1, _ _ . V
1 , . .
<
A
4 . , ,
,
1 .
i
. : ~
q. . , .
. c .
. .
. .
..
.1
. .7 .
E . . .
. a
r
:
. .
. .
.
.
_ .

1.942.... .5; ..

 

? o........ ...
. .2 .

 

OPERIOD Anni]

HOT

OF .P

ME

THE EFF

ARDI

OD H0

Ear

RAT;

2N .DEERM'ICHP

TEMPE

SOUS?) :f

ow

.ON “F0

1"!

ER:

\

é

afnonferthebeg

:

               

   

xmg .
"haw... “av“. £4.20. Gum”...
u. .1.

. . t u. .
,. . .Hxfi... V : 1.
. : 11%.“
1 1.11.“. a: .2:
. . w A_ Hi? .3. 4 3.2.;

     
                      
   

 

 

L ummlwm. $50.»...

:Il‘...

 

 

 

 

.' n. . “‘1‘..." {0|
min...» £.1....1.:t|..l {ADI-Io .lnsb .
. ‘r‘lvrl"_¢i|l '3'
4.455% I .1. r5 FUJIK l
.\ I 74 In! A?
§§y .5

. (11.13.: 1J1!“
..

F

    

.. . Q; . I
‘1'): ’JIOA’LNLID
. .v’l. '2: .n:.'....
‘1... \.Iy1\ I." t
\ 7‘\. n1

  

IIIIIIIIIIIIIIIIIIIIIIIIIIIIII

3 1293 10702 7082

‘3 4-;‘ ‘

1.13qu ze-
Midyigan 5n ,-

Uilivmi 1' y

 

 

 

ABSTRACT

THE EFFECT OF PHOTOPERIOD AND TEMPERATURE ON
FOOD HOARDING IN DEERMICE (PEROMYSCUS)

 

BY

William J. Barry

In northern latitudes, food hoarding has been inter-
preted as one means of providing for seasonal fluctuations
in the availability of food. The relative scarcity of food
during the winter months in these latitudes would pose a
special problem to a non-hibernating or non-migratory
animal in the absence of stored food. Many rodents,

including deermice (Peromyscus), solve this problem by

 

storing food in the form of body fat or free (unconverted)
items of food in or near their nest. It was hypothesized
that a short photoperiod and/or low temperature were the
environmental stimuli for deermice to hoard during the fall
in the northern latitudes. Previous published reports

indicated that flying squirrels (Glaucomys volans) responded

 

to experimentally shortened photoperiods by intensively
hoarding their food. Studies on the laboratory rat also
suggested that low temperatures may induce hoarding in at

least some species of rodents.

William J. Barry

Five taxa of deermice including P. lgucopus nove—

 

boracensis, g. l. castaneus, E: maniculatus bairdi. g, m,
blandus, and £3 eremicus eremicus were tested in the labo-
ratory for the amount of food hoarded in seven consecutive
days under various combinations of test temperature (7°C
and 27°C) and photoperiod (9L:15D or 15L:9D). Hoarding was
operationally defined as the number of pellets carried from
a food bin at the end of a 43cm.-long alley attached to the
animal's home cage in a 24 hour period. All mice prior to
testing were acclimated for 4 weeks either to 7°C or 27°C,
with half of the mice tested at the same temperature to
which they had been acclimated and half to the opposite
temperature. Mice were kept under a constant photoperiod,
either a "short day" (9L:ISD) or a "long day" one (15L:9D),
for both the acclimation period and testing. All mice were

laboratory reared and, except for E. l. castaneus, were the

 

first generation from wild caught parents.

The results indicated that the two northern subspecies,
E. l. noveboracensis and g. m. bairdi. hoarded significantly
more under the low temperature and short photoperiod than
they did under the high temperature and long day photo-
period. A similar response to these environmental vari—
ables in the wild would lead to a higher level of hoarding
in the fall of the year than in the summer. It was also
found that photoperiod and temperature were not of equal

influence on both E. l. noveboracensis and E. m. bairdi.

 

Test temperature had more of an effect than photoperiod on

William J. Barry

the hoarding of E. l. noveboracensis, while in E, m,
bairdi the opposite was true.
Comparisons of the mean hoarding scores of E, l,

noveboracensis with E, l. castaneus, and E: m, bairdi with

 

 

g. m, blandus revealed that in both cases the northern
subspecies hoarded significantly more than the southern one
under the cold test temperature. Under the warm test
temperature there was no significant difference in their
scores. These findings indicate a selection for hoarding
in northern subspecies of deermice living in environments
where a scarcity of food is associated with seasonally low
temperatures.

Test temperature also had a significant effect on the
hoarding of g, m, blandus, which hoarded more under 27°C
than under 7°C. The increase in hoarding under the warm
temperature by this subspecies may have been due either to
an inhibition of activity under the cold temperature or to
the possibility that warm temperatures in its environment
(i.e., southwestern deserts) indicate a summer shortage of
food. The high level of hoarding observed for a. g.
eremicus, in contrast to the other groups studied, was inter-
preted as perhaps related to the occurrence of summer torpor
in this desert species and its reliance on stored food while
restricted to burrows during extremes in temperature.

In all groups studied, except possibly 3. g, bairdi,
there was no evidence found to support a model of food

hoarding in deermice based on long term nutritional needs

William J. Barry

or deficits (i.e., losses in body weight) as has recently
been proposed for laboratory rats. In a, m, bairdi there
was only a small correlation between weight change and

hoarding scores. In all other groups this correlation was

non-significant.

THE EFFECT OF PHOTOPERIOD AND TEMPERATURE ON

FOOD HOARDING IN DEERMICE (PEROMYSCUS)

BY

William J. Barry

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Zoology

1974

ACKNOWLEDGMENTS

I wish to express my appreciation to John A. King, my
advisor, for his help and guidance throughout the duration
of this study. I am also grateful to the members of my
guidance committee including Rollin H. Baker, James C.
Braddock, Walter H. Conley, and Richard W. Hill for their
constructive comments during this research. A special
thanks must go to my wife, Mary Ellen, for her incalculable
patience during all phases of my study. This research was
supported in part by NIH Animal Behavior Training Grant
# 5 T01 GM01751-05 BHS and by a teaching assistantship

provided by the Department Zoology.

ii

TABLE OF CONTENTS

LIST OF TABLES . . . . . . . .
LIST OF FIGURES. . . . . . . .
INTRODUCTION. . . . . . . . .
LITERATURE REVIEW . . . . . . .

Phylogeny . . . . . . . .
Collection and Transportation .
Preparation of the Hoard . .
Nature of the Hoard.
Seasonality . . .
Evolution . . .

Experimental Studies

Organismic Factors . . . . .

Deprivation . . . . .
Dietary and Energetic Deficienci
Sex Differences. . . . . .
Neural Control . . . . .
Learning and Early Experience
Genetic Variation . . . .

Situational Factors . . . . .

Competition-Social Interaction.
Cage, Lighting, and Familiarity
Stimulus Characteristics. . .
Temperature . . . . . . .
Photoperiod . . . . . . .

METHODS AND MATERIALS. . . . . .

Subjects . . . . . . . . .
Apparatus . . . . . . . . .

iii

es

Apparatus

Page

vii

12
13
15
16
19
25

25

25
28
31
33
36
38

40

40
42
43
46
51

51

51
55

Procedure .
Analysis .

RESULTS . .

Comparisons Within Each Subspecies.
Comparisons Among Subspecies.

Analysis of Weight Changes

DISCUSSION .

Test Temperature.

Photoperiod

Possible Physiological Mechanism for

Torpor and Food Hoarding .

Weight Change and Hoarding

Conclusions
SUMMARY . .
APPENDIX . .

BIBLIOGRAPHY.

iv

Page

O O O O O 55
O O O O O 59
. . . . . 62
O O O O O 62
O O O O O 77
O O O O O 91
O O O O O 108
O O O O O 109
O O O O O 116
Hoarding . . 118
. . . . . 119
O O O O O 121
O O O O O 123
O O O O O 126
O O O O O 128
O O O O O 142

LIST OF TABLES

Systematic Relationships, Sex and Number of
Mice Used in This Study . . . . . . .

Description of the Geographical Origins of
the Mice Used in This Study. . . . . .

The Experimental Design for This Study. . .

Mean Daily Log Hoarding Scores for the Main
Effects of Sex, Acclimation Temperature,
Photoperiod, and Test Temperature. . . .

Student's t Tests of the Difference Between
the Mean Daily Log Hoarding Scores Under 7°C
and the Mean Scores Under 27°C for Both
Subspecies of E, lgucopus and g,
maniculatus . . . . . . . . . . .

Multiple Comparisons Among the_Mean Log
Hoarding Scores (i.e., Log (X+l) for Each
Mouse) i_ls.E. for All Subspecies of
Peromyscus Used in This Study . . . . .

Mean Log Hoarding Scores (i.e., Log (2+1) for
Each Mouse) for the Five Stocks of
Perom scus Under All Possible Combinations
of Test Temperature and Photoperiod . . .

Planned Comparisons of the Mean Log Hoarding
Scores (i.e., Log (X+l) for Each Mouse) of
Mice Tested Under a Short Photoperiod and
7°C With Those of Mice Tested Under a
Long Photoperiod and 27°C . . . . . .

Correlation Coefficients Between the Hoarding
Scores of All Five Stocks of Peromyscus
and Their Age, Initial Weight, Weight
Change, and Final Weight. . . . . . .

 

Page

52

53

58

63

71

85

92

93

94

Table
10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

Student's t Test Comparisons of the Mean
Weight Changes (In Grams) After Acclimation
to 7°C and 27°C for the Significant
Interaction of Subspecies x Acclimation
Temperature . . . . . . . . . . .

Weight Changes (In Grams) of Mice Under All
Combinations of Acclimation and Test
Temperature . . . . . . . . . . .

Analysis of Variance of the Log Hoarding
Scores of a, ieucopus noveboracensis. . .

Analysis of Variance of the Log Hoarding
Scores of £3 leucgpus castaneus . . . .

Analysis of Variance of the Log Hoarding
Scores of £3 maniculatus bairdi . . . .

Analysis of Variance of the Log Hoarding
Scores of £3 maniculatus blandus . . . .

Analysis of Variance of Log Hoarding Scores
of £3 eremicus eremicus . . . , , , ,

Combined Analysis of Variance for the Mean Log
Hoarding Scores (i. e., Log (X+1) for Each
Mouse) of P. 1. noveboracensis, P. l.
castaneus, -P. -m.’ba1rdi, andIP. m.

Blandus . . . . . . . . . . .

Combined Analysis of Variance for the Mean
Weight Changes in P. l. noveboracensis,
P. l. castaneus, P. m. bairdi, and P. m.
Biandus . . . . . . . . . . . .

 

Analysis of Variance of the Weight Change in
a. eremicus . . . . . . . . . . .

vi

Page

100

107

129

130

132

134

136

137

139

141

LIST OF FIGURES

Figure Page

1. Mean log hoarding scores for the two sub-

species of both P. leuco us and E,

maniculatus on each of tEe seven test days . 65
2. Mean daily log hoarding scores for P. ‘1.

noveboracensis under the two test-
temperatures . . . . . . . . . . . 68

3. Mean daily log hoarding scores for P. L
castaneus under the two test temperatures . 70

4. Mean daily log hoarding scores of P. 1.
castaneus for the significant interactions
of (a) sex x acclimation temperature, and
(b) photoperiod x acclimation temperature . 73

5. Mean daily log hoarding scores for P. m.
bairdi under the two test temperatures . . 76

6. Mean daily log hoarding scores for P. m.
blandus under the two test temperatures . . 79

7. Comparisons of the mean log hoarding scores
per test day for all five stocks of
Peromyscus studied . . . . . . . . . 82

 

8. Graphs of the mean log hoarding scores per
test day for the interaction of subspecies x
test temperature. . . . . . . . . . 84

9. Graphs of the mean log hoarding scores per
test day for the significant interaction
of subspecies x photoperiod x acclimation
temperature in P. leucopus under (a) a
short photoperiod, and (b) a long
photoperiod . . . . . . . . . . . 87

 

vii

Figure

10. Graphs of the mean log hoarding scores per
test day for the significant interaction of
subspecies x photoperiod x acclimation
temperature in g, maniculatus under (a) a
short photoperiod, and (b) a long photo-
period . . . . . . . . . . . .

ll. Graphs of the mean weight changes for all
five stocks of Perom scus after acclimation
to either 7°C or 275% . . . . . . .

12. Graphs of the mean weight changes for the
significant interaction of subspecies x
sex x acclimation temperature . . . .

13. Graphs of the mean weight changes for the
significant interaction of subspecies x
sex x photoperiod . . . . . . . .

14. Graphs of the mean weight changes for the

significant interaction 6f subspecies x
sex x test temperature. . . . . . .

viii

Page

89

97

99

102

105

INTRODUCTION

Hoarding has been defined by McCleary and Morgan (1946)
as the "acquisition of property to an extent that exceeds
the needs of the moment or the immediate future." Hoarded
items are typically "cached" or hidden away where they can
be used during times of scarcity (Hamilton, 1939). In
northern latitudes, the hoarding of food has been interpreted
as one means of providing for seasonal fluctuations in the
availability of food. The relative scarcity of food during
the winter months in these latitudes would pose a special
problem to a non-hibernating or non-migratory animal in the
absence of stored food. Many rodents, including deermice

(Peromyscus), solve this problem by storing food in the

 

form of body fat or free (unconverted) items of food in or
near their nest. At other times of the year, a wider
variety of food, including insects, fungi, tree buds, and
some fruits can be obtained. Many of these food items,
however, are unsuitable for storage in caches because they
cannot be readily transported or stored without spoiling.
This dissertation examines two environmental cues,

temperature and photoperiod, that may influence food

hoarding in five taxa of deermice. Deermice were selected
because (1) they will hoard food both in the wild and in the
laboratory, (2) they can easily be raised under standardized
laboratory conditions to control for learning, and (3) they
have undergone considerable speciation in a broad geographic
section on the North American Continent with most species
demonstrating geographic variation and having numerous
recognizable subspecies. Deermice are found in practically
every sort of available habitat including deserts, grass-
lands, chaparral, mixed deciduous forests, conifer forests,
and even in the tropical rain forests. Caching of food is
common in all species of deermice that have been studied,
and the items assembled vary depending upon the species and
the seasonal availability of food (Eisenberg, 1968). Their
food is all very similar and includes grains, seeds,
fruits, insects, and largely the mast crops of trees and
shrubs (Baker, 1968). Seeds and nuts are commonly stored,
although a variety of other items have been found in food
caches including the remains of grasshoppers, beetles, and
other insects, deer pellets, fox scats, leaves, flower
parts, and even lichens (Howard & Evans, 1961). The caches
are frequently very large and may contain several pounds of
material (Criddle, 1950).

The wide-spread geographical distribution of deermice
suggests that their food storage may vary with climate and
photoperiod. Food storage by deermice in the northern

latitudes appears seasonal in its occurrence, with the

greatest amount of activity in the fall months of the year
(Burt, 1946; McCabe & Blanchard, 1950; Criddle, 1950;
Howard & Evans, 1961). Other seasonal behaviors such as
reproduction, hibernation, and migration are influenced by
both photoperiod and temperature. Recent information
suggested that perhaps hoarding was likewise under the
influence of these environmental cues. Muul (1968) found
that flying squirrels (Glaucomys volans) responded to a
shortened photoperiod by intensively hoarding the ripening
mast crops. According to Muul this capability allows a
precise timing of food storage to the season of the year
coinciding with the greatest abundance of food suitable
for hoarding. Laboratory studies on food hoarding in the
laboratory rat also suggested that perhaps the colder
daily temperatures of autumn might similarly increase the
intensity of hoarding in at least some rodent species
(McCleary & Morgan, 1946).

If food storage in deermice is under the influence of
both temperature and day length, the response of any
particular species ought to be predictable from a knowledge
of their habitat and latitudinal origin. A comparative
study of several subspecies along a latitudinal gradient
should reveal differences that can be explained on the
basis of differential selection for certain behavioral
adaptations appropriate in each environment (King, 1968).
One would not expect that food storage by deermice from the

tropics would vary to any appreciable extent with

photoperiod. The seasonal differences in day length
increase with latitude, and mice from latitudes close to
the equator would experience little change in their day
length throughout the year. On the other hand, mice from
northern latitudes (i.e., above 35° North Latitude) Could
use photoperiod as a highly reliable cue to seasonal
fluctuations in their food supply. As the photoperiod at
each latitude has essentially been constant throughout the
evolutionary history of the genus Peromyscus, there would
have been adequate time to have evolved the capability of
responding to the appropriate changes in day length.
Climate, however, has been less constant with the relatively
recent retreat of the glaciers in what is now the northern
range of both E: maniculatus and P, leucopus. As climate
also varies with elevation, this cue might generally be
considered less reliable than photoperiod. Nevertheless,
climatic factors such as temperature and rainfall may have
been sufficiently predictable in the low-elevation habitats
of the southern latitudes (i.e., deserts) for some races
or species to have evolved a capability of using these
factors as cues to the availability of food.

Recent experimental work on the hoarding behavior of
laboratory rats has tended to emphasize a physiologic
mechanism for food storage. Herberg and Blundell (1970)
have proposed a hypothalamic model for hoarding whereby the
behavior is motivated by long-term nutritional deficits

reflected by losses in body weight or a reduction in fat

reserves. The observation that laboratory rats increase
their hoarding after food deprivation and under cold
temperatures would tend to support their model. Electrical
stimulation of the lateral hypothalamic "feeding center" was
also found to initiate hoarding in both naive and satiated
rats (Blundell & Herberg, 1973). However, there is little
evidence to suggest that such a mechanism may explain
hoarding in deermice and other rodents. Most wild species
of rodents hoard when food is abundant and in the northern
latitudes may even be gaining weight in preparation for
winter. The body weights of wild E. leucgpus noveboracensis
apparently do increase during autumn and early winter
(Sealander, 1951), which would coincide with their maximal
hoarding activity. The weight changes of deermice in my
own study were examined under all experimental conditions
to see if a correlation did exist between the level of
hoarding and weight loss. Since body weight is known to
vary with such factors as diet, breeding condition, and
age, special procedures were used to control for these
variables.

In summary the following three general hypotheses were
examined in this investigation:

1. Photoperiod and/or temperature affect the amount

of food stored by deermice.
2. Their effect will vary within a single deermouse
species depending upon the latitudinal and

climatic origin of the subspecies being studied.

Subspecies from the northern latitudes (i.e.,

above 35° North Latitude) and not those from
southern latitudes will hoard more under conditions
simulating winter (i.e., 9L:lSD photoperiod and
7°C) than under conditions simulating summer

(i.e., 15L:9D photoperiod and 27°C).

The hoarding of food by deermice is not related to

changes in their body weight.

LITERATURE REVIEW

The literature on food storage by mammals can generally
be divided into two major categories: experimental and
field observations. Experimental studies on food storage
have typically been done in the laboratory or with caged
animals. In general, the goal of these studies has been
to elucidate the mechanisms (both endogenous and exogenous)
of the food storage response with the subjects being

laboratory rats (Rattus norvegicus), mice (Mus musculus),

 

or hamsters (Mesocricetus auratus). More recently Morris
(1962), Ewer (1965, 1967), Muul (1968), and Lockner (1972)
have provided additional information from laboratory
studies on wild caught animals that have been called
"natural hoarders." Some of these studies suffered from
the use of very few animals or animals of one sex. The
second major category of studies includes a miscellany of
published reports on the occurrence of food storage in
animals in the wild. Many of these observations were made
by the early naturalists around the turn of this century on
the mammals of North America. These reports provide

information on which species store what kinds of food as

well as the possible function of food storage in the
animal's own habitat. Many of these early explanations of
food storage were speculative, although the results of more
recent laboratory studies have occasionally been given
similar "naturalistic" interpretations (for example, see
Thorpe, 1963).

In the review which follows I have separated the
literature on food storage into roughly the two categories
just described. First, I will survey the occurrence of
this behavior in the various mammalian orders, with special
emphasis on the food "hoarding" of rodents. The unique
characteristics of rodent food hoarding will be examined,
and the possible mechanisms for its evolution will be
discussed. Second, I will attempt to provide a complete
and critical review of the experimental work done on

hoarding behavior in the laboratory.

Phylogeny

A survey of the Class Mammalia immediately reveals
that certain of the orders characteristically show no food
storage behavior and for obvious reasons. Orders with an
entirely aquatic mode of existence (i.e., Cetacea, Sirenia,
Pinnipedia) or a truly aerial one (Chiroptera) typically
make no attempt either to cache or accumulate food beyond
what is of immediate use. Among aquatic mammals, the
absence of a permanent home or place in which to store

food as well as the type of food itself probably rule out

this behavior as a possible adaptation in their environment.
Only easily preserved food is suitable for storage and the
diet of some species of bats may be solely of insects, fish,
or fruit, which cannot be readily stored.

The seemingly inconsistent pattern of food storage in
the remaining terrestrial orders of mammals is a little
more difficult to explain. The absence of food storage
among the ungulates (Perissodactyla and Artiodactyla) and
elephants (Proboscidea), could be due to their nomadic
existence and the widespread dispersion of their readily
spoilable food. Similarly, the absence of the storage
habit among non-human primates can probably be attributed
to both their type of food (often soft, fleshy fruits) and
their lack of a permanent "homesite." A relatively concen-
trated source of preservable food and a permanent homesite,
den or place to store it, appear to be essential conditions
of the storage response. The ability to manipulate items
of food is also important and may account for the absence
of this habit in some mammaliam orders.

Food storage is relatively common among carnivores in
all latitudes even though one would expect their food to
spoil rapidly. Apparently the problem of food spoilage can
often be successfully circumvented through a variety of
novel behaviors practiced by those animals which normally
store their food. An animal's propensity to eat spoiled food
with impunity may also be of some influence on whether it

will store its food or not. Characteristically, carnivores

10

will cache surplus and partially eaten prey in small pits

or cover the cache with soil, ground litter, or even snow

in the northern latitudes. This burial or covering response
may actually be an attempt to preserve the food and not
simply to conceal or "hide" it, as has frequently been the
interpretation. Among North American carnivores, food
caching has been reported in the red fox (Vulpes £2123;
Murie, 1936; Scott, 1943), common badger (Taxidea 33x22;

Snead & Hendrickson, 1942), black bear (Ursus americanus;

 

Seton, 1926), wolf (£2213 lupus; Seton, 1926), mink
(Mustela yiggg; Seton, 1926), and both the shorttail and
longtail weasel (Mustela erminea and M. frenata; Seton,
1926).

The absence of food storage in all but a few species
of insectivores appears to be related to their diet and
feeding behavior. Insectivores are among the smallest
mammals and they normally have high basal metabolic rates
(Morrison, Ryser, & Dawe, 1959). As a consequence, they
require large numbers of insects and other invertebrates
and feed almost constantly. It is not surprising, then,
that among the few species in which food storage has been
reported, it appears to be highly developed. The habit of

short tailed shrews (Blarina brevicauda) storing live

 

snails in their runways is apparently quite common (Shull,
1907; Hamilton, 1930; Ingram, 1942). In fact, Ingram
(1942) found that their runways often have two distinct

chambers, one being a "culture" chamber where live snails

11

are kept and one a "discard" chamber where eaten snail
shells are placed. Ingram pierced the shells of snails in
the culture chamber and found that after 24 hours some of
them had been eaten and, as predicted, had been placed in
the discard chamber. The European mole (23123) also stores
earthworms in its runways (Degerbol, 1927). Apparently the
moles either decapitate or "paralyze" the worms by biting
them in the head before storing them. There have been no
reports of American moles storing food, prompting Hamilton
(1939) to ask why Blarina would practice storage when food
ought to be available all year round in its habitat.
Clearly more information on the generality of this habit
among insectivores would be instructive and might equal a
phenomenon as interesting as the classical ant and aphid
relationship in insects.

The habit of food storage is probably best developed
in the order Rodentia and is frequently referred to as

"hoarding." The dictionary definition of a "hoard" is an
accumulation of something for preservation or for future

use" (American College Dictionary, Random House, 1963).

 

The key word is "accumulation,' which is of food in this
case and which probably best distinguishes this form of
storage from those described for the other orders of
mammals. The general tendency for rodents to be either
seed eaters or herbivores, suggests that their food is

more suitable for long term storage than would be soft,

fleshy fruits or animal matter. It appears that the

12

grass-eating rodents generally do not store food, while
the seed-eaters do. The collection, transportation, and
preparation of hoard are all highly developed in many

rodents and will be discussed below.

Collection and Transportation

The sequence of behaviors involving the collection and
transportation of food items to be cached varies with both
the species and the type of food. Tree squirrels will
normally cut a large quantity of nuts from the top of mast
trees before returning to the ground for their collection
(Allen, 1943; Muul, 1968). Usually all twigs, stems, and
leaves are removed from the nuts before they are cached.
Some squirrels will also discard any nuts which are worm-
infested (Mailliard, 1931). Many rodents collect seeds
from the ground after they have fallen from trees, shrubs,
and other plants. Some species, in particular the kangaroo

rats (Dipodomyg) and pocket mice (gerognathus), have the

 

ability to glean small seeds from loose soil at phenomenal
rates. Burns (1958) reported that kangaroo rats could

place small millet-size seeds in their pouches at an average
rate of one seed every 0.2 second. The forepaws in

kangaroo rats are extremely important in the collection
process and are used to rake and sift small seeds from the
loose sand. Both the actual harvesting of the food and

the sequence of behaviors involving the caching of it

13

usually involve some use of the front limbs in rodents
(Morris, 1962; Eibl-Ebesfeldt, 1963; Ewer, 1965, 1967).
Collected food items are normally carried either in the
teeth or in special cheek pouches. These pouches may be
either external and fur lined as in the heteromyids (i.e.,
pocket mice, kangaroo rats) and geomyids (i.e., pocket
gophers); or internal and membranous as in some sciurids
(i.e., chipmunks, ground squirrels) and a few cricetids
(i.e., deermice, hamsters). The evolution of these special
pouches suggests that there has been a strong selection

pressure for food hoarding in some species.

Preparation of the Hoard
The prevention of food spoilage has given rise to many
novel behaviors among the rodents that store food. Western

pine squirrels (Tamiasciurus) normally cache pine cones in

 

moist debris or logs and sometimes even in pools of water
(Shaw, 1936). Shaw observed that this unique behavior
prevents the cones from drying out and that he was able to
keep cones for over two years if they were kept continually

under water. The habit of red squirrels (Tamiasciurus

 

hudsonicus) storing mushrooms in trees to dry is apparently

 

also quite common (see Hatt, 1929). Mushrooms and other
fungi are placed in the branches of trees where they will
dry and may later be transported to dry cavities for winter
storage (Buller, 1920; Hardy, 1949). Klugh (1927) has

proposed that the red squirrel stores "hard" and "soft"

14

objects differently. He observed that hard things such as
seeds and nuts were either buried or stored centrally,
while soft things such as fungi, fruit, or pieces of meat
were placed individually in the branches of trees. Klugh
noted that this squirrel normally does not bury or put into
piles things which could either mold or decay readily.
After the winter rains in the semi-arid regions of

California, the giant kangaroo rat (Qipodomys ingens)

 

stores the still green seeds of pepper grass (Lepidium

 

nitidium) and filarie (Oenothera) in small pits above its

 

 

burrows (Shaw, 1934). The seeds dry in these pits before
their removal to the underground burrows. Shaw dyed the
seeds in some of the pits with mercurochrome and replaced
others with kernels of rice, and discovered that the
contents of these pits were usually transferred into the
burrow by early June. It has also been reported that the

pika (Ochotona princeps), the only lagomorph known to store

 

food and which inhabits the talus slopes of western
mountains, frequently constructs sizeable "haypiles" which
cure in the sun before being removed to the shelter of
large rocks (Dalquest, 1948). However, this curing be-
havior is now of some question (see Barash, 1973). A
similar food-curing behavior has also been reported in the

mountain beaver (Aplodontia rufa), a colonial rodent found

 

in the moist, forested areas at high elevations in the
Pacific Northwest (Grinnell & Storer, 1924). Piles of

"hay" consisting of narrow—leaved lupine (Lupinus logipes),

15

which had been cut green and then piled supposedly for
curing, were observed along a stream in October of the
year. Grinnel and Storer noted, however, that these piles
could have been nesting material. As this plant was not
the mountain beaver's predominant food, this latter inter-

pretation is probably the more correct one.

Nature of the Hoard

Food hoarding in rodents is often referred to as
either a "larder" or a "scatter" type depending upon the
nature of the hoard (Morris, 1962). A larder hoarder
transports food back to the security of its home nest, den,
or burrow for storage. In contrast, a scatter hoarder
stores food not in its nest but rather in the general
vicinity of its homesite and not all in the same place.
Attempts to correlate these two types of hoarding with a
taxonomic classification have, however, failed for the most
part. Morris (1962) had proposed that larder hoarding was
typical of the myomorphs, while at least one hystricomorph,
the green acouchi (Myoprocta pratti), was a scatter hoarder.
Within the sciuromorphs, there appears to be no consistent
pattern to the occurrence of either type of hoarding. The

African ground squirrel (Xerus erythropus) is a scatter

 

hoarder, although it lives in a burrow and might be
expected to be a larder hoarder (Ewer, 1965). Ewer
proposed that scatter hoarding in this species may be a

vestige from their ancestral mode of arboreal life, as

16

their pattern of food burying is almost exactly like that

described for the European red squirrel (Sciurus yulgaris)

 

by Eibl-Eibesfeldt (1963). Scatter hoarding does seem to
be typical of most North American tree squirrels including
fox squirrels (Sciurus 23235; Allen, 1943), gray squirrels
(S. carolinensis; Brown & Yeager, 1945), and flying
squirrels (Glaucomys volans; Muul, 1968). However, the
pine squirrel (Tamiasciurus) may be either a larder or
scatter hoarder depending upon the type of food (Klugh,
1927). Small conifer seeds are an important dietary item
to Tamiasciurus and there is very little competition for
this food with other rodents. Neverthless, the pine
squirrel stores the complete pine cones in the center of
its territory, which enables it to defend its food supply
from conspecifics (Smith, 1968). The level of hunger in
some species will also influence whether the food will be
larder or scatter hoarded. In a laboratory study of
hoarding by chipmunks (Eutamias guficaudus), it was observed
that the animals cached more food outside their nests when
they were satiated than when deprived (Lockner, 1972).
Other sciuromorphs such as the prairie dog (Cynomys)
apparently never store food (Scheffer, 1947; Tileston &

Lechleitner, 1966).

Seasonality

 

Food storage has classically been interpreted as an

adaptation to the seasonal fluctuations in the animal's

17

food supply, with the assumption that the animal is capable
of "anticipating" or "foreseeing" an imminent shortage to
come and a need to store food (see Hamilton, 1939). In
northern latitudes, food supplies show a pronounced
seasonality in abundance that is directly related to
changes in both temperature and photoperiod. Most rodents
at these latitudes do in fact store food in the fall of the
year and use it during the following winter. Muul (1968)
found that flying squirrels are capable of using the change
in photoperiod as a highly reliable cue to begin intensive
storage of the mast crops during the fall of the year. He
proposed that this prevented both premature harvesting of
unripe nuts as well as the "wasting" of energy searching
for food at times of the year when it would either not be
present or unsuitable for storage.

Although food hoarding by deermice in northern lati-
tudes apparently occurs at all times of the year, it is
most intense during the fall (Criddle, 1950; Burt, 1946).
In a study carried out in a 300 acre area of grassland in
southern Michigan, it was observed that the prairie deermice
(E. m, bairdi) began hoarding in artificial nest boxes
about the middle of October while only a few seeds were
found in the boxes during the summer months (Howard & Evans,
1961). Seed storage was maximal in this study during
November and, surprisingly, most of the items were gone by
late December and all by March. However, when some prairie

deermice were kept in an outdoor enclosure, they hoarded

18

large quantities of food at all times of the year when it
was consistently plentiful (Howard & Evans, 1961). A
slightly different finding and one more consistent with the
field evidence, was reported by McCabe and Blanchard (1950)
for two species of deermice in California. They found that
when 2, 2522; were held in large outdoor cages, the greatest
amount of hoarding occurred during the season of the year
associated with the "time of the first rain and cool fall
weather" (McCabe & Blanchard, 1950, p. 22). Similarly,
caged E. californicus tended to hoard more in "late fall"
but only sporatically at other times. As food was apparently
continually available in the McCade and Blanchard study,
this would seem to suggest a seasonality in the intensity of
food hoarding by some species of deermice that is unrelated
to the mere abundance of their food.

In the southern latitudes seasonal changes in vegeta-
tion and food supplies tend to be less striking, although
this will vary depending upon the elevation of a particular
area and its closeness to large bodies of water. In desert
areas food supplies are closely synchronized with the
seasonal pattern of precipitation. In these regions, brief
periods of rain are normally followed by longer periods of
drought. After the rains, large numbers of small annual
plants germinate, grow, and then flower, leaving an
abundance of seeds that will remain dormant in the soil
until the next rains. Rodents inhabiting these arid

regions, in particular the heteromyids and probably also

19

deermice, store most of their food after these rains so
that it will be available during the subsequent dry periods.
In contrast, the tropical southern regions show no regular
seasonal changes in vegetation (Richards, 1973). Flowering
can occur at any time of the year and tends to vary with
each particular plant species. Food hoarding has been
reported in several tropical rodents under artificial

conditions, including the green acouchi (Myoprocta pratti;

 

Morris, 1962), the African ground squirrel (Xerus erythropus;

Ewer, 1965), and the African giant rat (Cricetomyg

 

gambianus; Ewer, 1967), but nothing is known about the
occurrence of this behavior in the wild. Since a variety
of food ought to be available at all times of the year in
the tropical habitats, there is little reason to assume
that food hoarding would show any pronounced seasonal

pattern.

Evolution

 

In environments where food supplies undergo seasonal
fluctuations, the storage of food during times of abundance
would have obvious adaptive advantage. Presumably those
animals which store food would survive better during
subsequent periods of food scarcity. These animals should
be selected for and would supposedly contribute more to the
next generation's gene pool than individuals that did not

store food.

20

Even if food supplies did not undergo seasonal
fluctuations, food storage may impose an adaptive advantage
for other reasons. An animal which can store food will
naturally have to venture forth from the safety of its home
less frequently than one that has no food reserves. If
each of these foraging trips exposes the animal to predation,
then it follows that the fewer the excursions the less
likely the animal might be caught. This may be a somewhat
tenuous argument, however, considering that all animals
which store food must during some particular period or season
spend a great deal more time collecting food than ones that
do not store. One possible explanation is that the threat
of predation during the time when food is most actively
stored is in fact less than it is when the food is scarce.
An abundance of available food for storing corresponds in
most habitats with a particularly dense vegetative cover
providing protection to small mammals that forage within it.
Also, in the northern latitudes, food storage among rodents
is most frequently done in the fall for use during the
winter. Presumably animals searching for food on the snow
covered ground of winter would be more heavily preyed upon
because of the lack of natural cover than they would be in
the fall.

Another possible eXplanation for food storage is that
its adaptive value is measured directly in the survival of
offspring which might utilize food caches near the nest.

As in the argument given above, juvenile animals could

21

avoid predation if they obtained their first solid food
from stores in the nest or burrow. This could be the case
in desert-inhabiting species where food storage is greatest
after winter rains and during or just before the reproductive
season. In northern latitudes, however, food storage does
not appear to be related to the reproductive season, and
food is probably not stored for the young.

The storage of food may also have adaptive value from
a bioenergetic standpoint. Obviously the time, and hence
the energy, required to collect and store food when it is
abundant is much less than the amount needed when food is
scarce. Another fortuitous aspect of food storage is that
the items assembled are usually high in caloric content.
Seeds and nuts are commonly stored, and these contain high
concentrations of energy, especially in regards to carbo-
hydrates (Davis & Golley, 1963). Animals living in the
colder climates normally undergo high metabolic demands
during the winter season. In the absence of hibernation
or seasonal dormancy, stored food would meet this demand
without necessitating additional expenditures of energy to
search for it. Some small rodents which do hibernate even
store food in their burrows prior to entry into hibernation
(Lyman, 1954). Apparently the food is used during those
periods when the animal awakes and also in the spring.

All of the above arguments for the evolution of
hoarding have centered around the thesis that food storage

frequently shows a seasonal pattern in frequenty. However,

22

several theories have been proposed in an attempt to
explain the initial motivation of the hoarding response.
The proposal that larder hoarding may have evolved as a
consequence of an animal's desire to eat in some secure
place (Morris, 1962; Ewer, 1965) would seem to be supported
by earlier findings that laboratory rats apparently hoard
to areas of maximum security (Bindra, 1948c). The obser-
vation that submissive wild rats in outdoor enclosures show
more larder hoarding than dominant animals (Calhoun, 1963),
also suggests that the original motivation in some species
might have been for security and protection.

Morris (1962) has proposed that under special circum-
stances scatter hoarding may have developed out of larder
hoarding. He suggested that in the green acouchi, a South
American hystricomorph, scatter hoarding may have evolved
concurrently with the species' propensity to become more
mobile and to shun a fixed home site. An animal which is
relatively mobile could seek refuge at any temporary site
within its home range if it had a scattered food supply
cached nearby (Morris, 1962). Ewer (1965) has also offered
the explanation that scatter hoarding in a species like the
African ground squirrel may have evolved as a result of
"food envy" or the practice of satiated animals attempting
to hide food from members of their own species. This
latter interpretation is very appealing in light of the
finding by Lockner (1972) that chipmunks will show more

scatter hoarding when satiated than when deprived. One

23

must consider, however, that the nature of the food probably
always has had an important bearing on whether it was larder
or scatter hoarded, as already pointed out for pine
squirrels.

There are several additional advantages unique to
scatter hoarding that would also act to aid its selection.
Ewer (1965) offered the novel explanation that storing food
away from the nest forces the animal to search for it every
time the animal is hungry and assures that stored items are
not consumed prior to a real shortage of food. If these
scattered hoards are not eaten, the seeds will eventually
germinate, providing future generations with a new source
of food. Undefended scattered hoards can also be utilized
by all members of the same species, regardless of which
individuals made the cache.

The discussion of the evolution of hoarding up to now
has centered around the explanation that this seasonal sort
of behavior is apparently "adaptive." Like most evolutionary
explanations of behavior there is a paucity of good experi-
mental evidence either to support or refute such ideas.
Adaptiveness can be defined as anything that increases the
animal's ability to survive and leave fertile offspring.

The proposal that food storage by deermice in northern
latitudes leads to higher survival has never actually been
tested, although several observations have been made that
might be construed as evidence for this. Howard (1949) has

reported that cold temperatures together with shortages of

24

food probably contribute to winter mortality of deermice in
Michigan. He observed prairie deermice (E, m, bairdi) dead
in next boxes after cold waves in March which had apparently

died from starvation. In contrast, E'.$' noveboracensis

 

when placed outdoors without protection from wind or snow
during February and March could survive if given ample food

(Howard, 1949). Under laboratory conditions, £3 E: bairdi

 

seems to be able to tolerate cold better than 3. l.

noveboracensis (Sealander, 1951), which would indicate they
both can survive if given sufficient food. The presence of
daily torpor in both species probably also leads to higher
survival and the prolongation of food stores (Howard,
1951).

An obvious experimental approach to studying whether
food storage is critical to these animals, would be the
removal of the food stores from some individuals but
allowing others to retain and accumulate their food. The
field situation described by Howard and Evans (1961), in
which nest boxes were placed in a 300 acre area of habitat,
would be ideal for such an experiment. If the boxes were
left out long enough, the deermice would eventually get
accustomed to using the boxes, especially if other speCies
were symstematically excluded from them by the eXperi-
menter. Two identical test areas could be set up where in
one area the mice would be allowed to accumulate food in
the boxes, but in the other area the food would be periodi—

cally removed (preferably as often as possible). Survival

25

and reproduction on the two areas could then be measured
and compared. A higher survival and reproductive rate on
the area with food stores would be good evidence for the

"adaptiveness" of this trait.

Experimental Studies

Since Wolfe's pioneer work in 1939 entitled "An
exploratory study of food-storing in rats" published in the
Journal of Comparative Psychology, numerous investigators
have attempted to answer the question of what are the
critical variables affecting hoarding behavior. Ross,
Smith, and Woessner (1955) in their review of the hoarding
literature, arbitrarily classified all of the variables into
two groups: organismic and situational factors. In the
review below, I have used an outline of topics similar to
theirs and have retained their use of the words "organismic"
to describe all of the possible endogenous motivating
factors of hoarding, and "situational" to include all of
the environmental factors. The two groups will not always

be mutually exclusive.

Organismic Factors

errivation.-~Studies on the effects of food deprivation

 

on hoarding have led to some contradictory results depending
on the species. Morgan 35 31. (1943) found that Lashly
brown rats deprived of food for periods of 12, 24, 36, and

48 hours showed higher hoarding scores as compared to scores

26

during satiation. Although their results were ambiguous
due to a poor experimental design, they concluded that an
accumulation of a physiological deficit was necessary to
produce hoarding because deprived rats took several days
before they began to hoard. They believed hoarding was an
"all or none" behavior as the hoarding scores were
essentially the same for all four experimental groups (12,
24, 36, and 48 hour deprivations). They also found that
there was some competition between the tendency to eat and
hoard. Hoarding scores were higher when measured in the
second half hour of the test situation than when measured
in the first half hour, which was spent in eating. Females
also hoarded more than the males and boarded sooner during
deprivation than males. Stellar and Morgan (1943) verified
these results in another study in which they compared the
hoarding of rats kept on an lB-day deprivation schedule
(given food for only 1 hour daily) with ones kept on a
satiation schedule (unlimited food). In the subsequent 7
days of hoarding trials under deprivation conditions for
both groups, the deprived rats showed high levels of
hoarding almost immediately, whereas those that had been
satiated took several days before their scores reached
maximal levels. They concluded "that the effects of a
deprivation cycle accumulate over a period of days until
they reach some critical level of deficit which instigates
hoarding . . . (and) . . . the amount of hoarding does not

vary quantitatively with the amount of deprivation but

27

rather appears at its maximal amount when deficits in the
body accumulated through deprivation reach a 'threshold'
level" (p. 53). This theory is commonly referred to as the
"deficit hypothesis" of hoarding. Smith and Ross (1950a)
found essentially the same results with Wistar rats on a
35-day deprivation schedule during which rats were given
one-half hour each day to eat food weighing 1/8 of their
body weight and then were tested for 1/2 hour. Once
hoarding began in the second week of the schedule it did
not drop off significantly for the rest of the 35 days.
The results of eXperiments on food hoarding in the
golden hamster agree with those found for rats. Smith and
Ross (1950b) reported that hamsters hoarded significantly
more food pellets during a deprivation schedule of only 1'
pellet a day as compared to satiation of 3 pellets daily.

Lockner (1972) also reported that in the chipmunk (Eutamias

 

ruficaudus) deprivation enhances both the amount of food

 

hoarded and the Speed at which it is hoarded.

However, not all of the studies have supported the
deficit hypothesis. Bindra (1948b) demonstrated that
deprivation was not a necessary condition to elicit
hoarding when he found that rats would hoard saccharin-
treated mash placed in bottle caps even when non—deprived.
Bindra altered the experimental situation by using closed
alleys, and his conclusions are therefore not directly
comparable to those of Morgan st 31. (1943). Ross and

Smith (1953) also reported that the deficit hypothesis

28

failed to explain hoarding in laboratory mice. Using C3H
strain mice they found that deprivation tended to have an
inhibitory effect on hoarding. They used three experi-
mental regimes of 15 days each in which satiation was
alternated with deprivation and found that the hoarding
scores were highest when satiation was the initial con-
dition. In an attempt to resolve these conflicting results
the adequacy of the laboratory diets was investigated, and

these studies are discussed in the next section.

Dietary and energetic deficiencie§.--Wolfe (1939) was
the first to study the importance of a nutritional deficiency
on hoarding behavior. He found that when the rats' normal
diet of food pellets was supplemented with milk, lettuce,
and cod liver oil, their hoarding scores were significantly
less than when rats were given only the food pellets but in
sufficient quantity to maintain their weight. Unfortunately
this study may have compounded the factor of deprivation or
not enough food with that of nutritional deficiency. It is
very plausible that the rats kept on a diet "sufficient to
maintain body weight" were in fact being deprived of their
normal amount of food. Wolfe may have been comparing
supplimented rats with deprived rats which would confuse
his results. In any case, he concluded that "the absence
of certain food elements provides a distinct motivation

for hoarding" (Wolfe, 1939, p. 107).

29

A similar effect of a supplemented diet on food
hoarding in mice was reported by Ross and Smith (1953).
They placed mice on an experimental regime of 5 days of
satiation (food 29.ll§3) followed by 5 days of satiation
plus supplemental lettuce and sugar cubes, and then finally
on 5 days of satiation again. During the satiation-
supplementation phase hoarding dropped off by 54%,
apparently as some undefined need was satisfied. When
supplementation was stopped, the amount hoarded in the next
5 days rose by 51%, or to the level at which it had
previously been.

Several approaches have been used to study whether
deprivation of specific food elements leads to a body
tissue deficit which in turn triggers hoarding. Stellar
(1943) manipualted the sugar metabolism of rats by injecting
insulin, epinephrine, and glucose. It was predicted that
insulin would increase hoarding because of its lowering of
the blood glucose levels and the production of a glucose
"deficit." On the other hand injections of epinephrine
(i.e., adrenalin) and glucose were expected to produce the
opposite effect on hoarding. Only epinephrine showed the
expected effect, and therefore his hypothesis is still
equivocal.

Bindra (l948a,b) investigated the importance of
specific components of the diet including fats, vitamins,
minerals, salts, proteins, and carbohydrates on the

hoarding response. He found that hoarding increased when

30

both carbohydrates and proteins were given together. There
was no evidence of an effect by any of the other components
which were studied.

In an attempt to investigate whether a dietary
deficiency was affecting overall metabolism rather than a
Specific tissue.deficit, Stellar (1951) manipulated thyroid
functioning in three groups of rats. He proposed two
contrasting hypotheses for the role of metabolism on
hoarding: (1) that hoarding served as a heat production
activity like any other bodily activity, or (2) that
hoarding was a heat conservation activity similar to nest
building. He also theorized that any condition which would
either raise or lower metabolism ought then to affect
hoarding. In his experiment one group of rats received a
thyroid depressent, one group had their thyroids removed,
and one group received thyroid injections. There were no
significant differences among the hoarding scores of the
three groups and Stellar concluded that "general metabolic
changes do not underlie the hoarding behavior of the
laboratory rat" (Stellar, 1951, p. 297).

In summary, any general conclusion about the effects
of both deprivation and dietary deficiency upon the
hoarding response is still equivocal. Clearly there is a
need for a more standardized approach to these questions
using comparable test conditions and subjects of similar

age, weight, sex, and feeding experience.

31

Sex differences.--Many researchers have reported
significant sex differences in the hoarding response (i.e.,
in rats, Wolfe, 1939, Morgan 22.31., 1943, Smith & Ross,
1950a; in mice, Smith & Ross, 1953a; and in hamsters,

Smith & Ross, 1950b). As already mentioned, hoarding by
rats is typically quite high if the animals are kept on
some sort of deprivation schedule. However when food is
continuously available, a sex difference usually emerges
with the females hoarding more than males. Wolfe (1939)
attributed this to some sort of malnutrition in females
caused by previous lactation, while Calhoun (1963) reported
increased hoarding by female rats during lactation. Marx
(1950b) proposed that high levels of female hoarding might
possibly be due to heightened activity during their estrous
cycle. Studies by Lapetite and Soulairac (1951) and
Herbert 23.21' (1972) have also implicated the estrous
cycle as being a factor, but not because of an increase in
sexual arousal or general activity as Marx predicted.
Lapetite and Soulairac (1951) brought female rats into
permanent diestrous by the administration of an anti-
thyroxine drug called methyl thioruracil and noted increased
hoarding. Male subjects showed a decrease in hoarding.
Both sexes also decreased their food intake. Lapetite and
Soulairac hypothesized that the mechanism underlying
hoarding behavior was different in each sex with males

hoarding under a physiological condition related to hunger,

32

and females hoarding because of some sort of extra-thyroid
condition which might be called "sexual equilibrium."

The study by Herberg g£_gl. (1972) also confirmed that
female rats tend to hoard more during the diestrous days of
their cycle. Apparently when sexual and general activity
(i.e., wheelrunning) is maximal, hoarding behavior falls
off. This same sort of phenomenon was found by Kinder
(1927) for nest building by rats. The greatest amount of
nest building was also observed during the diestrous days
of the cycle. The possibility that hoarding behavior is
modified or displaced nest building was examined by Herberg
35 21. (1972). They found that satiated females hoarded a
much greater proportion of cotton-wool pledgets than did
food-deprived female and male control groups, although food
pellets predominated. As the satiated subjects still
preferred food pellets to the nesting material on all but
the first 2 test days, it was concluded that hoarding was
not simply a manifestation of nest building activity.
Instead, Herberg gt 31. (1972) proposed two alternative
explanations for the sexual difference seen in hoarding.
They hypothesized that perhaps estrogen imposed some sort
of restriction on food intake in females, resulting in the
release of the hoarding response in much the same fashion
as seen in food deprived males. Kennedy (1969) had
previously given evidence that estrogen reduces food intake
and limits body weight in female rats to a level below

males. This hypothesis, however, would seem inconsistent

33

with Lapetite and Soulairac's (1951) observations that their
treatment of anti-thyroxine drug resulted in reduced food
intake in both sexes but with a resultant increase in only
the hoarding of females. The alternative hypothesis

offered by Herberg 35 21. (1972) is that the difference in
hoarding behavior is yet another "long term repercussion of
perinatal androgen activity" (p. 190).

Not surprisingly, consistent sex differences were not
observed in several of the reported studies of hoarding
behavior. Hunt gt‘gl. (1947) and Stellar (1951) found that
sex differences contributed little if anything to the total
variance in their studies. Marx (1951) also found that the
percentage of rats which hoard was about the same for both
sexes and in fact male rats would hoard more than females
when the rats were given only 30 minutes in which to hoard.
When given 24 hours, then females tended to hoard more than
males. Apparently the experimental situation frequently
interacts with sex leading to some of the differences

reported in the literature.

Neural contr01.--One would naturally expect that

 

because the hoarding response depends upon sensory input,
any interference with the cortical sensory areas should
affect hoarding activity. Both the early and recent
research has centered around attempts to isolate areas of
cortex directly controlling this response and not sur—

prisingly the results have often been contradictory.

34

Zubeck (1951) found that small lesions in the neocortex
regardless of location increased hoarding in laboratory
rats. However, Stamm (1953) found the opposite effect that
bilateral cortical lesions reduced the level of hoarding in
rats. He also found that small lesions of the median
cortical surface affected the level of hoarding to a
greater degree than those in other larger areas on the
cortical surface (Stamm, 1954a). Vanderwolf (1967) has
also demonstrated that anterior thalamic lesions result in
the depression of hoarding by rats.

Most recently Wallace and Tigner (1972) have confirmed
that neocortical lesions lead to increased hoarding in rats
under both 29.l£2° and food deprivation conditions. They
explain the apparent contradiction of both Zubeck's and
their own findings with that of Stamm's (1953, 1954a) as
probably due to Stamm's selection of animals for study
which showed maximal preoperative hoarding. As a result
these animals would have been less likely to show any
further increase after the operation. In contrast, for
both the studies of Zubeck (1951) and of Wallace and Tigner
(1972), animals were selected under ad lib, conditions and
probably were hoarding at a very low preoperative level.
Appropriate experimental manipulation therefore would be
expected to significantly increase hoarding over that of
controls. Wallace and Tigner (1972) also found that rats
with hippocampal lesions did not hoard significantly more

than controls under either ad lib. or food deprivation.

35

Again, however, this is an apparent contradiction of the

work by Wishart 25 El. (1969) in which they found the exact

Opposite effect.

The role of thelateral hypothalamus in food hoarding
was first implicated by Herberg and Blundell (1967) when
they found that electrical stimulation of that area
elicited hoarding in satiated rats. Previous work had
already demonstrated the importance of this area on
feeding, and lesions there diminish both feeding and
hoarding (Blundell, 1970). In another study Herberg and
Blundell (1970) also found that although lesions in the
ventromedial nucleus of the hypothalamus increased food
intake, there was no increase in hoarding. Later when their
lesioned subjects became very obese, hoarding could not be
elicited even after 16 hours of deprivation. As a conse-
quence of their findings, Herberg and Blundell (1970) have
presented a model for hypothalamic control of hoarding.
First they propose that hoarding is motivated by long term
nutritional needs or "deficits" (i.e., body weight loss or
changes in the size of fat reserves) mediated ultimately
though the lateral hypothalamus feeding area. Secondly,
they propose that hoarding, unlike feeding, is not under
the inhibitory control of the ventromedial satiety area
which probably responds to elevated blood glucose levels or
some other short term change following feeding. Stellar's
early work involving the injection of massive doses of

glucose and its failure to have any effect on hoarding

36

behavior would seem to give additional support to this
model. More recent work by Blundell and Herberg (1973) has
likewise tended to support their model as they have
demonstrated that electrical stimulation of hypothalamic
sites other than the lateral area did not induce hoarding
even though locomotor activity was increased.

Whether the model described above by Herberg and
Blundell (1970) holds for small mammals other than laboratory
rats remains to be seen. In fact, this may not be the case
in laboratory mice which, as already noted, do not increase
their hoarding after food deprivation (Ross & Smith, 1953).

The results of my work with deermice (Peromyscus) would also

 

tend to contradict this model, since there was no consistent

relationship between weight loss and hoarding.

Learning and early experienc§,--Wolfe (1939) was the

 

first to study the relative importance of early experience
on adult hoarding. He compared rats raised with food
pellets to those fed only powdered food, and found that the
rats with previous pellet experience hoarded more. Marx
(1950a) interpreted these results as suggesting that
hoarding was a stimulus-response behavior formed primarily
from the fusion of "earlier and independently learned
'homing' and pellet seizure habits" (p. 91) aided by some
form of reinforcement. Smith and Ross (1953a) attempted to
verify this hypothesis by comparing C3H mice reared on

pellet food to those reared on a liquid diet (similar to

37

Wolfe's original study with rats). When hoarding trials
were conducted, the two groups showed no significant
difference in the level of hoarding. They concluded that
prior experience with the transportation of food was not
necessary in mice.

Stellar and Morgan (1943) approached the problem a
little differently and instead studied the effect of prior
experience with the hoarding apparatus itself. They con-
cluded that "emotional adaptation" to the apparatus was
what appeared in their data as "experience." Bindra
(l948c) later disagreed with this conclusion and stated
that in fact the opposite was true, with some form of
anxiety (emotion) being necessary to even produce the
hoarding response. More recently and probably much to the
chagrin of these earlier workers, Blundell and Herberg
(1973) have demonstrated that they can elicit the hoarding
response in naive rats which have never previously hoarded
by electrically stimulating the lateral hypothalamus. My
own observations of deermice (Peromyscus), indicate they
too can hoard almost the instant they are given food to
carry even without any sort of prior experience.

The effect of early feeding frustration on hoarding
also captured the interest of many workers. Hunt (1941)
found evidence that depriving rats in infancy augmented the
amount of hoarding exhibited by adults after food depri-
vation. Later work by Hunt 32 31. (1947) also provided

what was thought to be additional support for such an

38

effect on hoarding. However, McKelvey and Marx (1951) took
exception to the statistical procedures used in the original
study. In an attempt to repeat Hunt's study, they were
unable to find any evidence to support his conclusions.
Nonetheless, Marx (1952) proposed that increased rates of
hoarding by infantile-deprived rats might be the result of
their retention in adulthood of an increased feeding rate
with consequent effects on hoarding.

Albino and Long (1951) also repeated Hunt's experiment
but modified it by insuring that all groups prior to adult
deprivation tests were kept on a satiation schedule long
enough to insure equivalence of pre-test "hunger." Their
results did agree with Hunt's but they pointed out that
their ”feeding frustration effect" may have been a conse-
quence of the differential experience their experimental
animals had in transporting food. Manosevitz (1970) con-
curred with this latter explanation and also attempted to
replicate Hunt's original study, although this time with
inbred strains of mice. He found no evidence to support

Hunt's earlier conclusions.

Genetic variation.--For many years there was a con-

 

tinuous debate concerning whether hoarding was instinctive
(McDougall, 1923; Morgan, 1947; Thorpe, 1963) or learned
(Holland, 1954; Marx, 1950a; Mowrer, 1960). Today, how-
ever, one can only conclude that the debate was inconclusive

and that more attention needs to be paid to the relative

39

contribution of both the environment and genotype to the
hoarding response.

Stamm (1954b) was the first investigator to report
strain differences in the amount of hoarding by rats. In
1956, he pursued this further by mating a high hoarding
strain with a low hoarding strain and then backcrossed some
of the Fl's with the low strain. Although this method was
not complete enough to clearly demonstrate the mode of
inheritance of this trait, Stamm (1956) did conclude that
hoarding was dependent on a single dominant gene. The Fl's
were similar in hoarding scores to the high strain and the
backcross generation had a bimodal distribution in their
hoarding scores. Koski (1963) also reported that a single
gene theory could possibly explain inheritance of this
trait in golden hamsters. Although he found no significant
differences in the scores of the two hamster strains he
studied, he did get a bimodal distribution of scores for
both strains. Smith and Powell (1955) found a significant
difference in the hoarding scores of several inbred strains
of mice but attributed this to each strains' characteristic
"emotionality."

More recently the contribution of genetic factors to
hoarding has been investigated by Lindzey and Manosevitz
(1964), in which they too found significant differences in
the level of hoarding among several strains of inbred
laboratory mice. Significant interactions between genotype

and the environment have also been found (Manosevitz, 1965;

40

Manosevitz gt 21,, 1968). Manosevitz and Lindzey (1970)
have emphasized that the environment can influence the
"heritability" of the genetic factors involved and that if
one attempts to select for high or low strains, the results
are dependent upon the specific environmental situation.
They concluded that the determinants of hoarding are complex
and involve the genotype, components of the environment,

and their interaction.
Situational Factors

Competition-social interaction.--The effect of con-
specifics on the hoarding response was investigated by
Miller and Postman (1946). In their experiment they used a
central hoarding bin from which rats from several cages
could hoard. When the rats were permitted to hoard
simultaneously some preferred to take food from the bin
whereas others hoarded more from the cages of the other
rats. The rats whose food was stolen showed no resistance
to the thefts suggesting that the goal of hoarding is the
activity itself rather than the result of such activity
(Miller & Postman, 1946).

Hoarding by rats in a group situation was also studied
by Ross st 21. (1950). They found that a dominance-
submission relationship was established in which submissive
rats hoarded less under the group condition as compared to
when they were isolated. The best hoarder was the

dominant rat. Denenberg (1952) extended this study and

41

raised rats in a group and then habituated them to indi-
vidual living conditions. When tested under isolation
conditions, they tended to hoard more than they did under
the group situation where more time was spent eating than
hoarding. Calhoun (1963), however, has found that low
ranking Norway rats hoard more than high ranking ones when
kept together in large outdoor enclosures.

In another study, Lindzey and Manosevitz (1964)
examined the effect of group competition in mice using two
experimental conditions of deprivation. One condition
consisted of depriving the mice as a group and forcing them
to compete for food. For the other condition, mice were
deprived while isolated in individual cages. When the mice
from the two groups were later tested under identical
conditions in individual hoarding cages, there was no
significant difference between the two groups' hoarding
scores. Lindzey and Manosevitz (1964) concluded that there
was little evidence for the direct effeCt of group competi-
tion for food upon the hoarding response.

Recently Terman (1974) has reported on a study of
several laboratory pOpulations of prairie deermice (P.

maniculatis bairdi) and the possible correlation of

 

hoarding behavior and population growth. He observed that
hoarding behavior always began prior to population asymptote,
and hypothesized that increasing intraspecific competition
or social pressure may have been a causative factor.

Typically a founding female was the first and most

42

consistent hoarder, although a few other mice also partici-
pated in the hoarding. Terman (1974) also speculated that
the "establishment and maintenance of the hoard appears to
be an effort on the part of the hoarders to spatially
organize the populations with respect to the location of

food storage and consumption" (p. 144).

gage, lighting! and familiarity of apparatus.—-The
importance of lighting or illumination on the hoarding
apparatus and its effect on hoarding scores was first
noticed by Wolfe (1939). He found that lighting the
hoarding alleys more than doubled the number of pellets
retrieved by rats. Waddell (1951) also found this to be
true when measuring hoarding in hamsters. When he lighted
their food bin, the hoarding rate was doubled. Stellar
(cited by Morgan, 1947) found that differential lighting of
the cage and alley decreased hoarding when the source of
light was over the cages. Apparently if the home cage was
shaded, hoarding was increased. Manosevitz (1965) likewise
found that if the home cage is kept dark, higher hoarding
scores can be obtained.

The familiarity of the cage is also important as
pointed out by Stellar and Morgan (1943) and Viek and
Miller (1944). Optimum hoarding was obtained in their
studies of rats when a familiar cage, alley, and food were
used. This familiarity was apparently related to olfactory

as well as visual cues. They also concluded that rats tend

43

to transport food to places of maximum security, in this
case their home cage. Bindra (l948c) also found that the
type of alley was important. "Open" alleys increase the
frequency of hoarding as compared to "closed" ones. Bindra
noticed that the latency to enter the alleys increased with
the use of open ones and interpreted this as indicating
insecurity or "shyness" in the rats. Partial support for
this hypothesis has been provided by Tigner and Wallace
(1972) who observed greater hoarding in anosmic rats
(olfactory bulb removed) than in sham operated or blind
rats. They noticed an increase in the latency to enter the
hoarding alley by the anosmic rats and attributed this to
their increased shyness.

Several studies have demonstrated that the length of
the alley is an important factor affecting the amount of
hoarding (Miller & Viek, 1950; Smith & Ross, 1950a, 1953a;
Wolfe, 1939). In general shorter alleys result in higher
hoarding scores apparently because the distance over which
the pellet has to be transported is less. The shorter the
alley the less the "anxiety" experienced by the animal as he

retrieves the food to his home cage (Ross SE al., 1955).

Stimulus characteristics.--Several studies have
attempted to investigate the stimulus characteristics of
hoarded material to see whether there was some sort of
relationship between the animal's need and what is hoarded.

Although relevance to a need has an effect on the intensity

44

of the food hoarding response in rats (Wolfe, 1939;
Stellar & Morgan, 1943), it apparently is not a necessary
characteristic of the hoarding stimulus ESE.§E¢ Bindra
(l948a) found that rats hoarded saccharin flavored food
even when not deprived. He also found that they would hoard
cotton dental plegets soaked in water when they had been
satiated with water (Bindra, 1947). Smith and Ross (1953b)
similarly found that laboratory mice would hoard more wet
or dry cotton pellets than food pellets, even when they had
experienced neither food nor water deprivation. As there
was also plenty of nesting material, the explanation that
possibly the cotton was needed for nest building seemed
unlikely.

The relevance of need as an important characteristic
of the hoarded material has been further investigated by
McCann 22.2l' (1964). They found that food deprived rats
hoarded fewer paper strips than did satiated rats. Severe
water depression likewise tended to depress hoarding of
both paper strips and food. In conclusion they state that
severe deprivation of food or water depresses the hoarding
of materials which are incapable of alleviating that
deprivation. Tigner and Wallace (1972) have similarly
found that food deprived rats hoard more food pellets than
wooden blocks. When not under food deprivation, they
hoard more blocks than food. Licklider and Licklider
(1950) in an earlier study, had also shown that non-

deprived rats hoarded more tin foil wrapped pellets than

45

plain ones when given a choice. A different finding has
been reported for prairie deermice, however, as they
apparently will not hoard wooden pellets or cotton balls
when non-deprived in the laboratory (Rice, 1972). Although
"novelty" seems to be important in rats, this may not be
the case in deermice.

The stimulus cues by which an animal might identify
objects to be hoarded have also been studied to a limited
extent. The role of vision and olfaction has been studied
by Tigner and Wallace (1972) in rats. They found that
neither visual nor olfactory cues were necessary for the
elicitation of hoarding. In fact anosmic rats (olfactory
bulbs removed) hoarded significantly more food pellets
under both 22.l£2 and deprivation than did blind and sham
operated rats. Tigner and Wallace (1972) concluded that
several cues are available to an animal and that there is
no evidence that cues for only one sensory modality are
being used.

Muul (1968, 1970) has observed that flying squirrels
use olfactory cues to recognize nuts previously stored by
themselves or other individuals. In his study "new" nuts
were preferred 4 to 1 over previously stored ones. This
behavior apparently prevents the wasteful restoring of
food by this species and insures that a large accumulation
of nuts will be gathered for winter. This is not the case
in prairie deermice, however, for in the laboratory they

show no apparent preference for either fresh food pellets

46

or previously hoarded ones, when given a choice of both
kinds to hoard (Rice, 1972).

Studies on the stimulus characteristics of objects
hoarded by hamsters in the laboratory have implicated
novelty or variety as being an important determinant as to
what things are preferred for hoarding. Scelfo and Hammer
(1969) showed that non-deprived hamsters preferred sucrose
pellets to plain chow pellets for hoarding purposes. They
concluded that taste was probably responsible for the
preference. However, in a follow-up study Hammer (1972)
found that variety was probably all that was necessary and
that citric acid flavored pellets were preferred over plain
chow, sucrose over citric acid, and glass beads even over
sucrose. The glass beads were highly preferred. In each
test there was a decrease in preference by the fifth day of
the experiment indicating that "novelty" was probably
important. Hammer (1972) speculated that this apparent
"innate liking for (glass beads) and disposition to hoard
these articles may be comparable to the reaction to innate
releasing stimuli in other behavior systems" (p. 140). In
fact these stimuli may even be "super normal releasers" as

compared to ordinary food items or nesting materials.

Temperature.-—McCleary and Morgan (1946) were the first

 

to investigate the effect of temperature on food hoarding
in laboratory rats. In their preliminary work, rats were

tested on alternate days either under temperatures between

47

lO°-15°C or between 25°-30°C. When tested under the colder
temperature range, the level of hoarding was almost seven
times higher than it was under the warmer temperatures. As
the experimental procedures were different under the two
temperature conditions, the test was repeated except that
the lower temperature range was from 25°-30°C and the

higher range from 40°-45°C. Again tests were conducted on
alternate days with the rats again hoarding more under the
lower temperatures (although heat stress was almost cer-
tainly an important factor). In their next experiment,

they housed 6 animals (4 males and 2 females) in a room
where the temperature was allowed to fluctuate with outside
ambient temperature with a range from 8°-34°C. The rats
were given unlimited food and the number of food pellets
carried from an alley into a home cage in a 24 hour period
was the measure of hoarding. As predicted they found an
inverse linear relationship between the logs of the hoarding
scores and the mean daily temperatures. McCleary and

Morgan (1946) provided three possible explanations of the
observed increases in hoarding. It may have been an attempt
on the part of the rats to (1) increase heat production
through muscular activity, (2) insure a supply of food to
meet increased metabolic demands, or (3) supply insulation
material for the home cage. Although McCleary and Morgan
raised many interesting questions, their study has apparently

never been repeated with rats.

48

Ross and Smith (1953) did perform a similar experiment
to McCleary and Morgan's but with female laboratory mice
(C3H) as the subjects. They had predicted that because
mice respond differently than rats to the stress of food
deprivation, mice would also respond differently to changes
in ambient temperatures as it effects their hoarding. Their
test temperatures varied from about -2° to 29°C, but there
was no relationship between the lowest daily temperature
and the highest hoarding scores. In fact on 17 of the 20
days of testing, hoarding varied directly with temperature.
Temperatures below approximately 15°C appeared to inhibit
hoarding.

The results just reported for the Ross and Smith
(1953) study might tend to cast some doubt on the naturalis-
tic interpretation that mice hoard when the temperature
drops in order to have food for subsequent cold weather
periods. A criticial appraisal of their study suggests,
however, several possible explanations for this apparent
contradiction to the more classical interpretation. First,
their experimental situation consisted of a variable
temperature which could fluctuate nearly 27°C from one day
to the next. A more appropriate test might have been to
keep the test temperature constant to see whether it was
temperature change or the temperature per se that was
important. Secondly, their subjects, besides being all of
one sex (females), were also inbred mice on which the

selection for hoarding had been relaxed for many generations.

49

Genetic variability was greatly diminished in these mice.
Lindzey and Manosevitz (1964) have shown that there are
strain differences in hoarding by laboratory mice and it is
possible that another strain might have demonstrated an
opposite effect of temperature. The results of my own work
to be discussed later, show that the response of deermice
to ambient temperature apparently has some relationship to

their latitudinal and climatic origins.

Photoperiod.--Muul (1968) was the first investigator

 

to hypothesize that photoperiod had an effect on hoarding in
at least some rodents. In his study of flying squirrels

(Glaucomy§_volans), Muul found that he could increase their

 

level of hoarding by either giving the squirrels an
appropriate photoperiod for fall or a shorter than normal
one. In fact the level of hoarding was directly related to
the length of the dark'phase of his light-dark cycle. Time
tests indicated, however, that it was not merely a conse-
quence of having more time to hoard. The response appeared
to be completely independent of all temperature cues. In
fact, cold temperatures tended to inhibit the squirrels'
activity and probably would similarly affect their amount
of hoarding.

According to Muul (1968, 1970) the capability of
sensing photoperiod changes is advantageous because it
allows a precise timing of food storage to the time of year

when food suitable for hoarding is readily available. In

50

the northern states, hickory nuts and acorns are the main
items found in caches made by flying squirrels. This food
is used until the next mast season and during the summer,
insects and vegetable matter often supplement their diet.
During mid-September, the intensity of the storage response
increases and reaches a peak in November. Flying squirrels
apparently hoard at all times of the year, but the greatest
intensity is in these autumn months. Muul proposes that
their use of photoperiod changes for a cue to begin inten-
sive hoarding prevents premature gathering of unripe nuts
which might spoil or be less nutritional. It also assures
that they will not "waste" their energy searching for food
that would be unsuitable for storage at other times of the
year. The results of my own study of hoarding in deermice
(geromyscus) indicates that the response of some subspecies
to photoperiod might be a purely opportunistic one in which
they use the longer periods of darkness to be active

longer.

METHODS AND MATERIALS

Subjects
The mice used in this study represented 5 different
taxonomic groups of 3 species (Table 1). Two subspecies

were chosen from within both a, lgucopus and a, maniculatus

 

 

so that north-south comparisons of their hoarding scores
could be made. 33 l. castaneus were selected because of

their proximity to the equator, whereas E! l. noveboracensis

 

were used because they were available locally from brush
and forest habitats around East Lansing, Michigan. Both of
these subspecies are found at practically equivalent
altitudes. Within £3 maniculatus, stocks of 33 m, blandus
were available from the deserts of the southwest, whereas

23 g, bairdi were trapped in cornfields bordering the campus
of Michigan State University. 2, eremicus eremicus were
also utilized because they represented an entirely differ-
ent subgenus than the other two species and are found in
the same habitat as 3. m, blandus. The latitude, elevation,
habitat, and climatic conditions from which each of the

above stocks was obtained are summarized in Table 2.

51

52

 

 

 

mom «mm mampoe

NH mm msOHEmum macaamum m>EOH>Eonmm
«m «o msmcmummo msmoosma msmoooma msomxsoumm
em we mflmcmomuonm>0c msmoosma msmooswa msomxeonmm
em vo moccmHn msumasoflcms mopmHsoHcmE msom>Eonmm
we vm Houfimn msumHSUHcmE msumasoflcme msomAEoumm

mmHmEmm mmamz mEmz mEmz msouo mEmz
Hmnfidz camaommmndm owmwommm mmwommm oaumcmmnsm

 

.mcsum mass as pomp mos: mo Honesz can xmm .mmwanOflpmamm oflumEmummmun.H mabme

53

.omoom mo ummmcuuosnnuuoc moafifi mm a

.oowxmz .mmmouwomm “out

Houmsaxoummm oowxmz 3oz .musnmouoq Hows

 

 

 

 

 

 

 

 

 

 

Apmcoocmnm Hm>mq z .hmoma OUflxmz

Agony Aumdv a o>auomv mom can .couomsmnu msmcmummo
Ha m\H ma .m.mH m.om moamamauoo ummz z .omoma new mmmouwomm mamoosmfl .a

Aomov Awasbv coaxmz msoflEmum
0H m\H «a .H.mn m.mm unmmoo omma z .Hmomm 3mz .omuom msoflamum .a

Aomav Audubv oowxmz moccmHn
0H m\a va 4H.ml m.mm unmmwo omma z .Hmomm 3oz .owcom mopmasoflcme .d

Annoy Audubv .Hz mflmcmomuonm>oc
m m\H ma m.mu m.s~ ammuom sum 2 .mwomv .mcamcmq .m msmoosma .a

lawns issues Am>fluomc .az Henson
m n\H ma m.mu ~.p~ mcamamcuoo arm 2 .mv.mv .ocnmcmq .m mspmasuacme .a

umwuuonm ummmcoq .cwz .xmz consummu Amnopmfiv opsuflumq cflmwuo mmflommmndm
Amway numcmq woo “oovmmusumnomEme muons opsuwuam
manna can: “nuance

 

.aosum mane cw pom: cow: 0:» mo msamwuo Hmofinmmumowo

ms» mo coflumfluomwouu.m wands

54

All mice were raised under laboratory conditions in the
colony maintained on the Michigan State University Campus.
The colony temperature was kept at approximately 21°C
under a 15:9 hour light-dark cycle. The mice were housed
in plastic cages measuring 28.6 cm. X 12.7 cm. X 15.2 cm.,
provided with wood shavings and a cotton Nestlet (Anicare).
Food (Wayne Mouse Breeder Blox) was provided ad libitum by
means of a wire mesh hopper in the top of each cage. Water
was also always available ad libitum from a suspended drop.

Except for g, l. castaneus all mice were the first

 

laboratory generation from wild caught parents. 33 l.

castaneus probably consisted of individuals of the 3rd—4th

 

generation of laboratory raised mice. All mice were weaned
at approximately 21 days of age and kept in groups of 3-4
of one sex until their use in the experiment. The average
ages in days (i 1 standard error) of the mice when placed
in the experiment were: 33 1, noveboracensis 146.3 i.1'8'

P. 1, castaneus 165.4 i.4°7 (the ages of 33 of these mice

 

were unknown and were arbitrarily assigned ages of 250
days), a, m, bairdi 141.6 i 1.9, a. m. blandus 157.6 : 4.1,

and g, eremicus eremicus 146.9 1 6.5. The average weights

 

in grams (i’l standard error‘ of the mice before testing

*were: 3, l. noveboracensi§_22.6 : 0.5, 3' l. castaneus

 

29.1 i 0.8, g. _r_n_. bairdi 17.8 i 0.3, g. m. blandus 27.0 i

0.5, and E. g. eremicus 26.0 i 0.8.

55

Apparatus

The hoarding apparatus was similar in design to that
described by Smith and Ross (1953a). The test cages were
the same as those in which the mice were raised and had
hardware cloth (l/4-in. mesh) alleys approximately 43 cm.
long and 7.62 cm. square attached to them. Access to the
alleys was provided by a hole (2.54 cm. in diameter) drilled
in the end of each cage. A hardware cloth food bin measuring
about 7.62 x 7.62 x 2.54 cm. was placed at the opposite end
of each alley from the test cage and filled with food. The
food pellets were of the same type (Wayne Mouse Breeder
Blox) as used for rearing and had been previously sorted to
a uniform size of about 3.67 Z.°-°7 grams each. All food
was removed from the normal food hoppers atop the cage

during experiments.

Procedure

The experimental situation was designed so that the
effects of four variables on food hoarding could be studied:
(1) ambient temperature, (2) acclimation to temperature,

(3) photoperiod, and (4) latitude or subspecies.

Four weeks prior to testing, a group of mice (normally
about 48) were randomly selected from the various available
stocks and divided into four equal experimental groups.

Each mouse was weighed and housed singly in a plastic cage
of the same type used for previous rearing. One group of

mice was then maintained at 7.0 : 0.8°C and under a 15:9

56

hour light-dark photoperiod (henceforth to be called a
"long" day photoperiod); a second group at 7.0 :.0.8°C and
under a 9:15 hour light-dark photoperiod (a "short" day
photoperiod); a third group at 26.7 : 1.0°C and under a
long day photoperiod; and a fourth group at 26.7 : 1.0°C
and under a short day photoperiod. Both of the "warm"
temperature groups (27°C) were kept in separate rooms with
independent light control systems and with the temperature
controlled by a single thermostat. The two "cold" temper-
ature groups (7°C) were kept in a refrigerated chamber with
the shelves for both groups again having their own inde-
pendent light control systems. The cages were kept on
steel and wood shelving, with approximately 12.7 cm. between
cages. Lighting was provided by two 60 watt light bulbs
approximately 43 cm. above each shelf of 6-8 cages. A
7 1/2 watt red bulb provided the only light in each room
during the dark phase of the photoperiods. Relative
humitity in the warm rooms was approximately 40-50% and in
the cold rooms 75-85%. Food and water were available ad
libitum.

In order to distinguish between the effects of
immediate ambient temperature at the time of testing
versus the effects of a long term acclimation to the warm
or cold temperatures, each of the above four experimental
groups were further subdivided on the day before testing so
that half of the animals would be tested at the opposite

temperature (but the same photoperiod) to which they had

57

been acclimated for the previous 4 weeks. This resulted in
a total of 8 experimental groups (4 temperature groups
under each of the 2 photoperiods). This design is outlined
in Table 3.

On the day prior to testing, the appropriate cages
were switched from one temperature control room to another
as outlined above and the hardware cloth alley was attached
to the end of each cage. All food was removed from the
normal food hoppers atop the cages. On the first day of
testing, the number of food pellets that had been carried
into the home cage during the previous night was recorded.
Good pellets that were not appreciably eaten were returned
to the food bins at the end of the alleys and if necessary
additional fresh pellets were added to keep the bins full.
Any pellets left in the alley but not in the bins were also
replaced (but not counted) and any nesting material in the
alley was removed. This procedure was followed for seven
consecutive days while the mice had continued access to the
alleys and the food. At the end of the test, all mice were
again weighed and then destroyed.

A total of 552 mice were tested using the above
procedure from January 1973 to March 1974. An attempt was
made to randomize the months during which individuals of
any one stock were tested. As equal numbers of mice of
suitable ages from the same stock were not always available,
the test periods necessarily contained disporportionate

numbers of mice from various stocks. The only exception to

58

 

.cowuomm mpocume mom .mDOAEmHm .M ummoxm mmsoum Ham MOM combo

 

 

 

 

 

 

m m w m m m m m m m m m m m m mHmEflsd
mo umnEsz

n

o

m

w

m

N
H m%MQ
m m momomomom xmm

m m m I. + 4 k w

UOFN Uob UoFN Uob UOFN Uofi UOFN Uoh wHDMMHGQEGB UmmH
Oahu Ooh Oohm Ooh musumummEme
cofiumEHHood
unwed ma uxumo m xuma m “usmflq ma coaummouonm

 

m.>osum mane now cofimmo HmucmEHquxm one:|.m magma

59

this procedure was the testing of the g. eremicus mice
which were all tested simultaneously from November to
December 1973 because of their unavailability earlier in
the study. The small number of avilable a. eremicus also
necessitated their testing under the same temperature to
which they had been acclimated in order to insure an

adequate sample size for the comparisons of interest.

Analysis

The dependent variable chosen for analysis within each
subspecies was the number of food pellets hoarded on each
of the seven days of testing. For comparisons among the
various subspecies an average daily hoarding score (i.e.,
the sum of all pellets hoarded by each mouse divided by 7
days) was used. In both cases the common logarithm of the
score plus 1 was taken in order to reduce the variance among
scores and to prevent negative scores (Morgan, 1945). In
summary, for all comparisons within a subspecies involving
days as a factor, the log (X + l) for each mouse was used;
for comparisons among subspecies the log (2 + l) was
analyzed. The latter procedure results in slightly higher
mean scores, but the relationship between them remains the
same.

The eXperimental design for the comparisons within

I'U

. l. noveboracensis, P. 1. castaneus, P. m. bairdi, and

 

 

E. m. blandus represented a 2x2x2x2x7 factorial arrangement

of treatments with one repeated measure for each taxonomic

60

group (2 sexes x 2 acclimation temperatures x 2 photo-
periods x 2 test temperatures x 7 days) (Winer, 1962). For
the comparisons among these same subspecies the design was
a 4x2x2x2x2 factorial (4 subspecies x 2 sexes x 2 accli-
mation temperatures x 2 photoperiods x 2 test temperatures).
The hoarding data for P. eremicus were analyzed in a 2x2
(2 test temperatures x 2 photoperiods). The actual analysis
was done on Michigan State University's CDC 6500 computer
using the MSU Stat System Version 3.04 Program for Analysis
of Variance. Interaction cell means were compared using t
tests. Simple correlation coefficients were also calculated
using the MSU Stat System's Bastat Program. Homogeneity of
variance was tested by using both the Bartlett's Test
(Sokal and Rohlf, 1969) and Cochran's Test (Winer, 1962).
The change in body weight of the mice during the
eXperiment was also analyzed. A score for each mouse was
obtained by subtracting the initial weight of each mouse
from their final weight after testing, and adding 20.00 to
this difference in order to prevent negative values. Scores
less than 20.00 therefore indicate a loss of weight during
the 35 days of the experiment, and scores greater than
20.00 indicate a gain in weight for the same period. The

experimental design for all the subspecies of 3' maniculatus

 

and g. leucopus again represented a 4x2x2x2x2 factorial
arrangement of treatments (4 subspecies x 2 sexes x 2
acclimation temperatures x 2 photoperiods x 2 test temper-

atures). For the analysis of weight change in B. eremicus,

61

the design was simply a 2x2 factorial (2 temperatures x 2

photoperiods) because of the small sample size.

RESULTS

The following description of the results has been
divided into three basic parts to coincide with my three
original hypotheses. First, the effects of the treatments
on each subspecies will be examined, followed by comparisons
among all the subSpecies. Special emphasis will be placed
on comparisons between the northern and southern sub-
species of both 33 leucopus and £3 maniculatus. Finally
the results of the analysis of the weight changes under the
various experimental conditions will be examined. The

complete analysis of variance tables are given in the

appendix.

ngparisons Within Each Subspecies
The analysis of variance of the daily hoarding scores

of E, l. noveboracensis indicated a highly significant main
effect of test temperature (F=6.39, df=l/112, p=0.01), with
mice hoarding more under 7°C than under 27°C (Table 4).
Neither photoperiod nor any of the other main effects
between subjects gave a significant F value. However, the
within subjects factor of days was highly significant

(F=12.76, df=6/672, p<0.001) and is graphed in Figure 1.

62

63

 

 

 

 

 

 

 

 

homo. mvbv. bNmm. vomm. mflmv. Uobm

boon. Gown. Obwm. ¢mh¢. momm. Ooh
.3
@mom. Nmbm. 0mmv. hmmm. NNHm. mGOA
wmmh. Hmmv. ammo. Omhv. monm. uuonm
wowummouosm

. . ombm. OmHm. Hmam. Nhom. Uohm

. . hmmv. mamm. wmmfi. mamm. Uoh
.3
. . mmmm. Ohvm. ommm. Ommm. mHmem
. . vmav. mNmm. mowv. hmom. mHmE
.mmw

mdofiEwuo .M maps—wan am .M «panama Am .M mamcmummo .m. .M .Qm>oc .H

 

cum memos 039 on» moumoaocHav
ummoxm .mHmEHc< vb mo ommuo>4 onu mucmmmummm coo: noun

x.ucmummmao saucmoflwflamflm

 

.omucmmmumom mud mHmEHsd om haze scans cw mooHEmuw

OM CH

.ousumuomEme umme can .pofiummouonm

.musumuomeme cofiumsfiaoos .xom mo muoommm can: can now mmuoom mcwoumom moa xafimo cmwznn.v magma

64

Figure 1. Mean log hoarding scores for the two subspecies
of both P. leuco us and £3 maniculatus on each
of the seven test days. Each point represents

the average of 128 animals over all experimental
conditions.

65

{on .8...

H musmwm

 

 

«50:03 a
:53 a
q
._.

 

 

.m
.m
3323“. .m
.n...

flap—$095502:-

I
I
I

I .m

 

pepiooH s:a||od.(|+)'.oN 00"] upon

66

The scores for the first two days were typically the
highest and then leveled off at a lower level for the
remaining days of testing. None of the interactions either
between or within subjects was significant, although the
interaction of test temperature x days was graphed

(Figure 2) for comparison with P. l. castaneus.

 

The analysis of the mean daily hoarding scores for
£3 1, gastaneus revealed a significant main effect of sex
(F=4.03, df=l/112, p=0.05), with females hoarding more than
males (Table 4). Both the main effects of temperature and
photoperiod were nonsignificant, although the within sub-
ject factor of days was again significant (F=23.l9, df=
6/672, p<0.001). As already pointed out for £3 1, 22327
boracensis, the mean daily hoarding scores for £3.l'

castaneus were highest in the first two days of testing

 

with a leveling off in the remaining five days (Figure 1).
There was also a highly significant interaction between
days and test temperature (F=4.52, df-6/672, p<0.001;
Figure 3). On day l of testing, mice under 7°C, hoarded
significantly more than those under 27°C (Table 5). How-
ever, on all subsequent days, more hoarding was done under
the warmer temperature with the difference becoming signi-
ficant on the last three days of testing.

The interaction of sex x acclimation temperature in

E, 1, castaneus was also significant (F=6.31, df=l/112,

 

=0.0l; Figure 4a). After acclimation to 7°C, females

hoarded significantly more than males (t=3.20, df=62,

67

Figure 2. Mean daily log hoarding scores for P. l.
noveboracensis under the two test temperatures.

Each point represents the average of 64
animals.

68

 

 

 

2.5 :3... N .33..
b m 0 ¢ n N _
j 1 HO
Hm.
Icicle, .. v.
0‘ (O
o\ J m.
u/o\o/v/Av mm.
1 h.
01.... I
Ooh I

2223th So».
3239.830: 2.323. m

 

 

 

popmoH smled (|+)°oN 601 uoaw

69

 

Figure 3. Mean daily log hoarding scores for §.'l.
castaneus under the two test temperatures. Each
pOint represents the average of 64 animals.

70

£25 So...

m musmwm

 

 

“vohmwdflld.
Huck. dflld
228380... .3...

 

2.8230 2533. .M

 

H. W
.1.
in...”
u
.m.
1V6
m
[nor
.7
1m.“
1 m
33
H
nv
m
p
a
p

 

71

Table 5.--Student's t Tests of the Difference Between the
Mean Daily Log Hoarding Scores Under 7°C and the
Mean Scores Under 27°C for Both Subspecies of a.
leuco us and E, maniculatus. Negative Values
Indicate More Hoarding Was Done Under 27°C Than
Under 7°C. The Level of Significance is Given in
Parenthesis Beside Each t Value; df=126.

 

 

 

 

 

 

 

 

 

 

 

 

3' _l_. noveboracensis E.“ l. castaneus
Days 2 Difference t Value 2 Difference t Value
1 .2874 5.38(.001) .1526 3.42(.01)
2 .1401 2.62(.02) “.0576 1.29(N.S.)
3 .1879 3.51(.001) ‘.0529 1.19(N.S.)
4 .1357 2.53(.02) ’.0625 1.40(N.S.)
5 .1700 3.18(.01) ‘.1107 2.49(.02)
6 .0962 1.80(.10) -.1358 3.05(.01)
7 .1291 2.41(.02) -.0901 2.02(.05)
.13: m. bairdi a. m. blandus
Days 2 Difference t Value 2 Difference t Value
1 .1234 2.38(.05) .0433 .86(N.S.)
2 -.0247 .48(N.S.) -.1535 3.04(.01)
3 .0444 .86(N.S.) ’.1972 3.91(.001)
4 .0229 .44(N.S.) “.2086 4.13(.001)
5 ”.0951 1.83(.10) “.1602 3.17(.01)
6 -.0476 .92(N.S.) -.1674 3.31(.01)

7 -.0629 1.21(N.S.) -.O963 1.9l(.10)

 

Figure 4.

72

Mean daily log hoarding scores of P. l.
castaneus for the significant intefacfions of
(a) sex x acclimation temperature, and (b)
photoperiod x acclimation temperature. Each
point represents the average of 32 animals.

73

(a)
(b)

Femalas

s

W

10 n.
O

.M L

7°C
Acclimation Temperature

Figure 4

A.
o
a.

 

 

fireflih”. emerging».
62:31 32.2. 2.... vogue: 32.3. 2.:
.oz 03 zoos. .oz 3.. .322

27°C

74

p=0.01), although after 27°C there was no significant
difference between their scores. The interaction of
acclimation temperature x photoperiod was likewise highly
significant (F=1l.52, df=1/112, p=0.001; Figure 4b). Under
a short photoperiod, g, l. castaneus hoarded significantly
more after acclimation to 27°C than after 7°C (t=2.72,
df=62, p=0.01). However, under a long photoperiod just the
opposite was true (t=2.08, df=62, p=0.05). Further exami-

nation of this interaction will follow when E, l. castaneus

 

and £3.1- noveboracensis are compared with each other.

 

The analysis of variance of the hoarding scores of
E, m, bairdi showed that photoperiod had a significant main
effect (F=4.01, df=l/112, p=0.05), with mice hoarding more
under a short day photoperiod than under a long day one
(Table 4). There was, however, no significant difference
in the rate of their hoarding (i.e., the log number of
pellets hoarded plus 1 divided by the number of hours of
darkness available to hoard them) in contrast to the amount
under either photoperiod (i rate=.0411 per hr. under
short, versus i=.0537 per hr. under long). The main effect
of test temperature was non-significant. Unlike both a. l.

noveboracensis and 3°.1° castaneus, P, m, bairdi did not

 

 

differ in hoarding across days. The interaction of test
temperature x days was, however, on the borderline of
significance (F=2.06, df=6/672, p=0.056) and is graphed in
Figure 5. There was a significant difference between the

scores for mice tested under 7°C and those under 27°C only

75

 

Figure 5. Mean daily log hoarding scores for E. m, bairdi
under the two test temperatures. Each pOint
represents the average of 64 animals.

 

76

2:5 :6...

h m n ¢ m N

 

ollllo.‘ lo
2.11. “Kw
1’ (o

Cox-NI
nuoh_.Yllo

9.29.380... 3:.

:52. 2:23.56 mu

 

m wusmflm

 

pepmoH sienedu-I-rou 601 uoew

77

on the first day of testing (Table 5). None of the other
interactions for a. m, bairdi was significant.

The results of the analysis of variance for a, g,
blandus disclosed that test temperature had a significant
main effect (F=6.77, df=l/112, p=0.01), with mice tested
under 27°C hoarding more than those under 7°C (recall that
this is the exact opposite of what was found for £3.1'
noveboracensis). Like 23 m, bairdi, there was no signifi-
cant effect across days, although the interaction of test
temperature x days was highly significant (F=2.92, df=
6/672, p=0.008: Figure 6). On every test day but the
first, a, m, blandus hoarded more under the warm 27°C
temperature than under the cold 7°C temperature (Table 5).
This is very similar to what was found for P, l, castaneus.
No attempt was made to interpret the significant four way
interaction of sex x acclimation temperature x photoperiod
x days (F=2.21, df=6/672, p=0.05).

The analysis of variance of the mean hoarding scores

for g. eremicus eremicus indicated that both main effects

 

of temperature and photoperiod were non-significant, as
was their interaction. The within subjects factor of days
was not analyzed because of the small sample size and

simpler experimental design (see "Methods" section).

Comparisons Among Subspecies

 

The analysis of variance for the combined hoarding

scores for all subspecies except 3. eremicus indicated a

78

 

Figure 6. Mean daily log hoarding scores for g. m. blandus
under the two test temperatures. Each pOint
represents the average of 64 animals.

79

goo to...
s o 0 ¢ n .o.

7 d

1 q . «I!
\/q\ 14/6

 

.Crlllllllll‘Pllr

llL‘IIIIIIIIIL‘fIIIIIIIIIaPIII A.

093 I
0.... old

9.29.350... BS.

 

2.2.2.. «223.5... m:

m mucosa

.
F0

 

nun

 

an.

pepmoH ssened(|+)°oN 60-1 uoew

80

highly significant main effect of subspecies (F=4.4l, df=
3/448, p=0.005). The highest mean hoarding score was for

£°.$° noveboracensis, with E. m, bairdi, £.'1. castaneus,

 

and P. m. blandus following in that order (Figure 7).

 

Comparisons of these means using the Student-Newman-Keul's
Test and a lsd test for the mean of g, eremicus (because of
its much smaller sample size) indicated that eremicus
hoarded significantly more food than both a. 1' castaneus
and g, m, blandus (Table 6). The mean scores for E, l.
noveboracensis, 33 m, bairdi, and g,‘£. castaneus were all
significantly greater than the mean for E, m, blandus, but

were not significantly different from each other.

Table 6.--Multiple Comparisons Among the Mean Log Hoarding
Scores (i.e., Log (X+l) for Each Mouse)+1S.E. for
All Subspecies of Peromyscus Used in This Study.
Any Two Means Not Joihed’by the Same Vertical Line
Are Significantly Different at the .05 Level.

 

 

 

 

 

 

 

Subspecies Mean :1 S.E.
g. eremicus .7123 i .0586
a. 1. noveboracensis .6435 i .0360
E, m, bairdi .6331 i .0359
3°.l' casataneus .5794 i .0299
3° m, blandus .4964 i .0290

 

The interaction of subspecies x test temperature was
also highly significant (F=4.85, df=3/448, p=0.002;
Figure 8). Consistent with my original hypothesis, E. 1.

noveboracensis hoarded significantly more than 3. l.

 

81

 

Figure 7. Comparisons of the mean log hoarding scores per
test day for all five stocks of Peromyscus
studied. Each bar represents the average for
128 animals, except for E. eremicus in which 40
animals are represented. The vertical lines are
:1 standard error.

 

82

 

 

Wu

 

 

 

 

{I

b

3 7. 6. 5 A 3 2
23m 9.6.3...

es per

'for
ich W

nes a“

WIS. 238.3

m 3 63.3...

 

E Banach

R as sausages?»

V. .3858

Figure 7

 

83

Figure 8. Graphs of the mean log hoarding scores per test
day for the interaction of subspecies x test
temperature. Each point represents the average
of 64 animals.

84

m madman

 

 

 

 

   

 

 

 

 

 

8282.60... 8288...:
.5 :5
Oahu Ooh Oahu Ooh
- 1 AU . q AV
Lr_. LI_.
4? 4f H
61m MW
9.6% 1 0. 4 m. m-
m.
b
r 0. . 2.2.280 .. 0. W3
$8.841. m
. s. . s.
«223.5... d. L m. «2.33. m... 5.

 

r W

est
V9139?

85

castaneus under a test temperature of 7°C (t=2.79, df=
126, p=0.01). As might be predicted, there was no differ-
ence between the subspecies under the warm, 27°C, test
temperature. As already pointed out in the preceding
section, 33 l. noveboracensis hoarded significantly more
under the cold test temperature than it did under the warm
temperature. In contrast, test temperature had no signifi-
cant effect on the hoarding scores of g. 1. castaneus.
Comparisons of the hoarding scores for both subspecies
of of g, maniculatus under the two test temperatures

(Figure 8) revealed that under 7°C, g, m, bairdi also

hoarded significantly more than its southern subspecies,
£3 E, blandus (t=3.36, df=126, p=0.01). The difference
between their scores under 27°C was not significant. As
previously pointed out, test temperature had no effect
within 3, m. bairdi, although it did within 23 m, blandus.
The interactions of subspecies x acclimation temperature
x photoperiod (F=3.44, df=3/448, p=0.02) and subspecies x
sex x acclimation temperature x photoperiod (F=2.76, df=
3/448, p=0.05) were both significant. Only the three way
interaction was examined because no hypotheses had been
formulated to explain possible interactions with sex, and
secondly, the level of significance was greater in the
three way interaction than in the four way interaction. The
cell means for this interaction are graphed in Figures 9
and 10. As illustrated in Figure 9a, under a short day

photoperiod and after acclimation to cold temperature,

Figure 9.

86

Graphs of the mean log hoarding scores per test
day for the significant interaction of sub-
species x photoperiod x acclimation temperature
in g. leuco us under (a) a short photoperiod,
and (b) a long photoperiod. Each point
represents the average of 32 animals.

87

6.26.366... 5:08:84
ooh ooh

 

   

 

.3338.»

.\

Masseusosesok IV

 

 

62.3295 93.. .625 .n a

test

 

2.303. M m 9.33..
0.29.2.6»... 3:05:36.

 

 

 

 

 

0.5 o... o
1..“
sly
co .0
pm
coo
.o.
3.36.06 .6 96.6 Q10 .. N.
62.3225 :25 30.5.... .m.

tare
id,

91003 bugplooH

Figure 10.

88

 

Graphs of the mean log hoarding scores per
test day for the significant interaction of
subspecies x photoperiod x acclimation temper-
ature in E- maniculatus under (a) a short
photoperiod, and (b) a long photoperiod. Each
point represents the average of 32 animals.

 

89

 

33.3...qu

8223.5... co..oE..oo<

Oahu Ooh

’1

I
53.6%

63.3225 93.. .895...

I.

o

.H
e

j '

 

OH wuomHm

9.22356... 3:68:02

 

Oahu Ooh

 

3:332... :25 .6256

_j

 

 

91003 5U!pJDOH

90

g.‘l. noveboracensis hoarded significantly more than a, l,
castaneus (t=2.93, df=62, p=0.01). After acclimation to
the warm temperature (27°C) no difference was found between
their scores. 3, l. gastaneug also hoarded more after
acclimation to 27°C than after acclimation to 7°C (t=2.65,

df=62, p=0.02), while the hoarding of g, 1, noveboracensis

 

was apparently unaffected by prior temperature acclimation
under this photoperiod. Under a long day photoperiod
(Figure 9b) previous temperature acclimation had more of an
effect on the hoarding of both subspecies. Although both
subspecies tended to hoard more after acclimation to 7°C,

the effect was only significant in g. 1, noveboracensis

 

(t=2.07, df=62, p=0.05).

Figure 10a illustrates the means for both subspecies of
£3 maniculatus under a short day photoperiod. As in 33
leucoEus, the northern subspecies, g, m, bairdi, hoarded
significantly more than the southern subspecies, P, m,
blandus, after acclimation to 7°C (T=2.22, df=62, p=0.05).
However, the difference between their scores after accli-
mation to 27°C was not significant at the .05 level (t=
1.91, df=62, p=0.10). Under a long photoperiod (Figure
10b), there was no significant difference between any of
the means for both a. m. bairdi and E! m, blandus.

My original hypotheses (see "Introduction") suggested
several a priori comparisons that could be made inde-
pendently of the AOV design and regardless of the signifi—

cance level of the F values obtained (Steel & Torrie,

91

1960). The means of mice tested under 7°C and a short
photoperiod were compared individually by subspecies with
the means for mice tested under 27°C and a long photoperiod
(Tables 7 and 8). Consistent with my original hypotheses,
both 23 l, noveboracensis and B! m, bairdi hoarded signifi-
cantly more under a short day photoperiod and cold temper-
ature than under a long day photoperiod and a warm temper-
ature. The means within E. l. castaneus, g, m, blandus,
and g, eremicus under these two test conditions were not
significantly different.

The simple correlation coefficients between the mean
hoarding scores for all subspecies and their age, initial
weight, weight change, and final weight are given in
Table 9. There were no significant correlations with age
since this variable was fairly well controlled for in this
study. There were also no significant correlations with
either initial or final weight in any of the subspecies.
There was, however, a significant correlation between the

mean hoarding score of 2: m. bairdi and their weight change

 

(r=-0.21, p=0.05). The correlation indicated that there
was a tendency for hoarding scores to increase slightly with

loss in weight in this subSpecies.

Analysis of Weight Changes
The combined analysis of variance of the weight changes
of all the subspecies except 3' eremicus revealed that the

main effects of both acclimation temperature (F=6.40,

92

Table 7.--Mean Log Hoarding Scores (i.e., Log (2+1) for
Each Mouse) for the Five Stocks of Perom scus
Under All Possible Combinations of Test *emper—
ature and Photoperiod. Each Mean Represents the
Average of 32 Animals, Except in P. eremicus in
Which 10 Animals are Represented.-'Means in
Parenthesis are for Mice Acclimated and Tested at
the Same Temperature.

 

Test Temperature

 

 

 

 

Photoperiod
7°C 27°C

§.'l. noveboracensis

Short .7705(.754l) .6064(.5450)

Long .7320(.7955) .4651(.3407)
P_._ l. castaneus

Short .5239(.3922) .5895(.6802)

Long .6207(.6615) .5835(.4657)
g. m. bairdi

Short .7004(.7179) .7ll4(.5943)

Long .6004(.5837) .5203(.5049)
P. m, blandus

Short .4078(.4344) .6316(.5502)

Long .4621(.4675) .4842(.4697)
3° eremicus*

Short .7596 .6872

Long .8346 .5678

 

*For g. eremicus, all mice were tested at the same
temperature to which they had been acclimated.

93

Table 8. --Planned Comparisons_ of the Mean Log Hoarding
Scores (i. e., Log (X+1) for Each Mouse) of Mice
Tested Under a Short Photoperiod and 7°C With
Those of Mice Tested Under a Long Photoperiod

 

 

 

 

and 27°C.
Subspecies X(7°+short) X(27°flong) t Value(prob.)a

a. 1.

novaoracensis .7705 .4651 3.60(.001)
§.'l.

castaneus .5239 .5835 .66(N.S.)
P. m.

Bairdi .7004 .5203 1.99(.05)
P. m,

Biandus .4078 .4842 1.14(N.S.)
g. eremicus .7596 .5678 1.15(N.S.)

 

adf=62 for all

df=18.

subspecies except a.

eremicus in which

94

 

 

 

 

 

 

 

smH.u mac. oHH.n Hmo. mmo. usages Hagan
mmo. meo. .mom.u 000.: moo.u mmcmno unmflmz
Hm.au mmo. moo. omo. «so. unmflmz HmfluHcH
mm.Hu mmo.- omo. «No.. 050.- was
mSOHEmHo .M msccmHQ .m .M HUuHmn .mm .M mumsmummo .fl .M mHmcwomuonm>oc .IH. .M

 

..Hm>mq mo. man um

oUGMOHMHcmHm moumoH©:H«. uanoz HmsHm use .mmcuso uanmz .uanmz HmHuHcH .mmm uHmze can
msommeomumm mo mxooum o>Hm HHa mo mouoom mchumom may com3umm mucmHonmmou coHuMHmuuoouu.m meme

95

df=l/448, p=0.01) and subspecies (F=22.68, df=3/448, p=
0.001) were significant as was their interaction (F=5.14,
df=3/448, p=0.002; Figure 11). Although both 33 l.
noveboracensis and a. l. castaneus lost weight under the
two acclimation temperatures, they lost significantly more
weight under 7°C than under 27°C (Table 10). There was
also a significant correlation in both subspecies between
their initial weight and weight change (r=-0.62, p=0.01,

for 33.1“ noveboracensis; r=-0.58, p=0.01, for E! l.

 

gastaneus) indicating that the greater the initial weight

 

the larger the loss. The negative sign should be ignored
as it is a result of the method in which the weight changes
were scored (see "Methods" section). The change in weight
of g, eremicus was also significantly different under the
two acclimation temperatures (F=7.98, df=l/36, p=0.008)
with individuals losing weight if acclimated to 7°C, but
gaining weight under 27°C. The correlation between initial
weight and weight change was also significant (r=-0.Sl,
p=0.01). The difference in weight change for both a. m.
bairdi and g. m, blandus were non-significant (Table 10),
but again the correlation between initial weight and weight
change was significant (r=-0.56, p=0.01 for a. m. bairdi;
=-0.6l, p=0.01 for E! m. blandus).
The interaction of subspecies x sex x acclimation

temperature was also significant (F=2.74, df=3/448, p=
0.05) and is graphed in Figure 12. Comparisons of the

means using t tests indicated that it was the females of

96

Figure 11. Graphs of the mean weight changes for all five
stocks of Peromyscus after acclimation to
either 7°C or 27°C.

 

97

as enemas

0.5 U
o... s
2223.5... 3:66:34

 

\

 

 

r rirm

 

 

«SQSEQ .6. 32633 \Ilu 6.4332326: Ell 3333 E “a. Essa E m

l

 

mw+

m7?

(swmb) ebuoqg :qbgeM

98

Figure 12. Graphs of the mean weight changes for the
significant interaction of subspecies x sex x
acclimation temperature.

99

 

 

\S

 

 

 

 

 

NH enemas

OobN U

Ooh

V\.

82235: 3.8.5.3...

 

33836 \IQI

6.4286332: \ m

 

mabtuxfill

J

l

 

s94

 

.33

Q

VII

”In

an

N...

m...

 

.v...

(swmb) ebunqo (qbgeM

100

Table lO.--Student's t Test Comparisons of the Mean Weight
Changes (In Grams) After Acclimation to 7°C and
27°C for the Significant Interaction of Sub-
species x Acclimation Temperature.

 

 

 

 

 

 

Subspecies 2(7°C) 2(27°C) t Value(prob.)
a. l. noveboracensis -2.16 -0.92 2.09(.05)
3. 1° castaneus -3.26 -0.92 3.94(.001)
g, m, bairdi +1.20~ +0.41 1.35(N.S.)
g, m, blandus +0.24 +0.46 .38(N.S.)
g. eremicus -l.72 +0.79 2.83(.01)

 

 

both subspecies of g. leucopus which accounted for the
significantly greater loss in weight under the colder of
the two acclimation temperatures (t=2.13, df=62, p=0.05
for female 3, l, noveboracensis; t=3.88, df=62, p=0.001
for female E: l. castaneus). The losses in weight observed
in males of both of these subspecies were not significantly
different under either of the two acclimation temperatures.
Consistent with what was mentioned above, acclimation
temperature also had no significant effect within either
sex of E! m, bairdi or P, m, blandus, although female
33 m, blandus gained weight under 7°C, while males lost
weight under the same temperature (t=2.48, df=62, p=0.01).
The interaction of subspecies x sex x photoperiod was
significant (F=3.53, df=3/448, p=0.02) and the means are
graphed in Figure 13. Photoperiod had no significant

effect on weight change in either subspecies of E. leucopus

101

Figure 13. Graphs of the mean weight changes for the
significant interaction of subspecies x sex x
photoperiod.

 

 

 

 

 

 

I —

 

\

 

 

 

W1,

 

 

 

 

_l

 

a Long(|5L290)
D ShorHSLZ (5 D)

Photoperiod

 

+2-

I n
O -’ N
I I

(swam) ebuoqa :qbgeM

-3 -
Fi

103

but did interact with sex in both E: m, bairdi and £3 a,
blandus. Under a short photoperiod female 3, m, bairdi
gained weight, whereas males showed a small average loss

in weight (t=2.35, df=62, p=0.05). Under a long photo-
period there was no significant difference in the amount of
weight change in either sex of g, g, bairdi. On the other
hand, female g, m, blandus gained significantly more weight
under the long photoperiod than males, which reacted just
the opposite and lost weight (t=2.72, df=62, p=0.01).
Photoperiod had no significant effect on weight change in
a. eremicus.

The interaction of subspecies x sex x test temperature
was also highly significant (F=5.29, df=3/448, p=0.001) and
is graphed in Figure 14. One would normally expect that
changes in body weight would be more influenced by the 4
week acclimation period to each temperature than by the
final seven days of testing. This was in general true for
all groups except 33 m, blandus, which did show some signi-
ficant adjustment in their weight during the seven days of
the testing. 33 m, blandus males which had lost weight
during acclimation to 7°C, tended to gain it back if
tested at 27°C. Likewise, females which had gained weight
under 7°C, tended to lose it if tested at 27°C. The same
sort of changes went on in the other groups but not to the
same extent. Apparently more than seven days is needed
before an adjustment is complete in these groups and there-

fore Figure 14 may be misleading. The weight changes of

104

Figure 14. Graphs of the mean weight changes for the
significant interaction of subspecies x sex x
test temperature.

105

N

58 §_

 

 

 

 

 

&

«H ousmflm

 

 

 

3328.» a haemooxoomsoe \ .1

 

b

_

 

. mTI
Oohw D
Ooh. an
2333.5... «no... 4 NI
. .I
o
s . . _,,
w . N...

w

«attoxs E Q

 

 

 

>3on St a

m+

(swam) ebuoqo mbgaM

106

all mice under the various combinations of acclimation
temperature and test temperature are shown in Table 11,
which serves to illustrate this point.

The significant interactions of sex x test temperature
(F=4.34, df=l/448, p=0.05), sex x acclimation temperature x
photoperiod (F=7.ll, df=1/448, p=0.008), and sex x accli-
mation temperature x photoperiod x test temperature (F=4.38,
df=l/448, p=0.05) were not examined for several reasons.
First, I was not interested in "average" changes over all
the different subspecies, but only those significantly
different from one subspecies to the next. Second, I had
no hypothesis concerning weight change under various
environmental conditions. The relationship of interest,
namely whether hoarding and weight change are related in

some fashion, will be discussed in the next section.

107

Table ll.--Weight Changes (In Grams) of Mice Under All
Combinations of Acclimation and Test Temperature.

 

 

 

 

Acclimation Test Temperature _
Temperature X
' 7°C 27°C
a. _1_. noveboracensis
7°C -2.28 -2.04 -2.16
27°C -0.99 -0.86 —0.92
2 -1.64 -1.45
11.1. castaneus
7°C -2.98 -3.54 -3.26
27°C -l.85 0.00 -0.92
i -2.42 -1.77
a. 1_n_. bairdi
7°C +1.53 +0.88 +1.20
27°C +0.43 +0.38 +0.41
i +0.98 +0.63
3' m. blandus
7°C +0.12 +0.35 +0.24
27°C +0.51 +0.41 +0.46
2 +0.31 +0.38
g. eremicus
7°C -l.72 . . —1.72
27°C . . +0.79 +0.79
i —1.72 +0.79

 

DISCUSSION

The thesis of this study has been that deermice are
capable of utilizing multiple environmental cues to
"anticipate” seasonal changes in their food supplies by
beginning intensive hoarding. Five taxonomic groups
representing three species of geromyscus were studied under
laboratory conditions of photoperiod and temperature
simulating fall and summer to mice from northern latitudes.
My three working hypotheses and their fates upon testing
were: (1) both photoperiod and temperature affect the
amount of food stored by deermice. This hypothesis was
confirmed in regards to test temperature, but the evidence
for photoperiod was only conclusive in one subspecies of
g, maniculatus. (2) The effect of photoperiod and test
temperature will vary depending upon the latitudinal and
climatic origins of the subspecies being studied. This
hypothesis was confirmed since the northern subSpecies of
both E, maniculatus and 3° leucopus hoarded significantly
more under experimental conditions simulating fall or
early winter than under summer conditions. The two

southern subspecies did not show this difference, but did

108

109

hoard more under a warm test temperature than under a cold
one, except during the first day of testing. (3) The
hoarding of food by deermice is not related to changes in
their body weight. This hypothesis was confirmed in all
groups except 2, g, bairdi, in which the relationship may

have been spurious.

Test Temperature

The effect of ambient temperature upon the hoarding
behavior of two subspecies of EB leucopus was consistent
with the original hypothesis that their response would be
predictable on the basis of their latitudinal or climatic
origins. £3 1, noveboracensis, a northern subspecies found
predominantly in forest-type habitats, hoarded significantly
more under the cold temperature (7°C) than under the warm
temperature (27°C). The southern subspecies, g,‘l.
castaneus, which inhabits an entirely different habitat of
cultivated cornfields in tropical Mexico, did not show a
significant difference in the average amount hoarded under
either temperature. However, when the hoarding of £3 1.

castaneus was analyzed on a day by day basis it was found

 

that the mice on day l hoarded significantly more under
7°C than under 27°C. On all subsequent days, their
hoarding under the cold dropped so that by the fifth day
it was below the amount hoarded in the warm temperature.
The interaction of acclimation temperature and test

temperature was non—significant in both subspecies and

110

indicated that these two factors acted independently of
each other.
The effect of temperature on the two subspecies of £3

maniculatus contrasted with what was found in E: leucopus.

 

g, g, bairdi, the northern subspecies which occurs
predominantly in cultivated fields and grassland habitats,
did not show the same consistent response to test temper-
ature that £3.l' noveboracensis did. Only on the first
day of testing did g, m, bairdi hoard significantly more
under the cold temperature than under the warm. On the
remaining six days, hoarding was approximately the same
under both temperatures with a slight tendency to hoard
more under the warmer temperature, especially in the last
three days. 23 m, blandus, the southern subspecies which
occupies the desert-shrub habitat of the southwest, hoarded
even more consistently in the warm temperature. On all
days but the first, 3, m, blandus hoarded significantly
more under the warm temperature. When the two subspecies
were compared under both test temperatures, it was found
that E, m, bairdi hoarded significantly more than E: m,
blandus under 7°C as would be predicted on the basis of my
hypothesis. Under 27°C there was no significant differ-
ence between their scores. The absence of any significant
interaction of test temperature and acclimation temperature
again illustrated the independence of these two factors.
Several possible explanations might be proposed to

interpret the effects of test temperature on hoarding in

111

deermice: (l) hoarding is directly related to food con-
sumption under all ambient temperatures, (2) hoarding scores
reflect changes in general activity under various ambient
temperatures, (3) ambient temperatures act as cues to the
mice to begin intensive hoarding at certain times of the
year. The first proposal that hoarding is related to food
consumption under all ambient temperatures is not supportable
in light of my findings. The metabolic rate of rodents,
and hence their food consumption, varies inversely with
ambient temperature except while in torpor or hibernation
(see Hart, 1971). If changes in metabolic rate were
responsible for hoarding, one would expect then a similar
inverse relationship between food storage and temperature.
However, my results indicate that in deermice the effect of
test temperature varies with both subspecies and species,
even though its effect on their metabolic rates is in
general the same (Hart, 1953; McNab & Morrison, 1963). The
finding that a cold temperature increases hoarding in a. l.
noveboracensis but has the opposite effect on g, m, blandus
would tend to contradict any explanation of hoarding based
solely on metabolic rates. This agrees with the earlier
findings by Stellar (1951) that metabolic rate, at least
when manipulated through experimental interference of
thyroid functioning, has little effect on hoarding.

The second proposal that hoarding scores reflect
changes in activity with ambient temperature cannot be

entirely refuted by my study and may account for some of

112

the observed subspecific differences. It is possible that
in g, l. castaneus, the cold test temperature was suffi-
ciently stressful so as to inhibit their hoarding under the
cold temperature on all days except the first, when they
hoarded more under 7°C than 27°C. During the early stages
of this study, several 33 l, castaneu§_(approximate1y 7 out
of 135) died when placed under the cold temperature,
indicating that they may lack the ability to adjust as
quickly to cold as does a, l. noveboracensis, in which no
deaths were observed. The deaths of g, l. castaneus always
took place 24-48 hours after being placed at 7°C and
apparently were due to their inability to construct a nest
in a sufficiently short period of time to provide adequate
insulation from the cold. When g,‘l. castaneus individuals
were given sufficient time to construct a nest in the test
cage before being placed at 7°C, no deaths were observed.
The latter procedure was incorporated into the experiment
as soon as this fact became known.

An explanation similar to the one suggested above for

E,‘l. castaneus could possibly be used to interpret my

 

findings for B! m. blandus. The overall level of hoarding
in a, m, blandus was low, and perhaps their hoarding under
the cold temperature was inhibited sufficiently to cause a
significant difference in their scores under the two test
temperatures. Winter nights in the southwestern deserts
from which the original stocks of a. m, blandus were

obtained are cool, with minimum temperatures falling below

113

freezing much of the time from mid-November through early
March (Lordsburg, New Mexico Weather Station, 0.5. Depart-
ment of Commerce). My results sugest that blandus might
be less active during these periods.

The failure of E, m, bairdi to reSpond to either test
temperature except on the first day, probably indicates
temperature was not significantly affecting their activity.
Stinson (1952; cited by Falls, 1968) has reported that
laboratory reared g, m, bairdi tended to be most active
between 15°C and 27°C and this may account for the slightly
greater hoarding response under 27°C test temperature in
the last three days of testing.

If inhibition of activity by cold is used to interpret
the results in g, l, gastaneus and perhaps a, m, blandus,
it cannot be used to explain the effect of cold temperature.
on the hoarding of g, l. noveboracensis. In fact the
opposite may be true in this subspecies. Orr (1959) has
reported that E: l. noveboracensi§_kept in outdoor
enclosures showed the greatest amount of activity in
crossing treadles between approximately 4° and 10°C, which
encompass the cold test temperature of my own study. It
is possible that the increases in food storage by this
particular subspecies under 7°C were a consequence of their
increased activity at that temperature. This finding
agrees with what has already been reported on the hoarding
behavior of laboratory rats (McCleary & Morgan, 1946), but

remains contrary to observations on laboratory mice

114

(Ross & Smith, 1953). The response of a, m, blandus to
test temperature is in more agreement with the latter
study.

By far the most interesting explanation of the effect of
test temperature on hoarding is that it acts as a cue to
the mice to begin intensive hoarding during certain seasons
of the year. The substantial difference between the two
subspecies of g, leucopus under 7°C and the significant
effect of cold temperature on the hoarding of g, 1.
noveboracensis strongly suggests that cold ambient temper-
atures are at least in part responsible for initiating the
intensive food storage reported for this species during the
fall and early winter in the northern latitudes. Mean
daily temperatures in the area of Michigan from which the

stocks of E, 1, goveboracensis were derived, do undergo

 

seasonal changes from a high of about 27°C in July to a low
of -9°C in January (East Lansing, Michigan Weather Station,
U.S. Department of Commerce). Mean daily temperatures for
the fall months of September, October, and November range
between 3° and 16°C. Consistently low temperatures could,
therefore, serve as a reliable cue to begin intensive food
hoarding.

The failure of E, m, bairdi to be similarly affected by

cold temperature is surprising, considering that both E, m,

 

bairdi and g. 1. noveboracensis are sympatric in Michigan.
Whether this difference has any relationship to their

different habitats remains to be resolved. The greater

115

influence of photoperiod on g, m, bairdi than on B°.$°
noveboracensis will be discussed in the next section.

The response of g, m, blandus to the warm test temper-
ature might also be explained on the basis that a hot
temperature may be a cue to begin intensive hoarding in
its habitat. The continental climate of the southwestern
deserts is characterized by large annual and diurnal ranges
in temperature, but with distinct seasonal temperature
changes. Summers are quite warm with mean daily maximum
temperatures ranging above 32°C from mid-May until
September (Lordsburg, New Mexico Weather Station, U.S.
Department of Commerce). Half the yearly moisture normally
falls in the summer months of July and September, providing
good stands of grasses and other annual plants. It is
probably during this time of year that g, m, blandus does
the most hoarding, although there is no field evidence to
either support or refute this conclusion. It is reasonable
to assume that the long photoperiod and warm temperature
under which the if m, blandus were tested may have simulated
summer conditions to them and the need to hoard in prepara-
tion for the relatively mild but dry winter season.

Another possibility that will be discussed more thoroughly
later is that both a. eremicus and E! m. blandus may hoard
during certain seasons so as to have food for use during
prolonged periods of torpor induced by extremes in daily

temperature.

116

Photoperiod

Day length had by far the greatest effect on g, m,
bairdiz which hoarded significantly more under the short
day photoperiod than under the long one. All other groups,
except g, l. castaneus, showed small but non-significant
increases in hoarding under the short day photoperiod.
These findings do suggest that at least one northern sub-
species, 3, m, bairdi, may be using the length of the
photoperiod as a cue to increase hoarding. However, an
alternative explanation might be that the longer nights
could have afforded a longer period of time in which to
hoard, since there was no significant difference in the
rate of hoarding (in contrast to the amount) by g, m,
bairdi under the two photoperiods. This is contrary to
Muul's (1968) finding that flying squirrels do increase
their rate of hoarding under a shortened photoperiod.

If E: m. bairdi hoard whenever food and time are
available, their accumulation of food during autumn and
early winter may be a fortuitous consequence of their

remaining active under longer periods of darkness. Both

3: m, bairdi and g, l. noveboracensis do show longer

 

periods of nocturnal activity in the winter than in the
summer (Johnson, 1926; Orr, 1959). During the winter, the
nocturnal activity pattern of 3° m. bairdi is characterized
by two major peaks in activity both early and late in the
night (Falls, 1968). In contrast, during the summer (when

nights are shorter) the activity of E. m. bairdi is more

117

continuous, with only a single major peak early in the
night. If their fall and early winter hoarding activity
follows a similar pattern, the two peaks in activity would
result in larger caches of food at that time of the year.
In fact, Falls (1968) has suggested the two peaks probably
represent a feeding pattern which would further suggest a
relationship to their hoarding.

The activity of g, 1, goveboracensi§_also shows two
peaks during winter nights, although the second peak is
typically not as high as the first (Orr, 1959). In E: m,
bairdi both peaks are substantial and this may account for
the failure to find any significant difference in the
amount of hoarding done by g, 1. noveboracensis under either
photoperiod. The small but non-significant increase in the
amount of hoarding under a short day photoperiod in all the
other groups, except a. 1. castaneus, may also suggest a
similar pattern of nocturnal activity and hoarding in these
mice. However, the failure of £3 1, gastaneus to show this
trend suggests that a study of their activity patterns
might be fruitful.

Obviously, the above offered explanation for the
increase in hoarding under a short day photoperiod does
not explain the intensive food storage by many diurnal
rodents during the shortening days of fall and early
winter. A study on the hoarding behavior of such species
as chipmunks, ground squirrels, or one of the diurnal tree

squirrels under a natural photoperiod would provide

118

conclusive evidence of whether the mechanism which has been
proposed as acting in flying squirrels (Muul, 1968) has any
generality among the other rodents or even in other
sciurids. It will also be absolutely essential that all
learning be controlled for, as this was the one weakness
of Muul's study in which he used wild caught animals.
ggssible Physi21ogical Mechanism
for Hoarding
Both 13, 1. ggveboracensis and g. g. bairdi showed a
significant difference in the amount of food hoarded under
experimental conditions simulating autumn or early winter
as compared to the amount hoarded under summer conditions.
The highly significant response of g, 1, ggveboracensis to
the "fall" conditions in my study, agrees well with the
finding by Lynch (1973) that this subspecies when kept
under a short photoperiod (9L:1SD) and a cold temperature
(5°C) will show physiological changes associated with the
fall of the year. Starting approximately 4 weeks after
acclimation to these conditions, some of the mice in
Lynch's study began to show both a "fall" molt and gonadal
regression. Photoperiod was apparently the more important
influence, since cold tended to accelerate molting (as
compared to groups under the same photoperiod but at 27°C)
but had no appreciable effect on the rate of gonadal
regression. Both males and females responded similarly to
the photoperiod and temperature conditions in regards to

both their molting and gonadal regression. Weight changes,

 

119

if any, were not reported. Lynch (1973) proposed this as
case in which multiple environmental cues permit a more
reliable assessment of critical changes in the environment
leading to a higher probability of survival. Since some of
the g,‘1. noveboracensis in my own study (exactly 50% of
those tested) were also acclimated under a 9L:15D photo-
period and under 7°C, it is reasonable to assume that the
physiological changes reported by Lynch were occurring.
Table 7 illustrates that under a short photoperiod, g, 1,
noveboracensis kept at 7°C over a period of 5 weeks hoarded
more than twice as much as those mice kept at 27°C and
under a long photoperiod. The finding that test temperature
had more of an influence than photoperiod appears to be the
one major difference in the findings of Lynch and this
study. Obviously, there is yet insufficient evidence to
suggest that the physiological mechanisms underlying molt
and gonadal regression are the same as those for hoarding.
Additional study is also needed on g, m, bairdi to see
whether similar changes occur in this species under con-
trolled laboratory conditions of photoperiod and temper-

ature .

Torpor and Food Hoardigg

 

The high level of hoarding by 3‘ eremicus even in
contrast to the northern subspecies, 3° 1, goveboracensis
and 3° 3. bairdi, is at first a little perplexing. I have

attributed the intensive hoarding of both E. 1.

120

noveboracensis and g. m, bairdi as being a necessary

 

 

response in the northern latitudes where seasonal fluctu-
ations in food supplies can be severe. The fact that food
supplies in other environments may also fluctuate has only
been given brief mention up to now. Food supplies in the
desert habitats from which both E, eremicus and g, m,
blandus were derived to undergo seasonal changes, although
at different times of the year than in the northern
latitudes. MacMillen (1965) has reported that g, eremicus
will enter torpor during both the winter and summer in their
habitat. During the winter, torpor can only be induced
through food restriction, but during the summer either food
restriction or negative water balance is capable of inducing
it. MacMillen has prOposed that summer torpor probably
allows E, eremicus to remain in its humid burrows for
several weeks, if necessary, to avoid the demanding exter-
nal desert environment. He also mentions that torpor could
be one method of prolonging the food stores which are

needed during these periods of forced inactivity. Un-
fortunately, no field evidence exists to confirm whether

a, eremicus hoard prior to summer torpor. The exceptionally
high level of hoarding of this species in the present study
suggests, however, that this trait has been heavily selected
for. This species' obligatory restriction to burrows

during the summer would be a strong selection pressure on
the development of the hoarding habit. As pointed out

earlier, the fact that E. m. blandus hoard significantly

 

121

more under a warm temperature than under a cold one, may
indicate that it, too, hoards for periods of forced
inactivity during the exceptionally hot summer months.
Until someone studies whether torpor is also the normal
reaponse of blandus to food and water deprivation in both
winter and summer, the reasoning given here remains viable.
Whether the presence of daily torpor in both E, g, bairdi
(Fuller, Stebbins, & Dyke, 1969) and g, 1, ggveboracensis
(Gaertner, Hart, 5 Roy, 1973) also has any relationship to
their hoarding behavior remains to be investigated.
Apparently both of these species will undergo temperature
induced torpor even in the presence of stored food (Howard,

1951; Gaertner 21 al., 1973).

Weight Change and Hoarding
The finding that a, 1, goveboracensis lose weight
during acclimation to a cold temperature appears to be
consistent with what has already been reported for this
subspecies (Hart, 1953). Although in this study both a. 1,

noveboracensis and a, 1, castaneus lost weight under both

 

acclimation temperatures, they did lose significantly more
weight under 7°C than under 27°C. The general loss in
weight under both temperatures could have been a response
to the experimental conditions, perhaps induced by their
social isolation while in the experiment. Animals that
were obese showed the greatest changes in weight in all the

subspecies. The increases in weight of E, m. bairdi under

 

122

the two acclimation temperatures demonstrate, however, the
diversity of responses possible within this genus to both
temperature and social isolation. The slightly greater
gain in weight by bairdi under 7°C is in agreement with
Sealander's finding that laboratory raised members of this
subspecies show a significant increase in weight when
acclimated to 8.5°C versus 30.5°C (Sealander, 1951). The
highly significant loss in weight by E, eremicus under the
cold acclimation temperature and their gain under the warm
temperature indicates a lack of stability in their body
weight that may have some ecological significant in their
desert environment. The inconsistent effect of acclimation
temperature and sex on weight change in g, m, blandus, a
sympatric species with B. eremicus, deserves further
investigation.

The results of this study suggest that there is little
evidence of any relationship between weight change (either
weight loss or gain) and hoarding behavior in deermice. The
correlation between weight change and hoarding in a, m,
bairdi was small, whereas the correlations in the other
groups were all non-significant. There was also no apparent
relationship between the factors which influenced weight
change and the factors influencing hoarding. Under a short

day photoperiod, E. 1. castaneus hoarded significantly more

 

after acclimation to 27°C than after 7°C, even though it

lost more weight under the colder temperature. 3. 1.

noveboracensis, which had similarly lost weight under the cold

 

123

temperatures, failed to show any significant difference in
the amount of food hoarded after acclimation to either
temperature under the same photoperiod. Test temperature,
which could have only a minimal effect on weight change
because it lasted for only 7 days, had more of an influence
on hoarding than acclimation temperature in all the groups.
Changes in weight under the two photoperiods also tended

to be an equally poor predictor of hoarding. E, m, bairdi
hoarded significantly more under the short day photoperiod
than under the long, but tended to gain weight under the
longer photOperiod. These findings would appear to con-
tradict the model of Herberg and Blundell (1970) in which
they proposed that hoarding in rats at least, is motivated
by long term nutritional needs or "deficits" (i.e., losses
in body weight or changes in the size of fat reserves)
acting through the lateral hypothalamus. The report that
E, 1, noveboracensis apparently gain weight during fall and
early winter (Sealander, 1951) when they would be most

actively hoarding, tends to further dispute their model.

Conclusions

 

The hypothesis that the hoarding behavior of deermice
is under the influence of environmental factors has been

confirmed by my study. The two northern subspecies of

 

both E, leucopus and E. maniculatus found in Michigan do
hoard more under laboratory conditions of a low temperature

and short photoperiod, than they do under a high

124

temperature and long photoperiod. A similar response in
the wild would lead to the accumulation of food in'the fall
of the year which could then be used during a winter
scarcity. Similarly, my results suggest that the response
of E, m, blandus to a warm temperature may have some
relationship to the seasonal changes in its food supplies
in the southwestern deserts.

My results also indicate that temperature and photo-
period are not of equal importance to both E, leucopus

noveboracensis and g, maniculatu§_bairdi. E, m, bairdi

 

was more influenced by photoperiod than by temperature, but
in g, 1, noveboracensis just the opposite was true. As
both of these species are sympatric in Michigan, further
research is needed to see whether this finding has any
relationship to their habitat requirements.

There remains a great deal of work to be done on the
hoarding behavior of geromyscus, especially in regards to
the collection of good field data. With one exception
(Howard & Evans, 1961), field reports of food storage by
members of this genus have been only casual in nature with
little information as to how much food is hoarded, when or
where it is hoarded, or by how many individuals. The lack
of experimental interest in their hoarding up till now is
surprising in view of the considerable attention given to
other aspects of the behavioral ecology of this genus (for
example, see King, 1968). However, recent advances in the

use of radioactive tracers ought to facilitate field

125

studies of hoarding as radioactively labeled food could be
provided and then relocated after storage. This technique
could be used in a variety of habitats with equal efficiency
and might provide answers to the many questions raised in

this dissertation.

SUMMARY

1. Five taxa of deermice including a, 1, goveboracensis,

 

3° 1. castaneus, g. m, bairdi, 13, m, blandus, and 1’,
eremicus were tested in the laboratory for the amount of
food hoarded on seven consecutive days under various com-
binations of test temperature (7°C or 27°C) and photoperiod

(9L:15D or 15L:9D).

2. The two northern subspecies of both a, leucopus and

 

g, maniculatu§_hoarded significantly more food under a low
temperature and a short day photoperiod than they did under
a high temperature and long day photoperiod. A similar
response to these environmental variables in the wild would
lead to a higher level of hoarding in the fall of the year

than in the summer.

3. Comparisons of 3°.lw noveboracensis with E, 1.
castaneus, and a, m, bairdi with E, m, blandus revealed
that in both cases the northern subspecies hoarded signifi-
cantly more than the southern one under the cold test
temperature. Under the warm test temperature, there was

no significant difference between their scores.

126

127

4. Photoperiod and temperature were not of equal
importance to both E, 1. noveboracensis and E, m, bairdi.
Test temperature had more of an effect than photoperiod on
the hoarding of E, 1, noveboracensis, while in E, m,
bairdi photoperiod was the more important factor. Test
temperature also had a significant effect on E, g, blandus
which hoarded more under the 27°C test temperature than

under 7°C.

5. The high level of hoarding by E, eremicus in
contrast to the other groups studied, may be related to
the occurrence of summer torpor in this species and its
restriction to burrows during extremes in temperature in

its desert habitat.

6. In all of the groups studies, except perhaps E, m,
bairdi, there was no evidence found to support a model of
food hoarding in deermice based on long term nutritional
needs or deficits (i.e., losses in body weight) as proposed
by Herberg and Blundell (1970) for laboratory rats. In
E, m, bairdi there was only a small correlation between
weight change and hoarding scores. In all other groups

this correlation was non-significant.

APPENDIX

Table l2.-—Analysis of Variance of the Log Hoarding Scores
of E, leucopus noveboracensis.

 

 

 

Source of Variance df MSS F Prob.
Between Subjects
Sex 1 1.3383 1.4034 .239
Accl. Temp. 1 1.2347 1.2949 .258
Photoperiod l .9278 .9730 .326
Test Temp. 1 6.0909 6.3875 1211
Sex x Accl. Temp. 1 .0314 .0329 .856
Sex x Photop. 1 .0283 .0297 .863
Sex x Test Temp. 1 .2119 .2222 .638
Accl. Temp. x Photop. 1 1.4556 1.5264 .219
Accl. Temp. x Test Temp. 1 2.4868 2.6079 .109
Photop. x Test Temp. 1 .3847 .4034 .527
Sex x Accl. Temp. x Photop. 1 2.9956 3.1414 .079
Sex x Accl. Temp. x Test
Temp. 1 .7922 .8308 .364
Sex x Photop. x Test Temp. 1 .1690 .1772 .675
Accl. Temp. x Photop. x
Test Temp. 1 .1192 .1250 .724
Sex x Accl. Temp. x Photop.
x Test Temp. 1 .2375 .2491 .619
Error 112 .9536

128

129

Table 12.--continued.

 

 

 

Source of Variance df MSS F Prob.
Within Subjects
Days 6 1.1672 12.7589 <1222§
Sex x Days 6 .0601 .6573 .684
Accl. Temp. x Days 6 .1016 1.1108 .354
Photop. x Days 6 .0245 .2673 .952
Test Temp. x Days 6 .1247 1.3627 .227
Sex x Accl. Temp. x Days 6 .0555 .6067 .725
Sex x Photop. x Days 6 .0486 .5315 .785
Sex x Test Temp. x Days 6 .0925 1.0109 .417
Accl. Temp. x Photop. x Days 6 .0518 .5665 .757
Accl. Temp. x Test Temp. x
Days 6 .0798 .8724 .515
Photop. x Test Temp. x Days 6 .1164 1.2721 .268
Sex Accl. Temp. x Photop. x
Days 6 .0985 1.0767 .375
Sex x Accl. Temp. x Test
Temp. x Days 6 .0593 .6487 .691
Sex x Photop. x Test Temp.
x Days 6 .0106 .1154 .995
Accl. Temp. x Photop. x
Test Temp. x Days 6 .0805 .8801 .509
Sex x Accl. Temp. x Photop. x
Test Temp. x Days 6 .0837 .9149 .483
Error 672 .0915
Total 895 (SS=200.3108)

 

130

Table 13.--Analysis of Variance of the Log Hoarding Scores
of E, leucgpus gastaneus.

 

 

 

Source of Variance df MSS F Prob.
Between Subjects
Sex 1 2.6227 4.0335 1211
Accl. Temp. 1 .1336 .2055 .651
Photoperiod l .4666 .7175 .399
Test Temp. 1 .5826 .8960 .346
Sex x Accl. Temp. 1 4.1030 6.3100 1211
Sex x Photop. 1 1.5196 2.3370 .129
Sex x Test Temp. 1 1.1225 1.7263 .192
Accl. Temp. x Photop. 1 7.4910 11.5204 .001
Accl. Temp. x Test Temp. 1 .7044 1.0833 .300
Photop. x Test Temp. 1 .3801 .5846 .446
Sex x Accl. Temp. x Photop. 1 .0804 .1236 .726
Sex x Accl. Temp. x Test
Temp. 1 .0000 .0000 .993
Sex x Photop. x Test Temp. 1 1.6297 2.5062 .116
Accl. Temp. x Photop. x
Test Temp. 1 .2086 .3209 .572
Sex x Accl. Temp. x Photop.
x Test Temp. 1 1.0076 1.5495 .216
Error 112 .6502

Table 13.--continued.

131

 

 

 

 

Source of Variance df MSS F Prob.
Within Subjects
Days 6 1.4744 23.1871 <.0005
Sex x Days 6 .0498 .7839 .583
Accl. Temp. x Days 6 .0624 .9820 .436
Photop. x Days 6 .0181 .2845 .944
Test Temp. x Days 6 .2875 4.5217 <.0005
Sex x Accl. Temp. x Days 6 .0293 .4604 .838
Sex x Photop. x Days 6 .0549 .8629 .522
Sex x Test Temp. x Days 6 .0736 1.1579 .327
Accl. Temp. x Photop. x Days 6 .0433 .6812 .665
Accl. Temp. x Test Temp. x
Days 6 .0228 .3579 .905
PhotOp. x Test Temp. x Days 6 .0515 .8092 .563
Sex x Accl. Temp. x
Photop. x Days 6 .0645 1.0141 .415
Sex x Accl. Temp. x Test
Temp. x Days 6 .0789 1.2416 .283
Sex x Photop. x Test Temp.
x Days 6 .0824 1.2962 .257
Accl. Temp. x Photop. x Test
Temp. x Days 6 .0372 .5856 .742
Sex x Accl. Temp. x Photop.
x Test Temp. x Days 6 .0968 1.5215 .168
Error 672 .0636
Total 895 (SS=152.7749)

 

132

Thble l4.--Analysis of Variance of the Log Hoarding Scores

of E, maniculatu§_bairdi.

 

 

 

Source of Variance df M85 P Prob.
Between Subjects
Sex 1 .0076 .0078 .930
Accl. Temp. 1 .9124 .9316 .337
Photoperiod 1 3.9295 4.0125 °Qi§
Test Temp. 1 .0072 .0074 .932
Sex x Accl. Temp. 1 1.9706 2.0123 .159
Sex x Photop. 1 .3376 .3447 .558
Sex x Test Temp. 1 .0219 .0219 .883
Accl. Temp. x Photop. 1 1.3783 1.4074 .238
Accl. Temp. x Test Temp. 1 .4548 .4644 .497
Photop. x Test Temp. 1 .2131 .2175 .642
Sex x Accl. Temp. x Photop. 1 1.2986 1.3260 .252
Sex x Accl. Temp x Test Temp. 1 .0430 .0439 .834
Sex x Photop. Test Temp. 1 .6259 .6391 .426
Accl. Temp. x Photop. x
Test Temp. 1 .0800 .0817 .776
Sex x Accl. Temp. x
Photop. x Test Temp. 1 1.2745 1.3015 .256
Error 112 .9793

133

Table l4.--continued.

 

Source of Variance df M38 P Prob.

Within Subjects

Days 6 .0599 .6942 .654
Sex x Days 6 .1495 1.7337 .111
Accl. Temp. x Days 6 .1125 1.3036 .253
Photop. x Days 6 .0298 .3458 .912
Test Temp. x Days 6 .1780 2.0628 ,EEQ
Sex x Accl. Temp. x Days 6 .0086 .0994 .996
Sex x Photop. x Days 6 .1347 1.5610 .156
Sex x Test Temp. x Days 6 .1082 1.2542 .277
Accl. Temp. x Photop. x Days 6 .0672 .7790 .587
Accl. Temp. x Test Temp. x Days 6 .0299 .3468 .912
Photop. x Test Temp x Days 6 .0881 1.0217 .410
Sex x Accl. Temp. x Photop.

x Days 6 .0218 .2528 .958
Sex x Accl. Temp. x Test

Temp. Days 6 .0781 .9057 .490
Sex x Photop. Test Temp.

x Days 6 .0832 .9640 .449
Accl. Temp. x Photop. x

Test Temp. x Days 6 .1469 1.7023 .118
Sex x Accl. Temp. x Photop.

x Test Temp. x Days 6 .1159 1.3437 .235
Error 672 .0863

Total 895 (SS=188.6874)

 

134

Table 15.--Ana1ysis of Variance of the Log Hoarding Scores
of E, maniculatus blandus.

 

Source of Variance df MSS F Prob.

 

Between Subjects

Sex 1 .0611 .1025 .749
Accl. Temp. 1 .8940 1.4986 .223
Photop. 1 .8583 1.4387 .233
Test Temp. 1 4.0390 6.7706 "211
Sex x Accl. Temp. 1 .1210 .2029 .653

Sex x Photop. 1 .2751 .4612 .498
Sex x Test Temp. 1 1.1721 1.9647 .164
Accl. Temp. x Photop. 1 .2066 .3464 .557
Accl. Temp. x Test Temp. 1 .0150 .0251 .874
Photop. x Test Temp. 1 1.2532 2.1007 .150
Sex x Accl. Temp. x Photop. 1 .2506 .4201 .518
Sex x Accl. Temp. x Test

Temp. 1 .0127 .0212 .884
Sex x Photop. x Test Temp. 1 .5474 .9176 .340
Accl. Temp. x Photop. x

Test Temp. 1 .2163 .3626 .548
Sex x Accl. Temp. x

Photop. x Test Temp. 1 .3625 .6076 .437

Error 112 .5966

135

Table 15.--continued.

 

 

 

Source of Variance df MSS F Prob.
Within Subjects
Days 6 .1189 1.4569 .191
Sex x Days 6 .0202 .2481 .960
Accl. Temp. x Days 6 .0726 .8895 .502
Photop. x Days 6 .0675 .8272 .549
Test Temp. x Days 6 .2379 2.9151 .EEE
Sex x Accl. Temp. x Days 6 .0707 .8660 .519
Sex x Photop. x Days 6 .0744 .9111 .486
Sex x Test Temp. x Days 6 .0586 .7180 .635
Accl. Temp. x Photop. x Days 6 .0270 .3310 .921
Accl. Temp. Test Temp. x Days 6 .0728 .8918 .500
Photop. x Test Temp. x Days 6 .0603 .7394 .618
Sex x Accl. Temp. x
Photop. x Days 6 .1800 2.2055 .311
Sex x Accl. Temp. x Test
Temp. x Days 6 .0267 .3269 923
Sex x Photop. x Test Temp.
x Days 6 .0124 .1519 .989
Accl. Temp. x Photop. x
Test Temp. x Days 6 .1312 1.6080 142
Sex x Accl. Temp. x Photop.
x Test Temp. x Days 6 .1002 1.2276 .290
Error 672 .0816
Total 895 (SS=139.9269)

136

Table 16.-~Analysis of Variance of Log Hoarding Scores
of E. eremicus ggemicug.

 

 

Source of Variance df M88 P Prob
Temperaturea 1 .0049 .0357 .351
Photoperiod 1 .2876 2.0797 .158
Acclimation Temp. x
Photoperiod l .0944 .6828 .414
Residual error 36 .1383
Total 39 (SS=5.3647)

 

aAll mice were tested at the same temperature to which

they had been acclimated.

137

Table 17.--Combined Analysis of Variancg for the Mean Log
Hoarding Scores (i.e., Log (X+l) for Each Mouse)
of E. 1. pgveboracensig, E. 1. gasganeus, P. m.
bairdi, and E. m. blandus. - _

 

 

 

Source of Variance df MSS F Prob.
Subspecies 3 .5795 4.4134 .222
Sex 1 .2217 1.6887 .194
Subspecies x Sex 3 .1858 1.4152 .238
Accl. Temp. 1 .3704 2.8212 .094
Subspecies x Accl. Temp. 3 .1230 .9371 .423
Sex x Accl. Temp. 1 .0409 .3114 .577
Subspeciex x Sex x Accl. Temp. 3 .2689 2.0477 .106
Photoperiod l .4476 3.4083 .065
Subspecies x Photop. 3 .2080 1.5845 .192
Sex x Photop. l .0634 .4829 .487
Subspecies x Sex x Photop. 3 .1220 .9288 .427
Accl. Temp. x Photop. 1 .1803 1.3729 .242
Subspecies x Accl. Temp.

x Photop. 3 .4516 3.4395 .211
Sex x Accl. Temp. x Photop. 1 .0769 .5860 .444
Subspecies x Sex x Accl.

Temp. x Photop. 3 .3618 2.7553 .222
Test Temp. 1 .1018 .7753 .379
Subspecies x Test Temp. 3 .6374 4.8545 .222_
Sex x Test Temp. 1 .0458 .3486 .555
Subspecies x Sex x Test Temp. 3 .1782 1.3571 .255
Accl. Temp. x Test Temp. 1 .4084 3.1102 078
Subspecies x Accl. Temp.

x Test Temp 3 .0092 .0703 976

Table 17.--continued.

138

 

Source of Variance df MSS F Prob.
Sex x Accl. Temp. x Test
Temp. 1 .0126 .0959 .757
Subspecies x Sex x Accl.
Temp. x Test Temp. 3 .0316 .2407 .868
Photop. x Test Temp. 1 .4967 3.7827 .052
Subspecies x Photop. x
Test Temp. 3 .0213 .1626 .922
Sex x Photop. x Test Temp. 1 .0942 .7177 .397
Subspecies x Sex x Photop.
x Test Temp. 3 .1552 1.1819 .316
Accl. Temp. x Photop. x
Test Temp. 1 .0177 .1347 .714
Subspecies x Accl. Temp. x
Photop. x Test Temp. 3 .0160 .1220 .947
Sex x Accl. Temp. x Photop.
x Test Temp. 1 .2864 2.1812 .140
Subspecies x Sex x Accl.
Temp. x Photop. x Test Temp. 3 .0576 .4390 .725
Residual Error 448 .1313
Total 511

 

139

Table 18.--Combined Analysis of Variance for the Mean Weight
Changes in P. 1. noveboracensis, E. 1. castaneus,
E. m. bairdi, and E. m. blandus.

 

 

 

Source of Variance df M88 P Prob.
Subspecies 3 255.6920 22.6762 .222§_
Sex 1 21.2063 1.8807 .171
Subspecies x Sex 3 10.8070 .9584 .412
Accl. Temp. 1 72.1500 6.3987 .212
Subspecies x Accl. Temp. 3 57.9114 5.1359 ._22
Sex x Accl. Temp. 1 .1188 .0105 .918
Subspecies x Sex x Accl. Temp. 3 30.9308 2.7431 .222
Photoperiod 1 .8945 .0793 .778
Subspecies x Photop. 3 3.0120 .2671 .849
Sex x Photop. 1 .1070 .0095 .922
Subspecies x Sex x Photop. 3 39.7912 3.5289 .212
Accl. Temp. x Photop. l .4395 .0390 .844
Subspecies x Accl. Temp.

Photop. 3 15.6312 1.3863 .246
Sex x Accl. Temp. x Photop. 1 80.1695 7.1099 .222
Subspecies x Sex x Accl.

Temp. x Photop. 3 15.0153 1.3316 .263
Test Temp. 1 2.3926 .2122 .645
Subspecies x Test Temp. 3 5.3758 .4768 .699
Sex x Test Temp. 1 48.8813 4.3351 .222
Subspecies x Sex x Test Temp. 3 59.6087 5.2864 .221
Accl. Temp. x Test Temp. 1 13.1970 1.1704 .280
Subspecies x Accl. Temp. x

Test Temp. 3 12.3776 1.0977 350

140

Table 18.--continued.

 

Source of Variance df MSS F Prob.

 

Sex x Accl. Temp. x Test

Temp. 1 17.0382 1.5110 .220
Subspecies x Sex x Accl.

Temp. x Test Temp. 1 28.1301 2.4947 .059
Photop. x Test Temp. 1 .1876 .0166 .897
Subspecies x Photop. x

Test Temp. 3 24.7451 2.1945 .088
Sex x Photop. x Test Temp. 1 17.9251 1.5897 .208

Subspecies x Sex x Photop.
x Test Temp. 3 1.8783 .1666 .919

Accl. Temp. x Photop. x
Test Temp. 1 11.8220 1.0484 .306

Subspecies x Accl. Temp. x
Photop. x Test Temp. 3 18.0512 1.6009 .188

Sex x Accl. Temp. x Photop.
x Test Temp. 1 49.3770 4.3790 . 37

Subspecies x Sex x Accl.

Temp. x Photop. x

Test Temp. 3 7.2797 .6456 .586
Residual Error 448 11.2758

Total 511

141

 

 

 

Table 19.--Ana1ysis of Variance of the Weight Change in
E, eremicus.

Source of Variance df MSS F Prob.
Temperaturea 1 62.7503 7.9834 .008
Photoperiod 1 2.6523 .3374 .565
Temperature x Photoperiod 1 26.0823 3.3183 .077
Residual Error 36 7.8601
Total 39 (SS=374.4478)

aAll mice were tested at the same temperature to which

they had been acclimated.

BIBLIOGRAPHY

BIBLIOGRAPHY

Albino, R. C. and M. Long. 1951. The effect of infant
food deprivation upon adult hoarding in the white rat.
Brit. J. Psychol. 42:146-154.

Allen, D. L. 1943. Michigan fox squirrel management.
Michigan Dept. Cons., Game Div. Publ. 100. Lansing.
404pp.

Baker, R. 1968. Habitats and distribution. In: J. A.
King (ed.), Biology of Peromyscus (Rodentia). Spec.
Publ. No. 2, Amer. Soc. Mbmm. pp. 98-126.

Barash, D. D. 1973. Territorial and foraging behavior of
pika (Ochotona princeps) in Montanna. Amer. Midl. Nat.
89:202-207.

Bindra, D. 1947. Water hoarding in rats. J. Comp.
Physiol. Psychol. 40:149-156.

Bindra, D. 1948a. Hoarding behavior of rats: nutritional
and psychological factors. Unpubl. Doctoral Disserta-
tion, Harvard University.

Bindra, D. 1948b. The nature of motivation for hoarding
food. J. Comp. Physiol. Psychol. 41:211-218.

Bindra, D. 1948c. What makes rats hoard? J. Comp.
Physiol. Psychol. 41:397-402.

Blundell, J. E. 1970. An investigation of some moti-
vational functions of the lateral hypothalamus. Ph.D.
Thesis, University of London. (Quoted by Blundell,
1971).

Blundell, J. E. 1971. Possible mechanism for the effect

of anorexic agents on feeding and hoarding behaviour
in rats. Psychopharmacologia (Berl.). 22:224-229.

142

143

Blundell, J. E. and L. J. Herberg. 1973. Effectiveness of
lateral hypothalamic stimulation, arousal, and food
deprivation in the initiation of hoarding behaviour in
naive rats. Physiol. Beh. 10:763-767.

Brown, L. G. and Yeager, L. E. 1945. Fox squirrels and
gray squirrels in Illinois. Bull. Ill. Nat. Hist.
Survey. 23:499-536.

Buller, A. H. R. 1920. The red squirrel of North America
as a mycophagist. Trans. British Mycological Soc.
6:355-362.

Burns, R. D. 1958. Foraging in kangaroo rats and its
adaptive significance. Ph.D. Thesis, Mich. State Univ.
56pp.

Burt, W. H. 1946. The mammals of Michigan. The University
of Michigan Press. Ann Arbor. 288pp.

Calhoun, J. B. 1963. The ecology and sociology of the
Norway rat. U.S. Dept. Health, Ed., and Welfare.
Public Health Service Publ. No. 1008. Bethesda,
Maryland. 288pp.

Criddle, S. 1950. The P. m. bairdi complex in Manitoba.
Canad. Field-Nat. 61': 189-177." '

Dalquest, W. W. 1948. Mammals of Washington. Univ.
Kansas Publ. Mus. of Nat. Hist. Vol. 2. 444pp.

Davis, D. E. and F. B. Golley. 1963. Principles of
Mammalogy. Reinhold Publ. Co., New York. 355 pp.

Degerbol, M. 1927. Vidensk. Meddel. Dansk. Naturhist.
84:195-202.

Denenberg, V. H. 1952. Hoarding in the white rat under
isolation and group conditions. J. Comp. Physiol.
Psychol. 45:497-503.

Eisenberg, J. F. 1968. Behavior patterns. In: J. A.
King (ed.), Biology of Eeromyscus (Rodentia). Spec.
Publ. No. 2, Amer. Soc. Mamm. pp. 451-495.

 

Eibl-Eibesfeldt, I. 1963. Angeborenes und Erworbenes im
Verhalten einiger Sauger. Z. Tierpsychol. 20:705-754.

Ewer, R. F. 1965. Food burying in the African ground
squirrel Xerus erythropus (E. Geoff.). Zeitschrift
fur Tierpsychologie. 22:321-327.

144

Ewer, R. F. 1967. The behaviour of the African giant rat
(Cricetomys ambianus Waterhouse). Zeitschrift ffir
Tierpsychologie. 24:6-79.

Falls, J. B. 1968. Activity. In: J. A. King (ed.),
Biology of Peromyscus (Rodentia). Spec. Publ. No. 2,
AMI. SOC. Mam. Pp. 543-570.

Fuller, W. A., L. L. Stebbins, and G. R. Dyke. 1969. Over-
wintering of small mammals near Great Slave Lake,
northern Canada. Artic. 22:34-35.

Gaertner, R. A., J. S. Hart, and O. 2. Roy. 1973. Seasonal
spontaneous torpor in the white-footed mouse, Peromyscus
1gucopus. Comp. Biochem. Physiol. 45A:169-181.

Grinnell, J. and T. I. Storer. 1924. Animal life in the
Yosemite. Univ. Calif. Press. Berkeley. 752pp.

Hamilton, W. J. 1930. The food of the soricidae. J.
Mamm. 11:26-39.

Hamilton, W. J. 1939. American mammals. Their lives,
habits, and economic relations. McGraw-Hill. New
York. 434pp.

Hammer, L. R. 1972. Further hoarding preferences in
hamsters. Psychon. Sci. 26:139-140.

Hardy, G. A. 1949. Squirrel cache of fungi. Canad.
Field-Nat. 63:86-87.

Hart, J. S. 1953. The relation between thermal history
and cold resistance in certain species of rodents.
Can. J. 2001. 31:80-98.

Hart, J. S. 1971. Rodents. In: G. Whittow (ed.),
Comparative Physiology of Thermoregulation. Vol. II.
Acad. Press. New York. pp. 1-149.

Herberg, L. J. and J. E. Blundell. 1967. Lateral
hypothalamus: hoarding behavior elicited by electrical
stimulation. Science. 155:349-350.

Herberg, L. J. and J. E. Blundell. 1970. Non-interaction
of ventromedial and lateral hypothalamic mechanisms in
the regulation of feeding and hoarding behavior in the
rat. Q. J. Exp. Psychol. 22:133-141.

Herberg, L. J., J. G. Pye, and J. E. Blundell. 1972. Sex
differences in the hypothalamic regulation of food
hoarding: hormones versus calories. Anim. Behav.
20:186-191.

145

Holland, J. G. 1954. The influence of previous experi-
ence and residual effects of deprivation on hoarding
in the rat. J. Comp. Physiol. Psychol. 47:244-247.

Howard, W. E. 1949. Dispersal, amount of inbreeding, and
longevity in a local population of prairie deermice on
the George Reserve, southern Michigan. Contrib. from
the Lab. Vert. Biol. Univ. Mich. 43:1-50.

Howard, W. E. 1951. Relation between low temperature and
available food to survival of small rodents. J. Mamm.
32:300-312.

Howard, W. E. and F. C. Evans. 1961. Seeds stored by
prairie deer mice. J. Mamm. 42:260-263.

Hunt, J. McV. 1941. The effects of infant feeding-
frutration upon adult hoarding behavior. J. Abn. Soc.

Hunt, J. McV., H. Schlosberg, R. L. Solomon, and E. Stellar.
1947. Studies of the effects of infantile experience
on adult behavior in rats. I. Effects of infantile
feeding-frustration on adult hoarding. J. Comp.
Physiol. Psychol. 40:291-304.

Ingram, W. M. 1942. Snail associates of Blarina brevicauda
talpoides. J. Mamm. 23:255-258.

Johnson, M. S. 1926. Activity and distribution of certain
wild mice in relation to biotic communities. J. Mamm.
7:245-277.

Kinder, E. F. 1927. A study of the nest building activity
of the albino rat. J. Exp. 2001. 47:117-161.

King, J. A. 1968. Psychology. In: J. A. King (ed.),
Biology of Peromyscgg (Rodentia). Spec. Publ. No. 2,
Amer. Soc. Mamm. pp. 496-542.

Klugh, A. B. 1927. Ecology of the red squirrel. J. Mamm.
8:1-32.

Koski, J. 1963. The bimodal distribution of hoarding
scores in the golden hamster. Proc. Penn. Acad. Sci.
37:291-294.

Lapetite, D. and A. Soulairac. 1951. Action du methylthio-
uracile sur 1e comportment d'amassement du rat.
Seance. 145:505.

146

Licklider, L. C. and J. C. R. Licklider. 1950. Obser-
vations on hoarding behavior of rats. J. Comp.
Physiol. Psychol. 43:129-134.

Lindzey, G. and M. Manosevitz. 1964. Hoarding in the
mouse. Psychonomic Sci. 1:35-36.

Lockner, F. R. 1972. EXperimental study of hoarding in
the red-tailed chipmunk, Eutamias ruficaudus. z.
Tierpsychol. 31:410-418.

Lyman, C. P. 1954. Activity, food consumption and
hoarding in hibernators. J. Mamm. 35:545-552.

Lynch, G. R. 1973. Effect of simultaneous exposure to
differences in photoperiod and temperature on the
seasonal molt and reproductive system of the white-

.footed mouse, Perom scus leucgpug, Comp. Biochem.
Physiol. 44:1573-Ig76.

MacMillen, R. E. 1965. Aestivation in the cactus mouse,
Peromyscus eremicus. Comp. Biochem. Physiol. 16:

Mailliard, J. 1931. Redwood chickaree testing and storing
hazel nuts. J. Mamm. 12:68370.

Manosevitz, M. 1965. Genotype, fear and hoarding. J.
Comp. Physiol. Psychol. 60:412-416.

Manosevitz, M. 1970. Prolonged aperiodic feeding and
adult hoarding in mice. J. Comp. Physiol. Psychol.
70:228-234.

Manosevitz, M., R. B. Campenot, and C. F. Swencionis.
1968. Effects of enriched environment upon hoarding.
J. Comp. Physiol. Psychol. 66:319-324.

Manosevitz, M. and G. Lindzey. 1970. Genetic variation and
hoarding. In: G. Lindzey and D. D. Thiessen (eds.),

Contributions to behavior-genetic analysis: the mouse
as a prototype. Appleton-Century-CrOfts. New York.
336pp.

Marx, M. H. 1950a. A stimulus-response analysis of the
hoarding habit in the rat. Psychol. Rev. 57:80-94.

Marx, M. H. 1950b. Experimental analysis of the hoarding
habit in the rat: 1. Preliminary observations. J.
Comp. Physiol. Psychol. 43:295-308.

147

Marx, M. H. 1951. Experimental analysis of the hoarding
habit in the rat: II. Terminal reinforcement. J.
Comp. Physiol. Psychol. 44:168-177.

Marx, M. H. 1952. Infantile deprivation and adult
behavior in the rat: Retention of increased rate of
eating. J. Comp. Physiol. Psychol. 45:43-49.

McCabe, T. T. and B. Blanchard. 1950. Three Species of
Eeromyscus. Rood Associates. Santa Barbara. l36pp.

McCain, 6., G. L. Garrett, C. Reed, G. Mead, and R.
Kuenstler. 1964. Effects of deprivation on hoarding
of objects other than the deprived material. Anim.
Behav. 12:409-415.

McCleary, R. A. and C. T. Morgan. 1946. Food hoarding in
rats as a function of environmental temperature. J.
Comp. Psychol. 39:371-378.

McDougall, W. 1923. Introduction to social psychology.
Robert B. Luce, Inc. Washington, D.C.

McKelvey, R. K. and M. H. Marx. 1951. Effects of infantile
food and water deprivation on adult hoarding in the
rat. J. Comp. Physiol Psychol. 44:423-430.

McNab, B. K. and P. R. Morrison. 1963. Body temperatures
and metabolism in subspecies of Perom scus from arid
and mesic environments. Ecol. Monogr. 33:63-82.

Miller, G. A. and L. Postman. 1946. Individual and group
hoarding in rats. Amer. J. Psychol. 59:652-668.

Miller, G. A. and P. Viek. 1950. Hoarding in the rat as
a function of the length of the path. J. Comp.
Physiol. Psychol. 43:66-69.

Morgan, C. T. 1945. Statistical treatment of hoarding
data. J. Comp. Psychol. 38:247-256.

Morgan, C. T. 1947. The hoarding instinct. Psychol. Rev.
54:335-341.

Morgan, C. T., E. Stellar and 0. Johnson. 1943. Food
deprivation and hoarding in rats. J. Comp. Psychol.
35:275-295.

Morris, D. 1962. The behaviour of the green acouchi
(M o rocta pratti) with reference to scatter hoarding.
Proc. Zoo . Soc. London. 139:701-732.

148

Morrison, P. R., F. A. Ryser, and A. R. Dawe. 1959.
Studies on physiology of the masked shrew, Sorex
cinereus, Physiol. 2001. 32:256-271.

Mowrer, O. W. 1960. Learning theory and behavior. John
Wiley and Sons, Inc. New York.

Murie, A. 1936. Following fox trials. Univ. Mich. Mus.
of 2001. Misc. Publ. No. 32. 45pp.

Muul, I. 1968. Behavioral and physiological influences on
the distribution of the flying squirrel, Glaucom s
volans. Univ. Mich. Mus. of 2001. Misc. PEBI No. 134.

66pp.

Muul, I. 1970. Day length and food caches. In: American
Museum of Natural History, Field Studies in Natural
History. Van Nostrand. New York. pp. 78-86.

Orr, H. D. 1959. Activity of white-footed mice in relation
to environment. J. Mamm. 40:213-221.

Rice, L. Z. 1972. A study of food hoarding in freely
growing laboratory populations of prairie deermice.
Masters Thesis. The College of William and Mary,
Virginia. 46pp.

Richards, P. W. 1973. The tropical rain forest. Scienti-
fic Amer. 229:58-67.

Ross, S. and W. I. Smith. 1953. The hoarding behavior of
the mouse II. The role of deprivation, satiation, and
stress. J. Genet. Psychol. 82:299-307.

Ross, 8., W. I. Smith, and V. H. Denenberg. 1950. A
preliminary study of individual and group hoarding in
the white rat. J. Genet. Psychol. 77:123-127.

Ross, 8., W. I. Smith, and B. L. Woessner. 1955. Hoarding:
An analysis of epxeriments and trends. J. Genet.
Psychol. 52:307—326.

Scelfo, L. M. and L. R. Hammer. 1969. Stimulus prefer-
ences in hoarding. Psychon. Sci. 17:155-156.

Scheffer, T. H. 1947. Ecological comparisons of the
plains prairie-dog and the Zuni species. Trans.
Kansas Acad. Sci. 49:401-406.

Scott, T. G. 1943. Some food coactions of the northern
plains red fox. Ecol. Monogr. 13:427-479.

149

Sealander, J. A. 1951. Survival of Eeromyscus in relation
to environmental temperature and acclimation at high
and low temperature. Amer. Midl. Nat. 46:257-311.

 

Seton, E. T. 1926. Lives of Game Animals. Vol. II.
Doubleday, Doran, and Co., Inc. New York. 746pp.

Shaw, W. T. 1934. The ability of the giant kangaroo rat
as a harvester and storer of seeds. J. Mamm. 15:
275-286.

Shaw, W. T. 1936. Moisture and its relation to the cone-
storing habit of the western pine squirrel. J. Mamm.
17:337-349.

Shull, A. F. 1907. Habits of the short-tailed shrew.
Amer. Nat. 41:495-522.

Smith, C. C. 1968. The adaptive nature of social organi-
zation in the genus of tree squirrels Tamiasciurus.
Ecol. Mongr. 38:31-63.

Smith, W. I. and E. K. Powerll. 1955. The role of
emotionality in hoarding. Behav., Leiden. 8:57-62.

Smith, W. I. and S. Ross. 1950a. The effect of continued
food deprivation on hoarding in the albino rat. J.
Genet. Psychol. 77:117-121.

Smith, W. I. and S. Ross. 1950b. Hoarding behavior in
the golden hamster (Mesocricetus auratus auratus). J.
Genet. Psychol. 77: - .

Smith, W. I. and 8. Ross. 1953a. The hoarding behavior
of the mouse: I. The role of previous feeding eXperi-
ence. J. Genet. Psychol. 82:279-297.

Smith, W. I. and 8. Ross. 1953b. The hoarding behavior of
the mouse: III. The storing of "non-relevant" material.
J. Genet. Psychol. 82:309-316.

Snead, E. and G. O. Hendrickson. 1942. Food habits of the
badger in Iowa. J. Mamm. 23:380-391.

Sokal, R. R. and F. J. Rohlf. 1969. Biometry. The
principles and practice of statistics in biological
research. W. H. Freeman and Co. San Francisco.
776pp.

Stamm, J. S. 1953. Effects of cortical lesions on
established hoarding activity in rats. J. Comp.
Physiol. Psychol. 46:299-304.

150

Stamm, J. S. 1954a. Control of hoarding activity in rats
by the median cerebral cortex. J. Comp. Physiol
Psychol. 47:21-27.

Stamm, J. S. 1954b. Genetics of hoarding: I. Hoarding
differences between homozygous strains of rats. J.
Comp. Physiol. Psychol. 47:157-161.

Stamm, J. S. 1956. Genetics of hoarding: II. Hoarding
behavior of hybrid and backcrossed strains of rats.
J. Comp. Physiol. Psychol. 49:349-352.

Steel, R. G. D. and J. D. Torrie. 1960. Principles and
procedures of statistics. McGraw-Hill. New York.
481pp.

Stellar, E. 1943. The effect of epinephrine, insulin, and
glucose upon hoarding in rats. J. Comp. Psychol. 36:
21-32.

Stellar, E. 1951. The effects of experimental alterations
on metabolism on the hoarding behavior in the rat. J.
Comp. Physiol. Psychol. 44:290-299.

Stellar, E. and C. T. Morgan. 1943. The roles of experi-
ence and deprivation in the onset of hoarding behavior
in the rat. J. Comp. Psychol. 36:47-55.

Stinson, R. H. 1952. Effects of some environmental
factors on the behavior of Peromyscgg. Unpubl. Ph.D.
Dissertation, Univ. Toronto.

Terman, C. R. 1974. Behavioral factors associated with
cessation of growth of laboratory populations of
prairie deermice. Res. Pop. Ecol. 15:138-147.

Thorpe, W. H. 1963. Learning and instinct in animals,
2nd ed. Methuen. London. 493pp.

Tigner, J. C. and R. 8. Wallace. 1972. Hoarding of food
and non-food items in blind, anosmic, and intact
albino rats. Physiol. Behav. 8:943-948.

Tileston, J. V. and R. R. Lechleitner. 1966. Some com-
parisons of the black-tailed and white-tailed prairie
dogs in north-central Colorado. Amer. Midl. Nat.
75:292-316.

Vanderwolf, C. 1967. Effects of experimental diencephalic
demage on food hoarding and shock avoidance behavior
in the rat. Physiol. Behav. 2:399-402.

151

Viek, P. and G. A. Miller. 1944. The cage as a factor in
hoarding. J. Comp. Psychol. 37:203-210.

Waddell, D. 1951. Hoarding behavior in the golden hamster.
J. Comp. Physiol. Psychol. 44:383-388.

Wallace, R. J. and J. C. Tigner. 1972. Effect of cortical
and hippocampel lesions on hoarding behavior in the
albino rat. Physiol. Behav. 8:937—942.

Winer, B. J. 1962. Statistical principles in experimental
design. McGraw-Hill. New York. 672pp.

Wishart, T., L. Brohman, and G. Mogenson. 1969. Effects
of lesions of the hippocampus and septum on hoarding
behaviour. Anim. Behav. 17:781-784.

Wolfe, J. B. 1939. An exploratory study of food-storing
in rats. J. Comp. Psychol. 28:97-108.

Zubek, J. P. 1951. Effects of cortical lesions upon the
hoarding behavior in rats. J. Comp. Physiol. Psychol.
44:310-319.