THE LIBRARY OF

YORK

UNIVERSITY

ANIMAL MOTIVATION

EXPERIMENTAL STUDIES ON
THE ALBINO RAT

Prepared under the auspices of
COLUMBIA UNIVERSITY COUNCIL FOR RESEARCH
IN THE SOCIAL SCIENCES

ANIMAL MOTIVATION

EXPERIMENTAL STUDIES

ON

THE ALBINO RAT

BY

C. J. WARDEN

THE ANIMAL LABORATORY, DEPARTMENT OF PSYCHOLOGY
COLUMBIA UNIVERSITY

WITH THE COLLABORATION OF

T. N. JENKINS MARION JENKINS

L. H. WARNER E. L. HAMILTON

H. W. NISSEN

New York
COLUMBIA UNIVERSITY PRESS
1931

Copyright, 1931
COLUMBIA UNIVERSITY PRESS

Published March, 1931

Printed by The Torch Press
Cedar Rapids. Iowa

ACKNOWLEDGMENTS

To the Council for Research in the Social Sciences of Columbia
University, generous thanks are due for their continued support of
this project, extending to the publication of the final report in its
present form. The project could not have been brought to success-
ful completion without the hearty support, from year to year, of
Dean F. J. E. Woodb ridge, of the Graduate Faculties, and Profes-
sor A. T. Poffenberger of the Department of Psychology.

As a fitting recognition of their valuable contributions, the names
of the five student co-workers who labored faithfully in carrying
out the several major assignments have been placed upon the title
page as collaborators, in the order in which they became connected
with the project. The project owes its success in no small way to
the willing cooperation of this group, and to the earlier work of
Frances Holden, whose monograph has been included in the appen-
dix. Miss Mercy Aylesworth deserves mention in connection with
the study of motivation in discrimination work, which will also be
found in the appendix.

Many valuable suggestions were offered by Professor R. S. Wood-
worth and Professor A. T. Poffenberger, of this department, relative
to the preparation of the material for publication in the present
volume.

C. J. Warden

PREFACE

In the spring of 1925, the writer submitted to the newly organized
Council for Research in the Social Sciences of Columbia University
the tentative plan of a comprehensive experimental investigation
of animal motivation to extend over a period of several years. The
plan called for the measurement and comparison of the reaction
tendencies or drives associated with hunger, thirst, maternity, vari-
ous sex conditions, etc., by means of the Columbia Obstruction
Method, which had been devised by T. N. Jenkins and C. J. Warden
in 1924, and was at the time being used by Holden in connection
with her doctorate research. It was proposed to use the white rat,
at first, in working the field over systematically and then, if pos-
sible, to duplicate "with the monkey and other higher types of
animal," such parts of the work as offered the most promise.

The plan covering the contemplated work on the white rat was
approved by the Council and an initial grant of $1000.00 for 1925-
26 became available on July 1, 1925. A like amount was provided
for the continuation of the work through the following year ; the
sum of $2000.00 was voted for further experimentation as planned
for 1927-28, and a final grant of $1000.00 made to cover the 1828-29
budget. Of the allotment for 1928-29, one-half was provided to
cover the cost of such clerical and statistical work as might be
necessary to bring the results of the various studies together (Part
VII) and to prepare the final report for publication in the present
volume. The Council also generously assumed financial responsibil-
ity for the printing of this report. A summary of expenditures for
the entire period shows the disbursement of the $5000.00 grant to
be as follows: (1) for animals, $3,173.14; for apparatus and equip-
ment, $752.13; (3) for food, bedding, etc., $647.71; (4) for clerical
and statistical help, $427.02.

As will be seen, the above fund covered only the cost of the neces-
sary animals and materials, and made no provision for stipends

vii

viii

PREFACE

for those who should care for the animals and do the actual experi-
mental work. From the beginning, it had been planned to make this
project on animal motivation the major line of research in the
departmental animal laboratory and to assign specific topics to
approved graduate students, who should work under the super-
vision of the writer. Under this arrangement the animals were to
be cared for, in the main, by the regular laboratory assistant and
the data taken by approved students who should be given assign-
ments. This plan has proven to be a wise economy and has worked
out satisfactorily in every way. In most cases the first topics
assigned were doctorate problems, and by encouraging each to take
additional assignments it has been possible to limit the experimental
work to a few mature and carefully selected students. By training
each worker thoroughly in connection with his first assignment, the
high level of uniformity in procedure necessary to warrant the
widest interpretation of the results, taken as a whole, was main-
tained.

Much of the material of this volume has already appeared from
time to time in the form of separate papers and monographs. In
reprinting these various publications it was found desirable, in some
cases, to modify the introductory portions so as to avoid needless
repetition. Since the general method had been given detailed treat-
ment in Part I, all diagrams and descriptions of the apparatus were
deleted from papers included in the other parts. Aside from such
modifications, the papers and monographs have been reprinted in
full.

The author is under special obligations to the editor and pub-
lishers (The Williams and Wilkins Company, Baltimore, Md.) of
The Journal of Comparative Psychology and Comparative Psychol-
ogy Monographs for permission to reprint Part I, 1 and 2 ; Part II,
1 ; Part IV, 1 ; and Appendices 1, 2 and 4 ; and to the editor and
publishers (Clark University Press) of The Journal of Genetic
Psychology and the Genetic Psychology Monographs for permission
to reprint chapters Part II, 2 ; Part III, 1 ; Part IV, 2 and 3 ; Part
V, 1 ; Part VI, 1 ; and Appendix 3. The short introductory sections
are, of course, new, and are intended to orient the reader with

PREFACE ix

respect to the general topic of each main part. The opening section
as well as Part VII, 1, in which the results are analyzed and inter-
preted, is also new.

C. J. Warden

CONTENTS

PART I

The Columbia Obstruction Method C. J. Warden 3

1. Standard Apparatus for the Study of Animal Motiva-
tion T. N. Jenkins, L. H. Warner, and C. J. Warden 17

2. An Experimental Analysis of the Obstruction Method

of Measuring Animal Drives

C. J. Warden and H. W. Nissen 34

PART II

The Hunger Drive C. J. Warden 53

1. A Study of Hunger Behavior in the White Rat by
Means of the Obstruction Method L. H. Warner 56

2. The Effect of Delayed Incentive on the Hunger Drive
in the White Rat (Experiment 1: The Obstruction
Method) E. L. Hamilton 81

PART III

The Thirst Drive C. J. Warden 99

1. A Study of Thirst Behavior in the White Rat by

Means of the Obstruction Method L. H. Warner 100

PART IV

The Sex Drive C. J. Warden 117

1. A Study of Sex Behavior in the White Rat by Means

of the Obstruction Method L. H. Warner 119

2. The Effect of Segregation on the Sex Behavior of the

White Rat as Measured by the Obstruction Method

Marion Jenkins 179

xii

CONTENTS

3. The Effects of Gonadectomy, Vasotomy, and Injections

of Placental and Orchic Extracts on the Sex Behavior

of the White Rat H. W. Nissen 262

PART V

The Maternal Drive C. J. Warden 333

1. A Study of Maternal Behavior in the White Rat by

Means of the Obstruction Method H. W. Nissen 334

PART VI

The Exploratory Drive C. J. Warden 353

1. A Study of Exploratory Behavior in the White Rat by

Means of the Obstruction Method H. W. Nissen 354

PART VII

A Comparison of Normal Drives C. J. Warden 371

1. The Relative Strength and Persistence of the Normal

Drives in the White Rat C. J. Warden 372

APPENDICES

Introduction C. J. Warden 401

1. A Study of the Effect of Starvation upon Behavior by

Means of the Obstruction Method Frances Holden 403

2. The Effect of Short Intervals of Delay in Feeding upon

Speed of Maze Learning

C. J. Warden and E. L. Hamilton 446

3. The Effect of Delayed Incentive on the Hunger Drive
in the White Rat (Experiment 2 : The Learning Method)
E. L. Hamilton 455

4. The Relative Value of Reward and Punishment in the
Formation of a Visual Discrimination Habit in the
White Rat C. J. Warden and Mercy Aylesworth 488

Index 501

PART I

THE COLUMBIA OBSTRUCTION METHOD

Part I

THE COLUMBIA OBSTRUCTION METHOD
C. J. Warden

I. Previous Methods of Studying Animal Motivation

The measurement and analysis of the motivation factor in infra-
human behavior has been more or less neglected in the modern
experimental movement in comparative psychology. The reason for
this can hardly be due to lack of interest in the dynamic aspect of
behavior, since theoretical discussions on this topic have not been
wanting. Doubtless one important cause of this neglect has arisen
from the obvious difficulties of applying experimental methods in
this field. It is simple enough to get rough measures of learning
performance under various conditions which may be compared and
may thus lead to an understanding of the learning process. But
precisely how can one measure the strength of tendencies in the
organism aroused by hunger, thirst, sex deprivation, etc., and thus
come to know these primary determiners of animal activity as they
operate together in the daily life of various types of organism?
The problem is made difficult by reason of the fact that one is here
dealing directly with fluctuating physiological conditions that are
difficult to control even roughly. Then too, those internal states
which underlie various types of motivation carry with them the
likelihood of a highly excitable organism, which predisposes to
emotional disturbances during the testing process. When these and
other factors (Part I, 1, 2) are taken into account, it is small won-
der that experimentalists, interested in the quantitative analysis of
infra-human behavior, should choose to work in the field of learning
rather than in that of motivation.

Nevertheless several attempts have been made to investigate some
of the problems of animal motivation in the laboratory, although
usually these have been more or less preliminary and sporadic

3

4

COLUMBIA OBSTRUCTION METHOD

ventures. A brief account of the methods employed in this work
will be given in the present section. This survey will be followed by
a more detailed account of a new and thoroughly standardized
method — the Columbia Obstruction Method — devised by T. N.
Jenkins and C. J. Warden of this laboratory, in 1924, and, as later
modified in certain indicated details, employed throughout the com-
prehensive project reported in this volume.

The Revolving Wheel Method seems to have been among the first
used in the study of animal activity. As early as 1898, Stewart (15)
applied the method to small mammals. The apparatus consists
essentially of a revolving wheel attached to the living cage, after
the manner of the old-fashioned squirrel cage, to be set in rotation
by the animal inside whenever it is predisposed to be active. A
measure of the state of activity of the animals at different times and
under various conditions, can be secured by means of a kymograph
record, a revolution counter, or some other such mechanical device.
Slonaker (12) from about 1907 and onward, and Richter (10) and
his pupils from about 1922 and onward have made no little contri-
bution to our knowledge of motivation in small mammals by their
extensive use of this method. Many problems intimately associated
with the more analytical aspects of motivation can be studied by
this method, although as we have before noted (Part I, 1) the
indices obtained are of general activity which can seldom be related
to specific incentives or incentive-seeking tendencies. Thus noc-
turnal-diurnal and other rhythms of activity can be measured, and
the influence of gland extirpation, injections of various secretion
extracts and of drugs, etc., can be determined by comparison of
the indices secured with those of the normal animal. While the
method does seem to give a fairly trustworthy index of the general
tendency of the animal to be active or inactive under various condi-
tions, it does not appear to lend itself readily to the isolation of
many of the more important animal drives. Unfortunately for our
present interest, the activity is always the same — running the
wheel — regardless of the internal state induced in the organism ;
there is nothing characteristic in the behavior from one condition
to another to give us the cue as to what may be the objective of the

COLUMBIA OBSTRUCTION METHOD 5

activity in the aroused organism. In the study of specific animal
drives, it is desirable to set the conditions so that the amount
of activity is registered, but also in such wise that this activity is
exhibited as seeking behavior directed toward the specific type of
incentive that the experimentally induced physiological state may
call for. Certainly, the knowledge that a hungry rat is excited by
food in the offing and is bending its energies to secure the food, to
the exclusion of other types of activity, presents a more adequate
biological and psychological picture of its behavior than the knowl-
edge that, when equally hungry the same rat will turn a wheel
round and round so many times per hour. The fact that the former
statement indicates the relation of the behavior to a specific group
of stimuli (food) is a vital point in the understanding of the nature
of motivation in the situation.

The same general criticisms have been urged (Part I, 1) against
the Activity Cage Method which was devised by Szymanski (17)
and applied to a large number of animal types, and later used on
small mammals by Richter and his pupils in America, This method
has so far yielded much valuable information regarding some of
the more fundamental activity rhythms in these organisms, and
also concerning the manner in which these rhythms come to be
overlaid by habit patterns. References to studies of activity of this
type — "spontaneous" activity — are given in the bibliography,
and in connection with a recent article by Richter (11). In this
later article Richter has, indeed, applied the term "internal drives"
to the behavior revealed by this method as well as the method dis-
cussed in the preceding paragraph, but it would seem wise to
restrict the use of the term "drives" to activities depending upon
the arousal of internal physiological states which can be definitely
related to specific incentives, such as food, sex objects, etc., in the
manner of approaching or avoiding responses.

The Choice Method of determining the relative potency of differ-
ent incentive-drive situations may be considered an extension of the
Preference Method, as widely used by Lubbock (5), Graber (2) and
others in the study of sensory problems, during the latter half of
the nineteenth century, to the field of motivation. The animal is
stimulated by a pair of incentive objects (food-female, for exam-

6

COLUMBIA OBSTRUCTION METHOD

pie) and, with equally easy access to both, supposedly goes to the
one having the highest incentive value at the moment. It is difficult
to say who first applied this type of test, but certainly it has been
used extensively by various workers on lower forms in approaching
the problem of the relative potency of tropisms, and of instincts
under different conditions. The literature of the last several decades
contains so many studies of this sort that it is hardly necessary to
cite specific investigations. Mast has perhaps used the method more
widely than any other. More recently (1924) Moss (8) compared
the food-sex drives in the white rat by this method, using a simple
technique, which was greatly improved by Tsai (18) the following
year. Jenkins (Part I, 1, last 4 pages) developed a standardized
apparatus for testing animals by this method in 1925, but we have
made no use of this device in our research work on motivation at
Columbia, because we have felt that the method, even as improved
by Jenkins, is of slight value in this field. In describing the latter
apparatus, the writer (4) (Part I, 1, last 4 pages) pointed out the
more obvious and fundamental defects of any form of Choice
Method, as a means of analyzing dynamic behavior. It is quite
impossible to isolate a single drive, since the method itself requires
that two physiological states, corresponding to the two incentives
used (food-sex, for example) be simultaneously aroused in the
organism. But the two states induced are likely to be interdepen-
dent and the corresponding drives, unless they are genuinely antag-
onistic, thus influence one another in various manners and degrees
instead of operating as distinct internal tendencies. Furthermore,
it is impossible to say in our present state of knowledge precisely
how much starvation should be balanced against a given amount of
sex deprivation in such a case. The Method of Choice appears to be
ill suited to quantitative, analytical work on motivation.

The Learning Method (or Training Method) of studying moti-
vation has been employed with some success within a fairly limited
field. This method is an adaptation of the usual learning experi-
ment in that different incentives are used in the setting up of the
same habit, and the various incentive-drive situations ranked in
terms of some common index of learning efficiency (speed, accuracy,
etc.). This method has been used in a score or more of studies

COLUMBIA OBSTRUCTION METHOD 7

beginning with the early work of Yerkes (1907) on punishment vs.
reward; examples of the application of the method to certain
special problems will be found in the last three sections of the
appendix of the present volume. While the method must always
be considered an important one in connection with the analysis of
the incentive-drive factor in the learning process, it is doubtful
whether it furnishes reliable indices of drives as such. For, it is
clearly possible that a given internal physiological state may arouse
an organism to activity without increasing thereby its efficiency in
the adjustment-fixation process involved in habit formation. In-
deed, too great internal arousal may lead to decreased learning
efficiency, as when an animal is rendered so excited by a prolonged
period of acute starvation as to be unable to run a maze efficiently.
Conditions of motivation influence profoundly the stimulability of
the animal and the entire repertory of movemnts, as well as the
process of fixating new movements. Any comprehensive account of
animal motivation must, of course, include the influence of incen-
tive-drive factors on learning, but the more fundamental problem
of the relation of incentives and drives to activity and movement
must be obtained by some method that tests rather than trains.
Aside from the narrow limitations of the learning method, it has
also the very grave defect of introducing into the motivation index
obtained, the various sources of error from chance factors that
decrease the reliability of all training scores. It is common knowl-
edge that learning scores in both the animal and human field
involve gross inaccuracies because of the inability to control the
experimental conditions adequately over the usual long training
period. Then the very nature of the typical learning experiment
in so far as it requires any considerable amount of repetition of the
act to be learned, tends to lower rather than raise the value of the
motivation factor in the process. All these criticisms suggest that
the general problem of motivation should first be approached by
means of some direct test in which the training factor is eliminated
as far as possible. The learning method, however, may very well
take care of the special problems as to the relation of motivation to
different types of habit formation.

Several attempts have been made to develop a test method, as dis-

8 COLUMBIA OBSTRUCTION METHOD

tinguished from a training method of measuring motivation. The
basic principle employed in all these attempts is the general one in-
volved in the usual learning experiment, which has come down to
us from Lubbock. As early as 1882, Lubbock (6) laid down the
principle that the best method of testing an animal as to its intelli-
gence, etc., is to " interpose some obstacle" between the animals to
be tested and some definite incentive object, thus requiring the
animal to surmount the obstacle in order to secure the incentive.
Nearly all our laboratory techniques rely upon the application of
this principle in some form to arouse the animal to undergo the test
or to work upon the problem. This direct stimulation by incentives of
various sorts is, indeed, the sine qua non of experimental work on
animals, whatever the main objective of the investigation may
be. That the principle can be utilized in a test technique — that is,
under conditions in which the training element is reduced to a mini-
mum — as well as in the more usual learning technique is evident
enough.

The use of this 1 1 obstacle ' ' principle in a motivation test is fairly
simple. By using a standard obstacle and varying the incentive
from time to time, or from group to group, motivation indices that
are comparable for various incentive-drive conditions may be se-
cured. This procedure may be clearly distinguished from the
training method for here the animal is required to learn nothing
whatsoever; it must merely react to the incentive in accordance
with the degree of arousal effected by its physiological condition
under the circumstances. It is unnecessary to discuss in this connec-
tion the various applications of this general methodology to the
problem at hand. As examples of its use on lower forms we may
cite the work of Szymanski (16) and of Turner (19) on the cock-
roach, and that of Wodsedalek (23) on the May fly nymph. In
1922, Morgan (7) reported some tests made on the white rat which
made use of this principle. He argued that 1 ' the amount of inhibi-
tion necessary to overcome any tendency may be used as a measure
of the strength of that tendency" and suggests the possibility of
thus measuring all tendencies including those classified as instincts.
In 1924, Moss (8) applied this principle to the white rat more
extensively than Morgan had done and stated his thesis in language

COLUMBIA OBSTRUCTION METHOD 9

similar to that of Morgan. Moss used as an inhibiting1, or ' ' opposing
stimulus," as he speaks of it (the " obstacle" of Lubbock) a pair
of electric grills. The more important defects in the apparatus and
technique of Moss are indicated in Part I, 1, and require no further
mention' here. In no case had investigators carried their work far
enough to develop anything that properly deserves to be called a
general method of measuring animal motivation. The fundamental
principle underlying the various techniques employed had been
known for half a century and occasionally had been applied to
motivation problems. The basic principle was the old one of Lub-
bock, and one that had been used by both biologists and comparative
psychologists in the study of motivation. The matter of developing
a standardized technique apparently received scant attention, even
in the more recent work of Morgan and Moss. Nevertheless, it would
seem to be desirable to refer to the technique of Moss as the Resist-
ance Method in order to avoid confusing his apparatus and tech-
nique with that developed by Jenkins and Warden and called by
them the Obstruction Method.

The term 1 ' obstruction method" was introduced by the writer in
1926 (4) to designate the general method of measuring animal
drives which had been developed during the preceding two years in
the Columbia laboratory. Several writers, evidently confused as
to the facts in the case (I, 9), have spoken of the " obstruction
method" of Moss, mentioning our own method as a modification of
his technique. As before stated, our method was developed with
direct reference to the "obstacle" principle of Lubbock, the term
"obstruction" being used instead of obstacle because the latter
seemed to us to have a much wider application than any single
laboratory method. From the facts already cited it would seem
that, so far as basic principle is concerned, neither ourselves, Moss,
Morgan, Szymanski or other workers who adopted, knowingly or
otherwise, the "obstacle" principle can lay any just claim to orig-
inality. It seems equally clear, on the other hand, that each of these
workers, ourselves included, should be credited with the particular
application of the principle which each has made in the way of
apparatus and procedure. And especially so, when, as in the case
of our own apparatus, the similarities with any other previously

10 COLUMBIA OBSTRUCTION METHOD

devised apparatus are of the most superficial sort. Naturally, we
made use of the work of previous investigators in this field, includ-
ing that of Moss, but only in a general way, and in the manner of
rejecting outright rather than modifying the relatively simple
apparatus and technique which others had used. It is hoped that
this explanation will prevent any further confusion of the Columbia
Obstruction Method with the Resistance Method of Moss, which
have nothing important in common except that both apply the
"obstacle" principle of Lubbock, as many earlier techniques have
also done.

II. The Columbia Obstruction Method of Measuring
Animal Drives

In 1924 the writer began the development of what was planned
to be a general method of measuring motivation possessing a wide
application in the animal field. It was the original intention to build
a test method around the "obstacle" principle of Lubbock rather
than to apply the principle as a learning technique as most previous
workers in the field of motivation had done. It was considered not
impossible that the experimental conditions could be so simplified
that not only the different incentive-drive indices obtained for the
same species, but also corresponding indices from species to species
would be directly comparable.

The "obstacle" principle naturally calls for a three compart-
ment test arrangement: an entrance or reaction compartment and
an incentive chamber separated by an appropriate obstacle. The
planning of this part of the apparatus thus presented no very great
difficulties, except in the matter of devising automatic controls for
the doors. Since we desired to carry out work on the white rat at
first, it was necessary to construct the different compartments of
this section of a suitable size for this animal ; also to be careful not
to introduce into it any device that could not be utilized in a cor-
responding testing section for animals of different species.

The matter of the general type of obstacle to be used was not so
easily disposed of. To begin with it was felt that a general method
of the sort here proposed should not be limited to a single type.
Szymanski, Morgan, Moss and several others had used some form

COLUMBIA OBSTRUCTION METHOD 11

of electric shocking device as the obstacle, with its obvious advan-
tages of ease of control in quantitative units of determinable stimu-
lation value. After trying out a number of other types of obstacle,
it was finally decided to use an electric grill, and to concentrate
our efforts toward the development of a more adequate shocking
system than had heretofore been used. However, it should be noted
that, as a general method, the Obstruction Method is not limited to
the shock type of obstacle, although it does seem that the electric
shock, when properly applied, offers the greatest possibilities in
the case of most types.

The problem remained as to how to make possible the selection of
a suitable, standard shock by empirical tests to be used in as wide
a range of conditions as possible. Any form of induction coil, such
for example as that used by Szymanski and Moss would not do here,
since in order to get the widest range of stimulation conditions from
which to select a standard, amperage and voltage should be capable
of independent change. Obviously the demand here was for a more
or less elaborate system whereby a series of currents of graduated
amounts could be had in any desired voltage.

Expert advice was needed at this point and, fortunately, T. N.
Jenkins offered to devise and construct the electrical system needed.
Too much credit can hardly be given to Dr. Jenkins for the months
of experimental effort expended in perfecting this electrical system
whereby a wide range of shocks of known stimulation value can be
supplied to the obstruction grill. He found it necessary to con-
struct resistances to be used, since ordinary commercial resistances
were inadequate to carry the range of high voltages we wished to
use at a constant value for long periods of time. The superiority
of the Obstruction Method over previous techniques depends in no
small degree upon the standardization of the "obstacle," or obstruc-
tion grill, by means of this system.

The apparatus as originally constructed by Jenkins and Warden
in the fall of 1924, was first used in the doctorate research of
Holden (Appendix 1), on the hunger drive in 1924-25. She em-
ployed three degrees of shock — low, medium and high — and from
her results it was seen that a current of low amperage and of
medium voltage would be best suited for use as a standard obstruc-

12 COLUMBIA OBSTRUCTION METHOD

tion. Her study suggested also several other changes in apparatus
and technique. The use of a tunnel over the grills to prevent the
animals from attempting to jump across without getting a shock
has proven to be more satisfactory than the arrangement of glass
partitions employed in the original apparatus. The incentive com-
partment was found to be somewhat too short to contain the incen-
tive object and allow room for the animal to react properly to it.
This fact led the writer to devise a small fourth compartment, in
direct line with the other three, in which the incentive object could
be placed. When animals (male, female, litter, etc.) were to be
used as incentive objects it would be necessary, of course, to close
this small incentive chamber by a door. This called for some form
of mechanical device to operate the door that should relieve the
experimenter of the task, and also eliminate a possible source of
error from human operation of the door. The automatic door
release, illustrated * in Figure 2, Part I, 1, was worked out by
Jenkins and Warden and constructed by Warner in connection
with his doctorate research in the spring of 1926. The grill used by
Holden was also unsatisfactory on account of the posibility of short-
ing in case of micturition ; this defect was remedied by a new type
which allowed drops of urine to fall through and thus prevented
shorting. The plan of this improved grill was suggested by Jenkins,
and the grill itself constructed by Warner in the spring of 1926.
Dr. Jenkins should also be credited with the device illustrated in
Fig. 1, Part IV, 2, added in 1927 to facilitate the control of the
door to the entrance compartment. The illumination hood, designed
to eliminate any possible visual cue from the experimenter without
the use of a curtain, was contributed by Warner in the spring of
1926. No further modifications were made in the apparatus, which,
as thus standardized, was used throughout the project on animal
drives reported in this volume. The first study of this series was
that of Warner on sex behavior which was begun in the summer of
1926.

Several modifications were made in the technique originally
planned for the method in connection with the work of Holden.
Among the more important may be mentioned (1) the procedure
whereby the preliminary and "shock" crossings precede immedi-

COLUMBIA OBSTRUCTION METHOD 13

ately the test period proper ; (2) the lengthening of the test period
from 10 to 20 minutes, and (3) the slow and careful manner of
transferring the animal from the incentive compartment, after
crossing, back to the entrance compartment, so as to prevent the
possibility, of emotional disturbances. The details of these changes
can be had by comparing the technique used by Holden (Appendix
I) with the new technique standardized for the project (Part 1, 1, 2 ;
Part IV, 1). Although the writer must be held responsible for the
methodology finally adopted for use in connection with the obstruc-
tion apparatus, Holden, and later Warner deserve no little credit
for carrying out extensive preliminary tests in the laboratory
covering possible lines of procedure and these results were drawn
upon in determining specific points of methodology. Both worked
untiringly, especially in the matter of testing out group after
group of animals, in the effort to determine the most suitable shock
to be used in the project. Nissen also deserves special mention, not
only for making the actual tests reported in Part I, 2, but also for
his valuable suggestions in the planning of the analysis of the
method there reported.

In view of the fact that the next two sections deal with the method
in considerable detail, it seems unnecessary to discuss the matter
further here. A complete description of the apparatus and general
procedure will be found in Part I, 1, and additional points on
procedure in Part II, 1. The function of the obstruction and of the
incentive as related to this method are indicated in the experimental
findings of Part I, 2, together with the effect of re-testing on indi-
vidual and group scores. In both of these papers attention is called
to the fundamental advantages of this method over other methods
that had been previously used in the study of animal motivation.

In certain parts of the experimental work it was necessary to
make use of special techniques lying outside the field of psychology
proper. Each experimenter was required to master under competent
workers in the appropriate field such special techniques as should
be necessary to carry out his part of the project, such as the making
of vaginal smears, both quick and permanent, and the technique of
administering injections and of doing the necessary surgical work.
Special credit has been given in connection with the individual

14 COLUMBIA OBSTRUCTION METHOD

papers to the several workers who took special interest in our pro-
ject and gave advice and counsel on points lying outside our field.

III. A Note on Terminology

Unfortunately the recent attempts of comparative psychology to
adjust itself to a natural science, or biological viewpoint has not as
yet progressed to the point where a consistent objective terminology
has come into general use. No field is more beset by difficulties in
this regard than that of animal motivation, in which the phenomena
have been described for ages past in anthropomorphic and teleolog-
ical phraseology. Because of its historical antecedents, even the
term motivation itself appears to most of us to involve a distinctly
subjective connotation. Still, the present writer knows of no term
less harmful in this respect which might be used instead.

As experimentalists, we are not especially interested in theoretical
controversies on questions of terminology. Nevertheless, it is of
some importance that we describe our findings and discuss the
meaning of our results in language that will prove to be as accurate
and as understandable as the present circumstances permit. In the
absence of a generally accepted terminology in our field, we have
found it necessary to select such terms as seemed best suited to our
purpose, and have attempted to use these consistently throughout
the various reports included in this volume. In order to avoid mis-
understanding it seems desirable at this point to define as clearly
as possible several of the more important terms employed.

By an incentim is meant an external object such as food, water,
animal of the opposite sex, etc., that is capable of operating as a
': stimulus and arousing some fairly definite internal physiological
state. While recognizing that any external object, in order to be
classed as a stimulus at all, must be capable of arousing the organ-
ism in some manner and to some extent, we have purposely limited
the term incentive in this report to those objects or classes of objects
that satisfy, from the observer's point of view some fairly definite
drive or seeking tendency.

By a drive we mean an aroused reaction tendency which is char-
acterized primarily by the fact that the activity of the organism is
directed toward or away from some specific incentive, such as food,

COLUMBIA OBSTRUCTION METHOD

15

water, animal of the opposite sex, etc. If the appropriate incentive
object is not present, what appears to be seeking behavior to an
outside observer is exhibited. While recognizing that all behavior
represents in some manner and to some degree the dynamics of the
organism, or the drive of the organism, still there appears to be
some justification for limiting the term drive to reaction tendencies
that are of sufficient biological importance to represent seeking
behavior directed toward some fairly definite class of incentive
objects, such as food, water, animal of the opposite sex, etc.

The term drive does not refer to the physiological state or system
aroused either by an incentive or by deprivation of some sort, but
to the behavior tendency resulting from the internal arousal. For
example, hunger drive refers to the tendency of the animal to
approach or to search for food and not to the physiological changes,
such as stomach contractions, etc., that may be taking place at the
time. In most cases we have not attempted to secure an independent
index of the physiological state underlying the drive, so as to cor-
relate various aspects of physiological change with corresponding
changes in the reaction tendency, although something of the sort
has been done in certain of the studies on the sex drive. Attention
is called to this fact in order to emphasize the necessity of avoiding
a commonly made confusion between the reaction tendency as a
behavior phenomenon and the underlying physiological state.

By the term incentive-drive situation, and similar usage, we have
meant to stress the fact that the internal and external factors may,
and usually do operate together in behavior as induced in the usual
motivation experiment. The methodology of the Obstruction Method
actually requires that the incentive be either present or close at
hand. In a sense, then, all our drive indices are in reality incentive-
drive indices ; but this must always be true if we define a drive as a
reaction tendency directed toward an incentive, since this implies
that the latter shall be present in some sense. The word drive has
been used in most cases because the hyphenated term incentive-drive
is more or less awkward; the latter could be substituted for the
former throughout, however, without changing the essential mean-
ing of the text.

Finally, the term motivation has been employed in the experi-

16 COLUMBIA OBSTRUCTION METHOD

mental sections to cover both the incentive and the drive aspects of
the behavior investigated. In the discussion of The Columbia
Obstruction Method (Warden) the meaning of the term has been
extended to include studies of general or "spontaneous" activity
by various methods as well as certain types of experimentation on
"tropisms" and "instincts," which seem to deal primarily with the
dynamics of behavior.

1. STANDARD APPARATUS FOR THE STUDY OF
ANIMAL MOTIVATION 1

T. N. Jenkins, L. H. Warner, and C. J. Warden

In the first section of the present report a description is given
of an apparatus for the measurement of drives by the Method of
Obstruction. The apparatus has been developed during the past
two years in the Columbia laboratory in connection with a research
project on motivation in the white rat. It can be readily adapted,
however, to most common mammalian forms, and thus offers pos-
sibilities in comparative studies within this field. Great pains have
been taken to standardize the essential features by the use of auto-
matic devices when possible and by exactitude of detail in construc-
tion. The control procedure as developed will also be indicated.

The second section is a description of an improved apparatus
of the choice method type, following the general pattern used by
Tsai (18). The improvement consists, in the main, in the intro-
duction of automatic electrical devices permitting a more adequate
control of the animal and of the incentive stimuli. Standard prob-
lems and procedure are suggested.

I. The Obstruction Method Apparatus

As previously stated, this method involves the placing of an
obstruction of some sort between the organism and the incentive
stimulus. In order to obtain the latter and satisfy the dominant
demand of the moment, the organism is required to surmount the
obstruction. In the present apparatus the obstruction employed is
an electric grill supplied with a constant current of quantitatively
determined attributes. Various other types of obstruction might be
used but the electric stimulation has the advantage of being more
convenient to control and administer than most others would be.

1 Beprinted with modifications from The J ournal of Comparative Psychol-
ogy, 1926, 6: 361-82.

17

18 COLUMBIA OBSTRUCTION METHOD

For the sake of clarity in description the obstruction apparatus
will be divided into two parts and each treated in order. First, the
box for the control of the animal and the administering of the
shock; and second, the device for the control of the electrical cur-
rent supplied to the box in the obstruction section, permitting
definite systematic variation of the same. This description of the
apparatus will be followed by a discussion of methods of controlling
the intra -organic state of the organism for this type of investiga-
tion, kinds of data that may be obtained by this general method,
and standard problems to which the method offers an experimental
approach.

1. The animal control box

The ground plan of this part of the apparatus is shown in
figure 1. The box consists essentially of three compartments as
follows : entrance compartment A, obstruction compartment, B, and
incentive compartment, C-D, divided into two sections for reasons
indicated later. The body of the box is built of whitewood and is
finished both without and within in dull black. It rests upon a
table 30 inches in height which also serves to support the elaborate
electrical equipment necessary to furnish the shock to the grill in
compartment B.

The entrance compartment A is 10 inches square and 10 inches
deep, inside measurements. It is entirely separated from compart-
ment B by a wooden partition except for a doorway 4 inches square

A

c

D

Fig. 1. Diagram of Floor Plan of the Obstruction Box

A, entrance compartment; B, obstruction compartment; C, D, divided in-
centive compartment; E, release plate; ch, manually operated door of entrance
compartment; d2, automatic door (operated by release plate) between two divi-
sions of incentive compartment.

COLUMBIA OBSTRUCTION METHOD

19

leading into the tunnel that crosses the latter compartment. This
doorway is fitted with a door of thin sheet steel, dx, which slides in
felt-lined grooves, and is therefore practically noiseless. It is oper-
ated by hand by means of a cord which passes through a bent glass
tube (instead of the usual pulley device) to a point outside the box,
so that its operation cannot be observed by the animal in the com-
partment. In closing, the door sinks into a narrow felt-lined groove
in the floor. The purpose of the door is obviously to restrain the
animal in the entrance compartment until such time as the experi-
menter may wish to test its reaction to the obstruction-incentive
situation beyond. The compartment is covered by a panel of heavy
plate glass, but may be opened by sliding the panel, which rests in
a felt-lined groove, back over compartment B.

The dimensions of compartment B are exactly the same as
those above enumerated for compartment A, except for the fact
that the effective part of the compartment is restricted to a tunnel
4 inches wide and 4 inches high connecting the entrance and incen-
tive compartments. The entire floor of compartment B consists of
an electric grill set in the wooden base of the box so that it comes
up even with the floor of the two adjacent compartments. The
conducting elements in the grill are number 18 copper wires wound
upon a slab of bakelite -J inch thick. The turns are i inch apart,
and are wound in such a way that the turns from the two terminals
of the transformer alternate. A rat in crossing the grill will close
the circuit, therefore, by stepping upon any two consecutive turns.
This practically insures stimulation from the instant the animal
steps upon the grill. Moss used, as a shocking device, two plates
placed some distance apart along a passageway, simultaneous con-
tact with both plates being necessary to produce the shock. The
grill above described offers a much more dependable and uniform
source of stimulation.

The presentation of the electrical stimulation to the animal offers
certain difficulties. After trying out several other schemes the
tunnel device was finally selected as being the most satisfactory.
At first it was thought desirable to place the grill in the center of
the box without any further separation into definite compart-
ments. This would allow the animal to make contact at any point

20 COLUMBIA OBSTRUCTION METHOD

along the edge of the grill, and to have free passage from A to C
across any portion of it. This arrangement was found to be unsatis-
factory. Many of the animals would attempt, and sometimes with
success, to jump from A to C, a distance of 10 inches, without touch-
ing the grill. To prevent this two pieces of plate glass, extending 6
inches from either wall at a point 3 inches from the edges of the
grill, were so placed that the animal must pass through a winding
pathway approximately 4 inches wide in crossing the grill. This
arrangement eliminated the jumping behavior quite successfully
and also aided in marking off the obstruction part of the box from
the entrance part.

However, a further difficulty arose which led us to separate the
box into 3 definite compartments, as above described, by partitions,
and to install the tunnel device. It became clear from observing the
behavior of the animals, especially when a high degree of shock was
used, that they did not discriminate readily between the entrance
and obstruction portions of the box when not so divided. Usually
they would rush pell mell across the grill to the incentive portion of
the box the instant they were placed at A, and after rushing to C
paid no heed whatsoever to the incentive stimulus (food), but
attempted to escape by jumping up against the glass panel covering
the box or, perhaps crouched sullenly in a corner. Rats that had
been starved for long periods of time (one to three days) responded
in this manner. The animals seemed to be dominated by fear of the
combined A-B situation rather than by hunger. They were appar-
ently avoiding the entrance and obstruction sections rather than
seeking the incentive section. This escape response would be reen-
acted promptly, over and over again, as soon as the animal was
returned to the entrance end of the box, in preparation for a suc-
ceeding test.

When A and B were separated into compartments by partitions,
and the animal restrained in compartment A by the door, dlt for a
brief interval of time between successive tests, no such behavior
occurred. This suggests that the animal probably did not discrim-
inate, in the former arrangement, between the entrance and obstruc-
tion portions of the box. A-B as a whole rather than just B was
avoided, since the two portions were not sufficiently isolated to be

COLUMBIA OBSTRUCTION METHOD 21

readily discriminated as separate sources of stimuli which might
serve as the basis of a differential response. The copper wires on
the B portion of the floor constituted the only important difference
between that and the A portion, and the line of junction between
the two was not very marked since the grill had been made to fit
very neatly into the floor of the box. A second possible explanation
for this escape behavior can be made in terms of a conditioning
process of the substitute stimulus type as indicated in the following
diagram :

S (shock at B) B (escape into C)

S (sight of B while '
still in A)

This explanation assumes that the A portion of the box had been
actually discriminated from the B portion at first, but later came
to serve essentially as a single situation, by virtue of the substitu-
tion indicated above. At any rate the separating of the box into
definite compartments, as in the present apparatus, together with
the introduction of a brief interval of time between successive tests
remedied the difficulty.

The tunnel across the grill is 4 inches wide, 4 inches high and 10
inches long. The two sides are of wood, and the top of plate glass
to permit the same degree of illumination in this as in the other two
compartments of the box. The use of the tunnel has the effect of
making the response to the obstruction more definite, since there is
only this single point of approach and contact. Accidental contacts
with the grill are now much less likely to occur than under the old
arrangement. Furthermore, the movements of different animals in
passing over the grill are more uniform — they must either walk or
run. This makes possible relative constancy in the amount of elec-
trical stimulation involved in the act of crossing, this latter being
the chief unit of measurement on the behavior side.

The incentive section includes a double compartment, C-D,
as indicated in the diagram. The division into two parts would
not be necessary with such incentives as different kinds of food, or
other non-living stimuli. However, when the incentive stimulus is
a living animal (such as a female in testing the male sex drive)

22

COLUMBIA OBSTRUCTION METHOD

it is usually necessary to confine the stimulus animal in order to
prevent it from crossing over to the test animal. The method of
restraining the stimulus animal should be as natural as possible so
as not to arouse in it behavior likely to reduce its efficacy as a
normal positive stimulus. The restraint should cease at the moment
the test animal has completed the act of crossing the grill, and the
release of the stimulus animal should be automatic, if possible.

Moss restrained the female rat in the incentive chamber by a
simple door that must be raised by the experimenter, presumably as
the animal was crossing the grill. Such a method involves the error
of inexact timing — of opening the door either too soon or too late.
Furthermore, the moving door might be a factor of central impor-
tance in the stimulation situation during the test. A door of
ordinary opaque material would decrease the visual, and probably

Fig. 2. Diagram of Side Elevation of the Obstruction Box

A, entrance compartment; B, obstruction compartment; C, D, divided in-
centive compartment; E, release plate; d1} manually operated door of entrance
compartment; automatic door (operated by release plate) between two divi-
sions of incentive compartment; F, a hinge and G, an adjustable coil spring,
both supporting the release plate ; H, a rod rigidly secured to the release plate ;
I and J, silver contacts; K, a rod pivoted at L, the adjustment of which deter-
mines the level of the release plate when the latter is not depressed ; M, a block
carrying the bolt adjusting the depth of depression of the release plate; N, elec-
tro-magnet supporting free end of the pendulum, P, prior to its release; 0, in
circuit with the magnet and acting as resistance and as a signal; Q, adjustable
pendulum weight; B, cord conveying movement of pendulum to the door, d~.

COLUMBIA OBSTRUCTION METHOD

23

the olfactory stimulation in connection with the animal in the
incentive chamber. The movement of the door would likely serve
also as a distraction to the test animal.

We have endeavored to avoid all these difficulties by the use of
a separate restraining chamber, D, the entrance to which is guarded
by the translucent door, d2 (Figs. 1 and 2) which is opened auto-
matically by the test animal in the act of crossing the grill. The
door is 4 inches square, constructed of celluloid i inch in thickness
and is therefore semi-transparent. Three small windows, each 1 inch
square, cut in the upper half of the door, permit the free passage of
olfactory stimuli, whether of food or of sex object, from the
restraint chamber.

The mechanism of the automatic door release is illustrated in
figure 2. An opening 4 inches wide and 6 inches long is cut in the
floor of compartment C directly in front of the exit of the tunnel.
In this opening the release plate, E, is set flush with the floor.
The plate is of wood and painted a dull black so as to be distin-
guished with difficulty from the adjacent floor. It is supported
by a hinge, F, and the adjustable coil spring, G. Fastened rigidly
to the lower surface of the release plate is the L-shaped rod, H,
at the tip of which is the silver contact point, /. The tension of the
spring, G, is just sufficient to keep this point in contact with the
small silver plate, J. This latter is mounted on the rod, K, which is
adjustable in the vertical plane, being pivoted at the wooden block,
L, The V-shaped wooden block, M, spans the rod, H, and in the
base of this block is a rubber pad which can be raised or lowered by
an adjustable nut. The rod, H, rests on this pad when the release
plate is fully depressed.

In circuit with the points I and J, is an electro-magnet, N, of
1200 ohms resistance, and a 10-watt electric lamp, 0, both of which
are mounted on the underside of a 15-inch shelf fastened on the
outside of the box. Suspended from cone bearings at the other end
of the shelf is a pendulum, P, made of J inch square brass tubing
which carries an adjustable weight consisting of a brass collar held
in place by a set screw. A small piece of iron is soldered to the
pendulum at such a point that when the latter is raised the iron
will make contact with the pole of the magnet, N, and be held in

24 COLUMBIA OBSTRUCTION METHOD

contact if the current is passing through the magnet. A piece of
felt fastened over the pole of the magnet serves to deaden the click
produced by this contact. By means of a cord and 3 small pulleys,
the falling pendulum lifts the door, d2. Any tendency to rebound is
entirely overcome by the friction within the system.

The operation of this automatic door release is exceedingly simple
and practically noiseless. At the first touch of the release plate by
the animal in crossing the grill, the contact between I and J is
broken, to be re-made only after the animal has entirely cleared the
plate. The maximum depth of depression of the release plate is
controlled by the device at M and need not be more than 2 or 3 mm.,
the rod, K, being always adjusted so that the plate is flush with the
floor of the compartment. The tension of the spring, 6r, can be easily
and quickly adjusted for animals of any weight. We have found
that a tension of 40 grams is sufficiently delicate when 80-gram rats
are being tested.

Even when the white rat dashes through the tunnel at high
speed, the action of the automatic door release is so prompt that
the animal never comes into collision with the rising door. Indeed
the door does not seem to be noticed by the animal at all. The
stimulus animal is visible in the restraint chamber through the win-
dows of the celluloid door even when the latter is closed. The
restraint chamber is 4 by 4 by 6 inches in size and is covered with
plate glass. It can be removed by the simple turning of a thumb-
screw for purposes of cleansing or deodorizing. All incentive
stimuli, whether living animals or food, etc., are placed in the
restraint chamber, so that the automatic door device is utilized in
testing the strength of any drive.

The obstruction apparatus is placed for use in a dark room in
which the only source of light is an illumination hood. This latter
is made of wood and is 36 inches long and 4 inches wide and deep.
It is heavily coated with flat white inside so as to make a good
reflecting surface, and finished on the outer surface in dull black.
The hood extends the full length of the control box at a point
approximately 12 inches above the box in the central position. It is
illuminated by 8 lights of 3 CP., set at equal intervals (4 inches)
along the top of the hood within. The light is diffused through a

COLUMBIA OBSTRUCTION METHOD 25

sheet of translucent paper lying above the "Celotex" screen that
covers the lower open side of the hood facing the box. A piece of
ground glass of the proper grade would likely do as well. The hood
device gives a uniform illumination of the interior of all compart-
ments of the box, the outside environment being, at the same time,
relatively unilluminated. This arrangement permits the experimen-
ter to observe the behavior of the animal freely during the test
period, while remaining practically invisible to the animal.

2. The mechanism for the control of the electrical stimulation

The obstruction used in the present apparatus is an electric grill
forming the floor of a tunnel 4 inches wide and 10 inches long,
through which the animal must pass in order to obtain the incentive
stimulus. The scientific value of the apparatus depends very largely
upon the exactitude of control exercised over the current supplied
to the grill in compartment B. It is necessary, in the first place,
that the significant attributes of the physical stimulus be specified
in quantitative terms, if duplication by other investigators is to be
made possible. Among these are the form, intensity and frequency
characteristic of the stimulus. The dependence of stimulation value
of an electrical stimulus upon the rate at which the energy is
expended is well known. Waller (20) in 1899 showed that stimula-
tion value is dependent not only upon both voltage and amount of
current, but that there is an optimal ratio of energy used to the
rate at which it is used. In other words, there is an optimal minimal
stimulus or "optimum gradient of mobilizing energy producing
greatest excitation by least energy. ' ' Moss seems to have made the
mistake of supposing that the voltage value is a sufficient specifica-
tion of an electrical stimulus, for he fails to state the current values
used in his work. In our own apparatus we have taken great care to
avoid any possible error in providing a constant physical stimulus
value in the grill.

A schematic diagram of the electrical system, arranged to provide
for two control boxes to be placed side by side, is shown in figure 3.
The system would not be complex except for the fact that it was
considered desirable to provide for a series of 8 degrees of electrical
stimulation, either of which could be switched into the grill by

26 COLUMBIA OBSTRUCTION METHOD

simply turning a button, as will be explained later. If only a single
degree of stimulation were desired the apparatus would be very
simple in construction.

The voltage used is generated in the secondary of a General
Electric potential transformer shown at T, and voltage changes are
secured by varying the input into the primary by means of the
potentiometer, P. The electrical pressure can be measured at any
time by means of a high tension voltmeter, V, by closing a snap
switch, K.

Current regulation is secured by means of variable resistance
units, each controlled by multipoint switches, C, and C2. Each unit
contains eight elements (a, b, c, etc.), each having a resistance of
1,250,000 ohms. A common resistance, r, of 100,000 ohms, is in

Fig. 3. Diagram of Mechanism for Production of Electrical Stimuli

G, A.C. generator; P, potentiometer for controlling the voltage impressed
upon the primary of the potential transformer at T; K^, switch used for taking
voltage readings off voltmeter, V; A, microammeter for measuring stimulus
current (the current through either grill, gx or g2, is measured by throwing the
knife switch K2, to either pole 1 or pole 2) ; C\ and C2, multipoint switches for
throwing different amounts of resistance (a, b, c, d, e, f, g, h) into the stimulus
circuit ; Sx and S2, switches for closing the stimulus circuit through the grills.

COLUMBIA OBSTRUCTION METHOD

27

circuit with both resistance units, so that when all the elements (a,
&, c, etc.) are cut out there will remain approximately 100,000 ohms
of resistance in the circuits. The multipoint switches enable the
operator to vary the resistance cumulatively in equal steps from
100,000 to 10,000,000 ohms. The elements are liquid resistance in
glass tubes 2 feet long. Some of the better types of standard radio
resistance were first tried, but these were found to be too unstable
to be of service for the present purpose.

The current is measured by a micro-ammeter located at A. The
method of wiring is such that by means of the double-throw knife
switch, K, the current in either grill (g1 or g2) can be measured by
one fixed measuring instrument, thus insuring the same amount to
both control boxes, when they are used at the same time. The
measuring instrument used adds only about 30 ohms of resistance
to the circuit, and this amount is so small in comparison to the total
resistance of the circuit that it may be altogether disregarded.

In using the apparatus, the voltage is first determined by closing
the switch, K, and any necessary adjustment to bring the voltage
to the point desired is then made at P. The current values for each
step can be measured after first short circuiting the grills. The
current is thrown into either grill by closing the switches S1 or S2.

This arrangement enables us to specify the form, intensity, and
frequency characteristics of the physical stimulus in quantitative
terms. The writers mention form here because commercial alternat-
ing current is approximately sinusoidal. Stimulation value ought
to be practically constant, therefore, for a given voltage since
amperage is kept practically constant throughout.

The problem arises in this connection regarding the factor of
individual difference in the resistance of the skin of different ani-
mals, and of the same animal from moment to moment. In order to
be certain that we are obtaining an adequate measure of motivation
in terms of the obstruction overcome by the animal, not only must
the electrical stimulus in B be kept constant, but the effect of the
electrical stimulation on the animal must be controlled as definitely
as possible.

Moss attempted to insure a constant, uniform stimulation by
filling compartment A with water, or partly filling it. This would

28 COLUMBIA OBSTRUCTION METHOD

make certain a more ready contact with the grill, but does not
adequately control the factor of individual difference in skin
resistance. Furthermore, the presence of water in the entrance
compartment complicates the drive factor under investigation,
especially since there was no water in the incentive compartment.
The water may have aroused an avoiding reaction which operated
along with the hunger drive. Moreover, the effect of the water may
have been different for different animals, and for variations of the
hunger drive in the same animal.

We have attempted to control the factor of difference in skin
resistance by the use of high terminal voltages and by introducing
large resistances into the circuit to reduce the current values to
the extent required to give the appropriate stimulation value in the
grill. The logic of this procedure is fairly obvious. If we let r
be the variable skin resistance of the rat, and B the remaining resis-
tance in the circuit, we have the equation :

E

I =

B + r

where I is the current and E the impressed voltage. If B is very
large in comparison to r, it follows that a relatively large change in
r will produce relatively small changes in I.

Let us suppose, for example, that we are operating the system
at a pressure of 500 volts through a resistance of 4,000,000 ohms,
and that the skin resistance of a sample of rat varies from 100 to
5000 ohms. Then the current flowing for a skin resistance of 100
ohms would be 124.99687 microamperes, and that for 5000 ohms
would be 124.84394 microamperes. The maximal change in cur-
rent value is only 0.122 per cent. In our work last year we used
1200 volts. In such a case the differential in skin resistance becomes
a negligible quantity. It is not necessary to resort to the use of
water to insure uniform contact.

Moss used a current of 28 volts or less in his work, and in such
case the difference in skin resistance would likely be an important
factor in determining the results. Let us consider, for example,
the values for a current of 28 volts, with a secondary coil resistance
of 100 ohms, within the limits stated above. In this case a skin

COLUMBIA OBSTRUCTION METHOD 29

resistance of 100 ohms gives a current of 0.14 amperes, and a skin
resistance of 5000 ohms gives a current of 0.00549 amperes. This
represents a change in current value of 96.07 per cent. If the
E.M.F. is 20 instead of 28 volts the corresponding value will be
99.6 per cent. This argues very clearly for the use of high voltage
in this type of work where a uniform stimulation value is one of
the primary requisites.

The quantity of energy that strikes the excitable tissue each time
is determined by the voltage, amperage, and frequency of the alter-
nating current. As noted above, Waller has shown that stimulation
value is dependent upon both the amount of energy that strikes
the excitable tissue, and the rate of impact. Knowledge of the
voltage, amperage, and frequency furnishes us a means of determin-
ing both the amount of energy and the rate with which it acts. The
direction of flow of the current changes at every cycle, so that the
excitable tissue is struck once at every cycle. Most probably
the effect of the initial blow differs somewhat from that of succeed-
ing blows, inasmuch as the latter are conditioned to some extent by
the refractory period of the nerve.

The procedure to be followed in using the Obstruction apparatus
is exceedingly simple. The electrical supply to the grill is first set
at the point desired by an adjustment of the potentiometer. The
incentive is placed in chamber D and all doors and panels of the
control box closed. The test animal is then placed in compartment
A, and after a brief interval allowed to the animal for adjustment
to the situation, dx is raised and the stopwatch started. When the
animal crosses the grill he automatically releases d2, and thus
obtains access to the incentive stimulus. The animal should be con-
ditioned to the box without the electric stimulation on each of
several days preceding the test, the incentive stimulus being present
in this preliminary work.

Several units of behavior may be secured in using the apparatus.
The number of times contact with the grill is made, without actu-
ally crossing within a test period of specified length, and the num-
ber of successive crosses within such a stated interval are among
the most important data obtainable. Jumping and similar activity
apparently connected with the escape motive can also be quanti-

30

COLUMBIA OBSTRUCTION METHOD

tatively determined by the experimenter. It will be seen, further-
more, that by keeping the stimulation value of B constant, the
general problem of individual differences in a given drive may be
attacked, and different drive situations may also be directly com-
pared. This is possible because of the fact that the motivation of
the animal is in all cases measured in terms of its response to an
identical obstruction stimulus.

Perhaps a word should be said concerning the control of the
intra-organic conditions that represent the physiological basis of
animal drives. The use of standardized environmental conditions,
and especially that of nutrition, offers the only possible method of
insuring a normal animal in any experimental work, and the chief
requirement in studies of motivation is simply for a more careful
and rigorous control of these objective conditions. The present
apparatus offers no special difficulties in this direction, and the
application of general physiological principles and results furnishes
the only sound basis of control of this complex factor.

Systematic variation of the physiological state underlying a given
drive cannot be had, of course, in any very ideal sense. The best
method of isolating a specific drive, in so far as this is possible, and
of obtaining varying degrees of it for measurement, is to deprive
the normal animal of the appropriate incentive stimulus for vary-
ing periods of time, keeping the general physiological state of the
organism as constant as possible. For example, an objective index
of the hunger drive can be had by varying the interval of food
deprivation. The same method will probably apply in studying the
male sex drive. The female sex drive, on the other hand, must be
investigated with reference to the oestrus cycle, and the determina-
tion of the latter requires the use of appropriate histological tech-
nique. In all cases objective criteria should be employed in deter-
mining the physiological status and of obtaining varying degrees of
the same.

II. The Choice Method Apparatus

The ground plan of the apparatus designed for the study of
motivation by the choice method is shown in figure 4. It consists
essentially of an entrance section, B, and two incentive chambers,

COLUMBIA OBSTRUCTION METHOD 31

L and B, together with the necessary connecting pathways. The
control box measures 19^ inches in length by 10 inches in both
width and depth (inside measurements). The body is built of wood
and finished in dull black, and the partitions are of the same con-
struction, except as otherwise stated.

The entrance compartment, B, is 3J inches wide and extends
forward 7 inches to the door, Db. The latter is of thin sheet metal
sliding in felt grooves and operated manually. It serves to restrain
the animal in the entrance compartment until the release to the
incentive compartments is desired. The pathways (Al, and Ar)
leading to the incentive chambers are 3 inches wide. The incentive
compartments are indentical in size (5 inches wide by 7 inches
long) and in construction. The walls surrounding each compart-
ment are made of wood except for a 3-inch strip of wire netting
(i inch mesh) which forms the lower portion of the wall all around.
The purpose of the netting is to allow the test animal a more ade-
quate stimulation from the incentives in the two chambers. The
smooth wooden sides above tend to keep the incentive animal (when
such is used) from climbing out of easy range of the test animal in
the entrance compartment. The incentive chambers have been made

D* AR

L

B

R

?.

Fig. 4. Diagram of Floor Plan of the Choice Box

B, entrance compartment; L and R, incentive compartments; Pl and Pr, re-
lease plates for the pendulum, the operation of which opens the doors Dl and
Dr and closes Dx and D2; Ajj and A-r, runways from entrance compartment to
incentive compartments; Dr, door for retaining animal in compartment B.

32 COLUMBIA OBSTRUCTION METHOD

small for the purpose of bringing the incentive animal as near to
the test animal as possible and keeping the relation between the
two as constant as possible.

The incentive chambers are closed by the two doors Dl and Dr,
which serve to keep the incentive animal (when such is used)
within the chamber occupied. Both of these doors will be released
automatically by the animal in running across the appropriate
release plate (Pl or Pr) which operate noiselessly. At the points
where the release plate is situated, a wire netting similar to that
forming the lower portion of the side wall of the incentive cham-
ber, is made an integral part of the plate. This arrangement
prevents the incentive animal from operating the release plate by
crawling about on the netting.

The mechanism of the release plate operates as follows: When
the animal steps upon the plate, a circuit is broken through an
electromagnet which releases a pendulum. To this pendulum 4
sectors are attached in such wise that they can move through slots
in the floor of the box (Dl, Dr, Bx and D2). When the pendulum
drops, the attached sectors are carried along with it so that the door
(Dl or Dr) drops out of sight through the floor. The doors are
made of translucent mathematical celluloid to make them as incon-
spicuous to the approaching test animal as possible. As the doors
Dl and Dr drop out of sight, thus opening the incentive compart-
ments simultaneously, the doors D1 and D2 move into place and
prevent the animal from re-entering compartment B, or from run-
ning around to the opposite incentive chamber.

This automatic door release has the merit of being practically
noiseless, and we believe it to be superior to any of the usual sliding
door devices. An electric light placed in series with the magnet
goes out when the animal steps on the release plate. The release
plates rest firmly upon levers attached to a common beam which
operates in cone bearings mounted on the under side of the box.
This type of mounting is preferred because the bearings are far
enough removed to prevent particles from the box from falling into
them and hindering the free movement of the mechanism.

The control box is equipped with plate glass panels on top to
prevent the escape of the animal from the various compartments.

COLUMBIA OBSTRUCTION METHOD 33

It is used in the dark room and the lighting controlled by an
illumination hood similar to that employed in connection with the
Obstruction apparatus. The experimenter is kept outside the visual
range of the animal while the latter can be conveniently observed
by the experimenter during the test.

The technique of operating the choice method is simple. The
test animal is placed in B and, after a specified period of stimula-
tion, is released into the runways leading to the two incentive
chambers by opening the door Z>b. If the animal turns to the
right and stops on the plate Pr in approaching compartment R,
the door D1 leading to the opposite compartment is closed. At the
same time Dr opens permitting the animal to enter the incentive
chamber and secure the reward. If the animal turns to the left
when released from the entrance compartment, a similar mechanism
is set in operation.

The control of the intra-organic condition of the organism is
less satisfactory under this method than under the Obstruction
Method because of the fact that simultaneous stimulation by two
incentive stimuli always occurs. This means that systematic varia-
tion of both drive conditions must be made previous to the test in
which the two incentive stimuli are utilized. Both food and sex
starvation in measurable amounts must be brought about during
the period previous to the test, if food and sex object are to be used
as incentive stimuli. Naturally the control cannot be as satisfactory
when two sets of intra-organic conditions are varied instead of only
a single one. For this, and other reasons, the choice method seems
much less promising than the Obstruction Method, as a means of
investigating animal motivation.

2. AN EXPERIMENTAL ANALYSIS OF THE
OBSTRUCTION METHOD OF MEASURING
ANIMAL DRIVES 1

C. J. Warden and H. W. Nissen 1

The increasing use of more or less complicated apparatus in the
study of animal behavior brings with it the very serious problem
as to the extent to which the data obtained are a function of
the particular experimental conditions imposed rather than of the
variable selected for investigation by the method. Perhaps not
enough attention has been given to this obviously important source
of error in our laboratory studies. The attitude is all too common
of assuming that a given apparatus adequately serves the purpose
for which it has been designed if it seems to do so in the opinion
of the experimenter, or other human censor, without recourse to
thoroughgoing experimental tests to determine its actual fitness
upon the animal to be studied, or upon the specific function to be
measured.

The general purpose of the present study was to determine how
adequately the Obstruction Method, as finally standardized for our
project, gives a valid measure of drive-incentive behavior in the
white rat. The hunger drive in the male was chosen for study
largely as a matter of convenience. One is able to avoid altogether
thereby the complicating factor of activity rhythms associated with
the oestrus cycle in the female. That the hunger drive is thoroughly
typical is clearly shown by results already published or awaiting
publication. The present study should be thought of as method-
ological and as applying to the measurement of drives in general
rather than to the hunger drive in particular.

The same general conditions as those obtaining in our regular
drive work were maintained as far as possible. Younger and less
expensive animals were used, however, and therefore the absolute

i Reprinted with modifications from The Journal of Comparative Psychology,

1928, 8: 325-42.

34

COLUMBIA OBSTRUCTION METHOD 35

indices secured cannot be directly compared. In the drive project
we have used throughout the experimental colony strain of the
Wistar Institute at 185 days of age, whereas the rats employed in
our present study were approximately 120 days of age and were
secured from a local dealer. Our regular time in testing drives is
from 8 p.m. to 3 a.m., the white rat being more active then, as a
rule, than in the early forenoon (6 a.m. to 9 a.m.) when our present
tests were made. McCollum's standard diet, to which was added
bi-weekly rations of greens was used throughout except that when
our procedure required a long series of starvation periods alternat-
ing with short feeding periods this diet was supplemented by milk-
soaked bread. The animals were kept in the laboratory for at least
ten days before being tested. The regular diet (McCollum's) was
always used as the incentive. Each group was starved for forty-
eight hours immediately preceding the test in cages similar to the
regular living cage with an ample supply of water. Unless specifi-
ally stated otherwise the standard shock as used in our regular
drive work was employed. This shock is produced by an alternating
current of 60 cycles, with a terminal pressure of 475 volts, an
external resistance of 10,100,000 ohms in the circuit and a current
of 0.047 milliamperes. The shock produced is considerably lower
than the lowest degree used by Holden and was selected after much
painstaking effort to find the shock most suitable to be employed as
a constant throughout the drive project. A higher degree of shock
(Shock II) was made use of in a few instances in this investigation
(475 volts, 8,750,000 ohms, 0.054 milliampere) .

We have attempted to isolate by the usual experimental method-
ology the two most important features of the apparatus in order
to determine the influence of each in the measurement of drive
behavior : the obstruction, or shock and the incentive arrangement.
The method must be judged in the last analysis by the adequacy
with which these parts of the apparatus serve the functions for
which they were designed, when used in the proper manner. A
third problem attacked is concerned with the effect of practice
when groups are re-tested at regular intervals, and the validity of
the method in the study of individual differences in drive. These
topics will be discussed in the order mentioned.

36 COLUMBIA OBSTRUCTION METHOD

The Obstruction (Shock)

The value of the obstruction placed between the test animal and
the incentive depends upon its effectiveness in bringing out a rea-
sonably wide range of differential behavior. It represents a con-
stant against which the individual, various drives, and different
conditions of the same drive are placed for direct comparison. Our
reasons for selecting a high voltage electric shocking device in pre-
ference to other possible types of obstruction have been fully set
forth in a former paper (4). The standard shock was determined
upon after a large amount of preliminary testing of animals of the
age and weight that we had planned to use in the project. While
our testing out of many different degrees of shock was fairly system-
atic, we have omitted the detailed data from this report. The
important fact is that the standard shock finally selected has proven
highly satisfactory in the measurement of all drives so far studied.
In connection with the method as a whole, and for the mature white
rat, it fulfils exceedingly well the fundamental requirement of a
test ; it brings out an adequate range of individual and group dif-
ferences in the functions which it supposedly measures, and yields
valid indices.

TABLE 1

Showing the effect of the standard shock, with and without incentive

GROUP

NUM-
BER OF
ANI-
MALS

OBSTRUCTION
CONDITION

APPROACHES

CONTACTS

CROSSINGS

INCENTIVE
CONDITION

Aver-
age

Range

Aver-
age

Range

Aver-
age

Range

A

8

No shock

No Food

13.0

1-26

3.2

0-5

3.4

0- 5

B

7

No shock

Food

13.1

10-16

1.1

0-4

24.0

12-38

C

8

Standard shock

Food

22.9

15-30

3.6

1-7

11.4

0-22

Only one animal in each of groups A and C did not cross at all during the
test period.

Our primary problem, in the present instance, was to determine
the extent to which the standard shock as administered operates as
a deterrent to the normal tendency of the animal to approach the
incentive when placed in the entrance compartment. Accordingly
two groups were tested under standard conditions except that in
the case of one group the shock was omitted after the preliminary
conditioning. The data are given in table 1, in which is also
included that of a group which was tested with both shock and
incentive omitted.

COLUMBIA OBSTRUCTION METHOD

37

All of our studies indicate that the number of actual crossings
furnish the best index of the drive function. It will be noted that
when neither shock nor incentive is included in the set-up the
tendency to cross over and explore compartment C is very weak.
When an incentive (food) is present an average of 24.0 crossings
occur when no shock is involved. Only 11.4 crossings occurred in
a comparable group with the standard shock in operation, a reduc-
tion of somewhat more than one half. The P.E. of the difference is
4.299 and the obtained difference (12.6) is 2.93 times the P.E. of
the difference, so that the difference is quite reliable although the
small size of the groups should not be overlooked. These results
illustrate the normal operation of the obstruction ; the shock inhibits
without stopping altogether the tendency of normally motivated
animals to approach the incentive. It is doubtful whether the
standard shock affords much real "punishment" to the animal in
crossing since the stimulation value is quite low, being about the
same as we have used in visual discrimination work with mature
white rats in the Yerkes-Watson apparatus.

Instead of using a shock of constant value as we have consistently
done in our drive studies it would be possible with the apparatus
as constructed to begin with a low degree of shock and increase the
same either during a given test, or at successive tests, to the point
where the animal would refuse to cross. By means of a multipoint
switch (4) it is possible to vary the resistance in the circuit cumu-
latively in eight equal steps each representing 1,250,000 ohms. Two
of the higher steps had been used by Holden and were found to be
too severe at 1200 volts. However, it appeared to be worth trying
to begin with the first step (standard shock at 475 volts) and
increase by one step at each successive test. This method of using
the apparatus proved to be wholly unsatisfactory. A group of four
animals showed an average of 15.2 crossings on the initial test when
the standard shock was employed, but not one of them crossed when
re-tested with the switch set at shock II. The plan had to be
abandoned, especially since it was found that the animals still
refused to cross even after eight or more successive tests at this
setting were made. They seemed to be very strongly conditioned
against the grid. Of course a finer gradation of increasing steps

38 COLUMBIA OBSTRUCTION METHOD

could be arranged but this method of testing does not seem to offer
much. Even if a high degree of shock did not inhibit the activity
of the animal altogether, as seems to be quite generally true, it
would be likely to disturb the drive in some way so that the results
would tell us little concerning the normal drive. The use of a mild
shock kept constant throughout the test period and consequently
throughout the study of a given drive, or series of drives is much
to be preferred, especially since it has been found to be highly
satisfactory in actual application. In brief, we feel the data here
presented justifies our use of a constant versus an increasing shock
as an obstruction in the apparatus, and indicates that our standard
shock, while not necessarily the best possible one, does fulfil in a
highly satisfactory manner the purpose for which it was designed
in connection with the method. No attempt has been made as yet to
work out a series of standard shocks for the white rat at various age
or weight levels, although this could be done readily enough and
would add greatly to the general usefulness of the obstruction
method. Our project as planned is limited to the measurement of
drives in the mature animal and the age-weight factor has been
kept constant within very narrow limits.

The Incentive (Food, Water, Sex-Object, Etc.)

In an apparatus designed to test the drive-incentive aspect of
behavior the incentives should be so placed with respect to the
animal under test as to be highly stimulating if not actually
dominating in the test situation. This idea was effectually carried
out in developing the obstruction method. The smaller compart-
ment D which contains the incentive until the test animal crosses
over the grid and automatically opens the door, d2, is situated
directly in line with the tunnel. Visual stimulation by the incentive
is further aided by the fact that the door itself is made of trans-
lucent mathematical celluloid, and the whole apparatus is flooded
with diffused light during the test by the illumination hood, con-
trasting sharply with dark room conditions outside. The door is
also amply perforated to permit the free passage of odor from the
incentive which naturally passes through the tunnel into the
entrance compartment since all compartments are closed, or nearly

COLUMBIA OBSTRUCTION METHOD 39

so during the test. Every opportunity is thus afforded for normal
stimulation of the distance receptors directly.

The nature of the incentive is further emphasized by the plan
of having 4 preliminary crossings without shock and 1 with shock

- TABLE 2

Showing the effect of introducing incentive on group previously tested twice)

without incentive

GROUP

NUM-
BER OF
ANI-
MALS

OBSTRUCTION
CONDITION

INCENTIVE
CONDITION

APPROACHES

CONTACTS

CROSSINGS

Aver-
age

Range

Aver-
age

Range

Aver-
age

Range

D—

-Initial test

8

No shock

No food

13.0

1-26

3.2

0-5

3.4

0-5

D-

-Ee-test 1-

8

No shock

No food

7.5

1-18

0.6

0-1

3.0

0-13

Dl-

-Ke-test 2

4

No shock

No food

7.7

2-12

0.7

0-2

1.2

0-2

D2-

-Ee-test 2

4

No shock

Food

13.7

8-21

1.0

0-3

13.7

7-22

After the second test, group D was dvided into 2 sub-groups of 4 animals
each, one of which was tested with, the other without incentive. In 2 later
re-tests of D2, crossings rose to 21.2 and 26.2 respectively.

TABLE 3

Comparison of incentive and non-incentive group scores (Warnei sex drive)

APPROACHES

CONTACTS

CROSSINGS

o

■s

o

fl >»

"g «

"3 >»

GROUP

INCENTIVE

P

* H

CD

S.2

© 2i

©

CONDITION

SP

u

11

I'!

0

©

be

|l

II

a

©

1*

'-5

a

>

03 >

<U

8 'S3

!3

S

S

©
>

si >
*j ©

©

S3

©

03 >

3

<

o $

&

«

20 males

Female (cor-

1

1

1

nified)

1.5

0-10

2.3

2.35 j

102.2

Li

0-6

1.5

1.5 1

100.0

15.0

4-20

13.45

4.03

29.9

20 males

No incentive

1.0

0-3

1.2

1.03 1

85.8

l.D

0-4

1.05

1.181 112.3

3.0

0-7

2.95

1.8

61.0

21 females

1

1

1

(cornified)

Male

5.0

0-14

5.43

3.63 j

1

66.9

3.0

0-9

3.14

2.73 1

86.8

16.0

2-24

14.14

5.14

36.0

22 females

1

1

1

(cornified)

No incentive

3.0

0-10

3.23

2.7 |

83.6

2.0.

0-6]

2.27

1.91 j

84.1

5.0

1-9

5.05

2.55

50.5

precede immediately the test period proper in every case. Doubtless
the animal has been well conditioned to the specific incentive, as well
as to the electrified obstruction by this time. The more or less
frequent crossings during the test must result in a cumulative
enhancement of the specific incentive since contact with the incen-
tive stimulus is allowed after each crossing. It is difficult to see
how a better arrangement could be devised to bring into relief the
specific incentive factor since in the present case these successive
re-conditionings are aided by the direct stimulation of the appro-
priate distance receptors throughout the test.

That the incentive does operate as the dominant factor in causing
the animal to cross into compartment C, both when the grid is

40 COLUMBIA OBSTRUCTION METHOD

electrified and when it is not, can be demonstrated easily enough.

Certain data of table 1 show the value of the incentive, in this
case food, when no shock is administered during the test. The
normal exploratory activity of group A resulted in an average of

TABLE 4

Showing reliability of the difference between incentive and non-incentive scores
(average number crossings, Warner — sex drive)

GROUPS

DIFFER-
ENCE

STANDARD
DEVIATION
OF THE
DIFFER-
ENCE

DIFFERENCE
S. D. OF DIF-
FERENCE

CHANCES
IN 100 OF
A TRUE
DIFFER-
ENCE
GREATER
THAN 0

Ma les : incentive — non-incentive

Females: incentive — non -incentive

10.5

9.1

1.14
1.35

9.2

6.74

100.0
100.0

TABLE 5

Comparison of temporal distribution of crossings during test, percentile scores,
of incentive and non-incentive groups (Warner — sex drive)

GROUP

INCENTIVE
CONDITION

0-5 TH
MINUTE

6th-
10th

MINUTE

11TH-
15TH
MINUTE

16th-
20th

MINUTE

20 males

20 males _

21 females (cornined) _

22 females (cornined)

Female (cornined)
No incentive
Male

No incentive

28.2
47.6
19.2
43.2

23.4
28.9
14.7
32.4

22.3
15.3
26.4
15.3

26.0
8.5

39.7
9.9

only 3.4 crossings into compartment C whereas group B crossed
to the food incentive 24.0 times. Further evidence on this point is
furnished by the results presented in table 2. The number of cross-
ings steadily decreased at each re-test when food was not present in
the incentive compartment, whereas crossings jumped to 13.7 when
food was suddenly introduced at the second re-test of group D2,
and rose steadily to 21.2 and 26.2 at the third and fourth re-tests
of this group. This shows that the incentive compartment calls out
little exploratory activity in the absence of some sort of specific
incentive.

The dominating influence of the incentive even in the presence
of the standard shock has been clearly indicated by the work of
Warner (22) on the sex drive, and tables 3, 4, and 5 of our report
have been arranged from his data. We will discuss here only the
scores covering crossings inasmuch as such scores have been found
to be by far the most significant indices of drive obtained by the
obstruction method. The incentive factor represents the sole differ-

COLUMBIA OBSTRUCTION METHOD 41

ence in the general and test conditions of the groups compared, and
hence the differences in score may be taken as a measure of the
incentive factor in the standard test set-up. The facts in the case
are so clear cut that little discussion is necessary. Both male and
female animals were tested when in the physiological condition at
which the normal sex drive has been shown by Warner to be at its
maximum, and the incentive animals as used likewise represented
the best possible condition for giving the maximum sex stimulation.

Under these conditions, the males crossed on the average 13.45
times when the proper incentive was present in the compartment
and only 2.95 times when no incentive was present. The relative
variability of the former group is also very much lower. The
difference between incentive and non-incentive groups is almost as
great in terms of average number of crossings for the females,
although in this case the tendency to cross even when no incentive
is present is stronger than in the males. This latter fact, as
Warner has suggested, may be due to some special oestrous deter-
minant in the female which also shows itself in the higher absolute
score in crossings when the incentive is present. The relative
variability is also measurably lower in the case of the females for
the incentive group. The validity of the difference between incentive
and non-incentive scores for both male and female test groups is
extremely satisfactory as a glance at table 4 will show.

In table 5 the temporal distribution of crossings during the test
period, when the latter is divided into 4 equal divisions of five
minutes each, are given. It will be noted that the percentage of
crossings shows a marked tendency to decrease as the testing
proceeds when no incentive is present in the case of both male and
female groups. When an incentive is present, the percentage of
crossings steadily rises as the testing proceeds for the females and
remains practically constant for the males. This would seem to
indicate that the more or less frequent contact with the incentive
following crossings during the test period tends not only to main-
tain but also to increase the stimulation value of the specific incen-
tive being used at the time. Another possible interpretation of the
data of table 5 would be an increasing tendency to become nega-
tively adapted to the grid as the test proceeds. Probably both

42 COLUMBIA OBSTRUCTION METHOD

factors are at work here and are jointly accountable for the tem-
poral distribution of crossings and the opposite tendency shown
under incentive and non-incentive conditions. The behavior of the
test animal must result from some sort of balancing of the positive
influence of the incentive and the negative effect of the obstruction.
The two factors are so interpenetrated in their actual operation
that it is difficult if not impossible to evaluate them independ-
ently in a satisfactory manner. The important fact is that, as both
are used in our standard technique, valid indices of the drive-
incentive aspect of behavior can be secured as in no other apparatus
so far devised.

Effect of Re-Testing

It seemed desirable, in connection with this general study of the
obstruction method, to determine the effect of repeating the test on
the same group a number of times. Two questions arise at this
point. The first is that regarding the practice effect when a given
test period is considered a practice period and a number of these
are given in succession as in a learning experiment. We should
expect the animal to become adapted to the experimental condi-
tions, particularly the grid, each successive test or practice period.
But at any rate the speed at which this adaptation takes place
is a matter of some concern. The second question concerns the
validity of the apparatus as a measure of individual differences
in drive behavior. Do the individual indices obtained in an initial
test hold their relative positions in later re-tests under the same
general conditions ? If so, then this is so much evidence in favor of
the validity of the method as a means of securing dependable indi-
vidual as well as group scores of drive.

The re-test procedure was in general the same as that of the
initial test although certain important modifications were necessary.
The starvation period was kept constant for each re-test at forty-
eight hours by running these on every third day, food being sup-
plied on the day following each test period. The test animal was
placed in the starvation cage, however, for at least one hour
immediately after testing before being transferred to the regular
living cage, in order that no direct connection should be formed

COLUMBIA OBSTRUCTION METHOD

43

eSBiaAy

1-1 CM

r— (

3

rH OS

co y
o o

i-l d

88UB}J

lO CO

e3BJ9Ay

ft ^MMMNNMhhh

C» t— lr-(r-li—(i—(i— !>—(»— (r-Hi—l

jo uaqnin^

OOOOQOCOOOOOOOOOOOOO

e3aB}j

e28J9Ay

e3uB£[

©SBiOAy

eSuB^j

<M *o t-i ©
(N CO rJH

o ©

9

co jjij ^fj y—i co

I III I IT

O00^ 00 05OO(NOM

—< (N CO CO M< Tjl <m

^OOOOr-t^(M(Mos»ot>

i-H i— l CM CO CO ~ ~ TP

bp^OpCMCOCMrHrylrH Jt>. CO

i-iOOOOOOOOO(MC^

e3Bj3Ay

3q3t8M.
eStsjaAy

BIBOTIUB

jo aaquin^i

OiOOuOO^i-iCOr-iOfNt-

MHNodddddciwjN

OiOOOO^(Oh(OM <M CO

•P^S i T i .i .i j

"5iOCOOOOOOOOOO»>
(N"*XMCChOhOOiOh

g ^rt<COCO(M(N<M(N<N(N<N»-i

Ca ,_ <r— <r-lr— It— It— Ir-Hr-tr-Hr— It— <r-4

00 00 00 00 00 00 (30 00 00 00"^"*
rH t-1

::::::::: i< n

;r-((NCO"*lOONOOO)HH
. ] "* i ^ i a i a i ** i ** | ~» | j ) HH H™l

• aitiinQggisioBnuu

03IIIIIIIIIOO

44 COLUMBIA OBSTRUCTION METHOD

between the test and the 23-hour feeding period. The schedule in
re-testing was thus as follows: starved two days — initial test —
starved one hour, feeding allowed twenty -three hours; starved 2
days — second test — starved one hour — feeding allowed twenty-
three hours, etc. Inasmuch as the normal food was a powder (Mc-
Collum's diet) which does not lend itself to gorging, opportunity to
feed for a day (twenty-three hours) between successive starvation
periods seemed necessary. And even this precaution did not prove
sufficient to keep the animals up to their normal weight as will be
noted later.

A second difference between test and re-test involved the omission
of the regular preliminary crossings except for the fifth, or condi-
tioning crossing. It did not seem necessary to repeat the series of
4 crossings without shock after the initial test since the purpose of
these crossings is to establish a connection between the incentive
and the act of crossing and this connection has been well formed
by the end of the initial test. Furthermore the animals have been
negatively conditioned to the grid in the first test and will not
readily pass over it even when it is not electrified in later tests.
Each re-test proper began then, after a single re-conditioning cross-
ing with shock and lasted the usual twenty minutes. Shock II was
used in the tenth re-test, the standard shock being employed in the
first 9 re-tests. Food was always present in the incentive compart-
ment.

If the results of table 6 be taken at their face value then it is
quite clear that the animal becomes rapidly adapted to the grid so
that it no longer inhibits him from crossing. However, the results
given in the first section of the table are complicated by the fact
that, quite unintentionally on our part, the animals were being
seriously underfed. This would mean, of course, that the physiolog-
ical basis of the drive under investigation was being varied from
test to test and that the effects of chronic starvation were probably
influencing the results. The animals were losing weight on our
schedule which allowed one day of normal feeding between every
two-day starvation period (see table 6) whereas they should have
been growing and taking on weight. After the sixth test milk-
soaked whole wheat bread was added to the diet of the animals on

COLUMBIA OBSTRUCTION METHOD 45

the interpolated feeding day but even this did not stop altogether
the losing of weight, although a few remained practically at this
level thereafter.

The results in the first section of table 6 involve not merely the
factor of practice growing out of successive re-tests but also that
of quantitative stunting. In order to determine the influence of
the factors separately a control group was given the same feeding
schedule for twenty-seven days as the test group had had and then
tested on what corresponded to the eighth re-test period of this
latter group. The loss in weight was approximately the same in
both test and control groups. It is quite apparent that the effect
of the stunting was to increase the drive, since the control group
crossed 23.1 times instead of 11.4 times as the test groups show on
their first test. It appears then that the factor of stunting was in
no small degree responsible for the increase in number of crossings
at successive tests in the case of the test group. The effect of stunt-
ing and of practice, as involved in re-testing, is, obviously, to run
up the score.

When due allowance has been made for the influence of stunting
on the test group, the data of table 6 show clearly that adaptation
to the obstruction does take place rather rapidly under our condi-
tions of re-testing. The average number of crossings jumps from
11.4 on the initial test to 19.6 on the first re-test made only three
days later. Also the effect of practice is evident from the fact that
the crossings rise from 23.1 to 35.9 on a second test of the control
group. The most convincing evidence, however, is the fact that the
first test of the control group shows a score of only 23.1 crossings
as compared with 42.2 crossings for the corresponding period
(re-test 8) of the test group. The latter value represents both
practice effect and stunting, while the former is a measure of the
stunting factor alone.

It is interesting to note in passing that the average number of
crossings came finally to approach very closely the theoretical limit
for a twenty-minute test period in the case of the test group. That
is, under the influence of stunting and re-testing conditions the
animals crossed continuously as soon as replaced in compartment A.
The theoretical limit is 44 crossings in a twenty-minute test period

46 COLUMBIA OBSTRUCTION METHOD

and is based upon the speed of the animal in crossing and the time
required to shift the animal back to the entrance compartment
again. This maximum number of crossings was actually reached in
many individual cases and became the upper limit of the range in
re-tests, 6, 7, 8, and 9. The fact that the average number of crossings
was above 40 on re-tests 5 to 9 shows that the obstruction did not
inhibit the tendency to cross over at this stage of practice under
the conditions of stunting then obtaining. It does not mean that
complete adaptation to the obstruction would follow a half dozen
re-tests under normal feeding conditions. Adaptation to the grill
probably does take place rather rapidly in successive testings and
on this account we recommend that the data of the initial test only
to be taken as measures of drive. The learning factor entering into
re-test scores make them less and less valuable as exact measures of
drive if the method be thought of as a test rather than a learning
technique. In our opinion the learning method of measuring drives,
as used by Simmons, Szymanski and others gives a much less ade-
quate index of drive function than the obstruction method. The
directness and simplicity of the response in the latter reduces
immeasureably the large element of chance factors that is always
prominent in maze learning and similar learning situations. There
is every reason for using a test rather than a learning set-up in
quantitative studies of drive, although interesting corroborative
data may be secured by determining the relative value of various
incentives in habit formation.

On the ninth re-test the animals of the test group were crossing
the grid continuously and we decided to give them shock II on the
next re-test. The 8 animals were divided into 2 groups, A and B.
Shock II was given to group A at the beginning of re-test 10 and
to group B at the middle of the corresponding re-test period. In
both cases the higher shock practically inhibited all crossings, the

TABLE 7

Rank of individual animals on successive tests

ANIMAL

A

B

c

D

E

p

G

H

Initial test

2.5

2.5

4

5

6

7

8

Re-test 1 —

2.0

4

3

5

6

7

8

Ee-test 2—

3.5

3.5

2

6

5

7

8

Re-test 3__

4.5

4.5

2.5

6

4.5

7

8

Re-test 4 —

4

4

2

4

6

7

8

COLUMBIA OBSTRUCTION METHOD

47

value recorded for group B representing the number of crossings
that occurred before shock II was thrown in, i.e., during the first
half of the test period. Both groups were given a re-test, using
shock II, three days later but not one of the 8 animals crossed the
grid. Four of the 8 rats were held over for thirty days and given
another re-test and again not one' of them crossed into compartment
C. These results show that shock II is altogether too severe for
practical use in connection with the grid, or obstruction. It also
argues against the technique of increasing the shock in a graded
series of steps in testing the same animal and getting the strength
of drive in terms of the degree of shock that inhibits crossing alto-
gether, unless much finer gradations of shock were used than the
series afforded by our present apparatus. A shock strong enough
to condition the animal completely against the grid is also likely to
bring about a shift in physiological condition and in the drive
under investigation, and thus defeat the purpose of the test alto-
gether.

We come now to the question of the validity of the method as a
measure of the drive function in the individual. No extensive
investigation of this aspect of the method has been made but such
data as we are able to present are all favorable. In table 7 we give
the ranking of a group of 8 animals on the initial test and on four
successive re-tests. The rat which crossed the largest number of
times was assigned the rank of I, etc. It will be seen that animals
A, G, and H kept the same relative position throughout, while there
was comparatively little shifting around in rank among the other
animals of the group. So far as the evidence goes, the method seems
to have a relatively high self correlation, although such a conclu-
sion should finally rest upon a larger body of data.

Our general conclusion must be that the obstruction method as
standardized and used in connection with our project on animal
drives fulfills most adequately the purpose for which it was
designed. In its present form it can only be used in the measure-
ment of drive-incentive behavior in the white rat. However, we
plan to work out such modifications of the animal control section as
may be necessary in order to extend its use to a wide variety of
animal forms of different sizes and habits. Inasmuch as the essential

48 COLUMBIA OBSTRUCTION METHOD

response is that of normal locomotion, and the set-up relatively
simple so far as the animal is concerned, its use as a method in
comparative psychology would seem to be almost unlimited so far
as adaptation to divergent morphological types goes. We plan in
the near future to build and standardize the apparatus of the
proper size to be used on dogs, cats, raccoons and monkeys.

Bibliography

(1) Dashiell, J. F. 1928. Fundamentals of objective psychology. New
York, Houghton, Mifflin, pp. 257, 258.

(2) Graber, V. 1884. Grundlinien zur Erforschung des Helligkeits —
und Farbensinnes der Tiere. Prag, Tempsky, 322 p. (Excellent
summary of preference method work of Bert, Graber, Lubbock, etc.)

(3) Holden, Frances. 1926. A study of the effect of starvation upon
behavior by means of the obstruction method. Comp. Psychol. Mono-
graphs, 3, No. 17, 45 p.

(4) Jenkins, T. N., L. H. Warner, and C. J. Warden. 1926. Standard
apparatus for the study of animal motivation. J. Comp. Psychol.,
6 : 361-82.

(5) Lubbock, J. (Avebury). 1882. Ants, bees, and wasps. New York,
Appleton, 448 p.

(6) Lubbock, J. (Avebury). 1882. Ants, bees, and wasps. New York,
Appleton, p. 247.

(7) Morgan, J. J. B. 1922. The measurement of instincts. Proc. 31st.
Ann. Meeting of the Amer. Psychol. Assoc., Cambridge, Mass., p. 94.

(8) Moss, F. A. 1924. Study of animal drives. J. Exp. Psychol, 7 : 165-85.

(9) Murphy, G. 1929. An historical introduction to modern psychology.
London, Kegan Paul, p. 378.

(10) Richter, C. P. 1922. A behavioristic study of the activity of the
rat. Comp. Psychol. Monographs. 1, No. 2, 55 p.

(11) Richter, C. P. 1927. Animal behavior and internal drives. Quart.
Rev. Biol, 2:307-43.

(12) Slonaker, J. R. 1907. The normal activity of the white rat at dif-
ferent ages. J. Comp. Neurol, 17 : 342-59.

(13) Slonaker, J. R. 1908. Description of an apparatus for recording
the activity of small mammals. The Anatomical Record, 2 : 116-22.

(14) Slonaker, J. R. 1912. The normal activity of the albino rat from
birth to natural death, its rate of growth and the duration of life.
J. Anim. Behav., 2 : 20-42.

(15) Stewart, C. C. 1898. Variation in daily activity produced by
alcohol and by change in barometric pressure and diet with a descrip-
tion of recording methods. Amer. J. Physiol, 1 : 40-56.

COLUMBIA OBSTRUCTION METHOD 49

(16) Szymanski, J. S. 1912. Modifications of the innate behavior of
cockroaches. J. Anim. Behav., 2 : 81-90.

(17) Szymanski, J. S. 1920. Aktivitat und Ruhe bei Tieren und Men-
schen. Z. allg. Physiol., 18 : 105-61.

(18) Tsai, C. 1925. The relative strength of sex and hunger motives in
the albino rat. J. Comp. Psychol., 5 : 407-15.

(19) Turner, C. H. 1912. An experimental investigation of an apparent
reversal of the responses to light of the roach (Periplaneta orientalis
L.). Biol. Bull, 23:371-86.

(20) Waller, 0. D. 1899. The characteristic of nerve. Proc. Royal Soc,
65 : 207.

(21) Wang, G. H. 1923. The relation between "spontaneous" activity
and oestrous cycle in the white rat. Comp. Psychol. Monographs, 2,
No. 6, 27 p.

(22) Warner, L. H. 1927. A study of sex behavior in the white rat by
means of the obstruction method. Comp. Psychol. Monographs, 4,
No. 22, 67 p.

(23) Wodsedalek, J. E. 1912. Formation of associations in the may-fly
nymphs (Heptagenia interpunctata Say). J. Anim. Behav., 2:1-19.

PART II

THE HUNGER DRIVE

Pakt II

THE HUNGER DRIVE
C. J. Warden

The two papers included in Part II deal with very different
aspects of the hunger drive. The first covers a repetition of the
work of Holden, using the Obstruction Method in its final form in
the measurement of the normal hunger drive in the male and female
white rat. By normal hunger drive we mean the tendency to
approach and secure food after a given period of starvation when
the general physiological condition of the animal has been kept
undisturbed, except in the matter of starvation. By controlling
other possible drives in the manner indicated below, the hunger
drive was made dominant during the time of the test — though,
perhaps, it was not isolated from all other tendencies in the strict
sense.

The thirst drive was eliminated by the simple procedure of allow-
ing the animal to have access to its regular water supply up to the
moment when it was taken from the cage to be tested. The sex drive
offered more difficulties because the method of control for the
different sexes would necessarily be different. The sexual factor
in the males was eliminated as far as possible by using animals that
had had no opportunity to copulate for 35 days, at which time the
male, while normally active, shows little or no tendency to be stim-
ulated by the female in oestrum (Part IV, 1). For the females,
selection of only those in dioestrum for testing solved the problem
of control of the sex factor. The additional precaution was taken
to test animals of the same sex only on a given day, and to cleanse
the apparatus thoroughly after the series of tests for each day had
been given. The maternal drive was not involved, inasmuch as preg-
nant females and females with litters were not included. General
exploratory behavior could not be strictly controlled. It was, how-

53

54

THE HUNGER DRIVE

ever, kept as constant as possible by testing all animals at the same
period of the day, and at the time (between 9 P.M. and 4 A.M.)
when the white rat is known to be most active. The small size of
the incentive compartment should have the effect of making the in-
centive object stand out and thus would tend to discourage explo-
ration in the main. This general procedure of standardization of
both living and testing conditions was effective, we feel confident,
in isolating to a high degree the hunger drive — at least as far as
such is possible without introducing controls so artificial that they
themselves would be a disturbing factor important enough to in-
validate the indices of drive behavior obtained.

Part II, 2, opens up a large field of motivation as tested by this
method, i.e., the effect of immediacy of reward and strength of
drive. In all other work included in the main body of this report the
animal is given access to the incentive as soon as he passes over
the obstruction and enters the incentive compartment. In the present
section, however, the animal is held up after making the usual
attempt to secure the incentive (food) for various intervals of time.
The logic of such procedure might be either that the delay tends to
weaken the drive or that the delay tends to intensify the stimula-
tion value of the incentive situation in a cumulative manner. This
study attempts to determine the matter so far as short intervals of
delay are concerned. The thirst, sex, maternal, and exploratory
drives were controlled in the same manner as indicated above in
connection with the normal hunger drive. In selecting a certain
degree of hunger on which to study the effects of delayed incen-
tive, the 48-hour starvation period was decided upon because this
had been found (Part II, 1) to represent the stage at which the
hunger drive was at its maximum under normal conditions. The
experiment is thus limited to the study of the effect of delayed in-
centive when the hunger drive is at its maximum. Further work
covering the effect of delayed incentive on other stages of the hun-
ger drive, should be carried out before wide generalization is war-
ranted on this aspect of motivation.

The term ' ' hunger drive ' 1 as used in connection with this project
has reference to the tendency aroused in the animal by starvation to
approach food in the incentive compartment. In our work the

THE HUNGER DRIVE

55

strength of this tendency is indicated by the indices of behavior
obtained. Since no independent check was had of actual stomach
conditions, no implications are intended as to the presence or ab-
sence of "hunger contractions" during the state of arousal. The
term ' ' incentive ' ' refers to the food itself, while ' ' incentive-value ' '
or ' ' incentive-index ' ' denotes the stimulation effect of the incentive
in terms of the behavior score. None of the terms used carry any
implications whatsoever regarding the subjective state of the or-
ganism.

1. A STUDY OF HUNGER BEHAVIOR IN THE
WHITE RAT BY MEANS OF THE OBSTRUCTION

METHOD 1

L. H. Warner

In a recent monograph (13) were published the results of an ex-
perimental study of sex behavior in male and female white rats.
The method used in that study has been termed the obstruction
method (4). The most important data consist of the records of
the number of times in a test period of constant duration that an
animal will cross an electric grid to reach a sex object (an animal
of the opposite sex). The effect upon this tendency to approach a
sex object or, in brief, the effect upon the sex drive of variation in
physiological condition was studied. In the female it was found
that the sex drive was dominant during only a rather restricted
phase of the entire oestrous cycle, that during this period (the
cornified stage) the crossing was decidedly and reliably more fre-
quent than during the dioestrous stages. The average number of
crossings per unit time made by the animals tested when in the
cornified stage, at the peak of sex activity, can be considered an
index of the sex drive in the normal female.

In the case of the male the factor influencing most profoundly
the sex drive is the length of time which has elapsed since the last
previous mating period. Of the seven sex deprivation intervals
studied the twenty-four hour interval resulted in the maximum
drive. The results for this group may be considered an index of
the sex drive in the normal male.

A second fundamental type of behavior to be studied is that
directed toward food. Such behavior has previously been studied
by Moss (6) and by Holden (2) using the obstruction method. The
results of neither of these investigations can be compared with
those of the sex study mentioned above because of decided dissim-
ilarities in method. These will be enumerated in a later section.

i Reprinted from The Journal of Comparative Psychology, 1928, 8:273-99

56

THE HUNGER DRIVE

57

The present article is a report of the experimental study of food-
seeking behavior made under such conditions as to render the re-
sults directly comparable with those presented in the monograph
on sex behavior. It also includes a comparison of these results with
those on the sex drive.

The animals used were of the same stock and raised under the
same conditions in every particular as were those used in the study
of sex behavior. The apparatus and general procedure were the
same. There will be taken up here consideration of only such mat-
ters as refer especially to feeding conditions. For further details
see Warner (13).

The animals had from the time of weaning been supplied with a
diet which was adequate qualitatively and quantitatively. The de-
tails of the diet during their first hundred and fifty days during
which they were in the care of the Wistar Institute cannot be given,
but the long and successful experience of this organization in the
care and breeding of normal animals is ample guarantee of its ade-
quacy. During the thirty -five days each animal was in this labora-
tory prior to testing the following diet was supplied it continu-
ously :

This was mixed, dry, in an electric mixer for at least ten min-
utes. Greens were supplied in abundance once a week. There was
always an ample supply of water from inverted bottles which ren-
dered fouling impossible. More than a sufficient supply of food
was kept in the cages at all times (except, of course, during the
starvation periods mentioned below). These feeding conditions
differed in no way from those which held in the study of the sex
drive but are described here since they are closely related to the
hunger drive here under investigation.

Methods and Procedure

Per cent

Whole wheat flour. .
Whole milk powder.
Casein

70.0
.16.6
.11.0
. 1.4
. 1.0

Sodium chloride. _
Calcium carbonate

58

THE HUNGER DRIVE

Just as in the study of sex behavior the physiological condition
of the animal as related to food consumption was kept as constant
as possible (and in that condition accompanying a minimum of
food-seeking activity), so in this study we have determined and
kept constant the animal's condition as this related to sex behavior
and have used those conditions accompanying a minimum amount
of sex behavior. In no other investigation of either of these types
of behavior has such control been attempted or, perhaps, even the
need for it recognized. This criticism holds for even the more
recent work (Simmons (9), Moss (6), Tsai (11), Holden (2)).

The most dominant factors affecting sex activity in the female
are, apparently, (1) the present stimulus situation and (2) the
oestrous condition. To take up the first of these factors: there was
no element in the experimental situation which seemed at all re-
lated to, and thus likely to arouse sex behavior. Although it seems
well nigh impossible to control the olfactory situation when ani-
mals of the two sexes are used on the same apparatus we at least
spared no effort in attempting to do so. Animals of the opposite
sex were not used in the apparatus on the same night (all experi-
mentation took place between 9 P.M. and 4 A.M., as in the case of
the sex study).

Immediately at the conclusion of a period of experimentation
the apparatus was thoroughly washed out with soap solution and
allowed to air. At no time was an animal given the opportunity of
associating the apparatus with another animal. In other words,
but one animal at a time was in the apparatus, the object in the
incentive compartment being not another animal but food. The
second factor was controlled by determining the oestrous condition
of the animal at the time of its testing. This determination was
based upon the histological characteristics of the vaginal secretion.
Sampling of the secretion was never made until after the animal
had been tested ; thus there was eliminated the possibility that this
process might interfere with the normal behavior of the animal
during testing. Sex activity in the female is rather rigidly re-
stricted to a relatively brief portion of the oestrous cycle. During
the remainder of the cycle the animal rarely shows any interest in

THE HUNGER DRIVE

59

the male as a sex object and will, in fact, actively avoid mating in
most cases. By using females in the inactive period or dioestrum
the sex factor is reduced to the minimum. The data reported herein
concern only such females as were in the following stages as de-
fined in the study on sex behavior : Recuperative, Early inactive,
Inactive and Late inactive. Several females were in oestrum at the
time of testing (as determined afterward by smear examination)
in spite of the fact that the females were selected for testing on the
basis of behavior observation which seemed to indicate that they
were in dioestrum. Data on such animals were discarded. There
was not a sufficient number of them to make it worth while to con-
sider them as a separate group.

The dominant factors affecting sex behavior in the male are (1)
the present stimulus situation and (2) the period of time which
has elapsed since the last previous mating period. The first of these
was controlled as in the case of the female. The second was kept
constant in the case of all males used. The sex deprivation interval
immediately preceding testing was thirty-five days. This long
period was determined upon after careful consideration of the re-
sults of the study on the sex drive. When studying food-seeking
behavior it was not merely our purpose to keep conditions affect-
ing the sex drive constant but so to arrange them that the sex drive
should be at its minimum at the time of the testing. This cannot be
done as simply with the male as in the case of the female since the
male is not subject to a cycle involving a period of sex inactivity.
It might be suggested that we use males immediately after they
had been given a period of mating, after sex 1 'satiation" had re-
sulted. While it is true that the sex drive appears to be at its
minimum at this time our observations upon rats tested (in the
obstruction apparatus with a female in oestrum in the incentive
compartment) when in this condition indicate that their behavior
is so profoundly affected that a fair measure of the food-seeking
drive could not be obtained. Such males were found to be ex-
tremely inactive. The usual type of behavior displayed was sleep
(or what appeared to be such).

Again it might be suggested that we use males whose sex drive

60

THE HUNGER DRIVE

had been partially satiated. But casual laboratory observations
appear to indicate that the drive remains at a high point under this
condition although it is displayed rather more spasmodically.

Nor would the use of males after a short deprivation interval
prove satisfactory. As has been shown in the sex study, recovery
of the sex drive is very rapid. It is apparently well along the road
to recovery by the time (varying with different animals) when the
male becomes aroused from the lethargy following a prolonged
period of mating. It thus seems that if we should use a rather
short sex deprivation interval we would not be dealing with a
homogeneous group of animals but with a group containing both
lethargic and active animals.

A sex deprivation period of one day has been shown to result in
sex behavior at or near the high point. But a glance at figure 4 in
the sex monograph shows that beyond this point there is a gradual
drop in the drive. The longest sex deprivation interval used in
that study was one of twenty-eight days. A decided falling off in
the strength of the drive was found in this group of twenty males.
Just as significant for our present purpose is the consideration of
the behavior displayed by these twenty-eight day sex-deprived
animals when placed in a situation involving sex stimulation. Of
the twenty males so tested only two crossed to the female during
the first minute. This is the lowest number recorded for any group,
being even lower than that for the control group (no sex object or
other incentive). This indicates that these two crossings were very
probably determined by other stimulation than that related to sex.
During the first five minutes but sixteen crossings were recorded
for the twenty-eight day group as contrasted with seventy-six for
the one day group. After the first five minutes the activity became
progressively greater indicating what might be described as a
reawakening of interest in the sex object.

As has been pointed out in the sex study, the sex drive in the
male appears to be more determined by the external stimulation
and less by the internal stimulation than in the case of the female.
Thus sufficiently long deprivation periods in the male produce a
drop in the drive such as would be much less true of the female

THE HUNGER DRIVE

61

because of the rhythmic occurrence of an internal stimulus situation
in the latter. It is true that the male after a long period of sex
deprivation can recover in a few minutes its normal interest in a
sex object when it is placed in the presence of a female in oestrum.
This fact does not, however, prevent such males from being the
ideal animals for our present purpose since there is nothing in the
situation involved in the measurement of the hunger drive which
would directly, or by association arouse the sex drive.

Thus by using long time (35-day) sex deprived males in the
present study of the hunger drive we feel that we are keeping con-
stant and at a minimum the sex drive. These animals may in fact
be considered somewhat analogous to females in dioestrum.

The incentive stimulus in the present study was food, about
twenty grams of the mixture described above. This food was con-
tained in a black metal receptacle firmly fixed in the smaller section
of the incentive compartment, i.e., where the incentive or stimulus
animal was placed in the study of the sex drive. (See either of the
following references for cuts and description of the apparatus used :
(4) or (13)). The automatic door, which had for its purpose the
restraint of the incentive animal in the study of sex, operated in
just the way it did in that study although it was of course not
necessary to restrain the food. In brief, except for the fact that a
food object was substituted for a sex object the experimental con-
ditions were no different from those which held in the study of the
sex drive.

The behavior which determined when the animal should be
returned to the entrance compartment after crossing consisted in
the taking of a nibble of food. This criterion was adhered to with the
following exceptions. If during the preliminary training period
the animal did not take one nibble of food within one minute it was
returned anyway. This period was reduced to thirty seconds during
the test period. This exception had a particular bearing on the
groups of animals which were run immediately after being taken
from cages containing food, i.e., after no starvation period. These
animals, as would be expected, failed to respond to the food any
more than to other details of the incentive compartment. As a

62

THE HUNGER DRIVE

matter of fact, crossing was far less frequent in animals of this
group even though they were sometimes returned prior to sampling
the food.

It was originally planned that a control group corresponding
to that group showing the greatest hunger drive should be tested.
This group was to have differed from the test group only in that
there would no food or other incentive object in the incentive com-

TABLE 1

Showing grouping of animals and the number of each sex in group

LENGTH OF STARVATION
PERIOD

NUMBER OF ANIMALS

Male

Female

days

0

10

10

2

10

10

3

10

10

4

10

10

6

10

10

8

10

10

partment while the animals were tested. Our results showed, how-
ever, such reliable differences between the behavior of animals
recently fed and those starved for a period that it was felt that
additional indication that the differential behavior was a function
of the animal's condition as related to food was unnecessary. It
might be noted parenthetically that the control groups run in the
sex study cannot serve as controls for this investigation in spite
of the fact that these groups consisted of animals tested with no
object in the incentive compartment. This is true because the
female control consisted of females in oestrum (whereas the females
used in the present study were in dioestrum) and the male control
consisted of males taken immediately from cages supplied with food
(and would thus not be comparable with the four day starvation
group, which in the present study showed the hunger drive at its
height in the male).

The periods of starvation first studied in this investigation were
0, 2, 4, 6 and 8 days. The 3-day group was later added as a result
of data already obtained since these seemed to indicate the possibil-
ity that the peak of food-seeking activity for the two sexes com-
bined might lie between 2 and 4 days. Standard groups of twenty
animals each were used as was done in the study of the sex drive

THE HUNGER DRIVE

63

and as is being done in other studies on drive to be published from
this laboratory. The grouping is tabulated in table 1.

Results

Table 2 gives the distribution of the approaches, contacts and
crossings for the various groups. The medians, ranges, averages,
standard deviations, and coefficients of variability are found in
table 3. Table 4 gives the reliability of the differences between the
average crossings for various groups. The approaches, contacts
and crossings as distributed minute by minute throughout the
twenty minute test period are given in table 5. In table 6 these data
are combined in five-minute units and are also reduced to per-
centages. Figure 1 shows graphically the effect of variation in the
starvation period upon the average number of approaches, contacts
and crossings, the data for the two sexes being combined. Figure 2
is like figure 1 except that the two sexes are treated separately.
Figure 3 graphically demonstrates the effect of variation in the
starvation period upon the temporal distribution of crossing during
the test period. The writer is deeply indebted to Dr. Frances
Holden for the statistical treatment of the data and the construc-
tion of the tables and figures.

The results will be discussed in four main sections dealing with
(1) the effect of variation in the length of the starvation period
upon the hunger drive, (2) sex differences in the hunger drive,
(3) a comparison of the results here reported with previously
reported data on food-seeking behavior, and (4) a comparison of the
strength of the hunger drive as herein measured and the sex drive
as measured in the work reported previously by the writer (10).

1. The effect of variation in the length of the starvation period
upon the hunger drive

The groups will be considered in order from the shortest to the
longest starvation period.

The 0 group consists of animals taken immediately from the
living cage, where food was always available, to the obstruction
apparatus where they were tested. About ten minutes elapsed
between their removal from the cage and the commencement of the

64

THE HUNGER DRIVE

CO

3

9FW

CM

CO

l-H

i-H

— H

o~

CM

I— 1

i-H

CM

1—1,

1—1

CO
>*
<

■ m

9lBH

1— (

CO

CM

T-H

1— 1

l-H

1— 1

a

co

Ji

oreuiaj;

CM

l-H 1

i-H

i-H

l-H

l-H

1— 1

e

ftf

ft

9lBN

i—i

CI

r— 1

i— l

i— i

CM

CM

II

9IBW

05

1 CO

9lBJV

a

o

JI

9JBUI9J

proaches

9IBI/\l

i— i

CO

CM

i-H

l-H

i-H

CM

CI

CM

CO go

9IBW

CM

i-H

CM

o

i—i

1— 1

i— 1

i— l i— 1

05
r"
•<

Con-
j tacts

9IBW

i-H

CI

CI

O

M<

CO

i-H

s

• XI

ft y

CI

i-H

C)

ca

i-H

i-H

< o
c
ft

0[TJUJ0j

i— 1

C4

OJ

CM

l-H

ii

9lBIM

i-H

l-H

i-H

l-H

1— 1

0-

00

111

i— 1

CI

CM

<:

Q
co

c3l

CI

CI

»— 1

co
9
• -a
a&

91BTAT

CI

Cl

<S
a

91TJTJU9 J
L a

9TB TAT

r-H

CM

l-H

o-S

9JBCU9J

l-H

1— 1

co
>-

1 CO

9IBTAT

i— i

1— 1

i-H

i-H

CO

l-H

•<

Q
CM

Ji

gjBTjag^

i— i

CI

CM

co

9

9IBTAT

CM

co

i-H

<§
ft

OjtjUJ.9^

CO

1— I

CI

i-H

CM

& »

91BUT

iO

I— 1

1— (

i-H

^H

2c

9JBUJ9^

CO

co

l-H

i-H

9I«I\[

CM

CM

CO

CJ

O

9JBUI9J

CM

CM

I-H

Ap- I

proaches

9IBW

lO

CO

9I1JUI9J

CO

co

CM

THE HUNGER DRIVE

65

8 DAYS

Cross-
ings

Con-
tacts

•l«W

3

1 -C

a

•!•!!

6 DAYS

Cross-
ings

y—i »— t

1— 1 1—1

Con-
tacts

03

5

a

4 DAYS

Cross-
ings

i— < »H iH iH i— 1

I— 1 »— 1 1— 1 T— (

Con-
j tacts

Ap- 1
proachesj

•mi

3 DAYS

Cross-
ings

•Ibn| <N iH »h

Con-
tacts

are ma j

Ap-

proaches

•i*w

aremaj

2 DAYS

Cross-
ings

WW

iH T-4 fH

arcmaj

Con-
tacts

Ap-
proaches

•lBW

O

Cross-
ings

•I«JI

Con-
tacts

•mi

erBUiaj

8
a

arBUiaj

Ift CO

66

THE HUNGER DRIVE

TABLE 3

Showing results for each starvation period

APPROACHES

Female

Male

Total

Female

Male

Total

Female

Male

Total

Female

Male

Total

Female

Male

Total

Female

Male

Total

0-7
0-7
0-7

0-6
0-9

0- 9

1- 8
1-11

1- 11

2- 9
0-7

0- 9

1- 12

0- 9

1- 12

0-12
0-10
0-12

2.0
1.5
1.75

3.4
3.3
3.35

5.2
5.0
5.1

5.2
3.6
4.4

6.1
3.6
4.85

6.8
4.6
5.7

"2 >
a o

1.34
1.34
1.36

1.56
1.47
2.81

1.25
1.41
2.46

2.04
2.20
2.37

1.77
1.59
3.58

1.
1.79
3.77

I1

CONTACTS

0-6
0-6
0-6

0-15
0-12
0-15

6-14
0-10

0- 14

4-8

1- 5
1-8

0-13

0-9

0-13

0-16
0-16
0-16

3.5
2.2
2.85

6.9
4.9
5.9

8.0
5.7
6.85

5.7
2.8
4.75

5.3
3.0
4.15

7.3
5.5
6.4

c ©

ji"o

1.55
1.20
1.42

1.87
1.59
3.92

41

58
2.93

10
17
1.91

.80
.55
3.61

2.29
2.18
5.71

CROSSINGS

0-7
0-9
0-9

3-29
3-25
3-29

1910-29
18 4-29

4-27
3-25
3-27

0-16
0-19
0-19

2.1
2.
2.

2.07
.37
.78

73

42

19.
16.1
17

08

9-3618.47.63
52
58

4-2617.05.92
5.87
5

19.1
18.1

a

5.2
•o >
a <d
at)

CO

67

91
6.56
.78

07

27

14.07.18
14.25.98

14.1

6.61

6.04.69
9.85.33
7.95.07

11

125
116

47
41
44

41
42
42

35
31
33

51
42
47

78
54
64

TABLE 4

Showing reliability of the differences between the average crossings of each group

GROUPS

STANDARD
DEVIATION

OP THE
DIFFERENCE

DIFFERENCE
BETWEEN

THE
AVERAGES

DIFFERENCE
S. D. OF DIFFERENCE

CHANCES IN
100 OF A
TRUE
DIFFERENCE

2.43

15.1

6.21

100

2.36

.7

.3

62

2.17

.1

.05

52

2.00

4.0

2.0

98

1.86

6.2

3.33

100

2.40

15.8

6.58

100

2.22

4.1

1.85

97

THE HUNGER DRIVE

67

twenty-minute test period. This time was occupied by the pre-
liminary training in the apparatus. These animals crossed to the
incentive compartment almost as readily as did starved animals
during this preliminary training (during which the grid was not
electrified). But once in the incentive compartment they were
occupied mainly with exploring, not with eating. With the begin-
ning of the test period (during which the grid was electrified) their
behavior was greatly altered. They continued to cross occasionally,
but only after much hesitation. During the first five minutes of the
test period a total of only 31 crossings was recorded for the twenty
animals while 28 contacts were recorded for the same period show-
ing that the animals retreated from the grid after touching it
almost as often as they continued on across. The second five-minute
period shows only 9 crossings, the third, 5, and the fourth, 3. Thus
there is a continuous diminution in the tendency of the animals to
cross. The average number of crossings for this group was only 2.4,
approaches, 1.75 and contacts, 2.85. This group represents the low
point for all three forms of behavior. No animal in this group
crossed more often than nine times.

The two-day group presents a striking contrast to the preceding.
The average number of approaches increases to 3.35, the average
number of contacts increases to 5.9 while the average number of
crossings leaps from 2.4 to 17.6. The difference between the
average number of crossings for the 0 and the two-day group is
highly reliable being over six times the standard deviation of the
difference. Animals of this group were, obviously, decidedly
more active. There was still occasional hesitancy as is shown by
the approaches and contacts, but for the most part the animals
recrossed the obstruction thirty to sixty seconds after being replaced
in the entrance compartment. The crossing was more uniformly
distributed throughout the entire test period than in the case of
the 0 group although this activity was still rather more frequent
during the early part of the period.

The three-day group does not differ markedly from the two-day.
There is a slight further increase in all three forms of behavior
recorded but, especially in the case of crossings, the differences are
not reliable. The chief difference between this and the preceding

68

THE HUNGER DRIVE

group is to be found in the temporal distribution of crossings
during the test period. These are now almost uniformly distributed
throughout the period.

The four-day group shows a slightly lower average number of
approaches and contacts while the crossings remain practically the
same. The crossings are still rather uniformly distributed.

The average number of approaches and contacts for the six-day
group is almost the same as for the four-day group but there is a
definite drop in the number of crossings, the average being 18.1
for the four-day group and 14.1 for the six-day group. This differ-
ence being twice the standard deviation of the difference is not
entirely reliable, the chances of a true difference being 98 in 100.
When it is noted that the succeeding group shows a further drop
there seems reason to suppose that this difference is a real one.

The eight-day group shows a further and decided reduction in
the number of crossings. Animals of this group were very weak as
shown by their gait and general behavior. There is some selection
in the composition of this group since three females and two males
died in the course of filling out the group which as usual consisted

DHTML 2 0Af3 3d *y* 4<iAy* t*~f*

Fig. 1. Showing average number of crossings (solid line), contacts (dash
line) and approaches (dot-dash line) for each interval of starvation, the data
for the two sexes being combined. Each point represents the average of twenty
animals. ("Control" should read "O," i.e., no starvation period.)

THE HUNGER DRIVE 69

0O»ifj Z daft 9t/~f* * "04fs £ doys fdoyi

Fig. 2. Showing diagrammatieally the same data illustrated in Fig. 1, but
here the two sexes are treated separately. The abscissa represents length of
starvation period. Average number of crossings for the female (solid line), for
the male (dot-dash line) ; average number of contacts for the female (long
dash-short dash line), for the male (dotted line) ; average number of approach-
es for the female (short dash line), for the male (dash-two dot line).

to

Cor*nOL ~2doyS 3doys **0—fS 6a«/s fCyS

Fig. 3. Showing temporal distribution of crossings, data for the two sexes
being combined. The upper line represents the percentage of crossing which
occurred during the first five minutes of the test period. The middle line repre-
sents the percentage which occurred during the first ten minutes. The lower line
represents the percentage which occurred during the first fifteen minutes.
(" Control' ' should read "O," i.e., no starvation period.)

70

THE HUNGER DRIVE

of ten animals of each sex. In spite of their weakness these animals
showed a definite orientation toward the incentive compartment.
This is indicated by the fact that although they crossed less fre-
quently than did the animals of the six-day group they approached,
and made contact without crossing even more frequently. After
observing the behavior of these animals one is tempted to say that
the hunger drive is as strong as ever and that reduction in crossings
is due to decreased capacity to resist the shock. Starvation may
possibly effect not only the animals tendency to approach food but
its capacity to undergo electric stimulation.

2. Sex differences in the food-seeking drive

Although the behavior of the two sexes was sufficiently similar
to justify their combined treatment in the above section there are
certain sex differences worth noting. Figure 1 shows diagram-
matically the effect of variation in the starvation period upon the
approaches, contacts and crossings, the two sexes being considered
together. In terms of crossings, the most significant data, the peak
of activity extends from two to four days, there being but insignifi-
cant differences between the two, three and four-day groups, and
all three groups showing decidedly higher values than those pre-
ceding or succeeding them. When, however, the two sexes are treated
separately it is seen that each presents a rather clear cut peak, that
for the female being at two days and that for the male being at four
days. This is shown in figure 2. Thus the plateau seen in figure 1
is merely the result of combining the curves of the two sexes which
show peaks at different points. It may well be supposed that the
drop for the longer starvation periods is due to the lowering of
vitality through the lack of nourishment. If so, then it might be
said that this weakening influence effects the female "joth sooner
and more severely than it does the male. Starvation periods of
three and four days lead to reduction of crossing in the female
whereas in the male the crossings increase up to four days. This
does not explain why it is that a two-day starvation results in some-
what greater activity in the female than in the male.

A sex difference is also seen in the data concerning contacts and
approaches.

THE HUNGER DRIVE

71

Approaches. The 0, two and four-day groups show almost no sex
difference. From this point on, however, the female shows decidedly
more activity of this sort than does the male.

Contacts. The trend in the number of contacts with various
degrees of starvation for the male and the female runs almost
exactly parallel, with the female at all times showing considerable
more activity of this sort than the male.

In terms of approaches and contacts it can thus be seen that the
male at no time displays more activity than the female, and, in
fact, that the reverse is usually the case, especially in the groups of
longer starvation periods. It might then be said that the orienta-
tion of the female was just as surely directed toward the food as in
the case of the male; furthermore, that the more decided drop in
crossings in the case of the female might therefore be due not to less
internal stimulation but to a greater sensitivity to the shock, due to
a more rapid lowering of physical vigor.

Data on the temporal distribution of activity during the test
period show no significant sex differences and so are not considered
here.

3. A comparison of the results here reported with previously
reported data on food-seeking behavior

The present work cannot be compared in any great detail with
any of the work done previously since none have made use of
technique sufficiently like that used here. Holden's study is the
most nearly comparable.

Moss (6) used so few animals that his quantitative data are of
little significance. The procedure which he adopted is not well
suited to the problem of the rapidity of the increase in the hunger
drive and the period of starvation resulting in a maximum. His
longer starvation groups were composed merely of those animals
which had not crossed the obstruction when tested after shorter
periods of starvation. Nevertheless he appears to conclude that the
drive constantly increases with an increase in the period of starva-
tion. Presumably we can conclude, on this principle, that it reaches
its maximum at the moment the animal dies of starvation. The

72

THE HUNGER DRIVE

present study however, clearly indicates an optimum followed by a
decline.

Szymanski (10), studying the effect of the length of starvation
upon the rate of learning a problem involving a food reward found
increased efficiency with increased starvation periods. The longest
period he used was, however, twenty-four hours so it seems doubtful
whether his studies included the optimal period.

Dodson (1), studying the effect of variation in the length of
starvation upon the formation of a visual habit found an optimal
point at 41 hours. Animals starved for this period learned more
rapidly than those starved for periods of 24, 31 or 48 hours. The
differences, however, are not great. Furthermore he made use of
animals of but 78 days of age. It might well be expected that such
young animals would feel the deleterious effects of starvation
sooner than would the 185 day old animals used in the present
study.

The effect of starvation periods upon the amount of activity
registered by a rat in a freely revolving drum has been noted by
Richter (8) . Of the eight animals used in the starvation experiment
four were deprived of water as well as food and are therefore not
comparable to ours. Of the other four animals, two showed a peak
of activity after two days of starvation, one after one day while the
last animal showed constant decrease in activity after the removal
of the food. Unfortunately neither the ages nor the weights of the
animals are given.

Nicholls (7) using the same method on the guinea-pig found the
peak of activity on the second day followed by a decrease.

The work of Holden (2) is more nearly comparable to the present
investigation than any other. Since her results are quite different it
seems worth while to enumerate the differences in methodology
between her work and this.

The apparatus as used in the present work included the follow-
ing features not found in Holden 's: a tunnel-shaped obstruction
compartment provided with manually operated doors at the end,
a non-shortcircuiting stimulation grid, a restraining compartment
and automatic door built for the purpose of restraining the
stimulus animal in the study of the sex drive (but used throughout

THE HUNGER DRIVE

73

the present work), an illumination hood providing a constant source
of light which threw no shadows.

The animals used in the present work were of the experimental
colony strain of the Wistar Institute of Anatomy. They had been
raised until the age of 150 days with the sexes unsegregated. They
were fed while in this laboratory on a modification of McCollum's
standard mixture, which mixture was used as the incentive in the
experimentation. They were 185 days of age when tested. Holden 's
animals were purchased from Douredoure, Philadelphia. They were
raised with the sexes segregated. They were fed on bread and milk,
which food was used as the incentive in the experimentation. They
were 50-60 days of age when tested.

Several differences in procedure should be noted. Holden gave the
animals preliminary training on three successive days prior to
the test, rather than immediately preceding it, and to groups of ten
animals at once rather than to each singly. The test period was
ten minutes rather than twenty minutes in length. Holden used
three degrees of shock, the lowest of which was much greater than
that used in the present study, 1200 volts rather than 475 (the
external resistance in the circuit being the same in both cases) . Only
those results of Holden's concerned with the lowest degree of shock
will be considered here since the other degrees of shock did not yield
reliable data from the present standpoint.

Using starvation intervals of 0, 12, 24, 36, 48, 60 and 72 hours
Holden found a constant increase in the number of crossings up to
the 36 hour period which represents the peak of the curve. Further
increase in the interval gave constantly decreasing crossing. The
absolute values reported are entirely different from those found in
the present investigation. For example, the animals of the 36 hour
group crossed on the average 40.3 times in the ten-minute test
period. Using test periods twice this long we found that the most
frequently any animal of any group crossed was 36 times and that
the highest group average was 19.1. The decided difference in
absolute values reported must be due to the radical differences in
apparatus and procedure. That Holden found the peak of activity
at 36 hours and almost complete reduction of activity for those
starvation periods which in the present investigation resulted in the

74

THE HUNGER DRIVE

greatest activity is probably accounted for by the fact that the
animals she used were immature while those used in our work were
full grown.

4. A comparison of the hunger drive ( on the basis of the data here
reported) and the sex drive (on the basis of data previously
reported)

No effort has been spared in these two studies to obtain data
which could justly be compared. Animals of the same stock and
age were used. The general conditions, feeding, social and so on
were the same. The same apparatus was used and no differences
in procedure or technique were introduced. All the data for both
studies were taken by the same investigator and he made every
effort to handle the animals and the apparatus in a uniform way
throughout. The method of handling animals in experimental work
has been given little consideration in the past. Perhaps in studies
of maze learning it matters little whether the rat is lifted by the tail
or otherwise. It is the writer 's opinion that in the study of animal
drives too much care cannot be taken on this point. Slight emotional
disturbances may not seriously prolong the rat 's learning of a maze
but it will serve decidedly to alter the behavior of the animal toward
an incentive object. A criticism to be raised against the obstruction
method in particular is that results obtained by it are too readily
influenced by the technique of the individual investigator in hand-
ling the animals. If comparable data are to be obtained by two
investigators it seems almost essential that they work together for
a time in order that they may adopt the same manner and rhythm
of handling the animals. The details of this technique are difficult
to transmit exactly in writing. An attempt has been made to enum-
erate the important points in the report of the study of sex
behavior.

The value of the obstruction, i.e., the physical characteristics of
the current in the grid, was the same in the study of the two drives.
In brief, the only variable consisted in the type of incentive used,
food in the one case and sex object in the other.

In the study of the hunger drive the physiological condition most
closely related to this phenomenon (the effect of starvation) was

THE HUNGER DRIVE

75

varied in order that the optimal condition might be determined. The
same holds for the study of the sex drive although the physiological
conditions most intimately involved in this case were not the same
for the male and the female. With the male it depended upon the
length of the sex deprivation interval, with the female it depended
upon the oestrus condition.

In the study of the hunger drive the sex factor was kept constant
and at the minimum in the two sexes, in the case of the male by

TABLE 5

Showing temporal distribution of approaches, contacts and crossings during the
twenty-minute test period, in one-minute i/ntervals

GROUP

ACTIVITY

1st minute 1

K
EH

p

3rd minute 1

B

EH

p

EH

-r

5th minute 1

E-

W

Eh

<£>

7th minute 1

8th minute 1

9th minute 1

10th minute 1

1 11th minute 1

1 12th minute 1

13th minute 1

! 14th minute 1

s
§

Eh

| 16th minute 1

1 17th minute 1

| 18th minute J

H

EH
fa

si
EH

OS

| 20th minute 1

days

Approaches

2

4

5

1

3

0

3

2

1

4

0

1

2

4

0

0

1

0

2

0

0 ■

Contacts

7

11

3

5

2

3

6

3

2

4

1

0

0

3

0

0

2

3

1

1

Crossings

10

6

8

3

4

5

0

2

1

1

1

0

2

0

2

1

0

0

2

0

Approaches

7

2

4

1

3

2

0

3

1

2

4

2

8

2

2

4

4

4

7

3

•

Contacts

4

3

2

3

3

6

3

4

7

3

7

9

15

7

4

5

5

8

12

7

Crossings

29

34

10

14

34

30

22

24

18

8

11

23

26

10

16

13

7

9

3

1

Approaches

6

11

13

4

8

8

4

7

1

5

4

3

6

4

0

2

6

3

2

5

3 I

Contacts

12

8

11

8

15

11

8

13

4

1

7

3

4

5

9

6

2

4

3

3

Crossings

24

10

14

15

12

17

14

31

13

16

19

24

24

20

13

17

21

26

26

S

Approaches

1

8

6

4

4

6

8

5

7

8

3

0

4

1

2

5

4

1

7

4

'1

Contacts

15

11

7

4

4

5

4

8

1

4

2

0

4

1

1

4

5

0

3

2

Crossings

20

18

24

21

11

15

21

20

18

15

13

17

18

16

23

22

26

19

14

10

Approaches

4

7

9

3

5

4

2

7

6

1

1

3

7

4

3

5

3

9

5

9

.1

Contacts

6

5

2

2

3

2

3

0

4

5

3

4

1

1

4

4

10

7

4

8

Crossings

14

11

9

15

15

13

8

16

18

10

17

22

13

16

9

7

11

14

21

23

Approaches

2

8

6

11

3

6

8

2

7

5

3

4

1

3

5

8

6

14

12

0

'I

Contacts

8

3

4

7

5

8

4

0

7

4

5

1

8

7

6

8

7

12

11

8

Crossings

6

5

9

4

7

G

9

14

7

4

7

11

8

4

12

13 10

9

5

8

using a long deprivation interval and in the case of the female by
using only animals in dioestrum. Similarly in the study of the sex
drive the food-seeking factor was kept constant and at the minimum
by testing the animals immediately after removal from a cage in
which there had always been sufficient food.

a. Comparison of the two drives in the male animal. Fairly to

76

THE HUNGER DRIVE

compare the two drives in the male we must compare those two
groups which represent each drive at the optimum, i.e., those groups
which displayed the maximum activity directed toward the incen-
tive in each case. Naturally, it would not do to compare a food-
seeking group of a given starvation period with a sex group of a sex
deprivation period of the same length, since our results show that
the two drives do not recover their strength after a period of satis-
faction at the same rate. The sex drive recovers much the more

TABLE 6

Showing temporal distribution in five-minute intervals of approaches, contacts
and crossings during the twenty-minute test period

STARVATION

Jr ±jltlUJJ

ACTIVITY

0-5TH
MINUTE

6-10TH

MINUTE

11-15TH

MINUTE

16-20TH
MINUTE

O

Number

Percent

Number

Percent

Number

Percent

Number

Percent

days

Approaches

15

42.8

10

28.6

7

20.0

3

8.6

35

. ' 1

Contacts

28

49.0

18

31.6

4

7.0

7

12.6

57

Crossings

31

64.5

9

18.8

5

10.4

3

6.2

48

Approaches

17

26.2

8

12.6

18

27.6

22

33.8

65

■ (

Contacts

16

13.5

23

19.5

42

35.5

37

31.4

118

Crossings

130

37.0

102

29.1

86

24.5

33

9.4

351

Approaches

42

41.3

25

24.5

17

16.7

18

17.6

102

Contacts

54

39.4

37

27.0

28

20.4

18

13.1

137

Crossings

75

20.6

91

25.0

100

27.5

98

26.9

364

Approaches

23

26.2

34

38.5

10

11.6

21

23.8

88

' {

Contacts

41

48.2

22

25.9

8

9.4

14

16.5

85

Crossings

94

26.0

89

24.7

87

24.2

91

25.2

361

Approaches

28

28.8

20

20.6

18

18.6

31

32.0

97

« {

Contacts

18

23.1

14

17.9

13

16.7

33

42.3

78

Crossings

64

22.7

65

23.0

77

27.4

76

26.9

282

Approaches

30

26.3

28

24.6

16

14.1

40

35.1

114

Contacts

27

20.6

29

22.2

29

22.2

46

35.1

131

Crossings

31

19.6

40

25.3

42

26.6

45

28.5

158

rapidly, reaching its maximum again in about 24 hours, whereas
the hunger drive does not reach its peak until about the fourth day.
It is the peaks of the two curves which should be compared or at
any rate those two groups which, of those tested by us, represent
the drives at their maximum. Our interest is in the limits of
behavior as influenced by each of these factors and how these limits
compare. The four-day starvation group in the case of the hunger

THE HUNGER DRIVE

77

drive and the one day sex deprivation group in the case of the sex
drive represent these limits in the male. These data are brought
together in tables 7, 8 and 9. It will be seen that the average
number of crossings to the sex object was 13.45, to the food, 19.1.

TABLE 7

Distribution table of those groups ivhich represent the sex drive and the hunger
drive in the male and in the female at the optimum. Sex drive in the male is
represented by the one-way sex deprivation group; sex drive in the female, by
the group of animals tested when in oestrum (the stage of cornified cells);
hunger drive in the male, by the four-day starvation group ; food-seeking drive
in the female, by the two-day starvation group. Distributions for the two groups
representing the sex drive have been combined as have also the distributions for
the two groups representing the hunger drive. These combined distributions
have been placed side by side in the center for ready comparison.

INTERVAL

6EX DRIVE
IN MALE

SEX DRIVE
IN FEMALE

SEX DRIVE,
TWO SEXES
COMBINED

HUNGER
DRIVE,
TWO SEXES
COMBINED

HUNGER
DRIVE
IN* MALE

HUNGER
DRIVE
IN FEMALB

o

1

2

1
X

3

1

l

1

4

1

1

2

5

1

i

X

1
I

1

6

1

1

7

g

9

2

1

3

10

2

2

11

2

1

3

12

1

1

13

2

2

14

3

3

15

4

2

6

2

1

1

16

3

3

6

17

1

2

3

1

1

18

2

1

3

3

2

1

19

2

2

20

1

1

21

2

2

2

1

1

22

1

1

23

1

1

24

1

1

25

2

1

1

26

27

28

1

1

29

3

1

2

78

THE HUNGER DRIVE

This difference is quite a reliable one, being 2.8 times the standard
deviation of the difference. We feel justified, then, in saying that,
using the obstruction method, the hunger drive is stronger than the
sex drive in the male animal.

b. Comparison of the two drives in the female animal. Here
again we must compare those groups representing the two drives at
their high points. These are the one-day starvation group in the
case of the hunger drive and the cornified group in the case of the
sex drive (see tables 7, 8 and 9). As in the case of the male there
is a difference in favor of the hunger drive. This difference is not

TABLE 8

Showing the number of animals in each group, the average number of crossings
and the standard deviations for those groups tafcen to represent the sex and the
hunger drives in the two sexes

NUMBER OF

AVERAGE

STANDARD

GROUPS

ANIMALS IN

NUMBER OP

DEVIATION

THE GROUP

CROSSINGS

20

13.45

4.3

Sex drive in female _ _ _

21

14.14

5.5

Sex drive in two sexes combined-

41

13.8

4.9

Hunger drive in male _ _ _ _

10

19.1

5.87

Hunger drive in female _ — _

10

19.0

8.91

Hunger drive in two sexes combined-

20

19.05

7.6

TABLE 9

Showing reliability of differences in the average number of crossings for the
sex and the hunger drives in the two sexes taken separately and combined

STANDARD

CHANCE8

DEVIATION

THE

DIFF./S.D.

IN 100 OF

GROUPS

OF THE

DIFFER-

OF DIFFER-

A TRUE

DIFFER-

ENCE

ENCE

DIFFER-

ENCE

ENCE

Sex drive in male and hunger drive in

2.0

5.65

2.8

99.74

Sex drive in female and hunger drive

2.9

4.91

1.7

96

Sex drive in the two sexes and hunger

drive in the two sexes. _

1.86

5.25

2.8

99.74

entirely reliable being 1.7 times the standard deviation of the differ-
ence (chances of a true difference: 96 in 100).

c. Comparison of the two drives, data for the two sexes being
combined. As would be expected, when the data for the male and
the female groups are combined the difference between the strength
of the two drives as here measured is a reliable one. The hunger
drive is clearly stronger than the sex drive (tables 7, 8 and 9).

THE HUNGER DRIVE

79

This conclusion is in accord with that of Tsai (11) who used the
choice method and with that of Simmons (9) who used the learning
method. Since all data, in so far as they are interpreted in terms
of general tendencies such as drives, are functions of the method
used there is good reason for attacking problems by means of any
and all methods at our disposal. Nevertheless we feel that the
results here reported are important not merely because they have
been obtained by a different method but also because they have been
obtained by a more reliable method than any others which have been
employed.

Using the learning method as Simmons did the problem to be
solved is a distraction which serves to reduce the prominence of
the incentive object in the entire stimulus situation. This reduces
the possibility of various incentives producing clear cut differences
in the animal's behavior. A minimum amount of learning is
involved in the obstruction method as here used. A very simple
association is formed in the preliminary trials, that between the
incentive object and the incentive compartment. In the course of
developing the technique used in these two investigations data
were obtained which indicated that the brief preliminary training
period used is quite as effective as is a much longer one. As tested
by the obstruction method the animal is presented with a stimulus
situation in which the incentive is dominant. It is not distracted
by the complexities of threading a maze or solving other problems,
before being stimulated by the object which will call out the drive
under investigation.

The choice method does not suffer so much from involving con-
siderable learning. Two simple associations must be formed, how-
ever, instead of the one demanded by the obstruction method, and
the actual movement of the animal is less simple. Furthermore it
seems probable that chance or uncontrolled factors must always
play an important part in the choice method. The animal fre-
quently responds simply to whichever incentive happens to stimu-
late it first in spite of the experimenter's efforts to have the two
incentives stimulate it simultaneously and equally. The choice
method suffers from a further and more serious difficulty which
was apparently not recognized by either Moss or Tsai. To obtain

80

THE HUNGER DRIVE

a choice which is to represent a fair comparison between the two
drives each acting at its optimum it would be necessary to control,
simultaneously, the related physiological conditions. In the case of
the female this would involve prediction of oestrous condition, no
simple task, and one in which manipulation of the animal prior to
testing could scarcely be avoided. In the case of the male it is true
that, perhaps on the basis of the results reported herein, one could
start the period of starvation at one time and the period of sex
deprivation, in the same animal, at another in such a way that
theoretically the two drives would reach their optimal points simul-
taneously. It is improbable, however, that the physiological condi-
tions effecting the two drives are entirely independent of each other.
Starvation would effect sex behavior; mating would effect food-
seeking behavior. Each drive would be tested under the artificial
conditions imposed by the control of the other.

Summary

1. The tendency of the male white rat to approach food as
measured in terms of the number of times it will cross an electrical
obstruction within a given period of time is at its low point when
the animal is tested immediately after being removed from a cage
containing food. This tendency increases with an increase in the
length of the starvation period, up to a period of four days, and
from this point on, decreases.

2. In the case of the female white rat this tendency reaches its
high point much earlier. The tendency after one day of starvation
is apparently almost as strong as it is in the male after a four-day
period of starvation. After one day of starvation the tendency in
the female declines gradually and then more rapidly, never after
the second day being as strong as in the male.

3. Comparison of the groups which, of those tested, represent
the hunger and the sex drives at their maximum, indicates that the
tendency for a white rat to approach a food object is stronger than
its tendency to approach a sex object. This is true of rats of both
sexes although the difference is not so great in the case of the
female.

2. THE EFFECT OF DELAYED INCENTIVE ON
THE HUNGER DRIVE IN THE WHITE RAT
(EXPERIMENT 1: THE OBSTRUCTION
METHOD)1

E. L. Hamilton
I. Introduction

In experimental studies of animal motivation, the organism is
usually required to perform some task in order to obtain the incen-
tive. Ordinarily, the latter is accessible to the animal immediately
upon completion of the performance. This is perhaps the best pro-
cedure for most experimental work on animals, since a more favor-
able emotional adjustment to the artificial laboratory situation tends
to be made under this condition. An interesting problem from the
standpoint of general motivation, however, is that of separating
the incentive stimulus from the required performance by various
intervals of time, thereby introducing the factor of delayed incen-
tive. The question arises as to the effect of various intervals of
delay between the performance of the task and the incentive on the
strength of the drive under investigation. Is the drive weakened,
strengthened, or in no way affected when the reward does not imme-
diately follow the performance?

In the previously reported studies of drives (hunger, thirst,
normal sex, segregated sex) under this project, the matter of
delayed incentive has not been considered. In all of these investi-
gations the animal was allowed contact with the incentive object as
soon as it crossed the grill and entered the incentive compartment.
The present study is an attempt to measure the effect of delayed
incentive on the strength of one of these drives — the hunger drive.
In Appendix 3, experiment 2, making use of the maze and the

1 Reprinted with modifications from Genetic Psychology Monographs, 1929,
5, No. 2, 137-66. For balance of monograph, covering experiment 2 on maze
learning, see Appendix 3.

81

82

THE HUNGER DRIVE

learning method of measuring the influence of delay on behavior,
will be found. Although the obstruction method has, we believe,
many advantages over the learning method as a means of determin-
ing the effect of delay upon strength of drive, it will be seen that
the results obtained by the two methods show marked agreement in
general.

II. Experiment

The obstruction method

Method and procedure. The apparatus employed in experiment 1
was the Columbia Obstruction Apparatus which has been described
in detail in Part I, 1, of this volume. It consists of four compart-
ments: A, the entrance compartment; B, the obstruction com-
partment; C, the reaction compartment; and D, the incentive
compartment. In the present study, the animal was delayed in com-
partment C for various intervals of time as described later. The
electrically operated grid device, developed by Dr. T. N. Jenkins
(see Part IV, 2) was used in giving the animal the initial shock
on the fifth trial. The standard shock was, of course, employed
in the present experiment. For Fig. 1 of this monograph, showing
apparatus, see Part I, 1, of this volume.

A total of 100 albino rats of the "Experimental Colony strain"
of the Wistar Institute of Anatomy were tested. Until shipped to
us at about 150 days of age, they were kept in litter groups unsegre-
gated as to sex, an approximately equal number of males and
females making up each group. Upon arrival at our laboratory,
they were segregated to the extent of placing males and females in
separate cages in which condition they were kept until they reached
the age of about 185 days. All animals fell within the age range of
175 to 196 days at the time of experimentation. The age range at
the time of segregation was 143 to 157 days. The segregation period
extended from 32 to 39 days. This period provided ample time for
the animals to become accustomed to the laboratory conditions and
also insured the availability of non-pregnant females for use in the
experiment, since litters are born about three to four weeks after
the animals have become pregnant.

The animals were fed a well-regulated diet before they were

THE HUNGER DRIVE

83

received by us. In our laboratory they were fed McCollum's Stan-
dard Diet (see 11, p. 27) with the addition of a weekly ration of
greens. More than a sufficient amount of food was kept in the cages
to insure uniform conditions with respect to hunger. We could be
practically certain that, when the starvation period began, the ani-
mals were all at about the same stage of hunger. A small dish of
McCollum's diet was also used as the incentive in the experiment.
Fresh water was supplied daily in inverted bottles with nozzles to
prevent fouling. The animals were thus always supplied with water,
even during the 48-hour starvation period preceding the test. At
the beginning of a starvation period, fresh sawdust was provided to
insure complete absence of food. The general laboratory conditions
were precisely the same as those obtaining in all other studies com-
ing under the project.

Four periods of delay, 15 seconds, 30 seconds, 1 minute, and
3 minutes, were investigated, groups of 20 animals being used in
each cage. The 48-hour starvation group representing the maximum
strength of the hunger drive (14) was employed as a control. The
exactitude of the procedure followed, in connection with the Ob-
struction Apparatus in this series of researches clearly warrants
the use of the scores from one experiment to another as here
employed (12). The grouping of the animals is shown in Table 4.
It will be seen from the table that, although 11 males were used in
the delay groups, only 9 females were employed in these groups.
The reason for this is a practical one. The females could be used
only when in dioestrum, a stage in the oestrom cycle, determination
of which will be discussed below in connection with procedure. This
stage could not be detected before testing since the smears were not
made until after the animal was tested. A large number of females
which had been run could not be used. Only 36 females out of 101
tested proved to be clearly in dioestrum. Naturally, the records of
those in some stage of oestrum had to be discarded since the scores
on such animals involved both hunger and sex drive. This difficulty
does not arise with the males and hence male groups could be easily
made at any desired size.

The starvation period was constant for all groups, consisting of
a 48-hour interval during which the animals received no food,

84

THE HUNGER DRIVE

directly preceding the test period. This interval was employed,
since it represented the maximum strength of the hunger drive, as
above stated, and individual and group differences should best be
brought out by the use of maximal drive conditions.

TABLE 4

Grouping of Animals

NUMBER OF ANAMALS

LENGTH OF DELAY PERIOD

Male

Female

Total

0

(Control) _

10

10

20

15

Seconds _ _

11

9

20

30

Seconds _ _

11

9

20

1

Minute

11

9

20

3

Minutes

11

9

20

Procedure. The sex drive was controlled in the males by use of
the segregation period of 35 days, which insured uniform sex
deprivation. Since Warner (13) found that the sex drive in the
male is at its minimum with the longest interval of sex deprivation
he studied (28 days), we assumed that with a still longer interval
of 35 days it would probably be still weaker or at least as weak as at
28 days. This would insure little interference or complication by
the sex drive. In the case of the females, the matter of dioestrum
was determined by the use of the smear technique developed by
Long and Evans (5) and employed in the drive studies in the Co-
lumbia laboratory. Only those animals in the inactive period (di-
oestrum) were used. As mentioned above, this necessitated discard-
ing the data on a large number of females. Determination of the
oestrous condition was made from both the quick smear and the
permanent slide by the experimenter. Dr. Warner, who has had
considerable experience with the technique, checked the determina-
tions. In all cases the determination was made on the basis of the
histological picture and without consideration of the behavior data.
To eliminate olfactory stimulation connected with the opposite sex,
animals of only one sex were run on any one night and the box
was washed with soap and water and allowed to air for each night 's
experimentation.

Time of day was controlled by testing all animals between the
hours of 9 P.M. and 4 A.M. Since the laboratory was practically
deserted at this time, noise distractions were also reduced to a min-

THE HUNGER DRIVE

85

imum. The temperature of the laboratory was kept fairly constant
by thermostatic regulation.

The method of handling the animals was made uniform by pre-
vious training in which the method of picking up the animal and
transferring it at a constant rate to the entrance compartment was
reduced to a habit. The time for moving the animal from the in-
centive compartment back to the entrance compartment was about
30 seconds. Naturally, the fact of emotional arousal in motivation
situations makes it necessary to take every precaution to avoid
distractions of every sort.

The preliminary training given the animals did not differ from
the usual procedure Avith the Obstruction Apparatus. The animal
was placed in the entrance compartment and after 10 seconds was
allowed to cross the grid, which was not electrified, to the food in
the incentive compartment. After the animal had obtained a nibble
of food, it was removed and again placed in the entrance compart-
ment. After four crossings to the food in this manner, the animal
was placed in the entrance compartment, but this time, upon clear-
ing the door between the entrance and obstruction compartments,
the animal received a shock from the grid. The shock was given by
means of the indirect method of electrifying mentioned above in
connection with description of apparatus, the closing of the door
dj_) automatically causing current to flow in the grid. After the
animal had obtained a nibble of food on the fifth trial, it was re-
moved from the incentive compartment and the test period began.
An exception to this procedure was made in the case of such ani-
mals as did not take a nibble of food upon reaching the incentive
compartment but engaged in exploratory activity by climbing up
the sides of the box, etc. If the animal had not taken a nibble of
food at the end of 60 seconds, it was removed, nevertheless, and re-
placed in the entrance compartment. The exception was also ob-
served in the study of the normal hunger drive (14) from which
our control group was taken.

The procedure for the test period for the control group consisted
in placing the animal in the entrance compartment and after 10
seconds releasing it into the obstruction compartment as in the case
of the preliminary training. However, the animal now received a

86

THE HUNGER DRIVE

shock as soon as its feet made contact with the grid, since the direct
method of electrification described in the preceding section had
been set in operation for this and all later tests. The following
types of behavior, which are characteristic for the Obstruction
Method, were recorded on suitable blanks minute by minute: ap-
proaches, contacts, crossings.

An "approach" refers to orientation of the animal before the
grid with its head projecting into the obstruction compartment fol-
lowed by withdrawal without touching the grid. A "contact" re-
fers to the touching of the grid by the animal followed by imme-
diate withdrawal again. A "crossing" refers to the crossing of
the grid by the animal and its entrance into the delay compartment.
Quantitative data on these three types of behavior during a period
of 20 minutes were recorded. Whenever a crossing occurred, the
animal was removed from the incentive compartment after the
usual nibble of food and returned to the entrance compartment.
It was not detained for 10 seconds before being released into the
obstruction compartment as before, but was allowed immediate
access to the grid.

The procedure for the test period in the case of the delay groups
was similar to that for the control group with the exception that
the animal was not allowed immediate access to the food. Upon
crossing the grid, the animal was held in the delay compartment
for an interval of time, at the end of which it was permitted access
to the food. In order to retain the animal during the interval of
delay, d3 was disconnected from the automatic delease (E) and
operated manually. The operation of the door by the experimenter,
after the period of delay, soon became mechanical, so that as soon
as the stop-watch indicated the end of the delay period, the door
was opened immediately. The animal was allowed a nibble of food
and was returned to the entrance compartment after each crossing
and delay period.

Quantitative behavior in the form of approaches, contacts, and
crossings for a period of 20 minutes was recorded, as in the case of
the control group. This time was exclusive of the delay periods. A
cumulative stop-watch was used to measure the time of the 20-
minute period. A second stop-watch was used to time the delay

THE HUNGER DRIVE

87

TABLE 5

Frequency distributions covering approaches, contacts, and crossings for the
various periods of delay

Number of
responses

6 S &

15 SECONDS 30 SECONDS 1 MINUTE 3 MINUTES

g s £ s £ s

s

3

2

1

1

1

1

1

1

2

1

3

1

1

0

3

l

J

1

1

2

1

1

3

1

2

1

2

1

1

1

1

2

4

1

2

1

1

2

1

0

2

1

1

1

1

i'

3

1

1

g

1

2

2

2

2

1

3

1

2

1

1

0

0

0

2

2

1

1

1

1

1

1

0

0

1

0

0

1

2

3

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

0

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

2

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

2

1

2

1

1

3

1

1

1

1

1

2

1

2

1

1

1

1

1

1

1

2

1

2

1

1

1

2

1

1

1

1

1

1

1

1

1

2

1

1

1

1

1

1

1

1

1

1

2

1

1

1

2

1

1

1

2

1

1

1

1

1

1

1

1

1

88

THE HUNGER DRIVE

periods. The cumulative watch was started at the beginning of the
test period. When an animal entered the delay compartment, the
watch was stopped and the second watch started. At the end of the
delay period the second watch was stopped and the cumulative
watch started again. When the latter registered 20 minutes, the
test period was considered completed. Thus, regardless of the num-
ber of delays, the time during which the test animal was exposed to
the incentive and might cross to it was the same for all groups, in-
cluding the control group. In addition to the quantitative data,
notes were made on the behavior of the animals during the delay
period. Each animal's weight was taken and recorded before the
period of starvation and again after testing.

Results

The results are presented in Tables 5, 6, 7, 8, and 9 and in Fig-
ures 2 and 3. Table 5 gives the frequency distributions covering
approaches, contacts, and crossings for the various periods of de-
lay. It reads as follows: In the control or 0-delay group, there
were 3 females and 2 males who made no approaches, 1 female and
1 male who made no contacts, and no females or males who made
no crossings, etc.

Table 6 represents the measures of central tendency and vari-
ability for each delay period, including the median, range, average,
standard deviation, and co-efficient of variability of approaches,
contacts, and crossings for each sex and for the combined (male
and female) groups.

Table 7 gives the measure of reliability of the difference between
the average crossings of various groups.

Table 8 shows the distribution of approaches, contacts, and cross-
ings during the 20-minute test period, for the various groups, min-
ute by minute. It reads as follows: The control group during the
first minute made 7 approaches, 4 contacts, and 29 crossings, etc.

Table 9 gives the distribution of approaches, contacts, and cross-
ings during the 20-minute test period in 5-minute intervals, in
terms of absolute scores and percentages.

The graph presented in Figure 2 shows the average number of

THE HUNGER DRIVE

XjinqBTJBA r-< Tt<

jo -}U9p^ao3 ^

•a -9

OHO | OS i-l O

OS wt> aoo

0> lO 05
CO CO CO

iTBipaj\[

S3

3

jo ^ue'iDiyaoo

oq io co

co oo co
t-oo>

t- o o

i-l CO (M

I I I

<M CnI C-q

a -s

©SBJ3AV

CO IO Tfi

05 HU5
IO IO LO

t*- CO CO

oq tr- io

UBip3J\[

io lo lo in io to

O T-H o
CM i-H CM
I I I
O T-H O I CM

LO LO
l> H H
I I I

O tH o

AJIIiqBUBA ^

jo ^uaioyjaoQ ^ ^ oo

•a -s

to

<£> LOU5

I 3

I oo oo

_<*_.,,., I w as as | co co co

I o o o o c- t-

I I iH

LO LO LO

UBip8J\[ CO CM CO I t-^O "*LOH

I rH t— I i— i I fqHN

I I I

LO 00

O LO
CM

I I

— CM

2 <e J

hi

90

THE HUNGER DRIVE

crossings, contacts, and approaches at each period of delay for the
combined (male and female) groups.

Figure 3 presents graphically the distribution of crossings dur-
ing the 20-minute test period in 5-minute intervals.

If we examine the scores on average crossings (Table 6) for the
combined groups and the graph (Figure 2), we note that there is a
decided drop with a 15-second delay. An animal subjected to such

Fig. 2. Average number of crossings (solid line), contacts (dot-dash line),
and approaches (dash line) at each period of delay for the combined (male and
female) groups.

a small amount of delay crosses on an average of 10.0 times, while
an animal which is allowed to proceed immediately to the incentive
crosses about 17 times, indicating a decrease in the drive with de-
lay.

The scores on average approaches show the following tendency :
An animal with 15 seconds' delay approaches about 18 times, while
one which is not delayed approaches about 3 times. This tendency
toward an inverse relation of approaches to crossings has been
found in the other studies on drives with the Obstruction Appara-
tus. The reason the number of approaches is larger for the delayed
group which crosses fewer times is that the control group was prob-

THE HUNGER DRIVE

91

ably affected by the incentive in such a manner that activity in the
direction of the incentive compartment nearly always resulted in
the full response of crossing, while the conditions of delay resulted
in the activity's taking the less complete from of approaching.
Stated in terms of drive, the hunger drive was decreased by the
conditions of delay, or the incentive value was not great enough to
elicit the complete response of crossing the grid so often.

Fig. 3. Temporal distribution of crossings for the combined (male and fe-
male) groups. The lower line represents the percentage of crossing which oc-
curred during the first 5 minutes of the test period. The middle line represents
the percentage which occurred during the first 10 minutes. The upper line
represents the percentage which occurred during the first 15 minutes.

The scores on average contacts were not materially different
under conditions of delay than under the normal incentive situa-
tion, indicating that th delay did not influence significantly par-
tial crossings.

The scores on average crossings for the 30-second and 1-minute
groups are only slightly different from those of the 15-second
group. The differences may perhaps be regarded as fluctuations
due to chance factors, so that we may conclude that there is little
difference between the effects of 15-second, 30-second, and 1-minute
delays. The scores on average approaches and contacts tend to
support this conclusion.

92

THE HUNGER DRIVE

However, there is a decided drop in average crossings with the
3-minute group, an animal delayed for this length of time crossing
on an average of only about 4 times as compared with 10 for the
smaller periods of delay and 17 for the 0-delay group. This indi-
cates a marked decrease in drive with delays as long as 3 minutes.
There is some indication that the animals in the 3-minute group
were less active than the other groups, since the number of ap-
proaches did not increase appreciably with the further decrease in
crossings and the number of contacts actually decreased slightly.
The incentive value was evidently so weak under this long delay
that there was elicited less general activity (approaches and con-
tacts), since an increased number of approaches might have been
expected with decrease in number of crossings.

The difference between the average crossings of the control group
and each of the delay groups (Table 7) is in all cases very reliable.

TABLE 7

Reliability of the differences between the average crossings of each group

GROUPS

DIFFER-
ENCE

S. D. OP THE
DIFFERENCE

DIFFERENCE
S. D. OF THE
DIFFERENCE

CHANCES IN
100 OF A TRUE
DIFFERENCE
GREATER
THAN 0

0

and 15 Seconds

7.6

2.03

3.74

100

0

and 30 Seconds

9.3

2.23

4.17

100

0

and 1 Minute

8.3

2.10

3.95

100

0

and 3 Minutes

12.9

1.95

6.62

100

15

Seconds and 30 Seconds

1.7

1.74

.98

84

15

Seconds and 1 Minute

.7

1.57

.45

67

ao

Seconds and 1 Minute

1.0

1.82

.55

71

15

Seconds and 3 Minutes

5.3

1.37

3.87

100

30

Seconds and 3 Minutes

3.6

1.66

2.17

98.6

1

Minute and 3 Minutes

4.6

1.44

3.19

100

We find a tendency for the reliability^ to increase as the period of
delay increases in length. There is also a tendency for the reliabil-
ity of the difference between the groups to increase as the differ-
ence between periods of delay increases. Thus, when we compare
delay groups in which the difference in length of period of delay is
relatively small, as in comparisons between the 15-second, 30-second,
and 1 -minute groups, little reliability is found. This supports
our conclusion that there is little difference in effect between de-
lays of 15 seconds, 30 seconds, and 1 minute. In delay groups in
which the difference in length of the periods of delay is relatively
large, as in the comparisons between the 3-minute group and the

THE HUNGER DRIVE

93

TABLE 8

Temporal distribution of approaches, contacts, and crossings during the 20-
minute test period in 1-minute intervals

DELAY
PERIOD

ACTIVITY

1st minute 1

Eh

C

CM

3rd minute I

4th minute I

5th minute

6th minute

-

3
H

r*
t>

8th minute 1

9th minute 1

10th minute I

1 11th minute I

12th minute I

13th minute 1

14th minute 1

15th minute 1

■ 16th minute 1

| 17th minute I

| 18th minute I

| 19th minute

| 20th minute 1

Approaches

7

4

1

3

2

u

3

1

2

4

2

8

2

2

4

4

4

7

3

0

Contacts

4

3

3

3

3

0

3

4

7

3

7

9

15

7

4

5

5

8

12

7

Crossings

29

34

19

14

34

30

22

24

18

8

11

23

20

10

16

13

7

9

3

1

A nnvnn r>hp<i

20

19

30

26

17

20

19

24

10

IS

24

17

8

9

8

15

15

14

15 Sec.

Contacts

7

9

7

10

12

12

3

5

6

7

5

9

4

4

1

2

10

4

5

5

Crossings

14

12

11

15

8

14

9

8

11

14

13

8

8

10

12

3

7

8

5

10

33

i -
1 (

22

25

29

22

q -

24

17

30

30

19

18

21

17

16

19

10

11

18

30 Sec.

Contacts

10

8

9

12

9

8

6

12

2

2

9

2

2

5

3

3

2

0

2

13

Crossings

15

9

12

4

8

7

6

11

7

7

6

9

5

8

0

9

8

10

9

9

Approaches

18

27

14

17

14

18

12

31

18

18

14

12

17

13

11

10

15

IS

13

20

1 Min.

Contacts

Q

o

8

5

.7

5

1

0

7

4

2

4

4

0

4

5

7

4

5

9

9

Crossings

18

12

13

11

13

8

8

8

8

11

5

9

10

6

9

8

7

5

7

9

Approaches

25

23

26

8

16

13

11

18

24

27

22

17

17

21

10

10

13

17

13

23

3 Min.

Contacts

10

4

13

7

0

4

1

2

0

3

2

2

0

1

2

0

2

0

1

0

Crossings

13

5

8

4

6

3

5

4

6

3

5

3

3

0

5

4

5

1

3

2

TABLE 9

Temporal distribution of approaches, contacts, and crossings during the 20-
minute test period in 5-minute intervals

0 -5th

6TH-10TH

11TH-15TH

16th-20th;

minute

MINUTE

MINUTE

MINUTE

DELAY PERIOD

ACTIVITY

fa

u

u

u

<

Eh

V

c

<v

fl

<D

O
Eh

1

o

n

«

u

«

1

O

u

(0
Ph.

1

(V

Ph

525

Ph

Approaches

17

26.2

8

12.6

18

27.6

22

33.8

65

0

Contacts

16

13.5

23

19.5

42

35.5

37

31.4

118

Crossings

130

37.0

102

29.1

86

24.5

33

9.4

351

Approaches

122

33.6

97

26.7

83

22.9

61

16.8

363

15

Seconds

Contacts

45

35.4

33

26.0

23

18.1

26

20.5

127

Crossings

60

30.0

56

28.0

51

25.5

33

16.5

200

Approaches

126

29.0

130

29.9

105

24.1

74

17.0

435

30

Seconds

Contacts

48

44.0

30

27.5

21

19.3

10

9.2

109

Crossings

48

29.1

38

23.0

34

20.6

45

27.3

165

Approaches

90

26.8

97

28.9

67

20.0

82

24.4

336

1

Minute

Contacts

28

26.7

20

19.0

23

21.9

34

32.4

105

Crossings

67

36.2

43

23.2

39

21.1

36

19.5

185

Approaches

98

26.8

93

25.4

93

25.4

82

22.4

366

3

Minutes

Contacts

34

56.6

16

26.7

7

11.7

3

5.0

60

Crossings

36

38.3

21

22.3

22

23.4

15

6.0

94

94

THE HUNGER DRIVE

other delay groups, we find that the differences are highly reliable.
Hence the drop in the curve with the 3-minute group seems to rep-
resent a genuine tendency.

There is little evidence of sex differences from the scores on aver-
age crossings (Table 6) with the exception of one group. In the
case of the 30-second group, the average of the females is consider-
ably lower than that for the males. Reference to Table 5 shows
that there were three females in the 30-second group which crossed
only once. In view of the high average score, this is an unusually
large percentage of animals with only one crossing. The notes on
behavior indicate nothing unusual in the behavior of these animals.
Their general behavior was perfectly normal and they were quite
active in exploring the entrance compartment. This apparent sex
difference for the 30-second group would seem to be due to some
accidental factor rather than to a genuine difference between the
sexes. At any rate, the other three groups show no such differences
so that no consistent sex differences can be claimed as resulting
from the delayed incentive factor.

Tables 8 and 9, which show the temporal distribution of ap-
proaches, contacts, and crossings during the 20-minute test period,
and Figure 3, showing crossings, do not reveal any significant differ-
ences between the control group, which may be called the "nor-
mal" incentive group, and the delayed incentive groups. Table 9
and Figure 3 indicate that the approaches, contacts, and crossings
occurred rather uniformly throughout the 20-minute period for all
of the groups. That is, delay apparently had no effect in bringing
about a change in the relative number of crossings, contacts, and
approaches during the early or later stages of the test.

Summary of Results

(1) The hunger drive, as measured by the Obstruction Method,
was markedly decreased by a period of delay as short as 15 seconds
between the crossing of the grid and the obtaining of the incentive
(43 per cent decrease in number of crossings).

(2) Longer delay periods of 30 seconds and 1 minute did not fur-
ther decrease the drive index.

THE HUNGER DRIVE

95

(3) A period of delay as long as 3 minutes effected a further
marked decrease in number of crossings (73 per cent from control
group).

(4) The data on approaches and contacts corroborated the con-
clusions based on crossings. The effect of delay tended to increase
the number of approaches as would be expected since the number
of crossings was decreased. The number of contacts remained ap-
proximately the same for all the groups.

Bibliography

(1) Dodson, J. D. 1917. Relative values of reward and punishment in
habit formation. Psychobiology, 1:231-76.

(2) Holden, F. 1926. A study of the effect of starvation upon behavior
by means of the obstruction method. Comp. Psychol. Monographs, 3,
No. 17, 45 p.

(3) Jenkins, M. 1928. The effect of segregation on the sex behavior of
the white rat as measured by the obstruction method. Genet. Psychol.
Monog., 3 : 455-571.

(4) Jenkins, T. N., L. H. Warner, and C. J. Warden. 1926. Standard
apparatus for the study of animal motivation. Comp. Psychol.,
6 : 361-82.

(5) Long, J. A., and H. M. Evans. 1922. The oestrous cycle in the rat
and its related phenomena. Mem. Univ. Calif., 6 : 148 p.

(6) Moss, F. A. 1924. A study of animal drives. J". Exp. Psychol.,
7:165-85.

(7) Nicholls, E. E. 1922. A study of the spontaneous activity in the
guinea-pig. J. Comp. Psychol., 2 : 303-30.

(8) Richter, C. P. 1922. A behavioristic study of the rat. Comp.
Psychol. Monog., 1, No. 2, 55 p.

(9) Simmons, R. 1924. The relative effectiveness of certain incentives
on animal learning. Comp. Psychol. Monog., 2, No. 7, 79 p.

(10) Szymanski, J. S. 1919. Die Abhangigkeit der Lerngeschwindigkeit
von der Antriebsstarke. Pflug. Arch. f. d. gesamte Physiol., 173 :
141-48.

(11) Tsai, C. 1925. The relation of the sex and hunger motives in the
albino rat. J. Comp. Psychol., 5 : 407-15.

(12) Warden, C. J. and H. W. Nissen. 1928. An experimental analysis
of the obstruction method of measuring animal drives. J. Comp.
Psychol, 8 : 325-42.

(13) Warner, L. H. 1927. A study of sex behavior in the white rat by
means of the obstruction method. Comp. Psychol. Monographs, 4,
No. 22, 68 p.

96 THE HUNGER DRIVE

(14) Warner, L. H. 1928. A study of the hunger drive in the white rat
by means of the obstruction method. J. Comp. Psychol., 8 : 273-99.

(15) 1928. A study of the thirst drive in the white rat by means

of the obstruction method. J. Genet. Psychol., 35 : 178-92.

PART III
THE THIRST DRIVE

Part III

THE THIRST DRIVE
C. J. Warden

The primary purpose of the project was to secure comparable
norms of the more important drives in the white rat, and in the
case of thirst no more than this could be done. In Part III, 1, will
be found the indices representing a series of thirst conditions with-
in such limits as to exhibit the maximum thirst drive in the white
rat. The hunger drive was controlled by allowing the animal access
to its usual diet (McCollum's powdered mixture) during the
periods of water-deprivation. The sex, maternal, and exploratory
drives were controlled in the same mariner as was indicated (see
Warden, The Hunger Drive). It was planned to determine the in-
fluence of delayed incentive on the thirst drive paralleling the
study on the hunger drive (Part II, 2) but this had to be given up
in order to carry through other work more in line with the general
purpose of the project.

1. A STUDY OF THIRST BEHAVIOR IN THE WHITE
RAT BY MEANS OF THE OBSTRUCTION
METHOD 1

L. H. Warner

The effect of water-deprivation upon animals has not been subject
to such extended investigation as has inanition following food-
deprivation. Kudo (2, 3) has made an anatomical study of the effect
of thirst upon the development of the white rat and especially
upon the relative loss of weight by various organs and systems.
His interest did not extend to any degree to the behavior of the
animal. Eichter (4) appears to be. the only investigator to have
observed the influence of water-deprivation upon the behavior of the
white rat. Using the rotating wheel apparatus for measuring
general activity he found that three animals " starved but permitted
to have water all the time showed a definite increase in activity
for the first two to three days after the beginning of starvation and
then a steady marked decrease to a. point of almost complete in-
activity on the eighth day." In contrast to this it is to be noted
that four animals deprived of both food and water showed a steady
marked decrease in activity immediately. This group reached
the point of complete inactivity already on the fifth day. " The
animals were presumably adult. Since no data are given on the
food- and water-deprived animals after the fifth day it is supposed
that they died soon after that point. Richter reports no data on
animals supplied with food but deprived of water. In a more recent
article (5) he makes a brief statement regarding the detection of
a thirst cycle in the animal : "A recess large enough for the animal
to enter and drink, but too small for it to lie down or turn around,
was built on one side of the cage and an inverted watering tube was
placed at the end away from the cage. The bottom of this recess
was made of a piece of aluminum pivoted at one end and supported

i Reprinted from The Journal of Genetic Psychology, 1928, 35: 178-92.

100

THE THIRST DRIVE

101

on a tambour at the other, so that whenever the animal entered to
drink, a mark was recorded on a smoked drum. The records . . .
show that the rat drinks about ten times a day at intervals of two
and a quarter hours. The periodicity of the thirst response is quite
as remarkable as that of the hunger response, in view of the current
conception that the rats eat and drink at very frequent and irregu-
lar intervals."

One other passage in the same article relates to thirst behavior as
casually observed in the case of a single animal : ' 1 This animal had
habitually deposited its feces in the water-cup. Usually the water
was changed every day, but on one occasion, by some neglect, it
was not changed for several days, so that the resulting odor became
very unpleasant. At this point the animal started to cover the hole
over the water-cup. It first removed part of the upper layer of the
cardboard bottom of the large central cage, and dragged it into the
water-box. It placed the cardboard over the cup and smoothed it
down on all sides until the hole was perfectly covered. Then from
the bottom of the central cage it lifted stones larger than its head
three inches into the drinking cage and placed them over the card-
board cover. Besides the large stones numerous pebbles and sticks
were used until the water-box was completely blocked. The animal
had cut off its only water supply by this performance. Since we
wished to see what it would do when it became very thirsty, the
material was left undisturbed and no other water was given. After
three days, the animal pushed all of the sticks and stones from the
drinking cage into the large central cage, tore up the cardboard
seal, and drank its fill of the polluted water. ' '

The present paper reports the third of a series of studies on
drives in the normal white rat. In two of the several investigations
on animal drives undertaken in this laboratory recently, data were
obtained on the sex and on the hunger drive in the normal male
and female white rat (7, 8). What may, for convenience, be termed
the thirst drive is the subject of the present report. Since the
standardized method, apparatus, and animals have been thoroughly
discussed in the above-mentioned articles, the writer will here
restrict himself to a very brief description of such general condi-
tions and will detail only such matters as relate specifically to the

102

THE THIRST DRIVE

study of the thirst drive. The term " thirst drive" is used merely to
describe the tendency of animals to approach water under certain
conditions.

The Method

The method used in these studies has been termed the obstruction
method (1). It rests on the principle that the energy released by
internal stimulation can be directed against an obstruction of known
value and measured by its ability to overcome this obstruction. As
the method is applied in these studies the obstruction consists of the
stimulation resulting from contact with an electric grid which is so
placed that the animal must of necessity cross it in order to reach the
object which will "satisfy" the drive under investigation. In the
case of the thirst drive this object is, naturally, water. The most
important data consist of the records of the number of times in a
test period of constant duration that an animal will cross the grid
and reach the object in question. Perfect specificity in our measure
of a given drive cannot be obtained. It is, however, possible to
arrange conditions so that, at the time of testing, the drive in
question dominates the behavior of the organism. This is accom-
plished, first by the establishment of a physiological condition con-
ducive to the operation of that particular drive and non-conducive
to the operation of other drives. For example, in the study of the
thirst drive, the animals are water-deprived for various periods
of time prior to testing but are not deprived of food, nor tested in
those conditions (oestrum in the female, relatively short period
of sex-deprivation in the male) which previous study has shown to
result in a strong sex drive. Besides this control of the internal
condition of the animal, the external situation is so arranged as
to call out the drive under investigation only, or, at any rate,
dominantly. Thus in the present study the animal reached water
upon crossing the electric grid but did not find an object appeal-
ing to any other strong drive, such as food, or a receptive animal
of the opposite sex. In a. preliminary training series the animal had
the opportunity of associating water with the apparatus, but not
any other incentive object. Needless to say, a given animal is used
in the study of but a single drive.

THE THIRST DRIVE

103

Apparatus

The only modification made in the obstruction apparatus relates
to the smaller section of the reward compartment (for cuts see 1
or 7). It was in this small compartment that food had been placed
in the study of the hunger drive and a sex object in the study of the
sex drive. A piece of wire mesh similar to that used in the animals '
living* cages was constructed across the end of this compartment
shortening it about two inches. Hung on the outside of this was a
small inverted bottle whose curved nozzle projected in through a
hole in the mesh about one inch above the floor of the compartment.
This exactly duplicated the drinking situation which the rats were
accustomed to in their living cages. The copper nozzle was, how-
ever, so manipulated that the water would not flow from it as
copiously when the end of it was licked by an animal as was the
case ordinarily. The rat could little more than moisten its tongue on
the nozzle's end.

In other respects the apparatus was similar to that used in former
work.

Animals

White rats of the Experimental Colony Strain of the Wistar
Institute of Anatomy were used. They were shipped to us from the
institute at the age of 150 days and were used by us at the age of
185 days. While in this labor sdory they were fed on a modified
McCollum's mixture, plus greens. These greens were withheld only
during the water-deprivation period described below. This was
done because of their high water content. The dry food mixture
was at no time withheld, being supplied in a more than sufficient
amount even during the water-deprivation periods. The water con-
tent of this mixture was negligible. The animals were thus kept
constantly provided with sufficient food up to the very moment of
testing for the purpose of reducing the hunger drive to the mini-
mum. Even though so provided, the food intake fell off decidedly
in the longer deprivation periods. There seems no way of attaining
water deprivation without involving a certain degree of food-
deprivation. Forced feeding was impossible for practical reasons
and would probably result merely in upsetting the animals. That

104

THE THIRST DRIVE

the hunger drive was as a matter of fact operating far less effec-
tively than the thirst drive in animals under these conditions was
indicated beyond question by observations upon animals other than
those actually used in this work that were similarly water-deprived
but with food present. Such animals removed to a fresh cage con-
taining both a food dish and a water bottle invariably fought for
the privilege of licking the water nozzle and neglected the food
during the first half hour — a period as long as the entire testing
required.

The males were tested after a 35-day period of sex-deprivation.
The females were tested only if in dioestrum, i. e., after an animal
was tested, a vaginal smear was taken and examined under the
microscope and the data on the animal were not used unless this
examination showed it to be in dioestrum. For the reasons for
adapting these conditions see (8).

Procedure

This varied from that used in the other studies with respect
to the criterion for the return of the animal to the entrance com-
partment after a crossing, and here it was quite analogous to that
used in the hunger study. Instead of being returned after one
nibble of food it was returned after it had once licked the water
nozzle. With respect to preliminary training, length of test period,
use of the automatic door, amount of electric shock, etc., the proce-
dure of the former studies was followed exactly. As before, an
' ' approach ' ' refers to orientation of the animal before the grid with
its head projecting into the obstruction compartment followed by
withdrawal without touching the grid; "contact" refers to the
touching of the grid by the animal followed by withdrawal again ;
"crossing" refers to the crossing of the grid by the animal and its
entrance into the incentive compartment.

Grouping was based upon length of the water-deprivation period
and is indicated in Table 1. It will be noted that each group con-
tained 20 animals, 10 of each sex. Deprivation periods were begun
at about eleven in the evening by the removal of the water bottles.
All animals were tested at night, between 9 p.m. and 3 a.m.

THE THIRST DRIVE

105

TABLE 1

Showing groups of animals and the number of each sex in each group

LENGTH OF WATER-DEPRIVATION PERIOD

0

1 day
2
4

6 days

NUMBER OF ANIMALS

Male

Female

10

10

10

10

' 10

10

10

10

10

10

TABLE 2

Distribution table covering approaches, contacts, and crossings for the various
periods of water-deprivation or male and female rats

•* •> «■

O re
< 6 O

§ 2

u o

£ -

£ "m £ "re £ "re

en 4> ^ OJ ~ <V

£"73 E ~ £ "cj

*^ o " 4> 3 <u
£ *^ £ "w £ "H

<TJ 4) <"3 <y rn ^

£ H £ E "3

o
i

2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
30
31
32
33
34
35
36
37
38
39
40
41
42

2 1 5
2 2
2 4 3
1

1 1

1 3

1 2
1

1 1
1 1

1 2

2 1

2 1

1 1
1

1 1

1 1 1
1

1 1

1 1

1

1 1
1 1
1

1

1

2 3 1
1

1 1 2

1 2

1 1 1

2 I
1 1

1 1

1 5 4

2 2 2
1 1 1

1 1

2 1 1

1 1
1

1

1

2 1

4 4
2 4
1 2

2
1

106

THE THIRST DRIVE

Results

Table 2 gives the distribution of the approaches, contacts, and
crossings for the various groups. The medians, ranges, averages,
standard deviations, and coefficients of variability are found in
Table 3. Table 4 gives the reliability of the differences between the
average crossings for various groups. The approaches, contacts,
and crossings as distributed minute by minute through the twenty-
minute test period are given in Table 5. In Table 6 these data are
combined in five-minute units and are also reduced to percentages.
Figure 1 shows graphically the effect of variation in the water-
deprivation period upon the average number of approaches, con-
tacts, and crossings, the data for the two sexes being combined.
Figure 2 graphically demonstrates the effect of variation in the
water-deprivation period upon the temporal distribution of cross-
ings during the test period.

It will be seen at once that the average number of crossings is
affected markedly by the length of the water-deprivation interval.
At fiTst it was planned to use the intervals, 0, 2, 4, 6, 8 days, to
correspond with the groups studied in the case of the hunger drive.
The 8-day group had to be given up because of the high mortality
rate. During the taking of the data on the other four groups casual
observation of the animals during the deprivation period suggested
that the point of greatest activity lay between the 0-, and the 2-day
groups. For this reason the 1-day group was run and proved to
represent the optimal interval of those studied, for the production
of thirst drive. The average number of crossings for the 0 group is
4.15, somewhat but not reliably higher for the corresponding
group in the hunger drive. The average leaps to 20.4 after one day
of water-deprivation. This figure represents the highest average
thus far obtained in the study of normal drives although it is not
reliably higher than the high point for the hunger drive. These
animals crossed and recrossed the grid to the water almost as
rapidly as they could be returned, especially during the early part
of the test period.

In fact the frequency of their crossing approaches the maximum
possible under the conditions imposed by our procedure as deter-

THE THIRST DRIVE

107

mined recently in this laboratory (6). Approaches and contacts
are also higher for this group than for any other.

The average number of crossings for the 2-day group shows a de-
cided though not entirely reliable drop, a decline which is continued
by the 4-day group. This result stands in decided contrast to that
for the hunger drive which remains virtually on a plateau for 2, 3,
and 4 days of starvation. In general appearance, too, there is a
difference between animals water-deprived for 4 days and animals

5 d a a

Fig. 1. Showing average number of crossings (solid line), contacts (dot-
dash line), and approaches (dash line) for each interval of water-deprivation.

starved for the same period. The latter are the more active. The
former customarily rest in the corners of their living cages with
their heads curled up in under their chests. Whereas after but one
or two days of water-deprivation they are often observed to stick
their noses and tongues out of the opening where they had formerly
found the water nozzle, after four days' deprivation such seeking
behavior largely disappears. Small blood clots are often to be ob-
served in their nostrils. After six days of water-deprivation the

108

THE THIRST DRIVE

same symptoms are seen not merely in a few but in practically all
animals. In three cases there was indication of semi-circular
disturbance but whether this was due to the lack of water or not,
cannot be said. Two deaths occurred in this group, two other ani-
mals being water-deprived and tested in order to bring the group
up to the regular size. Had the dead animals been scored as not

10

20

30
40
50
60
70
80
90
100

Fig. 2. Showing temporal distribution of crossings. The upper line repre-
sents the percentage of crossing which occurred during the first five minutes of
the test period. The middle line represents the percentage which occurred dur-
ing the first ten minutes. The lower line represents the percentage which oc-
curred during the first fifteen minutes.

having crossed, the average would, of course, have been lowered.
As it is, the average number of crossings for the 6-day group is
reliably lower than that for the shorter period of water-deprivation.
It is interesting to observe that there is a rather consistent decline
in approaches and contacts with an increase in the deprivation
period beyond one day. This is to be contrasted with the results

THE THIRST DRIVE

21

Cu
c

<

JO 4U3I0m3O3
U0I}BIA3p
p.lBpUBJS

33BJ3AV

aSuBa
uBipap^

XjIJTqBIJBA

jo )U3jom303

OOIJBIASp
pJBpUBJS

33BJ3AV

3Sub^j
uBipaj^

XjIJTqBUBA

jo }uYpuj303
aopBtAap

pjBpiIBJS
33BJ3AY

3SUB^

uotjBAudap-jajBM

M09>

T-t

NO T*» N©

vo vo vo

NO ^ VO

^< •+

in N no

O to to

osoo

CO _ ON

^ NO ^

to

to r~-

On vo O

m ^ t»i

4 CM CM

00 ON 00

NO vo no

VONh

to t>.

vo r^- no

On vo CM

4 to 4

a H 6
*-* CM CM

W NO NO

4 CM to

NONts

7 7 7

o o o

777

to CM CM

CM O CM

777

4 CM CM

CM On CM
to CM to
1 1 1

M H H

0-20
0-19
0-20

OO

VO VO

to vo

to

~ CM CM

NO N N

-h N O

M VO ^
00 On *

on vo

OO ON ON

Is «h ts
CM « -<

VO tO ON

O to — <

Ov H fl

on r^-

to On

On CM no

ON h

w

to tJ- to

4 t\5 4

vo

no 2

vri

00 tf\ 1H

vo

vo

00 ON to

©4-1

4 NO VO

to 4 4

4 4 4

0-11
0-19
0-19

O NO VO

7 7 7

^ in \a

NO ^ NO

XX X

ooo

o o o

ooo

1 1 !
OOO

vo

vo

vo

rl- -H- ri-

(1 «1 (fl

^ N

^ vo n©
VO >^

oo rt oo

WON
O no 00

i— I OO CM

ff\ SO M

N M »h

>T N NO

WOOts

ON t-» VO

CM 4 CM

4 to 4

to to to

to 4 CM

vo

CM > i-J

vo

ts O0 CM

t>I

4 vo 4

to to to

h'"6h

ON vo ©n
1 1 1

NO CM so

7 7 7

CM CM
1 1 7

op to oo

vo cm vo

ooo

O CM O

OOO

OOO

1 1 1
ooo

vo

VO

vo

vo

CM CM CM

NO NO

4 •+

CM to to

U

femal
male
total

femal
male
total

femal
male
total

fema
male
total

femal
male
total

o

to

PJ

91

to
CM tS

on
to
PJ

<«- -o

on
to
PJ

NO *0

110

THE THIRST DRIVE

for the hunger drive. In the latter case after prolonged starvation,
the number of crossings was reduced, but the number of approaches
and contacts increased. This was interpreted as indicating a strong
drive to get the food but probably a physical weakness rendered
the animals unable to overcome the resistance offered by the shock.

The animals oriented definitely in the direction of the food but
often hesitated to cross. In the case of the thirst drive a number
of 6-day water-deprived animals did not even show this definite
orientation after having crossed several times.

TABLE 4

Showing reliability of the difference between the average crossings of Adjacent

Groups

GROUPS

STANDARD
DEVIATION
OF THE
DIFFERENCE

DIFFERENCE
BETWEEN THE
AVERAGES

DIFFERENCE

S D. OF
DIFFERENCE

CHANCES
IN 100 OF A
TRUE DIF-
FERENCE

0 and 1 day

2.69

16.25

6.0

100

1 and 2 days

3.8

4.4

1.16

87

2 and 4 days

3.4

2.4

2.70

76

4 and 6 days

2.36

6.4

2.7

99.7

It will be noted that the temporal distribution of crossings is
uneven throughout. From 28% to 51% of the crossings always
occurred during the first five minutes of the twenty-minute test
period. Especially in those groups where the drive was operating
vait its strongest is there a predominance of crossing early in the
period. This is quite different from what was observed for the
groups in which the hunger drive was operating strongly. The
writer is inclined to think that the internal stimulation at the basis
of the thirst behavior was more readily rendered ineffective by the
moistening of the tongue which accompanied each crossing than
was the internal stimulation underlying the hunger behavior by the
nibble of food which accompanied each crossing in that case. The
sum total of the minute bits of food obtained would scarcely suffice
to inhibit the hunger contractions of the stomach. If the stimula-
tion related to the thirst behavior is at least partially derived from
dryness of the mouth parts, several crossings would be sufficient to
reduce the efficacy of such stimulation to a considerable degree.
If this is the case, the conditions which we have set up operate in
favor of the hunger drive. If we were to take as our measure of

THE THIRST DRIVE

111

to SO PN

y—l

tt NO VI

H H rt

mi

hOvo

tO tO CM

^- to

tO O NO

»N H W

mi

© CM

^ W NO

to NO 00

i-i to On

to © to

WI

O p*

to O M
1-1

NO O NO

CM O no

O to to

V9I

CM O O

NO tO T*-

CM

T*- tO

ONN

CM CM no

VSI

so o o

to OO ©

CM

*N i-« to

CM to

O «0 vrs
CM

tJ- O h
ph

O i-i On

— —

* ON
i— i

to ri- no

NON

CM O no

oo^

CM CM rj-

tJ- O On

i-i i-i CM

*Oo

»0 «<■

CM CM

NO — O

O H O

voi

to CM rf-

Wn5n
CM

00 NO o

*Ooo

O CM OO

O '-' >o

on t>s cm

CM

NO CM On

O ^ ON

to O to

m

tO CI f\J

to CM t*-

00 On
rH

CM O CM

T—

*f *H tO

NO O to

NO -4- Tt-
i— <

tO -H

1-*

N50*

O CM C*»

V9

NO O rj-

OO «0 on
CM

NO * CM
CM

to CM

HO00

© CM no

l>» «o «>
CM

CM On
CM

CM * CM

CM

—« O «o

On r-t

On vo o
to

NO WN
i—(

H H W

HHOO

OO w
tH to

tO NO NO

to

CM O 00

CM CM to

NO © CM

NO 00 CM
CM CM

'-' o
i~< to

On CM to
to

to ^ CM

^" OS

i-H CM no
CM —

^- to ^

CM

Approaches

Contacts

Crossings

Approaches

Contacts

Crossings

Approaches

Contacts

Crossings

Approaches

Contacts

Crossings

Approaches

Contacts

Crossings

day

days

days

days

o

CM

NO

THE THIRST

DRIVE

Tot5

1-4 NO tO

SO CM oo

NO CM oo

HMO

On O0 CM
to

CM OS CM
NO CM N
CM

N1N<t

CM CM 1-
th

-20th
inute

N&

so oo so

tO ^H VO

CM to OO
CM CM

^ CM h

O SO CM
CM CM CM

2 e

1-H 1-H

On O no

H H NO

O On no
CM i-H CM

Cn. NO On

to

no CM
to

-15th
inute

i-H N© 00
CM *H

NO CM 1-1

i-i CM CM

OO On
1-H i-H 1-H

CM i-H OO

to CM 1-H

NO NO OO

i-i ~ CM

CO Tt- t^-
1—1

NO*
CM CM 0O

NO O
1-H 1-H SO

O NO 00
CM

CM

6-10th
minute

O no
to CM CM

oo so — <
CM

hvOnO
CM CM CM

1-1 ^ NO

to CM O

1+ SO NO

to CM CM

1-H CM 1-H

to CM 00

NON9
CM i-H CM

NMO

CM CM CM
to CM CM

OO SO 1-H

to

-5th
inute

to CM i-i

no r*- to

SO T*- OO

CM to t*-

On OO CM
CM rf tJ-

CM OO
to to cm

° E

O 00 CM
CM *t

^ oo to
to VO

^ oo to
CM CM no

oo vn

00 o o

1-H tO

Activity

Approaches

Contacts

Crossings

Approaches

Contacts

Crossings

Approaches

Contacts

Crossings

Approaches

Contacts

Crossings

Approaches

Contacts

Crossings

Period
of water-
deprivation

o

1 day

2 days

4 days

6 days

THE THIRST DRIVE

113

the drive the amount of crossing during the first five minutes of the
test period, the thirst drive would stand out as clearly stronger
than the hunger drive. During the first five minutes the twenty
animals of the 1-day water-deprived group made a total of 153
crossings and those of the 2-day group a like number. The largest
number of crossings made in any five minutes by any hunger group
was 130, the record for the first five minutes in the testing of the
2-day starvation group.

The results for the two sexes have not been separately discussed
because, as the data show, the differences are negligible.

Summary

The tendency of the white rat to approach water as measured in
terms of the number of times it will cross an electrical obstruction,
as here employed, within a given period of time is at its slow point
when the animal is tested immediately after being removed from a
cage containing available water. Of the periods studied, that of
one day of water-deprivation results in the greatest demonstration
of this tendency. From this point on, the tendency diminishes con-
stantly until the death of the animal. Sex differences appear to be
negligible. This tendency appears to be at least as strong in the
white rat as that to approach food as previously measured by the
same method.

Bibliography

(1) Jenkins, T. K, L. H. Warner and C. J. Warden, 1926. Standard
apparatus for the study of animal motivation. J. Comp. Psychol., 6,
361-82.

(2) Kudo, T., 1921. Studies on the effects of thirst: I. Effects of thirst
on the weights of the various organs and systems of adult albino rats.
Amer. J. Anat., 28, 339-430.

(3) Kudo, T., 1921. Studies on the effects of thirst: II. Effects of thirst
upon the growth of the body and of the various organs in young al-
bino rats. J. Exper. Zool., 33, 435-461.

(4) Richter, C. P., 1922. A behavioristic study of the activity of the rat.
Comp. Psychol. Monog., 1, No. 2, 55 p.

(5) Richter, C. P., 1927. Animal behavior and internal drives. Quart.
Rev. Biol, 2, 307-43.

114

THE THIRST DRIVE

(6) Warden, C. J., and H. W. Nissen. An experimental analysis of the
obstruction method of measuring animal drives. Unpublished.

(7) Warner, L. H., 1927. A study of sex behavior in the white rat by
means of the obstruction method. Comp. Psychol. Monog., 4, No.. 22,
68 p.

(8) Warner, L. H., 1928. A study of hunger behavior in the white rat
by means of the obstruction method. J. Comp. Psychol., 8.

PART IV
THE SEX DRIVE

Part IV

THE SEX DRIVE
C. J. Warden

The sex drive offers a most fruitful field for intensive experimen-
tation. The work of the project on this drive comprises three major
studies — each a doctorate monograph. The first study, by Dr.
Warner, is concerned with the determination of norms for the sex
drive in male and female white rats which had been reared together
from birth in litter groups and thus had had a normal development
of sexual life. The second study, by Dr. Marion Jenkins, deals
with the effects of segregating the sexes at various stages of the
life cycle previous to the tests. The evidence is clear that early
segregation (pre-puberty) tends to disturb the "normal" sex
life and to introduce homosexual tendencies. The third study, by
Dr. Nissen, also deals with variations from the normal sex condi-
tions — gonadectomy, vasotomy and injections of prepared extracts
of the male and female gonadal hormones.

In each of these studies the general methods of control were the
same, so that direct intercomparisons can be made between the dif-
ferent special conditions introduced and the norms determined by
the first study of the series. The hunger drive was controlled pre-
cisely as in the work on the thirst drive, by allowing the animal
access to its usual diet (McCollum's powdered mixture) up to the
moment of the test. The thirst drive was controlled in a similar
manner by having the usual water supply in the cages continuously
during the pre-test period. The maternal drive was eliminated al-
together by using only non-pregnant females and those having no
litters at time of test. The exploratory drive was kept as constant
as possible in the manner described above (see Warden, The Hun-
ger Drive). Variations in the male sex drive were obtained by a
series of sex-deprivation intervals during which the animals were

117

118

THE SEX DRIVE

given no opportunity to copulate. This period began in the case of
each group immediately following a period of copulatory satiation
(2 hours) which reduced the sex drive to something like a zero
level, so that the sex-deprivation intervals represent periods of re-
covery from a common physiological zero point. Variations in the
female sex drive corresponded in general to the different stages of
the oestrous cycle, these stages being determined in all cases by
vaginal smears (quick and permanent) which were classified by
experts in this field who knew nothing whatsoever of the behavior
indices with which these stages were to be correlated.

The generous aid of Professor G. N. Papanicolaou of the Cornell
Medical College in connection with the several studies on the sex
drive deserves special mention. He very kindly consented to exam-
ine and classify the permanent smears in the first study of this
group, and offered many valuable suggestions concerning tech-
niques and procedures from time to time. He also gave the parts
of the study dealing with the various special physiological condi-
tions investigated a critical reading. Dr. E. F. Kinder was also
kind enough to demonstrate operative techniques in connection with
the work reported in Part IV, 3. Dr. Warner checked the classifi-
cation of permanent smears made by Jenkins in her study of seg-
regation, Part IV, 2.

The plan to carry out work on the effect of delayed incentive on
the normal male and female sex drive, corresponding to the study
of Hamilton (Part II, 2) on the hunger drive, had to be given up
in order to complete other lines of work.

1. A STUDY OF SEX BEHAVIOR IN THE WHITE
BAT BY MEANS OF THE OBSTRUCTION
METHOD 1

L. H. Warner
I. Introduction

The general purpose of this investigation was the study of the
sex behavior of the normal male and female albino rat. More spe-
cifically, the attempt was made to determine the effect upon such
behavior of varying the physiological conditions most intimately
related to it.

The female mammal is subject to periodic changes in the repro-
ductive tract which are, apparently, paralleled by definite changes
in the behavior of the animal toward a sex object. In the female
white rat ovulation is supposed to occur at rather regular intervals
of from four to six days. Observation of behavior has shown that
the animal " comes into heat," i.e., it will accept copulation, dur-
ing only a few hours at a given time, and that these periods of heat,
or oestrum, also recur rather regularly at intervals of from four
to six days. At other times the female will avoid mating by fleeing
or fighting.

Although the relationship between the changes in the reproduc-
tive tract and the animal's behavior has been noted it has not been
worked out in great detail. It is not known, for example, whether
the onset and termination of active sex behavior is abrupt or grad-
ual. One reason for the lack of exact knowledge has been the lack
of a method of observing any sex behavior other than that at the
two extremes: Active mating and active avoidance of mating. A
second reason has been the lack of an accurate method for the de-
termination of the oestrous condition of the animal.

1 Reprinted with modifications from Comparative Psychology Monographs,
1927, 4; No. 22, 68 p.

119

120

THE SEX DRIVE

No rhythm analogous to the oestrous cycle (the periodic changes
in the reproductive tract of the female) has been observed in the
male rat. In the case of the male the effect upon behavior of vari-
ations in the physiological condition due to various periods of seg-
regation involving sex deprivation has been studied. The data ob-
tained show the speed of recovery of the copulative ability of the
male after a period during which it has mated freely and may be
said to be " satiated" sexually. So far as the writer knows no data
bearing directly upon this aspect of male sex behavior have pre-
viously been reported.

II. Physiological Conditions Related to the Sex Drive

A. In the female white rat

The study of the sex drive in the female must take into consider-
ation the well known fact that sex activity in the female mammal
is not displayed, at least to the same extent, at all times but is con-
fined rather rigidly to certain recurring periods. The regularity
and period of this rhythm, the oestrous cycle, varies from species
to species.

Heape (7) classified mammals from this standpoint as either
monoestrous or polyoestrous. The former group is supposed to in-
clude those animals which experience a single oestrum separated
by comparatively long periods of inactivity, that is, but a single
oestrum in a single breeding season. Polyoestrous animals are those
experiencing more than one oestrum in each breeding season. The
latter group is subdivided according to the presence or absence of
the anoestrum in the reproductive cycle. Those animals experienc-
ing an anoestrum, or long period of sexual inactivity between
breeding seasons, usually breed during but one period of the year.
Those animals whose breeding seasons are not so separated breed
throughout the year.

Heape 's classification is not entirely satisfactory. The facts are
not so clearcut and simple as his terminology seems to indicate.
Not all mammals are found to fall into one group or another. The
dog, for example, does not strictly belong to any of the above
groups. However, the white rat, the animal used in the present

THE SEX DRIVE

121

investigation, may be said to belong to the group of polyoestrous
mammals experiencing no anoestrum. This animal, under
laboratory conditions, breeds throughout the year. Oestrum, in the
normal, unmated female rat, recurs every four or six days from
pubescence to menopause. The success of the present investigation
depended upon a knowledge of the nature of the rhythm in the ani-
mal used and upon a reliable method for determining the condition
of the animal with respect to the oestrous cycle.

Until recently, observations upon the periodic changes which
take place in the generative organs of non-pregnant female mam-
mals have been inexact and largely qualitative. In most of the
early papers which have touched upon the oestrous cycle the inter-
est in this phenomenon has been merely incidental. Among such
papers are those dealing primarily with the problems of ovulation
and of pregnancy. Studies of this kind were made for the most part
upon the smaller mammals commonly found in laboratories. The
fact that detection of the oestrous rhythm is much more difficult
in the case of such animals than in the case of the larger domestic
and wild animals is no doubt the main reason why exact knowledge
regarding this phenomenon has been for such a long time delayed.

Several methods have played a part in the detection of an oes-
trous rhythm in the smaller laboratory mammals and in the deter-
mination of the length of the complete cycle and of the various
phases of the cycle. The most important methods or types of ob-
servation will be taken up in turn, together with mention of rep-
resentative investigators who have made use of each. The results
obtained will be considered, especially in so far as they throw light
upon the reliability of the various methods.

1. Macroscopic observation of the living animal. In many mam-
mals the oestrous cycle is accompanied by changes in the amount
and character of the vaginal secretion so pronounced as to be easily
observable without instrumentation. There is often, also, such con-
gestion that the vulva and other parts become noticeably swollen.
Such signs are commonly observed among the larger animals but
they are not so evident among the animals used in the laboratory.
Nevertheless certain workers have found such signs convenient in
the absence of more exact ones. Lataste (14), for example, says

122

THE SEX DRIVE

that the approach of heat may be recognized in rodents by detec-
tion of swelling in the vaginal region :

Mais, a l'approche du rut . . . les bords de la vulve s'epaississent con-
siderablement. Le changement d'aspect que subit alors cet organe est tel-
lement caracteristique, qu'il me permettait de prevoir, a un jour pres le
moment ou une femelle, meme femelle meme jeune et precedemment un-
pubere, allait se trouver apte au rapprochement sexuell et a la feconda-
tion.

Lataste was not primarily interested in the oestrous cycle. Nev-
ertheless he apparently recognized its rhythmic nature although he
gives no data on its exact duration in any single mammal. He gives
it as his opinion that in rodents the cycle is usually about ten days
in length.

Sobota (29) noted a thin septum closing the vaginal opening
of the mouse between heat periods. This was ruptured when the
vaginal region became distended upon the approach of heat. His
estimate, however, that the cycle has a duration equal to that of
gestation (twenty to twenty-one days) has since been found to be
false as have most estimates based upon the general method of
gross observation.

Rubaschkin (23) noticed closure of the vaginal opening of the
guinea pig within three or four days after parturition when mat-
ing was not allowed during the oestrum which in most cases imme-
diately follows parturition. This closure prevented mating until
the next heat period, according to this writer. It might thus be
taken as a sign of the dioestrous interval, but it was evidently not
sufficiently clearcut to enable him to detect a typical length for the
dioestrum. He disagrees, however, with Bischoff's report that the
period averages about five weeks, since he noticed in certain cases
an open and flushed condition of the vaginal orifice from ten to
twelve days after parturition.

Another, and more recent worker who should be mentioned in
this connection, since he has found useful a method of detecting
the oestrous condition which involves no instrumentation, is Hart-
man (6). In the study of the reproductive cycle in the opossum
he has found that by palpation of the mammary glands it is pos-
sible to recognize the pseudopregnant growth of these organs which

THE SEX DRIVE

123

occurs prior to each ovulation period. In conjunction with this
method Hartman employed the method of microscopic examination
of the vaginal secretion in the living animal (a method to be taken
up in detail below). Hartman states that this latter method is the
more accurate for the determination of the oestrum, but that "at
other periods sometimes one, sometimes the other method has
proved the safer guide." The palpation method, depending as it
does on the unusual development of the mammary glands of the
opossum, would scarcely be applicable to the smaller rodents where
such development is lacking.

Probably the most skillful use of the general method of macro-
scopic observation has been at the hands of Stone (32, e). In the
study of the initial copulatory response of the young female rat
Stone was unwilling to use the vaginal smear method, although
admitting it to be the most reliable, since he feared that the manipu-
lation involved might affect the initial copulation. Out of twenty-
one females he selected as being receptive, twenty mated, a tribute
more to the skill of the experimenter than to the general reliability
of the method.

The general conclusion regarding the method of macroscopic
observation would seem to be that, while it possesses the advantage
of necessitating a minimum amount of interference with the animal,
it is not sufficiently reliable to be depended upon alone when deal-
ing with the smaller laboratory mammals. It should be considered
to be a supplementary method, helpful in the differentiation be-
tween the general oestrous and the general dioestrous periods only.
As such it has been used in the present investigation in selecting
from a large group of female rats those most likely to be in or ap-
proaching oestrum.

2. Post-operative and post-mortem examination of the generative
organs. It is this general method which has been used in a large
number of researches. In spite of this fact the results of these in-
vestigations would seem to indicate that the method is no more
reliable than the preceding one. Using it, certain workers failed
to detect the existence of a periodic cycle in the guinea pig, while
others have claimed that there is such a cycle and yet do not agree
among themselves as to its duration.

124

THE SEX DRIVE

This method furnishes data at only one point in the oestrous
cycle of an individual animal. It is thus necessary to determine
accurately at least one other point, previous to the operation or
to the sacrifice of the animal, in order to obtain data upon the
duration of the cycle. It has often been observed that the guinea
pig and the rat, among other mammals, experience oestrum and
ovulation within a few hours after parturition. This first post-
parturition oestrum has, therefore, been used as the starting point
by most investigators. Different animals were operated upon or
sacrificed at different intervals after this oestrum and the condition
of the ovaries and related parts noted. Since this method offers no
possibility of securing successive data upon a single animal and
thus of tracing the cycle continuously in the individual it is useless
in an attack upon many important problems. Successive cycles in
the same animal might vary slightly or considerably and there still
might be uniformity within the species as to the average duration
of the cycle. On the other hand, although successive cycles within
the individual might be remarkably uniform there might be notable
individual differences regarding the duration of the cycle. These
problems could scarcely be attacked by this method. Certain other
problems could be, but only by using large numbers of animals
since data upon but one cycle could be obtained from each. Prob-
lems of latter sort would include the effect upon the cycle of
changes in the environment and in the organic condition of the
animal.

One of the earliest investigators to use the method of post-mortem
examination was Bischoff (2). He was primarily concerned with
the problem of whether ovulation is spontaneous or dependent
upon copulation. He was inclined to hold to the former view
although he did not deny that copulation might have some effect
upon the time of ovulation. Although in connection with this work
he examined the reproductive tracts of a large number of guinea
pigs killed at various intervals after the first post-parturition oes-
trum he was unable to detect an oestrous rhythm. He gave it as his
opinion that the return of oestrum is dependent upon such factors
as age, diet, time of year and individual differences.

THE SEX DRIVE

125

Hensen (8), using the same animal, noted that if pregnancy did
not follow the oestrum which ordinarily occurs shortly after par-
turition, ovulation took place in the four cases reported, 17, 18, 35
and 37 days later. We assume that an unfruitful copulation took
place during the post-parturition oestrum. This would probably
have influenced the time of subsequent ovulation periods. Like
Bischoff, he concludes that the heat periods follow no sharp perio-
dicity.

While the conclusions of these writers were based largely upon
examination of the ovaries, the next group of workers based their
observations upon post-mortem examination of the vaginal epi-
thelium and secretion.

Morau (18) studied microscopically the vaginal condition of
mice killed at various intervals of from one to twenty-one days
after mating. In the cases of those females which had not become
pregnant he noted histological changes which appeared to indicate
a cycle of ten days duration. (Recent work has shown the cycle in
the mouse to be about half that long.)

Retterer (22) made similar observations on the guinea pig but did
not report estimates on the length of any cycle other than the first
post-parturition cycle. He usually found cornification of the super-
ficial layer of the vaginal epithelium on the fifteenth day following
parturition.

Lataste (14) states clearly his opinion that the modifications in
the secretion are independent of gestation :

. . . tout au moins chez les Rongeurs de la famille des Murides (rats et
souris, gerbilles, merions, etc.) les modifications epitheliales du vagin sont
directement liees a un tout autre phenomene que celui de la gestation.
. . . C'est done au rut, e'est-a-dire a l'ovuiation . . . et nullement a la ges-
tation, qu'est liee revolution de l'epithelium vaginal.

Lataste noted the cornified stage in the guinea pig from the fifteenth
day to the twentieth day following parturition (mating being
prevented). He carried his studies no farther than the twentieth
day.

Konigstein (13), although chiefly interested in the phenomena
accompanying gestation, noted histological changes in the vagina

126

THE SEX DRIVE

of the non-pregnant rat during the first several days following
parturition. These observations were largely incidental and were
not followed up by more extensive work.

Smith (28) examined the ovaries of mice killed at various in-
tervals after parturition, mating being prevented. He concluded
that the average duration of the oestrous cycle was seventeen and a
half days, but that the variability was great. Commenting upon
this work Allen (1, a) states that "estimates of previous oestrous
periods arrived at from histological comparisons of the corpora
lueta are not reliable since . . . many mice when isolated from
males do not ovulate spontaneously during oestrus."

This survey indicates that using the method of post-operative or
post-mortem examination of the generative apparatus several in-
vestigators failed utterly to recognize an oestrous rhythm and that
in no case was the duration of the rhythm determined with exacti-
tude. These facts indicate the difficulty of using this method.

A type of observation which might have led to the discovery of
the oestrous rhythm in the smaller rodents had it not been ante-
dated by a method to be mentioned later is the observation of the
spontaneous contraction of the uterine musculature. It is mentioned
under this heading since it involves sacrifice of the animal. Blair
(3) recorded the contraction of strips cut from the uteri of the rat
and suspended in oxygenated Locke's solution at thirty-eight de-
grees Centigrade. He discovered the contractions to be rhythmic
and noted that their period depended upon the oestrous condition
of the animal. During oestrum the contractions occurred at the
rate of twenty-four per hour, increasing to eighty per hour during
the dioestrum with intermediate rates during the transition stages.
Further work on the contraction of the uterine muscle, and also
of the Fallopian tube has been reported by Keye (12) and by Seck-
inger (24).

3. Observation of mating. Since mating occurs in most animals
only during oestrum the simplest method of studying the cycle
would seem, at first sight, to be the analysis of data on the periods
of mating. Unfortunately this method is rendered almost useless
because of the frequent occurrence of pregnancies in any normal
animals which are permitted to mate freely.

THE SEX DRIVE

127

Bischoff, using the guinea pig, separated females from males dur-
ing the first post-parturition oestrum. He then restored the males
and noted the time of the first subsequent matings. These occurred
at 40, 50, 51 and 51 days after parturition. Bischoff doubted the
existence of periodicity. There seems no reason to suppose that this
method would not have given at least an approximation of the
length of the cycle. A much larger number of animals would have
to be used and, since certain of the animals might first mate during
the second post-parturition oestrum, some during the third, etc.,
it would be necessary to determine, if possible, the approximate
largest common divisor for the intervals obtained. Supposedly
this would represent the normal duration of an oestrous cycle.

A method resembling this to some extent was used by Danforth
in the study of the mouse. This is reported by Allen (1, a) as fol-
lows:

Danforth in unpublished observations in 1914-1915 attempted to check
the length of the cycle by tabulating the number of days between two
successive litters and computing the greatest common divisor of these
intervals. In his records of sixty-six animals, twenty-three were excluded
because of the possible complication due to the mother lactating while
pregnant. The greatest common divisor of the modes of his curve was
found to be four to six days.

Since the greatest objection to the method of the observation of
mating has been the complication of pregnancies there have been
attempts to avoid such complication. Two methods of avoiding
pregnancy have been used: (1) by operation upon the female,
and (2) by the use of males which have been rendered sterile. In
addition to the usual post-mortem examination of the generative
organs Loeb (15, a) has used observation of the mating reaction of
female guinea pigs upon which double ligation of the tubes had
been performed. Although this method may possess certain advan-
tages over others it possesses the decided disadvantage of using
other than strictly normal animals. Loeb concludes that the cycle
in the guinea pig varies from thirteen and a half to nineteen days.
This range of five and a half days is so much greater than that
found by Stockard and Papanicolaou in an investigation to be re-
ported below that it may reasonably be supposed that the extreme
cases represent other than normal variations from the average.

128

THE SEX DRIVE

The second method of preventing conception, and one which
does not involve operation upon the female, was used by Long and
Evans (16). In this case the mating of vasectomized males with
normal females was observed. Long and Evans do not report in
full this attempt to determine the normal oestrous cycle since it
proved unsuccessful. Although the complication of pregnancy was
avoided a condition which has been termed pseudopregnancy was
found to follow mating with vasectomized males. The oestrous
rhythm was disturbed, successive oestra being usually delayed.
That this resulted from the mechanical stimulation involved in the
mating act seemed indicated by the fact that a similar condition
could be produced by stimulation of the cervical region by means
of a spatula or glass rod. Wang's observations agree with those of
Long and Evans on this point.

In conclusion it is seen that the disturbing influence of both fruit-
ful and unfruitful copulation upon the oestrous cycle tends to in-
validate the general method of observation of mating.

4. Observation of behavior other than that of mating. As long
ago as 1883, Rein (21), working on the guinea pig, claimed that the
heat periods could be recognized with great probability, though not
with absolute accuracy, by observation of the animal's behavior
even in the absence of the male. He describes running and jump-
ing and the occurrence of acts resembling those of normal copula-
tion. Nevertheless he states that he failed to notice a periodicity
in the recurrence of heat.

More recently, Stone (32, a, e) has described in considerable de-
tail the behavior of the female rat during oestrum but he does not
claim that such observation was sufficient to permit positive deter-
mination of the oestrus condition of the animal.

In the limited experience of the writer such observation has
been found sufficient in the case of many but not all rats to permit
differentiation of the general oestrous period from the general
resting period. But there seem to be decided individual differ-
ences with respect to the prominence of such overt manifestation
of the physiological condition. Certain animals fail to display the
behavior characteristics of heat at any time; an occasional animal
has been found which displayed such characteristics to some ex-

THE SEX DRIVE

129

tent at all times, not confining them to any one part of the cycle.

The type of behavior observation (other than that of mating)
which reveals most clerly the oestrous rhythm is the measurement
of the animal's voluntary or spontaneous activity. The activity
wheel has been in use for many years, but only recently have
studies been made which relate a rhythmic variation in activity to
the oestrous cycle. Wang (34) found that normal mature female
rats showed pronounced rhythms of from four to six days duration.
This rhythm was absent in males, and in females which were preg-
nant, pseudop regnant (following artificial cervical stimulation),
lactating, ovariectomized or sexually immature. Using the method
of the examination of the vaginal secretion in the living animal
(to be considered below) he found that the maximum activity oc-
curred during the stage of cornified cells, i.e., the stage during
which the animal will normally mate.

Slonaker (27, a) found an even more pronounced activity
rhythm than did Wang and found that as a rat approached meno-
pause the activity rhythm gradually disappeared. In general the
results of the two investigators agree. A more detailed analysis of
the activity at the various stages of the cycle has recently been
made by Slonaker (27, b, c). This writer supports the activity
method as the most satisfactory means of determining the oestrous
rhythm in the rat.

Our experience seems to indicate that the making of the vaginal smears,
due to the necessary handling, etc., of the animal, tends to disturb the
regularity of oestruation. We feel, therefore, that the normal sequence is
more accurately shown by activity curves than by any other method ad-
vanced as it leaves the animal undisturbed by unavoidably varying con-
ditions.

A very real objection to the activity method is that its use
involves a separate piece of apparatus for each animal. Now that
animals are used in larger numbers than in the past this would
often lead to insurmountable practical difficulties. The first step
toward a determination of the relative merits of the activity and
the vaginal smear methods would be the careful comparison of
results obtained from the two methods upon the same or similar
animals. Since the evidence thus far obtained reveals very little

130

THE SEX DRIVE

difference between the results obtained by the two methods there
seems no reason why the choice between them should not be
governed by their relative convenience and economy.

5. Histological examination of the vaginal secretion in the living
animal. As we have seen above the existence, and to some extent,
the histological character of the vaginal flow has long been recog-
nized. Sobotta (29) noted the secretion in the mouse during heat.
Heape (7) makes this general statement regarding the phenomenon:

Following the swelling and congestion of the external generative or-
gans, there is, in most animals, a discharge from the generative canal. The
discharge may consist merely of mucus from the uterine glands and from
the glands of the cervex and from those in the neighborhood of the vaginal
orifice, of the products derived from the breaking down of epithelial tissue
and of fragments of small masses of pavement epithelium from the va-
gina; such a discharge is usually to be seen in the rat and mole. In addi-
tion, fragments or small masses of columnar uterine epithelium may be
observed in various animals. Again, to the above, blood may be added for
a large number of animals, some of which rarely, some frequently, and
some always suffer from loss of blood. While, finally, more or less com-
pact masses of uterine stroma tissue are included in the discharge of the
primates and some of the lower mammals.

In only a few cases has the vaginal mucosa and epithelium been
made a chief object of study. Examples of such cases are the papers
of Morau and of Retterer, already cited. In no case was there made
a systematic study of the secretion in the living animal until the
work of Stockard and Papanicolaou (31, a, b). These investigators,
stimulated by the need of a method for determining with exactitude
the time of ovulation in the guinea pig, developed a method which
has proven to be superior to the others we have mentioned. To
quote from their earlier article :

The observations were made by using a small nasal speculum which was
introduced into the vagina and the arms opened apart by means of the
thumb screw. The speculum permits an examination of the entire surface
of the vaginal canal. In this way the vaginae of a number of females
have been examined daily and smears made from the substance that hap-
pened to be present in the lumen.

Using this method a definite oestrous period of about twenty-four
hours' duration was found to recur with striking regularity at
intervals of fifteen or sixteen days.

THE SEX DRIVE

131

According to the appearance and consistency of the secretion
the following stages of the general oestrous period were distin-
guished :

Stage IV they considered to be sometimes lacking and perhaps
not typical. The above stages compose the general oestrous period.
During the early part of the dioestrous interval the secretion is
very scant while during the second week following oestrum a slight
mucuous discharge is noticed which increases as the new oestrum
approaches.

Microscopic examination of the smears (stained with haemo-
toxylin and eosin) made during the various stages show the follow-
ing types of cells to be present. During stages I to II squamous
epithelial cells predominate, with "a certain number of elongate,
cornified cells without nuclei" during the latter part of the first
and the early part of the second stage. Stage III is marked by the
appearance of leucocytes which seem to have a destructive influ-
ence upon the second stage cells, often entering such cells during
the process of dissolving them. During stage IV there are found
red blood corpuscles in addition to the cells of stage III. During
the dioestrous interval the typical cells found in the scant secretion
are leucocytes.

These findings were based upon the daily examination of twenty-
six sexually mature virgin females, from one to seven cycles being
observed in each. In all, sixty-seven complete cycles were observed.
The average length of the cycle in the normal animal was found to
be 15.65 days. The work was carried on throughout the year. The
seasonal change was found to have very little effect on the length
of the cycle, probably because the animals were kept in a room
heated rather uniformly to seventy degrees Fahrenheit during the
winter months. The individual differences in the length of the cycle
were also very slight, the average length for a cycle ranging from
fifteen to sixteen and a half days. In no case was a cycle of less
than fifteen or more then seventeen days reported in a normal

I. Mucous secretion

II. Cheese-like secretion

III. Serous

IV. Bloody discharge

6-12 hours or more
2-4 hours
4-6 hours

f

132

THE SEX DRIVE

female. In their second paper Stockard and Papanicolaou state
that in older females the cycle is occasionally as long as eighteen
days, but that although they had records of observations upon
several hundred cycles they had yet to find a cycle of less than
fifteen or more than eighteen days. In this paper they also report
that careful observations lead them to believe that copulation takes
place only during stage I. The secretion at this time is a clear
foamy mucous and contains no leucocytes. They state that 1 'both
the nature of the vaginal fluid at this time and the absence of
leucocytes contribute to the success of copulation and fertilization. ' '

Using the method introduced by Stockard and Papanicolaou
Long and Evans (16) have made an extended investigation of the
oestrous cycle and related phenomena in the rat. Records were
made of the cycles of one hundred sexually mature virgin females.
In all, the lengths of 1999 cycles were determined. While the
average of these cycles was 5.4 days the distribution of the cases
was such that Long and Evans feel justified in designating 4.6
days as the average of those cycles which may be considered normal.
This is the average of those cycles (82 percent of the total), with a
duration of from 4.5 to 6 days.

Changes in the cell content of the vaginal secretion as detected
microscopically permits the subdivision of the cycle into five stages,
according to Long and Evans. Table 1 lists these stages together
with the various characteristics of each. It was compiled merely
for my own convenience and is based partly upon the observations
of Long and Evans and partly upon my own.

It will be seen that the changes correspond in a general way to
those found in the guinea pig by Stockard and Papanicolaou.
Long and Evans devote several pages of their monograph to a
comparison of the cycle in the rat and in the guinea pig, noting
that the chief differences lie in the length of the dioestrous interval
(which is much longer in the guinea pig) and in the condition of
the lumen during the period of receptivity (dry in the rat; charac-
terized by abundant mucous in the guinea pig). An attempt to
correlate in more detail Stockard and Papanicolaou's work on the
guinea pig with Long and Evan's work on the rat has been made
by Selle (25).

uouviiaa

Will not
mate un-
less at end

Actively
solicits
mating;
excitement

Usually
will not
mate

Will not
mate

Will not
mate

CELL CONTENT

Cornified

None

Thin, trans-
parent
non-nu-
cleated,

. scale-like,
single

Same, but
in sheets

A few still
remain

None

Epithelial

Small, round
nucleated;
very uni-
form, some
times small
sheets

None

None

Variable
number
may appear

t
a
i-
i-

a

ular shaped
always
si ngle
more nu-
merous in
later stages

Leucocytes

None

None

None

Quantities
of poly-
morpho-
nuclear
cells
appear

•

Z
t-
&

nate es-
pecially
in early
stages

QUANTITY
OF

SECRETION

Small drop

Scant

Profuse

Usually
rather less
profuse

Variable

CHARACTER
OF

SECRETION

Translucent;
jelly-like

Opaque
whitish
granular

Same, but
more dry
and cheesy

More fluid;
creamy

Gruel-like

DISLODGING
OF

SECRETION

Difficult,
usually
must be
smeared

Usually
may be
tapped off

Same as
above

As above

As above

INSERTION
OF SPATULA

With ease

Often some
resistance

Less re-
sistance

With ease

With ease

APPEARANCE
OF MUCOSA

Dry; less
transpar-
ent; some
turgescence

Very dry;
opaque;
maximum
turges-
cence

As above

Less turges-
cence

Clear, moist
gJistening,
trans-
lucent,
pinkish

GROUPING
USED IN THIS
WORK

Early con-
gestive

Cornified

Late corni-
fied

Post-ovula-
tive

Recupera-

ative.
Early inac-
tive, Inac-
tive, and
Late inac-
tive

Nouvana
aivivixoHddv

hrs.
9-21

27-51

3-15

42-75

STAGE (AS
NUMBERED

BY LONG
AND EVANS )

Pro-oestrum

'Oestrum

Metoestrum

Dioestrum

n.
in.

*

134

TEE SEX DRIVE

Long and Evans' technique in obtaining and identifying the
sample of the vaginal secretion differs somewhat from that used by
Stockard and Papanicolaou. A narrow spatula is inserted into the
vagina and withdrawn with a small drop of the secretion adhering.
This is dislodged in a small drop of physiological saline solution
and is immediately examined under eighty diameters. Under these
conditions samplings from individuals in different stages present
sufficiently different pictures to permit quick identification (at least
in most cases) .

Allen (1, a) has applied the method of vaginal smear examina-
tion to the study of the oestrous cycle of a third laboratory rodent,
the mouse. This investigator finds that the histological changes in
the vaginal secretion during the cycle are not markedly dissimilar
from those previously observed in the guinea pig and rat. He finds
for the length of the oestrous cycle a mode of four and a half days
and an average of from four to six days. Data on 563 complete
cycles were taken.

This review of the more important works upon the oestrous cycle
serves to emphasize the significance of the method of examination
of the vaginal secretion from the living animal. Prior to the intro-
duction of this method there was disagreement among investigators
concerning even a fact of such primary importance as the duration
of the normal cycle in the various laboratory animals. Since the
introduction of this method not only has the length of the normal
cycle been determined rather exactly in several mammals but the
cycle has been subdivided according to the histological picture.
Furthermore, the means of determining the oestrous condition of
the animal has led to the study of the relation of this phenomenon
to various other internal and external conditions, with the promise
that work of this kind will be widely extended in the future.
Among such conditions may be mentioned normal pregnancy, pseu-
dopregnancy, lactation, temperature, humidity, diet, hormones and
internal secretion.

This method is of value to the animal psychologist as well as to
the physiologist. It seems inevitable that behavior data of any sort
taken on the mature female animal are affected by the oestrous
cycle. The dominating influence of the oestrous condition upon

THE SEX DRIVE

135

activity has been demonstrated by Wang. It would not be surpris-
ing to find that learning, and perhaps discrimination, are also
affected by this phenomenon.

It is obvious that without the use of this method the present study
of the sex drive in the female would have been carried out with
much more difficulty and with less promise of significant results.

B. In the male white rat

If there exists a rhythmic fluctuation of the sex activity in the
male analogous to the rhythm in the female it should be recognized
and controlled in a study of this kind.

Heape (7) made the following classification of males with respect
to their mating periods:

1. Those which rut

2. Those which are sexually capable but which do not rut throughout

the year either because :

(1) The females of the species have no anoestrum (prolonged

period during which mating does not take place), or
because :

(2) The anoestra of the females are asynchronous

This classification leads one to assume that mating habits in the
male are determined by the type of oestrous cycle in the female
rather than by some rhythmic change within the male. If this is
the case, then males which normally rut but once a year would
mate at any time during the year if it were possible to bring the
female in heat at other than the one period. There is apparently no
experimental work bearing directly on this problem. Animals com-
monly brought under experimental observation belong to the second
class, i.e., they do not rut. The guinea pig and the rat belong to
II A, the dog belongs to II B. Thus we cannot say that the existence
of a sexual cycle in the male has been disproved but merely that
none has been observed. By sexual cycle we mean here a periodic
fluctuation in the intensity of the sex drive and the capacity of the
male to copulate, this fluctuation being caused by changes within
the male and independent of the female.

It has been noted that captivity tends to modify the rutting
season. This may be due to the effect of the more uniform environ-

136

THE SEX DRIVE

ment upon the cycle in the female rather than to its effect upon the
male directly.

Sexual behavior in the male albino rat has been the subject of
careful investigation by Stone (32, a). His particular interest has
been in the following points:

1. The pattern of the sex response

2. The development of this pattern in the young male

3. The age at the first mating

4. The senses involved in the arousal of the act

5. The number of copulations during a given period

Stone reports nothing which seems to indicate a periodic fluctua-
tion of the sex drive in the male.

Using the activity method neither Wang nor Slonaker observed
periodic variations in spontaneous activity in the male such as are
characteristic in the normal, mature female. The males show much
less activity than do the females, and the slight irregularities which
appear do not seem to be rhythmic. A recent article of Slonaker 's
(27, c) presents data hinting at, but not definitely establishing a
fluctuation in the amount of activity which follows a two to four
month cycle. It is very unlikely that such a rhythm is related to
the reproductive process in the male since it was found in both
sexes and since, in the male, it was found to continue on into the
period of senility.

It seems, then, that we are justified in assuming that rhythmic,
internally determined fluctuation in the strength of the sex drive
of the white rat in captivity either does not exist or is so slight
that it would not materially affect the results of the present
investigation.

III. Method and Procedure
A. Age and care of animals

Since the purpose of this experiment was the study of the sex
behavior in the normal male and female white rat, care was taken
to use animals which had from birth been raised under normal
sexual conditions. Such animals were raised especially for us by
the Wistar Institute of Anatomy, Philadelphia. The animals (of
the strain known by them as the ' ' Experimental Colony Strain")

THE SEX DRIVE

137

were raised in groups of 5 to 10, the sexes being unsegregated,
there being an approximately equal number of males and females
in each group. As a rule 2 litters were thrown together at the time
of weaning and raised together until they reached the age of 150
days at which time they were shipped to us. Upon their receipt the
sexes were segregated. Segregation, or sex deprivation intervals
involved merely the elimination of all tactual contact between an
animal and one of the opposite sex. Olfaction was not controlled,
it being impossible to obtain separate rooms for the two sexes. The
animals were not used until about 5 weeks after we had received
them and segregated the sexes, or until they had reached the age of
about 185 days. No animal under 175 or over 196 days of age was
used (except the males of the 28-day group which averaged 28 days
older than this). The segregation period prior to the experimenta-
tion varied in length from 33 to 39 days. This 5-week period
permitted the animals to become accommodated to the laboratory
conditions. It also gave them time for all pregnant females to come
to term and for the new litters to reach an age at which they
could be weaned. Assuming a female to have become pregnant
just prior to the segregation of the sexes, her litter would be born
at the end of 3 weeks and in 2 weeks more could be weaned. (It is
better however, not to wean until a litter is 3 weeks old.) These
litters were not used in the present experiment but were added to
the laboratory colony for future work.

The Wistar rats were used as being the best and most uniform
animals available. It was felt that the variables studied in this
experiment might have such slight effects that every care should
be taken to rule out sources of error which might cloud the results.
The individual variation in the animals was slight. The range in
weights was small and in general appearance the animals were
highly uniform. There seems no reason to suppose that these ani-
mals were anything but normal from the standpoint of sex experi-
ence. They were sexually mature and had copulated freely as was
evidenced by the large number of litters produced. Being raised
in groups containing never less than 2 animals of each sex all ani-
mals, it may be assumed, had had normal sex experience from ma-
turity.

138

THE SEX DRIVE

The writer feels that the most serious criticism that is to be
raised against this work concerns the 35-day period of segregation
of the sexes just prior to the experiment. In defense of this pro-
cedure the following points can be made :

1. The animals had matured approximately 3 months prior to
segregation and had all, doubtless, mated many times. The normal
sex drive was thus given full opportunity to develop.

2. Only rarely was behavior noted among the segregated groups
which might have been interpreted as homosexual.

3. For practical reasons we were forced to segregate the females
in order that they be non-pregnant when used. In order to keep
the conditions constant for the two sexes it was necessary to segre-
gate the males for the same interval. Several means of avoiding
this procedure were considered :

A. A colony of vasectomized males could have been established.
Such animals living with the females during this period would
mate without the ensuing complication of pregnancy. On the basis
of the work of Long and Evans, however, it is assumed that such
mating, even though not resulting in pregnancies would have pro-
duced irregularities in the oestrous cycle.

B. The animals could have been kept singly, in individual cages
during this period. Although this procedure would eliminate the
possibility of the development of homosexual tendencies it is not
improbable that it would have had at least as great an influence
upon the results since it would affect the animals' responses to
other animals of the same species regardless of sex.

C. Probably the ideal plan would be to leave the sexes unsegre-
gated throughout, using only those females which happened to be
non-pregnant when they reached the age of approximately 185
days. The first difficulty that this procedure would present would
be that of detecting pregnancies. It is very likely that a number of
animals would be used which would later be found to have been,
at the time, in an early stage of pregnancy. Thus a large amount of
data would have to be discarded so far as the present investigation
is concerned. Furthermore, the percentage of non-pregnant fe-
males at any given time would be low since rats customarily mate
shortly after parturition. In order, therefore, to use a large pro-

TEE SEX DBIVE

139

portion of the animals most of them would have to be held over
until one or probably several litters had been born before they
could be used in a non-pregnant state. This would introduce a wide
age difference, a factor which we desired to keep uniform since it
may be related to sex behavior. Of course, if it were possible to
secure animals far in excess of the numbers actually needed with
the intention of discarding the bulk of them, data could be then
secured on a sufficient number of unsegregated, non-pregnant fe-
males. Existing facilities rendered this method impractical. If this
plan were followed it would be interesting to note whether the
drive during the first oestrum following parturition is noticeably
stronger than that during subsequent oestra. Observations of the
animal 's behavior have suggested this possibility.

4. Unpublished data (from this laboratory) of Marion P. Jenkins
seem to indicate that the influence of the five weeks' segregation
period upon the results is negligible. This investigator succeeded
in securing a few unsegregated, non-pregnant females which were
about 185 days of age. These she tested exactly as was done in the
present investigation, with results almost identical to those reported
here.

Sufficient and suitable food had been provided the animals
throughout their development. Although the exact content of the
diet given them at he Wistar Institute cannot be reported long and
successful experience of this organization with the breeding of rats
is a guarantee of its adequacy. During the five weeks spent in our
laboratory the diet consisted of "McCollum's mixture" plus
weekly rations of greens. Although McCollum's standard diet is
described in several available sources it may be inserted here for the
convenience of the reader :

Per cent

Whole wheat flour 70.0

Whole milk powder 16.6

Casein 11.0

Sodium chloride 1.4

Calcium carbonate 1.0

This was mixed in an electric mixer for ten minutes or more and
fed dry. There was at all times a supply available, i.e., food con-
tainers were refilled before they were entirely emptied. Water

140

THE SEX DRIVE

was supplied by means of inverted bottles and nozzles. Fouling
of the water supply was impossible. For description of this device
and for other valuable information concerning the care of rats see
G-reenman and Duhring (5).

Our purpose in keeping more than a sufficient amount of food
continually before the animals was to keep uniform and constant
the hunger drive, or food-seeking activity, at the time when the
animal was used in the experiment. While a rat, even under these
circumstances does not eat as uniformly throughout the twenty-
four hours as does a guinea pig, for example, still we found that
eating was much more generally indulged in throughout the day
and night than is the case when rats are fed but a single day's ra-
tion at a time and are fed at a constant time, day after day, as is
usually done in animal laboratories. We may assume that the rats
were not in an active state of hunger when experimented upon
since they were taken directly from a cage provided with food.

The animals were housed in wire mesh cages of about 420 square
inches floor area and 5,900 cubic inches capacity. There were from
five to ten animals in a cage. Bedding of coarse saw dust was pro-
vided.

B. The apparatus

The Columbia Obstruction apparatus (11) was employed in this
study. As noted, the box for the control of the animals consisted
of three compartments which may be called the entrance compart-
ment, the obstruction compartment and the incentive compartment.
A small section of the latter was separated off by an automatic
door. The inside dimensions of the various sections were as fol-
lows :

WIDTH

LENGTH

HEIGHT

inches
10
4

inches

inches
10
4

Entrance compartment--.
Obstruction compartment.
Incentive compartment:

10
10

Main section

10
4

10
4

10
4

Small section.

The entire floor of the obstruction compartment consisted of an
electric grid. The grid which was described in the article cited

THE SEX DRIVE

141

above was found to possess one serious disadvantage. A drop or
two of water could easily short circuit it, interfering with the elec-
trical stimulation value of the grid. A micturition response of the
animal occasionally had this effect, necessitating interruption and
delay. To avoid such possible irregularities a second grid was
built. Unlike the first, which consisted of copper wires wound
around bakelite, this grid was formed of copper strips, one-eighth
inch thick and three-eighths inch in width, running diagonally
across the floor of the compartment. The strips were separated
from each other by a uniform gap of one-eighth inch. This gap
was of such width that it was impossible for a drop to hang sus-
pended between two strips. The strips were so wired that those
connected with the positive and the negative poles alternated. The
mounting and wiring of the strips was from beneath in such a way
as to secure insulation and protection from dirt and moisture.
This grid proved to be perfectly satisfactory. Apparently the ani-
mals never failed to make simultaneous contact with strips from
opposite poles and thus to receive the shock. They experienced no
difficulty in walking across the grid, their habit of walking with
outstretched toes preventing their catching their feet in the nar-
row gaps.

Except for the grid, the celluloid door and the glass covers, the
entire apparatus was painted flat black. It was set up on a table
in an improvised dark room. Since the experiments were all done at
night we were not faced with the difficulty of eliminating day-
light. The only source of light was an illumination hood 11 inches
above the center of the apparatus. This cast uniform light in all
parts of the apparatus. Since the illuminating surface extended
without break over the whole length of the apparatus the partitions
and the doors of the apparatus cast no shadows. The hood was so
adjusted that light was thrown only into the box and not upon the
other objects in the room nor upon the experimenter. Since the ani-
mal stood in the light and the experimenter in the shadow it did not
seem necessary to use a screen and peephole as is usually done in
well controlled animal experimentation.

Only one degree of shock was used in the present investigation.
Alternating current of 60 cycles was used. The terminal pressure

142

THE SEX DRIVE

was 475 volts. The external resistance in the circuit totalled 10,-
100,000 ohms and the current was thus 0.047 milliamperes.

The determination of a suitable degree of shock was no small
problem. It was our purpose to adopt a degree which could be used
not only in the study of the sex drive but also in the study of other
drives, such as the hunger drive. Dr. Holden 's study of the hun-
ger drive in the white rat by means of the obstruction method (9)
gives clear indication of the effect Upon the drive of variation in
the degree of shock. Of the three degrees she used the lowest gave
the best results. Although the specific results of Dr. Holden 's work
could not be directly applied to the present study because of im-
portant differences in apparatus and technique her experience was
helpful in guiding the preliminary work done for the purpose of
determining the degree of shock, to be used. In this preliminary
work data on both the hunger and sex drives were taken using va-
rious degrees of shock. The shock finally settled upon as the most
satisfactory is less than the lowest shock used by Dr. Holden. The
present work indicates that it is suitable for the study of the sex
drive under the conditions here employed and further work from
the Columbia laboratory, soon to be reported, proves its suitability
in the study of the hunger drive. For Figures 1 and 2 of this mon-
ograph showing appartus, see Part I, 1, of this volume.

C. Detailed procedure

The male. Variation in the physiological condition of the male
was obtained by varying the interval separating the test and the
last previous mating. In every case the male was placed in a cage
with a female in heat for two hours and the mating behavior ob-
served. Distribution of the number of matings occurring during
this period is shown in Table 2. The accuracy of these data cannot
be vouched for since the experimenter was usually engaged in stain-
ing vaginal smears and in other laboratory duties while one or more
such mating periods were in progress. It is unlikely, however, that
the data seriously misrepresents the facts. The chief purpose of
keeping a record of the mating was to determine whether the male
in question behaved normally toward the female and to ascertain
whether full opportunity for "satiation" was afforded.

THE SEX DRIVE

143

After this two-hour period the male was placed in a cage with
other males for the specified deprivation interval, at the end of
which it was tested in the obstruction apparatus. Animals in the
zero group were transferred directly from the cage with the recep-
tive female into the apparatus and tested. The deprivation in-
tervals studied were 0, 6 and 12 hours, 1, 4, 7, and 28 days. There
were twenty animals in each of these groups. There was also tested

TABLE 2

Distribution table covering number of mati/ngs during the two-hour period
prior to segregation of males

NUMBER OF COPULATIONS

FREQUENCY

6-10

7

11-15

5

16-20

11

21-25

10

26-30

13

31-35

16

36-40

17

41-45

18

46-50

16

51-55

19

56-60

11

61-65

6

66-70

4

71-75

7

76-80

3

81-85

2

160

Average 40.6.

Standard deviation 18.0.

a control group of twenty animals. The procedure in the case of
this group was exactly like that of the one-day group except that
no stimulus animal or other incentive was used. The stimulus com-
partment was empty. In the case of all the other groups the stim-
ulus was a female rat in oestrum, i.e., in the cornified stage. The
stimulus animal was placed in the smaller section of the incentive
compartment and the automatic door shut, preventing her from
crossing the obstruction to the male. The male to be tested was
placed in the entrance compartment, the door leading into the ob-
struction compartment being closed. The glass slides above all
compartments were closed except when an animal was being intro-
duced or removed. The male being restrained in the entrance
compartment and the female in heat being restrained in the smaller
section of the stimulus compartment the experiment proper could

144

THE SEX DRIVE

begin. This consisted of a short preliminary training period im-
mediately followed by a twenty-minute test period. The purposes of
the perliminary training were

(1) To accommodate the animals to the apparatus

(2) To permit the male to associate the stimulus compartment with

the female in heat

(3) To permit the male to associate shock with the obstruction com-

partment, and with the obstruction compartment only

A considerable amount of data were taken in an effort to deter-
mine what type of preliminary training best served these ends.
It was found that the final results were profoundly affected by the
nature of the preliminary training. The procedure finally adopted
is believed to fulfill the three purposes as adequately as this can
be done.

Ten seconds after a male was placed in the entrance compart-
ment the door leading into the obstruction compartment was
opened. There was at this time no current in the grid. The test
animal usually passed through the obstruction compartment and
into the stimulus compartment within the next quarter minute or
less. After he entered the stimulus compartment the door separat-
ing it from the obstruction compartment was quietly closed so that
he could not retrace his steps. Meanwhile the automatic door had
opened and the two animals were together in the stimulus com-
partment. At this moment the experimenter was always alert. The
male was to be removed before he mated and yet not until he had
had opportunity to identify the stimulus animal as a female. Since
the most usual response of the male was the nosing of the female,
and in particular of the genital region this behavior was taken as
the criterion. Not until the male had done this was it removed and
replaced in the entrance compartment. At the same time the stim-
ulus animal was replaced behind the automatic door, and this
closed. The animals and the doors were now in the positions they
had been before. This procedure was repeated, four times in all.
In each successive crossing the male tended to show less hesitation,
to cross more promptly to the female, and, once in the stimulus
compartment, to devote less time to exploration of the compartment
and more to attention to the female.

THE SEX DRIVE

145

It was not until the fifth time that the animal was allowed to
cross to the female that electric shock was introduced. In this fifth
crossing, the last of the preliminary series, the switch controlling
the current to the grid was closed after the animal had entered the
obstruction compartment. At the same instant the door was lowered
behind him, cutting off possible retreat into the entrance compart-
ment. The only way for the animal to end the electrical stimulation
was by continuing on into the stimulus compartment. Again, as be-
fore, he was left there until his interest in the female crystallized

#

sx

WT

DR

GR

DATE

//

1

2

3

4

5

6

7

8

9

10

T

A

T

C

14

A

11

12

13

15

16

17

18

19

20

T

T

C

Fig. 3. Reproduction of Data Card Used in the Present Investigation
Numerals represent the minute in the twenty -minute period during which the
activity took place. A stands for approaches; T, for contacts; C, for crossings.

into the specific behavior of nosing the genital region. He was then
replaced in the entrance compartment and the test period was im-
mediately begun.

The procedure in the test period was like that in the preliminary
period with the following exceptions:

1. The grid was electrified throughout the test period.

146

THE SEX DRIVE

2. The test period was twenty minutes in length, irrespective of
the behavior of the animals.

3. The behavior of the test animal (the male in this case) was
observed and recorded in terms of approaches, contacts and cross-
ing's together with the time during the test period when each such
act accurred. These data were recorded on a card such as shown in
figure 3.

The three types of behavior recorded are defined as follows t

An approach: Some part of the animal (actually, the tip of its nose)
crosses the threshold between the entrance and the obstruction compart-
ment, but instead of proceeding and thus coming into contact with the
grid, the animal withdraws.

A contact: Some part of the animal (its feet, occasionally its nose)
touches the grid, but instead of proceeding across the grid, the animal
retires into the entrance compartment again. The withdrawal is usually
more abrupt in this case since it is aroused by the electrical stimulation.

A crossing : The animal passes through the entrance compartment and
into the stimulus compartment, the door between the obstruction and the
stimulus compartment being dropped behind him.

4. The final difference between the preliminary and the test
period procedure lies in the criterion which determines that the
male is to be returned to the entrance. compartment after a crossing.
In the preliminary trials he was not returned until he had nosed
the genital region of the female even though this occassionally
involved a delay of several minutes during which the male explored
the stimulus compartment. During the test period such general
exploratory activity was not so often observed and the attention of
the male appeared fixed primarily upon the female. Nevertheless
the definite nosing response did not always occur promptly and the
criterion was broadened to include any specific contact between the
two animals so that one or more of the following acts was some-
times substituted for the nosing act on the part of the male ; nosing
of other than genital regions; nipping or biting, most often of the
back and neck regions. Such behavior on the part of the male was
usually accompanied by similar behavior on the part of the female.
In the case of a very few males who were extremely sluggish and,
in most cases, crossed but a few times such behavior on the part of
the females, only, was required since the males failed to make any

THE SEX DRIVE

147

appropriate response within a reasonable amount of time. In gen-
eral the test animal was returned after from five to fifteen seconds
in the stimulus compartment.

The female. The physiological phenomenon most intimately re-
lated to the female sex drive is, of course, the oestrous cycle. The
female is said to be in oestrum when it will accept the male. The
other periods of the cycle are defined as related to this period.
Mere observation of the behavior of a female in the presence of a
mature male will indicate only that it will or that it will not mate.
The obstruction method of measuring drives supplements such
rather meager data since it permits us to rank those females which
will mate in the order of the strength of the mating impulse. Just
as the vaginal smear method developed by Stockard and Papani-
colaou yields a much more exact and detailed picture of the physio-
logical changes during the oestrous cycle than was theretofore
obtainable so we feel that the obstruction method gives a more sen-
sitive index of the behavior of the animal during the cycle than can
be obtained by other behavior observation. We have used the vag-
inal smear method and the obstruction method in conjunction
enabling us to follow the rise and fall of the strength of the sex
drive during the cycle, thus correlating physiological condition and
behavior in a more detailed way than has been possible in the past.

In testing the female the procedure was exactly like that for the
male with the following exception :

1. The previous two-hour mating period and period of depriva-
tion were omitted, the females being transferred directly from the
living cage to the apparatus.

2. The stimulus animal was always a male (except in the con-
trol group, with which there was no stimulus animal used).

3. Vaginal smear records were taken for each female imme-
diately following the test period.

Females for testing were selected at random from the living cage
except that whenever a female was noticed, which, judging from
its behavior and from macroscopic observation (non-manipulative),
was in oestrum that animal was tested at once. If the animals had
been selected entirely at random the majority of them would have
been in the various phases of the dioestrum. Since we wished to

148

THE SEX DRIVE

test at least forty animals which were in heat, twenty of these to
form the control group, selection with this in mind was necessary.
In no case, however, was a vaginal smear taken prior to the testing
of an animal.

A vaginal smear record was taken immediately after the removal
of the female from the apparatus at the end of the test period.
These were stained with haemotoxylin and eosin and permanently
mounted in balsam. The classification of the females as to oestrous
condition, and thus their grouping, was wholly on the basis of these
histological records. After the slides had been examined and
classified they were taken to Dr. G. N Papanicolaou who repeated
this process (with no knowledge of the behavior data). Dr. Papan-
icolaou's grouping of the slides according to the stage of the cycle
represented checked the previous classification but went further
than it had in the way of subdividing the cycle. He found that he
could distinguish eight groups, representing as many phases of the
complete cycle. The behavior data was grouped accordingly as will
be seen by referring to the results. Discarding the rather loosely
defined terms such as Metoestrum, Postoestrum and so on, Dr.
Papanicolaou has suggested the descriptive terms which we have
used for the eight groups.

These microscopic records are open to examination at the Animal
Laboratory, Columbia University.

The chief point in which our own observations do not verify
those of Long and Evans concerns that stage of the cycle termed
by them stage IV. The smears which most nearly resemble this
stage are those placed by us in the post ovulative group. These do
not show the clearcut picture described by Long and Evans. Both
leucocytes and epithelial cells seem to have returned while a few
cornified cells still remain. This would appear to be a rather mixed
stage suggesting a gradual transition from the cornified stage to
the stage of epithelial cells and leucocytes. Thus it contrasts with
the more abrupt and clearcut onset of oestrum when the epithelial
cells give place rather suddenly for the cornified, if we may judge
from the fact that the two are not ordinarily found in the same
smear. As will be seen below the behavior data likewise show a sud-
den onset and a gradual cessation of the characteristics of oestrum.

THE SEX DRIVE

149

An important part of the technique, but one which baffles de-
scription, consists in the handling of the animals. The animals be-
ing tested must be lifted from the incentive compartment after
crossing and replaced in the entrance compartment without arous-
ing a fear response. Considerable preliminary work was required
not merely for the determination of a suitable degree of shock, ap-
propriate preliminary training, and so on but also to familiarize
the experimenter with the apparatus, procedure and method of
handling the animals. Careful observation of the animals during
the experiment proper led to the impression that they were not
seriously disturbed by the necessary handling. On the several oc-
casions when such a disturbance was noted the animal was dis-
carded. The chief points noted in the handling of animals when
using the obstruction method were the following :

1. Strict silence was observed at all times.

2. All movements of the experimenter were slow.

3. A rhythm of removing and replacing the animals was ac-
quired so that the process occupied very nearly the same amount of
time in every case.

4. An animal was never lifted in such a way that, being insuffi-
ciently supported in the hand it would reach out for other support
to which to cling. The animal never touched any object outside the
apparatus except the hand of the experimenter.

5. When released from the hand the animal was not dropped
into the entrance compartment but was lowered in the hand to the
floor of the compartment and there gently released.

6. Manually operated doors and slides were manipulated with
due regard for the position of the rat and in particular of its tail.

IV. Results

A. The male sex drive

Table 3 indicates the grouping of the one hundred and sixty male
rats tested. The distribution of approaches, contacts, and crossings
for the eight groups is given in Table 4. The medians, ranges, aver-
ages, standard deviations, and coefficient of variability are found
in Table 5. Table 6 shows the reliability of the differences between
the averages. The approaches, contacts, and crossings as distrib-

150 THE SEX DRIVE

TABLE 3

Showing male groups

NUMBER OF ANIMALS

PERIOD OF SEX DEPRIVATION

OBJECT IN INCENTIVE COMPARTMENT

20

24 hours

None (control group)

20

0

Female in heat

20

6 hours

Female in heat

20

12 hours

Female in heat

20

1 day

Female in heat

20

4 days

Female in heat

20

7 days

Female in heat

20

28 days

Female in heat

TABLE 4

Distribution table covering approaches, contacts and crossings for the various periods
of sex deprivation in the male rat

CON-
TROL*

0

6 HOURS

12

HOURS

1 DAY

4 DAYS

7 DAYS

28 DAYS

TEBVAL

>proache3 |

) ntacts

so

S)
C
'w

8

^proaches 1

>ntact8

■ossings 1

> proaches |

>n tacts

ossings

^proaches j

mtacts

ossi ngs

jproaches

mtacts

CO

c

'§?
2

pproaches |

) ntacts

ossings

)proache3

>ntacts

ossings

jproaches j

mtacts

B

a

8

Z

<

O

<

0

O

<

O

O

<

O

O

<;

O

<

0

O

<

O

Q

<

O

Q

0

5

8

2

8

15

6

3

2

3

5

4

1

4

6

7

5

1

5

3

2

6

6

2

1

9

7

4

6

4

2

2

1

1

2

3

6

6

3

6

3

3

2

5

4

2

2

2

3

2

2

3

2

2

3

3

3

4

5

6

2

3

1

4

1

3

4

4

3

1

3

1

2

1

3

3

1

2

1

2

5

4

2

1

4

2

3

3

2

1

4

2

1

4

1

1

2

1

2

4

3

1

3

2

1

5

4

1

1

5

5

2

3

1

1

1

1

1

6

1

3

2

2

1

1

2

1

2

7

1

1

1

1

8

1

1

1

1

9

1

2

2

2

1

1

10

2

1

1

1

11

1

1

3

3

2

1

2

1

12

1

1

3

1

2

1

13

2

1

2

1

1

14

1

3

2

3

15

1

3

4

3

2

16

1

3

4

2

17

1

1

1

1

18

1

1

2

3

19

1

1

20

2

1

21

1

1

22

1

1

1

23

24

1

1

* Control group, twenty-four hours ' sex deprivation ; no female in the incentive
compartment.

THE SEX DRIVE

151

TABLE 5

Showing results of the male groups

Control
0

6 hours
12 hours

1 day
4 days

7 days
28 days

APPROACHES

1
1

4

2.5
1.5
1.5
2.5
1

0-3

0-5

0-12

0-7

0-10

0-6

0-5

0-10

1.2

1.2

4.15

2.7

2.3

1.8

2.25

2.5

5*o

02

1.03

1.36

3.26

2.03

2.35

1.9

1.5

2.9

85.8
113.3

78.5

75.2
102.2
105.5

66.6

1
0

5

2.5

116.0jl._

0-4

0-3

0-11

0-6

0-6

0-4

0-9

0-6

1.05

0.35

4.5

2.65

1.5

1.85

3.1

2.0

CROSSINCS

e

TJ.2

ent of
bility

e

TJ-J

ent of
bility

Standa
devia

Coeffici
varia

Median

Range

Averag

Standa
devia

Coeffici
varia

1.18

112.3

3

0-7

2.95

1.8

61.0

0.7

200.0

2

0-13

3.6

4.08

113.5

2.77

61.5

9.5

0-18

8.1

5.2

64.2

1.98

74.7

12

0-20

12.2

4.8

39.3

1.5

100.0

15

4-20

13.45

4.03

29.9

1.56

84.3

14

0-22

12.6

6.24

49.5

2.5

80.6

14.5

0-24

12.25

7.07

57.7

1.97

98.5

12.5j0-22

10.55

7.1

66.4

TABLE 6

Showing reliability on the difference between the average crossings of the male groups

GROUPS

THE DIFFER-
ENCE

STANDARD
DEVIATION

OF THE
DIFFERENCE

DIFFERENCE
S.D. OF THE DIFF.

CHANCES IN
100 OF A
TRUE DIF-
FERENCE
GREATER
THANO

0 hours and 6 hours

4.5

1.47

3.06

100.0

6 hours and 12 hours

4.1

1.58

2.59

99.5

12 hours and 1 day

1.25

1.4

0.89

81.0

0.85

1.65

0.51

69.0

0.5

2.1

0.24

60.0

1.55

2.23

0.695

76.0

2.9

1.8

1.6

94.0

10.50

1.14

9.2

100.0

uted during the twenty-minute test period (in one-minute inter-
vals) is given in Table 7. In Table 8 these data are reduced to the
percentages of such activity occurring during successive five min-
utes of the twenty-minute test period. Figure 4 shows graphically
the effect of variation in the sex deprivation interval upon the ap-
proaches, contacts and crossings. Figure 5 shows graphically their
distribution during the test period in terms of the percentage fall-
ing within each successive five minutes.

We will first take up the male groups in order from the shortest

152

THE SEX DRIVE

1 g

ainNin hxOo

HOO

<Nh»0

CO CO h

o

1—1

HO"*

h CO CO
CO

co oo

axxixim Hxgx

MOM

OOiO

00 CO 00

COiO O

CO

iG

H T— < CO

CO h CO
CO

f h CO

CO

axn.viK Hxgj

OhO

h ON

co co co

O CO O

t>

H i-H CO

CO 00 h
CO

•HH O f

CO

(N r-t H

NOiO

00 rt< o

o

o

01

OOH

h CO o

CO

MOO

H

axiiNiiv hxqj

WOh

h h co

HLO00

CO

H H ^

CO lO h
CO

CO

axriNiK HXgx

CO CO f

HOM

CO CO H

O.f Tf<

00

01

o

CO O CO

CO CO
r— 1

h co co

axriNiPi hx^x

rH OH

f)H(M

CO CO CO

H H t>.

o

co

M

CO CO CO

CO f 05

axnMiK Hxgx

WHN

NhN

COCO CO

co oo n

CO

©

co

CO CO CO

CO CO H

CO O h

3XX1NIW hxjx

H O H

O CO

LO CO CI

co o h

CO

H H O

Of N

H CO H

axnMiiv hxxx

i-H Oh

coo-*

f co o

co h oo

CO

o

CO ^ H

1— 1

INNH

HTfN

axriNUM hxox

O h h

O O CO

00 <tf h

H

h CO iO

CO

o

HHN

CO CO "tf

H "HH CO

axfiNiiv HXg

H HK)

h 0*0

CO f if?

COM< O

CO

1 — '

(NON

CO h io

f CO h

3XJ1NIIV HX8

CO 00 00

O h f

■^cco »o

O H OS

MON

axuxiK Hi£

OCO CO

H

iO iO Ci

00

iO

CO

'O

co h co

h 00

h CO CO

3XHMIK HXQ

OO"*

HOP)

f CO H

00

CO CO O

CO

ro

H CO H

h co co

O H H

3XHKIIV HXg

H H f

O O CO

CO H CO

CO

Ol

oa

HHC5

CO CO CO

O O CO

aXQNIIV HX^

OH©

(Nhh

CO O CO

coco co

CM

CO

01

H

CO CO o

co co o

co o h

axaxiw aug

CO CO 'O

H H(N

h«N

^

co

CO CO 00

CO CO

o h co

axnxiw aNj

0^05

OOh

lOTt*0

o co oo

CO

h-

co h o

H H ^

MON

axnxiw xsx

©©H

h OCO

OHN

co co o

CO

00

CO CO Tf*

hc<5h

co co

GO

o3 o C

a c o

" -H>

O o3

a o

CO

CO

o3 o G

£§ 2

c3 o £3

aa g

CmO £

g C3 C

"3

CO «J

-3 So
d

d o -
O cd '55

O- C o
C O fcn

c3 « C

II'3

aoti

12 M

&1§

03
T3

THE SEX DRIVE

153

TABLE 8

Showing temporal distribution of approaches, contacts and crossings of the male
groups during the twenty>-minute test period in five^minute intervals

GROUP

ACTIVITY

0 TO 5th
MINUTE

6th to 10th

MINUTE

11th to 15th

MINUTE

16th to 20th

MINUTE

TOTAL

Num-
ber

Per
cent

Num-
ber

Per
cent

Num-
ber

Per
cent

Num-
ber

Per
cent

Approaches

3

12 0

4

16.0

9

36 0

9

36.0

25

Control \

Contacts

8

38 1

7

33.0

3

14.0

3

14.0

21

Crossings

28

47.6

17

28.9

9

15.3

5

8.5

59

Approaches

4

16 6

3

12.5

12

50.0

5

20 9

24

o \

(!r»n t,$} r»f,s

2

28 6

2

28.6

2

28.6

1

14 3

7

Crossings

10

13 6

19

26.4

18

25.0

25

34 7

72

Approaches

26

31.2

17

20.4

24

28.8

16

19 2

83

6 hours ^

Contacts

31

34.4

22

24.4

19

21.1

18

19 9

90

Crossings

52

32.1

61

37 6

28

17.3

21

13.0

162

Approaches

6

12.0

22

44.0

15

30.0

7

14.0

50

12 hours {

Contacts

13

24.7

13

24 7

8

15.2

19

36.1

53

Crossings

74

30.3

76

31 .0

44

18 0

50

20.5

244

Approaches

19

41.8

16

35 2

6

13.2

5

11.0

46

1 day \

Contacts

9

29 9

9

29 9

8

26 6

4

13.3

30

Crossing

76

28.2

63

23 4

60

22.3

70

26.0

269

Approaches

15

42 0

8

22 4

9

25.2

4

11.2

36

4 days i

Contact

12

43 2

5

18 0

8

28.8

3

10.8

28

Crossings

62

24 6

55

21 8

65

25.7

70

27 7

252

Approaches

12

27 6

11

25 3

11

25 3

9

20.7

43

7 days <

Contacts

17

27 2

9

14 4

24

38.4

12

19.2

62

Crossings

47

19 2

33

13. 5

58

23 7

107

43 7

245

Approaches

9

18

8

16.0

15

30.0

18

36.0

50

28 days j

Contacts

4

10 0

9

22.0

22

55.0

5

12.5

40

Crossings

16

7 6

40

18 9

54

25.6

101

47.9

211

to the longest deprivation interval, noting the influence upon the
behavior of varying this factor.

Animals of the 0 group, that is, those which had mated freely
for two hours immediately preceding the test period, were not
inclined — either to approach or to make contact with the grid at
first. During the first five minutes a total of but four approaches,

154

THE SEX DRIVE

two contacts and ten crossings were recorded for the twenty ani-
mals in the group. The animals became more and more active as
the test proceeded, a total of twenty-five crossings being recorded
for the last five minutes of the period. Even so, the average num-
ber of approaches, of contacts and of crossings were lower for this
group than for any other but the control. Six of the animals did

Fig. 4. Showing average number of crossings (solid line), contacts (dot-
dash line) and approaches (dash line) for each interval of sex deprivation in
the male rat. Kesults for the control group are shown on the one day abscissa
since animals of this group were tested after a one day sex deprivation period,
the average number of approaches being represented by the point labeled 1,
contacts by 2, crossings by 3.

not cross at all, while fourteen crossed three times or less. Notes
taken during the course of the experiment indicate that the char-
acteristic behavior of the animals of this group, and especially of
the fourteen least active animals was, apparently, sleep. At any
rate these animals spent considerable time lying practically motion-
less in a corner of the entrance compartment. Five of the six did
not cross at all during the test period, apparently slept the entire
twenty minutes and were sleeping when removed.

THE SEX DRIVE

155

The preliminary training period for this group involved con-
siderable temporal irregularity due to the extreme inactivity of
many of the subjects. Although all of them crossed at least once
and all but three crossed at least twice during the first fifteen
minutes devoted to the preliminary training, it is doubtful whether
certain of the animals would have crossed the five times (as re-
quired by the technique adopted) in less than two hours or more.

Fig. 5. Showing temporal distribution of crossings for the male groups. The
upper line represents the percentage of crossing which occurred during the
first five minutes of the test period. The middle line represents the percentage
which occurred during the first ten minutes. The lower line represents the per-
centage which occurred during the first fifteen minutes. The points 1, 2 and 3
represent the percentages of crossing during the first five, ten and fifteen min-
utes for the control group.

For practical purposes it was necessary to make the following ex-
ception to the usual procedure. If an animal had crossed but two
or three times during the first thirty minutes of the preliminary
training period the electric shock was introduced during the next
crossing (as was done normally during the fifth crossing) and im-
mediatly following this the test period was begun. It seems un-
likely that this irregularity should vitiate the results. If an animal
was so lethargic as not to cross to the female even though no shock

156

THE SEX DRIVE

was introduced it is quite unlikely that it would cross when the
electric obstruction was present. Had we not made this exception
the preliminary training period would have required, in certain
cases, sufficient time to permit a partial recovery from the inactive
state which normally follows a period of mating. An examination
of the data for the six-hour group leads to the supposition that this
recovery is relatively rapid.

Quite a different picture is presented by the six-hour group. The
average number of crossings leaps from 3.6 to 8.1 and the ap-
proaches and contacts show a corresponding increase. A glance at
Table 6 shows that the difference between the average number of
crossings for the 0 and for the six-hour groups is reliable, being
3.06 times the standard deviation of the difference. The two groups
differ also in that the animals of the six-hour group were decidedly
more active during the first half of the test period, whereas those
of the 0 group were more active during the last half. This fact,
together with the substantial increase in the average number of
crossings can be taken as indicating that even the short interval of
six hours permits a decided recovery of the tendency of the male
to approach a sex object.

In terms of crossings the twelve-hour group shows still further
increase in activity. The average number of crossings increases to
12.2. The chances that there is a true difference between this aver-
age and the average for the six-hour group are 99.5 in 100. As in
the six-hour group more crossings wer erecorded during the earlier
part of the test period. The average number of approaches is decid-
edly less than that for the six-hour group. The same is to be said
for contacts. This would at first seem to be -contradictory. It must
be remembered, however, that by an ' ' approach ' ' is meant 1 1 an ap-
proach and withdrawal ' ' and that by a" contact ' ' is meant 1 t a con-
tact and a withdrawal." It is then quite to be expected that as
the animals cross more frequently they will approach or make
contact without crossing less frequently.

A note should be made with respect to the twelve-hour group.
On the average these animals were tested later at night than were
the animals of the other groups. Most of them were tested between
2 and 4 a.m. Any effect that this difference in conditions might

THE SEX DRIVE

157

have had would probably be in the direction of decreasing the ac-
tivity of the animals of this group, since the rats become less active
with the approach of dawn.

Animals of the one-day group did not react very differently
from those of the twelve-hour group. The average number of cross-
ings, 13.45, is but slightly larger, the difference not being reliable.
The approaches and touches show a further, but hardly significant
decrease. The chief way in which this group differs from the pre-
vious relates to the time during the test period in which the cross-
ing occurred. Whereas the crossings were most frequent during
the first half in the six- and the twelve-hour groups, in the twenty-
four group this advantage was eliminated and the crossing was
remarkably uniform throughout the twenty-minute period. A def-
inite and increasing tendency for the animals to cross more often
in the second half of the period was found in all groups of more
than twenty-four hours deprivation. Table 8 and Figure 5 show
this increasing tendency clearly.

In other respects than that of the temporal distribution of
activity just mentioned, the four and the seven-day groups differ
but slightly from the twenty-four hour group and from each other.
The slight drop in the average number of crossings from the
twenty-four-hour to the four-day group and from the latter to the
seven-day group is not reliable. When taken in conjunction with
the more pronounced drop found in the twenty-eight-day group
this tendency may be considered to have some significance. The
average number of crossings in the twenty-eight-day group is only
10.55. The difference between this average and that for the twenty-
four-hour group is 1.6 times the standard deviation of the differ-
ence, i.e., the chances that this difference is a true difference are
94 in a 100. It may be that there is an optimum deprivation
interval (optimum from the standpoint of the tendency of the ani-
mal to cross the shock to a sex object) and that beyond this
interval further deprivation interferes with the operation of the
normal sex drive. Perhaps sex activity is dependent to a slight
extent upon habit elements which are weaned by a long period of
inactivity. It might be that homosexual tendencies began to develop
during the longer deprivation intervals since the males were kept

158

THE SEX DRIVE

with other males during these intervals, and that these interfered
with the response to a heterosexual stimulus. Examination of tables
7 and 8 shows that the animals of the twenty-eight-day group
crossed more and more as the test proceeded, crossing more often
in the last six minutes than during the first fourteen minutes. This
might be taken to indicate a gradual reawakening of interest in a
heterosexual stimulus.

The sex deprivation interval used in the male control group
was twenty-four hours since it had been found that of the inter-
vals used this one had resulted in the maximum activity. The
results for the control group are therefore compared with those for
the twenty-four-hour group. These two groups differed in but one
respect. In the case of the control group there was no object in the
reward compartment while in the case of the twenty-four-hour
group there was always a female in heat in this compartment. The
results for the two groups are strikingly and reliably different.
The average number of crossings for the twenty-four-hour group
was 13.45, for the control 2.95. The difference between these aver-
ages is 9.2 times the standard deviation of this difference. This
proves conclusively that the presence of the female in heat was the
dominating factor in determining the activity of the males of the
twenty-four-hour group. Twenty-eight of the fifty-nine crossings
recorded for the control group occurred during the first five
minutes of the period and forty-five during the first ten minutes.
In the twenty-four-hour group, on the other hand almost as many
crossings occurred during the second half as during the first half of
the period. An interpretation might be that animals of the control
group underwent the punishment in response to a comparatively
less powerful and more transient drive related to exploring activity
whereas those of the twenty-four-hour group showed by their much
more frequent crossing and by the fact that they crossed with
practically undiminished vigor through the test period that they
were influenced by a drive that was both stronger and more lasting.
Obviously it should not be inferred from these remarks that it is
supposed that this experiment gives a fair and comparable measure
of the exploring drive, if we may use the term, since the region was
too limited and simple to provide suitable "reward" for such a
drive.

THE SEX DRIVE

159

To summarize : The male sex drive appeared to be at its lowest
point immediately after a two-hour period during which mating
had taken place freely. Recovery was rapid during the first six
hours and almost as rapid during the succeeding six. By twenty-
four hours the tendency of the males to cross the grid to the
female had reached its high point. During the following six days
there was apparently little change, but what change there was was
in the direction of a reduction in the strength of this tendency.
Three more weeks of deprivation resulted in a more certain indica-
tion of such reduction. If the reduction in the strength of the drive
is a true one it is in all probablity capable of rapid restoration
as indicated by the increasing number of crosses during the final
minutes of the test period in the animals sex deprived for twenty-
eight days.

TABLE 9

Showing female groups

NUMBER OF ANIMALS

22

12
6
4
7

21
3
6

10

OESTROUS CONDITION

Cornified
Early inactive
Inactive
Late inactive
Early congestive
Cornified
Late cornified
Post-ovulative
Eecuperative

OBJECT IN INCENTIVE COMPARTMENT

None (control group)

Male

Male

Male

Male

Male

Male

Male

Male

B. The female sex drive

Table 9 indicates the grouping of the 91 female animals used.
The distribution of approaches, contacts and crossings for the nine
female groups is shown in table 10. The medians, ranges, averages
and average deviations are given in table 11. In table 12, four of
the female groups (recuperative, early inactive, inactive, late
inactive) have been combined. This composite group, containing
32 animals, represents the general dioestrous or inactive period of
the oestrous cycle. This composite group is compared with the
cornified group (21 animals) and the control group (22 animals)
in this table (table 12). The medians, ranges, averages, standard
deviations and coefficients of variability are given. Table 13 deals
with the reliability of the difference between the average number

160

THE SEX DRIVE

of crossings for the cornified and control groups and for the corni-
fied and composite dioestrum group. The approaches, contacts and
crossings for all the female groups as distributed during the twenty

TABLE 10

Distribution table covering approaches, contacts and crossings of the females as
correlated with the histological character of the vaginal secretions*

CONTROL

(22)

(NO SEX

object)

EARLY
INACTIVE
(12)

INACTIVE

(6)

LATE
INACTIVE

(-4)

EARLY
CONGES-
TIVE
(7)

CORNI-
FIED
(21)

LATE
CORNI-
FIED

13)

POST-
OVULA-
TIVE
(6)

Grouping is based upon histological characteristics of the vaginal secretion.

minute test period (in one-minute units) is given in table 14. In
table 15 these data are given in terms of the percentage of each
activity occurring during the four successive five minutes of the

THE SEX DRIVE

161

twenty-minute test period. Figure 6 shows graphically the effect
of variation in oestrous condition upon the average number of
approaches, contacts and crossings. Figure 7 shows the temporal

TABLE 11
Showing results of the female groups

APPROACHES

CONTACTS

CROSSINGS

GROUP

a

©

So

c

b£ cS

c

©

c
_o
a>°-3
Si «

a

©

a

o
©'•«

2

©

CO

.2

©
bp

to

c3

2>

3

©

bC

bO
a

£■£

9

a
«s

©

£ ©

-5

c

©

C

©

s

n

>

<

<

i

«?

>

<

<

©

cj

K

>
<

>T3

Control

3.0

0-10

3.23

2.34

2.0

0-6

2.27

1.66

5.0

1-9

5.05

2.2

Early inactive

0.0

0-2

0.42

0.63

1.0

0-3

1.25

0.75

1.0

0-5

1.5

1.42

1.0

0-4-

1.5

1.3

1

0-2

1.17

0.56

0

0-4

0.83

1.11

Late inactive

2.5

0-5

2.5

1.5

1.5

0-2

1.25

0.75

2.0

0-3

1.75

0.87

Early congestive

7.0

0-14

6.71

4.3

4

1-9

4.86

1.84

14.0

0-22

11.14

4.85

Cornified

5.0

0-14

5.43

2.93

3.0

0-9

3.14

2.1

16.0

2-24

14.61

4.21

Late cornified

2.0

1-4

2.33

1.11

4.0

2-5

3.66

1.11

10.0

2-14

8.66

4.43

Post-o vulative

0.5

<M

1.17

0.83

3.5

0-9

4.0

2.33

2.5

0-12

4.66

4.21

Recuperative

1.0

0-4

1.3

0.99

1.5

0-5

2.0

1.4

0.5

0.9

1.3

TABLE 12

Showing results of the cornified group, the control group and a composite dioes-
trous group composed of the recuperative, the early inactive,
the inactive and the late inactive groups

APPROACHES

CONTACTS

CROSSINGS

NUMBER OP

c

t of

lity

c

c

o >>

ANIMALS

O

S

■oj

«.2

11

C

bfl

ard
iati<

©.o
'S.2

C

©

bO

•g.2

c5 eS

3ien
abil

.2
-a

©

9
bfl

a

eg
%*
©

"S >

C c9

» ~o

oeffi<
vari

CO

t3

CP

U)

c

oaj

eS

-o >

© g
o >

03

T3

©

to
c

eS

>

S »

oeffii
vari

&

<

CO

O

a

«

<

O

<

GO

O

Cornified (21)

5

0-14

5.43

3.63

66.9

3

0-9

3.14

2.73

86.8

16

2-24

14.61

5.14

36.

Composite

(dioestrum)

(32)

1

0-4

1.16

1.42

117.3

1

0-5

1.47

1.27

86.3

1

0-5

1.34

1.57

117.

Control (cor-

nified, no

sex object)

(22)

3

0-10

3.23

2.7

83.6

2

0-6

2.27

1.91

84.1

5

1-9

5.05

2.55

50.

distribution of such activity in terms of the per cent occuring dur-
ing successive five minutes.

162

THE SEX DRIVE

The data on the behavior of the female animals were grouped
according to the interpretation of the vaginal smears taken immedi-
ately after the test periods. Of the 69 animals tested with a male in
the reward compartment 21 were found to have been in the corni-
fied stage at the time of testing. It has been commonly observed

TABLE 13

Showing reliability of the difference between the average crossings of the con-
trol, the cornified and the composite dioestrous group

GROUP

STANDARD
DEVIATION
OB" THE
DIFFERENCE

THE DIF-
FERENCE

DIFFERENCE
S.D. OF DIFF.

CHANCES
IN 100 OF
A TRITE
DIFFER-
ENCE
GREATER
THAN 0

Cornified and control (1)

1.35

9.1

6.74

100

Cornified and composite (dioes-

trum) (2)

1.16

12.8

11.0

100

Composite and control _ _

0.61

3.71

6.08

100

(1) These two groups differ only in that the incentive compartment con-
tained a male during the testing of the cornified group and was empty during
the testing of the control group. Animals of both groups were in the cornified
stage.

(2) These two groups differ only with respect to oestrous condition, the first
representing oestrum, the second dioestrum. A male was in the incentive com-
partment when both groups were tested.

that the most active sex behavior on the part of the female occurs
at this time. Our results are in accordance with such observation.
Animals in this group crossd, on the average, more than ten times
as often as those in the four groups which together may be called
the composite dioestrous group. The difference between the aver-
ages is a true one being over five times the standard deviation of
that difference. Animals in dioestrum not only showed little ten-
dency to cross (13 of the 28 did not cross at all) but were not
inclined to approach or make contact with the grid more than once.
The average number of approaches was 1.16, of contacts, 1.47. For
the cornified group the figures are 5.43 and 3.14. The differences
are reliable though not so great as in the case of the crossings.
With respect to the time during the testing period when the animals
crossed most often the cornified and dioestrous differ. Animals in
the former group tended to cross more and more as the test period
proceeded. Over half of the crossings occurred during the final six
minutes. Just the reverse is true for animals in dioestrum. Such

THE SEX DRIVE

163

animals not only crossed less often but were less and less inclined to
cross as the test period progressed. Over half of the crossings
occurred during the first six minutes and less than a tenth during
the final six minutes.

Fig. 6. Showing average number of crossings (solid line), contacts (dot-
dash line) and approaches (dash line) for the female groups (grouping being
based upon histological character of vaginal smears). Eesults for the control
group are shown on the abscissa of the Cornified group since animals of the
control group were tested when in the cornified stage. The average number of
approaches for the control group is represented by the point labeled 1, contacts
by 2, crossings by S.

Since we believe that the physiological condition related to the
oestrous rhythm was the only factor which was different in the two
groups we feel justified in the conclusion that the decided difference
in behavior which has just been noted is related to the oestrous
cycle.

The relationship between the cycle and the behavior recorded is
demonstrated clearly by consideration of the progressive changes

164

3 in NIK HX()g

THE SEX DRIVE

Ohh OOO OOO O i— < CO iflH"* O h Oho

H «— 1

ainNiw HXgx

OO-i OOO OOO OM(N COC0O0 Ohm OOh

CM

3XHNIK HXgJ

OOO OOO OOO MCON CO 00 O OOh O O h

CO

3XQNIK HX^J

OOh OOO OOO C0*O)C© h lo h cooco ooco

CO

axnxiK H19X

O1-1O i-H O O — 1 O O 1-1 CO 00 H 0 CO O h H OO'-i

CO

3XQXIW HX<JJ

OOO O h O OOO MhN C0CO*O O CO CO OOO

CO

axriNiw Hif]

OOH OOO HOO rHrHM C0"^CO HOM OOO

3XANIW HXgJ

Ohh OhO hOO CO h CO O CO h OOh h CO O

3X0NIIV HX£J

h O O OOO O h i-h tThco b~00 Oho O H 1— i

3XQKIJV HXJJ

OOh h O h OhO CO h lO CO CO O OOh O CO CO

3XQNIIV HXOI

OOh h O h OhO h O l>» iO H 0 OOO O h 1-1

HlflNIK HXg

h ^ CO HOO hOO "«^hH COCOOS OhO CO CO CO

SIilXIK HX8

OCOh OCOO COOh 10 h CO rococo OOH ocoo

OCOO COOO OOh CO h h 00COO0 HON O CO CO

CO CO CO OOO h O h h 0 h t^COb- OOO Ohio

Ohm OOh OOCO CO — 1 CO OCOO h CO CO h i— < CO

1—i 1—1

HX^

OOh OOh OhO h H H CO CO CO hOh h r-i h

3XQNIW QHg

h h O OCOO Oho CO CO CO h—iCO OhO h CO CO

3XANIK ONJJ

OOO h~hO h O O CO O h rH 0 h OOO OHCO

sxriNiw xsi

OC0C0 COOh (NOh CO h CO CO H _ _ CO OOCO

ACTIVITY

Approaches

Contacts

Crossings

Approaches

Contacts

Crossings

Approaches

Contacts

Crossings

Approaches

Contacts

Crossings

Approaches

Contacts

Crossings

Approaches

Contacts

Crossings

Approaches

Contacts

Crossings

GROUP

....

Early inac- \
tive I

Inactive

Late inac-
tive

Early con-
gestive

Cornified j

Late corni-
fied

Post-ovula-
tive

THE SEX DRIVE

165

5 £-

•rf'fc

*C3
o

i

axaNiw moz

axnNiw Hi6l

O rH O rH r— ( rH © r— 1 •— 1

3inNIW HX81

O O O O i— ( o <_> v^J

axnNiK hx^i

rH i— 1 O O rH rH rH i-H rH

axnxiw Hxgi

■ /"s , >■», ■ —44 /vi /^s
<— > h o 1X3 f~i ~ CN CN <W

3XOMIW HXSI

O 1 rH O CO CO O C3 rH

axnNiw hxi-i

j-H©rH HON OOM

axn.Miw Hxgi

© HH CI O d rH CO d

aXflKIK hxji

© © © lO t^" CO »-H t-H i-H

3XQKIK HlJl

© C3 rH © tO t~** rH CO CO

3XQNIK HXOI

l-H ^H rH (^ft /VI fVl Cl?l

rH CI © © CO CO CO id

3iaMIR HXg

r^ t t »-v r"i la J r"H ^nj

3X0XIK HX£

rH CI © l^* *0 Is" CO ^ rH

aiamn hxq

V*N 1»N J ^ J V.' J tlj

CM

3xqnin Hxg

wN *^ UN JT"* UN U»> UN UN

axntiiK nif

Ml , i »n rO k_
V— UN i±J VU Is* U Cv

axflNiw aag

UN 41 J UN ^1 CN

3XANIK anz

©©© ©rH© NHO

axnNiw X9i

(NCO-H (M CO rH »OiOiO

ACTIVITY

Approaches

Contacts

Crossings

Approaches

Contacts

Crossings

Approaches

(J 11 1 <-lv I/O

Crossings

GROUP

Recupera- \
tive [

Control

Composite f
(dioe- <
strum) (

166

THE SEX DRIVE

TABLE 15

Showing temporal distribution of approaches, contacts and crossings of the female
groups during the twenty-minute test period in five-minute intervals

ACTIVITY

0 to 5th

MINUTE

6th to
10th

MINUTE

11th to
15th

MINUTE

16th to
20th

MINUTE

TOTAL

Number

Per cent

Number

Per cent

Number

Per cent

Number

Per cent

Approaches

19

26.8

32

45.1

14

19.7

6

8.5

71

Contacts

17

34.0

13

26.0

15

30.0

5

10.0

50

Crossings

48

43 9

3fi

39 4

17

1 5 3
xO. o

1U

Q Q

ii i
in

Approaches

1

20.0

3

60.0

1

20.0

0

0.0

5

Contacts

4

26.6

8

53.3

1

6.6

2

13.0

15

Crossings

Q

33.3

A
\j

33.3

3

Ifi fi
xO. D

3

ifi fi

18

XO

Approaches

3

33.3

4

44 4

1

11.1

1

11.1

9

Contacts

3

42.9

2

28.6

2

28.6

0

0 0

7

Crossings

3

60.0

1

20.0

I

0

V/

0 0

Approaches

3

30.0

4

40.0

2

20.0

1

10.0

10

Contacts

2

40.0

1

20.0

2

40.0

0

0.0

5

Crossings

3

42.9

3

42 Q

I

14.3

o

0.0

7

Approaches

11

23.4

17

36.2

13

27.7

6

12.8

47

Contacts

6

17.7

3

8.8

8

23.5

17

50.0

34

Crossings

H

14. 1

18
xo

23. 1

21

26.9

28

35.9

78

Approaches

49

43.0

29

25.4

23

20.2

13

11.4

114

Contacts

8

12.1

16

24.2

16

24.2

26

39.4

66

Crossings

ov

1Q 9

4s*

to

14 7

It . 4

81
ox

9ft 4
^o. t

199
ILL

3Q 7

307

Approaches

3

42.9

1

14.3

1

14.3

2

28.6

7

Contacts

4

36.4

1

10.0

3

27.3

3

27.3

11

Crossings

5

19.2

3

11.5

7

26.9

11

42.3

26

Approaches

3

42.9

2

28.6

1

14.3

1

14.3

• 7

Contacts

5

20.8

9

37.5

9

37.5

1

4.2

24

Crossings

9

32.1

10

35.7

3

10.7

5

21.4

28

Approaches

5

38.5

6

46.2

1

7.7

1

7.7

13

Contacts

6

30.0

7

33.0

4

20.0

3

15.0

20

Crossings

6

46.2

3

23.1

4

30.7

0

0.0

13

Approaches

12

32.4

17

45.9

5

13.5

3

8.1

37

Contacts

15

31.9

18

38.3

9

19.1

5

10.7

47

Crossings

18

41.9

13

30.2

9

20.9

3

7.0

43

GROUP

Control.

Early inactive.

Inactive.

Late inactive.

Early congestive.

Cornified.

Late cornified.

Post-ovulative.

Recuperative.

Composite.

THE SEX DRIVE

167

in behavior accompanying the physiological change. As noted
above, the interpretation of the vaginal smear record has led to the
classification of the animals into eight groups. Four of these have
been treated together (table 11), as representing the general dioes-
trous period and have been compared with a fifth, the cornified
group representing oestrum, or heat. These groups represent the
extremes, the periods of the least and the greatest sex activity. The

Fig. 7. Showing temporal distribution of crossings for the female groups.
The upper line represents the percentage of crossing which occurred during
the first five minutes of the test period. The middle line represents the per-
centage which occurred during the first ten minutes. The lower line represents
the percentage which occurred during the first fifteen minutes. The points 1,
2, and S represent the percentages of crossing during the first five, ten and
fifteen minutes for the control group.

remaining three groups represent transitional stages. The early
congestive group contains animals which were no longer in dioes-
trum but had not yet reached the stage when only cornified cells
were present in the smear. This period could be described as one of
preparation. The two remaining groups, the late cornified and the
post-ovulative, represent the transition from the condition of the
greatest sex activity to that of the least. These groups are small

0

ID

168

THE SEX DRIVE

and so might well have been combined into a single post-oestrous
group. But since they can be definitely separated into two groups,
the one slightly more closely related to the cornified period and the
other to the inactive period, there seems no harm in letting them
stand as originally classified.

Since the complete oestrous cycle is a continuously recurring
phenomenon a description might be started at any point in the
process. We will start with the consideration of the early con-
gestive group.

This group contained 7 individuals. The average number of
crossings was 11.14, with a standard deviation of 4.85. This large
standard deviation indicates the lack of uniformity of the group
with respect to behavior. One animal crossed not at all and one
crossed but 3 times. Such records are typical of the dioestrous
period just preceeding this. On the other hand, 4 of the animals
crossed 12 or more times, one of these crossing 22 times. Records
such as these resemble those found in the cornified group repre-
senting a stage immediately following this. The conclusion that
is suggested is that, so far as behavior is concerned, there is not a
gradual transition from dioestrum into oestrum. That is, there
is no pre-oestrum period during which the female rather half-
heartedly seeks the male. The sex drive in the female is held entirely
in abeyance until the time when it is released and it then acts with
full force.

Results for the cornified group have been mentioned above where
they have been contrasted with those for the dioestrous groups.
Of the 21 animals in the cornified group 16 crossed 14 or more
times. The average number of crossings, 14.14 is the highest group
average, verifying the assumption that the maximum sex activity
accompanies the early cornified stage. The average number of ap-
proaches, 5.43 and of contacts, 3.14 are higher than the corre-
sponding averages for any of the dioestrous groups, though not
quite so high as found in certain transitional groups. That the
average number of approaches and contacts is not decidedly greater
than in the case of the transitional groups might be explained as
follows. Whereas approaches, contacts and crossings are all meas-
ures of general activity and are all, therefore, likely to show higher

THE SEX DRIVE

169

values for the general oestrous period (including transitional
stages) than for the general inactive period, the crossings, rather
more than the approaches and contacts, measure also the specific
orientation to this activity toward the sex object. Such orienta-
tion is, of course, characteristic of the cornified stage, or oestrum
proper.

The late cornified group contains but 3 animals. Thus the aver-
ages are not particularly significant (approaches, 2.33; contacts,
3.66; crossings, 8.66). As far as the data go, the indication is that
the drive is less intense or specific than earlier during the cornified
stage.

The post-ovulative group is also rather small (6 animals). Since,
however, the average number of crossings for the late cornified, the
post-ovulate and the recuperative groups shows a consistent decline
it seems reasonable to suppose that the downward trend is a true
one and, moreover, a gradual one especially when compared with
the more sudden onset of heat.

The recuperative group, containing ten animals, shows a low de-
gree of activity, as 5 animals refused to cross and 2 crossed but
once. The average is 0.9. Approaches and contacts are also low.
This stage may be said, then, to mark the beginning of the dioes-
trum.

The early inactive group contains 12 animals. The average num-
ber of approaches and of contacts are lower for this group than for
the preceding one, but there is a slight increase in the average
numfber of crossings. While this increase is not great, there is a
suggestion that it may possibly represent something other than a
chance irregularity in the data regarding the temporal distribu-
tion of crossings. (Fig. 7). It is seen that in the other three groups
representing the dioestrum (recuperative, inactive and late inac-
tive) about half of the crossings occurred during the first five min-
utes, about a quarter during the second five minutes, the remaining
quarter during the third five minutes and none at all during the
final five minutes. In the early inactive group, on the other hand,
there is found crossing rather more uniformly distributed through-
out the twenty minutes. In other words, this group resembles the
cornified group more than does any of the other dioestrous groups.

170

THE SEX DRIVE

The suggestion, and it is only that, is that there may be a slight
recurrence of oestrum-like behavior occurring at about the middle
of the dioestrous period.

The inactive group, containing 6 animals, represents the lowest
point in the cycle, so far as average number of crossings are con-
cerned, although it is but a shade lower than the recuperative; 4
of the 6 animals failed to cross. The average number of contacts is
also at its lowest point in this group. The average number of ap-
proaches shows a slight increase, an increase which is continued in
the two succeeding groups.

The late inactive group (4 animals) may be said to close the
dioestrous period. All three forms of activity recorded show a mod-
erate increase, but the group is clearly a dioestrous group, giving
a very different behavior picture from the rather hybrid group
which follows it (the early congestive) with which this descrip-
tion of the cycle commenced.

The control group differs from all the other female groups in that
there was no male in the incentive compartment during either the
preliminary or the test period. This compartment was entirely
empty. This group resembles the cornified group in all other re-
spects, i.e., the animals were in the cornified stage and any differ-
ences in the behavior of the two groups must have been due to the
presence of the sex object in the one case and its absence in the
other. There are decided and reliable differences. The control
group was much less active in terms of approaches, contacts and
crossings. The average number of crossings for the control group
was 5.05, for the cornified group, 14.14. The difference between
these averages is 6.74 times the standard deviation of that differ-
ence. In the cornified group 17 of the 21 animals crossed more
frequently than did any of the 22 animals in the control. The con-
clusion is that the behavior observed in the cornified group was
largely determined by the presence of the male.

A further comparison, however, shows that important though the
presence of the male is in determining the behavior of the animals
of the cornified group, the physiological condition of the animals
also plays an important part. It should be noted that the average
number of crossings for the control group is higher than that for

THE SEX DRIVE

171

any other group except the cornified. In other words, a female in
oestrum is more likely to undergo the electric shock in crossing to
an empty compartment than is a female not in oestrum in crossing
to a compartment containing a male. The implication is that the
internal stimulation accompanying oestrum furnishes more drive
than the external stimulation of the presence of the male. No claim
is made for the specificity of the drive furnished by such internal
stimulation. Wang and others have noticed the great increase in
" spontaneous ' ' activity accompanying oestrum. Such activity, ir-
respective of its source and normal mode of ''satisfaction," would
be reflected in data taken using the obstruction method.

These data indicate a definite relationship between the histolog-
ical characteristics of the vaginal smear and the behavior of the fe-
male rat in a situation involving crossing an electric grid to reach
the male. The onset of activity directed toward a sex object is
apparently rather abrupt and corresponds to the very early part
of the cornified stage. The cessation of this activity appears to be
more gradual. During the period when the vaginal secretion con-
tains predominantly epithelial cells and leucocytes activity is much
reduced. None of the 32 females which were in this condition when
tested crossed more than 5 times whereas only 3 of the 21 females
of the cornified group crossed as few as 5 times or less while 16
crossed 14 times or more.

C. Comparison of the male and the female sex drive

To obtain a fair comparison of the sex drive in the male and in
the female we should compare the two groups which represent the
drive at its maximum in each sex. These are the twenty -four hours
sex deprivation group for the male and the cornified group for the
female. In terms of crossings there is not much difference. The
average number for the females, 14.14 in twenty minutes is slightly
but unreliably greater than that for the males, 13.45. In terms of
approaches and contacts the female group has a more decided ad-
vantage, indicating, perhaps, that the behavior in the females was
more definitely oriented toward the incentive compartment even
though they did not cross much more frequently than did the males.

An important influence operated in the favor of the male drive.

172

THE SEX DRIVE

This consisted of the behavior of the stimulus animal. When the
females were tested the stimulus animals were, of course, males.
These stimulus males were selcted for their docility and activity.
No male that showed pugnacious tendencies or that was lethargic
was used. Nevertheless, as stimulus objects they were hardly com-
parable to the female stimulus animals used in the testing of the
males. These female stimulus animals were in the cornified stage
when used. They were much more active than were the male stim-
ulus animals and their activity was of the type which arouses and
encourages sex behavior in the male. Were these stimulus females
as passive as the male stimulus animals it seems not unlikely that
the number of crossings of the male test animals would have been
less and that there might even have been a reliable difference be-
tween the average number of crossings for the two sexes in the
favor of the female.

For a detailed description of the type of female behavior re-
ferred to above see C. P. Stone (32, a. e).

A further difference in the operation of the sex drive in the male
and in the female is suggested by an examination of the results for
the control groups for the two sexes. We have noted above that
the female control group (cornified, empty incentive compartment)
crossed much more than did the female dioestrous groups (inactive
stages, male in incentive compartment) indicating that the internal
stimulation accompanying oestrum plays an important part in the
behavior. That this is not the case for the males is seen by glanc-
ing at the results for the male control group. Animals in this
group, although in the physiological condition during which the
maximum sex activity is apparently displayed (twenty-four hours
sex deprivation) did not display as much activity in terms of either
approaches, contacts or crossings as did the animals in the other
male groups. Even the " satiated" males (0 group) showed more
activity than did the control. The indication is, then, that the ex-
ternal stimulus situation, the presence of the female in heat, is the
dominating factor in the determination of sex activity in the male.

THE SEX DRIVE

173

V. Summary of Conclusions

A. Male

1. The tendency of a male rat to approach a female rat in oes-
trum may be measured in terms of the number of times it will cross
an electrical obstruction to reach the female within a given period
of time. That the behavior of such males was dominated by the
presence of the female in the present experiment was indicated by
the results of a control group tested under conditions exactly like
those of one of the test groups except that there was no female
present in the incentive compartment. The males in this control
group crossed the obstruction to the incentive compartment decid-
edly and reliably less often than did the animals in the comparable
test group.

2. The tendency of a male rat to cross an electrical obstruction
to a female rat in oestrum is at its low point immediately after a
period (two hours in this case) during which the male has had
access to and mated frequently with a female in oestrum. Recov-
ery of the tendency from this low point is rapid during the first
six hours after such a period of mating, almost as rapid during the
second six hours after which it has reached a point- only slightly
below the maximum manifestation of this tendency which is found
in an animal twenty-four hours after a period of mating.

3. Intervals of sex deprivation longer than one day do not in-
crease the tendency of the male to cross. The data at hand suggest
that this tendency decreases slightly from this point on to a depri-
vation interval of twenty-eight days, the longest interval studied.
This decrease is not statistically reliable, however, and the only
conclusion upon this point is that there is no increase in the ten-
dency after the first day.

B. Female

1. The tendency of a female rat to approach a male rate may be
measured in terms of the number of times it will cross an electrical
obstruction to reach the male within a given period of time. That
the behavior of the females in oestrum was dominated by the pres-
ence of the male in the present experiment was indicated by the
results of a control group run under conditions exactly like those

174

TEE SEX DRIVE

of one of the test groups except that there was no male present in
the incentive compartment. The females in this control group
crossed the obstruction to the incentive compartment decidedly and
reliably less often than did the animals in the comparable test
group.

2. There is a very definite relationship between the histological
character of the vaginal secretion of a female rat and its tendency,
at the moment, to cross an electrical obstruction to reach a male,
i.e., between the oestrous rhythm and behavior toward a sex object.
The group of animals, which judging by the vaginal smear, were
tested during the early part of the stage during which cornified
cells only are found in the secretion displayed a decidedly and re-
liably greater amount of crossing and other activity oriented
toward the male than did the group whose vaginal secretion was
characterized by the presence of epithelial cells and leucocytes. In
other words, activity directed toward a male is confined rather
rigidly to a single period during the oestrous cycle, the period
usually known as oestrum.

3. From the standpoint of behavior the onset of oestrum is sud-
den while its cessation is relatively gradual. The behavior of the
group representing the transition into oestrum indicates that the
group is not homogeneous but contains certain animals whose be-
havior resembles that commonly seen in dioestrum, and others
whose behavior resembles that commonly seen in oestrum. When
oestrum sets in it is apparently with practically full force. The
groups representing transition out of oestrum show, in general,
behavior midway between that seen in oestrum and that seen in
dioestrum. The transition out of oestrum appears to be gradual.

C. Comparison of the male and the female

1. Since the strength of the tendency of the rat to cross an elec-
tric obstruction to reach a sex object is related to quite different
physiological factors in the two sexes, the only fair comparison of
the sexes from this standpoint is found in the comparison of those
groups in which the physiological condition is such that the ten-
dency was at its maximum for each sex. For the male this group
is that tested after one day of sex deprivation. For the female it is

THE SEX DRIVE

175

the group of animals tested when in the early cornified stage. The
behavior record shows that of these two groups the female group
was the more active although the difference is not statistically re-
liable. When it is considered that the incentive stimulus in the case
of the male (a female in oestrum) was in most cases extremely ac-
tive, displaying that type of activity commonly preceding mating
whereas the stimulus animal for the female (a male) was decidedly
less active and less frequently made such advances, the assumption
seems justified that were it possible to equate the activity of the
incentive stimuli the female group would be found to show a re-
liably greater amount of activity directed toward a sex object.

2. The behavior data indicate that sex activity is initiated rather
more by the external stimulus situation in the male than in the
female and rather more by internal stimulation in the female than
in the male. The female rat in oestrum displays more activity in
the form of crossing the electrical obstruction to the incentive com-
partment even though that compartment be empty than does the
female in dioestrum even though the compartment contains a male.
A male rat which has not mated for twenty-four hours (the inter-
val which, of those studied, found sex activity at its maximum) will
cross electrical obstruction less often to an empty incentive com-
partment than will a male of any of the other sex deprivation
intervals studied (0, 6, 12 hours, 4, 7, 28 days) to an incentive com-
partment containing a female in oestrum.

Bibliography

(1) Allen, E. 1922. The oestrous cycle in the mouse. Amer. Jour.
Anat., xxx, 297-348.

1923. Racial and familial cyclic inheritance and other evi-
dence from the mouse concerning the cause of oestrous phenomena.
Amer. Jour. Anat., xxxii, 293-304.

(2) Bischoff, Th. L. W. 1852. Entwicklungsgeschichte des Meer-
schweinchens. Giessen.

(3) Blair, E. W. 1922. Contraction rate of the uterine musculature of
the rat with reference to the oestrous cycle. Proc. Amer. Assoc.
Anat., Anat. Bee. xxiii, 9-10.

(4) Corner, G. W. 1921. A review of some recent work on the mam-
malian reproductive cycle. Jour. Mammal, ii, 227-31.

176

THE SEX DRIVE

Corner, G. W. 1921. Cyclic variation in uterine and tubal contrac-
tion waves. Amer. Jour. Anat., xxxii, 345-51.

(5) Greenman, M. J., and F. L. Duhring. 1923. Breeding and care of
the albino rat for research purposes. The Wistar Institute of Anat-
omy and Biology, Philadelphia.

(6) Hartman, C. 1923. The oestrous cycle in the opposum. Amer.
Jour. Anat., xxxii, 353-95.

(7) Heape, W. 1900. The sexual season of mammals and the relation
of the "Pro-oestrum" to menstruation. Quar. Jour. Micr. Sci. xliv,
1-70.

(8) Hensen, V. 1876. Beobachtungen iiber die Befruchtung und Ent-
wickelungen des Kaninchens und Meerscheinchens. Zeit. f. Anat. u.
Entwick., i, 213-72; 353-423.

(9) Holden, F. 1926. A study of the effect of starvation upon behavior
by means of the obstruction method. Comp. Psych. Mon., 3, No. 17,
45 p.

(10) Ishii, 0. 1920. Observations on the sexual cycle of the guinea-pig.
Biol. Bull., xxxviii, 237.

1922-. Observations on the sexual cycle in the white rat.

Anat. Bee, xxviii, 311.

(11) Jenkins, T. N., L. H. Warner, and C. J. Warden. 1926. Stand-
ard apparatus for the study of animal motivation. Jour. Comp.
Psych., vi, 361.

(12) Keye, J. D. 1923. Periodic variation in spontaneous contraction of
uterine muscle, in relation to the oestrous cycle and early pregnancy.
Bull. Johns Hopkins Hosp., xxxiv.

(13) Konigstein, H. 1907. Die Veranderungen der Genital Schleimhaut
wahrend der Graviditat und Brunst bei einigen Nagern. Arch. f.
Physiol., cxix, 553-70.

(14) Lataste, F. 1892. Transformation periodique de l'epithelium du
vagin des rongeurs. Mem. de la Soc. de Biol., xliv, 765-69.
1893. Rhythme vaginal des Mammiferes. Ibid., xlv, 135-46.

(15) Loeb, L. a. 1911. The cyclic changes in the ovary of the guinea-
pig. Jour. Morph., xxii, 37-70.

b. 1914. The correlation between the cyclic changes in the

uterus and the ovaries of the guinea-pig. Biol. Bull., xxvii, 1-44.
c. 1923. The mechanism of the sexual cycle with special ref-
erence to the corpus luteum. Amer. Jour. Anat., xxxii, 305-43.

(16) Long, J. A., and H. M. Evans. 1922. The oestrous cycle in the rat
and its related phenomena. Mem. Univ. Calif., vi, 1-148.

(17) Marshall, F. H. A. 1922. Physiology of reproduction. London:
Longmans, Green & Co.

(18) Morau, H. 1889. Des transformations epitheliales de la muqueuse

THE SEX DRIVE

177

du vagin de quelques rongeurs. J. de VAnat. et de la Physiol., xxv,
275-97.

(19) Moss, F. A. 1924. A study of animal drives. Jour. Ex. Psych., vii,
165.

(20) Papanicolaou, G. N. 1923. Oestrus in mammals from the com-
parative point of view. Amer. Jour. Anat., xxxii, 284-92.

(21) Rein, G. 1883. Beitrage zur Ke'nntniss der Reifungserscheinungen
und Befruchtungsvorgange am Saugethierei. Arch. f. Mikr, Anat.,
xxii, 233-70.

(22) Retterer, E. a. 1892. Sur la morphologie et revolution de epithe-
lium au vagin mammiferes. Mem. Compt. rend. Soc. de biol., xliv,
101-7.

b. 1892. Evolution de Perithelium du vagin. Ibid, 566-8.

c. Sur les modifications de la muqueuse uterine Tepoque du

rut. Ibid., 637-42.

(23) Rubaschkin, W. 1905. Uber die Reigungs- und Befruchtungs-
prozesse des Meersehweincheneies. Anat. Hefte, Wiesbaden, xxix,
507-53.

(24) Seckinger, D. D. 1923. Spontaneous contractions of the fallopian
tube of the domestic pig with reference to the oestrous cycle. Bidl.
Johns Hopkins Hosp., xxxiv.

(25) Selle, P. M. 1922. Changes in the vaginal epithelium of the
guinea-pig during the oestrous cycle. Amer. Jour. Anat., xxx,
429-49.

(26) Simmons, R. 1924. The relative effectiveness of certain incentives
in animal learning. Comp. Psych. Monographs, 2, No. 7.

(27) Slonaker, J. R. 1924. The effect of pubescence, oestruation, and
menopause on the voluntary activity in the albino rat. Amer. Jour.
Physiol., lxviii.

1925. Analysis of the daily activity of the albino rat. Ibid.,

Ixxiii.

1926. Long fluctuations in voluntary activity in the albino

rat. Ibid., lxxvii.

(28) Smith, H. P. 1917. The ovarian cycle in mice. Jour. Roy. Mic. Soc,
p. 252.

(29) Sobotta, J. 1895. Die Befruchtung und Furchung des Eies der
Maus. Arch. f. Mikr. Anat., xlv, 15-93.

(30) Stockard, C. R. 1923. The general morphological and physiolog-
ical importance of the oestrous problem. Amer. Jour. Anat., xxxii,
227-83.

(31) Stockard, C. R., and G. N. Papanicolaou, a. 1917. The existence
of a typical oestrous cycle in the guinea-pig, with a study of its his-
tological and physiological changes. Amer. Jour. Anat., xxii, 225-83.
b. 1919. The vaginal closure membrane, copulation, and the

178

THE SEX DRIVE

vaginal plug in the guinea-pig, with further consideration of the
oestrous rhythm. Biol. Bull., xxxvii, 222-43.

(32) Stone, CP. a. 1922. The congenital sexual behavior of the young
male albino rat. Jour. Comp. Psych., ii.

b. 1923. Further study of sensory function in activation of

sexual behavior in the male rat. Jour. Comp. Psych., iii.

c. 1924. The awakening of copulation ability in the male

albino rat. Amer. Jour. Physiol., lxviii.

d. 1924. Delay in the awakening of copulatory ability in the

male albino rat incurred by defective diets. Jour. Comp. Psych.,
iv and v.

e. 1926. The initial copulatory response of female rats reared

in isolation from the age of twenty days to the age of puberty.
Jour. Comp. Psych., vi, 73-83.

(33) Tsai, C. 1925. The relative strength of sex and hunger motives in
the albino rat. Jour. Comp. Psych., v, 407.

(34) Wang, G. H. 1923. Spontaneous activity and the oestrous cycle.
Comp. Psych. Monographs, ii, No. 6.

2. THE EFFECT OF SEGREGATION ON THE SEX
BEHAVIOR OF THE WHITE RAT AS MEASURED
BY THE OBSTRUCTION METHOD 1

Marion Jenkins
I. Introduction

Through various experimental conditions it is possible to prevent
the normal expression of sex behavior. Environmental conditions,
for example, may be imposed which interfere with the consumma-
tion of the normal sex drive. Normal is here used in the biological
sense to designate that form of sex behavior which ordinarily cul-
minates in reproduction; or more specifically, the response of a
male to a receptive female, or the response of a receptive female
to a male. Interference with the operation of the normal sex drive
is commonly brought about by the use of two methods : namely,
isolation and segregation. Of these two methods, isolation is prob-
ably by far the greatest variation from the normal condition. It
has usually been found impracticable to isolate completely. In fact,
isolation rarely means more than eliminating visual and tactual
stimuli from other animals of the same species. Usually little or no
control of odors, sounds, etc., is exercised. The usual condition of
isolation is separation from other members of the same species. In
such a case, it may be possible for objects in the cage or human
beings who care for the animals to elicit the sex response. Sex
segregation, as we use the term, differs from isolation in that, al-
though the normal outlet for the consummation of the sex drive has
been prevented, nevertheless segregated animals still have an op-
portunity to exhibit sexual behavior toward members of their own
sex.

These two methods of environmental control have been used by

i Eeprinted with modifications from Genetic Psychology Monographs, 1928,
3: 457-571.

179

180

THE SEX DRIVE

various investigators, but, in most cases, the studies have been
limited to cage observations without experimental measurement.
Studies of isolation will be considered first.

Interesting among these are the cage studies of Craig (2), who
observed the behavior of doves which had been isolated long before
puberty. He states that, with his conditions of isolation, the doves
could "sometimes hear other doves, but they could never touch or
see them." The four cases which he discusses in some detail did
not manifest "masculine display behavior" until socially stimulated
by the presence of a female dove. Then it appeared suddenly, but
' ' human beings seemed to be the symbol ' ' ; the sex response appear-
ed to be elicited chiefly by the hand and shoe. Three of the four
finally "gave up their intimate friendship for human beings. But
they gave it up slowly and gradually, showing interesting division
of attention between human and dove companions. ' ' The other one
never entirely gave up cooing to human beings. Indeed Craig says,
"when I came near his cage, he still showed a desire to get out to
me, and jealousy of other doves in my presence."

More extensive and detailed observations of the effect of isolation
have been made by Stone (13). He observed the initial copulatory
response of 21 female rats reared in isolation from the age of twenty
days to the age of puberty. Nineteen individuals copulated within
a few seconds after being put with the males. He therefore con-
cludes that "the immediacy of the copulatory response in these
females would seem to justify the conclusion that no environmental
influences or factors beyond those necessary to insure normal
somatic development are required to bring about sexual maturity
as manifested by ability to perform the copulatory act during the
receptive stage of oestrus. " While these results show that the normal
response is possible after isolation, they do not disprove the pos-
sibility of the occurrence of an abnormal response when the normal
stimulus is not present.

Loutitt (7) made a study of the effects of isolation on the guinea
pig, using both receptive and non-receptive females as the sex
object ; that is, he put males with both receptive and non-receptive
females. He isolated males and females at 70 days of age and
tested them at 150 to 300 days of age. He does not state the number

THE SEX DRIVE

181

of animals which he used. His results indicate that the receptive
female is a more effective stimulus to sex activity than the non-
receptive female. The following characteristics in the behavior
of the male to a receptive female were observed : ' ' periods of quiet
are less and shorter, copulatory strokes are long and relatively slow,
the male succeeds in intromission,' he makes the first mount sooner
and subsequent mounts at shorter intervals." Although Loutitt
finds more activity in response to a receptive female, he states that
after a period of isolation mounting occurs in response to the
homosexual as well as to the normal stimulus. In these instances,
both males attempted to play the part of the male. Of course this
means very little since he does not state the length of the period of
isolation. Homosexual behavior among females, on the other hand,
he finds to be limited chiefly to a " short period before and during
the receptive stage .... She takes the part of aggressor." When
put with a male, the female exhibits ' ' the ordinary behavior of the
receptive female."

Cage studies on segregated groups show that conditions of segre-
gation lead to the development of forms of sex behavior which one
would naturally be led to predict on the basis of observation of
animals kept in isolation. In a study of the congenital sexual be-
havior of the young male albino rat, Stone (9) segregated fifteen
males from weaning till about the onset of puberty, or longer, six
of which were normal, the others having the function of certain
receptors destroyed by operation. In addition, eight animals were
isolated at 21 days. While he does not discusss these two groups
separately, he states in his general conclusions that ' ' on the whole,
the evidence brought forward indicates that the chief elements of
the primary reproductive act appear within a sort period of time
at puberty." Stone makes no attempt to test for the presence of
homosexual behavior. His investigations were limited entirely to
heterosexual stimuli.

The presence of homosexuality is discussed, however, by most
other investigators who employ methods of segregation. Loutitt
(7) states that when six male guinea pigs were kept together "at
different times the same pig might play the part of male or fe-
male . . . The male having the part of female did not submit will-

182

THE SEX DRIVE

ingly but usually ran or fought with his aggressor. ' ' He considers
that the "exhibition of mating behavior toward animals of the
same sex, at certain times or under certain conditions, is a normal
response." However, since he at no point states that he has used
other than isolated and segregated animals, this conclusion does not
seem to be justified by his data, for it appears that his conclusions
are not based upon observations of unsegregated animals — a con-
dition which probably most closely approximates the normal.

Homosexuality among segregated guinea pigs has been ob-
served by Avery (1). "Male guinea pigs," he says, "ofttimes
mount each other. This occurs most frequently, however, when
several males are kept together or when small young males are in-
troduced into the pens of sexually mature males." Like Loutitt
(7), he finds homosexuality among females chiefly at oestrum.
"Experimental observation directed to some 60 pigs over a period
of about three months showed that homosexual behavior occurred in
no less than ten to fifteen per cent of animals coming into heat
during that time. ' 9

Homosexuality has also been noted among pigeons. Several cases
are found among the scattered observations of Whitman (16, p. 9).
In one case, when two males were paired, one "attempted courting
during an entire month. He then began the course of incubation
on the floor of the cage." Whitman (16, pp. 98-99) has further
emphasized the importance of environmental conditions in the
development of abnormal sexual behavior. "Young birds raised
under foster-parents of a different species are very apt to prefer a
mating not with their own kind, but with a member of the species
among which they have been reared. ' ' Confinement of two animals
alone has a similar effect (16, pp. 99-100). "The continued confine-
ment and isolation of two birds will secure a mating when other-
wise it would not occur. Confinement alone is sometimes efficacious,
while at other times isolation from the sight and sound of other
birds is necessary. Even inter-species mating can be secured in this
manner. ' ' Whitman does not give the previous history of his ani-
mals nor the length of isolation or segregation period.

A few observers have noted that even unsegregated animals may
exhibit homosexual behavior (16, p. 35). Whitman noted an in-

THE SEX DRIVE

183

stance of this in the case of passenger-pigeons. "Two of my male
passenger-pigeons," he states, "mated with each other, notwith-
standing they were in a pen where there were several unmated
females desirous of mating. ' ' Such behavior has also been observed
by Hamilton and Kempf in the case of unsegregated monkeys.
Hamilton (3) states his conclusions in the following passage:
"Homosexual behavior is normally an expression of tendencies
which come to expression even when opportunities for heterosexual
intercourse are present. Sexually immature male monkeys appear
to be normally impelled toward homosexual behavior by sexual
hunger. The fact that homosexual tendencies come to less frequent
expression in the mature than in the immature male suggests the
possibility that in their native habitat these animals may wholly
abandon homosexual behavior ... on arriving at sexual maturity. ' 1
He finds that homosexuality among females is "rarely manifested in
response to sexual hunger." Kempf 's (5) findings are substantially
in agreement with Hamilton's although he had but six monkeys in
contrast with Hamilton's group of 28. Hamilton specifically states
that he used both segregated and unsegregated animals under vary-
ing conditions of confinement (18 situations in all).

Nothing on mammals beyond the simple method of cage study
was developed until Moss (8) introduced an experimental method
for the objective study of animal drives. He utilized the principle
of separating the test animal from the incentive by an electric grill,
the animal being required to cross the same in order to obtain access
to the incentive. The apparatus, technique, and results of Moss
have been criticized by Jenkins, Warner, and Warden (4), who,
in the same connection, describe an apparatus, which they have
named the Obstruction Apparatus. The technique for the measure-
ment of the sex drive has been described in detail by Warner (14)
whose study on the normal sex drive was the first of the series to
appear.

Warner was interested in the effect of sex deprivation on the
sexual behavior of the male white rat and in relating the various
stages of the oestrus cycle to the sexual behavior of the females.
His animals were raised to maturity (150 days) unsegregated and,
after a short period of segregation (35 days) necessary to his ex-

184

TEE SEX DRIVE

perimental technique (for details see his account), were tested by
the obstruction method. In his investigation Warner found that
long periods of sex deprivation seemed to weaken the normal sex
drive in the male — a fact which led him to suspect the development
of homosexual tendencies. However, Warner's procedure did not
enable him to determine definitely whether changes in the strength
of the normal sex drive are due to a transfer of the sex response to
the homosexual stimulus. Futhermore, since he did not employ both
sexes in the incentive compartment, he was unable to determine
whether the response to the normal sex stimulus is essentially
sexual, or primarily social in character.

The present investigation is an attempt to determine the effect
of sex segregation upon the strength of the sex drive, and the rela-
tive incentive value of various kinds of stimuli (males, and recep-
tive and non-receptive females). Animals were segregated before
and after puberty, in order to determine the importance of the dis-
integration of sex habits in the development of homosexuality2 and
its effect upon variations in the strength of the drive. To this end,
a mating period was provided for some groups one day before the
test, to ascertain whether an interval of sex activity will effectively
reinstate old sex habits or develop new ones, or whether, on the
contrary, it only will tend to intensify homosexual habits formed
as a result of sex segregation.

The present investigation involves two sets of essentially differ-
ent conditions. First, there is the variation in the conditions pre-
ceding the test period, which is represented by sex segregation and
mating periods. Second, there is the variation of the conditions in
the test period, which is achieved by using three types of incentive
animals — namely, males, and receptive and non-receptive females.
Varying the incentive stimuli in this manner enables one to deter-

2 By homosexuality is meant merely that certain behavior definitely sexual
in character is called out by an animal of the same sex, even though at the
same time similar behavior may be exhibited toward an animal of the opposite
sex. The term has not been employed in the extreme sense in which homosex-
uality is supposed to preclude all normal or heterosexual behavior. Perhaps, the
term homosexual tendency would be, in general, more exact than homosexuality
and we have used the former term exclusively in dealing with the male, since
here the tendency was somewhat less clearly marked than in the female.

THE SEX DRIVE

185

mine not only the strength but the relative incentive value of dif-
ferent types of stimuli. A determination of the incentive value of
various stimuli helps to account for the variations in the intensity
of the rsponse to the normal sex stimulus ; it also helps to ascertain
whether the drive is predominantly sexual or whether it is due pri-
marily to social or to other factors.

II. Method and Procedure

A. Age and care of the animals

The laboratory conditions which Warner has described in his
study of sex behavior were observed in this experiment (cf. Warner,
pp. 24-27), inasmuch as both are part of a comprehensive investi-
gation of animal drives as above mentioned. The * ' Experimental
Colony Strain" of albino rats of the Wistar Institute of Anatomy,
Philadelphia, was used throughout. The ages at which the animals
were tested were identical in both researches: that is, approxi-
mately 185 days, the range being from 175 to 196 days. The diet
consisted of McCollum's mixture, plus weekly rations of greens.
In the case of the unsegregated groups, an attempt was made to
achieve constancy of conditions by always keeping an equal number
of males and females in the same cage. This, of course, does not
insure an equal number of receptive females in each cage at all
times, but if a long period of time is considered, the average num-
ber of receptive females in a cage would be about the same.

B. The apparatus

The apparatus used in this investigation has been fully described
in a recent article by Jenkins, Warner, and Warden (4). The modi-
fied grill (see Warner, pp. 28-31), and an improved device de-
signed by Dr. T. N. Jenkins for the better manipulatory control
of the apparatus to be described later (see Figure 1), were used.

The general procedure as presented by Warner was followed and,
for the convenience of the reader, will be briefly outlined here. In
the preliminary training period the test animal was placed in the
entrance compartment and allowed to cross the grill five times to
the stimulus animal. In the first four trials no current was passing

186

THE SEX DRIVE

through the grill. On the fifth trial the electric shock was intro-
duced. The test animal was returnd as usual to the entrance com-
partment after the fourth trial. At the end of ten seconds the door
leading into the obstruction compartment was opened, as in the
first four trials. However, a change in procedure occurred at this
point as a result of the introduction of the electric shock for the
first time. When the animal entered the obstruction compartment,
the door was lowered behind him and the switch closed. In this
way, retreat to the entrance compartment was prevented and elec-
trical stimulation continued until the animal had crossed the grill
into the incentive compartment. As soon as the rat had traversed
the grill, the door to the incentive compartment was closed behind
the rat.

Since the function of the fifth trial is to permit the animal to
associate shock with the obstruction compartment, it is probably
more important to have conditions Uniform for all individuals in
this trial than in any other. The technique of closing the door be-
hind the animal and throwing in the switch should, of course, be
well standardized. Without this automatic device, the operation
of closing the two doors was affected by the use of one hand, the
other hand being required in throwing the switch which electrified
the grill. Much skill is necessary in order to perfect this movement
with even a fair degree of uniformity from one test to another.
It is very important that the switch and entrance door be simul-
taneously closed, since each animal should receive the electric
stimulation immediately upon leaving the entrance compartment.
Otherwise it would be quite possible to condition the animals un-
equally. That is, there would be more or less stimulation according
to whether the switch were closed before or after the door. Divided
attention on the part of the experimenter between observing the
animal and operating the various devices would tend to introduce
widely differing conditions of stimulation for different animals.
The automatic device (see Figure 1) successfully eliminates the ir-
regularities arising from the performance of three separate opera-
tions by modifying the manipulatory procedure so as to minimize
the importance of the personal equation in the conduct of the ex-
periment.

TEE SEX DRIVE

187

First, this device permitted the switch to be closed before the
fifth trial began. Second, simultaneity of the electrification of
the grill and the closure of the entrance door was assured, since the
closing of the door automatically turned on the current. The actual
procedure followed in the fifth trial in this investigation was as
follows: Switch #x is first closed. After the animal has remained

Fig. 1. Diagram of mechanism to facilitate the manipulatory control of the
obstruction box. S1 and S2, switches through which current is conducted from
the transformer T to the grill. D, manually operated door of the entrance com-
partment. I, I, contact springs. C, C, contacts. Z, Z, sponge rubber stops for
manually operated door. K, E, sponge rubber tips of the contact springs.

188

THE SEX DRIVE

in the entrance compartment for ten seconds, the door D is opened.
When open, the door rests against the stops Z, Z. In this position
the pressure of the door against the rubber tips K, K, breaks the
electrical circuit at contact C, C, so that no voltage is impressed
upon the grill. As soon as the animal is on the grill, the door is
immediately lowered, and the initiation of the movement closes the
circuit through contacts C, C. This prevents the rat from return-
ing to the entrance compartment. The reaction-time of the rat is
such that the animal does not have time to react to the current be-
fore the door is shut — thus precluding any possibility of return to
the entrance compartment. As soon as the rat has crossed the
grill, the door to the incentive compartment is closed. The method
of manipulating the doors is simplified by the fact that each door
is operated by a single hand, instead of operating both doors by
movements of one hand. As soon as the fifth trial is completed,
Switch $2 is also closed, so that the grill remains electrified without
interruption for the entire twenty-minute test period.

C. Detailed procedure

Variation in the intra-organic condition of the test animal was
obtained in part by varying the length of the period of segrega-
tion. Four such periods were used: o-days (unsegregated), 1-day,
35-days, and 155-days. The unsegregated group were of course
never segregated. The 1-day group was formed by giving one-
half of the unsegregated males a 2-hour interval of sex activity,
after which they were segregated for one day and then tested. The
35-day group were unsegregated from birth until 150 days, at
which time they were segregated for 35 days and tested at 185
days. The 155-day group were unsegregated for the first 30 days of
their lives. When they were 30 days of age they were weaned and
segregated, being kept segregated until tested at 185 days of age.

It is important to note here that all groups except the 155-day
group have had, up to the time of segregation, an opportunity to
copulate freely. Copulatory activity of the 155-day group was
precluded by segregating them at 30 days, since the earliest age at
which copulation has ever been observed is 37 days. Stone's work
is probably the most extensive study which has been done concern-

THE SEX DRIVE

189

ing the awakening of copulatory activity in the male albino rat. He
used 56 males. He finds (11) "the mean age at which copulatory
ability was first manifested was 48.06 days and the ages ranged
from 37 to 72 days."

Segregation not only prevented copulatory activity, but pre-
cluded the possibility of any tactual stimulation by members of
the opposite sex. No attempt was made to rule out olfactory stim-
ulation. Inasmuch as the males were placed in the top row of
cages and the females in the lower row, this tended to minimize
visual stimulation. The cages were wire mesh, resting on heavy
galvanized removable trays which extended out about two inches
on all sides.

An endeavor was made to secure constancy of temperature in
the laboratory, but seasonal changes could not be avoided. The un-
segregated, 1-day, and 35-day groups were tested between January
20 and May 15, the 155-day animals from June 8 to July
19. In spite of this, the temperature conditions did not vary
greatly. During the cool season the temperature was kept at ap-
proximately 75 degrees F. as detected by thermostat readings.
In summer the laboratory was quite cool as it is located in the
northeast corner of a modern brick building with cement floors.
Owing to artificial heat in the building, the range of variation in
humidity during the cool months may have varied somewhat from
that of the summer months.

The male test animal

1. Two-hour interval of sex activity. At the outset of the in-
vestigation it was intended to have at least 20 animals in each
group. However, in order to determine whether the effects of seg-
regation are lasting, it was decided to give half the animals in each
group a 2-hour interval of sex activity 24 hours before the test
period. The 2-hour interval was placed 24 hours before the test pe-
riod since Warner has shown (13) that this is approximately the
point at which recovery from complete satiation (2-hour mating
period) takes place. The writer had intended to combine the re-
sults of these two subgroups. It was found, however, that this dif-

190

THE SEX DRIVE

ference in conditions so changed the results for these groups that
it did not seem fair to combine them.

In this paper the data used include two groups of males which
were not tested by the writer — namely, Warner's 1-day and 28-day
groups. The 28-day group was the longest interval which he used,
but was considered sufficiently near to 35 days to be used here.
However, there is one very important difference between this
group and those used by the writer. Warner's 28-day group was
first segregated for 35 days, then given 2 hours of sex activity, and
again segregated for 28 days. The writer's groups, on the other
hand, were tested at the end of the 35-day period.

The 2-hour interval of sex activity requires further description.
Half of the males for each segregation period was caged for two
hours with an isolated receptive female. The males were chosen at
random from among the group which happened to be 185 days of
age at the time. Receptive females were chosen on the basis of be-
havior which is characteristic of oestrum, and the arousal of the
copulatory response in indicator males.

It will be well, at this point, to explain the function of indicator
males. These indicator males were selected because they showed
exceptionally strong sex behavior, especially in the matter of fre-
quent copulation. For example, when placed in a cage of females,
an indicator male would manifest a minimum of general explora-
tory behavior; almost immediately he would begin to nose the
genital region of the females and would copulate as soon as the
female displayed characteristic mating behavior, such as the short
run described by Stone (9). The use of indicator males for select-
ing receptive females for the 2-hour interval was found to have
one rather serious disadvantage. Male sex behavior, such as nosing
the genital region of the female, was often omitted. After a time,
varying from two or three days to one month, the experienced indi-
cator would no longer wait for the manifestation of the mating
behavior on the part of the receptive female. When placed in a
cage of females, he often mounted any female, and copulated. Con-
sequently, if the arousal of the copulatory response in an indicator
had been used exclusively as a criterion for the selection of recep-
tive females, non-receptive, and even pregnant females would some-

THE SEX DRIVE

191

times have been chosen. In order to overcome this difficulty, only
those females were used which had aroused the copulatory response
in at least two, and in the majority of cases three, indicator males.

To insure further the selection of a receptive female for the
2-hour interval, great care was used in observing their own active
mating behavior. As the investigation proceeded the writer became
more experienced in discriminating intensely active from fairly
active receptive females. Since the 155-day group were the last
tested, it is quite possible that the receptive females used in the
2-hour interval for the 155-day group were somewhat more intense
in their mating behavior than those used with the 1-day and 35-day
groups.

Having selected the receptive females to be used in the 2-hour
interval, they were each placed in separate wire mesh mating cages
(15" x 8" x 6"), the galvanized metal floor of which was covered
with a thin layer of sawdust. The females were cage-adapted for
at least 15 minutes before the male was admitted. The purpose of
this period was to allow the female to explore thoroughly the cage
so that she would display mating behavior as soon as the male was
admitted. After being cage-adapted, any new stimulus, such as a
male, tends to elicit exploratory behavior from the female.

These 2-hour intervals of sex activity usually occurred between
1 :30 a.m. and 3 :30 a.m., but a few were given as early as 11 :30 p.m.,
or as late as 4 :30 a.m. A close examination of the data showed that
there was no correlation between the hour of sex activity and the
number of copulations. Table 1 shows the distribution of the copu-
latory acts which occured in the 2-hour interval of sex activity.
The writer was careful to secure accurate data for this interval.

It will be noted that, of the 79 animals given the 2-hour oppor-
tunity for sex activity, 58 did not copulate (Table 1). This should
not be interpreted to mean that these animals did not manifest any
sex activity whatever. In many cases they persistently nosed the
genital region of the female and actively pursued her about the
cage. They would sometimes stop for a moment and lick the penis,
and then continue the pursuit. Often, this behavior very closely
resembled that of males which were about to copulate. These males
seemed on the verge of mounting and yet never actually mounted

192

THE SEX DRIVE

(although this described behavior lasted for some time). Mean-
while, the receptive female would usually display as intense mating
behavior as those females which actually elicited the copulatory
response. Indeed the receptive female would usually continue to
display active mating behavior for some time (perhaps five min-

TABLE 1

Distribution table covering number of mating s during the two-hour period prior
to segregation of males

NUMBER OF COPULATIONS

FREQUENCY

No copulatons

58

1— 5

1

6—10

3

11—15

3

lo — Z\J

2

21—25

1

26—30

4

31—35

0

36—40

0

41—45

0

46—50

2

51—55

2

56—60

1

61—65

0

66—70

0

71—75

0

76—80

0

81—85

0

86—90

1

91—95

1

79

Average for those which copulated

31.8

Average deviation

20.9

utes) after the male had relapsed again into a state of inactivity.
After this, the female would continue to display activity, but of the
random variety, as exploring the cage, washing her body, or nosing
the male genitals.

The writer was astonished by the lack of copulatory activity on
the part of the male in response to the aggressive mating behavior
of the female. Two explanations may be advanced for this lack of
activity. First, the male had not been cage-adapted before the
2-hour interval. Second, the mating behavior of the female may
have lacked the persistence and specific pattern necessary to elicit
copulation on the part of these males, although the females used in
the mating period were apparently quite normal. In order to deter-
mine whether such conditions were of any great importance, a

THE SEX DRIVE

193

group of fourteen sluggish males, which showed no tendency to
copulate, were again given an opportunity for sex activity after
they had been tested in the obstruction box the next day. These
males were given approximately one-half hour of cage-adaptation
before admitting the receptive female. Furthermore, an attempt
was made to elicit copulation from these males by placing them
with an exceptionally excitable and persistent receptive female. In
about half the cases the female used was one which was extremely
exciting, having at one time elicited 88 copulations from a single
male in two hours. Out of this group, only one copulated at all and
after two copulations this male was quiescent for the rest of the
period. The other males nosed the genitals for varying lengths of
time and sometimes pursued the female, but never attempted to
mount. The mating behavior of the female continued for some time
after the male had ceased to respond.

The disparity between these results and Warner's as regards the
number of males which did not copulate is somewhat puzzling.
Much of Warner 's work was done in a warm summer in a laboratory
which had southeastern exposure and was therefore considerably
warmer than the laboratory used in the present experiment. Wea-
ther conditions then may have been effective in causing a difference
in the copulatory activity of the males of the two groups.

A second explanation should be considered. Dr. Warner has
privately informed the writer that he only tested males which
copulated on the ground that only males which copulate are normal.
In this investigation, on the other hand, every male was tested
which had had a 2-hour opportunity for sex activity. The reason
for this difference in technique becomes apparent when the prob-
lems of the two investigators are considered. Warner, as he states
in his monograph, was studying the sex behavior of the "normal"
rat. The present investigation, on the other hand, was an attempt
to determine the effect of segregation upon all rats regardless of
whether they possessed the normal drive or not. Furthermore, we
might expect that segregation, since it is an abnormal condition
(that is, imposed by an external agency) would produce abnormal
behavior. One of the important problems under investigation,
therefore, was the development of abnormal behavior as the result

194

THE SEX DRIVE

of segregation. Obviously, then, it was impossible for the writer to
ignore the behavior of rats which displayed a deficency in the type
of behavior usually considered characteristic of the normal sex
drive.

It is interesting to note in this connection that the distribution
of copulations for the males which actually copulated is quite similar
to that obtained by Warner. Furthermore, the present investiga-
tion is not unique in the observation that all males are not sexually
aggressive. Avery (1) has clearly stated the same finding: "Ex-
tremes in sexual aggressiveness of adult males are readily observable
in breeding experiments wherein one has an opportunity to study
the behavior of many males. Some males will exhibit sexual aggres-
siveness day after day even though allowed to copulate almost daily,
whereas others are sluggish for a day or two after repeated copula-
tions with a female; still others are never sexually aggressive
although in good health and in other respects quite vigorous. These
points are of great importance in breeding experiments when an
assumption is made by the experimenter that the male will copulate
at the first opportunity given by the female. Such assumptions are
precarious at all times and should be made only, if at all, when the
male in question has been thoroughly and recently tested with
receptive females." (Italics by the writer.)

2. The test period. The majority of the tests in the present
investigation occurred between 11 :30 p.m. and 3 :30 a.m. However,
some animals were tested as early as 9 :00 p.m., and others as late as
5 :00 A.M., but the range of hours was approximately constant for
each group. The procedure of the test period has been discribed in
considerable detail by Warner (p. 32-39).

However, it was not possible to follow this technique in every
detail. It was necessary in the present work to quantify the tem-
poral aspect of the criterion for sex behavior, although Warner had
not done so. He states that "in the preliminary trials (the male)
was not returned until he had nosed the genital region of the
female, even though this occasionally involved a delay of several
minutes. ' ' It will be recalled that, since he was studying the normal
sex drive, he tested only animals which copulated in the 2-hour
interval of sex activity. If the animals which did not copulate

THE SEX DRIVE

195

had been tested, it is quite probable that he would have discovered
some extraordinary delays in nosing the genitals. It has previously
been stated that the writer tested all animals regardless of whether
they actually copulated or not during the 2-hour interval. Since
many animals were tested which did not copulate it can be under-
stood why the writer discovered in the preliminary experiment that
animals very often failed to nose the genitals of a female stimulus
even after a protracted period of delay. In one case a delay of 30
minutes was insufficient time for a stimulus animal to elicit a nosing
response from the male test animal. For practical reasons a time
limit was set for the period of delay in nosing the genitals. One
reason for setting a definite limit upon the time spent by the test
animal in the incentive compartment was to make it possible for
the experimenter actually to carry out the investigation, since all
animals must be tested between the age limits of 175 and 196 days.
Furthermore, a technique should be broad enough to include all
animals in the group. Finally, it is necessary to quantify the tem-
poral aspect of the criterion in order that all tests might be compar-
able. A limit of three minutes in the incentive compartment was
allowed for the nosing of the genitals. This interval of time was
selected because, in the preliminary investigation, it was found
that an animal which failed to nose in three minutes was not very
likely to nose, even if given a much longer time.

In the actual test (20 minutes) Warner broadens his criterion to
include 1 1 any specific contact between the two animals. ' ' It should
be noted, however, that he always used a normal sex object as a
stimulus animal whereas in the present investigation the incentive
animal in some cases was an animal of the same sex. It was found
that males which show marked homosexuality repeatedly nose the
genital region of one another, whereas males which are definitely
heterosexual do not continue such behavior long. The criterion in
the preliminary period (nosing the genitals) was also employed in
the test. A time limitation has the additional advantage of quanti-
fying the temporal aspects of the criterion. A limit of 30 seconds
was employed largely on the basis of preliminary work regarding
the length of time at which nosing occurs. If neither animal nosed
the genital region of the other within this interval of time, the test

196

THE SEX DRIVE

animal was removed and again placed in the entrance compartment.

The female test animal

With the exception of the 1-day series, the same segregation
groups were used as in the case of the males — namely, unsegre-
gated (0-day), 35-day and 155-day groups. It was found imprac-
tical to test a 1-day group of females because the oestrus cycle is
such that it would involve an excessive waste of animals, for a
female if given an interval of sex activity when in oestrum would
usually not be in oestrum at the time of the test 24 hours later as the
conditions of this group required.

The data presented in this study include only test animals which
were in the cornified stage of oestrum. The criteria for the cornified
stage were satisfied by means of the vaginal smear. Since the
vaginal smear could not be taken until after the test (cf. Warner,
p. 37), the female to be tested could not be chosen by this method.
On the contrary, the experimenter relied on behavior criteria to
determine whether the female was probably in the cornified stage
of oestrum, in order not to waste an excessive number of females.
With the exception of the unsegregated group, the behavior of the
females toward indicator males could not be used as a criterion,
since any association with a male would amount to interpolating an
unsegregated interval into the segregation period. The only feasible
method then was to choose test females on the basis of their general
behavior. The writer would sometimes spend an entire hour in
observing a single female to determine whether she was in oestrum.
Various cues were found helpful in deciding which females were in
oestrum, such as the gait of the animal, the intensity of her activity,
jerky walk, nervousness when approached, etc. Under such cir-
cumstances, it was inevitable that an occasional female would be
tested, and later found not to be in the cornified stage as indicated
by the vaginal smear. All such animals were eliminated entirely.

1. Copulatory interval. On account of the short duration of the
receptive period in the female, it was impossible to test animals
which had had a 2-hour interval of sex activity 24 hours before the
test period. Consequently, no female groups could be tested which
were exactly comparable to the male groups which were given a

THE SEX DRIVE

197

2-hour opportunity for sex activity. However, in the ease of the
155-day group, an attempt was made to introduce an interval of
sex activity before the test period. Since this group was segregated
before puberty, there had never been any opportunity for the
female to associate the male with the sex response. Consequently,
the male would be a novel, but not necessarily a sex, stimulus for
the female. In addition to a group which had had no sex activity, it
was thought desirable to test also a group which had had an oppor-
tunity to become familiar with the male as a sex stimulus. When
this group was formed, the intention was to combine the results
of this group with the regular 155-day group. However, it was
found that this interval of sex activity had such a profound effect
upon the behavior of the females during the test period, that it was
deemed inadvisable to combine the data for these two groups.

The procedure of giving the females an interval of sex activity
prior to the test period is, in some respects, similar to that used in
the case of the male groups. The female was placed in the regular
mating cage and allowed at least fifteen minutes for cage-adaptation.
Sex activity for each female was provided by one of two indicator
males. In order to equalize as far as possible the opportunities for
sex activity for all members of the group, provision for copulatory
activity was restricted to the alternate use of the two indicator
males.

Since this interval was given in order that the female might asso-
ciate the male with the copulatory response, it became essential for
actual copulation to take place. Five copulations were allowed each
female. After the five copulations had taken place the indicator
was removed and the female left undisturbed for one hour at the
end of which she was removed and tested in the usual manner.

2. The test period. The conditions of testing the females were
precisely the same as for the males. Immediately following the
test period a vaginal smear was taken and a permanent slide made.
These were later examined by Dr. Warner and are on file in the
Animal Laboratory. The records of all animals whose smears did
not definitely show the cornified stage were excluded from the data
herein presented.

198

THE SEX DRIVE

The incentive animals

For both the male and female groups, three types of stimulus
animals were used in the present investigation — namely, male,
receptive female, and non-receptive female. Throughout this paper,
the term receptive is used interchangeably for female in the corni-
fied stage, female in heat, and female in oestrum; the term non-
receptive is used to designate a female in the recuperative stage of
dicestrum. These stages were determined on the basis of mating
behavior and the usual vaginal smear.

In order to keep the male stimulus as constant as possible, only
three male stimulus animals were used throughout the entire inves-
tigation. These animals will be designated as No. 1, No. 2, and
No. 3. In the case of the female groups, No. 3 was always used as
the stimulus. This animal was selected because it possessed a very
stereotyped form of behavior, and therefore the advantage of being
a singularly constant stimulus. When the door of the restraining
compartment opened, he would come out at once, and immediately
approach the female and nose the genital region. As soon as the
experimenter removed the female from the incentive compartment,
he would return to the restraining compartment and remain there
while the door was being closed.

The function of the incentive animal is to stimulate the test
animal. In order to carry out this function to the greatest possible
extent, the incentive animal was always allowed a period of adapta-
tion in the obstruction box before the test period began. The object
of introducing this period of adaptation was to reduce exploratory
behavior to a minimum and hence increase its value as a stimulus.
The incentive animal was placed in the box at least one-half hour
before the test animal was admitted. The restraining door was
kept open in order to permit the incentive animal to explore freely
the incentive compartment.

The experience of testing over 300 rats has thoroughly convinced
the writer that a period of adaptation to the apparatus itself (no
incentive animal present) should also be allowed in the case of the
test animal, in order to minimize the importance of exploratory
behavior during the test period. The primary function of the pre-
liminary trials is to permit the test animal to associate the stimulus

THE SEX DRIVE

199

with the incentive compartment. An attempt should be made to
rule out exploratory behavior before the experiment begins, so that
when the incentive animal is introduced, it will immediately become
the dominant stimulus in the situation. The desirability of eliminat-
ing as far as possible exploratory behavior in the test period is
supported by the following reasons : (1) Exploratory behavior prob-
ably tends to increase the dispersion around the averages — a fact
which would decrease the reliability and introduce an element of
heterogeneity into the results; (2) Exploratory behavior is likely
to complicate the situation and render an analysis of the causes
underlying variations in the strength of the sex drive less easy or
valid; (3) In the case of small differences, the exploratory factor
may even shift the relative positions of a set of averages to such an
extent as to minimize the significance of the trends indicated on the
face of the data.

The question might be raised as to the manner in which the three
types of incentive animals could operate as differential incentives
under our conditions. For, after all, the different types (receptive
female, non-receptive female, and male) appear in general very
much alike. From what we know of the visual capacity of the white
rat it is hardly reasonable to suppose that one type could be dis-
tinguished from another on the basis of visual cues. It is true that
the males were quite noticeably larger to the human eye than either
type of female (receptive and non-receptive approximately the same
size), but there is little or no reason to supposee that this difference
served as a cue. Differences in odor seem to offer a more reasonable
basis of explanation. Long and Evans (6) state that during the
cornified stage of the oestrus cycle ' 'there is usually a disagreeable
odor attached to the vaginal secretion, which perhaps is a means of
attracting and exciting the male." Also Watson (15) found that in
the maze "adult rats showed preferences for entrances that con-
tained the odor of the opposite sex." The obstruction apparatus is
constructed so as to favor the passage of odor through the tunnel to
the test animal — both incentive compartments (C, D) being closed
by a glass cover during the test, and the door between perforated
so as to permit the odor of the incentive animal to pass freely
through. However, we have no evidence of differences in odor

200

THE SEX DRIVE

between male and non-receptive female sufficiently marked to serve
as a cue. There may or may not be such. Whether, then, differences
in odor between the three types of incentive animals operating
directly upon the test animal through the 10-inch tunnel was an
important factor in determining the nature of the response, cannot
be certainly known from our results. We were interested primarily
not in an analysis of the incentive, animal as a stimulus (although
this is a problem of importance in itself) but in the behavior of the
test animal to these three types of incentive stimuli, assuming for
the moment their differential character.

The specific manner in which the incentive operated as an excitant
becomes a problem in this connection only when the fact is taken
into account that the test animal is in the entrance compartment at
the moment of response and is thus separated from the incentive
animal. This does not mean, however, that we must suppose that
the response of the test animal to the incentive must be based upon
necessarily internal sensory cues operating at the moment. It is
much more reasonable to believe that during the five preliminary
crossings to the incentive (each animal was tested but once and to
only one of the three types of incentive animals), the incentive
compartment and the specific type of animal contained therein
became associated with the act of crossing over. The incentive-drive
value of the incentive animal is thus operating as an excitant to the
test animal before the test proper is begun and continues through-
out the test as the animal crosses over and comes in direct contact
with the incentive animal again and again. The excitation so
aroused by the incentive animal is thus cumulative, the nature and
amount of the excitation depending upon the type of incentive
animal used in testing any given animal or group. Under our con-
ditions no animal was required to discriminate between incentive
types in making a response since only a single type was present
during a given test and for a given animal.

The writer believes that it is quite possible to furnish a still
simpler interpretation of differences in incentive power of the
various types of stimulus animals (as indicated by number of
crossings, etc.). It is quite reasonable to suppose that the prelimi-

THE SEX DRIVE

201

nary trials, as well as those that follow, cause changes in internal
stimulation — an after-effect consequent to stimulation in the incen-
tive compartment. An incentive which calls forth sex behavior on
the part of the test animal probably sets up internal activity which
continues or perseverates for a time after the animal has been
returned to the entrance compartment. The effect of this increased
excitement is greater activity. This would tend to elicit the crossing
response, even if no association were formed between the incentive
compartment and crossing over, because the obstruction apparatus
is so constructed that crossing over to the incentive compartment
is one of the simplest and most natural responses for the animal to
make.

III. Results for Males

The main results for males will be presented in four general
tables. Table 2 indicates the grouping of the 197 male rats tested.
The distribution of approaches, contacts, and crossings for the
eighteen groups is given in Table 3. The medians, ranges, averages,
mean deviations, coefficients of variation, and temporal medians are
shown in Table 6. The approaches, contacts, and crossings as dis-
tributed during the 20-minute test (in one-minute intervals) are
given in Table 4. To these general tables, special tables will be
added when necessary.

The four tables mentioned above give an inadequate picture of
the situation. The many interrelated problems involved require
the designation of the various trends and their subsequent evalua-
tion through an analysis of the data by means of intercomparison
and special statistical and graphic methods. Such an analysis will
be made by treating the results under four headings :

A. Variations in the strength of the drive and the relative incen-
tive values of the stimuli. This topic involves a study of the effects
produced by two sets of variables. First, there are the changes in
the strength of the drive which result from varying the conditions
(segregation periods) preceding the test. Secondly, there are the
effects produced by varying the conditions (incentive stimuli) in
the test, due to the use of the three types of stimuli for each period
of segregation. Varying the incentive stimulus, as was done in this

202

THE SEX DRIVE

experiment, furnishes data which are of great importance in the
explanation of the drive for various segregation periods.

B. Individual differences.

C. Persistence of the drive as shown by the temporal distribution
of drive behavior during the test.

TABLE 2

Showing male groups

Number

of Period of sex deprivation Object in incentive compartment
animals

13

0 hours

Female —

cornified stage*

10

Female in
staget

dioestrum — recuperative

10

n

Male

15

0 hours plus 2 hours sex ac-

Female —

cornified stage

tivity plus 1 day deprivation

dioestrum — recuperative

11

Female in
stage

10

»

Male

20

28 days:}:

Female —

cornified stage

10

35 days

Female in
stage

dioestrum — recuperative

10

»>

Male

20

35 days plus 2 hours sex ac-

Female —

cornified stage

tivity plus 1 day deprivation^

dioestrum — recuperative

10

Female in
stage

10

»

Male

5

155 days
»>

Female —

cornified stage

10

Female in
stage

dioestrum — recuperative

10

»>

Male

5

155 days plus 2 hours sex ac-

Female —

cornified stage

tivity plus 1 day deprivation

Female in
stage

dioestrum — recuperative

8

10

Male

♦Throughout this paper the writer has used "female in cornified stage"
and "receptive female" interchangeably.

tThroughout this paper the writer has used "female in dioestrum" and
"non-receptive female" interchangeably.

$These groups were tested by L. H. Warner and are reported in his
monograph on sex behavior.

D. The effect of the mating period upon the test 24 hours later.

The standard measures dealt with in these topical discussions are
average number of crossings, indices of the strength and incentive
value of the drive, variability, and persistence. In view of the fact
that frequent reference will be made to these measures in different

T3i

8
to

I

Co

<

w
<

2
O
h
<
>

ft,

w
Q

M
<3

lis

£8

to «
•O

( -7S

» i s &J

J e 3

J? 5 ^ .

s ^ >

- 2 2 «0

&3g

is 2 «<
£8"

: 11

° is

p 2

sSuissoj^

«•» m- — to to —

S10EJUO~

vOvO<MCMCMCM

to to — — c\J

O

CM

s8uissoj3

CM CO — — — M

SJDBJUO^

^ TO CM —

satpEOiddy

r»» cm —

s3urssoJ3

— cm cm cm —

— —

SIDElUCr;

cmto — cm cm

saipeoaddy

vO — — — —

s3uissoj3

TOTOCMCM — TO— .—

— —

to\OCMt!-_ —

satpEOJddy

GO WO TO CM

s8urssoJ3

— CM

SJDEJUCV}

— TO — CM

00

saipEOJddy

s3uissoj3

CM r_ <VJ

—

CMTO — CM —

saipEoaddy

WO CM CM —

— — txj cm CM

__

SIDE JU03

to— wo _

saqDeoiddy

OOM

sSuissojq

— — CM

SJDEIUCV}

CM — CM

saipecuddy

s3uissoj3

— — — —

SJOEJUO^

— TO —

wo

saipEOJddy

•<*• —

s3uissoj3

SJDEJU03

CM "^l- CM —

0

saipEcuddy

OO "> — CM CM — —

CM

s3uissoj3

— — — CM — CM

SJDEJUCT}

Tl-^CM —

O

saipEOJddy

M" <M — _

s3uissoj3

SJDEIUCQ

00 *M

saqoBOiddy

-»J- TO CM —

O

s3uissoj3

— CM — <0 — — CM —

— CM — CM CM

SOOBIUOQ

CM CM —

O

saijDEOJddy

^"""CM-^

CM

s3u;ssoj3

— — — _

— CM —CM

SIOBIUCQ

00 <M

O

saipBOJddy

■o«~> —

s3uissoj3

saipBOJddy
s3uissoj3

S10BJUO3

sat[DBOiddy

S3U!SS0J3
SJDEJU03

saipecuddy
s3uissoj3

SJ0BJUO3

saijOBOJddy
s3uissoj3

SJDEJU03
S3L[DBOiddy

s3uissoj3

S1DE1U03

upECuddy
s3uissoj3

SJ0B1UO3 Z

satpBOiddy £J
s3uissoj3

SJDBJU03

saijDEOiddy
s3uissoj3

SJ0BJUO3

satpEOiddy
s3uissoj3

S13EJU03

saqDBoaddy
s3uissoj3

SJ0BJUO3

saipBcuddy
s3uissoj3

SJDBJU03

soqDBoiddy

TO CM — "f
^VOtJ-CM —

CM CM — „ ,
VO^-Tj- — <m,
NO 10 — CM to

— — CM TO'

-CM Tf- TO — CM

1 CM CM m 1

TO CM — —

— CM CM — ■
•fCM —
NO CM — —

\OCM — —

s z

I 0
* 2

to

CM — CM '

> CM CM —

CMVOCM

TO CM — TO CM

• — CM — CM

— -«*-CM

CM — TO— 1

" TO TO TO —

CM -<f vr» CM — CM

CM —
CM —
— CM

O — Ni

ivOr^ooo^O — CMto^-ionci

■ 00 0*0 — CM
. — — CM CM CM 1

a.

204

THE SEX DRIVE

TABLE 4

Showing temporal distribution of approaches, contacts, and crossings of the male
groups during the twenty-minute test period in one-minute intervals

<v

o

u

V

aj

V

u

u

ti

U

U

IV

o

<u

c

0

1 1

3

a

3
C

3
3

3

c

3

c

3
3

3
C

3
3

3
C

3
3

3
3

3
3

3
3

3
3

3
3

a

3
3

3
3

3
C

3
3

I 11

° gs

o «

' >

■|

g

'1

g

g

g

g

g

1

g

1

'1

1

6

e

C

E

e

'1

a w

<

•o
c

TJ

.3

J3

_3

-3

-3

-3

.3

-3

—

-3

JS

-3

-3

JS

JS

-3

Q «J

CM

>o

oo

O

r»

oo

O
CJ

Female —

cornified

stage

0 days

Approaches

Contacts

Crossings

0
0

1 1
1 1

2
4

2
I

a

O

1
1

7

1

r

•»
j

2
0

2
0
o

2
3

1

2

1

0

C

0
0

3

2
0

7

3
0
-1

3
3

«r

6
3

c
O

4

6

5
3

6
3
4

3
1

4

5

2

_

1 dav

Approaches
Contacts
Crossi ngs

1

0
22

0
1

1 6

0
0
1 5

1
1

1 2

0

i

1 1

2
0
] 4

!

2
1 0

0
0
1 3

1
1

1 2

0
0
5

0
2

1

0
9

0
0

g

2
1

g

2
0

Q

o

0
0
q

0
I

c
3

l

0

E
3

0
0

c
5

1

0

i

•r

-a

S Female in
SfdicEstrum —

0 days

Approaches

Contacts

Crossings

0
2
g

0
0

2
0

0
0
5

0
0

5

0
1

5

0

1

0
0
2

0
0
2

0

1

2

0
0

1
J

0
0

1
3

0
0

0
0
c

0

1

0
0

0
0
i.

0

l

&

~cf

0

1
J

0

->

Sirecupera-
£ tive stage

1 day

Approaches

Contacts

Crossings

1

1
1 i

1

0
1 i

1

2
1 1

1

0
6

()

0
7

1

0
g

1

0
4

1

0
g

1

2
5

0
1

5

l

4

3

l

2

0
0
5

1

1

■j

0
0
4

l
l

4

l

0
4

0
0

5

0
0

-2

J

0
0

5

0 days

Approaches

Contacts

Crossings

0
1
7

0
0
7

2
1

g

1

0
1

0
0

5

0
0
2

1

0

0
0
2

0
0
5

0
0
1

0
0

0
l
4

0
0

■J
J

0
0

■J

3

0
0
4

0
0
J

0
0

4
0

3

0
0

0
0

0

Male

1 day

Approaches

Contacts

Crossings

2
1

9

0
0
g

0
1

7

0
0

3

0
1

1
1

4

1

3
1

1
1

2

1
1

3

0

1

4

0
0

I

0
0
4

0

1

]

0
2
2

0

1

1

0
0
2

0
4

o

1

2
2

0
2
1
1

0
0
1

Female —

cornified

stage

Approaches
28 days Contacts
Crossings

4
3
2

j

0
7

0
1

3

3
0
1

0
0
2

0
1
4

1

2
5

2

0
7

4
2
1 1

1

4
1 3

I

4

7

l

6
I j

3
5
j j

6
4
9

4

3

1 9
1 c

s

4
2 I

2
0
]9

4
0

14

4
1

22

3
0
19

1 day

Approaches

Contacts

Crossings

7
3
1 2

2
1

17

1
I

l J

3
2
1 6

6
2
1 9

6
3
17

5
3
1 5

1
1

9

2
1
1 |

2
1

1 0

l

2
10

l

l

l 3

2
0
1 6

0

3

1 J

2
2

i n

l

2

0
0
l 2

1
1

1 7

2
1

1 r
i ->

I

0
1 7

-o

^ Female in
^dicEstrum —

35 days

Appioaches

Contacts

Crossings

2
0
7

0
1
g

o
0

0
0
4

2
2
3

2
0
4

0
0
5

0
0

3
1

5

2
0
1

0

0
5

2
1

■»

3

0
0

2

0

1

2

1
1

2

0
0
2

0
l

2

I

0

3

1

2

0
0
2

"2 recupera-
tive stage

00

5)

1 day

Approaches

Contacts

Crossings

1

0
9

0
0
g

0
0
g

0
0
5

0
0
5

0
0
7

0
n

q

1

1
4

0
0
3

0
0
2

I

0
7

o

0

^

0
0
4

0
0
3

2
0

0
I

5

0
0
2

2
0
3

0
0
3

1

0
5

u

35 days

Approaches

Contacts

Crossings

4
0
7

1
1

4

0
0
5

2
0
4

I

0
5

0
0
I

0

0
3

0
0
4

0
0

3

0

1

2

0
0
2

2
0
2

2
0
5

0
0
2

2
0
2

I

0
2

I

0

3

(1
0

3

0

0

3

0
0
1

Male

1 day

Approaches

Contacts

Crossings

2
0
1 1

0
1

7

0
1

g

1
4

3

1

2
9

n
3
7

2
0
7

2
2
4

3
2
3

2
2
2

2
0
2

1
1

3

3
1
4

1

0

g

1
1

5

I

2
4

0
0

0

n
3

0
0
4

2
3
t;

Female —

cornified
„ stage
>.

•3

155 days

Approaches

Contacts

Crossings

(l
0
c

J

0
0

3

0
0
4

0
0
5

0
0
2

(1

0
2

0
2
n

0
0
3

0
1
1

0
0

o

0
0
4

0
1

3

0
0
I

0

1

o

0
1

q

I

0

2

0
2

j[

0
1
j

0
0
1

0
0

0

1 day

Approaches

Contacts

Crossings

0

1

5

0
0
4

0
I

0
0
1

1
1

2

Q
0

3

0
n

4

0
0

3

0
0

1

0
0

3

0
0

l

0
0

3

0
0
1

0

1
1

0
0

3

0
0
2

0
l

0

0
0
1

0
0

3

0
0

0

^1 Female in
diozstrum —

155 days

Approaches

Contacts

Crossings

0
2
6

0
0
7

0
0

6

1

0
4

0
0
4

0
2
5

0

1

4

2
1
4

0

1

3

1

2
1

I

2
2

0
2
1

1

0
2

0

1

3

0

1

3

0
0

l

I
I

2

0
0
2

0
4
I

2
6
2

-% recupera-
te tive stage
to

V

1 day

Approaches

Contacts

Crossings

2
1

6

0

1

3

0
2
2

1

0
4

1
1

2

it
2
4

1

0
3

(l
0
3

Q
1

2

(1
2
0

0
0

l

0
0
1

0
0
3

1
1
1

0
1
0

0
0
2

l

2
0

0
3
1

0
0
2

1
1
1

u
6
oo

155 days

Approaches

Contacts

Crossings

0
0

9

l

2
6

0
3
5

n
i

6

1
1

s

0
2
4

n

0
6

2
(1
2

0
0
4

1

3
3

2
0

6

0
0
3

1

4

5

0

1

3

0
0
4

I

0
2

0
l
l

0
2
2

1

3

3

3

Male

1 day

Approaches

Contacts

Crossings

2
3
6

0
1
4

0
1

4

0
o
5

0
0
2

0
0
2

0
0

1

0
0
1

0
0

1

1
1

2

0
0
l

0
1

3

0
1
2

1
!

2

0
0
0

0
0
3

0
0
0

0
0

!

I

0
4

0
1

2

THE SEX DRIVE

205

TABLE 5

Showing temporal distribution of approaches, contacts, and crossings of the
male groups during the twenty -minute test period in five-minute intervals

I IJ

> ° - R

U81

0 to 5th
Minute

Jum- Per
ber Cent

6th to 1 0th
Minute

Num-
ber

Per
Cent

Uth to ISth
Minute
A

Num- Per
ber Cent

16th to 20th
Minute —

Num- Per £
ber Cent

Female —
cornified
stage

Approaches
days Contacts
Crossings

6 11.7

7 21.2
36 35.6

15.7
15.1
23.7

14 27.4
6 18.1
23 22.8

23 45.1 51
15 45.4 33
18 17.8 101

Approaches
day Contacts
Crossings

2 15.4

3 30.0
76 37.6

4 30.8
3 30.0
54 26.7

5 38.4
3 30.0
44 21.8

2 15.4
1 10.0
28 13.8

13
10

202

JJ Female in 0

Jf dicEstrum —

at) recuperative
Sj stage ]

3

Approaches
days Contacts
Crossings

2 66.6
2 28.6
31 40.3

0 0
3 42.
16 20.

0 0

1 14.3
18 23.4

1 33.3
1 14.3
12 15.6

Approaches
day Contacts
Crossings

4 30.1
3 21.4
46 38.3

4 30.1
3 21.4

32 26.6

3 23.1
7 50.0
21 17.5

2 15.4
1 7.0
21 17.5

13
14
120

Male

Approaches
days Contacts
Crossings

3 37.5
2 66.6
26 32.1

1 12.5
0 0
17 21.0

0 0

1 33.3
20 24.7

4 50.0
0 0
18 22.2

Approaches
day Contacts
Crossings

2 28.6

3 13.6
33 53.2

4 57.1
7 31.8
14 22.6

0 0
4 18.2
9 14.5

1 14.3

8 36.5
6 9.7

Female —
cornified
stage

Approaches
days Contacts
Crossings

9 18.0
4 10.0
16 7.6

8 16.0

9 22.0
40 18.9

15 30.0
22 55.0
54 25.6

18 36.0
5 12.5
101 47.9

Approaches
day Contacts
Crossings

19 41.8
9 29.9
76 28.2

16 35.2
9 29.9
63 23.4

6 13.2
8 26.6
60 22.3

5 11.0
4 13.3
70 26.0

7
22
62
50
40
211

46
30
269

£ Female in 35

dicEstru m —

*]3 recuperative

I "age 1
u

Approaches
days Contacts
Crossings

4 25.0
3 30.0
28 38.9

7 43.7
1 10.0
21 29.2

3 18.7
3 30.0
14 19.4

2 12.5

3 30.0
9 12.5

Approaches
day Contacts

Crossings

1 12.5
0 0
35 30 1

1 12.5
1 50.0
25 21.5

3 37.5
0 0
26 22.4

3 37.5
1 50.0
18 15.5

16
10
72
8
2
104

35

Male

Approaches
days Contacts
Crossings

8 50.0
1 50.0
26 40 6

0 0

1 so.o

13 20.3

•6 37.5
0 0
13 20.3

2 12.5
0 0
12 18.7

Approaches
day Contacts
Crossings

4 16.7
8 32 0

38 36.8

9 37.5
9 36.0
23 22.3

8 33.3
3 12.0
20 19.4

3 12.5
5 20.0
22 21.3

16

2
64
24
25
_103
1
9
38

Female —
cornified
stage

155
1

Approaches
days Contacts
Crossines

0 0
0 0
19 50.0

0 0
3 33.3
6 15.8

0 0
3 33.3
8 21.0

1 100.0
3 33.3
5 13.2

Approaches
day Contacts
Crossings

1 100.0
3 60.0
15 32.6

0
0

34.8

0 0

1 20.0
9 19.6

0 0

1 20.0
6 13.0

Female in 155

dicestrum —

recuperative
stage i

Approaches
days Contacts
Crossings

1 11.1

2 7.7
27 42.8

3 33.3
7 26.9
17 27.0

2 22.2
6 23.1
11 17.5

3 33.3
II 42.3
8 12.7

Approaches
day Contacts
Crossings

4 50.0

5 27.8
17 41.5

1 12.5
5 27.8
12 29.3

1 12.5

2 11.1

6 14.6

2 25.0
6 33.3
6 14.6

155

Male

Approaches
days Contacts
Crossings

2 18.1
7 29.2
31 37.8

3 27.3
5 20.8
19 23.2

3 27.3
5 20.8
21 25.6

3 27.3
7 29.2
11 13.4

Approaches
day Contacts
Crossings

2 28.6
5 .50.0
21 45.6

1 14.3
1 10.0
7 15.2

3 42.8
3 30.0
8 17.4

1 14.3
1 10.0
10 21.7

206

THE SEX DRIVE

pq o

JUBipSUJ

jBJOdoiaj^

O; —

1#>CM

oo ■»»■'

14.9

9.6

OO CM

vbf-'

ui O
t^sb

© rn

veil*;

in m

oo

r^sb

uodbuba

r- cm m

OO^f^

•* — m

©vo-
odoo
mvom

71.6
29.9

59.1
39.7
52.7

rnt^f*
O^NO

in© cd

CM CM CM

mm oo
— — in
vomm

woo*

rn©ob
enr^-*

Crossings

A

uoueiAdp
ue3jaj

oo —

OOinOO

©cnoo
t*»— e\i

m© *n

m cn en

— cm —

od vdr-^

.10.6 6.1
13.4 3.3

tncncn

CM -tOO

*nom
in m

<tt»in

vodoo

VO CM •«*-

r«.'doo

— — in

en CM CM

m — sO

so in in

CM sOm
CM CM CM

CM SO ^

oo io

0—19
7—23
0—23

TTT

OOCMOO

TTT

— CM —

0—22
4—20

0—14
2—15
0—15

0—15
5—24
0—24

TTT

mr*»m

so — s©

III

4—14

0—8

0—14

^ UBipajAJ

in — ©
r-'^cM

tsl —

©m©

12.5
15.0

f»'cMC>

©Om
uifjitN

in cm in

r«^c>oo'

m © in
so in in

qqq

co in t-^

UOIJBUBA

105.2
161.4
142.0

142.8
125.3
125.0

167.0
130.9
145.8

100.0
100.0

100.0
190.0
146.6

188.0
140.4
163.8

48.8
1000
80.7

101.1

85.4
94.1

94.1
100.0
102.9

UO|JBIA9p

UBaj^

— om-^
cm —

00 m©

CM —

■O CM

oo mr-

MOON
CM —

— moo

oooo*

Contacts

»3BJ3Ay

in in
CM ' —

mCM CM

"cm —

©in
CM —

qcM\o

CM —

OO©'*

sO CM •»*;
CM CM CM

CM — —

»8ub^

III

ill

m

2!

222

0—1

0—14

0—14

mCM cn

111

0—11

0—7

0—11

m enm

ill

UBipaj^

CN» ' —

OONO

WJ©

n\coo

so in Co

mm cm
CM — CM

o©o

CM CM CM

m*n\q
cm — —

r uoneiJBA

>■

ju9iDigao3

109.2
132.9
130.4

210.0
104.1
141.2

cm —

r-» CMsO

116.0
91.3

93.7
150.0
125.0

125.0
130.4
131.5

©©©

odd

O- O O

m©m
in oo in

CM CO CM

148.1
178.5
166.6

UOUBIASp
UB3J\[

O
cn cm

inOO*

— r^o

CM —

CMsO©

SO in —
— CM CM

m m tn

m © CM

>roaches

33BJ3Ay

m
Oaocn
m cm

m CM OO

oof- r-»

tn «n
CM CM

— CM CM

CM CM CM

OOOs

— f>0^

<

38ub>i

0—17

0—4

0—17

HI

in en in

ill

OO

22

ill

U1CCOO

111

111

222

mmm

111

UBipSJ^

t^C>m
CM —

sCnO

©in

O CO CM
CM —

mw-«-

VO vO SO

© m CM

oor^oo

ibuiiub is»j

*JO

UOUEAUd^Q

0 hours

1 day
Total

0 hours

1 day
Total

0 hours

1 day
Total

28 days
1 day

35 days
1 day
Total

3 5 days
1 day
Total

155 days
1 day
Total

155 days
1 day
Total

155 days
1 day
Total

JEU1IOB

f Female—
cornified
stage

Female in
diozstrum —
recuperative
stage

f Female—*
cornified
stage

Female in
diozstrum
recuperative
stage

"re

s

Female —
cornified
stage

Female in
' diozstrum —
I recuperative
stage

J£
"re

Y

p»lc8aj8»sof)

V

s/Bp 5£ pairSajSag

Y

sXEp JSI P3Jc89J3as

2i

o S
S 2

£-5

5^

6o
— a.
2 E

o a

a

£ «>

THE SEX DRIVE

207

parts of this paper, a little space will be given at this point to their
definition and elucidation.

1. Discussion of the number of crossings will involve two kinds
of averages which will be designated respectively as 1 ' total average ' '
and "partial average." The former is the average number of cross-
ings in the entire 20-minute test; the latter includes only the
crossings that occurred during the last fifteen minutes of the test.
One object for introducing the partial average is to furnish an
additional check upon the trends indicated by the total averages.
The primary reason, however, is that it tends to eliminate the dis-
turbing effects of other forms of dynamic behavior, such as explora-
tory, which may be present in the early part of the test period. If
we assume for the moment that the exploratory drive has not been
completely exhausted during the preliminary period, more explora-
tory behavior would likely be exhibited during the early part of the
test period than later. One might argue that this would tend to
vitiate the results for the sex drive. The writer, however, believes
that this is not the case. In the first place, since the conditions pre-
ceding the test period were the same for all the groups, the surplus
exploratory behavior remaining after the preliminary period ought
to be about the same for all the groups. The effect of exploratory
behavior would probably be to increase the size of the average for
each group ; the relative differences between the groups should
remain approximately the same. However, in order to furnish
some statistical evidence on this point, the "partial average" was
computed. The first five minutes were excluded on the assumption
that exploratory behavior would rapidly diminish in the early part
of the test period so that the sex drive would become progressively
more dominant. A comparison of the total and partial averages
shows quite strikingly that exploratory behavior, even if present to
some extent, does not affect the order of the differences between the
various groups (Table 8). In brief, there is a fairly high correla-
tion between the two sets of averages.

2. The indices used in this paper are designed to depict two
different aspects of behavior. First, there are the changes in
behavior due to variations in conditions of segregation. Secondly,
there are the changes that are due to variations in the incentive

208

THE SEX DRIVE

stimulus. The relative changes occurring as a result of varying the
segregation period (incentive stimulus constant) will be indicated
by a drive or motivation index. Those that occur as a result of vary-
ing the incentive stimuli (segregation period constant) will be
designated by an incentive index.3 The general formula for the
calculation of these indices is :

Y — X

Index =

X + Y

where X is the standard, and Y is the comparison value. In other
words, it is the difference between the averages for these two sets
of conditions divided by their sum. The reason for using the sum of
the standard and comparison values in the denominator instead
of taking the standard as the unit was to keep the index from fluctu-
ating between infinite values. By using the sum, the maximum
numerical value is unity, so that the index can fluctuate only
between the extreme values of + 1.0 and — 1.0. Positive values of
the index indicate that conditions (segregation or incentive) are
modified in such a way as to intensify the behavior represented by
the standard; negative values indicate a decrease in the form of
behavior represented by the standard. One advantage implied by
this index is that the value of the index is in a sense independent
of the averages, so that given relative differences appear equally
significant. In other terms, the greater the magnitudes compared,
the less significant the same absolute difference. Using absolute
values, on the other hand, would tend to give too much weight to
differences when the averages are large, and insufficient weight to
differences when the averages are very small.

In the case of the motivation index, the average for the unsegre-
gated group was always taken as the standard. This step was taken
on the assumption that the unsegregated state approximates most
closely to the normal condition for the animals in their natural
habitat. In some respects, a better standard would have been an
unsegregated group which had been given a short period of sex
deprivation before the test, on the ground that sex deprivation

3 The use and limitations of these indices and the formulae for their cal-
culation were suggested by Dr. T. N. Jenkins.

THE SEX DRIVE

209

would more nearly equalize the intra-organic factors involved (cf.
p. 485). Since optimal sex activity occurs after 24 hours of sex
deprivation (Warner), the 1-day group, according to this criterion,
would be the best norm for the males. However, the writer wished
to compare the results for males and females ... a fact which
stressed the desirability of a norm common to both sexes. In the
present case, this purpose is best answered by the unsegregated
group, for no 1-day group of females was tested. In view of the
temporal characteristics of the oestrus cycle of the female, it would
be impractical to test a 1-day group of females on account of the
excessive waste of animals as discussed in the previous section.

As the numerical value of the motivation index approaches nearer
and nearer to unity, the change in the intra-organic factors, as
represented by the form of behavior measured, becomes more and
more complete. In the course of most work on drives by the
obstruction method, it is not very likely that an index of +1-0 or
— 1.0 would occur, since this would indicate a complete change of
the intra-organic state of the organism. However, in an extreme
case of negative conditioning, it may be quite possible to obtain a
motivation index of + 1.0 or — 1.0. When the motivation index is
positive, it indicates that the intra-organic state of the animal has
been modified so as to intensify the behavior represented by the
average for the unsegregated group.4

In the case of the incentive index, the behavior toward the normal
sex stimulus was always taken as the standard.

Then, for males,

X = average for the normal sex stimulus (receptive female), and

Y = average for the abnormal sex stimulus (male or non-receptive

female).
For receptive females,
X — average for the normal sex stimulus (male), and

Y = average for the abnormal sex stimulus (receptive or non-receptive

female).

An incentive index of zero indicates that the two stimuli com-
pared are equally effective in eliciting a response. As the incentive

* To generalize beyond the data presented in this paper, these indices would
require a statement of their probable errors. The writer relied, instead, upon
the reliability of the difference as presented in Tables 7 and 10.

210

THE 8EX DRIVE

index approaches nearer and nearer to unity, the change in the
incentive value of the stimulus, as represented by the behavior
measured, becomes greater and greater. If there are no crossings to
the abnormal stimulus, the index is — 1.0. A positive index points
towards homosexuality. In fact it would be possible to define homo-
sexuality in terms of a positive incentive index . . . the greater
the index, the more specific the homosexual drive. Small negative
values, on the other hand, would serve to indicate the development
of homosexual tendencies. An index of +1.0 would indicate that
the heterosexual stimulus is wholly ineffectual in eliciting the
measured response.5 Other data, of course, have been employed
throughout to determine the presence of homosexual tendencies
(cf. sections on Individual Differences and the Effect of the Mating
Period).

3. The magnitude and nature of the dispersion or variability is
indicated by four measures. The magnitude is indicated by the
average deviation, the coefficient of variation, and the range; and
the nature of the scatter is indicated by the relative density. The
first three measures are well known and need no further comment.
The relative density6 for each quartile of the distribution is cal-
culated as follows: The first, second, and third quartiles are com-
puted. From these values and the range for the entire distribution
considered as a continuous series the range for each quartile is com-
puted. Then the relative density is computed by finding the recip-
rocal of the quartile range and multiplying by 100. It may be
expressed in the following form :

10 0

Relative density =

quartile range (continuous series)

A table of relative densities is convenient in pointing out deviations

6 The discussion of motivation and incentive indices does not imply that such
indices have such a limited application as to apply only to problems in sex
behavior. It is believed that they could be used equally well in connection with
other drives, such as the hunger and the thirst drive. In the case of hunger,
different periods of food deprivation would furnish one basis for a motivation
index. On the other hand, the incentive index could be used in a study of the
incentive value of various foods by using one kind of food, such as McCollum's
diet, as the standard of comparison.

8 Suggested by Dr. T. N. Jenkins.

THE SEX DRIVE

211

from the normal form of the distribution, such as cases of bimo-
dality.

4. The data for determining the persistence of the drive during
the test period are furnished by the temporal distributions (Table
5 and Figure 6). If the number of crossings for a given animal
falls off toward the end of the period, the drive may be said to be
less persistent, as judged by behavior criteria, than when the rate
of crossing is relatively constant throughout the test period, or
actually increases.

An attempt was also made to determine the general or average
effect of different periods of segregation by means of a combined
score, obtained by pooling the results for the three incentive con-
ditions. Such scores, however, involve several factors which tend
to nullify their significance. These factors in certain cases tend to
lower the average independently of changes in the average that can
be assigned to variations in the strength of the sex drive. First,
there is the factor of the physiological limit to the number of cross-
ings which a rat can make in a test period of given length. In other
words, we might say that the number of crossings is not a linear
function of the strength of the drive. As the drive becomes
stronger, the number of crossings approaches the physiological
limit so that each additional increment to the strength of the drive
is paralleled by smaller and smaller increments in the number of
crossings. Therefore the absolute number of crossings as a measure
of the strength of the sex drive tends to minimize the significance
of the drive in intensely motivated animals. A second factor is
negative adaptation to certain incentives used. For example, the
35-day females failed to cross as frequently to a receptive female
as to an empty incentive compartment. In fact, it is the writer's
opinion that the avoidance of the females in this case was evidence
of a strong heterosexual sex drive in the animals of this group.
Here it is quite plain that negative adaptation would tend to lower
the value of the pooled result even though the animals may actually
be quite active. Consequently, the method of pooling gives results
which are spurious as measures of activity.

212

THE SEX DRIVE

A. Variations in the strength of the drive and the relative incentive
values of the stimuli

One day of segregation nearly doubles the strength of the normal
sex drive,7 as measured by the average number of crossings to a
receptive female (Tables 7 and 8, and Figure 2). Unsegregated
males, on the average, cross only 7.8 times, but when segregated for
one day their average rises to 13.5. After one day the sex drive
seems to diminish slowly in strength as the segregation period
lengthens until at 155 days the activity is no greater than that of
the unsegregated group. In fact, the motivation index shows a
slight negative value for the 155-day group (Table 9 and Figure

TABLE 7

Reliability of the difference between the average number of crossings of males
for various deprivation periods

GROUPS

(deprivation periods)

STIMULUS ANIMAL

THE DIFFERENCE

STANDARD DEVIA-
TION OF THE DIF-
FERENCE

m

o

9

W

h

s

S. D. OF THE DIF-
FERENCE

CHANCES IN 100 OF
A TRUE DIFFERENCE
GREATER THAN 0

Unsegregated and 1 day

Cornified

Dicestrum

Male

5.7
3.2
1.9

1.94
1.27
1.81

2.93
2.51
1.04

99.8
99.4
85

1 day and 35 days

Cornified

Dicestrum

Male

2.9
3.7
.2

1.97
1.43
2.40

1.47
2.58
.08

93

99.5
53

35 days and 155 days

Cornified

Dicestrum

Male

3.0
.9
1.8

2.00
1.81
2.27

1.50
.49
.79

93
69
79

Unsegregated and 35 days

Cornified

Dicestrum

Male

2.8
.5
1.7

2.37
1.78
2.51

1.18
.29
.68

88
62
75

Unsegregated and 155 days

Cornified

Dicestrum

Male

.2
1.4
.1

1.97
1.69
1.63

.10
.83
.06

54
79
52

1 day and 155 days

Cornified
Dicestrum
Male

5.9
4.6
2.0

1.45
1.31
1.46

4.07
3.51
1.36

100
100
91

7 By normal sex drive the writer means the sex drive to the normal stimulus
or incentive animal. In the case of a male the normal incentive stimulus is the
receptive female; in the case of the female the normal stimulus is the male.

THE SEX DRIVE

213

if

1-H CO

it

-3 *
IS u

«° CO

13

II

CO «

04 C4

s °

13

1-SS

^ -S So

? bo

stop ssx

AJIAIJDB X3g

6A*ep sg

sA*ep sg

£

3

h -<

XBp i

AJIAtJOB X3§

—stop SSI

stop SSI

jCjjaijdb xag
— stop S£

stop SS

top I

XjIAJJDB X3g

—stop SSI

stop SSI

X)iaipb xa§
—stop S£

stop %z

top i

214

THE SEX DRIVE

Fig. 2. Graphs of the average number of crossings and average deviations
for males for various deprivation periods. Incentive stimulus indicated in
upper right-hand corner of each diagram.

THE SEX DRIVE

215

3). This becomes of more significance when it is considered that
the plane of activity is low as compared with that of the unsegre-
gated group. The number of approaches and contacts indicates
that there is a preponderance of activity in the case of the unsegre-
gated group, although the small number of crossings would suggest
that they are quite effectually deterred by the grill.

The response to a non-receptive female shows somewhat similar
relationships. The motivation index takes on a substantial positive
value after one day of segregation; for longer periods it becomes
increasingly negative. The rise in the number of crossings to a
non-receptive female, which occurs after one day of segregation is
practically as great as the drop from the first to the thirty-fifth
day, the average for the unsegregated group being 7.7, the 1-day
group 10.9, and the 35-day 7.2. After 35 days, the decline in the
strength of the drive seems to be very slow, for the average falls
only to 6.3 as the result of 155 days of segregation.

The response to a male incentive shows an almost complete re-
versal of trend from the results obtained in the case of the normal
incentive. One day of segregation substantially decreases the in-
tensity of the response to a male incentive animal. In fact, the
motivation index shows a fairly large negative value, whereas those
for one day of segregation in the case of the female incentives are
strongly positive. Thereafter, the strength of the drive to a male
gradually increases until at 155 days, the motivation index takes
on a slight positive value. In other words, as the segregation
period lengthens beyond one day, there is an increase in the stim-
ulation value of the male.

A comparison of the responses to the three types of incentive
stimuli shows some pronounced differences in the incentive values
of the stimuli for various periods of segregation. In the case of
the unsegregated group, all three types of incentive stimuli are
about equally effective — a fact which indicates that there is not a
very specific sex drive. One day of segregation, on the other hand,
increases enormously the stimulus value of the receptive female as
compared with the male (Table 10 and Figure 4). An inspection
of the incentive indices shows that there is also a substantial in-
crease in the incentive value of a non-receptive female (Table 11

216

THE SEX DRIVE

THE SEX DRIVE

217

and Figure 5). Although the three types of incentive animals rank
in the same order after 35 days of segregation, the effectiveness of
both types of female has diminished, the loss being greatest for the
non-receptive female. The increase in the incentive value of the
male, however, is somewhat obscured by the bimodality of the dis-

TABLE 10

Reliability of tJw difference between the average number of crossings of males
for various stimulus animals

GROUPS
(stimulus
animals )

DEPRIVATION

THE DIF-

STANDARD
DEVIATION

DIFFERENCE

CHANCES
IN 100 OF
A DIFFER-

PERIOD

FERENCE

OF THE
DIFFERENCE

S. D. OF THE
DIFFERENCE

ENCE
GREATER
THAN 0

Cornified

and
Dicestrum

0 days

1 day
35 days

155 days

.1

2.6

3.4
1.3

2.03
1.11
2.17
1.61

.04

1.34
1.56
.80

52

91
94
79

Dicestrum
and
Male

0 days

1 day
35 days

155 days

.4
4.7

.8
1.9

1.80
1.27
2.49
1.50

.22
3.70

.32
1.26

58
100
62
89

Corniiied
and
Male

0 days

1 day
35 days

155 days

.3
7.3
4.2

.6

2.15
1.54
2.70
1.37

.14
4.74
1.55

.43

56
100
93
66

tribution. In the case of the 155-day group, differences in incen-
tive value are not so great as in the 1-day and 35-day groups. In
fact, the high degree of specificity to the receptive female has prac-
tically disappeared. Indeed the incentive index suggests homosex-
ual tendencies.

The number of approaches and contacts (Table 6) for the 155-
day group indicates that the degree of activity is practically the
same and uniformly low for all incentive stimuli — a fact which
shows among other things that the sex drive is rather weak as the
result of this very long period of segregation.

An inspection of Table 8 shows that the trend of the partial
averages parallels very closely that of the total averages. This
fact, in conjunction with the reliability of the difference as pre-
sented in Tables 7 and 10, indicates that the directions of the ob-
served differences are true.

It will be noted that the number of crossings is approximately
the same for all three incentive conditions for the unsegregated

218

THE SEX DRIVE

THE SEX DRIVE

■ss

to «h
•3 co

1 X

"c3

8

s

n

-a <

« T3

>.2

'u. i-

D

E

«-

in

8

CB

220 THE SEX DRIVE

THE SEX DRIVE

221

group, being low in all cases. The explanation seems to lie in the
fact that the females are constantly present in the home cage, pro-
viding opportunity for frequent sex activity. That some copulatory
activity had occurred recently is suggested by the fact that the
average number of crossings to the receptive female is 7.8, whereas
for satiated males it is only 3.6 and 12.2 after a 12-hour recovery
period (Warner). Evidently the behavior of the unsegregated
males is not wholly dominated by the sex drive, but involves ex-
ploratory and social factors ; these males have had sufficient oppor-
tunity for sexual activity preceding the test to decrease materially
the strength of the sex drive.

In the 1-day series, the trend is distinctly upward for both re-
ceptive and non-receptive, and downward for male incentives. The
conditions during the mating period were precisely the same for all
three groups. In the case of the receptive female incentive, the rise
was in all likelihood due to (1) the fact that a 24-hour recovery
priod after mating was allowed, and (2) the fact that during this
interval the males were segregated for the first time and thus de-
prived of the opportunity for stimulation by the females in the
living cage. However, it should be noted that during the time pre-
vious to the mating period a receptive female may not have always
been present in the living cage, since occasionally all females may
have been in the dioestrus stage. The rise in the crossings to a non-
receptive female offers greater difficulties since the male usually
does not copulate with a non-receptive female. The fact that the
rise is not so great in this case shows that the non-receptive female
is less of a stimulus than the receptive and this is what would be
naturally expected.

The behavior of the test animals suggests that the reaction to the
non-receptive female might be looked upon as a case of substitute
stimulus. Indicator males, as used by the writer, at first go through
the entire repertory of sex reactions, but after a time (varying
from two or three days to a month) omit such preliminary sex be-
havior as the exploratory nosing of the external genetalia of the
female. The more experienced indicator males do not wait for the
females to display such characteristic reactions as the short run,
but mount immediately and finally come to mount both receptive

222

THE SEX DRIVE

and non-receptive females alike. The present writer has observed
eight cases in which experienced indicator males mounted and ap-
peared to copulate with pregnant females. In no case, however,
were indicator males, after segregation for 24 hours, observed to
mount males although nosing sometimes occurred .

The number of crossings to the male incentive is less after segre-
gation for one day than for the unsegregated group. This may be
due to negative adaptation to males during the segregation period.

The effect of increasing the segregation period from 1 day to 35
days is to lessen the strength of the drive to receptive and non-
receptive females in terms of the number of crossings, while the
average is approximately the same to the male incentive. This de-
crease is evidently due in part to a waning of the sex drive. How-
ever, there is evidence of a disintegration of sex habits and a
development of homosexual tendencies, as shown by the fact that the
response to the male was almost as high (in absolute number of
crossings) as to the non-receptive female and much closer to the
value for the receptive female than in the 1-day group.

The value for the 155-day series is uniformly low as might have
been expected since these males had been segregated before puberty.
The female incentives (both receptive and non-receptive) appear to
be merely a novel, rather than a sex stimulus. As will be noted, the
crossings to the male incentive stimulus are greater than to the
non-receptive, and measurably greater than to the normal stim-
ulus. This suggests that segregation at so early an age brings about
the development of a homosexual tendency. It is true that the drive
toward the male incentive is, on the whole, weaker than that to the
normal incentive. As indicated by the incentive index, however, it
is relatively stronger after a very long (155 days) segregation
period than a short one (1 day or 35 days). This may be accounted
for in part on the assumption that after 155 days there is a devel-
opment of definite homosexual tendencies. That these are in evi-
dence will be more clearly indicated when Individual Differences
and the Effect of the Mating Period are discussed.

It might be argued that the small differences between the various
incentives, in terms of number of crossings, indicate not so much
homosexuality as desexualization. Such a difference in interpreta-

THE SEX DRIVE

223

tion becomes very largely a matter of definition of terms. If homo-
sexuality be thought of as precluding normal sex behavior alto-
gether, then desexualization might conceivably be the better term
to use in this instance. But as stated in the introductory section we
have used the term throughout this paper to include all specifically
sexual behavior displayed toward an animal of the same sex even
though at the same time similar behavior be exhibited toward an
animal of the opposite sex. If this definition be allowed, the 155-
day group did show a homosexual tendency, when averages, vari-
ability, etc., are all taken into account.

The observations of Stone (8) support the view that homosexual
habits might thus be built up in the segregated male rat. He says :
" Associated with scratching of the body is the nibbling of the ex-
ternal genitals which, in the case of the males, eventually comes to
be chiefly licking of the penis. The stimulating effect of licking
probably causes the act to be repeated and to become habitual.
Whether or not this act may be considered a form of masturbation
in the rat is an open question ... No case of orgasm has been dis-
covered as a result of the licking of the organ, but the difficulties
of detecting the orgasm are such that one cannot be sure that it is
not effected." Stone, however, has observed homosexual behavior
among males. ''Upon rare occasions adult male rats have been
observed to mount one another under ordinary conditions of cage
confinement. This occurs most frequently when strange males are
introduced into a cage of males, or when several males still in a
state of sexual excitement following prolonged copulation are
brought together in their home cage."

Observations by the writer corroborate these facts and add fur-
ther support to the suggestion that homosexuality may develop in
the segregated male. Males of both the 35-day and 155-day groups
were seen to go through the copulatory act as defined by Stone
(11) — namely, pursuit and mounting of a male, palpation of the
sides of the sex object, cessation of palpation and backward lunge,
and licking of the penis. Furthermore, three indicator males which
had been allowed to copulate rather freely with receptive females
every night for over a month were observed to go through the cop-
ulatory act with each other on the second night after they had been

224

THE SEX DRIVE

confined to their cages and allowed no sex activity with females.
Prior to this short segregation period they had been segregated
during the day but never seen to display sexual behavior toward
each other.

B. Individual differences

The magnitude and character of the scatter around the various
averages are shown in Table 12 and Figure 2. These measures of
variability confirm the conclusion that homosexuality has developed
in certain individuals of the 35-day group and has become general
in the 155-day group.

It will be noted that the relative variability is uniformly high in

TABLE 12
Measures of variability for males

Relative Density*

c

c

a*

Stimulus

atio

a!

ient
iatio

i rtile

artil

artil

'£

animal

Deprh
period

Averaj
deviati

Quarti
deviati

Coeffic
of var

Range

1st qua

2nd qu

3rd qu

s
cr

0

days

4.8

4.5

77.0

0—19

31

23

20

13

day

3.1

2.6

28.7

7—23

34

31

47

12

28

days

6.1

6.5

71.6

0—22

28

10

28

15

Female —

35

days —

cornified

sex

activity

3.3

2.6

29.9

4—20

13

27

66

23

stage

155

days

1.9

1.2

25.0

5—12

83

76

83

23

155

days —

sex

activity

1.5

2.8

20.4

7—13

83

100

166

23

0

days

3.0

1.6

48.8

3—15

40

66

55

13

1

day

1.3

1.4

14.9

9—13

111

62

83

76

Female in

35

days

3.4

2.5

59.1

0—14

18

66

28

22

dioestrum —

35

days —

recuperative

sex

activity

3.3

3.0

39.7

2—15

15

28

40

66

stage

155

days

3.1

2.5

61.5

2—16

66

33

50

11

155

days —

sex

activity

2.1

1.7

51.5

2—11

66

66

50

12

0

days

3.5

3.0

54.0

1—18

22

40

28

13

day

3.0

3.4

60.6

2—12

58

55

20

40

35

days

5.3

5.5

103.7

0—15

66

28

13

28

Male

35

days —

sex

activity

3.9

1.6

47.4

5—24

40

66

58

7

155

days

2.2

1.4

33.5

4—14

37

76

66

11

155

days —

sex

activity

2.6

1.6

70.8

0—8

23

125

40

66

1 0 0

* Relative density = — - —

quartile range (continuous series)

THE SEX DRIVE

225

the unsegregated series (0 days). This is very likely due to the
fact that some of the animals have copulated more recently and
more often than others previous to the test, inasmuch as the test
animals were taken directly out of the living cages which included
both sexes. The 1-day segregated series show marked homogeneity
in comparison. This is not surprising in view of the fact that the
same opportunity for copulation was given in the mating period,
which was followed by the constant segregation interval, allow-
ing precisely the same condition for recovery. It should be noted,
however, that although the average number of crossings of the
1-day group to males is less than it is for the unsegregated group,
nevertheless the variability continues to be relatively high. An
examination of the relative densities discloses the presence of
bimodality. It has previously been pointed out that the decrease
in the number of crossings suggests negative adaptaion ; the bimo-
dality would indicate that part of the males react in this way. The
greatest individual variation occurs in the 35-day series — in fact,
the coefficient of variation is so tremendously high in the male in-
centive group that some explanation for this heterogeneity must
be sought. As before mentioned, the distribution of this group is
bimodal. When this fact is taken in conjunction with the low
average, the only feasible explanation would seem to be that with
certain animals the normal stimulus alone will elicit a sex response
while in others the drive takes the homosexual direction. About 60
per cent of the group crossed to the male incentive stimulus be-
tween zero and five times, the modal value being two. The remain-
ing 40 per cent crossed between eleven and fifteen times, the modal
value being thirteen. The disproportionately large average devia-
tion of this group when the incentive is a receptive female again
suggests the disintegration of sex habits and the development of
homosexuality.

The small amount of variability found in the 155-day group in-
dicates the operation of causes which are common to all members of
the group. One reason for this low degree of variability is that
there has been no radical change in living conditions since puberty
which would tend to induce different modes of adaptation, as has
already been found in the case of the 35-day group. As previously

226

THE SEX DRIVE

stated, the number of approaches and contacts shows that activity-
is uniformly low for the 155-day group as compared with the un-
segregated group, but the average number of crossings is about the
same. Since the number of crossings is about the same, in spite of
the fact that the approaches and contacts in the 155-day group are
apparently much lower, it appears that the incentive stimulus must
have been the effective factor in eliciting the crossing response.
Variability is also greater in the case of the unsegregated group —
a fact which corroborates the conclusion that general or random
activity is high for the unsegregated group, and comparatively low
in the 155-day group. Indeed, the very low variability in the case of
the 155-day group indicates that one factor is relatively predom-
inant in the drive to cross to the incentive compartment. Since the
male is not a novel stimulus, the test animal must have reacted to
it as a sex stimulus.

Thus we find that the average number of crossings to the homo-
sexual stimulus is only slightly greater than that to the females.
The relatively large number of crossings to the females is prob-
ably due to the fact that they are novel stimuli. The fact that the
drive is apparently stronger to a receptive than to a non-receptive
female is probably a consequence of the greater activity of the fe-
male in oestrum.

C. Persistence of the drive as shown by the temporal distribution of
drive behavior during the test

Table 5 and Figure 6 indicate the persistence of the drive during
the 20-minute test as shown by the per cent of crossings for the
four five-minute intervals. According to this measure, the unseg-
regated groups are quite persistent in crossing to all three incen-
tives. The 35-day series, on the other hand, show wide differences
depending upon the incentive employed. The drive to the normal
incentive seems to gather increasing force as the test proceeds,
while that to the non-receptive female diminishes in persistence.
The drive to the male is the most constant. With the 155-day group,
the drive to the non-receptive female shows the greatest decrease in
persistence during the last fifteen minutes of the test.

THE SEX DRIVE 227

45
30
15

UNSSOREGATED

45

n

i

30
15
0

! §S

I ^

1

■

1 1

- o Says 1 Bey SA 0 Days 1 Day SA 0 Days 1 Day SA

45

90

15
P.

rrri

35 DATS

45
80
15

f

ri -

ivi

H

1

It

28 Days 35 D-SA 36 Days 35 P-SA 35 Days 35 D-SA

45

P

3§

155 DATS

45
30

iri
\^

■ i

HI

30
15
Q

•*„>

Lv

k

"V"

mm

k

15

156 Days 156 D-SA 155 Days 155 D-SA 155 Days 155 D-SA

COHKIFIED | DIOESTBDM | MACE
No Ssx. Activity Sex Activity |

Fig. 6. Column diagram showing temporal distribution of crossings of the
male groups during the twenty-minute test period in five-minute intervals.

228

THE SEX DRIVE

D. The effect of the mating period upon the test twenty-four hours
later

1. Variations in the strength of the drive and in the relative
incentive value of the stimuli. It will be recalled (section on
method) that all male groups, except the 1-day, were separated
into two subgroups, one of which was given a mating period 24
hours before the test, and in the other the test animals were taken
directly from the living cages to be tested. The effect of the inter-
polated mating period can be found by comparing the indices of

TABLE 13

'Reliability of the difference between the average number of crossings of males
with and without sex activity for various deprivation periods

a

o

• w
«5

H

s

o o

fc

« a

O

H H <

e W

M

w

GROUPS

STIMULUS ANIMAL

3

B
r*

Eh

s|

gel

n

K

5 * £

« B 3

i<

cc

. w

00 fe

5 < o

Unsegregated

Cornified

5.7

1.94

2

.93

99.8

and

Diosstrum

3.2

1.27

2

.51

99.4

Unsegregated — Sex activity

Male

1.9

1.81

1

.04

85

35 days

Cornified

2.8

1.92

1

.45

93

and

Dicestrum

3.2

1.88

1

.70

96

35 days — Sex activity

Male

3.9

2.60

1

.50

93

155 days

Cornified

1.6

1.35

1

.18

88

and

Dicestrum

1.2

1.53

.78

78

155 days — Sex activity

Male

3.6

1.35

2

.66

99.6

the two subgroups (Tables 13, 14, and 15, and Figures 7 and 8).
When the normal incentive (receptive female) is used, the effect
of segregation is counterbalanced to a considerable extent by such
a period, there being a decrease in the effectiveness of the mating
period in passing from the short to the long periods of segregation
(0 — 35 — 155 days). When a non-receptive female is used, a sim-
ilar tendency is noted, except for the 155-day group which shows a
slightly opposite direction. When a male incentive is employed, the
effect of the interpolated mating period does not seem consistent.
As would be expected the mating period lowered the effectiveness
of the male incentive for the zero group. A decided increase in

TEE SEX DRIVE

229

crossings to the male in the 35-day group, apparently due to the
mating period, would seem to mean that the opportunity to mate
with a female stirred up greater sex activity even toward a male
incentive animal 24 hours later. The mating period in the case of
the longer segregation interval (the group which was segregated

TABLE 14

Reliability of the difference between the average number of crossings of males

after sex activity

ft

o
ft

2VIA-
S DIP-

w

o

ft
8

ft ft
o o

2 ft

S«ft

rt ft <!

ft
8
ft

35

ft

1

ft
a

ftg g

GROUPS

STIMULUS ANIMAL

ft
ft

s

eh

§ ft ft
3 © o

Q ft

H m ft
ca Eh ft

ft
ft
ft
3

EH

si

CHANCES J
A TRUE Di:
GREATER 1

ft

w

Eh

. ft

0 «

. ft

co ft

Unsegregated — Sex activity
and

35 days — Sex activity

Cornified
Dicestrum
Male

.1

.5
4.1

1.33
1.40
1.94

2

.07
.35
.11

53
64
98

35 days — Sex activity
and

155 days — Sex activity

Cornified
Dicestrum
Male

4.2
5.3
5.7

1.22
1.60
1.85

3.44
3.31
3.08

100
100
100

Unsegregated — Sex activity
and

155 days — Sex activity

Cornified
Dicestrum
Male

4.3
5.8
1.6

1.31
1.05
1.56

3.28
5.52
1.02

100
100
84

before puberty) seemed to arouse the normal sex drive, causing a
partial breakdown of the homosexual tendency as indicated by the
rather marked decrease in crossing to the male. Indeed, the fact
that sex activity has operated to decrease the average number of
crossings to a male from 8.2 to 4.6, points rather significantly to
the conclusion that homosexual tendencies were present before the
mating period was given.

2. Individual differences. The subgroups which were given the
mating period showed less variability (except for the 155-day
group crossing to a male in which case the variability is less when
the last fifteen minutes are considered) regardless of the nature of
the incentive, than the corresponding groups in which the mating
period was omitted. The general effect of the 2-hour period of sex
activity seems therefore to operate as a constant in reducing indi-
vidual variation.

230

THE SEX DRIVE

THE SEX DRIVE

231

232

THE SEX DRIVE

3. Persistence of the drive as shown by the temporal distribution
of drive behavior during the test (Tables 5 and 6, and Figure 6).
An examination of the temporal distribution of crossings during
the 20 minutes shows that the mating period had little effect on
persistence except in the case of the 35-day group and herein slight
significance attaches to the criterion.

IV. Results for Females

The experimental technique in the study of segregation is a more
difficult problem in the female than in the male on account of the
oestrus cycle. The external conditions for both sexes, however,
were essentially the same. Table 15 indicates that the same segre-
gation periods and incentive stimuli were employed with each sex.
As explained under the section on method (p. 485), all females
tested were in the cornified stage of the oestrus cycle. One impor-

TABLE 15

Showing female groups

NUMBER OF

PERIOD OF SEX

ANIMALS

DEPRIVATION

10

0

hours

15

0

hours

10

0

hours

21

35

days*
days*
days*

12

35

15

35

10

155

days

7

155

days

10

155

days

9

155

days plus 5 copula-
tions plus 1 hour de-
privation

8

155

days plus 5 copula-
tions plus 1 hour de-
privation

6

155

days plus 5 copula-
tions plus 1 hour de-
privation

OBJECT IN INCENTIVE COMPARTMENT

Male

Female in dioestrum — recuperative stage

Female — Cornified stage

Male

Female in dioestrum — recuperative stage

Female — Cornified stage

Male

Female in dioestrum — recuperative stage

Female — Cornified stage

Male

Female in dioestrum — recuperative stage
Female — Cornified stage

* This group was tested by L. H. Warner and is reported in his monograph
on sex behavior.

tant difference was necessitated by the character of the oestrus
cycle, i.e., it was impossible to give the females the 2-hour mating
period 24 hours before the test, as in the case of the males. The
receptive stage is not long enough to allow a 24-hour lapse of time

THE SEX DRIVE

233

between such a period and the test, since the test animal must be in
the cornified stage. No 1-day group, therefore, could be tested.
However, a variant to this condition was introduced by giving half
of the females of the 155-day group a brief mating period one hour
before the test (p. 196). In each case, the amount of sex activity
allowed was five copulatory responses, which usually occurred
within a few minutes, although fifteen minutes was required in
some cases.

In order to facilitate a comparison, the results for the females
will be treated under the same topical plan as that followed in the
discussion of the results for males. The main topics are :

A. Variations in the strength of the drive and the relative in-
centive value of the stimuli.

B. Individual differences.

C. Persistence of the drive as shown by the temporal distribu-
tions of drive behavior during the test.

D. The effect of copulatory activity (five responses) upon the
test one hour later.

A. Variations in the strength of the drive and the relative incentive
values of the stimuli

The changes in the strength of the sex drive as a result of vary-
ing the segregation conditions preceding the test are best indicated
in the graphs of Figures 9 and 10. The average number of cross-
ings of the unsegregated females to the normal incentive stimulus
( male ) is greater than the number of crossings for any other group
tested. However, there is practically no difference (14.2 and 14.1)
between the unsegregated group and the 35-day groups, there being
only 52 chances in 100 of a true difference. This is convincing evi-
dence that the sex drive, after 35 days of segregation, in females
raised unsegregated to maturity, showed no apparent loss to the
appropriate stimulus. The data for the 155-day group furnish a
reliable indication that segregation from before puberty has re-
sulted in a substantial decrease in the strength of the normal sex
drive.

The response to the non-receptive female, like that to the normal
incentive, indicates that the drive is not altered appreciably as a

234

THE SEX DRIVE

Fig. 9. Graphs of the average number of crossings and average deviations
for females for various deprivation periods. Incentive stimulus indicated on
each diagram.

THE SEX DRIVE

235

236

THE SEX DRIVE

result of 35 days of segregation. One hundred and fifty-five days
of segregation, however, results in an enormous increase, the aver-
age rising from 8.5 to 13.4, and the motivation index rising to
+.201.

The response to the receptive female, unlike that to the non-
receptive female, shows a very striking decrease in the strength of
the drive as a result of 35 days of segregation. The average number
of crossings falls from 12.0 for the unsegregated group to 3.8 for
the 35-day group, a change which is not only large but reliable, the
probability that there is not a true difference being less than one
chance in a million (cf. Table 20). The relative importance of this
decrease in the strength of the drive to a receptive female is indi-
cated by the motivation index which drops to — .518 (cf. Table 22).
However, in the case of the 155-day group, the number of crossings
to a receptive female rises again to practically the same figure as
that of the unsegregated group. This rise is contrary to what one
would be led to predict on the basis of the trend up to 35 days of
segregation.

It will be easier to interpret these changes in the strength of the
sex drive after examining the effects produced by varying the in-
centive conditions. This procedure enables one to determine the
relative importance of the three types of animal as sex stimuli. A
glance at Figure 11 shows that there are some striking changes in
the relative incentive value of the various stimuli. The magnitude
and direction of these differences are indicated in Figure 12. As
might naturally be expected, the normal sex stimulus in the case
of the unsegregated group is more effective than either type of
female, although the incentive value of the non-receptive female is
less than that of the female in oestrum. An examination of the
averages of the approaches and contacts for the unsegregated group
shows that the activity is much higher in the case of the male than
with the other incentive stimuli — the sum of the approaches and
contacts for the male being 5.5, that for the receptive female being
only 0.8 (cf. Table 19). The 35-day group also react more strongly
to the normal stimulus than to the other incentive conditions. But
there is a significant change in the specificity of the drive ; in fact,
differences in incentive value reach a maximum in this group,

238

THE SEX DRIVE

THE SEX DRIVE

239

the incentive index dropping in one case (receptive females)
to the low figure of — .575. This increase in value might
lead one to expect a still greater degree of specificity as the segre-
gation period is lengthened. But the data for the 155-day group
indicate that the contrary is true. Not only is the specificity very
low, but what little there is seems to be directed toward a homo-
sexual incentive (non-receptive female).

An examination of Table 21, shows that the order of magnitude
and direction of the differences between the various partial aver-
ages follow the same general trend as those in the case of the total
averages. This is additional evidence, other than that found in
Tables 20 and 23, that the tendencies which have been pointed out
above are warranted by the data.

In spite of the apparent reversals in trend, a careful interpreta-
tion of the data shows conclusively that most of these changes can
be explained on the basis of the formation and disintegration of sex
habits and that segregation from before puberty to 185 days is
probably accompanied by the development of homosexual ten-
dencies. One point which merits special examination is the striking
drop in the number of crossings to a receptive female in the case of
the 35-day group. This is all the more interesting when we note
that the number of crossings for the non-receptive female is prac-
tically the same as for the unsegregated group. In fact, the average
number of crossings to a receptive female is five less than that for
any other group, the next lowest being 8.5. Seventy -five per cent
of the animals cross less than five times and only one animal more
than seven times. The number of crossings are even fewer than for
Warner's control group, in which no incentive animal whatever
was used in the incentive compartment. The average crossings for
his control is 5.0, whereas the average for the present group is only
3.8. Since the same electrical obstruction was used in both groups,
and since both groups had been segregated for the same length of
time under the same conditions of care, the presence of the recep-
tive female must have been the deterrent which reduced the num-
ber of crossings from 5.0 to 3.8.

This conclusion is further supported by the fact that there is a
considerable increase in the number of approaches and contacts as

240

THE SEX DRIVE

pagmjco

3AI)BJ3dn33J

— uinj)sa>ip

01 3|ELU9J

aSos
payiiuoo
— ajBUiaj

— uiiuissip

Ul 3]PUJ3j

aSeis
payiinco

— 'uinJiS3jip

o 5

s3uissoj3
saqsBOJddy

&3uissoj3
saipeojddy

sSuissojq
«oejtio3
sstpBOJddy

sStHSSOJQ

swbiuoq
ssipEOJddy

ssqsBOJddy

«3BJTK>3

s»i|3cojddy

sSuissoj3
svbiucq
saipecuddy

sSuissar)
saqsetuddy
$Suiss<xi3

saipsoiddy
sSausojQ

satppcuddy
sSujssojq

ssipeojddy
sSuissojq

SWBlUCr)

sstpecuddy
s3uissoj3

saipBOJddy
s8uissoj3

saipccuddy
sSatssar)

wipeojddy

r^- en r\j i

■f — —

\c ^ — *m —

m «vj — — —
«*■ <M — — eg —
rsi pg — <ni rj — pm

■ <si <\j ««■> «*i <

r ) — r- 1 _ Psi

<\» so >

m Ul — t-r, '

— — CM — —

— i CM rsj — eg.

THE SEX DRIVE

241

compared with those of the other groups, indicating that, even
though the number of crossings are few, nevertheless there is ma-
terial evidence of considerable activity. These facts show conclu-
sively that there has been no positive conditioning to the receptive
female as a result of 35 days of segregation. On the other hand,
certain facts about mating behavior support the hypothesis that
negative conditioning may have been set up. When a female is in
the receptive stage, sensitivity to tactual stimulation in the genital
region seems to be increased. For example, Stone (8) observed that
an application of acid to the base of the tail of a non-receptive fe-
male causes it to exhibit mating behavior, due to increased sensi-
tivity. These facts suggest the possibility that, since these females
are orientated toward the normal stimulus alone and yet are ex-
tremely sensitive to any tactual stimulation, the increased activity
of the receptive female induces an avoiding reaction. In other
words, the females tend to avoid sexual stimulation from all stim-
uli other than the normal one. In this manner, negative condition-
ing is perhaps developed to such an extent that a receptive female
in the incentive compartment may become a definite deterrent.

It may seem strange that negative conditioning does not also
occur in the unsegregated group when a receptive female is used
as an incentive. This can probably be accounted for in either of
two ways. The first explanation is a statistical one. Since the
number of animals in the living cage is kept constant (the other
half being males) , there are only half as many females present in
a cage as in the case of the 35-day group. Therefore, when a given
female is in oestrum, the chances that another female will be in
oestrum at the same time are correspondingly lowered. The chances
of another female being in oestrum would be four times as great
for the 35-day group as it would be for the unsegregated group.
In the second place, if two receptive females should happen to be
present in the cage at the same time, they would have little oppor-
tunity to stimulate each other because they are probably reacting
entirely to the males. These two conditions seem able to account for
the fact that there is no evidence of negative conditioning to recep-
tive females in the unsegregated group.

A word should also be said concerning the apparent lack of neg-

242

THE SEX DRIVE

ative conditioning in the case of the 155-day group. Since negative
conditioning occurs in the 35-day group, still more would it nat-
urally be expected to be found in the case of the 155-day group,
but, as already noted, quite the contrary is true. The number of

TABLE 17

Showing temporal distribution of approaches, contacts, and crossings of the female
groups during the twenty -minute test period in one-minute intervals

2 S

•2 E

Q 2

EEEEEEEE

EEEEEEEEEES

•5 •£ •£ M •£ -5 "5 *S ■£

Male

Approaches
0 days Contacts
Crossings

2 3 4 2 4 3 2 2 2 1
0102100200
15 12 67958776

1 0 0 1 0 1 4 3 2 3
0 0 1 1 0 1 0 4 1 1

6 6 9 7 5 5 6 5 5 6

m Female in
gdicEStrum —
5> recuperative
<2 stage

Approaches 0101002004
0 days Contacts 12 10 0 10 10 1

Crossings 988 10 511 7484

1 10 10 10 0 0 1
1001000000
7933747667

3 Female —
cornified
stage

Approaches
0 days Contacts
Crossings

0 0000010010001001000
10002000100000001000
10 8576737674755754566

'Male

Approaches
35 days Contacts
Crossings

12 4 11 12 10 787256762244225
30122333343034695821
14 11 12 12 10 7 8 10 9 9 9 10 11 16 35 26 24 30 28 2

£ Female in

dicEstrum —
"3 recuperative
Z sta8e

Approaches 20001001010002210140
35 days Contacts 00000001 120222100001

Crossings 12 10 858765235525434342

H Female-
s'cornified
w stage

Approaches
35 days Contacts
Crossings

6 4
4 3
14 10

3 3 5
1 3 4
6 1 3

3 8 8
0 3 0

4 2 2

4 5
4 3
2 1

5 2

0 0

1 1

5 0
1 2
1 1

2 3 8 7
0 2 13
2 0 11

0 2

1 1

2 2

Male

Approaches
155 days Contacts

Crossings

0 0
0 I
10 10

0 1 0

1 2 I

7 3 7

0 0 1

4 0 0

5 9 5

0 0
0 0

8 6

1 0

0 0

3 8

0 0
0 0

6 7

0 0 0 0
0 0 0 1
6 6 4 4

Approaches
1 hour Contacts
Crossings

2 0
0 2
7 7

0 0 0

1 I 3

7 7 6

0 2 1
0 0 1
5 4 5

0 2
0 2
5 4

0 0

1 2

5 2

0 0
2 0
4 4

0 0 0 0
10 0 0
4 4 13

^2 Female in
"*" dicFstrum —
~Z recuperative
« stage

Approaches
155 days Contacts
Crossings

0 1
0 0
7 7

0 0 0

1 0 1

5 7 5

0 0 0
2 1 1
4 5 2

0 0
0 1

6 5

0 0
0 0

5 3

0 0
0 0

3 5

0 10 0
0 0 0 0
6 0 4 4

Approaches
1 hour Contacts
Crossings

0 I

0 0
8 7

0 1 0

1 2 0

7 6 4

0 0 I

1 1 0

4 6 5

I 0
0 2
4 5

0 0
0 0

5 6

0 0
0 0
3 4

0 0 0 0
0 0 0 0
4 8 4 5

Female —
cornified
stage

Approaches
155 days Contacts
Crossings

0 1 0

1 1 2

5 6 7

2 0 0
I 0 1

5 5 6

0 0
0 1

8 7

0 I

1 1

7 7

0 0
0 0

6 3

0 0 0 0
10 0 1

7 6 6 5

1 1

8 3

Approaches
I hour Contacts
Cro.rjinKS

0 0
0 0

6 6

0 1 1
0 0 0
5 1 4

0 2 0

2 0 0

3 1 5

0 0

0 0

1 3

0 0
0 0

2 3

0 0

1 0
0 3

0 0 10
0 0 0 0
0 111

1 1

0 0

1 0

crossings to a receptive female as the result of 155 days of segre-
gation is practically the same as that for no segregation. This
change, however, will be understood when we are careful to note
the differences in the life history of the animals prior to segrega-
tion. The 155-day group have had no opportunity to become con-
ditioned to males. On the other hand, they have been surrounded

THE SEX DRIVE

243

continuously by females since weaning. In other words, they had a
chance to become habituated to sexual stimulation by receptive fe-
males but have had no chance whatever to become adapted to stim-
ulation by males. The general activity due to internal stimulation
would naturally cause a large number of crossings and the number

TABLE 18

Showing temporal distribution of approaches, contacts, and crossings of the
female groups during the twenty-minute test period in five-minute intervals

0 to

5th

otn to

lUtn

11th

to 15th

16th

to 20th

m

Minute

Mir

lute

Minute

Minute

i i|

> o

4-1 "

A

A s

'£g
<u.2

s.§

w c

r

Num-

\

Per

Mum

.[Num-

Jrer

s

Num

- Per

XT

Num'

- Per

Q ~

CO

<

ber

Cent

ber

Cent

ber

Cent

ber

Cent

Approaches

15

37.5

10

25.0

2

5.0

13

32.5

41/

^Male

0

days

Contacts

4

26.7

2

13.3

2

13.3

7

46.7

t r

I J

Crossings

49

34.5

33

23.2

33

23.2

27

19.0

142

2j Female in
£ dicEstrum —

Approaches

2

15.4

6

46.1

3

23.0

2

15.4

13

0

days

Contacts

4

44.4

3

33.3

2

22.2

o

o

9

j> recuperative
stage

Crossings

40

30.1

34

25.5

29

21.8

30

22.5

133

Female —

;

Approaches

0

0

—

2

66.6

1

33.3

0

0

3

comified

o

ays

Contacts

3

60.0

1

20 0

0

0

1

20.0

5

Crossings

36

30.0

30

25.0

28

23.3

26

J 1 .6

120

Approaches

49

43.0

29

25.4

23

20.2

13

11.4

1 14

£Male

35

ays

8

12.1

16

24.2

16

24.2

26

39.4

66

Crossings

59

19.2

45

14.7

81

26.4

22

39.7

307

£ Female in
dicEstrum —

Approaches

3

20.0

2

13.3

4

26.7

6

40.0

IS

35

days

Contacts

0

0

4

33.3

7

58.3

1

8.3

12

12 recuperative
n stage

Crossings

43

41.7

23

22.3

21

20.4

16

15.5

103

Z Female —

Approaches

21

25.3

28

33.7

14

16.9

20

24.1

83

S? comified
1/3 stage

35

days

Contacts

15

41.7

10

27.8

3

8.3

8

22.2

36

Crossings

34

59.6

11

19.3

6

10.5

6

10.5

57

Approaches

1

25.0

1

25.0

1

25.0

1

25.0

4

155

days

Contacts

5

50.0

4

40.0

0

0

1

10.0

10

Male

Crossings

37

29.8

33

26.6

30

24.2

24

19.3

124

Approaches

2

25.0

5

62.5

0

0

1

12.5

8

I

hour

Contacts

7

43.7

3

18.7

6

37.5

0

0

16

>i

n

T3

Crossings

34

36.9

23

25.0

19

20.6

16

17.4

92

Approaches

1

50.0

0

0

0

0

1

50.0

2

vr> ,

"> Female in

155

days

Contacts

2

28.5

5

71.4

0

0

0

0

7

~" dicEstrum —

Crossings

31

33.0

22

23.4

22

23.4

19

20.2

94

recuperative

Approaches

2

50.0

2

50.0

0

0

0

0

4

« stage

60

1

hour

Contacts

3

33 3

4

44.4

0

0

2

22.2

9

V

Crossings

32

30.8

24

23.0

22

21.1

26

25.0

104

6C

V3

Approaches

2

40.0

2

40.0

1

20.0

0

0

5

Female —

comified

stage

155

days

Contacts

6

40.0

3

20.0

3

20.0

3

20.0

15

Crossings

34

27.6

31

25.2

30

24.4

28

22.7

123

Approaches

2

28.5

2

28.5

0

0

3

42.8

7

I

hour

Contacts

0

0

2

66.6

1

33.3

0

0

3

Crossings

22

46.8

13

27.6

8

17.0

4

8.5

47

of crossings is probably not lowered by negative conditioning. Per-
haps it may even be increased by a slight positive conditioning to
receptive females.

The lack of a strong differential response to the three stimuli in
the case of the 155-day group can be accounted for by the fact that

244

THE SEX DRIVE

f #UBip3LU

UOIJEIJEA

iu9ioiy903

UOUBIAap
33BJ3AV

UBipap^

UOUBUBA

juap 53303

uoijEiAap

'uE3JAJ

~ -s
c

c3

33BJ3A y
33UB^[

UOIJBUBA

UOUBIAOp
*UE3y\[

»3BJ3Ay

UB!P3W

jbuiiue isa;

UOi;BAUd3Q

TEUItUE

snjnuj'us

sdnoif)

VO

<VJ

to

CO

OO

a*

VO

vO

00

-

cn

to

to
to

to

m

vO

CMj

CV1

&>

c->

to

to

>+

to

CVJ

oo'

p

LO

<Mj

OO

6—19

0—15

O

CM

i

2—24

in

I

m

o

O

to

id

d

vq

O

oo

p

vo
vo

©

esj

CO

oo

vq

CO

VO

■ m

CM

0—5 1.5

0—4 6

0—2 .5

0—9 3.1

0—7 1.0

o

r>»

CO

p

vq

to

CM

p

vO

LO

O

VO

CM*

oo

vo

VO

oo

in

Ov

vq

csi

p

CO

to

~t-

CM

0—14

CM

J

0—14

to

1

vo

q

vq

co

It

a

>>

cj
T3

re

©

O

O

to

Male

Female in
idicEstrum —
, recuperative

stage

Female —

cornified

stage

■Male

Female in
dicEstrum —
recuperative
stage

p3jB3ai3asuQ

tvj

oo vq

oo t>

oocni

Oo'ov

in in
Om

— vOOv

©intMj

oo vq ^.

00

**CMCM

tvivc^r

— to

t^inod
CM

^9
t\i

— Ov
CM in to

in«M^-

r^vq —
—

00

Tj-CvjTt-
CM© —

•*p«vj
to to to

to 00 vO
CM f>« O

0—16

vO CT> Ov

TTT

CO <M CM

vo in vo

III

o o

TTT

CM CO fM

to

OON

CM in 00

O cm r«-

to ON g\

0«ol««.

vq © vq

roO

WON

r-roo

oo C> co
Om —

oo

\or^«vj

vONOv

to vqp

CM

o vq

p — —

in in —

I

in

in

in CM m

^in-t;

■*p«v)

•street

to

inoqvq

vOOm

O co in

■*

CO

r^oo'vd
into-*-

vo'vot^
vo CM in

©vo'od

CM — —

in©f»

^■inm

OO CM O

in
in

^t- 0s vq

tom ^

in — r^.

0—14

III

III

CO CO to

Hi

t-- O CO

vqpr>.

15 days

at v- n
rs O O

ti u a
in

>. 3 2

•axH

m

Female —

cornified

stage

1 v

a \ >
•- c —

JB 2 2

"s 5 & «

£ 3 3 60

U. U to

SXBp JC paiB33J33§

sXsp 5S| paaB33i33§

THE SEX DRIVE

245

the sex life of these females has never been complicated by the
stimulative effect incident to copulation. Since the copulatory re-
sponse has never been elicited, they have never had an opportunity
to become specifically orientated toward a given stimulus. The sex
habit can never have developed beyond nosing or mounting. Al-
though the specificity is low, nevertheless, it seem that sex activity
is directed toward the homosexual stimulus, instead of the hetero-
sexual — in fact, the incentive index for the non-receptive stimulus
is slightly positive. Furthermore, the indication of homosexuality
may be greater than is shown by the numerical data, since the male
is a novel stimulus to the test animal. The introduction of this ele-
ment of novelty into the situation may produce enough additional

TABLE 20

Reliability of the difference between the average number of crossings for
females for various deprivation periods

GROUPS
(deprivation periods)

STIMULUS AN I M Ali

THE DIFFERENCE

STANDARD DEVIA-
TION OF THE DIF-
FERENCE

DIFFERENCE

S. D. OF THE DIF-
FERENCE

CHANCES IN 100 OF
A TRUE DIFFERENCE
GREATER THAN 0

Unsegregated and 35 days

Male

Dicestrum

Cornified

.1

.4

8.2

1.82
1.71
1.50

.05
.23

5.46

52
59
100

35 days and 155 days

Male

Dicestrum

Cornified

1.7
4.9
8.5

1.48
1.55
2.03

1.15
3.16
4.18

87
100
100

155 days and 155 days —
copulatory activity

Male

Dicestrum
Cornified

2.2
.4
4.5

2.33
.88
2.03

.94
.45
2.21

83
67

98.6

Unsegregated and 155 days

Male

Dicestrum

Cornified

1.8
4.5
.3

1.71
1.22
2.24

1.05
3.68
.13

85
100

55

activity on the part of the test animal to raise appreciably the aver-
age number of crossings in response to a male, in spite of the fact
that the male may have very little stimulation value as a sex object.

In the 155-day group the slight advantage of the non-receptive
over the receptive female as concerns their capacity to elicit a re-
sponse from the female test animal is what should be expected as a

246

THE SEX DRIVE

So

.5 e

OO SO

o\ r>»

T3
C

!§

o

O

« 3

a.

3
O

a

o

<3> ci
?3 03

eo p

J So

>

•a

In
D,

O

o

«4

Is

K_ — +■»

-o

-a

s> O ^

-a 3-3
£ 0 «

m

i U

•3 3 "5

a.

-a

►1

^3

C

.2

aj "3

>.2
'Z ^

s-s.

THE SEX DRIVE

247

result of the conditions of segregation. The opportunity for con-
tact with receptive females is less than that for non-receptive fe-
males. Since occasionally two receptive females would be in oes-
trum at the same time, the test animals may have been somewhat
conditioned to receptive females. But, on account of the limited
number of animals in a cage, this would be the exception and not
the rule. Consequently, segregation would facilitate sexual condi-
tioning to the non-receptive female.

TABLE 23

Reliability of the difference between the average number of crossings of females
for various stimulus animals

GROUPS
( STIMULUS ANIMALS )

Male
and
Picestrum

Dicestrum

and
Cornified

Male
and
Oornified

DEPRIVATION PERIOD

THE DIFFERENCE

STANDARD DEVIA-
TION OF THE DIF-
FERENCE

DIFFERENCE

S. D. OF THE DIF-
FERENCE

O O
_ te O

^ w <

«3 Q k>

ill

w 64 «
o < o

0 days
35 days
L55 days

155 days — eopulatory act.

5.3
5.6
1.0
2.8

1.73
1.80
1.18
2.19

3.06
3.11
.84
1.27

100
100
80
89

0 days
35 days
155 days

155 days — eopulatory aet.

3.1
4.7
1.1
5.2

1.61
1.61
1.98
.97

1.92
2.91
.55
5.36

97

99.8
71
100

0 days
35 days
155 days

155 days — eopulatory act.

2.2
10.3
.1
2.4

1.90
1.41
2.08
2.28

1.15
7.30
.04

1.05

87
100
51
85

B. Individual differences

The changes in variability tend to support the conclusions drawn
from an analysis of the differences in the average number of cross-
ings for the various groups. The magnitude and characteristics of
the scatter around the averages is shown in Table 25. The graphs
of Figure 9 show the relation of the changes in variability to the
changes in the average number of crossings for various periods of
segregation. Figure 11 is designed to facilitate a comparison of the
variabilities for the different incentive stimuli.

Variability of response to the normal incentive, as measured by

248

THE SEX DRIVE

2 Q

1 s

mi

3 "2

«u *- £ .5 c
Q a <55 «

h a.

THE SEX DRIVE

249

TABLE 25
Measures of variability for females

RELATIVE

DENSITY*

^£
2
S t*

STIMULUS ANIMAL

DEPRIVATION PERIOD

^ERAGE
5VIATIO

LTARTIL
5VIATIO

£1 <

!>

H
O

;t quart

id quai

'd quar

,h quar

<! O

o o

i

IN

CO

0 days

3

G

3.5

31.6

6—19

18

25

33

66

35 days

4

2

2.3

37.1

2—24

8

47

40

15

Male

155 days

2

4

2.8

24.1

8—16

40

40

32

111

155 days —

eopul. act.

5

1

6.4

62.6

2—19

38

15

15

38

Female in

0 days

Z

1

2.5

43.8

0 — 15

14

50

47

23

dicestrum —

35 days

3

9

3.5

56^4

0—15

18

22

40

28

recuperative

155 days

1

5

1.2

14.0

10—16

37

66

100

55

stage

155 days —

copul. act.

1

2

1.5

11.5

11—15

100

66

66

100

Female —
cornified
stage

0 days
35 days
155 days

3
2
4

2
6
7

2.9
1.5
5.3

33.3
84.2
47.8

6—20
0—16
2—20

37
45
15

23
76
22

66
58
16

15
9

55

155 days —

copul. act.

1

6

.5

25.6

3—9

18

142

250

250

♦Eelative density = _

quartile range (continuous series)

both the coefficient of variation and the average deviation increases
slightly with 35 days of segregation as compared with the unsegre-
gated group. The range also indicates increased variability, being
6 — 19 for the unsegregated group, and 2 — 24 for the 35-day group.

Although the central tendencies of the unsegregated and 35-day
groups are very similar in terms of the average number of crossings
to the normal incentive, individual differences are increased slightly
by segregation. It may be that the drive of some females is waning,
while for others the presence of a male again during the test rein-
forces a drive made more intense through inability to reach con-
summation.

As judged by all three measures — namely, average deviation,
coefficient of variation, and range, for the normal sex drive, the
155-day group are less variable than either the unsegregated or
35-day females. The reason for the lack of variability of the 155-
day group is better appreciated when we are reminded of the fact
that this group has lived under conditions less favorable than the
other groups for the development of individual differences. Let us
examine these conditions for the various groups. First, all groups,

250

THE SEX DRIVE

other than the 155-day group, were segregated after puberty — a
condition which has permitted the formation of very definite sex
habits in response to a normal stimulus. The 155-day group, on the
other hand, were segregated before puberty — a condition which
precludes any possibility of the formation of highly integrated sex
habits in response to a normal and, therefore, definitely oriented
sexual stimulus. The first copulatory experiences at puberty would
probably have a tendency to condition the drive of different fe-
males unequally. But the 155-day group have never had an op-
portunity for the stimulative effect of copulation, whereas both the
unsegregated and 35-day group have had an opportunity to copu-
late since puberty. Consequently, differences in aggressiveness in
the case of the unsegregated and 35-day groups are probably ac-
centuated and are manifested objectively in the form of mating
behavior. By aggressive mating behavior is meant responses on the
part of the female which tend to stimulate the male and induce
copulation.

The response of the unsegregated group to a normal stimulus
varies more than that of the 155-day group for another reason —
namely, the unsegregated females may have had widely differing
amounts of copulatory activity before being tested, because the ex-
perimenter had no way of measuring how long the animal had been
in oestrum before the test. This should tend toward greater varia-
bility in the response to the normal stimulus than would be found
in the case of the 155-day group. The amount of sex activity im-
mediately previous to the test period might be just sufficient with
some animals to reinforce the drive ; in others enough to lessen the
drive.,

Neither of these conditions obtains for the 155-day group. We
may, therefore, safely conclude that the dispersion of the 155-day
group for the response to a normal stimulus is less than that of the
other groups because, not having had a chance to become habituated
to the normal stimulus, the possibility of unequal conditioning
through practice and the inequalities in the rate of disintegration
of sex habits resulting from lack of exercise, are thereby eliminated.

Differences in variability of response to non-receptive females
follow the same trend as those for the normal stimulus, although

THE SEX DRIVE

251

the differences between the various segregated groups tend to be
greater for the non-receptive female than for the normal stimulus.
The greatest variability, as measured by average deviations and
the coefficient of variation, occurs with the 35-day group and the
least variability with the 155-day group. Since a strong drive pro-
duces homogeneity in a group, an increase in the relative impor-
tance of other drives may be the cause of variability. This may
help to explain the increased variability of the 35-day group.

From the large number of crossings of the 155-day group and
from the lack of opportunity to become conditioned to the normal
stimulus, we have concluded that they may be homosexual. The
unity of the drive to cross to a non-receptive female further sup-
ports this conclusion, for we have noted repeatedly in the case of
the male groups (see section on Males) that a strong drive causes
a unified response so that individual differences approach a mini-
mum.

When a variability of the drive to receptive females is exam-
ined, it is noted that the order of the differences for the average
deviations is not the same as that for relative variability. To be
specific, the average deviation is smallest for the 35-day group, but
the coefficient of variation is by far the largest for this group. The
absolute variability, i.e., average deviation, is important in itself,
but, since groups are being compared, relative variability is the
most suitable measure. When the groups are compared in terms
of relative variability, the unsegregated animals are the least vari-
able, the 155-day group more variable, and the 35-day group ex-
tremely variable.

As previously explained, there seems to be considerable justifica-
tion for assuming that a strong drive is indicated by homogeneity
of response. Likewise, when variability is great, the drive is weak.
Upon this basis the increased variability of the 155-day group to
receptive females can be explained. If the drive is weak, the cross-
ings which occur are in response to various incentives and, there-
fore, cause increased variability for the groups. In other words,
even though the fairly large number of crossings indicates a large
amount of activity due to internal stimulation, the variability in-
dicates that the specific drive to a receptive female is weak, thus

252

THE SEX DRIVE

enhancing the relative significance of exploratory behavior and
general activity. The increased variability of the 35-day group,
on the other hand, indicates considerable individual variability in
negative conditioning. In other words, some females have been neg-
atively conditioned to receptive females more strongly than others.

A brief survey of the differences in relative effectiveness among
the three types of incentive stimulus, as measured by variability,
lends further support to conclusions previously drawn. Figure 11
reveals the fact that the variability of the unsegregated group is
fairly constant for the three incentives. The 35-day group resem-
bles the unsegregated group in terms of absolute variability. The
coefficient of variation, however, discloses genuine differences. The
drive to males is most homogeneous (37), that to non-receptive fe-
males is next (56), while that to receptive females is extremely
heterogeneous (84). The data of the 35-day group would, there-
fore, seem to suggest rather conclusively that the normal sex drive
is very strong, while the drive to receptive females is very weak.
Since receptive females are so ineffective, the crossings are prob-
ably due to the functioning of various other drives.

The differences in variability of the 155-day group are perhaps
more significant than those of the other groups. The drive to non-
receptive females is most homogeneous, that to males less homo-
geneous, and that to receptive females quite heterogeneous. These
differences support the view, treated more fully in the previous
section, that the response to the non-receptive female is due to con-
ditioning toward the non-receptive female ; that to males, is prob-
ably due largely to the novelty of the stimulus ; and that to the re-
ceptive female is due to the combined effect of several drives.

It is interesting to note that, although the averages of the 155-
day group in general do not differ widely from those of the unseg-
regated, nevertheless the indices of variability indicate that the
155-day group is in general more homogeneous than the unsegre-
gated group. This would seem to indicate that the unsegregated
group exhibits more general activity than the 155-day group. That
this long period of segregation has had a tendency to make these
animals sluggish is also evidenced by the relative number of

THE SEX DRIVE

253

approaches and contacts for the two groups — the number of ap-
proaches and contacts is uniformly low for the 155-day group.

C. Persistence of the sex drive as shown by the temporal distribu-
tion of drive behavior during the test

The third measure of the strength of the sex drive, which has
been employed in this investigation, is its persistence throughout
the twenty-minute test period. The constancy of the drive is indi-
cated by the percentage changes for the five-minute intervals (cf.
Table 18 and Figure 13). On the strength of the data, not much
can be said definitely about persistence. The differences, as indi-
cated by the variations in the number of crossings for the five-
minute intervals, are not great enough to warrant any sweeping
conclusions. The number of crossings is usually greater at first and
then tends to fall off somewhat before the end of the period.

The data seem to indicate that the drive of both the unsegregated
and the 155-day groups is quite persistent regardless of the nature
of the incentives. For both these groups external conditions have
been constant. That is, no outside agency has interfered with the
development of the sex drive.

The 35-day group was originally conditioned to males and allowed
to copulate freely until 150 days of age. In the last 35 days before
the test period segregation has prevented the normal functioning
of the sex drive. The result, as indicated by the temporal distribu-
tion, has been to cause the drive to the normal stimulus to become
increasingly stronger as the test period continues, that to the non-
receptive female to be less persistent, and that to the receptive
female to approach zero after the first five minutes (over 60 per
cent of the crossings occur in the first five minutes). An examina-
tion of the temporal distribution for evidences of persistence in
response to the three types of stimuli gives no significant differ-
ences, except for the 35-day group. Persistence, then, is a poor
measure for differences in the relative potency of the three types
of incentives to elicit a response on the part of the female test
animal.

254

THE SEX DRIVE

45

30

15

Ml

UNSBGREGAT2D

0 Day 8

0 Daya

0 Days

85 DAYS

cn v!

18^

86 Days

Day

35 Day

155 Days 166DayeCA

155 Daya 155 D-CA

156Day8 155Days-CA

MALE

DI0B3THUM :

C0B5IFISD

Vo Copulatory Activity £^D CojwUtory Aotivity»

Fig. 13. Column diagram showing temporal distributions of crossings of
the female groups during the twenty-minute test period in five-minute intervals.

THE SEX DRIVE

255

D. The effect of copalatory activity (five responses) upon the test
one hour later

1. Variations in strength of the drive and the relative incentive
values of the stimuli. Under the section on method, the reasons for
giving an interval of copulatory activity (five copulations one hour
previous to the test) to the females of the 155-day group were stated.
The results for this group are presented (Tables 19, 21, 22, 23, 24,
and 26, and Figures 9, 10, 11, 12, and 13).

One of the significant results is that copulatory activity causes a
decrease in the number of crossings. The average number of cross-
ings to a male decreases from 12.4 to 10.2. The average number of
crossings to a receptive female after copulatory activity decreases
even more, dropping from 12.3 to 7.8. Copulatory activity, however,
produces practically no effect upon the response to the non-
receptive female, the average with copulatory activity being 13.0,
and without, 13.4. Partial averages further emphasize these rela-
tionships. The incentive indices also show very clearly that the
tendencies, as indicated by the direction of the small relative differ-
ences for the receptive and non-receptive stimuli, are accentuated
by copulatory activity. The index for the non-receptive female
becomes still larger in the negative direction, while that for recep-
tive females becomes even larger but in the positive direction. A
negative index indicates that the incentive value is higher for the
non-receptive female than it is for the normal stimulus. This fact
in itself suggests the development of homosexuality. The receptive
female, on the other hand, loses in effectiveness after copulatory
activity. When the index is calculated for the last fifteen minutes,
these differences are even more strikingly exhibited.

All of the quantitative measures indicate that five copulatory
acts, one hour before the test, effect a decrease in the drive to cross
to males. An examination of the individual records suggests that
copulation was painful. More than 50 per cent of the females which
crossed less than eleven times gave evidences of some attempt to
avoid the copulatory advances of the male. On the other hand, all
of the animals which crossed more than eleven times were decidedly
aggressive in stimulating the male during the copulatory interval.
Escape behavior is well exemplified in the following example :

256

THE SEX DRIVE

1st min. Male copulates. Female remains inactive.

2nd min. Male approaches. Female runs.

Male noses female genitals. Female kicks.

3rd min. Male noses female genitals. Female lies on her back and kicks.

4th min. Male licks his genitals. Female approaches.

5th min. Male copulates. Female runs and squeals.

6th min. Male licks his genitals. Female remains inactive.

7th min. Male washes his body. Female remains inactive.

8th min. Male copulates twice although female kicks and squeals.

9th min. Male noses female genitals. Female retreats.

10th min. Male noses female genitals. Female kicks.

11th min. Male noses female genitals. Female kicks.

12th min. Male attempts to mount. Female turns and fights.

13th min. Male copulates. Female kicks and squeals.

The following is a fair example of aggressive behavior of female
(group crossing more than eleven times) during the copulatory
interval :

1st min. Male noses female genitals. Female takes short run character-
istic of mating.
Male copulates. Female walks about the cage.
Male copulates — then licks his genitals. Female approaches,
noses male genitals.
2nd min. Male licks his genitals. Female again approaches and noses
male genitals.

Male explores cage. Female follows and noses genitals of male
Male stands still. Female approaches and then gives character-
istic run.

3rd min. Male copulates. Female approaches again and then hops a

short distance.
Male copulates. Female noses and then gives short run.
4th min. Male licks his genitals. Female approaches and then gives a

short run.

5th min. Male stands. Female noses male genitals and then hops a short
distance.

6th min. Male copulates. Female again noses male genitals.

From these records it seems justifiable to conclude that the first
copulatory experience of mature animals is very effective in condi-
tioning the female either positively or negatively to the male.

While five copulatory acts may induce or inhibit a sex drive to
the male, it does not serve to break down the homosexual drive to the
non-receptive female. In other words, after copulatory activity

THE SEX DRIVE

257

the homosexual tendencies seem to be about as firmly fixed as ever.
Not only is there a decrease in the number of crossings to a male,
but there is also a quite decided decrease in the number of crossings
to a receptive female, facts which indicate considerable increase
in specificity of response to non-receptive females. The number of
cases is too small to generalize from with assurance. We can at
least state that, for the cases tested to a receptive female as the
stimulus, the average fell from 12.3 to 7.8 after copulatory activity
and that no female crossed over nine times, while 70 per cent of the
animals which were tested without having sex activity crossed more
than nine times. The effect of copulatory activity, therefore, in the
155-day group, is to increase differences in incentive values. It
may be possible that if the 155-day females were given more copu-
latory activity in successive periods of oestrum, they might be
reconditioned, so that they would give an avoiding reaction to non-
receptive females and display aggressive mating behavior in the
presence of a male.

2. The effect of previous copulatory activity as measured by
variability. The measures of variability shed some light upon the
conclusions previously drawn. We find that five copulations increase
the variability of response to a male, supporting the hypothesis
that copulatory activity may condition some females positively to
a male and others negatively. An inspection of the table of relative
densities will show that this conclusion is still further supported by
evidence of bimodality.

Copulatory activity does not affect variability of response to a
non-receptive female, a fact which supports the conclusion, already
drawn, that the homosexual drive to a non-receptive female is not
changed by previous copulatory activity. Copulatory activity, on
the other hand, lowers variability of response to a receptive female,
a fact which lends support to the conclusion that copulatory activity
tends to enhance the differences in incentive value. The decrease in
variability is one indication of greater homogeneity of response
and shows the more predominant operation of a single drive. This
single drive is probably manifested in the form of exploratory
behavior, the decrease in the number of crossings being due to a
diminution of sex behavior (already small) toward the receptive

258

THE SEX DRIVE

female. The evidence for specificity of response to the non-receptive
female merely adds further support to the argument for homo-
sexuality in the 155-day group.

3. The effect of previous copulatory activity as measured by
persistence. Examination of the temporal distributions indicates
that previous copulatory activity tends to make the behavior toward
males and receptive females more transient in character. There is
a considerable drop in persistence in the case of the male stimulus
and a very decided drop in persistence in the case of the receptive
female. It seems to the writer that this is still further evidence that
the chief effect of copulatory activity, in the case of the 155-day
group, is to increase the specificity of response and, since it is
shown especially toward the non-receptive female, there is very
strong evidence of homosexual behavior.

V. Summary of Conclusions

A. Male

Effect of segregation

1. Unsegregated males are fairly active but do not seem to be
motivated strongly by the sex drive when tested at 185 days of age,
all three incentives (male, non-receptive and receptive female)
being about equally effective.

2. One day of segregation, at 185 days, has the effect of doubling
the number of crossings to the normal incentive (receptive female),
and of increasing very materially the number of crossings to the
non-receptive female.

3. Thirty-five days of segregation (150-185 days) decrease the
normal sex drive, due probably to a waning of the drive in some
animals and to the development of homosexual tendencies in others.
The non-receptive female appears to act to some extent as a sub-
stitute stimulus for the receptive female in both this and the 1-day
group.

4. Segregation before puberty (at 30 days) seems to lower the
activity of the group toward both the receptive and the non-
receptive female. The sex drive in general appears to be subnormal
but more crossings occur to the male than to either the receptive or
the non-receptive female.

THE SEX DRIVE

259

Influence of the mating period on segregation

1. A 2-hour mating period interpolated 24 hours before the test,
seems to counteract in part the decrease in the normal drive result-
ing from segregation. However, the longer the animals have been
segregated, the less effective is the mating period.

2. The mating period seems to cause the 35-day group to become
sexually more active to the male as well as to the female incentive.

3. The mating period evidently decreases homosexuality in the
155-day group.

4. During the 2-hour mating period the number of males which
actually copulate was small, regardless of segregation conditions.
However, the mating period has a decided effect on the number of
crossings in the test 24 hours later.

B. Female

Effect of segregation

1. The 35-day period of segregation does not appear in the 185-
day animals to decrease the sex drive to the male. The number of
crossings was much lower to the non-receptive female and extremely
low to the receptive female. In the latter case, there was strong
evidence of negative conditioning having taken place in the living
cage previous to the test; more than half of the crossings occur
within the first three minutes of the test and the total crossings
are fewer than when no incentive whatever was employed.

2. Segregation before puberty (at 30 days) in general weakens
the drive for the normal, or male, incentive. The female appears to
be stimulated most by a non-receptive female and least by a recep-
tive female (which reacts against sex advances) thus showing the
operation of the homosexual factor.

Influence of the mating period on segregation

In females segregated before puberty (see 2 above) five copula-
tory responses one hour before the test apparently have no effect in
changing the direction of the homosexual drive. Mating in this
manner decreases the drive to the normal incentive and also
increases variability. The detailed records suggest that copulation

260

THE SEX DRIVE

was painful with the result that an avoiding reaction to males was
noted in the test one hour later.

C. Comparison of the male and the female

A detailed comparison of the sexes from group to group seems
hardly necessary in view of the analysis already made. However,
the following generalizations regarding sex differences seem to be
warranted.

1. The values for the females are consistently higher than for
the males under both segregated and unsegregated conditions,
although the difference is not great in certain cases. This suggests
that the female drive tends to be stronger than the male under the
conditions imposed. The difference between sexes is practically
negligible when the points of maximal sex activity are considered.
The unsegregated female and the 1-day segregated male, represent-
ing such optimal drive conditions, differ only slightly (14.2 and
13.5 respectively) .

2. The relative effect of segregation is greater for males than for
females in cutting down crossings to the normal incentive. Segre-
gation seems to bring about the appearance of the homosexual
tendency earlier in the male, although when segregation is begun
before puberty, the effect is approximately the same for both sexes.

Bibliography

(1) Avery, G. T. 1925. Notes on reproduction in guinea pigs. J. Comp.
Psychol, 5, 373-96.

(2) Craig, W. 1914. Male doves reared in isolation. J. Animal Behav.,
4, 121-33.

(3) Hamilton, G. V. 1914. A study of sexual tendencies in monkeys
and baboons. J. Animal Behav., 4, 295-319.

(4) Jenkins, T. N., L. H. Warner, and C. J. Warden. 1926. Stand-
ard apparatus for the study of animal motivation. J. Comp. Psy-
chol., 6, 361.

(5) Kempf, E. J. 1917. The social and sexual behavior of infra human
primates with some comparable facts in human behavior. Psycho-
anal. Rev., 4, 127-54.

(6) Long, J. A., and H.. McC. Evans. 1922. The oestrus cycle in the
rat and its associated phenomena. Berkeley, Univ. of Calif. Press,
148 pp. (See pp. 20 and 71.)

THE SEX DRIVE

261

(7) Loutitt, C. M. 1927. Reproductive behavior of the guinea pig. I.
The normal mating behavior. J. Comp. Psychol., 7, 247-65.

(8) Moss, F. A. 1924. A study of animal drives. J. Exper. Psychol.,
7, 165.

(9) Stone, C. P. 1922. The congenital sexual behavior of the young
male albino rat, J. Comp. Psychol., 2, 95-153.

(10) Stone, C. P. 1923. Further study of sensory functions in the ac-
tivation of sexual behavior in the young male albino rat. J. Comp.
Psychol, 3, 469-75.

(11) Stone, C. P. 1924. A note on "feminine" behavior in adult male
rats. Amer. J. Physiol, 68, 39-42.

(12) Stone, C. P. 1924. The awakening of copulatory ability in the male
albino rat. Amer. J. Physiol, 68, 407-25.

(13) Stone, C. P. 1926. The initial copulatory response of female rats
reared in isolation from the age of twenty days to the age of pu-
berty. J. Comp. Psychol, 6, 73-83.

(14) Warner, L. H. 1927. A study of sex behavior in the white rat by
means of the obstruction method. Comp. Psychol. Monog., 4, No. 22,
1-68.

(15) Watson, J. B. 1903. Animal education. Univ. Chicago Contrib.
Philos., 4, No. 2, 53.

(16) Whitman, C. 0. 1919. Orthogenic evolution in pigeons, 3. The be-
havior of pigeons. Publ. Carnegie Instit. Wash., No. 257.

3. THE EFFECTS OF GONADECTOMY, VASOTOMY,
AND INJECTIONS OF PLACENTAL AND ORCHIC
EXTRACTS ON THE SEX BEHAVIOR OF THE
WHITE RAT 1

H. W. Nissen

I. Introduction

A. Nature and scope of the problem

The general aim of this investigation was to measure the effects
of the internal secretions of the gonads on sex behavior of male
and female albino rats. More specifically, the attempt was made to
determine the influences of gonadectomy, of injections of the female
hormone into males and females, of injections of several types of
testicular extract into males, and of vasotomy on such behavior.
Among the more important problems of endocrinology, at which
these experimental procedures were especially directed, may be
mentioned the following: (1) The function of the gonads in con-
trolling sex behavior and the probable manner in which this control
is exercised in the male and female, respectively. (2) The relatively
immediate effect of the female hormone on the male; sex-specificity
of the gonadal substance and the relation between the heterologous
sex gland hormones. (3) The analysis of the testicular substance
into its efficient and incidental components. (4) The question of the
efficacy of vasotomy in restoring sexual vigor in old males.

This study differs from two others (59, 146) included in the same
general project on animal drives and also dealing with sex behavior
in that the latter concerned themselves with what may be termed
-normal" variations in intra-organic and environmental conditions,
whereas in the present series of tests arbitrary and " artificial"
physiological conditions were imposed.

Reprinted with modifications from Genetic Psychology Monographs, 1929,
5: 451-547.

262

THE SEX DRIVE

263

The present investigation differs in an important respect, also,
from most or all previous physiological studies in the field. Earlier
workers have contented themselves, for the most part, with one or
more of the following types of criteria for judging the effects of
their experimental conditions: (1) Morphological and histological
changes, including general growth, distribution of fat, distribution
and nature of hair or feathering, histology of the vaginal and
uterine mucosae, size and histological character of other organs.
(2) Mating activity and non-quantitative observations of cage
behavior; measurements of spontaneous activity. (3) Miscellaneous,
including clinical findings and reports of patients, metabolism
determinations, physiological effects on various organs and on
contractibility and fatigue of isolated muscles or muscle groups. In
the work here reported, on the other hand, the effects of the various
physiological conditions induced were measured in terms of the
intensity of the drive of the test animal towards a normally ade-
quate sex object. The indices obtained, therefore, are neither of
spontaneous activity nor of ability to copulate, but are measures of
the strength of the internal stimulation (motivation) causing the
test animal to cross an obstruction (constant electric shock) in
order to reach a sexually active animal of the opposite sex. The
scores yielded by the Columbia Obstruction Method are, of course,
the resultants of two kinds of stimulation: intra-organic and
environmental; since the latter, including as its most important
parts the shock and the incentive, was kept constant, for each sex,
throughout the study, the differences in scores obtained can be
attributed only to the changes produced in the intra-organic factor.

It has been long recognized (88) that the sex drive of most of the
vertebrates can be divided into two distinct behavior sequences,
commonly termed the contrectation and detumescence drives (39).
The former brings the animal into the proximity of a sex-object
and, in rats, leads to nosing of the genitals, gentle biting, and other
"sex play," whereas the latter culminates in coitus. The term
' ' detumescence drive ' ' is, obviously, more descriptive of the activity
of the male, who actually ejaculates the accumulation of seminal
fluid from the genital tract in copulation, than it is of the female
who, in certain species, ovulates some time after mating (86).

264

THE SEX DRIVE

Nevertheless, the descriptive division of the sex drive into two
constituent parts is applicable to both sexes, and, although "copu-
latory drive" probably is a more exact designation for the second
component, the writer has retained here the more conventional
terms already mentioned.

It is possible, of course, to consider the contrectation drive as
forming merely the ' 1 preparatory reactions" of the true sex drive
(150) or to think of the two components as constituting successive
links in a chain drive or instinct (82, p. 481). Some of the observa-
tions described later in this report (page 283) lead the writer to
conclude, however, that these "preliminary" activities have an
independent end or consummatory reaction of their own. They are
brought to a conclusion — the contrectation drive is "satisfied" —
apparently, by the olfactory, tactual, visual, and perhaps even
auditory and kinaesthetic stimulation which the sex-object affords.
This viewpoint receives support from the observation made by two
other writers (59, 146), as well as by the present author, that there
is no correlation between copulatory strength or ability and the
indices of the contrectation drive obtained by the Obstruction
Method. The contrectation drive, at least in rudimentary form, is
probably a necessary antecedent to the detumescence drive, but the
latter certainly does not always follow the manifestations of the
former. It seems probable that the contrectation drive has its prin-
cipal biological significance in its relation to the copulatory or
detumescence drive. It may well be that phylogenetically or even
ontogenetically the detumescence drive comes first ; the preparatory
reactions of this drive, then, may develop their own self-sufficient
end-reaction in accordance with some such principle as Wundt's
"heterogeny of ends." As far as the writer is aware, the present
study is the first one to attempt the quantitative determination of
gonadal influences on the contrectation drive.

B. Physiology of the sex glands

The present section aims to give only a brief account of those
facts known about the sex glands which have a direct bearing on the
results of this investigation and their interpretation. In view of the
great volume of the literature on the gonads — at least six hundred,

THE SEX DBIVE

265

and probably over a thousand articles and books on the topic have
appeared during the last decade alone — a comprehensive survey of
the field obviously cannot be undertaken here. More or less com-
plete compilations have been made by Lipschiitz (72), Biedl (12),
Marshall (86), and Harms (47).

1) The female gonads. It is well known that among mammalia
other than primates, the female will mate only during the rhythmi-
cally recurrent periods of rut or heat ; in the intervals between such
periods the animal is sexually quite inactive. These alternations in
sex behavior are only one item in a syndrome of changes going on
in the female organism during the oestrous cycle, among which
modifications are the regular growth and rupturing of the Graafian
follicles and formation of corpora lutea in the ovary, and the cyclic
distension and degeneration of the uterine epithelium. The cor-
related changes in the histological nature of the vaginal mucosa
form the basis of a convenient and accurate method, developed by
Stockard and Papanicolaou (130, 131), for determining the various
stages of the oestrous cycle. An excellent description of the cycle
in the rat is given by Long and Evans (80), who have shown that
the periods of heat are characterized by a vaginal mucosa composed
exclusively of cornified cells.

After extirpation of the ovaries the phenomena of the female
cycle come to an end (80, p. 94). The effect of spaying on behavior
has long been recognized, as is apparent from the following state-
ment by Aristotle (6) : "The ovaries of sows are excised with the
view of quenching in them sexual appetites. " It is clear, therefore,
that mating behavior, in common with other oestrual manifestations,
is dependent on the ovaries. The proof that the ovarian influence is
chemical rather than nervous is found in experimental and thera-
peutic work in which the effects of ovariectomy were overcome by
transplantations of the female sex glands or by injections of ovarian
extracts. Knauer (62), using rabbits, was apparently the first (in
1895) to perform successful auto-transplantations of ovaries. Sub-
sequent work with homoio- and hetero-transplants has quite con-
clusively demonstrated the chemical nature of the ovarian control
of the oestrous cycle. The grafted ovary can maintain its normal
structure and the rhythmical character of its functioning; some

266

THE SEX DRIVE

cases are reported (42) in which the ova of the transplant have been
fertilized and have developed normally. Feeding and injection of
ovarian substance have been tried, with varying degrees of success,
for the past 30 to 35 years. Definite proof that the oestrous cycle
can be re-established in spayed mammals by injections of an extract
of the follicular fluid has been given by Allen and Doisy (2, 3) and
others. Steinach (128) has shown that extracts of ovary and of
placenta will induce oestrum in infantile spayed rats who have
never had a normal cycle, and that such injections will cause the
reappearance of the sexual cycle in senile females whose ovaries
had ceased their rhythmical functioning two to five months earlier.
Slonaker (116) was able to produce the typical vaginal changes of
oestrum by injections of the fresh follicular hormone into female
albino rats 538 to 971 days old, some of whom had been previously
spayed. Doisy (28) points out that an overdose of the hormone is
not harmful (one rat was injected with 20 rat units), and Courrier
(26) found that varying the amount of the injection did not influ-
ence the nature of the changes in the vaginal epithelium.

The exact part of the ovary responsible for the production of the
oestrous-inducing hormone and the mechanism of its elaboration
have not been conclusively determined. The fact that this substance
is regularly found in the follicular cavity and walls has led many
writers, including Allen and Doisy (2), to the conclusion that the
hormone is produced by the follicles, perhaps under stimulation
from the maturing ova. Pugh believes that the cystic condition of
the larger follicles, which he found in nymphomaniac cows and
heifers, may have stimulated the growth of the follicular epithelial
cells. Recent work by Parkes and his associates (17, 100), indicat-
ing that the hormone is produced after the germinal epithelium has
been destroyed by irradiation, leads this investigator to conclude
that the cortex of the ovary is responsible for its elaboration.
Steinach (122), Lipschiitz (72), and others believe that the Pubcr-
tatsdruse or interstitial cells of the ovary secrete the female hor-
mone. The oestrous-inducing hormone has been found in the
placentae of woman, of the cow, mare, and other animals. The
source of the placental hormone is not definitely known; Parkes
and Bellerby (101) suggest that it is derived partly by withdrawal

THE SEX DRIVE

267

from the maternal circulation and partly by elaboration of the pla-
centa itself. The female hormone manifests complete lack of species-
specificity. Thus extracts of human placenta have been injected
with positive results into mice and rats (27, 151). Allen and his
associates (4) were able to produce oestrous growth in the genital
tract of spayed rats by injections of an extract of the ovaries of
laying hens. Fellner (32) had positive results with extracts of bird
eggs and fish ovaries injected into rabbits. Loewe (77) has demon-
strated the presence of the oestrous-inducing substance in the blood
of women, rabbits, and cows during the pro-oestrous period. Con-
cerning the little-known chemical composition of the female hor-
mone, it may suffice to note that Allen and Doisy (2) found that the
active principle followed the lipoid fraction of the follicular liquid.

Several authors (35) have reported the presence of the oestrous-
inducing substance in the corpora lutea. There is good evidence,
also, that this part of the ovary produces a substance which has,
among other possible functions, an inhibitory influence on the
follicular hormone. Loeb (76) and others have been able to hasten
the occurrence of oestrum by destruction or removal of the corpora
lutea, and Papanicolaou (99) has suspended or suppressed the
female cycle for months at a time by weekly injections of a purified
corpus luteum extract. It seems, therefore, that the corpus luteum
is either a " mixed gland," producing two distinct hormones, or
that, while actively secreting only a non-follicular agent, it also
contains some of the hormone produced by other parts of the ovary.

Wang (143) has shown that correlated with the other manifesta-
tions of the oestrous cycle there are rhythmical variations in the
spontaneous activity of female rats. During dioestrum the animals
are relatively quiet whereas at oestrum they show great activity.
After ovariectomy these fluctuations cease and the activity level
remains constantly low.

The influence of the ovarian hormone or hormones extends beyond
those conditions primarily associated with the oestrous cycle. The
morphological secondary sex characters are gradually lost after
spaying, the animal tending towards an asexual or even masculine
type. The uterus and other parts of the genital tract not only no
longer show the rhythmical variations associated with the female

268

THE SEX DRIVE

cycle but undergo a progressive atrophy. There is often an increase
in weight and decrease in basal metabolism after ovariectomy. Lee
(66) found a decrease of about 10% in heat production, following
extirpation of the ovaries, and interprets his result as demonstrat-
ing the stimulating effect of the female gonads on the thyroid gland.
Loewy and Richter (79) report a decrease of 14-20% in metabolism
following ablation of the ovaries in dogs ; the same writers state that
injection of ovarian substance into spayed animals results in raising
the metabolism 30-50% above normal. Livingston (75) found that
the pituitary gland often enlarges after ovariectomy of rabbits.

2) The male gonads. Animals castrated before puberty usually
show no signs of a sex drive at any time of life. Steinach (119)
states, however, that male rats, prepuberally castrated, may show
slight and weak signs of heterosexual interest at the normal age of
maturity. Steinach (119), Lichtenstern (67), and Busquet (20)
have found that male rodents and horses castrated after the time
of puberty continue to manifest sex activity for some time after the
operation. This was not an altogether new finding, for we find in
Aristotle (6) the following statement: "If a full-grown bull be
mutilated, he can still to all appearance unite sexually with the
cow." Rieger (108) quotes a number of authors, as far back as
Juvenal and Lucian, to the effect that the sex drive (in man) is not
extinguished after castration. Jager (58) differentiates between
the eunuch and the castrate, the former being deprived of both
testes and penis, the latter lacking only testes ; he remarks that the
eunuch seems incapable of any sexual feeling, whereas the castrate
does not lose the ability or desire to have intercourse. Stone (134)
has recently made an exact study of the length of time during
which copulatory ability continues in rats castrated when 90 days
old. (According to Stone's figures (133), the upper age limit for
the first copulatory act in male rats is 72 days.) He found that
three months after the operation about 43% of his animals were still
copulating. Unfortunately, Stone does not tell us what kind of rats
he used, so that his results are not directly comparable with our
own.

Probably the most generally accepted explanation for the gradual
extinction of the sex drive following castration is the one sum-

THE SEX DRIVE

269

marized in the following quotations from Lipschiitz (72) : "Per-
sistence of some signs of the sexual activity which diminish very
slowly after postpuberal castration is by no means incompatible
with the contention that the erotization of the nervous system
(Steinach, 1910) depends on the sexual glands." In speaking of
the slower extinction of sexual libido in man as compared to ani-
mals, Lipschiitz says, "I think that this is to be explained by the
reflex actions responsible for erection and coitus being fixed and
brought into play in the normal man by more manifold external
factors than in the lower mammal. " If I interpret the implications
of these statements correctly, this means that the sexual responses,
originally stimulated or facilitated directly by the testicular hor-
mone, become 1 1 attached, ' ' in accordance with the laws of learning,
to certain external stimuli which thereafter can evoke sex behavior
independently.

The most obvious effect of gonadectomy is, probably, the loss of
secondary sex characteristics. Pelikan (102), Tandler and Gross
(135), Koch (63), and others have described the results of castra-
tion among the Skopecs, a religious sect of Russia and Roumania,
which includes this operation as part of its ritual. Supplementing
these accounts by the great mass of clinical literature on eunuchoid-
ism and on operative or accidental castration, we find among the
more salient and common consequences of human male gonadectomy
the following: abnormal bodily proportions, often but not always
accompanied by obesity, feminine distribution of fat, scanty or
absent facial, axillary, and pubic hair, but usually luxuriant head
hair, very small penis, sallow complexion, and, if the testes were
extirpated before puberty, a childish treble voice. Birds, having
pronounced secondary sex characters, show the effects of castra-
tion in a striking manner; the caponized rooster loses his male
feathering, often ceases to crow, and, if the operation was carried
out early in life, does not develop comb and wattles.

In 1849 Berthold (10) transplanted the testes of cocks from their
normal site to other parts of the body and found that the usual
consequences of castration did not appear. Although John Hunter
(57) had done similar work in the preceding century, Berthold 's
results are generally accepted as giving the first definite proof that

270

TEE SEX DRIVE

the testicular influence is chemical rather than nervous. In 1889
Brown-Sequard (19) reported remarkable physical and mental im-
provements, following injections, into himself, of emulsions of
testicular substance. Since this time a large amount of work has
been done with auto-, homoio-, and hetero-transplants and with im-
plantations and injections of testicular substance and extracts.
Many investigators report excellent results in overcoming symp-
toms of senescence and eunuchoidism with testicle transplants and
with implantations of orchic substance. Among the beneficial
results to human patients, mentioned by such workers as Vor-
onoff (140, 141), Stanley (118), Thorek (138), Hunt (55),
Lydston (81), Retterer (105), Falcone (31), Muhsam (94), and
others, are : loss of obesity, growth of facial and body hair, in-
crease in size of penis, return of potency and erection, lower pitched
voice, improvement in memory and general intelligence, increased
strength, libido and feeling of well-being, and improvement in
asthenia, acne vulgaris, functional disorders, and psychoses. It
should be noted, as some of the above authors themselves point out,
that suggestion may have been responsible for some of the im-
provements reported ; in some cases, on the other hand, it is claimed
that the patients did not know what results to expect. The effects
of transplants often do not survive very long, although Voronoff
(141) reports the indefinite survival of anthropoid grafts in man,
and recently Moore (93) has been able to preserve the normal
structure of testes, transplanted into animals, for long periods of
time. Implanted testicular tissue soon becomes necrotic but exerts
its influence during absorption. The effect of a single injection of
testicular extract is at best transitory, and for practical reasons
this method has not been employed extensively in human work.

In 1911 Pezard (103) demonstrated the growth of comb and
wattles and the retention of normal sex behavior in the castrated
cockerel following injections of an extract of pig testicles. Since
that time many workers, including especially Steinach (120),
Sand (109), Lipschutz (72), and Voronoff (140) have performed
experiments on animals in which the effects of castration, senes-
cence, and eunuchoidism were overcome by testicular grafts or
injections. A few examples from the more recent work may be

TEE SEX DRIVE

271

cited: Wilhelm (149) reports that homoio-transplants into three
senile dogs caused increase of weight, return of sexual potency,
and partial rejuvenation of skin. Busquet (21) was able to produce
male characteristics in capons and senile cocks by injections of a
serum from young bulls, stallions, and rams. Harms (49) found
rejuvenescence in old dogs following testicular transplantations.
Thorek (138) castrated six apes and after all of them had lost
sexual potency (7 to 11 months after the operation) made homoio-
and hetero-transplantations. Foar of the apes regained completely
the ability to copulate, one recovered partially, and the other gave
negative results. Bouin and Ancel (14, 15) were able to minimize
the effects of castration in guinea pigs by subcutaneous injections of
an extract of retained testicles prepared with glycerine and water.
Steinach (120) brought back the clasping reflex in castrated frogs
by injections of a pulp of testicular substance ; the reflex appeared
12-24 hours after the injections and disappeared again in 3-4 days.
Loewy (78) was able to prevent some of the effects of gonadectomy
in capons by feeding testes, but Steinach (120) found oral admin-
istration of testicular substance to be without effect in castrated
rats.

There are three different kinds of tissue in the testicle which
might produce the male hormone: the germinal epithelium, the
interstitial cells of Leydig, and the Sertoli cells. Of these, the first
two have usually been considered the most probable source of the
efficient substance. Reinke (104) was apparently the first, in 1896,
to attribute to the interstitial cells the function of producing the
internal secretion controlling sex behavior. Without going into the
details of the controversy regarding the " interstitial gland"
(Bouin and Ancel, 13), the principal lines of evidence favoring the
view that the interstitial cells elaborate the male hormone may be
briefly summarized as follows : All of the secondary sex characters
may remain after the process of spermatogenesis has been stopped
by (a) spontaneous or artificial cryptorchidism (72), (6) by liga-
ture of the vas deferens (127), (c) by irradiation (1, 17), and (d)
by disease (72). Lillie and Bascom (69) point out that the free-
martin, whose occurrence is supposed to depend on the influence
of the male hormone, comes into being at a time of foetal life when

272

THE SEX DRIVE

the male germinal epithelium is not yet functioning but when the
male interstitial cells are active. Grafted testes exerting a mascu-
line effect often show an absence of germinal epithelium which is
accompanied by a hypertrophy of the interstitial tissue (114).
Histologically, interstitial cells look more like typical gland cells
than does the epithelial tissue (47). Certain animals showing
seasonal periods of sex activity with elaboration of secondary sex
characters at this time manifest a synchronous proliferation of
the interstitial cells.

Although the majority of workers today probably accept the in-
terstitial cells as the source of the male hormone, this view is by no
means universal. Champy (23) points out that potent testicular
transplants occur which show only a very few interstitial cells and
that some fish with definite sex characters have none of this tissue,
whereas the mole, with many interstitial cells, has no marked sec-
ondary sex characteristics. Harms (48) points out that testicular
grafts do contain unripe spermatogenes and Sertoli cells. Oslund
(96) found germinal tissue in the indifferent stage present in cryp-
torchic testicles. Moore (92) and Oslund (97) were unable to con-
stitute hypertrophy of the interstitial tissue after vasectomy. Ceni
(22) reports a case of hyperplasia of interstitial cells with loss of
secondary sex characters in a rooster. The negative viewpoint is
summarized in the following statement by Kingsbury (61) : " Aside
from the fact that the interstitial cells look like gland cells, failure
to appreciate that the 'indifferent' cells composing the tubules of
the testes in the absence of any spermatogenic process are neverthe-
less parenchyma and a possible source of substances of endocrine
function is responsible for the acceptance of the interstitial gland
interpretation. As far as I can ascertain, in no instance has a
gonadal effect been demonstrated in the absence of parenchyma."

In addition to the changes in copulatory ability, noted above,
castration is followed by a significant diminution in spontaneous
activity. Hoskins (53) found that when male rats were gonadec-
tomized at the age of 70 days, the activity was reduced to about
one-fifth that of normal animals. Gans (37) showed, however, that
if the castration is performed during the first 12 days of life this
diminution does not follow. Hoskins (54) was unable to raise the

THE SEX DRIVE

273

activity level of his castrated animals by grafting- testes into them,
although he did succeed in increasing the activity of two senile
males by such transplants. Smith and Boukalik have demonstrated
a slight increase following implantations, by the Stanley syringe
method, of testicular substance from young males into castrated
and senile animals. Feeding of dessicated testis and subcutaneous
injections of testis extracts — saline, glycerine, and lipoid solvent
extracts and dialysates of several kinds — were tried by Brown
and his collaborators with no significant results. Gans and Hoskins
(38) showed that the gastrocnemius muscle of castrated rats is
heavier than that of normals, although the proportion to total body
weight was the same in both cases. The original strength of the
muscle was the same, but those of castrated animals fatigued more
quickly and, therefore, did less work. These authors state that their
results need not be ascribed to lessened muscular efficiency per se,
but are more probably due to inadequacy of supporting functions,
such as circulation or respiration. Athanasiu and Pezard (7)
found that the action current of the gastrocnemius muscle during
voluntary walking in capons is only 20 per cent of that in normal
cocks. Bukalik and Hoskins (16) gave repeated implantations of
fresh rat testicle substance to castrated and senile rats, but found
an increase in spontaneous activity in the case of only one senile.

Loewy and Richter (79) found the same decrease in basal metab-
olism in gonadectomized male dogs as has been reported above for
females. Tandler and Gross (135) report that the thyroid glands
of the Skopecs are extremely small. Bertschi (11) found no change
in respiratory exchange after castration or following subsequent
injections of testicular extracts. Korenchevsky (64) found the
gaseous metabolism unchanged after injections of testicular emul-
sions. Tsubura (139), on the other hand, reports a diminution of
respiratory exchange following gonadectomy; this effect could be
overcome by implantation of testes but not by feeding testicular
substance. Wheelon (147) reports a fall of blood-pressure after
castration and a decrease of blood-pressure reactions to nicotine ;
he interprets this result as showing that extirpation of the testicles
affects the sympathetic nervous system itself. In a later article
(148) the same author reports that these consequences of castration

274

THE SEX DRIVE

are partially overcome by implants of slices of testicles. Fichera
(33) found that after gonadectomy the pituitary enlarges to about
twice its normal size in many animals. Hatai (52) also found this
increase, but Livingston (75) reports that it is not an invariable
accompaniment of castration. Steinach and Kun (129) report that
sexual development could be hastened by injections of pituitary
extract into infantile rats ; the characteristics of eunuchoidism in
rats were overcome by the same treatment. Smith (117) obtained
similar results with transplantation of the pituitary gland but
found that the results were negative when the host had been pre-
viously castrated. Donaldson and Hatai (29) and Hatai (51) found
a diminution in the weight of the nervous system following extir-
pation of the male gonads.

Atrophy of the penis, the prostate gland, the seminal vesicles
and Cowper's gland,2 following castration, has been found by many
investigators (40, 72, 86, 112, 119). The seminal vesicles probably
have an external secretion (34), but the writer has found no evi-
dence of an internal secretory function of this organ. There is
some evidence indicating that the prostate gland elaborates a hor-
mone which passes into the blood stream. Serrelach and Pares
(113) found that ejaculations ceased after prostatectomy in dogs.
Ott and Scott (98) report that injections of prostatic extract are
followed by erections. Various psychical disturbances, including
neurasthenia, following extirpation of the prostate in man have
been interpreted as indicating an internal secretion of this gland
(41). Macht and Ulrich (84) found that prostatectomized white
rats learned a problem involving muscular coordination and effi-
ciency more slowly and less well than did normal animals. Hunt
(56) believes that testicular transplantation is more successful
when accompanied by injections of prostatic extract. Wallace
(142) and Griffiths (40) point out that the prostate exhibits sea-
sonal variations in those animals which pair only at certain times
of the year. Steinach (127) reports hypertrophy of the prostate
after vasectomy, the latter operation, according to this author, in-
creasing the internal secretion of the testes. Macht (83) found

2Nagel (95), however, reports that Cowper's gland does not undergo hypo-
plasia after gonadectomy.

THE SEX DRIVE

275

that feeding of prostate accelerates the metamorphosis of frog and
salamander larvae. Korenchevsky and Carr (65) state that injec-
tions of testicular emulsions caused a slight decrease in nitrogen
metabolism, whereas if prostate emulsion was injected at the same
time, there was an increase of as much as 11 per cent.

3) Hermaphroditism and sex reversal. In true hermaphroditism,
rare occurrence among mammals, the gonads of both sexes are
present in the same organism. One or the other sex may predomi-
nate in regard to behavior, although various degrees of bisexuality
are often present. The morphological sex characters are usually
mixed, in varying proportions. Sand (109, 111) has succeeded
in producing hermaphroditism experimentally, demonstrating that
testes and ovaries may survive in the same animal without inhibit-
ing each other in development or function. These hermaphrodites,
Sand reports (111), often showed rapid alternations in behavior,
changing from the character of a placid female to that of an ag-
gressive male within an hour. Lipscmitz and Vinals (74) trans-
planted ovaries into two non-castrated male guinea pigs; at a time
when their mammary glands were actively secreting, these animals
impregnated three normal females. Steinach (125) reports that if
ovaries are transplanted into normal males the behavior for the
first few weeks is definitely male but then passes to a periodical
alternation between that characteristic of male and that charac-
teristic of female.

In pseudo-hermaphroditism the gonads of only one sex are ap-
parently present, but the organism has also some of the secondary
characters of the other sex. Whether these cases arise as the result
of a perversion of some of the hormone-producing gonadal tissue or
whether there are present, in addition to the gonads of one sex,
some of the heterologous secreting cells, has not been definitely
determined (45). Steinach (126) has reported that in a number
of human cases, showing no marked abnormalities except homosex-
ual behavior, he found unusual elements in the testes, resembling,
in several respects, cells of the corpus luteum.

Steinach (121, 122, 123, 125) gonadectomized male and female
guinea pigs and rats and grafted into them the heterologous sex
glands. After a period of time these animals developed the mor-

276

THE SEX DRIVE

phological and behavior characteristics belonging to the sex of the
grafts. Moore (90, 91) has confirmed the possibility of feminizing
male and masculinizing female guinea pigs and rats. By the same
method Brandes (18) has produced sex reversals in deer, and
Guthrie (42), among others, in fowl. Lipschiitz (71) and others
have found that when a spayed female is masculinized, the clitoris
develops into a penis-like organ. Wang, Richter, and Guttmacher
(144) were able to raise markedly the spontaneous activity level of
castrated rats by transplanting ovaries into these animals; they
report that a number of these feminized males showed the rhyth-
mical variations in activity characteristic of normal females.

A few cases are reported in which the characteristic sex activity
of male animals has been facilitated by the female hormone. Stein-
ach (120) found that the clasping reflex of castrated male frogs
could be elicited by injections of triturated ovaries of females in
heat. Harms (46) and Meisenheimer (87) have confirmed this
finding.

II. Method and Procedure

A. Age and care of animals

This study is one of a series of investigations which together con-
stitute a comprehensvie survey of drive behavior in the white rat.
In all of these studies laboratory conditions and such factors as the
strain and age of animals used have been kept as much alike as pos-
sible. A complete description of such conditions is given in the
first monograph of the series (146, pp. 24-27) and will not be re-
peated at length here.

The animals used were albino rats from the "Experimental Col-
ony Strain ' ' of the Wistar Institute of Anatomy, Philadelphia. All
of the females, and all but 30 of the 168 males tested, were born at
the Wistar Institute and were kept there until at least two months
old. The other 30 males were of the first generation born of Wistar
animal parents in this laboratory. No animal was used which had
not been living in the laboratory for at least 35 days before the
test. The animals included in the two senile groups ranged in age
from 16 months 0 days to 16 months 22 days, averaging 16 months
9 days, at the time of the test. All other animals were approxi-

THE SEX DRIVE

277

mately 185 days old when tested, the range being from 175 to 196
days.

An over-sufficient supply of food (McCollum's mixture) was
kept in the cages at all times ; lettuce was given weekly. This diet
was never supplemented during the 35 days immediately preceding
the test, but prior to this time, especially in the case of the very
young animals and the seniles, a meat and bone powder, made into
a paste by adding boiling water, was given in small quantities every
7-10 days. An ample supply of fresh water was available at all
times.

B. Aparatus and procedure

(1) Apparatus. A complete description of the apparatus used
in this investigation is given by Jenkins, Warner, and Warden
(60). The modified grid (146, pp. 28-31) and the improved con-
trol mechanism, used on the fifth preliminary crossing and de-
scribed by Jenkins (59, pp. 471-474), were used. An experimental
analysis of the most important features of the apparatus is given
by Warden and Nissen (145). For Figure 1 of this monograph
see Part I of this volume.

(2) Social factors. The animals were kept in the laboratory
for at least 35 days before being tested. There were never less than
5 and never more than 12 rats in one living cage ; the usual number
was 6 to 8. The cages were wire mesh, of 5,900 cubic inches ca-
pacity. We hoped to reduce to a minimum, by this precedure, the
purely social or gregarious drive to other animals, regardless of
sex. That we were successful in this attempt is shown by the work of
Jenkins (59), who found that the number of crossings of males
to a male incentive, and of females to a receptive female incentive,
is not significantly different from the number of crossings to an
empty incentive compartment.

Thirty-five days before the test the sexes were segregated; that
is, for about a month preceding the test, males and females were
kept in separate cages. Warner (146, pp. 25-26) has given the
practical reasons for adopting this procedure, and Jenkins (59)
has shown that this period of segregation does not materially affect
the results. Before the time of segregation the two sexes were

278

THE SEX DRIVE

present in about equal numbers in each cage. This means that for
about three months following the average age of puberty (see
Stone, 133) all of our animals had what is probably normal op-
portunity for sex activity. In the case of the two senile groups this
period of normal sex life was, of course, much longer.

(3) Incentive animals. The incentive for the female test ani-
mals was always a sexually aggressive male. In order to keep the
incentive for the females as constant as possible, only four different
males were ever used for this purpose. These males were chosen for
uniformity of behavior and for sexual vigor. The writer has found
that a receptive female3 is usually considerably more active than
the average male, so that in order reasonably to equate the incen-
tives for the two sexes it was necessary to select especially ener-
getic incentive males. When a number of females were being tested
in succession it was often found advisable to change incentive ani-
mals occasionally, since these rats would sometimes become sluggish
after being in the apparatus for several hours. It was found that
allowing the incentive male to copulate a few times, outside the ap-
paratus, would often revive his activity. Before beginning a test,
the incentive animal was always given about 20 minutes to adapt
itself to the apparatus; it would then direct its entire activity to
the test animal when the latter was introduced.

The incentive for all male test animals was a receptive female.
Since 60 or more animals were usually available for this purpose,
there was, with occasional exceptions, not much difficulty in secur-
ing two or more females who were in heat. A female was used as
incentive only after it had copulated several times with at least two
males. When two or more tests were made in succession, the female
incentive animal was taken out of the apparatus between tests and
given opportunity for copulation. If this activity did not occur in
typical and characteristic fashion (indicating that the animal was
passing out of oestrum), it was discarded and another incentive
female substituted. The individual differences in mating behavior
on the part of receptive females are greater than would be supposed

3 Throughout the paper the terms "receptive female/' "female in heat,"
"female in cornified stage," and "female in oestrum" are used interchange-
ably.

THE SEX DRIVE

279

from cursory observation (see page 281). Care was taken to select
always those females which showed the most typical and what
seemed to be the most provocative pattern of behavior. In this con-
nection it should be noted that the incentive animals used for the
first 61 males tested (see page 281) were about 1 year old whereas
the incentive females used with the other males were from 3£ to 9
months old. Although all incentive females used in the study were
active and vigorous, showing the typical pattern of copulatory ac-
tivity, it is not impossible that the difference in age mentioned had
some slight effect on the results, probably in the direction of re-
ducing the scores of the first 61 test males.

(4) Control of normal physiological variations in the female.
Warner (146) has found that the contrectation drive is at its maxi-
mum at the time when the vaginal mucosa is composed exclusively
of cornified cells. It is obvious that the effects of our control oper-
ation (laparotomy without extirpation of glands) would be most
apparent if the animals were tested during oestrum. In order to
determine when these females were in the cornified stage of the
oestrous cycle, it was necessary to rely on behavior criteria not
involving the use of a male animal. It would have been easier to
select the animals on the basis of their reactions to a male, but such
a procedure would have been, in effect, the interpolation of a
period of sex stimulation before the test, which would have made
our results non-comparable with those of Warner and Jenkins.
Another possible and very definite basis of selection would have
been the taking of a vaginal smear before the test, but this would
have involved the danger of interfering with the normal expression
of the sex drive during the test period immediately following.
Careful observation of cage behavior, therefore, formed the sole
criterion for selecting these females, but immediately after the test
period a vaginal smear was taken from each animal. These smears
were stained with haemotoxylin and eosin and permanently
mounted in balsam; they are on file and open to inspection at the
Columbia Laboratory. The results of any animal whose smear did
not show the typical picture of cornification, with absence of all
other types of cells, were discarded.

The test females which were spayed and injected with placenta

280

THE SEX DRIVE

extract were uniformly tested 48 hours after the second of two in-
jections, preliminary work having shown that with this procedure
the animals were quite regularly in oestrum at the time of the test.
Vaginal smears as above were taken immediately following the test
period, but the requirement of complete absence of all elements
other than cornified cells was not as rigidly adhered to here as in
the control group. Three of the smears show a few leucocytes and
epithelial cells, although in all cases cornified cells predominate
overwhelmingly. The writer would have eliminated the records of
the animals giving these slightly atypical smears had it not been
found that such elimination would not have changed the results in
any significant way. Each animal of this group was placed in a
cage with a sexually aggressive male a short time (10 to 60 min-
utes) after the test; in every case copulations took place within 8
minutes.

The spayed, non-injected females show no cyclic variations in the
histological character of the vaginal mucosa. It was found that the
smears taken from these animals conform closely to the following
description given in a private communication to the writer by Pro-
fessor J. A. Long: " After ovariotomy and the subsequent cessa-
tion of the cycle the vaginal smear picture is that of the interval,
that is, the cells are epithelial and leucocytes. As in the ordinary
interval there will be considerable variation in the proportion of
these two kinds of cells. 9 1

(5) Control of normal physiological variations in the male.
Spontaneous, rhythmical variations in the physiological conditions
underlying sex behavior, such as those found in the oestrous cycle
of the female, are apparently absent in the male. The " normal' '
factor determining the intensity of the drive in the latter seems to
be the length of time elapsed since the last previous period of sex
activity (146), provided, of course, that contributing factors, such
as age and diet, are kept constant. It has been found (146) that the
sex drive of the male albino rat, as measured by the Columbia Ob-
struction Method, is at its maximum 24 hours after a two-hour
period of copulatory activity. Therefore, in order to keep the fac-
tor of normal variation in the drive at a constant and optimum
level, every animal included in the male groups was given a two-

THE SEX DRIVE

281

hour period of opportunity for copulatory activity 24 hours be-
fore the test. The procedure was as follows: A receptive female
which had copulated several times with at least two indicator4
males was placed by itself in a small cage (15x8x6 inches) and
given 10 to 15 minutes in which to adapt itself to the new situa-
tion. At the end of that time the test male (one to be tested the
following night) was introduced into the small cage. Careful ob-
servation was made of the behavior occurring during the following
two hours, all copulations taking place being recorded. The latter
are summarized in Table 1. These mating periods always took
place between 7 :30 p.m. and 4 :30 a.m. ; during the summer months
they were never begun before 10 p.m.

Interesting individual differences came to light in the course of
these mating periods. Certain males were relatively passive and
could be 11 teased" into copulatory activity only after prolonged
activity of the female. A short little run with a sudden stop, often
followed by depression of the lumbar, and elevation of the sacral
regions of the back, was the most frequent and characteristic be-
havior pattern on the part of the female. Often she would nose the
male genitals repeatedly, and, if this failed to evoke a response, she
might straddle her hind legs over the back of the male and then
move forward, thus bringing her vulva into direct contact with the
nose of the male. On the whole, the females were far more active
than the males and would only rarely resist copulation. Such re-
sistance was sometimes induced when an especially virile male had
mated a large number of times in quick succession. Occasionally,
also, a male, after mounting, would maintain his hold on the flanks
of the female for an abnormally long time instead of releasing
her, as usual, a second or so after gaining intromission ; this often
elicited squeaking and a temporarily aggressive attitude on the
part of the female. Whether or not intromission continued during
the entire duration of these long embraces could not be ascertained

4 Indicator males were animals selected for exceptionally strong sex be-
havior; their function was to pick out receptive females with a minimum loss
of time. When placed in a cage of females, these males would manifest very-
little exploratory behavior, but would begin almost immediately to nose the
genital region of the females and would copulate as soon as a receptive animal
had been found.

282

THE SEX DRIVE

4> O TJ

o S g

••■a

i-4 ^

— M -4 1-1

3 £ 'a ja
" " S

©*/->© r nj C <u ^ j

0*^0"^0"~iOu^ou~iO"~iO>-oO>-riOvriOO'— i £ b0>«

l«rtiHNNWt»Ht^m>n\0*ONNOOMffiOvHHiH _ C fl (I) CJ

I i i i I I I I i i I I I I I i i i i I i I — '. V — - i- «» — «

'rt\OHVOHVOH<OHVOH\OHVOHVOHVOH^H\0 CT O 3 35 3

»-« _V O r" O

THE SEX DRIVE

283

with certainty. Such behavior on the part of the male usually ap-
peared at the end of a long-continued series of copulations, al-
though with certain individuals it was a regular occurrence. The
behavior of the males who copulated most often was usually quite
independent of the provocative activity of the female. These ani-
mals often mated 5 to 20 times in quick succession and then re-
mained quiescent for a period of 2 to 10 minutes. It was found
that some of the more passive males would not mate with one fe-
male although they would copulate with another ; this in spite of
the fact that both females were receptive and both mated freely and
typically with other males. The writer, therefore, made it a prac-
tice, when a given male had not copulated at all during the first 30
minutes of the period, to put a different female in the cage. Three
cases are recorded in which copulation followed such a substitu-
tion, although in all instances the number of matings occurring was
small.

Very striking was the behavior of many animals, both castrated
and non-gonadectomized, who seemed very eager to sniff at the
genital region of the female and to engage in all kinds of "sex
play" but who showed no attempts at copulation. This sort of be-
havior often continued, with intermissions, throughout the mating
interval. Similar observations are reported by Jenkins (59, pp.
479-480). Other males would partly mount the female without
actually copulating, this pattern of response being observed in ani-
mals which did not mate at all, as well as in those who, at other
times during the period, did copulate. At times it was difficult to
determine with certainty whether a real mating had occurred; for
the sake of uniformity in recording results, all of the following
criteria were required before an act was considered a true copula-
tion:

(1) The typical depression of the lumbar, and elevation of the
sacral parts of the back by the female.

(2) Complete mounting by the male so that intromission seemed
physically possible and probable.

(3) Licking of the penis by the male before another mounting
was attempted.

Any mounting which did not fulfil all three of these require-

284

THE SEX DRIVE

ments was recorded as an incomplete attempt and was not included
in the figures of Table 1, which shows only true copulations. It is
possible that licking of the penis does not follow every complete
mating, but it is such a frequent accompaniment of the other
criteria mentioned that the figures would not have been changed
significantly had this sign been disregarded.

A comparison of the scores made during the mating interval and
those obtained in the test period shows that there is no correlation
between these two sets of indices. This finding corresponds with
the results of other studies in which measures of both drives were
obtained from the same animals (59, 146). The only reasonable
interpretation of this fact seems to be that there is no correlation
between the contrectation and detumescence drives; that a male
may be strongly motivated by the stimulation afforded by an ani-
mal of the opposite sex without necessarily having a drive to copu-
late, and vice versa. In females the correlation between contrecta-
tion and detumescence drives is apparently very high (146).

(6) Test procedure. The precise technique of testing animals
by the Columbia Obstruction Method has been described in detail
in previous publications (146, 59) ; this technique was followed as
closely as possible in the present study. The test animal was allowed
to make four crossings from the entrance compartment to the incen-
tive chamber without receiving any shock in the interposed ob-
struction compartment. By this procedure the association " incen-
tive — other side ' ' was presumably formed. On the fifth or last
preliminary crossing, the circuit was closed while the animal was
still on the grid but after the door leading back to the entrance
compartment had been noiselessly closed. The animal was thus
forced to pass from the obstruction (shock) to the third compart-
ment where it came into contact with the incentive. For an ex-
tended discussion of the reasons for adopting this procedure as well
as for a complete description of the various precautions taken to
remove the experimenter from the situation and to avoid disturb-
ing the animal in any way, the reader is referred to the above-
mentioned monographs. Immediately after the last preliminary
crossing the test proper began. The test period was always 20 min-
utes in duration, and the behavior of the animal during this time

THE SEX DRIVE

285

was recorded in terms of approaches (the animal brings its head
into compartment B), contacts (the animal steps on the grid or
touches it with its nose but then retreats into the entrance compart-
ment), and crossings (the animal crosses the grid and reaches the
incentive ) .

The test animal was returned to the entrance compartment as
soon as there was any specific contact between it and the incentive
animal. This " specific contact" took several forms, the most com-
mon being nosing of the genital region, biting the neck of the other
animal, the short jerky run and sudden stopping on the part of the
female test animal, and the straddling movement which forcibly
brought the genitals of the one animal (male or female) into con-
tact with the nose of the other. If none of these criteria was real-
ized within 30 seconds after the test animal had reached the incen-
tive compartment, it was nevertheless returned at the end of that
time. (For reasons and justifications see 59, pp. 483-485.) It is clear,
at any rate, that the incentive used in this study is directed to ap-
peal primarily and immediately to the contrectation drive and only
secondarily and indirectly to the detumescence drive.

All animals were tested between the hours of 8:30 p.m. and 4
a.m. At no time was a test begun earlier than two hours after sun-
set. The tests were made at the following times of year :

Females

Control December, 1927

Spayed at 5 months December, 1927

Spayed at 5 months, inj.
placenta extract December and January, 1927-1928

Males

Control (20 animals) December, 1927

Control (7 animals) July, August, September, 1928

Castrated at 5 months (20 animals) December, 1927

Castrated at 5 months (5 animals) July, August, September, 1928

Castrated at 3 months September, October, November, De-
cember, 1928

Castrated at 3 months, inj. November, December, 1928
placenta extract

Castrated at 3 months, inj. December, January, 1928-1929
orchic 1, 2, and 3

Normal inj. placenta 6 hours December, January, 1927-1928
and 2-3 days

Seniles, normal and vasotomized August, September, 1928

286

THE SEX DRIVE

C. Operations and injections

(1) General conditions. All operations were performed with
general ether anaesthesia. Air was forced through liquid ether in
a jar, the vaporized ether passing through a tube to a rubber cone
which was held over the face of the animal. The animal was held
firmly in one hand until it became completely quiet. During the
operation just enough anaesthetic was used to keep the animal
motionless and relaxed. Injections were made without local or
general anesthesia.

Aseptic conditions were maintained as far as possible. All in-
struments were boiled in a sterilizer for at least 20 minutes. The
site of the incision was cleared of hair by the use of sodium sul-
phide, after which the part was thoroughly washed with alcohol.
After completion of the operation the wound was painted with
iodine. In the first few operations sterilized surgical silk was used
for ligating and sewing, but in the great majority of the operations
fine catgut, sterilized in a solution of corrosive sublimate, was em-
ployed.

(2) Ovariectomy. There are two general methods for perform-
ing this operation, both of which are frequently used. In the first, a
single incision is made along the ventral median line, the intestines
being pushed first to one side and then to the other, successively
exposing the right- and left-uterine horns and ovaries. Preliminary
work convinced the writer that this method involved unnecessary
manipulation of the intestines with its attendant danger of visceral
adhesions, and all animals whose results are included in this report
were spayed by the dorsal route. Two skin incisions, each about
one-half inch long, were made on either side of the dorsal median
line just above the location of the ovaries. Then a somewhat
shorter incision was made through the underlying musculature
into the peritoneal cavity, exposing a mass of fat in which the
ovary was imbedded. This fat was pulled out of the cavity with
forceps, and the ovary freed from the surrounding tissue. A liga-
ture was placed over the uterus, just below the end of the ovi-
duct, so that when the ovary was ablated the oviduct and about
one-sixteenth inch of the uterus were removed also. The fatty
tissue was then replaced into the cavity and both incisions were

THE SEX DRIVE

287

sewed up. In all cases both ovaries were removed. The animals
recovered quickly from the operation ; cage behavior seemed normal
within a day or two and at the time of the test, about 35 days later,
the wounds were hardly visible. At postmortem the only adhesions
found were a few between skin and underlying musculature; in-
fections or regenerations were never found. Considerable atrophy
of the uterus was noted 35 days after the operation.

In the control operations the technique was exactly the same as
that outlined above, except that after the ovary and its surround-
ing fat had been drawn out of the cavity no ligature was applied
and the glands were left uninjured.

(3) Castration. Two incisions were made parallel and lateral
to the median ventral line of the scrotum. The incision through the
integument was about three-eighths inch long and a second incision
through the tunica vaginalis or muscular sac containing the testi-
cles was of about the same length. Pressure of the thumb from the
abdomen caudally served to bring the testis out of its scrotal
cavity. The relatively frail mesentery by which the testis and its
appendages are suspended from the tunica was torn to a point just
cranial from the caput epididymis. Here a ligature was placed
round the entire spermatic cord (tying off the vas deferens, sper-
matic arteries and veins), the latter being sectioned caudally from
the ligature. In this way the entire testis together with all parts
of the epididymis were ablated. The incision in the tunica was not
closed, but the skin incision was sewed in such a way as to reduce
considerably the size of the scrotal cavity; this was done in
order to reduce the chances of a subsequent hernia. In all cases
both testes were extirpated. The animals were apparently not
seriously affected by the operation, copulations being observed in
several cases 12 hours later.

In the control operation the same technique was followed except
that the testes were not removed from the scrotal cavity nor other-
wise mutilated. Care was taken in sewing up the incisions not to
draw the skin so tightly as to cause cryptorchidism.

(4) Vasotomy. Two short incisions were made through the
skin just caudal from the inguinal canal. A slit was then made
through the tunica vaginalis, and the vas deferens was pulled out

288

THE SEX DRIVE

of the opening so produced. The vas was ligated in two places,
about one-half inch apart, and a piece one-fourth inch long was
excised from between the ligatures. Care was taken not to injure
the blood vessels. The incisions were closed with a few stitches.
In all cases of vasotomy both canals were treated as described.

(5) Injections. The materials for injection were prepared by-
Eli Lilly and Company and were given to the writer by Dr. G. N.
Papanicolaou, of the Cornell Medical College. All injections were
made subcutaneously, through the skin of the back. Care was
taken not to get the substance near the mammary glands or the
axillary lymphatics. When two or more injections were made into
the same animal, the needle was applied to different parts of the
skin each time.

The placental substance used was an oily preparation of ether-
soluble, lipoid extract of cow placentae. It had been previously
tested and found potent in producing oestrous phenomena in
spayed guinea pigs by Dr. Papanicolaou. The timing of these in-
jections into spayed females was determined by preliminary ex-
periments : one-half cc, 72 hours before the test, one cc, 48 hours
before the test. Under these conditions all but two animals of the
group were in oestrum when tested; the results of the two excep-
tions, of course, were excluded from the tabulated data. Deter-
mination of the oestrous stage was made by the vaginal-smear
method as described above.

A group of males castrated when three months old and a group
of normal, non-operated males were given injections of placental
extract under the same time relations as those described for fe-
males. A third group of males, non-castrated, was given one and
one-half cc. of the extract in one injection, six hours before the
test.

Three kinds of orchic extract were used, all of them being pre-
pared from the entire testis and epididymis of immature cattle.
The three types of extract are designated Orchic 1, Orchic 2, and
Orchic 3 ; below is given, in reversed order, a short description of
each:

Orchic 3 — The ether soluble, lipoid fraction of the testis-epididymis sub-
stance; oily preparation.

TEE SEX DRIVE

289

Orchic 2 — The first water soluble fraction made from residue remain-
ing after removal of Orchic 3. Light in color.

Orchic 1 — The water soluble residue remaining after removal of Or-
chics 2 and 3. Dark in color.

Three groups of castrated males were given orchic extracts, one
group for each type of extract. One-half of each group was given
the injections on the following schedule : one-half cc, 11 a.m. ;
three-fourths cc, 3 p.m. ; three-fourths cc, 6 p. m. ; all on the day of
the test. This arrangement was adopted because it was thought that
the testicular hormone exerted its effect rapidly and for a short
time only. However, after observing that there were no positive
results from this temporal arrangement of the injections, the fol-
lowing schedule was adopted for the other half of each group :
one-half cc, 48 hours before the test ; three-fourths cc, 24 hours
before the test; and three-fourths cc, at 11 a.m., on the day of
the test.

III. Results and Discussion

Of the three types of indices yielded by the Columbia Obstruc-
tion Method, approaches, contacts, and crossings, the latter is
clearly the most specifically indicative of the strength or intensity
of the drive. While it is probably true that an approach or a con-
tact shows a tendency in the animal to cross over to the incentive,
such behavior may be merely the manifestation of general activity.
A crossing, on the other hand, quite definitely manifests the opera-
tion of a specific drive. Most of the following discussion, there-
fore, will be based on scores in crossings.

A comparison of the figures of Tables 4 and 6 shows that there is
a high correlation between the average number of crossings and
persistence of the drive as indicated by the percentage of scores
occurring in each of the successive five-minute intervals. That is,
when the number of crossings is relatively high it will be found
that toward the end of the test period these scores are as frequent
as, or more frequent than they were at the beginning. On the other
hand, when the number of crossings of a group is low, only a small
proportion of the total number usually occurs during the last 5 to
10 minutes of the period. It will be seen that this general corre-

290

THE SEX DRIVE

« ,3

e-s g

3 O 3

i; 04 1

(» P «i

« o w
OuO

e ©

■*-> aj

53

t> 4) 4>

c c c
o o o
2Z2

s 1 5

<u 5 ^ ^

g.>aBS

° ° M O O

C o S

» * o^.S* «

<u aj ~ — fc« «->

C C c *- ? w w

o o o rt > > >

zzu-S.Soo

THE SEX DRIVE

291

TABLE 3

Frequency distribution covering approaches, contacts, and crossings of the

female groups

Group

Normal
cornified

Control

Spayed at
5 raos.

Spayed at 5 mos.,
inj. placenta ext.

° o
. a

o v>

a.
<

<

u

on

on

to

JS

CJ

on

an

U)

proa

ntact

ossin

proa

ntact

ossin

o

o

<

rj

rj

<

<->

u

1

2

6

2

5

2

3

1

4

3

7

2

2

1

2

6

1

2

1

1

1

1

1

2

1

1

1

1

2

1

2

1

0

1

2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24

spondence holds also for the male groups reported below, as well as
for the findings of other investigators using the same method (146,
59). In the persistence of the drive, therefore, a check on small
differences in averages is available ; if two groups differ in average
number of crossings by only a few points, and if, at the same time,
the group with the lower average also shows less persistence, the
disparity of the groups in the indicated direction is emphasized.

A. Results for females

In Table 2 are shown the special conditions pertaining to the
female groups. All operations were carried out when the animals

292

THE SEX DRIVE

<

UOIJBUBA JO
O0l}BIA3p
U0pBIA3p

dSBJdAy
nBtpaj^

OOUBUBA JO

VuaiDgjaoo

UOI}BIA9p
pjBpOBJS

aopBiAsp

9§BJ3Ay
33BJ3AV

93UB^
UOpBIJBA JO

^aaiDiyao^

aopBiAsp
pjBpiiB^s

aopBiAsp
dSBJdAy

aSUBtf

OBipaj^

a
a
o
u

o

vo
CO

U-v

to

o

!*

1-H

VO

SO

to

CM
CM

vn

vn

CM

vn

»-*

CM

OO

CM
O

On

r-.

to

CM

vn

vn
to

On

OO

*+•

to

i-h

to
,"H

CM

CM

to
CM

oo

O
CM

vn
■

CM

■

vn

o

CM

O

6

O

vn

vn

©

vo
»-h

CM

rH

oo

to

oo

CM

VO
00

to

to

ON

ON

to
On

CM

CM

CM

o

VO

o

to

O
to

CM

CM

rt-

O
to

OS

VO

to

CM

CM

CM

on oo »-< vn
i i i i i
o o o o o

O O O vn O
M* N Cvi H rH

to CM rt- to

CM CM rf- to
to vn vn oo vo

*vO WtvH

o o o o o

O O O vn O
vn vn vo vo i-h

a g

V 3

gj J3
— - en

O

o £ *>

c 2 o

I

4-. O

«« ca-

rt o f "9 <vt

JJ 2.2.J

Z Uc/5 CO Z

i

^3

53 80

Wz
H*6I

MJZl
V9I
VST

qm

VII
VOI
V6
V8
TO
V9
VS
V*
pj£
pui

a
a
o

o

»-« CM

O

1-H

VO

vn -h o

to

tO CM

CM OO
CM

CM

O

f>

1-H

vn CM O

+

O 00

oo o
to

to

vn

1-H

vo O CM

CM

O Ov

«n

CM

O

CM

to CM i-H

vn

O CM

ON VO
CM

CM

CM

00

VO CM O

O ft

so vn
to

CM

1-H

to

VO to 1-H

vn

O *H

rl- VO

<#■

fH

oo

vo f O

CM

tO H

CM

to o

o o

to

to

CM CM i-H

vn

*^ to

to On

1-H

CM

tO O 1-H

CM ^

ON

00

CM

00 rj- to

VO 00

to On

CM

1-H

to vn i-H

t>»

to CM

vn

CM

vn

1-H

OO tO t-H

oo oo

to oo

**

t-» tJ-

o

vo vo

to Is*

vn

to

o

VO ^H tO

00

CM CM

CM O

>— <

to

vn

CM

rj* to CM

vn

to o

CM CM

<=>

1-H

to

o

Tj- 1-H tO

VO

~« CM

vn

On

O O vo

VO

O vn

O

VO

00

vn

1-H

CM oo

to tJ-

CM
r-<

to

ON vo O

JS cnj3 <r> J3 enJS on

S s.s g tj.s s t.s gg.s

C"*5 Eeac'1 C"'" U™00

g-S o %G O %G O O.C O
O.OU. D.OW CI. o in D-oih

<uu <uu <uu <uu

Zo U CO vn CO vn CX

THE SEX DRIVE

293

c3 a

2.1

a a

so

<S 0)
O C

IS

so

©.a

fa

jaqiunu

33BJU3DJ3J

iaqoin^[

HO\0\ OO OO OS ONrJ-vO t^NO^
^« to CO © © CM »-•»-<© HON

CM CM rj; VONM CM © CM Os ©0

O^VO CM «/i »/i OO rt" OO <-4 rf^

CM CM CM CM —i CM ^ t-. O CM »-< CM

ro o -< >-l (N OO CO rf- N N (fl
Nh« N VO N CM wo

S5U4U33J3J v^Tt-* t^^w to ^ tX

<M CM «-« CM CM CM CM CO CM to to

»«2 £ ° *t CM t^. CM 5svOH
N H ^ OJhvo Mihh co CM

O <H N On ^ Ov ©0*0 ©r*»
t*i CM OS ©^i-5 Noool CO
tH»H to CM CO co to CM CM

J 9 Q LU tl M 0\ oo 0\ oo to o N ff\ oo to ^ oo

T N. ^. tocOCMtOrJ-^CMCM — ^5

<

a

3
O

o

cj w 6a o

2 o a 5

9 rS •= P

a o

in c- o u a o £ a

to

tS.e

S ©
O u

o

£ 8

c
o
U

a a. 2 c

GO to

i-i, — a.

O <u
© G

2 £
— -a

cj 3

T3 C g

*- o a» c

•O « **
CO -o

to OS CM
«*• rt- ~ OS
* On ^

OS © ^- SO

VO i-c to tO
" H CA H

I ill
fsf>:

(five'
-g

>s

T

ctj "o "o
C u, aj u

O O o. o

ZUcwU

294

THE SEX DRIVE

296

THE SEX DRIVE

were approximately 150 days old (range 143-157 days) ; the tem-
poral arrangements relating to the injections have been fully de-
scribed in the preceding section. The distribution of approaches,
contacts, and crossings for each of the groups is given in Table 3.
Table 4 shows the medians, ranges, averages, average and standard
deviations, and coefficients of variation of each group. The ap-
proaches, contacts, and crossings, as distributed in 1-minute inter-
vals during the 20-minute test period, are shown in Table 5. In
Table 6 are given the absolute and relative number of scores occur-
ring during each of the four successive five-minute intervals of the
test period. Table 7 shows the reliability of the differences between
the average number of crossings of the several groups. In Figure 2
is shown graphically the effect of the various conditions imposed,
on approaches, contacts, and crossings. In Figure 3 the percentages
of Table 6 are presented in diagrammatic form.

It is apparent that the control operation had only a slight effect
on the scores of the females. The actual difference in crossings be-
tween the control and normal5 groups is .69 and is statistically
quite unreliable. In variability the two groups are closely similar ;
in approaches and crossings the controls show somewhat higher
scores. The persistence of the drive is about the same in both
groups, over 25 per cent of the crossings occurring during the last
five minutes of the period in each case.

Ovariectomy (spaying) effects a large reduction in scores (Table
4 and Figure 2). In view of the fact that the oestrous cycle and
its related phenomena seem to be quite dependent on the ovarian
hormone, this result is, perhaps, not surprising. Nevertheless, the
outcome could not have been predicted with certainty, for although
we know that the female rat will copulate only during oestrum and
that oestrum does not occur in the absence of the ovarian hormone,
this would not necessarily imply that the contrectation drive also
disappeared after ovariectomy. The finding (146) that the sex
drive, as measured by the Obstruction Method, is practically non-
existent in normal females during dioestrum would not have made
prediction of the result any more possible, for there is reason to

6 The term 4< normal' ' is used throughout this report to designate non-
operated animals and is not meant to imply anything beyond that fact.

THE SEX DRIVE

297

believe (see Section I) that the ovaries produce two quite different
hormones ; the possibility was, therefore, given for a non-f ollicular
ovarian hormone which might have actively inhibited expression
of the contrectation drive during dioestrum. The present finding
rather definitely establishes the fact that the depressed sex activity
of the female during dioestrum can be attributed only to a slight
extent, if at all, to the inhibitory action of an ovarian hormone.

A comparison of the ovariectomized group with normal animals
in dioestrum shows a slightly greater number of crossings and a
considerably higher score in approaches and contacts for the for-
mer. Two interpretations for this finding present themselves. We
may, in the first place, consider the differences as expressive of a
higher level of general vitality and activity on the part of the
spayed group, and this, in turn, might be explained on the basis of
either one of two assumptions: (a) The recurrent phases of tre-
mendous activity accompanying oestrum may leave, in their wake,
periods of generally lowered physiological vigor, but these alter-
nations of great activity and subsequent relapse into inactivity
would not be found in spayed animals; (b) following elimination
of the ovarian influence there may be a hyperplasia of some other
gland, such as the thyroid, which raises the level of general activ-
ity. Although this hypothesis is very attractive it is inconsistent
with the work of Wang (143) and others showing that, after spay-
ing, spontaneous activity remains on the low level of dioestrum.
There is, furthermore, good evidence of lowered metabolism fol-
lowing ovariectomy (79).

The second interpretation is based on the consideration that ap-
proaches and contacts are really expressions of a specific drive
rather than merely of general vigor. Apparently the animals were
motivated quite strongly and were actively deterred from crossing
over to the incentive by the shock; it is not improbable that the
operation made these animals more sensitive to the obstruction.
These considerations imply, of course, that the sex drive of the
spayed rats was greater than that of normal animals in dioestrum.
As was pointed out above, it is not impossible that the ovary does
secrete a hormone during dioestrum, which has a slightly inhibitory
effect on the expression of the contrectation drive. A possible and

298

THE SEX DRIVE

more probable interpretation of such a difference would be that
normal animals exercise their sexual mechanisms to such an extent
during oestrum, stimulated by the follicular hormone, that these
mechanisms, especially the neutral, go through a subsequent phase
of fatigue or reduced activity. Such rhythmical recurrences of
great stimulation not occurring in ovariectomized rats, the mech-
anisms would remain on a medium level of activity, higher than
that of dioestrum and lower than that of oestrum in normal ani-
mals.

The injections of placental extract several days before the test
served to raise the number of the crossings of spayed females to
a point only 1.34 below that of the control group. In other words,
the injection of this extract overcame almost entirely the effect of
ovarian extirpation. The slight difference between the groups is
probably explained by the fact that all animals were given the
same amount of extract and were tested the same length of time
after the injections; possibly there are individual differences
(based on weight, for instance) which would vary the optimum con-
ditions in respect to dosage and time relations. The persistence of
the drive in the injected group is slightly lower than that in the
control group, 24.3 per cent of the total number of crossings oc-
curring during the last five minutes of the period. The extract
used was obtained from cows, illustrating again (see Section I) the
species non-specificity of the hormone.

B. Results for males

In Table 8 is given a summary of the male groups, comprising
168 animals ; all of these rats with the exception of the two senile
groups were tested when 185 days old. Table 9 gives the distribu-
tion of approaches, contacts, and crossings for each group. Medians,
ranges, averages, average and standard deviations, and coefficients
of variation are given in Table 10. The approaches, contacts, and
crossings, as distributed in 1-minute intervals during the 20-minute
test period, are shown in Table 11. In Table 12 are shown the
absolute and relative number of scores occurring during each of
the four successive 5-minute intervals of the test period. Table 13
shows the reliability of the differences between the averages of the

THE SEX DRIVE 299

o

<u

T3
N

0> V

O o

X; XI

O

« r* a

u g o i i

K « rt W x x

C i: «- >2 « <u

— K X K _ _

S W *» W rtrt

c o o o c c

s c

o o

v u v v

a „
o o

Cu

o

3 I

W) «
c o.

3 X
3 «U

C eo en « v>

m a c c c c
_ o o o o

So
a >
o .c c

c c
o o

s e

o

v© ~
<-< C

s n d n n w _

c c

o o

u u u. u,

tf) {fi (O ffl

C C *-
O O 3

ZU-°UUU<J

S »

« rt O

« t» 4* o
(f> <r> C C C y>
« O O O t«

UUZZZ>

300

THE SEX DRIVE

a 0
a o

<n 2

.3 «

e? 2
U 5

U E

J O 9

cj S "3,

■3 x
6.2

dncuo.

s3uissoj3
esqoBoaddy

63UI8SOJ3
6J0BJUO3

saqoccuddy
s3ni880.t3

saqOBO-tddy

8JDBJU03

6oqOBOjddv
sSuissojo,

6J0BJUO3

eaqoEO.tddy
s2urssoi3

8J0BJUO3

saqoccuddy

6SUI860J3
8JDBJO03

6aq3BOJddy
e9ui88o.i3

SJ3BJU03

saqoBoiddy
6Sni380J3

SJDBJU03

6aqoBO.tddy
83ntssoJ3

SJDBJUO^

saqoFOJddy
sSaisso.io,

SJOBJUO^

soqoDOJddy

SJDBIUOQ

eoqoBOJddy

60SUOds3J

jo -on

CM _| rl

HMN^N CM W h

NMNN i-i
CM CM t*- CM CM CM <mi
i-4 i-H i-4 -H n i-H
VO VO N H «-i

•H ^ M *4 «-« f4 f* ■* <M

«M to CM

N H H V> 1-4 t-4

<-H i-l N H H rH

to hhn CM t-»

CM T-l CM ~ *H H ^
t-fv* t>4 th CM CM ,h

CM HH^- 1-4 fH

CM CM »-« CM ii i-< i-i

CM HHM H H —4

»»« tO 1-4 1-H CM 1-4 1-4

M M H CM 1

1-4 —I i— I CM i— l CM -I H H N INN

Tj- 1-4 tO

NMrtNH «NN — >

CM MNHH^H-HH CM «»H

to ij- «o to CM CM ih

CM CM to f— I CM <M — p* n i4

i CM CM CM CM <N *•* to to rj> wm

CM rhto^-^wcM'-tt-'—i pH r-i

«H »-« »-t »h n i-i rj- n CM CM CM to CM CM »h Ni4

00 NO CM —« to

to— (Tl-vOT-trhCMr-f to i-4 t-h

»-« »H CM CM ^ CM * to »h CM — <

\£) ^ N -H fN
+ VOMi-^rt 1-4

O-lNM'j-tfMOSOO^O'HNMrHrtVONoOCAO'-N
HH14HHHHHHHNNN

THE SEX DRIVE

301

UOIJE ti E A JO

juapyjaoa
uotjBiAap
piBpuejs

UOIJBIAap

a3BJ9Ay

33BJ3Ay

33UB^

UBipajA]

UOIJBIJBA JO

jua'pujao^
aoijBtAap

pjBpUB}§

UOIJBIAap
33BJ3AV

33BJ3AV

33ub^
UBipap^

UOIJBIJBA JO

iu3'puj303

UOI}BIA3p
piBpUBJS

UOI}BIA3p
9SBJ3AV

33BJ3AV

3§ub^
OBipap^

a

3

2

o

O M <t M N OO Os.jn <i- q n

uisOOOcoOOOOtJ-oOI^

^^CM^cocsj»*st>.©ssq»osq

O^HOu"»OOOOcot^CMOsu">
CMCMCM— i»-hCM»-H'-**-hCMi-it^

tJ-OOOCM'-'OO-hCOOO
OOOOOOOOOOOO

O N + "i M <fl N h vo N
OhM?ihOCiOMOO»N +

OsoOOOOOOOOOSO

OOOOOOOOOOOO

oooooo»^«^w^ooo

o o o o o o

o

Os

00

o

»-<

OS

CM

~ 4-

os

oo
°^

SO

o

o

OS

OS
CO

00 o

CO

sn

CO

CO

Csj CM

oo

CM

oo

o

o
°5

V/S
CO

00

CM so

co

co

to

<N

cm

CM

o

O

o

SO

o

CO

O

ON

CO

CO O
i-l 00

vX

so

CO CM

ON

CO

CM

O

CM O O O

o q
co

oo";oia

so co co

inqwo
so to CM CM

t-4 CM

o

™ o o
°* o o

co ^

!3 •*> co
^CM

sac

r en w

2 o o

S S S

co ro co

.|_> +J

C3 CJ C3

tn m v> v>

to a c >-< </i
^ e a

sftnaanaawsis

fl o o

302

THE SEX DRIVE

S

to

to
a

I

Wx

WX
H*9X
Wx

MUX
WX
MUX
Wl

mi

q*9
■W

vn

5SI

O N N W ^ to VO

Nrtxri \ONw vONO\ N»hih m t^CSI <~i

H 'H N O u-»M « u^NN NNN - to

O O CM i— iCM»— i Or- iO to CM to to 0\ h h N

h N S N N ^ W V3 * N ^ rt vO*m^Csj

NNO "O w © u-v rt- to u*i c\| f<>
©to»oto«.«CM©to©\'<i"i~ivOto O CM

NOvfl WNN N s VONM M OO to »h

_ — CM CM oo

wNo ^>«^ >*iAOO N-<vfl n «-« CM
<M O * vO<tNO*NwOviN N N

N«»< ff**N 60M» - V* to

<fMN NWp< *>^00 OOmN « to

vO^ts t\tO»N C(W« MNN M vO^^NTj-

»C N os to ^ O N « 00 to SO r+- w oo N iH

mNvO NNfr, wins <rt On vO M N

— ' m N w vO CO O N 00 vOvrt^

CM O^t^ N N S OnvOOn _ O N vfl 00

to N 00 CM OO 00 <0 rj- vr»\rvO

« ™ Cl - - ~ ^ "

uificCoiooCu/ttuxaujibiluMct

Y. W f> W c« il n u ra u " f l 15 »l

c o 5« c o

<UU <UU <UU <UU <UU <UU

E *3 +2 i; u- m t- o
o o ca_w «« c

2 U Urt UraUw"aUtoo

THE SEX DRIVE

303

q*oz

3jnuip^

qi/i

VSI

3jnuip\[

qm

H'£I
MUI

qioi
qi6

q*2

qJ9

q>s

qit
P-i£

5SI

3

O

o

CM CM CO CM CM

sO CM — • u->

— i h m N i—i CM ^- CO oo CM CO u-i
^ CM i— < CM CM ^ ^t- CM CM ^ CM CO

r*- u-> CM CM i-<

CM M Nnifl

CM^h^h n m w rt s ifl vn

CM

CM — < w»>

CM CM CO

^ CO N D N -f- O Nf< <f N W

—i — CM CM rt- CM CM »0 -« w-i

CO CO t-I CO CM CO SO CO lO VO CM T- I CM

CO co CM co-i-fotoco^coCM CO — < ^

to pn i-. t*- io ^^CMOO^-CM^M 00

Oirm «hO -tN + + +sO W CO SO

hhM «NS -h tJ- vo N N ^- CM CO CM »0

NhwhhNvOia^ CM ^ so CM OO — <r*-CM

SO t*- u-i uo CO r^- CO OO CM OS i-l CM O CO ^ OO

(7i M CO CM ^O CO CO SO CO On N N N M OO

u*><4-ts% r4 N >fl N W h Oi N CO CM CM t^.

iflMN iahN v^NN m O CM ^ CO co so O

CJ 05

11

C w o C «tJC w o G « o G «?

•s 2 «~ 2 2 2

a o t, a. o *

a. c

D. o

lh a- o c c- o u a. o i- d- o u «

u <uu <uu <uo <uu <

CJ {fl

c o
o »-

T3 jn'S "U

SSl

T3

N
1

aT P

05 " . co-- . — — —

O « .5 U « .5 Z a so Z ccm co c

cu cJ
CO >

304

THE SEX DRIVE

i

s3j

3

2S

©

jaquinu |ejojl

dSE)U93J9J
jaqiun\

a

3
O

o

so © Os
CM

Omo

~i tO SO

«« * o

^1 °i ^

00 OS fs)
rt ffi «

ff> ff» \fl

CM WV 00
CM

M 1H »A
1-1 CM

On OO so
On O

~« CM CM

00

O w
NNN

csj »-; r>.

>0 Is N
CM

w ifl <n
to ©
ii CM

oo so o
rst cm

r*. cm ws

O CM O
CM CM CM

© * Os
to ^ CM
CM CM CM

© ts
sOwO

00 «A OS
CM OS to

OO to Hfjifl SO CM O

© rj- © r>» tN. so

Os O Is*
O so

v» h to ts O N

SO W» ©

>rt NO N
CM CM CM

O to so

r>C ^ to

© f-4 tO

© OO Os

Tt" S- tO

OO to tO to CM 00

si

w> jS an w
M y g M u

O

Co r* c o

c o

o o

g* g S g* c b
com cq

u com Com Com o. o m com

y <uu <uu <uu <uu <uu

Ss

e

U vs

T3 T3 C «o

I S S S8 8

U to U to c u

THE SEX DRIVE

305

-2

c

Is

jaqmnu ibjojl

jaquun^

jaqtun^
aSBjuaojaj
jaqiunfsj

a

3
O
hi

a

ON 00 Oj
00*0 00

to © ©
no © o\

Vi-i ^- Tf N O *C

t> oo

O to *N

HMO*

«Hifl OWN

CM fx CM

cm ^ cm

vo ^
N N M

th N

O

^ ^ 04

O

CM T*" CM

o o oo
d *4 t»;

h ^ oo
CM CM

to o o o

CM N ©0 \0 M to O to to

H « 08

CM CM

q ^
n© cm cm

CO w> rf-

vo q vo
no ©

CM to
no on
CM — i CM

NOtsW
00 NO M
CM no ri-

hmm OO O O CM**-»0

u

03

a a
a o

•5 S

en £

&») JS
bfi cj

•5 g

O 5" C O

C- o u CI. o u Cu o u C o in U O h Q- © u,

<uu <uu <uu <uu <uu <uu

« tj C

Cu O

C w 1/3
a o

CJ so bX)

01 cn
O

a o

cd

T3

T)

!s

u

CM

cj

to

id

u

en
O

cd

o

ast

E

rch

ast

E

rch

U

to

0

U

CO

o

_* 03 cn

_ >> — s

2 C « <U 03 <u O

E oj-o — g o

O rt to '£ £ 'S £

4» ~ i <u O oj cd

Z O.CM C/D C C/j >

THE SEX DRIVE

hi
HI

1

13

1

0

I

aasssssassssssa

« h N m

im

Hill tell

h M W M

S3 3 SSS

liilliiiii

oob««ofl««o o o q. o 4)

THE SEX DRIVE

307

several groups. Figure 4 shows graphically the effect of the various
conditions imposed on the average scores. Figure 5 presents the
percentages of Table 12 in diagrammatic form.

It appears from consideration of the average number of cross-
ings that the control operation had a marked effect on the animals
of this group ; there is a difference of almost three points between
the scores of the normal and control groups. This finding is espe-
cially surprising in view of the fact that the control operation per-
formed on the females, which is apparently much more severe than
is the corresponding male operation, caused only a very slight
diminution in crossings. It is true, of course, that the scrotal
wound is subjected to more friction and other irritations than is
the dorsal wound in the female. Thus the males may have become
more sensitive to traumatic stimuli of all kinds, and especially to
those mediated through the scrotum; it seems quite possible that
the animals felt the electric shock in the apparatus through this
part of the body. Such an interpretation of the decreased num-
ber of crossings by the control males gains support from the fact
that the approaches and contacts of this group are more numerous
than are those of the normal animals ; it may be supposed that the
drive was equally great in both groups but that in virtue of the
scrotal wound the physically constant shock had a more intense
effect on the controls. Futhermore, it is seen from Figure 5 that
the diminished number of crossings is not paralleled, as usual, by
less persistence. In view of what has been said above (page 289),
this fact mitigates the difference between the two groups.

There is, however, another possible explanation for the relatively
low score of the controls. It was originally planned to have 20
animals each in the control group and in the group castrated one
month before the test. The unexpected result in regard to the for-
mer group caused the writer to add 7 animals to the controls and
5 animals to the second group ; these 12 animals were tested in the
summer of 1928. The average of the first 7 animals, computed
separately, is 12.0 (compared to an average of 10.56 for the entire
27 rats) , and the average of the 5 animals added to the ' 1 castrated
at 5 months" group is .6 higher than the average for the entire
group of 25. This suggests that there was some factor operative in

THE SEX DRIVE

309

312

THE SEX DRIVE

the winter of 1927 which tended to reduce the crossings of animals
tested at that time. Two such factors may be suggested: (1) the
season of the year. It is well known that in the wild state, rodents,
like many other animals, considerably reduce or suspend altogether
their breeding activities during the winter months. Rats housed in
an evenly heated laboratory, however, continue to reproduce
throughout the year. It has been noted in this laboratory that
litters are born about as frequently in the winter as in summer.
Donaldson says in this connection: " Constancy of (moderate) tem-
perature is in general favorable to continuous reproduction. " It is
not beyond the bounds of the possible, however, that a seasonal
rhythm persists in laboratory animals, this periodicity finding ex-
pression not in the more fundamental detumescence drive but
manifested in the contrectation drive. (2) The female incentive
animals used in the winter of 1927 were older than those employed
later (see page 279). Although we have no objective way of
knowing whether this greater age was accompanied also by lowered
activity and less "provocative" behavior, this may well have been
the case. The writer is quite firmly convinced that there was, actu-
ally, a general factor operative during the tests made in December
and January, 1927-1928, tending to reduce the number of crossings,
and believes that this factor is to be found in the nature of the
incentive females used at that time. This means, of course, that the
scores of all male groups tested at this time (see page 279) would
ordinarily have been a few points higher.

Castration one month before the test served to reduce the cross-
ings 2.36 points below the average of the control group. Although
this difference is not entirely reliable statistically, the indicated
tendency is substantiated by the decreased persistence of the drive
as shown in Figure 5 and Table 12. Castration two months earlier,
i.e., at the age of three months, resulted in a further reduction of
2.1 points in the average number of crossings ; the persistence of the
drive is low, as shown by the regular decrease in crossings during
each successive five-minute interval. It appears, therefore, that
castration has a detrimental effect on the strength or intensity of
the contrectation drive and that this effect is greater the longer the
interval between the operation and the test. It should be noted,

THE SEX DRIVE

313

however, that extirpation of the gonads does not reduce the drive
as quickly and completely in the male as it does in the female.
Table 1 shows that copulations also decrease in number, but do not
entirely cease, after castration. The latter finding is in accordance
with the results reported by Steinach (119), Stone (134), and
others.

Six possible explanations for the slow extinction of the male sex
drive after gonadectomy may be suggested :

1) The stimulating hormone may persist in the blood stream
for a long period of time, becoming gradually less and less effica-
cious. This hypothesis is made less probable by studies showing that
the effects of injected testicular extract are very transitory.

2) Castration diminishes the stimulation of other glands,
especially of the thyroid; the decrease in sex drive is only one
manifestation of a generally lowered activity level following thyroid
hypo-function. This hypothesis seems very improbable in view of
the fact that other performances, also depending on the general
activity of the organism, do not seem to suffer nearly as much as
the sex drive in particular.

3) The testicular hormone originally stimulates sex behavior,
but after the nervous system has mediated such activity for a long
period under its influence, the nervous patterns involved continue
to function for a time without the facilitation of the hormone. This
is probably the most difficult of all the hypotheses to test and can be
accepted only after the others have been proven inadequate.

4) The hormone stimulating the nerve centers controlling sex
behavior issues from some gland other than the testis, but this
' 'other gland" is in hormonal or other physiological dependence
on the gonads. The prostate, composed of several parts and
atrophying after castration, may be suggested as the source of the
directly efficient hormone. Several lines of evidence, both for and
against this view, have been briefly mentioned in Section I.

5) Tarchanoff (137) suggested that the clasping reflex in the
male frog is stimulated (or facilitated) by the afferent impulses
initiated in the distending seminal vesicles. Allport (5, p. 69) has
applied this hypothesis to the human male and states that in man
"the original stimulus for sex responses ... is the gradual dis-

314

THE SEX DRIVE

tension of the seminal vesicles, a condition requiring a fairly
periodic discharge of their contents. ' ' This explanation is attractive
in several respects, but is not consistent with the experimental
findings of Steinach (120) who has shown that the clasping reflex
is present in the frog before the vesicles are distended with fluid,
and that extirpation of the gland does not result in disappearance
of the reflex. The same author reports that removal of the seminal
vesicles has no effect on sex behavior in rats.

6) The writer suggests the following generalized statement of
the fundamental principle on which the hypothesis of Tarchanoff
and of Allport is based.6 Some organ, other than the gonads, but in
structural and functional dependence on the testes, effects a tumes-
cence or tension in itself or in another tissue or organ ; this tension
or tumescence initiates the afferent impulses which stimulate sex
activity. This hypothesis shares, with the one outlined in (5) above,
the advantage of supposing that the stimulation of the central
nervous system is in the form of sensory nervous impulses. It is
true that certain hormones as well as various foreign chemical
agents (narcotics, stimulants, poisons) are known to exert a selec-
tive influence on different parts of the nervous system, but quite
generally this selectivity follows the lines of more or less well-
defined neuroanatomical divisions. While it is quite impossible to
disprove the possibility of an hormonal action affecting only those
parts of the nervous system which together control a given behavior
pattern, it would appear equally or more probable that the nervous
system is directly affected through afferent channels, mechanically
stimulated. Any attempt at particularizing our hypothesis is clearly
a speculative procedure, but since it may prove directive or provoc-
ative of further research, we may suggest the following : The penis
is the organ whose tumescence or turgor initiates the afferent
impulses in question. It will be recalled that reports of the " black
eunuchs" of the Orient, who have been deprived of both testes and
penis, indicate that the latter is of considerable influence on the sex
drive. Since it is known that the penis atrophies after castration,
two alternative explanations of the sex-stimulating mechanism pre-
sent themselves: (1) The normally developed intromissive organ

6 Compare, also, the similar viewpoint of Havelock Ellis (30).

THE SEX DRIVE

315

responds to generally accelerated physiological vigor or to specific
erotic exteroceptive stimuli by turgescing. After gonadectomy,
this organ, having depended for its full development on the testicu-
lar hormone, becomes gradually less capable of tumescing and so
of initiating the sensory impulses resulting in sex behavior. (2)
Some other gland, such as the' prostate, which maintains its normal
structure and function in virtue of the gonadal influence stimulates
the tumescence of the penis.7

The complications arising out of the fact that sex behavior
apparently is often initiated by an external stimulus must be
"explained," provisionally, by the laws of learning or by the
assumption that this form of activity occurs normally in response
to a proper combination of external and internal stimuli.

Placental extract injected two and three days before the test
into castrated males caused an increase of 3.23 points in average
number of crossings as compared to the score of castrated non-
injected animals. As shown in Figure 5, the persistence of the
drive, also, was greater in the injected group. Since it has been
shown (144) that transplantation of ovaries into castrated male
rats raises the spontaneous activity level of such animals consider-
ably above the pre-grafting level, the present finding may be inter-
preted as the manifestation of augmented motility produced by the
female hormone. This explanation is possible in spite of the fact
that crossings are indicative of a definite drive rather than of non-
specific random movement, for it is obvious that the measure of any
specific drive includes always a certain value contributed by the
general activity level of the test animal.

If, on the other hand, we wish to interpret the above result as
indicative of a facilitating effect of the female hormone on the
male sex drive, a synthesis of the results and theories of other inves-
tigators and the reconciliation of a number of apparently contra-
dictory facts will have to be attempted. The resulting viewpoint is
confessedly speculative but has the virtue of correlating a con-

7 Professor E. S. Woodworth has suggested that if we consider the stimula-
tion for penis turgescence as being primarily external (mediated through the
exteroceptors), the function of the prostate — or its equivalent in this scheme
— may be that of raising the irritability or lowering the threshold of the re-
ceptors in this organ.

316

THE SEX DRIVE

siderable array of otherwise disconnected data. The points to be
considered may be listed as follows :

(1) From what has already been said, it appears very probable
that the testicular hormone does not directly influence the nervous
system but exerts this effect through an interpolated mechanism.
The arguments in favor of this viewpoint have already been dis-
cussed (page 313) and need not be repeated here. That the
ovaries also control sex behavior indirectly seems probable when
we consider the fact that the follicular hormone is present in the
blood stream for some time preceding oestrous (77) and that injec-
tions of placental extract manifest their action several days later,
rather than immediately.

2) The interpolated mechanism is undoubtedly under the con-
trol of the internal gonadal secretions. All other secondary sex
characters depending on the sex glands for their existence, there
is no reason for supposing that these particular mechanisms form
an exception.

3) The soma is apparently equipotential as regards its secondary
sex characteristics. The work of Steinach (123) and others in
experimentally producing sex reversals demonstrates clearly that
the gonads determine the distinguishing sex structures.

4) It follows, then, that the specific sex behavior expressed by
an animal is under the direct control not of the gonads themselves
but of the interpolated mechanism above mentioned. The sex glands
determine sex behavior indirectly by dictating the nature of the
development of the mediating mechanism.

5) There is no crucial evidence, so far as the writer is aware,
against the assumption that the activating function of the gonads
is non-specific as to sex. Halban's theory of the complete non-
specificity of the gonads is no longer tenable ; the present suggestion
is merely that a given structure or mechanism — whose function is
sex-specific — may be stimulated to its characteristic activity by
either the ovarian or testicular hormone.

6) In the work on sex reversals and experimental hermaphro-
ditism with alternations of male and female behavior hitherto
reported, the effects were always produced by a transplanted gland
or by a long-continued series of injections. Opportunity was

THE SEX DRIVE

317

thereby given for the heterologous sex hormone to develop its sex-
specific interpolated structure, through which the gonadal stimu-
lation effected its corresponding behavior.

7) In the present work the mediating mechanism of the castrated
males probably atrophied to some extent, but the corresponding
structure of the opposite sex was not stimulated to development.
When, three months after castration, the placental extract was
injected, it exerted its stimulating, activating influence on the only
mechanism present, that of the male.

To summarize : The gonads exert their influence on sex behavior
through the mediation of an interpolated mechanism8 which is sex-
specific in its function but whose structure is in the first place
determined by the specificity of the gonadal hormone ; this mecha-
nism may be stimulated to its characteristic function by either sex-
gland.

Two groups of normal (uncastrated) males were given injections
of placental extract. If we compare the results of these groups
with normal animals, it appears that one injection six hours before
the test diminished the drive, whereas two injections, given several
days before the test, had no appreciable effect. The former result
might be attributed either to an emotional disturbance resulting
from the injection a few hours before the test or to a rapidly
effective inhibitory influence of the female hormone on the male
drive.

In view of what has been said (pages 278-279) of the incentive
animals used when these two groups were tested, and especially if
the evidence of a general factor operating to diminish crossings at
this time is considered, it seems more fair to compare the results
of the groups with those of the control animals. If this is done, it is
seen that the injection six hours before the test caused a slight
reduction in the intensity of the drive, whereas injections two and
three days preceding the test produced an increase in the number

s Such a mechanism for the male has been already suggested (pages 313-
14); testis hormone (prostrate function) — penis tumescence — afferent im-
pulses stimulating sex behavior. In view of the more immediate dependence of
female behavior on gonadal influence, we may very tentatively propose the fol-
lowing as the female mechanism : follicular hormone — tension in uterine or
vaginal epithelium — sensory impulses initiating sex behavior.

318

THE SEX DRIVE

of crossings. Thus the findings on the normal, injected males corrob-
orate the view presented that either male or female hormone can
activate the mechanism stimulating the sex behavior particular to
that mechanism, and that the hormonal influence is indirect, requir-
ing several days before it manifests itself.

The or chic extracts injected into castrated males apparently had
no material effect on the drive, as shown either in crossings or in
persistence. Each group tested with a given orchic extract was
divided into two sub-groups, 5 animals receiving the entire quantity
of extract on the day of the test, the other 5 receiving injections
beginning 48 hours before the test. To see whether the temporal
arrangement of the injections had any effect, averages and indices
of variation and persistence were calculated separately for each
sub-group. No significant differences were found in the case of the
animals tested with orchic extracts "2" and "3." The five animals
given orchic extract "1," beginning 48 hours before the test, aver-
aged 9.8 crossings, the other sub-group, 4.6. The number of animals
used is obviously too small to allow any definite conclusions to be
drawn, but the results suggest, first, that type "1" of orchic extract
alone has a facilitating effect on the male drive, and secondly, that
this extract does not exert its effect immediately after injection.
Thus our hypothesis of the indirect effect of the gonadal hormone
receives additional evidence in its favor. It is curious to observe
that the female hormone is found in the lipoid fraction of placental
extract whereas the male hormone is apparently present in the non-
lipoid, water soluable part of the testicular substance.

The unoperated group of senile males were a few days over 16
months old when tested. As seen from Table 1, the copulatory
activity of the rats is still about as frequent and vigorous as in the
younger animals.9 The number of crossings in the Obstruction
Apparatus, however, is considerably decreased when compared to
normal rats of 185 days and is even slightly lower than the score
for the younger animals castrated a month before the test. On the
whole, these findings are quite in harmony with the observations

9 In this connection it is necessary to bear in mind that selective factors,
operative between the sixth and sixteenth month of life, may have eliminated
(by death) more of the sexually weak than of the sexually strong animals.

THE SEX DRIVE

319

made during the two-hour copulatory interval. On this occasion
the old males paid slight attention to the female but, when they did
occupy themselves with the latter, the activity usually terminated
in copulation. The low number of approaches and contacts shown
by this group is indicative of depressed activity and lack of motiva-
tion. The percentages of crossings occurring during each successive
five-minute interval indicate lack of persistence of the drive. In
male rats, therefore, the contrectation drive as measured by the
Obstruction Method considerably decreases in intensity with
increased age. Whether this diminution is relative as well as abso-
lute, that is, whether the sex drive in relation to other drives is
weaker in the senile animals than in males six months old, is not
shown by our results. Richter (107) and others have found that the
spontaneous activity of rats is at a much higher level when the
animals are approximately 6 months old than it is 10 months later.

The vasotomized seniles scored 1.71 points less than the unoper-
ated males of the same age. The difference is unreliable and is not
supported (nor disproved) by the data on the persistence of the
drive. It is clear, however, that the occlusion of the vas deferens
did not serve to increase the contrectation drive in these animals;
the slight decrease found may be attributable to the greater sensi-
tization of the animals after the operation. Our results, therefore,
do not agree with those of Steinach (127) and his followers. The
negative findings obtained here have four possible explanations:
(1) Vasotomy is not followed by interstitial cell hypertrophy. (2)
The interstitial tissue does not produce the male hormone. (3) The
testicular hormone follows the " all-or-none " principle, and the
increase of its production above the necessary minimum is without
effect. (4) It may be that the sex drive of our vasotomized animals
increased shortly after the operation but that it had lapsed to a
lower level of intensity by the time the test was made. Futakawa
(36) found an accelerated desire for sexual intercourse in albino
rats soon after vaso-ligation was performed, but this effect passed
away very rapidly. Macht and Teagarden (85) report a transitory
improvement in muscular coordination and in appearance following
vasectomy.

320

THE SEX DRIVE

IV. Conclusion

The sex drive of albino rats, and probably of other higher verte-
brates as well, includes two fairly distinct behavior sequences or
patterns, each of which has its own end or consummatory reaction.
In the male rat there seems to be no correlation between the contrec-
tation drive which leads to various forms of sex play and the
detumescence drive which culminates in copulation. In the female
the correlation between the two drives is high.

In the present study, quantitative indices of the influence of
gonadal factors on the contrectation drive of white rats have been
obtained by use of the Obstruction Method.

A. Summary of results

1) Female rats spayed when five months old manifest complete
absence of a detumescence drive and only a very weak contrectation
drive when tested one month after the operation.

2) Ovariectomized females show a slightly greater contrectation
drive than do normal animals during dioestrum.

3) Two injections of placental extract (cow), prepared for the
follicular hormone, given 72 and 48 hours, respectively, before the
test, serve to restore the contrectation drive of ovariectomized
females almost to normal.

4) Castration of male rats results in a slow and gradual diminu-
tion in the intensity of the contrectation drive. The score of a con-
trol group was 10.56; of a group castrated when five months old,
8.20 ; and of a group castrated when three months old, 6.10 ; all of
these animals were tested at the age of six months.

5) Injections of placental extract apparently increase the con-
trectation drive — to the usual sex-object — of normal and castrated
males. This result is obtained when two injections are made, 72
and 48 hours, respectively, before the test.

6) Injections of the lipoid fraction of testis-epididymis extract
into castrated males are without effect. There is some indication
of a facilitating influence on the sex drive of gonadectomized males
by injections of the water-soluble residue of the testis-epididymis
substance.

THE SEX DRIVE

321

7) There is no significant difference between the detumseence
drive of males 16 months old and those 6 months old. The contrec-
tation drive, on the other hand, is much more intense in the younger
animals.

8) Vasotomy performed, on males 14 months old had no appreci-
able effect on the contrectation drive measured 60 days later.

9) Individual differences in sex behavior observed during the
two-hour " satiation periods" are described (page 281).

B. Summary of discussion

1) The fact that spayed females show a very weak contrectation
drive favors the view that the lack of sex activity in normal animals
during dioestrum is caused by the absence of a stimulative factor
rather than by the presence of an inhibitory ovarian hormone.

2) The following explanation is suggested for the finding that
the contrectation drive is somewhat more intense in ovariectomized
females than in normal animals during dioestrum: In normal ani-
mals the great stimulation, during oestrum, of the nervous mechan-
isms involved in sex activity is followed by a period of depressed
functioning; in spayed animals the ovarian stimulation is absent,
and sex activity remains on a level which is low, but not as low as
that of normal animals in dioestrum. The fact that there is a sex
drive at all after extirpation of the ovaries may be explained on the
basis of nervous conditioning.

3) It is suggested that the observed delay of two or three days
between the injection of placental hormone into spayed females and
the manifestation of an effect on behavior indicates that the female
hormone does not influence the nervous system directly but exerts
its effect through a mediating mechanism.

4) The writer offers the following hypothesis for explaining the
results reported in paragraphs 4 and 5 of the preceding section:
The gonadal hormone does not directly stimulate nor facilitate sex
behavior but controls the nature of the development of an interpo-
lated mechanism or structure which is sex specific in function —
that function being the stimulation of male or female sex behavior
— but which may be activated by both ovarian and testicular
hormones. Applied to the male, two alternative particularizations

322

THE SEX DRIVE

of the hypothesis are tentatively suggested, (a) The penis, which
has developed normally under the influence of the testicular hor-
mone, becomes partially tumescent in response either to general
physiological vigor or to specific external erotic stimuli. This
tumescence initiates afferent nervous impulses which stimulate sex
behavior. (6) The testicular hormone dictates the development of
the prostate gland and penis and ordinarily stimulates the prostate
to its characteristic activity. The latter causes a tumescence in the
penis and thereby initiates the afferent impulses resulting in male
sex behavior. After castration the prostate gradually atrophies but
it may be reactivated by either male or female hormone.

Bibliography

(1) Albers-Schonberg, — . 1903. Ueber erne bisher unbekannte Wir-
kung der Rbntgenstrahlen auf den Organismus. Munch, med.
Woch., 50, 1859.

(2) Allen, E., and E. Doisy. 1923. An ovarian hormone. J. Amer.
Med. Asso., 81, 819-21.

(3) Allen, E., E. Doisy, B. F. Francis, L. L. Robertson, C. E. Col-
gate, C. G. Johnston, W. B. Kountz, and H. V. Gibson. 1924.
The hormone of the ovarian follicle; its localization and action in
test animals, and additional points bearing upon the internal secre-
tion of the ovary. Amer. J. Anat., 34, 133-68.

(4) Allen, E., J. W. Whitsett, J. W. Hardy, and F. L. Kneibert.
1924. The follicular hormone of the hen ovary. Proc. Soc. Exper.
Biol, and Med., 21, 500-3.

(5) Allport, F. H. 1924. Social psychology. Boston: Houghton Mif-
flin, 453 pp.

(6) Aristotle. Historia animalium. (Trans, by D. W. Thompson.)
1909. The works of Aristotle, vol. 4. Oxford : Clarendon Press.

(7) Athanasiu, I., and A. Pezard. 1924. Influence de la castration sur
Penergie nerveuse motrice. Compt. rend. Acad, sci., 178, 874-76.

(8) Battelli, F., and J. Martin. 1922. La production du liquide des
vesicules seminales en rapport avec la secretion interne des testi-
cules. Compt. rend. Soc. biol., 87, 429-31.

(9) Benoit, J. 1922. Sur les conditions physiologiques relatives a la
parure nuptiale periodique chez les oiseaux. Compt. rend. Acad.
Sci., 174, 701-4.

(10) Berthold, A. A. 1849. Transplantation der Hoden. Midler's
Arch. Anat. Physiol, 42-46. (Cited by Vincent, S. 1924. An intro-
duction to the study of secretion. London: Arnold.) 168 pp.

THE SEX DRIVE

323

(11) Bertschi, H. 1920. Beitrage zur Physiologie der Driisen. Unter-
suchungen iiber den respiratorischen Stoffwechsel kastrierter Kan-
inchen. Biochem. Zentbl., 106, 37-55.

(12) Biedl, A. 1913. Innere Sekretion. Berlin: Urban & Schwarzen-
berg.

(13) Bouin, P., and P. Ancel. 1903. Recherches sur les cellules inter-
stitielles du testicule des mammiferes. Arch. zool. exper., 1, 437-
523.

(14) 1906. Sur l'effet des injections d'extrait de glande inter-

stielle sur la croissance. Compt. rend. Acad, sci., 142, 298.

(15) 1906. Action de l'extrait de glande interstitielle sur le de-

veloppement du squellette et des organes genitaux. Compt. rend.
Acad, sci., 142, 232.

(16) Boukalik, W. F. and R. G. Hoskins. 1927. Further studies on
testicular grafting. Endocrin., 11, 335-37.

(17) Brambell, F. W. R., and A. S. Parkes. 1927. Changes in the
ovary of the mouse following exposure to X-rays. III. Irradiation
of non-parous adult. Proc. Roy. Soc. London, (B), 101, 316-28.

(18) Brandes, — . 1914. Berlin Tagebl., June 7. (Quoted by Sharpey-
Schafer, 114, p. 391.)

(19) Brown-Sequard, C. E. 1889. The effects produced on man by sub-
cutaneous injections of a liquod obtained from the testicles of ani-
mals. Lancet, 2, 105-7.

(20) Busquet, H. 1910. La fonction sexuelle. Paris, 380 pp.

(21) 1927. Determination ou retour des caracteres de masculinite,

chez les chapons et les viex coqs, par le serum de jeunes animaux
males. Compt. rend. Soc. biol., 97, 1463-1465.

(22) Ceni, C. 1922. Der Einfluss der Sehkraft auf die Funktion des
Hodens und auf die auszeren Geschlechtscharaktere. Arch. f. Ent-
wickmech., 51, 504-8.

(23) Champy, C. 1924. Les caracteres sexuels considered comme phe-
nomenes de developpement et dans leurs rapports avec l'hormone
sexuelle. Paris, Doin, 376 pp.

(24) Courrier, R. 1921. Glande interstitielle du testicule et caracteres
sexuels secondaires chez les poissons. Compt. rend. Soc. biol., 85,

42- 43.

(25) 1923. Sur le cycle de la glande interstitielle et revolution

des caracteres sexuels secondaires chez les mammiferes a sperma-
togenese periodique. Compt. rend. Soc. biol., 89, 1131.

(26) 1926. Sur Taction quantitative de Thormone folliculaire.

Compt. rend. Acad, sci., 182, 1492-94.

(27) Doisy, E., E. Allen, J. O. Ralls, and C. S. Johnson. 1924. Prep-
aration and properties of an ovarian hormone. J. Biol. Chem., 59

43- 44.

324

THE SEX DRIVE

(28) Doisy, E., J. 0. Ralls, and C. N. Jordan. 1926. Some chemical
and physiological properties of the hormone of the liquor folliculi.
Endocrin., 10, 273-85.

(29) Donaldson, H. H., and S. Hatai. 1911. Note on the influence of
castration on the weight of the brain and spinal cord in albino rats
and on the percentage of water in them. J. Comp. Neur., 21, 155-
60.

(30) Ellis, H. 1905. Studies in the psychology of sex. Philadelphia;
Davis. Vol. 3, 2d ed., 1913. Vol. 4.

(31) Falcone, R. 1920. Sugli innesti della ghiandola interstiziale. Bi-
tot, med., 36, 1177-80.

(32) Fellmer, O. O. 1925. Ueber das Vorkommen des femininen Sexu-
allipoids in Vogeleiern und den Eirstocken der Fische. Klin.
Woch., 4, 1651-52.

(33) Fichera, G. 1905. Sur l'hypertrophie de la glande pituitaire con-
secutive a la castration. Arch. ital. biol., 43, 405-26.

(34) Fisher, N. F. 1923. The influence of the gonad hormones on the
seminal vesicles. Amer. J. Physiol., 64, 244-51.

(35) Frank, R. T., R. G. Gustavson, J. Hollow ay, D. Hyndman, H.
Krueger, and J. White. 1926. The occurrence and present chem-
ical status of the female sex hormone. Endocrin., 10, 260-72.

(36) Futakawa, M. 1925. Influence of the ligation of the seminal ducts
on the internal secretion of the testicles. J. Okayama Med. Soc,
423.

(37) Gans, H. M. 1927. Effect of early castration on voluntary activ-
ity of male albino rats. Endrocrin., 11, 141-4.

(38) and R. G. Hoskins. 1926. Studies on vigor. VIII. The

fatigability of castrated rats. Endocrin., 10, 56-63.

(39) Gerhardt, U. 1924. Versuch einer vergleichenden Analyse des
mannlichen Geschlechtstriebes der Tiere. Zsch. f. d. ges. Anat., Abt.
III., 25, 661-95.

(40) Griffiths, J. 1890. Observations on the function of the prostate
gland in man and the lower animals. J. Anat. Physiol., 24, 236-46.

(41) Guisy, B. 1912. Fernkomplikationen bei transvesikalen und peri-
nealen Prostatektomien. Pra- und post-operative Geistes-storungen.
Zsch. Urol, 6, 124-30.

(42) Guthrie, C. C. 1908. Further results of transplantation of ovaries
in chickens. J. Exper. Zool., 5, 563-71.

(43) Halban, J. 1903. Die Entstehung der Geschlechtscharaktere. Eine
Studie iiber den formativen Einfluss der Keimdriise. Arch. Gyn'd-
kol, 70, 205-308.

(44) Hallion, L., L. Morel, and E. Papin. 1913. Action vaso-dilatrice
penienne de l'extrait prostatique. Compt. rend. Soc. biol., 74.

(45) Hammett, F. S. 1920. Gynecomastia. Endocrin, 4, 205-20, 401-3.

THE SEX DRIVE

325

(46) Harms, W. 1910. Hoden- und Ovarialinjektionen bei Rana Fusca-
kastraten. Pfliig. Arch. f. d. ges. Physiol., 133, 27-44.

(47) 1914. Experimentelle Untersuchungen iiber die innere Se-

kretion der Keimdriisen. Jena: Fisher, 368 pp.

(48) 1922. Keimdriisen und Alter zustand. Fortschr. Natur-wis.

Forsch., 77, 189-298. .

(49) 1924. Morphologische und experimentelle Untersuchungen

an alternden Hunden. Zsch. f. Anat. u. Entwickmech., 71, 319-81.

(50) Hartman, C, C. Dupre, and E. Allen. 1926. The effect of fol-
licular and placental hormones upon the mammary glands and
genital tract of the opposum. Endocrin., 10, 291-300.

(51) Hatai, S. 1913. The effect of castration, spaying or semi-spaying
on the weight of the central nervous system and of the hypophysis
of the albino rat; also the effect of semi-spaying on the remaining
ovary. J. Exper. Zool., 15, 296-314.

(52) 1915. The growth of organs in the albino rat as affected by

gonadectomy. J. Exper. Zool., 18, 1-67.

(53) Hoskins, R. G. 1925. Studies on vigor. II. The effect of castra-
tion on voluntary activity. Amer. J. Physiol., 72, 324-30.

(54) 1925. Studies on vigor. IV. The effect of testicle grafts on

spontaneous activity. Endocrin, 9, 277-96.

(55) Hunt, H. L. 1922. Experiences in testicle transplantations. En-
docrin., 6, 652-54.

(56) 1925. New theory of the function of the prostate deduced

from gland transplantations in physicians. Endocrin., 9, 479-89.

(57) Hunter, J. 1928. (Quoted by C. R. Stockard in Chemistry in
Medicine. New York: Chemical Foundation, 264 pp.)

(58) Jager, G.. 1880. Die Entdeckung der Seele. (Cited by Rieger,
108.)

(59) Jenkins, M. 1928. The effect of segregation on the sex behavior
of the white rat as measured by the obstruction method. Genet.
Psychol. Monog., 3, 457-571.

(60) Jenkins, T. N., L. H. Warner, and C. J. Warden. 1926. Stand-
ard apparatus for the study of animal motivation. J. Comp. Psy-
chol, 6, 361-82.

(61) Kingsbury, B. F. 1924. The endocrine organs: a point of view.
Endocrin., 8, 91-102.

(62) Knauber, E. 1896. Einige Versuche iiber Ovarientransplantation
bei Kaninchen. Zentbl. Gynakol., 20, 524-28.

(63) Koch, W. 1921. Ueber die Russisch-Romanische Kastratensekte
der Skopzen. Jena : Fischer, 38 pp.

(64) Korenchevsky, V. 1925. The sexual glands and metabolism.
III. The influence of injections of testicular or ovarian emulsions

326

THE SEX DRIVE

upon the nitrogen and gaseous metabolism of dogs and rabbits.
Brit. J. Exper. Path., 6, 158-71.

(65) Korenchevsky, V., and M. Carr. 1925. The sexual glands and
metabolism. II. Influence of emulsions of testis and prostate upon
the nitrogen metabolism of rabbits. Brit. J. Exper. Path., 6, 74-83.

(66) Lee, M. 0. 1927. Basal metabolism in the rat during the oestrous
cycle. Amer. J. Physiol., 81, 492-93.

(67) Lichtenstern, — . 1916. Untersuchungen iiber die Funktion der
Prostata. Zsch. Urol, 10, 1-24.

(68) Lillie, F. R. 1923. Supplementary note on twins in cattle. Biol.
Bull, 44, 47-78.

(69) and K. F. Bascom. 1922. An early stage of the freemartin

and the parallel history of the interstitial cells. Science, 55, 624-25.

(70) Lipschutz, A. 1917. On the internal secretion of the sexual glands.
J. Physiol, 51, 283-86.

(71) 1918. Umwandlung der Clitoris in ein penisartiges Organ

bei der experimentellen Maskulierung. Arch. f. Entwickmech., 44,
196-206.

(72) 1924. The internal secretions of the sex glands. Baltimore:

Williams & Wilkins, 513 pp.

(73) B. Ottow, and K. Wagner. 1922. Ueber Eunuchoidismus

beim Kaninchen, bedingt durch Unterentwicklung des Hodens.
Arch. f. Entwickmech., 51, 66-78.

(74) Lipschutz, A., and E. Vinals. 1927. Fecondation par le cobaye
experimentalement hermaphrodite. Compt. rend. Soc. biol, 97,
1400-1.

(75) Livingston, A. E. 1916. The effect of castration on the weight of
the pituitary body and other glands of internal secretion in the
rabbit. Amer. J. Physiol, 40, 153-85.

(76) Loeb, L. Deutsch. meal. Woch. (Cited by Papanicolaou, 99)

(77) Loewe, S. 1925. Nachweis brunsterzeugender Stoffe im weiblichen
Blute. Klim. Woch., 4, 1407.

(78) Loewy, A. 1903. Neuere Untersuchungen zur Physiologie der
Geschlechtsorgane. Erg. d. Physiol, 2, 130-58.

(79) Loewy, A., and — Richter. 1902. Zur Frage nach dem Einfluss
der Kastration auf den Stoffwechsel. Zentbl. Physiol, 16, 449.

(80) Long, J. A., and H. M. Evans. 1922. The oestrous cycle in the rat
and its associated phenomena. Berkeley : Univ. Calif. Press, 113 pp.

(81) Lydston, G. F. 1921. Two remarkable cases of testicle implanta-
tion. N. Y. Med. J., 113, 232-33.

(82) McDougall, W. 1926. An introduction to social psychology. Bos-
ton: Luce, 513 pp.

(83) Macht, D. I. 1921. Endocrinological studies of the prostate.
Amer. J. Physiol, 55, 311-12.

THE SEX DRIVE

327

(84) Macht, D. I., and J. L. Ulrich. 1922. Effects of prostatectomy
on integration of muscular movements of the white rat. Amer. J.
Physiol, 59, 482.

(85) Macht, D. I., and E. J. Teagarden. 1923. Rejuvenation experi-
ments with vas ligation in rats. J. Urol., 10, 407-13.

(86) Marshall, F. H. A. .1922. The physiology of reproduction. Lon-
don: Longmans, Green, 770 pp.

(87) Meisenheimer, — . Experimented Studien zur Soma- und Ge-
schlechtsdifferenzierung. Vol. 2. Jena: Fischer, 1912.

(88) Moll, A. Untersuchungen iiber die Libido Sexualis. Berlin:
Fischer, 1897. 310 pp.

(89) Moore, C. R. 1919. On the physiological properties of the gonads
as controllers of somatic and psychical characteristics. I. The rat.
J. Exper. Zool., 28, 137-60.

(90) 1921. On the physiological properties of the gonads as con-
trollers of somatic and psychical characteristics. IV. Gonad trans-
plantations in the guinea pig. J. Exper. Zool., 33, 365.

(91) 1921. Sex gland transplantation and the modifying effect

in rats and guinea pigs. Anat. Bee, 20, 194.

(92) 1924. The behavior of the testis in transplantation, experi-
mental cryptorchidism, vasectomy, scrotal insulation and heat ap-
plications. Endocrin., 8, 493-508.

(93) 1926. On the physiological properties of the gonads as con-
trollers of somatic and psychical characteristics. IX. Testis graft
reactions in different environments. (Rat.) Amer. J. Anat., 37,
351-416.

(94) Muhsam, R. 1922. Endergebnisse der Hodeniiberpflanzung. Dtsch.
med. Woch., 48, 1341-43.

(95) Nagel, W. 1906. Physiologie der mannlichen Geschlechtsorgane.
Nagel's Handbuch der Physiologie des Menschen, Vol. XL Braun-
schweig: Vieweg.

(96) Oslund, R. M. 1926. Cryptorchid testes and testicular hormone
production. Amer. J. Physiol., 77, 76-82.

(97) 1926. Ligation of vasa efferentia in rats. Amer. J. Physiol.,

77, 83-90.

(98) Ott, I., and J. C. Scott. (Cited by Marshall, 86.)

(99) Papanicolaou, G. N. 1926. A specific inhibitory hormone of cor-
pus luteum. J. Amer. Med. Asso., 86, 1422-24.

(100) Parkes, A. S. 1927. On the occurrence of the oestrous cycle after
X-ray sterilization. II. Irradiation at or before birth. Proc. Boy.
Soc. London (B), 101, 71-95.

(101) Parkes, A. S., and C. W. Bellerby. 1927. Studies on the internal
secretion of the ovary. IV. The significance of the occurrence of
oestrin in the placenta. J. Physiol., 62, 395-96.

328 THE SEX DRIVE

Pelikan, E. 1876. Gerichtlich-medizinische Untersuchungen iiber
das Skopzenthum in Russland. (Russian text, xx, 1872.) German
trans, by N. Iwanoff. Giessen.

Pezard, A. 1911. Sur la determination des caracteres sexuels sec-
ondaires chez les Gallinaces. Compt. rend. Acad, sci., 154, 1183.
Reinke, F. 1896. Beitrage zur Histologic des Menschen. I. Ueber
Krystalloidbildungen in den interstitiellen Zellen des menschliehen
Hodens. Arch. f. mikr. Anat., 47, 34-44.

Retterer, E. 1927. Nouvelle observation d'un testicle de singe
greffe a l'homme. J. urol. med. et chir., 24, 97-115.
Retterer, E., and S. Woronoff. 1921. La glande genitale male et
les glandes endocrines; etudes histophysiologique. Paris: Doin,
240 pp.

Richter, C. P. 1922. A behavioristic study of the activity of the
rat. Comp. Psychol. Monographs, 1, No. 2, 55 pp.
Rieger, C. 1900. Die Castration in rechtlicher, socialer und vitaler
Hinsicht. Jena: Fischer, 35 and 113 pp.

Sand, K. 1919. Experiments on the internal secretion of the sex-
ual glands, especially on experimental hermaphroditism. J. Phys-
iol., 53, 257-63.

1921. Vasectomie pratiquee chez un chien dans un but de re-
generation. Compt. rend. Soc. biol., 85, 1201-5.

1922. L'hermaphrodisme experimental. J. physiol. path.

gen., 20, 473-87.

Schneidemuhl, G. 1883. Vergleichende Untersuchungen iiber den
feineren Bau der Cowperschen Druse. Dtsch. z. Tier-med., 6.
Serralach, N., and N. Pares. 1907. Quelques donnees sur la phy-
siologie de la prostate et du testicule. Compt. rend. Soc. biol., 63,
790.

S harpe y- S chafer, E. 1926. The endocrine organs. Part 2. Lon-
don: Longmans, Green.

Slonaker, J. R. 1912. The normal activity of the albino rat from
birth to natural death, and its rate of growth and the duration of
life. J. Animal Behav., 2, 20-42.

1927. The effect of the follicular hormone on old albino rats.

Amer. J. Physiol, 81, 325-35.

Smith, P. E. 1926. Hastening the development of the female geni-
tal system by daily homoplastic pituitary transplants. Proc. Soc.
Exper. Biol. Med., 24, 131-32.

Stanley, L. L. 1922. An analysis of one thousand testicular sub-
stance implantations. Endocrin., 6, 787-94.

Steinach, E. 1894. Untersuchungen zur vergleichenden Physiolo-
gic der mannlichen Geschlechtsorgane insbesondere der accessor-
ischen Geschlechtsdriisen. Pfliig. Arch. f. d. ges. Physiol., 56, 304-38.

THE SEX DRIVE

329

(120) 1910. Gesehleehtsrieb und echte sekundare Geschlechtsmerk-

male als Folge der inneren sekretorischen Funktion der Keim-
driisen. Zentbl. Physiol, 24, 540-66.

(121) 1911 Umstimmung des Geschlechtscharakter bei Saugertieren

durch Austausch der Pubertatsdriisen. Zentbl. Physiol, 25, 723-25.

(122) 1912. Willkiirliche Umwandlung von Saugetiermannchen in

Tiere mit ausgepragt weiblichen Geschlechtscharakteren der Keim-
driisen. Pfliig. Arch. f. d. ges. Physiol, 144, 71-108.

(123) 1913. Feminierung von Mannchen und Maskulierung von

Weibchen. Zentbl. Physiol, 27, 717-23.

(124) 1916. Experimentell erzeugte Zwitterbildung beim Sauge-

tier. Anz. Akad. Wiss., Wien, No. 12.

(125) 1916. Pubertatsdriisen und Zwitterbildung. Arch. f. Ent-

wickmech., 42, 307-32.

(126) 1920. Histologische Beschaffenheit der Keimdriise bei homo-

sexuellen Mannern. Arch. f. Entwickmech., 46, 29-37.

(127) 1920. Verjiingung durch experimentelle Neubelebung der

alternden Pubertatsdriise. Arch. f. Entwickmech, 46, 557-618.

(128) Steinach, E., H. Heinlein, and B. P. Wiesner. 1925. Activa-
tion of the sexual cycle, development of sex characters, reactivat-
ing effect on the senile organism by ovarian and placenta extract.
Pfliig. Arch. f. d. ges. Physiol, 210, 598-611.

(129) Steinaoh, E., and H. Kun. 1928. Die entwicklungsmechanische
Bedeutung der Hypophysis als Aktivator der Keimdruseninkretion.
Med. Klin,, 24, 524-29.

(130) Stockard, C. R., and G-. N. Papanicolaou. 1917. The existence
of a typical oestrous cycle in the guinea-pig, with a study of its
histological and physiological change. Amer. J. Anat., 22, 225-83.

(131) — 1919. The vaginal closure membrane, copulation, and the

vaginal plug in the guinea-pig, with further consideration of the
oestrous rhythm. Biol. Bull, 37, 222-43.

(132) Stone, C. P. 1923. Experimental studies of two important factors
underlying masculine sexual behavior: the nervous system and the
internal secretion of the testis. J. Exper. Psychol, 6, 85-106.

(133) 1924. The awakening of the copulatory ability in the male

albino rat. Amer. J. Physiol, 68, 407-24.

(134) 1927. The retention of copulatory ability in male rats fol-
lowing castration. J. Comp. Psychol, 7, 369-87.

(135) Tandler, J., and S. Gross. Einfluss der Kastration auf den Or-
ganismus. II. Mitteilung; Die Skopzen. Arch. f. Entwickmech.,
30, 236-53.

(136) 1913. Die biologischen Grundlagen der sekundaren Ge-

schlechtscharaktere. Berlin: Springer, 169 pp.

330

THE SEX DRIVE

(137) Tarchanoff, J. R. 1887. Zur Physiologie des Geschlechtsappara-
tus des Frosches. Pfliig. Arch. f. d. ges. Physiol., 40, 330-51.

(138) Thorek, M. 1924. Experimental investigations of the role of the
Leydig, seminiferous and Sertoli cells and effects of testicular
transplantation. Endocrin., 8, 61-90.

(139) Tsubura, S. 1923. Beitrage zur Kenntnis der inneren Sekretion
der Keimdriisen. II. Keimdrusen und respiratorischer Gaswechsel.
Biochem. Zentbl., 143, 291-322.

(140) Voronoff, S. 1923. Greffes testiculaires. Paris: Doin, 83 pp.

(141) 1924. Quarante-trois greffes du singe a Phomme. Paris:

Doin, 255 pp.

(142) Wallace, C. S. 1907. Prostatic enlargement. London: Frowde,
215 pp.

(143) Wang, G. H. 1923. The relation between "spontaneous" activity
and oestrous cycle in the white rat. Comp. Psychol. Monog., 2, No.
6, 27 pp.

(144) Wang, G. H., C. P. Riciiter, and A. F. Guttmacher. 1925. Ac-
tivity studies on male castrated rats with ovarian transplants, and
correlation of the activity with the histology of the grafts. Amer.
J. Physiol., 73, 581-99.

(145) Warden, C. J., and H. W. Nissen. 1928. An experimental analysis
of the obstruction method of measuring animal drives. J. Comp.
Psychol., 8, 325-42.

(146) Warner, L. H. 1927. A study of sex behavior in the white rat by
means of the obstruction method. Comp. Psychol. Monographs, 4,
No. 22, 68 pp.

(147) Wheelon, H. 1914. Extirpation of testes and vaso-motor irrita-
bility. Amer. J. Physiol, 35, 283-91.

(148) Wheelon, H., and J. L. Shipley. 1916. The effects of testicular
transplants upon vasomotor irritability. Amer. J. Physiol., 39,
394-400.

(149) Wilhelm, R. 1923. Beitrag zur histologisch-physiologischen Er-
forschung der sogenannten Erscheinungen der Verjiingung. Rev.
med. de Chile, 51, Nos. 7 and 8. (Abstracted in Ber. d. ges. Phys-
iol, 1924, 23, 261

(150) Woodworth, R. S. 1921. Psychology. New York: Holt, 580 pp.

(151) Zondek, B., and S. Aschheim. 1926. Ovarialhormon, Wachstum
der Genitalien, sexuelle Fruhreife. Klim. Woch., 5, 2199-202.

PART V

THE MATERNAL DRIVE

Part V

THE MATERNAL DRIVE
C. J. Warden

The limitations of the project allowed only a single study of the
maternal drive. This is especially unforunate inasmuch as this
drive offers so many interesting angles of approach. Among the
many possible lines of investigation, a few of the more important
are, (1) the effect of successive pregnancies on strength of drive,
(2) the relative strength and persistance of the maternal drive
from the birth of the young onward through the normal period of
lactation and beyond, (3) the relation of this drive to age, (4) the
analysis of the physiological factors underlying the maternal drive
into glandular, neural, and other components, (5) the relation of
suckling and other activities of the young to the development and
maintenance of the normal drive, and (6) the effect on the drive of
separating the young from the mother for various periods of time.

The study here reported deals with only certain of these lines of
approach in a preliminary way. In order to carry out any general
comparison of drives, it was necessary to test out a group of ani-
mals of standard age. In addition to this, a test of the incentive
value of first litters at from 8 to 16 hours after birth was made. The
effect of depriving the mother of the litter for some 3 or 4 hours
before testing was also determined, as well as the variation in
strength of drive with successive litters. The fact that the mater-
nal drive proved to be so strong suggests the desirability of more
systematic work than could be attempted at this time.

333

1. A STUDY OF MATERNAL BEHAVIOR IN THE
WHITE RAT BY MEANS OF THE OBSTRUC-
TION METHOD 1

H. W. Nissen
I. Introduction

The maternal behavior of the white rat has been the subject of
numerous observational studies, but of only a few quantitative in-
vestigations. The most commonly used criteria for the determina-
tion of the presence or absence of a maternal tendency or drive
have been the following: (a) nursing activity; (b) general care of
the young; (c) reactions of the mother animal to removal or pres-
entation of the young. Such crude methods are obviously unsatis-
factory and certainly cannot hope to give a reliable index of the
effect of slight variations in conditions, such as size of litter, age
of litter and of mother, and so forth. For the analysis of maternal
behavior into its several component factors and for the determina-
tion of the relative importance of such conditions, more exact
methods are required.

Szymanski (10) and Simmons (8) have compared the effective-
ness of the litter as an incentive in learning situations with that of
other incentives, such as various kinds of food, escape, return home,
and sex object. Szymanski gave each of three female white rats of
unknown age two trials per day in a maze, the goal box of which
contained the litter in its living cage. Training was begun on the
second day following parturition. Two of the animals did not
learn the maze at all; they were at all times quite indifferent to
their litters, the author reports. The third rat, which learned the
problem quite rapidly, was much more concerned about her young
and bestowed much care upon them. Szymanski continued giving
this animal trials even after it had learned the maze and found
that when the young had opened their eyes and moved about rather

iKeprinted from The Journal of Genetic Psychology, 1930, 37: 377-93.

334

THE MATERNAL DRIVE

335

independently (age about 20 days) the mother began to make more
errors and require more time. He concludes that the maternal
drive decreases in intensity as tine litter grows older. It is inter-
esting to note that the animal which succeeded in the maze had only
six young, whereas those which failed had 7 and 12, respectively.

Simmons (8) used 12 female albino rats between 3 and 4 months
old; presumably, therefore, they were all primiparous animals.
Beginning from 12 to 18 hours after parturition massed trials in
a simple maze were given until the animals had learned the prob-
lem or had conclusively demonstrated that they would not learn
it. The litters were all born in the goal box of the apparatus so
that, as in Szymanski's work, the incentive was really a double
one : litter and return home. Six of the 12 animals did not learn the
maze at all ; the remaining members of the group made scores ap-
proximately as good as animals tested to a bread and milk incen-
tive, although the variability of the former was higher.

The general disadvantages attendant on the learning method for
measuring drive behavior have already been pointed out in a paper
by Jenkins, Warner and Warden (2). The maternal tendency pre-
sents a special objection to this method for, unless an exceedingly
simple learning situation be chosen, the time elapsing since par-
turition changes, and, as will be shown below, this factor seems to
be of great importance in determining the intensity of the drive.

Wang (11) and Slonaker (9) have shown, according to Richter
(7), that "pregnancy causes a 60 to 95 percent decrease in activity
which lasts through the entire gestation and lactation period. ' ' The
reference, here, is to "spontaneous" activity. The significance of
this finding will be discussed in a later section.

Moss (4) tested five female rats to their "newly born litters"
by noting whether the}' would cross an electrified grid in order to
approach the young. One of the group crossed. Moss failed to
control so many vital factors (2, 13), such as equalizing or com-
pensating for individual variations in skin resistance, age of rats
and period after parturition, that his data — aside from the essen-
tial differences between his "resistance" method and the Columbia
Obstruction Method — are in no way comparable with those pre-
sented below.

336

THE MATERNAL DRIVE

The number of factors which materially influence the intensity
of the maternal drive is probably very great ; below are listed some
of the more obvious ones:

(a) Age of test animal

(b) Length of interval between parturition and test period

(c) Serial order of litter used (primiparity vs. multiparity2)

(d) Segregation of mother animal from litter

(e) Familiarity — or the opposite — of setting in which litter is found

(f ) Time relations between test and nursing periods

(g) Size of litter (conditions a and c being constant)

(h) Weight of test animal; nutritional factors

(i) Tameness of animal and experimental conditions
(j) Paternity (?)

(k) Various combinations of preceding factors

In none of the studies reviewed above has the attempt been made
to isolate any one of these factors, and in most cases many of them
have been left uncontrolled. In the experiments described below
the writer has endeavored to secure some preliminary data on the
first five items of the list. The field is large and complex, and ob-
viously the results yielded by the present work must be considered
mainly suggestive and perhaps of chief value as directive for fur-
ther, more systematic, experimentation.

II. Animals and Procedure

The data presented in this report are based on 34 tests conducted
with 24 female albino rats of the same strain as the animals used
in other studies of this project. It is unfortunate that so few ani-
mals could be brought into the study, but this circumstance was
beyond the control of either the writer or of the Laboratory;
we were obliged to make use of what time and materials were
available. The rats were of the Wistar Institute ' 1 Experimental
Colony" strain, although all of them had been born in this Labora-
tory. The diet used was McCollum 's mixture, available in the cages
at all times, supplemented by weekly rations of greens and occa-
sional feedings of a meat and bone powder made into a pasty

2 As here used this term refers not to number of young in the litter, but de-
notes that the female had had one or more litters previous to litter used in the
test.

THE MATERNAL DRIVE

337

substance by the addition of water. Generally from three to four
females were kept in a living cage together with an equal number of
males. The pregnant animals were not removed from the regular
living cages until from 1 to 7 days before parturition. At that time
they were transferred to a specially constructed enclosure which we
shall designate hereafter as the " maternity cage."3 Three such
cages were available, and each one housed, at a given time, only
one animal — together with the litter which it delivered there. The
maternity cages were of trapezoid shape, bases 4 and 7 inches re-
spectively, altitude 9 inches, giving a floor area of 49J square
inches. They were 6 inches high. The floor was constructed of wood
covered to a depth of one-half inch with wood shavings; the walls

A

c

////////

D

Fig

l

were of ^-inch wire mesh and the top was a piece of mathematical
celluloid. The wire mesh wall arising from the 4-inch base was so
arranged that it could be removed just before a test was to be made,
and the cage could then be brought into the relation with the Ob-
struction Apparatus. (Fig. 1.) Before bringing the maternity cage
into this position the mother rat was removed, being started, in the
test, in compartment A. All of the details of procedure which were
observed in other studies included in the project on animal drives
of which this is one, were followed here. For a complete description
of the apparatus and of the testing technique employed, the reader
may be referred to earlier reports of the series, especially (1), (2)
and (13). Briefly stated, the test animal was given 5 preliminary
crossings from A to the incentive compartment D, four of these
being without shock and the fifth with shock. That is, on the last
of the preliminary crossings the grid, composing the floor of com-

3 For one exception to this rule see Group C, Table 1.

338

THE MATERNAL DRIVE

partment B, was electrified- Immediately thereafter the test period
proper, twenty minutes in duration, was begun. Scores were taken
in terms of the number of times the test animal approached,
touched (but retreated) and crossed the electrified grid during
this period. Of these three types of scores, the latter is certainly
the most significant (5).

It will be noted that our conditions permitted unrestricted mat-
ing and therefore provided as natural a situation as possible, in a
laboratory setting, for the occurrence of pregnancies. No attempt
was made to ascertain just when copulations occurred; daily macro-
scopic examination of the animals proved sufficiently trustworthy
in determining an impending delivery.

The criticism may be raised that our incentive was not the litter
alone, but this incentive plus the added incentive of a "return
home." The reason for adopting this double incentive is that previ-
ous observations in the laboratory clearly indicated to the writer
that a mother rat often or usually behaves quite differently to her
young when these are in the environment in which they were born
than when in strange surroundings. It may be questioned whether
a litter is a litter, to the mother rat — or at least whether it has the
same stimulative effect on the mother — outside of a certain setting.
It will be recalled that Szymanski (10) and Simmons (8) also had
their litter incentive in the enclosure in which delivery had taken
place; in this connection Simmons says: "Observations of the ma-
ternal behavior of other rats had shown that when a litter is moved
from its original place the mother often refuses to have anything
more to do with it." In order to discover just what part the "re-
turn home" of the incentive situation played in influencing our
results, a control group, consisting of animals which had delivered
their young in the large living cages and which were tested to their
litters in the maternity cages, was included.

Our animals were divided into five groups, the various conditions
obtaining in each being summarized in Table 1. The letter "X"
indicates variability in the particular condition so marked. Group
A consisted of ten females approximately 185 days old (range 175
to 197 days) tested from 12 to 20 hours after parturition. All but
one were multiparous, having had one or two litters before the one

THE MATERNAL DRIVE

339

to which they were tested. None of the animals had had any pre-
vious contact with the Obstruction Apparatus (or with any other
apparatus, for that matter). This group was formed in order to
provide opportunity for comparing the intensity of the maternal
drive with that of other drives, tested under strictly comparable
conditions. One factor in our procedure, indeed, differed from that
adopted in studying other forms of dynamic behavior: Our animals
were not segregated (by sexes) 35 days before the test as had been
done in previous studies. This variation, however, cannot be con-
sidered very consequential, since, in the first place, our animals
were pregnant for about 22 days before the test and therefore had
not copulated for that length of time and, secondly, because Jen-
kins (1) has shown the unimportance of a 35-day post-pubertal seg-

TABLE 1

Summarizing the conditions under which each of the five groups of females

were tested

GROUP

NUMBER
OF ANI-
MALS IN
GROUP

AGE IN
DAYS

SERIAL
NUMBER OF
LITTER
USED

TIME INTER-
VAL AFTER
PARTURI-
TION
(HOURS)

SEGREGATED
FROM
LITTER

PREV-
IOUSLY
TESTED

PLACE OF
DELIVERY

A

10

185

X

(lst-3rd)

15

No

No

Maternity
cage

B

9

X

(79-150)

1st

15

No

No

Maternity
cage

C

5

X
(about
185)

X

(lst-3rd)

15

No

No

Large
living

cage

D

5

X
(Avge.
195)

X

(lst-4th)

168

No

Yes

Maternity
cage

E

5

X
(Avge.
188)

X

(lst-3rd)

168

Yes

Yes

Maternity
cage

regation period even when the sex drive itself is being measured.
The information supplied by Table 1 probably indicates sufficiently
well the points of interest involved in each of the remaining four
groups; further details will be supplied in the discussion of the
section following.

III. Results and Discussion

The results for each of the five groups are presented in Tables 2,
3, 4 and 5. Each group is described (1) by a letter which corre-

340

THE MATERNAL DRIVE

sponds to the one used in the summary of conditions found in
Table 1, and (2) by three symbols which designate, in order, (a)
age of test animals, (b) serial number of litter, i.e., whether the
litter used in the test is the first, second, third or fourth which the
animal delivered, (c) number of hours elapsing between parturi-
tion and test period. Again the letter "X" indicates that the cor-
responding factor varied among the individuals of the group.
"Con" refers to the control condition of having the litters born in
a cage other than the one used during the test. The letter "3 " in-
dicates that the mother rats were separated from their young for
from 3 to 5 hours before the test period.

In Table 6 the averages of the several groups are compared and
the reliabilities of the differences between these averages are shown.
Objections may be raised against the first three comparisons made ;
it may well be argued that the intensity of one drive cannot be com-
pared with that of another, even considering the scrupulous care
which has been taken in the several studies of this series to keep
all conditions rigidly constant. For how can we say that a bit of
food is of the same relative incentive value in the hunger drive as
is superficial contact with a sex object in the sex drive, or as a few
seconds' contact with the litter is in the maternal drive? Indeed,
this criticism cannot be completely refuted, but the following con-
siderations should materially reduce its severity. Although it is
true that the incentives used in various drives are "equated" only
by common sense, it seems improbable that a more precise method
for the purpose is, or will soon become, available. And probably no
one will deny that such comparisons as we have made are of some
value, at least from the "practical" viewpoint. Finally we may
take refuge in the statement that the comparisons are valid "under
the conditions of our experiment. ' '

The average number of crossings of Group A, composed of ani-
mals about 185 days old, is higher than that of any group in any
drive previously tested by the Columbia Obstruction Method. The
difference is statistically reliable in only one case, however, because
of the relatively small number of animals used and the variability
in performance shown. Bimodality, such as we might expect in the
light of the results of Simmons and Szymanski, is not indicated

THE MATERNAL DRIVE

341

TABLE 2

Distribution table covering approaches, contacts and crossings for each group

A

185-X-15

B

X-l-15

C

X-X-15, C

D

X-X-168

E

X-X-168, S

(Table 2) but the range in scores, from 7 to 37, is very large. It
should be remembered that we are comparing here the maximum
scores obtained in other drives with the single score — available for
comparison — in the maternal drive. Were we to vary the several
factors which influence the tendency, at the same time increasing

THE MATERNAL DRIVE

CROSSINGS

UOIJBI.IBA

rH t- rH tJ< CO

<M 00 00 30

pjBpuB;g

9.14
7.73
9.35
9.77
4.63

"U0i;BlA8<J
9SBJ9Ay

7.92
6.52
7.92
8.32
4.32

aSajeAy

22.4
28.33
11.60
11.60
5.60

eSuB'jj

7-37
13-40

0- 27

1- 28
0-12

UBip9J\[

24.0
32.0
10.0
10.0
4.0

CONTACTS

UOIJBI.IBA

CO rH C~ CO ^

Tr- 00 OS OS CO
rH

UOI^BrA9Q
pjBpilB^g

1.90
.63
1.36
2.42
1.02

U0I}BIA9(J
93BJ9AV

1.48
.52
1.04
1.92
.88

9SBJ8AV

1.30
.78
1.40
2.60
1.60

9J3UB'}J

W N * t- «
1 1 1 1 1
O O rH O O

UBtp9J\[

id O O O O
O tH rH CO CO

APPROACHES

UOIJBIJBA

jo^uatoujaoQ

C5 OS OS lO O
00 CO CO Tr- CD

U01JBJA9Q

pjBpuB;g

3.67
2.68
2.33
1.79
2.65

U0lJBtA9Q
9SBJ9AV

2.76
2.32
2.08
1.60
2.08

9SBJ9AV

4.10

3.89

3.40

4.0

4.40

9SUBU

0-11

0- 9

1- 7
1-6
1-9

UBip9J\[

o o o o o

CO CO CO 1/5

ACTIVITY

Group

A

125X15
B

X-1-15
C

X-X-15, C

X-X-168
E

X-X-168, S

THE MATERNAL DRIVE

343

I

i

a I!

3J.X1NIJM HX03

i 7 ] rH —

1

rH LO
rH

rH CO

rH rH

rrj A fcTT TIT UTCT

r-l r- 1
r-l

CO
rH

CO

CO CO

rH rH

TTflMTTV* TJTOT

CO rH
rH

rH

CO LO

CO CO

rH CO

T T fl W T T\T TIT 1 T
J J_-_.1_.V1 J\. ±±<lijj I

rH Tf*
rH

rH rH
rH

CO

rH CO

rH rH

li.L-l_X.MV HJ-9X

CM —
rH

l—tr-tr-i
rH

rH CO

CO

rrr j r\ KT T TAT UTPT

co 00

H H N
rH

CO

rH CO

rH CO

T T (I KTT W UTfjT
_LJ..1.\1 IV xlJ-T/ L

HNN
rH

LO
rH

CO

CO

Tt* rH

IT d WT W UT OT
0._L_1 AIJV Xlili{3 L

CO CO
rH

CO co

rH

LO

rH

CO

fJT fl kTT TAT UT7T

H 00

rH

rH

rH CO

CO CO

CO rH

rH CO
rH

rH CO

CO CO

CO

CO

T T f. KTT TAT TJT.AT
> 1 J. . J _ x 1 J V XXilt\J L

CO Oi

Tt< CO
rH

rH CO

CO rH CO

rH

M T fl KTT TV XX X d
.±J-._l_VlJv UlLq

•<f rH CO
rH

NHOi

rH rH -HH

rH rH

rrj fi fcTTTlf TTTO

CO CO
rH

rH rH

CO CO

IT T fl KTTTAT TIT .

H M
rH

Tt< CO
rH

rH rH CO

rH CO CO

rH CO

IT fi KTTTAT XXXCl

d._L..i _v l j v muy

-j< H H
rH

CO CO
rH

CO -HH

rH LO

tH

rjT fl KTT TAT XX J O

LO CO
rH

rH CO
rH

CO rH CO

rH CO

rH rH CO

T.T fl KTTTAT TT T.hR
JJ-U Alii lliJL-T/

rH O
rH

CO rH CO
rH

CO "tfi

rH rH CO

1-4 r4

a Jj_1 AUV VJ.CIJ3

CO rH
rH

N H H

CO rH

CO CO Tt*

CO rH

rH rH CO
rH

rH CO LO
rH

rH CO CO

CO Tf<

rH rH CO

CO t-
rH

LO CO
rH

CO LO

CO

CO rH

ACTIVITY

Approaches

Contacts

Crossings

Approaches

Contacts

Crossings

Approaches

Contacts

Crossings

Approaches

Contacts

Crossings

Approaches

Contacts

Crossings

GROUP

A

185-X-15

B

X-1-15

C

X-X-15, C

X-X-168

E

X-X-168, S

344

THE MATERNAL DRIVE

CM

lOI>U5 t- C- 00
CO U5 iH IQ

^ B

jaquinft

w *

W

P

o

^3 CD
O DO tX)

<loo

>G oo
w ■ on

n fl °

00

7 O Jh

00

O 00 bX)

ft fl ©
<!00

us
oo

M
M

0><

2 5 ^
5 o c

o

S5 P

c3 "no

$86

oo

M

THE MATERNAL DRIVE

345

TABLE 6

Showing the reliability of the differences between the average number of cross-
ings of the groups

GROUPS

DIFFERENCE
BETWEEN THE
AVERAGES

STANDARD
DEVIATION
OF THE

T» T V TT "P T? "NT P V

DIFFERENCE
DIFFERENCE

CHANCES IN
100 OF A TRUE

DIFFERENCE
GRATER THAN
0

A and Female sex drive

at maximum*

8.26

3.09

2.67

99.6

A and Female hunger

drive at maximum*

3.40

4.03

.84

80

A and Female thirst

drive at maximum*

2.70

4.54

.59

72

A and B

5.93

3.87

1.53

93.7

A and C

10.80

5.07

2.13

98.3

A and D

10.80

5.23

2.07

98.1

A and E

16.80

3.55

4.73

100

D and E

6.00

4.65

1.29

90

* These three groups were tested by L. H. Warner (13, 14, 15). All three of
the comparison groups were composed of female animals of the same age and
strain as those of Group A.

the size of our groups, it is not at all improbable that the maternal
drive, under the most favorable conditions, would prove to excel
other drives in intensity or perseverance. That, however, is merely
a possibility. What the data yielded by the experiment do indicate
is that the maternal tendency is at least as strong, as measured by
crossings in the Obstruction Apparatus, as are the hunger and thirst
drives and rather definitely more so than the sex drive, when the
latter are at their maximum.

It is interesting, in this connection, to consider the significance
of the finding reported by Wang (11) and Slonaker (9), i.e., that
1 1 spontaneous ' ' activity is diminished in amount during the time of
lactation. It seems to the writer that the spontaneous activity
method, insofar as it aims to give indices of the strength of specific
drives is based on the assumption that any intra-organic stimulation
tends to express itself in non-directed activity when the particular
external stimulus situation which is related to it, is absent. To a
limited extent this assumption seems quite justifiable, as is seen in
the work of Richter, Slonaker, Hoskins and many others. The
present study, however, considered in connection with the results
of Wang and Slonaker, above mentioned, shows its limitations, for
here we apparently have a specific drive correlated with a period
of reduced spontaneous activity. It is true, of course, that both

346

THE MATERNAL DRIVE

Wang and Slonaker allowed the litters to remain with their mothers
while the activity records of the latter during the lactation period
were taken ; under these conditions the external stimulus situation
which ' ' satisfies ' ' the maternal drive was present, so that the intra-
organic stimulation could find more or less complete release in
directed behavior (nursing and general care of the young) . If the
litters had been removed soon after parturition the activity records
of the mothers might have been quite different. One caution to be
observed, then, in interpreting spontaneous activity scores is the
control of the possibility of directed behavior. A second point to be
considered is that the manifestation of one kind of intra-organic
stimulation may be entirely obscured in spontaneous activity
records by the influence of another physiological factor. Kinder (3) ,
for instance, has shown that the tendency towards nest building
(in the non-pregnant rat) is correlated with a period of greatly
reduced spontaneous movement. The influence of the oestrous
cycle, apparently, completely overshadows the effect — on spontane-
ous activity — of the nest-building tendency. The obstruction
method obviously isolates to a far greater extent the operation of
the several drives and can bring each to its maximum degree of
measureable expression.

The animals of Group B were younger than those of A, ranging
from 79 to 150 days at the time of the test. The average score of the
former is 5.93 points higher, this difference being statistically unre-
liable. In variability Group A somewhat exceeds Group B. If the
difference between the averages is considered significant — and
after all it is rather large — the question may be raised as to what
factor was responsible for the disparity, age or the matter of
primiparity vs. multiparity. Obviously the data at hand do not
make it possible to answer this question ; either factor or a combina-
tion of the two may have determined the result. In order to solve
this problem one could test another group of animals 185 days old
reared, until about 160 days old, with vasotomized males. Checking
through the individual records of all of the groups, no significant
correlation was discovered, within a given group, between either
age or number of litter and scores. Because of the small size of the
groups such correlations, even if found, would not be very valuable.

THE MATERNAL DRIVE

347

The five animals comprising Group C delivered their young
(first, second or third litter) in the large living cages and had their
first contact with the maternity cages when they found their young
in the latter during the test. In other respects (excepting a slightly
wider age range) these rats were tested under conditions strictly
comparable to those obtaining in Group A (Table 1). The results
show rather definitely (Table 6) the influence of the variable factor.
The fact that these animals were tested to a novel situation may
have provided the incentive for an auxiliary drive: exploration.
(For study of this drive see (6)). Thus the difference found
between the averages of C and A — 10.8 points — might have been
greater if this factor could have been properly controlled. Whether
this finding means that the crossings of the other groups are to be
attributed largely to the " return home" part of the double incen-
tive, or that the mother rat does not "recognize" its litter in a
novel setting, cannot be decided here. One who has observed the
behavior of the A and B animals — how they dragged their young
from one part of the cage to the other and how they struggled
against removal from the maternity cage — and who has contrasted
this with the "indifferent" behavior manifested by the C animals,
certainly would be inclined to accept the second of these possibili-
ties.4 At first thought it might seem that a better control would
have been given by a group of animals which delivered their young
in the maternity cages and which were then tested to these cages
minus the litters. The writer considered this possibility but rejected
it because such a situation would doubtless have resulted in a large
amount of "searching" behavior, thus giving deceivingly high
scores.

The animals of Groups D and E were tested to their litters from
5 to 7 days after parturition ; those of the latter group were sepa-
rated from their young for about 4 hours immediately preceding
the test period, whereas the D animals, as those of the other three
groups, were tested within 5 minutes after separation from the

* On the record card of one of the animals of group B I find the following
notation: "Lost considerable time trying to make mother leave young so as
to replace her in entrance compartment. Score might have been higher if she
had not been so determined to stick to litter. "

348

THE MATERNAL DRIVE

litter. In age and number of litters delivered the animals of these
two groups were closely similar to each other and to those of Group
A (Table 1). The differences between the averages of these several
groups are large (Table 6) but because of the great scatter in
scores are not always reliable. If we take into consideration the
facts (1) that the D and E animals had had previous contact with
the apparatus (having been used in groups A or C), and (2) that
Warden and Nissen (12) have shown that scores tend to increase
materially as trials are repeated, even when the external and
internal stimulating conditions are kept constant, it seems probable
that with more carefully equalized conditions the differences
between A and D or E scores would be even greater. The results
indicate, at any rate, that the maternal tendency is more intense
about 15 hours after parturition than it is after an interval of
from 5 to 7 days, and that the drive diminishes in strength if the
mother rat is separated from her young for several hours preceding
the test. To a limited extent the former finding corroborates the
conclusion drawn by Szymanski as noted above. It may be men-
tioned, incidentally, that within Group D a high correlation was
found between the length of the interval: parturition-test (range
from 5 to 7 days), and scores, the lower indices being associated
with the longer intervals.

The fact that a segregation interval of several hours just before
the test serves to decrease rather than to increase the number of
crossings suggests that the internal stimulation for the maternal
drive does not arise in afferent impulses initiated by the distending
mammary glands. For it would seem reasonable to suppose that
these glands had greater turgor after an enforced restraint from
nursing (Group E) than if the young were allowed to suckle until
just before the test (Group D). This is not a necessary interpreta-
tion, however, since it is possible that the interval of separation
counteracted the glandular effect by weakening the recognition of
the mother animal for her young.

IV. Summary

1. A study was made of the dynamic aspect of maternal behavior
in twenty-four female albino rats by measuring the number of times

THE MATERNAL DRIVE

349

these animals crossed an electrified grid in order to approach their
respective litters (Columbia Obstruction Method). Because of the
relatively small number of animals used and the variability in per-
formance found, the results have their chief value in suggesting
systematic problems for future investigation.

2. The animals were divided into five groups corresponding to
as many sets of conditions under which the tests were conducted.
The results indicate that the intensity of the maternal drive

a) is slightly greater than that of the hunger and thirst drives at
their maximum (80, 72) f

b) is greater than that of the sex drive at its maximum (99.6) ;5

c) decreases as the age of the animals increases, when litters are
being dropped with normal frequency (94) ;5

d) decreases considerably as the age of the litter increases (98) ;5

e) decreases if the mother is separated from her litter for about
four hours immediately preceding the test (90). 5

Bibliography

(1) Jenkins, Marion. 1928. The effect of segregation on the sex
behavior of the white rat as measured by the obstruction method.
Genetic Psychol. Monographs, 3, 457-571.

(2) Jenkins, T. N., L. H. Warner, and C. J. Warden. 1926. Standard
apparatus for the study of animal motivation. J. Comp. Psychol., 6,
361-82.

(3) Kinder, Elaine Flitner. 1927. A study of the nest building activ-
ity of the albino rat. J. Exp. Zool., 47, 117-61.

(4) Moss, F. A. 1924. Study of animal drives. J. Exp. Psychol., 7,
165-85.

(5) Nissen, H. W. 1929. The effects of gonadectomy, vasotomy, and
injections of placental and orchic extracts on the sex behavior of the
white rat. Genetic Psychol. Monographs.

(6) Nissen, H. W. 1929. A study of exploratory behavior in the white
rat by means of the obstruction method. J. Genetic Psychol.

(7) Richter, C. P. 1927. Animal behavior and internal drives. Quart.
Rev. Biol, 2, 307-43.

(8) Simmons, Rietta. 1924. The relative effectiveness of certain incen-
tives in animal learning. Comp. Psychol. Monog., 2, No. 7.

(9) Slonaker, J. R. 1925. The effect of copulation, pregnancy, pseudo-

6 These figures represent the statistical reliability of the indications stated,
in terms of the chances in 100 of a true difference (Table 6).

350

THE MATERNAL DRIVE

pregnancy, and lactation on the voluntary activity and food con-
sumption of the albino rat. Amer. J. Physiol, 71, 362-94.

(10) Szymanski, J. S. 1918. Versuche iiber die Wirkung der Faktoren,
die als Antrieb zum Erlernen einer Handlung dienen konnen. Pfliig.
Archiv f. d. ges. Physiol., 171, 374-85.

(11) Wang, G. H. 1923. The relation between "spontaneous" activity
and oestrous cycle in the white rat. Comp. Psychol. Monographs, 2,
No. 6.

(12) Warden, C. J., and H. W. Nissen. 1928. An experimental analysis
of the obstruction method of measuring animal drives. J. Comp.
Psychol., 8, 325-42.

(13) Warner, L. H. 1927. A study of sex behavior in the white rat by
means of the obstruction method. Comp. Psychol. Monographs, 4,
No. 22.

(14) Warner, L. H. 1928. A study of hunger behavior in the white rat
by means of the obstruction method. J. Comp. Psychol., 8, 273-99.

(15) Warner, L. H. 1928. A study of thirst behavior in the white rat by
means of the obstruction method. J. Genetic Psychdl., 35, 178-92.

PART VI
THE EXPLORATORY DRIVE

Pabt VI
THE EXPLORATORY DRIVE
C. J. Warden

The question may arise as to whether or not the tendency to
explore should be considered a specific drive, since the concept of
drive as previously developed (Warden, The Columbia Obstruction
Method) involves the notion of activity directed toward a specific
incentive. Exploratory behavior is usually more or less random
rather than so directed. In speaking of an exploratory drive, we
must mean something more than a general tendency to be active,
and the specific element may be introduced into the activity situa-
tion by providing an incentive which will induce exploratory
behavior.

This condition seems to be fulfilled, at least in some degree, in the
experimental conditions to be described in Part VI, 1. Instead of
the small incentive compartment used in the measurement of the
other drives, a large box designed especially to bring out explora-
tory behavior was provided. The results obtained under this
arrangement have been directly compared with the various control
tests of other drives in which compartment C was always empty.
The difference in behavior under the two conditions would seem to
be due to the lure of the exploratory box as thus utilized. All other
drives were controlled in the same manner as in the previous work
(see Warden: The Hunger Drive; The Thirst Drive; The Sex
Drive; The Maternal Drive). The work could not be carried
further than necessary in order to obtain an index of the explora-
tory drive in animals of the standard age selected for use in the
project as a whole.

353

1. A STUDY OF EXPLORATORY BEHAVIOR IN THE
WHITE RAT BY MEANS OF THE OBSTRUC-
TION METHOD 1

H. W. Nissen
L Introduction

The tremendous amount of activity expended by rats in ' ' explor-
ing'' a novel situation or environment can hardly escape the notice
of anyone working with these rodents in the laboratory. When
one of these animals is placed into a new cage or when it is merely
replaced into its old, unaltered cage after an absence of ten or
fifteen minutes, it spends a considerable length of time in a hurried
albeit rather thorough survey of the enclosure. Dashiell (1) has
recorded the common observation that even a hungry animal will
forego food until after it has completed its investigation of the
renovated nest box. The writer (4), in the course of his study of
the sex drive, found that a certain amount of exploratory activity —
sometimes continuing for ten minutes or more — usually preceded
any attempts at copulation on the part of sexually vigorous animals
placed in an unfamiliar situation.

In spite of the frequency with which this characteristic form of
behavior is encountered, few if any studies have been directed
specifically towards its analysis or measurement. It is taken for
granted and it is exploited; certainly the ' ' curious behavior" of
rats was a factor not of the least importance in determining the
widespread adoption of these animals for laboratory use, especially
in studies of learning.

Even the terminology to be employed in designating the behavior
here under consideration presents difficulties, for it is not known
exactly what intra-organic and environmental conditions are requi-
site for its manifestation. Apparently a novel situation — one which

i Reprinted from The Journal of Genetic Psychology, 1930, 37: 361-76.

354

THE EXPLORATORY DRIVE

355

does not frighten the animal — provokes exploration, the expres-
sion of other drives, as illustrated above, being temporarily deferred.
Thus by a crude application of the method of choice we have some
indication of the intensity of the exploratory tendency. Although
such an uncontrolled observation cannot be taken too seriously, it
suggests to the writer that exploration is more than a mere general
activity drive which finds its outlet in the most common activities
in the repertory of the animal, such as running, climbing, sniffing
and moving the vibrissae. If exploration is in fact a dynamic form
of behavior akin to the hunger, thirst, sex and maternal drives, it
seems that when put to the test the animal would overcome a certain
obstruction in order to explore . If, on the other hand, exploratory
behavior is nothing more than the chance manifestation of a ten-
dency towards the activity or exercise of the effector mechanisms, it
does not seem reasonable to suppose that the animal would repeat-
edly overcome its negative reaction to such an obstacle or obstruc-
tion in order to reach an external situation which is especially
favorable to exploration. In the latter case, indeed, the term " ex-
ploratory activity" would prove to be a misnomer, a result of mis-
interpretation.

It was in the light of these considerations that the present experi-
ment was planned. The Columbia Obstruction Method was em-
ployed ; a group of animals was given the opportunity to cross an
electrified grid in order to reach a novel and " interesting' ' situa-
tion, and the scores of this group were compared with those of other
animals, in a similar physiological condition, tested to a situation
in which the novelty factor of the incentive was relatively insignifi-
cant. A further discussion of the theoretical aspects of the problem
will be deferred until after the details of procedure and of results
have been presented.

This study is one of a series of investigations of drive behavior
in the white rat using the Columbia Obstruction Method. As far as
possible all factors, with the exception of the variable being tested,
have been kept constant throughout this larger project. The appa-
ratus and testing technique employed have been fully described in
earlier reports of the series, (2, 3, 5,) hence only a short account of
the more important features will be given here, together with a

356

THE EXPLORATORY DRIVE

description of certain modifications introduced because of the
special nature of the present problem.

In the accompanying diagram (Fig. 1), the Obstruction Appa-
ratus proper is indicated by the portions lettered A, B, and C. The
test animal must pass from the entrance compartment A, through
B, which contains the electric grid, to C, the incentive reaction com-
partment. Instead of the usual compartment D in which such
incentives as food, water, etc., are placed, a new sort of compartment

E

Fig. 1. A, B, C, comprise the Obstruction Apparatus proper. D, is the
exploratory incentive compartment, which contained sawdust and objects as
detailed in the text.

D, designed to stimulate exploratory behavior was used in this
experiment. This compartment was constructed after a plan by
Dashiell (1) of an apparatus by which to demonstrate spontaneous
activity in the white rat in a simple manner. Several modifications
of Dashiell 's design were introduced. The walls (solid lines of inner
portion of D) were made of galvanized sheet iron rising 5£ inches
above the pine wood base. All pathways were 4 inches wide, the
floor consisting of 32 squares each having an area of 16 square
inches. In this way the total area of the apparatus to which the test
animal had access after crossing the grid was approximately six
times as great as it was in previous studies of the series. The entire
top of the incentive compartment was covered by a large sheet of
rigid wire mesh which was so hinged that it could be easily lifted

THE EXPLORATORY DRIVE

357

up by the experimenter for removal of the test animal. Small pads
of felt were fastened to the underside of the mesh cover so that its
movement did not cause any appreciable amount of noise.

Preliminary observations with several rats confirmed the writer 's
guess that the many pathways and corners in compartment D would
provoke a great amount of exploratory activity. Several features
were added to further enhance the "interest value" of the situa-
tion: Wood shavings were piled up to the top of the walls in the
back part of the incentive chamber (shown by stippling in Figure
1). In two of the square areas, blocks of wood (b) were placed, and
a cork (c) in each of two others. There was a small rubber mat (r)
in one alley-way; two walls (x) were formed of wire mesh. In the
experience of the writer all of these items were of the nature to
stimulate exploratory behavior in rats.

The automatic device (see Part I, 1) for the control of the door
(d) leading to compartment D was used throughout the tests — not
because it was necessary for our purposes but in order to keep
whatever effect this mechanism may have the same as in the other
drive studies in which this general method has been used.

II. Animals and Procedure

Twenty male albino rats about 185 days old (range : 179 to 191
days) were used in the experiment. All of them were born in this
Laboratory, being the descendents (first and second generations) of
animals procured from the Wistar Institute of Anatomy, Philadel-
phia. Males were used exclusively in order to avoid the complica-
tions introduced by the oestrous cycle of female animals, but also
because it was planned to use all females available in the Laboratory
at the time this study was made, for another investigation. The
living conditions of our animals, described below, were planned
with two requirements in view: (1) Strict comparability with the
groups to be used as controls for our results; (2) Reduction to a
minimum of the operation of other drives (hunger, thirst, sex, gen-
eral activity) in the test situation. The success of our attempt to
meet these demands can be judged from the following account and
by reference to previous reports in which are described the condi-
tions of testing the control groups (5, 6, 7).

358

THE EXPLORATORY DRIVE

The animals were not weaned (artificially, at least) until about
60 days old; from that time until about the 150th day of life they
were kept with an equal number of females in cages containing 6
to 8 animals each. Beginning approximately 35 days before the test
the males were separated from animals of the other sex; Warner
(5) has shown that the sex drive of male albino rats is at a fairly
low level of intensity after such a period of segregation.

The cages were supplied at all times with an abundance of the
standard food, McCollum 's mixture. Greens were given once a week
and, until the animals were 135 days old, a meat and bone powder
was fed about every ten days. Plenty of fresh water was always
available in the cages. Since the animals were taken directly from
their living cages to the testing apparatus, it would seem that the
hunger and thirst drives did not enter as contributing factors into
our results. Any social drive which may be operative in these ani-
mals in addition to the specific sex drive was probably kept at a low
degree of intensity by the living conditions above described. The
cages, each housing about 7 animals, were large enough (5900 cubic
inches capacity) to insure ample opportunity for exercise so that
any activity drive which may (and I believe does) exist in addition
to the other, more specific tendencies, should have been pretty well
' ' satisfied ' ' at the time of the test.

Excepting the fact that care was taken not to clean the living
cages or to otherwise disturb the animals for a period of two days
immediately preceding the test, no special arrangements were taken
to control the little known intra-organic conditions influencing the
exploratory tendency. It is quite possible, of course, that the past
history of the animal — the opportunities which it has had for
exploration, the kind of environment in which it has lived, and so on
— is an important influence in determining the intensity of this
drive, but this is a matter for future work to decide. Regarding
the present experiment we can say that the test animals had lived
in a highly uniform environment giving little chance for unusual
exploration and affording no definite rewards or incentives for such
activity. It is probable, therefore, that our pre-testing conditions
were generally unfavorable to the drive being measured.

All tests were conducted in the evening, usually between 9 and 11

THE EXPLORATORY DRIVE

359

o 'clock and never before 8 o 'clock. The same technique of handling
the animals and the same precautions to avoid emotional disturb-
ances were observed as in previous studies (2, 5) of this general
project.

Preliminary crossings. In order (1) to adapt the test animal to
the experimental situation, (2) to get the connection, other-side-of-
the-grid-incentive, established, and (3) to exhaust as far as possible
the influence of a conceivable auxiliary drive (exploration), the
usual technique of the Columbia Obstruction Method calls for
five preliminary crossings from the entrance to the incentive com-
partment; four of these crossings are allowed without any current
passing through the grid, and on the fifth the shock is introduced
after the animal is already in the obstruction chamber. After each
of these preliminary crossings the animal gets a certain amount of
stimulation from the incentive, food, water, sex-object or litter, as
the case may be. The analogous procedure in the present study
would have been, perhaps, to allow the animal to advance a certain
distance — or for a certain length of time — into compartment D, at
the end of each preliminary crossing, as well as after each crossing
during the test proper. The following considerations, however,
caused the writer to modify the preliminary procedure to the extent
of allowing the animal to proceed only as far as d3, (see Figure 1)
after each of the first three crossings and to enter E only after each
of the last two preliminary runs: (a) The novelty of the incentive
obviously decreases with each opportunity to explore it. The pro-
cedure adopted preserved the stimulating effectiveness of compart-
ment D inasmuch as when the test proper began the animal had
been in it only twice, (b) In order legitimately to compare our data
with those of other studies (thus giving the only possible basis for
evaluating our results) it was necessary to equate approximately
the length of the preliminary period. If door d3 had been left open
throughout this period — so it was found in preliminary work —
the animals would have gone right through to D, stopping in C
hardly at all. Then it would have been necessary to remove the
animals rather quickly in order to prevent too much exploration of
D (reason (a) above) and this, of course, would have meant that
the preliminary period of our experimental animals would have

360

THE EXPLORATORY DRIVE

been much shorter than that of their controls, (c) Preliminary-
work had demonstrated that if the animal were given no access to
D at any time before the test period proper, it would often not even
attempt to cross the grid during the test. Naturally the animal
would need to have been previously stimulated in D if the latter
was to function as an incentive and induce crossings in the test,
(d) It had been found in studies of other drives that during the
first few preliminary crossings the animal paid attention to the
specific incentive only after it had thoroughly investigated compart-
ment C. It seems, therefore, that our modified procedure gave suf-
ficient opportunity for the formation of the association between the
further side of the grid and the incentive, and at the same time
equated fairly well the preliminary practice given to the experi-
mental and control groups.

At the end of each of the first three preliminary crossings the test
animal was given 3 minutes in C. This was more time than was
given to most of the animals of the control groups but the differ-
ence, if effective at all, influenced the results in a direction unfavor-
able to the exploratory drive. After each of the last two preliminary
runs the animal was allowed to explore D for ten seconds. During
the test proper the animal was removed after 8 to 10 seconds in D ;
during this time it was able to traverse from 5 to 15 of the square
areas (see Figure 1). If the rat made no attempts to explore it was
removed anyway at the end of 30 seconds (compare (2) page 485).
It may be said that on the average the amount of time allowed the
test animal in the incentive compartments after each crossing, both
in the preliminary period and during the test itself, was about the
same in this study as in previous investigations of this series.

III. Results

The results for the group of twenty male albino rats tested are
given in Tables 1, 2, 3 and 4. Of the three types of scores yielded
by the method used, crossings probably have the greatest signifi-
cance ; they show how many times the animals actually crossed the
electrified grid and reached the incentive. Contacts and approaches,
while probably indicating a tendency to cross, may be, conceivably,
only the expression of spontaneous, non-directed activity. The

THE EXPLORATORY DRIVE

361

major portion of our discussion, therefore, will be based on scores
in crossings.

In Table 5 are shown the differences — and the reliabilities of
these differences — between the average score of the exploratory
drive group and the average of each of the following: (A) A group
of ten males tested to a food incentive immediately after being
removed from a living cage well supplied with food. (B) A group
of ten males tested to water immediately after being taken from a
cage in which an ample supply of fresh water was available. (C) A
group of twenty males tested to a small empty incentive compart-

TABLE 1

Distribution table covering approaches, contacts and crossings

NUMBER OF RESPONSES

APPROACHES

CONTACTS

CROSSINGS

0

3

6

1

4

2

2

2

4

1

3

3

2

3

4

2

2

5

5

1

2

1

6

4

2

7

1

8

1

9

2

1

10

1

11

1

12

1

1

13

14

1

15

16

17

18

19

20

21

22

23

1

TABLE 2

Showing measures of central tendency and of variability

APPROACHES

CONTACTS

CROSSINGS

> 2

3-3

O c8

a, >

CQfi

O 03

> a

9 O

0-14

5.20

3.12

3.85

74

1.5

0-5

1.80

1.40

1.66

92

4.0

1-23

6.00

3.40

4.89

81

362

THE EXPLORATORY DRIVE

TABLE 3

Showing temporal distribution of approaches, contacts and crossings during the
twenty-minute test period in one-minute intervals

a

a

a

K

a

a

a

a

a

a

a

a

E<
P

(3

&
P

&

&
P

&

P

E-

&

&

&

E-

&

Eh

E-

E-

P
g

g

P
g

g

g

P
g

g

g

P
g

P
g

p

g

P
g

p
g

P
g

p
g

P
g

P

g

ACTIVITY

9

3

3

5

i

i

i

s

S

«

H

H

w

&

w

M

E-

H
&

H

H

H

W

H

H

W

K

K

M

M

r<
10

&

t-

00

a

o

&

&

&
oa

e*

■<*

E-
1C

Eh

t-

E-

X

E-

a

Approaches

12

7

9

1

9

G

2

11

5

5

2

4

4

5

4

2

4

2

4

6

Contacts

3

4

4

1

2

2

6

1

1

1

1

2

4

1

1

2

Crossings

12

14

7

7

8

6

S

7

5

3

G

4

3

6

4

3

3

4

4

6

TABLE 4

Showing temporal distribution of approaches, contacts and crossings during the
twenty-minute test period in five-minute intervals

ACTIVITY

0 TO 5th

MINUTE

6th to 10th

MINUTE

llTH TO 15TH
MINUTE

16th TO 20TH
MINUTE

TOTAL
NUM-
PER

Num-
ber

Per
cent

Num-
ber

Per
cent

Num-
ber

Per
cent

Num-
ber

Per
cent

Approaches

38

36.5

29

27.9

19

18.3

18

17.3

104

Contacts

12

33.3

12

33.3

4

11.1

8

22.3

36

Crossings

48

40.0

29

24.1

23

19.2

20

16.7

120

TABLE 5

Showing the reliability of the differences between the average number of cross-
ings of the exploratory group and of the three comparison groups

EXPLORATORY GROUP AND COMPARISON
GROUP

DIFFERENCE
BETWEEN
THE AVER
AGES *

STANDARD
DEVIATION
OF THE
DIFFERENCE

DIFFERENCE

S. D. OF
DIFFERENCE

CHANCES IN
100 OF A
TRUE DIF-
FERENCE
GREATER
THAN 0

A — Ten males tested to food
without previous starvation
period (6)
B — Ten males tested to water
without previous water-depri-
vation interval (7)

C — Twenty males tested to
small empty incentive com-
partment during period of
maximum sexual vigor (5)

3.30
2.30

3.05

1.52
1.67

1.16

2.16
1.38

2.63

98.46

91.62

99.57

* In all cases the score of the exploratory drive group is higher than those
of the several comparison groups.

ment at a period of maximum sexual vigor (24 hours after a
two-hour copulatory period). In all cases the members of the com-
parison groups were of the same age and sex as were the animals
tested to the exploratory incentive; the physiological condition of

THE EXPLORATORY DRIVE

to e
.§3

it

IS1

!i

^ ft,

I I

+

+

S o

QQ O

a 5

.2^

S.2 S

I " I

364

THE EXPLORATORY DRIVE

the animals of the first two comparison groups was as similar to that
of our animals as possible. A summary of the internal and external
stimulating conditions pertaining to the several groups is presented
in Table 6; a plus, minus or equal sign indicates that the given
condition is presumably present in greater, lesser or equal measure,
respectively, as compared to the exploratory group.

It appears from Table 6 that in each of the three principal com-
parisons to be made we are dealing with more than one variable.
The difference in intra-organic stimulation found in comparison
group C, indeed, makes any very significant conclusions impossible
here; we can only say that the external factor: exploratory incen-
tive, results in a greater number of crossings than the internal con-
dition of sexual excitation. The two variables in comparison groups
A and B, on the other hand, are both external factors, and in each
of these groups one variable may be disregarded; food and water
may probably be considered as of zero stimulating value to animals
which are neither hungry nor thirsty. Arguments denying this
point will, of course, strengthen the conclusion drawn by the writer
below. The only significant difference to be considered between the
exploratory and the first two comparison groups, then, is the
incentive which presumably is related to the exploratory tendency.
This one important disparity in the experimental conditions
resulted in a difference in scores in both cases favorable to the
group tested with the exploratory incentive. In other words, our
animals overcame their negative reaction to an electrified grid
more often, within a stated period of time, if by crossing that grid
they reached a situation affording opportunity for exploration than
if the further side of the obstruction compartment was less conduc-
ive to exploratory activity. This suggests — although it does not
prove, since complete statistical reliability is lacking — that explora-
tion is a definite form of dynamic behavior similar to the hunger,
thirst, sex and maternal drives.

It should be remembered that the physiological-neurological con-
dition related to exploration, whatever it may be, was not controlled
in such a way as to certainly obtain a maximum expression of the
drive. Probably the scores of the exploratory group would have
been higher had we restrained the animals in a small opaque enclo-

THE EXPLORATORY DRIVE

365

sure, giving a minimum of extero-ceptive stimulation, for some time
before the test. The writer had considered this procedure but aban-
doned it because of the necessarily concomitant variation in the
possible activity tendency — the drive towards the exercise of the
neuro-muscular mechanisms — whose importance in influencing be-
havior in the obstruction apparatus is not known. It may be argued
that even under our actual experimental conditions the incentive
used was more favorable not only to exploration but also to mus-
cular activity (running, etc.). This is true to a certain extent, but
consideration of the following facts makes it appear improbable
that the activity drive entered as an important factor into our re-
sults: (1) Our animals had ample opportunity for muscular exer-
cise in their cages until just beore the test (see page 6) ; (2) the
entrance compartment of the apparatus, while not as large as the
living cages, is extensive enough (10x10x10 inches) to permit free
movement. We discuss below the possibility of considering ex-
ploration as an activity drive of a special kind.

In behavior tendencies like those of hunger and sex there are
certain fairly definite physiological conditions related to each spe-
cific drive. Is there a comparable intra-organic mechanism in the
exploratory drive — assuming that our results justify the use of
this term ? The answer to this question is so utterly remote that a
few speculations may be permissible. Just as the muscles appar-
ently require a certain optimum amount of exercise, may not the
receptor organs and their more central connections also require
stimulation ? If this be considered a reasonable assumption we may
go a step further and suggest two possible modes of origin of this
tendency or trend toward stimulation: (a) A fundamental ten-
dency inherent in all living tissue toward the expression of its
characteristic activity; a basic drive toward functioning, (b) A
derivative tendency; in the early life history of the animal (or of
the species) specific intra-organic stimuli (hunger contractions, for
instance) led to random movements and these, in turn, to various
kinds of external stimulation and to " satisfaction of internal
needs." With repetition of such sequences the novel situation —
i.e., novelty — became conditioned to the primary drives and thus
became of itself efficient in arousing a new type of drive behavior.

366

THE EXPLORATORY DRIVE

It has been shown (2, 5) that sex behavior is dependent upon
two factors: environmental and intra-organic stimulation. The
same is true of hunger drive (6) and thirst drive (7) behavior. To
a limited ext'ent the intra-organic factor may be measured in the
absence of the related external stimuli, since an internal tension
(or condition of excitation) has the tendency to express itself in
spontaneous, non-directed activity when the proper incentive object
is not present or available. Without both factors in the experi-
mental situation, however, drive behavior, capable of measure-
ment in terms of an obstruction overcome, is apparently impossible.
We cannot, therefore, agree with Warner (5, page 4) when he
says, speaking of the physiological condition of the animal and
of the related external stimuli, that "of these two factors implied
by the term, drive, the first is essential, the second non-essential. ' 1
The writer holds that both are essential.2 Nevertheless it seems to
be true that in some cases the external factor is of greater relative
importance, in other cases the internal factor. Warner (page 65)
appears to mean this when he says, "The behavior data indicate
that sex activity is initiated rather more by the external stimulus
situation in the male than in the female, and rather more by inter-
nal stimulation in the female than in the male. ' ' In the exploratory
drive we have, perhaps, the case in which the external stimulus sit-
uation is of the greatest relative importance, whereas the activity
drive (tendency toward the exercise of the neuro-muscular mech-
anisms) possibly offers one of the best illustrations of the predom-
inance of the intra-organic factor.

Summary

1. The tendency of twenty male albino rats approximately 185
days old to explore a novel situation was measured in terms of the
number of times they crossed an electrified grid in order to reach
an incentive especially designed to furnish favorable opportunity
for exploration. The results of this group were compared with

2 1 do not mean, of course, that the incentive object must be immediately
present to the animal. It is sufficient if the immediately present situation is
capable of initiating fairly specific 1 'preparatory reactions" which lead to the
necessary environmental factor.

THE EXPLORATORY DRIVE

367

those of other animals tested under conditions varying only in
respect to the exploratory "value" of the incentive employed.

2. The results indicate that use of an incentive offering oppor-
tunity for exploration increases the tendency of the animals used
to cross the standard shock employed in the Columbia Obstruction
Method. This indication is not entirely reliable statistically, vari-
ability in performance being very great.

3. In so far as the finding reported in the preceding paragraph
is considered reliable, we are justified in speaking of an exploratory
drive in rats of the age, sex and strain used.

Bibliography

(1) Dashiell, J. F. 1925. A quantitative demonstration of animal drive.
J. Comp. Psychol, 5, 205-8.

(2) Jenkins, Marion. 1928. The effects of segregation on the sex be-
havior of the white rat as measured by the obstruction method. Genetic
Psychol. Monographs, 3, 457-571.

(3) Jenkins, T. N., L. H. Warner, and C. J. Warden. 1926. Standard
apparatus for the study of animal motivation. J. Comp. Psychol., 6,
361-82.

(4) Nissen, H. W. 1929. The effects of gonadectomy, vasotomy, and
injections of placental and orchic extracts on the sex behavior of
the white rat. Genetic Psychol. Monographs.

(5) Warner, L. H. 1927. A study of sex behavior in the white rat by
means of the obstruction method. Comp. Psychol. Monographs, 4,
No. 22.

(6) Warner, L. H. 1928. A study of hunger behavior in the white rat
by means of the obstruction method. J. Comp. Psychol., 8, 273-99.

(7) Warner, L. H. 1928. A study of thirst behavior in the white rat by
means of the obstruction method. J. Genetic Psychol., 35, 178-92.

PART VII

A COMPARISON OF NORMAL DRIVES

Part VII

A COMPARISON OF NORMAL DRIVES

C. J. Warden

The primary purpose of the project as a whole was to make pos-
sible a direct comparison of the normal drives in the white rat. By
the term 11 normal" we mean merely that the conditions so desig-
nated were as natural as is possible in a laboratory environment
except for the fact that the animal was deprived of the specific
incentive under investigation. This general situation is normal as
contrasted with many other conditions tested, such as those in-
volving segregation, gonadectomy, vasotomy, injections, delayed
incentive and the like. These normal conditions are of special in-
terest in connection with our main purpose in determining the rela-
tive influence of the different drives under natural conditions. It
is clear that the maximum scores obtaied under normal conditions
may be made the basis of a ranking of the drives for this species,
within the limits of the particular test method employed. The com-
parison of normal drives (Part VII, 1) is based upon data drawn
from Part II, 1 ; Part III, 1 ; Part IV, 1 ; Part V, 1 ; Part VI, 1 ;
and covers the hunger, thirst, and sex drives, both male and female,
and the maternal and exploratory drives.

No attempt will be made to summarize and compare the results
reported in the other studies; since these are concerned with an
extension of the test to special conditions of a given drive and have
no general comparative significance. Whatever comparisons are
possible were indicated in each paper, and a summary of these
would be little more than a repetition of the main conclusions in
each case.

In bringing together the results of the five studies cited for pur-
poses of comparison, only such scores as seem relevant to the dis-
cussion of the rank order of the various drives have been tabulated
and used. The main emphasis throughout Part VII, 1, will be
upon the maximum scores for the normal drives, since the ranking
of the drives must rest upon these scores.

371

1. THE RELATIVE STRENGTH AND PERSISTENCE
OF THE NORMAL DRIVES IN THE WHITE RAT

C. J. Warden

The central purpose of the project, as stated in the preceding sec-
tion, was to determine the relative strength and persistence of the
more important drives in the white rat when tested under the most
natural conditions, e.g., with all conditions normal except for some
definite and measurable deprivation of the incentive. We shall
speak of the drives as tested under these conditions as " normal"
drives. The value of the present study over previous studies of
dynamic behavior lies not merely in the fact that a more direct and
controlled method (The Columbia Obstruction Method) was used
in the measurement of the several drives, but even more in the
fact that the project was so standardized throughout that the
scores of the various drives may be directly compared with one
another. This standardization makes possible a fairly definite rank-
ing of the drives and thus gives us a picture of the natural dynamic
behavior patterns of the white rat, within the limitations of a
single test method.

Before entering upon an analysis and comparison of the actual
scores, it will be necessary to direct attention to the extent to which
standardization was carried in method and procedure — since the
validity of comparing the scores directly from drive to drive must
depend, in the last analysis, upon the thoroughness of such stand-
ardization. This aspect of the problem will be covered in the first
sub-section, and this will be followed by a second sub-section deal-
ing with the comparison of the scores indicating the ranking of the
different drives.

Standardization of Experimental Conditions

Our interest here is not to establish the validity of the Columbia
Obstruction Method as an instrument for the measurement of ani-

372

COMPARISON OF NORMAL DRIVES

373

mal drives; for such a treatment the reader may be referred to
Part I of this volume. Our present concern is merely to show that
the method was adapted to the measurement of the various drives
under such uniform conditions that the scores from drive to drive
may be directly compared. For the sake of clearness the conditions
of uniformity adopted will be presented, in so far as possible, in
tabular form. The conditions enumerated below were uniform for
all the drives except for certain necessary variations as indicated.

A. Apparatus conditions (See Part I, Fig. 1, for general diagram

of apparatus)

1. Compartment A (Entrance) empty except for test animal.

2. Compartment B (Obstruction) grid giving standard shock;
alternating current of 60 cycles, with terminal pressure of 475
volts, external resistance of 10,100,000 ohms, and current of 0.047
milliamperes.

3. Compartment C (Incentive response-chamber) empty except
for test animal.

4. Compartment D (Incentive container). Identical box used
in tests on hunger, thirst, and sex ; for special maternity cages, see
Fig. 1, Part V, 1 ; for exploration box, see Fig. 1, Part VI, 1.

5. Doors separating the different compartments, including the
automatic release door between compartments C and D, were oper-
ated in the same manner in testing all drives, regardless of whether
the incentive required the use of a given door or not.

B. Animals employed

1. Wistar Institute experimental colony strain throughout.

2. Animals reared at Wistar Institute until approximately 150
days of age, or first and second generation of this strain reared
in our own laboratory.

3. Animals weaned at approximately 30 days ; males and females
reared together until 150 days of age, at which time they were
segregated as to sex for 35 days (range, 33-39 days) to eliminate
pregnancies. Exception: groups used in study of the maternal
drive were not segregated until near term, normally the fact of
pregnancy being sufficient to prevent copulation during this pe-

374 COMPARISON OF NORMAL DRIVES

riod ; this exception was necessary in order to make the primiparous
and muciparous groups strictly comparable.

4. Standard age for testing was 185 days (range, 175-196 days)
or the age reached at the end of the 35-day segregation period. Ex-
ceptions: (a) Male sex drive, 28-day group (Part IV, 1) were of
standard age but for the longer period of sex deprivation (28
days) involved; this group can be eliminated without effecting the
comparisons to be made in this section, (b) First litter, maternal
drive group in which age of testing was necessarily determined by
the factor of primaparity (range, 79-150 days).

5. Animals shipped to us from the Wistar Institute at approxi-
mately 150 days were thus allowed about 5 weeks to become ac-
customed to living conditions in our laboratory before reaching the
standard test age. See Part IV, 1, sub-section on Method and Pro-
cedure, for detailed statement of general living conditions in our
laboratory — conditions which were strictly maintained through-
out the testing of all drives.

C. Incentive conditions

1. The general plan was to deprive the animal of some one of
the factors (food, water, sex, litter, etc.) of the standardized set of
living conditions, keeping all other factors of the normal living
conditions constant and uniform from drive to drive.

2. A summary of incentive conditions is presented in Table I of
this study. The incentives used in the apparatus were samples of
usual food (McCollum's diet), water supply (metal nipple bottle
system), sex object, etc., so that the incentive response was always
to the sort of stimulus to which the animals had been accustomed
in cage life.

3. Isolation of a given drive for testing was accomplished, in so
far as isolation could be secured, by keeping drives, other than the
one being tested quiescent, or at a physiological minimum during
the test period. A summary of the controls utilized are shown
in the last five columns of Table I of this study. The conditions for
keeping any one of the given drives at a minimum while testing
some other drive was constant and uniform throughout. In the test-
ing of any drive, four of these controls were operative, so that,

COMPARISON OF NORMAL DRIVES

375

METHOD OF KEEPING OTHER DRIVES QUIESCENT

Exploratory

Specific in-
centive ob-
jects pres-
ent; no
place to ex-
plore

Same as
above

Same as
above

Same as
above

Same as
above

Drive be-
ing tested

Maternal

Only non-
pregnant
females
without
litters

Same as
above

Same as
above

Same as
above

Drive be-
ing tested

Males only

Sex 1

Males seg.
35 days ;
Females in
dioestrum

Same as
above

Drive be-
ing tested

Drive be-
ing tested

Females
with litter

Males seg.
35 days

Thirst

Regular
water sup-
ply to
hour of
test

Drive be-
ing tested

Same as
thirst
above

Same as
above

Same as
above

Same as
above

Hunger

Drive be-
ing tested

Regular
diet, ex-
cept greens
omitted

Regular
diet to
hour of
test

Same as
above

Same as
above

Same as
above

INCENTIVE
RESPONSE

Nibble of
powdered
food; 30
sec. limit

Moistening
tongue on
damp nip-
ple; 30
sec. limit

Nipping,
biting, nos-
ing gen-
italia,
mounting,
etc.; 30
sec. limit

Nipping,
biting, nos-
ing gen-
italia,
exposing
vulva; 30
sec. limit

Attending
to litter ;
30 sec.
limit

Random
activity ;
30 sec.
limit

INCENTIVE

( Compart
ment D)

Sample of

regular
diet (Mc-
Collum's
mixture)
in compart.

Regular

water
bottle in
compart.

Female in
cornified
stage, ex-
cept com-
part, emp-
ty for con-
trol group

Normal
male, ex-
cept com-
part, emp-
ty for con-
trol group

Litter in
maternity
cage

Explora-
tion box

§

o

h
o

H

eg

6Q

Comb.

o o o o o o

CO CO CO CO CO CO

o o o o o

CO CO 03 CO CO

ft

O O O O O O

OOOOO
rH rH rH rH tH

t— rH CO CO SO <M
CO CO CO

Oi o

O O O O O O
rH iH iH rH iH rH

OOOOO
rH iH rH rH rH

oooooooo

C0C0COC0COCOC0C0

o

CO

DRIVE CONDITION
TESTED

03

d M SI 10 « to 10

l_j 03 c3 c3 o3 o3 o3
rln X) Tj

O CO CO ^ CD 00

(Thirst)

0 days

1 day

2 days
4 days
6 days

5 I

s §

1 £££ CO CO ©3

» O o O d c5 S « «i

O^CXNrHrH-^OOO
rH CO

(Sex — female)
Early congestive
Oestrum (cornified)
Same (control)
Late cornified
Post-ovulative
Dioestrum (4 stages)

(Maternal)
First litter

Multiparous (stand, age)

(Exploratory)
Males (stand, age)

376 COMPARISON OF NORMAL DRIVES

even if the controls should be considered inadequate, they were
uniform for all normal drive conditions investigated. The ade-
quacy of the various controls will be discussed at some length later.

D. Procedure

1. The experimenter was eliminated by making the tests in a
dark room with observation made possible by means of a specially
constructed illumination hood (see Part I, 1) which operated to
place the experimenter outside the visual field of the test animal.

2. All tests were made between the hours of 9 p.m. and 4 a.m.,
or during the period when the white rat is normally most active ;
it was felt that maximum drive scores were most likely to be se-
cured at this time when diurnal activity is also at a maximum.
This time of day also favored better general laboratory conditions
for testing.

3. The test period proper was standardized at 20 minutes im-
mediately following the preliminary period to be next described.
The advantages of a temporal over other forms of limiting the test-
ing will be shown in the later discussion.

4. During the preliminary period, the incentive was indicated
to the animal by allowing four crossings from compartment A into
compartment C, with access to the incentive object in compartment
D, with no shock present in compartment B. This tended to make
the specific incentive dominant in the stimulation conditions af-
forded by the apparatus. The presence and nature of the shock was
indicated to the animal by a fifth crossing, with the standard cur-
rent turned on in compartment B. The testing proper followed
immediately.

Exception: In testing the exploratory drive, the animal was not
allowed to proceed into compartment D, except on the fourth and
fifth crossings of the preliminary period, in order not to lower the
value of the exploration box as an incentive and to allow compart-
ment C to be thoroughly explored in the preliminary period and
thus enhance the incentive value of the exploration box in the test
proper. This change in procedure favored the exploratory drive
somewhat, perhaps, but the scores for this drive were extremely
low anyway, so that the ranking of drives is not disturbed thereby.

COMPARISON OF NORMAL DRIVES

377

5. The general method of handling' the animals was kept uni-
form; the mode of lifting the animal from incentive compartment
and the rate of movement in transferring it back into the entrance
compartment was developed into an automatic habit by the experi-
menter before being allowed to take data in connection with the
project.

6. Scores were taken during the test period, covering the fol-
lowing types of response to the obstruction-incentive situation
(see Fig. 3, Part IV, 1, for copy of score card) :

(a) Approaches — Tip of nose across threshold of door between com-
partments A and B, but without touching the grid, facing incentive.

(b) Contacts — Touching grid with nose or stepping upon it; partial
crossings involving shock.

(c) Crossings — Complete passage over grid into incentive compartment.

7. Only one sex was tested on a given day and the apparatus
washed out thoroughly at the close of each test period.

It will be noted, in checking over the above summary of condi-
tions, that strict uniformity was had from drive to drive, with the
following exceptions: (1) The maternal groups were not segregated
as to sex 35 days previous to testing. This variation in conditions
may well be ignored in view of the fact that they were pregnant
for a period of 22 days of the 35-day period, on the average, during
which time copulation would naturally be avoided. Furthermore,
as Jenkins (Part IV, 2) has shown, a segregation period of 35 days
is of slight consequence in the case of the adult white rat. (2) One
group of males (28-day sex deprived) were, on account of the long
period of sex deprivation beyond the standard age at the time of
testing. We have eliminated this group from the comparisons of
this section, so that this variation in conditions can be entirely
ignored. (3) The first litter, maternal group was younger than the
standard age at time of testing, since age was necessarily deter-
mined by the factor of primiparity. In ranking the drives, this
group may also be eliminated, and the standard age, maternal
group score be taken as the maximum score for the maternal drive,
under our standard conditions. (4) Passing the threshold into
compartment D was allowed only on the fourth and fifth crossings,
in testing the exploratory drive. As we have elsewhere stated, this

378 COMPARISON OF NORMAL DRIVES

variation in conditions favored the exploratory drive, and since
the score for this drive is extremely low even with this advantage,
this variation does not influence the rank order of the drives at all,
and so may be disregarded.

Much more important than the above mentioned slight varia-
tions in procedure certain problems raised by the methods of con-
trol utilized to keep other drives at a minimum while a given drive
was being tested, the comparability of the incentive-response
(Table 1) from drive to drive, and the validity of a test period of
set length. The questions raised here do not arise in so far as the
method is utilized in the measurement of a single drive, but are of
major consequence in any attempt to compare the scores and rank
the different drives. These points will now be discussed in order.

It is manifestly impossible to isolate, in any absolute sense, a
single drive in a complex organism like the white rat. Obviously,
the different physiological systems upon which drives are based are
more or less interdependent and operate simultaneously when the
organism is maintained in a normal or natural state. The best that
can be done is to keep drives, other than the one being tested in a
minimum status of quiesence, and such a status must be empiri-
cally determined for each drive and with reference to the general
living conditions which prevail during the investigation. Stand-
ardization of method covering the manner of securing the quies-
cence of a given drive is of as much importance as standardization
of test procedure. Furthermore, in a project such as the present
one in which the central aim is the ranking of drives, it is necessary
to make use of the same method of allaying a given drive through-
out the entire investigation.

Such standardization was accomplished in the present case in
the following manner: (1) Except when the hunger drive was itself
being tested, the animal was taken from a cage which had been
continuously supplied with the usual diet of the laboratory. (2)
Except when the thirst drive was itself being tested, the animal
was taken from a cage which had been continuously supplied with
water. (3) Except when the male sex drive was itself being tested,
the male had been segregated from females for 35 days, and at ap-
proximately this time the male sex drive is at a minimum within

COMPARISON OF NORMAL DRIVES 379

the limits of our tests on male sex deprivation. (See score for 28-
day sex deprived group, Part IV, 1.) (4) Except when the female
sex drive was itself being tested, only females in dioestrum, at
which stage the female sex drive is at a minimum, were used. This
does not apply, obviously, to the maternal groups, but the same
principle holds here, since females are sexually inactive while carry-
ing the young/ (5) Except when the exploratory drive was itself be-
ing tested, the animals were prevented from exploratory excitation
by the small size of compartment D, by the preliminary crossings
into compartments C and D, and by the dominance of appropri-
ate incentive objects in compartment D, whenever they entered
these compartments. The chief value of the Columbia Obstruction
Method as a means of measuring definite incenitve-drive conditions
lies in the directness and dominance of the incentive-object stimu-
lation. (See Warden, The Columbia Obstruction Method, and Part
I, 2.) Even if these controls were inadequate in maintaining at an
absolute minimum all drives, other than the one being tested, what-
ever lack of control there may have been operated uniformly
throughout and hence does not argue against the comparability of
the scores from drive to drive.

Some question may arise as to the comparability of the incentive-
response from drive to drive. The response of an organism to food,
water, sex object, etc., naturally differs considerably not only qual-
itatively, but also as to the directness with which it is called out
when brought in contact with the incentive object, as in compart-
ments C and D. The hungry or thirsty animal reacts promptly
enough to food or water, and likewise the maternal and exploratory
response is usually prompt ; at least this was true under our condi-
tions in which the animals had gotten over the normal tendency to
investigate a novel place during the five preliminary crossings into
these compartments. But the completed sex act is not prompt in
either male or female and we found it necessary to use some lati-
tude in deciding when the animal had made an appropriate sex re-
sponse so that it might be returned to the entrance compartment
again. Such preliminary sex activities as biting the sex object or
nosing the external genitalia rather than the act of copulation were
the criteria adopted. In the case of the female, the familiar re-

380 COMPARISON OF NORMAL DRIVES

sponse of elevating" the vulva under the nose of the male was con-
sidered sufficient. Mounting, or attempted mounting, in the male,
if it occurred before the usual nosing or biting was also checked as
a response to the sex object. The white rat exhibits no small degree
of individual difference in the matter of preliminary sex activities,
and it was necessary to use common sense in the matter of a cri-
terion in the case of both male and female sex drive. It might be
argued that, since the complete sex act was not permitted, the test
conditions for the sex drive differed in an important manner from
that for the other drives. But, as a matter of fact, deprivation dur-
ing the test is necessary was the rule for all the drives; the incen-
tive object was utilized as a stimulus to excite the drives rather
than as a means of satiation. In the case of hunger, only the merest
nibble of powdered food was permitted upon entering the incentive
compartment; only the damp nipple of the water bottle could be
licked in testing the thirst drive; only the merest attention to the
young was allowed in the maternal group, and only the slightest
running about was permitted in the exploration box in testing the
exploratory drive. We cannot, of course, contend that the stimula-
tion conditions from drive to drive were of the same degree, in so
far as the incentive-response in compartments C and D are con-
cerned, but we made use of the best criteria of response to the dif-
ferent types of incentive that we could devise after a large amount
of preliminary observation of the animals in the apparatus. Fur-
ther observation during the progress of the testing confirms us in
our opinion that, when taken in connection with our method of
scoring, to be next discussed, the criteria were sufficiently com-
parable as not to involve any question regarding the validity of
ranking the different drives. This view of the matter is further
strengthened by the fact that the ranking of the drives, in the last
analysis, will be based upon the maximum scores for each drive;
but, when a given drive is strongly aroused, the response to the in-
centive tends to be definite and prompt as might be expected.

The advantages of a test period of definite length has many ob-
vious advantages, and among these, the matter of ease in scoring is
of considerable importance. The amount of work done in a given
time, or the number of repetitions accomplished within a definite

COMPARISON OF NORMAL DRIVES 381

practice period are well established methods of scoring human per-
formance. The only questions that might arise regarding the valid-
ity of such a set test period as we used in this project (20 min.)
would be: (1) Was the test period long enough to include a suffi-
cient amount of behavior characteristic of a given drive to yield an
adequate index of the strength of the drive; (2) since the incen-
tive responses varied in promptness from drive to drive, was a test
period of the same length equally fair to the various drives inves-
tigated ?

The first question can be answered clearly in the positive. Pre-
liminary work indicated that a test period of only 10 minutes by
this method would give valid indices of drive conditions, the curves
following the same general lines as those obtained by our standard
20-minute test period. We have checked through the results of
several drives and have found that the ranking of drive conditions
would not be disturbed at all by discarding the results obtained
during the last half (10 min.) of the standard test period. In the
case of each drive, approaches, contacts, and crossings were checked
on the record card (see Fig. 3, Part IV, 1) for each minute of the
20-minute test period so that this matter is open to the most com-
plete analysis. The test period was set at 20 minutes so as to in-
sure the inclusion of any drive behavior of significance in the score.

The second point cannot be so clearly disposed of. For, as we
have said, the incentive response varied not only as to general type
from drive to drive, but also as to promptness, and this latter fac-
tor would be expected to influence the score since the test period
was limited. In order to equalize the value of a crossing to the in-
centive from drive to drive, the animal was allowed to remain in
the incentive compartment, after crossing over the grid, for not
longer than 30 seconds, regardless of whether he had met the cri-
terion of an appropriate incentive response for the given drive or
not. That is to say, the time allowed for an incentive response was
limited just as was the total period of the test. This was done with
a view to standardizing the method so as to take account of differ-
ences in promptness in responding to the incentive from drive to
drive. Again we must suggest that, in ranking the drives, we are
directly comparing the maximum scores only, and that when

382

COMPARISON OF NORMAL DRIVES

strongly aroused the response to the incentive tended to be prompt
and uniform in the case of each of the drives. As a matter of fact,
the act of crossing and making an appropriate response to the in-
centive did not seem to vary any more from drive to drive than
from one condition to another within the same drive. In any case,
the incentive response could not last longer than 30 seconds, and in
most cases actually occurred much sooner than that in each of the
drives, and especially so when these were at a maximum.

From the above discussion it will appear that our apparatus and
method was sufficiently well standardized to warrant the use of the
scores in ranking the normal drives as proposed.

Analysis of Results

Our main interest, in the present section, will be in a comparison
of the maximum scores of the different drives with a view to rank-
ing them in a definite order on the basis of strength and persis-
tence. Nevertheless, it seems best as a matter of convenience to the
reader to bring together in tabular form the results of each normal
drive condition investigated, and so to present the maximum scores
in their appropriate setting.

The averages for each of the normal groups, covering approaches,
contacts, and crossings, are given in Table 2, together with the
usual statistical values for the several averages. These scores rep-
resent the full 20-minute test period. Separate indices have been
entered for the two sexes in the case of hunger and thirst, in addi-
tion to the combined scores for these two drives.

In Table 3 will be found a set of values representing a grouping
of the scores on the basis of successive five-minute parts of the 20-
minute test period, and percentage scores, following the same
grouping comprise Table 4. A finer temporal distribution of the
scores (minute by minute) is available in the original reports, but
no summary of this analysis has been included at this point, since
even the five-minute grouping seems to be of little significance. The
data were taken and reported in these temporal groupings in the
hope that such temporal distribution would throw some light upon
the persistence of the drive during the test period. It was thought
possible that some useful measure of the persistence of a given

COMPARISON OF NORMAL DRIVES 383

drive might be thus secured, corresponding to the strength score
based upon the total score. However, it appears to be very doubtful
whether the distinction between strength and persistence of drive
is defensible — at least in so far as these may be indicated by total

TABLE 2

Showing results for the various drive conditions

APPROACHES

CONTACTS

CROSSINGS

DRIVE CONDITION TESTED

© h

ientof
ion

<D

IS:

* S

S3 B

ientof
ion

f
to

3

c
o

ientof
ion

8§

ho
■

SS

flic

•iat.

S3
U

H

.2

11

3 s

<

J'l

«Q

P *

<u

|1

WQ

g S

CD
>.

<

i5 «

1 3

O <3

of>

(Hunger-

—male)

1 9c:

0 period

10

1

5

1

34

2

2

1

20

55

2.7

3

37

2 days

10

3

3

47

44

4

9

1

59

32

16.1

6

5G

44

3 days

10

5

0

i

L

41

28

5

7

1

58

28

18.0

7

K2
o —

42

4 days

10

3

6

9
u

2n

61

2

8

1

17

42

19.1

5

87

34

6 days

10

3

6

1
1

59

44

3

0

1

55

52

14.2

5

98

42

8 days

10

4

G

1
X

7<i

on

oy

5

g

2

18

40

9.8

5

33

04

(Hunger-

—female)

0 period

10

2

0

1

34

f?7

3

5

1.55

44

2.1

2

07

QQ

ao

2 days

10

3

4

4

56

46

G

9

1.87

27

19.0

g

91

47

3 days

10

5

2

4

25

24

8

0

1

41

76

18.4

7

G3

44

4 days

10

5

2

2

04

39

5

7

1

10

19

17.0

5

92

35

6 days

10

G

1

4

77

9Q

zy

5

3

1

80

34

14.0

7

18

01

8 days

10

6

8

1

69

9K

7

3

2

29

31

6.0

4

G9

7C

(Hunger — combined)

■

116

0 period

20

1

75

i

X

3G

79

2

85

1

42

50

2.40

2

78

2 days

20

3

35

2

81

84

5

90

3

92

66

17.60

7

78

i o

44

3 days

20

5

10

2

4G

6

S5

2

93

43

18.20

7

58

42

4 days

20

4

40

2

37

£4

4

75

1

91

40

18.10

<j

J o

QO

oo

6 days

20

4

85

3

58

< 4

4

15

3

Gl

87

14.10

6

61

4.7
1 <

8 days

20

5

70

3

77

DO

G

40

5

71

89

7.90

5

07

04

(Thirst-

-male)

108

0 period

10

3

.0

1

6

53

1

7

2

1

123

3.7

4

0

1 day

10

7

3

3

3

45

5

6

5

4

96

21.1

11

6

54

2 days

10

5

.0

3

.7

74

A
-±

c
. O

4

3

95

i a 7
lb . i

12

.3

74

4 days

10

3

.0

1

8

60

1

8

2

0

111

12.7

9

.0

71

6 days

10

0

.8

0

.7

87

0

9

1

2

133

7.5

5

.5

73

(Thirst-

-female)

94

0 period

10

3

.1

2

9

0

.8

0

9

112

4.6

3

.8

83

1 day

10

7

.3

4

7

64

4

7

3

9

83

19.7

11

.1

56

2 days

10

4

0

3

5

87

3

8

3

4

89

15.3

11

.7

76

4 days

10

3

.2

3

3

103

1

.1

1

4

127

14.5

8

.3

57

6 days

10

1

.7

1

9

111

1

.8

1

9

105

6.9

G

.9

100

(Thirst-

-combined)

0 period

20

3.05

2

.3

75

1

25

1

3

104

4.15

3

.9

94

1 day

20

7

30

4

1

56

5

15

4

7

91

20.40

11

.4

56

2 days

20

4

50

3

6

80

4

.45

3

9

94

16.0

12

.0

75

4 days

20

3

10

2

7

87

1

.45

1

7

117

13.6

8

.7

64

6 days

20

1

25

1

5

120

1

.35

1

6

119

7.2

6

.0

83

384 COMPARISON OF NORMAL DRIVES

TABLE 2 (Continued)

(Sex — male)

0 period

20

1.2

1.36

113

0.35

0.7

200

3.6

4.08

114

6 hours

20

4.15

3.26

79

4.50

2.77

62

8.1

5.20

64

12 hours

20

2.70

2.03

75

2.65

1.98

75

12 .2

4. 80

39

1 day

20

2.30

2.35

102

1.5

1.5

100

13.45

4.03

30

1 day (control)

20

1.20

1.03

86

1.05

1.18

112

2.95

1.80

61

4 days

20

1.80

1.90

106

1.85

1.56

84

12.60

6.24

50

7 days

20

2.25

1.50

67

3.10

2.50

81

12.25

7.07

58

28 days

20

2.50

2.90

116

2.00

1.97

99

10.55

7.10

66

(Sex — female)

Early congestive

7

6.71

5.39

80

4.86

2.31

48

11.14

6.08

55

Oestrum (cornified)

21

5.43

3.63

67

3.14

2.73

87

14.14

5.14

36

Same (control)

22

3.23

2.70

84

2.27

1.91

84

5.05

2.55

51

Late cornified

3

2.33

1.39

60

3.66

1.39

38

8.66

5.55

64

Post-ovulative

6

1.17

1.04

89

4.00

2.92

73

4.66

5 .28

113

Dioestrum (4 stages)

32

1.16

1.42

117

1.47

1.27

86

1.34

1.57

117

(Maternal)

First litter

9

3.89

2.68

69

0.78

0.63

81

28.33

7.73

27

Muciparous (stand, age)

10

4.10

3.67

89

1.30

1.90

146

22.40

9.14

41

(Exploratory)

4.89

81

Males (stand, age)

20

5.20

3.85

74

1.8

1.66

92

6.00

score and temporal distribution respectively. It is true, as an ex-
amination of Tables 3 and 4 will show, that when a drive is at its
maximum as indicated by the maximum score in crossings, the
score tends to be uniformly maintained throughout the test period ;
the one exception to be noted is the maximum score for the thirst
drive. Scores representing other than maximum conditions of the
drives usually show either an increase or a decrease through suc-
cessive five-minute parts of the test period. But what this may
mean in terms of the loose concept of persistence of drive is not at
all clear, and we shall omit further discussion of these data from
this section.

In comparing the various normal drives we shall limit ourselves
to the scores for crossings, in which a complete response to the ob-
struction-incentive situation occurred. As a matter of fact, this
plan was followed in the original reports on the several drives,
since the other two types of response (approaches, contacts) ap-
peared to be of little or no significance as between different drive
conditions. From the first, approaches and contacts were checked
in the hope that these partial crossings would reveal something of
importance regarding strength of drive, and the practice was con-
tinued in the interests of uniformity in all later studies. However,

COMPARISON OF NORMAL DRIVES 385

TABLE 3

Showing temporal distribution in five-minute intervals of approaches, contacts,
and crossings during the twenty-minute test period

DRIVE CONDITION TESTED

AP

l'ROACHES

CONTACTS

CROSSINGS

Minutes of
test, period

Minutes of
test period

Minutes of
test period

0-5

6-

10

il-
ls

16-
20

0-5

0-
10

il-
ls

16-
20

0-5

6-
10

il-
ls

16-
20

(Hunger — combined)

—

—

0 days
2 days

15

in

7

3

28

IS

4

7

31

9

5

3

1 7

0

i s

99

•)•)

- o

49

37
o i

1 QO

1 OQ

iuy

oO

OO

3 days

4 days
6 days
8 days

42

25

17

18

54

37

28

18

75

91

100

98

23

34

10

21

41

22

8

14

94

89

87

91

28

20

18

31

18

11

13

33

64

65

77

76

30

28

16

40

27

29

29

46

31

40

42

45

(Thirst — combined)

0 days

1 day

2 days
4 days
6 days

20

18

13

10

8

0

4

7

42

21

7

13

74

31

22

19

38

21

20

10

153

106

84

65

24

31

16

20

28

22

14

19

153

81

60

26

1 8

\ 7

90

7

1 4

-LI

D
•J

a
u

1 1 ^

J- LO

70

48

3Q

8

s

4

5

10

G

4

7

39

31

42

32

(Sex — male)

—

—

0 hours

6 hours
12 hours

X day

1 day (control)
4 days

7 days
28 days

4

3

12

5

2

2

2

1

10

19

18

25

2G

17

24

.16

31

22

19

18

52

61

28

21

0

22

15

7

13

13

8

19

74

76

44

50

19

10

6

5

9

9

8

4

76

63

60

79

o
6

15

8

Q

y
9

Q

y
4

o
o

12

i

5

o
O

8

Q
O

3

98

62

1 i

55

Q

y
65

0

70

12

11

11

9

17

9

24

12

47

33

58

107

9

8

15

18

4

9

22

5

16

40

54

101

(Sex — female)

Early congestive

11

17

13

6

6

3

8

17

11

18

21

28

Oestrum (cornifled)

49

29

23

13

8

10

16

Of?

26

59

45

81

22

Same (control)

19

32

14

6

17

13

15

5

48

36

17

10

Late cornified

3

1

1

. 2

4

1

3

3

5

3

7

11

Post-ovulative

3

2

1

1

5

9

9

1

9

10

3

5

Dioestrum

12

17

5

3

15

IS

9

5

18

13

9

3

(Maternal)

First litter

Muciparous (Standard age)

11

12

1G
14

29

5
8

3
7

4
1

1

5

1

5

1

2

64

63

59
50

65
53

67
58

20

(Exploratory)
Males (Standard age)

38

19

18

12

12

4

8

48

29

23

we can find no indication of any significance in these two types of
partial crossing to the incentive. From an inspection of the values
for approaches and contacts in Table 2, it appears that they exhibit
relatively slight variations from one condition to another within
the same drive, and from one drive to another, sharply contrasting
with markedly variable crossing scores. There seems to be no legiti-

386 COMPARISON OF NORMAL DRIVES

mate method by which the three types of reaction to the obstruc-
tion-incentive situation might be combined into a single index of
strength of drive. A crossing always includes, of course, both an
approach to, and a contact with the obstruction (grid). It is logical
to suppose that any type of score for partial crossings would be
reduced by actual crossings, as when the drive is strong. There is

TABLE 4

Showing temporal distribution in five^minute intervals of approaches, contacts,
and crossings, (percentages) during the twenty -minute test period

DRIVE CONDITION TESTED

APPROACHES

Minutes of
test period

6-

ii-

16-

6-

il-

16-

6-

il-

16-

0-5

10

is

20

0-5

10

ls

20

0-5

10

ls

20
—

(Hunger — combined)
0 days

2 days

3 days

4 days
6 days
8 days

43

28

20

9

49

32

7

12

65

19

10

6

26

12

28

34

14

19

36

31

37

29

25

9

41

24

17

18

39

27

21

13

21

25

27

27

ZD

SO

1 9
1Z

OA

Z4

lb

Zo

y

1 7
1 I

on
ZD

25

24

OK
ZD

OQ

91

lo

9 9

3Z

99

Zo

1 7
1 (

A 9

4Z

23

23

2 1

97

z (

OO

zb

9C
40

-1 A

14

35

91

Zl

99

2Z

O O

2_

35

20

25

27

oo
Zo

(Thirst — combined)

0 days

1 day

2 days
4 days
6 days

33

30

21

16

32

24

16

28

51

25

8

16

51

- 1

15

13

41

26

22

11

37

26

21

16

26

a 4

18

zz

O A

34

26

17

23

48

25

19

8

29

28

32

11

48

in

21

21

42

26

18

14

32

32

16

20

37

22

15

26

28

22

28

22

(Sex — male)

50

29

0 hours

6 hours
12 hours

1 day

1 day (control)
4 days

7 days
28 days

17

12

21

29

28

14

14

26

25

35

31

21

29

19

35

24

21

20

32

38

17

13

12

44

30

14

24

25

15

36

30

31

18

21

41

35

13

11

30

30

27

13

28

24

22

26

12
42

16
22

36
25

36
11

38
43

34
18

14

28

14
11

48
24

29
22

15
26

8
28

28

26

25

21

27

15

39

19

19

13

24

44

18

16

30

36

10

22

55

13

8

19

25

48

(Sex — female)
Early congestive
Oestrum (cornified)
Same (control)
Late cornified

23
43

36
26

28
20

13
11

18
12

9
24

23
24

50
40

14
19

23
15

27
26

36
40

27

45

20

8

34

26

30

10

43

32

15

10

43

14

14

29

36

10

27

27

19

12

27

42

Post-ovulative

43

29

14

14

21

38

37

4

32

36

11

21

Dioestrum (4 stages)

32

46

14

8

32

38

19

11

42

30

21

7

(Maternal)

First litter

29

34

20

17

8

38

38

16

28

22

24

26

Multiparous (Standard age)

31

46

14

9

58

14

14

14

25

23

26

26

(Exploratory)

Males (Standard age)

37

28

18

17

34

33

11

22

40

24

19

17

Minutes of
test period

CROSSINGS

Minutes of
test period

COMPARISON OF NORMAL DRIVES 387

some evidence for an inverse relationship between scores for partial
crossings, particularly approaches, and actual crossings, although
this relationship is not very consistent. This argues against any
method of weighting the three types of incentive response and re-
sorting to pooling in order to reduce all to a common index of
strength of drive, aside from the element of arbitrariness that
would enter into any such weighting of scores that are so disparate
in significance. We did pool the averages for approaches, contacts,
and crossings, without weighting, as a matter of information to
ourselves, comparing these pooled scores with the scores for cross-
ings, and while such a pooling is certainly not legitimate, we may
as well report the result. There was no displacement of rank among
the group averages of a given drive, except in the case of the male
and female sex drive, where the grouping represents finer divisions
in experimental analysis, and even here the displacements were
based upon slight differences. This comparison serves at least to
show the neutral character of partial crossing scores and their con-
sequent insignificance in any index of drive. We feel justified in
suggesting that, in later studies of drive in the white rat by means
of the Columbia Obstruction Method, scoring be limited to cross-
ings only, since it seems a mere waste of time to take account of
partial crossings of any sort.

Since we are interested at this point primarily in the maximum
scores as these may be useful in ranking the normal drives, further
discussion of Tables 2, 3, and 4 will be unnecessary. In so far as
the values there summarized relate to a particular drive, detailed
analyses and comparisons will be found in the full reports under
the appropriate sections. We may turn at once to an examination
of the maximum scores covering the number of crossings to the in-
centive, considering the same a common index for the ranking of
the various drives. It will be convenient to deal separately with
the group of male and of female drives, as well as with the group
of drives in which the sexes are combined.

The rank order of the several drives, according to these group-
ings will be found in Table 5, together with the maximum score in
crossings upon which the ranking was based. Since, as we shall see,
the sex differences for the hunger, thirst, and sex drives are too

388 COMPARISON OF NORMAL DRIVES

small to be of much significance, the ranking of the drives in the
last grouping of the table (male and female combined) are of most
interest to us.

Before dealing with the reliability of the maximum scores in
comparisons from drive to drive, it may be well to discuss briefly
the significance of the maximum scores for a given drive. In each
case these indicate the number of crossings to the incentive when
the drive was most strongly aroused. Except for the maternal and
the exploratory drives, no less than five conditions, representing
systematic changes in the variable, were tested in the case of each
drive. The curves covering the successive changes in the variable
(water, food, sex deprivation, etc.) comprise Figures 1, 2, and 3,
of this study. The first value on each curve represents an approxi-
mate physiological zero for the given drive. Thus hunger and thirst
begin with zero deprivation; male sex with a zero established by
copulating with a female in heat until completely satiated (see
Part IV, 1), while dioestrum constituted a zero point for the fe-
male sex drive. In thirst, hunger, and male sex, deprivation was
carried far beyond the point of the maximum drive condition as
the curves clearly indicate ; each of the eight recognized stages of
the oestrus cycle were tested in the female. We are able to say,
then, that for each of these four drives, we secured the true maxi-
mum score under our conditions, within the limits of the unit of
change adopted for the several variables. A finer unit of change
might show the true maximum point to lie somewhere near rather
than precisely at the points we indicate in each instance. For ex-
ample, it is possible that a finer working of the field might result
in showing that the maximum point for thirst is somewhat less or
somewhat more than the one-day point where our maximum score
was found ; and so with the other drives. But, in any case, it must
be admitted that our maximum points are approximately true, and
are clearly established as true by the general trend of the scores
downward in either direction from these maximum points, as may
be observed in Figures 1, 2, and 3.

Only a single condition was tested in the case of the maternal
drive, and the exploratory drive, under our normal standardiza-
tion, and therefore, we cannot insist that the scores included for

C 031 PARIS ON OF NORMAL DRIVES

389

390 COMPARISON OF NORMAL DRIVES

Dioettrwn

o I . I __| I , , , ,

0 ** *± «U Si. bd Id. i4r

Fig. 2. The curves represent the different drive-conditions tested in the female. The
several periods of deprivation (thirst, hunger) are indicated on the abscissa and the
average number of crossings on the ordinate.

these two drives are true maxima. So far as the maternal drive is
concerned, it is quite doubtful whether the score in Table 5 is the
highest possible maternal score, even under our standard conditions.
The score for first litter is considerably higher (28.33) but this
score could not be used here because the animals tested were nat-
urally younger than our standard age. However, there is one im-
portant factor that might be varied in animals of standard age
which might give a higher score than we secured, e.g., length of
time between parturition and testing. Our standard age, maternal
group were tested on the average 15 hours (range 12-20 hours)
after casting the litter, and this may not have been the best time

392 COMPARISON OF NORMAL DRIVES

to have tested them so far as strength of maternal drive is con-
cerned. But, the admission of the possibility of an even higher
score than we secured for the maternal drive does not disturb our
table of rankings, since the maximum score we show is high enough
to place the maternal drive at the top of the list any way. On the
other hand, the score for the exploratory drive is so low that fur-
ther tests are not likely to shift it from the position of the lowest
in our table of ranks. So far, then, as general considerations are in-
volved, the ranking of the drives as shown in Table 5, and as indi-
cated by the graphs of Figures 1, 2, and 3, must be regarded as
based upon sound principles. This is of importance as establishing
the basic method of comparing drives in terms of a common type
of response to a well standardized obstruction-incentive situation.

The relation of the maximum scores for each drive to the scores ,
obtained when the drive was tested at other than maximum strength
is shown in Table 6. In every instance, the difference between the
maximum score and some one or more scores on a given drive curve
at points on either side of the maximum point are large enough to
pass the most rigorous test of statistical validity. It is true that
the differences between the maximum point and points near the
maximum are, in many cases, too small to be statistically valid, but
such small differences must be interpreted not independently but
in the light of the general trend of the scores on either side of the
maximum. The clear cut trends in every instance establish the
validity of the maximum score within the limitation of the unit of
change in the variable, or the fineness of the experimental analysis,
for a given drive.

Accepting the maximum scores as valid indices of the strength of
the several drives, we must now inquire as to the significance of the
differences between these scores from drive to drive. As previously
mentioned, the ranking of the drives upon the basis of the maxi-
mum score is indicated in Table 5, and can also be seen by an in-
spection of the graphs of Figures 1, 2, and 3. The reliability of the
differences between the maximum scores for the several male drives
will be found in Table 7; similar values for the several female
drives comprise Table 8, while Table 10 covers the same point with

COMPARISON OF NORMAL DRIVES

393

the drives grouped together without regard to sex. The relations
of these reliabilities to rank order are summarized in Table 11.

Before discussing these data we shall examine Table 9 and dis-
pose of the matter of sex differences in strength of drive. Separate
scores for the sexes are available in the thirst, hunger, and sex
drives only. The difference in both hunger and thirst favors the
male, but is too small to be considered significant, while the sex
drive is higher in the female by only a small margin of questionable
validity. These findings refer only to the height of the curve for
the drives and do not imply that the maximum point was identical

TABLE 5

Showing maximum scores (crossings) for the various drives, arranged in ranlc

order

CONDITION

SIZE OF GROTTP

COEFFICIENT

DRIVES TESTED

OF

AVERAGE

STANDARD

OF

MAXIMUM DRIVE

M.

F.

Comb.

DEVIATION

VARIATION

(Male)

Thirst

2nd day

10

21.10

1.1.60

54

Hunger

4th day

10

19.10

5.87

31

Sex

1st day

20

13.45

4.03

30

Exploratory

Only 1 tested

20

6.00

4.89

81

(Female)

Maternal

Standard age

10

22.40

9.14

41

Thirst

1st day

10

19.70

11.10

56

Hunger

3rd day

10

19.00

8.91

47

Sex

Oestrum

21

14.14

5.14

36

(Combined)

Maternal

Standard age

10

22.40

9.14

41

Thirst

1st day

10

10

20

20.40

11.40

56

Hunger

3rd day

10

10

20

18.20

7.58

42

Sex

1st day & Oestrum

20

21

41

13.80

4.64

34

Exploratory

Only 1 tested

20

6.00

4.89

81

for the two sexes. As will be seen by comparing the graphs of Fig-
ures 1 and 2, the peak of the thirst drive came at the same point (1
day) for both sexes, but this was not true in the case of the hunger
drive. The peak for the females, in the latter drive, came after 2
days of starvation, while that for males came after 4 days of star-
vation. From the showing in Table 9, we may conclude that the
hunger, thirst, and sex drives are about equally strong in the two
sexes when at their maximum.

A detailed examination of Tables 7, 8, and 10 does not seem to be
especially called for, since the main results in so far as the matter
of the validity of the rank order of the various groupings of drives

394 COMPARISON OF NORMAL DRIVES

TABLE 6

Showing the reliability of the difference between the maximum score (cross-
ings) for each drive and all the other scores for the given drives

DRIVE CONDITIONS
TESTED

S. D. OF
DIFFERENCE

DIFFERENCE
BETWEEN
AVERAGES

DIFFERENCE

CHANCES
IN 100 OF

A TRUE
DIFFERENCE

S. D. OF
DIFFERENCE

(Hunger — male)
4 days and 0 days
4 days and 2 days

A rlfi'va onrl 3 ^ovq
tc tlcl V £> ctlXU. O Ud Vo

4 days and 6 days
4 days and 8 days

2.14
2.78
3 09
2.65
2.51

16.4
3.0
x.x
4.9
9.3

7.66
1.08
o q«

1.85
3.71

100

86

RA
04:

97
100

(Hunger — female)
2 days and 0 days
2 days and 3 days

4 tldj'H ctllU. "± tidy a

2 days and 6 days
2 days and 8 days

2.89
3.71

3 38

O.OO

3.62
3.18

16.9
0.6
9 o

5.0
13.0 ,

5.84
0.16
o

1.38
4.08

100

56

79

92
100

(Hunger — combined)
3 days and 0 days

o udj'o diiti u tidy a

3 days and 4 days
3 days and 6 days
3 days and 8 days

1.81

9 A.1

2.16
2.25
2.04

15.8
o r

0.1
4.1
10.3

8.73

0 9^

0.05
1.82
5.05

100

R(\

52
97
100

(Thirst — male)
x tidy diiti u tidys
1 day and 2 days
1 day and 4 days
1 day and 6 days

O.OO

5.35
4.64
4.06

17 A
X 1 .t

4.4
8.4
13.6

A AH

0.82
1.81
3.35

1 00

80
96
100

(Thirst — female)

X tidy dlltl u (Id jo

1 day and 2 days
1 day and 4 days
1 day and 6 days

o. 1 X

5.10
4.38
4.13

xo.x
4.4
5.2

12.8

A 017

0.86
1.19
3.10

1 00

xw
81
88

100

Thirst — combined)
1 day and 0 days
1 day and 2 days
1 day and 4 days

X tidy dlltl U tldjS

2.69
3.70
3.21
9 88

16.3
4.4
6.8

13 9

6.06
1.19
2.12

A ^8

100
88
98

1 00

(Sex — male)
1 day and 0 hours
1 day and 6 hours
1 day and 12 hours
1 day and 4 days
1 day and 7 days
1 day and 28 days
1 day and control

1.28
1.47
i ac\

1.65
1.82
1.83
0.99

9.9
5.4

i 3
x.o

0.9

1.2

2.9

10.5

7.68
3.64

0 80

0.51
0.66
1.60
10.61

100
100

81

69
75
95
100

(Sex — female)
Cornified and control
Cornified and dioestrum
Control and dioestrum

1.25
1.16
0.61

9.1

12.8
3.7

7.28
11.00
6.08

100
100
100

(Maternal)
First litter and standard age

3.87

5.9

1.53

94

COMPARISON OF NORMAL DRIVES 395

have been summarized in Table 11. In column 1, of this table, the
rank order for the several male drives tested, based upon the max-
imum scores is given; following this the rank order of the female
drives, and finally the rank order for the drives as such without
regard to sex. That is, the maximum scores for the two sexes cov-
ering the hunger, thirst, and sex drives are here combined ; we have
already found, in connection with the discussion of Table 9, that
there is apparently no significant sex difference in maximum score,

TABLE 7

Showing the reliability of the difference between maximum scores (crossings)
of the different drives in males

DRIVE CONDITIONS
TESTED

S. D. OF
DIFFERENCE

DIFFERENCE
BETWEEN
AVERAGES

DIFFERENCE

S. D. OF
DIFFERENCE

CHANCES
IN 100 OF

A TRITE
DIFFERENCE

Hunger and thirst

4.11

2.00

0.49

69

Hunger and sex

2.06

5.65 +

2.74

99.7

Hunger and exploration

2.15

13.10 +

6.08

100

Thirst and sex

3.78

7.65 +

2.03

98

Thirst and exploration

3.83

15.10 +

3.95

100

Sex and exploration

1.42

7.45+

5.26

100

Legend: The plus sign indicates that the- first drive of the pair has the
higher score.

TABLE 8

Showing the reliability of the difference between the maximum scores (cross-
ings) of the different drives in females

DRIVE CONDITIONS
TESTED

S. D. OF
DIFFERENCE

DIFFERENCE
BETWEEN
AVERAGES

DIFFERENCE

S. D. OF
DIFFERENCE

CHANCES
IN 100 OF

A TRUE
DIFFERENCE

Hunger and thirst

4.50

0.70

0.16

56

Hunger and sex

3.03

4.86 +

1.60

95

Hunger and maternal

4.04

3.40

0.84

80

Thirst and sex

3.69

5.56+

1.51

93

Thirst and maternal

4.55

2.70

0.59

72

Sex and maternal

3.10

8.26

2.66

99.6

Legend: The plus sign indicates that the score is higher than the female
score.

TABLE 9

Showing the reliability of the difference between the maximum scores (cross-
ings) of the male and female groups

DRIVE CONDITIONS
TESTED

S. D. OF
DIFFERENCE

DIFFERENCE
BETWEEN
AVERAGES

DIFFERENCE

S. D. OF
DIFFERENCE

CHANCES
IN 100 OF

A TRUE
DIFFERENCE

Hunger

3.37

0.10 +

0.03

51

Thirst

5.08

1.40 +

0.28

61

Sex

1.44

0.69

0.48

68

Legend : The plus sign indicates that the male score is higher than the female

score.

396 COMPARISON OF NORMAL DRIVES

so that such combination seems thoroughly justifiable. The reliabil-
ity of the difference between a given drive score (maximum) and
that for each drive of lower rank, taken singly, in terms of proba-
bility is shown for each drive in the next four columns of the table.
The position of the sex drive, both male and female, as lower than
either hunger or thirst, appears to be practically certain by this
method of comparison. The exploratory drive is shown to be cer-
tainly lower than any of the other drives. The probability that
thirst is higher than hunger is the least satisfactory, while the
chances that maternal is higher than hunger and thirst, taken
singly, are not any too good.

The problem of the validity of the rank order may be dealt with
by the method of combining averages (see, Yule, pp. 115, and 142)
rather than by the method of single comparisons. By this method
the score of a given drive is compared, not with the score of the
next lowest drive, but with the average score of all the drives of
lower rank. This method is more pertinent to the present discussion
because it deals simply with the factor of rank position within a
group of values, such as we have here. The probability of the or-
dinal positions worked out by this method is much higher than that
obtained by comparing the obtained differences singly, as will ap-
pear from an inspection of the last column of Table 11. The rank
order as listed in this table amounts to practical certainty for the
group of male drives, and for the group of drives in combination,
while the chances that the listed rank order for the female drives
is a true ranking is very high. The most general statement that
can be made, then, is that for the white rat, tested under our con-
ditions, and within the limitations of the scope of our investigation,
the normal drives rank as follows in strength: Maternal, thirst,
hunger, sex, exploratory.

The fact that the maternal drive ranks highest comes as some-
thing of a surprise, since several investigators using other methods
have rated the maternal tendency rather low (see Part V, 1). In
some cases, at least, the normal maternal drive has been more or
less disturbed by the method employed. A common error in tech-
nique is to have the litter born in the living cage, and then transfer
it for test to the apparatus used. We tested a group, following this

COMPARISON OF NORMAL DRIVES 397

procedure, and obtained a score only about half as high (11.6, see
Table 3, Part V, 1) as in our regular group in which the small
maternity cage in which the litter had been born was attached di-
rectly to the apparatus, as compartment D. It may be argued that,
in such a case, we are testing the maternal drive plus the tendency
to return to the home nest. But the truth seems to be that when
the normal conditions of maternity are disturbed by transferring

TABLE 10

Showing the reliability of the difference between the maximum scores (cross-
ings) of the different drives with sexes combined

DRIVE CONDITIONS
TESTED

S. D. OF
DIFFERENCE

DIFFERENCE
BETWEEN
AVERAGES

DIFFERENCE

S. D. OF
DIFFERENCE

CHANCES
IN 100 OF

A TRUE
DIFFERENCE

Hunger and thirst

3.06

2.20

0.72

76

Hunger and sex

1.84

4.40 +

2.39

99

Hunger and maternal

3.35

4.20

1.25

89

Hunger and exploration

2.02

12.20 +

6.03

100

Thirst and sex

2.65

6.60 +

2.49

99.4

Thirst and maternal

3.85

2.00

0.52

70

Thirst and exploration

2.77

14.40+

5.19

100

Sex and maternal

2.98

8.60

2.89

99.8

Sex and exploration

1.31

- 7.80 +

5.95

100

Maternal and exploration

3.09

16.40 +

5.31

.100

Legend: The plus sign indicates that the first drive of the pair has the
higher score.

TABLE 11

Showing the reliability of the differences between maximum scores, as related
to rank order. The difference between a given score and that for each drive of
lower rank, taken singly, is treated in columns 2, 3, 4, and 5 ; in the last column
the scores for all the drives of lower rank have been combined

DRIVE TESTED

RANK ORDER

CHANCES IN 100 OF A TRUE DIFFERENCE GREATER THAN 0

THIRST

HUNGER

SEX

EXPLORATORY

COMBINATION

(Male)

69

Thirst

1.

98

100

99.4

Hunger

2.

99.7

100

100

Sex

3.

100

100

Exploratory

4.

(Female)

72

80

Maternal

1.

99.6

96

Thirst

2.

56

89

86

Hunger

3.

95

95

Sex

4.

(Combined)

Maternal

1.

70

89

99.8

100

99.6

Thirst

2.

76

99.4

100

99.7

Hunger

3.

99

1,00

100

Sex

4.

100

100

Exploratory

5.

398 COMPARISON OF NORMAL DRIVES

the litter to an unfamiliar apparatus, the tendency to explore the
novel surroundings of the litter interferes with the maternal drive.
Then, too, many animals ignore young that have been moved about
— this is true of many species of birds such as the quail. Our pur-
pose was to test the maternal drive, along with the other drives dis-
cussed in this section, under normal conditions, in order to secure
maximum indices, and we feel that our ranking of the maternal
drive is the only defensible one from this point of view. The fact
that the score for first litter, maternal drive, was higher than the
score for standard age, maternal is a further argument for the
placing of the maternal drive at the top of the list. We have not
used the score for first litter in the comparisons of this section rel-
ative to ranking the several drives, since the animals were younger
than standard age.

The conclusions drawn above regarding the relative rank of the
several normal drives in the white rat, must be, of course, qualified
by the general and specific conditions representing the standardi-
zation of the project as a whole. In the strict sense, these conclu-
sions apply only to the adult white rat, of the strain and age em-
ployed, and reared unsegregated as to sex until approximately the
standard test age. The fact should be recognized that relative
strength of drive is very likely a function of the age of the animal
within wide limits, and there may be other factors also that enter
here, such as strain, the precise incentive object used, and specific
aspects of our method of testing and scoring. Our results, at any
rate, justify the extension of the Columbia Obstruction Method, as
a means of analyzing the dynamics of behavior, to other species,
and further work in this direction has been planned by this labora-
tory for the immediate future. The importance of indices of mo-
tivation for typical animal forms, and secured by the same method
of testing, can scarcely be overestimated for systematic compara-
tive psychology.

APPENDICES

APPENDICES

INTRODUCTION
C. J. Warden

The following four papers have not been included in the main
body of this report for the reason that the Columbia Obstruction
Method as standardized for use in the general project was not em-
ployed. They are, however, studies of motivation in the white rat,
and, except for the first study, were financed out of the regular
project allotment.

As noted (see Warden, The Columbia Obstruction Method), the
original obstruction apparatus, as devised by Jenkins and Warden
was first used by Holden in 1924-25 in the study of the hunger
drive, previous to the beginning of the project itself. Important
changes in both apparatus and procedure were introduced later
in further standardizing the method for use in the project. Never-
theless, this study of Holden was closely connected with the de-
velopment of the project and belongs very properly in the present
collection.

The paper by Warden and Hamilton (Appendix 2) and the later
paper of Hamilton (Appendix 3) are concerned with the effect of
delayed incentive in maze learning. The shorter paper may be con-
sidered as a preliminary attack upon this problem, which was more
systematically and adequately dealt with in the longer report. The
latter is of special interest inasmuch as it was done in connection
with, and parallels the experiment on the effect of delayed incentive
on the hunger drive as measured by the Columbia Obstruction
Method (Part II, 2). This fact makes possible a comparison of
these two essentially different methods, with respect to the influ-
ence of delay in the function tested.

The final paper is concerned with the relative value of reward
and punishment as motivation situations in the Yerkes-Watson

401

402

APPENDICES

discrimination apparatus. This study was intended to be the be-
ginning of a comprehensive comparison of these two opposite types
of incentive, but this plan was later discontinued when it appeared
desirable to concentrate the work along the main line of investiga-
tion.

1. A STUDY OF THE EFFECT OF STARVATION
UPON BEHAVIOR BY MEANS OF THE
OBSTRUCTION METHOD 1

Frances Holden
I. Introduction

The present study is an attempt to measure the drive,2 or moti-
vation value, of starvation in the white rat in terms of objective
behavior. For this purpose the Obstruction Method has been em-
ployed. This method involves the principle of placing an obstruc-
tion of some sort between the animal to be tested and an incentive
stimulus, selected with reference to the drive under investigation.
The animal is required to pass over the obstruction (in the present
case an electric grill) in order to reach the incentive stimulus and
thus satisfy the drive dominant at the moment. A quantitative
determination of the strength of the drive can thus be secured
in terms of the behavior of the animal toward the obstruction-
incentive situation. By keeping the obstruction constant and vary-
ing systematically the physiological state of the organism from which
the drive results (as would occur for example in the case of the
hunger drive with a series of starvation periods of different length)
it is possible to obtain indices of the motivation value of different
organic states in relation to a specific stimulus. By keeping the
obstruction constant it is also possilbe to obtain comparable data
concerning the relative strength of different drives. A complete
description of this method, including certain modifications intro-

1Eeprinted with modifications from Comparative Psychology Monographs,
1926, 3: No. 17, 45 pp.

'2 The term hunger drive as used in this study, means nothing more than that
the animal displayed a food searching response. The term carries no impli-
cations whatsoever concerning the presence or absence of so called 11 hunger
contractions 1 1 of the stomach, or concerning the subjective state of the organ-
ism.

403

404

APPENDICES

duced after my study was completed, has recently been published
(6).

Very little work has been done up to the present in the field of
animal motivation. The learning method has been employed for
the most part. The recent work of Moss (7) in which he developed
the Resistance Method bears the most direct relation to the present
study. Both the Resistance Method of Moss, and the Columbia
Obstruction Method, as devised by Jenkins and Warden for use in
my work are similar in that both make use of the principle of plac-
ing an obstruction of some sort between the test animal and the
incentive. The results of Moss are open to criticism not only because
his work was not well controlled, but also because he used very
small groups. He does not state the age of his animals, which is
doubtless an important factor in the matter of strength of drive.
Instead of using the method of equivalent groups, he tested the
same animals over again, after increasing the length of the starva-
tion period, so that his conclusions regarding the influence of this
factor can hardly be considered valid. His investigation must be
considered, however, very suggestive and valuable.

Moss also made use of the method of choice, in which food and a
sex object were placed before the animal simultaneously, with free
access to either allowed. This method has been further developed
by Tsai (15) and used by him in comparing a 24-hour hunger
period with the sex drive. The choice apparatus has been greatly
improved by Jenkins of this laboratory, and the new design along
with a criticism of the method has been recently published (6).
, Moss further considered the relative effect of different drives on
maze learning. This method had, however, been previously employed
by Szymanski (13, 14) and Simmons (10). Szymanski used the
maze both for the determination of the relative effects of different
drives on learning and for the study of the effect of different hunger
periods on learning. He found in the latter case that with a 30-
minute, 4-hour, and 24-hour hunger period, the most rapid learning
occurred with the 24-hour period, while the 30-minute period was
least effective.

Simmons used both a simple and complex maze for a comparison
of the effect of different food incentives and of different drives on

APPENDICES

405

learning. Since a constant hunger period was used throughout her
study the results have little bearing on the present work.

Dodson (3) studied the effect of different periods of starvation
on the development of a dark-light discrimination habit in the white
rat. He used starvation periods of 24, 31, 41, and 48 hours and
found that for rats 78 days old, the optimum starvation period for
habit formation of this type was 41 hours; however for retention
of the maze habit after 21 days, the 48-hour period appeared to be
more effective.

Studies of the effect of different periods of starvation on general
bodily activity as measured by an " activity cage ' ' have been made
by Richter and Nicholls. Richter (9) found that white rats lived 8
days when water but no food was provided, reaching a maximum
degree of activity between 2 and 3 days. After three days the
activity decreased gradually until death which occurred on the
eighth day. When neither water nor food was provided the rats
showed an immediate decrease in activity which was continuous
until death which occurred on the fifth day. A similar study was
made by Nicholls (8) upon the guinea pig. She found that the
maximum degree of activity occurred on the second day of starva-
tion and was followed by a marked decline on the third day.

Of the above investigators it appears that Szymanski, Dodson,
Richter, Nicholls and Moss have each considered the effect of a
series of starvation periods of different length upon behavior. Moss
alone, however, used a method in any way comparable with the
present one. The other investigators either measured general activ-
ity by means of some form of " activity cage" or investigated the
relation of starvation to speed of habit formation. Essential differ-
ences in method make a comparison of these results impossible.
From the small amount of work so far reported it is evident that
the relative motivation value of different starvation periods has not
been adequately determined.

The present investigation includes two main problems. The first
is the determination of the effect of systematic variation of the
period of starvation upon drive behavior when a constant amount
of electrical stimulation is employed in the obstruction section. The
second deals with the relative effect of three different degrees of

406

APPENDICES

electrical stimulation in the obstruction compartment upon the
drive behavior resulting from a series of starvation periods.

II. Apparatus and Procedure

The apparatus used in this study consisted of a control box for
testing the animal and a mechanism for the control of the current
supplied to the grill. The general plan of this apparatus is shown

Fig. 1. Diagram of Control Box and Accessory Mechanisms (Wiring

Omitted)

A, entrance compartment; B, obstruction compartment; C, incentive com-
partment; D and E, glass plates; F, potentiometer; G, voltmeter; H, switch
controlling voltmeter; I, multipoint switches controlling each resistance unit;
J, microammeter ; E, intake switch; L, L switches throwing current into grills.

in figure 1. A second control box was also provided in order that
two animals might be tested simultaneously. Each control box was
divided into 3 sections, each of which was 10 inches wide, 10 inches
long and 10 inches high. The entire box was painted a dull black.
Sections A and C were made identical in every respect in order that
any possible preference for either section might be obviated.

APPENDICES

407

Section A was used throughout as the entrance compartment,
section C as the incentive compartment. The floor of section B was
made of a slab of bakelite 0.5 inches thick, wound with no. 18
copper wire at intervals of 0.25 inches in such a way that the wires
from each terminal alternated, thus insuring a shock when any two
adjacent wires were touched. The terminals of these wires were
connected with the electric control apparatus. There were no divi-
sions between the three sections (A, B, C) of the box but at D and
E glass plates 6 inches wide and 12 inches high were placed as indi-
cated in figure 1, in order to keep the animals from leaping across
the grill. Section A was covered with a pane of glass to keep the
animals from jumping out of the box.

In the box used by Moss the entry and incentive compartments
were not similar. In the first place the entrance compartment was
larger than the incentive compartment. In the second place the
floor of the entrance compartment was covered with water, whereas
the floor of the incentive compartment was dry. Since the degree
of motivation resulting from the conditions of the box was not
determined by a control group with which the food incentive was
lacking, it cannot be ascertained in this case whether or not the food
stimulus was the sole factor in causing the rats to cross the grill.
Instead of a grill of the type described above, Moss used as a shock-
ing device two brass plates both of which had to be touched simul-
taneously for the animals to experience a shock. A grill presenting
a continuous stimulation surface, such as was used in the present
study has certain obvious advantages over this simpler device.

The current supplied to the grill was controlled by means of a
General Electric potential transformer and voltage changes were
secured by means of the potentiometer, F (fig. 1). The electric
pressure was measured at the beginning of each experiment by
means of a high tension voltmeter, G, in order to obviate any devia-
tions in the current supplied, adjustment being made by means of
the potentiometer. The voltmeter when not in use was thrown out
of the circuit by means of a snap switch, H. The current was regu-
lated by meaus of variable resistance units each of which was con-
trolled by means of a multipoint switch, I. Each resistance unit
contained eight elements each having a resistance of approximately

408

APPENDICES

1,250,000 ohms. The multipoint switches made possible a cumula-
tive variation of the resistance in equal steps from 100,000 to
10,000,000 ohms. The current supplying the grills was measured
by a microammeter, J. The method of wiring made possible, by
means of the double throw knife switch, K, the measurement of the
current in either grill (one for each control box) by one fixed
measuring instrument. The current could be thrown into either or
both of the grills by closing the switches, L, L. This apparatus
makes possible an exact determination of the current supplied to
the grills in terms of voltage and resistance and provides an ade-
quate means of controlling the current and of varying it in syste-
matic amounts. This control is a decided improvement over that of
Moss, since the latter merely considers the current used in terms of
voltage and disregards the resistance factor. Moreover, since the
skin resistance of the animals undoubtedly varies greatly from rat
to rat the amount of stimulation received must also vary, unless
this factor is in some way eliminated. Moss attempted to overcome
this difficulty by placing water in section A so that the feet of the
animals were moistened on approaching the grill. This method is
not entirely satisfactory however, as was indicated above. In the
present study this factor was eliminated to as great an extent as
possible by using a very large resistance within the electric circuit
and thus rendering negligible the resistance of the rats. Since the
resistance here used ranged from 3,850,000 to 10,100,000 ohms, it is
believed that any variability in resistance in the rats is rendered
practically negligible, thus insuring a constant degree of stimula-
tion.

The following current values were used in this study :

MILLIAMPERES

RESISTANCE (OHMS)

VOLTAGE

Shock

1

0.12

10,100,000

1,200

Shock

2

0.19

6,350,000

1,200

Shock

3_ _ _

0.32

3,850,000

1,200

A total of 803 albino rats were used in this investigation.3 The
size of each of the 21 groups studied is indicated in table 1 which

s In view of the large number of rats used in this experiment an autopsy of
the animals to determine the presence or absence of food in the stomach could
not be made.

APPENDICES

409

also shows the arrangement of starvation periods. With the 60-
and 72-hour starvation periods many animals died due to the
length of the starvation period, thus introducing a selective factor
into these groups, which makes them not entirely comparable with
the earlier groups. This factor is, however, not great except in the
case of the 72-hour groups.

The rats used with shock 1 were highly uniform weighing be-
tween 80 and 100 grams and were approximately 8 weeks old.
Those used with shock 3 were slightly more variable in size weigh-
ing between 80 and 110 grams but were of approximately the same
age as those of shock 1. Those used with shock 2 were in general
similar to those of shock 1 with the exception of half of the 12-,
24-, 48-, 60- and 72-hour groups which were similar to those of
shock 3.

TABLE 1

Slwwing number of rats used in each starvation group
A indicates number of rats given preliminary training; B, number used in
experiment proper; C, number that died during starvation period.

SHOCK 1

SHOCK 2

SHOCK 3

GROUP

A

B

C

A

B

C

A

B

C

Control

20

20

20

20

20

0~

2(T

20

0

12-hour

37

37

0

40

40

0

40

40

0

24-hour

40

40

0

39

39

0

40

40

0

36-hour - -

40

40

0

40

40

0

40

40

0

48 -hour

40

40

0

40

40

0

40

40

0

60 -hour _ _

41

35

6

40

40

0

53

40

13

72-hour _

40

21

19

44

21

23

49

37

12

The rats were kept in the laboratory for one week before being
given the preliminary training, in order that they might become
adjusted to the laboratory conditions. During this period they were
fed milk-soaked bread in their cages at noon each day.

The preliminary training consisted of placing the animals in
groups of ten in section A and allowing them to run to the food
in section C or explore freely within the box for 10 minutes. Dur-
ing this preliminary training there was of course no current in the
grill. This procedure was repeated for three consecutive days, in
order that the animals might become accustomed to the conditions
of the control box.

In order to determine whether the food incentive was the sole

410

APPENDICES

factor causing the rats to cross the grill, a control group was used
with each degree of shock. The preliminary training of these
groups was the same as for the other groups except for the fact
that section C contained no food while they were in the box. The
control groups were fed immediately after being removed from the
box.

In the experiment proper each rat was placed individually in
section A with the current turned on in the grill and a dish of
milk-soaked bread in section C. The behavior of each rat was re-
corded during the test period of 10 minutes. If any animal crossed
the grill during the test period it was allowed a nibble of food and
immediately returned to A. If a rat crossed the grill continuously
(i.e., with only momentary pauses upon being returned to section
A) a nibble was allowed after every third cross.

The procedure for the control group was the same as that for
the starvation groups except for the fact that there was no food
stimulus in section C. The experiment proper, occurred in the case
of the control group, 12 hours after the last period of preliminary
training. The control group is consequently equivalent to the 12-
hour group in every condition but the absence of the food incen-
tive.

The following types of behavior were recorded for each animal
during the 10-minute test period: (a) crossings, (b) contacts, (c)
jumps. The term crossing indicates that the animal actually crossed
the grill and entered the incentive compartment. This measure ap-
pears to be the most significant of the three types, inasmuch as it
represents the only reactions which resulted in satisfaction of the
hunger drive.

By the term contact is meant that the animal merely touched the
grill and withdrew upon being shocked, remaining in the entrance
compartment. Partial crossings of the grill, followed by return
to the entry compartment (instead of proceeding into the incen-
tive compartment) were scored as contacts. Data concerning con-
tacts are obviously important not only as indicating an attempt of
the animal to react positively to the incentive stimulus but also
as a measure of the conditioning process resulting in many cases
in a negative reaction to the grill. Contacts are in reality frus-

APPENDICES

411

trated crossings and thus are closely related to the latter type of
behavior. It has, however, seemed advisable to consider the two
kinds of activity separately.

The term jump refers to the behavior of the animal in the en-
trance compartment. A jump was checked only when the animal
cleared the floor with all four feet. This type of activity seems to
indicate an attempt to escape from the entrance compartment.
This reaction was probably in most cases the result of the shock,
since jumping did not occur until after the first experience with
the grill.

The test period of every group in the experiment proper was so
arranged as to fall between 3 and 6 p. m. The last of the three
periods of preliminary training preceded the experiment proper
by the number of hours of starvation requisite for each group.
Consequently the preliminary training of the control, 12-, 36- and
60-hour groups was given as 4:30 a.m. while that of the 24-, 48-
and 72-hour groups was given at 4 :30 p.m.

III. Presentation of Results

The quantitative results of this study, covering both the effect
of different periods of starvation and different degrees of shock on
drive behavior are presented in tabular and graphic form as fol-
lows:

Tables 2, 3 and 4 are distribution tables for shocks 1, 2 and 3
respectively, covering in each case the three types of behavior —
crossings, contacts, and jumps — for each starvation group.

Tables 5, 6 and 7 give the main results of shocks 1, 2 and 3 re-
spectively, for each starvation period. Columns 2 to 5 of these
tables give the measures of central tendency for crossings (me-
dian, average, S.D. and C.V.). Columns 6 to 10 give measures of
the extreme tendencies for crossings. Column 6 indicates the range
of crossings with each starvation group. Column 7 shows the per
cent of rats that were immediately conditioned in each group (i.e.,
of those rats that did not cross the grill after their preliminary
experience with it. The first experience with the grill may be
either a crossing or a contact). This figure indicates the percentage
of rats with which the resistance offered by the grill was at the

412

APPENDICES

TABLE 2

Distribution table covering crossings, contacts and jumps under shock 1

, o

2 n
H

CROSSINGS

CONTACTS

JUMPS

Q

12

24

36

48

60

72

c

12

24

36

48

60

72

Q

12

24

36

48

60

72

0

6

8

4

1

2

0

6

7

7

2

1

15

8

7

8

20

9

8

1

6

5

1

1

3

8

7

15

6

8

9

7

8

4

6

6

4

6

2

6

3

2

5

1

7

7

7

3

9

6

6

2

6

7

3

3

2

1

3

1

2

2

4

1

2

3

2

6

4

10

6

1

3

4

.1

4

1

4

1

3

3

2

1

5

3

6

3

2

1

2

2

4

2

1

1

5

3

2

1

1

2

4

7

4

3

5

6

1

3

1

6

1

1

2

4

3

1

1

3

1

3

3

3

4

7

3

3

4

1

1

1

3

4

3

4

1

8

1

1

2

2

2

2

1

1

1

1

9

1

2

1

1

1

1

1

1

2

10

1

1

1

1

1

11

1

1

2

1

4

1

1

12

1

1

1

3

13

1

2

2

1

14

1

15

1

1

1

16

1

17

2

1

18

1

1

1

2

19

20

1

1

2

1

21

1

1

1

22

1

1

1

23

1

1

1

2

1

24

2

3

2

25

1

1

1

26

27

2

1

28

3

1

29

1

1

30

1

1

31

1

1

32

1

1

33

34

1

1

1

1

1

35

1

36

1

1

2

1

37

3

1

38

1

1

39

1

40

1

4

1

2

1

45

1

3

3

4

1

APPENDICES

413

TABLE 2 — Continued

CROSSINGS

C

12

24

36

48

60

72

3

3

3

1

2

2

1

1

1

1

2

3

1

1

1

1

1

1

1

1

3

1

1

1

« o

50
55
60
65
70
75
80
85
90
95

CONTACTS

12

24

36

48

60

72

12

24

36

48

60

72

The notation C refers to the control group, the numbers 12, 24, 36, 48, 60 and
72 to the series of starvation periods.

The interval of the distribution from 0 to 40 is 1, from 40 to the end of
the scale, 5.

This table should read as follows : Control group : 6 rats crossed 0 times, 6
rats crossed 2 times, 1 rat crossed 3 times, etc.

outset stronger than the hunger drive. Column 8 gives the per-
centage of animals that crossed the grill only 5 times or less. The
number 5 was chosen because the distributions of the groups indi-
cated that the majority of rats with which the resistance soon
proved stronger than the hunger drive, were conditioned by 5
crossings or less. Column 9 gives the precentage in each group that
crossed 40 or more times. This number was arbitrarily chosen be-
cause few of the animals crossed more than 40 times within the test
period. Column 10 shows the percentage of rats that failed to be
conditioned against the grill (i.e., were still crossing the grill at
the end of the 10-minute test period). This measure indicated the
number of animals within each group with which the hunger drive
was still stronger than the resistance offered by the grill after a 10-
minute experience with it. Columns 11 to 14 give the measures of
central tendency for contacts (median, average, S.D. and C.V.).
Columns 15 to 18 give the same measures of central tendency for
jumps.

Tables 8, 9 and 10 for shocks 1, 2 and 3 respectively, give the re-
liability of the difference between each two starvation groups with
the same degree of shock. (Reliability is stated in terms of the
ratio of the difference between two groups to the S.D. of that dif-
ference.)

414

APPENDICES

TABLE 3

Distribution table covering crossings, contacts and jumps under shock 2

i

CROSSINGS

CONTACTS

JUMPS

63

|

C

12

24

36

48

60

72

c

12

24

36

48

60

72

c

12

24

36

48

60

72

0

~~

10

7

8

2

2

3

7

8

1

1

1

5

1

8

23

12

19

15

18

12

1

7

2

9

3

5

5

1

5

5

8

3

11

11

3

3

5

5

10

5

11

2

8

7

6

6

1

8

2

5

8

12

12

5

7

4

3

3

5

4

6

1

1

3

2

4

3

8

1

6

2

5

9

12

7

7

4

1

4

1

3

2

4

2

4

2

1

3

5

3

1

7

3

4

6

6

1

2

1

3

1

2

1

1

5

1

2

1

3

3

2

1

6

2

2

1

2

4

1

1

1

6

1

1

4

1

4

1

3

4

1

2

2

1

7
8

1

2

1
1

1

1

1
1

2
1

1

1

1

3

1

1

3

1

1
1

1

1

9

1

1

1

1

10

1

1

1

1

1

1

1

1

11

1

1

1

1

1

1

1

12

1

1

13

2

2

1

1

14

2

2

1

15

1

2

16

1

1

17

2

18

1

19

1

1

20

1

1

1

1

1

21

1

1

1

1

22

1

1

23

24

25

2

26

1

27

1

2

1

28

1

1

29

1

1

30

31

1

1

32

33

2

1

34

OO

36

1

1

1

37

38

39

1

1

40

1

2

3

2

APPENDICES 415

TABLE 3 — Continued

<

>

CROSSINGS

CONTACTS

JUMPS

INTE

C

12

24

36

48

60

72

C

12

24

36

48

60

72

C

12

24

36

48

60

72

45
50
55
60
65
70
75
80
85
90
00
05

1
1

2
1

1

1

1
1

3
2
1

1

1

1

See legend accompanying table 2.

Table 11 indicates the relative effect of the 3 degrees of shock
upon the various starvation groups.

Table 12 shows the reliability of the difference between any two
shocks with a given starvation group.

Figures 2, 4 and 5 show the cumulative percentile distributions
for crossings for shocks 1, 2 and 3 respectively. These figures make
possible a point to point comparison of any decile in the distribu-
tion of one group with any decile of any other group with which
the same degree of shock was used.

Figure 3 shows graphically the effect of the different starvation
periods upon the central and extreme tendencies of each group
with shock 1.

Figure 6 shows graphically the median number of crossings
made by each starvation group with each degree of shock.

The results of this study will be considered in two sections. The
first section is concerned with the effect of the series of different
starvation periods upon behavior in the obstruction-incentive sit-
uation. The second section will consider the effect of different de-
grees of shock upon behavior when the preceding starvation period
was the same for each degree of shock.

The effect of different periods of starvation

This section is concerned with the effect of a series of different

416

APPENDICES

TABLE 4

Distribution table covering crossings, contacts and jumps under shock 3

<
>

CROSSINGS

CONTACTS

JUMPS

INTER

C

12

24

36

48

60

72

c

12

24

36

48

60

72

C

12

24 •

36

48

60

72

o

o

L

o
o

ft

9

A
o

A
ft

o

o

o

A

o

1
1

A
D

4

ft

o
o

7
4

y

14

1 1
1 1

19
1Z

Q

y

13
lo

1

7

A

4
1

7

4

7

A
ft

o

7

in

1U

Q

y

1 1

XX

A
D

7

9

7

4

Q
o

4

ft

D

A

o

K
O

2

A
t

1 9

o
o

A
t

A
D

Q

y

Q
o

4,

ft

xo

A
Q

in

1U

c
O

C

o

4

ft

9

K
O

4

ft

9

o

O

3

1

0

J.

A
ft

l

p.
D

9
_

Q

o

A
D

in

7

in
xu

o
o

O

o

1

K
0

1

1
1

Q
O

4

1

K

o

o

9

ft

9

X

4

4

ft

A
D

9

Q

o

1
X

9

1
X

9

4

4

Q
O

5

1
1

1
1

1
1

9

ft

i

X

1

X

Q
O

c

o

O

9
-

4

9

Q
O

9

X

6

1

1
1

A
ft

o

1

X

1
X

1

X

9

9

Q
o

Q
o

1

9

4

9

4

9

7

1
1

1

1
X

1

X

1

X

Q
o

1

X-

1
X

1

X

1

9

1

9

9
—

9

2

1
X

8

1
1

9
—

1

X

1

2

1

1

3

2

2

9

i

X

2

2

9

9

£t

10

1

X

1

2

]

\

1
X

11

1
1

9

9

1

1

2

12

1

2

]

2

2

\

4

13

1
1

9
_

1

14

1

1
1

1
X

'9

1

1

15

1

X

1

X

i

X

1

J

1

3

16

1

i

X

17

1

X

1
X

9
—

18

1
X

19

1

1

1

X

1

20

1
X

1

X

1
X

i

X

1

21

22

1

2

23

1
1

i

X

1

24

1

X

1

25

1

X

26

1

1

27

1

X

28

1

X

1

29

1

X

1

30

1

X

31

1

1

1

X

32

1

3

1

1

33

1

34

1

35

1

36

1

37

1

1

2

1

38

39

1

1

40

1

1

1

APPENDICES

417

TABLE 4 — Continued

<
>
es

CROSSINGS

H

Q

12

24

36

4S

GO

72

45

2

1

1

2

50

1

3

1

55

1

60

65

1

70

1

75

2

1

1

80

85

90

1

95

12

24

3G

48

GO

72

C 12 24 36 48 60 72

See legend accompanying table 2.

starvation periods upon the behavior of the white rat in the ob-
struction-incentive situation. The effect of the different starvation
periods will be considered when the obstruction involves (a) a low
degree of electrical stimulation — shock 1; (6) a medium degree
— shock 2 ; and (c) a high degree — shock 3.

The effects of the different periods of starvation will be dis-
cussed in respect to five main types of data. First, in relation to the
number of crossings made by each group, as indicated by the me-
dian,4 or by the percentage of animals within each group that
crossed the grill 40 or more times ; or more fully by a comparison of
the cumulative percentile graphs showing the distribution of each
group in regard to the number of crossings. Second, by the per
cent of animals within each starvation group that were imme-
diately conditioned, i.e., that failed to cross the grill after the first
experience with the shock. This figure probably indicates the num-
ber of animals in which the hunger drive was at the outset weaker
than the resistance offered by the grill. Third, by the percentage
of rats that failed to be conditioned within each group, i.e., were
still crossing the grill at the end of the ten-minute test period. This
figure probably indicates the number of animals in which the hun-
ger; drive was at the end of the test period still stronger than the

4 The median rather than the average will be referred to throughout this
discussion. Owing to the fact that the extreme deviations always occurred at
the same end of the distribution scale, the average does not in this case give
a reliable measure of central tendency.

418

APPENDICES

-buba jo ^uapiyaoQ

M O h O fl) i—i

^ CO O CO N O

00 CO i-M t^* T^H

M IM N O OJ 00 CM

M >0 Tt< N CO rfi CO

93BI9AV

h M N N N N CO
H CO Tji iO M< CM*

tmpaj^

tO O 00 o (O ^
© ci »o CO i-h CO* i-J

-tjixba jo ^uaiotyaoQ

— I CI l-H CO C5 --H

CO (O O N O0 N

N O (N C-l N O
O h CO ^ M O

O CO (MM (N h

83BJ3Ay

SO iO CO (M N 00
rn rfi N CO W <N

UBipajv

O O Cl h O h
CM r-i CO CM CO

panot}
-ipuoo %ou %UQ0 sad

OhOOOOCOOICON

HHIOCOOIHCOH

ajoui jo saint}

0I> 3UI86OJ0 %UB0 J8J

©OCO©COO>t>.»0*0
h CO lO CO CM

SS3| 10 SOlUt)

§ Suissojd ?uao iaj

lOiOiOOiOO^^O
CO iO h rtiiOCMCO

pauoi^ipuoo X(

©COCO*O©CiCM©00
CO CO —i CM ^

CMt^GOOOOOOOOOOO
HNClOOOOOiOCOtO

II 1 1 1 1 '1 1

OOOOOi-HOCMO

-BUBA jo ^uapiyaoQ

COiONHTt^OOO'*
LOCONCOCOIN^OO'O

uot^BiAap pjBpuc^g

iOCOCOTt<COl>CMOOCO
(NCONiOO^^iO^

83BJ8Ay

NINHtONOOHOiO
hNOhOhOOOO

l-H CM T}< CO l-H l-H l-H

utjipajv

(OiOtOOONtO'OO

i-H-^OO^t^cO-^COCM
CM CO CM i-t

(0 CO W W 60

t_ S_ S_ S_ t-i

3 3 3 3 3

o o o o o

Ji JS

° CM co 00
U i-h CM CO

APPENDICES

419

-BtJBA jo inaioijjeoo

M Oi O >0 05 »0
M N <-i ^ N CD iC

t-H i-H <M r-< »-l

uoi^BrAap pjBpuecjg

CO N 05 © N CO h
h t}( CO t}( CO O CO

M CN M ^ CD N CO

t*< CO CO t> 00 lO
M H « H M H M

tmpaj^

N CO «5 h O CN CO
»-h O <N i-H (N i-t O

o

o

o

CO

CO

CO

o

-buba jo ^neioijjaoQ

a>

OS

00

CO

t«-

CM

r-

h-

o

1

uoi^BLvap pjnpu^g

-*

I>

CO

CO

CM

CO

CM

g

o
o

CO

C

CM

t-

«>

i—i

CM

CO

CO

CO

CM

CO

co

C5

o

CO

CO

CO

CM

CM

CO

CO

CM

CO

paaop

-ipUOO }0U J8J

©LOiOOOOOOOlO
CO lH

ojora jo sawi%
Of Suiseojo ;nao ja<j

lO 00 CO

H H N

SB3I JO SOUIt}

g Sujssojo cjaeo iaj

ID»ONIDOOCOCOCO
OOCDCDNCOCD^N

pauot^ipuoo A\
-e^Biparaiu'i %udo jaj

1OOOC0C00005H

00

CO©"HOt^cOO©
WhNcON^iOiO

1 1 1 1 1 1 1 1

'-'OOOOOOO

-bijba jo ^neioijjeoo

I>OOOCOOi>-Hi—iO
NO5CDCD00COH00
i-i i-t hh W

UOI^BIAep pjBpaB^s

^^OJCOOCNCON

83BJ8AV

coco-^TfiO'^ajio

i"H CM »H

impajij

^COCOiOOHOO
NCOWMO^CDO

CO CO CO

+3 o

as

O 3

O O

420

APPENDICES

rfi CM CM «-h CO
^ N N O! t H CO

I-H 1— 1 »— 1 »— 1 T— 1 T— 1

OS 00 (N H rH CO

o in to o h i-i

N lO CO CO CO W lO

N N CO 00 CO lO Ci
iO IM CO rf t)* CO

lO 00 CO CO iO O CM

cm cm «-* co cm co cm'

-buba jo inapgjaoo

h O) N N N © OS
00 CO N CO N O CO

UOl^BIAQp pJBpUB^g

1.14
1.80
1.82
2.40
1.84
3.69

2.18

93BJ9AV

^ O >0 CO rj( (M
H N N CO N CO CO

UBipaj^

N CO 00 »0 CD CM CO
i-H CM <M CO CM CO CO

paaot}
-ipaoo %ou %uqo iaj

"300000000

9J0tn jo sauiii
Of 8ai6S0iD ?U8D J9<j

lOCNMONNCNOO©

CM r-l

669] JO Sdmi)

5 Sntssoao }U90 jgj

0>OiOCOCOOOOHiO
00NCS»Ocp^cO^»O

paaot^ipaoo A\
-a^Bipara'aif ?u90 J9j

lOOOOMMOO^iO
t)<h(MNM(N^hcO

98ubh

OOOOOOOOO

-bijba jo ^aaioyjeoQ

OOJINOihNOOOOQ
HiONMiONiOHCO

U0t^BtA9p pJBpUB^g

OOOOCOCNCDOCOCOiO

i-tO"^OOCOCOCMCOO
^Hi-iCMi-ii-i'-i'-trHf-i

93BJ9Ay

COOOOi-hOCOOOOCM
iOCDO^hOCO^h

N H H H 1-t ^

uBipgjv

COCNO»OCOCOC£)iON
CM CO* 00 *0 tJ< CC » CM t>^ CO*

CO CO GO

£ 3 3 3 3 3

43 O O O O O

C ^3 ^ X -C J

° (N to CO O

U H CN CO t CO

bfl
d

c
a

S

5
o

Q
cj

a

a

s

APPENDICES

421

TABLE 8

Showing the reliability of the differences between the various starvation periods
(average crossings) under shock 1

12

24

36

48

60

72

STARVATION
GROUP

SDD

D
SDD

SDD

D
SDD

SDD

D

SDD

SDD

D

SDD

SDD

D

SDD

SDD

D

SDD

Control

3.87

4.03

3.57

7.70

4.06

9.85

3.15

9.23

2.55

3.33

2.38

3.31

12

5.21

2.27

5.55

4.40

4.93

2.73

4.57

1.55

4.48

1.72

24

5.34

2.34

4.70

0.09

4.31

4.40

4.22

4.64

36

5.07

2.15

4.59

6.49

4.63

6.93

48

3.81

5 41

3.87

5.48

60

3.20

0.19

TABLE 9

Showing the reliability of the differences between the various starvation periods
(average crossings) under shock 2

12

24

36

48

60

72

STARVATION
GROUP

SDD

D
SDD

SDD

D

SDD

SDD

D

8DD

SDD

D

SDD

SDD

D

SDD

SDD

D
SDD

Control

3.59

0.90

3.82

1.83

3.23

1.27

3.80

4.52

2.85

1.43

2.56

1.09

12

4.31

0 72

3.80

0.03

4.30

3.07

3.49

0.03

3.26

2.08

24

4 02

0.75

4.49

2.24

3.72

0.81

3.51

2.82

36

4.00

3.28

3.12

0.00

2.85

2.42

48

3.70

3.54

3.48

5.74

60

2.42

2.85

TABLE 10

Showing the reliability of the differences between the various starvation periods
(average crossings) under shock 3

STARVATION
GROUP

12

24

36

48

60

72

SDD

D
SDD

SDD

D

SDD

SDD

D
SDD

SDD

D

SDD

SDD

D

SDD

SDD

D

SDD

Control

3.14

0.38

4.66

3.09

3.90

2.17

3.72

1.45

3.16

0.76

2.81

1.63

12

4.20

3.14

3.35

2.18

3.13

1.34

2.43

0.49

2.80

1.57

24

4.80

1.22

4.65

1.93

4.21

2.85

4.42

1.98

36

3.89

0.80

3.36

1.81

3.63

0.80

48

3.14

0.96

3.43

0.06

60

2.81

1.13

422

APPENDICES

.._ Control

[Z Hoars

24 HoufS

3& HecrS

. — 4$ Hours
. • . - 6>0 Hours
72 I burs

..J

r

J

r

_j

r

. i

H

Fig. 2. Percentile Distribution of Each Starvation Group Covering
Crossings with Shock 1

Ordinates indicate the number of crossings, abscissas the deciles of each
distribution.

— . 7>.r c<=^v o( cr&^> Not oimd.xtfned (V^ad i>rr|oJ«J_

?trc»t c^Sra-J, enOSs<mf 40*rcrwe e^U^

"Rst <re»> of <Sr<xf> iToai.M.* 4o at moretmrsCJnKl • »clv>4«d,)

Fig. 3

a. Showing median number of crossings made by each starvation group.
Ordinates indicate number of crossings, abscissas periods of starvation.

b. Showing percentage of animals immediately conditioned and percentage
crossing 5 times or less in each starvation group. Ordinates indicate percent-
ages, abscissas periods of starvation.

c. Showing percentage of animals crossing 40 times or more and percentage
that failed to be conditioned in each starvation group. Ordinates indicate per-
centages, abscissas periods of starvation.

424

APPENDICES

J2i

.Control
JZ Hours
. 2A Hours
. 3fo Hours
.48 Hours"
,.60 Hours
_72Hour&

Fig. 4. Percentile Distribution of Each Starvation Group Covering

Crossings with Shock 2
Ordinates indicate the number of crossings, abscissas the deciles of each
distribution.

APPENDICES

425

Conrrox.

1Z Hours

a Hours

3b Hours

.. 48 Hours

_ ^ Hours

72HourS

II

r
i

ii £

if.

s£ — I

| —
I

i

i

I

J, JErd*

Fig. 5. Percentile Distribution of Each Starvation Group with Shock 3

Covering Crossings
Ordinates indicate the number of crossings, abscissas the deciles of each
distribution.

426

APPENDICES

APPENDICES

427

resistance offered by the grill. Fourth by the median number of
contacts and fifth by the median number of jumps made by each
starvation group.

Shock 1. The results of the control group considered by any of
these measures indicate that the tendency to cross the grill with
shock 1, in the absence of the food incentive is negligible, the me-
dian number of crossings in this case being 1.6, while only 1 rat
crossed the grill more than 3 times (table 5, row 1). In general the
tendency to touch the grill and withdraw is greater than that of
crossing it.

The number of crossings for the 12-, 24- and 36-hour groups in-
creases with the length of the starvation period. When this period
is increased beyond 36 hours there is a consistent decrease in the
number of crossings. This holds true of the 72-hour group only if
the dead are included in the results as failing to cross the grill.5
When the dead are excluded, however, the number of crossings
made by the 72-hour group is greater than that for the 60-hour
group. These results are clearly indicated by the medians of each
group (table 5, column 1) and are shown graphically in figure 3,
a, The results of the 24-hour group are in this respect approxi-
mately the same as those of the 48-hour group, while those of the
12-hour group are similar to those of the 60-hour group. The rela-
tive effect of the different starvation periods as shown by the per-
centage of rats crossing 40 or more times (table 5, column 9) is
similar to that shown by the median number of crossings with the
exception that the number of animals crossing very frequently is
greater in the 12 than in the 60-hour group (see Fig. 3, c).

The variability of the different starvation groups (crossings —
table 5, column 5) is in general inversely related to the strength of

6 The results of the groups in which rats died of starvation are stated in
Tables 5, 6 and 7 both excluding and including the dead. Neither of these
measures are entirely satisfactory for a direct comparison with the results of
the other groups, but it is probable that the results in which the dead are
included give a better measure of the tendency of the group than those which
exclude them. This is due to the fact that the dead in all probability corre-
spond to the least active rats in the other groups. Consequently any results
excluding them probably indicate only the tendencies of the upper end of the
distribution scale.

428

APPENDICES

the drive. In the control group in which the hunger drive ap-
pears to be a negligible factor, the variability is extremely large
(C.V. 153). With the series of starvation groups the variability is
in every case smaller than with the control group. In general the
variability tends to diminish as the starvation period is increased
to 36 hours and to increase with the longer periods.

The distribution covering the crossings for each of the starvation
groups is shown in Figure 2 by means of cumulative percentile
graphs. It is apparent from this figure that the differences be-
tween the starvation groups indicated above by the median and
per cent crossing 40 or more times holds true between the cor-
responding deciles of each group. The differences between the
control group and any of the starvation groups are immediately
apparent. The fourth decile of the 12-hour, the second decile of the
24-hour, and the first decile of the 36-hour groups show as frequent
crossings as the ninth decile of the control group. The 12- and 60-
hour groups are simliar throughout the first 6 deciles of their dis-
tributions, but from the sixth through the tenth deciles the 12-
hour group shows more frequent crossings than the 60-hour group.
The distributions of the 24- and 48-hour groups are similar from
the fourth through the ninth deciles, but from the first through the
third decile the 48-hour group crossed considerably more frequently
than the 24-hour group. In the tenth decile the 24 crossed more
frequently than the 48-hour group. From the second through the
ninth decile the 36-hour group crossed more frequently than any
of the other groups. These results again indicate the similarity of
the 12- and 60-hour groups but further emphasize the fact that
the more active animals of the 12-hour group cross more frequently
than those of the 60-hour group. It is also shown that the least
active animals of the 48-hour group cross considerably more fre-
quently than those of the 24-hour group although the results of
these two groups are otherwise very similar.

The same results considered in relation to the percentage of rats
that failed to cross the grill after their first experience with the
shock, i.e., were immediately conditioned (table 5, column 7) indi-
cate that increases in the period of starvation from 12 to 48 hours
result in corresponding decreases in this type of behavior. After a

APPENDICES

429

starvation period of 48 hours all the rats crossed the grill at least
5 times. The results of further increases in the period of starva-
tion, in this respect, cannot be determined because of the large pro-
portion of animals that died in the 60- and 72-hour groups. Within
these two groups all the living animals crossed at least once while
in the 72-hour group no rat crossed less than 2 times. These re-
sults may indicate that the hunger drive of rats starved from 48
hours to the point of physical exhaustion is after the first experi-
ence with the shock still stronger than the resistance offered by the
grill. The percentages of rats crossing the grill 5 times or less
(table 5, column 8) give results similar to the percentages imme-
diately conditioned for starvation periods of 12 to 48 hours (see
Fig. 3, b). With starvation periods of 60 and 72 hours, however,
this measure seems to indicate that in many cases the hunger drive
tends to become weaker than the resistance offered by the grill after
5 experiences with the shock. This appears to be more often the
case with the 60-hour than with the 24-, 36-, or 48-hour groups.

The percentage of rats that failed to be conditioned indicates
that the hunger drive is strongest with the largest proportion of
any group at 36 hours, since with this period of starvation 50 per
cent of the group failed to be conditioned after 10 minutes experi-
ence with the shock, i.e., were still crossing the grill at the end of
the test period. From this fact it might be concluded that the mo-
tivation resulting from 36 hours starvation is equivalent for this
group to the resistance offered by 10 minutes experience with the
grill under the conditions of this experiment. A further increase
in the period of starvation to 48 hours reduced the number of ani-
mals that failed to be conditioned to 30 per cent. Further increases
in the starvation period, however, do not appreciably alter this
percentage. The number of rats failing to be conditioned in any
group starved longer than 36 hours, is in every case, greater than
that of any group starved less than 36 hours (see Fig. 3, c).

From these results it appears that, measured by the number of
crossings, the hunger drive tends to increase in strength as the
period of starvation is increased from 12 to 36 hours and to de-
crease as it is further increased from 36 to 72 hours. It is prob-
able, however, that any determination of these results in terms of

430

APPENDICES

the number of times the grill was crossed measures not only the
hunger drive but also the factor of general activity since physical
exhaustion necessarily results from long starvation periods. As has
been stated above the tendency to cross the grill very frequently is
considerably greater with the 12- than with the 60-hour group,
whereas the tendency to persist in attempts to reach the food stim-
ulus, as shown by the per cent of each group that failed to be con-
ditioned and the per cent immediately conditioned, is considerably
greater with the 60- than with the 12-hour group. As indicated by
the number of times the grill was crossed the hunger drive appears
to diminish from a maximum at 36 hours to the point of physical
exhaustion which occurred with the majority of these rats after 72
hours' starvation. However, a rat that was too weak to cross the
grill still made attempts to do so, thus indicating that the hunger
drive had not been entirely overcome, even- though extreme physi-
cal exhaustion had set in. Consequently the number of crossings
within the test period does not measure the hunger drive alone but
also includes the factor of general bodily activity. Richter (9)
found that the general activity of the white rat increases with the
period of starvation, reaching a maximum at between 2 and 3
days, and then continuously decreases to the point of physical ex-
haustion at 8 days. Richter 's results would appear similar to those
obtained in the present study when measured in terms of the num-
ber of times the grill was crossed by each group. It is probable
therefore that with the longer hunger periods (48 to 72 hours) this
measure indicates the physical capacity of the rats to attain the
incentive stimulus rather than the actual motization resulting
from this stimulus. The latter factor is probably better indicated
by the percentages of animals immediately conditioned and fail-
ing to be conditioned, since these measures are based upon per-
sistence in reaching the food stimulus even when the physical
capacity to do so is relatively slight.

Considering these results together it appears that the hunger
drive reaches a maximum strength at 36 hours. After 48 hours this
drive seems to be slightly diminished, but further increases in the
length of the starvation period to 60 and 72 hours do not seem to
result in any further weakening of the hunger drive, although the

APPENDICES

431

capacity to reach the stimulus is greatly reduced owing to the on-
set of physical exhaustion.

The reliability of the differences in crossings between any two
groups in the shock 1 series will be found in table 8. It is apparent
that the differences between any starvation group and the control
group are completely reliable. The differences between the 12- and
36-hour groups and between the 36- and 60- or 72-hour groups are
also extremely reliable. The differences between the 12- and 24-,
or between the 24- and 36-hour groups are less reliable as are also
the differences between the 36- and 48-, or that between the 48-
and 60-hour groups.

The tendency to touch the grill without crossing as indicated by
the median number of contacts made by each group (table 5, col-
umn 11) does not appear to be related directly to the length of the
starvation period. With the 12- and 60-hour groups, in which
the contacts tended to deter the rats from further crossings, the
tendency to touch the grill is great. It. was also large with the 36-
hour group, but in this case the contacts generally immediately
preceded the crossings and were not as effective in conditioning as
in the other groups. The tendency to jump was in general greater
with starvation periods of 36 hours or less than with the longer
periods.

Shock 2. The results of the shock 2 series are difficult to inter-
pret because of the lack of consistent tendencies. These results seem
to indicate that the shock used in this series was so great as to
arouse drive factors other than those of hunger. Some animals ap-
peared to be reacting against the electric grill in section B rather
than to the food stimulus in section C. After rushing across the
grill into section C they did not eat but sat in a corner with sides
palpitating as if in fear, or attempted to jump from the box. Ap-
parently they did not discriminate between sections A and B for
upon being replaced in A after crossing they would immediately
dash across the grill.

The type of reaction thus described was evident in the control
group, as shown by the fact that 15 per cent of the animals crossed
between 10 and 36 times, in spite of the fact that section C con-
tained no food. The median number of crossings (2.4), however,

432

APPENDICES

indicates that the majority of these animals were easily conditioned.

The differences between the control group and any of the starva-
tion groups in this series, with the exception of the 48-hour group,
are too small to be reliable. The median number of crossings of the
control, 12-, 24-, 36-, and 60-hour groups all fall between 2.4 and
4.1. The median number of crossings is 6.0 for the 72-hour group
excluding the dead and 0.0 when they are included. The percentage
of animals crossing the grill 40 or more times is not large with
shock 2; the largest number (25 per cent) occurred in the 48-hour
group. The tendency to cross the grill continuously is slightly
greater with the 24- and 36-hour than with the 60- and 72-hour
groups. The greatest differentiation between the groups is shown
by the percentage of animals that were immediately conditioned
(table 6, column 7). The results of the control, 12-, and 24-hour
groups are quite similar in this respect ranging from 30 to 40 per
cent. There is a decrease, however, to 28 per cent with the 36-hour
group and to 18 per cent with both the 48- and 60-hour groups and
to 19 per cent with the 72-hour group. The number of animals that
failed to be conditioned (table 6, column 10) never rises above 15
per cent in any but the 48-hour group in which 38 per cent were
still crossing the grill at the end of the test period.

The variability (crossings) of the 12-, 24-, and 36-hour groups
(table 6, column 5) is very great ranging from 160 to 198. With
the 48-hour group it falls to 89 and with the 60-hour group to 131.

In general it may be said that the individual differences between
the animals of each group are so great as to obliterate any group
differences that might result from the various starvation periods.
It was observed that each group included both animals that regu-
larly ate after crossing and animals that failed to do so. In the lat-
ter case the escape reaction was evident. Upon being returned to
section A these rats either withdrew as far as possible from the
grill making no further attempt to cross, or immediately recrossed.
In the latter case the crossings occurred in rapid succession and
then ceased altogether. In the former case there were frequent
pauses between crossings and the animals gave no indications of
fear.

The cumulative percentile graph (Fig. 4) for shock 2 indicates

APPENDICES

433

clearly the difference between the 48-hour and the other groups of
this series. The tendency to cross the grill in the 48-hour group is
evident in the fourth decile and very marked from the fifth through
the tenth decile. The majority of the animals in this group ate
after crossing the grill and showed little evidence of fear. They
seemed to be reacting directly to the food stimulus rather than at-
tempting to escape. Pauses within section A and contacts followed
by withdrawal without crossing were frequent in this group. This
type of reaction, shown by the majority of the 48-hour group, oc-
curred in only a few cases in the other groups of this series. The
tendency to cross the grill is very slight with any other group. In
fact from the first through the sixth decile the number of crossings
made by any starvation group other than the 48-hour differs only
slightly from that of the control group.

In the ninth and tenth deciles the tendency toward extreme reac-
tions is considerable and greater in every case with the starvation
groups than with the control group. This fact, however, does not
warrant the conclusion that these extreme reactions were due to the
hunger drive alone, since in many cases the animals failed to eat
after each crossing, although given ample opportunity to do so.

It appears from these results that the 48-hour group is the only
one of the shock 2 series in which the strength of the hunger drive
is at all comparable with the resistance offered by the grill. Through-
out the other groups this shock appears to disturb seriously the
hunger drive of the majority of the animals. This disturbance
seems to be greater with groups that have been starved less than
36 hours than with those starved more than 36 hours, inasmuch
as the number of animals immediately conditiond is considerably
less with the longer starvation periods. This fact considered in
conjunction with the results of the 48-hour group above discussed
suggests that the degree of shock sufficient to overcome the hunger
drive (as indicated by a failure to react to the food stimulus when
in the presence of it) is not identical for all periods of starvation
but is dependent upon the length of the starvation period preceding
the shock.

The tendency to touch the grill without crossing in the shock 2
series appears to bear some relation to the length of the starvation

434

APPENDICES

period although the differences between the groups are not large.
The number of contacts made by each starvation group is in every
case twice as great as the number made by the control group. The
median number of contacts (table 3, column 11) increases from 2.9
with the 12-hour group to 3.4 with the 48-hour group and falls to
2.6 with the 60-hour group. Jumping behavior seems to bear no
consistent relation to length of starvation period under the condi-
tions of shock 2. The number of jumps (table 3, column 15) is
greatest with the 24-hour group and slightly smaller with 48 -hour
group, but the other groups do not differ greatly from the control
group in this respect.

Shock 3. The tendency to react against the grill rather than to
the food stimulant is noticeable to a much greater extent with
shock 3 than with shock 2. The results of the control group (cross-
ings (of this series are extremely variable (C.V. 210), and seem to
indicate that a strong degree of shock results with a few animals in
a decided tendency to cross the grill. The only rat that failed to be
conditioned by shock 3 occurred in the control group, in the absence
of any food stimulus. This rat showed evidence of fear and seemed
to be fleeing from section A.

The tendency to cross the grill with shock 3 is relatively slight
except in a few extreme cases. Every rat in each starvation group
was conditioned against the grill by shock 3 within the test period
(table 7, column 10). The greatest tendency to cross the grill fre-
quently, as shown by the percentage of animals crossing 40 or more
times (table 7, column 9) occurred in the 12-and 24-hour groups.
With the other groups this tendency does not differ noticeably from
that of the control group. The median number of crossings (table
7, column 2) increases from 2.3 with the control group to 8.0 with
the 24-hour group and then falls to 4.3 as the period of starvation
is increased to 48 hours. The effect of a 60- and 72-hour starvation
period upon the shock 3 groups cannot be determined owing to the
fact that so many animals in these groups died. If the dead are
excluded from the results these groups appear to cross the grill
more frequently than the 48-hour group, whereas if they are
included the medians are smaller than that of this group. The per-
centage of rats immediately conditioned and the percentage cross-

APPENDICES

435

ing the grill 5 times or less (table 7, columns 7 and 8) show the same
trend as the median number of crossings.

The percentile graph for shock 3 (fig. 5) indicates that the differ-
ence between the various starvation groups from the first through
the fifth decile is negligible. Moreover there is no apparent differ-
ence between the starvation and control groups within this range.
From the sixth through the tenth decile, however, the 24-hour group
crosses more frequently than any of the other groups, while from
the seventh through the tenth decile the 36-hour rats cross more
frequently than those of any group with the exception of the
24-hour. Since the differences between the control and starvation
groups (excepting the 24-hour group) are very small and unreli-
able, and since animals usually failed to eat after crossing, shock 3
cannot be said to offer a measure of the hunger drive alone. That
some factor other than the incentive stimulus operates with shock 3 is
shown by the fact that animals crossed frequently in the absence
of food and after refusing to eat. This additional factor appears
to be the tendency to escape, which is stimulated by the shock itself
and results from an association of the grill with section A. The fact
that the number of crossings increases with an increase in the star-
vation period from 12 to 24 hours and decreases with further
increases from 24 to 48 hours indicates that there is some relation
between the tendency to cross the grill and the length of the starva-
tion period. This relation is probably due to differences in the
organic state of the animals resulting from the different starvation
periods. Richter (supra) found that general activity bore a direct
relation to differences in organic state resulting from different
starvation periods. Consequently it is probable that shock 3 indi-
cates more clearly the factor of general activity than the motivation
resulting directly from the incentive stimulus.

The median number of contacts in this series increases from 2.3
with the 12-hour group to 3.5 with the 36-hour group and then
decreases to 2.6 with the 48-hour group. There appears to be an
increase in the number of contacts with the 60- and 72-hour groups
but owing to the number of deaths in these groups, the results are
not comparable with the others of this series. The jumping activity
(table 7, column 15) shows no relation to the length of the starva-

TABLE 11

Showing the effect of different degrees of shock upon the same starvation

groups

CROSSINGS

c

•- >>
o

II

II

u >>

§ s

•8
SI
« »-

^ o

C as

4) 4>

° a

Cm

CONTACTS
(MEDIAN)

Control group

Shock 1

1.6

153

0-12

60

uO

2.0

0 fifi

Shock 2

2.4

177

1-36

35

85

0

0

1.6

1.7

Shock 3

2.3

210

0-54

45

80

5

5

1.7

2.5

12-hour

group

Shock 1

4.5

135

0-77

33

55

19

11

4.6

2.4

Shock 2

3.3

198

0-108

30

65

5

5

2.9

0.87

Shock 3

3.2

159

0-44

18

75

2.5

0

2.3

2.8

24-hour

group

Shock 1

28.6

77

0-98

13

25

33

18

1.9

5.0

Shock 2

2.6

160

0-74

40

67

15

5

2.9

2.5

Shock 3 , . .

8.0

122

0-93

20

25

23

0

2.8

1.8

36-hour

group

Shock 1 , . ,

37.0

61

0-89

5

10

50

50

3.7

3.8

Shock 2

3.5

166

0-60

28

75

10

8

3.3

1.1

Shock 3.

5.5

129

0-79

23

53

10

0

3.5

3.6

48-hour group

Shock 1 ,

27.0

64

5-

-80

0

5

33

30

2.6

1,0

Shock 2

20.0

89

0-

-77

18

3S

25

38

3.4

2.0

Shock 3.,

4.3

151

0-

-70

33

68

7.5

0

2.6

2.5

60-hour group

Shock 1

6.7

120

1-58

9

46

9

23

3.1

3.6

4.6

140

0-58

22

54

7

19.5

2.6

1.8

Shock 2

4.1

131

0-46

18

68

2.5

0

2.6

1.2

Shock 3

6.3

127

0-69

20

48

2.5

0

3.2

3.0

*

2.6

158

0-69

40

60

2

0

2.4

1.9

72-hour group

Shock 1 . .

16.5

88

2-68

0

24

5

33

2.4

1.4

*

2.0

154

0-68

48

60

2.5

17.5

0.0

0.0

Shock 2

6.0

111

0-50

19

43

4.8

10

3.6

0.89

0.0

389

0-50

61

73

2.3

5

0.0

0.0

Shock 3

7.5

110

0-70

14

41

8.1

0

3.6

2.2

3.7

138

0-70

35

55

6.1

0

2.6

0.0

* Results in this row include the dead.

APPENDICES

437

tion periods. It is rather large in every instance but that of the
24-hour group.

The effect of different degrees of electrical stimulation

This section in concerned with the effect of different degrees of
electrical stimulation upon drive behavior resulting from a given
starvation period. The relative effect of the three degrees of shock
upon the control group and upon each of the six starvation periods
will be considered, as measured by crossings, contacts, and jumps.

Control group. In the absence of a food stimulus in section C,
the central tendencies of the groups indicated by the medians
(table 11, column 1) are not materially affected by increasing the
shock. The extreme tendencies of the groups are however, con-
siderably affected. The percentage of animals immediately condi-
tioned by shock 1 is greater than that by either shocks 2 or 3.
Frequent crossings of the grill increase directly with an increase in
shock as shown by the fact that only one rat crossed the grill more
than 3 times, and that only 12 times, with shock 1, whereas 3 rats
crossed between 10 and 36 times with shock 2, and 3 between 8 and
54 times with shock 3. Moreover the only animal that failed to be
conditioned in any control group occurred with shock 3. As indi-
cated by the coefficient of variability an increase in shock directly
increases the variability of the groups (table 11, column 2).

The number of contacts made by the control groups was greater
with shock 1 than with either shocks 2 or 3 (table 11, column 8).
This appeared to be due to the fact that the rats moved much more
rapidly and less cautiously with shocks 2 and 3 and consequently
upon touching the grill did not stop but proceeded across the grill,
whereas with shock 1 the animals approached the grill rather slowly
and generally withdrew after experiencing the shock. The number
of jumps made by the control groups appears to be directly
increased by an increase in shock, the median number of jumps
being 0.66 with shock 1, 1.7 with shock 2, and 2.5 with shock 3.

In brief, in the absence of the food stimulus an increase in shock
while affecting the central tendencies (crossings) of the groups
only slightly produces a marked effect upon the extreme tendencies
of the groups. The lowest degree of shock appears to be most effec-

438

APPENDICES

tive in immediately deterring the animals from crossing the grill,
whereas a heavy shock may result in repeated crossings. Inasmuch
as the degree of shock constitutes the only variable condition in the
control groups, it must be assumed that the increased tendency to
cross the grill is due to the shock itself. Observations of the behavior
of the animals indicated that an increase in the degree of shock
used resulted in an increase in the reactions indicating fear. Such
reactions were not evident with shock 1 but occurred frequently
with shock 2. In such cases the rats after crossing, ran wildly about
section C or tried to jump from the box. Upon being replaced in
section A they would dash immediately across the grill or bury
their heads in a corner of the box, in which position they would
remain, breathing very rapidly until finally removed. With shock
3 the number of animals showing these reactions was greater and
the conditions described more exaggerated. Jumping also occurred
more frequently with an increase in shock. From these facts it
appears that in the absence of the incentive stimulus shocks 2 and 3
were both strong enough to frighten the animals. The increase in

TABLE 12

Showing the reliability of the difference between the various shocks (average
crossings) for a given starvation group

STARVATION
GROUP

SHOCKS

1 AND 2

SHOCKS

2 AND 3

SHOCKS

1 AND 3

SDD

D

SDr

SDD

D

SDD

SDD

D

SDD

Control

2.17

1.70

3.37

0.09

2.70

1.48

12

4.81

1.64

3.37

0.74

4.19

2.48

24

4.75

3.51

4.99

1.69

5.21

1.75

36

4.70

6.85

3.78

1.24

4.94

5.56

48

4.43

1.85

4.11

2.80

4.06

4.85

60

3.14

0.02

2.59

0.54

3.03

0.69

72

2.73

2.56

2.65

3.28

3.20

0.53

the number of crossings made by some rats with an increase in
shock seemed to be due to the fact that fear aroused random move-
ments resulting in a failure to discriminate between sections A and
B so that A was reacted to in the same way as B.

Twelve-hoar group. With the introduction of the incentive
stimulus in section C the relative effect of the three degrees of shock
upon the behavior of the rats is changed. The differences between

APPENDICES

439

the central tendencies of the groups are, as with the control group,
small and unreliable (table 12). An increase in shock in this case,
however, appears to decrease the number of times the rats will
cross the grill, the medians being 4.5 with shock 1, 3.3 with shock 2,
and 3.2 with shock 3.

The results of the three degrees of shock in regard to immediate
conditioning indicate a somewhat different effect. Thirty-three
per cent were immediately conditioned by shock 1, 30 per cent by
shock 2, and 18 per cent by shock 3. Animals reacting to shock 3
are thus more likely to cross the grill at least twice than when the
lighter shocks were used. On the other hand shock 3 is the most
effective of the three shocks in finally conditioning the animals
since no rats failed to be conditioned in this case, whereas 2.5 per
cent remained unconditioned with shock 2 and 11 per cent with
shock 1. Moreover only 25 per cent of the group crossed more than
5 times with shock 3, while 35 per cent did so with shock 2 and 45
per cent with shock 1. The number of rats crossing 40 or more
times was also directly decreased by an increase in shock.

The number of contacts in the 12-hour group decreased with an
increase in shock (table 11, column 8). The tendency to jump does
not vary consistently with the degree of shock as will be seen from
table 11.

While the control and 12-hour groups had been without food
exactly the same length of time, the results are quite opposed in
respect to the relative effect of the three degrees of shock. It will
be recalled that in the control group the animals reacting to shock
1 showed the least tendency to cross the grill, while in the 12-hour
group they show the greatest tendency to do so. Moreover whereas
the greatest tendency to cross in the control group occurred with
shock 3, in the 12-hour group this shock is most effective in deterring
the rats from crossing.

Twenty-four-hour group. The difference between the results
(crossings) with shocks 1 and 2 after 24-hour period of starvation
is large and reliable (table 12). The number of crossings made with
shock 1 is considerable as shown by a median of 28.6 and the fact
that 75 per cent of the group crossed more than 5 times. The ten-
dency to cross with shock 2 is, on the other hand, very slight as

440

APPENDICES

shown by a median of 2.6 and the fact that only 23 per cent of this
group crossed more than 5 times. Only 13 per cent of the animals
used with shock 1 were conditioned immediately, while 40 per cent
of the group with shock 2 were immediately conditioned.

The number of crossings made with shock 3 is in this case greater
than with shock 2, but considerably less than with shock 1, the
median number of crosses being 8.0. The tendency to cross 40 or
more times with shock 3 was not quite as marked as with shock 1 as
shown by the fact that 23 per cent of the group with shock 3 crossed
40 or more times whereas 33 per cent did so with shock 1. The
number of rats immediately conditioned by shock 3 is only half as
great as that by shock 2.

The results of the 24-hour group are least variable with shock 1
and most variable with shock 2 (table 11, column 2).

The number of contacts made by the 24-hour groups was smallest
with shock 1. The number of jumps in this case decreased directly
with an increase in shock.

The behavior of the rats in response to the food after crossing to
section C varied considerably with the different shocks. With shock
1, the rats ran immedately to the food and started eating with all
possible speed. With shock 2 some rats reacted in like manner
showing practically no evidence of fear, while others were obviously
frightened and paid no attention to the food, but tried to jump out
of section C. With shock 3 hardly any of the animals noticed the
food. They either cowered in a corner of section C or attempted to
jump from the box. The speed in crossing the grill when replaced
in section A also differed with the different shocks. With shock 1
hesitation for several seconds at the edge of the grill followed by a
crossing, frequently occurred, whereas with shock 3, if they crossed
at all, the tendency was to dash immediately across the grill after
being placed in A. Very little hesitation at the edge of the grill
occurred with shock 2 although the temporal interval between
crossings varied considerably from rat to rat.

Thirty -six-hour group. The relative effect of the three degrees of
shock upon crossings with the 36-hour group is in general similar to
that noted with the 24-hour group. The largest number of crossings
was made with shock 1 (median 37.0) and the smallest number with

APPENDICES

441

shock 2 (median 3.5). The difference between the results with
shocks 1 and 2 is in this case extremely reliable (table 12). The
number of crossings made with shock 3 (median 5.5) is slightly
greater than, that with shock 2.

The number of animals that failed to be conditioned is largest
with shock 1 (50 per cent), while none failed to be conditioned by
shock 3. The number of rats immediately conditioned was again
smallest with shock 1 (5 per cent) and greatest with shock 2 (28
per cent).

The number of contacts made with shock 1 was rather large
(median 3.7). In this case, however, a contact was generally a
preliminary reaction immediately preceding a crossing. With
shocks 2 and 3 the contacts did not bear as close a relation to the
crossings but were only slightly less frequent than with shock 1.
Jumping activity was considerable with shocks 1 and 3 but was
relatively slight with shock 2.

Forty -eight -hour group. With the 48-hour group the tendency to
cross the grill is directly related to the degree of shock. The group
with shock 1 crossed most frequently (median 27.0) and that with
shock 3 least frequently (median 4.3). The number of crossings
with shock 2 was in this case only slightly less (median 20.0) than
with shock 1. With shock 1, however, only 5 per cent crossed 5
times or less whereas 38 per cent did so with shock 2 and 68 per cent
with shock 3. The number of rats failing to be conditioned was
slightly greater with shock 2 than with shock 1, but the number of
crossings 40 or more times was slightly less than with shock 1.

The number of contacts was approximately the same with shocks
1 and 3 (medians 2.6) but slightly greater with shock 2 (median
3.4). The jumping activity increased directly with an increase in
shock.

The variability (crossings) of the 48-hour group is least with
shock 1 and greatest with shock 3 (table 11).

Sixty-hour group. Due to the fact that some of the 60-hour rats,
with shocks 1 and 3, died, the results for the three degrees of shock
are not entirely comparable. The general effect of an increased
shock appears, however, to be very slight with the 60-hour group.
Considering these results in terms of the medians the differences

442

APPENDICES

between the groups are negligible and have no reliability. The varia-
bility of the three groups is approximately the same (table 11,
column 2).

If the results are considered in respect to the percentage of ani-
mals that were immediately conditioned, there appears to be a
slightly greater tendency for all the rats to cross with shock 1 than
with shocks 2 or 3. More rats crossed 5 times or less with shock 2
than with shocks 1 and 3. The chief difference between the 3 degrees
of shock occurs at the upper end of the scale as 9 per cent of the
group with shock 1 crossed 40 or more times, whereas only 2.5 per
cent did so with shocks 2 and 3. 23 per cent of the group with
shock 1 failed to be conditioned while none failed to be with shocks
2 and 3.

The number of contacts is relatively large with shocks 1 and 3
(medians 3.1 and 3.2, respectively) but is somewhat less with shock
2 (median 2.6). The tendency to jump is also considerably greater
with shocks 1 and 3, (medians 3.6 and 3.0) than with shock (me-
dian 1.2).

It appears from these results that an increase in shock with rats
that have been starved 60 hours was only effective in respect to the
animals that crossed the grill most frequently, since an increase in
shock diminished the number of extreme cases but did not noticeably
affect the central tendency. In general, the activity shown by touch-
ing the grill and jumping was greater with shocks 1 and 3 than with
shock 2.

Seventy-two-hour group. With these groups the factor of selec-
tion is again operative so that the results are not entirely compar-
able. In general the tendency to cross the grill is greatest with
shock 1 since no animals were immediately conditioned in this case
whereas 19 per cent were by shock 2 and 14 per cent by shock 3. The
percentage of rats crossing 5 times or less is also smaller with shock
1 than with shocks 2 or 3. The number crossing 40 or more times is
approximately the same with the three degrees of shock, but the
number failing to be conditioned decreased directly with an increase
of shock. As shown by the median the number of crossings was
greatest with shock 1 and considerably less with shocks 2 and 3
(table 11).

APPENDICES

443

In general the number of contacts is greater with shocks 2 and 3
than with shock 1. The number of jumps in this case seems to bear
no relation to the degree of shock.

Summary. The results of the three degrees of shock upon the
same starvation periods show that, in the absence of the incentive
stimulus, the tendency to cross the grill is relatively slight with any
degree of shock. In this case, however, the general tendency of an
increased shock, as has been stated in the discussion of the control
group, is to increase the tendency to cross the grill. This fact indi-
cates that a strong shock in itself produces a certain degree of
motivation.

These results show that when the food stimulus is introduced
into the otherwise identical situation the relative effect of the three
degrees of shock was dependent upon the length of the starvation
period that had preceded the shock. The relative effect of the three
degrees of shock on each starvation group can be clearly observed in
figure 6. With either a short (12 hours) or long (60 and 72 hours)
starvation period the tendency to cross the grill is not markedly
different with the three degrees of shock, although there is a slight
tendency to cross more frequently with shock 1 than with shocks 2
or 3, except in the case of the 72-hour groups where the results of
the three degrees of shock are not comparable owing to the deaths
occurring in these groups. With a median period of starvation
(24 or 36 hours), however, the number of crossings made with
shock 1 is very great, but is relatively slight with shocks 2 and 3.
With these shocks the tendency to cross the grill in response to the
food stimulus does not appear to differ radically from that observ-
able when the food stimulus is absent. This fact indicates that
shocks 2 and 3 do not give adequate measures of the relative effect
of different periods of starvation. Shock 2 appears to divide the
animals into two groups, those motivated largely by the food
stimulus and those motivated chiefly by fear. With shock 3 the fear
motivation appeared to become dominant in practically all cases.
With shock 1, however, all the animals seemed to be motivated by
the food stimulus. Consequently the data of shock 1 alone, can be
considered as affordig measures of the hunger drive.

444

APPENDICES

IV. Summary and Conclusions

1. Of the three degrees of electrical stimulation used in compart-
ment B as an obstruction against which to measure the hunger
drive, only the lowest proved to be suitable for use with this method.
Both of the higher shocks were severe enough to disturb the normal
food response. Shock 2, in general, had the effect of separating into
two classes the animals of each starvation group: in one case the
animals appeared to be motivated mainly by hunger and in the
other mainly by fear. Shock 3 was so severe that it seemed to intro-
duce an additional source of motivation which disturbed and
obscured the normal hunger drive.

2. The lowest degree of shock furnished a degree of stimulation
adequate for the measurement of the hunger drive by the Obstruc-
tion Method, in that it was high enough to bring out differentiation
in drive behavior resulting from different periods of starvation, and
low enough not to arouse other conflicting tendencies.

Inasmuch as shock 1 alone proved suitable with this method for
the measurement of the hunger drive the following statements are
based solely on data obtained with this shock.

3. As measured by the number of crossings within the test
period, the hunger drive appears to increase from 12 to 36 hours
and to decrease with further increases in the length of the starva-
tion period.

4. The hunger drive as indicated by the percentage of animals
immediately conditioned and by that failing to be conditioned,
apparently reaches its maximum at 36 to 48 hours and remains
fairly constant from this point to 72 hours.

5. As measured by the number of crossings the results of the
24- and 48-hour drives appear to agree, while the 12- and 60-hour
drives have highly similar values. As indicated by the per cent that
failed to be conditioned, however, the hunger drive, is stronger
with a 48- and 60-hour starvation period than with 12- and 24-hour
periods. The difference between the results of these two measures
is probably in large part due to the factor of physical exhaustion
occurring with the longer starvation periods.

6. The number of contacts and jumps do not appear to bear any
consistent relation to the length of the starvation period.

APPENDICES

445

Bibliography

(1) Carlson, A. J. 1919. Control of Hunger in Health and Disease.
Chicago Univ. Press.

(2) Cole, L. W. 1911. The relation of strength of stimulus to rate of
learning in the chick. Jour. Animal Behav., i.

(3) Dodson, J. D. 1917. Relative values of reward and punishment in
habit formation. Psychobiology, i.

(4) Donaldson, H. H. 1924. The Rat. Philadelphia.

(5) Hoge, M. A., and R. J. Stocking. 1912. A note on the relative
value of punishment and reward as motives. Jour. Animal Behav., ii.

(6) Jenkins, Warner and Warden. 1926. Standard Apparatus for the
Study of Animal Motivation. Jour. Comp. Psych., 6 : 361-82.

(7) Moss, F. A. 1924. A study of animal drives. Jour. Exp. Psych., vii.

(8) Nicholls, E. E. 1922. A study of the spontaneous activity of the
guinea pig. Jour. Comp. Psych., ii.

(9) Richter, Curt P. 1922. A behavioristic study of the rat. Comp.
Psych. Monographs, i, No. 2.

(10) Simmons, Rietta. 1924. The relative effectiveness of certain incen-
tives in animal learning. Comp. Psych. Monographs, ii, No. 7.

(11) Slonaker, J. R. 1907. The normal activity of the white rat at dif-
ferent ages. Jour. Comp. Neur. and Psych., xvii, No. 4.

(12) Szymanski, J. S. 1918. Abhandlungen zum Aufbau der Lehre von
den Handlungen der Tiere. Pfliiger's Archiv. f. d. ges. Physiol., clxx.

(13) Szymanski, J. S. 1918-1919. Die Abhangigkeit der Lerngeschwindig-
keit von der Antriebsstarke. Pfliiger's Arch. f. d. ges. Physiol., clxxii.

(14) Szymanski, J. S. 1918. Versuche iiber die Wirkung der Faktorett,
die als Antriebs zum Erlernen einer Handlung dienen konnen. Pflii-
ger,s Arch. f. d. ges. Physiol., clxxi.

(15) Tsai, Chiao. 1925. The relation of the sex and hunger motives in
the albino rat. Jour. Comp. Psych., No. 5.

(16) Yerkes, R. M., and J. D. Dodson. 1908. Relation of strength of
stimulus to rapidity of habit formations. Jour. Comp. Neur. and
Psych., xviii.

2. THE EFFECT OF SHORT INTERVALS OF DELAY
IN FEEDING UPON SPEED OF MAZE
LEARNING 1

C. J. Warden and E. L. Hamilton

In most experimental situations in which animals are employed
it is necessary to arouse the test animal to the desired state of
general activity by the use of an incentive stimulus of some sort.
The incentive stimulus, if it is to be adequate, must be an object
or situation which normally evokes in the organism positive
responses of the locomotor type with respect to the stimulus in
question. For obvious reasons food has been used most extensively
in thus bringing about a more or less general stimulation of the
organism, although other kinds of stimuli, such as sex objects for
example, have been employed on occasion. In any case the same
general principle applies. In order to reach the incentive stimulus
the animal performs, or attempts to perform such feat as the
experimental situation may require.

As a rule the animal is allowed immediate access to the incentive
stimulus after performance so that the latter response is continuous
with the response to the incentive stimulus. And it is, of course,
quite proper that such procedure should be followed in the ordinary
laboratory situation, since rewarding the animal immediately after
performance most probably favors its general emotional adjustment
to the artificial laboratory conditions. One interesting aspect of the
general problem of motivation, however, is that concerning the
effect upon behavior of separating the test response from the re-
sponse to the incentive stimulus by intervals of time of various
length. What will be the effect of thus delaying the latter response
upon the learning of the former? Does the fixation of a pattern
response, such as a maze habit, proceed more effectively and
speedily when the reward immediately follows the performance?

i Published by C. J. Warden and E. L. Haas (now Mrs. E. L. Hamilton) in
The Journal of Comparative Psychology, 1927, 7: 107-16 and reprinted from
that periodical.

446

APPENDICES

447

What, in fact, are the results of a progressive series of intervals of
delay upon the rate of learning 1 The present study is a preliminary
report of a systematic investigation of this general problem.

The present experiment was suggested by the work of Watson (1)
who studied the effect of a 30-second interval of delay in feeding
upon the rate of learning the sawdust type of problem box in the
white rat. He used 6 animals in both test and control group and
employed food as the incentive stimulus. The experimental condi-
tions were kept constant and identical for both groups except for
the fact that the animals of the test group were prevented from
obtaining food after each daily performance by covering the food
dish with a perforated lid which could be conveniently raised by
the experimenter at the close of the 30-second interval of delay.
In both groups, moreover, the animals were allowed to feed for
5 seconds from the dish and then transferred to their cages where
the daily ration was provided. Watson found that the animals were
very active during the period of delay. The time curves for the two
groups are closely similar, leading Watson to conclude that the
delay of 30 seconds did not alter the speed of forming the habit.

The work of Watson has been extended in two directions in the
present study. In the first place a different type of performance
has been selected, a simple maze replacing the problem box. This
made necessary a different kind of apparatus for restraining the
animal from securing the food during the period of delay, a detailed
description of which will be given later. In the second place much
longer intervals of delay were employed, i.e., a 1-minute and a
5-minute interval.. A change in procedure was also introduced.
Instead of feeding the animals for only 5 seconds in the apparatus,
and then giving them their main ration in their living cage, we
allowed our animals to feed for 5 minutes in the apparatus after
daily performance before removing them to a feeding cage close by,
where they were allowed to feed for an additional 5-minutes, and
then removed to their living cage. Any possible influence of the
incentive stimulus, or of the response to it, upon the previous per-
formance of running through the maze was thus given the best
possible conditions in which to operate. In most cases active feed-
ing had ceased before the animals were removed from the food box

448

APPENDICES

The design of the maze employed is shown in figure 1. It proved
to be a relatively easy pattern in spite of the fact that it contains 8
culs-de-sac. It was constructed after the usual manner of wood,
with 4-inch runways, and was provided with a sliding door in front
of the food box which prevented the animals from re-entering the
maze after having reached the food box. The food dish consisted of
the metal bell from an alarm clock, approximately 6 cm. in diam-
eter, screwed down to the floor at the center of the food box.

Fig. 1. Maze Pattern

The device utilized for restraining the animal from obtaining
the food during the period of delay is illustrated in figure 2. The
wooden screen was formed by extending the sides of the food box
upward for a distance of 18 inches. This screen served the purpose
of isolating the animal during the period of delay from possible
distracting stimuli from without, to a large extent. It also elim-
inated the experimenter from the visual field of the animal quite
effectually while permitting easy observation of the animal. The
screen was provided with a door, C, for the removal of the animal
from the food box after daily practice and feeding.

APPENDICES

449

The restraining device proper consisted of a heavy metal funnel
9 cm. in diameter, lined inside with thin felt, which could be
dropped down over the food dish at will. The funnel was heavily
weighted at the top with lead so that it could not be removed from
the dish by the animals, when once in place. It was operated by
means of a wire cable as indicated in the diagram, the cable being
coated with vaseline to prevent the rats from attempting to climb

Fig. 2. Eestraining Device
A, bell-shaped food dish; B, perforated metal funnel, attached to picture
wire; C, hinged door, through which the animals were removed from the food
box; D, food box (shown in broken lines).

it. Perforations in the funnel permitted normal olfactory stimu-
lation during the period of delay. The operation of the device was
practically noiseless.

A total of 43 white rats of approximately 80-gram weight were
used. The control group included 15 animals and the two test
groups 14 each. The diet consisted of whole wheat bread and milk
and a sample of the regular diet was used as the incentive stimulus.
The general illumination of the room was kept as constant as pos-

450

APPENDICES

sible and was identical for all groups. Seven rats of each group
were tested daily at 3 p.m., and the balance at 7 p.m., thus equal-
izing the time of day factor for the groups. The animals were al-
lowed to feed, on each of 3 days previous to the training period, for
5 minutes in both entrance box and food box and then for a like
period in the adjacent feeding cage. During this preliminary con-
tact with the apparatus, the funnel was suspended about 6 inches
above the food dish in which food had been placed. The animals
were inserted and removed repeatedly during this period through
the door, C, in order to acustom them to this method of removal
which it was necessary to use in the later training series.

Only one trial per day was given, and this always at the same
time of day for each animal. Time was taken with a stopwatch, and
both forward and backward errors were checked. The two types
of error have been combined in the tables, however, since no sig-
nificant group differences appeared. An error was checked when
the ears of the test animal were seen to project into the entrance
of a cul-de-sac. The training was continued until 4 out of 5 trials
were made without error (norm B). A second norm of mastery
(norm A) — the first perfect trial — was introduced as a check on
the more severe norm, the computations being made of course from
the same records.

The essential difference in procedure for the different groups
centered around the food box, since conditions elsewhere were kept
uniform throughout. In the case of the control group the funnel
was raised exposing the food at the instant the door closed behind
the animal. After 5 minutes of feeding the rat was removed
through the door, C, and placed in a small cage, with food, for an
additional 5 minutes and then retured to its living cage. The pro-
cedure in the case of the two test groups was precisely the same,
except for the fact that the funnel was raised only after the inter-
val of delay, in each case, was over.

The results are presented in tables 1, 2, and 3 for both norms of
mastery and for all criteria of performance (trials, errors, time).
The statistical treatment of the average (norm B) appears in the
last four columns of each table, this average representing probably
on the whole the best group index. The median for the same norm

APPENDICES

451

agrees in general with the average, and the scores of norm A are
also in essential agreement with those of norm B. In fact these four
indices of central tendency exhibit a remarkable uniformity
throughout, and the data for trials, errors and the time all point in
the same direction.

Let us compare, first, the control and the 5-minute groups. If
we take the scores at their face value it is evident that we must
conclude that a 5-minute interval between the pattern response to
be learned and the feeding activity has no disadvantageous effect
upon the fixation of the maze habit. In fact, the values for both
trials and errors are — in most instances — slightly lower than
those of the control group, although this difference in favor of the
5-minute group is in no case significant. The 5-minute interval of
delay then did not increase the number of trials required to mas-
ter the maze, nor did it increase the number of errors made during
the period of mastery.

The average total time spent within the maze was, however,
somewhat greater for the 5-minute group, as will appear from an
inspection of table 3. The difference is not large but is consistent
regardless of norm of index of central tendency employed. Inas-
much as the number of trials, and of errors, was practically identi-
cal for the two groups, this difference in time would seem to re-
duce itself to speed of locomotion (average speed) in running
through the passage ways. The animals of the 5-minute group evi-
dently ran a little less speedily, or else made more frequent or
longer stops, on the whole. Whether this difference in time per
trial was due to the influence of the interval of delay cannot be
certainly known from the data at hand. It may have been due
instead to an accumulation of rats in this group that naturally
run at a slower than average pace. There is a large individual dif-
ference in natural speed of running in the white rat, as anyone
knows who has attempted to make use of a temporal norm of mas-
tery in connection with either maze or problem box. Even if the
slowing up in speed of locomotion were clearly due to a difference
in motivation arising from the interval of delay it would still be
difficult to connect this in any important way with the process of
fixation. There is little reason to suppose that this slight difference

452

APPENDICES

TABLE 1

Showing mean number of trials required in mastery of maze

GROUP

NORM A

NORM B

.Median

Average

Median

Average

Q

CO

S. D. (average)

D

S. D. (D)

V

Control

7.0

6.9

10.0

10.6

2.2

0.6

C. @ O.M. 2.00

20.6

One-minute

7.5

7.3

12.0

12.4

2.8

0.7

O.M. @ F.M.

22.6

2.40

Five-minute

5.5

5.6

10.0

10.2

2.2

0.6

C. @ F.M. 0.04

21.6

TABLE 2

Showing mean number of errors during mastery of maze

NORM A

NORM B

GROUP

verage)

D

C

©

a

C

i

8>

3.

S. D. (D)

V

-5

©

2

9

Q

Q

1

<

3

<

CO

CO

Control

22.0

23.8

22.0

24.7

6.8

1.8

C. @ O.M. 2.3

27.5

One-minute

29.5

29.6

29.5

31.2

8.4

2.2

O.M. @ F.M. 2.9

26.9

Five-minute

22.0

21.7

24.5

24.0

4.7

1.3

C. @ F.M. 0.3

19.6

TABLE 3

Showing mean number of seconds required in mastery of maze

NORM A

NORM B

GROUP

C

S

©

g

C

a

©

§
2

. (average)

s

D

. D. (D)

V

-5

©

"5
©

©

>

Q

Q

©
S

>

a

<

CO

CO

Control

208.0

195.9

237.0

232.5

34.8

9.0

C. @

O.M. 2.8

15.0

One-minute

229.5

233.5

263.5

294.9

75.3

20.1

O.M.

@ F.M. 0.7

25.5

Five-minute

211.0

222.1

263.5

275.6

61.5

16.4

C. ©

F.M. 2.3

22.3

APPENDICES

453

in speed of locomotion would bear any essential relation to rate of
fixing the maze habit.

The results of the 1-minute group do not yield themselves to so
obvious and simple an interpretation. The data are consistent
throughout in showing a slower rate of learning for this group in
comparison with either of the other two, regardless of norm of
mastery, or index of central tendency employed. Moreover when
the average for norm B is considered the difference is large enough
to be fairly reliable in most cases, and always so as between the 1-
minute group and the control. The acceptance of this result at its
face value would lead to the conclusion that although a 5-minute
interval of delay is not disadvantageous, a shorter one such as one
minute may be so. Such a conclusion, while not impossible, does not
appear to fit the logic of the situation very well. Nor is it sup-
ported by any fact of observation; there was no noticeable differ-
ence either qualitative or quantitative in the behavior of the 1- and
5-minute groups during the period of delay. They appear equally
active in both cases.

The extreme precautions taken to insure uniformity of experi-
mental conditions practically eliminate the possibility of any
constant error in the results. Nor do the data afford any clue con-
cerning a possible accumulation of a variable error, save, perhaps,
in the matter of sampling. As already noted the 1-minute group
showed, on the whole, a greater tendency to variability than the
other two groups. Moreover, a single animal in this group extends
the upper limit of the range in trials (norm B) from 16 to 20, as
indicated in table 4. These facts suggest that perhaps the 1-minute
group is less homogeneous than the other two, and hence not as fair
a sample. A standard error of sampling formula can hardly be ap-
plied to such small groups. However, a rough check on this point
seemed to be furnished by the following simple procedure. Each
of the three groups was divided into two sub-groups of equal size
by chance, and averages for these sub-groups computed for all
criteria. Fifteen numbers were drawn from a hat and placed alter-
nately into X or Y pile, each number representing a given individual
animal. A large difference in the averages of the two sub-groups
should mean lack of homogeneity in the major group, although

454

APPENDICES

TABLE 4

Distribution table covering number of trials required for complete mastery

(norm B)

7

8

9

10

11

12

13

14

15

1G

17

18

19

20

Control

1

2

2

3

1

4

0

1

1

0

0

0

0

0

One-minute

0

0

1

3

1

5

0

2

0

1

0

0

0

1

Five-minute

2

1

1

5

2

1

0

2

0

0

0

0

0

0

TABLE 5

Averages based upon a chance division of each group into sub-groups X and Y

TRIALS

ERRORS

SECONDS

GROUP

X

Y

X

Y

X

Y

Control -

10.1

11.1

24.0

25.3

215.4

252.0

One-minute

10.9

14.0

29.6

33.9

250.9

339.0

Five-minute _

10.0

10.4

23.9

24.1

250.1

301.0

such a simple method gives no more than a rough indication.
An inspection of table 5 will show that the difference between the
averages of the sub-groups, arranged on a chance basis, is very
slight for both the control and 5-minute group, but is quite con-
siderable for the 1-minute group. This is additional evidence that
the 1-minute group is less homogeneous than the other two groups,
and raises the question of the validity of the various indices of this
group.

It seems best, therefore, to leave the matter of the effect of the
1-minute interval open for further investigation. This of course
does not disturb in any way the clear cut evidence that a 5-minute
interval of delay between the act to be fixated and the feeding
sponse had no measurable effect upon the rate of fixation. This
latter result must be considered as so much evidence against the
law of effect, insofar as the latter insists that the value of the ' ' sat-
isfying state" is a function of its " nearness in time" to the act, or
series of acts to be fixated.

Bibliography

(1) Watson, J. B. The effect of delayed feeding upon learning. Psycho-
biology, i, 51-59.

t

3. THE EFFECT OF DELAYED INCENTIVE ON THE
HUNGER DRIVE IN THE WHITE RAT 1
(EXPERIMENT 2: THE LEARNING METHOD)

E. L. Hamilton

I. Introduction

While the Columbia Obstruction Method doubtless offers the
most direct and satisfactory method of measuring' animal drives,
the marked influence of the incentive-drive situation in animal
learning suggested the use of this more indirect method as a check
on our results reported in Part II, 2.

Watson (14) was perhaps the first to attack the problem of deter-
mining definitely the effect of delayed incentive on rate of learning.
The apparatus employed was a modification of the sawdust box
provided with a device for restraining the animal from obtaining
the food until an interval of time had passed. The food was placed
in a cylindrical container 5 cm. in diameter and 3 cm. high, sup-
plied with a lid perforated with several mm. holes to allow for ol-
factory stimulation. A small vertical rod was screwed into the
center of the lid and passed through a hole in the wire mesh of the
problem box, passing out through the top. By means of this rod
the experimenter could allow the animal to get food after a given
interval.

He used 6 male rats approximately 60 days of age and 6 male
rats about 100 days of age. The twelve animals were divided into
two groups, six of which had to learn the problem by the usual
method of immediate feeding and six others by the method of de-
layed feeding. The animals were distributed at random into the
two groups, with the only precaution that each group contained
three young animals and three older ones.

xEeprinted with modifications from Genetic Psychology Monographs, 1929,
5, No. 2, 137-66.

455

456

APPENDICES

As preliminary training, all twelve animals were allowed to get
their food in the box a number of times before the training series.
At all times during the tests the lid to the food box was left on in
the case of both groups. The lid was lifted by the time the animals
reached the food box for the control group. The animals of either
group were allowed to eat for five seconds on each trial after enter-
ing the box and then were lifted out and taken back to their living
cages. Only one trial per day was given. The same method was
adopted for the delayed incentive group except that in this case the
lid to the food box was held down for 30 seconds. Watson found
that the animals were very active during the period of delay. Table
1 shows the average time per trial for the two groups. Examina-
tion of the data indicates that a delay of 30 seconds between the
solving of the problem and obtaining the incentive (food) has no
effect upon speed of forming this habit. Such decided changes in
procedure as the increase of trials to two trials per day and an
interchange of incentive conditions between the immediate and
delayed groups produced no noticeable effect upon the results.

Simmons (6), in her study of the effectiveness of various incen-
tives in aimal learning, made a brief test of the effect of delayed
feeding on the learning of a complex maze. She used five male and
five female albino rats between three and four months of age. The
rats were selected by chance so that the two groups represented
random samples with the exception that each group had approxi-
mately an equal number of males and females. As preliminary
training, the rats were fed in a special feeding cage, for 7 to 10
days before the experiment was begun, at the same time of day at
which they were later fed during the experiment proper. Two
groups were run, a standard and a delayed group. The former were
fed, during the training period, immediately after practice each
day, as is usual in ordinary laboratory procedure. The delayed
group were fed in the feeding cage one and one-half to two hours
after the last run. Work was done in the late afternoon by electric
light. Two trials per day were given to each group, and the maze
was considered mastered at that trial after which nine out of ten
successive runs were made without error. The results appear in
table 2. The immediate feeding group had records slightly su-

APPENDICES

457

TABLE 1

Table from Watson showing average time per trial

Trials

Average delayed feeding
Seconds

Average immediate feeding
Seconds

1

385

923

2

58

111

3

84

89

4

30

24

5

18

36

6

17

23

7

13

9

8

12

13

9

9

8

10

9

15

11

13

9

12

8

13

13

8

9

14

9

7

15

11

24

16

7

8

17

6

9

18

8

6

19

10

5

20

9

5

21

8

6

22

6

6

23

8

7

24

6

7

25

6

8

26

6

7

27

7

6

28*

5

4

29

5

5

30

5

5

31

4

5

32

4

4

n

5

A
T

34t

6.3

3.6

35

4.0

4.3

36

4.6

4.3

37

6.0

5.3

38

3.3

3.3

39

3.3

4.3

40

5.0

5.3

41

3.3

4.3

* At this trial the two-trials-per-day schedule was introduced.

t At this trial the delayed group was fed immediately and the control group
delayed 30 seconds. Only three animals (60 days of age) of each group were
included.

perior in respect to number of trials, but inferior in respect to time
and errors. Simmons does not consider these differences large nor,
in view of the large mean deviations and lack of agreement between
the three criteria, very significant. Also the difference in error

458

APPENDICES

score between the groups was confined largely to the first five trials
and the records were almost coincident thereafter. She concludes
that there is no significant difference between the records obtained
under the two conditions.

The more recent study of delayed incentive on rate of learning
by Warden and Haas [Hamilton] (8) will be found in Appendix
2 of this volume. Larger groups were used and a better control of

TABLE 2

Table from Simmons showing averages with initial trial omitted

CONDITIONS

TRIALS

.... . ,

ERRORS

TIME

Immediate feeding _ _

Delayed feeding .

27.8
33.3

127.7
93.3

1761.44
1105.10

conditions was secured. The data appeared to warrant the conclu-
sion that delays of one and five minutes had no appreciable effect
on the rate of learning.

It will be seen from the above review that very little work has
been done on this topic. Such results as have been obtained must
be considered inconclusive. Watson's study was limited to a single
delay interval of 30 seconds, and Simmons' to one of approxi-
mately 2 hours. The need for a more systematic investigation of
the effect of delayed incentive on rate of learning is clearly indi-
cated.

A. Method and procedure

Apparatus. The apparatus used with the learning method of
studying delay was the standard unit maze designed by Warner
and Warden (13). Figure 4 shows the types of unit of which the
maze is constructed. The material used in building the maze was
18-gauge, galvanized, Armco sheet metal covered with aluminum
paint to insure uniformity of appearance. As is seen in the figure,
the units consist of walls and a roof, with no floor. They were set
up on a platform 7x8 feet in size, which had been covered, after
very smooth finishing, with battleship linoleum (cemented down).
This platform served as a base or floor. The angle of junction
of cul-de-sac to pathway is the same as the angle at the end of the
cul-de-sac, so that the animal can get no cues from looking down

APPENDICES

459

into the latter as to whether it is a blind alley or not. The roof of
both pathway unit and cul-de-sac is made of very heavy wire mesh
neatly riveted into the frame, while the two ends of the cul-de-sac
are also made of wire mesh in order to secure lighting conditions
in the alley similar to those of the pathway.

Several problems arose in connection with the apparatus. The
first concerned the pattern to be used in the present study. It was

Fig. 4. Photograph showing types of unit of the Warner-Warden Standard

unit maze

(For Figures 1, 2 and 3 of this monograph see Part II, 2)

desired that the pattern be sufficiently difficult to bring out indi-
vidual and group differences and yet simple enough to be learned
in a reasonable time. The pattern finally chosen (Figure 5) met
these requirements very well. The conditions of the study made
the use of a screen and peep-hole impossible for practical purposes,
since the noiseless operation of the doors was especially important.
Great care was needed to prevent the animal from becoming con-
ditioned against the restraining compartment, and one of the most
common causes of fear in experiments with rats is the matter of
occasionally pinching their tails in the doors of an apparatus.
Automatically controlled doors were impractical and doors oper-
ated by a system of levers and strings proved to be too slow for
the rat 's quick movements. In many animal mazes the entrance and
food boxes are a considerable distance from each other, necessitat-
ing the use of two experimenters, one to introduce the animal and
the other to remove it, or else the experimenter must change his
position during a given trial. In the pattern used, the entrance and
food boxes were in close proximity to each other, making the change
of position of the experimenter unnecessary. The experimenter

460 APPENDICES

took a position midway between the two boxes and remained quietly
throughout a given trial operating manually the doors for both the
entrance and the food boxes.

The fact that the pathways were somewhat deeper than those of

Fig. 5. Maze Pattern

E, entrance compartment; F, food box; B, restraining compartment; dlt
door separating entrance compartment from maze; d2, door separating restrain-
ing compartment from maze; o and c represent door (cL) in open and closed
positions, respectively; d3, door separating restraining compartment from food
box; xx, rubber washers on brass rod to which door is fastened; y, piece of thin
metal used to close opening in R.

most rat mazes (5 inches in the clear), together with the fact that
a wire mesh was used as a covering for the maze, aided in the elim-
ination of the experimenter from the visual field of the animal. A

APPENDICES

461

stop-watch of the lever type was used in the experiment and proved
to be practically noiseless.

Several slight modifications were made in the apparatus as de-
scribed by Warner and Warden (13). In previous work with this
maze in the Columbia laboratory it was found that the animals
hesitated as if frightened upon coming from the covered pathways
into the food box, which was not covered with wire mesh. To pre-
vent this behavior, a piece of mesh, of the same size as that forming
the roof of the pathway and cul-de-sac units, was fastened over the
long, narrow portion of the food box, so that the latter would more
nearly resemble the other parts of the maze. A thin metal door
(Figure 5, dlf d3) of the swinging type replaced the heavier metal
sliding door at both the entrance and food compartments. The
door, in each case, was fastened to a brass rod which turned in holes
in the sides of the unit. Rubber washers (xx) prevented any noise
which might result from the rod turning on the metal, and felt
fastened to the bottom of the doors made their operation practically
noiseless.

The problem of restraining the animal from the food during the
delay period was met by using a pathway unit as a restraining com-
partment (R). This was placed between the last pathway unit of
the pattern and the food box. Since this involved joining a path-
way unit with another pathway unit instead of a cul-de-sac, as in
the ordinary maze pattern, two openings were left in the restrain-
ing compartment. The opening nearest the food box was closed
up by a thin piece of metal (y) especially cut to fit it. The other
opening was used to good advantage. A thin metal door (d2) of
the swinging type with vertical pivot was made and served two
purposes. When it was open it formed the wall of the maze between
the last pathway unit and the restraining compartment; when
closed, it served to prevent the animal from retracing its steps into
the maze, thus forming an effectvie restraining compartment. A
wire hook attached to the door provided a secure fastening. The
door (c?3) to the food box was held closed by a brass stop while the
animal was in the restraining compartment.

A piece of plate glass was used, to cover that part of the food
box in which the food was placed, to prevent the animals from

462

APPENDICES

climbing out during the allotted time for feeding. The experi-
menter soon developed a noiseless method for replacing and remov-
ing this glass.

Animals. The subjects for this experiment were 105 male albino
rats of approximately 100 grams in weight and 2 months of age at
the beginning of the experiment, obtained from the Albino Supply,
Inc., of Philadelphia. Upon arrival at our laboratory, the animals
were placed in large cages, 30x15x15 inches, about 12 to 15 animals
in a cage, and kept for two weeks before they were used. This two-
week period previous to testing enabled them to become accustomed
to laboratory conditions and to the experimenter, who handled
them daily in feeding. All the animals remained in normal health
throughout the experiment.

The animals were fed whole-wheat bread and milk throughout
the experiment, with the addition of a weekly ration of greens.
This diet was chosen instead of McCollum's Standard Diet because
of the high incentive value of bread and milk (6). The animals
were given their ration after the daily trial in the maze, the same
food being used both for incentive and for regular feedings. Fresh
water was supplied daily from inverted bottles with nozzles to
prevent fouling. The animals were supplied with water even dur-
ing the 24-hour starvation period. Before the starvation period,
which marked the beginning of the experiment, fresh sawdust was
supplied to insure complete absence of any food in the cage.

Five groups of animals were used. Table 10 shows the number
of animals in each group. In addition to the 0-delay, or control
group, four groups were used with the following intervals of de-
lay : 1 minute, 3 minutes, 5 minutes, and 7 minutes.

The starvation period was 24 hours for all groups. The number
of trials per day was constant for the groups, only one trial per
day being given to each animal.

Procedure. The sex drive was kept constant for all groups by
using males. It seemed advisable to eliminate all oestrous cycle
complications by this restriction in the interests of a more ade-
quate control of extraneous drive conditions. They were segre-
gated as to sex during the two weeks previous to the beginning of
the experiment and throughout the period of learning. The factor

APPENDICES

463

of time of day was controlled by running all the animals between
the hours of 9 p.m. and 4 a.m. Since the laboratory was practically
deserted at this time, noise distractions were reduced to a minimum.
The room in which the experiment was conducted was lighted in-
directly by electric light, the same number of lamps being lighted
each night. The illumination was thus uniform in all parts of the
room. The maze was kept in a constant position in the room,
throughout the experiment and care was taken not to shift the pat-
tern to new positions on the platform. The temperature was kept
fairly constant by means of a thermostat.

TABLE 10

Grouping of animals

LENGTH OF DELAY PERIOD

NO. OE ANIMALS

0 (Control)

16

1 Minute

22

3 Minutes

22

5 Minutes

20

7 Minutes

25

The preliminary training consisted in feeding each animal singly
in the entrance and food compartments 1 minute each and in a small
feeding cage nearby for 10 minutes thereafter. A glass plate was
kept on the food box to prevent the animal from climbing out dur-
ing the 1-minute feeding period. This preliminary feeding was
given the animals for two days. The animals were starved for 24
hours before each feeding period, as during the training series.
The first trial, which occurred on the third day, was included in
the preliminary training. The arguments for this procedure, which
is the rule in the Columbia laboratory, have been developed by
Cook (2).

The procedure for the 0-delay, or control group, was that usually
employed in maze experiments, except that the animal was fed for
a period of 1 minute in the food box and for 9 minutes in an adja-
cent feeding cage after the daily trial. The door between the maze
proper and the restraining compartment (Figure 5, d2) was open
when the rat was placed in the maze. The stop-watch was started
at the same time the animal was inserted in the maze. When the
animal's tail cleared the door (d2) at the restraining compartment,

464

APPENDICES

the door was closed and the watch stopped. The door (d3) between
the restraining compartment and the food box was raised imme-
diately, allowing the animal to pass to the food box, and then
closed. ds was raised, since this procedure was necessary in the case
of the delay groups. The watch was started again, when c?3 was
closed, to time the 1-minute feeding period in the food box. After
the animal had fed for 1 minute, it was removed to the adjacent
feeding cage and allowed to remain for 9 minutes.

The procedure for the delay groups was the same as that for the
control group, with the following exception: Upon entering the
restraining compartment, the animal was delayed for a period of
time (1 minute, 3 minutes, 5 minutes, or 7 minutes) by closing the
door (d2) behind it. After the period of delay was over, the ani-
mal was permitted to enter the food box through d3, which was
raised, as in the control group.

Data were taken in terms of trials, errors, and time in seconds.
An error was checked when the animal's nose projected past the en-
trance to a cul-de-sac. This criterion of error was chosen because
the entrance of the nose in a cul-de-sac was more easily observed,
and hence could be more accurately checked than other possible
movements of the animal. Backward errors were checked for each
unit of the pathway. Notes were made on the general activity of
the animals in the delay compartment during the period of delay.

A change in the procedure was introduced at the 100th trial in
the case of those animals which had not learned after running 99
trials and which had shown no indication of learning for 50 trials.
On the 100th trial and thereafter, these animals were allowed im-
mediate access to the food box, as in the case of the control group.
This change was introduced to ascertain whether these animals
were lacking in ability to learn the maze or whether failure to learn
up to that time was due to the factor of delay.

The animals were run until they had attained the norm of mas-
tery of four perfect trials out of five (Norm B), except the nine
animals which had not reached the norm at the 99th trial. The data
were also computed on the basis of the norm of two perfect trials
out of three (Norm A), and under this norm all of the animals had

APPENDICES

465

learned by the 99th trial. In computation of scores the last four
trials involved in the norm of mastery were omitted.

B. Results

The main results are based on the scores obtained by using the
norm of two perfect trials out of three (Norm A) for the reason
that only under this norm did all animals complete the learning
process. The scores obtained under Norm B will be treated in a
later section.

Norm A. The results are presented in Tables 11, 12, 13, 14, 15,
16, and 17, and in Figures 6, 7, and 8. Table 11 gives the fre-
quency distributions covering trials, errors, and time for the vari-
ous groups. Beginning at the fourth line, it reads as follows : one
animal in the 0-delay group, one in the 1-minute group, and one in
the 3-minute group, none in the 5-minute, and one in the 7-minute
group learned in four trials. Two animals in the 0-delay group
and none in any of the other groups, made between 13 and 16 er-
rors in learning. Four animals in the 0-delay group, one in the
1-minute group, none in the 3-minute group, one in the 5-minute
group, and none in the 7-minute group required from 91 to 120
seconds to learn, etc. Table 12 gives the frequency distributions
covering trials, errors, and time for the various groups, using larger
intervals (1-5 for trials, 1-15 for errors, 1-150 for seconds). Table 13
is a cumulative frequency table showing percentage of group having
learned by the end of each successive five trials. Figure 6 shows
the same thing in graphic form. Table 14 presents the measures of
central tendency and measures of variability for each period of
delay. Table 15 gives the measure of reliability of the difference
between the averages of various groups for trials, errors, and time.
Table 16 gives the average errors per trial for each five trials. Figure
7 shows the same thing graphically. Table 17 presents the aver-
age time per trial for each five trials. Figure 8 gives the same
data in graphic form.

When we examine the averages for the groups based on the scores
for trials (table 14), we note that the averages and medians of
the delayed incentive groups are considerably higher than those

466

APPENDICES

Fig. 6. Cumulative Frequency Curves Showing Percentage of Group Hav-
ing Learned by the End of Each Successive Five Trials (Norm A)
Control group (dash-line) ; 1-minute group (solid line) ; 3-minute group
(dot line) ; 5-mnute group (dot-dot-dash line) ; 7-minute group (dot-dash line).

of the control group. It will be seen from the table that the effect
of a 1-minute delay is practically to double the number of trials
required to learn the maze. The difference between the averages of
the 1-minute group and the control group is highly reliable (3.82),
as is shown in Table 15. The scores on errors and time show very
much the same tendency.

The scores on trials and errors for the 3-minute, 5-minute, and
7-minute groups are only slightly different from those of the 1-
minute group (Table 14), and in no case is the difference very re-

APPENDICES

467

J CO

CO

s

2

O C
. O

CU to

3

O

u

o h

2 5
O d

w N N ^ »-« 04 -< to

NO ^- ^

OoOOOOOOOOOOOOOOOOOOoOOOOOO

h h h m 1-1

HNH HNN V* M ^ »h N N H

H H H H CO NHH v-l *-« 1-t

CSJ t-f C\J »-< — I CN> »H ~*

,tooN>eo<fcoN'«o + ooN\£io*ooNvoothooNvooTt-oo

tH i-< ^ tO t-< CO tH i-H

*H tH ,~| H OJ

i-t to CM <N to

468

APPENDICES

^.5

s

2 d

s

an

D. C

rs

O d

o .£

ooooooooooooooooooooooo
^•NOM^e>iN»ooo«>+soMvoe>N<«ooiH»f-NO

NV0O400N<OO*»NveO*00N'«O + 08N^O

i-i ^ tN w

APPENDICES

469

Fig. 7. Average Errors per Trial for Each Five Trials
Control group (dash line) ; 1-minute group (solid line) ; 3 -minute group
(dot line) ; 5-minute group (dot-dot-dash line) ; and 7-minute group (dot-dash
line).

liable (the range in reliability index is .12 to 1.44), as will be seen
from Table 15. This means that there is little difference between
the effects of 1-minute, 3-minute, 5-minute and 7-minute delays
when trial and error scores are considered.

The error curves (Figure 7) indicate that the animals in the de-
layed incentive groups did not differ markedly from each other in
errors per trial. They consistently made more errors than the con-

Fig. 8. Average Time per Trial for Each FIve Trials
Control group (dash line); 1-minute group (solid line), 3-minute group
(dot line); 5-minute group (dot-dot-dash line); and 7-mnute group (dot-dash
line).

APPENDICES

o c
. o
o v

u">00^00©u">00»-*

i-« i-h

OoOOOOOOO
^©w>©u»io*/">©U"i

i-< r> t/-, m M M i— -n^ho^»-i

OHlflHNUIO + HOH

N S w <h m

APPENDICES

471

trol-group animals. Thus, these results corroborate our conclusions
based on scores for trials and for errors.

In the case of the time scores, however, the averages tend to in-
crease with increase in length of delay period. The differences be-
tween the averages also tend to be slightly more reliable (the range
in reliability is .12 to 2.26), as is shown in Table 15. In general,
time scores are not considered as valid an index of learning in maze

TABLE 13

Cumulative frequency table showing percentage of group having learned by the
end of each successive five trials (Norm A)*

GROUPS

NO. OF TRIALS

Control

1 Minute

3 Minutes

5 Minut-es

7 Minutes

1-5

6.3

13.6

9.1

0.0

8.0

10

62.5

27.3

22.7

20.0

40.0

15

81.3

31.8

36.4

40.0

56.0

20

100.0

54.5

54.5

50.0

68.0

25

59.1

63.6

55.0

76.0

30

.77.3

72.7

70.0

76.0

35

90.9

81.8

90.0

88.0

40

95.5

95.5

95.0

92.0

45

100.0

100.0

95.0

100.0

50

100.0

* Two perfect trials out of three.

experiments as trial and error scores, and hence more weight will
be given to the latter. The inconsistency of the time scores with
the general tendency of similarity among the delay groups seems
to indicate that the longer delays caused the animals to run more
leisurely through the maze. The lengthened time score arising in
connection with delay is supported by behavior observations made
by the experimenter. The animals appeared to be less motivated
and ran slowly through the maze in a hesitant manner, while the
control-group animals ran more quickly and directly through the
maze to the food box. The time curves (Figure 8) show this same
tendency — increase in time per trial with increase in length of
delay period. The 7-minute group curve strikingly illustrates this
tendency, rising much more suddenly than the other curves and re-
maining at a higher level. Thus the ainmals in the delay groups
took a longer time per trial with increase of length of delay period.
The notes made on the behavior of the animals in the delay com-

472

APPENDICES

■<*- r*- O

H oo d oo 6

M N OO N 0\
11 M W «

© © © © ©

Os N OO ^ w
O t*» i-H OO ©\
th N w to N

»/% »-< Cs| tr\
t>. c4 \A W

oo ro w N

^ 4* M N 6
W « NVO

O ui O ^ ©
ff! 6 N 6 ui
CM OO rj-

_H ^ _

'ttOTj-vOrj-

O «^» O O

<u i> <v cj
D 3 3 D
©.g.S.S.S

0 UBqj

anjj b jo 001
aqj jo -q-s

3DU3J3JJIQ

»v jo crs

0 UBqj J3JB3J3

3DU3J9jy|p

anjjB jo ooi

33U3J3^ip

aqj jo -q's

93U9Jdgip

»V J0 'Q S

3DU3J3JJIQ

0 UBq) J3JB3J3
33U3J9Jfip

snj) b jo ooi

UI 83DUBq3
33U3J3gip

aqj jo a' s

3DU3J3jyip
3qj JO -q s

3DU3J3JJia

to w> rf- ih C^J

't«ri^vO^OONNNOO

MOOOvOvOvO

ooooHNHViMTt-Ovm

0^0;N^O;oo>ANN

C G C C C C

SS5SSS

c c c cro*r,tvW^t^t^

S LT IT -O XI X -O X

W n rj n oj R
hw^N

xi xi xi x .5 .5 .S .5 .5 .5
cs « « w^SSSSS

APPENDICES

473

TABLE 16*

Average errors per trial for each five trials

No. of

Groups

trials

Control

1 Min.

3 Min.

5 Min.

7 Min.

1 c
i- j

c
j

c
o

c
o

7

e

10

2

2

2

3

3

15

1

2

2

2

2

20

1

2

2

2

3

25

1

2

2

2

3

30

1

3

2

2

3

35

2

2

3

2

40

2

1

2

2

45

1

2

2

2

50

1

3

2

2

* In most learning experiments it is not possible to obtain an adequate pic-
ture of the average time per trial or of the average errors per trial for any
group, due to the fact that individual animals reach the norm of mastery after
different numbers of trials and are dropped out of the experiment. Thus the
latter part of the curve must be based on smaller and smaller numbers of ani-
mals. In our experiment, however, we continued to run the animals in an effort
to have them reach the higher norm (Norm B.) We, therefore, had data on a
large number of animals beyond the point where they had learned under the
lower norm (Norm A), and in constructing our tables (16, 17) and curves
(Figures 7, 8) for time and errors per trial, we used this available data so as
to increase in number of animals making up the average at a given point. This
method probably gives a closer approximation to a true curve than possible
statistical methods. The exact number of animals upon which the average at
any point is based may be ascertained by reference to Figure 11.

TABLE 17

Average time per trial for each five trials

No.of
trials

Control

1 Min.

Groups
3. Min.

5 Min.

7 Min.

1-5

18

21

27

30

24

10

7

9

12

12

13

15

6

9

13

14

17

20

7

11

16

16

24

25

8

13

15

18

25

30

7

15

18

18

25

35

16

21

24

22

40

18

19

21

27

45

13

23

20

23

50

16

29

20

22

partment reveal no noticeable differences between the groups. All
animals were equally active, running back and forth, biting and
clawing at the doors, etc. In no case did an animal remain quies-
cent or merely wait for the door to be opened. When released from
the compartment, the animals in all cases ran directly to the food

474

APPENDICES

APPENDICES

475

and ate for the full minute allowed them in the food box, except on
the first few trials during which they exhibited the usual explora-
tory activity which is natural for the rat in a relatively new situa-
tion.

Norm B. An attempt was made to run all of the animals until
they attained the norm of four perfect trials out of five. Since 9
animals had not reached the norm at the 99th trial, and had shown
no indication of learning for 50 trials, the change in conditions
from delayed to immediate incentive was introduced. The effect of
this change will be discussed in this section.

The results are presented in Tables 18, 19, and 20 and in Fig-
ures 9, 10, 11, and 12.

Figure 9 presents the distribution of animals according to trials,
errors, and time taken to learn, each square representing one ani-
mal. The animals which had not learned by the 99th trial are indi-
cated in solid black.

Table 18 is a cumulative frequency table showing percentage of
group having learned by the end of each successive five trials. Fig-
ure 10 presents the data in graphic form.

Table 19 shows the effect of change in procedure from delayed to
immediate incentive conditions, in terms of average time per trial,
on the 9 animals which had not reached Norm B by the 99th trial.
Figure 11 shows the same thing in graphic form.

Table 20 shows the effect of change in procedure from delayed to
immediate incentive conditions, in terms of average errors per
trial, on the 9 animals which had not reached Norm B by the 99th
trial. Figure 12 shows the same data graphically.

We note from Figure 9 that 9 of the animals failed to learn,
under this norm, in 99 trials. Obviously true averages and me-
dians could not be obtained for the groups in which these animals
occurred (3-minute and 5-minute groups). The averages and me-
dians of the remaining groups are indicated in Figure 9. It will
be seen from the distributions according to trials, errors, and time
(Figure 9) that the delay groups all learned more slowly than the
control group, showing the same general tendency as under Norm
A. It appeared to be of some interest to see what would happen,
under conditions of immediate feeding, to the 9 animals which had

476

APPENDICES

APPENDICES

477

H-

hh

3^

s

*

%

St

t>

*o

-

\

\

N

-

\

\

\

V

3

-1-

u

S>

s

u

-

■

—

I

Fig. 11. Effect of Change in Procedure from Delayed Incentive to Imme-
diate Incentive Conditions in Terms of Average Time per Trial

not learned. Accordingly, on the 100th trial and thereafter the
interval of delay between running the maze and being fed was
eliminated and the animals were allowed to proceed at once into
the food box, as in the case of the control group. Although the
animals had shown no progress in learning for 50 trials past, they
speedily began to learn under the new conditions. As will be seen
from Figure 10, by the 105th trial all 9 animals had reached the
degree of mastery required under Norm B. The change in condi-
tion thus brought about almost immediate learning. This indicates
that the animals were not lacking in ability to learn the maze and
hence failure to learn up to the 99th trial was probably due to the
factor of delay.

Figure 11 shows that the average time per trial decreased very
markedly after the change in procedure was introduced. On the
101st trial the curve descended to a point lower than had been

478

APPENDICES

TABLE 18

Cumulative frequency table showing percentage of group having learned by the
end of each successive five trials (Norm B)*

NO. OF TRIALS

GROTTNS

Control

1 Minute

3 Minutes

5 Minutes

7 Minutes

1-5

0 0

ft n

ft n
u.u

A ft
4.U

10

43.8

18 9

u.u

k ft
o.u

1 R ft
JLO.U

15

50.0

18 9

i7.JL

i n (i

J.U.U

iiU.U

20

81.3

36.4

Q 1

if. J.

1 ft

39 0

,25

87.5

A^ ^

^u.u

/in ft

30

100.0

1

ov.x

1 3 £

JLO.U

3ft n

OU.U

Ad. ft

35

XO,£i

fin n

uu.u

^9 ft

40

68.2

31 8

7ft 0
/u.u

fift ft
uu.u

45

72*. 7

36.4

75.0

64.0

50

81.8

40.9

80.0

68.0

55

81.8

45.5

85.0

72.0

60

86.4

50.0

90.0

84.0

65

90.9

50.0

90.0

88.0

70

100.0

54.5

90.0

96.0

75

59.1

90.0

100.0

80

63.6

90.0

85

63.6

90.0

90

68.2

90.0

95

68.2

90.0

99

68.2

90.0

Change to immediate incentive conditions

105 | | 100.0 100.0 |

* Four perfect trials out of five.

reached in the 95-99 trials (delayed incentive conditions). This
indicates that the longer time per trial was probably due to the
conditions of delay and not to a naturally slower speed of running
in the case of these animals. Behavior observations made by the
experimenter support these conclusions. During the trials in which
the incentive was delayed, the animals appeared to be less moti-
vated, running slowly through the maze in a somewhat hesitant
manner as above mentioned. Under the immediate incentive condi-
tions, however, they ran quickly and directly through the maze to
the food box, resembling in this respect the control-group animals
in their behavior.

The error curve (Figure 12) shows the same general tendency
as the time curve. The curve descended on the 102nd trial to the
zero line. The general level is considerably lower than that of the
95-99 trials (delayed incentive conditions). This supports our con-
clusion that failure to learn up to the 99th trial was probably due

APPENDICES

479

|

-§

\

a

-1

A

-

■ft

-4

J"

J

Fig. 12. Effect of Change in Procedure from Delayed to Immediate
Incentive Conditions in Terms of Average Errors per Trial

to the conditions of delay rather than to inability to learn. The
nature of the errors made previous to the 100th trial also offers
corroborative evidence. The animals persisted in making one or two
errors, going into the same cul-de-sac each time, which would sug-
gest that the motivation was not strong enough to bring about per-
fect learning. In several cases, animals would run three perfect
trials out of four and then make errors on the following trials so
that the norm was not reached for many trials later. Evidently
the delayed-incentive conditions worked against consistency in er-
rorless runs since all animals reached Norm A by the 50th trial,
while under the higher norm 9 had not learned by the 99th trial.
That is, by reducing the motivation, delay operated to lower the
quality or degree of perfection to which the habit was developed.

In relating our results to those of previous investigators, we find
that few studies are comparable. Simmons (6) used a 2-hour period
of delay, so naturally no comparison with the present work, which
was restricted to much shorter periods, is possible. Watson (14)
found that a delay as short as 30 seconds had no effect upon speed
of learning. However, the interval he used was only half as long
as our shortest, so the results are not comparable.

In a former study made by the writer (8), intervals of delay of
1 minute and 5 minutes were found to have no effect on speed of

480

APPENDICES

TABLE 19

Effect of change in procedure from delayed to immediate incentive conditions
in terms of average time per trial

(The 9 animals which had not reached Norm B* by the 99th trial)

TRIAL

AVERAGE TIME IN SECONDS

95

22

96

23

97

27

98

35

99

35

Change to immediate incentive conditions

100

26

101

19

102

11

103

8

104

9

105

8

* Four perfect trials out of five.

TABLE 20

Effect of change in procedure from delayed to immediate incentive conditions
in terms of average errors per trial
(The 9 animals which had not reached Norm B* by the 99th trial)

TRIAL

AVERAGE ERRORS

95

2

96

2

97

2

98

3

99

2

Change to immediate incentive conditions

100

1

101

1

102

0

103

0

104

1

105

1

* Four perfect trials out of five.

learning. In the general tendency shown, this finding would agree
with Watson's conclusion, but is not supported by the present ex-
periment in which all periods of delay employed (including a 1-
minute and 5-minute period) decreased the speed of learning.
However, the previous study differed from the present one in one
important respect: the animal was delayed in the same compart-
ment with the food. That is, in the former experiment the animal
was directly stimulated by the presence of the unobtainable food,
whereas in the present case the direct stimulation of the food could

APPENDICES

481

occur only after the animal was released from the delay compart-
ment into the food box. In the present experiment the matter of
place of delay was much better controlled — the delay compart-
ment was entirely separated from the food compartment. Evidently
the situation of being delayed in the same place with the food, even
though the latter be inaccessible, has little or no effect on the fixa-
tion process. While such a situation involves temporal delay, the
pattern of the performance of running the maze is not definitely
broken since the animal is not delayed until it comes upon the spot
where the food is located. Our present technique involved not
merely temporal delay but temporal delay at a spatial position
which interrupted the general performance pattern of reaching
the food.

Perhaps the conditions in our present experiment may be said to
be analogous to the situation of children, at a birthday party, who
are not allowed to enter immediately the room where the birthday
cake is with its candles and decorations, but are made to wait in an
anteroom. The former study might then be analogous to the situa-
tion in which the same children are not detained in an anteroom
but are allowed immediate access to the refreshment room. How-
ever, they are prevented from touching or eating anything for an
interval of time. The latter situation has the added element of
direct stimulation of the desired objects, while the former involves
mere knowledge about them. Watson, who also did not use the
anteroom technique, but delayed the animals in the same compart-
ment with the food, obtained similar results with those of the pre-
vious study of the writer, but different from those in the present
experiment.

The results here obtained seem to show clearly that delay de-
creased the drive as measured by rate of fixation. In the case of all
of the delay groups, the rate of learning was reliably slower than
that for the control group, in terms of trials and errors. The fact
that running time was increased somewhat, with longer delays, in-
dicates a motivation difference which, however, did not affect meas-
urably the rate of fixation. Obviously, trials and errors are rather
gross units of measurement of efficiency and if finer units could
have been used, a decreased efficiency reflecting the slowing down

482

APPENDICES

in running time might have been found. On the other hand, there
is no need to suppose that speed of maze running is perfectly, or
even highly, correlated with rate of fixation.

There is some indication that delay affected the quality of the
learning, i.e., the habit was not formed as perfectly under the delay
conditions, as seen by a comparison of results under the two norms
of mastery. Although the animals in all groups had attained the
lower norm (Norm A) 2 by the 50th trial, 9 of them had failed to
learn under the higher norm (Norm B)3 by the 99th trial. Delay
•had apparently decreased the drive in these animals and there was
insufficient motivation to bring about further perfection of the
habit. This interpretation is borne out by the fact that the 9 ani-
mals immediately developed the habit to the higher level (Norm
B)4 when the delay was eliminated.

It would seem to be impossible to relate our findings in any very
important way to the law of effect as formulated by Thorndike.
Our results agree with an increasingly large body of data showing
the advantageous effects of motivated over unmotivated learning
situations. If the terms "satisfying state" and "annoying state,"
as employed in the statement of the law of effect (7, p. 4), be taken
in their usual psychological connotation as indicative of the sub-
jective states of the organism, then naturally our findings could
not be stated in terms of this law since we have no index of the
subjective states of our animals. It might be argued that they
"enjoyed" the delay in some anticipatory way with as much
cogency as that they were "annoyed" by it. The drive and incentive
values obtained serve as objective indices of certain tendencies of
the organism under the various conditions of motivation employed
— and that is all that can be said. The argument covering this
point has been made by Carr (1) and other writers.

As a corollary to the law of effect, Thorndike develops the prin-
ciple that the strength of the bond is determined by the temporal
closeness of the response and the satisfying or annoying state which
follows. If we assume, as Thorndike seems to, that the incentive
(food) produces a satisfying state, then our results would suggest

2 Two perfect trials out of three.

3 Four perfect trials out of five.

APPENDICES

483

that very definite restriction must be placed upon the value of the
temporal factor in this type of learning. For, while any interval,
within the limits investigated, cuts down very markedly the effi-
ciency of the learning process, the longer intervals were hardly
more effective than the shorter. Our results do not afford us a basis
for conjecture as to whether delay operated as a positive "punish-
ment" or merely as a distraction from the incentive situation in a
purely negative sense, nor do they offer any clues as to the physi-
ological (or neurological) processes underlying the behavior called
out.

C. Summary of results
Norm A4

(1) An interval of delay as short as 1 minute between the run-
ning of the maze and the feeding response markedly decreased the
rate of learning, in terms of the scores for trials, errors, and time.
(The increase in trials taken to learn was 97 per cent; in errors, 85
per cent; in time, 112 per cent.)

(2) Longer delay periods of 3 minutes, 5 minutes, and 7 min-
utes did not further decrease the speed of learning, as indicated by
the scores on trials and errors. The time scores, however, were
considerably lengthened as the interval of delay increased in
length, indicating a slower rate of progress through the maze.

Norm B 5

(1) When a high norm of mastery (four perfect trials out of
five) was used, 9 of the 105 animals failed to learn by the 99th
trial, although all had learned by the 50th trial under Norm A.

(2) Change of conditions from delayed to immediate incentive
had a markedly favorable effect on the performance of the above-
mentioned 9 animals. Time and error scores decreased immediately
and all of the animals learned within 5 trials after the change in
procedure, although they had made little or no progress for ap-
proximately the last 50 trials under the delayed incentive condi-
tions.

* Two perfect trials out of three,
s Four perfect trials out of five.

484

APPENDICES

D. Comparison of obstruction and learning methods

In a comparison of the results obtained in the two experiments
it is essential to have clearly in mind the method and procedure
employed in each experiment. The most important differences in
method have been arranged in tabular form as follows:

(1) The length of the test period was arbitrarily limited to 20
minutes in the obstruction experiment, whereas time per trial was
determined by the animal and varied from time to time in the case
of the maze experiment.

(2) Only a single test period was given each animal in the Ob-
struction Method, hence only a short time required per animal. No
animal in the longest delay group (3-minute) required a longer
time than one hour for the complete test. In the maze study, how-
ever, one trial a day was given until the maze had been learned,
and thus the experiment extended over a much longer period of
time per animal. An animal in the 3-minute group might require
as long as three hours of testing time.

(3) The period of starvation was 48 hours in the obstruction
study. This period was used in order to get the maximal state of
the hunger drive which should best bring out the effects of delay.
In the maze experiment a period of 24 hours was used, since one
trial per day was given.

(4) The number of intervals of delay was determined by the
number of crossings possible within the 20-minute period in the
obstruction study. None of our animals reached the theoretical
limit which has been shown to be about 44 crossings (9). On the
other hand, the number of delay periods per trial in the learning
experiment naturally was limited to one, since only a single daily
trial was given.

(5) In reacting to the incentive situation in the obstruction
experiment, the act of securing the incentive did not need to be
learned. That is, a more direct measure of drive was obtained,
largely uncomplicated by the chance factors that always operate in
a learning situation of the maze type.

(6) The obstruction study involved ' * punishment ' ' (shock),
whereas the maze involved no punishment, but was a type of free
activity which seems to be natural to the rat.

APPENDICES

485

(7) Fewer chance factors enter into the obstruction test, since
such factors as illumination, sex drive, emotional disturbance, etc.,
could be more carefully controlled. The fact that only a single test
period involved but one incentive situation, while the maze involved
many incentive situations distributed over a much longer period of
time, also made possible better control of conditions in the ob-
struction experiment than in the maze experiment.

It is apparent from the differing nature of the methods employed
in this study that only general comparisons of the effect of delay

Fig. 13. Main Results in Obstruction and Maze Experiments.
The solid line represents the average number of crossings for the combined
(male and female) groups (Obstruction Method). The dash line represents the
average number of trials required to learn (Maze Experiment).

can be made. In order to compare the general trend of the results
in the experiments, curves were constructed on the basis of the
main results in each experiment as shown in Figure 13. Since no
comparison is possible beyond the 3-minute interval, the curve for
the maze was not extended beyond this point. It will be seen from
the curves that a delay of 1 minute markedly decreased the drive
and to much the same extent in both experiments. We do not know

486

APPENDICES

what effect shorter delays such as 15 seconds and 30 seconds might
have had in the maze-learning experiment. During the progress of
the maze experiment, no further decrease in the drive was obtained
with the longer delays of 3 minutes and 5 minutes, so it appeared
to be of more interest to use a longer interval of 7 minutes in an
attempt to determine whether the plateau effect was a real tendency
or not. It would have been interesting also to study the effect of
intervals shorter than 1 minute, but this was impossible without
undue extension of the experiment.

In general, the same tendency is apparent in both experiments;
namely, that even a short delay decreases the hunger drive, as
measured by the methods used. The longer delays studied had
little effect except that of a second drop in the curve of some mag-
nitude in the obstruction experiment. The study of still longer
periods of delay in the maze-learning experiment might reveal a
tendency to decreased drive in the learning scores similar to that
noted in the obstruction experiment. It is clear, however, that the
most important decrease in drive is induced by very short delays,
since the rate of decrease as between 15 seconds and 3 minutes in
the obstruction experiment is only about 50 per cent as great as
that between 0 and 15 seconds, even when no allowance is made
for the difference in increments of delay in the two cases (2| min-
utes in the former and 15 seconds in the latter).

From the standpoint of both the method and procedure, and the
results, it will be seen that the Obstruction Method of measuring
drive is more adequate than the learning method in many respects.
As above noted, the Obstruction Method is superior in that it fur-
nishes a more direct measure of drive, uncomplicated by the factor
of fixation ; the technique is better standardized and controlled than
the maze method and hence fewer chance factors enter into the
drive index. Furthermore, the time consumed in obtaining an in-
dex of drive is much less than half as long as that required in the
direct maze-learning method.

Bibliography

(1) Carr, H. 1914. Principles of selection in animal learning. Psychol.
Rev., 21, 157-65.

APPENDICES

487

(2) Cook, S. A. 1928. The effect of various temporal arrangements of
practice on the mastery of an animal maze of moderate complexity.
Arch. Psychol, 15, No. 98, 33 pp.

(3) Jenkins, M. 1928. The effect of segregation on the sex behavior of
the white rat as measured by the obstruction method. Genet. Psy-
chol. Monographs, 3, 455-571.

(4) Jenkins, T. N., L. H. Warner, and C. J. Warden. 1926. Stand-
ard apparatus for the study of animal motivation. J. Comp. Psy-
chol, 6, 361-82.

(5) Long, J. A., and H. M. Evans. 1922. The oestrous cycle in the rat
and its related phenomena. Mem. Univ. Calif., 6, 148 pp.

(6) Simmons, R. 1924. The relative effectiveness of certain incentives
in animal learning. Comp. Psychol Monographs, 2, No. 7, 79 pp.

(7) Thorndike, E. L. 1913. Educational psychology, II. New York,
Teachers College, 452 pp.

(8) Warden, C. J., and E. L. Haas. 1927. The effect of short intervals
of delay in feeding upon speed of maze learning. J. Comp. Psy-
chol, 7, 107-15.

(9) Warden, C. J., and H. W. Nissen. 1928. An experimental anal-
ysis of the obstruction method of measuring animal drives. J. Comp.
Psychol, 8, 325-42.

(10) Warner, L. H. 1928. A study of the hunger drive in the white rat
by means of the obstruction method. J. Comp. Psychol, 8, 273-99.

(11) 1927. A study of sex behavior in the white rat by means of

the obstruction method. Comp. Psychol Monographs, 4, No. 22,
68 pp.

(12) 1928. A study of the thirst drive in the white rat by means

of the obstruction method. J. Genet. Psychol, 35, 178-92.

(13) Warner, L. H., and C. J. Warden. 1927. The development of a
standardized animal maze. Arch. Psychol, 15, No. 92, 64 pp.

(14) Watson, J. B. 1917. The effect of delayed feeding upon learning.

Psychobiol, 1, 51-59.

4. THE EELATIVE VALUE OF REWARD AND PUN-
ISHMENT IN THE FORMATION OF A VISUAL
DISCRIMINATION HABIT IN THE
WHITE RAT 1

C. J. Warden and Mercy Aylesworth 2

One of the major problems within the general field of animal
motivation is that concerning the relative incentive value of re-
ward and punishment. The terms reward and punishment as here
used carry no subjective connotation whatsoever, but are used
merely as convenient classificatory terms for stimuli that normally
evoke positive and negative reactions respectively when employed
as incentives in experimental situations. The chief interest in this
topic centers around the relation of this factor to speed or efficiency
of habit formation, and more intensive work has been done on the
discrimination type of habit than upon either problem box or maze.

The present study was suggested by the investigation of Hoge
and Stocking (1) who found some evidence that punishment has a
higher incentive value, either alone or in combination with re-
ward, than reward alone in the building up of a simple visual dis-
crimination habit in the white rat. Their conclusion to this effect
was based upon very slight data, only two animals being tested in
each group. In one case correct responses were rewarded by a
morsel of food; in a second, correct responses went unrewarded
while wrong responses were followed by a shock, while in the third
instance a correct response was followed by food and an incorrect
response by a shock. They found that both rats in the reward-
punishment group perfected the habit — one in 490 trials and the
other in 550 trials, while one rat in the punishment group learned
in 550 trials and the other failed to learn in 620 trials, and neither

1 Reprinted from The Journal of Comparative Psychology, 1927, 7: 117-27.

2 The problem was suggested and the data gathered by the junior author;
the senior author collaborated in technical details and in writing this report.

488

APPENDICES

489

of the animals in the reward group had learned after 590 trials.
The problem involved a discrimination between a 2 c.p. and 16
c.p. light, the norm of mastery being 15 correct responses on each
of two successive days, and the apparatus a modification of the
Yerkes discrimination box. The shock used was furnished by a Por-
ter inductorium and applied in the usual manner.

The purpose of the present investigation was to repeat the work
of Hoge and Stocking, using larger groups and modifying the
technique in certain details. The main improvement in method,
perhaps, consisted in a more standardized shock than that used by
Hoge and Stocking, the mechanism for the control of which being
identical with that employed in connection with the Obstruction
box (2) of this laboratory. In the present work a current of 0.11
m.a. with a resistance of 1000 volts was used, and readings were
taken daily before beginning the taking of data. The resulting
shock was strong enough to cause the animals to lift their feet ener-
getically and sometimes squeal a little in scampering off the grills,
without causing them to be too inactive for our purposes after
getting off. We used a group of 10 white rats approximately 3
months of age in connection with each of the three incentive con-
ditions previously studied by Hoge and Stocking, i.e., reward, re-
ward-punishment, and punishment.

The apparatus employed was a modification of the Yerkes-
Watson design (see their description, Behavior Monographs No. 2)
and need not be figured in detail. The source of light was a 75-
watt Mazda Westinghouse bulb (120 volts) set 10 inches behind the
plate at an angle of 45 degrees, in such wise that the center of the
bulb coincided with the center of the circular opening (6 cm.) of
the plate. The distance from center to center of the openings on
either side of the stimulus box was 5J inches, and that from the
center of the opening to the floor of the control box was 4 inches. A
bulb was placed on either side of the box so that a shift from light
to dark could be made without disturbing the shutter, by the simple
expedient of pressing an electric button conveniently placed on the
outside of the stimulus box. The current was supplied to the two
bulbs through a common intake. The ground plan of the control
section of the apparatus in which the animal was placed for testing

490

APPENDICES

is shown in figure 1. The box is 37 inches long-, 29 inches wide, and
11^ inches deep (inside dimensions) and the partitions as figured
are drawn to scale. The electric grills were equipped with inde-
pendent switches so that either could be turned off at will by the
experimenter, stationed directly in front of the entrance compart-
ment. Since a dark room was not available, the control box was
entirely screened in by a hood of black cloth extending 24 inches
above the top of the box. Inside the hood and in the exact center

Fig. 1. Floor Plan of Control Box
E, entrance box; D, door leading into reaction compartment; X and Y,
electric plates; L, central partition separating X and Y ; A and B, position of
lights used as stimuli; F and G, doors used to restrain animal (F open and G
closed) ; H and I, food compartments; C, reaction compartment.

above the door, D, sl 25-watt blue Mazda lamp was placed and kept
burning constantly during the experiment. Small automobile mir-
rors were fastened to the top of the box above the grills, X and Y.
The animals were visible in the dim light at all times under this
arrangement and were observed through a peep hole in the hood
above the center of E.

The problem was to learn to discriminate between the stimulus
patches outlined by the circular openings in the plates, when one

APPENDICES

491

was illuminated from behind and the other not. The animals were
trained to the dark patch inasmuch as 8 of the 10 in each group
showed either a preference for the illuminated patch or no pref-
erence at all, the other two in each case showing only a slight pref-
erence for the dark patch. The preference series consisted of 10
trials per day for two days with food in both compartments (I and
H), regardless of later incentive conditions. As the data show, the
discrimination was much less difficult than that required in the
Hoge and Stocking experiment.

The procedure was kept as uniform for the three groups as
possible. The characteristic response of the animal under the differ-
ent conditions of motivation made necessary minor variations in
procedure as will appear from the description that follows. When
the response was correct, in the reward group, a nibble of milk-
soaked bread was allowed in the appropriate goal box ; when incor-
rect the animal was lifted from the goal box, which contained no
food, by the experimenter. Food was placed in neither goal com-
partment in the punishment series; when a correct response was
scored the animal merely escaped the shock and was removed from
the goal compartment as above. An incorrect response resulted in
a shock and, as a rule, the animal retreated from the grill and
remained in the reaction compartment, C, since there was no incen-
tive for it to continue on across the grill into the goal compartment
for that side. Accordingly the animals of this group were removed
from the reaction compartment after such retreat from the grill, or
from the appropriate goal compartment if they actually crossed the
grill instead of retreating. Contact with the grill beneath the light,
resulting in a shock, was thus the basis of scoring errors for this
group regardless of the nature of the more or less insignificant
response that followed this experience. The same procedure was
followed likewise when an animal of the reward-punishment group
made an error; it was removed from either position, the goal or
reaction compartment, depending upon whether it retreated from,
or crossed over the grill. In certain instances in both the reward
and reward-punishment groups the animals would dash around to
the correct goal compartment after making a wrong response before
the restraining door (F or G) could be closed. This occurred in

492

APPENDICES

approximately 5 per cent of the trials of the former, and 2 per cent
of the trials of the latter group.

The original plan was to give 15 trials per day and the order of
shifting the lights was arranged for such a schedule. This order
which was repeated over and over, without regard to the number of
trials given per day, was as follows for the dark stimulus patch:
left, right, right, left, right, left, right, right, left, left, left, right,
right, left, right. Five trials per day were given for the first 7 days,
ten trials per day thereafter until the 27th day, after which the 15
trial per day schedule was followed in the case of the reward group,
which showed little or no evidence of learning at this stage. The
reward-punishment group had completed the problem by the 15th
day ; the punishment group were so nearly finished on the 27th day
that they were continued on the 10 trial per day schedule. The
change to the 15 trial schedule on the 27th day for the reward group
was made in the hope that they might master the problem within the
limit of time available for the study.

During the first day 10 minutes were allowed for each trial pro-
vided the animal did not make a ' ' choice ' ' before. From the second
day onward this time was reduced to 5 minutes. Animals of the
reward group usually reacted within a few seconds. However, the
use of the shock in the other two groups induced hesitation and
inactivity to such an extent that many of the animals did not react
within the 5-minute period. In the reward-punishment group the
animals had to be removed without having reacted in 59 out of 760
trials (7.8 per cent) while in the punishment group the correspond-
ing value was 425 out of 1500 trials (28.3 per cent). The effective-
ness of the food in inducing an increased tendency to react promptly
in the former group is thus clearly indicated. Failure to react
within the 5-minute period was checked as an error, although the
data on this point are given separately from the errors, in table 1,
under the heading "no reaction."

The positive value of the reward incentive stimulus (food) in
bringing about activity, and especially the definite act of going
promptly to the goal was also very evident. When reward alone was
used the complete reaction of going to one or the other goal com-
partments occurred in practically every case. Naturally there was

TABLE 1

Showing the percentage of right, dwrong, and "no reaction" responses for

each series of trials

REWARD

PUNISHMENT

REWARD-PUNISHMENT

DAYS

Kignt

Wrong

Right

Wrong

"No re-
action"

Right

Wrong

"No re-
action"

1

OKJ

70

42

44

14

34

62

4

o
A

9ft

72

10

10

80

40

26

34

Q
O

3fi

fU

24

14

62

42

26

32

OA*

rift

44

18

38

62

18

20

0

3ft

62

46

18

36

68

20

12

0

32

68

56

14

30

64

30

6

7

44

56

52

8

40

84

14

2

o
o

42

58

57

11

32

84

15

1

Q

45

55

51

7

42

80

20

in

1U

47

53

55

10

35

95

5

1 1

n

43

57

59

6

35

86

12

2

19

47

53

56

4

40

87

10

3

to

48

52

54

6

40

90

10

14.
1£

45

55

71

9

20

100

10

54

46

67

12

21

100

1 A

47

53

80

10

10

17
1 /

48

52

86

6

8

lo

52

48

75

13

12

1Q

55

45

60

12

28

57

43

85

12

3

91

Ail

63

37

83

7

10

99

58

42

80

15

5

OQ
AO

52

48

80

10

10

94.

49

51

70

20

10

9|!i
AO

58

42

100

9fi
^O

51

49

70

30

97

55

45

50

40

10

9ft

57

43

50

30

20

9Q

57

43

80

20

ou

60

40

90

10

31

Ol

59

41

70

30

39

55

45

80

10

10

v\

Oo

57

43

100

34
0*t

56

44

100

oo

61

39

36

57

43

37

62

38

38

55

45

39

61

39

40

65

35

41

62

38

The term "no reaction" means that the animal did not make a "choice"
within the five-minute test period. Such failure to react did not occur in the
case of the reword group.

494

APPENDICES

little or no incentive for the animals of the punishment group to go
to the goal compartments. In 376 trials out of 1500 (25 per cent)
they were removed without having reached the goal — they reacted
to the grill-stimulus patch situation and stopped at that, although
they had been fed during the 20 trials of the preference series in
whichever goal compartment they happened to enter. This type of
response occurred in only 62 out of 760 trials (8 per cent) in the
case of the reward-punishment group which had, of course, a more
definite incentive to keep running till they reached the goal. When
both ' 1 no reaction ' ' and failures to go on to the goal after reacting
are combined, we find that in 63.3 per cent of the trials the punish-
ment group did not actually reach the goal while the corresponding
value for the punishment-reward group is 15.8 per cent.

Bread and milk constituted the main article of diet both before
and during the experiment. The animals of each group were allowed
to feed for 10 to 15 minutes (or until they were satiated) one hour
after the daily schedule of trials for the group had been com-
pleted. This interval had the effect of separating the ordinary
taking of nourishment from the incentive situation presented in the
apparatus. The groups were tested at the same time of day in so far
as the conditions of the experiment permitted, the work beginning
at 11 a.m. each day. Each trial was timed with a stop watch but
these data have not been included in this report since they appear
to add nothing of significance.

The results are given in table 1 and the graph of figure 2 in terms
of the percent of right, wrong, and "no reaction" responses per day
throughout the 41 day training period. At the end of this period
the reward group were scoring only about 60 per cent right
responses and had made little gain in the preceding 300 trials (20th
day onward). The punishment group made a perfect record on the
33rd day and the reward-punishment group on the 13th day, these
differences being so great that no question of interpretation arises.
The reward-punishment motivation is by far the best ; reward alone
is markedly inferior to punishment alone, just how much so we
cannot tell since the former group did not perfect the habit at all.
The rise in the error curve of the punishment group on the second
and third days was due to the large number of "no reaction"

APPENDICES

495

responses occurring on those days. The marked rise in this curve
on days 26-28 and on several days thereafter was due in large part
to the erratic behavior of rat no. 4. This rat after making almost
perfect records for several days suddenly took to jumping over the
grill on the right hand side, without regard to the stimuli, and this
habit had to be broken up before further progress could be made.

The data on number of trials required to form the habit are
given in table 2 for each animal of the three groups as well as the
averages for the groups. The records were checked against three
norms of mastery, i.e., 9 correct out of 10, 18 correct out of 20, and

496

APPENDICES

27 correct out of 30 responses. None of the animals of the reward
group reached the highest norm and only one (no. 8) reached the
norm of 18 correct responses out of 20 consecutive trials, doing so
only after 265 trials. The data for the lowest norm only, therefore,
are included in the table for the reward group. A comparison of
the averages for the three groups is thus possible only on the
basis of the data of norm A. The average number of trials to learn
for the groups, when so checked, is as follows: reward, 293.5,

TABLE 2

Showing number of trials required to learn under three norms of mastery

(A, B, C)

REWARD

PUNISHMENT

REWARD-PUNISHMENT

RAT

Norm

Norm A

Norm B

Norm C

Norm A

Norm B

Norm C

1

439

36

89

126

39

64

72

2

242

50

60

68

32

37

47

3

173

47

137

145

46

52

62

4

362

60

125

. 301

23

99

108

5

145

22

141

174

13

49

57

6

446

32

42

70

24

75

84

7

276

81

134

170

39

56

59

8

255

127

136

147

19

40

45

9

279

35

66

75

49

56

60

10

318

72

114

183

44

69

79

Average

293.5

56.2

104.4

145.9

32.8

59.7

67.3

Norm A, 9 correct out of 10 trials ; norm B, 18 correct out of 20 trials ; norm
C, 27 correct out of 30 trials. None of the rats of the reward group reached
norm C and only one (no. 8, after 265 trials) reached norm B. When rat no. 4
of the punishment group is not included (see text) the average for the group
is 128.6 instead of 145.9 as above.

punishment, 56.2, and reward-punishment, 32.8. It is noteworthy
that the number of trials is more than doubled in the case of the two
latter groups when the norm is raised to 27 correct responses out of
30 trials (norm C).

It is evident that norm A is not high enough to insure depend-
able results in work of this sort. It does not rule out the possibility
of apparent success in discrimination which is, in fact, only appar-
ent and due to chance factors. For, as appears from table 1, the
reward group as a whole never averaged on any given day higher
than 65 per cent correct responses, and yet each of them made a
record of 9 correct out of 10 trials at least once during the training

APPENDICES

497

period. That this record was due to chance factors mainly is shown
by the fact that only one animal of the group (rat number 8) was
able to reach norm B and none norm C, although in many cases
the requirement of norm A had been met before one-half of the
training series (223 trials) was completed. The use of a high norm
of mastery appears to be more necessary in discrimination work
than in habit formation of the maze type (3). The order of shifting
the lights, and the greater likelihood of the setting up of position
habits in a simple situation of this sort, probably operate as impor-
tant factors in this connection.

The general principle suggested by the work of Hoge and Stock-
ing seems to be fully established by the present work, for this type
of problem. This conclusion should, however, be limited to simple
and easy discrimination, inasmuch as several investigators (Yerkes
and Dodson, Dodson, Cole, etc.) have shown that in more difficult
discrimination too much punishment (shock) must not be employed
if the best results would be obtained. Before an unqualified general-
ization can be made concerning the relative incentive value of
reward, punishment, and reward-punishment these must be studied
in relation to a series of discriminations ranging between wide
extremes of difficultness.

A criticism of the control box used by ourselves and others, in so
far as its use in studying the relative value of incentives is con-
cerned, should be made by way of offering a suggestion regarding
further work on this topic. The shock is usually administered by
means of a grill placed directly in front of the stimulus patches
while the food or reward is placed at some distance from the patches
and can be reached only by making a turn after crossing the grill.
Doubtless the value of the punishment incentive is greatly increased
by the fact that it tends to stop the animal directly in front of the
discrimination stimuli while the reward incentive (food) does not.
In order to obtain strictly comparable data the reward and punish-
ment factors should be administered at points equidistant from the
stimulus patches. That is, either the food (or other reward object)
should be located in the front end of the control box near the grill,
or perhaps better, the grill should be placed in, or near the goal
compartments rather than directly in front of the stimulus patches.

498

APPENDICES

So far as we know, this matter of the identity of location of incen-
tive stimuli has not been adequately controlled in any work so far
done on the relative value of reward and punishment in this type
of habit formation.

Bibliography

(1) Hoge, U. A., and R. J. Stocking. 1912. A note on the relative
value of reward and punishment as motives. Jour. Anim. Behav., 2,
43-50.

(2) Jenkins, T. N., L. H. Warner, and C. J. Warden. 1926. Standard
apparatus for the study of animal motivation. Jour. Comp. Phychol.

(3) Warden, C. J. 1926. A comparison of different norms of mastery
in animal maze learning. Jour. Comp. Psychol., 159-79.

INDEX

INDEX

Columbia Obstruction Method: appa-
ratus, description of, 17-30, 140-42;
development of, 4, 9, 10-14, 17, 406-
8; effect of incentive in, 38, 42;
effect of practice on, 42, 47; influ-
ence of shock in, 36-38; scores, ef-
fect of re-test on, 47-49; standard
shock employed in, 36, 142

Comparison of drives: basis of, 373-
82; rank order obtained, 396-98;
sex factor in, 393 ; use of maximum
scores in, 382

Delayed feeding: effect on hunger
drive, 94-95; effect on maze-learn-
ing scores, 493 ; previous work of
Warden and Haas on, 450-54; prev-
ious work of Watson on, 447, 455-57

Diet, McCollum's standard, 139

Discrimination test: incentive condi-
tions used in, 489 ; procedure em-
ployed in testing, 489-92; relative
value of reward, punishment, and
reward-punishment, 496-98

Exploratory drive: controls employed
in testing, 353 ; correlation with
other drives, 393 ; physiological
basis of, 364-66; procedure used in
testing, 356-60 ; summary of test
results of, 366; term denned, 353,
354-55

Gonadectomy: control operation em-
ployed, 307; effect of, on extinction
of sex drive, 313 ; procedure used
in test, 286-88, 296-97

Homosexuality: evidences of, in fe-
males, 259-60; in males, 258-59;
term defined, 184

Hunger drive: controls employed in
testing, 53-55; correlation with
other drives, 74-78, 393; procedure
used in testing, 57-63; sex differ-
ences in, 70-71; starvation intervals

investigated, 63-70; work of others
on, 72-79

Individual scores, rank order of, on
re-test in obstruction apparatus, 47

Injections: or chic extract, effect of,
on castrated males, 318, 320; on
senile males, 318-19, on vasotomized
males, 319-21; placental extract, ef-
fect of, on females, 320-21, on
males, 320-21

Lubbock, Sir John, obstacle principle
of, 8

Maternal drive: activity manifested,
345; age as factor in, 346; condi-
tions tested, 339 ; controls employed
in testing, 333 ; correlation with
other drives, 349 ; factors that exert
influence on, 336; primiparity, in-
fluence on, 349; procedure in test-
ing, 336-39

Methods of testing motivation: ac-
tivity cage, 5; choice, 6, 17, 31-33;
learning, 6-7, 484-85; resistance, 8,
9; revolving wheel, 4, 129. For use
of obstruction in, see Columbia Ob-
struction method

Moss, F. A., use of resistance method
by, 8, 9, 22, 27-29, 79, 404-8

Morgan, J. J. B., resistance method
of, 8

Motivation: defined, 15-16; drive and
incentive factors distinguished, 14,
15; incentive -index defined, 55, 208-
9; normal drives defined, 371; ob-
stacle principle, as related to, 5, 8

Oestrus cycle, table showing subdivi-
sions of, 133

Papanicolaou, G. N., examination of
vaginal smears by, 118

Practice: effect of, in obstruction ap-
paratus, 42-47

Reactions to obstruction: approaches,

501

502

INDEX

contacts, and crossings, defined,
146

Richter, C. P., activity cage method
of, 4, 5, 405

Segregation: effect of, on homosexua]
tendencies, 182-85, 223; influence
of, on female sex drive, 259-60, on
male sex drive, 258-59 ; periods of,
investigated, 188-89, 196; proce-
dure used in testing, 189-97

Sex drive: controls employed in test-
ing, 117-19; correlation of, with
other drives, 393; homosexual ten-
dencies in, 138, 182-85, 223; inter-
vals of deprivation tested in, 150,
188-89, 196; physiological basis of,
in female, 120-35, 159 ff., in male,
135, 265-68; procedure used in
testing, 136-40, 142-49; sex differ-
ences in, 171, 172, 174, 175; sum-
mary of test results of, on normal
females, 170-74, on normal males,
159, 173; work of others on, 6, 8-9,
17, 180 ff. See also Gonadectomy,
Injections, Segregation, Vasotomy

Simmons, R, learning method of, 7,
79, 404, 456-57

Slonaker, J. R., revolving wheel meth-
od of, 4

Standard test conditions: animals,
age, and strain factor in, 373-74;
apparatus, 373; incentives in, 374-
76; procedure in, 376-82; scores of,
382

Stone, C. P., work of, on effect of
sex isolation, 180-81

Thirst drive: controls employed in
testing, 99; correlation with other
drives, 393; procedure used in test-
ing, 104; sex differences in, 113;
summary of test results of, 113;
work of others on, 100-2

Tsai, C, choice method of, 6, 18, 79,
404

Vasotomy, effect of, on sex drive, 320-
21

Waller, O. D., work of, on stimula-
tion value of electric current, 25

Watson, J. B., work of, on effect of
delayed feeding on learning, 447,
455-57

COLUMBIA UNIVERSITY PRESS
Columbia University
New York

FOREIGN AGENT

OXFORD UNIVERSITY PRESS
Humphrey Milford
Amen House, London, E. C.

Date Due

— B — «*-

«#

!

Apr 1 91

&

STLAC1E

Annas

mm

—

CAT. NO. 23 233

PRINTED IN U.S.A.

mm

W UNIVERSITY
LIBRARIES

BOOK CARD

YOU ARE RESPONSIBLE
FOR THE LOSS OF THIS CARD

413542