Journal of Mammalogy, 83(1):267–279, 2002
RESPONSE OF TWO LOW-DENSITY POPULATIONS OF
PEROMYSCUS LEUCOPUS TO INCREASED FOOD AVAILABILITY
JOHN A. YUNGER*
Department of Biological Sciences, Northern Illinois University, DeKalb, IL 60115
Studies of food supplementation in small mammals often result in significant increases in
population density. These numerical responses to increased food availability may interact
with other variables, such as predation, territoriality, and emigration, to limit population
size. In this study, the demography of Peromyscus leucopus was compared between an oak
wood undergoing a high mast year and a reference wood. At a smaller spatial scale, replicated plots were supplemented with rodent chow in an attempt to reproduce patterns
occurring at the larger woodlot scale. Some of the food supplemented plots also had predators excluded to test for interactions between top–down and bottom–up effects. Very lowdensities of P. leucopus responded numerically to increased food availability. Results for
a relatively large ‘‘natural’’ experiment and a more controlled manipulative experiment
were similar. In both cases, the increase in numbers was primarily the result of immigration,
not reproduction. Increased food availability in the spring resulted in earlier reproduction
but a decline in numbers—probably because of territorial defense. The food supplementation led to increased predator activity and decreased survivorship of the prey. This resulted
in a rapid turnover of individuals and a significant predator 3 food interaction.
Key words: food supplementation, mast production, northern Illinois, Peromyscus leucopus, whitefooted mouse
The most direct way to test for resource
or bottom–up limitation in mammals is to
provide additional nutrients or supplemental food, an approach which has been used
by numerous investigators (Desrochers et
al. 1988; Dijkstra et al. 1982; Duquette and
Millar 1995; Galindo-Lealindo and Krebs
1998; Hansen and Batzli 1978; Licht 1974;
Schweiger and Boutin 1995; Taitt 1981;
Taitt and Krebs 1981; Wolff 1985). Boutin
(1989) reviewed 62 such studies with small
mammals. Supplemental feeding commonly
resulted in increased duration of the breeding season, increased adult body weight,
decreased home range size, and an increase
in density. However, as Boutin (1989)
pointed out, there is a need to relate responses to food supplementation to a variety of environmental conditions. Distur-
bance or periodic perturbations in the environment often results in substantial decreases in density (Huston 1979; Reice
1994). Yet most food supplementation studies generally have not been carried out during periods of very low population densities
of the target species, presumably because of
the limited information on which a study is
based.
In addition to the role of disturbance, interactions between bottom–up forces, such
as food, and top–down forces, such as predators, can be important in determining the
dynamics of populations (Leibold 1989;
Matson and Hunter 1992; Power 1992;
Strong 1992). Such forces have been well
studied with invertebrates and in aquatic
systems (Hunter and Price 1992; Power
1992), yet have received little attention
among terrestrial vertebrates (see Desy and
* Correspondent: [email protected]
267
268
JOURNAL OF MAMMALOGY
Batzli 1989 and Krebs et al. 1995 for 2 exceptions). There are several reasons to suspect that a multifactor experimental approach could prove useful in elucidating
mechanisms controlling small mammal demography (Gaines et al. 1991; Lidicker
1991, 1994). First, the work with invertebrates and aquatic systems has shown that
predation and food availability may interact
to regulate population dynamics. Second, in
a review of food supplementation studies,
Boutin (1989) found that a 2-fold increase
in population size is often observed in small
mammals. One possible explanation is the
pantry effect (sensu Ford and Pitelka 1984),
where compensatory responses by predators limit population size. Also, it is unlikely that the complex small mammal population dynamics observed at midlatitudes
(Getz et al. 1987) are the result of interactions at only a single tropic level.
Boutin’s review (Boutin 1989) also illustrated the importance of scale when conducting food supplementation studies with
vertebrates. Increased densities in the areas
being manipulated may result from immigration. Any subsequent increase in reproduction could result in a source habitat from
which individuals disperse to surrounding
areas (Pulliam 1988). Boutin (1989) recommended that the spatial scale of supplementation experiments be increased. However, this is not always logistically feasible.
An alternative approach is to use both natural and manipulative experiments. Natural
experiments have the disadvantage that they
are less rigorous than manipulative experiments but have the advantage that they can
occur over much larger spatial scales (Diamond 1986). A combined approach can
help to offset the inherent weaknesses of
each design.
As part of a long-term study, I had the
opportunity to address questions on the role
of food availability in affecting demographic fluctuations of the white-footed mouse
(Peromyscus leucopus). The duration of the
study included extremely low population
densities resulting from disturbance, a se-
Vol. 83, No. 1
vere ice storm. Data from 2 oak woodlots
were used to compare demographic fluctuations of P. leucopus. Because of the differences in mast phenology and species
composition, 1 of the 2 woodlots underwent
a high mast year; mast has been shown to
be an important food source for P. leucopus
(Ostfeld et al. 1996; Wolff 1996). Concurrent with this natural experiment, P. leucopus populations were provided supplemental food on replicated plots, some of
which excluded predators. This permitted
the following questions to be addressed:
After a severe disturbance and dramatic decline in numbers, 1) does P. leucopus increase numerically following the increased
availability of naturally occurring food? 2)
can observed responses to changes in the
natural food supply be replicated using manipulative experiments? and 3) does predation interact with food to regulate population densities of P. leucopus?
MATERIALS
AND
METHODS
Study site.—The study site was located at Fermi National Accelerator Laboratory (henceforth
Fermilab) in northern Illinois. Fermilab is a heterogeneous landscape of woodlots, wetlands, agricultural fields, restored prairie, shrubby-oldfields, nonnative grasslands, and developed sites
totaling 3,200 ha (Illinois Natural History Survey 1985). The food supplementation experiment was located in a 32-ha prairie restoration.
Prior to 1989 the site was a cornfield; in the
spring of 1989 it was plowed, disked, and a mix
of seeds of prairie grasses and forbs was evenly
broadcast across the site. Thus, the management
history of the site provided a relatively homogeneous environment for large-scale, replicated
field experiments. Plant coverage at the site was
dominated by big bluestem (Andropogon gerardii) and tall goldenrod (Solidago altissima).
Two oak woodlots, located 1.8 km apart and
separated by an 8-m-wide body of water, were
used to compare the dynamics of small mammal
populations between a high mast site and a reference site. The 1st woodlot (Woodlot I, 6.2 ha)
was dominated by bur oak (Quercus macrocarpa) and had a large mast crop. The 2nd woodlot
(Woodlot II, 10.1 ha) was dominated by black
oak (Quercus velutina) and did not have a large
February 2002
YUNGER—SMALL MAMMAL FOOD AVAILABILITY
mast crop. These 2 woodlots were part of a larger study monitoring small mammal population
dynamics at 6 locations across Fermilab.
A diverse assemblage of predators has been
recorded at the site, of which coyotes (Canis latrans), followed by red fox (Vulpes vulpes),
were the most common terrestrial small mammal
predators, red-tailed hawks (Buteo jamacensis)
and great-horned owls (Bubo virginianus) were
the most common avian small mammal predators (Yunger 1996). Of these, data collected on
predator diets (Cooper 1997, Randa 1996) indicated that coyotes were the most significant
predators on Peromyscus at the study site.
Experimental design.—During the late summer and early fall of 1992, eight 0.6-ha plots
(78 by 78 m) were delineated $80 m apart at
the prairie site. Each plot had a trapping grid
arranged in an array (6 by 6) with 12-m spacing.
One Sherman live trap (7.5 by 9 by 23 cm),
baited with rolled oats and peanut butter, was
placed at each station and checked in early
morning and late afternoon, daily, during trap
sessions. Polyester fiber was used in the traps as
bedding to reduce stress and provide insulation
during winter months. Preliminary trapping was
conducted for 3 consecutive nights on each grid
during October and November of 1992. Trapping was reinitiated in March of 1993 and continued uninterrupted at monthly intervals
through January 1998. Captured individuals
were identified to species, marked with a
uniquely numbered ear tag, and data taken on
the date of capture, plot, grid location, mass,
sex, sexual condition (males: testes abdominal
or descended; females: perforate or imperforate), and reproductive condition (females only:
pregnant, dilated pubic symphysis, lactating).
Females were identified as reproductively active
using any 1 of the 3 previous criteria. Peromyscus were aged as juvenile, subadult, or adult using pelage characteristics (Conventry 1937;
Gottschang 1956; Layne 1968).
In the spring of 1993 construction of vertebrate predator exclosures was initiated on plots
2, 4, 6, and 8, and completed by July 1994. The
exclosures were constructed using 2-m high,
2.5-cm mesh poultry fencing. The top 0.5 m of
fencing was angled outward 458 and had a 0.25m overhang to prevent predators from climbing
over. At ground level, the fencing projected outward for 0.5 m and was staked down to prevent
predators from digging under. The exclosures
269
were effective for all predators (canids, longtailed weasel, and raptors) but the smallest (least
weasel—Yunger 1996). Low-fencing (0.25-m
high), over which the predators were readily
able to pass, was constructed around the control
grids to test for a fence effect on the small mammals (Krebs et al. 1969; Ostfeld 1994). Trap
lines inside the plots were periodically compared
with 2 control lines on the north and south end
of each control plot. An equal trap effort both
in and out of the control plots revealed that the
fencing had no observable effect on small mammal species composition or numbers (Yunger
1996). The lack of a fence effect (sensu Hestbeck 1982, 1988) was corroborated by visual
signs, including direct observations of small
mammal movements, tracks in the snow, and
microtine tunnels, providing further support that
the fencing did not impede Peromyscus movements. Raptors were excluded using nylon netting (12.5 by 12.5 cm mesh) suspended above
the plots. Polyvinyl chloride and bamboo poles
were placed on the control plots to duplicate the
perches for passerines on the exclusion plots that
resulted from the posts used to support the netting. The netting excluded raptors, but permitted
most passerines to pass through freely.
Food supplementation was initiated in October 1994, immediately following the monthly
small mammal census and terminated in May
1995 immediately prior to the monthly small
mammal census. Rodent chow (Rodent Breeder
Chow, PMI Feeds, Saint Louis, Missouri) was
evenly broadcast by hand on plots 1, 2, 3, and
4 at the rate of 11.4 kg per week until the May
1995 small mammal census. Work with a previously established laboratory colony showed
that the chow readily supported both growth and
reproduction of P. leucopus.
As part of a larger study, in March of 1994,
small mammal trapping was initiated at 6 additional locations across Fermilab, representing
prairie, oldfield, woodland, and shrub habitats.
Three 192-m transects, spaced 50 m apart, in
each of the 6 locations, were trapped for 2 consecutive nights at monthly intervals. These transects sampled the 2 oak woodlots which were
examined for small mammal responses to a high
naturally occurring mast crop. The same markand-recapture techniques used on the experimental plots were applied to the transects. Sherman live traps (7.5 by 9 by 23 cm), baited with
rolled oats and peanut butter, were placed at 12-
270
JOURNAL OF MAMMALOGY
m intervals along each transect (total of 51 traps
per location).
Acorn density was estimated in the 2 oak
woodlots to examine the effects of a naturally
occurring food supply. Individual acorns were
collected and weighed to determine mean mass
and multiplied by the mean number of caps 1
acorns/caps counted in 51, 1-m2 quadrate, to
compare mast biomass between woodlots.
Data analysis.—Small mammal density, reproductive status, and mean body weight per
plot for each monthly census were summarized
using the program Capture-Mark-Recapture (Le
Boulengé 1987). Juveniles and subadults were
not included in analyses of reproductive status
and weight. Densities were determined using
minimum number known alive, as the number
of marked individuals recaptured (Ri) commonly
did not meet the assumptions (Ri $ 10 per trap
session) of probability estimation techniques
(Pollock et al. 1990; Seber 1982). The trap records for October 1994 through June 1995 from
the experimental site were used to test for top–
down versus bottom–up effects limitation of P.
leucopus numbers. The following abbreviated
codes were used for the 4 treatments: food not
supplemented, predator present 5 2F 1P; food
supplemented, predators present 5 1F 1P; food
not supplemented, predator excluded 5 2F 2P;
food supplemented, predator excluded 5 1F 2P.
Trap records from March 1994 through March
1995 from the 2 oak woodlots were analyzed for
the effects of mast availability on P. leucopus
demography.
For the continuous response variables (density
and weight), regressions were fitted to the data,
and F-values were calculated to see if higher
order polynomials added significant explanatory
strength to the models. The best-fit models were
then tested for autocorrelation using the Durbin–
Watson statistic (Ostrom 1990; PROC REG,
SAS Institute Inc. 1990). As the time series were
relatively short there was no significant autocorrelation, and treatment effects were tested with
a covariate analysis using general linear models
(PROC GLM, SAS Institute Inc. 1990). The
covariate approach also permits the testing of
significant curvilinear responses and their interactions, which is difficult to do with repeated
measures analysis of variance. For all parametric
analyses, residuals were inspected for departures
from normality using box-plots and normal
probability plots (Tukey 1977) and inferences
Vol. 83, No. 1
based on type III sums-of-squares. The biomass
of acorns in both oak woodlots was highly
skewed. Consequently, a nonparametric Wilcoxon rank test was used to test for a significant
difference (Zar 1984). All analyses were considered significant at a 5 0.05.
Survival analyses of adults were conducted
using a nonparametric log-rank test (PROC
LIFETEST, SAS Institute Inc. 1990), which is
robust to differences in survival curve shape and
does not assume any particular underlying distribution (Fox 1993; Lee 1980). Two preplanned
comparisons, 1F 1P versus 1F 2P and 1F 1P
versus 2F 1P, were conducted to test specifically the effects of food supplementation alone
and the effects of predators when food is supplemented. Data are presented as mean 6 1 SE.
RESULTS
Population dynamics.—A severe disturbance, an ice storm, in January 1994 resulted in ca. 2.5 cm of ice covering much
of northern Illinois, including the field site.
Prior to the ice storm, densities of Peromyscus at the experimental site averaged
16.7 (61.43) individuals per plot; for the 3
months following the ice storm, densities of
P. leucopus declined to an average of 0.79
(60.41) individuals per plot. Sampling at
the 6 additional locations across Fermilab
showed effects of the ice storm were not
restricted to the experimental site but were
widespread over several thousand hectares.
By late summer, trap success never exceeded 3.9 individuals per 100 trap nights at any
of the field sites.
In the oak woodlots, densities of P. leucopus were similar between March and August 1994 (Fig. 1). Following high production, acorn biomass on the forest floor was
nearly 6 times greater in Woodlot I (114.72
6 21.13 g) as compared with Woodlot II
(19.12 6 4.33 g), a significant difference (S
5 3,225.00, Z 5 4.032, P , 0.001). Acorns
began falling in August, which was followed by a dramatic increase in P. leucopus
numbers. Acorns ceased falling in approximately mid-October, with mouse numbers
reaching their peak in November. Following
the peak, P. leucopus declined dramatically,
February 2002
YUNGER—SMALL MAMMAL FOOD AVAILABILITY
271
TABLE 1.—Analysis-of-covariance table of
Peromyscus leucopus densities in high-mast
(Woodlot I) versus low-mast (Woodlot II) oak
wood lots.
Source
Woods
Time
Time2
Time 3 woods
Time2 3 woods
Error
FIG. 1.—Population dynamics of P. leucopus
in oak woodlots with high and low mast production (61 SD). Stippling indicates period
when acorns were dropping.
with the number of individuals recorded in
January being similar to premast abundance; numbers remained relatively stable
following the decline. In oak Woodlot II,
mouse abundance remained relatively stable during the time period of large fluctuations in Woodlot I (Fig. 1).
The change in P. leucopus numbers over
time was significant and occurred as a quadratic response (Table 1). Overall, numbers
were also significantly greater in Woodlot I
than in Woodlot II. Of particular importance are the 2 highly significant interaction
terms, time and time2 crossed with woods
(Table 1). The interactions indicate that the
number of Peromyscus significantly diverged and then subsequently converged
(Fig. 1).
Following the initial increase in P. leucopus densities in Woodlot I, food supplementation was initiated at the experimental
site (October 1994). Between October and
December, there was a general trend toward
increased densities (Fig. 2), a seasonal pattern commonly observed among Peromyscus in the Midwest (Baker 1983; Hoffmeister 1989). There was a decrease in numbers
during January as evidenced by low trap
success after several days of continued
snow. A similar trend was observed at other
d.f.
MS
F
P
1
1
1
1
1
84
53.217
350.512
324.023
145.827
112.944
10.153
5.24
34.52
31.91
14.36
11.12
0.025
0.001
0.001
0.001
0.001
locations around Fermilab. For the duration
of the supplementation experiment the
number of individuals on 2F 1P and 2F
2P plots remained relatively constant.
However, by February, there was an approximately 4-fold increase in densities on
the food supplemented plots as compared
with nonsupplemented plots (Fig. 2). Subsequent to the increase, numbers declined
steadily until densities converged on the
supplemented and nonsupplemented plots
in May 1995.
Overall, densities were significantly
greater on the 1F plots as compared with
that of the 2F plots and the number of individuals changed significantly over time in
FIG. 2.—Population dynamics of P. leucopus
on 4 treatments: control (2F 1P), food supplemented (1F 2P), predator excluded (2F 2P),
and food supplemented and predator excluded
(1F 2P). Food supplementation was initiated
following October census and terminated following May census.
272
JOURNAL OF MAMMALOGY
Vol. 83, No. 1
TABLE 2.—Analysis of covariance for densities and body mass of Peromyscus leucopus on predator
exclusion versus predator access and food supplemented versus unsupplemented plots.
Density
Body mass
Source
d.f.
MS
F
P
MS
F
P
Predation
Food
Time
Time2
Pred 3 food
Time 3 pred
Time 3 food
Time2 3 pred
Time2 3 food
Error
1
1
1
1
1
1
1
1
1
54
0.001
133.683
545.615
526.251
107.640
1.537
274.216
2.768
242.590
26.274
0.01
5.09
0.77
20.03
4.10
0.06
10.44
0.11
9.23
0.996
0.028
0.001
0.001
0.048
0.810
0.002
0.747
0.004
2.861
18.098
41.146
81.891
18.473
0.115
19.478
0.235
15.955
3.891
0.74
4.66
10.58
21.06
4.75
0.03
5.01
0.06
4.10
0.395
0.036
0.002
0.001
0.034
0.864
0.029
0.817
0.048
a quadratic fashion (Table 2). Interactions
of both time 3 food and time2 3 food were
significant, reflecting the strong increase
and subsequent decrease in number of Peromyscus on food supplemented plots (Fig.
2). There was also a significant interaction
of predator 3 food (Table 2), which was the
only indication of a predator exclusion effect on P. leucopus. An interaction plot was
constructed by collapsing the 4 treatment
means over time (Fig. 3a). Densities were
lowest on 2F 1P plots, intermediate on 1F
2P, and 2F 2P plots, and highest on 1F
1P plots.
Reproduction and survivorship.—During
September all individuals captured on both
plots assigned to the food supplementation
treatment (5 males, 4 females) and those assigned to unsupplemented plots (7 males, 3
females) were reproductively active. In October, 2 of 3 males (40%) and the 1 female
on plots assigned to the food supplementation were reproductively active, whereas
4 of 5 males (80%) and 3 of 5 females
60%) on unsupplemented plots showed
signs of reproductive activity. Following
the October census, no reproductively active individuals were recorded until the following spring. Thus, food supplementation
did not maintain reproduction going into
the winter. The first reproductively active
females on supplemented plots were captured in March, 2 months earlier than on
nonsupplemented plots (Fig. 2).
Reproduction in the high mast production Woodlot I followed a pattern similar to
that on the experimental site (Table 3). The
majority of males were scrotal in September
and October, following which no reproductively active males were recorded until May
1995. Similarly, the majority of females
were either pregnant or had swollen teats in
September and October. In November 31%
were reproductively active, following
which no reproductively active females
were recorded until May 1995. The cessation of reproduction was not because of
new immigrants. All of the males and 69%
of the females captured in October were recaptured following the decline in reproduction. The small sample size in Woodlot II
precluded a statistical comparison of P. leucopus reproduction between the 2 woodlots.
In Woodlot II, 2 females in September, 3 of
4 females and 1 male during October, 2 of
5 females and no males during November
were reproductively active. Reproductively
active individuals were not recorded again
until May 1995. Thus, high mast production
did not sustain reproduction into the winter
and there was no notable difference when
compared with reference woods.
The low number of captures in Woodlot
II also precluded an interwoodlot comparison of juveniles and subadults. When captures from Woodlot II, the experimental
prairie site, a 2nd prairie site, and a shrubby
site (the additional locations previously de-
February 2002
YUNGER—SMALL MAMMAL FOOD AVAILABILITY
scribed) were pooled (Table 4), the proportion of subadults did not differ significantly
(x2 5 3.898, d.f. 5 3, P 5 0.273) from
Woodlot I. Consequently, it is unlikely that
in situ recruitment was responsible for the
numerical increase of mice following mast
production.
There was no significant difference in
survivorship between the 2 woodlots (x2 5
0.045, d.f. 5 1, P 5 0.831; Fig. 4a). Similarly, there was no overall significant difference in survivorship between the 2F 1P,
1F 1P, 2F 2P, 1F 2P plots (x2 5 1.348,
d.f. 5 3, P 5 0.718; Fig. 4b). Mean survivorship on the 4 treatments was 2F 1P,
41.03 days; 1F 1P, 30.77 days; 2F 2P,
38.10 days; 1F 2P, 40.72 days; with the
maximum ranging from 157 to 186 days.
The lack of an overall predator-exclusion
effect on P. leucopus densities was corroborated by one of the planned survival comparisons: there was no significant difference
between survivorship on the 1F 2P and
1F 1P plots (x2 5 1.091, d.f. 5 1, P 5
0.296). For the 2nd planned comparison,
survivorship was significantly greater (x2 5
6.386, d.f. 5 1, P 5 0.012) on the 2F 1P
plots compared with the 1F 1P plots.
Body weight.—There was no significant
effect of increased mast availability on P.
leucopus body weight (F 5 3.13; d.f. 5 1,
47; P 5 0.083) nor significant changes in
body weight over time in the woods (F 5
0.01; d.f. 5 1, 47; P 5 0.945). The beginning (April–July) and end (March–May) of
the time series should be interpreted with
caution (Fig. 5a) because of relatively small
sample sizes. During the central portion of
the time series, body weights between the
two woodlots fluctuated similarly. Individuals tended to weigh more in the low mast
wood lot (Fig. 5a).
Peromyscus leucopus body weights fluctuated significantly during the course of the
food supplementation experiment (Table 2).
On 2F 1P and 2F 2P plots body weights
declined between October and December
while remaining relatively constant on 1F
1P and 1F 2P plots (Fig. 5b). Between
273
FIG. 3.—Plots of significant interactions between food supplementation and predator exclusion. Response variable a) densities and b)
weights were collapsed over time.
March and April mean body weight increased substantially. The significant quadratic term (time2) is a result of the overall
decrease and subsequent increase in mean
body weight. Food supplementation resulted in a significant increase in mean body
274
JOURNAL OF MAMMALOGY
Vol. 83, No. 1
TABLE 3.—Reproductive activity of adult Peromyscus leucopus in Woodlot I, the high-mast woods.
Males
Females
Month
Testes
scrotal
Testes
abdominal
Reproductively
active adults
(%)
Pregnant or
lactating
Not pregnant
or lactating
Reproductively
active adults
(%)
September
October
November
December
6
6
0
0
2
4
27
19
75
60
0
0
8
11
6
0
2
2
13
13
80
84
31
0
weight (Fig. 5b). There were also significant interactions of food 3 time and food
3 time2; body weight increased at a greater
rate on supplemented plots, particularly between March and May. Food supplementation and predator exclusion resulted in a
significant interaction for body weight (Table 2). When collapsed over time, weight
on the 2F 1P treatment was significantly
lower than the other 3 treatments (Fig. 3b).
DISCUSSION
In January of 1994 a severe ice storm
resulted in a drastic decline in Peromyscus
numbers at the field sites. By late summer,
populations had still not approached prestorm abundance. Low numbers were not
isolated to 1 or 2 field sites; the same pattern was observed at several other locations
being trapped across Fermilab (Randa
1996). Populations did not appear to be limited by food, as precipitation and presumably plant productivity were average (Yunger 1996). Also, there was no indication of
a numerical effect resulting from predator
exclusion at the experimental site. Meadow
voles (Microtus pennsylvanicus), generally
considered to be competitively dominant
over Peromyscus (Abramsky et al. 1979;
Grant 1971; Redfield et al. 1977), averaged
8.6 (61.72) individuals per plot prior to the
ice storm. A total of only 11 M. pennsylvanicus were captured at the experimental
site between February 1994 and April 1995,
when food supplementation was completed.
Microtus remained similarly rare across
Fermilab (L. A. Randa and J. A. Yunger, in
litt.). Thus, competitive effects of Microtus
on Peromyscus were probably negligible.
One possible explanation for the absence of
a numerical increase in P. leucopus density
in 1995 is the Allee effect: under extremely
low population densities, dispersion of individuals across the landscape reduces their
likelihood of finding mates (Allee 1938).
However, in late summer 1994, P. leucopus populations increased dramatically as
a result of high mast production on a 6.2ha woodlot. There was no indication that
increased in situ recruitment was responsible for the numerical response. Immigration
remains the best explanation. Even though
there was an adequate food supply for Peromyscus, individuals detected and responded to a particularly abundant patch of food.
As emphasized, the overall abundance of
mice was very low, including oak Woodlot
I prior to mast production. Thus, there was
no social fence (sensu Hestbeck 1982,
1988) to preclude rapid immigration and
subsequent increase in numbers. What is
unclear is why there was not a significant
increase in reproduction, expressed as either
the percentage of reproductively active females or an extension of the breeding season. Such changes in reproduction have
been observed previously following high
mast years (Ostfeld et al. 1996; Wolff
1996). The only difference between my
study and the previous 2 studies is the extremely low densities as a result of the
large-scale disturbance in my study.
The demographic response at the experimental site was similar to that observed in
Woodlot I. No reproductively active individuals were recorded in November; this
%
Subadult
1
1
32
31
Total
%
8
10
54
62
14.8
16.1
Subadult
Total
%
1
1
24
21
4.2
4.8
Subadult
Total
%
Total
22
20
Location
Woodlot I
Alternative sites
Subadult
18.2
10.0
December
November
October
September
TABLE 4.—Proportion of subadults in the total population of Peromyscus leucopus in Woodlot I compared to 4 alternative sites.
275
3.1
3.2
YUNGER—SMALL MAMMAL FOOD AVAILABILITY
4
2
February 2002
FIG. 4.—Survival distribution functions, the
probability of an individual age t having a lifetime exceeding t, for Peromyscus leucopus. Below the abscissa, each symbol for a censored
observation (i.e., right censored) represents an
individual captured at the last trap session and
its fate unknown. a) Survivorship in high and
low mast oak woodlots. b) Survivorship for the
food supplemented and predator excluded experiment.
was also the last month juveniles were observed until the following spring. Consequently, the numerical response may again
be attributed to immigration. The main difference was that the response at the experimental site was substantially more protracted. Two explanations may account for
the difference. First, the January trap census
may not have accurately portrayed the true
densities. There was a noticeable drop in
the number of individuals captured on all 4
treatments. A similar decrease in the number of individuals (compared with December and February) was observed at several
other locations at Fermilab. The decreased
number of captures was accompanied by
moderate, regular snowfall, which, I have
276
JOURNAL OF MAMMALOGY
FIG. 5.—Mean body weight of P. leucopus in
a) oak woodlots with high and low mast production and b) on control (2F 1P), food supplemented (1F 2P), predator excluded (2F
2P), and food supplemented and predator excluded plots (1F 2P). Arrows indicate start of
acorns dropping and food supplementation for
a) and b), respectively; acorns stopped dropping
in mid-October.
observed, can reduce trap success of P. leucopus. Although reduced probability of trap
success contributed to the protracted response, an alternative explanation is responses of small mammals in the area receiving supplemental food. The comparatively large size of Woodlot I (6.2 ha) substantially increased the probability of
individuals locating the abundant food
source, as compared with the 0.6-ha manipulative plots. Following the peak in February, P. leucopus numbers declined steadily
Vol. 83, No. 1
until densities on the supplemented and
nonsupplemented plots converged in May.
Accompanying the decrease in numbers
on food supplemented plots was the earlier
onset of reproductive activity, occurring 2
months before that on nonsupplemented
plots. P. leucopus can use group nests during winter months to help thermoregulate
(Madison et al. 1984; Vogt and Lynch
1982) because individuals do not defend
territories at this time. Ostfeld (1985, 1990)
has shown theoretically that female small
mammals establish territories to defend
patchily distributed food. Pregnant and lactating females have high-energy demands,
and the defense of an abundant food source
can aid in the necessary procurement of energy. Several studies support Ostfeld’s
model (Ostfeld 1985, 1990), including
work with Peromyscus (Wolff 1989). Alternatively, females may defend territories
during the reproductive season to help prevent infanticide (Agrell et al. 1998; Wolff
1993). Irrespective of the mechanism, I
suggest that the ultimate reason individuals
were able to immigrate at a rapid rate onto
the supplemented plots was not only lowdensities but also the absence of territorial
defense. With the onset of reproduction, individuals began to reestablish territories.
This led to the exclusion of conspecifics
and the observed decrease in density. An
analysis of capture points supports this conclusion. During December through February, individuals showed a significantly
clumped distribution on the food supplemented plots (Yunger 1996). During May
through September 1995, a period of active
reproduction, individuals were more evenly
distributed with a mean weighted overlap of
only 17% on the control plots (Yunger
1996).
The limitation of food supplementation
studies to detect only a 2-fold increase in
densities may have been the result of predation. The predator response (pantry effect, sensu Ford and Pitelka 1984) assumes
predators respond to an abundant patch of
prey, much the same as small mammals im-
February 2002
YUNGER—SMALL MAMMAL FOOD AVAILABILITY
migrating into abundant patches of food.
Desy and Batzli (1989) and Desy et al.
(1990) studied the effects of food supplementation and predator exclusion on seeded
populations of prairie voles (Microtus ochrogaster) in a central Illinois bluegrass
field. After 12 and 24 weeks, vole densities
increased significantly in response to food
supplementation and predator exclusion,
with additive and equal effects on density.
The significant interaction was attributed to
predators inhibiting the numerical response
to supplemental food. Krebs et al. (1995)
tested the effects of mammalian predators
and supplemental food on populations of
snowshoe hare (Lepus americanus) in the
Yukon. Snowshoe hare densities showed
similar increases on food supplemented and
predator exclusion plots (1.4–6-fold) with a
multiplicative effect of food 3 predator interaction (X̄ 5 11-fold).
Contrary to these previous studies, the
combined food addition and predator exclusion had neither additive nor multiplicative
effects on P. leucopus densities. Rather,
densities were the highest on 1F 1P plots,
lowest on 2F 1P plots, and intermediate
on the 1F 2P and 2F 2P plots. The pantry
effect may explain the observed responses,
although not in the manner usually expected (i.e., densities greatest on 1F 2P treatments). If the increase in densities on the
1F plots was the result of immigration, and
predators removed a substantial portion of
these individuals as they entered the population, a rapid turnover would result in a
greater number of individuals being captured over time. Two lines of evidence support this conclusion. First is the pattern of
coyote foraging. When Peromyscus densities were at their highest on the 1F 1P
plots in February, so was canid activity; activity declined in relation to a decrease in
prey (L. A. Randa and J. A. Yunger, in litt.).
Over the same period of time, canid activity
and prey densities remained relatively constant on 2F 1P plots. Thus, predators may
have detected the abundant food source.
Second, the significantly greater survivor-
277
ship on 2F 1P as compared to 1F 1P
plots is contrary to usual predictions that an
increase in food availability will result in
increased survivorship. However, it does
support the idea that an increase in predator
foraging resulted in a rapid turnover of individuals and the significant predator 3
food interaction.
Food and predation also significantly influenced P. leucopus body mass. On average, individuals weighed more in supplemented plots. It is unlikely that significant
changes in body mass were an artifact of
the significant numerical responses. The
highest mean body weight was also the
treatment that reached the highest density,
whereas the lowest average body weight
was observed on treatments with the lowest
average density. This differs from the oak
woods where increased mast availability
did not increase body weight. The distinction is that the experimental site was a press
experiment (sensu Bender et al. 1984)
where manipulation is maintained for the
duration of sampling, and the increased
mast production was a pulsed experiment,
where sampling is maintained following a
one-time manipulation. The press provided
additional time for the average weight to
increase. What is surprising is that the increase in body weight was so protracted.
Unlike several previous investigations
(e.g., McShea 2000; Wolff 1996), this study
did not find increased densities the year following greater food availability. The Fermi
study differed from previous ones by low
population densities of Peromyscus following a large-scale disturbance. Future work
simultaneously manipulating densities and
food availability may help elucidate interactions between population dynamics and
disturbance events (Fortier and Tamarin
1998).
ACKNOWLEDGMENTS
I am grateful to P. Meserve who provided
continuous and diverse support throughout this
investigation. Special thanks go to L. Randa for
untiring assistance in the field and support with
278
JOURNAL OF MAMMALOGY
both research and making graduate school enjoyable. P. Klingensmith, P. Meserve, and L.
Randa offered valuable comments that helped to
improve the manuscript. D. Cooper, V. Lagos,
and J. Bell provided valuable field assistance. B.
Patterson, S. Scheiner, P. Sörensen, and C. von
Ende provided support and advice during the design, implementation, and synthesis of the research. Fifteen undergraduate students, in particular C. Joos, assisted with a variety of different
field and laboratory aspects of the investigation.
R. Walton coordinated activities with Fermilab
and helped obtain funding. Funding for this project was provided by the Department of Energy
(MO-RS023), Sigma Xi, the Department of Biological Sciences, Northern Illinois University,
and the Graduate School, Northern Illinois University.
LITERATURE CITED
ABRAMSKY, Z., M. I. DYER, AND D. HARRISON. 1979.
Competition among small mammals in experimentally perturbed area of the short grass prairie. Ecology 60:530–536.
AGRELL, J., J. O. WOLFF, AND H. YLÖNEN. 1998. Counter-strategies to infanticide in mammals: costs and
consequences. Oikos 83:507–517.
ALLEE, W. C. 1938. The social life of animals. Norton,
New York.
BAKER, R. H. 1983. Michigan mammals. Michigan
State University Press, East Lansing.
BENDER, E. A., T. J. CASE, AND M. E. GILPIN. 1984.
Perturbation experiments in community ecology:
theory and practice. Ecology 65:1–13.
BOUTIN, S. 1989. Food supplementation experiments
with terrestrial vertebrates: patterns, problems, and
the future. Canadian Journal of Zoology 68:203–
220.
CONVENTRY, A. F. 1937. Notes on the breeding of some
Cricetidae in Ontario. Journal of Mammalogy 18:
489–496.
COOPER, D. 1997. Foraging responses of the greathorned owl (Bubo virginianus) to changes in prey
abundance. M.S. thesis, Northern Illinois University,
DeKalb.
DESROCHERS, A., S. J. HANNON, AND K. E. NORDIN.
1988. Winter survival and territory acquisition in a
northern population of black-capped chickadees.
Auk 105:727–736.
DESY, E. A., AND G. O. BATZLI. 1989. Effect of food
availability and predation on prairie vole demography: a field experiment. Ecology 70:411–421.
DESY, E. A., G. O. BATZLI, AND J. LIU. 1990. Effects
of food and predation on behavior of prairie voles:
a field experiment. Oikos 58:159–168.
DIAMOND, J. M. 1986. An overview: laboratory experiments, field experiments, and natural experiments.
Pp. 3–22 in Community ecology (J. Diamond and
T. Case, eds.). Harper and Row, New York.
DIJKSTRA, C., L. VUURSTEEN, L. DAAN, AND D. MAS-
MAN.
Vol. 83, No. 1
1982. Clutch size and laying date in the kestrel
Falco tinnunculus: effect of supplementary food.
Ibis 124:210–313.
DUQUETTE, L. S., AND J. S. MILLAR. 1995. The effects
of supplemental food on life-history traits and demography of a tropical mouse Peromyscus mexicanus. Journal of Animal Ecology 64:348–360.
FORD, R. G., AND F. A. PITELKA. 1984. Resource limitation in populations of the California vole. Ecology
65:122–136.
FORTIER, G. M., AND R. H. TAMARIN. 1998. Movement
of meadow voles in response to food and density
manipulations: a test of the food-defense and pupdefense hypothesis. Journal of Mammalogy 79:337–
345.
FOX, G. A. 1993. Failure-time analysis: emergence,
flowering, survivorship, and other waiting time experiments. Pp. 253–289 in Design and analysis of
ecological experiments (S. M. Scheiner and J. G.
Gurevitch, eds.). Chapman and Hall, New York.
GAINES, M. S., N. C. STENSETH, M. L. JOHNSON, R. A.
IMS, AND S. BONDRUP-NIELSEN. 1991. A response to
solving the enigma of population cycles with a multifactorial perspective. Journal of Mammalogy 72:
627–631.
GALINDO-LEALALINDO, C., AND C. J. KREBS. 1998. Effects of food abundance on individuals and populations on the rock mouse (Peromyscus difficilis).
Journal of Mammalogy 79:1131–1142.
GETZ, L. L., J. E. HOFMANN, B. J. KLATT, L. VERNER,
F. R. COLE, AND R. L. LINDROTH. 1987. Fourteen
years of population fluctuations of Microtus ochrogaster and M. pennsylvanicus in east central Illinois.
Canadian Journal of Zoology 65:1317–1325.
GOTTSCHANG, J. L. 1956. Juvenile molt in Peromyscus
leucopus noveboracensis. Journal of Mammalogy
37:516–520.
GRANT, P. R. 1971. Experimental studies of competitive interaction in a two-species system. III. Microtus and Peromyscus species in enclosures. Journal of
Animal Ecology 40:323–350.
HANSEN, L. P., AND G. O. BATZLI. 1978. The influence
of food availability on the white-footed mouse: populations in isolated woodlots. Canadian Journal of
Zoology 56:2530–2541.
HESTBECK, J. B. 1982. Population regulation of cyclic
mammals: the social fence hypothesis. Oikos 39:
157–163.
HESTBECK, J. B. 1988. Population regulation of cyclic
mammals: a model of the social fence hypothesis.
Oikos 52:156–168.
HOFFMEISTER, D. F. 1989. Mammals of Illinois. University of Illinois Press, Urbana.
HUNTER, M. D., AND P. W. PRICE. 1992. Playing chutes
and ladders: heterogeneity and the relative roles of
bottom-up and top-down forces in natural communities. Ecology 73:724–732.
HUSTON, M. 1979. A general hypothesis of species diversity. American Naturalist 113:81–101.
ILLINOIS NATURAL HISTORY SURVEY. 1985. Evaluation
and recommendations concerning biological assessments of habitats and related endangered, threatened, and federally proposed flora and fauna of Fermi National Accelerator Laboratory. Final Report.
Champaign, Illinois.
February 2002
YUNGER—SMALL MAMMAL FOOD AVAILABILITY
KREBS, C. H., ET AL. 1995. Impact of food and predation on the snowshoe hare cycle. Science 269:
1112–1115.
KREBS, C. J., B. L. KELLER, AND R. H. TAMARIN. 1969.
Microtus population biology: demographic changes
in fluctuating populations of Microtus ochrogaster
and M. pennsylvanicus in southern Indiana. Ecology
50:587–607.
LAYNE, J. N. 1968. Ontogeny. Pp. 451–495 in Biology
of Peromyscus (Rodentia) (K. A. King, ed.). American Society of Mammalogists, Special Publication
10:1–593.
LE BOULENGÉ, E. 1987. Package CMR User’s Guide.
Université Catholique de Louvain, Louvain la Neuve, Belgium.
LEE, E. T. 1980. Statistical methods for survival data
analysis. Lifetime Learning, Belmont, California.
LEIBOLD, M. A. 1989. Resource edibility and the effects of predators and productivity on the outcome
of trophic interactions. American Naturalist 134:
922–949.
LICHT, P. 1974. Response of Anolis lizards to food supplementation in nature. Copeia 1974:215–221.
LIDICKER, W. Z., JR. 1991. In defense of a multifactor
perspective in population ecology. Journal of Mammalogy 72:631–635.
LIDICKER, W. Z., JR. 1994. Population ecology. Pp.
323–347 in Seventy-five years of mammalogy
(1919–1994) (E. C. Birney and J. R. Choate, eds.).
American Society of Mammalogists, Special Publication 11:1–433.
MADISON, D. M., J. P. HILL, AND P. E. GLEASON. 1984.
Seasonality in the nesting behavior of Peromyscus
leucopus. American Midland Naturalist 112:201–
204.
MATSON, P. A., AND M. D. HUNTER. 1992. The relative
contributions of top-down and bottom-up forces in
population and community ecology. Ecology 73:
723.
MCSHEA, W. J. 2000. The influence of acorn crops on
annual variation in rodent and bird populations.
Ecology 81:228–238.
OSTFELD, R. S. 1985. Limiting resources and territoriality in microtine rodents. American Naturalist
126:1–15.
OSTFELD, R. S. 1990. The ecology of territoriality in
small mammals. Trends in Ecology and Evolution 5:
411–415.
OSTFELD, R. S. 1994. The fence effect reconsidered.
Oikos 70:340–348.
OSTFELD, R. S., C. G. JONES, AND J. O. WOLFF. 1996.
Of mice and mast: ecological connections in eastern
deciduous forests. BioScience 46:323–330.
OSTROM, C. W., JR. 1990. Time series analysis, regression techniques. SAGE Publications, Newbury Park,
California.
POLLOCK, K. H., J. D. NICHOLS, C. BROWNIE, AND J. E.
HINES. 1990. Statistical inference for capture-recapture experiments. Wildlife Monograph 107:1–97.
POWER, M. E. 1992. Top-down and bottom-up forces
279
in food webs: do plants have primacy? Ecology 73:
733–746.
PULLIAM, R. H. 1988. Sources, sinks, and population
regulation. American Naturalist 132:652–661.
RANDA, L. A. 1996. Prey selectivity and foraging activity of Canis latrans and Vulpes vulpes in response
to prey fluctuations in a heterogeneous landscape.
M.S. thesis, Northern Illinois University, DeKalb.
REDFIELD, J. A., C. J. KREBS, AND M. J. TAITT. 1977.
Competition between Peromyscus maniculatus and
Microtus townsendii in grasslands of coastal British
Columbia. Journal of Animal Ecology 46:607–616.
REICE, S. R. 1994. Nonequilibrium determinants of biological community structure. American Scientist
82:424–435.
SAS INSTITUTE INC. 1990. SAS/STAT user’s guide,
Version 6. Volume 2, 4th ed. SAS Institute Inc.,
Cary, North Carolina.
SCHWEIGER, S., AND S. BOUTIN. 1995. The effects of
winter food addition on the population dynamics of
Clethrionomys rutilus. Canadian Journal Zoology
73:419–426.
SEBER, G. A. F. 1982. The estimate of animal abundance and other related parameters. Griffin, London.
STRONG, D. R. 1992. Are trophic cascades all wet?
Differentiation and donor-control in speciose ecosystems. Ecology 73:747–754.
TAITT, M. J. 1981. The effect of extra food on small
rodent populations: I. Deer mice (Peromyscus maniculatus). Journal of Animal Ecology 50:111–124.
TAITT, M. J., AND C. J. KREBS. 1981. The effects of
extra food on small rodent populations: II. Voles
(Microtus townsendii). Journal of Animal Ecology
50:125–137.
TUKEY, J. W. 1977. Exploratory data analysis. Addison-Wesley Publishing Co., Reading, Massachusetts.
VOGT, F. D., AND G. R. LYNCH. 1982. Influence of ambient temperature, nest availability, huddling, and
daily torpor on energy expenditure in the white-footed mouse Peromyscus leucopus. Physiological Zoology 55:56–63.
WOLFF, J. O. 1985. The effects of density, food, and
interspecific interference on home range size in Peromyscus leucopus and Peromyscus maniculatus. Canadian Journal of Zoology 63:2657–2662.
WOLFF, J. O. 1989. Social behavior. Pp. 271–291 in
Advances in the study of Peromyscus (Rodentia) (G.
L. Kirkland, Jr. and J. N. Layne, eds.). Texas Tech
University Press, Lubbock.
WOLFF, J. O. 1993. Why are female small mammals
territorial? Oikos 68:364–369.
WOLFF, J. O. 1996. Population fluctuations of masteating rodents are correlated with acorn production.
Journal of Mammalogy 77:850–856.
YUNGER, J. A. 1996. Population dynamics of small
mammals. Ph.D. dissertation, Northern Illinois University, DeKalb.
ZAR, J. H. 1984. Biostatistical analysis. Prentice-Hall,
Englewood Cliffs, New Jersey.
Submitted 31 July 2000. Accepted 3 May 2001.
Associate Editor was John G. Kie.