Full text of "Physiological mammalogy"

See other formats

PHYSIOLOGICAL MAMMALOGY

VOLUME I

Mammalian Populations

Physiological Mammalogy

VOLUME I
Mammalian Populations

CONTRIBUTIONS BY

JOHN B. CALHOUN J. J. CHRISTIAN

VOLUME II
Mammalian Reactions to Stressful Environments

PHYSIOLOGICAL
MAMMALOGY

EDITED BY

WILLIAM V. MAYER

Department of Biology, Wayne State University, Detroit, Michigan

and
RICHARD G. VAN GELDER

Department of Mammalogy, The American Museum of Natural History
New York, New York

VOLUME I
Mammalian Populations

1963

ACADEMIC PRESS— New York and London

\0 PART OF THIS BOOK MAY BE REPRODUCED IN ANY FORM,
BY PHOTOSTAT, MICROFILM, OR ANY OTHER MEANS, WITHOUT
WRITTEN PERMISSION FROM THE PUBLISHERS.

ACADEiMIC PRESS INC.

1 1 1 Fifth Avoniie, New York o. New York

United Kingdom Edition published by
ACADEMIC PRESS INC. (LONDON) LTD.
Berkeley' Square House, London W.l

Library of Congress Card Number: 63-23422

PRINTED IN THE UNITED ST.\TES OF AMERICA

Contributors to Volume I

JOHN B. CALHOUN

Laboratory of Psychology

National Institute of Mental Health

Bethesda, Maryland

J. J. CHRISTIAN

Division of Endocrinology and Reproduction

Research Laboratories, Albert Einstein Medical Center

Philadelphia, Pennsylvania

PREFACE

The field of mammalogy has, until very recently, been largely concerned
with morphology and systematics; investigators have done little to bring
together experimental work on the widely divergent mammalian types of
which we have knowledge. The treatise "Physiological Mammalogy" was,
therefore, conceived as a device to bring together the existing knowledge
of an experimental nature on those animals usually regarded as "wild."
Approximately fifteen themes, on which a considerable body of evidence
has been accumulated over the years, provide the central organizing core
around which various authors have been asked to make their contribu-
tions. The amount of material available on the non-classic laboratory
animals has been difficult to synthesize ; but the experts who were asked to
contribute to the planned original single volume have proven so knowl-
edgeable about the areas of their specialities and have made such a truly
impressive survey of the experimental literature on their topics that it
has become necessary to expand the work from the original single volume
of relatively short essays to a minimum of three and perhaps even more
additional volumes.

The plan of this work is such that anyone working with mammals will
find it an indispensable reference. It is particularly valuable to experi-
mentalists working with mammals in the areas of physiology, mammalogy,
and ecolog3^ Within these volumes will be found comprehensive essays on
specific topics in physiological mammalogy, as well as a cogent analysis of
the experimental field developing both what is known and what needs yet
to be done. The research worker will find, in perusing these pages, chal-
lenging obser\'ations to which he might well address future researches.
The student will find reference material and previous observations which
will make these volumes useful as a baseline from which additional studies
can be begun. ]\Iost of the authors have, in addition to providing a com-
prehensive review article, intercalated their own observations and com-
mentary to the point where the articles themselves constitute original
contributions to the field. This treatment provides a comprehensive analy-
sis of the over-all specified topic.

These volumes will provide the investigator with information that
will enable him to choose experimental animals previously little used in the
laboratory because their specific physiological properties and behavior
were formerly not well known. As a compendium on physiology of nor-
mally non-laboratory mammals, this treatise will be of value to anyone
who works with mammals and to any experimental animal biologist.

viii Preface

The first volume consists of two comprehensive articles dealing with the
physiology of populations. Calhoun's "The Social Use of Space" presents
many interesting new ideas on the behavior of animals in populations and
the effects of grouping of individuals upon the physiology of the organism.
Where many workers have thought of the experimental animal only as an
individual apparently divorced from his environment and other members
of the species, Calhoun points out the fallacy of this view in providing
experimental data that demonstrate the effects of numbers of individuals
on the behavior of each individual. Christian's article on population growth
treats the problem largely from an ecological viewpoint in dealing with
limiting factors of natural populations and population interrelationships.

Volume II will consist of three contributions dealing with natural
populations and their adaptations to stressful environments. Dr. Charles
Kayser treats the mammalian phenomenon of hibernation as a mechanism
for avoiding periods of unfavorable environment, and Dr. Robert Chew
deals with water balance in desert rodents. Inasmuch as reproduction is
considerably affected by the environment, it too can be considered a process
modified by environmental stresses, and is discussed in the second volume.
Subsequent volumes will include articles on such topics as temperature and
metabolism, physiological genetics, photoperiod, and orientation by echo-
location. The Editors will conclude the series with a summary article on
the phylogeny of physiology.

The Editors feel particularly fortunate in that they have received fine
cooperation from outstanding authorities in the specific subject matter
topics covered. It is the caliber of the individual author on which these
volumes base their contribution to science; and while the Editors assume
any responsibility for defects of organization or inadvertent errors, the
credit for the success of the volumes, as a whole, rests on the indi^'idual
contributors.

October, 1963 William V. Mayer

Richard G. Van Gelder

INTRODUCTION

The development of the biological sciences has been largely through
the observational method. In the nineteenth century descriptive biologists
came to occupy a preeminent and leading position in the field of biology.
With the advent of the twentieth century, however, biology began to
emphasize the experimental; and today, biology is an experimental science
in practically all of its branches. However, experimentation in animal
biology has very largely concentrated on a relatively few species which
have come to be classic. These animals were selected because of their
tractability and availability, and because of the fund of existing knowledge
about them. Therefore, animal experimental biology has been largely
oriented around the dog, cat, rabbit, mouse, guinea pig, and white rat.
Despite the fact that again and again it has become obvious that there
are wide varieties of metabolic patterns and inter-specific difl^erences in
physiology as well as morphology, the bulk of experimentation continues
to be done with relatively few classic types. Despite this concentration of
effort, over the years numerous researchers have investigated the use of
different animals for experimental purposes, including mammals ranging
from aardvarks to zebras.

The writings of these investigators, however, are scattered widely both
in literature and in time, making it diflScult for an experimental biologist
to draw on the fund of already available knowledge about animals other
than those most frequently used. The basic thesis guiding the preparation
of "Physiological Mammalogy" has been to make available to the experi-
mental biologist the wealth of data in the work of widely geographically
scattered research scientists of diverse experimental interests.

:S\ca7

TABLE OF CONTENTS

LIST OF CONTRIBUTORS V

PREFACE Vii

INTRODUCTION ix

CHAPTER 1

The Social Use of Space
John B. Calhoun

I. Introduction 2

II. The Bivariate Normal Type of Home Range 4

III. Behavioral Origins of the Bivariate Normal Type of Home Range 8

IV. Use of a Two-Dimensional Field 19

V. Summary of the Concept of Home Range 25

VI. Continuous Removal Trapping of Small Mammals 26

VII. Toward a General Theorj^ of Interspecific and Intraspecific Use

of Space 34

VIII. Interpretations of Observed Data Derived from Removal Trapping

of Small Mammals 52

IX. A Theoretical Conceptualization of the Evolution of a Social

Hierarchy among Species in the Utilization of Space 70

X. Psychological Dominance as the Primary Component of the Niche 77

XI. An Induced Invasion 80

XII . Derivation of Compact Colonies from Constellations 86

XIII. A Formulation of Group Dynamics 101

XIV. Consequences and Examples of Social Interaction Systems 148

XV. Conclusion 184

References 185

xii Table of Contents

CHAPTER 2

Endocrine Adaptive Mechanisms and the
Physiologic Regulation of Population Growth

J. J. Christian

General Introduction 189

PART 1. THE ENDOCRINE ADAPTIVE MECHANISMS

I . Introduction 191

II. The Endocrine Glands of Adaptation 192

PART 2. PHYSIOLOGIC ADAPTION AND MAMMALIAN POPULATIONS

I. Introduction 261

II. Endocrine Responses to Social Pressures and to Population Density 263

III. Conclusion 325

References 328

AUTHOR INDEX 355

SUBJECT INDEX 363

The Social Use of Space

JOHN B. CALHOUN

Laboratory of Psydwlogy, National Institute of Mental Health, Bethesda, Maryland

TABLE OF CONTENTS

I. Introduction 2

II. The Bivariate Normal Type of Home Range 4

III. Behavioral Origins of the Bivariate Normal Type of Home Range. _ 8
A. Activity in a One-Dimensional Habitat 8

IV. Use of a Two-Dimensional Field 19

A. Theoretical Origin 19

B. Travel-Path Home Range 24

V. Summary of the Concept of Home Range 25

VI. Continuous Removal Trapping of Small Mammals 26

A. Rich Lake Island, New York, 1952, Sixty-Day Removal Study —

Data Contributed by William L. Webb 27

B. Chadwick Woods, Montgomery County, Maryland, Removal

Study, 1958-1959— Data Contributed by Kyle R. Berbehenn 29

C. Comparative Catches, Huntington Wildlife Forest, 1952-1953 —

Data Contributed by Earl F. Patric and WilUam L. Webb_ . . 29

D. Comparative Catches, Huntington Wildlife Forest, 1951 — Data

Contributed by William L. Webb 30

E. Comparative Catches in Maine (1950) and Maryland (1953) ... 31

F. Comparative Catches of Peromyscus and Clethrionomys 32

VII. Toward a General Theory of Interspecific and Intraspecific Use of

Space 34

A. A Two-Species System 36

B. The Nature of the Inhibitory Influence 38

C. The Learning of Signals 39

D. The Distance between Neighbors of the Same Species 42

E. Methods of Calculating Data Relative to the Distance between

Neighbors 44

F. Further Comment on the Impact of All Individuals on the

Environment 44

G. Contacting Neighbors 45

H. Sign Field of All Neighbors 47

I. Signal Field of Neighbors 47

J. Hum Field 50

K. General Conclusion Concerning the Distance between Neighbors 50

L. The Number of Neighbors Perceived 51

John B. Calhoun

VIII. Interpretations of Observed Data Derived from Removal Trapping

of Small Mammals 52

A. The Relationship between Two Dominant Species 52

B. Removal Captures of Socially Dominant Species 55

C. Constellation Formation — An I ntraspecific Phenomenon 57

D. Expected Variability in the Number of Individuals Forming

Constellations 62

E . Social Rank and Intraspecific Associations 64

F. The Instability of Social Relations 67

IX. A Theoretical Conceptualization of the Evolution of a Social

Hierarchy among Species in the Utilization of Space 70

X. Psycholo.2;ical Dominance as the Primary Component of the Niche. . 77

XI. An Induced Invasion 80

XII. Derivation of Compact Colonies from Constellations 86

A. Compact Colony Formation in the Norway Rat 87

B. Howler Monkeys, a Compact Colony Living Species 90

C. Behavioral Sink D;>velopment by the Norway Rat 92

D. Yarding by Deer in Northern Wisconsin 95

E. Concerning Basic Numl;ers, A^d, for Man 97

XIII. A Formulation of Group Dyi amies 101

A. The Model of Social Intera.'tion 101

B. Basic Processes Involved in Social Interaction 116

XIV. Consequences and Examples of Social Interaction Systems 148

A. Velocity Reduction in a Hierarchy of Mice 148

B. The Choosing of a Partner 154

C. The Response-Evoking Capacity Circumplex 155

D. Conformity, Withdrawal, and Creativity 162

E. Velocity and Home Range 164

F. Velocity in High-Density Rat Societies 168

G. Exploratory Behavior 175

XV. Conclusion 184

References 185

I. Introduction

During recent years many investigators have considered the physiological
consequences to the individual of altering the size of the group of which it
is a member. Inherent in such studies is the assumption that for a particular
species there is some optimum group size, above or below which the altered
frecjuency or type of interactions are either stressful or fail to elicit optimum
physiological states. My purpose wall be to develop formulations con-
cerning the social use of space to determine whether there might be certain
optimum group sizes.

I shall consider only indirectly the physiology of individuals. Instead,
emphasis will be upon relationships which determine the "physiology" of
the community. The basic particle is the individual mammal. In any total

1. The Social Use of Space 3

assembly of such particles inhabiting a particular environment, taxonomic
categories, such as species and genera, represent general classes of particles.
Although several classes may share certain characteristics, each possesses
characteristics peculiar to itself, which on the average differentiate it from
all other classes. Furthermore, the properties of any particular particle
may change through time as a function of maturation and experience.
Such changes are reflected in the individual's internal milieu. Only at this
level are we concerned with physiology in its classical sense.

A social system consists of particles moving through space and time. In
the course of evolution and maturation, such particles may develop en-
hanced capacities for affecting others and, in turn, for being more affected
by them. This inquiry seeks that essential nature of these particles which
influences the course of social evolution. It concludes that each category of
particles which we designate as a species must develop a basic group size,
designated as its basic number, Nb. Then, assuming the validity of the
concept of an Nb, the inquiry is extended to encompass the following
questions :

1. How do changes in TV from Nb alter the social system?

2. Given a species, whose groups are of the optimum Nb type, how do
changes in the environment or in the nature of the particles affect
the social system?

3. What are the basic organizational and interactional properties of the
group?

I found very few prior studies in the literature suitable for guiding me
in developing formulations adequate to satisfy these objectives. Therefore,
this discourse will be neither a review of the literature nor even an adequate
presentation of empirical data. Instead, the major theme concerns develop-
ment of a logically sound, theoretical framework of processes underlying
social phenomena.

In the course of this development some empirical data will be presented.
These data are not meant to prove the correctness of the formulations.
They are presented simply as background for the origin of ideas or to show
that there at least exist some data which are harmonious with the developed
concepts. Much of such data is presented here for the first time. I am in-
debted to many persons for their permission to cite such original data.

I have attempted to place many of the concepts in sufficiently precise
mathematical form to permit a better understanding of the dynamics in-
volved. My inadequacy in the field of mathematics has been buffered by
the advice of several competent mathematicians : James U. Casby, Murray
Eden, Samuel W. Greenhouse, Seymour Geisser, Clifford Patlak, and

4 John B. Calhoun

John Gilbert. However, I assume full responsibility for any errors, in-
adequate presentation, or overextension from their initial guidance.

I have found this effort a rewarding one for the development of insight
into complex social systems, and I can only hope that in some small meas-
ure it may serve as a bridge for others in their design of experiments or in
their further theorizing.

II. The Bivariate Normal Type of Home Range

Home range denotes the area covered by an individual in its day-to-
day activities. Field studies of many species of mammals have revealed
that each individual customarily stays within a restricted area for long
periods. The individual utilizes the center of such an area most inten-
sively. With increasing radial distance from this home range center (HRC)
the relative frequency of visitation per unit of area decreases. Calhoun
and Casby (1958) found that the bivariate normal distribution function
adequately describes home range. The following is a summary of their
analyses.

In home range studies, "density function" is a mathematical expression
representing the probability of an animal being present in some arbitrarily
small area. Three assumptions are made :

The home range is fixed. In other words, the statistics of the home range
are stationary or time independent.

There is a true center of activity although the apparent center, the mean
coordinate point of capture, of activity may deviate from it.

The probability of an animal being in a unit of area decreases with in-
creasing distances from the true center of activity. This and the second
assumption suggest a bivariate normal distribution of the density function :

f{x, 7j) dxdy = — — - exp [- (a:^ + if)/2a^'] dxdy (1)

where o- is the standard deviation of the distances in the x and y direction
and is assumed to be equal for both, and x and y are measured from their
respective means. This density function may be used to represent the
percentage of time spent in the area dxdy located at the Cartesian coordi-
nates X, y, or in polar coordinates:

/(/•, 6) rdedr = — — exp (-r'^/'Ia^) rdddr (2)

zcrV

Here, the area rdddr is determined by r.

The density function in terms of the Cartesian coordinates is more

1. The Social Use of Space

meaningful from an ecological standpoint because it states in comparative
terms the amount of time spent by an animal in a small standard area at
any position in the home range. However, for the initial mathematical
manipulation, it was found more convenient to express the density func-
tion in terms of polar coordinates. Then the probability of finding the
animal between the radii r and r -\- dr about the true center of the home
range is:

f{r)dr = — - exp ( — r~/2a'^) rdr

If Eq. (3) is integrated over the range 0 to o- we have

r 2r

/ T- exp (-rV2(r2) rfr = 1 - g-'/^ = 0.3940

'o -^0-

(3)

(4)

In the above eciuations a, the standard deviation of the normal distribu-
tion function, is the value of a radius within w^hich the probability of the
animal being present is 39.4%, if its movements can be described by a
bivariate normal density function.

Similarly, integrating Eq. (3) over the range 0 to 2o- gives

1 - e-4/2 ^ 0.8645 (5)

Similarly, integrating Eq. (3) over the range 0 to 3o- gives

1 _ p-9/2 = 0.

f6)

The above sigma thus delineates a single distance term by which home
range may be described. The term "home range sigma" will be so utilized
in following sections.

Although this sigma may be calculated from a series of coordinate points
of capture by equations presented in the original paper, use of recapture
radii provide a more direct means, adequate for most purposes. Calculate
the mean coordinate point of capture, the approximate home range center.
Then on a large scale grid map of the study area measure recapture radii,
r, from this mean coordinate point of capture. Unbiased estimates of sigma,
s and Si may be calculated by the following equations:

s =

&i =

_2(A^ - n).

Ki
3=1

_2(/:.- 1)

(7)

(8)

6 John B. Calhoun

where :

s = unbiased estimate of the home range sigma for all the animals

in a sample
St = unbiased estimate of the home range sigma for any particular
animal
Ki = number of captures of ith animal
n = number of animals

N = total captures = ^ Ki

ij = jih observation of the iih. animal

A detailed analysis of the home range for 25 male harvest mice
(Reithrodontomys) on which there were 10 to 24 captures each indicated
that there was a significant variation of sigma among animals. In other
words, some animals had significantly larger home ranges than others.
Therefore, in order to compare the observed recapture radii with the
theoretical (Table II in Calhoun and Casby, 1958), each recapture radius
was normalized into a standard measure denoted by Z in which the home
range sigma for each animal was assigned a value of 1.0 and all recapture
radii expressed as a proportion of this.

As may be seen from Fig. 1, the theoretical closely approximated the
observed. Although this detailed analysis has been applied only to this one
species, it shall be assumed for the purpose of developing further formula-
tion that the bivariate normal distribution function adequately describes
fixed home ranges of other species.

Comparison of observed and theoretical distribution of home range
radii required viewing home range as a probability of capture which changes
with radial distance from the home range center. Bands of equal width
increase in area with radial distance from the home range center, while
probability of capture per unit area decreases with increase in radial dis-
tance. Interaction of these two factors results in more captures at about
one sigma from the home range center than at any other distance (Fig. 1).

However, the ecologically important aspect of the bivariate normal dis-
tribution as an expression of home range is the relative probability that an
animal will be in a unit of area with respect to the radial distance of that
unit area from the home range center (Fig. 2). For any given sigma char-
acterizing a particular species, its density function in terms of area curve
may be obtained by multiplying the relative sigma value on the abscissa
by the observed sigma and dividing the density function values on the
ordinate by the square of the observed sigma.

1. The Social Use of Space

70 1 r

Rodius from center of home range

Fig. 1. Observed (histogram) and theoretical distribution of 348 recapture radii
(Z) of 25 male harvest mice from the center of their home range. Z here represents a
normalized measure of the bivariate normal home range sigma.

0 .6 1.2 1.8 2.4 3.0

Radius from center of home ronge in units of cr

3.6

Fig. 2. Cross section of the density function of home range in terms of area. Rotation
of this curve about its axis reveals the mountain-shaped topography of home range.

8 John B. Calhoun

III. Behavioral Origins of the Bivariate Normal Type of Home Range

The fact that a particular equation happens to describe home range
enables derivation of several principles regarding the use of space by an
entire community. Discussion of these principles follows in Section VII.
However, as a background to this discussion it will be advantageous to
seek an understanding of the biological basis for the bivariate normal type
of home range.

A. Activity in a One-Dimensional Habitat

Admittedly, animals rarely live in essentially one-dimensional environ-
ments. However, I suspected that if animals were placed in such environ-
ments certain regularities of behavior might be revealed which would pro-
vide insight into their use of two-dimensional environments. To this end,
four 14-foot long alleys were constructed. Each had a channel 8 inches
wide. Each 8 X 12-inch segment of the floor was so suspended that when
a domesticated Norway rat, used as a subject, stood on such a segment a
microswitch closed. This closure initiated a signal such that the exact
position at every point in time was recorded on a recording oscillograph. A
partition between the first and second treadles formed a home compart-
ment. A 3 X 3-inch opening through this partition provided access to the
rest of the alley. Food and water placed in this compartment further en-
hanced the role of this compartment as a "home." A ground glass plate,
through which shone the light from a 100-watt lamp, formed the opposite
end of the alley. This light served to concentrate the activities of the rats
emanating from the home compartment. It was as if every foot of the alley
were several feet long. Details of the effect of varying light intensity at the
end of the alley on explorations will be presented elsewhere. This ap-
paratus is referred to as the Ferguson Activity Alley.

Suffice it to consider the results from 73 rats, each run for 72 hours in
the alley. Each rat made from 10 to 30 excursions out into the alley each
night. Despite the presence of the bright light at the end of the alley, one-
fourth of the trips terminated at the end of the alley. In other words, the
end of the alley formed a barrier. Most of the trips thus terminated at the
end of the alley presumably would have represented trips of greater length
had the alley only been longer. Most of the time a rat would go out to some
intermediate distance, stop momentarily, and then turn around and go
directly back home. Occasionally, a rat would wander back and forth from
the point of initial termination. All trips with such vacillations and those
ending at the barrier were excluded from the initial analysis.

1. The Social Use of Space 9

1. Trip-Terminations in an Unstructured One-Dimensional
Habitat

The initial investigation focused upon examination of the distances from
the home compartment at which nonvacillating trips terminated in the
unstructured alley. "Unstructured" denotes the absence along the alley
of any stimuli likely to elicit responses and so induce a rat to stop. Rats
were placed in the alley during the middle of the afternoon, a time of
minimal activity within their normal 24-hour rhythm of activity. And yet
when placed in the alley every rat exhibited a 2- to 3-hour period of hyper-
activity. Further details of the decay curve of this hyperactivity are dis-
cussed in Section III, A, 3.

For 73 rats complete records (Table la) were available for all trip-

Tablk la
Number of Trips Terminated with Reference to Distance from Home

Unstructured alley

Structured alley

T^' A

feet of

During first

During next 3

Strips of

Pellets

termination

2.5 hours

6 A.M.-6 P.M.

periods

paper

of food

274

558

1053

955

204

407

839

643

150

315

723

509

104

277

302

375

294

194

288

219

230

258

139

151

163

115

149

137

112

119

113

104

1112"

2535"

3810

3551

Barrier and vacillat-

502

1458

59*

127*

ing trips

Total trips

1614

4023

3869

3678

Trips/rat/hour

8.844

1..531

" Only nonvacillating trips included.

* Vacillating trips and trips at whose end an object was transported home are not
included since the termination of trips determined was solely by the number of tran.s-
I)ortt'd objects.

John B. Calhoun

terminations during both the initial 150 minutes of hyperactivity and the
following three, 6 p.m. to 6 a.m., 12-hour periods of normal heightened
nocturnal activity. The frequency, y, of terminating trips as a function of
distance, x, from home is described by the equation :

y = exp (o 4- bx) (9)

where h is the slope. The slope for trip-terminations during the hyper-
active period, 6i, is —0.2099; while 62, the slope for trip-terminations during
the 36 hours of normal nocturnal activity, is —0.1924. The t test

61 — 62

= -0.682

VVar. (61 - 62)

has a p value of 0.051 which indicates that slopes 61 and 62 do not differ
significantly.

Therefore, it is concluded that the neural mechanism producing termina-
tion of trips is unaltered by the nearly sixfold increase in the incidence of
initiating trips accompanying initial exposure to a strange environment.
For this reason, the mean slope of —0.20115, i.e., (61 + 62) /2, was fitted to
both these sets of data shown in Fig. 3 as trips per rat per hour for compara-
tive purposes to emphasize the hyperactivity of initial exposure to a strange
environment.

2.0

1.0

.02

"eMOTlONAL "
I si. 2.5 hours
of Hyperaclivily

TERMINATION OF TRIP IN FEET

Fig. 3. Frequency of terminating trips at successive distances from the home com-
partment in the unstructured Ferguson Alley. See Table la.

1. The Social Use of Space 11

2. The Role of a Structured Environment on the Termination
OF Trips

Natural habitats possess structures which elicit responses. Items of food
and nesting material represent structures normally causing animals such
as rats to terminate trips. When such items are transported home the trip
resembles the nonvacillating ones in the one-dimensional alley in the sense
that there is a direct outward phase, terminated by the object being picked
up, followed by a direct homeward trip transporting the item. In order to
explore the effect of such structuring in the one-dimensional habitat upon
termination of trips, one of two procedures was followed; At each one-
foot interval from home along the alley, there was placed a pad of paper
strips or an open hopper of food pellets. During any particular rat's stay
of 3-12 days in the alley, only nesting material or only food pellets were
available. Periodic replenishment of each source ensured a continuous
supply at each distance. Nevertheless, the rats removed items from each
distance (Table la) even though this necessitated passing by opportunities
to respond while on the outward journey. Each item removed at a particular
distance from home is considered to indicate a trip-termination at that
distance. Examination of the oscillograph record confirmed this inter-
pretation.

The frequency of termination of such trips as a function of distance is
also described by the equation, y = exp (a -\- bx) . The slope for trips termi-
nated by picking up paper strips, 63, is —0.3027; while 64, the slope relating
to securing food pellets, is —0.2481. The t test,

63 — hi

= -2.128

VVar. (63 - 64)

has a p value between 0.05 and 0.01 which indicates a statistically signifi-
cant difference between these two slopes. However, examination of Fig. 4
reveals a marked dispersion about the best-fit line of the observed points
relating to nesting material. For this reason, the interpretation that the
63 and 64 slopes differ statistically is open to question that this difference
in slope implies biological significance. I therefore believe it wisest to as-
sume that 63 and 64 are really identical, or nearly so.

If this is so, we may compare the slopes of the mean of 61 + 62 with that
of 63 -f 64. Here the t test

Uh + 62) - 1(63 + 6.) ^ ^^^25

VVar. Uh + 62) - K&3 + 64)
with a p value less than 0.001.

John B. Calhoun

o 1000

1 1 1 1 1 1 I 1 1
/I Paper strips

B Food pellets

\-v

—

•

1 1 1 1 1 1 1 1 1

DISTANCE FROM HOME IN FEET
FROM WHICH OBJECTS WERE TAKEN

Fig. 4. Frequency of transporting nesting material and food into the home com-
partment from points at successively greater distance from it. A, 3810 strips of paper;
B, 3551 pellets of food. See Table la.

It is therefore concluded that structuring the environment with items
inducing responses leads to a reduction of the distance from home at which
trips are terminated.

3. The Prob.\bility of Terminating Trips

The two prior sections merely demonstrate an effect produced by struc-
turing the environment. They do not further our understanding of the
underlying biological process.

The behavior of rats in the structured environment provides the clue.
During any period of intensive transportation, one trip almost immediately
followed the preceding one. And yet the distance at which a particular
trip terminated bore no relationship to the distance at which the previous
or following one terminated. It was as if the rat w^as blind to its surroundings
on the outward trip until some neural switching mechanism became acti-
vated in a random fashion with reference to the time of the trip's initiation.
This switching on (or off?) placed the rat in a perceptive phase at which
time it responded by picking up the nearest relevant object and trans-
porting it into the home compartment. Therefore, it will be helpful to deter-

1. The Social Use of Space 13

mine the probability of this switching, which is synonymous with the
probability of terminating a trip.

Let: tj = the number of trips reaching any jth distance from home.
Nj = the number of trips that stop at the jth. distance.
Pj = probability of stopping at the jth distance.

Then:

tj~i - tj = Vi-itj-i (10)

Nj = p,tj (11)

Nj - Nj_i =pjtj - pj-it.j^i (12)

If Pj = p (a constant independent of j), then:

Nj - Nj^i = p(ij - ij_i)

= -p(tj-i - l.j) (13)

Nj - iVy_i = -piptj^i) (14)

And by analogy to Eq. (11):

iVy_i = pti^i (15)

Substituting Eq. (15) into Eq. (14):

Nj - Nj^i = -pNj^i (16)

Therefore

p = (Nj_r - Nj)/Nj_r (17)

This p represents a constant probability of terminating trips which
arrive at a point regardless of the distance from home. Rigorous proof that
this p actually is a constant is difficult from present data because of the
barrier produced by the relatively short alley. However, the validity of a
constant p, independent of distance, may be arrived at intuitively since an
equation of the form y = exp (a + hx) best represents the observed data.
In other words, log ij plotted against x forms a straight line. Whenever
this is so, Eq. (17) must be true.

Utilizing Eq. (9) stated in the form :

loge?/ = a - bx (18)

the expected number of trips terminating at Nj^i and Nj, where j = 2,
were found to be as shown in Table lb, along with the p values calculated
from Eq. (17). Thus, the probability of 0.182 of terminating trips arriving
at any distance in the unstructured alley is increased to 0.24 by structuring.

John B. Calhoun

Table lb

Probability of Terminating Trips with Respect to Environmental Structure

AND Level of Activity

Alley

Secondary condition

Expected number of
trips terminating at:

A^,-i

A^i

Mean

During first 2.5 hr. of hyper- 237.5 194.3 0.182«

activity
Unstructured 0 . 182

During normal nocturnal ac- 510 417 0.182°

tivity

Structured

Paper transportation
Food transportation

999

783.3 0.261

837 653.5 0.220

0.240

" The mean b value of —0.20115 used in the calculation of p.

1000

i 100

~>s^

— 1

A ■

1 1 1 1 I 1 — —

/f£ Trips terminated Z

RE Objects transported _

\ —

. 1. ..

1 . 1 . . 1 ~

DISTANCE FROM HOME IN FEET

Fig. 5. Theoretical curves depicting relative likelihood of trips going beyond suc-
cessive distances from home based upon an initial 1000 departures. A, in an unstructured
alley (see Fig. 3) ; B, in an alley structured with objects available for transport back
home.

1. The Social Use of Space 15

The amount of structuring used here increased the probability of stopping
by 32% [i.e., (0.240 - 0. 182) /0. 182]. The extent to which this difference
modifies the use of a field may be visualized by considering the following:
Of every 1000 trips arriving 1 foot from home in the unstructured alley, 30
would stop at the tenth foot and 134 would continue farther, whereas in
the structured environment only 20 would stop at the tenth foot and only
64 would continue farther (see Fig. o).

In closing this section it should be emphasized that the probability of
stopping is basically a function of time. In the particular situation, the p's
refer to the time required for the rats to travel a distance of 1 foot.

A further question may be asked: "How does structuring increase the
probability of stopping if the animals are perceptually 'blind' to specific
stimuli to the extent of being unable to exhibit directed responses to them?"
One may visualize the situation in general terms: There is some assembly
of neurons which provides a signal terminating an ongoing behavior, such
as outward locomotion from home. The magnitude of this signal neces-
sary for behavioral termination reciuires simultaneous firing of some x
number of neurons. This assembly of neiu'ons may be called a "governor."
It must be located anatomically in some subcortical portion of the brain
precluding conscious awareness of its functioning. During any ongoing
behavior all perceivable stimuli unrelated to the ongoing behavior initiate
impulses which arrive at the behavior terminating governor. Each unit of
impulse causes an increase in the rate of firing of the neuronal net forming
the governor. The greater the intensity of such stimuli, or the greater the
number of stimuli of a given intensity, the more rapidly will the neurons
of the governor fire and thus the shorter will be the interval between emis-
sion of effective signals by the governor. This signal both terminates the
ongoing behavior and produces awareness of stimuli appropriate to eliciting
those responses appropriate to initiating some other behavior.

4. Initial Hyperactivity in a Strange Environment

For the rats discussed in Section III, A, 1 the level of activity in terms
of trips per rat per hour (Fig. 3 and Table la) was approximately six times
as high during the first 150 minutes of exposure to a strange environment
in comparison with the later periods following adjustment. This observa-
tion requires postulation of a second governor, one which determines the
probability of initiating trips. Initiating trips is here considered to be
synonymous with initiation of periods of diffuse undirected motor activity.
Examination of the expression of hyperactivity during the first 2-2|
hours of exposure to a strange environment provides insight into the func-
tioning of this governor of trip-initiations.

John B. Calhoun

For 71 of the 73 rats previously discussed with reference to the distaiu^o
of trip-terminations, a count was made of the total feet traveled each half-
hour during the first 2| hours of exposure to the alley. A similar analysis
was prepared for the 26 rats, considered further in Section III, A, 5, which
emerged into the alley more than momentarily and for which the recording
system functioned properly.

For each set of data, results are quite similar (Fig. 6). Activity con-
tinuously declines during initial exposure to a strange environment. But
the important point is that this decline begins from an extremely elevated
state of hyperactivity. From Table la it may be calculated that the average

z i 100

K I

I- U. —

llJ tt

^ X

*- <

Q 2

1 1 1 1 1

- 4 By ^/ Flats

^\. B By 26 Rats

—

\\.

—

"^"^^ >v^

1 1 t I 1

12 3 4 5

CONSECUTIVE 30 MINUTES IN ALLEY

Fig. 6. Hyperactivity in a strange environment: Curve A represents the mean
activity level of 71 rats immediately after first exposure to the Ferguson Alley. Curve B
similarly gives the mean response of 26 rats to the NIH Emotional Activity Alley.

round trip approximates 10.44 feet. During the normal nocturnal period of
activity, trips occur at the rate of 1.58 per hour. This means that in this
one-dimensional alley rats travel 8 feet per half-hour on the average when
adjusted to their environment. As may be seen from Fig. 3 the activity is
in(;reased nearly twenty times normal during the initial half-hour in this
strange environment. Projection of the curves, shown in Fig. 6, indicates
that a normal level of activity will be reached by 3—4 hours of adjustment.
Such decay curves of hyperactivity suggest that the rats secrete some
humoral agent upon their initial exposure to strange stimuli such as repre-
sented by the activity alley. This humoral substance increases the rate of
firing of the neuronal net comprising the governor which determines onset
of diffuse motor activity, such as trips out into the alley. As this substance
is degraded, the frequency with which this governor emits signals initiating
trips decreases until normal behavior is achieved.

1. The Social Use of Space 17

5. Avoidance of a Strange Field

During the pursuit of these studies on behavior in a one-dimensional
habitat, a few rats failed to emerge into the alley until after the lapse of
several hours. ]VIost frequently such rats were members of subgroups
having had less opportunity to adjust to novel stimuli. This suggested
that the less opportunity an animal had to make adjustments to strange
stimuli, or the more novel was the strange environment, the more likely
it would be avoided.

New alleys were constructed to permit exploration of this hypothesis.
These alleys consisted of an 8 X 8-inch channel 15 feet long. At one end a
home nest box could be attached from which the rat gained access to the
alley by way of a 3 X 3-inch door. At the opposite end of the alley, light
from two 60-watt lamps shone through a ground glass plate. Unless other-
wise modified, the floor was stationary. A photoelectric cell at each 3-foot
interval along each alley initiated a signal to an Esterline-Angus event
recorder each time a rat passed. This apparatus is called the "NIH Emo-
tional Activity Alley."

Each of the 7(i subjects, Osborne-]Mendel male rats, was housed alone
for 3 months from weaning in a smaD cage precluding the visual perception
of any object outside the cage. Cages were not opened. Food and water
were delivered into the cage through channels making it unnecessary to
open them. These procedures minimized opportunity for adjusting to new
configurations of stimuli. In the terminology of Section XIV, G, 1, the
isolation cage represents an Ei configuration of stimuli. Under this termi-
nology El • • ' En represent a series of discrete configurations of stimuli
which an individual may encounter for the first time in that order.

Twenty rats were transferred directly from the isolation cages into the
activity alley for a 2-hour exposure. Of the remaining 56 rats, 24 were ex-
posed to an E2 configuration for 2 hours for 10 days prior to being placed
in the alley. Similarly, 16 others were exposed to an E-^ configuration, while
the remaining 16 were exposed to E2 for 2 hours, then immediately exposed
to Es for 2 hours for 10 days. Three days after these exposures, which pro-
vided opportunity to adjust to the new stimuli of E2, E:^, or Ei and £"3,
each rat was given a 2-hour trial in the activity alley, which represented
an Ei configuration. Half of each of these four groups were placed in the
alley with a stationary floor, an Eu. configuration; while for the other half
the alley contained sections of tilting floor, an E^v, configuration, which
clanged when the rats passed over them.

Greater detail of these studies are presented in Section XIV, G, 4 and in
Table lib. However, the results presented in Table Ila suflfice for the
present purposes. Both aspects of the hypothesis were confirmed. The

John B. Calhoun

Table Ila

Effect of Training and Degree of Strangeness of a Field upon Its Being
Entered during 2 Hours of Exposure by Each of 76 Rats"

'Training" in adjusting to

Response

strange stimuli

Degree of strangene.ss of alle}'

alley

Yes

High Low
(tilting floor) (stationary floor)

Entered
Avoided

37
19

4
16

13 30
25 8

-A data: x^ = 10.8045; p less than 0.001; B data; x'' = 13.7110; p less than 0.005.

stranger a field or the less opportunity there has been to adjust to novel
stimuli, the greater will be the tendency for an animal to avoid that field.

6. Probability of Vacillating at the Termination of Trips

In the unstructured Ferguson Activity Alley, rats usually went directly
out to the point where the trip terminated, then turned around and went
directly back to the home compartment. Yet occasionally when a rat
made an excursion out into the alley, it would wander back and forth

Table Hb

Effect of Kind and Sequence of "Training" in Adjustment to Strange Stimuli
UPON Avoidance of an Activity Alley

Group"

E,A Alley

EiB Alley

Total
trials

Trials
entered

Proportion
entered

Total
trials

Trials
entered

Proportion
entered

A
B

C
D

40
48
32
32

11
33
31
24

0.275
0.688
0.969
0.750

48
32
32

0.050
0.500
0.688
0.188

Total

152

0.651

152

0.355

" See Section XIV, G, 4 for details of treatment of the four groups.

J. The Social Use of Space

about the place where the trip terminated. Two hundred and fifty-six of
the total 2357 trips studied with reference to the initial period of hyper-
activity, terminated in such vacillations. As graphically shown in Fig. 7
the probability of vacillating is proportional to the length of the trip.

0 5 10

LENGTH OF TRIP IN FEET
Fig. 7. Probability of trips including a period of vacillation, wandering back and
forth about the point of termination. Curve based on 2,357 trips by 136 rats during their
first 150 minutes in the Fergu.son Activity- Alley.

These results are given here because of their bearing on the question of
the proportion of time an animal is capable of responding to static compo-
nents of its environment. As I have stated above, I am proceeding with this
formulation, utilizing the hypothesis that directed responses can be made
only at the ends of periods of diffuse activity, which are represented here
by the outward termination of trips. If this responsiveness at ends of trips
persists throughout the wandering about at the end of trips, then it follows
that the farther a familiar type of object is from home the more likely it
will be responded to in proportion to trips terminating there. This conclu-
sion is subject to the (jualification that the object is so sparsely distributed
in space as to require more searching than mere detection of the surround-
ings at the point of terminating a trip.

IV. Use of a Two-Dimensional Field

A. Theoretical Origin

We are here considering animals whose movements emanate from a fixed
home base. For them a two-dimensional field differs from a one-dimensional

20 JoJui B. Calhoun

one in that in a two-dimensional field the area available at successive radial
distances from home increases with radius. Considering this fact, will the
behavior exhibited by rats in a one-dimensional field lead to an equation
for home range closely resembling the bivariate normal distribution
[Eq. (3)]?

When Casby and I originally found that the bivariate normal distribution
did conform with the observed home range resulting from captures, we
were merely culminating a search for a means of describing the distribution
of captures about the mean coordinate point of capture. This conformity
revealed nothing about the biological mechanisms involved. We shall now
inquire whether the phenomena of (a) decreasing frequency of arriving at
successively greater distances from home, and (b) the probability of
wandering increasing with distance from home suffice to explain the origin
of the bivariate normal type home range.

A critical issue concerns the origin of the observation or "capture." Two
types of observation are possible. First, the observer may record the
physical presence of the animal at successive points independent of the
activities of the individual. Second, the observation may arise as a conse-
quence of the animal responding to an object placed by the investigator.
Captures in traps represent this type of observation.

An assumption is made regarding where responses, such as entering
baited traps, will be made. This is that such responses to continuously
present and unvarjdng stimuli occur only during the period of wandering
at the end of trips. This assumption implies that the animal remains in a
perceptually blind state during the outward and return phases of a trip.
The circumstantial evidence suggesting this assumption will not be con-
sidered here.

Let : Pi = probability of terminating a trip at radius r.
t = time spent wandering at r if it stops there.
C = probability of capture at r, which equals tPi times geometry
factor of two-dimensional space.

Then:

Pi(r) = Ae-^'-^ (19)

t{r) = Br (20)

These two equations, in which .4, B, and a are constants, represent the
two basic assiunptions regarding use of one-dimensional space. Then
considering the geometry factor:

C = Kre-'i" ' rdrde (21)

1. The Social Use of Space 21

And normalizing, it is found that:

C = ^ — rh'-'-'" dr (22)

Where r/a = R, the cumulative probability of capture, Cum, as a func-
tion of the radius R from home becomes :

Cumii = 1 - f — + /? + 1 j e-« (23)

Equation (23) above will be called Curve II as shown in Table III and
Fig. 8. It may be compared to a similar cumulative probability curve for
the bivariate normal distribution, which will be called Curve I, and which
has the form:

Cumi = 1 - exp (-rV2) (24)

Curve II, Eci. (23), may be compared to Cui've I, Eq. (24), by con-
verting R into units of r, where r, the radius from the center of the home
range, is measured in the a units of the bivariate normal distribution, pro-
vided the constant a. oi R = r/a is known. This conversion was arbitrarily
accomplished as follows: It can be shown that the cumulative probability
of "capture" of Eq. (23), when expressed in terms of a and r, has the form :

Cum = 1 - ^-— (r2 + 2ar + 2a2)e-'-/a (25)

2a:-

By Eq. (24), Curve I, when r = 1.2, Cum = 0.513. Therefore, by succes-
sive approximations, utilizing Eq. (25), it was found that when a = 0.44,
Cumii = 0.511 atr = 1.2. Therefore, 0.44 is the a conversion factor applied
to ^ = r/a, so that Curves I and II may be compared in terms of the bivariate
normal home range cr distance.

The values for these two curves, as shown in Table III, are shown in
Fig. 8. Note that up to about 2<t radius these two curves are so nearly
identical that they are either likely to approximate actual field data equally
well. There is considerably more "tail'' to Curve II, but since so few ob-
servations occur in the greater than 2a range, it will still be difficult to
decide which of these two curves most nearly approximates actual field
data for the longer recapture radii.

However, the objective was to determine how well phenomena observed
in the use of one-dimensional space could lead to a curve approximating
the bivariate normal distribution. One of the assumptions was that the
wandering responsive phase was proportional to radius from home. In
the analysis of wanderings, vacillations at ends of trips in the one-dimen-

John B. Calhoun

s
o
W

s
o

5 2

■<

"■ s

^ 5

^ «

« ^

^ pq

Pi w

s
p
o

^ +

rH|N

(-1

p:; .2

lOOCOOODOJt^OGCiCO-Ht^O OOl

o d o o d o o d c o o c c d d

1^ fO O O 05
'-lCCCOI>00OiO5O5O5O5O5O5Oi 005

ooooooooooooo

o o

lO CO o

O lO iM lO QO lO 05 O
C^3 t^ t^ ^ ^1 CO »C CO CO O O t'- CO

COiOCO-— iOt^iCO-*0(X'C«DC2
0^00 0 0(^00050205050505

coocoooo

o o o o

00 CO o
i—(l>.050502CO'^lO

Ttiiocofoi-oo405cocoooioooo5

'*COC0<-(I^O'30'tlt^00O5O505
O^COIOC00000050505050505

CO rt<
05 <0

CO 00

ooooooooooooo

Tt<l:^.-i0505t^cOCOi-H-«tii-(0
C^CO(MOOCO':t<COOOr}<^t^O
iOcOCO'<*iOO(M05'tit^05050
O'-HC0l0C0000005C50505O

doodddd oooo--!

COC005^JiCOC^'*it^OCO«005 oo

odd— lrHr-H(N'(M(MC0CO00CO i-HC^

1. The Social Use of Space

< 0.5
m

• Ma/e

O Female S Fawns

4.0"

.3 .6 .9 1.2 1.5 1.8 2.1 2.4 2.7 3.0 3.3 3.6 3.9
RADIUS, r, FROM CENTER OF HOME RANGE IN 0" UNITS

Fig. 8. Accumulated probability of "capture." Curve / is the expected from the
bivariate normal distribution function. Curve // is the expected based upon (a) ter-
mination of trips according to the equation, y = e«+''* in which the slope h is negative,
(b) probability of v/andering at end of trips is proportional to radius, and (c) captures
only occur during the wandering phase at end of trips. Curve /// is the accumulated
travel path with no wandering at end of trips and "captures" represented by observa-
tion of the animal in motion.

sional alley shown in Fig. 7, it was only recorded whether or not a wander-
ing occurred at the end of trips. No measure was made of the length of the
wandering. However, it is my impression that the longer a trip, the greater
will be the amount of wandering at the end of a trip if wandering is initiated.
This would mean that the amount of responsive wandering at the end of
trips is proportional to r- rather than just to r. If this is really the case,
then we can derive a fourth cumulative probability curve, Curve IV,
which has the form:

Cumiv = 1

+ - + /2 + 1

\e~^

(26)

By successive approximations utiUzing Eq. (25), it was found that when
a = 0.32, Cumiv = 0.516 at r = 1.2. Therefore, 0.32 is the a conversion
factor apphed to R = r/a, so that Curves I and IV may be compared in
terms of the bivariate normal home range a distance. Note that thus there
is forced conformity of both Cumn and Cumiv to Cumi at the 1.2 value
of r.

Near the home range center. Curve IV predicts fewer captures than is
imphed by Curve II or found by observed captures. Table III. However,
at the longer radial distances from the home range center, Curve W more

24 John B. Calhoun

closely approximates Curve I and the observed data than does Curve II.
At present parsimony demands assuming only that the farther an animal
moves from his home range center, the more likely it will terminate an
outward trip and respond to stimuli near the place of stopping. Extremely
careful observation is required to determine if animals tend to wander still
farther about the points of termination of trips as these points occur
farther from home.

B. Travel-Path Home Range

Captures or responses represent only one method of assessing home
range. I have indicated that one assumption regarding behavior is that on
the outward trip from home and on the return trip to home, the animal is
in a preceptually blind state during which static stimuli fail to elicit
responses. And yet it is possible to observe and record the presence of such
nonresponsive individuals on this outward and return trip. Utilizing Eq.
(19) and considering the effect of the geometry factor in the sense that the
observer in a two-dimensional field can record an animal only if it passes
directly by the observer, and assuming that the amount of wandering at
end of trips is minimal, the equation for the cumulative probability of ob-
serving an animal during its travels. Curve III, becomes:

Cumin = 1 - (1 + 7?)e-« (27)

Data for Curve III with the a constant = 0.44 are shown in Table III and
on Fig. 8. Obviously, if "captures" represent such observations, the animal
will appear to spend more time closer to home. Leopold et al. (1951, Fig. 50)
provide data on home range based upon visual observation of marked mule
deer. They presented their data in terms of number of observations within
successive 100-yard bands from the site of capture. The actual home range
center will be on the a\'erage somewhere to the right or left of the hne of
length d connecting the point of capture and the point of later observation,
the actual radial distance from the true home range center would be K • d/2
when K is between 1 .0 and 1.414. Without going into the origin of K, it is
still apparent that the distances given by Leopold et al. (1951) can be
utilized as approximating proportionality to radial distances of observa-
tion from the home range center. The cumulative probabilities of observa-
tion for 102 observations of males and 103 observations of does and fawns
are with distance, respectively: 100 yards (0.363,0.495), 200 yards (0.500,
0.815), 300 yards (0.678, 0.952), 400 yards (0.726, 0.980), 500 yards
(0.862, 1.0), 600 yards (0.862,—), 700 yards (0.932, — ), 800 yards
(1.0, — ). Noting that 0.50 of the males were observed within 200 yards

J. The Social Use of Space 25

of the point of capture, and that 0.495 of the females and fawns were ob-
served within 100 yards of the point of capture, and utilizing Eq. (27) and
the fact that R = r/a, a, in the sense of d, for males is 270 yards and for
females and fawns is 137 yards. On this basis the observed cumulative
probabilities of observation were converted into sigma units of distance
from the point of capture and posted on Fig. 8.

There is forced conformity with Curve III, Eq. (27), at 0.74(7. However,
the further proximity of the observed points, up through l.oa, to Curve
III provides credence to the general formulation of Eq. (27). Observa-
tions may occur when the animal is on either the outward or return por-
tions of a trip or during the wandering at the end of a trip. The more an
animal wanders at ends of trips relative to the cumulative travel path,
excluding wandering at the end of trips, the farther the observed points
may be expected to diverge from Curve III and approach Curves I and II.
In fact, were it possible to obtain adequate assessments of home range in
terms of observations made of the individual during its travels as well as
of data derived from responses such as entering traps, it would be possible
to calculate wandering distance at r.

V. Summary of the Concept of Home Range

Developing an attachment to a restricted region to which an animal
returns after wandering outward from it requires that the individual belong
to a species which has evolved the capacity to retain the memory of prior
experience. If an animal remains at the site of its birth there is no question
that the immediate surroundings of this site will be known better than
more distant ones. However, many individuals are forced from the site of
their birth. The term "forced" is here used as a postulate derived from the
observed avoidance of fields containing strange stimuli. Parsimony de-
mands the assumption that an animal will remain at home unless forced
away, until it can be shown that an animal will change its home in the
absence of any factors which might tend to expel it.

Even if an animal is expelled from the site of its prior residence and
begins to wander at random, it will be chance cover some areas more fre-
quently than others. As it thus becomes familiar with a restricted area,
it will increase its avoidance of less frequented areas. This process will lead
to a repeated frequenting of a particular restricted location which can be
defined as "home." Home may acquire other attributes, such as the con-
struction of a den, but such additional attributes are not necessary for a
site to become a home.

At irregular, probably random, intervals the individual engages in ex-

26 John B. Calhoun

cursions from home. During these outward excursions the animal remains
in a refractory state during which it is unable to respond to stimuli en-
countered. Only at the termination of trips, including the period of wander-
ing at end of trips, will the animal be able to respond to objects or situa-
tions such as it had already passed on the outward trip. However, those
objects which were bypassed on the outward trip do alter the probability
of terminating trips. The greater the number of such objects, or the greater
the intensity of stimuli emanating from them, the greater will be the
probability of a trip stopping after reaching every distance from home.
Should novel stimuli be encountered during a trip, the shorter will be the
interval after reaching home before initiating another trip. This holds in
so long as the novel stimuli are not excessive. In this case, the animal will
remain at home for unusually long periods before again venturing forth.
The probability of wandering increases with distance from home. These
several aspects of locomotion and responsiveness to stimuli lead to a
probabiUty of responding within the region about an animal's home,
which so closely resembles that of the bivariate normal distribution func-
tion that the latter may be used to represent home range.

VI. Continuous Removal Trapping of Small Mammals

When a large number of snap-traps are set within a habitat where mice
and shrews live, it is usually observed that more individuals are captured
on the first than on the second or third day, and that fewer animals enter
traps on the third than the second day. It has generally been assumed
that this decline in catch through time would continue were the traps left
set for a longer period. Based upon this assumption, various equations
(e.g., Zippin, 1956; Calhoun and Casby, 1958) have been developed for
estimating the number of individuals which would eventually enter the
traps.

Furthermore, it has also generally been assumed that the catch for a 3-
or 4-day period would reflect the relative species composition of the com-
munity being sampled. This assumption served as the basis for establishing
the North American Census of Small Mammals (Calhoun, 1949-1957).
During the nine years, cooperators provided results (see NACSM Release
No. 9) for 1615 standard trap-lines consisting of 20 stations, usually 50
feet between stations, 3 traps per station, run for three consecutive nights.
Despite the laudable objectives of this cooperative endeavor, this tre-
mendous effort has, for reasons we shall shortly see, provided inadequate
data for insight into either relative density or species composition.

Questioning the usefulness of short-term removal trapping developed

1. The Social Use of Space 27

slowly. It began in 1950 when I asked A. Dexter Hinckley, then a summer
student at the Jackson Memorial Laboratory, to continue eight NACSM
lines for 15 days (see Section VI, E and Fig. 13A). Despite the expectation
of a continually declining catch after the first 3 days, it turned out that the
number taken from day 4 onward actually increased each successive day
until on day 15 three times as many individuals entered traps as did on
day 1.

These results initiated an intensive effort to explore the results obtained
from prolonging the period of removal trapping. I am particularly indebted
to Drs. William L. Webb and Earl F. Patric of the Huntington Wildlife
Forest, and to my colleague. Dr. Kyle R. Barbehenn. Without their per-
mission to utilize here some of the data from their extensive studies of
removal trapping, it would have been impossible to develop the concepts
elaborated in succeeding sections of this paper. I have also profited from
a number of 30-day census contributed by other NACS]\I cooperators.

Our concern here will be with the contribution of this program, and
allied research developing from it, to furthering our understanding of
home range and the organization of small mammal communities. The
following few sections merely present results. Note that within each study
presented, the several species present markedly different patterns of cap-
tm*e by traps over time. Interpretations based upon the data in Section
VI are given in Sections VIII, IX, and X.

A. Rich Lake Island, New York, 1952, Sixty-Day Removal Study— Data
Contributed by William L, Webb'

Four NACSM traplines were run on this island in the Huntington
Forest for 60 consecutive days, beginning August 16. The red-backed
mouse, Clethrionomys, and the shrew Sorex formed the majority of the
catch. Figure 9 gives the results for the first 30 days. Note that
Clethrionomys, which entered the traps in large numbers during the first
few days, was by the end of 30 days represented by only half the total
catch as that for Sorex, Avhose peak in catch per day did not come until
the ninth day. Both species exhibited a secondary minor period of increase
in captures.

During the same summer, 34 NACSM lines were run for the standard
3-day period in similar forests on the mainland (Table IV). As on the
island, comparatively few Sorex entered traps on mainland areas during
the first 3 days of trapping. And yet, continuous removal trapping for an
80-day period (Section XI) in one tract of the mainland forest produced
()17 Clethrionomys and 1225 Sorex.

John B. Calhoun

30-DAY
TOTAL
• • Clefhrionomys 97

m- ■ Sorex

10 16 20 25

DAY OF TRAPPING

Fig. 9. Daily removal captures of the two most abundant species inhabiting a 30-
acre island; recorded by Dr. William L. Webb.

Table IV

Comparison of Results between Short-Term and Long-Term Removal
Trapping on the Huntington Forest during 1952

Genus

Location

Clethrionomys Sorex Perornyscus

80-Day total for continuous removal 617
trapping on a 30-acre mainland area
(see Section XI)

1225

34 NACSM lines each for 3 days on main- 753 (22 . 1 ) «
land

17 (0.5) 38 (1.1)

4 NACSM lines on Rich Lake Island
60-day total
Total 1st 3 days

133
53 (13.3)

243

4 (1.0)

27
0

Mean 3-day totals shown in parentheses.

1. The Social Use of Space

B. Chadwick Woods, Montgomery County, Maryland, Removal Study,
1958-1959— Data Contributed by Kyle R. Barbehenn

Five circular traplines, each with a radius of approximately 300 feet,
were used. The number of traps per station and the interval between
stations varied among the lines. However, for the present purposes these
differences are unimportant. As with the Rich Lake Island study, the
genus, here Peromyscus, with the initial most rapid input was represented
by only half as many (57) captures as another genus, Blarina, which
entered the traps more slowly (116 taken) (see Fig. 10). A third genus,
Sorex, began its period of maximum captures only after the majority of the
other two genera had been removed, but only 48 were captured. Note that
Peromyscus exhibited a fairly rapid rate of capture until approximately
65% of the 30-day total had been taken. Following an intervening period
with very few captures, there ensued a secondary period of increased
captures.

CHADWICK WOODS- MONTGOMERY COUNTY, MARYLAND

10 15 20

DAY OF TRAPPING

Fig. 10. Differential "schedules" of entering traps expressed by the three species
recorded by Dr. Kyle Barbehenn in his study of continuous removal trapping in an
upland hardwood forest.

C. Comparative Catches, Huntington Wildlife Forest, 1952-1953— Data
Contributed bv Earl F. Patric and William L. Webb

During these two years Peromyscus and Blarina were universally scarce
on the Huntington Forest. In order to determine the general pattern of
their input, data from five separate plots must be pooled.

John B. Calhoun

Four plots consisted of concentric circular traplines, 75 feet between
circles, with one trap each 10 feet along the lines. Two plots consisted of
three lines, and two of four hues. The fifth plot is represented by the first
15 days' results of the Rich Lake Island study presented in more detail in
Section VI, B. All areas were sampled for at least 15 days (Fig. 11). Each
day all animals captured were removed.

Note that whereas the catch of Clethrionomys initially declines through

Clethrionomys
Sorex

Peromyscus
Blarina

15 -DAY

TOTAL

286

377

--a

5 10

DAY OF CAPTURE

Fig. 11. Summated results from five continuous removal studies conducted by Dr.
Earl Patric and Dr. William Webb during 1952 and 1953 on the Huntington Forest,
New York.

time, all other species exhibit an increase in catch per day for the first 4
or 5 days.

D. Comparative Catches, Huntington Wildlife Forest, 1951 — Data Con-
tributed by William L. Webb

Nine NACSM lines were each run for at least 24 days during September
of 1951 (Fig. 12). In contrast to the following two years, Peromyscus
nearly equaled Clethrionomys in abundance. Likewise, Blarina was abun-

1. The Social Use of Space

iCIethrionomys

DAY OF CAPTURE

Fig. 12. A continuous removal study conducted by Dr. William L. Webb on the
Huntington Forest in 1951. This year represented a time when the dominance relation-
ship of the three most abundant species was not yet clarified, the small mammal com-
munity was in a state of social flux. See Table V. Points shown are 3-point moving
averages.

dant. Only Sorex was rare. The relative likelihood of capture during the
initial and terminal days of trapping forms a most interesting series (Table
V) . Initial and terminal likelihood of capture are inversely related despite
the fact that usually over 50% of each genus were taken during the middle
period of trapping.

Table V
Huntington Forest (1951) 24-Day Continuous Removal Trapping

Genus

24-Day total

Proportion of 24-day total during

First 5 day.s

Last 5 days

Peromysctis
Clethrionomys
Blarina
Sorex

217

267

229

0.418
0.311
0.188
0.024

0.042
0.079
0.166
0.585

E. Comparative Catches in Maine (1950) and Maryland (1953)

Eight NACSM lines within an 80-acre tract of a much more extensive
continuous forest opposite the Hamilton Station of the Jackson Memorial

John B. Calhoun

Laboratory on Mt. Desert Island formed the trapping procedure in the
Alaine study. Four concentric circular traplines, each of a 75-feet greater
radius than the next innermost one, having one trap each 10 feet along the
lines, formed the trapping procedure in the Maryland study. During the
years these studies were conducted, I was still working under the assump-
tion (see Calhoun and Webb, 1953) that the large numbers of animals
taken shortly after the first 3 days of trapping resulted primarily from in-
vasion from beyond the periphery of the trapping area. These relatively
short-term removal studies are presented here (Fig. 13) since they provide
further insight into the differential rate of input for the several species
making up the small mammal community. Total lo-day catch by genus

B (MARYLAND)

Peromyscus ^

1 1 Blarina

7/ /

Pitymys

]/ J

1 1

DAY OF TRAPPING

5 10 15

DAY OF TRAPPING

Fig. 13. Two short-term removal studies. The more dominant a genus happens to
be the sooner will 50% of its members be trapped. In Fig. 13.\: B = Blarina; P =
Peromyscus.

for the Maine study (Fig. loA) : Clethrionomys 80, Sorex 53, Peromyscus
75, and Blarina 132. Total 14-day catch by genus for the Maryland study
(Fig. 13B) : Peromyscus 45, Blarina 76, and Pitymys 62. The Maine study
was conducted by A. Dexter Hinckley under the author's direction, and
the Maryland study was conducted by the author.

F. Comparative Catches of Peromyscus and Clethrionomys

Relationships between these two genera, revealed by NACSM census
data, have been particularly helpful in developing insight concerning com-
munity organization. Census from three localities, where both genera
occur, are represented by a large number of traplines. Those run in each

J. The Social Use of Space

locality encompass a period of several years, and so should provide a
representative picture of relationships in the respective habitats. Results
in Table VI derive from 37,080 "trap-days" of effort.

Drs. Earl F. Patric and William L. Webb provided the New York data
from the Huntington Wildlife Forest at Newcomb, New York. The majority
of the animals shown were taken in the years of high density, 1953-1954,
although the few taken during the low density years of 1955 and 1956 are
included. Dr. J. E. Aloore of the University of Alberta, Edmonton, pro-

Table VI
Comparative Catch of Two Genera in Three Localities

Genus Location Number

of lines

Day"

Total, Mean

3 days 1-3 per line

New York

Clethrionomys Alberta

Maine

90 586 470 380

(0.408) (0.327) (0.265)

1436

36 117 103 83 303

(0.390) (0.338) (0.272)

184

(0.489) (0.315) (0.196)

" Proportion of 3-day 2 shown in parentheses.

16.0

2.3

New York

(0.274)

117
(0.344)

130
(0.382)

340

3.8

Peromyscus Alberta

241

(0.577)

122

(0.292)

(0.132)

418

11.6

Maine

132

(0.434)

103

(0.399)

(0.227)

304

3.8

vided the Alberta census including the years 1948-1956. Dr. John A. King
and two U. S. National Park Rangers, Clifford Senne and L. S. Winsor,
and the author conducted the ]\Iaine census on Alt. Desert Island between
1949 and 1952.

Figure 14 shows the decline in catch from day 1 through day 3 for these
two genera for Maine and New York.

In Maine where both genera have low densities, their respective patterns
of decline in catch through time are very similar. However, the relatively
greater catch of Clethrionomys on day 1 should be noted with reference to

John B. Calhoun

the trends of the 15-day trap-out as shown in the left-hand graph of Fig.
13 for eight NACSM lines also run on Mt. Desert Island.

In New York, where Peromyscus had the same density as in Maine but
Clethrionomys was four times as numerous as Peromyscus, a striking differ-
ence in the trends resulted. Clethrionomys exhibited a typical, though some-
what slow, decline through the three consecutive days. In contrast, despite
removal trapping and thus fewer deer mice available for entering traps, a
greater number of Peromyscus entered traps each successive day. I wish
to emphasize that competition for entering traps contributed negligibly
to these trends. Three times as many traps were set each day as there were
total animals caught for the entire 3-day period.

• • Clethrionomys

A -A Peromyscus

■5(-

B. NEW YORK

DAY

Fig. 14. Capture rates for Clethrionomys and Peromyscus in Maine and New York.
Nearly codominance, or lack of dominance, is reflected by the Maine data, whereas in
New York Peromyscus is clearly subordinate to Clethrionomys. See Table VI.

In Alberta, where both species are on the average relatively abundant
(Table VII), there exists a marked seasonal difference (Fig. 15) in the
trend of input over time between these species. During the spring, when
low densities characterize both species, each exhibits a rapid rate of de-
cline. In contrast, by fall when high densities have developed for both
species, Clethrionomys shows a relatively constant input. Note the reversal
of the trends of input for Clethrionomys and Peromyscus when Fig. 1 1 and
Fig. 15B are compared.

VII. Toward a General Theory of Interspecific and Intraspecific Use of Space

The data presented in Section VI reveal that a continuous decline in
catch from one day to the next during removal trapping is the exception

1. The Social Use of Space

Table VII
Mean Catch per Trapline om Dr. Moore's Alberta Study Area

Season

Spring

Fall

Clethrionomys
Peromyscus

5.00
8.06

11.90
15.56

2.18
1.93

rather than the rule. Even where the numbers taken per day do initially
decline, a secondary increase usually ensues by the fifteenth day of trapping.
Furthermore, many species actually exhibit an increase in catch per day
as their associates are removed. Any trend of increase in catch after previ-
ous removal of associates can only mean that the survivors have in some
way altered their behavior so as to increase their exposure to traps. Further-
more, we must conclude that those caught earlier in time must, while still
alive, have suppressed this change of behavior. We shall now consider how
these results make possible a conceptualization of spatial, temporal, and
social organizations of the small mammal community.

A. SPRING

m • Clfthrionomys

▲ A Peromyscus

ALBERTA I, I-n-B NACSM LINES 1949-1957

DAY

B. FALL
i ^ • • Clethrionomys

N A— — -A Peromyscus

Y- N

\
\
\
\
\

DAY

Fig. 15. Seasonal capture rates for two species. With the increase in density from
spring to fall (see Table VII) Peromyscus apparently becomes dominant to Clethrionomys.

36 John B. Calhoun

A. A Two-Species System

The relationship between Clethrionomys and Sorex on the Rich Lake
Island (Section VI, A, Fig. 9) and between Peronujscus and Blarina in
the Chadwick Woods study (Section VI, B, Fig. 10) form the basic data
leading to this formulation. The shrews — Sorex and Blarina, respectively, —
in these two studies exhibited a delay in entering traps until many of the
mice, Clethrionomys or Peromyscus, had been removed. Yet despite this
slowness of entering traps, twice as many shrews as mice were taken during
the 30 days of trapping in each study.

These data pose two questions :

1. Why were there twice as many shrews as mice?

2. In what way (and why) did the behavior of the shrews change so
that after se\'eral days of trapping they were more exposed to traps
than initially?

The first assumption will be that the larger the home range the greater
the likelihood an animal will encounter a trap and be caught. Since during
the first few days of trapping many mice but few shrews were taken, despite
the greater abundance of shrews, the mice must have had considerably
larger home ranges than the shrews. Furthermore, since the number of
shrews taken per daj^ increased during the first 10 days of trapping, it
follows that their home range expanded as the mice were killed off. This
leads to the conclusion that the mice in some way inhibited the extent of
home range of these shrews. We may now designate the mice as being
dominant or alpha species and the shrews as subordinate or beta species.

Formulation of a theory depicting the social and spatial aspects of such
a two-species system recjuires the assumption of a uniform distribution of
centers of home range for the alpha species. This represents the simplest
assumption leading to a 1:2 ratio of number of alpha and beta species.
In the preparation of Fig. 16, a field of uniformly spaced dots (not showTi
in Fig. 16) was plotted. These dots represented home range centers for
members of the alpha species. A circle of radius half the distance between
centers was drawn about each center. Each of the larger circles in Fig. 16
encompasses some portion of an alpha individual's home range.

Now w^e can ask : Where is it most logical to find the home range centers
of beta species? They should be located at points minimizing encounter
by members of the beta species with members of the alpha species. The
interstices formed by juncture of each set of three neighboring home ranges
of alpha species represent such locations.

Here a beta individual is equidistant from three alpha individuals. Dis-
placement of the home range of a beta species member from such a point

1. The Social Use of Space 37

will increase its probability of encountering at least one member of the
alpha species. About each such home range center a smaller circle was
drawn (Fig. 16). This smaller circle represents the same proportion of the
beta species home range as does the larger circle for the alpha species.

Examination of Fig. 16 shows that according to this formulation there
will be in an ideal steady state exactly twice as many individuals of the
subordinate beta species as there are of the dominant alpha species. This
is because there are twice as many interstices between uniformly distributed
alpha home ranges as there are alpha home ranges.

Fig. 16. Spatial distribution of a dominant and a subordinate species. Large circles
represent a uniform distribution of one sigma radius portions of the home ranges of a
dominant, alpha, species, while the smaller circles represent a similar proportion of the
home range for members of a subordinate, beta, species. See text for other details,
Section VII, A.

Now, suppose that a few traps are placed at random within a habitat
characterized by inhabitation by such an alpha and a beta species. The
fifteen dots in Fig. 16 represent such random points. It is readily seen that
there is an alpha individual exposed to nearly every trap, but few in-
dividuals of the beta species are exposed to traps.

Now suppose we do set traps in such a system and remove the individuals
caught. It is the general experience in continuous removal trapping that
25-50% of the members of the alpha species which are taken during 30
days are actually taken during the first 3 days.

This can only mean that by the end of 3 days there must be many mem-
bers of the beta species whose neighbors of the alpha species are then no

38 John B. Calhoun

longer present. In the absence of the inhibiting influences emanating from
the former alpha neighbors, the beta individuals then make a comple-
mentary expansion of their home range. As they do so, some beta individuals
come into contact with traps and are also removed. Inspection of a number
of continuous removal census in which there is an alpha species and one or
more subordinate species indicates that maximum expansion of home ranges
of the subordinate species is generally reached by the fifteenth day of
trapping. After this time, the catch by day for subordinate species also
declines over time since fewer and fewer remain to be caught.

B. The Nature of the Inhibitory Influence

Both Sorex and Blarina in Fig. 10 and Sorex in Fig. 9 exhibit an increase
in catch even beginning on the second day of removal trapping. Similar
results apply to the subordinate species included in Figs. 11-14, although
not so apparent in those graphs where the ordinate represents accumulated
catch. These results indicate that even removal of a small proportion of
the alpha species is sufficient to induce home range expansion by sub-
ordinate species. Thus, the means of communication through which in-
hibition operates must be sufficiently effective and repetitive that a change
in the general field intensity (or frequency) of stimuli emanating from an
alpha species is detected within a few hours at least by subordinate species.

Bodily contact by random movement is unlikely to be effective. By the
same token that it takes several days before all alpha species are taken in
traps, it follows that in many instances subordinate species would be un-
aware of the absence of their alpha neighbors if this detection were a con-
sequence of a change in frequency of contact. Production and detection of
scent, at least where scent signposts are concerned, would likely operate
to inhibit home range expansion because of the persistence of scent beyond
the death of alpha individuals. Sight is unlikely to be an effectual means
of detection of alpha by beta species, both because of concentration of
activity during the night by many species and because many of these beta
species actuality spend considerable time under the leaf mold.

There remains vocalization and audition as the means of communication.
Although there is as yet no proof that such is the means of communication
whereby individuals can detect the presence of unseen neighbors, it stands
out as the most likely possibility. iVIost small mammals do vocalize.

Fewer barriers exist that might prevent or distort the passage of sound
through the environment than is true with regard to light stimuli or odors.
In the following discussions vocalization and audition will be assumed to
be the means of communication within and between species. However,

1. The Social Use of Space 39

identification of the means is unessential to the general argument; only-
recognition of the existence of some effective means of communication is
necessary.

C. The Learning of Signals

There exists the possibility that the response of one individual to a
signal emitted by another has become through evolutionary processes
one which does not require a learned association between the signal and
some act on the part of the emitter for its development. In the prior history
of such species there must have been the opportunity for associating the
signal with its emitter and there must have been survival value in the re-
ceptor developing an innate response to detection of the signal. However,
until such responses to signals are demonstrated to be innate, it shall be
assumed that they are learned.

We may now ask, "How may the members of a species learn a signal
when the individuals are characterized by fixed home ranges which may be
described by the bivariate normal distribution function?" In order to gain
insight into this question, we shall consider two neighbors, A and B. A's
home range center is fixed whereas B, who lives some distance away,
gradually shifts its home range center toward that of A. When the home
range centers are six home range sigma or more apart, it is apparent that
the probability of their meeting by chance will be extremely remote. This
relative probability of meeting is proportional to the product of their
density functions at any particular point (see Table 2 in Calhoun and
Casby, 1958).

However, as the home range center of B approaches that of A, these
two individuals will meet by chance on very rare occasions. Three examples
of the relative probability of A and B contacting are given in Fig. 17.
When the home range centers (HRC's) are 3.9 sigma apart, one peak is
1.5 sigma from ^'s HRC and the other is 1.5 sigma from B's HRC. In
examining Fig. 17 it is well to keep in mind that we are considering the
probability of contact at points along the line connecting the two home
range centers. At all distances intervening between home range centers,
from slightly over 3 sigma up to 6 sigma, there are always two peaks in
this curve of probability of contact between two neighbors. As the home
range center of B approaches 3 sigma to that of A , these two peaks approach
each other until at 3 sigma they coincide for the first time. This single peak
of highest probability of contact of two neighbors, which lies exactly half-
way between the two home range centers, characterizes all distances less
than 3 sigma intervening between the home range centers.

John B. Calhoun

Let us assume that when .4 and B meet there is some interaction between
A and B. That is, A responds to B and B responds to ^. At the same time,
each emits a signal. If such chance contact occurs frequently enough there
exists the opportunity of each individual associating the other with the
signal emitted at the time of interaction. Furthermore, we may assume that
learning is enhanced by other factors of the environment being constant

CONCERNING THE LEARNING OF SIGNALS

0.6 1.2 1.8 2.4

er DISTANCE FROM As H.R.C.

3.0

Fig. 17. Concerning the learning of signals. Not unless home range centers are
3 sigma or less apart will there be a single point of most probable contact of two animals
meeting by chance. This point lies halfway between the line connecting the two home
range centers. The closer home range centers approach, the higher will be the proba-
bility of chance contact and thus the more likely the association of any simultaneous
signal with the consequences of meeting.

at the time of interaction. When home range centers are more than 3
sigma apart, the two points of greatest probability of contact are separated
from each other and therefore are unlikely to have identical surroundings.
This nonidentity of surroundings, that is the absence of identical secondary
reinforcers, may be expected to retard learning.

However, at a 3-sigma interval between home range centers, there is
only a single point of greatest probability of contact; thus, at this inter-
vening distance between home range centers there is not only an increased

7. Tfw Social Use of Space 41

probability of contact because the home range centers are closer together,
but there exists a greater constancy of secondary reinforcers at the single
point of greatest probability of contact. It is for this reason that I suspect
that learning of signals is not likely to be effective unless home range
centers are 3 sigma or less apart.

Once an animal has learned to associate a signal with the animal which
emitted it, there then exists the opportunity that the detecting individual
can perceive (hear) the signal at some distance from the emitter and make
the appropriate response of approach or withdrawal. How far the signal
may be detected depends upon both the intensity of the signal emitted
and the ability of the detector to hear it. Presumably, sensory capacities
for detection and motor capacities for emission have e\'olved simultaneously
and in harmony. We may then wonder as to the distance over which such
evolution of capacities permits the detection of an emitted signal. In the
absence of any experimental data, introspection suggests that one might
anticipate evolution of capacities to the point that an individual can just
detect a signal emitted at the maximum distance between home range
centers which still permits the learning of such signals. As we have seen,
this distance is equivalent to 3 home range sigma. In other words, when
animal A is at its home range center, a signal emitted by animal B at the
border of A's home range (as represented by a 3-sigma distance) reaches
A in just the sufficient intensity to elicit a response by A. As B moves
farther than 3 sigma away from A, the signal exhibits further decrease in
intensity. Such reduced intensity may well be perceived by A until B
gets at least 6 sigma away. Thus, between 3 and 6 sigma, it is suspected
that the signal itself is perceived but is below the threshold necessary for
eliciting a response by the receptor. Signals arising between 3 and G sigma
from the receptor are here designated as contributing to what I shall call
"hum."

These characteristics of the signal are represented schematically in Fig.
18. A signal emitted by one animal when nearly in contact with another
may be given a rating of 1.0. For the purpose of later calculations, it is
assumed that there is an inverse decrease of intensity of the signal at suc-
cessive distances from the emitter until at a distance of 3 sigma it has
reached one-tenth of the intensity that might be recorded at the emitter.

Investigations of this formulation requires that the sound signals emitted
by typical individuals of a species be recorded and other individuals of the
same species trained to exhibit a response upon perception of a recorded
vocalization. Then, in the native habitat of the species, the trained subject
must be moved continually farther away from the sound source until it no
longer exhibits the characteristic response which it had been trained to
perform following presentation of the signal. This distance may then be

John B. Calhoun

compared with the animars home range sigma, as determined by Hve trap-
ping and related observational procedures.

D. The Distance between Neighbors of the Same Species

From the observed tendency of a 1:2 ratio between alpha and beta
species when coexisting in the same habitat, it was concluded that the
home range centers for the alpha species should approximate a uniform
distribution. However, this analysis provided no insight into whether there

^ 0.04 -

°^ (0-0.3O-)

DISTANCE FROM H.R.C.
AT WHICH SIGNAL IS PERCEIVED

Fig. 18. Theoretical conceptualization of signal characteristics with reference to
emission at a constant intensity at the home range center and detection and response
by others at distances from it.

might be some ideal distance which should intervene between home range
centers. In order to seek this insight we shall consider a single species com-
munity in which all individuals have exactly the same-sized home range,
centers of home range are uniformly distributed, and the bivariate normal
distribution function describes the home range.

Answers to two questions will be sought with reference to the relative
distance between home range centers:

1. How does distance between home range centers affect the impact of
the community on the environment?

2. How does the interval between home range centers affect the proba-
bility of one individual meeting or detecting its neighbors?

The impact of an individual upon its environment should be proportional

J. The Social Use of Space

to the amount of time per unit area it spends at successive distance from
its home. This relative impact is described by the bivariate normal dis-
tribution function (Fig. 2). For any particular animal this means that at
3 sigma distance from its home its impact per unit area will be only 0.011
of what it was adjacent to its home. We may visualize the impact of any
one animal upon its enviroimient as having a mountain-shaped topography.
Where home range centers are at least 6 sigma apart, there lies between
them a "valley" where neither animal has a significant effect on the en-

IMPACT ON THE ENVIRONMENT
(PER UNIT AREA)

>-
z
< o

o °^.

I- I

Ul to

> K

-I <

UJ UJ

Q: z

i<
"* 1

> z

CD <

0.4 I —

-• /.5tr

• • » »

-• • 2./<r

3.0

0 0.6 1.2 1.8 2.4

o- DISTANCE FROM ANY H.R.C.

Fig. 19. Impact on the environment (per unit area). The value 1.0 represents the
effect one individual will have near its home range center. Since home ranges increasingly
overlap as their centers approach each other, i.e., density increases, the summated
impact of all animals on any one point not only increases, but the relative impact on
all points becomes more nearly equivalent.

vironment. As soon as home range centers get closer than 6 sigma to each
other, the home ranges overlap and neighbors can both affect those por-
tions of the environment falling within both neighbor's home ranges. As
soon as home range centers become less than 3 sigma apart, some portion
of the environment can be affected by more than two individuals.

At any point in the environment, the impact of all animals which can
arrive at that point during this normal ranging about their home is pro-
portional to the sum of their separate density functions at that point.
Utilizing the normative data of density function as a function of the sigma
radius from home, given in Table 2 of Calhoun and Casby (1958), several
curves of summated density function were calculated (Fig. 19).

44 John B. Calhoun

E. Methods of Calculating Data Relative to the Distance between
Neighbors

In Fig. 16, which illustrates the uniform distribution of an alpha species
and of a beta species lying in the interstices between the home ranges of
the alpha individuals, we can select any single alpha individual and note
certain characteristics of the geometric distribution of its alpha neighbors.
One such individual, whose home range center is indicated by a triangle,
is shown in Fig. 16. A line drawn between the home range centers of its
adjoining nearest neighbors forms a hexagon about this individual. Just
as there are six nearest neighbors, there are twelve next-nearest neighbors.
Lines connecting the home range centers of these next-nearest neighbors
also form a hexagon. Successively more distant neighbors form concentric
hexagons, each containing six more indi\nduals than the next innermost
hexagon. For the purpose of investigating the effect of neighbors on each
other or upon the environment, a system of four "concentric" hexagonal
sets of neighbors was prepared on a large sheet of graph paper. This pro-
cedm'e was repeated three times, forming spatial sets in which home range
centers between nearest neighbors were respectively 1.5, 2.1, and 2.7
sigma apart. Ruler scales representing density functions (Table 2, Calhoun
and Casby, 1958) at successive sigma distances from the home range
center, as well as ruler scales representing intensity of signal (Fig. 18)
were prepared. Using these ruler scales, several types of events were calcu-
lated with regard to their changes in intensity or freciuency along a 3-
sigma route such as is shown by the heavy dashed line in Fig. 16.

At each of eleven points along this typical route of travel, a sum of the
density functions of all neighbors whose home ranges overlapped one or
more of these eleven points was calculated (see Fig. 19).

F. Further Comment on the Impact of All Individuals on the Environment

Each of these sums of density functions were divided by 0.159, the rela-
tive density function of an animal near its own home range center. By so
doing we can obtain a fairly good idea of the impact of all individuals who
may arrive at any particular point with reference to the effect that one
individual would have near its home range center. It may be seen that
when home range centers are 2.7 sigma apart, considerable inequality
between points exists. In other words, points near home range centers are
relatively intensively used in comparison to distances about halfway
between home range centers. This inequality of usage of the environment
is even more pronounced when home range centers are more than 2.7

1. The Social Use of Space 45

sigma apart. However, by the time home range centers are uniformly dis-
tributed at 2.1 sigma apart, all portions of the environment are approxi-
mately equally utilized although every point is more intensively utilized
than when home range centers were farther apart. Every further increase
in density, as represented by home range centers coming closer together,
merely increases the intensity of usage of every part of the environment and
all parts continue to be equally utilized.

These curves (Fig. 19) are particularly instructive in gaining an insight
into an "ideal" interval between home range centers. It is logical to assume
that portions of the environment which are less utilized than others serve
as a trap to catch wandering individuals who have not yet established a
home range. As long as the process of eciualizing distance between adjoining
home range centers continues, no remaining pockets of less utilized habitat
will occur by the time adjoining home range centers are nearly 2.1 sigma
apart.

Any increase in density, that is any shortening of the interval between
home range centers below 2.1 sigma, will merely increase the probability
that available objects will be overutilized.

At the maximum interval between HRC's at which uniform utilization
of the environment arises, aggressi^'e actions exhibited by individuals with
resident home ranges may be expected to prevent excess members of the
populations from settling down within such an established area. Such
wandering individuals may be expected to wander through and out of such
established areas and into marginal habitats.

Thus, if minimizing the opportunity for aggressive encounters and the
development of a uniform utilization of resources represent forces affecting
evolution, we may anticipate development of capacities for communication
which will most readily assure that the members of a population of a single
species will be able to distribute themselves uniformly through space with
an approximate 2.0 home range sigma distance intervening between any
two adjoining home range centers.

G. Contacting Neighbors

The product of the density functions of any two individuals at a particu-
lar point determines the relative probability that these individuals will
meet by chance. Similarly, the product of the density function of any one
individual at a point with the sum of the density functions of all other
individuals determines the relative probability that this one individual
will contact neighbors at that point. Such latter calculations were made
for uniformly distributed home range centers at 2.7, 2.1, and 1.5 sigma

John B. Calhoun

(see Fig. 20) . As might be expected, as home range centers get closer to-
gether the probability of any individual contacting neighbors at every
distance from its home range center increases. There may be some fre-
quency of contacting neighbors which becomes so unbearable to the in-
dividual that his resultant aggressive actions prevent further contraction
of the interval between home range centers. However, we have no basis
for gaining insight as to what this frequency might be.

0.1 r-

•- -^ .01

L. <

O t-

-I q:

m "^

.001

.0002 ■-

0.6 1.2 1.8 2.4 3.0

<r DISTANCE FROM ANY H.R.C.

Fig. 20. Contacting neighbors. The relative probability of any one individual meet-
ing others with reference to its distance from its own home range center, and to the
distance between home range centers of all individuals.

One characteristic of these curves of relative probability of contacting
neighbors does lend itself to suggesting a condition leading to an optimum
interval between home range centers. When HRC's are 2.1 sigma apart
the probability of contacting neighbors up to half the distance between
home range centers is for all practical purposes constant. Applying this
insight to all members of the population, it is apparent that with HRC's
this distance apart, the probability of contact between neighbors becomes
relatively constant everywhere. If we accept the principle elaborated by

1. The Social Use of Space 47

Fredericson (1951) that animals attempt to make their environment
predictable, and if we accept constancy of consequences as assuring greater
predictability, then it follows that where the members of a community
have their home range centers approximately 2.1 sigma apart, greatest
predictability with regard to contacting neighbors characterizes this
interval.

H. Sign Field of AlljNeighbors

Urination, defecation, and activities relating to the removal of materials
used for food or nests represent signs by which one individual might recog-
nize the presence of neighbors. Signs left by neighbors may be expected to
be proportional to the sum of the density functions of neighbors at points
considered. Unless home range centers are extremely close together, there
will arise a noticeable increase in signs of neighbors as the individual moves
away from its own home range center. Obviously, the closer home range
centers are to each other the relatively greater will be the sign of neighbors
at any particular radius from the individuals's own home range center.
We may then wonder what standard a particular individual may utilize
in judging the intensity of signs left by its neighbors. Any individual's
own sign is maximal near its own home range center. Therefore, an in-
dividual may resort to comparing the relative amount of sign of neighbors
at any point to that which it would leave in a similar area near its own home
range center.

Dividing the sum of the density functions of all neighbors at a particular
point by the density function of a particular individual near its home range
center provides such an index of the relative intensity of sign of neighbors
(Fig. 21). The optimum interval between home range centers with regard
to the sign field should be that interval at which throughout the home range
of a particular individual the total sign left by neighbors nearest approxi-
mates that individual's own standard and in which there is greatest pre-
dictability with regard to sign, that is in which there is the least variation
in intensity of neighbors' signs from point to point. Judging from the three
curves presented in Fig. 21 an inter-home range center interval of some-
where near 2.1 sigma would lead to the development of an optimum sign
field.

I. Signal Field of Neighbors

Following the formulation presented in Section VII, C, it is assumed
that the signal emitted by one individual can be perceived by another in-

John B. Calhoun

dividual with sufficient intensity to produce a response by the latter in so
long as the indiA'iduals are separated by a distance no greater than 3 home
range sigma. In all probability, signals in the sense of vocalizations are
emitted by each individual periodically as they wander through their
home range. In order to simplify calculation of the signal field of neighbors,
the particular condition was taken where all signals are emitted only from
the home range centers. Thus, along a typical route of travel, as shown by
the heavy dashed line in Fig. 1(), the sum of the intensity of signals from
all neighbors was calculated.

0.6 1.2 1.8 2.4

a DISTANCE FROM ANY H.R.C.

Fig. 21. Sign field of neighbors. Signs are considered as any persisting indication of
an animal having made a response, i.e., defection, gnawings, or removal of food items.
Thus Fig. 21 essentially represents the subtraction of the density function of one indi-
vidual from the sum of the density functions of all individuals as shown in Fig. 19.

Again, we might wonder what standard the individual might utilize in
judging the total intensity of signals received. Since the learning of the
signal presupposes emission by one individual and detection by the other
when they are in contact, this level of intensity with an assigned value of
1.0 may be taken as the standard. Since the intensity of signals probably
drops off inversely proportional to distance, the sum of signals at any
point in place and time may be less than 1.0. A further complication to the
problem is that all neighbors may not emit signals simultaneously. Simul-
taneity other than by chance will arise regularly only if the detection of the

1. The Social Use of Space

signal by one individual elicits a similar response by the perceiver. However,
if this is so, and if each individual after emitting a burst of sequential
signals enters a refractory period (see Section XIII, A) of some given
mean length before it can emit signals again, then we have a situation in
which there occur recurrent periods during which most nonsleeping in-
dividuals in the community emit signals nearly simultaneously. Although
no proof of the validity of this assumption can be offered at present, my
formulation will accept the existence of such a process. Such an assumption
is inherent in the utilization of P'ig. 22 in arriving at some insight as to the

2.0

1.0

0.3 >—

/.5ff-

0.6 1.2 1.8 2.4

o- DISTANCE FROM ANY H.R.C.

3.0

Fig. 22. Signal field produced by neighbors. If all neighbors emit signals simul-
taneously at their home range centers, and these signals have the properties shown in
Fig. IS, then their summated intensity will form a "topography" as here shown with
reference to any particular animal moving through its home range. The value 1.0 repre-
sents the intensity of a .signal at the point of emission.

influence of the signal field on determination of an optimum interval be-
tween home range centers.

As with the sign field of neighbors, the signal field of neighbors is pre-
sumed to influence interval between home range centers through the mem-
bers of the community seeking that interval between home range centers
which will ensure most closely attainment of both constancy of the signal
field, leading to predictability and to approximation of the standard signal
intensity. Judging by the three curves in Fig. 22, operation of these two
criteria indicates an optimum interval between home range centers slightly
less than 2.1 sigma.

John B. Calhoun

J. Hum Field

In Section VII, C it was suggested that when the emitter is between 3
and 6 sigma from the receptor the signal given by the emitter can be per-
ceived by the receptor but is insufficient in strength to elicit the appropriate
responses. The sum of all such signals below threshold for inducing a
response is here termed "hum." When this value exceeds the standard
intensity of 1.0 (see Section VII, I), the receptor will become restless even
though perhaps not exhibiting a specific response to the signals. What
effect this general state of restlessness may have upon the receptor is un-
known, but it is logical to assume that the members of the community

3.0 I — •_

> 20

> 1.0

t-
<

LiJ

0.4

_• ■ a-

/.5 a-

2.IO-

27a-

\
Distance

between

H. R. C. 's

0.6 1.2 1.8 2.4

a DISTANCE FROM ANY H.R.C.

Fig. 23. Hum field refers to the sum of the intensity of all signals, any one of which
is below that minimum intensity required to elicit a specific response. The value, 1.0,
denotes the intensity of a signal at the point of emission.

will attempt to adjust the interval between the home range centers in such
a way as to reduce the likelihood of the "hum" exceeding 1.0. To do so
implies that home range centers must be of the order of 2.4 sigma apart
(Fig. 23).

K. General Conclusion Concerning the Distance between Neighbors

Effective learning of signals probably does not begin until home range
centers come at least 3.0 sigma from each other, and even further shorten-
ing of this interval must increase the effectiveness of learning. Several

1. The Social Use of Space 51

factors examined all suggest that the optimum interval between home
range centers should be of the order of a 2 home range sigma distance.
From an evolutionary standpoint, the most important of these probably is
the fact that utilization of the environment becomes uniform at slightly
more than 2 sigma distance between home range centers. This factor should
have been the major one in the evolution of those characteristics of in-
dividuals pertaining to the frequency of trips, the velocity of the individuals,
and the aggressive acts elicited when two individuals meet. It is probably
strictly coincidental that the properties of the sign and signal field are also
likely to be such as to make an interval of about 2 sigma between home
range centers optimum. It is at this distance between home range centers
that both signs and signals become most constant and thus more pre-
dictable. Also, the intensity of the signs and signals most nearly approxi-
mate any receptor's own behavior near its home, which can serve for it as a
standard in evaluating the intensity of the actions of its neighbors impinging
on it.

L. The Number of Neighbors Perceived

When home range centers are 2.7 sigma apart, an indi\ddual at its home
range center can just perceive all its six nearest neighbors. However, as it
begins to move away from home it begins to lose contact with those nearest
neighbors lying on the opposite side of its home range center. By the time
it reaches half the distance to its nearest neighbors in the direction in
which it is traveling, it can detect only three of its nearest neighbors.
Toward the periphery of its home range, it can detect only two of its nearest
neighbors. In addition, it can detect one, and only one, of its next-nearest
neighbors, w^hich are members of the group of twelve forming the second
hexagonal tier of neighbors about its home range center. Thus, with this
fairly large interval between home range centers, any one individual has
poor contact with its associates. When home range centers are 2.1 sigma
apart, any one individual can maintain contact with all six of its nearest
neighbors out to about 0.7.") sigma. Even when an individual has journeyed
halfway in a direction of a nearest neighbor it is still in contact with four
of its six nearest neighbors, in addition to one of the twelve neighbors
lying in the next tier. As he proceeds still farther, losses of nearest neighbors
are compensated for by next-nearest neighbors. Thus, at all times when
home range centers are 2.1 sigma apart, an individual is in contact or po-
tential contact with five or six of its neighbors.

By the time home range centers are of l.o sigma apart, an individual is
in potential contact with ten other indi\'iduals though not always the same

52 John B. Calhoun

ten, no matter where it is in its home range. Since the dynamics of the use
of space relating to uniformity of utihzation of resources and the character-
istics of the sign and signal field all point to an optimum interval between
home range centers of somewhere near 2 sigma, it follows that there should
have been evolutionary adjustment of tolerance to simultaneous or near
simultaneous communication with five to ten others.

VIII. Interpretations of Observed Data Derived from Removal Trapping of

Small Mammals

At this stage in the development of a concept of community organiza-
tion, one must resort to a certain amount of quasi circular reasoning.
Regularities in observed results lead to theoretical formulations. Then these
formulations can be used to reexamine the data for further insight. This is
my present intent. In time, many aspects of the concept may be subjected
to more rigorous study. However, for the present we must content ourselves
with a search for a best approximation to a very complex set of phenomena.

Section VI, "Continuous Removal Trapping of Small Mammals,"
presented results from several extensive studies. Specific interpretations
follow\

A. The Relationship between Two Dominant Species

The dominant species in the community reveals itself during removal
trapping through its members having such large home ranges that every
individual living near a trap has a high probability of encountering it.
Thus, for them, few^r days lapse from initiation of trapping until 50% of
the population is caught. As can be seen from the two Alaryland studies
presented (Fig. 10 and Fig. 13B), Peromyscus fulfills this criterion. For
species associated with Peromyscus, whether they be Blarina and Sorex, or
Blarina and Pitymys, the dates of 509(' removal arrive much later. The
later the date of 50% removal, the more subordinate a species, and the
more slowly its members expand their home ranges as the dominant
species is removed.

In more northern forest habitats, Peromyscus rarely is found in the
absence of Clethrionomys. In fact, it is as if the red-backed mouse is just
superimposed upon the simpler Peromyscus-Blarina- Sorex community of
more southern forests. Typical dominance of Clethrionomys over Peromyscus
may be seen in Figs. 9, 11, and 13A. Although I am convinced that
Clethrionomys usually can develop the ability to inhibit the home range of

1. The Social Use of Space 53

Peromyscus, there exist conditions limiting the extent to which this abihty
may develop.

The typically low relative density of both species (Table VI) on Alt.
Desert Island, Maine, represents such a condition. As revealed in Fig. 14,
both species decline at about the same rate from day 1 through day 3.
This can happen only when no alteration in home range size transpu-es
over time or when the survivors of each species make equivalent but slight
increases in extent of home range. Appreciation of why inhibition of home
range fails to develop at low densities demands knowledge of variables we
lack.

It demands that we know actual densities. The NACSAI census provide
only relative densities. However, we can make approximations. Run long
enough (30 days), the 950-foot-long B-type NACSM census procedure
should take all residents wdthin 3 home range sigma on either side of the
trapline and for a radius of this distance about the end of the line. If we
take 50 feet as approximating the average home range sigma of small
mammals, uninhibited by dominants, then approximately 8 acres are ex-
posed to such a trapline. Furthermore, examination of 30-day censuses re-
veals that for species with uninhibited home ranges, 25-50% of the
residents are taken during the first 3 days of trapping. On this basis, there
was on the average less than one Clethrionomys and less than two Peromys-
cus per acre in this Maine study. Since juveniles, with as yet probably little
influence on the spatial distribution of associates, comprised a portion of
the catch, it is quite likely that the average distance between home range
centers for each species exceeded 3 sigma. As shown in the prior theoretical
sections, learning of signals would most likely be fairly ineffective here
because contact between neighbors w^ould be infrequent.

Under these circumstances the signals emitted by each species should
have acquired little in the way of negatively stimulating characteristics
for its own members. It follows that inhibition of home range size will
have been negligible and thus the two species, which probably have nearly
the same size home range, should encounter traps with nearly equal fre-
quency, and thus the rate of decline in catch from days 1 through 3 should
be nearly equal. However, in the one Mt. Desert Island, Maine, study where
eight NACSM lines were run for 15 days (Fig. 13A), it is apparent that
home ranges of Peromyscus were slightly contracted. Fifty per cent of the
15-day total for Clethrionomys was attained by day 6, but not until day 10
for Peromyscus. Thus, where both species regularly occur at low densities,
Peromyscus is only moderately subordinate to Clethrionomys.

In the Adirondacks, where Peromyscus most frequently has a low density
and Clethrionomys a much higher one (Table VI), Peromyscus is markedly
subordinate. Its home ranges not only are markedly contracted, but also

54 John B. Calhoun

they expand immediately as Cleihrionumys are removed. The vahdity of
this interpretation is revealed in Fig. 14B, which shows that during the
first 3 days of trapping as the Clethrionomys population is reduced, as in-
dicated by fewer numbers taken on successive days, the numbers of Pero-
myscus taken increases. This increasing catch can only result from sufficient
expansion of home range by survivors to bring about an increased prob-
ability of encountering traps.

One might argue that these latter data for Pewmyscus merely indicate
that the snap-trap is initially a sufficiently strange object to elicit avoidance.
As time elapses these mice become accustomed to the presence of the trap
and thus later in time more individuals will enter traps. Two lines of evi-
dence of reasoning suggest the fallacy of this interpretation. If it were cor-
rect, we must conclude that Pewmyscus in Maine lack this strange object
response but those in New York have it highly developed (see Fig. 14).
There exists no logical basis for believing that such a difference character-
izes the populations of these two areas. Furthermore, in the trapping of
both Pewmyscus and Clethrionomys it is not an uncommon experience to
find either of these mice dead in a trap with a bloody stump of one hind
leg, while another trap, 2-5 feet away, is covered with fresh blood and fur.
The conclusion as to what happened is clear. The mouse happened to get
caught in one trap by one leg, it chewed or pulled itseh loose, then went
fairly directly to another trap, bit at the bait on the treadle and was thus
killed. If these mice have a strange-object reaction it must be of a suffi-
ciently low order of magnitude that even the recent loss of a leg in one trap
fails to increase it to the point of avoiding the next trap encountered.

In the third area. Alberta, from which adequate data are available for
these two genera, both are relatively abundant but Pewmyscus exceeds
Clethrionomys (Table VII). Dming the fall season, nearly twice as many
of each genus are trapped as during the spring. During the period of low
spring densities, these genera exhibit nearly identical rates of decline
(Fig. 15 A) accompanying removal trapping. As with the Maine data, such
trends may be interpreted as indicating that at such densities neither
species is capable of markedly inhibiting the home range of the other.
However, by fall many Clethrionomys have contracted their home ranges
as a response to their exposure not only to more of their own kind but also
to more Peromyscus. Home ranges of Peromyscus remained unaltered, as
indicated by the similarity of rate of decline during both spring and fall
(Fig. 15). However, the daily catch for Clethrionomys remained nearly
identical through three successive days of removal trapping. Sufficient en-
largement of home ranges by surviving Clethrionomys must have taken
place each day to lead to an equivalent frequency of traps being encountered
on the following day bj^ red-backed mice, despite their fewer numbers

1. The Social Use of Space 55

than on the preceding day. At this Alberta site, the inhibition of Clethri-
onomys home ranges by Peromyscus must have been less than the inhibition
of home ranges of Peromyscus by Clethrionomys in the Adirondacks.

Despite the paucity of areas from which extensive comparable data are
available, it looks as though Clethrionomys has a slight advantage over
Peromyscus in gaining psychological ascendency. At this point one may
suspect that the characteristics of some environments will markedly favor
the reproduction and survival of one of these genera. Whichever genus
this happens to be will then become psychologically dominant to the other,
as evinced by the contraction of home range of the less numerous genus.
Consideration of other aspects of the relationship between these two
genera requires familiarity with the concept of the constellation dealt
with in the following sections.

B. Removal Captures of Socially Dominant Species

In several studies already presented (Figs. 9-13) we have seen that one
species tends to be caught in large numbers during the first few days, and
that the time of maximum input for the remaining species comes during a
successively later period. Very frequently a secondary increase in daily
catch starts near the 1 oth day of trapping for the species with initially the
greatest rate of capture. Such species will henceforth be designated as the
socially dominant or alpha species of the small mammal community. For
example, see Cleihrionoviys in Fig. 9 and Perotnyscus in Fig. 10. Where
there are several species taken, usually only one is characterized by this
secondary input. Four censuses examined included one such species and
a fifth included two with definite secondary inputs.^

Although several species are involved in this phenomenon, the assump-
tion is here made that they all so behave because of similar properties
leading to their alpha rank in the community. If this is so, we are justified
in pooling the data. A table of the total catch per day of trapping was

^ The five censuses utilized in preparing Fig. 24:

1. By Dr. J. E. Moore, Sept. 1959, Edmonton, Alberta: 128 Peromyscus maniculatus
borealis, 65% of 30-day total by day 14-15.

2. By Dr. A. I. Roest, Oct.-Nov. 1959, San Luis Obispo, California: 75 Dipodomys
heermanni, 65% of 30-day total b}^ day 15-16.

3. By Dr. William L. Webb, Fall 1952, Rich Lake Island, Newcomb, New York:
97 Clethrionomys g. gapperi, 65% of 30-day total by day 7.

4. By Dr. Earl F. Patric, Fall 1953, Arbutus Area, Newcomb, New York: 86
Clethrionomys g. gapperi, 65% of 30-day total by day 12-13.

5. By Dr. Kyle R. Barbehenn, Nov.-Dec. 1959, Chadwick Woods, Montgomery
County, Maryland: 57 Peromyscus leucopus, 65% of 30-day total by day 14;
116 Blarina brevicauda, 65% of 30-day total by day 20-21.

John B. Calhoun

prepared. These data are shown as a three-point moving average in Fig. 24.
There results a continuous decHne in catch until about the 12th day.
After this, the daily rate of capture increases, reaching a second maxi-
mum five to seven days later. Following this, the number of animals taken
continuously declines, but it is not until about the 27th day of removal
trapping that the numbers taken per day reaches the low level character-
izing the 12th day.

1 1

MAINLY aa/3

1 1 1
MAINLY Y

INDIVIDUALS

g 50

S 40
o

1 •

•

o 30
IT

\»

§20

\ *
^

• •
7 • *N,^

•X.

1 1

1 1 1

0 5 10 15 20 25 30

DAY OF TRAPPING

Fig. 24. Removal captures of 559 small mammals who are representatives of species,
socially domiuant in their community. Alpha and beta represent the intraspecific domi-
nant individuals with large home ranges. The gamma individuals represent the intra-
specific subordinates which enlarge their home ranges following removal of the alphas
and betas.

If the assumption that practically all the resident population exposed to
the traps is removed diu-ing 30 days of trapping is correct, then the accu-
mulated catch plotted as the proportion of the 30-day total over time will
reveal both the proportion of the total comprising the initial input and the
time at which the secondary input begins (Fig. 25). This shows that the
intersection of the two rates of input occurs on day 15 after removal of
64% of the resident population.

1. The Social Use of Space

C. Constellation Formation — An Intraspecific Phenomenon

After the removal of 64% of the resident population of socially dominant
species, whose home ranges are sufficiently large to give them a higher
probability of encountering traps, there arises a secondary input. These
latter individuals must have enlarged their home ranges as a response to
the absence of their former associates. It follows that certain individuals

i 1

' ' .V"*'^

< 0.9

1 —

j§ —

g 0.8

V»

¥ 0.7

Ll.

Z 0.6

— ^y^

jm^

(£

S 0.5

_ /•

v) 0.4

— 1

—

< 0.3

_ L

So.

-A

o 0.1

t—

1 1

I 1

15
DAY

Fig. 25. Removal capture.s of socially dominant species. Same data as in Fig. 24,
but here shown as an accumulated total.

have the ability to restrict the home ranges of other members of their own
species. At this point in my analysis I began to wonder whether there were
circumstances relating to the spatial distribution of the population which
might lead to a 64:36 ratio of larger : smaller home ranges.

In the previous discussion we have seen that one possible spatial equilib-
rium is that in which home range centers become uniformly distributed
with an inter-home range center interval of about 2.0 sigma. It was further
pointed out that the mechanics of moving through space are such as to
favor the evolution of a signal emitted by one individual which has the

John B. Calhoun

potentiality of eliciting a response by another up to a maximal distance of
about 3 sigma between individuals. We may then ask, "How many other
individuals in such a uniformly distributed population can any one in-
dividual know?" Figure 26 shows that a line connecting the home range
center of any one individual's six nearest neighbors forms a hexagon. With
the hypothesized communication system operating, this one individual
can remain in constant communication contact with all its six nearest
neighbors except under the condition when it and one of them happen to

Fig. 26. Spatial distribution of home range centers during the preconstellation phase.
Small open circles represent centers for individuals destined to become alpha members.
Dots represent home range centers for all other individuals. Home range centers are
uniformlj' distributed.

move in opposite directions away from the line connecting their respective
home range centers. Such actions can increase the distance between them
to greater than 3 sigma.

Lines connecting the home range centers of this individual's next-nearest
neighbors also form a hexagon. Home range centers for all these twelve
next-nearest neighbors lie nearly 4.0 sigma from the selected individual.
Therefore, each of these next-nearest neighbors can be in contact with the
selected individual when they approach each other, such that the distance
intervening between them becomes less than 3.0 sigma. Any more-distant
neighbors, those living G.O or more sigma from the selected one, will have

1. The Social Use of Space 59

such a low frequency of communication with the selected individual that
mutual effects will be of little significance.

Examination of the spatial distribution of home range centers in a uni-
formly distributed population reveals that one-twelfth of the members
each have six nearest neighbors whom they can influence without competi-
tive influence being exerted by any other individuals in this one-twelfth
population. Such individuals will hereafter be designated as alpha indi-
viduals. Home range centers for these alpha individuals are designated by
small open circles in Fig. 26. It will further be noted that two such alpha
individuals, who live closest to each other, share next-nearest neighbors.

Antagonistic relations, uniform utilization of the environment, and
attainment of uniform sign and signal fields approximating in intensity
the sign and signal intensity expressed by an individual at its own home
range center, and eciualization of contact rate with all neighbors throughout
any one individual range, all serve as forces leading to a uniform distribu-
tion of home range centers at near 2.0 sigma between centers (Section

VII, K) . Yet, we may anticipate an opposing force of attraction between
members in close communication with each other. This will lead to all
nearest neighbors shifting the centers of their home ranges slightly toward
their single alpha associate. Such nearest neighbors will hereafter be
designated as beta individuals.

Now we can turn our attention to the alpha's next-nearest neighbors.
It may be seen from Fig. 26 that in each group of twelve next-nearest
neighbors, six are eciuidistant to two neighboring alphas. Therefore, since
the probability of each of these six nearest neighbors is 0.5 of being at-
tracted to either of two alpha individuals to which they are eciuidistant,
each alpha on the average will have attracted to it three of these next-
nearest neighbors. The remaining six next-nearest neighbors to any alpha
are ecjuidistant to three alphas. It similarly follows that on the average
any one alpha will have attracted to it two of six such next-nearest neigh-
bors. Next-nearest neighbors w^hich so move toward an alpha associate
will hereafter be designated as gamma individuals. Thus, each alpha will
have attracted toward it six beta individuals and on the average five
gamma individuals. This process leads to a clumping of the population
into groups ranging in size from 7 to 19, with a mean of 12 (see Section

VIII. D).

The central alpha indi\'idual because of its more favored position in the
communication network may be assumed to be dominant to both its beta
and gamma associates. Gamma individuals, because of their peripheral
location in the developing clump, here designated a constellation, will be
subordinate to both the beta and alpha associates. Now let us consider the

60 John B. Calhoun

situation where home ranges contract as individuals become more
subordinate.

How much beta and gamma members may be expected to contract
their home ranges requires consideration of phenomena treated in Section
XIV, E, titled "Velocity and Home Range." There, it is shown that in
terms of o-„ units of distance, the home range sigmas for alpha, beta, and
gamma individuals become, respectively, 1.0, 0.9575, and 0.6457. Since
an optimum uniform utilization of the environment requires a 2a interval
between home range centers, the distance between alpha and beta home
range centers wall become 1.9575o-a, and 1.6032o-a between beta and gamma
home range centers.

The periphery of each constellation (Fig. 27) may be arbitrarily desig-
nated as that radius from the center of the alpha member's home range
extending to one aa beyond the center of each gamma individual's home
range. This radius is 3.1478a-a. And since the home range centers of the
alpha members of adjoining constellations are 6.92o-a apart, it is obvious
that between constellations there lies what may be called an interconstella-
tion matrix, receiving very little usage from the dominant species forming
constellations. Furthermore, this interconstellation matrix must markedly
reduce communication between members of adjoining constellations. How
extensive this reduction becomes must be viewed against the requirement
of a 2a inter-HRC interval for an optimum state. Yet, the HRC of any
beta or gamma member of one constellation will lie on the average 4a units
of distance away from the nearest beta or gamma members of an adjoining
constellation with reference to their own reduced home ranges.

On the average, such constellations will consist of twelve individuals, 1
alpha, 6 betas, and 5 gammas. Of these, the gammas with smallest home
ranges form 0.417 of the population. These are the individuals who, after
removal of their dominant alpha and beta associates, will enlarge their
home ranges. Actually, some gammas will be trapped before the 15th day
and some alphas and betas will be taken after it. Furthermore, Fig. 25
clearly shows that a few more gammas would have been taken after the
30th day of trapping. Therefore, the observed proportion of the population
consisting of gamma individuals, estimated at about 36%, reasonably well
approximates the theoretical expected of 41.7%-

To date, this approximation stands as the only direct supporting proof
(see Section XI) that populations of dominant species of small mammal
communities do, in fact, tend to form such constellations. If they really
do, we may anticipate that there has been evolution relative to physiology
such that its optimum state is most compatible with interactions among
individuals transpiring in groups with a mean size of twelve adults. In

7. The Social Use of Space

later sections, I shall present data which support the hypothesis that a
group size of twelve has been phylogenetically retained in more highly
evolved forms.

For the present, let us examine fm'ther implications of constellation
formation upon the structure of the small mammal community. Con-

FiG. 27. The theoretical constellation phase of intraspecific community organization
of dominant species. A 3-sigma radius home range for alpha individuals is z-epresented
by the large circle. Small circles represent 1-sigma radius portions of all members of a
constellation. Contraction of home range by beta and alpha members permits the more
intensively used portion of the home ranges of all individuals to fall mostly within the
3-sigma radius home range of the dominant alpha members. Crosshatched circles
represent the 1-sigma portion of the home range of alpha individuals.

62 John B. Calhoun

stellatioii formation presupposes passing through a stage of uniform dis-
tribution of home ranges. In the interstices between the home ranges of
alpha species there is the opportunity for very subordinate species with
highly contracted home ranges to establish themselves. As the constellation
forms, certain members of the subordinate species will be retained within
the interstices of the constellation while others will lie at its periphery.
However, as constellations form, there develops an interconstellation
matrix rarely frequented by members of the alpha species. We may then
wonder what members of the community will live here.

At this point, the Huntington Forest censuses of 1952 and 1953 (Fig. 11)
are particularly instructive. The red-backed mouse, Clethrionomys, is ob-
viously the alpha species. Just as obviously, the shrew Sorex is the most
subordinate species, which has highly contracted home ranges lying in the
interstices between those of Clethrionomys. Had all these census been run
for longer than 30 days, it looks as though the 2 : 1 ratio of the alpha species
to one of its most subordinate associates would have been realized. Censuses
shown from other areas indicate that both Peromyscus and Blarina are
dominant to Sorex in the community. And yet, in these Huntington
Forest censuses, relatively few of either were caught. What I suspect has
happened is that as the constellations developed, both species were relegated
to the interconstellation matrix, where they existed in somewhat contracted
home ranges, as indicated by the catch per day increasing from day 1 to
days 4-6. In fact, for both these species it looks as though a certain portion
of their members had even more markedly contracted home ranges. This
is indicated by the secondary increase in catch per day starting betw^een
days 8 and 10.

D. Expected Variability in the Number of Individuals Forming
Constellations

In terms of the formulation of Section VIII, C the six nearest neighbors
to any alpha individual will always be attracted toward it as the members
of a population wuth a uniform distribution of home range centers begin
to form diffuse clumps, termed constellations. The twelve next-nearest
neighbors to any alpha individuals may be divided into two types, a and h,
according to their probability of being attracted toward any given alpha
neighbor.

Let: a represent the six individuals who have a probability, pa = 2, of
being attracted to any given alpha.
h represent the six individuals who have a probability, 'Pb = k, of
being attracted to any given alpha.

1. The Social Use of Space

The probability, Pa{i), that i number of the six a individuals will be at-
tracted to the given alpha will be :

Pa(i)

(A /I

(28)

Similarly the probability, pb(i), that i number of the six b individuals will
be attracted to the given alpha wdll be:

Pb(i) =

il \3/ \3

' /9\^— *

(29)

Therefore, the probability, p{m), where m = 0 to 12, of m members of
the 12 next-nearest neighbors joining any given alpha and its six nearest
neighbors to form a constellation will be:

p{m) = ^ Paii) ' Pbim - i)

(30)

For example, if m = 4 this becomes:

P(4) = PaiO) ' Pbi^) + Pail) ' Pt(:i) + Pa{'2) ' p,{2) + p„(3)

• P5(l) +p„(4) . PbiQ) (31)

p(S)) • • • p{12) were calculated and are shown in Table VIII as p{7 + 0)

Table VIII
Expected ^'ARIABILITy ix the Number of Individuals Formixg a Constellation

Xumbor of animals in group

Prol:)abilit3' of group developing

9
10
11
12
13
14
15
16
17
18
19

0.00137200
0.01234568
0.05041152
0.12345679
0.20190329
0.23225308
0.19266546
0.11612654
0.05047582
0.01543210
0.00315072
0.00038580
0.00002143

Z 1.00000023

64 John B. Calhoun

• • • p (7 + 12) since the constellation becomes composed of the addition
of these next-nearest neighbors, or gamma individuals to the basic core of
the one alpha and its six beta nearest neighbors.

E. Social Rank and Intraspecific Associations

In the 14-day removal study shown in Fig. 13B, trapping results indi-
cated large home ranges for Peromyscus, moderate-sized ones for Blarina,
and highly contracted home ranges for Pitymys. Estimated home range
sigmas, o-, of 50, 25, and 12.5 feet, respectively, probably closely enough
approximate the real values to permit their use in a study of these data.
If we knew the actual home range centers of all animals trapped, a 1-sigma
radius circle plotted about each on a map of the study area should pro\'ide
more insight into spatial relationships.

Since such centers were not known, they were approximated by making
the following assumptions:

1. Peromyscus maximize inter-home-range-center distance from other
Peromyscus.

2. Blarina similarly maximize distance, not only from others of theu'
own kind, but also from Peromyscus.

3. Pitymys maximize distance from both Peromyscus and Blarina, as
well as from others of their own kind.

4. The later an animal was trapped the farther its home range centers
were located from the trapline.

For Peromyscus some of the HRC's were shifted to one side of the line
of capture, and the remaining to the other side, until ever}'' adjacent 3
HRC's approximated equilateral triangles. A similar procedure was ap-
plied to Blarina captures except that where possible their HRC's also
were placed in the center of triangles formed by the HRC's for Peromyscus.
Then the HRC's for Pitymys w^ere, insofar as possible, placed in the centers
of triangles connecting the HRC's of the other two species.

Utilizing these assumptions, centers of home ranges for Peromyscus
were plotted first, Blarina second, and Pitymys last. Originally these were
all plotted on a single figure, but for the sake of greater clarity in examining
intraspecific relations the home ranges, in terms of 1-sigma circles, are
shown separately in Figs. 28-30.

Moisture conditions in the habitat varied markedly. Steep xerophytic
slopes of oak and pine covered the three sectors encompassed between the
E and SW radii. The sector between the N and NW radii was quite steep
and dry. A damp drainage area lay roughly along the W and NE radii.

L The Social Use of Space

Fig. 28. Schematic home ranges for Peromyscus. The octagons represent traplines
along which removal trapping was conducted.

<r«25 FEET N

Fig. 29. Schematic home ranges for Blarina.

John B. Calhoun

Away from the nari-ow drainage area the other four sectors were moderately
mesophytic. Deep leaf mold covered the entire forest floor. No attempt was
made to obtain a detailed cover map. Total timbering a year later for a
housing development revealed that many of the trees in the forest exceeded
150 years in age.

It may be noted that both Peromyscus and Blarina tended to avoid the
more xerophytic areas, whereas Pitymys was more abundant in these
drier areas. However, it is interesting to note that no Blarina occurred in
the N to NW sector and yet quite a number were taken in the even more

a = l2.5 FEET N

O jO o

s
Fig. 30. Schematic home ranges for Pitymys.

xerophytic E to SW sectors. The marked clumping of Pitymys in the N to
NW sector suggests that a tendency toward colony formation in this very
subordinate species may actually serve to exclude from that region the
more dominant Blarina.

However, despite such a local tendency for Pitymys to cluster, no over-
lapping of the 1-sigma radius home ranges occurred anywhere in the study
plot. In fact, the mean interval between home range centers of nearest
neighbors is of the order of 4 sigma. This must mean that the probability of
one Pitymys associating with a neighbor is very low. With their larger
home ranges, there is more opportunity for Blarina to associate with others

1. The Social Use of Space 67

of their own kind. Yet even with them their movements are sufficiently
inhibited as to reduce contacts far below that otherwise possible. How-
ever, with Peromyscus so much overlapping of home ranges existed that
most individuals must have had frequent associations with others of their
own kind. Reduction of communication with others of their own kind ac-
companies interspecific social subordination. The lower the rank of a
species, in terms of the degree other species in the community cause it to
restrict its home range, the more its intraspecific systems of communication
will be reduced. Even, as the analyses shown in Figs. 28-30 suggest, if
there develops some compensatory clumping of home ranges by members
of subordinate species, there must still be a greater degree of isolation
between such clumps than between any comparable number of groups of
the dominant species.

F. The Instability of Social Relations

Results from the extensive censusing of small mammals conducted by
Drs. Patric and Webb and their associates at the Huntington Forest form
a major key in unraveling the process involved in community structure.
Fortunately, their records (Patric, 1958) include some of the years before
1952 (see Table VI for 1952-1956) .

An effort of 9650 trap-nights during 1940 and 1941, in which traps were
set for five consecutive nights, caught 173 Clethrionomys and 1280 Peromys-
cus, or 7.4 Peromyscus for each Clethrionomys; in contrast, during the
years of 1952-1956, 4.1 Clethrionomys were captured for each Peromyscus
taken (Table VI). However, during these two years the trends of capture
over time were so nearly identical that only the greater numbers of Pero-
myscus can argue for its having been more dominant. Actually, both
showed increases in catch per day as associates were removed. The day 1
to day 5 captures were: for Peromyscus, 117, 225, 313, 317, 308; for Clethri-
onomys, 17, 19, 61, 46, 30. Apparently many members of both species were
characterized by reduced home ranges, which they expanded as associates
were removed. There being so few Clethrionomys present, they could not
represent the species producing the inhibition of home ranges. Thus at this
time we must suspect that actions by members of the genus Peromyscus
not only caused many of its own kind to contract their home ranges, but
also caused similar contraction by Clethrionomys. In the light of the ap-
parent reversibility of social roles of these two genera, previously discussed
when comparing different localities, these comparisons between different
eras within the same habitat suggest that Peromyscus and Clethrionomys
in this habitat are really codominants, but that one will nearly exclude the

68 John B. Calhoun

other. At such times whichever one became more numerous than the other
would cause members of the less numerous species to contract their home
ranges and live within the interconstellation matrix of the more abundant
species.

A small amount of trapping during 1946-1948 revealed that Peromyscus
still held a 3:1 relative abundance over Cleihrionomijs. The intensive trap-
ping was resumed during 1951, an apparently critical year in the social
balance of the population. Forty-one NACSM traplines, each run for three
consecutive days, for a total of 7380 trap-nights of effort (NACSM Re-
lease No. 5) provided total catches for the three consecutive days as
follows: Clethrionomys: 143, 114, 77; Peromyscus: 167, 132, 108; Blarina:
52, 58, 50; Sorex: 2, 0, 2. Although Peromyscus was slightly more numerous
than Clethrionomys, 9.93 versus 8.15 per trapline per three days, its slower
rate of decline in captures from days 1 through 3 suggests that it was
slightly subordinate and that some of its members were enlarging home
ranges as their associates were being trapped off. However, the nearly
constant total catch per day for Blarina clearly indicated its subordinate
status to the other two genera.

The 24-day continuous removal study of 1951 (Table V, Fig. 12) pro-
vided further insight into this realignment of social relations. Despite
nearly equal numbers taken for the three most abundant species, Blarina
is clearly subordinate to both Peromyscus and Clethrionomys. Its more
marked contraction of home ranges is revealed by the continuously in-
creasing daily catch over the first few days of trapping (Fig. 12) . Compari-
son of the proportion taken the first 5 days with that during the last 5
days provides an index of expansion of home range. The relatively more
that are taken during the first 5 days, the less has been the expansion of
home range, and thus the more dominant the species. On this basis, the
four genera are listed in order of decreasing rank in Table V.

One of the most remarkable aspects of this set of data is the nearly con-
stant daily catch for each of the three more abundant genera between the
4th and 17th day of trapping. This means that within each genus, survivors
increase their home range each day sufficiently to result in as many en-
countering traps as on the previous day despite their fewer numbers.
Furthermore, an individual member of each of these genera must be re-
ceptive to the inhibitory signals from the other two genera as well as of
others of its own kind. Otherwise, the capture curves would have resembled
that for clearly alpha species (Fig. 9).

This set of data also suggests that in the presence of several more domi-
nant species actively contending for rank status, the very subordinate
Sorex not only is markedly reduced in numbers, but also is much slower in
expanding its home ranges. Peak captures, and thus maximum expansion

1. The Social Use of Space 69

of home range, did not take place until day 21, which was just after the
final maximum expansion of home range for the other three species.

By the following year (Table lb, NACSIM Release No. 6) this uncertain
social state had completely clarified. Three-day totals for 34 NACSM
lines were Clethrionomys (753), Peromyscus (38), Sorex (17), and Blarina
(1). Other details of the resultant social organization have*already been
treated in Section VIII, A.

The studies (e.g., Figs. 10 and 13B) conducted during the past few
years in Montgomery County, Maryland, by Dr. Barbehenn and me,
further substantiate the indeterminancy of the small-mammal community
as a dynamic system. We have mostly sampled woodlands of 50 to 1000
acres. Peromyscus, Blarina, Sorex, and Pitymys are the more abundant
species. In every case Peromyscus is the dominant species. For it, the largest
catches per day occur during the first few days, and usually from day 1
there is a continuously declining catch per day with the exception of the
slight secondary increase resulting from expansion of home ranges by
gamma individuals (Fig. 10). Blarina is also nearly universally present.
The date by which 509f of the total is trapped consistently arrives several
days later for Blarina than for Peromyscus. In actuality, peak numbers
taken per day usually occur several days after initiation of trapping. Thus,
in these communities the home ranges of most Blarina are socially con-
tracted. When either Sorex or Pitymys is present, they are definitely sub-
ordinate to both Peromyscus and Blarina in terms of the degree to which
their home ranges are contracted. Their peak captures per day never occur
until after most of the two dominants have been removed. Relative num-
bers are quite another matter. It seems to be purely a matter of chance
whether either Pitymys or Sorex is present in any particular woodlot.
Either, both, or neither may be present. Their absence appears not to be
due to absence of requirements for food and shelter, but merely due to
failure to reproduce under circumstances of spatial isolation, when the
processes of social adjustment within a particular woodlot happens to
markedly reduce the numbers of some one species. These woodlots in subur-
ban to semirural Montgomery County, adjacent to the district of Columbia,
exist as ecological islands which must be characterized by a rather low
probability of receiving colonizers of these subordinate species. Further-
more, any one of the three subordinate species can become the most
abundant species in the community. In each case the most abundant species
has small home ranges whose centers lie within the interstices of the larger
home ranges of the dominant Peromyscus (as shown in Fig. IG). Blarina
was such a species in the study shown in Fig. 10. Had trapping been con-
tinued longer in the study shown in Fig. 13B, Pitymys would undoubtedly
have had a nearly 2:1 ratio of abundance to the dominant Peromyscus.

70 John B. Calhoun

In other unpublished studies by Dr. Barbehenn, Sorex catches nearly
doubled those of Peromyscus.

IX. A Theoretical Conceptualization of the Evolution of a Social Hierarchy
among Species in the Utilization of Space

The concept of social inhibition of home range represents an inference
derived from the observed differential probability of capture by snap
traps. Reduction of extent of home range must result from an increase in
the probability of terminating trips away from home. We have seen that
an increase in the structuring of the environment, with stimuli which
elicit responses, does lead to an increase in the probability of terminating
trips. This structuring of the environment represents an increase in the
amount of stimuli impinging on the organism. For this reason I believe
that one is justified in making the following tentative generalization: Any
increase in the frequency or intensity of stimuli to which the animal has
responded in the past will lead to an increase in the probability of termi-
nating trips.

It follows that genera such as Sorex or Pitymys frequently are char-
acterized by such small home ranges that we may conclude that they ex-
perience a high frequency or intensity of relevant stimuli. At the same time
their associates, Peromyscus or Clethrionomys, hsixe large home ranges.
This being so, we may conclude that the latter are not unduly exposed to
an excess of relevant stimuli. And yet we may suspect that any stimulus
which Sorex or Pitymys can detect can also be detected by Peromyscus or
Clethyrionomys. Thus, their differential response to stimuli must involve
some internal mechanism through which stimuli produce an effect indicating
relevancy or irrelevancy.

So far it has appeared that auditory stimuli of the class represented by
vocalizations of members of the small mammal community might represent
the class of stimuli producing alterations of home range. If this inference
proves correct, it means that some species respond to a broad spectrum of
different auditory stimuli whereas others "ignore" all except those emitted
by their own species.

Broadbent (1958) elaborates a theory of perception and communication
which may serve in conceptualizing how vocal communication functions in
leading to an organization of the small mammal community. Briefly this
theory is as follows:

The central nervous system may be conceptualized as a signal flow
system possessing the following connections and characteristics. Stimuli
impinging upon the sense organ generate signals which pass into a tempo-

1. The Social Use of Space 71

rary store. Storage here is in terms of seconds only. From this short-term
store, signals must pass through a hmited capacity channel before they
can serve to initiate immediate responses or reach a long-term store where
the signal can be preserved to affect later action. Signals passing through
the limited capacity channel from the temporary store may be passed
back through another circuit and reenter the temporary store. Likewise,
responses made to a stimulus or a sequence of stimuli, in turn, serve as
stimuli which generate signals entering the temporary store. Furthermore,
several signals may arrive simultaneously at the temporary store through
separate sensory channels. Only a portion of these signals in the tempo-
rary store can get through the limited capacity channel.

There has evolved a neural mechanism which Broadbent calls a "filter,"
intervening between the temporary store and the limited capacity channel.
This filter "selects" which signals may get through the limited capacity
channel and thus be available for (a) recirculation into the temporary
store, (b) inducing immediate response, or (c) entering the long-term
store.

The following conditions affect the probability of signals passing through
the filter:

1. The signal is of the same class as that of the prior signal. That is,
the related stimulus has similar characteristics in terms of frequency, in-
tensity, pattern, or location of origin. In other words, the filter tends to
pass in sequence several signals from stimuli with related characteristics.

2. However, the longer a given category of signals has been passing the
filter, the more likely the filter will switch to signals arriving from a differ-
ent sensory channel.

3. Signals generated by intense or infreciuent (novel) stimuli e.xhibit a
high probability of passing through the filter.

4. Given any three signals in the temporary store and one is passed
through the filter, the one of the remaining two most likely to follow it is
the one which followed it most frequently on prior occasions.

A special case will particularly concern us. An animal may exhibit both
bodily response and vocalization to a given external stimulus. Each of
these responses also becomes a stimulus with a high probability of associa-
tion, each with the other, and each with the external stimulus. As the ex-
ternal stimulus becomes weaker, only the bodily response is preserved.
Presumably the reason for this is that the bodily response represents
a more intense stimulus and for this reason develops a higher conditional
probability of association with the external stimulus. That is, the signal
from the bodily response stimulus is more likely to pass through the filter
immediately after the signal from the external stimulus.

72 John B. Calhoun

Although I have treated Broadbent's theory only sketchily here, its
importance in the context of the present discussion is his documentation
of the necessity for postulating the existence of some neural mechanism
having the properties he attributes to the "filter." His thesis evolved
primarily from studies with human subjects, and to a much lesser extent
from studies of rats and dogs. He accepts the existence of a filter, and even
that there are intraindividual differences in the effectiveness of its matura-
tion. His concern with the evolution of the filter extends only to his belief
that animals with a smaller cortex probably also have a less well developed
filter. He does not discuss how a less well developed filter would differ
from a more highly evolved one. We might anticipate the four attributes
of filters listed above to be less well developed. That is, signals sequentially
passing the filter represent a class of stimuli having a wider range of varia-
bility; shifts from one sensory channel to another occur with greater fre-
ciuency; a stimulus need be only slightly more intense or novel to generate
signals capable of having priority over other signals in passing through the
filter, and two stimuli must be associated in time much more frequently
for their signals reaching the temporary store to have a higher conditional
probability of passing in sequence through the filter.

Let us turn to a consideration of how the social use of space may have
encouraged the evolution of animals with more efficient neural filters. At
the dawn of mammalian evolution, we can visualize a type having close
equivalence both morphologically and physiologically to contemporary
shrews of the genus Sorex, with the exception that they had developed
essentially no neural filter mechanism for screening signals passing from
the temporary store to the limited capacity channel. All stimuli arriving
separately would get through the limited capacity channel and it was
purely a matter of chance which of two simultaneously arriving stimuli
might find passage. Under these circumstances, no discrimination may be
made between aversive stimuli emitted by neighbors at a distance. This
would lead to a uniform distribution of home range centers at approximately
2.0 sigma distance between centers.

Paleontological evidence suggests that these early diminutive mammals
preyed upon insects and other small invertebrates. Furthermore, since
among present day shrews Blarina is dominant to Sorex in the hierarchy
of use of space, it follows that evolution permitting such differential con-
trol of space must have proceeded prior to further marked alteration with
reference to feeding habits. Therefore, it appears that early in the history
of shrew-like mammals there was sufficient evolution to permit one species
to inhibit the extent of home range of another. For simplicity's sake let
us designate the earlier form as A, and the later derived one as B. Two
characteristics gradually become fixed in B. It evolved an altered vocaliza-

1. The Social Use of Space 73

tion. This ^'Ocalization preserved much of the characteristics of A, but
entailed an addition of components. Following Broadbent's analysis of
stimulus characteristics we may suspect that an attribute of some portion
of the added vocal components included an increase in intensity. Likewise,
B's filter developed alterations which enabled B to filter out selectively
those vocalizations not including the new attributes developed by B. Thus,
members of the new species B could develop conditioned associations with
the vocal signals emitted by its own kind while ignoring those emitted by
species A. At maximum stability of such a two-species community there
would exist a 2:1 ratio of A:B in a similar fashion exemplified by the
Blarina: Peromyscus community previously described for Dr. Barbehenn's
Chadwick Woods study (Section VI, B) and Dr. Webb's Rich Lake
Island study (Section VI, A) for Sorex: Clethrionomys.

The interesting aspect of this 2 : 1 ratio of species A:B is that it enabled
three animals to live where only one lived before. In other words, the e\'olu-
tion of dominant species B not only enabled as many of species B to live
in the habitat as was formerly the case with reference to the time when
species A only existed there, but it also enabled twice as many of the more
primitive species A to live in the habitat as had been the case when B was
absent. For such a pattern of evolution to have transpired, it means that
intraspecific dispersal of home range centers in a one-species community
as a consequence of the repulsive character of vocal stimuli must have been
sufficient to ensure an average utilization of resources far below maximum
carrying capacity. To clarify further what is intimated above: Members
of A, as a result of antagonistic interactions with others of its own kind,
develop conditioned avoidance responses to intraspecific vocalizations.
Furthermore, the greater the frequency of these vocalizations, as repre-
sented by increases in density of the species, the greater is the probability
of outward excursions from home being terminated, thus the smaller home
range. Lacking a sufficiently effective neural filter, .-1 responds to B's
vocalizations as if they were their own. If B emits signals with the same
frequency as does A, it follows that in a stable two-species system A will
be responding to three times the signal load as B. For this reason, ^'s
home range becomes markedly reduced in contrast to its area when B
was absent.

The next step in the evolution of the social hierarchy of space utilization,
resulting in species C, entailed similar alterations to vocalizations and to
enhancement of the filter in restricting the spectrum of stimuli which
would likely be associated with intraspecific interactions. Judging by the
fact that the omnivorous mouse Peromyscus is dominant to both Sorex
and Blarina, one may conclude that evolution of altered food preference
facilitated further evolution of interspecific social domination of space.

74 John B. Calhoun

In fact, it was this difference in food preference which led me to conchide
that the interspecific aspects of dominance in the utihzation of space most
hkely fails to involve direct physical interaction between members of
different species. Objects desired by species holding opposite ranks in the
hierarchy gradually became more and more different. At this level of evolu-
tion species A responds not only to vocalizations of its own kind but also
responds to vocalizations of both B and C as if they were by members of
species A. Species B can ignore signals from A but treats both its own
signals and those from C as B signals, while C "filters out" A'& and 5's
signals and responds only to those of its own species. In other words, C
functions as if it alone were in the environment.

The fourth step in\'olving evolution of species of type D, which is similarly
dominant to species types A, B, and (', again is accompanied by further
specialization toward a nearly total use of plant material as food. Judging
by the results of field studies presented earlier in this paper, the red-
backed mouse, Clethrionunujs, represents a species at the D level. In most
situations where both it and Peromyscus are present, Clethrionamys domi-
nates. Following the previous line of reasoning we may anticipate that
such a D-type species emits vocalizations having not only the basic char-
acteristics of species A, B, and C, but, in addition, possesses vocalization
characteristics peculiar to itself.

At each level a species responds not only to its own vocalizations, but
also to those of all species ranked above it, as if they were emitted by its
own kind. The lower the rank of a species, the greater are the number of
vocal stimuli to which it responds and thus the smaller its home range.
Reasoning back from present day simall-mammal communities to the
probable course of evolution, it appears that there is a correlation between
(a) social rank in the domination of the use of space, and (b) the shift
from carnivorous to herbivorous diet. I do not believe that an herbivorous
diet per se contributes in any way to social dominance. Rather, it has
relevance only because of the later development of flowering plants and
grasses. Evolution of more advanced types of plant permitted evolution of
small mammals specialized to utilize these new resoiu'ces. To a certain
degree such feeding specialization would enable an incipient species to
avoid direct competition with its progenitor. Once removed from overt
competition with its progenitor, psychological dominance by the incipient
species could then proceed through the process of increased complexity of
vocalization and development of a more effective filter.

A major aspect of my thesis is that psychological dominance, resulting
from a greater complexity of vocalization and an increased effectiveness
of the neural filter, far outweighs all niche specializations in determining
the relative abundance of species comprising the small mammal commimity.

1. The Social Use of Space 75

Such psychological dominance also develops among members of a single
species, at least at the higher levels of the interspecific hierarchy. We have
already presented the data and logic which gave rise to the concept of
constellation formation. On the average, each constellation has one central
alpha member with a very large home range, six beta members with home
ranges slightly smaller than for alpha individuals, and five gamma members
with markedly restricted home ranges. Such intraspecific differences in
home range size suggests that, among alpha (C or D types in above dis-
cussion) species at least, developmental alterations in vocalization arise.
Gamma members of alpha species exhibit a minimum complexity of
vocalization. Successively beta and alpha members increase the complexity
of their vocalizations. All members of the species presumably have the
same level of filter development. Even so, the differences in complexity of
vocalization should result in alpha members mainly ignoring vocalizations
of beta and gamma individuals insofar as these signals have a negative
valance. Beta members respond not only to other beta individuals but to
their alpha associates. Gamma members not only treat the vocalizations
of other gamma members as inhibitory stimuli, but are likewise similarly
influenced by those from their beta and alpha associates.

It must be borne in mind that this whole discussion of vocalizations and
filters in the context of the small mammal community is strictly theoretical.
However, it not only provides a conceptual framework offering one inter-
pretation of empirical data, but also enables formulations capable of ex-
perimental analysis.

A study of the complexity and intensity of vocalizations is suggested as
having priority in testing the theory. Sorex, Blarina, Peromijscus, and
Clethrionomys, respectively representing theoretical types A, B, C, and D
discussed above, should serve as particularly useful subjects, especially
since they all may occur in the same small-mammal community. The vocali-
zation of each higher member of the series should include the basic char-
acteristics of all lower ones and in addition possess characteristics not held
by lower members in the series. Furthermore, if each of these species is
experimentally exposed to a conditioned avoidance situation where vocali-
zations of their own species serve as a conditioned stimulus, one may
anticipate that vocalizations of the other members of the series will equally
well induce avoidance upon their replacement of the intraspecific stimulus
only if the vocalization represents a higher member of the series. Unfortu-
nately, the extreme paucity of om* knowledge of vocalizations of small
mammals necessitates these suggestions in lieu of any firm experimental
evidence.

Development of cryptic behavior forms an ancillary aspect of this con-
cept. The lower-ranked shrews typically spend much of their time in under-

76 John B. Calhoun

ground runways. Presumably such behavior enables these species to reduce
the frequency or intensity that they experience by the vocalizations of
their dominant associates. When Dr. Barbehenn first joined me in these
studies of small-mammal communities he insisted that the only effective
way of trapping Blarina was to set traps at points where excavations in
humus revealed underground runways. However, it had been my experi-
ence with continuous removal trapping, where traps were set at fixed
intervals from stations without any regard to underground runways,
that once the dominant Peromyscus has been removed Blarina were caught
on surface sets with equal ease as earlier for Peromyscus. Therefore, when
he initiated his Chadwick Woods study (Fig. 10) he set his traps on the
surface without regard to runways. During the initial days while many
Peromyscus still survived, most Blarina were taken only by traps acci-
dentally set by underground runways. As the number of Peromyscus
became fewer and fewer by the removal trapping, not only were more
Blarina caught, but also an increasing percentage of these had so entered
traps as to indicate clearly that they had been wandering about the surface
and were not emerging through the leafmold below the trap. Thus, a re-
duction in crypticism accompanied enlargement of home range. This same
change in behavior characterizes the typically subterraneous mouse Pitymys
following removal of its dominant associates. The prior discussion of the
special case of both bodily movement and vocalizations of another in-
dividual become important here. We might thus expect that vocalization
of shrews would become reduced as they become more cryptic in the
presence of dominant mice. Also as mice are removed from the habitat,
shrews should not only spend more time out on the surface, but they
should vocalize more.

This section cannot be closed without brief reference to the meadow
mouse Microtus. Data from a recent unpublished study by Dr. Barbehenn
in an abandoned orchard in Maryland suggest that this genus represents
one terminal phase in this evolution of types, which on the psychological
level enables successively evolved types to acquire a more dominant posi-
tion with reference to the use of space. At points isolated from each other
by at least 600 feet he placed covered feeding stations which contained
rolled oats mixed with dyes which stained the fur of animals eating there.
Daily removal of food increased for nearly a month, after which the daily
removal fluctuated about an asymptote for another month. At the end of
this time removal trapping was conducted along a circular trapline having
a radius of 150 feet from the feeding station. Practically all the Peromyscus
taken were marked with the red dye, but no Microtus were so marked.
This indicated that Microtus living at this distance from the feeding sta-
tions were not only unaffected by it, but their home ranges were unaffected

1. The Social Use of Space 77

by Perovnjscus, even those living much farther out in the habitat, altering
their movement and passing with increasing frequency and in an altered
pattern through the home ranges of Microtus. This trapping was followed
by additional removal trapping along several concentric circular traplines
between the original one and the bait station. Many more Microtus were
trapped, but practically all Peromyscus had been removed in the initial
trapping. Furthermore, the drop-off in catch of Microtus in this later
trapping, as the concentric lines came nearer the initial one, indicated that
only Microtus living within 37 feet of the initial circular line had been taken
by it. It furthermore indicated that Microtus were not wandering about,
as we had supposed they might be, but regardless of distance from the
feeding stations had maintained fixed home ranges despite many strange
Peromyscus focusing their movements toward this one central spot. Paren-
thetically, it might be added that this altered behavior by Peromyscus did
disrupt Sorex home ranges. They were taken in relatively large number
beginning with the very first day of trapping along the initial circular
trapline.

Now the question arises, "What do these data indicate?" Since Peromys-
cus altered its behavior by readjusting its movements toward the feeding
station as if Microtus was not there, I conclude that one of Microtus'^
adaptations has been the loss of vocalization. On the other hand, since
Microtus appears to be unaffected by the altered movements of Peromyscus
and the periodic concentration of many Peromyscus in one spot, which must
increase the vocalizations at that spot, I conclude that Microtus has
evolved a filter system so effective that they can ignore signals from other
species. This leads me to suspect that Microtus represents what might be
called a secondary herd-type species. See later discussion on the evolution
of colonialism and herding.

X. Psychological Dominance as the Primary Component of the Niche

Hutchinson (1957) includes the relationships one species has with others
as comprising aspects of its niche requirements equally to be considered
along with food, shelter, and climatic factors. Implicit in his inclusion of
interspecific social factors in niche characterization is the relative capacity
for any two given species to compete for some given environmental com-
modity or condition. The more nearly the identity of their nonsocial niche
requirements, the more important become social relationships both as a
component of the niche and in leading to the elimination of one species by
the other via the principle of "competitive exclusion" (Hardin, 1960).

In terms of this viewpoint and the marked difference in diet, morphology.

78 John B. Calhoun

and behavior of such distantly related genera as Cleihrionomys, Peromyscus,
Blarina, and Sorex, we would suspect minimum niche similarities and thus
little influence of one genus upon the other.

If their niche reciuirements are really quite dissimilar, if within any
individual's home range each requirement is represented at many points,
and if in most home range-sized plots most of the niche requirements for
each genus abundantly occur, then, depending upon the intraspecific
factors leading to fluctuations in density, we might in fact anticipate that
one or several of these genera might occur simultaneously in high densities.
In fact, observed relative densities do behave in such a fashion. Were we
content to rest our case solely on relative densities, we would remain con-
tent with the satisfactoriness of such logic in revealing true relationships.

However, the time sequence analysis of removal captures leads to a
formulation of community dynamics suggesting the nearly complete
invalidity of the concept of lack of influence of one species upon another if
their niche reciuirements markedly differ. In fact, the concept of competition
has little relevance. Instead, "home range inhibition" becomes the most
useful concept. Home range inhibition is the consequence of processes
through which the individual reacts to a signal as if it were identical to a
different signal with which it shares certain physical characteristics. In
other words, an animal may react to a signal emitted by a member of an-
other species as if it were the signal emitted by another member of its own
species with whom its interactions have led to its characterizing the signal
as noxious. Such a species becomes a subordinate member of the community.

Signals of the dominant species must contain not only the basic char-
acteristics of the subordinate but some characteristics peculiar to itself.
Thus, when both a dominant and a subordinate species occur simultane-
ously in a habitat, members of the dominant species will come to associate
no.xious qualities only to that portion of the signal which is species-specific
since it is the only portion of the signal which always accompanies a nega-
tive intraspecific interaction. If we consider a commiuiitj^ composed of
four species A, B, C, and D (following the nomenclature of Section IX)
in which D is most dominant and A most subordinate, and signal compo-
nents a, b, c, and d are observed, then Z)'s signal should include components
a, b, c, and d; C's should include a, b, and c; 5's only a, and b; and A's
only a.

Even though A learns to define signal a as noxious only by reacting with
other A associates, he will respond similarly to detection of the a compo-
nents of his B, C, and D associates. I have cited data indicating that the
greater the frec^uency of rele^'ant stimuli in the environment the greater
will be the probability of the origin of a neural signal leading to termination
of the behavior of outward movement from home. In this situation, every

1. The Social Use of Space 79

time A emerges and starts on an outward excursion ho experiences a bom-
bardment by so many a signals that his trip is shortly terminated and he
returns home. In a similar fashion, 5's home range will become contracted
but not so much as A's because A does not emit h signals, whose sum at
any one time is contributed to only by species B, C, and D. C will have
only a slightly inhibited home range since it is responsive only to signals
contributed by itself and D. At the apex of the system members of species
D are influenced only by d.

At the psychological apex of the community, further differentiation
occurs among the members of the dominant, D species. According to the
formulation developed in Section VIII, C, members of this species have the
capacity to differentiate into alpha, beta, and gamma members. Their
basic c?-type signal becomes differentiated into (/„, dis, and dy components.
Alpha individuals possess all three; beta members only d^ and dy] while
gamma members emit only dy.

This purely theoretical formulation predicts that home range size and
complexity of signals emitted are positively correlated. The observed
data on removal captures lead to inferences of home range expansion
following removal of associates. Intra- and interspecific inhibitions are so
apparent as to demand the minimum assumptions made above for main-
taining such a complex spatial organization of the small-mammal com-
munity. Compare these formulations with those of "velocity" developed
in Sections XIII, A; XIV, A; and XIV, E and F. Such comparison leads
to the conclusion that the greater an individual's velocity, the more com-
plex will be the pattern of signals he emits.

Psychological dominance is then the ability to inhibit the home range of
others resulting from the fact that the dominant shares certain signal
characteristics with the subordinate, but in addition possesses signals
which the subordinate lacks. The sharing of a signal by a species with
another species which it usually dominates may lead to mutual inhibition
of home range or actual reversal of roles.

Consider the Clethrionomijs and Peromyscus relationship. The greater
frequency with which Clethrionomijs appears to be dominant to Peromyscus
suggests that it has the c-rZ-type signal while Peromyscus has only c. We
may ignore other shared characteristics of their signals. In fact, we may
focus only on the shared c component. Chance vagaries of the system may
from time to time, after a crash in the populations of both species, result in
a marked preponderance of Peromyscus over Clethrionomys. Similarly, at
the southern periphery of its range, we can expect Peromyscus frequently
to be more dense than Clethrionomys. Clethrionomys will then meet too
infrequently to make any associations with their own species' specific
signal. However, encounters will occur more frequently between Clethri-

80 John B. Calhoun

onomys and the more numerous Peromyscus. Regardless of which species
dominates in the actual encounter, the interspecific common characteristic
of the signal may be expected to assume only a negative quality by Clethri-
onomys. Some interactions among Peromyscus must be of a positive nature.
Thus, even though both species may detect the interspecies common com-
ponent of their signals with equal frequency, Peromyscus might be expected
to exhibit less contraction of its home range because this signal is less
aversive than it is for Clethrionomys.

This line of reasoning applies also at times to relationships among three
species. In the unusual situation (Fig. 12) in which Blarina, Peromyscus,
and Clethrionomys were all c^uite abundant, the seciuential trapping data
clearly indicate that inhibition of home range nearly equally characterizes
all three species.

In such a system, the extent and center of an animal's home range de-
pends not so much on characteristics of the habitat as on the current
density and origins of signals, and the temporal and spatial history of
interaction among members of the community during the most recent
generations. The absence of an animal in a locality cannot be construed
to mean unsuitability of the habitat. It is conceivably possible to delineate
microhabitat characteristics contributing to the animal's niche even under
such circumstances. However, it requires that we know an individual's
home range center and that we mark out stations along a circumference of
a circle having a radius which will result in a high probability of the animal
crossing that circumference. A 1-sigma radius should prove effective.
Examination of stations from such a series, which have a greater than
chance frequency of visitation, might lead us to isolate those conditions
which do contribute to an animal's niche.

XL An Induced Invasion

The major portion of this section presents the author's interpretation of
a study conducted by Webb and Rosasco (1953). It describes the response
of the red-backed mice, Clethrionomys, surrounding a 30-acre tract within
which continuous removal trapping was conducted for 80 consecutive
days. In this account it will become apparent that the concepts elaborated
remain inconclusive. Even so, their implications, when taken in conjunc-
tion with the other sections of this paper, warrant their presentation.

A brief history of events leading up to Dr. Webb's study provides a
background for appreciating the objective. During the summer of 1950,
while I was in residence at the Roscoe B. Jackson Memorial Laboratory
as a National Institute of Mental Health Special Fellow, Dr. A. Dexter

1. The Social Use of Space 81

Hinckley, then a sophomore at Yale University, was assigned to me for
supervision on a research project. I assigned him the task of placing eight
NACSM traplines in an 80-acre central portion of a much more extensive
tract. He ran the 480 traps for 15 consecutive days. At that time it was
generally accepted that if one plotted a regression line through points
representing catch per day as a function of total prior catch, this line would
intersect the abscissa at a point denoting the total population (see Calhoun
and Casby, 1958, pp. 15-16 for a summary of this procedure). Previously,
Hayne (1949) had made such estimations on the basis of 3 days of consecu-
tive trapping as employed by the North American Census of Small Mam-
mals. Thus, it was my anticipation that Hinckley's longer-term trapping
would merely result in the anticipated continuous decline from day to day,
and that by the 15th day essentially no more animals would be entering
the traps. Furthermore, by the end of the loth day the total should ap-
proximate that predicted by the intersection of the abscissa by the regres-
sion line as noted above. During the first 3 days the catch per day did
decline. However, on the fourth day Hinckley reported a larger catch than
on the first day. Well, I thought this was just due to the vagaries of chance
or some unrecognized climatic factor. On every successive day his report
was the same — more animals than yesterday. On the 15th day, three times
as many animals were taken as on the first day. All of this was very
perturbing.

After several months of reflection I came up with this formulation:
During the first 3 days the number of residents in the 80-acre tract had
been drastically reduced. Animals at the periphery of this tract would then
find themselves with the normal number of neighbors centrifugal to the
trapped area but with very few remaining centripetally. Now, suppose
that in the normal state maintenance of a uniform distribution through the
environment is facilitated by vocalizations and audition. Each individual's
customary state would be that of detecting an equal intensity of signals in
all directions from the center of its home range. In these terms mice and
shrews at the periphery of the trapped area would detect few signals
toward it but comparatively many away from it. Their normal response
being to move so as to e(iualize signals coming from all directions, their
response to the neighboring depleted area would be to move in toward it.
They then became exposed to traps still set in the central area and many
of these invaders were killed. This left the next peripheral group of animals
in the same situation so that they also began moving inward toward the
trapped depleted area. By this process, a chain reaction was set in motion
in which the entire population for a great distance from the trapped area
began moving toward it. If we visualize the trapped area as a circle, rather
than its actual rectangular shape, it becomes apparent that if animals are

82 John B. Calhoun

moving in at a constant speed of tra\'el per day, those invading the trapped
area each day will represent residents from successive bands of equal
width. Each successive day the majority of the captures will be from a
more distant band than on the prior day. In a system of concentric bands
of equal width, each band more distant from the center contains a larger
area than the next innermost one. Thus, with density proportional to area
and invasion into the central area transpiring from a constant rate of move-
ment toward it, catch per day should increase with time.

At the time when this formulation was just crystallizing in the fall of
1950, Dr. Webb wrote me concerning their developing plans for long-
range studies of small-mammal populations of the Huntington Forest near
Newcomb, New York. I mentioned the interesting results obtained by
Hinckley, and the hypothesis generated by them. His response was to
replicate Hinckley's study.

Details of these two studies have been included here (Figs. 12 and 13A).
However, at the time of these two studies the failure of catch per day to
decline over time blinded us to the story which the differences of input
for different genera could tell us. That is, it was not realized at that time
that expansion of home range by subordinate members of the dominant
species and by all members of the remaining subordinate species could
lead to results by removal trapping in which catch per day did not decline
even though there were no invasions. Xot recognizing this possibility the
results of these two studies were described (Calhoun and Webb, 1953) as
supporting the hypothesis that continuous removal trapping did in fact
lead to invasion of the trapped-out area by residents from surrounding
areas.

By that time I had become associated with the Neuropsychiatry Divi-
sion of the Walter Reed Army Institute for jMedical Research. If this
hypothesis were correct it might have relevance both to lemming migration
(Elton, 1942) and to certain panic phenomena of troops (Ranson, 1949;
Caldwell, et al., 1951). In order to explore this phenomenon in more detail,
it was possible for Dr. Webb to negotiate Contract Number DA-49-007-
j\ID-325 between the ^Medical Research and Development Board (Office of
the Surgeon General, Department of the Army) and the College of Forestry-,
State University of New York. This enabled him and his associate to execute
a large series of studies on the response of small mammals to removal trap-
ping. The major details of these studies will be published elsewhere by Dr.
Webb. I merely wish here to present a brief outhne of one of these studies
because of its importance to the general theme being developed in this
paper.

In the center of an extensive forested tract they established a circular
trapline with a 562-foot radius. Along this trapline 781 snap-traps were

1. The Social Use of Space

placed, one at approximately every .5 feet. Along two diameters at right
angles to each other, 156 additional traps were placed, 3 to a station and
with a 50-foot interval between stations. This central cross of traps was
intended to facilitate removal of residents and afterward capture any in-
vaders "filtering" through the peripheral circular trapline. All 937 traps
were run for 80 consecutive days. Xo Clethrionomys entered the traps on
the 21st day. Prior to that, 101 were caught and, in addition, six died in a
preceding period of live trapping. These 107 individuals presumably
represent the majority of residents. Between the 21st and 80th day of
trapping, 501 additional redbacks entered traps. Furthermore, these 501
represented four fairly definite waves, roughly 15 days elapsing between
the beginning of each wave and the start of the next one. Each later wave
exceeded in numbers that of the preceding one (Fig. 31).

Let us now return to a consideration of the hypothesis, originally con-
ceived from reflecting upon the results of Hinckley's 1950 Maine study.
Although it is now recognized that the hypothesis does not apply to the
study from which it originated, w^e shall now ask the question: "Does it

140

O 130

21-35 36-50 51-65 66-80

PERIOD OF CAPTURE (DAYS)

Fig. 31. Invasion of redbacked mice into an area from which most of the residents
had presumably been removed.

84 John B. Calhoun

apply to the present data with regard to captures following presumed
removal of all residents during the first 20 days?" If it does, we shall also
wish to know whether the apparent waves might reflect some basic property
of intraspecific social organization.

The first problem concerns determination of the area inhabited by
those animals taken during the first 20 days. It must include all the rr^
area delimited by the trapline. Furthermore, animals from some distance
outward from the trapline must also have been caught. On first thought,
it might be logical to anticipate that all mice whose home range centers
lay within 3.0 home range sigma of the circular trapline, and away from it,
would be the only ones to exposed. However, as we shall see, the distance
outward from the trapline to which animals are affected by it more likely
approximates the maximum distance at which they can perceive signals
from other mice.

But, this is getting ahead of the analysis. For the present let us assume
that each wave of mice entering the traps represents the inhabitants of a
band of width w. Furthermore, assume that during the first 20 days mice
from a band of width, w, immediately outward from the trapline, get
caught by it in addition to those internal to the circular trapline.

The radius from the center of the area being trapped to the trapline was
562 feet. Thus, the area sampled during the first 20 days equals
7r(562 + w)^. Since each wave of invaders is presumed to represent a band
of equal width, w, then the entire area sampled during the entire 80 days
equals 7r(o62 + ow)-. One hundred and seven mice were taken from the
central area, and 608 from the total area. Thus, to the extent that number
of mice is proportional to the area they inhabit, 5.626 as many mice in-
habited the total area as the central area. It follows that:

5.626[x(562 + w)^] = 7r(562 + nwy

Thus w = 302 feet.

Radii to the limits of the central area and the four successive bands
become 864, 1166, 1468, 1770, and 2072 feet (Table IX). From these the
areas within the central area and the four bands may be calculated. These
are areas as proportions of the total within a circle having a radius of 2072
feet as given in Table IX. These proportions can then be utilized to calcu-
late the expected number of mice residing within the central area or in-
vading it during later successive periods.

The assumptions force identity between observed and expected for the
central area, but not for the four bands. However, the observed catch for
the four bands will approximate that of expected only to the extent that
the formulation is in harmony with reality. As may be seen from Table IX,
observed and expected numbers approximate each other so closely as to

1. The Social Use of Space

Table IX

Invasion of Redbacked Mice into an Area Subjected
TO Continuous Removal Trapping

Radius to Proportiou of total
Location outer edge area from which

of area trapped mice originated

Catch

Expected Observed

Central area,

r = u) + 562 ft.

864

0.174

107

Band No. 2

1166

0.145

Band No. 3

1468

0.183

111

107

Band No. 4

1770

0.227

138

135

Band No. 5

2072

0.271

165
609

163

1.000

608

make a test of significance superfluous. This, of course, only proves that
the formulation is not wrong; it does not prove its validity. In the absence
of any evidence to the contrary, the formulation will be accepted as ap-
proximating reality, and I will proceed with exploring further questions.

First, what about w = 302 feet? Webb and Rosasco (1953) and Patric
(1958) offer considerable evidence indicating that the home range sigma
for Clethrionomijs must be close to 50 feet. Thus, w = 6. Oct. According to
the independently arrived at formulation, vocal signals can be detected
up to a distance O.Oo- from the emitter (Fig. 18 and Section VII, C). Also,
the diameter of the constellation approximates G.Oo- (Fig. 27, Section
VIII, C). Recall that the constellation represents the inferred basic unit
of social organization among such animals. Thus all members of constella-
tions which overlap the circular line, or even just touch it although the
home range centers of their members all live peripheral to the trapline,
can detect a signal void centripetal to the trapline.

Before going on, I would like to emphasize that the diameter of 1124
for the circular trapline was dictated by the desire to ensure that mice
arriving at the trapline would not be able to hear others across the void on
the other side. In fact, if 302 feet does represent the threshold distance of
perception, then the angle of signal void facing a mouse arriving at the
trapline would be that angle subtended by chords of 302 feet from that
point, in this case 159 degrees.

The existence of constellations assumes prior development of bonds of
attachment among its members. Once the central cross of traps and the
circular trapline had removed all mice exposed to it, survivors of constella-

86 John B. Calhoun

tioiis ill coutjuit with the traplinc would then tend to move simultaneously
toward it. There would then arise the situation in which out to a iv distance
from the trapUne few mice remained. At this time the mice in the second
band would begin moving more or less together after a lag in time resulting
from their attachment to home and to each other. No mouse could leave
home until several of its associates were ready to move together. Similarly,
residents of bands 3, 4, and 5 would start inward as soon as most of their
neighbors of the adjoining inner band had moved w distance toward the
central trapping area. In this way a "chain-reaction" was set in motion in
which several bands simultaneously were moving toward the central area.

For the present this interpretation seems the most likely one. Whether
the phenonemon has any analogies to lemming migrations or troop panic
need not concern us here. Of importance are the indications that the in-
fluence of a signal void extends for approximately (i.O home range sigma.
The wavelike nature of invasion suggests that most of the mice from a
band of 6.0 sigma move together. That they should do so is in harmony
with the concept of constellation formation into groups whose mean size
is 12 individuals and whose spatial diameter is about 6.0 home range sigma.

In the context of the present paper this additional support of the con-
cept of a constellation as a real phenomenon represents the prime impor-
tance of this study on an induced invasion. For it is the evolution of the
constellation and its later condensation into the compact colony way of
life that I believe represents the reason why a basic group size of 12 adults
is the most important one in the mammalian series.

XII. Derivation of Compact Colonies from Constellations

Accentuation of the social bonds among members of a constellation
must have increased reproductive effectiveness and permitted survival of
individuals exhibiting a decreased antagonism toward those neighbors who
become familiar through repeated contacts. As the aggressiveness became
reduced, beta and gamma members gradually shifted their home range
center toward that of their dominant alpha associate until at last theirs
coincided with his. A compact colony will then be formed in which all in-
dividuals share the same home range. Furthermore, the aggressive actions
of the single dominant alpha individual of a compact colony serves to
protect the other members of his colony from intrusion by extracolony
members of the same species. Whereas constellation formation may be ob-
served to take place within one or a few generations from a prior uniform
distribution of home range centers, the development of compact colonies
from constellations must have entailed considerable genetic change over

1. The Social Use of Space

sufficient generations to permit the evolution of a new species or genus.
Stages intermediate between that of constellations and compact colonies
may be recognized by the mean interval between adjoining home range
centers decreasing from 2 sigma toward zero sigma.

A. Compact Colony Formation in the Norway Rat

Over a 27-month period I observed the development of social organiza-
tion in a population of Norway rats confined in a one-ciuarter acre enclosure.
Preliminary details have already been presented (Calhoun, 1949, 1952),

Table X
Characteristics of Norway Rat Colonies

Mean

Mean weight,

Number

Proportion

weight,

nonpregnant

Proportion

X young

Colony

female

males

females

of females

per

rats

(gm.)

reproducing

female

0.928

548

457

0.769

4.3

0.666

511

449

0.750

3.8

c + d

0.600

500

488

0.428

3.4

0.642

512

435

0.555

2.3

0.534

456

427

0.250

2.0

0.812

432

413

0.153

0.500

477

357

0.000

442

—

0.100

448

—

1.0

0.000

429

—

and furthei- details will shortly be published (Calhoun, 1963). By the end
of the study the population had increased essentially from a single repro-
ducing female to 120 adults. These formed 11 clear-cut local colonies
(Table X). Each colony inhabited a single burrow or group of neighboring
boxes, placed below the surface, to which the rats had access by a drain
tile from the surface. Each such burrow or cluster of inhabited boxes was
separated from adjacent ones by an average distance of about 35 feet.
These are rank-ordered from a to k in a descending order of social rank.
For the present purposes we may equate social rank with sex ratio and
reproductive success. High-ranking colonies had few males and many
females, most of whom successfully reared litters or were pregnant at the
time of terminating the study. As social rank of the colony decreased there

88 John B. Calhoun

gradually ensued a change toward more males per female, and these females
were less successful in reproduction. At the lowest level a colony consisted
of only males, or if females occurred they were essentially asexual with
regard to reproductive effectiveness. The lower the colony's rank, the lower
the mean weight.

Members of each colony represented more than one place of birth in the
pen. Members of the highest-ranked colony, a, mostly still lived at the
place of their birth or had come there from adjacent colonies. As social
rank of the colony decreased, its members represented ever more different
places of birth. Despite this disparity in places of birth characterizing each
colony, the total number of adults forming each colony varied little from
the mean number of 12 characterizing the loosely knit constellation of such
simple social types as Peromyscus and Clethrionomys.

At this point we may consider possible events which fostered the evolu-
tion of the compact colony from the loosely knit constellation. The major
impact of the constellation way of living is that the most frequent group
size would be 12 individuals. It might, therefore, be anticipated that
evolutionary processes would adjust the physiology and behavior of such
species to be most effective and appropriate to interactions transpiring in
such a sized group. In accordance with the conservatism of evolution we
might anticipate restriction of group size about this optimum of 12 as
other factors caused the constellation to contract into a compact colony.
Restriction of location of food stands is the most likely candidate for an
appropriate environmental change. Provided the location of food became
restricted but abundant and relatively permanent at these locations, we
could expect types like the Norway rat to develop. They build burrows
not too far away from such spatially restricted sources of food and all
members of the colony participate in transporting this food into the burrow
where large caches are formed.

Compact colony evolution produced a situation which necessitated
further evolution of the nervous system. If groups of 12 adults assembled,
either as a consequence of being born at the same place or from random
mixing, a more nearly equal sex ratio would characterize most colonies
than was the case in my study within the experimental enclosure. This
presents the opportunity for development of marked aggression among
males and consequent stress experienced by associated females. Actually
this was the initial situation always characterizing an incipient colony of
Norway rats. In the process one or more males were driven off and suffi-
cient females remained, although some always left to keep the colony at
near 12 individuals. The rejected males either joined another developing
colony lacking a male as dominant as the one at the colony they left or
they joined to form an all-male colony. Where there was a single very

1. The Social Use of Space 89

dominant male (as in colony a, Table X) his actions kept away most other
males at times when his harem females were in estrus. This reduced the
stress experienced by females in estrus, such as followed the thousands of
mountings or attempted mountings experienced by females in colonies
lacking such dominant males. In this sex-ratio restructuring of the popula-
tion, a few colonies contribute most of the young.

Although this readjustment ensures the survival of the species, it is not
the conseciuence most important with regard to further social evolution.
Each time a rat is excluded from one colony it attempts to join another
aggregate in order that it will again find itself in a group of a size most
compatible with its physiology. As soon as a group exceeds the optimum
size, some of its members are excluded from it. This results in a marked
reshuffling of the population from the time of puberty of young born in
one season up to the beginning of the next breeding season. At this time
the population is relatively stable with regard to membership of each
colony.

In the process of attaining colony stability, the social environment is
in a constant state of turmoil. Each individual is forced to make many
adjustments to such changes. It continually has to learn new social rela-
tionships. These are learned so well that a group can shift its place of resi-
dence over a sufficient distance that passage by or through other colonies
is necessitated. Even so, they can maintain their group integrity. I am
convinced that the necessity of making such changes of membership from
one colony to another required for reproductive survival of the species
has resulted in the evolution of the Norway rat into a species which is not
only highly perceptive of changes within the environment but has the
capacity to learn required adjustments of behavior.

Although such capacities for perception and learned adjustment must
have arisen in the context of a changing social environment, these same
capacities then become available for perceiving and adjusting to nonsocial
changes in the environment. Among ecologists this extreme awareness to
changes in their environment by Norway rats has been termed the "strange
object reaction" (e.g., Chitty and Southern, 1954). Among psychologists
it is reflected in the studies falling under the broad rubric of ''open-field
emotional behavior" (e.g.. Hall, 1934, Schneirla and Tobach, 19G2).

If I were to make my evolving thesis concerning the social use of space
complete, I should substantiate the role of vocal communication at the
compact colony level of social evolution. Unfortunately, I cannot cite any
adequate proof regarding the nature of its function. Norway rats do have
a wide scope of vocalization ranging from the loud signal accompanying
the termination of a fight to the low chirping and whining one can detect
if one lies on the surface of a burrow with his ear at an entrance hole. Many

90 John B. Calhoun

other vocalizations also occiii- when rats are on the surface near their
burrows and also while at the source of food. I can only suspect that among
these there is specific communication among alpha members, the dominant
males, neighboring colonies, and that rats at the food source emit a signal
which might be termed a "here is food" signal. Unless there are such signals,
it is difficult for me to understand many of the observed behaviors of rats
which clearly indicated that they were aware when other rats were or were
not at the food source, even though they could not see it.

B. Howler Monkeys, a Compact Colony Living Species

Carpenter (1962) summarized the results of field research on this species
{Alouatta palliata) during the past thirty years. Tabular data on 136 dis-
tinct groups show the number of adult males and females and the number
of immature individuals in each group. I derived Fig. 32 from these data.
Two large groups containing 27 and 31 adults, respectively, were omitted
from the analysis.

Regardless of group size all groups contained more females than males.
Excluded males live in a state of near isolation and have very little associa-
tion either with each other or the organized groups. As with the Norway
rat, reproduction within compact colonies apparently requires a reduction
in the number of males in groups for effective reproduction to take place.

One of the central hypotheses in my formulation of the social use of
space is that constellation formation must have served as a mold which
so guided evolution that beha\'ior and physiology would become fixed so
that they would have optimum expression in a group of 12 adults. For
howler monkeys, groups of 9 to 11 are encountered more frequently than
smaller and larger ones (Curve A, Fig. 32). And although the decline in
frequency of groups containing more than 1 1 adults is not so rapid as ex-
pected (Table \'III), the observed data do reveal a marked decrease in
frequency of larger groups. Furthermore, only 3% of the 136 groups ex-
ceeded the maximum of 19 anticipated by the theory. The theory indicates
no expectation of groups containing less than 7 adults, yet 26.5% of the
136 groups of howler monkeys did contain less than 7 adults. However, it
must be pointed out that such theory presupposes completion of all social
processes culminating in a group having considerable stability of member-
ship. Incipient groups formed from the fragmentation of larger ones and
larger ones approaching the point where fragmentation is imminent should
logically both be excluded in comparing observed and theoretical fre-
quencies of group sizes. However, present knowledge prevents such a
comparison.

1. The Social Use of Space

Production of young is fairly ineffective in small groups in comparison
with those containing 10 or 11 adults (Curve C, Fig. 32). In general, the
trend of young per female also decreases as groups get larger than 12 in-
dividuals. The two very large groups not included in this figure and which
had 27 and 31 adults were characterized by only 0.18 and 0.29 young per
female, respectively. We must conclude that group size does affect those
behaviors and physiology culminating in the production and survival of
young.

UJ ■—

2 Q. 0.10

P < 0.05
o >.

Vng/j

'''"' o° ^ ° ° ° ° °o , o „ - CO

I 1 I I I I I I I 1 1 I I I I I 1 1 1

5 10 15 20

NUMBER OF ADULTS IN GROUP (HOWLER MONKEYS)

Fig. 32. Group dynamics of howler monkeys based upon Carpenter (1962) .

Another way of looking at this problem is to ask: "What proportion of
the total young are contributed by each group size?" For any group size
the proportion will depend upon (a) the proportion of females in the
group (Curve D, Fig. 32), (b) the number of young per female (Curve C,
Fig. 32), and (c) the number of groups of each size (Curve A, Fig. 32);
where such data for each group are weighed against similar data for all
other group sizes. Resultant data are given in Curve B, Fig. 32. This
curve clearly shows that more individuals gain their initial social experience

92 John B. Calhoun

in groups of about the size anticipated by theory as most nearly optimum
than they do in any smaller or larger sized groups. In fact, the mean number
of adults with which the 838 young in these 136 groups had their early
social experience was 12.22!, even though Curve C, Fig. 32, is flatter than
predicted by the theory of Section VIII, D and Table VIII.

My selection of the Norway rat and the howler monkey as examples of
compact colony types is open to the criticism that I selected those species
which would support my theory that group size in higher evolved types is
dependent upon their evolution from species which had been characterized
by the loose constellation form of social use of space. At present, this
criticism cannot be avoided. I wish merely to say that my intensive study
of a few groups of Norway rats, and the extensive study by Carpenter and
his colleagues of a large number of howler monkey groups, are the only
ones known to me that appear adeciuate for the present purpose.

All we can really say at present is that available data show that optimum
group size in some compact colony living species appears to approximate
12 adults and that this number is in harmony with that number antici-
pated by the physics of communication characterizing more primitive
and more dispersed types.

C. Behavioral Sink Development by the Norway Rat

IVIuch of the prior sections have been devoted to documenting theory
and e\ddence supporting the hypothesis that groups of 12 individuals
represent a major category of optimum density. Yet many species cus-
tomarily live as aggregates much larger than this. Why such large groupings
should have evolved remained a puzzle until insight developed from an
unexpected phenomenon arising in the situational content of some experi-
mental populations of rats I was studying (Calhoun, 1962a).

The upper portion of Fig. 33 presents in diagrammatic fashion the salient
aspects of the environment affecting the rats. Four 35 square-foot pens,
separated by 2-foot high partitions, formed a linear communication net-
work through the opportunity of access between adjoining pens via the
V-shaped ramps, R, surmounting barriers between pens. F and W represent
a superabundant supply of food and water in each pen. "Apartment"
houses, H, connected to the floor by ramps provided ample place of retreat
and rearing young by most residents. The //'s of the left-hand pens I and
11 were 3 feet from the floor, while in pens III and IV a 6-foot distance
separated the i/'s from the floor. Height formed an intentional environ-
mental factor designed to produce a 2:2:1:1 ratio of density across pens
1:II:III:IV as a consequence of the inverse-proportionality-to-effort
usage principle.

1. The Social Use of Space

In addition the endedness of the environment biased movement. After
some period of time every rat tended to leave the pen it was then in and
go into an adjoining pen. Rats in an end pen could only go to the adjoining
center pen, while a rat in one of the two center pens could go to the other
center pen or to the adjoining end pen. In other words, when a rat is in an
end pen it has a probability of 1.0 of going into the adjacent center pen,
but if it is in a center pen it has a probability of 0.5 of going into the adja-
cent end pen or 0.5 of going into the other center pen. Repetition of shifting
by all members by the operation of this principle alone soon leads to a

PEN I

PENH

PEN m

PEN IS

F W
H 3 Ft.

==3 RC^

F W
H 3 Ft.

H 6 Ft

z> 0.5

PEN
lor HI

I *********** ******* *'

ASONDJFMAMJJ
1958 MONTH 1959

Fig. 33. The upper portion of the figure represents in schematic fashion the environ-
ment in which large social groups of albino rats were maintained. See text for details.
The lower figure contrasts the amount of food consumed in the most used and the least
used pen from the inception of the behavioral sink during the sixth month (August,
1958) of the study.

steady state in which a 1:2:2:1 ratio will characterize the density of rats
across pens I: II: III: IV. See pages 298-299 of Calhoun (1962a) for details
of the mathematics involved in the origin of this ratio.

Populations in four such 4-pen environments were studied from Febru-
ary 1958 to July 1959. By the eighth month, September 1958, each popu-
lation consisted of three generations, each artificially fixed at near 30 in-
dividuals. The first and second generations were sexually adult, and the
thu-d was recently weaned. Distribution of adults at this time proved to
be of considerable importance in producing a phenomenon I have termed
a "behavioral sink." Five surveys of place of residence gave a total count
by pen of 343 for pen I, 467 for pen II, 331 for pen III, and 245 for pen IV.

94 John B. Calhoun

If the two movement biasing principles described above operated inde-
pendently and equally, the expected ratio of density would be 3:4:3:2
across pens I: II: III: IV, thus giving an expected distribution for the
September 1958 observation of 347:462:347:232. The observed and ex-
pected values are so nearly the same as to support strongly the belief that
these two movement biasing principles were in fact the only effective ones
operating at this time. These same factors also affected the third generation
then maturing.

Taking mortality into consideration, each population consisted of about
80 rats with on the average 20 living in pen I, 27 in pen II, 20 in pen III,
and 13 in pen IV. Only in pen IV did the density approximate the ideal of
12. Elsewhere, particularly in pen II, density far exceeded this. From this
time on a remarkable change in the differential use of space, particularly
as reflected by food consumption, set in. In one pen (in three instances it
was in pen II and in the fourth in pen III) food consumption increased at
the expense of that in the other three pens. See the lower half of Fig. 33.
By the time seven months had elapsed, most rats were eating all their
food in this "favored" location and all rats were doing most of their eating
there.

The explanation for this change appears rather simple. Gnawing food
through the wire mesh of the food hoppers required considerable time. In
the one pen w^here more rats fed than in the other pens, the probability of
two rats eating side by side increased. Gradually rats redefined the eating
situation as requiring presence of other rats. Thus, all rats shifted most of
their eating to that pen where this condition was most likely to be met. It
must be kept in mind that such a system is stochastic and not deterministic,
so it was not unexpected that pen III became the favored place of eating
by one of the four groups. However, the likelihood of pen IV ever becoming
the favored pen is extremely remote.

The learned need for social proximity while engaging in an act which
might have been expressed alone assumed priority over the simple original
hunger drive. Food was not food without the presence of a comrade. This
whole process of developing excessive aggregations in order to satisfy a
secondarily acquired social drive is what I mean by a "behavioral sink."
Gradually more rats also shifted residence to this favored place of eating.
Such behavioral sinks result in every member encountering more associates
than the ideal, and even more than necessitated by the operation of those
principles of spatially structuring the environment which biases movement.

In this situation marked alterations in mortality and behavior resulted.
Males became pansexual in the sense that they mounted other rats ir-
respective of sex or age. Nest building and maternal behavior became so
disrupted in most females as to preclude the possibility of most young

1. The Social Use of Space 95

surviving. In each experimental setting the rats experienced less disturb-
ance in one pen, usually pen IV, than elsewhere, since they were somewhat
less trapped in the behavioral sink. And yet even here only half the young
born survived to weaning and their growth was markedly retarded. In
contrast, in the pen where most rats assembled only half as many young
were born and only 1% of these survived to weaning.

In addition, the abnormal frequency of social interaction resulted in
marked disturbance to female reproductive physiology. Near-term fetuses
died. Some females with such dead fetuses shortly succumbed from ap-
parent toxemia. Others died from massive hemorrhaging in many organs,
an accentuation of the event likely to have been associated with fetal
death. Many females who survived such events later died as a conseciuence
of a site of resorption of a near-term fetus becoming the focal point for the
development of a large abscess. Normally death occurred by the time the
abscess reached a diameter of 50 mm. As an example of this scourge, 56%
of second-generation females died by a year of age, by which time only
10% of males had died.

Were a species to survive for many generations in an environment
fostering development of a behavioral sink, it is obvious that selection
must proceed to produce individuals whose behavior and physiology were
in harmony with such a heightened frequency of social interaction. Wher-
ever an environmental resource which was formerly so widely distributed
as to be readily available within each individual's or group's home range
becomes restricted, then conditions are ripe for production of a behavioral
sink. It is my belief that just such happenings have been the usual altera-
tions which have forced the evolution of horde or herd type species from
one previously characterized by an optimum group size of 12 adults.

D. Yarding by Deer in Northern Wisconsin

Characteristically since 1935 white-tailed deer (Dahlberg and Guettinger,
1956) in Northern Wisconsin assemble during the winter in a restricted
area known as yards. These cover only 5-10% of their range. Conifers,
which comprise the major cover in the yard, provide protection from deep
snows. However, such cover provides only a secondary quality food. One
or more feeding stations were established in most yards. During the 1930's
food supplements represented a small amount of total food requirements.
By 1953, when artificial feeding was largely terminated, most of the food
reciuirements were supplied at these feeding stations in many yards.
Nevertheless, many deer died of "starvation" even in yards where the most
food was provided.

96 John B. Calhoun

The restricted locations where food was provided and the striking ag-
gregations of deer in their vicinity reaching 350 per square mile strikingly
resemble my experimental populations of rats from which developed the
concept of the behavioral sink. Even though many deer died in the yards,
the question stands: "Is this really an instance of a behavioral sink?"
Gaining insight into this question has proved to be a difficult detective
job. Shiras (1921), Sanders (1939), Swift (1948), Rabat et al. (1953),
Schorger (1953), and Dahlberg and Guettinger (1956) proved to be
particularly helpful.

Before the days of lumbering, deer were so scarce in the primeval forests
of Northern Wisconsin as to contribute very little to the diet of Indians.
Between 1860 and 1880 a marked increase in deer followed lumbering
operations with the consequent development of openings and second growth
which provided abundant food. Although the deer did not reach the den-
sities of 1935 to 1953, they supported a major industry as a commercial
source of meat. There is some mention during this era of herds up to 200
being seen, of aggregations about salt licks, and about concentrations in
white cedar swamps during heavy snows. However, one gets the impression
from Schorger's citations that it was more customary for deer to be scat-
tered. After heavy snows the commercial hunters trailed the deer until
they found them exhausted and trapped by the deep snow. Schorger
(1953, p. 210) writes, "It is stated by Harvey Braein that about Christmas,
1857, a crust about one-half inch in thickness formed on the deep snow in
Buffalo County, and that nearly every deer perished. Following the spring,
their bodies were found in nearly every coulee." Unfortunately, this is the
sort of data one has to rely on. Even so, it suggests a typical pattern of
scattering.

As early as 1920 when the Northern Wisconsin deer herd was well on
its way to recovery after its prior decimation by forest fire and unrestricted
hunting, private hunting clubs and the operators of tourist camps had be-
gun the practice of feeding deer during the winter months. After 1935
Civilian Conservation Corps camps and the Wisconsin Conservation De-
partment greatly increased this artificial feeding. However, the intensive
artificial feeding characterizes only the 1943-1953 period. Swift (1948)
states that yarding had not conmienced very extensively until after 1941
even though astonishingly high populations existed in many locations. It is
difficult to escape the conclusion that the accentuation of yarding was a
direct outgrowth of the artificial feeding.

Daily movements rarely exceeded one-quarter mile from the feeding
stations. Thus, available food outside the yards remained unutilized. Even
with the advent of warm weather deer exhibited considerable reluctance
in leaving the yard despite increasing new growth outside it. Even cessa-

1 . The Social Use of Space 97

tion of artificial feeding has not disrupted the marked yarding tendencies
of Wisconsin deer. Through many generations they have developed a
culture demanding an excessive frequency of contact with others during
the winter season when they would otherwise be more scattered.

I will freely grant that this interpretation of the origin of yarding may
be oversimplified. Nevertheless, available observations warrant considera-
tion of the concept of the behavioral sink as helping us understand the
historical development of yarding to an excessive degree.

E. Concerning Basic Numbers, Nb, for Man

During the past half million years, density of Homo sapiens has exhibited
a continued increase over the inhabited portions of the earth (Deevy,
1960; von Foerster et al, 1960) . Most who have concerned themselves with
studying such change restrict their emphasis to changes in density accom-
panying advance in extractive efficiency of natural resources. Less atten-
tion, even by anthropologists, has been devoted to determining sizes of
social groupings forming partially closed systems. Such partially closed
systems range from a male-female pair to that of a nation such as the
United States among which interactions among all members may be con-
ceived of in terms of population potential (Stewart, 1948; Calhoun, 1957).
It will not be my purpose here to treat the entire range. Rather, I shall
merely present a few highly selected examples of some of the smaller
groupings which presumably reflect evolutionary limitations to group
structure. On the assumption that these group sizes represent the conse-
quences of underlying basic forces, they will be used in later sections
(XIII, A and B) as data for developing a general formulation of group
size and social interaction.

For about 98% of his history during the past half-million years, simple
food gathering limited man's economy and social life. One of the earliest
known settlements at the Star Carr site in east-central England of nearly
10,000 years ago consisted of five families, ten adults (Braid wood and
Reed, 1957) . Thus, it appears that an adult group size not diverging greatly
from my hypothesized ideal of 12 may have characterized the human
species up until at least 10,000 years ago.

The Australian aborigines (Birdsell, 1953, 1957) provide further insight
into the basic numbers of human groups. In terms of the typical number of
adults, five levels are recognizable: (a) the family with 2 adults; (b) the
"horde" or extended family with 16 adults; (c) the supra-horde of 50
adults; (d) the tribe with 200 adults; and (e) the supra-tribe with 2200
adults. The horde forms the most basic social group, ranging generally

98 John B. Calhoun

between 10 and 20 adults. When the group exceeds 20 adults a budding
process occurs, 10 adults forming a minimal-sized horde. These approxima-
tions of BirdseU's again suggest a basic group size not diverging greatly
from the 12 presumably fixed by much earlier evolution. Occasionally,
the horde may fragment temporarily into single family groups when
scarcity of food demands such dispersal.

Also, occasionally an average of five hordes, 50 adults, may temporarily
assemble into a supra-horde. However, this grouping appears to be a less
basic one than the other four. The third grouping, the tribe, lacks any
form of authority, and only on rare occasions do the 12 or so hordes forming
the typical tribe assemble. However, common bonds of culture, their cus-
toms and value systems, clearly delineate the tribe as a social entity.
Furthermore, marriages are primarily restricted to those between in-
dividuals of different hordes within the same tribe. Only in one local region
of Australia has a higher-order grouping evolved. Three supra-tribes,
averaging 2200 adults, represent an assembly of tribes bound together by
a more advanced type of political organization characterized by matrilineal
descent.

These data suggest that cultural evolution has proceeded by saltatorial
steps, each characterized by some accretion to the culture. It wnll be my
hypothesis, to be developed in more detail in Sections XHI, B, 2 and B, 3,
that the sole function of culture is to provide a mold which enables inter-
actions to transpire in a larger group such that their physiological conse-
quences to the average individual closely approximate those that would
result were the individuals still living in a closed social group of about 12
individuals.

Hallowell (1960, pp. 345-846) states that "... a normative orientation
becomes an inherent aspect of the functioning of all socio-cultural systems,
since traditionally recognized standards and values are characteristic of
them. Techniques are appraised as good or bad; .... Knowledge and
beliefs are judged true or false. Art forms and linguistic expression are
evaluated in relation to ethical values. All cultures are infused with ap-
praisals that involve cognitive, appreciative, and moral \'alues," and "if
the total ramifications of the normative orientation of human societies
are taken into account, we have a major clue to the kind of psychological
transformation that must have occurred in hominoid evolution w^hich
made this level of adaptation possible and some measure of its depth and
significance for an understanding of the dynamics of human social systems
of social action." Culture so conceived as normative orientation in which
individuals play sanctioned roles provides the structure which allows in-
dividuals to reap the maximum rewards (the theta, 9, of Section XHI, A)
from participation in the social system.

1. The Social Use of Space 99

If culture really does permit individuals to function in the context of a
larger social group as if they were still only in the basic A^'t = 12 group com-
patible with their physiology, then any disruption in the culture should
reduce the group size since its unstable state would then no longer buffer
the individuals from the excessive contacts with their associates. I am as-
suming that, depending upon the extent of the cultural disturbance,
physiological disturbances comparable to those of my rats caught in the
behavioral sink (Section XII, C) would arise.

In fact, Birdsell (1953) demonstrates that such a phenomenon has
characterized Australian aborigines in recent times. Tribes which have
recently adopted the rites of circumcision or subincision generally have a
size less than one-third that of tribes which have either not been exposed
to these practices or adopted them long ago. Furthermore, the historical
records indicate that tribes once reduced in numbers after they first adopted
these rites now after several generations have recovered their typical
numbers.

These data on the Australian aborigines further suggest that an in-
dividual can shift his participation from one level of social organization to
another, provided there are cultural means for channeling such participa-
tion. Duff and Kew (1957) provide an account of the recently extinct
Kunghit Haida Indians of British Columbia, which enables similar in-
sights into basic group sizes in a food-gathering people.

Their winter village consisted of 16 to 20 large houses (1600 sq. ft. of
floor space each). From various of the accounts it appears that the tribe
totaled about 500 individuals, of which slightly over 200 were adults. This
means about 10-12 adults on the average per house. Each house was in-
habited by a kinship group or lineage. During the warmer months of the
year each lineage group left the winter village for its own hunting territory.
Like the Australian aborigines, these British Columbia Indians also appear
to have a basic group size not diverging far from 12 and an assembly of
these into a tribe of around 200 adults.

Incipient agriculture, in which plow and draft animals are absent, repre-
sents an even more ad\'anced efficiency of food extraction, characterized
by a permanent village. The Jarmo site in Iraq, inhabited some 6700 years
ago, presumably represents a typical village at this level (Braid wood and
Reed, 1957). Braidwood and Reed estimate that 150 persons (50 adults)
inhabited the 25 houses located there. This type of village structure ex-
tends into the present. The mean size of 185 villages in this part of Iracj
is 140, which presumably represents 46-56 adults.

From the scanty examination of lower-order basic group sizes in man
we shall skip to the urban society of a modern nation, the United States.
The social organization represented by Australian aborigines and the

100 John B. Calhoun

Kunghit Haida suggests that each larger semiclosed social system includes
within it all the culturally limited basic group numbers. But even if some
are skipped or unrecognizable, the one group structure which must be
preserved is that of 12 adults. Recent studies by Zimmerman and Broderick
(1954) and Zimmerman and Cervantes (1960) confirm this suspicion.
Their approach has been to focus on any given family, designated the ego
family, and then to determine with how many other families its members
have frequent and close associations. These latter are designated as friend
families. Absence of divorce or desertion, juvenile arrest, or children not
completing high school comprised criteria for judging a family as "good"
or "successful." Presence of these traits were used to delimit the "bad" or
"unsuccessful" families. Values held by a family were judged on the basis
of their religion, region of origin, income level, and kinship bonds. The
good ego families typically have five friend families with whom they have
a high coincidence of values, and furthermore, if the ego family is char-
acterized as good most of the friend families are likely also to be so char-
acterized. On the other hand, bad ego families generally have fewer friend
families and they are likely to differ from them with respect to the value
traits. The fewer the values shared by the several families forming such a
cluster, the smaller the cluster will be and the greater the probability that
each family will be characterized by one or more of the traits denoting it
as an unsuccessful family.

The ideal state then appears to be six families, 12 adults, composed of
an ego family and five friend families. Shared values bind such a cluster
despite the dispersal of the member families through the local community.
Furthermore, each friend family in a particular cluster is, as an ego family,
the center of another cluster. In this way an extension of the cluster de-
velops to include 26 total families. Although similar bonds between families
may include a larger network, insofar as any particular family is concerned
the 25 friend families and extended friend families form the limit of de-
pendence and social support relationships. This approximation of 50 adults
of the family-friend cluster further argues for the reality of Nb = 50 as a
basic grouping revealed also in Birdsell's supra-horde of Australian abo-
rigines and of the incipient agricultural village of the Jarmo type. Reduction
of the size of the family cluster below the optimum of six when values held
by member families diverge from each other represents another example of
the principle of group fragmentation, enunciated by Birdsell, which follows
a clash in values. Zimmerman and Cervantes refer to this conflict as a
"confusion of values."

All the information in this section, when viewed as a whole and in the
context of the earlier sections concerning the evolution of a basic group
size, suggests the following tentative generalization: Modern man derived

1. The Social Use of Space 101

from his primate and preprimate ancestors a physiology transpiring in
groups within the range of 10-20 adults. This physiology was fixed some
half a million years ago and has not significantly diverged from it since.
Development of a larger social group is made possible by a culture in which
a normative orientation prescribes values, and sanctions roles of behavior
such that the total effect of participation in a larger group so buffers the
individual that at any particular time the individual functions socially as if
he were a member of a group of 12 individuals. Furthermore, genetic
changes of the central nervous system making learned value systems of
cultures possible must have arisen under circumstances which prevented
division of the basic group size when it reached twice this level. Either an
ecological-psychological trap like the behavioral sink (Section XII, C) or
any isolated but very abundant source of a needed resource would be ade-
quate to demand either a genetic change of physiology making life in large
groups tolerable, or a genetic change endowing the central nervous system
with the capacity to learn and culturally transmit values. Each increase in
group size is associated with a reorientation of the value system. Such in-
creases in group size are saltatory. The theoretical basis of why such
changes must be saltatory and not transitional is discussed in Section
XIII, B, 3.

XIII. A Formulation of Group Dynamics

Twelve individuals represent the approximate optimum group size for
certain species (Sections VIII, C and XII, E) . Furthermore, such a sized
group might be expected to have evolved from home range dynamics.
Evidence could readily be assembled that other basic A^'s, NbS, characterize
other mammahan species. Some typically live as pairs while others assemble
in herds exceeding 100 or 1000 individuals. However, circumstances may
force N to diverge markedly from A^6. Elaboration here of the model of
social interaction presented in pages 349-354 of Calhoun (1957) provides
insight into the consequences of such divergence of A^ from Nb.

A. The Model of Social Interaction

On a presumptive basis there are three variables which should determine
the mechanics of contact and interaction. These are (a) the number of
animals moving about and having opportunity of contacting each other,
(b) the length of the refractory period following the response of one animal
upon contacting another until it is again capable of exhibiting a similar

102 John B. Calhoun

response, and (c) the amount of space in which the movement of A'' in-
dividuals takes place. Random distribution of positions of individuals at
any moment in time is assumed. We choose to ignore a small correction
factor arising from the fact that all individuals move. Velocities of all in-
dividuals are initially assumed to be a constant. Furthermore, we assume
that all individuals are identical. Thus our concern is not which individuals
meet, but rather which state, responsive or refractory, the contacting in-
dividuals happen to be in.

N = Number of animals forming the group.

d = The diameter of interaction for each animal, that is, that distance
between the centers of two individuals at which a physical or
psychological collision or contact occurs. In the simplest case
animals may be considered equivalent to billiard balls. Then d
represents the diameter of the ball, the individual. See Section
XIII A, 1 for further elaborations.

Assume an animal moving in some direction on the plane in a population,
A^ — 1, of other animals.

Each of these other individuals presents a target of dimension d, normal
to the X direction. The expectation that the incoming animal will make a
collision while moving a distance A.r (in time t) is the ratio,

d(N - 1)A.T

of surface covered by the targets to the total surface, where A is the area
available to the animals.

It should be emphasized that the unit of time must be sufficiently large
so that the number of collisions in that time interval is large enough to
justify using the statistical law of large numbers in the derivation. For
similar reasons, it must be assumed that the mean free path of the in-
dividuals must be large in comparison with the target diameter.

Since the velocity v may be considered ec^ual to Ax/t, the average num-
ber of contacts w,-, per individual in time /. is

djN - l)vt

Tie = (o2j

For present purposes we are concerned only with the average ric in t and
not in the variability in contacts in t. The frequency of contacts by a given
individual will be :

^. . - . rf(iV-i)_. (33)

{ A

1. The Social Use of Space 103

Since d, v, and A will be considered constants for this presentation, we set

(dv/A) = n so that /. = m(A^ - 1) ('^^i

fjL reflects the ease of communication in the sense of contacts per unit ot
time. Basically, d, v and A may be specified in terms of linear unit, L.
Therefore,

dv L • Lt~^ 1 ..,_,

" = 1^-6^ = 7 ''•^'

The symbol ^ is here used in the sense of ''dimensionally equivalent to."
So by selecting appropriate units of time, n can be made equal to 1.0. In
following discussions n will be considered equal to 1.0 in this sense whenever
the basic A^, Nb, of a species is in an evolutionarily steady state.

We will assume that the population of individuals can be divided into
two classes: [Na], those individuals who are in a responsive state, and
[Np] those individuals who are in a refractory state. We further assume
that the individuals in [Na] will be rendered refractory either after a
contact with a member of the same class or with a member of [Np]. After
any such contact, an individual will remain in the refractory state for a
length of time, a, the refractory period, and after this time has elapsed
return to membership in [N^]. It is also assumed that any contact that
an individual undergoes while it is in the refractory state has no influence
on the duration of its refractory period.

The duration of such refractory periods must be a function of the be-
havior of each member of the contacting pair toward the other. It is as-
sumed that the critical aspect of this behavior is its intensity. At the steady
state of an Nb we shall first consider every individual to be identical with
reference to the intensity of its behavior toward others.

Let:

la be the intensity of action of any member of [A^„] toward every associ-
ate it encounters.

ip be the intensity of action of any member of [Np] toward every associ-
ate it encounters.

It is further assumed that the most likely way that the duration of the
refractory period, a, becomes a function of the behavior of two individuals
toward each other is that it results from the product of the intensities of
their behaviors. Furthermore, there must be some factor, which will be
called B, which governs whether a refractory period will result from the
interaction.

104 John B. Calhoun

Let:

Ba = 1.0 be the value of this factor in all members of [Na].
Bp = 0 he the value of this factor in all members of l^Np].
aaa represent the refractory period resulting in each of two members of

[_Na] who meet.
a,rp represent the refractory period resulting in each member of [^Na']

which encounters a member of [A^p].
aj,a represent the refractory period resulting in each member of [iVp]

which encounters a member of [_Na].
app represent the refractory period resulting in each of two members of

[_Np^ which meet.

It follows that:

iaia = oiaaBa = idp = dapBa, aud all are real values (36)

and that:

^p^a ^pa-*^p ^p^p ^pp^ p V7 \*J i )

In all following discussion Ba and Bp will be omitted in discussing a,
but every mention of aaa and Uap will assume the action of Ba, and likewise
any mention of apa and app will assume the action of Bp.

It may be objected that no distinction is being made between the two
kinds of contacts, responsive-responsive, and responsive-refractory. It is
perfectly feasible to introduce two refractory periods, aaa and aap, of differ-
ent duration to answer this objection. At the present juncture the experi-
mental data are so scanty that it does not appear to be fruitful to introduce
additional complexity, and we have chosen to consider ana = cinp for the
workmg model insofar as duration is concerned.

An alternative model would have been to choose aap = 0, that is, the
only contact inducing a refractory period being a contact between two
individuals both of whom are responsive. Under such an assumption the
number of contacts between responsive individuals in a unit interval of
time w^ould increase asymptotically to the value 1/a, whereas, as we shall
show, the model adopted provides that the number of contacts between
responsive individuals passes through a maximum as A^ increases.

Given sufficient proximity of an individual in the responsive state to
some other indi\'iduals requisite to the usual elicitation of an interaction
or response to denote a contact, evidence from certain mammals suggests
a mechanism capable of blocking a social response. Such a mechanism

1. The Social Use of Space 105

which defines the probability of a contact being socially "perceived'' shall
be called n'.

ria == number of contacts made while the given animal is in the "re-
sponsive" state.

fa = (ria/t) is the frequency of responsive contacts when the animal in
question is in the "responsive" state over all time.

Since each contact between two individuals, at least one of whom is in
the responsive state, is followed by a refractory period a, characteristic of
each of the responsive individuals, and since there are ?ia such contacts in
time t, then the individual is in the refractory state for a total time ana.
Clearly, the total time ta in which the individual is in a responsive state
is t — an'Ua.

Since na = tfa then

ta = t - aix'Ua = t - afx'tfa = t{l - OCuJa) (38)

In this sum of refractory intervals, ta, contacts will be made at frequency
/c, but all such contacts are made while the animal is responsive so that

Ua = fda = fct{l — CXIl'fa) = tfa

or, since /a = ria/t

fa = /.(I - V/a) (39)

Thus

1 — anja
It will also be helpful to rearrange Eq. (39) to obtain /„ as a function of /d
/. = /„ + ccn'fafc =/«(!+ an%)

/« = , / ., (41)

We may also define the frequency of refractory contacts by each in-
dividual over all time

fp =fc-fa (42)

Contacts between individuals will be of three kinds: (a) both individuals
responsive; (b) both refractory; and (c) one individual responsive and
the other refractory. A given individual meets fc other individuals in unit

106 John B. Calhoun

time. Of these J\ contacts, fa are with individuals in a responsive state.
Hence, the probabiHty pa that any given encounter will be with a responsive
individual will be

Pa = (fa/f.) (43)

Therefore, of all the encounters /„ in unit time which the given individual
makes while it is responsive, the number

faa = Pa fa (44)

will be with other responsive animals. Thus, /„<, may be considered the
(absolute) frecjuency of responsive-responsive encounters. Substituting
(43) into (44) we obtain

faa = y (45)

Using Eq. (41)

•^'"' ^ /.(I + an%r ^ (1 + «M7e)2 ^^^^

In like manner we can define fpp as the frequency of contacts of individuals
both of whom are refractory and of fap for the freciuency of contacts in
which one individual is responsive and the other refractory.

By an argument analogous to that given above we arrive at the
formulation

fr>.='^ = ^^^^-^ (47)

Je Je

Using Eq. (41)
J pp ~

^' Vl + V/JJ/ -^ -^ \ l+V/e /

(1 -^an%y

Again:

(48)

/.. = -^ = ?^^^I^ (49)

1. The Social Use of Space

And using Eq. (41)

Jap

107

2/.

1 + an%
2

1 + an'fc

(1 +a/z70-

•/.(I + an% - 1)
1 + an'fc

f50)

We can obtain an explicit relation between /„p and N and /aa and N by-
substituting Eq. (34) respectively into Eqs. (50) and (46) :

And

Jap

J aa

2aix'tx-{N

[1 +aMM'(iV - \)J

y.{N - 1)

[1 +a,x,i'{N - \)J

(51)

(52)

The function /«« = 0 when A^ = 1 and also faa approaches 0 as A" tends to
infinity. Since faa is continuous and differentiable for all positive values of
A", it has a maximum A'^, for some value of fan at which the derivative of
faa w'ith respect to A" is zero

dfa

dNr

= 0 =

and
Hence

m[1 + ann.'{N - !)][! - ann'jN - 1)]

[1 +«m/(A^- 1)?

1 - auitx'{N„. - 1) = 0

Nm- \ =

afiix

ocuii

Nm - 1

(53)

(54)

(55)

Thus, the larger a, the smaller Nm. In other words, the position of the
maximum shifts to the left as a increases. In order to find the maximal
value of faa, faV, we may substitute Eq. (55) into (52) obtaining

/i:^ = 1/4V (56)

In other words, the number of responsive-responsive contacts in unit
time decreases as the refractory time increases. Since the refractory time.

108 John B. Calhoun

a, is assumed to increase when the intensity of interaction, i-, increases,
faa decreases as intensity of interaction increases.

When the A^" of a species has attained an evolutionary steady state,
designated as Nb, n and ix' will each have values of 1.0. When n = 1.0 it
will be designated tib. Existence of an A''^ steady state does not mean that
the temporal A^ camiot fluctuate within the lifetime of a species or the
history of a population. Rather, it means there is a particular A^ compatible
with ixb- At this Nb, with its /X6, all contacts are perceived. That is, n' = 1.0,
and whenever /jl' = 1.0 it will be designated fxb- Not only will all contacts
by responsive individuals be perceived, but each member of A^6 will interact
with the same average intensity and, thus, a becomes ab. Obviously Nb
is the Nm toward which a species "strives." In this "striving," which may
be either maturational or evolutionary in terms of units of time, A^ may
vary as a function of ex, or a may vary as a function of A''. At that A^ it
follows from Eqs. (55) and (56) that:

a5/i:^ = 0.25 (57)

This holds for all A^;,.

<*b/ia^ defines the maximum satisfaction from social interaction and
will hereafter be referred to as db.

The usual intensity, i, of interaction, which determines a, since P = a,
may be considered as basically under genetic control. Similarly, n and m'
may be considered to be normal expressions of genetic factors in so long
as A^6 is approximately realized and the members of N experience conditions
in harmony with their genetic constitution, that is to say that the environ-
mental conditions approximate those usually experienced by the species
for many prior generations. However, abnormal environmental circum-
stances may so alter physiology and condition behavior that i, n, and n'
diverge from the u, Hb, and Mb' appropriate to A^'fe. In these circumstances n
and n' no longer each equal 1.0, nor is u in harmony with A''^ in the sense
that afaa will lead to maximum satisfaction from social interaction. Yet,
regardless of how i, fi, and fi' have diverged during maturation, this maxi-
mum may be attained if the species adjusts by attaining that A^, different
from Nb, such that :

a/i:^ = 0.25 = d':' (58)

Theta, the maximal and also optimal satisfaction from social interaction,
is here designated as d^"^ or just 6o to indicate its possible attainment at
some other N than A^6.

Interactions whose frequency has been designated by fap require special
consideration.

1. The Social Use of Space 109

For clarification :/ap = fap + f^a

And f^p = frequency with which a given responsive individual interacts

with refractory individuals, while it is itself in the responsive

state.
And fpa = frequency with which a given refractory individual interacts

with responsive individuals, while it is itself in the refractory

state.

It can be demonstrated that f,'^ = f^^ .

Therefore

Zip = 0.5/„, (59)

If we let a„, represent that a appropriate to A^„, and /i™\ then from Eq.
(55)

Similarly, for Nb with /x and /x' = 1-0

A^6 - 1

(60)

(61)

If we assign /^"^ as the fap characterizing Nm when /i^^ represents the
maximal value of /„„ (e.g., see Eq. (56)), then by utilizing Eqs. (51),
(52), (55), and (60) and considering the fact that /x' = 1-0, it follows that

/S^ = 2/i:^ (62)

And from Eq. (59) it follows that, when n and m' each equals 1.0,

fT' = fa:' (63)

For clarification, it is to be noted that Eqs. (62) and (63) refer to the con-
dition when fjL and m' remain unchanged at the fxb and (Xb values appropriate
to Nb, but a adjusts to the existing A^ according to Eq. (60) so that the
existing N becomes an A^^*") differing from Nb in most instances.
As already demonstrated

eo = af'-' (64)

where do represents the maximal, and for this special case also the optimal,
amount of time an individual can remain in that refractory state denoting
satisfaction from social interaction. At iV„„ where Bo is realized a/„'j"'^

110 John B. Calhoun

amount of time is spent in frustrating refractory periods.
Here

e'r' = af:i-' (65)

And from Eqs. (63), (64), and (65) it is obvious that

qCw) _ n(m)

(66)

Since Nb is a special, and the most important, case of Nm, Eq. (66)
represents a significant consequence of evolution, as well as adjustment to
current group size different from Nb. It means that when members of a
group attempt to maximize satisfaction from social interaction, they will
of necessity spend an equivalent amount of time experiencing frustration
from social interaction. Evolution having transpired in such a system of
social physics, physiology must be in harmony with this normal degree of
frustration. Likewise, any marked decrease or increase of /„p from /^p"'^
should prove stressful.

From Eqs. (51), (59), and (61), when mm' and a remain appropriate
to Nb but A'' fluctuates, it follows that

^ MN - ly-

As N approaches zero, /„p approaches zero. As N approaches infinity,
lap approaches the fc characteristic of Nb, that is when/^ = Nb — I = l/ab.
Yet at the same time (see discussion following Eq. (52)) /„„ approaches
zero as N approaches infinity with reference to its divergence from Nb-
At N's much larger than Nb the frequency of contacts resulting in refrac-
tory periods (i.e., faa and f^p) comes to approximate the total contacts
transpiring in Nb] however, practically all of such contacts are of the type
frustrating to individuals having returned to the responsive state.

1. Terms and Equations

The following assembly of definitions will facilitate understanding later
discussions. Insofar as possible the N animals in the group will serve as the
basis of the definitions. Some terms utilized in later sections will also be
included here.

N = Total niunber of individuals in the group. In the strictest sense,
a group is defined by habitation of an exclusive area in which
each resident member has a good chance of contacting all others.
d = Target diameter of an individual. In the simplest sense, d
specifies the actual physical diameter with the "animal" having
no more d properties than a billiard ball. Included under d are

J. The Social Use of Space ^

any characteristics such as bright color, vocalizations, odoi, or
upright posture which enhance the hkelihood of an individual
being perceived by its associate. Through evolution and matura-
tion certain species, particularly man, acquire the capacities to
utilize nonphysical characteristics to alter target diameter.
These nonphysical characteristics include attitudes and values
whose possession influences the likelihood of the holder being
perceived and responded to by his associates.
V = "Velocity" with which an individual ''moves" through its en-
vironment. It includes all properties which enhance the likeli-
hood of one individual approaching its associates. Thus, in addi-
tion to including actual velocity, it includes all sensory mecha-
nisms which extend the individual's perception of others in any
direction along its travel path. Thus, where r is the radius of
perception beyond the physical bounds of the individual, v be-
comes rLt-\ see Eq. (35). Furthermore, v = rU-^ must actually
become more complex than this. Animals further vary in the
number of trips per unit time. See previous discussion in Sec-
tions III, A, 1 and A, 4 which deal with how emotionality alters
the frequency of trips. Therefore, if we let:

D^ = r = radius of perception

^2 = number of trips per unit time; or any time or place
pattern of movement which alters probability of con-
tacting others

v-i .-^ Lt-^ = actual velocity

Then biological velocity, v, becomes:

Note : Here the product is used in the sense of a function
of
When V is considered in later discussions it will have all these
connotations.

r = radius of perception as discussed above.
A = area inhabited by the N individuals, each of whom has
a good opportunity of contacting any other member

oiN.
n = (dv/A) is a communication-enhancing or contact-
producing factor. By considering v in its simplest sense,
n becomes (drv/A) as soon as the perception sivath

112 John B. Calhoun

determined by the individual's capacity to perceive
beyond its own physical bounds comes into play.
n' = K communication-inhibiting or contact-blinding factor.
It reflects a psychological property permitting the in-
dividual to ignore a contact resulting from n. n' must
derive from {d'r'v' /A') factors. See Section XIII, B, 4
for further treatment of n' .
i = Intensity of action of one individual toward another

upon contact.
a = ^■^ the duration of the refractory period following the
contact of a responsive individual with some other
individual. Contacts made by an individual while it is
in a refractory state have no influence upon its a. In
some way the refractory period is a consequence of the
intensity of interaction. It is thus the result of the
interplay between the action of each individual toward
the other. I have, therefore, assumed that the product
of these intensities of action represents a first approxi-
mation of a proportionality to the duration of the
refractory period.

faa = The frequency with which one individual, while in the
responsive state, meets other individuals, who are also
in the responsive state.

/„p = The frequency with which an individual, while in the
responsive state, meets nonresponsive ones (i.e., those
in the a refractory state) .

aaa = Refractory period produced in each individual after
each of the faa interactions in which it is involved. «„«
produces satisfaction.

cxap = Refractory period produced in the responsive individual
after each f^p interaction. a„p produces frustration at
least in the sense of being a nonspecific stressor of
physiology.

ocaa = ocap wlth regard to duration.
da or

0W = afaa, the amount of time per unit time spent in satis-
fying refractory periods, da represents the consequences
of positively affective interaction.
Of or

^(/) = olJ'^j,, the amount of time spent in the frustrating and
physiologically stressful state.

df approaches zero as N approaches I.O

1. The Social Use of Space 113

df approaches 0.25 at Nb

df approaches 1.0 as A^ approaches infinity with

reference to Nb-
Of represents the consequences of negatively affective
interactions.
0<»') = amfi":^^ = maximal da = 0.25; see Eq. (57).
j^(m) ^ That A^ at which S^"'^ results. In other words, at N""
satisfaction from social interaction is maximized, but
at A'"^'"^ df"^ = d^"'\ that is, there is as much frustra-
tion as satisfaction from social interaction.
A^6 = The basic group size of a species living under those
conditions to which it is most adapted. A^6 is a special
case of N^'"\ 0f \ e^/\ ab, u, ixb, and Mb' represent values
appropriate to Nb- Here 6^^^ and df^ always = 0.25,
as may be seen from Eqs. (57) and (63). At Nb both
Mb and nb must equal 1.0. d^J"\ dj'"^ represent values
appropriate to A^c-) in which ^i'"> - ^f ^ and 0)'"> =

However, a^"'^ i^"'\ m^""^ and m'^'"^ at A^^'") may all
differ from comparable values appropriate to Nb.
Q(o) _ Q{m) £q^. ^ij ^ Q^j^gj. ^j^g^j^ ^^^ although quantitatively

9^^°\ d^J"\ and da'"^ all = 0.25.
No = Any A^ when da = d^°\ No may equal Nb, but when it
differs from Nb, some alteration in n, ij.', or a permits
attainment of the optimum 6 a, that is dj .

2. Interaction P'unctioxs Stated in Terms of A^

The number of indi\iduals inhabiting an area is more readily measured
than any other function relating to this model of social interaction. There-
fore, it will be helpful to state all other functions in terms of A^:

/. = n{N - 1) (34)

If intensity of interaction is labile to the point that 0^'"* can always be
attained, then

^ (60)

which means that :

jlmlJ-m {N m — 1)

\n,ny^n^'{Nm " 1)/

(68)

JJ4 John B. Calhoun

A core aspect of this thesis is that whenever ju increases above m6 there
will be compensatory shifts in m' such that mm' will again eciual 1.0. There-
fore, a and i will gradually become a function of N„„ or we might rather
say that the members of the group attempt to adjust their intensities of
interaction to make any existing A'', regardless of how much it has diverged
from Nb, become A„,. Thus, Eqs. (60) and (68) become:

Oim —

N - 1

(69)

(70)

At Nb it follows from Eqs. (56) and (69) that the maximum frequency
of interaction of one responsive individual with other responsive ones
becomes :

Hrn) ^ ^f - 1 (71)

J aa A

Accepting the logic above that in time all individuals will attempt to ad-
just their intensity of interaction compatible with any existing N, it
follows that:

A — 1

fm) ^ f! i (72)

Jaa ^ ^ ^

It further follows from Eqs. (02) and (69), where mm' tend to adjust to
1.0 and i adjusts to maximize <?„ regardless of change in N, that:

.(.) . ^(^ - 1) = ^(^ - '^ (73)

•''"' [1 + a(N - i)y 4

And similarly at/i"'\ it follows from Eqs. (51), (59), and (69) that:

^ a,HN - 1)^ ^ ,HN - 1)
•'"^ [1 + a(N - l)y 4

Equations (73) and (74) must be kept in mind while reading Section
XIII, B. It has already been pointed out in the discussion following Eq.
(66) that in the evolutionary steady state dj"'^ = di'"\ Here [see Eq.
(63)], f^a^ = fai'"\ This is a major premise of this paper, that animals
"strive" to experience equal amounts of satisfaction and frustration from
social interaction. But note what happens according to Eqs. (73) and
(74) when ju varies. If m increases and the members attempt to optimize

1. The Social Use of Space 115

satisfaction, to attain 6^„"'\ then

-T-, = fi, or df = ixd^o"'\

aim) ^' J r-

Thus, excess frustration will increase proportional to the increase in m- For
this reason, animals will always be conservative in that they will attempt
to reject any changes leading to an increase in fj..

Where /j. = fib = 1.0 and A'' becomes Nm [also refer to Eq. (2;")) ]

N - 1

fim) ^ jfUn) ^ 1 /75)

J aa J ap j \ "/

In some circumstances n may be more labile than i as an adjust! ve mecha-
nism to changes in A^ from Nb- Where intensity of interaction remains
constant at that level appropriate to Nb, a remains ab. And yet Eq. (55)
reveals that the N differing from A^6 can become No provided:

No = l + -^, (76)

(XblJio^

Also by analogy to Eq. (69) :

Nb - 1
Then substituting Eq. (77) into (76) :

(77)

iVo - 1 + ^^^ (78)

Therefore :

, Nb

MoM =

No - 1

(79)

In the original change of No from Nb, m and /x' were //& and nb and each
was therefore equal to 1.0. However, we are here concerned with the case
when n is labile, that is, it can become different from /xb. Furthermore,
fib can be ignored since any change in fib must await some stability in the
change of fx. Thus, when a remains at ab

,o = ^^^ (80)

^ No-\

Where intensity of interaction remains constant, and N i represents the

116 John B. Calhoun

value of N at the inflection point of /^^, that is at the point where the
second derivative of /„<, as a function of N is zero, it may be shown that

Ni= \+— (81)

abHo

Then inserting Eq. (77) into Eq. (81)

;,,.!+ ^J^hjZ^ (82)

Mo
Then inserting Eq. (80) into Eq. (82)

Ni= \ + 2(No - 1) (83)

Where A^ is Ni, Eq. (83) becomes

iV, = 1 + 2(iV6 - 1) - (84)

= 2N,- I (85)

B. Basic Processes Involved in Social Interaction

1. Satisfaction and Frustration as a Function of Group Size

Satisfaction and frustration from social interaction are by definition
measured, respectively, by da and 6/. Full satiation attains at d^""^ and
optimum frustration at dj"'\ These equivalent cjuantities are equally
necessary for the individual to persist in an optimum state. We shall here
be concerned with the effects upon 6a and df resulting from varying A''
when ab, fib, and fxb remain constant and appropriate to Nb. It must be
recalled from the statement preceding Eq. (65) that df''^ is not used in
the sense of the maximum af'^p but rather as the amount of time spent in
frustrating refractory periods at iV„t, that N where the maximum amount
of time, 0^J^\ is spent in satisfying refractory periods. As implied in Table
XI, B and Fig. 35, af'^p attains a maximal value when A^ = infinity.

As Nb increases i and a must decrease in order to maintain d^'"^ [see
Eqs. (56), (61), and (69-71)]. Likewise, for any arbitrary series of a,
such as 1.0, 0.75, 0.5, 0.25, 0.1, 0.05, 0.025, and 0.01, there must be respec-
tive Nb at which a/^^^ = 0f \ Each such a with its corresponding Nb
might be considered as representing a distinct species. For each species
circumstances may cause N to diverge from its Nb. In any such divergence
da diminishes, and for the species where a remains constant, /„« will exhibit
changes proportional to da. Thus, faa may be taken as an index of the degree
to which changes in A^ from Nb diminish satisfaction from social interac-

1. The Social Use of Space

117

tion. Such reductions of /„„ are shown in Fig. 34; f^J^\ Eq. (75), is that
faa denoted by the point where the dashed hne intersects each soUd hne
curve. Dropping vertically to the abscissa from each such intersection
defines the Nb for which that a is appropriate in the sense of optimizing
satisfaction, d^J"\ In each case, regardless of the size of Nb, af["^^ = 0.25.

Fig. 34. Frequency of satisfactory social interactions {faa) as a function of density
(iV) and refractory period (a) . The intersectiori of the dashed line and any solid line
curve defines the basic group size, Nh, appropriate to that a. All «/„„ defined by these
intersections represent a constant, e^^ , which defines the optimum amount of satiation
which can be experienced by any member of an iVt- The values on the ordinate indicated
by N intersecting the dashed line represent /i™ ^ = (iV - l)/4 when n and m' each
equals 1.0 or tin' = 1.0. For any given /^^\ a can be determined by finding the A^ at
which /ao intersects the dashed line. Here a = \/{N — 1).

Each A^6, so defined, represents a distinct species in the sense that there
have arisen genetic alterations in i, such that there exists the highest
probability of a steady state in which «/„„ optimizes satisfaction from social
interaction.

Some of the insights revealed in Fig. 34 may be more explicitly compre-
hended by examining Table XL In the left-hand part A of this table, sue-

118 John B. Calhoun

cessive doublings of Nb are presented. For each Nb those values for fc, u,
ab, and f^a^ are given which are requisite for each member to attain on the
average an optimum satisfaction, do = afaa\ from social interaction. It is
apparent that as Nb increases, ab and ib decrease, while fc and f^^^ increase.
As the Nb group size increases, each individual will have more fc contacts
with associates, of which one-fourth wall be satisfying (i.e., fit'a^/fc = i).
For clarification, I might add that for each individual on the average at
every Nb, another one-fourth of the contacts are of the frustrating f'^p
type, while the remaining one-half of the contacts transpire while in the
refractory state which involve fp^ or fpp contacts. Values for A^6 = 12 are
shown in italics for reference because of the apparent importance of groups
of this size. It may be seen that i for Nb = 12 is only one-third that for
Nb = 2. Nb must increase from 12 to 121 for a similar decrease in intensity
of interaction to be necessary.

In the now voluminous literature on "stress," many papers deal with the
physiological repercussions accompanying change in group size. Yet these
reveal little concerning how much physiological disturbance might be ex-
pected to result from a given change in group size. The prior model of
social interaction will now be examined to determine what insight the
model provides, under the assumption that it approximates reality.

Table XI
Normative Relative Values of Interaction Factors"

A. When N =

B. When a = 0.091 ''

Am)
J aa

faa

0..5fa%

Oa "dd"

1.000

0.25

0.84

0.08

0.07644 0.1736

0.007

0.333

0.577

0.75

1.85

0.51

0.16835 0.0816

0.046

0.143

0.378

1.75

2.62

1.67

0.2384 0.0116

0.142

0.091

0.302

2.75

0.250 0.00

0.25

0.067

0.258

3.75

2.68

3.66

0.2439 0.0061

0.333

0.032

0.180

7.75

2.12

5.99

0.1929 0.0571

0.545

0.016

0.126

15.75

1.39

7.97

0.1265 0.1235

0.725

128

127

0.008

0.089

31.75

0.80

8.71

0.0728 0.1772

0.793

° All values are relative to the intensitj' of interaction, 1.0, appropriate to a group of
2 individuals, when n = 1.0.

'' The a appropriate for Nb = 12.

' 0.5 fap when a remains constant approaches 1/a as N approaches infinity. In this
case with a = 0.091, 1/a = 11.0. In other words 0.5 fap approaches iVj — 1. 0.5 fa,, is
used in the sense of fap-

1. The Social Use of Space 119

Two deviations from the consequences of social interaction appropriate
to Nb stand out as the logical candidates as physiological stressors. First
there is the situation in which an existing da is less than the optimum ^^"'^
"or do. This difference is designated as the satiation deficit, dd, where

0d = abfil'^ - CChfaa = do- da (86)

The second stressor is df. It presents a philosophical problem with regard
to its assessment as a stressor. At A^6, df = do. The mechanics of interaction
according to the model are such that optimizing (i.e., maximizing) satis-
faction from social interaction leads to an equivalent amount of frustra-
tion. During evofution, physiology must have been altered such that it
became compatible with this amount of frustration. 6; may be thought of
as a nonspecific stressor whose presence in d^"^ amount, that is the amount
which will arise when Nm is also Nh, is necessary for stimulating physiology
to an optimum level. When df is below optimum, df < d^/"'\ there will be a
deficit in the nonspecific stressors required to maintain physiology at
normal levels. Above optimum levels, when df > dj"'\ df may be considered
truly as a stressor to the extent that it exceeds d^^K As TV becomes greater
than Nb, df becomes greater than the optimum value of 0.25, and as N
approaches infinity df approaches 1.0. This means that nonspecific stressors
in the sense of frustration from social interaction can never exceed four
times the optimum level. For these reasons I choose to examine merely
how df varies as a function of A^ rather than making any effort to evaluate
any possible differential effect resulting from df being greater or less than
the optimum.

To see directly how these two stressors, dd and df, vary as N changes, a
specific case for Nb = 12 is given in Part B of Table XI. Here again, as in
Fig. 34, it may be seen that the fr.equency of satisfactory interactions,
faa, declines following either decreases or increases in N from Nb. However,
frustrating interactions, f^p decline as A^ declines below Nb, and likewise
increase as N increases above Nb. There results an approximately 70%
deficit in satiation when N declines to 2 or increases to about 122 from the
Nb = 12. Somewhat more marked changes from the optimum frustrations
follow changes in N from Nb.

In order that the change in df as a function of the deviation of A^ from
Nb may be visualized, they were calculated (Fig. 35) for Nb = 2 and
Nb = 12, for which appropriate intensities of interaction are, respectively,
1.0 and 0.302. Only increases in df at N's above Nb are shown. As Nb in-
creases from 2 to 12, the respective intervening curves for df shift to the
right, that is, it takes slightly greater relative increases from Nb to produce
an equivalent increase in df. For all practical purposes the df curves for
all Nb above 12 are identical with that of Nb = 12. Initial increases in N

120

John B. Calhoun

above Nb produce the greatest increase in 6/. Later equivalent increases in
N produce less and less increments to Of, frustration.

Satiation deficit, dd, increases (Fig. 35) in a somewhat similar fashion
as df, but it takes somewhat larger increments in N to produce comparable
increments in dd. Whereas df most likely represents a quantity of social
nonspecific stressors, Od most likely reflects emotion of a kind which on the
human level we call sadness, foreboding, apprehension, or home sickness.
It represents the physiological consequences resulting from needed and
perhaps known social interaction.

8=017424

Fig. 35. Satiation deficit and frustration at greater than optimum group size.
The inset figure defines the N of ma.ximum decrease in the satiation d, per unit increase
oiN.

A word of explanation is in order to reveal why I selected the nonfinite
Nh = 12.11 instead of 12.0 for examining Od in Fig. 35. In brief, I reasoned
that there must be some Od not compatible with maintaining social life at
that corresponding A". When this A^ is reached there must be some genetic
change transpiring which so reduces intensity of interaction that do may be
restored. On a rather arbitrary basis, I selected dd = 0.174, which is equiva-
lent to a 70^f decrease in 9a from do. Reacquisition of do at this N reciuires
i to decrease to 0.3 of its former level. By such criteria, successively larger
Ni, starting with Nb = 2, will form a series of 2, 12.11, 124.45, etc. Any

'(3)

such saltatorial series of Nb's may be designated N^ \ Ny, A"

]V^"\ The dd curves for all A^6 of 12.11 and above will cross the 0.174

horizontal dd line within the small black rectangular area superimposed on

I. The Social Use of Space 121

this line in Fig. 35. Further elaboration of such saltatorial series of Nb is
given in Section XIII, B, 3.

It must be kept in mind that a basic assumption underlying this general
formulation of social interaction is that the mean free path of an individual
must be large in comparison with the target diameter represented by any
other individual. That is, animals must not be so crowded that one in-
dividual becomes so physically hemmed in or surrounded by a few others
that opportunity to contact many of its associates becomes markedly re-
duced. When an experimental study violates this condition, what has been
said in the above statements will not apply.

2. The Buddixg Off of Social Groups

As the group size increases beyond Nb both the deficit in satiation, dd,
and the amount of frustration, 9f, increase in so long as intensity of inter-
action remains constant. Members of the group will find participation in
it both less and less satisfactory and more and more stressful. The question
arises, "At what point will members find conditions so unbearable that
they will leave or at which the group will split?" Changes in da, or da, as a
consequence of changes in A^, are proportional to /„„. Therefore, we would
like to know if there is some N from which any given change in N brings
about a greater change in faa than a similar change from any other N. The
second derivative, f^'a = 0, occurs when: A^ = 1 + 2/(a6Mo) [see Eq.
(81) ]. This N shall be referred to as Ni.

For Nb = 12.11 this arithmetic inflection point comes &t N = 23.22.
As may be seen from the inset graph in Fig. 35 this N marks the point of
maximal change in da with a given change in N. At this point an increase
in N produces a greater decrease in da than a similar change at any other
N greater than Nb. Furthermore, at Ni, dd, as calculated by Eq. (86), for
any N will always be 0.0278, which represents an 11% deficit in do.

For howler monkeys and man, where the basic A'' appears to be about 12
adults, the social group size rarely exceeds 2Nb unless, as in the case with
man, the next well-defined larger group is much larger. It is for this reason
that I suspect that N will split or bud off another group when N approaches
(2Nb — 1) , the point of maximal rate of change in satiation deficit as given
by Eq. (85). In essence, this line of reasoning says that by the time a
group nearly doubles in size from its basic A^, its members will begin to feel
uncomfortable in the sense that they do not find participation in the group
sufficiently satisfactory. This will lead to enough members leaving the
former group so that within each of the two new groups interaction will
produce near optimal results.

122 John B. Calhoun

3. Saltatorial Chaxges in the Basic Group Size

Within most orders, and many lesser taxonomic categories, related
species may be found between which there exist marked differences in the
typical group size. Caribou and elk characteristically maintain large
herds in contrast to the small groups or even isolated pattern of living by
mule deer or moose. During the active breeding season, bats of the species
Myotis lucifugus and M. yumanensis roost singly or in small clusters,
whereas M. grisescens and il/. velifer maintain large assemblies even during
the breeding season. Woodchucks, Marmoia monax, tend to live in isola-
tion, whereas black-tailed prairie dogs, Cynomys luchviciamis, live in large
colonies.

Obviously these represent a select group of comparisons. Although I
shall not attempt to substantiate here the typical group sizes found within
any fairly closely related series of species, examination of many series
suggests that there are within each series several discrete basic group sizes
with an extensive range between any two nearest sizes not represented by
any species. For the present purpose, this conclusion will be accepted as
approximating reality. Then the question follows: "What characteristics
of physiology and group interactions might lead to saltatorial steps in
group size which become fixed by natural selection or cultural evolution?"

In the first place, there must be some condition which induces animals
to assemble in far greater group sizes than their Nb, and this condition must
remain sufficiently strong to prevent splitting of the group as it approaches
2Nb — 1 [see Eq. (85)]. Any spatially restricted but locally abundant
resource might well so act, particularly if response at the source favored
the establishment of a behavioral sink as described in Section XII, C. In-
crease in group size beyond 2A^6 — 1 would accentuate social discomfort
and stress in the sense of increasing 6d and 6/ (Fig. 3o) . At some point these
factors must become so intense as to produce sufficient decrements in re-
production and survival to threaten the survival of the species. There is
no a priori basis for judging what this threshold might be. Beyond 2A^6 — 1
each increment in N produces a smaller increment of 9d and df. Examination
of the curves in Fig. 35 reveals that when dd is about 0.70 of its maximum
and df is slightly over twice its optimum level, any further increments to
N produce little further change in dd and 6/. The horizontal line through all
curves defines this point on each curve and shall be considered empirically
as a limit beyond which further increases in A'^ cannot be tolerated.

If Ni'^ = 12, the limit is Ni^^ = 82 for Of and A^f ^ = 123 for dd. It will
be recalled from Section XII, A that Nb for the Norway rat appeared to
be about 12. Although local colonies approximated this number, all mem-
bers of all colonies were forced to interact at the single source of food and

1. The Social Use of Space 123

•

water. Furthermore, the entire population in the quarter-acre pen sur-
rounded by a rat-proof fence made a closed system out of the entire popu-
lation. At 123 adults (Table X), marked disturbance was in evidence,
witli only a minority of the females reproducing successfully. In the closed
systems contained within a smaller area, described in Section XII, C,
severe reproductive disturbance characterized an A'' of slightly less than 80
adults. Therefore, for the Norway rat at least, when the actual A'' ap-
proaches the increase above Nb presumed to represent the tolerance limit,
a degree of physiological disturbance of sufficient magnitude arise as to
indicate a necessity for some evolutionary adaptation to the increase in
group size for continued survival. Parenthetically, I might add that this
tolerance limit for dd and 6/ was arbitrarily assigned simply with reference
to the slope of the curve, and without prior knowledge that this level
would lead to tolerance limit A^'s so closely approximating those observed
in my experimental studies.

One type of evolutionary change which will reinstate 6 to its optimum
level, that will eliminate dd, involves reducing intensity, i, of interaction.
It will be recalled that i- = a, the refractory period following interaction.
Now if Nb = 12, and the tolerance limit A^ is 123, a reduction of a from
0.091, the a appropriate to Nb = 12, to 0.0082, the a appropriate to Nb =
123, will return d to its optimum value of 0.25. The change in i is 0.3 of its
level at the former A^'b, that is, from i = 0.3 to i = 0.09.

With a hereditary change in behavior amounting to a reduction of i to
0.3 its former level, A^6 changes from 12 to 123. These would then represent
two species, the stem one having an A^^^^ = 12 and residing under those
environmental conditions not necessitating an evolutionary change in
physiology and behavior, and the derived one having an A^^^^ = 123 and
residing in the presence of those environmental conditions forcing the
maintenance of group size far above that of the stem species. If this process
is repeated each time 9d reaches a tolerance limit of approximately 0.607 of
the minimum dd possible, there arises what might be termed a "satiation
deficit saltatorial series of basic N's" which are as follows, starting with
A^,^^^ = 12.

Ni'

123

Ni'

1,359

Ni'

15,088

Ni'

167,645

Ni'

1,862,544

Ni'

20,695,365

A^r

229,937,917

124 John B. Calhoun

Each successive Nh will have an intensity of interaction approximating
0.3 that of the preceding. At the 8th and last listed Nb, intensity of inter-
action would be only 0.0002 that when Nh = 12. It seems rather patent
that no meaningful behavior could transpire with such a reduced intensity
(duration) of interaction. Two-hundred thirty million adults in a semi-
closed social system can only apply to the world as a whole for the human
species. Reduction of intensity of activity as a means of recovering satia-
tion from social interaction could, in evolutionary terms, likely suffice in
mammals to the third stage of 15,000 adults which entails a reduction of i
to 0.09 of that appropriate to A''^ = 12.

A similar series of Nb can be calculated with reference to Of = 0.775 or
the tolerance limit involved in shifting Nb from 12 to 82. Optimum df =
0.25 can be regained if at this limit i is reduced to about 0.36 of its intensity
at the former Nb. Such a "frustration saltatorial Nb series" becomes:

Nt'^

Ni'^

^(3)

597

m^^

4,491

^(5)

32,343

238,153

iV^'

1,753,772

^(8)

12,914,892

Again this series becomes rather absurd at the upper limit because of
the great demand for reducing intensity of interaction. Since semiclosed
systems, at least on the human level, and occasionally with other mammals
do approach some of these A^6, we must ask what other avenues of evolu-
tion exist.

For this we must assume that intensity of interaction remains constant
at some level approximating that for Nb = 12, but that a tolerance limit
for dd and df exists. At the A^ of these limits a change in behavior may take
place which insulates the individual by producing subaggregates in which,
for all practical purposes, the individual at any particular time is a mem-
ber of a subgroup in which Nb = 12, even though many other subgroups
exist in the environs. The individual may be a member of several such
groups but participates in only one at a time. Such changes in behavior
can be considered to be of either genetic or cultural origin. In either case,
so long as any tolerance limit for 6/ and/or 6d exists, there must be salta-
torial steps between successive Nb, and only a few such steps are possible
even if the tolerance limit arises at a somewhat lower level than hypothe-
sized above.

If later research supports this hypothesis it will haxe considerable

7. The Social Use of Space 125

bearing on our understanding of the course of evolution involving change
in group size. It will mean that gradual changes in heredity or culture will
rarely have transpired. Rather, from the pool of gene variability accumu-
lated in the species, there will be rapid shifts in gene frequencies of many
genes, thus resulting in a new phenotype. In so long as environmental
conditions facilitate maintenance of its Nb by a species, its gene pool may
become quite diverse through the accumulation of mutant genes. Then,
once environmental circumstances force the species to maintain an ele-
vated A'' near its tolerance level for dd or 9/, an extreme selection pressure
will arise for reducing the frequency of all genes except those which adapt
the species to its new N. A genetically variable A^^^^ species will thus
rapidly be transformed into a genetically rigid Nl"^^ species.

On the cultural level such a process of saltatorial change in basic group
size demands that the value system which dictates acceptable roles of
action and communication be preserved even after the usual group size
has far exceeded the Nb appropriate for that value system. At the same
time, under the pressure of increases in 6d and 6/, small segments of the
closed system will develop values divergent from the main group. At the
tolerance limit of dd and 6/, when A^ has so diverged from Nl^\ there will
arise a marked and rapid shift to the prevalence of those newer values
appropriate to d^'"^ and 9f"'^ at N^'^K Value frequencies and gene frequencies
become isomorphic in these two avenues through which there can be a
saltatorial evolution from one basic group size to another.

Basic group size for adults only in the primary steps of human cultural
evolution seem to include the 10-16 range, 50, 200, and 2,000. This series
resembles neither of the hypothetical saltatorial group size series except in
its saltatorial character. The hypothetical series merely demonstrated the
kind of changes following from stated assumptions. The exact series
followed by any line of change depends upon the threshold tolerance limit
for dd and 6/ as well as three factors ignored in our discussion up to the
present. Discussion up to this point assumes /x = (dv/A) = 1.0, where
d represented the target diameter of other individuals, A the area inhabited
by the N individuals, and v the velocity of movement of individuals. In
essence, m represented the likelihood in time t of one individual encounter-
ing another.

It can readily be shown from Eqs. (52) and (60) that da, the satiation
from social interaction, i.e., afaa, can remain constant regardless of changes
in fjL. At least this is so if the physiology and behavior of the species is com-
pletely adjustive. From the general form of Eq. (60) where n' = 1.0,
a = 1/[m(A^ — 1 ) ], it follows that each doubling of /x, that is doubling the
likelihood of one indi\'idual meeting another, necessitates a halving of
a, and thus reduces intensity of interaction from (a)^'^ to (a)^/-/2. If we

126 John B. Calhoun

follow the prior assumption that each species has an optimum intensity of
interaction, then each increase in n will have an analogous effect to in-
creasing group size. In other words, increasing n above 1.0 will increase
dd and df. When we are concerned with the effects of changes in m but as-
sume a remains static at the value appropriate to m = 1-0, then a must be
calculated from Eci. (60) with /x' = 1-0 and faa calculated with this a by
using Eq. (52) above and some value of n different from (Xb = 1.0.

For example, consider Nb = 12. Then ab = 0.091, and db = 0.25 (see
Table XI). If A'' doubles and /x remains 1.0, d„ becomes 0.219, but if A^
remains constant at Nb but m doubles to 2.0, 6 becomes 0.195. Thus, a
comparable increase in ^ produces a greater deficit in satiation, 6^, from
social interaction, than does a double of A^.

Thus, saltatorial evolution of A^^^^^ to A^^"^ may be necessitated either
by an increase in A^ or an increase in m- The rate of change in N and ijl
may well offset the tolerance limit of da or 6/ and thus affect the magnitude
of the shift from A''^^^ to N^^^K dv essentially measures the rate of com-
munication and A the space within which this communication takes place.
Thus, n will increase if A remains constant and dv increases, or if dv re-
mains constant and A decreases. If both the rate, that is means, of com-
munication increases and the distance over which communication must
take place decreases, m will increase very rapidly. Detailed consideration of
communication is given in the following section.

4. The m Communication Function

We have already seen that n = (dv/A), as defined by the previous
Eqs. (35) and (80), is a communication-enhancing or contact-producing
factor. (See prior discussion under Terms and Equations, Section XIII,
A, 1.) Other than for pointing out in the latter part of Section XIII, B, 3
that altering /x has much the same consequences as altering N, we have
been content to consider consequences of variability in other functions
when IX remains constant at that value /X6 = 1.0 appropriate to A^6.

I was led to examine the question of the consequences of varying /x
as a result of the observation by Birdsell and by Zimmerman and Cervantes,
cited in Section XII, E. They observed that where a conflict of values
arises in a group there results a reduction in group size. Here, we are con-
cerned with the special case where attitudes or values comprise a major
aspect of the target diameter d. Each member of the group holds some n
number of values by which others recognize it as an appropriate object
for interacting. When some particular ^-aIue is shared by all members, it
may be said to possess a unitary value in contributing to target diameter.
In other words, under this circumstance all individuals possess the same

7. Tlw Social Use of Space 127

target diameter, d = 1.0. With reference to Eq. (:i5), a unity value for
target diameter merely means that there has been genetic or cultural
adaptation to the actual magnitude of d, such that ju = dv/A = 1.0. How-
ever, if an individual expresses a value shared by only a few of his associ-
ates, he will by this fact be much more likely to be perceived by his as-
sociates, and thus more likely to be reacted to by them. His target diam-
eter will be increased. Furthermore, it is logical to assume that the larger
an individual's target diameter, the greater will be the response evoked
from associates. In so long as all other d xqXwq characteristics remain
identical among the members of the group, the one which does vary among
members will assume the sole role of influencing target diameter. As a
first approximation this response-evoking capacity, which I will call S,
of a particular d value can be taken as being inversely proportional to the
probability j) of its being encountered among the members of the group of
N individuals.
Therefore :

S = l/v (87)

Where only one component of d varies, and since those shared compo-
nents of d may be ignored, d ^ S. For the special case where all members
have the same d:

M = {dv)/A = {Sv)/A = 1.0 (88)

Furthermore, where area. A, remains unchanged at the value appropri-
ate to Nb it has the relative value of 1.0. Therefore:

Sv = 1.0

S = l/v (89)

And considering Eqs. (87) and (89)

V ^ p (90)

Lastly:

V = l/S (91)

A'ariability of the target diameter d among individuals means that d
comprises an assembly of traits, physical size, color, vocalizations, behavior,
and attitudes or values, d is the total complex. Components shared by all
members will be referred to as {d) . Those remaining traits, through which
an individual differs from its associates, represent a genetically and cul-
turally determined phenotype to which the response evoked from associ-
ates is a function. By response I here refer solely to actions reflecting the
choosing or rejecting of an associate. Such a response may be a function

128 John B. Calhoun

of the trait itself or may be a function of a recognizable degree of differ-
ence between one individual and its most similar associate. 1 believe that
such degrees of difference form the primary basis for the maturation of
social behavior and social organization within a group. Further treatment
of this topic follows in Sections XIII, B, 5, a and b; XIV, A and B.

Such traits or degrees of difference comprise the units influencing social
behavior. These units will here be called rf-genes. As stated above they may
be of either hereditary, or cultural origin. Any rf-gene, g^^\ may develop
an allelic series of differing or "mutant" forms ^i^' • • • gl^K When degrees
of difference, and not the absolute amount or kind of difference, forms a
(/-gene there can only be two forms of a particular t/-gene, g]l^ and g^^\
where g\]^^ represents a degree of difference from the ideal type, the
ideal d, and g^2^ represents the retention of the ideal traits for which g'^^^
represents the divergence. c?-genes of the type g*"^ will be called dominant
c?-genes, while those of the type g]p will be called recessive c?-genes. d-
genes of the latter type are treated in detail in Sections XIII, B, 5, a and
b. Without specifying the allelic nature of any rf-gene it is obvious that the
target diameter d is a function of (rf), g^^\ g^-\ • • •, ^^"\

Let

Si^ represent the response-evoking capacity of any ith. individual with
reference to the probability of its being chosen by associates as
an object of affection. S'"^^ is related to Schaeffer's love-acceptance
referred to in Section XIY, C.

aS,^"^ represent the response-evoking capacity of any zth individual
affecting the probability of his being rejected by associates. *S^"^
is related to Schaeffer's hostility-rejection referred to in Section
XIV C.

<S^^^ is a function of both (d) and the assembly of dominant d-genes,
while <S^"^ is solely a function of recessive c?-genes. The probability of en-
countering the common (d) assembly of traits will be 1.0. Therefore, from
Eq. (87) the positively affective stimulus-evoking capacity of this com-
monly held assembly of traits will contribute to the *S^-^^ of an individual
inversely proportional to the probability of its being encountered within
the V individuals forming the group. Thus where pg represents the prob-
ability of encountering a particular c?-gene and A, B, C, • • • , represents the
dominant "allele" of c?-genes (1), (2), (3), •••, and there are N — 1
d-genes of the degrees of difference type, then:

S\^' =— + -77 + 47+ ••• +-7^17 (92)

P(d) Pa PoB Pff(Ar_i)

Similarly where a, b, c, • • ', N — 1 represent recessive d-genes of the

1. The Social Use of Space 129

degree of difference type:

Pga Puh Pg(.N-i)

Equation (92) applies strictly only for that single individual which has
preserved all the dominant or ideal traits. For every other individual one
or more terms in Eq. (92) will be missing, depending upon replacement of
the dominant d-genes by recessive ones. Similarly Eq. (93) applies strictly
only to that individual in which all traits, other than those commonly
held (d) , have diverged from the ideal, that is in that individual in which
all non-(d) rf-genes are recessive. For all other individuals one or more
terms in Eq. (93) will be missing.

Consider the case where all d-genes in a group had been identical up
until a particular point in time, at which a particular d-gene, g^^\ "mutated"
to g[^'^ in half the members. In this mutation gf^^' diverged sufficiently
from g[^^ to make quite distinct the derived from the original. Then the
probability of each in the group will be only 0.5 and thus the response-
evoking capacity of each will rise to a relative value of 2.0. For the average
individual S will have increased from 1.0 to 2.0.

The total Nh members of a basic sized group will consist of Ni type 1
individuals possessing g[^^ and A^2 type 2 individuals possessing g[^\ Thus
ATj, = ATi + Ni. Any individual will be considered as being able to encounter
itself in the sense of being aware of its own characteristics.

Considering this premise it follows that the probability pi of any type 1
individual being met by associates becomes:

p, = N,/N, (94)

Similarly, the probability, pa, of type 2 individuals being met by as-
sociates becomes:

p, = N2/N, (95)

From Eq. (87) it follows that response-evoking capacity, *Si, of any
type one individual, and S2 of any type two individual will be respectively:

S, = Nb/Ni (96)

S, = Nb/N2 (97)

Therefore, the mean response-evoking capacity, S, of the Nb individuals
becomes:

Ni-^ N,

130 John B. Calhoun

Substituting E(|8. (9()j and (97) with (98)

S = 2.0 (99)

The S of Eqs. (96) to (99) is essentially that of Eq. (93) in which g[^^
and ^72^' become, respectively, recessive d-genes, g^^^ and gl^K

As may be seen from Eci. (98), whenever A^i or A^2 is zero, that is, all
members of Nb have the same target diameter, the response-evoking
capacity of each member of the group has a relative value of 1.0. However,
Eq. (99) shows that as soon as Nb becomes divided into subgroups A^i and
A''2, even though the divergent A''2 has only one member, the average
response-evoking capacity doubles. The probable consequence of this
doubling depends upon the relative numbers of A'": and A^2- Consider

Table XII

The Influence of Relative Size of Subgroups of Nb = 12
ON Response-Evoking Capacity

«•:

1.0

1.0909

2.0

1.2

2.0

1.3333

2.0

1.5

2.0

1.7143

2.4

2.0

Nb = 12, then when A''! and A''2 have the sizes given below, members of
each will have Si and S2 as shown in Table XII.

The jS of the members of larger subgroups can never exceed twice the
optimum level, but the S of the members of smaller subgroups has a
maximum of N times that where all members of Nb have the same target
diameter. To understand the consequences to an individual resulting from
possession of a large S, we must inquire further as to its implication. In
the first place, it may evoke more frequent responses from associates. If
the group is essentially an Nb one, such an individual will experience more
contacts than otherwise would be anticipated. This will have the same
deleterious consequences to him of being in too large a group. S in this
case may be thought of as increased target diameter, d. On the other hand,
a heightened S may evoke more intense reaction, i, from associates at
time of contact. This will have the consequence of increasing the refractory
periods, a, and thus wath frequency of contacts maintained harmonious

J. The Social Use of Space 131

with A^6, the thetas, both those of satiation and frustration, will be increased.
In this situation the critical point becomes the attitude of associates toward
such indi\'iduals. If the attitude toward this individual possessing rarely-
encountered characteristics is one in which the desired characteristic is
venerated or desired by the majority not possessing it, then the resulting
intense interaction will have the consequence of more frequently resulting
in an aaa satiation type refractory period. However, it is much more likely
that the strange, rare characteristic will elicit an aggressive or rejection
type response leading to an aap frustrating type refractory period for the
individual with the heightened S.

When the iVi subgroup represents a majority, they can achieve a return
of their n and 9's to more nearly normal levels by ejecting the aberrant
A''2 individuals. However, the minority A''2 members of the A^i + A^2 group
suffer most from heightened m and ^'s. Therefore, their seeking escape from
the group becomes a motivating force greater than that of ejection by their
associates.

If neither A^i or A''2 form a clear majority, the most likely result will be
a splitting of the group in half, but with retention of both A''! and A''2 type
individuals in each smaller group. This consequence derives from the
following ;

When S doubles, nb doubles, the new n = 2.0.

The "effort" of the group will be to make the easiest adjustment which
will make ^t = 2.0 = mo- This route lies in reducing the A^ which was an
Nb to an No. From Eq. (78) it follows that:

A^i iV2 1
A^o=l+^ + ^-- (100)

The best approximation any group can make is to divide in half. Each
half must contain nearly equal numbers of A^i and A''2. If all A^i formed a
group spatially distinct from the iV2 members of the former Nb, the m of
every member in each group would automatically return to 1.0 since within
each new and smaller group all members would have the same target
diameter. Thus, with the A'" of each new group being only O.oNb, every
member would experience a marked reduction in satisfaction and frustra-
tion thetas below optimum since the frequency of contacts /c [Ecj. (34)]
would automatically be reduced.

In discussing this concept of a few divergent individuals or even a single
one doubling the n of the entire group, one of my colleagues remarked that
such an increase appears unreasonable. Three examples will suffice to
demonstrate the I'easonableness of this assumption. Barnett (1955) main-

132 John B. Calhoun

tained established groups of Norway rats in large cages. Despite or because
of the existing hierarchy characterizing the group, all rats exhibited rela-
tive amicable relations, one to another. All presumably had developed
nearly identical target diameters. Introduction of a single alien rat im-
mediately produced a state of turmoil within the group, particularly in-
tense actions being directed toward the alien by all members of the estab-
lished group. In such experiments the ahen rat, the one with the markedly
differing target diameter, frequently died within a day or two. This death
came, not as a result of wounds received, but as a result of inability to
accommodate physiologically to the intense action directed toward him.
His df exceeded a threshold compatible with survival.

Or we may take any one of the several incidents publicized by the press
and television during 1960 and 1961 in which "Freedom Riders" engaged in
action which challenged the established value systems of certain segments
of the socially dominant whites in the Deep South of the United States.
Such actions freciuently evoke drastic reaction, including physical \1o-
lence, from members of the established group. Furthermore, many cross
currents of elevated intensity of interaction became generated within
members of the established group as they considered appropriate adjust-
ments to the threat to their held values and to their prior target diameter.

The point I have been making in both the theoretical formulation and
in these examples is that a qualitative change in the target diameter of a
portion of the group will produce both an increase in n and an increase in
intensity of interaction. Another pertinent example is one I have previously
given (Calhoun, 1956, pp. 87-88). That example concerns the establish-
ment of a new group, designated as "C57 Colony IB," from two smaller
groups of mice not previously having contact with each other. After pro-
vision of an access door through the wall previously separating the two
groups, the males from the socially more integrated group invaded the
living space of the other males. Not only did the former attack the latter,
but the males from the more socially integrated group began a period of
intense fighting among themselves such as had never previously been
observed.

Admittedly, all three of these examples include introduction of aliens
into an established group rather than the hypothesized divergence in
target diameter of a portion of a single group. However, it is difficult to
imagine a portion of a group altering their target diameter without tempo-
rary partial isolation from the remainder. So, in effect, the consequences
of divergence of target diameters among members of a group will be the
same regardless of how the group is assembled.

In actuality

M = {Sv/A) (101)

1. The Social Use of Space 133

However, in most instances, I shall continue to consider fj. = (dv/A) as
previously. But when so doing, it must be understood that d is used in the
sense of its S response-evoking capacity.

When S changes from 1.0 to 2.0 for the reasons relating to Eq. (99), n
will no longer be m6 = (dv/A) = 1.0, but /x will then become (2dv/A) =
2.0. Inserting fx = 2.0 into Eq. (78), in which fx' remains 1.0, for the special
case where A^";, = 12, then No becomes 6.5. This means that in the presence
of conflicting values group size must be reduced for each individual to
maintain its do, its optimum level of satiation from social interaction. In-
crease in fi follows increases in d or v, or decrease in A. Regardless of the
origin of the increase in /z, reduction in group size should follow.

Such reduction in group size should not be instantaneous. Consider
Nb = 12, ab = 0.091 and n = 2.0 and the group has not yet fragmented.
From Eq. (82) it is obvious that Nb = Ni, when Ho = 2.0 and a = ab,
and that da, the deficit in satiation from social interaction, will be as great
as if /x had remained unchanged at 1 .0 and Nb increased to [1 + 2 ( A^6 — 1 ) ].
[Refer to Section XIII, B, 2.] This is a very interesting consequence for
it means that when n increases to 2.0, Nb = Ni. Recall that Ni is that N
at which an increment in N brings about the greatest change in dd. Since
groups do resist division and since any increase in /x is likely to be gradual,
the most likely time for fragmentation of the group is when fx becomes 2.0
and Nb = Ni. Then Nb will divide into two groups approximating No
determined by Eq. (78). Roughly, this says that when the ease of com-
munication doubles as a result of a doubling of the response-evoking
capacity S, the group will approximately divide in half if it is to optimize
satiation from social interaction.

This process of halving the basic group size each time the ease of com-
munication becomes twdce as efficient cannot continue long if Nb = 12,
because by the fourth doubling of fx, sexual reproduction could no longer
be tolerated. That is, No would be less than two individuals. The practice
of divorce by the human animal reflects this process. We now have another
question raised: "What avenue of adaptation or adjustment is open if Nb
remains 12 and i remains unchanged at uV

Although Id theoretically may be defined in terms of attributes of d, v,
and A external to the organism, any solution to this question demands
that n must in effect be reduced back to 1.0 by some compensating
mechanism.

This mechanism which alters the probability of a contact being socially
perceived has been called n'. In Eqs. (38) to (55) it was shown that mm'
represents the appropriate interaction between these two factors. So far
/x' has been elaborated no further, /xju' then becomes the communication
constant, more explicitly stated as {dv/A)ix. Since m can vary as a result
of any one of its contained factors, d, v, or A, fluctuating alone, one cannot

134 John B. Calhoun

escape the conclusion that during evolution a separate compensating
mechanism for each must have arisen. This means that there is a d', a v',
and an A', and that /jl' = id'v' / A'). Furthermore, when mm' = 1-0, [i =
\/\x. Having arrived at these insights, one is logically lead to ask: "What
do d\ v' , and A' most likely represent biologically?" Although answering
this ciuestion is not necessary for the general theoretical formulations, an
attempt to specify their more likely nature may be helpful in searching
for their identification.

d' represents the degree to which the stimuli emanating from any con-
figuration pass unselectively from the sense organs into the memory store
of the central nervous system (CNS). Thus, an increase in d' means
facilitation of passage of stimuli into the CNS, while a decrease in d' indi-
cates impeding or preventing stimuli from getting to the CNS. When the
target diameter of associates increases through evolution by acquiring
more (/-genes, a compensating evolution of a d'-mechanism wall permit a
discrimination among the <i-genes such that in that brief span of time re-
quired for psychological contact only a portion of the c?-genes of the other
individual will be responded to. It must be kept clearly in mind that an in-
crease in the efficiency of the mechanism which serves to alter d' , decreases
d'. Such a decrease in d' imphes the evolution of a filtering device which re-
duces the amount of information about others per unit time arriving at
the sense organs, which is permitted to pass from them into the integrative
centers of the nervous system. Without specifying what CNS structure
serves the d' function, it meets the reciuirements hypothesized by Broad-
bent for his CNS "filter." See Section IX.

v' also represents a process internal to the indi\'idual. It cannot have
any influence upon the motor component of v. Therefore it must affect
the consequences of those sensory capacities which enable the individual
to achieve a psychological contact prior to an actual physical one. This is
the r component of velocity mentioned in Section XIII, A, 1. Just as an
increase in r increases v by decreasing the time required for a contact, so
must a decrease in v' function to increase the time from the moment of
input of a signal from a d-gene at the sensory organ until this transformed
signal reaches and evokes a response at an effector. Thus v' could represent
either a structural or biochemical alteration in the time required for an
impulse to pass over a synapse, or it could be represented by an alteration
in the number of neurons in the circuit which will also alter transmission
time. But we must not confuse the magnitude of v' with the efficiency of
the mechanism involved. An increase in v' means a decrease in the efficiency
of the mechanism, that is increased synaptic transmission, while a decrease
in v' follows from an increase in the efficiency of the mechanism in impeding
the passage of the signal along the circuit between the sense organ to the
effectors.

J. The Social Use of Space 135

A word is required to differentiate clearly d' from v'. d' governs the prol)-
ability of a signal relating to c?-genes getting through the Broadbent typo
hypothesized "filter," while v' applies to the speed of transmission from the
filter to effectors.

In a similar fashion, conceptualization of A' must be in terms of counter-
acting A. Where an A''^ group is living under optimum conditions, A may
be considered to be equal to 1.0. Optimum conditions will continue even
though A changes in so long as A A' = 1.0. For the sake of simplicity we
are considering the case where the only change in the system pertains to
A. Recall that A represents the area which the group shares. In essence
then A alters the time between contacts. Therefore A' must operate in a
similar but opposite direction to .4. Suppose that A increases. In effect
this is equivalent to a decrease in density. Under such circumstances there
can be no internal mechanism enhancing the probability of an actual
contact. Therefore, when A increases, an A' compensating mechanism in-
volves an imagined contact. To the extent that such an imagined contact
leads to an equal a refractory period, an A' mechanism will be effective.
Since an increase in A implies a decrease in A', a decrease in A' means an
increase in the capacity to store memories of associates, which can com-
pensate for their absence. Such an increased storage of memories implies
an increase in cortical mass.

On the other hand, suppose that A decreases. This will reduce the time
elapsing between contacts, and since this will have the same consequence
as increasing N w^hen we are concerned with an Nh group, its members
will experience an increased da and an increase in 6/ above the optimum
level. In this situation A' must function to increase the relative time be-
tween contacts. The only way for this to happen is for the intensity, i, of
interaction to decrease. Recall that intensity of interaction has been
measured in terms of its duration, and that i = (a)^ when a represented
the duration of the refractory period following interaction. Furthermore,
Eq. (60) becomes:

fJLfx'iN - 1)

when we consider the several factors in the general sense where a variation
in one may influence any other.
Then

an' = ^- (103)

^ fxiN - 1)

The reason for stating the equation in this fashion is that n' cannot alter
/i as such but can only alter its effects through changing something else.

136 John B. Calhoun

What Eq. (103j implies is that where A' is the factor which produces the
variabihty iu /x', it can be effective in maintaining 60 but not necessarily
an optimum 6/ (see discussion following Eq. (74)), provided it acts as a
governor on the intensity of interaction independent from the influence
upon i exerted by d. A decrease in the A component of n indicates a de-
crease in the home range a (see Sections II-V). The significance of this
hne of reasoning is that this A governor, which controls i, is likely to be
identical to the one previously postulated for determining the duration of
an outward trip from home. In the general sense, this governor controls
the duration of behaviors. Social interaction merely represents one specific
category of behaviors. Related to the above discussion, it may be noted
that Eqs. (69) and (103) are equivalent since n' = l/n, so long as m = 1-0
or fjL > 1.0. As originally formulated in the discussion before Ec^. (38), n'
simply operated as a probability of an actual contact being perceived.
However, the above and following discussion indicate that m' can exceed
1.0. By imagining contacts, which actually do not occur, through a de-
crease of the A' component of /i', n' can exceed 1.0.

In terms of the model proposed in Section III, A, whenever A decreases,
the rate of firing of the neuronal net of the governor will increase. Thus,
an increase in A' represents an increased rate of firing of the neuronal net
of the governor.

Previously I pointed out that a decrease in A' to compensate for an in-
crease in A might be visualized as an hallucinatory process. There is no
reason why A' camiot equally involve the governor of intensity of activity
when A increases above .-l^ normally appropriate to Nb', a decrease in A'
would represent a slowing of the firing of the neuronal net of the governor,
which change would then permit a longer duration of interaction. This
longer duration of interaction would compensate for the fewer /«« interac-
tions possible in a larger A.

The concept of social perception of contacts subsumed under n' thus
includes a wdde variety of processes. It includes (1) selective acceptance of
stimuli to be integrated in the central nervous system; (2) impedance or
facilitation of passage of signals across neuronal synapses; (3) alteration
in the length of a neuronal circuit; (4) hallucinating a contact; and (5)
governing the intensity of interaction. These m' capacities for adjustment
are most likely to be found farthest advanced in those species in which d,
V, and A fluctuate most markedly with reference to all members of a group
within the life span of every individual. However, over long spans of time
encompassing many generations, a gradual increase in d, v, and A should
increase the complexity of social life possible. At least this conclusion
holds to the extent that enhanced synaptic transmission, increased dura-
tion and intensity of behavior, and increased discriminatory capacity
foster more effective social behavior.

1. The Social Use of Space

137

The following summary of presumed relationships may serve as a basis
for evaluating the consequences of change in the components of ju and m'-

External

Internal

change

Increase in

Decrease in

= Increase in discriminatory capacity

= Impeded synaptic transmission

= Increased duration and intensity of be-
haviors, or
Increased memory storage of social stimuli
permitting hallucinated interactions which
indicate increase in cortical mass

Decrease in Increase in
d d'

= Decreased discriminatory capacity (i.e.,
less selection of stimuli passing the
"filter")

V v' = Enhanced synaptic transmission

A A' = Decreased duration and intensity of be-

havior, or decreased memory storage of
stimuli which are available for producing
hallucinated social interactions which in-
dicates decrease in cortical mass

These relationships represent intuitive logical deductions, d' and A'
serve as fairly satisfactory first approximations. With progressive social
evolution, d and A generally increase. Along with this trend, discriminatory
capacity increases, cortical mass increases, and ability to maintain a be-
havior for longer periods of time increases, all of which tend to harmonize
with the model. However, I am less satisfied with my formulation of v'
(Section XIII, A, 1). The reason for this opinion is that with advancing
social evolution there presumably arises an increase in Vi, that sensory ex-
tension of the seK to produce psychological contact before bodily contact.
This should produce the opposing phenomenon of impeded synaptic trans-
mission. For the present, I can see no rational basis for reconciling impeded
synaptic transmission with social advances.

v' merely connotes alteration of sensory perceptions of others involving
some distance between the two individuals concerned. My supposition
that alteration of synaptic transmission subserves v' may well be wrong.
The important point for the present is simply recognition of the likelihood
of some such mechanism whose evolution is influenced by social conditions.

138 John B. Calhoun

I have pointed out that since fx can change as a result of independent
change in d, v, and A, conseciuently there must be discrete phenomena in
the individual representing what I call d', v' , and A' , which can change
independent of each other. Nevertheless, it is ciuite likely that the compo-
nents of n and /x' do have interactions. The concept of intensity (duration) ,
i, of social action includes control of i by both internal and external factors.
The internal factor is A', the "governor" previously discussed. The ex-
ternal factor is the d of the other. The greater d, the greater i. Thus, d
can influence the governor. This means that an increase in the d of associ-
ates can decrease the A' of self.

In studies with rats recently completed (Calhoun, 1962b) some rats
develop a high v while others develop a low v. Those with a very high v
exhibit high i in terms of both intensity and duration. This suggests that
in some way an increase in v leads to a decrease in A' . Furthermore, male
rats with very low v commonly respond as though they did not make ade-
quate discrimination of the cues emanating from associates. They sexually
mount associates without regard to their age, sex, or sexual receptivity.
These observations suggest that lowering of the motor components of self's
V increases self's d' , which suggests that somehow when an animal decreases
its velocity its ability to discriminate among available social stimuli also
becomes reduced. All I have attempted to do in the preceding paragraphs
is to lay the groundwork for understanding the meaning of the contact-
modifying factor jj.' .

Decreases of n re increase in N : We are here concerned with the special
case where the area. A, remains constant as numbers of individuals, N,
increase. This means that density increases. We have already seen that an
increase in A^ with ^ held constant leads to a deficit, da, in satisfaction from
social interaction as well as an increase above optimum of the frustration,
df, from such interaction. As density increases one should anticipate n
changing before y.' . Therefore we shall consider ix' as remaining constant
at the 1.0 value appropriate to Nb but let N increase. In each instance we
wish A^ to become No, which means that do and 6/"'^ will be optimum. Con-
sider the case where A^b = 12 and intensity of interaction remains at u, it
may be seen from Eq. (80) that successive doublings of No demands suc-
cessively slightly more than halving of jXo'-

IJ'O

1.0

0.478

0.234

0.116

192

0.057

384

0.028

1. The Social Use of Space 139

Where fx = Ho and N = No it may be seen by substituting Eq. (80) into
Eq. i'-U) that the frequency of contacts will remain constant at (Nb — 1)
regardless of the increase in density. Since we are considering the special
case where area, A, remains constant, the constancy in the absolute num-
ber of contacts despite increases in density can only derive from decreases
in the d or the v component of ju = dv/A.

Where decreases in ju transpire within the lifetime of an individual as an
adjustive change to increase in density, and where this decrease in m results
solely from a decrease in the target diameter, d, this change must be re-
flected through reductions in the frequency of usage of signahng mechanisms
or of the intensity of such signals. Vocalizations should occur less frequently,
be less complex, and of reduced intensity. Bodily display characteristics
under voluntary control should similarly be reduced, as should also the
use of chemical signals. Similarly, when a species, A, now exhibits an Nb
of 80-120 or 800-1400 (see Section XIII, B, 3) but there is reason to sus-
pect that at some earlier era it had had an Nb of 2 or 12, such as is still
expressed by a related species (or genus), B, then a comparative examina-
tion of species of types A and B should reveal that members of species B
more frequently resort to the utilization of vocal, display, and chemical
signals and that these are of greater complexity and intensity than in
species A. In making any such comparisons it is well to keep in mind the
earlier hypothesis that the change of an Nb = 2 to an Nb = 12 may lead
to an increase in complexity of the signal. For this reason, it is more ap-
propriate to compare a species w^hose A^6 has a typical range of 7-19 with
one which rarely falls below 80. Such comparisons will provide critical
tests of the general formulation.

In like fashion, velocity, v, may be the factor which becomes reduced as
density increases. This reduction may be expressed in any of the three
aspects of v listed in Section XIII, A, 1. Sense organs may become less
effective with reference to the animals' perception. This reduces ^i. Such
a tactic must be effective only through genetic changes and so can serve
only as a long-term adaptive mechanism. The animal may also reduce its
V2 by w^ay of reducing the number of trips it makes, or by altering the time
and place of its activities. This strategy includes initiating activity while
others are resting, or by becoming cryptic in the sense of becoming arboreal
or subterranean. Finally, the animal may reduce its Vs, its actual rate of
movement.

Although such reductions in d and v will lead to a reduction in m which
will compensate for increase in density, it is obvious that n must decline
so markedly as to present biological limits of effectiveness. Further accom-
modation to increases in density must require m' to decrease also. Reduc-
tion in this communication-inhibiting or contact-blinding factor means
(a) decreased duration or intensity of behaviors or decreased memory

140 John B. Calhoun

storage (i.e., increase in A'); (b) increase in discriminatory power in the
sense of screening out portions of those stimuli of the d of others requisite
for ehciting responses (i.e., decreasing d'); or (c) impeded synaptic trans-
mission (i.e., decreasing v').

5. Behavioral Origin of Response-Evoking Capacity, S

a. The target diameter genotype as determined by varibility of behavioral
traits. I now wish to present the logic of why variabiHty of behavioral
traits becomes inevitable. In fact, as aminals become more social, varia-
bility in physical traits must become of less importance in determining
the kind and intensities of interaction. So let us start with the case where
all indi\'iduals possess identical heredity and therefore identical physical
characteristics. Even for so simple an organism as the house mouse, marked
differences in capacities for social involvement develop despite the fact
that the members of the group come from a stock made genetically homo-
zygous by nearly a hundred generations of brother-to-sister inbreeding
(Calhoun, 1956).

The initial formulation of social interaction dealt with a deterministic
model in which all individuals w^ere identical. It showed that half the time
an individual was in the responsive state it would meet another responsive
individual and half the time it would meet another in the refractory state.
Thus, even under ideal conditions, an individual would be frustrated as
frequently as it would be satisfied from social interaction. But satisfaction
will not hkely precisely alternate with frustration. Furthermore, if we
consider some arbitrary relatively short span of time when the group first
forms, determined by the average individual having, for example, a total
pool of interactions equivalent to 2-5 times the number of individuals in
the group, then something like the following will have transpired :

Each individual's behavior toward another may be characterized by its
form or pattern and by its timing with regard to whether the other in-
dividual involved in the interaction is also in the responsive state (the a
state) or whether it is in the opposite or nonresponsive state (the p state) .
Initially the form of the behavior of all individuals in the responsive state
will be identical. Identical form denotes possession of the entire assembly
of traits, d, by every individual. With each individual contacting its as-
sociates in a random sequence over time, it is inevitable that some, who
are in the responsive a state, will purely by chance more frequently en-
counter others who happen to be in the nonresponsive p state. Each such
encounter will throw the responsive individual into an a^^ frustrating type
refractory period. Thus, the appropriate behavior of this individual will
not only not be rewarded, it w^ill actually be punished. After this individual

1. The Social Use of Space 141

passes through its a^p refractory period and again enters the responsive
state, some random change may typify its behavior. The more freciuently
it is frustrated, the more hkely will its behavior become deviant simply
because there has been so infrequent reinforcement of its original appropri-
ate form. Conversely, those individuals which, when they are in the re-
sponsive state, have met another also in the responsive state, will have the
original appropriate behavior rewarded or reinforced. The behavior of
such individuals will remain much in its original form.

In this way the members of a group may be rank ordered according to
the degree to which their behavior has deviated from the original. Each
recognizable unit of deviation represents a recessive d-gene. Each unit of
retention of the original behavior pattern from which theirs has been a
deviation represents a dominant c?-gene. The more dominant c?-genes an
individual possesses, the more intensely will associates respond to him in a
positive affective manner, and the more likely will he be chosen as a
partner or leader. Conversely, the more recessive c?-genes an individual
possesses, the more intensely will associates impose restraints or sanctions
on him, and the more likely will they reject him.

It is useful to borrow terminology from genetics which deals with he-
redity. Let upper case letters represent dominant d-genes, and lowercase
letters represent recessive d-genes. Such a system for an A^ = 11 is shown
in Table XIII. Here (d) represents the common traits shared by all mem-
bers of N. (d) plus the remaining dominant and recessive genes specify
an individual's rf-genotype. That individual which has been exposed to the
least number of circumstances producing changes in its c?-genotype may
be said to possess the "ideal" d-genotype. All other individuals will diverge
more or less from this ideal. They may be rank ordered from the alpha in-
dividual with the ideal d-genotype to the omega nth ranked individual,
which differs most from the alpha one. Each can then be assigned a simi-
larity rank, R. The alpha individual is represented by Ri; the one who
differs least from the alpha has R2; while the individual who differs most
from the alpha has Rn. In Eqs. (92) and (93) i = R (Tables XIII to XV).

Consider R4. By utilizing Eq. (93) its c?-genotype, by which we simply
mean its d, becomes:

d = (d) + gi'^ + r/r- + r/;^ + g\,'' + g'/'

+ g'o'' + gk'' + g'l'' + g'^ (104)

Obviously the d for each similarity rank, R, will be different for every
other one. Therefore S for each individual will be unique. This response-
evoking capacity, S, represents the d-phenotype.

142

John B. Calhoun

Table XIII
Trait, d-GENE, Differentiation in a Similarity Rank Hierarchy

Common
Similarity N^ Nj traits

rank, R {d)

Differentiating traits, d-genes

11 1

2 10

id)

Dominant d-genes

ABCDEFGHI J

10 2 (d) a

B CDEFGHI J

9 3 (d) a b

CDEFGHI J

S 4 (rf) a b c

D E F G H I J

7 5 {(}) abed

E F G H I J

G 6 ((/) a I) c d e

F G H I J

id) a b c d e f G H I J

4 8 id') a b c d e f g

H I J

3 9 {d) a b c d e f g h

I J

(d) a b c d e f g h i I J

1 11 {d) a b c d e f g h i j

Recessive d-genes

1. The Social Use of Space 143

b. Response-evoking capacities. Prior formulations regarding the typing of
behavior and personality, exemplified by Schaefer (19r)9, 1961), suggest that
the nature of one's own S and the nature of the response evoked from others
is influenced by the mood or attitude of the other individual involved. This
mood or attitude determines how one views the target diameters of others.
It determines whether one focuses on the dominant rf-genes or the recessive
<:/-genes of others. That is, when an individual responds to another does he
look at the other's desirable or undesirable characteristics. It will be a
prime premise of my formulation that one can be afTected only by the good
or by the bad side of another at a particular moment in time, but not by
both simultaneously. Furthermore, I shall show that there are two ways of
assessing the good qualities, the dominant d-genes, of another. Likewise,
there are two ways of assessing the undesirable finalities, the recessive
rf-genes, of another. Any individual's (/-genotype in a social setting produces
four types of r/-phcnotypos, that is four kinds of S.

Let

*S(-^> represent the response-e\-oking capacity of an individual affecting
the probability of his being chosen by associates as an object of
affection. Sa is related to Schaefer's love-acceptance. Sa depends
upon both the common target diameter, (d) , and on dominant
d-genes.

gia) represent the response-evoking capacity of an individual affecting
the probability of his being rejected by associates. S'^"^ is related
to Schaefer's hostility-rejection. S'-"^ depends solely on recessive
rf-genes. It is dependent in no way upon the common traits, (d) .

S'^^^ represent the response-evoking capacity of an individual affecting
how intensely he will be loved or approved of by associates. Love
here implies intensity of positive response and thus connotes
increase in probability of an individual realizing the objective of
his behavior. *S'^'^ determines the extent to which an individual's
desires will be facilitated by associates. In this sense, S'-^^ is
related to Schaefer's autonomy. S'-^^ depends upon both the
common target channels, (d), and on dominant rf-genes.

*§(") represent the response-evoking capacity of an individual affecting
the degree to which its velocity is altered. S'-'^ is related to
Schaefer's control. *S^''^ is dependent solely upon recessive d-
genes. It influences the intensity of negative sanctions imposed by
associates.

Equations for ;S^^' and S'f'' have already been given; i.e., Eqs. (92)
and (93) subject to the restrictions there stated.

144 John B. Calhoun

Let:
Pi^^ = the probability of choosing any ith individual.
Pi"^ = the probability of rejecting any zth individual.

Then

Vr = -W^ (105)

Vr = -W^— (106)

i here, and in Eqs. (92) and (93) and (105) to (113), refers to specifica-
tion of individuals by similarity rank, R. See discussion before Eq. (104).

Equations (105) and (106) in essence state that the probability of any
other individual choosing or rejecting the ith. individual depends upon
what proportion of the total dominant d-gene pool, or recessive c?-gene
pool, of the entire N individuals is encompassed by this ith. individual.
Note that these equations include evaluation of one's entire experience
with members of the group, including awareness of one's own traits. This
topic of self-awareness will be discussed later.

Conceptually, it is somewhat more difficult to understand S^^^ and S^''\
although the equations for their calculation are rather simple. Let us con-
sider iS^"^ first, since earlier reference simply to S was usually in the re-
stricted sense of ^S'^''^

Consider the individual in Table XIII with similarity rank 6, Re. When
individuals Ri to R^ are in that state where they tend to impose restraints
or sanctions on others, they will view Re as being more different from the
ideal type than they themselves are. In this sense, Ri to ^5 are type 1
individuals, in the sense of Eq. (94). Similarly, R^ to Rn will perceive ^7
as being like themselves in that they all share the recessive rf-gene, g[^K
Thus, they along with Re may be considered as type 2 individuals, in the
sense of Eq. (95), with reference to calculating the S^"^ of Re by Eq. (97).
Nj for Re is 6. By a similar logic the S'^'"'> of each individual may be calcu-
lated. See Table XV for S'^"^ calculated for every member of an A^ = 11
as depicted in Table XIII.

Each individual will belong to a different-sized A'' of type 2 indi^nduals.
This N will hereafter be referred to as Nj to differentiate it from the A^2
given in Eq. (97) . By analogy to Eq. (97) :

^(.-) = N/Ni (107)

1. The Social Use of Space 145

And from Eq. (91) it follows that

Vi = N-JN (108)

Equation (108) has proved a most useful one in the study of social
groups of experimental animals because it leads to predicting the degrees
of social withdrawal expected among any group of known size. In the dis-
cussion following Eq. (99) I pointed out that where /x increases as a result
of S^"^ becoming greater than 1.0, acconmiodation might be through
ejection of those members with the largest ^S^"^ or by a splitting of the
group. Each of these possibilities presumes unused area A into which the
appropriate individuals may immigrate. However, when surrounding
groups maintain territories, or other circmnstances preclude emigration,
then the A component of m = {Sv/A) remains constant. Thus, reduction
in velocity, v, becomes the only avenue for reducing /x back to the 1.0
value appropriate to No.

By a similar line of reasoning to that leading to Eci. (107), A''j represents
the number of individuals with which the individual in question possesses
a given uniqueness of dominant c?-genes.

Reference to Table XIII will clarify the meaning of Nj. For example,
Rh belongs to an iVj = 5 since it may be recognized by sharing the domi-
nant c?-gene, E, wdth four other individuals. A^j and the similarity rank, R,
will always have identical numerical values.

By analogy to Eq. (107)

S^P = N/Nj (109)

And although I do not for the present see how one identifies T' in biological
or social terms, although it may represent the seeking for positive affec-
tion, it is obvious that

Vi = Nj/N (110)

In this sense behavioral c?-genes do not represent retention or deviation
from specific behaviors. Characterization by two individuals of possessing
at least three degrees of deviation does not mean that these degrees of
deviation are identical.

Now consider a group consisting of four individuals, the pertinent data
and calculations for which are given in Table XIV. The probabilities of
the dominant and recessive c?-genes are:

Pa = 1/4 Pa = 3/4

Pb = 2/4 p, - 2/4

PC = 3/4 p. = 1/4

Pid) = 4/4

146

John B. Calhoun

f^ .

fa z

° S

CU ID

^ J

O C

« o

^ w

-< >

z 9

H-c n

z ^

O tc

> K

a H

'^ z

^ H

►J s

>■

-< <;

t— t

PS CL,

(-1 K,

> 5
-!;

K to

a «

'-^ a

C Ch

£-1 ei-

2 M

H Z

< O

z -

fi K

H t

■»

u .

"OQ*

S:^ <

1. The Social Use of Space

147

The pool of shared behaviors, (d), acts as a single dominant d-gene.
Where R represents the rank of the animal in terms of the degree of reten-
tion of dominant d-genes, that is how closely it expresses the ideal proto-
type manner of behaving, S'-'^^ and S'-"^ may be calculated by Eqs. (92)
and (93) utilizing the above probabilities of dominant and recessive
d-genes.

(4/1 + 4/2 + 4/3 + 4/4) = 8.33
(4/2 + 4/3 + 4/4) = 4.33

(4/3 + 4/4) = 2.33

(4/4) = 1.00

Then

^R=4

16.00 = 4A^

Restatement of the concept of behavioral c?-genes in the context of a
hypothetical example of their relationship to the origin of the various
stimulus-evoking capacities and of velocity, v, will assist in clarifying the
relationships. As described in Section XIII, B, 5, a the members of an es-
tablished group may be ranked in terms of the degree to which deviation
has developed from the ideal prototype. This ideal prototype will always
be that individual characterized by the least absolute deviation. There will
be A^ — 1 degrees of deviation in a group of N individuals.

Let:

Ai,
Then:

S2, •••, Sn-1 represent the characterization of an individual by 1,

2, ■ ■ • , N — 1 degrees of deviation
A2, • • • , A.v_i represent the absence of de\'iations 5i, 82, • • • , 5.v_i

5i = recessive d-gene a

82 — di = recessive d-gene b

8;i — 82 = recessive rf-gene c
etc.

And

Then

Ai = dominant d-gene A

A2 — Ai = dominant d-gene B

A3 — A2 = dominant c?-gene C

etc.

^R=l

= 0

C(a) _

(4/3) = 1.33

C(n) _
^R=i —

(4/3 -h 4/2) = 2.33

S'r% =

(4/3 + 4/2 -K 4/1) = 6.33

j: .s,^«)

10.00 = 3A^ -

ff=i

(A^/2)

148 John B. Calhoun

N.1 and A^j in the sense of type one and type two individuals discussed
in relation to Eqs. (94) and (95) and (107) are given in Table XIV. When
A^j = 2 it means that the second ranked, R = 2, individual is a member of
a subgroup of two individuals which share the dominant d-gene B. Like-
wise when N j = 2 it means that the third ranked, R = 3, individual be-
longs to a subgroup of two individuals each of whom possesses recessive
d-gene b. Every individual belongs to a unique A^j and N,. Given these
uniquenesses of A^'j and A^j as shown in Table XIV, the respective S^^\
Sl"^ and Vi were calculated respectively by Eqs. (109), (107), and (108)
and presented in Table XIV.

Up to the present I have intentionally maintained the discussion of
response-evoking capacity, S, on a nearly strictly theoretical plane for
the purpose of clarifying concepts. Even though such theoretical formula-
tions may be justified in their own right, regardless of how well they ap-
proximate reality, still it is desirable to ascertain whether they are in
harmony with observed data. To this end I selected two sets of empirically
obtained data. One involves measurements for mice from which velocity,
V, may be derived. The second involves the choosing of table partners
among groups of delinquent girls. These latter data permit determination
of how well the observed choosing can be predicted by Eq. (105).

XIV. Consequences and Examples of Social Interaction Systems

A, Velocity Reduction in a Hierarchy of Mice

I have previously described (Calhoun, 1956) the patterns of social rela-
tionships which develop among members of small groups of inbred domes-
ticated mice. Whenever two mice passed within a few inches of each other
they were recorded as having a contact, regardless of whether or not a
detectable social interaction could be detected. During a "contact" two
mice merely came within that range of each other for which contentions
for status did at times develop. We shall consider the example provided
by a group of eleven C57 black, inbred male mice which had developed a
stable hierarchy prior to recording the frequency and kind of associations
of each individual with his associates. Some pairs of mice contacted each
other much more frequently than anticipated on a chance basis, while
members of other pairs very infrequently met. These mice are rank ordered
in Table XV according to the number of contacts each had with associates.
For all practical purposes this order also represents the observed ability
to dominate an encounter which precipitated in a fight or flight. That is.

1. The Sdicial Use of Space 149

the most active mouse was the most dominant one and the least active
one the most subordinate.

It can readily be shown that with N's as large as eleven, the number of
contacts, ric, is so nearly proportional to velocity, v, as to justify utilizing
number of contacts as a measure of velocity in the present example. From
Eq. (91) it is seen that S^i'''> varies inversely with Vi, and therefore on a
theoretical basis mice with few contacts should be those wdth large response-
evoking capacities.

Table XV
Velocity in a Hierarchy of Mice

5 6

Rank
R

iVi

Number of contacts, ric, expected

Mouse's
number

Obs.

Exp. V

Eq. (113) Eq. (108)

Eq. (107)

315 1 11 123 128.17 1.000 1.00

311 2 10 116 116.63 0.910 1.10

282 3 9 109 104.84 0.818 1.22

286

93.18

0.727

1.38

319

81.52

0.636

1.57

303

69.85

0.545

1.83

321

58.19

0.454

2.20

317

46.65

0.364

2.75

290

34.99

0.273

3.67

288

23.33

0.182

5.50

301

20
769

11.66

0.091

11.00

2 = 6.001

= (N + l)/2

Where v represents relative velocity as calculated by Eq. (108) for any
A^, empirical calculation will show that:

Then where wj°^^) and 7ilf-p^ represent respectively, observed and ex-
pected number of contacts in any arbitrary period of time, the number of

150

John B. Calhoun

contacts, Hd, for any ith animal l)ecomes:

(exp) _

{N + 1)/2J t^

Zn^

bs)

= Vi

2 E n(f «)

iV + 1

(112)

And by substituting Eq. (108) into Eq. (112) we obtain an equation
more convenient for calculation:

„(exp) _
"'ci ~

2 E nT'

t=i

N + 1

(113)

nfl^\ so calculated, are given in Table XV. Where

i2=l

, (exp)

X" = 8.001, which with 10 degrees of freedom has a p of 0.629. On this
basis the observed certainly does not deviate significantly from the expected.

1. Awareness of Self

Three-fourths of the contribution to the above x" come from the single
omega, nth. ranked indi^'idual. Considering only the highest ten ranked
individuals, x" = 2.006 which with 9 degrees of freedom has a p of 0.99! I
have already shown (Calhoun, 1956) that the paired contacts in this
group diverged markedly from randomness, and so the divergence must
reflect some fixed social system such as elaborated here with regard to
reduction in velocity. Therefore, the marked divergence of this single
omega individual is likely to reflect a basic process, not just a random
variation; so I asked, "How would self-awareness afTect the present formu-
lation?" By self-awareness I mean that an individual recognizes and im-
poses self-sanctions which are of sufficient intensity to reduce his velocity
just as much as do the sanctions imposed upon him by his associates. For
this to happen it means that an individual can "meet" himself.

Equations (96), (97), (107), and (109) imply that an animal can meet
itself. That an individual meets himself means that he must recognize
himself. This raises the question of how an individual recognizes himself.
One way is by comparison. Considering degrees of difference depicted by
Table XIII, an individual can say, "I am at least as different as those

1. The Social Use of Space 151

which possess certain recessive c?-genes that I possess." By this method of
comparison with individuals who differ more than he does, an individual
can know the entire extent of his difference. This is true for all except the
7iih. ranked omega individual. As may plainly be seen from Table XIII,
7?ii has no basis for comparing his d-gene ^"^j"^ since no other individual
shares it. He can only be aware of his divergence including gi^^---g\^^
shared in entirety by Rw. Therefore, by Eq. (107) his

SM = 11/2 = 5.50

andby Eq. (108) his

V = 2/11 = 0.182
Therefore, by Eq. (113) his

n^;-p) = 2.3.33

Using this \'alue in Table XX the x" for the entire group of 11 mice
becomes 2.522, which with 10 degrees of freedom has p = 0.99. I realize
that this is only one case and I may justifiably be accused of making a
"conceptual mammoth out of a mouse." Yet if animals do recognize their
individuality, the only confirmation of such recognition can come through
observing that the omega individual in an Nh group exhibits twice the
velocity or twice the number of contacts he would otherwise be expected
to have.

It is realized that if the iVj of the iVth ranked animal, the omega in-
dividual, is equivalent to that of the (iV — l)th ranked animal, the n^®^^^
for the omega animal will actually be slightly less than 23.33. Likewise
the n^^^^^ for all higher ranked animals would be slightly less than stated
in Table XV, since

N N

R=l S=l

2. Sanctions and Facilitations

When the target diameters of all members of N are identical, then the
S, in the sense of S'^''\ is identical for all individuals and is equal to 1.0. As
target diameters vary in accordance with the system illustrated in Table
XIII, S increases except for the alpha first-ranked individual. Such in-
creases in S mean that the intensity (duration) of the response evoked
from others will be greater. Therefore, intensity of response, the i of the
prior formulation, will increase by a factor = aS*"'.

Let 2^'' represent the intensity of response evoked from every member,

152 John B. Calhoun

e, of A^ compatible with one's own »S'''\ Then it follows from Eq. (70) that
for the iih individual

,■(«) _ C'C'')

N - 1

1/2

(114)

{(V) represents the imposed restraint or sanction. Sanctions in this sense
lead to velocity reduction as exemplified by the mice, i^"^ = i'f^ with
reference to self action.

Where the target diameters of all individuals are identical, Eq. (70)
with an AT" of 11 gives an i = 0.316. However, where target diameters differ
as given in Table XIII, it may be seen from column 7 in Table XV that
S''"^ increases from 1.0 to 5.5. (Here, I am also assuming that the A^'th
ranked individual has the same S as the (N — l)th ranked individual for
reasons discussed in Section XIV, A, 1.) The mean *S^"^ will be 2.52. There
fore, the mean i^"^ becomes 2.52 X 0.316 = 0.816.

We are considering intensities of interaction involved during imposition
of sanctions in the sense of restraints. Therefore, when we consider any
individual in the a responsive state, the other individuals from which it
evokes an {'-'"'> sanction-type interaction must be considered to be in the p
nonresponsive state since the resulting interaction will not contribute to
this individual's satiation from social interaction. Every member, e, of A'',
which is in the responsive state, a, will exhibit an i^'f intensity of response
toward any ith individual, also in the a responsive state, where

iiV = S)''

N - \

(115)

i^^^ represents the intensity of interaction of the individual searching
for satisfaction. Similarly the mean z^^^ will also be 0.816 instead of the
0.316 which exists when all individuals in an A^ of 11 have the same target
diameter.

= 0.666

This a^j, represents the mean frustration-type refractory period resulting
from the imposition of sanctions in an A^ = 11. Since when all individuals
have the same target diameter, z^ = (0.3 1 6) ^ = 0.1, hierarchy formation
increases the average amount of stress more than sixfold. No wonder that
the i^gi elicited from others became so eflFective in reducing velocity so
that each individual's m again equals 1.0. That is, r^S" = 1.0. Although

1. The Social Use of Space 153

intermittent application of sanction, t^''^ may be required to reinforce the
V behavior, sanctions still must be primarily a phenomenon of hierarchy
formation or role assumption.

When two individuals, x and y, are both in the a responsive state, their
resulting interaction, ii^H^^^ will lead to a satisfying aaa refractory period.
Being in the responsive state implies a need for satisfaction from social
interaction. Furthermore, during such interactions each must be focusing
its attention on the S^'^^ aspects of the other's target diameter. In like
fashion, when individual x in the responsive state approaches 2j in the non-
responsive state, X exhibits behavior appropriate to ys *S^^\ while y
retaliates with a response appropriate to x's *S^"\

One individual which is in the responsive phase will have its behavior
toward another judged by the latter as appropriate or inappropriate, de-
pending upon whether or not the individual approached is also in the re-
sponsive state. Appropriate social behavior becomes synonymous with
what I have termed "dominant (/-genes." These behaviors must be ap-
propriate both in form and in timing, such that the approached individual
will be in a similar need state for obtaining satisfaction from social inter-
action. Even though a social behavior may seem appropriate with regard
to form, if the timing of its expression is not harmonious with the refractory
nonresponse state of the individual approached, then the approached in-
dividual will judge this behavior as being inappropriate. Inappropriate-
ness in this sense becomes what I term a "recessive c?-gene."

In a perfectly random system, with all individuals exhibiting identical
behavior with regard to its form, some individuals will by chance more
frequently encounter others who are in a refractory nonresponsive state.
Thus, such individuals will be responded to by their associates in identical
fashion as they would have been if the form of their behavior were actually
inappropriate. To the degree that such structurally appropriate behaviors
fail to be reinforced by similar behavior from associates, they may be ex-
pected to vary in some random fashion until their form becomes relatively
distinct from the original. To the extent that these alterations in behavior
become established, they represent clearly recognizable recessive d-genes.
Due to this modification of behavior S^''^ will increase and S^^'^ decrease,
so even when the approached individual is also in the responsive state the
approaching individual will experience less satisfaction from the inter-
action than it would have had these random processes not transpired.

Once M becomes restabilized to 1.0 following the reductions in velocity
accompanying hierarchy formation, it will then be possible for two in-
dividuals, each in the socially receptive a state, to encounter each other
and mutually contribute to each other's need state by an i^^H^^^ - ««»
interaction. In this case, each individual's own i^^^ represents its searching

154 John B. Calhoun

for social satiation, while the other's i'^^ represents a social facilitation, the
opposite of a social sanction.

I shall leave the concept of social facilitation at this theoretical level
without seeking empirical confirmation. However, excellent data have been
presented by Aloreno (195o; Moreno and Jennings, 1960) which permit
an exploration of the probable validity of *Sj-'^', Eq. (92).

B. The Choosing of a Partner

We have already seen that >S.^"), Eq. (107), and Si^\ Eq. (109), relate
to those response-evoking capacities of an individual's target diameter
which determine the intensity with which associates will respectively im-
pose sanctions or facilitations. S\-'^\ Eq. (92), and S^^^ Eq. (93), like-
wise represent aspects of one's target diameter influencing the response of
others. One's own S'-'^^ determines the probability of being chosen by others
as an appropriate object for social response, while one's own S'^"^ similarly
determines the probability of being rejected. No doubt there are excellent
empirical data for testing the validity of S'-'^K However, I shall confine
myself to S'-'^''. If I can show the likelihood of aS^^^ being an approximation
of reality, it follows that «S^°^ can be similarly justified as a concept.

Moreno (1953, 1960) presents a set of data for which there has been no
adequate formulation of their origin. In seven cottages each containing
exactly 26 delinquent girls, he asked each girl to choose three others in
their own cottage whom they would most like to sit close to at the dining
table. This instruction presents marked complications in determining
whether Eq. (105) wall account for the observed results. However, Eq.
(105) includes the possibility that one wdll choose oneself as a partner;
that is, one will choose to eat alone. Moreno by his instructions excluded
this possibility. Further, Moreno's instructions precluded the possibility
of choosing the same person two or three times, which Eq. (105) permits
on successive independent choices. Dr. Clifford Patlak worked out for me
the full set of equations required to determine how many times each in-
dividual would be chosen, considering Moreno's restrictions, after the
probability, p, of being chosen was calculated by Eq. (105) for each iih
individual in an N = 26. Moreno's restrictions so complicated the calcula-
tions that it was concluded that a simple lottery would adequately test
the applicability of the present theory, and at the same time avoid the
time-consuming job of developing a computer program to the same end.
This was done as follows:

1. Sl''^^ was calculated by Eq. (92) for each of the 26 members of N,
from Si'^'> for the alpha-ranked individual to Sii^ for the omega-
ranked individual.

1. The Social Use of Space 1^5

2. Then the p of being chosen was calculated for each individual by
Eq. (105). For example:

p(-4)

= 0.152,

Pi'^

= 0.121

p1^^

= 0.092,

p[V

= 0.039

pir

= 0.012

3. We then prepared 1000 -pi tokens for each individual.

4. All these 1014 tokens were placed in a large glass jar. For any "in-
dividual" to make three choices, three squares at random were
picked out. If an "individual" picked himself, the token was returned
to the jar and another "choice" made at random. Similarly, if another
individual was chosen twice, one of the two tokens was returned to
the jar and another selection made until each "individual" had
chosen three different "others." Between drawings all tokens were
returned to the jar and the contents mixed. Each such test consisted
of 78 "choices." At the end of each test it was possible to tally how
many of the 26 "individuals" had never been chosen, chosen once,
chosen twice, etc.

5. Fourteen such independent tests were performed and the mean number
of individuals in each choice category was calculated. This mean
represents an estimate of the expected.

These data in Table XVI are plotted in Fig. 36. A smoothed curve ap-
proximating the means of the observed and the lottery presumably ap-
proximates that conforming to Eq. (105), considering the restrictions im-
posed by Moreno. This smoothed curve is called "theoretical" in Fig. 36.
Each of the two sets of points varies so closely about this curve as to pro-
vide confirmation that my formulation is adequate to account for the ob-
served sociometric phenomenon of choice.

IMoreno and his associates make much of the chains, triangles, etc. of
reciprocal or nonreciprocal choices that became apparent in such a socio-
metric system. While such patterns may become fixed realities, there is no
reason to attribute any condition other than chance to their origin.

C. The Response-Evoking Capacity Circuraplex

Schaefer (1959, 1961) presents a conceptual model capable of describing
the attitudes, personality or behavior of an individual in a two-dimensional
space. This latter is determined by two orthogonal axes of polar opposites.
As may be seen in Fig. 37, one axis consists of the polar opposites love and
hostility, while the other is represented by control and autonomy. About

156

John B. Calhoun

Table XVI
Empirical and Theoretical Data Regarding the Choosing of a Partner

Mean number of individuals

"Theoretical"
mean of

Times chosen

Moreno (1953,

Present Lottery

A and B

1960)

based on S^"^'

5.00

5.36

5.18

4.15

5.36

4.76

4.29

3.82

4.06

3.72

2.29

3.01

2.29

2.50

2.40

2.00

1.79

1.90

1.14

1.50

1.32

1.43

0.86

1.15

0.86

0.72

0.79

0.57

0.14

0.50

0.32

0.43

0.36

0.40

0.00

0.29

0.15

0.00

0.07

0.04

1 1

1 { 1 1
"Theoretical "

1 1

^.^

•
▲

Moreno 1953
Lottery

—

^ 4

•

•
\

^ 3

u.
o

LU
OD

i 2

•

^^^^^ •

—

1 1

1 Vl^-i-

0 5 10

T TIMES CHOSEN
Fig. 36. The choosing of a partner. Moreno's empirical data are compared with a
lottery conforming both to Moreno's procedure and the probabilities of being chosen as
predicted by Eq. (105) in the present paper.

1. The Social Use of Space

157

or within the circle determined by the extremes of these polar opposites,
the typical characteristic of an individual may be oriented.

There exist remarkable similarities between Schaefer's model and that
implied so far by my concept of response-evoking capacity, S, and the
resultant or concomitant change in the behavior of the individual, such as
reflected by change in velocity.

*§<'■', S^^\ S^"^', and S^"'' represent factors not specifically treated by

TM. QUADRANT

Choleric "
Ouorrelsomness
Irritability
Impulsive
Aggressiveness
Delinquent

I SJ- QUADRANT

Songume "

Intellectual efficiency
Social participation
Friendliness
Leodership

(HOSTILITY )
REJECTION (ii)s'°>

(LOVE)
ACCEPTANCE

Ego strengtit
Sympathetic
Trusting

3 RD. QUADRANT

' Melancholic "
Schizophrenio
Psychosthenia
Social withdrawal

CONTROL

Introversion
Repressed anxiety
Social opprehensiveness
Neuroticism

2 ND. QUADRANT

Phlegmatic"
Obedient
Conscientious

Fig. 37. The circumplex depiction of behavior and personality superimposed upon
the coordinates of response-evoking capacity, S.

Schaefer. They relate to his formulation as follows: They represent those
characteristics of the individual that determine the probability of accept-
ance or rejection as a social object and that determine the kind and in-
tensity of response elicited. Schaefer considers the second logical step, the
evoked response, such as the mother's response toward her children. He
also considers the third-order phenomenon, that of the personality de-
veloped as a result of being the target of such responses.

I shall now attempt to place these throe orders of phenomena in perspec-
tive. For an A^ = 11, S^'-'\ S^^\ S^^\ and S^"^ were calculated by the

0.091

1.000

0.000

0.166

0.244

0.220

1.000

0.000

0.030

1.000

158 John B. Calhoun

above equations. These values were then recalculated as proportion of the
maximum value. For examples, see tabulation.

Similarity rank S'"' S^^ S^^'> S(«)

/S*"^ and iS^^^ with values of 1.0 form polar opposites, as likewise do
^S^-^^ and aS^"\ These polar opposites are shown as a two-dimensional co-
ordinate system in Fig. 37. The four "response-evoking capacity" coordi-
nate points for each individual are connected by lines. Thus, a square may
be delineated for each ranked individual which represents its "life-space"
with regard to eliciting responses from associates.

Schaefer's two-dimensional circumplex description of behavior and
personality replaces the relative intensity values of my axis of correlation
coefficients. In most comprehensive sets of measures of behavior or per-
sonality, he regularly found that two measures which can be equated with
the terms love (acceptance) and hostility (rejection) are highly negatively
correlated and so form polar opposites which may thus be plotted as polar
coordinates at 180 degrees from each other. Similarly, concepts identified
by the terms "control" and "autonomy" form polar coordinates opposite
each other. Control and autonomy have zero correlation with acceptance
and rejection and so the control-autonomy axis lies at right angles to the
acceptance-rejection axis. From the center zero point each of the four axes
extend outward to represent a maximum correlation coefficient of 1.0 at
its extremity. Every other concept in the set is then correlated with each
of the four "key" concepts. It is regularly observed that every other con-
cept in the set has positive correlations with two of the neighboring polar
concepts, and so each may be plotted in this two-dimensional behavior-
personality field.

Such points approximately fall on the circumference of a circle inter-
secting the polar coordinates. Schafer calls such a set of points a circumplex.
His schematic representation of types of maternal behavior is shown by
dots in Fig. 37. Terms connected by arrows to these points denote various
kinds of maternal behavior. These maternal behaviors are the types I
would anticipate as being directed toward, and derived from, the response-
evoking-capacity circumplex determined by *S^^\ ^S^^^ S^^^, and >S^"\

Schaefer also reviews many studies by others which harmonize with this
conceptualization of behavior and personality. I have included selected

1. The Social Use of Space 159

terms in boxes which carry the "flavor" of what each pole or each t^uadrant
represents.

It will be noted that the response-evoking-capacity life-spaces of the
members of any group, as I have described their origin, mainly fall in the
first and third quadrant of this two-dimensional space. I shall, therefore,
call these two quadrants the "primary life-space." We may inquire how
individuals become identified with the other two quadrants, the second and
the fourth, which may be called the secondary life-space.

INIy colleague. Dr. Kyle Barbehenn provided the solution. In examining
Table XIII he noted its bilateral symmetry with reference to dominant
and recessive c?-genes. In any group recessive d-genes merely represent
degrees of divergence from some ideal prototype. The alpha-ranked in-
dividual possesses only dominant prototype characteristics, whereas the
omega-ranked individual, except for commonly held (d) traits, possesses
only divergent and therefore recessive d-genes. But suppose in the history
of the group some circumstance led the omega's associates to consider his
characteristics as being more desirable than the alpha's. Such a change in
attitude would reverse the roles of all individuals, except for the median-
ranked individual. The individual who formerly was socially withdrawn
and had a low velocity would immediately become a high velocity in-
dividual, participating in many social interactions. Likewise, the former
alpha individual would become the omega one, losing his "leadership''
role and becoming socially withdrawn.

This transformation of recessive rf-genes into dominant ones, and vice
versa, will still place most of the individuals again within the first and third
quadrants of the circumplex life-space. The critical question involves the
transition period. In this transition Eqs. (92) and (93) change roles with
reference to their application to dominant and recessive c?-genes, as like-
wise do Eqs. (107) and (109). There arises the hkelihood that previous to
this transition choosing and rejecting will be a more predominant activity
than controlling or granting autonomy. If so, S'-^'' will switch to pertain
to recessive rf-genes and /S^"^ to dominant d-genes while *S^^^ and S^"''
will retain their orientation toward dominant and recessive c?-genes, re-
spectively. Or the reverse could happen if the group were primarily oriented
toward the control-autonomy axis. The consequence of these changes are
summarized in Table XVII.

1. Shifts into the Second and Fourth Life-Space Quadrants
BY Rats

No intent is here implied of proving the validity of the general formula-
tions. All I have hoped for is to evolve a logical and reasonable formula-
tion that may later pro^'e to be a fair first approximation of processes that

160

John B. Calhoun

do in fact exist. This elaboration of Schaefer's circumplex life-space im-
mediately brought recall of a striking transitory change in behavior which
regularly occurs in structured, dense, socially closed systems of domesti-
cated Norway rat populations.

When these populations have been permitted to attain a density of 60
to 100 adults in a space ideal for 40 or fewer adults, an extremely rigid
social structure develops. All rats "know" their place and aggressive ac-
tions terminating in some individuals receiving even moderate-sized
wounds become markedly reduced. Threat and avoidance becomes the
predominant pattern where aggression and imposition of sanctions di-

Table XVII

Change of "Attitude" toward d-GENEs Affecting Shifts into the Secondary
Life-Space of the 2nd and 4th Quadrants

Initial change

cZ-Gene Interaction of

involved neighboring poles Quadrant

5'"*^ oriented to recessive Recessive Control-acceptance 2nd
<i-genes

S'^"'! oriented to dominant Dominant Autonomy-rejection 4th
d-genes

S^^^ oriented to recessive Recessive Autonomy-rejection 4th
d-genes

(S'"' oriented to dominant Dominant Control-acceptance 2nd
d genes

rected toward maintaining status roles is involved. And yet in six of seven
such populations studied, one or several males went temporarily berserk.
Each such male abruptly began attacking all other members of the popu-
lation except those that behaved as they did. They inflicted deep slashing
gashes on the bodies and tails of associates of all ages and both sexes until
fresh blood could persistently be observed splattered about the habitat.

For any particular rat such episodes persisted from one to several days
and since usually several males were involved the total period of such dis-
turbance might last up to 6 weeks. In every instance these bursts of males
going berserk followed a period of at least three months during which the
investigator had removed all young prior to weaning. Then a new genera-
tion of young were permitted to survive. The period of males going berserk

1. The Social Use of Space 161

coincided with the initial post-weaning integration of these young rats into
the society when they ranged between 4") and 90 days of age.

All the males going berserk belonged to the high velocity, generally
dominant, segment of the society, and so belonged in the "sanguine"
most desirable first quadrant of the life-space circumplex. This meant that
dominant d-genes predominated in their target diameter. The juvenile
rats emerging into the society, being less dilTerentiated, were therefore now
prototypic and therefore resembled dominant adult males in their posses-
sion of mostly dominant d-genes. And yet these juveniles would still possess
traits lost during maturation by most adults.

Due to the typical response of rejecting strange objects, the young would
be rejected while still being permitted autonomy of action due to domi-
nant c?-genes shared with high-ranking adults. There being more juveniles
in the population than adults belonging clearly in the first quadrant of the
life-space circumplex, the general response of rats to these adults would
be to react to them as they did to juveniles by rejection. This rejection
triggered the release of the muted aggressive capacities of dominant males
to the extent that it was expressed with great intensity even toward others,
such as juveniles and adult females, who normally were not bitten.

Such an origin of an aberrant behavior in a rat society is patently an
interpretation lacking the complete documentation to carry the conviction
of its reality. Yet my intensive studies of rat societies permit identification
of so many behavior-personality types as to suggest that rats are equally
as complex as humans in this regard. The eight societies previously studied
have been commented upon in general terms elsewhere (Calhoun, 1962a, b) .

In the early history of a rat society, while its numbers and density are
low, most individuals seem rather clearly to fall into quadrants one and
three of the circumplex. This is a typical expectation when a straight-line
hierarchy develops, as it always does in initial stages of social organization.
Later on, histories and situations become more complex. Other types
develop which may clearly be assigned to the second and fourth quadrant.

For example, there is the type I call a "prober," which appears to repre-
sent a rat having shifted from the third to the fourth circumplex quadrant.
Earlier in their history they clearly belong within the lower echelons of the
social hierarchy. Later on they are generally ignored by dominants with
whom they live most closely. In consequence, they develop a marked per-
sistent state of hyperactivity indicating autonomy of action. They seem
to generalize this autonomy of action as permitting them freedom of ac-
tion anywhere. Consequently, they persist in invading the domains of
territorial males whenever members of their harems are in estrus. During
such invasions they rarely contest the status of the territorial male, but
in the process of being rejected by him receive wounds. These are received

162 John B. Calhoun

so frequently that their entire posterior becomes a mass of scar tissue
de\'oid of hair. Like other males who still receive sufficient sanctions from
associates to maintain their velocity at a low level, they become pansexual
in the sense of including adult males and juveniles of both sexes as objects
for sexual advances. They also share with the berserk males of (juadrant
four the property of heightened intensity of interaction. Theirs, however,
reveals itself in sexual behavior toward adult females. Mounts, instead of
lasting the usual 1 to 3 seconds, may continue for several minutes. This
persistence of the mounting, without intromission, resembles that of frogs.

Low velocity male rats, those which belong in circumplex quadrants
two and three, fall mostly into two distinct categories, those which have
received many wounds and those which have received very few. The former,
which belong in quadrant thi-ee, present no conceptual problem as to their
origin. Straightforward operation of the four S factors will always place
the lower-ranked members of the hierarchy in the third quadrant. How-
ever, the latter, "phlegmatic" types must have arisen from a secondary
180 degree shift in response evoked by the d-genes. Their lack of wounds
during their entire history indicates S^'^'> involves dominant d-genes. Their
low velocity indicates that now as adults S^^'> also involves dominant d-
genes. They are rats for whom we may infer that their associates have
always been "overindulgent" and "overprotective." Such rats are quite
fat and have relatively small adrenals, ventricles, and kidneys. Though
they exhibit some displaced sexual behavior, their "personality type" is
one most characteristically involving failure either to elicit or initiate social
interaction. They are types lacking social involvement.

This brief discourse on rat types suggests that my elaboration of
Schaefer's (1959, 1961) circumplex behavior-personality complex may
prove to be a fruitful framework for pursuing studies in comparative social
psychology.

D. Conformity, Withdrawal, and Creativity

When N increases above Nb or n increases above jU6 it has been shown
that fragmentation of N to appropriately sized discrete subgroups may lead
to reacquisition of 6^°^ and 6j°\ Even when A^ remains at Nb, the unavoid-
able variability in the four types of S leads to yu differing among members
of the group. Reduction of v as S'-''^ increases enables the individuals to
prevent excessive increments of 6/ above d'/\ Even so, many individuals
with reduced v must experience either reduced da or excessive Of. We may
inquire as to possible avenues for escaping these deviations from dj^"'' and
6^/'> without leaving the group.

1. The Social Use of Space 163

For those individuals who in their behavior do not differ markedly from
the alpha-ranked, Ri, member, there exists the possibility of adopting the
outward behavior of Ri. Such acquired conformity should permit eleva-
tion of velocity to basic levels and should produce a discontinuity in the
range of velocities observed among members of the group. Whyte (1956) in
his "The Organization Man" has emphasized the role of conformity as an
adjustive mechanism accompanying increases in A''.

At the opposite extreme of original behavioral divergence there exist
individuals with such reduced velocity that their contacts with higher-
ranked individuals proves insufficient to permit their developing conformity
through emulating the behavior of their superiors. Recently I have had the
opportunity of studying the behavior of all members of three populations
of domesticated Norway rats for every member of which assessments of
velocity had been made. Low-velocity rats develop the capacity to move
about without engaging in interactions with their associates. They rarely
initiate interactions nor do they elicit actions from associates. This social
withdrawal becomes so complete that, despite being in the presence of many
associates, they are characterized by small adrenals and small ventricles,
just as are rats which have lived all their lives as members of A'''s of 3.
David Riesman et al. (1953) has movingly described such isolation and
oblivion of surroundings among humans in his "The Lonely Crowd."

In between the velocity levels producing conformity and withdrawal,
there lies a narrow but important range of velocity permitting a process
which can terminate in creativity. Why this is so requires recall that ac-
cording to my model or social interaction, interaction with self must be
included along with interaction of self with others. One can choose oneself,
one can reject oneself, one can facilitate one's own behavior, or one can
restrain or impose sanctions on one's own behavior. All are possible.

Consider sanctions. These are mostly the i['i' , Eq. (114), of others
directed against oneself. Let i^"'^ represent self-sanctions. The self-inter-
action becomes:

N - 1

l/2\2

= ^ ' ^ (116)

N - 1

The lil"'^y, which initially solely represents self-control but may evolve
into creativity, is proportional to [ij^"'^]-. As may be seen from column 7
in Table XV, the intensity of this self-control mounts ever more rapidly
as lower ranks in the hierarchy are approached. In so long as self-interac-

164 John B. Calhoun

tion only represents self-control, there can be no creativity. However, in
this system what were recessi^•e c?-genes can become considered as dominant
d-genes. When this happens for the whole group, a complete reversal of
the social rank ordering develops. But such reversal can also take place
within a single individual. In essence, this means that [ij-"'^]", which is
equivalent to an a'^p frustrating experience, became transformed into an
[^i^'^]"; which is equivalent to an aaa satisfying experience, without losing
any intensity in the process.

There need not be any outward manifestation of this transformed self-
control. \_i\^'^^ only implies cortical associations among stored traces of
external events lacking any aura of negativism. Other than this, there are
no limits to the kinds of traces which may by chance attain a high condi-
tional probability of association. External manifestation of p,-^'^]", that
is creativity, implies sufficient contact or awareness of external events to
permit symbolic or behavioristic alteration of the enivornment in harmony
with these heightened conditional probabilities of association, [z'i^'^]-
must not be confused with learning. I shall not attempt to go into the ques-
tion of learning here other than to say that it should be most effective in
high velocity individuals.

It is well recognized that a feeling of ecstasy, of extreme well being, ac-
companies any instance of creativity. A unit of ^^°^ arising from a single
interaction represents such ecstasy. Recall that di°'> = 0.25. Examination
of Table XV reveals that not until R^. is reached in the descent through such
a hierarchy of A^ = 11 individuals will li^-'^J = IS^/^J/iN- 1) exceed
0.025. It will exceed 0.25 for Rs • • - Rn in an A^b = 11. And yet I have
already indicated that some of these, probably at least the last two, will
be so withdrawn from reality as to preclude any opportunity of creativity.
At most, we can therefore anticipate only 2/11 or 18% of the group to
possess potentialities of creativity. As N increases, proportionately more
individuals will accommodate by social withdrawal and relatively fewer
will possess potentialities of creativity.

Realization of these potentialities requires another set of conditions,
opportunity for having made many satisfactory accommodations to new
configurations of stimuli. As discussed in the later Sections XIV, G, 1 to
3, this means that there must have transpired an increase in one's psycho-
logical area. A", as a compensation for the reduced A associated with a
lowered v.

E. Velocity and Home Range

In my search for adequate formulations of the social use of space and
time I have been guided by several competent mathematicians. With

1. The Social Use of Space 165

regard to home range, Mr. James U. Casby (see Calhoun and Casby,
1958, pp. 16-17) derived a function, K/'Iira^-. He called this term "visita-
tion frequency." It describes the relative frequency of visiting a particular
place in the environment. Later, though published earlier (Calhoun, 1957),
Dr. Murray Eden derived the function, ^l = dv/A to represent a communi-
cation function defining the relative probability of one individual meeting
another. K/2Ta~ concerns arrival at a stationary point, while dv/A con-
cerns arrival at a moving point. It was only after I began this elaboration
of concepts concerning social use of space that I realized the isomorphism
of these two functions. A and 2ira^ become equivalent expressions as like-
wise do K and dv. d here is used in the sense of S'~''\ Eq. (107).

I have already pointed out in the discussion pertaining to Eqs. (35),
(80), and (88) to (91) that when a group is in a steady state in harmony
with its heredity and environment, m = 1.0, provided appropriate units of
time are considered. It followed that v, S^'^ and A also have relative
values of 1.0 at Nb, the harmonious steady state TV. Therefore

t;*S(^V27rc72 = 1.0 (117)

Given this relationship, can it assist us in determining relative home range
0-? We have already seen in Sections VI-VIII that contractions and ex-
pansions result from both intraspecific and interspecific interactions. Let
us consider home range of members of a constellation as discussed in Sec-
tion VIII, C.

On the average the constellation consists of one alpha, 6 beta, and 5
gamma individuals, ranked in this order:

Nb = N. + N0^ Ny,

where N^ = 1, N^ = Q, Ny = o. The a individual represents the ideal
prototype. Each beta diverges an equivalent amount from the alpha. Like-
wise, each gamma also diverges from the alpha, but more so than the beta.
By analogy to the discussion pertaining to Table XIII, the rf-genotypes of
the 12 individuals in a constellation will be represented as shown in Table
XVIII. as in the table denotes alteration in the home range a resulting
from the individual's S'-''\

The Ni for the alphas, betas, and gammas, according to the formula-
tion of Section XIII, B, 5 become respectively 12, 11, and 5. From Eq.
(107) it follows that their S^''^ are, respectively, 1.0, 1.091, and 2.40.
This increase in response-evoking capacity, S^''\ among beta and gamma
individuals will lead to their reduction of velocity, v, respectively to 0.9167
and 0.4167, as given by Eq. (108), when compared with v = 1.0 for the
alpha member.

166 John B. Calhoun

In terms of home range the aS'^'^ function in Eq. (117) may be ignored,
which means that:

l^/27rc72 = 1.0 (118)

For this relationship to maintain in the face of reduced v for beta and
gamma individuals, a must correspondingly decrease, such that

a = (y/27r)i/2 (119)

Since 2t is a constant we may ignore it for the purpose of determining

Table XVIII
Relative Home Ranges of Constellation Members

Constellation

member d-Genotype N-, v (Eq. (lOS)) <ts (Eq. (120))

a (d) A B 12 1.0 1.0

/3, (d) a B 11 0.9167 0.9575

/32

id)

0.9167

0.9575

id)

0.9167

0.9575

/34

(d)

0.9167

0.9575

id)

0.9167

0.9575

id)

0.9167

0.9575

id)

0.4167

0.6457

id)

0.4167

0.6457

id)

0.4167

0.6457

(d)

0.4167

0.6457

id)

0.4167

0.6457

home range a relative to that of the alpha individual. Then

as = (vy (120)

In essence, this line of reasoning predicts that as velocity is reduced the
area encompassed by the home range will develop similar reductions. Rela-
tive home range sigmas, derived by Eqs. (107) and (120), as given in
Table XVIII formed the basis for the relative sizes of home ranges within
a constellation (Fig. 27).

In considering the home range of the individual in the context of mem-
bership within a constellation of Nb individuals, it is obvious that the term
"area" possesses dual meaning. (2TrSaa^) approximately measures the .4

1. The Social Use of Space 167

of the constellation (Section VIII, C), while (27r(js^) represents that
portion of each individual's home range within which it spends 0.394 of
its time. Each individual simultaneously "inhabits" two life spaces, his
own and that of the group of which it is a member.

For all members but the alpha member, tiualitative differences in S^'-'''
lead to reductions in v, which in turn bring about reductions in A such that
the reduced A is proportional to v. This accommodation will result in /x
again exceeding 1.0 and will thus expose the individual to stress from an
increased frequency of contacts. Further accommodation, again reducing
M to the relative value of 1.0, can come only through quantitative reduction
of the target diameter, d, since the S'-"^ qualitative aspect has become a
stable factor. It will be recalled that the concept of the constellation pre-
supposes a capacity for emitting and receiving signals sufficient for any
member to gain contact with any other member despite their spatial separa-
tion. Such signals amount to an increase in d at the moment of this emis-
sion. The easiest strategy for regaining a m = 1-0 will be for each individual
to reduce the frequency of signaling as much as it had reduced its velocity.
This line of reasoning culminates in the conclusion that v, a", and d (in the
sense of freciuency of signaling) will all be reduced to values inversely
proportional to S^"K With respect to the unchanged relative values of 1.0
for the alpha individual, these values for beta members become respec-
tively 0.9167, (0.9167)- and 0.9167, while for gamma members they be-
come respectively 0.4167 (0.4167)2, and 0.4167.

Consistent with my objective of developing formulations adequate for
acquiring data in concrete experimental studies, the above theoretical
conclusions may be tested for their application to groups more compact
than represented by constellations. Given an experimental N = 12 re-
stricted to an area A, one can estimate relative velocity be determining
the proportions of a series of time samples in which each individual is
active and exposed to situations where social interactions do occur. This
velocity we can designate by v. Velocity so estimated will reflect true
velocity more accurately than the cruder measure of number of contacts,
Tie, used in the specific case in the discussion pertaining to Eq. (113). This
is because an animal may be active when all others are at rest.

The more velocity is reduced, as predicted by Eq. (108), the more the
individual should restrict his travels to a smaller portion of the area utilized
by the group. Of an originally common home range shared by members
of a ''compact" group, low velocity members will come to utilize only a
portion. Thus, the ideal design of an area within which social organization
is being studied must provide for many subareas to which visitations may
be recorded. Otherwise, reduction in home range cannot be detected.

Likewise, the more an individual reduces his velocity, the more he should

168 John B. Calhoun

reduce the emission of signals eliciting social awareness of him by associ-
ates. Such signaling may involve other than vocal modes and may even
involve reduction of S'-"^ itself. In the end, such reductions of d and S'^"^ will
produce an individual which, when encountered, is judged by associates as
being dull, lifeless, and with flat affect, lacking the attributes of an appro-
priate object for social interaction.

F. Velocity in High-Density Rat Societies

While these formulations of social dynamics were being developed I
was simultaneously pursuing empirical studies of social dynamics revealed
by large groups of domesticated Norway rats. We shall examine those
data which indicate (a) that densities greater than appropriate for Nb
suppress velocity, and (b) that an increase in vitamin A above normal
levels buffers the social system against the velocity-suppressing force of
increased density.

During a 16-month period of 1960-1961 further studies in the habitat
(Fig. 33), discussed in Section XII, C, w^ere conducted. In this second series
of studies the only habitat change involved altering the method of providing
food and water. This change precluded the development of the behavioral
sink discussed in Section XII, C. [See Calhoun (1962b) for a general ac-
count of these two series of studies.] However, our concern here will be
with results not previously presented.

During the 13th, loth, and 16th months of this study estimates of
velocity, v were made for each of the 32 adult males in each of the two
societies considered here. All males were fully mature, ranging in age from
10 to 15 months of age.

Procedure for velocity estimation: As illustrated diagrammatically in
Fig. 33, each of the four pens in the room defining area A contained two
areas w^here social interaction occurred most frequently. One was on top
of the elevated artificial burrows; the other was on the floor in the immediate
vicinity of the sources of food and water. During each half hour of observa-
tion, each rat was given one velocity score for each of the eight locations
visited. On 2 days, not more than 3 days apart, Dr. Kyle R. Barbehenn
and I each recorded such velocity scores for 16 half-hour periods during
each of the three months mentioned above. For a particular month, the
estimated velocity thus consisted of the sum of the velocity scores for 32
half-hour periods of observation. The estimated velocity, v, is here taken
as the mean for three 32 half-hour sums.

The two societies were designated as lA and 2A. Thirty-two males in
lA and 32 males in 2 A survived through the 15th month. A few males died

1. The Social Use of Space

169

just before the third set of velocity estimations due to injections of a mon-
amine oxidase inhibitor (Catron), which substance had no influence on
the velocity of survivors. Thus, for a few individuals, v is based on the mean
of only two estimates of velocity for the 13th and loth months.
■ These estimates of velocity are shown as data in Fig. 38. Approximate
regression lines were fitted by eye through these points for both lA and
2A. Velocities of lA males are markedly lower than those for 2A males.
According to the general formulation of velocity, one would anticipate
that males in lA had been exposed to many more social restraints or sanc-
tions in the form of aggressive actions from associates. This, in fact, was

10 20

VELOCITY RANK

Fig. 38. Velocity, velocity-rank relationships among male rats in a closed society.

the case as is reflected by the amount of scar tissue developed over the
lumbar-sacral area from wounds received in fighting. On the basis of a 5-
point rating scale (0 = none, 5 = most) , the mean scar tissue index for
lA and 2A males was, respectively, 3.24 and 2.49.

The lowered intensity of fighting by 2A males was associated with a
higher vitamin A content of their diet. Both lA and 2 A rats were given an
dentical synthetic diet except for vitamin A. lA rats received 3 interna-
tional units per gram of diet, a high normal level in comparison to natural
foods. However, the 2A rats were given 1 2 international units per gram of
diet, a level comparable to that given to humans in high potency vitamin
pills. Although it is outside the objectives of the present discourse to detail

170 John B. Calhoun

the effects of increased levels of vitamin A on behavior, it suffices to note
that increases in vitamin A above normal levels acts as a kind of "tran-
quilizer" which reduces fighting but increases the prevalence of abnormal
behavior (i.e., females become poorer mothers while males increase the
frequency of exhibiting inappropriate sexual behavior) .

Previously we have seen (Section XIV, A) that in an A'' = 11 the ob-
served velocities for mice, as calculated by Eq. (113), closely approximated
the theoretical. We may ask the same equation of the present data. For
2A the sum of velocity indices [^iLi^i"*"^^] foi' IG hours of observation
was 1089.2, and for lA it was 804.7. Let v^^''^^ and v^^^p^ represent the
velocities of the alpha, Ist-ranked, individual and the omega, A^th ranked,
individual, as calculated by Eq. (113). For lA, v^^^p^ = 48.77 and v^f^'P^ =
1.52; while for 2A, yj^^p) = 66.01 and yjj^^p) = 2.06.

Had all the intermediate v^^^p'> been calculated, as would have been
represented by a straight line connecting these extremes on Fig. 38, it is
quite obvious that the observed w^ould differ significantly from the ex-
pected. One could drop the inquiry at this point and conclude that veloci-
ties for rats do not accord with theory as for the mice cited in Section
XIV, A. However, there are two reasons for not dropping the inquiry at
this stage. First, 32 males represent an V almost three times the theoretical
basic N, Nb = 12 while the A" = 11 for the mice closely approximated this
Nb. Second, the regression curve of velocity for the lA and 2A males con-
verge at the omega-ranked individual. Furthermore, this convergence is
at a velocity 6 to 8 times that anticipated by Eq. (113) . This fact suggests a
minimum velocity, Vm, below which rats cannot reduce their velocity and
long survive. Two such individuals with excessively reduced velocity are
shown on Fig. 38. I can only say that, on the basis of the very few individ-
uals which did develop such unusually low velocities, such individuals
usually become bloated and usually shortly died.

Accepting the indication of the reality of Vm, is there a logical basis for
recognizing its relative value? In the course of evolution, where the group
becomes adjusted to an Nb, there will in any stabilized group be an Nb-
ranked omega individual. In terms of Eq. (107) it is readily apparent that,
where R = rank in the sense used in Table XIII:

Ni = 1 -\- (N - R) (121)

Therefore, for any ith. individual where i is equivalent to its rank R, Kq.
(108) becomes

Z. The Social Use of Space

171

Since the rank, Ra, of the omega individual is A^;,:

Va =

(123)

Since Vn is the lowest velocity achieved in the normal Nb selected by evolu-
tion, Vq most likely also represents v,n. Given a v,„ observed, y£°^^ we may
calculate a v^J^'p^ appropriate to Nb. Recall that from Eq. (108) v^'^^^ in
relative terms = 1.0 and the relative velocity, v^'^^\ of any other ranked
individual is by this equation represented as proportions of the alpha's
velocity, v^^^^^ or v^^^''^ from Fig. 38 is 12. Therefore

12 = v^'^^^

i(exp)

And thus at Nb, and utilizing Eq. (123) :

yCexpat.Vft) = ^(obs) /y (rel) — y(obs) ^ jy^

(124)

However, solving Eq. (124) requires that Nb be known. The rats used were
a domesticated albino strain, Osborne-Mendel. Nb still might be 12, as we
can expect it to be for the wild type, but we have no way of knowing
directly how domestication has altered Nb. Furthermore, the artificial
environment imposed possible changes on the area. A, factor in fj. = dv/A.
So all that can be expected is that there is some optimum A^, No, har-
monious with the existing spatial structure of the environment and any
changes arising through domestication. Now, using Nb in the sense of A'',,,
values in relative (rel) terms become:

For N ^ Nb, R = l:v„ = 1

For A^ = A^6 or iV < or > Nb, R = Nb or A^: vq = Vm = l/Nb

'N - Nb

For N > Nb, R = I: vj'^^^ = 1 -

(1 - v„,)

(125)

When A^ = oo : y„ = t;^ = 1/A^6

The general equation for v, where v„ = 1/A^6, and R = velocity rank
becomes:

_ {1 _ [(jV - A^,)/A^](l - vj]
A^ - 1

,(rel)

R - N

+ v„, (126)

172

John B. Calhoun

Assuming the validity of all assumptions inherent in Eqs. (124) to
( 126) we can now approximate the No of the lA society of rats. The regres-
sion curve for lA shown in Fig. 38 gives v^°^^^ = 38. Insertion of successive
values of Nb in Eqs. (124) and (125) shows that when No = 9,
yjexpatA-6) = 1Q8 and z'i-^^" = 0.36. Since 0.36 X 108 = 38.9, it follows
that, for the strain of rat under the existing environmental conditions, 9
individuals approximate No for male rats.

Now we may return to the "tranquilizing" effect by which vitamin A
"buffered" the 2A males from the velocity-inhibiting consequences of
A'' > Nb. There needs to be a correction factor in Eq. (126) which, as a
"tranquilizing" factor, Z, increases the slope of the velocity — velocity
rank curve also increases, "pivoting" about Vm. At present there is no

IBJ

\ /V--/^

N Varies

/ -- 0 0

"^^A^-'i"*

1 1

N'48

I 1

I Z v-0.5

Fig. 39. Hj^jothetical effect of group size, A^, and tranquilizer, Z, on velocity, v,
when A'^6 = 12.

a -priori basis for determining this Z factor, which can draw upon empirical
evidence. However, one can visualize a likely formulation of Z. Z implies a
factor "blinding" awareness of the d or ^("^ of associates. Such perceptual
blinding conforms with the cV factor of ix' = d'v'/A', previously alluded to.
The critical issue concerns the influence of Z upon r„. If the velocity of the
first-ranked alpha individual never exceeds that appropriate at Nb, that is,
if

: = I'M = 1.0,

(A' 6) =

then we would have a partial basis for understanding how Z alters v.

In the absence of adequate empirical data regarding the function of Z
on V, consideration of hypothetical relationships (Fig. 39) will facilitate
our understanding to the point of enabling the design of critical experi-

1. The Social Use of Space 173

ments. First consider Eq. (125). Increasing A^ reduces v of all members.
By the time N is only a few times Nb, velocity of all members will be so
reduced as to bias the probability of the social system surviving. Minimal
velocity implies withdrawal from social interaction and restriction of
activities to independent acciuisition of food and water.

Now, assuming a vm, even though it might exceed 1.0, Fig. 39 suggests
that increasing Z will eventually elevate Va to Vm, while v^ remains at Vm
as shown in Fig. 38. Once Va reaches Vm, further increase of Z should reduce
awareness of the ^^"^ of others to the point that vq departs from Vm and
begins to approach vm- At Z = 'x , v^ = Vm. Far before Z = cc , the velocity
of all members of N will be maximal for all practical purposes. A state of
maximum conformity will then have been attained in which each individual
views every other one as so like himself that no individual imposes re-
straints on the actions of any associate. All social organization must col-
lapse, leaving a state of maximally moving independent particles, the
random contact between any two of which will be eciually satisfactory in
consummating any interaction in which two individuals are necessary.
This state demands equipotency of capacities. Complexity of behaviors will
be limited to that degree possible by every individual having identical
learning experience. Maximizing Z becomes incompatible with a high
state of learning and culture.

The human species appears to be embarked upon a journey of both
maximizing A^ and maximizing Z. If we are to avoid one of the other of
the nirvana-like states of uniform v^ or Vm, it behooves us to seek further
insight from experimentation with animal groups.

Returning from theory to reality, we may consider some correlates of
velocity. Alost of the lA and 2A rats discussed above survived to autopsy
during the 17th month of the study. Each set of males was divided into
five velocity class intervals, with as nearly as possible the same number of
rats in each velocity range. Associated conditions are graphed in Fig. 40.
As velocity increases, the amount of scar tissue derived from fighting
increases. The somewhat S-shaped character of this curve conforms with
historical events. In general, the more rats withdrew from social interaction
by reducing their velocity, the fewer wounds they received. However, a
few individuals, though having a very low velocity in late adulthood, were
characterized by extensive scar tissue because they failed to withdraw as
early in life as had their low velocity comrades. At the other extreme of
velocity, territorial males or highly dominant individuals, who were
territorial in the time dimension but not in space (i.e., the "changing-of-
the-guard" phenomenon through which several males share the dominant
role in a particular area) , by their status avoided attack and thus avoided
being wounded even though they were extremely active and inflicted

174

John B. Calhoun

wounds on associates. Their lesser degree of scar tissue contributed to a
reduction in the mean scar tissue index for high-velocity males.

Weights of adrenals, kidneys, and heart all tend to increase as velocity
increases. Despite lack of histological studies, what I suspect has happened
is that as velocity drops as animals withdraw from social interaction, organ
size decreases in accordance with decreased demands made upon them.

20 30 40

VELOCITY

Fig. 40. Some major characteristics of rats affected In-^ conditions that determine
velocity.

Fig. 40(B) simply shows the product of the weight of these three differ-
ent organs.

Most sensitive to velocity and easy to measure is fat. Fig. 40(C). Fat here
represents those abdominal deposits most easily removed: the dorsal
lumbar-sacral deposit, that in the genital mesentery, and those in the
mesenteries of the gut. Though these deposits are rarely as large among
males as for females, it is nevertheless quite clear that as rats slow down
by social withdrawal they exhibit greater propensities for converting food-
stuffs into fat.

1. The Social Use of Space 175

G. Exploratory Behavior

Ultimate exposure to some new configuration of stimuli represents the
common factor in the three phenomena encompassed by the term, "ex-
ploratory behavior." These three phenomena are: (a) the rise and decline
of a hyperactive state following exposure to a new configuration of stimuli
(see Section III, A, 4) ; (b) the rejection of new configurations (see Sec-
tion III, A, 3) ; and (c) the seeking of new configurations. We shall now
examine how the opportunity for expressing such behaviors alters an in-
dividual's attitude toward its physical and social environment.

1. The Hyperactivity Phexomenon

In an animal's normal habitat this phenomenon may be anticipated to
follow an encounter with a new configuration of stimuli at places in the
normal home range where it has not occurred during customary travels.
Field studies directed toward the elucidation of the consequences of such
encounters are essentially nonexistent. Pearson (1960), by photographing
marked mice as they move along their trailway systems, has found that,
following the experience of being trapped and handled, mice not only be-
come more active but also visit places within or near their normal home
range which are normally infrequently visited. My study of the reaction of
domesticated Norway rats following exposure to an activity alley (Sec-
tion III, A, 4) represent this same type of situation with the exception that
the induced state of hyperactivity must take place in the presence of the
new configuration of stimuli represented by the alley.

Recall that the rat is placed in a compartment with an access door at
one end of the alley. It does not have to enter, but many rats do so rather
immediately. For example consider the 73 rats involved in the analysis of
distance of termination of trips shown in Fig. 3. Analyses (Fig. 6) have
been made of their hyperactivity during the initial 2.5 hours of their resi-
dence in the activity alley. From an initial high level, activity declines
exponentially over an approximate 3-4 hour period nearly to a base level
maintained on the average through each of the 12 hours of normal height-
ened activity during the next 3 days. This pattern is shown diagram-
matrically in sketch (1) of Fig. 41. Such heightened diffuse motor activity
lacks any aura of goal direction and will be designated by the symbol,
DMA. Such a configuration of new stimuli as is represented by the alley
will be designated as E. If a rat is placed again in the alley for 2 hours on
each of several consecutive days, no appreciable amount or duration of
DMA occurs on any day. This means that there has transpired an adjust-

176

John B. Calhoun

ment, A, to E by the end of the initial 2 hours of exposure to E. After A,
motor activity persists on the average at a base line intensity, b, unless
some other new configuration E is encountered. Thus, DMA represents
increments of activity above b.

In the alley, E can only include nonsocial physical stimuli. And yet in a
social milieu of others of an animal's own species, the responses of an as-
sociate also represent an E. Provided such a social E has not previously

(1)

log
GSS

Th-

log
DMA

"_v^

(2)

log
DMA

Tob

• -Th Drift

fa-

I— 7— "op 4—

(E,)2

XE,

Time

log

(3)

t
No chonge in
threshold

GSS

aop ♦ (

log
DMA

ao"p"'|

dMA 1

Jap \

aop

i "■

t A, t
E2 2 E

t
2E4

\ t

Eg EjRemoval

to home cage

DMA DECLINE ONLY IN
HOME CAGE, THUS NO
A ADJUSTMENT TO E3

Removal

to home cage

DMA DECLINE ONLY IN
HOME CAGE, THUS NO
A ADJUSTMENT TO E4

Fig. 41. Adjustment to configurations of stimuli. See text for comment.

been encountered or if the associate is in the nonresponsive, [p], refractory
state, then the latter's behavior, though perhaps experienced before,
amounts to a previously unencountered E in the sense that it does not
permit directed motor activity leading to a satisfactory refractory period
by the other individual involved in the encounter. For this reason, it will
here be presumed that the period of decline of hyperactivity is equivalent
to the Gap frustrating type refractory period involved in normal social
intercourse.

1. The Social Use of Space 177

If rats are maintained in isolation for several weeks or months and then
placed in an emotional activity alley, a large proportion of them never
go out into the alley from the "home" compartment. From this I infer that
there is some upper threshold, Th, of DMA which "overloads" that neural
circuit permitting its expression. The three dots in sketch (2) denote such
overbadings. In any sample of subjects, most of whom avoid entering the
alley, there are a few with various lengthened latencies, causing delays of
up to nearly the end of the 2-hour test period before initiation of the initial
phase of intense hyperactivity. From this fact it is apparent that the normal
DMA becomes replaced by some generalized stress state, GSS, which
follows a similar (but lower?) rate of decline than the DMA. GSS lacks a
striated muscle component. Once it has dechned to Th, GSS becomes
transformed to DMA. Such subjects, owing to their having been protected
from impinging external stimuli for such a long time during their isolation,
view the alley configuration as an increased intensity, E. An inference from
this is that there is a slow drift downward of Th toward 6 during the weeks
of isolation when opportunities for adjustments Ai • • • An in response to
El • ' • to En are absent. Thus, depending upon how extensive this drift
has been, any particular E, such as represented by the activity alley, raises
the GSS different amounts above Th. I suspect, though my data are not
conclusive, that the more elevated GSS is above Th the lower will be its
rate of decline. These differences are diagrammatically shown in sketch
(2) of Fig. 41. Given sufficient time with no interference by other E's,
DMA will eventually reach h and an adjustment, A, transpires.

Now we may consider the more normal coiu-se of maturation, experience
with a sequence of ^i • • • En [sketch (3) in Fig. 41]. One general observa-
tion first. The greater has been the experience of rats in the sense of a larger
number of different E's to which adjustments, A, have been made, the
lower will be the probability of withdrawal in the form of failing to enter
the alley from the starting "home" compartment. An explicit experiment
concerning this point has already been cited with reference to Table lb.
Such results indicate an elevation of Th to Tha following each A, and this
elevation will be proportional to the magnitude of E provided E does not
elicit a DMA exceeding Th. Thus, at some later time an intense E, desig-
nated as 2Ei in sketch (3), will result in adjustment Ai, although had a
2E configuration occurred earlier it would have resulted in the undesirable
consequences accompanying elevation of DMA above Th. There is another
conclusion, somewhat more tentative, though some of my results do sup-
port it. This is that with each successive E, the rate of decline of DMA
increases. In other words, the aap refractory period decreases. Thus, the
more adjustments an individual makes, the better he will be al)le to curtail
oif„p = df. I have already mentioned the failure of a second exposure to a

178 John B. Calhoun

particular E to induce any material increase in DMA. This means that if
E merely represents a configuration of strange stimuli with which the
individual can interact in no meaningful way, the second exposure to E,
that is (£'1)2, will evoke no response. (£"1)2 will merely be ignored. However,
if some aspects of the Ei configuration permit meaningful interaction,
then (£'1)2 will result in an interaction having an «„„ refractory period
proportional to the evoked directed motor action, dAIA. This a„„ will
normally be of shorter duration than a„p to (£"1)1 since some components
of the configuration are likely to be irrelevant to dAIA. Furthermore,
dMA may be expected to be maintained at near its initial intensity until
the evoked behavior terminates. No further elevation of Th accompanies
(£1)1 ••• {Ei)n. Persistent recurrence of any specific E merely serves to
prevent the drifting downward of Th.

Each adjustment to a new E configuration resulting in an elevation in
Th represents a contribution to the individual's psychological area. A" .

2. The Seeking of New. Configurations

Consider two groups of individuals, A and B, of which the members of
each for a fairly long period merely reexperience particular sets of £"s.
Members of group A differ from those of B in that they are exposed to a
larger assembly of different £"s. Two such groups have been considered in
Section III, A, 5 and Table I la. The fact that more of the A-type individuals
entered the alley when exposed to this new E configuration of stimuli in-
dicates that reexperience of a larger assembly of different £"s does maintain
Th at higher levels despite some downward drift. The A-type individuals
obviously have a larger psychological area. A", than the B-type ones. A"
essentially connotes capacity to adjust. In any environment presenting
frequent necessity for adjustment, a high A" will prove advantageous. So
we need to consider the question of maximizing A" .

Persistence in repeated interaction with certain £"s is necessary or de-
sirable because of acquired reward value accompanying the directed motor
activity, dMA, appropriate to these £"s. Such dMA can only preserve Th
at a given level. Response to these £"s involved in this dMA consumes time.
If all w^aking time becomes relegated to dMA-evoking £"s, the individual
will develop a static A". Furthermore, the more waking time becomes
filled with repetitions of response to any given E, the more restricted will
be A" . It is thus obvious that the best strategy for maximizing A" will be
to reserve a portion of one's waking hours simply for responding to new
£"s. Some as yet unknown but probably fairly long time, certainly of the
order of several days even for rats, must elapse between one exposure to
such a new E and a reexposure, permitting sufficient extinction of the A

1. The Social Use of Space 179

adjustment so that at the re-exposure the E eUcits a DMA comparable to
that of the initial exposure. Because of the limitations of time and space in
which any individual's activities must transpire, maximizing A" demands
an active seeking of new £"s. This seeking, I term v". It represents a kind
of velocity difficult to distinguish from the normal velocity, v, unless one
is aware of the history of an individual with reference to its pattern of
repeating specific dMA. In empirical experimental situations the relative
magnitude of v'/ may be determined by observing the probability of ap-
proaching a new E introduced into an individual's home range so that nor-
mal movements will produce exposure. Ecologists concerned with con-
trolling the density of species which damage human property or serve as
hosts for diseases transmittable to man have been aware of v" in a negative
sense. They (e.g., Chitty and Southern, 1954) have noted the avoidance of
new objects, such as poison baits or traps. This avoidance has been termed
"the strange-object response."

3. Active Rejection of New Configurations

Distinct from the relative attraction to or avoidance of a new configura-
tion is the phenomenon of physically rejecting or psychologically blocking
awareness of new £"s. Processes included under this phenomenon of rejec-
tion may be designated as d". A grasp of the -types of phenomena subsumed
under d" may be obtained through considering a case observed in my
laboratory.

Large "life-space" cages, LSC, were designed to provide an optimum
situation for the breeding of the very sensitive wild Norway rats in the
laboratory. A 16 X 25-inch floor provided access to an activity wheel on
one side and a lever on the other, which when pressed, provided a drop of
water. From this floor two ramps led to a partitioned second floor of eciual
dimensions. From one side of the second floor rats had access to one 8 X
8 X 6-inch nest box, while two next boxes w^ere accessible from the other
side of the second floor. One male and two female adult rats lived in each
of six cages. At the time in question three of these cages each also contained
a recently weaned litter. Up to this time the water-providing lever apparatus
had not been delivered by the manufacturer. In its place the adults were
provided water through a drinking tube from a bottle, as had been the
practice since they were captured in the wild as juveniles. When the lever
apparatuses became available, one was inserted into each cage and the
water bottle was remo^xd. By the following morning when the cages were
next examined, the situation in each cage was identical; all movable ob-
jects available to the rats, paper used as nesting material and orange peels,
had been piled over the lever, completely hiding it.

180 John B. Calhoun

In those cages containing recently weaned young, the young soon scat-
tered the pile of objects, exposing the lever, and in so doing accidentally
pressed the lever and gradually learned its function. There then followed a
repeated process of covering the lever by the adults and its removal by the
young. Through this process the adults were forced to face E, represented
by the lever and its attached water reservoir, sufficiently to permit an A
adjustment to E thi-ough the DMA decline process. Several points may be
deduced from these observations. Th for the adults had previously drifted
downward toward b so that the E lever configuration caused DMA to
exceed it. Furthermore, at weaning Th is sufficiently removed from b that
many ^'s will fail to evoke DMA elevation above Th. Had evolution not
resulted in such a balance between neurology and physiology, animals just
emerging out into the many E's of their environment would immediately
be forced into a withdrawal state. Although I shall not go into this problem
here, it is obvious that retardation of "weaning" increases the probability
of withdrawal .

In the three cages lacking recently weaned young, the pile of material
covering the levers remained undisturbed for several days until the rats
were so weak from lack of water that it was apparent that their rejection
of the lever was so complete that they would die before getting the op-
portunity to learn its function through chance depression of it. Replacing
the former water bottle merely satisfied their thirst but failed to alter their
rejection of E. "Teaching" the rats the lever was finally accomplished by
taking all movable objects from the cage and gradually increasing the
interval during which the water bottle was removed. It took 3 weeks to
reach the same level of lever pressing by these rats that was obtained within
3 days by adults when young not only made rejection impossible but also
set an example of adjusted interaction with the E lever configuration.

Other examples of such d" active rejection, but toward social E's, have
already been given in Section XIII, B, 4 in connection with the three
examples involving Barnett's Norway rats, the "Freedom Riders," and
the formation of the Co7 Colony IB of house mice. Thus, in terms of active
rejection, d" becomes essentially synonymous with intensity of action
tow^ard another, i'-^f as given by Eq. (114). But d" must also encompass
the more strictly psychological phenomena of psychological deafness or
psychological blindness such as characterizes the "malingering type" of
individual.

Note that this consideration of exploratory behavior has lead to formu-
lations of d", v", and A". As for prior comparable terms:

/' = (d"v"/A") (127)

Here /x" represents a third contact modifying function when relating to

1. The Social Use of Space 181

a social group. Just as with /x aud n', so it is apparent here that the magni-
tude of v" and A" will normally change by comparable degrees in the same
direction and that d" will approximately vary inversely with v".

4. The Effect of Interval between E's, on Velocity

Further examination of the data in Table lib, in the light of the formu-
lations relating to exploratory behavior and configurations of stimuli, pro-
vides additional insight into the origin of reductions in velocity, both v
and v". Refer to Section III, A, 3 for other comments. Presentation of the
experiment in terms of the present formulations is as follows :

The subjects consisted of male albino Osborne-Mendel strain rats iso-
lated at weaning. At this time each rat was placed in a 6 X 6 X 8 inch
cage from which it could not see out. Water and food were introduced
through channels from the outside which prevented the rats from seeing
the experimenters or being handled by them. AH rats remained in such
isolation for approximately three months before further treatment. These
isolation cages may be termed an Ei configuration of stimuh. Due to the
long absence of opportunity for adjustment to new configurations of
stimuli, the Th of all rats probably drifted toward b.

At the end of the three months of isolation the subjects were divided
into four groups: A (20 rats), B (24 rats), C (16 rats), and D (16 rats). On
each of 10 days during the next two weeks each member of Group C was
exposed for two hours to a new configuration of stimuli, E2, which was a
Wahman activity wheel; all rats so exposed entered the wheel and ran
during each of the 10 days. Similarly, members of Group C were exposed
to an E3 configuration of stimuh. This exposure consisted of placing the
rats in a 2 X o-foot pen where they had the opportunity to climb onto a
central platform where a lever could be pressed to receive a drop of water.
Each day of this 10-day training period half the members of Group D
were exposed for two hours to E2 and then, immediately following, for
two hours to Ez. The other half of Group D were similarly treated but were
exposed to E3 just prior to E2. Group A remained in their isolation cages
during these two weeks.

During the third experimental week every rat in all four groups was given
a two-hour exposure to the NIH Emotional Activity Alley on each of
four successive days. This alley represented an E4 configuration of stimuli.
For half the rats in each group Ei had a stationary floor, a condition we
may designate as Ei^. For the remaining rats Ei had a tilting floor which
clanged as the rats ran across it. This modification is designated as Eis,
which represents a much more intense or strange configuration than EiA.
Many rats avoided entering the alley.

182 John B. Calhoun

Regardless of the amount of prior opportunity to adjust to novel con-
figurations, the more intense £"48 elicited a more marked avoidance than
did Ei\. (See Section III, A, 3.) However, our present concern is with
a different aspect of the results in Table lib.

I'pon exposure to EiA it appears that prior experiences with E^ was
much more effecti\'e than with E-i in reducing avoidance of the EiA con-
figuration. But members of both Groups B and C evinced much less
avoidance of Ei than did members of Clroup A. This supports the formula-
tion that prior opportunity to adjust to new configurations elevates Th
so that at a following exposure to another new configuration of stimuli,
DMA is less likely to exceed Th. Avoiding entering the alley is taken as
evidence of DMA exceeding Th.

These results confirmed prior hypotheses. However, it was further as-
sumed that rats of Group D would exhibit the most marked accommoda-
tion to Ei since they would ha^-e had twice the opportunity for making
adjustments to new E's. And yet e\ei\ to E^a, the rats of Group D showed
little better capacity for adjustment than did members of Group B, and
much less than did rats of Group C. Upon elevation of the intensity of
Ei to EiB, members of Group D exhibited an extremely more marked reduc-
tion in capacity to adjust than did the rats of Groups B and C, that pre-
sumably had less opportunity for "training" in making adjustments.

These results apparently contradict the theory. But consider the follow-
ing. For rats exposed to new configurations of stimuli, such as £"4, but
permitted to remain for several rather than for 2 hours, it has been noted
that many individuals reciuire up to 3 hours for DMA to decline to h. In
the 2000-odd tests where rats have been exposed to the AVtype alley con-
figuration to test for emotionality, the tacit assumption has been made
that the remaining decline in DMA will take place after the return of the
rats to their accustomed environment. However, this opportunity did not
prevail when, after 2 hours in E2 or E^, rats were transferred to the
opposite E.

Events presumably transpiring are diagrammed in the left-hand side of
sketch (4) in Fig. 41. Upon exposure to Es after only partial decline of
DMA following exposure to E2, the same increment in DMA is elicited,
but its rise starts at the point of a still fairly high level of DAIA. Thus,
this second increment in DMA forces it above the threshold, Th, where
DjVIA is transformed from difi'use motor activity into the generalized stress
state, GSS. Although GSS and DMA were not measured during E3 (or
during E2, if it came second for Group D) , both must have eventually com-
pletely disappeared after the usual return to the home cage. And yet the
very failure of many members of Group D to enter the Ei alley when given

1. The Social Use of Space

183

exposure to it suggests that for each degree of GSS induced by too close
spacing of consecutive new E's, there had transpired a drop in Th and B
such that Th drops relatively more than h. Each of the ten opportunities for
consecutive exposure to E2 and Es must have narrowed the gap between h
and Th. Thus the later exposure to Ei must have caused "overloading"
for most Group D rats to the extent that GSS had not declined to Th by
the end of the two hours in Ei. This meant that decline in DMA to base
level took place in Ei. Return to Ei for 2 hours on each of the following 3
days was characterized by persistence in avoiding entry into the Ei alley
configuration by most rats avoiding it on first exposure. Avoidance, v", of
any new E will thus be proportional to the degree of GSS "overloading"
elicited by the new E's.

Although new E's have been considered above in the sense of physical
nonsocial configurations, we may consider the consequences of too closely
spaced new or undesirable £"s in the social sense of i^'^'s of associates. Re-
call that such i^^^'s represent social restraints or sanctions imposed by
associates. For such sanctions to become effective in reducing velocity, v,
that is for causing a drop in baseline of activity, consecutive sanctions by
the same or different associates must be sufficiently closely spaced to induce
a GSS. No opportunity for v" avoidance is possible. In fact, with the drop
in velocity v (synonymous with h), psychological area A" must be re-
stricted through failure to make adjustments A to E while still in the
presence of E. And as we have seen, as A" decHnes so will v". This means
that as velocity, v, declines, the individuals have even less capacity to
avoid strange stimuli. However, recall that as v" declines d" increases.

I can cite no quantitative data to support this conclusion of d" increasing
as y" decreases. However, the following observations support its reality.
While making the observations on velocity of rats summarized in Fig. 38, I
was consistently impressed by the manner in which most very low-velocity
rats moved "through" their associates. During those rare times when
active, they would pass by associates as if completely psychologically
blind to their presence. Furthermore, their blase, unaffective mode of
posture equally failed to elicit response from associates. The completeness
of this psychological rejection of reality is reflected in their smaller organ
weight and larger amount of fat (Fig. 40), corresponding to states char-
acterizing rats maintained in approximate isolation by restrictions to
groups of 1 male with 2 females in small cages or pens.

A further corroborative observation comes from the study referred to
in Section XH, A. Among the wild Norway rats in that study was a small
group designated as possessing an array of aberrant symptoms and be-
havior which I called the "syndrome of the social outcast." From compari-

184 John B. Calhoun

son with rats described in Sections XII, C and XIV, F, I now know that
these social outcasts must have been very low-velocity rats. Of all the rats
involved in the study referred to in Section XII, A, they were the only
ones ever caught in "Havahart" traps. These large, shiny wire-meshed
traps with a door opened at either end and shiny metal treadles on the
center floor were regularly placed on trials. All other rats invariably ran
around these traps. Yet the social outcasts apparently ran into them with-
out ever sensing their presence. This sensory unawareness is the low v"
factor characterizing low-velocity rats.

XV. Conclusion

Man did emerge from the trials, successfully overcome, of a myriad of
ever more simple forms. I have attempted to formulate some phenomena
which have affected man's social evolution. Some of these phenomena
appear no longer directly operati^'e on the human animal. They neverthe-
less left their imprint on man's capacity to adjust physiologically and
psychologically to the social system in which he lives. Foremost among these
presumed legacies are the limitations imposed upon him from his origin
out of an evolutionary line in which optimum adjustment demanded living
in small groups, not exceeding twice twelve individuals.

Only yesterday, as one may measure evolutionary time in units of ten
thousand years, did man begin his attempt to escape this evolutionary
bond. But cultural evolution has not produced escape from this bond,
merely accommodation to it. Human society has developed the form of a
many-layered chain link armor. Each link is composed of not much less
than, nor many more than, twelve individuals. The links have a fluid
character. Through time, any one individual shifts his membership back
and forth among several joining links. This poetic view embodies the es-
sence of reality.

In contrast to such evolutionary legacies, there exist certain principles
of social physics which must affect all social animals, man included. These
principles derive from certain universals I have called velocity, target
diameter, area, basic group size, the social refractory period, threshold for
tolerance for change, and the like. I am fully cognizant that my formulations
represent only crude approximations to reality, and may in fact contain
several errors of logic. Yet we must develop adequate formulations along
the lines I have attempted or all efforts to gain insight into the individual's
involvement in social action will prove sterile.

This sterility will crown the endeavors of both classical physiology and

7. The Social Use of Space 185

psychology. We can no longer afford to ignore the impact of the social
setting on the individual's behavior and physiology. And without knowl-
edge of evolutionary limitations and universal principles of social physics,
consideration of the social setting will also prove of little avail. The search
for conceptualizations, adequate for furthering this objective, serves as the
justification for inclusion of my effort to introduce the more strictly physio-
logical discussions by the other authors in these volumes.

References

Bamett, S. A. (1955). Competition among wild rats. Nature 175, 126.

Birdsell, J. B. (1953). Some enivronmental and cultural factors influencing the structur-
ing of Australian Aboriginal populations. Am. Naturalist 87, 169-207.
Birdsell, J. B. (1957). Some population problems involving Pleistocene man. Symp

on Quant. Biol. 22, 47-69.
Braidwood, R. J., and Reed, C. A. (1957). The achievement and early consequences of

food-production. Symp. on Quant. Biol. 22, 19-31.
Broadbent, D. E. (1958). "Perception and Communication." Pergamon Press, New York.
Caldwell, J. M., Ranson, S. W., and Sacks, J. G. (1951). Group panic and other mass

disruptive reactions. U.S. Armed Forces Med. J. 2, 541-567.
Calhoun, J. B. (1949). A method of self-control of population growth among mammals

living in the wild. Science 109, 333-335.
Calhoun, J. B. (Ed.) (1949-1957). Annual reports— North American Census of Small

Mammals.
Calhoun, J. B. (1952). The social aspects of population dynamics. J. Mammal. 33, 139-

159.
Calhoun, J. B. (1956). A comparative study of the social behavior of two inbred strains

of house mice. Ecol. Monographs 26, 81-103.
Calhoun, J. B. (1957). Social welfare as a variable in population dynamics. Cold Spring

Harbor Symp. on Quant. Biol., 22:339-956.
Calhoun, J. B. (1962a). A behavioral sink. In "Roots of Behavior" (E. L. Bliss, ed.).

Chapter 22. Harper Hoeber, New York.
Calhoun, J. B. (1962b). Population density and social pathology. Sci. Am. 206,

139-146.
Calhoun, J. B. (1963). The ecology and sociology of the Norway rat. U. S. Public Health

Service Pub. No. 1008, Government Printing Office. 288 pp.
Calhoun, J. B., and Casby, J. U. (1958). Calculation of home range and density of small

mammals. U. S. Public Health Monograph No. 55.
Calhoun, J. B., and Webb, W. L. (1953). Induced emigrations of small mammals.

Science 117, 358-360.
Carpenter, C. R. (1962j. Field studies of a primate population. In "Roots of Behavior"

(E. L. Bliss, ed.). Chapter 21, Harper (Hoeber), New York.
Chitty, D., and Southern, H. N. (1954). "Control of Rats and Mice." Oxford Univ.

Press (Clarendon), London and New York.
Dahlberg, B. L., and Guettinger, R. C. (1956). The white-tailed deer in Wisconsin.

Wisconsin Conservation Dept. Tech. Wildlife Bull. No. 14.
Deevey, E. S., Jr. (1960). The human population. Sci. Am. 203:194-204.

186 John B. Calhoun

Duff, W., and Kew, M. (1957). Anthony Island, a Home of the Haidas. Report of the

Provincial Museum of Natural History and Anthropology.
Elton, C. (1942). "Voles, Mice and Lemmings — Problems in Population Dynamics."

Oxford Univ. Press, London and New York.
Fredericson, E. (1951). Time and aggression. Psychol. Rev. 58, 41-51.
Hall, C. S. (1934). Defecation and urination as measure of individual differences in

emotionalit}" in the rat. Psychol. Bull. 31, 604.
Hallowell, A. L (1960). Self, society, and culture in phylogenetic perspective. In "The

Evolution of Man." (S. Tax, ed.), pp. 309-372. Univ. of Chicago Press, Chicago,

Illinois.
Hardin, G. (1960). The competitive exclusion principle. Science 131, 1292-1297.
Hayne, D. W. (1949). Two methods for estimating population from trapping records.

/. Mammal. 30, 399-411.
Hutchinson, G. E. (1957). Concluding remarks (to Population Studies: Animal Ecology

and Demography) Cold Spring Harbor Symposia on Quant. Biol. 22, 415-427.
Kabat, C, CoUias, N. E., and Guettinger, R. C. (1953). Some winter habits of white-
tailed deer and the development of census methods in the Flag yard of northern

Wisconsin. Wisconsin Conservation Dept. Tech. Wildlife Bull. No. 7.
Leopold, A. S., Rine.y, T., McCain, R., and Tevis, L., Jr. (1951). The jawbone deer herd.

Slate of Calif. Dept. Nat. Resources, Game Bull. No 4.
Moreno, J. L. (1953). "Who Shall Survive?" Beacon House, Beacon, New York.
Moreno, J. L., and Jennings, H. H. (1960). Statistics of social configuration, pp. 19-51.

In "The Sociometry Reader" (J. L. Moreno et ah, eds.). The Free Press, Glencoe,

Illinois.
Patric, E. F. (1958). Some Properties of the Small Mammal Populations of the Hunting-
ton Wildlife Forest. Ph.D. Thesis, College of Forestry, New York State Univ.,

Syracuse, New York.
Pearson, O. P. (1960). Habits of Microtus calif ornicus revealed by automatic photo-
graphic recorders. Ecol. Monographs 30, 232-249.
Ranson, S. W. (1949). The normal battle reaction: Its relation to the pathologic battle

reaction. In "Combat Psychiatry," Bull. U.S. Army Med. Dept. 9, SuppL, 3-11.
Riesman, D., Glazer, N., and Denney, R. (1953). "The Lonely Crowd." Doubleday,

Garden City, New York.
Sanders, R. D. (1939). Results of a study of the harvesting of white-tailed deer in the

Chequamegon National Forest. Trans. North A)n. Wildlife Conf. 4, 549-553.
Schaefer, E. S. (1959). A circumplex model for maternal behavior. J. Abnormal Social

Psychol. 59, 226-235.
Schaefer, E. S. (1961). Converging conceptual models for maternal behavior and for

child behavior, pp. 124-146. In "Parental Attitudes and Child Behavior" (J. C.

Glidewell, ed.), C. C. Thomas, Springfield, Illinois.
Schneirla, T. C, and Tobach, E. (1962). Eliminative responses in mice and rats and the

problem of "emotionahty." In "Roots of Behavior" (E. L. Bliss, ed.). Chapter 15.

Harper (Hoeber), New York.
Schorger, A. W. (1953). The white-tailed deer in early Wisconsin. Trans. Wisconsin

Acad. Sci., Arts and Letters 42, 197-247.
Shiras, G. III. (1921). The wildlife of Lake Superior, past and present. Natl. Geograph.

Mag. 40(2), 113-204.
Stewart, J. Q. (1948). Concerning social physics. Sci. American 178, 20-23.
Swift, E. (1948). Wisconsin's deer damage to forest reproduction survey — final report.
Wisconsin Conservation Dept. Publ. 347.

1. The Social Use of Space 187

Swift, E. (1946). A history of Wisconsin deer. Wisconsin Conservation Dept. Publ. 323.

von Foerster, H., Mora, P. M., and Amiot, L. W. (1960). Doomdsay: Friday, 13 Novem-
ber, A. D. 2026. Science 132, 1291-1295.

Webb, W. L., and Rosasco, E. M. (1953). An Investigation of the Properties of Induced
Mass Movements of the Small Mammals of the Genera, Peromijscus, Clethrionomys,
Blarina and Sorex. 100 pp. mimeographed. College of Forestry, State University of
New York. Re Contract DA-49-007-MD-325, Medical Research and Development
Board, Office of the Surgeon General, U. S. Army.

Whyte, W. H. (1956). "The Organization Man." Simon and Schuster, New York.

Zimmerman, C. C, and Broderick, C. B. (1954). Nature and role of informal family
groups. Marriage and Family Living 16, 107-111.

Zimmerman, C. C, and Cervantes, L. F. (1960). "Successful American Families."
Pageant Press, New York.

Zippin, C. (1956). An evaluation of the removal method of estimating animal populations.
Biometrics 12, 163-189.

Endocrine Adaptive Mechanisms
and the Physiologic Regulation
of Population Growth*

J. J. CHRISTIAN V^^

Division of Endocrinology and Reproduction, Research Laboratories,
Albert Einstein Medical Center {North), Philadelphia, Pennsylvania

General Introduction 189

Part 1. The Endocrine Adaptive Mechanisms 191

I. Introduction 191

II. The Endocrine Glands of Adaptation 192

A. The Adrenal Glands 192

B. The Thyroid Gland 228

C. Other Endocrine Adaptive Factors 240

D. General Measurements of the Endocrine Adaptive Responses 242

Part 2. Physiologic Adaptation and Mammalian Populations 261

I. Introduction 261

II. Endocrine Responses to Social Pressures and to Population Density- - 263

A. Experiments in the Laboratory with Populations of Fixed Sizfc_-_ 263

B. Freely Growing Populations 281

C. Natural Populations 300

III. Conclusion 325

References 328

General Introduction

Endocrine adaptive responses have become of particular interest to the
mammalogist in recent years because of the likelihood that they play an
important role in the regulation of the growth of mammalian populations.
Sufficient evidence has accumulated from the field and laboratory to war-
rant stating with fair certainty that these adaptive mechanisms are opera-
tive in and related to changes in the size of mammalian populations.
However, there is still considerable uncertainty about their precise role and
relative importance in the regulation of population growth, especially with
regard to generalizing to a large number of species from the few species for
which data are presently available. There also exists a great deal of un-

* Note added in proof. For additional references pertaining to recent work on endocrines
and population, the reader is referred to Christian (1961, 1963a & b).

189

190 /• /• Christian

certainty about the relationships of various environmental factors to the
adaptive mechanisms. Then, besides the interest in adapti\-e mechanisms in
relation to population growth, there is the frequently overlooked fact that
these same phj^siologic reactions may effect profound morphologic changes
in the members of a population and therefore directly affect the taxonomist
who must use morphologic criteria for distinguishing species and subspecies.
It is entirely possible that many subspecific descriptions have been based
on morphologic differences resultmg from differences in the densities of the
populations on which the descriptions are based. The mammalogist inter-
ested in reproduction in mammals must take adaptive mechanisms into
consideration, as alterations in reproductive functions are an integral part
of these same adaptive responses. Therefore there is adequate justification
for this chapter on the endocrine adaptive responses, their effects, and their
relationships to the densities of mammalian populations.

No matter how well an animal may be genetically adapted to its general
environment, it still must have sufficient adaptive flexibility to meet the
daily and seasonal environmental changes, as well as emergency situations,
to which it will normally be subjected, and still maintain a constant internal
environment. Nothing in the daily life and external environment of an
animal remains constant; on the contrary there frequently are very sudden,
often extreme, shifts in the environment which are stimuli that, if unop-
posed, would alter the internal environment of the animals. But the internal
environment must remain constant if the animal is to survive. Therefore
there must be a constantly active system of physiologic feedback mecha-
nisms to compensate immediately for any tendencies to shift the internal
environment. However, these adaptive responses do not take place without
producing measurable effects in the organs and glands primarily responsible
for meeting the altered demands. Compensation for a life-maintaining
change frequently occurs at the expense of some function less immediately
important for survival, for example, reproduction. Consequently reproduc-
tive function declines measurably in the face of a need to maintain a con-
stant internal physiologic state in the presence of adversity. The adaptive
responses are changing constantly in degree to meet constantly changing
daily circumslances, and it is generally thought that a certain amount of
change is necessary to maintain the integrity of the system so that it will
be capable of responding to more demanding circumstances. It is not sur-
prising that at any given moment the physiologic status of a mammal
reflects its total environment and that the whole system is in a constant
stage of change, but for these same reasons it becomes difficult to study
such a dynamic system in the complex environments of natural populations.
Therefore a great deal of the existing evidence on the adaptive mechanisms
of mammals, especially in relation to population density, has been gained
by studies in the laboratory.

2. Endocrines and Populations 191

Past research on the adaptive mechanisms has emphasized the adrenal
cortices and to a lesser extent the thyroid gland and their hormones. Un-
questionably the adrenal cortex is essential to life and plays a basic role in
the adaptive responses, nevertheless there is a real tendency to overlook
the paramount importance of other systems and organs which also respond
to adverse circumstances. In fact their actions are simultaneous with and
inseparable from the actions of the adrenal cortex in many instances. A
great many responses on the part of the organism act in concert to prevent
any alteration in the basic physiology of the animal and to meet emergency
needs. The central nervous system is a major and integral part of these
adaptive mechanisms. Our understanding and interpretation of the physio-
logic changes taking place under a given set of circumstances too frequently
suffer from a tendency to think statically and in terms of isolated organs,
systems, or hormones — a result of the kind of experimental approach
necessary to understand the actions of various glands and their secretions.
The isolated organ concept must give way to thinking in terms of dyna-
mically interacting systems. However, to describe these mechanisms and
their effects in dynamic inclusive terms is extremely difficult and has been
made even more so by the recent elucidation of the key role played by the
central nervous system in regulating the activities of the glands of internal
secretion, as well as by the realization that we are dealing with an enor-
mously complex interacting system further complicated by the complex
temporal relationships of the responses of these systems to the applied
stimulus and to each other.

The balance of this chapter will be devoted to a more detailed discussion,
first of the physiologic adaptive mechanisms themselves, and then with
particular attention to the evidence implicating these responses in the
regulation of the growth of mammalian populations. An attempt will be
made to clarify some of these responses and to indicate areas where further
research is needed, especially in relation to behavior and population density.
Subjects which are adequately covered in the usual textbooks of physiology
will be omitted or only briefly summarized.

Part 1. The Endocrine Adaptive Mechanisms

I. Introduction

When a mammal is subjected to a stimulus which, if unopposed, would
result at least in a change in its internal physiology, and more likely produce;
a circulatory collapse, a series of neural, neuroendocrine, endocrme, and
vascular responses follow which counteract the deleterious effects of the

192 J- J- Christian

stimulus and also supply the increased needs of many tissues in order to
meet the situation. Selye (1950) introduced the phrase "alarm stimulus" to
describe such a stimulus which produces shock and evokes the usual
physiologic responses to shock. In the present account an alarm stimulus is
defined as any stimulus which, when applied to a mammal, tends to alter
fluid and circulatory homeostasis, and therefore necessitates a physiologic
adaptive response. This definition is somewhat circular insofar as it is in
terms of a response, but it is not restrictive, and it does not imply that the
adrenal cortex (at least that part responsible for the secretion of carbo-
hydrate-active corticoids) is an essential participant, as is so often assumed.
There may be qualitative similarities in the responses to different stimuli,
but detailed studies suggest that there are all degrees of variation in the
degree of participation of various systems and organs to a gi\'en stimulus.
Probably the prime objection to the current concept of "nonspecific"
response is the practical one that uncritical usage has tended to obscure
important differences in the physiologic responses to different stimuli.

It should be pointed out that the degree of these responses appears to be
relative, as the same responses qualitatively are essential for daily life, but
must increase quantitatively in the face of adverse circumstances. A pri-
mary function of the endocrine adaptive responses is to insure an adequate
circulation with an adequate supply of glucose and oxygen to tissues essen-
tial for emergency situations. Part of this function is the maintenance of an
adequate circulatory volume and proper electrolyte and fluid balances.
These adaptive responses will be discussed in greater detail in the following
account.

II. The Endocrine Glands of Adaptation
A. The Adrenal Glands

1. Introduction

This discussion is primarily for the benefit of those who are interested in
the physiological and comparative aspects of mammalogy. Therefore what
is known of the endocrine adaptive mechanisms will be outlined without
dwelling on details or becoming involved in the minor details or contro-
versies of today's frontiers in endocrinology.

Research on adaptive mechanisms to a large extent has centered around
the adrenal glands, especially the cortex. One of the factors tending to
synonomize "stress" with adrenocortical activity has been the measure-
ment steroid secretion, weight, ascorbic acid depletion, and cholesterol

2. Endocrines and Populations 193

content of the adrenals to determine whether and to what degree a stunulus
produces ''stress." Although the adrenals are of primary importance in
physiologic adaptation to changing needs, it is important not to equate
adaptation solely with adrenal function or to assume that the only function
of the adrenals is to enable an organism to meet new and sudden demands.
It is appropriate, especially for the mammalogist, to discuss adrenal glands
in some detaU because of their importance and because of the convenience
of using them as indices of the degree of adaptive response to particular
situations or stimuli. However, judgment must be used in interpreting the
results of measurements of adrenal function, and one must realize that
there are many other responses which are measured with extreme difficulty,
and yet others may be completely masked by extraneous factors.

a. General morphology of the adrenal glands. The anatomy of the adrenal
glands is discussed in detail in many texts and papers on histology, gross
anatomy, and comparative anatomy. Attention is called to the books by
Bourne (1949), Hartman and Brownell (1949), Bachman (1954), and
Jones (1957) for general treatments, especially from the comparative point
of view.

The adrenal glands are yellowish paired organs lying at or near the
anterior poles of the kidneys. Their position and form vary considerably
from species to species. For example, in rabbits (Stjlvilagus, Oryctolagus)
they are oval discoid organs closely applied to the vena cava; in wood-
chucks (Marmota) they are sausage-shaped and lie between the kidneys
and the midline, usually closer to the latter; in mice and voles of almost all
species they are round, oval, or pyramidal and lie approximated to the
poles of the kidneys; and in the bats Myotis and Pipistrellus they lie be-
neath a layer of the renal capsule. These examples simply serve to illustrate
the wide variations that occur in their gross shape and position.

Two distinct portions of the adrenal are discernible when they are sec-
tioned and examined grossly: a dark reddish brown or gray central core,
the medulla ; and a wide outer portion, the cortex, which is usually yellowish
but may be gray or even translucent reddish brown, dependmg on the
activity of the gland. The yellow color is imparted by lipids contained in
the cortical cells; thus color will vary with changes in the lipid content.
Usually the cortex is quite wide, comprising from one-half to two-thirds of
the radius of the gland. However, in some of the adult soricid shrews (Sorex
jumeus, S. cinereus, S. palustris, S. dispar, and Microsorex hoyi) the gland
consists almost entirely of medulla and has a very narrow cortex only a few
cells wide. The extreme narrowness of the cortex is especially pronounced
in mature male shrews.

The adrenal gland is surrounded by a connective tissue capsule from
which a stromal framework of connective tissue descends into the cortex.

194 /• /• Christian

The amount of cortical stroma may vary considerably; it is inconspicuous
in most rodents whereas it is marked in most carnivores.

b. Zonation of the adrenal cortex. Three distinct major zones usually are
identifiable histologically in the cortex, although the zonation is difficult to
discern in a number of species (Bourne, 1949) . An outer thin zona glomeru-
losa lies just beneath the capsule and consists of loops or balls of rather
large cells with relatively clear cytoplasm. Central to the zona glomerulosa
is a wide central zona fasciculata, which is composed of radially arranged
straight cords of polyhedral cells that usually contain numerous cytoplasmic
lipid vacuoles. Lipid vacuoles occur in the cortical cells of most species,
but they may be absent in some, for example, the golden hamster (il/c.so-
cricetus auratus) (Alpert, 1950; Knigge, 1954a; Schindler and Knigge,
1959a). Little or no lipid is present in the cortices of cattle, sheep, and
pigs (Deane and Seligman, 1953). When present, the vacuoles may vary
considerably in size and number, depending on variations in the activity
of the cortex. The cells of the outer half of the zona fasciculata usually are
larger and contain more lipid than those in the inner half of the zone. The
fascicular cords are arranged as paired columns of cells lining vascular
sinusoids in man and monkeys (Elias and Pauly, 1956), but is continuous
in rats, the sinusoids penetrating the continuum (Pauly, 1957). The latter
normally contain large amounts of blood circulating from the arteries in
the capsule to the medullary venous sinusoids and adrenal vein. There are
variations in the circulatory arrangement with species, and it is more com-
plex in detail than has been described here, but these matters are thoroughly
covered elsewhere (Gersh and Grollman, 1941; Hartman andBrownell,
1949; Harrison, 1951, 1957; Elias and Pauly, 1956; Pauly, 1957). The cells
of the fasciculata, when stamed by routine procedures, bear a marked
resemblance to the luteal cells of the ovary, interstitial cells of the testis,
and, although less closely, the cells of the "brown fat" or "hibernating
gland" in its usual functional state. The zona reticularis forms a fairly wide
cortical band between the medulla and the zona fasciculata in most species,
but it is not always present (Hartman and Brownell, 1949). Its cords (or
cortical continuum) are more or less continuous peripherally with those of
the zona fasciculata, but they rapidly break up into a reticular network as
they proceed centrally toward the medulla. The cells are generally smaller
than other cortical cells and usually contain no vacuoles. However, when
vacuoles are present, they are usually very large.

There is need for a detailed, well illustrated, and thorough discussion of
the comparative morphology of the adrenal glands which would include a
wide variety of species and a sufficient number of animals of each species
to describe age and sex, as well as seasonal and environmental, relation-
ships. It is not the purpose of the present discussion to dwell on the anatomy

2. Endocrines and Populations 195

of the adrenal glands, but a brief summary of the particularly useful and
more recent publications on the subject will be given. The books which
already have been listed discuss the adrenals of a large number of species,
but the descriptions and illustrations are limited. However, the histology
and histochemistry of the adrenals of a few species are discussed in con-
siderable detail in a number of papers. The recent publications of Elias and
Pauly (1956) and Pauly (1957) describe the microscopic anatomy of the
adrenal glands of laboratory rats and humans. These papers are well
illustrated, and the stereographic reconstructions of serial sections are
helpful in understanding the adrenal morphology of these two species. One
of the important facts brought out in these papers is that the adrenal cortex
of the rat is not arranged in cords as it is in humans and monkeys. The
parenchyma of the rat adrenal cortex is a continuum which is tunneled by
vascular channels. A number of additional papers deal with the anatomy,
circulation, or histochemistry of the adrenals of laboratory rats, especially
with regard to function, zonation, and reactions to various stimuli (Howard,
1938; Flexner and Grollman, 1939; Greep and Doane, 1947, 1949a; Deane
et al., 1948; Deane and Morse, 1948; Cain and Harrison, 1950; Feldman,
1950, 1951; Cater and Stack-Dvmne, 1953, 1955; Josimovich et al, 1954;
Jones and Spalding, 1954; Jones and Wright, 1954a, b; Christianson and
Jones, 1957), and other more general papers on the histochemistry and
function of the adrenals are based largely on material from laboratory rats
(Dempsey, 1948; Greep and Deane, 1949a, b; Sayers and Sayers, 1949).
The differences in morphology between the adrenals of wild rats {Rattus
norvegicus and Rattus alexandrinus) and those of Norway rats from the
laboratory have been described by Rogers and Richter (1948), and the
histology of wild and laboratory Norway rats has been described and com-
pared by Mosier (1957). A comparative study of the vascularization of
the adrenals of rabbits, rats, and cats has been made by Harrison (1951)
and followed by a description of the adrenal circulation and its regulation
in the laboratory rabbit (Onjctolagus) (Harrison, 1957).

The histology of the adrenal glands of the prototherians Onuthorlnjnchus
and Tachyglossus has been described in considerable detail by Wright et al.
(1957). The bulk of the chromaffin tissue was found in the lower pole in
these species rather than in the more usual central position. The cortices of
these species also differ considerably in their histologic appearance from
those of eutherians. We have mentioned above that the adrenals of North
American soricids have strikingly little cortical tissue, although a critical
study of this material has not been made (J. J. Christian, unpublished) .
Lanman (1957) has described the fetal zones of the adrenals of the fol-
lowing fetal or neonatal primates: macques (Macaca midatta) , potto
{Perodicus putto) , chimpanzee (Pan sp.), hybrids of Cercopithecus {Cerco-

196 /• J- Christian

pithecus sp.), marmoset {Callothrix argentata) , slow loris (Loris sp.), colo-
bus monkey {Colohus polykomos), and humans. The anatomy of the
adrenals of the macaque has been described by Harrison and Asling (1955) .
Variations in the histochemistry of the adrenals of cows, rats, and monkeys
followmg "stress" or treatment with adrenocorticotropin, cortisone, or
deoxycorticosterone were the subject of a paper by Glick and Ochs (1955) .
Additional descriptions of the adrenals of cows and other domesticated
ungulates are subjects of papers by Elias (1948) and Weber et al., (1950).
Finally, Zalesky (1934) described in considerable detail the seasonal
histologic changes in the adrenals of thirteen-lined ground squirrels ( Citel-
lus tridecemlineatus) .

The morphology and histochemistry of the adrenals of laboratory and
wild house mice have been thoroughly studied, largely because of the
endocrine relationships of the transitorj^ X-zone which was first described
by Howard (1927). Tamura (1926) wrote a detailed description of the
changes during pregnancy in the adrenals of mice. This was followed by
Howard's (1927) description of the X-zone and Waring's (1935) descrip-
tion of the development of the adrenal glands of the mouse. Following
these there was a spate of papers describing the X-zone and its reactions
to various hormones and experimental treatments (Gersh andGrollman,
1939; Waring, 1942; McPhail and Read, 1942a, b; McPhail, 1944; Jones,
1948, 1949a, b, 1950, 1952; Miller, 1949; Benua and Howard, 1950; Howard
and Benua, 1950; Jones and Roby, 1954; Allen, 1954; Allen, 1957). The
histology and histophysiology of the adrenals of hamsters (Mesocricetus
auratus) have been described by Alpert (1950) and Holmes (1955), and
the effects of hypophysectomy and starvation on their adrenals by Knigge
(1954a, b).

The adrenals of a number of species of European small mammals have
been studied and described by Delost, particular attention being paid to
the presence or absence of an X-zone and its relationships to the sexual
cycle and sex accessories. The mammals in these studies included Microtus
arvalis (Delost 1951; 1952a, b; 1954; 1956a, b) Microtus agrestis (Delost
and Delost, 1955), Clethrionomijs glareolus (Delost and Delost, 1954),
Pitymys (Delost and Delost, 1955), Sorex araneus (Delost, 1957), and
Crocidura (Delost, 1957).

Immature male and young nuUiparous female house mice (Mus muscu-
lus) have an adrenocortical juxtamedullary zone, the X-zone, which un-
equivocally shows sex relationships (Howard, 1927; Deanesly, 1928; Jones,
1957) . This zone is absent from mature male and parous or old females.
Cortical X-zones have been described for a number of other species of
small mammals including meadow voles (Microtus agrestis and Microtus
arvalis), red-backed voles (Clcthrionomys glareolus, pine voles (Pitymys

2. Endocrines and Populations 197

suhterraneiis) , and shrews (Sorex araneus and Crocidura russula) (Delost,
1951, 1952a, 1954, 1957; Delost and Delost, 1954, 1955), but there is some
question whether the X-zones of these species are entirely analogous to the
X-zones of the house mouse. Delost (1954, 1956b) reports that cortisone
involutes the so-called X-zone of voles, which is a response not seen in
house mice. The X-zone of the shrew behaves like that of the house mouse
with respect to its involution, but apparently has not been subjected to
critical experiments in the laboratory (Delost, 1957). The X-zone consists
of cords of small, deeply acidophilic cells with intensely basophilic nuclei
and which are about one-half the size of those of the zona fasciculata
(Howard, 1927; Deanesly, 1928; Jones, 1949a, b, 1950, 1957; Benua and
Howard, 1950) . The cytoplasm of these cells, besides being more acido-
philic than those of the fascicular cells, is unvacuolated ordinarily and
lacks the sudanophilia of the other zones of the adrenal cortex (Jones,
1957) . Criteria for critically distinguishing the X-zone have been reviewed
by Benua and Howard (1950) and Holmes (1955). The uniqueness of this
zone rests on the fact that it is involuted by androgens and appears to de-
pend on pituitary luteinizing hormone for its maintenance (Howard, 1927,
1959; McPhail and Read, 1942a, b; McPhail, 1944; Waring, 1942; Jones,
1949a, b, 1950, 1952, 1957). The function of this zone, if there is a specific
function, is unknown. The so-called X-zone of voles which has been de-
scribed by Delost (1951, 1952a, 1954;) and Delost and Delost (1954, 1955)
reappears after castration or after the hibernal periods of sexual inactivity
in the males and persists through gestation and lactation in the females,
and in these respects it differs markedly from the X-zone of house mice.
This zone may confound the use of adrenal weight as an index of increased
cortical activity in the house mouse (Christian, 1956) and other species
which possess it, but it provides a useful measurement for determining
histologically the onset of androgen production, therefore puberty, in male
house mice (Christian, 1956). A poorly defined X-zone has been described
in mature nulliparous female hamsters, but not in males (Holmes, 1955),
differing from the X-zone of house mice in this respect. It is likely that an
X-zone will be described for other species when enough material from all
age groups of both sexes has been critically examined, and that a variety of
manifestations of this zone will be found.

The morphology and size of the adrenal cortex varies with its functional
status (see also discussion under reticularis) . The cortex undergoes rapid
hyperplasia and hypertrophy in response to stimulation by adrenocorti-
cotropin (ACTH) from the anterior pituitary. At first there is a rapid
diminution in the size and number of lipid vacuoles, ascorbic acid, and
cholesterol of the cortical cells (Sayers and Sayers, 1949). The vacuoles
soon increase in number and size, providng the stimulation is not too

198 ./. J. Christian

severe (Dempsey, 1948; Sayers and Sayers, 1949; Greep and Deane,
1949b) . The lipid vacuoles disappear first from the fasciculata next to the
reticularis, so that it becomes indistinguishable from the latter. As stimula-
tion continues, the disappearance continues centrifugally, and at the same
time enzymes normally absent from the fasciculata, but present in the
reticularis, make their appearance in the cells of the fasciculata, the inner-
most portion moving outward (Symington et at., 1958). Upon withdrawal
of the stimulus of ACTH the lipid vacuoles increase considerably in size
and may become very large. This stage presumably represents lipid storage.
The cellular hyperplasia and hypertrophy mainly account for increases in
the size and weight of the adrenal glands. Initially the cortex responds to
stimulation with a marked decline in its cholesterol, neutral lipids, and
ascorbic acid content (Greep and Deane, 1949b; Sayers and Sayers, 1949).
These soon return at least partially to their original state, and in the inac-
tive gland they may exceed their original levels. These matters are dis-
cussed in detail in the cited references in addition to discussions therein of
the relationships of the cortex and its activity to various stimuli for varying
lengths of time and with varying intensity.

c. The adrenal medulla. The adrenal medulla consists of rather irregular
masses of polyhedral chromaffin cells derived, along with the ganglia of
the sympathetic nervous system, from the primitive neuroectoderm. The
medulla is homologous with the sympathetic ganglia and receives myeli-
nated cholinergic preganglionic fibers from the greater splanchnic nerve.
The medulla itself serves as the ganglion and the postganglionic tracts.
There apparently are several types of cells in the medulla; these are dis-
cussed in more detail elsewhere (Hartman and Brownell, 1949; Eranko and
Raisanen, 1957) . The cytoplasm of the medullary cells contains numerous
minute deeply basophilic granules which stain blue with ferric chloride
and brown with chromic acid (chromaffin) and which appear in some way
to be related to secretory function.

The adrenal medulla generally is not thought to hypertrophy following
stimulation in the same way that the adrenal cortex does. Rogers and
Richter (1948) reported the absence of medullary hypertrophy with
changes in adrenal size in rats. However, there is good evidence that the
medulla does hypertrophy, at least in some species and under some circum-
stances, even though it may not contribute significantly to an increase in
the total weight of the gland, as a consideration of its geometry will show.
House mice have been shown to exhibit a marked medullary hyperplasia
and hypertrophy during pregnancy (Tamura, 1926) or chronic stimulation
due to crowding (Bullough, 1952). Medullary hypertrophy also has been
observed in a variety of species of captive wild ungulates subjected to
conditions in a zoological garden similar to the crowding of mice reported

2. Endocrines and Populations 199

by Biillough ( 19o2) (Ratcliffe, unpublished; J. J. Christian, unpubUshed) .
These conditions which resulted in which medullary hypertrophy all
constituted prolonged, chronic stimuli. Medullary hypertrophy due to
hyperplasia probably occurs simultaneously with cortical hypertrophy in
many species but perhaps requires a more sustained stimulus and develops
at a much slower rate. There also seems to be some suggestion that emo-
tional stimuli may be important in this effect. Finally, it has been shown
that treatment with pituitary growth hormone will produce a marked
hypertrophy of the adrenal medulla (Moon et at., 1951; Lostroh and Li,
1958) , and may eventually result in medullary tumors (Moon et at., 1950) .
The role of sympathicomedullary function in physiologic adaptation re-
quires more investigation, especially in regard to chronic stimulation, such
as is produced by sociopsychologic pressures, and for a variety of species.

2. Hormones Secreted by the Adrenal Cortex: Their Actions and
THE Regulation of Their Secretion.

a. The zona glomerulosa. (1) The hormones. The adrenal cortex secretes
two steroid hormones, aldosterone (18-aldocorticosterone) and deoxycorti-
costerone (11-deoxycorticosterone), which have their primary effects on
salt-electrolyte and water metabolism. However, aldosterone is the only bio-
logically important sodium-retaining corticoid secreted by the adrenal cor-
tex and it is many times more powerful than deoxycorticosterone in its
effects on electrolyte metabolism (Farrell et al., 1955; Gaunt et at., 1955;
Gross and Lichtlen, 1958) . Also aldosterone is an important secretory pro-
duct of the adrenal cortex, whereas deoxycorticosterone is secreted only in
trace amounts (Farrell et at., 1955; Jones, 1957) and is probably a precursor
in the formation of aldosterone (Giroud et al., 1958). The actions of these
two hormones are very similar within the physiological range of dosages for
each, but their actions with overdosage differ considerably: overdosage
with aldosterone does not lead to the excessive sodium retention and the
diabetes insipidus-like state which are seen after overdosage with deoxy-
corticosterone (Gross and Lichtlen, 1958).

It is appropriate at this point to comment on the general classification of
the adrenal corticoids into the two broad categories which are used in the
present account. The hormones of the adrenal cortex have been loosely
grouped as "sodium-retaining" or ''carbohydrate-active" according to
whether their prunary actions are on salt-electrolyte metabolism or if they
are among those steroids having marked effects on carbohydrate meta-
bolism. The sodium-retaining steroids include aldosterone, deoxycorti-
costerone, and, to a much lesser extent, 17-hydroxy-ll-deoxycorticosterone
(Reichstein's compound S) . The principal carbohydrate-active steroids are

200 J. J. Christian

hydrocortisone and cortisone (ll-oxy-17-hydroxycorticoids). Corticoste-
rone is included in this latter group, although it has moderate effects on
both salt-electrolyte and carbohydrate metabolism. It is considerably
weaker in all these actions than the principal corticoids in either of the
categories (Jones, 1957). This classification into primarily carbohydrate-
active and primarily sodium-retaining corticoids is useful, but by no means
does it reflect the entire spectrum of activities of the hormones; in many
instances there is a considerable overlap in these activities for a particular
hormone, for example, corticosterone.

Recent morphologic and direct evidence shows that the secretion of
aldosterone is a function of the zona glomerulosa, whereas the carbohy-
drate-active corticoids, except corticosterone, and probably the C19 steroids
are secreted by the zona fasciculata and zona reticularis. Probably the
most conclusive evidence for the relationship between specific secretory
function and zonation of the adrenal cortex has been provided by the in
vitro incubation and determination of the secretory products of selected
segments of the adrenal cortex. Aldosterone was found to be secreted only
by incubated portions of the zona glomerulosa of the adrenals of rats and
beef cattle (Ayres et al., 1956; Giroud et al., 1956; Giroud et at., 1958);
hydrocortisone was produced only by the zonae fasciculata-reticularis, and
corticosterone was produced at approximately equal rates by all three
zones of the adrenals of beef cattle (Ayres et al., 1956; Giroud ct al., 1958
Stachenko and Giroud, 1959a, b) . It was also shown in these experiments
that ACTH or corticotropin peptides or other steroids were without effect
on the production of aldosterone by the zona glomerulosa but that they
markedly increased the production of total corticosteroids and of hydro-
cortisone by the fasciculata-reticularis (Stachenko and Giroud, 1959b).
Additional evidence of functional zonation, less direct, has been obtained
by relating changes in the composition of the secretory product with mor-
phologic changes in the various zones of the adrenal cortex. A sodimn-
deficient diet produces extreme hypertrophy of the zona glomerulosa and
atrophy of the zona fasciculata of the adrenal cortices of rats (Hartroft and
Eisenstein, 1957), and these changes are associated with a marked increase
in the secretion of aldosterone and decreases in the secretion of corticoste-
rone (Eisenstein and Hartroft, 1957). In somewhat comparable experi-
ments it was shown that (1) sodium deprivation markedly increased the
aliesterase activity of the zona glomerulosa but had no effect on its activity
in the fasciculata of the adrenals of mice; (2) deoxycorticosterone or sodium
flooding depressed the aliesterase activity of the zona glomerulosa and
increased it in the zona fasciculata; and (3) injected ACTH markedly
increased the aliesterase activity of the zona fasciculata but did not affect
it in the zona glomerulosa, whereas blocking ACTH secretion with cortisone

2. Endocrines and Populations 201

depressed the fascicular aliesterase activity (Allen, 1957). The results of
the foregoing experiments provide convincing evidence in support of the
hypothesis of Greep and his co-workers that the zona glomerulosa secretes
the electrolyte-active, and the fasciculata the carbohydrate-active, corti-
coids (Greep and Deane, 1947, 1949b; Deane et at., 1948). This hypothesis
was based on observations that (1) increased sodium intake or injections
of deoxycorticosterone produced histochemical changes indicative of de-
creased activity in the zona glomerulosa of the adrenals of rats, and that
(2) a reduction in sodium or increase in potassium produced cytological
changes indicative of increased activity of the zona glomerulosa. There is
little doubt that the glomerulosa is responsible primarily for the secretion of
aldosterone and that the carbohydrate-active corticoids are secreted by
the zona fasciculata and possibly by the zona reticularis. Convincing evi-
dence of a functional separation between the zonae fasciculata and reticu-
laris is not available, but the reticularis generally is not believed to be as
active a secretory zone as the fasciculata.

The chief action of aldosterone is on sodium-potassium transport in the
tubular cells of the renal nephron, and it is relatively more effective in pro-
moting sodium retention than in promoting potassium excretion or water
retention (Gaunt et al, 1955; Bartter, 1956; Jones, 1957; Gross and
Lichtlen, 1958; Stanbury et al., 1958). It apparently stimulates the ionic
exchange between potassium and sodium ions in the renal tubular cells
(Bartter, 1956; Stanbury et al., 1958), although an overdosage of aldoste-
rone will not produce excessive sodium retention and the animal therefore
stays in sodium balance (Bartter, 1956; Gross and Lichtlen, 1958). In
general the sodium-retaining corticoids act on the nephric tubular cells to
promote an ionic exchange between sodium and potassium; so that sodium
is retained and potassium is excreted (Bartter, 1956, 1957). Water is re-
absorbed with the sodium or independently under the action of neurohy-
pophyseal antidiuretic hormone (ADH) (Bartter, 1957). Proper fluid and
electrolyte balance is mamtauied by these homeostatic endocrine activities
acting in concert with water and salt intake and with hemodynamic and
neural factors which affect fluid volume, blood pressure, and renal glomeru-
lar filtration. The apparent anomaly of overdosages of aldosterone failing
to produce excessive sodium retention depends on the fact that the blood
pressure is raised and therefore the glomerular filtration rate is increased
and sodium is lost accordingly (Stanbury et al, 1958) . Proper fluid and
electrolyte balance is vital to any animal, and adrenalectomized animals
can be maintained with injected deoxycorticosterone or aldosterone, al-
though they cannot adapt to added stress (Gaunt et al, 1955) . The adrenal-
ectomized laboratory rat or mouse also can be maintained alive by sup-
plying W( sodium chloride in its drinking water to replace the sodium loss

202 J. J. Christian

accompanying adrenalectomy. However, adrenalectomized wild Norway
rats (Rattus norvegicus) cannot be maintained in this fashion, even with
the NaCl content of the drinking water as high as 4 %(Richter et al., 1950) .
These facts emphasize the wide divergence between laboratory and wild
strains of the same species. Evidently the requirements for adrenocortical
hormones are much greater in mammals under feral conditions than for
those raised or maintained in the laboratory or zoo. There is a marked
disparity in the adrenal weights of mammals raised in the laboratory and
in the same species under natural conditions, the differences due largely to
differences in the amount of cortical tissue (Rogers and Richter, 1948;
Nichols, 1950; Christian and Ratcliffe, 1952; Christian, 1955a) ; some of this
difference, however, may be associated with the unconscious selection in
breeding colonies for docility and good breeding performance.

(2) Regulation of aldosterone secretion} Since aldosterone acts pri-
marily to maintain fluid and electrolyte homeostasis, it is not surprising
that the secretion of this hormone is regulated largely by these factors.
Changes in the volume of extracellular fluid (probably mainly the intra-
vascular volume) , and the level of body potassium affect the rate of aldo-
sterone secretion (Liddle et al., 1956; Bartter, 1957; Bartter et at., 1959),
but to some extent the secretion of aldosterone in vivo can be stimulated
by adrenocorticotropin (Farrell et al., 1955; 1958; Liddle et al., 1956), but
apparently not in vitro (Stachenko and Giroud, 19596). This discrepancy
may be explained by the increased production of the precursors of aldoste-
rone by the f asciculata which then become accessible to the zona glomerulosa
in the intact adrenal. Even though the secretion of aldosterone is mode-
rately stimulated by ACTH, the stimulation is not maintained in spite of
continued treatment with ACTH (Liddle et al., 1956), and the response is
considerably less than that seen following changes in the volume of extra-
cellular fluid or in body potassium (Bartter et al., 1959). The glomerulosa
will respond to increased ACTH with increased secretion of aldosterone for
only about 3 or 4 days, and then the rate of secretion declines in spite of
continued ACTH and reaches base levels or even lower levels of secretion
in about a week (Liddle et al., 1956). After this period, continued ACTH
will not increase the secretion of aldosterone (Bartter et al., 1959). Finally,
the secretion of aldosterone is only slightly depressed by suppressing ACTH
secretion (Farrell et al., 1955; Liddle et al., 1956; Bartter, 1957) or by hy-

' Since completion of this chapter, there has been marked progresses in understanding
the regulation of aldosterone secretion in response to hemodynamic changes. It is fairly
certain that in response to decreased arterial pressure there is increased relea.se of renin
from the kidney. The end product of this release is angiotensin II which, in the presence
of basal levels of ACTH, stimulates aldosterone secretion. [For a review see J. (). Davis
(1963).]

2. Endocrines and Populations 203

pophysectomy (Farrell ei al., 1955). These results, in addition to those
discussed under the relationship between secretory function and zonation
of the adrenal cortex, clearly indicate that the secretion of aldosterone is
largely independent of adrenocorticotropin and the adenohypophysis and
that its responses to ACTH may reflect the increased availability of aldoste-
rone precursors from other parts of the cortex. Nevertheless, the secretion
of aldosterone in dogs appears to be dependent on an intact pituitary gland
(Davis, et al., 1959a). However, as one might expect, a variety of stimuli
which produce an increase in the secretion of ACTH may also stimulate
an increase in the secretion of aldosterone. Farrell (1958) lists position,
surgery, emotional factors, hypertension, insulin shock, and other stimuli
among those resulting in an increased secretion of aldosterone, but prob-
ably none of these are without an effect on fluid and electrolyte balances
which in turn would effect directly the mechanisms regulating the secretion
of aldosterone. On the other hand, there is a marked increase in the secre-
tion of aldosterone in those diseases which are characterized by striking
disturbances in fluid and electrolyte metabolism, such as congestive heart
failure, hepatic cirrhosis, and nephrosis (Liddle et al, 1956) . It seems likely
that the increase in aldosterone secretion is slight in those circumstances
which produce a marked increase in the secretion of ACTH and of the
carbohydrate-active corticoids unless there is also involvement of fluid and
electrolyte balances. It has been found that only one, the A-1 fraction, of the
several distinct fractions of ACTH has an appreciable effect on the secretion
of aldosterone, and this fraction is a relatively small proportion of the total
amount of ACTH which may be secreted (Farrell et al, 1958; Farrell,
1959a) .

The principal regulation of aldosterone secretion seems to be by a com-
bination of neural and neurohumoral factors in response to changes in the
volume of extracellular fluid or body potassium. However, there can be
little doubt that a hormonal factor is involved in aldosterone secretion, as
recently demonstrated with cross-circulation experiments by Yankopoulos
et al (1959) . Recent experiments have indicated that the brain may secrete
a hormone, glomerulotropin, not as yet isolated and characterized, from the
region of the pineal body which stimulates the secretion of aldosterone
from the adrenal zona glomerulosa (Farrell, 1959a) .

Small changes in blood volume can effect striking changes in the rate of
secretion of aldosterone (Bartter, 1957) possibly by affecting changes in
pulse pressure (Bartter and Gann, 1960). Changes in blood volume elicit
maximal reciprocal responses in the secretion of aldosterone and it appears
that this system is the most sensitive, as well as the most important, of
those involved in the regulation of the secretion of aldosterone (Bartter,
1957; Bartter et al, 1959). A rise in blood volume reflexly depresses the

204 /. /. Christian

secretion of aldosterone via stretch receptors in the region of the right
striiim or adjacent vena cava (Davis ct al., 1956, 1957, 1958; Liddle ct al.,
1956; Bartter et al, 1958, 1959; Farrell, 1959a; Anderson et al, 1959), the
vagus nerve (Mills et al, 1958), and central pathways possibly to depress
the secretion of glomerulotropin from the pineal region of the brain al-
through Davis et al (1959b) indicated that the vagus is not involved in the
afferent pathways of this control. Conversely, a decrease in blood volume
stimulates the secretion of aldosterone (Bartter et al, 1959), although the
exact pathways and mechanism by which this is achieved is unknown.
Bartter and Gann (1960) have suggested that pulse pressure is a factor in
changes in blood volume which affects aldosterone secretion. A drop in
pulse pressure stimulates the release of aldosterone and a rise inhibits its
release. These changes apparently come about through changes in the rate
of tonic impulses over receptor nerves in the region of the thyrocarotid
artery.

Another system that regulates the secretion of aldosterone involves the
levels of potassium in the body. A deficiency of potassium, therefore a
lowered concentration of body potassium, results in a lower rate of secretion
of aldosterone if it was originally elevated, whereas an increase in body
potassium results in an increase in the secretion of aldosterone (Bartter,
1956; Bartter et al, 1959). A rise in serum potassium, either absolute or
relative to the concentration of sodium, is associated with an increase in the
secretion of aldosterone, but it is not known whether a fall m potassium
actively inhibits its secretion or permits it to return to base levels passively
(Farrell, 1958) . It has been shown that these changes in the rate of secre-
tion of aldosterone in response to changes in body potassium are inde-
pendent of sodium concentration in the serum or the total amount of so-
dium in the body and are also independent of the sodium: potassium ratio
in the serum (Bartter, 1956; 1957; Bartter et al, 1959). Similarly, there is
no evidence to suggest that altered renal hemodynamics are responsible for
the altered secretory rates of aldosterone (Bartter et al, 1956; Cole, 1957).
It is not known yet whether the regulation of the secretion of aldosterone by
the body potassium is directly on the cells of the adrenal zona glomerulosa
or is mediated through central channels (Bartter, 1956; Bartter ct al, 1959) .
It cannot be said whether serum potassium, intracellular potassium, or a
combination of both effects the control of the secretion of aldosterone, but
there is evidence that the adrenal cortical cells themselves may respond
directly to this type of stimulus (Bartter, 1956) . On the other hand, Farrell
(1958) suggests that the effect is through central channels. However,
changes in potassium are not as important in the regulation of the secretion
of aldosterone as changes in the volume of the extracellular fluid (Bartter,
1957; Bartter efaL, 1959).

2. Endocrines and Populations 205

In summary, three mechanisms are involved in regulating the secretion
of aldosterone. The first and most important regulating factor is the volume
of the extracellular, probably intravascular, fluid, changes in which act
through atrial stretch receptors and other as yet unknown pathways to
effect reciprocal changes in the rate of secretion of aldosterone. Decreased
pulse pressure also stimulates increased aldosterone secretion and may be
one way in which changes in blood volume act. Depression of the secretion
of aldosterone by increases in blood volume requires an intact vagus nerve.
The second mechanism responds to changes in body potassium; a rise in
potassium resulting in elevation of the rate of secretion of aldosterone and a
fall in potassium permit the secretion of aldosterone to fall back to normal.
Finally, adrenocorticotropin, or at least a fraction thereof, is capable of
stimulating the secretion of aldosterone in the intact animal, but only to a
moderate degree and for a relatively short period of time, although the
secretion of aldosterone or its regulation and the functional integrity of
the adrenal zona glomerulosa apparently do not depend upon adrenocorti-
cotropin. Glomerulotropin, a recently described hormone from the pineal
complex region of the brain which stimulates the secretion of aldosterone,
may be an important link in the regulating system depending on the volume
of the extracellular fluid or body potassium or both, but this work requires
confirmation.

The actions of aldosterone are essential in combatting incipient shock
mammals, and this hormone apparently plays a vital role in the daily
maintenance of fluid and electrolyte homeostasis. Aldosterone also may be
more directly responsible for maintaining blood pressure and counteracting
hemoconcentration through its activity in correcting alterations in blood
volume.

h. The zona fasciculata. (1) The hormones. This zone of the adrenal
cortex normally secretes hydrocortisone (Kendall's compound F), corti-
costerone (Kendall's compound B), small amounts of cortisone (Kendall's
compound E), 11-deoxycorticosterone (Kendall's compound A), 11-de-
oxycorticosterone (DOC, DCA, or DOCA), ll-deoxy-17-hydrocorti-
costerone (Reichstein's compound S), and C19 ketosteroids, usually andro-
genic, the amounts and proportions depending on the species and the circum-
stances. Although modification of this concept is required in the light of the
work of Symington and his co-Avorkers (1958) (c/. above). Their experi-
ments indicate that the reticularis is the part of the cortex that normally
produces corticoids and 17-ketosteroids at rest, and that the fasciculata be-
comes functional with increased stimulation. In other words, the reticularis
is the active part of the cortex and the fasciculata is a resting portion.
Actually this work indicates that the morphologic separation of the cortex
into fasciculata and reticularis is unjustified. In addition to aldosterone, the

206 /• /• Christian

normally important adrenocortical hormones are corticosterone and hydro-
cortisone, and their respective ratios vary from species to species (Bush,
1953; Nelson, 1955) and possibly with the degree of cortical stimulation
(Bradlow and Gallagher, 1957). The ratio of hydrocortisone to corticoste-
rone (F:B ratio) may vary from less than 0.05 in rats and rabbits to greater
than 20 in monkeys (Bush, 1953; Reif and Longwell, 1958; Dorfman,
1959) . Most species lie between these two extremes (Bush, 1953) . However,
there is little doubt that in most species studied, exclusive of rats and mice,
these two steroids form from 80% to 95% of the total adrenal secretion of
corticoids (Jones, 1957). Corticosterone is the principal carbohydrate -
active corticoid secreted by mice, rats and, rabbits (Bush, 1953; Hofmann,
1956; 1957; Reif and Longwell, 1958; Wilson et al, Bloch and Cohen, 1960) .
whereas hydrocortisone is the principal corticoid in other species, including
guinea pigs, hamsters, ferrets, cats, monkeys, sheep, and humans (Bush,
1953; Nelson, 1955; Jones, 1957; Peron and Dorfman, 1958, Schindler and
Knigge, 1959a, b). The adrenals of house mice and rats apparently secrete
large amounts of 1 l-hydroxy-**-androstene-3 , 17-dione ( 1 1-0H4AD) 1 1-hy-
droxyandrostene-3,17-dione, and other closely related steroids as major
components of their natural adrenal secretory product in addition to
corticosterone and very small amounts of hydrocortisone and other corti-
coids (Sweat and Farrell, 1952; Bush, 1953; Hofmann, 1956; Bahn et al.,
1957; Poore and Hollander 1957; Wilson et al., 1958 Bloch and Cohen
1960). Probably all species secrete 11-0H4AD and closely related Cig
steroids, but usually in proportionately small amounts (Bradlow and Gal-
lagher, 1957; Gallagher, 1958). However, it has been shown recently that
the adrenal androgen, dehydroepiandrosterone, comprises about 50% of
the total secretion of steroids by the human adrenal cortex (Vande Wiele
and Lieberman, 1960) . The general problem of the secretion of sex steroids
by the adrenal cortex has not been studied until recently in the same
detail as the carbohydrate-active corticoids, especially with regard to
differences among species, but there is no doubt that they are secreted b}^
the cortex (Gallagher, 1958) . These Cig steroids may be normal products,
metabolites, or intermediate metabolites in the synthesis of other steroids
(Dorfman and Shipley, 1956; Gallagher, 1958). Apparently there is con-
siderable variation with species with respect to the secretion of androgens
and androgen precursors (Bush, 1953; Jones, 1957; Gallagher, 1958; Wilson
et al., 1958) . In any event, the adrenal cortex is the starting point of C19
steroids which may act as weak androgens (Dorfman and Shipley, 1956;
Gallagher, 1958) .

Normally cortisone, hydrocortisone, and corticosterone appear in the
urine as metabolites which can be identified and related to the parent corti-
coid by the appropriate procedures (Gallagher, 1958; Dorfman, 1960).

2. Endocrines and Populations 207

However, there are other metaboHtes in the urine in smaller (juantities
which cannot be related specifically to particular adrenocorticoids without
radioactive labeling, but these ordinarily are not produced in appreciable
ciuantities (Gallagher, 1958). The types and quantities of the metabolites
of a particular hormone which appear in the urine usually provide a good
index of adrenocortical activity for relatively longer periods of time than
can be obtained by the measurement of the corticoids in the plasma, which
only reflect the immediate situation (Nelson, 1955). However, the urinary
metabolites do not always reflect the actual adrenal secretory pattern, as
has been shown for mice (Bradlow et at. 1954; Wilson ct al., 1958) although
in many instances this may be surmised with confidence (Dorfman, 1960).
In summary it may be said that hydrocortisone or corticosterone and C19
weak androgens are the major secretory components of the adrenal fascicu-
lata, but that other carbohydrate-active corticoids, sodium-retaining corti-
coids, and adrenal androgens are also secreted, although usually not in
appreciable quantities.

It is impossible to make hard and fast statements about the quantitative
relationships of the adrenocortical hormones to one another because of a
certain degree of inherent variability and because the techniciues for their
measurement are not sufficiently refined and certain for such detailed
comparisons. A large number of steroids have been isolated from the adre-
nals of various species, frequently from perfused glands. Some of these may
be biochemical artifacts, but many are probably intermediate products in
the biosynthesis of the normal secretory products, or possibly steroids which
are secreted only under unusual conditions (Jones, 1957; Bradlow and Gal-
lagher, 1957; Gallagher, 1958) . It is not known whether or to what extent,
some of these steroids are secreted naturally. The picture is complicated
further by the fact that the liver and other tissues metabolize the steroid
hormones to new steroids which appear in the circulation and urine and
which may have biological activity to varying degree (Gallagher, 1958).
Therefore, the specific roles of the various adrenocortical steroids and their
metabolites, especially those that appear in very low concentrations, in the
economy of the w^hole mammal, and the variations in their secretory pat-
terns from species to species and under normal and abnormal circumstances,
needs clarification. Some of the discrepancies that appear in the literature
regarding the relative amounts of various steroids secreted by the adrenals
of a particular species seem to depend on whether the measurements were
made in vivo or on perfusates of isolated glands (Bush, 1953; Jones, 1957).
It seems evident that appreciably higher proportions of steroids which
normally are secreted in low concentrations are found in perfusates than in
vivo. However, it suffices for the present to know that the carbohydrate-
active corticoids, hydrocortisone and corticosterone, are the major na-

208 /. /. Christian

turally secreted corticoids of the zona fasciculata and to discuss the actions
of these hormones as a class.

(2) Adiojis of the fascicular hormones. Hydrocortisone, cortisone,
and corticosterone have important effects on carbohydrate metaboHsm
and therefore are classed loosely as carbohydrate-active corticoids. They
have in common either a hydroxyl group or ketonic oxygen at the carbon-11
position, and those with the most pronounced effects on carbohydrate
metabolism, hydrocortisone and cortisone, have a hydroxyl group on the
C-17 of the steroid nucleus. Corticosterone has a weaker action on carbohy-
drate metabolism than hydrocortisone or cortisone (Dorfman, 1949; Ingle,
1950; Parmer et al., 1951; Santisteban and Dougherty, 1954; Dougherty
and Schneebeli, 1955; Kass et al., 1955; Noble, 1955) , but it has appreciably
more effect on salt-electrolyte metabolism than either of the others (Noble,
1955; Farrell et al., 1955; Jones, 1957) . Because of these facts, the relatively
small amounts of hydrocortisone which are normally secreted by the adre-
nals of mice and rats, along with corticosterone, have been held responsible
for most of the carbohydrate-active corticoid activity, such as involution
of the thymus, which has been observed in these animals (Wilson etal.,
1958). The designation "carbolwdrate-active corticoids" for this group of
steroids by no means reflects all their activities. These corticoids have sup-
pressive effects to varying degrees on inflammation and therefore are
classified also as anti-inflammatory (antiphlogistic) hormones (Selye, 1950;
Dougherty 1953) . As a class they have profound effects on protein metabo-
lism, fat metabolism, growth, oxygen consumption, and a number of other
physiological functions (Noble, 1955; Jones, 1957). Hydrocortisone and
cortisone are the most powerful of the fascicular carbohydrate-active
corticoids and corticosterone the least powerful with respect to the enume-
rated activities (cf. above), although cortisone is not produced in bio-
logically important quantities in any of the species so far investigated
(Bush, 1953; Nelson, 1955). As a general rule, the degree of activity of a
corticoid on carbohydrate metabolism is related inversely to its sodium-re-
taining ability. Finally, it should be noted that other steroids may affect the
actions of the corticoids; for example, testosterone and estradiol potentiate
the anti-inflammatory action of the carbohydrate-active corticoids (Tauben-
haus, 1953), and testosterone enhances the thymolytic activity of cortisone
(Selye, 1955; Dorfman and Shipley, 1956) .

The carbohydrate-active corticoids, secreted by the zona fasciculata, will
maintain life in adrenalectomized mammals (Ingle, 1950) ; nevertheless the
exact functions of the adrenocortical hormones in the intact normal animal
are difficult to delineate precisely, as these hormones are integral elements
in a complex system of endocrine and neural responses which form a feed-

2. Endocrines and Populations 209

back system to maintain homeostasis or to meet emergencies. Nevertheless,
much has been learned about the specific activities of these hormones by
the classic experimental approaches to such a problem : the substitution of
pure hormones or extracts into intact and adrenalectomized animals and
refinements of these procedures. The effects of injected carbohydrate-active
corticoids are closely paralleled by those produced by injecting adreno-
corticotropin (ACTH), the hormonal protein of the anterior pituitary
responsible for stimulating the secretion of the carbohydrate-active and
Ci9 steroids from the adrenal cortex (Poore and Hollander, 1957; Li et al.,
1957; Lostroh and Li, 1957; Wilson, et al., 1958; Farrell et al., 1958).

The carbohydrate-active corticoids stimulate gluconeogenesis from pro-
teins, and this activity is reflected by hyperglycemia and glycosuria (Ingle,
1949; 1950; Jones, 1957). The increased levels of glucose in the blood and
urine are also partly due to the inhibition of glucose utilization (Jones,
1957) . These hormones also increase glycogen deposition in the liver by
accelerating its formation and depressing its release (Ligle, 1950; Jones,
1957) . Glycogen deposition commonly is used to bioassay steroids for their
gluconeogenic activity and other effects of carbohydrate metabolism (Dorf-
man, 1949). The carbohydrate-active steroids not only increase protein
catabolism, but they also depress protein anabolism (Engel, 1952). These
two actions on protein metabolism are reflected by an increase in the non-
protein nitrogen of the blood as well as an increase in the excretion of uri-
nary nitrogen (Selye, 1950) . Lipogenesis is inhibited by the carbohydrate-
active corticoids, but their effects on lipid metabolism are poorly understood
(Jones, 1957) . There is considerable evidence to indicate that hydrocorti-
sone and cortisone increase the sensitivity of blood vessels to the actions of
epinephrine and norepinephrine, and that these steroids perform an essen-
tial function in maintaining normal tonus of the vasculature (Zweifach
et al., 1953; Ramey and Goldstein, 1957). The carbohydrate-active hor-
mones also decrease capillary permeability and fragility and antagonize the
spreading action of hyaluronidase, presumably by their effects on the
ground substance; these corticoids appear to decrease permeability of the
ground substance, and their ability to decrease capillary permeability may
be dependent on this effect (Seifter et al., 1953; Zweifach et al., 1953). In
these effects the carbohydrate-active corticoids are opposed by the actions
of the sodium-retaining corticoids and growth hormone (Seifter et al., 1953;
Kass et al. 1953b; Dougherty and Schneebeli, 1955; Kramer et al., 1957) . In
addition to their catabolic effect on protein, the carbohydrate-active corti-
coids have specific suppressive effects on osteogenesis, chondrogenesis,
mitosis, growth in general, connective tissue growth, inflammation, phago-
cytosis, granulation, and antibody formation (Taubenhaus and Amromin,
1950; Baker, 1950, Selye, 1951; Dorfman, 1953; Dougherty, 1953; Tauben-

210 /. /. Christian

haus, 1953; Bullough, 1955; Dougherty and Schneebeli, 1955; Kass et al.,
1955; Irving, 1957) . The effects on inflammation result from a failure of the
usual inflammatory cells, lymphocytes and fibroblasts, to appear at the site
of injury (Dougherty, 1953; Dougherty and Schneebeli, 1955). The lack of
an adecjuate inflammatory response together with the failure of adetiuate
granulation to take place markedly delays wound healing (Dougherty,
1953; Dougherty and Schneebeli, 1955) . These effects, coupled with inhibi-
tion of phagocytosis and antibod}^ formation, result in a marked decrease
in resistance to infections, so that an animal may be rapidly overwhelmed
by an infection (Thomas, 1953). There is ample experimental evidence to
show that cortisone and hydrocortisone decrease host resistance to infection
by a wide variety of pathogenic viruses, bacteria, protozoan and metazoan
parasites (Thomas, 1953; Shwartzman and Aronson, 1953; LeMaistre^ia/.,
1953; Kass et al. 1953b; Robinson and Smith, 1953; Whitney and Anigstein,
1953; Pollard and Wilson, 1955) . Animals resistant to particular organisms
may be made nonresistant by these steroids, and usually mild infections
may become highly virulent.

High physiological doses of cortisone or hydrocortisone in the pregnant
mammal may result in the development of malformations, especially cleft
palate, in the fetus, the particular anomaly apparently depending on the
stage of development of the fetus when it is subjected to the actions of the
hormone (Glaubach, 1952; Fraser et al, 1953; Davis and Plotz, 1954; Kal-
ter, 1954; Moss, 1955). Cortisone and hydrocortisone both produce cleft
palates and other congenital defects in the fetus when injected into pregnant
mice, the incidence of these anomalies being greater when the injections
were made on the tenth day than when later ( Fraser et al., 1953) . The tera-
togenic effects of cortisone in mice have been shown to be decreased with
increased maternal body weight and to be affected by maternal genotype
(Kalter, 1954; 1956). Treatment of pregnant rats with high physiologic
doses of cortisone results in a significant increase in intra-uterine mortalit3\
occurring minly at mid-term, and later (Seifter et al., 1951 ; Davis and Plotz,
1954) . High doses of cortisone administered to nursing mice 9-12 days after
parturition depress the growth of progeny, whereas ACTH and low doses
of cortisone were without effect on the offspring, except to abolish the
difference in growth rate normally seen between male and female mice
(Glaubach, 1952). Cortisone, and to a lesser degree ACTH, depresses the
growth of infant rats, stimulates the eruption of teeth, opening of the eyes,
and development of the gingivae (Parmer et al, 1951) . Cortisone in a total
dose of 0.5 mg. given to newborn rats during the first week produced long-
term damage, as indicated by the failure of the animals to attain normal
body weight after three months (Parmer et al., 1951). Corticosterone and
pregneninolone were without effect in these experiments. Cortisone treat-

2. Endocrines and Populations 211

mciit of neonatal rats can also result in marked morphologic changes in
the brain and skull (INIoss, 1955). Cortisone or hydrocortisone may stimu-
late lactation (Selye, 1954), but the mechanism by which this is accom-
plished is unknown. These effects cannot be attributed to an inhibition of
the secretion of gonadotropin by cortisone, as it has been shown that even
relatively large doses of cortisone are without effect on the production of
gonadotropins (Byrnes and Shipley, 1950), although it is well known that
enormous doses of corticoids do exert some antigonadotrophic activity.

The carbohydrate-active corticoids also involute lymphoid tissues by
producing degeneration and actual fragmentation of the lymphoid cells,
inhibition of differentiation, and depression of lymphocytopoiesis (Selye,
1950; Dougherty, 1953; Santisteban and Dougherty, 1954; Gordon, 1955;
Weaver, 1955). These effects are also seen following injection of ACTH,
with an increase in endogenous corticosteroid secretion (Baker ei al., 1951).
Lj^mphocytolysis evidently serves to release a readily available store of
amino acids and may serve to provide a sudden flood of stored antibodies,
which are normally produced m the lymphoid tissues (Keuning et al.,
1950; Dougherty, 1953; Kass et al., 1953a; Sundberg, 1955). These actions
result in involution of the thymus, lymph nodes, and malpighian corpuscles
of the spleen. Therefore weights of those organs may provide useful indices
of adrenocortical activity when they are used along with other indices of
adrenal activity, such as adrenal weight, and appropriate controls. It
should be remembered, however, that androgens, and to a somewhat lesser
extent estrogens, are capable of involuting the thymus (Burrows, 1949;
Weaver, 1955) ; therefore cognizance must be taken of this fact when using
thymic involution as a means of appraising adrenocortical activity. How-
ever, the lymph nodes lose weight only after treatment of the animal with
ACTH or carbohydrate-active corticoids (Weaver, 1955). Estrogen, tes-
tosterone, thyroid extract, adrenalectomy, thyroidectomy, and gonadec-
tomy were without effect on the lymph nodes in these experiments (Weaver,
1955) . The adrenal carbohydrate-active corticoids also depress the numbers
of circulating eosinophils and lymphocytes, so that counts of these cells are
frequently used to assess the functional integrity of the pituitary-adreno-
cortical system (Speirs and Meyer, 1949; 1951; Gordon, 1955; Speirs,
1955). In using counts of eosinophils or lymphocytes as indices of adreno-
cortical activity in wild mammals care must be taken (1) to standardize
the procedures so that the results are completely comparable from count
to count, and (2) not to elicit an adrenocortical response during the process
of handling the animal.

The biological activity of ll^-hydroxyA^-androstene-3,17-dione (llOH-
4AD) and the closely related steroid 11/3-hydroxytestosterone, as well as
other related C19 steroids, deserve further comment, as one or the other of

212 J. J. Christian

thefirst two, probably the first, is a major secretory product of the cortex,
presumably of the zona fasciculata-reticiilaris of house mice, rats, and very
possibly other rodents (Bush, 1953; Wilson ei al., 1958), and to these must
be added dehydroepiandrosterone, which is now known to account for half
of the steroid product of the adrenal cortex of human beings (VandeWiele
and Lieberman, 1960) . These steroids, as a group, are very weak androgens
with insufficient activity to maintain the seminal vesicles and ventral
prostate in hypophysectomized mice (Bahn et al., 1957), although they
evidently are sufficiently androgenic to produce histologically detectable
changes in the epithelium of these organs, if not in changes in gross weight
(Davidson and Moon, 1936; Lostroh and Li, 1957), and evidently, if they
are secreted in large enough ciuantities, they can produce masculinization
in humans (Dorfman and Shipley, 1956). In addition, Howard (1959) has
shown that these weak androgens are more strongly androgenic if their
activity is measured in terms of other assays, such as stimulation of the
preputials and os penis. However, the C19 steroids with weakly androgenic
activity, as measured by their ability to stimulate growth of the prostate
or capon comb, can inhibit the secretion of gonadotropins in rats, especially
in immature animals (author's italics) , although it has been shown that the
carbohydrate-active corticoids are incapable of producing this effect
(Byrnes and Shipley, 1950; Byrnes and Meyer, 1951; Wilson et al., 1958).
Therefore, it is possible that an increased secretion of these androgens
steroids by the adrenal cortex in mice and rats can account, at least in part,
for the suppression of reproduction commonly associated with circum-
stances which increase the secretion of ACTH and carbohydrate-active
corticoids, as described by Selye (1939).

These, in brief, are the actions of the important adrenal cortical hor-
mones. Many questions remain unanswered regarding the functions of the
cortical hormones, especially wath respect to their relationships to each
other and to other endocrines, such as the thyroid and pancreatic islets.
The carbohydrate-active corticoids in many respects are antagonistic to
insulin and probably suppress thyroid activity, but these topics will not be
dealt with here. The actions already listed are the major activities of the
cortex which will enable an interpretation to be made of, as well as to
anticipate, the results in other species. All the above effects have been
duplicated by injecting adrenocorticotropin into intact mammals and
thereby stimulating an increased secretion of endogenous adrenocortical
steroids. They also have been produced by alarming stimuli, which increase
the secretion of endogenous ACTH and in turn endogenous corticoids.
Among these stimuli are cold, emotional trauma, physical trauma, toxic
agents, and many others, although the general response to these stimuli is
not necessarily quantitatively, or even quantitatively, similar in every case.

2. Endocrines and Populations 213

There is much to learn about the differences in hormonal response to
different stimuli. There are excellent reasons to believe that the various
adrenocortical hormones are not always secreted in any fixed proportion to
each other independently of the nature of the stimulus or its degree of
severity (cf. above). Therefore it seems advantageous to speak of the
responses evoked by the specific stimulus in each case rather than of a
generalized nonspecific response to "stress."

(3) Regulation of secretion of the fascicular hormones. The regulation
of the secretion of the hormones of the cortical zona fasciculata has been
investigated intensively, and a great deal is known about the mechanisms
involved, although many questions still remain unanswered. It is well
established that adrenocorticotropin (ACTH), a protein hormone secreted
by the anterior pituitary, stimulates corticoid production and release from
the adrenal fasciculata and to a lesser extent stimulates the secretion of
aldosterone from the zona glomerulosa (Wexler et al., 1955; Allen, 1957;
Farrell ct al, 1955; Jones, 1957; Li et al, 1957; Farrell et al, 1958, and
other references cited earlier) . ACTH also stimulates cellular hyperplasia
and hypertrophy of the cortex, although pituitary growth hormone may
play an important role in stunulating cortical hypertrophy (Cater and
Stack-Dunne, 1953, 1955; Jones, 1957; Lostroh and Li, 1958), possibly
acting synergistically with ACTH (Lostroh and Woodward, 1958). How-
ever, growth hormone has no effect on the release of corticosteroids (Rosen-
feld and Bascom, 1956; Guillemin et al, 1958) . ACTH also has a number of
extra-adrenal actions; for example, it stimulates the preputial glands and
other sex accessories, but not to the same extent as does testosterone
(Davidson and Moon, 1936; Davidson, 1937; Jacot and Selye, 1951;
Lostroh and Li, 1957) . The regulation of the release of ACTH from the
pituitary is not clearly understood, although there no longer can be any
doubt that for the most part a neurohumor from the hypothalamus has a
major role in regulating the production and release of ACTH (IMcCann,
1953; McCann and Brobeck, 1954; Harris, 1955a, b; Fortier, 1957; Wood-
bury, 1958). The regulation of the release of ACTH has been reviewed by
Harris and Fortier (1954), Harris (1955b), and Fortier (1956), but advances
in this field are rapid, and much additional information is available which is
not included in these reviews. Some doubt has been cast on the role of the
hypothalamus in regulatmg ACTH secretion in response to surgical trauma
(Story efai., 1959).

The adenohypophysis has a double arterial supply: a portal system
draming from the hypothalamus and a direct systemic arterial supply
(Green, 1951). Branches from the superior hypophyseal arteries form a
tufted plexus of capillaries in the median eminence of the hypothalamus in

214 /. /. Christian

intimate association with nerve fibers of the supraopticohypophyseal tract.
These vessels drain into the adenohypophysis, where they break up to
form a series of sinusoids. Here they are joined by capillaries from the
systemic arterial supply. The direction of the flow is from the median
eminence to the adenohypophysis. Neural fiber tracts from hypothalamic
nuclei descend through the median eminence and give off branches which
terminate in intimate association with the arterial portal plexus in the
median eminence (Scharrer and Scharrer, 1954). The remaining fibers
continue on their way, to terminate in close approximation to capillaries
in the neurohypophysis. Neurosecretory material is believed to traverse
the sheaths of these nerves from hypothalamic secretory nuclei (Scharrer
and Scharrer, 1954; Rennels and Drager, 1955). There is evidence that a
fraction of this neurosecretory material is responsible for stimulating the
release of ACTH from the anterior pituitary (Drager, 1955; Rothballer,
1953; Saffran et al., 1955). Evidently the material is released by the appro-
priate stimulus from the terminations of the neurosecretory fibers into the
pituitary portal vessels and is then borne, via the portal system, to the
cells of the adenohypophysis proper (Scharrer and Scharrer, 1954; Rennels
and Drager, 1955). It is generally accepted that no nerve fibers terminate
in the adenohypophysis and that there are no direct neural, only vascular,
connections between the adenohypophysis and neurohypophysis (Green,
1951). The identity of the hypothalamic corticotropin-releasing factor
(CRF) has not been determined, although it appears to be a small protein
often associated with Pitressin (Saffran et al., 1955; Porter and Rumsfeld.
1956). One group of investigators believes that the ACTH-releasing ac-
tivity is associated with Pitressin, in particular with its pressor activity
(ADH, vasopressin) (McCann, 1957; jMcCann and Fruit, 1957). ADH
exerts a powerful antidiuretic action on the kidney at many times the
dilution that is necessary for it to exert a pressor effect or to stimulate the
release of the ACTH (McCann and Fruit, 1957) . It has been suggested for
this reason that the concentration of ADH, when it is released from the
posterior pituitary into the systemic circulation, is sufficient to effect anti-
diuresis, but insufficient to stimulate the release of ACTH, but that when
it is released directly into the pituitary portal circulation, it reaches the
adenohypophysis in a sufficient concentration to effect the release of ACTH
(McCann and Fruit, 1957). It should be noted that ADH evidently is
released simultaneously into the pituitary portal system from the hypothal-
amus and from the neurohypophysis into the systemic circulation in re-
sponse to stimulation.

Nevertheless, as attractive as this theory may seem, good evidence is
accumulating from a number of sources which indicate's that the corti-
cothopin-releasing factor is an entity separate from vasopressin, although

2. Endocrines and Populations 215

the site of its origin in the hypothalamus and its exact identity remain mi-
known (Saffran et al., 1955; Porter and Jones, 1956; Porter and Rumsfeld,
1956; Clayton et al., 1957; Schally and Guillemin, 1959), and that the
release of ADH may be completely unrelated to the release of ACTH (Mc-
Donald et al, 1957; Schapiro et al, 1958) . It appears that this substance is a
small protein which is recoverable from hypophyseal portal blood, from the
neurohypophysis, or from extracts of appropriately selected portions of
the hypothalamus, and that it is probably loosely bound to and travels with
a much larger protein which is inactive with regard to the release of ACTH
(Porter and Rumsfeld, 1956; Guillemin et al, 1957; Schally and Guillemin,
1959). The activity of the corticothopin-releasing factor may be augented
by the simultaneous action of epinephrine (Saffran et al, 1955) . As is so
often the case, the solution to this problem may prove to be midway be-
tween, or to be a combination of, the two opposing views. It is possible
that there are hypothalamic substances, such as vasopressin, other than a
specific corticotropin-releasing factor which are capable of effecting the
release of specific ACTH fractions from the adenohypophysis. Recent-
evidence indicates that different stimulating factors are involved in ac-
tivating different adrenocortical functions (Fortier, 1956; Guillemin et al,
1958 ; Slusher, 1958 ; Nowell, 1959) .

In spite of the existing uncertainty with regard to the nature of the speci-
fic factor or factors from the hypothalamus which stimulate the release of
ACTH, there is little doubt that hypothalamic substances carried to the
adenohypophysis by the portal system can stimulate the release of ACTH
in response to alarming stimuli, although Nowell (1959) has suggested
that emotional and systemic stress involve different pathways for the
release of ACTH. However, there are many other aspects of the overall
regulation of the secretion of adrenocorticotropin which are not clear at
the present time. One of these aspects is the control of the normal daily low-
level secretion of ACTH. There probably is a basal release rate of ACTH
which is independent of the hypothalamic-hypophyseal portal system and
which is sufficient for normal maintenance of adrenocortical function
(Fortier, 1957) . A number of workers maintain that the release of ACTH is
subject to hypothalamic control only in response to acute stressful stimuli
(Woodbury, 1958) . The level of circulating corticoids undoubtedly exerts a
control over the rate of secretion of ACTH, as circulating adrenocorticoids
are capable of inhibiting the release of ACTH under most circumstances,
although the more severe the stimulus, the greater the level of circulating
corticoids must be in order to block ACTH secretion (Sydnor and Sayers,
1954; Farrell and Laqueur, 1955; Sydnor et al, 1955; Richards and Pruitt,
1957). Fvu'thermore, Fortier (1959a, b) has shown that the corticotropin-
releasing effect of stress is influenced markedly by the level of circulating

216 /• /• Christian

corticoids. Withdrawal of the corticoids enhances both the release and
synthesis of ACTH, its predominant effect being on the synthesis (Fortier,
1959a, b) . The steroids appear to act on the pituitary itself, but they may
act also on the hypothalamic centers, or on the releasing mechanism or
mechanisms themselves, or even may act to some extent on the adrenal
gland directly (Birmingham and Kurlents, 1958). However, the corticoids
evidently are not capable of blocking the release of ACTH in response to all
types of tissue damage, unless the steroids are present in very large quan-
tities. It has been suggested in the case of tissue damage, such as burns,
that substances are released from the site of the injury (Hume, 1953;
Woodbury, 1958; Share and Stadler, 1958), which are able to stimulate
directly the release of ACTH by direct action on the pituitary, although
final proof of such a suggestion is lacking. It is unlikely that histamine,
epinephrine, serotonin or a variety of other compounds which are released
during tissue-breakdown are responsible for directly stimulating the release
of ACTH (Sandberg et al, 1953; Guillemin, 1955; Guillemin et al, 1957;
Woodbury, 1958) as has been suggested. Nervous impulses from the site
of injury cannot be responsible for initiating the release of ACTH following
injury, as complete denervation of the part or transection of the cord fails
to block the release of ACTH (Hume, 1953). The mechanism by which
adrenocorticotropin is released following tissue damage needs to be deter-
mined.

Finally, there is evidence that two or more substances from the hypo-
thalamus may regulate different activities of the adrenal cortex. Slusher
(1958) has shown that lesions in one region of the hypothalamus will block
the usual fall in adrenal ascorbic acid which follows cortical stimulation
without preventing the release of cortical steroids. On the other hand,
lesions in a nearby region can prevent the release of corticoids without pre-
venting the fall in ascorbic acid. Guillemin ct al, (1958) have shown that
there may be a functional separation of corticoid secretion from ascorbic
acid depletion. These investigators found that a marked secretion of corti-
coids could occur in the absence of detectable changes in adrenal ascorbic
acid concentration when both variables were measured simultaneously fol-
lowing stimulation by ACTH in hypophysectomized rats. Finally, there is
other evidence which suggests that the factors controlling cortical hyper-
trophy (adrenal weight factor: AWF) and ascorbic acid (ascorbic acid
factor: AAF) reduction are separate entities and may respond differentially
to different stimuli (Nowell, 1959) . Nowell (1959) has suggested that stim-
ulation of the release of various ACTH factors from the pituitary may in-
volve different mechanisms for different stimuli; for example, emotional
stimuli may require activation of the hypothalamic centers, whereas adre-
nal regeneration may not require hypothalamic activity.

2. Endocrines and Populations 217

Plainly there are many problems still to be solved before a complete and

detailed explanation of the regulation of the release of ACTH and the
fascicular corticoids will be possible, but the evidence for the mechanisms
which have been described is strong. There is no doubt that the hypothala-
mus has a major role in the regulation of ACTH release and of adenohy-
pophyseal function in general, and there is no reason to question the
existence of such a neurohumoral mechanism. The evidence for some of the
details, however, is less conclusive. The cortical glomerulosa and the re-
lease of aldosterone are largely independent of this system of regulation, as
indicated earlier, except for a very moderate response to ACTH for a
limited time.

c. The junction of the zona reticularis. The zona reticularis has been
thought to secrete sex steroids (Selye, 1947), especially androgens, but
the evidence for such a function is poor and largely inferential (Dempsey,
1948; Creep and Deane, 1949b; Jones, 1957). This zone is made up of a
continuation of the cords or continuum, as the case may be, of the zona
fasciculata, the cells of which are in varying degrees of activity and respond
accordingly to stimulation by ACTH (Jones, 1957). By and large they
have the appearance of cells of the fasciculata which are declinmg in ac-
tivity and becoming obsolete, but there is certainly some functional capa-
city in the reticular cells, as indicated by various histochemical procedures
and the presence of mitosis, although it is apparently much less than that
of the zona fasciculata (Dempsey, 1948; Creep and Deane, 1949b) . Syming-
ton et al. (1958) have presented convincing morphologic and histochemical
evidence, coupled with secretory studies, that the reticularis normally
secretes the hormones usually attributed to the zona fasciculata and that
the latter is in a resting state until fvnther stimulated by ACTH. This sug-
gestion is well worth considering especially as morphologic changes are
related to steroid production by direct measurement. These studies necessi-
tate a revision of the classic concepts of the functional roles of the reticularis
and fasciculata which have been based only on morphologic evidence.

d. The problem of the function of X-zone. The X-zone of the adrenal cortex
of the immature house mouse (Mus musculus) is maintained by gonado-
tropin rather than by ACTH or growth hormone (Jones, 1949b, 1950, 1952,
1957) . Its function is unknown, and histochemically it does not present the
appearance of being a secretory zone (McPhail, 1944; Howard, 1939; 1946;
McPhail and Read, 1942a, b) . The cells do not give the usual histochemical
reactions for steroids or other lipids, although they do contain ascorbic
acid (Jones, 1949a, 1950) . There is no evidence that the X-zone of the
house mouse secretes sex steroids (McPhail, 1944; Jones, 1957), although
such a function has been postulated by several investigators (Jones, 1957) .
In fact the large amounts of Cig steroids secreted by the adrenals of mice in

218 ./• ./• Christian

response to ACTH-secrcting pituitary tumors were from the cortictes of
adrenals without X-zones (Bahn et al., 1957). On the other hand, Delost
(1951) suggests that the X-zone may be that part of the cortex responsible
for secreting the adrenal androgens which maintain the activity of the vas
deferens and ventral prostate of meadow voles (Microtus arvalis) in winter
and in adrenalectomized voles (Delost, 1954). The evidence for this sug-
gestion is tenuous and requires further investigation and confirmation. The
function of the X-zone, if there is one, remains undiscovered at the present
time.

3. Adrenocortical-Gonadal Interrelationships

This problem has been discussed by Jones (1957), and the reader is re-
ferred to this work for a more detailed discussion of this intriguing problem.

There is a very close relationship between the cells of the adrenal cortex
and those of the gonadal cells with respect to their origin and function.
Both groups of cells arise from the embryonic genital ridge of the coelomic
mesothelium, both secrete steroid hormones in response to stimulation by
pituitary hormones; and there is a definite overlap in their functional
capacities. Even though the adrenal cortex is believed to be a source of sex
steroids, gonadectomy in normal animals is followed by complete involution
of the sex accessories to the immature condition (Jones, 1957) . The involu-
tion occurs even in mammals such as mice and rats in which weakly andro-
genic Ci9 steroids normally are an appreciable portion of the cortical secre-
tory product (Dorfman and Shipley, 1956; Wilson et al, 1958). Similarly,
the sex accessories and secondary sex characteristics which are under hor-
monal control fail to develop in the prepubertal castrate mammals (Bur-
rows, 1949; Jones, 1957) . Finally, the transplantation of the adrenal to the
uterus does not alter the morphology of that part of the uterus or of the
vagina which is in direct contact with the graft (Sakiz, 1956) . These facts
cast considerable doubt on their ability of the normal adrenal cortex of any
species to secrete steroids with strongly androgenic or estrogenic actions.
There is much evidence, however, that under certain circumstances the
adrenal cortex is capable of assuming sex steroidal activity, for example, in
the androgenital syndrome and similar virilizing conditions in man (Dorf-
man and Shipley, 1956; Gallagher et al, 1958). A striking increase in the
secretion of weakly androgenic adrenal steroids may account for these
effects (Jones, 1957) , but they may also result from the conversion of adre-
nal androgens into testosterone in these circumstances (Dorfman, 1960).
These conditions are usually associated with functional tumors of the cortex
or with a marked increase in the secretion of ACTH with an accompanying
cortical hyperplasia, but without an appreciable increase in the secretion of

2. Endocrines and Populations 219

the cortieoids, although with a marked increase in the secretion of C19
steroids (Dorfman and Shipley, 1956; Bradlow and Gallagher, 1957, Dorf-
man, 19G0) . It is also possible that these effects arc a result of aberrant bio-
chemical pathways in the synthesis of adrenal steroids which produce
greatly increased amounts of weakly androgenic steroids or possibly some
compounds with more strongly androgenic activity (Dorfman and Shipley,
1956; Bradlow and Gallagher, 1957; Gallagher, 1958, Dorfman, 1960).
Finally there may be abnormal or greatly increased androgenic metabolites
of adrenal steroids (Gallagher et at., 1958). It is well known that gonadec-
tomy is followed by a marked increase in the production of gonadotropins ;
therefore if the production of sex steroids by the adrenal cortex normally
was controlled by gonadotropins, one would expect an increased secretion
of these sex steroids to follow gonadectomy and to be reflected by con-
tinued maintenance of the sex accessories. This is not the case, as the sex
accessories involute nearly completely following gonadectomy (Burrows,
1949; Jones, 1957). However, it has been shown that increased secretion
of ACTH is followed after a time by a marked increase in the secretion of
Ci9 steroids (Bush, 1953; Dorfman and Shipley, 1956; Bradlow and Gal-
lagher, 1957; Jones, 1957; Wilson et al, 1958, Vande Wiele and Lieberman,
1960). Apparently slight stimulation of the sex accessories by adrenocorti-
cal androgens occurs following ACTH treatment of gonadectomized rats,
but these effects can be detected only histologically (Davidson and Moon,
1936; Davidson, 1937; Lostroh and Li, 1957, Desclin, 1959) . In marked con-
trast to the general tenor of these results, Delost (1951) reports that the
adrenal cortex secretes hormones with androgenic actions on the sex acces-
sories of male Microtus arvalis, as determined by morphologic criteria, but
this work has not been confirmed. Delost (1956a) maintains that cortisone
will stimulate some of the secondary sex organs in these mammals, but
confirmation of these res ilts also is needed before acceptance of his conclu-
sions is warranted. It is best for the present to assume that the adrenal
cortex of most species normally does not secrete appreciable amounts of
steroids with pronounced androgenic or estrogenic activity and that, it
usually represents abnormal function when appreciable quantities of hor-
mones with these activities are secreted by the cortex, (Bradlow and Gal-
lagher, 1957). The ability of the adrenal cortical and gonadal endocrine
cells to secrete steroids with overlapping activity is an intriguing problem,
and the reader is referred to Dorfman (1960) for a more detailed discussion
of the steroidal biosynthetic pathways involved and the elucidation of some
of these paradoxes. There is certainly variation between species in the types
of adrenal steroids normally secreted. Under abnormal conditions there is
no doubt that the adrenals produce increased amounts of steroids which may
have sex steroidal activity and that the ovaries, for example, can be induced

220 /. J. Christian

to secrete steroids with cortical or androgenic activity. Hill (1948) has
shown that under the proper conditions, the ovaries of mice can assume
adrenocortical function when transplanted to the ears. Under other circum-
stances the ovaries can be made to secrete androgens (Hill, 1937; Delost,
1955) . It must be apparent that the whole field of adrenal-gonadal relation-
ships is poorly understood, especially with regard to species differences.

There is one area of adrenal-gonadal relationships, however, which seems
to be on relatively solid ground. Selye (1939) called attention to the fact
that a variety of stimuli which elicited increased adrenocortical activity
also inhibited reproductive activity. Female rats subjected to a variety of
alarming stimuli exhibited ovarian atrophy and more or less permanent
anestrus. These observations have been confirmed and extended for a
variety of species, including mice, voles, man, and others, and for a wide
variety of stimuli (Christian, 1959b). The entire and exact mechanisms by
which inhibition of reproductive function is brought about are not known
with certainty. It may be due in part to a direct suppressive action of the
carbohydrate-active corticoids on the reproductive organs; for example, it
has been shown that cortisone and hydrocortisone depress the ovarian re-
sponse to chorionic gonadotropins in intact rats (Smith, 1955), possibly
by direct action on the vasculature. However, it is more likely that the
suppression results primarily from an inhibition of the secretion of gonado-
tropins from the anterior pituitary when there is an increased production
of adrenocorticotropin. Selye (1939) suggested that there is a "shift" in
pituitary function; the production of gonadotropins is sacrificed in order
to increase the production of "life-maintaining" ACTH. But it is equally,
if not more, likely that the secretion of gonadotropin is inhibited by in-
creased amounts of cortical androgens following stimulation by ACTH.
The ability of androstenedione and other related weakly androgenic
steroids to suppress pituitary gonadotropic activity has already been
discussed and it has also been stated that these steroids appear to be major
natural secretory products of house mice, rats, man and probably other
species. It is particularly interesting, especially to those studying popula-
tion phenomena, that immature mice are significantly more susceptible
than mature mice to the suppressive effects of steroids on the secretion of
gonadotropins (Byrnes and Shipley, 1950; Byrnes and Meyer, 1951) . There
is also the possibility that nervous stimuli, acting directly through the
hypothalamus, can depress gonadotropin secretion.

Whether one or all of these mechanisms is involved, there is a decrease in
the secretion of gonadotropin with the result that the secretion of sex
steroids by the gonads is inhibited and the secondary sex organs may ap-
proach the appearance seen following gonadectomy or hypophysectomy
(Selye, 1939; Christian, 1959b). The germinal cells of the gonads also

2. Endocrines and Populations 221

reflect the withdrawal of stimuhition by the gonadotropins and sex steroids.
Inhibition of these endocrine functions can result in a depression of all
reproductive activities, including the onset of puberty, lactation, fertility,
normal mamtenance of the embryos in utero, size of the secondary sex
organs, and other functions associated with full functional competence of
the pituitary-gonadal system (Christian, 1956, 1959a, b). It is pertinent
here to note that androgen injected into female mice or rats under 10 days of
age will produce permanent sterility ; this finding suggests a method for the
production of sterility or markedly delayed maturation in natural popula-
tions (Barraclough, 1961; Barraclough and Gorski, 1961). Most of these
effects presumably are brought about by a decreased secretion of gonado-
tropins, but some of the failures in reproductive function in these situations
undoubtedly result from the direct effects of increased levels of corticoids
follow^ing increased secretion of ACTH. Certainly the ability of the carbo-
hydrate-active corticoids to increase protein catabolism and suppress
growth, especially of the connective tissue and its products, and to suppress
mitoses, must have appreciable effects on the reproductive process, espe-
cially on the developing fetus, as we have indicated earlier. There is con-
siderable evidence to indicate that alarming stimuli during pregnancy, pre-
sumably with a marked increase in the secretion of adrenocortical steroids,
can result in congenital anomalies dependent on the stage of fetal develop-
ment (Fraser et al., 1953; Aycock and Ingalis, 1946; Ingalls, 1956; Curley
and Ingalls, 1957; Ingalls and Philbrook, 1958).

Finally it should be borne in mind that the degree of reproductive sup-
pression will vary in some proportion to the severity of the inducing stimu-
lus and that there w411 be some variation with species with respect to the
particular part of the reproductive cycle which will be most severely cur-
tailed. For example, intra-uterine mortality and resorption of the embryos
is marked in house mice, but apparently is inconsequential in Norway rats,
whereas the postparturitional loss of young is as great or greater in rats
than it is in mice (Christian, 1959b). Voles evidently are quite susceptible
to depression of fertility and development of maturity (Kalela, 1957) , but
house mice are affected similarly (Crowcroft and Rowe, 1957). These
variations are of particular importance to the investigator involved in
comparative studies, especially in studies of phenomena relating to popula-
tion density. It should be kept in mind that the severity and duration of
the stimulus and the age of the affected mammal also will have an important
bearing on the particular stage of the reproductive cycle affected as well as
the degree of its inhibition. There are numerous examples of modification
of reproductive function in response to various environmental stimuli,
probably representative of activity of physiologic adaptive mechanisms.
Baker and Ransom (1932) found that winter temperatures (5°C.) signifi-

222 /. /. Christian

cantly reduced the number of births on Microtus agrestis even though they
were maintained on summer food. Chitty and Austin (1957) have called
attention to the effects of environmental factors, especially social "stress,"
on changes is estrous pattern and behavior exhibited by Microtus agrestis.
Additional examples of the effects of deleterious environmental factors on
reproduction will be given later.

There is another aspect of the suppression of reproduction which should
be mentioned at the present time, although it is not directly pertinent to
adrenocortical-gonadal relationships. Inanition, starvation, protein defi-
ciency, and probably other dietary inadequacies are capable of depressing
or totally inhibiting reproduction, the degree depending on the severity of
the inadequacy (Lutwak-]\Iann, 1958), without stimulating increased
secretion of ACTH and therefore increased adrenocortical activity (Baker,
1952; Christian, 1959b, c; Eisenstein, 1959), and adrenocortical secretion is
actuallj^ depressed in rats deficient in pantothenic acid (Eisenstein, 1957).
The experimental evidence on which this statement is based has been
derived from laboratory mammals and therefore does not necessarily apply
to other species or to natural populations, but the burden of proof lies with
those who claim otherwise. Recent experiments with house mice have
demonstrated rather conclusively that inanition does not stimulate the
pituitary-adrenocortical system, although its effect on the reproductive
system is striking ( Lutwak-Mann, 1958; Christian, 1959b, c). Further-
more, a deficiency of vitamin Be or partial starvation does not increase the
secretion of corticoids nor impau- the ability of the adrenal cortex to secrete
them, even though they may be a marked cortical hypertrophy (Eisenstein,
1959). These results are in agreement with those from a large number of
earlier experiments (cf. Baker, 1952; Christian, 1959c) . In addition Srebnik
et al. (1958) have shown that a protein deficiency results in a failure in the
secretion of gonadotropins, and jVIarrian and Parkes (1929) had shown
earlier that anterior pituitary extracts could correct the anestrus produced
by a dietary insufficiency of vitamin B with the return of normal estrus and
normal changes in the accessory organs of reproduction with the estrous
cycle. One might hypothesize that an inadequate diet would impair the
ability of the anterior pituitary to synthesize its protein hormones. How-
ever, the ability of the pituitary-adrenocortical system of mice to exhibit
the usual responses to alarming stimuli, and therefore the ability to elabo-
rate adrenocorticotropin, in spite of inanition indicates that this is probably
not the case (Christian, 1959c). Therefore, the decreased production of
gonadotropins by animals subjected to inanition, starvation, or other
dietary inadequacies probably operates through an unknown mechanism,
probably at the hypothalamic level, which does not reflect an inability of
the anterior pituitary to synthesize protein hormones.

2. Endocrines and Populations 223

4. Epinephrine and Norepinephrine ; the Hormones of the Aduexal
Medulla and Sympathetic Nervous System

The adrenal medulla and its hormones, epinephrine and norepinephrine,
have been the subjects of numerous and ^"oluminous reviews and are also
discussed in considerable detail in most good texts of physiology and phar-
macology. Therefore, except for a few aspects, a detailed account of these
hormones and their physiologic and pharmacologic actions will not be given
here. The reviews of the following investigators may be referred to for a
more detailed coverage of the subject: Hartman and Brownell, 1949; von
Euler, 1951; Hagen and Welch, 1956; Gaddum and Holzbauer, 1957;
Ramey and Goldstein, 1957; Elmadjian et al., 1958.

The adrenal medulla is an integral part of the sympathetic nervous sys-
tem. The medulla is homologous with the sympathetic ganglia and is in-
nervated by cholinergic preganglionic fibers of the splanchnic sympathetic
nerves. Upon stimulation of the sympathetic nervous system, the adrenal
medulla discharges norepinephrine and epinephrine into the systemic circu-
lation, the proportions of these two compounds varying with the species
and with the nature of the stimulus (von Euler, 1951; Hagen and Welch,
1956; Gaddum and Holzbauer, 1957; Gray and Beetham, 1957; Elmadjian
et al., 1958; Goldfien et al., 1958). Norepinephrine is also secreted by the
postganglionic sympathetic nerves and by the extra-adrenal chromaffin
tissue of the sympathetic nervous system (von Euler, 1951; Hagen and
Welch, 1956). Norepinephrine probably is the neurohumoral transmitter
substance of the postganglionic sympathetic nervous system, and appar-
ently is released on nervous stimulation at the sympathetic nerve endings
(von Euler, 1951 ; Hagen and Welch, 1956; Richardson and Woods, 1959).
These two hormones have profound effects on the circulatory system and
glucose and fat metabolism, but by and large their effects are short lived,
owing to their rapid destruction in the body by the cytochrome oxidase
system or by amine oxidases (Bell et al., 1950; Gaddum and Holzbauer,
1957). Norepineplirine and epinephrine have a variety of effects over the
entire body which are brought about largely by their actions on smooth
muscle and which in general parallel the effects of stimulating the sympathe-
tic nervous system (Hartman and Brownell, 1949).

Epinephrine and norepinephrine both have profound pressor effects on
the cardiovascular system and on the levels of blood sugar; but epinephrine
in general has a greater effect on carbohydrate metabolism and produces a
greater hyperglycemic response than norepinephrine, whereas norepineph-
rine has a greater pressor effect than epinephrine (Gaddum and Holzbauer,
1957). In many ways the actions of these two hormones are similar, but in
others they have opposing actions. Norepinephrine in general produces a

224 /• /. Christian

greater rise in blood pressure than epinephrine because it increases overall
peripheral resistance, largely by constricting the vasculature of the muscles
as well as of the skin (Bell et al., 1950) . Epinephrine produces a greater
constriction of the vasculature of the skin, but dilates the vessels of the
skeletal muscles and increases the cardiac output by increasing the rate and
strength of the heart beat. The effects of norepinephrine on cardiac output
are variable. Both of these amines decrease the formation of urine, produce
relaxation of the gut by inhibition of its smooth muscle, produce splenic
contraction, dilate the bronchi, inhibit the bladder, and produce pupillary
dilatation (Bell et al., 19501 . Both produce a rise in blood sugar but, as we
have mentioned, epinephrine produces a greater rise than norepinephrine.
The rise in blood sugar and subsequent glucosuria result from the mobiliza-
tion of glucose from the readily available stores in the liver, and secondarily
from the muscles. The immediate effect of epinephrine is to release glucose
from the available stores of liver glycogen ; therefore the magnitude of the
resulting hyperglycemia depends on the amount of glycogen in the liver
(Hartman and Brownell, (1949). The eventual effect of epinephrine, after
an initial depletion of liver and muscle glycogen, is to shift carbohydrate
from the muscles to the liver, as the uptake of glucose by muscle is de-
pressed and it is well known that lactic acid derived from muscle glycogen
is used by the liver to synthesize glycogen.

The actions of these hormones are the classic preparations for "fight or
flight" in response to emergency situations (Cannon, 1915, 1932) . The com-
bined activity of epinephrine and norepinephrine ensure adequate blood
and glucose to the muscles, increased oxygenation, and adequate blood
flow. An increased supply of oxygen to the tissues is ensured by an increase
in respiratory rate, bronchial dilatation, and contraction of the splenic
capsule with release of stored red blood cells into the circulation. Other
activities, unneeded in an emergency, are suppressed. The adrenal medulla
and sympathetic nervous system respond to cold, fear, rage, trauma, pain,
blood loss, anoxia, emotional tension, and a variety of additional alarming
stimuli. A variety of chemical agents, such as potassium and serotonin, will
release the catechol amines from the medulla (Gaddum and Holzbauer,
1957). The sympathico-adrenal system represents a major and immediate
reaction system of the body to prepare for, or to counteract the effects of,
an emergency situation. The acute response is relatively short lived and
serves to maintain life and counteract shock until the emergency passes or
until longer-acting adaptive systems, such as the pituitary-adrenocortical
system, take over and aid in physiologic adaptation to the situation.

Recent evidence has shown that norepinephrme is normally found in the
walls of the arteries (Schmiterlow, 1948), and that it plays a major, per-
haps decisive role in the maintenance of normal vascular tonus and reac-

2. Endocrines and Populations 225

tivity, achieving the latter by diminishing the sensitivity of the arterial
musculature to epinephrine and norepinephrine by maintaining a constant
low level of pressor amines in the arterial wall (Burn and Rand, 1958a, b).
The source of the norepinephrine in the arterial walls is apparently the
chromaffin tissue including the adrenal medulla, or sympathetic neural
terminations which appear to release a low level of these cathechols amines
constantly into the circulation (Bell et al., 1950; Gaddum and Holzbauer,
1957). Experiments with reserpine (Burn and Rand, 1958a, b; Eranko
and Hopsu, 1958) , which depletes the epinephrine and norepinephrine from
the adrenal medulla and sympathetic chromaffin tissue, depletes the con-
tent of catechol pressor amines from the arterial walls and thereby makes
them excessively sensitive to circulating epinephrine and norepinephrine.
However, the arteries are unresponsive to other noncatechol pressor amines
which apparently exert their usual effects by releasing the norepinephrine in
the arterial walls (Burn and Rand, 1958a, b). There is also evidence that
the adrenal carbohydrate-active corticoids have a part m the maintenance
of arterial tonus and reactivity by increasing the sensitivity of the vascula-
ture to the action of epinephruie and norepinephrine (Ramey and Gold-
stein, 1957) .

Reserpine is a pharmacologic agent which causes the disappearance of
the catechol pressor amines from the chromaffin tissue and subsequently
from the arteries (Burn and Rand, 1957, 1958b), but stimulation of the
sympathetic nervous system also can exhaust the pressor amines from the
sympathetic ganglia and adrenal medulla, although the stimulus must
persist for 30 minutes or longer to achieve exhaustion of the adrenal medul-
las of dogs (Gaddum and Holzbauer, 1957). Therefore, it is conceivable
that prolonged and intense emotional stimuli, such as one might expect as
a result of social interactions between animals in populations of excessive
density, might exhaust the stores of pressor amines, especially in the sub-
ordinate animals. If such an event occurs, one might anticipate that there
would be a subsequent depletion of the arterial norepinephrine and loss of
arterial tonus which might account for the occasional deaths due to the
shock seen in mice shortly after they are first placed together (Christian,
1955b) or following more protracted periods of social strife (Frank, 1953).
A loss of vascular tonus with a subsequent hypotension, and eventually
shock with circulatory collapse, could explain the symptoms observed by
Frank (1953) in dense populations of Microtus in the wild or in captivity
or might be a part of the picture of "shock disease" (Green and Larson,
1938; Green et al, 1939; Christian and Ratcliffe, 1952). There is also the
possibility of a simultaneous exhaustion of readily available glucose re-
serves by the action of epinephrine, especially in animals with a high
metabolic rate, or, perhaps more likely, a loss of the ability to mobilize

226 /. /. Christian

reserves due to exhaustion of the supplies of epinephrine. Such a mecha-
nism, albeit conjectural, may provide a better explanation for the immedi-
ate and precipitate cause, the proximate cause, of "shock disease" than the
previously postulated pituitary- adrenocortical exhaustion (Christian,
1950b) , although adrenocortical hyperactivity probably plays an additive
or even synergistic role in the cause of the immediate mortality in "shock
disease." These conjectures are not meant to relegate the pituitary-adreno-
cortical-gonadal system to a secondary role in the more prolonged and
chronic effects of increased population density or in the control of popula-
tion growth, as we shall see later. However, the available experimental
evidence places the sympathico-adrenal medullary system in the forefront
of the mechanisms w^hich respond acutely and which need investigation in
relation to "shock disease" and the sudden and mass mortality associated
therewith, as well as in relation to those sudden deaths, resembling hypo-
glycemic shock, which occur on first placing strange mammals together.

The development of techniques to measure the secretion of the catechol
amines has led to a number of investigations on the secretion of epinephrine
and norepinephrine in response to a variety of stimuli. One can almost pre-
dict which of these two amines will be secreted in response to a particular
stimulus by knowing which has the greater effect on blood sugar or on blood
pressure. Norepinephrine appears to be released preferentially by the adre-
nal medulla during rest (Gaddum and Holzbauer, 1957). The plasma
concentration of norepinephrine rises sharply with acute muscular work,
but the response of epinephrine varies from no change to a marked rise,
depending on the individual (Gray and Beetham, 1957). Both return to
normal levels within 15 to 30 minutes after cessation of work. Hypoglyce-
mia is followed by a marked and sharp rise in the medullary secretion of
epinephrine with a much less marked rise in norepinephrine (Gaddum and
Holzbauer, 1957; Goldfien ct al., 1958). Infusion of glucose promptly re-
turns their secretion to normal levels. Repeated production of hypoglyce-
mia with insulin eventually leads to a decline in the secretion of epinephrine,
evidently due to medullary exhaustion (Elmadjian ct al., 1958). Hypoten-
sion produces a marked rise in the secretion of norepinephrine, but little or
no rise in epinephrine (Elmadjian et al., 1958). Surgical shock or a change
in position from recumbent to standing leads to a sharp rise in the secretion
of norepinephrine with or without a rise in epinephrine. (Elmadjian et al.,
1958). Tense, anticipatory but passive emotional situations produce a
marked rise in the secretion of epinephrine, norepinephrine being secreted
in normal amounts, but active, aggressive emotional situations are related
to a rise in norepinephrine (Elmadjian et al., 1958) . If the emotional display
is intense enough, both epinephrine and norepinephrine are elevated. It is
of particular interest that in adrenalectomized patients the secretion levels
of norepinephrine and their diurnal variations are completely normal.

2. Endocrines and Populations 227

indicating that the normal daily secretion of norepinephrine is largely from
extra-adrenal chromaffin tissues, whereas the secretion of epinephrine is
largely from the adrenal medulla. These facts are particularly pertinent to
investigations of the physiologic responses of mammals to sociopsychologic
factors.

The secretion of epinephrine and norepinephrine evidently is controlled
by separate hypothalamic centers (Elmadjian et al., 1958). The secretion
of norepinephrine apparently is related to an area in the posterior hypothal-
amus and that of epinephrine to a lateral area. However, the vasomotor
center is in the region of the floor of the fourth ventricle in the medulla
oblongata, and it is the most sensitive area relating to the secretion of
epinephrine (Elmadjian d o/., 1958).

In addition to the actions of epinephrine and norepinephrine listed
above, these compounds of the sympathico-adrenal system have important
activity relationships with other endocrine organs and their hormones. The
medullary hormones, thyroid, and pituitary growth and thyrotropic hor-
mones have a number of interrelated and interdependent actions. We have
already mentioned that growth hormone stimulates adrenal medullary
hypertrophy (Moon ct al., 1950, 1951) . Hypertrophy of the medulla with a
pronounced increase in its epinephrine content also follows chronic poi-
soning of the thyroids of male and female rats with thiouracil (Marine
and Bauman, 1945), whereas chronic nicotine poisoning causes a marked
medullary hypertrophy mainly owing to an increase in the norepinephrine-
containing cells (Eranko et al., 1959). The latter response cannot be
elicited in mice or guinea pigs, although the adrenal medulla of the mouse
has both epinephrine and norepinephrine-containing cells (Eranko et al.,
1959). This difference clearly demonstrates the kind of difference one
may anticipate between species, even as closely related as arc the rat
and mouse (Rattus norvegicus and Mus musculus). The ability of epi-
nephrine to mobilize depot fat and produce a rise in unesterified fatty
acids in intact animals is thoroughly established (Hartman andBrownell,
1949), but its ability to mobilize depot fat, as well as its hyperglycemic
action, seems to depend on the integrity of the adrenal cortex and the carbo-
hydrate-active corticoids (Levy and Ramey, 1958; DeBodo and Altzuler,
1958). These results have led Levy and Ramey (1958) to suggest that the
adrenal steroids and epinephrine may act in concert to regulate metabo-
lism of fat cells. On the other hand, the ability of epinephrine to mobilize
unesterified fatty acids depends on optimal thyroid function, but appar-
ently is unrelated to adrenocortical function (Goodman and Knobil, 1959) .
The mobilization of fatty acids evidently provides a readily available source
of metabolites for the formation of glycogen (Hartman and Brownell,
1949) or for direct utilization by the muscles (Fredrickson and Gordon,
1958).

228 /• /• Christian

It is well known that epinephrine produces a marked rise in calorigenesis
and oxygen consumption in intact animals (Hartman and Brownell, 1949;
Gaddum and Holzbauer, 1957). This calorigenic activity of epinephrine
and norepinephrine is important in adaptation to cold (Hsieh ct al., 1957a,
b) and is potentiated strikingly by thyroxine (Swanson, 1956, 1957) and
further intensified by the action of growth hormone, apparently by the
effect of the latter in increasing thyroid function (Evans ct al., 1958).
Evidently both epinephrine and thyroxine (and an intact pituitary-thyroid
system) are essential for survival in cold exposure and for adaptation to
cold, as, in the absence of thyroxine, epinephrine does not exert its calori-
genic action in rats (Swanson, 1956). It was found that oxygen consump-
tion in thyroidectomized rats increased in proportion to the dose of epi-
nephrine when thyroxine was supplied at a fixed standard dose (Swanson,
1957). On the other hand, oxygen consumption varied linearly with the
log-dose of thyroxine when the animals were kept on a standard dose of
epinephrine (Swanson, 1956). Epinephrine apparently is essential for in-
creased calorigenesis, but requires thyroxine for its activity. Swanson
(1957) has expressed the opinion that since reactivity to epinephrine is di-
rectly and intimately dependent on the level of thyroxine, the main role of
the increased secretion of thyroxine in acclimitization to cold may be to
potentiate the calorigenic activity of epinephrine. In any event, both
epinephrine from the adrenal medulla and an intact properly functioning
pituitary-thyroid system are essential for increased calorigenesis and adap-
tation to cold. The necessity of the adrenal medulla and an intact thyroid
is shown by the fact that the calorigenic response to cold is abolished either
by adrenal demedullation (Morin, 1946a, b) or by thyroidectomy (Swan-
son, 1957) .

It should be obvious from the foregoing discussion that the adrenal
medulla and sympathetic nervous system and their hormones, epinephrine
and norepinephrine, are vital components of a variety of adaptive mecha-
nisms and, if anything, their importance has tended to be underestimated.
It appears from the available evidence that the sympathico-adrenal hor-
mones may play a key role in the physiologic responses to sociopsychologic
factors associated with changes in population density, and therefore deserve
more investigation.

B. The Thyroid Gland

1. Introduction

The thyroid gland and its hormones, primarily thyroxine and to a lesser
extent triiodothyronine, are important components of the internal mecha-
nisms which provide the organism with sufficient physiologic flexibility to

2. Endocrines and Populations 229

be able to maintain a constant internal environment in the face of stimuli
from and changes in the external environment. The thyroid is intimately-
associated with a number of adaptive mechanisms and has important
interractions with the adrenal cortex, as well as with the adrenal medulla
and its hormones. This is not the place for a detailed account of thyroid
physiology, but a brief review of the functions and actions of the thyroid
and its hormones will be given, largely derived from standard accounts
and reviews. Emphasis will be placed on the role played in physiologic
adaptation to environmental changes, especially in response to adverse
stimuli.

It has long been known that thyroid hormone is essential for normal
growth and development, as well as for the normal metabolism of most
tissues. Furthermore it has a vital role in permitting a mammal to adapt
to changes in the temperature of the external envirormient, especially in
adaptation to cold, by acting synergistically with the calorigenic action of
epinephrine, as we have seen in the preceding section, as well as by playing
an important physiological role of its own (Swanson, 1957) . The evidence
that will be discussed subsequently shows that the thyroid also is involved
intimately in the physiological responses to alarming stimuli. Therefore
the thyroid, like the adrenal gland, assumes particular importance in a
discussion of adaptive mechanisms.

2. The Thyroid Hormones and Their Actions

The thyroid hormones, tetraiodothyronine (thyroxine) and Z-3,5,3'-
triiodothyronine, have an overall action of increasing heat production by
increasing the oxidative processes of many tissues and therefore their
oxygen consumption (SoUman, 1957). This metabolic effect is in part
brought about by the augmentation or facilitation of the calorigenic action
of epinephrine, and it has been suggested that thyroxine and epinephrine
influence consecutive rate-limiting reactions in the metabolic cycle, thyrox-
ine acting at a later stage than epinephrine (Swanson, 1956) . The meta-
bolic effect of thyroxine, however, is not exerted equally on all tissues. The
rate of oxidation by brain tissue, for example, is not influenced at all (Tata
et al., 1957), whereas the metabolism of the liver is increased more than of
the body as a whole (Barker and Schwartz, 1953). However, the electro-
encephalogram excitability of the brain, electrolyte distribution, and circu-
lation of the brain are profoundly affected by thyroid hormones (Tata
et al., 1957). Increased thyroid hormone first affects carbohydrate, then
fat, and finally protein metabolism (Sollman, 1957) . Conversely, a defi-
ciency in thyroid hormone reduces the oxidative activity of tissues in
general. It has been suggested that all the actions of the thyroid hormones

230 /. /. Christian

on metabolic processes may reflect a primary action at one biochemical site,
possibly on cytochrome c (Rawson et al., 1955). Sollman (1957) has listed
the following additional effects of thyroxine. Increased levels of thyroid
hormones usually are accompanied by an increased pulse rate, increased
nervous excitability, weight loss, and decreased liver glycogen. Thyroxine
also sensitizes the tissues, especially the blood vessels, the actions of
sympathomimetic compounds such as epinephrine (see above) as well as
to the toxic effects of poisons. The increased sensitization apparently occurs
at the receptor mechanisms. Thyroxine also effects the circulation, but
mainly as a result of increased heat production. Thyroxine has a direct
effect on the heart in increasing its oxygen consumption, but it also has an
indirect effect on the heart and the rest of the circulatory system in the
following way. The delayed, indirect, effect is due to an increased metabolic
demand of the tissues which results in an increase in carbon dioxide and
decreases in oxygen at the arteriolar level. These effects result in a subse-
quent decrease in peripheral resistance, increased venous return, and, via
cardiac reflexes such as the Bainbridge, an increased cardiac output and
increased pulse pressure. The increased heat resulting from increased oxida-
tion must be dissipated, and this is accomplished by dilatation of the vessels
and opening of the arteriovenous anastomoses in the skin and other tissues
with an accompanying increase in the amount of heat loss due to radiation.
The normal calorigenic action of the thyroid hormones is essential for
normal growth, maturation, and tissue differentiation.

Proper functioning of the thyroid and the production of thyroid hormones
is completely dependent on an adequate dietary intake of iodine. Inorganic
iodine is essential for the formation of the thyroid hormone by the thyroid
gland, as the thyroid is incapable of trapping organic iodine (Salter, 1949;
Halmi et al., 1953; Rawson et al., 1955) . The way in which trapped inorganic
iodide and tyrosine are converted into the thyroid hormone (s) has been
critically reviewed by Rawson et al., (1955). Inorganic iodide is trapped
and presumably momentarily oxidized to active iodine in the thyroid epi-
thelium. Pituitary thyrotropic hormone promotes the trapping of iodide
by the thyroid gland, although the gland has some autonomy in this ac-
tivity (Vander Laan and Greer, 1950; Halmi et al., 1953; Vander Laan and
Caplan, 1954; Vander Laan, 1955). The trapped and activated iodine is
then used in converting tyrosine to diiodotyrosine, probably in the presence
of peroxidases and cytochrome oxidases. Two molecules of diiodotyrosine
are condensed to form a single molecule of thyroxine (tetraiodothyronine),
which combines with thyroid globulin (thyroglobulin) and is stored as
such in the colloid of the thyroid follicles. Other iodinated thyronines are
found in the thyroid, but in much smaller amounts than thyroxine. These
probably result from partial iodination and may represent other pathways

2. Endocrines and Populations 231

in the formation of thyroxine than the one given above (Rawson et al.,
1955). However, the metabohc pathway given is the usually accepted
scheme for the synthesis of thyroxine. Thyroxine presumably is released
from thyroglobulin by the action of proteolytic enzymes (SoUman, 1957).
Thyrotropin promotes the iodination of tyrosine and the release of thyrox-
ine from thyroglobulin. A small amount of thyroxine loses one of its
iodine atoms to become /-.3,5,o'-triiodothyronine in the thyroid, probably
more in the peripheral tissues, and small amounts of this compound are
found in the thyroid gland and in circulation (Gross, 1955). /-3,5,3'-
Triiodothyronine has a more pronounced action on oxidative processes
than thyroxine, and its action is much more rapid, but less prolonged
(Sollman, 1957). It has been postulated that the triiodothyronine provides
the immediate thyroid response and is the substance which enters the cells
and exerts the ultimate thyroid action, whereas thyroxine is the circulating
form of the thyroid hormone (Gross, 1955). The differential distribution
of these two hormones in the circulation and in the cells of various tissues
supports this hypothesis (Gross, 1955). Since inorganic iodide is essential
for the normal functioning of the thyroid gland, the rate at which radio-
active iodine is trapped and accumulated by the gland is a good index of
thyroidal activity provided there is not an increased renal excretion of
iodine. The rate of release of radioactive iodine from the thyroid is con-
sidered a more reliable and reproducible index of thyroid activity, as each
animal can serve as its own control in experimental procedures (Brown-
Grant et al., 1954a).

3. Factors That Regulate the Activity of the Thyroid

The secretory activity and hormone synthesis by the thyroid is largely
under the control of the protein hormone thyrotropin (TSH) secreted by
the basophils of the anterior pituitary (D'Angelo, 1955). However, the
thyroid is capable of a low level of autonomous function without stimulation
by the pituitary (Brown-Grant et al., 1954a). The rate of release of TSH
from the anterior pituitary is apparently related to the level of circulating
thyroid hormone (D'Angelo, 1955), although the mediation of the hypo-
thalamus appears to be required (Harris and Woods, 1958; D'Angelo and
Traum, 1958) ; therefore a low level of circulating thyroid hormone stimu-
lates an increased secretion of TSH and a high level inhibits its secretion.
This is the classic concept of an endocrine feedback mechanism which
operates to regulate the release of hormone from the target gland and to
maintain a more or less constant level of circulating hormone under normal
circumstances. However, there is recent evidence to suggest that the secre-
tion of TSH may not be a direct response to the level of circulating thyrox-

232 /. /. Christian

ine, but rather to one or more of its peripheral metabohc actions (Goldberg
et al., 1957) . The secretion of thyrotropin from the pituitary is undoubtedly
under the control of the hypothalamus (D'Angelo and Traum, 1958;
Harris and Woods, 1958, Harris, 1959). Like the secretion of adrenocorti-
cotropin, the secretion of thyrotropin appears to be controlled by a humoral
factor from the hypothalamus (Brown-Grant et al., 1957). Section of the
pituitary stalk leads to a loss of the inhibitory response of the thyroid to
restraint or pain, and regeneration of the pituitary portal vessels is accom-
panied by a return of this response (Harris, 1955a, b; Reiclilin, 1957a, b).
The thyroid remains at least partially responsive to exogenous TSH when
the pituitary stalk is sectioned (Reichlin, 1957a, b; Harris and Woods,
1958). There is also evidence that the hypothalamus is necessary for a
decreased blood concentration of thyroid hormone to effect an increased
secretion of TSH (Greer, 1951; 1952; Bogdanove and Halmi, 1953; Harris
and Woods, 1958; D'Angelo and Traum, 1958). Increased activity of the
thyroid accompanies electrical stimulation of the anterior portion of the
median eminence, but not other parts of the hypothalamus (Harris and
Woods, 1958). The hypothalamic factor responsible for stimulating the
release of TSH does not appear to be contained in or associated with
Pitressin (Reichlin, 1957a; D'Angelo and Traum, 1958), but little else is
known of the nature of this factor. In addition to these regulating factors,
the release of TSH is also affected by cortisone and ACTH (Brown-Grant
et al., 1954c) . Cortisone and ACTH inhibit the release of radioiodine from
the thyroid of rabbits and rats, probably by inhibiting the release of TSH,
although the mechanism by which this is accomplished is unknown (Brown-
Grant et al., 1954b, c; Brown-Grant, 1955). Cortisone and ACTH have
also been reported to inhibit the uptake of radioiodine by the thyroid of
rats (Money et al., 1950; Albert et al., 1952; Perry, 1951; Verzar and Vido-
vic, 1952) and of humans (Perry, 1951; Kuhl and Ziff, 1952; Albert et al.,
1952; Berson and Yalow, 1952). However, reports of decreased uptake of
radioiodine must be interpreted with caution, as it may reflect a lowered
concentration of circulating iodide as a result of increased renal clearance
of iodide following treatment with cortisone or ACTH (Brown-Grant et al.,
1954c) . Adrenalectomy inhibits the release of iodine and therefore of
thyroid hormone from the thyroid glands of rats (Fllickiger and Verzar,
1955). Bastenie and Ermans (1958) have shown that cortisone, in addition
to its effect on the secretion of TSH, inhibits the stimulating effect of
thyroxine on oxygen consumption and phosphorus turnover but fails to
inhibit these actions of triiodothyronine. These authors concluded that
cortisone inhibits the peripheral degradation of thyroxine to triiodothyro-
nine. Nevertheless, there is little doubt that cortisone and ACTH do inhibit

2. Endocrines and Populations 233

thyroid activity, probably by inhibiting the release of TSH from the pitui-
tary.

Epinephrine, in addition to its synergism with thyroxine with respect to
calorigenesis, appears to have more direct effects on thyroid activity, al-
though the reports are conflicting. Badrick and his co-workers (1954, 1955)
reported that epinephrine inhibited the uptake of radioiodine by the thy-
roids of intact and hypophysectomized rats, and similar results were re-
ported for intact rats by Money et al. (1950). Brown-Grant et al. (1954b)
found that epinephrine decreased the release of radioiodine by the thyroids
of intact rabbits. On the other hand, epinephrine has been reported to in-
crease the release of radioiodine by the thyroid (Williams et al., 1949) and
to increase the content of radioiodine in the thyroid (Botkin and Tew,
1952), although the latter could result from an inhibition of its release.
More recently Ackerman and Arons (1958) have reported that epinephrine
increases the release of radioiodine from the thyroids of intact and hypo-
physectomized dogs. These divergent results are difficult to reconcile, al-
though the fact that epinephrine acted in hypophysectomized as well as in
intact rats and dogs (Badrick et al, 1955; Ackerman and Arons, 1958)
suggests that the effects were directly on the thyroid gland. It is possible
that these varying results may reflect the effects of epinephrine on the
thyroidal vasculature in different animals and different circumstances, as
was suggested by Badrick et al. (1955) .

From the foregoing account it should be clear that the regulation of
thyroid function is complex and involves a number of factors. However,
there seems to be little doubt that the hypothalamus has a relatively basic
role in regulating the secretion of thyrotropin from the pituitary, and evi-
dently is capable of inhibiting its release (Brown-Grant et al., 1954b; Reich-
lin, 1957a; Harris and Woods, 1958; D'Angelo and Traum, 1958) . D'Angelo
and Traum (1958), in an extensive series of experiments on hypothalamic-
hypophyseal-thyroidal relationships in rats, concluded that the hypothala-
mus functions as a modulator, modifying the production and release of
thyrotropin from the adenohypophysis not only under conditions of in-
creased demand, but also in meeting day-to-day needs. Normal blood levels
of TSH cannot be maintained in the absence of stimulation by the hypo-
thalamus even with decreased levels of thyroid hormone. However, the
adenohypophysis keeps its capacity to produce and release TSH, but regu-
lated by the level of circulating thyroid hormone or by peripheral meta-
bolic actions thereof (cf. above). The action of thyroid hormone may be
directly on the pituitary (D'Angelo and Traum, (1958). The injection of
minute amounts of thyroxine into the hypothalamus is followed by a rela-
tively long latent period before the release of TSH is inhibited, whereas

234 /. J. Christian

injection into the anterior pituitary is followed by a more immediate re-
sponse (Yamada and Greer, 1959; Yamada, 1959).

The thyroid was blocked in these studies (D'Angelo and Traum, 1958) by
hypothalamic lesions or by treatment with propylthiouracil. Thyroid func-
tion was evaluated by the rate of turnover of radioiodine, bioassay of the
pituitaries and plasma for their content of TSH and thyroid hormone, and
was finally correlated with the results of detailed histologic studies of the
glands of experimental and bioassay animals. The results appear to be
definitive. These workers were further able to demonstrate the independ-
ence of ovarian, adrenocortical, and thyroid-regulating parts of the hypo-
thalamus. How the adrenal corticoids and adrenocorticotropin fit into the
regulation of thyroid function is unknown at present.

4. Thyroid Responses to External Stimuli

Probably the best-known action of the tl\yroid is its ability to increase
the secretion of thyroid hormone in response to cold and to decrease its re-
lease in response to heat (Ring, 1942; Brown-Grant et at., 1954a; Stevens
et al., 1955; Hellman and Collins, 1957). The gland also becomes hyper-
plastic with prolonged exposure to cold (Money, 1955). Upon exposure to
low temperatures, there is an increased release of thyroid hormone from
the thyroid gland (Brown-Grant et al., 1954a; Woods and Carlson, 1956),
apparently in response to increased secretion of TSH by the pituitary
(Stevens et al., 1955) . However, Ring (1942) found that the thyroid per se
can account for only a small part of the increase in heat production required
to meet the demands of mice exposed to 2-4° C, and that epinephrine and
thyroid hormone are required for a maximum continued metabolic response
to cold. He therefore concluded that the principle action of thyroid in re-
sponse to cold exposure was to sensitize the animal to the calorigenic action
of epinephrine. Swanson (1956, 1957) essentially confirmed and extended
Ring's findings and conclusions, as we have seen above. Thyroidectomy
abolishes the response to cold and deprives the animal of the ability to
survive exposure to cold (Swanson, 1957).

Exposure to cold increases pituitary-adrenocortical as well as thyroidal
activity, contrary to the more usual reciprocal relationship between ACTH
and TSH secretion by the pituitary described below. Increased thyroidal
activity apparently is the more vital of these two adaptive responses to
cold. Thyroxine permits survival of hypothyroid rats subjected to cold,
whereas adrenocorticotropin and adrenocortical extract do not (Freedman
and Gordon, 1955). Similarly, minute amounts of thyroxine permitted
100% survival of rats treated with thiouracil upon exposure to cold,
whereas adrenocortical extracts failed to increase survival (Ershoff, 1948).

2. Endocrines and Populations 235

Heat, on the other hand, diminishes thyroid activity (Hellman and Collins,
1957) . Rats and mice have been shown to have reduced thyroid activity in
the warm summer months (Hurst and Turner, 1947) . Puntriano and Meites
(1951) suggested that the seasonal changes in thyroid activity might be
due to changes in day length rather than to seasonal changes in tempera-
ture, on the basis of their findings that prolonged exposure to light inhibits
thyroid activity in mice and exposure to darkness increases thyroid activity.
However, these results could not be confirmed in rabbits (Brown-Grant
et al., 1954b) . The rate of release of radio iodine by the thyroids of rabbits
was unaffected by prolonged exposure either to darkness or to light. There-
fore the validity of the hypothesis that day length affects thyroid activity
is dubious.

In contrast to the thyroidal-adrenocortical relationship in response to
exposure to cold, there is abundant evidence indicating that the activity of
the thyroids is inhibited in response to alarming stimuli which evoke an
increased pituitary-adrenocortical activity. In fact it has come to be
generally accepted that there is a reciprocal relationship between ACTH
and TSH secretion in response to "stress" (Harris, 1955a). Anoxia, neph-
rectomy, vitamin deficiencies, tourniquet shock, fasting, injected typhoid
vaccine, swimming in cold water (15° C), injections of formalin, spinal
cordotomy, electroshock, and gastrointestinal and peripheral trauma all
inhibit the activity of the thyroid glands of rats (Williams et al, 1949;
Paschkis et al, 1950; Bogoroch and Timiras, 1951; Hamolsky et al, 1951;
Van Middlesworth and Berry, 1951; Badrick et al, 1954, 1955). Similar
results were obtained with rabbits, using the release of radioiodine as the
basis of measurement (Brown-Grant et al, 1954b) . Hemorrhage, anesthesia,
laparotomy, intraperitoneal injections of turpentine, draining subcutaneous
abscesses, or emotional stress in the form of restraint, subcutaneous faradic
stimulation, or sudden changes from light to dark or vice versa were all
found to inhibit thyroid activity in rabbits for 1-2 days. After this time
the rabbits evidently became used to the procedures and were no longer
alarmed by them, as thyroid function gradually returned to normal with
continued exposure to these emotional stimuli. In some instances of these
experiments the release of radioiodine from the thyroid was totally inhi-
bited (Brown-Grant et al, 1954b). Emotional stimuli also were followed
by a time lag of somewhat less than 3 hours between their first application
and the first thyroid response. These experiments demonstrate that emo-
tional factors, visual stimuli, and the central nervous system can affect
thyroid function, presumably via the hypothalamus, just as they can in-
crease the secretion of ACTH or adrenal medullary hormones (Brown-
Grant et al, 1954b). Kracht (1954) described a "true thyrotoxicosis" in
wild rabbits stimulated by fear in response to being chased by ferrets and

236 /• /• Christian

stated that this reaction was a "model of a thyrotrophic alarm reaction."
His evidence was, for the most part, indirect and his conclusions that there
was increased thyroid activity on the basis of increased renal excretion of
injected radioiodine is untenable, especially for animals with a marked
increase in adrenocortical activity. Finally, Brown-Grant et al. (1954b)
found that the release of radioiodine by the thyroids of wild rabbits was
inliibited by emotional stress, just as it was in the usual domestic rabbit.
In view of these facts, the concept of a "fright thyrotoxicosis" is inaccepta-
ble. The inhibition of thyroid activity by emotional stress was not as
consistent in adrenalectomized as it was in intact rabbits, otherwise their
responses were similar. There is little doubt that the inhibition of thyroid
activity in these experiments was due to a decrease in the secretion of
thyrotropin (Brown-Grant et al., 1954a, b). It is true that cortisone and
ACTH inhibit the release of TSH from the pituitarics of rats (Brown-
Grant, 1955) and rabbits (Brown-Grant ct al., 1954a, b; Harris and Woods,
1958), but the adrenal corticoids cannot be responsible for the major part,
if any, of the inhibition of the release of thyrotropin in the above experi-
ments, as adrenalectomy did not affect the inhibition of the thyroid, and
cortisone was without effect on the release of radioiodine by the thyroids
of rabbits injected with TSH. Finally, injected TSH abolishes the inhibition
of the thyroid produced by ACTH and cortisone (Harris and Woods, 1958) .
There has been some confusion in the literature about the effect of
alarming stimuli on the thyroid. The experiments which have been dis-
cussed so far indicate that the thyroid is depressed under these circum-
stances, but it is also well known that emotional factors can precipitate
thyroid crises in humans (Selye, 1950). These apparently contradictory
experiences may depend on species differences, dose-time relationships, or
both, as the recent work of Gerwing et al. (1958) has indicated. These
investigators have shown that thyroid activity in mice, rats, and rabbits
subjected to chronic toxic "stress" (injection of small doses of bacterial
endotoxins repeatedly for 24 days) is inhibited at first, but returns to
normal and eventually exceeds the original normal level as the stimulation
continues (Gerwing, 1958; Gerwing et al, 1958) . On the other hand, guinea
pigs and rhesus monkeys exhibit increased thyroid activity from the be-
ginning. Injected thyrotropin stimulates the thyroids of mice, rats, and
rabbits with inhibited thyroids following injection of the toxin, indicating
that the depressed thyroid function was due to a diminished secretion of
TSH. It may be coincidental, but it is nevertheless interesting that these
differences in thyroid function between species in response to bacterial toxin
coincide with the differences in the secretory patterns of the major corti-
coids in these same species (cf. above) : rats, mice, and rabbits secrete pri-
marily corticosterone, and the other two species, hydrocortisone. These

2. Endocrines and Populations 237

investigators (Gerwing, 1958; Gerwing et al., 1958) conclude that the
hypothesis of a reciprocal relationship between the secretion of TSH and
ACTH by the anterior pituitary is compatible with the evidence for rats,
mice, and rabbits, but not for guinea pigs and monkeys. They suggest that
in the latter two the inhibitory effect of increased corticosteroids does not
occur. However, we have seen that the corticosteroids probably have no
effect in the inhibition of TSH release and thyroid function in rats and
rabbits in response to alarming stimuli; so that some other explanation
must be sought. Man is probably similar to guinea pigs and rhesus monkeys
in the above responses of the thyroid (Gerwing et al., 1958) . These studies
emphasize the importance of careful comparative work as well as the im-
portance of dose-time relationships in physiologic functions in different
species of mammal.

Starvation causes a profound depression of thyroid activity, probably
by suppressing the secretion of TSH, as it has been shown that acute starva-
tion markedly depresses the blood level of TSH in mice and rats (Monej'-,
1955). However, if starved animals are subjected to cold, the degree of
thyroid depression is inversely related to the degree of the reduction in
environmental temperature (Reichlin, 1957a). The thyroidal iodine release
rates at the lower temperatures, even though reduced, are appreciably
higher than they are in starved animals at higher temperatures. The mainte-
nance of body temperature appears to take precedence over the conserva-
tion of nutritive reserves and tissues in the regulation of thyroid activity
activity (Reichlin, 1957a).

In contrast to some of the above relationships, increased thyroidal ac-
tivity increases adrenocortical activity in white rats (Wallach and Reineke,
1949). Administration of thyroxine decreases adrenal ascorbic acid to
minimal levels after 2-4 days and is followed, upon continued treatment
with adequate doses of thyroxine, by a progressive increase in adrenal size
and ascorbic acid content which reaches maximum in 4 weeks. The in-
crease in adrenal weight in these circumstances is roughly proportional to
the dose of administered thyroxine. There is a narrow dose range of thyrox-
ine in which there is no effect on the adrenals. Perhaps this dose range repre-
sents the normal physiologic daily secretion rate. It may be that 'the
adrenal hypertrophy observed in rats exposed to cold is in part a result of
the increased thyroid activity. On the other hand, adrenocortical function
does not seem to depend on the presence of a normally functional thyroid,
as thyroidectomy does not cause adrenal atrophy in rats although the zona
fasciculata is reduced in width (Hess, 1953), nor does thyroidectomy de-
crease the ability of the adrenals to respond to stress (Hess and Finerty,
1952) . The adrenals of rats treated with antithyroidal compounds (thioura-
cil) also (were) responsive to injected adrenocorticotropin (Freedman and

238 /. J. Christian

Gordon, 1955), although others report adrenal atrophy following blockade
of the thyroid with antithyroid compounds (Seifter et al., 1949). Finally,
thyroxine augments the adrenal hypertrophy produced by growth hor-
mone, although it does not augment the adrenal hypertrophy produced
by ACTH (Bois and Selye, 1957).

There are many other factors which affect thyroid function. A high
dietary salt intake can increase thyroid activity, even with a reduced iodine
intake, in a rather circuitous way (Isler et al., 1958). The mechanism ap-
pears to be that with increased NaCl intake, and consequently an increase
in the excretion of NaCl, there is apparently an increase in the excretion of
iodine with the sodium chloride. The loss of iodine produces a low serum
concentration of iodine and secondarily a low output of thyroxine. The low
thyroxine in turn stimulates increased secretion of TSH from the pituitary
and a resulting stimulation of the thyroid gland.

5. Thyroidal-Gonadal Ixterrelatioxships

In addition to its more or less reciprocal relationship with the adrenal
cortex, the thyroid, and therefore thyroidal adaptive responses, has im-
portant effects on reproduction and the reproductive organs. These effects
are of particular interest here, because they may occur in response to
alarming stimuli and must be evaluated as possible factors producing the
changes in reproductive function seen in mammals, especially in relation
to the density of the population. A thyroidal-gonadal relationship is indi-
cated by the tendency for goiters to occur at puberty, during pregnane}^,
or during estrus; also the serum protein-bound iodine is elevated during
pregnancy, and stilbesterol increases the release of iodine by the thyroids
of rats and mice (Sollman, 1957) . There is considerable evidence indicating
that hypothyroidism is associated with abnormal ovarian function in
humans (Rawson et al., 1955). Cold exposure causes a marked increase in
the length of the estrous cycle in rats which is prevented by administering
thyroxine (Dempsey and Uotila, 1940; Denison and Zarrow, 1955) . Thyrox-
ine also augments the recover}^ of function of the testes of immature rats
from the atrophic effects of starvation (Horn, 1955) . The recovery of the
testes in this case is not related to the recovery in body weight, and the
recovery of the sex accessories following treatment with thyroxine was only
partial and required a greater period of time. jMaqsood and Reineke (1950)
found that high temperatures (30° C.) decreased the weights of the testes
and seminal vesicles of mice, but the administration of thyroprotein in-
creased the weights of these organs at that temperature. Alild hj^perthy-
roidism stimulated, and hypothyroidism inhibited, sexual development in
growing male mice (Maqsood and Reineke, 1950). Higher doses of thyroid

2. Endocrines and Populations 239

hormone decreased the weights of the reproductive organs. However, thy-
roidectomy was followed by an increase in the relative weights of the testes
and seminal vesicles of rats (Hess, 1953). Hess (1953) concluded that
thyroidectomy appears to sensitize the gonads and secondary sex organs
to the actions of gonadtropins. Hyperthyroidism is associated with ab-
normal reproductive function in female rats (Denison and Zarrow, 1955).
Prolonged exposure to 2° C. resulted in a marked increase in the length of
the estrous cycle: proestrus and estrus were markedly prolonged, while
metestrus and diestrus were shortened. Treatment with 50 fxg. of thyroxine
corrected these changes. On the other hand, brief (3 days) treatment of
female rats with thyroid hormone just prior to mating apparently resulted
in an increase in subsecjuent litter size, but the same amount of thyroid
hormone given during hot weather or an increased amount of thyroid
hormone adversely affected reproduction (Kraatz, 1939) . These results
seem to indicate that there is an optimal level of thyroid function for proper
functioning of the reproductive system, and that any appreciable deviation
from this level in either direction leads to diminished reproductive function.

The role of the thyroid in reproduction has been investigated by using
radiothyroidectomized mice (Bruce and Sloviter, 1957). Radiothyroidec-
tomy apparently had no effect on male fertility, female fertility, or litter
frequency, but estrus and gestation were prolonged. Litter size possibly
was reduced owing to increased resorption of embryos, but the matter
requires further investigation. In contrast to these results, Chu (1944)
found that surgical thyroidectomy in rabbits during early pregnancy re-
sulted in total loss of fetuses due to resorption or abortion. The same opera-
tion in later pregnancy resulted in the young being stillborn. Induction of
pregnancy in thyroidectomized rabbits was followed by resorption or
abortion of the embryos or a prolongation of gestation, the young subse-
quently being born dead. Feeding desiccated thyroid tended to prevent
the adverse effects of thyroidectomy (Chu, 1944) . Hypothyroidism induced
by thiouracil in rats apparently was without effect on fertility in males and
females but interferred with gestation in pregnant females, causing resorp-
tion in 100% of the animals (Jones et al., 1946). Thiouracil treatment of
rats for shorter periods of time, less than 100 days, did not prevent some
of the rats from delivering litters, but the young exhibited abnormal
growth and development. It is quite evident that pregnancy is adversely
affected by severe hypothyroidism with a marked increase in intrauterine
mortality.

In general reproductix'e function is altered by marked increases or de-
creases in thyroid activity, and it seems that the depression in thyroid
activity in response to alarming stimuli must play a role, along with in-
creased adrenocortical activity, in the commonly observed suppression of

240 /. /. Christian

reproductive function. However, a great deal more investigation is neces-
sary to clarify thyroidal-adrenal-gonadal relationships.

6. Other Thyroid Relationships

In addition to its relationship to the adrenals and gonads and to its
general metabolic effects, the thyroid hormones play specific roles in the
maintenance of other organs, and these may offer means of evaluating
thyroid function in intact mammals in the field, providing they are used
in conjunction with other information. For example, the harderian glands
of rats and mice are maintained by thyroid hormones and as yet unknown
pituitary factors (Boas and Bates, 1954; Hellman and Collins, 1957), al-
though the thyroid may not be essential for the maintenance of these
glands in guinea pigs (Smelser, 1943). Thyroid hormone, in conjunction
with androgens, is required for the maintenance of the size and granulation
of the cells of the serous tubules of the submaxillary glands of male rats
(Grad and Leblond, 1949) . The thyroid alone restores the number of these
cells to normal, but does not affect the size or granulation.

Thyroid function in general declines moderately with age and increased
size (Hurst and Turner, 1947), but there are no marked effects of this
decline on growth or other functions.

It should be clear from the above account that the thyroid and its
hormones participate actively in the adaptive responses and probably play
a significant role in many aspects of these responses. Investigations on the
reactions of the thyroid with changes in the size of mammalian populations
and its relationship to adrenal and reproductive functions are needed.

C. Other Endocrine Adaptive Factors

1. Pancreatic Islets

A discussion of the physiology of the pancreas with respect to glucose
metabolism is not properly the function of this chapter, but it should be
mentioned as one of the important endocrine organs regulating glucose
utilization. The hormone best known of the islets of Langerhans is insulin,
which is derived from the beta cells. More recently glucagon, presumably
from the alpha cells, has been described. Insulin, in brief, increases the
utilization of glucose and in many of its actions is antagonized by cortisone.
Glucagon, on the other hand, is a hyperglycemic factor. The function of
these hormones wdth respect to the generalized adaptive responses has been
little studied, but they certainly must play an important part in the general
economy of the mammal under adverse circumstances. Anything affecting

2. Endocrines and Populations 241

the mobilization and utilization of glucose is inevitably going to have an
important bearing on the adaptive responses when a more or less final
picture is available of the entire scheme of responses. The frequency of
diabetes in, and its effect on, natural populations is unknown, although
recent work indicates that under certain circumstances it is far from being a
rare disease.

2. Posterior Pituitary

The posterior pituitary, or neurohypophysis, has been mentioned re-
peatedly in the discussions on the regulation of the secretion of the hor-
mones of the anterior pituitary, but in addition the posterior pituitary has
important functions of its own in physiologic adaptation. A lengthy discus-
sion will not be indulged here, as these functions more properly come under
the purview of discussions elsewhere or are insufficiently well known to
warrant elaboration in detail.

a. Antidiuretic hormone. The posterior pituitary secretes an antidiuretic
factor (ADH) which is important in regulating the reabsorption of water
from the renal tubules. Section of the pituitary stalk with a resultant
denervation of the posterior pituitary is followed by diabetes insipidus as a
result of the lack of antidiuretic hormone. ADH is apparently manufac-
tured in the hypothalamus and traverses the fibers of the supraopticohy-
pophyseal tract and is released into the systemic circulation in the posterior
pituitary. A vasopressor activity is associated with the antidiuretic factor
of the pituitary, but the neurohypophyseal hormone is much less effective
in this activity than in its antidiuretic action. There is evidence that the
secretion of ADH is increased in response to alarming stimuli as well as to
dehydration, but there is also evidence that its secretion is independent of
the release of ACTH from the anterior pituitary (cf . above) . For a more
detailed account of the pituitary antidiuretic hormone, the reader is re-
ferred to the recent review by Thorn ( 1958) .

h. Lipid mobilizing factor. The recent work of Seifter and his colleagues
( 1959) has indicated the existence of a posterior pituitary lipid-mobilizing
factor (LMF) which is released in response to adrenal corticoids. The
hormone appears to be a peptide capable of mobilizing triglycerides from
the mesenteric fat depots following injections of cortisone, exposure to cold,
or subjection to other stimuli which induce an increased adrenocortical
secretion. The lipid mobilization is blocked by adrenalectomy or hy-
pophysectomy. The pathway for lipid mobilization by LMF in response to
alarming stimuli appears to be anterior pituitary, adrenal cortex, and
posterior pituitary (Seifter et al., 1959).

242 /• J- Christian

3. Other Gland Systems

Many other glands, organs, and organ systems participate in physio-
logic adaptation, either directly or in response to increased adrenal or
other hormones or diminution thereof. The essential integrative role of
the central nervous system has been implicit throughout the foregoing
discussion. The gastrointestinal tract and its appendant organs such as the
pancreas, liver, and salivary glands also actively participate in adaptation
(Ehrich and Seifter, 1948; Selye, 1950; Baker and Bridgman, 1954; Baker
and Abrams, 1954). The responses of the stomach to the adrenocortical
hormones and to "stress" have been reviewed recently by Gray and Ramey
( 1957) and will not be discussed further here. One example of the participa-
tion of the digestive organs in the adaptive responses is the response of the
serous glands to increased adrenocortical secretion, first, with a loss of
zymogen granules followed by a depletion of cytoplasm nucleic acids
(Ehrich and Seifter, 1948) . However, as most of these reactions are second-
ary to the increased activity of the pituitary and adrenal cortex, it does
not seem appropriate here to dwell on them in detail since the entire organ-
ism responds to one degree or another to the physiologic alterations subse-
quent to increased activity of the primary adaptive mechanisms.

D. General Measurements of the Endocrine Adaptive Responses

1. General

The actions of and responses to specific hormones and adaptive mecha-
nisms have been discussed, but often there are responses in the intact ani-
mal that cannot be attributed to a particular hormone or system. The
splenic hypertrophy and enlargement of the nucleus pulposus which is seen
in mice and voles are examples of such responses (Clarke, 1953 ; Chitty et al.,
1956; Christian, 1959c). The lymphoid tissue of the spleen is involuted by
the adrenal glucocorticoids, but the resultant atrophy is more than offset
by increased hematopoiesis in the spleens of mice and voles subjected to
social stress or other alarming stimuli (Dawson, 1956). However, it is not
known at present what specifically is responsible for the increased hemato-
poiesis. This seems to be the proper place to review the physiologic adaptive
reactions and to do it in terms of these responses which are found in the
intact animal in response to specific stimuli. Wherever possible, the re-
sponses will be related to the specific mechanisms responsible for their
occurrence.

It is appropriate first to discuss the stimuli which are known to elicit
endocrine adaptive responses. These stimuli have typical dose-time-re-
sponse relationships: the severer or the longer the stimulation, the greater

2. Endocrines and Populations 243

the response (Sayers and Sayers, 1949). Nevertheless, it should not be
assumed that all adaptive responses are qualitatively similar. Some stimuli
may elicit quite similar responses in kind and degree, whereas others will
be manifested in qualitatively and quantitatively different manners. This
statement should be apparent from the preceding accounts of the mecha-
nisms involved. All have in common that if uncompensated they will pro-
duce widespread physiologic changes which come under the heading of
shock (Selye, 1950). However, whether or not the symptoms commonly
associated with shock are elicited depends on the severity of the stimulus
(Selye, 1950). In short, almost any change in the external environment of
the animal, physical damage to the animal, or emotional upset is po-
tentially capable of producing profound deleterious effects on the animal,
but within reasonable limits the adaptive mechanisms provide it with the
flexibility to accommodate most changes or adverse stimuli. It is not until
the latter reach rather serious proportions and evoke marked responses that
we customarily consider the animal to be subject to "stress" or that the
stimuli are labeled "alarming."

These stimuli act through or are received through a number of receptor
pathways in the host. For example, physical damage, including burns,
surgery, acute physical trauma of all sorts, will evoke adaptive responses
probably by direct neural pathways as well as via unknown chemical
mediators, resulting from tissue damage which may stimulate the release
of ACTH or other adaptive responses directly. Hemorrhage, by decreasing
blood volimie, can effect the release of aldosterone directly. Many stimuli
act directly through the central nervous system, such as light, noise, fear,
and rage. Heat, cold, and similar environmental factors act through recep-
tors in the individual and produce physiologic responses directly. Regardless
of the receptor or pathway, all these primary environmental stimuli, if
severe enough or chronic enough, have the ability to produce the symptoms
of shock through several common pathways. In so doing they produce
physiologic changes, either directly or through the central nervous system,
which in turn activate a chain of responses leading to defensive reaction
against any adverse physiologic shift which may have been produced bj^
the initial stimulus. The variety of stimuli which have been found to elicit
adaptive reactions is legion (Selye, 1950). In fact, one could conclude that
almost every experience encountered in daily life by any mammal, if severe
enough, is capable of producing shock and evoking marked adaptive re-
sponses, but for the purpose of the present discussion, the importance of
emotional stimuli in evoking these reactions will be emphasized. These
may be acute stimuli, and all of us are personally familiar with some of the
reactions that they are capable of producing. However, it should be noted
that emotional stimuli of a more chronic nature can produce profound

244 /. /. Christian

long-term physiologic reactions which may have widespread effects on the
host (Elmadjian ct at., 1958; Ratcliffe and Cronin, 1958; Christian, 1959b;
Mason, 1959) . Close confinement, for example, can result in atrophy of the
adaptive mechanisms with a resultant marked susceptibility to any subse-
quent alarming stimulus (Christian and Ratcliffe, 1952). Social competi-
tion, social pressure, or chronic anxiety have been shown to produce marked
physiologic responses (Christian, 1959a; Mason, 1959) Emotional factors,
of necessity, have their begmning in the central nervous system as a result
of sensory stimuli received from the external environment via the visual,
aural, and olfactory sensory receptors. Therefore they are part of the com-
plex system responsible for integrating the mammal with its immediate
environment.

Whatever the particular adverse stimulus may be and regardless of the
pathway of its reception by the animal, whether sensory, traumatic, or
otherwise, all stimuli have the ability to produce shock and evoke physio-
logic adaptive responses in the mammal. Shock is a vaguely defined entity,
and there is no general agreement with regard to its exact nature or cause
(see Selye, 1950; Scudder, 1952; Bing, 1952; Zweifach, 1952; Engel, 1952;
Agate, 1952; Randall, 1952). Moreover, its primary effects and the re-
sponses to it are often difficult to separate. Fundamentally the principal
manifestations of shock involve changes in the circulatory system leading
to inadequate circulatory function and eventually circulatory collapse. If
one keeps in mind that the fundamental changes seem to be blood sludging
(Knisely et at., 1947), changes in selective permeability of the peripheral
vascular bed (Zweifach, 1952), capillary and arteriolar atony and hypo-
reaction of the vascular musculature to epinephrine and even direct stimu-
lation with a resultant visceral vasodilatation (Zweifach, 1952), then the
secondary effects of hemoconcentration, hypotension, fall in blood volume,
decreased cardiac output, and increased hematocrit are more easily under-
stood. In general shock seems to involve a generalized failure of circulatory
integrity. With a superimposed relative circulatory stasis and hypoxia,
these alterations became more profound and the whole process becomes a
vicious circle (Zweifach, 1952) . It is not known by what specific means
these alterations in circulation are effected, but that they involve both
neural and humoral components can hardly be questioned. We have re-
peatedly emphasized that the basic role of the adaptive reactions is to
maintain vascular integrity in opposition to forces tending to destroy it as
well as to prepare the organism to meet emergency situations. Most
alarming stimuli also produce an immediate discharge of the medullary
hormones epinephrine and norepinephrine with mobilization of glucose
reserves, increased heart rate and strength of beat increasing cardiac out-
put, and constriction of visceral and cutaneous blood vessels with a re-

2. Endocrines and Populations 245

suiting shift of the blood to the musculature and central nervous system. If
at the same time the output of ACTH is stimulated, adrenal glucocorticoids
will also be increased, and among their effects is an increase in the reactivity
of the blood vessels to the effects of epinephrine and norepinephrine. None
of the reactions is completely isolated, but the balance between them may
shift considerably depending on the nature of the stimulus. There are also
temporal relationships. In many ways the immediate and short-lived ac-
tivity of epinephrine is mimicked more chronically by the glucocorticoids.
The response of the adrenal medulla is immediate, whereas the cortex re-
sponds somewhat more slowly and can be sustained for a great deal of
time.

It is worth describing some of the measurable changes effected in a
mammal by activation of the adaptive responses. Many of these have been
described earlier under actions of the various hormones, but it is appropri-
ate to discuss these as measurable changes in such a way that they might be
useful in detecting and interpreting the effects of physiologic adaptation to
potentially harmful stimuli. Many of the actions of the adrenocortical and
other hormones serve specifically to restore the equilibrium of the internal
environment after an alarming stimulus. These actions may be a decided
disadvantage to the animal when they are prolonged. For example, in-
creased secretion of the carbohydrate-active corticoids serves to maintain
proper fluid and electrolyte balances and to provide readily available
glucose reserves in an emergency, but they also suppress inflammation,
granulation, and antibody formation and thereby reduce host resistance to
infection. The following list of adaptive responses is by no means complete,
those effects having been selected which might prove useful or of basic con-
ceptual importance to the investigator wishing to study the effects of
physiologic adaptation.

2. Measurements Indicating Increased Adrenocortical Function

a. Adrenal weight. This is a presumptive measurement of ACTH activity
and response to an acute or chronic distress. Weights are the best available
index of adrenocortical activity for many studies. For many long-term
changes, weights are much more useful and much less subject to pitfalls
than techniques such as changes in adrenal ascorbic acid, lipids, cholesterol,
the production of plasma corticosteroids, or circulating eosinophils or
lymphocytes. All these measurements are labile and reflect rapid changes
in adrenal function: they indicate the status of the animal at the moment
of making the measurement, thus they often may reflect nothing more
than the process of trapping or handling an animal and may completely
obscure longer-term changes, especially with excitable, highl}^ reactive

246 J. J. Christian

wild mammals. The problem of adrenal hypertrophy with weight changes
and the relationship of these changes to the length and severity of the
stimulus has been discussed at length by Sayers and Sayers (1949).

Adrenal weight changes largely reflect changes in the zona fasciculata,
which undergoes hyperplasia and hypertrophy following stimulation by
ACTH and therefore is indicative of changes in the secretion of glucocorti-
coids. Under different circumstances, already discussed, the zona glomeru-
losa may hypertrophy, but changes in its size are relatively unimportant
with respect to changes in the total weight of the gland because of its rela-
tively minor contribution to the total mass of the adrenal. The same state-
ment is generally true for the adrenal medulla. Fmally, changes in adrenal
weight, when used as an index of cortical activity, should be evaluated in
the light of histologic studies, as the amount of lipid in the cells of the
fasciculata may vary sufficiently with various functional states to have a
marked effect on adrenal weight. For example, stimulation of sufficient
intensity to deplete the cortex of visible lipids may result in a decrease in
weight whereas activity may actually be greater than in a heavier gland
containing a large amount of stored lipid (Christian, 1959b).

The adrenal medulla in some species may hypertrophy and contribute
to changes in adrenal weight. Although the contribution of the medulla to
adrenal weight changes is usually negligible, it would be appreciable for the
Soricidae, in which the adrenal is composed largely of medulla. Changes in
the medulla and cortex of the long-tailed shrews (soricids) have not been
studied, although these medullary-cortical relationships have been ob-
served in mature individuals of Sorex cinereus, S. dispar, S. fumeus, S.
palustris, and Microsorex hoyi. There may be a relationship between body
size, adrenocortical mass, and metabolic rate in these minute animals.

Adrenal weights are useful for field studies, as they may be obtained on
fresh or fixed material. From a practical standpoint it is almost essential to
obtam adrenal weights from fixed material in mammals the size of mice or
smaller, as it is very difficult to clean the glands properly and obtain
reliable weights on the fresh glands. Rapid water loss from fresh glands of
such small size further complicates the problem. In most circumstances
increased adrenal weight is an acceptable indicator of increased adreno-
cortical activity, but an effort should be made to obtain the adrenals from
suddenly killed specimens, not from captive animals or from those held in a
live trap for extended periods, if one wishes to assess adrenocortical func-
tion in the animal in its natural state. Animals which have been killed sud-
denly are not subject to changes in adrenal weight resulting from capture
or handling

A word of caution is appropriate here with regard to sample size and
adrenal weight-body weight relationships. Adrenal weight varies consider-

2. Endocrines and Populations 247

ably from individual to individual in the same species, as well as with sex,
size, and therefore presumably age, in most mammals. Consequently, it is
necessary to obtain large enough samples of each sex in each age category
(at least mature and immature) to have reliable criteria for evaluating
adrenal changes in a population and with time.

In most studies of natural populations the investigator does not have
the privilege of selecting animals for size, weight, and sex each time a
sample is collected, and it is therefore necessary to determine adrenal
weight-body weight, adrenal weight-body length or some similar relation-
ship for the species and population with which he is working in order to
establish a common baseline for all samples. For example, it has been found
that the logarithm of adrenal weight in milligrams on body length (exclu-
sive of the tail) gives the best straight-line relationships for all sizes and
for both sexes of Norway rats (Christian and Davis, 1955). In other
species, for example Microtus pennsylvanicus, there are marked differences
in body-adrenal relationships between the sexes. The oft-used relationhsip
of milligrams of adrenal weight per 100 grams of body weight is quite
arbitrary and usually overcorrects for lighter and undercorrects for heavier
animals, therefore this relationship is useful for only a very narrow range
of body weights. Too often adrenals are collected as an afterthought along
with other data for which the investigation was originally designed. As a
result the data on adrenal weights usually are inadequate. A study of
changes in adrenal weight must be designed specifically to obtain this and
the necessary related data. Samples should be consistent with respect to
time, size, and sex and age composition. Many potentially useful field
studies involving an immense amount of effort have been of little value
because of poor sampling, usually a result of treating the collection of data
on adrenal weights as a secondary consideration in the investigation.
Sampling usually does not alter the populations, as the normal rate of re-
cruitment more than offsets losses due to sampling, in addition to which
compensatory changes in mortality rate will usually compensate for sam-
pling losses, unless the sampling is very frequent and intense.

b. Routine Idstologic measurements. It is very useful in evaluating adrenal
activity to measure the widths of the various zones and to coimt the number
of cells in cortical cords in an area where cords are in straight columns
(Zwemer, 1936; Zwemer et al., 1938; Christian, 1956). A variation of this
is the somewhat more precise technique of projecting a section through the
center of the gland and outlining the various zones, cutting them out and
weighing them or else measuring their areas with a planimeter. It is possible
with serially sectioned glands to repeat such a procedure every so many
sections and in this way obtain precise information on the contributions
made by the individual zones to the total weight of the gland. These tech-

248 J- J' Christian

iiiques necessitate selection of appropriate sections and therefore require
serial sections, especially on animals with adrenals weighing less than 20 mg.
For glands heavier than this it is possible to select a portion through the
middle of the gland grossly. All these measurements can be made on sec-
tions stained routinely with hematoxylin and eosin.

c. Adrenocortical sudanophilia. Fat stains, such as Sudan IV, are useful
in evaluating the secretory activity of the adrenal cortex. The normally
active cortex ("resting cortex") usually exliibits marked sudanophilia of
the fasciculata and glomerulosa with the fat in vacuoles of moderate size.
The sudanophilia becomes markedly reduced in acutely stimulated glands,
and upon continued stimulation the sudanophilia may return to some
degree, but the size of the vacuoles is reduced. Upon cessation, sudanophilia
becomes intense and the vacuoles large. Sudanophilia may be completely
absent from some species, such as the hamster, under all conditions. These
details are discussed in greater detail elsewhere (Dempsey, 1948; Sayers
and Sayers, 1949; Greep and Deane, 1949a, b; Symington ct al., 1958).

d. Adrenal ascorbic acid declines on stimulation with ACTH, and this has
been used as a means of assaying ACTH in blood and other fluids (Sayers
and Sayers, 1949; Greep and Deane, 1949b). In general adrenal ascorbic
acid is a useful histochemical means of assaying adrenal stimulation, al-
though it may not reflect cortical secretory activity in the intact animal
(Slusher, 1958).

e. Adrenal cholesterol also declines on stimulation and can be used to
assess cortical activity (Dempsey, 1948; Sayers and Sayers, 1949; Greep
and Deane, 1949b) . IVIeasurement of adrenal steroids in the adrenal venous
affluent, plasma, or their urinary metabolites can be used to assess cortical
activity directly. These are subject to daily variation and also respond
rapidly to stimulation. In addition, they are involved techniques to carry
out properly. There is also some question concerning the biologic signifi-
cance of the particular steroids measured. These measurements are there-
fore of limited use for investigations outside the laboratory and are subject
to numerous pitfalls in dealing with species which cannot be handled in the
laboratory without stimulating adrenal activity. Measurements of urinary
corticosteroids when practical and appropriate would probably be more
useful than blood levels for studies relating to mammalian populations.
Measurement of the products of corticosteroid metabolism in the urme is
somewhat intermediate between weight and direct functional measure-
ments, such as plasma steroids as a measure of chronic cortical stimulation.
Urinary steroids may be collected daily ad infinitum under laboratory
conditions and are unquestionably ^'ery valuable. However, steroid deter-
minations are complex and difhcult to interpret, and in the last analysis
represent only those steroids which escape as such via the urine and consist

2. Endocrines and Populations 249

largely of metabolites of steroids. Furthermore, a large proportion of the
steroids are excreted in the feces in some species (Barry et al., 1952; Brad-
low et al., 1954) . Finally, techniques for the measurement of urinary steroids
have not been developed for use in the field. It should be apparent from
the foregoing discussion that assessing adrenocortical activity by physio-
logic or biochemical means in living wild animals or animals which have
been live-trapped is fraught with difficulty, and under most circumstances
measurements as such of the blood steroids will probably reflect the im-
mediate situation of the animal. In general, weight and histologic criteria
from animals killed suddenly seem to offer the most dependable informa-
tion at the present time for studying population phenomena, although the
other procedure should be explored further. Even though there is always a
question of the presumptive relationship between hypertrophy, morpho-
logic change, and function, in most circumstances it is generally accepted
that cortical hypertrophy reflects functional change. The probable depar-
tures from this general statement have been discussed above.

/. Thymicolymphatic system. In the foregoing account it has been stated
that the carbohydrate-active corticoids produce involution of the thymico-
lymphatic system, mainly by effecting involution of the lymphoid elements
proper. Therefore the weights of the thymus, lymph nodes, and spleen
provide useful indicators of increased adrenocortical activity, especially if
appropriate histologic checks are used.

g. Thymus weight. Cortisone, hydrocortisone, and endogenous adrenal
corticoids cause thymic involution (Dougherty, 1952a, b; Weaver, 1955).
Increased phagocytosis, edema of the connective tissue stroma, and hyper-
plasia of the reticulum are associated with the destruction of the lymphoid
tissue proper (Gordon, 1955). If the reaction is severe enough, only the
stroma and a modified reticulum may remain, with no distinction between
the cortex and medulla of thymus and lymph nodes in degree of involution
(Weaver, 1955) . The weight of the thymus therefore will be greatly de-
decreased. Within a few hours following stimulation there is a marked
edema, and during this period of edema there is a marked reduction in the
number of lymphocytes; those remaining exhibit degenerative changes
(Dougherty, 1952b). The effect is greatest on the small and medium-sized
lymphocytes and thymocytes. These exhibit pycnosis, karyolysis, frag-
mentation, and other degenerative changes. The fragments arc phagocy-
tized by macrophages in reaction centers of the lymph nodes and are
carried off in the lymphatics. Mitosis ceases in the lymphoid organs with
treatment with cortisone, hydrocortisone, or with chronic stimulation of
the adrenal cortices by endogenous or exogenous ACTH (Weaver, 1955;
Gordon, 1955) . These changes are common to the lymphoid follicles where-
ever they occur — spleen, nodes, gastrointestinal tract — but are most

250 /. /. Christian

marked in the thymus. The greater sensitivity of thymocytes over the
l3'mphocytes of the lymph nodes and spleen may be due to their greater
rate of proliferation. Hydrocortisone is more effective in involuting lymph-
oid tissue than cortisone or corticosterone and tends to produce degenera-
tive changes among immature lymphocytes of the lymph nodes to a much
greater degree than cortisone (Dougherty, 1953; Santisteban and Dough-
erty, 1954) . On the other hand, pituitary growth hormone and deoxycorti-
costerone appear to promote the growth of lymphoid tissue. The apparent
increase in lymphoid tissue in gonadectomized or adrcnalectomized animals
(or in those animals with adrenocortical atrophy from inactivity or over-
feeding) (Christian and Ratcliffe, 1952) apparently results from protection
against involuting agents rather than a true hyperplasia (Dougherty,
1953; Santisteban and Dougherty, 1954) .

The thymus is also involuted by androgens, estrogens, and to a variable
degree by thyroidectomy (Weaver, 1955) ; androgens also potentiate the
ability of the carbohydrate-active corticoids to effect thymic involution
(Selye, 1955; Dorfman and Shipley, 1956) . The cortex and medulla remain
as distinct zones after this type of in\'olution, although they are less clearly
defined than in normal (Weaver, 1955). All elements of the thymus share
equally in the involutional process. Thymocytes are reduced, but no acute
destructive changes are noted. Therefore involution of the thymus by sex
hormones is a nonspecific action (Weaver, 1955) . The so-called involution
due to aging of the thymus is prevented by gonadectomy in either sex
(Selye, 1947) ; it is therefore logical to assume that the normal involution
of this gland, seen especially at puberty (Christian, 1956) , is due to the sex
steroids.

It should be clear that thymic mvolution, and therefore thymic weight,
can be a useful index of adrenocortical activity providing adequate con-
sideration is taken of the action of the sex steroids.

h. Lymph node weight is a useful index of adrenocortical activity as their
lymphoid tissue is involuted by adrenal corticoids as discussed above, but
to a less marked degree than for the tlwmus (Weaver, 1955). The iliac
lymph nodes are involuted by prolonged administration of ACTH, acute
administration of cortisone (Weaver, 1955), or a variety of stimuli evoking
increased adrenocortical activity, but the sex steroids or thyroidectomy are
without effect on the lymph nodes (Weaver, 1955) or maj^ even cause a
hypertrophy of the nodes (Money et al., 1950) . A weight loss of the lymph
nodes should therefore more specifically reflect increased adrenocortical
activity than the thymus. However, tlwmic weight, properly controlled, is
more frequently used because of its much greater sensitivity to the carbo-
hydrate-active corticoids. Selye (1950) has indicated that the lymph nodes

2. Endocrines and Populations 251

even may hypertrophy in infectious disease or other circumstances which
demand increased phagocytosis of particulate matter.

i. Splenic weight. The lymph follicles of the spleen also are involuted by
corticoids. A decrease in splenic weight is sometimes used to indicate in-
creased adrenocortical secretion. However, contraction of the splenic cap-
sule further decreases splenic weight. This may be a valid procedure in the
laboratory with injected hormones, particularly in hypophysectomized or
adrenalectomized animals, but often may not be a valid indication of in-
creased adrenocortical activity, especially in intact small mammals. Voles
(Microtus agrestis) and house mice, and possibly other species, exhibit a
marked splenic hypertrophy when exposed to stimuli which cause in-
creased pituitary-adrcnocortical activity (Clarke, 1953; Chitty, 1957;
Christian, 1959c). Ln-olution of the lymphoid follicles occurs in these
circumstances as might be expected, but the decrease in weight from this
cause is overridden by the hypertrophy resulting from congestion and
markedly increased hematopoietic activity. The increased hematopoiesis in
these animals is accompanied by a reticulocytosis of the circulating blood.
The cause of increased hematopoiesis in these animals has not been demon-
strated.

j. Lymphocijte counts. After an alarming stimulus there is an immediate
rise in the number of circulating lymphocytes, which is then followed by a
characteristic lymphopenia and eventually a return to normal levels. The
initial rise in circulating lymphocytes probably is at least in part due to
release of lymphocj-tes from lymphoid tissues in response to an initial re-
lease of epinephrine, but this release by no means accounts for the entire
rise (Gordon, 1955). The prolonged phase of lymphopenia which begins in
the early stages of the increased pituitary-adrenocortical activity probably
results from the destruction of medium and small-sized lymphocytes within
the lymphatic organs by the corticoids (Gordon, 1955) . However, there is
also some reason to believe that lymphocytes may migrate to depot situa-
tions. Over a prolonged period the main cause of the lymphopenia appears
to be due to the reduced amount of lymphoid tissue and decreased lympho-
cytopoiesis with a decreased delivery of lymphocytes into the circulation.
However, with small amounts of adrenal factors there may be an acutal
hyperplasia of lymphoid elements, so that interpretations must be made
with caution.

Lymphocyte counts as criteria of stress in wild mammals are subject to
the same criticism as other measurements made on the living mammal : the
response is rapid enough and the counts labile enough so that there is
real danger that counts may reflect alarming stimuli induced by handling,
thus masking any other effects which may be the main point of the study.

252 /. J. Christian

k. Eosinophil counts. The numbers of circvilating eosinophils are beyond
doubt diminished by adrenal factors and eosinophil counts are commonly
used to assess the response of the adrenal cortex to ACTH or indicate in-
creased pituitary-adrenocortical activity (Speirs and Meyer, 1949; Speirs,
1955). The reduction in eosinophils appears to result mainly from their
increased destruction under the influence of cortical hormones. A reduction
in eosinophils can be effected by epinephrine as well as by cortical hormones,
and there is evidence that the presence of cortical hormone is required for
epinephrine to produce an eosinopenia (Gordon, 1955). It is mainly with
respect to whether or not cortical hormones are necessary for the eosino-
penic action of epinephrine that the specificity of eosinopenia as an indica-
tor of increased adrenocortical activity has been questioned. Nevertheless,
if epinephrine produces a marked eosinopenia in the intact animal, eosino-
phil counts in mammals, especially wild, appear to have limited value.
Fear resulting from handling, trapping, etc., not only would produce
eosinopenia, but could produce it without necessarily having an increase
in adrenocortical activity. Nevertheless Louch (1958) used eosinophil
counts to assess adrenocortical function in relation to changes in popula-
tions of voles and with adequate precautions the use of eosinophil counts
for assessing adrenocortical activity in natural populations seems to have
considerable value in these experiments. However it is not possible to
state definitely whether the declines in eosinophils were due to cortical or
medullary hormones, although it seems logical that the former were respon-
sible and there was cortical hypertrophy. Southwick (1959) has also related
declines in eosinophil counts in mice to increased adrenocortical functions,
but again the role of epinephrine has not been evaluated. Eosinophils de-
cline about an hour after acute stimulation of the pituitary-adrenocortical
systems (Louch, 1958) ; therefore using counts as criteria of cortical func-
tion with respect to chronic stimuli has many inherent hazards. Acute re-
sponses to handling, fear, or line trapping could easily mask any effects pro-
duced by the chronic stimulus of changing population density. The relative
roles of epinephrine and of the cortical hormones in producing the observed
declines in eosinophils following emotional stimuli in mice must be assessed
before one can state with finality that the cause was increased cortical
activity.

I. Neutrophil counts. A rise in circulating neutrophils accompanies in-
creased pituitary-adrenocortical activity, but similar changes can be in-
duced by so many factors that, even though they may involve increased
adrenocortical secretion, neutrophil counts are not very useful indices
with which to measure pituitary-adrenal activity.

m. Liver glycogen. The accumulation of glycogen by the liver has been
used as a means of assaying various steroids for their activity on carbohy-

2. Endocrines and Populations 253

drate metabolism in adrenalectomized rats (Dorfman, 1949; Ingle, 1950).
This response is a reflection of the ability of adrenal corticoids to stimulate
gluconeogenesis. Glycogen in the liver can be measured chemically (Marto-
rano, 1957) or can be visualized histochemically by the appropriate proce-
dures (Dempsey, 1948; Lillie, 1954). Liver glycogen presumably falls as a
result of exhaustion of the animal's ability to adapt (Frank, 1953) . Exactly
what becomes exhausted under these circumstances is problematical, but it
apparently is not the ability of the adrenal cortex to secrete steroids (Rosen-
feld, 1958) . Nevertheless, it is true that blood glucose and liver glycogen
eventually fall to extremely low levels following intense and prolonged
stimulation by alarming stimuli. This result may follow prolonged stimula-
tion of the sympatho-adrenal system. Exhaustion of glycogen, consequently
of glucose reserves, has been observed in natural populations of snowshoe
hares (Green and Larson, 1938; Green et al, 1939) and voles (Frank, 1953)
during episodes of mass mortality and may possibly be explained on the
basis of exhaustion of the adaptive reserves.

A word of caution should be inserted on the use of liver glycogen as an
indicator or activity of the carbohydrate-active corticoids. Under carefully
controlled conditions in the laboratory such measurements are very useful,
but they do require rigid controls and precise, highly standardized proce-
dures. Martorano (1957) has studied the variables having important effects
on liver glycogen. The amount of time between killing and enzymatic
immobilization, the manner of killing, and a variety of other factors can
alter glycogen levels. Glycogen and glucose levels in the liver decline rapidly
after death. There also is a daily cycle in the levels of liver glycogen which
is associated with feeding and activity (Martorano, 1957). Therefore, re-
ported liver glycogen levels must be critically examined in relation to the
procedures used and their reliability.

n. Changes in the digestive organs and gastrointestinal tract. Alarming
stimuli, ACTH, or cortisone produce a marked loss of the acidophilic
zymogen granules and basophilic cytoplasmic pentose nucleic acids from
the acinar cells of the pancreas and serous salivary gland cells (Ehrich and
Seifter, 1948; Selye, 1950). The pituitary adrenocortical system also pro-
foundly affects the function of the mucosa of the stomach and intestinal
tract (Baker and Abrams, 1954; Baker and Bridgman, 1954; Gray and
Ramsey, 1957). The adrenocortical hormone apparently stimulates the
secretion of pepsinogen by gastric zymogenic cells (Gray and Ramsey,
1957; Mason, 1959), which is reflected morphologically by a loss of the
intracellular pepsinogen granules and increased cytoplasmic basophilia
(Baker and Bridgman, 1954). Additional changes may not be noted histo-
logically in the mucosal cells of the intestine (Baker and Bridgman, 1954).

254 /. /. Christian

3. Generalized Effects of Physiologic Adaptation

a. Growth. Associated with acute adaptation there is usually a suppres-
sion of growth (Selye, 1950) . Selye (1950) has suggested that during actua-
tion of the pituitary-adrenocortical system there is a decreased production
of the other hormones of the adenohypophysis, including growth hormone.
However, injected corticoids or ACTH into intact animals also results in
suppression of growth and ACTH will inhibit growth produced by growth
hormone in h^Tpophysectomized rats (Jones, 1957). It was pointed out
earlier in this chapter that the adrenal carbohydrate-active corticoids in-
crease protein catabolism, prevent the gro^^iih of bone, lymphoid tissue,
skin, connective tissue with the production of collagen, and block mitoses
in general. Therefore the suppression of growth by alarming stimuli may
reflect decreased secretion of growth hormone as well as the direct suppre-
sive action of the carbohydrate-action corticoids on mitoses and therefore
growth and development. Diminished thyroid activity during periods of
actively increased adrenocortical activity may also play a role in this phe-
nomenon. Selye (1950) has pointed out that most of the observed changes
following an alarmmg stimulus serve to maintain life and that other, less
immediately important, functions are suppressed.

b. Inflammation and granulation. The adrenocorticoids, cortisone, hydro-
cortisone, and to a lesser extent corticosterone, exert powerful anti-
inflammatory effects which stem from the suppression of growth of connec-
tive tissue, the depression of Ij'mphocyte activity, and interference with
the phagocj^tic process (Dougherty, 1953; Robinson and Smith, 1953;
Thomas, 1953), but in addition they prevent the mobilization of all of the
usual elements of an inflammatory response around the site of injury
(Taubenhaus and Amromin, 1950; Dougherty, 1953; Robinson and Smith,
1953; Gordon, 1955; Dougherty and Schneebeli, 1955). Androgens and
estrogens potentiate these anti-inflammatory responses. They decrease the
destruction of fibroblasts and the invasion of polymorphonuclear leuco-
cytes and macrophages. The appearance of epithelioid macrophages, giant
cells, and formation of new fibroblasts and macrophages are suppressed by
cortisone, either injected or implanted as pellets (Baker, 1954) . Dougherty
and Schneebeli (1955) explain the inhibition of the inflammatory response
around the site of injury in the following way: When there is cellular injury,
substances are released from the injured cells which trigger a series of re-
sponses which comprise inflammation (Menkin, 1955; Dougherty and
Schneebeli, 1955). Cortisone or hydrocortisone inhibit the inflammatory
response by protecting the surviving cells from the actions of the released
products of cellular injury. Growth hormone, deoxycorticosterone, and
aldosterone appear to exert a stimulating effect on inflarmnation and the

2. Endocrines and Populations 255

development of granulation tissue (Selye, 1955; Dougherty and Schneebeli,
1955) . These hormones, however, also inhibit the anti-inflammatory action
of cortisone, hydrocortisone, and corticosterone and may enhance the local
inflammatory response by increasing the susceptibility of the cells to the
inflaming stimulus (Dougherty and Schneebeli, 1955). Therefore, if there
is a reduction in the secretion of growth hormone with a simultaneous in-
crease in the production of ACTH and the adrenocortical steroids, there
will not only be a direct suppression of the inflammatory responses to infec-
tion or injury, but also a withdrawal of the factors which ordinarily would
stimulate such activity. The effects of various alarming stimuli or hormones
on inflammation and granulation in the intact animal have been studied
and measured by using experimental granulomas (Meier et at, 1950; Selye
and Bois, 1954; Robert and Nezamis, 1957; Christian and Williamson,
1958) or other means of inducing inflammation and granulation. It has been
shown that in addition to suppressing inflammation the carbohydrate-
active corticoids also suppress the formation of granulation tissue and
healing, primarily by preventing connective tissue growth. There can be
little doubt that the normal defenses against infection are severely depressed
in stressed animals.

c. Antibody formation. Antibodies are formed mainly in the lymphatic
tissues (Kenning and Van der Slikke, 1950; Kass et al., 1953a; Kelsall and
Crabb, 1958) . One school maintains that this function resides primarily in
the plasma cells, w^hile another group holds that lymphatic cells in general
are capable of manufacturing antibodies (Kenning and Van der Slikke,
1950; Dougherty, 1953; Kelsall and Crabb, 1958). We do not intend to
enter into this controversy at the present time, but it seems relatively
certain that the lymphoid tissues are primarily responsible for the produc-
tion of antibodies. A variety of experiments have shown that injected
corticoids or the increased secretion of endogenous corticoids, brought
about either by injected ACTH or in response to alarming stimuli, markedly
suppress the formation of antibodies (Kass et al., 1953a) . Protein manufac-
ture, and therefore the formation of antibodies, requires the presence of
nucleic acids in the cells, especially in the cytoplasm, and antibody forma-
tion is normally associated with an increase in nucleic acid content of the
lymphoid organs (Kass et al., 1953a; Kelsall and Crabb, 1958). Therefore,
when there is interference with nucleic acid metabolism or its formation,
there is an accompanying reduction in the rate of formation of antibodies
(Kass et al., 1953a; Kelsall and Crabb, 1958). We have seen that in addi-
tion to actually destroying lymphoid tissue, the glucocorticoids reduce the
PNA content of the remaining lymphatic cells. The ability of the reticulo-
endothelial system, and possibly other cells, to dispose of phagocytized
particulate material is also impaired (Thomas, 1953) even though phago-

256 /. /. Christian

cytosis may be stimulated (Gordon and Katsh, 1949; Thomas, 1953). By
these several mechanisms the production of antibodies may be seriously
impaired following activation of the adrenal cortex although there appears
to be a dose-response relationship (Dougherty, 1953; Dougherty and
Schneebeli, 1955). While most of these effects were demonstrated most
clearly by injecting adrenocortical hormones, the same effects have been
shown repeatedly following stimulation of adrenocortical secretion in the
intact animal.

d. Resistance to infection. The three immediately preceding topics all
deal with factors involved in the resistance to infection. It stands to reason
that reducing the inflammatory response to, and depressing the formation
of antibodies against, infectious agents will inevitably impair the ability of
an animal to resist infection. Cortisone, hydrocortisone, and ACTH have
been shown to decrease resistance to a variety of experimental infections
caused by a variety of infectious agents including streptococcal, pneumo-
coccal, tuberculosis infections in mice, rats, and guinea pigs, brucellosis,
malaria in monkeys, and others (Kligman etal., 1951 ; Schmidt and Squires,
1951 ; Selye, 1951 ; Kass et al., 1953b; Le Maistre et al., 1953; Robinson and
Smith, 1953). The pathogenicity of various agents has been increased by
cortisone injection. For example, the virulence of Coxsackie infections in
mice was greatly enhanced by cortisone (Boring et al., 1955), and polio-
myelitis can be made a paralytic disease in the normally resistant hamster
by cortisone or hydrocortisone (Shwartzman and Aronson, 1953). Vire-
mias may likewise be prolonged appreciably by the adrenal glucocorticoids
(Whitney and Anigstein, 1953; Pollard and Wilson, 1955). The list of
experimental infections which have been made more virulent, prolonged,
or otherwise increased in their pathogenicity by treatment with carbohy-
drate-active adrenocorticoids or ACTH (Selye, 1951) is long, and there is
no point in listing them in detail here.

Whenever experiments with injected hormones are considered the ques-
tion arises whether or not the same events may occur as a result of in-
creased endogenous secretion of the same or similar hormones. A criticism
frequently made of experiments with exogenous hormones, especially with
large doses, is that the results are pharmacologic rather than physiologic.
However, it is by using isolated hormones in highly controlled situations
that an understanding of the basic mechanisms is gained. Nevertheless,
before one can extrapolate from these data to natural events, comparable
changes must be shown to occur in natural or seminatural conditions.
Changes in host resistance may result from adverse environmental stimuli,
possibly as a result of adrenocortical activity. It has long been common
knowledge that excess fatigue, chilling, and a variety of comparable stimuli
increase the susceptibility of humans to colds and other infections. It should

2. Endocrines and Populations 257

be apparent now that most of these same stimuH also increase the secretion
of adrenocortical steroids. However, commonly accepted truisms still do
not constitute experimental evidence and proof of such conclusions, but
several experiments have shown that host resistance is decreased by ex-
posing the animals to stimuli which are known to increase adrenocortical
activity. When mice are exposed to 4° C. for a period of time, Coxsackie
infections become much more pathogenic, spreading especially to the heart
and liver in adult mice (Boring et al., 1956) . Duninished resistance of mice
to trichinosis and to tuberculosis has been demonstrated by procedures
which also produce increased pituitary-adrenocortical activity and depress
inflammation and granulation (Christian and Williamson, 1958; Tobach
and Block, 1956; Davis and Read, 1958) . Some aspects of these studies will
be considered in more detail later.

e. Reproduction. Suppression of reproduction is a very important aspect
of the endocrine adaptive responses (Selye, 1939, 1950). The decrease in
reproductive function is in many ways a more sensitive measure of the
existence of altered physiologic functions in response to adversity than the
increase in adrenocortical activity and some of its sequelae (Christian,
1955a, b, 1956, 1959b, c) . The bulk of the experimental evidence indicates
that, like growth hormone, the secretion of pituitary gonadotropins is sup-
pressed in response to alarmmg stimuli which evoke an increased secretion
of ACTH (Selye, 1939, 1950; Christian, 1956, 1959b, c). It is another
indication that the immediate restoration of the normal internal environ-
ment takes precedence over functions which are less important to the im-
mediate siu-vival of the individual. Cold, heat, disease, trauma, severe
emotional stress, and other stimuli will depress normal reproductive func-
tions (Marrian and Parkes, 1929; Selye, 1939; Bohanan, 1939; Poindexter,
1949; Denison and Zarrow, 1955; Barnett and Manly, 1956; Christian and
LeMunyan, 1958; Christian, 1959b). The suppression of gonadotropin
secretion is evidently the primary cause of the inhibition of reproductive
function (Mulinos et al, 1939, Selye, 1950; Srebnik et al, 1958), although
there may be direct effects of increased adrenocortical activity and altered
thyroid function, as mentioned earlier (Brynes and Shipley, 1950; Baker
et al, 1950; Smith, 1951; Brimblecombe et al, 1954). The secretion of
gonadotropins seems to be regulated principally by the hypothalamus
(Anderson and Haymaker, 1948a, b; Markee et al, 1952; Everett and
Sawyer, 1953; Hammond, 1954; Critchlow and Sawyer, 1955; Nalbandov
et al, 1955; Laqueur et al, 1955; Fortier, 1957; Greer, 1957; D'Angelo and
Traum, 1958 Everett, 1959) , although there is evidence that the secretion
of luteotropin by the pituitary may be independent of the hypothalamus
(Everett, 1956). The hypothalamic centers involved in the regulation of
the secretion of the gonadotropins apparently are distinct from those

258 /. /. Christian

responsible for the regulation of ACTH and TSH secretion (Laqueur et al.,
1955; Greer, 1957; D'Angelo and Traum, 1958). In addition there are
neurogenic factors involved in ovulation which consist of adrenergic and
neurogenic components (Markee et al., 1952; Everett and Sawyer, 1953;
Nalbandov et al, 1955) and probably also in the milk let-down reflex
(Grosvenor and Turner, 1959a). Finally, as in other endocrine regulating
mechanisms, the level of circulating sex steroids seems to exert a regulating
effect on the release of the particular gonadotropin responsible for their
release (Selye, 1947; Sturgis, 1950; Byrnes and Shipley, 1950; Byrnes and
Meyer, 1951). Whether these steroids exert their effects on the hypothala-
mus or on the anterior pituitary itself is not known. In any event, with-
drawal of stimulation by the gonadotropins leads to atrophy of the gonads
and decline in the production of their respective steroids (Burrows, 1949).
The decline in the production of sex steroids is in turn followed by atrophy
of those accessory organs and secondary sex characteristics which depend
on the sex steroids for their activity and maintenance (Burrows, 1949).

If a male animal is subjected to adverse circumstances for any length of
time, there is a marked decrease in spermatogenesis and a reduction in the
secretion of androgens as indicated by atrophy of the seminal vesicles,
prostate, and preputial glands (Christian, 1956; 1959b; Christian and
LeMunyan, 1958) . These changes may all be followed by the changes in
their weights, but weight changes should be assessed by appropriately
selected histologic examination of the organs. In young animals there may
be either a delay in the onset of puberty or a total suppression of the de-
velopment of puberty, as indicated by the development of spermatogenesis
and the sex accessories or, in the mouse, by a failure of the X-zone to
involute at the usual time (Christian, 1956, 1959a, b). If mature animals
are subjected to severe stimuli, there may be easily discernible degenerative
changes in the tubules of the testes and cells of the spermatogenic series.

Changes in the female reproductive function are less easily seen than in
the male, as organ weights do not provide as useful a criterion of changes.
Nevertheless, changes in estrus are usually evident and can be detected by
vaginal perforation or by vaginal smears. Depending on the severity of the
inducing stimulus, estrus may be prolonged or totally suppressed. Uterine
weight may reflect reproductive suppression, and would be especially
valuable in demonstrating a delay in the onset of puberty in female animals.
In some circumstances there may be a total suppression of reproductive
activity. It is in actual reproductive performance that the effects of sup-
pression of the reproductive endocrine system becomes most evident. De-
pending on the severity of the stimulus and on the species involved, there
may be a complete failure to become pregnant, which may be due to a
failure of ovulation or a failure of the shed ova to implant in the uterus or,

2. Endocrines and Populations 259

if the females do become pregnant, there may be a marked increase in fetal
mortality with increased resorption of the embryos (Christian, 1959a, b) .

The causes of fetal mortality in these circumstances have not been
explored in detail, but a number of factors may be involved. It is well
known that cortisone and hydrocortisone have serious effects on the fetus,
apparently by inhibiting growth and development (Glaubach, 1952; Fraser
et al, 1953; Davis and Plotz, 1954; Kalter, 1954). The resultant defect
probably depends to a large extent on the developmental stage of the fetus
when these hormones are active, as is the case with nutritional deficiencies
(Lutwak-Mann, 1958) . A variety of congenital defects have been produced
experimentally by the injection of these hormones. However, high doses are
required to produce these effects; furthermore, with chronic injection of
cortisone some fetuses do not seem to be affected (Seifter et al, 1951).
Hydrocortisone and cortisone increase the incidence of fetal mortality in
rats during the second half of pregnancy. The mechanism is unknown, but
it has been suggested that it may be due in part to premature ''aging" of
the vasculature of the placenta (Seifter et al, 1951). Increased mortality
may reflect fatal defects in the growth of the embryos due to the action of
these hormones. In other words, there may be time-dosage relationships
which determine whether the effects of these glucocorticoids will be fatal or
will result in "congenital defects," such as cleft palate or cardiac anomalies.
However, these explanations leave unanswered the question why many
embryos subjected to the same influences are born viable and free of de-
fects, even from the same pregnancies. However, there are many other
factors to consider when discussing intrauterine mortality resulting from
activation of physiological adaptive systems. One must consider the in-
creased secretion of adrenal androgens by those species in which androgens
or proandrogens constitute a major secretory product. These compounds
possibly may directly inhibit the action of estrogens and progestins. For
example, it has been shown that testosterone can completely inhibit the
feminizing action of estrogens on developing rat embryos (Greene et al,
1941). More importantly, the adrenal androgens can inhibit the secretion
of gonadotropins, especially of FSH (Byrnes and Shipley, 1950; Byrnes
and Meyer, 1951 ; Dorfman and Shipley, 1956). Whether the decline in the
secretion of gonadotropins with increased ACTH secretion is due to inhibi-
tion by increased circulatory levels of adrenal androgens or whether it is
independent of the androgens, cannot be said. However, it seems probable
that the androgens are not important or, if so, only in a limited number of
species. The adrenal secretion of androgens or their precursors in many
species does not appear to be sufficient to account for the effects seen. For
example, in humans pituitary blockage requires much higher doses of
testosterone than is required to produce overt androgenic responses. Very

260 J' J' Christian

likely the reciprocal relationship between ACTH and gonadotropin secre-
tion is independent of the adrenal androgens in most species, although the
latter possibly may enhance a pre-existing inhibition of the gonadotropins.
However, they may be of considerable importance in house mice, rats, and
other species which secrete appreciable amounts of adrenal androgens.
There undoubtedly are many other factors contributing to intra-uterine
mortality which have not been discussed here. In general, however, it is
evident that factors increasing pituitary-adrenocortical activity are asso-
ciated with increased intra-uterine mortality, although the details of the
mechanisms are largely unknown.

Finally, if the young are born, there may be a failure of lactation due to
deficiency in the hormones normally required for the maintenance of
adequate lactation (Christian and LeMunyan, 1958). The changes in re-
productive function following inanition have been show^n to be due pri-
marily to a decrease in the secretion of gonadotropins, as the gonads remain
responsive to injected gonadotrophin (Srebnik et al., 1958). Most of these
alterations in reproductive function will be discussed in more detail later.

/. Acute visceral degenerative changes. Acute degenerative changes of the
liver, pancreas, and parotid glands have been described as occurring during
an alarm reaction to a variety of stimuli (Selye, 1950; Ehrich and Seifter,
1948). Selye (1950) states that similar changes also occur during the
"exhaustion" phase of adaptation.

The most constant hepatic changes are pycnosis, cloudy swelling, and
stromal edema. Occasionally there is fatty infiltration, severe atrophy, focal
necrosis, and leucocytic infiltration. However, many of these changes are
duplicated by a variety of specific stimuli and cannot be considered as
diagnostic in any sense unless all other possible factors, other than increased
pituitary-adrenocortical activity, have been ruled out.

The salivary glands, especially the parotid, also exhibit degeneration and
necrosis of the parenchyma, often with suppurative inflammation, during
an alarm reaction (Ehrich and Seifter, 1948). In addition to the zymogen
discharge already alluded to, the pancreatic parenchyma may exhibit focal
necrosis and inflammation comparable to that seen in the salivary glands
(Selye, 1950). The cytoplasmic nucleic acids are markedly decreased
during acute adrenocortical stimulation or in response to injected adreno-
corticoids, and may precede the degenerative changes. Loss of cytoplasmic
basophilia in hepatic cells under the same circumstances reflects the same
kind of influence of adrenal steroids on cytoplasmic nucleic acids.

g. Other effects. In addition to those effects of physiologic adaptation
which have been discussed, there are others which stem from a general shift
in physiologic functions or as yet have not been related to the activity of
any specific hormone. Among these is enlargement of the intervertebral

2. Endocrines and Populations 261

discs (nucleus pulposus) which has been found in voles {Microtus orcaden-
sis) subjected to emotional stress (Chitty et al, 1956) . The enlargement of
the intervertebral discs could result from altered fluid and electrolyte
balances brought about by increased activity of adrenal corticoids, or
might follow alterations in the ground substance by cortical hormones;
however, these explanations are conjectural as the mechanism is as yet
unknown.

Physiologic adaptation may be accompanied by a variety of shifts in the
composition of the blood with changes in hematocrit, electrolyte concentra-
tions, and various metabolites. Blood lipids and sugar may also shift mark-
edly, as we have mentioned earlier. However, these factors are discussed
in considerable detail elsewhere (e.g., Selye, 1950; Hartman and Brownell,
1949; Jones, 1957) and will not be considered further in this discussion.

Part 2. Physiologic Adaptation and
Mammalian Populations

I. Introduction

The foregoing account dealt largely with the basic endocrine and other
physiologic adaptive mechanisms that serve to maintain physiologic homeo-
stasis in the face of a variety of stimuli tending to alter the internal environ-
ment and to equip the animal to meet the demands of emergency situations.
In general, any stimulus which imposes physiological demands on an organ-
ism beyond those ordinarily met in undisturbed idyllic daily life calls into
play a series of feedback mechanisms that regulate the secretion of hor-
mones responsible for the maintenance of a relatively constant internal
environment. These mechanisms act upon the distribution of the internal
environment via the circulatory channels, the composition of the internal
environment with respect to fluids, electrolytes, glucose, fats, and a variety
of other metabolites and metabolic products, and on the supply of readily
available nutrients and oxygen, especially for the skeletal muscle and ner-
vous system. We have seen that these effects are not achieved withovit
sacrificing functions less immediately vital to the individual, such as re-
production, growth, and resistance to infectious disease and parasitism.
We also have pointed out the error in thinking that the responses to all
adverse stimuli are the same and that all necessarily are associated with
increased secretion of adrenal carbohydrate-active corticoids.

262 /. /. Christian

The mechanisms, hormonal effects, and general responses so far described
have been confined largely to the results of experiments in the laboratory
with the usual laboratory species. Very little work has been done on native
mammals in the laboratory, and comparative studies are certainly needed.
In addition, the work has been to a great extent limited to studies of the
effects of injected hormones or of subjecting laboratory animals to ex-
tremely severe conditions.

The role that physiological adaptive mechanisms play under natural
conditions was not investigated in most of these studies. However, it was
postulated in 1950 that these same responses could be evoked by increased
population density and that these same physiological reactions could serve
as a feedback to regulate the growth of mammalian populations, their
declines, and the mass mortality which occasionally terminates the build-up
of a natural population to extremely high densities (Christian, 1950b). It
was first suggested that the intraspecific strife and social competition that
force animals into adverse circumstances, together with all of the other
adversities which become aggravated by high population densities, would
elicit adaptive responses such as those which have been described. There-
fore one would anticipate a direct relationship between adrenocortical
activity and a more or less reciprocal relationship between reproductive
function and population density. The hypothesis that physiological mecha-
nisms were active in all populations in response to changes in density and
that most environmental deficiencies acted through this mechanism was
inherent in the original postulate. This hypothesis implied that social
competition or pressure was the sole factor, always present in all popula-
tions, which could logically be expected to elicit the gamut of adaptive
responses in every population. However, these relationships had to be
demonstrated, and it was necessary to show that changes in population
alone could induce a proportional increase in pituitary- adrenocortical
activity, decrease in reproductive activity, decrease in resistance to disease,
or even death from shock, and all the other reactions and responses which
have been described in the preceding section.

Since that time considerable evidence has been accumulated from the
laboratory and from natural populations which indicates that these re-
sponses to population density do occur and that they can regulate popula-
tion growth. The balance of this chapter will be devoted to a presentation
of the experimental evidence for the response of physiological adaptive
mechanisms to social competition and therefore population density, and
the evidence implicating these mechanisms, acting as a feedback system, in
the regulation of mammalian populations. Finally there will be a discussion
of the pertinence of this evidence to the regulation and control of natural
populations.

2. Endocrines and Populations 263

The general plan of the following discussion will be to explore first the
ability of purely social factors to affect adrenocortical and reproductive
function. Next the relation between population density and adrenocortical,
reproductive, and other functions as well as alterations in resistance to
disease, will be investigated. In general the plan will proceed from popula-
tions of fixed size in the laboratory to freely growing populations in the
laboratory, and finally to natural populations. Under each of these experi-
mental categories adrenocortical function, reproductive function, disease
resistance, and mortality will be discussed with all appropriate experiments
and species. The effect of population density on growth will be discussed
where appropriate, and the effects of food and other environmental factors
on social interactions and endocrine function will be discussed. It is not
possible to stay strictly within this framework, as it is somewhat artificial,
but it does seem to offer the most logical means of presenting the available
information as it progresses from the most artificial but most highly con-
trolled experiments to natural populations which are controlled with great
difficulty, if at all.

II. Endocrine Responses to Social Pressures and to Population Density

A. Experiments in the Laboratory with Populations of Fixed Size

One of the basic tenets in the theory that physiological feedback mecha-
nisms can regulate population growth is that a fundamental regulating
factor must be present and active in all populations. Whether or not this
particular factor is the proximate factor limiting population growth in a
given instance is not important if it is universally present. The only known
element common to all populations is social interaction, or intraspecific
competition. Basically, competition depends on the behavioral characteris-
tic of the species, but some sort of social organization or mutual intolerance
is exhibited by all species of mammals.

1. Social Factors, Adrenocortical and Reproductive Functions

Social interactions may arise from two kinds of situations: one in which
there is invasion of the private territory of one animal by another; and
another in which there is conflict involved in the establishment and mainte-
nance of a hierarchical situation. The first requirement of the hypothesis
that physiologic mechanisms can and do control population growth is to
show that purely behavioral or social interactions, acting through the
central nervous system, can induce endocrine responses, especially of the
pituitary-adrenocortical and reproductive systems, and to be able to rule

264 /. J. Christian

out other factors which are known to produce the observed changes. That
pure sociopsychologic factors can produce these effects is clearly demon-
strated by the following experiments.

When a strange vole {Microtus agrestis) of either sex was placed daily for
27 days in a cage containing a resident pair of voles, and therefore in what
was essentially their private territory, vicious fighting ensued (Clarke,
1953). The introduced voles lost weight and exhibited a significant hyper-
trophy of the adrenals and spleens and atrophy of the thymus compared
to the residents. In similar experiments with Microtus orcadensis there were
significant increases in the size of the livers, adrenals, intervertebral discs,
and spleens, and decrease in the thymus in the "stressed" animals, those
introduced as strangers to resident pairs (Chitty et al., 1956). Clarke
(1953) attributed the changes in organ weight to the actual fighting,
stating that fighting is a very effective form of stress since it involves
vigorous muscular exercise in addition to the trauma of wounds (Selye,
1950) . However, the fact that the resident voles had to fight as much as
those which were introduced, and yet did not reflect this with increased
adrenocortical activity, apparentlj^ was overlooked. This problem may be
explored by data on albino mice (Christian, 1959d). An analysis of the
relationship between scars from fighting and adrenal weight from 280 male
albuio house mice from 55 populations of 4, 5, or 6 each showed that, al-
though the mean adrenal weight of every population increased appreciably
with respect to isolated controls, there was no relationship between the
amount of scarring, as an indication of the severity of fighting, and adrenal
weight. Fm'thermore, the presence or absence of scarring in a population
made no difference in adrenal weight. The results were similar irrespective
of whether the analysis involved only the differences between populations
or the difference between individuals. Adrenal weight increased the same
amount in populations in which there was no fighting, or so little that none
of the mice had injuries from biting, as it did in populations in which
fighting was severe enough that most of the animals were badly scarred. It
should be pointed out that these injiu'ies were superficial and for the most
part represented bites through the skin only. These results indicate quite
clearly that fighting or injury per se are not the stimuli responsible for
stimulating increased adrenocortical activity with thymic involution when
animals are placed together. Conclusive evidence that sociopsychologic
pressures alone are mainly responsible for evoking these physiologic adap-
tive responses lies in the following experimental results (Davis and Chris-
tian, 1957) .

When house mice are placed together in groups of six, there is immediate
fighting which soon ceases with the establishment of a social hierarchy with
one mouse dominant over the others and another subordinate to all the

2. Endocrines and Populations 265

others. The remaining mice arrange themselves in some sort of hierarchy in
between. It was found that adrenocortical hypertrophy was greatest in the
most subordinate animals and was slight or absent in the dominant mice.
The adrenals of those in between tended to fall in line in between in a
reciprocal relationship to their dominance rank. Fighting cannot have been
an important stimulus to mcreased pituitary-adrenocortical activity in
these experiments, as the dominant animals fought as much as, or more
than, any of the subordinate animals. The mice in these experiments also
exhibited changes in the weights of their reproductive organs consistent
with suppressed secretion of gonadotropms coincident with increased
pituitary-adrenocortical activity. There was no clear-cut decline in the
weights of the reproductive organs with decreasing social rank, but the
dominant mice had much heavier reproductive organs, especially the pre-
putial glands, then the subordinate animals (Davis and Christian, 1957).
These results were confirmed and intended in dogs by Eik-nes (1959), who
found that the dominant dogs in groups secreted about half the amount
of corticoids that the subordinate dogs secreted. Therefore there can be
little question that there are significant differences in adrenocortical func-
tion associated with differences in social rank.

In another series of experiments Southwick ( 1959) demonstrated that
moving mice into a new environment daily could induce a marked increase
in adrenocortical activity as determined by eosinophil counts. That the
eosinopenia was not a result of handling was shown by the fact that mice
handled in the same way but not placed in a strange situation responded
with only a slight fall in circulating eosinophils. The mice transferred to new
cages for a period of time every day adapted to the situation, as the eosino-
phil count returned to normal levels by the end of the 8-day experimental
period. A third series of anunals were placed in groups of four each day and
these animals exhibited an 80% mean decline in their eosinophil counts
and these counts remained low as long as the animals were placed in groups.
Presumably the mice responded to grouping with a marked increase in
adrenocortical activity and did not adapt to the situation. It is clear from
these results that merely placing mice in a strange situation is an emotional
stimulus sufficient to result in a decline in circulating eosinophils and pre-
sumably in adrenocortical activity and that grouping constitutes a more
profound stimulus to which animals fail to adapt. As we have pointed out
previously, it cannot be concluded finally that the eosinopenia was due to
increased adrenocortical activity, although it seems likely, imtil increased
secretion of epinephrine is ruled out as a causative factor.

Similar results were obtained when male Norway rats were placed in
groups (Barnett, 1955) . The subordinate males, subjected to severe fighting
for short periods, at first showed marked decreases in adrenocortical sud-

266 /. J- Christian

anophilia, whereas prolonged exposure to less severe fighting resulted in
normal sudanophilia with adrenal hypertrophy in the subordinate animals.
None of the dominant rats showed adrenocortical hypertrophy or changes
in sudanophilia in spite of the fact that they fought as much or more than
the subordinate animals. Barnett (1958) has recently published additional
evidence on a smaller number of rats from which he concluded that the
adrenal cortices of both subordinate and dominant animals hypertrophy.
Nevertheless his data show that the mean adrenal weight, relative or abso-
lute, of subordinate animals in a group of rats introduced into a colony was
appreciably greater than that of dominant animals. Barnett used very small
numbers of rats and based his conclusions on absolute rather than on rela-
tive adrenal weight, in spite of the fact that his animals varied from 250 to
400 gm. at the start and 170 to 420 gm. at the end of the experiment. He
concluded, on the basis of thirty rats, that there was no relationship between
adrenal weight and body weight for rats weighing more than 150 gm. This
conclusion may be ciuestioned for several reasons. In the first place the
mean adrenal weight of his animals increased with increasing body weight,
athough the differences were not significant. However, had larger numbers
of animals been used it is more than likely that a significant increase in
adrenal weight with increased body weight could have been shown. Other
data on over 1200 wild Norway rats shown conclusively that there is a
definite increase in adrenal weight with increasing body weights varying
from 50 through 600 gm., and that there was a significant linear relationship
between the logarithm of the adrenal weight and body length or weight
(Christian, 1954; Christian and Davis, 1955) . Finally, it is well known that
there is a definite tendency for the larger animals to be dominant, therefore
adrenal hypertrophy in smaller, subordinate rats would tend to make the
adrenals of dominant and subordinate animals weigh the same. It is of
considerable interest, however, that rats from colonies of mixed sex were
appreciably larger than those from all-male colonies. In general the amount
of sudanophilia in the zona fasciculata coincided with the weight data.
Nevertheless, Barnett's results appear to agree with those established earlier
as well as with the results of the experiments of other investigators.

The above experiments warrant the general conclusion that in house mice
and Norway rats adrenal weight tends to be inversely related to social
rank in that dominant animals exhibit little or no increase in adrenal weight,
while subordinate animals show a marked increase. A word of caution should
be directed with regard to the interpretation of results with adrenal weights.
A hyperactive adrenal with a loss of cortical lipids may weigh less than a
less active gland containing a large amount of lipids. This has been shown
to be the case in mice (Christian, 1955a, 1959b) and evidently is also true
in Barnett's (1958) experiments in which the adrenals of his "interloper"

2. Endocrines and Populations 267

rats weighed somewhat less than those of subordinate rats, but the adrenals
of the interlopers contained little or no lipid.

The relationship between social dominance and adrenal activity is,
however, not as clear cut as these experiments at first indicate and suggests
that a great deal of work needs to be done on the factors which affect
dominance-subordinance relationships and w^hat constitutes "social stress"
in mammals. The adrenals of the dominant animal in a group do not always
weigh the least and occasionally may be the heaviest, even though the
average weight of the adrenals of the dominant animals is appreciably less
than those of other ranks. Furthermore, the amount of fighting varies
greatly from group to group. Finally, the amount of scarring on a mouse in
a group is in a general way a measure of its rank. Observation has shown
that it is generally safe to assume that the unscarred mouse in a group is
the dominant animal, and yet, as we have shown, there is no relationship
between the amount of scarring and absolute adrenal weight in groups of
albino male mice although there was clearly a tendency for adrenal weight
relative to body weight to increase with increased scarring (Christian,
1959d) . These results apparently indicate that fighting is a poor measure of
social rank and is related only indirectly to it. These results appear to
contradict the results of other experiments, although these differences
would probably be reconcilable if more detailed information were available
on social behavior in these animals. Rather rigid and simple criteria are
used to determine rank in animals, such as physical dominance-subordi-
nance relationships, and it is quite clear that social interactions are far
more complex than this. However, the apparently contradictory results
with relation to fighting and dominance in no way invalidate the conclu-
sions that purely psychological social pressures are responsible for stimu-
lating increased adrenocortical and decreased reproductive activity in
groups of mice and rats and that fighting per se has little or no effect on
the adrenal hypertrophy observed in groups of animals. One can speculate
with some reason that a massive stimulation of the nervous system and
adrenal medulla occurs in subordinate animals when they are suddenly
confronted with a dominant, aggressive male. Observation shows that
these animals are aware of their rank and cower in front of the dominant
animal. When there is no escape from constant contact, it seems inevitable
that the subordinate animals must suffer from emotional anxiety resulting
from a desire to escape from the situation and the inability to do so. Chronic
stimulation of the pituitary-adrenocortical system presumably results from
the chronic continuation of such a situation. Admittedly this somewhat
anthropomorphic interpretation of the situation is largely speculative, but
observations of mice in groups inevitably lead one to such a conclusion.

Finally, Mason (1959) has shown that when rhesus monkeys are kept in

268 /. /• Christian

groups for prolonged periods of time there is a significant increase in their
urinary tetrahydroxycorticoids, the metabolites of the carbohydrate-active
corticoids, excreted by grouped monkeys is greater than their combined
daily production of corticoids when individually caged. Furthermore, the
production of corticoids remains high for the entire period of grouping. It
has also been observed that when human bomber crews are housed as a
group, their production of urinary corticoids is increased over the combined
individual production of corticoids by the same men, paralleling the findings
for monkeys (Mason, 1959) . In the case of monkeys and men there can be
no question that the stimulus to increased corticosteroid production is
psychological, resulting from social interactions.

There can be little doubt that social pressures can increase pituitary-
adrenocortical activity. We have also suggested that there is a depression of
reproductive function in male mice, as indicated by the weights of the sex
accessories, coinciding with increased adrenocortical function in relation to
social factors. These results are in agreement with the earlier work of Crew
and Mirskaia ( 1931 ) and Retzlaff ( 1938) , who showed that increased popu-
lation density depresses reproduction in female mice. The reproductive
performance of female mice was inversely related to population sizes in
populations of 1, 4, 8, or 12 pairs. Retzlaff (1938) also indicated that repro-
ductive performance was best in the socially dominant females in each
population. He made several additional observations of interest in these
experiments. He noted that there were aggressively dominant females that
attacked and viciously fought introduced females or females which had
been removed and were later replaced. The subordinate females suffered
death or injury, and, of particular interest, any infections that they had
were greatly exaggerated. This is one of the early experunental indications
of decreased resistance to disease following social stress, and is further
evidence that resistance is decreased primarily in the subordinate animals.
He also noted that in the largest populations, 12 pairs of mice, there was a
sufficient confusion among the mice to offer the subordinate animals partial
protection from attack. Similar effects were later noted in populations of
32 male mice (Christian, 1955a, b). Finally, it was found that a reduction
in environmental temperature of approximately 16° C. resulted in a signifi-
cant decline in mean litter size for mice from populations of 1, 2, or 8 pairs,
but not from populations of 4 or 12. The protection against a reduction in
litter size by decreased temperature probably reflects huddling as a means
of maintaining body heat and therefore diminishing the need for increasing
thyroid activity. Retzlaff could not explain the reduction in populations of
8, but evidently it was due to severe social strife rather than the reduction
in temperature. It is conceivable that severe strife prevented huddling due
to mutual intolerance; so that both factors could play a causative role in

2. Endocrines and Populations 269

reducing reproductive performance. The preceding experiments usually-
used populations consisting entirely of male or mixed male and female mice
or rats (Crew and Mirskaia, 1931 ; Retzlaff, 1938; Barnett, 1958) . However,
a depression of reproductive function occurs when only female mice are
placed in groups (Andervont, 1944). In these latter experiments estrous
cycles began at an earlier age, were more frequent, and lasted until a
greater age in segregated female mice than in their littermates kept in
groups of 8 each. Bullough (1952) showed that "overcrowding" mice (16 to
a cage) for 3 weeks resulted in a 30% increase in the cross-sectional area of
the adrenal cortex and an 80% increase in medullary area. These changes
were accompanied by a 60% reduction in epidermal mitoses, which was
attributed to an increased secretion of adrenocortical steroids. Finally,
Chitty (1955) showed that liver glycogen was appreciably lower in voles
maintained in the laboratory in groups than in those maintained under
segregated conditions.

2. Adrenocortical and Reproductive Responses to Population
Density

The preceding experiments indicate that sociopsychologic factors in-
volved in social interactions between mammals can elicit physiologic
adaptive responses with increased pituitary-adrenocortical and decreased
reproductive function in voles, house mice, both wild and albino, Norway
rats, rhesus monkeys, and humans. However, with the exception of the
experiments of Crew and Mirskaia (1931) and Retzlaff (1938) there was
no indication that these functions were altered in relation to population
density. Although an inverse relation between population size and reproduc-
tive performance was shown in the experiments of Crew and Mirskaia
(1931) and Retzlaff (1938), these authors did not investigate adrenocorti-
cal function. If social competition, as a stimulus to increased pituitary-
adrenocortical and decreased reproductive activity, is responsible for regu-
lating the growth and decline of mammalian populations, then there must
be a relationship between population density and the magnitude of the
endocrine responses. The existence of such a relationship has been demon-
strated in a variety of experiments.

If male mice which have been segregated since weaning are placed to-
gether in groups of 1, 4, 8, 16, or 32 per cage for a week, there is a hyper-
trophy of the adrenal glands and atrophy of the gonads and sex accessories
which progresses more or less linearly as the logarithm of the population
increases (Christian, 1959b) . There is a decline in thymus weight from that
of the isolated controls, but the decrease is not related to the population
size. The increase in adrenal weight was found to result primarily from

270 /• J- Christian

cellular hyperplasia and hypertrophy of the zona f asciculata, although the
glands were not critically examined for medullary hypertrophy. The decline
in the weights of the testes was a reflection of the generalized reduction in
body weight in one series of experiments (Christian, 1955a), but in another
series the testes declined in relative testicular weight as well as in absolute
weight with increasing population size (Christian, 1955b). Body weight
was significantly less in populations of 4, 8, and 16.

These experiments were repeated using wild house mice raised in the
laboratory in populations of 1, 3, 4, 6, 8, 9, and 17 with similar though
more pronounced results (Christian, 1955b). The thymus weight of these
mice decreased markedly as population size increased. The adrenocortical
response was also much more pronounced in these mice than in the albino
mice, although the mean weights of the adrenals of segregated albino and
brown male mice were identical. The increase in adrenal weight of albino
mice reached a maximum of 8% above the control levels in populations of
16, whereas that of the wild stock attained a maximum increase of 21%
above the control levels in populations of 9. In populations of 32 albino
mice and 17 wild mice the mean adrenal weight was less than in the next
smaller population size in each case. This was found to be due to a loss of
lipid and a marked decrease in the size of cells in the zona fasciculata al-
though the degree of hyperplasia was greater than in the adrenals of mice
from the preceding population size (Christian, 1959b). The wild mice are
much more alert, reactive, and aggressive than albino mice, and the
differences in adrenal reactivity probably reflect such behavioral differences.

The decline in the weights of the accessory reproductive organs, seminal
vesicles, and preputial glands indicated a diminished secretion of androgens
from the testes with increasing population size, based on the assumption
that these organs accurately reflect androgen levels (Burrows, 1949; Ren-
nels ct al, 1953) . This conclusion is strengthened by the fact that relative
testicular weight also declined in one group of experiments (Christian,
1955b) and absolute weight declined in all experiments with increased
population size. Since all indications point to a decline in the secretion of
androgens, the decline in thymus weight must therefore represent involu-
tion by increased amounts of circulating corticoids.

The progressive hypertrophy of the adrenals and atrophy of the thymus
and reproductive organs do not reflect a diminished space per mouse per
se, as the results were essentially identical where populations of 1, 4, 6, 8,
and 17 each of male wild mice were placed in cages with 42 times the area
of the cages used in the preceding experiments (Christian, 1959b). The
similarity of the results from these two series of experiments with vastly
different amounts of area per mouse also suggests that the amount of
exercise or activity was not a factor in the observed endocrine responses.

2. Endocrines and Populations 271

It has been established that close confinement with adequate feeding can
result in adrenocortical atrophy and lymphoid hypertrophy in a variety of
species of mammals (Christian and Ratcliffe, 1952), but these limits of
confinement evidently were not approached in the above experiments, as
indicated by the adrenal weights in the large and small cages.

These results apply to male mice. When female mice are grouped, there
is evidence of increased adrenocortical activity, increased corticoids and
androgens being secreted (Christian, 1960). The degree of hypertrophy is
small, however being nowhere nearly the amount seen in grouped male mice
or even in females from populations of mixed sex. The inference is that the
cortical hypertrophy seen in females from populations of mixed sex must to
a large degree reflect a situation created primarily by the males.

3. Food and Social Competition; Splenic Hypertrophy

In all the experiments with mice so far discussed, food and water were
provided ad libitum. Food was scattered over the cage and water was
available from several sources in order to avoid competition for food which
might result in inanition in the subordinate animals and might constitute
stimuli to the endocrine adaptive responses. These precautions were taken
even though observation had indicated that feeding and drinking were
more or less random and on an individual basis, and that there was no
observable competition for these items. Also Uhrich (1938) and Strecker
and Emlen (1953) had indicated that a limited supply of food did not in-
crease competition among house mice. Nevertheless, the question whether
or not the location and amount of food was a major factor in eliciting the
observed changes in grouped mice in the pituitary-adrenal and pituitary-
gonadal systems was answered by specifically designed experiments. In
one series of experiments, male mice, some in groups of four each and some
segregated, were provided food either ad libitum or limited to 4.0 gm. per
mouse per day for a 7-day experimental period. This amount of food re-
stricted weight gain but did not produce weight loss. The food given the
grouped animals, whether limited or ad libitum, was scattered for half of
the populations and supplied from a feeder for the other half. Therefore
food was given the mice in one of the four following ways: scattered and
limited in amounts, from a feeder and limited, scattered and ad libitum, or
from a feeder ad libitum. The experiment was repeated using groups of
five each and a food limitation of 3.5 gm. per mouse per day, an amount
which produced an appreciable weight loss in all the mice during the week
of the experiment. This amount of food per mouse per day can therefore be
said to produce inanition. It was found that neither the location of food,
whether scattered or from a feeder, nor the amount of food had any effect

272 J. J. Christian

on adrenal weight in the segregated or grouped mice. (Irouping produced
significant adrenal and splenic hypertrophy along with atrophy of the
thymus and reproductive organs, and these changes were not altered by
the source or amount of food. The splenic hypertrophy produced in the
albino house mice in these experiments by grouping paralleled the splenic
hypertrophy in voles following social "stress" (Clarke, 1953; Chitty et at.,
1956). However, restricting the amount of food resulted in an atrophy of
the reproductive organs, thymus, and spleen which was related to the
degree of food restriction. The source of food was without effect on the
adrenals, thymus, spleen, or reproductive organs. Therefore the source or
amount of food did not increase the level of social competition above that
already present as a result of the establishment of a social order. The
amount of food provided had no effect on the adrenal glands, therefore it
was concluded from these and the experiments of others (Mulinos and
Pomerantz, 1941; D'Angelo ct al., 1948; Baker, 1952) that inanition does
not constitute a stimulus to the pituitary-adrenocortical system of house
mice and rats.

On the other hand, Frank's (1953) experiments with confined popula-
tions of meadow voles (Microtus arvalis) suggest that a deficient supply of
food may increase social competition in this species, as a marked increase
in fighting occurred following the development of a food shortage. Whether
or not a food shortage will increase competition very likely depends on the
time relationships of the feeding behavior of the species. One would not
expect to find increased competition among animals which feed randomly
with respect to time, otherwise one would have to attribute to these animals
the ability to predict, ahead of time, the development of a shortage, as
there would be no appreciable increase in competition for food at any given
moment. This consideration also implies that dominant and subordinate
animals would lose weight equally in the presence of a deficient food supply.
Such was actually the case in the experiments with house mice (Christian,
1959c) . On the other hand, an increase in competition would be expected to
follow the development of a food shortage among mammals which habi-
tually feed during the same period of time every day. In this situation one
would not expect the dominant animals to lose weight to the same degree
as the subordinate animals, if at all. Frank's (1953) results coincide with
this latter situation. If these conjectures are correct, it is apparent that the
temporal feeding relationship and behavior of any species will assume para-
mount importance with respect to the production of competition and the
physiologic responses to it, and should therefore be investigated critically
for a variety of mammals. The experiments with house mice should be re-
peated with voles and other species and any increases in competition deter-
mined by observation and the accompanying adrenocortical responses.

2. Endocrines and Populations 273

Although inanition and starvation may not be stimuli to increased
pituitary-adrenocortical activity in mice and rats, they may be in guinea
pigs (D'Angelo et al., 1948) and white-footed mice {Pero7nyscus leucopus)
(Sealander, 1950). However, a limited amount of food does depress repro-
ductive activity in house mice as well as in other species which have been
investigated (Lutwak-Mann, 1958), possibly as a resvilt of a protein de-
ficiency, as it has been shown that a dietary deficiency of protein diminishes
the secretion of gonadotropins (Srebnik et al., 1958; Lutwak-Mann, 1958).
The striking declines in the weights of the preputial glands, and especially
of the seminal vesicles, brought about by a limited amount of food (Chris-
tian, 1959c) indicate that the secretion of androgens by the testes was
markedly depressed. Whatever the mechanism by which a food deficiency
depresses the secretion of gonadotropins, it was not by eliciting a generalized
adaptive response involving the pituitary-adrenocortical system. These
results lead to several conclusions, at least as far as mice are concerned.
One is that when food is supplied ad libitum, competition for food is not a
factor in eliciting pituitary-adrenocortical responses in relation to popula-
tion. A second is that inanition and starvation per se are not stimuli to
increased pituitary adrenocortical activity in mice or rats. The third is
that food restriction depresses reproduction by dimmishing the secretion
of gonadotropins without eliciting a more widespread response, and there-
fore a limited supply of food can limit population growth specifically by
depressing reproduction without operating through the pituitary-adreno-
cortical system. However, we have seen that there is some indication that
these relationships may not be universally true. Perhaps food supplies and
social competition can act independently to limit population growth. Food
conceivably may not be important to natural populations as long as sub-
ordmate animals are free to move elsewhere. Calhoun (1949, 1950) has
indicated that social competition is a more important factor than food
supply per se for Norway rats. One fact is abundantly clear as a result of
these experiments, and that is that sociopsychologic factors stimulate in-
creased pituitary-adrenocortical function and depress reproductive func-
tion in proportion to population density ; this reaction system is therefore
active at all levels of population in the control of population growth,
whereas a limitation in the food supply will exert its effects on reproduction
only when it results in inanition in the members of a population. These
considerations are of obvious importance to the investigator interested in
the control and regulation of mammalian populations.

The production of splenic hypertrophy in highly inbred albino mice by
increased population density is especially interesting in view of Chitty's
(1957) conclusion that the splenic hypertrophy in voles {Microtus agrestis
and M . orcadensis) was due to inherited genetic factors. This conclusion

274 J. J. Christian

was based on the fact that splenic hypertrophy, which had been observed
previously only in natural populations of high density, had made its ap-
pearance recently in laboratory stocks of these voles. The only factor which
seemed to account for this appearance was the introduction into the labora-
tory breeding population of voles from natural populations exhibiting
splenic enlargement. The fact that a similar hypertrophj' was observed in
response to increased population density in a highly inbred strain of mice,
maintained under constant conditions for a good many years with no expo-
sure to natiu-al conditions, makes a genetic explanation of the sort postu-
lated by D. Chitty unlikely in such an inbred strain of presumptively gene-
tically stable mice. Similarly, as was pointed out earlier, a comparable
hypertrophy of the spleen was observed in a few inbred rats subjected to
alarming stimuli. The hypertrophy, when examined critically, has been
found to be due to increased hematopoiesis (Dawson, 1956), possibly in
response to the increased stimulation of erythropoiesis by hormonal factors.
It is also well known that splenic erythropoiesis can be stimulated by any
stimulus that produces anemia. Therefore, conclusions regarding the causes
of splenic hypertrophy, due to increased erythropoiesis, must be inter-
preted with caution; although it does not seem likely that it is genetic in
origin, at least in inbred albino mice with splenic hypertrophy following
increased population density.

4. Reproductive Fuxctiox in Female Mice: Lactation, Reproduc-
tion

Female mammals have frequently been observed to respond to adverse
stimuli with a reduction of reproductive function. When rats are suddenly
moved from a temperature to which they have become accustomed to a
different temperature, either higher or lower, there is retardation of growth
and prolongation of the estrous cycle (Bohanan, 1939). Selye (1939) has
listed a variety of agents which will inhibit o^'arian function and estrus.
Reproduction is also depressed in female mice by increased population
density. We have noted that Crew and Mirskaia (1931) and Retzlaff (1938)
found that reproductive performance of female albino mice declined with
increasing population density. In another series of experiments no young
were produced and no females became visibly pregnant when mice were
crowded 20 males and 20 females to a cage for 6 weeks (Christian and Le-
Munyan, 1958) . It is not known whether there was a marked suppression
of ovarian function with diminished ovulation, a failure of the ova to im-
plant, or intra-uterine loss early in pregnancy, but more than likely all
these factors were involved. When the population size was reduced to 10
males and 10 females, all the females became pregnant, but the number of

2. Endocrines and Populations 275

implanted ova was reduced significantly and only 7 of the 10 females
delivered young. The remainder lost their progeny in utero during the early
stages of pregnancy. The onset of pregnancy also was considerably delayed
in these animals. Therefore there was decreased fertility, decreased im-
plantation, and a marked increase in intra-uterine mortality. These results
indicate that female mice respond to increased population density with a
depression of reproductive function at all stages of the processes. These
results correspond to those seen in males as indicated by the weights of
their reproductive organs. Reproductive suppression was also observed
when female mice were grouped without males (Andervont, 1944; Whitten,
1959; Christian, 1960).

Chitty (1952) noticed that young voles from natural populations of high
density were reduced in size and were unusually susceptible to increased
mortality. He hypothesized that the young were adversely affected in utero
by the physiologic derangements in the mothers which resulted from high
population densities (Chitty, 1952, 1954). Later it was shown in a limited
series of experiments that social "stress" diminished lactation in voles, as
measured by the weights of progeny nursed by the mothers subjected to
social pressures which had previously been shown to result in increased
pituitary-adrenocortical activity (Chitty, 1955). These experiments were
repeated on a larger scale using laboratory white mice (Christian and Le-
Munyan, 1958) . It was found that progeny nursed by previously crowded
mothers weighed appreciably less at weaning than those nursed by mothers
which had always been segregated. The effect on the progeny was greater
in the larger litter sizes, suggesting a quantitative rather than a qualitative
deficiency in the supply of milk. When these young which had been nursed
by crowded mothers were themselves bred, the progeny which they in turn
nursed were significantly lighter at weaning than their controls. Again the
defect was greatest in the larger litters, but the difference was not manifest
until a larger litter size had been reached than in the preceding generation.
These differences in the second filial generation cannot be attributed to
grouping. Crowding depresses all the other reproductive functions, therefore
it is not surprising that lactation is also suppressed, as it is to a large degree
under the control of pituitary gonadotropins and sex steroids in addition
to oxytocin, thyroxine, and growth hormone (Folley, 1956; Grosvenor and
Turner, 1959a, b, c,). These results are consistent with others which show
that lactation can be inhibited by a variety of stimuli which stimulate in-
creased pituitary-adrenocortical activity and diminish increased reproduc-
tion and growth (Selye, 1954) ; in fact, lactation can be limited by a defi-
ciency of any one of a variety of hormones necessary for its fulfillment
(Grosvenor and Turner, 1959a, b, c). Since the young mice in litters of
small size were unaffected in the above experiments, it is unlikely that any

276 /. /. Christian

substance, such as adrenal hormones, contained in the milk were responsible
for the decrease in weights of the progeny. The mechanism by which these
effects were carried over into the second generation of progeny is not known.
Presumably the second generation of young may reflect the inanition suf-
fered by the first generation of progeny as a result of deficient lactation.
However, the fact that androgens can produce subsequent permanent
sterility when injected into mice less than 10 days old suggests the possibil-
ity that increased amounts of adrenal androgens may reach the nursing
young via the milk and exert similar partial effects if some escape metabo-
lism in the liver. The weights of the young at birth were unaffected by the
earlier crowding of their mothers. Although the precise mechanisms are not
understood, these results support and extend D. Chitty's results with voles
and help in providing an explanation for the prolonged effects of high
density on surviving young observed in natural populations, such as
Chitty's (1952, 1954) observations on young voles from natural popula-
tions.

5. Growth

The effects of increased population density on growth have not been
studied to the same degree as its effects on other aspects of endocrine
physiology, although suppression of the secretion of pituitary growth
hormone presumably is a part of the response to stimuli which also result
in the increased secretion of ACTH (Selye, 1950) . However, there are a few
experiments which clearlj^ indicate that the growth of house mice is de-
pressed in response to increased population density (Vetulani, 1931;
Christian, 1955b) . It has been shown also that there is suppression of
growth in all but the dominant and second-ranking mice in a group of six
(Christian, 1961). The degree of inhibition was related to rank.

6. Inflammation, Resistance to Infection, and Population Den-
sity

In earlier sections of this account the inhibition of the inflammatory re-
response, granulation, and antibody formation, and therefore resistance to
disease, by adrenal carbohydrate-active corticoids was discussed. Most of
the experimental evidence cited was based on the results of injecting corti-
cal hormones. A few examples of decreased resistance to infectious agents
brought about by stimuli, such as cold, which stimulated increased adreno-
cortical secretion were given. However, even though these experiments
indicated that such effects might occm- under natural circumstances, they
did not establish this possibility, and especially they did not show that
increased population density could stimulate a sufficient increase in the

2. Endocrines and Populations 277

secretion of adrenal carbohydrate-active corticoids to inhibit inflammation,
granulation, and antibody formation sufficiently to decrease resistance to
disease, although Retzlaff (1938) had indicated that there was decreased
resistance to infection in the subordinate mice in his population studies.
Experiments have been conducted which do establish these points to a
limited degree, but more experiments with more species are needed.

An efficient method of inducing an inflammatory response and the forma-
tion of granulation tissue in rats or mice is to implant subcutaneously cotton
pellets moistened with turpentine (Meier et at., 1950; Christian and Wil-
liamson, 1958). Later these pellets and the surrounding tissue can be re-
moved and weighed. In this fashion the degree of the inflammatory response
and formation of granulation tissue can be assayed under a variety of
conditions. This technique has been used to show that ACTH and the
adrenal carbohydrate-active corticoids suppress these responses (Meier
ct al., 1950) . By this procedure it was found that when mice w^ere placed
in groups of 5 each there was approximately a 20% reduction in the amount
of granulation tissue formed in 1 week compared to the amount formed by
the segregated control mice (Christian and Williamson, 1958). These
experiments clearly indicated that increased population densit}^ is a suffi-
cient stimulus to the pituitary-adrenocortical system to decrease signifi-
cantly inflammation and granulation.

Davis and Read (1958) conducted a series of related experiments in
which they demonstrated that placing wild-stock house mice in groups
markedly increased the susceptibility of the mice to invasion by the larvae
of TrichincUa spiralis. Each mouse was infected parenterally with approxi-
mately 125 embryonated Trichinella larvae. Each mouse was maintained in
a separate cage, but from day 3 through 11 after infection 11 of the mice
were placed in two groups, one of 5 and the other of 6 mice, for 3 hours a
day, while 11 others were left segregated. The mice were sacrificed the 15th
day after infection, the gastrointestinal tracts were digested, and the larval
worms were recovered. Only 3 of the segregated mice were infected with an
average of 9 worms apiece, whereas all the grouped mice were infected with
an average of 32 worms each, a 250% increase. The experiment was re-
peated with 6 segregated and 5 grouped mice, but they were not sacrificed
until the 30th day post infection, and the encysted larvae were recovered.
All the mice, segregated and grouped, were infected, but the grouped mice
had 48% more encysted larvae than their segregated controls. In both
experiments the differences in the number of worms between segregated
and grouped mice were highly significant and in both experiments the mean
adrenal weight was greater in the grouped than in the segregated mice, 8%
iP < 0.20) in the first experiment and 20% (P < 0.01) in the second
experiment.

278 /. /, Christian

In earlier experiments it had been shown that either cortisone or ACTH
increased the invasiveness of Trichinella larvae by suppressing the defen-
sive inflammatory response of the host's intestinal wall and possibly by
prolonging the sojourn of the adult females in the gut by suppressmg im-
mune responses to the worms (Stoner and Godwin, 1953). Thus, these
hormones decreased the resistance of mice to invasion by the larvae of
Trichinella spiralis by inhibiting inflammation and possibly antibody
formation. Grouping evidently stimulated a sufficient increase in the secre-
tion of adrenal corticoids to produce similar effects. This conclusion is sup-
ported by the demonstration that grouping was sufficient stimulus to
pituitary-adrenocortical activity to diminish granuloma formation appreci-
ably. These experiments with trichinosis in house mice have been repeated
with albino mice with similar results, confirming the original results and
demonstrating that albino and wild-strain house mice react similarly (Davis
and Read, unpublished) .-

The effects of crowding on trichinosis in mice is presumably primarily due
to the effects of increased adrenocortical secretion on the inflammatory
response to the worms. Therefore a similar series of experiments were per-
formed which were more specificaUy designed to demonstrate the effects of
grouping on antibody formation (Davis and Prudovsky, 1959). In this case
mice were injected with tetanus toxoid and challenged 10 days later with
13 mouse MLD of tetanus toxin. The dosage of toxin was selected to cause
death in 50% of segregated mice. Mice were placed in groups either 5 or 3
days prior to giving the toxoid and left in groups until 5 days after the
toxoid. Control mice were left segregated in individual cages throughout
the experimental period but otherwise were treated identically to the experi-
mental animals. Grouping appreciably decreased resistance and increased
mortality of mice to the challenge dose over the control levels. This effect
was less marked when the mice were grouped 3 days before administering
the toxoid. The effects of grouping resembled the effects of injected corti-
sone at the appropriate times and in appropriate doses. These results indi-

^ A report was published recently which implied that cortisone or ACTH are without
effect on the course of trichinosis in mice, but which showed that cortisone markedl}'
increased mortality (Lord, 1958). These experiments require some critical comment.
The procedure used to infect the mice was similar to that used by Davis and Read
(1958), but injections of cortisone and ACTH were not begun until 6 days after infection.
One would not anticipate that these hormones would affect the course of trichinosis at
that late date, as invasion by the larval worms, the inflammatorj' response to them in
the intestinal wall, and immune responses to them are fairly well accomplished facts
by that time. Therefore one would not anticipate an alteration in the resistance to
invasion by the worms. The doses of cortisone used in these experiments were extremely
high pharmacologic doses: 76 mg. in 39 days, beginning with 3 mg. a day per mouse.
It is not surprising that cortisone increased mortality.

2. Endocrines and Populations 279

cate that the adrenocortical response to grouping is sufficient to depress
antibody formation as well as inflammation. From these results one can
easily sec that behavioral factors associated with increases in population
density, or grouping, can sufficiently increase adrenocortical activity to
profoundly affect host resistance to infectious disease and parasites.

The effects of hormones on miu-me tuberculosis are somewhat complex.
Cortisone enhances tuberculosis infections m mice, especially by converthig
a smoldering chronic infection into a fulminating acute process (Hart and
Rees, 1950) . However, ACTH was without effect on the early development
of tuberculosis in mice (LeMaistre et al., 1953). From these experiments it
would appear that an increased secretion of endogenous adrenal corticoids
was ineffective in enhancing tuberculosis in mice whereas injected carbohy-
drate-active corticoids markedly enhanced the infection. Therefore, it is
especially interesting that Tobach and Block (1956) were able to show that
crowding significantly altered the courses of acute and chronic tubercu-
losis m mice. Crowding after mfection decreased the survival time of mice
of both sexes suffering from an acute tuberculous infection. Crowding after
infection had essentially no effect on the course of chronic tuberculosis in
female mice, whereas it enhanced the chronic disease in male mice. These
experiments also indicate that crowding prior to infection may enhance
host resistance.

These experiments, although quite limited in number, all confirm the
fact that crowding (increased population density) can reduce host resist-
ance to disease. The evidence from experiments with increased popula-
tion density is completely in accord with the results of experiments with
injected corticoids. These results, considered in the light of the mass
of this evidence on the effects of corticoids and ACTH on experimental
infections, can only mean that increased population density sufficiently
stimulates the pituitary-adrenocortical system to lower resistance to disease
by inhibition of inflammation, granulation, and probably antibody forma-
tion by endogenous corticoids.

7. Production of Mortality Directly

In the preceding section the role of behavioral factors and increases in
population density in decreasing resistance to disease was discussed. It
therefore follows that grouping can increase mortality by this means and
the more subordinate animals will be the ones most often affected. However,
grouping is often followed in a very short tune by sudden death preceded by
alternate prostration and convulsions of many of the animals (Christian,
1955b), and the greater the number which are placed together, the greater
the number which succumb in this fashion. (J. J. Christian, unpublished).

280 J. J. Christian

although there is considerable variation from population to population in
the numbers which die in populations of the same size. The more reactive
and aggressive wild-stock house mice are much more prone to succumb in
this fashion than the usual albino mouse in the laboratory (Christian,
1955b). The proximate cause of sudden death in these animals has not been
investigated, although its onset, behavior, and general symptomatology arc
highly suggestive of hypoglycemic shock and closely resembles the deaths
from "shock disease" observed in natural populations at peak densities
(Christian, 1950b). Frank (1953) observed similar deaths in voles {Micro-
tus arvalis) after placing them in groups, as well as in natural populations,
and was able to demonstrate that glucose could prevent the fatal termina-
tion of the syndrome, and also showed that the symptomatology of insulin
hypoglycemia was identical to that observed in voles dying naturally
following "crowding." Furthermore, he showed that the voles which were
on the verge of dying had markedly decreased levels of blood sugar and
liver glycogen. These results make it appear probable that the convulsions
and death which follow shortly on the social strife produced by grouping
mice are due to hypoglycemic shock. There is reason to suggest that a
massive discharge of the adrenal medulla, with its subsequent exhaustion
may lead to a failure to further mobilize glucose and the animals die in a
hypoglycemic episode. However, this hypothesis requires further testing.

In any event, social strife and the physiologic responses to it may result
in mortality directly as well as by affecting host resistance.

8. Summary of Results from Populations of Limited Size

Experiments with populations of fixed size have demonstrated that socio-
psychologic interactions can and do stimulate increased pituitary-adreno-
cortical and decreased reproductive activity and growth in mice, voles, and
Norway rats. Increased adrenocortical activity is related to social domi-
nance-siibordinance relationships, the more subordinate animals exhibiting
a greater response than the dominant animals. These responses apparentl}'
are unrelated to fighting per se. Fighting is evidently a sign of social interac-
tion, just as is the endocrine response, and not a causative factor. There is
a progressive increase in the pituitary-adrenocortical response and de-
crease in reproductive function in male and female mice with increasing
population size. All phases of reproductive activity are depressed, including
fertility, implantation of blastulae, intra-uterine survival, estrus, lactation
(with persistent effects for at least two generations) , and the size and
activity of the male gonads and sex accessories. Inhibition of reproductive
fimction apparently stems from a depressed secretion of pituitary gonado-
tropins. Growth is also inhibited by increased population density.

2. Endocrines and Populations 281

In addition to these primary effects of physiologic adaptation to the
social pressures associated with increased population density, there is a
decrease in resistance to parasitism and infection. It was shown that in-
creased density produced an inhibition of inflammation, granulation, and
probably antibody formation. A decrease in host resistance to trichinosis
and tuberculosis was observed which apparently was due to these factors.

It was also shown that food was not a factor in these effects, either in-
directly by increasing social competition or directly by inanition in the
subordinate animals. A deficiency of food is not a stimulus to increased
adrenocortical activity in mice, or rats, although it may be in other species.
A food deficiency, however, does result in a marked suppression of reproduc-
tive function, probably by decreasing the production of pituitary gonado-
tropins.

B. Freely Growing Populations

1. Introduction

Experiments with populations of limited size, although indicative, do not
establish that these same responses occur in freely growing populations. In
the first place, experiments with limited populations were conducted for
limited periods of time, usually too brief to permit adaptation to the situa-
tion. Furthermore, in every case the experimental approach involved sud-
denly placing strange animals together under rather artificial circum-
stances. Therefore, these experiments provided no evidence that mammals
which have been more or less in contact with each other from birth would
respond similarly. House mice {Mus musculus) and voles {Microtus penn-
sylvanicus, Microtus agrestis) from free growing populations show the same
physiologic responses to increased population density as do mice from
populations of fixed size but to an even greater degree.

2. Population Density and Adrenocortical Function

Increased pituitary-adrenocortical activity has been related to increased
population density in a number of experiments with freely growing popula-
tions of house mice and voles (Christian, 1956: Louch, 1956; Christian,
1959a, b). A number of experimenters have shown that the growth of
freely growing confined populations is self-limited in spite of the fact that
food, water, nesting material, and nesting space were provided ad libitum
and well scattered (Strecker and Emlen, 1953; Clarke, 1955; Southwick,
1955a; Christian, 1956; Louch, 1956; Crowcroft and Rowe, 1957; Christian,
1959b) . Competition for food or a lack of availability of food by particular
individuals was not a factor (Christian, 1956; Crowcroft and Rowe, 1957).

282 /. /. Christian

The growth form of these populations was approxunately sigmoid, indi-
cating that an intrinsic damping factor was operating to regulate and limit
the growth of populations throughout their histories. These populations
were started by introducing a small number of animals of both sexes into
confined quarters and allowing the population to grow of its own accord.
The populations were either sacrificed, at maxunal and submaximal levels,
to obtain organ weights and histologic material, or blood samples were
taken for eosinophil counts to assess functional changes, especially of the
adrenal cortex.

The zonae fasciculatae of the adrenal cortices were hyperplastic and
hypertrophic in house mice of both sexes from populations of maximum
(asymptotic) size. The number and size of the cells of the fasciculata were
increased. Adrenal weight reflected the fascicular hypertrophy by increases
of 25% in the males and 14% in the females. The adrenal cortical hyper-
trophy was approximately half as great in mice from populations of approxi-
mately one half the maximiun size.

The presence in mice of an adrenocortical X-zone which is involuted by
androgens has been described. This zone complicates the interpretation of
adrenal weight from immature or puberal male or nulliparous female house
mice unless histologic studies accompany the data on weights. In the case
of the freely growing populations of house mice there was a pronounced
hump in the adrenal weights in relation to body weights for mice from the
experimental populations in the 13-19.0-gm. weight range, even though the
fascicular hypertrophy was proportionately constant for all weights. The
curve of adrenal weight on body weight for the segregated control mice was
more or less regular with no pronounced irregularities, although there is a
tendency for the slope to decrease with increasing body size. This appar-
ently excessive adrenal hypertrophy in 13-19-gm. mice from intermediate
and high populations resulted from a failure of the X-zone to involute
normally rather than from a true hyperplasia or hypertrophy. The width of
the X-zone in segregated male house mice begins to decline in mice in the
10-12.9-gm. weight group, and involution is essentially complete by the
time a body weight of 16 gm. is reached. The decline did not begin in the
experimental animals until a weight of 16 gm. or more had been reached
and was not complete until a weight of 19 gm. Therefore, along with a
marked hyperplasia of the zona fasciculata, suggesting an increased secre-
tion of corticoids with increased density, there was a delay in the onset of
puberty, presumably with an insufficient production of androgens to in-
volute the X-zone of male mice. A great variety of steroid hormones, in-
cluding corticoids, have been tested for their ability to involute the X-zone
in male house mice (McPhail and Read, 19-42b; Antopol, 1953; Allen, 1954;
Christian, 1954) , but only those with pronounced androgenic activity (e.g.,

2. Endocrines and Populations 283

testosterone) have been effective, although Delost (1954) has reported that
cortisone involutes the X-zone in voles. This delay in androgen production
presumably indicated an inhibited secretion of gonadotropins from the
anterior pituitary, although there apparently were sufficient gonadotropins
to maintain the X-zone, as luteinizing hormone presumably is responsible
for maintaining this zone (Jones, 1949b, 1950, 1952).

Thymic involution is effected by carbohydrate-active corticoids, estro-
gens, and androgens, with variations in the mode of involution (cf . above) .
Therefore thymic involution, as a measure of increased adrenocortical
secretion, must be interpreted with caution if there is reason to suspect
differences in the levels of circulating sex steroids. Such was the case in
the experiments under discussion, but additional information makes it
possible to state with reasonable certainty that the weights of the thymus
reflect increased corticoid secretion. The thymuses of 13-19-gm. male mice
from high populations weighed more than their segregated controls, whereas
the mean thymus weight of 19-23-gram mice from the high populations was
less than that of the segregated mice. The greater thymus weight coincides
with the greater width of the adrenal X-zone in the experimental mice and
probably, as in the case of the X-zone, represents inhibition of androgenic
activity which is not overridden by the increased amounts of circulating
corticoids. It has been shown that the natural adrenal secretory products of
mice will produce thymic involution and lymphopenia (Bahn ct al, 1957;
Wilson ct al, 1958) . However, depression of the thymus weight to below
the control levels in the larger mice from high populations can only reflect
increased adrenocortical activity. The mean thymus weights of mice from
the intermediate populations were greater than those of the controls or
experimental mice from high populations in the 16-19-gm. body weight
range. Data from the X-zone indicate that the secretion of androgen (or
at least its activity) was inhibited to the same degree in the intermediate
and high populations, whereas adrenal weights and width of the zona fas-
ciculata were less in the intermediate than in the high populations. There-
fore the greater thymus weights in mice weighing 16-19 grams from the
intermediate populations may have resulted from a less marked increase in
adrenocortical activity than occurred in the high-density populations,
whereas androgen secretion was depressed equally in populations of both
sizes. These results might be interpreted to mean that at increased popula-
tion densities the younger, and presumably subordinate, animals are the
ones predominantly affected by increased density. Since these results are
obtained by sacrificing an entire population at one time, such a conclusion
would be valid if it were not for the fact that the evidence indicates that the
increase in pituitary-adrenocortical activity involved all weight groups and
therefore all ages. Nevertheless, as we shall see below, a few of the heaviest

284 /. /. Christian

animals, probably including the dominant animals, were less affected by
increased density than those in any other weight group.

It was mentioned earlier than in response to stimuli which evoke in-
creased adrenocortical activity there is often inhibition of growth resulting
either from direct inhibition of growth by carbohydrate-active corticoids,
by inhibition of the secretion of pituitary growth hormone, or both. Evi-
dence has been presented indicating that there is inhibition of growth with
increased population density. Therefore, one might question the comparison
of organ weight data from mice from populations of high density with those
from segregated mice, as mice from the high-density populations may be
older than segregated controls for the same body weight. However, if it
were possible to correct for age, the differences with respect to body weight
between high-density populations and segregated controls would be even
greater. Therefore, comparisons with respect to body weight will err on the
conservative side.

The zonae fasciculatae of female mice of all sizes from high populations
were appreciably wider than their controls as a result of cellular hyper-
plasia and hypertrophy, although the presumptive increase in adrenocorti-
cal secretion was not indicated by the thymus weights. If anything, the
thymuses of females from the experimental populations were heavier than
those of the segregated controls, possibly a reflection of a diminished secre-
tion of the sex steroids.

On the other hand the preputial glands respond to stimulation by ACTH
(Jacot and Selye, 1951; Hess et al, 1952, 1953; Rennels et al., 1953), and
the preputials of female mice from high-density populations were heavier
than those from their segregated controls. Therefore, it is possible that the
increase in the weights of the preputial glands of these mice may have re-
sulted from an increased secretion of ACTH. This problem will be discussed
in more detail subsequently.

The effect of increased population density on adrenocortical function of
voles {Microtus pennsylv aniens) in freely growing confined populations has
also been studied. The experimental procedure used for these populations
was essentially the same as for the house mice except that eosinophil
coimts were used to measure adrenocortical function in the three popula-
tions of voles (Louch, 1956). The use of eosinophil counts has an obvious
advantage over adrenal weight for assessing adrenocortical function in
that the animals do not have to be sacrificed, especially when various factors
contributing to variation in the counts are taken into consideration ( Louch
et al., 1953) . In such long-term studies as this with repeated counts, there
can be little doubt that the eosinophil counts reflect adrenocortical function
rather than medullary. There was a significant negative correlation between

2. Endocrines and Populations 285

eosinophil counts and population density, indicating a progressive decline
decline in the number of circulating eosinophils with increasing population
density, but the correlation between eosinophil counts and population
density was not significant in the third population. The latter population,
however, never reached 30 animals in size. The published figures indicate
a striking parallel between the rate of population growth and eosinophil
counts for all three populations, and in all three there were significantly
fewer circulating eosinophil when the populations of these voles were greater
than 30 than when they were less than 30 (Louch, 1956). These results
indicate that adrenocortical function increased progressively with increasing
population density. The variability in Louch's data reflect to some extent
the difficulties in obtaining precise eosinophil counts, even though he was
fully aware of these problems and took every step possible to a\^oid the
usual pitfalls (Louch et al, 1953; Louch, 1956). The problems involved in
using eosinophil counts as indices of adrenocortical function have been dis-
cussed fully elsewhere (Thorn et al, 1953; Louch et al, 1953; Rosemberg
et al, 1954; Speirs, 1955; Visscher and Halberg, 1955; Louch, 1956; and
earlier in this chapter) . Handling will cause an adrenal medullary and corti-
cal discharge due to fear, excitement, and possibly rage in wild mammals
and therefore can effect eosinophil counts (Southwick, 1959) . Conseciuently
it is possible for variability to result from handling, as well as by individual
and perhaps more importantly, by unknown factors, unless appropriate
precautions are taken (Louch, 1958). However, in spite of these problems,
eosinophil counts offer a promising means of investigating adrenocortical-
population density relationships until simple, more direct tests become
available. Probably the best procedure at the present time is to follow the
population with eosinophil counts and to substantiate the changes with
adrenal weights at the termination of the study. Louch's results provide
evidence of increased adrenocortical function in voles with increased popu-
lation density.

Li a similar experiment, a marked increase in adrenal weight, as an index
of cortical activity was found in both male and female voles (Microtus
pennsijlvanicus) from a freely growing confined population which had
reached its maximum size (Christian, 1959b) . The adrenals of mature male
voles were increased 39.6%, and those of mature females 36.6% over their
segregated controls. It was pointed out earlier that adrenal weight-body
size relationships may vary with species and with sex; therefore, if animals
covering a wide range of sizes are to be used, the correct relationship in
these must be determined. It was found that in male voles over 115 mm
long, adrenal weight did not vary with further increases in the size of the
animal, therefore the absolute adrenal weights could be used for compara-

286 /. /. Christian

tive purposes. However, in female voles adrenal weight increased with in-
crease in body size; therefore for females it was necessary to use relative
adrenal weights for purposes of comparison.

3. Reproduction

a. Male. Reproductive function was depressed in proportion to increases
in population size in both male and female house mice from populations of
fixed size. The results of experiments with freely growing confined popula-
tions of house mice and voles also show a progressive decrease in reproduc-
tive function with increasing density. We have seen that there was a delay
in the onset of puberty in male house mice from freely growing populations
of high density, as demonstrated by the delayed involution of the adrenal
X-zone and the development of the testes, seminal vesicles, and preputial
glands. The size of the gonads and sex accessories was less than their con-
trols from segregated mice of all but the heaviest body weight. Rapid de-
velopment of the seminal vesicles and preputial glands did not begin until a
body weight of 16 gm. was reached, which coincides exactly with the be-
ginning of X-zone involution in animals. These results strengthen the con-
clusion that androgen secretion was delayed. In the segregated controls, the
sex accessories had begun rapid growth by the time the animals had reached
a weight of 13 gm. These results were confirmed by histologic examination
of the testes. The size of the testes primarily reflected the advancement of
spermatogenesis and the development of the seminiferous tubules. Sperma-
togenesis was inhibited to the same degree as the sex accessories (Christian,
1956) .

Male mice from freely growing populations of intermediate size exhibited
an inhibition of the reproductive organs which was approximately inter-
mediate between that exhibited by the males from the high populations and
the segregated controls (Christian, 1956).

It is of additional interest that testes and sex accessories of male mice
weighing 28 gm. or more, the heaviest weight group in the experimental
populations, were about the same size as those of the isolated controls.
Furthermore, there was no inhibition of spermatogenesis in these animals.
In general, this weight group contained the dominant animals in the four
high-density populations examined and was represented by a total of 21
animals. These mice represented the initially introduced males and those
males from the first litter or two that had established their dominance early
in the histories of the populations. These observations are confirmed by
Crowcroft and Rowe (1957), who observed that the productive females in
in confined populations of house mice were the introduced animals or those
from their first few litters. Detailed observations relating the weights of

2. Endocrines and Populations 287

reproductive organs to specific animals were not made, but it is likely that
these results parallel those from populations of limited size in which it was
found that the weights of the reproductive organs were greatest in the
socially dominant animals. We have indicated that the weights of the
accessory reproductive organs appear to be the most sensitive indications
of changes in population density as well as social pressures and differences
in rank. This also appears to be the case in these experiments with freely
growing populations. Reproductive function of male mice has not been
examined in detail, usually not at all, but other investigators studying
freely growing populations, although Strecker and Emlen (1953) did find
all the males in a self-limited population with epididymal sperm 6 months
after population growth had ceased.

It might reasonably be asked why, during the 6 months or so that most
of these populations lasted, there was not adaptation to the situation with
a diminution of the effects of population density on the adaptive responses
of the adrenals and reproductive organs. Brown (1953) and Southwick
(1955b) have both pointed out that in a well-stabilized social hierarchy of
mice there is a relatively low level of fighting, usually used as an indication
of the amount of sociopsychologic pressure and interaction, but that when
shifts or disruptions in the hierarchy occur, due to the maturation of new
individuals or death of old ones, social pressures increase, as indicated by an
increase in the amount of fighting. Southwick (1955b) has pointed out that
these factors are constantly disrupting the social order in rapidly growing
populations, therefore, as the population increases it is inevitable that social
pressures increase apace. Comparable results have been observed in other
freely growing populations, but there is not necessarily fighting (Christian,
1956, 1959a, b). Furthermore, female mice become particularly aggressive
prior to parturition (Brown, 1953; Crowcroft, 1954), and so the total
amount of aggressiveness contributed by females would tend to increase
with the population. For these reasons it appears that the amount of social
pressure in a population will increase with the size of population. It has
been observed also that the self-limited maxunum size varies greatly from
population to population (Southwick, 1955a, b; Christian, 1956) appar-
ently as a function of the amount of social pressure within the population,
the variation being contributed by individual differences, the stability of
the social structure, and similar factors. For example, Southwick (1955b)
has described the individual differences between males in the amount of
territory they will fight over. Southwick ( 1955b) has discussed in consider-
able detail some of the factors involved in the composition of social compe-
tition between mice. One population has been observed in which the social
order of the population was disrupted at about half the estimated maximum
size by the death of an old, tyrannical male (Christian, 1956, and unpub-

288 /• /. Christian

lished) . Severe fighting broke out in this population and the growth of the
population ceased completely until social order was reestablished 6 weeks
later. The cessation of growth resulted from a complete cessation of all
aspects of reproductive activity. These considerations, plus the fact that
the physiologic responses are equivalent in populations of maximum self-
limited size irrespective of the number of mice (Christian, 1956), make it
likely that the growth of populations of house mice, and possibly voles, is
limited by the total amount of social pressure rather than by the number
of animals.

The reproductive competence of males has not been examined in detail
for any species from freely growing populations other than house mice ; and
no studies have adequately explored the problem. In most studies on popu-
lation density the criteria of male fertility are position of the testes, gross
size of the testes, visibility of the epididymal tubules, and occasionally the
presence or absence of sperm in the epididymes (Strecker and Emlen, 1953;
Brown, 1953 ; Southwick, 1955a; Crowcroft and Rowe, 1957) . These criteria
are actually poor indicators of relative fertility and only suggest whether
or not an animal is mature. Detailed morphologic studies provide more
evidence of fertility (Christian, 1956), but they do not provdde conclusive
evidence. No information on subtle changes in fertility is provided by any
of these criteria. The ability of males to fertilize females with respect to
population density has not, to my knowledge, been investigated.

Many influences which depress reproductive activity cause striking
degenerative changes in the germinal cells in the testes (Selye, 1947). The
formation of giant cells from cells of the spermatogenic series is a common
indicator of such changes (Selye, 1950; Steinberger and Dixon, 1959).
Furthermore, influences which can produce severe degenerative changes
or inhibit spermatogenesis completely can produce more subtle changes if
the damaging stimulus is less severe (Steinberger and Dixon, 1959). It
seems likely that the conspicuous changes that occur in the testes of house
mice as a result of increasing population density [a delay in the onset of
spermatogenesis, reduction in the number of mature spermatoza, and
formation of giant cells (Christian, 1951, 1956)] can also be extended to
include less obvious abnormalities of the sperm, such as decreased motility
and viability. There also may be a decrease in the nmnber of sperm pro-
duced, and therefore in the number of sperm in the ejaculate. All these
factors presumably affect fertility (Chang and Pincus, 1951). To these
must be added the likelihood of an altered medium for the sperm which
would be suggested by the changes in the sex accessories associated with
testicular changes, as indicated by the decline in their weights (Burrows,
1949; Lutwak-Mann et al., 1949; Leathem, 1950; Mann, 1954) . Atrophy of
sex accessories carries the implication that their secretory products may be

2. Endocrines and Populations 280

abnormal, deficient, or decreased in amount, and that the composition of
the ejaculate will in all probability not provide an optimal medium for
sperm (Burrows, 1949; Lutwak-Mann et al., 1949; Cavazos and Melampy,
1954). Therefore it is possible that male fertility is depressed appreciably
by increased population density, although the gross examination or rela-
tively minor changes in weight would provide no indication of such a reduc-
tion in male fertility. Male fertility is assessed in most studies on small
mammals by (1) the position of the testes, (2) whether or not the epididy-
mal tubules are clearly visible, or (3) possibly by epididymal smears to
determine the presence or absence of spermatozoa (Southwick, 195oa;
Louch, 1956; Strecker and Emlen, 1953; Crowcroft and Rowe, 1957).
Degenerative changes can be detected in testicular smears (Christian,
1950a, 1951) , but with much less assurance than properly prepared, stained,
and critically examined sections of the testes. Morphologic studies of the
testes, if properly done, can be extremely revealing, but the weights of the
testes and accessory organs, while providing a useful index of reproductive
function in general, do not tell whether or not the male is actually fertile
unless inhibition is severe. In the absence of good evidence to the contrary
it is generally assumed that the males in a population are fertile — an as-
sumption that may be misleading. Studies of male fertility in relation to
population density are needed that make use of sperm counts, determina-
tion of the motility and viability of the sperm, and possibly direct determi-
nation of fertility by mating with proven females. Admittedly such a pro-
gram would present problems, but they are by no means insurmountable.
Furthermore, once the techniques are worked out for obtaining ejaculates
from the males in populations of small mammals, it should be a valuable
and useful procedure for routinely assessing male fertility and reproductive
competence in freely growing populations of small mammals. But it is in-
correct to assume that male fertility is not affected by increased population
density, especially in view of the evidence indicating that the secretion of
testicular androgens is extremely sensitive to changes in population density
or to social rank.

Southwick (1955a) observed a decrease in fertility in female mice from
populations which were slowly declining from peak levels. He also noted, as
have others (Southern, 1948; Calhoun, 1949), that in populations of high
density or with poor social organization (Calhoun, 1949, 1950) that "copu-
lation pressure" on females in estrus was high; that is, a number of males
gathered around her and attempted to copulate although they were fre-
quently pushed off by others, so that the mean copulation time per male
was reduced. "Copulation pressure" was used as a possible explanation
for the decline in fertility. This explanation overlooks several facts. The
populations were progressively declining from peak levels and did so for six

290 J. J. Christian

months, and so did pregnancy rates and fecundity rates (correspondence in
the dechnes in these latter two was used as an argument against the possi-
bility of increased intra-uterine mortality) . It is therefore unlikely that a
steadily declining fertility can be explained by copulation pressure in a
constantly decreasing number of animals. Male fertility was assessed by
uncritical procedures and female fertility by performance. Assessing fer-
tility by female performance of course does not provide a way to assess the
possible role of depressed male fertility. It seems much more likely that
both male and female fertility were depressed. Crowcroft and Rowe (1957)
noted a decline in fertility in female house mice with increasing density as
well as during the period of asymptotic stability in populations of house
mice. These authors indicate the failure of population growth resulted en-
tirely from inhibition of female reproductive function. Louch (1956) was
unable to correlate diminished fertility with "copulatory pressure," al-
though it was observed in his population of voles. We have seen that the
females in extremely dense populations became pregnant but failed to bear
litters and that recovery from the effects of chronically depressed fertility
was extremely slow (Christian and LeMunyan, 1958). Increasing age may
have been a factor in these freely growing populations, especially if fertility
was already partially curtailed.

This discussion is meant neither to be a criticism of the observations
made by these investigators nor to provide the correct explanation for
them. Rather, these examples have been used to indicate pitfalls in experi-
mental procedures presently in general use to assess male fertility and the
effects of various factors on fertility in populations of small mammals, and
to call attention to reasonable and perhaps more probable explanations
other than those usually provided. The literature contains many examples
which could have been used. Finally, it should be noted that the popula-
tions in Southwick's studies had reached peak densities and had been de-
clining slowly for 6 months before the animals were examined, so it would
be difficult to determine the effects of increasing density on fertility in mice
of either sex. The results of those studies made during the period of increase
of a population or at the time it reached maximum density are not com-
parable to studies made so much later, and it is not correct to extrapolate
from one to the other, as was done in several instances. The results of other
experiments in which male fecundity was assessed during the period of
population increase indicate that male fertility, in terms of the whole
population, has a striking negative correlation with density (Christian,
1956, 1959a, b; Crowcroft and Rowe, 1957) . The secretion of gonadotropins
is apparently depressed with increasing pituitary-adrenocortical activity
in the greater proportion of subordinate animals associated with increases
in density. Evidence has been presented which shows that increased adreno-

2. Endocrines and Populations 291

cortical activity in mice results in an increase in the secretion of weakly
androgenic steroids, and that these can depress the secretion of gonado-
tropms.

h. Female. Reproductive competence of female mammals can be assessed
grossly more easily than that of males diu-ing studies of growing popula-
tions simply because the number of young produced can be counted. Some
comments have been made about the relationships of female fertility and
reproduction to population density in animals from freely growing popula-
tions. Most of the pertinent evidence indicates that fertility and reproduc-
tion in female mammals is depressed by increased density. Two kinds of
information has been gathered on female productivity in relation to popula-
tion density. In one case the number of young born at intervals throughout
the history of the population has been determined, and in the other female
fertility is assessed by external examination of the vagina, which is closed
by a thin membrane during anestrus in many species, or by examination of
the animals at autopsy. Information of both kinds is desirable, although
neither is sufficient by itself.

It has been shown in studies with house mice and voles in freely growing
confined populations from which adequate data are available that the birth
rate for mature or potentially mature females in the population declines
steadily with increasing population size (Clarke, 1955; Christian, 1956,
1959b, 1961; Crowcroft, and Rowe, 1957). The decline in the number of
young born per female appears to decline approximately linearly as the logio
of the population increases (Christian, 1959b),

The declines in productivity apparently resulted from suppression of all
phases of reproductive functions. Crowcroft and Rowe (1957) showed
there was complete suppression of estrus, as indicated by closure of the
vaginal orifice, in an increasingly larger proportion of females as the size of
a population increased. It was shown that closure of the vaginal orifice was
a reliable index of fertility in female house mice in these experiments by
the very high correlation between the condition of the vaginal orifice and
the development of the reproductive organs. In these populations the
majority of the young were produced by the introduced females initially or
by females from the first few litters. These results correspond to those from
populations of limited size in which it was indicated that the reproductive
competence of female mice was related to their social status. Similar results
were also observed in confined populations of Norway rats (Calhoun, 1949,
1950) . Reproductive function was completely suppressed in the majority
of female born into these populations.

These results are comparable to those from experiments with freely
growing populations of house mice (Christian, 1956), thus indicating that
reproductive function in female mice was depressed for all body-weight

292 J. J. Christian

groups. In these experiments the weights of the preputial glands of mice
from populations of high density were significantly greater than those from
segregated controls. For their stimulation to full development and function,
the preputial glands depend primarily on the more potent androgens (Bur-
rows, 1949), but the weakly androgenic steroids (Huggins et al., 1955),
ACTH (Jacot and Selye, 1951; Asling et al., 1951), and pituitary growth
hormone (Huggins et al., 1955) all have a mild stimulatory action on the
preputials. The evidence for the effects of estrogens on the preputials is
conflicting. On the one hand, estrogens appear to have a mild stimulatory
action (Burrows, 1949; Beyler and Szego, 1954) whereas other experiments
indicate that they inhibit the preputials (Rennels et al., 1953). Since other
evidence from the experiments with mouse populations indicates that the
production of sex steroids by the gonads and growth hormone by the pitui-
tary is inhibited at high densities, it seems probable that the increased size
of the preputials in these intact females must have been due to increased
ACTH and possibly increased adrenal C19 steroids. In any event the data
on the preputials are strongly suggestive of increased pituitary-adrenocorti-
cal function. Attention should be directed to the fact that the preputials
were appreciably larger than those from segregated mice with full reproduc-
tive competence. Therefore there is no reason whatsoever to implicate estro-
gens in these effects ; so ACTH and adrenal androgens must be responsible
for stimulating preputial development to well above the control levels.
Presumably the levels of estrogens declined in these females as a result of
a decreased secretion of pituitary gonadotropins in association with in-
creased secretion of ACTH. These conclusions were supported by the fact
that there was also (1) a marked increase in the proportion of females of
adult size with infantile uteri and ovaries, and (2) a 20% decline m
number of females with mature reproductive organs that were also preg-
nant, (3) a 13% decrease in the mean number of viable embryos per
pregnancy, and (4) a 58% increase in the number of resorbing embryos per
pregnancy. Thus there was diminished fertility and increased losses between
o\'ulation and implantation, and implantation and birth. Decreased fertility
was shown in some animals by a total inhibition of reproduction and in
others by a markedly diminished number of viable embryos per pregnancy.
These results agree with those of Crowcroft and Rowe (1957). The same
phenomena apparently occurred in populations of voles (Clarke, 1955;
Louch, 1956; Christian, 1959b) . Birth rates declined steadily as the popula-
tions increased for house mice and voles (Christian, 1959b, 1961). There-
fore it seems that reproductive function is inhibited in female mice and voles
in proportion to population density, either in terms of the proportion of
reproductively competent adult females in a population or of the degree of
reprodu(•ti^'e function in individuals. Stated another way, reproductive

2. Endocrines and Populations 293

function in female mice and voles is reciprocally related to population
density. The degree of inhibition apparently may vary from none to com-
plete in the females of a given population, within limits depending on social
relationships of the particular individuals.

We have already criticized the concept of "copulation pressure" which
has been used to explain the decline in female fertility seen in populations
of high density. As an explanation it is inadequate because it does not
coincide with the facts, even in those cases in which it has been used as an
explanation (Southwick, 1955b). Furthermore, no correlation could be
shown between "copulation pressure" and fertility in other studies (Louch,
1956) . In the light of those studies in which female reproductive function
was more precisely assessed, it appears that reproductive function was at
least partially inhibited in those populations in which "copulation pressure"
was used to explain the decline in fertility, and that the means used to assay
female fertility were not sufficiently sensitive. "Copulation pressure" con-
ceivably could have an effect on female reproductive performance, but
until the problem is reexamined more critically it must remain an unaccept-
able concept, especially in view of the preponderance of evidence indicating
that reproductive function is inhibited, either individually or on a popula-
tion-wide basis, as part of the adaptive responses to increased density. A
broad view of this evidence suggests considerable variability from popula-
tion to population in the degree of reproductive inhibition with respect to
the numbers of females involved as well as to the magnitude of inhibition
in particular individuals. These variations are evidently related to social
factors, and this aspect of the problem needs examination in greater detail
than heretofore.

Prenatal mortality of the fetuses was an appreciable factor in reducing
the number of births in dense populations, presumably largely owing to a
suppression of gonadotropins, but also possibly to the effects of increased
adrenal corticoids and androgens. There seems to be some kind of "dual
response" to density in depressing female productivity. Judging from the
results described in this and preceding sections, it appears that the least
"dose-response" effect occurs when fertility is partially diminished, as by
decreased numbers of ova and increased losses between ovulation and birth,
without a total suppression on the ability to bear young. A greater effect
apparently is total inhibition of estrus and reproduction in potentially
mature animals. Delayed attainment of maturity probably reflects partial
inhibition. The possible mechanisms involved in producing intra-uterine
mortality have been discussed in an earlier section.

c. Litter survival. A marked decline in litter survival with increasing
population density was observed in all but a few experiments with freely
growing populations of house mice or voles (Rtrecker and Emlen, 1953;

294 /• /• Christian

Brown, 1953; Clarke, 1955; Southwick, 1955a, b; Christian, 1956; Louch,
1956). Litter survival apparently declines approximately linearly as the
logarithms of the population increases (Christian, 1959b). However, litter
mortality is not always a factor in diminishing the rate of population
growth. Crowcroft and Rowe (1957) found very little mortality of young
mice in their populations and furthermore were able to attribute all the
observed mortality to interference on their own part in making routine
censuses. The decline in productivity in their populations was due almost
entirely to prenatal loss or complete inhibition of all reproductive activity.
It has been suggested (cf. above) that diminished fertility probably was
more of a factor than was realized in several other investigations; neverthe-
less litter mortality has been an unportant factor in many populations.
Decreased litter survival may be due to diminished lactation (Christian,
1956; Christian and LeMunyan, 1958) or to behavioral factors (Brown,
1953; Strecker and Emlen, 1953; Southwick, 1955b).

Partial inhibition of lactation with increased density was described
earlier. The results suggested that lactation, controlled by complex endo-
crine mechanisms, was partially inhibited owing to inhibition of gonado-
tropin secretion and a resultant diminished production of estrogens and
progestins. Lactation therefore reflected population density in a similar
fashion to other reproductive functions. It was also pointed out that the
progeny nursed by mothers subjected to increased density were profoundly
affected by the inhibition of lactation. Experiments with freely growing
populations of house mice suggest that deficient lactation may be an im-
portant factor in decreasing litter survival (Christian, 1956). These young
were stunted and were weaned early. At autopsy their stomachs contained
little or no milk. Calhoun (1949, 1950) suggested that the socially sub-
ordinate rats in freely growing confined populations were incapable of
raising then* young, even though they were protected from disturbance.
He indicated that physiologic and psychologic disturbance in the sub-
ordinate females affected the young through either poor fetal nutrition or a
breakdown of maternal instincts.

Dimmished lactation may have been an appreciable factor in reducing
the survival of young in the above experiments, and diminished lactation
is a logical consequence of a generalized inhibition of reproductive function
presumably due to diminished gonadotropin secretion.

Decreased litter survival was attributed entirely to social or behavioral
factors in a number of experiments with freely growing populations of
house mice or voles. Brown (1953), Southwick (1955b), and Louch (1956)
related litter survival to the type of nest constructed and maintained.
Brown and Louch each noted that litter survival was good so long as the
female mice or voles were able to maintain covered or bowl types of nests,

2. Endocrines and Populations 295

but that this was frequently impossible because of destruction of the nests
and interference with the attempts of the females to maintain or reconstruct
them. Only 49 of the 72 attempts to build and maintain either of these two
types of nest were successful in these populations (Brown, 1953). South-
wick (1955b) noted a similar relationship between the type of nest and
litter survival, but he also found that litter survival was good in bowl or
covered nests regardless of the number or sex of the adults which were also
present. Litter survival was zero with no nests. These and Brown's (1953)
results imply that not all the females were prevented from maintaining
proper nests and nourishing then* litters irrespective of the conditions and
interference. These results are revealing, as they suggest strongly that more
is involved in litter mortality in these experiments than simple interference
or the type of nest maintained. Similar results were noted in other experi-
ments (Christian, 1956) except that in this case even those with no nests
sometimes survived. In these latter experiments it was also noted that if
litter mortality occurred when the populations were less than one-half of
maximum density, it occurred within the first four postnatal days and
might be attributed to interference which resulted in too much movement
of the litter by the mothers. Of additional importance is the fact that litter
survival declined regularly with increasing density and that this began to
occur well before interference could be significant. If the young survived
the first 4 days at these lower densities, they almost always were success-
fully weaned. However, at higher densities the increased litter mortality
usually occurred after the young were 10 days old. Until this age survival
equaled that seen at lower densities. These facts suggest that the defect was
primarily nutritional, even though there still were some deaths of young
less than 4 days old attributable to interference. The age of death of the
young is not known for most of the reported experiments, so that this
distinction was not possible. However, the fact that crowding and inter-
ference was not a factor in litters with proper nests suggests that perhaps
the behavioral changes which resulted in poor nest maintenance may be
based on an underlying alteration in maternal behavior, possibly due to
endocrine changes. Admittedly this is speculative, but the available evi-
dence suggests some such explanation. The problem certainly must be
investigated m greater detail. It is possible that the capacity to lactate
properly and the abilitj^ to maintain proper nests are closely related by a
common causal basis. However, until evidence for or against such a hy-
pothesis is available, it must be assumed that postnatal litter mortality at
high population densities may result from (1) inhibition of lactation, (2)
interference by other animals, or (3) changes in maternal behavior, which
apparently decreases with increasing population density.

We are inclined to the view that interference of various sorts probably

296 /. ./. Christian

plays a more or less constant, probably unimportant, role in causing litter
mortality at all levels of population density, whereas inhibition of lactation
and changes in maternal behavior are density dependent and probably
reflect physiologic (endocrine) responses to changes in population density.
Southwick (19o5b) noted that aggressiveness, as indicated by fights per
mouse per unit of time, was reciprocally related to litter survival at high
densities, but that this relationship did not necessarily hold at lower
densities. On the other hand, there was a correlation between fighting and
density in only one of three populations of voles (Louch, 1956). The be-
havioral changes observed in relation to litter survival may be outward
manifestations or "symptoms" of more profound physiologic changes at
higher population densities.

A final consideration of importance is that fecundity and litter survival
are sufficiently sensitive to changes in the population to be responsive to a
variety of different circumstances affecting density. Litter survival was
markedly increased in a population of house mice in which there was an
appreciable mortality of adults (Christian, 19o9b). The growth form and
eventual size of the population were comparable to those of populations
without appreciable adult mortality. The loss of adults in this population
was compensated for by increased litter survival. Similar relationships were
observed in three populations of voles (Louch, 1956). There was a signifi-
cant correlation between fertility and population density in one population
of voles in which there was no correlation between density and adult
mortality. In a second population there was a good correlation between
density and adult mortality, but none between density and fertility. The
third population was intermediate between these two. Therefore, there
appears to be sufficient flexibility in the physiologic adaptive responses to
compensate for losses of adults from populations of voles or mice. This
compensation apparently can occur at any stage of the reproductive pro-
cess from ovulation to weaning.

d. Effects of food supply. It is often assumed without question that food
shortage is responsible for limiting population growth. However, we have
seen that the growth of a population may be entirely self-limited without a
shortage of food. The role of a limited food supply in the regulation of
population growths has been investigated (Strecker and Emlen, 1953;
Strecker, 1954). Clarke (1955), Christian, (1955b, 1956, 1959b), South-
wick (1955a, b), and Louch (1956) have emphasized the fact that in
their experiments food was always abundant and usually scattered so that
any animal could obtain food irrespective of other animals present. Further-
more, experiments with populations of fixed size indicated that inanition
was not a stimulus to increased adrenocortical activity in house mice, nor
did it constitute a factor for increasing competition when it (food) was

2. Endocrines and Populations 297

limited either in source or amount (Christian, 1959c). These results are
consistent with those from experiments with freely growing confined
populations of house mice or voles in which competition for food in every
case was rare or absent (Strecker and Emlen, 1953; Louch, 1956; South-
wick, 1955a, b; Christian, 1956). Strecker and Emlen (1953) provided two
confined populations with a limited amount of food but the growth of only
one of these was limited by the limited food supply. Growth of this popula-
tion was rapid and ceased abruptly when the limit set by the food supply
was reached. Growth ceased because reproduction stopped abruptly
with involution of the reproductive organs, especially of the females, all of
which were reproductively quiescent at autopsy, although the animals
otherwise appeared to be in excellent condition and usually excessively fat.
Several points of interest emerge from these experiments. One is that
population growth ceased very abruptly with a truncated growth curve,
indicating that a limited food supply does not act as a damping factor, but
exerts its effects at one point on the growth curve. Therefore, it is unlikely
that a typically sigmoid growth curve would result from a food shortage.
The second point of interest is that reproduction was so sharply limited by
the limited food supply. These results agree with those from populations of
fixed size and indicate that the food limitations completely inhibited re-
production, probably by suppressing the secretion of gonadotropins, with-
out producing increased activity of the pituitary-adrenocortical adaptive
system. The last point is that the population immediately adjusted to the
supply and the general condition of the mice was unaltered by the limitation
in food. The inhibition of reproductive activity in all the females contrasts
sharply with the results of other experiments conducted on self-limited
freely growing populations of mice or voles supplied with an excess of food.
In the latter populations there were always some reproductively active
females, although the actual proportion in the population varied with each
population. Data on adrenal activity were not collected in these experi-
ments on the efTects of food limitation on freely growing populations. It is
probable that there would have been density-dependent increases in adreno-
cortical activity, but the effects of food limitation were independent of any
density-dependent effects. In another series of experiments, the food supply
was limited similarly, but egress from the population was allowed (Strecker,
1954) . In these populations there was a low constant rate of egress until
the food limitation was reached and then a high rate of egress began and
continued until the experiment was terminated three months later. The
egress apparently involved all segments of the population except the
youngest. Reproduction continued at a good rate in this population in
sharp contrast to the confined population without egress. The implications
of these experiments are clear: f1) food limitation can limit population

298 J- J- Christian

growth by inhibiting reproduction, but the growth curve becomes sharply
truncated; (2) if egress from a population is permitted, food limitation
has no effect on reproduction; (3) feeding must be random and food cannot
have been a factor increasing competition, otherwise all the mice would not
have been affected equally. Extrapolating from these results to natural
populations, it is evident that a limited supply of food will not affect repro-
duction or otherwise affect the population so long as emigration is possible,
but if emigration becomes impossible, owing to a complete saturation of
the available habitat, reproduction will cease without any other apparent
effect. Whether or not there is competition for food in species other than
house mice must be determined, as increased competition would increase
pituitary-adrenocortical activity with all its sequelae. There may be in-
creased competition for food in voles (Frank, 1953), although it seems un-
likely that in general there will be an increase in competition over that
already present due to social factors (cf. above) . However, the usual result
of emigration from a population is increased mortality: the animals which
leave enter strange territories and very rapidly become mortality statistics
(Errington, 1943, 1954b; Calhoun, 1948; Davis, 1953). Physiologic adap-
tive mechanisms apparently are always operative in relation to population
density and therefore would reflect the relative density whether or not the
growth of a population was limited by food, and a distinction between food-
limited and self-limited populations should be possible by comparing the
reproductive activity of the populations and the general conditions of the
individual animals. Adrenal weights and other parameters of increased
pituitary-adrenocortical activity should also differentiate between the two
unless a population was food-limited at a point close to its maximum self-
limited size.

These statements and conclusions are based on the results of experiments
with house mice and seem fairly conclusive for this species, possibly even
for voles, but it is difficult to believe that food is not a competitive factor
for all species and for all populations. The results of experiments to date
with populations of fixed size or freely growing indicate that food is not
even a competitive factor within the existing social structure, as all anmials
are apparently equally affected irrespective of their social rank (Strecker
and Emlen, 1953; Christian and LeMunyan, 1958; Christian, 1959b). It
should be remembered that if ever and when ever food becomes an object
of competition, shortages then necessarily will act through the generalized
physiologic adaptive responses and limit population growth in a density-
dependent fashion just as does purely social competition. The net conclu-
sion from these results is that a shortage of food per se probably rarely
limits population growth as the peculiar combination of effects which
result from a food shortage is seldom seen. However, this statement is not

2. Endocrines and Populations 299

to be interpreted to indicate that acute starvation cannot ensue in local
populations or that a limited food supply is not at times important. The
relationships between food, competition, reproduction, and the growth of
populations require much more investigation, especially for different
species. Until additional critical information is available, any conclusions
regarding food and population growth must remain tentative.

4. Growth

In a recent experiment with a freely growing population of house mice,
it was shown that there was appreciably inhibited growth in all but the
mice originally introduced and those from the first litters (Christian,
1961) . In this experiment all mice were toe-clipped so that their ages were
known. In general the results from the reproductive organs and adrenal
glands confirmed those of earlier experiments and were even more pro-
nounced because the results could be equated with age rather than body-
weight categories.

5. Summary of Results from Freely Growing Experimental Popu-
lations

We have seen that, for the most part, freely growing populations of
house mice and voles are self-limited, primarily as a result of a density-
dependent activity of physiologic adaptive mechanisms. The activity of
these mechanisms is reflected in both sexes by a progressive inhibition of
reproduction and stimulation of the pituitary-adrenocortical activity with
increasing population size. The results agree with those obtained from
populations of fixed size. Diminished reproductive function in the female
may be apparent at any one or all phases of reproductive activity : inhibi-
tion of estrus, increased intra-uterine mortality, reduced nvmibers of im-
planted ova, increased postparturient mortality due to suppression of
lactation, and possibly by diminished maternal behavior. Inhibition of male
reproductive ability is shown by decreases in the weights of the sex acces-
sories and depressed spermatogenesis, although the problem of male fertility
needs further investigation. The decreases in reproductive function in both
sexes apparently stem from decreases in the secretion of pituitary gonado-
tropins with a secondary decline in the production of estrogens, progestins,
and androgens. Evidence for and against the role of "increased copulation
pressure" in diminishing the fertility of female mice and voles was dis-
cussed. Without much more conclusive evidence, it is likely that "copula-
tion pressure" is an unimportant factor and that fertility is depressed by
increased density operating through sociopsychologic and physiologic path-
ways. Increased adrenocortical activity was shown by increased adrenal

300 J' J- Christian

weights, eosinophil counts, and histologic studies. Thymus weights re-
flected these changes. Increasing infant mortality with increasing density
may be due to diminished lactation, but may also result from social and
behavioral factors. Food limitation may restrict population growth by
completely inhibiting reproduction, but is without any other apparent
effects.

Finally, it is evident that the growth of confined populations of mice and
voles is regulated and limited by density-dependent physiologic responses
which are activated by sociopsychologic pressures. Whether or not these
relationships hold for all species remains to be determined.

C. Natural Populations

Evidence gathered in the laboratory relating physiologic adaptive mecha-
nisms to changes in population density cannot prove that these responses
occur in natural populations. Even for freely growing confined populations
in the laboratory, the environmental conditions are quite altered by the
removal of the effects of climatic and seasonal variability from those found
in natural situations. Preventing egress is another artificially imposed
condition; although it seems more likely that confinement parallels the
situation when populations reach densities where egress into neighboring
areas becomes possible. Clearly, evidence from natural populations is neces-
sary before any conclusions are justified regarding the role of physiologic
adaptive mechanisms in the regulation of mammalian populations under
natural conditions. Conclusive results are much more difficult to obtain from
natural populations than from confined populations, as might be expected,
due to the complexity and variability of these populations and techincal
difficulties in collecting the required data. Adequate controls and proper
assessment of environmental factors are often severe obstacles, and ob-
taining samples of adequate size from populations at very low densities is
intrinsically very difficult. Nevertheless, evidence has been obtained to
indicate that the density-dependent physiologic responses in natural popu-
lations are similar to those seen in experimental populations. Finally, social
strife has been implicated (Kalela, 1957) as an important factor in producing
the effects associated with increased density ; territorial mutual intolerance
increases greatly with sexual maturity in Clethrionomys rufocanus, which
serves to increase tension even with decreased density (Kalela, 1957).
Therefore behavioral factors again appear to be unportant.

1. Adrenocortical Activity

The available evidence relating adrenocortical activity to changes in
population density has been obtained largely from studies on populations

2. Endocrines and Populations 301

of Norway rats, both urban and rural, and voles {Microtus sp.), although
suggestive results have been obtained for a few other species.

Adrenal weight in urban Norway rats of both sexes was shown to be
related to population "density" in a study of 21 populations in Baltimore,
Md. (Christian and Davis, 1956). Each population was confined to a city
block, the latter acting as an island (Davis, 1953). The capacities of the
blocks for rats varied considerably, as did the sizes of the populations them-
selves, which resembled confined freely growing populations of house mice
in this respect. Therefore, the populations were categorized by their posi-
tion on their own growth curves at each sampling. A hypothetic growth
curve was divided into low stationary, low increasing, high increasing, high
stationary, and decreasing, the progression from low increasing through
decreasing being a progressive increase in relative density, although the
"density" status of low stationary is equivocal (Christian and Davis, 1956) .
Each time a sample was collected, the population from which it came was
put in one of these categories. "Density" is obviously not strictly in terms
of area, but in relation to the carrying capacity of the block and the social
characteristics of the population. As populations increased in relative
density from low increasing to decreasing, the mean adrenal weight was
increased with each relative density; so that rats of both sexes from de-
creasing populations had 19% heavier adrenals than those from low, in-
creasing populations. Thymus weight was reciprocally related to adrenal
weight in female rats, but there was no apparent relationship between
thymus weight and adrenal weight, with respect to population density, in
male rats. The weights of the pituitary glands of male rats were positively
related to the adrenal weights. That is, changes in adrenal weight were re-
flected by changes in pituitary weight in the male rats (Christian and Davis,
1956) . However, the functional significance of these changes in the weight
of the pituitary glands is not known. The weights of the pituitaries of fe-
males and thyroid glands of both sexes bore no apparent relationship to
population density. The data from the low stationary populations are
difficult to evaluate. The adrenals of female rats were almost as light as
those from low increasing populations, and their thymus glands were the
heaviest found in any category. On the other hand, the adrenals of male
rats were heavier than in any other category. Low, stationary populations
are extremely difficult to evaluate, as the actual numbers of rats are so
small that even proportionately large changes are difficult to detect. How-
ever, these difficulties with their inherent errors in properly assigning
populations to a category do not explain the divergence in the male and
female adrenal weights, nor do they alter the fact that these populations
were at very low levels of density. It should be pointed out that the de-
creasing populations were at maximum "density" because they were de-
clining naturally and therefore were at or above the environmental capa-

302 /. J. Christian

city. It should also be emphasized that food was available in excess of needs
for all populations and from numerous sources ; so that food cannot be con-
sidered to be an imprtant factor with respect to these populations.

Some of the variability between urban populations of rats may be ex-
plained by the fact that the rat population in each block is divided into
several discrete colonics when the total population is at a relatively low
density (at high density these colonies coalesce and the population evidently
becomes a single unit throughout the block). The colonies comprising
several of the block-populations were evaluated individually with respect
to relative density in the same way as the populations for the entire blocks
(Christian and Davis, 1956). Relative density values were assigned and
compared with the adrenal weights of the rats from these colonies. A sug-
gestive correlation between adrenal weights and relative population den-
sities of these colonies was found, although the differences were not signifi-
cant, probably owing to the small number of samples. The sample size was
limited by the relatively few colonies from which a sufficient number of rats
were collected for comparative purposes.

Although adrenal weight is greatest in naturally declining populations
(Christian and Davis, 1956) , it was shown in another study that artificially
reducing populations of rats produces an immediate and proportional reduc-
tion in adrenal weight (Christian and Davis, 1955). Adrenal weight in both
sexes of rats from three populations was reduced 38% by intensive trapping.
The adrenal weights were maintained at this level for 5 months by main-
taining the populations at the reduced level.

In another study a rural population of Norway rats was followed by
monthly sampling for two years (Christian, 1959b). An index of the size
of the population was obtained each time a sample was collected by using
a standardized trapping procedure. The weights of the adrenals, pituitaries,
and thyroids were determined for each sample. There was a highly signifi-
cant correlation between adrenal weight and population index for both
sexes, and an even more significant correlation between pituitary and
adrenal weight for both sexes for the 24 monthly samples in the two-year
period. The functional significance of the pituitary changes is not known;
it can only be inferred from data on weight. Evidently changes in the rate
of ACTH secretion, as indicated by adrenal weight, are accompanied by
parallel changes in pituitary weight. Changes in thyroid weight and in
thymus weight were not correlated with changes in the population size,
adrenal weight, or pituitary weight.

The results of the foregoing experiments indicate that adrenocortical
activity in Norway rats from natural urban or rural populations is related
directly to changes in population density and that both sexes respond in the
same way. Therefore it appears that physiologic adaptive mechanisms are

2. Endocrines and Populations 303

operative in natural populations of Norway rats, just as they are in freely
growing confined populations of house mice or voles.

Another series of experiments with populations of rats produced results
which at first glance appeared to contradict the foregoing conclusions. Alien
rats were introduced into stationary and increasing populations to verify
earlier observations that such a procedure profoundly affected the popula-
tion (Calhoun, 1948). When stationary populations of rats were increased
20% by adding alien rats of one sex or the other, the populations declined
abruptly to about three-fifths their original values. When large numbers of
rats were substituted (native rats were removed and replaced by an equal
number of aliens) into increasing populations, the populations promptly
ceased growing. These procedures presumably produced severe social strife
with a marked increase in mortality (Calhoun, 1948; Davis, 1953). How-
ever, the adrenal weights of rats taken at regular sampling intervals failed
to reflect the increased strife and closely followed the course of the popula-
tion size. The explanation of this apparent paradox is clear. The introduced
rats were, of course, aliens — hence subordinate rats — in a strange environ-
ment (see Clarke, 1953; Barnett, 1955, 1958) and therefore became mor-
tality statistics. The native rats which succumbed were probably subordi-
nate animals in the original population. The adrenal weights were only from
survivors and therefore probably from dominant animals. It already has
been pointed out that adrenal weight is least in dominant animals, and so
the increase in adrenal weight which might have been anticipated from a
superficial examination of the situation obviously and simply failed to
appear. Had the adrenals of the rats that died been obtainable, they un-
questionably would have exhibited profound changes.

Natural populations of voles (Microtus) have also been studied to deter-
mine the relation between adrenal function and population density. Adams,
Bell, and Moore (Christian, 1959b) periodically collected frequent samples
from a natural population of Microtus montanus in Montana for four years
to obtain indices of population density, adrenal weights, and data on re-
productive activity. The number of males collected was inadequate for
valid comparisons. However, population density and the adrenal weights
of female voles were closely correlated. The population was at peak densities
in late summer of 1952 after which there was a continued general overall
decline in the population until the summer of 1955. The adrenal weights of
females were maximum in the late summer of 1952 and then underwent a
gradual overall decline, generally following the population. However, within
each year there were marked seasonal changes in the population which were
reflected by equally marked changes in adrenal weights. The population
reached its annual maximum density in late summer or early fall and then
declined sharply to very low levels where it remained until late winter or

304 /. ,/. Christian

early spring when breeding recommenced and the populations increased
sharply. Changes in adrenal weight closely followed these animal cycles in
population density. Peak adrenal weights were found in early fall and coin-
cided with a cessation of reproduction and the beginning of the autumnal
decline in the populations. Adrenal weights were minimal by November
and remained there until March, ^ when the population again began to in-
crease. Adrenal weights were maximum in the last summer of 1952, aver-
aging about 15% greater than in any other year for a comparable time of
year. The mean weights in winter were roughly 60% less than the maxi-
mum. These results indicate that the supposedly severe winter conditions of
the Rocky Mountain Region of Northern Montana do not impose physio-
logic hardships on voles. Climatic conditions, if they affected the animals
at all, apparently were unimportant compared to changes in population
density. The sharp declines in the population every fall evidently were due
to a cessation of breeding and continuing random mortality, not to a de-
clining environment, otherwise the adrenal weights would not have de-
creased so spectacularly (Christian, 1959b). It is possible that physiologic
responses to high densities in late summer contributed to the annual de-
clines in population density, but it seems unlikely except in 1952, In that
year it is probable that physiologic adaptive responses were at least partly
responsible for the decrease in the size of the population. It is noteworthy
that the rapid decline from peak densities in the late summer and early
fall of 1952 was not caused by increased mortality from disease (Adams et
al.; cited in Christian, 1959b).

Adrenocortical-density relationships were studied by Louch (1958) in
two natural populations of voles (Microtus pcnnsylvanicus) . He used
eosinophil counts, supplemented by adrenal weights, as indices of adreno-
cortical activity. Relative adrenal weights were high and eosinophil counts
remarkably low in one population at peak density. Following a sharp de-
cline in this population, indicated by an 85% drop in the population index,
the eosinophil counts rose sharply about 500% and continued to rise more
gradually for the subsequent nine months of the study. Adrenal weight
averaged 59% lower during the period of low density than during the pre-
ceding period of high density: 14.5 mg./lOO gm. compared to 23.0 mg./lOO
gm. The size of the second population remained low tliroughout the experi-
mental period. The eosinophil counts began high and gradually rose to
higher levels. Adrenal weight declined about 27% (from a mean of 19.4 to
a mean of 15.2 mg./lOO gm.) in the second population over the same period

3 Subsequent work has suggested that the low winter adrenal weights are partly due
to sexual inactivity and regression, although not entirely as a number of the voles were
not sexually regressed. In other words the adrenal weights should be low due to density
factors (cf. October), but these were too low [see Christian (1961, 1963b)l.

2. Endocrincs and Populations 305

as the 59% decline in the first population. However, the decline was not
nearly so great and the initial weight was 19% less than in the first popula-
tion. The lower adrenal weight was reflected by a higher mean eosinophil
for the same period of time. Even though the reason for the lower adrenal
weight in the second population during the second period is not known,
adrenal weight and eosinophil counts in these studies generally reflected
differences in population density (Louch, 1958).

Methods for estimating most natural populations of mammals are at
best relatively crude, insensitive, and subject to many errors. At best they
can detect only relatively large changes in population density. The catch
per trap-night or any comparable index of population density obtained by
some sort of uniform trapping effort is probably as good an index as any
readily usable procedure. The catch per unit of time was used in these
studies of vole populations by Louch (1958) and Adams et al. (Christian,
1959b) , as well as in the studies of rural populations of Norway rats (Chris-
tian, 1959b) . Errors in estimating population density in this fashion may
account for most of the observed discrepancies. If the existence of physio-
logic responses to population density is established for a variety of species,
it is likely that the magnitude of these responses will provide a much more
precise index of relative population density (the important figure for practi-
cal purposes) than any existing indirect method for determining population
size.^

Limited studies with natural populations of other species have been made.
A 68% decline in the index of the size of a population of white-footed mice
{Peromyscus leucopus) from one July to the next was accompanied by a
58% decline in their mean adrenal weight (Christian, 1959b). Similar, but
less conclusive, results have been obtained with natural populations of
short-tailed shrews [Blarina) (Christian, 1954, and unpublished). Beer
and Meyer (1951) studied the seasonal changes in the endocrine and
i-eproductive organs of muskrats and found a marked peak in adrenal
weight in adults of both sexes in early fall and a second minor peak in
March and April, especially in adult females. These seasonal changes are
similar to those we have noted in Microtus montanus and may be related
to behavioral and social changes.

Preliminary work on the relationships of the weights of deer adrenals to

" The persistence of immature zonation in the adrenals of males and lack of hyper-
trophy due estrogenic stimulation in females may explain some of the discrepant re-
sults with Microtus, and possibty Lemmus, as these would seriously confound adrenal
weight relationships if the population under study is composed of an appreciable per-
centage of sexuallj' immature, regressed, or otherwise sexually inactive animals. Winter
or high-density populations will consist largely of such animals [see Christian (19(il,
1963b)].

306 J- J- Christian

various factors indicates that adrenocortical physiology in this species, as
in others, reflects environmental changes, and the adrenal was therefore
said to be an indicator of conditions as in any other species. (Hughes and
Mall, 1958). However, sika deer (Cervus nippon) respond to changes in
population density in a manner quite similar to that seen in rats (Christian
et al., 1960) . A 60% decline in the population of these deer was accompanied
by a proportional decline in adrenal weight. The decline of this population
was brought about by a mass mortality which apparently was due to
metabolic disturbances resulting from the prolonged adrenocortical hyper-
activity associated with a high density of the population. There is evidence
that potassium deficiency resulting from cortical hyperactivity may have
contributed directly to the mortality. Marked stunting of growth was also
seen during the period of high density and especially during the year of
die-off. Other possible causes of the die-off, such as malnutrition, could be
ruled out.

More experimental work is needed to relate adrenal function with popula-
tion density for a number of species in natural populations. The problems
are numerous and not the least of these is being able to rule out extraneous
factors or else to assess their role in producing the measured effects.

2. Reproductive Function

A large number of studies suggest that reproductive function is depressed
with increasing density of natural populations, but studies sufficiently
discriminative to attribute changes in reproductive function to changes
in density per se without additional complications are few. Therefore a
great deal of the evidence is circumstantial and tentative at best. The
present discussion is limited to those studies which have been conducted
with suflficient care and attention to a variety of details so that one is
confident that the factors have been properly considered which possibly
might affect reproduction. However, a brief discussion of various problems
involved in evaluating reproductive function is in order before considering
the evidence implicating physiologic responses to population density in
the suppression of reproductive function.

First there is the problem of the food supply of the experimental popula-
tion. As we have seen, there is ample evidence that food deficiencies can
curtail reproductive function independently of other adaptive mechanisms.
However, the effects of food and of increased density on the reproductive
system and growth are so similar that unless careful assessments are made
of the food supply (preferably these should be accompanied by an evalua-
tion of adrenocortical and other adaptive fvmctions, no separation can be
made between the effects of food and of density, especially at relatively

2. Endocrines and Populations 307

high population densities. Furthermore, it is very likely that the adaptive
reactions to density and their effects may overlap food shortages, so that
their effects are mutually augmentive at critical densities. However, the
aim of the present discussion is to assemble the evidence that adaptive
reactions to density occur irrespective of whether or not the effects of in-
adequate food or other environmental factors are superimposed. The growth
of populations unquestionably can be limited by environmental factors
which may either act through the physiologic adaptive mechanisms or in-
dependently of them, especially in localized populations. However, the
basic concept of physiologic adaptation to population density is that these
mechanisms are always operative, and will regulate and can limit popula-
tion growth. Finally, limitations of enviroimiental factors, even of food,
may increase competition directly and therefore indirectly produce in-
creased activity of the physiologic adaptive mechanisms. Pitelka's (1957b)
statement that "the interest in the stress mechanisms has led some students
to overlook the point that such a mechanism does not evolve without
linkages to critical variables extrinsic to the population" bear repeating
and is emphasized and supported by an immense amount of work with rat
populations (Davis, 1953). Nevertheless, it is extremely miportant to dis-
tinguish between the direct action of environmental variables on popula-
tions and their indirect action through sociopsychologic and physiologic
mechanisms; and experimental evidence to date emphasizes the general
importance of the latter mode of action. Evidence of the direct action of
food shortages in depressing reproduction in natural populations of Pero-
myscus boylii and P. maniculatus is cited by Jameson (1953, 1955), and
population declines evidently followed the shortages as a result of con-
tinuing normal mortality in the presence of a lull in reproductive activity
with no recruitment into the populations. The evidence presented by Jame-
son, supported by the controlled experiments previously described, strongly
favors the interpretation that inadequate food supplies were directly re-
sponsible for the inhibition of reproduction. However, the possibility that
the shortages induced increased competition and increased activity of
adaptive responses were not ruled out completely. Davis (1951c), in a
study of Norway rats from natural populations, provided definitive evi-
dence that a deficient food supply can inhibit reproduction under natural
circumstances. However, w^hether the effects were direct or produced by
increasing competition with stimulation of generalized physiologic adaptive
response is unknown. It was determined later that the adrenals of these
rats were responsive to changes in population density (Christian, 1959b).
Other examples could be cited, but the majority could serve only to empha-
size the fact that discrimination between the effects of food and other factors
is usually not attempted.

308 /. J. Christian

Another, perhaps more serious problem in evaluating the effects of
density on reproduction is the lack of adequate criteria to determine the
age of most small mammals; so that weight and length measurements are
usually used as criteria of age. Since one of the major effects of increased
activity of physiologic adaptive responses is diminished growth, as well as
inhibition of reproductive function, there is no valid way to separate normal
immature animals from older animals which should be mature but in which
both growth and reproductive maturity are inhibited. Several serious
misinterpretations can result from this situation : ( 1 ) a shift of age composi-
tion of the population toward greater age will be obscured; and (2) repro-
ductive competence is usually assessed on the basis of the prevalence of
pregnancy, prevalence of lactation, and similar criteria in those animals
which are obviously mature, therefore total suppression of reproductive
activity with a delay in the onset of puberty will be missed altogether. Of
course in such a situation, the reproductively active and obviously mature
animals represent only the portion of the population which is least affected
by density factors, a situation to which attention has been called earlier.
The studies of Crowcroft and Rowe (1957) offer a clear illustration of
this situation in controlled populations in which the ages of the animals
were known. How would the nonreproductive mice in these populations
have been classified in the usual studies of natural populations with trapped
animals? Kalela's (1957) studies on the effects of population density on
Clethrionomys rufocanus from a natural population are of singular value
because (1) he had a means of determining the age of the animals by the
rooting of their molars and was therefore able to state with certainty that
growth and maturity were totally inhibited, and (2) he eliminated the
possibility that food was an etiologic agent in producing these phenomena.
Of course, all gradations of these effects may occur to further confuse the
situation. Most studies which evaluate changes in reproductive function of
small mammals, for which there is no adequate way of determining age, in
relation to density in populations must be viewed with considerable skepti-
cism. It has already been pointed out that the usual means of determining
reproductive competence do not detect changes short of almost total
inhibition, especially in males.

A third problem, less frequent in its occurrence, is the failure to take into
consideration changes in age composition when evaluating changes in litter
size, although this frequently results from the inability to determine age.
Hamilton (1937) reported that litter size in Microtus pennsijlvanicus in-
creased with increasing population density and yet there was increased
intra-uterine mortality at the higher densities. It is well known that litter
size increases with parity for the first few litters and that there usually is a
direct relationship between body weight and litter size for most litter-

2. Endocriues and Populations 309

bearing mammals (Watt, 1934; Hoffman, 1957; Flick et at., 1959). There-
fore it is entirely possible that the apparent increase in litter size noted by-
Hamilton was a result of a shift in age composition, and even if this is not
the case, the point is generally valid and is well illustrated by Hamilton's
data. These data bring up another point which often proves confusing;
that is there is no word in use to designate an in utero group of embryos.
"Litter" used for both postpartum and antepartum young with the result
that ''litter size" is frequently determined from embryo counts and conclu-
sions drawn therefrom regarding litter size at birth. Although the authors
are themselves usually quite clear on these differences, their descriptions in
the literature often are not, and the reader will arrive at false conclusions
or else be unable to make any at all. This situation can be corrected by the
use of the collective noiui gravidum to describe the in utero counterpart of
"htter" (Snyder and Christian, 1960).

Finally, there is a wide divergence in the descriptions of the effects of
population density on reproduction. It seems clear from experimental and
field data that the reproductive processes can be inhibited at a number of
different points, evidently depending on the severity of the stimulus, dura-
tion of the stimulus, age at which the stimulus was first effective, the posi-
tion of any given animal or groups of animals in the social structure, prob-
ably individual physiologic and psychologic differences, as well as other
unknown factors. The details of these various relationships are not known,
but experimental results suggest that the smallest stimulus inducing an in-
creased adaptive response with inhibition of reproduction in mature ani-
mals will affect lactation, increase litter mortality, and possibly depress
fertility; a greater stimulus will result in increased intra-uterine mortality
and an even greater stimulus will totally suppress reproductive function.
It is suggested that as the alarming stimulus increases in severity there is
a progressively greater suppression of the secretion of gonadotropins. It is
equally evident that young mammals reaching maturity are far more sus-
ceptible to a given alarming stimulus than an animal which has already
reached maturity. It is known that immature rats and mice are many
times more sensitive to suppression of gonadotropins by steroids than the
mature animals (Byrnes and Meyer, 1951). The results of experiments
discussed in the preceding sections show that these statements apply to the
effects of population density. The effects of vitamin and other nutritional
deficiencies follow a similar pattern; furthermore, the earlier in the process
of reproduction the deficiency begins, the more profound is its effect on the
developing fetuses (Lutwak-Mann, 1958). Usually a deficiency beginning
at the start of pregnancy or earlier will produce a very high percentage of
resorptions. These decrease the later the deficiency begins. It is noteworthy
that only "borderline" deficiencies appear to produce congenital anomalies.

310 J. J. Christian

So it appears to be with population density. Only those female mice which
were moderately affected by density produced young which were then
permanently affected by reduction in lactation, whereas those mice more
seriously affected failed to produce at all and exhibited 100% resorption of
embryos (Christian and LeMunyan, 1958). Therefore, when one considers
the enormous number of variables affecting a population, it is not sur-
prising that there is so much variation among individuals and among
populations wdth respect to the effects on reproduction. Reproduction
should include the entire process of producing young from maturation of the
parents to weaning when discussing or measuring the effects of physiologic
adaptive mechanisms and the effects on the endocrine organs.

In spite of these problems, a number of studies on natural populations of
small mammals have been sufficientlj'' detailed and critical to permit exami-
nation of the relationships between population density and reproduction.
The most conclusive of these have been carried out on Norway rats (Rattus
norvegicus) and voles {Microtus sp. and Clethrionoimjs) , but there have been
others on muskrats (Ondatra), cotton rats {Sigtnodon) , woodchucks {Mar-
mota) , and hares (Lepus americanus) . Most of these suffer to one extent
or another from a lack of dependable criteria for determining age, as dis-
cussed above.

The reproducti^'e performance of Norway rats from increasing, de-
creasing, and stationary populations was investigated (Davis, 1951a).
These rats have a major reproductive season in late winter and early
spring and a secondary minor peak in reproductive activity in the fall.
Increasing populations have a higher prevalence of pregnancy than either
decreasing or stationary populations. These differences are especially
marked in the spring breeding season when the prevalence of pregnancies
was 41.6% in increasing, 25.3% in decreasing, and 14.4% in stationary
populations. The incidences of pregnancy were, repsectively, 6.1, 6.3, and
3.8 per year. However, there were no apparent differences in litter size and
the prevalence of lactation Avas essentially the same in all three categories
although the data on lactation could not be analyzed separately for the
two breeding seasons. However, an analysis of lactation by seasons could
not account for the similarity of the overall figure and one must conclude
paradoxically that, although the pregnancy rate was highest in increasing
populations, there was also a greater parturitional mortality and at the
same time a better survival of those young which were not lost soon after
birth.

Similar results were obtained when reproductive performances of Norway
rats from a rural population were compared at two different relative popula-
tion densities. The pregnancy rate was 48.3% a month after artificially
reducing the population a third, compared to 14.4% before the reduction.

2. Endocrines and Populations 311

These results indicate that reproductive performance is inversely related
to population density in Norway rats, at least for increasing and stationary
populations. The status of decreasing populations is equivocal and may
represent a partial recovery of reproductive function in declining popula-
tions, although decreasing populations are presumably at maximum den-
sity. However, we already have noted that adrenocortical hypertrophy is
greatest in populations spontaneously decreasing from asymptotic levels.
These results represent a large number of populations and carry consider-
able weight; nevertheless in another experiment in which three urban
populations of Norway rats were artificially reduced about 50% there was
no corresponding increase in the rates of pregnancy (Davis and Christian,
1958). Two factors may have obscured such an occurrence. One was that
the populations were increasing and therefore had a fairly high pregnancy
rate at the time of the reduction (30.2%) , and the other was that the main
breeding season was declining at the time the next samples were collected
(April, May, and June) from which a mean pregnancy rate of 32.2% was
obtained.

Experiments designed to analyze reproductive function in relation to
density of populations of rats suffer from the inability to determine the
age of the animals, therefore attainment of maturity and other reproduc-
tive end points are based on the sizes of the animals. There is evidence to
indicate that rats in increasing populations grow faster, and so are larger
at a given age, than rats from stationary populations (Davis, 1951b).
Therefore inhibition of growth in populations of high density may tend to
obscure evidence of partial inhibition of reproductive function, such as
delayed puberty, since it is likely that reproductive function and growth
are equally inhibited. Thus, these results are evidence to support the earlier
comments regarding the desirability of determining age in mammals when
evaluating reproductive function.

One of the more conclusive studies to date on the relationship between
density of population and reproduction was conducted by Kalela (1957)
on red-backed voles {Clethrionomijs rufocanus) in Finland. A major factor
contributing to the value of this study was the ability to determine and
approximate age of the animals by the root development of the first molars.
These molars are rootless at the end of September of the year of birth, but
the roots are visible by the following spring and continue to develop further
(Zimmermann, 1937; Kalela, 1957). The question might be raised whether
the growth and rooting of the molars would be suppressed along with sup-
pression of growth, but molars were rooted in animals in which growth and
maturity were definitely suppressed (Kalela, 1957) ; so that any suppression
of molar growi^h and rooting would make the results even more dramatic.
[]This and other means of estimating age in voles has been more fully re-

312 /. /. Christian

viewed by Bourliere (1951).] Kalela also determined that mature sperma-
tozoa are present in the testes of these voles when the testes are seven or
more millimeters long and defined maturity on this basis. Females which
were pregnant or parous were defined as fecund.

During the first summer of Kalela's studies in 1954, moderate population
densities were achieved after a spring characterized by a very small popula-
tion which had overwintered from the preceding fall. Males and females
from the litters born early in the breeding season matured without excep-
tion. Four-fifths of the mature females of late summer were young of the
year. A large number of the young of the year which reached maturity had
two litters and some even had thi-ee during their first summer. The old
overwintering females had as many as four litters. The following spring,
1955, the population started with a much larger number of overwintering
animals. The rate of growth of this population at first was as high or higher
than the rate of growth of the population in the preceding year, but it
rapidly declined as the density of the population increased. Peak densities
were reached in this year. A good many of the males born in the first part
of the breeding season reached maturity, but males with enlarged testes
were rarely found in August, and there was an accumulation of immature
males weighing between 20 and 24 gm. at this time (mature red-backed
voles normally weigh more than 24 gm., usually between 30 and 40 gm.).
This weight range represents the maximum size achieved by male or fe-
male red-backed voles which fail to mature, irrespective of their age. There
was no accumulation of voles in this weight category during the preceding
year of low population densities. Female young of the year exhibited the
same cessation of growth and failure to mature, except for those born early
in the breeding season. Less than half of the mature females in late summer
were young of the year and there was a marked accumulation of immature
females in the 20-24 gm. M^eight class. Furthermore, the pregnancy rate fell
sharply and had fallen to well below the 1954 levels by mid-August. The
overwintering females had up to three litters, but none had four litters.
Those young of the year which reached maturity had no more than two
litters and only a few had this many. One area (Malla) had twice the
density of the others under study at this time, and none of the young of the
year, male or female, attained sexual maturity or grew beyond the 20-24
gm. category.

The population had undergone a major decline in numbers by the spring
of 1956 and was well below the 1954 levels for the same period, and yet the
rate of growth of the population was extremely slow. The overwintering
animals, as well as the young of the year, were distinctly smaller than
normal. Reproductive performance was essentially identical to that of the
preceding summer of 1955 when the density of the population was extreme.

2. Endocrines and Populations 313

The rates of fecundity, pregnancy, and other measurements of reproductive
competence were unchanged from the preceding year. There was no appar-
ent change in htter size or intra-uterine mortaUty throughout the period of
the study of this population. There was no shortage of food at any time
during these studies, and especially not during periods of active reproduc-
tion. Therefore, the changes in reproductive function are not attributable
to ^'arying food supplies. Climatic factors, although variable, did not corre-
late with the functional changes and were probably noncontributory in
producing functional changes.

This study shows, beyond reasonable doubt, that the shifts in reproduc-
tive function were reciprocally related to the density of the population in
1954 and 1955 and that increased density was the factor responsible for
the decline in reproductive function (with a total suppression of reproduc-
tion in the young of the year in one area) and attainment of maturity in
1955. It also shows that the primary effect of increased density was to in-
hibit maturation so that large numbers of immature animals accumulated
by late summer. At the same time, growth beyond a general level of 20-24
gm. was suppressed in both sexes. The data from the area of extreme
density, in which no animals matured are especially convincing. No direct
measures of endocrine function were made, but these occurrences in natural
populations of Clethrionomys rufocanus conform with the results from the
more highly controlled experiments with populations of mice or voles in
the laboratory.

The apparently paradoxical failure in 1956 of reproduction to return to
the 1954, or higher, levels, together with the apparent stunting, also fits
the conclusions discussed earher. The effects of increased density on lacta-
tion and subsequent growth of the young and in turn their inability to
lactate adequately were discussed earlier, and attention was called to the
profound long-range effects of increased density on the animals experiencing
the increase as well as two generations of their progeny (Chitty, 1955;
Christian and LeMunyan, 1958) . The red-backed voles which overwintered
until the spring of 1956 had been subjected to increased density and had
experienced its effects. It was noted that these voles were unable to mature
in the year of their birth and were stunted. The following year their repro-
ductive performance and that of their offspring reflected a similar situation,
closely resembling the effects of increased density on laboratory popula-
tions. The population dechne was consequently due to the high mortahty
of those young born during the period of maximum density and probably
reflects decreased resistance to a variety of stimuli coincident with increased
activation of adaptive mechanisms. A very high rate of infant and juvenile
mortality was a major factor in the cessation of growth in laboratory popu-
lations, but in these there was no appreciable mortaUty of the adults.

314 J- J- Christian

There is a continuous high rate of mortality of all age groups in natural
populations of these small mammals so that a failure of reproduction and
a sharply increased mortality rate of juvenile animals make a collapse of
the population easily understood. The prolonged recovery from this decline
also can be attributed to the effects of increased density, apparently through
defects in lactation as well as other endocrine reproductive functions, al-
though the details of these mechanisms require much more investigation.
It should be noted in these populations; that while all phases of reproduc-
tion were affected the particular functional aspect which was most severely
affected appeared to depend on the level of density and on the age of the
animals involved — the young being most severely affected, as was the case
in laboratory populations. It appears that the responses of this natural
population to density were identical in every respect to those seen in labora-
tory populations with the exception of the ever-present mortality which
occurs in the wild.

Reproductive function e\ddently is inversely related to density in natural
populations of red-backed voles {Clethrionomys) , and the inhibition of re-
production and growth probably resulted from a suppression of the secre-
tion by the anterior pituitary of gonadotropins and growth hormone. There
is no other explanation evident which fits the known facts, even though
there was no direct assessment of pituitary function. The apparent paradox
of a low population exhibiting a marked inhibition of reproduction and
growth following a decline appears to result from the prolonged effects of
high population densities on these animals. These conclusions derived from
Kalela's studies carry particular weight, as food, and apparently other
environmental factors, appear not to have been causative agents. Kalela
(1957) discusses the social problems in these populations.

A quite similar study was conducted on Microtus montanus (Hoffmann,
1958) with comparable results. This population was followed for 3 years
and there was a peak between the second and third breeding seasons which
was followed by a marked decline in the size of the population. The popula-
tion density was moderate in the spring of 1952, increased during the sum-
mer and early autumn, and was followed by the annual decline in density
subsequent to the annual cessation of breeding. The population in the
spring of 1953 began moderately and increased sharply until September.
The spring population of 1954 was higher than in 1953, but the population
declined rather than increased during the breeding season and was at a
very low level in the spring of 1955. Thus the history of this population of
Microtus montanus was in many respects similar to that of Kalela's popula-
tion of Clethrionomys. The published data indicate that Microtus montanus
may differ from most other small mammals in that litter size remains
constant with respect to age, weight, and parity (Hoffmann, 1958) . There-

2. Endocrines and Populations 315

fore changes in mean litter size apparently do not need to be corrected for
these variables. During each breeding season of the study there was an
appreciable decline in litter size as measured by embryos and uterine im-
plantation scars and this decrease was also paralleled by a decline in the
number of corpora lutea. There was also an increase in the general level of
the population for comparable months from 1952 to 1953. This general in-
crease coincided with an overall decline in mean litter size which, however,
was not significant. These changes were inversely related to the annual
increase in population density for the first two years. However, the popula-
tion declined sharply from spring to fall in the third year and yet the mean
litter size also declined.

Prenatal mortality was directly related to population density in the first
breeding season, inversely in the third and bore no conspicuous relationship
in the second. Prenatal mortality was greatest after the population had
declined in September 1954. The mean prevalence of pregnancy also de-
clined with each successive year.

The peak of the population was evidently in September of 1953, and the
decline followed shortly thereafter. Hoffmami noted that at this time the
proportion of fertile males began to decline and at the same time 24% of
the mature females weighed less than 33 gm. This figure increased to ap-
proximately 45% in June and July of the following season. The proportion
of mature males weighing less than 35 gm. also rose to reach a maximum of
33% in July of 1954. These morphologic functional changes resemble those
observed by Kalela and probably reflect suppression of growth at high
densities. The fact that there was a decline in male fertility as well as in the
incidence of pregnancy further supports such a conclusion. Hoffman called
attention to the fact that prenatal mortality was lowest at peak densities
and highest the following year. However, examination of his data
reveals that the proportion of multiparous females rose steadily throughout
the 1953 season to reach 100% in September, undoubtedly accounting for
for the low proportion of prenatal losses for two reasons: (1) it is likely
that the older parous females suffer less prenatal loss at any time; and (2)
more important, they probably represent the dominant animals and there-
fore those least affected by high densities in accordance with the results of
laboratory experiments (Retzlaff, 1938; Christian and LeMunyan, 1958).
The age of the animals could not be determined in these studies ; immature
animals were classed as such and were not further divided into those old
enough to have matured and young animals. Had it been possible to deter-
mine age, it is likely that these results would have been more decisive.
There was a low mortality of weanlings and juveniles during the period of
build-up in this population, but apparently there was a marked increase in
mortality in all age groups during the decline.

316 /• /• Christian

These data, like those from the preceding study, indicate that the effects
of high densities on reproduction and growth persist for some time after
the peak has been reached and passed. It should be remembered that all
animals beginning the breeding season after a peak are sur^-ivors from the
peak densities. Although the highest numerical size of Hoffmann's popula-
tion was in September of 1953, the actual peak could just as easily have
been the relatively high spring population of 1954. The data suggest this
to be the case, as there was a brief period at the beginning of the 1954
(June) breeding season in which prenatal and infant mortality were lower
than usual and then climbed sharply. Hoffmann concluded that ovulation
rate and Utter size tend to vary inversely with population density and thus
to run counter to the cychc trend of the population, except during the
crash decUne period in Microtus montanus, when lowered Utter size may
contribute to the drop in density. He also concluded that speed of attain-
ment of reproductive maturity remained rather constant despite changes
in density, which is quite the reverse of Kalela's results and conclusions.
This conclusion is, however, based on the assumption that there is no
inhibition of growth or that it is totally independent of changes in reproduc-
tive function. Since the age of these voles could not be determined, age
was assumed to be reflected in body weight.

In view of the likelihood that inhibition of growth and reproduction are
not independent, but rather are two manifestations of one reaction, as we
have seen for Clethrionomys as weU as for Mus in experimental populations,
the conclusion is not justified that the speed of attainment of maturity
remains constant in the presence of changes in density. Furthermore,
Hoffman's figures indicate that a high proportion of mature animals were
much smaller in peak and declining populations than usually found at
lower densities. Of course these data could be interpreted to indicate an
early attainment of maturity with respect to size, but this interpretation is
most unUkely in view of the other data on reproduction and the results of
other experiments. This problem could be settled by determining the age
of these animals; a problem ah-eady discussed at some length.

Chitty (1952) also studied a population of voles {Microtus agrestis) and
obtained results similar to those of Kalela. During the peak breeding season
of 1937 he found that young males did not increase in weight above 22.2
gm., whereas normally adult males seldom weigh less than 22.3 gm. after
May. Young females weighed more than 22.2 gm. only while pregnant;
otherwise the heavier females in this population were old adults with a few
exceptions in September and October. The prevalence of pregnancy in the
peak year was apparently less than in other years and appeared to be due
mainly to a failure of maturation in the young of the year, as older animals
exhibited a uniformly high pregnancy rate. In one area all age groups ceased

2. Endocrines and Populations 317

breeding in August of the peak summer and the survivors did not breed
the following spring after the population had declined. Other populations
which were at peak density the following year continued to breed from
May to October and survivors bred again the following year, after the
dechne, although no recovery in population occurred. D. Chitty concluded
that lowered fertility was not invariably associated with a decline in num-
bers. The decreases in size in all these populations primarily was due to an
excessive juvenile mortality before August which was attributed to some
adverse effect of high population density on the ability of the young to
survive.

Hoffmann (1958) also studied a population of Microtus calif ornicus and
noted a general inverse relationship between density and reproductive
function which occurs on a seasonal basis. In other words, the annual in-
crease in the population also has density effects on its members. Fitch (1957)
noted a decline in litter size in Microtus ochrog aster with progression
of the breeding season and increase in the population, as noted by other
authors, but he attributed this to an increasing number of young of the
year reaching maturity and having small first litters. However, this explana-
tion does not account completely for the observed progressive decline in the
population. A similar relationship has also been noted in muskrats (Ondatra
zibethica) (Errington, 1948, 1951, 1954a). Errington (1954a, 1957) has also
shown that reproduction may be depressed during the low years following
high densities and increase during an increasing phase of the population.
The depressed reproduction probably is another example of the prolonged
effects of high population densities on succeeding generations.

The pregnancy rate was high early in breeding season during the peak
year in a population of voles (Microtus montanus) studied by Adams, Bell,
and Moore (Christian, 1959b) , but dropped precipitously after June so that
the percentage of pregnancy was below 30% for August and September —
months of peak breeding activity in subsequent years with lower popula-
tion densities.

Hamilton (1937) reported an increase in litter size with increasing popu-
lation density and a decrease in litter size during the decline. He also re-
ported increased intra-uterine mortality. However, although the ages of
these animals were not determined, it is probable that the age composition
of the population shifted so that it is consisted of a higher proportion of
older animals as the peak approached. The litter size of Microtus pennsyl-
vanicus increases with parity (Hatfield, 1935; Martin, 1956; Fitch, 1957)
differing from M. montanus in this respect; so that the increase in mean
Utter size could reflect such an occurrence. The decline in litter size during
the decrease in population size is consistent with the preceding studies.

Odum (1955) also reports a greater mean Htter size for Sigmodon in

318 /. /. Christian

"high" density years than in "low" density years. Once again weight was
used as an age criterion and so the possibihty exists that Odum's figures for
age composition in "low" versus "high" density populations may reflect
complete suppression of reproduction and growth in young animals at high
densities, rather than a shift in age composition toward younger animals as
his figures suggest.

Tanaka (1956) reported complete inhibition of reproductive function in
Anteliomys smithii {Clethriononiys smithii) irrespective of the age or sex
of the voles during a season of peak densities. The reproductive organs were
in a condition of complete quiescence. The possibility that a deficiency of
food may have been instrumental in producing these effects cannot be
ruled out. However, Tanaka (1956, 1957) points out that maximum den-
sities, or vole "outbreaks," may occur either in concurrence with or inde-
pendently of widespread flowering of bamboo grasses. The fact that the
voles apparently were equally numerous and showed similar reproductive
inhibition regardless of whether they were living in areas of living or areas
of dead bamboo strongly suggests that the inhibition of reproduction was
density dependent.

Lemmings, because of the conspicuousness of their population cycles,
have been the subjects of study by a number of investigators. Lenimus
trimucronatus has been reported to experience a complete suppression of
reproduction during periods of peak densities of their populations (Rausch,
1950) . In the spring of 1949, prior to a precipitous decline in their popula-
tion, there was an absence of young animals and no evidence of reproduc-
tive activity: none were pregnant, there were no uterine implantation scars,
few females had perforate vaginas, and the testes of the males were small
and incompletely descended. Similar results were obtained by Barkalow
(1952) in a study of the tundra mouse (Microtus oeconom.us) and lemmings
(Lemmus trimucronatus and Dicrostonyx groenlandicus) at Barter Island,
Alaska. Barkalow in summarizing the results of his and other studies on
vole and lemming populations in the Arctic states that cessation of breeding
prior to and during a cyclic decline is in his opinion a population character-
istic and that large litters are characteristic of the build-up phase of a
population cycle, especially in its early stages. Rausch (1950) stated that
although there was a reduction in the tundra vegetation there was no sug-
gestion that the decimation of the population resulted from a deficiency of
food. Thompson (1955) takes the view that exhaustion of the supply of
food and cover is the important factor, and well it may be in a local situa-
tion. However, the bulk of the evidence suggests that a limited supply of
food is not usually the factor limiting population growth. Lemmings have
an annual decline in reproduction with the annual increase of density, espe-
cially in prepeak or peak years of density similar to that seen in the other

2. Eudocrines and Populations 319

microtines discussed (Pitelka, 1957b). Pitelka (19o7a, b) also makes the
point that the cessation of reproduction at peak densities may occur and
produce a decline in the lemmings before their predators have a chance to
exploit the lemming population. Lemming populations have not been
studied for a long enough period of time or with sufficiently detailed evalua-
tion of the various factors which might inhibit reproduction to warrant any
conclusions at present. However, it is apparent that there can be inhibition
of reproduction with increased density, but the extent to which food sup-
plies may also contribute has not been sufficiently appraised. The evidence
presently available suggests that failure of reproduction in response to ex-
treme densities is the major cause of the cyclic declines in lemming popula-
tions.

Many studies indicating a decline in reproduction with high population
densities in other species, such as hares (Preble, 1908; MacLulich, 1937),
but in none of these is it possible to ascribe the alteration in reproduction
solely to density factors, even though it may well be that such is the case.
Many other studies could be listed which illustrate the same sort of thing,
but no useful purpose would be served by doing so here. More recent studies
with Microtus pennsylvamcus indicate that in this species the annual re-
productive cycle is governed by population density, but this does not appear
to be true for Peromyscus manicidatus (Christian, 1961).

The main point of interest is that reproduction does respond to changes
in population density in natural populations of voles and rats and probably
in other .species as well. In some studies there was a reciprocal relationship
between adrenocortical and reproductive function, in others only one or
the other of these general responses was studied, but the results conform
to this general relationship. It seems undeniable, especially in the light of
studies on laboratory populations, that physiologic adaptive responses are
operative in natural populations and are reflected by curtailment of repro-
duction and growth and stimulation of increased pituitary-adrenocortical
activity in natural populations. However, detailed and well correlated
studies on a variety of species still are required and for a variety of different
populations in order to assess the relative importance of these mechanisms
in curtailing free-for-all population growth.

3. Disease Resistance and Mortality

One of the major points of the hypothesis that physiologic mechanisms
could control population growth as a feedback system acting in response to
changes in density was that the sudden mass mortality seen at the end of a
build-up of populations to peak densities could be accounted by exhaustion
of the adaptive responses (Christian, 1950b). It was suggested that the

320 /. /. Christian

symptoms so frequently observed (Hamilton, 1937; Green and Larson,
1938 ; Green, et al., 1939) were the result of hypoglycemic shock as a result of
pituitary-adrenocortical exhaustion. The subject remains at that point to-
day, except that Frank (1953) has succeeded in showing that hypoglycemic
shock, which is uncorrected by epinephrine but is corrected by glucose in-
jection, occurs in Microtus. Frank (1953) also succeeded in producing mass
mortality from hypoglycemic shock by crowding voles in enclosures. How-
ever, he believes that in addition to increased density alone, which sets the
conditions for a mass mortality, a trigger, such as competition for food must
also be present. This seems a reasonable suggestion which perhaps can be
modijfied to be included in density-dependent factors. On the basis of
Frank's and other experimental data it appears that competition for any
factor is also a matter of relative population density and inseparable from
it. In this case it seems that the voles were reducing the environmental
capacity at high densities and were therefore themselves setting up the
conditions for increasing competition among themselves above the level
that was already present.

It has been mentioned that mass mortality in sika deer was probably a
result of electrolyte imbalances brought about by prolonged adrenocortical
hyperactivity. There were indications of a prominent role of the zona
glomerulosa, presumably with increased secretion of aldosterone, to explain
the evidence of potassium deficiency (Christian et al., 1960; Christian,
1963).

Frank (1953) was able to eliminate infectious disease and climatic
factors as causative agents in precipitating the mortality in these animals,
as were Adams, Bell, and Moore (Christian, 1959b) in Microtus montanus.
Rausch (1950) also indicated that neither parasitism nor infectious disease
could account for the sudden decline in the lemming population he was
studying. In a study of mortality in an experimentally induced epidemic
in Norway rats, Davis and Jensen (1952) found that mortality rate was not
significantly changed and probably would be so only under special circimi-
stances. A precipitous population decline due to mass mortality from hypo-
glycemic shock is probably only one of several mechanisms which may
operate to reduce population density drastically and it may be a rather
special occurrence. Our knowledge of the physiologic mechanisms involved
in invoking "shock disease" in natural or experimental populations is little
better understood now than previously. It was originally postulated that
adrenocortical exhaustion was the cause, but it is just as likely, if not more
so, that readily mobilizable sources of glycogen and glucose become totally
exhausted with continued overstimulation of the adrenal medulla and cor-
tex. There may be exhaustion of other parts of the responsive system, such
as the pituitary, hypothalamus, or of the adrenal medulla. None of these

2. Endocrines and Populations 321

possibilities has been explored and in retrospect the idea of attributing
hypoglycemic shock solely to adrenocortical exhaustion seems naive. That
the hypoglycemic shock syndrome does occur cannot be denied, but its
physiological causes remain to be determined. Such indirect evidence as is
available indicates that it is a very complex affair. Mice and voles have
been observed to die of hypoglycemic shock shortly after being placed
placed together in groups (Frank, 1953; Christian, 1955a), but if moribund
animals are removed from their cagemates, they recover (Frank, 1953; J. J.
Christian, unpublished) . The rapidity with which this occurs suggests some-
thing other than cortical exhaustion, perhaps medullary exhaustion or
exhaustion of readily available glycogen reserves. In addition, there is
evidence that adrenal cortical exhaustion may not occur per se, but may
be an apparent result of exhaustion elsewhere in the adaptive system (Ro-
senfeld, 1958).

At the present time it appears that a failiu-e in reproductive function
may account for a decline in a population much more frequently than shock.
It has been observed in several species that the young born during peaks
of density appear to be much more susceptible to various mortality factors,
and their deaths largely account for precipitous population declines (Green
and Evans, 1940a, b, c; Chitty, 1952, 1954; Godfrey, 1955). This dispro-
portionate mortality of young animals may result from three different
factors that have already been discussed: (1) they are the subordinate
animals in a population and therefore, as we have seen, are much more
severely exposed to sociopsychologic pressures than older and more domi-
nant animals; (2) they appear to be more sensitive to these pressures and
respond accordingly; and (3) they are adversely affected by deficient
lactation on the part of their mothers, and then- progeny in turn are
affected. The combination of these factors would be expected to have pro-
found effects on the young, and we have already seen that their reproduc-
tive function is severely depressed. Excessive stimulation of the adaptive
responses in these animals would also be expected to result in exhaustion
and shock. Frank (1953) has shown that socially inferior voles are the
ones primarily affected and that succumb to shock.

In addition to these effects of social pressure and increased activity of
physiologic adaptive responses, one would anticipate a marked decline in
resistance to infection, parasitism, or other harsh environment experiences.
It is not surprising that epidemics are frequently observed following peak
densities and a general decrease in host resistance. It is particularly signifi-
cant that a variety of agents may be found to produce death in a given
population, apparently as a result of decreased resistance (Chitty, 1954) .

It is not the intent of this account to dwell on the causes of mortality in
natural populations except insofar as a decreased resistance may result from

322 J- J' Christian

increased physiologic adaptation to increased density. A causal relation-
ship between population density and parasitism in hares is suggested by
the studies of Erickson (1944). Parasites were scarce when population
densities were low and maximiun parasitism was found in populations of
hares at peak densities. The hare populations declined but the parasites
remained abundant in the surviving animals. There was no evident correlation
between shock disease and the parasites in these hares other than that
parasitism was at high levels when the population began to decline owing
to shock. Similarly, a close correspondence between the level of parasitism
of rabbits by Eimeria stiedae and population density was found in New
Zealand (Whittle, 1955) and it was suggested that the differences in the
incidence of severe infections were correlated with differences in host re-
sistence, which in turn may have been a function of host density (Bull,
1955, 1957, 1958). It is of interest that young rabbits were the ones pri-
marily affected in these populations (Tj^ndale-Biscoe and Williams, 1955;
Bull, 1958). In a study on parasitism by lice on populations of Microtus
penrisylvanicvs and Peromyscus maniculatus it was found that the level of
parasitism varied with season in both the voles and the deer mice, but that
the size of the louse population changed significantly with season only in
the male voles (Cook and Beer, 1958). It was suggested that the spring
maximum in the louse populations on the male voles came at the period of
greatest "stress" and therefore the increase in the louse population at this
time resulted from decreased host resistance.

The evidence for a decrease in host resistance to specific diseases or
parasitism in natural populations is not conclusive, but is certainly sugges-
tive. However, there can be little (luestion that there are severe physiologi-
cal derangements associated with high population densities which result in
high mortality rates in voles, lemmings, and hares (Green and Evans,
1940b; Elton, 1942; Rausch, 1950; Chitty, 1952, 1954; Frank, 1953; God-
frey, 1955), and that this mortality cannot be accounted for by infectious
disease, even though the prevalence of infection and parasitism may be
increased. It is equally evident that young animals are more severely af-
fected than older animals and that the effects are prolonged and account
for continued declines in populations for appreciable periods following a
sharp reduction in population density (Chitty, 1952, 1954, 1955; Christian
and LeMunyan, 1958; Christian, 1959a, b).

In summary, there is much indirect evidence indicating that a decline of
a population from peak densities may be caused by a density-dependent
activation of physiologic adaptive mechanisms and exerting their effects
in one or all of several ways:

1. There may be direct mass mortality due to physiologic exhaustion
following prolonged, excessive stimulation of the adaptive mechanisms.

2. Endocrines and Populations 323

2. There may be increased mortality due to parasitism or infectious
disease due to decreased host resistance (presumably from suppression of
inflammatory and immune defense mechanisms by increased adreoncortical
activity) .

3. There may be partial or complete inhibition of reproductive function,
especially in the younger animals and with prolonged effects on the next
generations of progeny, apparently initiated by deficient lactation on the
part of mothers subjected to mcreased density, but also owing to the fact
that the young probably are most severely affected by increased density.

It is assumed that in the absence of mass mortality normal or moderately
increased mortality rates can produce striking declines in density when
reproduction is partially or wholly suppressed. The bulk of evidence in the
literature indicates that this is more usually the case than mass mortality.
Furthermore, food shortages, epidemics, or predation may be the immediate
cause of the mortality, although it is equally clear that the mortality still
occurs when these factors are absent (Elton, 1942; Rausch, 1950; Chitty,
1954; Godfrey, 1955; Pitelka, 1957a, b). It is also possible that some pre-
cipitating factors may act through the physiologic adaptive mechanisms by
increasing intraspecific strife, as was shown to be the case in experimental
populations of voles (Frank, 1953).

4. Summary: Natural Populations

There is considerable evidence to indicate that there is a density-de-
pendent increase in the activity of physiologic adaptive mechanisms in
natural populations of Norway rats, voles {Microtus, Clethrionomys, Dicro-
stonijx, Lemmus, Ondatra), hares (Leptis), rabbits {Onjctolagus) , white-
footed mice (Peromijscus) , and probably other species. A relationship be-
tween population density and adrenocortical activity has been demon-
strated for Norway rats, meadow voles {Microtus niontanus and M.
pennsylvanicus) , white-footed mice (Peromyscus leucopus), and sika deer
(Cervus nippon). An inverse relationship between density and reproductive
function exists for Norway rats, red-backed voles (Clethrionomys) , meadow
voles {Microtus agrestis, M. montanus, M. californicus, M. pennsylvanicus),
muskrats {Ondatra), cotton rats {Sigmodon), and hares (Lepus). However,
in few of these studied were reproductive function and adrenocortical func-
tion correlated in the same population. Furthermore adequate assessment
of other envu-onmental factors has not been made. In many investigations
there were no adequate means of determining age, so that a true indication
of the degree of inhibition of reproductive function was not possible, espe-
cially with a simultaneous suppression of growth. Determinations of
age should be possible in a variety of species if a concerted effort is made
to find the proper criteria. There is evidence that, as in the laboratory,

324 /. /. Christian

there is suppression of lactation, as well as of other reproductive functions,
with increased density which may account for the prolonged effects of high
density on the young. However, the effects of high density may also be
exerted on the young animals directly. There is also evidence that resistance
to disease and parasites is decreased, especially m the young, at high popula-
tion densities. In extreme cases, there may be mass mortalities of popula-
tions due to hypoglycemic shock, presumably following exhaustion of some
part of the entire physiologic mechanism or of the reserves of readily
available necessary metabolites or both.

None of the data from natural populations is conclusive evidence that
density-dependent responses of physiologic adaptive mechanisms are active
in the regulation and control of the growth of mammalian populations.
However, when the available evidence is viewed in the light of the results
obtained from carefully controlled studies on laboratory populations, there
is sufficient reason to conclude that these mechanisms are operative in
natural mammalian populations, and more precisely in populations of
rodents and lagomorphs. It is equally apparent that sociopsychologic pres-
sures are the stimuli to the various physiologic responses involved. There-
fore, there is a wide range in responses in the individuals of a population,
the subordinate, and therefore younger, animals being the most severely
involved. Finally, it appears that these sociopsychologic-physiologic mecha-
nisms are basic controlling factors in populations of mammals and other
factors are secondarily important in most cases.

There is a great need for a coordinated study on a small rodent for which
there are criteria for determming age in which adrenocortical function, re-
productive function, resistance to disease, mortality, social factors, and
enviroilmental factors, can be adequately studied for a complete cycle of
the population. Until the results of such a study are available, conclusions
regarding density-dependent physiologic mechanisms must lean heavily on
extrapolations from data gathered from experimental populations in the
laboratory. There are in addition to the general problem many specific
problems that need investigation. Many of these have been mentioned m
the course of the discussion, but a few have not. A major problem is what
factors determine where the reproductive process will be affected in a
particular population. For example, why is it intra-uterine mortality in one
case, total inhibition of reproductive function in another, and depressed
lactation in a third? The explanations are largely conjectural at present,
but that these differences exist cannot be questioned, yet all represent a
common mechanism and effect: inhibition of reproduction. The answer to
another question may provide the answers to those already asked : how do
social pressures affect the individual members of a population?

2. Endocrines and Populations 325

III. Conclusion

The first part of this chapter reviewed briefly the endocrine adaptive
mechanisms which are of unquestionable importance in enabling mammals
to meet and adapt to their constantly changing environment and the visic-
situdes of daily existence. The mechanisms discussed are by no means the
only adaptive responses which are evoked by environmental exigencies,
but they are the better known ones and at the present time those which are
most Hkely to be of interest to mammalogists, ecologists, and students of
mammalian population dynamics. The first part of the chapter was in itself
a summary review of these responses; so no useful purpose would be served
by further condensing and .summarizing at this point. However, it should
be emphasized that one is not justified in regarding any single response, for
example increased adrenocortical activity, as an isolated phenomenon
complete unto itself and independent of any other changes. Any adaptive
response initiates and is a part of an extremely complex series of physio-
logic changes which probably involve every aspect of the host's physiology
and metabolism. Furthermore, the adaptive responses are not static affairs,
but a system of dynamic changes in the nature of feedback mechanisms
which operate to maintain a constant internal environment and life. As
stated earlier, these mechanisms are undoubtedly constantly active and
responding to even minor changes encountered in daily life. Emphasis is
lent to this statement by the hypoactive physiologic state observed in
closely confined, inactive, and overfed mammals. Therefore one should
regard the adaptive responses that are customarily studied and reported
merely as quantitative deviations from "normal" daily experience. Finally,
we should regard physiologic adaptations as flexible in that the available
evidence indicates that qualitatively different stimuli probably do not
elicit qualitatively similar responses even though cursory examination may
seem to contradict this statement. The components of the responses prob-
ably differ proportionally from each other with differing stimuli. It is true
that certain organ systems are primarily responsive to the demands of
external change, but their responses appear not to be identical, either
quantitatively or qualitatively, to all stimuli. For example, we have seen
that heat and cold elicit quite different responses, while emotional stimuli
evoke a third and different set of reactions, although all three may have
certain features in common. Therefore a plea is again made to examine
adaptive responses in the light of the specific stimuli by which they were
elicited. Generalization serves an extremely useful purpose conceptually,
but it may be misleading in the interpretation of results from critical
studies designed to explore physiologic adjustments to specific stimuli.

326 J- f- Christiau

iMiuill}', what has been presented here only represents a selected stopping
point in a very rapidly advancing field of research and must, perforce, be
used as such and modified in the light of newer developments.

The second part of this chapter attempted to review critically the evi-
dence from the laboratory and from the field relating physiologic adaptive
responses to changes in population density, largely in response to poorly
defined sociopsychologic factors. There seems to be little doubt that endo-
crine adaptive responses to sociopsychologic pressures are of basic import-
ance in the regulation of population growth, at least for a limited number
of species. It also is obvious that a great many gaps remain to be filled
before the evidence for the physiologic regulation of natural populations
can be considered conclusive one way or the other except for a few species
of rodent. These gaps have been pomted out in the appropriate portions
of the foregoing discussions. Nevertheless, considerable support has been
derived in the last ten years for the original hypothesis that factors inti-
mately related to population density are stimuli to physiologic feedback
mechanisms and that population growth and decline are largely controlled
by changes in density. It has also become evident that "density" in terms
of mammalian populations is related only indirectly to numbers of mammals
per unit of area, being more directly related to intraspecific competition,
social strife, sociopsychologic pressures, or whatever other comparable
designation one may choose to use for the interactions between mammals
in the same population. It is obvious that these factors require a great deal
more study in order to define them precisely.

Figure 1 is presented as a schematic summary of the physiologic feedback
n^gulation of population growth as it is envisioned today in the light of the
available experimental evidence. It should be re-emphasized that this is a
dynamic and flexible system and that the importance of various compo-
nents may vary with respect to each other. The broad ascending arrows on
the left of the figure indicate that as population size increases, social pres-
sures increase accordingly. The dotted lines above and below the arrow for
population increase serve to indicate that the actual population size may
vary considerably from population to population for a given degree of social
pressure, which we see hypothetically as the fundamental growth-regulating
and growth-limiting factor in all mammalian populations and the factor to
which physiologic adaptive mechanisms respond. However, that this factor
is always present and always operative does not necessarily imply that it is
always the factor which limits population growth. Environmental factors
probably operate through this mechanism by increasing or decreasing social
pressures, although conclusive evidence for this statement presently is
lacking. There is good e^'idence, howe\^er, that the social factors will limit
population growth despite an abundance of all environmental requirements
and that the degree of social density is related to the behavioral composition

2. Endocrines and Populations

327

i lie

O * E <

Itul

w O"

; =

<= 2 -

• c £

<
o

^ c ^

< -o -

328 /• J- Christian

(A the population. As social pressure increases, it acts as a progressively
stronger stimulus to physiologic adaptation via the mechanisms on the
right of the broken vertical line. These in turn act to decrease reproduction
and increase mortality, and therefore the rate of population growth declines
with increasing density and eventually ceases, as indicated by the"reproduc-
tion" and "mortality" arrows on the left. We have already seen that repro-
duction may cease altogether at maximmii densities and that mortality
may achieve precipitous proportions. An alteration in age composition of
the population is assumed to be inherent m these effects. The details of the
various mechanisms have been discussed earlier, and will not be discussed
further here. However, it is hoped that this diagram will help in visualizing
the dynamic mterrelationships between the endocrine responses in the
members of the population to changes in the population as a whole. It will
be observed that a food deficiency has been shown as decreasing directly
the production of gonadotropins, but, as we have discussed earlier, it may
serve to increase competition and therefore effective density under some
circumstances. Theoretically it seems unlikely that a food deficiency would
reach proportions in natural populations which would enable it to exert its
effects dhectly on the pituitary without competitive factors first becoming
operative. This diagrammatic hypothesis appears to fit the facts available
for a Imiited number of studies on a few species of rodent, and undoubtedly
it will be necessary to modify it as information from more critical studies
become available. It is of primary importance to obtain results from careful
studies on ungulates, similar to those already conducted on rodents, before
regulation of population growth of ungulates by sociopsychologic-physio-
logic feedback mechanisms can be considered to be more than hypothetical.
However, at present it is very probable that these mechanisms are basically
important in regulating the growth of populations of rodents and lago-
morphs.

An attempt has been made ui this chapter to point out the areas where
critical data are lackuig and some of the problems and pitfalls which may
be encountered in investigating these physiologic phenomena in relation to
population density. It is hoped that it will serve as a useful guide for further
research by investigators who wish to explore the mechanics of population
dynamics.

References

Ackerman, N. B., and Arons, W. L. (1958). The effect of epinephrine and norepinephrine
on the acute thyroid release of thyroid hormones. Endocrinology 62, 723-737.

Agate, F. J., Jr. (1952). The possible role of the adrenal cortex in fluid shifts in the ham-
ster following renal damage. Ann. N. Y. Acad. Sci. 55, 404-411.

Albert, A., Tenney, A., and Ford, E. (1952). The effect of cortisone and corticotropin on
the biologic decay of thyroidal radioiodine. Endocrinology 50, 324-326.

2. Endocrines and Populations 329

Allen, J. C. (1954). Effect of cortisone on the morphology of the mouse adrenal with

special reference to changes in the X-zone. Bull. Johns Hopkins Hasp. 94, 349-350.
Allen, J. M. (1957). A chemical and histochemical studj' of aliesterase in mouse adrenal

cortex under hormonal influences and altered sodium intake. Endocrinology 61, 368-

374.
Alpert, M. (1950). Observations on the histophysiology of the adrenal gland of the

golden hamster. Endocrinologij 46, 166-167.
Anderson, C. H., McCally, M., and Farrell, G. L. (1959). The effects of atrial stretch on

aldosterone secretion. Endocrinology 64, 202-207.
Anderson, E., and Haymaker, W. (1948a). Influence of the hypothalamus on sexual

functions, Part I. J. Am. Med. Women's Assoc. 3, 402-406.
Anderson, E., and Haymaker, W. (1948b). Influence of the hypothalamus on sexual

functions, Part II. /. Am. Med. Women's Assoc. 3, 457-461.
Andervont, H. B. (1944). Influence of environment on mammary cancer in mice. /.

Natl. Cancer Inst. 10, 579-581.
Antopol, W. (1953). Anatomical changes with excessive doses of cortisone in the normal

mouse, the pregnant mouse, the fetus, and the offspring. Ain. J. Pathol. 29, 612-613.
Ashng, C. W., Reinhardt, W. O., and Li, C. H. (1951). Effects of adrenocorticotropic

hormone on body growth, visceral proportions and white blood cell counts of normal

and hypophysectomized male rate. Endocrinology 48, 534-547.
Aycock, W. L., and Ingalls, T. H. (1946). Maternal disease as a principle in the epido-

miology of congenital anomalies. Am. J. Med. Sci. 212, 366-379.
Ayres, P. J., Qould, R. P., Simpson, S. A. S., and Tait, J. F. (1956). The in vitro demon-
stration of differential corticosteroid production within the ox adrenal gland.

Biochem. J. 63, 19p.
Bachman, R. (1954). Die nebenniere. In "Handbuch der mikroskopischen Anatomic

des Menschen" (W. v. Mollendorft', ed.). Vol. 6. Springer, Berlin.
Badrick, F. E., Brimblecombe, R. W., Reiss, J. M., and Reiss, M. (1954). The influence

of stress conditions on the uptake of V^^ by the rat thyroid. /. Endocrinol. 11,

305-313.
Badrick, F. E., Brimblecombe, R. W., and Reiss, M. (1955). Responses of hypophy-

sectomized rats to stress. J. Endocrinol. 12, 205-208.
Bahn, R., Wilson, H., Kedda, L., Furth, J., and Anderson, E. (1957). Metabohc changes

associated with transplantable ACTH-producing pituitary tumors of the mouse.

Can. J. Biochem. and Physiol. 35, 1057-1065.
Baker, B. L. (1949). Modification of body stmcture by adrenocortical secretions with

special reference to the regulation of growth. Am. Assoc. Advance. Sci. Symposium

on Pituitary- Adrenal Function, pp. 88-95.
Baker, B. L. (1952). A comparison of the histological changes induced by experimental

hyperadrenacorticalism and inanition. Recent Progr. in Hormone Research 7, 331-373.
Baker, B. L. (1954). The connective tissue reaction around implanted pellets of steroid

hormones. Anat. Record 119, 529-539.
Baker, B. L., and Abrams, G. D. (1954). Effect of hypophysectomy on the cytology of

the fundii glands of the stomach and on the secretion of pepsin. Am. J. Physiol.

177, 409-412.
Baker, B. L., and Bridgman, R. M. (1954). The histology of the gastro-intestinal mucosa

(rat) after adrenalectomy or administration of adrenocortical hormones. Am. J.

Anat. 94, 363-397.
Baker, B. L., Schairer, M. A., Ingle, D. J., and Li, C. H. (1950). The. induction of in-
volution in the male reproductive system by treatment with adrenocorticotropin.

Anat. Record 106, 345-359.

330 /. /. Christian

Baker, B. L., Ingle, 1). J., and Li, C. H. (1951). The histology of the lymphoid organs of

rats treated with adrenocorticotropin. Am. J. Anal. 88, 313-349.
Baker, J. R., and Ransom, R. M. (1932). Factors affecting breeding of the field mouse

{Microtus agresiis). Part II. Temperature and food. Pi-oc. Roy. Soc. B112, 39-46.
Barkalow, F. S., Jr. (1952). Life history and ecologic observations on the tundra mouse

{Microtus oeconomus) and lemmings (Lemmus trimucronatus and Dicrostonyx groen-

landicus) at Barter Island, Alaska. J. Elisha Mitchell Sci. Soc. 68, 199-205.
Barker, S. B., and Schwartz, H. S. (1953). Further studies on metabolism of tissues

from thyroxine-injected rats. Proc. Soc. Exptl. Biol. Med. 83, 500-502.
Barnett, S. A. (1955). Competition among wild rats. Nature 175, 126-127.
Barnett, S. A. (1958). Physiological effects of "social stress" in wild rats. I. The adrenal

corte.x. J . Psychosoniat. Research 3, 1-11.
Barnett, S. A., and Manly, B. M. (1956). Reproduction and growth of mice of three

strains, after transfer to -3°C. J. Exptl. Biol. 33, 325-329.
Barraclough, C. A. (1961). Production of anovulatory, sterile rats by single injections of

testosterone propionate. Endocrinology 68, 62-67.
Barraclough, C. A., and Gorski, R. A. (1961). Evidence that the hypothalamus is respon-
sible for androgen-induced Sterility in the female rat. Endocrinology 68, 68-79.
Barry, M. C, Eidinoff, M. L., Dobriner, K., and Gallagher, T. F. (1952). The fate of Ci^-

progesterone in mice and rats. Endocrinology 50, 587-599.
Bartter, F. C. (1956). The role of aldosterone in normal homeostatsis and in certain

disease states. Metabolism Clin, and Exptl. 5, 369-383.
Bartter, F. C. (1957). The role of aldosterone in the regulation of body fluid volume and

composition. Scand. J. Clin. & Lab. Invest. 10, 50-61.
Bartter, F. C., and Gann, P. S. (1960). On the hemodynamic regulation of the secretion

of aldosterone. Circulation 21, 1016-1023.
Bartter, F. C., Liddle, G. W., Duncan, L. D., Jr., Barber, J. K., and Delea, C. (1956).

The regulation of aldosterone secretion in man: The role of fluid volume. /. Clin.

Invest. 35, 1306-1315.
Bartter, F. C., Biglieri, E. G., Pronove, P., and Delea, C. S. (1958). Effect of changes in

intravascular volume on aldosterone secretion in man. In "Aldosterone," Intern.

Symposium, Geneva, 1957 (A. F. Muller and C. M. O'Connor, eds.), pp. 100-110.

Little, Brown, Boston, Massachusetts.
Bartter, F. C, Mills, I. H., Biglieri, E. G., and Delea, C. S., (1959). Studies on the

control and physiologic action of aldosterone secretion. Recent Progr. in Hormone

Research, 15, 311-344.
Bastenie, P. A., and Ermans, A. jVI. (1958). Effect of cortisone on the action of thyroxine

in hypophysectomized rats. Endocrinology 62, 245-250.
Beer, J. R., and Mej-er, R. K. (1951). Seasonal changes in the endocrine organs and

behavior patterns of the muskrat. J. Mammal. 32, 173-191.
Bell, G. H., Davidson, J. N., and Scarborough, H. (1950). "Textbook of Physiology and

Biochemistry." Williams & Wilkins, Baltimore, Maryland.
Benua, R. S., and Howard, E. (1950). A carbonyl reaction differentiating the fetal zona

reticularis of the human adrenal cortex from the mouse X-zone. Bull. Johns Hopkins

Hasp. 86, 200-213.
Berson, S. A., and Yalow, R. S. (1952). The effect of cortisone on the iodine accumulating

function of the thyroid gland in euthj'roid subjects. J. Clin. Endocrinol, and Meta-
bolism 12, 407-422.
Beyler, A. L., and Szego, C. M. (1954). Steroid hormone interactions as evidenced by

modification of j8-glucuronidase activity of the preputial glands of the rat. Endo-
crinology 54, 334-337.

2. Endocrines and Populations 331

Bing, R. J. (1952). The role of coronary circulation in shock. Ann. N. Y. Acad. Set.
55, 367-369.

Birmingham, M. K., and Kurlents, E. (1958j. Inactivation of ACTH by isolated rat
adrenals and inhibition of corticoid formation bj- adrenocortical hormones. Endo-
crinology 62, 47-60.

Bloch, E., and Cohen, A. I. (1960). In vitro steroid production by normal and adrenal
tumor-bearing mice. J. Natl. Cancer Inst. 24, 97-107.

Boas, N. F., and Bates, R. W. (1954). Role of the thyroid and anterior pituitary glands
in the maintenance of the Harderian glands of the rat. Endocrinology 55, 601-612.

Bogdanove, E. M., and Halmi, N. S. (1953). Effect of hypothalamic lesions and sub-
sequent propylthiouracil treatment on pituitary structure in the rat. Endocrinology
53, 274-292.

Bogoroch, R., and Timiras, P. (1951). The response of the thyroid gland of the rat to
severe stress. Endocrinology 49, 548-556.

Bohanan, E. H. (1939). Effects of environmental factors on the length of estms cycles
in the rat. Am. J. Hyg. 29A, 1-10.

Bois, P., and Selj-e, H. (1957). Enhancement b}- thryoxine of the adrenal-stimulating
effect of somatotrophic hormone. Endocrinology 15, 171-179.

Boring, W. O., Angevine, D. M., and Walker, D. L. (1955). Factors influencing host-
virus interactions. I. A comparison of viral multiplication and histopathology in
infant, adult, and cortisone-treated adult mice infected with the Conn .-5 strain of
Coxsackie virus. /. Exptl. Med. 102, 753-766.

Boring, W. O., ZuRhein, G. M., and Walker, D. L. (1956). Factors influencing host-virus
interactions. II. Alterations of Coxsackie virus infection in adult mice bj' cold.
Proc. Soc. Exptl. Biol. Med. 93, 273-277.

Botkin, A. L., and Tew, J. T. (1952). Tlu'roid and adrenal respon.se to varying dose
levels of epinephrine. Proc. Soc. Exptl. Biol. Med. 79, 378-379.

Bourliere, F. (1951). "Vie et moeurs des mammiferes," pp. 1-250. Payot, Paris.

Bourne, G. H. (1949). "The Mammalian Adrenal Gland," pp. 1-239. Clarendon, Oxford.

Bradlow, H. L., and Gallagher, T. F. (1957). Metabolism of l]-/3-hydroxy-A^-androstene-
3,17-dione in man. J. Biol. Chem. 229, 505-518.

Bradlow, H. L., Dobriner, K., and Gallagher, T. F. (1954). The fate of cortisone-T in
mice. Endocrinology 54, 343-352.

Brimblecombe, R. W., Halkerston, I. D. K., and Reiss, AI. (1954). The effect of various
ACTH extracts on the gonads. /. Endocrinol. 10, 291-301.

Brown, R. Z. (1953). Social behavior, reproduction, and population changes in the
house mouse. Ecol. Monographs 23, 217-240.

Brown-Grant, K. (1955). Inhibition of the release of thyroidal radioiodine in the rat by
cortisone. Endocrinology 56, 607-609.

Brown-Grant, K., Von Euler, C, Harris, G. W., and Reichhn, S. (1954a). The measure-
ment and experimental modification of thyroid activitj- in the rabbit. /. Physiol.
{London) 126, 1-28.

Brown-Grant, K., Harris, G. W., and Reichlin, S. (1954b). The effect of emotional and
physical stress on thjToid activit}^ in the rabbit. J. Physiol. (London) 126, 29-40.

Brown-Grant, K., Harris, G. W., and Reichlin, S. (1954c). The influence of the adrenal
cortex on thyroid activity in the rabbit. J. Physiol. (London) 126, 41-51.

Brown-Grant, K., Harris, G. W., and Reichlin, S. (1957). The effect of pituitary-stalk
section on thyroid function in the rabbit. J. Physiol. [London) 136, 364-379.

Bruce, H. M., and Sloviter, H. A. (1957). Effect of destruction of thyroid tissue by
radioactive iodine on reproduction in mice. ./. Endocrinol. 15, 72-82.

Bull, P. C. (1955). Population regulation in rabbit nematodes. Natwe 175, 218.

332 /. /. Christian

Bull, P. C. (1957). Changing incidence of parasites in a declining population. Proc. New

Zealand Ecol. Soc. 5, 11-12.
Bull, P. C. (1958). Incidence of Coccidia (Sporozoa) in wild rabbits Orydolagus cuniculus

(L) in Hawke's Bay, New Zealand. New Zealand J. Set. 1, 289-328.
BuUough, W. S. (1952). Stress and epidermal mitotic activity. 1. The effects of the

adrenal hormones. /. Endocrinol. 8, 265-274.
BuUough, W. S. (1955). Hormones and mitotic activity. Vitamins and Hormones 13,

261-292.
Burn, J. H., and Rand, M. J. (1957). Reserpine and adrenaline in artery walls. Lancet

273, 1097.
Burn, J. H., and Rand, M. J. (1958a). Noradrenaline in artery walls and its dispersal.

Brit. Med. J. I, 903-908.
Burn, J. H., and Rand, M. J. (1958b). The action of sympathomimetic amines in animals

treated with reserpine. /. Physiol. (London) 144, 314-336.
Burrows, H. (1949). "Biological Actions of Sex Hormones," 2nd ed. Cambridge Univ.

Press, London and New York.
Bush, I. E. (1953). Species differences in adrenocortical secretion. /. Endocrinol. 9, 95-

100.
Byrnes, W. W., and Meyer, R. K. (1951). The inhibition of gonadotropic hormone

secretion by physiological doses of estrogen. Endocrinology 48, 133-136.
Byrnes, W. W., and Shipley, E. G. (1950). Estrogenic and gonadotrophic hormone

inhibiting activity of some adrenal cortical substances. Proc. Soc. Exptl. Biol. Med.

74, 308-310.
Cain, A. J., and Harrison, R. G. (1950). Cytological and histochemical variations in the

adrenal cortex of the albino rat. J. Anat. 84, Pt. 2, 196-226.
Calhoun, J. B. (1948). Mortality and movement of brown rats in artificially super-
saturated populations. /. Wildlife Management 12, 167-171.
Calhoun, J. B. (1949). A method for self-control of population growth among mammals

living in the wild. Science 109, 333-335.
Calhoun, J. B. (1950). The study of wild animals under controlled conditions. Ann. N . Y.

Acad. Sci. 51, 1113-1122.
Cannon, W. B. (1915). "Bodily Changes in Pain, Hunger, Fear and Rage." Appleton,

New York.
Cannon, W. B. (1932). "The Wisdom of the Body." Norton, New York.
Cater, O. B., and Stack-Dunne, M. P. (1953). The histological changes in the adrenal of

the hypophysectomized rat after treatment with pituitary preparations. J. Pathol.

Bacterial. 66, 119-133.
Cater, O. B., and Stack-Dunne, M. P. (1955). The efi'ects of growth hormone and

corticotrophin upon the adrenal weight and adrenocortical mitotic activity in the

hypophysectomized rat. J. Endocrinol. 12, 174-184.
Cavazos, L. F., and Melampy, R. M. (1954). Cytological effects of testosterone propio-
nate on epithelium of rat seminal vesicles. Endocrinology 54, 640-648.
Chang, M. C, and Pincus, G. (1951). Physiology of fertilization in mammals. Physiol.

Revs. 31, 1-26.
Chitty, D. (1952). MortaUty among voles (Microtus agrestis) at Lake Vj^mwy, Mont-
gomeryshire in 1936-9. Phil. Trans. Roy. Soc. London B236, 505-552.
Chitty, D. (1954). Tuberculosis among wild voles with a discussion of other pathological

conditions among certain mammals and birds. Ecology 35, 227-237.
Chitty, D. (1955). Adverse effects of population density upon the viability of later

generations. In "The Numbers of Man and Animals," pp. 57-67. Oliver & Boyd,

Edinburgh.

2. Endocrines and Populations 333

Chitty, D. (1957). Self-regulation of numbers through changes in viability. Cold Spring

Harbor Symposia Quant. Biol. 22, 277-280.
Chitty, H., and Austin, C. R. (1957). Environmental modification of oestrus in the vole.

Nature 179, 592-593.
Chitty, D., Chitty, H., Leslie, P. H., and Scott, J. C. (1956). Changes in the relative

size of the nucleus in the intervertebral disc of stressed orkney voles (Microtus

orcadensis). J. Pathol. Bacterial 72, 459-470.
Christian, J. J. (1950a). A field method of determining the reproduction status of small

male mammals. J . Mammal. 31, 95-96.
Christian, J. J. (1950b). The adreno-pituitary system and population cycles in mammals.

J. Mammal. 31, 247-259.
Christian, J. J. (1951). The mammals of the Mazinaw lake region of Ontario; their

reproduction and population dynamics. Ann. Carnegie Museum 31, 339-386.
Christian, J. J. (1954). The Relation of the Adrenal Cortex to Population Size in Rodents.

Doctoral dissertation, Johns Hopkins School of Hygiene and Public Health, Balti-
more, Maryland.
Christian, J. J. (1955aj. Effect of population size on the weights of the reproductive

organs of white mice. Am. J. Physiol. 181, 477-480.
Christian, J. J. (1955b). Effect of population size on the adrenal glands and reproductive

organs of male mice in populations of fixed size. Am. J. Physiol. 182, 292-300.
Christian, J. J. (1956). Adrenal and reproductive responses to popultaion size in mice

from freely growing populations. Ecology 37, 258-273.
Christian, J. J. (1959a). Control of population growth in rodents by interplay between

population density and endocrine physiology. Wildlife Disease 1, 1-38.
Christian, J. J. (1959b). The roles of endocrine and behavioral factors in the growth of

mammalian populations. In "Comparative Endocrinology," Proc. Columbia Uni-
versity SjTnposium, 1958 (A. Gorbman, ed.), pp. 71-97. Wiley, New York.
Christian, J. J. (1959c). Adrenocortical, splenic and reproductive responses to inanition

and to grouping. Endocrinology 65, 189-197.
Christian, J. J. (1959d). Lack of correlation between adrenal weight and injury from

fighting in grouped male albino mice. Proc. Soc. Exptl. Biol. Med. 101, 166-168.
Christian, J. J. (1960). Adrenocortical and gonadal responses of female mice to increased

population density. Proc. Soc. Exptl. Biol. Med. 104, 330-332.
Christian, J. J. (1961). Phenomena associated with population density. Proc. Natl.

Acad. Sci. U. S. 47, 428-449.
Christian J. J. (1963a). The pathology of overpopulation. Military Med. 128, 571-603.
Christian, J. J. (1963b). Interrelated social and endocrine factors in populations of

rodents. (Abstract.) Proc. 16th Intern. Congr. Zool., Washington, D. C, 3, 8-13.
Christian, J. J., and Davis, D. E. (1955). Reduction of adrenal weight in rodents by

reducing population size. Trans. N. Am. Wildlife Conf. 20, 177-189.
Christian, J. J., and Davis, D. E. (1956). The relationship between adrenal weight and

population status in Norway rats. J. Mammal. 37, 475-486.
Christian, J. J., Flyger, V., and Davis, D. E. (1960). Factors in mass mortality of a herd

of sika deer. Chesapeake Sci. 1, 79-95.
Christian, J. J., and LeMunyan, C. D. (1958). Adverse effects of crowding on lactation

and reproduction of mice and two generations of their offspring. Endocrinology 63,

517-529.
Christian, J. J., and Ratclifte, H. L. (1952). Shock disease in captive wild mammals.

Am. J. Pathol. 28, 725-737.
Christian, J. J., and Williamson, H. O. (1958). Effect of crowding on experimental

granuloma formation in mice. Proc. Soc. Exptl. Biol. Med. 99, 385-387.

334 J. J. Christian

Christianson, M., and Jones, I. C. (1957). The interrelationships of the adrenal glands of

mother and fetus in the rat. /. Endocrinol. 15, 17-42.
Chu, J. P. (1944). The influence of the thyroid on pregnancy and parturition in the

rabbit. /. Endocrinol. 4, 109-114.
Clarke, J. R. (1953). The effect of fighting on the adrenals, thymus, and spleen of the

vole {Microtus agrestis). J. Endocrinol. 9, 114-126.
Clarke, J. R. (1955). Influence of numbers on reproduction and sui-vival in two experi-
mental vole populations. Proc. Roy. Soc. B144, 6S-S5.
Clayton, G. W., Bell, W. R., and Guillemin, R. (1957). Stimulation of ACTH-release in

humans by non-pressor fraction from commercial extracts of posterior pituitarj'.

Proc. Soc. Exptl. Biol. Med. 96, 777-779.
Cole, D. F. (1957). Effect of aldosterone on renal excretion of intravenously administered

saline. Endocrinology 60, 562-565.
Cook, E. F., and Beer, J. R. (1958). A study of louse populations on the meadow vole

and deer mouse. Ecology 39, 645-659.
Crew, F. A., and Mirskaia, L. (1931). The effects of density on an adult mouse popu-
lation. Biol. Generalis 7, 239-250.
Critchlow, B. \., and Sawjer, C. H. (1955). Electrical activity of the rat brain cor-
related with neurogenic stimulation of the adenohypophysis. Federation Proc. 14,

32-33.
Crowcroft, P. (1954). Aggressive behavior in wild house mice {Mus musculus L.)

Ministrj' of Agr. and Fish Infestation Control Division, Research Rept. No. 35.
Crowcroft, P. and Rowe, F. P. (1957). The growth of confined colonies of the wild house

mouse {Mus vnisculus L.). Proc. Zool. Soc. (London) 129, 359-370.
Curley, F. J., and Ingalls, T. H. (1957). HjiJoxia at normal atmospheric pressure as a

cause of congenital malformations in mice. Proc. Soc. Exptl. Biol. Med. 94, 87-88.
D'Angelo, S. A. (1955). Pituitary regulation of thyroid gland function. Brookhaven

Symposia in Biol. 7, 9-29.
D'Angelo, S. A., and Traum, R. E. (1958). An experimental analysis of the hypothalamic

hypophyseal-thyroid relationship in the rat. Ann. N. Y. Acad. Sci. 72, 239-270.
D'Angelo, S. A., Gordon, A. S., and Charipper, H. A. (1948). A differential response of

the rodent adrenal gland to acute starvation. Proc. Soc. Exptl. Biol. Med. 68, 527-

531.
Davidson, C. S. (1937). Effect of adrenotropic extract upon the accessory reproductive

organs of castrated rats. Proc. Soc. Exptl. Biol. Med. 36, 703-705.
Davidson, C. S., and Moon, H. D. (1936). Effect of adrenocorticotropic extracts on

accessory reproductive organs in castrate rats. Proc. Soc. Exptl. Biol. Med. 35,

281-282.
Davis, D. E. (1951a). The relation between level of population and pregnancy of Noiwaj-

rate. Ecology 32, 459-461.
Davis, D. E. (1951b). The relation between level of population and size and sex of

Norway rats. Ecology 32, 462-464.
Davis, D. E. (1951c). The relation between the level of population and the prevalence of

Leptospira, Salmonella, and capillaria in Norway rats. Ecology 32, 465-468.
Davis, D. E. (1953). The characteristics of rat populations. Quart. Rev. Biol. 28, 373-401.
Davis, D. E., and Christian, J. J. (1957). Relations of adrenal weight to social rank of

mice. Proc. Soc. Exptl. Biol. Med. 94, 728-731.
Davis, D. E., and Christian, J. J. (1958). Population consequences of a sustained j-ield

program for Norwaj' rats. Ecology 39, 217-222.
Davis, D. E., and Jensen, W. L. (1952). Mortality in an induced epidemic. Trans. N.

Am. Wildlife Conf. 17, 151-160.

2. Endocrines and Populations 335

Davis, D. E., and Prudovsky, S. (.1959). Effect of behavior on d<'veloi)ment of resistance

to tetanus toxin. Unpublished.
Davis, D. E., and Read, C. P. (1958). Effect of behavior on development of resistance in

trichinosis. Proc. Soc. Exptl. Biol. Med. 99, 269-272.
Davis, J. O. (1963) The role of the adrenal corte.x and the kidney in the pathogenesis of

cardiac edema. Yale J. Biol. Med. 35, 402-428.
Davis, J. O., Goodkind, M. J., Pechet, M. M., and Ball, W. C, Jr. (1956). Increased

excretion of aldosterone in urine from dogs with right-sided congestive heart failure

and from dogs with thoracic inferior vena cava constriction. Am. J. Phyniol. 187,

45-50.
Davis, J. O., Pechet, M. M., Ball, W. C, Jr., and Goodkind, M. J. (1957). Increased

aldosterone secretion in dogs with rightsided congestive heart failure and in dogs

with thoracic inferior vena cava constriction. /. Clin. Invest. 36, 689.
Davis, J. O., Kliman, B., Yankopoulos, N. A., Peterson, R. E., and Casper, A. (1958).

Increased aldosterone secretion following acute constriction of the inferior vena

cava. /. Clin. Invest. 37, 1783-1790.
Davis, J. O., Bahn, R. C, Yankopoulos, N. A., KHman, B., and Peterson, R. E. (19.59a).

Acute effects of hypophysectomy and diencephalic lesions on aldosterone secretion.

Am. J. Physiol. 197, 380-386.
Davis, J. O., Yankopoulos, N. A., and Holraan, J. (1959b). Chronic effects of carotid

sinus denervation cervical vagotomy and aortic depressor nerve section on aldo-

.sterone and sodium excretion. Am. J. Physiol. 197, 207-210.
Davis, M. E., and Plotz, E. J. (1954). The effects of cortisone acetate on intact and

adrenalectomized rats during pregnancy. Endocrinology 54, 384-395.
Dawson, J. (1956). Splenic hj-pertrophy in voles. Nature 178, 1183-1184.
Deane, H. W., and Morse, A. (1948). The cytological distribution of ascorbic acid in the

adrenal cortex of the rat under normal and experimental conditions. Anat. Record

100, 127-141.
Deane, H. W., and Seligman, A. M. (1953). Evaluation of procedures for the cyto-
logical localization of ketosteroids. Vitamins and Hormones 11, 173-204.
Deane, H. W., Shaw, J. H., and Creep, R. O. (1948). The effect of altered sodium of

potassium intake on the width and cytochemistry of the zona glomerulosa of the

rat's adrenal cortes. Endocrinology 43, 133-153.
Deanesly, R. (1928). A study of the adrenal cortex in the mouse and its relation to the

gonads. Proc. Roy. Soc. B103, 523-546.
DeBodo, R. C, and Altzuler, N. (1958). Insulin hypersensitivity and physiological

insulin antagonists. Physiol. Revs. 38, 389-445.
Delost, P. (1951). Structure histologiciue et etude I'activite phosphatasique alcahne du

canal deferent du campagnol des champs (Microtus arvalis, P.) etat du cortex

surrenal. Compt. rend. soc. biol. 145, 373-377.
Delost, P. (1952a). Le cortex surrenal du campagnol des champs {Microtus arvalis, P.)

et ses modifications aprcs castration. Compt. rend. soc. biol. 146, 27-31.
Delost, P. (1952b). Modifications histochimiques du cortex surrenal du campagnol des

champs {Microtus arvalis, P.) apres castration. Compt. rend. soc. biol. 146, 1080-1083.
Delost, P. (1954). Action involutive de la cortisone sur la zone X surrenalienne du cam-
pagnol des champs {Microtus arvalis. P.). Compt. rend. soc. biol. 148, 1161-1163.
Delost, P. (1955). Role de I'ovaire et du cortex surrenal dans la stimulation de la prostate

femelle chez les rongeurs sauvages. Compt. rend. soc. biol. 149, 1336-1339.
Delost, P. (1956a). Les correlations genito-surrenaliennes chez le campagnol des champs

{Microtus arvalis, P.). Ann. Sci. nat. Zool. 11. 18, 392-567.

336 J. J. Christian

Delost, P. (1956b). Donnees experimentales comparees sur la zone X surrenalienne.

Action de la cortisone et de la testosterone. Compt. rend. soc. biol. 150, 1330-13332.
Delost, P. (1957). Recherches sur le cortex surrenal et la zone X dcs musaraignes {Sorex

araneus L. et Crocidura russida H.). Compt. rend. soc. biol. 151, 484-487.
Delost, P., and Delost, H. (1954). Existence d'une zone X surrenalienne chez le cam-

pagnol roussatre {Clethrionomys glareolus, S.). Compt. rend. soc. biol. 148, 1788-1793.
Delost, P., and Delost, H. (1955). Modifications ponderales et histologiques des glandes

surrenales du campagnol agreste {Microtus agrestis, L.) en fonction de I'etat sexual.

Compt. rend. soc. biol. 149, 910-914.
Dempsey, E. W. (1948). The chemical cytolog}- of endocrine glands. Recent Progr. in

Hormone Research 3, 127-157.
Dempsey, E. W., and Uotila, U. U. (1940). The effect of pituitary stalk section upon

reproductive phenomena in the female rat. Endocrinology 27, 573-579.
Denison, M. E., and Zarrow, M. X. (1955). Changes in estrous cycle of rat during pro-
longed exposure to cold. Proc. Soc. Exptl. Biol. Med. 89, 632-634.
Desclin, J. (1959). Masculinisation spontanee et experimentale des glandes sous-maxil-

laires chez la souris femelle. Compt. rend. acad. sci. 248, 597-600.
Dorfman, A. (1953). The effects of adrenal hormones on connective tissues. Ann. N. Y.

Acad. Sci. 56, 698-703.
Dorfman, R. I. (1949). The bioassay of adrenocortical steroids. Ann. N. Y. Acad. Sci.

50, 556-575.
Dorfman, R. I. (1959). Comparative biochemistry of the adrenocortical hormones. In

"Comparative Endocrinology," Proc. Columbia University Symposium, 1958

(A. Gorbman, ed.), pp. 613-623. Wiley, New York.
Dorfman, R. I. (1960). Steroid metabolism in endocrine tumors. In "Biological Activ-
ities of Steroids in Relation to Cancer" (G. Pincus and E. P. VoUmer, eds.), pp. 93-

110. Academic Press, New York.
Dorfman, R. I., and Shipley, R. A. (1956). "Androgens. Biochemistry Physiology and

Clinical Significance." Wiley, New York.
Dougherty, T. F. (1952a). Studies of the antiphlogistic and antibody suppressing

functions of the pituitary adrenocortical secretions. Recent Progr. in Hormone

Research 7, 307-330.
Dougherty, T. F. (1952b). Effect of hormones on Ij-mphatic tissue. Physiol. Revs. 32,

379-401.
Dougherty, T. F. (1953). Some observations on mechanisms of corticosteroid action on

inflamnation and immunologic processes. Ann. N. Y. Acad. Sci. 56, 748-756.
Dougherty, T. F., and Schneebeli, G. L. (1955). The use of steroids as anti-inflammatory

agents. Ann. N. Y. Acad. Sci. 61, 328-348.
Drager, G. A. (1950). Neurosecretion following hypophysectomy. Proc. Soc. Exptl.

Biol. Med. 75, 712-713.
Ehrich, W. E., and Seifter, J. (1948). Role played by salivary glands in the alarm re-
reaction. A. M. A. Arch. Pathol. 45, 239-245.
Eik-nes, K. (1959). Comments on a paper by J. W. Mason. Recent Progr. in Hormone

Research 15, 380-381.
Eisenstein, A. B. (1957). Pantothenic acid and adrenocortical hormone secretion.

Endocrinology 60, 298-302.
Eisenstein, A. B. (1959). Relationship of vitamin Be to adrenocortical function in the

rat. Proc. Soc. Exptl. Biol. Med. 100, 111-114.
Eisenstein, A. B., and Hartroft, P. M. (1957). Alterations in the rat adrenal cortex

induced by sodium deficiency: Steroid hormone secretion. Endocrinology 60, 634-640.

2. Endocrines and Populations 337

Elias, H. (1948). Growth of the adrenal cortex in domesticated Ungulata. Am. J. Vet.

Research 9, 173-189.
Ehas, H., and Pauly, J. E. (1956). The structure of the human adrenal cortex. Endo-
crinology 58, 714-738.
Elmadjian, F., Hope, J. M., and Lamson, E. T. (1958). Excretion of epinephrine and

norepinephrine under stress. Recent Progr. in Hormone Research 14, 513-553.
Elton, C. (1942). "Voles, Mice and Lemmings." Clarendon, Oxford.
Engel, F. L. (1952). The significance of the metabolic changes during shock. Ann. N. Y.

Acad. Sci. 55, 381-393.
Eranko, O., and Hopsu, V. (1958). Effect of reserpine on the histochemistry and content

of adrenaline and noradrenaline in the adrenal medulla of the rat and the mouse.

Endocrinology 62, 15-23.
Eranko, O., and Raisanen, L. (1957). Adrenaline and noradrenaline in the adrenal

medulla during postnatal development of the rat. Endocrinology 60, 753-760.
Eranko, O., Hopsu, V., and Raisanen, L. (1959). The effect of prolonged administrations

of nicotine on the medullary volume and the distribution of noradrenaline in the

adrenals of the rat, the mouse, and the guinea pig. Endocrinology 65, 293-297.
Erickson, A. B. (1944). Helminth infections in relation to population fluctuations in

snowshoe hares. J. Wildlife Management 8, 134-153.
Errington, P. L. (1943). An analysis of mink predation upon muskrats in North-central

United States. Iowa State College Agr. Expt. Sta. Bidl. No. 320.
Errington, P. L. (1948). Ecology of the muskrat. Report on Agr. Research for year-ending

June 30, 1948. Iowa State College Agr. Expt. Sta. pp. 230-232.
Errington, P. L. (1951). Concerning fluctuations in populations of the prolific and

widely distributed muskrat. Am. Naturalist 85, 273-292.
Errington, P. L. (1954a). On the hazards of overemphasizing numerical fluctuations in

studies of "cyclic" phenomena in muskrat populations. /. Wildlife Management 18,

66-90.
Errington, P. L. (1954b). The special responsiveness of minks to epizootics in muskrat

populations. Ecol. Monographs 24, 377-393.
Errington, P. L. (1957). Of populations, cycles and unknowns. Cold Spring Harbor

Symposia Quant. Biol. 22, 287-300.
Ershoff, B. H. (1948). Effects of thiouracil feeding on resistance to low environmental

temperature. Endocrinology 43, 36-39.
Evans, E. S., Simpson, M. E., and Evans, H. M. (1958). The role of growth hormone in

calorigenesis and thjToid function. Endocrinology 63, 836-852.
Everett, J. W. (1956). Functional corpora lutea maintained for months by autografts of

rat hypophyses. Endocrinology 58, 786-796.
Everett, J. W. (1959). Neuroendocrine mechanisms in control of the mammalian ovary.

In "Comparative Endocrinology Proc. Columbia University Symposium," 1958

(A. Gorbman, ed.), pp. 74-186. Wiley, New York.
Everett, J. W., and Sawyer, C. H. (1953). Estimated duration of the spontaneous

activation which causes release of ovulation hormone from the rat hypophysis.

Endocrinology 52, 83-92.
Farrell, G. L. (1958). Regulation of aldosterone secretion. Physiol. Revs. 38, 709-728.
Farrell, G. L. (1959a). The physiological factors which influence the secretion of aldo-'

sterone. Recent Progr. in Hormone Research 15, 275-310.
Farrell, G. L. (1959b). Steroidogenic properties of extracts of beef diencephalon. Endo-
crinology 65, 29-33.

338 J. J. Christian

Farrell, G. L., and Laqueur, G. (1955). Reduction of pituitary content of ACTH by
cortisone. Endocrinology 56, 471-473.

Farrell, G. L., Rauschkolb, E. W., and Royce, P. C. (1955). Secretion of aldosterone by
the adrenal of the dog. Effects of hypophysectomy and ACTH. Am. J. Physiol. 182,
269-272.

Farrell, G. L., Fleming, R. B., Rauschkolb, E. W., Yat.su, F. M., McCally, M., and
Anderson, C. H. (1958). Steroidogenic properties of purified corticotropins. Endo-
crinology 62, 506-512.

Feldman, J. D. (1950). Histochemical reactions of adrenal cortical cells. Anat. Record
107, 347-358.

Feldman, J. D. (1951). Endocrine control of the adrenal gland. Anat. Record 109, 41-69.

Fitch, H. S. (1957). Aspects of reproduction and develoi)ment in the prairie vole {Micro-
tus ochrogaster) . Univ. Kansas Museum Nat. Hist. Publ. 10, 129-161.

Flexner, L. B., and Grollman, A. (1939). The reduction of osmic acid as an indicator of
adrenal cortical activity in the rat. Anat. Record 75, 207-221.

Flick, S., Davis, D. E., and Christian, J. J. (1959). Effect of constant length of day on
reproduction in albino mice. Brit. J. Hyg. 57, 193-197.

Fliickiger, E., and VerZc4r, F. (1955). Thyroid function after adrenalectomy. J. En-
docrinol. 13, 39-45.

Folley, S. J. (1956). "The Physiology and Biochemistry of Lactation." C. C. Thomas,
Springfield, Illinois.

Fortier, C. (1956). Hypothalamic control of the adrencorticotrophic function SOth
Intern. Congr. Physiol., Brussels, Abstr. of Revs. 1-18.

Fortier, C. (1957). Neuroendocrine relationships. Progr. in Kexirol. and Psychiat. 12,
252-261.

Fortier, C. (1959a). Pituitary ACTH and plasma free corticosteroids following bilateral
adrenalectomy in the rat. Proc. Sac. Exptl. Biol. Med. 100, 13-16.

Fortier, C. (1959b). Effect of hydrocortisone on pituitary ACTH and adrenal weight in
the rat. Proc. Sac. Exptl. Biol. Med. 100, 16-19.

Frank, F. (1953). Untersuchungen liber den Zusammenbruch von Feldmausplagen
{Microtus arvalis Pallas). Zool. Jahrh. Jena. 82, 95-136.

Fraser, F. C, Fainstat, T. D., and Kalter, H. (1953). The experimental production of
congenital defects with particular reference to cleft palate. Neo-Natal Studies 2,
43-58.

Fredrickson, D. S., and (iordon, R. S., Jr. (1958). Tran.sport of fatty acids. Physiol-
Revs. 38, 585-630.

Freedman, H. H., and (cordon, A. S. (1955). Endocrine factors in response of the hypo-
thyroid rat to cold. Endocrinology 56, 135-142.

Gaddum, J. H., and Holzbauer, M. (1957). Adrenaline and noradrenaline. Vitamins and
Hormones 15, 152-205.

Gallagher, T. F. (1958). Steroid hormone metabolism and the control of adrenal secretion.
Harvey Lectures Ser. 52, 1-22.

Gallagher, T. F., Kappas, A., Hellman, L., Lipsett, M. B., Pearson, O. H., and West,
C. D. (1958). Adrenocortical hjperfunction in "idiopathic" hirsutism and the
Stein-Leventhal syndrome. Abstr. Endocrine Soc. Program, p. 15.

Ganong, W. F. (1959). Adrenal-hypoph3^seal interrelations. In "Comparative Endo-
crinolog}^," Proc. Columbia University Symposium, 1958 (A. Gorbman, ed.),
pp. 187-201.

Gaunt, R. Renzi, A. A., and Chart, J. J. (1955). Aldosterone: a review. J. Clin. Endo-
crinol, and Metabolism 15, 621-646.

2. Endocrines and Populations 339

Gersh, I., and Grollman, A. (1939). The nature of the X-zone of the adrenal gland of

the mouse. Anat. Record 75, 131-154.
Gersh, I., and Grollman, A. (19-41). The vascular pattern of the adrenal gland of the

mouse and rat and its physiological response to changes in glandular activity.

Carnegie Inst. Wash. Publ. No. 525, 111-126.
Gerwing, J. (1958). The effect of continued toxic stress on the activity of the thyroid

gland in the rat and the guinea-pig. J. Physiol. (London) 144, 243-249.
Gerwing, J., Long, D. A., and Pitt-Rivers, R. (1958). The influence of bacterial exotoxins

on the activity of the thyroid gland in different species. J. Physiol. [London) 144,

229-242.
Giroud, C. J. P., Stachenko, J., and Piletta, P. (1958). Ln vitro studies of the functional

zonation of the adrenal cortex and of the production of aldosterone. In "Aldro-

sterone," Intern. Symposium, Geneva, 1957 (A. F. Muller and C. M. O'Connor,

eds.), pp. .56-72, Little, Brown, Boston, Massachusetts.
Giroud, C. J. P., Stachenko, J., and Venning, E. H. (19.56). Secretion of aldosterone by

the zona glomerulosa of rat adrenal glands incubated in vitro. Proc. Soc. Exptl.

Biol. Med. 92, 154-1.58.
Glaubach, S. (1952). Effect of excessive doses of cortisone, ACTH and prolactin in

pregnant and nursing mice. /. Mt. Sinai Hosp. N. Y. 19, 84-90.
Ghck, D., and Ochs, M. J. (1955). Studies in histochemistry: Quantitative histological

distribution of cholesterol in adrenal glands of the cow, rat and monkey, and effects

of stress conditions, ACTH, cortisone and desoxvcorticosterone. Endocrinology 56,

285-298.
Godfrey, G. K. (1955). Observations on the nature of the decline in numbers of two

Microtus populations. /. Mammal. 36, 209-214.
Goldberg, R. C, Wolff, J., and Creep, R. O. (1957). Studies on the nature of the thyroid-
pituitary interrelationship. Endocrinology 60, 3S-52.
Goldfien, A., Zileh, M. S., Despointes, R. H., and Bethune, J. E. (19.58). The effect of

hypoglycemia on the adrenal secretion of epinephrine and norepinephrine in the

dog. Endocrinology 62, 749-757.
Goodman, H. AL, and Knobil, E. (1959). Mobilization of fatty acids by epinephrine in

normal and hj-pophysectomized rhesus monkeys. Proc. Soc. Exptl. Biol. Med. 100,

195-197.
Gordon, A. S. (1955). Some aspects of hormonal influences upon the leukocytes. Ann.

N. Y. Acad. Sci. 59, 907-927.
Gordon, A. S., and Katsh, (J. F. (1949). The relation of the adrenal cortex to the struc-
ture and phagocytic activity of the macrophagic system. Ann. N. Y. Acad. Sci.

52, 1-30.
Grad, B., and Leblond, C. P. (1949). The necessity of testis and thyroid hormones for

the maintenance of the serous tubules of the submaxillary glands in the male rat.

Endocrinology 45, 250-266.
Gray, I., and Beetham, W. P., Jr. (1957). Changes in plasma concentration of epine-
phrine and norepinephrine with muscular work. Proc. Soc. Exptl. Biol. Med. 96,

636-638.
Gray, S. J., and Ramey, C. G. (1957). Adrenal influence upon the stomach and the

gastric responses to stress. Recent Progr. in Hormone Research 13, 583-617.
Green, J. D. (1951). The comparative anatomy of the hypophysis, with special reference

to its blood supply and innervation. Am. J. Anat. 88, 22.5-312.
Green, R. G., and Evans, C. A. (1940a). Studies on a population of snowshoe hares on

the Lake Alexander area. I. Gross annual censuses 1932-1939. J. Wildlife Manage-
ment 4, 220-238.

340 /. /. Christian

Green, R. G., and Evans, C. A. (1940b). Studies on a population of snowshoe hares on

the Lake Alexander area. II. Mortalitj^ according to age groups and seasons. J.

Wildlife Management 4, 267-278.
Green, R. G., and Evans, C. A. (1940c). Studies on a population of snowshoe hares on

the Lake Alexander area. III. Effect of reproduction and mortality of young hares

on the cycle. /. Wildlife Management 4, 347-358.
Green, R. G., and Larson, C. L. (1938). A description of shock disease in the snowshoe

hare. A)n. J. Hyg. 28, 190-212.
Green, R. G., Larson, C. L., and Bell, J. F. (1939). Shock disease as the cause of the

periodic decimination of the snowshoe hare. Am. J. Hyg. SOB, 83-102.
Greene, R. R., Burrill, M. W., and Ivy, A. C. (1941). Experimental intersexuahty: The

effects of combined estrogens and androgens on the embryonic sexual development

of the rat. /. Exptl. Zool. 87, 211-232.
Greep, R. O., and Deane, H. W. (1947). Cytochemical evidence fo the cessation of

hormone production in the zona glomerulosa of the rat's adrenal cortex after pro-
longed treatment with desoxycorticosterone acetate. Endocrinology 40, 417-425.
Greep, R. O., and Deane, H. W. (1949a). Histological, cytochemical and physiological

observations on the regeneration of the rat's adrenal gland following enucleation.

Endocrinology 45, 42-56.
Greep, R. O., and Deane, H. W. (1949b). The cj^tology and cytochemistry of the adrenal

cortex. Ann. N. Y. Acad. Sci. 50, 596-615.
Greer, M. A. (1951). Evidence of hypothalamic control of the pituitar}^ release of thrj^o-

trophin. Proc. Soc. Exptl. Biol. Med. 77, 603-608.
Greer, M. A. (1952). The role of the hypothalamus in the control of thyroid function.

/. Clin. Endocrinol, and Metabolism 12, 1259-1268.
Greer, M. A. (1957). Studies on the influence of the central nervous system on anterior

pituitary function. Recent Progr. in Hormone Research 13, 67-104.
Gross, F., and Lichtlen, F. (1958). Further evidence for a qualitative difference between

aldosterone and cortexone. In "Aldosterone," Intern. Symposium, Geneva, 1957

(A. F. MuUer and C. M. O'Connor, eds.), pp. 39-55. Little, Brown, Boston, Massa-
chusetts.
Gross, J. (1955). The distribution of radioactive th^-roid hormone in tissues. Brookhaven

Symposia in Biol. No. 7, 102-110.
Grosvenor, C. E., and Turner, C. W. (1959a). Effect of growth hormone and oxytocin

upon milk yield in the lactating rat. Proc. Soc. Exptl. Biol. Med. 100, 158-161.
Grosvenor, C. E., and Turner, C, W. (1959b). Effect of growth hormone upon thj'roid

secretion rate in the rat. Proc. Soc. Exptl. Biol. Med. 100, 70-72.
Grosvenor, C. E., and Turner, C. W. (1959c). Thj-roid hormone and lactation in the rat.

Proc. Soc. Exptl. Biol. Med. 100, 162-165.
Guillemin, R. (1955). A re-evaluation of acetylcholine, adrenaline, noradrenaline and

histamine as possible mediators of the pituitary adrenocorticotorphic activation by

stress. Endocrinology 56, 248-255.
Guillemin, R., Hearn, W. R., Cheek, W. R., and Housholder, D. E. (1957). Control of

corticotrophin release: Further studies with in vitro methods. Endocrinology 60,

488-506.
Guillemin, R., Clayton, G. W., Smith, J. D., and Lipscomb, H. S. (1958). Measurement

of free corticosteroids in rat plasma: Physiological validation of a method. Endo-
crinology 63, 349-358.
Hagen, P., and Welch, A. D. (1956). The adrenal medulla and the biosynthesis of pressor

amines. Recent Progr. in Hormone Research 12, 27-44.

2. Endocrines and Populations 341

Halmi, N. S., Spirtos, B. N., Bogdanove, E. M. and Lipner, H. J. (1953). A study of

various influences on the iodide concentrating mechanism of the rat thryoid. Endo-
crinology 52, 19-32.
Hamilton, W. J., Jr. (1937). The biology of microtine cycles. /. Agr. Research 54, 779-

790.
Hammond, J., Jr. (1954). Light regulation of hormone secretion. Vitamins and Hormones

12, 157-206.
Hamolsk}-, M. W., Gierlach, Z. S., and Jensen, H. (1951). Uptake and conversion of

radioactive iodine (P^^) b}^ the thyroid gland in vivo and in vitro in tourniquet shock

in rats. Am. J. Physiol. 164, 35-53.
Harris, G. W. (1955a). The reciprocal relationship between the thyroid and adreno-
cortical responses to stress. CIBA Foundation Colloq. on Endocrinol. 8, 531-550.
Harris, G. W. (1955b). "Neural Control of the Pituitary Gland." Arnold, London.
Harris, G. W. (1959). Neuroendocrine control of TSH regulation. In "Comparative

Endocrinology." Proc. Columbia University Symposium, 1958 (A. Gorbman, ed.),

pp. 202-222. Wiley, New York.
Harris, G. W., and Fortier, C. (1954). The regulation of anterior function, with special

reference to the secretion of adrenocorticotrophic hormone. In "4th Annual Review

of Stress" (H. Selye and G. Heuser, eds.), pp. 106-127. Acta, Montreal.
Harris, G. W., and Woods, J. W. (1958). The effect of electrical stimulation of the

hypothalamus or pituitary gland on thyroid activity. /. Physiol. (London) 143,

246-274.
Harrison, R. G. (1957). The adrenal circulation in the rabbit. J. Endocrinol. 15, 64-71.
Harrison, R. G. (1951). A comparative study of the vascularization of the adrenal gland

in the rabbit, rat, and cat. /. Anat. 85, 12-23.
Harrison, R. G., and Asling, C. W. (1955). The anatomj^ and functional significance of

the vascularization of the adrenal gland in the rhesus monkey {Macaca mulatta).

J. Anat. 89, 106-113.
Hart, P., and Rees, R. J. W. (1950). Enhancing effect of cortisone on tuberculosis in

the mouse. Lancet ii, 391-395.
Hartman, F. A., and Brownell, K. A. (1949). "The Adrenal Glands." Lea & Febiger,

Philadelphia, Pennsylvania.
Hartroft, P. M., and Eisenstein, A. B. (1957). Alterations in the adrenal cortex of the

rat induced by sodium deficiency: Correlation of histologic changes with steroid

hormone secretion. Endocrinology 60, 641-651.
Hatfield, D. M. (1935). A natural history study of Microtus calif ornicus. J. Mammal.

16, 261-270.
Hellman, K., and Collins, J. (1957). Thyroid, salivary and Harderian glands in mice

exposed to heat. J. Endocrinol. 15, 145-150.
Hess, M. (1953). Effect of thyroidectomy in adrenal weight in adult male rats. Proc.

Soc. Exptl. Biol. Med. 84, 127-130.
Hess, M., and Finerty, J. C. (1952). The effect of hypo- and hyper-thyroidism on the

pituitary-adrenal response to stress. Texas Repts. Biol, and Med. 10, 554-570.
Hess, M., Hall, O., Hall, C. E., and Finerty, J. C. (1952). Endocrine factors effecting the

weight and ascorbic acid content of rat preputial glands. Proc. Soc. Exptl. Biol. Med.

79, 290-292.
Hess, M., Rennels, E. G., and Finerty, J. C. (1953). Response of preputial and adrenal

glands of hypophysectomized rats to ACTH. Endocrinology 52, 223-227.
Hill, R. T. (1937). Ovaries secrete male hormone. I. Restoration of the castrate tj-pe of

seminal vesicle and prostate glands to normal by grafts of ovaries on mice. Endo-
crinology 21, 495-502.

342 J. J. Christian

Hill, R. T. (1948). Grafted mouse ovaries and their adrenal cortical function. Endo-

crinoloyy 42, 339-351.
Hoffmann, R. S. (195S). The role of reproduction and mortality in population fluctuations

of voles [Alicrotus). Ecol. Monoc/raphs 28, 79-109.
Hofmann, F. CJ. (19.)6). Olweivations on in vitro adrenal steroid synthesis in the albino

mouse. Endocrinology 59, 712-715.
Hofmann, F. G. (1957). The in vitro hydroxjdation of 21-carbon steroids by rat adrenal

glands. Endocrinology 60, 382-389.
Holmes, W. N. (1955). Histological variations in the adrenal cortex of the golden hamster

with special reference to the X-zone. Anat. Record 122, 271-294.
Horn, E. H. (1955). Nutritional and hormonal influences upon reproductive maturation,

organ weights and histochemistr}- of immature male rat. Endocrinology 57, 399-40S.
Howard, E. (1928). A transitory zone in the adrenal cortex which shows age and sex

relationships. Am. J. Anat. 40, 251-293.
H(jward, E. (1938). The representation of the adrenal X-zone in rats, in the light of

observations on X-zone variability in mice. Am. J. Anat. 62, 351-373.
Howard, E. (1939). A note on the reactions of the mouse uterus to ovariectomy in tlu'

presence of the X-zone. Bull. Johns Hopkins Hosp. 65, 341-347.
Howard, E. (1946). The effect of adrenalectomy on the accessory reproductive glands of

mice castrated for short periods. Endocrinology 38, 156-164.
Howard, E. (1959). A complementary action of corticosterone and deh^^droepiuandro-

sterone on the mouse adrenal with observations on the reactivity of reproductive

tract structures to dehydroepiandrosterone and 11-hydroxy-androstenedione.

Endocrinology 65, 785 801.
Howard, E., and Benua, R. S. (1950). The effect of protein deficiency and other dietary-
factors on the X-zone of the mouse adrenal. /. Nutrition 42, 157-174.
Hsieh, A. C. L., and Carlson, L. D. (1957a). Role of adrenaline and noradrenaline in

chemical regulation of heat production. Am. J. Physiol. 190, 243-246.
Hsieh, A. C. L., Carlson, L. D., and Gray, G. (1957b). Role of sympathetic nervous

system in the control of chemical regulation of heat production. A))i. J. Physiol. 190,

247-251.
Huggins, C, Parsons, F. M., and Jensen, E. V. (1955). Promotion of growth of preputial

glands b}^ steroids and the pituitary growth hormone. Endocrinology 57, 25-32.
Hughes, E., and Mall, R. (1958). Relation of the adrenal cortex to condition of deer.

Calif. Fish Game 44, 191-195.
Hume, D. M. (1953). The neuro-endocrine response to injury-, present status of the

problem. Ann. Surg. 138, 548-557.
Hurst, v., and Turner, C. W. (1947). The thyroid secretion rate of growing and mature

mice. Am. J. Physiol. 150, 686-692.
Ingalls, T. H. (1956). Causes and prevention of developmental defects. J. Am. Med.

Assoc. 161, 1048-1051.
Ingalls, T. H., and Philbrook, F. R. (1958). Monstrosities induced by hypoxia. Netv Engl.

J. Med. 259, 558-564.
Ingle, D. J. (1949). Some studies on the role of the adrenal cortex in organic metabolism.

Ann. N. Y. Acad. Sci. 50, 576-595.
Ingle, D. J. (1950). The biologic proj>erties of cortisone: A review J. Clin. Endocrinol.

10, 1312-1354.
Irving, J. T. (1957). A compar son of the influence of hormones, vitamins, and other

dietary factors upon the formation of bone, dentine, and enamel. Vitamins and

Hormones 15, 291-232.

2. Endocrities and Populations 343

Isler, H., Leblond, C. P., and Axelrod, A. A. (1958). Mechanism of the thyrod stimu-
lation produced by sodium chloride in the mouse. Endocrinology 62, 159-172.
Jacot, B., and Sel.ye, H. (1951). A non-adrenal mediated action of ACTH on the preputial

glands. Proc. Soc. Exptl. Biol. Med. 78, 46-48.
Jameson, E. W., Jr. (1953). Reproduction of deer mice (Peromyscus maniculatus and P.

hoylii) in the Sierra Nevada, California. J. Mammal. 34, 44-58.
Jameson, E. W., Jr. (1955). Some factors affecting fluctuations of Microlus and Pero-

mycus. J. Mammal. 36, 206-209.
Jones, G. E. S., Delfa, E., and Foot, E. C. (1946). The effect of thiouracil hypothyroidism

on reproduction in the rat. Endocrinology 38, 337-344.
Jones, I. C. (1948). Relationship of the X zone of the mouse adrenal. Proc. Soc. Exptl.

Biol. Med. 69, 120-121.
Jones, I. C. (1949a). The action of testosterone on the adrenal corte.x of hypophysecto-

mized prepuberally castrated mouse. Endocrinology 44, 427-438.
Jones, I. C. (1949b). The relationship of the mouse adrenal cortex to the ])ituitaiy.

Endocrinology 45, 514-536.
Jones, I. C. (1950). The effect of hj^pophysectomy on the adrenal cortex of the immature

mouse. Am. J. Anat. 86, 371-403.
Jones, I. C. (1952). The disappearance of the X zone of the mouse adrenal cortex during

first pregnancy. Proc. Roy. Soc. B139, 398-410.
Jones, I. C. (1957). "The Adrenal Cortex." Cambridge Univ. Press, London and New

York.
Jones, I. C, and Rob}-, C. C. (1954). Some aspects of zonation and function of the

adrenal cortex. I. The effects of hypophysectomy on the adrenal cortex of the adult

male mouse. ./. Endocrinol. 10, 245-2.50.
Jones, I. C, and Spalding, M. H. (1954). Some aspects of zonation and function of the

adrenal cortex. II. The rat adrenal after enucleation. J. Endocrinol. 10, 251-261.
Jones, I. C, and Wright, A. (19.54a). Some aspects of zonation and function of the

adrenal cortex. III. Self-selection after adrenal enucleation in rats. /. Endocrinol

10, 262-265.
Jones, I. C, and Wright, .\. (1954b). Some aspects of zonation and function of the

adrenal cortex. 1\. The histology of the adrenal in rats with diabetes insipidus.

./. Endocrinol. 10, 266-272.
Josimovich, J. B., Ladman, .\. J., and Deane, H. W. (1954). A histophysiological study

of the developing adrenal cortex of the rat during fetal and earl.\' post natal stages.

Endocrinology 54, 627-639.
Kalela, O. (1957). Regulation of reproduction rate in subarctic populations of the vole

Clethrionomys rufocamis {Sund.). Ann. Sci. Fennicae, Ser. A. Iv. Biol. 34, 1-60.
Kalter, H. (1954). The inheritance of susceptibility to the teratogenic action of cortisone

in mice. Genetics 39, 185-196.
Kalter, H. (1956). Modification of teratogenic action of cortisone in mice h\ maternal

age, maternal weight and litter size. Am. J. Physiol. 185, 65-68.
Kass, E. H., Kendrick, M. I., and Finland, M. (1953a). Effect of certain corticosteroids

and of growth hormone on nucleoproteiiis of Ivmph nodes. Ann. .V. Y. Acad. Sci.

56, 737-741.
Kass, E. H., Lundgren, M. M., and Finland, M. (1953b). Observations on the effect of

corticosteroids and growth hormone on resistance to experimental pneumococcal

and influenza virus infections. Ann. N. Y. Acad. Sci. 56, 76.5-771.
Kass, E. H., Kendrick, M. I., and Finland, M. (1955). Effects of corticosterone, hydro-
cortisone and corticotroi)in on production of antibodies in ral)bits. J. Exptl. Med.

102, 767-774.

344 ,/. J. Christian

Kelsall, M. A., and Crabb, E. D. (1958). Lymphocytes and plasmacj'tes in nucleoprotein

metabolism. Ann. N. Y. Acad. Sci. 72, 293-338.
Keuning, F. J., and Van der Slikke, L. B. (1950). The role of immature plasma cells,

lymphoblasts, and lymphocytes in the formation of antibodies, as established in

tissue culture experiments. /. Lab. Clin. Med. 36, 167-182.
Kligman, A. M., Baldridge, G. D., Rebell, G., and Pillsbury, D. M. (1951). The effect of

cortisone on pathologic responses of guinea pigs infected cutaneously with fungi,

viruses, and bacteria. J. Lab. Clin. Med. 37, 615-620.
Knigge, K. M. (1954a). The effect of hypophysectomy on the adrenal gland of the

hamster {Mesocricetus auratus). Am. J. Anat. 94, 225-271.
Knigge, K. M. (1954b). The effect of acute starvation of the adrenal cortex of the

hamster. Anat. Record 120, 555-582.
Knisely, M. H., Bloch, E. H., Ehot, T. S., and Warner, L. (1947). Sludged blood. Science

106, 431-440.
Kraatz, C. P. (1939). Effect of brief experimental hyperthyroidism on reproduction in

the rat. Proc. Soc. Exptl. Biol. Med. 40, 499-502.
Kracht, J. (1954). Fright-thyrotoxicosis in the wild rabljit, a model of thyrotrophic

alarm-reaction. Acta Endocrinol. 15, 355-367.
Kramer, J., Meyers, V. M., McCarthy, H. H., and Simay Kramer, M. (1957). Further

study on the endocrine relations of capillary resistance. Endocrinology 60, 589-596.
Kuhl, W. J., Jr., and Ziff, M. (1952). Alteration of thyroid function by ACTH and

cortisone. /. Clin. Endocrinol, and Metabolism 12, 555-559.
Lanman, J. T. (1957). The adrenal fetal zone: Its occurrence in primates and a possible

relationship to chorionic gonadotropin. Endocrinology 61, 684-691.
Laqueur, G. L., McCann, S. M., Schreiner, L. H., Rosemberg, E., Rioch, D., and Ander-
son, E. (1955). Alterations of adrenal cortical and ovarian activity following hypo-
thalamic lesions. Endocrinology 57, 44-54.
Leathern, J. H. (1950). Seminal vesicle response as an assessment of androgenic activity.

Trans. N. Y. Acad. Sci. [2], 12, 234-235.
LeMaistre, C. A., Tompsett, R., and McDermott, W. (1953). The effects of cortico-
steroids upon tuberculosis and pseudotuberculosis. Ann. N. Y. Acad. Sci. 56,

772-778.
Levy, A. C, and Ramey, E. R. (1958). Effect of autonomic blocking agents on depot

fat mobilization in normal adrenalectomized animals. Proc. Soc. Exptl. Biol. Med.

99, 637-639.
Li, C. H., F0nss-Bech, P., Geschwind, I. I., Hayashida, T., Hungerford, G. F., Lostroh,

A. J., Lyons, W. R., Moon, H. D., Reinhardt, W. O., and Sideman, M. B. (1957).

Corticotropins (ACTH) X. Biological investigations of a-corticotropin. /. Exptl.

Med. 105, 335-360.
Liddle, G. W., Duncan, L. E., and Bartter, F. C. (1956). Dual mechanism regulating

adrenocortical function in man. Metabolism 5, 369-383.
Lillie, R. D. (1954). Histopathologic Technic and Practical Histochemistry. McGraw-
Hill (Blakiston), New York.
Lord, R. A. (1958). Studies on the use of cortisone and ACTH in trichinosis. Am. J.

Trop. Med. Hyg. 7, 611-617.
Lostroh, A. J., and Li, C. H. (1957). Stimulation of the sex accessories of hypophysecto-

mized male rats by non-gonadotrophic hormones of the pituitary gland. Acta

Endocrinol. 25, 1-16.
Lostroh, A. J., and Li, C. H. (1958). Effect of growth hormone and thyroxine on body

weight of hypophysectomized C3H mice. Endocrinology 62, 484-492.

2. Endocrines and Populations 345

Lostroh, A. J., and Woodward, P. (1958). Changes in the adrenal of hypophysectomizc^d
C3H mouse with a-corticotropin and growth hormone. Endocrinology 62, 498-505.
Louch, C, Meyer, R. K., and Emlen, J. T. (1953). Effect of stress on diurnal fluctuations

in eosinophils of the laboratory mouse Proc. Soc. Exptl. Biol. Med. 82, G68-671.
Louch, C. D. (1956). Adrenocortical activity in relation to the density and dynamics of

three confined populations of Microlus pennsylvanicus. Ecology 37, 701-713.
Louch, C. D. (1958). Adrenocortical activity in two meadow vole populations. J.

Mammal. 39, 109-116.
Lutwak-Mann, C. (1958). The dependence of gonadal function upon vitamins and other

nutritional factors. Vitamins and Hormones 16, 35-75.
Lutwak-Mann, C, Mann, T., and Price, D. (1949). Metabolic activity in tissue trans-
plants. Hormone induced formation of fructose and citric acid in transplants from

accessory glands of reproduction. Proc. Roy. Soc. B136, 461-471,
McCann, S. M. (1953). Effect of hypothalamic lesions on the adrenal cortical response

to stress in the rat. Am. J. Physiol. 175, 13-20.
McCann, S. M. (1957). The ACTH-releasing activity of extracts of the posterior lobe of

the pituitary in vivo. Endocrinology 60, 664-676.
McCann, S. M., and Brobeck, J. R. (1954). Evidence for a role of the supraopticohypo-

physeal system in regulation of adrenocorticotrophin secretion. Proc. Soc. Exptl.

Biol. Med. 87, 318-324.
McCann, S. M., and Fruit, A. (1957). Eflfect of synthetic vasopressin on release of

adrenocorticotrophin in rats with hypothalamic lesions. Proc. Soc. Exptl. Biol. Med.

96, 566-567.
McDonald, R. K., Wagner, H. N., Jr., and Wiese, V. K. (1957). Relationship between

endogenous antidiuretic hormone activity and ACTH release in man. Proc. Soc.

Exptl. Biol. Med. 96, 652-655.
MacLulich, D. A. (1937). Fluctuations in the numbers of the varying hare. Univ. Toronto

Studies Biol. Ser. 43, 1-136.
McPhail, M. K. (1944). Androgenic activity of the adrenal gland with special reference to

the X-zone. Rev. can. hiol. 3, 312-327.
McPhail, M. K., and Read, H. C. (1942a). The mouse adrenal. I. Development, de-
generation, and regeneration of the X-zone. Anat. Record 84, 51-73.
McPhail, M. K., and Read, H. C. (1942b). The mouse adrenal. II. The action of certain

hormonal substances on the adrenal gland of the mouse with particular reference to

their action on the X-zone. Anat. Record 84, 75-89.
Mann, T. (1954). On the presence and role of inositol and certain other substances in the

seminal vesicle secretion of the boar. Proc. Roy. Soc. B142, 21-32.
Maqsood, M., and Reineke, E. P. (1950). Conference of environmental temperatures and

thyroid status on sexual development in male mouse. Am. J. Physiol. 162, 24-30.
Marine, D., and Bauman, E. J. (1945). Hypertrophy of adrenal medulla of white rats in

chronic thiouracil poisoning. Am. J. Physiol. 144, 69-73.
Markee, J. E., Everett, J. W., and Sawyer, C. H. (1952), The relationship of the nervous

system to the release of gonadotrophin and the regulation of the sex cycle. Recent

Progr. in Hormone Research 7, 139-163.
Marrian, G. F., and Parkes, A. S. (1929). The effect of anterior pituitary preparations

administered during dietary anestrus. Proc. Roy. Soc. B105, 248-258.
Martin, E. P. (1956). A population study of the prairie vole {Microtus ochrogasler) in

Northeastern Kansas. Univ. Kansas Museum Nat. Hist. Publ. 8, 361-416.
Martorano, J. J. (1957). Variation in liver glycogen levels of intact and adrenalectomized

mice. Naval Med. Res. Inst. Res Rept. 15, 191-198.

346 ./. J. Christian

Mason, J. W. (1959). Central nervous system factors in the regulation of endocrine

secretion. Recent Progr. in Hormone Research, 15, 345-3(S9.
Meier, R., Schuler, W., and DesauUes, P. (1950). Zur Frage des Mechanismus der

Hemmung des Bindegewebswachstume durch Cortisone. Experientia 6, 469-471.
Menkin, V. (1955). Factors concerned in the mobilization of leukocytes in inflammation.

Ann. N. Y. Acad. Sci. 59, 956-958.
Miller, R. A. (1949). Cytological criteria of activity and inactivity in the glomerular

fascicular zones of the adrenal cortex in mice. Anat Record 103, 489.
Mills, I. H., Casper, A., and Bartter, F. C. (1958). On the role of the vagus in the control

of aldosterone secretion. Science 128, 1140-1141.
Money, W. L. (1955). The interrelation of the thyroid and the adrenals. Brookhaven

Symposia in Biol. No. 7, 137-168.
Money, W. L., Kirschner, L., Kraintz, L., Merrill, P., and Rawson, R. W. (1950). Effects

of adrenal and gonadal products on the weight and radioiodine uptake of the

thyroid gland in the rat. J . Clin. Endocrinol. 10 1282-1295.
Moon, H. D., Simpson, M. E., Li, C. H., and Evans, H. M. (1950). Neoplasms in rats

treated with pituitary growth hormone. II. Adrenal gland. Cancer Research 10,

364-370.
Moon, H. D., Simpson, M. E., Li, C. H., and Evans, H. M. (1951). Neoplasms in rats

treated with pituitary growth hormone. V. Absence of neoplasms in hypophysecto-

mized rats. Cancer Research 11, 535-539.
Morin, G. (1946a). MeduUo-surrenale et regulation thermique. I. Action calorigene de

I'adrenaline. Demonstration, signification. Rev. can. hiol. 5, 121-134.
Morin, G. (1946b). MeduUo-surrenale et regulation thermique. II. Action adrenalino-

secr^trice du froid. Rev. can. hiol. 5, 388-400.
Mosier, H. D. (1957). Comparative histological study of the adrenal cortex of the wild

and domesticated Norway rat. Endocrinology 60, 460-469.
Moss, M. L. (1955). Morphological changes in the growing rat skull following administra-
tion of cortisone acetate. Proc. Soc. Expll. Biol. Med. 89, 648-650.
Mulinos, M. G., and Pomerantz, L. (1941). Hormonal influences on the weight of the

adrenal in inanition. Am. J . Physiol. 132, 368-374.
Mulinos, M. G., Pomerantz, L., Smelser, J., and Kurzrok, R. (1939). Estrus-inhibiting

effects of inanition. Proc. Soc. Exptl. Biol. Med. 40, 79-83.
Nalbandov, A. V., Moore, W. W., and Norton, H. W. (1955). Further studies on the

neurogenic control of the estrus cycle liy uterine distension. Endocrinology 56,

225-231.
Nelson, D. H. (1955). The significance of plasma corticosteroids as a measure of adrenal

cortical function. In "Symposium on Adrenal Functions in Infants and Children,"

N. Y. Univ. Med. Ctr., Syracuse, 1954. NMRI (Naval Med. Res. Inst.) L and R

55-1.
Nichols, J. (19.50). Effects of captivity on adrenal gland of wild Norway rat. Am. J.

Physiol. 162, 5-9.
Noble, R. L. (1955). Physiology of the adrenal cortex. In "The Hormones" (G. Pincus

and K. V. Thimann, eds.), \o\. Ill, pp. 685-819. Academic Press, New York.
Nowell, N. W. (1959). Studies in the activation and inhibition of adrenocorticotrophin

secretion. Endocrinology 64, 191-201.
Odum, E. P. (1955). An eleven year history of a Sigmodon population. ./. Man^iud. 36,

368-378.
Parmer, L. G., Katonah, F., and Angrist, A. A. (1951). Comparative effects of ACTH,

cortisone, corticosterone, desoxycorticosterone, pregnenolone on growth and de-
velopment of infant rats. Proc. Soc. E.rptl. Biol. Med. 77, 21.5-218.

2. Endocrines and Populations 347

Paschkis, K. E., Cantarow, A., Eberhard, T., and Doyle, D. (1950). Thyroid funftion in

the alarm reaction. Froc. Soc. Exptl. Biol. Med. 73, 116-118.
Pauly, J. E. (1957). Morphological observations on the adrenal cortex of the laboratory

rat. Endocrinology 60, 247-264.
Peron, F. G., and Dorfman, R. I. (1958). Quantitative analysis of steroids isolated from

the urine of normal and ACTH treated guinea pigs. Endocrinology 62, 1-8.
Perry, W. F. (1951). The action of cortisone and ACTH on thryoid function. Endo-
crinology 49, 284-288.
Pitelka, F. A. (1957a). Some aspects of population structure in the short-term cycle of the

brown lemming in Northern Alaska. Cold Spring Harbor Symposia Quant. Biol.

22, 237-251.
Pitelka. F. A. (1957b). Some characteristics of microtine cycles in the arctic. Biol.

Colloq. Proc. 18th Colloq. 73-88.
Poindexter, H. A. (1949). Experimental animal colony in tropical West Africa. Public

Health Repts. {U. S.) 64, 57-62.
Pollard. M., and Wilson, B. R. (1955). The prolongation of murine typus rickettsemia by

cortisone. Texas Repts. Biol, and Med. 13, 126-128.
Poore, W., and Hollander, V. P. (1957). Plasma ketosteriod studies in normal mice.

Endocrinology 61, 651-656.
Porter, J. C, and Jones, I. C. (1956). Effect of plasma from hypophyseal-portal vessel

blood on adrenal ascorbic adid. Endocrinology 58, 62-67.
Porter, J. C, and Rumsfeld, H. W., Jr. (1956). Effects of lyophilized plasma and plasma

fractions from h3-pophvseal-portal vessel blood on adrenal ascorbic acid. Endo-
crinology 58, 359-364.
Preble, E. A. (1908). A biological investigation of the Athabaska-Mackenzie region.

U. S. Dept. Agr. N. Am. Fauna No. 27.
Puntriano, G., and Meites, J. (1951). The effects of continuous light or darkness on

th^-roid function in mice. Endocrinology 48, 217-224.
Ramey, E. R., and Goldstein, M. S. (1957). The adrenal cortex and the sympathetic

nervous system. Physiol. Revs. 37, 155-195.
Randall, H. T. (1952). The shifts of fluid and electrolytes in shock. Ann. N. Y.Acad.

Sci. 55, 412-428.
Ratcliffe, H. L., and Cronin, M. T. I. (1958). Changing frequency of arteriosclerosis in

mammals and birds at the Philadelphia Zoological Gardens. Circtilation 18, 41-52.
Rausch, R. (1950). Observations on a cyclic decline of lemmings {Lemmus) on the arctic

coast of Alaska during the spring of 1949. Arctic 3, 166-177.
Rawson, R. W., Rail, J. E., and Sonenberg, M. (1955). The chemistry and physiology of

the thyroid. In "The Hormones" (G. Pincus and K. V. Thimann, eds), Vol. Ill,

pp. 433-519. Academic Press, New York.
Reichlin, S. (1957a). The effect of dehydration, starvation, and pitressin injections on

thyroid activity in the rat. Endocrinology 60, 470-487.
Reichlin, S. (1957b). The effect of hypothalamic lesions upon the thyroid response to

partial thyroidectomy. Endocrinology 60, 567-569.
Reif, A. E., and Longwell, B. B. (1958). Extraction of corticosteroids from adrenal vein

blood of rats and rabbits. Endocrinology 62, 573-586.
Renuels, E. G., and Drager, G. A. (1955). The relationship of pituicytes to neurosecretion.

Anat. Record 122, 193-203.
Rennels, E. G., Hess, M., and Finerty, J. C. (1953). Response of preputial and adrenal

glands of the rat to sex hormones. Proc. Soc. Exptl. Biol. Med. 82, 304-309.
Retzlaff, E. G. (1938). Studies in population physiology with the albino mouse. Biol.

Generalis 14, 238-265.

348 /. /. Christian

Richards, J. B., and Pruitt, R. L. (1957). Hydrocortisone suppression of stress-induced

adrenal 17-hydroxycorticosteroid secretion in dogs. Endocrinology 60, 99-104.
Richardson, J. A., and Woods, E. F. (19.59). Release of norepinephrine from the isolated

heart. Proc. Soc. Exptl. Biol. Med. 100, 149-151.
Richter, C. P., Rogers, P. V., and Hall, C. E. (1950). Failure of salt replacement therapy

in adrenalectomized recently captured wild Norway rats. Endocrinology 46, 233-242.
Ring, G. C. (1942). The importance of the thyroid in maintaining an adequate production

of heat during exposure to cold. Am. J. Physiol. 137, 582-588.
Robert, A., and Nezamis, J. E. (1957). The granuloma pouch as a routine assay for

antiphlogistic compounds. Acta Endocrinol. 25, 105-112.
Robinson, H. J., and Smith, A. L. (1953). The effect of adrenal cortical hormones on

experimental infection. Ann. N. Y. Acad. Sci. 56, 757-764.
Rogers, P. V., and Richter, C. P. (1948). Anatomical comparison between the adrenal

glands of wild Norwaj-, wild alexandrine and domestic Norway rats. Endocrinology

42, 46-55.
Rosemberg, E., Cornfield, J., Bates, R. W., Anderson, E. (1954). Bioassay of adrenal

steroids in blood and urine based on eosinophil response: a statistical analj'sis.

Endocronilogy 54, 363-383.
Rosenfeld, G. (1958). The function and capacity of the adrenal cortex immediately

before radiation sickness death. J. Lab. Clin. Med. 51, 198-210.
Rosenfeld, G., and Bascom, W. D. (1956). In vitro studies of the influence of corticotropin,

somatotropin, thj-rotropin and gonadotropin on adrenal steroidogenesis. Endo-
crinology 59, 497-499.
Rothballer, A. B. (1953). Changes in the rat neurohypophysis of painful stimuli with

particular reference to neurosecretory material. Anat. Record 115, 21-41.
Saffran, M., Schally, A. V., and Benfey, B. G. (1955). Stimulation of the release of

corticotrophin from the adenoh3'pophysis by a neurohypophj'seal factor. Endo-
crinology 57, 439-444.
Sakiz, E. (1956). Le cortex surrenal exerce-t-il une action hormonale sexuelle? Corn-pi.

rend. soc. biol. 150, 2172-2174.
Salter, W. T. (1949). The metabolic circuit of thyroid hormone. Ann. N. Y. Acad. Sci. 50,

358-376.
Sandberg, A. A., Nelson, D. M., Palmer, J. G., Samuel, L. T., and Tyler, F. H. (1953).

The effects of epinephrine on the metabolism of 17-h3'dro.xycorticosteroids in the

human. /. Clin. Endocrinol, and Metabolism 13, 629-647.
Santisteban, G. A., and Doughert}-, T. F. (1954). Comparison of the influence of adreno-
cortical hormones on the growth and involution of lymphatic organs. Endocrinology

54, 130-146.
Sayers, G., and Sayers, M. (1949). The pituitary-adrenal system. Ann. N. Y. Acad. Sci.

50, 522-539.
Schally, A. V., and Guillemin, R. (1959). Concentration of corticotropin releasing factor

by chromatography on carboxymethylcellulose. Proc. Soc. Exptl. Biol. Med. 100,

138-139.
Schapiro, S., Marmorston, J., and Sobel, H. (1958). Mobilization of the antidiuretic

hormone and the secretion of ACTH following cold stress. Endocrinology 62, 278-282.
Scharrer, E., and Scharrer, B. (1954). Hormones produced by neurosecretorj-- cells.

Recent Progr. in Hormone Research 10, 183-240.
Schindler, W. J., and Knigge, K. M. (1959a). Adrenal cortical secretion by the golden

hamster. Endocrinology 65, 739-747.
Schindler. W. J., and Knigge, K. M. (1959b). In vitro studies and adrenal steroidogenesis

by the golden hamster. Endocrinology 65, 748-765.

2. Endocrines and Populations 349

Schmidt, L. H., and Squires, W. L. (1951j. The influence of cortisone on primate malaria.

/. Exptl. Med. 94, 501-520.
Schmiterlow, C. G. (1948). The nature and occurrence of pressor and depressor sub-
stances in extracts from blood vessels. Acta Physiol. Scand. 16, Siippl. 56, 1-113.
Scudder, J. (1952). Introduction to shock syndrome. Ann. N. Y. Acad. Sci. 55, 3-49-350.
Sealander, J. A., Jr. (1950). Effect of environmental temperature and starvation on

adrenal glands of the white-footed mouse. Am. J. Physiol. 163, 92-95.
Seifter, J., Ehrich, W. E., and Hud^-ma, G. M. (1949). Effect of prolonged administration

of antithyroid compounds on the th^-roid and other endocrine organs of the rat.

A. M. A. Arch. Pathol. 48, 536-547.
Seifter, J., Christian, J. J., and Ehrich, W. E. (1951). The effect of cortisone and other

steroids on the hibernating gland of the pregnant white rat. Federation Proc. 10,

334.
Seifter, J., Ehrich, W. E., Baeder, D. H., Butt, A. J., and Hauser, E. A. (1953). Evidence

for the direct effect of steroids on the ground substance. Ann. N. Y . Acad. Sci. 55,

696-697.
Seifter, J., Baeder, D. H., Zarafonetis, C. J. D,, and Kalas, J. (1959). Effect of adrenals,

pituitar}^ hver, and mucopolysaccharides on blood lipids. In "Hormones and

Atherosclerosis" (G. Pincus, ed.), pp. 265-2S1. Academic Press, New York.
Selye, H. (1939). The effect of adaptation to various damaging agents on the female sex

organs in the rat. Endocrinology 25, 61.5-624.
Selye, H. (1947). "Textbook of Endocrinology." Acta, Montreal.
Selye, H. (1950). "Stress." Acta, Montreal.
Seh'e, H. (1951). The influence of STH, ACTH and cortisone upon resistance to infection.

Can. Med. Assoc. J. 64, 489-494.
Sel3-e, H. (1954). Stress and lactation. Rev. can. hiol. 13, i,4)) 377-384.
Selye, H. (1955). Interactions between corticoid and foUiculoid hormones in the regu-
lation of anabolism, lymphatic tissue and inflammation. Acta Endocrinol. 20, 1-18.
Selye, H., and Bois, P. (1954). On the design of experiments to appraise the role of the

adrenals in biologic reactions. In "4th Annual Report on Stress," (H. Selye and G.

Heuser, eds.), pp. 533-552. Acta, Montreal.
Share, L., and Stadler, J. B. (1958). Alterations in sodium and potassium metabolism

following hind-leg fracture in the rat: Role of the adrenal cortex. Endocrinology 62,

119-128.
Shwartzman, G., and Aronson, S. M. (1953). Pohomj^elitis infection by parenteral routes

made possible bj^ cortisone. Ann. N . Y . Acad. Sci. 56, 793-798.
Simpson, M. E., Evans, E. S., and Rosenberg, L. L. (1959). Re-evaluation of the evidence

for a pituitary erythropoietic hormone. Endocrinology 64, 592-599.
Slusher, M. A. (1958). Dissociation of adrenal ascorbic acid and corticosterone response

to stress in rats with hypothalamic lesions. Endocrinology 63, 412-419.
Smelser, G. K. (1943). Changes induced in the Harderian gland of the guinea pig b^- the

injection of Iwpophyseal extracts. Anat. Record 86, 41-56.
Smith, E. K. (1955). Interrelationships of anterior pituitary and adrenal cortex in the rat

ovarian hyperemic reaction. Endocrinology 56, 567-574.
Smith, R. W., Jr. (1951). Augmented excretion of urine gonadotrophins during ACTH

administration. Proc. Soc. Exptl. Biol. Med. 78, 868-872.
Snyder, R. L., and Christian, J. J. (1960). The reproduction cycle and the litter size of

the southern woodchuck. Ecology 41, 647-656.
Sollman, T. (1957). "A Manual of Pharmacology," 8th ed. Saunders, Philadelphia.
Southern, H. N. (1948). Sexual and aggressive behaviour in the wild rabbit. Behaviour 1,
173-194.

350 J. J. Christian

Southwick, C. H. (1955a). The population dynamics of confined house mice supplied with

unlimited food. Ecologrj 36, 212-225.
Southwick, C. H. (1955b). The regulatory mechanisms of house mouse population;

social behavior affecting litter survival. Ecology 36, 627-634.
Southwick, C. H. (1959). Eosinophil response of C57BR mice to behavioral disturbance.

Ecology 40, 156-157.
Speirs, R. S. (1955). Physiological approaches to an understanding of the function of

eosinophils and basophils. Ann. N. Y. Acad. Set. 59, 706-731.
Speirs, R. S., and Meyer, R. K. (1949). The effects of stress, adrenal and adrenocorti-

cotrophic hormones on the cirulating eosinophils of mice. Endocrinology 45, 403-429.
Speirs, R. S., and Meyer, R. K. (1951). A method of assajnng adrenal cortical hormones

based on a decrease in the circulating eosinophil cells of adrenalectomized mice.

Endocrinology 48, 316-326.
Srebnik, H. H., Nelson, M. M., and Simpson, M. E. (1958). Response to exogenous

gonadotiophins in absence of dietary protein. Proc. Soc. Expll. Biol. Med. 99, 57-61.
Stachenko, J., and Giroud, C. J. P. (1959a). Functional zonation of the adrenal cortex:

Pathways of corticosteroid biogenesis. Endocrinology 64, 730-742.
Stachenko, J., and Giroud, C. J. P. (1959b). Functional zonation of the adrenal cortex

site of ACTH action. Endocrinology 64, 743-752.
Stanbury, S. W., Gowenlock, A. H., and Mahler, R. F. (19.5S). Interrelationships of

potassium deficiency and renal disease. In "Aldosterone," Intern. Symposium,

Geneva, 1957 (A. F. Muller and C. M. O'Connor, eds.), pp. 155-166. Little, Brown,

Boston, Massachusetts.
Steinberger, E., and Dixon, W. J. (1959). Some observations on the effect of heat on the

testicular germinal epithelium. Fertility and Sterility 10, 578-595.
Stevens, C. E., D'Angelo, S. A., Paschkis, K. E., Cantarow, A., and Sunderman, F. W.

(1955). The response of the pituitary thyroid system of the guinea pig to low en-
vironmental temperature. Endocrinology 56, 143-156.
Stoner, R. D., and Godwin, J. T. (1953). The effects of ACTH and cortisone upon

susceptibility to trichinosis in mice. Am. J. Pathol. 29, 943-950.
Story, J. L., Melby, J. C, Egdahl, R. H., and French, L. A. (1959). Adrenal cortical

function following stepwise removal of the brain of the dog. Am. J. Physiol. 196.

583-588.
Strecker, R. L. (1954). Regulatory mechanics in house mouse populations: The effect of

limited food supply on an unconfined population. Ecology 35, 249-253.
Strecker, R. L., and Emlen, J. T. (1953). Regulatory mechanisms in house mouse popu-
lations— the effect of limited food supply on a confined population. Ecology 34,

375-385.
Sturgis, S. H. (1950). The mechanism and control of primate ovulation. Fertility and

Sterility 1, 40-52.
Sundberg, R. 1). (1955). Lymphocytes and plasma cells. Ann. N. Y. Acad. Sci. 59(5),

671-689.
Swanson, H. E. (1956). Interrelationships between thyroxin and adrenalin in the regu-
lation of oxygen consumption on the albino rat. Endocrinology 59, 217-225.
Swanson, H. E. (1957). The effect of temperature in the potentiation of adrenalin by

thyroxine in the albino rat. Endocrinology 60, 205-213.
Sweat, M. L., and Farrell, G. L. (1952). Resolution and quantitative analysis of steroids

in adrenal-vein blood. /. Clin. Endocrinol, and Metabolism 12, 168.
Sweat, M. L., and Farrell, G. L. (1954). Decline of corticosteroid secretion following

hypophysectomy. Proc. Soc. Exptl. Biol. Med. 87, 615-618.

2. Endocrines and Populations 351

Sydnor, K. L., and Sayers, G. (1954). Blood and pituitary ACTH in intact and adrenalec-

tomized rats after stress. Endocrinology 55, 621-636.
Sydnor, K. L., Burks, R., and Brown, D. (1955). Blood ACTH in the stressed adrenalec-

tomized rat after intravenous injection of hydrocortisone. Endocrinology 56, 204-208
Symington, T., Currie, A. R., O'Donnell, U. J., Grant, J. K., Oastler, E. G., and Whyte,

W. H. (1958). Hyperplasia and tumours of the human adrenal cortex: histologj^,

coenzymic changes and corticoid production. CIBA Foundation CoUoq. on Endo-
crinology 12, 102-121.
Tamura, Y. (1926). Structural changes in the suprarenal gland of the mouse during

pregnancy. Brit. J. Exptl. Biol. 4, 81-92.
Tanaka, R. (1956). Fluctuations of vole populations following the widespread synchro-
nous flowering of bamboo-grasses on Mt. Turugi. Bull. Kochi Women's College 4,

61-68.
Tanaka, R. (1957). An ecological review of small-mammal outbreaks with special ref-
erence to their association with the flowering of bamboo grasses. Bull. Kochi Women's

Univ. Nat. Sci. Ser. 5, 20-30.
Tata, J. R., Rail, J. E., and Rawson, R. W. (1957). Metabolism of L-thyroxine and

L-3:5:3'-ti'iiodothyronine by brain tissue preparations Endocrinology 60, 83-98.
Taubenhaus, M. (1953). The influence of cortisone upon granulation tissue and its

synergism and antagonism to other hormones. Ann. N. Y. Acad. Sci. 55, 666-673.
Taubenhaus, M. and Amromin, D. (1950). The effects of the hypophysis, thyroid, sex

steroids, and the adrenal cortex upon granulation tissue. J. Lab. Clin. Med. 36,

7-18.
Thomas, L. (1953). Cortisone and infection. Ann. X. Y. Acad. Sci. 56, 799-814.
Thompson, D. Q. (^1955). The I'ole of food and cover in population fluctuations of the

brown lemming at Point Barrow, Alaska. Trans. N. Am. Wildlife Conf. 20, 166-176.
Thorn, G. W., Jenkins, D.. and Laidlaw, J. C. (1953). The adrenal response to stress in

man. Recent Progr. in Hormone Research 8, 171-215.
Thorn, N. A. (1958). Mammalian antidiuretic hormone. Physiol. Revs. 38, 169-195.
Tobach, E., and Block, H. (1956). Eft'ect of stress by crowding prior to and following

tuberculous infection. Am. J. Physiol. 187, 399-402.
Tyndale-Biscoe, C. H., and Williams, R. M. (1955). A study of natural mortality in a

wild population of the I'abbit, Oryctolagus cunicuhis (L.). New Zealand J. Sci.

Technol. B36(6), 561-580.
Uhrich, J. (1938). The social hierarchy in albino mice. J . Comp. Psychol. 25, 373-413.
Vander Laan, W. P. (1955). The biological significance of the iodide concentrating

mechanism of the thjroid gland. Brookhuven Symposia in Biol. No. 7, 30-39.
^'ander Laan, W. P., and Caplan, R. (1954). Observations on a relationship between

total thyroid iodine content and the iodide concentrating mechanism of the th^Toid

gland of the rat. Endocrinology 54, 437-447.
Vander Laan, W. P., and Greer, M. (1950). Some effects of the hypophysis on iodine

metabolism by the thyroid gland of the rat. Endocrinology 47, 36-47.
\'andeWiele, R., and Lieljerman, S. (1960). The metabolism of dehydroisoandiosterone.

In "Biological Activities of Steroids in Relation to Cancer" (CJ. Pincus and E. P.

Vollmer, eds.), pp. 93-110. Academic Press, New York.
Van Middlesworth, L., and Berrj', M. M. (1951). Iodide metabolism during anoxia,

nephrectomy, trauma, avitaminoses and starvation in the rat. .4/?;. /. Physiol. 167,

576-580.
Verzar, F., and Vidovic, V. (1952). Action of T.S.H. and corti.sone on uptake of'-" I

by thyroid gland of hypophysectomized rats. /. Endocrinol. 8, 321-328.

352 /. /. Christian

"\^etulani, T. (1931). Untersuchungen iiber das Wachstum der Saugetiere in Abhangigkeit

von der Anzahl zusammengehaltener Tiere. Biol. Generalis 7, 71-98.
Visscher, M. B., and Halberg, F. (1955). Daily rhythms in numbers of circulating

eosinophils and some related phenomena Ann. N. Y. Acad. Sci. 59, 834-8-49.
von Euler, U. S. (1951). The nature of adrenergic nerve mediators. Pharmacol. Revs. 3,

247-277.
Wallach, D. P., and Reineke, E. P. (1949). The effect of varying levels of thj-roidal

stimulation on the ascorbic acid content of the adrenal cortex. Endocrinology 45,

75-81.
Waring, H. (1935). The development of the adrenal gland of the mouse. Quart. J.

Microscop. Sci. 78, 329-366.
Waring, H. (1942). Effect of hormones on degeneration of the x-zone in the mouse

adrenal. /. Endocrinol 3, 123-131.
Watt, L. J. (1934). Frequencj' distribution of litter size in mice. J. Mammal. 15, 185-189.
Weaver, J. A. (1955). Changes induced in the thj-mus and lymph nodes of the rat by the

administration of cortisone and sex hormones and by other procedures. J. Pathol.

Bacterial. 69, 133-139.
Weber, A. F., McNutt, S. H., and Morgan, B. B. (1950). Structure and arrangement of

zona glomerulosa cells in the bovine adrenal. J. Morphol. 87, 393-416.
Wexler, B. C, Runfret, A. P., Griffin, A. C, and Richardson, H. C. (1955). Evidence of

pituitary control of the lipid content of the zona glomerulosa of the rat adrenal

cortex. Endocrinology 56, 120-123.
Whitney, O. M., and Anigstein, L. (1953). Effect of cortisone on experimental rickett-

sioses and on antibody formation. Texas Repts. Biol, and Med. 11, 561-569.
Whitten, W. K. (1959). Occurrence of anoestrus in mice caged in groups. J. Endocrinol.

18, 102-107.
Whittle, P. (1955). An investigation of periodic fluctuation in the New Zealand rabbit

population. New Zealand J. Sci. Technol. B37(2), 179-200.
WiUiams, R. H., Jaffe, H., and Kemp, C. (1949). Effect of severe stress upon thyroid

function. Am. J. Physiol. 159, 291-297.
Wilson, H., Borris, J. J., and Bahn, R. C. (1958). Steroids in the blood and urine of female

mice bearing on ACTH-producing pituitary tumor. Endocrinology 62, 135-149.
Woodbury, D. M. (1958). Relation between the adrenal cortex and the central nervous

system. Pharmacol. Revs. 10, 275-357.
Woods, R., and Carlson, L. D. (1956). Thyroxine secretion in rats exposed to cold.

Endocrinology 59, 323-330.
Wright, A., Jones, I. C, and Phillips, J. G. (1957). The histologj' of the adrenal gland of

the prototheria. /. Endocrinol. 15, 100-107.
Yamada, T. (1959). Studies on the mechanism of hypothalamic control of thyrotropin

secretion: Effect of intrah^-pothalamic thyroxine injection on thj'roid hypertrophy

induced by propjdthiouracil in the rat. Endocrinology 65, 216-224.
Yamada, T., and Greer, M. A. (1959). Studies on the mechanism of h.vpothalamic control

of thjTotropin secretion: Effect of thyroxine injection into the hypothalamus or the

pituitary on thyroid hormone release. Endocrinology 64, 559-566.
Yankopoulos, N. A., Davis, J. O., Kliman, B., and Peterson, R. E. (1959). Evidence

that a humoral agent stimulates the adrenal cortex to secrete aldosterone in ex-
perimental secondary hyperaldosteronism. J. Clin. Invest. 38, 1278-1289.
Zalesky, M. (1934). A study of the seasonal changes in the adrenal gland of Citellus

with particular reference to its sexual cycle. Anat. Record 60, 291-321.

2. Endocrines and Populations 353

Zimmermann, K. (1937). Die markische Rotenmaus. Analyse einer population. Markische

Tierewelt 3, 24-40.
Zweifach, B. W. (1952). Functional deterioration of terminal vascular bed in irreversible

hemorrhagic shock. Ann. N. Y. Acad. Sci. 55, 370-380.
Zweifach, B. W., Shorr, E., and Black, M. M. (1953). The influence of the adrenal

cortex on behavior of terminal vascular bed. Ann. N. Y. Acad. Sci. 55, 626-633.
Zwemer, R. L. (1936). A study of adrenal cortex morphology. A7n. J. Pathol. 12, 107-114.
Zwemer, R. L., Wooton, R. M., and Norkus, M. G. (1938). A study of corticoadrenal

cells. Anat. Record 72, 249-264.

Author Index

N umbois ill italics indicate the pages on which the complete references are listed.

Abrams, G. D., 242, 253, 329

Ackerman, N. B., 233, 328

Agate, F. J., Jr., 244, 328

Albert, A., 232, 328

Allen, J. C, 196, 282, 329

Allen, J. M., 196, 201, 213, 329

Alpert, M., 194, 196, 329

Altzuler, N., 227, 335

Amiot, L. W., 97, 187

Amromin, D., 209, 254, 351

Anderson, C. H., 202, 203, 204, 209, 213,

329, 338
Anderson, E., 206, 212, 218, 257, 258,

283, 285, 329, 3U, 348
Andervont, H. B., 269, 275, 329
Angevine, D. M., 256, 331
Angrist, A. A., 208, 210, 346
Anigstein, L., 210, 256, 352
Antopol, W., 282, 329
Arons, W. L., 233, 328
Aronson, S. M., 210, 256, 349
Asling, C. W., 196, 292, 329, 341
Austin, C. R., 222, 333
Axelrod, A. A., 238, 343
Aycock, W. L., 221, 328
Ayres, P. J., 200, 329

Bachman, R., 193, 329

Badrick, F. E., 233, 235, 329

Baeder, D. H., 209, 241, 349

Bahn, R., 212, 218, 283, 329

Bahn, R. C, 203, 206, 207, 208, 209, 212,

218, 219, 283, 329, 335, 352
Baker, B. L., 209, 211, 222, 242, 253, 254,

257, 272, 329, 330
Baker, J. R., 221, 330
Baldridge, G. D., 256, 344
Ball, W. C, Jr., 204, 335
Barber, J. K., 204, 330
Barkalow, F. S., Jr., 318, 330
Baiker, S. B., 229, 330
Barnett, S. A., 131, 185, 257, 265, 266, 269,

303, 330

Barraclough, C. A., 221, 330

Barry, M. C, 249, 330

Bartter, F. C, 201, 202, 203, 204, 330, 344,

346
Bascom, W. D., 213, 348
Bastenie, P. A., 232, 330
Bates, R. W., 240, 285, 331, 348
Bauman, E. J., 227, 345
Beer, J. R., 305, 322, 330, 334
Beetham, W. P., Jr., 223, 226, 339
Bell, G. H., 223, 224, 225, 330
Bell, J. F., 225, 253, 320, 340
Bell, W. R., 215, 334
Benfey, B. G., 214, 215, 348
Benua, R. S., 196, 197, 330, 342
Berry, M. M., 235, 351
Berson, S. A., 232, 330
Bethune, J. E., 223, 226, 339
Beyler, A. L., 292, 330
Biglieri, E. G., 202, 203, 204, 330
Bing, R. J., 244, 331
Birdsell, J. B., 97, 99, 185
Birmingham, M. K., 216, SS/
Black, M. M., 209, 353
Bloch, E. H., 206, 244, 331, 344
Block, H., 257, 279, 351
Boas, N. F., 240, 331
Bogdanove, E. M., 230, 232, 331, 341
Bogoroch, R., 235, 331
Bohanan, E. H., 257, 274, 331
Bois, P., 238, 255, 331, 349
Boring, W. O., 256, 257, 331
Borris, J. J., 206, 207, 208, 209, 212, 218,

219, 283, 352
Botkin, A. L., 233, 331
Bourliere, F., 312, 331
Bourne, G. H., 193, 194, 331
Bradlow, H. L., 206, 207, 219, 249, 331
Braidwood, R. J., 97, 99, 185
Bridgman, R. M., 242, 253, 329
Brimblecombe, R. W., 233, 235, 257, 329,

355

356

Author Index

Broadbent, D. E., 70, 1S5

Brobeck, J. R., 213, 345

Broderick, C. B., 100, 187

Brown, D., 215, 351

Brown, R. Z., 287, 2S8, 294, 295, 331

Brownell, K. A., 193, 194, 198, 223, 224,

227, 228, 261, 341
Brown-Grant, K., 231, 232, 233, 234, 235,

236, 331
Bruce, H. M., 239, 331
Bull, P. C, 322, 331, 332
Bullough, W. S., 198, 210, 269, 332
Burks, R., 215, 351
Burn, J. H., 225, 332
Burrill, M. W., 259, 340
Burrows, H., 211, 218, 219, 258, 270, 288,

289, 292, 332
Bush, I. E., 206, 207, 208, 212, 219, 332
Butt, A. J., 209, 349
Byrnes, W. W., 211, 212, 220, 257, 258,

259, 309, 332

Cain, A. J., 195, 332

Caldwell, T. M., 82, 184

Calhoun, J. B., 4, 6, 26, 32, 39, 43, 44, 81,
82, 87, 92, 93, 97, 101, 132, 138, 140,
148, 150, 161, 165, 168, 185, 273, 289,
291, 294, 298, 303, 332

Cannon, W. B., 224, 332

Cantarow, A., 234, 235, 350

Caplan, R., 230, 351

Carlson, L. D., 228, 234, 342, 352

Carpenter, C. R., 90, 91, 185

Casby, J. U., 4, 6, 26, 39, 43, 44, 81, 165,
185

(::asper. A., 204, 335, 346

Cater, O.B., 195, •1\:^,332

Cavazos, 289, 332

Cervantes, L. F., 100, /57

Chang, M. C, 288, 332

Charipper, H. A., 272, 273, 334

Chart, J. J., 199, 2m, 388

Cheek, W. R., 215, 2X^,340

Chitty, L)., 89, 179, 185, 242, 251 , 2(\\ , 264,
272, 273, 275, 276, 313, 316, 321, 322,
323, 332, 333

Chitty, H., 222, 242, 261, 2o4, 272, 333

Christian, J. J., 189, 197, 202, 210, 220,
221, 222, 225, 226, 242, 244, 246, 247,
250, 251, 255, 257, 25S, 259, 260, 2(i2,
264, 265, 266, 267, 268, 269, 27.), 271,

272, 273, 274, 275, 276, 277, 279, 280,
281, 282, 285, 286, 287, 288, 289, 290,

291, 292, 294, 295, 296, 297, 298, 299,
301, 302, 303, 304, 305, 306, 309, 310,
311, 313, 315, 317, 319, 320, 321, 322,
327, 333, 334, 349

Christianson, M., 195, 334

Chu, J. P., 239, 334

Clarke, J. R., 242, 251, 263, 272, 281, 291,

292, 294, 296, 303, 334
Clayton, G. W., 213, 215, 216, 334, 340
Cohen, A. I., 206, 331

Cole, D. F., 204, 334

Collias, N. E., 96, 186

Collins, J., 234, 235, 240, 341

Cook, E. F, 322, 334

Cornfield, J., 285, 348

Crabb, E. D., 255, 344

Crew, F. A., 268, 269, 274, 334

Critchlow, B. v., 257, 334

Cronin, M. T. I., 244, 347

Crowcroft, P., 221, 281, 286, 287, 288, 289,

290, 291, 292, 294, 308, 334
Curley, F. J, 221, 334
Cun-ie, A. R., 198, 205, 217, 248, 351

DahllxMg, B. L., 95, 96, 185

D'Angelo, S. A., 231, 232, 233, 234, 257,

258, 272, 273, 334, 350
Davidson, C. S., 212, 213, 219, 334
Davidson, J. N., 223, 224, 225, 3,30
Davis, D. E., 247, 257, 264, 265, 266, 277,

278, 298, 301, 302, 303, 306, 307, 309,

310, 311, 320, 333, 334, 335
Davis, J. O., 202, 203, 204, 335, 352
Davis, M. E., 210, 259, 335
Dawson, J., 242, 274, 335
Deane, H. W., 194, 195, 198, 201, 217, 248,

335, 340, 343
Deanesly, R., 196, 197, 334
DeBodo, R. C, 227, 335
Deevey, E. S., Jr., 97, 185
Delea, C. S., 202, 203, 204, 330
Delea, C, 204, 330
Delfa, E., 239, 343
Delost, H., 196, 336
Delost, P., 196, 197, 218, 219, 220, 283,

335, 336
Dempsey, E. W., 195, 198, 217, 238, 248,

253, 336
Deeney, R., 163, 186

Author Index

357

Denison, iM. E., 238, 239, 257, SS6

Desaulles, P., 255, 277, 346

Desclin, J., 219, 336

Despointes, R. H., 223, 224, 339

Dixon, W. J., 288, 350

Dobriner, K., 207, 249, 330, 331

Dorfman, A., 209, 336

Dorfman, R. I., 206, 207, 208, 209, 212,

218, 219, 250, 253, 259, 336, 347
Dougherty, T. F., 208, 209, 210, 211, 249,

250, 254, 255, 256, 336, 348
Doyle, D., 235, 347
Drager, G. A., 214, 336, 347
Duff, W., 99, 186
Duncan, L. D., Jr., 204, 330
Duncan, L. E., 202, 203, 204, 344

Eberhard, T., 235, 347

Egdahl, R. H., 213, .350

Ehrich, W. E., 209, 210, 238, 242, 253, 259,

260, 336, 349
Eidinoff, M. L., 249, 330
Eik-nes, K., 265, 336
Eisenstein, A. B., 200, 222, 336, 341
Elias, H., 194, 195, 196, 337
Eliot, T. S., 244, 344
Elmadjian, F., 223, 226, 227, 244, 337
Elton, C, 82, 186, 322, 323, 337
Emlen, J. T., 271, 281, 284, 285, 287, 288,

289, 293, 296, 297, 298, 345, 350
Engel, F. L., 209, 244, 337
Erachkoo, ()., 198, 225, 227, 337
Erickson, A. B., 322, 337
Ermans, A. M., 232, 330
Errington, P. L., 298, 317, 337
Ershoff, B. H., 234, 337
Evans, C. A., 321, 322, 339, 340
Evans, E. S., 228, 337, 349
Evans, H. M., 199, 227, 228, 346
Everett, J. W., 257, 258, 337, 345

Fainstat, T. D., 210, 221, 259, 338
Farrell, G. L., 199, 202, 203, 204, 206, 20s,

209, 213, 215, 329, 337, 338, 350
Feldman, J. D., 195, SS5
Finerty, J. C., 237, 270, 284, 292, 341, 347
Finland, M., 208, 209, 210, 211, 255, 256,

343
Fitch, H. S., 317, 338
Fleming, R. B., 202, 203, 209, 213, 338
Flexner, L. B., 195, 338

FHck, S., 309, 338

Flueckiger, E., 232, 338

Flyger, V., 306, 320, 332

Foenss-Bech, P., 209, 213, 344

FoUey, S. J., 275, 328

Foot, E. C., 289, 343

Ford, E., 232, 328

Fortier, C., 213, 215, 216, 257, 338, 341

Frank, F., 225, 253, 272, 280, 298, 320, 321,

322, 323, 338
Eraser, F. C., 210, 221, 259, 338
Fredericson, E., 47, 186
Fredrickson, D. S., 227, 338
Freedman, H. H., 234, 238, 338
French, L. A., 213, 350
Fruit, K., 21^,345
Furth, J., 206, 212, 218, 283, 329

Gaddum, J. H., 223, 224, 225, 226, 227,

338
Gallagher, T. F., 206, 207, 218, 219, 249,

330, 331, 338
Gann, P. S., 203, 204, 330
Ganong, W. F., 338
Gaunt, R., 199, 201, 338
Gersh, I., 194, 196, 339
Gerwing, J., 236, 237, 339
Geschwind, I. I., 209, 213, 344
Gierlach, Z. S., 235, 341
Giroud, C. J. P., 199, 200, 202, 339, 350
Glaubach, S., 210, 259, 339
Glazer, N., 103, 186
c;iick, D., 196, 339
(iodfrey, G. K., 321, 322, 323, 339
Godwin, J. T., 278, 350
Goldberg, R. C, 232, 339
Goldfien, A., 223, 226, 339
Goldstein, M. S., 209, 223, 347
Goodkind, M. J., 204, 335
Goodman, H. M., 227, 339
Gordon, A. S., 211, 234, 238, 249, 251, 252,

254, 256, 272, 273, 334, 338, 339
Gordon, R. S., Jr., 227, 338
Gorski, R. A., 221, 330
Gould, R. P., 200, 329
Gowenlock, A. H., 201, 350
Grad, B., 240, 339
Grant, J. K., 198, 205, 217, 248, 351
Gray, G., 228, 342
Gray, I., 223, 226, 339
Gray, S. J., 242, 253, 339

358

Author Index

Green, J. D., 213, 214, 339

(iiven, R. G., 225, 253, 320, 321, 322, 83.9,

340
Greene, R. U., 259, 340
Greep, R. ()., 195, 198, 201, 217, 232, 248,

335, 339, 340
Greer, M., 230, 351

Greer, M. A., 232, 234, 257, 258, 340, 352
Griffin, A. C, 213, 352
Grollman, A., 194, 195, 196, 338, 339
Gross, F., 199, 201, 340
Gross, J., 231, 341
Grosvenor, C. E., 258, 275, 340
Guettirigor, R. C, 95, 96, 185, 186
Guillemin, R., 213, 215, 216, 334, 340, 348

Hagen, P., 223, 340

Halberg, F., 285, 352

Halkerston, I. D. K., 257, 331

Hall, C. E., 202, 284, 341, 348

Hall, C. S., 89, 186

Hall, 0., 284, 341

Hallowell, A. I., 98, 186

Halmi, N. S., 230, 232, 331, 341

Hamilton, W. J., Jr., 308, 317, 320, 341

Hammond, J., Jr., 257, 341

Hamolsky, M. W., 235, 341

Hardin, G., 77, 186

Harris, G. W., 213, 231, 232, 233, 234, 235,

236, 331, 341
Harrison, R. G., 194, 195, 196, 332, 341
Hart, P., 279, 34I
Hartcroft, P. M., 200, 336, 341
Hartman, F. A., 193, 194, 198, 223, 224,

227, 228, 261, 341
Hatfield, D. M., 317, 341
Hauser, E. A., 209, 349
Hayashida, T., 209, 213, 344
Haymaker, W., 257, 329
Hayne, D. W., 81, 186
Hearn, W. R., 215, 216, 340
Hellman, K., 234, 235, 240, 341
Hellman, L., 218, 219, 338
Hess, M., 237, 239, 270, 284, 292, S4I, 347
Hill, R. T., 220, 341, 342
Hoffmann, R. S., 314, 317, 342
Hofmann, F. G., 206, 309, 342
Hollander, V. P., 206, 209, 347
Holman, J., 204, 335
Holmes, W. M., 196, 197, 342

Holzhauei', M., 223, 224, 225, 226, 227, 338
Hope, J. M., 223, 226, 227, 244, 337
Hopsu, v., 225, 227, 337
Horn, E. H., 238, 342
Housholder, D. E., 215, 216, 3^0
Howard, E., 195, 196, 197, 212, 217, 330,

342
Hsieh, A. C. L., 228, 342
Hudyma, G. M., 238, 349
Huggins, C, 292, 342
Hughes, E., 306, 342
Hume, D. M., 216, 342
Hungerford, G. F., 209, 213, 344
Hurst, v., 235, 240, 342
Hutchinson, G. E., N., 77, 186

Ingalls, T. H., 221, 329, 334, 342

Ingle, D. J., 208, 209, 211, 253, 257, 329,

330, 342
Irving, J. T., 210, 342
Isler, H., 238, 343
Ivy, A. C, 259, 340

Jacot, B., 213, 284, 292, 343

Jaffe, H., 233, 235, 352

Jameson, E. W., Jr., 307, 343

Jenkins, D., 285, 351

Jennings, H. H., 154, 186

Jensen, E. V., 292, 342

Jensen, H., 235, 341

Jensen, W. L., 320, 335

Jones, G. E. S., 239, 343

Jones, I. C, 193, 195, 196, 197, 199, 200,
201, 206, 207, 208, 209, 213, 215, 217,
218, 219, 254, 261, 283, 334, 343, 347,
352

Josimovich, J. B., 195, S.^5

Kaljat, C., 96, 186

Kalas, J., 241, 349

Kalela, O., 221, 300, 308, 311, 314, 343

Kalter, H., 210, 221, 259, 338, 343

Kappas, A., 218, 219, 335

Kass, E. H., 208, 209, 210, 211, 255, 256,

343
Katonah, F., 208, 210, 346
Katsh, G. F., 256, 339
Kedda, L., 206, 212, 218, 283, 329
Kelsall, M. A., 255, 344
Kemp, C, 233, 235, 352

Author Index

359

Kendrick, M. I., 208, 210, 211, 255, S4S

Keuning, F. J., 211, 255, 344

Kew, M., 99, 186

Kirschner, L., 232, 233, 250, 346

Kligman, A. M., 256, 344

Kliman, B., 203, 204, 336, 352

Knigge, K. M., 194, 196, 206, 344, 348

Knisely, M. H., 244, 344

Knobil, E., 227, 339

Kraatz, C. P., 239, 344

Kracht, J., 235, 344

Kraintz, L., 232, 233, 250, 346

Kramer, J., 209, 344

Kuhl, W. J., Jr., 232, 344

Kurlents, E., 216, 331

Kurzok, R., 346

Ladman, A. J., 195, 343

Laidlaw, J. C, 285, 351

Lamson, E. T., 223, 226, 227, 244, 337

Lanman, J. T., 195, 344

Laqueur, G. L., 215, 257, 258, 338, 344

Larson, C. L., 225, 253, 320, 340

Leathern, J. H., 288, 344

Leblond, C. P., 238, 240, 339, 343

LeMaistre, C. A., 210, 279, 344

Le Munyan, CD., 257, 258, 260, 274, 275,

290, 294, 298, 313, 315, 322, 333
Leopold, A. S., 24, 186
Leslie, P. H., 242, 261, 264, 272, 333
Levy, A. C, 227, 344
Li, C. H., 199, 209, 211, 212, 213, 219, 227,

257, 292, 329, 330, 344, 346
Lichtlen, F., 199, 201, 340
Liddle, G. W., 202, 203, 204, 330, 344
Lieberman, S., 206, 212, 219, 351
Lillie, R. D., 253, 344
Lipner, H. J., 230, 341
Lipscomb, H. S., 213, 215, 216, 340
Lipsett, M. B., 218, 219, 338
Long, D. A., 236, 237, 339
Long well, B. B., 206, 347
Lord, R. A., 278, 344
Lostroh, A. J., 199, 209, 212, 213, 219, 289,

344, 345
Louch, C, 284, 285, 345
Louch, C. D., 252, 281, 284, 285, 290, 292,

293, 294, 296, 297, 304, 305, 345
Lundgren, M. M., 209, 210, 256, 343

Lutwak-Mann, C, 222, 259, 273, 288, 289,

309, 310, 345
Lyons, W. R., 209, 213, 344

McCain, R., 24, 186

McCally, N., 202, 203, 204, 209, 213, 329,

338
McCann, S. M., 213, 214, 257, 258, 344,

345
McCarthy, H. H., 209, 344
McDermott, W., 210, 279, 344
McDonald, R. K., 215, 345
Mac Lulich, D. A., 319, 345
Mc Nutt, S. H., 196, 352
McPhail, M. K., 196, 197, 217, 282, 345
Mahler, R. F., 201, 350
Manly, B. M., 257, 330
Mann, T., 288, 289, 345
Mall, R., 306, 342
Magsood, M., 238, 345
Marine, D., 227, 345
Markee, J. E., 257, 258, 345
Marmorston, J., 215, 348
Marrian, G. F., 222, 257, 345
Martin, E. P., 317, 345
Martorano, J. J., 253, 345
Mason, J. W., 244, 253, 267, 268, 346
Meier, R., 255, 277, 346
Meites, J., 235, 347
Melampy, R. M., 289, 332
Melby, J. C, 213, 350
Menkin, V., 254, 346
Merrill, P., 232, 233, 250, 346
Meyer, R. K., 211, 212, 220, 252, 258, 259,

284, 285, 305, 330, 332, 345, 350
Meyers, V. M., 209, 344
Miller, R. A., 196, 346
Mills, I. H., 202, 203, 204, 330, 346
Mirskaia, L., 268, 269, 274, 334
Money, W. L., 232, 233, 234, 237, 250, 346
Moon, H. D., 199, 209, 212, 213, 219, 227,

334, 344, 346
Moore, W. W., 257, 258, 346
Mora, P. M., 97, 187
Moreno, J. L., 154, 186
Morgan, B. B., 196, 352
Morin, G., 228, 346
Morse, A., 195, 335
Mosier, H. D., 195, 346

360

Author Index

Moss, M.L., 210, 211,3-^6
Mulinos, M. G., 257, 272, 346

Nalbanov, A. V., 257, 258, 346

Nelson, D. H., 206, 207, 208, 346

Nelson, D. M., 216, 348

Nelson, M. M., 222, 257, 260, 273, 350

Nezamis, J. E., 255, 348

Nichols, J, 202, 346

Noble, R. L., 208, 346

Norkus, M. G., 247, 353

Norton, H. W., 257, 258, 346

Nowell, N. W., 215, 216, 346

Oastler, E. G., 198, 205, 217, 248, 351
Ochs, M. J., 196, 339
O'Donnell, U. J., 198, 205, 217, 248, 351
Odum, E. P., 317, 346

Palmer, J. G., 216, 348
Parkes, A. S., 222, 257, 345
Parmer, L. G., 208, 210, 346
Parsons, F. M., 292, 34^
Paschkis, K. E., 234, 235, 347, 350
Patric, E. F., 67, 85, 186
Pauly, J. E., 194, 195, 337, 347
Pearson, O. H., 218, 219, 338
Pearson, O. P., 175, 186
Pechet, M. M., 204, 335
Peron, F. G., 206, 347
Perry, W. F., 232, 347
Peterson, R. E., 203, 204, 335, 352
Philbrook, F. R., 221, 342
Phillips, J. G., 195, 352
Piletta, P., 199, 200, 33i>
Pillsbury, D. M., 256, 344
Pincus, G., 288, 332
Pitelka, F. A., 307, 319, 323, 347
Pitt-Rivers, R., 236, 237, 339
Plotz, E. J., 210, 259, 335
Poindexter, H. A., 257, 347
Pollard, M., 210, 256, 347
Pomerantz, L., 257, 272, 346

Poore, W., 206, 209, 347

Porter, J. C., 214, 215, 347

Preble, E. A., 319, 347

Price, D., 288, 289, 345

Pronove, P., 204, 330

Prudovsky, S., 278, 335

Pruitt, R. L., 215, 348

Puntriano, G., 235, 347

Raeisaenen, L., 198, 227, 337

Rail, J. E., 229, 230, 231, 238, 347, 351

Ramey, E. R., 209, 223, 225, 227, 344, 347

Ramey, C. G., 242, 253, 339

Rand, M. J., 225, 332

Randall, H. T., 244, 347

Ransom, R. M., 221, 330

Ranson, S. W., 82, 185, 186

Ratcliffe, H. L., 202, 225, 244, 250, 271,

333, 347
Rausch, R., 318, 320, 322, 323, 347
Rauschkolb, E. W., 199, 202, 203, 208, 209,

213, 338
Rawson, R. W., 229, 230, 231, 232, 233,

238, 250, 346, 347, 351
Read, C. P., 257, 277, 278, 335
Read, H. C., 196, 197, 217, 282, 345
Rebell, G., 256, 344
Reed, C. A., 97, 99, 185
Rees, R. J. W., 279, 341
Reichlin, S., 231, 232, 233, 234, 235, 236,

•2-il,331,347
Reif, A. E., 206, 347
Reinhardt, W. O., 209, 213, 344
Reineke, E. P., 237, 238, 345, 352
Reinhardt, W. O., 292, 329
Reiss, J. M., 233, 235, 329
Reiss, M., 233, 235, 257, 329, 331
Rennels, E. G., 214, 270, 284, 292, 34I, 347
Renzi, A. A., 199, 201, 338
Retzlaff, E. G, 268, 269, 274, 277, 315, 347
Richards, J. B., 215, 3^5
Richardson, H. C, 213, 352
Richardson, J. A., 223, 348
Richter, C. P., 195, 198, 202, 348
Riesman, D., 163, 186
Riney, T., 24, 186
Ring, G. C., 234, 348
Rioch, D., 257, 258, 344
Robert, A., 255, 348
Robinson, H. J., 210, 254, 256, 348
Roby, C. C., 196, 343
Rogers, P. V, 195, 198, 202, 348
Rosasco, E. M., 80, 85, 187
Rosemberg, E., 257, 258, 285, 344, 348
Rosenberg, L. L., 349
Rosenfeld, G., 213, 253, 321, 348
Rothballer, A. B., 214, 348
Rowe, F. P., 221, 281, 286, 288, 289, 290,
291, 292, 294, 308, 334

Author Index

361

246,
248,

Royce, P. C, 199, 202, 203, 208, 213, .338
Rumsfeld, H. W., Jr., 214, 215, 347
Runfret, A. P., 213, 3S3

Sacks, J. G., 82, 185

Saffran, M., 214, 215, S^S

Sakiz, E., 218, 348

Salter, W. T., 230, 348

Samuel, L. T., 216, 348

Sandberg, A. A., 216, 3^5

Sanders, R. D., 96, 186

Santisteban, G. A., 208, 211, 250, 348

Sawyer, C. H., 257, 258, 334, 337, 345

Sayers, G., 195, 197, 198, 215, 243,

248, 348, 351
Sayers, M., 195, 197, 198, 243, 246,

348
Scarborough, H., 223, 224, 225, 330
Schaefer, E. S., 143, 155, 162, 186
Schairer, M. A., 257, 329
Schally, A. V., 214, 215, 348
Schapiro, S., 215, 348
Scharrer, B., 214, 348
Scharrer, E., 214, 345
Schindler, W. J., 194, 206, 348
Schmidt, L. H., 256, 349
Schmitenloew, C. G., 224, 349
Schneebeli, G. L., 208, 209, 210, 254,

256, 336
Schneirla, T. C., 89, 186
Schorger, A. W., 96, 186
Schreiner, L. H., 257, 258, 344
Schuler, W., 255, 277, 346
Schwartz, H. S., 229, 330
Scott, J. C., 242, 261, 264, 272, 333
Scudder, J., 244, 349
Sealander, J. A., Jr., 273, 349
Seifter, J., 209, 210, 238, 241, 242, 253,

260, 336, 349
Seligman, A. M., 194, 335
Selye, H., 192, 208, 209, 211, 212, 213,

220, 236, 238, 242, 243, 244, 250,

254, 255, 256, 257, 258, 260, 261,

274, 275, 276, 284, 288, 292, 331,

349
Share, L., 216, 349
Shaw, J. H., 195, 201, 335
Shipley, E. G., 211, 212, 220, 257, 258,

309, 332

259.

217,
253,
264,

343,

259,

Shipley, R. A., 206, 208, 212, 218, 219, 250,

259, 336
Shiras, G., 96, 186
Shorr, E., 209, 353
Shwartzman, G., 210, 256, 349
Sideman, M. B., 209, 213, 344
Simay, Kramer, M., 209, 344
Simpson, M. E., 199, 222, 227, 228, 257,

260, 273, 346, 349, 350
Simpson, S. A. S., 200, 329
Sloviter, 239, 331

Slusher, M. A., 215, 216, 248, 349

Smelser, G. K., 240, 349

Smelser, J., 346

Smith, A. L., 210, 254, 256, 348

Smith, E. K., 220, 349

Smith, J. D., 213, 215, 216, 340

Smith, R. W., Jr., 257, 348

Snyder, R. L., 309, 349

Sobel, H., 215, 348

SoUman, T., 229, 230, 231, 238, 349

Sonenberg, M., 230, 231, 238, 347

Southern, H. N., 89, 179, 185, 289, 349

Southwick, C. H., 252, 265, 281, 285, 287,

288, 289, 293, 294, 295, 296, 297,' 3.50 '
Spalding, M. H., 195, 343
Speirs, R. S., 211, 252, 285, 350
Spirtos, B. N., 230, 341
Squires, W. L., 256, 349
Srebnik, H. H., 222, 257, 260, 273, 350
Stachenko, J., 199, 200, 202, 339, 350
Stack-Dunne, M. P., 195, 213, 332
Stadler, J. B., 216, 3^5
Stanbary, S. W., 201, 350
Steinberger, E., 288, 350
Stevens, C. E., 234, 350
Strecker, R. L., 271, 281, 287, 288, 289,

293,296,297,298,350
Stewart, J. O., 97, 186
Stoner, R. D., 278, 350
Story, J. L., 213, 350
Sturgis, S. H., 258, 350
Sundberg, R. D., 211, 350
Sunderman, F. W., 234, 350
Swanson, H. E., 228, 229, 234, 350
Sweat, M. L., 206, 350
Swift, E., 96, 186, 187
Sydnor, K. L.. 215, 351
Symington, T., 198, 205. 217. 248, 351
Szego, CM., 292, 330

362

Author Index

Tait, J. F., 200, 339

Tamura, Y., 196, 198, 351

Tanaka, R., 318, 351

Tata, J. R., 229, 351

Taubenhaus, IM., 208, 209, 210, 254, 351

Tenney, A., 232, 338

Tevis, L., Jr., 24, 186

Tew, J. T., 233, 331

Thomas, L., 210, 254, 255, 256, 351

Thompson, D. Q., 318, 351

Thorn, G. W., 285, 351

Thorn, N. A., 241, 351

Timiras, P., 235, 331

Tobach, E., 89, 186, 257, 279, 351

Tompsett, R., 210, 279, 344

Traum, R. E., 231, 232, 233, 234, 257, 258,

334
Turner, C. W., 235, 240, 258, 275, 340, 342
Tyler, F. H., 216, 348
Tyndale-Biscoe, C. H., 322, 351

Uhrich, J., 271, 351
Uotila, U. U., 238, 336

Vander Laan, W. P., 230, 351

Van der Slikke, L. B., 211, 255, 344

Vande Wiele, R., 206, 212, 219, 351

Van Middlesworth, L., 235, 351

Venning, E. H., 200, 339

Verzar, F., 232, 338, 352

Vetulani, T., 276, 352

Vidovic, v., 232, 352

Visscher, M. B., 285, 352

Von Elder, C, 231, 331

Von Elder, U. S., 223, 234, 236, 352

Von Foerster, H., 97, 187

Wagner, H. N., Jr., 215, 345
Walker, D. L., 256, 257, 331
Wallach, D. P., 237, 352
Waring, H., 196, 197, 352
Warner, L., 244, 344
Watt, L. J., 309, 352

Weaver, J. A., 211, 249, 250, 352

Webb, W. L., 32, 80, 82, 85, 185, 187

Weber, A. F., 196, 352

Welch, A. D., 223, 340

West, C. D., 218, 219, 338

Wexler, B. C, 213, 352

Whitney, D. M., 210, 256, 352

Whitten, W. K., 275, 352

Whittle, P., 322, 352

Whyte, W. H., 163, 187, 198, 205, 217,

248, 351
Wiese, V. K., 215, 345
WiUiams, R. H., 233, 235, 352
Williams, R. M., 322, 351
Williamson, H. 0., 255, 257, 277, 334
Wilson, B. R., 210, 256, 347
Wilson, H., 206, 207, 208, 209, 212, 218,

219, 283, 329, 352
Wolff, F., 232, 339
Woodbury, D. M., 213, 216, 352
Woods, E. F., 223, 348
Woods, J. W., 231, 232, 233, 234, 236, 341
Woodward, P., 213, 345
Wooton, R. M., 247, 353
Wright, A., 195, 343, 352

Yalow, R. S., 232, 330

Yamada, T., 234, 352

Yankopoulos, N. A., 203, 204, 335, 352

Yatsu, F. M., 202, 203, 209, 213, 338

Zalesky, M., 196, 353
Zarafonetis, C. J. D., 241, 349
Zarrow, M. X., 238, 239, 257, 336
Ziff, M., 232, 344
ZUeli, M. S., 223, 226, 339
Zimmerman, C. C, 100, 187
Zimmermann, K., 311, 353
Zippin, C, 26, 187
ZuRhein, G. M., 257, 331
Zweifach, B. W., 209, 244, 353
Zwemer, R. L., 247, 353

Subject Index

AAF, 216
Aborigines

Australian, 99
ACTH, 197, 200, 209, 214, 220, 232, 243

assajdng, 248

basal release rate, 215

A-1 fraction, 203

endogenous, 249

exogenous, 249

inhibiting the release, 215

mice, 237

rabbits, 237

rats, 237

regulation of the release, 213

release, 216, 217

release fraction, 215

secretion, 237, 258

secretion of ADH, 241

secretion of aldosterone, 202

sj-nthesis, 216
Activity

hyperactivity, 9

minimal, 9

normal level, 16

normal nocturnal period, 16

24-hour rhj-thm, 9

endocrine glands of, 192

evolutionary, 123
Adaptive mechanisms, 190

atrophy, 244
Adaptive responses

measurements of the endocrine, 242
Adenohypophysis, 214, 233

arterial supply, 213
ADH, 201, 214, 241
Adirondacks, 55
Adjustment

learned, 89
Adrenal

androgen, 206

effect on nursing young, 276

atrophy, 238

ascorbic acid, 248
depletion, 192

Callothrix nrgentata, 196

cat, 195

Cerropithent^, 195, 196

chimpanzee, 195

cholesterol, 192, 193, 248

circulation, 195

Ciiellus tridecemlineatus, 196

Cleihriononiys glareolus, 196

colobus monkej', 196

Colohus polykomos, 196

cortex, 191, 192, 193, 217, 245, 246

hormones, 199

human beings, 212

lipid vacuoles, 194

measurements of increased function,
245

Microsorex hoyi, 246

parenchjina, 195

Sorex cinereus, 246

Sorex dispar, 246

Sorex fumeus, 246

Sorex palustris, 246

secretorj' activity, 248

sex steroids, 219

sex steroidal activity, 218

sudanophilia, 197

X-zone, 217

X-zone of mice, 282

zona fasciculata, 194

zona glomerulosa, 194, 199

zona reticularis, 194

zonation of, 194
demeduUation, 228
cow, 196
Crocidura, 196
gland, 192

effects of h\T5ophysectom\- on, 196

fetal zones of the, 195

general morphology of, 193

ground squirrels, 196

hamsters, 196

histologic measurements, 247

histology, 195

humans, 195, 196

hypertrophj', 238, 246

lipid content, 193

Loris, 196

Macaca mulatta, 195

macaque, 195, 196

marmoset, 196

363

364

Subject Index

Adrenal — continued

medulla, 193, 198, 223, 224, 245, 246
hypertrophy, 199
Microsorex hoyi, 246
Sorex dispar, 246
Sorex fumerus, 246
/Sorex cinereus, 246
Sorex palustris,
tumors, 199
Mesocricetus auratus, 196
mice, 196

Microtus agrestis, 196
Microtus arvalis, 196
monkeys, 195, 196
Ornithorhynchus, 195
Orydolagus, 195
Pan, 195

Perodicus potto, 195
pituitary, 197
Pilymys, 196
potto, 195
primates, 195
rabbits, 195

laboratory rabbit, 195
rats, 195, 196, 198
Norway rat, 195
wild rats, 195,
Rattus alexaadrinus, 195
Rattus norvegicus, 195
regulation of hormones, 213
reticularis, 198
samples, 247
slow loris, 196
Sorex araneus, 196
starvation, 196
steroids, 219
steroid secretion, 192

androgenic metabolites, 219
synthesis, 219
Tachyglossus, 195
transplantation, 21 X
ungulates, 196
vascularization, 195
weight, 192, 245
factor, 216
fighting, 264
variation, 246
X-zone, 196
Adrenalectomized animals, 201
Adrenocortical activity
indices, 21 1

in natural populations, 300
sudanophilia, 248
Adrenocortical-gonadal interrelationships,

218
Adrenocorticotropin, 196, 197, 205, 213,
220, 222
release, 216
secretion, 215

secretion of aldosterone, 202, 205
inaccuracy of determining, 308
Age

determination by teeth, 311
Alarm stimulus, 192
Alberta, 54
Albino mice {see also mice)

trichinosis, 278
Aldosterone, 199
release, 217, 243
secretion

decrease in blood volume stimulates,

204
and pulse pressure, 204
stimulation by decrease in blood

volume, 204
regulation, 202, 203
renal hemodynamics, 204
stimulation of, 203
Aliesterase, 200
Alley

emotional activity, 177
Alouaita palliata, 90
Alpha individuals, 59

central, 59
Amines

pressor, 225
Amine oxidases, 223
Androgen, 217

adrenocortical, 219
cortical, 220
precursors, 206
Androgenic steroids
house mice, 220
man, 220
rats, 220
Androstenedione

gonadotropic activity, 220
Animal {see also specific animal names),
239
laboratory, 262
.4 ntelioinys sni ithii

inhibition of reproductive function, 318

Subject Index

365

Antibody formation, 255, 278

inhibition, 210
Antibodies

stored, 21 1
Antidiuretic hormone, 201
Antigonadotrophic activity, 211
Area

determination, 84
Ascendency

dominant, 55

psychological, 55
Ascorbic acid, 217, 237

adrenal, 216, 248

factor, 216
Associations

intraspecific, 64
Audition, 38

Australian aborigines, 97
Avoidance

of a strange field, 17
Awareness

of self, 150
AWF, 216

Bainbridge

cardiac reflex, 230
Balance

electrolyte, 192

fluid, 192
Barter Island, Alaska, 318
Basic group size, 113, 125
Basic number, 3
Bats (see also generic names such as

My Otis), 193
Behavior, 15, 158

alterations, 94

cultural origin, 124

exploratory, 175

genetic origin, 124
Behavioral sink, 93

development by Norway Rat, 92
Beta individuals, 59
Bivariate normal distribution, 20

function, 4, 39
Blarina, 29, 30, 31, 32, 36, 38, 52, 61, 64,
66, 68, 69, 72, 75, 76, 78, 80

population, 305
Brain

subcortical portion, 15

British Columbia

Indians of, 99
Broadbent's theory, 70, 72
Brown fat, 194

Cages

"life-space," 179
Callothrix argentata, 196
Capture

cumulative probability, 21
Carbohydrate-active corticoids

suppressive effects, 209

Cl9

steroids, 206, 217, 219

ketosteroids, 205
Caribou, 122
Catch

comparative, 29, 30, 31, 32

decline, 56

Huntington WildUfe Forest, 29, 30

Maine (1950) and Maryland (1953), 31

Peromyscus and Clethrionomys, 32
Cattle, 194
Cells

inflammatory-, 210
Census

30-day, 53
Central nervous system, 191

integrative role, 242

memory store, 134
Cercopithecus, 195, 196
Cervus nippon, 323

population density, 306
Chadwick Woods, 76
Changes

environmental, 190

liver, 260

pancreas, 260

parotid, 260

visceral degenerative, 260
Chimpanzee, 195
Chromaffin

cells, 198

tissue, 195, 223, 225
Circulatory collapse, 244
Citellus tridecemlineatus, 196
Clethrionomys, 27, 30, 31, 32, 33, 34, 35, 36,
52, 53, 54, 55, 61, 67, 68, 69, 70, 75,
78, 80, 83, 85, 88, 314, 323

inhibition of growth and reproduction,
316

366

Subject Index

Clethrionomys — continued

relationship between population densitj'

and reproduction, 310
reproductive function, 314
Clethrionomys glareolus, 196
Clethrionomys rujocanus, 311
population density, 308, 313
territory, 300

inhibition of reproductive fun(^tion, 318
Close confinement

effects of, 271
Codominants, 67
Cold

adaptation, 228
exposure, 234
Colobus polyl'omos, 196
Colony

compact, 86
evolution of, 88
Howler monkeys, 90, 92
Norway rat, 92
formation, 87
stability, 89
Communication, 38, 125

low frequency, 59
Communication
reduction, 67
vocal, 70
Communication constant, 133
Communication function (/*), 126
Community, 2, 8, 32, 35, 49, 55, 74, 78
Blarina: Peromyscus, 73
one-species, 73
psychological apex, 79
small-mammal, 75
Sorex: Clethrionomys, 73
subordinate member, 78
Competition, 78
food and social, 271
social, 262
"Competitive exclusion," 77
Configuration

active rejection of new, 179
physical nonsocial, 183
seeking of new, 178
Conformity, 162
Constellation, 85
adjoining, 60

derivation of compact colonies from, 86
expected viability in, 61
formation, 57

individuals forming, 63
loosely knit, 88

number of individuals forming, (il
periphery, 60
theoretical, 62
Contact

between individuals, 105
decrease of responsive-responsive, 107
encounters, 106
frequency, 38, 106, 139
frequency of responsive-responsive, 106
individual, 106
refractory, frequency of, 105
responsive-refractory, 103
responsive-responsive, 103
social perception, 136
variables determining, 101
Copulation pressure

criticized, 293
Cortex {see also adrenal cortex), 72
adrenocorticotropin, 197
cellular hj^perplasia, 213
fasciculata, 198
h^-pertrophy, 213
Cortical glomerulosa, 216
Cortical hormones

relationships, 212
Cortical hyperplasia

cortical, 218
Cortical mass

increase, 135
Cortical stroma
carnivores, 194
rodents, 194
Corticoids

carbohydrate-active, 220, 245
secretion, 222
Corticosteroids

release, 213
Corticosterone, 205, 206
mice, 208
rats, 208
Corticotropin peptides, 200
Corticotropin-releasing factor, 214
Cortisone, 196, 197, 205, 232

tetratogenic effects in mice, 210
Cotton rat, {see also Sigmodon), 323
Cow, 196
Creativity, 162
Crocidura, 196

Subject Index

367

Crocidura russula, 197
Cultural disturbance, 99
Cynomys ludovicianus, 122
Cytochrome, 230
Cytochrome oxidase, 223, 230

d-gene, 134

differentiation in a similarity rank
hierarchy, 142

dominant, 128, 141

recessive, 128, 141

transformation of recessive, 159
DCA, 205
DOC, 205
DOCA, 205
Deer

adrenals, 305

Northern Wisconsin, 95

white-tailed, 95

yarding, 95
Deer mice (see also Peromyscus), 322
Dehydroepiandrosterone, 206, 212
Density

actual, 53

high, 53

increase, 138

low, 53

low spring, 54
Density function, 4, 6, 44, 45, 47

Cartesian coordinates, 4
Deoxycorticosterone, 196, 199
Diabetes, 241

Dicrostonyx {see also lemming) 323
Dicrostonyx groenlandicus

population cycles, 318
Diestrus, 239
Differences

morphologic, 190
Diffuse motor activity (DMA), 175
Digestive organs

effect of stimuli, 253
Diiodotyrosine, 230
Disease resistance

in dense populations, 319
Distance

between neighbors, 44

methods of calculating data, 44
Distribution

spatial, 58, 59
uniform, 81

Dog

adrenal medulla, 225

dominant, 265

hypophysectomized, 233

release of radioiodine, 233

subordinate, 265

secretion of aldosterone, 203
Dominance

psychological, 74, 75, 77, 79

Effector, 134
Eimeria stiedae, 322
Electrolyte balance, 201
ll-Deoxycorticosterone, 205
ll-Deoxy-17-hydrocorticosterone, 205
IIB-Hydroxy testosterone, 211
11-0H4AD, 206, 211
Elk, 122
Embrj^o

resorption of, 310
Emotional activity alley, 181
Encounters, 106
Endocrine

adaptive mechanisms, 191

adaptive responses, 189

function

inhibition, 221

organs

relationships, 227
Endogenous corticosteroid seci-etion, 2 1 1
Environment

abnormal, 108

external, 190

impact on, 43

of all individuals, 44

initial hyperactivity in a strange, 15

internal, 190, 261

optimum uniform utilization, 60

shifts, 190

structured, 11

two-dimensional, S
Environmental factors, 190
Eosinopenia, 252
Eosinophil

as index of adrenocortical activity, 211

counts, 252, 285

as index of population density, 284
Epinephrine, 209, 216

action, 215

effects, 223, 224

secretion, 226, 227

368

Subject Index

Estradiol, 208
Estrus, 239

changes, 258
Evolution, 51

cultural, 98, 122

of a new species or genus, 87

of compact colony, 88

of a filtering device, 134

probable cause, 74

man's social, 184

social, 3, 137

P^acilitation, 151
Factor

communication-inhibiting, 112, 139

contact-Winding, 112, 139
Families

human, 100
F:B ratio

monkeys, 206

mice, 206

rabbits, 206

rats, 206
Feedback mechanisms

physiologic, 190
Female

asexual, 88

estrus, 89
Ferguson activity allc}', S, 18
Fertility

density, 290

female, 239

male, 239

mouse, male, 290
Fetus

congenital defects, 210

growth and development, 259

prenatal mortality, 293

resorption, 95
Fighting

as a measure of social rank, 267

as form of stress, 264
Filter

Broadbent, 71

neural, 71, 74

system, 77
Food

limiting population growth, 298

Food

scarcity, 98
Food supply

effects on population growth, 296
Frequency

of meeting of responsive individuals, 112

of meeting of responsive-nonresponsive
individuals, 112
Frustrations

optimum, 119
FSH, 259
Functions

interaction, 113

Gene

d-, 128

frequency, 125

mutant, 125
Gene pool, 125
Gamma individuals, 59
Gastrointestinal tract

effect of stimuli, 253

integrative role, 242
Glomerular filtration rate, 201
Glomerulotropin, 203, 205

depression, 204

secretion, 204
Glucagon, 240
Glucocorticoids, 245
Glucose utilization

inhibition, 209
Glj'cogen

exhaustion, 253

showshoe hares, 253

voles, 253

levels

factors altering, 253
Glycosuria, 209
Gonadal endocrine cells

steroids, 219
Gonadectomy, 219
Gonadotropin, 217, 219

inhibition, 211, 220

production, 211

secretion

suppression, 257, 309
Granulation

effect of corticoidson, 254

Subject Index

369

Gravidum, 309
Ground squirrels

thirteen-lincd, 196
Group dynamics

formulation, 101
Group size

basic, 3

change, 125

effect on reproduction, 91

optimum, 2

saltatorial changes in the basic, 122

satisfaction and frustration as a function
116
Growth, 254

effects of increased population density,
276

house mice, 299

population, 1S9
Guinea pig

harderian glands, 240

hydrocortisone, 236

medullary hypertrophy, 227

pituitary-adrenocortical activity, 273
Guinea pig

pneumococcal infection, 256

streptococcal infection, 256

thyroid activity, 236

tuberculosis infections, 256

Habitat, 42

marginal, 45

one-dimensional, 8

unstructured one-dimensional, 9

unsuitabilitj', SO
Hamster, 196

golden, 194

poliomj-elitis, 256

sudanophilia, 248
Hare (see also Lepus), 319, 323

mortality, 322

population density and parasitism, 322

relationships between population den-
sity and reproduction, 310
Harvest mice, 6
Hematopoiesis, 251
Hibernating gland, 194
Hierarchical situation, 263

Hierarchy formation

reductions in velocity of, 153
Histamine, 216
Homeostasis
circulator}', 192
electrolyte, 202, 205
fluid, 202, 205
Home range, 20, 25, 36, 38, 42, 164

behavioral origins of the bivariate

normal, 8
bivarate normal, 4
center, 4, 23, 39, 42, 43, 47, 57, 58, 72

actual, 64

"ideal" interval between, 45

optimum interval, 51

uniform distribution, 59
clumping, 67
contracted, 54
enlargement, 76
expansion, 68, 79
fixed, 77

inhibition, 53, 78, 80
intraspecific differences in size of, 75
mutual inhibition of, 79
overlapping, 67
periphery, 51
reduction, 70
relative, 165

relative, of constellation members, 166
schematic

for Blarina, 65

for Peromycus, 65

for Pitymys, 66
sigma, 5, 6
size, 53

social inhibition, 70
travel-path, 24
Homo sapiens, 57
Hormone

antidiuretic, 241
anti-inflammatory, 208
fascicular action, 208
regulation of secretion of the fascicular,
213
Host resistance
decrease, 210
effect of crowding, 279
infectious disease, 279
parasites, 279

370

Subject Index

House mice {see also mice, mouse, Mus),
140, 148, 180, 197, 251
ACTH and gonadotropin relationship,

260
adrenals, 206
adrenal cortex, 217, 282
adrenal weight, 266
albino, 269
birth rates, 292
competition, 271
competition for food, 297
crowding mortality, 280
decrease in reproduction, 286
food, 296
growth, 276
inanition, 272
inhibited growth, 299
intra-uterine mortality, 221
litter survival, 294
nests, 294

pituitary-adrenocortical activity, 273
resorption of the embryos, 221
responses to increased population den-
sity, 281
responses to sociopsychologic factors,

269
self-limited populations, 299
social pressure, 288
spermatogenesis, 288
splenic hypertrophy, 272
trichinosis, 278
wild, 269
wild-stock, 280
Howler Monkeys, 90, 121
Humans, 196
Humans

dominant, 132

hyi)othyroidism, 238

responses to sociopsychologic factors,

269
thyroid, 232, 236
Hum field, 41, 50
Humoral substance, 16
Huntington Forest, 27, 61, 82
Hyaluronidase, 209
Hydrocortisone, 205, 236
Hydrocortisone
cats, 206
ferrets, 206
guinea pigs, 206

humans, 206

mice, 20^

monkeys, 206

rats, 208

sheep, 206
HjTJeractivity, 175

in a strange environment, 16
Hyperglycemia, 209, 224, 226
Hyperthyroidism, 239
Hj^pertrophy

cortical, 216
Hypoglycemic shock, 320, 321
Hypophysectomy

secretion of aldosterone, 202, 203
Hypothalamic centers, 216

Indians

Kunghit Haida, 99
Individuals

alpha, 59, 61, 165
beta, 59, 165
gamma, 59, 165
Infection

resistance to, 256
Inflammation

effect of corticoids on, 254
Inflection point

arithmetic, 121
Inhibitory influence, 38
Input

rates, 56
Insulin, 240
Intensity of action, 112
Interaction
constant, 115
frequency, 108
frustrating, 119
intensity, 115
maximum frequency, 114
positively affective, 112
refractory, 109
responsive, 109
variables determining, 101
Interconstellation matrix, 60, 68
Intraspecific competition, 263
Intrauterine mortality, 259
Invasion

induced, 80
Iodine, 230
Iraq, 99
Islets of Langerhans, 240

Subject Index

371

Isolation

spatial, 69
Isolation cage, 17

Jarmo site, 99

Kendall's compound A, 205
B, 205

E, 205

F, 205

Lactation
failure, 260
inhibition, 294
Lemming (see also Dicroslonyx and
Lemmus), 320
populations

cj'clic decline, 319
mortality, 322
population cycles, 318
Lemmus, 305, 323
Lemmus trimucronalus

population cycles, 318
Lepus (see also hare), 323
Lepus americanus

relationships between population den-
sity and reproduction, 310
Life-space quadrants

shifts into the second and fourth by
rats, 159
Lipid mobilizing factor, 241
Lipogenesis, 209
Litter mortality
factors causing, 295
mice, 294
size, 239
survival

effect of increasing density on, 295
effect of nests on, 295
factors, 294
mice, 293, 294
voles, 293, 296
Liver, 260

glycogen, 252
LMF, 241
Loris, 196
Lymph nodes

weights, 249, 250
Lymphocytes

as indices of adrenocortical activity, 21 1

criteria of stress, 251

counts, 251
Lymphocytolysis, 211
Lymphocytopoiesis

depression, 211
Lymphocytopoiesis

depression, 211
Lymphoid organs

nucleic acid content of, 255
Lymphoid tissue

growth, 250

involution, 250

Macaca mulatia, 195
Macaque, 196
Maine, 32, 54
Male

dominant, 89

rats, pansexual, 162

reduction of, 90

territorial, 161
Mammals {see also specific names)

adrenalectomized, 208

European, 196
Man {see also human) 121, 173, 194

anestrus, 220

basic numbers for, 97

corticosteroid production, 268

ovarian atrophy, 220
^Marking

with red dye, 76
Marmoset, 196
Marmota, 193

relationships between population den-
sity and reproduction, 310
Marmota monax, 122
Maryland, 32

Montgomery County, 69
Meadow mouse (see Microtus)
Medulla {see also adrenal)

adrenal, 198

basophilic granules, 198

chromaffin cells, 198
Medullary hyperplasia

mice, 198

ungulates, 198
Mesocriceius auratus {see also hamster),

194, 196
Metestrus, 239

372

Subject Index

Mice {see also house mouse, AIus, white-
footed mice, mouse), 26, 148, 170, 193,
196, 19S, 212, 218, 222, 257, 264, 277,
293, 309

adrenals, 200

adrenal weights, 246

adrenocortical responses, 270

albino, 264

androgen, 221

anestrus, 220

antibody formation, 278

blood level of TSH, 237

corticosterone, 236

Coxsackie infections, 256, 257

decline in birthrate, 291

during pregnancy, 196

effect of new environment, 265

effect of overcrowding, 269

effect of population density, 274

effect of social pressures, 268

eosinophil counts, 252

fighting social competition, 287

gonadotropins, 309

harderian glands, 240

hierarchy, 265

hypoglycemic shock, 321

hypophysectomized, 212

inanition, 222

inbred, 274

inbred albino, 273

laboratory, 196

lactation, 274

medullary hypertrophy, 227

method of inducing an inflammatory
response, 277

nursing, 210

ovarian atrophy, 220

ovaries, 220

pituitary-adrenocortical activity, 273

pneumococcal infection, 256

pregnant, 210

preputial glands, 292

radiothyroidectomized, 239

release of iodine, 238

reproduction, 274

reproductive suppression, 275

resistance to infection, 277

shock death, 225

social hierarcy, 264

sociopsychologic interactions, 280

sjjlenic hypertrophy, 242, 273, 274

streptococcal infection, 256

susceptibility to trichinella, 277

testes, 238

thymus, 283

thj'roid activity, 235

trichinosis, 257

tuberculosis, 257, 279

tuberculosis infections, 256
Microsorex hiyi, 193, 246
Microtus (see also voles), 77, 301, 305, 323

hypoglycemic shock, 320

natural populations, 303

relationships between population den-
sity and reproduction, 310

"shock disease", 225
Microtus agrestis, 196, 251, 273, 323

behavior, 222

estrous pattern, 222

inhibition of growth and i('])roduction,
316

number of births, 222

respons(>s to increased pf)pulation den-
sity, 281

sociopsj-chologic factors, 264

splenic hypertrophy, 273
Microtus arvalis, 196, 218

confined populations, 272

crowding mortality, 280

sex accessories, 219
Microtus californicus, 323

reproduction, 317
Microtus montanus, 305, 316, 317, 320, 323

population density, 303, 314

population fluctuation, 303
Microtus ochrogaster

reproduction, 317
Microtus oeconomus

population cycles, 318
Microtus orcadensis

nucleus pulposus enlargement, 261

sociopsychologic factors, 264

splenic hyjjertrophy, 273
Microtus pcnnsylvanicus, 323

adrenocortical-densitj' relationships, 304

adrenal weight, 285

body-adrenal relationships, 247

effect of increased population density,
284

Subject Index

373

lice, 322

litter-size, SOS, 317
population fluctuations, 304
productive cycle, 319
responses to increased population den-
sity, 281
Migrations

lemming, 86
Milk let-down reflect, 2r)S
Monkey (see also Rhesus monkey), 194

brucellosis, 256

colobus, 196

corticoids, 268

malaria, 256
Moose, 122
Mortality

alterations, 94

effect of crowding on, 279

immediate cause, 323

increase, 323

in dense populations, 319

mass, 322

of young
factors, 321

prenatal, 315
Motor activity, 16
Mt. Desert Island, 33, 53
Mouse (see also mice), 196, 170, 175

adrenalectomized laboratory, 201

adrenal gland, 196

adrenal weight, 197

albino, 280

copulation pressure, 289

crowding mortality, 280

development of, 196

dominant, 264

fertility, 288

fights, 296

male fertilit\-, 289

Xoninovolution of X-zone, 258

puberty, 197
Mouse (see also Clethrionomys)

red-backed, 27, 52, 54

subordinate, 264

tundra, 318

velocity in a hierarchy, 149
Mule deer, 122

does, 24

fawns, 24

males, 24

Mus (see also mice and mouse)

inhibition of growth and reproduction,
316
Muskrat (see also Ondatra), 323
endocrines, 305

relationships between population den-
sity and reproduction, 310
reproduction, 317
reproductive organs, 305
Mus musculus {see also mice and mouse),
196, 217
medullar}^ hypertrophj^ 227
responses to increased population den-
sity, 281
Myotis, 193
Myotis grisecens, 122
Myotis lucifugus, 122
Myotis yiunanensis, 122
Myotis velifer, 122

Natural populations, 323
Natural selection, 122
Neighbors

contacting, 45, 46

distance between, 50

number of, 51

perception of, 51

signal field, 47

sign field, 47
Nerves

postganglionic sympathetic, 22:]

splanchnic sympathetic, 223
Nervous system

sympathetic, 223
Neurohypophysis, 214, 241
Neutrophil

count, 252
New York, 33, 54
New Zealand, 322
Niche

characterization, 77

l)rimary component of, 77

specializations, 74
NIH Emotional Activit.y Alley, 17
Norepinephrine, 209

effects, 223, 224

secretion, 226, 227

source, 225
North American Census of Small Mam-
mals, 26

374

Subject Index

Norwa}^ Rat (see also rat, Rattus), 87, 180,
195, 320, 323

adrenalectomized, 202

adrenal weight, 266, 301

behavioral sink development by, 92

body-adrenal relationships, 247

domesticated, 163, 175

inhibition of reproduction, 307

intra-uterine mortality, 221

laboratory, 195

relationships between population den-
sity and reproduction, 310

reproductive depression, 291

resorption of the embryos, 221

responses to sociopsychologic factors,
269

rural population, 302

social competition, 273

sociopsj'chologic interactions, 280

subordinate males, 265

wild, 179, 183, 195, 265
Number

total of individuals, 110

Oak, 64

Ondatra (see also muskrat), 323

relationships between population den-
sity and reproduction, 310
Ondatra zibethica

reproduction, 317
Open-field emotional behavior, 89
Orders, 122
Organ

isolated, 191
Ornithorhyn chics, 195
Oryciolagvs, 193, 323
Os penis

stimulation, 212
Ovary, 194
Ovaries

androgens, 220

Pan, 195
Pancreas, 260
Pancreatic islets, 240
Pantothenic acid

deficient, 222
Parasitism, 322
Parotid. 260

Particle

basic, 2

general classes, 3
Partner

choosing of, 154

data regarding the choosing of, 156
Perception, 89, 139
Perodicus potto, 195

Peromyscus, 29, 30, 31, 32, 33, 34, 35, 36,
52, 53, 54, 55, 61, 64, 66, 67, 68, 69,
70, 75, 76, 77, 78, 79, 80, 88, 323
Peromyscus boylii

inhibition of reproduction, 307
Peromyscus leucopus

pituitary-adrenocortical, 273

population, 305
Peromyscus maniculatus

inhibition of reproduction, 307

lice, 322

productive cycle, 319
Personality, 158
Phagocytosis

inhibition, 210
Physiologic adaptation

generalized effects, 254
Physiology

classical, 184

evolution, 60

reproductive, 95
Pigs, 194
Pine, 64
Pineal, 204

Pipistrellus (see also bat), 193
Pitressin, 214, 232
Pituitary

anterior, 197, 220

antidiuretic, 241

gonadotropins, 257

hormones, 218

lipid mobilizing factor, 241

posterior, 241

secretion of aldosterone, 203
Pituitary-adrenocortical exhaustion, 226

tumors, 218
Pitymys, 32, 52, 64, 66, 69, 70, 76, 196
Pitymys subterraneus, 196, 197
Population, 189

collapse, 314

densities, 190, 238

fertility, 290

Subject Index

375

fixed size, 263

food-limited, 298

freely growing, 281

individuals of, 103

limited size, 280

methods for estimating natural, 305

natural, 190, 262, 300, 306

physiologic adaptation, 261

reproductive function, 306

resident, 56

self-limited, 298

size, 240

uniforml}' distributed, 58
Population density, 226, 262, 287, 305

adrenocortical and reproductive re-
sponses, 269

effect on adrenocortical function, 281

endocrine responses, 263

index of relative, 305

inflammation, 276

resistance to infection, 276

schematic summary, 326, 327
Potassium, 204

excretion, 201

levels, 204

level of body, 202

secretion of aldosterone, 204
Potto, 195

Preconstellation phase, 58
Pregneninolone, 210
Preputial development

adrenal androgens, 292

stimulating, 292

stimulation, 212
Probability

formulas, 144
Pi(A), 144
PiW, 144

tenninating trips, 12

vacillating at the termination of trips, 18
Proestrus, 239
Psychology

classical, 184, 185

Rabbit, 193, 323

adrenalectomized, 236
corticosterone, 236
domestic, 236
embryo

resorption, 239

parasitism, 322

release of radioiodine, 233

thyroid, 232

thyroid activity, 235

thyroidectomy, 239

thyrotoxicosis, 235

wild, 235, 236
Radial distance, 623
Radioiodine

release in domestic rabbits, 236

release in wild rabbits, 236

uptake, 232
Rat (see also Norway rat, Rattus), 138,
159, 182, 194, 196, 212, 218, 309

ACTH and gonadotropin relationship,
260

adrenals, 201, 206

adrenal atrophy, 237

adrenal cortex, 200

adrenalectomized, 253

adrenalectomized laboratory, 201

adrenal hypertrophy, 237

adrenal weight, 302

adrenocortical secretion, 222

androgen, 221

anestrus, 220

albino Osborne-Mendel strain, 181

blood level of TSH, 237

chorionic gonadotropins, 220

corticosterone, 236

domesticated albino strain, 171

domesticated Norway, 8

effect of temperature changes on, 274

gonadectomized, 219

gonadotropins, 212, 309

growth, 210

harderian glands, 240

house, 212

hypertrophy of medulla, 227

hypophysectomized, 216, 233, 254

hypothalamic-hypophyseal-thyroidal re-
lationships, 233

hypothyroid, 234, 239

inanition, 272

in a strong environment, 303

inbred, 274

infant, 210

length of estrous cycle, 238

method of inducing an inflammatory
response, 277

376

Subject Index

Rat — continued

newborn, 210

Nonvay (see Norway rat)

Osborne-Mendel, 17

ovarian atrophy, 220

pituitarj^ 236

pituitary-adrenocortical activity, 273

pneumococcal infection, 256

postparturitional loss, 221

pregnant, 210

release of iodine, 238

splenic hypertrophj^ 274

streptococcal infection, 256

submaxillary glands, 240

subordinate, 303

testes, 239

of immature, 238

thiouracil treatment, 239

thyroid, 232

thyroid activity, 235, 237

thyroidectomized, 228

tuberculosis infection, 256

uptake of radioiodine, 233

white, 237

wild, 195
Rat embryos

action of estrogens, 259
Rat society, 161

velocity in high-density, 168

velocity-rank relationships in, 168, 169
Rattus (see rat, Norway rat)
Rattus alexandrinus, 195
Rattus norvegicus, 195

adrenalectomized, 202

medullar}' hypertrophy, 227

relationships between population den-
sity and reproduction, 310
Recapture radii, 5, 6
Receptor, 50
Refractory period, 49, 141, 153, 176

duration, 103, 112

frustrating, 110

frustrating type, 131

frustrating producion, 112

increasing, 130

satisfaction producing, 112

time spent in satisfying, 112
Refractory state

maximal, 109

optimal, 109

Regression line, 81

Reichstein's compound S, 199, 205

Reithrodontomys, 6

Rejection

psychological, 183
Relationships

spatial, 64

summary of presumed, 137
Relative densitj', 78
Relative probability, 45
Removal study

Chadwick Woods, Montgomery Count}',
Maryland, 29

Rich Lake Island, New York, 27
Removal trapping

observed data, 52
Renal nephron

tubular cells, 201
Reproduction, 190, 286-291

evaluating the effects of densit}' on, 308

inhibition, 309

mice, female, 291

mice, male, 286

sexual, 133

suppression, 222, 257
Reproductive

inhil)ition, 220

function, 292
evaluation, 306

inhibited by population density, 292
inhibition, 323

organs

accessory weights, 287
Reserpine, 225
Response

avoidance, 93

degree, 192

endocrine, 191

inflammatory, 210

intensity, 151

neural, 191

neuroendocrine, 191

nonspecific, 192

vascular, 191
Response-evoking capacities, 143, 146
Response-evoking capacity

behavioral origin, 140

circumplex depiction of behavior and
personality superimposed upon, 157

Subject Index

377

circumplex, 155
formulae, 128, 1-13
S'-^\ 143
S^^\ 143
SjA, 128
Si^^', 128
S^'^), 143
S('), 143
Responsive state, 140
Rhesus monkey (see also monkey). 267
hj'drocortisone, 236
responses to sociopsychologic factors,

269
thyroid activity, 236
Rodents (see also specific categories), 212
Runaways

underground, 76

Salivary glands, 260

Sanctions, 151, 163

Scent, 38

Schaefer's autonomy, 143

Schaefer's control, 143

Schaefer's hostility-rejection, 128, 143

Schaefer's love-acceptance, 128, 143

Serotonin, 216

Sex accessories, 238

involution, 218
Sex steroids, 206, 217
Sheep, 194
Shock, 192, 243, 244
"Shock disease," 226, 280, 320
Shrews (see also Sorex and Blarina), 26, 27,
75, 193, 197

long-tailed, 246

population, 305

short-tailed, 305
Sight, 38
Sigmodon {see also Cotton rat), 323

litter size, 317

relationships Ix^twecn })o})ulation den-
sity and reproduction, 310
Sign, 47
Sign field, 49
Signal, 53, 57, 167

detection, 41

inhibitory, 68

intensity, 48, 49

learning of, 39, 50

neural, 78

noxious, 78

simultaneous, 48

sound, 41
Signal field, 47, 49
Sika deer, 323

mass mortality, 320

population density, 306
Size

basic group, 86

effect on response-evoking capacity, 130
Slow loris, 196

Snowshoe hare (see also hare), 253
Social behavior, 128

more effective, 136
Social dominance, 74

adrenal activity, 267

changing, 89
Social factors

adrenocortical functions, 236

interspecific, 77

reproductive functions, 263
Social groups

budding off, 121
Social hierarchy

evolution, 70
Social interaction, 133

basic processes, 116

consequences and examples, 148

frequency of satisfactory, 117

frustration from, 110

as a function of density, 117

heightened frequency, 95

intensity, 152

model, 101

optimum satisfaction, 118

and refractory period, 117

satisfaction from, 110

withdrawal from, 173
Social phenomena, 3
Social pressure

endocrine responses, 263
Social rank, 64
Social relations

instability, 67
Social response

blocking, 104

mechanism, 104
Social subordination

interspecific, 67

378

Subject Index

Social system, 3, 4

semiclosed, 100
Social withdrawal, 145
Society

human, 184

rat, 161
Sodium: potassium ratio

serum, 204
Sodium retention, 201
Sorex, {see also shrew), 27, 29, 31, 32, 36,

' - 38, 52, 61, 68, 69, 70, 72, 75, 77, 78
/Sorex araneus, 196, 197
Sorex cinereus, 193, 246
Sorex dispar, 193, 246
Sorex fumeus, 193, 246
(Sorex palustris, 193, 246
Soricidae {see also shrew, Sorex, and
Blarina)

adrenal, 246
Space

general theorj- of use, 34

interspecific and intraspecific use, 34

social use, 2

theoretical conceptualization of use, 70
Spatial equilibrium, 57
Species, 3

alpha, 36, 37

l)eta, 36, 37

distance between neighbors, 42

dominant, 36, 52, 67, 68, 78

incipient, 74

rf^lationship between two dominant, 52

removal captures, 55

socially dominant, 55, 57

subordinate, 36, 38, 52, 61, 67, 68, 78

survival, 122
Spermatogenesis,

decrease, 258
Spleen

house mice, 251

h^'pertrophy, 251

Microtus agrestis, 251

volves, 251

weights, 249, 251
Splenic hypertrophy, 271
Standard area, 5
State

nonresponsive P, 140
Stations

feeding, 95

Status

subordinate, 68
Steady state

evolutionary, 108
Steroid hormones, 218
Steroids, 248

sex, 258

site of action, 216

urinary, 249
Stimuli

alarming, 212, 215
Stimuli

auditor}^ 70

emotional, 243

conditioned, 75
Strange object reaction, 89
Strange-object response, 179
Stress, 88, 193, 196, 213, 235

generalized state, 182

resulting from group size change, 118
Stressful state

amount of time spent, 112
Stressors

nonspecific, 119
Stretch receptors

secretion of aldosterone, 204
Strife

intraspecific, 262
Sudanophilia, 248
Supraopticohypophyseal tract, 214
Survival

threshold of, 132
Sylvilag^is {see also rabbit), 193
SjTnpathetic nervous system

ganglia, 198
System

Iwpothesized communication, 58

two-species, 36

Tachyglossus, 195
Target diameter

determined by variabihty of l^diavioral
traits, 140

genotype, 140

hj'pothesized divergence, 132

of individual, 110

variability, 127
Territory, 263

of male mice, 287

Subject Index

379

Testis

interstitial cells, 194
Testosterone, 208
Thetas

satisfaction and frustration, 131
Thiouracil, 237

Thymicolymphatic system, 249
Thymus, 283

cortex, 249

involution, 208, 211, 249, 250

medulla, 249

weights, 249
Thja-ocarotid artery, 204
Thyroglobulin, 230
Thyroid

activity

factors regulating, 231
index, 231
mice, 235, 236
NaCl intake, 238
rabbit, 236
rat, 235, 236

gland, 191, 228

honnone
actions, 229
formation, 230

inhibition, 236

responses to external stimuli, 234
Thyroidal-gonadal interrelationships, 238
Thyroidectomy, 228, 234
Thyrotropin, 231

inhibition, 236
Thyroxine, 228, 229, 231

effects, 230
Tonus

vascular, 224
Trap, 20, 37

"Havahart," 184

probability of encountering, 52
Trapline, 53, 82

standard, 26
Trapped-out area

invasion, 82
Trapping, 76

comparison of results between short-
term and long-term removal, 28

continuous removal, 26, 37

removal, 76

short-term removal, 26

Tribe

as a social entity, 98
Trichinella

increased invasiveness, 278
Trichinella spiralis, 277, 278
Trichinosis

effects of crowding, 278
Triiodothyronine, 228, 229, 231
TSH

mice, 237

rabbits, 237

rats, 237

rate of release, 231

release, 232, 236

secretion, 237, 258
Tuberculosis

crowding, effect on, 279

murine, 279
Two-dimensional field

use, 19
Tyrosine, 230

Ungulates

captive, 198

domesticated, 196
Urine

metabolites, 207

Vaginal orifice

closure, 291
Vagus nerve, 204
Vasopressin, 214
Velocity, 164

biological, 111

effect of interval between ^'s on, LSI

effect of tranquilizer on, 172

effect on fat, 174

effect on weights of adrenals, kidney;
and heart, 174

formulae, 111, 112, 113
a, 112
aaa, 112
aap, 112
afaa, 112
af'ap, 112

«mfaa^"\ 113

d, no

faa, 112
f'ap, 112
i, 112

380

Subject Index

12, 112
/i', 112
N, 110
Nb, 113
N^""', 113
No, 113
0a, 112
^a^""), 113

0a^»', 113

6>(''), 112
0f, 112
0"), 112

^a'""*, 113

V, 111

minimal, 173

rank, 171

reduction in hierarchj^ of mice, 148
Visitation frequency, 165
Vitamin A

effects on behavior, 170
Vitamin B

insufficiency, 222
Vocalization, 38, 48, 75

loss of, 77
Voles, (see also Microius, Clethrionomys,
and Pitymys), 193, 196, 253, 261, 293,
317, 318, 322, 323

adrenalectomized, 218

adrenal weight, 285, 286, 301

adrenocortical-densit}' relationships, 304

anestrus, 220

birth rates, 292

competition for food, 297

confined populations, 272

copulatory pressure, 290

crowding mortality, 280

decline in birthrate, 291

decrease in reproduction, 286

depression of fertility, 221

effects of high density, 276

effect of increased population density,
284

effect of overcrowding, 269

effect of population densities, 275

eosinophil counts, 252

female, 286

fights, 296

hypoglycemic shock, 320, 321

inhibition of growth and reproduction,
316

intraspecific strife, 323

laboratory, 274

litter survival, 294

meadow, 196, 218, 272, 323

northern montana, 304

mortality, 322

natural populations, 303

nests, 294

ovarian atrophy, 220

pine, 196

population fluctuation, 312

red-backed, 196, 311, 312, 314, 318, 323

relationships between population den-
sity and reproduction, 310

reproductive function, 314

responses to increased population den-
sity, 281

responses to sociopsychologic factors,
269

self-limited populations, 299

social pressure, 288

sociopsychologic factors, 264

sociopsychologic interactions, 280

splenic hypertrophy, 242, 272, 273, 274

X-zone, 283

White-footed mice (see also Peromyscris),

273, 323
White-footed mice

population, 305
Withdrawal, 162
Withdrawal state, 180
Woodchuck (see also Marmota), 193

relationships between population den-
sity and reproduction, 310
Wound healing

delay, 210

X-zone

acidophilic cells, 197

androgens, 197

basophilic nuclei, 197

Crocidura russula, 197

experimental treatments, 196

function, 217

hormones, 196

pituitary luteinizing hormone, 197

Pitymys subterraneus, 196, 197

reactions to, 196

shrews, 197

Subject Index

381

Sorex araneiis, 197
voles, 196

Zona fasciculata, 217

carbohydrate-active corticoid.s, 2C0
hormones, 205
hyperplasia, 246
hypertrophy, 246
regulation of hormones, 213

Zona glomerulosa

aldosterone, 213

beef cattle, 200

hypertrophy, 246

rats, 200
Zona reticularis

carbohydrate-active corticoids, 200

function, 217

sex steroids, 217

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Books

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "Physiological mammalogy"

See other formats