
was Piaget who pointed out that the child is not merely moving in an
externally given space-time system, but also constructing it by his own
movements (Piaget, 1936, 1946a,b, and Piaget et al., 1964; see also
Poincare, 1963). In the realm of the study of biological space, a longitudinal
description of the changing kinetic field may provide us with an epistemo-
logical view of the structuring, rather than with the mere structure, of indi-
vidual space. In the realm of the study of social and sexual behavior, the
forms of kinetic fields and the nature of the contribution of the various ani-
mals to these forms can serve as a basis for the abstraction of various bio-
logical variables and for their quantification. For example, Beach's sugges-
tion (personal communication) to replace the "male chauvinistic" concept
of female receptivity with concepts of "attractivity," measured through the
male's active contribution, "proceptivity," measured through the female's
active contribution, and "receptivity," measured through the female's
"passive" contribution, could be further elaborated to include all logical
combinations of the "active" and "passive" contributions of both animals,
as well as the exact nature of this contribution to the kinetic field both in
terms of final end positions and the movements that converge to them.
It has been proposed recently that the distinction between the two types
• Medical School, Department of Psychiatry and Behavioral Sciences, Tel-Aviv University.

114 lIan Golani
of space is also valid on a neurophysiological level. O'Keefe and Nadel

(1976) have suggested that mammals have two distinct functional and ana-
tomical brain structures that deal with each type of space: the Kantian,
external, homogeneous, "map" type space is represented in the hip-
pocampus and ancillary structures such as the medial septum; and the epis-
temological, "route" type space is represented throughout most of the brain
in all the perceptual and motor systems.
C. Anatomical Connectedness
The head, the neck, and the torso can be conceived of as a group of
serially connected limb segments which, through coordinated movements,
carry the head to specific positional values that are maintained by freezing
or fixation, until the performance of the next set of movements that carry
the head to a new positional value. In a very crude way, the movements of
these serially connected limbs may be compared to the movements of the
limb segments of a leg, which lead the foot to a specific positional value
that is maintained by freezing or fixation until the next set of coordinated
movements. In both cases, the flow can be conceived of as a continuous se-
quence of alternating functional groups of movements that either converge
to a new position, or maintain the new position by fixation or freezing. The
fact that the limbs of the leg converge to "real" contact points of the foot
with the ground, whereas the head, neck, and torso movements converge to
"imaginary" contact points of the head in absolute space, is crucial in a
biomechanical description, and subordinate in a purely formal description.
In terms of the above criteria, the form of the coordination of movement is
analogous: "imaginary" or "real" contact points are established, steadily
maintained, and released, successively. The movements of the head, the
neck, and the torso-like the movements of the upper leg, the lower leg, and
the foot-may be conceived of as a continuous stepwise movement from
one fixated contact point to another.
When viewed in an absolute frame of reference, the establishment of
a contact point reverts the mechanical interdependence of the group of
serially connected limbs. Consider the movements of the limbs of a leg:
before the establishment of contact, the movement of a proximal limb car-
ries along passively the more distal limbs, disregarding their own bodywise
movements. The root of the movement is in the hip joint, and once the
upper leg is moving, the lower leg and the foot are carried along. When
contact of the foot is established, the root of the movement shifts to the
"joint" between the ground and the foot. A movement of the foot around
this "joint" results in the "passive" carrying along of the lower and upper
Homeostatic Motor Processes in Mammalian Interactions us
leg. In terms of the E-W notation language, this change in the mechanical
interdependence of the limbs is expressed as a change in the hierarchy oj
heavy and light limbs. Before the establishment of contact, the upper leg is
the heaviest limb (not in terms of actual weight), the lower leg is lighter, and
the foot is the lightest. When contact is established, the foot becomes the
heaviest limb and the hierarchy is reversed. In the same way, the torso may
be regarded as the heaviest limb during release of contact of the head, and
the head may become the heaviest limb during the maintenance of contact.
The neck and the torso may then be conceived of as being carried along by
the movements of the head, which is fixated to a particular position or
"joint" in space. The head then "wags" the body, rather than vice versa.
These heavy and light limb segment hierarchies are merely mechanical
layouts for future activity. They are neither hypothetical constructs of
central hierarchies, nor do they describe the patterning of actual behavior.
They merely describe the mechanical interdependence of the limbs: a priori,
before any movement actually takes place, a heavier limb segment is in a
mechanical position such that its movements will result in a displacement of
the lighter limb segments attached to it in absolute space. The proximal~
distal heavy and light limbs hierarchy may, after the establishment of
contact, revert to or coexist with a distal-proximal hierarchy. What we
have then is a mechanical heterarchy, i.e., the coexistence of several
mutually compatible mechanical hierarchies (compare with. McCulloch's
1945 suggestion of heterarchies of nervous nets). For example, during a
fixation of an animal's head to a particular point on the mate's head, the
heaviest limbs may at the same time be the legs and the head of the display-
ing animal. The animal's movements are then seemingly constrained
between two "joints": one in the contact point between the legs and the
ground, the other in the contact point between the head and the fixated
position. All that the animal can "do," being subjected to two opposing
heavy and light limb hierarchies, is twist its body between the two fixated
contact points. When, during a motor interaction sequence, the other
animal is also fixating its head to a particular position on the first mate's
head so that its movements are also constrained between two similar
contact points, the two animals form together a system that moves in
unison. Such a system is termed a kinetic chain in the E~ W movement nota-
tion. The reader may gain a better grasp of the concept by clasping his
palms and moving both his hands simultaneously. The movement of the two
hands as one kinetic chain is constrained between the two shoulders,
whereas the movement of each hand separately is constrained between the
shoulder and the contact point with the other hand. The shoulder joints are
analogous to the contact points of the two animals with the ground, and the
contact point of the two hands is analogous to the head-to-head "contact"
116 lIan Golani
001 005 033
044 061 070
212
217 232 249

between the two animals. This contact changes in absolute space, but is
kept steady in relation to coordinate systems imposed on the animals' own
bodies. The series of drawings presented in Fig. 24 are comprised of one
such kinetic chain, during which one and the same joint is maintained by
the two animals during a multitude of movements. In a later part of the se-
quence, not represented in the drawings, the "joint" is released and the two
animals "slide" along each others' bodies, ultimately forming a new joint
which is again maintained for a while and then further released to form
another joint, etc. The formation, maintenance, and release of joints is by
definition done in a stepwise manner.
Thus, when viewed in absolute or superimposed on the animals' coor-
dinate systems, mammalian display may often take the form of a stepwise
kinetic chain. Sometimes, as in the example presented in Fig. 24, the two
animals, while connected by a particular head-to-head contact, may release
their fixed contact with the ground and move in unison as one
"superorganism," as if they were articulated by a joint somewhere between
their heads. The joint concept is further elaborated and exemplified in a
recent review of Marsupial behavior (Eisenberg and Golani, 1976) and in a
paper on the behavior of Tasmanian devils (Golani, in preparation).
This concept of the establishment and release of contact as a key for
viewing certain behaviors may also prove fruitful in the examination of such
behaviors as described by Wilson and Kleiman in South American rodents
(1974). These authors focus on jerky swift head shakes, jumps, and head
and body rotations performed during what they call play behavior. They
demonstrate that these behaviors occur predominantly as a response to new
olfactory stimuli, or at the end of play sequences. These behaviors are, of
course, always associated with an abrupt release of contact, which could
possibly be interpreted as their function. To claim that the function of these
behaviors is to cut off input, in the sense employed by Chance (1962), is at
our present state of knowledge an overstatement, since such hypotheses can
and must be experimentally verified. (See also below, for a discussion of
hypotheses relating to behavior as management of input.)
(
Fig. 24. The heads of a pair of Tasmanian devils "wag" their bodies into a multitude of pos-
tures and movements. In a context of cheek-to-cheek joint maintenance, the two animals move
in unison as one kinetic chain. The homeostatic character of this behavioral process would, of
course, be lost by describing the process as a sequence of fragmentary behaviors such as cir-
cling, mutual upright, and rolling on back. During the entire interaction, the male (in back-
ground until frame 147, and foreground from frame 212) maintains the female's head on his
right cheek, until he extricates himself by shifting the female to his pelvis (frame 249). The fe-
male keeps the male on her right cheek until she shifts him to her snout (frame 217) and to her
left cheek (frame 249). Drawings were made from a 16 fps 16 mm film. Numerals indicate
frame numbers of drawings and are attached to a constant position on the floor so that dis-
placement in absolute space can be inferred.
118 Han Golani
The particular view of motor interaction sequences as being controlled

by joints is only brought to attention in a description that considers the
contact points as given facts of motor behavior. The given positions seem-
ingly objectify themselves in movement: starting from them we proceed by
describing how the animal gets around their givenness by "twisting" into
specific forms. Once a joint has been established, the animal has numerous
options for "twisting" around this joint, as well as for "extricating" itself
from the joint's "grasp." This depends partly on the animal's anatomy. In
animals that can move their eyes, once eye contact has been established the
eyes may become the heaviest limb segment, and the rest of the animal then
moves "around" the joint between the eyes and the fixed position in the en-
vironment. After release of contact, the eyes again become the lightest limb,
and move on the head until the next joint is formed, etc. The stepwise
formation and release of joints may then be mainly carried out by the eyes.
Such is the case in animals that have the nystagmus reflex, for example. In
animals with restricted eye movements, such as marmoset monkeys or pi-
geons, it is the head that alternately becomes the heaviest and the lightest
limb segment. Since the head and the neck are the most distal limbs that
are anatomically moveable, the jerky stepwise movements from joint to
joint have to be carried either by the head, or by both the neck and the head
(for pigeons, see Friedman, in preparation). In other animals, such as small
rodents or hopping birds, the stepwise jerky movements which carry along
the head and the eyes to new contact points with the environment are
performed by the whole body through fast shifts of front. Notwithstanding
the ultimate causes for these jerky movements (e.g., antipredatory behavior,
in the case of rodents), in all of the above examples we have been dealing
with various strategies of movements of serially connected limbs with
regard to the heavy and light limb segments hierarchy. Both during joint
maintenance and during movements to new positions, the animal can in
principle superimpose on the mechanical hierarchy or heterarchy of heavy
and light limbs a large number of spatiotemporal strategies. Some strategies
are eliminated by the anatomy, but what are the limiting constraints on
other strategies? The study of the topological spatiotemporal ordering of
actual movements on top of mechanical hierarchies may prove to be quite
fruitful in sQrting out various inter- and intraspecific movement strategies in
both animals and humans. (Berger,5 personal communication).
The concept of the kinetic chain, the concept of twisting movements
within joints, and the concept of extricating movements that break joints
and carry the serially connected limb segments into new joints, all conceive
behavior from the point of view of given contact points. The other option is,
5 Dr. Meir Berger, M.D., Mental health clinic, Ramat Chen, Israel.
of course, to consider the actual movements as the given facts of behavior.

Their significance or coherence is then disclosed by the final positions to
which they converge, hence the description of motor behavior in terms of
routes of convergence. Both approaches contemplate the connectedness of
motor behavior unidirectionally; both start (or end) with some given, stub-
born facts. Only by vacillating from one form of description to the other
will we be able to gain some grasp of the cyclic organization of the fixed
and the flowing components of motor behavior.
VI. FURTHER HYPOTHESES AND FUTURE FIELDS OF STUDY
In a fluent, extremely variable universe of movements always in the

process of becoming, always on the verge of seeming chaos, the final fixed
positions to which movements converge can serve as immutable landmarks
in an everchanging scenery. It is no wonder that these recurring positions,
which are maintained for longer spans of time and therefore easy to ob-
serve, were historically used in the science of animal behavior to facilitate
the reduction of the flux of behavior to impoverished static simplicity. For
many years, students of animal behavior were content with descriptions that
viewed behavior as a sequence or a mosaic of static, fragmentary units (for
review, see Slater, 1973), thereby reducing the problem of form to the sim-
plistic problem of combinatorics, stochastics, etc. Although globographic
notation techniques have been available for quite a long time (e.g., Albert,
1876; Strasser and Gassman, 1893), and although the E-W movement nota-
tion was fully developed some seventeen years ago (Eshkol and Wachmann,
1958), these techniques are seldom used by students of behavior (see,
however, Prost, 1967, 1974 a,b, and in press).
In the area of comparative psychology, students of behavior have in-
terpreted some final consequences as goals (for a review and criticism, see
Hinde and Stevenson, 1970), thereby immediately subordinating the study
of the behavioral structures that "lead" toward a "goal," and shifting their
attention to the actual or hypothetical physiological mechanisms involved
(e.g., Miller et al., 1960). In the domain of actual or hypothetical
physiology, they had access to languages that disclosed the properties of be-
havior as a process, and it was these properties which they became aware of
and which they were eager to point out. Notwithstanding the tremendous
insights into behavior as a process which they managed to express, the study
of the orderliness of the rich, transient, highly variable movements that oc-
cur between final end positions was neglected. These movements can cer-
tainly not be disposed of with the excuse that they are enormously
120 lIan Golani
heterogeneous. They, too, must be subjected to at least some limitations; as

soon as limitations exist, there must be decision between alternatives, which
implies some kind of definite ordering (Emmet, 1966, p. 142). Bohm (1969)
has convincingly shown that the more "chaotic" the structure, the more
intensive is its order. Randomness may seem an apparent disorder, while
actually indicating the existence of a large number of ordering principles.
Apparent conflict can be removed by including more relevant conditions in
the definition of the system (Holt, in Bateson, 1972, p. 190).
The multidimensionality of display, and the several ordering principles
(both organismic and environmental) which operate at the same time on the
spatial, temporal, and body dimensions all enrich the behavior and make
demonstration of its orderliness difficult. It is much easier to show the or-
derliness in its fixed stable aspects than to sort out the orderliness during
the transient flow from one fixed position to another. The availability of a
conceptual apparatus that can deal with the various dimensions of move-
ment, with the help of concepts which are themselves unitary abstractions
of quite complex relationships, makes the analysis of this extraordinarily
rich domain feasible.
In the present paper, I have been mainly preoccupied with the task of
demonstrating the directedness of display. As long as the demonstration of
directedness does not become an end in itself, it certainly facilitates further
analysis by providing points of reference for the examination of the flux. In
order to do so, I have had to shift attention to the consequences of behavior,
and these consequences are mainly disclosed in a description in the absolute
frame of reference and in the description of the kinetic field. These coor-
dinate systems yield a relatively simple description because they describe
atomistic, particulate processes which show a greater degree of invariance
than their individual components. In order to demonstrate the actual
mechanisms of the flow and the ways they are coupled to produce various
particulate processes within the same route of convergence, one has to
mainly concentrate on a bodywise description. This was partly done for the
first five routes of convergence and for the sixth route. It was not done for
the far more complicated routes of convergence to positional values on the
animals' bodies. Nor was it done in an orderly manner for the modulations
of the various routes, their mutual modification by each other, and their or-
derliness of concatenation.
The examination of instances in which routes of convergence are
violated, possibly due to higher organismic and environmental ordering
principles, has to be preceded by a description of the degree of specificity
and plasticity within unmodulated routes of convergence. This work, which
still has to be carried out, is connected with the concept of commutable
groups of movements. If a particular final position is looked upon as the
end result of a set of concatenates, the problem becomes which movements

are interchangeable with each other to the extent that they end up in the
same final position. This problem was discussed by Lashley (1930, p. 194)
as the problem of motor equivalence. His assertion that "the problem of
equivalence of motor responses has been less studied than that of equiva-
lence of stimuli, but the phenomenon seems to be equally common," is as
pertinent today as it was 45 years ago. A rigorous definition of commu-
tability is an essential part of a description of the degree of plasticity of a
system.
By now, some readers might have suspected that I am heading toward
an interpretation of display in terms of management of input (see Powers,
1973; Chance, 1962; Gibson, 1966). A fixation, according to this line of
thought, might indicate that an animal is moving in such a way as to keep
certain aspects of the input steady. This is indeed possible, but since I had
no access to physiological data in the present study, I am not in a position
to verify or refute such hypotheses. Thus, a joint between two animals may
in principle result from the specific visual, tactile, or olfactory requirements
of one or both animals, but it may also be a compromise between several
such input requirements. For example, when a male jackal or a male
Tasmanian devil approaches his respective female, he often shifts his weight
forward toward the female, and at a certain point starts shifting front so
that the flank of his pelvis approaches the female's head, whereas the head
is kept steady and his whole body is pivoting around an imaginary joint
between his head and the female's head. These movements may ultimately
lead to bumping with the pelvis into the other animal while the approached
animal darts with an open or closed mouth toward the approaching
animal's head or neck; it might alternatively lead to freezing in a broadside
position in front of the other animal's head. Bumping with hindquarters is
abundant in quadruped mammals which are as remote from each other as
genets (Genetta tigrina) and civets (Civettictis civetta) (Wemmer, 1972),
meerkats (Suricata suricatta) (Ewer, 1963), wombats (Vombatus ursinus)
(Wunsch mann, 1970), polar bears (Thalarctos maritimus) and striped
hyenas (Hyaena hyaena), (Wemmer, personal communication), and many
others. Freezing in a broadside position is abundant in some of the above
species, as well as in cats, who do not practice bumping during agonistic be-
havior (Leyhausen, personal communication) and may at the same time use
their forepaws as a weapon. When considered logically, these behaviors are
the only possible way for a quadruped mammal to (1) continuously ap-
proach another animal, while (2) keeping its head as far as possible from
the other animal's teeth, and at the same time (3) have some visual access to
the other animal's head. In cats, where forepaws may be used as a weapon,
bumping is not used, and the approaching cat stops approaching at a "fore-
122 Han Golani
legs distance." Ending up in these particular relative positions could then be

interpreted as a compromise between several input requirements. However,
the difference is small between such an assumption and the assumption that
a mammal is nothing but a highly sophisticated moth, in Loeb's sense
(1918). To assume that the apparent fixity of a display is the mere result of
an interaction between a given anatomy and reiterating circumstances in the
animal's environment is tantamount to assuming that ontogeny and
phylogeny play no role in the formation of this behavior, unless through the
animal's anatomy.
An alternative hypothesis might be that the contingency of shifting
weight and front in the above manner becomes fixed during ontogeny, still
due to reiterating circumstances, and is performed henceforth by the animal
as a unitary piece of behavior that is only partly dependent on circum-
stances. [The possible formation of behavioral contingencies during ontogeny
was suggested by Klopfer (personal communication).] The question then be-
comes not "what is this behavior?" (in an Aristotelian classificatory sense),
but rather "from what moment is this behavior?" (in a Hegelian sense of
process description) (Fried, personal communication). Hypotheses in this
form may open up the domain of display for studies in the spirit of Kuo
(1967) and Schneirla (1972), which have already been successfully applied in
other ethological contexts (e.g., Gottlieb and Kuo, 1965; Bateson, 1966;
Klopfer and Hailman, 1964; Hailman, 1967). As pointed out by Simpson
(1973), an ontogenetic approach to display is long overdue.
Still another possible set of hypotheses may assume that some display
forms are highly specific and resistant to ontogenetic modification. Some
evidence along this line is presented by Dieterlen's (1959) isolating experi-
ments with golden hamsters. The fact that Kaspar Hauser hamsters display
fully on their first encounter with other conspecifics rules out the relevance
of particular experiences for the development of this behavior in hamsters.
However, the study does not take a stand on whether the display is centrally
programmed as a gestalt and triggered by the presence of another animal as
one unitary piece of behavior, or whether it is a contingency of responses of
a given anatomy or reflexology to the particular immediate circumstances.
If indeed some parts of a display are integrated so that the contingencies of
actual movements are centrally rather than peripherally programmed, and
so that this programming is unaffected by ontogeny, it might be of interest
to look for possible physiological explanations that would account for the
fact that a large component of mammalian display, when examined within
the limits of simultaneous occurrence, appears to be composed of successive
overlapping routes of convergence. (For suggestions of central program-
ming, see Kennedy, 1976.)
To neurophysiologists, the more accurate description obtained through
the use of the E- W notation may provide a better formulation of the de-
mand made upon the central nervous system on the level of movements of
whole limb segments. Once the demand has been rigorously formulated,
there are only a definite number of ways by which the organism can cope
with this demand. Unless such a demand is first described, the
neurophysiologist is at loss to explain the mechanism he is dealing with:
In so far as the neurophysiologist is concerned to understand how the brain works, he
must equip himself with a nonphysiological account of the tasks which the brain and its
peripheral organs are able to perform. Only then can he form mature hypotheses as to
how these tasks are carried out by the available "hardware." (Granit, citing Longuet-
Higgins, 1973)
Thus, the same movement during a particular motor display at one stage in
ontogeny may depend on continuous sensory feedback (the so-called
"ramp"), and at another may be the result of a ballistic preprogrammed
command. Kornhuber (1974) suggests that ramps are generated by the basal
ganglia, and that ballistic preprogrammed movements are generated by the
cerebellum. Similarly, a fixation or a reversal of the heavy and light limbs
hierarchy may be the result of environmental control exerted within a context
of a voluntary movement, of a reflex ring which involves the maintenance of
a perceptual constancy (Bernshtein 1957), of a reflex arch, or even of a
closed loop within the organism that does not involve the environment at
all. An example of such a closed loop is described by Robinson (1974) in
cats, who can shift their eyes to a particular position in complete dark-
ness and then maintain a fixation through compensatory eye and head
movements.
Another concept that might be of interest to neurophysiologists is the
concept of routes of convergence, which involves localization within a
context of coordination and integration. One way to look at the interaction
between the levels of motor behavior, perception, and neurophysiology is to
look for perceptual and neurophysiological routes of convergence that cor-
respond to behavioral ones. As elaborated very eloquently by Bernshtein
(1957), this will undoubtedly open a whole new approach to the study of
perception in the context of regulation of motor behavior, and, vice versa,
to the management of perceptual constancies through behavioral con-
stancies (Powers, 1973). When Bernshtein (1935) writes about "the
structural neurophysiology of movement," and Luria (1966) discusses cere-
bral organization, one of the key concepts they use is that of dynamic local-
ization, which should replace the concept of anatomical localization and
also should convey the correspondence between neurophysiological and be-
havioral activities. Luria (1966, p. 28) cites Ukhtomskii in describing dy-
namic localization as a "constellation of harmoniously working ganglionic
areas," which coordinate in time, speed, and rhythm of action to create a
124 lIan Golani
functionally unified "center." This organization, which they conceive of as

dynamic localization within a context of coordination and integration, is
what I try to convey in the concept of a route of convergence. As I pointed
out earlier, the concepts of localization, coordination, and integration are
abstracted distinctions which bring into attention various aspects of the
same reality of movement: they all stem from the twofold description of a
movement in terms of both a trajectory and a final abstracted position (see
above). Once the trajectory is taken as given and the final position is
described as its consequence, localization is conceived of as being the conse-
quence of coordination and integration. On the other hand, if the final posi-
tion is taken for granted, localization becomes a stubborn, given "fact,"
and coordination and integration are then conceived of as its conse-
quences-the way by which the organism arrives at certain localizations,
"twists" around their givenness, and "extricates" itself from them. Thus,
each of the three concepts is implicit in the others, and therefore does not
stand only for itself but also for the other two from which it was distin-
guished.
One way to examine routes of convergence is through the study of the
biological shape they have generated in absolute space. As demonstrated
earlier (Fig. 11 and its accompanying text), by transcending time, a con-
catenate may be conceived of as a crystalized spatial form in absolute
space. The study of motor behavior on the level of limb segments may then
be conceived of as the comparative study of such spatial structures. Deriva-
tion of the properties of behavioral processes from the structure of bio-
logical shape is not in any way limited to motor behavior. Blum, in an at-
tempt to create a new model for global brain function in terms of
spatiotemporal convergence of signals (1967), ended up in developing a new
geometry for biological shape (1973); and Nelson (1973), in a most original
paper that is possibly ahead of its time, demonstrates a striking structural
similarity between things as remote as bird song, plant phyllotaxy, and
music.
The present paper indicates that mammalian motor display should not
be regarded initially as an exact somatic projection of central input. Once
we do not use as a first approximation the hypothesis that the moving pe-
riphery is an exact reproduction of central programs of whatever nature, we
are faced with the task of looking for special peripheral physiological and
environmental structures that can account for the organization and coordi-
nation of motor behavior. (For an excellent exposition of this thesis, see
Bernshtein, 1940, p. 104.) The description of the changing topography of
the interacting partner can provide the environmental dynamic structuring
factors. After all, motor display often occurs in the highly dynamic environ-
ment of a partner, and every student of behavior will agree that the environ-
ment often exerts some control on behavior. The other component of the
peripheral physiological organization of motor b:!havior on the level of
whole limb segments has still to be accounted for by neurophysiologists.
One step toward a definition of such peripheral coordinating systems is
Flynn's (1967) discovery of "state-dependent reflexes." Flynn has shown
that electrical stimulation of the hypothalamus, whose activity can be
modulated by electrical stimulation of other parts of the brain, can modu-
late the proneness of the cat to perform certain reflexes, such as paw strik-
ing, head orienting that brings the object to the lips, mouth opening, strong
bite and mouth closure. These reflexes cannot be triggered by the appro-
priate visual and tactile stimuli unless the appropriate site in the brain is si-
multaneously stimulated. Flynn suggests that the electrical stimulation
modulates both the motor disposition of certain reflexes and the sensory re-
ceptive fields through which these motor acts can be triggered. Thus, the
central state may create a disposition for action, whereas the action itself is
carried out only when triggered by stimuli that meet the requirements es-
tablished by the modulated sensory filters. The possibility that a part of
mammalian motor display is patterned by similar interactions between
central states and peripheral patterning mechanisms should not be over-
looked (however, see also Kennedy, 1976).
Whatever the ultimate and proximate correlates of display are, it is of
interest to point out the directedness of its subsets. By trying to stay on a
descriptive level, and by limiting myself to behavioral phenomena that oc-
cur within the limits of organismic simultaneity, I point out what certain
structures of motor behavior are. This is indicated without implying
particular ontogenetic, phylogenetic, or physiological mechanisms.
Whereas most ethological studies operate on a much broader time
scale of ontogeny or phylogeny, the present paper deals with behavioral
mechanisms of actualgenese, in the sense applied to this term by the
Gestaltists, within the limits of what Whitehead described as "transitory im-
mediacy," and on the level of actual movements, positions, and con-
catenates.
Part of the present section deals with the limitations, or rather, qualifi-
cations of the present study, so it is necessary to point out three more such
qualifications. First, I did not attempt to describe routes of convergence
that occur during fast movements. Behaviors such as mutual upright, fast
circling, darting toward and away from an opponent, and sideways jumps,
abound in typical routes of convergence and kinetic fields not yet described.
Their systematic discription is essential for the understanding of the actual
genesis of agonistic and play sequences.
126 lIan Golani
Another group of behaviors not dealt with in the present paper are
facial expressions, ear movements, tail movements, bristling, and vocaliza-
tions. These behaviors could be examined in relation to the animals' move-
ments in various kinetic fields. There is no doubt in may mind that at least
some of these behaviors "switch on and off" in a regular manner during
movement from one part of a kinetic field to another. A change in the ki-
netic trajectory is often "cross-referenced" by a change in some of the
above behaviors. ("Cross-referencing" is a term suggested by Birdwhistell,
1970, to indicate that the occurrence of one behavioral event may qualify
the significance of another occurrin~ at the same time.) For instance, when
two jackal pairmates face each other and then proceed slowly in an anti-
parallel trajectory in the direction of each other's hindquarters, the male
"switches on," almost always growling, on crossing the female's shoulder
line. When the two partners align in parallel and the male then proceeds
toward and around the female's face (Fig. 18), gaping and growling are
again "switched on" when the male crosses the female's shoulder line. A
method for the description of facial expressions with the E-W notation is
presently being devised by Eshkol (personal communication).
The last qualification has to do with the question of serial order.
Sequential order is accounted for in the present study only as long as the
movements concerned belong to a particular route of convergence. If, for
example, a route extends over one second of behavior, then the serial order
within this second is accounted for. If a route extends over one whole
minute due to the formation of a joint between two animals, then serial
order is accounted for during the whole minute. Yet, even within this
minute, the sequence of events is predicted only to some extent, since within
the same process of joint maintenance the animals have quite a few, though
a limited number of, equivalent "strategic options." Thus, the present paper
deals only with the demarcation of particulate processes, and to some extent
with the indication as to what options are available for the animal in terms
of serial order: the joints are demarcated particulate processes, and th~ ki-
netic fields indicate what options are available for the animals in terms of
the sequencing of joints. In a way, what we have is a series of demarcated
"fields," in the context of a broader kinetic field. The fact that, at a given
moment, the animal chooses only one specific strategy out of several op-
tions is a "merely given," a stubborn spontaneous choice. The points in
time in which such spontaneous decisions are made are described by R. and
M. Dawkins (1973) as "decision points," or moments of high information
content: once a decision is made, the rest of the behavior until the next deci-
sion point is relatively predictable. [See also Fentress (1972, p. 90) for
nonstationarity, and Nelson (1973, p. 300) for nonlinearity in behavioral
processes.]
VII. TOGETHERNESS WITHIN SHARED ROUTES OF

CONVERGENCE
The convergence to a particular relationship between two communicat-

ing organisms is at least partly comprised of a moment-to-moment partici-
pation within shared common routes of convergence. Individual motor
routes of convergence are embedded within a broader context of shared mo-
tor routes of convergence, themselves embedded within a still broader
context of higher-level behavioral convergences, which ultimately converge
to particular relationships and social structures.
Communication as a sharing of a common organization is not limited
to animal interactions. It has long been maintained by Birdwhistell, and
more recently by Brazelton, Chapple, Condon, and others, that participa-
tion within shared organizational forms constitutes an essential component
of human communication. Birdwhistell (1970) maintains that the structur-
ing of human infant body motion communication is achieved through a
process of entrainment of the infant to given cultural body motion struc-
tures. Chapple (1970) can judge the nature of the interaction by the tem-
poral form of the entrainment. Condon and Ogston (1967), and Condon and
Sander (1974), limit themselves to synchronization in the strict sense, i.e.,
condurance o====n, and demonstrate it between verbal behavior and body
motion behavior in interacting adults, as well as in one day old neonates
who move in precise and sustained segments of movement that are
synchronous with the articulated structure of adult speech. Brazelton et al.
(1973) describe human mother-infant interactions in terms of the nature of
the synchronization of cycles of contact maintenance between the two
interactants. For each individual, each such cycle may be described as initia-
tion of contact, steady maintenance of contact, and extrication from
contact and withdrawal. (Actually the cycle is further divided into seven
phases for the infant, in terms of the intensity and quality of the contact,
and into five strategies for management of contact for the mother.) Each
phase is cross-referenced to specific behaviors such as smiling, laughing, vo-
calizing, reaching, etc. Sensitive mothers establish contact with their babies
right after the baby's initiation of contact, and break contact right before
the baby breaks its contact; i.e., the mother's contact behavior entdures
(~) the baby's contact behavior. These mothers increase the contact
behavior of the baby in time. Insensitive mothers bombard their babies with
attempts at contact initiation, i.e., encase (~) or concede (~) the
baby's contact behavior thereby causing him to withdraw.
Finally, Rosenberg (1971), and Rosenberg and Medini (1973), describe
psychotherapeutic encounters as the process of steady maintenance of a
128 lIan Golani
precise, precariously balanced surface contact, in a context of sharing and

converging to common values. They suggest that the moment-to-moment
interaction, which they describe as "moving on a razor's edge" and as the
"grasp of the quick flow of changing images," is more relevant to the
interactants than "stereotypes, rigidity of concepts, old values, and even
beliefs." These insights into the nature of human communication lend some
perspective to the description of mammalian display as a togetherness in the
sharing of common routes of convergence within the limits of transitory im-
mediacy.
VIII. SUMMARY
To summarize, this paper suggests a possible integrative approach,

rather than a theory, through which motor behavior may be studied. The
emphasis in this type of analysis is on the following:
1. The actual genesis of motor behavior is viewed as a process of con-
catanation of actual movements and actual positions, performed by the
various limb segments of one or more animals.
2. The directedness of the flow is described in terms of (a) anatomy
(heavy and light limb hierarchies), (b) space (trajectories and final posi-
tions), and (c) topological temporal properties (concatenations). The focus
is on the transitory forms of routes rather than on their fixed final positions.
3. Actual movements and actual positions are extensively connected.
Extensive connection can be described by an appropriate conceptual ap-
paratus in order to understand the structure of organismic simultaneity.
4. Four coordinate systems are used for a description. Vacillation
between these coordinate systems is essential for the description of both
mechanism (actual genesis) and function (significance).
5. Connectedness and the cyclic organization of the fixed and fluent
aspects of motor behavior can only be grasped by vacillating between two
approaches: (a) proceeding from the assumption that the movements are
given, and that the final positions are then actively generated by the actual
movements and actual position, and (b) proceeding from the assumption
that the final positions are given, and that the movements are then ways by
which the animals get around these given positions.
6. The use of four coordinate systems, and the option to choose either
the positions or the movements as given, help in making the role of the ob-
server in the description as explicit as possible.
7. By rigorously defining the limits of organismic simultaneity, time is
transcended in the formation of particulate compound body-spatiotemporal
structures (concatenates). The study of movement then becomes a special

case of the study of biological shape.
8. Several hypotheses that relate to a much wider temporal domain of
ontogeny and phylogeny are suggested, but not solved. The demarcation of
transitory immediate processes is accompanied by a suspension of judgment
as to their physiological mechanism and actual formation during the life
history of the animal or the species. Hypotheses are characteristically sug-
gested at the end rather than at the beginning of the study.
9. By splicing behavior to minute events, dimensional complexity is
reduced rather than increased (see also Nelson, 1973). Several directive
principles account for most of the observed richness, not only in the ob-
served animals, but in other mammals as well. On the one hand, the focus is
again shifted, as in the early days of ethology, to the interface between
anatomy and behavior; on the other hand, the detailed description of limb
segment movements and positions can serve as a more appropriate specifi-
cation of the demand made upon the muscular and neurophysiological
levels.
The similarity between the behavior of the eutherian golden jackal and
the remotely related metatherian Tasmanian devil, the prevalence of similar
behaviors in other quadrupedal mammals, and the directive properties of
the display, all suggest the possibility of obtaining a baseline for quadru-
pedal mammalian display.
IX. ACKNOWLEDGMENTS
This paper is a result of years of common work with my close friends

and teachers. Meir Berger 6 urged me for years to study behavior on the
fine-grained level on which I have been studying it only recently. We have
discovered and marvelled together at the potentialities of the E-W notation
and its appropriate use, and many of the ideas presented in this paper are
his. Yehuda Fried's6 methodological approach and ideas, which I have
learned to understand through years of friendship, are enmeshed throughout
this paper, and it is impossible to list them here. Eddie Rosenberg6 has
taught me to listen to what Whitehead describes as "the rush of transitory
immediacy." Amos Hetz, my teacher of the movement notation, and Noa
Eshkol, on whose work and insights this study is based, were and are for me
endless sources of knowledge and understanding of movement. H.
Mendelssohn of the Department of Zoology, Tel-Aviv University, and J.
• Medical School, Tel-Aviv University and Mental Health Clinic, Ramat Chen, Israel.
130 lIan Golani
Eisenberg of the Smithsonian Institution, Washington D.C., who have

taught me what zoology is all about, serve as my examples and are the best
zoologists I have met. I also want to thank Ann Weisler, who was a full
participant in the phase which analyzes the movement notation material
and in the early stages of developing the concepts of concatenates and
routes of convergence. My discussions with Wolfgang Schleidt and Leon
Davies of the Department of Zoology, University of Maryland, were very
valuable. Graham Goddard, Robert Adamec, and John Fentress of the De-
partment of Psychology at Dalhousie University, Halifax, N.S., introduced
me to the physiology of movement. Maya Stavi was my assistant
throughout the present study, Gershon Harel did the computer program-
ming, and Shmuel Zeidel notated a massive amount of jackal interaction se-
quences. Leon Davies planned and made the drawings and made many
valuable remarks concerning the text. Thanks are also due to Devra
Kleiman and John Eisenberg for critical reading of the manuscript. Uri
Marder, the director of the Tel-Aviv University research zoo, the
zookeepers of the Tel-Aviv University research zoo, and the zookeepers and
staff of the research department of the U.S. National Zoo in Washington,
D.C., kept the animals and helped in every possible way. Tasmanian devil
films were taken by J. F. Eisenberg and Larry R. Collins.
Special thanks are due to the Smithsonian Institution and the Depart-
ment of Psychology, Dalhousie University, under whose auspices this paper
was written. The phase of data collection and analysis was supported by a
Ford Foundation grant B-9.
X. REFERENCES
Albert, E. (1876). Zur Mechanik des Huftgelenkes, Med. Jahrb., Wien, pp. 107-129.
Altman, S. A. (1962). A field study of the soci!'lbiology of rhesus monkeys, Macaca mulatta.
Ann. N. Y. A cad. Sci. 102:338-435.
Altman, S. A. (1965). Sociobiology of rhesus monkeys. II. Stochastics of social communication.
J. Theoret. Bioi 8:490-522.
Baerends, G. P., Brouwer, R., and Waterbolk, H. T. J. (1955) Ethological studies on Lebistes
reticulatus (Peters): I. An analysis of the male courtship pattern. Behaviour 8:249-334.
Barlow, G. W. (1968). Ethological units of behavior. In Ingle, D. (ed.), Central Nervous System
and Fish Behavior, Univ. of Chicago Press, Chicago.
Bateson, M. C. (1972). Our Own Metaphor, Alfred A. Knopf, New York.
Bateson, P. P. G. (1966). The characteristics and context of imprinting. Bioi. Rev. 41:177-220.
Bernstein (Bernshtein), N. (1935). The problem of the interrelation of co-ordination and integra-
tion. Published in English in: Bernstein, N. (1967). The Co-ordination and Regulation of
Movements, Pergamon Press, New York.
Bernshtein, N. (1940). Biodynamics of Locomotion. Published in English in: Bernstein, N.

(1967). The Co-ordination and Regulation of Movements. Pergamon Press.
Bernshtein, N. (1957). Some emergent problems of the regulation of motor acts. Published in
English in: Bernstein, N. (1967). The Co-ordination and Regulation of Movements.
Pergamon Press.
Birdwhistell, R. L. (1970). Kinesics and Context; Essays on Body Motion Communication.
University of Pennsylvania Press, Philadelphia.
Blum, H. (1967). A new model of global brain function. Perspectives in biology and medicine.
Vol. 10, No.3, pp. 381-408.
Blum, H. (1973). Biological shape and visual science (part I). J. Theor. BioI. 38:205-287.
Bohm, D. (1969). Some remarks on the notion of order. In Waddington, C. H. (ed.), Towards
a theoretical biology. 2. Sketches: An IUBS Symposium. Aldine Publishing Company,
Chicago.
Bohm, D. (1971). An inquiry into the function of language and thought. A. Fragmentation and
wholeness. B. The rheomode: a new mode of language. Unpublished manuscripts obtained
from the author, Birkbeck College, University of London.
Brazelton, T. B., Koslowski, B., and Main, M. (1973). The origins of reciprocity. The early
mother-infant interaction. In Lewis, M., and Rosenblum, L. (eds.), Origins of Behavior.
Vol. I., Wiley, New York.
Chance, M. R. A. (1962). An interpretation of some agonistic postures, the role of "cut off"
acts and postures. Symp. Zool. Soc. Lond. 8:71-89.
Chapple, E. D. (1970). Culture and Biological Man. Holt, Rinehart and Winston, New York.
Condon, W. S., and Ogston, W. D. (1967). A segmentation of behavior. Journal of Psychiatric
Research 5:221-235.
Condon, W. S., and Sander, L. W. (1974). Neonate movement is synchronized with adult
speech: interactional participation and language acquisition. Science 183 (4120):99-101.
Craig, W. (1918). Appetites and aversions as constituents of instincts. Biological Bulletin
34:91-107.
Dawkins, R., and Dawkins, M. (1973). Decisions and the uncertainty of behavior. Behaviour
XLV 1-2:83-103.
Dempster, W. T. (1955). The anthropemetry of body action. Ann. N.y. A cad. Sci. 63:559-
585.
Dieterlen, F. (1959). Das Verhalten des Syrischen Goldhamsters (Mesocricetus auratus Wat-
herhouse): Untersuchungen zur Frage Seiner Entwicklung und seiner Angeborener Anteil
durch Geruschsisolierte Aufzuchten. Z. Tierpsychol. 16:47-103.
Eisenberg, 1. F., and Golani, I. (1976). Communication in Metatheria. In: Seboek, T. A. (ed.),
How Animals Communicate. Indiana University Press, Bloomington, Indiana.
Emmet, D. (1966). Whitehead's Philosophy of Organism. McMillan St. Martin's Press, New
York.
Eshkol, N., and Wachmann, A. (1958). Movement Notation. Weidenfeld and Nicholson. Move-
ment Notation Society.*
Eshkol, N. et al. (1968). Classical balIet. Movement Notation Society.*
Eshkol, N. et al. (1969). Physical training. Movement Notation Society.*
* The Eshkol-Wachmann movement notation publications can be obtained in the U.S.A. and
Canada from Dr. Anneliz Hoymans, Dept. of Physical Education, University of Illinois at
Urbana-Champaign, Illinois. In other countries publications can be obtained from the Move-
ment Notation Society, 75 Arlozorov Str., Holon 58327, Israel.
132 lIan Golani
Eshkol, N. et al. (l970a). Notation of movement. BCL report No. 10.0, Department of
Electrical Engineering, University of Illinois, Urbana, Illinois.*
Eshkol, N. et al. (1970b). Folk dances of Israel. Movement Notation Society.*
Eshkol, N. et al. (1971). The Hand Book. Movement Notation Society.*
Eshkol, N. et al. (1972a). The Yemenite dance: 142 steps. Movement Notation Society.*
Eshkol, N. et al. (I 972b). Twenty-five lessons by Dr. Moshe Feldenkrais. Movement Notation
Society.*
Ewer, R. F. (1963). The behaviour of the Meerkat, Suricata suricalla (Schreber). Z.
Tierpsychol.20:570-607.
Fentress, J. C. (1972). Development and patterning of movement sequences in inbred mice. In:
Kiger, John A. Jr. (ed.), The Biology of Behavior. Proceedings of the thirty-second annual
biology colloquium.
Fisher, H. (1965). Das Triumphgeschrei der Graugans (Anser anser). Z. Tierpsychol.29:247-
304.
Flynn, J. P. (1967). The neural basis of agression in cats. In: Glass, David C. (ed.),
Neurophysiology and Emotion. Biology and Behavior Series, The Rockefeller University
Press and Russell Sage Foundation, New York.
Fried, Y., and Agassi, J. (1975). Paranoia (Monograph). In preparation.
Friedman, M. B. (in preparation). Patterns of avian eye and head movements. Presentation at
the XI 11th International Ethological Conference, Washirigtoil, D.C.
Gibson, J. J. (1966). The Senses Considered as Perceptual Systems, Houghton Mifflin, Boston.
Golani, I. (1973). Non-metric analysis of behavioral interaction sequences in captive jackals
(Canis anreus). Behaviour XLIV 1-2 :89-112.
Golani, I., (in preparation). The display of Tasmanian Devils Sarcophilus harrisii with special
reference to the "joint" concept.
Golani, I., Eshkol, N., and Seidel, S. (1969). The Golden Jackal. Movement Notation So-
ciety.*
Gottlieb, G., and Kuo, Z. Y. (1965). Development of behavior in the duck embryo. J. Compo
Phys. Psychol. 50:183-188.
Granit, R. (1973). Demand and accomplishment in voluntary movements. In: Stein, R. B.,
Pearson, K. G., Smith, R. S., and Redford, J. B. (eds.) Control of Posture and Locomo-
tion. Advances in Behavioral Biology. Vol. 7, Plenum Press, New York and London.
Hailman, J. P. (1967). The ontogeny of an instinct: the pecking response in chicks of the laugh-
ing gull Larus atricilla I and related species. Behaviour Supplement 15.
Harries, J. G. (1969). Shapes of movement. Movement Notation Society.*
Heinroth, O. (1911). Beitrage zur Biologie, namentlich Ethologie und Psychologie der
Anatiden. Proc. 5th. Int. Om. Congr. (1910): pp. 589-702.
Hinde, R. A., and Stevenson, J. G. (1970). Goals and response control. In: Aronson, L. R.,
Tobach, E., Lehrman, D. S., and Rosenblat, J. S. (eds.), Development and Evolution of Be-
havior. Essays in Memory of T. C. Schneirla.
Kennedy, D. (1976). Properties of neural elements in relation to network function. In: Fentress,
J. C. (ed.), Simpler Networks and Behavior, Sinauer Associates, Sunderland, Mass.
Klopfer, P. H., and Hailman, J. P. (1967). An Introduction to Animal Behavior. Ethology's
First Century. Prentice-Hall, Englewood Cliffs, New Jersey.
Klopfer, P. H., and Hailman, J. P. (1964). Basic parameters of following and imprinting in
precocial birds. Z. Tierpsychol.21:755-762.
Kornhuber, H. H. (1974). Cerebral cortex, cerebellum and basal ganglia: An introduction to
their motor functions. In Schmitt, F. 0., and Worden, F. G. (eds.), The Neurosciences.
Third Study Program, The MIT Press, Cambridge, Massachusetts, and London, England.
Kuo, Z. Y. (1967). The Dynamics oj Behavior Development; an Epigenetic View, Random

House, New York.
Laban, R. (1956). Principles oj Dance and Movement Notation, Macdonald and Evans Ltd.,
London.
Lashley, K. S. (1930). Basic neural mechanisms in behavior. In: Beach, F. A., Hebb, D.O.,
Morgan, C. T., and Nissen H. W., (eds.) (1960). The Neuropsychology oj Lashley. Selected
Papers oj K. S. Lashley, McGraw-Hill, New York.
Loeb, J. (1918). Forced Movements, Tropisms. and Animal Conduct. With a new introduction
by Jerry Hirsch. Dover publications, New York, 1973.
Lorenz, K. (1949). Er redete mit dem Vieh, den Vogeln. und den Fischen. Borotha, Wien, pp.
205-209.
Lorenz, K. (1953). Verstandigung unter Tieren. Forum (Fontana, Zurich) 1:47-48.
Lorenz, K., and Tinbergen, N. (1939). Taxis und Instinkthandlungen in der Eirollbewegungen
der Graugans.l. Z. Tierpsychol.2:1-29.
Luria, A. R. (1966). Higher Cortical Functions in Man. Basic books Inc., Consultants Bureau,
New York.
Maturana, H. R. (I 970}. Biology oj Cognition. Biological Computer Laboratory Report no.
90, Department of Electrical Engineering, University of Illinois, Urbana, Illinois.
McCulloch, W. S. (1945). A heterarchy of values determined by the topology of nervous nets.
Bull. oj Math. Biophysics 7:89-93.
Miller, A. G. Galanter, E., Pribram, K. H. (1960). Plans and the Structure oj Behavior. Holt,
Rinehart and Winston, New York
Nelson, K. (1965). The temporal patterning of'courtship behavior in the glandulocaudine fishes
Ostariophysi. Characidae. Behaviour 24:90-146.
Nelson, K. (1973). Does the holistic study of behavior have a future? In: Bateson, P. P. G., and
Klopfer, P. H. (eds.), Perspectives in Ethology. Vol. I., Plenum Press, New York and
London.
O'Keefe, J., and Nadel, L. (1976). The Hippocampus as a Cognitive Map. In preparation.
Piaget, J. (1936). La Construction du Reel chez {,Enjant. Delachanz and Niestle, NeucMtel and
Paris.
Piaget, J. (1946a). Le Development de la lVotion de Temps chez {'Enjant. P. U. F., Paris.
Piaget, J. (1946b). Les Notions de Mouvement et de Vitesse chez {,Enjant, P. U. F., Paris.
Piaget, J., Inhelder, B., and Szeminska, A. (1964). L'Epistemologie de {,Espace, P. U. F.,
Paris.
Poincare, H. (1952). Science and Method, Dover Publications, New York.
Poincare, H. (1963). Dernieres pensees, Flammarian, Paris.
Powers, W. T. (1973). Behavior: The Control oj Perception, Aldine Publishing Company,
Chicago.
Prost, J. H. (1967). Bipedalism of man and gibbon compared using estimates of joint motion.
Am.J. Phys. Anthrop. 26,:135-148.
Prost, J. H. (1974a). Varieties of human posture. Human Biology 46(1}:1-19. Wayne State
University Press.
Prost, J. H. (1974b). An experiment on the physical anthropology of expressive gestures. In
Leaf, Murray J. (ed.), Frontiers oj Anthropology, D. Van Nostrand, Princeton.
Prost, J. H. (in preparation). Postural variety in baboons and humans.
Robinson, D. A. (1974). The effect of cerebellectomy on the eat's vestibuloocular integrator.
Brain Research 71:195-207.
Rosenberg, E. H. (1971). A changing view of experience. American Journal oj Psychotherapy
25:117-124.
134 lIan Golani
Rosenberg, E. H., and Medini, G. (1973). Comments on interaction in psychotherapy.

Contemporary Psychoanalysis 9:471-480.
Schenkel, R. (1967). Submission: Its features and function in the wolf and dog. Am. Zool.
7:319-329.
Schneirla, T. C. (1972). In: Aronson, Lester R. et al. (eds.), Selected Writings oj T. C.
Schneirla, W. H. Freeman, San Francisco.
Simpson, M. J. A. (1973). Social displays and the recognition of individuals. In: Bateson, P. P.
G., and Klopfer, P. H. (eds.), Perspectives in Ethology, Vol. I., Plenum Press, New York
and London.
Slater, P. J. B. (1973). Describing sequences of behavior. In: Bateson, P. P. G., and Klopfer, P.
H. (eds.), Perspectives in Ethology, Vol. I, Plenum Press, New York and London.
Spencer Brown, G. (1969). Laws oj Form. George AlIen and Unwin Ltd., London.
Strasser, H., and Gassman, A. (1893). Hulfsmittel und Normen zur Bestimmung und Verans-
chaulichung von Stellungen, Bewegungen und Kraftwirkungen am Kuge1gelenk, insbe-
sondere am Huft und Schalterge1enke des Menschen. Anat. Hejte 2:398-434.
Tinbergen, N. (1951). The Study oj Instinct, Oxford University Press, New York.
Waddington, C. H. (1966). Principles oj Development and Difjerentiation. Macmillan, New
York.
Wemmer, C. (1972). The behavior of genets Geneta tigrina and civets Civettictis civetta. Ph.D.
Thesis, Smithsonian Institution, Animal Research Department, National Zoo, Wash-
ington, D.C.
Whitehead, A. N. (1923). Relativity, logic, and mysticism. Proceedings oj the Aristotelian
Society jor Systematic Study oj Philosophy, Suppl. Vol. III: 34-41.
Whitehead, A. N. (1929). Process and Reality, an Essay in Cosmology, Gifford Lectures,
Edinburgh, 1927-8, Cambridge University Press, Cambridge.
Whitman, C. O. (1899). Animal behavior. Biological lectures, Marine Biological Laboratory,
Wood's Hole, Massachusetts.
\vilorf, B. L. (1956). Language, Thought, and Reality, Wiley, The Technology Press, M. I. T.
Wilden, A. (1972). System and Structure: Essays in Communication and Exchange, Tavistock
Publications, London.
Wilson, S. c., and Kleiman, D. G. (1974). Eliciting play: A comparative study. Amer. Zool.
14:341-370.
Wunschmann, A. (1970). Die Plumbeutler (Vombatidae). A. Ziemsen Verlag, Wittenberg
Lutherstadt.
Chapter 3
BIOLOGICAL RHYTHMS AS INFORMATION

CHANNELS IN INTERPERSONAL
COMMUNICATION BEHAVIOR
Paul Byers
Teachers College
Columbia University
New York, N.Y.
I. INTRODUCTION
In this paper I begin with the following assumptions which are implicit in a
General Systems or communication theoretical epistemology.
1. Implementing, maintaining, or changing the nature of human (and
animal) relationships is a process involving feedback.
2. The relating process, then, requires that each interactant have
continuous information about the state of the other.
3. The "state of the other" is multilevel. That is, there are states of
cells, organs, organ systems, the organism, etc.
4. Language, at best, functions only in relation to information avail-
able to the thinking-symbolizing component of humans, and may be a
construct of "thinking" rather than a verbal readout of actual states. In any
event, imagine two persons trying to develop a relationship with only se-
mantic information exchanged via teletypewriters.
In this framework I will describe research which supports the following
assertions:
1. In every human (and animal) interaction we can discover that the
motor onsets underlying movement or vocal behavior are of the nature of a
modulated rhythm and that, on one level of organization, the nature of the
relationship can be stated in terms of the relationships between or among
the modulated rhythms.
2. The basic underlying rhythms are related to a few brain rhythms or
oscillations.
135
136 Paul Byers
3. There are rhythms or rhythmic relationships on ascending levels of

organization. The ballisic motor impulses of the intercostal muscles, for
example, show rhythm at the level of syllable pulsing. On a higher level, the
back-and-forth conversation between two people also can be seen as an
oscillation.
4. The variation from purely mechanical, exact oscillation represents
information about the state of one actor which is continuously available (in
vis-a-vis situations) to the other(s). In music the variation from mechanical
rhythm is information called "expression."
5. There is a biological response in humans such that certain rhythmic
relationships are felt or perceived as good (sometimes referred to as "good
vibes") and others, mismatching, are biologically disturbing. We recognize
that one usually feels uncomfortable if he is late, out of step, etc. This also
operates at lower levels (higher frequency). We recognize that "timing" is
important in team sports, music, acting, etc.
6. Except for cultural conditioning, which may reinterpret "good"
feeling as bad and vice versa (e.g., pleasure can be seen as seductively dan-
gerous and suffering as righteous and ultimately rewarded), the nature of a
human (or animal) relationship is isomorphic with the nature of the
rhythmic relationship between or among interactants.
I will report on specific research, including my own, which identifies
rhythms underlying performed and communication behavior; I will discuss
the methodological and theoretical implications of this research; and I will
suggest the implications for this emerging line of research for an emerging
new view of man and human nature.
II. BRAIN RHYTHMS
In 1951, at a symposium concerned with "Cerebral Mechanisms in Be-

havior," Karl Lashley presented a paper titled "The Problem of Serial
Order in Behavior" (1951, pp. 112-136). He pointed out that both associa-
tive chain theories and the reflex arc theory of brain function were "doomed
to failure because they start with the assumption of a static nervous system"
(1951, p. 135). He said that the available experimental evidence showed the
brain to be either a dynamic, constantly active system or a composite of
many interacting systems. He suggested, although without adequate empir-
ical evidence to support the hypothesis, that internal and external rhythms
should be considered. Three quotations from Lashley (1951) are particularly
Biological Rhythms as Information Channels in Communication Behavior 137
relevant to this research:

... the problems raised by the organization of language seem to me to be characteristic
of almost all other cerebral activity. There is a series of hierarchies of organization; the
order of vocal movements in pronouncing the word, the order of words in the sentence,
the order of sentences in the paragraph, the rational order of paragraphs in a discourse.
Not only speech but all skilled acts seem to involve the same problems of serial order-
ing, even down to the temporal coordination of muscular contractions. (pp. 121-122)
I believe that there exist in the nervous organization, elaborate systems of interrelated
neurons capable of imposing certain types of integration upon a large number of widely
spaced effector elements. (p. 127)
The remaining alternative (to the reflex arc or associative chain theories) is that the
mechanism which determines the serial activation of the motor uqits is relatively inde-
pendent, both of the motor units and of the thought structure. (p. 118) (Parentheses
mine)
A. Sollberger, who compiled a massive volume titled Biological

Rhythm Research (1965), wrote:
All living organisms are set into the solid framework of the physical world, the structure
of which pulses with an abundance of "external rhythms": diurnal, tidal, seasonal,
solar, sidereal. The capacity to follow them, to oscillate, would certainly enhance the
survival potential of a species ... The construction of our bodies (randomly,
genetically, and selectively determined) should therefore include components which are
capable of rhythmic function. They do not necessarily have to be formed for this pur-
pose but they must exist. Neither need they be absolutely conditioned by the environ-
ment. Once being present, for whatever reason, they might sometimes be expected to os-
cillate even if the external world is removed. (p. 3)
Sollb.:;rger cites the 10 c/sec "alpha" rhythm, but without reference to

any behavioral correlates, as well as a rhythm with a frequency of 7 c/sec,
which is found both in EEG traces and in assorted experimental situations.
He says, "The highest rate of nervous feedback oscillation in Homo would
seem to lie around 7 c/sec simultaneously setting the upper frequency for
rhythmic movements" (1965, p. 81).
Mary Brazier (1960) analyzed brain waves by an autocorrelation
method and found a 7.3 c/sec rhythm (from electrodes over the temporal
area and the vertex) and a 9.6 c/sec rhythm (from the occiput). These two
rhythms correspond to the two cited from Sollberger.
Lindsley (1961), Wolff (1968), and others believe that at least some
brain rhythms are related to a mechanism which paces and/or integrates
motor behavior. It is reasoned that one learns to make relatively complex
movements, or pronounce single sounds, that may involve the coordination
of dozens of muscles as single units. This was the problem Lashley dis-
cussed as "serial ordering" or "serial activation."
If we accept the idea that brain waves reflect internal biological
rhythms which pace and/or integrate muscle activation, we are still left
with the problem of demonstrating the relationship empirically and of
138 Paul Byers
explaining the process by which a few basic rhythms of fixed frequencies

can fit the wide range of human activities, which clearly proceed at rates re-
lated to tasks and are not limited to the few basic rhythmic rates that have
been derived from analyses of brain waves.
Peter Wolff (1968) has investigated this development from simple to
complex rhythms, and, in particular, has shown that the stereotypic be-
havior (e.g., rocking) of mentally retarded or pathologically regressed
children is sufficiently regular to suggest that such behavior is the
expression of simple, undifferentiated infantile rhythms. He says,
We are a long way from demonstrating how complex phase interactions develop from
simple repetitive rhythms, or how the contributing oscillators can be assembled in ever
new combinations at a rapid rate to conform to shifting demands of the intended task.
(\968, p. 481)
It has long been understood that the sense of time in children becomes
evident around age 4 and is "mature" by early adolescence, and that this is
paralleled by concomitant changes in EEG traces. Whatever the develop-
mental mechanisms, it seems likely that the "maturing" found in the sense
of time and in brain waves is also in some way parallel to the development
of motor coordination skills. Perhaps training in skills (tennis, table tennis,
playing musical instruments, etc.) might usefully be seen as a matter of-
using Wolff's language-assembling the contributing oscillators into com-
binations which conform to the demands of the intended task. Perhaps this
is ~he process referred to by our word "practice."
Although he has not been concerned primarily with brain rhythms,
Karl Pribram (1971, pp. 206-208) cites evidence that the predominant
synchronized brain rhythm, the 10 c/sec "alpha," which is most evident in a
resting state, becomes desynchronized at arousal and differentiates into
multiple rhythms. Lindsrey (1961) believes that, when desynchronized, the
neural elements become functionally independent of each other and are
available as separate information processing channels.
III. OTHER RHYTHMS
To discuss the matter of rhythms in information processing we must, in

addition to the above, examine the nervous system's capacity to use a single
beat of a rhythm as a comparator or reference in feedback loops, which
may embrace (1) internal structures or organs, or (2) two or more
organisms. And to do this it will be necessary to look at rhythms of longer
periods or slower frequencies, where experimental research has managed to
shed some light on temporal properties of events.
Zelenii (1907), working in Pavlov's laboratory, discovered that if the

salivary conditioned reflex in dogs is reinforced at regular intervals of ten
minutes, the salivation will appear in the same regular intervals without the
. presentation of any stimuli. This is now known as the temporally-condi-
tioned reflex. Five years later, Feokritova (1912) reported that an important
aspect of the temporally-conditioned reflex is its great rhythmic accuracy-
about 3 percent-and that, unlike the usual conditioned reflex which is ex-
tinguished gradually, the temporally conditioned reflex is extinguished
abruptly. A few years later, Deriabin (1916) found that the nature of the
sound used as stimuli is immaterial. Dogs learned to differentiate a stimulus
on the basis of its rhythm or duration without regard to the particular
quality of the sound, whether a whistle, bell, or trumpet, etc.
Another elegant experiment that demonstrates the significance of dif-
ferent time durations in relation to brain function was performed by Eugene
Sokolov (1960, pp. 187-276). He presented subjects with a horn beep of a
given loudness and duration. The subjects oriented toward the beep stimulus
and, apart from the body orientation, there were concomitant changes in
the subject's EEG components, increased blood flow to the head at the
expense of flow to the fingertips, changes in the electrical resistance of the
skin, and changes in heart and respiratory rate. The stimulus was repeated
again and again until the response disappeared. The subject was said to be
"habituated" to the stimulus. At this point the stimulus was again pre-
sented, bllt at a lower intensity, and the full-blown response reappeared.
This established that a critical component of the stimulus was its intensity
or loudness. Then the subject was again habituated to the horn beep stim-
ulus, and at that point the stimulus was presented again, but this time the
duration of the beep was shortened. And again the full-blown response oc-
curred. From this experiment we can see that both intensity and duration
are crucial (structural) components of the event in Sokolov's experiment.
We can also see the structural significance of duration in common
social behavior. We can easily recognize that eye contact may be a signifi-
cant component of interpersonal relatedness, but within that component we
can also see that the duration itself may be critical. A person walking in
Times Square in New York City, for example, is permitted only fleeting eye
contact with passing strangers. The duration of this allowable eye contact
between strangers could conceivably be measured. If, then, a woman in the
street prolongs her eye contact with a passing male stranger, he may read
into that prolongation a different message. Prostitutes can use this prolon-
gation to identify themselves to perceptive passing men, and it is probable
that men looking for prostitutes search the female passersby for this prolon-
gation. Although such signaling behavior will have components other than
prolongation of eye contact, if all the other behavior were held constant, the
140 Paul Byers
time dimension-the duration of eye contact-would alone probably

identify the special signal in this context.
People and the animals who live with them can, to some extent, recog-
nize shifts (from some baseline) that signal state changes. In both people
and animal companions we can recognize changes in qualities of movement
behavior that we interpret as signaling a new state-of tiredness, illness,
depression on one hand, or excitement, impatience, or restlessness on the
other. I believe that we could demonstrate that the temporal qualities of
movement behavior are more fundamental and universal than the spatial
configurations of movement and gesture. I recognize that at the level of in-
formation transmission in the nervous system, even spatial information is
temporally encoded.
If a stage actor wants to imitate, or mime, a crow, an elephant, or a
terrier, he cannot look very much like these animals, but he can get ap-
plause from his audience if he actually reproduces the temporal characteris-
tics of the movements of the animal he is imitating. The lumbering walk
and head swing of the elephant, the sudden head shifts of the crow, or the
rapid restless shifts of the terrier's feet or his rapid tail wag, all suffice to
identify the particular animal and require the addition of only minimal
spatial cueing.
Musicians know that it is possible to recognize a particular piece of
music by hearing only its rhythm tapped out-without the tune or harmony.
But it is not possible to hear or even imagine a tune without a rhythmic
frame.
When, then, we conceptualize behavior or events in terms of Lashley's
"hierarchies of organization," we can see that a rhythm can be conceived as
a whole event with its own properties (on one level of organization). But it
may also be seen as a component, part, or quality of events or behavior on
the next higher level of organization. To use a simple, mechanical analogy,
the wheels in a clock may exist independently of the clock (which requires a
particular structural relationship among its wheels), but the clock cannot
exist independently of its component wheels.
The relation of poetry to its underlying rhythm was considered in a
small monograph by R. H. Stetson (1903), entitled Rhythm and Rhyme, in
which he argued that rhythm was a thing apart from the content that ac-
companied it. Stetson believed that the meter of poetry and the rhythm of
music were, in some sense, separate from the words or melody even though
they were packaged together.
In 1905, Stetson wrote a long two-part paper entitled "A Motor Theory
of Rhythm and Discrete Succession" (Stetson, 1905, pp. 250-270, 293-350).
In this paper, he disagreed with the followers of Wundt, who conceived of a
"pure" rhythm which was modified by the nature of the "ideational
content" superimposed on it. Stetson said that the prevailing concept was
incorrect:
The material rhythmized is conceived (by followers of Wundt) as an antagonistic force
which destroys the regularity and therefore the "purity" of the rhythm. The more elabo-
rate the "ideational content," therefore, the less regular the rhythm and the more must
groupings depend on a course foreign to the rhythm. (1905, p. 251)
In the same article he wrote, "The most important natural rhythm-produc-

ing apparatus is the vocal apparatus" (1905, p. 257). Stetson distinguished
between the mechanical cycling of exact intervals and "rhythm," which he
acknowledged was inexact since it was reflecting rapidly shifting inner af-
fect-states. But he did not believe that it was the "ideational content" of the
speech or music or poetry which changed the rhythm.
Forty-six years later, Lashley was to say" ... the mechanism which de-
termines the serial activation of the motor units is relatively independent,
both of the motor units and of the thought structure" (Lashley, 1951, p.
118). In the language of his report the statement is that the rhythm and the
context simply exist on different levels of organization.
In his last work, Motor Phonetics (1951), published the year after his
death, Stetson wrote of the significance of rhythm in an even larger frame:
The rhythm is certainly one of the most fundamental characteristics of the utterance of
a language, and is often most difficult for a foreigner to acquire .... The tremendous
changes which have taken place in the "sounds" of the French as it developed from the
vulgar Latin must have been due to something besides the ordinary changes to which a
language is subjected ... There can be little question that a profound change in rhythm
underlies the extensive alterations during the transition from Latin to French. A new
rhythm has produced a new language, reshaping the ancient words, eliminating syllables
and shifting the stress to an alternation from syllable to syllable ... such a change of
rhythm and of stress is apparent in the Hebrew over against the Arabic, and in
the Czech over against the Lithuanian, and in the French over against the Italian.
(pp. 124-5)
In contrast to descriptive linguists, who provide elegant structural

descriptions of man's languages as symbol systems, Stetson analyzed lan-
guage as rhythmic motor behavior on ascending levels of organization. Al-
though he did not relate these rhythms to biological rhythms, his minimal
vocal unit, the syllable pulse, which rides on the energy from an intercostal
muscle ballistic contraction, is also a unit which emerged in the research to
be described later in this report.
In 1940, Eliot Chapple and Conrad Arensberg introduced the idea to
anthropology that human relations (interaction) could be measured.
Interaction was conceived as a sequence (rhythm) of temporally interrelated
. actions and inactions between (or among) individuals. The quantitative
analysis of durations of action-inaction in paired interactions was seen to
provide the basis for a description of the nature of the interaction, as well as
142 Paul Byers
providing "fruitful aids to the study and analysis of personality" (Chapple

and Arensberg, 1940, p. 47). This was the beginning of what has come to be
called "interaction chronography." Apart from the measurement aspect of
the Chapple and Arensberg paradigm, this was also one of the beginnings of
a shift from description in terms of the presence or absence of content items
of culture to a concern for the dynamics or sequencing of interpersonal rela-
tions, a shift from an emphasis on content to a concern for the temporal
structure of process.
A second direction was taken by Chapple himself, who designed the
first interaction chronograph, as well as a method for assessing certain per-
sonality characteristics of individuals (Chapple, 1949, 1953). Chapple used
the interaction chronograph to measure the interaction behavior of a sub-
ject in a controlled interview. He found that certain properties of the
interaction behavior of individuals were invariant in given personalities.
In 1942, Chapple wrote:
Personality depends on a combination of variables: (I) the amount of interaction which
a person requires, (2) the frequency of his habitual interactions, (3) his origin-response
ratio, (4) the rhythm of his interaction rate, and (5) his ability to synchronize, or adjust,
to others. An individual's status in each of these variables is acquired by a combination
of conditioning and genetic inheritance (Chapple and Coon, 1942, p. 69).
IV. INFORMATION PROCESSING
Now that we have ranged from "primary" 10 c/sec brain rhythms to

higher levels of rhythmic organization that are seen as the underlying tem-
poral frames for events ranging from single muscle activations to dyadic
interactions and even to entire languages, we can look at the processes
which represent "information processing."
When a generalized, steady, mechanical brain rhythm differentiates
into multiple rhythms (even of the same frequency or rate) which are avail-
able to "neural elements," the state of the neural element or structure will
be reflected in the modulation that is impressed on the rhythm of that struc-
ture. In a general sense, the study of electroencephalography is a matter of
inferring relationships between evoked potentials (electrical activity
recorded at the scalp) and brain activity, and of making inferences about
the state of brain structures from variations in the emitted brain waves at
various sites.
I believe that this relationship holds at all levels of organization. Each
individual's rhythm, at a given level of organization and/or as related to
specific neural structures, contains information-in the form of modula-
tion-concerning the state of the structure. And each communicating
"other," whether that "other" is another organ within an individual or

another person in a communicative relationship, has its own rhythm, as a
comparator for sensing, reading, or decoding the information inherent in
the modulated rhythm of the "other."
A familiar example can be seen in musical performance. As formally
written in a score, music would appear to be mechanically regular in
rhythm (with the exception of written indications for pausing, holding, slow-
ing, etc.). But if music were performed mechanically, it would be said to
have no "expression." Expression, or "meaning," is impressed on the un-
derlying mechanical rhythm by modulating or varying the temporal rela-
tionships between beats.
This example from music can be extended to illustrate a problem in
rhythm research. How does one go about identifying the underlying
mechanical beat or rhythm when, in expression, it is almost always modu-
lated or varied from beat to beat by the state of the organism (i.e., the
performer)? In the case of music, except for some contemporary music, the
listener is almost never in doubt. He may tap his finger or foot "in time
with the music" even though the music is, to some extent, always
expressively "out of time." I assume that the range of "expressive" varia-
tion-beyond which one "loses the beat"-is limited. I have processed
musical performances in such a way as to measure precisely the intervals
between performed beats, and have found that one can, indeed, measure the
variation. But I have found no mathematical procedure for establishing the
rate of the underlying rhythm, except by averaging intervals.
This leads to an alternative view: that the reference or comparator is
always in the "head" of the listener, whose own reference beat is never itself
mechanically regular but always reflects his own state at the moment. I will
discuss this matter more fully later in this paper, but at this point it is useful
to call attention to the methodological problem of establishing frequencies
or rates of underlying rhythms from observed behavior, and to point out,
again, that whereas externally imposed rhythms (e.g., circadian day-night
oscillations) provide an available reference or comparator, internal rhythms
are idiosyncratic, and the reference or comparator rhythms of individuals
are normally always themselves modulated. In human relations, then, when
the rhythms involved are endogenous-as in all the examples involving
human relationships described above-there is no absolute measurement
against an external referent; there are only relationships.
To summarize: there are at least two primary brain rhythms, with fre-
quencies of 10 c/sec and 7 c/sec, found in the human brain. The 10 c/sec
rhythm has been found in animals as far away from man as the water flea.
These rhythms, acting as pacers and/or integrators of motor activity, lie at
one level of organization, but all behavior can be seen as lying along a
144 Paul Byers
hierarchy of levels of organization and the rhythms involved as being

specific to levels. A conversation between two persons can be described as
an oscillation which has an irregularity or asymmetry assignable to the dif-
fering personalities of the individuals and to the state of their immediate
relationship. The states of the individuals and the state of their relationship
can be described in terms of temporal measurements of their behavior,
singly and conjoint. This information is continuously available to persons in
vis-a-vis (communication) relationships, although the only reference or
comparator that a person has available is his own state-modulated rhythm.
The implications of this for group behavior will be discussed later.
I believe that these interpersonal rhythm relationships are what is
popularly referred to as "getting it together" or sharing "vibes."
v. RESEARCH PROCEDURES AND DATA
A few years ago, I set out to explore the temporal dimensions of

human vis-a-vis interaction. Condon and Ogston (1971) had reported an
extremely precise self-synchrony and synchrony between interactants, but
had offered no explanation that related this synchrony to the qualitative na-
ture or the structure of the interpersonal relationship under observation. Al-
though I am an anthropologist, I reasoned that our scientific responsibility
required a form of explanation that would span the spectrum of behavioral
organization stretching from brain structure to social structure. This in-
volved crossing the (imaginary?) line dividing biology from culture.
To begin with, I acquired copies of film data from three cultural
groups: the Netsilik Eskimo, the Bushmen of the Kalahari Desert in Africa,
and the Maring of New Guinea. There is a large archive of film for each of
these peoples, and I selected about 2500 feet of film from each. The selec-
tion criteria were simply, (1) that two or more interacting individuals be
continuously visible, whole bodies if possible, for at least 30 seconds, and
(2) that synchronous sound tapes be available for at least some of the film.
These criteria were difficult to satisfy, since most cinematographers regard
prolonged sequences as dull and either move the cameras around the scene,
zoom in for close-ups of individuals, or take only brief shots. For data pur-
poses, the requirement is for an unmoving camera-on-a-tripod and long,
unbroken sequences of the whole bodies of all interactants.
I used an analyzer projector, with which it is possible to move the film
back and forth at any speed, even frame-by-frame. The film was frame-
numbered with consecutive five-digit numbers on successive frames.
Many of us who have worked on various kinds of behavioral analyses
of film-recorded behavior have observed that changes of one sort or another

(changes in direction of movement, duration of movement, repetitions)
often occur at five or ten frame intervals. To my knowledge no one has at-
tempted to explain the significance of this duration. In my own research it
was immediately apparent that the meaning or significance of a single tem-
poral unit or its octave would not emerge from any simple, linear counting
procedure, but might have significance when the multilevel structure of the
behavior could be described.
There is now a large literature demonstrating the multilevel structural
organization of behavior (indeed, of all phenomena). In the analysis of body
motion communication there are two other matters to consider. First, body
motion may be multichannel. That is, the actor may be acting in relation to
more than one "other." He may be talking to one person, glance at another,
wave to a third, caress a child, and carry out a task, all channels which may
be operating simultaneously or woven closely together. There is some sense
in which these channels will be interrelated, and some sense in which they
will be independent. Second, some part of body motion is related simply to
gravity opposing, balance maintaining, inertia, or coming to rest.
It is far easier to "analyze" another person's behavior intuitively-i.e.,
to communicate or interact appropriately with him-than it is to segment
and level the multichannel data for purposes of scientific description.
There are two procedures which can somewhat reduce the behavioral
complexity. One method is to select scenes or events in which the par-
ticipants are, as nearly as possible, "doing nothing"-i.e., where task in-
volvements are minimal or absent and where interpersonal involvements are
present but minimal-or are, as nearly as possible, single-focused, as in inti-
mate relationships or in formal events (where the form of the performance
takes precedence over the content). A second method processes interper-
sonal speech in such a way that the underlying motor impulsing can be use-
fully distinguished from the content material (speech sounds). Both of these
"screening" procedures are illustrated below.
Figure 1 was derived from a scene in which three Eskimos are "doing
nothing." They have dragged a seal from a hole in the ice to a place on
land, and one man is working on the seal while the other two are simply
looking on. One onlooker is standing, the other sitting on a rock. The third
is on his knees working on (skinning?) the seal. There is a minimum of move-
ment. The standing man moves one hand from his head to a limp, hanging
position alongside his body. He then brings the hand up to grasp a staff he
is holding vertically with his other hand, and this other hand moves to his
head where it moves in two circles (scratching his temple) and then comes
back down to the staff. He then shifts his weight from one foot to the other,
lifting each slightly in turn. The second person, a woman, is looking into
146 Paul Byers
AI
I
c I I
SUM I I I I I I
10 cps rhythm
3 Eskimos
Fig. I. A notational record of movements of three Eskimos. The squares, reading left to
right, represent individual motion picture frames. The blackened squares are those on which
onsets of movement for each of three persons are seen on slow projection. When all the move-
ment onsets of the three persons are summed in the bottom line, it can be seen that all are
moving in relation to a common rhythm. The frequency is 10 cps. There is no external pacer.
space; at one point she turns her head suddenly, and otherwise is idly mov-
ing a small stick up and down. The third person, working on the seal, sud-
denly rises on his knees, lifts his head, turns toward the camera, turns
back, sinks back down, and continues working on the seal. The entire
filmed "take" lasts only about 15 seconds.
In Fig. I I have noted the time, i.e., the specific film frames, at which
changes in motion occur for each person. A change is defined as the point
at which visible movement of a body part begins, ends, or changes direction
or velocity. The three uppermost lines of Fig. I each represent a single actor
in the scene. The bottom line sums all the onsets~changes of all three
MALEI
FEMALE
SUM.
20 30 40
I
• •
50
I -- 60
I
••
70
I
80
•
•••
90
I
I
100 110
••
I:!O
I
10 cps rhythm 2 Rhesus monkeys
Fig. 2. A record made in the same way as that shown in Fig. I except that the interactants
are a consort pair of rhesus monkeys. The frequency of the rhythm is also 10 cps. It appears
different from Fig. I only because the film of Fig. I was made at 24 fps and the film of the
monkeys was made at 16 fps.
persons. From the bottom line we can see that all three people are moving
in relation to a commonly shared beat or rhythm, and that the rhythm is
proceeding at a rate of 10 frames (approximately·Oo4 sec).
This duration of 004 sec. is the same (recognizing that "same"
embraces a small range of possible variation) as that of the military march
rate, is half that of the accepted heart rate at rest, and is equivalent to four
cycles of the familiar "alpha" rhythm of encephalography. The same
rhythm sharing phenomenon is found with such regularity in interpersonal
and group behavior in all cultures I have examined (using film records) that
I believe it to be a human communication universal. I will show evidence
that there is an underlying biological rhythm with a rate of 10 c/sec. This is
the rate of the alpha rhythm. There are other possible brain rhythms, and
human behavior is performed at rates appropriate to tasks, but it appears
that human communication or interaction always shows evidence of an un-
derlying rhythm of 10 c/sec.
I define human communication as the process or processes which im-
plement, maintain, or change the state of relatedness between or among
people. And since the "state of relatedness" can be derived from observa-
tions of rhythm sharing (which I will discuss in more detail later), I find it
useful to say that individuals are "in communication" when this shared
rhythmic relationship, at this 10 c/sec rate or frequency, is occurring.
A piece of film from the Marin~ of New Guinea can be used to
illustrate and refine the relation of rhythms to communication. In the film
about a dozen people are standing or seated in a clearing near a small settle-
ment. There are two clans represented in the group, and the clans are hav-
ing an argument concerning the killing of a dog belonging to a member of
one clan by a member of the other clan. Clan members support each other.
This much information was available in the ethonographic notes ac-
companying the film. I could not identify the clan membership of the indi-
viduals in the film by dress, spatial distribution, or postural or seating orien-
tation, etc. But I found that as each of the many speakers talked, about half
of the group moved with~i.e., in a rhythmic relation to~a given speaker.
By noting these relationships, it was possible to discover that each of the
two halves of the group moved together and that any two people from dif-
ferent halves did not. My guess is that this procedure produced an identifi-
cation of clan membership, but I have not been able to verify that guess.
Had it been possible to verify the clan membership identification, it
would have been possible to proceed to the more difficult question: Where
in the interpersonal behavior is the synchrony that is required for any com-
munication or interpersonal involvement? And where is the dissynchrony
that is the concomitant of the argument or disagreement? I would expect
both to exist and to lie on different levels of organization, since, for
148 Paul Byers
example, one person must be synchronously related to another's rhythm in

order to "interrupt" him-i.e., violate that synchronous relationship.
Some years ago, for purposes other than the study of rhythmic rela-
tionships, I took film of the paired interaction of rhesus monkeys at Cayo
Santiago, an island near Puerto Rico. I have applied the observational
methods used in the Eskimo film (Fig. 1) to a stretch of film of a consort
pair of interacting monkeys. The results are shown in Fig. 2. The rhythm
sharing is clearly present and the rate appears to be the same. In this
monkey film, however, the problems of multiple channeling were clearly
present. The female was intermittently approached by her own infant and
some of her interaction was involved in chasing the infant away; she was
also intermittently diverted by interested members of her troupe and, occa-
sionally, by my presence. The chart in Fig. 2 was made of behavior between
"distractions." Unless the behavioral analyst studying rhythms from
recorded data has the necessary contextual information about the interac-
tion, he can be confused by this problem of multichanneling.
I would think it important to examine the rhythmic relations in animal
behavior within and between, or among, species. Research of this kind with
some animals may require film records in which the frames-per-second rate
is faster than the usual 24 fps.
Because of the limitations of 24 fps film, and the problems introduced
by the multichanneling of movement relationships, this research moved in
the direction of processing and analyzing interpersonal speech. behavior.
Condon and Ogston (1971) and Birdwhistell (1970) have shown that both
speech and movement behavior are part of the structure of communication
and are not processually different systems of communication. I will show a
series of charts generated from speech behavior, but first it is necessary to
describe the rationale and procedures involved.
The goal of speech processing is to produce an electronically generated
pen tracing of the energy (loudness) of the ongoing speech so that the motor
bursts underlying the vocal activity, as well as the points of speech stress,
can be pinpointed in time. The chart record should be of such dimensions
that intervals of tenths of a second will be far enough apart on the chart for
accurate measurement.
The first component of this rig is a tape recorder with an output
(earphone) jack and with four-speech capability. The speech to be processed
is recorded (or rerecorded) on reel-to-reel tape at a speed of 7Y2 inches per
second. The speed is then reduced to 1%6 ips for playback-processing. This
drops the frequencies by a factor of eight, and into a range that can be han-
dled by a simple analogue pen recorder. Normal speech frequencies lie in
the 250-350 c/sec range (ignoring the harmonics or upper partials which
can be filtered out or, in any event, will not seriously distort the trace). The
inertia of the pen in simple analogue recorders will not allow a linear trace
at these (250-350) frequencies, but when the tape recorder speed is reduced
eight times the reduced frequencies (about 31-44 cjsec) are manageable.
The slowed speech, then, is fed to the recorder. The trace that emerges
shows speech packages or bursts (see Fig. 3) that contain information about
both loudness and frequency. To eliminate the frequency component and
increase the linearity of the loudness component, it is possible to insert a
diode in one leg of the circuit between the tape recorder and the pen
recorder, and to put a small (0.2 to 2 microfarad) capacitor across the cir-
cuit to smooth out the small fluctuations in loudness and reduce or
eliminate the remaining frequency component. Figure 4 shows the trace of
the same speech produced by this method. This illustration is of Bushman
speech, and I will now examine it in some detail.
Storytelling is a common cultural practice among Bushmen. In the
present example a person known to be a particularly good storyteller has
been asked to tell a series of stories "to" another Bushman. In practice one
Bushman does not talk or tell stories "to" another. He talks the "story
line," while the other person or persons participate by imitating part of his
gestures in a prescribed way, by adding comments on top of his speech, and
by repeating syllables of his speech at the end of certain segmental or
syntactic units. This practice of repeating the speaker's last few syllables at
points of pause in his speech is commonly found throughout the world.
In the southern, and perhaps other parts of the United States, a listener
may insert "uh huh," "yes," or some other sound into the speaker's
pauses, usually accompanied by a head nod and perhaps other gestural
acknowledgment.
Figures 3 and 4 are pen tracings of the same stretch of the Bushman
storytelling, i.e., the first five seconds of a sound film record made by John
Marshall. Figure 3 is a tracing of both frequency and amplitude, and Figure
4 is a tracing of the amplitude alone. In these traces one second of real time
is represented by 80 mm on the chart. One tenth of a second of real time
equals 8 mm on the chart.
We can first observe, in Fig. 3, that the speech sound is not continuous,
but emerges in bursts with silence between. We can also observe that there
is some suggestion of regularity, or even spacing, in part of the vocal
stream. If we place a strip of paper with marks representing tenths of a
second intervals alongside (below) the charts, and if we position the strip
with the marked mechanical rhythm such that the first speech onset of the
first speaker (the storyteller) matches one of the beats on the marked
reference strip, we discover that some of the other bursts do, indeed, coin-
cide with the beat points on the reference strip. Now, if we reposition the
reference strip so that one of the beat points falls on an onset or burst that
~
-
Fig. 3
~ '" \ ~ ~ aJ I r~ ~\ /\ ~'If.~ ~ ~ ~ \ .
LM..\ ~~ ··lV\\~ l' ~ l; ~ l\ ~ \..: V\)~ \~ \;\: U~~L,~
Fig. 4
,-
:I .'1 .lj\ - f' II I\ , "~ '-Icn ~ I[J, ,\ ' ~-

-f~~ \~\ ,\\: l:
I
l~ ~ \jill ~ \~ Ut\!YU~~ \~~\~\-!~l~~\U,,-
H H
Fig. 5
Figs. 3, 4, and 5. Strip-chart records of 5 sec of talk of two Kung Bushmen in Africa. One man is telling a traditional story and the second interjects .,'"I:l
listener-comments. In Fig. 3 the tracings show as clumps, minimal sound packages, or syllables, and show both frequency and amplititude. In Fig. 4 e.
the frequency component is filtered out so that the exact points of motor impulsing can be seen. In Fig. 5 the sounds of each man are found to lie on a =
~
\0 cps rhythm train. ;;l
did not fit the first positioning, we can see that the onsets of those bursts
which did not fall on the first positioning do fall on the second positioning.
That is, there are two intermixed, out-of-phase sets of speech bursts, each
falling on its own 0.2 sec rhythm train.
When we listen to the audible speech in relation to Fig. 3
(electronically slowing down the tape speed somewhat to make the listen-
ing-matching more manageable) we discover that those bursts belonging to
one speaker fall on the same rhythm train (i.e., points on the reference
strip), and that those bursts belonging to the other speaker fall on the other
rhythm train. Figure 5 shows the same stretch of speech with both rhythms
(one for each speaker) marked on the chart trace.
This continuous rhythm train will continue in the normal flow of the
discourse for the duration of a clearly marked segmental unit. Since the
segmentation or punctuation of the behavior stream is also the identifier for
levels of organization of the behavior, it will be useful to digress long
enough to discuss the matter of segmentation and its relationship to
rhythms.
The epistemological, and consequently the methodological, frame of
this research is one in which the search is for the processes which connect or
relate the parts or variables in a system. The concern is not for "what hap-
pens," but only for "how it happens." The syntactic sentence of the linguist
is, in this sense, a system requiring a particular arrangement of parts to
form S0111e whole (a syntactic sentence). Phonemes combine to form
morphemes, morphemes combine to form phrases, phrases combine to form
sentences, etc. The linguist has operational procedures for segmenting the
vocal stream into units and for further dividing these units into subunits.
The linguist studies "an arbitrary system of vocal symbols by means of
which people cooperate" (Bloch and Trager, 1942, p. 1). He cannot, then,
include movement behavior in the same framework. This research focuses
on the organization of vocal or movement behavior, and not on language
per se. In this way it is possible to study the organization of vocal and
movement behavior (i.e., the processes) in the same methodological frame,
and to demonstrate that both speech and movement share certain underly-
ing organizational characteristics or properties. In the present case, both
speech and movement share the same limitations set by the preprogram-
ming of acts in the brain (see a later discussion of research on this point by
Clynes, 1970), and these minimal brain programs combine in part-whole
fashion to form larger segmental units. The present discussion of a
particular segmental unit, then, is directed toward the demonstration that
both speech and movement share underlying organizational properties and
that, at some levels of organization (if not at all levels), the segmentation
can be derived from either the vocal behavior or the movement behavior.
152 Paul Byers
The segmental syntactic unit that I have referred to is identifiable in

the Bushman case by a number of "marker" behaviors:
1. The speaker (and possibly the listener) takes a breath. This is audi-
ble only on carefully made sound recordings, usually using lapel or lavalier
microphones.
2. In ordinary Bushman conversation (I am told by ethnographers) the
"listener" repeats the last few syllables of the speaker's utterance at certain
pause points. In the storytelling instance, the speaker ends each of these
segmental units by uttering an "Mmmm" sound which is repeated by the
listener in overlap, i.e., the listener's "Mmmm" sound begins before the
speaker's ends. Because this canon-overlapped "Mmmm" sound has a
particular shape on chart traces, the storytelling can be segmented at this
level of behavioral or linguistic organization by noting these shapes on the
traces.
3. In the film, the listener is seen to repeat the last element or
component of the speaker's gesture. (The part repeated also represents a
segmental unit on a lower level.)
4. As this gestural sequence by the speaker ends, he brings both arms
close to his chest, crossed at the wrists, and at the same time bobs his head
downward.
5. The listener follows the speaker by repeating the wrist-crossing and
head bob.
I would assume that there is a specific stress and/or intonational
contour (of a limited number of possible contours) to these segmental units.
I assume this only because syntactic units of comparable length have such
contours in other languages.
This particular segmental unit, then, is marked or signaled with
considerable redundancy. Within this unit, of course, there are subunits and
within those are sub-subunits. If I were to say (and I have no linguistic evi-
dence for this) that the unit described above is a syntactic sentence, then
within that there would be units of phrase size, and within phases other
units, down to syllable bursts and even syllable parts. This follows a form of
hierarchical description used by Stetson, who studied the motor activity un-
derlying speech, and not the symbol system itself.
Each of these units can be identified by markers that have both vocal
and/or movement concomitants. In Fig. 3, for example, the small bursts
that are bounded by silence or pause are syllable bursts. And when we
measure the timing or rhythm of these bursts, we discover that no syllable
burst of a single speaker begins until at least 0.2 sec after the beginning of
the preceding syllable burst. That is, a single syllable burst requires a 0.2 sec
time-slot regardless of its actual duration.
This 0.2 sec interval is represented by five frames of film, and at this
Biological Rhythms as Information Channels in Communication Behavior IS3
point we begin to fit the five-frame unit mentioned earlier into a larger tem-
poral structure. When the Bushman storyteller waves his arm back and
forth in this first (segmental) unit of the storytelling, each swing of his arm
requires 10 frames, and other gestural parts in this (and probably any)
human communication situation show this interval. This interval is, of
course, the one cited earlier for the Eskimo shared rhythm and that of the
rhesus monkeys. It is evidence of an underlying biological rhythm.
Manfred Clynes (1970, p. 327) asked subjects to press one button if a
red light appeared (in an experimental situation), and to press a different
button if a green light appeared. By. moving these stimulus lights closer and
closer together, he found that approximately 200 msec (0.2 sec) was the time
that must elapse before the subject could change his mind and shift from one
response to another. Clynes (1970a) called this interval a "present moment,"
and wrote:
We observe the physiologic fact that the course of a simple movement is preprogram-
med by the brain before it begins. (There is no time in the 200 msec for feedback to af-
fect the decision.) In all these movements there is a decision which precedes the begin-
ning of the movement. This decision element or event is of precise nature and controls
the subsequent movement according to the program it has set; i.e., it preprograms the
course of the movement. The preprogramming is of the nature of an algorithm. During
approximately 200 msec after a programming decision is made, another programming
decision affecting the same limb or muscle cannot be made. Movements begun under
such a decision cannot be reversed or controlled within this time. (p. 327)
Evidence can be found, then, in both speech and movement, for a

rhythm with a period of 0.2 sec by observing both this interval and its
octave (0.4 sec or 10 frames). We can now return to a consideration of Figs.
3, 4, and 5, and find evidence that the underlying rhythm is actually 0.1 sec
and that the longer intervals are octaves of 0.1 sec.
The two Bushmen (Figs. 3, 4, 5) are each on an observable rhythm
train. The two rhythm trains have the same frequency but are not
synchronous. I have not yet accounted here for this phase-offset between
them. This phenomenon-the two parallel but offset rhythms-was en-
countered regularly in human dialogue. An explanation follows.
In the Bushman storytelling, the first speaker, the storyteller, begins
with a few syllables followed by a brief pause, and the listener begins his in-
termittent interjections at this point. The listener's first speech onset begins
an exact number of tenths of a second after the first speaker's preceding
stress peak. This is illustrated in Fig. 6. It seems obvious, after the fact, that
a listener has no accurate access to the speaker's onsets, but that the
listener's rhythm train is triggered, or entrained, by the immediately preced-
ing stressed sound (which only rarely coincides with the onset of a syllable).
That is, the onset of a syllable burst of one person is triggered, or entrained
by a syllable-part of the other. (A more formal systems statement would be:
154 Paul Byers
a configuration on one level can be "explained" by the arrangement of parts

on a lower level of organization. This, indeed, is the only form of "explana-
tion" in a systems epistemology.)
Figures 6 and 7 illustrate this relationship in other languages and
cultures.
When we find that the stress-peak to onset interval is not always a
multiple of 0.2 sec intervals, but is always a multiple of 0.1 sec intervals, we
can see that the underlying rhythm is actually one with a period of 0.1 sec.
Furthermore, if a syllable (burst) is longer than 0.2 sec, either because
it is uttered with prolongation or because more than the usual content is
compressed into it (e.g., pronouncing "hello" as "h'lo"), the next burst may
begin at the next beat-point on the underlying 10 c/sec (0.1 sec) rhythm
train. This observation also demonstrates that the underlying rhythm has a
frequency of 10 c/sec, even though the minimal programs require a 0.2 sec
time slot.
There are occasions, perhaps in every society, when groups of people
engage in synchronous or phase-related behaviors of one sort or another.
Human communication is, clearly, one of these occasions. There are also
common familiar examples, such as dancing, singing, and marching, and we
can extend the beat-frequency (i.e., levels of organization) in either a faster
or a slower direction. Relatively slow oscillations in social behavior may be
related to such exogenous pacing as the earth's orbital progression around
the sun (annual cycles), or the circadian rhythms paced by the earth's rota-
tion. Higher frequency synchronicity is observable in a group singing a
single note "in tune" at a frequency measured in hundreds of cycles per
second. Chapple (1970) points out that interpersonal synchrony-dissyn-
chrony has biological concomitants which the individual experiences as feel-
ing good (synchrony) or bad (dissynchrony). It seems reasonable to me to
assume that insofar as there are "feelings" (seen either as a biochemical or
a psychological state), this phenomenon would have the significance, in evo-
lutionary terms, of supporting, reinforcing, or leading a creature toward
adaptive relationships and away from maladaptive or lethal relationships.
That is, the creature's intrapersonal or internal balance would reflect an in-
terpersonal or social-ecological balance. Feeling good (biologically or
psychologically) is not only an internal matter, but is a concomitant of be-
ing in a synchronous, or phase-related, relation with the environment, both
physical and social.
In New Guinea, the Maring (and other groups) look forward to travel-
ing a considerable distance to collect together a few times a year to sing and
dance. On such occasions the fact of coming together at a particular time is
one level of synchrony. The fact of singing and dancing "in time" together
is another level. And the fact of singing (or chanting) the same or har-
1:1:1
i:
~.
"
e.
iI:I
~'. ~ , '\
...,=-
~ [
..... .\"\...,.,'\.?...: ~,\ .~ , \'-' \ '"'\ j
, .!.'/'-'-..' II 1
",-_,",-,r' "-"~ ~ ~
~ ."-oN
;-
-.
Fig, 6. Record of the first half-second of the Bushman storytelling. It is "enlarged" by speeding up the paper travel in the recorder. Here, as in the iii
following examples, the secoQd speaker's initial syllable/motor impulse is triggered by the preceding speaker's last stress peak. This accounts for the "g.
phase offset. ..
('")
=-
"
..
!2.
'"5'
('")
Q
a
a
..=
;:j'
~
g'
\ ~ .... - ~ '\,
=- ~
:\' \·\::~\·~-·~/\·\\/·\v~'-~';'~~'''~~/\~-:'--J·'·\:~·~~\ \ ~\'--, .. is'
- - --
M"'-' ~
Fig. 7. Record of dialogue between two Gabra, a camel nomad people in East Africa. Each speaker's initial onsets are on a 10 cps rhythm train
which is initiated or triggered by the preceding speaker's last stress peak. ~
-
156 Paul Byers
monically related note together is a third level of synchrony. The fre-

quencies range from the few times a year of the occasions themselves to the
few hundred cycles per second of the notes sung in unison.
I have found one cultural form, in my very limited search, in which two
people come together and engage in full talk synchrony while using different
semantic content. It is interesting to examine both the temporal structure of
this talk form-one which produces a high degree of multilevel synchrony-
and the social context in which it is found.
This unusual talk form is found among the Yanomamo, who live on
the upper Orinoco in South America. Figure 8 is a chart of the machine-
processed speech. The temporal structure and the mechanics of this talk
form are:
1. The first speaker begins by shouting, typically, three syllable bursts.
The onsets are 0.2 sec apart.
2. The second speaker begins one tenth of a second after the first onset
of the first speaker. He also shouts three syllable bursts. The im-
pressionistic sound is that of two men shouting at each other si-
multaneously.
3. The first speaker begins again (typically) three tenths of a second
after the last stress peak, presumably of the other speaker, although
it is difficult to assign particular peaks to individual speakers in this
"shouting match."
Since the second speaker is triggered, or entrained, by the onset of the
first speaker (or at least they are phase-locked in full synchrony), it appears
that an initial shout (after silence) suffices as the trigger and that by shout-
ing they can achieve full synchrony. I have seen no other instance" in which
such a full phase-locked synchrony occurs and is sustained for minutes at a
time between two interactants without an exogenous pacer.
If we know something of the Yanomamo and of the social circum-
stances in which this talk form occurs, it is possible to make an educated
guess about the function of this talk form.
The standard ethnography of the Yanomamo is titled The Fierce
People (Chagnon, 1968). Chagnon writes, "Yanomamo culture calls forth
aggressive behavior, but at the same time provides a regulated system in
which the expressions of violence can be controlled" (1968, p. 118). I am
told that the talk form described above occurs in two circumstances. One is
when two headmen of different villages come together for a feast, and they
begin the ritual greeting by participating in this shouted talk form. This is
an occasion when a display of fierceness is required, but when violence must
be avoided. And second, I am told, any conversation that escalates to the
point of near-violence can turn into this talk form. And violence is avoided.
...,... ,'" ~ ~ ~ :~""~
§:
=
:v-vt\' '''- co
IIQ
\,.,1~'~''''\' i';'
~
::c
:r
'<
~
~
~
=
S-
Fig. 8. Record showing the same phenomenon in dialogue between two Eskimos.
-a
"t'A:'I:O\lt\\1 004s.D4=''-_ _ __
g-
=
("'J
:r
ill
[
til
,r
S~;u'(1' \ 0,''10.: 1, "'; ;Io,".t.kC"r \ (h~I' '\PC~l.~1 \ On'lC"u
~
~
=
n'
1!'.
is'
:\I\~i '.'~\\\\~ =
~
=
~
is'
.,
~roo.·~ " rt II 0tIt.:1~ "r....Il.er U 0"11.:1\ '\ll,;a.l.~ t lJ o.1!'''·~
Fig. 9. Example from the Yanomamo who live on the upper Orinoco in South America. In this talk form two men are shouting. One begins and
typically shouts two or three syllables. The second person begins almost exactly one-tenth of a second after the first begins and also shouts two or ....
three syllables. The shcmted onset of the first speaker serves to put the two men in almost full phase-locked synchrony. ~
158 Paul Byers
I suggest, then, that behaved animosity has a nonsynchronous con-

comitant, but that close synchronization between interactants creates a
homeostasis between them that precludes violence; this is one observable
expression of what Chagnon calls the "control of violence." I suggest that
this self-imposed synchrony precludes violence, or conflict, through the
regulation of biological processes. I believe that those who examine the
"politics" of a society or a group may overlook the lower-level interactional
and biological processes involved in maintaining "political" systems.
So far in this report, the rhythms of communication movement and
speech have fitted the frequency of what I consider to be an underlying bio-
logical rhythm, or oscillation, with a frequency of 10 c/sec. But there is evi-
dence from Mary Brazier's autocorrelation analysis of brain waves (1960)
that there is also a 7 c/sec rhythm (7.3, c/sec to report her findings
exactly). According to Sollberger's (1965) encyclopedic report on biological
rhythms, this frequency of 7/sec seems to be the upper limit of human
movement frequency, where the units of movement involve feedback. He
cites the work of psychologists who have found that such behaviors as tap-
ping, chewing, etc., cannot be performed faster than 7/sec. It is possible for
pianists or typists to strike more than seven keys per second, of course, just
as a speaker can put more than one linguistic syllable into a single pulse or
motor unit (e.g. "H'lo"), by preprogramming two or more notes or
typewritten letters into a single unit.
I have found a 7 c/sec rhythm in speech traces when the speaker is
performing speech instead of talking with another person. When, for
example, a Maring at a ceremony shouts to his dead ancestors to come to
the ceremony, the underlying rhythm is 7 c/sec. When one Yanomamo
shouts to a group of others to help him sweep the village plaza in prepara-
tion for a feast, the rate is 7 c/sec. But when a villager shouts back and the
two "talk," the rates are then 5-10 c/sec. I have recorded and processed
radio news broadcasts, and found again that this performed speech shows
an underlying 7 c/sec rhythm. When, however, a person anywhere in my
film data is working at a task-chopping, hitting, paddling a canoe, hoeing,
etc.-the rhythm appears to be task-related and is not otherwise predicta-
ble. But even When a person is carrying out a repetitive task at a task-
related rate, and another person comes along and engages the first in
conversation, the shared rhythm invariably shifts to fit the 5-10 c/sec
frequency.
In the course of this research, a friend offered me access to a collection
of sound tapes of stuttering speech and assorted other speech pathologies.
More out of curiosity than an interest in investigating these pathologies, I
ran several of the tapes through the processing apparatus and made the
following observations:
1. In one instance of stuttering-in which initial consonants were
repeated-the speech before and after the stuttered repetitions showed the
usual 5-10 c/sec underlying rhythm, but the onsets of the stuttered repeti-
tions fell clearly on a 7 c/sec train. (This could be measured accurately only
when the repetitions were voiced stops, which have sharp onsets, e.g., b, d,
g.) Furthermore, when the 10 c/sec train from the preceding speech was
projected across or through the stuttered component, the speech following
the stuttering was still falling on the rhythm beats of the original train. That
is, the stuttering episode, at 7 c/sec, had not interrupted the underlying 5-
10 c/sec train. Also, the 7 c/sec stuttered consonant was repeated until a
beat of this 7 c/sec train coincided with the projected 10 c/sec train, at
which point the speech proceeded normally back on the 10 c/sec train until
the next stuttering episode. Those instances of stuttered onsets which were
very close to points on the projected 10 c/sec train were very much greater
amplitude than the others.
2. Other stutterers showed the 5-10 c/sec frequency, or rhythm, un-
derlying both the normal and stuttered components of the speech, but the
stuttered component was on a rhythm train which was out of phase with the
rest of the speech. It appeared on the chart traces that the speaker had two
out-of-phase rhythms underlying his speech, one of which underlay the
normal speech, and another which underlay the stuttering.
3. One example of pathological speech had no stuttered repetitions,
but was, impressionistically, monotonous and expressionless. Chart traces
showed that the onsets of the syllabic bursts showed almost no (expressive)
variation, and were nearly mechanical in their regular 5 c/sec spacing.
VI. SUMMARY AND DISCUSSION
I have reported here on rhythmic regularities found in instrumentally-

processed records of interpersonal movement and speech behavior, and I
have interpreted the findings as evidence of underlying biological rhythms
which pace and integrate the behavioral expression of brain programs. Al-
though two rhythms have been identified, one frequency (10 c/sec) has in-
variably been found in interactions (communication) both between humans
and between a pair of rhesus monkeys.
As one examines progressively larger segmental units of behavior, i.e.,
behavioral packages, or units, on ascending levels of organization, each seg-
mental level can be seen to have its own rhythm. Since, in interaction, each
person is phase related to the other on multiple levels, and since the rhythms
are modulated by and thus reflect the state of the human organism, in-
formation is constantly available to each interactant about the multilevel
states of the other. But because the reference, or comparator, of each
160 Paul Byers
person is his own state-modulated rhythm, his inference (largely out of

awareness) is actually one concerning the relationship between his own and
the other's states. Our present folk epistemology, however, tends to see this
information about the relationship as a "message from the other." The con-
cept of messages and transmission, useful as it is in electronic media, has sup-
ported the popular view of the human species as a collection of individuals
who send and receive messages to and from each other consciously as their
primary or perhaps only form of relating.
On the other hand, if we find that the variations or modulations of un-
derlying rhythms at multiple levels are systematically related to biological
processes, we can envision a human and animal world that is communica-
tionally related through the sharing of time forms on multiple levels of be-
havioral organization.
Although one organism may not directly perceive his phase relation-
ship to another, I believe that, in the case of humans, the phase relationship
is systematically related to the biochemical (endocrinological) states which
are reflected in or perceived as feelings. It is particularly interesting to ob-
serve that the information carried by interpersonal rhythms does not move
directionally from one person to another. This information cannot easily be
conceptualized as "messages," since the information is always si-
multaneously shared and always about the state of the relationship.
While the 10 c/sec rhythm found underlying speech and movement be-
havior is clearly a biological rhythm, these temporal units are the building
blocks for such culturally determined behavior as pauses between speakers
or listener interjections in a speaker's pauses. I have suggested that the
Yanomamo's special speech form shows an unusual degree of interpersonal
synchrony, and that, considering its cultural use, it represents part of the
"regulated system in which the expressions of violence can be controlled"
(Chagnon, 1968). That is, two interactants in a tight synchrony necessarily
and biologically are brought into the same state (of consciousness) by virtue
of their mutual entrainment. Coberly (1972) has examined the sequences, or
processes, involved in shamanistic curing ceremonies in ten cultures, and
has shown that, despite the variety of cultural content, the process is always
the same: the shaman collects a group, synchronizes the group through
dance, movement, song, clapping, chanting, etc., and then brings in the
state-deviant patient who joins-is entrained by-the group and is brought
to their state through synchronous participation. In eastern religions, and in
numerous primitive religions, instrumental means (chanting, breathing,
dancing, singing, etc.) are used to change the state of consciousness of indi-
viduals or to bring a group to the same state of consciousness.
As an anthropologist, I believe that it would be useful to review our
ethnographic data to discover what inventory of states (of some total possi-
ble spectrum) is found in each cultural-ecological context, which states are
encouraged or discouraged by a culture, and which cultural customs

represent instrumental means of implementing or avoiding those states. In
our own society, drugs, meditation, jogging, surfing, some sports, some
kinds of dance and music, all may have significance in implementing,
maintaining, or changing states of consciousness, and consequently the re-
latedness of participants, and all can be examined for the adaptive signifi-
cance in a particular social-environmental context. There is some evidence
that human functioning requires intermittent experience of some spectrum
of states. We may come to see good mental health as a matter of meeting
these state-quota requirements.
In this paper I have created a rather large conceptual map-hopefully
one with systematic integrity-from a relatively small body of observational
data. Most of the process connections are still obscure. We do not know,
for example, how only a few primary rhythms in the brain can be quickly
differentiated to meet the demands of everyday skills. Nor do we under-
stand the mechanism by which motor-expressed brain programs are fitted
to the integrating and pacing rhythms. We do not understand the processes
by means of which states of consciousness (defined by Wolff, 1966, as
semistable organizations of the nervous system) are instrumentally imple-
mented, maintained, or changed.
To put all of this in a still larger framework, it appears to me that we
can see the interrelatedness of life in terms of at least three contrasting
kinds of processes. I have dealt here only with those processes in which
state-modulated rhythms intersect in interaction to make information about
the relationship available to the interactants. This process functions on
multiple levels embracing state information related to cells, organs, organ
systems, organisms (states of consciousness), and even groups or whole
societies. It applies as well to both human and animal life.
We have been, in our world, more explicitly concerned with communi-
cation processes which involve the semantic content of human speech. We
have certainly focused on the evolutionarily unique capabilities inherent in
speech capacity (the capability, for example, of constructing an elaborate
technology and civilization), and have almost ignored the fact that the se-
mantic content of speech lies on a single level of organization and is largely
a product of conscious thinking. The modulations of rhythms are a reflex of
states; the modulated rhythm is an "accurate" representation of states. But
no such simple isomorphism or "accuracy" exists between the state of the
human organism and the organism's verbal messages. If we were,
hypothetically, to quantify the information carried by the modulated
rhythm process vs. the verbal message exchange process, we would readily
conclude that the information carried by modulated rhythms is precise and
preponderantly related to interpersonal relationships, whereas the informa-
tion carried by verbal messages is subject to great imprecision, and is only a
162 Paul Byers
reflex of the individual's perception of his relationship to his environment,

with only a most unwieldy possibility of corrective feedback at the semantic
level.
While much of the focus of the recent research in biofeedback has been
concerned with its orthophysiological possibilities (e.g., learning to relax a
frontalis muscle, control heart rate, etc.), there is the implication in biofeed-
back research that the calibration of biological control systems at multiple
levels is mediated by verbal-symbolic-cognitive processes in man. The two
processes, involving state-modulated rhythms and verbal messages ex-
change-sharing, are linked. The mind and the body are not separate.
There is, I believe, a third process which I would consider more
"primitive" than the two discussed above. There is much popular concern
with, and some scientific exploration into, phenomena associated with the
terms "bioplasma," "psi phenomena," "new energies," "auras," "psychic
phenomena," "extrasensory perception," etc. It is probably too early to
guess whether this line of inquiry will lead to the identification of a new
form of matter-energy ("bioplasma"), the identification of a heretofore un-
recognized paraneurological system, the identification of a new order of
signals and their processing apparatus, or, conceivably, the emergence of a
radically different epistemology which would comfortably embrace
"mystical" observations without requiring much in the way of radical scien-
tific discovery. Whatever the scientific eventuality, the human will move
toward a more explicit concern for his relationship to extraterrestrial
phenomena, i.e., a whole that embraces a larger system of parts than we
now recognize.
And again, whatever the scientific eventuality, the human species is ac-
quiring the conceptual tools with which to move toward a more explicit
concern with relationships than with individual properties, and toward con-
ceptual wholes that will embrace the cosmos. The human notion of human
nature is certain to change radically.
VII. REFERENCES
Bateson, G. (1971). Introduction. The Natural History of an Interview, University of Chicago

Library Microfilm Collection of Manuscripts, in Cultural Anthropology, series 15, Nos.
95-98.
Birdwhistell, R. L. (1970). Kinesics and Context, University of Pennsylvania Press,
Philadelphia.
Bloch, G., and Trager, G. L. (1942). Outline of Linguistic Analysis, Special Publication of the
Linguistic Soc. of America.
Brazier, M. (1960). Long-persisting electrical traces in the brain of man and their possible rela-
tionship to higher nervous activity. In Jasper, H. H., and Smirnov, G. D. (eds.), The
Moscow Colloquium on Electroencephalography of Higher Nervous Activity, EEG J.
Suppl. 13:347-358.
Byers, P. (1971). Sentics, rhythms, and a new view of man. Unpublished paper presented to
AAAS, Philadelphia, December.
Byers, P. (1972). From biological rhythm to cultural pattern: A study of minimal units. Ph.D.
dissertation, Columbia University.
Chagnon, N. A. (1968). Yanomamo: The Fierce People, Holt, Rinehart, and Winston, New
York.
Chapple, E. D. (1939). Quantitative analysis of the interaction of individuals. Proc. Nat. Acad.
Sci. 25:295-307.
Chapple, E. D. (1949). The Interaction Chronograph: Its Evolution and Present Application,
American Management Association, New York.
Chapple, E. D. (1970). Culture and Biological Man: Explorations in Behavioral Anthropology,
Holt, Rinehart and Winston, New York.
Chapple, E. D., and Arensberg, C. (1940). Measuring human relations: An introduction to the
study of the interaction of individuals, Genetic Psychology Monographs 23:3-147.
Chapple, E. D., and Coon, C. S. (1942). Principles of Anthropology, Holt, Rinehart and Win-
ston, New York.
Chapple, E. D. (1953). The standard experimental (stress) interview as used in interaction
chronograph investigations. Hum. Organ. 12:23-32.
Clynes, M. (l970a). Emotion communication in the living moment. Symposium of the American
Association for the Advancement of Science: Biocybernetics of the Dynamic Communica-
tion of Emotions and Qualities.
Clynes, M. (l970b). Toward a view of man. In Clynes, M., and Milsum, J. H. (eds.),
Biomedical Engineering Systems, McGraw-Hill, New York.
Coberly, L. (1972). An Interactional analysis of ten curing ceremonies. M.A. thesis,
Colum1:.;:l University.
Condon, W. c., and Ogston, W. D. (1971). Speech and body motion synchrony of the speaker-
hearer. In Horton, D. L., and Jenkins, J. J. (eds.), The Perception of Language, Charles E.
Merrill Pub. Co., Columbus, Ohio.
Davis, H. (1956). Space and time in the central nervous system. EEG Clin. Neurophysiol.
8:185.
Deriabin, V. S. (1916). Further studies on the physiology of time as a conditional stimulus of the
salivary glands. Dissertation, St. Petersburg (in Russian).
Feokritova, J. P. (1912). Time as a conditional stimulus for the salivary glands, Dissertation,
St. Petersburg (in Russian).
Lashley, K. S. (1951). The problem of serial order in behavior. In Jefress, L. A. (ed.), Cere-
bral Mechanisms in Behavior, Jdhn Wiley and Sons, New York.
Lindsley, D. B. (1961). The reticular activation system and perceptual integration. In Sheer,
D. F. (ed.), Electrical Stimulation of the Brain, University of Texas Press, Austin.
Pribram, K. H. (1971). Languages of the Brain, Prentice-Hall, Englewood Cliffs, N.J.
Sokolov, E. N. (1960). Neuronal models and the orienting reflex. In Brazier, M. A. B. (ed.),
The Central Nervous System and Behavior, Josiah Macy, Jr. Foundation, New York.
Soli berger, A. (1965). Biological Rhythm Research, Elsevier Pub. Co., Amsterdam, London,
New York.
Stetson, R. H. (1903). Rhythm and rhyme. Psychological Rev. Monographs 4:413-466.
Stetson, R. H. (1905). A motor theory of rhythm and discrete succession I and II. Psych.
Rev. 12:250-270, 293-350.
164 Paul Byers
Stetson, R. H. (\951). Motor Phonetics: A Study oj Speech Movements in Action, North-

Holland Pub. Co. (published for Oberlin College), Amsterdam.
Wolff, P. H. (1966). The causes, controls, and organization of behavior in the neonate. Psychol.
Issues, V:I, Monograph 17.
Wolff, P. H. (1968). Stereotypic behavior and development. Canadian Psychologist 4:474-484.
Zelenii, G. P. (1908). Studies on the question of the response of the dog to auditory stimuli.
Dissertation. St. Petersburg. (In Russian.)
Chapter 4
ANIMAL PLAY: PROBLEMS AND

PERSPECTIVES
Marc Bekoff
Department oj Environmental, Population, and Organismic Biology
Ethology Group
University oj Colorado
Boulder, Colorado 80309
When people talk about play, they tend to say what it is not-"it is not real" or "it is
not serious" -and then the rest of the sentence gets rather vague when the speaker
realizes that play is serious (Bateson, 1956, p. 145).
I. ABSTRACT
Social play behavior is receiving increasing attention from ethologists

and other scientists who are interested in social development and social
organization. To date, however, few quantitative data have been amassed
concerning play, and it is not uncommon to read about unsupported conten-
tions being accepted as fact. The purpose of this chapter is to summarize
some aspects of play behavior, and to show that more rigorous quantitative
analyses are needed in order for interested workers to gain a fuller under-
standing of this category of behavior.
II. INTRODUCTION
With respect to social play, a great deal has been written (Loizos, 1966,
1967; Millar, 1968; Berlyne, 1969; MUller-Schwarze, 1971; Bekoff, 1972;
Ellis, 1973; Symons, 1973; Fagen, 1974a, b), but little detailed research has
actually been done. Definitions or characterizations of play are scarce, yet
165
166 Marc Bekoff
Fig. I. A hypothetical relationship between the

amount of play one would expect an individual to
hi engage in and age. Sex differences would not be
unexpected, and constancy of conditions would be
important to maintain. Area A could be considered
to be the "period of socialization." For example, in
Wso c/ullzu,flon many canids there is an increase in the frequency of
J' per/ad"
play at 1-3 months, and this time span is known as
the socialization period, during which intraspecific
(and interspecific) attachments are formed (Scott
and Fuller, 1965). Data could easily be gathered
that would fit to these axes, and the effects of
various regimes of social development on the
10 L -_ _ _ _ _____________
amount of play performed during A could then be
measured, and subsequent effects on social be-
birth
lage havior determined.
most people agree that animals do play both in captivity and in the field,l
that there are seasonal, sex, species, and individual differences in the fre-
quency and form of what are called play activities, that there often are
preferred play partners, and that play usually occurs only after more im-
mediate physiological drives have been satisfied. In a comparative study of
social development in various canids and their hybrids, the frequency of oc-
currence of social play has been used as one "behavioral phenotype" in the
construction of a taxonomic scheme. The taxonomic relationships based on
behavioral characters were in full agreement with taxonomies based on ana-
tomical characters (Bekoff et af., 1975). In addition, it can be
demonstrated, using mathematical models, that play is a " ... unique
optimal adaptive strategy in certain environments," and when such condi-
tions are met, play will be selected and will subsequently evolve (Fagen,
1974b, 1976). Finally, Dissanayake (1974) even has hypothesized that
artistic activities have evolved from play.
But what is play? Our lack of knowledge concerning the defining
characteristics of social play is unfortunate, since careful description of
what is called "play" will certainly shed some light on other aspects of this
behavior, such as its motivation and adaptiveness. Also, play typically is ob-
served in its highest frequency in young animals [a prediction of Fagen's
(1976) model], a time in life during which behavior (and the central nervous
system: Bekoff and Fox, 1972; Barnett, 1973) appears to be most modifiable
(Fig. 1). The high frequency of play during the period when young
1 Play has been described in approximately 45 species of birds, representing 13 orders, and in
140 mammalian species, representing 13 orders, and 35 families (Robert Fagen, personal
communication). One of the most detailed analyses of play in free-ranging animals has been
conducted by Owens (1972, 1975) on Anubis baboons (Papio anubis).
Animal Play: Problems and Perspectives 167
organisms are becoming socialized, and are acquiring the skills necessary to
become "card-carrying" members of their respective species, is probably
more than just mere coincidence.
In this chapter, I shall consider briefly three theories of play that pre-
dominate in the literature, and then discuss some of the ways in which play
may be characterized quantitatively. Some of the suggestions presented
below may appear to be "simplistic" and "obvious"; however, a naive ap-
proach appears to be a good way both to review a field replete with qualita-
tive studies and to suggest ideas for further research. As Schleidt (1973, p.
365) recently has written, regarding a little-studied aspect of animal
communication,
A lack of data invites speculations and theories about how things are supposed to be.
This may be a waste of time, but it could be heuristically valuable if it enables us to
develop a set of hypotheses which can be tested by comparing the predictions with data
from carefully controlled observations and experiments.
III. THEORIES OF PLAY
Anyone theory of play is likely to be incomplete, since play appears to

be so highly variable in both form and function (Gulick, 1920; Beach, 1945).
Of the rr.~ny theories that have been proposed (see Groos, 1898; Winch,
1906; Gulick, 1920; Beach, 1945; Bateson, 1956; Gilmore, 1966; Millar,
1968; Berlyne, 1969; Ellis, 1973; Fagen, 1974a, b) three have had particular
influence on persons interested in the motivation and functions of play.
These are the "surplus energy" theory, the "practice" theory, and the
"arousal" theory, and fall into thejunctionalist camp (Fagen, 1974a).
A. The Surplus Energy Theory
The poet, Friedrich Schiller, suggested that play was the result of an
overflow of excess energy, energy not needed by the organism for perform-
ing activities necessary for survival. Popular support for this theory is not
difficult to find (e.g., Morris, 1969, p. 181). Herbert Spencer made a claim
similar to that of Schiller, and this theory has become known eponymously
as the Schiller-Spencer Surplus Energy Theory. Alexander (1958) also has
written that play is an exercise of surplus libidinal energy, energy not
needed for the grim task of survival! Support for this theory of play stems
from the common observation that infants typically play more than adults,
168 Marc Bekoff
the explanation being that young organisms have more energy to "blow off"
than do adults, and that this energy is released through play.2
Data relevant to the testing of the Schiller-Spencer theory are difficult
to collect. The most obvious way to test this theory is to deprive an animal
of play, and then to look for subsequent increases in the amount of play.
However, the difficulty lies in depriving the animal of play without also de-
priving it of general activity and other types of experience (see
below).
The two published studies of the effects of "play deprivation" are
rather preliminary and have produced contrasting results (not necessarily
unexpected, since different species and small samples were observed). Play
deprivation was brought about by experimenter interference, and whether or
not this method had any confounding effect on the behavior of the indi-
vidual animals is important to assess. MUller-Schwarze (1968) found no sig-
nificant increase in the frequency of the occurrence of actions usually ob-
served during social play in the two black-tailed deer (Odocoileus hemionus
columbianus) he play-deprived, except for an increase in the running speed
of one of the animals. Chepko (1971), on the other hand, reported that play
deprivation in infant Toggenberg goats (Capra hircus) resulted in an
increase in subsequent play activities. The effect of deprivation on social
play is not clear in this study, and her categorizations of play are question-
able. For example, she included grooming under the same heading as the
picking up of inanimate objects. Furthermore, limited sample size and
failure to control for variables that "seem to have a profound effect on play
behavior" (Chepko, 1971, p. 525) make this study preliminary and incon-
clusive. In fact, Chepko (personal communication), in a follow-up study
and controlling for additional factors, subsequently failed to find a statis-
tically significant difference between the amount of play in play-deprived
kids and the amount of play in kids who were not play deprived. In addition
to the fact that energy models are extremely difficult to test, the attributing
of play to an excess of energy does not lead to any further clarification of
the characteristics of play.
B. The Practice Theory

Karl Groos (1898), in a delightful book written at the turn of the
century, suggested that young animals play in order to pre-exercise instincts
in preparation for later life. That is, animals have a childhood in order to
• A very interesting suggestion has recently been put forth by Altman et al. (1973), relating in-
fant-type behavior to hippocampal maturation. These authors suggest that maturation of the
hippocampus transforms "exuberant," "reckless," and "innattentive" juveniles into "placid,"
"cautious," and "observant" adults (but see Nadel et al., 1975, for a sharp criticism of this
hypothesis).
play. This theory has had much appeal for many workers, and Aldis (1975),
after reviewing the liter~ture, still feels that the main function of play is to
prepare an animal for future survival (however, see below).
In order to test Groos's theory, social deprivation procedures are again
called for. The effects of social isolation are much more far-reaching than
the simple removal of the possibility for social play (see Lorenz, 1965;
Ewer, 1968, pp. 309, 323: Harper, 1970, p. 87ff), and as the methods of
deprivation become more drastic and increasingly depressing (both to the
subject and reader) (e.g., Suomi and Harlow, 1969, 1972), the absence of
play (and social experience) becomes a relatively less critical factor.s Post-
deprivation procedures are also very important to consider. The results of
deprivation studies in which animals were socially isolated have made quite
an impact, appearing in a vast array of journals (see Bronfenbrenner, 1968;
Sackett and Ruppenthal, 1973, for reviews), and are probably the reason
why Steiner (1971), in his otherwise excellent study of play in Columbian
ground squirrels (Spermophilus columbian us columbian us), made the state-
ment: "whether play may affect some characteristic of adult behavior in
Columbian ground squirrels remains to be determined in the laboratory, us-
ing the classical deprivation experiments" (p. 259).
There is some evidence from studies on naive or socially deprived
organisms (no attempt was made to deprive the animals solely of play) that
these animals are able to kill prey in species-typical fashion (Wlistehube,
1960) and successfully copulate (Beach, 1968; Harper, 1968) (see Harper,
1970 for a review of the literature). Kittens reared in darkness in the
absence of their mother show "hunting play" when they are first shown a
dummy mouse (Thomas and Schaller, 1954; cited by Tinbergen, 1969). In
addition, Kruuk (1972) observed that a number of adult behavior patterns
were absent from the play repertoire of young spotted hyenas (Crocuta
crocuta), and Schaller (1972) noted that stalking was relatively unimportant
in the play of the lions (Pantera leo massaicus) he observed, but that stalk-
ing was very important in hunting. In four infant coyotes (Canis latrans),
neither (1) the frequency of participation in play, (2) the frequency of partici-
pation in agnostic interactions, nor (3) the frequency of occurrence of motor
patterns from the "prey-killing repertory" (with one exception) in the context
of play or agonistic behavior were significantly correlated with success in
prey-killing. Only the frequency of pouncing (rearing on the hind limbs and
striking at the prey or other object with the fore limbs) in play was signifi-
cantly correlated with success in prey-killing (r = 0.996, p < .05; Vincent
and Bekoff, 1976).
3 Kruijt (1964), in his study of Burmese red junglefowl (Gallus gallus spadiceus), has
demonstrated clearly that an organism cannot be isolated from itself, and that this experience
is important to consider. For a similar suggestion as regards predatory behavior, see Eisen-
berg and Leyhausen (1972, p. 82).
170 Marc Bekoff
So, although there is good evidence that prior play or practice is not
mandatory for the development and successful performance of various skills,
there does remain that possibility that "practice makes more perfect" (e.g.,
in the perfection of predatory techniques: Eisenberg and Leyhausen, 1972;
Leyhausen, 1973; Rasa, 1973), particularly in those species in which preda-
tory or sexual actions, for example, first are observed during play. In these
cases, play experience may increase the "smoothness," or finesse, with which
a particular sequence is carried out. One must also keep in mind that labora-
tory studies probably oversimplify conditions, and that in the wild one must
consider not only whether an animal can kili prey or successfully copulate,
etc., but with what "efficiency" it performs. Rasa (l97Ia) suggests that
practice leads to an increase in corporal agility, and that repetition results in
improved performance. It may be that playing with a conspecific does not
specifically improve, for example, predatory behavior(s), but rather, that
play experience might have an effect on the performance of any number of
mot9r actions, and may help to improve perceptual-motor coordination. In
summary, it is obvious that neither field nor laboratory studies fully support
the practice theory of play. Detailed comparisons between "players" and
"nonplayers" are needed, and further studies will also be helpful in deter-
mining the extent to which phylogenetic differences play a role in the ap-
plication of this theory.
At this point, we may also consider the development of species-typical,
or intraspecific, communicatory skills. Many authors, including me, have
suggested that developing organisms acquire species-typical communicatory
skills during early life, and that social experience is indispensable for the ac-
quisition of such skills (see Harlow, 1969; Bekoff, 1972; Sackett and Rup-
penthal, 1973; Symons, 1974, for reviews). This hypothesis is diagrammed
Fig. 2. A hypothetical relationship

between the amount of an organism's
social experience and its ability to
T
',ocll1l1ztld" interact. Interact in this case refers to
.. -L the animal's ability to send the
"proper" signal when called for, to
react "appropriately" to messages
_-- --- -- -- ---- "unsoclollztld" directed toward it, and to predict
social behavior (Altmann, 1962;
Fedigan, 1972; also, see OUe, 1974,
pp. 396 ff and Fig. 8, for a discussion
of the hypothetical relationship
between experience, the plausibility of
fage deceit, and the ability to "misin-
(fexperience) form").
in Figure 2. Social deprivation procedures are usually used to test this

hypothesis, and there is little doubt that socially-isolated animals often are
deficient in social skills (e.g., Miller et al., 1967), the deficiency apparently
being some function of the seriousness of the deprivation, or of the "com-
plexity" (Sackett and Ruppenthal, 1973) of the environment in which the
organism is reared.
There is only one reported study, to my knowledge, providing a
detailed analysis of the relationship between social play experience and the
acquisition of a specific communicatory skill, that does not use deprivation
procedures. Symons (1974) has asked the question whether or not rhesus
monkeys learn, refine, or practice agonistic communicatory skills in ag-
gressive play. Based on detailed film analyses, he concluded that " ... the
context of aggressive play does not seem to provide the necessary opportu-
nities or contingencies for the learning, practicing, or refining, in the or-
dinary sense of the words, of agonistic signals." His is only one study, and
much more work is needed concerning whether or not complex communica-
tion patterns are developed during social play" Perhaps, as suggested by
Baldwin and Baldwin (1974), a modicum of social behavior can develop
without social play experience, play experience resulting in an increased
flexibility of the behavioral repertoire and an increase in the ability of a
organism to respond to subtle social stimuli. Fedigan (1972, p. 361) suggests
that in vervet monkeys play might be important for the development of
"social perception," " ... the ability to predict other individuals' behavior
and to act appropriately ... " Gilbert (1968, p. 875), studying social
development in fallow deer (Dam a dama) , noted that "the most striking
change demonstraied by hand-rearing is the irreversible cleavage of the
social bonds between the fawn and its own kind. The elements which are
essential to herd organization (the cohesion of individuals, the awareness oj
slight changes in mood by exchange oj signals) have disintegrated" (my em-
phasis). In changing social environments, flexibility of the behavioral
repertoire would be favored, and the time and energy "cost" of playing
might have significant "pay-off." However, until appropriate studies are
conducted, Schaller's (1972, p. 165) claim, that it is by no means clear that
• One must also consider whether social play, or simply other types of contact and/or exposure
are the critical variables (Dolhinow and Bishop, 1970). There is some evidence that some type
of contact is necessary (Sackett, 1973), although brief periods of contact during a particular
time of life appear to be adequate for "normal" social development in some mammals.
Observational learning may also be important (e.g., Adler, 1955; John et 01., 1968). In a
recent study of an intact litter of coyotes (Bekoff and Jamieson, 1975), we observed that the
lowest-ranking female was virtually noninteractive during the first three months of life.
However, when she did interact both as an infant and adult, she showed absolutely no
noticeable deficiency in social communication.
172 Marc Bekoff
information derived from play actually benefits the animal, must be kept in
mind.
C. The "Arousal" Theory of Play
A more modern theory of play seeks to explain playas being related to

"arousal," and is based on the conceptual framework provided by Hebb
(1955) and Berlyne (1960). Basically, play is viewed as stimulus-seeking be-
havior "that is motivated by the need to elevate the level of arousal towards
the optimum" (Ellis, 1973, p. 110). Ellis (1973, p. 109) also writes that " ...
play is an artificial category. It is merely that part of stimulus-seeking be-
havior to which we cannot ascribe a prepotent motive." First of all, it
should be stressed that, in some social situations, play may also lead to a re-
duction in "arousal." For example, when high-ranking individuals "self-
handicap" themselves during the solicitation of play from their subor-
dinates, or during a play bout (e.g., Altmann, 1962), the overall effect may
be a reduction in "arousal." The same effect may occur during play
between adult stumptail macaques (Macaca arctoides) (Rhine, 1973).
The major shortcoming of this point of view is that the term "arousal"
represents a hypothetical construct that may be used in a myriad of ways,
and for which evidence is usually indirect (but see Rasa, 1971b). Ellis recog-
nizes the vagueness of the term (1973, p. 111), and it is possible that the
very nature of the concept makes it easily applied to play.5
There are many other functionalist theories in the literature. For
example, I have not considered the possible role of play in the development
and maintenance of social relationships, a function that has been suggested
in a plethora of studies. It is obvious that theorizing about play is getting us
nowhere, especially at the motivational level, mainly because we have as yet
only a slight appreciation of the phenomenon.
5 Mason (1973, p. 21) recently has offered the following definition of arousal: "Arousal is
considered to be a hypothetical physiological variable that is correlated with overt behavior.
It is assumed that a given level of arousal establishes a predisposition to engage in certain pat-
terns of behavior rather than serving as a precipitating cause. Specific behaviors are jointly
determined by arousal level and by presence of specific objects. Level of arousal cannot (my
emphasis) be equated with level of overt behavior inasmuch as such activities as crouching,
clinging, freezing, screaming, and frantic running are all expressions of high arousal in spite
of gross differences in motor output." There also are marked species differences in the be-
havioral correlates of arousal (Pfaffmann, 1973). The reader is referred to Glickman and
Schiff's (1967) paper, in which reinforcement theory is offered as an alternative to the arousal
paradigm. See also Glickman's (1973, p. 222, ff) reconsideration of his earlier paper. Andrew
(1974) recently has written an excellent, long overdue, critical summary concerning the use of
the word "arousal" to explain the causation of behavior.
IV. THE CHARACTERIZATION OF SOCIAL PLAY
A. How Has Play Been Characterized?
There appear to be few motor patterns that are specific to play, one
possible exception being play-soliciting actions (see below). Sade (1973, p.
541) has observed a postural component (rotation of the head and torso in
the tranverse plane) in rhesus monkeys (Macaca mulatta) that seems to be
unique to play, and has proposed that " ... play should be considered a
mood in its own right and not simply derived from other aspects of be-
havior." However, play has most often been characterized by reference to
the ways in which motor actions usually observed in other contexts-such
as aggression, sex, or predation-may be altered when transferred to a play·
ful context; this is the structuralist approach (Fagen, 1974a). Some sug-
gested and often-quoted characteristics are listed (after Loizos, 1966, p. 7, in
slightly different order): I. The sequence may be reordered. 2. The sequence
may be broken off altogether by the introduction of irrelevant activitIes, and
resumed later. This may be called fragmentation. 3. The individual move-
ments making up the sequence may become exaggerated. 4. Certain move-
ments within the sequence may be repeated more than they would usually be.
5. Movements may be exaggerated and repeated. 6. Individual movements
within the sequence may never be completed, and this incomplete element
may be repeated many times. This applies equally to both the beginning of a
movement (the intention element) and to its ending (the completion element).
Unfortunately, there have been very few rigorous quantitative tests of Loizos'
ideas.
B. The Analysis of Temporal Sequences
One way to begin to describe play is by analyzing the temporal se-

quencing of motor action patterns. 6 If it is accepted, for practical purposes,
that a given organism has a finite behavioral repertoire, and that the actions
making up its repertoire are combined in a finite number of permutations,
then the way in which these actions are linked during agonistic, sexual, or
• For suggestions and representative studies on a wide range of organisms, see Cane (1961),
Wiepkema (1961), Nelson (1964, 1973), S. A. Altmann (1965), Bobbitt et al. (1969), Delius
(1969), Hinde and Stevenson (1969), Slater and Ollason (1972), D. Altmann (1973), Golani
(1973), Hailman and Sustare (1973), Maurus and Pruscha (1973), Andersson (1974), Maurus
et al. (1974), Metz (1974), Bekoff (1976a), and MUller-Schwarze (1975). Also, see Hazlett and
Bossert (1965), Dingle (1972), and Steinberg and Conant (1974), for application of informa-
tion theory to the analysis of behavior sequences, and Poole (1973), for suggestions on doing
film and computer analyses of rapid behavior sequences.
174 Marc Bekoff
predatory behavior may be analyzed. After establishing what the typical se-
quences are in these nonplay situations, one may begin to think about com-
paring these sequences with those observed during social play.
Struhsaker (1967, p. 33), studying vervet monkeys (Cercopithecus ae-
thiops), wrote that "although the data necessary for a strong argument are
not available, the distinct impression was gained that the temporal and
sequential patterning of play behavior were very different from nonplay be-
havior ... " MUller-Schwarze (1975) has data that clearly demonstrate that
the temporal sequences of play in juvenile black-tailed deer differ signifi-
cantly from temporal sequences observed during nonplay situations; the
same is true in various infant canids (Bekoff, 1976a).
Little study has been devoted to whether or not the sequences observed
during play are less predictable than those observed during nonplay. Loizos
(1967, p. 231) has suggested that" ... play has no formalized sequence of
events ... "
Markov models and measures of conditional uncertainty are two ways
by which behavioral diversity may be analyzed (Bekoff, 1975a). If play se-
quences are more random than nonplay sequences, Markov models of play
should be of a lower order than Markov models of nonplay. Likewise, using
information-theoretic analysis, we may ask whether or not there will be sig-
nificant differences in the uncertainty with which a subsequent act (or acts)
can be predicted-knowing that a given act had been just performed-when
comparing play with nonplay. That is, we can ask whether
(HBIA play =) - L PI} log2 Pj/i >-L p' ij log2 p' jli (= HBIA nonplay)
where H BIA is the conditional uncertainty; i and j are actions performed by

animals A and B respectively; Pij and p' ij are the joint probabilities of the
two act sequence (i. j) for play and nonplay respectively; and Pj/i and P' iii
are the conditional probabilities for the occurrence of act j, given that act i
has just taken place, for play and nonplay respectively. This analysis can be
extended to include those instances in which it has been demonstrated that
it is necessary to consider more than just one prior act (e.g., a Markov
model of greater than first-order).
No published papers dealing with play have used the above methods. In
fact, problems in meeting some of the underlying assumptions (e.g., sta-
tionarity) of the above mathematical procedures may, in fact, preclude their
use on many behavioral problems, and these procedures often have been
abused (Chatfield, 1973; Slater, 1973; Baylis, 1975; Bekoff, 1975b, 1976b).
[Alternative methods for the analysis of behavior sequences are suggested in
Bekoff (1976b.)] Using other methods to analyze behavioral sequences (z
scores), Poole and Fish (1975) found that play in young rats does have
definite sequences, but they are not the same sequences as those observed in
adults. MUller-Schwarze's (1975) data (analyzed by information-theoretic

analyses) indicate that juvenile play is more random than juvenile nonplay
behavior, and that both juvenile play and nonplay are more loosely struc-
tured than adult nonplay behavior; in one study on infant canids, play se-
quences were found to be less predictable than sequences performed during
agonis1ic interactions (Bekoff, unpublished data).
There is, therefore, evidence that there is some reordering of actions
during play,7 but there are not yet sufficient data concerning whether or not
play sequences are more (or less) random than those sequences observed dur-
ing nonplay behavior to tell whether this is a general characteristic of play.
C. Are Actions Exaggerated When Performed During Play?
There are few supporting or contradictory data relevant to this sug-

gested characteristic of play. Overall, gross movements performed during
play appear to be performed in an "exaggerated" fashion, as are play-so-
liciting actions (Bekoff, 1972, 1974). Wilson and Kleiman (1974), in their
study of play in six unrelated and diverse mammals, labelled a group of be-
havior patterns as locomotor rotational movements, and these were
considered to be play " ... because they are exaggerated forms of normal
locomotor and rotational body movements found in other functional
contexts ... " They operationally defined exaggeration in terms of the ap-
parent amplitude of the movement and the rapidity of execution. Quantita-
tive data (for six motor patterns) on exaggeration (in this case duration) are
provided by Henry and Herrero (l974) for black bear cubs (Ursus ameri-
canus). For only one of the six actions was the duration greater in social
play than during nonplay situations, for two actions the duration was less,
and for three actions there was no significant difference. They correctly
concluded that "motor patterns are not performed in an exaggerated or
prolonged manner in social play when compared to the same motor patterns
in adult nonplayful behavior." In a study of social development in New
England canids or eastern coyotes (c. latrans var.) (Hill and Bekoff, 1976),
the duration of three acts was measured both in play and agonistic contexts.
For two of the three acts the duration was less in play and for one act there
7 Some suggestions as to why there is reordering of actions during play have been provided by
Morris (1956) and Leyhausen (1965, 1973). Morris (1956, p. 643) suggested that play involves
the switching off of some type of mechanism, and as a result" ... there is no control over the
types and sequences of motor patterns in the usual sense ... " Leyhausen (1965, 1973) feels
that the ordering of various actions is controlled by whether or not the motivation for the end
action is in control (see Ewer, 1968, p. 291). When motivation for the end action is absent,
ordering of the individual actions breaks down, and the temporal sequencing of actions
changes.
176 Marc Bekoff
Table I. The Duration of Three Acts (seconds) in Play and Agonistic Contexts in
New England Canids(Eastern Coyotes)
Action" Play Agonistic Significance levela
Body bite X = 2.21 4.64 .001

Stand-over X = 3.36 6.10 .01
Scruff bite X = 3.17 4.98 NSc
a Data were analyzed using t-tests.

" See Bekoff (1974) for descriptions of the actions.
C Not significant (p > .05).
was no significant difference (Table 1). Whether or not sounds or odors are
exaggerated in some manner during play remains to be studied. However, due
to the apparent great variability in the performance of individual actions,
even in nonplay situations, a very large number of observations would be
needed in order to decide whether actions performed during play are different
from or the same as equivalent actions performed in nonplay situations (T. B.
Poole, personal communication). A similar argument above could be made
concerning the incompleteness of actions (point 6).
There are few data available that refute or support the other suggested
characteristics. The fragmentation of sequences by the performance of "ir-
relevant" acts (point 2, p. 173) does not seem unique to play [e.g., displace-
ment activities observed during courtship and aggression are frequently (and
pta haps erroneously) also termed "irrelevant"]. And, in the absence of
detailed analyses, how can one say that any behavior is either "irrelevant" or
"relevant" !
D. Play and Anthropomorphism: Can Play Be Characterized as Being

Pleasurable?
One characteristic of play often presented is that it seems to be a

pleasurable activity (e.g., Bertrand, 1969; Csikszentmihalyi and Bennett,
8 There is always a temptation to anthropomorphize the behavior of nonhuman animals (or
zoomorphize the behavior of man) (see Callan, 1970). Discussions of play are often replete
with anthropomorphic interpretations (e.g., play is "fun," or "pleasurable"). However, a
large number of explanations of other categories of behavior also contain anthropomorphic
thought (e.g., discussions of "shy" animals, "depressed" animals, "angry" animals, "emo-
tional" animals, and of castes, queens, workers, or soldiers; the adjectival terminology runs
the full gamut; see for example, Buirski et al., 1973.) It is questionable whether or not there is
any necessity to anthropomorphize a nonhuman animal's experience (see Huxley, 1962, cited
by Callan, 1970, p. 46), but as Hebb (1946, p. 88) has written, " ... a rigid refusal to
'anthropomorphize' may have its scientific disadvantages ... a complete rejection of all con-
cepts derived from experience with man would leave a vacuum ... " See also Lorenz (1974),
for a defense of anthropomorphism.
1971; Bekoff, 1974; Csikszentmihalyi, 1975; Reynolds, 1975) or is em-

phatically gay (Meyer-Holzapfel, cited by Tembrock, 1958).8 It has been
demonstrated that play is a preferred activity, and has a rewarding effect
(see Mason, 1967) on an animal. Leyhausen (1973, p. 235) has written that
"anything with such strong overtones of pleasure naturally develops power-
ful appetences which cause the animal actively to seek out or manufacture
play situations." Finally, Geist (1971, p. 253) has noted that, in play, indi-
viduals repeat what was found to be satisfying. Bertrand (1969, p. 59ff)
stressed that it would be premature, and scientifically unfruitful, to reject
the hypothesis that play is self-rewarding or pleasurable, simply because
necessary techniques for studying this have not been yet developed. She
goes on to write:
From a mere consideration of probabilities and of the evolutionary viewpoint, is it wise
to consider that man is the only animal, and in particular, the only primate, who has the
faculty to enjoy an activity for its own sake? One of the best evolutionary tricks to
ensure that a given behavior pattern becomes part of the behavioral repertoire of a
mammal may be that it affords pleasure to its performers. Sexual behavior, preferences
for certain food types, comfort activities, etc. cannot be considered only from the point
of view of the future of the species. Given two behavior patterns having the same selec-
tive value, the one which will have the best chance to become part of the repertoire of
the species will quite likely be the one which is most pleasurable to the individual. (pp.
59-60)
Suffice it to say, young organisms (and on certain occasions adults)
voluntarily spend a lot of time and (presumably) energy both searching out
playmates and engaging in social play.9 There are "pleasure centers" in the
brain associated with other behaviors (e.g., Olds, 1956), and many
vertebrates seem to experience the emotion of pleasure (Maier and Maier,
1970, p. 336). It appears that animals do "enjoy" playing, although cur-
rently this would be a most difficult statement to prove unequivocably.
Neurophysiological studies and accurate monitoring of the autonomic
nervous system (Young, 1973) may provide some clue.
V. THE COMMUNICATION OF PLAY INTENTION
In this section I shall discuss how animals communicate their intent to

engage in social play (e.g., metacommunication, after Bateson, 1955, 1956)
and make some suggestions for future research.
Since actions from various contexts are performed during play, it is
important that the interactants be able to communicate their intent to
9 The energy costs of various activities can be calculated (see Schmidt-Nielsen, I 972a,b;
Yousef, et ai., 1973, for examples), and energy cost data for play would provide very interest-
ing information.
178 Marc Bekoff
engage in play. This typically is accomplished through the use of certain

signals that carry the message, "what follows is play" (Figure 3). This form
of precommunication has been termed metacommunication (Bateson, 1955,
1956; see Bekoff, 1972, for review). Play signals have been described for a
wide range of animals (especially mammals) and appear to be "prescrip-
tors" (Marler, 1961, p. 301) in that they convey requirements for specific
responses, and in that they appear to " ... dispose the receiver to perform
certain response sequences rather than others." In effect, play signals alter
the "meaning," or significance, of subsequent behaviors (Altmann, 1967).
For example, in canids, predatory-like shaking of the head from side to side
while delivering a bite, aggressive-like baring of the teeth, and/or sexual-
like mounting, appear to take on a different "meaning" when preceded by
a play-soliciting signal.
Play signals are observed both before and during play bouts. In canids,
it appears that play signals precede play bouts when the likelihood of misin-
terpreting another behavior is high, and when fighting would otherwise be
likely to occur (Bekoff, 1974). During a play bout, these signals probably
serve as assurances that "this is still play." Synchrony between the players
is thus made more likely. Very often these signals are very subtle and fleet-
ing (at least to the human observer).
\()'i"...
",\'/~ ATMOSPHERE"
o~O.f.( \ """not shored

""'~\ / shored ~
'f" '" "'NOPLAY
PLAY SIGNAL--"'REC/P/ENT~PL A Y
(''R")
(SENDER)
("S ")
S (r) R. ..
R~S. ..
Fig. 3. A diagrammatic representation of the process of communicating
play intention (see text for explanation). Play signals are also observed during
play. The bottom figure shows how the roles of sender (S) and receiver (R)
may alternate during the course of an interaction. The DNA analogy is from
Goyer (1970).
Although visual play signals are most often studied, vocalizations (e.g.,
Loizos, 1967; Winter, 1968; Steiner, 1971; Jolly, 1972), odors (e.g., Wilson,
1973; Wilson and Kleiman, 1974), and probably touch, are also important
to consider.lO Bleicher (1963), doing spectrographic analyses of vocaliza-
tions in domestic dogs, found no differences in growls recorded during play-
fighting and real aggression, and suggested that facial expressions and
postures are more important in differentiating play from nonplay. It would
also be interesting to study, for example, if and how the "play-pant," audi-
ble during the play of various mammals, differs from the "exertion-pant"
(Aldis, personal communication). In the African dwarf mongoose (Helogale
undulata), there is a "play vocalization" in both social and object play that
starts as soon as the animal starts playing, and continues throughout play-
ing (A. Rasa, personal communication). It is distinct from aQY other vocal-
ization, and definitely helps to establish the "play-mood." An animal giving
the play call can dash at another and bite it on the neck without being at-
tacked in return. The same bite without a play call would result in a fight.
The play call appears in ontogeny as soon as recognizable play behavior
emerges (approximately 22 days of age). Wilson (1973) has demonstrated
that a scent is important in stimulating play in voles (Microtus agrestis),
and that differences in the amount of play shown by young born in different
seasons (autumn vs. spring) can be traced to the fact that autumn-born
young do not produce the play-stimulating scent.
All of the above examples provide evidence that there is a temporal
correlation between the appearance of a particular action and subsequent
play, but a temporal correlation is not sufficient datum for demonstrating
that such signals alter the "meaning" or significance of subsequent be-
havior(s). Therefore, the acceptance of a "metacommunicative" function
for what have been called "play signals" would be premature. If certain
signals do function to alter the signifiance of a subsequent act or acts, one
should be able to do a simple analysis to demonstrate that there are dif-
ferences between the probability distribution of subsequent responses to
specific acts when there is a preceding "play signal" and when there is not.
In a study of play between two infant coyotes (c. latrans), the follow-
ing specific question was asked: Does the performance of a "play signal"
alter the significance of a subsequent offensive threat, so that the response
of the recipient of the threat is different from the response that would typi-
,. There has been little work done using information theory to analyze mammalian vocaliza-
tions. However, I. Hirsch (1971), analyzing vocalizations in dogs, found that the rate of in-
formation transfer was less during play (2.5 bits/sec) than during agonistic interactions (3.92
and 4.54 bits/sec), and that, during play, there was phonic redundance (G. Tembrock, per-
sonal communication). Dingle (1972, p. 139) feels that there may be some limitations in the
application of information theory to metacommunication.
180 Marc Bekoff
cally follow a threat in the absence of a prior "play signal"? (See Bekoff,
1976c, for details.) A detailed analysis of 92 interactions demonstrated, at
least in this pair of coyotes, that the probability of submission following a
threat preceded by a "play signal" was significantly lower than if the threat
had not been preceded by such a signal. Therefore, it was demonstrated that
the "play signal" somehow altered the significance of a subsequent behavior.
This is the only quantitatively supported example of "metacommunication"
that I could find in the literature; however, those persons doing sequential
analyses on their own play data should also be able to perform similar
analyses.
It is obvious that many aspects of the communication of play intention
require further detailed study. Some questions that may be considered in the
planning of further research include:
1. Are play-soliciting actions specific to play? Does the particular

species (or individual) use primarily visual, auditory, or olfactory
signals? How are various signals combined? Careful description,
movie film, sonographic, and scent analyses are necessary.
2. If there are distinguishable play-soliciting actions, are any of them
more successful in initiating play than others? Is there any correla-
tion between their success and frequency of use? In infant coyotes,
certain signals are more successful than others in the initiation of
play, and they are used more frequently (Bekoff, 1974). This rela-
tionship between frequency and success appears to be closely linked
with the early ontogeny of aggression in this species. The relation-
ship between success and frequency of use does not hold for infant
beagles (c. jamiliaris), or for wolves (c. lupus), or for the groups of
free-roaming domestic dogs that we have been studying for the past
four years.
3. What is the relative (temporal) position of these signals within a
bout: only at the beginning of the bout, or interspersed throughout?
Is there any regularity in their place of appearance? That is, do play
signals occur more frequently following particular nonplay actions
(such as baring of the teeth) (Figure 4)? Simonds (1974, p. 198) sug-
gests. that there are two sets of play signals, one to initiate play, and
another to reinforce continually that the ongoing activity is play. He
further suggests (without an example) that " ... usually the initiating
signals are repeated whenever a clear statement of purpose is
necessary." This is not the case in infant canids.
4. What is the probability that if play occurs, it was previously so-
licited? Which individual or individuals do the most soliciting?
Animal Play: Problems and Perspectives lSI
Fig. 4. This figure shows some possibilities for the appearance a) P2141323 .. ,
(positioning) of play signals (P), either before or during play bouts.
(P = NO observable play signal; 1,2,3, and 4 = various actions) (a) p\- - -
This represents a play bout that has been solicited previously and b) /214P132P3 ...
during which no additional play signals are observed. The play P
signal appears to have a long term (tonic? see Schleidt, 1973) effect.
(b) In this case, P occurs regularly after every three actions. There p\-- - -
might also be a temporal factor, P occurring every n seconds. (c) P c) ,..21P41P323 ...
occurs after a particular action (in this case, I) is performed. It is j5
also possible that P may regularly precede a particular action. (d) P
occurs randomly throughout a bout. In none of the above cases am I P
\-~'~
implying that play sequences are either random or predictable. For d) ,..2P141P323 ...
infant coyotes, beagles, and wolves, cases (a) and (d) apply. j5
The situation becomes more complex when considering two (or more) individuals.
Consider a dyadic interaction between animals A and B. A bites B's face and then A sends the
message, "That was play, so let's keep on playing." It is also possible that B might signal, "I
think that was in play, so if you want to continue playing, then so do I." All interact ants must
be accounted for.
Answers to these questions would shed some light on how important

it is for an animal to indicate, before a bout that it wants to engage
in play. In one study of pairs of infant coyotes, it was determined
that 90% of all play bouts had been previously solicited, and that the
subordinate animal did most of the soliciting (Bekoff, 1974).
5. Are any individuals in a group (litter) more or less successful in so-
liciting play? We (Bekoff and Jamieson, 1975) have found a signifi-
cant relationship between the rank of an individual coyote and its
success in soliciting play from littermates; the more dominant indi-
viduals are the least successful (see also Bekoff, 1974).
6. Are there more occurrences of play-soliciting signals during an
initially unsolicited bout than during one that previously had been
solicited? In other words, if two animals begin to play without first
indicating "this is play," do they have to be more "convincing" of
their intent as the bout proceeds? Data collected on infant coyotes,
beagles, and wolves indicate the following trend. Play signals are not
observed in increased frequency during unsolicited play bouts, but
unsolicited bouts tend to be shorter in length (duration and number
of actions) than previously solicited boutsY
7. At what age do play-soliciting actions first occur? And in what order?
(see Tembrock, 1958; MUller-Schwarze, 1968; Bekoff, 1974).
11 There are also some differences, as yet unexplainable, in the sequencing of actions during un-
solicited bouts.
182 Marc Bekoff
I am certain that other questions may be asked in addition to the above

suggestions. Answers to these will provide a start in gaining a better
understanding of the communication of play.
VI. CONCLUSION
A lengthy conclusion is not necessary. There are enough "hard qualita-

tive" data on a limited variety of mammals, and quantitative studies of
social play on a wider range of organisms are sorely needed. I have re-
viewed some suggested characteristics of play, and have also offered sugges-
tions for further research. As West (1974) has stated, some researchers have
concluded before sufficient data are in that play is not a useful classs of be-
havior. Berlyne's (1969, p. 843) suggestion that the category of play should
be given up " ... in favor of both wider and narrower categories of be-
havior" is not helpful, since wider categories certainly will be more difficult
to handle adequately.
The kaleidoscopic nature of the concept of play has resulted in con-
troversy, and refutation, instead of leading to further research (Popper,
1965), seems, for some, to have halted progress. It appears to be a very
simple matter: to date, social play-or whatever it is called-represents a
category of behavior that has not yet been studied sufficiently. It deserves
b-:-tter.
VII. ACKNOWLEDGMENTS
I would like to thank the following people (listed alphabetically) for

their comments on an earlier draft of this paper, or on particular sections:
Anne C. Bekoff, R. F. Ewer, Dave Henry, Buford Holt, Robert Jamieson,
Devra G. Kleiman, Wayne Lazar, Manfred Maurus, Dietland MUller-
Schwarze, Anne Rasa, Peter C. Reynolds, Robert and Linda Sussman,
Chris Wemmer, and Susan Wilson. Robert Cadigan and Jerry Corcoran
were of great help in the development of the necessary computer programs.
I would also like to thank Diane Chepko-Sade, Robert Fagen, Dietland
MUller-Schwarze, Trevor Poole, Anne Rasa, and Gunther Tembrock for
providing me with some of their unpublished data. The editorial assistance
of Patrick Bateson and Peter Klopfer was very helpful. The advice of the
above people was not always followed, and I remain responsible for the
final form of this chapter.
VIII. REFERENCES
Adler, H. E. (1955). Some factors of observation learning in cats. J. Genet. Psych. 86:159-177.
Aldis, O. (1975). Play Fighting, Academic Press, New York.
Alexander, F. (1958). A contribution to the theory of play, Psychoanal. Quart. 28:175-193.
Altman, J., Brunner, R. L., and Bayer, S. A. (1973). The hippocampus and behavioral matura-
tion. Behav. Bioi. 8:557-596.
Altmann, D. (1973). Modell einer diadishen Sequenzanalyse des Verhaltensinventars bei Muf-
flons. Forma et Functio 6:213-246.
Altmann, S. A. (1962). A field study of the sociobiology of rhesus monkeys, Macaca mulatta.
Ann. N. Y. Arad. Sci. 102 (article 2) 338-435.
Altmann, S. A. (1965). Sociobiology of rhesus monkeys. II. Stochastics of social communica-
tion. J. Theor. Bioi. 8:557-596.
Altmann, S. A. (1967). The structure of primate social communication. In Altmann, S. A.
(ed.), Social Communication Among Primates, University of Chicago Press, Chicago.
Andersson, M. (1974). Temporal graphic analysis of behavior sequences. Behaviour 51:38-48.
Andrew, R. J. (1974). Arousal and the causation of behavior. Behaviour 51:135-163.
Baldwin, J. D., and Baldwin, J. I. (1974). Exploration and social play in squirrel monkeys
Saimiri. Amer. Zool. 14:303-315.
Barnett, S. A. (ed) (1973). Ethology and Development, Clinics in Developmental Medicine No.
47, Spastics International Medical Publications, J. B. Lippincott Co., Philadelphia (W.
Heinemann Medical Books Ltd., London).
Bateson, G. (1955). A theory of play and fantasy. Psychiat. Res. Rep. A. 2:39-51.
Bateson, G. (1956). The message "this is play." In Schaffner, B. (ed.), Group Processes, Josiah
Macy, Jr. Foundation, New York.
Baylis, J. (1975). A quantitative analysis of long-term courtship in two sympatric species of
cichlid fish. Ph.D. dissertation, University of California, Berkeley.
Beach, F. A. (1945). Current concepts of play. Amer. Nat. 79:523-541.
Beach, F. A. (1968). Coital behavior in dogs. III. Effects of early isolation on mating in males.
Behaviour 30:217-238.
Bekoff, M. (1972). The development of social interaction, play, and metacommunication in
mammals: An ethological perspective. Quart. Rev. Bio. 47:412-434.
Bekoff, M. (1974). Social play and play-soliciting by infant canids. Amer. Zool. 14:323-340.
Bekoff, M. (l975a). Animal play and behavioral diversity. Amer. Nat. 109:601-603.
Bekoff, M. (1975b). Uses and abuses of stochastic processes in analyses of social behavior.
Paper presented at an international workshop on social interaction processes, University
of North Carolina, Chapel Hill.
Bekoff, M. (1976a). A sequence analysis of social interaction in infant canids: Social play and
aggression. In preparation.
Bekoff, M. (1976b). Statistical analyses of behavior sequences: Uses and abuses. Submitted for
publication.
Bekoff, M. (1976c). The communication of play intention: Are play signals functional? Semio-
tica (in press).
Bekoff, M., and Fox, M. W. (1972). Postnatal neural ontogeny: Environment-dependent and/or
environment-expectant? Dev. Psychobiol. 5:323-341.
Bekoff, M., Hill, H. L., and Mitton, J. B. (1975). Behavioral taxonomy in canids by discriminant
function analyses. Science 190:1223-1225.
184 Marc Bekoff
Bekoff, M., and Jamieson, R. A. (1975). Social behavior in coyotes during the first year of life.
In preparation.
Berlyne, D. E. (1960). Conjlict, Arousal, and Curiosity. McGraw-Hill, New York.
Berlyne, D. E. (1969). Laughter, humor, and play. In Lindzey, G., and Aronson, E. (eds.), The
Handbook oj Social Psychology, Vol. 3, Addison-Wesley, Reading, Mass.
Bertrand, M. (1969). The behavioral repertoire of the stumptail macaque. Biblio. Primat. No.
11.
Bleicher, N. (1963). Physical and behavioral analysis of dog vocalizations. A mer. J. Vet. Res.
24:415-427.
Bobbitt, R. A., Gourevitch, V. P., Miller, L. E., and Jensen, G. D. (1969). Dynamics of social
interactive behavior: A computerized procedure for analysing trends, patterns, and se-
quences. Psych. Bull. 71:110-121.
Bronfenbrenner, U. (1968). Early deprivation in mammals: A cross-species analysis. In
Newton, G., and Levine, S. (eds.), Early Experience and Behavior, C. C. Thomas, Spring-
fieid, Ill.
Buirski, P., Kellerman, H., Plutchik, R., and Weininger, R. (1973). A field study of emotions,
dominance, and social behavior in a group of baboons Papio anubis. Primates 14:67-78.
Callan, H. A. (1970). Ethology and Society, Oxford University Press, New York.
Cane, V. (1961). Some ways of describing behavior. In Thorpe, W. H., and Zangwill, O. L.
(eds.) Current Problems in Animal Behavior, Cambridge University Press, New York.
Chatfield, C. (1973). Statistical inference regarding Markov chain models. Appl. Stat. 22:7-
20.
Chatfield, c., and Lemon, R. E. (1970). Analysing sequences of behavioural events. J. Theor.
BioI. 29:427-445.
Chepko, B. D. (1971). A preliminary study of the effects of play deprivation on young goats. Z.
Tierpsychol. 28:517-528.
Csikszentmihalyi, M. (1975). Flowing: A general model of intrinsically rewarding experiences.
In preparation.
Csikszentmihalyi, M., and Bennett, S. (1971). An exploratory model of play. Amer. Anthro.
73:45-58.
Delius, J. D. (1969). A stochastic analysis of the maintenance behavior of skylarks. Behavior
33:137-178.
Dingle, H. (1972). Aggressive behavior in stomatopods and the use of information theory in the
analysis of animal communication. In Winn, H. E., and Olla, B. S. (eds.), Behavior oj
Marine Animals, Vol. \, Plenum Press, New York.
Dissanayake, E. (1974). A hypothesis of the evolution of art from play. Leonardo 7:211-217.
Dolhinow, P. J., and Bishop, N. (1970). The development of motor skills and social rela-
tionships among primates through play. Minn. Symp. Child. Psychol. 4:141-198.
Eisenberg, J. F., and Leyhausen, P. (1972). The phylogenesis of predatory behavior in mam-
mals. Z. Tierpsychol. 30:59-93.
Ellis, M. J. (1973). Why People Play, Prentice-Hall, Englewood, New Jersey.
Ewer, R. F. (1968). Ethology oj Mammals, Plenum Press, New York.
Fagen, R. (1974a). Selective and evolutionary aspects of animal play. A mer. Nat. 108:850-
858.
Fagen, R. (1974b). Theoretical bases for the evolution of play in animals. Ph.D. dissertation,
Harvard University.
Fagen, R. (1976). Selection for optimal age-dependent schedules of play behavior. Amer. Nat.
(in press).
Fedigan, L. (1972). Social and solitary play in a colony of vervet monkeys. Primates 13:347-
364.
Geist, V. (1971). Mountain Sheep. University of Chicago Press, Chicago.

Gilbert, B. K. (1968). Development of social behavior in the fallow deer (Dama dama). Z.
Tierpsychol. 25:867-878.
Gilmore, J. B. (1966). Play: A special behavior. In Haber, R. N. (ed.), Current Research in
Motivation. Holt, Rinehart, and Winston, New York.
Glickman, S. E. (1973). Responses and reinforcement. In Hinde, R. A., and Stevenson-Hinde,
J. (eds.), Constraints on Learning. Academic Press, New York.
Glickman, S. E., and Schiff, B. B. (1967). A biological theory of reinforcement. Psychol. Rev.
74:81-109.
Golani, I. (1973). Non-metric analysis of behavioural interaction sequences in captive jackals
Canis aureaus L. Behaviour 44:89-112.
Goyer, R. S. (1970). Communication, communicative process, meaning: Toward a unified
theory.J. Commun. 20:4-16.
Groos, K. (1898). The Play oj Animals. D. Appleton and Co., New York.
Gulick, L. H. (1920). A Philosophy oj Play. C. Scribner, New York.
Hailman, J. P., and Sustare, B. D. (1973). What a stuffed toy tells a stuffed shirt. BioScience
23:644-651.
Harlow, H. F. (1969). Age-mate or peer affectional systems. Adv. Study Behav. 2:333-383.
Harper, L. V. (1968). The effects of isolation from birth on the social behavior of guinea pigs
in adulthood. Anim. Behav. 16:58-64.
Harper, L. V. (1970). Ontogenetic and phylogenetic functions of the parrent-offspring relation-
ship in mammals. Adv. Study Behav. 3:75-117.
Hazlett, B. A., and Bossert, W. A. (1965). A statistical analysis of the aggressive communica-
tion systems of some hermit crabs. Anim. Behav. 13:357-373.
Hebb, D. O. (1946). Emotion in man and animals: An analysis of the intuitive processes of
recognition. Psychol. Rev. 53:88-106.
Hebb, D. O. (1955). Drives and the C.N.S. (Conceptual Nervous System). Psych. Rev. 62:243-
254.
Henry, J. D., and Herrero, S. M. (1974). Social play in the American black bear: Its similarity
to canid social play and an examination of its identifying characteristics. Amer. Zool.
14:371-389.
Hill, H. L., and Bekoff, M. (1976). Characterizing social play in young canids. In preparation.
Hinde, R. A., and Stevenson, J. G. (1969). Sequences of behavior. Adv. Study Behav. 2:267-
296.
Hirsch, I. (1971). Metrische analyse der Bellsequenzen das Haushundes. Thesis, Humboldt
University.
John, E. R., Chesler, P., Bartlett, F., and Victor, I. (1968). Observation learning in cats.
Science 159: 1489-1491.
Jolly, A. (1972). The Evolution oj Primate Behavior. Macmillan, New York.
Kruijt, J. P. (1964). Ontogeny of social behavior in Burmese red junglefowl Gallus gallus
spadiceus. Behaviour Suppl. 12.
Kruuk, H. (1972). The Spotted Hyena, University of Chicago Press, Chicago.
Leyhausen, P. (1965). Ober die Funktion der relativen Stimmungs hierarchie Dargestellt am
Beispiel der phylogenetischen und ontogenetischen Entwicklung des Beutesfangs van
Raubtieren. Z. Tierpsychol. 22:412-494. (translated in Lorenz, K., and Leyhausen, P.
(1973). Motivation oj Human and Animal BehaVior. Van Nostrand Reinhold, New York).
Leyhausen, P. (1973). Verhaltensstudien an Katzen. Z. Tierpsychol. Suppl. 2.
Loizos, C. (1966). Play in mammals. Symp. Zool. Soc. Land. 18: 1-9.
Loizos, C. (1967). Play in higher primates: A review. In Morris, D. (ed.), Primate Ethology.
Aldine, Chicago.
186 Marc Bekoff
Lorenz, K. (1965). Evolution and Modification oj Behavior, University of Chicago Press,

Chicago.
Lorenz, K. (1974). Analogy as a source of knowledge. Science 185:229-234.
Maier, R. A., and Maier, B. M. (1970). Comparative Animal Behavior. Brooks/Cole,
Belmont, California.
Marler, P. (1961). The logical analysis of animal communication. J. Theor. Bioi. 1:295-317.
Mason, W. A. (1967). Motivational aspects of social responsiveness in young chimpanzees. In
Stevenson, H. W., Hess, E. H., and Rheingold, H. L. (eds.), Early Behavior, John Wiley
and Sons, Inc. New York.
Mason, W. A. (1973). Regulatory functions of arousal in primate psychosocial development.
In Carpenter, C. R. (ed.), Behavioral Regulators oj Behavior in Primates, Bucknell
University Press, Lewisburg, Penn.
Maurus, M., and Pruscha, H. (1973). Classification of social signals in squirrel monkeys by
means of cluster analysis. Behaviour 47:106-128.
Maurus, M., Hartmann, E., and Kiihlmorgen, B. (1974). Invariant qualities in communication
processes of squirrel monkeys. Primates 15:179-192.
Metz, H. (1974. Stochastic models for the temporal fine structure of behaviour sequences. In
McFarland, D. (ed.). Motivational Control Systems Analysis, Academic Press, New York.
Millar, S. (1968). The Psychology oj Play, Penguin, Baltimore, Maryland.
Miller, R. E., Caul, W. F., and Mirsky, I. A. (1967). Communication of affects between feral
and socially isolated monkeys. J. Pers. Soc. Psychol. 7:231-239.
Morris, D. (1956). Discussion of K. Lorenz, Plays and vacuum activities. In L'instinct dans Ie
comportement des animaux et de ['homme, Masson, Paris.
Morris, D. (1969). The Human Zoo, Dell, New York.
MUlier-Schwarze, D. (1968). Play deprivation in deer. Behaviour 31: 144-162.
MUlier-Schwarze, D. (1971). Ludic behavior in young mammals. In Sterman, M. B., McGinty,
D. J., and Adinolfi, A. M. (eds.), Brain Development and Behavior, Academic Press, New
York.
MUller-Schwarze, D. and C. (1975). Sequential analysis of motor patterns in mammalian
"play" behavior. In preparation.
Nadel, L., O'Keefe, J., and Black, A. (1975). Slam on the brakes: A critique of Altman, Brun-
ner, and Bayer's response-inhibition model of hippocampal function. Behav. Bioi. 14:151-
162.
Nelson, K. (1964). The temporal patterning of courtship behavior in the glandulocaudine
fishes. Behaviour 24:90-146.
Nelson, K. (1973). Does the holistic study of behavior have a future? In Bateson, P. P. G., and
Klopfer, P. H. (eds.), Perspectives in Ethology, Vol. I, Plenum Press, New York.
Olds, J. (1956). Pleasure centers in the brain. Sci. Amer. 195:105-117.
Otte, D. (1974). Effects and functions in the evolution of signaling systems. Ann. Rev. Ecol.
Syst.5:385-417.
Owens, N. W. (1972). The development of behaviour in free-living baboons, Papio anubis. D.
Phil. dissertation, University of Cambridge, England, 176 pp.
Owens, N. W. (1975). Social play behaviour in free-living baboons, Papio anubis. Anim. Behav.
23:387-408 and 757-765.
Pfaffmann, C. (1973). The comparative approach to physiological psychology. Ann. N. Y.
Acad. Sci. 223:57-64.
Poole, T. B. (1973). Cine film and computer analysis of rapid sequences of behavior. Med.
Bioi. Illus. 23:170-175.
Poole, T. B., and Fish, J. (1975). An investigation of playful behavior in Rattus norvegicus and
Mus musculus (Mammalia).J. Zool. Lond. 175:61-71.
Popper, K. (1965). Conjectures and Refutations: The Growth of Scientific Knowledge. Harper,
New York.
Rasa, O. A. E. (197Ia). Social interaction and object manipulation in weaned pups of the
northern elephant seal, Mirounga angustirostris. Z. Tierpsychol. 29:82-102.
Rasa, O. A. E. (197Ib). Appetence for aggression in juvenile damsel fish. Z. Tierpsychol.
Suppl. 7.
Rasa, O. A. E. (1973). Prey capture, feeding techniques, and their ontogeny in the African
dwarf mongoose, Helogale undulata refula. Z. Tierpsychol.32:449-488.
Reynolds, P. C. (1975). Playas a state of mind. In preparation.
Rhine, R. J. (1973). Variation and consistency in the social behavior of two groups of
stumptail macaques Macaca arctoides. Primates 14:21-35.
Sackett, G. P. (1973). Innate mechanisms in primate social behavior. In Carpenter, C. R.
(ed.), Behavioral Regulators of Behavior in Primates. Bucknell University Press,
Lewisburg, Penna.
Sackett, G. P., and Ruppenthal, G. C. (1973). Development of monkeys after varied
experiences during infancy. In Barnett, S. A. (ed.), Ethology and Development. Spastics
International Medical Publication with Heinemann Medical Books Ltd. London.
Sade, D. S. (1973). An ethogram for rhesus monkeys. l. Antithetical contrasts in posture and
movement. Amer. J. Phys. Anthro. 38:537-542.
Schaller, G. B. (1972). The Serengeti Lion. University of Chicago Press, Chicago.
Schleidt, W. M. (1973). Tonic communication: Continual effects of discrete signs in animal
communication systems. J. Theor. Bioi. 42:359-386.
Schmidt-Nielsen, K. (1972a). Locomotion: Energy cost of swimming, flying, and running.
Science 177:222-228.
Schmidt-Nielsen, K. (1972b). How Animals Work. Cambridge University Press, New York.
Scott, J. P., and Fuller, J. L. (1965). Genetics and the Social Behavior of the Dog. University
of Chicago Press, Chicago.
Simonds, P. E. (1974). The Social Primates. Harper and Row, New York.
Slater, P. J. B. (1973). Describing sequences of behavior. In Bateson, P. P. G., Klopfer, P. H.
(eds.), Perspectives in Ethology. Vol. I, Plenum Press, New York.
Slater, P. J. B., and Ollason, J. C. (1972). The temporal pattern of behavior in isolated male
zebra finches: Transition analysis. Behaviour 42:248-269.
Steinberg, J. B., and Conant, R. C. (1974). An informational analysis of the intermale be-
havior of the grasshopper Chortophaga viridifasciata. Anim. Behav. 22:617-627.
Steiner, A. L. (1971). Play activity of Columbian ground squirrels. Z. Tierpsychol.28:247-261.
Struhsaker, T. T. (1967). Behavior of vervet monkeys Cercopithecus aethips. Univ. Calif. Pub.
Zool., Berkeley.
Suomi, S. J., and Harlow, H. F. (1969). Apparatus conceptualization for psychopathological
research in monkeys. Behav. Res. Meth. Instr. 1:247-250.
Suomi, S. J., and Harlow, H. F. (1972). Depressive behavior in young monkeys subjected to
vertical chamber confinement. J. Compo Physiol. Psychol. 80: 11-18.
Symons, D. A. (1973). Aggressive play in a free-ranging group of rhesus monkeys (Macaca
mulatta). Ph.D. dissertation, University of California, Berkeley. 241 pp.
Symons, D. (1974). Structure and function in the aggressive play of rhesus monkeys. A mer.
Zool.14:317-322.
Tembrock, G. (1958). Speilverhalten beim Rotfuchs. Zool. Beitr. Berl. 3:423-496. (English
translation by B. Piddack).
Tinbergen, N. (1969). Ethology. In Harre, R. (ed.), Scientific Thought 1900-1960. Clarendon
Press, Oxford.
188 Marc Bekoff
van Hooff, J.A.R.A.M. (1967). The facial displays of the catarrhine monkeys and apes. In
Morris, D. (ed.), Primate Ethology, Aldine, Chicago, 471 pp.
Vincent, L. E., and Bekoff, M. (1976). Ontogeny of predatory behavior in coyotes, Canis latrans.
Submitted for publication.
West, M. (1974). Play in the domestic cat. A mer. Zool. 14:427-436.
Wiepkema, P. R. (1961). An ethological analysis of the reproductive behavior of the bitterling
Rhodeils amarus Bloch. Arch. Neerl. Zool. 14:103-199.
Wilson, S. (1973). The development of social behaviour in the vole Microtus agrestis. Zoo!. J.
Linn. Soc. 52:45-62.
Wilson, S., and Kleiman, D. G. (1974). Eliciting and soliciting play: A comparative study:
Octodon, Octodonomys, Pediolagus, Phoca, Choeropsis, Ailuropoda. Amer. Zool. 14:341-
370.
Winch, W. H. (1906). Psychology and philosophy of play. I and II. Mind 15:32-52 and 177-
190.
Winter, P. (1968). Social communication in the squirrel monkey. In Rosenblum, L. A., and
Cooper, R. W. (eds.), The Squirrel Monkey, Academic Press, New York.
WUstehube, C. (1960). Beitrage zur Kenntnis besonders des Spiel und Beutefangverhaltens
einheimischer Musteliden. Z. Tierpsycho!. 17:579-613.
Young, P. T. (1973). Emotion in Man and Animal. Robert E. Krieger Publishing Co., New
York.
Yousef, M. K., Robertson, W. D., Dill, D. B., and Johnson, H. D. (1973). Energetic cost of
running in the antelope ground squirrel Ammospermophilus leucurus. Physiol. Zool.
46:139-147.
Chapter 5
EXERCISE, PLAY, AND PHYSICAL

TRAINING IN ANIMALS
Robert M. Fagen
Provisional Department of Ecology, Ethology, and Evolution
The University of lllinois
Urbana, Illinois 61801
I. ABSTRACT
Since animal play includes vigorous physical exercise, such behavior has
been hypothesized to serve in part as physical training for young animals.
Citing recent physiological studies of training responses of muscle, bone,
and the cardiopulmonary, endocrine, and nervous systems, I offer seven
testable, physiologically based predictions about play, including its be-
havioral composition, interruption and repetition, age dependence, and
phylogenetic distribution. 1 also discuss manipulative, social, and diversive
aspects of play which the physical training hypothesis alone does not fully
explain.
II. INTRODUCTION
Animal play behavior has been defined by Welker (1961) as

a wide variety of vigorous and spirited activities: those that move the organism or its
part through space such as running, jumping, rolling and somersaulting, pouncing upon
and chasing objects or other animals, wrestling, and vigorous manipulation of body
parts or objects in a variety of ways. The goals and incentives of vigorous play consist
of certain patterns of variable or changing stimulation of the sensory surfaces.
Clearly, as Groos (1896), Brownlee (1954), West (1974), and others point
out, vigorous play will also stimulate muscle, bone, and the
cardiopulmonary system, which fact led Brownlee (1954) to propose that
189
190 Robert M. Fagen
play serves as physical training for young animals, exercising and thereby
maintaining the somatic muscle groups used in escape, fighting, and re-
production. In the absence of play, according to Brownlee, these muscle
groups would not develop adequately, as they are seldom used in a func-
tional context until adulthood.
Brownlee's argument is further supported by recently obtained data on
the physiology of muscular exercise. Important advances in exercise
physiology and neurobiology have been achieved since Brownlee's paper ap-
peared, and enduring adaptive physiological, cytological, and biochemical
responses to vigorous and repeated stimulation ("training responses") are
now known to occur not only in skeletal muscle, bone, and the
cardiopulmonary system (A strand and Rodahl, 1970; Basset, 1972;
Burleigh, 1974; Goldspink, 1970; Holloszy, 1973; Keul et al., 1972) but also
in the endocrine and nervous systems (Greenough, 1974, 1975; Rose, 1969;
Tharp and Buuck, 1974; Valverde, 1967). Lacking this additional informa-
tion, Brownlee could only discuss adaptive physiological responses to
exercise in broad and general terms and was· forced to view all neurobio-
logical, perceptual, and social consequences of play behavior as incidental.
In addition, Brownlee failed to explain why play is almost universal in
homoiotherms and virtually absent from the rest of the animal kingdom.
Aims of the present paper are to offer testable predictions about playas
exercise, based on physiological information not available when Brownlee's
physical training hypothesis appeared; to resolve the abovementioned defi-
ciencies in that hypothesis; and to clarify a number of additional points re-
lating to play and muscular activity.
PhysiologicaHy defined training effects in the nervous and endocrine
systems were not found until recently, nor were the biochemical correlates
and age- and system-specificity of training well understood prior to the
development of modern histological, biochemical, and catheterization tech-
niques. Hence this paper implies no criticism of the pioneering work of
Brownlee and others who attempted to interpret play behavior in a physio-
logically meaningful way.
III. PHYSICAL EXERCISE IN ANIMAL PLAY
As mentioned above, Groos (1896), Brownlee (1954), and other au-

thorities, including Beach (1945), Hinde (1971), Loizos (1966), and Miiller-
Schwarze (1971), point out that the goals and incentives of vigorous play
need not be restricted to sensory stimulation but may also include muscular,
skeletal, and cardiovascular training effects resulting from vigorous mus-
Exercise, Play, and Physical Training in Animals 191
cular exercise frequently observed to occur in play. Indeed, motor patterns

producing vigorous exercise and occurring in a nonfunctional context are
common in the behavior of young mammals (Table I) and young birds
(Table II).
In this paper I will use the term "exercise play" to denote sequences of
motor patterns, such as those listed in Tables I and II, that result in
vigorous exercise in a nonfunctional context (either solitary or social).
These motor patterns would include the· "locomotor-rotational move-
ments" described by Wilson and Kleiman (1974). The terms "locomotor
play" (MUller-Schwarze, 1971) and "acrobatics" (Loizos, 1967) have been
used to describe particular types of exercise play.
Play behavior has been observed and fully described most frequently in
predators (e.g., carnivores, pinnipeds, raptors) and in animals having highly
developed visual and manipulative abilities (e.g., monkeys and apes, par-
rots, corvids), but is also found in numerous other species of homoiotherms,
as the above discussion indicates. In general, poikilotherms do not play
(Beach, 1945; Meyer-Holzapfel, 1956; Thorpe, 1966). Possible physiological
reasons for this sharp phylogenetic discontinuity will be suggested in a later
section.
IV. EXERCISE PHYSIOLOGY AND TRAINING RESPONSES
Regular performance of physical work by an animal tends to increase

the organism's capacity to perform that work. Such "enrichment" of the
animal's environment through regular work and exercise results in profound
and long-lasting adaptive physiological changes (termed "training
responses") in mammals of several species, and particularly in young mam-
mals (A strand and Rodahl, 1970; Holloszy, 1973; Keul et al., 1972).
The above meaning of the word "training" should be distinguished
from a very different psychological use of the same word. In psychology,
"training" is sometimes used to indicate a form of directed learning in
which "responses are channeled and are followed by immediate reinforce-
ment" (Beck, 1975). In this paper I use "training" solely in the physio-
logical sense indicated above.
Strength and endurance of particular muscle groups, maximum work
rate, and endurance in whole-body work such as running and swimming.
have all been shown to adapt to a program of regular exercise. Muscle mass
(A strand and Rodahl, 1970), heart size (Poupa et al., 1970) and capillary
density (Bloor and Leon, 1970; Tomanek, 1970), and vital capacity
(Ekblom, 1969a) can also adapt to exercise stress. Indeed, growth rate
192 Robert M. Fagen
Table I. Representative Patterns of Playful Exercise in Mammals
Species Play patterns References
Human children Manipulate briskly, run, chase, flee, Hutt,1970

(Homo sapiens) jump, swim
Chimpanzees Jump and leap wildly, chase, wrestle van Lawick-Goodall,
(Pan troglodytes) 1968
Rhesus monkeys Chase, flee, wrestle, stalk and pounce Harlow and Harlow,
(Macaca mulatta) 1965
Infant canids Wrestle, jaw-wrestle, chase their tails Bekoff, 1974
(Canis sp.)
Polecats Mount, jump on, attack, and defend with Poole, 1966
(Mustela putorius) inhibited bites and rapid reversal of roles
Northern elephant Rear, wrestle, thrash Rasa, 1971
seals (Mirounga
augustirostris)
American black bear Chase, climb, rear, lunge, roll Henry and Herrero,
(Ursus americanus) 1974
Giant pandas Roll, somersault, perform supported Wilson and Kleiman,
(Ailuropoda mel- headstands and handstands, butt, swat 1974
anoleuca)
African dwarf mon- Bound, fling, smash backwards, roll Rasa, 1973
goose (Helogale
undulata)
Spotted hyenas Roll, pull at the ears and tail of a con- van Lawick and van
(Crocuta crocuta) specific, chase, run in circles, tug at ob- Lawick-Goodall,
jects held by conspecifics 1970; Kruuk, 1972
Domestic cats Spring, chase, kick, climb, leap, fling, Leyhausen, 1973a, b;
(Felis Calus) rear, bat West, 1974
Domestic cattle Gallop, buck, kick, head-butt, gore Brownlee, 1954
(Bas taurus)
Red deer Chase and reverse-chase, stand on hind Darling, 1937
(Cervus elaphus) legs, kick, jump, leap into the air
Ponies Push, kick, mount, gallop rapidly back Tyler, 1972
(£quus cabal/us) and forth, toss and paw at objects, bite,
prance, chase
Warthogs Run and flee, toss the head, spring, hop Fradrich, 1965
(Phacochoerus and jump
aethopicus)
(Ekblom, 1969a) and possibly even length of life (Arshavsky, 1972) can
respond adaptively and positively to a program of regular physical exercise,
provided that the exercise program begins sufficiently early in ontogeny,
before growth ceases in adulthood.
Physiological mechanisms that produce training responses, and indeed
Table I-Continued
Pygmy hippopotami Push, shake and toss the head, splash leap Wilson and Kleiman,
(Choeropsis out of the water, pirouette on hind legs 1974
Iiberensis )
Tapirs Gallop, head toss, chase wildly around Fradrich and Thenius,
(Tapirus sp.) their mothers 1972
Hedgehogs Nose, bite, push, pick up and carry ob- Dimelow, 1963
(Erinaceus jects, and run back and forth
europaeus)
Olympic marmots Rear, push, lunge suddenly, chase, somer- Barash, 1973
(Marmota olympus) sault
Norway rats Run, box, chase MUller-Schwarze,
(Rattus norvegicus) 1971
Salt desert cavies Jump, turn, hind leg kick-back, run head Wilson and Kleiman,
(Dolichotis sali- shake, head toss, body twist 1974
nicola)
Degus Head shake, body twist, jump, run Wilson and Kleiman,
(Octodon degus) 1974
Bottle-nosed dolphins Chase, hurl objects into tlie air, toss, drop Tavolga, 1966;
(Tursiops truncatus) and snatch objects. Tavolga and Essapian,
1957
Vampire bats Wing-flap and chase Schmidt and Manske,
(Desmodus ro- 1973
tundus)
African ele:-oants Slap each other with their trunks, roll Sikes, 1971
(Loxodonta afri- over, gambol
can us)
Whiptail wallabies Hop rapidly back and forth, bat, leap, hit, Kaufmann, 1974
(Macropus parryi) and jump around
Southern shortnosed Actively chase, leap, burrow, climb Heinsohn, 1966
bandicoot
(lsoodon obesulus)
Tasmanian devils Chase, fight, vigorously manipulate Ewer, 1968;
(Sarcophilus har- Hediger, 1958
risii)
Wombats Butt, rear, jump and roll over Grzimek, 1967;
(Family Hediger, 1958;
Phascolomidae) Troughton, 1941;
Wunschmann, 1966
the training responses themselves, vary with type, frequency, intensity, and
duration of exercise and with age. "Work overload" exercises, such as
weightlifting or static muscle contraction, which only stress certain muscle
groups, will develop (increase the mass of) only those muscles which are
exercised, while whole-body "endurance" exercise, such as running, cycling,
194 Robert M. Fagen
Table II. Representative Patterns of Playful Exercise in Birds
Frigate-birds Swoop down over rock-pools, attempt to Gibson-Hill, 1947;

(Fregata aquila, break tree branches, catch feathers and Stonehouse and
Fregata minor) strands of seaweed from each other in Stonehouse, 1963
midair while flying together
Krestrels Pounce on paper wads, sticks, and other Battersby, 1944;
(Falco tinnunculus) inanimate objects Brosset, 1973;
Thorpe, 1963
Domestic chicks Run briefly with wings raised, jump up Guhl,1958
(Gallus domesticus) and down
Inca Terns Fly vigorously, pick up objects, fly with Ashmole and Tovar,
(Lacosterna inca) and drop and retrieve objects 1968
Keas Splash and float objects in running water, Porter, 1947
(Nestor notabilis) chase and flee conspecifics, seize and Jackson, 1963b
manipulate objects
Hornbills Wrestle, chase, and flee Moreau and Moreau,
(Bycanistes brevis, 1944
Bycanistes bucina- Swynnerton, 1908
tor)
Hairy woodpeckers Perform wild, erratic flights, dodge and Kilham,1974
(Dendrocopos shift in a lively fashion around tree
vil/osus) and Downy trunks
woodpeckers (D.
pubescens)
Ravens Lie on their backs and shift objects from Gwinner, 1966;
(Corvus corax) beak to claw, throw objects into the air Hf/ljgaard, 1954
and catch them, hang upside down,
balance on light branches, chase a con-
specific that is carrying an object, catch
the object in midair if it should be drop-
ped by the other animal
jogging, or swimming, tends to increase the biochemical potential (but not

the mass of) a wide variety of muscles and to improve cardiopulmonary
system function (Astrand and Rodahl, 1970). Dynamic exercise may
improve dynamic strength while leaving isometric strength unchanged
(Hansen, 1967). In humans, the immediate response to endurance exercise
stress involves cardiopulmonary adjustments on a time scale of several
weeks, but if exercise continues for a longer period of time, say several
months, other physiological parameters will also adapt, further improving
the organism's endurance (Astrand and Rodahl, 1970; Ekblom, 1969b). A
significant component of this second, long-term response to exercise is pro-
found cellular and biochemical adaptation. Rats subjected to a program of
endurance exercise show increased steady state amounts of enzymes and

respiratory pigments. In addition, such trained animals exhibit greater use
of fats during exercise than do untrained controls, and it is believed that this
alteration in substrate source, which results in a reduced rate of glycogen
breakdown, is related to the increased endurance capacity and delayed
onset of fatigue seen in trained rats (Holloszy, 1973; Keul et al., 1972). To
cite one such response whose detailed dynamics are known, the steady-state
cytochrome content of rat quadriceps muscle approximately doubles in
response to regular endurance exercise, due primarily to a decrease in
cytochrome breakdown rate (Terjung et al., 1973).
The nature and magnitude of training responses can differ strikingly in
organisms of different ages (Bloor and Leon, 1970; Ekblom, 1969a, b;
Eriksson, 1972; Hartley et al., 1969; Poupa et al., 1970; Rarick and Larsen
1959; Rowell, 1974). For example, overloading of the heart in young rabbits
leads to accelerated growth both of cardiac cells and of the cardiac
capillary bed, but adults exhibit only changed cell volume (Poupa et aI.,
1970). Young men, but not older individuals, exhibit increased maximal
oxygen extraction from blood (mixed arteriovenous oxygen difference) after
exercising regularly for several months (Hartley et al., 1969). Cardiac
hypertrophy in response to endurance exercise occurs both in young white
rats and in adults, but adults require greater frequency and intensity of
exercise than do juveniles in order to achieve the same magnitude of
response (Bloor and Leon, 1970). Adult humans and ll-13 year old boys
increased their maximum oxygen consumption by the same fractional
amount after corresponding programs of training, but responses of different
physiological systems produced this increase in boys and men, suggesting
that "if physical training is begun as early as in prepuberty," greater
physical capacity "may be attained in adulthood than if training were
initiated later in life," since both systems rather than just one will be trained
in the former case (Eriksson, 1972). This physiological evidence suggests
that training programs should begin early in life in order to be maximally
effective, for the following three reasons:
1. Some training responses occur only in young animals.
2. Young animals often respond more rapidly or more strikingly to
training than do adults.
3. Even when a training response is not age-specific, the physical ca-
pacity of an animal which has trained since a very early age will at
least temporarily exceed that of an agemate that begins training in
adulthood, if only because the first animal has trained for a longer
total period of time.
Decreased muscular activity also causes physiological change, indeed
196 Robert M. Fagen
physiological deterioration, in animals. After prolonged inactivity (bed rest,

hospitalization, weightlessness of space flight, or simply a sedentary mode
of life) endurance, work capacity (A strand and Rodahl, 1970; Rowell, 1974;
Saltin et al., 1968), and muscular strength all tend to decrease. Cytological
and biochemical correlates of this "detraining" response are analogous but
opposite in sign to those observed after endurance training (Saltin et al.,
1968).
Evidence cited above suggests that young organisms, at least young
mammals, must be viewed as possessing significant physiological plasticity.
This evidence further suggests that behavior in young mammals should be
viewed not only in terms of adaptation to the immediate environment or in
terms of traditional ontogenetic and learning paradigms, but also in terms
of possible impact on the young organism's developing physical capacity.
Whenever a young animal exercises its muscles, it produces feedback which
will affect the growth and development of those muscles, perhaps at the
expense of other muscles that are exercised less frequently or less intensely.
One can even imagine different muscle groups competing for limited nu-
trients. In this physiological view of behavioral ontogeny, patterns of ju-
venile motor behavior are viewed as a mechanism for allocating exercise-
and therefore training and nutrients-to particular muscle groups. Note
that such plasticity would be extremely important if young animals
developed individual behavioral strategies for vigorous activities such as
predation, predator escape, fighting, or courtship, since each individual
would then tend to exhibit idiosyncratic sequences of motor behavior. As a
result, each individual would levy a different set of physiological perform-
ance requirements on its muscles, and practice or repetition of idiosyncratic
patterns of behavior would implement these requirements by inducing train-
ing responses in the particular muscle groups which such motor sequences
stressed most heavily.
How might such molding of physiological plasticity occur in young
animals? It is likely that a young animal deprived of all opportunities for
regular vigorous exercise would not deVelop the physiological machinery
necessary for later survival and reproduction. [That physical capacity itself,
rather than prey abundance or time, may sometimes be a limiting factor for
at least some animals, is suggested by an observation of Eaton's (1969) on
apparent physical exhaustion occurring in a cheetah following an unsuccess-
ful hunt.] Brownlee (1954) argues that an animal that would not be expected
to regularly exercise certain muscle groups (e.g., those needed for predation,
escape, fighting, or reproduction) in a functional context until adulthood
might well suffer from the physiological consequences of its inactivity, while
an animal that regularly exercised such muscle groups in a nonfunctional
context ("play") would train itself physically in preparation for adulthood.
Brownlee cites additional muscle groups that are regularly used by young
animals in functional contexts but which are not used in play. As argued
above, current understanding of the physiology of exercise in mammals
further supports Brownlee's hypothesis, suggesting that regular vigorous
exercise must take place before adulthood to ensure adequate training.
Once growth ceases, exercise is less effective and is therefore more costly in
terms of time and energy; in adulthood, exercise will often occur in a func-
tional context; moreover, training of some physiological functions in
adulthood appears to be impossible. For these reasons alone, young animals
should play.
In the next section I will discuss properties and characteristics of play
behavior which may be explained in terms of physical training: "playful"
form (including behavioral composition, interruption and repetition, rough-
and-tumble, and approach-withdrawal); age-dependence; play deprivation
and exercise deprivation; "warmup" play before foraging; play with fa-
miliar objects; and the phylogenetic distribution of play behavior. I will then
consider other aspects of play for which physical training does not appear
to furnish a sufficient explanation.
V. CONSEQUENCES OF THE PHYSICAL TRAINING

HYPOTHESIS
A. Why Playful Rather than Serious Forms of Activity?
From a strictly physiological point of view, vigorous exercise in a

nonplay context does not differ from vigorous exercise in play. (Differences
in muscle tone have been suggested, but such differences remain to be
demonstrated electromyographically.) Functional forms of fighting, escape,
and predation could be extremely costly in terms of time, energy, and sur-
vivorship, since they would tend to expose the young organisms to very real
dangers. In addition, the partner in such serious interactions should be
selected to fight back, attack, or escape in ways which were not only func-
tional but also produced the smallest possible physical training response in
the animal which initiated the interaction. In play, on the other hand,
choice of training schedule and intensity is left to the individual, and natural
selection should act on the behavioral composition (as argued by Brownlee),
frequency, duration, and intensity of play bouts in order to produce optimal
age-specific, species-specific, and sex-specific exercise programs. That is,
play should train muscles otherwise used only in vigorous adult activities
such as prey-catching, fighting, predator avoidance, and mating (Ewer,
198 Robert M. Fagen
1968; Wilson, 1971). Muscle groups exercised in frequently occurring

activities performed in functional contexts by the young animal (nursing,
mastication, vocalization, lying down or rising up, and traveling with the
group) are not used specifically in play, at least in domestic cattle
(Brownlee 1954). Nonstrenuous adult behavior patterns, including agonistic
and sexual signals and displays, also tend to be absent from play. For
example, the physical training hypothesis correctly predicts the absence
from rhesus monkey play of specialized agonistic visual and auditory
signals, reported by Symons (1974). As Symons points out, hypotheses
which view playas a medium for practice and perfection of communicative
skills fail to account for this observation.
While exercise physiology cannot yet predict the exact frequency, dura-
tion, and intensity of optimal training programs, certain physiological prin-
ciples of optimal exercise may be cited. These principles lead to testable
predictions about play, as the following discussion will indicate.
1. Interruption and Repetition
To be most efficient, strenuous exercise should be both interrupted and

repeated (Astrand and Rodahl, 1970). During the periods of interruption,
light exercise should be performed (Astrand and Rodahl 1970). As one
would expect from these principles, interruption and repetition are com-
monly observed in play (Henry and Herrero, 1974; MUller-Schwarze, 1971)
and have indeed been considered to be among its defining characteristics
(Loizos, 1966, 1967).
2. Types oj Training
In humans, three distinct types of optimal exercise programs and three

corresponding types of training responses have been identified by exercise
physiologists (Keul et af., 1972). Brief (30 sec) periods of strenuous
"overload" exercise involving high-intensity and even maximal muscle
contractions mainly increase the strength and speed of the specific muscles
exercised, through training effects which consist mainly of muscular
hypertrophy. Slightly longer (1-3 min) periods of high-intensity exercise
might be expected to increase aerobic capacity and local muscular endu-
rance by acting on glycolytic and oxidative enzyme levels. Prolonged
exercise periods (several minutes to an hour or more) at lower intensities
("endurance exercises") build endurance and aerobic capacity by training
those cardiovascular, respiratory, and biochemical system components
associated with oxygen transport and aerobic metabolism. While absolute
duration of these three types of exercise periods may depend on body size or
on some other characteristic of the organism, relative magnitudes of the

three types of exercise period should not change since underlying
biochemical mechanisms are the same across species. In animals, one would
therefore predict three types of playful exercise. The first type should
involve maximal, usually static (sometimes called "isometric") exercise:
pushing, pulling, shoving, biting, hanging by a limb, springing, jumping,
squeezing, butting, pushing, pulling an immovable object, etc. This type
of exercise should occur in brief but intense bouts that are frequently
interrupted by rest periods. The third type should involve prolonged bouts of
locomotion at submaximal intensity, including motor patterns such as run-
ning, chasing and fleeing, swimming, or swinging from tree to tree. This
form of play has been termed "locomotor play" by MUller-Schwarze
(1971). The second type of exercise would be distinct from the first only
under ideal conditions, since the difference between the two types is es-
tablished by the exact intensity of exercise, while the sequences of motor
patterns used are the same in both cases (at least in humans). I suggest that
the first and second types of play correspond to rhesus monkey "rough-and-
tumble" play (Harlow and Harlow, 1965), that the third corresponds to
rhesus "approach-withdrawal" play (Harlow and Harlow, 1965), and that
these terms should also prove useful (for the physiological reasons cited
above) in describing the play of animals other than rhesus monkeys.
As mentioned above, the first type of exercise exerts highly specific ef-
fects, training only those particular muscle groups stressed by the exercise
program. Given the variety of muscle groups which must be exercised in
young animals, and given the physiological fact that only a few muscle
groups can be exercised at anyone instant of time, animals needing playful
exercise necessarily must have developed a diverse set of exercise patterns;
but the very diversity of these patterns has often caused play to be viewed as
a wastebasket, or catch-all, category of behavior. This view becomes mean-
ingless once the common physiological and evolutionary basis of these be-
havior patterns is understood. The power of Brownlee's hypothesis is that it
serves to unify and explain a collection of seemingly unrelated and seem-
ingly inexplicable behavior patterns, and does so in physiologically mean-
ingful terms.
B. Age Dependence
1. Allocation of Time and Energy to Play in Young and Adult Animals
Young animals tend to play more than do adult animals (Baldwin and
Baldwin, 1974; Dolhinow and Bishop, 1970; Hinde 1971; Loizos, 1966,
200 Robert M. Fagen
1967; MUller-Schwarze, 1971; West, 1974). This age-dependent scheduling

of play may be explained as a consequence of natural selection, assuming
that play results in immediate costs (e.g., diversion of time and energy from
other activities) and eventual benefits (due to increased physical capacity
caused by training effects of play) (Fagen, 1974b, 1976). As argued above,
exercise physiology supplies a well-defined mechanism by which play be-
havior can produce such eventual benefits. Indeed, the known age-speci-
ficity of many training responses suggests that play must occur during the
juvenile period if it is to occur at all, since these responses often tend to be
greatly curtailed in adulthood. In addition, as West (1974) points out,
vigorous exercise in functional contexts, such as feeding, predator avoid-
ance, social interaction, reproduction, and care of offspring, may be suffi-
cient to maintain the adult organism's physical capacity. Play in adulthood
would then be superfluous. This argument is consistent with exercise
physiology. The amount of exercise needed to enhance a given physical ca-
pacity level is considerably greater than that needed to maintain that
enhanced level (Astrand and Rodahl, 1970). Once an animal has attained a
given level of physical capacity after months or years of juvenile play, adult
activities, though less vigorous or less frequent than that play, may still
constitute adequate exercise for maintenance of that level of physical condi-
tion. Adults might still be expected to play in order to maintain physical
condition during periods of food abundance or low predator pressure. And
indeed, if adult mammals have been selected to exercise spontaneously in
oroer to maintain physical condition, the phenomenon of spontaneous
activity in caged individuals (e.g., wheel-running), which Collier
(1970) considers "poorly understood" despite the fact that it "has been
studied longer, more persistently, and with less profit than any other
single phenomenon in animal psychology," becomes transparent and fully
understandable.
Adults should also be selected to play with their offspring in order to
increase the offspring's physical capacity and therefore the adults' own in-
clusive fitness. However, since the interests of parent and offspring in mam-
mals and birds are not identical (Trivers, 1974), one would predict parent-
offspring conflict over the parent's participation in play with its offspring,
as follows. Initially, parent and offspring should agree to play together.
After a certain point in ontogeny, the offspring will continue to solicit play
from its parent, but the parent will respond less and less frequently to these
invitations. In domestic cats, the waning of the mother-kitten bond appears
to result from parent-offspring conflict over play participation and from
disagreement over access to milk (Rosenblatt et ai., 1962; Rosenblatt,
1971 ).
Parent-offspring play is known in a number of species (e.g., rhesus
monkeys: Redican and Mitchell, 1974; meerkats: Wemmer and Fleming,

1974; chimpanzees: van Lawick-Goodall, 1968; certain caviomorph rodents:
Wilson and Kleiman, 1974). It is interesting that the extent to which the
father is involved in such play can vary greatly from species to species
within a given taxonomic group (Wilson and Kleiman, 1974).
2. Thermoregulation, Play in InJancy, and the Development oj Locomotion
Play tends to begin shortly after walking begins, in late infancy in altri-
cial mammals and during the first days of life in precocial mammals (Table
III; Fig. 1). Man and the great apes seem to be the only clear exceptions to
this rule. Absence of play in early infancy in altricial mammals is under-
standable if one considers the severe functional requirements placed on the
infant. The time and energy demands of feeding (including possible costs of
competition with siblings, as discussed below) and of maintaining body
temperature through proper orientation to and contact with the parent, lit-
termates, or nest material (particularly when young are not fully en-
dothermic: Dawson, 1972) would appear to preclude play at this stage of
development. These demands (and therefore the cost in fitness due to the
diversion of time and energy from feeding and temperature maintenance)
should decrease with age, paralleling decreasing growth rate (grams weight
increment per gram body weight per unit time) and decreasing resting me-
tabolism ~;:;alories per gram body weight per unit time). Exercise play will
begin at that stage in ontogeny when the long-term benefits of exercise out-
weigh the short-term costs of such diversion of time and energy.
Figure 2 presents the above argument graphically. It may be seen from
Fig. 2 that the following two assumptions are sufficient to explain the
absence of play in early infancy in altricial species:
1. At birth, the cost of play, defined as the decrease in fitness resulting
from diversion of one unit of time or energy from feeding and
temperature maintenance, is greater than the benefit of play, defined
as the fitness increment resulting from the allocation of one unit of
this resource to play.
2. During infancy and during the juvenile period, the cost of play
decreases more rapidly with time than does the benefit resulting
from play. (This cost function may increase again in early
adulthood: Fagen, 1976.)
Arshavsky (1972) claims that the heavy demands placed on skeletal
muscle groups in early infancy by the need for temperature maintenance
preclude the ongoing use of these muscles in other contexts, and that when
the animal attains a standing posture "the skeletal muscles are fully or
202 Robert M. Fagen
Table III. Development of Locomotion and Play in Some Homoiotherms
Locomotion begins
Species (walking or flying) Play begins
Homo sapiens 11-13 mo. (Hurlock, 1972) 3 mo. (Hurlock, 1972)

11-15 mo. (Johnson and 1'/2 mo. (Johnson and Me-
Medinnus, 1965) dinnus, 1965)
Pan troglodytes 4-6 mo. (lab) (Hinde, 1971) 3-4 mo. (Hinde, 1971)
Pan troglodytes c. 12 mo. (field) (Hinde, 3-4 mo. (Hinde, 1971)
1971)
Gorilla gorilla 5-6 mo. (Hinde, 1971) 3-4 mo. (Hinde, 1971)
Macaca mulatta 2-3 wk. (Hinde, 1971) 2 wk. (Meier and Devanney
1974)
Macaca mulatta 2-3 wk. (Hinde, 1971) 3 wk. (Hinde et al., 1964)
Cercocebus albigena 1-2 mo. (Hinde, 1971) 2-3 mo. (Hinde, 1971)
Papio sp. c.3 wk. (Hinde, 1971) 4 wk. (Hinde, 1971)
Papio sp. c. 3 wk. (Hinde, 1971) 4-6 mo. (Hinde, 1971)
Saimiri sciureus IY, mo. (Hinde, 1971) I mo. (Ploog et al., 1967)
Saimiri sciureus IY2 mo. (Hinde, 1971) 2-2Y2 mo. (Baldwin, 1969)
Canis jamiliaris 3-3Yz wk. (Fuller and Fox, 3Y2-4 wk. (Fuller and Fox,
1969) 1969)
Canis latrans 2Y, wk. (Bekoff, personal 4 wk (Bekoff, personal
communication) communication)
Canis lupus 2Y2 wk. (Bekoff, personal 3-3Y2 wk. (Bekoff, personal
communication) communication)
Felis catus 3 wk. (Kling et al., 1969) 3 wk. (West, 1974)
Felis catus 3 wk. (Kling et al. 1969) 5 wk. (complete) (Kling et
al., 1969)
Equus caballus 1 day (Tyler, 1972) I day (Tyler, 1972)
Sciurus carolinensis 3 wk. (Horwich, 1972) 6-8 wk. (Horwich, 1972)
Sciurus vulgaris c. 5 wk. (Horwich, 1972) 5-6 wk. (Horwich, 1972)
Tamiasciurus hudson ius 2Y2 wk. (Horwich, 1972) 5-8 wk. (Horwich, 1972)
Eutamias amoenus 2Y2 wk. (Horwich, 1972) 7 wk. (Horwich, 1972)
Rattus norvegicus 1'/2 wk. (Horwich, 1972) 2 wk. (Horwich, 1972)
Mus musculus 1Y2 wk. (Horwich, 1972) 3-4 wk. (Horwich, 1972)
Dolichotis (Pediolagus) sali- day ("coordinated") I day (Wilson and Kleiman,

nicola (Wilson and Kleiman, 1974)
1974)
Octodontomys gliroides and day ("uncoordinated") 2-3 days (Wilson and
Octodon degus (Wilson and Kleiman, Kleiman, 1974)
1974)
Turdus merula 2Y2 wk. (Messmer and 4 wk. (Messmer and
Messmer, 1956) Messmer, 1956)
Gallus domesticus 1 day (Guh1, 1958) during the first week (Guhl,
1958)
, ,,
,,
200
, ,,
,,
," ..
Gorilla
-,,,,
,..,.. ' _Humar
- ,'Chimpanzee
-
Human
-- -,,
-,. ,,'- Squirrel monkey
- ,-,-
,
, ,/
,
_ " -Rhesus monkey
, ,,
,
, ,,
,
4
, ,,
,
2
, ,,
,
,,
I'
2 3 4 5 6 18910 20 30 4050 1090 200 300 500
60 80 100 400 600
ONSET OF LOCOMOTION (DAYS)

Fig. I. Age of onset of play (ordinate) plotted against average age of onset of the adult loco-
motion (abscissa) in some homoiotherms. Data from Table III. (Averages were computed if a
range of times appeared in the Table, e.g., if play was reported as beginning at age 4-6 mo.,
then a value of 5 mo. was plotted.)
partly relieved of the heat-regulating function they formerly performed."

Slonim (1972) likewise relates onset of play to the development of ther-
moregulation. These arguments further support the evolutionary cost-
benefit argument presented above. However, more data are needed on the
amount and apparent functional significance of skeletal muscle activity in
the early infancy of altricial mammals.
The intrinsic speed of contraction of skeletal muscle is fixed early in
ontogeny for a given muscle group and a given adult body size (Close,
1972). As a result, juvenile animals have muscles whose intrinsic speed is
not appropriate to their current body size, but rather is suited to their adult
204 Robert M. Fagen
BorC
&
~
~
~
~
~
\I)
~~ B
1i;~
~ ~
GCb
.... ~
~,
~Cb
~.~
Cb"::::
Il:l .....
~
t::: C
.~
~
ts
~
:.5?
-..;;
oq:
o B<C. no ploy B>C. play occurs

t
Time Since Birth
Fig. 2. Cost, C(t) (decrease in fitness resulting from diversion of one unit of time or energy
resources from feeding and temperature maintenance) and benefit, 8(t) (fitness increment
resulting from alloc\ltion of one resource unit to play) of play in infancy, plotted as a function
of time t, with t = 0 corresponding to birth or hatching. Play occurs for all times t such that
8(t) > C(t).
body size. For example, an adult rat is not physiologically equivalent to a

young carnivore whose current body size is the same as that of the rat, even
though their body sizes are equal. The adult rat's muscles contract at a
speed appropriate to the functions they perform, while the like-sized young
carnivore's muscles contract much more slowly under the same external
load. The implications of this difference, both for day-to-day functional

performance in the young and for future motor skill development, have not
yet been spelled out in any detail; however, it should be possible to under-
stand many previously unexplained features of juvenile movement patterns
and motor function in the light of this difference. For example, the apparent
relaxed, loose body tone characteristic of juvenile play may stem from the
relatively slow contraction velocity of the muscles being used.
Although little seems to be known regarding ontogeny of play behavior
in birds, the evolutionary and possible physiological considerations stated
above suggest that young precocial birds should begin to play shortly after
hatching (as in the case of the domestic chick: Guhl, 1958), while young al-
tricial birds should start playing soon after the acquisition of endothermy,
or, if Arshavsky's argument is correct, soon after they begin flying. Young
blackbirds (Turdus merula), belonging to an altricial species, learn to fly at
18 days and begin play-chases at 30 days (Messmer and Messmer, 1956);
similar data for other bird species would be of interest.
During infancy in altricial homoiotherms, the heavy demands of feed-
ing, thermoregulation, and, often, competition, leave little time or energy
for play and may also provide sufficient physical exercise. Play fails to oc-
cur in infants even when muscle groups later used in play are already func-
tional, as is the case in piglets, who fight with each other over access to nip-
ples during the first days of life (McBride, 1963; Hafez and Signoret, 1969).
Such early fighting also occurs to some extent in domestic cats (Rosenblatt,
1971) and in hedgehogs (Herter, 1965). Likewise, feeding by nestling chaf-
finches (Fringilla coelebs) may involve vigorous struggles for position
(Newton, 1973). In the species cited, skeletal muscle has clearly developed
to the point of being functional in early infancy, yet the animals cited do
not play until they are older and have begun to walk or fly (e.g., play-
chases of fledgling chaffinches: Marler, 1956).
In most altricial homoiotherms the infant, like the adult, simply works
and rests. In the great apes and man, however, an infant learning to walk
has already been playing for several months. The relatively early onset of
play in these primates suggests that their play behavior may indeed serve
some function or functions other than physical training, since this behavior
begins to appear at a time when skeletal muscle may require no additional
exercise.
C. Play Deprivation and Exercise Deprivation
It has been suggested (Morris, 1964) that the tendency for adult ani-
mals to play in captivity more frequently than in nature stems from a need
206 Robert M. Fagen
for sensory stimulation or drive release. (Thorpe, 1966, finds no evidence

for this tendency; further studies are needed.) It may be hypothesized that
increased play in captive adults results from regression to an immature
ontogenetic stage. Such regression may involve psychological dependence of
the animal on its human captors (a nonevolutionary explanation), or may
constitute manipulation of the humans by the animal: if humans use per-
formance of play activities as a cue indicating an early stage of ontogeny
and, therefore, a need for greater "parental" investment in the animal
exhibiting play behavior, the animal may be selected to play in order to ob-
tain increased benefits from its captors. (I imply an evolutionary explana-
tion of regression here; see Trivers, 1974.) But if certain animal species have
been selected to keep in physical condition by playing when the environment
itself does not provide adequate exercise, specialized psychological events
need not be invoked in order to explain the phenomena described by Morris
(1964) and others.
Extreme confinement, including physical restraint, removes the op-
portunity for physical exercise. One would therefore predict play upon
release from confinement, as is indeed the case (Brownlee, 1954; Chepko,
1971; Cooper, 1942). Incidentally, this argument may help to explain the
differences between the published results of Chepko's (1971) and MUller-
Schwarze's (1971) play deprivation experiments. Chepko deprived domestic
goats (Capra hircus) of play using physical restraint and found a substantial
increase in play after the deprivation period, while MUller-Schwarze (1968)
deprived black-tailed deer (Odocoileus hemionus) of play, but not of other
opportunities for exercise, and found that their tendency to exercise, as
expressed in "longer activity periods, more locomotion, and more explora-
tion" (MUller-Schwarze, 1971), increased as the result of his manipulation.
A more "biological" example of long-term confinement is the winter
rest of certain mammals (e.g., bears and woodchucks). Do physiological de-
training effects occur during such long periods of inactivity? If so, one
would expect play to occur when hibernating mammals and other animals
that are inactive, for the most part, during the winter emerge from rest. I
have observed an instance of vigorous rolling and running in a woodchuck
(Marmota monax) that had recently emerged from a burrow in early spring,
but I cannot state whether this behavior is a general phenomenon during
this season in this species. Krott and Krott (1963) likewise describe vigorous
approach-withdrawal play in young, free-living brown bears (Ursus arctos)
in March following several months of inactivity during winter.
Additional instances of long-term play deprivation are known in mam-
mals. Baldwin and Baldwin (1974) observed no play in two troops of squir-
rel monkeys in 261 hr of observations over 10 weeks. This fact, and other
evidence cited by the Baldwins, indicates that "social organization and
many normal social behaviors can develop in squirrel monkeys without

social play." However, the young squirrel monkeys in these troops appeared
to spend an unusually large amount of time foraging and traveling, and the
physical exercise provided by these activities may have adequately trained
the animals (though perhaps not optimally, as argued above). On the other
hand, as MUller-Schwarze (1971) notes, it is often the case that "young ani-
mals do not play for days or weeks because of disease or bad weather," and
such natural play deprivation will necessarily tend to cause physiological
deterioration if the animals do not exercise adequately in some other
context.
D. Warmup
Attempts to demonstrate physiologically beneficial consequences of so-

called "warmup" (moderate exercise preceding strenuous exercise) have
produced both positive and negative results (Arshavsky, 1972; Barger et al.,
1956; Barnard et al., 1973; Elbel and Mikols, 1972). In healthy humans,
sudden exercise without warmup may produce cardiac ischemia (local blood
deficiency) and abnormal electrocardiographic responses, indicating
potentially hazardous cardiovascular stress; but with prior warmup exercise,
sudden exercise produced no ischemic changes, and ECG responses were
normal in all subjects (Barnard et al., 1973). Play behavior which can be in-
terpreted as "warmup" is not unknown in animals. African wild dogs
(Lycaon pictus) play before starting to forage (Estes and Goddard, 1967;
Kuhme, 1965), as do timber wolves (Mech, 1970) and the choz-choz
(Octodontomys gliro ides) , a South American caviomorph rodent (Wilson
and Kleiman, 1974). If preforaging play serves as "warmup," other less
physiologically meaningful functions need not be postulated for this be-
havior, although there is no evidence at present that such additional func-
tions do not sometimes exist. [Warmup need not involve complex forms of
play behavior. Bumblebees (Heinrich, 1975) and certain Lepidoptera (Kam-
mer, 1968) warm up by exercising their wings vigorously before flying,
but-possibly for historical reasons-such exercise has not been termed
"play. "]
E. Play with Familiar Objects
It has sometimes been suggested that play functions to provide sensory

stimulation, perceptual novelty, and an enriched cognitive environment
(Bruner, 1973; Morris, 1964; Welker, 1961). However, play also occurs
208 Robert M. Fagen
frequently in familiar circumstances with familiar objects. Leyhausen

(l973a) describes favorite toys "owned" by a Fishing Cat (Prionailurus vi-
verrinus) and by some servais (Felis serval). Why should a human child
react negatively to the substitution of a shiny new toy for a familiar old toy
if the sole object of play is stimulus novelty and the recombination of motor
sequences? The perennial popularity of certain toys is understandable,
however, if play objects serve as exercise aids. Indeed, some human athletes
claim that particular sports apparatus and playing fields enable them to
train and perform particularly well.
F. Why Only Homoiotherms?
Except in a very few intriguing cases (Komodo monitor, Varanus

komodoensis: Hill, 1946; fish: Bergmann, 1968; Gunter, 1953; van Iersel
1953; Meder, 1958; Meyer-Holzapfel, 1956) poikilotherms are not known to
play. However, play occurs in many species of birds and mammals. Using a
graphical cost-benefit model of the time and energy costs of feeding in ani-
mals having different metabolic requirements, and assuming that no selec-
tive consequences other than physical training result from play, one may
show that individuals of species having low metabolic requirements will tend
less to be selected to exhibit play behavior than will animals of the same
body size having high metabolic requirements. Briefly, the argument is as
foilows. In a given environment, consider the relative amount of time
needed to satisfy metabolic energy requirements, and the relative number of
different food types taken by a homoiotherm, H, and by a poikilotherm, P,
of equal size. Using a graphical method of analysis derived from Schoener
(1971), it may be demonstrated that H must feed for a longer time than P
and will tend to take a greater variety of food items than will P, including
"less desirable" food items which contain fewer gross calories, which re-
quire a longer time to obtain, or which require more energy to obtain than
do the most desirable items in the diet. These results imply greater selective
pressure for increased physical capacity in homoiotherms than in like-sized
poikilotherms. These differential selective pressures are interpreted as pro-
ducing stronger tendencies to play in homoiotherms than in poikilotherms.
This argument may also be applied to homoiotherm groups having different
characteristic levels of weight-specific metabolism. Basal metabolic rate
(kcaljkg3/4) tends to increase in the series, monotremes: marsupials:
edentates and insectivores: other placentals (Dawson, 1972). Therefore,
marsupials, whose position in the metabolic scale is intermediate between
that of poikilotherms and that of placental mammals, would be expected to
play less than like-sized placentals. In the light of this prediction, it is highly
interesting that while many species of marsupials do exhibit tendencies to

play (Table I), Kaufmann (1974) has noted that play in the whiptailed
wallaby (Macropus parryi) was "conspicuously inconspicuous" compared
with play in primates and carnivores.
Reliable data on poikilotherm play are difficult to locate, but a possi-
ble instance of play has been reported from the Komodo monitor (Varanus
komodoensis) (Hill, 1946). Physiological response to exercise stress in this
large reptile may resemble that of a mammal, since (unlike other reptiles)
monitor lizards may increase their aerobic metabolism during vigorous
exercise (Gordon, 1972). It would be of interest to look for behavior re-
sembling play in the American alligator (Alligator alligator) which, accord-
ing to McIlhenny (1935), exhibits a period, several months in duration, of
postnatal paternal care.
VI. ASPECTS OF PLAY WHICH THE PHYSICAL TRAINING

HYPOTHESIS ALONE DOES NOT EXPLAIN
The physical training hypothesis views play in terms of the known

physiological responses of skeletal muscle, bone, and the cardiopulmonary
system and offers several valid predictions. Yet, as Loizos (1967) points out,
physical tr;:ining alone cannot account for "the range and variety of all play
behavior." In particular, manipulative and social play require further dis-
cussion. After considering these phenomena I will suggest that known
neurobiological training responses may aid in interpreting the behavioral
and functional significance of many forms of play which cannot be
explained by the physical training hypothesis.
A. Manipulative Play
Manipulative play, i.e., persistent, oriented, variable manipulation of

objects, prey animals, or parts of the body, occurring without vigorous
physical exertion and in a nonfunctional context, is found in young primates
(Dolhinow and Bishop, 1970; Hutt, 1970; van Lawick-Goodall, 1968), carni-
vores (Leyhausen, 1973a) ceteceans (Tavolga, 1966; Tavolga and Essapian,
1957), corvids (Gwinner, 1966; Jones and Kamil, 1973), and parrots
(Derschied, 1947; Jackson, 1963b; Low, 1970). In man and in the great
apes, manipulative play first occurs during early infancy (Hurlock, 1972;
JohIison and Medinnus, 1965; van Lawick-Goodall, 1968; Schaller, 1963).
As already mentioned, early onset of play in these species suggests that such
210 Robert M. Fagen
early play activity must serve some function other than physical training.
Possible functions of manipulative play, including motor practice, develop-
ment of tool use and problem-solving strategies, experimentation on the en-
vironment, and discovery of novel behavior patterns, are discussed
elsewhere (Fagen, 1974a,b; Fedigan, 1972; van Lawick-Goodall, 1970;
Leyhausen, 1973a; Jolly, 1966; Sylva, 1974; Tinbergen, 1963).
B. Social Play
Play behavior is very often a social phenomenon. Playing animals typi-

cally orient to, solicit play from, and respond to play soliciting by conspeci-
fics (Bekoff, 1974, 1975; Dolhinow and Bishop, 1970; Hinde, 1971; Loizos,
1966, 1967, 1969; Wilson and Kleiman, 1974). Indeed, friendly, physically
vigorous social interaction offers an ideal environment for physical training,
since bone and muscle are trained "in context" without the risk of serious
injury. But there may be more to social play than physical training. Al-
though some forms of social interaction, particularly agonistic displays,
tend not to occur in the social play of many mammals (see below), and
while maternal status, rather than outcomes of social play, appears to de-
termine dominance rank in young rhesus (Sade, 1967), much other evidence
appears to indicate that social play furnishes a context for social com-
~~rison, i.e., for assessment of one's ability relative to that of agemates
(Dolhinow and Bishop, 1970; Fagen, 1974a). While testing and comparison
of agonistic communication skills seems not to occur in social play (e.g.,
the rhesus monkey: Symons 1974; the domestic cat: Fagen, 1975; the black
bear: Henry and Herrero, 1974), social play appears to be well suited to
testing and comparing developing physical strength, endurance, and skill
without the risks which would accompany serious fighting. Vigorous chas-
ing during courtship in adult animals (Bastock, 1967; Marler, 1956) would
perhaps serve an analogous function, namely, to test the physical condition
(strength and endurance) of the partner before mating. Of course, as men-
tioned above, these social play activities may simultaneously induce training
responses; play behavior need not be viewed as functioning in only one way.
C. Diversive Play
Regular, vigorous "exercise" induces adaptive responses in muscle,

bone, circulation, and respiration. When appropriately and repeatedly
stimulated, the central nervous system likewise exhibits physiologically
measurable training responses both on the cellular and on the biochemical
levels (Ferch min et al., 1975; Greenough, 1974, 1975; Rose, 1969). Detrain-
ing in response to prolonged stimulus deprivation has been reported in the

central nervous system (Valverde, 1967), as one might expect from the
presence of analogous detraining responses in the somatic functions cited.
Since animals must repeatedly and vigorously exercise muscle, bone,
circulation, and respiration in order to develop and maintain these systems,
might not the same be true for the nervous system, at least for those neural
components and structures displaying physiological training responses? If
so, what might constitute an optimal level of stimulation for these
components and structures? To what extent must the animal actively seek
and create its own "enrichment"? How do these requirements vary with
age? Hutt (1970), Morris (1964), Welker (1961), and many others have sug-
gested that play serves to provide increased perceptual stimulation or
increased arousal, but these authors do not specify the physiological
mechanisms whose properties might give rise to the need for such "enrich-
ment." Such hypotheses about play would be more convincing if known,
physiologically defined training responses occurring in structures displaying
nervous activity had been cited. An experimenter could then attempt to
demonstrate that play stimulates such responses, or that detraining occurs
in the absence of play, keeping in mind the fact that training and detraining
responses sometimes require weeks or even months to occur. A necessary
condition for the occurrence of such play-induced responses would be that
play (in nature or in the laboratory) should produce significant increases in
the quantity, quality, diversity, or variability of stimulation. Experiments by
Barnett (1974) suggest that this may be the case in human children. That
such increases would stimulate some component or components of the
nervous system is evident, although the occurrence and magnitude of train-
ing responses would depend on the intensity, duration, and frequency of
stimulation.
The natural history of play behavior is replete with observations sug-
gesting that stimulation-seeking activity may occur in some forms of play.
In general, these observations cannot be explained merely by citing somatic
training, skill development, tool mastery, purposeful but vigorous experi-
mentation on the environment, or social comparison. Hutt (1970) has
termed such play "diversive." Some examples of diversive play are as
follows. Human children who had already thoroughly investigated a novel
object (a rectangular box bearing a lever which, when manipulated, causes a
bell to ring) patted the lever repeatedly, leaned on the lever making the bell
ring continuously while looking around the room, and ran around the ob-
ject, carrying a toy truck and ringing the bell each time the object was
passed (Hutt, 1970). A young vervet monkey (Cercopithecus aethiops) was
observed to leap on and off a lilac-colored feather, place the feather in its
mouth, hop about for several seconds holding the feather in its mouth, and
finally discard the object (Struhsaker, 1967). Chimpanzees often play
212 Robert M. Fagen
together with inanimate objects: they carry and attempt to grasp twigs and
fruits, and one individual ran trailing a palm frond behind him, just out of
reach of a younger sibling (van Lawick-Goodall, 1968). Bottlenose dolphins
(Tursiops truncatus) suspend small objects on water jets, balance fish on
their snouts, transfer objects from fluke to fluke, carry objects and swim
with them (Tavolga, 1966). A four month old domestic kitten was observed
to spring on and grasp a paper wad, lie on his side and kick it, bat it away,
pick it up in his mouth and drop it, dribble it a few feet, pick it up in his
mouth and carry it across the room, drop it into a slipper, and continue to
manipulate and carry the wad to and from the slipper, varying the pattern
of movements several times before lying down next to the slipper (personal
observation). Such behavior appears capricious, frolicsome, surprising, and
delightful to human observers. Judging from human experience, the par-
ticipant may be experiencing similar enjoyment, although this subjective,
essentially "private" aspect of play is hardly subject to scientific analysis
(West, 1974). Perhaps diversive play best merits the term "play." In any
case, this form of behavior appears susceptible to interpretation in the light
of developing knowledge of physiological training responses in the nervous
system.
VII. SUMMARY OF PREDICTIONS
To summarize, the physical training hypothesis offers the following pre-

dictions:
1. Behavioral composition. Exercise play and social play should in-
clude behavior patterns that exercise muscles otherwise used only in
physically vigorous adult activities (e.g., prey-catching, fighting, predator
avoidance, and mating). However, the abundance in play of behavior pat-
terns already having an immediate function in the young, or of behavior
patterns requiring little or no muscular exertion, should not be greater than
the abundance of those patterns in other juvenile behavior. This prediction
is not expected to hold for manipulative play or for diversive play, i.e., play
which appears to serve primarily as a source of perceptual and central
nervous system stimulation.
2. Interruption and repetition. Exercise and social play should be inter-
rupted and repeated. During the periods of interruption, light exercise
should be performed.
3. Types oj exercise play. Two distinct categories of exercise and social
play should be found in animals. The first type (rough-and-tumble play)
should involve maximal, usually static exercise occurring in brief but
intense bouts that are frequently interrupted by rest periods. The second
type (approach-withdrawal play) should involve prolonged bouts of locomo-

tion at submaximal intensity. Mean bout length for rough-and-tumble play
should tend to be less than mean bout length for approach-withdrawal play.
4. Age dependence. Young animals should play more frequently than
adults or infants. Exercise play should begin when endothermy and adult
forms of locomotion appear (e.g., following the development of walking
ability in mammals) and should decrease significantly when growth ceases
in adulthood.
5. Parent-offspring conflict over play. Initially, parent and offspring
should agree to play together. After a certain point in ontogeny, the off-
spring will continue to solicit play from its parent, but the parent will
respond less and less frequently.
6. Play after winter rest or after confinement. Mammals that are
physically inactive for long periods should exhibit exercise play following
periods of inactivity. This prediction applies in particular to hibernators, to
other winter-resting species, and to animals that have been physically
confined.
7. Phylogenetic distribution. Homoiotherms, but not poikilotherms,
should exhibit exercise play. Within the homoiotherms, exercise play should
be more prominent in eutherians than in like-sized marsupials, and should
be more prominent in "advanced" eutherians than in like-sized "primitive"
eutherians (edentates and insectivores).
VIII. ACKNOWLEDGMENTS
I thank R. L. Terjung, H. K. Knuttgen, and C. R. Taylor for introduc-

ing me to the exercise physiology literature, and M. Bekoff, S. Wilson, and
R. L. Terjung for their comments on earlier drafts. The model of play in
animals with different metabolic levels is based on discussions with C. R.
Taylor, to whom I am greatly indebted for his suggestion that I consider the
behavioral implications of recent exercise physiology research. Responsi-
bility for interpretations and possible misinterpretations of the physiological
literature is mine.
IX. REFERENCES
Arshavsky, l. A. (1972). Musculoskeletal activity and rate of entropy in mammals. Adv.

Psychobiol. I: I-52.
Ashmole, N. P., and Tovar, S. H. (1968). Prolonged parental care in Royal Terns and other
birds. Auk 85:90-100.
214 Robert M. Fagen
Astrand, P. -0., and Rodahl, K. (1970). Textbook of work physiology, McGraw-Hill, New
York.
Baldwin, J. D. (1969). The ontogeny of social behavior of squirrel monkeys, Saimiri sciureus.
in a seminatural environment. Folia Primatol. 11:35-79.
Baldwin, J. D., and Baldwin, J. I. (1974). Exploration and social play in squirrel monkeys
(Saimiri). A mer. Zool. 14:303-315.
Ball, S. D. (1943). Chimney Swifts at play? Auk 60:269-270.
Barash, D. P. (1973). The social biology of the Olympic marmot. Animal Behavior
Monographs 6:171-245.
Barger, A. C., Richards, V., Metcalfe, J., and Gunther, B. (1956). Regulation of the circula-
tion during exercise. A mer. J. Physiol. 184:613-623.
Barnard, R. J., Macalpin, R., Kattus, A., and Buckberg, G. (1973). Ischemic response to sud-
den strenuous exercise in healthy men. Circulation 48:936-942.
Barnett, L. (1974). An information processing model of children's play. M. S. Thesis,
University of Illinois at Urbana-Champaign.
Basset, C. A. L. (1972). Biophysical principles affecting bone structure. In Bourne, G. (ed.),
The Biochemistry and Physiology of Bone, 2nd ed., Academic Press, New York.
Bastock, M. (1967). Courtship, Heinemann, London.
Battersby, E. (1944). Do young birds play? Ibis 86:225.
Beach, F. A. (1945). Current concepts of play in animals. Amer. Natur. 79:523-541.
Beck, B. (1975). Primate tool behavior. In Tuttle, R. (ed.), Antecedents of Man and After. III:
Primate Socioecology and Psychology, Mouton, The Hague.
Bekoff, M. (1974). Social play and play-soliciting by infant canids. Amer. Zool. 14:323-340.
Bekoff, M. (1975). Animal play: Problems and perspectives. (This volume.)
Bergmann, H. (1968). Eine deskriptive Verhaltensanalyse des Segelflossers (Pterophyllum
scalare Cuv. & Val., Cichilidae, Pisces). Z. Tierpsych. 25:559-587.
Bloor, C. M., and Leon, A. S. (1970). Interaction of age and exercise on the heart and its
blood supply. Lab. Invest. 22:160-165.
Brosset, A. (1973). Etude comparative de I'ontogenese des comportements chez les Rapaces
Accipitrides et Falconides. Z. Tierpsych. 32:386-417.
Brownlee, A. (1954). Play in domestic cattle: an analysis of its nature. Brit. Vet. J. 110:48-68.
Bruner, J. S. (1973). Organization of early skilled action. Child Dev. 44:1-11.
Burleigh, I. G. (1974). On the cellular regulation of growth and development in skeletal
muscle. BioI. Rev. 49:267-320.
Chepko, B. D. (1971). A preliminary study of the effects of play deprivation on young goats. Z.
Tierpsych.28:517-526.
Close, R. I. (1972). Dynamic properties of mammalian skeletal muscles. Physiol. Revs.
52: 129-197.
Collier, G. (1970). Work: a weak reinforcer. Trans N.Y. Acad. Sci. 32:557-576.
Cooper, J. B. (1942). An exploratory study of African lions. Compo Psychol. Monogr. 17:1-48.
Darling, F. F. (1937). A Herd of Red Deer, Oxford University Press, London.
Dawson, T. J. (1972). Primitive mammals and patterns in the evolution of thermoregulation.
In Bligh, J., and Moore, R. (eds.), Essays on Temperature Regulation, North-Holland,
Amsterdam, and London, and American Elsevier, N.Y.
Deitrick, J. E., Whedon, G. D., and Short, E. (1948). Effects of immobilization upon various
metabolic and physiologic functions of normal men. Am. J. Med. 4:3-26.
Derscheid, J. M. (1947). Strange parrots. A vicultural Magazine 59:44-49.
Dimelow, E. J. (1963). The behavior of the hedgehog, Erinaceus europaeus L., in the routine of
life in captivity. Proc. Zool. Soc. Lond. 141:281-289.
Dolinhow, P. J., and Bishop, N. (1970). The development of motor skills and social rela-
tionships among primates through play. Minnesota Symposia on Child Psychology 4:141-
198.
Donaldson, C. L., Hulley, S. B., Vogel, J. M., Hattner, R. S., Bayers, J. H., and McMillan, D.
(1970). Effect of prolonged bed rest on bone mineral. Metabolism 19:1071-1084.
Eaton, R. L. (1969). Notes on breathing rates in wild cheetahs. Mammalia 33:543-544.
Eibl-Eibesfeldt, I. (1950). Beitrage zur Biologie der Haus- und der Ahrenmaus nebst einigen
Boebachtungen an anderen Nagern. Z. Tierpsych.7:558-587.
Ekblom, B. (1969a). Effect of physical training in adolescent boys. J. Appl. Physiol. 27:350-
355.
Ekblom, B. (1969b). Effect of physical training on oxygen transport in man. Acta Physiol.
Scand. Supp. 328.
Elbel, E. R., and Mikols, W. J. (1972). The effects of passive or active warm-up upon certain
physiological measures. Int. Z. Angew. Physiol. 31:41-52.
Eriksson, B. (1972). Physical training, oxygen supply and muscle metabolism in 11-13 year old
boys. Acta Physiol. Scand. Supp. 384.
Estes, R. D. and J. Goddard. (1967). Prey selection and hunting behavior of the African wild
dog.J. Wildl. Man. 31:52-70.
Ewer, R. F. (1968). Ethology oj Mammals, Plenum, New York.
Fagen, R. (1974a). Selective and evolutionary aspects of animal play. Amer. Natur. 108:850-
858.
Fagen, R. (1974b). Theoretical bases for the evolution of play in animals. Ph.D. diss., Harvard
Univ.
Fagen, R. (1976). Selection for optimal age-dependent schedules of play behavior. A mer. Natur.
In press.
Fedigan, L. (1972). Social and solitary play in a colony of vervet monkeys (Cercopithecus ae-
thiops). Primates 13:347-364.
Ferchmin, P. A., Bennett, E. L., and Rosenzweig, M. R. (1975). Direct contact with enriched en-
vironmt:ut is required to alter cerebral weights in rats. J. Compo Physiol. Psych. 88:360-367.
Fradrich, H. (1965). Zur Biologie und Ethologie des Warzenschweines (Phacocheorus ae-
thiopicus Pallas), unter Berucksichtigung des Verhalten anderer Suiden. Z. Tierpsychol.
22:328-393.
Fradrich, B., and Thenius, E. (1972). Tapirs. In Grizmek, B. (ed.) Animal Life Encyclopedia,
Vol. 13, (Mammals IV), Van Nostrand Reinhold, New York.
Fuller, J. L., and Fox, M. W. (1969). The behavior of dogs. In Hafez, E. S. E. (ed.), The Be-
havior oj Domestic Animals, 2nd ed., Balliere, Tindall and Cassell, London.
Gibson-Hill, C. A. (1947). Notes on the birds of Christmas Island. Raffles Mus., Singapore,
Bull. 18:87-165.
Goldspink, G. (1970). Morphological adaptation due to giOwth and activity. In Briskey, E. J.,
Cassens, R. A., and Marsh, B. B. (eds.), The Physiology and Biochemistry oj Muscle as a
Food, 2, University of Wisconsin Press, Madison.
Gordon, M. S. (1972). Animal Physiology: Principles and Adaptations, 2nd ed., Macmillan,
New York.
Greenough, W. T. (1974). Enduring brain effects of differential experience and training. Paper
prepared for Conference on Current Research Approaches to the Neural Mechanisms of
Learning and Memory, Asilomar, California.
Greenough, W. T. (1975). Environmental effects on brain connectivity and behavior. In
Mostofsky, D. (ed.), Behavior Control and Modification of Physiological Activity.
Groos, K. (1896). Die Spiele der Tiere, Fischer, Jena. (The play of animals, 1898, Appleton,
New York.)
Grzimek, B. (1967). Four-legged Australians, Collins, London.
216 Robert M. Fagen
Guhl, A. M. (1958). The development of social organization in the domestic chick. Anim.
Behav. 6:92-111.
Gunter, G. (1953). Observations on fish turning flips over a line. Copeia 3:188-190.
Gwinner, E. (1966). Uber einige Bewegungsspiele des Kolkraben (Corvus corax L.) Z.
Tierpsych.23:28-36.
Hafez, E. S. E., and Signoret, J. P. (1969). The behaviour of swine. In E. S. E. Hafez (ed.), The
Behaviour oj Domestic Animals. BaIliere, Tindall, and Cassell, London.
Hansen, J. W. (1967). Effect of dynamic training on the isometric endurance of the elbow
flexors. Intern. Z. Angew. Physiol. 23:367-370.
Harlow, H. F., and Harlow, M. K. (1965). The affectional systems. In Schreier, A. M.,
Harlow, H. J., and Stollnitz, F. (eds.), Behavior oj Nonhuman Primates. Vol. 2,
Academic Press, New York and London.
Hartley, L. H., Grimby, G., Kilborn, A., Nilsson, H. J., Astrand, I., Bjure, J., Ekblom, B., and
Saltin, B. (1969). Physical training in sedentary middle-aged and older man. III. Cardiac
output and gas exchange at submaximal and maximal exercise. Scan. J. Clin. Lab. Invest.
24:335-344.
Hediger, H. (1958). Verhalten der Beuteltiere (Maruspiala). Handb. d. Zool. 8:10 (9), 1-28.
Heinrich, B. (1975). Thermoregulation in bumblebees. II. Energetics of warm-up and free
flight.J. Compo Physiol. 96:155-166.
Heinsohn, G. E. (1966). Ecology and reproduction of the Tasmanian bandicoots. Univ. Cal.
Pub. Zool. 80.
Henry, J. D., and Herrero, S. M. (1974). Social play in the American black bear. Amer. Zool.
14:371-389.
Herter, K. (1965). Hedgehogs. Phoenix House, London.
Hill, C. (1946). Playtime at the zoo. Zoo-LiJe 1:24-26.
Hinde, R. A. (1971). DevcJopment of social behavior. In Schreier, A. M., and Stollnitz, F.
(eds.), Behavior oj Nonhuman Primates. Vol. 3, Academic Press, New York and London.
hinde, R. A., Rowell, R. E., and Spencer-Booth, Y. (1964). Behavior of socially living rhesus
monkeys in their first six months. Proc. Zool. Soc. Lond. 143:609-649.
H0jgaard, A. M. (1954). Nogle Iagttagelser over en tam Gr0nlansk Ravn. (Corvus corax princi-
palis). Dansk om. Foren Tiddskr. 48:38-47. (Cited in Thorpe, 1966).
Holloszy, J. O. (1973). Biochemical adaptations to exercise: aerobic metabolism. Exercise and
Sport Sciences Reviews 1:45-71.
Horwich, R. (1972). The ontogeny of social behavior in the Grey Squirrel (Sciurus
carolinensis). Z. Tierpsych. Supp. 8.
Hurlock, E. B. (1972). Child Development. 5th ed., McGraw-Hill, New York.
Hutt, C. (1970). Specific and diversive exploration. Adv. Child Dev. and Behav. 5:119-180.
Jackson, J. R. (1963a). Nesting of Keas. Notornis 10:319-326.
Jackson, J. R. (1963b). Studies at a Kaka's nest. Notornis 10:168-176.
Johnson, R. c., an.d Medinnus, G. R. (1965). Child Psychology: Behavior and Development.
Wiley, New York.
Jolly, A. (1966). Lemur social behavior and primate intelligence. Science 153:501-506.
Jones, T. B., and Kamil, A. C. (1973). Tool-making and tool-using in the Northern blue jay.
Science 180:1076-7.
Kammer, A. (1968). Motor patterns during flight and warm-up in lepidoptera. J. Exp. Bioi.
48:89-109.
Kaufmann, J. H. (1974). Social ethology of the whiptail wallaby, Macropus parryi. in
northeastern New South Wales. Anim. Behav. 22:281-369.
Keul, J., Doll, E., and Keppler, D. (1972). Energy Metabolism oj Human Muscle (Med. &
Sport 7), University Park Press, Baltimore, London and Tokyo.
Kilham, L. (1974). Play in hairy, downy and other woodpeckers. Wilson Bull. 86:35-42.
Kling, A., Kovach, J. K., and Tucker, T. J. (1969). The behaviour of cats. In E. S. E. Hafez (ed.),
The Behaviour of Domestic Animals, Balliere, Tindall, and Cassell, London.
Krott, P., and Krott, G. (1963). Zum Verhalten des Braunbaren (Ursus arctos L. 1758) in den
Alpen. Z. Tierpsych. 20:160-206.
Kruuk, H. (1972). The Spotted Hyena, Univ. of Chicago Press, Chicago.
Kuhme, W. (1965). Freilandstudien zur Soziologie des Hyanenhundes. Z. Tierpsych. 22:495-
541.
Leyhausen, P. (1973a). Verhaltensstudien an Katzen. Z. Tierpsych. Supp. 2, 3rd ed. Paul
Parey, Berlin and Wartburg.
Leyhausen, P. (1973b). On the function of the relative hierarchy of moods (as exemplified by
the phylogenetic and ontogenetic development of prey-catching in carnivores). In Lorenz,
K., and Leyhausen, P., Motivation of Human and Animal Behavior, Van Nostrand Rein-
hold, New York.
Loizos, C. (1966). Play in mammals. Symp. Zoo I. Soc. Lond. 18:1-9.
Loizos, C. (1967). Play behaviour in higher primates: a review. In Morris, D. (ed.), Primate
Ethology, Aldine, Chicago.
Loizos, C. (1969). An ethological study of chimpanzee play. In Carpenter, C. R. (ed.),
Proceedings of the Second International Congress on Primatology, Atlanta, Georgia, Vol.
I. Karger, New York.
Low, R. (1970). The Massena's Parrot. Avicultural Mag. 76:153-4.
Marler, P. (1956). Behaviour of the chaffinch, Fringilla coelebs. Behaviour Supp. 5.
McBride, G. (1963). The "teat order" and communication in young pigs. Anim. Behav. II :53-
56.
McIlhenny, E. A. (1935). The Alligator's Life History, Christopher Publishing House, Boston.
Mech, L. D. (1970). The Wolf: The Ecology and Behavior of an Endangered Species, Natural
History Press, Garden City, New York.
Meder, E. (l ;58). Gnathonemus petersii (Gunther). Z. Vivaristik (Mannheim) 4: 161-171.
Meier, G. W., and Devanney, V. D. (1974). The ontogeny of play within a society: preliminary
analysis. Amer. Zool. 14:289-294.
Messmer, E., and Messmer, I. (1956). Die Entwicklung der Lautausserungen und einiger
VerhaltensweiseI;! der Amsel (Turdus merula merula L.) unter naturlichen Bedingungen
und nach Einzelaufzucht in schalldichten Raumen. Z. Tierpsych. 13:341-441.
Meyer-Holzapfel, M. (1956). Das Spiel bei Saugetieren. Handb. d. Zool., Berl., 8 (10) No.5,
1-36.
Moreau, R. E., and Moreau, W. M. (1944). Do young birds play? Ibis 86:93-4.
Morris, D. (1964). The response of animals to a restricted environment. Symp. Zool. Soc. Land.
13:99-120.
MUller-Schwarze, D. (1968). Play deprivation in deer. Behaviour 31: 144-162.
MUller-Schwarze, D. (1971). Ludic behavior in young mammals. In Sterman, M. B., McGinty,
D. J., and Adinolfi, A. M. (eds.), Brain Development and Behavior, Academic Press, New
York.
Newton, I. (1973). Finches, Taplinger, New York.
Ploog, D., Hopf., S., and Winter, P. (1967). Ontogenese des VerhaItens von Totenkopf-Affen
(Saimiri sciureus). Psych. Forschung 31: 1-41.
Poole, T. B. (1966). Aggressive play in polecats. Symp. Zool. Soc. Lond. 18:23-44.
Poole, T. B. (1967). Aspects of aggressive behaviour in polecats. Z. Tierpsych. 24:351-369.
Porter, S. (1947). The breeding of the Kea. Avicultural Mag. 53:50-55.
Poupa, 0., Rakusan, K., and Ostadal, B. (1970). The effect of physical activity upon the heart
of vertebrates. Medicine and Sport 4:202-233.
218 Robert M. Fagen
Rarick, G. L., and Larsen, G. L. (1959). The effect of variations in the intensity and frequency
of isometric muscular effort on the development of statis muscular strength in pre-
pubescent males. Int. Z. Angew. Physiol. 18:13-21.
Rasa, O. A. E. (1971). Social interaction and object manipulation in weaned pups of the
Northern elephant seal Mirounga angustirostris. Z. Tierpsych. 29:82-102.
Rasa, O. A. E. (1973). Prey capture, feeding techniques, and their ontogeny in the African
dwarf mongoose, Helogale undulata rujula. Z. Tierpsych. 32:449-488.
Redican, W. K., and Mitchell, G. (1974). Play between adult male and infant rhesus monkeys.
Amer. Zool. 14:295-302.
Rose, S. P. R. (1969). Neurochemical correlates of learning and environmental change. FEBS
Letters 5:305-312.
Rosenblatt, J. S. (1971). Suckling and home orientation in the kitten: a comparative develop-
mental study. Pages 345-410. In Tobach, E., Aronson, L. R., and Shaw, E. (eds.), The
Biopsychology oj Development, Academic Press, New York.
Rosenblatt, J. S., Turkewitz, G., and Schneirla, T. C. (1962). Development of suckling and re-
lated behavior in neonate kittens. In Bliss, E. L. (ed.), Roots oj Behavior, Harper & Bros.,
New York.
Rowell, L. B. (1974). Human cardiovascular adjustments to exercise and thermal stress.
Physiol. Revs. 54:75-159.
Sade, D. S. (1967). Determinants of dominance in a group of free-ranging rhesus monkeys. In
Altmann, S. A. (ed.), Social Communication among Primates, University of Chicago
Press, Chicago.
Saltin, B., Biomqvist, G., Mitchell, J. H., Johnson, R. L. Jr., Wildenthal, K., and Chapman,
C. B. (1968). Response to submaximal and maximal exercise after bed rest and training.
Circulation 38, Suppl. 7.
Schaller, G. B. (1963). The Mountain Gorilla, University of Chicago Press, Chicago.
Schmidt, U., and Manske, U. (1973). Die Jugendentwicklung. der Vampirfledermause (Des-
modus rotundus). Z. Saugetierk, 38:14-33.
Schoener, T. W. (1971). Theory of feeding strategies. Ann. Rev. Ecol. Systemat. 2:369-404.
Sikes, S. K. (1971). Natural History oj the Ajrican Elephant, American Elsevier, New York.
Sionim, A. D. 1972. Ecophysiologic studies of innate behavior in mammals. Int. J. Psychobiol.
2:37-64.
Stonehouse, B., and Stonehouse, S. (1963). The frigate bird Fregata aquila of Ascension Is-
land. Ibis 103b:409-422.
Struhsaker, T. T. (1967). Behavior of vervet monkeys (Cercopithecus aethiops). Univ. oj Calif.
Pub. in Zool. 82.
Swynnerton, C. F. M. (1908). Further notes on the birds of Gazaland. Ibis 2:391-443.
Sylva, K. (1974). The relationship between play and problem-solving in children 3-5 years old.
Ph.D. diss., Harvard Univ.
Symons, D. (1974). Aggressive play and communication in rhesus monkeys (Macaca mulatta).
Amer. Zool. 14:317-322.
Tavolga, M. C. (1966). Behavior of the bottle-nose dolphin (Tursiops truncatus): Social
interactions in a captive colony. In Norris, K. S. (ed.), Whales, Dolphins and Porpoises,
University of California Press, Berkeley and Los Angeles.
Tavolga, M. C., and Essapian, F. S. (1957). The behavior of the bottle-nosed dolphin (Tursiops
truncatus): Mating, pregnancy, parturition and mother-infant behavior. Zoologica 42:11-
31.
Terjung, R. L., Winder, W. W., Baldwin, K. M., and Holloszy, J. O. (1973). Effect of exercise
on the turnover of cytochrome c in skeletal muscle. J. Bioi. Chem. 248:7404-7406.
Tharp, G. D., and Buuck, R. J. (1974). Adrenal adaptation to chronic exercise. J. Appl.
Physiol. 37 :720-722.
Thorpe, W. H. (1963). Learning and Instinct in Animals, 2nd ed. (1st. ed., 1956), Methuen,
London.
Thorpe, W. H. (1966). Ritualization in ontogeny: I. Animal Play. Phil. Trans. Royal Society.
Ser. B, 251:311-319.
Tinbergen, N. (1963). On aims and methods of Ethology. Z. Tierpsych. 20:410-433.
Tomanek, R. J. (1970). Effects of age and exercise on the extent of the myocardial capillary bed.
Anat. Rec. 167:55-62.
Trivers, R. L. (1974). Parent-offspring conflict. Amer. Zool. 14:247-262.
Troughton, E. (1941). Furred Mammals oj Australia, Angus and Robertson, Sydney and
London.
Tyler, S. J. (1972). The behaviour and social organization of the New Forest ponies. Anim.
Behav. Monog. 5:85-196.
Valverde, F. (1967). Apical dendritic spines of the visual cortex and light deprivation in the
mouse. Exper. Brain Res. 3:337-352.
van Iersel, J. J. A. (1953). An analysis of the parental behavior of the male three-spined
stickleback. Behav. Supp. 3:1-159.
van Lawick, H., and van Lawick-Goodall, J. (1970). Innocent Killers, Collins, London.
van Lawick-Goodall, J. (\970). Tool-using in primates and other vertebrates. Adv. Study Behav.
3:195-249.
van Lawick-Goodall, J. (\968). The behaviour of free-living chimpanzees in the Gombe Stream
Reserve. Anim. Behav. Monog. 1:161-311.
Wagner, H. O. (1954). Versuch einer Analyse der Kolibribalz. A. Tierpsych. 1l:182-212.
Welker, W. I. (1961). An analysis of exploratory and play behavior in animals. In Fiske, D.
W., and Maddi, S. R. (eds.) Functions oj Varied Experience, Dorsey, Homewood, Ill.
Wemmer, C., and Fleming, M. J. (1974). Ontogeny of playful contact in a social mongoose,
the mec:kat, Suricata suricatta. Amer. Zool. 14:415-426.
West, M. (1974). Social play in the domestic cat. A mer. Zool. 14:427-436.
Whedon, G. D., Lutwak, L., Rambaut, P. C., Whittle, M. W., Smith, M. C., Reid, J., Lynch,
C., Stadler, C. R., and Stanford, D. D. (ms.) Mineral and nitrogen metabolic studies, ex-
periment M071. NASA Technical Report.
Wilson, E. O. (197\). Competitive and aggressive behavior. In Eisenberg, J. F., and Dillon, W.
(eds.), Man and Beast: Comparative Social Behavior. Smithsonian Institution Press,
Washington.
Wilson, S., and Kleiman, D. (1974). Eliciting play: a comparative study. Amer. Zool. 14:341-
370.
Wunsch mann, A. (1966). Einige Gefangenschaftsbeobachtungen am Breitstirn-Wombats
(Lasiorhinus latifrons Owen 1945). Z. Tierpsychol. 23:56-71.
Chapter 6
MY DESCENT FROM THE MONKEY

Department of Psychology
Clark University
Worcester, Massachusetts
I want to explain why a credentialed specialist in primate social behavior

should have wanted to become an ornithologist, just when the Age of Pri-
mates was at its height. Could it have been only an obstinate inclination to
be always where the action isn't, or could it be that primatology ceased to
generate the kind of intellectual action that makes for sustained progress in
a scientific domain?
A few of my readers may be primatologists of one sort or another and,
as such, may have written funding requests to granting agencies during the
early 60's. If so, you probably wrote that dreaded paragraph which
explained to the review panel why it should put aside 20 inexpensive studies
of local and/or easily maintained species, such as the rat, the frog, the
mourning dove, the German shepherd, or the American housewife, in favor
of your primate study, which demanded either a Landrover and a ticket for
eight to Nairobi or 20 custom-made stainless steel cages and an air-condi-
tioned laboratory lavishly outfitted with equipment for maintaining exotic
animals. In such paragraphs one traditionally argued that the monkey
should be studied-despite the exorbitant costs-because the monkey is
closer to man. In the course of funding my graduate career I wrote many
such paragraphs, and each made me more uncomfortable than the one
before. The logic was obviously devious. If "closeness to man" was such a
certain warranty of theoretical importance, then why not study man
himself? I always thought that one day I would be punished for these hypoc-
risies, but never did I realize how fitting and devastating the punishm~nts
would be.
I was, by coincidence of geography and age, one of the first graduate
students to specifically train himself for a career studying the social be-
221
222 Nicholas S. Thompson
havior of primates. I believed deeply in field study, but because of an irra-

tional fear of leeches and lions, I set about doing a laboratory study in
partial satisfaction of the requirements for a Ph.D. in Comparative
Psychology. For three years, off and on, I and my monkeys performed
together in the beautifully appointed, windowless basement laboratories of
Tolman Hall at the University of California at Berkeley. My days were
spent from nine to five testing and filming monkeys, while my evenings were
devoted to coding and organizing the hundreds of thousands of frames of
timelapse film which my labors were generating.
From the horror of those days, I remember almost nothing about
simian social organization. I remember the constant battle to keep man,
monkey, and environment in their separate places. I remember the
paraphernalia of sanitation: the stainless steel buckets, the steam hoses
which behaved like sulky snakes when one tried to coil them, and the anti-
T.B. foot tray which exuded a pleasant odor of iodine but which left one's
foot faintly wet for hours. I remember the virus B. anxiety, helped along by
occasional ambiguous reports of deaths at other laboratories, deaths from
monkey bite, or monkey scratches, or monkey spit, or perhaps even just
monkey breath. Each of these hazards required a piece of garb to protect
the investigator, so that the official standard outfit in our laboratory
consisted of boots, coveralls, rubber gloves, paper caps, face masks, and
goggles. Each worker in the laboratory omitted some part of this armor,
hqsed on a personal calculation which included as variables how much he
hated the particular item and how recently he had read of a Monkey B.
death. The goggles were a favorite omission because they fogged on the in-
side from one's breath and on the outside from the steam hoses. Those of us
who omitted this particular item developed a covert, subconscious squint,
which we deployed in the presence of monkeys to guard us against a lethal
dose of monkey spit. Thus did we who had pledged our professional lives to
understanding monkeys spend the majority of our energies in isolating
ourselves from monkeys.
What is significant about this period is that I came to hate monkeys.
These were not the lush-furred, loving citizens of DeVore's or Carpenter's
troops; these were incorrigible prisoners, and I was both warden and sole
guard. All the while I worked with my subjects, I was promising myself
never again to look at a confined monkey; since leeches and lions still held
their terrible power over me, I had to recognize that I was promising myself
never again to look at a monkey. Readers may conclude at this point that
my reasons for leaving primate social behavior were too personal to be of
general relevance. But personal reasons have a way of gathering logic, and
many a scientist has been forced to make observations of catholic im-
portance because of a personal bind. The trivial decision to change from
My Descent from the Monkey 223
monkeys to something else was not interesting in itself, but rather the reac-
tion this decision provoked in those with whom I subsequently had dealings.
I took my degree and went looking for a job as a comparative
psychologist. On every side my disclaimers of interest in primates were
greeted with disbelief. Like every young primate specialist in those happy
days of a lush job market, I was courted by universities hoping to start pri-
mate laboratories and thus partake in the seemingly limitless flow of funds
that was to follow. Over coffee and brandy in the new faculty clubs of many
a burgeoning campus, provosts and chairmen inquired solicitously after my
needs. How many square feet for my laboratories? How many technicians?
How many secretaries would be needed to type the myriad grant requests
that I would undoubtably be filing? And best of all, (sotto voce) How did I
project my initial publication thrust? Graduate students of the 1970's may
think this discussion self serving, but only because you cannot imagine how
inflated the stocks of primate social behavior were in the 60's. No sane
man-particularly an accredited primatologist-would forsake the monkey
for a lesser species. What did I have to offer instead? came the ultimate in-
credulous query. In desperation, I grasped at the crow. At that time, I
had learned only two facts about crows. At the age of 23, I had learned from
a graduate student colleague that crows had an interesting vocal system. At
the age of 10, I had learned that you can't shoot a crow with an air rifle. For a
desperate man, these two pieces of information were enough.
Desperate choices-like personal reasons-have a way of gathering
logic. You can no doubt imagine the consternation of my interviewers when
I announced that, no, I wouldn't need technicians, no, I wouldn't need help
preparing a grant proposal, NO, all I would require, thank you, was a good
pair of binoculars (8 X 35 or 7 x 25 would do) and a few years to think. I
began to make this announcement later and later in the interview, since I
began to sense a high correlation between those interviews in which I men-
tioned crows early and those interviews in which I found myself back at the
airport without the customary dinner and brandy at the faculty club.
But time was passing. In that year in which major universities were
fighting over M.A.'s, I with my Ph.D. was not finding a job. I decided that
early confession of my conversion to crows was essential: at the worst, it
might save me unnecessary travel. At best, it might favorably dispose
chairmen and provosts toward me, if not for my wisdom, at least for my
forthrightness. The first letter in this program went to a department
chairman at the major campus of a growing eastern university, who con-
cluded that I was pulling his leg. He called me long distance, with several
colleagues lurking on extensions to check me out. They rang up bright and
early one morning all full of what I later came to identify as the bustle and
good cheer of an eastern university on a rare sunny morning. In their
consternation, they forgot about the time differential between the east and
west coasts. Thus the call arrived in the predawn of an all-too-typical gray,
foggy northern California morning. I remember very little about the con-
versation which ensued. My wife tells me that she woke up, gradually and
painfully, to see her husband, naked and silhouetted against the chilly dank-
ness, phone in hand, touting the merits of the American crow as a subject for
coordinated field and laboratory research.
I finally took a temporary position as a comparative psychologist at a
small liberal arts college. Was I sure that I didn't want space for monkeys?
I thanked them for their concern, borrowed some binoculars from the
biology department, and entertained myself by following crows about the
woods. The job market cooled. I shamelessly published a crow paper.
Provosts stopped calling me. My mail consisted entirely of entreaties for
payment from Brill. I had become (as one of my colleagues put it) a
zoologist.
In the peaceful hours which followed, I had ample opportunity to
repent the "closeness of man" argument with which I had so often financed
my monkey researches. If I had been funded in 1963, '64, '65, and '66 to
study monkeys because they were so "close to man," how could I now in
good conscience proceed to study so· "distant" a member of the animal
kingdom as the crow? One solution to this quandry was to argue that the
crow is closer to man than the monkey. I used to require a drink in my
hand in order to make this argument, but now I can do it cold sober. It en-
taIls, as you can imagine, a rather elaborate redefinition of the word
"close."
When people use the word "close" in this sense they normally have in
mind a phylogenetic tree. Two animals are said to be "close" when they are
on offshoots of the same twig, "far" when they are on different major
limbs. The phylogenetic tree is basically a classificatory device. Things are
"close" or "distant" in the tree because of the degree to which they share
similarities. Thus, man and monkeys are on the same twig because of the
vast numbers of anatomical similarities which they share, but they are on
separate twiglets because of some important behavioral and morphological
differences. Morphologically, the evolution of man involves some novelties
in the hind limb, pelvis, and brain; behaviorally, man displays a communi-
cation system which has no parallels among other primates and a social or-
ganization which is unique in its complexity. Man's communication system
is useful for conveying detailed information between two individuals with
different experiences. His social organizations display the constantly shift-
ing social groupings which would make such information exchanges valu-
able. Two hunting parties divide up and go in different directions; at night-
fall they meet again and exchange notes. Hunting is good to the east, bad to
the west. The next day, both parties exploit the better site. When they return
to home that night, they are able to relate in detail their experiences to all
who will listen and learn.
In comparison with human language, monkey communication systems
seem to be most useful for conveying immanent social intentions. Messages
such as "I am going to hit you," "I am going to mount you," "I am ready
to be groomed," and "I am going to run away," predominate in monkey
communication. Monkey groupings ordinarily make an exchange of in-
formation about the outside world unnecessary. Most sorts of monkeys live
out their days in stable, closely knit social groupings with a minimum of
coming and going. Messages about the environment need only consist of
such directives as "look there," or "attention, danger," or "danger on the
ground." The vocalizations of most monkeys consist in rather unelaborate
graded series apparently not far from the mental capacity of the average
dog to decipher and emit. In fact, nothing about the complexity or sophisti-
cation of macaque vocalization or social organization would exclude from
the alpha male role an intelligent collie dog.
Crows, on the other hand, display many characteristics of communica-
tion and social organization which, on first impression, seem to mimic their
human equivalents. The social organization seems to shift constantly
between large groups and small. In one day crows can be seen alone, in
family groups, in owl mobs of 50 to 100 individuals, and in roosts of several
thousand. The transitions between these various groupings are accompanied
by incredibly complex vocalizations, ranging from yips and coos to growls,
rattles, and sniggers. These sounds seem to be organized at two levels,
analogous to syllables and words in human speech. Some of the vocaliza-
tions do not bear any obvious relationship to immediate social events, and
thus are possible exchanges of information about circumstances displaced in
space or in time. Thus, while nobody would argue that the crow is close in
morphology and general behavior, I might argue that the crow is "close" in
those features which seem to make man special. The crow may be "close"
where it counts.
But the Zeitgeist was against crows. The 1960's were the Age of the In-
nate Depravity Book. This phrase was applied by Ashley Montague to a
class of books which traced man's griefs to his relationship with animals,
and particularly primates. Over the last decade, dozens of books and
magazine articles on this theme have appeared, as one by one noted-and
not so noted-ethologists, anthropologists, sociologists, and just plain in-
tellectual dilettantes cashed in. The common conceptual thread in these
books was the idea that man's true animal nature is covered by a veil of
civilization. If we are truly to know man, we must not study him but rather
his poor relations who wear their true natures on their hairy sleeves. This
argument is the latency gimmick, used to protect a false principle from the
facts which would disprove it. The manifest innocence of the accused
heretic is seen as evidence for the wiliness of the devil; the manifest sexless-
ness of a dream is seen as evidence for the power of the repressive forces
against sexuality; the manifest loyalty of the State Department official is
seen as evidence for the guile of Communism; and the manifest civility of
most humans is seen as a veneer over the unacceptable animal nature deep
within.
The public affection for the innate depravity idea is truly phenomenal.
During this period, I taught an adult education course under the naive
presumption that I would have to sell the significance of animal behavior to
the average man-in-the-street. On the contrary, my chief problem was to
convince the man-in-the-street not to take every monkey anecdote to heart
in the raising of his children. My students were New York citizens who rode
the subways and daily saw the spectacle of hundreds of people of every age,
sex, race, religion, mode of dress, smoking habit, eating habit, and degree of
wakefulness compressed peaceably into an intimacy rarely shared between
husband and wife in our society, much less by strangers. Yet despite this
spectacle, my students were eager to interpret the handful of aggressive in-
cidents which occur daily in the subways as evidence that man's simian na-
ture ill suits him to live in an urban environment. To an ethologist, what is
surprising about people in subways is not their hostility; on the contrary, it
is the degree of coordination and habituation which permits thousands of
people to move daily through an environment so physically hostile as to
stampede the herds of any sane animal. But these mature, intelligent New
Yorkers rejected their daily personal experiences of this miracle in favor of
the distant evidence of some knifing presented by a tabloid. The extent to
which members of the public reach out for this sort of lurid ~nformation
suggests that the innate depravity idea is more deeply rooted in their minds
than could possibly have been accomplished in ten years of ethological
popularizing. Indeed, it suggests that the ethological popularizers have been
reflecting popular beliefs, rather than the other way around.
We in the animal behavior business have felt very ambivalent about
these books. The books have resulted in a kind of attention from society in
general, and from government in particular, which was personally en-
couraging and financially gratifying. One year, the Smithsonian wined and
dined the ethological profession in a four day conference entitled "Man and
Beast." The conference started with cocktails in the National Zoo, com-
plete with a madrigal group in an empty giraffe cage and a string quartet in
an empty zebra cage. Cocktails and hors d'oeuvres were served on damask
before the hippo cage, while the hippo looked on. Gracious Washington
ladies with patronizing pictures of LBJ on their mantles gave dinner parties
in honor of ethologists; earnest roundtable conferences were held, and im-

passioned pleas for the preservation of the hunt were made by lobbyists
from the National Rifle Association. Hands were shaken by Chief Justice
Warren, and advice was sought on foreign policy by William Fulbright. A
gala held in the National Gallery was attended by international statesmen.
It was all very glamorous. I confess that when I saw my colleagues getting
all this attention, I wanted very badly to write an innate depravity book.
Why shouldn't I trace man's ills to his relationship to the crow? After all,
man is vertebrate, too! I rushed home to my typewriter. "The black nature
of man," I wrote, "harks back to the black plumage of his cousin, the
crow." My electric typewriter jammed in protest, but I wrote on. "Have
you ever wondered why social prejudice, man's greatest social evil, is called
Jim Crow? Have you ever wondered why, in the mythology of men, both the
crow and the devil speak with forked tongues? Have you ever wondered why
a man who humbles himself-returns to his origins, so to speak-'eats
crow?' Surely [and here my typewriter set about producing a profusion of
x's and had to be subdued with a sharp blow] ... Surely, these expressions
are not mere coincidences. Surely they embody the deep folk wisdom, the
certain race knowledge, that man is descended Jrom the crow!"
But my typewriter and my conscience prevailed in the end. The crow
was not to be defended against the monkey by the closeness argument. How
was it to be defended? How was I ever to get the monkey off my back?
At this time I attended a symposium on the "Social Significance of
Animal Si.udies." Any reader who has recently been a graduate student in
psychology will recognize the title as that of the classic D. O. Hebb and W.
R. Thompson entry in the Encyclopedia oj Social Psychology. The main
theme of this article is that the significance of animal studies is not in the
generalizations one can make about man, but in the power of such studies
to cause human investigators to see human behavior in a new light. A com-
parative investigation is like a good vacation; when it's all over you can
never again quite get your tires into the old ruts. The symposium was con-
vened in one of those plastique baroque ballrooms of a Philadelphia hotel.
A standing-room-only crowd of social and clinical psychologists was wait-
ing eagerly for the beast to be unlocked from the cabinet of man's soul. The
speakers were magnificent; they urged on the audience a new set of expecta-
tions about the relevance of animal studies. One by one they spoke suc-
cinctly about their animal researches and went on to offer their results as a
standpoint from which to view some aspects of human life. In each case, the
animal research caused the investigator to ask some probing questions
about how human investigators thought about human behavior. If the order
and timing of events are critical in establishing harmonious mother-infant
relationships in rats, what about order and timing as determinants of the
mother-infant relationship in man? If birds have some sort of genetic limit

on the songs they can learn, what might be the genetic limits upon human
language learning? If environmentally deprived monkeys rock rhythmically
in their cages, what does this suggest about the etiology of autism, a syn-
drome which sometimes includes similar rocking behavior?
What made these presentations so useful was not the superficial resem-
blance they pointed out between animal and human phenomena; what made
them of interest was that the animal phenomena they described indicated a
contradiction inherent in scientists' ideas about the world. Their presenta-
tions showed that the notion of experience deprivation is simply a
contradiction in terms: how is one to measure the passage of time inherent
in the concept of deprivation except in terms of some sort of experience? To
deprive of one experience is necessarily to provide with another experience,
and one is hard put to tell if the odd behavior produced is due to the
experience that was withheld or to the experience that was provided. The
"experience-deprived" animal has not learned "nothing"; on the contrary, it
has usually learned something that is incompatible with effective existence
in the normal, "experience-rich" world.
By this time, a less precarious defense of the crow against the monkey
was beginning to suggest itself: a defense on epistemological grounds. A
traditional view of the scientist holds that he is collecting information about
the universe and that the more he collects, the more accurate is his under-
standing of the universe. According to this sort of view, the pursuit of
science is like pasting tiny trading stamps of knowledge in giant books of
ignorance. The partially filled books are passed on from generation to
generation, and every once in a while somebody fills a book and cashes it in,
sometimes for a Nobel prize. (Darwin is a particular hero of this school of
thought, because he is one of the few scientists who has ever singlehandedly
collected enough stamps to fill a book.) The "trading stamp" view of
science fails to take into account the fact that most people who collect in-
formation have ideas about the meaning of that information, and that for
Charles Darwin, in particular, the idea of evolution was family history long
before the voyage of the Beagle. A more modern view of science holds that
we always have. ideas about the world no matter how ignorant we may be,
and that we seek facts when our ideas come into conflict. According to this
view, our ideas about nature become the central figures in science, not na-
ture itself.
The insight that science arises from conflict among concepts is a useful
one for explaining characteristic patterns of birth, growth, and decay in the
sciences. Initially, a phenomenon is brought sharply into focus by its rela-
tionship to a conceptual problem. A first generation of imaginative investi-
gators is attracted to the phenomenon in the hope of casting light on the
related conceptual issue. These investigators generate a lot of argument, a

little progress, and a lot of publicity. Then a second generation of scientists
is attracted, who are drawn to the problem more by the sound of battle than
by any genuine interest in the original issue. By then the conceptual issue
has been straightened out, the good people have left, and those who re-
main devote their time to swirling in even tighter eddies of technological
perfection.
If one takes this view of science, then one sees the whole "closeness"
argument as spurious. A study does not get to be significant because it is on
an animal phylogenetically close to man; it gets to be significant because it
relates to an important question. Even studies of man are insignificant if
they do not relate to an important question. The early field studies on pri-
mate social behavior were not significant because their subjects were close
to man; they were significant because the scientists were addressing
themselves to interesting ambiguities in their conceptions of man. Previous
generations of anthropologists had anchored their conceptions of man as a
unique species upon particular features of man's way of life, such as lan-
guage, tool using, fire building, meateating, and upon certain anatomical
specializations, such as a pelvis for erect walking and a large brain. The
data discovered by the early field workers indicated that nonhuman pri-
mates possessed many behaviors previously thought to be unique to hu-
mans, and that hominids possessed many characteristics previously thought
to be beneath them. These data raised questions concerning the borders of
our conceptions of man: hominid, anthropoid, and primate. Do monkeys
communicate through vocalizations? (Yes!) Do they normally use a lan-
guage? (No! On the contrary, most mature nonhuman primate communica-
tion systems are rather unelaborate.) Do primates eat meat? (Yes, some-
times! And under circumstances particularly conducive to theories about
the evolution of human meat eating and weapon using.) Do primates ever
use tools? (Yes, but in ways that are so uninteresting as to cast doubt upon
the importance of the original question.) To what extent do primate social
organizations resemble man's? (To no great extent; prairie dogs, wolves,
and beavers provide parallels that are equally interesting.)
As these questions have been answered, no new ones have been forth-
coming. Some may regard this absence of emotion-laden questions as a
healthy sign of a disciplined and selfless empiricism. To me, it is a sign that
primate social behavior has skipped from infancy to pedantry. A good con-
ceptual problem is a resource, like a virgin forest. Theoretical development,
like forestation, must be planned for. Each author who solves a theoretical
problem has a responsibility to seed the literature with several more. Pri-
mate social behavior is like a forest that has been sloppily and opportunis-
tically logged. Some important discoveries have been harvested, and their
theoretical stumps stand rotting. Strewn about is a tremendous quantity of

empirical litter and secondary growth. No new important growth is es-
tablishing itself.
Why has so vigorous a field suffered so rapid an atrophy? Possibly be-
cause of its uncritical acceptance by the public and by funding agencies.
Whereas ethologists of other species have had to justify their researches in
terms of important conceptual issues, primatologists have always been able
to fall back on the "closeness" argument. So, then, why should such a rene-
gade primatologist want to study crows? Because at this time crows present
a greater potential challenge to our ideas about man. What is a language?
What sorts of information communicated in what sorts of ways constitute a
language? What is a society? What sorts of components formed in what
sort of a structure make a society? How necessary is the Big Brain to the
possession of language and society? Is the Big Brain essential, or can a lan-
guage and a society be observed in an organism with a brain the size of half
of a small walnut meat?
Of course, the presence of such challenges to basic conceptions of man
and nature is not unique to the crow. Indeed, a vast array of lower species
may provide equally interesting challenges. A fall from the top of the scala
natura will always be a fall into grace, if the faller lands with his feet firmly
set on fertile theoretical ground. This ethologist's lament is thus a call for a
revolution by the phylogenetic proletariat. "Rise up my fellow ethologists of
the lower orders! No longer need ye doff your caps to the evil-tempered and
incontinent monkey. Go forth with a light heart and study the crow, the
frog, the paramecium, and the newt. The meek shall inherit the earth!"
Chapter 7
THE ORGANIZATION OF AGGRESSION AND

FEAR IN VERTEBRATES
John Archer 1
The University oj Sussex
Ethology Group, School oj Biological Sciences
Brighton, Sussex, U.K.
I. ABSTRACT
This chapter considers what factors influence the occurrence of attack and
fear behavior (e.g., immobility, flight, distress signals). It is argued that,
with the exception of conditioned attack and fear behavior, the same types
of basic situations are capable of evoking either type of response.
These situations are as follows: (l) Pain. (2) A moving object coming
too close to an animal (individual distance intrusion). (3) A novel object or
animal entering another's familiar area: this includes not only the typical
territorial situation, but also a wide range of other situations. (4) Entrance
into an unfamiliar area: most commonly, fear behavior occurs, but attack
can occur if there is a suitable target. (5) A familiar object in an unfamiliar
place. (6) Frustrative nonreward. (7) Thwarting (a physical object prevent-
ing completion of a learned response). (8) Low reinforcement schedules,
which involve delayed reward. (9) Morphine withdrawal, which is similar to
frustrative nonreward.
It is suggested that all these situations can summate with one another,
up to a certain limit, to increase the likelihood and intensity of attack, but
that above this limit fear behavior occurs.
It is then suggested that all the situations involve the common property
of a discrepancy between observed and expected stimulation. Different
types of "neuronal models" may represent the different types of expect a-
1 Present address: Division of Psychology, Preston Polytechnic, Preston, Lancashire, U.K.
231
232 John Archer
Hons the animal may have in relation to pain, novelty, and frustration. The
situations can also be viewed as an evolutionary series of increasing com-
plexity, involving increasing reliance on cognitive representation of expected
stimulation.
A number of factors are suggested as influencing whether attack or
fear behavior occurs in a given situation. The following are regarded as
internal variables influencing the general predisposition to attack or show
fear behavior: (1) The degree of discrepancy between observed and expected
stimulation; at low levels, attack is more likely, but this tends to be replaced
at higher levels by fear behavior. (2) The strength of ongoing motivation: if
this is high, any discrepancy is more likely to result in attack than in fear
behavior. (3) Hormonal factors: androgens tend to facilitate attack, and the
possible role of ACTH is also discussed. (4) Experiential factors: early
experience, social isolation, and previous reinforcement, all provide ways of
biasing the decision toward attack or fear behavior.
The following factors are also discussed as affecting the precise be-
havioral outcome: (1) Target properties: whether the discrepancy involves a
small localized stimulus, or whether one is nearby; movement and size (rela-
tive to the attacker's size) appear to be important target properties. (2) The
predominance of active or passive escape appears to be related to several
variables, including the intensity and localization of stimulus discrepancy.
(3) If escape is physically blocked, attack may occur instead.
The final discussion raises the following points: (1) the dynamic and
functional nature of attack and fear behavior; (2) a comparison of the
present classification of aggression with that of Moyer; (3) an outline of
predictions from the present model and a discussion of possible future
research.
II. INTRODUCTION
The purpose of this paper is to review a representative sample of the

literature on vertebrate aggression, with the following aims in view. First, to
classify the situations in which aggression occurs; second, to argue that
these situations are also the ones which evoke responses concerned with
avoidance, fear, and communication of distress (henceforth referred to as
"fear behavior"); third, to suggest that certain common features underlie all
these situations, and that this enables a general theory of the causation of
aggressive and fear behaviors to be formulated; fourth, to consider which
other variables (concerned with internal motivational and physiological fac-
Aggression and Fear in Vertebrates 233
tors, previous experience, and external variables) affect the probability of

occurrence of aggressive and fear behaviors.
The first problem is to define s()me of the terms to be used and to indi-
cate the types of studies to be considered in the review. "Aggression" is a
vague, imprecise, and inclusive term, which can refer either to an in-
terpretation of intent (e.g., Moyer, 1973), or to a state of mind, or to a
hypothetical motivational system; or it may merely be used in a descriptive
sense to indicate forcefulness. "Fighting" is a more operational term,
describing the expression of aggression involving two or more animals
(Kuo, 1960a). It does, however, only refer to some of the situations where
aggressive motivation may be inferred. For example, it would not include
"threat," which describes the symbolic expression of the intent to fight. A
further term, "dominance," refers to the modification of behavior which oc-
curs as a result of the outcome of aggressive encounters. Finally, "anger"
and "rage" have been used to refer to the emotional state underlying ag-
gressive behavior. Moyer (1973) provides a more detailed discussion of the
terminology used in studies of aggression.
The present paper is principally concerned with the causation of "at-
tack," a term referring to a rapid movement directed towards a localized
stimulus, usually culminating in physical damage to that stimulus (e.g., by
biting, kicking, pecking, wing flapping, or some other means, but excluding
such movements where they are concerned with the acquisition of food).
Thus we exclude "predation," which is generally regarded as motivationally
distinct from "aggression," although it does involve many of the same mo-
tor components (cf. Moyer, 1968, who considers predation to be a form of
aggression). We also exclude doubtful forms of "predatory aggression"
(Moyer, 1968), such as mouse-killing by rats (Karli, 1956). These exclusions
do not imply that the motivational systems for attack and predation may
not show partial overlap under some circumstances (e.g., in cannibalism by
the chimpanzee: Bygott, 1972; in mouse-killing by rats: Karli, 1956); it
does, however, maintain that the two classes of behavior have essentially
evolved independently, and may be found to coincide only in some special
circumstances, mainly owing to their shared anatomical and motor
components. There is in fact good evidence for their neural and motiva-
tional separation (Moyer, 1968).
Different factors may affect initial attack and other aspects of fighting
behavior. The details of the motor components of fighting, the final out-
come of the encounter, subsequent dominance relations, etc., may all be in-
fluenced by factors other than those which affect the probability of initial
attack. For example, in a study by Hutzell and Knutson (1972) it was shown
that social or isolation rearing influenced the total amount of shock-induced
234 John Archer
fighting by rats, but did not influence the amount of shock-induced biting.
The present paper discusses mainly what initiates attack, rather than how
its final form, outcome, or duration are determined.
A second type of behavior, which will be considered in relation to the
causation of attack, is termed "fear behavior": this refers to escape and
avoidance (e.g., rapid locomotion away from a source of stimulation), im-
mobility or protective responses, and distress signals (e.g., calls, odors,
facial expressions). The term is, therefore, used to describe a loose category
which includes some behaviors traditionally referred to as escape and
avoidance (in a descriptive sense), and others which are commonly regarded
as indicative of such central states as "emotionality" (e.g., Denenberg,
1969) or fear (e.g., Candland and Campbell, 1962). The term "fear" has
also been used either in the sense of an emotion (e.g., Hebb, 1946), or as an
intervening variable (e.g., Miller, 1948a), or as a tendency (e.g., Andrew,
1956). However, its present use as "fear behavior" is intended as a loose
description, and does not therefore imply a unitary motivational relation-
ship between the various measures (for a further treatment on this subject,
see Archer, 1973d).
Situations evoking attack and fleeing (a form of "fear behavior") have
been fairly extensively studied by ethologists, with particular reference to
the causation of threat displays. A major part of this work has been
directed to the question of whether these displays result from conflicting
tendencies to attack and escape aroused by the same stimulus (e.g., Tin-
bergen, 1952; Blurton-Jones, 1968; see also Hinde, 1970, Chapter 16), since
both types of behavior occur in similar situations. In another context, Hebb
(1946) commented on the interrelationship of "fear" and "rage," suggesting
that similar immediate external causal factors can evoke either emotion,
and that flight and attack both serve the similar function of restoring a dy-
namic equilibrium. Similarly, Scott and Fredericson (1951) remarked on
the functional similarity of these two types of behavior in rodents. More
recently, Galef (1970d) suggested that attack and avoidance behavior are
two varieties of a single type of response to novelty by rats: he showed that
the two responses were associated with similar environmental antecedents
and were mediated by common neural structures.
Theoretical and experimental investigations of "aggression" in man
and animals have focused on a number of different types of immediate
causal factors: e.g., pain (Ulrich, 1966), frustration (Dollard et al., 1939;
Berkowitz, 1962, 1969), specific social stimuli (e.g., Lack, 1939; Tinbergen,
1952; Mackintosh and Grant, 1966), and innate internal drives (e.g., Eibl-
Eibesfeldt, 1961; Lorenz, 1966; for criticisms of this approach, see Hinde,
1967; Johnson, 1972; Berkowitz; 1973). Moyer (1968) has proposed a tenta-
tive classification of aggression, but this is based on an inconsistent variety

of internal and external causal factors (see final section of this paper).
The authors referred to above have generally considered only the cau-
sation of aggression, rather than relating it to the causation of fear be-
havior, along the lines suggested by Hebb, Galef, and the other writers
mentioned above. In the present paper, evidence is considered concerning
the causation both of attack and of fear behavior, in a variety of vertebrate
species.
To summarize the argument briefly, it is suggested that, with the ex-
ception of conditioned attack and fear behavior resulting from previous
reinforcement (e.g., "instrumental aggression": Moyer, 1968), the same
situations can evoke either attack or fear behavior: these situations can be
classified into a series of increasing complexity, from the simplest form,
involving pain-induced attack or fear behavior, to the most complex, involv-
ing different forms of "frustration." It is suggested that they rely increas-
ingly on some form of central representation of the external environment,
and it is further argued that they all involve a discrepancy of some mag-
nitude between an observed event and what was predicted or expected on
the basis of this central representation, although this common type of
mechanism is not meant to imply a single neurophysiological system.
It is then argued that a number of variables influence the probability of
either attack or fear behavior occurring in a given situation. It is suggested
that thes" variables act in two stages, first, to determine the general predis-
position of the animal toward either attack or fear behavior, and second, to
determine the more precise behavioral outcome.
It is suggested that the outcome of the first stage of this hypothetical
dual mechanism (called "decision processes") is determined by several
internal variables, one of which is the degree of discrepancy between ob-
served and expected events; it is postulated that the tendency for both at-
tack and fear behavior to occur increases with increased discrepancy, but
that the tendency for fear behavior to occur increases at a faster rate.
The second internal variable is the strength of motivation associated
with the expected outcome: if a highly motivated, goal-directed activity is
prevented from occurring by a discrepancy in stimulation, attack behavior
is more likely to occur than when a less-motivated type of behavior is
involved.
A third class of variables refers to internal chemical factors, such as
the level of testosterone, and a fourth to the animal's past history, including
positive or negative reinforcement of attack or fear behavior; the latter
exerts a particularly strong influence on the subsequent probability of one
of the two types of behavior occurring.
John Archer
It is then suggested that the second stage of the hypothetical

mechanism ("decision process 2") is influenced by external variables, for
example, whether the discrepancy in stimulation is localized as a stimulus
whose general appearance (e.g., shape, size, movement) makes attack an
appropriate response for removing the discrepancy; and also whether the
most readily emitted behavior can occur unimpeded by physical obstacles
(e.g., if escape behavior is blocked, attack becomes more likely).
The remainder of the paper discusses in more detail the evidence for
this theoretical scheme.
III. CONDITIONED ArrACK AND FEAR BEHAVIOR
Thompson (1969) distinguished between three types of "aggression" on

the basis of the degree of conditioning they involved: he described uncon-
ditioned (e.g., pain-induced), classically conditioned, and operantly condi-
tioned aggression. Unconditioned attack is considered in the following
sections. Classically-conditioned attack occurs when a stimulus eliciting un-
conditioned attack (UCS), e.g., an electric shock, is paired with a neutral
stimulus which subsequently comes to elicit attack in the absence of the
UCS. This process has been demonstrated in rats by Vernon and Ulrich
(1966); who paired shock with a tone; by Creer et al. (1966), who paired
shock with a loud buzzer; and in Siamese fighting fish by Thompson (1969),
who paired a mirror image of a fish with a light. Lyon and Ozolins (1970)
trained male rats on a discrimination procedure, pairing shock with one
tone (S+) but also presenting another tone in the absence of shock (S-).-
Discrimination and also reversal learning were demonstrated. Berkowitz
(1970) has argued that the classical conditioning paradigm provides an ap-
propriate model for much impulsive human aggression.
Operantly-conditioned attack involves either the reinforcement of at-
tack by conventional reinforcers, which themselves are not related to attack
(e.g., food, water, etc.), or it involves reinforcement through an outcome of
attack itself. In the first case, it was shown by Reynolds et al. (1963) that
when food reinforcement was made contingent upon attack of a model, pi-
geons formed a stable pattern of attack which could be brought under the
control of exteroceptive stimuli, such as illumination. Similarly, using a
fixed interval schedule, Azrin and Hutchinson (1967) also conditioned and
extinguished attack for a food reward. Ulrich et al. (1963) successfully
made water reinforcement contingent upon attack of another rat. Again, it
is considered that much of human aggression can usefully be considered in
terms of the instrumental conditioning paradigm (Johnson, 1972, p. 141;

Buss, 1971).
The consequence of attack itself can also act as a reinforcer: thus an
animal can learn to perform a response in order to obtain an object to at-
tack. However, in such a situation, it is difficult to demonstrate that attack,
rather than the presence of the attacked stimulus, provides the reinforce-
ment. For example, in one study, Azrin et at. (1965b) showed that, when
shocked, Squirrel monkeys would learn to pull a chain to produce an inani-
mate object which was then attacked. In another, Hogan (1967) showed
that, in Siamese fighting fish, the opportunity to perform aggressive display
and fighting movements to their mirror image provided reinforcement for
swimming through a ring and down a runway. Nevertheless, as remarked
before, it is uncertain whether studies of this type involve only aggressive
motivation, particularly since Johnson and Johnson (1973) have shown that
Siamese fighting fish will perform an operant response in order to view a
variety of objects, including a conspecific male, a mirror, other species of
fish, a marble, and even an empty chamber.
Several other studies of the reinforcing properties of aggression have
used mice. For example, Lagerspetz (1964) found that these animals would
cross an electric grid to reach a goal box where they had the opportunity to
attack and beat another mouse, although the incentive value was slight.
Generally, mice from a more aggressive strain crossed the grid more readily
than did those from a less aggressive strain, and a recent fighting experience
further increased their readiness to cross the grid. These findings suggested
that in this case attack itself, rather than the mere presence of the attacked
stimulus, was providing the incentive for grid crossing.
Legrand (1970) showed that mice would learn to run a runway for the
opportunity to defeat a submissive mouse, and that those rated as "highly
aggressive," or those given a pretest fight, showed faster running times than
did less aggressive mice or than did those not given pretest fights. Tellegen
et at. (1969) also found that male mice acquired, extinguished, and reversed
a position preference in aT-maze when provided with the opportunity to at-
tack a "victim" mouse as a reinforcer. Again, mice given a pretest fight
proved superior in choice performance during acquisition and reversal
learning, and showed slower extinction than mice which had not had pretest
fights, suggesting that attack itself rather than the presence of the attacked
mouse was again providing the incentive. These findings were extended by
Tellegen and Horn (1972) to three strains of mice, and compared with run-
ning for a food reward. In contrast to the findings of Lagerspetz (1964),
mentioned above, Tellegen and Horn found that learning in the two better-
performing strains approached that involving a food reward, indicating that
238 John Archer
the opportunity for attack provided a relatively strong incentive in these

cases. Perhaps the different results are partly attributable to the different
methods used, since it is likely that the shocks administered by Lagerspetz
would have aroused tendencies to escape or "freeze," which are incompati-
ble with attack.
Attack can also be reinforced negatively, by means similar to those
described above, but using conventional punishment procedures (Ulrich et
01., 1969; Baenninger and Grossman, 1969). Both Azrin (1970) and Roberts
and Blase (1971) also used shock to punish shock-induced attack, in
monkeys and rats, respectively, and found that attacks decreased as a func-
tion of the intensity of contingent shock. Other studies have involved arti-
ficially induced defeat following aggressive encounters (e.g., in mice: Hud-
gens and MacNeil, 1970; Kahn, 1951; Lagerspetz, 1964), with the result
that attack became markedly reduced or absent on subsequent occasions.
On the other hand, fear behavior (distress vocalizations and escape be-
havior) was facilitated by such treatments. Fear behavior can, of course, be
classically or instrumentally conditioned (e.g., during active or passive
avoidance), so that stimuli other than primary or unconditioned ones
(considered in the following sections) come to evoke these responses.
IV. SITUATIONS EVOKING UNCONDITIONED ATTACK OR

FEAR BEHAVIOR
Having briefly described the various forms of conditioned attack and fear
behavior, we now consider the situations which evoke these types of be-
havior in the absence of conditioning. The present classification differs from
that of Moyer (1968) in that it does not refer to predation (see Introduc-
tion), and in that it is based solely on the different situations in which ag-
gression and fear behavior occur, without reference to inferred internal
states (e.g., Moyer's "irritable aggression"). Differences between the two
schemes are discussed further in the final section of this chapter.
A. Pain
Electric shock or sudden contact with a heated floor evokes attack in

pairs of rats (Ulrich, 1966; Ulrich et 01., 1965; Ulrich and Azrin, 1962) and
the aggressive postures involved are those shown in socially induced fighting
(Reynierse, 1971). Lloyd and Christian (1967) observed that a bite from a
conspecific often preceded attack in mice housed in dense colonies. A
physical blow has been shown to evoke attack in the squirrel monkey
(Azrin et al., 1965c).
Pain-elicited fighting has been reported in a number of species, such as
ground squirrels (Turner et al., 1973), gerbils (Boice and Pickering, 1973),
hamsters, cats, monkeys, snakes, turtles, and chickens (reviewed by Ulrich,
1966; Ulrich et al., 1965).
In all of these studies, attack is directed toward a member of the same
species or toward an inanimate object (Ulrich, 1966; Ulrich et al., 1965;
Azrin et al., 1964a; Azrin et al., 1965c; Galef, 1970a,b).
Electric shock can, of course, also evoke fear behavior. In rats, both in
the presence of a conspecific (e.g., Ulrich, 1966; Reynierse, 1971; Conner et
al., 1971) and in its absence (e.g., Friedman et al., 1967; Savage, 1960;
Galef, 1970a), jumping, continuous vocalization, and defecation have been
described in response to shock. Grossen and Kelley (1972) found that
"freezing" was the predominant type of behavior shown by rats shocked in
a novel environment, whereas this response was much less frequent in the
absence of shock.
Recent work has shown that different physiological changes occur after
shock, depending on whether fighting or escape behavior has occurred (e.g.,
Williams and Eichelman, 1971; Conner et al., 1971). Conner et al. suggest
that these different physiological changes (e.g., decreased blood pressure
and ACTH secretion) represent an adaptive consequence of fighting, in that
they lessen the internal response to the stressor.
B. Individual Distance Intrusion
A wide variety of vertebrate species typically attack when an "in-

truder" (not necessarily a conspecific or even another animal) comes into
close proximity. Most of the studies of this type of aggression have been
carried out on birds, but it is clear from Hediger (1950) that its occurrence
is also widespread in mammals.
In bunting flocks, Andrew (1957) reported that aggressive responses
were usually aroused by another bird approaching too close, although other
small approaching objects would also arouse attack. Similarly, threat pos-
tures or attack commonly occur when two -birds are perched together or
feeding near each other on the ground (in black-headed gulls, tufted ducks,
wigeons, starlings, meadow pipits, curlews, lapwings, and oystercatchers:
Conder, 1949; in blue tits and great tits: Hinde, 1952; in chaffinches: Hinde,
1953, Marler, 1956b; in Quelea: Crook and Butterfield, 1970). Such an area
of close proximity is often referred to as an "individual distance" (Conder,
1949), and in the chaffinch, Marler (1956a) found that its size varied with
240 John Archer
a number of the intruder's individual characteristics, such as its sex and

behavior.
Marler (1956a) suggested that individual distance may be the simplest
form of defended area, and that perhaps it is the evolutionary precursor of
other types which are centered on an object or collection of objects essential
for reproduction or feeding (Le., a territory). Crook (1968) made a similar
point when he referred to territory as "a special case of spatial defence not
easily separable from the maintenance of personal space."
An alternative response to violation of individual distance is escape or
avoidance; e.g., this has been described for a variety of species (see above)
by Conder (1949), for Quelea by Crook and Butterfield (1970), and for
chaffinches by Hinde (1953). One factor that encourages attack rather than
avoidance appears to be familiarity with the surroundings (Marler, 1956a,
for chaffinches; Stokes, 1962, for blue tits at a winter feeding station).
It is concluded, therefore, that when a small unfamiliar object, espe-
cially another animal, approaches too close to an animal, attack or fear be-
havior is likely to occur.
C. "Territory": The Presence of a Novel Object in a Familiar Situation
Attack occurs commonly in a wide variety of vertebrate species when a

strange conspecific enters an established colony, or the territorial area of an
established male, or (more generally) another animal's home or familiar
area. There are two important environmental variables to consider in con-
nection with this type of attack. First, the familiarity of the home area to
the attacker, and second, the stimulus properties of the intruder which
evoke attack.
Several studies have demonstrated the importance of familiar environ-
mental cues for the attacker. Kuo (1960d) showed this rather dramatically
in quail which had been trained to be highly aggressive, using a method used
by quail gamblers (involving prolonged "frustration" of feeding): these ani-
mals appeared to lose all tendency to attack once their familiar ground was
replaced by unfamiliar cues. A similar result was obtained for imprinted
ducklings, which showed attack in the presence of the familiar imprinted
stimulus, but not in its absence (Hoffman and Boskoff, 1972).
Jones and Nowell (1973) compared the latencies to attack and total at-
tacking times towards standard castrated opponents, for male mice in envi-
ronments differing from their home cage. The mice were found to attack less
according to the degree of dissimilarity between the new environment and the
home cage, and familiar olfactory cues were found to be more important
than familiar visual ones.
Thus, the first important characteristic of this type of attack is the at-
tacker's familiarity with the surrounding area. This is, of course, one of the
important characteristics of a territory (Moyer, 1972, provides further dis-
cussion of this point). The second important environmental variable is the
class of stimuli which evokes attack. In some cases, specific characteristics
are important: e.g., color in the side-blotched lizard (Ferguson, 1966), the
red breast of robins (Lack, 1939), a.c. electric fields in Gymnotus carapo
(Black-Cleworth, 1970), and the red abdomen of three-spined sticklebacks
(according to Tinbergen, 1952; but see Peeke et aI., 1969). More generally,
however, the evidence suggests that unfamiliarity with the intruding con-
specific plays an important role in evoking attack. For example, fighting is
relatively uncommon in established colonies of animals, but occurs when a
stranger is placed into such a colony, especially if the stranger is an adult
male (in mice: Andrzejewski et al., 1963; in wild rats: Barnett, 1955, 1958a;
Alberts and Galef, 1973; in domestic fowl chicks: Horridge, 1970; in rhesus
monkeys: Southwick, 1967a). Females and young may also be attacked, in
some instances, when they are placed in an established colony (in wild rats:
Galef, 1970d; Alberts and Galef, 1973; in rhesus monkeys: Southwick,
1967a; in polecats: Poole, 1973).
Studies of territorial rodents have demonstrated that they direct fewer
attacks toward their immediate neighbors (i.e., familiar animals) than
toward complete strangers (Healey, 1966; Wolfe and Summerlin, 1968);
similarly, ovenbirds and white-throated sparrows show a lesser response to
song from their immediate neighbors (Falls and Brooks, 1965; Falls, 1969).
Waning of aggressive responses toward a conspecific presented
repeatedly over several days has been demonstrated in convict cichlid fish
(Peeke et al, 1971), in three-spined sticklebacks (Peeke, 1969), and in
Siamese fighting fish (Peeke and Peeke, 1970). Peeke et al. (1971) suggested
that this process-presumably habituation-enabled neighboring animals to
maintain a low level of aggression toward one another. In a more recent
study, Peeke and Veno (1973) illustrated further the importance of novelty
in territorial attack by three-spined sticklebacks. Territorial males showed a
waning of their attack toward a novel conspecific after thirty minutes, but
would again attack at their initial intensity if a different conspecific was
then presented. Thus, habituation was demonstrated, and the fish were
shown to respond to one another on an individual basis.
Specific stimuli and novelty of the intruding animal may interact as
factors evoking attack. For example, male mice secrete an odor or odors
which increase the probability of their being attacked by another male
(Mugford and Nowell, 1970, 1971 a), and female mice produce an odor
which decreases the probability of their being attacked (Mugford and
Nowell, 1971 b). However, males which are familiar with one another show
242 John Archer
little or no fighting, despite the presence of the male odor, but fighting does
occur when a strange male odor is present (Mackintosh and Grant, 1966;
Archer, 1968; Mugford, 1973). Thus, in mice, there seems to be some
degree of specificity, and yet unfamiliarity is also important.
A number of studies demonstrate that a wide range of strange objects
placed in an animal's familiar area may evoke attack, especially if these ob-
jects resemble another conspecific in properties such as size and movement:
other species may be attacked (e.g., in Peromyscus: Baenninger, 1973; in
the robin: Lack, 1939; in the cod: Brawn, 1960; in the chimpanzee: van
Lawick-Goodall, 1968, p. 175), and so might a lame, deformed, or
paralyzed conspecific (in Uganda waterbucks: Spinage, 1969, p. 353; in
chimpanzees: van Lawick-Goodall, 1968, p. 279; in wild rats: Alberts and
Galef, 1973). Several species (great tits, rats, mice, domestic fowl chicks)
may direct attack toward an unfamiliar human hand or finger placed in
their cage (e.g., Blurton-lones, 1968; Galef, 1970d; Horridge, 1970; Mur-
phey, personal communication). In contrast, no attack is shown toward a
familiar human hand (Galef, 1970d; Horridge, 1970). Horridge also
reported that in cases where a chick had habituated to the human hand, at-
tack could be elicited by using a colored glove. In other studies, inanimate
objects placed in an animal's familiar area have been attacked: e.g., a mov-
ing cardboard box (domestic fowl chicks: Bateson, 1964), moving black
cardboard rectangles (chicks of several species: Schaller and Emlen, 1962),
a bottle brush (mice: Lagerspetz and Mettala, 1965), a pencil or a paper roll
(great tits: Blurton-lones, 1968), a green wooden triangle (domestic fowl
chicks: Evans, 1967), and a red wooden model (sticklebacks: Peeke et al.,
1969). Brawn (1960) reported that even a sudden noise outside the tank
could elicit threat in the cod.
Two of these studies provided more precise data showing that attack
waned (presumably habituated: see above) with successive presentations of
the stimuli. Evans (1967) reported that threat and aggressive pecks waned
over 44 consecutive trials (repeated at two minute intervals), and Peeke et
al. (1969) reported that the biting response of the territorial male three-
spined stickleback declined markedly over 12 two minute daily exposures to
a model.
Two conclusions can be derived from the studies reviewed above; first,
that novelty appears to be an important stimulus property in evoking terri-
torial attack of one conspecific by another, and second, that a variety of
novel stimuli that are different in form from another conspecific may also
evoke attack.
It is therefore suggested that, whereas individual stimulus properties
are undoubtedly important in inducing attack, and may be crucial for some
species, one common factor in "territorial" and similar forms of fighting is
the presence of an unfamiliar object in an area with which the animal is fa-
miliar. It is also likely that the probability of attack is, up to a point,
increased with increasing novelty of the unfamiliar object, and decreased
with decreasing familiarity of the familiar area.
Considering territorial attack in this way, the importance of specific
properties of the novel object (e.g., red breast, male odor, etc.) are still
recognized, but they are regarded as special forms of novel stimuli with a
high probability of being attacked in the particular species concerned. More
generally, it is suggested that a wider range of objects than is often supposed
may provide a substitute for such stimuli.
Thus it is argued that the presence of a novel object in a familiar envi-
ronment forms the basis of the causation of attack in territorial situations.
These situations can also evoke fear behavior, which may sometimes occur
in conjunction with attack; e.g., Blurton-lones (1968) reported that escape
behavior occurred in response to stimuli which were also evoking attack by
great tits; Schaller and Emlen (1962) found that a moving black cardboard
rectangle placed in the home cage of chicks of several different species
could evoke both avoidance (fleeing, crouching) and vigorous pecking. In
an imprinting study, Bateson (1964) also reported that some chicks would
violently peck the moving cardboard model after the cessation of avoidance
behavior, and before following had begun. Similarly, Evans (1967) found
that withdrawal responses could be elicited in the chick by the same
stimulus which evoked attack (a green wooden triangle), and that both
withdrawal and attack responses showed a decline with successive presenta-
tions.
More commonly, however, fear behavior provides an alternative
response to attack. When a strange conspecific enters an animal's familiar
area, the resident may in some cases flee from or avoid this conspecific. For
example, Kahn (1951) found that when a strange mouse was dangled in the
home cage of a previously defeated mouse, the latter typically responded with
squealing, escape, and defense behavior. Horridge (1970) reported that
strange male chicks with noticeable comb growth were generally avoided
when placed in an established group.
The human hand placed in the home cage of a rat or chick can also
evoke fear behavior (e.g., Ader, 1968; Ader et al., 1967; personal observa-
tion). Strange inanimate objects may also do the same; e.g., illumination of
a flashlight in a chick's home cage produces peeping and immobility
(Broom, 1969). Blurton-lones (1969) found that fleeing was readily evoked
in the great tit by a small lamp, after brief training during which the lamp
was flashed on whenever the bird was near to it. A strange feeding container
placed in the cage of a wild rat also produced avoidance and inhibition of
feeding (Barnett, 1958b).
244 John Archer
In many of the examples referred to above, fear behavior waned with

repeated presentations or prolonged exposure (e.g., Evans, 1967; Bateson,
1964; Schaller and Emlen, 1962). Thus, novelty is again likely to be the
main stimulus property responsible for producing fear behavior in these cir-
cumstances. Nevertheless, specific factors are also undoubtedly important
determinants for many species; e.g., movement in the rat (Blanchard et at.,
1975), alarm calls in the domestic fowl chick (von Uexkull, 1934), alarm
substances in zebra fish (Gandolfi et at., 1968), and the spatial configuration
of predators in various birds (Tinbergen, 1952). Even so, it is again
envisaged that the basic situation evoking fear behavior is the occurrence of
a novel event in a familiar environment, and that species-typical responses
to particular stimuli have evolved from this more general situation.
In conclusion, therefore, it is suggested that the presence of a novel
stimulus in an animal's familiar area is one of the basic situations evoking
attack or fear behavior.
D. Presence of an Unfamiliar Environment
It is further suggested that attack or fear behavior can be evoked by an

animal entering or being placed in an unfamiliar situation. Attack must, of
course, be directed at some form of target; several studies have investigated
the fighting which occurs when strange animals are placed in a novel envi-
ronment (e.g., Levine et at., 1965; Brain and Nowell, 1969a), but these
instances involve both a novel environment and a novel target source. There
is, however, some evidence that animals which had hitherto been familiar
with one another and did not fight show increased likelihood of attack when
placed in a novel environment: Willis (1966) placed pigeons which had pre-
viously been living in a common loft into different-sized strange cages, and
found that a higher frequency of fighting occurred there than in the loft, al-
though this may have been caused by the decrease in cage size (Archer,
1970). Archer (1968) found increased fighting between pairs of mice that
had been taken from their home cage and placed in a clean cage, although
fighting was much less frequent in this case than when they were placed in a
cage also containing a strange male odor. It was also observed that single
male mice placed into a cage containing the odor of a group of males be-
came very "aggressive" after about three hours (Archer, unpublished
observations; procedure described in Archer, 1969); in the absence of a
suitable stimulus for attack, these mice would rattle their tails (a
component of fighting behavior: Mackintosh and Grant, 1966) and they
would attack a hand or a metal forceps placed in their cage. Similarly,
Poole (1973) found that male polecats which were familiar with one another
and which lived together amicably would fight when introduced into a large
arena, although biting was more inhibitied lhan that occurring between un-
familiar polecats.
When a single animal is placed in a strange area by an experimenter,
some type of fear behavior may occur. Thus, in rats and mice placed in a
strange area, an initial period of fast movement has been observed (Livesey
and Egger, 1970; Blizard, 1968), and rats will cross an electric grid in order
to escape to their home cage (Blanchard et al., 1974). Jumping at the sides
has also been reported in such novel arenas, for mice (e.g., Lieberman,
1963; van Oortmerssen, 1971), rats (Brannigan, 1972), pigs (Fraser, 1975)
and chicks (Archer, 1973a,c). Freezing and crouching, alternative forms of
fear behavior, have also been reported when rats are first placed in a novel
environment (e.g., Woods, 1962; Bindra and Spinner, 1958; Hutchings and
Gibbon, 1970), and similarly for chicks (Candland et al., 1963; Sluckin et
al., 1970; Archer, 1973b), which also emit distress calls in such circum-
stances (e.g., Candland et al., 1963; Sluckin et al., 1970; Archer, 1973a,c).
Animals often show high rates of defecation in novel environments (studies
on rats and mice reviewed by Archer, 1973d) and this has typically been
regarded as an indication of "emotionality" or sympathetic nervous activity
(e.g., Hall, 1934, 1938; Broadhurst, 1969; Gray, 1971), although the sup-
posed relationship of this to other behavioral and physiological responses is
not as simple or consistent as is often supposed (Archer, 1971, 1973d;
Russell, 1973).
Thus, it is argued that entering into or being placed in an unfamiliar
area is also a situation which can evoke attack or fear behavior, although
the evidence for attack is less clear-cut than it is for the other situations dis-
cussed so far.
E. Familiar Object in an Unfamiliar Place
Peeke and Veno (1973) found that after a territory-holding three-spined

stickleback had habituated to the presence of a conspecific, it would show
an increase in attack if the same conspecific were presented in a different
place. A change of location also increased attack toward a strange con-
specific. Similarly, Galef (1970b) found that wild rats showed more pain-in-
duced attack toward a familiar inanimate object when the object was placed
in a new location than when it was in a familiar place. In this case, the
change in location alone was insufficient to produce a direct facilitation of
attack, but when combined with pain, an additive effect occurred.
As far as the present writer is aware, there have been no reports of how
such changes might affe~t fear behavior.
246 John Archer
F. Frustrative Nonreward
Omission or reduction of a customary reward is a well-known and well-

studied form of frustrating situation, likely to evoke aggression in both
animal and human subjects (e.g., Berkowitz, 1962).
The technique most commonly used in animal studies has been to cease
rewarding an animal that had previously been responding on a continuous
reinforcement (CRF) schedule. Attack is then directed toward a ~ariety of
possible sources, according to the stimuli made available (e.g., a rubber
hose, a model of a conspecific, a mirror, a conspecific, a food dish). Male pi-
geons (Azrin et al.• 1965a, 1966; Knutson, 1970), rats (Thompson and Bloom,
1966), chicks (Archer, 1974b), and squirrel monkeys (Hutchinson et al.•
1968) have been used in these studies.
Gallup (1965) used another form of frustrative nonreward, first training
pairs of female rats on continuous reinforcement, and then transferring
them to partial reinforcement. He found that attack of a second, also
partially reinforced, conspecific occurred on the nonrewarded, but not on
the rewarded trials of partial reinforcement.
Other studies show that fear behavior can also occur during frustrative
nonreward: e.g., escape behavior in rats (Adelman and Maatsch, 1956;
Daly, 1969), defecation, jumping, and distress calling in chicks (Archer,
1974b), and "freezing" in mice (Slater, 1972).
There is, therefore, evidence that frustrative nonreward can evoke both
attack and fear behavior. This conclusion is similar to the hypothesis of
Gray (1972) that the state of "frustration" is physiologically and func-
tionally similar to that of mild "fear," both terms being used in this case as
intervening variables (see below for further discussion of Gray's theory).
G. Thwarting
"Thwarting" refers to the situation in which an animal is prevented by

a physical barrier from completing a previously reinforced response. It has
often been described as a form of "frustration" (e.g., Berkowitz, 1962;
Lagerspetz and Nurmi, 1964; Duncan and Wood-Gush, 1971), and has been
shown to produce attack of a nearby conspecific in both sexes of domestic
fowl (Duncan and Wood-Gush, 1971), and in male mice (Lagerspetz and
Nurmi, 1964).
In a similar situation, but using single hens, Guiton and Wood-Gush
(1967) reported the occurrence of escape behavior, and Duncan and Wood-
Gush (19.72) reported attempted escape behavior and alarm calling.
Under field conditions, in many species, fighting may occur in com-
petition for a localized food source (e.g., in blue tits: Stokes, 1962; in chaf-
finches: Hinde, 1953; in chimpanzees: van Lawick-Goodall, 1968) although
there is evidence from other species that food deprivation per se does not
generally increase readiness to attack (e.g., Duncan and Wood-Gush, 1971,
for the domestic fowl; Mykytowycz, 1961, for the rabbit; Strecker and
Emlen, 1953, for mice; Andrew, 1957, for buntings; Blurton-Jones, 1968,
for the great tit; Marler, 1956a, for the chaffinch; but see also Fredericson,
1950, for a different finding in mice).
Food competition may be loosely categorized as "thwarting," in that
one animal physically prevents access to a food source or to the most-
preferred food. Tests based on such a situation have from time to time been
devised to measure "aggressiveness" in mice (e.g., Fredericson, 1952), and
one variation of these is the so-called "dominance test," where two animals
are trained to run for a food reward, but only one is allowed access to that
reward (e.g., Kanak and Davenport, 1967).
Other types of rewarded responses that are physically prevented from
completion may also evoke attack in social situations: e.g., the possession of
nesting material (in Quelea: Crook and Butterfield, 1968), or attempts to
mount unreceptive females (in the Norwegian lemming: Arvola et a/., 1962).
H. Low Reinforcement Schedules
Delay between initiation and completion of an instigated response se-

quence is a further form of "frustration" likely to evoke "aggression" (for
human evidence: Berkowitz, 1962, chapter 2).
Several studies have reported attack in male pigeons of a target animal
or model during a high-ratio fixed-ratio (FR) schedule (e.g., 50-120), but
not during continuous reinforcement or a low-ratio RF schedule (e.g., 2, 6,
15, or 25) (Gentry, 1968; Flory, 1969a; Cole and Litchfield, 1969; Knutson,
1970; Cole and Parker, 1971). Similar results were found by Hutchinson et
at. (1968) for male squirrel monkeys, and Cohen and Looney (1973) have
reported FR-induced attack of a mirror-image in pigeons. Most of the at-
tacks found in these studies occurred during the postreinforcement pauses.
Gentry and Schaeffer (1969) report rather different findings for female rats
trained on water reinforcement. Attack increased from FRI to FR20, but
at FR40 and 60 declined to baseline levels; in addition, attacks were found
to occur not only in the postreinforcement pauses. Also using female rats,
Hayes et at. (1969) found no attack of a target animal, but they used an FR
6 schedule. Variables associated with species, sex, and the nature of the
reinforcement may have accounted for these different findings.
Three other types of low reinforcement schedules have been associated
248 John Archer
with attack of a conspecific model. Knutson and Kleinknecht (1970) found

that male pigeons attacked a restrained target during DRL 20 but not dur-
ing CRF, and that attack was not restricted to the postreinforcement
pauses. Flory (1969b) trained male pigeons on a fixed-time (FT) schedule,
whereby food was delivered at fixed intervals after previous reinforcement
irrespective of the animal's response: he found that as the interfood interval
was increased from 15 to 960 sec, attack increased to a maximum at FT
60-120, and then declined. Cherek and Heistad (1971), Richards and Rill-
ing (1972), and Flory and Ellis (1973) found that pigeons attacked target
birds or a stuffed model during FI (fixed interval) food reinforcement, and
that the attacks occurred during the postreinforcement pauses.
There is no direct information on whether such schedules also evoke
fear behavior, but both Azrin (1961) and Thompson (1964), using pigeons
and rats respectively, found that their subjects showed an increased
tendency to press a "time-out" button (and hence escape the contingencies
of the schedule) as the ratio of the FR schedule became larger. There was
also a substantial increase in time-out frequency accompanying a shift from
CRF to FR responding. The time-out periods occurred during the postrein-
forcement pauses (Azrin, 1961; Thompson, 1964), the time at which attack
was found to occur in other studies (see above). Azrin explained his results
in terms of the aversive nature of high ratio schedules.
It seems likely, therefore, that these high ratio schedules can evoke
both attack and fear behavior, although the relevant direct observations
have not been made for the latter.
I. Other Situations
Boska et al. (1966) reported increased fighting behavior during mor-

phine withdrawal in rats. Hinde (1970) remarked that this may be a
phenomenon similar to extinction-induced fighting, but since the animals
were also paired for the first time when observed by Boska et aI., the fight-
ing could have been caused by this unfamiliarity rather than by morphine
withdrawal.
More recently, other studies of morphine withdrawal and fighting be-
havior (e.g., Davis and Khalsa, 1971), using more adequate controls for
other possible causes of fighting, have demonstrated that male but not fe-
male rats show an increased number of aggressive encounters during mor-
phine withdrawal. Several other variables, such as age, strain, cage space,
and season of the year also affect the occurrence or extent of this
phenomenon.
Morphine withdrawal has also been reported to induce jumping in

single mice (Marshall and Weistock, 1971), and again this suggests that
similar situations produce both attack and fear behavior.
J. Other Types of Responses Evoked by These Situations
It could be argued that an increase in a more general drive state is in-

volved in all of the situations described above, and that to describe such
situations as specifically causing attack and fear behavior is misleading. In
particular, sexual behavior, irrelevant activities (e.g., displacement preen-
ing, drinking, eating), and exploratory behavior commonly occur in many
of the situations mentioned. However, if one examines the situations
associated with each of these alternative types of behavior, their range is
found to be more restricted than that of situations which evoke attack and
fear behavior.
Sexual behavior can occur in response to an unfamiliar stimulus placed
into a familiar area: e.g., in response to a female entering the territory of a
territory-holding male, or to a human hand placed in the cage of a young
male chick (e.g., Andrew, 1966). Sexual behavior can also be given in
response to males by other males (e.g., in laboratory rats: Grant, 1963).
Electric shock sufficiently intense to cause rearing and squealing can
facilitate copulation by male rats when a female is present (Barfield and
Sachs, 1970; Caggiula, 1972); it is important, however, to note that this fa-
cilitation occurs after the shock has ceased, and therefore it should be
regarded as an "aftereffect" rather than as a direct result of such stimula-
tion (see Fentress, 1973, for a general discussion of this point). Copulation
does not seem to be facilitated by frustrating situations such as those
described under headings G., H., and I. above (Falk, 1971). In fact, Duncan
and Wood-Gush (1971) reported that when thwarted in a feeding situation,
male fowl attacked nearby females, even though these were animals with
which they normally copulated.
There is a partial overlap of the conditions which can produce fear be-
havior and attack with those which can produce displacement and irrelevant
activities. The latter occur particularly readily in "frustrating" situations
(see above), and usually take the form of behavior which is common in the
animal's repertoire anyway (McFarland, 1966), e.g., drinking (Panksepp et
al., 1972) and grooming or preening (e.g., Duncan and Wood-Gush, 1972).
McFarland (1966) proposed an attention-switching hypothesis to account
for the phenomenon, suggesting that displacement activities arise as the by-
product of a mechanism for switching attention from a frustrating situation.
250 John Archer
Displacement activities are not, however, common features of the other

situations evoking attack and fear behavior, i.e., those involving either pain
or novelty. There are reports of increased grooming after presentation of a
novel stimulus (e.g., in Microtus: Fentress, 1968), but the more general find-
ing is a decrease compared with the home cage or undisturbed condition
(e.g., in Clethrionomys: Fentress, 1968; in the rat: Bolles, 1960). It should
also be noted that the increased grooming found in Microtus occurred after
the stimulus presentation (cf. pain-induced copulation, referred to above;
and see Fentress, 1973).
Novel stimuli and novel situations also evoke approach and explora-
tory responses, such as active investigation, manipulation, approach, and
following. Typically, exploration occurs after fear behavior has waned. At-
tack, which appears to occur when the degree of novelty is less pronounced
(see below), may overlap with exploration of a novel stimulus; e.g., an in-
truder may be investigated before being attacked (e.g., in mice: Banks,
1962; in the Norwegian lemming: Arvola et al., 1962). Bateson (1964) found
that attack of a moving object by chicks often occurred suddenly at the
change-over point between avoidance and following. Exploratory responses
do not generally occur during schedule and extinction-induced frustration
(see Falk, 1971, for a discussion of the types of behavior which do occur in
such situations).
In conclusion, it is argued that while there may be an overlap in some
of the conditions producing fear behavior, attack, and the other responses
discussed in this section, this overlap is only a partial one for each of these
possible alternative forms of behavior. It is therefore argued that a "general
drive" interpretation of situations producing fear behavior and attack is not
supported by the evidence.
V. INTERACTION OF SITUATIONS EVOKING ATTACK AND

FEAR BEHAVIOR
In addition to arguing that the situations described in the previous sec-

tions increase the probability that attack or fear behavior will occur, it is
also suggested that the interaction of any of these situations is, within
certain limits, an additive one. (It will be argued in subsequent sections that
they all involve a similar type of mechanism.)
Thus novelty and pain in combination are found to be more effective in
eliciting attack than either situation alone (in rats: Galef, 1970b,c; in mice:
Legrand and Fielder, 1973), providing that pain is not too severe (Galef,
1970c; Legrand and Fielder, 1973). Ulrich (1966) also reported that no
fighting behavior occurred when pairs of rats were placed on a floor which
became gradually hotter, but that it did occur when rats were placed on a
preheated floor of the same temperature. Thus, stimulus change in com-
bination with noxious stimulation elicited attack, but the noxious stimula-
tion alone, introduced gradually, did not produce attack.
Similarly, "frustration" and pain appear to exert an additive effect.
Hayes et al. (1969) found that a low fixed-ratio schedule was insufficient to
produce attack in female rats, but that the addition of a mild electric shock
did produce attack under such circumstances. Frustration can also increase
the probability of attacking an unfamiliar conspecific; thus, Lagerspetz and
Nurmi (1964) reported that thwarting increased the level of attack and bit-
ing shown by male mice toward an unfamiliar male conspecific. Finally, the
combination of two different forms of novelty-changing the conspecific at-
tacked and changing its location-produced a greater increase in attack by
the male three-spined stickleback than did either change by itself (Peeke
and Vene, 1973).
The precise way in which the situations referred to above interact with
one another under a wide variety of circumstances cannot be deduced with
certainty from the few studies described in the previous paragraphs, but the
evidence so far does clearly indicate an additive interaction between the fac-
tors involved, rather than independent effects on different types of ag-
gression (as might be predicted by Moyer, 1968, for example), or inhibition
of one form of aggression by another. It therefore seems plausible to sug-
gest that tile effects of pain, novelty, and frustration operate on a common
mechanism at some point in the system (although this may be on the output
side.) The question of possible common factors being involved in the causa-
tion both of attack and of fear behavior is now discussed in more detail.
VI. COMMON PROPERTIES OF SITUATIONS EVOKING

ATTACK AND FEAR BEHAVIOR
In this section we consider whether situations evoking attack or fear

behavior possess any readily identifiable common properties. Hinde (1970)
remarked on the similarity between the various situations eliciting ag-
gression (i.e., pain, frustration, and the proximity of a conspecific), and he
suggested that a common definition might be made to include all of them.
Gray (1972), discussing the hypothetical central state "fear" induced by
novel and aversive situations, argued that this state was qualitatively
equivalent to "frustration." Evidence to support this hypothesis was derived
from several sources: e.g., that similar responses occurred following both
lSl John Archer
fear-evoking and frustration-evoking stimuli; that a CS presented during ex-

tinction acquired aversive properties; and that, like punishment, frustration
increased the vigor of appetitive responses.
Thus, Hinde has suggested that there is a similarity between situations
evoking attack and Gray has essentially suggested the same for "fear." In
presenting the classification proposed in the previous sections, it was argued
that fear behavior and attack both occur in response to similar situations.
From the suggestions of Hinde and Gray, the next logical step is to look for
common factors in the causation of both attack and fear behavior. Gray
(1972) suggested four categories of stimulus properties which evoke "fear,"
of which intensity and novelty are the most meaningful. We might therefore
begin by suggesting that the situations listed above all involve high intensity
or high novelty of stimulation, but this would not cover the various frustra-
tion situations, which involve only a minor degree of novelty; the important
point here would seem to be that there is a change in the stimulus situation
from that which the animal was expecting.
Hebb (1946) suggested that "fear" and "anger" arise not from a
particular set of stimulus properties, such as novelty, but from the dis-
crepancy between what had been expected or had frequently been
experienced by the animal. Such a definition could, with some qualifica-
tions, cover all the situations described in the previous section.
First, we would assume that the animal maintains a continuous com-
plex representation of expectancies based on: (a) the sum total of its
cAperiences, i.e., what types of relations particular stimuli generally have to
one another, what types of relationships commonly Qccur, what types occur
less often, and what types never occur; (b) more precise spatial representa-
tions of particular habitually used areas of the environment, so that certain
habitually experienced stimuli become expected; (c) temporal representa-
tions of the expected outcome of particular sequences of previously
rewarded responses-thus a sequence culminating in a rewarding event
would come to evoke a fairly precise expectation when the animal again
found itself in a similar situation. It is also suggested that just as rewarding
events become easily incorporated into the animal's expectation model, the
model is particularly resistent to incorporating intense or painful stimuli.
Second, we would assume that such a model or models of the animal's
expectations are continually being compared with information derived from
its perception of incoming events, and any large discrepancy is the factor
initiating a motor command to attack or show fear behavior.
The main objection that. might be raised concerning such a possible
type of mechanism initiating all forms of attack and fear behavior involves
the case of pain-elicited aggression, which, it has been argued (e.g., by Ul-
rich and Azrin, 1962; Ulrich, 1966), is a reflexive response to pain that
persists over many repeated trials (e.g., Ulrich and Azrin, 1962) and often
shows a gradual increase over repeated sessions. Such an increase contrasts

with the typical habituation of aggressive responses found for territorial
fighting (e.g., Peeke et al., 1969) and would seem to suggest, at first sight,
that a more specific mechanism, and not a general discrepancy between ob-
served and expected stimulation, has to be postulated to explain this type of
attack. However, Crosby and Cahoon (1973) pointed out that most previous
studies have employed shock-intervals of constant length, and they argued
that superstitious learning, related to the consistency of shock-offset, might
account for the continuation and persistence of the so-called "reflexive
fighting." In other words, the rats were being reinforced for attacking by
the offset of shock. Crosby and Cahoon (1973) themselves found that attack
by rats that were given sessions of shock presentations of a constant dura-
tion increased over sessions, whereas attack by rats that were given shocks
of variable duration declined over sessions. These results therefore support
Crosby and Cahoon's suggestion; and they further suggest that when the
subject is unable to predict the timing of shock-offset, shock-elicited attack
does show the decline over test sessions characteristic of habituation in
other forms of aggression. The ability of rats to maintain this type of fight-
ing for 10,000 responses, lasting 7.5 hr (Ulrich and Azrin, 1962), would
seem to eliminate the possibility that the waning of attack could be at-
tributed to fatigue.
It is therefore suggested that animals can habituate even to the aversive
stimuli employed in studies of pain-induced attack. Ulrich's study (1966), as
noted above, showed that no attack occurred if the rats were placed on a
gradually heated floor, but that when they were placed on a preheated floor
of the same temperature, they did fight. This again suggests that habituation
to stimuli evoking pain-induced attack occurred. In the theoretical scheme
suggested here, these results can be explained in terms of the animal being
able to incorporate the input from the aversive stimulus into the complex ex-
pectation model referred to above. The model may be more resistant to the
incorporation of painful rather than neutral stimuli, and they may take
longer to incorporate, but it is nevertheless envisaged that the same general
comparison process is involved with all forms of input.
The evidence discussed above suggests that the sudden presentation of
a painful stimulus constitutes the most important factor for the evocation of
pain-induced attack. From this, one would predict that acute pain presented
at infrequent intervals would be more likely to evoke attack than would
more frequent presentations of less intense pain. A study by Hutchinson et
al. (1971) supports this prediction. They found that pain-elicited biting by
squirrel monkeys decreased within and across test sessions, when the shock
was of a low intensity and high frequency. In contrast, when a high-intensity
and low-frequency shock was used, a slight increase in biting occurred
within and across sessions.
254 John Archer
It is therefore suggested that the common factor involved in situations

evoking attack is that they provide a discrepancy of some magnitude from
the animal's expectation model, or models of its environment. This process is
regarded as essentially the same as that suggested by Sokolov (1960) in rela-
tion to the orienting response: he proposed that when the sensory input did
not coincide with the animal's previously established "neuronal model,"
certain vegetative changes, comprising the orienting response, occurred (such
changes have the effect of maintaining the sense organs in contact with the
particular sensory input and preparing the animal for further more specific
responses). Thus, it is suggested, first, that any perceived discrepancy ac-
tivates the orienting response, which has the effect of increasing the input in
relation to the relevant stimulus source; and second, that if the discrepancy is
sufficiently large, a system capable of producing fear behavior or attack also
becomes activated. A more detailed diagrammatic representation of these
suggestions is presented in Fig. 1: the feedback loop on the top left-hand
corner represents part of the orienting response; "discrepancy" is the dis-
crepancy between observed and expected events, which if above a certain
value (y) activates the "attack-fear system" (by a motor command).
The present hypothesis postulates a common type of mechanism for
situations which can evoke either attack or fear behavior. It does not,
however, necessarily suggest that the same neurophysiological structures or
the same types of neuronal models are involved in all the different types of
attack-evoking and fear-evoking situations. The sharing of common
properties between the various situations lies in the way the different
processes can be represented at the behavioral level by a single model.
It would appear, from the literature reviewed above, that there are
three basic variations on this model, representing pain, novelty, and frustra-
tion. These involve, respectively, a form of aggression involving a very
simple comparison process, a comparison with the animal's representation
of its environment in spatial terms, and a comparison with the animal's
representation of its environment in temporal terms (Le., its expectations).
It may be helpful to amplify this type of reasoning by placing it into a
possible evolutionary perspective. It is envisaged that the first stage in the
evolution of attack and fear-avoidance systems occurred because animals
had to counteract stimuli in their environments that were capable of pro-
ducing physical damage. Thus, most animals from the Protozoa onwards
show some form of escape or avoidance responses, and the selective ad-
vantage of these is obvious. An alternative to avoiding or escaping from a
noxious' stimulus would be for the animal to remain where it is and to
remove the noxious stimulus from its vicinity. Such aggressive responses
(found mainly in arthropods and vertebrates: Scott, 1971) would have
clear advantages over escape where the noxious stimulus could readily be
Aggression and Fear in Vertebrates 2SS
EXPECTATION +
copy
+
if discrepancy> 0
if discrepancy> 0
+
command for
command for
orienting response threshold gate
tendency toward
tendency toward fear behavior
attack (magnitude bl (magnitude _I Internal factors
biasing animal
toward octive (AI
or passive fear
behavior
localization of
stimulus
discrepancy (Ll.
and other
+ +
Environmental consequences
of behavior +
Fig. I. Diagrammatic representation of factors influencing the occurrence of aggression and

fear behavior.
256 John Archer
removed, and where escape would have placed the animal in a suboptimal
environment.
As animals evolved more sophisticated sensory equipment, the problem
of deciding which stimuli were potentially noxious became important, be-
cause it would be advantageous for the animal to react to such stimuli
in advance rather than waiting for actual damage to occur before taking
action.
It is therefore possible that pain-induced attack represents the simplest
form of attack; i.e., any damage to the animal's body is counteracted by at-
tack or escape responses, and pain receptors provide a way of monitoring
such changes. To fit such a mechanism into the discrepancy model would
entail a central monitoring process which matches incoming stimuli (pain
signals) with a neuronal model (representing no signal), so that in this case
the resulting discrepancy would represent the extent of the pain stimulus. A
more sophisticated mechanism would entail changes in the neuronal model as
a result of chronic pain, so that the neuronal model could be changed to ex-
pect a low level of chronic pain, and would not register a discrepancy when
this level of pain occurred. Thus habituation could occur in response to
prolonged painful stimuli.
Where the animal possessed the neurosensory equipment to react to
potential rather than to real noxious events, the comparison process would
become more important. Here the animal's "defended area" would no longer
cease at the border of its body and the environment, and therefore it could
not rely on responding merely to tactile stimuli of sufficient intensity to ac-
tivate pain receptors. The simplest way of assessing potential danger would
appear to be for the animal to respond to any major change in its immediate
environment. Thus the animal builds up a neuronal representation of the sta-
ble properties of its surrounding area-defined either in terms of its own body
("individual distance") or in relation to a fixed area of the environment ("a
territory")-and the input is repeatedly compared with this representation.
Again, any large discrepancy evokes attack or fear behavior.
A more subtle form of assessing environmental change occurs in ani-
mals capable of building up temporal as well as spatial expectations of what
will occur in their environment. Thus, the animal builds up a representation
of what is expected on the basis of past reinforcement. Again, the input is
matched with this neuronal representation as the animal performs the
learned response sequence, and the discrepancy between the input and the
neuronal model may occur in one of several ways; e.g., by the absence or
delay of the expected reward, or by the animal's being physically prevented
from completing the response. Other forms of mismatch are also possible,
but these have not usually been investigated in studies of frustration-induced
attack. Examples include the introduction of a red flashing light into a
runway which doves had learnt to run for water (McFarland, 1966), and
changes in the visual properties of a runway which chicks ran for a food
reward (Archer, 1974a).
In this way, we can suggest that the situations evoking attack or fear
reviewed above form a graded sequence, involving increasing reliance both
on cognitive interpretation and on matching an external event to a central
representation of the environment (summarized diagrammatically in Fig. 2).
I t is therefore suggested that these different forms of aggression and fear
evoking situations involve a similar type of comparison process, but it is not
suggested that the same types of neural structures are involved. In fact, it is
(1) Pain
•t Intrusion into
body surface
(2) Individual Intrusion into

distance area surrounding body
.
intrusion
(3) 'Territorial'
or familiar
(
/
--- '"
\
-- Intrusion into larger area
located in space independent
of the animal's body
.
area intrusion
\ J
'--t--/
(4) Anima,
surrounded by /----" Animal's familiar space
a
is replaced by novel
novel area ( \ area
\ ~
'- ~./
/
(5) Thwarting O-reward
t 0- '"thwarting'"
obstacle
f I - activated expectation
copy for reward Goal-directed
responses
involving
(6) Frustrative
non· reward (extinction) t
,-,-previously rewarded
situation
hypothetical
'expectation copy'
,,_expectation copy for reward, which
for reward is prevented from
being matched
,(7) Time out from positive O-reward by environmental
t
reinforcement (low events.
O-temporal
reinforcement schedules) barrier
.. _ activated expectation
copy for reward
Fig. 2. Diagrammatic representation of situations evoking attack and fear behavior, shown as
a series of increasing complexity.
258 John Archer
suggested that the causation of pain-induced attack may be essentially (but

not entirely) different from the causation of novelty-induced and frustra-
tion-induced attack.
We now turn from a consideration of the possible type of mechanism
involved in initially determining that the animal will show attack or fear be-
havior, to a consideration of what other factors may decide which of the
two types of behavior occurs in a given situation. A two-stage process is
envisaged, since many internal variables seem to affect the general predis-
position of an animal to attack or show fear behavior, whereas other,
mainly external, factors seem to influence the details of the final outcome.
The remainder of this paper is therefore devoted to reviewing evidence con-
cerning factors that determine whether attack or fear behavior is likely to
occur in a given situation, presented as factors affecting the two
hypothetical decision processes, 1 and 2 (shown diagrammatically in Fig. 1).
VII. INTERNAL FACTORS AFFECTING DECISION PROCESS 1
A. Degree of Discrepancy between Observed and Expected Events
Berkowitz (1962) argued, from human data, that as the intensity of

noxious stimulation or anticipiated noxious stimulation increased, "fear"
tt:llded to replace "anger," because the tendency for fear to occur rises
more rapidly than does the tendency for anger to occur. A similar type of
relationship is envisaged in the present theoretical scheme: as discrepancy
(see Fig. 1) increases, the tendencies for both fear behavior and attack
increase, but the tendency for fear behavior rises more rapidly than that for
attack. These relationships are shown diagrammatically in Figs. 3,4, and 5.
There is some experimental evidence from pain-induced aggression ex-
periments to support such a hypothetical relationship. Fighting behavior in
rats increased as shock-intensity increased from 0.5 to 2.0 rnA, but at
higher intensities the response rate was reduced (Ulrich, 1966). A similar
relationship occurred with increased shock duration (Azrin, et a/., 1964b;
Ulrich, 1966) and over repeated trials (Reynierse, 1971). Since it is observed
that at higher intensities and durations escape behavior competes with at-
tack (e.g., Ulrich, 1966), there seems to be a maximum intensity and dura-
tion for eliciting pain-induced attack, which, if exceeded, results in the re-
placement of attack by fear behavior.
A similar relationship has been reported in other species. Azrin et a/.
(1964a) reported that biting in squirrel monkeys increased as shock
intensity increased, for the lower shock intensities they used. Melvin and
Anson (1969), and Melvin and Ervey (1973), found that moderate shock
/
/
/
/
>-
u A
z
w
o
z
W
I-
W
III /
Z /
oCL /
/
/
III ;'
/'
W ./
cr
DISCREPANCY MAGNITUDE
Fig. 3. Graph showing the hypothetical relationship between discrepancy mag-

nitude and tendency to attack (A) or show fear behavior (E).
:s::
U
<{
l-
I-
<{
lJ...
o
>-
I-
...J
CO
<{
CO
ocr
CL
DISCRE PANCY MAGNITUDE

Fig. 4. Graph showing the hypothetical relationship between discrepancy mag-
nitude and probability of attack.
260 John Archer
~
o
>
<l:
I
W
ro
0::
<l:
w
lJ...
"DECLINE IN
lJ...
o ATTACK PROBABILITY
..J
(l)
<l:
CO
o0:: RISE IN ATTACK
CL PROBABILITY
DISCREPANCY MAGNITUDE
Fig. 5. Graph showing hypothetical relationship between discrepancy magnitude
and probability of fear behavior.
intensIties intensified the aggressive display of Siamese fighting fish,

whereas high intensities suppressed the display (see also Adler and Hogan,
1963). In response to a wing-shock of 5.5 rnA, hens attacked the electrode
wires, whereas at a shock intensity of 8.5 rnA immobility was the typical
response (Bracco and Miller, 1972).
Results obtained by Legrand and Fielder (1973) from an experiment
involving both shock and novelty are also consistent with the relationship
suggested above. In mice, they found that unfamiliarity with the opponent
increased attack at low shock intensities, but depressed attack at higher
shock levels.
Further research is required to determine whether this type of relation-
ship can be extended beyond pain-induced attack, as suggested by the
present model. A study by Bateson (1964) does, however, provide some sug-
gestive evidence from a situation not involving shock. He found that when
chicks were tested in an imprinting situation they initially avoided the mov-
ing object. With increasing duration of exposure, more birds changed from
avoidance to approach of the object and often vigorously pecked at the time
of change-over. In terms of the present theory, it is envisaged that the object
initially represented a high degree of novelty, and that this waned as time
passed: thus, discrepancy between observed and expected events was cor-
respondingly high at first, rapidly declining thereafter, and fear behavior oc-
curred until the discrepancy had declined to a point sufficient for attack
behavior to become the dominant response tendency. It was soon replaced,
however, by a third type of response-following-which presumably occurs
when the discrepancy has reached a lower level.
Other studies can also be interpreted in terms of these hypothetical
relationships. Kuo (1960d), for example, found that aggressive quail that
would attack any conspecific in familiar surroundings would not do so in an
unfamiliar setting, but merely made undirected running and jumping
responses. Thus the low or moderate stimulus discrepancy involved in at-
tacking other con specifics in a familiar area was replaced by a much larger
discrepancy in a novel environment, and this was therefore sufficient to
produce fear behavior instead of attack, despite other factors (notably, rein-
forcement of attack) which would generally bias the decision process fairly
heavily in favor of attack.
Similarly, ducklings raised in isolation, and imprinted to a moving
stimulus, attacked socially reared ducklings when the imprinted stimulus
was present, but not in its absence, in which case attack was replaced by
distress calling and avoidance (Hoffman and Boskoff, 1972). Here, the
presence of the socially-reared bird provided an unfamiliar stimulus, pro-
ducing a moderate and localized discrepancy from the expected environ-
mental conditions; but the absence of the familiar imprinted stimulus
constituted a much larger discrepancy, producing fear behavior.
These few examples, although not conclusive, do suggest that the
increase in a general factor, such as stimulus discrepancy, can account for
changes from attack to fear behavior in situations where the other causal
factors remain constant.
B. Strength of Ongoing Motivation
From studies of frustration-induced attack, it is reasonable to suggest

that if an animal is pursuing a goal-directed activity, a discrepancy between
observed and expected events which interferes with such an activity is likely
to produce attack behavior; on the other hand, a discrepancy which does
not interrupt such a goal-directed activity (e.g., the appearance of a novel or
unexpected event unrelated to the animal's current response) is less likely to
evoke attack.
When studying this variable, the combined effect of the animal's degree
of food deprivation and the number of reinforced trials (i.e., strength of mo-
tivation) are the important factors to manipulate. Both are necessary, since
deprivation alone will not induce attack (see a later section, and Berkowitz,
1969, in relation to "frustration").
262 John Archer
It is therefore suggested in the present theoretical scheme that a factor

termed "reward expectancy" (Fig. 1) is activated when an animal is in a
suitably motivated state (e.g., food deprived), and pursues any activity
which has habitually resulted in a reward on previous occasions. This factor
affects the expectation copy (Fig. I), i.e., it influences what the animal ex-
pects from its environment at that moment; but it also affects the decision
mechanism for the attack-fear system, biasing this in the direction of at-
tack: this influence, of course, only affects overt behavior when a command
for decision process 1 (Fig. 1) has activated that decision mechanism.
C. Hormonal Factors
1. Sex Hormones
Males typically show greater readiness to attack than do females, in

proximity-induced and pain-induced attack (Table I), but sex differences
have not been investigated for frustration-induced fighting (Gallup, 1965,
Table I. Typical Sex Differences in Aggression Reported in the Literaturea
Sex showing
f_ '.thor and date Species Type of aggression more aggression
Evans, 1936 Anolis lizard P M

Lack, 1939 Robin P M
Seward, 1945 Lab. rat P M
Clarke, 1956 Vole P M
MarIer, 1956a, b Chaffinch P M
Conner and Levine, 1969 Lab. mouse S M
Watson, 1970 Red grouse P M
Norris and Adams, 1972 Gerbil P M
Davis and Khalsa, 1971 Lab. rats M no difference
(SD & LE)
Knutson and Hynan, 1972 Lab. rats S M
(LE)
Hutzell and Knutson, 1972 Lab. rat S M
(LE)
Milligan et 01., 1973 Lab. rat S M
(LE)
Powell et 01., 1971b Lab. rat S no difference
(SD)
Milligan et 01., 1973 Lab. rat S no difference
(SD)
a SD: Sprague-Dawley strain, LE: Long-Evans strain, P: Proximity·induced, S: Shock-

induced, M: Morphine-withdrawal.
using female rats, remarked on how little fighting occurred after nonrein-
forcement, compared with other studies, which generally used males). An-
nual variations in "aggression," connected with the breeding cycle and with
differences in testis activity, are also widespread in many bird and mammal
species; thus the male only holds territories or shows aggressive behavior at
certain times of the year when testicular activity is high (e.g., in the red
grouse: Watson, 1970; in red deer: Lincoln et al., 1972; in the elephant:
lainudeen et al., 1972).
In the cases both of sex differences and of seasonal variations, the
higher incidence of aggression is generally associated with the presence of
physiological levels of androgens. The link between androgens and readiness
to attack is well established for a number of species (evidence summarized
in Table II). From studies of mice, it appears that exposure to androgens
during the developmental period is an additional essential factor in mam-
mals, without which later physiological levels of androgens are ineffective
(Edwards, 1968; Bronson and Desjardins, 1968, 1970; Conner and Levine,
1969; Edwards and Herndon, 1970). Some important qualifications to these
findings are that there is an interaction with the genotype (Vale et al., 1972)
and that the critical period for exposure to early hormones is less restricted
than was originally supposed (Edwards, 1970).
There are a number of exceptions to this typical picture of the male, as
a result of his higher androgen levels, being more aggressive than the fe-
male. Tht"~e exceptions can usefully be considered as two main classes: first,
modification of the typical enhancing effects of androgens by habit or
experience; and second, the existence of entirely different hormonal in-
fluences in some other species.
The former class involves past experience overriding an experimentally
induced increase or decrease in androgen levels, which would in other cir-
cumstances have facilitated or inhibited attack. Such a phenomenon has
been reviewed by Guhl (1964) in relation to the peck order of chickens: he
described how the typical increase in aggressiveness produced by androgen
treatment when the injected bird is matched with a total stranger is masked
in the bird's usual social setting by habits developed in relation to the peck
order (he termed this phenomenon "social inertia"). Other studies, carried
out on mice, have demonstrated "inertia" of the opposite type: if castration
is carried out after maturity or after the development of isolation-induced
fighting (discussed in a later section), the readiness to attack continues for
some time afterwards (Uhrich, 1938; Beeman, 1947b; Sigg et al., 1966). In
another study on mice, carried out to determine the combined and separate
effects of castration and "competitive training," Bevan et al. (1960) showed
that pretest experience (whether the animal won or lost encounters) had a
far greater influence on later aggressive behavior than did androgen levels.
Other studies have also shown that the relationship of androgens to fighting
264 John Archer
Table II. Summary of Typical Studies on Androgens and Aggression"
Type of
Author and date Species aggression Main finding
Fredericson, 1950 Lab. mice P Maturation of aggressive be-

Levy and King, 1953 Lab. mice P havior coincided with rise
Lagerspetz and Talo, Lab. mice P
in androgen level at pu-
1967
Hutchinson et al.,
1965
Lab. mice S J berty
Lincoln et al., 1972 Red deer P

Uhrich, 1938 Lab. mice P
Beeman, 1947a Lab. mice P
Sigg et al., 1966 Lab. mice P Castration reduced readiness
Barfield et al., Lab. rats P to attack or prevented
1972 development of aggressive
Vandenberg, 1971 Hamsters P behavior
Payne, 1973a Hamsters P
Powell et al., 1971 a Lab. rats S
Milligan et al., 1973 Lab. rats S
]
Christenson et al., Gerbils P
Castration increased readi-
1973
ness to attack
Anisko et al., 1973 Gerbils P
Lagerspetz et al. Lab. mice P
~967 Aggressiveness correlated
Kuo,1960c Quail F positively with testis weight
1
Rose et al., 1971 Rhesus P or plasma testosterone
monkeys levels
Persky et al., 1971 b Man Q
Beeman, 1947 a Lab. mice P
Levy and King, 1953 Lab. mice P
Lagerspetz and Talo, Lab. mice P
1967
Barfield et al., Lab. rats P
1972 Injected androgens increased
Kiryukhin, 1972 Lab. rats p readiness of castrates to at-
Vandenberg, 1971 Hamsters P tack, or produced early
Payne, 1973a Hamsters P development of aggressive
Horridge, 1970 Domestic Fowl P behavior in immature
chicks males
Lincoln et al., Red deer P
1972
Andrew, 1972a Domestic Fowl P
chicks
" P: Proximity-induced, S: Shock-induced, Q: Questionnaire measures, F: Frustration-induced.

b Not confirmed by other work (Doering et al., 1974).
is more complex than merely providing a facilitating influence whatever the

experimental circumstances (e.g., in mice; Bevan et al., 1957; in pigeons:
Lumia et al., 1973; in quail: Kuo, 1960d). These studies show that past
experience is particularly important when assessing hormone effects on
fighting; indeed, it can override such hormonal effects in many cases (see
Gottier, 1972, and Bevan et al., 1960, for further discussions of this point).
We now turn to a brief consideration of species in which hormones
other than androgens have been found to influence aggressive behavior.
Studies reviewed below, particularly those by Wootton (1970) on the
stickleback, and by-Crook and Butterfield (1968) on Quelea, suggest that, in
these species, androgens only influence aggressive encounters concerned
with nest possession. Further, it seems likely that the male hormone exerts
this effect indirectly, through a more general influence on motivational
mechanisms concerned with nesting, since it also facilitates nest building it-
self (Hoar, 1962; Crook and Butterfield, 1968).
Thus, in the three-spined stickleback, Hoar (1962) and Baggerman
(1966) provided indirect evidence to suggest that gonadotropic hormones
facilitate attack during the pre nesting phase, whereas during nestbuilding
the level of aggression seemed to be largely sensitive to the level of
androgens. More direct evidence by Wootton (1970) showed that during the
nest building phase, castration selectively reduced the attack shown by nest-
holding males to other males, but had no effect on other forms of attack.
Comparable findings have been reported for two bird species. In star-
lings, a gonadotropic hormone, luteinizing hormone (LH), is more
associated with aggressive behavior than is testosterone (Davis, 1957; Mat-
thewson, 1961). LH also appears to be important in maintaining attack
associated with individual distance in Quelea (Crook and Butterfield, 1968;
Lazarus and Crook, 1973), although, as with the stickleback, testosterone
produced a higher level of encounters over nest material (Crook and But-
terfield, 1968).
Although males are generally more ready to attack than are females
(see Table I), a number of exceptions have been found. Moyer (1972) re-
viewed some naturalistic accounts of territorial defense where females as
well as males attacked intruders; examples were taken from gibbons, rhesus
monkeys, lions, and voles. However, the observation that the female some-
times aids the male in territorial defense does not provide data on whether
in these cases the female shows an equal or greater tendency to attack com-
pared with the male.
It is known that, in many species, females show a form of territorial at-
tack, but this is restricted to the period of infant protection. Such "ma-
ternal aggression" (Moyer, 1968) has been studied experimentally in mice.
Dennis st. John and Corning (1973) found that lactating females were just
as ready to attack a restrained target male as were unrestrained males, but
266 John Archer
that there was substantially less attack in a strange cage than was the case
for males. Gandelman (1972) also reported that maternal aggression was
strongly associated with the lactation period, which suggests that the
hormone prolactin might play an important role in facilitating this type of
attack.
In other rodent species, nonlactating females have been observed to
show high levels of aggression. For example, Frank (1957) reports that in
European Microtus species, the female occupies a territory surrounding her
burrow and drives away strangers of either sex except during the estrus period
(Frank, 1957). Laboratory studies of paired encounters between strangers in
a novel environment have shown that nonreceptive female hamsters and
gerbils show as much tendency to attack one another as do males (Payne
and Swanson, 1970; Swanson, 1973). In heterosexual encounters, even when
matched for body weights, diestrus female hamsters win more often than do
males. This finding was partly due to a lesser readiness of males to attack
females than to attack other males (Payne and Swanson, 1972a,b; Payne,
1973b), which has been found in several other species (e.g., chaffinches:
Marler, 1956a; mice: Mugford and Nowell, 1971b). Circulating levels of
progesterone were found to be responsible for this phenomenon in the
hamster (Payne and Swanson, 1972a), and also to exert a direct enhancing
effect, thereby increasing aggression toward females as well as male op-
ponents (Payne and Swanson, 1971 a, 1972a). Injected progesterone also
enhances male aggression in the same manner (Payne and Swanson, 1972b).
Androgens also facilitated attack in encounters between male hamsters
in one of the pair's home cages (Vandenberg, 1971; Payne and Swanson,
1971b, 1972b, Payne, 1973a), but did not facilitate attack toward females
(Payne and Swanson, 1972b). It was also shown that ovarian implants
(Payne and Swanson, 1971b), or injections of estradiol (Vandenberg, 1971;
Payne and Swanson, 1972b), enhanced the aggressiveness of castrated males
toward a male opponent, but that neither estrogen nor testosterone affected
the aggressiveness of ovariectomized females (Vandenberg, 1971; Payne and
Swanson, 1971a, 1972a).
Finally, it was shown that neonatal injections of testosterone increased
the aggression of adult males so that it was similar to that of normal fe-
males (Payne and Swanson, 1972c, 1973), and it was suggested from this
and from other studies that the typical "masculinization" of the developing
nervous system by androgens is incomplete in the hamster, compared with
rats and mice.
There are, therefore, some substantial variations in the hormonal con-
trol of aggression by sex hormones within the few species of birds and mam-
mals which have been studied in any detail. Although it is important to bear
in mind these variations, androgens have nevertheless been shown to com-
monly exert an enhancing effect on attack. Little consideration has been
given, however, to the possible ways in which androgens could affect ag-
gression. One possible way is by a direct action on brain mechanisms
facilitating attack; in the present theoretical scheme, this would involve de-
cision process 1, which determines the general disposition towards attack or
fear behavior, so that attack is made more likely and fear behavior less
likely. In domestic fowl chicks, Archer (l973a,b) found that testosterone
lowered the amount of distress calling in a novel environment, and reduced
the duration of "freezing" when behavior had been interrupted by a loud
bell, thus tending to support this hypothesis. However, more recent results
show that testosterone can also prolong immobility in novel, startling, and
frustrating situations (Archer, 1973c, 1974b, unpublished data), and may in
these cases also decrease the frequency of active escape.
One possible way in which androgens might affect the type of fear be-
havior which occurs, as indicated by these results, is through an attentional
change. Andrew and Rogers (1972), Andrew (1972a,b), and Rogers (1971,
1974) found changes in search behavior in testosterone-injected male chicks
which suggested that their attention to particular stimuli was more
persistent once they had begun to attend. Andrew and Rogers (1972) have
suggested that such testosterone-induced changes could provide an indirect
way for the hormone to increase fighting behavior, by making attack more
persistent and hence more prolonged.
Attentional changes could also be involved in the initial decision
whether attack or immobility, rather than active escape and associated be-
havior, should occur in response to a particular stimulus. An initial period,
when attention is focused on the object to be attacked, occurs before most
instances of attack; thus, inhibition of any tendency to switch attention
from such an object would provide one indirect means of facilitating attack,
which could alternatively result in immobility if the discrepancy from what
was expected were high. Conversely, the probability of the occurrence of fear
behavior associated with frequent shifts of attention, such as active escape
and peeping (Andrew, 1973; Archer, 1973c), would be decreased. This sug-
gestion, although at present speculative for animals other than the chick, is
made plausible by the commonly observed periods of visual fixation that oc-
cur as threat signals before fighting (Vine, 1970).
A further possible way in which androgens could facilitate attack is
again an indirect one. Androgens are anabolic, promoting muscular
development, which would enable an animal to be more successful in a
fight. Thus, reinforcement of attack would be more likely, and it would tend
to occur more often on similar occasions in the future. Even discounting
such an effect of testosterone on reinforcement, the anabolic effects of
steroids could facilitate fighting directly. Kuo (1960b) found that isolated
male and female quail became better fighters when given thiamin or pro-
tein-enriched diets, whereas a deficiency of these nutrients produced poorer
268 John Archer
fighters. ··Better" and "poorer" referred to increased or decreased

probabilities of attack, restrictiveness of what would be attacked, and where
it would be attacked.
In other studies, body weight and success in fights have been found to
be positively correlated (e.g., Uhrich, 1938; Payne and Swanson, 1970), and
Evans (1964) found in wild rats that body weight was positively correlated
with threats, but not with attacks. Bevan et al. (1960) found that the effects
of castration on the attack tendencies of mice with differing competitive
histories related more to how body weight was affected by castration than
to a lack of male hormone per se. Direct effects on muscle growth
may therefore provide an important-although usually neglected-means
whereby androgens can influence aggression.
These possible ways in which androgens could facilitate attack have
been discussed in some detail. For the purposes of the present theoretical
scheme, outlined in Fig. 1, these details are omitted; androgens are shown
as increasing the probability of attack by acting on decision process 1, but
the precise way in which this occurs is not stated.
2. Hormones oj the Pituitary-Adrenocortical System
There is clear evidence that in isolated male mice, manipulations of the

pituitary-adrenocortical axis can influence readiness to attack (e.g., Brain
et al., 1971; Kostowski et al., 1970; Harding and Leshner, 1972). More
precisely, ACTH appears to be the important hormone, high levels causing
reduced aggressiveness and low levels causing increased aggressiveness
(Candland and Leshner, 1974; Leshner et al., 1973). This effect is inde-
pendent of the influence of androgens discussed in the previous section
(Candland and Leshner, 1974; Leshner, 1974; Leshner et al., 1973).
The evidence at present suggests the following general pattern. Various
procedures which affect the level of ACTH, including prolonged social
isolation (which lowers basal adrenopituitary activity: Welch and Welch,
1969), also influence aggressiveness. This influence may involve either
several threshold effects (represented in Fig. 6) or a continuous modulation.
Thus, group-housed animals (higher ACTH levels than isolates) show a
lower level of aggressiveness than isolates do (Welch and Welch, 1969).
Various procedures which raise the ACTH levels of isolates, e.g., adrena-
lectomy or ACTH injection, lower their aggressiveness, so that it ap-
proaches that of grouped animals (Harding and Leshner, 1972; Brain et al.,
1971; Leshner et al., 1973; Candland and Leshner, 1974). Replacement
doses of 200 /J-g of corticosterone a day restore the aggressiveness of adrena-
lectomized isolates to that of intact isolates (Candland and Leshner, 1974).
This can be further raised by procedures which lower ACTH levels, e.g.,
LOW
<J)
<J)
GO GAb
w
z ,----
w IAor IT ,
>
<J)
<J)
w IorIAC
a::
(9
(9
<l:
IDorIH
HIGH
L OW -----------~) HIGH
ACTH LEVELS
Fig. 6.. The possible relationship between ACTH level and readiness to attack ("aggressive-
ness") in male mice. Abbreviations: I, isolation-housed; A, adrenalectomized; D, dexameth-
asone-injected; H, hydrocortisone-injected; C, corticosterone-injected; T, ACTH-injected.
(Notes: a It appears from Harding and Leshner, 1972, that G mice fought less than IA mice,
but no statistical treatment was given for these results. b Burge and Edwards, 1971, found no
difference between the aggressiveness of adrenalectomized and intact mice isolated for only a
short period of time, i.e., perhaps equivalent to grouped mice, suggesting that there is a limit to
the lowering "1 aggressiveness by high ACTH levels).
dexamethasone treatment (Brain et al., 1971; Candland and Leshner, 1974),

or injections of hydrocortisone (Kostowski et al., 1970).
Figure 6 summarizes the possible relationship between ACTH and ag-
gressiveness, based on the evidence reviewed in the previous paragraph.
There do, however, appear to be a number of limitations to this rela-
tionship. For example, Erskine and Levine (1974) found that complete sup-
pression of pituitary-adrenal activity in rats produced no decline in their
shock-induced aggression. This result can be accounted for either by dif-
ferences in the type of aggression (shock-induced as opposed to isolation-in-
duced) or by the different species used (rat as opposed to mouse). This
second variable is clearly the one responsible for another set of negative
findings: that ACTH administered to hamsters during a period of isolation
produced no modification of aggressive behavior (Brain and Evans, 1973).
It has also recently been found that manipulation of the adrenal-
pituitary system after, rather than during, an isolation period, fails to in-
fluence aggressiveness in the same way (Brain and Evans, 1973). A number
270 John Archer
of other findings (Poole and Brain, 1973), concerning the relationship

between the time of adrenalectomy or hormone injection, the period of
isolation, and the type of testing (cf. Burge and Edwards, 1971), suggest
that these variables can interact to produce some complex limitations on
the appearance of any modulating effects of the pituitary-adrenal system on
aggressiveness. Since this work has not yet reached a satisfactory conclud-
ing point, it is difficult to draw any firm conclusions from it at present. We
are nevertheless left with the question of whether ACTH is a factor in-
fluencing the probability of attack under the wide range of circumstances
discussed in this review, or under a narrower set, involving limitations by
sex, strain, housing conditions, and a variety of more specific factors. Only
future research will resolve this question.
A final point, of some importance within the context of the present
theory, concerns the suggestion made by Svare and Leshner (1973) and
Candland and Leshner (1974) that pituitary-adrenal manipulations in-
directly affect the readiness to attack by influencing the level of "fear."
Thus, ACTH would decrease aggressiveness by increasing the readiness to
show fear behavior. A number of studies using active and passive avoidance
learning support this suggestion in general terms: thus passive avoidance is
enhanced by ACTH in male rats (Levine and Jones, 1965), and extinction
of active avoidance responding is inihibited (e.g., Murphey and Miller,
1955; Greidanus and DeWied, 1971). DeWied et al. (1970) showed that
ACTH also improved the rate of acquisition of an avoidance response when
hypophysectomized rats were used (cf. little or no effect on intact animals:
Murphey and Miller, 1955; Weijnen and Slangen, 1970).
A major difficulty with this general explanation is that the portion of
the ACTH molecule shown to affect conditioned avoidance responding, the
4-10 residue (DeWied et al., 1970), has been shown to exert no effect on
fighting in isolated mice (Brain, 1972; Brain and Evans, 1973). Another dif-
ficulty is that the work on ACTH and avoidance behavior has involved rats,
whereas that on aggression has used isolated mice. Clearly, these inconsis-
tencies provide major objections to the theory that ACTH is exerting a
single effect in the two types of experiments. Only future experiments,
designed to test the hypothesis rigorously, will resolve the question.
3. Catecholamines
Earlier clinical work suggested that, in human subjects, anger directed

outwards is associated with high levels of noradrenalin, and anxiety is
associated with high levels of adrenal in (Funkenstein, 1955; Gray, 1972,
chapter 5, for reviews), although one cannot be certain from these studies
that the difference does not correspond to a less specific "active-passive"
dimension involved in preparation for the two sorts of behavioral response.
A more recent study by Marrone et al. (1966) suggested a similar rela-

tionship in Siamese fighting fish: they found an increased frequency of fin
extension and gill flaring when visually isolated fighting fish were swimming
in a noradrenalin solution, but not when they were swimming in an
adrenalin solution. Baenninger (1968), however, found that mirror-elicited
aggressive displays in isolated fighting fish were suppressed by both cate-
cholamines, when compared with controls in sodium chloride solution or
water. Adrenalin did cause some dominant fish to behave like subordinates,
in that they did not display to live submissive opponents. Baenninger's
results with mirror-images and conspecifics, therefore, clearly do not sup-
port the results of Marrone et al. with isolated fish, and do not support the
general hypothesis of an association between noradrenalin and aggression.
There is, however, some indirect evidence, in rodents, for an associa-
tion between noradrenalin and aggression. E.g., Eichelman et al. (1972) and
Thoa et al. (1972) showed that there was a significant increase in shock-
induced aggression in rats after depletion of brain noradrenalin by
6-hydroxydopa. Also, Lagerspetz et al. (1967) found higher levels of
noradrenalin levels in the brains of mice that were selectively bred for
aggressiveness.
We must therefore conclude that, while there is evidence to implicate
noradrenalin in events determining or associated with aggressiveness in
three very different vertebrate species, further research is required before
any more definite statement can safely be made.
D. Experiential Factors
Earlier in this paper, the various forms of conditioned aggression were

referred to. In this section, we consider how a wider range of experiential
variables, including those concerned with the outcomes of previous fights,
can affect the probability of a particular situation producing attack or fear
behavior. Early social experience, isolation housing, and learning associated
with the outcome of previous attacks or fear behavior, are all considered.
1. Early Social Experience
Studies on this subject have been concerned with the effects of the
cross-fostering of mice to a mother of another strain, or with the way in
which interactions with conspecifics of a similar age affect later readiness to
attack.
Southwick (1967b, 1968) cross-fostered, at birth, young adult male
mice from a nonaggressive strain to mothers from an aggressive strain, and
found that their adult aggressive behavior was significantly increased, com-
272 John Archer
pared with mice fostered to mothers from their own strain. Denenberg
(1973) has reviewed studies by himself and his co-workers involving foster-
ing mice to rat mothers; C57B 1 strain mice fostered in this way showed
pra·ctically no adult aggression, irrespective of whether they were raised
with or without peers (see below), whereas Swiss albino strain mice showed
no comparable absence of aggression as a result of fostering. These studies
thus demonstrated a very marked interaction between genotype and ma-
ternal treatment. Further studies of this type are required for a wide range
of species, since the whole question of how early interactions between
mother and infant can influence adult aggression may have important im-
plications for the understanding of factors affecting human aggression.
There have been few studies of this type using the other varieties of early
experience that are known to affect adult fear behavior (e.g., handling,
stress procedures); according to the present theory, these should also affect
aggression. Levine (1959) did, however, show that neonatal handling of rats
produced an increased readiness to attack when they were tested as adults.
Further studies have involved the influence on their later aggressive be-
havior of raising animals with or without other conspecifics. Kuo (1960d)
found that quail isolated from an early age were more ready to attack a
wide range of stimuli in a variety of environmental conditions than were
group-reared birds, after they had all experienced two months of training
for aggressiveness. Denenberg (1973) reviewed studies, on mice from his
laboratory, which showed that the absence of peer group interaction
between birth and weaning led to a significant increase in adult aggression,
regardless of post weaning experience. Thus the presence of peers during the
neonatal or posthatching period seemed to inhibit later aggression in these
two cases. In contrast, mice housed in isolation after they had been weaned
were found to be less aggressive than socially reared animals, when tested
with conspecifics in adulthood (King and Gurney, 1954; King, 1957; Denen-
berg, 1973). Similar results have been found for pain-induced attack in rats
(Hutchinson et ai., 1965; Hutzell and Knutson, 1972). In the study by King
(1957), it was further shown that ten days of social stimulation after weaning
(not necessarily involving competition or learning of aggression) was an im-
portant factor in accounting for the higher incidence of aggression in the
socially-reared animals.
These relatively few investigations serve to demonstrate the powerful
influences on adult aggressiveness of developmental variables associated
with social interactions. Such influences do not appear to be of a single
type, as shown by the different efferts of peer group interaction before and
after weaning, and they can involve interactions both with the mother and
with other developing conspecifics ( usually siblings). That such early in-
fluences on later aggression are not restricted to stimuli associated with

social behavior was shown by Levine's (1959) study, and it would seem that
the whole area of developmental influences on aggression should prove a
fruitful one for future research.
2. Social Isolation
In rats and mice, long-term isolation has been shown to produce an

increased readiness to attack a strange conspecific (e.g., Yen et al., 1962;
Hatch et al., 1963; Sigg et al., 1966). It is also associated with a number of
physiological changes, including lowered basal adrenocortical activity,
lowered basal adrenal medullary function, slower turnover and release of
brain catecholamines and serotonin, and hypersensitivity to sym-
patheticomimetic drugs (Welch and Welch, 1969). Isolated mice are also
reported to be hypersensitive to physical stimuli: Welch and Welch (1969)
argue that heightened "aggressiveness" is the result of a general increase in
sensitivity to environmental stimuli, and that the overall effect of isolation
can be regarded in terms of arousal theory and the general level of environ-
mental stimulation (Welch, 1964). If this reasoning is applied to the present
theory, a neuronal model of the environment encountered under conditions
of isolation would be a very restricted one, and therefore the subsequent
mismatch when a new set of stimulus conditions is encountered would be
proportionately greater. If this mismatch involves only one part of the
animal's environment, and is localized (i.e., that it is a strange mouse), the
discrepancy should be of sufficient magnitude to evoke attack in many
cases, rather than always evoking fear behavior: Krsiak and Borgesova
(1973) report that defense or flight do occur as common alternatives to at-
tack in isolated mice. In most studies of isolation-induced fighting,
however, a series of tests are given in the isolated animal's home cage, and
therefore, once attack has taken place, it is very likely that the isolate, after
winning its first encounter (as would be very likely in its home cage), would
be reinforced and hence would attack more vigorously on future occasions.
An alternative explanation for the occurrence of isolation-induced
fighting is also possible, in terms of chemical changes induced by isolation.
Isolation is known to be associated with increased gonadal function in male
mice (Brain and Nowell, 1969b, 1970, 1971), and also with higher levels of
brain noradrenalin and lowered adrenopituitary function (Welch and
Welch, 1965, 1966, 1969; A. S. Welch, 1965a,b); it is therefore possible that
isolation exerts its influence on aggressiveness through changes in the
neuroendocrinological mechanisms which facilitate fighting, e.g., by
increasing either androgen, or ACTH, or noradrenalin levels (see above).
274 John Archer
3. Reinjorcement oj Attack and Fear Behavior

Previous experience in aggressive encounters is of primary importance
in determining how an animal will respond to a situation capable of evoking
attack or fear behavior. In the section on the effects of sex hormones, it was
emphasized that previous experience can override such potential chemical
influences on aggressiveness. Various studies have demonstrated that the
winning of encounters increases the probability that an animal will respond
by attack to a subsequent situation that evokes attack or fear behavior
(pain-induced: Kimbrell and Chesler, 1971, for rats; Legrand and Fielder,
1973, for mice; intermale fighting: Ginsburg and Allee, 1942; Scott, 1946;
Lagerspetz, 1964; Bevan et al., 1960, for mice). Conversely, defeat increases
the probability that an animal will subsequently respond to such a situation
with either distress calls, escape avoidance, or immobility (Kimbrell and
Chesler, 1971, for rats in response to shock; Kahn, 1951, for mice), and
defeat also facilitates the learning of an avoidance task involving escape in
response to signaled shock (Hudgens and MacNeil, 1970).
Therefore, since success reinforces attack (e.g., Legrand, 1970;
Tellegren et al., 1969), it subsequently occurs more readily when the animal
next encounters a situation that is potentially capable of evoking either at-
tack or fear behavior. Conversely, if an animal experiences defeat, it will be
more likely to react to this type of situation with fear behavior.
E. Interaction of Variables Affecting First Stage of Decision Process
All the factors described above, and included in the diagram of the
model (Fig. I) as influencing decision, process 1, are regarded as influencing
the commands (a or b in Fig. 1) which reflect the probability of the animal
attacking or showing fear behavior, for a given set of environmental condi-
tions of the type which can affect decision process 2. In this sense, the
values of a and b might be regarded as equivalent to the animal's fear or
anger state.
Factors affecting decision process 1 clearly do not occur in isolation
from one another, and therefore we should consider briefly how they might
interact with one another. In general, the answer to this question is un-
certain at present, but there is some evidence in the literature on which to
base speculation. For example, it has been shown how previous reinforce-
ment of attack or fear behavior can override the chemical influence of
androgens (e.g., Guhl, 1964, for "social inertia"; Bevan et al., 1960, for an
experimental study of previous reinforcement), so that the effects of pre-
vious reinforcement must be given larger weightings than androgens and
perhaps then other chemical factors (Kuo, 1960d, for thiamin). The factor
which does, however, emerge as very important-even more so than previous
reinforcement-is the degree of mismatch discussed earlier; thus, if the sur-
roundings are sufficiently novel, even animals with a long history of positive
reinforcement for attacking may show fear behavior instead of attack (e.g.,
studies by Kuo, 1960d; Hoffman and Boskoff, 1972, described above).
The weightings and relationships of the other variables affecting the de-
cision mechanism must await further experimental evidence.
VIII. EXTERNAL FACTORS AFFECTING DECISION PROCESS 2
It is suggested that once the decision from process 1 has occurred (Fig.
1), the final behavioral outcome is then determined by a number of other
variables, most of which are related to external factors. This separation of
the overall decision process into two parts is suggested in order to account
for observations such as the following: that in the presence of a suitably lo-
calized stimulus, attack may occur even though fear behavior might have
occurred in the absence of sllch a stimulus; that the physical prevention of
escape may evoke attack, even though the animal is showing more general
indications that it is fearful; that "fear behavior" may take one of two main
forms, immobility or escape, depending on factors such as stimulus
intensity and localization. These observations, and their implications for the
present model, are now discussed in more detail.
A. Attack and Target Properties
It is suggested that the initial decision from stage 1 can be expressed as

a predisposition toward one of two types of behavior, attack or fear (and we
have discussed in previous sections variables which affect the predominance
of either). If attack is the predominating tendency, it is further suggested
that information from a subsystem, concerned with analyzing the extent to
which the stimulus discrepancy provides a suitable localized stimulus to at-
tack, then influences the part of the second decision process concerned with
attack: if this analysis indicates that there is such a suitable localized stim-
ulus, the motor command for attack occurs. Thus if the discrepancy in-
volves another animal, attack is likely. This would occur, for example, when
a strange conspecific is placed in an animal's familiar area, when thwarting
involves another conspecific (e.g., in food competition tests: Kanak and
Davenport, 1967), and when pain is inflicted by another conspecific (e.g., in
276 John Archer
mice: Lloyd and Christian, 1967). Similarly, inanimate objects of a suitable

size are attacked when they are the source of the discrepancy, e.g., a human
hand placed in an animal's familiar area (rats: Galef, 1970d; chicks: Hor-
ridge, 1970; great tits: Blurton-Jones, 1968), and a bottle-brush similarly
placed (Lagerspetz and Mettala, 1965).
If analysis of the stimulus discrepancy reveals that this does not involve
a suitable localized stimulus for attack, it is suggested that other aspects of
the input are then analyzed in a similar manner, so that attacks can be
directed to nearby stimuli that have characteristics suitable for attack.
Thus, if another animal is in spatial proximity when a discrepancy occurs, it
may become the object of attack. For example, rats attack other rats which
are nearby when foot-shock occurs (e.g., Wolfe et al., 1971); when mice or
polecats familiar to one another are placed into a novel environment, attack
occurs (e.g., Poole, 1973), but a single animal does not attack another part
of the novel environment; when another mouse is present during the thwart-
ing of a learned response, attack also occurs (Lagerspetz, 1964), although it
is not directed to any other part of the environment in the absence of the
other mouse. Similarly, inanimate objects, not themselves the source of dis-
crepancy, can also be attacked; for example, in squirrel monkeys, a rubber
hose (Ulrich, 1966), and in rats, a hardboard ball (Galef, 1970a,b), a small
metal, wood, or rubber object (Azrin et al., 1968), or a celluloid doll
(Miller, 1948b).
It is likely that the factor of the potential target's proximity to the
potential attacker is the important characteristic for attacks which are "dis-
placed" from the original source of discrepancy to a nearby object or
animal. Ulrich (1966) reported that the further away rats were situated
from each other in a pain-induced fighting situation, the less the likelihood
of their attacking one another. "Displacement" of aggression has been
studied in relation to human aggression resulting from frustration: if stimuli
directly associated with the frustrating situation are not suitable for attack,
others in the surrounding environment with characteristics in common with
the frustrating object tend to evoke more aggressive responses than features
of the environment with nothing in common with the object of frustration
(Berkowitz, 1969). We might suggest, therefore, that the spatial proximity
and the degree of similarity to the source of the discrepancy both influence
whether an object becomes the target of displaced aggression.
So far, we have referred vaguely to "properties suitable for attack" in a
localized stimulus. It is now suggested that size and movement are generally
the most important target characteristics in determining whether attack oc-
curs. Few studies of animal aggression have investigated target size.
Edwards (1969) found that when mice were provided with a small (prepu-
bertal) opponent in their home cage, a high frequency of attack occurred,
even among females and castrated males (which do not usually attack full-
sized intruders). These findings suggest that the typically lower frequencies
of aggression found in female and castrated mice, as opposed to intact adult
males, are partly due to an inhibition of attack by larger-sized targets. Simi-
larly, White et al. (1969) found that female mice would attack much smaller
prepubertal female opponents, but would not attack larger mature males.
The size of target obviously has to be considered in relation to the size
of the attacker. Many animals will direct attacks toward small parts of a
larger object or animal; e.g., in rats and chicks, the human hand will be at-
tacked, whereas mice and great tits will attack the finger, and small dogs
and cats may attack the human leg. Inanimate targets used in experimental
studies are generally about the size of another conspecific or smaller (e.g., a
tennis ball for squirrel monkeys, a bottle-brush for mice, and a small green
wooden triangle for chicks). Where larger inanimate objects have been used,
fear behavior occurs instead of, or more frequently than, attack (e.g., a 5
inch X 8 inch cardboard triangle for chicks, by Schaller and Emlen, 1962; a
10 inch x 5 inch x 7 inch box for chicks, by Bateson, 1964; an 8 inch x 8
inch card for rats, Blanchard et aI., 1975). It is suggested, therefore, that
as the target becomes larger, the probability of fear behavior replacing at-
tacking increases.
The second important stimulus property is movement. Some animals
will attack motionless animals or inanimate objects (e.g., the side-blotched
lizard: Ferguson, 1966; Gymnotus: Black-Cleworth, 1970), but generally a
moving target is more effective (Berkowitz, 1969). Lagerspetz and Mettala
(1965), for example, found that a moving bottle-brush was attacked more
readily than a static one, and Lagerspetz (1964) reported, as a result of her
experiments, that movement appeared to be an important characteristic for
eliciting attack in mice. Rapid movement was also important in evoking
territorial attack in the side-blotched lizard (Ferguson, 1966). Alberts and
Galef (1973) observed that the cessation of movement by an opponent in-
hibited attack in male wild rats; however, an anesthetized strange male rat
still elicited attack by a resident male, although it was not attacked in
preference to a moving intruder.
More specific characteristics may determine whether or not a living
conspecific is attacked. Studies of schedule-induced attack in pigeons (e.g.,
Azrin et al., 1966) show that only a low incidence of attack occured when a
stuffed animal replaced a living one. When one member of a pair of rats
was insulated from footshock, the frequency of aggression was far lower than
that generally found (Knutson, 1971). Knutson and Hynan (1972; cited in
Knutson, 1973) also found that the behavior of a shocked target rat was im-
portant in determining whether or not another shocked rat would attack it.
If there is no object suitable for attack when a discrepancy between ob-
278 John Archer
served and expected stimulation occurs, it is likely that some form of fear
behavior will occur. Thus, in the absence of a conspecific, fear behavior is
evoked by shock in rats (e.g., Galef, 1970a), whereas attack is common
when a conspecific is present (Ulrich, 1966). Escape behavior may occur in
response to extinction of a rewarded response (Adelman and Maatsch,
1956), but when a target is present in such a situation, attack occurs (e.g.,
Thompson and Bloom, 1966).
B. Two Types of Fear Behavior
The general term "fear behavior" includes two contrasting types of

responses, "freezing" and fleeing. Freezing presumably evolved as an anti-
predator device, since many animals perceive only moving prey (von
Uexkull, 1934).
Recent studies have separated some of the factors that determine which
of the two types of fear behavior will occur in a given situation. Fentress
(1968), for example, found, in voles, that when locomotion preceded
overhead presentation of a moving object, fleeing was more likely to occur
than freezing, which occurred after prior immobility. It therefore seems that
in this case the immediately preceding behavioral state predisposes the
animal to show one or the other type of fear behavior. Another example of
the animal's internal state biasing it towards showing immobility or active
escape is the observation that young male chicks treated with testosterone
show more frequent "freezing" and less frequent active escape in response
to a novel and startling situation than do untreated chicks (Archer,
unpublished).
Other recent studies have investigated external factors. In rats, Blan-
chard and Blanchard (1968, 1970) found that a more localized shock source
evoked active avoidance, whereas a more diffuse one produced immobility.
Ratner (1967) has suggested that animals will freeze in response to a preda-
tor if it is some distance away, but will show active escape if it comes
nearer. It is also possible that localization of the stimulus is important in
this case. Barcik and Collins (1972) investigated the effect of intensity of
stimulation. They found, over a range of shock intensities from 0.05 to 4.0
mA, that low and medium intensities produced immobility, whereas higher
intensities produced flight.
As a tentative conclusion from these few studies, it can be suggested
that low-intensity diffuse stimulation is more likely to produce immobility,
whereas high-intensity localized stimulation is more likely to evoke active
flight. Clearly, this should be regarded only as a working hypothesis for fu-
ture studies, since the subject has not been adequately investigated so far.
The temporal sequence of fear responses also deserves investigation, since in

many cases both immobility and active escape occur following a fear-evok-
ing stimulus. For example, in the author's own studies of interrupting be-
havior in an open field by a loud bell, it was found that the typical response
while the bell was sounding was active escape, but that this was replaced by
"freezing" when the bell had stopped ringing (chicks: Archer, 1973b,
unpublished data; rats: Archer, 1974c; mice: unpublished data). Further,
more detailed studies of the sequence of behavior changes following the
interruption of ongoing behavior by different types of startling stimuli are
required, in order to investigate not only which type of fear behavior is
more likely to occur in response to a particular stimulus, but also the
precise temporal relationship between the two types of behavior when they
both occur in response to one particular stimulus (see also Fentress, 1968,
1973).
C. Blocked Escape Behavior
One final addition to the present model is required, in order to take

into account the frequently observed phenomenon that when escape is
physically blocked attack is likely to occur, even though the animal still
shows general indications of fear behavior (e.g., distress vocalizations, au-
tonomic responses). In this case, it is suggested that in the absence of a
space in Which to escape, the motor command for escape is changed to at-
tack, even though no general tendency to attack exists (see Fig. 1). It is
presumed that if attack is 5uccessful in removing the blocking stimulus, es-
cape can then occur unimpeded.
In connection with this point, Hediger (1950) showed that animals
respond by flight when a potential predator comes closer than a certain
distance (the "flight distance"); however, if escape is not possible, and the
potential predator ventures even closer (exceeding the "critical distance"),
attack is likely to occur.
D. Other Situational Factors
There is little evidence to suggest that the presence of food or the lack
of it can either cause or directly enhance aggressive behavior, except when
some form of "frustration" is involved (various species: review by Gottier,
1972; mice: Clark and Schein, 1966; Strecker and Emlen, 1953; rabbits:
Mykytowycz, 1961; buntings: Andrew, 1957; chaffinches: Marler, 1956a;
blue-tits: Blurton-Jones, 1968). Andrew (1957) did, however, find that
280 John Archer
aggressive encounters became more frequent after food had been removed in
bunting flocks, but this was attributable to the resulting increased activity
rather than to the effects of hunger per se. Encounters thus became more
frequent, and since aggressive responses were usually aroused by a fellow bird
approaching too close, the level of aggression was increased. A similar expla-
nation could account for reports of food deprivation increasing the tendency
to attack in other species (e.g., in mice: Fredericson, 1950).
There is some evidence that the presence of females may influence
male aggression in rodents; e.g., in wild rats, Barnett (1955) and Evans
(1964) found that fighting was facilitated in the presence of females, and a
similar result was found by Ginsburg and Allee (1942) and Bronson (per-
sonal communication, 1966) for Peromyscus. On the other hand, Levine et
al. (1965) found that aggressiveness was reduced in male mice of the CBA
strain when females were present. Poole (1973), however, reported that the
presence of females in a group of polecats exerted little detectable influence
on the attack of strange conspecifics.
IX. DISCUSSION
One aspect of the present model has still to be mentioned. In Fig. 1, the
overt behavior-either "attack," "immobility," or "escape"-is shown as
producing an outcome, or "environmental consequences of behavior,"
which then modifies the "input." In this way a feedback loop is completed,
and the occurrence of either attack or fear behavior can be regarded as dif-
ferent ways of modifying the input, so that the discrepancy between ob-
served and expected events is reduced to within tolerable limits. Eventually
the attack or the fear behavior ceases. This is essentially saying what has
been realized about animal fighting for many years. Thus Craig (1928, p.
267) wrote: "When an animal does fight he aims, not to destroy the enemy,
but only to get rid of his presence and his interference."
To account for observations such as the continuation of immobility or
escape after the cessation of the eliciting stimulus, we have to assume that
once the systems underlying these types of behavior are activated they
remain so, even though the input may have changed. Variable delays in the
feedback loop, depending on the strength of the original discrepancy, could
be introduced into the model to explain these findings.
The model depicted in Fig. 1 is therefore regarded as a dynamic one,
the values of the different parts changing with time; it is therefore possible
for the precise outcome to be fairly stable, or to vary between the different
possible responses, according to the circumstances at anyone time. The
neuronal model, or expectation copy, must also be regarded as dynamic,

but over longer periods of time, in that it incorporates newer aspects of the
animal's experience.
The model for aggression suggested in this paper is an extension of
results obtained in certain specific situations with rodents (e.g., by Ulrich,
Azrin, and their co-workers, and by Galef) to include a wider range of
situations in which aggression can occur. More generally, it is an extension
of previous ideas concerning the essential unity of situations producing at-
tack and fear behavior, as put foward by Hebb (1946), Hinde (1969, 1970),
and Berkowitz (1962). These and the other studies reviewed above have been
incorporated into a general basic plan showing the factors influencing ag-
gression and fear behavior. Further modifications of this basic plan can oc-
cur, both on an individual and an evolutionary basis. Learning can play an
important part in affecting attack and fear behavior, as indicated at the be-
ginning of the paper. In inframammalian vertebrates in particular, it is
envisaged that species-typical biasing factors can be introduced on the input
side, so that these more readily evoke attack or fear behavior, without
necessarily involving a high-stimulus discrepancy. The most parsimonious
way of incorporating this into the present model is to suggest that the ex-
pectation copy is resistant to incorporating certain forms of stimulation (as
suggested earlier for pain).
Moyer (1968) also proposed a theoretical scheme for aggression, but
his starting point was that the stimulus situations evoking aggressive be-
havior are diverse, and that therefore their physiological substrates must
also be diverse. In the present paper it was argued that even if these situa-
tions may appear diverse at first, common properties do exist. Two of
Moyer's classes of aggression, "predatory" and "instrumental," were not
considered in the present paper's main model. It was argued earlier that
predation is rather different from aggression, as defined at the beginning of
this chapter. Indeed, this difference is generally recognized, and
neurophysiological evidence reviewed by Moyer strongly suggests that
predation has neural bases different than those in the forms of aggression
considered here. "Instrumental aggression" was also considered separately
in the present scheme, together with classically-conditioned aggression, as
"conditioned attack."
There remain, according to Moyer's scheme, five further types of
aggression, which refer to the main forms considered as unconditioned
aggression in the present paper. Moyer's criteria for classifying these five
types were not consistent; e.g., "irritable aggression" was vaguely defined,
on the basis of inferred mood, referring to pain-induced attack and the three
forms of frustration-induced attack in the present scheme. Moyer's "fear-
induced aggression" was defined as aggression preceded by attempts to
282 John Archer
escape; this clearly refers to attack produced by blocking of escape (see

previous section and Fig. 1). Moyer's remaining three classes are called
intermale, territorial, and maternal aggression. In the present paper, these
were all regarded as variations on the basic situation of an animal's attack-
ing a stimulus with appropriate target properties which had intruded into
the animal's familiar area (used here in a wide sense to cover both indi-
vidual distance and territory).
Neurophysiologically, Moyer's (1968) paper did not contain any evi-
dence which can be regarded as inconsistent with the present model, but the
findings he reviewed were sparse in most cases; e.g., there was little definite
evidence on the neural substrates underlying "intermale aggression," "terri-
torial aggression," and "maternal aggression," and "irritable aggression"
was shown to be associated with parts of the amygdala different than those
associated with "fear-induced aggression." It should be noted that in the
present, behaviorally-based, model these two types of aggression also show
very different causal bases (Fig. 1).
The present model reflects both the diversity of "aggression" (em-
phasized in Moyer's neurophysiological approach), in that different
external, internal motivational, and hormonal factors can be involved, and
also its unity, in that mechanisms for attack and fear behavior are regarded
as being activated by a common type of variable (discrepancy).
Since the present theoretical scheme has been deduced from many
studies which were not originally designed to test its particular predictions,
there are many gaps in the evidence that is required to support, refute, or
amend various aspects of it. It remains to outline some of the predictions
derived from the present model which could in the future be tested experi-
mentally.
A more complete investigation is first required of those situations
which have hitherto been regarded as producing predominantly attack or
fear behavior, but not both. For example, it is predicted that the various
forms of frustration, including schedules such as FR, FI, and DRL, can in-
duce fear behavior, but there is little experimental evidence on this point at
present. There is some evidence for thwarting and extinction situations, but
this is very restricted. Thus, future studies directed at this general point
would be helpful. These would need to involve a more ethological or
observational approach to schedule-induced behavior than has hitherto been
attempted.
Conversely, more studies are required of situations which typically in-
duce fear behavior, to determine whether aggression is also a possible al-
ternative. There is little evidence on this point for novel environments; it
was suggested earlier that attack is rarely observed in novel environments
because of first, the high discrepancy involved, and second, the lack of a
suitable target. Thus a type of study is envisaged in which there is a lo-

calized "target" (possibly based on those used in existing studies of
schedule-induced attack), with the animal being observed during the course
of repeated exposures to the novel environment (so that novelty, and hence
discrepancy, is reduced sufficiently for attack to be the predicted outcome).
The next stage would be to investigate the possible interaction between
the different types of situations capable of inducing attack or fear behavior,
i.e., pain, novelty, and frustration, in order to provide more evidence on the
precise nature of their interaction; it was suggested earlier that the existing
evidence indicates an additive effect, but, since this is an important point,
confirmatory studies should be undertaken. The question of how different
causes of attack and fear behavior interact is closely related to the idea,
crucial to the present model, of a common type of mechanism for the
various forms of fear and attack. This common type of mechanism involves
the idea of a discrepancy between observed events and a model based on
past experience. It is therefore also crucial to attempt to manipulate the
degree of discrepancy; for example by systematically varying the animal's
housing and testing conditions. In particular, it is important to determine
the extent to which experimental evidence supports the theoretical rela-
tionships which were suggested earlier between the degree of discrepancy
and the tendency for fear behavior and attack to occur. (Figs. 3,4, and 5).
Such relationships were based on Berkowitz's (1962) suggestions from
human work, and also from animal experiments on pain-induced ag-
gression; It is important to determine whether such relationships hold for
"discrepancy" in general (i.e., can be extended from pain-induced ag-
gression studies to ones involving novelty and frustration).
Among other factors that were suggested as influencing the probability
of the occurrence of attack or fear behavior one factor was the strength of
ongoing motivation. This could be investigated using food or water rein-
forcement, by varying the extent of deprivation, and by varying the number
of continuous reinforcement trials that take place before extinction sessions
are introduced.
The precise way in which chemical factors, such as testosterone,
ACTH, and catecholamines, affect aggression is still unclear, and there are,
of course, a large number of studies being carried out on hormonal in-
fluences on aggression, in a variety of species. When more is known about
their site or sites of action, the present model will presumably have to be
amended to incorporate such findings; at present they have been represented
simply as increasing the probability of attack or decreasing the probability
of fear behavior, but this representation is more a convenience than a
realistic deduction based on strong evidence.
It was remarked in the sectIOn on developmental influences that this
284 John Archer
research area would provide a fruitful one for future investigations of

aggression, from a general and practical viewpoint. Such investigations
would also be important in the context of investigating the present theoretical
model, e.g., to test explanations of isolation-induced aggression, and to de-
termine effects of early experience on both attack and fear behavior.
Overall, it is more important in theoretical terms to investigate how the
various environmental factors-long term, such as developmental in-
fluences, and shorter term, such as previous reinforcement-interact with
internal factors, in order to obtain an assessment of the weightings and
interactions of these variables in influencing the tendency to attack or show
fear behavior, as represented here by decision process 1 (Fig. 1).
Finally, the suggestions made in reference to situational or external
variables (in relation to decision process 2), such as target properties,
provide some fairly precise hypotheses concerning the localization of the
target source, the importance of target size and other properties such as
movement in influencing whether attack or fear behavior occurs, and con-
cerning the precise point at which the probability of one declines and the
other increases. There is also a whole ar:ea of study concerning factors that
affect the different types of fear behavior-immobility and active escape-
and whether or not blocking escape with a suitable target for attack induces
attack in all cases, as suggested in Fig. 1 and in one of the previous sections.
The present model does, therefore, provide a number of fairly precise
predictions, as well as indicating in a more general way what might prove
fl uitful avenues for future research. If such research is forthcoming, the
author's main purpose in outlining this theoretical scheme will have been
fulfilled.
X. ACKNOWLEDGMENTS
This chapter was written while I was supported by a grant from the
Medical Research Council of the U.K. to Professor R. J. Andrew. I thank
Professor Andrew, Dr. P. P. G. Bateson, Dr. P. J. B. Slater, Dr. H. H.
Swanson, Dr. P. Brain, and Dr. D. K. Candland for their helpful comments
during preparation of the manuscript.
XI. REFERENCES
Adelman, H. M., and Maatsch, 1. L. (1956). Learning and extinction based upon frustration,
food reward, and exploratory tendency. J. Exp. Psychol. 52:311-315.
Ader, R. (1968). Effects of early experiences on emotional and physiological reactivity in the
rat. J. Compo Physiol. Psychol. 66:264-268.
Ader, R., Friedman, S. B., and Grota, L. J. (1967). "Emotionality," and adrenal cortical func-
tion: Effects of strain, test, and the 24 hr corticosterone rhythm. Anim. Behav. 15:37-44.
Adler, N., and Hogan, J. A. (1963). Classical conditioning and punishment of an instinctive
response in Betta splendens. Anim. Behav. II:351-354.
Alberts, J. R., and Galef, B. G. (1973). Olfactory cues and movement: Stimuli mediating in-
traspecific aggression in the wild Norway rat. J. Compo Physiol. Psychol. 85:233-242.
Andrew, R. J. (1956). Fear responses in Emberiza sp. Anim. Behav. 4:125-132.
Andrew, R. J. (1957). The influence of hunger on aggressive behavior in certain buntings of the
genus Emberiza. Physiol. Zool. 30:177-185.
Andrew, R. J. (1966). Precocious adult behavior in the young chick. Anim. Behav. 14:485-500.
Andrew, R. J. (1972a). Changes in search behaviour in male and female chicks following dif-
ferent doses of testosterone. Anim. Behav. 20:741-750.
Andrew, R. J. (1972b). Recognition processes and behavior with special reference to effects of
testosterone on persistence. In Lehrman, D. S., Hinde, R. A., and Shaw, E. (eds.), Ad-
vances in the Study oj Behavior, Vol. 4, Academic Press, New York and London.
Andrew, R. J. (1973). The evocation of calls by diencephalic stimulation in the conscious
chick. Brain, Behav. Evol. 7:424-446.
Andrew, R. J., and Rogers, L. (1972). Testosterone, search behaviour and persistence. Nature
(Lond.) 237:343-346.
Andrzejewski, R., Petrusewicz, K., and Walkowa, W. (1963). Absorption of newcomers by a
population of white mice. Ekol. Polska, Ser. A. II: 223-240.
Anisko, J. J., Christenson, T., and Buehler, M. G. (1973). Effects of androgen on fighting be-
havior in male and female Mongolian gerbils Meriones unguiculatus. Horm. Behav.
4:199-208.
Archer, J. (1968). The effect of strange male odor on aggressive behavior in male mice. J.
Mammal. 49:572-575.
Archer, J. (1969). Adrenocortical responses to olfactory social stimuli in male mice. J. Mam-
mal. 50:839-841.
Archer, J. (1970). Effects of population density on behaviour in rodents. In Crook, J. H. (ed.),
Social Behaviour in Birds and Mammals, Academic Press, London and New York.
Archer, J. (1971). Sex differences in emotional behaviour: A reply to Gray and Buffery. Acta
Psychol.35:415-429.
Archer, J. (1973a). The influence of testosterone on chick behavior in novel environments.
Behav. Bioi. 8:93-108.
Archer, J. (1973b). Effects of testosterone on immobility responses in the young male chick.
Behav. Bioi. 8:551-556.
Archer, J. (1973c). A further analysis of responses to a novel environment by testosterone-
treated chicks. Behav. Bio!. 9:389-396.
Archer, J. (l973d). Tests for emotionality in rats and mice: A review. Anim. Behav. 21:205-
235.
Archer, J. (1974a). The effects of testosterone on the distract ability of chicks by irrelevant and
relevant novel stimuli. Anim. Behav. 22:397-404.
Archer, J. (1974b). Testosterone and behaviour during extinction in chicks. Anim. Behav.
22:650-655.
Archer, J. (1974c). Sex differences in the emotional behavior of three strains of laboratory rat.
Anim. Learn. Behav. 2:43-48.
Arvola, A., limen, M., and Koponen, T. (1962). On the aggressive behaviour of the Norwegian
lemming Lemmus lemmus with special reference to the sounds produced. Ann. Zool. Soc.
Vanamo. 17:80-101.
286 John Archer
Azrin, N. H. (l961). Time-out from positive reinforcement. Science 133:382-383.

Azrin, N. H. (1970). Punishment of elicited aggression by a physical blow. J. Exp. Anal.
Behav. 14:7-10.
Azrin, N. H., and Hutchinson, R. R. (l967). Conditioning of the aggressive behavior of pi-
geons by a fixed-interval schedule of reinforcement. J. Exp. Anal. Behav. 10:395-402.
Azrin, N. H., Hutchinson, R. R., and Sallery, R. D. (1964a). Pain-aggression toward inani-
mate objects. J. Exp. Anal. Behav. 7:223-228.
Azrin, N. H., Ulrich, R. E., Hutchinson, R. R., and Norman, D. G. (1964b). Effect of shock
duration on shock-induced fighting.J. Exp. Anal. Behav. 7:9-11.
Azrin, N. H., Hutchinson, R. R., and Hake, D. F. (l965a). Extinction-induced aggression.
Am. Psychologist 20:583.
Azrin, N. H., Hutchinson, R. R., and McLaughlin, R. (1965b). The opportunity for aggression
as an operant reinforcer during aversive stimulation.J. Exp. Anal. Behav. 8:171-180.
Azrin, N. H., Hake, D. F., and Hutchinson, R. R. (l965c). Elicitation of aggression by a
physical blow,J. Exp. Anal. Behav. 8:55-57.
Azrin, N. H., Hutchinson, R. R., and Hake, D. F. (l966). Extinction-induced aggression. J.
Exp. Anal. Behav. 9:191-204.
Azrin, N. H., Rubin, H. B., and Hutchinson, R. R. (l968). Biting attack by rats in response to
aversive shock. J. Exp. Anal. Behav. 11:633-639.
Baenninger, L. P. (1973). Aggression between weanling Peromyscus and Microtus. Anim.
Behav. 21 :335-337.
Baenninger, R. (l968). Catecholamines and social relations in Siamese fighting fish. Anim.
Behav. 16:442-447.
Baenninger, R., and Grossman, 1. C. (l969). Some effects of punishment on pain-elicited ag-
gression.J. Exp. Anal. Behav. 12:1017-1022.
Baggerman, B. (1966). On the endocrine control of reproductive behaviour in the male three-
spined stickleback Gasterosteus aculeatus L. Symposia oj the Society jor Experimental
Biology 20:427-456.
Banks, E. M. (l962). A time and motion study of pre fighting behavior of mice. J. Gen.
Psycho!. 101:165-183.
Barcik, 1. D., and Collins, D. E. (1972). Shock-elicited defensive behavior and passive avoid-
ance performance. Psychon. Sci. 28:37-40.
Barfield, R. 1., and Sachs, B. D. (l970). Effect of shock on copulatory behavior in castrate
male rats. Horm. Behav. 1:247-253.
Barfield, R. 1., Busch, D. E., and Wallen, K. (l972). Gonadal influence on agonistic behavior
in the male domestic rat. Horm. Behav. 3:247-259.
Barnett, S. A. (1955). Competition among wild rats. Nature 175:126-127.
Barnett, S. A. (1958a). An analysis of social behaviour in wild rats. Proc. Zool. Soc. Lond.
130:107-152.
Barnett, S. A. (1958b). Experiments on "neophobia" in wild and laboratory rats. Brit. J.
Psychol.49:195-201.
Bateson, P. P. G. (l964). Changes in chicks' responses to novel moving objects over the sensi-
tive period for imprinting. Anim. Behav. 12:479-489.
Beeman, E. A. (I 947a}. The effect of male hormone on aggressive behavior in mice. Physiol.
Zool. 20:373-405.
Beeman, E. A. (1947b). The relation of the interval between castration and first encounter to
the aggressive behavior of mice. Anat. Rec. 99:570-571.
Berkowitz, L. (l962). Aggression, McGraw-Hill, New York.
Berkowitz, L. (1969). The frustration-aggression hypothesis revisited. In Berkowitz, L. (ed.),
Roots oj Aggression, Atherton, New York.
Berkowitz, L. (1970). The contagion of violence: An S-R mediational analysis of some effects
of observed aggression. In Arnold, W. J., and Page M. M. (eds.) Nebraska Symposium on
Motivation, University of Nebraska Press, Lincoln.
Berkowitz, L. (1973). The case for bottling up rage. New Society 25:761-764.
Bevan, W., Levy, G. W., Whitehouse, J. M., and Bevan, J. M. (1957). Spontaneous aggressive-
ness in two strains of mice castrated and treated with one of three androgens. Physiol.
Zool.30:341-349.
Bevan, W., Daves, W. F., and Levy, G. W. (1960). The relation of castration, androgen
therapy and pre-test fighting experience to competitive aggression in male C57Bl/10
mice. Anim. Behav. 8:6-12.
Bindra, D., and Spinner, N. (1958). Response to different degrees of novelty: the incidence of
various activities. J. Exp. Anal. Behav. 1:341-350.
Black-Cleworth, P. (1970). The role of electrical discharges in the non-reproductive social be-
haviour of Gymnotus carapo (Gymnotidae, Pisces). Anim. Behav. Monogr. 3:1-77.
Blanchard, R. J., and Blanchard, D. C. (1968). Escape and avoidance responses to a fear elicit-
ing situation. Psychonom. Sci. 13:19-20.
Blanchard, R. J., and Blanchard, D. C. (1970). Dual mechanisms in passive avoidance. Psy-
chonom. Sci. 19:1-4.
Blanchard, R. J., Kelley, M. J., and Blanchard, D. C. (1974a). Defensive reactions and ex-
ploratory behavior. J. Compo Physiol. Psycho I. 87: 1129-1 133.
Blanchard, R. J., Mast, M., and Blanchard, D. C. (1975). Stimulus control of defensive reac-
tions in the albino rat.J. Comp. Physiol. Psychol. 88:81-88.
Blizard, D. (1968). Autonomic and behavioural correlates of emotionality. Ph.D Thesis,
University of Wales, Cardiff.
Blurton-Jones, N. G. (1968). Observations and experiments on causation of threat displays of
the Great Tit Parus major. Anim. Behav. Monogr. 1:75-158.
Boice, R., and Pickering, A. (1973). Shock-elicited behaviors in paired gerbils. Bull. Psychon.
Soc. 2:2"'1-224.
Bolles, R. C. (1960). Grooming behavior in the rat. J. Compo Physiol. Psychol. 53:306-310.
Boska, S. c., Weisman, H. M., and Thor, D. H. (1966). A technique for inducing aggression in
rats utilizing morphine withdrawal. Psychol. Rec. 16:541-543.
Bracco, H. F., and Miller, L. L. (1972). Effects of varying shock intensities on pain-elicited ag-
gression in the domestic fowl. Percept. Mot. Skills 34:719-723.
Brain, P. F. (1972). Study on the effect of the 4-10 ACTH fraction on isolation-induced inter-
male fighting behaviour in the albino mouse. Neuroendocrinol. 10:371-376.
Brain, P. F., and Evans, C. M. (1973). Some recent studies on the effects of corticotrophin on
agonistic behaviour in the house mouse and the golden hamster. J. Endocrinol. 57:xxxix-
xL.
Brain, P. F., and Nowell, N. W. (1969a). Some behavioral and endocrine relationships in adult
male laboratory mice subjected to open field and aggression tests. Physiol. Behav. 4:945-
947.
Brain, P. F., and Nowell, N. W. (1969b). Some endocrine and behavioral changes in the
development of the albino laboratory mouse. Comm. Behav. Bioi. A 4:203-220.
Brain, P. F., and Nowell, N. W. (1970). The effects of differential grouping on endocrine func-
tion of mature male albino mice. Physiol. Behav. 5:907-910.
Brain, P. F., and Nowell, N. W. (1971). Isolation versus grouping effects on adrenal and
gonadal function in albino mice. I. The Male. Gen. Compo Endocrinol. 16:149-154.
Brain, P. F., Nowell, N. W., and Wouters, A. (1971). Some relationships between adrenal
function and the effectiveness of a period of isolation in inducing intermale aggression in
albino mice. Physiol. Behav. 6:27-29.
288 John Archer
Brannigan, C. R. (1972). Exploratory behaviour in the laboratory rat. M.Sc. Dissertation,

University of Birmingham, England.
Brawn, V. M. (1960). Aggressive behaviour in the cod Gadus cal/arias L. Behaviour 18:107-
147.
Broadhurst, P. L. (1969). Psychogenetics of emotionality in the rat. Ann. N.Y. A cad. Sci.
159:806-824.
Bronson, F. H., and Desjardins, C. (1968). Aggression in adult mice: Modification by neonatal
injections of gonadal hormones. Science 161:705-706.
Bronson, F. H., and Desjardins, C. (1970). Neonatal androgen administration and adult ag-
gressiveness in female mice. Gen. Compo Endocrinol. 15:320-325.
Broom, D. M. (1969). Behavior of undisturbed 1-10 day old chicks in different rearing condi-
tions. Develop. Psychobiol. 1:287-295.
Burge, K. G., and Edwards, D. A. (1971). The adrenal gland and the pre- and post cast rational
aggressive behaviour of male mice. Physiol. Behav. 7:885-888.
Buss, A. H. (1971). Aggression pays. In Singer, J. L. (ed.), The Control of Aggression and
Violence, Academic Press, New York and London.
Bygott, J. D. (1972). Cannibalism among wild chimpanzees. Nature 238:410-411.
Caggiula, A. R. (1972). Shock-elicited copulation and aggression in male rats. J. Compo
Physiol. Psychol. 80:393-397.
Candland, D. K., and Campbell, B. A. (1962). Development of fear in the rat as measured by
behavior in the open field. J. Compo Physiol. Psychol. 55:593-596.
Candland, D. K., and Leshner, A. I. (1974). A model of agonistic behavior: Endocrine and au-
tonomic correlates. In Di Cara, L. (ed.), Advances in Limbic and Autonomic Nervous
System Research, Plenum, New York.
Candland, D. K., Nagy, Z. M., and Conklyn, D. H. (1963). Emotional behavior in the
domestic chicken (white leghorn) as a function of age and developmental environment. J.
Compo Physiol. Psychol. 56:1069-1073.
Cherek, D. R., and Heistad, G. T. (1971). Fixed interval induced aggression. Psychon. Sci.
25:7-8.
Christenson, T., Wallen, K., Brown, B. A., and Glickman, S. E. (1973). Effects of castration,
blindness and anosmia on social reactivity in the male Mongolian gerbil Meriones ungui-
culatus. Physiol. Behav. 100:989-994.
Clark, L. H., and Schein, M. W. (1966). Activities associated with conflict behaviour in mice.
Anim. Behav. 14:44-49.
Clarke, J. R. (1956). The aggressive behaviour of the vole. Behaviour 9:1-23.
Cohen, P. S., and Looney, T. A. (1973). Schedule-induced mirror responding in the pigeon. J.
Exp. Anal. Behav. 19:395-408.
Cole, J. M., and Litchfield, P. M. (\969). Stimulus control of schedule-induced aggression in
the pigeon, Psychon. Sci., 17:152-153.
Cole, J. M., and Parker, B. K. (1971). Schedule-induced aggression: Access to an attack able
target bird as a: positive reinforcer. Psychon. Sci. 22:33-35.
Conder, P. J. (1949). Individual distance. Ibis. 91:649-655.
Conner, R. L., and Levine, S. (1969). Hormonal influences on aggressive behaviour. In
Garatini, S., and Sigg, E. B. (eds)., Aggressive Behaviour, Excerpta Medica Foundation,
Amsterdam.
Conner, R. L., Vernikos-Danellis, J., and Levine, S. (1971), Stress, Fighting and
neuroendocrine function, Nature 234:564-566.
Craig, W. (1928). Why do animals fight? Int. J. Ethics. 31:264-278.
Creer, T. L., Hitzing, E. W., and Schaeffer, R. W. (\966). Classical conditioning of reflexive
fighting. Psychon. Sci. 4:89-90.
Crook, J. H. (1968). The nature of territorial aggression. In Montagu, A. (ed.), Man and Ag-
gression, Oxford University Press.
Crook, J. H., and Butterfield, P. A. (1968). Effects of testosterone propionate and luteinizing
hormone on agonistic and nest building behaviour of Quelea quelea. Anim. Behav.
16:370-384.
Crook, J. H., and Butterfield, P. A. (1970). Gender role in the social system of Quelea. In
Crook, J. H. (ed.), Social Behaviour in Birds and Mammals, Academic Press, London
and New York.
Crosby, R. M., and Cahoon, D. D. (1973). Superstitious responding as an artifact in investiga-
tions of shock-elicited aggression. Psychol. Rec. 23:191-196.
Daly, H. B. (1969). Learning of a hurdle-jump response to escape cues paired with reduced
reward or frustrative non-reward. J. Exp. Psychol. 79:146-157.
Davis, D. E. (1957). Aggressive behavior in castrated starlings. Science 126:253.
Davis, W. M., and Khalsa, J. H. (1971). Some determinants of aggressive behaviour induced
by morphine withdrawal. Psychon. Sci. 24:13-15.
Denenberg, V. H. (1969). Open field behavior in the rat: What does it mean? Ann. N. Y. A cad.
Sci. 159:852-859.
Denenberg, V. H. (1973). Developmental factors in aggression. In Knutson, J. F. (ed.), The
Control of Aggression, Aldine, Chicago.
Dennis St. John, R., and Corning, P. A. (1973). Maternal aggression in mice. Behav. Bioi.
9:635-639.
DeWied, D., Witter, A., and Lande, S. (1970). Anterior pituitary peptides and avoidance ac-
quisition of hypophysectomized rats. In DeWied, D., and Weijnen, J. A.W. M. (eds.),
Pituitary, Adrenal and the Brain, Elsevier, Amsterdam.
Doering, C. H., Brodie, H. K. H., Kraemer, H., Becker, H., and Hemburg, D. A. (1974). Plasma
testosterone levels and psychologic measures in men over a two-month period. In Friedman,
R. C., Richert, R. M., and Vandewiele, R. L. (eds.), Sex Differences in Behavior, Wiley,
New York.
Dollard, J., Doob, L. W., Miller, N. E., Mowrer, O. H., and Sears, R. R. (1939). Frustration
and Aggression, Yale University Press, New Haven.
Duncan, I. J. H., and Wood-Gush, D. G. M. (1971). Frustration and aggression in the
domestic fowl. Anim. Behav. 19:500-504.
Duncan, I. J. H., and Wood-Gush, D. G. M. (1972). Thwarting of feeding behaviour in the
domestic fowl. Anim. Behav. 20:444-451.
Edwards, D. A. (1968). Mice: Fighting by neonatally androgenized females. Science 161:1027-
1028.
Edwards, D. A. (1969). Early androgen stimulation and aggressive behavior in male and female
mice. Physiol. Behav. 40:333-338.
Edwards, D. A. (1970). Post-neonatal androgenization and adult aggressive behaviour in fe-
male mice. Physiol. Behav. 5:465-467.
Edwards, D. A., and Herndon, J. (1970). Neonatal estrogen stimulation and aggressive be-
havior in female mice. Physiol. Behav. 5:993-995.
Eibl-Eibesfeldt, I. (1961). The fighting behavior of animals. Scient. Am. 205: 112-122.
Eichelman, B. S., Jr., Nguyen, B. T., and Ng, K. Y. (1972). Facilitated aggression in the rat
following 6-hydroxydopamine administration Physiol. Behav. 8:1-3.
Erskine, M. S., and Levine, S. (1973). Suppression of pituitary-adrenal activity and shock-in-
duced fighting in rats. Physiol. Behav. 11:787-790.
Evans, C. S. (1964). Aspects of social stress in wild rats. Ph.D. Dissertation, University of
Glasgow, Scotland.
Evans, L. T. (1936). Behavior of castrated lizards.J. Genet. Psychol. 48:217-221.
290 John Archer
Evans, R. M. (1967). Early aggressive responses in domestic chicks. Anim. Behav. 16:24-28.
Falk, J. L. (1971). The nature and determinants of adjunctive behavior. Physiol. Behav. 6:577-
588.
Falls, J. B. (1969). Functions of territorial song in the white-throated sparrow, In Hinde, R. A.
(ed.), Bird Vocalizations, Cambridge University Press, Cambridge.
Falls, J. B., and Brooks, R. J. (1965). Studies of intraspecific recognition by song in birds. Am.
Zool. 5:225 (abs.).
Fentress, J. C. (1968). Interrupted ongoing behaviour in two species of vole Microtus agrestis
and Clethrionomys britannic us. 1. Response as a function of preceding activity and the
context of an apparently "irrelevant" motor pattern. Anim. Behav. 16:135-153.
Fentress, J. C. (1973). Specific and nonspecific factors in the causation of behavior. In
Bateson, P. P. G., and Klopfer, P. (eds.), Perspectives in Ethology, Vol. I, Plenum, New
York.
Ferguson, G. W. (1966). Releasers of courtship and territorial behaviour in the side-blotched
Lizard, Uta Stansburiana. Anim. Behav. 14:89-92.
Flory, R. (l969a). Attack behavior in a multiple fixed-ratio schedule of reinforcement.
Psychon. Sci. 16:156-157.
Flory, R. (I969b). Attack behavior as a function of minimum inter-food interval. J. Exp. Anal.
Behav. 12:825-828.
Flory, R. K., and Ellis, B. B. (1973). Schedule-induced aggression against a slide-image target.
Bull. Psychon. Soc. 2:287-290.
Frank, F. (1957). The causality of microtine cycles in Germany. J. Wildl. Mgmt. 21:113-121.
Fraser, D. (1975). Vocalizations and other behaviour of growing pigs in an "open field" test.
Appl. Anim. Ethol. In press.
Fredericson, E. (1950). The effects of food deprivation upon competitive and spontaneous
combat in C57 black mice. J. Psychol. 29:89-100.
Fredericson, E. (1952). Aggressiveness in female mice. J. Comp. Physiol. Psychol. 45:254-257.
Friedman, S. B., Ader, R., Grota, L. J., and Larson, T. (1967). Plasma corticosterone response
to parameters of electric shock stimulation in the rat. Psychosom. Med. 29:323-328.
Funkenstein, D. H. (1955). The physiology of fear and anger. Scient. Am. 192:74-80.
Galef, B. G. (1970a). Target novelty elicits and directs shock-associated aggression in wild
rats. J. Comp. Physiol. Psychol. 71:87-91.
Galef, B. G. (1970b). Familiarity of target location as a factor in the shock-associated ag-
gression of wild rats. Psychon. Sci. 19:299-300.
Galef, B. G. (1970c). Stimulus novelty as a factor in 'the intraspecific pain-associated ag-
gression of domesticated rats. Psychon. Sci. 18:21.
Galef, B. G. (1970d). Aggression and timidity: Responses to novelty in feral Norway rats. J.
Comp. Physiol. Psychol. 70:370-381.
Gallup, G. G. (1965). Aggression in rats as a function of frustrative nonreward in a straight
alleyway. Psychon. Sci. 3:99-100.
Gandelman, R. (1972). Mice: Postpartum aggression elicited by the presence of an intruder.
Horm. Behav. 3:23-28.
Gandolfi, G., Mainardi, D., and Rossi, A. C. (1968). The fright reaction of zebra fish. AUi
Soc, Ital. Sci. Nat. 107:74-88.
Gentry, W. D. (1968). Fixed-ratio schedule-induced aggression. J. Exp. Anal. Behav. 11:813-
817.
Gentry, W. D., and Schaeffer, R. W. (1969). The effect of FR response requirement on ag-
gressive behavior in rats. Psychon, Sci., 14:236-238.
Ginsburg, B., and Allee, W. C. (1942). Some effects of conditioning on social dominance and
subordination in inbred strains of mice. Physiol. Zool. 15:485-506.
Goodall, J. van Lawick (1968). The behaviour of free-living Chimpanzees on the Gombe
Stream Reserve. Anim. Behav. Monogr. 1:161-311.
Gottier, R. F. (1972). Factors affecting agonistic behavior in several subhuman species. Genet.
Psychol. Monogr. 86:177-218.
Grant, E. C. (1963). An analysis of social behaviour of the male laboratory rat. Behaviour
21:260-281.
Gray, J. A. (1971). Sex differences in emotional behaviour in mammals including man:
endocrine bases. Acta Psychol. 135:29-46.
Gray, J. A. (1972). The Psychology of Fear and Stress, Weidenfeld & Nicholson, London.
Greidanus, T. B. W., and de Wied, D. (1971). Effects of systemic and intracerebral administra-
tion of two opposite acting ACTH-related peptides on extinction of conditioned avoidance
behavior. Neuroendocrinol. 10:291-301.
Grossen, N. E., and Kelley, M. J. (1972). Species-specific behavior and acquisition of avoid-
ance behavior in rats.J. Compo Physiol. Psychol. 81:307-310.
Guhl, A. M. (1964). Psychophysiological interrelations in the social behavior of chickens.
Psychol. Bull. 61:277-285.
Guiton, P., and Wood-Gush, D. G. M. (1967). Studies on thwarting in the domestic fowl. Rev.
Compo Anim. 5:1-23.
Hall, C. S. (1934). Emotional behavior in the rat. 1. Defecation and urination as measures of
individual differences in emotionality. J. Compo Psychol. 18:385-403.
Hall, C. S. (1938). The inheritance of emotionality. Sigma Xi Quart. 26:17-27.
Harding, C. F., and Leshner, A. E. (1972). The effects of adrenalectomy on the aggressiveness
of differently housed mice. Physiol. Behav. 8:437-440.
Hatch, A., Balazs, T., Wiberg, C. S., and Grice, H. C. (1963). Long term isolation in rats.
Science 142:507.
Hayes, J. G., Daley, M. F., and Cheney, C. D. (1969). Effects of induced aggression on a low
fixed-ratio schedule of food reinforcement. Psychon. Sci. 17:259-260.
Healey, M. C. (1966). Aggression and self-regulation of population size in deermice. M.Sc.
Dissertation, University of British Columbia, Canada.
Hebb, D. O. (1946). On the nature of fear. Psychol. Rev. 53:259-276.
Hediger, H. (1950). Wild Animals in Captivity, Butterworth, London.
Hinde, R. A. (1952). The behaviour of the great tit parus major and some other related species.
Behaviour Suppl. 2.
Hinde, R. A. (1953). The conflict between drives in the courtship and copulation of the chaf-
finch. Behaviour 5:1-31.
Hinde, R. A. (1967). The nature of aggression. New Society 231 :302-304.
Hinde, R. A. (1969). The bases of aggression in animals.J. Psychosom. Res. 13:213-219.
Hinde, R. A. (1970). Animal Behaviour: A Synthesis of Ethology and Comparative
Psychology, McGraw-Hill, New York.
Hoar, W. S. (1962). Hormones and the reproductive behavior of the male three-spined stickle-
back Gasterosteus aculealus. Anim. Behav. 10:247-266.
Hoffman, H. S., and Boskoff, K. J. (1972). Control of aggressive behavior by an imprinted
stimulus. Psychon. Sci. 29:305-306.
Hogan, J. A. (1967). Fighting and reinforcement in the Siamese fighting fish Bella splendens.
J. Compo Physiol. Psychol. 64:356-359.
Horridge, P. A. S. (1970). The development of copulatory and fighting behaviour in the
domestic chick. D. Phil. Dissertation, University of Sussex, England.
Hudgens, G. A., and MacNeil, D. A. (1970). Aggressiveness and learning ability: Effect of his-
tories of wins or defeats on avoidance learning in mice. Psychon. Sci. 20:51-53.
Hutchings, D. E., and Gibbon, J. (1970). Preliminary study of behavioral and teratogenic ef-
292 John Archer
fects of two "stress" procedures administered during different periods of gestation in the
rat. Psychol. Rep. 26:239-246.
Hutchinson, R. R., Ulrich, R. E., and Azrin, A. H. (1965). Effects of age and related factors
on the pain-aggression reaction. J. Compo Physiol. Psychol. 59:365-369.
Hutchinson, R. R., Azrin, N. H., and Hunt, G. M. (1968). Attack produced by intermittent
reinforcement of a concurrent response. J. Exp. Anal. Behav. 11:489-495.
Hutchinson, R. R., Renfrew, J. W., and Young, G. A. (1971). Effects of long-term shock and
associated stimuli on aggressive and manual responses. J. Exp. Anal. Behav. 15:141-166.
Hutzell, R. R., and Knutson, J. F. (1972). A comparison of shock-elicited fighting and shock-
elicited biting in rats. Physiol. Behav. 8:477-480.
Jainudeen, M. R., Katongole, C. B., and Short, R. V. (1972). Plasma testosterone levels in
relation to must and sexual activity in the male Asiatic elephant, Elphas maxim us. J.
Reprod. Fert. 29:99-103.
Johnson, R. N. (1972). Aggression in Man and Animals, Saunders, Philadelphia.
Johnson, R. N., and Johnson, L. D. (1973). Intra- and interspecific social and aggressive
behaviour in the Siamese fighting fish, Betta splendens. Anim. Behav. 21 :665-672.
Jones, R. B., and Nowell, N. W. (1973). The effect of familiar visual and olfactory cues on the
aggressive behavior of mice. Physiol. Behav. 10:221-223.
Kahn, M. H. (1951). The effects of severe defeat at various age levels on the aggressive be-
haviour of mice. J. Genet. Psychol. 79: 117-130.
Kanak, N. J., and Davenport, D. G. (1967). Between subject competition: a rat race. Psychon.
Sci. 7:87-88.
Karli, P. (1956). The Norway rat's killing response to the white mouse: an experimental
analysis. Behaviour 10:81-103.
Kimbrell, G. M .. and Chesler, D. (1971). Role of dominance status in shock-stimulated
agonistic behavior. Psychol. Rep. 29:768-770.
King, J. A. (1957). Relationships between early social experience and adult aggressive behavior
in inbred mice.J. Genet. Psychol. 90:151-166.
King, J. A., and Gurney, N. L. (1954). The effect of early social experience on adult aggressive
behavior in C57BI/1O mice. J. Compo Physiol. Psychol. 47:326-330.
Kiryukhin, V. G. (1972). The effect of central cholinoreceptor blocking on the aggressiveness
of male rats with varying levels of sex hormones. Farmakol. Toksikol. 35:541-544.
Knutson, J. F. (1970). Aggression during the fixed-ratio and extinction components of a
multiple schedule of reinforcement. J. Exp. Anal. Behav. 13:221-231.
Knutson, J. F. (1971). The effects of shocking one member of a pair of rats. Psychon. Sci.
22:265-266.
Knutson, J. F. (1973). Aggression as manipulatable behavior. In Knutson, J. F. (ed.), The Con-
trol of Aggression, Aldine, Chicago.
Knutson, J. F., and Hynan, M. T. (1972). Influence of upright posture on shock-elicited ag-
gression in rats. J. Compo Physiol. Psychol. 81 :297-306.
Knutson, J. F., and Kleinknecht, R. A. (1970). Attack during differential reinforcement of a
low rate of responding. Psychon. Sci. 19:289-290.
Kostowski, W., Rewerski, W., and Piechocki, T. (1970). Effects of some steroids on aggressive
behaviour in mice and rats. Neuroendocrinol. 6:311-318.
Krsiak, M., and Borgesova, M. (1973). Aggression and timidity induced in mice by isolation.
Act. Nerv. Super. 15:21-22.
Kuo, Z. Y. (1960a) ..Studies on the basic factors in animal fighting. I. General analysis of fight-
ing behavior. J. Genet. Psychol. 96:201-206.
Kuo, Z. Y. (1960b). Studies on the basic factors in animal fighting. II. Nutritional factors af-
fecting fighting behavior in quails. J. Genet. Psychol. 96:207-216.
Kuo, Z. Y. (1960c). Studies on the basic factors in animal fighting. III. Hormonal factors af-
fecting fighting in quails. J. Genet. Psycho!. 96:217-223.
Kuo, Z. Y. (1960d). Studies on the basic factors in animal fighting. IV. Developmental and en-
vironmental factors affecting fighting in quails. J. Genet. Psychol. 96:225-239.
Lack, D. (1939). The behaviour of the robin. Parts I and II. Proc. Zool. Soc. Lond. A
109:169-219.
Lagerspetz, K. (1964). Studies on the aggressive behaviour of mice. Ann. Acad. Sci. Fenn. Ser.
B 131 :(3).
Lagerspetz, K., and Mettala, R. (1965). Stimulation experiments on stimuli eliciting aggressive
behaviour in mice. Rep. lnst. Psychol., University of Turku, No. 13.
Lagerspetz, K., and Nurmi, R. (1964). An experiment on the frustration-aggression
hypothesis. Rep. lnst. Psychol., University of Turku No. 10.
Lagerspetz, K., and Talo, S. (1967). Maturation of aggressive behaviour in young mice. Rep.
lnst. Psychol., University of Turku, No. 28.
Lagerspetz, K. Y. H., Firri, R., and Lagerspetz, K. M. J. (1967). Neurochemical and endocri-
nological studies of mice selectively bred for aggressiveness. Rep. lnst. Psychol., University
of Turku, No. 29.
Lazarus, J., and Crook, J. H. (1973). The effects of luteinizing hormone, oestrogen and
ovariectomy on the agonistic behaviour of female Quelea quelea. Anim. Behav. 21:49-60.
Legrand. R. (1970). Successful aggression as the reinforcer for runway behavior of mice.
Psychon. Sci. 20:303-305.
Legrand, R., and Fielder, R. (1973). Role of dominance submission relationships in shock-in-
duced fighting of mice. J. Compo Physiol. Psychol. 82:501-506.
Leshner, A. I., Walker, W. A., Johnson, A. E., Kelling, J. S., Kreisler, S. J., and Svare, B. B.
(1973). Pituitary adrenocortical activity and intermale aggressiveness in isolated mice.
Physiol. Behav. II :705-711.
Levine, L., Diakow, C. A., and Barsel, G. E. (1965). Interstrain fighting in male mice. Anim.
Behav. !3:52-58.
Levine, S. (1959). Emotionality and aggressive behavior in the mouse as a function of infantile
experience. J. Genet. Psychol. 94:77-83.
Levine, S., and Jones, L. E. (1965). Adrenocorticotrophic hormone (ACTH) and passive
avoidance learning. J. Compo Physiol. Psychol. 59:357-360.
Levy, J. V., and King, J. A. (1953). The effects of testosterone propionate on fighting behavior
in young male C57BI/IO mice. Anat. Rec. II7:562-563.
Lieberman, M. W. (1963). Early developmental stress and later behavior. Science 141:824-
825.
Lincoln, G. A., Guinness, F., and Short, R. V. (1972). The way in which testosterone controls
the social and sexual behavior of the red deer stag Cervus elaphus. Horm. Behav. 3:375-
396.
Livesey, P. L., and Egger, G. J. (1970). Age as a factor in open-field responsiveness in the
white rat. J. Compo Physiol. Psychol. 73:93-99.
Lloyd, J. A., and Christian, J. J. (1967). Relationship of activity and aggression to density in
two confined populations of house mice Mus musculus. J. Mammal. 48:262-269.
Lorenz, K. (1966). On Aggression, Harcourt, Brace & World, New York.
Lumia, A. R., Rieder, C. A., and Reynierse, J. H. (1973). The differential effects of reinforce-
ment and testosterone on aggressive responding in pigeons: species typical and aversive
aspects of pigeon aggression. Bull. Psychon. Soc. 1:165-166.
Lyon, D.O., and Ozolins, D. (1970). Pavlovian conditioning of shock-elicited aggression: A
discrimination procedure. J. Exp. Anal. Behav. 13:325-331.
294 John Archer
Mackintosh, J. H., and Grant, E. C. (1966). The effect of olfactory stimuli on the agonistic be-
haviour of laboratory mice. Z. Tierpsychol. 23:584-587.
Marler, P. (1956a). Studies of fighting in chaffinches. (3) Proximity as a cause of aggression.
Anim. Behav. 4:23-30.
Marler, P. (1956b). Territory and individual distance in the chaffinch Fringil/a coelebs. Ibis
98:496-501.
Marrone, R. L., Pray, S. L., and Bridges, C. C. (1966). Norepinephrine elicitation of ag-
gressive display response in Betta splendens. Psychon. Sci. 5:207-208.
Marshall, I., and Weistock, M. (1971). Quantitative method for assessing one symptom of the
withdrawal syndrome in mice after chronic morphine administration. Nature 234:223-
224.
Matthewson, S. (1961). Gonadotrophic hormones affect aggressive behavior in starlings.
Science 134:1522-1523.
McFarland, D. J. (1966). The role of attention in the disinhibition of displacement activity.
Quart. J. Exp. Psychol. 18:19-30.
Melvin, K. B., and Anson, J. E. (1969). Facilitative effects of punishment on aggressive be-
havior in the Siamese fighting fish. Psychon. Sci. 14:89-90.
Melvin, K. B., and Ervey, D. H. (1973). Facilitative and suppressive effects of punishment on
species-typical aggressive display in Betta splendens. J. Compo Physiol. Psychol. 83:451-
457.
Miller, N. E. (1948a). Studies of fear as an acquirable drive. I. Fear as motivation and fear-re-
duction as reinforcement in the learning of new responses. J. Exp. Psychol. 38:89-101.
Miller, N. E. (l948b). Theory and experiment relating psychoanalytic displacement to stim-
Ulus-response generalization. J. Abnor. Soc. Psycho!. 43: 155-178.
Milligan, W. L., Powell, D. A., and Borasio, G. (1973). Sexual v~riables and shock-elicited ag-
gression. J. Compo Physiol. Psychol. 83:441-450.
Moyer, K. E. (1968). Kinds of aggression and their physiological basis. Comm. Behav. Bioi. A
2:65-87.
Moyer, K. E. (1972). Territorial aggression: Is this concept useful? Totus Homo. 4: 104-117.
Moyer, K. E. (1973). A physiological model of aggression: Does it have different implications?
In Fields, W. S. (ed.), Neural Bases oj Violence and Aggression, Plenum, New York.
Mugford, R. A. (1973). Intermale fighting affected by home-cage odors of male and female
mice. J. Compo Physiol. Psychol. 84:289-295.
Mugford, R. A., and Nowell, N. W. (1970). Pheromones and their effect on aggression in
mice. Nature (Lond.) 226:967-968.
Mugford, R. A., and Nowell, N. W. (l971a). The preputial glands as a source of aggression-
promoting odors in mice. Physiol. Behav. 6:247-249.
Mugford, R. A., and Nowell, N. W. (l971b). Endocrine control over production and activity
of the anti-aggression pheromone from female mice. J. Endocrinol. 49:225-232.
Murphey, J. V., and Miller, R. E. (1955). The effect of adrenocorticotrophic hormone
(ACTH) on avoidance conditioning in the rat. J. Compo Physiol. Psychol. 48:47-49.
Mykytowycz, R. (1961). Social behaviour of an experimental colony of wild rabbits. IV. Con-
clusion: Outbreak of myxomatosis, third breeding season and starvation, C.S.I.R.O.
Wildl. Res. 6:142-155.
Norris, M. L., and Adams, C. E. (1972). Suppression of aggressive behaviour in the Mongolian
gerbil, Meriones unguiculatus. Lab. Anim. 6:295-299.
Oortmerssen, G. A. van (1971). Biological significance, genetics and evolutionary origin of
variability in behaviour within and between inbred strains of mice. Behaviour. 38:1-92.
Panksepp, J., Toates, F. M., and Oatley, K. (1972). Extinction induced drinking in hungry
rats. Anim. Behav., 20:493-498.
Payne, A. P. (1973a). The effect of androgens on isolation-induced aggression in the male

golden hamster. J. Endocrinol. 57:XXXVI.
Payne, A. P. (1973b). The aggressive behaviour of isolated male golden hamsters towards indi-
viduals of differing hormonal status. Acta Endocr. Suppl. 177:287.
Payne, A. P., and Swanson, H. H. (1970). Agonistic behaviour between pairs of hamsters of
the same and opposite sex in a neutral observation area. Behaviour. 36:259-269.
Payne, A. P., and Swanson, H. H. (197Ia). Hormonal modification of aggressive behaviour
between female golden hamsters. J. Endocrinol. 51:xvii-xviii.
Payne, A. P., and Swanson, H. H. (197Ib). The effect of castration and ovarian implantation
on aggressive behaviour of male hamsters. J. Endocrinol. 51:217-218.
Payne, A. P., and Swanson, H. H. (1972a). The effect of sex hormones on the aggressive be-
haviour of the female golden hamster Mesocricetus auratus Waterhouse. Anim. Behav.
20:782-787.
Payne, A. P., and Swanson, H. H. (1972b). The effect of sex hormones on the agonistic be-
havior of the male golden hamster Mesocricetus auratus Waterhouse. Physiol. Behav.
8:687-691.
Payne, A. P., and Swanson, H. H. (I 972c). Neonatal androgenization and aggression in the
male golden hamster. Nature (Lond.) 239:282-283.
Payne, A. P., and Swanson, H. H. (1973). The effects of neonatal androgen administration on
the aggression and related behaviour of male golden hamsters during interactions with fe-
males. J. Endocrinol. 58:627-636.
Peeke, H. V. S. (1969). Habituation of conspecific aggression in the three-spined stickleback
Gasterosteus aculeatus. Behaviour 35: 137-156.
Peeke, H. V. S., and Peeke, S. C. (1970). Habituation of aggressive responses in the Siamese
fighting fish. Behaviour 36:232-245.
Peeke, H. V. S., and Veno, A. (1973). Stimulus specificity of habituated aggression in the
stickleback Gasterosteus aculeatus. Behav. Bioi. 8:427-432.
Peeke, H. V. S., Wyers, E. J., and Herz, M. J. (1969). Waning of the aggressive response to
male models in the three-spined stickleback Gasterosteus aculeatus L. Anim. Behav.
17:224-228.
Peeke, H. V. S., Herz, M. J., and Gallagher, J. E. (197 I}. Changes in aggressive interaction in
adjacently territorial convict cichlids Cichlasoma nigrojasciatum: A study of habituation.
Behaviour 40:43-54.
Persky, H., Smith, K. D., and Basu, G. K. (1971). Relation of psychologic measures of ag-
gression and hostility to testosterone production in man. Psychosom. Med. 33:265-277.
Poole, A. E., and Brain, P. F. (1973). Effects of adrenalectomy and treatments with ACTH
and glucocorticoids on isolation-induced aggressive behaviour in male albino mice. Paper
presented at 9th International Summer School on Brain and Behaviour, Amsterdam.
Poole, T. B. (1973). The aggressive behaviour of individual polecats Mustela putorius. M. juro.
and hybrids towards familiar and unfamiliar opponents. J. Zool. 170:395-414.
Powell, D. A., Francis, J., and Schneiderman, N. (l97Ia). The effects of castration, neonatal
injections of testosterone, and previous experience with fighting on shock-elicited fighting.
Comm. Behav. Bioi. A 5:371-377.
Powell, D. A., Silverman, T., Francis, J., and Schneiderman, N. (l97Ib). The effects of sex
and prior experience with fighting on shock-induced aggression. Comm. Behav. Bioi. A
5l51-56.
Ratner, S. C. (1967). Comparative aspects of hypnosis. In Gordon, J. E. (ed.), Handbook oj
Clinical and Experimental Hypnosis, Macmillan, New York.
Reynierse, J. H. (1971). Submissive postures during shock-elicited aggression. Anim. Behav.
19:\02-\07.
296 John Archer
Reynolds, G. S., Catania, A. C., and Skinner, B. F. (1963). Conditioned and unconditioned ag-
gression in pigeons. J. Exp. Anal. Behav. 6:73-74.
Richards, R. W., and Rilling, M. (1972). Aversive effects of a fixed interval schedule of food
reinforcement. J. Exp. Anal. Behav. 17:405-41l.
Roberts, C. L., and Blase, K. (1971). Elicitation and punishment of intraspecific aggression by
the same stimulus. J. Exp. Anal. Behav. 15:193-196.
Rogers, L. (1971). The effects of testosterone and isthmo-optic lesions on visual search in
chickens. D. Phil. Dissertation, University of Sussex, England.
Rogers, L. (1974). Persistence and search influenced by natural levels of androgens in young
and adult chickens. Physiol. Behav. 12:197-204.
Rose, R. M., Holaday, J. W., and Bernstein, I. S. (1971). Plasma testosterone, dominance
rank and aggressive behaviour in male rhesus monkeys. Nature (Lond.) 231:366-368.
Russell, P. A. (1973). Open-field defecation in rats: relationships with body weight and basal
defecation level. Br. J. Psychol. 64: 109-114.
Savage, R. D. (1960). Constitutional determinants of emotionality in the rat: An experimental
investigation of fear, frustration and conflict. Ph.D. Dissertation, University of London,
England.
Schaller, G. B., and Emlen, J. T. (1962). The ontogeny of avoidance behaviour in some preco-
cial birds. Anim. Behav. 10:370-381.
Scott, J. P. (1946). Incomplete adjustment caused by frustration of untrained fighting mice. J.
Compo Psychol. 39:379-390.
Scott, J. P. (1971). Theoretical issues concerning the origin and causes of fighting. In Elef-
theriou, B. E., and Scott, J. P. (eds.), The Physiology of Aggression and Defeat, Plenum,
New York.
Scott, J. P., and Fredericson, E. (1951). The causes of fighting in mice and rats. Physiol. Zool.
24:273-309.
Seward, J. P. (1945). Aggressive behavior in the rat. I. General characteristics, age and sex dif-
ferences. J. Compo Psychol. 38:175-197.
Sigg, E. B., Day, c., and Colombo, C. (1966). Endocrine factors in isolation-induced ag-
gressiveness in rodents. Endocrinol. 78:679-684.
Slater, E. (1972). Some aspects of exploration and learning in three strains of mice. M. Phil.
Dissertation, University of Sussex.
Sluckin, W., Fullerton, C., and Guiton, P. (1970). The development of aversive responses in
domestic chicks. Rev. Compo Anim. 4:73-79.
Sokolov, E. N. (1960). Neuronal models and the orienting reflex. In Brazier, M. A. B. (ed.),
Central Nervous System and Behavior, Macy Foundation, New York.
Southwick, C. H. (1967a). An experimental study of intragroup agonistic behaviour in rhesus
monkeys. Behaviour 28:182-209.
Southwick, C. H. (1967b). Effect of maternal environment on aggressive behavior of inbred
mice. Am. Zool. 7:794.
Southwick, C. H. (1968). Effect of maternal environment on aggressive behavior of inbred
mice. Comm. Behav. BioI. A 1:129-132.
Spinage, G. A. (1969). Territoriality and social organization of the Uganda defassa water-
buck.J. Zool. (Lond.) 159:329-36l.
Stokes, A. W. (1962). Agonistic behaviour among Blue Tits at a winter feeding station. Be-
haviour 19:118-138.
Strecker, R. L., and Emlen, J. T. (1953). Regulatory mechanisms in house-mouse populations:
The effect of limited food supply on a confined population. Ecology 34:375-385.
Svare, B. B., and Leshner, A. I. (1973). Behavioral correlates of intermale aggression and
grouping in mice. J. Compo Physiol. Psychol. 85:203-210.
Swanson, H. H. (1973). Sex differences in the agonistic behaviour of the Mongolian gerbil. J.
Endocrinol. 57 :xxxviii-xxxix.
Tellegen, A., and Horn, J. M. (1972). Primary aggressive motivation in three inbred strains of
mice. J. Compo Physiol. Psychol. 78:297-304.
Tellegen, A., Horn, J. M., and Legrand, R. G. (1969). Opportunity for aggression as a rein-
forcer in mice. Psychon. Sci. 14:104-105.
Thoa, N. B., Eichelman, B., Richardson, J. S., and Jacobowitz, D. (1972). 6-hydroxydopa de-
pletion of brain norepinephrine and the facilitation of aggressive behavior. Science
178:75-77.
Thompson, D. M. (1964). Escape from So associated with fixed ratio reinforcement. J. Exp.
Anal. Behav. 7:1-8.
Thompson, T. (1969). Aggressive behaviour of Siamese fighting fish. In Garattini, S., and
Sigg, E. B. (eds.), Aggressive Behaviour. Excerpta Medica Foundation, Amsterdam.
Thompson, T., and Bloom, W. (1966). Aggressive behavior and extinction-induced response-
rate increase. Psychon. Sci. 5:335-336.
Tinbergen, N. (1952). The Study oj Instinct. Clarendon, Oxford.
Tonda, L. M., and Daly, H. B. (1972). The combined effects of frustrative non-reward and
shock on aggression between rats. Psychon. Sci. 28:25-28.
Turner, J. L., Boice, R., and Powers, P. C. (1973). Behavioral components of shock-elicited
aggression in ground squirrels. Anim. Learn. Behav. 1:254-262.
Uexkull, J. von (1934). A stroll through the worlds of animals and man. Translated in Schiller,
C. H. (ed.), Instinctive Behaviour: The Development oj a Modern Concept. Methuen,
London (1957).
Uhrich, J. (1938). The social hierarchy in albino mice. J. Compo Psychol. 25:373-413.
Ulrich, R. E. (1966). Pain as a cause of aggression. Am. Zool. 6:643-662.
Ulrich, R. E., and Azrin, N. H. (1962). Reflexive fighting in response to aversive stimulation.
J. Exp. Anal. Behav. 5:511-520.
Ulrich, R. F Johnston, M., Richardson, J., and Wolff, P. (1963). The operant conditioning of
fighting behavior in rats. Psychol. Rec. 13:465-470.
Ulrich, R. E., Hutchinson, R. R., and Azrin, N. H. (1965). Pain-elicited aggression. Psychol.
Rec.15:111-126.
Ulrich, R. E., Wolfe, M., and Dulaney, S. (1969). Punishment of shock-induced aggression. J.
Exp. Anal. Behav. 12:1009-1015.
Vale, J. R., Ray, D., and Vale, C. A. (1972). The interaction of genotype and exogenous neo-
natal androgen: Agonistic behavior in female mice. Behav. Bioi. 7:321-334.
Vandenberg, L. G. (1971). The effects of gonadal hormones on the aggressive behaviour of
adult golden hamsters. Anim. Behav. 19:589-594.
Vernon, W., and Ulrich, R. E. (1966). Classical conditioning of pain-elicited aggression.
Science 152:668-669.
Vine, 1. (1970). Communication by facial-visual signals. In Crook, J. H. (ed.), Social Beha-
viour in Birds and Mammals. Academic press, London and New York.
Watson, A. (1970). Territorial and reproductive behaviour of red grouse. J. Reprod. Fert.
Suppl. 11:3-14.
Weijnen, J. A. W. M., and Siangen, J. L. (1970). Effects of ACTH-analogues on extinction of
conditioned behavior. In DeWied, D., and Weijnen, J. A. W. M. (eds.), Pituitary. Adrenal
and the Brain. Elsevier, Amsterdam.
Welch, A. S. (1965a). Isolation-induced changes in brain biogenic amines. Am. Zool. 5:633.
Welch, A. S. (1965b). Effect of isolation on the metabolism of brain norepinephrine. ASB
Bulletin 12:54.
Welch, B. L. (1964). Psychophysiological response to the mean level of environmental stimula-
298 John Archer
tion: a theory of environmental integration. In Rioch, D. M. (ed.), Symposium on

Medical Aspects oj Stress in the Military Climate. U.S. Government Printing Office.
Washington D.C.
Welch, B. L., and Welch, A. S. (1965). Effect of grouping on the level of brain norepinephrine
in white Swiss mice. Life Sciences 4:1011-1018.
Welch, B. L., and Welch, A. S. (1966). Differential effect of chronic grouping and isolation on
the metabolism of brain biogenic amines. Fed. Proc. 25:623.
Welch, B. L., and Welch, A. S. (1969). Aggression and the biogenic amine neurohumors. In
Garattini, S., and Sigg, E. B. (eds.), Aggressive Behaviour. Exerpta Medica Foundation,
Amsterdam.
White, M., Mayo, S., and Edwards, D. A. (1969). Fighting in female mice as a function of the
size of the opponent. Psychon. Sci. 16:14-15.
Wilfe, M., Ulrich, R., and Dulaney. S. (1971). Fighting and escape reaction in paired rats.
Psychol. Rec. 21 :59-68.
Williams, R. B., and Eichelman, B. (1971). Social setting: Influence on the physiological
response to electric shock in the rat. Science 174:613-614.
Willis, F. N. (1966). Fighting in pigeons relative to available space. Psychon. Sci. 4:315-316.
Wolfe, 1. L., and Summerlin, C. T. (1968). Agonistic behavior of organized and disorganized
cotton rat populations. Science 160:98-99.
Wolfe, M., Ulrich, R., and Dulaney, S. (1971). Fighting and escape reaction in paired rats.
Psychol. Rec. 21:54-68.
Woods, P. 1. (1962). Behavior in a novel situation as influenced by the immediately preceding
environment. J. Exp. Anal. Behav. 5:185-190.
Wootton, R. 1. (1970). Aggression in the early phases of the reproductive cycle of the male
three-spined stickleback Gasterosteus aculeatus. Anim. Behav. 18:740-746.
Yen, H. C. Y., Day, C. A., and Sigg, E. B. (1962). Influence of endocrine factors on develop-
ment of fighting behaviour in rodents. Pharmacologist 4:173.
Chapter 8
ON INDIVIDUALITY: THE CONSTITUENTS

OF DISTINCTIVENESS
Department oj Zoology
University oj Maryland
College Park, Maryland
I. ABSTRACT
The notion that each organism has a certain degree of uniqueness is not
new. Most biologists, however, do not much appreciate interindividual
variability. It seems to blur our view of the "typical" or "ideal" case, which
becomes more clear only after the raw data have been refined with statis-
tical procedures. In fact, much of the difficulty in analyzing living systems
seems to derive from an apparent lack of congruence within a set of re-
plicas. Genetics is one of the few fields in which variability is investigated
systematically. But, in spite of the fact that individuality is blatantly ob-
vious in most organisms, and found on all levels from the biochemical to
the morphologic (Williams, 1956), we tend to suppress its existence, and
keep it, as a professional secret, from entering into textbooks.
In the study of animal communication, it has been realized for some
time that the complexity and variability of a signal can be related to its use
for individual recognition (Thorpe, 1961), or to the preventing of habitua-
tion by the receiving animal (Hartshorne, 1956). Not much thought,
however, has been given to the particular problem of how the complexity
and variability of such signals could be analyzed in greater detail.
In an investigation of the communicative function of the "wet-my-lips"
call of the male migratory quail (Coturnix coturnix coturnix L.), it became
desirable to compare the distinctiveness of this call when given by any
particular male with the calls of its neighbors and of the population as a
whole. I found that this task is easily managed if one considers not only the
299
300 Wolfgang M. Schleidt
variability of each characteristic, but also the number of characteristics (the

"multiparticity"). I then realized that these notions of distinctiveness, varia-
bility, and multiparticity can be applied to various other biological concepts
(e.g., the ecological niche as an n-dimensional space: Hutchinson, 1957).
Therefore, I attempted a more general treatment of their interrelations. For
the sake of clarity I have chosen to exemplify the interrelations between
variability and multiparticity with the example in which this problem arose
originally, namely the distinctiveness of a bird's call, and how this distinc-
tiveness may potentially be used for individual recognition among the calls
of neighbors.
There are several statistical methods available which can be applied to
such a problem, especially the "discriminatory analysis" (e.g., Kendall
1957). Unfortunately, few ethologists have sufficient understanding of
statistics to apply these methods properly; the majority are forced to either
ignore them, or to risk misusing them. Therefore, I have tried to present the
problem here as simply as possible, and to keep the mathematics as ele-
mentary as possible.
II. DEFINITIONS
Calls of animals, as well as other displays used in communication, are

often referred to as signals, or sometimes as signs. Since both terms have
well-defined meanings in information theory, we must not misuse them. It
seems desirable, however, to have a succinct term for a behavioral event
and those of its physical correlates which are likely to carry information to
the receiving communicant. I have proposed to term this a "key sign"
(Schleidt, 1973).
The physical properties of a key sign can be described completely by
the particular values of all its characteristics (e.g., intensity, speed, number
of elements, their duration, and the duration of the intervening intervals).
The set of these characteristics constitutes the multiparticity, in the sense of
a whole being composed of parts.
When we look at repetitions or replicas of the same key sign emitted by
the identical individual, or by another individual of the same species, we can
find that each characteristic shows a certain degree of variability. Depend-
ing on the source of those replicas, we must distinguish between intraindi-
vidual and interindividual variability. If the samples are sufficiently large,
the intraindividual variability will be smaller or equal to the interindividual
variability, but will not exceed it.
As a hypothetical case, let us consider a species which uses a double
On Individuality: The Constituents of Distinctiveness 301
pulse as an acoustical key sign. The characteristics of this key sign are the
following: the form of the pulses as such; the interval x between the two
pulses of such a pair within the key sign (intra-key sign interval); and the in-
terval y between successive key signs, or pairs of pulses (inter-key sign in-
terval). Further, let us assume that the form of the particular pulse is the
same in all individuals, and that only the interval x is used to convey the
identity of an individual, while the interval y is used to convey the readiness of
the individual to engage in a fight. Within a population, one individual might
display consistently a very short intra-key sign interval, another individual a
medium length intra-key sign interval, and a third one a very long intra-key
sign interval. For our notation, the individual animals ct, ffi, ... , m will each
produce a sample A, B, ... , N of key signs, and each sample will consist of
the identifying intra-key sign intervals XA, XB, ... ,XN from the particular in-
dividual. The accuracy by which the interval within a sample is maintained
can be expressed conveniently by the standard deviation s relative to the
mean X, known as the coefficient of variation, V ~ six, for the particular in-
dividual as VA, VB, ... , V N • The variability of the whole population, V pop ,
can be expressed accordingly. The degree of individuality of the charac-
teristic intra-key sign interval of animal ct relative to the population of con-
specifics can be represented by an index of individuality (Williams, 1971):
(1)
When we now ask how well we can discriminate between the key signs
of any two individuals within a population, it is obvious that this distinctive-
ness depends not only on the distance between the means, d = XA - XB, but
also on the width of the distribution of XA and XB relative to d (Fig. 1).
Therefore, I propose to define distinctiveness, D, by the degree of intersec-
tion between the two distributions FA(x) and FB(X) as
(2)
and the intersection of these two distributions as:

i = 1- D (3)
Note that the sample size does not enter in these expressions. Ob-
viously, as the sample size increases, the standard error of mean and of stan-
dard deviation will decrease, and our estimates of D and i will become more
accurate. I do not attempt, however, to establish confidence intervals for D
and i at this point.
When, in a particular key sign, individuality is invested in several
O+---~--~~~~----~
o x
Fig. 1. Given that the areas of the two distributions FA (x)
and FB (x) are of equal size, the area of their intersection (i;
stipled) is inversely proportional to the distinctiveness of the
constituent populations A and B.
characteristics which are not correlated, then the degree of overall distinc-
tiveness, DE, is not only inversely proportional to the variability of the
particular characteristics, but is also directly proportional to the number of
characteristics. In this sense, overall distinctiveness can be viewed as a com-
posite of variability and multiparticity. There is no a priori notion as to how
both will interact. Consequently, I propose a geometric model in Cartesian
space for special consideration. This model not only has interesting
properties when applied to the problem of individual recognition, but it also
has the additional advantage of being easily visualized and convenient for
computational procedures. However, let us look first at alternative models.
The simplest assumption that we could make in constructing a model is
that the distinctiveness of each characteristic is evaluated separately, and
that the overall distinctiveness is the sum of the "partial distinctiveness."
Another possibility could be to assign a weight to each partial distinctive-
ness, e.g., on the basis of their numerical value, so that characteristics
which contribute most to the overall distinctiveness are emphasized even
more. In the first case, i.e., without weighting, it will be very important to
know the full set of characteristics in order to derive an estimate of the
overall distinctiveness, while in the latter case, the accuracy of the estimate
will depend on the weighting function. The definite advantage of the
geometric model is that it applies a weighting on the basis of the number of
characteristics, so that relatively few characteristics of high individuality
provide by far the greatest portion of the measure of distinctiveness; and, as

we will see, the geometric model does not require that we know the com-
plete set of characteristics to arrive at a rather accurate estimate of the
overall distinctiveness.
III. THE GEOMETRIC MODEL OF DISTINCTIVENESS
The basic notion of distinctiveness, as defined by the degree of intersec-

tion between two distributions (equation 2), can be applied to any number
of characteristics when each characteristic is assigned a separate dimension.
In the case of n characteristics, each key sign is represented as a point in an
n-dimensional space. Samples of key signs from two particular individuals
form two point clusters in this hyperspace, and the degree of intersection
between these two clusters is inversely proportional to their distinctiveness.
Before discussing concrete cases, I will restate the problem, and make
a few assumptions which simplify the analysis. We have two animals, and
each produces one sample of its "individualistic" key signs. We want to
compare these two samples, A and B, to find out how distinct they are with
respect to each other. Both samples are composed of the same number of
observations (cases of key signs), and each observation has the same set
of characteristics, corresponding to dimensions 1, 2, 3, ... , n of the
model. For both samples of observations we can calculate one mean each,
a1, 2, 3•... , n. 51, 2. 3, ... , n, and the Euclidian distance dE is then calculated ac-
cording to the Pythagorean theorem:
dE = V d~ + d~ + d~ + ... + d~ (4)
We can then introduce a new axis, E, which goes through the two
means, and find the normal projection of each point on this "E-axis." This
transformation yields the probability density distributions .of A and B along
the E-axis. The extent of their distinctiveness can now be calculated con-
veniently according to equation 2, and that of their intersection according
to equation 3. Evidently, this area of intersection will depend on the form of
the distributions of A and B within the various characteristics (dimensions),
but as long as they are all unimodal their particular shape should not matter
much, since the amount of intersection along the E-axis is determined em-
pirically, without making assumptions concerning their shape or generating
process. To better understand the influence of the constituent characteristics'
distribution, let us consider two special cases; that is, all characteristics are
distributed either uniformly or normally.
A. Uniform Distribution
For simplicity, let us first assume a case in which n independent

characteristics are uniformly distributed, each having the same range, which
I propose to signify by 2r (two times the radius of the hypersphere). The dis-
tinctiveness of these two clusters in hyperspace increases with the distance d
between the respective means of each characteristic (within each dimen-
sion), and with the number of characteristics (dimensions) employed. If we
assume that d is equal in all dimensions, equation 4 can be simplified:
dE = ~ (5)
Whenever the Euclidian distance dE exceeds 2r, complete separation is
achieved. As can be seen in Fig. 2a, and from equation 5, increasing the
number of characteristics from one to two results in a sizable improvement
in the distinctiveness, but adding an additional characteristic to a set of
many characteristics has only a minor effect (i.e., to double the value of dE,
n must be increased from 1 to 4, but to quadruple it, n must be increased
to 16).
Note that in such a hypersphere the points are most densely packed at
the surface, and least densely packed in the center, as soon as we have 2 or
more variates. If we assume instead that the hyperspheres are of uniform
density, then the distribution function of the variates is dependent on the
number of dimensions of the hypervolume. If n = 1, it is uniform, if n = 2,
it becomes half an ellipse, and as n increases further it becomes bell-shaped
(over the range of 2r), and, as n goes toward infinity, it degenerates into a
very narrow spike. If the degree of intersection of two such hyperspheres
would be depicted within the same coordinate system as in Fig. 2a, the two
endpoints of each line would be identical, but the line itself would be slightly
curved downward in the case of n = 2, and, as n is increased, this curvature
would become more and more striking.
B. Normal Distribution
If we assume that n independent characteristics are normally dis-

tributed, each having the same standard deviation 0', we are dealing with the
intersection of two multivariate normal distributions. If we assume further
(as in the uniform case) that d is equal in all dimensions, equation 5 applies
here as well, with the difference that the unit of separation is 0'. Although
complete separation cannot be achieved, because both distributions cover
the range from - 00 to + 00, the degree of intersection rapidly approaches
zero, as can be seen in Fig. 2b.
0.5
1
2
Fig. 2. The relation between the distance between the means d of two distributions in each di-
mension, and the intersection i of two hypervolumes, dependent on the number of dimensions n
of the two intersecting hypervolumes, given that the distributions of al1 variates (I) have the
same standard deviation, (2) have the same distance between the two means within each dimen-
sion, and (3) are not correlated between dimensions. (a) For the case that, within each dimen-
sion, the variate is distributed uniformly and has the range 2r. (b) For the case that, within
each dimension, the variate is distributed normal1y and has the standard deviation s. The
distance d between the two means is expressed in terms of rand u, respectively.
C. General Considerations
If we assume n independent characteristics, all having the same dis-

tribution but different standard deviations, the characteristics with a large
value of D contribute substantially to the overall "Euclidian" distinctive-
ness, DE, while even a large number of additional characteristics with small
D have only a minor effect. Consequently, if the characteristics are ranked
in decreasing order of their contribution to the overall distinctiveness (i.e.,
according to the numerical values of Dx , Dy , Dz , • . . ,Dn ), DE can be calcu-
lated for increasing numbers of characteristics, and the value of DE, which
is approached asymptotically, can be accepted as an estimate for the
maximal distinctiveness possible in this particular case, Dmax. Such a rank-
ing of characteristics in decreasing order of their partial distinctiveness
further enhances the importance of the highly distinct characteristics, and
diminishes to a greater degree the contribution of the characteristics of
moderate or low partial distinctiveness. If this principle is realized in a bio-
logical system in which distinctiveness is desirable, we can expect to find it
invested in a few essential characteristics, rather than in many characteris-
tics that equally impart a small measure of distinctiveness.
The maximal distinctiveness of two sets of characteristics can be im-
portant, when the relations between two particular individuals showing
these characteristics are compared with the average of a population of indi-
viduals, or with any two individuals drawn at random from the population.
This method could be applied to determine whether the individuals within a
local population resemble each other more closely than members of
another, geographically separated population. On a finer scale, however, it
would also allow us to investigate whether nearest neighbors, e.g., holders of
adjoining territories, are more or less similar than are members of the popu-
lation in the average.
The problem of individuality within a population can be approached
from a different angle: we can ask how the mean Dmax of a community
relates to the number of different individuals which each member must be
able to discriminate in a particular situation (i.e., dependent on density),
and we can use this relationship to estimate the theoretical individuality en-
coding capacity of the population. The latter can be expressed by the ratio
of the hypervolume determined by the multivariate distribution of all
characteristics displayed by the population to the mean hypervolume of the
distribution of all characteristics displayed by the individual member.
Of course, the same principle can be applied when we compare the
degree of distinctiveness between any two different key signs within the
repertoire of a species, to estimate the capacity of this repertoire, or to de-
termine how the members of a community share the available "space for
encoding" in their particular environment.
IV. DISCUSSION
To identify, that is, to assign a particular item to one of several discrete

categories, is a very basic ability of the human mind. Furthermore, it is
deeply ingrained in our scientific thinking, not only in the predominantly
descriptive sciences (e.g., systematics), but also in those sciences which
relate to dynamic properties of living systems (e.g., physiology), where
discrete categories provide a degree of security and rigidity in the con-
ceptual framework. Even when the circumstances force us to phrase a state-
ment probabilistically, we still tend to establish discrete limits (e.g., confi-
dence intervals) to satisfy our need for sharply defined and delineated cate-
gories. Hassenstein (1954) has pointed out that rigid definitions are often
inappropriate, especially in biology, and has suggested the creation of
graded, adjoining concepts (injunktive Begriffe; note that this term has a
different meaning from "injunctive," and relates to the primary meaning in
Latin, "to add to"). Injunktive Begriffe allow continual transitions between
concepts, and the use of mUltiple features for concept characterization.
More recently, Zadeh (1965) has formalized very similar ideas in his con-
cept of fuzzy set as a class of objects with a continuum of grades of
membership. Such attempts to cope with the problem of placing an item
along a gradient are relevant to the problem of distinctiveness in two dif-
ferent respects. First, epistemologically, the concepts of distinctiveness,
variability, and multiparticity are graded from zero (as defined by their op-
posites: indistinctiveness, constancy, and undividedness) to infinity. Second,
the features of characteristics which we use to describe the entities of our
investigation (e.g., key signs) are graded, and consequently any differences
between classes of such entities must also be graded.
The problem of overlap between categories of classification is not new
in biology. Probabilistic concepts to handle classification problems were
used in taxonomy for some time. Similar problems arose in a new form
when high speed electronic computers were used to model cognitive
processes and to perform the function of probabilistically identifying com-
plex patterns. Interestingly enough, the central issue in these studies seems
to have been how to decide in a concrete case whether a particular item is a
member of the set A or B. and what probability to assign to this identifica-
tion. This led to the development of various algorithms which allow us to
approach such problems very elegantly, and on a high level of sophistication

(e.g., Duda and Hart, 1973). However, little attention has been given to the
relations among classes with fuzzy boundaries. Calhoun (1963) was one of
the first to approach such a problem, when he proposed to describe home
ranges of small mammals by using bivariate normal distributions of the
presence of particular individuals within the community. When Hutchinson
(1957) suggested that we consider the ecological niche as a hypervolume, he
created a concept in which the peculiarities of overlap between the niches of
competing species and the relative width of these niches became important
properties. However, the "multidimensionality" of the niche, though
generally accepted as a concept, has only recently been explored in any de-
tail. If we apply the "geometric model of distinctiveness" to the niche con-
cept, we can expect that competing species will maximize their separation
not only by filling "their" segment of the niche uniformly, but also by doing
this in several dimensions, though limited in number.
As for the problem of individuality in humans, we generally take it for
granted that we can identify a large number of individuals by features of
their face, posture, movement, or voice, realizing the immense complexity
of this ability only when we try to establish objective criteria for such dis-
criminations (Goldstein et al., 1971; Hair and Rekieta, 1972). We know
from the striking resemblance of a good caricature that relatively few ele-
ments can be of overriding importance, but little is known about how the
particular features are ranked, or weighted, by our perceptual processes in
discerning a likeness. It is obvious to everyone that members of the same
family often resemble each other, and it is known that inhabitants of a
small village in a rural area are more similar to each other than to the
people of neighboring villages (e.g., Spielman, 1973), but little is known
about whether particular features of the human body, especially of the face,
have been selected specifically to enhance individual recognition (e.g., hair
pattern: Miller, 1931; Setty, 1971). A complicating factor in encoding indi-
viduality in a human face is that encoding cannot be accomplished simply
by a set of static f~atures, as in a fingerprint, but has the movement of
facial expression and speaking superimposed. The concepts of variability
and multiparticity as constituents of distinctiveness might help to overcome
these difficulties.
The central aim of this paper is to suggest that the notion of the dis-
tinctiveness of a structure or a behavior can be represented by the degree of
intersection of multivariate distributions. I wish to point out, however, that
this geometric model can also be helpful in representing the relationships
between other biological concepts in which variability and multiparticity are
constituents. For example, Vermeij (1973) has suggested that "the potential
versatility of a given higher taxon or body plan is determined by the number
and range of independent parameters controlling form." Translated into the

geometric model, versatility could be expressed as the product of multi-
particity and variability.
V. SUMMARY
The distinctiveness of two structures or behavior patterns is a function

of the number of characteristics associated with them, and of the variability
of these characteristics. This relation can be represented by a multidimen-
sional geometric model of two intersecting multivariate distributions, where
each characteristic common to both distributions is represented by a
variate. The degree of distinctiveness of these distributions can be expressed
by the inverse of their intersection. If we assume that both distributions are
multivariate normal, and that all variates have the same standard deviation,
it can be shown that as the number of variates increases their individual
contribution approaches zero asymptotically, and relatively quickly. If a
particular structure or behavior should serve to achieve a high degree of dis-
tinctiveness (e.g., to allow individual recognition among conspecifics), the
optimal compromise appears to be the use of several, but few, characteris-
tics of small variability, rather than of one characteristic of extremely small
variability or of many characteristics of moderate variability.
VI. ACKNOWLEDGMENTS
Special thanks go to my students, who forced me to clarify my fuzzy

concepts, and especially to Mike Shalter, Irene Magyar, and George Sa-
maras. Drs. Gerrat Vermeij, Azriel Rosenfeld, Piotr Mikulski, Han Golani,
and Walter Elsasser made many helpful suggestions. Valerie Maiorana, Bob
Mullen, and Dave Williams assisted in the process of data analysis, from
which my quantitative approach to the problem of individuality of bird song
developed. A part of this research was supported by the National Institute
of Mental Health (grant MH 20-867-01).
VII. REFERENCES
Calhoun, J. B. (1963). The social use of space. In Physiological Mammalogy, Mayer, W. V.

and van Gelder, R. G. (eds.), Academic Press, New York: pp. 1-187.
Duda, R. 0., and Hart, P. E. (1973). Pattern Classification and Scene Analysis, John Wiley &
Sons, New York, 482 pp.
Goldstein, A. J., Harmon, L. D., and Lesk, A. B. (1971). Identification of human faces. IEEE
Proc.59:748-760.
Hair, G. D. and Rekieta, T. W. (1972).Speaker Identification Research, Texas Instruments
Inc., Houston, 110 pp.
Hartshorne, C. (1956). The monotony threshold in singing birds. Auk 73:176-192.
Hassenstein, B. (1954). Abbildende Begriffe. Verh. Dtsch. Zool. Ges., Geest & Portig, Leipzig,
pp. 197-202.
Hutchinson, G. E. (1957). Concluding remarks. Cold Spring Harbor Symposia on Quantitative
Biology 22:415-427.
Kendall, M. G. 1957. A Course in Multivariate Analysis. Charles Griffin & Compo Ltd.,
London. 185 p.
Miller, G. S., Jr. (1931). Human hair and primate patterning. Smithsonian Misc. Coli. 85
(10):1-13.
Schleidt, W. M. (1973). Tonic communication: Continual effects of discrete signs in animal
communication systems. J. Theor. Bioi. 42:359-386.
Setty, L. R. (1971). Hair patterns of the face of White and Negro males. J. Nat. Med. Assoc.
63:128-13.
Spielman, R. S. (1973). Do the natives all look alike? Size and shape components of anthro-
pometric differences among Yanomama Indian villages. Amer. Natur. 107:694-708.
Thorpe, W. H. (1961). Bird-song, Cambridge Univ. Press, Cambridge, 143 pp.
Vermeij, G. J. (1973). Biological versatility and earth history. Proc. Nat. A cad. Sci. USA
70:1936-1938.
Williams, D. H. (1971). Individual variation in the bobwhite call of Colinus virginianus. M.S.
Thesis, Univ. of Maryland, College Park, 31 pp.
Williams, R. J. (1956). Biochemical Individuality, John Wiley & Sons, New York, 214 pp.
Zadeh, L. A. (1965). Fuzzy sets. Inform. & Control 8:338-353.
Chapter 9
THE PRINCIPLE OF OPTIMAL BEHAVIOR

AND THE THEORY OF COMMUNITIES
H. Ronald Pulliam
Department of Ecology and Evolutionary Biology
University of Arizona
Tucson. Arizona
I. ABSTRACT
The notion that behavior maximizes fitness can be used as a tool for
generating testable hypotheses concerning the expected behavior of animals.
In turn, the expected behavior of animals with respect to available resources
can be used to predict much of the biotic structure and dynamics of com-
munities. This paper attempts to illustrate the use of simple optimization
techniques for predicting optimal behavior and for relating behavior to dy-
namical models of ecological communities. The specific behavior optimiza-
tion problems which are considered all relate to the use of food and space.
Examples are diet optimization under nutrient constraints, optimal foraging
path and microhabitat selection, and optimal sociality in relationship to diet
and habitat selection. The criteria for habitat and diet specialization vs.
generalization are presented in terms of the behavioral attributes of the
animal in relation to the availability of resources. The section on optimal
sociality illustrates how social behavior can be predicted from reasonable
assumptions regarding the time budget priorities of animals.
The results of animal behavior upon encountering food and habitat
resources are considered as the 'microecological' phenomena which, when
taken all together, determine the 'macroecological' phenomena of commu-
nities. The final section of the paper attempts to show how models of
optimal behavior can be used to determine the form of mathematical
models of the dynamics of communities.
311
312 H. Ronald Pulliam
II. INTRODUCTION
In this paper, I argue the basic importance of developing and testing

models of optimal behavior as a possible foundation for new theories relat-
ing population ecology to community ecology. For the most part, com-
munity ecology and population ecology have developed independently over
the past two decades. One major attempt to construct a theory of com-
munity ecology based on the theory of population ecology has been the
work on the "community matrix" (see Levins, 1968; May, 1973). This ap-
proach, based on fundamental concepts of population ecology, presents
exciting possibilities for testing ideas concerning community diversity and
stability, but has the shortcoming of only considering communities near
equilibrium, where the interactions between populations are represented by
fixed interaction coefficients. Thus, the interaction between a predator
population and a prey popUlation is represented by a pair of constants, al-
though it is well-known that predator-prey interactions in nature are ac-
companied by a host of density-dependent and functional responses. I hope
that this paper will serve as an indication of the utility of a "principle of
optimal behavior" in predicting the precise nature of these density-de-
pendent and functional responses, and thus of the utility of such a principle
in building a needed bridge between population and community ecology.
The principle of optimal behavior is new only in name. Concisely
stated, it is the notion that behavior maximizes fitness. As a tool for
generating testable hypotheses, it has been most successfully applied, thus
far, to the theory of feeding strategies (for review see Schoener, 1971). The
principle has been used to generate testable hypotheses on a variety of prob-
lems of interest in behavioral ecology. These include habitat selection (Fret-
well, 1972), movement patterns (Cody, 1971, 1974), mate selection
(Downhower and Armitage, 1971), flocking (Gadgil, 1972), and altruism
(Hamilton, 1964). The list could be extended easily to include two score or
more references, but those above are chosen as representative of the range
of applications. The idea has been implicit in the thinking of ecologists and
ethologists at least as far back as Darwin.
Analogous to the principle of optimal behavior is the concept of the
adaptive morphology of organisms. However, discussion of the kangaroo
rat having an optimal length of the Henle's loop evokes less of a visceral
reaction in many people than does a discussion of the kangaroo rat exhibit-
ing an optimal foraging behavior. I suspect two reasons for the difference in
reactions. First, there appears to be a common notion that morphology is
somehow more objective than behavior, and therefore that it might be easier
The Principle of Optimal Behavior and the Theory of Communities 313
to decide a priori what an optimal Henle's loop length might be. I reject this
notion, and hope that this manuscript illustrates that optimal behavior can
be specified a priori. Second, there is the commonly held belief that natural
selection is better able to work on morphology than on behavior. This belief
is to some extent based upon a proper interpretation of the current litera-
ture on behavioral genetics, and I offer no argument against it. I simply ask
the reader to bear in mind two points: (1) the principle is not dependent
upon the genetic determination of behavior, and (2) the models presented
should be judged solely on their predictive ability, rather than on any
preconceived notion as to whether or not behavior optimizes anything.
Schoener (1971), in a discussion of feeding strategies, has given a clear
statement of the procedure used to predict optimal behavior a priori:
The primary task of a theory of feeding strategies is to specify for a given animal that
complex of behavior and morphology best suited to gather food energy in a particular
environment. The task is one, therefore, of optimization, and like all optimization prob·
lems may be trisected: (I) Choosing a currency: What is to be maximized or minimized?
(2) Choosing the appropriate cost-benejit junctions: What is the mathematical form of
the set of expressions with the currency as the dependent variable? (3) Solving jor the
optimum: What computational technique best finds the extreme of the cost-benefit
function? (p. 369)
Certainly step one is the hardest, since it requires a determination a

priori of what the animal is optimizing. In some respects this is akin to as-
cribing goal-orientation to behavior, and is open to the same pitfalls. To
avoid such pitfalls, I never assume what is being optimized, but rather ask a
question like "What would the behavior which optimizes food intake be?"
This leads then to a testable prediction that the observed behavior is the
same as that which optimizes food intake.
In this paper I will use the phrase "the animal chooses to" as being
interchangeable with the phrase "the animal was observed to." Thus, an
animal's choice is defined by its observed behavior and, as such, is the overt
analogue of the term "decision" as used by Dawkins and Dawkins (1973).
One goal of this paper is then to predict some of the choices made by ani-
mals in their natural environment.
I will consider optimal choice as regards diet, habitat selection, and so-
ciality. With regard to diet I consider the cases of time minimization and
caloric maximization, with and without constraints on nutrient intake. The
models of optimal habitat selection include the optimal foraging path and
micro- and macro habitat selection under considerations of optimizing diet
and minimizing risk of predation. Finally, the discussion of optimal so-
ciality shows relationships between 'diet, habitat selection, and sociality, and
includes a model of optimal group size.
III. MODELS OF OPTIMAL DIET
The criterion for inclusion of a potential prey item into the realized diet
was stated clearly by Robert MacArthur (1973) in his book Geographical
Ecology:
On locating an item, the consumer has two choices: to pursue or to search again for a
better item and pursue it instead. Either choice ends with the animal ready to start a
new search, the better decision is the one that promises to yield more per unit time. In
other words an animal should elect to pursue an item if and only if, during the time the
pursuit would take, it could not expect both to locate and to catch a better item. (p. 62)
Following the suggestion of MacArthur and Pianka (1966) and the notation
of Pulliam (1974), the foraging time can be partitioned into the time, T p ,
involved in pursuit and handling once a prey is located, and the search time,
T., which includes the time between the end of pursuit and handling of one
prey, and the locating of the next. The expected total time per prey in the
realized diet is given by
T = T. + Tp
If Ii is the pursuit plus handling time per prey of prey species i,
MacArthur's statement quoted above can be represented by the criterion
that prey species i is included in the realized diet if and only if It < T. + Tp.
As MacArthur (1973) shows, this simple criterion is useful as a conceptual
tool for relating diets to competition. However, the criterion is not specific
enough to be testable in most field situations. I outline below some more
specific and realistic criteri~ for inclusion of potential prey in the realized
diet.
Let Ai be the number of prey items of type i encountered per unit
search time by the predator. Upon each encounter, the predator may pursue
and eat the item (with probability Pi), or may ignore the item and continue
to search (with probability 1 - Pi)' Assume that if the predator pursues a
prey item, he is always successful in capturing and eating it, and that this
will require Ii units of time. (Ii is hereafter referred to as pursuit time.) The
number of prey of type i actually encountered and eaten per unit search
time is then given by Xi = PiAj, and the total number of prey of all types en-
countered and eaten per unit search time is 1: tXt. Thus, the search time per
prey eaten is l,l}:tXt. The total time required per prey actually eaten is the
sum of the search time per prey and the pursuit time per prey. The expected
pursuit time per prey is 1:iXiti,l}:iXi. Thus, the total time per prey item
eaten is
Consider the case where a predator has the choice between only two
possible prey types, such that
T = + Xltl + X2t2 (I)
Xl +X2 Xl +X2
Recalling that XI = PlAt. we can outline three possible strategies of preda-
tion as follows. (a) The predator can specialize only on prey type one, in
which case PI = 1, P2 = 0, and, from equation (1)
T= 1 P1 A1t 1 + P2 A2t 2 1
+ =-+ t1
PIAl + P2 A2 PIAl + P2 A2 A1
(b) The predator can specialize only on prey type two, in which case P1 = 0,
P2 = 1, and T = I/A2 + t2. (c) The predator can eat the two preys in the
same proportions as they are encountered, in which case P1 = P2 = 1 and
T= + t1A1 + t2A2
A1 + A2 A1 + A2
The optimal strategy is, by definition, the one which minimizes the
total time, T, required per prey eaten. Now, assume that the pursuit time
per prey of type 2 is greater than the same per prey of type 1 (that is, 12 >
(1). Strategy (c) is better than strategy (b) if, and only if, the total time per
prey eaten when feeding by strategy (b) is greater than the total time per
prey eaten when using strategy (c). This is stated mathematically as
1 1 I1A1 + 12A2
-+ 12> +----
A2 A1 + 11.2 A1 + A2
which reduces to I/A2 > t1 - 12. The latter is always true, since t2 is greater
than II. Thus, a predator feeding so as to minimize the time required per
prey item eaten should never specialize on the prey with a higher pursuit
time, regardless of the density of that prey type.
The choice whether to specialize on prey type 1 (strategy a) or to
generalize and eat both prey types (strategy c) depends only on the density
of prey type 1 relative to the difference in pursuit times for the two prey
types. That is, the specia:list diet is optimal if, and only if
which reduces to
1
- < 12 - 11 (2)
A1
This is an explicit and testable criterion for specialization.
The above results can be easily extended to include the caloric value of
the prey. Let C i be the caloric value of prey type i. Then the average
calories per prey item of a diet consisting of Xl prey of type 1, and X 2 prey
of type 2, is given by (CIXI + C2X 2 )/(XI + X 2 ). The calories per prey item
divided by the total time per prey item [equation (1)] gives the average
calories per unit time as
C = CIXI + C2 X 2 (3)
1 + tlXI + t2X 2
Now, assume that the calories per unit pursuit time for prey type are
greater than the calories per unit pursuit time for prey type 2 (that is, CI /
tl > C 2/t 2 ). If the predator were to specialize on prey type 1, C would equal
CIAd(l + Altl). If, instead, the predator ate every prey item as it was en-
countered (a generalist strategy), C would equal
CIA I + C2A2
1 + tlAI + t2A2
Thus, the condition for the specialist strategy to be optimal is that
CIA I CIA I + C2A2
>
1 + Aitl 1 + IIAI + t2A2
which reduces to
1 C l t2
-< (4)
Al C2 - tl
I
This means that the condition for the generalist diet to be optimal is that 1/
Al > C l t 2/C 2 - It. and the optimal switching point occurs at the prey
density which satisfies I/A I = C l tdC2 - t i •
In the above discussion, I have considered only the situations where the
probability of eating an encountered prey, Pi, is either 0 or 1. In more
general treatments of this problem (Pulliam, 1974, 1975), I have shown that
all optimal solutions, when there are no nutrient constraints, have the charac-
teristic that Pi is 0 or 1. This leads to the interesting conclusion that a preda-
tor feeding so as to minimize the total time required per prey item eaten, or so
as to maximize the rate of caloric intake, should eat a prey item on every en-
counter with it or else not at all. In the following, I show that this is not
necessarily true when a predator maximizes caloric intake under the
constraints of a balanced diet.
A diet which maximizes the rate of caloric intake, as outlined above,
might be expected to occur in certain stress situations such as extremely
cold weather. But in general an animal must also have a diet with a balance
of nutrients. The problem then becomes one of maximizing the rate of ca-
loric intake under the constraint that certain nutrients are supplied in at
points
-I------------::""!--'-+----~ X,
CASE I
points
X2 - - - - - - - - - - - -
,l, x, + 1l2X2~
Fig. l. Solutions to the optimal diet problem with and without nutrient constraints. The
hatched area represents the feasible set in which all solutions must lie to meet nutrient
constraints. See text for discussion of the cases illustrated.
least minimal amounts. Mathematically, this is a more difficult problem

than those considered above. For the general solution, see Pulliam (1975),
for applications to the diet problem, and Intrilligator (1971) and Whittle
(1971) for discussions of the general theory. A simple graphical solution is
given below (Fig. 1) for the case of only two potential prey species.
As before, let Xl and X 2 be the prey eaten per unit search time by the
predator. These amounts are under the obvious constraints that 0 < Xl < Al
and 0 < X 2 < A2, where Al and A2 are the encounter rates per unit search
time, as above. Consider the additional constraint that the diet must contain
some minimal amount, M, of a specific nutrient. If Pi is the amount of this
nutrient in one individual of prey species i, then this requirement can be writ-
ten as PIXI + P~2 2: M. The problem can then be written as the maximiza-
tion of
C1Al + C2A2
1 + I1Al + 12A2
under the constraints (a) 0 < Xl < A1> (b) 0 < X 2 < A2' and (c) PIXI + P~2
2: M. These three constraints define a feasible set (as shown by the hatched
area in Fig. 1) in which the solution must lie.
Case I in Fig. 1 shows the solution to the problem when the specialist
diet is in the feasible set, that is, when a diet composed of only prey type 1
satisfies the nutrient requirements. The solution is exactly the same as in the
unconstrained maximization, that is, specialize when l/AI < C1 /21 C2 - I},
and generalize otherwise.
In case II of Fig. 1, the specialist diet falls outside of the feasible set.
The solution in this case is to eat both prey in the same proportions as they
are encountered, if l/AI > C 1 / 2 /C2 - 11, and otherwise to eat prey type 1
every time it is encountered, but to eat prey type 2, X 2* out of A2 encounters.
Thus, P2 = X2* IA2 for the optimal solution; the constrained diet problem
does allow for partial preferences, whereas the unconstrained problem does
not.
Case III of Fig. 1 shows how three different constraints interact to
define the feasible set. These constraints might represent crude protein, cal-
cium, and vitamin B requirements. There is no limit to the number of
constraints which can be considered simultaneously, but with a sufficient
number of constraints the feasible set may be empty. Note that in case III
only two constraints actually define the feasible set, since the third is always
satisfied when the other two are satisfied.
IV. OPTIMAL MODELS OF HABITAT SELECTION
Habitat selection is discussed here in terms of microhabitat and

macro habitat selection. By microhabitat, I mean a homogeneous patch
which is much too small to provide a forager with all of its food require-
ments. A macrohabitat is a heterogeneous area containing many such mi-
crohabitats, or a homogeneous area large enough to supply all the forager's
food needs. Microhabitat selection results in a differential use of patches
within macrohabitats, whereas macrohabitat selection results in a dif-
ferential use of areas large enough to contain all foraging. For a large bird,
macro habitats might be deciduous woods and pine woods, and mi-
crohabitats might be oak trees and elm trees within the deciduous woods.
For a small predatory insect the macrohabitats might be oak and elm trees,
and microhabitats might be oak leaves and oak twigs.
The distinguishing characteristic for microhabitats is that the foraging
animal must continually cross one patch type to reach another patch type.
If patch type A provides more calories per unit foraging time than does
patch type B, the question whether or not to specialize on A is not as simple
as first appears. In the model that follows, I present a special case of mi-
crohabitat selection. The algebra used is simple, and the examples merely
provide guidelines for the modeling of specific field problems.
Figure 2 illustrates the situation modeled. Envision a bird foraging on a
limb and looking for insects. Assume the following:
1. That each leaf has a probability p of having an insect on it, and each
interleaf portion of the branch has a probability q of containing an
insect.
2. That the probability of more than one insect per leaf, or per
interleaf portion of the branch, is negligibly small.
3. That an insect when captured yields C calories and requires t
seconds to be captured and eaten, regardless of the site of capture.
4. That to search a leaf requires t8 seconds, but since the bird must
walk along the branch, at least T seconds are required whether or
not the bird stops to capture prey encountered.
Though the assumptions may seem unlikely for a real situation, they are
meant only to illustrate the modeling technique, and could be made more
realistic with only slightly more complex algebra. The main reason for set-
ting the assumptions as above is that they are precisely what is needed for a
simple test of optimal microhabitat selection. A simple experiment with one
prey type could be easily arranged meeting the above assumptions.
From the listed assumptions it is easy to derive the calories per second
expected for generalist and specialist diets. For each strategy, consider the
rate of caloric intake from the base of one leaf to the base of the next. The
leaf specialist can expect pC calories in T + t8 + pt seconds. Similarly, the
branch specialist can expect qC/(T + qt) calories per second. The generalist
would get the calories from both the leaves and the branches, which would
amount to pC + qC and would require T + t8 + pt + qt seconds. Thus, the
Fig. 2. A bird may forage only on the branches of a tree (branch specialist),
or may forage both on the limb and the leaves (generalist). A bird which
moves along the generalist path, but eats only prey encountered on the leaves,
is called a leaf specialist.
generalist diet yields more calories per second than the leaf specialist if
pC + qC pC
-----'--->-----
T + Is + pI + qt T + Is + pt
which reduces to
(p + q)(T + Is) > 0
which is always true. Since it never pays to specialize on the leaves, the
question becomes whether to specialize on the branches or to generalize.
The optimal predator would specialize on the branches if, and only if,
qC pC + qC
- - - > -....:....--"'----
T + qt T + t. + pt + qt
This condition reduces to
q ( ; +~ + t) > T
and in this form it is readily seen that specialization on the branches is more
likely as the percentage of branch segments with insects, q, increases and
the percentage of leaves with insects, p, decreases. This is intuitive even
without the model. What the model adds is a precise, testable criterion of
optimal microhabitat selection.
The problem of optimum macrohabitat selection is trivial in the case of
the criteria of caloric maximization. With use of the techniques described in
the preceding section, two habitats can be directly compared as to their ex-
pected caloric return to a predator feeding in an optimal manner. Under
this criterion, the optimal macrohabitat would be the one with the higher of
the two caloric returns. This gives a precise means for testing the hypothesis
that an animal is selecting habitats based simply on a consideration of ca-
loric maximization. I expect that nature is rarely so simple. Particularly
under conditions of energetic stress, criteria of caloric optimization-with
or without nutrient constraints--ml1Y be very predictive for diet selection
and microhabitat selection, but are probably of comparatively little im-
portance in macrohabitat selection. Certainly, in the birds I have studied,
and probably in most vertebrates, considerations of nesting possibilities and
predation risk are important in habitat selection. The models described in
the next section of this paper simultaneously consider the problems of de-
tecting predators and finding enough food. A yet more complex model
which also considers macrohabitat differences would not be difficult to
construct.
V. MODELS OF OPTIMAL SOCIALITY
In this section, I combine the results of the section on diet optimization

with some simple and reasonable assumptions about time budget priorities,
in order to outline conditions under which flocking is advantageous. The
reasoning presented is based largely on my observations of winter flocks of
sparrows. Two observations in particular have been important for the

development of these ideas. First, the birds frequently stop and cock their
heads to the side. Such behavior could obviously be adaptive to the extent
that it leads to the detection of predators. When a bird does detect a preda-
tor the entire group flies off together. Second, there appear to be two more
or less distinct kinds of agonistic encounters in flocks. The first is a simple
displacement which occurs at a local food clump. This displacement takes
place within a second or two, with one bird (called the "dominant") sup-
planting a second bird (called the "submissive"). A second kind of agonistic
encounter seems quite unrelated to the presence of food. It may occur when
the birds are feeding, but is also likely to occur when they are resting or
preening in shrubs. This type of encounter is characterized by its duration
(often 30 sec or more) and its apparent lack of a spatial reference, such as a
food clump. In what follows, I assume that the short-duration displacement
confers on the dominant bird an immediate benefit of access to a food
clump, and that the long-duration agonistic encounter confers a long term
benefit to the aggressor by decreasing flock size and thereby ensuring a long
term food supply.
The assumptions I make regarding time budget priorities are as
follows: (1) That time spent in predator detection takes priority to time
spent feeding. (2) That time spent retreating from aggressive birds takes
priority over time spent feeding. This is true almost by necessity since a bird
could hardly feed when under attack. (3) That the long duration aggression
mentioned above is of a lower priority than feeding. This is simply to say
that an animal puts greater priority on immediate food demands than it
does on insuring a food supply for later food demands. The time spent in
the short duration displacements is included in feeding time and results in
an unequal feeding rate for dominant and submissive birds. If, in a flock of
two birds, each bird locates food at a rate of X food items per unit search
time, then the flock discovery rate is 2X. If the dominant bird gets a pro-
portion p (where p > 1 - P = q) of all fo<;>d located, then the effective food
discovery rate of the dominant bird is 2pX, and that of the submissive bird
is 2qX, the total remaining 2X.
The time spent in predator detection as a function of flock size can be
determined either empirically or theoretically. Pulliam (1973) derives a
theoretical expression for the probability of predator detection versus flock
size. If T is the time necessary for a predator to make its final uncovered
approach to within striking range of a flock, and (3n is the number of head
cocks per minute for each flock member, the probability that the predator
is detected is given by
where n is flock size. If the bird is feeding alone, PI = 1 - e-{jlT, and the
time spent per day in predator detection is flICT, where C is the time per cock
and T is the length of daylight. If the same bird were to feed in a larger flock,
it could have the same probability of detecting a predator with much less time
spent in predator detection. Thus, if
Pn = 1 - e-n{jn T = PI = 1 - e-{jlT
taking log P n = log PI> we have nfln T = fll T, and solving for fln,
fln = ~
n
Thus, the time spent per individual per day in a flock is only (fllln) CT.
From the foregoing it would appear that the larger the flock the more
time could be spent feeding for every individual in the flock, and thus flocks
would always be advantageous and optimal flock size would be infinite.
This of course is not the case when we consider the previously assumed time
budget priorities. We first consider the case where there is no long-duration
aggression.
F or every bird in a flock, we can partition the time budget into time
spent in predator detection, TD , time spent feeding, TF , and the remaining
time, TR • Thus,
T = TD + TF + TR
For the simple case of two birds, we can calculate T D, T F, and T R. Time
spent in predator detection is simple. Letting a = flICT, then TD = an-I. If a
bird is in the flock (designated by the symbol ~ over the appropriate time
component) TD = a12, and if the bird feeds alone, TD = a. The time spent
feeding is only slightly more difficult to calculate.
Let X be the total number of prey of all types encountered per unit
search time (using the notation of previous sections, X = ~Xt). Let lbe the
average time required to pursue and eat a prey, given that it has been en-
countered [I = (~Xttt)/~Xtl. The total time per prey item eaten is
1 _ 1 + IX
-+t=--
X X
as derived earlier in equation (1). If D is the total number of prey required
to satiate the predator, then the total time spent feeding to satiate the
predator is the product of the total number of prey required and the time
per prey eaten, i.e.,
TF = (1 I X + l) D
This equation applies for a bird feeding alone, but for flocking birds the
dominant bird has an effective food discovery rate of 2pX, and the sub-
missive bird a rate of 2qX. Thus, for a bird in a flock of two birds,
t
F
=( 1 +Xf)D
2pX
for the dominant bird, and
t
F
=(1 2qX
+Xf)D
for the submissive bird. These equations assume that the bird has enough
time to meet its food demands. If there is not enough time, then by the
assumptions on priority in the time budget, the bird spends all of the time
left after predator detection as feeding time, that is
TF = T - Tv
For a flocked bird this becomes T - a/2, and for a lone bird T - a.
We have now specified the time spent in predator detection and the
time spent feeding for the flocked and unflocked situations. Defining the re-
mainder time as the total time available, T, minus the predator detecting
time, Tv, and the feeding time, TF, we have specified an exhaustive and
mutually exclusive time budget for the light hours of a bird's day. Table I
summarizes the time budgets of the dominant and submissive birds for
flocked and unflocked conditions.
Feeding rate, R, is defined as the total food consumption per day,
which is the product of the total time spent feeding and the food items eaten
per unit feeding time. Thus, feeding rate is given by
R = t ( 2pX )
F 1 +Xl
for the dominant bird, and
t ( 2qX )
F I + Xl
for the submissive bird. In each case, R is the product of the total time
spent feeding and the food items eaten per unit time.
Whether or not flocking is advantageous depends on the magnitude of
q. For if p = q = Y2, the submissive bird locates food just as fast as it
would feeding alone, but has more time in which to feed. However, as q ap-
proaches zero it is of obvious advantage to feed alone. Figure 3 illustrates
feeding rate, R, as a function of food encounter rate, X, for the situation
when flocking is advantageous for both the dominant and the submissive
birds. Note that at the value of food encounter rate indicated by Z*, the
it
"'ll
:!.
Q.
:!.
Table I. Time Budgets of Flocked and Lone Birdsa
.sa.
o
Feeding in flock Feeding alone 'S.
§.
Dominant Submissive Dominant Submissive
e:.
~
~
Time in predator {3,CT 0 {3,CT 0 s·
-2-=2 -2-=2 {3,CT= 0 {3,CT = 0 .,
detection (TD) ~
=
Co
;.
Time feeding (TF ) if =
I + Xl)
B=(1 +XT)D +XT)D B= ( - - D B ..,.
T - TD > B . 2pX X
B (I 2qX =c ~;/T)D ~
~
Time feeding (TF ) if T - TD = T - 0/2 T - TD = T - 0/2 T - TD = T - 0 T - TD = T - 0 sa.
T- TD SB Q
a
Remainder Time (TR) T - TD - TF T - TD - iF T - TD - TF T - TD - TF
a
§
~
~.
a Definitions: {3, = head-cocking rate of lone bird; C = time per head cock; T = total time in day; TD = time per day spent in predator detection;
TF = time per day spent feeding; TR = time in day neither spent in predator detection nor feeding; X = rate of encounter of bird with food items
per unit search time; p = proportion of food located by flock which dominant bird eats; q = proportion of food located by flock which sub-
missive bird eats; T = average handling time of a prey item; D = total number of prey required per day. Band 0 are defined by the relations indi-
cated in the table.
~
N
Ul
FEEDING
RATE
(R)
--~----------------~----------------------.x
FOOD ENCOUNTER RATE
Fig. 3. Feeding rate as a function of food encounter rate, when flocking is advantageous for
both submissive and dominant birds, and long duration aggression is not considered.
feeding rate is thereafter the constant value D, regardless of the food

density, and there is thus no longer any advantage (or disadvantage) to feed-
ing in the flock.
N ow consider the effect on the previous model of adding a term into
the time budget for time spent in long-duration aggression. The simplest
assumption is that the dominant bird devotes all its spare time after feeding
and predator detection to chasing the submissive bird. Thus, when food
abundance is low, and it is advantageous to the dominant bird to be in the
flock, it does not chase the submissive bird at all. But once the food abun-
dance is high, the dominant bird no longer gets a short-term advantage
from flocking (and flocking is to its long-term disadvantage), so it spends
all of its spare time chasing the submissive bird. This means that all of the
terms in Table I remain the same for the dominant bird, except that the re-
mainder time, T R , now becomes time spent in long-duration aggression, TA •
The changes in the time budget of the submissive bird are somewhat
more complicated. The time spent in long-duration aggression for the sub-
missive bird is the time during which it is being chased by the dominant
bird. This time takes precedence over time spent feeding, for the bird can
hardly feed when under attack. Table II gives the new time budget for the
submissive bird.
Figure 4 shows feeding rate as a function of food encounter rate when
time spent in long duration aggression is considered in the model. The
curves for the dominant bird, and for either bird feeding alone, are the same
as in Figure 3. The curve for the submissive bird is the same as in Figure 3
until the dominant bird begins to use its spare time in long-duration ag-
Table II. Time Budget for the Suhmissive Bird when Time Spent in Long Duration Aggression
is Added to the Model
Feeding in flock Feeding alone
Time in predator f3,CT/2 = a/2 f3,CT = a

detection (To)
Time in long-duration 1'* (equals time spent o

aggression (TA ) in long-duration
aggression by dominant
bird-given as TR in
Table I)
Time feeding (TF ) if

B= (
I + 2XT) D
Q I -
B= ( - +XT)
X D
.
T-To-TA>B X
Time feeding (TF ) if

T-To-TAS,B
gression. The figure is drawn such that the dominant begins to do so as soon
as it has spare time. This may not be the case, since there is yet a short-
term advantage to the dominant bird until food density reaches Z*. Because
of the complex interaction between long-term and short-term advantages, it
is impossible to say exactly where in this interval the dominant bird begins
to use spare time in this manner. At "ny rate, the essential features of the
D ---------- - --~------_:::::,...,..---------
FEEDING
RATE I
(R) Submissive
in Flock
--~------------~-------~---------+x
T z·
FOOD ENCOUNTER RATE
Fig. 4. Feeding rate as a function of food encounter rate when long duration aggression is
considered.
model are unchanged, regardless of exactly where the change occurs. I feel
that the model is most likely correct as drawn, since the bird is likely to be
able to distinguish "sufficient food" and "insufficient food," but is unlikely
to be able to assess Z* directly.
The food encounter rate, marked T in Figure 4, is the value at which
flocking and feeding alone result in equal feeding rates. T is thus the
threshold food encounter rate above which flocking is no longer advanta-
geous because the aggressive behavior of the dominant bird could cause the
submissive bird to leave the flock.
In this section, I have shown that conditions under which flocking is
advantageous can be deduced from simple assumptions concerning the time
budget priorities of individuals. The conclusions reached are entirely de-
pendent upon the accuracy of the assumptions. However, if these assump-
tions are shown to be false, a proper reordering of time budget priorities can
be used to deduce a new set of predictions concerning optimal social be-
havior. It is the deductive methodology which is important, not the example
used.
I have only treated the cases of feeding alone and feeding in a flock of
size two. The reason was purely practical: the algebra becomes much more
complicated as larger groups are considered. However, the same technique
can be used to consider flocks of any size. More importantly, one major
prediction of the models remains unchanged when larger flocks are
considered. This prediction is that flocking is less advantageous when food
is abundant, and its corollary is that flocks are largest when food is scarcest
and food demand is highest. For an empirical demonstration of this predic-
tion see P\llliam et al. (1974).
VI. SPECULATIONS AND DISCUSSION
In this paper, I have outlined procedures for predicting optimal diets,

habitats, and sociality. How do these relate to community structure? The
dynamics of community processes must ultimately depend on the interac-
tions of individuals with their environment. The behavior of an individual
ultimately determines how that individual changes its environment, and the
sum of all of the behavioral responses of all of the individuals largely de-
termines the dynamics of biotic communities. For example, the movement
patterns of a predator determine the frequency of encounter between that
predator and its prey. The predator's behavior upon encountering a
potential prey determines the effect of that individual predator on that indi-
vidual prey. The sum of all such encounters between predators and prey
determines the diets of the predators, the competitive interactions between

predators, and the dynamics of predator-prey systems. In a similar manner,
habitat choice, largely as mediated by social behavior, may determine the
spatial patterns of predation intensity. Below, I briefly outline one way in
which the principle of optimal behavior can be used to deduce community
dynamics.
Consider the classical Volterra predator-prey equations. If N
represents prey population size, and P represents predator population size,
the predator-prey dynamics are represented by the following two equations:
dN = rN _ aNP _ CN2
dt
and
dP = bNP _ dN (7)
dt
In these equations, r is the intrinsic rate of increase of the prey population,
c represents a crowding factor for the prey population, and d is the death
rate of the predator population. The symbol a can be thought of as the
constant relation of encounter rate (determined by the product NP) to
actual predation rate, and b the same as a, with an additional component
describing the efficiency with which the predator turns captured prey into
new predators. Thus b can be thought of as equivalent to a product e x a,
where e is often near 10% in natural biological systems.
If we consider a as a measure- of the functional response (Holling,
1966) of the predator to prey d!!nsity, rather than as a fixed constant, a be-
comes a variable which is proportional to the inverse of the time required
per prey item (equation 1). For only one prey type this becomes
(XN
a=--~ (8)
1 + tN
where (X is a proportionality constant, and N is the product of X in equation
(1) and the area covered by the prey popUlation in question. Substituting
equation (8) back into equation (7), one obtains
dN = rN _ (XN2P _ cN2
dt 1 + tN
dP = EN 2P _ dN
dt 1 + tN
where E = (Xe. The properties of uniqueness and stability of solutions to
these equations are beyond the scope of the present paper.
An even more interesting way to modify the basic Volterra equations is

to look at multiple-prey:-multiple-predator systems where the values of a
are determined by the optimization criteria discussed earlier in this paper.
Let aij * be the optimal value of a when predator j is feeding on prey i. and let
biJ* = eaij*. The following system of equations applies in the situation of m
prey species and n predator species:
dN =
dt 1
r 1N 1 - C1
N 12 -
"\"
~ aij*NIPj
J
Because of the threshold nature of the values of aij * and bi / , these equa-
tions are very difficult to approach analytically, but they can be readily
simulated on a computer. Simulation can be used to look at properties of
coexistence of both predators and prey, and by allowing the values of Ci to
approach zero, one can explore the ability of a switching predator to con-
trol prey populations.
Most models of community dynamics discussed in the literature to date
have considered only the situation where habitat utilization is fixed and ani-
mals compete for resources within habitats. However, there is growing evi-
dence (Rosenzwieg and Sterner, 1970; Pulliam and Enders, 1971) that,
within habitats, similar species may have identical diets, and that competi-
tion results from a very complex interaction between food and habitat
utilizations. Again, the situation could be simulated by having two or more
habitats differing in food resources, and by having several species of preda-
tors feeding on those food resources according to specified optimization
criteria.
The types of models of sociality developed in this paper indicate the
importance of social structure in habitat utilization and, thus, in community
dynamics. There is plenty of food for thought in the models concerning how
prey production, and environmental factors that influence predator food
requirements, such as temperature, might direct the course of competition

and the likelihood of coexistence. The way in which social behavior mediates
the rate of disappearance of food in different habitats has hardly been
explored, but could be of obvious importance to community structure and
dynamics.
New theories relating to community dynamics and the social structure
of populations will be well worth exploring if we discover that animals do
indeed use optimization criteria for food selection. However, it would be a
fruitless task to clutter more books and journals with a theory of com-
munity dynamics based on optimization principles if animals do not behave
according to such criteria. I hope'that this paper will help to demonstrate
the tremendous potential for theoretical ecology in the use of optimization
principles, but at the same time I plea to field-oriented biologists that there
is an urgent need to test the applicability of simple optimization criteria to
the behavior of animals in the real world. This should be done before com-
plex models of community structure and dynamics, based on optimization
principles, are pursued further.
VII. REFERENCES
Cody, M. L. (1971). Finch flocks in the Mohave Desert. Theoret. Pop. BioI. 2: 141-158.
Cody, M. L. (1974). Optimization in ecology. Science 183:1156-1164.
Dawkins, R., and Dawkins, M. (1973). Decisions and the uncertainty of behavior. Behav.
45:83-104.
Downhower, J. F., and Armitage, K. B. (1971). The Yellow-bellied Marmot and the evolution
of polygamy. Amer. Natur. 105:355-370.
Fretwell, S. D. (1972). Populations in a Seasonal Environment, Monographs in Population
Biology 5, Princeton Univ. Press, Princeton, New Jersey.
Gadgil, M. (1972). The function of communal roosts: Relevance of mixed roosts. Ibis 114:531-
533.
Hamilton, W. D. (1964). The genetical evolution of social behavior. J. Theor. Bioi. 7:1-52.
Holling, C. S. (1966). The functional response of invertebrate predators to prey density. Mem.
Entomol. Soc. Canada 48:1-86.
Intrilligator, M. D. (1971). Mathematical Optimization and Economic Theory, Prentice-Hall,
Inc., Englewood Cliffs, New Jersey.
Levins, R. (1968). Evolution in ChOlJging Environments, Monographs in Population Biology 2,
Princeton Univ. Press, Princeton, New Jersey.
May, R. (1973). Stability and Diversity in Model Ecosystems, Monographs in Population
Biology 6, Princeton Univ. Press, Princeton, New Jersey.
MacArthur, R. H. (1973). Geographical Ecology, Harper-Row, New York.
MacArthur, R. H., and Pianka, E. (1966). On the optimal use of a patchy habitat. A mer. Natur.
100:603-609.
Pulliam, H. R. (1973). On the advantages of flocking. J. Theor. Bioi. 38:419-422.
Pulliam, H. R. (1974). On the theory of optimal diets. A mer. Natur. 108:59-74.
Pulliam, H. R. (1975). Diet optimization with nutrient constraints. A mer. Natur. 108:765-768.
Pulliam, H. R., Anderson, K. A., Misztal, A., and Moore, N. (1974). Temperature dependent
social behavior in juncos. Ibis 116:360-364.
Pulliam, H. R., and Enders, F. E. (1971). The feeding ecology of five sympatric finch species.
Ecology 52:557-566.
Rosenzweig, M. L., and Sterner, P. N. (1970). Population ecology of desert rodent communities:
body size and seed-husking as bases for heteromyid coexistence. Ecology 51 :217-224.
Schoener, T. W. (1971). Theory of feeding strategies. Ann. Review oj Ecol. and Syst. 2:369-
404.
Whittle, P. (1971). Optimization with Constraints. Wiley Series in Probability and
Mathematical Statistics-Applied, John Wiley and Sons, Inc., New York.
INDEX
Acrobatics, 191 Arousal

ACTH, 268-270, 273 dermed,l72
Actualgenese, 125 level of, 172
Adolescence, 8, 51, 52 paradigm, 172
in primate, 32-50 play theory, 171-172
Adrenalin, 270, 271 Associative chain theory, 136
Adulthood, in primate, 32-50 Attack,233
African wild dog, 207 behavior reinforcement, 274
Aggression, 231-298, see also Attack by success, 274
and androgen, 263-268, 274 conditioned, 236-238
conditioned, classically, 281 classically, 236
displacement of, 276 operantly, 236
factors leading to, 255, 257 consequence of, 237
fear-induced,281 when escape is blocked, 279
instrumental, 235, 281 and familiarity, 240-241
intermale, 282 initial, 233
irritable, 238-250, 281 in interaction situation, 250-251
in man, 234 and movement, 277
maternal, 265, 282 pain-induced,252
in mouse, 237 properties, 275-280
predatory, 233 reinforced, negatively, 238
and sex, 262 unconditioned, 238-250
territorial, 282 Attention-switching hypothesis, 249
types of, 236, 281 Attractivity, 113
Ailuropoda melanoleuca, see Giant Avoidance, 234
panda
Alligator alligator, see Alligator, Bandicoot, Southern shortnosed, 193
American Baboon,4,5,25,26,32,38
Alligator, American, 209 Bat, Vampire, 193
Altruism, 312 Beagle, 179
Anatomy and movement, 81-84 Bear
Androgen and aggression, 263-268, black, 175, 192
274 brown, 206
Anger, 233, 252, 258 polar, 121
Animal Begriff, injunktiver, as a concept,
exercise, 189-219 307
play, see Play Behavior
training, physical, 189-219 "adequate", 7
333
334 Index
Behavior (cont'd) Cavies, 193

communication, interpersonal, 135-163 Cercopithecus aethiops, see Vervet monkey
as a dependent variable, 56-57 Chacma baboon, see Baboon
excitement, 39 Chaffmch, 205
of female in general, 45-46 Chain
homosexual, 40 associative, theory of, 136
idiosyncratic, 45 kinetic, 115, 118
imitation of adults, 30 stepwise, 117
interference, 29 Cheetah,196
interpersonal communication, 135- 163 Chick, domestic, 194,202,205
language as, 141 Chimpanzee,4,26,38,39,192,202,203,
maternal, distorted by isolation, 49 211-212,242
maximizes fitness, 312 Choz-choz, 207
"normal", 7 Chord of movement, 88-89
notification, 29 Chronography of interaction, 142
optimal, principle of, 311-332 Civet, 121
organization, multilevel, 145 Civetticitis civetta, see Civet
structural, 145 Clethrionomys, 250
and rhythm, biological, 135-163 Communication, 128
sexual, 6, 249 interpersonal,135-163
male-male, 40-41 Community
sociosexual, 1-67 ecology of, 312
species-specific stereotyped, 82 matrix, 312
Bestiality, 42 theory of, 311-332
Bird, play behavior, 205 Commutativity, 120-121
Black bear, 175 Concatenate of movement, 88-91
Blackbird, 202, 205 Condurance, 128
Bonnet monkey, 38 Connectedness of movement, 84-119
Bos taurus, see Cattle anatomical, 114-119
Bottlenose dolphin, 212 spatiotemporal,89-114
Brain function temporal, 85-89
reflex arc theory, 136 Consorting, 6,37,46
time duration, 139 Contact
Brain wave, 137 establishment, 117
autocorrelation analysis, 158 maintenance, 128
Brown bear, 206 mother-baby, 128
Bumblebee, 208 release, 11 7
Burmese red jungle fowl, 169 Context, 29
Bushmen conversation, 150, 152, 153, 155 Convergence, routes of, 69
Conversation
Call of animals, 300 between two persons, 144
Canis sp. at play, 192 Copulation,S
Canis aureus,see Jackal age of female, 46
C. familiaris, see Beagle cross-species, 41-42
C. latrans, see Coyote in female, 46
C. lupus, see Wolf age of, 46
Capra hircus, see Goat frequency, 35
Cat, domestic, 192,205 mature pattern in male, 36-37
at play, 202 Corticosterone, 268
Catecholamine, 270-271 Cotumix cotumix cotumix, see Quail,
Cattle, domestic, 192 migratory
Index 335
Coyote, play of, 179 Ejaculation, 5,41

Crocuta crocuta, see Hyena, spotted Elephant, African, 193
Cross-referencing, 126 Elephant seal, Northern, 192
Emotionality, 234
Dama dama, see Fallow deer Endurance
Decision point, 126 exercise, 194-195
Decision process, 235, 236 ofmusc1e, 191
1.258-275 of whole body, 193
factors affecting, 258-275 Environment, familiar, 240-244
II, unfamiliar, 244-245
factors affecting, 275-280 Escape, 234
attack properties, 275-278 blocked, leads to attack, 279
target properties, 275-278 Eshkol-Wachmann movement notation,
Deer, black-tailed, 168, 173, 206 69-75
fallow, 171 sphere, 74, 75, 79, 80
red,l92 Eskimo
Defeat reinforces fear, 274 dialogue, 157
Defecation in novel surroundings, 245 movements, 146
Deference gesture, 51 Estradiol, 266
Degus, 193,202 Estrogen, 266
Depravity idea, 226 Estrus, 6
Diet E-W, see Eshkol- Wachmann
optimal, model of, 314-318 Excitement behavior, 39
and prey, 314-318 Exercise
caloric value, 214-218 of animal, 190-191
search time, 314-318 deprivation, 205-207
Discrepancy between observed and ex- dynamic, 194
pected event, 258-261 endurance, 198
Displacement activity, 250 isometric, 199
Display locomotor, 199
. choreography of, 69-134 muscular, 190
fixity of, 122 overload, 198
directedness of, 120 physiology, 191-197, 200
motor-, 124 play, 191, 199,201
Distance pattern in birds, 194
critical, 279 in mammals, 192-193
of flight, 279 wheel running, 200
intrusion, individual, 239-240, 257 whole-body endurance, 193
territory concept, 240-244 work overload, 193
Distinctiveness, 299-310 Exhaustion
defined,302 of cheetah after unsuccessful hunt, 196
geometric model of, 302-307 Expectancy 9freward, 261
Distribution Experience, early social, 271-273
normal, 304 Eye contact, 139
uniform, 304 duration of, 140
Dog and prostitutes, 139
African wild, 207
beagle, 179 Fallow deer, 171
Dominance, 6,233 Familiarity
and mating, 34-36 and attack, 240, 241
Duckling, 261 of object in unfamiliar place, 245
336 Index
Fear behavior, 231-298 Genital (cont'd)

conditioned attack, 236-238 pink penis, 18
factors leading to, 255, 257 touching, 15-21
fleeing, 278-279 Giant panda, 192
freezing, 278-279 Goat, 168, 206
in interaction situation, 250-251 Gorilla, 5, 202, 203
reinforcement, 274 Grey-lag goose, egg-retrieving behavior, 110
response, temporal sequence of, 279 Ground squirrel, Columbian, 169
types of, 278-279 Gymnotus carapo, 241
unconditioned attack, 238-250
Feedback Habitat selection, 312, 318-321
oscillation, nervous, 137 macro-,318-321
sensory, 123 micro-, 318-319
Feeding Hamadrya, 25
rate of, 324-328 Hamster, golden, 122
strategy of, 313 Harem leader, 27
Felis catus, see Cat Head movement trajectory, 83
Felis serval, see Serval Heart, overloading, 195
Fighting, 233 Hedgehog, 193,205
Fishing cat, 208 Helogale undulata, see Mongoose
Fixation, 92, 121 Hippopotamus, 193
Fleeing, 234, 278-279 Homo Sapiens, see Humans
Flight distance, 279 Homosexuality in primates, 40-41
Flocking, 321-328 Hormone, 24-25,32-34,56,262-271
of sparrows, 322 and aggression, 262
Food competition, 275-276, 279-280 male, 32-34
Food deprivation, 280 sex-, 262-268
as thwarting, 247 Humans at play, 192,202
Footclasp, 12 Hyaena hyaena, see Hyena
double, 48 Hyena, spotted, 169, 192
Foraging striped, 121
strategy, 314-318
time, 314 Immediacy, transitory, 125
Freezing (motion), 278-279 Incest, 33
Frigate bird, 194 brother-sister matings, 41
Fringilla coelebs, see Chaffmch mother-son matings, 41
Frustration, 235, 246, 251-254 in primates, 41-42
leads to attack, 261 Individuality, 299-310
nomeward, 257 distinctiveness, 299-310
and pain are additive, 251 geometric model of, 302-307
index, 301
Gabra tribe of East Africa, 155 Inertia,social,263,274
dialogue, 155 Infancy
Gallus gallus spadiceus, see Burmese red mounting, 9-12
jungle fowl in primate, 9-25
Genet, 121 thrusting, 9-12
Genetta tigrina, see Genet Information processing, 142-144
Genital Inhibition, social
grooming, 16 and mating, 33
pink bottom, 15, 18 Injunktive Begriffe, see Begriff
Index 337
Innate depravity idea, 226 Male

Interaction, 141, 170 group, 32-34
chronography, 142 change between, 32-34
dimension, human, temporal, 144-159 status within, 32-34
human, temporal dimension of, 144-159 hormones, 32-34
mammalian -male sexual behavior, 4041
choreography of display, 69-134 March military, 147
motor processes, homeostatic, 69-134 Maring tribe of New Guinea, 147, 154,
Interference behavior, 29 158
Intromission, 12 Markov model of nonplay, 174
and thrusting, 12 of play, 174
Isolation Marmoset monkey, 118
and behavior, idiosyncratic, 45 Marmot, 193
early social, 4445,273 Marmota monax, see Woodchuck
leads to fighting, 273 Marsupial, 208,209
Masturbation, 18
Jackal,126 Mating, 6
behavior,73,77,82,90,102,111 and dominance, 34-36
kinetic lines of, 111 and rank, 35
Japanese monkey, 38,47 selection of mate, 312
Juvenile and social inhibition, 33
contexts, 29 Maturity, 36
partner selection, 26-27 Meerkat, 121
posturing, 27-29 Metacommunication, 177
in primate, 25-32 Microtus agrestis, see Vole
seasonal pattern, 26 Mirounga augustirostris, see Seals
sequential mounting, 25-26 Mongoose, African dwarf, 178, 192
Modulation, 98
Kea, 194 Morphine withdrawal in rat, 248-249
Kestrel,194 leads to fighting, 248-249
Kidnap of female, 37 Morphology, adaptive, 312
Komodo monitor, 208, 209 Mother-infant relation, 53
Motivation, strength of, 235, 261-262
Motor action
Language, 70, 71 homeostatic, 69-134
as rhythmic motor behavior, 141 temporal sequencing, 173
Langur, 5, 33 Mounting, 4-5,9-12
Limb hierarchy female-female, 4648
heavy and light, 115, 123 male-male, 37-38
mechanical, 115 rear-oriented, 11
Lion, 169 for reassurance, 31
Locomotion for reunion, 31
age of onset, 203 sequential, 25-26
and play, 201-205 Movement
Luteinizing hormone, 265 actual elementary, defmed, 76
Lycaon pictus, see Dog, African wild descriptions required, 76-81
and anatomy, 81-84
Macaca arctoides, see Stumptail monkey chords of, 88-89
Macaca mulatta, see Rhesus monkey commutative, 120
Macropus parryi, see Whiptail wallaby concatenate of, 88-89
338 Index
Movement (cont'd) Partner (cont'd)

connectedness, 84-119 selection, 13-15, 26-27
directedness of, 89 willing, importance of, 24
Eshkol-Wachmann notation, 69-75 Performance, musical, 143
flxed, 78 Peromyscus, 242
flow, 78 Pigeon, 118
locomotor-rotational, 175, 191 Pituitary adrenocortical system, 268-270
as multidimensional process, 72-76 Plasticity, physiological, 196
pattern, 312 Play of animals, 165-187,189-219
slicing of, 76-84 age dependence, 199-205
stimulus for attack, 277 scheduling, 200
study of, 69-134 in the young, 199-205
trajectory, 78, 83, 84 allocation of energy, 204
Mustela putorius, see Polecat of time, 204
approach-withdrawal, 199
Neck movement trajectory, 84 as arousal, 171-172
Net, nervous, heterarchy of, 115 behavior, 211
Nomeward, frustrative, 246 defmed, 166, 199
Noradrenalin, 270, 271 in captive adults, 206
Northern elephant seal, 192 chases, 205
Notiflcation, 15 and childhood, 168
behavior, 29 concept of, 181
defmed, 166, 199
Observer deprivation, 168, 205-207
both sexes required, 58 diversive, 210-212
generalization by, 59 and exercise, physical, 190-191
interpretation by, 59 with familiar object, 207-208
maturity of, 58 intention, communicated, 177-181
Octodon degus, see Degus locomotor, 191, 199
Octodontomys gliroides, 202, 207 manipulative, 209-210
Odocoileus hemonius, see Deer, black tailed mood, 178
Odor in male mouse, 241 pant, 178
Optimal behavior, principle of, 311-332 parent-offspring, 200
Organization, hierarchy of, 140 patterns in mammals, 192-193
Orgasm,S, 13 is it pleasurable? 175-177
by female, 48-49 poikilotherms do not play, 191
grunts during, 49 as practice, 168-171
Overdetermination upon release from confmement, 206
of actual movements, 103 rough-and-tumble, 199
of symptoms, 102 signal, 177-181
social, 165, 172-177,210
Pain, 238-239 characterization, 172-177
acute, 252 soliciting action, 175
elicits flghting, 239, 252, 254, 256, 257 as surplus energy, 167-168
Pan troglodytes, see Chimpanzee Schiller-Spencer surplus energy
Pant hypothesis, 167
during exertion, 178 theories of, 167-172
during play, 178 what is it? 166, 199
Pantera leo massaicus, see Lion Poikilotherms do not play, 191, 208
Partner Polar bear, see Bear
preference, 13 Polecat, 192
Index 339
Population ecology, 312 Rat killing mouse, 233

Posturing, 27-29 Raven, 194
Practice theory of play, 168 Rearing
Predation, 233 environment for
Predator, 314-318 physical, 55
-prey system, 314-318, 329, 330 social,55
multiple systems, 33 by human hand, 43
Volterra equations, 329 recovery from restricted, 44
Presentation, 4 restricted, 21-24, 30-31, 43-45, 54
see Presenting of female, 49-50
Presenting of male, 43-45
by infants, 17 Receptivity, 113
between females, 48, 50 Reflex
by mothers, 10 arc theory, 136
patterns, 4, 15-21 conditioned in dog, 139
pink bottom, 15, 18 state-dependent, 125
is politeness, 48 Reinforcement schedule law, 247-248,
rump lifting, 16 257
Prey Reward expectancy, 261
caloric value of, 316 Rhesus monkey, 25, 32, 172, 192, 202,
and diet, 314-318 203
maximizing caloric value of, 316 consort pair, 146, 148
minimizing total time per, 315 play, 199
total time per, 314 Rhyme and rhythm, 140
Primate, 1-67 Rhythm, biological, 135-163
adolescence, 32-50 alpha-,137
adulthood, 32-50 and behavior, 135-163
infancy, 9-25 of brain, 136-138
juvenile period, 25-32 external, 136
sociosexual development, 1-67 as information, 135-163
embrace, 2 internal, 136
hate, 1 Riding, dorsal, 10, 11
love, 1,2 Robin, 242
posturing, 3 Rocking behavior of mentally retarded, 138
copulatory,S Route of convergence, 69, 93, 123, 124
Prionailurus vive"irnus, see Fishing cat connectedness, 101-119
Proceptivity, 113 defIned,92
Prostitute and eye contact, duration of, fIxation, 92
139-140 interactional,99-114
Puberty, onset of, 32 locomotary, 93-99
of movements, 119
Quail,267 togetherness, 127-128
aggressive, 261
gamblers, 240 Saliva as conditioned reflex in dog, 139
migratory, 299 Sarcophilus hamsii, see Tasmanian devil
Quelea, 240, 265 Schedule
fIxed interval, 248
Rage, 233 fixed time, 248
Ramp, see Feedback reinforcement, 247
Rank and mating, 35 Selection
Rat, Norway, 193, 202 of community, 312
340 Index
Selection (cont'd) Tasmanian devil, 82, 193

of habitat, 312 contact management, 108, 109
of mate, 312 point, 110
Sequence behavioral, see z-Score kinetic field, 112-114
Serial activation, 137 postures, 116-117
ordering, 137 Tension reduction, 15
Serval,208 Tern, 194
Shock, electric Territory, 240-244, 257
leads to fighting, 239 Testosterone, 34, 265, 266, 267
Siamese fighting fish, 237,271 Thalarctos maritimus, see Bear, polar
Sign, 300 Thermoregulation by play, 201-205
Signal,300 Threat, 233
and mood, 171 Thrusting, 9-12
Simultaneity, 85-88 Thwarting, 246-247,257
concept, 85-86 food competition as, 247
structure, 85-88 Timber wolf, 207
Sociality, optimal, model of, 321-328, 330 Training, physical
Sociosexual, defmition of, 34 of animal, 199-219
Space hypothesis, 197-209
concept, 113-114 play should train muscles, 197
biological, 113 responses, 191-197
Kantian, 114 types, 198
types of, 80-81 Turdus merula, see Blackbird
Sparrow flocking,321-322 Tursiops truncatus, see Bottlenose dolphin
Speech
goal of, 148 Ursus american us, see Bear, black
rhythms of, 160-162 U. arctos, see Bear, brown
storytelling, 149, 153, 155
Vampire bat, 193
Spermophilus columbianus columbian us,
Varanus komodensis, see Komddo monitor
see Ground squirrel
Vervet monkey, 173,211
Squirrel monkey, 203, 206, 258
Vole, 178, 250, 266, 278
Sterility, adolescent, of female, 46
Volterra predator-prey equations, 329
Stickleback, three-spined, 241, 245, 265
Vombatus ursin us, see Wombat
Storytelling, 149, 153, 155
Strength, dynamic, 194 Warm-up, 207
isometric, 194 Warthog, 192
Stumptail monkey, 18,47, 172 Waterbuck, 242
presenting, 19-21 Wheel running, 200
Stuttering, 158-159 Whiptail wallaby, 193, 209
Subadulthood, see Adolescence Wolf, 179, 207
Submission, 6, 7 Wombat, 121, 193
Suricata suricatta, see Meerkat Wood chuck, 206
Surplus energy theory of play, 167-168 Woodpecker, downy, 194
hairy, 194
Tapir, 193
Target Yanomamo tribe of Orinoco region, 156,158
properties, 275-278 shouting as dialogue, 157
proximity to attacker, 276
size, 277 z-Score, 174

Perspectives in Ethology Volume - VVAA PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Perspectives in Ethology Volume - VVAA PDF

Uploaded by

Copyright:

Available Formats

PERSPECTIVES

PLENUM PRESS • NEW YORK AND LONDON

©1976 Plenum Press, New York

F. Incest and Cross-Species Copulation ........ . . . . . . . . . . .. 41

III. Other Rhythms .......................................... 138

VI. Aspects of Play Which the Physical Training Hypothesis Alone

VIII. External Factors Affecting Decision Process 2 ............... 275

INDEX ...................................................... 333

The development of sociosexual patterns in nonhuman primates is traced

Gestures of love and hate, or affectional and agonistic patterns, have

Sociosexual behaviors such as mounting, presenting, genital touching,

Sociosexual: Behavior elements or patterns that are usually present in

tern (pausing after a period of thrusting or genital stimulation) even though

A hits Band B never or seldom reciprocates, then A is dominant to B or B

Harlow has described stages in heterosexual sexual development and

This period is characterized. by initially complete dependence on a

A. Thrusting and Mounting

The newborn monkey or ape is in almost constant contact with the

Dorsal riding (Fig. 4) in nonhuman primates may facilitate the

needs to be followed by more data is that of the onset difference seen

While thrusting, mounting, and intromission become linked for most

More recent observations do not contradict Zuckerman's conclusion

males before they did females, and the mount-embrace-touching relation-

E. Presentation Patterns and Genital Touching

Most infants are of course first presented to by their mothers as a

contacting and climbing attempts. This hypothesis could be tested with

Skolnikoff (l97Ia, 1974a,b), (Fig. 6). This is nowhere so clear as in chim-

Fig. 6. Unusual infant behavior. The infant stumptail

is a gesture reported to appear quite early in the terrestrial, highly social

reference is made to the "normal" development of monkeys reared with

Primates reared in restricted conditions show delayed and often dis-

normal mothering is not enough to produce socially adequate offspring"

The statement that mothers can "substitute" for peers reflects

there is obviously a great difference between a partner who cringes or

We might expect this to be true of same-age infants in a small space

The influence of hormones during the period from birth to puberty is

not seem to be necessary for the expression of sociosexual patterns in

IV. JUVENILE PERIOD

Most of the elements of the copulatory pattern have been integrated by

Rhesus and Japanese macaques have relatively strict breeding seasons:

Posturing becomes more stereotyped in juvenile primates, and most

panied by bites on the shoulders or neck. While mounts on females tend to

regularly accompanies mounts by juveniles. Bertrand (1969) mentions teeth

The most frequent context for juvenile sociosexual patterns seems to be

F. Interference and Imitation

extent, gorillas (Fossey, personal communication) seem attracted to and

Peer only raised juvenile males, however, show inadequacies when

The juvenile period seems to be a time for further integration of the

v. ADOLESCENCE AND ADULTHOOD

The onset of puberty implies a surge of pituitary and gonadal hormones

A. Male Group Change, Status, and Hormones

Adolescent male macaques and baboons seem to be particularly mo-

or a number of, older, more dominant males), to the disinterest or avoi-

B. Dominance and Mating

The evidence for a positive relationship between dominance rank and

so it is not always clear whether exclusive access to a female midcycle has

C. Mature Copulatory Patterns in Males

Maturity is hard to define in terms of age for most primates. Mature

teristic "lip flip" in his courtship which disappeared when he became

D. Other Adult Sociosexual Patterns

In addition to the increasingly stable and individually characteristic

(a) (b) (c)

occurs as a prelude to other interactions in greeting, reunion, and "tense"