a...
.1..s....lr
1.... t...

4. 1! r

. .
,5...nn.:.:u..
0x:

I .

5.1.3.. €53;
. x E.

...

.1

12.....Iaé

1.).

r. (11....
; 1.2... _
x J... r.
.V
. 1
.3192... _..|.. 1: vi .0
3.1;}... I. .. .u
I .ts::4 {5.9531 Lttui’lr
ara‘yt- 0’ 34¢.tx
.....¢a-§.v6 .
.a I

ill,
.4... .
$53.
.4 {U 2-.I:~‘.c ‘y‘ \i
.. 19"....2 w |._
i 39...... . ‘
1p f!:3.:¢
. .1 . .

.-
.1134...
.3131:

r , I.
i~u<53§5 sixtiallaix’n
1. , . .. .
e. 3:: .. s‘.\n......uvu.

you! .. . . . .

3.: . . ‘ .

. , 35'331
. :t-.oz......tt:.9
1!

AK .
x 3:.
5b.!!! , V
. :2}.
'03: "Alll kibl’l ll“\-
93", 391..""0173'I , ‘0‘- Ql 5.1;)
7:1 3...).
51..
:n‘ a .ntl‘ithifia
)2 2:)! n) . 11.7.?) 100.14.
.7... 352..
32925.
.slia):a?e
0.3%.!!!)

\
c3.
:3...
X: I...

I
. ‘. {fr-Ar .
.Z..vs\. l I . ,
-
at: ‘1... . ,
?1le-! 49. x<11 . .
‘o~vl...l7\... .3 x:
fict; 1 It. . . . ,
1: t a .rfl
nv. gaaxata).54::a|~l a
1.4.4.. .3. tvarii 3‘13]:
1?. 1- :
1:14. VI :3.
:5..«Jl.: i
«a _ .L
1:13.
.03.}.
1132.}
33.311)»
tilt

: .c ... 1»
>.«..Il§..i.~ l
a).v.r‘.vbo4)9 I.

n.‘5.a:wt«aal

. \. ‘x . .
.3 ‘ y
a- .1}
3;.7....:..
a: 1.3.3:; 1.
. , a a .3. . . ;
. (3?
3:3.)13uv)
2:3!3; .
fl“... 1‘.
u if; (if, 15933:}...1‘?! v
.114 133.1435)... fio.1$§3)5.|v§.
- u a ( rtnst.‘ ) . VIAAitti 9'!
‘53:!!1‘:
5 . 1311:.2.
)sl ‘ . 951....
1! {2.3

i .v;.. . l

,1: I‘J.!;:.V..1§a\ ; u

1.54:). , 533.11... .3: .1 1.»)! t x
143‘ I )«i l. :‘«x)‘r

x311..-

.fi.» ‘ :v.» 11.;
53-11,..0)‘ tint.

I~»17:)7¢u: .31‘ I
..I...:.|

 

 

,dla ‘
3".3‘1V9 i Hxx . ‘ , .
. . , .133...‘ 55., 31.15 ‘..,-Lf.|,.t
I ‘ . is.\;lt\o§:2luié f
. 1:53.(135)339531

I‘uXI.

a re
.o.«4.#.£3.11:ev . .‘f ‘
i.‘¢4|lz I v V . ).,).I
~11. L ‘ , 1 y I
{not‘alvlltrtz rx .1
1.1. ‘
‘ . . \ u 2".) \l'Y"
. :2... Elli)
10:0“.3.) 0.08.}. 7:! .
{s3rnto-v.on‘ . Rioti‘ologv‘ib-n;9v.~uxx l . i: .t . ..
3 oti!;3-\.I~O:|b~filotl£\tu vb“«..\: x ‘
‘ i‘rz I’l-‘rlfi I).§.l‘tt.- ‘»I1“
1:1?!» :19. Itinl:
in.) I: 2

.1. VA
.Wflfinhrmm J43... .1... firm»
nw. u .. hi rag .
. ‘ . s

.1Il.v:

vr ll. . . ,
‘3‘“..{31vvn4 ‘ , .
IV.“ , i‘fit)
zl‘li;
‘5! in a . . ‘
lt‘tw!ut‘u . \t

XI.“
0 txl‘:
.‘rnO‘t-Ilu‘ ‘

vu‘icg

v»bulvl

 

THES'F

This is to certifg that the

thesis entitled
GENETICS, EXPERIENCE AND STRATEGY AS FACTORS IN

THE FOOD HABITS 0F PEROMYSCUS:
USE OF OLFACTION

presented bg
LEE C. DRICKAMER

has been accepted towards fulfillment
of the requirements for

in;— degree in M

 

Date M8,! 20, 1920

0-169 ‘3

2
g
p
b

n I” ‘ “\ 4"
3. .l. .‘lJR q R y
Michigan Sta :
Unzversxtv

é, .

 

 

ABSTRACT
GENETICS, EXPERIENCE AND STRATEGY AS FACTORS IN

THE FOOD HABITS OF PEROMYSCUS:
USE OF OLFACTION

BY

Lee Co Drickamer

Feeding behavior in Peromyscus maniculatus bairdi and
_g. leucgpus noveboracensis was studied by examining the
extent to which food habits could be varied by genetic and
eXperiential manipulation and by measuring the strategy mice
use in locating and returning to food, Since these mice use
olfaction to locate food, three essential oils were used as
odor stimuli with laboratory chow to provide three different
food-odor combinationso

Species (genetic) preferences of the mice were tested
among three taxa, two stocks of different breeding histories
(wild caught and domestic), and several age groups. When the
mice were individually presented with all three food-odor
combinations, the percentage of the diet consumed from each
combination revealed no group differences. All experimental
stocks showed a strong pine preference, which provided a
baseline for interpretation of the experience and strategy

studies.

Lee C. Drickamer

A series of control eXperiments determined that the mice
were not: (1) discriminating odor intensities; (2) using a
dish preference; (3) discriminating soiled from unsoiled
food; and (4) preferring pine because of pine shavings in
their rearing cages. Also, there were no seasonal shifts in
the preference pattern. Use of a position habit was pre-
cluded by rotating the positions of the dishes in all experi—
ments.

Modifiability of the food-odor preferences was examined
by providing young and adult mice with olfactory experience
that would affect their subsequent preferences. Mice were
conditioned for two weeks to associate one of the three odors
with food. Control mice were conditioned with laboratory
chow, but no odor stimuli. Later preferences were determined
by: (1) an appetitive test which measured the number of
sticks a mouse gnawed through to get its preferred food—odor
combination, and (2) a consummatory test which measured the
amount of food consumed at each of the three food-odor combi—
nations. Mice were tested both immediately and one month
after the conditioning. Throughout the experiment no mouse
was ever tested twice. Young g, m, bairdi and young and adult
2, leucoEus were significantly affected by the olfactory
conditioning, but adult g. m, bairdi were not. Groups that
showed a significant experience effect immediately also

showed a significant effect one month later.

Lee C. Drickamer

Feeding_strategies used by young and adult mice were
tested for 26 days by using an apparatus which automatically
monitored feeding activity at three food hOppers containing
the three food-odor combinations. Reaction to a novel food—
odor stimulus was tested by replacing the anise odored food
with a sassafrass food—odor combination for days 21—26.
Analyses of the strategy patterns showed that young mice
adOpted a more conservative strategy than adults during the
first ten days (1—10), but all groups used the same strategy
for days 11-20. Young 2, m, bairdi avoided the novel food-
odor combination, while the other test groups all consumed
more than half of their diet from the new source. Both spec-
ies shifted from a position habit (days 1—10) to a following
response (days 11-20), using an odor cue to locate food, but
the shift was more dramatic for g. m, bairdi. Adult g,
lGUCOEUS switched feeding sites more frequently than any other
group.

The principle conclusion was that g, leucoQus showed
more flexibility in feeding behavior than 3, m, bairdi.
Although young mice of both Species showed a pine odor pref—
erence and both species were affected by early olfactory
experience, the avoidance of a novel food stimulus by g. m,
bairdi indicated that it was already restricted in its feed-
ing behavior. Also, at maturity g, leuc0pus were affected

by olfactory experience, but not P. m, bairdi, and the

Lee C. Drickamer

g, leucogus switched feeding sites more frequently than the
other test groups. Food habits were discussed in relation

to the distributions of these two Species of mice.

GENETICS, EXPERIENCE AND STRATEGY AS FACTORS IN
THE FOOD HABITS OF PEROMYSCUS:

USE OF OLFACTION

BY

Lee C) Drickamer

A THESIS

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

DOCTOR OF PHILOSOPHY
Department of Zoology

1970

 

ACKNOWLEDGEMENTS

My most sincere appreciation is expressed to Dr. John
A. King, my major professor, for encouragement during this
research effort and for carefully editing this manuscript.
I also thank Drs. Martin Balaban, William COOper and Glenn
I. Hatton for their thought provoking ideas regarding my
research, and for their help in preparing this manuscript.
I thank my wife, Lucy, for help in preparing the figures and
for her continued support, and Mrs. B. Henderson for her many
favors. Without the encouragement of my parents Dr. and Mrs.
H. G. Drickamer my graduate studies and this thesis would not
have been possible. Lastly, I would like to express my ap-
preciation to my fellow graduate students in the Biology
Research Center for their assistance and criticism throughout
my research and the preparation of this manuscript. Two
sources of financial support are acknowledged: NDEA Title IV
predoctoral fellowship No. 67-05706 to the author, and NIH
training grant No. GM-Ol751 to the Animal Behavior Program,

Department of Zoology, Michigan State University.

ii

TABLE OF CONTENTS

Page
LIST OF TABLES . . . . . . . . . . . . . . . . . . . . v

LIST OF FIGURES. . . . . . . . . . . . . . . . . . . . viii

 

INTRODUCTION . . . . . . . . . . . . . . . . . . . . .

Genetics. . . . . . . . . . . . . . . . . . . . .
Experience. . . . . . . . . . . . . . . . . . . .
.Strategy. . . . . . . . . . . . . . . . . . . . .

GPPCN |—\

\1

LITERATURE REVIEW. . . . . . . . . . . . . . . . . . .

Food Habits and Distribution. . . . . . . . . . . 7

Distribution . . . . . . . . . . . . . . . . 7

Food . . . . . . . . . . . . . . . . . . . . 1O
Sensory Cues. . . . . . . . . . . . . . . . . . . 12
Genetics. . . . . . . . . . . . . . . . . . . . . 12
Strategy. . . . . . . . . . . . . . . . . . . . . 14
EXperience. . . . . . . . . . . . . . . . . . . . 17

GENERAL METHODS. . . . . . . . . . . . . . . . . . . . 21

EXperimental Subjects . . . . . . . . . . . . . . 21
Maintenance . . . . . . . . . . . . . . . . . . . 21
Odor Stimuli. . . . . . . . . . . . . . . . . . . 22
-Dependent Variable. . . . . . . . . . . . . . . . 25

GENETIC FACTORS. . . . . . . . . . . . . . . . . . . . 25

Purposes. . . . . . . . . . . . . . . . . . . . . 25
Test Subjects . . . . . . . . . . . . . . . . . . 25
Procedure . . . . . . . . . . . . . . . . . . . . 26
Analysis. . . . . . . . . . . . . . . . . . . . . 29
Results . . . . . . . . . . . . . . . . . . . . . 50

iii

 

 

TABLE OF CONTENTS-~continued Page

CONTROL EXPERIMENTS. . . . . . . . . . . . . . . . . . 56

Control for Odor Intensity. . . . . . . . . . . . 56
Control for Dish Preference . . . . . . . . . . . 59
Seasonal Control. . . . . . . . . . . . . . . . . 4O
Shavings EffeCt . . . . . . . . . . . . . . . . . 42
Soiled Versus Unsoiled Food . . . . . . . . . . . 44

EXPERIENCE . . . . . . . . . . . . . . . . . . . . . . 47

Purposes. . . . . . . . . . . . . . . . . . . . . 47
Test Subjects . . . . . . . . . . . . . . . . . . 47
Procedure . . . . . . . . . . . . . . . . . . . . 48
Results . . . . . . . . . . . . . . . . . . . . . 55
Appetitive Test. . . . . . . . . . . . . . . 55
Consummatory Test. . . . . . . . . . . . . . 59

STRATEGY.......... . . . . . . . ... . . . . . . . . . 72
Purposes. . . . . . . . . . . . . . . . . . . . . 72
Test Subjects . . . . . . . . . . . . . . . . . . 72
Apparatus . . . . . . . . . . . . . . . . . . . . 72

Procedure . . . . . . . . . . . . . . . . . . . . 75
Results . . . . . . . . . . . . . . . . . . . . . 74

DISCUSSION . . . . . . . . . . . . . . . . . . . . . . 94
Genetics. . . . . . . . . . . . . . . . . . . . . 95
Experience. . . . . . . . . . . . . . . . . . . . 98
Strategy. . . . . . . . . . . . . . . . . . . . . 101

SUMMARY. . . . . . . . . . . . . . . . . . . . . . . . 106

BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . 109

iv

 

TABLE

10.

LIST OF TABLES

. Mean percentage (i,1 S. E.) of the diet consumed

from each of the three food-odor combinations. .

Summary of the results of the intensity control
test. Mean percentage (i.1 S. E.) of the diet
eaten from each intensity. . . . . . . . . . . .

Summary of the dish preference control test.

The ratio presented is the number out of four
animals tested in each group showing a signifi—
cant dish preference . . . . . . . . . . . . . .

Results of the seasonal control test. Mean per—
centage (i_1 S. E.) of the diet eaten from each
food-odor combination. . . . . . . . . . . . . .

Results of the test for the effect of the pine
shavings on food odor preferences. Mean per-
centage (i.1 S. E.) of the diet eaten from each
food-odor combination. . . . . . . . . . . . . .

Results of the test for soiled versus unsoiled
food. Mean percentage (i,1 S. E.) of the diet
eaten from each food type. . . . . . . . . . . .

Design for the study of the effect of eXperience
on odor preferences. The same design was re-
peated for each of the four test groups. . . . .

Summary of the ages at which the mice were con-
ditioned and tested. Age given is that at the
start of the conditioning or test period . . . .

Mean ages (1.1 S. E.) of adult mice used in each
test group in the eXperience study . . . . . . .

Frequency of deaths during the olfactory condi-
tioning process. . . . . . . . . . . . . . . . .

Page

51

58

41

45

45

45

49

50

50

54

LIST

11.

12.

15.

14.

15.

16.

17.

18.

19.

20.

21.

OF TABLES-—continued

Results of the appetitive test. The ratio is

the number of mice out of six in each test group
chewing through the sticks in front of the con—
ditioned food-odor combination (A=anise;
W=wintergreen; P=pine). . . . . . . . . . . . . .

Results of the control mice in the appetitive

test. Early and late tests combined. The ratio
is the number out of 12 mice in each group chew-
ing through each food-odor combination, or dying.
(A=anise; W=wintergreen; P=pine). . . . . . . . .

Number of animals out of six that ate a greater
percentage of their diet from the conditioned

odor than controls during day one of the consumma-
tory test. (A=anise; W=wintergreen; P=pine). . .

Number of animals out of six that ate a greater
percentage of their diet from the conditioned
odor than controls during day four of the consum-
matory test. (A=anise; W=wintergreen; P=pine). .

Summary of the effects of early experience. Plus
(+) indicates a significant overall effect and a
minus (-) indicates no significant effect . . . .

Mean number (i_1 S. E.) of days, out of 10, that
all three food-odor combinations were visited
during the initial activity bout. . . . . . . . .

Analysis of variance of the food-odor combina—
tions visited during the initial activity bout. .

Mean ratio (i_1 S. E.) of position reSponses to
following responses . . . . . . . . . . . . . . .

Three way analysis of variance with one measure
repeated for the ratio of position responses to
following responses . . . . . . . . . . . . . . .

Mean number (i.1 S. E.) of food—odor combinations
visited per hour of activity per mouse per day. .

.Three way analysis of variance with one measure

repeated on the food-odor combinations visited
per hour. . . . . . . . . . . . . . . . . . . . .

vi

Page

56

58

68

68

71

78

80

82

85

85

86

 

LIST

22.

25.

24.

25.

26.

27.

28.

OF TABLES--continued Page

Duncan's New Multiple Range Test. Those means
not subtended by the same line are significant-
ly different (p <T.05). . . . . . . . . . . . . . 87

tMean number (i,1 S. E.) of switches between food-

odor combinations per day . . . . . . . . ... . . 89

.Three-way analysis of variance of the number of

switches between food-odor combinations per day . 90

Duncan's New Multiple Range Test. Those means
not subtended by the same line are significantly
different (p <1.05) . . . . . . . . . . . . . . . 92

Mean percentages (i_1 S. E.) of the diet eaten
from the novel food-odor combination. . . . . . . 92

Two—way analysis of variance of the percentage of
the diet eaten from the novel food—odor combina—
tion during all six test days . . . . . . . . . . 95

Duncan's New Multiple Range Test. Those means

not subtended by the same line are significantly
different (p <:.05) . . . . . . . . . . . . . . . 95

vii

LIST OF FIGURES

FIGURE

Schematic drawing of the double plastic cage
(top view with wire lid removed). One inch in
the diagram equals two inches of the actual
cage and equipment. . . . . . . . . . . . . . .

Mean percentage (: 1 S. E.) of the diet eaten
from each of the three food-odor combinations
(A=adult; Y=young; WC=wild caught; F1WC=first
generation laboratory reared; Dom.=domestic). .

rMean percentage (1.1 S. E.) of the diet con-

sumed from each food-odor combination for day
one of the test, both immediately and one month
after conditioning in young g, m, bairdi.
(Conditioning odors indicated as: A=anise;
W=wintergreen; P=pine; C=control) . . . . . . .

Mean percentage (i,1 S. E.) of the diet con-
sumed from each food-odor combination for day
one of the test, both immediately and one month
after conditioning in adult g, m, bairdi.
(Conditioning odors indicated as: A=anise;
W=wintergreen; P=pine; C=control) . . . . . . .

Mean percentage (i.1 S. E.) of the diet con-
sumed from each food-odor combination for day
one of the test, both immediately and one month
after conditioning in young g, leucopus.
(Conditioning odors indicated as: A=anise;
W=wintergreen; P=pine; C=control) . . . . . . .

Mean percentage (i.1 S. E.) of the diet con-
sumed from each food-odor combination for day
one of the test, both immediately and one month
after conditioning in adult g, leucopus.
(Conditioning odors indicated as: A=anise;
W=wintergreen; P=pine; C=control) . . . . . . .

Mean percentage (1 1 S. E.) of the diet cons-

sumed from each of the three food-odor combina—
tions (A=anise; W=wintergreen; P=pine). . . . .

viii

Page

28

55

61

65

65

67

76

 

INTRODUCTION

The thesis presented here is that variation in food
habits among species of small mammals are attributable to
differences in the inherited capacity of each Species to be
affected by ontogenetic eXperiences. Variations in food
habits are the differences in dietary preferences of each
species. Variation is also used here as the relative fixity
of food preferences between Species. Finally, variation
refers to the degree to which food preferences can be modi-
fied by prior eXperience in each Species. Ontogenetic
eXperienceS are in the form of the visual, gustatory and ol—
factory cues associated with foodS encountered during deve10p-
ment. The variety of food cues encountered may influence the
initiation and maintenance of species differences in flexi-
bility of food habits. For example, if an individual of a
less variable Species encounters a restricted range of food
cues during develOpment its food preferences will be less
flexible than individuals of the same Species encountering a
broad range of food cues, or a different and more variable
Species encountering a restricted or broad range of food
cues during deve10pment. This is a verbal statement of a

cross-classified design with two Species (variable and

non-variable) and two ranges of food cues (restricted and
broad). This dissertation was primarily concerned with
Species differences using restricted sets of food—cues.

I also tested the strategy used by small mammals to locate
food as a measure of the flexibility of feeding behavior.

My eventual goal is to investigate the role of food
habits in the distribution of small mammals. Animals are
not distributed randomly in Space. Among the factors which
have been used to explain the departures from randomness
are climatic parameters (Dice,.1922; Johnson, 1926), social
influences (Sheppe, 1961, 1967), vegetation type (Wecker,
1965; Harris, 1952), and food habits (Drickamer, 1970).
Species food preferences, or those modified by eXperience
could limit small mammal distributions to the habitats where
those foods occur. The flexibility in food habits exhibited
by a species may also be a predictor of the number of types
of habitats it will occupy.

The animals used in these studies were mice of the genus

Peromyscus: Peromyscus maniculatus bairdi and g, leucopus

 

noveboracensis. These two Species use primarily olfaction in

 

finding food (Holling, 1955, 1958; Howard and Cole, 1967;
Howard et al., 1968). The odor stimuli used in these studies
were three essential oils placed on cotton in the bottom of
food dishes filled with laboratory chow. The assumption be—

ing made here is that the three odors in combination with the

 

 

 

 

 

 

 

 

 

food are perceived by the mice as three different foods with

identical nutritional values.
Genetics

Genetic bases for food habits have been claimed by
Barnett and Spencer (1955), and Young (1955). These authors
report that rats have Specific affinities and aversions for
various foods and flavors. Their conclusions are based on
observational evidence and not on studies designed to investi—
gate the specific role of genetics in determining food pref-
erences. EXperiments have Shown that there is a genetic
component to the saccharin preferences of the rat (Nachman,
1959), and that the alcohol preferences of laboratory Mug
are inherited (McClearn and Rodgers, 1961). In pilot work
for the current study, 2, m, bairdi and g, leucopus, born in
the laboratory to wild caught parents, Showed the same
preference for seeds from their native field or forest habi-
tat as did their parents (Drickamer, 1970, unpublished).
This suggests a genetic component in determining food habits

in Peromyscus.

 

In the current study I used olfactory preferences in
the place of natural food preferences since the diets com—
prised of a few natural foods resulted in death and lack of
reproduction of the mice. It was necessary to establish a
baseline for the olfactory preferences of each eXperimental

stock. Since both young and adult mice of each species were

used, the develOpment and stability of the odor preferences
were tested by examining both cross-sectional and longitudinal
groups (King, 1969). These groups will distinguish whether
young mice develOp their preferences as they mature, or
whether the preferences exist from the time they first begin
to eat food. No differences in food-odor preferences between
cross-sectional and longitudinal test groups would indicate

a genetic component.

In addition to Species and age differences, other stocks
were used to test the hypothesis concerning the genetic basis
for the olfactory preferences. One of the factors contribut—
ing to domestication in g, m, bairdi has been relaxed selection
pressure in the laboratory (Wecker, 1965; Price, 1967).
Testing the odor preferences of wild caught and domestic
stocks of g, m, bairdi and g, m, gracilis could point to any
changes that have occurred in the laboratory due to relaxed
selection pressures for the olfactory preferences.

The first experiment tested the following Specific
hypotheses:

(1) All of the different Species, age, and genetic
stocks will exhibit the same olfactory preferences.

(2) Longitudinal and cross-sectional developmental

groups of young 3, m, bairdi and g, leucogus will
exhibit the same food-odor preferences.

EXperience

 

The role of early eXperience in feeding habits is a

controversial topic, and the evidence to date is equivocal.

 

On the one hand, early eXperience is claimed to be the sole
determinant of feeding preferences (Kuo, 1967). Chow dogs,
cats and mynah birds were fed on different early regimes and
then tested for up to six months. Animals given restricted
early feeding eXperience would not eat new foods, while
animals given a variety of foods at an early age would accept
new foods (Kuo, 1967). Pre-feeding rats just prior to testing
their food preferences altered their choice of foods (Young,
1940). Positive early eXperience effects have also been
reported for other groups of vertebrates including snakes,
turtles, and birds (Burghardt, 1967a, 1967b; Burghardt and
Hess, 1966; Rabinovitch, 1966, 1969).

In contrast, Bronson (1966) was unable to demonstrate any

early experience effect in white rats. He fed young rats
plain mynah bird pellets or the same pellets encased in a gel-
atin capsule. When the rats were tested later the early
feeding experience had no effect on their food preference.
In much of the work with early eXperience it has been assumed
that young animals were mOre modifiable than adults (see
Beach and Jaynes, 1954).

The Specific hypothesis tested in this experiment was:

(1) Young 2, m, bairdi and g, leucopus conditioned to

eat at one of three food-odor combinations for two
weeks will prefer that food-odor combination in
subsequent tests, while adult mice of both Species

given similar olfactory conditioning will not Show
any experience effect.

Strategy

The strategy that a mouse adopts in locating and return-
ing to feeding places can affect the types of foods eaten.
Mice could adOpt a completely random strategy. In fact, they
do not (Zimmerman, 1965; Whitaker, 1966; Holling, 1965). What
are the alternatives?

The mice could use a position habit. This would imply
that the mouse returns repeatedly to the same feeding site.
Use of this strategy would restrict the mouse to whatever
food(s) were available at that site. Another strategy would
be for the mouse to associate various odors (or other cues)
with foods, enabling the mouse to locate other similar food
items on the basis of the food-cue association. This type of
strategy is analogous to the visual search images reported
for birds preying on various insects (L. Tingergen, 1960;
Mook et al., 1960; N. Tinbergen, 1967). In Peromyscus the
search image would be olfactory rather than visual.

This eXperiment examined some of the possible strategies

that Peromyscus could adopt in feeding at the three food-

 

odor combinations. The following experimental hypotheses
were tested:
(1) g, m, bairdi and g, leuc0pus will both use a follow-
ing response as opposed to a position habit in

locating and returning to preferred food sources.

(2) Young mice will adOpt a different feeding strategy
than adult mice.

   

LITERATURE REVIEW

The literature on food habits and factors affecting
feeding behavior in Peromyscus will be reviewed by examining:
(1) food habits and distribution; (2) the sensory modalities
used in the location and selection of food; (5) the genetic
bases for food habits; (4) the effects of eXperience on food
habits and preferences; and (5) strategies used in food

location.

.Food Habits aneristribution

Distribution

The study of factors affecting the distributions of
small mammals has evolved from an examination of environmental
parameters as constraints on the animal, to considerations of
distributions as a function of the animal. The habitat segre—
gation of two Species of deermice, Peromyscus maniculatus
bairdi (grassland inhabitant) and P, leuc0pus noveboracensis
(woodland inhabitant) has been studied extensively.

One group of hypotheses tested have attempted to find
relationships between environmental parameters such as climate,
vegetation and soil type, and the distributions of small

mammals (Dice, 1922; Johnson, 1926; Hardy, 1945; Verts, 1957).

 

This type of approach has been termed habitat correlation
(Klopfer,1969). These studies provided predictive associ-
ations for habitat characteristics and the presence of a
given Species of small mammal, but have not provided evidence
that the animals are restricted to particular habitats. Both
2, m, bairdi and g, leucopus are capable of living in either
grassland or woodland habitat (Dice, 1922).

Methods of measurement of environmental parameters have
permitted refined correlations of habitat characteristics
and animal distributions, but the small rodent has been con-
sidered a machine receiving input from the environment. The
systems of the animal, in a predictable machine-like manner,
use the incoming sensations in directing the animal's course
of behavior, and in controlling its distribution.

Although environmental factors may provide the limits
for an animals distribution, a new concept emerged; the exist-
ence of psychological factors affeCting its_distribution within
those limits (Lack, 1955, 1957; Klopfer, 1969). According to
this hypothesis the animal receives incoming stimuli from
the environment (sensation), interprets the stimuli (percep-
tion), and makes an active choice of habitat. Psychological
factors affecting habitat selection in mice include:

(1) vegetation type and form; (2) amount of cover; (5) nest
sites; (4) social influences; and (5) food habits. In some
Species, such as Microtus (a herbivore) hunger may override

a preference for a particular habitat location, leading to

 

movement to a new habitat with more food. In Peromyscus

(an omnivore) foods appear to be so abundant in the habitats
that hunger is probably not a factor driving the mice into
new habitats in search of food. Only the general form and
types of vegetation have been examined in experimental
studies of habitat selection in small rodents. .SubSpecieS

of Peromyscus selected an artificial habitat which most nearly

 

resembled the vegetation form in their native environment
(Harris, 1952). In another study, Microtus selected an arti—
ficial habitat Similar to their normal one, but g. leucoEuS
failed to select between the two sets of grassland cues
presented (Wirtz and Pearson, 1960). This raises questions
concerning the validity of the two models of grassland cues
used to test a forest SpeCies.

Psychological factors in distribution are in part
determined by inheritance (Wecker, 1965). That this genetic
base may be modified by eXperience is indicated by the strong
preferences grass-reared g, m, bairdi Showed for a grassland
over a woodland habitat, whereas 3. m. bairdi reared in a
woodland enclosure did not prefer that habitat (Wecker, 1965).
These results may be explained by a genetic predisposition
(an inclination before the actual choice) to select a grass-
land habitat (Wecker, 1965, 1964). The early grassland
eXperience reinforced this prediSposition. Wecker also sug-
gested an apparent relaxation of selection pressures in a

stock of g, m, bairdi, maintained in the laboratory for 15-20

   

10

generations, as an explanation for their failure to select
the grassland habitat over the forest. It appears that for
vegetation type there is an interaction of a genetic pre-
disposition to respond to certain cues and the early eXperi—
ence of the mouse. In a similar manner, the food habits of
grassland and woodland Peromyscus could have a genetic basis

reinforced by eXperience.

£224

The correlation between the availability and abundance
of foods in the habitat and the diets of small rodents has
been determined by gut analyses (Hamilton, 1941; Whitaker,
1965, 1966; Thompson, 1965). Although predictions can be
made from these correlations, eXperimentS Should: (1) ex—
amine the feeding behavior per se, or (2) study the relation-
ship between diet and choice of habitat.

The diet of Microtus was comprised of a very few foods
selected from those that were abundant in the environment
(Zimmerman, 1965). Whitaker (1965, 1966) found that

Peromyscus and Mus were eating only a selected few of the

 

foods available in their habitats. This suggests orderly
processes underlying the feeding habits of small rodents.
These selective feeding behaviors, or strategies, adopted by
the mice, cannot be determined from gut analyses alone, but
require systematic observation of the feeding behavior and
pattern. Further evidence for an orderly feeding process

comes from Holling (1965). P, leuc0pus varied in their

   
 

11

selection of sawfly prepUpae depending upon the alternate
foods available and on the relative palatabilities of these
foods.

Selective feeding behavior may eXplain, in part, the
observed distributions of the mice. If mice learn to eat
the foods available in their natal habitat through eXperience,
they will remain in that habitat where those foods are avail—
able. Two hypotheses may be tested. One is that eXperience
can affect the food habits of the mice. This I am testing
in the present thesis. The other hypothesis is that adult
wild caught mice from different habitats will select foods
from their natural diets when given a choice among foods from
a variety of habitats. This hypothesis has been tested
(Drickamer, 1970). g, m, bairdi and g, leucogus preferred
seeds from their own graSSland and forest habitats respec-
tively in a choice test.

Studies of food habits and distributions of small
mammals have lead to the following conclusions: (1) distri-
bution is not only a function of the environmental parameters,
but also the animal actively selects its habitat; (2) studies
of habitat selection indicate that inherited prediSpositions
to reSpond to certain environmental cues may interact with
eXperience in affecting the animals' choice of habitat; and
(5) mice choose their diet selectively from the foods that are

abundant in their environment.

 

12
Sensory‘Cues

Olfaction in deermice was tested by their ability to
locate and dig up buried‘food under conditions of low light
and total darkness (Howard et al., 1968). Mice were able to
locate preferred foods and dig them up with equal facility
under either lighting condition, indicating that olfaction
could be used exclusively. Peromyscus also locate, differ—
entiate and dig Up prepupae of the EurOpean sawfly (Neodiprion

sestif§£)using olfaction (Holling, 1955, 1958). Healthy pre—

 

pupae were clearly selected over dead or parasitized animals.
Larger prepupae could also be differentiated from smaller
ones. Holling concluded, through a series of elimination
experiments, that the mice were using olfaction in locating
the buried food and in digging up the healthy prepupae. The
eating phase of the feeding sequence involved both the

gustatory and olfactory modalities.

w

Two approaches have been taken in the investigation of
genetic factors and food habits: (1) measures of Specific
preferences, particularly in young animals, for foods that
have not been previously encountered; and (2) genetic
manipulation studies involving selection and breeding eXperi—
ments.

Young rats with no prior eXperience exhibited preferences

for special food items when presented with a variety of foods

 

15

(Harlow, 1952). In similar studies Barnett and Spencer
(1955a, 1955b) and Young (1952) have reported preferences
and aversions for a variety of foods and tastes in the rat.
None of these studies have eliminated prior feeding or
maternal influences as experiential factors contributing to
the observed preferences. Genetic bases for food habits
have also been demonstrated in other organisms. .Newly
hatched snakes Showed preferences for foods from the natural
environment of their parents (Dix, 1968). Species differ—
ences in the amount of tongue-flicking toward animal skin
extracts by newborn snakes also correlated with adult food
preferences for the Species (Burghardt, 1967a). .For

Peromyscus it has been Shown that mice born in the laboratory

 

will select seeds from the natural habitat of their parents
when given a choice among a variety of seeds (Drickamer, 1970,
unpublished). Pilot work for the present studies Showed that
.3. m, bairdi and g. leucopus preferred pine odored food to
anise or Wintergreen, and that experience with an odor other
than pine would not completely eliminate the Species prefer-
ence for pine.

Genetic manipulation studies have been conducted with
taste preference. A genetic basis for taste preference in
rats has been demonstrated by successfully selecting two
lines, one for saccharin preference and the other for a lack

of preference (Nachman, 1959). McClearn and Rodgers (1961)

demonstrated, through the use of inter—strain crosses, that

 

 

 

 

 

 

 

 

 

 

 

 

14

alcohol preference was inherited in laboratorvaus musculus.

 

The genetic manipulation method is dependent upon observ—
able behavioral differences which may then be selected or

cross-bred to determine their genetic bases.

Strategy

The hypotheses proposed in my introduction suggest two
areas for reviewing the literature on feeding strategies:
(1) the alternative strategies which could be employed by
the mice, and (2) the changes in the strategy that occur when
a novel food source is introduced.

Data from field studies of food habits in Peromyscus
suggest that the mice are selecting their diets from the
food items that are abundant in the environment (Whitaker,
1965, 1966; Hamilton, 1941). Gut analyses Show that the bulk
of the diet of these mice is comprised of a very limited
number of food items (non—random feeding). Thus the diet of
the mouse reflects both the abundance of various food items
in the habitat, and a degree of selectivity on the part of
the mouse in choosing its diet. Zimmerman (1965) showed
that Microtus exhibited selectivity, even among the very
abundant food plants in their surroundings. What strategy(s)
were the mice using to select their diet from all the foods
available?

The use of a position habit is one alternative. Rats

will return to the same location repeatedly for food and may

 

 

15

fail to locate the food if its location is Shifted (Young,
1958, 1945; Barnett, 1965). The use of a position habit by
a small mammal in nature would restrict its food sources to
those food items that were available in the immediate area
to which it repeatedly returned for food.

The mice could be using a mechanism to find food that
is analogous to the visual search images reported for birds
locating invertebrate prey (L. Tinbergen, 1960; Mook et al.,
1960). The search image concept implies that the animal pays
selective attention to certain cues characteristic of the prey
Species, or in other words there is some change in the stimu-
lus filter, either peripheral, or central. .This means that
as the animal moves about randomly in its environment, it
encounters an edible prey item, makes a cue—food association
and then locates additional Similar prey in the area (Holling,
1965; N. Tinbergen et al., 1967). Prey at very low densities
have a low risk of being eaten in this system. Their numbers
are too low for the formation of a search image by the pre—
dator. At very high densities the predator drops the search
image in order to insure a varied diet, rather than a diet
comprised all of one prey Species (L. Tinbergen, 1960).
Prey at moderate densities are most subject to the formation
of search images in this system.

In a predation model Holling (1959) prOposed that mice
(3, leUCOpus) learn to locate and consume their prey. This

learning process is characteristic of the formation of search

16

images. -As Peromyscus learns the olfactory cues associated
with healthy prepupae, they develop a search image and find
more prey in each localized area where they feed. The diet
of Peromyscus (Whitaker, 1966) in the wild does not corre—
late with the list of all abundant foods in the habitat.
The fact that the mice select only a few of the foods to eat
suggests that they may use several search images at one time.
The long list of items which Show up irregularly in the
mouse's diet might be present in the environment at densities
too low for the formation of search images. Search images
might also be used by the mice to locate other clumps of prey
Similar to the ones already consumed.

The reaction of the mouse to novel foods in the environ-
ment is an important part of its strategy. Wild strains of

rats (Rattus norvegicus) will avoid novel stimuli, while

 

domesticated white rats explore and sample new foods (Chitty
and Shorten, 1946; Barnett, 1955, 1956; Richter, 1955). No
comparisons have been made between young and adult rats for
their reaction to novel, but natural, food stimuli. Wild
rats will learn to avoid poisoned food through experience by
consuming sublethal doses of the bait (Richter, 1955). This
tox0phobia is learned more rapidly by young rats than adults.
The reaction to novel foods may wane with time (Thompson as
cited in Barnett, 1965). Young chimpanzees avoided novel
objects for a period of time after they were placed in their

environment, while adults made contact with the new objects

 

17

soon after they were introduced (Welker, 1956). As the
chimpanzees acquired more eXperience their fear of new ob-
jects disappeared. If young mice use a conservative feeding
strategy when they first leave the nest, they will learn to
eat only the foods that are in the area of the natal home
site. Young mice Should then Show a strong avoidance to
novel foods. As the young mice diSperse, they encounter many

new types of foods and the neophobia wanes.

Experience

 

The effect of early feeding experience on food prefer-
ences is a controversial subject. Positive effects of early
feeding eXperience have been demonstrated in several vertebrate
groups, including turtles (Burghardt and Hess, 1966; Burghardt,
1967b), birds (Rabinovitch, 1966, 1969; Kuo, 1967), and cats
and dogs (Kuo, 1967), but in small rodents the evidence to
date is equivocal. There are at least two schools of thought
regarding early experience effects on food habits.

One school claims that the food habits of adult animals
are entirely a function of eXperience (Kuo, 1967). Cats,
dogs and mynah birds, given either a restricted or a varied
early eXperience, showed entirely different behaviors when
tested with a new food. Animals reared on a restricted re-
gime avoided new foods, while animals reared on varied diets
ate new foods readily (Kuo, 1967). Kuo's major failing was

that he did not account for any inherited preferences or

18

aversions for his various early feeding regimes and test
foods, that is he does not provide any baseline data on the
food preferences and reactions of naive animals. The most
that can be said from Kuo‘s statements (no data are presented)
is that the food habits of the animals tested were modified
by eXperience.

The second school of thought hypothesizes that each
Species exhibits baseline preferences (inherited) which may
be modified by eXperience. Zebra finches normally prefer
red millet or white millet to canary seed (Rabinovitch, 1969).
Birds reared on each of these different seeds exhibit differ—
ent preferences according to their rearing regime. Those
reared on either red or white millet select that seed in
choice tests, but finches reared on canary seed selected both
canary seed and millet in preference tests. The inherited
prediSpositions of the birds to prefer the millet interacted
with their early feeding experience with canary seed.

A number of studies have tested the effects of early
feeding and taste experience in small mammals, without adopt-
ing either of these schools of thought. The results of these
studies have been contradictory. Some authors (Harlow, 1952;
Kuo, 1967) have reported positive early eXperience effects.
Other investigators have shown no significant early experience
effects (Bronson, 1966; FOrguS and Hutchings, 1960). Still
another group of authors (Warren and Pfaffman, 1959; Sique—

land, 1965; Thios et al., 1962) have reported intermediate

19

results, with positive effects right after the eXperience,
but no lasting impact.

Rats given a choice between two completely adequate
diets, one of which had been their normal food prior to the
test, preferred the diet to which they were accustomed. .This
preference may have resulted from a combination of eXperience
effects, knowledge that the normal food was quite adequate,
and a neOphobia for the new food. The work of Kuo (1967),
outlined above, Showed strong and lasting effects of early
feeding experience.

Studies providing rats with early taste experience with
bitter substances like sodium-octa-acetate (SOA) and hydrogen
chloride (HCl) have demonstrated that the preferences of the
rats were positively affected immediately after the eXperi-
ence (Warren and Pfaffman, 1959;.Siqueland, 1965). These
same rats retested after an intervening period Showed no
effects of the eXperience.

Thios et al. (1962), provided young rats with experience
drinking water solutions of three different concentrations of
sodium chloride (NaCl). When the rats were given their choice
of all three solutions, each group preferred the concentration
at which they had been given early eXperience. The results
of these taste experiments all indicate that experience may
be effective for a short period, but no long-term effects
are recorded. This, I feel, may be the result of the aversive

taste stimuli used in the experiments. The use of neutral or

20

pleasant stimuli in Similar eXperimentS will be needed to
determine the true effect of early taste experience.

Rats given early drinking experience with kerosene
tainted water, did not Show a significant preference for the
tainted water when tested immediately after their experience
(Forgus and Hutchings, 1960). Lasting effects were not tested.
Young rats reared with mynah bird pellets, either plain or
in a gelatin capsule, Showed no Significant early feeding
eXperience effects (Bronson, 1966). Bronson also tested the
possibility that the earliest food might acquire special
reward value that was lasting in effect. Again the results
indicated no effect of the early feeding.

The question of the effects of early feeding and drink-
ing experience on food habits or taste preferences remains
unanswered. Several important criticisms pertinent to all of
the above studies could, I think, indicate the reasons for
the conflicting results which have been reported. These in-
clude: (1) lack of adequate sample sizes; (2) re—use of
animals from one test to another, thus confounding the second
test with the additional eXperience gained during the first
test eXposure; (5) failure to determine the baselines of the
Species preferences for the substances used; and (4) the use

of aversive taste and/er food stimuli.

GENERAL.METHODS

Experimental Subjects

The mice used in these studies were from three taxa:
Peromyscus maniculatus bairdi, g, m, gracilus, and g}
leuc0pus noveboracensis. Wild caught g, m, bairdi and
g, leucopus were trapped at the Rose Lake Wildlife EXperi-
ment Station, Bath, Michigan. Wild caught g, m, gracilis
were trapped at the Dunbar Forestry Station, Channel View,
Michigan. .First generation laboratory (F1WC) stocks used
in the studies were the offSpring of the above wild caught
mice. Domestic (dom.) Stocks of g, m, bairdi and g, m.
gracilis have been laboratory bred for nearly 20 years (see

King, Price and Weber, 1968).

Maintenance

 

Four separate rooms were used to house the mice during
these studies: (1) a breeding chamber maintained at 65-780
F; (2) a holding room maintained at 68-75OF; (5) an eXperi-
mental test room maintained at 68-75OF; and (4) an eXperi-
mental chamber used for conditioning the mice, maintained at
62-77OF. All chambers were fixed on a 14 hours light, 10

hours dark cycle throughout the eXperimentS.

21

 

 

 

22

Relative humidity in all the chambers varied from 20 to 60
percent. All rooms and chambers contained ventilation fans
to maintain the air circulation.

Mice were housed in plastic laboratory cages measuring
6" by 12" by 6" deep, with fitted wire mesh lids. Purina
Laboratory Mouse Chow and water were provided §d_libitum
throughout the studies. Prior to the eXperimental procedures
mice were provided with pine shavings bedding and cotton
nesting material in their cages.

Young mice were housed with their parents until weaning
at 21 days of age. Mice used as adults (90-150 days of age)
were housed as litter mates (bisexual groups) until the
eXperimental procedures were begun. All experimental tests

were made on mice housed individually without pine shavings.

Odor Stimuli

 

The odor stimuli for all of the studies were three
essential oils: pine, Wintergreen, and anise, obtained from
Magnus, Maybee and Rhynard Inc. (Paramus, New Jersey).

Three separate food-odor combinations were obtained by plac-
ing three to five drOps of oil on cotton in the bottom of a
round metal food dish (one inch deep and three inches in
diameter) filled with Laboratory chow. The food was covered
with a metal retainer with large holes which enabled the mice

to gnaw the food withdut removing it.

 

 

 

25

Dependent Variable

Food consumption has been used previously as a dependent
variable in studies of early experience effects (Rabinovitch,
1969; Burghardt, 1967b; Bronson, 1966). Some authors have
used absolute weights of food eaten as a dependent variable
in comparing animals and treatment groups. ,This measure
totally ignores the weight of the animal which may vary widely.
Thus, a larger animal, consuming more food, will contribute
more to an overall measure of preference than a small animal
eating lesser quantities of food. In order to avoid this
error, the total weight of food consumed must be divided into
the consumption from each food container for each mouse.

This produces a dependent variable which is the percentage of
the diet taken from each alternative food choice. A day by
day computation provided the dependent variable used through—
out these studies. Weights were taken to the nearest 0.1
gram on one of two scales; a Shadowgraph Scale (Exact Weight
Scale Company, Columbus, Ohio), or an Autogram 1000 Scale
(O'Haus Corporation, Union, New Jersey).

Use of the percentages of the diet eaten from each of
the three food-odor combinations introduces the problem of
dependence of each percentage. Since an animal has a limit
to the amount of food it can consume during any given time
period, consumption at one dish early in the activity period

limits the amount of food that can be consumed from the other

 

 

24

dishes during the same day. Thus, all of the analyses con-
ducted were done separately for each food-odor combination.
This method of analysis technically permits the use of only
two of the three food-odor combinations as there are only
two degrees of freedom. To obtain a clear and complete
picture of the odor preference patterns, I have analyzed all

three combinations.

 

 

GENETIC FACTORS

Purposes

 

The purposes of this first experiment were: (1) to
determine the Species Specific odor preferences among the
three food-odor combinations; (2) to test for Species and
age differences in odor preferences; (5) to determine whether
domestication has altered the odor preferences of the mice;
and (4) to test the hypothesis that young mice tested
repeatedly (longitudinal), or only once (cross-sectional),

at various ages, have the same odor preferences.

Test Subjects

 

Eight groups of 10 adult mice each were tested. These
eight stocks were: (1) wild caught 2, m, bairdi; (2) FlWC
.g. m, bairdi; (5) domestic P, m, bairdi; (4) wild caught
P. m, qracilia; (5) F1WC g, m, qracilis; (6) domestic g, m,

gracilis; (7) wild caught 2, leuc0pus; and (8) F1WC g, leu—

 

c0pus. Young mice used were all from F1WC stocks of g, m,
bairdi and g, leucogus. One test group of eight mice of each
Species was tested longitudinally and 10 groups of eight mice
each, from each Species, of various ages were tested cross—

sectionally. Throughout the experiment no two mice from the

25

 

26

same litter were used in any of the test groups and no mouse

was used more than once.

Procedure

 

Each mouse was housed individually in a double plastic
cage (Figure 1), comprised of two of the plastic cages previ-
ously described connected by a one inch section of one inch
diameter Plexiglas tubing. One side of the cage contained a
nest box, nesting cotton and water. The other side of the
cage contained the three food dishes, each with a different
food-odor combination.

Adult mice were weighed and placed in the cage with
three weighed food dishes. At three day intervals for 27
days the mouse and food dishes were weighed. At each weigh-
ing the cotton with odor stimulus was changed, and the posi-
tions of the food dishes were rotated in a random sequence
to prevent the development of a position habit. Food was
added where necessary to maintain adequate food levels in all
the dishes.

Young mice were tested identically except that the longi-
tudinal groups were weighed every two days. These mice were
tested from weaning at 21 days of age until they were 61 days
of age. Cross-sectional groups began at 21, 25, 29, 55, 57,
41, 45, 49, 55 and 57 days of age and mice in each group were

tested for a two day interval.

 

 

 

27

Figure 1. Schematic drawing of the double plastic cage
(top View with wire lid removed). One inch
in the diagram equals two inches of the
actual cage and equipment.

 

 

 

 

 

28

FIGURE

 

Food
Dish

Food
Dish

Food
Dish

:
o
.-
I
L
a
a

Pflosfic

A

 

Connecfing
Tube

1

Nest

Box

VVHh

Cotton

 

 

 

 

 

29

Food consumption was obtained by subtracting the weight
of the food dish at the end of the time interval from the
weight of the dish at the beginning of the interval. This
value was corrected for humidity changes by a factor deter-
mined from control dishes placed in the eXperimental chamber.
The corrected weight of food consumed from each dish was
then used to compute the dependent variable; the percentage

of the diet taken from each food-odor combination per day.
Analysis

For young mice tested longitudinally and all adult
groups a mean percentage and standard error was calculated
and graphed. The data were analyzed within each odor across
the ten test groups using a one way analysis of variance
followed by Duncan‘s New Multiple Range test (Li, 1964).
This analysis tested for differences in Species preferences,
ages, and between wild caught and domestic stocks. Variance
in preferences (called here fixity) was computed within each
test group across the three odors. A mouse with complete
fixity would take 100 percent of its food from one of the
food-odor combinations and no food from the other two combi—
nations, which would produce a variance across odors of
5555.55. For each test grOUp the variance computed was
divided by this figure of 5555.55 to produce a ratio. Any
deviation from complete fixity would result in a ratio less

than 1.00. The lower the ratio computed in this manner, the

 

 

 

 

50

less fixed are the odor preferences of that particular
group.

Differences between cross-sectional and longitudinal
test groups of young mice were tested within each odor using
a parametric paired t—test. The basis forIpairing was the
age variable. -Mice in the two groups were compared for the
same ages. These tests were made within each odor to insure

the independence of the measures on each food—odor combi-

nation.
Results

Results of the first experiment are Shown in Table 1
and Figure 2. .The one way analyses of variance within each
odor showed no Significant differences across the ten test
groups for any of the three-food-odor combinations (anise,
F=0.10, DF=9,86, p > .10; Wintergreen, F=0.95, DF=9,86,

p > .10; pine, F=0.41, DF=9,86, p.> .10). The Duncan's
Multiple Range tests run on each analysis failed to Show any
Significantly different grOups. In all groups the pine
odored food was preferred (Figure 2). Pine comprised from
58 percent of the diet for wild caught and domestic g, m,
bairdi, to 92 percent of the diet for the F1WC P, m, bairdi
stock. Young F1WC g, m, bairdi and P, leuc0pus also showed
a strong preference for the pine odored food. Wintergreen
was the second most preferred food-odor combination in eight

of the ten test groups. Mice ate from this odor for between

 

51

 

 

 

is.ovm.me 1m.ov>.ma Am.ova.m Hmuos
ma.o 1m.mvm.ms Aa.aavo.ma A>.mvs.m osdm pasoa .mmmmmmmm am
.mm.o is.svm.sm Am.avo.m 1m.mvm.mfi 63 “Have .mmmmwmma am
sm.o 1m.aavm.so Am.svs.m Am.oavo.mm ozam mazes mmmmwmww am
no.0 Aa.mva.mm Am.mva.ma Am.ovm.o .soo pago<.mmammmmm as am
sm.o Am.mvm.sm Am.mvm.ma Ae.avm.a ozdm pasca.mmwmmmmm as am
am.o As.ava.mm is.evm.aa is.mvm.m oz ussoa.mwammmmm as am
ma.o Ae.aavm.sm Ao.oavm.wm A©.mvm.s .509 bases .mmmwmm as am
”5.0 Am.mvm.sm Am.acm.m o.mva.m osdm assoa .wmmwmm as am
ms.o Ao.oavs.mm Ae.mvm.mm Aa.mvm.ma us pssoa .wmmmmm as am
omio 8.82% $4.56 3:353 osdm mazes filed .a. .m.
oflpmm ocflm cmmumhmpcflz omflsm cflmuum wm< mmflommm
OUGMHHm> sopo

 

 

.mcoflymcflnfioo Hopo poow
mmnnp onu mo Sumo Eosm pmasmsoo umflp mgu mo A.m .m fi.Hv ommucwonmm cmmz .fi OHQMB

 

 

 

Figure 2.

52

Mean percentage (1 1 S. E.) of the diet eaten
from each of the three food-odor combinations
(A=adult; Y=young; WC=wild caught; F1WC=first
generation laboratory reared; Dom.=domestic).

%

MEAN

%

MEAN

%

MEAN

50

25

75

25

55

FIGURE

 

 

2

WINTERGREEN

ANISE

mm...

 

100

7’5

50

25

 

 

WC
A

 

 

F‘WC Dom.
A A

Rm.boirdi

 

F‘WC
Y

 

 

+...|_

 

 

 

PINE

 

41+

 

 

 

WC

A

F WC Dom.

I
A A

P. nugracilis

WC FWC FWC

A

1
IA Y

P. leucopus

 

 

 

54

4 percent (young F1WC 2,.leucopus), and 55 percent (domestic
adult g, m, bairdi) of their diet. Anise odored food was
not eaten for more than 15 percent of the diet except by
adult wild caught g, m, bairdi (19 percent) and young F1WC
g, leucopus (29 percent). 'Across all groups the average
diet composition was 9.4 (i 0.2) percent anise, 15.7 (i 0.5)
percent Wintergreen, and 76.9 (i 0.7) percent pine. .These
data showed a strong preference for pine odored food and no
Significant differences in the preferences across all the
test groups within each odor. No differences were found
between young and adult mice, among Species, or between wild
caught and domestic stocks of either g, m, bairdi and g, m,
gracilis.

The mice in these groups were tested for periods of
time ranging from 27 days (adult mice) to 40 days (young
mice). .No measurable deviations from the observed preference
levels were recorded during these time periods.

Young mice with little or no prior feeding eXperience
Showed the preference for pine odored food immediately.
Comparisons of the cross-sectional and'longitudinal groups
of young mice, using t-tests, revealed no significant differ-
ences for any of the three food-odor combinations (pine,
t=0.8, p :>.10; Wintergreen, t=O.9, p >~.10; pine, t=1.1,

p )».10; all with:DF=7). Groups of young mice tested once
showed the same pine preference at each test age as mice of

Similar ages tested longitudinally.

 

55

Examination of the degree of fixity of these odor
preferences (Table 1), as measured by the within group
across odors variance, Showed that F1WC g, m, bairdi adults
showed the most fixity (ratio=0.77), while the other two
adult 3, m, bairdi groups showed the lowest degrees of
fixity (ratio=0.15 and ratio= 0.18). g, m, gracilis Showed
a consistently high preference fixity (ratiol> 0.57). The
ratios for the two groups of young mice were relatively low
among the test groups, being 0.50 for g, m, bairdi and only
0.51 for g. leucogus. Adult g, leucopus also showed a
lower degree of fixity than most of the other test groups

(ratio=0.48).

 

 

CONTROL EXPERIMENTS

A number of control eXperiments were necessary to estab-
lish: (1) that the mice were not using a position habit;
(2) that the mice were not differentiating intensity changes
within an odor; (5) that the pine Shavings in the rearing and
holding cages had no effect on the preferences; (4) that the
mice could not differentiate soiled (used by other mice)
food from unsoiled food; (5) that dish preferences were not
present; and (6) that seasonal Shifts in the Species odor
preferences did not occur. .The first control, that for posi—
tion habit, was accomplished by rotating the positions of
the dishes in a random sequence, at least once every three
days. This rotation method was used throughout all of the
eXperimentS. The other controls for dish preference, in-
tensity, seasonal Shifts, food condition and shavings eXperi-

ence were tested as described separately below.

Control for Odor Intensity

The cotton with the odor stimulus at the bottom of each
food dish was changed every three days. While I was unable
to control for possible differences in the vapor pressure
and intensity present for each odor, this method of replace-

ment kept the same intensity present for each odor. It was

56

 

57

necessary, however, to make certain that the mice could

not differentiate fresh odor, just placed in the bottom of
the food dish, from odor three days old. -As an extra margin
the determination was also made for odors Six days old.

Young and adult F1WC g, m, bairdi and P, leucopus were
tested with 12 mice per group. Each mouse was presented
with three food dishes in the double plastic cage apparatus
described previously (Figure 1). The three food dishes
contained the same odor stimulus which had been put on the
cotton at different days. one dish was supplied with food
and odor six days prior to the start of the test, one dish
was set up three days prior to the beginning of the test,
and one dish was prepared with odor stimulus and food on
the day the test began. In effect this created a graded
series of intensities representing 0, 5 and 6 day old odors.
Each dish and the mouse were weighed for two three day inter-
vals. A fresh dish was added in place of the 6 day old dish
at the middle weighing interval to maintain the same graded
series of intensities.

The dependent variable was the percentage of the diet
taken from each of the three odor intensities. The data
were the same for the two time intervals, so the results were
computed over the entire six day test period. The results
were analyzed separately within each odor. The twelve mice
in each age and Species group were randomly divided into three

groups of four mice each. For each mouse in the first group

 

 

 

58

 

 

 

 

 

 

Am.aav Ao.m v As.mav Ao.mav Ao.mav Aa.mav Aa.aav A>.afiv Aa.mav
N.mm ¢.mm w.¢m fi.wm m.mm m.~w m.mm o.wm m.>m ma waspfi
Am.mav Am.m v Am.mav Am.aav Aa.oav Am.mav Aa.wav Am.mav Aw.oav
m.mm w.mm >.mm m.mm ¢.mm ¢.>m m.w¢ m.mm m.mfi Na masow
msaoosma KM
$.m v 36”.: 8..va 8.23 8.”: SEN: $.Nav :30: ANJHS
fi.¢m m.>m o.wm o.m¢ «.mm m.fim m.om 0.0% m.mm Na waspd
Aa.aav Ao.mav Am.mav Am.mav Am.mav Ao.mav Aa.aav 14.8 v Am.aav
«.mm m.om m.o¢ m.¢m m.mm o.mN >.¢m >.mm w.fi¢ ma masow
6.33 .a .M
z msonw
m m o m m 0 m m o huHmCGDCH
mcflm smmumumpcflz mmflcd Hopo

 

 

.wuflmcmucfl zoom Eonw copmm pmflp map mo A.m .m a Hg

mmmucmouom coo: .pmmp Houucoo >uHmCODcH map mo mpasmou may wo whoafism

.N OHQME

 

 

 

59

only the percentages for the 0 day intensity were used.
Similarly the 5 and 6 day intensity data were used from the
mice in the other two groups. In this way the results could
be analyzed across the different intensities Without violat-
ing the assumption of independence. Within each odor a
three way analysis of variance was run using three levels of
intensity, two Species, and two ages. The three analyses
showed no Significant Species, age, or intensity differences
within any of the three odors. The F-ratios for intensity
differences within each odor were; pine, F=1.5, p >>.10;
wintergreen, F=1.4, p >>.10; anise, F;1.4, p.> .10; each
with DF=2,57. Thus, changing the cotton with the oil odor
stimulus at three day intervals controlled for constant in-
tensity within each odor and for constant relatiVe intensie

ties between different odors.
Control for Dish Preference

It might be argued that the preferences in the first
eXperiment were due to the mouse marking a preferred dish
in some manner, and then returning to the dish, regardless
of the odor stimulus present or its position. To test dish
preferences I presented each mouse with three food dishes,
all with the same odor and intensity and rotated the posi-
tions of the dishes every other day. The amount of food
consumed at each dish was measured for three intervals of

three days each. Young and adult F1WC g, m, bairdi and

 

 

40

.E- leuc0pus were tested with sixteen mice in each group.
From this total, groups of four mice each were tested with
anise, wintergreen and pine odor and with plain laboratory
chow, the control stimulus in later eXperiments.

The dependent variable was the percentage of the diet
taken from each dish per day. The results were analyzed for
dish preference within each mouse, using a Kruskal-Wallis
one-way analysis of variance (Siegel, 1956). The results
of these analyses are Shown in Table 5 as the ratio of the
number of animals exhibiting a dish preference out of the
four mice tested within each grOUp. The analyses indicated
that only 5 mice out of 64 tested showed a Significant
dish preference. There were no trends within Species, ages,
or the different food-odor combinations tested. The result
obtained, that is three Significant cases of dish preference
out of 64 tests, would be eXpected on the basis of chance.

The mice in these studies were not using a dish preference.

Seasonal Control

Experiments on species odor preferences were initially
conducted during January of 1969. The strategy experiments
were not concluded until September of 1969, and the bulk of
the eXperience study was conducted during the summer months.
A control eXperiment tested the hypothesis that the odor
preferences of the mice had not changed from January to July.

The procedures were exactly as described for the initial

41

 

 

 

 

 

Table 5. Summary of the dish preference control test. The
ratio presented is the number out of four animals
tested in each group showing a significant dish
preference; ,

Odor

Group N _Anise Wintergreen Pine Control
g, m, bairdi

Young 16 0/4 0/4 1/4 0/4

Adult 16 0/4 0/4 0/4 1/4
2, leucopus

Young 16 0/4 0/4 0/4 0/4

Adult 16 0/4 1/4 0/4 0/4

 

 

 

 

42

determinations of Species odor preferences, except that only
two groups, adult F1WC g, m, bairdi and g, leuCOQuS were
tested. AS before, 10 mice were tested in each group for
nine three-day intervals. .The dependent variable was the
percentage of the diet taken from each food-odor combination.
To test the hypothesis concerning seasonal shifts in odor
preferences, paired parametric t-tests were used. Each odor
was analyzed separately within each Species. The basis for
pairing the January and July test animals was the time inter-
val variable.

The results are shown in Table 4 where the means and
standard errors are presented along with the t values and
probabilities. Odor preferences in these groups did not
shift significantly (p > .10) from January to July. This
makes it unlikely that the odor preferences within any of
the stocks changed during the total eXperimental period of

10 months.

Shavings Effect

All of the mice used in these studies were housed with
pine Shavings bedding until the start of the test procedure.
Young mice were on shavings for up to 21 days and adult mice
from birth until 90-150 days of age. The pine shavings
smelled similar to the pine odor used in these studies.

This control experiment tested the hypothesis that mice

 

 

 

45

 

 

 

 

 

mz as.o Am.mvm.as Am.mvm.ms mafia
mz Na.o is.mvs.aa Am.mva.m cmmnmnmucsz
mz ms.o A¢.>Va.ma AN.mvm.oa mmflcm
om msaoosma .M.
mz mo.fi Am.mvm.m> Ao.mvm.wm mnflm
mz mm.fi A¢.¢Vm.ma A>.ovm.ma cmmsmnmpnflz
mz om.o Am.avw.m Am.dvm.m omflcg
om Hannah as 2M
m Amumnvu mash %HMDSMb Hope 2 msouw
.coflwmcHQEoo Mononpoom comm Eosm cmpmo umflp
on“ yo A.m..m fi.flv mmmncmoumm coo: .ummp Houusoo Hmcommmm may mo muasmmm .w magma

 

 

44

reared either with or without the pine Shavings in the cage
would show the same odor preferences.

Pregnant mice were placed in bare plastic cages with
only a small amount of nesting cotton and the necessary
laboratory chow and water. .The offSpring were raised from
birth in a room without any of the pine Shavings. At 21
days of age the mice were given the standard Species odor
preference test with all three food-odor combinations. Both

.3. m. bairdi and g, leucopus young were tested. The test

 

period lasted Six days. Four mice were tested in each grOUp,
each mouse from a different litter. These mice were compared
with another group raised under standard laboratory condi—
tions with pine Shavings and in the mouse colony room, where
the shavings odor and other possible olfactory stimuli were
always present.

Mann-Whitney U nonparametric t—tests were used to com-
pare these groups (Siegel, 1956). rAgain the odors were tested
separately because they lacked independence if tested in the
same analysis. None of the U values was significant, indi—
cating that the species preferences for pine odor were not

the result of being raised on pine shavings (Table 5).

Soiled Versus Unsoiled Food

The food used in these studies was stored for up to
three weeks in large bins, and re-used. Each bin contained

food of only one odor. The food in each bin was mixed before

 

 

 

 

45

Table 5. Results of the test for the effect of pine Shavings
on food-odor preferences. Mean percentage (1 1
S. E.) of the diet eaten from each food—odor combi-

 

 

 

 

 

nation.
Odor
Group N Anise Wintergreen Pine
3, m, bairdi
Shavings 4 8.5(1.1) 15.5(2.4) 76.4(5.2)
No shavings 4 7.0(1.2) 12.6(2.5) 80.4(1.6)
U(DF=5) = 5 7 5
Prob. NS NS NS
g, leuc0pus
Shavings 4 5.8(5.5) 11.9(2.5) 82.5(2.6)
No shavings 4 6.1(1.5) 15.7(1.7) 80.2(1.0)
U(DF=5) = 6 5 6
Prob . NS NS NS

 

Table 6. Results of the test for soiled versus unsoiled food.
Mean percentage (1 1 S. E.) of the diet eaten from

each food type.

 

 

 

Group N Unsoiled Food Soilethood
P m bairdi 8 44.5(10.6) 55.5(11.7)
P. leucopus 8 55.6(8.9) 46.4( 9.7)

 

 

 

 

 

46

re-use to insure a random mixture of the age of the food.
This control experiment tested the hypothesis that the mice
did not discriminate between re-used and unused food. Both
foods were a mixture from bags of food Opened one to three
weeks previously. Thus, the age of the food was not a
factor.

One group of eight adult mice was tested from each
species: 3, m, bairdi and g, leucopus. Two food dishes
were presented to each mouse, one with previously used food
and one with the unused food. The food dishes were weighed
for two time periods of two days each. The data were lumped
for the entire eXperimental period and the dependent variable
was calculated as the percentage of the diet taken from each
type of food.

Results of the analysis of variance performed on these
data (Table 6) showed no significant species effects (F=1.45,
p > .10, DF=1,12), food effects (F=0.97, p > .10, DF=1,12).
or interaction effects (F=1.68, p >-.10, DF=1,12). The re—
sults of this test indicated that the mice were not discrimi-

nating between previously used and unused food.

 

 

 

EXPERIENCE

Purposes

This eXperiment tested the effect of prior olfactory
eXperience on the preference for a particular food—odor
combination by examining: (1) whether young mice were
more affected by the eXperience than adult mice; and
(2) whether there were any Species differences in modifi-

ability of odor preferences.
Test Subjects

Four groups of mice were used in this study: (1) young
3, m, bairdi; (2) adult 3, m, bairdi; (5) young 2, leucoEus;
and (4) adult g, leucoEus. All mice were from F1WC stocks.
The young mice used were all 21 days of age at the beginning
of the eXperiment and the adult mice were between 90 and 150
days of age (mean = 121 i 2). A one-way analysis of variance
was used to test the hypothesis that the average age of the
adult mice was the same for each test group. The data and
the analysis are presented in Table 9. There were no Sig-
nificant age differences across the test groups (F=0.26,

p >>.10, DF=7,184).

47

48
Procedure

The design of the entire eXperiment is summarized in
Tables 7 and 8. The independent variables manipulated were:
(1) age at time of eXperience; (2) age at time of testing;
(5) quality of experience; (4) types of tests used; and
(5) genetics (see King, 1958).

The apparatus used in the conditioning process was the
double plastic cage previously described (Figure 1), with
food dishes on one Side and water and nesting cotton on the
other side. Each mouse was conditioned individually for two
weeks with all three food-odor dishes present, but only one
food-odor combination available for eating. Control mice
were presented with three dishes, all containing laboratory
chow, but no oil odors. A mouse was trained to associate
food with one of the odors by allowing it to eat from only
the dish with the conditioning odor. The other two food-odor
combination dishes were blocked by a wire mesh screen that
prevented the mouse from Obtaining food. The mouse was
positively reinforced for going to one odor for food, and
was negatively reinforced (no food) at the other two dishes.
Control mice were allowed access to only one food dish, the
other two being blocked by Similar wire mesh screens.

Genetics was manipulated by testing both 3. m. bairdi

and g. leuCOpus. Both young and adult mice were used to

 

determine any aging effects. The quality of the experience

was varied by using the three food-odor combinations and

 

 

 

 

49

 

omuz Ammm mo ammo omauomvuasoa
mmuz Ammo mo m>mp dmvmcsow ”newcmomuoao>oc msaoosma msomaaosom

 

mmuz Imam no mama omsuomvuasom
wmnz Ammo mo mhmp amvmcsow ”HUMHOQ msumHOOflcmE msom>EonOm

 

”mmsouw pmoe

 

@"C mug mug m”: HOHHCOU
@"C @"C mug mug mCHm
mus mus mus mus comumsowcflz
@"G mug @“C wuc mmflgd.
pmmB pmmB pmoe pmoe AmMOOB NV
O>Hpflpomm¢ %HODOEEDchU m>HOHummm¢ >H0pmfifismcoo nopo UOGOHDHUGOU
uwoa mumq pmoe oumflUmEEH

 

 

.mmsoum anon snow onu mo some now povmmmos.mm3 cmflmop ofimm one
.moocmnommsm Hopo so mucosuwmxo mo uommmo may mo SUDDm may now smflmon .s OHQOB

 

50

Table 8. Summary of the ages at which the mice were condi—
tioned and tested. Age given is that at start of
the conditioning or test period.

 

 

 

Age at Condi— Immediate Late Test
Age Class tioning Test Age Age
Young 21 55 65
Adult 90—150 104-164 119-194

 

Table 9. Mean ages (i 1 S. E.) of adult mice used in each
test group in the eXperience study.

 

 

Species P, m, bairdi g, leuc0pus
Test Stick Stick ‘Cage Cage Stick Stick Cage Cage
Time of

Test Early Late Early Late Early Late Early Late

Mean age 118.5 124.5 125.0 118.9 120.8 119.5 121.6 125.5
Std. error 5.6 5.8 5.1 5.6 5.2 5.4 4.5 5.2

F=0.26 DF=7,184 p > .10

 

 

 

 

 

51

laboratory chow. The age at the time of testing was manipu—
lated by giving mice tests immediately after the eXperience
and one month later.

The mice were conditioned for a two-week period. Every
three days the positions of the food dishes were rotated in
a random sequence to prevent the development of a position
habit, and the odor stimulus was refreshed with new oil, to
maintain the constant intensities of the different odors.

At the end of the conditioning period mice were either tested
immediately, or one month later. During the interim month
the mice were housed in a Single plastic cage with normal
laboratory chow, water and nesting cotton (no shavings).
These two groups were designed to test the immediate effects
of the olfactory eXperience and its perseveration. The same
mice were not used in both the early (immediate) and late
(one month) test groups.

Two types of tests were used. Behavioral sequences
usually involve both an appetitive searching phase and a
consummatory phase (Craig, 1918; Lorenz, 1960). The appeti—
tive phase is a variable sequence of searching behaviors,
i.e., for a nest, for food, or for a mate; and the consumma—
tory phase is a more stereotyped pattern of terminating the
sequence, i.e., eating, Sleeping, or c0pulating. The two
tests employed in the present study measured these two
phases for the feeding sequence. The consummatory phase was

tested by measuring the amount of each food—odor combination

 

 

 

52

consumed when all three food-odor combinations were present.
The appetitive phase was measured by the number of sticks
a mouse chewed through to get to the food behind the sticks.

Appetitive and consummatory tests were given to separate
groups of mice. Each test was given to separate groups of
mice tested immediately and to mice tested one month after
the conditioning experience. The consummatory test involved
presenting each mouse with three food dishes containing the
three food—odor combinations. The amount of food consumed
at each dish was determined by weighing the dishes daily for
four days. Control mice were tested with the three food-odor
combinations and provided the baseline percentages used for
comparing with the results of the experimental groups.

The appetitive test apparatus consisted of a ten gallon
aquarium partitioned at one end by a two inch thick board,
extending the height and width of the aquarium. The board
had three openings at the bottom. In each opening 14 balsa
wood sticks were arranged in three rows of 5, 4 and 5 sticks
each (see King et al., 1968). Behind each opening was one
of the three food—odor combinations. The test animal was
placed in the chamber at the beginning of its activity period,
approximately 14 hours Since it had last eaten. The depend-
ent variable was the food-odor combination behind the open—
ing through which the mouse chewed to obtain food. If experi-
ence was effective in conditioning the mouse to associate

food with a particular odor, then the mouse Should chew

   

 

 

 

55

through the sticks in front of that particular food-odor
combination. Control mice were tested in a Similar manner.
All tests lasted 24 hours, or until the mouse died. The
positions of the three food—odor combinations behind the
barrier were Shifted at random between test subjects to con-

trol for any Species specific position bias.
Results

During the olfactory experience a number of mice died
before completing the two weeks prescribed. Each dead mouse
was replaced by another of the same age and Species. The
frequencies of deaths in each group were compared against a
probability of equal deaths in each conditioning treatment
using a Chi—square test (Table 10). It was necessary to lump
the young and adult mice within each Species to obtain fre—
quencies large enough to test. The results were not signifi-
cant for g. leucopus (X2=7.46, DF=5, p ;>.O5), but were
significant for g. E. bairdi (x2=54..02, DF=5, p < .001).

For 2, m. bairdi more deaths occurred in the anise treatment
groups than would be expected on the assumption of equal
probability for all treatments. Fewer deaths occurred in

the control group than would be eXpected. Gut analyses of
the dead g. m. bairdi indicated that they were probably dying
from starvation. All of the deaths that occurred during the
conditioning phase occurred during the first three days of

the procedure. Body weights of the dead mice showed that they

54

 

 

Ofi. m¢.> m d we we mma msmoosma .M
a4»
«00. No.4m a e m mm Nmfi HUHHOQ .H .M
z msouo Emma
.Qonm Amummvmx
Hopo

Hosucoo msflm swoumumpcflz wmacd

 

 

.mmoooum mcHQOHqucoo wuouumwao onu mcflnsp mflpmmp Mo

wocosvwsm .oa magma

 

 

55

had lost up to 40 percent of their original pre-experiment
body weight. Mice surviving the conditioning eXperience

did not Show a weight loss. The first experiment showed

that F1WC g. m. bairdi had a strong preference for the pine
odored food (91.9 percent of the diet), and virtually avoided
the anise food—odor combination (2.4 percent of the diet).
This aversion may account for the high death rate in g. m.

bairdi.

Appetitive Test

In Table 11 the data from the appetitive test are pre—
sented as the number of mice out of six test subjects in
each group that chewed through the sticks in front of the
food—odor combination with which they had been given condi—
tioning eXperience. These ratios were added within each
Species and age group. The totals and the probability de—
termined by a Sign test (Siegel, 1956) are Shown in Table 10.
A one—tailed test was used because the effects of the experi-
ence could only be in one direction (pilot study). The data
were analyzed separately for each test period: immediate
(early) and late (one month later).

Young 2. m. bairdi strongly preferred the food-odor
combination to which they had been conditioned when tested
immediately (17/18, p <f.001), and they retained this prefer—
ence over one month (14/18, p <f.02).‘ Conditioning was
equally effective with each of the three food-odor combina-

tions. Adult g. m. bairdi were not significantly affected

56

 

 

No. mfi\¢fi ®\@ m\w m\w doo. ma\>fi ©\© w\m m\m DH5U<
No. wfi\¢fi ©\m m\w m\¢ fioo. mfi\>fi m\m m\m m\m masow
msmoosma 2M
mz mfi\m m\w m\m m\fi mz mfi\m ©\m m\m m\N OHSU¢
NO. mfi\¢fi ©\m ©\m m\w fioo. mfi\ha m\m ©\m m\m masofi
Hannah as am
msouw
.Qoum Hmuoa m 3 4 .Qonm HOOOB m 3 d Hopo
pmma mung umme maumm

 

 

.Aosflmum “cmmumuwpsflznz “mmflcmuév
pwcofluflpcoo may mo “COMM CH mxofluw esp cmDOHSD msHBOSU msoum “mow 30mm CH
me mo #50 OOHE mo H0985: may mH OHDOH 0L8

.COHOOCHQEOU HOUOIpoom

.nmmn m>nunnmmmm was go mussmmm .aa magma

57

by the conditioning experience, either immediately, or one
month later.

Both young and adult 3. leuc0pus were significantly
affected in their food—odor preferences by the prior condi—
tioning to exactly the same extent: immediate test (17/18,

p <:.001), and late test (14/18, F) <: .02). During the early
test all combinations were equally effective (5/6 or 6/6),
but during the test one month later the pine (6/6) was more
effective than either wintergreen (4/6), or anise (4/6).

The results from the appetitive tests given to control
mice (Table 12) showed that these mice were not associating
any of the test odors with food. Twenty—three of the 48 con—
trol mice died during the appetitive test without ever chew—
ing through the sticks to obtain food. The deaths during
the test were equally distributed among the four Species and
age groups. Of the 144 mice conditioned to associate one of
the three test odors with food, only 17 died during the ap-
petitive test, with the maximum number of deaths being two in
one cell, which occurred only once. These deaths were prob—
ably also due to the failure of the mice to associate food
with any of the test odors.

In summary, young g. m. bairdi and young and adult
2. leucopus all showed strong effects of the conditioning
experience in the appetitive test. Adult g. m, bairdi were
not Significantly affected by the odor conditioning. Test

groups that showed significant experience effects when tested

 

 

58

Table 12. Results of control mice in the appetitive test.

Early and late tests combined.
number out of 12 mice in each group chewing

The ratio is the

through each food-odor combination, or dying.

(A=anise; W=wintergreen; P=pine)

 

 

 

 

ReSponse

Group A W P Dead
2, m, bairdi

Young 1/12 2/12 1/12 8/12

Adult 1/12 1/12 5/12 5/12
E, leucogus

Young 1/12 2/12 4/12 5/12

Adult 2/12 1/12 4/12 5/12
Totals 5/48 6/48 14/48 25/48

 

 

 

 

 

59

immediately after the olfactory experience also showed sig-

nificant effects when tested one month later.

Consummatory Test

The results of the consummatory test were analyzed for
the first day (one) and the last day (four) of the test in
order to discern a Shift in preference over the test period.
The dependent variable used was the percentage of the diet
eaten from each food—odor combination. The mean percentages
and standard errors for all test groups on day one are shown
as bar histograms in Figures 5, 4, 5, and 6.

Since all groups preferred the pine odored food, the
experimental groups were compared with the control groups
within each food-odor combination. For example, an individual
mouse from each test group was compared with the control group
baseline to determine whether the experimental mouse had
eaten a greater percentage of the food-odor combination on
which it had been conditioned than had the control. The
number of mice out of six in each test group that ate a greater
percentage of their diet from the conditioned food—odor combi—
nation than the control mean was computed as a ratio, exactly
as in the analysis of the appetitive test.

The results of the analyses for day one (Table 15) and
day four (Table 14) were identical, indicating that there was
no waning of the experience effect during the four day test
period. Young 3. m. bairdi Showed a Significant experience

effect at both the early and late test periods, while adult

 

Figure 5.

 

60

 

Mean percentage (1 1 S. E.) of the diet consumed
from each food-odor combination for day one of
the test, both immediately and one month after
conditioning in young 2. m. bairdi.
odors indicated as:
P=pine; C=control.

Conditionhm
A=anise; W=wintergreen;

61

 

 

 

 

 

FIGURE 3
Rm.bairdi Young
'00 IMMEDIATE
75
V
O
2
< soJ
H
E
25
. __I.I u II
A W P C A W P C A W P
TEST ODOR ANISE WINTERGREEN PINE
100
ONE MONTH
75

 

 

 

x5
2 °I
<
2
25
. u __ II

A w P C A w P C A w P
TESTODOR ANISE WINYERGREEN PINE

 

 

 

 

62

Figure 4. Mean percentage (i 1 S. E.) of the diet consumed
from each food-odor combination for day one of
the test, both immediately and one month after
conditioning in adult g. m, bairdi. Conditioning
odors indicated as: A=anise; W=wintergreen;
P=pine; C=control.

65

FIGURE 4

V’ m.boirdi Adult

IMMEDIATE

50
25 AI

MEAN Z

 

 

 

 

 

 

 

 

 

 

 

 

 

w P c T I A 7 w 7 A w v c
TEST ooon ANISE wmnnonnu PINE
100
ONE MONYH
75
V
o
z
<
... so
2
25 [h
.. w P c I I A 7 7 A w p c I
1151’ coonr ANISE

WINTERGREEN PINE

64

Figure 5. Mean percentage (1; 1 S. E.) of the diet consumed
from each food-odor combination for day one of
the test, both immediately and one month after
conditioning in young g, leucoEus. Conditioning
odors indicated as: A=anise; W=wintergreen;
P=pine, C=control.

 

100
75

N

z

(

w

2 50
25

 

 

0L

P. leucOpus Young

 

 

 

 

65

FIGURE 5

IMMEDIATE

 

 

 

 

 

 

 

 

rI-
EIIA

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

P 7 w i — c I A w P c
YES? 0003 ANISE WINTERGREENI PINE
100
ONE MONTH
I
75
N
z | F
. s. r
H _
2 —---I
25
I
o _
I A w P A w P c A w p c
TEST oooa ANISE WINTERGREEN PINE

66

 

Figure 6. Mean percentage (1 1 S. E.) of the diet consumed
from each food-odor combination for day one of
the test, both immediately and one month after
conditioning in adult g. leucopus. Conditioning
odors indicated as: A=anise; W=wintergreen;
P=pine; C=control.

 

67

FIGURE 6

R leucopus Adults

I00
IMMEDIATE

75

.. "I

O
MEAN I:

 

25

 

 

 

 

 

 

 

 

 

 

 

 

    

O f r
A W P C A W P C
TEST ODOR ANISE WINTERGREEN PINE
I00

ONE MONTH

75

25

O F
A W P C A W P C

TEST ODOR ANISE WINTEROREEN

 

 

 

 

 

 

 

 

 

 

 

 

 

68

 

 

mz ma\mfi m\m m\m m\¢ No. ma\¢a m\m m\m m\¢ uH5p<
mz ma\oa o\m o\a m\a mo. ma\aa o\m o\m o\e mazes
msmoosma {M
mz mfi\oa o\m m\e ©\a mz mfi\oa m\m m\m m\m waspfi
No. ma\efi ©\m m\¢ m\¢ aoo. ma\ma m\m m\m m\m mssow
HUHHOQ 28 KM
msouw
.nonm Hmuos m z a .nonm Hmuos A s a nooo
umms mung umme manna

 

 

A.onflmnm unmmumnmunHBHB “omflnmn¢v

.ummu
wnoquEsmsoo mnu mo mso hop ocflnsp mHOanoo CMSu Hopo pmnoflpflpsoo 03p Eoum
Oman Hflwnu mo mmmuamonom Hmummsm m mum pony Nam m0 u50.mHmEHem mo HOQEDZ .mfi OHQOB

69

 

 

 

 

mz mfi\mfi m\m o\m w\w mo. mfi\¢a m\w w\¢ m\¢ wasps

mz ma\m o\o o\a ©\m mo. ma\¢a o\m o\m o\a memos
msmoosma 2M

mz ma\m m\m o\m o\a mz ma\oa o\m o\m o\m pasta

No. mfi\¢fi m\m m\m m\¢ No. mfi\¢a m\m m\w ©\¢ mcsow
HOHHOQ 2a 2M
msouw
.Qonm HmuOB m 3 4 .Qoum Hmuoa m 3 g Mono

some mpmq umme hanmm
A.wsflmum usmmumumpsHBHB “mmflcmudv .ummp

maoumfifismcoo msu mo Hsom hop meansp maonusoo Gmnp Mono UOCOHOHUCOU mnu Eonm
soap Hanna mo mmmpcmouom Hmummno m mew away me mo #50 mHmEHcm Mo HOQESZ .efi OHQOB

70

g, m, bairdi were not Significantly affected by the olfactory
conditioning. Both young and adult 2, leucopus Showed a
significant conditioning effect during the test period imme—
diately after the experience, but one month later the effect
was not Significant (Tables 15 and 14).

Figures 5, 4, 5, and 6 illustrate that the Significant
eXperience effects during the immediate test, particularly
with anise and wintergreen odored food, waned in magnitude in
the groups of mice tested one month after conditioning. In
some instances the effects of the eXperience were no longer

significant one month later, particularly in both young and

adult 3, leucopus, as was previously noted using the Sign

 

tests.

The results of both the appetitive and consummatory tests
are summarized in Table 15. Young g. m, bairdi were affected
by the early conditioning during both test periods and as
measured by both tests, whereas adult g, m, bairdi were not
Significantly affected by the eXperience. Both young and

adult 2, leucopus were Significantly affected by the eXperi—

 

ence as measured by both tests during the immediate test
period. During the test one month later, results showed that

both young and adult P, leucopus were significantly affected

 

as determined by the appetitive test, but not as determined

by the consummatory test.

 

71

Table 15. Summary of the effects of early eXperience.
Plus (+) indicates a Significant overall effect
and a minus (-) indicates no significant effect.

 

 

 

Appetitive Consummatory

Group Test (Test
P} m. bairdi

Young-immediate test + +

Young—late test + +

Adult-immediate test - -

Adult-late test - -
g, leuc0pus

Young-immediate test + +

Young-late test + -

Adult-immediate test + +

Adult-late test + —

 

STRATEGY

Purposes

The purposes of this experiment were: (1) to describe
the strategies used by g, m, bairdi and P, leuCOQus in ob-
taining food from three different food—odor combinations;

(2) to test the hypothesis that the mice would locate a
preferred food-odor combination by odor rather than position
habit; and (5) to test the reactions of the mice to a novel

food—odor combination.

Test Subjects

 

The mice used in this eXperiment were young and adult
F1WC g. m, bairdi and g, leucoEus. Four mice were tested in

each Species—age group.

Apparatus

 

The test apparatus was a wooden cage 1' by 1%” by 10"
deep with a clear Plexiglas top. Each of three Openings in
one side of the cage was fitted with a Short passageway con-
taining a swinging door. Removable food hoppers were at-

tached to the end of each passageway. Each hopper contained

72

 

 

 

75

a different food—odor combination. To obtain food, the

mouse swung open the door, activating a microswitch, which
remained activated throughout the period the mouse was eating
at the hOpper. The frequency and duration of feeding were
recorded on a Rustrak Model 92 event recorder (Rustrak Instru—
ment Company, Manchester, New Hampshire) with a chart speed

of one inch per hour, and with Sodeco four digit counters
(Landis and Gyr, Inc., New York, New York). Food consumption
was measured to the nearest 0.1 gram with a Shadowgraph scale
(Exact Weight Scale Company, Columbus, Ohio). The three
dependent variables were: (1) the number of times each food—
odor combination was visited; (2) the amount of food consumed
from each food—odor combination; and (5) the pattern and dura-

tion of visits to the different food-odor combinations.

Procedure

 

The mice were tested individually for a period of 26 days.
During the first 20 days of the test, three food hoppers with
anise, wintergreen and pine food-odor combinations were af-
fixed to the ends of the passageways. During the final Six
days of the test, the anise food-odor combination was replaced
by food with the odor of sassafrass oil.

At the beginning of the test the mouse and foods were
weighed. Each test day the foods were weighed, the number
of counts at each door were recorded, the event recorder

chart paper was marked with the time, and the positions of

74

the food hOppers were changed. The odor stimuli were re-
freshed every three days with new oil. At the conclusion of
the test the mouse was weighed again. Since weighing and
handling the mouse daily would probably have made the pat-
terns of feeding unrepresentative, the weights of the mouse
at the beginning and at the end of the test were averaged to

provide a mean body weight for the entire period.
Results

The percentage of the diet consumed from each of the
three food-odor combinations during days 1-20 was computed
for each day and correlations were run between the consumption
at each odor and the number of counts recorded from the
counter at that food hOpper. For each of the Sixteen mice
the correlation was highly Significant (.80 < r <:.92, DF=18).
The count data were, therefore, eliminated from the analyses.
The mean percentage of the diet eaten from each of the
food-odor combinations was computed for each group over the
first 20 days of the test period. These data are graphed as
bar histograms with standard errors in Figure 7. Mice in the
various groups did not differ significantly in the percentage
of the diet eaten from each of the food-odor combinations.
The standard errors for the four values at any one food—odor
combination were all overlapping. The histograms also show
that the mice in all four groups preferred the pine odored

food, as in the earlier preference tests.

 

75

Figure 7. Mean percentage (: 1 S. E.) of the diet consumed
from each of the three food-odor combinations
(A=anise; W=wintergreen; P=pine).

 

 

(iisu

MEAN PERCENTAGE

 

60

40

3O

10

2OJ

 

 

 

 

 

 

76

FIGURE 7

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ADULT

P ’ A

Tim-bairdi

w P A w P A w P
YOUNG ADULT YOUNG

P. leucopus

 

77

Strategy was analyzed by collecting the data from the
event recorder charts, and translating it onto sheets marked
off in hours and days. The amount of time Spent at each food-
odor combination was transformed to an ordinal scale by
rating the amount of activity each hour as follows: zero if
the mouse was not at the combination at all, 1 if the mouse
fed 0—5 minutes, 2 if the mouse fed for 5-15 minutes, 5 if
the mouse fed for 15-50 minutes, and 4 if the mouse fed for
more than 50 minutes of the hour at that combination. The
ordinal scale was chosen on the basis of the distribution of
the length of feeding activity periods. The data then
provided a running record of the amount of time spent per
hour at each of the three possible food-odor combinations for
each mouse. These records were analyzed to determine four
aspects of the feeding strategy.

(1) The number of days the mouse went to all three fOOd—
odor combinations during the initial activity bout (defined
as the active period prior to the first inactive period of
60 minutes or longer) was analyzed. The 20 day experimental
period was divided into the first ten days (1-10) and the
second ten days (11-20). The mean number of days that all
three food-odor combinations were visited during the initial
activity period was computed for each of the four groups of
four mice each. These means and standard errors are presented
in Table 16. A ratio was computed for each mouse by dividing

the mean number of days all three food—odors were visited

 

 

78

 

 

 

Mean Ratio

Table 16. Mean number (i 1 S. E.) of days, out of 10, that
all three food-odor combinations were visited
during the initial activity bout.

Species P, m, bairdi g,_leuc09us

Age Young Adult Young Adult

Time

Days 1—10 2.5(0.6) 5.0(1.2) 4.0(1.5) 5.5(0.6)

Days 11—20 6.5(0.9) 5.2(1.4) 7.0(1.5) 6.8(0.2)

1-10
11-20

 

0.46(0.18) 0.9700305) 0.56(0.15) 0.85(0.15)

 

 

 

79

during the first ten days, by the mean for the second ten
days. These ratios were then averaged for each group. The
mean ratios and their standard errors are Shown in Table 16.
An analysis of variance was performed on these ratios. The
main treatment factors were species and age, with four
replicates per cell (Table 17). Only the age factor was Sig—
nificant (F=8.88, DF=1,12, p <<.001). Young 3, m, bairdi and

g. leuCOpus did not visit all three food-odor combinations on

 

as many days during the first ten days of the test as did
adult mice. Table 16 shows that young mice did visit the
same mean number of combinations as adult mice during days
11—20.

(2) A mouse's use of a position habit was tested by re-
cording the initial feeding reSponse, each test day, after
the food hoppers had been weighed and replaced in different
positions. The response was recorded as a position habit
(P) if the mouse went to the door where its preferred food—
odor combination had been the previous day. The behavior was
recorded as a following reSponse (F) if the mouse went to the
preferred food—odor combination regardless of its position.
This implies that the mouse used the odor cue to locate the
combination it preferred. The reSponse was recorded as
neutral if the mouse went to the remaining position, neither
the positionhabit nor the fOllowing reSponse. Of 520 re—
sponses only 15 percent were neutral; this response was not

considered in the analysis.

 

 

 

80

Table 17. Analysis of variance on the food-odor combinations
visited during the initial activity bout.

 

 

 

Factor DF Mean Square F Prob.
Species 1 0.002 0.05 NS
Age 1 0.615 8.88 .001
Species x Age 1 0.062 0.90 NS
Error 12 0.069

Total 15 (ss=l.505)

 

 

 

 

81

The data were divided into two time blocks; the first
ten days of the test (1-10) and the second ten days (11-20).
The data were computed as a ratio for each mouse; the number
of position responses divided by the number of following
responses. The mean ratios and standard errors for each
species and age group are shown in Table 18. These ratios
were subjected to a three—way analysis of variance with one
measure repeated, and with four replicates per cell. The
main factors were age and Species and the repeated measure
was time blocks.

The analysis (Table 19) showed Significant species dif-
ferences (F=5.01, DF= 1,12, p <f.05), and a Significant
trials effect (F=12.49, DF=1,12, p <:.005). There were no
significant age or interaction effects. During the first ten
days 3. m. bairdi had a very high P/F ratio (>-5.4), indicat-
ing a strong use of position habit. During the second ten
days the ratio for g, m. bairdi was less (P/F=0.4 and P/F=1.1),
indicating the mice were using both position habit and follow-
ing responses. Young and adult g. leucopus used the position
habit only slightly more than the following response during
the first ten days, and switched to the following reSponse
during the second ten days of the test.

(5) Differences in feeding strategy were analyzed from
the number of food-odor combinations visited per hour of
activity per mouse per day. These were averaged for the four

mice in each test group for four—day blocks (1—4, 5-8, 9-12,

 

 

 

82

Table 18. Mean ratio (i 1 S. E.) of position responses to
following reSponses.

 

 

Days 1—10 11—20

Group N

 

g. m, bairdi

Young 4 5.4(1.2) 1.1(0.5)

Adult 4 4.6(2.0 0.4(0.l)

g. leuCOpus
Young 4 1.5(0.1) 0.5(0.1)

Adult 4 1.9(0.7) 0.6(0.2)

 

 

 

85

Table 19. Three—way analysis of variance with one measure
repeated for the ratio of position responses to
following responses.

 

 

 

Factor DF SS MS F P
Between Subjects 15 50.52

Age 1 0.98 0.98 0.54 NS
Species 1 14.58 14.58 5.01 .05
Species x Age 1 0.08 0.08 0.05 NS
Error Between Subjects 12 54.88 2.91 I
Within Subjects 16 89.85

Trials 1 57.84 57.84 12.49 .005
Trials x Ages 1 2.65 2.65 0.87 NS
Trials x Species 1 8.82 8.82 2.91 NS
Trials x Species x Ages 1 1.12 1.12 0.57 NS

Error Within Subjects 12 59.59 5.05

Total 51 140.57

 

 

 

 

84

15—16 and 17-20). This meant that there were five (trials
1—5) four-day blocks during the 20—day test. The maximum
number of food-odor combinations that a mouse could visit

in one hour would be three. Thus, the maximum value that
the mean could have would be 5.0. The higher the mean value,
the greater the tendency for the mice to visit all three
combinations during each hour of activity.

The means and standard errors for these data are pre—
sented in Table 20. These data were subjected to a threeeway
analysis of variance with one measure repeated and four
replicates per cell. The two main factors were Species and
age and the repeated factor was trials. The analysis (Table
21) showed that the age factor approached Significance (F=4.55,
DF=1,12, .10 >-p >>.05). The trials effect was Significant
(F=4.24, DF=4,48, p <:.001). Duncan's New Multiple Range
test breakdown of these means (Table 22) showed that trials
1 and 2 were significantly different from trials 5, 4, and 5,
and trials 2 and 5 were not different from each other. The
breakdown of the trials by age interaction (F=4.51, DF=4,48,

p <:.001), showed that the most critical difference in the
experiment was that young mice on trials 1 and 2 were visiting
Significantly fewer food—odor combinations than adult mice on
all trials and young mice on trials 5, 4, and 5 (Table 22).
This means that young mice visited fewer food—odor combinations
per hour during the initial 8 days of the test than adults,

but during the 12 days follbwing, the mice of all ages and

 

 

 

85

 

 

 

 

Afifi.ovm>.d ANfi.ovmw.fi Amo.ovaw.fi Ama.ovmm.fi m owlba
A©0.0VO>.a Amfi.ovm>.fi Amo.ovm©.a AOfi.ovm>.fi w wanma
Afifi.ovmw.d Ama.ov>>.fi Aaa.ovow.fi Awfi.ov>>.fi m Nfilm
AOd.ovom.a Ama.OVMd.fi Amfi.ovmm.fi Amo.ovdd.a N mum
Afifi.ov>>.a Amo.ovfim.fi Amo.ovmw.fi Afia.ov¢m.fi a «la
HmflHB meQ
waspd mcsow pasbd onsow wm<
mdmdwmwfl .M mmmmmm .8 .M mmflowmm

.>wp Mom omDOE Mom >DH>HOOM
mo H503 Mom Umpflmfl> mCOHDMCHQEOO Honoupoom mo A.m .m a HV NOQEDC snmz .ON OHQOB

 

 

 

86

.uoonuoo mum mosam> Ononp
new moEHu amno>0m pmxoosolon mm3 mammamsm one .BOH maamsmscs who monam> m Ononam

 

 

maa.m om Hmuoe

Nwo.o ooo.m we cflnuflz nonnm

mz mm0.0 >00.0 >N0.0 w 004 x moflowmm x mHmHNB
mz mao.o moo.o moo.o a monommm x mannns
aoo. am.a ama.o mms.o a wad x mannns
aoo. am.e msa.o oas.o a mamnue
moa.m em muomflnsm canons

oma.o mme.a ma cwmapwm nouns

mZ mmo.0 000.0 000.0 a moaommm x 00¢
oa. V av mo. mm... $20 $90 a mm...
m2 mfi.0 Mm0.0 mm0.0 a. wwfloomm
amm.a ma mnomflnsm cmmsnmm

m m m2 mm mm Hovomm

 

 

.HDOS Hem pouflwfl> mnoflumcHQEoo
MOUOIUOOM esp so pmpmwmmn onsmnofi 0:0 SOHB mosmHHn> mo mfimmamnm xmBIOOHQB .fim OHQOB

 

 

 

87

 

 

mm.fi am.fi Om.fi >>.fi >>.fi 0b.fi m>.fi 0>.fi N¢.fi hm.fi
m N m w m m a w N a
< < d w W M m 4 N M

waspmu< masomnw

 

0®.fi w>.a m>.a N0.d 0m.fi COOS
m w m N a .02 HOHHB

COOS
HOHHB
00¢

mod x mHmHHB ANV

mamnus lav

 

 

.Amo. V my usmHOMMHn waucmOHchmHm mum mafia
0EMm may m9 pmocwpnsw uon memos omOQB .pmma omcmm OHQHDHDE 302 w.nmon:0 .NN wanna

 

 

 

88

species performed at the same level, visiting on the average
1.7 to 1.8 food-odor combinations per hour of activity.

(4) The data were also analyzed by counting the number
of times each mouse switched food-odor combinations per day.
A switch was recorded every time a mouse fed for five
minutes at one hOpper with no excursions to other food-odor
combinations, followed by at least five minutes feeding at

a different hOpper. There was no prescribed time limit for

 

the interval between feeding at the two different food-odor

 

combinations. Intermittent feeding at the same food-odor
combination would, therefore, not involve any switching
behavior.

The data were divided into two ten-day blocks (1-10 and
11-20). The mean number of switches per day was calculated
for each mouse for each ten-day block, and was used as the
dependent variable. The species and age group means and
standard errors are presented in Table 25. These data were
subjected to a three-way analysis of variance with one
measure repeated, and four replicates per cell. The main
factors were age and Species and the repeated factor was
10-day blocks. The analysis (Table 24) showed that only the
trials effect was significant (F=8.01, DF=1,12, .01 < p 41.025).
This difference was due to the fact that all of the mice
Showed more switching behavior during the second ten days of
the test. The Species difference approached significance

(F=4.55, DF=1,12, .05 <Ip <(.06). A Duncan's New Multiple

 

 

 

89

Table 25. Mean number (i.1 S. E.) of switches between food-
odor Combinations per day.

 

 

Days 1-10 11—20

GrOUp N

 

P. m, bairdi

Young 4 1.5(0.2) 2.0(0.5)

 

Adult 4 l.3(0.1) l.8(0.5)

g, leucopus
Young 4 1.6(0.1) 2.4(0.5)

Adult 4 5.9(0.4) 4.0(0.4)

 

 

 

90

 

 

 

Nm.fim fin HMDOB

ma.o os.m ms canons uonum

mz ma.o oo.o oo.o a mamnnn x monooom x ooa

mz m>.0 mm.0 mm.0 a wHMHHB x mwfloomm

mz oa.o no.0 mo.o a unease x om<
mmouv.oo.ao. ao.o mm.m mo.m a unanus
aa.oa ma uuooflnsm canons

No.m am.am ma coosuom nouns

oz as.m mm.m mm.m a monooom x mom

oa.v dovmo. no.4 aa.m aa.m a monowom
oz oa.a am.m am.m a om<
am.aa ma unconnsm cooSnom

m m m2 mm ma Honomm

 

 

.wop Mom wsoflpmcHQEoo Hopo
IUOOM cwwaoQ monouHBm mo H0955: onu mo oosmflun> mo mflmhamcm >m3loosne .wN manna

tiII. I(III.II..III.IIIIIII

 

 

91

Range test (Table 25) showed that the adult 3. leucopus
switched food sources Significantly more often than any of
the other three test groups, which did not differ from each
other.

On day 21 of the test, a food hopper with sassafrass
oil as the odor stimulus replaced the anise food—odor combi-
nation. The last Six days of the test measured the responses
of the mice to this novel food—odor combination. The data
were computed as percentages of the diet eaten from the
novel combination, for the first day of the test and for all
Six days of the test combined. The means and standard errors
are Shown in Table 26. The reSponse did not change from the
first day to all six days combined. A two-way analysis of
variance was performed on the data for all Six days (Table
27). The main factors were species and age with four repli—
cates per cell. A significant Species by age interaction
(F=7.68, DF=1,12, .01 <:p <:.025) was broken down using
Duncan's New Multiple Range test (Table 28). This Showed that
the difference was due to the avoidance reaction of the young
2. m. bairdi. The young g. m. bairdi consumed only 18.8 per-
cent of their diet from the sassafrass odored food, while
young g. leucopus and adult mice of both Species ate at least
50 percent of their diet from the novel combination during

the six test days.

 

 

 

 

92

Table 25. Duncan's New Multiple Range test. Those means not
subtended by the same line are Significantly dif—
ferent (p <<.05).

 

(1) Species X Age

Species g..m. bairdi g. m, bairdi g. leuc09u3~g. leucopus
Age adult young adult young
Mean 1.55 1.75 2.00 5.95

 

 

Table 26. Mean percentage (i 1 S. E.) of the diet eaten from
the novel food—odor combination.

 

Group First Test Day All Six Test Days

 

g. m, bairdi
Young 21.0(7.9) 18.8(8.0)

Adult 76.5(14.0) 64.9(17.0)

g. leucoEus
Young 74.1(16.1) 61.8(5.6)

Adult 55.5(19.4) 50.1(7.1)

 

 

 

95

Table 27. Two-way analysis of variance of the percentage of
the diet eaten from the novel food—odor combina—
tion during all Six test days.

 

 

 

Factor DF MS F P
Species 1 791.02 1.81 NS

Age 1 1185.09 2.71 NS
Species x Age 1 5555.29 7 .68 .01 < p <.025
Error 12 456.91

Total 15 (ss=10574.52)

 

Table 28. Duncan's New Multiple Range Test. Those means
not subtended by the same line are signifi-
cantly different (p «(.05).

 

 

(1) Species x Age

Species g. m. bairdi P, leucopus g. leucopus g. m. bairdi
Age young adult young adult
18.8 50.1 61.8 64.9

 

 

DISCUSS ION

At the outset it was proposed that variation in food
habits among species of small mammals are attributable to
differences in the inherited capacity of each Species to
be affected by ontogenetic experiences. Variation was in-
vestigated with regard to fixity of Species food-odor
preferences, modifiability of food-odor preferences, and
the strategies used by the mice to locate food. The princi—
ple conclusion was that similar olfactory eXperiences dif—
ferentially modified the food—odor preferences of the two
Species. Using a restricted set of food cues, g. leucopus
showed more flexibility in its feeding behavior than g. m.
bairdi. Since both Species showed the same pine preference,
and since the same cues were used to study modifiability
and strategy in both Species, there must be an inherited
Species difference in the capacity to reSpond to these
stimuli.

Results from both the experience and strategy studies
support the conclusion that g. leucopus are more flexible
in their feeding behavior than g. m. bairdi. While both
young and adult g. leucopus were significantly affected by

olfactory conditioning, only young g. m. bairdi shifted

94

 

 

95

food-odor preferences in accord with the eXperience treatment.
Analysis of the feeding strategies indicated that young
g. m, bairdi avoided a novel food-odor combination, whereas
the other three test groups all consumed more than 50 percent
of their diet from the novel source. Thus, even early in
development the g. m. bairdi are more restricted in their
feeding behavior. At maturity g. leucopus switched feeding
sites more frequently than g. m. bairdi, and the g. leucopus
initially adopted a strategy using both position habit and
following responses, whereas 2. m. bairdi used a position
habit almost exclusively.

I will now briefly discuss the conclusions and explana—
tions for each separate experiment, to relate these findings
to the overall theme of species variation in feeding

behavior.
Genetics

The central question in the first experiment was whether
the food—odor preferences of the mice were inherited. Three
approaches were taken: (1) a deScription of odor preferences
for the various Species and age groups; (2) a comparison of
longitudinal and cross-sectional developmental groups of
young mice; and (5) a test for possible relaxed selection for
the odor preferences in domesticated stocks of mice.

The results (Table 1 and Figure 2) Showed a consistent

pine preference in all experimental stocks. There are

 

 

96

several explanations for these results. (1) The pine prefer-
ence may be a generalized trait inherited by all the mice.
(2) The pine Shavings on which the mice were housed prior to
eXperimentation may have produced a consistent experiential
effect resulting in the uniform pine preference. A test of
this hypothesis (Table 5) Showed that the pine shavings in
the cage did not affect the odor preference pattern.

(5) Lastly, the dam's milk may have contained some factor
which affected the olfactory preferences of the mice. This
hypothesis could be tested only if there were some means of
artificially feeding the mice from birth.

Longitudinal and cross-sectional groups of young mice
did not differ significantly in their odor preferences
(p > .10). The preference for pine odored food was immediate
and was not a function of maturation. The immediate (day 21)
preference for pine odored food could be used as an argument
for a genetic component in the olfactory preferences.
Further study is necessary to determine the odor preferences
of the mice from birth until 21 days of age.

The hypothesis that laboratory confinement relaxed
selection pressures on the odor preferences of domestic
stocks of g. m. bairdi and g. m. gracilis was not supported
by the data (Figure 2). Differences between wild caught and
domestic stocks would have provided evidence for the overall
genetic hypothesis. Alternative explanations for these re-

sults include: (1) selection pressures were not relaxed

 

 

97

under laboratory conditions which maintained the selection
for odor preferences, and (2) the odor preferences are en-
tirely eXperiential.

Individual differences in olfactory preferences provided
the best evidence for genetic components. Eleven of 96 mice
tested showed clear preferences (c.a. 80 percent of the diet)
for the anise or wintergreen odored food. Individual differ—
ences have been used previously as an argument for inheritance
of behavior patterns (Free, 1958; Hamburg, 1967; Vale and Vale,
1969). Experimental tests of the genetic hypothesis would
involve selectively breeding mice that showed a fixed prefer—
ence for anise or wintergreen (or even pine) odored food.
Selection for a particular food-odor combination would con-
firm a genetic hypothesis. In a Similar type selection study
Nachman (1959) showed that saccharin preference in rats was
inherited.

The major conclusion from this first experiment was the
consistent preference pattern for the pine odored food.

This preference may be inherited, but only as a generalized
character. Genetic variability among the three odors used

was low and not systematically related to Species, age, or
domestication. The consistent baseline was important for
interpretation of the experience manipulation and for describ-
ing age and Species Specific feeding strategies. It will

be necessary to conduct further investigations, such as the

selection eXperiment, to determine the extent of genetic

 

  

 

 

98

variation in these species and the degree of inheritance

of these food—odor preferences.

Experience

Against this consistent background, eXperience was
manipulated to determine the modifiability of the odor
preferences. The results of this experiment are summarized
in Table 15. The two Species differed in the degree of
preference modification. The key group in this species
variation was the adult g. m. bairdi. The olfactory prefer-
ences of both young g, leuc0pus and P, m. bairdi were sig—
nificantly (p <<.05) modified by early experience. This is
consistent with most previous findings on early experience
effects (see Beach and JayneS, 1954; Scott, 1962). Adult
mice of these two species differed radically in the modifi-
ability of their olfactory preferences. Adult g. leuCOpus
were significantly (p <<.05) affected, but g. m. bairdi
adults showed no modification of odor preferences. The only
odor to have a Significant effect on the adult 3. m. bairdi
was pine, but there was already a strong baseline prediSposi—
tion to prefer pine odored food. There are two hypotheses
which may explain this adult variation between the two
species: (1) an age Specific effect, with rigidity of food
habits developing differentially in both Species, and
(2) species differences in the perception of the laboratory
experiences which lead to rigidity or maintain flexibility of

food—odor preferences.

 

 

 

99

(1) Two parameters were manipulated in testing the
effects of olfactory experience; age at conditioning and
age at testing. It is my hypothesis that the age of experi—
ence was the critical variable. In support of my position,
mice 66 days of age (see Table 8) were morphologically and
reproductively adult. Also, outside of the experience pro—
vided in the study, mice of all ages were exposed to the same
uniform environment, the only variable being the length of
exposure to that environment. It could also be argued,
however, that the two weeks of olfactory conditioning consti—
tuted a larger proportion of the total experience of mice
only 66 days of age than for mice 105 days of age or older,
in which case, the age of testing might be important.

Assuming that the age of experience was the critical
variable, the overall effect may be described as a probability
curve with age as the abcissa and the probability of being
reinforced for selecting any food—odor combination and learn—
ing the food—cue association as the ordinate. For many
behaviors, in a variety of animals, this probability curve
is higher in young animals (see Scott, 1962; King, 1968b),
and declines with age. In this study P. p, bairdi appeared
to follow this pattern, with high modifiability in young mice,
but virtually no Shifts in adult preferences where the prob-
ability curve had declined to a low level. The probability
of learning the food-cue association was uniquely identical

for both young and adult g. leucopus.

 

 

 

100

This Species variation is supported by two further
points of evidence. One, already discussed was the relative-
ly high fixity of the pine preference in adult g. m, bairdi
contrasted with lower fixity in young g. m. bairdi and young
and adult 2. leucopus. The second point comes from the
deaths recorded during the conditioning phase. Among the
g. leucopus, the deaths (59) were equally divided between
young (18) and adult (21) mice. However, for g. m. bairdi
42 (82 percent) of the 51 deaths occurred in the adult test
group, indicating that young mice associated new odors with
food more readily than adults.

(2) The laboratory environment may have provided a
sufficiently broad set of experiential stimuli for adult
g. leucopus to retain a high degree of flexibility in their
food-odor preferences. The same laboratory environment may

not be broad enough for g. m. bairdi, so at 90 days of age

 

they are no longer very flexible in their food—odor preference.

Testing this hypothesis would require manipulation of the
perceptual worlds of the mice to provide young 3. m. bairdi
with a wider range of stimuli prior to the olfactory condi—
tioning experience. The Species differences in preference
modifiability could also have resulted from an interaction of
a Species aging effect (1) and the differential interpreta-
tion of the breadth of the experiential world provided by

the laboratory (2).

 

 

101

The fact that young and adult g, leucopus showed sig-
nificant experience effects during the late test period in
the appetitive test, but not in the consummatory test
(Table 15) also deserves eXplanation. Two alternative ex-
planations are: (1) the effect of the eXperience waned
during the 50 day interval between conditioning and testing;
and (2) there was some difference between the two tests.

Since the interval was the same for both tests, the second

 

alternative appears more reasonable. In the appetitive test
mice were forced to select a position at which to chew
through the sticks to obtain food. In this test there was

no immediate food reinforcement. In the consummatory test
the mice were free to sample all three food-odor combinations
immediately. The interaction of the 50 day interval between
conditioning and testing with the nature of the test may
explain the nonsignificant results obtained for these two

grOUpS in the late consummatory test.

Strategy

The last experiment was designed to investigate species
23nd age differences in feeding strategies of the mice. The
Inajor conclusions of this study SUpport the species variation
:in feeding behavior observed in the study of eXperience.

(1) Young mice adOpted a conservative strategy for the
ifirst half of the test, feeding each day at the first food-

Cndor combination encountered which provided nutritional

 

 

 

102

reinforcement. During days 11—20 of the experiment the
experiential worlds of the mice expanded, and so did their
strategy. Measures of the number of food—odor combinations
visited during the initial activity bout and during each
hour of activity increased during days 11—20. As I have
already discussed the adoption of a conservative strategy
by young mice would strengthen their early feeding habits
which were restricted to one or several foods.

(2) Young g. m. bairdi reverted again to a conservative
strategy when a novel food-odor combination was introduced.
This group consumed only 18 percent of its diet from the
sassafrass odored food, whereas each of the other three test
groups (adult g. m. bairdi; young and adult P, leucopus)
consumed more than 50 percent of their diet from the novel
combination (Table 26). Thus, the Species variation in feed—
ing behavior is already manifested early in the ontogeny of
the mice.

(5) Species differences in feeding strategy were also
Shown with adult mice. Two of the measures tested, use of
a position habit and switching behavior, support this state—
ment. In contrast measures of food-odor combinations visited
initially each day and per hour of activity Showed no species
variation. The adult 2. p. bairdi strategy shifted drama—
tically from a position habit during days 1—10 to a following
response during the second ten days of the test. Adult

P. leucopus Showed a similar shift, but the magnitude of the

 

 

 

 

105

change was much reduced. This Species difference cannot be
explained from these data unless g. m. bairdi had stronger
position habits during days 1—10.

Adult 3. lpucopus switched feeding sites signigicantly
more frequently than any other test group (Table 25). In
keeping with a conservative strategy, young mice of both
Species would be expected to switch feeding Sites less fre—
quently than adults, which they did. But, adult g. m. bairdi
exhibited the least amount of switching behavior of all the
test groups. This finding supports the conclusion that the
feeding behavior and strategy are more flexible in adult
g. leucopus, in agreement with the results of the experience
study.

Since my eventual goal is to relate food habit to dis-
tribution, I will briefly reiterate what is known of the
distributions and food habits of g. m. bairdi and g. leucopus.
3. my bairdi is found exclusively in grassland, cultivated
or open field habitats, while g. leucopus is found in a much
wider variety of habitat types (Nicholson, 1941; Blair,

1940; Linduska, 1950; Whitaker, 1966; personal trapping
records). g. leuc0pus is characteristically found in wood—
land habitats, but it is also readily trapped in fields of all
types, in brushy areas, and in and around buildings. The only
complete study of the food habits of these two species
(Whitaker, 1966) showed that g. leucopus tended to select

a more varid diet, particularly if all types of habitats were

considered.

 

 

 

 

104

From the distributions of these two Species I would
hypothesize that g. leucopus would exhibit a wider variation
in feeding behavior. For the parameters I investigated,
this hypothesis is valid. Young mice of both Species were
affected by early experience, but even at an'early age
3. m. bairdi avoided new foods. In adult mice there were
distinct species differences in the flexibility of feeding
habits. Adult 2. m. bairdi Showed a higher fixity for pine
odored food than any other group, they switched feeding
sites less often than any other group, and olfactory condi—
tioning experience did not signigicantly alter their food-
odor preferences. In contrast g. leUCOpus retained their
flexibility, switched feeding sites frequently and showed
strong modifiability of food—odor preferences. This argument
has proceeded from the distribution to the types of feeding
behaviors shown by the mice. A more meaningful approach
would be to argue from the behaviors of the mice to the
distributions observed in nature.

It appears from the current study that the degree of
variation exhibited by a species in its feeding behaviors,
might be a predictor of its distribution. That is, a Species
which, as a population, exhibits a wider spectrum of varia—
tion in its feeding habits could occupy more types of habi—
tats. The present study has determined that Species do vary
in the degree of variation exhibited in feeding habits and

the present study has tested several factors which can affect

 

 

 

 

 

105

this variation. Now, a proposed direct test of the hypothesis
that feeding behavior affects distribution may be presented.

Since I am concerned with food habits and distribution,
I will assume, for the purposes of this experiment, that the
mice live in food worlds. The critical question is whether
the species variation in feeding behavior in one world can
be used as a predictor of the number of different worlds in
which the mice will live. A series of interconnected pens
could be set up with, for example, four food types in each
pen. As one set of independent variables the types (quality)
and quanities of the foods available in each pen could be
manipulated. Two Species (g, leucopus and g. m} bairdi)
and two rearing treatments in a cross—classified design will
be used to produce the mice to be tested in the apparatus.
The two rearing treatments would provide the mice with either
a restricted or a varied early feeding environment. The two
dependent variables would be the amount of time spent in
each pen and the food consumption at the various food types
in each pen.

From the current study on Species variation in feeding
behavior I would predict the following results for this
suggested experiment: (a) g. leucopus will distribute their
time evenly among the different pens, while g. E. bairdi
will be restricted to fewer pens; and (2) within each pen
3. leuc0pus will select a more varied diet than g. m. bairdi.
The early eXperience treatment would be expected to signifi-

cantly modify the food preference pattern.

 

 

 

 

SUMMARY

1. It was prOposed that variation in food habits among
Species of small mammals are attributable to inherited
capacities which are affected by ontogenetic experiences

with food cues.

 

2. Peromyscus maniculatus bairdi and P. leucopus nove-

boracensis were used to examine preferences for three food-

 

odor combinations made from three essential oils combined
with laboratory chow. Food-odor preferences were determined
for three taxa, two stocks of different breeding history
(wild caught and domestic), and two age groups. All groups
exhibited a strong preference for the pine food-odor combi—
nation and there were no significant differences among the
groups.

5. The mice did not: (1) discriminate odor intensities;
(2) use a dish preference; (5) discriminate soiled from un-
soiled food; (4) prefer pine because of pine shavings present
in their rearing cages; nor did their preference change over
seasons.

4. Mice conditioned for two weeks to associate one of
the three odors with food and control mice given laboratory

chow without the experimental odors produced the following

106

 

 

 

107

results in an appetitive and consummatory test given imme-
diately after and one month after the olfactory conditioning
eXperience:

a. Young 3, m, bairdi and young and adult P} leuc0pus
were significantly affected by the olfactory eXperi—
ence, but adult g, m. bairdi were not.

b. Groups of mice that showed a significant effect dur-
ing the immediate test also Showed a Significant ef—
fect when tested one month later.

5. Feeding strategies were tested in young and adult
mice of both Species by recording the duration, frequency,
sequences, and amount of food consumed at the three food-odor
combinations.

a. Young mice initially adOpted a conservative feeding
strategy, but changed to that strategy used by all
mice later.

b. Young E, m. bairdi avoided a novel food—odor combi-
nation, whereas the other test groups consumed more
than 50 percent of their diet from the new source.

c. Both species Shifted from a position habit (days
1-10) to a following response (days 11-20), using
an odor cue to locate food.

d. Adult g, leucopus switched feeding Sites more fre-
quently than any other test group.

6. It was concluded that g. leucopus showed more flexi—

bility in feeding behavior than g. m. bairdi.

 

 

 

 

 

108

a. Food habits were discussed in relation to the dis—
tributions of the mice.

b. The variation in feeding behavior Shown by different
species may be a predictor of the number of habitats

it will occupy.

 

 

 

BI BLIOGRAPHY

 

 

 

BIBLIOGRAPHY

Barnett, S. A. 1955. Problems of food selection by rats.
Anim. Behav. 1:159-164.

Barnett, S. A. 1956. Behavior components in the feeding
of wild and laboratory rats. Behaviour 9:24-45.

Barnett, S. A. 1965. The Rat. Chicago, Aldine Publishing
Co., 288 pp.

Barnett, S. A. and M. M. Spencer. 1955a. Responses of wild
rats to offensive smells and tastes. Anim. Behav.
1352*57.

Barnett, S. A. and M. M. Spencer, 1955b. EXperiments on
the food preferences of wild rats (Rattus norvggicus
Berkenhout). Jour. Hygiene 51:16—54.

Beach, F. A. and J. Jaynes. 1954. Effects of early eXperience
upon the behavior of animals. Psychol. Bull. 51:259-265.

Blair, W. F. 1940. A study of the prairie deer-mouse popu-
lations in southern Michigan. Amer. Mid.-Nat.
24:275—505.

Bronson, G. 1966. Evidence of the lack of influence of
early diet on adult food preferences in rats. Jour.
Comp. Physiol. Psychol. 62:162—164.

Burghardt, G. M. 1967a. Chemical clue preferences of inex—
perienced snakes: comparative aspects. Science 157:
718-721.

Burghardt, G. M. 1967b. The primacy of the first feeding
eXperience in the snapping turtle. Psychonom. Sci.
7:585—584.

Burghardt, G. M. and E. H. Hess. 1966. Food imprinting
in the snapping turtle, Chelydra serpentia. Science
151:108-109.

Chitty, D. and M. Shorten. 1946. Techniques for the study
of the Norway rat. Jour. Mamm. 27:65-79.

Craig, W. 1918. Appetites and aversions as constituents
of instincts. Biol. Bull. 54:91-107.

109

  

IIIHIIIIHHI

 

 

 

 

110

Dice, L. R. 1922. Some factors affecting the distributions
of the prairie vole, forest deer mouse and prairie deer
mouse. Ecology 5:29-47.

Dix, M. W. 1968. Snake food preferences: Innate intra—
Specitic geographic variation. Science 159:1478—1479.

Drickamer, L. C. 1970. Seed preference in wild caught
Peromyscus maniculatus bairdi and g. leuc0pus noevbora—
censis. Jour. Mann. 51:191—194.

 

 

Forgus, R. H. and D. E. Hutchings. 1960. Effects of early
experience on flavor preference. Psychol. Reports
6:410.

Free, J. B. 1958. The ability of worker honeybees (Apis
mellifera) to learn a change in the location of their
hives. Anim. Behav. 6:219-225.

Hamburg, D. A. 1967. Genetics of adrenocortical hormone
metabolism in relation to psychological stress. Chap.
8 in Behavior-Genetic Analysis, ed. J. Hirsch. New York:
McGraw—Hill Publications.

Hamilton, W. J. Jr. 1941. Food of small mammals of the
eastern United States. Jour. Mann. 22:250-265.

Hardy, R. 1945. The influence of types of soil upon the
local distribution of some mammals in southwestern Utah.
Ecol. Monogr. 15:71-108.

Harlow, R. F. 1952. Food preferences of the albino rat.
Jour. Genet. Psychol. 41:450-458.

Harris, V. T. 1952. An experimental study of habitat selec-
tion in prairie and forest races of the deermouse
gpgomyscus maniculatus. Contrib. Lab. Vert Biol. Univ.
Mich. 5631—55.

 

Holling, C. S. 1955. The selection by certain small mammals
of dead, parasitized and healthy prepupae of the European
sawfly, Neodiprion sestifer. Can. Jour. Zool. 55:404-
419.

 

Holling, C. S. 1958. Sensory stimuli involved in the loca—
tion and selection of sawfly coccoons by small mammals.
Can. Jour. Zool. 56:655—655.

Holling, C. S. 1959. The components of predation as revealed
by a study of the small mammal predation of the European
sawfly. Can. Entomol. 91:295-520.

 

111

Holling, C. S. 1965. The functional reSponse of predators
to prey density and its role in mimicry and population
regulation. Mem. Ent. Soc. Canada, 45:5-60.

Howard, W. E. and R. C. Cole° 1967. Olfaction in seed
detection by deer mice. Jour. Mamm. 48:147—150.

Howard, W. E., R. E. Marsh and R. C. Cole. 1968. Food de—
tection by deer mice using olfactory rather than visual
cues. Anim. Behav. 6:15—18.

Johnson, M. S. 1926. Activity and distribution of certain
wild mice in relation to biotic communities. Jour. Mamm.
72245-277.

King, J. A. 1958. Parameters relevant to determining the
effects of early eXperience upon the adult behavior of
animals. Psychol. Bull. 55:46—58.

King, J. A. 1968a. Biology of Peromyscus (Rodentig).
Manhattan, Kansas: The Allen Press Inc., 595 pp.

King, J. A. 1968b. Species Specificity and early experience.
Chapter 11 in Early Experience and Behavior, eds. G.
Newton and S. Levine. Springfield, Illinois: Charles C.
Thomas.

 

King, J. A. 1969. A comparison of longitudinal and cross-
sectional groups in the development of behavior of deer
mice. Annals New York Acad. Sci. 195:696-709.

King, J. A., E. 0. Price and P. Weber. 1968. Behavioral
comparisons within the genus Peromyscus. Papers Mich.
Acad. Sci. Arts and Letters, 55:115—156.

Klopfer, P. 1969. Habitats and Territories. New York:
Basic Books Inc.

Kuo, Z. Y. 1967. The Dynamics of Behavior Development.
New York: Random House, 240 pp.

Lack, D. 1955. Habitat selection in birds with special
reference to the effects of’afforestation on the Breck—
land avifauna. Anim. Ecol. 23259-262.

Lack, D. 1957. The psychological factor in bird distribu—
tion. Brit. Birds 51:150—156.

Li, J. C. R. 1964. Statistical Inference I. Ann Arbor:
Edwards Brothers Inc., 658 pp.

 

Linduska, J. P. 1950. Ecology and Land-use relationships of
small mammals on a Michigan farm. Mich. Dept. Conserva—
tion—Game Division, 150 pp.

112

Lorenz, C. 1960. Methods of approach to the problems of
behavior. The Harvey Lectures (1958-1959):60—105.

McClearn, G. and D. A. Rodgers. 1961. Genetic factors in
alcohol preferences in laboratory mice. Jour. Comp.
Physiol. Psychol. 54:116-119.

Mook, J. H., L. J. Mook and H. S. Heikens. 1960. Further
evidence for the role of searching images in the hunt—
ing behaviour in titmice. Arch. Neerl. 2001. 15:448—
465.

Nachman, M. 1959. The inheritance of saccharin preference.
Jour. Comp. Physiol. Psychol. 52:451—457.

Nicholson, A. J. 1941. The homes and social habits of the
wood—mouse (Peromyscus leucopus noveboracensis) in
southern Michigan. Amer. Mid. Nat. 25:196-225.

Price, E. O. 1967. Behavioral changes in the deermouse,
Peromyscus maniculatus bairdi after seventeen years of
domestication: Reaction to novel stimuli. Ph. D.
Thesis, Michigan State University.

Rabinovitch, V. 1966. The role of early experience in the
development of food habits in gull chicks. Bull. Ecol.
Soc. Amer. 47:144.

Rabinovitch, V. 1969. The role of experience in the develop—
ment and retention of seed preferences in Zebra finches.
Behaviour 55:222—256.

Richter, C. P. 1955. Experimentally produced behavior re—
actions to food poisoning in wild and domestic rats.
Annals New York Acad. Sci. 56:225—259.

Rodgers, W. and P. Rozin. 1966. Novel food preferences in
thiamine deficient rats. Jour. Comp. Physiol. Psychol.
6131-4.

Sheppe, W. 1961. Systematic and ecological relations of
Peromyscus oreas and g, maniculatus. Proc. Amer. Phil.
Soc. 1052421—446.

Sheppe, W. 1967. Habitat restriction by competitive exclu—
sion in the mice Peromyscus and Mus. Can. Field—Nat.
81:81—98.

Scott, J. P. 1962. Critical periods in behavioral develop-
ment. Science 1582949—958.

 

 

 

 

115

Siegel, S. 1956. Nonparametric Statistics for the Behavioral
Sciences. New York: McGraw Hill Book Co., 512 pp.

Siqueland, E. R. 1965. Experimental modification of taste
preference. Jour. Comp. Physiol. Psychol. 59:166—170.

Thios, J., J. Derivera and E. Aronson. 1962. Modification
of the rats saline intake by experience with Specific
concentrations. Psychol. Reports. 10:487-490.

Thompson, D. 1965. Food preferences of the meadow vole in
relation to habitat affinities. Amer. Mid. Nat. 7:76-86.

Tinbergen, L. 1960. The natural control of insects in pine
woods I. Factors influencing the intensity of predation
by songbirds. Arch. Neerl. 2001. 15:265-555.

Tinbergen, N., M. Impekoven and D. Frank. 1967. An experi—
ment on spacing out as a defence against predation.
Behaviour 283507—521.

Vale, J. R. and C. A. Vale. 1969. Individual differences
and general laws in psychology: A reconciliation. Amer.
Psychol. 2431095—1108.

Verts, B. 1957. The pOpulation and distribution of two
Species of Peromyscus on some Illinois strip-mined land.
Jour. Mamm. 58:55-59.

Warren, R. and C. Pfaffman. 1959. Early experience and
taste aversion. Jour. Comp. Physiol. Psychol. 52:265-
266.

Wecker, S. C. 1965. The role of early experience in habitat
selection by the prairie deermouse, Peromyscus manicu—
latus bairdi. Ecol. Monogr. 55:507—525.

Wecker, S. C. 1964. Critical periods for learning in the
deer mouse habitat selection. Bull. Ecol. Soc. Amer.
453145.

Welker, W. I. 1956. Effects of age and experience on play
and exploration in young chimpanzees. Jour. Comp.
Physiol. Psychol. 49:225-226.

Whitaker, J. O. 1965. Food of 120 Peromyscus leucopus
from Ithaca, New York. Jour.‘Mamm. 44:418-419.

Whitaker, J. O. 1966. Food of Mus musculus, Peromyscus
maniculatus bairdi and Peromyscus leuCOpus in Vigo
County, Indiana. Jour. Mamm. 47:475-486.

 

 

 

 

114

Wirtz, W. and P. Pearson. 1960. A preliminary analysis of
habitat orientation in Microtus and Peromyscus. Amer.
Mid. Nat. 65:151—142.

Young, P. T. 1952. Relative food preferences of the white
rat. Jour. Comp. Psychol. 14:297-519.

Young, P. T. 1955. Relative food preferences of the white
rat II. Jour. Comp. Psychol. 15:149—165.

Young, P. T. 1958. Preferences and demands of the white
rat for food. Jour. Comp. Psychol. 26:545-589.

Young, P. T. 1940. Reversal of food preferences in the
white rat through controlled pre—feeding. Jour. General
Psychol. 22:55-56.

Young, P. T. 1945. Studies of food preferences, appetite
and dietary habit V. Techniques for testing food prefer-
ence and the significance of results obtained with the
different methods. Comp. Psychol. Monogr. 1931-58.

Young, P. T. 1946. Studies of food preferences, appetite
and dietary habit VI. Habit, palatability and diet as
factors regulating the selection of food by the rat.
Jour. Comp. Physiol. Psychol. 59:159—176.

Zimmerman, E. G. 1965. A comparison of habitat and food of
two Species of Microtus. Jour. Mamm. 46:605—612.

 

 

 

 

 

 

 

HIWNIH

I
"h
T“
H
"Ill
H
“
A“
H

!

IIHIHNIN

3 1293 03062 08 7