SOCEAL GRGANEZAUON 0F WEE WESTERN FOX SQUIRREL

Thesis for the Degree of M. S.
MECHMN STATE UNWERSEY
ROBERT .EOHN BERNARD

1&79
VI in

u IIJ3II mnmwmflwmIllgmuyl ,2;

 

 

 

PLACE IN RETURN BOX to remove this chockout from your record.
TO AVOID FINES return on or before duo duo.

DATE DUE RATE DUE DATE DUE
39.29199: 1” ‘

 

 

 

 

   

 

V“: ‘/l ‘..r i 1‘5—
/L’"
:nAZZT

 

mm 5941 BP/RiiOésStZGIB

 

12m. . , , '

APR18200E
/ - D74 06 05

 

 

 

 

 

 

 

 

 

 

 

 

Tsunmm u — -~

 

ABSTRACT

SOCIAL ORGANIZATION OF
THE WESTERN FOX SQUIRREL

By

Robert John Bernard

During a two-year study of a discrete wild population of fox
squirrels, the existence of an intricate and stable hierarchal form
of social structure was documented. Animals were ranked by both
the percentile and tabular method of hierarchy delineation, based
on interactions obtained at an artificial feeder during consecutive
winters. The hierarchy was a straight line mixed sex arrangement
with males generally more dominant than females. Hierarchal status
advanced with increased age. Extent of minimum home range did not
increase with social rank. It was concluded that where food and
shelter are adequate, spatial requirements would not limit population
numbers.

Inventories of two discrete populations of fox squirrels were
made by trapping at a single central feeder during the time of winter
food scarcity. Population estimates based on a seven day trapping
period at the feeder compared favorably with the actual woodlot

squirrel populations.

SOCIAL ORGANIZATION OF
THE WESTERN FOX SQUIRREL

By

Robert John Bernard

A THESIS

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

MASTER OF SCIENCE

Department of Fisheries and Wildlife

T972

ACKNOWLEDGEMENTS

I wish to express special thanks to Dr. Leslie w. Gysel, who
while providing guidance and support allowed freedom to pursue those
aspects of the study which seemed most significant.

For critical review of the manuscript I am grateful to Dr. John
A. King and Dr. Rollin H. Baker.

Particular thanks is accorded wayne Schmidt and James Kessel for
the many cold hours they spent assisting with behavioral observations.

My deepest appreciation goes to my wife, Christine, for without
her enthusiastic moral and physical support, completion of this study
could not have become a reality.

This research was financially supported by the Agricultural
Experiment Station of Michigan State University.

 

ii

TABLE OF CONTENTS

Page
PART I
SOCIAL ORGANIZATION OF
THE WESTERN FOX SQUIRREL
INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . 1
MATERIALS AND METHODS . . . . . . . . . . . ........ 3
RESULTS . . . . . . . . . . . . . . . . . . ........ 7
DISCUSSION. . . . . . . . . . . . . . . . . . . ...... 20
LITERATURE CITED. . . . . . . . . . . . . . . . . . . . . . 26
PART II
APPLICATION OF'A SIMPLE METHOD
FOR CENSUSING FOX SQUIRRELS
INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . 29
MATERIALS AND METHODS . . . . . . . . . . . . ...... . 3l
RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . 35
DISCUSSION. . . . . . . . . . . . . . . . ......... 36
LITERATURE CITED. . . . . . . . . . . . . . . . . . . . . . 40

111

TABLE

LIST OF TABLES

PART I

Social hierarchy of lO fox squirrels ranked by the
linear method, Hudson Hoodlot. Michigan State

University, 197l. . . . . . . . . . . . ......

Social hierarchy of lO fox squirrels ranked by the
percentile formula. Hudson Noodlot. Michigan State

University, 197l. . . . . . . . . . . . ......

Social hierarchy of 13 fox squirrels ranked by the
linear method, Hudson Noodlot, Michigan State

University, l972. . . . . . . . . . . .......

Social hierarchy of 13 fox squirrels ranked by the
percentile formula, Hudson Noodlot, Michigan State

University, l972. . . . . . . . . . . .......

Average social rank for categories of fox squirrels.

iv

Page

12

T3

T4

15
18

FIGURE

LIST OF FIGURES

PART I

A basket feeder filled with ear corn was used to
draw squirrels for both trapping purposes and

behavioral observations. . . . . . . .......

Bob (tap) initiates an aggressive threat and
displaces the subdominant Oscar from his

position on the feeder . . . . . . . . ......

Dominants were intolerant of other squirrels
on the feeder tree; however. on the ground they

usually fed together peaceably . . . . . . . . . .

Effect of age on social rank for the l972 Hudson

Noodlot population of fox squirrels, (P<D.Ol). . .

Effect of size of minimum home range on social
rank for the l972 Hudson Noodlot population of

fox squirrels. (n.s.). . . . . . . ........

PART II

Fluctuation of the Toumey Noodlot fox squirrel
population showing estimate based on feeder

trapp1ng O O O O O O 0 O O O O O O O O O 00000

Fluctuations of the Hudson Noodlot fox squirrel
population showing estimate based on feeder

trapping . . . . . . . . . . . ..........

Page

TO

19

20

33

34

PART I
SOCIAL ORGANIZATION OF
THE WESTERN FOX SQUIRREL -

vi

SOCIAL ORGANIZATION OF THE WESTERN FOX SQUIRREL

The behavioral aspects of wildlife management have received
increased attention and application in recent years. Of particular
interest and importance are the conspecific interactions between
individual animals and their relationship to the surrounding
environment. The usual consequence of social interaction or
aggression is the formation of intraspecific territories or social
hierarchies (Davis 1964, Fisler T969). Bakken (I952), Fiyger (l955),
and Pack gt_gl, (l967), in their studies of gray squirrels, concluded
that this animal demonstrates dominance behavior that results in the
formation of social hierarchies. Allen (1943) concluded that fox
squirrels were probably not territorial to any extent and recommended
that more research be done on the social behavior of the animal.

Hhile involved in a population study by live-trapping fox
squirrels (ngugy§_niggg., the author observed a significant overlap
of individual home ranges as well as aggressive behavior among
individuals.<:This behavior suggested that the species maintains a
social system similar to that of gray squirrels (Sciurus carolinensis),
rather than a strong territorial form of behavior like that of red
squirrels (Tamiasciurus spp.) (Smith 1968). Research was initiated to
test the hypothesis that fox squirrels probably have a social system
similar to that of gray squirrels£:]

Allen (I943) recognized the necessity of knowing to what extent

an animal is (or is not) territorial in habit and what, if any.

restrictions this might impose on populations and hunting season
productivity.(:Stokes and Balph (1965) describe the intimate
association that animal behavior has with population ecology and
management of the animal and its habitat. They state that an
understanding of behavior is needed to determine how social inter-
actions within the population in turn affect the animals' ability
to reproduce and survive either directly or through changes in their
physiology.)

The author wishes to thank J. Kessel and H. Schmidt for their
assistance in making observations. Special thanks is accorded to
L. H. Gysel, who while providing guidance and support allowed freedom

to pursue those aspects of the study which seemed most significant.

STUDY AREA

The research was conducted in Hudson Noodlot, an 18-acre
woodland on the farm property of Michigan State University, where
hunting is prohibited. The study area is a closed, mature, all-
aged beech-maple stand with only a very small number of other
species present. The largest trees attain a D.B.H. of 30 inches
and provide an abundance of den sites. Only 10 acres were con-
sidered to be fair squirrel habitat, since the southern 8 acres
of the woodlot consist largely of pole-size trees which are less
frequently used by squirrels. Along the north and west boundaries
of the stand is cropland usually planted to field corn, while the
open areas near the southern and eastern boundaries are used as
pasturage. The distance to the nearest woodlot exceeds k-mile in

any direction, and the squirrel population was considered to be discrete.

The woodlot is only fair squirrel habitat. since there is a
paucity of mast-producing trees, with the exception of the

sporadically producing American beech (Fagus grandifolia)

 

(Gysel 1971). Here, as in much of southern Michigan, wintering
squirrel populations may in many instances be dependent on waste

corn left by the fall harvesting Operations (Allen 1943).

METHODS

Trapping activities began in January of 1971 and continued
until May of 1972. The observation periods each year were January
through April, the time when food was relatively scarce and animals
would readily come to an artificial source of food. Mean temperature
for the January to May period in southern Michigan is 27.0° F during
an average winter.

An artificial feeder (Sharp 1968), constructed with chicken
wire and filled with ear corn, was used to concentrate squirrels
both for trapping purposes and to increase the possibility of social
interactions (Fig. l). The feeder was placed in the woodlot in
December of 1970 and left there until the study was complete in
April of 1972. The feeder, which held a bushel of ear corn, was
hung 6 feet high on the bole of a large maple tree. Selection of
the tree was based upon two criteria: It was located in the
center of the best squirrel habitat on a tree with a large crown
intertwined with other nearby trees, allowing arrivals or departures
via this route if the animals so desired. After a period of one
month most squirrels in the woodlot were visiting the feeder.

Six live-traps were placed in a semicircle at the base of the

. - we: i9""'
‘2'." . m

 

 

 

Figure 1. A basket feeder filled with corn was used to draw squirrels
both for trapping purposes and behavioral observations.

feeder and were prebaited for 2 weeks prior to trapping -- a method
described by Sharp (1958). The traps were of a collapsible metal

(wire) variety which measured 9 x 9 x 24 inches. Success was
extraordinary, and nearly all of the population was captured in

7 trap-days. (Later observations disclosed one unmarked animal
identified by her characteristic pelage. This animal, dubbed Nancy.

was captured later after behavioral observations were complete and
trapping resumed.) As a result of the trapping in 1971. nearly all

the population had been captured and marked when behavioral observations
resumed in January of 1972.

Trapped animals were sexed. weighed and aged where possible
according to techniques described by Allen (1943). As Allen noted,
however, aging live fox squirrels is not always a clear-cut
proposition, so not all animals could be clearly separated as
juvenile or adult particularly during the winter. Age determina-
tions of adults were speculative and were assigned as 1.5+. In
1972, after year-long trapping activity, ages were known for a
greater proportion of the animals, although the rapid turnover of
squirrels kept the age of some a well-guarded secret.

A numbered aluminum tag was placed in each ear in the fashion
described by Sharp (1958), and each animal was dyed in a character-
istic pattern using Nyanzol A (Fitzwater 1943). Females were given
black tail tips, and all animals were named to facilitate recording
of data and subsequent reference. The markings proved indelible
and quite conspicuous, staining the fur a purple-black hue that
persisted until the spring-summer moult. Laboratory work on

animals was greatly simplified by use of the anesthetic metofane

(methoxyflurane), a relatively non-volatile compound widely used by
veterinarians for small animal surgery. Metofane, an anesthetic
with a wide safety range between anesthetic and toxic levels, has
been used with favorable results by other investigators (V. F. Flyger,
personal comm.).

Most observations were made from a blind constructed in a down
treetop 130 feet from the feeder tree. A spotting scape magnified
the animals when necessary to verify their markings. Later a tent
blind was erected only 20 feet from the feeder in order to document
social interactions on film.

Grid trapping operations were begun in April of 1971 and were
designed to unearth data on squirrel movements and home range. In
August of 1971, four additional feeders were placed in the woodlot
and trapping continued to further establish home range parameters.
Observation of squirrels immediately following release gave no
indication of den tree locations. Most animals seemed confused and
distressed, and did not move toward the same tree or even in the same
cardinal direction as in previous releases. Animals took refuge in
the nearest tree in some cases.
<::; Social interactions were based upon aggressive activity around
the feeder. An individual was Judged to be dominant if it displaced
a squirrel already at the feeder or if it failed to be evicted by
the interloper. If the newcomer was dominant. it was credited with
winning the interaction. Agytimes aggressive behavior was observed
among squirrels on the ground below the feeder.

Pack gt_al, (1967) and Ozoga (1972) utilized both the tabular

method of hierarchal arrangement illustrated by Davis (1964) and

the percentile rank formula employed by Komai §t_al, (1959) in
arranging their peck order data. Application of the percentile
rank formula of Komai et_al, (1959) required that each animal
listed interact with at least five other animals in order to be
ranked. (Of 12 animals trapped and marked in 1971, 10 satisfied
these requirements, while 13 out of 21 animals marked could be
ranked by this method in 1972. The higher the rank of a squirrel.
the lower the ranking number, e.g., the highest ranking squirrel
would be number one.) Both populations were also ranked by the
tabular method.
An attempt was made to measure the association, if any. of
home range with hierarchal rank. Home range determinations were
made using the ”minimum home range" method described by Mohr (1947),
for squirrels that had been trapped or seen in more than five locations.
The Spearman rank correlation (Seigel 1956) was used to determine
the degree of association between the tabular method and the percentile
rank method of hierarchy determination, and also to measure the
relationship of rank to age and home range.
Since weather conditions have been documented as a controlling
influence on fox squirrel activity (Hicks 1942), an effort was made in

this study to correlate squirrel activity with weather and social rank.

RESULTS
General Behavior:

Squirrels came to the feeder freely, and if uncontested most
animals hung by their hind feet or sat on t0p of the feeder while

eating kernels of corn removed with forepaws or Jaws. Characteristically,

the germ portion of the kernel was eaten, the remainder drOpped to
waiting birds or other hungry squirrels on the ground.(:The dominant
animal nearly always assumed and maintained a position on the feeder
and did not allow other animals access while he was in possession:)
(Fig. 2). 0n the ground, however, aggressive activity was displayed
less frequently, and as many as three or four animals might feed
peaceably all in close proximity to one another (Fig. 3). A squirrel
feeding on the tree seemed to enjoy relative security as well as
abundant food.)::hough apparently completely engrossed in feeding, a
slight movement from the blind would often frighten a squirrel from
the tree. This position gave the dominant squirrel a decided advantage
over his counterparts on the ground; although they were vigilant, it
was the squirrel on the tree which always took alarm first., The response
to an alarm stimulus was usually immediate departure without tail flick-
ing or barking. When the dominant departed, the animals on the ground
would then Jump to the tree to try to identify the source of alarm.:7
Although these animals were not hunted, they were exposed to
normal predation, since the red fox (Vulpes fulva), redetailed hawk

(Buteo Jamaicensis) and great-horned owl (Bubo virginianus) were common

 

in and around the woodlot. Presence of these predators might account
for the extreme wariness among squirrels which made observations
difficult. Some squirrels would flee when an observer approached
within 150 yards of the feeder. One squirrel was killed by a great-
horned owl in mid-morning, and another was found dead near a fox den

in the woodlot.

 

 

Figure 2. Bo a es an aggressive threat and displaces the
subdominant Oscar from his position on the feeder.

 

 

 

.2:

 

Figure 3. Dominants were intolerant of other squirrels on the feeder
tree, however on the ground they usually fed together
peaceably.

11

Social Behavior:

Based on interactions, a social hierarchy was determined by
ranking most animals who came to the feeder. Of 12 animals marked
in 1971, 10 were ranked on the basis of 72 interactions made during
135 hours of observation. In 1972, 13 out of 21 marked were ranked,
on the basis of 198 interactions during 101 hours of observation.
The hierarchies were straight-line arrangements involving both sexes,
similar to those discovered by Flyger (1955) and Pack gt_al, (1967)
for gray squirrels. The data are arranged by the tabular method
in Tables 1 and 3, and by the percentile method in Tables 2 and 4.
In the 72 interactions of 1971, there were no reversals, so
apparently the hierarchy was very stable and had been well established
when observations began. During 1972 two reversals were noted between
Howard and Simone, two squirrels adjacent to one another on the
hierarchy. These reversals did not necessarily reflect a change in
rank, however, for Howard later restored himself to his usual dominant
position. The correlation coefficient between the percentile and
tabular methods was .95 for 1971 and 1.00 for 1972, the first of which
is highly significant at the 1 percent level, the second a perfect

correlation.

Trap Response:

Sharp (1958) noted that the dominance hierarchy of gray squirrels
could be used to advantage when employing an artificial feeder for
trapping purposes. He stated that a dominant squirrel may never get
caught in a trap, for it remains on the feeder and few animals

(or none) can drive it from this position. During the 7-day trapping

12

 

 

 

o a a Pam excom op
F F a z mam sop m
P i- p a n. mg 2.5. N
N F N m a z mam mmpvz N
m _ i- N 2.- a h. m 352 m
N s... .3. P i- _ < z N «Home m
N N p a -i- --- F a u N ocean e
m -i- -u- a N p -i- N < z mmm gnome m
2 N P m N 1- N m P < z m 2.3 N
NN N N m N p F m o N < 2 pp nom p
.02 xcmm
cacom eohi xswmx mmrr21.»o=~zr mmgomu. ocean coumo acme, gem, mm< xmm mnhi osmz baboon
zoz
monhoaamczH “xv 4<=QH>HQZH >m mmamoeua> ao >uzmsomma “Xv Sumxaaom S<DQH>HQZH

 

 

_ .FNoF .zuwmgo>r::
ouaam camrcurx .uopuooz comes: .uozuoe Lumcpp one an umxcac apoccrscm xom op we xgugucmv; paruom

.p mpnop

13

Table 2. Social hierarchy of 10 fox squirrels ranked by the percentile
formula, aHudson Noodlot, Michigan State University, 1971.

 

 

Individual Squirrel (X) No. of
No. of Squirrels
Squirrels Victori-

 

 

Social X Victori- ous

Rank Name Tag No. Sex Age ous Over over X

1 Bob 11 M A 9 0

2 Gene 9 M A 7 1

3 Oscar 989 M A 4 2

4 Irene 7 F J 4 3

5 Niles 996 M J 3 3

6 George 8 M A 2 3

7 Nancy 5 F J 2 3

8 Tom 993 M J l 5

9 Kathy 995 F J 1 7
10 Sonya 991 F J O 6

 

aPercentile social rank of individual calculated by formula: x - (A + B)/2

when A . percentage of squirrels encountered that X dominates
B - 100 percent minus the percentage of squirrels dominating X.

l4

 

 

 

o m.F m mom azuax mF
F F m. u so chcma NF
N N nun m. m wN Fggmsm FF
o F F a m. z mN usz oF
m is: F m m m. z mN cgoa m
m in: an: F e e o.F u mF mcosFm m
n is: F F F F m m.F u NN coma 5
mm F m 5 OF F e NF o.F z m¢ usazo: m
mm F m m w a m m m o.N 2 Ne coo m
Nm F a m m m N m m e o.N z m— vFogn: c
mN F F N ¢ m F m m N a o.N z mam sch m
RN F F F N m w m N m N m m.N : mma Laumo N
w nun F in: in: us: run n.- F in: in- F F m.N : m mama F
.02 3:3-
AKHMM mEcmb rsbwsm “PM: csmn. OCOEFM cow—a ULGIOI camp “:95: EOF Laumo mam” wm< xmm GE mez FmFUOm

zoz

mzoFFo<mNFzF va 4<=oF>FozF >m mmFNOFQF> mo >uzmzomxu va ANNNFaom 4<=oF>FozF

 

.NNmF .qumLm>F:=
macaw camFgqu .quuooz comes: .vozums LaocFF mg» »n cognac mFocLF:om on mF mo xgueucoF; FmFuom .N anoF

15

Table 4. Social hierarchy of 13 fox squirrels ranked by the percentile
formula, aHudson Uoodlot, Michigan State University, 1972.

 

 

 

 

Individual Squirrel (X) No. of
No. of Squirrels
Squirrels Victori-
Social Minimum Age X Victori- ous
Rank Name Tag No. Sex (years) ous Over over X
1 Gene 9 M 2.5 4 O
2 Oscar 989 M 2.5 11 1
3 Tom 993 M 2.0 10 2
4 Harold 16 M 2.0 9 2
5 Don 42 M 2.0 8 3
6 Howard 43 M 1.0 7 5
7 Jean 22 F 1.5 5 5
8 Simone 15 F 1.0 3 6
9 John 26 M .5 3 7
10 Halt 29 M .5 3 8
11 Sherri 28 F .5 l 8
12 Jennie 47 F .5 1 9
13 Kathy 995 F 1.5 O 7

 

aPercentile social rank of individual calculated by formula: X - (A + B)/2

when A 8 percentage of squirrels encountered that X dominates
B - 100 percent minus the percentage of squirrels dominating X.

16

period in 1971, a parallel to Sharp's observations evolved. With one
exception, the sequence of capture followed hierarchal rank fairly well,
with the subdominants being caught first and the dominants later in the
trapping period. Therefore, results obtained by trapping around a
feeder would probably not reflect the true age or sex structure of the
population being sampled early in the trapping period. A solution is to
allow the feeder to become depleted after a few days of trapping, thus

forcing even the most dominant animals to enter the traps for food.

Hierarchal Structure:

Hierarchal rank was markedly affected by both age and sex of the
members. Table 5 lists the average rank for four categories for 1971
and 1972. In both years adults ranked significantly higher than
Juveniles, and males ranked higher than females. Since more accurate
data regarding age structure was available for the 1972 population, a
correlation of age with social rank was made (Figure 4) and found to be

highly significant at the 1 percent level.

Home Range:

Allen (1943) concluded that any "cut and dried" approach to home
range determination is more likely to be misleading than helpful due in
part to seasonal modifications and food availability. The efficacy of
home range determination by trapping is also questioned by Hayne (1949a)
and Flyger (1960) who note that home range increases with the number of
captures. Few squirrels in the 1971 hierarchy were trapped or seen
frequently enough for home range determinations; however, in 1972 trap-
ping and observation data from November, 1971 to May, 1972 were used to

delineate home ranges for 11 animals. A correlation of home range with

social rank was made (Figure 5) and found to be nonsignificant.

17

Table 5. Average social rank for categories of fox squirrels.

 

 

Average Social Rank

 

1971a 1972b
Linear Percentile Linear Percentile
Adultsc 2.75 3.12 5.0 5.0
Juveniles 7.33 7.08 9.3 9.3
Males 4.50 4.25 5.0 5.0
Females 7.00 7.22 10.2 10.2

 

aBased on 10 squirrels
bBased on 13 squirrels

cAge 1.5 or older

18

 

 

l3m- 1|
12_- 1.
IIP- <0
rs=.787
th- To
ca» 59" i.
_1
LL]
E; T 8- 10
D
{532% l’—- 0»
t5;:::
§§6- -
as:
.J 3:: 5*" 1|
58
§°4~ -
3- To
2.. T.
l- I.
l j l J J
.5 II) L5 2K) 215
AGEIN YEARS

Figure 4. Effect of age on social rank for the 1972 Hudson Woodlot
Population of fox squirrels. (P<0.01).

SOCIAL RANK OF SQUIRRELS

19

ll - e
rs =.286

DECREASING RANK ---

l~ o

l 1 1 1 1 l 1'
1.0 2.0 3.0 4.0 5.0 6.0 7.0

HOME RANGE IN ACRES

 

Figure 5. Effect of size of minimum home range on social rank for the
1972 Hudson Noodlot population of fox squirrels.-(n.s.).

20

Breeding Behavior:

Unfortunately, breeding behavior was never observed during the
two winter observation periods. Dense sugar maple reproduction
significantly reduced visibility around the blind, and this, along
with the extreme wariness of the animals, reduced opportunities to
observe mating behavior.

Pack gt_al, (1967) noted that for gray squirrels, high ranking
males are involved in most sexual contacts and probably leave more
offspring than subdominant males. It is also quite likely that fox
squirrel dominants would, as in feeding, maintain precedence in

mating encounters.

Heather:

Attempts to correlate specific climatic conditions with movements
or social rank were unsuccessful due to the multiplicity of factors
induced by capricious Michigan winters. Activity on a given day seemed
to be inextricably linked to weather on the previous days and days to
come. A good example of this occurred on February 5, 1972, when winds
were gusting to 43 knots and several different animals of high and low
rank appeared at the feeder. This was the first day of a severe storm,
and squirrels seemed to be aware of the consequence of waiting it out
on an empty stomach. 0n the other hand, on many relatively warm days

few squirrels were seen at the feeder.

1

DISCUSSION
A well-defined hierarchy very likely exists in fox squirrel

populations, due to normal competition, though the feeder used to

21

concentrate squirrels may generate unusual social pressures and cause
significant refinement of the structure. Some animals whose home range
might infrequently overlap, are drawn to the feeder and exposed to one
another for the first time.(:The author witnessed this first meeting

on several occasions when two squirrels (of low rank) displayed fear
of one another. After a relatively long period of avoidance behavior
on the ground, one would rush at the other. Subsequent interactions
were accomplished without delay, and reversals were almost never noted::]
This implies that strong individual recognition is necessary to
maintain a dominance hierarchy, a tenet supported by Guhl and Ortman
(1953) in their work on domestic chickens, as well as by Flyger (1955)
and Pack 33 21, (1967) in their gray squirrel studies.

Overall weather conditions varied a great deal during the two
study periods. The winter of 1971 was harsh both in terms of protracted
cold temperatures and snowfall. Over a foot of hard crusted snow
covered the ground from mid-December to mid-March; a condition that
effectively denied squirrels access to their normal food supply.
However, during the second winter little snow fell, and frequent
warm spells reduced the snow cover to a minimum or to a nonexistent
condition, thus allowing animals to exploit their normal food sources.

Heather conditions over a long period of time seemed to affect
squirrel behavior in hierarchy formation to some extent. The
severity of the first winter prompted early formation and led to
very consistent behavior patterns. Interactions were always swiftly
accomplished and the victor nearly always took his position on the
feeder. In 1972, however, observations began earlier and overall

weather conditions were far less severe.

22

Another pertinent factor in hierarchy formation was the large
component of very young squirrels (6) in the 1972 population, animals
with less experience. These young animals were at first unsure of
their positions, but soon adjusted quite well into the social structure.
Frequently, low ranking animals were observed approaching the feeder,
but would leave when they saw dominants were already present. Often
low ranking animals were seen as much as 100 yards from cover in the
corn field searching for waste corn, and in this position were extremely
vulnerable to predation. On several occasions animals were run down
by the author on foot, a testimonial to their vulnerability. In all
probability, cobs returned to the woodlot from the field by low ranking
animals could be usurped by the dominants. This was not actually
observed, but circumstantial evidence substantiates that it occurs,
and lends credence to the belief that dominance is of strong survival
value.

Day-to-day weather conditions did not appear to affect squirrel
activity, at least in any consistent predictable manner, for both
high and low ranking animals would appear at the feeder during a
variety of weather conditions. while dominants would feed to
repleteness in one hour and then depart, low ranking animals would
sometimes remain three or four hours eating discarded fragments and
awaiting access to the feeder. Certainly, there is survival value
in dominance for exposure to cold and predation are significantly reduced.

During the second winter a few animals seemed to prefer eating on
the ground rather than on the tree, and careful Judgment was required
to discern who was truly the dominant. This tendency decreased as the

winter progressed and the "normal” behavior of feeding on the tree

23
became the rule. Much of the abnormal activity centered around Howard,
an adult male whose origin was probably associated with urban living.
He was the only squirrel during two years of observations who would be
aware and tolerant of human presence without rapid departure.

Irene, the highest ranking female in 1971, was extremely antago-
nistic toward other subdominants. Oddly enough, Simone, the highest
ranking female in 1972, was also a despot and would not allow some
subdominants to feed even on the ground under the feeder. These two
animals would sometimes chase other squirrels from the vicinity of
the feeder, behavior that could be mistakenly interpreted as territorial
to the casual observer.

Actual conflict between squirrels was rare, and interactions
were quickly and smoothly accomplished. Sometimes one animal would
chase another, but the subdominant was clearly established and always
exhibited avoidance behavior.

Of 12 animals captured in January, 1971, only four remained in
the January, 1972 population of 20 animals. This attrition rate is
certainly severe but quite plausible considering the number of
predators present in the woodlot. The extent of emigration was unknown,
but considered to be small.

Observation during the second winter enabled the gathering of
additional data to substantiate fox squirrel hierarchal behavior and
changes in the structure over the course of a year. In addition,
after one year of trapping, animals could be aged more accurately
and the relationship of age to hierarchal rank more easily determined.
Upward shifts in rank did seem to occur with age particularly with

males. Gene, who was second in 1971, became the most dominant animal

24

in the woodlot. Oscar moved from third to second place, and Tom, who
was number 9, moved up to third place the second winter. Kathy, however,
who was third from the bottom in 1971, dr0pped to the bottom position

in 1972. When trapped, Kathy appeared to be in good physical condition
and was known to have carried a summer litter of young animals, so was
apparently normal physiologically. The results are inconclusive, but
tenure is probably an important factor in achieving high social rank,
though aggressiveness is likely linked to genetic factors and increased
age and experience does not guarantee advancement.

The percentile method of hierarchy determination as pr0posed by
Komai 9531. (1959) and utilized by Pack $1391. (1969) and Ozoga (1972)
is a useful method of defining hierarchy structure. Although it is
easy to use, generally requires fewer observations, and takes into
account hierarchal inconsistencies, it does have one shortcoming; it
assumes that a given animal will have a heterogeneous exposure to other
members of the hierarchy. This situation may not occur because of
non-random feeding habits. For example, in the 1971 hierarchy, Tom
is ranked above Kathy by the percentile method, but in actuality was
dominated by Kathy. The tabular method in this case is correct while
the percentile method is in error. These inconsistencies, however,
are few and small in magnitude, as evidenced by the close correlations
obtained by the two methods.

The elaborate and precise hierarchal system developed by fox
squirrels allows the conclusion that they are social animals. Fisler's
(1969:28) contention that ”the concepts of territory and hierarchy

should be considered as opposite extremes of a continuum of mammalian

organizational systems," leads to the conclusion that fox squirrels

25

with their intricate hierarchies probably exhibit very little
territoriality or suffer intraspecific restrictions on their

home ranges. The poor correlation between social rank and home

range strongly implies that this conclusion is correct. One general
characteristic of territories or home ranges is that more food

exists within the occupied area than is utilized (Calhoun 1952).

This restriction of the privilege of occupying space irrespective

of food abundance imposes limitations to population growth (Davis
1950). Fox squirrels travel long distances during periods of winter
food shortage (Allen 1943), which would necessitate crossing
territorial boundaries, if they existed, to obtain access to a point
source of food such as a corn field. The social hierarchy arrange-
ment allows exploitation of an abundant food source by many individuals
rather than the restrictive use by only one animal as in a territorial
system. In short, a hierarchy system allows a greater number of
animals to occupy a given area than does a territorial system. If
adequate food and shelter are available, spatial requirements would
probably not restrict population numbers of fox squirrels, certainly

a very important management implication.

26

LITERATURE CITED

Allen, D. H. 1943. Michigan Fox Squirrel Management. Mich. Dept.
of Cons., Lansing, Michigan. 404 pp.

Bakken, A. A. 1952. Interrelationship of Sciurus carolinensis
(Gmelin) and Sciurus niger (Linnaeus) Tn mixed populations.
PhD Thesis. Univ. Hisc., Madison, Hisc. 188 pp.

Calhoun, J. B. 1952. The social aspects of population dynamics.
J. Mammal. 33(2):139-159.

Davis, D. E. 1950. Malthus - a review for game managers. J.
Hildl. Mgmt. l4(2):180-183.

Davis, D. E. 1964. The physiological analysis of aggressive
behavior. Pp. 53-74. In Hilliam Etkin (Editor), Social
behavior and organization among vertebrates. Univ. Chicago
Press. 307 pp.

Fisler, G. F. 1969. Mammalian organizational systems. Los Angeles
County Museum of Nat. Hist. Publ. 167. 32 pp.

Fitzwater, H. 0. Jr. 1943. Color marking of mammals with special
reference to squirrels. J. Hildl. Mgmt. 7(2):l90-192.

Flyger, V. F. 1955. Implications of social behavior in gray
squirrel management. Trans. N. Am. Hildl. Conf. 20:381-389.

Flyger, V. F. 1960. Movements and home range of the gray squirrel
Sciurus carolinensis, in two Maryland woodlots. Ecology

Guhl, A. M. and L. L. Ortman. 1953. Visual patterns in the
recognition of individuals among chickens. Condor 55:287-298.

Gysel, L. H. 1971. Beechnut production in Michigan. J. Hildl.
Mgmt. 35(3):5l6-519.

Hayne, D. H. 1949a. Calculation of size of home range. J.
Mammal. 30(1):l-18.

Hicks, E. A. 1942. Some major factors affecting the use of two
inventory methods applicable to the western fox squirrel,
‘ Sciurus ni er rufiventer (Geoffrey). Iowa State Coll. Jour.
SCI. I6: 9 -3 5.

27

Komai, T., J. V. Craig, and S. Hearden. 1959. Heritability and
repeatability of social aggressiveness in the domestic chicken.
Poultry Sci. 38(2):356-359.

Mohr, C. 0. 1947. Table of equivalent populations of North American
small mammals. Amer. Midl. Nat. 37(1):223-249.

Ozoga, J. J. 1972. Aggressive behavior of white-tailed deer at
winter cuttings. J. Hildl. Mgmt. 36(3):861-868.

Pack, J. C., H. S. Mosby and P. B. Seigel. 1967. Influence of
social hierarchy on squirrel behavior. J. Hildl. Mgmt.
31(4):720-728.

Sharp, H. M. 1958. The art and technique of live-trapping gray
squirrels. Penn. Coop. Hildl. Res. Unit Report No. 3,
Penn. State Univ., University Park, Pa.

Seigel. S. 1956. Nonparametric statistics for the behavioral
sciences. McGraw-Hill Book Co., New York. 312 pp.

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

Stokes, A. H., and D. F. Balph. 1965. The relation of animal
behavior to wildlife management. Trans. N. Am. Hildl. and
Nat. Resources Conf. 30:401-410.

PART II
APPLICATION OF A SIMPLE METHOD
FOR CENSUSING FOX SQUIRRELS

28

APPLICATION OF A SIMPLE METHOD FOR CENSUSING FOX SQUIRRELS

The development of efficient and expedient census methods has
been the subject of much concern and distress for wildlife investi-
gators. Arboreal mammals such as squirrels are particularly difficult
to census, because many of the usual methods relying on vocalization,
tracks or habitat disturbance are unworkable. Most commonly, squirrel
populations have been measured by direct trapping or some additional
modification involving sight records and/or tag returns from hunting,
generally time-consuming methods. ‘Friley (1955), using live traps to
capture squirrels, found the Lincoln Index (Davis 1963:107) method of
calculating preseason hunting populations of fox squirrels (Sgigrg§_
£1993) more reliable for indicating trends in animal numbers than
per-unit effort trapping success. Leaf nest counts as an index of
squirrel abundance, were used by Goodrum (1937), though he expressed
dissatisfaction with the results in some habitats, as did Allen (1943)
in all Michigan situations. Allen (1943) used in his studies direct
trapping generally applied in a grid pattern to monitor population
levels. Flyger (1959) compared three census methods, two using the
Schnabel (1938) modification of the Lincoln Index, and the third based
upon time-area counts as proposed by Goodrum (1940). The first was
based on trapping results, the second on the visually determined ratio
of marked to unmarked animals (trap-sight records) obtained during
observations. He found the latter gave the most satisfactory results.
Nixon 35 21, (1967) also used grid trapping results for their comparison
of various mathematical predictions of population numbers. They con-
cluded that assumptions must be made which may not be true in regard

29

30

to trap response but that, in lieu of better methods, mathematical
functions may be used to produce population estimates. Sharp (1958)
took advantage of the social hierarchy described by Flyger (1955)

and lack of territoriality in gray squirrels to attract animals with
a feeder to a central location for trapping. He emphasized that con-
ditioning animals by prebaiting is the key to success in trapping.
The intent of this paper is to report the results obtained using the
same procedure for censusing fox squirrels, where sight records of
previously trapped and marked animals provided a Judgment as to the

efficacy of the method.

STUDY AREA

Two woodlots on the farm property of Michigan State University in
Ingham County comprised the study area. These woodlots were nearly
equal in both size and species composition. Toumey Hoodlot, the first,
is a 20-acre old-growth beech-maple stand surrounded by pasture land,
while Hudson Hoodlot located R-mile away is an l8-acre all-aged beech-
maple forest. Field corn is usually planted on the north and west
boundaries of Hudson Hoodlot, while pasturage is the main use for the
areas adjacent to the south and east boundaries. Both woodlots contain
wild, relatively discrete, unhunted populations of fox squirrels. Their
wildness could be a result of predator pressure and infrequent human
activity in the area. The woodlots contain few mast trees other than
the sporadically producing American beech (Fagus grandifolia) (Gysel
1971) and consequently do not normally carry high populations of
squirrels. In Michigan wintering squirrels often rely on auxiliary

food supplies such as field corn (Fouch 1961).

31

METHODS

Trapping activities began in July, 1970 in Toumey Hoodlot and
continued until March of 1972, for a duration of 21 months. Hudson
Hoodlot trapping was initiated in January, 1971 and continued 17 months
until May of 1972.

Trapping was done on a year-round basis during nearly every month
of the study periods and was accomplished by two basic methods. The
first was an application of the feeder system advocated by Sharp (1958)
for gray squirrels to attract the animals to the traps at a given
location. The second was an evenly distributed grid system wherein
each trap serviced an area of .90 acres. The grid provided an
opportunity to test the distance squirrels would travel to some central
location such as a feeder for trapping.

The feeders were simply basket arrangements made of chicken wire
and wood which could hold a bushel of ear corn (Sharp 1958). When
hung on a tree, they were readily accessible and visible to squirrels
within the woodlot and could provide food without service for long
periods of time. Six traps were placed in a semi-circle below the
feeder and also baited with ear corn. The traps were of a collapsible
metal (wire) variety which measured 9 x 9 x 24 inches. As Sharp (1958)
noted, the behavioral mannerisms of squirrels at the feeders make their
capture possible. Fox squirrels have a hierarchal form of social
arrangement and do not appear to suffer intraspecific restrictions on
their home ranges (Bernard 1973). This makes feasible trapping at a
central location without fear that territorial boundaries might
restrict animal movements to a central location. Five feeding stations

were maintained for much of the time in each woodlot, and at selected

III I l I

32

intervals four were emptied to test whether animals would transfer
their activity to one central feeder. Allen (1943) observed that fox
squirrels usually travel over a ten-acre area on a seasonal basis.
Hhen five feeders were operating, each serviced approximately four
acres, coverage thought to be adequate to attract most squirrels in
the woodlots. Hhen only one central feeder was operating, the maxi-
mum distance a squirrel in the woodlot would have to travel became
800 feet. In addition to the grid and feeder methods of trapping,
another popu1ation monitor was maintained. Both woodlot populations
were being observed as part of a behavioral study, and the identity
of nearly all animals was established. All animals captured were
ear tagged and most were dyed with Nyanzol A (Fitzwater 1943) in a
characteristic pattern that allowed visual identification of
individuals. Much time was spent in the woodlots, particularly from
January through March, making behavioral observations, so that sight
ratios of marked to unmarked animals could be maintained. Hhen no
unmarked animals were seen or trapped, the population was considered
to be inventoried. Therefore, both trap and sight records were used
to determine what was considered to be an efficacious estimate of the
actual number of squirrels in the respective populations (Figs. 1 and
2). The minimum number of squirrels known to be alive in a given
month was determined by adding the number seen or trapped that month
to those marked during an earlier trapping period but not seen or
recaptured until some subsequent date (Figs. 1 and 2). It is felt
that the population estimates, especially during January through
March, very nearly represented the actual number of animals present

since trapping was frequent and observations nearly continuous.

33

Number of Squirrels
on
I

 

 

1 1 1 1 1 1 1 4 1 L 1 1 1 1 1 1 1 1
J A S 0 N D J F M A M J J A S 0 N D J F M
1970 l97l 1972

—— Actual population based on grid trapping,
feeder trapping,and visual observation.

Population assessment based on seven days
trapping at a single central feeder. '

 

Figure l. Fluctuation of the Toumey Hoodlot fox squirrel population
showing estimate based on feeder trapping.

34

20"

...........
.3. ..........
........

......
.........

 

Number of Squirrels

 

 

................
-.-.;.:.:.:. ....
!-:-.-.-.-.-. 1

I J J J J_ J_ J J L l J J x
J F M A M J J A S O N D J F M A‘ M
1971 [972

 

-———- Actual population based on grid trapping,
feeder trapping ,ond visual observation.

Papulalion assessment based on six days
trapping at a single central feeder.

 

Figure 2. Fluctuations of the Hudson Hoodlot fox squirrel population,
showing estimate based on feeder trapping.

35

The question to be answered was whether animals could be induced
to come to a single point source of food and thereby become trapped
with a minimum of effort, with the attendant conviction that a
representative proportion of the population had been captured. A11
grid stations and feeders except the central one in each woodlot
were depleted at least two weeks prior to trapping at the single
central feeder. This allowed squirrels time to transfer their feeding
forays to the single remaining feeder. Here, trapping periods varied
in length from four to seven days during the difficult winter period
for squirrels (Figs. 1 and 2). During the second half of these periods,
the central feeder was also emptied of corn to ensure that the only
source of corn was contained in the traps. The trapping periods were
not continuous but were in part chosen in response to weather conditions.
Trapping was always preceded by a conditioning period where traps were
propped open and animals allowed free access to bait (corn) within the

traps.

RESULTS

A short trapping period of 7 days or less during periods of winter
food scarcity proved efficient for capturing nearly all animals present
in the woodlots (Figs. 1 and 2). The Hudson Hoodlot papulation was
inventoried during 1971 and 1972, and the Toumey animals only in 1972.
In Toumey, trapping at the single feeder for seven days yielded seven
animals (88%), when the actual population was considered to be eight.
All the animals were taken during the first four days of trapping. The
January, 1971 Hudson Hoodlot population was actually twelve animals,

and the estimate based on six trap days was eleven animals or 92% of

36

the population. In March of 1972 the Hudson population was considered
to be 17 animals, and the feeder estimate was 14 animals, or 82% of
the actual population. This estimate was based on six trap—days,
though all squirrels were caught by the fifth day. Animals which had
previously been taken at one or more of the other four feeders readily
transferred their feeding forays to the central feeder when it alone
held corn. Grid trapping periods following the inventory based on

the single feeder resulted in the capture of only one new animal, a
strong indication that the feeder could draw squirrels from the entire
woodlot. Intentions were to attempt the method during other seasons,
but overall trapping success was inconsistent and a reliable estimate

could not be made.

DISCUSSION

As in any approach to censusing animals, there are advantages
and shortcomings to be considered. One major drawback of this method
is the effect food shortage or abundance may have on trapping success.
As Barkalow $5.31, (1970) noted, in an exceptionally good mast year
squirrels may be overwhelmed by food and thus be literally untrappable.
The application of direct trapping as a census method, therefore.
requires a food shortage to function efficiently. This condition
occurs with some certainty in Michigan nearly every year, for even
though autumn food may be abundant, crusted snow and ice usually make
it unavailable to foraging squirrels during the January through March
winter period. Examination of field conditions will aid the investi-
gator in deciding when and if these or other conditions of food shortage

exist. Application of the method during two winters of markedly varied

37

snow conditions showed quite favorable results. In 1971 deep crusted
snow made access to buried food supplies difficult if not impossible.
In 1972 the snow cover was far less permanent, and at frequent
intervals squirrels had free access to their usual food sources.
During the periods between thaws, however, animals came well to the
feeder and were quite easily trapped. By necessity the season for
censusing, using feeders and live traps, becomes the winter period
of food scarcity, not always the most desirable time to obtain
population data. However, early winter numbers of animals, though
somewhat less, may approximate fall numbers. The data in Figures 2
and 3 reflect higher numbers of animals in January than in November,
most certainly a result of increasing effectiveness of trapping as
food became more difficult to obtain. Hinter estimates are useful
since they reflect popu1ation numbers at a time when normally only
mortality is operating, as opposed to fall when mortality, migration,
and even natality may be acting.

Another pertinent factor to consider is the effect weather
conditions may have on trapping success. For example, severe weather
in very early winter may lead to reduced or complete cessation of
squirrel activity (Allen 1943); however, by mid-January most squirrels
in the woodlot became quite active. Severe cold spells or stormy
weather sometimes restricted movements on a day-to-day basis, but the
investigator can judge these conditions and trap during more "normal"
weather conditions.

Consideration should also be given to the possibility of catching

a predator such as the raccoon (Procygn_lotor) at the trapping site.

 

If this situation should occur, the location may be rendered ineffective

38

until the offender is removed. During winter in Michigan, this is
seldom a problem, though it could become a serious drawback in more
southern states where raccoons and other omnivores remain active
through the winter months.

Hith shortcomings considered, however, the method has worked
well for Michigan fox squirrels in winter, giving good results when
compared to known populations. The centrally located feeder appeared
to attract animals over an 18- to ZO-acre area, and perhaps could be
expected to be effective over an even larger area. Sharp (1958)
found that one feeder for 28 acres was adequate to attract most gray
squirrels for trapping, and fox squirrels have been known to travel
more than a mile during severe winters (Allen 1943). It is entirely
possible that one feeder for 30 acres might suffice, though the
efficiency of the method might decline. Another advantage is the
small number of traps needed and the length of time needed to service
them. An hour was required to service 20 traps on a grid system,
while only 15 minutes was needed to check the six traps at a feeder
station.

The social behavior, particularly the dominance hierarchy, plays
an important part in determining success of this method. Hhen two or
more animals appear at the feeder, the subdominants are forced from
the feeder to the ground and the traps, where their capture is virtually
assured. A trapped animal does not frighten others nearby, but rather
seems to reinforce them that all is well. On one occasion, all six
traps held animals. In woodlots with large populations, more traps
should be utilized to ensure that a trap is available for each

approaching squirrel.

39

Present evidence shows that feeder trapping is an efficacious
method of capturing squirrels, requiring relatively limited time and
equipment. Further investigation of the method is in order on a

larger scale, especially in different geographical situations where

effects of winter are not so predictable.

4O

LITERATURE CITED

Allen, D. H. 1943. Michigan Fox Squirrel Management. Mich. Dept.
of Cons., Lansing, Mich. 404 pp.

Barkalow, F. 5., Jr., R. 8. Hamilton, and R. F. Soots, Jr. 1970.
The vital statistics of an unexploited gray squirrel population.
J. Wildl. Mgmt. 34(3):489-500.

Bernard, R. J. 1972. Social organization of the western fox squirrel.
M.S. Thesis. Michigan State Univ., East Lansing. 41 PP.

Davis, D. E. 1963. Estimating the numbers of game populations.
Pp. 89-119. In H. S. Mosby (Editor), Wildlife investigational
techniques. 2nd edition. The Wildlife Society, Washington,
D. C. 419 pp.

Fitzwater, H. 0., Jr. 1943. Color marking of mammals with special
reference to squirrels. J. Wildl. Mgmt. 7(2):l90-192.

Flyger, V. F. 1955. Implications of social behavior in gray squirrel
management. Trans. N. Am. Hildl. Conf. 20:381-389.

Flyger, V. F. 1959. A comparison of methods for estimating squirrel
populations. J. Wildl. Mgmt. 23(2):220-223.

Fouch, H. R. 1961. Mast crops and fox squirrel populations at the
Rose Lake Wildlife Experiment Station. Mich. Dept. Cons. Game
Div. Report No. 2332.

Friley, C. E. 1955. Criteria for investigating fall fox squirrel
populations. J. Wildl. Mgmt. 19(1):89-93.

Goodrum, P. D. 1937. Notes on the gray and fox squirrels of eastern
Texas. Trans. N. Am. Wildl. Conf. 2:499-504.

Goodrum, P. D. 1940. A population study of the gray squirrel in
eastern Texas. Texas Agric. Exp. Sta., Bull. 591. 43 pp.

Gysel, L. H. 1971. Beechnut production in Michigan. J. Wildl.

Nixon, C. M., W. R. Edwards, and L. Eberhardt. 1967. Estimating
squirrel abundance from live-trapping data. J. Hildl. Mgmt.
31(1):96-10l.

41

Sharp, W. M. 1958. The art and technique of live-trapping gray
squirrels. Penn. Coop. Wildl. Res. Unit Report, No. 3, Penn.
State Univ., University Park, Pa.

Schnabel, Zoe Emily. 1938. The estimation of the total fish
populations of a lake. Am. Math. Monthly 45(6):345-352.

‘tittiiititiiit