Professional Documents
Culture Documents
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
Department of Zoology
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
Department of Zoology
Duke University
Durham, North Carolina
Volume 2
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
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
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
Chapter 6
MY DESCENT FROM THE MONKEy ....................... 221
Nicholas S. Thompson
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
Chapter 8
ON INDIVIDUALITY: THE CONSTITUENTS OF
DISTINCTIVENESS
Wolfgang M. Schleidt
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
SOCIOSEXUAL DEVELOPMENT IN
PRIMATES
J. Hanby
MRC Unit on the Development and Integration oj Behavior
Madingley. Cambridgeshire. u.K.
I. ABSTRACT
II. INTRODUCTION
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
A. Definitions
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).
Sociosexual Development in Primates 5
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
B. Age Categories
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.
Sociosexual Development in Primates 9
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
(a)
'-----""~~~ (b)
Fig. 3. Presenting by mothers and boarding mounts by infants. (a) Chimpanzees (from
Goodall, 1968b). (b) Japanese monkeys (from life).
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
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
Sociosexual Development in Primates 13
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
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)
(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
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
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
G. Hormones
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.
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
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
Sociosexual Development in Primates 27
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
(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).
Sociosexual Development in Primates 29
E. Contexts
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
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.
Sociosexual Development in Primates 31
H. Conclusion
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.
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
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
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.
Sociosexual Development in Primates 39
10
"25
5_
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
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).
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
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.
Sociosexual Development in Primates 43
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.
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
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
(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
K. Female Presenting
L. Female Orgasm
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.
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
Sociosexual Development in Primates 53
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
Sociosexual Development in Primates 55
need for reassuring contact at all ages, may alleviate some of the strains in-
duced by modern society.
B. Research Suggestions
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.
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
7. Observations: Observers
and sociosexual experience do not affect his or her observations and in-
terpretations.
C. Conclusion
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.
Sociosexual Development in Primates 61
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.
Sociosexual Development in Primates 63
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.
Sociosexual Development in Primates 65
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.
Sociosexual Development in Primates 67
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
I. ABSTRACT
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
II. INTRODUCTION
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.
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,
o~
, --- ---
" '-
---
----,---
"-
..........-,-- - /
-------~-
I
/
/'
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
Homeostatic Motor Processes in Mammalian Interactions 75
@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
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
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
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
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.
A. Temporal Connectedness
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
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
Fig. II. A stick figure of a side view of human locomotion. Reproduced from Bern-
shtein (1940).
Homeostatic Motor Processes in Mammalian Interactions 89
B. Spatiotemporal Connectedness
[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
fo"'" [6 -1' [8 ~~
[i] +
[0)
[bJ [~ '" (oj
[~ " I-- [g ~
(oj
[9' ,. [8
[0]
"' ...... [~
\&J
XI
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
ri-
[0] I
-
J
XIV
--,-
[t] -" -~r~-~· ~
~,-
,~
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
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.
~ll:1 IIII~I
Head
Bodywise
Neck
Absolute Head
('1 I I I I 1('+)1
III IV
~ ~ ~.;
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°.
(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
III VI VII
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
Homeostatic Motor Processes in Mammalian Interactions 97
VII VIII
IX XI XII
XIII XIV
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
Homeostatic Motor Processes in Mammalian Interactions 99
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
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
~
""_, ;, f' ,
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
Homeostatic Motor Processes in Mammalian Interactions 103
dH Ibl
00'
O~ ." .56
I~ 1-16
06' ...
~
076
Hj] ~
OBO ... 112
[0]
116
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.
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
Homeostatic Motor Processes in Mammalian Interactions 107
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.
Homeostatic Motor Processes in Mammalian Interactions 109
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-
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
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
212
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
5 Dr. Meir Berger, M.D., Mental health clinic, Ramat Chen, Israel.
Homeostatic Motor Processes in Mammalian Interactions 119
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
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.]
Homeostatic Motor Processes in Mammalian Interactions 127
VIII. SUMMARY
IX. ACKNOWLEDGMENTS
• Medical School, Tel-Aviv University and Mental Health Clinic, Ramat Chen, Israel.
130 lIan Golani
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.
Homeostatic Motor Processes in Mammalian Interactions 131
* 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.
Homeostatic Motor Processes in Mammalian Interactions 133
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
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.
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)
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.
Biological Rhythms as Information Channels in Communication Behavior 147
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
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
,-
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
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)
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
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
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.
VII. REFERENCES
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
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
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
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.
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.
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
• 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.
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.
Animal Play: Problems and Perspectives 173
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.
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)
Table I. The Duration of Three Acts (seconds) in Play and Agonistic Contexts in
New England Canids(Eastern Coyotes)
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" !
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
\()'i"...
",\'/~ ATMOSPHERE"
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).
Animal Play: Problems and Perspectives 179
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:
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.
11 There are also some differences, as yet unexplainable, in the sequencing of actions during un-
solicited bouts.
182 Marc Bekoff
VI. CONCLUSION
VII. ACKNOWLEDGMENTS
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.
Animal Play: Problems and Perspectives 185
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
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
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.
(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
Exercise, Play, and Physical Training in Animals 193
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
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.
2. Types oj Training
B. Age Dependence
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
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
Locomotion begins
Species (walking or flying) Play begins
, ,,
,,
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
BorC
&
~
~
~
~
~
\I)
~~ B
1i;~
~ ~
GCb
.... ~
~,
~Cb
~.~
Cb"::::
Il:l .....
~
t::: C
.~
~
ts
~
:.5?
-..;;
oq:
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
D. Warmup
A. Manipulative Play
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
C. Diversive Play
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.
VIII. ACKNOWLEDGMENTS
IX. REFERENCES
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-
Exercise, Play, and Physical Training in Animals 215
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.
Exercise, Play, and Physical Training in Animals 217
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.
Exercise, Play, and Physical Training in Animals 219
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
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
224 Nicholas S. Thompson
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
My Descent from the Monkey 225
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
226 Nicholas S. Thompson
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
My Descent from the Monkey 227
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-
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
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-
Aggression and Fear in Vertebrates 235
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
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.
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
Aggression and Fear in Vertebrates 243
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
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.
F. Frustrative Nonreward
G. Thwarting
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).
I. Other Situations
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.
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 +
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
Aggression and Fear in Vertebrates 257
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
.
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
\ ~
'- ~./
/
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
/
/
/
/
>-
u A
z
w
o
z
W
I-
W
III /
Z /
oCL /
/
/
III ;'
/'
W ./
cr
DISCREPANCY MAGNITUDE
:s::
U
<{
l-
I-
<{
lJ...
o
>-
I-
...J
CO
<{
CO
ocr
CL
~
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.
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.
C. Hormonal Factors
1. Sex Hormones
Sex showing
f_ '.thor and date Species Type of aggression more aggression
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
Type of
Author and date Species aggression Main finding
]
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
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
Aggression and Fear in Vertebrates 267
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
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).
3. Catecholamines
D. Experiential Factors
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-
Aggression and Fear in Vertebrates 273
2. Social Isolation
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
Aggression and Fear in Vertebrates 275
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.
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.
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).
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
Aggression and Fear in Vertebrates 281
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.
Aggression and Fear in Vertebrates 285
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
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
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.
Aggression and Fear in Vertebrates 291
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.
Aggression and Fear in Vertebrates 293
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.
Aggression and Fear in Vertebrates 295
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.
Aggression and Fear in Vertebrates 297
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
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
II. DEFINITIONS
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)
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
On Individuality: The Constituents of Distinctiveness 303
A. Uniform Distribution
B. Normal Distribution
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.
306 Wolfgang M. Schleidt
C. General Considerations
termine how the members of a community share the available "space for
encoding" in their particular environment.
IV. DISCUSSION
V. SUMMARY
VI. ACKNOWLEDGMENTS
VII. REFERENCES
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
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
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)
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
The Principle of Optimal Behavior and the Theory of Communities 315
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.
316 H. Ronald Pulliam
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
The Principle of Optimal Behavior and the Theory of Communities 317
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.
318 H. Ronald Pulliam
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
320 H. Ronald Pulliam
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
The Principle of Optimal Behavior and the Theory of Communities 321
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.
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
324 H. Ronald Pulliam
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
326 H. Ronald Pulliam
FEEDING
RATE
(R)
--~----------------~----------------------.x
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.
Table II. Time Budget for the Suhmissive Bird when Time Spent in Long Duration Aggression
is Added to the Model
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.
328 H. Ronald Pulliam
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).
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.
330 H. Ronald Pulliam
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
The Principle of Optimal Behavior and the Theory of Communities 331
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.
332 H. Ronald Pulliam
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