Scott CausesFightingMice 1951

The Causes of Fighting in Mice and Rats
Author(s): J. P. Scott and Emil Fredericson

Source: Physiological Zoology , Oct., 1951, Vol. 24, No. 4 (Oct., 1951), pp. 273-309
Published by: The University of Chicago Press. Sponsored by the Division of
Comparative Physiology and Biochemistry, Society for Integrative and Comparative
Biology
Stable URL: https://www.jstor.org/stable/30152137
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
and The University of Chicago Press are collaborating with JSTOR to digitize, preserve and
extend access to Physiological Zoology
This content downloaded from

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
All use subject to https://about.jstor.org/terms
PHYSIOLOGICAL ZOOLOGY
Volume XXIV OCTOBER 1951 Number 4
THE CAUSES OF FIGHTING IN MICE AND RATS'
J. P. SCOTT AND EMIL FREDERICSON2

Division of Behavior Studies, Hamilton Station, Roscoe B. Jackson Memorial Laboratory
INTRODUCTION various species of animals have various

LL behavior may be considered as special methods for adjusting to environ-
A an attempt of an organism to
adjust to changed conditions,
whether internal or external. When fight-
mental changes. The fighting behavior of
mice and rats has particular usefulness
for science because it can be readily and
ing behavior is analyzed, it is found to be easily studied. Mice especially can be
one of several patterns of behavior which readily induced to fight fiercely but, at
may provide some measure of adjust- the same time, do little damage to their
ment when two organisms come into con- opponents unless the fight is prolonged;
flict. Other common and closely related and they can be readily controlled and
alternate patterns are escape behavior, separated.
defensive behavior, and passivity. It is An attempt to understand patterns of
difficult to consider any of these without agonistic behavior in two closely related
the others, and it is with this general species should contribute much to a gen-
group of behavioral adjustments, which eral understanding of these phenomena.
may be given the name "agonistic be- Consideration of the factors affecting
havior," that this paper is concerned. fighting and its ultimate causes in rats
These patterns of behavior have prob- and mice immediately raises problems of
ably been most extensively and thor- the broadest general interest. For ex-
oughly studied in two laboratory ani- ample, what is the real use of fighting?
mals, Rattus norvegicus and Mus muscu- Darwin's hypothesis of advantageous in-
lus; and on the basis of facts collected dividual competitive behavior must be
from these two rather closely related re-examined in the light of modern
species it is possible to arrive at certain knowledge of social organization. Why
interesting generalizations. do males fight more than females, and
It almost goes without saying that are there other hereditary differences in
aggressiveness? Is there an inborn and
I This study was supported in part by a research
grant from the National Institute of Mental Health, unsuppressible "drive" for fighting? How
U.S. Public Health Service.
do training and heredity interact to mod-
2 The authors wish to thank the followinlg workers
ify fighting? Why do mice generally form
in the field for a critical reading of the manuscript:
dominance hierarchies with one domi-
W. C. Allee, Benson Ginsburg, Elizabeth Beeman,
nant and several subordinate animals in-
J. P. Seward, John B. Calhoun, and Marvin Kahn.
273

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
274 J. P. SCOTT AND EMIL FREDERICSON
stead of straight-line hierarchies?

are fought inHow
round-robin fashion (Bee-
can fighting be controlled? man and Allee, 1945). Before attacking,
The following paper is an attempt to around or near the op-
the mouse circles
ponent,
answer some of these questions by with
con-the body humped up and
ducting a systematic investigation of steps. This behavior
taking short, rapid
available knowledge concerning
may the vari- as a combination of
be interpreted
ous factors which affect agonistic
attacking and be-
defensive postures, or it
havior in the two species, tomayorganize
be a warning signal similar to the
these facts on a firm theoretical
above. It basis,
is usually found in animals
and to indicate points where whichnew re-
have experienced both victory and
search is needed. defeat.
B. ATTACK
DESCRIPTION OF PATTERNS OF
AGONISTIC BEHAVIOR If two mice attack at once, they

scratch, and wrestle, often so rapidly
The behavior patterns which may oc-
the eye cannot follow their movem
cur in conflict situations may be first
If one attacks and the other runs a
described for mice. This applies to situa-
the first
tions in which the conflict is between one will chase after the second
ing and striking with the teeth wh
mouse and another without involving a
catches up.
second species, and the behavior may be
classified into three groups. C. DEFENSIVE AND ESCAPE BEHAVIOR
A. PRELIMINARY BEHAVIOR
When a mouse is beaten, it sque
Under certain situations actualwhen attacked and, if the space is la
fight-
ing does not begin immediately enough,
but is will run away. If shelter is p
preceded by the following events:vided, it may hide from the victorio
mouse. In small barren cages wher
I. Hair-fluffing.-The hair literally
such
stands on end over most of the escape is possible, the mouse
body,
rear up on the hind feet, holding out
with no emphasis on particular regions,
as in the cat and dog. The amount front of
paws toward the aggressor and
hair-fluffing is usually difficult to deter- motionless until attacked, w
maining
mine, and a poorly groomed animal it may
maysqueak and jump. This is a v
give somewhat the same appearance. characteristic and easily recognized
tern of behavior and has been called the
2. Tail-rattling.-The mouse rapidly
"submission
twitches its tail from side to side, so that reaction" because it is a
it takes on a wavy appearance. Ifvery
the reliable
tail indicator of a dominance-
subordination
happens to strike against some hard ob- relationship (Ginsburg and
Allee, 1942).3 Scott (1946) interprets this
ject a loud rattling sound is produced.
Both this and hair-fluffing appear to be
3 Ginsburg and Allee described this attitude in
involuntary behavior patterns,
thewhich
same way, except that one foreleg is said to be
other mice may learn to observe close
as warn-
to the body and the other extended stiffly. The
authors of
ing signals, and occur in situations inthe present paper have never observed
this to occur consistently. However, this is a very
which the mouse is hesitating minor
between
point, and the general facts appear to be that
attack and escape reactions (Scott,
the attitude of defense exhibited by an animal in its
first fight becomes more and more stereotyped with
1947).
repeated defeats, until it assumes this posture in a
3. Mincing.--This type of behavior mechanical way similar to a reflex action which may
apparently occurs frequently when mice be called a "submission reaction."

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
FIGHTING IN MICE AND RATS 275
behavior as an attitude of defense. If a When any two animals are repeatedly

beaten mouse is driven into a corner or placed in conflict situations, their be-
small space, he may roll over on the backhavior is soon organized into social domi-
and hold all four feet in the air. This is
nance. This phenomenon may be defined
simply an extension of the defensive atti- habitual control of the behavior of one
as
tude. animal by another by means of force,
There are certain commonly observedfighting, or threats. This control in the
patterns of behavior in mice which maymouse usually includes the more or less
be confused with fighting. complete inhibition of the patterns of at-
i. Investigation.-When strange micetack, so that the subordinate mouse
meet, one animal may nose the other infights only when attacked or not at all
various parts of the body. and confines its behavior to defense and
2. Grooming.-Mice tend to clean oneescape reactions. The same definition
another's fur with the claws and teeth,will apply to the behavior of rats.
and this may appear very similar to bit- In the scientific papers on the subject,
ing. In some cases rough grooming is ap-considerable attention has been devoted
parently substituted for aggression. to so-called "behavior traits." A "trait"
3. Sexual behavior.--Sexual mountingmay be defined as the consistent appear-
and chasing may appear very similar toance of a certain pattern or group of pat-
an attack, and in some cases this mayterns of behavior when an individual is
start fights between males. put into a given situation. This con-
Essentially the same patterns of be-sistency may be theoretically due to
havior can be seen in rats, except thateither heredity or habit formation or a
tail-rattling has not been reported forcombination of the two.
these animals. The defense posture is Most authors have used popular ter-
very characteristic, but rats often rear up minology, and this may at first be con-
together, so that it is difficult to tell fusing to the reader in this field of scien-
which is the dominant and which the sub- tific literature. Fortunately, the terms
ordinate animal. Another difference is
have been used fairly consistently. Where
the in
that playful fighting is very common situation is one of conflict between
young rats and has never been observed
mice or between rats, the tendency to at-
in young mice. Calhoun (1948) hastack
re- is called "aggressiveness," and the
ported that rats under wild conditions
tendency to run away or escape is called
will administer "psychological drubbing"
"timidity." A lack of conflict behavior is
or mild attacks on the young animals.
called "peacefulness."
This also has not been reported for mice.
On the other hand, where the conflict
Fighting in most of the tame strainsliesof
between mice and people, or rats and
rats appears to be much less vigorous
people, the tendency of the nonhuman
than in tame strains of mice. animal to attack has been called "savage-
The behavior patterns described ap-ness," and the tendency to escape has
pear to be largely unlearned reactions, as been called "wildness," while an absence
they may be seen fully developed in miceof conflict behavior is called "tameness."
which are fighting for the first time. An Several interesting problems come out
experienced and successful fighter showsof a consideration of the trait approach
little change in behavior, except that heto agonistic behavior. One of these is the
tends to become more vigorous and toproblem of measuring native aggressive
bite more viciously and more effectively.tendencies, which obviously are modified

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
by social dominance as soon ashave

thechiefly
ani- been used. One is the
"round-robin" method, in which each
mal has had any experience in fighting.
Native aggressiveness can be mouse
measured
in a group is given the opportunity
only at the time of the first fight,
to fightand
with every other mouse in turn.
even here may be modified by This was adapted from the standard
previous
method
training (such as handling, in the case for
of measuring the dominance
mice). Another problem is that of in
order how
a group, and usually has the ef-
general such behavior traits may be.
fect of Is
setting up a dominance hierarchy.
an animal which is very aggressive also
The members of the group have different
experiences,
very savage? The fact that the same ani- varying from a majority of
mal may be dominant in one victories
situationto a majority of defeats. The
and subordinate in another indicates other method is to standardize the train-
that behavior traits may have to be ingde-for all individuals in a group, either
by introducing mice which are under con-
scribed only in terms of the special situa-
tions and social relationships which occur
trol or by stopping the fight before seri-
ous injury takes place. In the latter case
in a given animal's life. A third interest-
ing problem is that of the connectionbehavior
be- is measured through response
tween the traits which appear to belatency,
op- or the time which it takes for a
posite but are actually alternativefightre- to develop.
The test situations themselves can be
sponses to the same situation. Is an ani-
mal which is more timid in one situation subdivided into two classes: those which
also more aggressive in another situa- are noncompetitive in the sense that the
tion? Or are aggressiveness and timidity animals simply fight in bare cages and
modified independently by various fac- have no external object to fight for and
tors? These problems will be discussed those in which the animals compete for
under the section on heredity below. some object or substance. This latter
situation may be further subdivided into
METHODS AND TECHNIQUES
cases in which the animals can actually
Hall has summarized the techniques get possession of some object, such as a
for studying this problem in his small
1941 pellet of food, and situations in
which
paper. This evidence and that which is the animal can only crowd others
found in subsequent experimental work away from a favorable location, such as a
make it clear that two basic experimental
drinking fountain or a spot in front of a
methods have been used. One is to take food dish. As will be shown below, the
animals which have been trained to fightdifferences between these situations make
and vary the different factors which are a great deal of difference in the results
supposed to affect the amount and inci- (see Table i). For example, the noncom-
dence of fighting. The other basic method petitive type of fighting is usually much
is to take animals which are naive and more vigorous, a result which provides a
untrained and try out the effects of simi-major experimental problem.
lar experimental factors. These twoThe early work on behavior traits
(Yerkes, 1913; Utsurikawa, 1917) made
methods give very different results, since
the first tends to test the importance considerable
of use of rating scales, usually
based on the appearance of certain be-
training and habit formation and the
second tends to test for the presence havior
of patterns. This was necessary be-
innate predisposing tendencies. cause the reaction tested was often that
Among training methods two types
of savageness in response to people. In

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
this situation the experimenter

may be transformed
is ob-into a numerical
viously dominant and makes
score attempts to but this process
or scale if desired,
must
provoke the subordinate rat alwaysfighting
into be somewhat arbitrary.
back by restraint, teasing with a straw,
RESULTS OF EXPERIMENTS ON INEXPERI-
etc. Such a rating scale is also useful
ENCED OR NAIVE ANIMALS
when the amount of conflict behavior be-
tween two relatively peaceful animals is At the outset, it should be stated
measured and only the milder forms of no animal is "inexperienced," excep
conflict behavior are present. sibly at the instant of conception
On the other hand, in the more recent term is used here to refer to animals
experiments with fighting between mice, which have no known experience with
such rating scales have been found agonistic behavior.
TABLE 1
CONFLICT BEHAVIOR APPEARING IN VARIOUS SITUATIONS TO WHICH

EXPERIMENTAL FACTORS MAY BE ADDED
COMPETITIVE
TRAINING NONCOMPETITIVE
Possession Possible Possession Impossible
Primary situation Violent fighting di- Fighting directed Either no conflict

(naive animals) rected toward other toward object behavior or mild
animal pushing
Situation repeated Behavior associated Behavior associated Same as above

(experienced with presence of with object: domi-
animals) other animal: nance-submission
dominance-sub- relationship de-
mission relation- velops if fighting
ship develops if prolonged
fighting pro-
longed
unnecessary. True fighting behavior is A. FIGHTING IN NONCOMPETITIVE
SITUATIONS
unmistakable, and its presence or ab-
sence can, in most cases, be used as the I. Development of fighting behavio
result of the experiment. Thus there is no mice and rats.-It is a common observa-
question of validity and in most cases tion among experimenters that litter-
very little question of reliability, pro-
mates which are raised together may live
vided that the animals are not tested peacefully in the same box even at a quite
advanced age, and Scott (1946) has
oftener than once every 48 hours on pro-
longed fights. More frequent bouts tend
found considerable difficulty in training
to cut down the amount of fighting, these mice to fight. On the other hand,
Fredericson has raised isolated animals
though not if the fights are terminated
after a brief interval. (C57 black io strain) from the age of 21
days (195oa) and has found that, when
On the whole, the presence or absence
of certain behavior traits may be consid-
they were brought together for a short
ered a more objective and accurate period daily, all the animals had begun to
measure of behavior than using rating fight vigorously by 36 days. The tenden-
scales or graded impressions. The former
cy to fight had appeared only I or 2 days

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
ducing
before the 36-day age, and this defensive
period in fighting, as this type
life seems to correspond to the occurs
time whenvery early in development and
the sex hormones first become active will be seen in either males or females.
The male sex hormore appears to be
(Snell, 1941). Female mice (with a few
very important in producing aggressive
exceptions) in this and most of the other
inbred strains do not fight either at fighting,
this but its influence may be nulli-
time or later. These results support fied
Bee- by various sorts of training. Sudden
man's (1947a) conclusion that the male or threatening movements may also be
factors in inducing fighting, as restraint
sex hormone is a definite factor in causing
fighting in the absence of competition. may be.
It is also apparent that the training
In rats Seward (i945) has described
and previous experience of animals which the development of fighting behavior in
have never fought have an important young albino rats, starting at 35 days of
effect on their reactions. The results cited
age. Individuals of either sex indulge in
above in which animals raised together what appear to be playful wrestling
show a tendency not to fight may bematches,
ex- leaping, pouncing, and rolling
over and over. Such behavior is rarely if
plained as a case of passive inhibition.
However, if peaceful male mice with ever seen in mice.
no previous fighting experience are at- At about 45 days of age this playful
fighting may become more serious, with
tacked by others, they almost invariably
squealing and defensive postures. It
fight back; and, if the mice are separated
immediately so that no harm is done,
would appear that the start of the fight is
caused by pain, perhaps as one rat plays
they consistently fight back on subse-
quent occasions. Such results suggest
too vigorously. It would also appear that
the onset of fighting probably coincides
that pain is an important factor in pro-
ducing fighting in undefeated mice. with sexual maturation and the appear-
ance of the sex hormones.
An experiment was done (Scott, 1946)
on young mice by subjecting them to
After puberty fights occur between fe-
males but less often lead to dominance-
pain at various ages by gently squeezing
subordination status than in the males.
the tail with forceps. The results showed
that even the newborn animals reacted Sex differences in fighting appear to be
less extreme in rats than in mice.
by squeaking and making attempts at
escape. As soon as the eyes were opened, If male rats are allowed to meet for the
first time beyond the age of puberty,
the mice would turn and bite the forceps.
they apparently form dominance rela-
Two days later it was found that threat-
ening movements of the forceps wouldtionships much more quickly after the
age of ioo days is reached.
produce the defense posture. All this oc-
curred long before attacking behavior In wild rats, Calhoun privately reports
had been observed. that the onset of lactation appears to
The conclusion can be drawn from the increase the vigor of fighting in females.
above data that there are three impor- O'Kelly and Steckle (1939) have re-
tant factors which affect fighting in mice ported that repeated electric shock ap-
which have never fought before: train- plied to the feet of a group of six male
ing, the presence of the male sex hor-rats had the effect of starting up violent
mone, and pain (which may result fromfighting, which eventually became long-
an attack by another animal). Pain is ob-continued. Using a different strain and
viously the most important factor in pro-different apparatus, Daniel (1943) was

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
able to get fighting, butfor

notthe of a long-
possession of a food pellet. The
continued sort. However, this
behavior provides
was not described in detail, but
additional evidence that
it pain is also
is obviously inviolent type of
not the
rats an initial cause of at least defensive
fighting described by Seward (1945) in
fighting. noncompetitive situations.
There have been no experiments in Fredericson (I95oa, b) has found that
either mice or rats on the question of how mice which are not hungry will not com-
various physical environmental factors pete or fight for a piece of food intro-
might affect the first appearance of fight- duced into the cage. On the other hand,
ing. mice which have been starved for 24
2. Conflict with people: wildness and hours or longer will struggle for a piece of
savageness.---As indicated above, mice food and occasionally vigorously bite or
and rats tend to be wild and savage when wrestle with the opponent. This has been
first handled by people, and this may be observed as early as 29 days of age. This
interpreted as a reaction to a situation of sort of behavior is quite different from
conflict with a member of a different the type of fighting found in noncom-
species. Wildness includes chiefly defense petitive situations. Effort is directed to-
and escape reactions, and savageness in- ward the possession of the food and at-
cludes chiefly defensive fighting and pos- tacks on the opponent stop as soon as the
sibly in some cases aggressive fighting. food is gotten. This contrasts with the
No careful work has been done on the much more vigorous and violent type of
nature of the stimuli which elicit this re- fighting which occurs in the absence of
sponse; as in most cases, the authors werecompetition, where activity is directed
interested in the nature of the responseentirely toward injuring or escaping from
itself. Pain, noise, sudden movements,the opponent and fighting goes on for
and restraint have been obviously in-long periods. Furthermore, the competi-
volved in the experiments of Yerkes tive type of fighting is independent of the
(1913), Utsurikawa (1917), and Stonemale sex hormone, since it can be pro-
(1932), on rats, and Coburn (1922) on duced in females and in males as young
mice; but other factors may be discov-as 21 days of age.
ered by the use of a more-careful "re- This new technique opens up the pos-
leaser" technique of study. sibility for experiments in which other
types of things can be competed for, as,
B. FIGHTING IN COMPETITIVE SITUATIONS
for example, the possession of a female in
Relatively few experiments have heat. Though
been dependent upon the male
done until recently with competition hormone,in the noncompetitive type of
inexperienced animals; and the fighting most ob- may take place whether or not
vious fact is that the results of such ex- the female is present or receptive.
periments seem to depend upon whether A reference may now be made to situa-
or not the animal can actually get control tions in which complete possession of the
over the object competed for. These two object competed for is not possible. Gins-
sorts of experiment will be discussed burg and Allee (1942) found no effect of
separately. starvation up to 36 hours (it is not stated
Bruce (194j) found that thirsty rats whether or not food could be possessed).
would struggle over a water fountain at Mice thirsty for 6-8 hours did not fight
which only one animal could drink and over a water bottle with a small nozzle.
that hungry rats would similarly struggle In rats Bruce (1941) found that thirsty

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
animals which were running hormone,

in an alley
pain, threatening movements,
showed no competitive behavior over
and possibly noise. In competitive situa-
drinking at a fountain if it
tions,
couldwhere
be possession of a portable
shared, and Fredericson (1946)food
foundobject
notis possible, hunger may be
added, the male hormone is unimportant,
only mild pushing but some indications
of dominance where hungry rats
and pain were
and threatening movements do
not seem
given a food cup which permitted to have much significance.
only
one animal to eat at a time. One serious question raised by these
Seward (1945) performed experiments results is the reason for the existence of
with both types of situations, using pel-noncompetitive fighting. In mice raised
lets and a food-hole. In both cases the in cages it serves no obviously useful pur-
rats had previously established a domi- pose, and, if such noncompetitive fight-
nance order and were allowed to compete ing appears under natural conditions, it
for food later. The amount of fighting would seem to be a disadvantage to the
was about the same with the pellet species. Two suggestions may be offered
(which could be possessed) and was here. One is the possibility that it tends
markedly reduced with the food-hole. to keep a population so scattered that a
These results are discussed later in con-group of males cannot be concentrated
nection with trained animals. around a food supply in a given area
(there are no indications that mice ever
The evidence is not entirely clear, but
the following tentative conclusions maydefend a definite territory as birds do, al-
though Calhoun finds indications of this
be drawn: (i) in both rats and mice, hun-
gry animals will struggle for the posses-wild rats). The other suggestion, which
in
is based on certain observations to the
sion of a food pellet, but with consider-
effect that fights may start as the result
ably less violence than in noncompetitive
of attempts by one male sexually to ap-
fighting; (2) thirsty rats (but not mice)
proach another, is that the fighting may
will struggle for a place at a water foun-
tain; (3) in rats, a situation such asserve
a to prevent sterile sexual activity.
The confirmation of these hypotheses
food-hole, in which one animal can block
off another from food, tends to producemust await the results of experiments on
less fighting than a pellet which can behavior
be in relation to socially organized
possessed. The inconsistencies may populations.4
be
caused by hereditary differences between The data also bring into question the
rats and mice but are more likely caused
universality of the frustration-aggression
by differences in technique. The amount hypothesis. Under certain conditions
of struggling would appear to be propor-when the possession of an object is the
tional to its usefulness in the situation.
problem, frustration leads to aggression
To summarize this section on fighting in mice and rats. When possession of the
in inexperienced animals: the importanceobject is impossible, frustration does not
necessarily lead to aggression. Further-
of training and experience is clearly indi-
cated, as well as the importance of the
more, in noncompetitive fighting there is
type of situation to which the animalsno obvious way in which the animals
are exposed. Factors which are impor- to be frustrated, unless it may be
seem
tant causes for fighting in one situation
4 Calhoun suggests from his observations that in
"noncompetitive" fighting, rats and mice fight for
are ineffective in others. In noncompeti-
undisturbed possession of living space, rather than
tive situations the important factors be-
any particular bit of space. The result would be
sides early training are age, the male
dispersal of the population.

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
the frustration of a nonaggressive males is placed with

sexual appetite. Ex- a fe-
periments in this area
male in which involve
estrus, allowed to copulate, and
then is placed
sexually satiated animals would with the
seemother male,
to the
be indicated; but, on latter
thepursues
whole,the first
it andmaymay start a
be said that frustration leads
fight. This to
appears aggres-
to be a case where one
male reacts
sion where aggression provides thetoward another as ifsolu-
it were a
female,satisfactory
tion to a problem (i.e., and the resulting homosexualad-be-
justment). havior leads to a fight, rather than a case
of competition for possession of a female.
C. NEEDED EXPERIMENTAL WORK
Further experiments are needed along
It will be obvious from the above ac- these lines.
count that almost no work has been done
RESULTS OF EXPERIMENTS WITH EXPERI-
with factors which might be expected to
ENCED OR TRAINED ANIMALS
cause inexperienced animals to start
fighting. Males seem to start their non- Experiments of this sort have bee
competitive fighting at about the time signed to test the effect of various f
when the male sex hormones might be upon the behavior of animals whic
expected to appear first, but there has used to fighting. Experimental f
been no experimental analysis to see might therefore be expected either
whether this could be produced at earlier duce or to increase the usual amount of
ages. Experimental work also needs to be agonistic behavior. Theoretically, three
done on the effect of lactation and the types of situation might be employed:
various endocrine changes which go with
one in which two animals had been al-
it.
lowed to develop a relationship of domi-
Little work has been done on the ef- nance and subordination; one in which
fect of environmental factors which
animals have been trained to fight, but
might cause inexperienced males living
not allowed to develop dominance; and,
together to start fighting. Littermates
finally, one in which the two animals
brought up together and living undis-have been trained to be peaceful. The
turbed in the same cage usually dofirst
not of these has been used extensively
start fighting until well along into with
ma- noncompetitive fighting, the second
turity, if at all; and it would be interest-
with competitive fighting, but the use of
ing to see whether or not the stresses theof
third has been little explored.
cold, hunger, sexual frustration, etc.,
might cause such males to initiate fight- A. FIGHTING IN NONCOMPETITIVE
SITUATIONS
ing. Once the habit is established, of
course, these factors would probablyMost of the work which has been done
become unimportant. in this area has been directed toward
modification
With regard to fighting in competitive of the dominance-subordi-
situations, experiments have been nation
done relationship, which has been pre-
viously defined. An extensive and thor-
only with food and water, and the con-
ough series of researches on mice has
clusion seems to be that fighting results
been
only when possession of the object can be done in this field by Beeman and
assured. So far there is no indication that Allee, who have used the technique of
competition for a sexual object will cause round-robin encounters, in which a small
fighting in rats or mice. Ginsburg andgroup of mice, isolated except during the
Allee (I942) found that when one of two experimental period, are fought at regu-

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
lar intervals in rotation. After a few en- themselves out on a cool surface. In the
counters, dominance relationships of aabsence of detailed experimental work, it
fairly stable sort are set up and other fac- may be tentatively concluded that high
tors can be introduced. In such a tech- temperatures have an inhibiting effect on
nique the usual result is to place one ani- the amount of general activity.
mal on the top and another on the bot- The possibility exists that mice may
tom, but the intermediate animals may fight over nests, although this has not
not be clearly ranked among themselves, been observed. Scott (i944b) reported
probably because mice appear to havethat males were never found in nests with
difficulty in recognizing other individuals females and very young mice when the
except by their attitudes of attack or animals were given the opportunity to
defense. construct substantial nests. However,
i. Ecological factors (physical).-Bee-the females were never observed in the
man has privately reported no effect ofact of driving them away.
cold on the incidence and vigor of fight- A certain number of experiments have
ing. Males were kept at a temperature of been done with the factor of space. In the
170 C. for 24 days before round-robin en- usual laboratory set-up, mice or rats are
counters were begun. Scott and Marston kept in isolated cages and the usual ex-
(1946) did an experiment in which pairs periment involves transferring the mice
of male C57 black mice which had estab- to another cage or to that of another
lished dominance were alternately fought mouse. Uhrich (1938) reported that mice
at a temperature of 50' F. (approxi-left in their home cages were likely to
mately 160 C.) and at room temperature have more success in fighting than were
(700 F.). No significant difference wasintroduced mice. The observation of such
observed, although there were some indi- mice indicates that the stranger spends
cations that the low temperature might most of his time investigating the pen,
slightly lower the amount of fighting. while the home mouse concentrates on
This is contrary to casual observationsfighting. The effect might be due either to
made on people, who appear more ir- the defense of the home cage or to the
ritable when exposed to sudden tempera- fact that the strange mouse is upset and
ture changes, and it may be explained bynervous.
the fact that adjustment to temperature There appears to be no evide
in mice is probably a matter of thyroid the guardianship of specific terr
adjustment rather than changes in cir- mice. Using a multiple pen w
culation, so that no immediate physio- boxes connected by runway
logical effect would result from a sudden (I944b) was unable to find any in
change in temperature. Thus it would ap- of territorial defense. Howeve
pear that neither long exposure nor sud- possibility cannot be excluded,
den change to low temperature has any space were given. Calhoun (1949
appreciable effect on the habitual fight- served cases in rats living under
ing of mice. conditions but in a large field, w
On the other hand, Ginsburg and Allee tain animals appeared to be conf
(1942) and most other experimenters certain areas by the attacks of oth
have observed that mice which have been dominant animals ranging more
exposed to high temperatures tend to The result is a rather vague sort
fight less vigorously and attempt to ad- toriality, in which most animal
just by digging in shavings or flattening cluded from a given area where

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
effectThere
more dominant animals live. on the dominance-subordination
ap-
relationship
peared to be no cases in which by modifying the type of
definite
boundary lines were defended,
agonistic as re- and upon social or-
behavior
ported in birds, where oneganization
animal by allowing
is a vague sort of
dominant in his own territory and sub-
territorialism in large areas.
ordinate in another's. Such activities 2. Ecological factors (biological).-No
work has been done with biological fac-
would be very difficult in mice and rats,
which are largely nocturnal in habittors
and in the environment with the excep-
accustomed to living in areas which tion
pro-of using the behavior of one animal
vide a great deal of cover. to modify that of another. For example, a
Crew and Mirskaia (1931) havesevere
re- defeat will cause a dominant ani-
ported experiments in which spacemal wasto become subordinate not only to
the aggressor but also to other animals
kept constant, but the density of animals
which he meets soon thereafter (Gins-
varied. Varying numbers of pairs of mice
were placed in small cages of uniform burg and Allee, 1942). Conversely, Scott
size. Dominance was not recorded, (1944a)
but found that a habitually peaceful
the maximum death rate (50o per cent) relationship could be upset by training
was observed in cages containing 4 pairs
one of the animals to fight. Thus, while
of mice, with slightly more favorablethere
re- are indications that the biological
environment may be important, the na-
sults with either larger or smaller popula-
tions. ture of the factors is really social and
psychological and should be discussed
The results must be interpreted with
regard to density rather than to space.
under the section on training below.
Crowding tends to increase the probabil- The effect of predators and parasites
ity of fighting but, beyond a certain on fighting behavior has not been stud-
ied.
point, exerts a favorable effect on mor-
tality by permitting the beaten animals 3. Physiological factors.-The effect of
to hide and escape notice. However, it
nutrition and hunger on fighting of ani-
may be inferred that small space tends towhich have set up a system of domi-
mals
prevent the formation of a dominance nance has been studied extensively. Gins-
order, which would permit survival of and Allee (1942) found that hunger
burg
individuals. and thirst did not affect the tendency to
Scott (1946) experimented with domi-fight. Seward (1945) performed extensive
nance in two types of pens, one large andexperiments with hungry rats in which
one small, and reported that behaviorhe used situations in which the animals
between dominant and subordinate ani- could obtain possession of the food and
mals was significantly altered. In the
could not obtain possession of it. The re-
sult was to cut down the amount of
large pens the dominant animal did more
chasing and less attacking, while the habitual fighting, particularly in the case
subordinate animal did more running where possession was impossible.
away and adopted the defense postureSeward used these results to question
fewer times. This indicates the adaptive
the frustration-aggression hypothesis,
nature of these responses. arguing that the hungry animals should
In summary, heat and probably cold
have become more aggressive because
they were frustrated. However, since
tend to reduce the amount of fighting be-
havior. The factor of space has been
training seems to be the all-important
definitely shown to have an important
guiding factor in fighting, it will be seen

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
usingtried
that in his experiment he simply the round-robin
to technique. These
animals never fought. Replacement ther-
introduce a new type of countertraining
and stimulation, which tendedapyto inter-
with testosterone propionate pellets
fere with habits already formed. A fair
was followed by the development of the
usual amount
test of the frustration-aggression hy-of aggressiveness in the
pothesis would be to take mice
mice, or
and rats
removal of the hormone pel-
lets caused an almost immediate cessa-
which had been trained in a competitive
situation and then vary the tion
amount of
of aggressiveness.
frustration. Such an experiment would
However, the fact that some animals
be comparable to those in goats (Scott,
continued to fight led her to make further
1948) where positive resultsexperiments
have been in which animals were first
trained to
obtained. However, all these results fight and become dominant in
tend
to indicate that frustration as a cause ofsituation and then were
the round-robin
aggression is only important castrated
where andani-tested for fighting without
mals have been trained to react being given any rest. These animals con-
aggressively
after frustration. tinued to fight without interruption.
Beeman and Allee (1945) have
Thesedone
results amay be explained either as
series of experiments in which the Vitamin
continued effect of strong habit, or it
may have been that, once the habit of
B, deficiency was used as an experimen-
tal factor. It was found that, fighting
once domi-has been established, the small
nance had been established, amount of male hormone from the adre-
the domi-
nant animal could maintain hisnal position
gland is sufficient to maintain aggres-
sive activity.
almost to the point of death from av ta- Beeman is continuing work
to test
minosis, the subordinate animal there-
still adrenal gland hypothesis at
the present of
sponding to the threatening attitude time.
the feeble dominant male. It is The conclusion must be reached that
probable,
however, that the amount and thevigor
presence
ofor absence of the male hor-
mone has athe
fighting was affected, even though very important effect, at least
number of fights won was not. on the initiation of fighting under non-
From these experiments it maycompetitive
be con- situations. Whether or not
it has an may
cluded that nutritional deficiencies effect on a well-established
lower the amount of fighting, habit of fighting
either by is still in doubt.
Beeman
distracting the animal's attention, as in has also started an experimen-
tal program
Seward's rats, or by causing a diminution on the effect on fighting pro-
of vigor, as in Beeman's mice,duced
but by
thatstopping the function of vari-
the dominance order and habit ousofsense organs and by ablation of vari-
fight-
ing are very little affected. ous parts of the brain; but this work
cannot
Beeman (1947a, b) has also done be reported on as yet.
some
excellent work on the effect of It the
mustsex
be concluded that, once habits
hormones on fighting. Females ofin
fighting
the do- or running away have been
mestic strains of mice very seldom get strongly established, organic changes
into fights in the ordinary situations tohave very little effect unless they are so
which they are exposed, whereas fighting serious as to threaten the life or cripple
in males is much more common. She the functions of the animal. The one pos-
therefore castrated male mice at various sible exception to this is the male hor-
ages and allowed them to rest 25 daysmone, and even it seems to have much
before starting to train them to fight, more effect on the initiation of fighting

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
than upon strongly formed habits

EXPERIMENTS of
WITH TRAINING
fighting. Since pain (as pointed out in an

A. EFFECTS OF EARLY ENVIRONMENT
earlier section) has the effect of causing
Kahn to
young females, as well as males, (1949)bitehas found that m
back, the hypothesis maywhich are severely defeated at v
be tentatively
immature
suggested that the hormone has ages
the will less readily a
dangled mice
effect of lowering the threshold when given the oppor
of sen-
as adults than will adults with no such
sitivity to painful stimuli. This fits
previous
in very well with the time in experience.
develop-
Fredericson
ment in which noncompetitive (I950b) has shown a simi-
fighting
appears. lar effect with competitive fighting.
It is obvious that many more experi- Young hungry mice between 29 and 35
ments remain to be done in this field. days of age were allowed to compete for
In evaluating experiments of this type food
it and were later tested in a nonhungry
state at 72 days of age. They immedi-
is essential (because of the importance of
ately competed for food, while control
training) to keep in mind the point as to
whether or not the experimental factor animals did not. Even a single experience
was altered before or after the trainingappears to produce this effect.
period began, and the amounts of aggres- As pointed out under the section on
sive or escape behavior should be meas-development of fighting behavior, Fred-
ured as well as dominance. ericson has found that male mice which
are raised alone and periodically allowed
B. COMPETITIVE SITUATIONS
to investigate each other for short pe-
The only work in pure competitivewill
riods begin to fight at around the age
sit-
uations which has been done on mice is of 36 days. On the other hand, it is com-
that of Fredericson (i95oa), who has re-monly observed that males which are
peatedly allowed hungry mice to com- raised together in the same cage and left
pete for food, separating them before undisturbed often do not fight even after
dominance develops. After a period of they reach maturity.
training in the competitive situation the The fact that animals which are raised
animals will compete for food, eventogether do not fight may be explained
though they are not hungry, indicating by the hypothesis of passive inhibition.
that frustration is not necessary, once theIn Pavlovian conditioning the repetition
habit of fighting has been built up. How- of a stimulus to which there is no primary
ever, Fredericson has done no work as
reaction will cause that stimulus to be
yet on the decrease or increase in fightingassociated with lack of activity; and
as a result of frustration, and in no case when it and a stimulus which is associ-
has he worked with situations in which ated with activity are presented together
dominance has been developed. it can be shown to have a definite inhibit-
Obviously, the situation has not been
ing effect. Presumably the young mice
so thoroughly explored as has dominance
raised together which do not fight up to
in the noncompetitive situations; but30 days form a strong habit of not fight-
here again the tentative conclusion can
ing, which is sufficient to overcome stim-
ulation which under other conditions
be made that training is a very important
factor compared to physiological vari-may produce it. Thus there is good evi-
ables. Females respond in this situation
dence that early training and experience,
as well as males (Fredericson, 1951). even though fighting may not be in-

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
286 J. P. SCOTT AN'D EMIL FREDERICSON
volved, has an important effect upon

tinction).-If a secondary stimulus is re-
fighting in later life. peated in the absence of a primary stimu-
lus, it will tend to be associated with a
B. THE PRINCIPLES OF LEARNING
lack of response. Such a stimulus given
As intimated in the preceding
togetherpara-
with a secondary stimulus which
graph, the effects of experienceison
associated
fight-with a response will tend to
weaken
ing behavior can be related to the the response (inhibition). If a
vast
secondary stimulus associated with a re-
body of knowledge which psychologists
have collected under the heading of
sponse is repeated without reinforcement
learning phenomena. by a primary stimulus, the response will
These data need be treated only very
tend to die out (extinction). It should be
pointed
briefly here. The field has been veryout here that a secondary stimu-
ade-
quately summarized by Hilgard lus may be and
associated with a lack of re-
Marquis (1940) and by Hi'gard (1948).
sponse either before (passive inhibition)
The pioneer work of Thorndyke and
or after reinforcement has taken place-a
Pavlov has been expandedconcept by suchwhich is of considerable impor-
workers as Guthrie (1935), Hull (1943),
tance in the control of fighting.
5. Strength
and Skinner (1938), so that certain gen- of association (memory).
The immediacy
eral principles have been established. For with which a response is
given
the most part these may be to a secondary stimulus or the
termed
number
"laws," in the sense that they areofde-
times it will be repeated with-
scriptions of behavior rather out
thanreinforcement
causal is determined by the
manner in of
explanations. The most important which reinforcement is given.
Primacy, recency, and frequency of re-
these may be summarized as follows:
inforcement
I. Adjustment or adaptation.-All be- appear to be particularly
havior is an attempt to adjust important.
or adapt Once reinforced, an associa-
to change, whether the changes tion are
will in-
reappear after extinction, pro-
videdof
ternal or external. As a corollary that sufficient rest is given, indi-
this,
cating that
it may be said that behavior is not a re- the process of association is a
permanent
sponse to a condition but to a change in one.
conditions. 6. Generalization and discrimination.-
2. Primary stimulation or releasers.--When an association has been built up
Certain responses are automaticallybetween a stimulus and a response, there
given to rather specific stimuli. Such re-is a tendency to respond to all similar
sponses form the raw material of learn- stimuli (generalization). Responses to
ing, usually appearing early in develop-unreinforced stimuli are eliminated or
ment. extinguished by negative association
(discrimination).
3. Association, reinforcement (memory).
-If a change to which no response is au- Several basic research problems arise
tomatically given (secondary stimulus)out of these laws. The last four are essen-
tially based on Pavlov's work with condi-
is repeated together with a primary stim-
ulus, the response may thereafter tioned
be reflexes. Can the interaction of
these laws explain the results of experi-
given to the secondary stimulus alone
ments on complex learning and problem-
(association). The repetition of the pri-
mary and secondary stimuli together solving?
is A major controversy in psychol-
called "reinforcement." ogy is raging about this point, the Gestalt
psychologists and such leaders as Tol-
4. Negative association (inhibition, ex-

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
man (1932) pointing out thattion and

thediscrimination,
organ- it would appear
that these
izing capacity of the nervous system also apply.
may
lead to results which cannot The
befinal question of whether these
explained
by a knowledge of simple laws apply to social
elemental proc- behavior has been
esses. However, this controversy
most thoroughlyhas tested on the fighting
little bearing on the causesbehavior of rats and mice,
of fighting in with confirma-
rats and mice, since in mosttory
cases
results
nowhich
com-will be discussed in the
next section. are in-
plex problem-solving situations
volved.
C. TRAINING ANIMALS TO BE DOMINANT
A second major question is whether or
AND SUBORDINATE
not these laws are applicable to all spe-
cies of animals. At the present time it Experiments with dominance in m
may be safely said that the first two ap- have been done almost entirely in
ply to all species of animals and that thenoncompetitive situation, but, be
this is discussed in detail, it is essen
last four, which are chiefly laws of learn-
ing, apply to all vertebrates which havethat the term be thoroughly unders
been tested. Whether or not the process As defined at the beginning of this
ticle, a relationship of dominance
of learning in invertebrates is essentially
different has not been established. All subordination is one in which one of two
animals regularly fights or exhibits ag-
these laws should apply to rats and mice,
and there are many confirmatory experi-
gressive behavior and the other regularly
ments on rats. exhibits escape or defensive behavior. It
The third problem is whether or should
not be emphasized that the domi-
these principles can be extended from nance relationship is one which involves
re-
flex to voluntary behavior, and true from
agonistic behavior of some sort.
these to complex patterns of socialFrom
be- a theoretical point of view it
havior. The work of Skinner (1938)
may on
be assumed that the relationship is
what he has called "operant condition-
caused by simple conditioning, the domi-
ing" has abundantly confirmed the
nant
ex-animal having formed the habit of
attack
tension of these laws to voluntary be-and the subordinate animal hav-
havior. In his experiments the primary
ing formed the habit of escape or defense
stimulus is food, and the primary re-
(Ginsburg and Allee, 1942).
sponse is eating. However, in order to experience or training is assumed
Since
get food, the experimental animalto be the cause of dominance and sub-
must
ordination,
do some voluntary act, such as pressing a it would, of course, follow
lever, which may be called a "secondary
that the previous experience of the ani-
response." This becomes associated mal
withwould be the all-important factor.
This is
eating; and it is clear that Skinner expectation was brilliantly and
thoroughly
working with the association of two re- confirmed by the work of
sponses (both voluntary) ratherUhrich than (1940), and Ginsburg and Allee
with association of a stimulus and re- (1942), who first showed that previous
sponse. success tends to produce a dominant
The laws of association, negative as-
animal, whereas previous defeat tends to
sociation, and strength of association are
produce a subordinate animal. They also
abundantly confirmed; and, although the found that these roles could be reversed.
experimental technique does not lend it- A succession of bad defeats will cause a
dominant animal to become subordinate.
self so easily to the study of generaliza-

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
nate one escapes,

With much more difficulty, involving a it should not be hurt
andof
long period of rest and a series should
easynot attempt to escape so
victories, a subordinate animal
hard. can be
Such lessening of vigor of fighting
trained to be dominant. is commonly observed where contacts
Considered in detail, the experimentsare repeated after dominance is estab-
of Ginsburg and Allee are seen to have lished.
a
direct bearing on the laws of learning. A pair of mice which have been trained
The authors first established the fact to fight in this way will both have some
reinforcement of fighting behavior.
that three inbred strains of mice differed
in fighting ability, so that one could Therefore,
be one might expect that the
depended upon to defeat another insubordinate
an mouse, after a long period
experimental situation. Using the round- with no reinforcement, would revolt and
robin technique, groups of mice in each fight again and that a dominant mouse
strain were fought in neutral cages untilrecently defeated would continue escape
a dominance order was established in behavior because of recency of reinforce-
each. These processes may now be ana- ment. Such appears to be the case.
lyzed from a theoretical viewpoint. To take an example, Ginsburg and
There is some doubt as to what may be Allee took an albino mouse (strain C)
the primary stimuli that evoke fighting which stood at the top of a group of 4
other albinos and subjected it to severe
behavior. Certainly, one of these is slight
pain; another may be the mere sight or defeats twice a day for 8 successive days
contact with a strange male mouse. Theby the alpha mouse of the C57 black
sight of an animal running away also ap- group. The dominant albino fought back
pears to be effective. On the other hand,only during the first three battles. The
defensive and escape behavior are evoked next day he was matched against all
by intense pain. Once a fight is started,members of the albino group three times.
the two animals reinforce each other's On the first round he was completely
fighting by mutual injury, until one may submissive to all mice (showing escape
be badly hurt and respond with escape or defense reactions) but was not com-
behavior. The victor keeps attacking, re- pletely submissive by the third. On the
inforcing escape behavior with further next day, during the third round of
injury. If this situation is repeated, rein- fights, he attacked one of the white mice,
forcement of fighting occurs in one won, and thereafter assumed the domi-
mouse, which becomes dominant, and nant position in the group. When exposed
reinforcement of escape or defensive pos- to nonagressive black and agouti ani-
ture occurs in the other, which becomes mals at this point, he was submissive.
subordinate. Meanwhile, secondary stim- All this can be interpreted as reinforce-
uli become associated with the response. ment of fighting, extinction of fighting,
The change to the fighting cage results in and reinforcement of defensive behavior
preliminary fighting behavior by the while being beaten by the black, gen-
dominant animal, and passivity by the eralization of defensive behavior toward
subordinate one. all animals, including the whites, and,
finally, discrimination between them and
It may be pointed out that the estab-
colored animals. The only unusual fea-
lishment of dominance tends to lessen re-
inforcement of both animals. The domi-ture seen in applying the principles of
nant animal is not hurt so much and
learning to fighting behavior is that the
reinforcing stimulus for fighting is the
should not fight so hard; if the subordi-

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
The first thing

same as that for escape behavior, that he obtained was a
except
strikingextinction
in degree, and that therefore confirmation of the predicted
effect of lessening
of one response necessarily results in rein- of fighting because of
forcement of the other. lack of reinforcement as dominance was
established.
The lowest-ranking black In the males
mouse, the mean num-
ber of
which did not fight at all, "aggressions"
was isolated (pushing, mauling,
nipping, etc.) pernor
for 5 weeks (neither reinforcement bout fell from 24.0 in
extinction of fighting).the
Hefirstwas
round then
to 2.5 in the eighth, while
matched with a submissive in mouse, which
females it fell from 23.9 in the first to
he attacked after it had reacted defen- II.2 in the fourth, after which observa-
sively to grooming and had run away.tions were not continued. Other indexes
Afterward, he was able to beat all the
of the amount of fighting activity showed
other black mice except the dominant similar drops.
mouse; but, when he was returned to the Next, Seward was able to repeat Gins-
round-robin system of fighting, he gradu-burg and Allee's results with defeated
animals. Rats which had been beaten
ally went back to his former place in the
dominance hierarchy. once rose to the original level of aggres-
The reappearance of fighting after asions per bout after 14 exposures to non-
aggressive animals. However, beaten
rest is strikingly similar to Pavlov's de-
scription of the revival of a conditionedanimals which had been isolated i week
showed no rise in aggressiveness, indicat-
reflex following extinction, provided that
rest is given. Fighting was reinforced ining that the rise in aggressiveness was a
bouts with weaker mice for a short periodreal learning phenomenon (extinction)
but was later extinguished, and escape rather than the effect of disuse.
behavior was reinforced by encounters Using the same group of rats, Seward
fought io pairs together, then fought the
with a superior male. The only puzzling
winners together, then matched them
thing here is why he was unable to dis-
criminate between aggressive and sub-
again in the original pairs repeatedly.
missive black mice; but, since the mem-
The result again confirmed the expecta-
bers of an inbred strain are strikingly
tion that repeated matching of a pair to-
similar in appearance, this may not be
gether should lead to lessened activity
too surprising. because of lack of reinforcement. The
winners tended to become less aggres-
Ginsburg and Allee's results were em-
sive, and the losers more aggressive.
pirically obtained from an experiment
Seward, however, observed a special ef-
designed to test the results of defeat and
victory, and the resemblance to Pavlov's fect on the winning rats, after the second
results was noticed afterward. Seward fight (not with the usual partner), in that
(1945) afterward made a deliberate at-
they showed symptoms of fear and ap-
peared to be in an unstable balance be-
tempt to see whether Pavlovian effects
could be predictably obtained on a group
tween fear and aggression, presumably as
of rats. The principal differences werethe result of having had a hard fight.
that he used a group of albinos of unde-The chief difference between these re-
termined inbreeding and that he included
sults and those of Ginsburg and Allee is
both females and males. Moreover, that he no evidence of spontaneous recovery
used more elaborate measures of behav- of aggressiveness or of escape behavior
ior in addition to simple dominance andfollowing disuse is found. Further experi-
submission. ments, dealing with short rest periods of

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
24 hours or longer ones up evidence for dominance and subordina-

to approxi-
mately i week gave the same result.
tion starting in early life when younger
However, in the mouse experiments,
animals are given light beatings by the
older ones. were
much longer periods of isolation One of the chief results is to
used, which may explain theprevent the subordinate animals from
discrepan-
cies. having an undisturbed home range. Be-
cause
These theoretical details may of constant disturbance, the sub-
obscure
the very important practical aspect
ordinate of grow and reproduce less
animals
Ginsburg and Allee's (1942) successfully.
work, that
One interesting
motivation for fighting is increased by peculiarity of domi-
training in accordance withnance in caged rats and mice is that in
reinforce-
ment theory; i.e., a tendency noncompetitive
to fight isfighting the animals tend
to
increased by success in fighting. be either
This dominant
was over all the mem-
bers ofusing
confirmed by Fredericson (1949), a group aor, if not, subordinate to-
different technique. Pairs of ward
mice allin
other
ad-animals. This was origi-
nally discovered
joining cages were let in together onceby Uhrich (1938) in his
per day and were separatedstudiesio seconds
of caged groups of mice and tends
also to come out
after fighting began. The response in work carried on in the
la-
round-robin
tency, or time at which the fight started,method. This contrasts
went down very rapidly and markedly
reached with
a the results in chickens,
stable level after the third trial.
goats, etc., where there is a tendency for
a straight-line
Taken together, the researches of system to be developed
Ginsburg and Allee and those of Seward
within a group, in which the intermediate
form a major body of evidence members
that arethe
dominant over some ani-
mals and subordinate
principles of associative learning apply to others. It also
contrasts
to a major type of social behavior, as with
well Calhoun's (1948) results
with rats
as to simple reflexes and learning in large spaces.
tasks.
The precise
Further, at least some of the principles of explanation cannot be
more complex learning, such given at the present time, but two pos-
as problem-
siblebehavior,
solving, also apply to agonistic explanations have been suggested.
as shown in the work on tamingOne is that the animals cannot recognize
discussed
below. other individuals except by their atti-
It must be emphasized that the experi-tudes and behavior, which indicate either
ments on dominance in mice have been dominance or subordination. Inbred
done under very artificial conditions, instrains have been largely used in these ex-
small cages and often allowing the ani- periments, which would make individual
mals to meet only at the time when therecognition difficult in any case. Another
fight takes place. What sort of relation-possibility is that the fierceness of the
ship would be developed under more nat-fighting is so great that an animal is
ural conditions where the defeated mouseeither victorious or so soundly beaten
can escape cannot be predicted. and emotionally disturbed that he has no
Calhoun (1948) has recently madechance to discriminate when he meets an-
studies of rat behavior under seminaturalother individual. A third possibility is
conditions but with the restriction thatthat this result is caused in some way by
the defeated animals could not com- the limitations of a cage environment.
pletely escape from the situation, the It will be interesting to see whether or
whole area being fenced. He finds strongnot the milder type of fighting which

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
is associated with competition

any damagewill
occurs lead
(if no trained fighter
is available, the
to a different sort of dominance dangling procedure de-
organ-
ization. scribed below is almost as effective). On
the second day a helpless mouse is
D. TRAINING ANIMALS O10 FIGHT, TO RUN dangled in front of the fighter and
AWAY, OR TO BE PEACEFUL
bumped against him gently in such a
way that the fighter is encouraged to at-
For the practical purposes of research,
tack. Four different mice are used in this
Scott (1946) has developed systems for
way, the second and fourth being turned
training animals to fight or be peaceful,
loose in the pen for short periods. The
based on general theoretical principles.
These will first be considered in connec- helpless mice are not used long enough to
tion with the problem of producing abe badly hurt. This procedure is repeated
vigorous fighter: (i) To train an animal daily, and on the fifth day dangling is
to fight, he must be affected by a primary followed by releasing an untrained,
stimulus or releaser, the easiest one ofstrange male in the pen. The trained
which to use is slight pain. (2) The fighter fighter will now attack and in almost ev-
must be stimulated by a variety of meth-ery case will win the fight without dif-
ods, or by different animals which behave ficulty. After this time, the trained
differently, as it has been observed that fighter is given a warmup period before
when the same method or animals are each fight but no further training is nec-
used over and over again the intensityessary.
of He should not be fought oftener
fighting tends to diminish. This may bethan every other day, and longer periods
of
explained by the fact that a stimulus is arest will not be harmful.
change and that a repeated stimulus isFighters trained in this fashion be-
less of a change and, furthermore, per- come so ferocious that, as Kahn (1949)
mits the animals to adjust to it with de-has shown, they will attack females and
creasing effort. (3) Animals become infant mice, which they would ordinarily
fatigued after violent fights and do not leave alone. Kahn's experiments seem to
usually fight vigorously if they doshow it that they attack the juvenile mice
every day or more frequently. (4) The even more vigorously than the older ani-
mals, probably because the young ones
animals also fight more vigorously if they
have a sort of warmup period and are give more violent escape reactions. An-
other interesting fact is that fighters
thoroughly excited before the actual fight
trained in this way which have been uni-
starts. (5) The fighter must be successful
in the fight and not be beaten, as defeat
formly successful show little, if any, pre-
teaches it to run away rather than liminary
to emotional reactions but attack
fight back. Repeated success leads to quickly, savagely, and efficiently (Scott,
strong motivation for fighting, whereas1947).
repeated defeat produces a strong habit Using such animals as these, it is easy
of escape (see Ginsburg and Allee, 1942).
to train naive animals to respond habit-
Using all these principles, the follow-
ually to a conflict situation with escape
or defensive behavior. A naive animal is
ing practical method was devised for de-
veloping a vigorous, persistent, and suc-
placed with the trained fighter, who has
cessful fighter. The mouse is first at-considerable advantage and usually wins
tacked by a trained fighter to which the
he first fight. The defeated animal usu-
will normally respond by fighting back.ally ceases all resistance after two daily
The two animals are separated before
3o-minute periods and thereafter runs

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
away at the sight of the trained experiment

fighterbeen done with training
or, indeed, of any other mouse which
trained is to be peaceful, but pre-
fighters
seen under the conditions of sumably this could be done, using the
the experi-
ment. same methods but with the expectation
The same principles may be used that
in a much longer period of training in
reverse in order to train naive animals to
restraint would be required than with
develop peaceful relationships (Scott,
inexperienced animals.
i944a). There the principal problem is to E. TRAINING ANIMALS TO BE TAME
avoid or overcome the effect of the pri-
mary stimuli or releasers. After this, pas- It is the common experience of ex
menters that repeated handling res
sive inhibition will do the rest. Male mice
are placed with females with whom they in gradually decreasing vigor of resp
ordinarily do not fight. Then, on fourso or that the animals may in a short
five successive tests the males are taken give little or no reaction and may b
out and given rough handling, i.e., areto be "tame." This may be interp
held by the tail with the forceps, placed as adaptive behavior resulting from
on a wire screen and firmly stroked withsolution to a problem. Escape and d
the finger. This has the effect of givingsive behavior are not effective and
the animals slight pain, but they are re- even result in the animal's hurting
strained from biting or effectively attack- self. The ineffective pattern of beh
is therefore abandoned in favor of one
ing. As soon as the handling is over, they
are placed back with the females, withwhich results in good adjustment, name-
whom they are already used to not fight-ly, a passive attitude on the part of the
ing. Considering this treatment as condi-animal toward the experimenter.
tioning, the animals are learning to as- One of the best experiments of this
sociate slight pain (normally the primary sort was done by Stone (1932), who
stimulus for fighting) and the conditionstested several genetically different groups
of rats and rated them for wildness and
associated with it, with not fighting.
If two males which have been treated savageness. The rates for all groups de-
in this way for 4 or 5 successive days arescend in what appear to be fairly typical
then placed together in a pen with one of learning curves, and rats in "wild" and
the females, no fighting results, either at "tame" strains are much more closely
the time or for as long as several weeksalike in the end than in the beginning.
afterward, which was the period of obser- Similar results were obtained by Coburn
vation. Presumably, the longer this(1922) with mice, and, indeed, the same
peaceful relationship lasts, the strongerresults have been obtained by any ex-
the habit of not fighting becomes. Inperimenter who has tried repeated tests.
short, if animals can be prevented fromIt may thus be tentatively concluded
fighting at the first meeting, the prin-that principles of complex learning apply
ciple of passive inhibition will tend toto agonistic behavior.
keep them from fighting subsequently. EXPERIMENTS WITH HEREDITY
Such animals can be trained to fight
A. THEORETICAL EXPECTATIONS
subsequently, using the methods de-
scribed above; but no effort has been The principles of transmission a
made to measure whether or not fightingphysiology of hereditary factors h
is as severe. Presumably, it would be been so thoroughly worked out in re
somewhat less severe. Neither has anytion to various morphological and phy

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
various techniques
ological traits that it is possible to pre-supports or denies
dict with considerable certainty the re-
these expectations. To begin with, a be-
sults of genetic studies ofhavior
agonistic be-
trait is probably one of the least
havior. favorable for the analysis of a precise
mode of inheritance,
i. Since even simple characters are af- first because the
measurement
fected by many genes (at least of behavior traits is always
17 known
genes affect coat color in the
time-consuming
mouse) and andmay take anywhere
from activity
since behavior traits involve several minutes
ofto even hours on
one individual
the whole body, it is almost a foregone and, second, because
(since all variabil-
conclusion that any hereditary genetic experiments depend
upon the
ity of agonistic behavior will beanalysis
caused of variability) large
by multiple genetic factors and
numbers thatneeded. The thor-
are always
variability in a random-bred population
ough study of the inheritance of a be-
will fall into a normal curve.
havior trait, therefore, becomes an ex-
2. Since the great majority of mutant
tremely laborious procedure; and, before
genes are recessive to their normal or such studies are made, everything pos-
wild-type allelomorphs, it would be ex-sible should be done to cut down the
pected that crosses between animals work, and serious consideration should
showing "wild" and "tame" behavior
be given to the question of whether the
characteristics would show dominance of
experiment needs to be done at all.
the "wild" kind of behavior in the first
One way in which time can be saved is
generation. Where crosses are made be-
to perfect simple and brief tests of be-
tween two "tame" types, the effect of the
multiple factors would probably be inter-
havior, so that each individual can be
mediate in the first generation. classified in a short period of time. Great
care should also be taken to cut down
3. Since many genes are known to have
variability by standardizing the environ-
effects on many characters, the possibil-
ity that known genes whose major effectment and care of the animals as closely
is on other characteristics may also af-as possible. The more the environmen-
fect behavior must be kept in mind. tally caused variability can be reduced,
However, such effects would be expectedthe easier it will be to study the genetic
to vary in different genic backgrounds.variability.
4. Since behavior consists of responses It should be remembered that the
to and adaptation to environmental basis of all genetic experiments is the
changes, it will be expected that undercontrol of genetic variability, and this
most situations environmentally causedmeans that the parental stocks must be
variability in behavior will be larger than
thoroughly known from the genetic point
genetically caused variability. This doesof view. Maximum control over variabil-
not modify the possibility that geneticity is achieved by the use of highly inbred
variability may be very important under strains of animals; and, in making an ex-
special conditions, either because ofperiment involving fighting behavior, the
thresholds or because of uniform environ-first step would be to make a brief survey
mental conditions. of many inbred strains and select two
which have the desired contrasting char-
B. METHODS AND EVIDENCE
acteristics. For the purpose of genetic
The experimental evidenceanalysis may now the best strains would be those
be examined to see how far the use of which showed small variability within

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
genesbetween
themselves and large differences which occur in inbred stocks and
the strains. which affect morphological characteris-
tics. The
The most highly controlled mode of inheritance of these
experi-
ment of all is one which involves a muta-
genes can be easily worked out; and, if
tion within an inbred strain. The mutant the stock is known to be pure and uncon-
gene can be kept heterozygous whiletaminated, it is very unlikely that a sepa-
close inbreeding is continued, so that therate mutation affecting behavior would
only genetic variability in the experiment occur at the same time. Keeler has made
is that caused by the gene. a survey of mutant strains derived from a
If only partially inbred strains arepartially inbred strain of wild-type Nor-
available, such as colony-bred animalsway rats and finds wide differences in be-
which have been kept a long time with-havior traits. He particularly stresses the
out outbreeding, variability can be some-association of tameness with the color
what reduced by using a small sampleblack. These results are unfortunately
from each colony, such as taking all par-not conclusive, since the original strain
ent-animals from a single litter. The con-was not highly inbred and the differences
trol of heredity achieved here will, ofin behavior can be accounted for by
course, be much less; and, unless the twochance selection of other variable genes.
colonies are known to be widely and dis- As a minor piece of negative evidence,
tinctly different, cross-breeding experi-in making a survey of differences in ag-
ments are probably not worth while. gressiveness in mice, Scott (1942) ob-
Another method which is sometimes served a mutant which had originated in
used for study of hereditary factors is one of the highly inbred C 57 black strains
that of selection of individuals into (C57 leaden) and found no appreciable
strains according to the traits which differences
they between the original strain
exhibit. Since the behavior trait may andbeits mutant. Comparison of the C57
influenced by many genes, selection black and C3H agouti strains gave no in-
without close inbreeding may result in
dication of the pronounced taming effect
selection of many genetic types of of the black (aa) gene which Keeler re-
indi-
ported in rats. With respect to aggres-
viduals which appear similar in behavior,
siveness toward other mice, the C3H
and it would be very likely that there
showed more initial aggressiveness, but
would be very little control over variabil-
Ginsburg and Allee (1942) found the C57
ity and hence that cross-breeding experi-
blacks superior fighters over the C3H
ments would be largely worthless. How-
agouti and C albino strains, both of
ever, the degree of separation between
which
the two strains gives a good measure of carry the AA genes.
the amount of hereditary variability This
in technique remains as a possibil-
the original stock. ity, but the ideal experiment is yet to be
Several studies which have been done done. From general considerations
on this subject may now be discussed in pointed out above, it might be expected
that any gene might have multiple ef-
the light of the above considerations. In
the first place, important data may be fects, some of which might concern be-
obtained without using cross-breeding havior. On the other hand, it would also
experiments, and the most time-savingbe expected that most of these effects
technique is that suggested by Keelerwould be minor in nature.
(1942, 1945) and Keeler and King (1942) It should be remembered at this point
of studying effects on behavior of mutantthat all methods of study are dependent

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
dures
upon raising the animals in or extremely
a uniform aggressive by experi-
en-
vironment, since all behavior traits may
ence in fighting; and it may be concluded
that thein
be largely affected by changes kind of strain
the en- differences which
vironment. This at once eliminates theappear is dependent upon the environ-
pedigree method as a possibility, since
ment and training of the animals con-
the maternal environment in different cerned. It may also be concluded that,
families cannot be kept constant. where maximum variability of the en-
A second short method for studyingvironment is permitted, training is far
heredity is to compare the behavior more
of important as a cause of variability
animals from different genetic subpopu-
between these inbred strains than is ge-
lations under environmental conditions netics. It should be remembered, of
course, that the differences observed are
which are as identical as possible. If care-
not so great as between wild and tame
fully done, this demonstrates that heredi-
strains.
tary differences exist; and these differ-
ences may be used in experimental work
Hall (1934) selected two strains of rats
for differences in what he called "emo-
for various reasons. However, in ap-
praising the results, consideration musttionality" in an open-field situation,
be given to the possibility of differences
which was measured by the frequency of
in maternal environment in the two pop- defecation and.urination. The descrip-
ulations which may be passed along from tion of his experiments indicates that he
generation to generation, for such a sur-
did not use close inbreeding and that, as
vey may show up cultural as wellusual as with most selection experiments,
genetic heredity. maximum differences were achieved at
Scott (1942) made a survey of highly about 7 or 8 generations. He then tested
inbred mouse strains raised under the the two groups of animals for differences
same conditions, studying aggressivenessin behavior in a fighting situation (Hall
as exhibited in the first fight. Since the
and Klein, 1942; Billingslea, 1941). The
animals were segregated before they be- rats were graded for aggressiveness on a
gan fighting, it was assumed that the scale modified from that of Davis (i933),
chance of cultural heredity was very
each rat being matched for 5 minutes
small. He found that under the condi-
with every other in a group of ten. Each
tions given, which involve frequent hu-
group contained five rats of each strain.
man handling, the C3H strain was dis-
Hall and his associates found that there
tinctly aggressive when matched against was a good correlation between what he
C albinos, whereas the C57 black strain called "aggressiveness' and between less
(subline io) was distinctly peaceful and timid behavior, as measured in the open
that there was almost no overlap between field. These results are interesting but
the two strains. Later Ginsburg and Allee raise many questions. Hall does not seem
(1942) studied the same strains with re- to have paid any attention to dominance,
spect to their ability to win fights (as dis-
and there seems to have been very little
tinct from the initiation of fights) and really active fighting in either strain.
found that the C57 blacks developed Much of what he called "aggressiveness"
into the best fighters, the C3H next best,
might also be considered simply fearless
and the C albinos the poorest. social investigation.
Later on, Scott (1944a) showed that
However, since the animals were kept
the C57 blacks could be rendered very
isolated in cages, it may be presumed
peaceful by appropriate handling proce-
that there was no cultural effect present,

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
and it may be concluded that an The limitations of the method are that
impor-
it gives an estimate of the number of
tant hereditary effect was demonstrated,
whatever its precise nature. Thegenes actually involved in the particular
correla-
tion between fearlessness in the cross open
rather than the total possible num-
field and aggressiveness is what ber which be
might can affect the character. Fur-
expected if there is a consistentthermore,
trait of if the parental stocks are only
aggressiveness running through many
partially homozygous, results can be ob-
situations; but this correlation tained
may,onlyof for the homozygous genes.
course, be caused by accidental Needless
selectionto say, strict control over en-
of two independent traits. vironmental conditions must be main-
tained in of
A more precise and difficult method successive generations.
genetic analysis may be employed, Unfortunately,
using only one such experi-
ment(origi-
as a basis Wright's (1934) formula has been properly made, since most
nally modified from Castle) forofthe
the esti-
work of this sort was done before
mation of the number of genetic Wright's
factors formula was developed. This
involved where variability is continuous:
particular experiment was that done by
A2 Dawson (1932), who measured wildness
N =8 (02F2
-02)P and tameness as measured by speed in a
runway when the mice were urged from
This formula and its variations are behind with a movable barrier. He used
based on comparing the variance in the as parent-strains laboratory stocks of
parental and various hybrid generations. wild and mutant mice. These were prob-
If two homozygous strains are mated ably to- somewhat inbred, but the amount
gether, the F, individuals will be heter- was not measured. The fastest animals
ozygous but identical, so that any varia- from the wild strains were mated with
bility in the parental or F1 generations the slowest from the tame strains, and,
must be due to the environment. Cross- through reciprocal crosses, F, and F, gen-
ing the F1's among themselves or makingerations were produced. The results indi-
a backcross to one of the parental strainscate that there were two or three genes
will increase the variability because ofinvolved and that the genes for wildness
segregation, and the amount of increasewere dominant. There were no differences
produced will depend upon the number between reciprocal crosses, and in the
of genes involved. It will be noticed that,wild strain the females appeared to be
in order for the formula to work, therewilder than the males. By selection Daw-
must be sufficient numbers in the paren-son was able to show that the average of
tal strains so that an accurate estimationthe tame strain could be altered, showing
can be made of the average differences,that these strains were not homozygous
and sufficient numbers must be raised inand that several more genes were actual-
subsequent generations so that an accu- ly present. In making the crosses, several
rate estimate of variance may be made. tame strains were crossed with the wild,
For a behavior trait the preferred experi-so that the results probably give a pic-
ment is one in which the F, is composedture of the average number of genes
of approximately equal numbers fromwhich were homozygous in these stocks.
reciprocal crosses. This allows for theT'his work definitely confirms the expec-
effects of maternal environment and doestation of multiple factors listed in the be-
away with the problem of differentialning of this section.
variability in the parental strains. Most of the other studies have been

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
made in such a way that the

nowild strain was not homozygous
statistical
whereas the
analysis is possible, but they may tame strain
be was inbred.
mentioned because of historical interest
Hence genetic analysis of the sort indi-
and because all of them tend to demon-
cated above cannot be made.
strate that some hereditary A similar
effect crossbreeding
was experiment by
present. Coburn (1922) in mice may also be men-
The original experiment of this type tioned. Here both parent-strains were
was done by Yerkes (1913), and his re- noninbred, and the result is that there
sults inspired others to attempt similar are no significant differences in variabil-
experiments. He crossed Castle's hooded ity between the F,'s and the F2's. Conse-
rat with wild animals and graded them quently, no conclusions regarding in-
for savageness, wildness, and timidity in heritance may be made, although the
reaction to human beings. The results study has other interesting points.
showed an F, which was extremely wild, The conclusion may be reached that a
savage, and timid, and demonstrated in- systematic study of the role of heredity
creased variability in the F2, but with in producing variability in agonistic be-
most animals near the timid grades. havior still needs to be made. Since such
Most of the animals in the F, were grade a study must inevitably be laborious, it
5 in savageness, whereas in the F, most probably should not be made without
of them were in grade o. The results are considerable preparatory work and should
peculiar from a hereditary point of view follow the following steps: (i) thorough
but indicate that possibly there was study under various standard environ-
dominance of the wild traits in the F, to- mental situations of large numbers of
gether with hybrid vigor and that a mul- two highly inbred strains known to be
tiple-factor ratio was apparent in the F,. widely different; (2) development of tests
Since the strains were not pure and since of temperamental traits in several social
small numbers were used in the parental situations, so that some conclusions can
strains, no conclusions can be drawn. be reached as to whether we have a single
The next study was that of Utsurika- underlying behavior mechanism which
wa (1917), who attempted to study the when modified will be part of all or many
results of outbreeding and inbreeding, situations, or several underlying mecha-
using Yerkes' tests as a model. Since he nisms adapted for particular situations
used an extremely small sample (14 ani- and modified independently; (3) cross-
mals in either strain) and made no breeding and measurement of animals in
crosses, no significance can be attached F, and F, or backcross generations; and
to the results, except that the material (4) analysis of results both by Wright's
was very variable. method and by intercorrelation of traits,
Stone (1932) made a cross between probably by factorial analysis.
Wistar albinos (presumably inbred) and In summary it may be pointed out
descendants of trapped wild rats. Only io that, while many of the data are not con-
animals were measured in the wild paren- clusive, all results are consistent with the
tal strain, which means that an accurate
hypothesis of traits affected by many
estimate of differences between the hereditary factors with dominant genes
means cannot be made. The F1, groups producing the wild-type traits of savage-
show a consistently greater variance than
ness and wildness and with the expecta-
the backcross to the tame strain. The re- tion that environmental influences are
sults can be attributed to the fact that very important. Final proof of multiple

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
effects on agonistic behavior byattack violently. On the other hand, mice

a single
which have been defeated once and are
gene is still not available. Experiments
on modification of aggressiveness andwith the opportunity of fight-
presented
savageness described in a previous sec- these symptoms before start-
ing exhibit
tion support the expectation of ing
thetoim-
attack. The conclusion was drawn
portance of environmental variability.
that these reactions are part of a mecha-
nism which comes into effect when the
EXPERIMENTS WITH PHYSIOLOGICAL
animal is equally stimulated to escape
MECHANISMS OF AGONISTIC
and attack and that other mice might
BEHAVIOR
learn to interpret this behavior as a
Agonistic behavior, as well as other
warning.
types, can be subdivided according These results
to the apply to one part of the
results of training into two parts important problem of what happens
(Scott,
1947). The first of these, which within may
the bodybe
after environmental con-
called "nonmodifiable behavior' for ditions are such that conflict behavior
want of a better term, is essentially follows.
re- The nonmodifiable behavior re-
flex in nature. Learning affects it by as to above largely includes the sorts
ferred
sociation only; and such a reaction of maybehavior which are loosely termed
be conditioned either to appear or not "emotional."
to It is suggested that the
appear in response to certain stimuli, process of positive and negative learning
may be made stronger or weaker, but is in male mice makes possible situations in
never essentially changed in nature. which
Ex- an animal can make no adaptive
amples of such reactions in mice arereactions.
tail- As a protection against this
possibility there apparently exists a
rattling and hair-fluffing, which appear
externally, and the various accompany- mechanism whereby nonmodifiable be-
ing internal reactions, such as changeshavior
inconnected with conflict behavior
the heart rate. may do at least two things: (a) produce
On the other hand, the second type ofexternal reactions which may act as sig-
behavior, which may be termed "modi-nals to other animals and (b) in extreme
cases may possibly take control over
fiable," not only responds to learning of
the simple association and inhibitionorgans normally used in modifiable be-
havior.
type but is much more variable and
adaptable to a wide variety of conditions. The latter possibility is not one which
has been observed in mice or rats but is
In this type of behavior, useless motions
can be eliminated and the efficiency ofone which stems from Masserman's
the response improved. For example, (1943) experiment of directly stimulating
such behavior as biting can be modifiedthe hypothalamus of cats, which "lost
so that an experienced fighter may learncontrol" over themselves to the extent of
striking and scratching, actions which
to disable his opponent within a few sec-
onds by biting in the right place. Suchare usually under "voluntary" control
modifications are not found in organs and belong to the class of modifiable be-
havior. The experiments of Cannon
which are controlled entirely by the
autonomic nervous system. (1929) and Masserman (1943) point to
It has been noted that mice which the existence of a center of nervous con-
trol working through the hypothalamus
have been trained to fight and have been
and, in turn, chiefly through the auto-
uniformly successful almost never exhibit
hair-fluffing and tail-rattling but at nomic
once nervous system, the adrenal me-

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
dulla, and various motor nerves

temperature which
and muscular tone, decrease
are normally "voluntary."
in bloodThis center
sugar, etc. These symptoms are
has two functions in addition to those reversed and body efficiency increased in
the phase of countershock, which may
listed above: (a) increasing the efficiency
of physiological reactions by altering follow
the as soon as a few minutes later.
This may account for the fact that
rate of breathing, heart rate, etc., and
(b) providing sensations which may act
trained fighters which are "excited" be-
as additional stimuli for modifiable before a fight are much more efficient than
havior, particularly if learning is those which are not.
in-
volved. These results have not been con- It may be postulated that these re-
firmed on mice and rats, but the anatomi-actions of shock and countershock should
cal and physiological characteristics of
also cause sensations within the body
mammalian species are similar enough to which may contribute to the type of be-
warrant the tentative extension of these havior which results. The later phases of
findings. "resistance" and "exhaustion" probably
The work of Selye (1946) on the rat affect behavior only in the case of long-
points to the existence of a second center continued and habitual fighting but may
of control in the hypophysis, which may have some relationship to recovery by
or may not be linked with the above and long-defeated mice.
which operates largely through the hor- Before further conclusions can be
mones of the adrenal cortex, which, in made specific, experiments need to be
turn, have profound effects on metabo- done on the relationship between the
lism, the circulatory system, and the "general adaptive syndrome" of Selye
kidneys. and agonistic behavior. Do all or only
This center of control reacts in some
part of the components of the syndrome
way to any injury to the body (probablyappear in connection with fighting? What
in reaction to substances' from injured are the physiological states of a winning
tissues) and also to fear and rage. In afighter, his defeated subordinate, or a re-
defeated mouse which has been injuredpeated loser? These and many other
by its opponent, both sorts of reactionsquestions are still to be answered.
would be expected to be present. From
the nature of its effects, it is obvious that DISCUSSION AND CONCLUSIONS
the center can affect conflict behavior
A. THE ADAPTIVE NATURE OF FIGHTING
chiefly in two ways, by altering the ef-
BEHAVIOR
ficiency of physiological processes and by
producing sensations which act as nerv- One general problem mentioned i
ous stimuli.s introduction is that of the adaptiv
Selye has described the results of ac- ture of social or intraspecific fight
tion by this center as a syndrome; and, mice and rats, as distinguished fro
during the initial or shock phase follow- terspecific fighting, which may
ing stimulation, body efficiency is low- sumed to be useful in predator-prey
ered by tachycardia, decrease in body tionships. Considered from the po
s The existence of this new evidence may explain
view of the individual, there is occ
some of the discrepancies in early experiments, as, ally some advantage in competitive
for example, the fact that the injection of adrenalin ing which enables the winner to
does not produce sensations of anger and that
sympathectomized animals may still appear to be
piece of food. However, the most im
emotionally aroused in conflict situations. tant type of social fighting in mi

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
rats appears to be noncompetitive Calhoun

in found
na- that where male rats
were not organized
ture, and its existence may therefore be into a stable domi-
the product of the artificial nance
conditions
order, groups of males would fol-
under which those animals are lowraised,
a femaleorin estrus and repeatedly and
it may be an injurious traitbriefly
which has Such females showed a
copulate.
relatively
accidentally arisen in evolution, or it maylow rate of conception (pos-
sibly
be a trait which has some use for thebecause
spe- no single male was ever
cies as a whole (Allee, 1942b). able
Thetolatter
reach a state of ejaculation). If
appears to be the most likely this observation is correct, fighting and
hypothesis;
dominance
for, wherever rats and mice live inwould lead to a more favor-
able results
g oups, some social organization condition for reproduction among
rats.
and fighting may be of use to the society
as a whole, even though it may On the negative
appear to side it may be stated
be merely destructive for anthat
individual.
competition over food and sex does
Calhoun (1949) has found not
withseemhis
to be an important cause of
studies of rats under seminatural condi- fighting. Competition over females does
tions that fighting behavior appears tonot seem to occur, and competition for
result in controlled use of living space.food does not appear except in special ex-
His rats had unlimited supplies of foodperimental situations. As Calhoun has
and water, but, whereas caged rats can pointed out, the so-called "hoarding"
live with only 2 square feet of living (really, food scattering) of rats actually
space per individual, his population lev-tends to make more food available for the
whole population, since these deposits
eled off at 50 square feet per individual.
It may be supposed that, if his rats hadare not guarded by the individuals who
been able to escape, many of those indi- make them.
viduals would have emigrated from the Further studies of the natural social
area and that the result of fighting in organization of mice are needed before
these animals is to maintain a rather the role of social fighting can be evalu-
ated, as has been done with Calhoun's
fluid and expandable type of social organ-
ization, which is always ready to make rat studies.
use of any vacant living space but which In the case of rats, and to a lesser ex-
tent in mice, fighting behavior is un-
will not permit more than a certain maxi-
doubtedly
mum density of population, even with an of use in fighting off predators
unlimited food supply. and capturing small animals for prey.
Scott has suggested that fightingThe
inconclusion of Collias (1944) that a
decrease of fighting outside a group is ac-
rats and mice may also tend to prevent
homosexual activity and to insurecompanied
dis- by an increase within a group
suggests that fighting within the species
tribution of the sexes through the living
space, so that fertile mating behaviormight
is be solely a misuse of behavior
easily achieved. The fact that female which is otherwise adaptive, and domi-
nance might be considered as merely a
mice are less likely to fight and are easily
dominated by the males makes it dif- primitive method of controlling the un-
ficult for them to reject the advancesdesirable
of aspects of a necessary behavior
trait. While this is a possibility which
males. It is therefore possible to assume
that part of the mechanism of fightingmust be kept in mind, the existence of
fighting behavior within the species will
within the species is primarily concerned
with insuring fertile matings. alter social organization, whether for use

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
or for harm. Further, if lelomimetic behavior (two or more ani-

interspecific
fighting alone were useful, there would same
mals doing the be thing at the same
no explanation for differingtime, often involving mutual imitation)
fighting
tendencies in the two sexes. compared to animals like sheep or dogs
In summary of these various points, it(Scott, 1950), group attacks upon an in-
dividual or fighting between groups can-
may be said that interspecific conflict be-
havior appears to be adaptive in preda- not be studied. It is not possible to study
tor-prey relationships. Conflict behavior
sibling rivalry or sexual jealousy as
causes of fighting, since these factors ap-
within the species (social fighting) ap-
pear to be absent or negligible. And, be-
pears to have two uses: to produce an ex-
cause of the relatively small size of both
panding and fluid type of society without
species, there are limits to what can con-
definite territories, which makes for the
veniently be done with physiological
optimum utilization of living space, and
analysis of fighting.
to insure fertile sexual behavior. There
On the positive side it should be
has apparently been selection toward pointed out that their relatively small
lessened aggressiveness on the part of size
fe- makes both species excellent for
males, so that fighting will not interfere
neurophysiological experiments. Small-
with male-female sexual behavior. Domi-
ness makes operations difficult, but it
nance, or the social relationship springing
makes subsequent microscopic examina-
out of habitual conflict behavior, has thetion of tissues much easier.
function of regularizing space utilization The two species are likewise ideal for
and sexual relationships, as well as mini-
experiments with training and heredity.
mizing the harmful and destructive as-Inbred stocks are available in both spe-
pects of social fighting. cies, and there is a particularly wide va-
riety in mice. As a glance at the extensive
B. RFLATIVE USEFULNESS OF THE TWO SPECIES
psychological literature on the subject
FOR EXPERIMENTAL ANALYSIS OF
will show, rats and probably mice are
CONFLICT BEHAVIOR
capable of associative learning and can
Enough work has been done solveupon
problems the of considerable complex-
conflict behavior of the two ity. species so
that it is possible to evaluate them in
As has been pointed out above, consid-
terms of usefulness for laboratory erable progressinves-
has been made along the
tigation on this subject. It must be con-
lines of analyzing the effects of variabil-
cluded that mice are probably better
ity in heredity and variability in train-
than rats for laboratory experiments ing; and, if the aboveon
limitations are kept
fighting. Most of the domestic strains
in mind, it may beareconsidered that a
more aggressive than the rats, and(though
fairly sound they by no means com-
are much easier to raise and control. plete) foundation has been laid for the
However, because of their larger size,study
rats of abnormal behavior which may
are probably better suited than mice arise
forin situations of conflict in the two
field experiments and for physiological
species. Furthermore, these studies may
analysis of fighting behavior. be considered a guide to the extension of
It may be pointed out that both spe-
findings to other species, so that even-
cies have definite limitations with regard
tually we may have the basic material for
to certain aspects of fighting. Because
sound generalizations regarding fighting
behavior.
these animals exhibit relatively little al-

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
C. 'IIIF PHYSIOLOGICAL MECHANISMS
Impulses from the same or perhaps
UNDERLYING CONFLICT
another center activate the anterior pitu-
BEHAVIOR
itary, which, in turn, sends hormones to
We may now summarize the basic
the adrenal cortex. Hormones released
mechanisms which appear to lie from behind
the area modify the action of the
fighting behavior in the two species,
circulatorycon-system and produce changes
sidering first the mechanism involved
in metabolism in which result eventually in
reacting to an immediate situation
increased andphysiological efficiency for
then the mechanism through whichdealing pre-
with an emergency and probably
vious conditions affect behavioralso(see Fig.
in sensations which are returned to
i). For simplicity, the discussion
the will be cortex and may set up new
cerebral
limited to a noncompetitive situation.
impulses there.
Fighting behavior appears to startor not these activities are
Whether
with external stimuli or changes underby the same center which con-
controlled
specific conditions. These involve themediated through the hypo-
trols those
presence of a strange animal, the
thalamusof
pain and autonomic system is not
an attack, or the sight of another animal
known. Whether either group of reac-
running away. These stimuli affect
tions isthe
unitary is likewise a matter of
cutaneous nerve endings and the mecha-
doubt.
nism of sight' and then are carried
Thistobasic
the mechanism may be modi-
cerebral cortex and probably toby
fied some
other external factors. Limited
sort of lower controlling centerfood (it should
supplies may cause two animals to
be emphasized that these neural centers
compete, leading to pain and thus initiat-
and much of the internal mechanisms are
ing the mechanism of fighting. Physical
hypothetical and derived from knowl-factors such as heat may set up compet-
edge of other species). Nerve impulsesing adaptive reactions which tend to
from the cerebral cortex result in the
slow down or reduce severity of fighting.
modification of the voluntary part Variations
of in space may cause two ani-
mals to choose different reactions which
fighting, such as scratching, running, and
biting. are adaptive to the situation.
Impulses from the lower center may Hypothetical as this schematic system
have many functions, all affecting non- is in many respects, one very important
modifiable behavior. Some of these act
point stands out: There is no evidence
through the hypothalamus to set up
for any sort of spontaneous internally
adaptive responses in the heart, adrenal
arising stimulation which would cause a
glands, and other parts of the body; to"need for fighting" per se. Instead, we
produce external signaling behavior,
have a mechanism which will produce
such as tail-rattling and hair-fluffing; and
fighting in response to specific and pre-
under extreme conditions to cause auto- dictable external stimulation. Even the
matic biting and scratching of the reflexmale hormone apparently produces dif-
nature. Sensations of the physiologicalferences in response levels between indi-
changes in the body may be carried toviduals rather than creating stimulation
the cerebral cortex and set up associa-within an individual (see Fig. 2).
tions with fighting and probably give This conclusion apparently conflicts
rise to new impulses. with the possibility stated by one of the
6 Sight is not necessary for fighting to take authors in an earlier paper (Fredericson,
place,
according to Beeman. 1946), that "aggressiveness stems from a

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
ENVIRONMENT Ce ~Wc
OC Cefter-
Social Ie Lmov e
o . Hyyp
(PO.%n o0 LCoA)
ri Ani awaY
Prey"Phy ic4 1
0-F0
Space
Heo
FIG. .--Mechanisms affecting the origin of fighting behavior, particularly of the noncompetitive type.
Nervous stimulation is shown in solid lines, hormonal influences in dotted lines. All details are not shown
because of lack of space; but it is evident that there is no place in the diagram where spontaneous internal
activity might arise to create a need for fighting per se.

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
H EREDITAR Y VARIABI LITY
Sex Ch rornosom0 s Auto somes

XX XY (mrlti ple foctors)
AssociAtive learninhg Problem so

Cposif've or negai.tve) (success orf
E.yrl)y tra nin5 , Later traein

(protected or (oaffack, escape,
attacKe.d) or peacefLl)
I ENVIRONMENTAL VAR
FIG. 2.-Mechanisms affecting variability and consistency in agonistic
the major sources of variability in the behavior of animals tested under si

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
basic energy source in the organism." In-

male hormone produces this effect is still
deed, if a similar diagram were
unknown,madeand the
offield is open for de-
the causes of fighting in a tailed experiments on the local or general
competitive
situation, one of the arrows action
wouldof thearise
sex hormone. Does it in-
from a state of hunger within creasethepain sensitivity? Or general activ-

animal
and presumably from the physiological ity? Or does it affect the central nervous
changes accompanying this state. system itself?
However, the state of hunger On itself
the otherishand,
a all available evi-
result of a lack of available food in the dence indicates that there are multiple
hereditary factors carried on the auto-
external environment, nor does it neces-
somes which cause variability in agonis-
sarily result in fighting. It must be con-
tic behavior. Here again the exact mech-
cluded that any analysis of the causes of
anism through which this variability
fighting on the basis of external changes
acts is unknown. The behavior of C3H
and immediate internal responses, on the
and C57 black strains of mice suggests
one hand, and more remote and slow
that heredity in some way affects the re-
physiological changes, on the other,sponse
is to a balance between positive and
purely an artificial one and that the most
negative training. The experiments on
realistic analysis is one which considers
hereditary differences in tameness, which
the problem as an interaction process be-
is really a response to negative training
tween external and internal changes. Oneto fight, indicate the possibility that the
particular situation may cause hunger balance
to between the control by the cere-
be very important and external changes bral cortex and that by the more auto-
negligible in the causation of fighting,matic lower centers has been altered so
and another situation produce a reverse that the responses of the latter are weak-
ened. Richter (1950) has suggested that
effect. In either case training has an al-
most overwhelming modifying effect. the adrenal cortex is involved. Further
and more precise studies of effects of
D. MECHANISMS AFFECTING VARIABILITY AND
heredity will be necessary before it will be
CONSISTENCY IN CONFLICT BFHAVIOR
(FIG. 2)
possible to evaluate its physiological
effects more precisely.
In addition to the effects of environ- It may be added that the effects of
mental stimuli, the activity of the mech- small hereditary factors may be enor-
anism can be importantly altered by two mously magnified by training. For ex-
sorts of previous conditions:heredity and ample, the outcome of the first fight is
training, the latter of which is the result very important in determining domi-
of reacting or not reacting to the previous nance, and an animal such as a C3H
environment. Considering hereditary mouse, which is less vigorous and more
variability first, the most precise mecha- hesitant, may well lose this first fight to a
nism known is that involving sex. Fac- C57 black and continue to lose on subse-
tors on the sex chromosomes cause a dif- quent occasions, even though the initial
ferentiation between male and female behavior differences were relatively
hormones which may produce an enor- slight.
mous difference in the degree of response Regarding the effects of training itself,
to environmental stimulation. Animals there are at least two types of learning:
possessing the male hormone are, of that involving simple association (posi-
course, very easy to stimulate. How the tive or negative learning) and that in-

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
volving the solution of problems. most any physiological change short of

Simple
associative learning can affect both death. ac-
The greatest known differences
tivities under the control of the lower produced by heredity can be duplicated
centers and those under the control of the or exceeded by differential training ap-
cerebral cortex; and, as the studies of plied to a single genetic type of animal.
Ginsburg and Allee (1942) and Seward Important as other factors may be under
(1945) have brilliantly established, the special situations, it must be concluded
same principles apply here as those which that the most important cause and hence
Pavlov derived from his study on the the most important factor of control of
salivary reflex. Theirs was the first dem- fighting and agonistic behavior in rats
onstration that the laws of associative and mice is training, which includes the
learning apply to a major basic type of interaction of the two processes of associ-
social behavior, as was anticipated byative learning and problem-solving.
Miller and Dollard in their study of imi-
SUMMARY
tation (1941), and may be considered as
a model for extending this principle to I. Fighting may be considered one of a
other types of social behavior. Associa- group of related behavior patterns which
tive learning can cause an animal to bemay be termed "agonistic."
consistently aggressive or consistently 2. This paper is an attempt to organize
timid and fearful. This is one of the mostthe available data concerning the factors
noticeable effects of training, for inex-which affect agonistic behavior in two
perienced mice facing a conflict situationfairly closely related species, mice and
for the first time show extremely variable rats.
and unpredictable behavior. 3. Agonistic behavior patterns are de-
The process of problem-solving ac-scribed in detail, and a brief outline of
counts for this preliminary variability,important experimental methods is given.
and this appears to be its most important 4. In "noncompetitive" situations
effect. The animal simply tries out in suc- (i.e., not involving fighting over objects),
cession the various simple adaptive pat- developmental study indicates that the
terns of behavior which are peculiar to male hormone and pain are important
the species. No great complexity of be- factors affecting the initial appearance of
havior is involved, and associative learn-fights. Pain, noise, sudden movements,
ing or habit formation fixes the reactionand restraint have been used to produce
which gives most successful adjustment.agonistic behavior in response to direct
The evidence cited in this paper pointshuman stimulation.
overwhelmingly to the conclusion that 5. In competitive situations the impor-
training includes by far the most impor-tant causal factors appear to be states of
tant group of factors which affect agonis- hunger and possibly thirst and the degree
tic behavior. Passive inhibition causes to which fighting is an effective response
mice raised together to be peaceful in-in a given situation. Thus frustration
stead of combative; and previous successtends to give aggression only in situa-
or defeat makes the difference between ations in which aggression provides a sat-
mouse which will attack all comers and isfactory solution to a problem.
one which runs away from any other ani- 6. Experiments with experienced or
mal. The effects of primary stimuli or re- trained animals have been largely done
leasers can be overcome by training, with animals which have developed a
which also withstands the effects of al- dominance order when fought by the

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
round-robin technique. Cold

would behas no ef-
dominant in the first genera-
fect within the temperature ranges
tion; that pleiotropic and interaction ef-
tested; heat appears to inhibit aggres-
fects would be observed; and that in most
siveness. Increased space modifies
situations the
environmentally controlled
variability would
patterns of behavior and permits exceed genetic varia-
a vague
sort of territorialism. bility.
7. Where dominance is already I5. Methodsestab-
for the genetic study of
lished, competition among behavior
hungry are described,
ani-and it is con-
mals for food results in cutting cluded that down
an ideal experiment
the should
usual amount of fighting.include, It is among other things, a wide vari-
concluded
that frustration as a cause of ety of tests covering various
aggression is situations.
important only when animals 16. Allhave
results sobeen
far obtained agree
trained to react aggressively with the expectation
after frustra-in paragraph 14. In
tion. addition, the kinds of genetic difference
8. Avitaminosis and other nutritional obtained vary considerably with the na-
deficiencies may reduce the amount ture of of the test situation, indicating that
fighting of a dominant animal but have the traits involved may be very specific.
little effect on dominance, short of death. I7. Two controlling centers of physio-
9. Male mice trained by the round- logical reactions connected with agonistic
robin technique and then castrated con- behavior appear to exist-a nervous cen-
tinue to fight, indicating that, once the ter in the hypothalamus and an endo-
habit of fighting is established, the com-crine center in the pituitary body.
plete supply of male hormone is not nec-18. Fighting behavior which is not
essary to maintain fighting (some andro- object-centered has two adaptive effects:
genic hormone may be furnished by the to regulate living space and to insure fer-
adrenal glands). tile sexual behavior. The adaptive nature
io. Both males and females will com- of object-directed fighting is obvious.
pete for food, indicating that the male Dominance has the effect of regularizing
hormone is of little significance in thisspace utilization and sexual relationships,
as well as minimizing the harmful and
situation. Hunger is not necessary to pro-
duce fighting over food, once the habit destructive aspects of social fighting.
has been established. 19. The relative usefulness of rats and
i1. Training in early life produces im-
mice for research on fighting is discussed.
portant effects on fighting tendencies in 20. The network of causal relation-
later life, whether or not object-competi-
ships through which fighting is produced
tion is involved. is summarized. In some situations exter-
12. The principles of learning are de- nal changes are very important; in others
scribed; experimental data concerninginternal changes may be very important.
dominance confirms their application toNo center of internal change has been
fighting behavior in both mice and rats.discovered which might create a "need"
13. Based on these principles, methods for fighting per se.
are described for training mice to fight, 21. It is concluded that training in-
run away, or be peaceful. cludes by far the most important group
14. It would be theoretically expected of factors which affect consistency in
that heredity would affect fighting be- agonistic behavior, although hereditary
havior through multiple factors; that factors (and particularly the male hor-
factors for "wildness" and "savageness"mone) may be very important in certain

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
situations and various environmental tant in determining whether or not fight-

ing is started in the first place.
and physiological factors may be impor-
LITERATURE CITED
ALLEE, W. C. 1942a. Group organization among

experience upon adult behavior. Jour. Abnorm
vertebrates. Science, 95: 289-93. and Social Psychol. (in press).
---. 1951. Aggressiveness in female mice. Jou
---. 1942b. Social dominance and subordination
Comp.
among vertebrates. Biol. Symposia, 8: and Physiol. Psychol. (in press).
139-62.
GINSBURG,
BEEMAN, E. A. 1947a. The effect of male hormoneB.,
on and ALLEE, W. C. 1942. Some effect
of conditioning
aggressive behavior in mice. Physiol. Zo6l., on social dominance and sub
20:373-405. ordination in inbred strains of mice. Physio
Zoi61., i5:485-506.
------. 1947b. The relation of the interval between
GUTHRIE, E. R. 1935. The psychology of learnin
castration and first encounter to the aggressive
New York: Harper & Bros.
behavior of mice. Anat. Rec. (abstr.), 99:570-71.
BEEMAN, E. A., and ALLEE, W. C. 1945. SomeHALL,
effectsC. S. 1934. Emotional behavior in the rat.
of thiamin on the winning of social contactsDefecation
in and urination as measures of indi-
mice. Physiol. Zo6l., 18:195--221. vidual differences in emotionality. Jour. Comp.
Psychol., 18:385-403.
BILLINGSLEA, F. Y. 1941. The relationship between
emotionality and various other salients---.of 1941.
be- Temperament: a survey of animal
havior in the rat. Jour. Comp. Psychol., studies.
31: Psychol. Bull., 38:909-43.
69-77. HALL, C. S., and KLEIN, S. J. 1942. Individual
BRUCE, R. H. 1941. An experimental analysis of differences in aggressiveness in rats. Jour. Comp.
social factors affecting the performance of white Psychol., 33:37I-83.
rats. Jour. Comp. Psychol., 31:363-412. HILGARD, E. R. 1948. Theories of learning. New
CALHOUN, JOHN B. 1948. The development and role York: Appleton-Century-Crofts.
of social status among wild Norway rats. Anat. HILGARD, E. R., and MARQUIS, D. G. 1940o. Con-
Rec. (abstr.), Ioi:694. ditioning and learning. New York: Appleton-
---. 1949. Influence of space and time on the Century Co.
social behavior of the rat. Ibid. (abstr.), 105:508. HULL, C. L. 1943. Principles of behavior. New York:
---. 1950. The study of wild populations under Appleton-Century Co.
controlled conditions. Ann. New York Acad. Sci., KAHN, M. W. 1949. The effect of severe defeat at
51:1113-22. various age levels on the aggressive behavior of
CANNON, W. B. 1929. Bodily changes in pain, mice. Jour. Genet. Psychol. (in press).
hun-
ger, fear and rage. 2d ed. New York: Appleton KEELER, C. M. 1942. The association of the black
Co., 1929. (non-agouti) gene with behavior in the Norway
COBURN, C. A. 1922. Heredity of wildness and rat. Jour.
sav- Heredity, 33:371-84.
ageness in mice. Behavior Mono., 4:1-71. ---. 1945. Gene determined physical modifica-
COLLIAS, N. E. 1944. Aggressive behavior tions amongaffecting the sociometry of rats. Sociom-
vertebrate animals. Physiol. Zoil., 17:83-123. etry, 8:30-41.
KEELER,
CREW, F. A. E., and MIRSKAIA, L. 1931. The C. M., and KING, H. D. 1942. Multiple
effects
of density on an adult mouse population. effects
Biol.of coat color genes in the Norway rat,
generalis, 7: 239-50. with special reference to temperament and
DANIEL, W. J. 1943. An experimental note domestication.
on the Jour. Comp. Psychol., 34:241-
O'Kelly-Steckle reaction. Jour. Comp. Psychol.,
50.
35:267-68. MASSERMAN, J. H. 1943. Behavior and neurosis.
Chicago: University of Chicago Press.
DAVIS, F. C. 1933. The measurement of aggres-
sive behavior in laboratory rats. Jour. Genet.
MILLER, N. E., and DOLLARD, J. 1941. Social learn-
Psychol., 43:213-17. ing and imitation. New Haven: Yale University
DAWSON, W. M. 1932. Inheritance of wildnessPress. and
tameness in mice. Genetics, 17:296-326.O'KELLY, L. I., and STECKLE, L. C. 1939. A note
on long-enduring emotional responses in the rat.
FREDERICSON, E. 1946. The theory of psychomotion
as applied to a study of temperament. Jour.
Jour.Psychol., 8: 125-31.
Comp. Psychol., 39:77-89. RICHTER, C. P. 1950. Domestication of the Norway
rat and its implications for the problem of stress.
---. 1949. Response latency and habit strength
In:C57
in relationship to spontaneous fighting in WOLFF, HAROLD G., et al. (eds.), Life stress
black mice. Anat. Rec. (abstr.), Io5:509.and bodily disease, chap. ii, pp. 19-47. (Assoc.
---. 1950a. The effects of food deprivation forupon
Research in Nervous and Mental Diseases,
competitive and spontaneous combat in Vol.
C5729.) Baltimore: Williams & Wilkins Co.,
black mice. Jour. Psychol., 29:89-1oo. 1950.
SCOTT, J. P. 1942. Genetic differences in the social
--. 1950b. Competition: the effects of infantile

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00
II. An Heredity,
behavior of inbred strains of mice. Jour. attempt to establish a dominance hier-
33:11-15. archy; III. The role of frustration; IV. Submis-
------. 1944a. An experimental test of the theory sion as determined by conditioning, extinction,
that social behavior determines social organiza- and disuse. Jour. Comp. Psychol., 38:175-97,
tion. Science, 99:42-43- 213-24, 225-38; 39:51-76.
---. 944b. Social behavior, range and terri- SKINNER, B. F. 1938. The behavior of organisms.
toriality in domestic mice. Proc. Indiana Acad. New York: Appleton-Century Co.
Sci., 53: 188-95. SNELL, G. D. 1941. Reproduction. In: SNELL, G. D.
---. 1946. Incomplete adjustment caused by (ed.), Biology of the laboratory mouse, chap. ii.
frustration of untrained fighting mice. Jour. Philadelphia: Blakiston Co.
Comp. Psychol., 39:379-90. STONE, C. P. 1932. Wildness and savageness of rats
---. 1947. "Emotional" behavior of fighting in different strains. In: LASHLEY, K. S. (ed.),
mice caused by conflict between weak stimu- Studies in the dynamics of behavior. Chicago:
latory and weak inhibitory training. Jour. Comp. University of Chicago Press.
and Physiol. Psychol., 40:275-82. TOLMAN, E. C. 1932. Purposive behavior in animals
---. 1948. Dominance and the frustration aggres- and men. New York: Appleton-Century Co.
sion hypothesis. Physiol. Zoil., 21 :31-39. UHRICH, J. 1938. The social hierarchy in albino mice.
---. 1950. The social behavior of dogs and Jour. Comp. Psychol., 25:373-413.
--. 1940. Effect of experience on fighting be-
wolves: an illustration of sociobiological system-
atics. Ann. New York Acad. Sci., 51:oo1009-2 havior of albino mice. Ecology, 21: 100-101.
SCOTT, J. P., and MARSTON, M. 1946. Effect of UTSURIKAWA, N. 1917. Temperamental differences
temperature on the amount of fighting between between outbred and inbred strains of the albino
pairs of dominant and subordinate mice. Un- rat. Jour. Anim. Behavior, 7: 111-29.
published MS. WRIGHT, S. 1934. The results of crosses between
SELYE, H. 1946. The general adaptation syndrome inbred strains of guinea pigs differing in number
and the diseases of adaptation. Jour. Clin. of digits. Genetics, 19:537-51.
Endocrinology, 6: 117-230. YERKES, R. M. 1913. The heredity of savageness
SEWARD, J. P. 1945. Aggressive behavior in the rat. and wildness in rats. Jour. Anim. Behavior,
I. General characteristics;age and sex differences; 3:286-96.
FACTORS INFLUENCING THE STATE OF DISPERSION OF THE

DERMAL MELANOPHORES IN RAINBOW TROUT*
O. H. ROBERTSON
Department of Medicine, University of Chicago, Chicago, Illinois, and Dep

of Biological Sciences, Stanford University, Stanford, California
T IS a common observation that trout cidental observations revealed the fact
I not only vary in color according tothat these fish exhibit a wide variety of
the light intensity of their home en-responses to different stimuli with respect
vironment but also become paler or to rate and degree of color change. Fur-
darker as they move to backgrounds ofther study soon indicated that adequate
greater or lesser illumination. However,interpretation of such melanophore reac-
the rate at which such changes in shadetions, especially those of isolated skin
occur has apparently received little study,
strips, would necessitate the acquisition
notwithstanding the interesting implica- of much more detailed knowledge of the
tions, both ecological and physiological.conditions affecting the state of disper-
During the course of previous investiga- sion of these cells in the trout. Although
tions on the transformation in color of the literature contains extensive investi-
rainbow trout (Robertson 1948, 1949) in-
gations on the effect of different environ-
mental conditions and chemical com-
* This work was aided by grants from the Ameri-
pounds on the activity of a variety of tel-
can Academy of Arts and Sciences and the Ella
Sachs Plotz Foundation. eost melanophores (reviewed by Parker,

143.167.200.46 on Tue, 27 Feb 2024 18:12:35 +00:00

Scott CausesFightingMice 1951

Uploaded by

Copyright:

Available Formats

Scott CausesFightingMice 1951

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Scott CausesFightingMice 1951

Uploaded by

Copyright:

Available Formats

The Causes of Fighting in Mice and Rats

Author(s): J. P. Scott and Emil Fredericson

Stable URL: https://www.jstor.org/stable/30152137

This content downloaded from

THE CAUSES OF FIGHTING IN MICE AND RATS'

J. P. SCOTT AND EMIL FREDERICSON2

INTRODUCTION various species of animals have various

This content downloaded from

stead of straight-line hierarchies?

AGONISTIC BEHAVIOR If two mice attack at once, they

This content downloaded from

behavior as an attitude of defense. If a When any two animals are repeatedly

This content downloaded from

by social dominance as soon ashave

This content downloaded from

this situation the experimenter

CONFLICT BEHAVIOR APPEARING IN VARIOUS SITUATIONS TO WHICH

Possession Possible Possession Impossible

Primary situation Violent fighting di- Fighting directed Either no conflict

Situation repeated Behavior associated Behavior associated Same as above

unnecessary. True fighting behavior is A. FIGHTING IN NONCOMPETITIVE

This content downloaded from

This content downloaded from

able to get fighting, butfor

This content downloaded from

animals which were running hormone,

This content downloaded from

the frustration of a nonaggressive males is placed with

This content downloaded from

This content downloaded from

This content downloaded from

This content downloaded from

than upon strongly formed habits

fighting. Since pain (as pointed out in an

This content downloaded from

volved, has an important effect upon

This content downloaded from

man (1932) pointing out thattion and

This content downloaded from

nate one escapes,

This content downloaded from

The first thing

This content downloaded from

24 hours or longer ones up evidence for dominance and subordina-

This content downloaded from

is associated with competition

This content downloaded from

away at the sight of the trained experiment

This content downloaded from

This content downloaded from

This content downloaded from

This content downloaded from

This content downloaded from

made in such a way that the

This content downloaded from

effects on agonistic behavior byattack violently. On the other hand, mice

This content downloaded from

dulla, and various motor nerves

This content downloaded from