Download niche dynamics of deer mice in a fragmented, old-growth

NICHE DYNAMICS OF DEER MICE IN A FRAGMENTED,
OLD-GROWTH-FOREST LANDSCAPE
MELISSA
A.
SONGER, MARK
V.
LOMOLINO, AND DAVID
R.
PERAULT
Department of Zoology and Oklahoma Biological Survey,
University of Oklahoma, Norman, OK 73019
Peramyscus areas and P. maniculatus were non-randomly distributed across macrohabitats
(including continuous old-growth forests, old-growth corridors, old-growth fragments, and
clearcuts) of the Olympic National Forest, Washington. At this landscape scale, population
densities of P. areas were significantly higher in old-growth sites, particularly continuous
forest and corridors, while densities of P. maniculatus were highest in clearcuts. At a finer
scale, niche breadths measured across local habitats (within each macrohabitat) were broadest in preferred macrohabitat of each species. The potential importance of interspecific
interactions was reflected in an inverse correlation between densities of the two species
across all sites, niche segregation at the landscape (between macrohabitat) and local scales,
and by the relatively high densities of P. areas in forest fragments that lacked P. maniculatus. Relative densities of the two species combined were nearly as high in corridors as
they were in continuous old-growth, and were nearly double that found in fragments and
clearcuts. Given the importance of these rodents as prey for other vertebrates, including
the endangered northern spotted owl (Strix accidentalis caurina), maintaining stands of
continuous old-growth and forested corridors should remain a priority for conserving diversity of native forest communities.
Key words:
Peninsula
Peramyscus, deer mice, fragmentation, corridors, niche, old-growth, Olympic
As Wilcox and Murphy (1985:879) wrote
more than a decade ago, " ... habitat fragmentation is the most serious threat to biological diversity and is the primary cause
of the present extinction crisis." Habitat
fragmentation threatens thousands of plants
and animals now isolated on ever-shrinking
islands of native habitat. The ecological effects of deforestation and other forms of
fragmentation include those associated with
habitat reduction, increased isolation, and
increased edge effects (Harris, 1984; Saunders et aI., 1991; Shafer, 1990). The effects
of fragmentation may be especially severe
for non-volant mammals in comparison to
birds and other species with greater dispersal abilities (Laurance, 1989). For example, in the fragmented, tropical forests of
Queensland, Australia, mammalian diversity decreases as fragments become smaller,
older, more irregular in shape (i.e., more
Journal of Mammalogy. 78(4):1027-1039, 1997
edge), and more isolated (Laurance, 1989,
1990, 1991). Distributions of individual
species among fragments of these and other
forests tend to be correlated with area and
isolation of the fragment (e.g., Bronmark,
1985; Laurance, 1989, 1990; van Dorp and
Opdam, 1987; Verboom and van Apeldoorn, 1990). Fragmentation can directly
threaten many species through loss of habitat, food resources, and shelter, or indirectly by affecting popUlations of predators,
parasites, competitors, and other symbiotes
(Harris, 1984; Saunders et aI., 1991). Thus,
the effects of fragmentation and associated
changes in the landscape may cascade
through the community and result in the
loss of species across many trophic levels.
Understanding how species respond to
changes in these large-scale features is an
important component of both landscape
ecology and conservation biology (Lidick1027
1028
JOURNAL OF MAMMALOGY
Vol. 78, No.4
\l Study Sites
o
10
Kilometers
FIG. I.-Deforestation and fragmentation of old-growth forests of the Hood Canal District of the
Olympic National Forest, Washington (old-growth shown in gray, clearcuts and second-growth are
shown in white, triangles indicate locations of trap sites). Maps were compiled from GIS data provided by the United States Department of Agriculture, Olympic National Forest.
er, 1995). Given this, it may be fruitful, albeit more challenging, to adopt a broader
view of the ecology of fragmentation, one
that includes the effects of environmental
features and interspecific interactions from
local to landscape levels.
The fragmented landscape of the Olympic National Forest offers an excellent opportunity to study the effects of fragmentation on the structure of old-growth forest
communities. This temperate rainforest was
logged extensively between 1940 and 1980,
leaving a patchwork of old-growth forest
(Fig. 1). The Olympic National Forest consists primarily of mid-elevation, Douglas-fir
(Psuedotsuga menziesii) forests encircling
the unfragmented, higher elevation forests
of Olympic National Park. Most of the area
outside the National Park and National Forest is of low elevation and either cleared or
dominated by second-growth forests. The
Olympic National Forest itself is a complex
mixture of patches of stands including
clearcuts, young and old second-growth,
and old-growth. In addition, a few corridors
extend through this landscape along rivers
and streams (Fig. 1).
Across the Olympic Peninsula <20% of
the original old-growth forest is left (Morrison 1989; Norse, 1990). The declines of
at least two officially threatened species, the
northern spotted owl (Strix occidentalis
caurina) and the marbled murrelet (Brachyramphus marmoratus), have been directly
linked to the destruction and fragmentation
of old-growth forests (Carey et aI., 1992;
November 1997
SONGER ET AL.-SPECIAL FEATURE: LANDSCAPE ECOLOGY
Marshall, 1988). In addition, forests of the
Pacific Northwest support one of the most
diverse mammal faunas in the United States
(Corn and Bury, 1991; Norse, 1990), with
mammals comprising >25% of the vertebrate species in this area (>70 species of
mammals on the Olympic Peninsula).
Throughout the forested regions of Washington and Oregon the common deer mouse
(Peromyscus maniculatus) occurs in relatively low numbers. However, the forest
deer mouse (P. oreas) is abundant, especially in old-growth forests of the Olympic Peninsula (Carey and Johnson, 1995).
The presence of these two species of Peromyscus on the Peninsula provides an excellent opportunity to study interactions of
closely related species in a changing landscape.
The hypothesis that competition is a major force in determining distribution and
abundance of species is a persistent theme
in ecology (e.g., Connell, 1983; Dueser and
Shugart, 1978; Grant, 1972; Heske et ai.,
1994; Schoener, 1974, 1983; Scott and
Dueser, 1992; Adler, 1985). Many studies,
especially those of insular or otherwise isolated communities, have shown that species
may shift their realized niches in the absence of putative predators or competitors
(e.g., Crowell and Pimm, 1976; Lomolino,
1984; MacArthur et ai., 1972; Munger and
Brown, 1981; Schoener and Spiller, 1987).
Because fragmentation results in insularization and loss of key competitors or predators, species occupying fragments also
may exhibit niche shifts analogous to populations on real islands.
While many studies have focused on effects of area and isolation on the species
composition of forest fragments (Burkey,
1989; Cutler, 1991; Malcolm, 1988; Quinn
and Harrison, 1988; Simberloff and Abele,
1982), such changes may not account for
all effects of fragmentation on a forest community. Habitat quality of the remaining
forest patches, corridors, and the surrounding matrix may all influence the diversity
of old-growth-forest mammals. Perhaps just
1029
as important, clearcutting may well result
in range expansion of generalists and early
successional species, such as P. maniculatus, at the expense of P. oreas and other
species dependent on old-growth forests.
The purpose of this study is to test the
hypothesis that mammals inhabiting oldgrowth forests are strongly influenced by
both direct and indirect effects of fragmentation. Specifically, we focus on distributions and densities of deer mice and test the
following predictions: 1) when viewed
across the entire fragmented landscape (irrespective of habitat), densities of P. oreas
and P. maniculatus should be inversely correlated; 2) across macrohabitats (continuous
old-growth, corridors, fragments, or clearcuts), the two species should segregate their
spatial niches, with P. oreas exhibiting
higher densities in stands of old-growth
while densities of P. maniculatus should be
higher in clearcuts; 3) within macrohabitats,
the two species should segregate their realized niches, with P. oreas exhibiting
higher densities at sites with more closed
canopies while densities of P. maniculatus
should be higher at sites with open canopies
and generally lower forest development; 4)
when viewed across dominant macrohabitats of this landscape (here old-growth, corridors, forest fragments, and clearcuts),
niche breadths should be higher for the habitat generalist, P. maniculatus, than for the
old-growth specialist, P. oreas; 5) just as
on true islands, P. oreas should exhibit ecological release in fragments lacking its putative competitor (P. maniculatus). In addition to testing these predictions, we also
discuss the potential importance of oldgrowth corridors and the relevance of our
findings with respect to conservation of
old-growth mammals such as P. oreas and
dependent, predatory species including
spotted owls.
MATERIALS AND METHODS
Description of research area.-Studies were
conducted in the Hood Canal District of Olympic National Forest (Fig. 1; Songer, 1996). The
\030
Vol. 78, No.4
JOURNAL OF MAMMALOGY
dominant species of tree in this area are Douglas
firs (P. menziesii), sitka spruce (Picea sitchensis), western red cedars (Thuja plicata), western
hemlocks (Tsuga heterophylla), mountain hemlocks (Tsuga mertensiana), and white firs (Abies
concolor). Continuous old-growth forest, as defined by the Old-growth Definition Task Group
(1986) for the Pacific Northwest, constitute areas
>50 km2 , with ;:::8 trees/O.4 ha older than 200
years or > 15 cm dbh, and with a deep multilayered canopy having at least four conifer snags
of 15 m long.
Trapping surveys and habitat analyses.Field studies were conducted June-August, 1994
and 1995. Live-trapping was carried out within
each of four macrohabitats; continuous oldgrowth forest, old-growth fragments, old-growth
corridors, and clearcuts. Areas that met the
above definition of old-growth were considered
continuous old-growth forest. Fragments were
defined as old-growth forests completely surrounded by clearcuts and second-growth, and
ranged from 0.1 to 2.0 km2 • Corridors were defined as strips of old-growth forest <0.5 km
wide, at least 3 km in length, and connected to
continuous old-growth. Clearcuts were defined
as areas cut within the past 25 years.
During each trapping session, traps were set
at five sites in continuous old-growth. Forest
fragments also were trapped during most Sessions. Three to five sites were set up within each
fragment, covering the extent of all but the largest fragments. Two trap sites were set up within
corridors and paired with two trap sites in adjacent clearcuts. Corridor and adjacent clearcut
sites were always monitored during the same
session. To promote independence of samples,
all trap sites were spaced 75 m apart and recaptures were excluded from all statistical analyses.
Trap sites also were situated ;:::75 m from the
nearest forest edge. During the study, trapping
was conducted at a total of 35 sites in oldgrowth, 71 in fragments, 22 in corridors, and 22
in adjacent clearcuts.
At each site, four Sherman traps were set
within 5 m of the center of the site in a variety
of microhabitats and baited with peanut butter
and oats. The traps were locked open for 5 days
before each trapping session, then rebaited, unlocked, and checked daily for the next 7 days.
Three trapping sessions were conducted during
1994 and four during 1995. Each small mammal
caught was sexed, weighed, aged, measured for
'TABLE i.-Habitat characteristics used· in
principal-compoments analysis.
Abbreviation
CANC
EDGE
SLOPE
SNAG
Description
Canopy closure (measured with spherical densiometer)
Distance to nearest edge (forest or
clearcut)
Percentage of slope (measured with a
clinometer)
Number of snags present within lO-m
radius
T20,
T2040,
T40
L20,
L2040,
L40
S20,
S2040,
S40
Number of trees with a dbh of <20 cm,
20-40 cm, and >40 cm, respectively,
within a 10-m radius
Number of logs with a dbh of <20 cm,
20-40cm, and >40 cm, respectively,
within a lO-m radius
Number of stumps with a dbh of :s2Ocm,
20-40 cm, and >40 cm, respectively,
within a lO-m radius
MOSS
FERN
GRASS
ROCK
SHRUB
HERB
LITTER
TLOG
Frequency
Frequency
Frequency
Frequency
Frequency
Frequency
Frequency
Frequency
in plot
Frequency
in plot
SOIL
of moss at 22 points in plot
of ferns at 22 points in plot
of grass at 22 points in plot
of rock at 22 points in plot
of shrub at 22 points in plot
of herb at 22 points in plot
of litter at 22 points in plot
of a log or tree at 22 points
of exposed soil at 22 points
length, and assessed for reproductive status. The
animals were then marked by toe-clipping and
released.
At each trap site, 22 habitat characteristics
were recorded during the trapping session (Table
1). Two lO-m ropes, knotted at I-m intervals,
were placed along the cardinal directions crossing at 90° angles at the center of the site. Under
each knot the presence of dominant habitat components was recorded including litter, rock, fern,
moss, herbaceous plant, shrub, stump, log, or
tree. Also, the number and size of trees, logs,
and stumps were counted within a lO-m radius
of the plot center. Categories included snags and
trees, stumps, and logs that were <20 cm dbh,
20-40 cm dbh, and >40 cm dbh. Canopy closure
was measured by use of a spherical densiometer.
A clinometer was used to estimate slope and
canopy height. The distance from the site to the
nearest edge of the macrohabitat also was recorded.
November 1997
SONGER ET AL.-SPECIAL FEATURE: LANDSCAPE ECOLOGY
Statistical analyses.-Relative densities for
each species were calculated by dividing the
number of individual mice captured (excluding
recaptures) by the number of functional trapnights. Functional trapnights were calculated by
subtracting from the total potential number of
trapnights, 1.0 for traps that were not functional,
subtracting 0.5 for traps that were disturbed,
missing bait, or containing a recaptured individual (Songer, 1996).
To examine habitat selection across macrohabitats, we compared relative densities (individuals per functional trapnight) for each species
within macrohabitat categories (continuous oldgrowth, corridors, fragments, and clearcuts) to
the proportion of trapping effort within that
macrohabitat. Here we define habitat selection
as a significant difference between use and
availability of resource (Ludwig and Reynolds,
1988). We used goodness-of-fit tests to compare
the number of individuals captured in each habitat to the number expected. The expected number of individuals was calculated by multiplying
total number of individuals caught for a species
by the proportion of functional trapnights in a
particular macrohabitat. A significant chi-square
value would indicate that the species was not
randomly distributed among macrohabitats. We
then tested the null hypothesis that the two species of Peromyscus were selecting the same type
of macrohabitats with a test of independence
(Sokol and Rohlf, 1981). A significant G-value
would indicate that P. oreas and P. maniculatus
were selecting different habitat types (Le., segregating their realized niches).
To examine selection within macrohabitats,
we first used principal-components analysis
(SYSTAT, Inc., 1992) to summarize habitat variation among sites within each macrohabitat and
then assign each trap site to one habitat category. We then tested whether deer mice were
randomly distributed among habitat categories.
Factor scores were obtained for each variable
and principal components were calculated. For
the continuous old-growth sites, factor I explained 19.5% of the variation and was a measure of forest structure, loading strongly on
T2040, T40, S40, CANC, SNAG, and L2040,
and MOSS (for definitions of habitat variables
see Table 1). Several of the variables with the
highest loadings in analyses of continuous oldgrowth (e.g., snags and large logs) were those
used to define old-growth (Old-Growth Defini-
1031
tion Task Group, 1986). Factor 2 explained
12.9% of the variation and was a measure of
large trees and coarse woody debris in the understory, loading strongly on TLOG, T40, and
S20. Combined, the factor scores explained
32.3% of variation among old-growth sites. The
sites were then classified into one of three new
habitat categories using KMEANS cluster analysis (SYSTAT, Inc., 1992).
For the corridor sites, GRASS and MOSS
variables were removed because there was no
grass or moss found in the corridor sites. Factor
1 explained 25.0% of the variation and was a
measure of forest structure, loading strongly on
CANC, SNAG, L40, T2040, SHRUB, SOIL,
T40, LITTER, S20, and HERB. Factor 2 explained 15.7% of the variation and was a measure of openness of the understory, loading
strongly on L20, TLOG, S2040, and LITTER.
Together the factor scores explained 40.6% of
the variation among corridor sites. The sites
were then classified into one of three new habitat
categories using KMEANS cluster-analysis.
For the fragments, the MOSS variable was removed because there was no moss found in the
fragment sites. Factor 1 explained 14.9% of the
variation and was a measure of forest structure,
loading strongly on CANC, SOIL, L40, T2040,
and S40. Factor 2 explained 13.1 % of the variation and was largely a measure of coarse
woody debris in the understory, loading strongly
on S20, S2040, L2040, L20, and T2040. Together the factor scores explained 28.0% of the
variation. The fragments were divided into four
new habitat categories using KMEANS clusteranalysis.
For clearcuts, the variables T40, T2040, and
MOSS were eliminated because they were not
found in the clearcut sites. Factor 1 explained
21.9% of the variation and was a measure of
coarse woody debris and distance to the edge,
loading strongly on L2040, L20, EDGE, TLOG,
SOIL, S40, and S20. Factor 2 explained 15.3%
of the variation and was a measure of openness
of the understory, loading strongly on T20,
SLOPE, LITTER, and L40. Together the factor
scores explained 37.2% of the variation. Clearcut sites were then classified into one of three
new habitat categories using KMEANS clusteranalysis.
To calculate niche breadth for each species of
Peromyscus across or within macrohabitat cate-
JOURNAL OF MAMMALOGY
1032
0.6
Vol. 78, No.4
•
0.5 -
0
0
0
CI)
ctI
@
0
-
0.4 -
Q
0
0
a
~ 0.3 tn
~
c:
CI)
c
CI)
> 0.2 ~
co
CI)
0::
0.1 -
0.0 -
e
0
I
•
0
0
o
I
•
0
0
s
Ii
0
0
0
0
•
• •
0
0
0
0
0
em
0
I
I
I
I
I
0.00
0.04
0.08
0.12
0.16
•
0.20
Relative Density of P. manicu/atus
FIG. 2.-Relative densities of Peromyscus oreas and P. maniculatus were inversely correlated
across trap sites (rs = Spearman rank. correlation = -0.21, P < 0.05). Relative densities were
calculated by dividing the number of individuals captured by the number of functional trapnights.
Trap sites lacking both species were excluded from this figure and from tests of interspecific correlations. The MGLH procedure of SYSTAT (1992) identified seven outliers or points with undue
influence (triangles) that were excluded from statistical analysis.
gories, we used an electivity measure of niche
breadth (NB; Feinsinger et aI., 1981):
NB = 1 _ ~ ABS(<t - Pi)
i~l
2
where qi = the proportion of trap nights in habitat category i, Pi = the proportion of individuals
captured within category i, and Nh = the number
of habitat categories. Niche breadth ranges from
zero for a perfect specialist to one for a perfect
generalist.
RESULTS
We recorded a total of 1,414 captures of
980 individual small mammals during the
study. Sixty percent of newly caught indi-
viduals were Peromyscus. Species captured
in order of abundance were P. oreas
(50.2%), Clethrionomys gapperi (24.6%),
Tamias townsendi (8.8%), and P. maniculatus (5.8%). The remaining species combined made up 11 % of the captures, with
no single species accounting for >2% of
total captures. These included Sorex monticolus, S. trowbridgei, Mustela erminea,
Zapus trinotatus, Spilogale putorius, Glaucomys sabrinus, Neotoma cinerea, Tamiasciurus douglasi, Microtus longicaudus, and
Mustela Jrenata. Overall, P. oreas represented 89% of newly caught Peromyscus
individuals.
November 1997
>-
SONGER ET AL.-SPECIAL FEATURE: LANDSCAPE ECOLOGY
1033
0,
.t::
Vl
c
Q)
C
Q)
>
'';:::;
~
Q)
a:
P. oreas (NB = 0.84)
Clearcut
P. maniculatus (NB
Fragment
= 0.66)
Corridor
OldGrowth
FIG. 3.-Relative densities (number of individu'als per functional trapnight) for Peromyscus oreas,
P. maniculatus, and both species combined within each macrohabitat. Niche breadth (NB), calculated
from a modified electivity measure (Feinsinger et ai., 1981), is shown for each species across macrohabitats.
Population densities and distributions
among macrohabitats.-Relative densities
of P. oreas and P. maniculatus were inversely correlated when viewed across all
sites combined (rspearman = -0.21, P < 0.05;
Fig. 2). This result stems in large part from
niche segregation of these species at the
landscape level. That is, P. oreas and P.
maniculatus exhibited significant segregation of their spatial niches across macrohabitats (old growth, corridors, fragments,
and clearcuts; Fig. 3). Goodness-of-fit tests
showed significant macrohabitat selection
for both species (P < 0.001, K = 64.76 and
55.57 for P. oreas and P. maniculatus, respectively, d.f = 3). A test of independence
showed that the two species were selecting
different types of macrohabitats (P < 0.001,
G = 68.50, d.f = 3). As predicted, P. oreas
showed a strong preference for old-growth
and had its lowest density in the clearcuts
(Fig. 3). Densities of P. oreas also were rel-
atively high in corridors, intermediate in
fragments, and lowest in clearcuts. In contrast, relative densities of P. maniculatus
were highest in clearcuts, but low in all other macrohabitats. Counter to our predictions, however, P. oreas exhibited the
broadest niche breadth across macrohabitats
(NB = 0.84 and 0.66 for P. oreas and P.
maniculatus, respectively).
Habitat selection at the local scale.Patterns of habitat selection and niche segregation at the local scale differed markedly
among macrohabitats (Fig. 4). For example,
P. oreas failed to exhibit significant habitat
selection within old-growth and corridor
stands (NB = 0.918 and 0.99, K = 5.34
and 0.08, respectively; d.f = 2, P ~ 0.05).
In contrast, this species exhibited relatively
narrow niche breadths in fragments and
clearcuts, with its highest density occurring
in fragment sites with more open canopies
(ca. 35% canopy cover) and relatively few
1034
JOURNAL OF MAMMALOGY
snags, large logs, or trees, and in clearcut
sites located closest to old-growth forests
(cluster a in Fig. 4D; NB = 0.63 and 0.65,
K = 116.52 and 17.03, respectively; d.f. =
2, P < 0.001).
Just as P. oreas exhibited its broadest
niche breadth within its preferred macrohabitats (i.e., old-growth and corridors), P.
maniculatus exhibited its broadest niche in
clearcuts (NB of P. maniculatus = 0.89 in
clearcuts versus 0.53, 0.54, and 0.54 in
fragments, corridors, and old-growth, respectively). Sample sizes were too small to
justify statistical tests in all but clearcut
sites. Here, however, P. maniculatus exhibited highly significant selectivity for clearcut sites with few snags, large logs or trees,
as well as sites with a relatively steep slope,
a high coverage of small to medium-sized
logs and a limited amount of litter. Tests of
independence, which again could only be
conducted for sites in clearcuts, indicated
significant niche segregation between P. oreas and P. maniculatus in clearcuts (G =
8.92, d.f. = 2, P < 0.025).
Niche shifts and ecological release.Mammals often exhibit ecological release
on true islands, occupying atypical habitats
at relatively high densities. Here we tested
whether P. oreas exhibited ecological release in forest fragments lacking its putative
competitor, P. maniculatus. Although its
densities were consistently low (typically
<0.08 individuals/trapnight for all but one
site), P. maniculatus was detected in seven
of 16 fragments studied (1,104 and 781
trapnights in fragments with and without P.
maniculatus, respectively). As in the above
analyses, principal components and cluster
analyses were used to assign each trap site
to one of five habitat clusters. The treatment
with P. maniculatus present lacked sites in
two of the five clusters. Therefore, only the
three common clusters were used to assess
the effects of P. maniculatus on habitat distributions of P. oreas within isolated fragments.
In syntopy, both species exhibited higher
densities at fragment sites with more open
Vol. 78, No. 4
canopies (cluster a in Fig. 5 and Fig. 4C).·
Densities of P. oreas at these sites, however, were much higher in the absence of
P. maniculatus (relative density of P. oreas
= 0.37 versus 0.23 for similar sites without
versus those with P. maniculatus, respectively; cluster a in Fig. 5). Interestingly,
niche breadth of P. oreas actually decreased
in the absence of P. maniculatus (NB =
0.73 versus 0.52 in syntopy versus in allotopy, respectively).
DISCUSSION
The results of these studies generally are
consistent with the hypothesis that mammals inhabiting old-growth forests may be
strongly influenced by both direct and indirect effects of fragmentation. The results
also confirm the strong association of P. oreas with old-growth forests and P. maniculatus with clearcuts. Thus, these species
exhibit highly significant niche segregation,
primarily at the landscape level (i.e., across
macrohabitats). As may be typical of most
landscapes, the source for one species may
serve as a sink for others. Indeed, niche
segregation appears to be so strong that
these species may have only limited opportunity to compete in syntopy. It is possible
that competition may, however, contribute
to the relatively low densities of P. oreas
in clearcuts and fragments. The limited,
natural experiment conducted in fragments
suggests that P. maniculatus may be inversely related to densities of P. oreas, but
manipulative field experiments are required
to rigorously test any causal explanation for
this pattern.
Interestingly, P. oreas exhibited a much
broader niche breadth than P. maniculatus
when measured across all macrohabitats,
while both species exhibited relatively
broad niches within their principal habitats
(P. oreas in continuous old-growth and corridors, P. maniculatus in clearcuts). Relatively broad niches in these macrohabitats
may stem from high densities and a tendency for individuals to spill over into otherwise suboptimal, local habitats. While sur-
November 1997
SONGER ET AL.-SPECIAL FEATURE: LANDSCAPE ECOLOGY
1035
B - Corridors
A - Continuous Old-Growth
0.30 -r-r===== = = == = =1l
~ P. areas (NB = 0.92)
c:::J P. maniculatus (NB = 0.54)
0.25
0.30
I c=J
0.25 -
~0 . 20
en
c
0.20
oQ) 0.15
0.15
-
0.10
0.1 0
-
0.05
0.05
-
c::=J
P. areas (NB =0.99)
P. maniculatus (NB =0.54)
I
-----;
r-'""
r-
Q)
>
~
Q)
n:::
0.00
-'-----'=-'-'-----"-""---'---'--" ' - - - - - - '
a
..
--,
0.00
a
c
b
c
b
>
Forest Structure
D - Clearcuts
C - Fragments
0.30 -r-r==== = = == =1l
0.25
>
Forest Structu re
IL'2J P. areas (NB = 0.63)
c:::J P. maniculatus (NB =0.53)
>-
0.30 -r-r===========iI
~ P. areas (NB =0.65)
0.25
c::=J
P. maniculatus (NB
=0.89)
0.20
'V5 0.20
c
Q)
o
0.15
0.15
~
&0.10
0.10
Q)
>
0.05
a
b
c
a
d
>
Forest Structure
b
c
>
near
forest
logs
cut stumps
exposed soil
FIG. 4.-Pattems of habitat selection and niche segregation for Peramyscus areas and P. maniculatus within four macrohabitats. The variable forest structure increases with canopy closure and
with the occurrence of snags, moss, large trees, or logs (Figs. 4A, B, and C). The abscissa in Fig.
4-D is a direct measure of distance from nearest forest and exposed soil, medium- to large-sized logs
and large, and cut stumps (thus, sites included in clearcut cluster a were situated closest to old-growth
forests). Niche breadth (NB) within each macrohabitat was calculated from a modified electivity
measure (Feinsinger et aI., 1981). Niche breadth of each species was highest in their principal habitat
(continuous old-growth and corridors for P. areas, and clearcuts for P. maniculatus).
1036
JOURNAL OF MAMMALOGY
0.4
/
/
o Fragments with P. maniculatus
/
INB =0.73)
II)
III
I!!C)
0.3
~
-...
>';;;
I:
o Fragments without P. maniculatus
V
0
0.2
V
0.1
/
Vol. 78, No.4
r-
INB =0.52)
V-
Q)
0
Q)
>
.;;
ra
Qj
a:
0
/
/
a
.
/
/
b
/
.
/
I
/
I
I
0
c
Habitat Clusters
Open canopy
Closed canopy
FIG. 5.-Relative densities (number of individuals per functional trapnight) of Peramyscus areas
in forest fragments with and without its putative competitor, P. maniculatus. NB refers to niche
breadth, calculated using a modified electivity measure (Fein singer et aI., 1981). Relative densities
of P. areas equaled 0.37 versus 0.23 in fragment sites without, versus those with P. maniculatus,
respectively.
prising and counter to our prediction, these
results serve to emphasize the scalar nature
of ecology. Patterns in niche breadths, habitat affinities, and densities of these species
differ among macrohabitats and across spatial scales.
Viewed at the landscape scale, however,
these distributional and niChe dynamics are
not surprising. In fact, they are equivalent
to those of two subspecies of P. maniculatus in response to clearing of eastern forests of North America during the 19th century. During this earlier period of deforestation, P. m. bairdi expanded its range eastward in concert with expansion of its
grassland habitats (Baker, 1968; Brown and
Gibson, 1983 ; Harris, 1952; Hooper, 1942;
Wecker, 1963, 1964). In contrast, P. m.
gracilis contracted its range during this period as its principal habitat, mature mesic
forests, was reduced. In the case of oldgrowth rainforests of the Pacific Northwest,
however, deforestation is more recent, more
rapid (Fig. 1), and has more serious implications for loss of biodiversity.
Implications for conserving biodiversity
of old-growth forests.-Although corridors
typically are considered old-growth habitats, they differ in a number of important
features which, in tum, affect community
structure. In comparison to continuous oldgrowth, corridors are narrower and are connected to continuous old-growth forest,
have a higher ratio of edge to interior habitat, and are more susceptible to wind damage and other edge effects. Habitat analysis
in our study showed an absence of moss in
corridors, while it was common component
of continuous old-growth sites. Despite
these differences, densities of P. oreas in
corridors were nearly equal to those in continuous old-growth, and much higher than
those in fragments (0.20, 0.19, and 0.11, respectively). Apparently, connectivity and
dispersal may compensate for the subtle
differences in habitat quality of corridors
and continuous old-growth.
November 1997
SONGER ET AL.-SPECIAL PEATURE: LANDSCAPE ECOLOGY
In contrast, fragments, which differ both
in habitat quality and in their isolation from
corridors and continuous old-growth, harbor substantially lower populations of P.
oreas. Habitat quality of fragments seems
less suitable for this forest mammal. In addition, because they are isolated by clearcuts, populations inhabiting fragments are
less likely to be supplemented, or rescued,
by emigration from source populations
(sensu Brown and Kodric-Brown, 1977).
Thus, we predict that fragments farther
from corridors or continuous old-growth
will support lower densities of P. oreas and
other forest mammals. Additional studies
presently are underway to test this and related predictions on the role of riparian corridors in this fragmented landscape.
Finally, we consider the relevance of observed patterns in densities and distributions of deer mice for conservation of other
animals occupying old-growth forests, especially predators. Old-growth is the principal habitat of the endangered northern
spotted owl. About 20% of fecal pellets of
owls collected in the Olympic Peninsula
contain remains of P. maniculatus and P.
oreas (Carey et aI., 1992; Forsman et aI.,
1991). Unfortunately, it is sometimes assumed that clearcuts support more individuals of Peromyscus than do forested habitats (Algren, 1966; Corn and Bury, 1991;
Gashwiler, 1959, 1970; Martell and Radvanyi 1977; Sullivan, 1979; Tevis, 1956).
Other studies have reported similar densities and demographic characteristics of populations of Peromyscus between old-growth
and recently logged areas in British Columbia (Petticrew and Sadleir, 1974; Sullivan,
1979). However, P. oreas did not occur in
these regions.
Our research in the Olympic National
Forest revealed that the combined density
of these key prey species was highest in
continuous old-growth forest and oldgrowth corridors (Figure 3). In fact, densities of Peromyscus in the continuous oldgrowth were nearly twice as high as densities in the clearcuts (0.20 and 0.11, re-
1037
spectively). Not only are densities of deer
mice higher, but we suspect that availability
of these and other prey also may be higher
in forested habitats. As an arboreal species,
P. oreas also is likely to be a more accessible prey for the spotted owl than P. maniculatus, which typically restricts its movements to the lowest stratum of the forest.
Moreover, the northern flying squirrel
(Glaucomys sabrinus), which makes up the
greatest percentage of biomass in the diet
of the northern spotted owl (Forsman et aI.,
1991), spends most of its time in the trees.
In summary, deforestation and fragmentation of old-growth forests of the Olympic
Peninsula clearly have had a strong impact
on distributions, densities, and niche dynamics of deer mice. Admittedly, we have
much to learn as this is just the first product
of our community wide studies on this
complex system. It seems quite clear, however, that as this landscape becomes more
fragmented, the combined densities of these
key prey species will continue to decline,
setting the stage for cascading effects and
continued decline of old-growth-forest
communities in general.
ACKNOWLEDGMENTS
We thank T. A. Franklin, A. Leikam, P. Leimgruber, M. Rene, and B. Rosewell for their assistance in collecting field data. G. A. Smith, T.
A. Franklin, and two anonymous reviewers provided valuable comments on an earlier version
of this manuscript. M. L. Johnson, N. Czaplewski, J. Lowry, and C. C. Vaughn aided with valuable comments throughout the project. GIS data
were provided by W. Wettengel of the United
States Department of Agriculture, Olympic National Forest. E. Milliman assisted with logistical support from Olympic National Forest. Field
housing was furnished by T. Neilson at the Satsop Wells Environmental Learning Lodge of the
Grays Harbor Conservation District. Funding for
this project was provided by a National Science
Foundation grant (DEB-9322699) to M. V. Lomolino.
LITERATURE CITED
G. H. 1985. Habitat selection and species interactions: an experimental analysis with small
mammal populations. Oikos, 45:380-390.
ADLER,
1038
JOURNAL OF MAMMALOGY
ALGREN, C. E. 1966. Small mammals and reforestation following prescribed burning. Journal of Forestry, 64:614-618.
BAKER, R. H. 1968. Habitats and distribution, pp. 98126, in Biology of Peromyscus (J. A. King, ed.).
Special Publication, The American Society of Mammalogists, 2:1-593.
BRONMARK, C. 1985. Freshwater snail diversity: effeGts of pond area, habitat heterogeneity and isolation. Oecologia, 67:127-131.
BROWN, J. H., AND A. C. GIBSON. 1983. Biogeography.
C. V. Mosby Company, St. Louis, Missouri, 643 pp.
BROWN, J. H., AND A. KODRIC-BROWN. 1977. Turnover
rates in insular biogeography: effect of immigration
on extinction. Ecology, 58:445-449.
BURKEY, T. V. 1989. Extinction in nature reserves: the
effect of fragmentation and the importance of migration between reserve fragments. Oikos, 55:7581.
CAREY, A. B., AND M. L. JOHNSON. 1995. Small mammals in managed, naturally young, and old-growth
forests. Ecological Applications, 5:336-352.
CAREY, A. B., S. P. HORTON, AND B. L. BISWELL. 1992.
Northern spotted owls: influence of prey base and
landscape character. Ecological Monographs, 62:
223-250.
CONNELL, J. H. 1983. On the prevalence and relative
importance of interspecific competition: evidence
from field experiments. The American Naturalist,
122:661-696.
CORN, P. S., AND R. B. BURY. 1991. Small mammal
communities in the Oregon Coast Range. pp. 241256, in Wildlife and vegetation of unmanaged Douglas-fir forests (L. E Ruggiero, K. B. Aubry, A. B.
Carey, and M. H. Huff, technical coordinators).
United States Department of Agriculture, Forest Service, General Technical Report, PNW-GTR-285:1533.
CROWELL, K. L., AND S. L. PIMM. 1976. Competition
and niche shifts of mice introduced onto small islands. Oikos, 27:251-258.
CUTLER, A. 1991. Nested faunas and extinction in
fragmented habitats. Conservation Biology, 5:496505.
DUESER, R. D., AND H. H. SHUGART, JR. 1978. Microhabitats in a forest-floor small mammal fauna. Ecology, 59:89-98.
FEINSINGER, P., E. E. SPEARS, AND R. W. POOL. 1981.
A simple measure of niche breadth. Ecology, 62:2732.
FORSMAN, E. D., 1. OTTO, AND A. B. CAREY. 1991.
Diets of the spotted owls on the Olympic Peninsula,
Washington and the Rosenburg District, Bureau of
Land Management. P. 527, in Wildlife and vegetation of unmanaged Douglas-fir forests (L. F. Ruggiero, K. B. Aubry, A. B. Carey, and M. H. Huff,
technical coordinators). United States Department of
Agriculture, Forest Service, General Technical Report, PNW-285:1-533.
GASHWILER, J: S. 1959. Small mammal study in westcentral Oregon. Journal of Marnmalogy, 40:128139.
- - - . 1970. Plant and mammal changes on a c1earcut in west-central Oregon. Ecology, 70:1018-1026.
GRANT, P. R. 1972. Interspecific competition among
Vol. 78, No.4
-rodents. Annual Review of Ecology and Systematics, 3:79-106.
HARRIS, L. D. 1984. The fragmented forest: island
biogeographic theory and the preservation of biotic
diversity. The University of Chicago Press, Chicago
211 pp.
HARRIS, V. T. 1952. An experimental study of habitat
selection by prairie and forest races of the deer
mouse, Peromyscus maniculatus. Contributions of
the Laboratory of Vertebrate Biology, University of
Michigan, 56:1-53.
HESKE, E. J., J. H. BROWN, AND S. MISTRY. 1994.
Long-term experimental study of a Chihuahuan Desert rodent community: 13 years of competition.
Ecology, 75:438-445.
HOOPER, E. T. 1942. An effect on the Peromyscus
maniculatus Rassenkreis of land utilization in Michigan. Journal of Mammalogy, 23:193-196.
LAURANCE, W. F. 1989. Ecological impacts of tropical
forest fragmentation on non-flying mammals and
their habitats. Ph.D. assertation University of California, Berkeley, 501 pp.
- - - . 1990. Comparative responses of five arboreal
marsupials to tropical forest fragmentation. Journal
of Mammalogy, 71:641-653.
- - - . 1991. Edge effects in tropical forest fragments: application of a model for the design of nature reserves. Biological Conservation, 57:205-219.
LIDICKER, W. Z., JR. 1995. The landscape concept:
something old, something new. Pp. 3-19, in Landscape approaches in mammalian ecology and conservation (W. Z. Lidicker, Jr., ed.). University of
Minnesota Press, Minneapolis, 215 pp.
LOMOLINO, M. V. 1984. Immigrant selection, predation, and the distributions of Microtus pennsylvanicus and Blarina brevicauda on islands. The American Naturalist, 123:468-483.
LUDWIG, J. A., AND J. F. REYNOLDS. 1988. Statistical
ecology: a primer on methods and computing. John
Wiley & Sons, New York, 337 pp.
MACARTHUR, R. H., J. M. DIAMOND, AND J. KARR.
1972. Density compensation in island faunas. Ecology, 53:330-342.
MALCOLM, J. R. 1988. Small mammal abundances in
isolated and non-isolated primary forest reserves
near Manaus, Brazil. Acta Amazonica, 18:67-83.
MARSHALL, D. B. 1988. The marbled murrelet joins
the old-growth forest conflict. American Birds, 42:
202-212.
MARTELL, A. M., AND A. RADVANYI. 1977. Changes
in small mammal populations after c1earcutting of
northern Ontario black spruce. The Canadian FieldNaturalist, 91 :41-46.
MORRISON, P. H. 1989. Ancient forests of the Olympic
National Forest: analysis from a historical and landscape perspective. Wilderness Society, Washington,
D.C., 21 pp.
MUNGER, J. c., AND J. H. BROWN. 1981. Competition
in desert rodents: an experiment with semipermeable
exc1osures. Science, 211:510-512.
NORSE, E. A. 1990. Ancient forests of the Pacific
Northwest. Island Press, Washington D.C., 327 pp.
OLD-GROWTH DEFINITION TASK GROUP. 1986. Interim
definitions for old-growth Douglas-fir and mixed-conifer forests in the Pacific Northwest and California
November 1997
SONGER ET AL.-SPECIAL FEATURE: LANDSCAPE ECOLOGY
(Franklin, J. F., et al.). United States Department of
Agriculture Forest Service, Pacific Northwest Research Station Note, PNW-447:1-7.
PETTICREW, B. G., AND M. F. SADLEIR. 1974. The ecology of the deer mouse Peromyscus maniculatus in
a coastal coniferous forest. 1. Population dynamics.
Canadian Journal of Zoology, 52:107-118.
QUINN, J. F., AND S. P. HARRISON. 1988. Effects of
habitat fragmentation and isolation on species richness: evidence from biogeographic patterns. Oecologia, 75:132-140.
SAUNDERS, D. A., R. J. HOBBS, AND C. R. MARGULES.
1991. Biological consequences of ecosystem fragmentation: a review. Conservation Biology, 5: 1832.
SCHOENER, T. W. 1974. Resource partitioning in ecological communities. Science, 185:27-39.
1983. Field experiments on interspecific
competition. The American Naturalist, 122:240285.
SCHOENER, T. W., AND D. A SPILLER. 1987. High population persistence in a system with high turnover.
Nature, 330:474-477.
SCOTT, D. E., AND R. D. DUESER. 1992. Habitat use
by insular populations of Mus and Peromyscus; what
is the role of competition? The Journal of Animal
Ecology, 61:329-338.
SHAFER, C. L. 1990. Nature reserves: island theory
and conservation practice. Smithsonian Institution
Press, Washington, D.C., 189 pp.
SIMBERLOFF, D., AND L. G. ABELE. 1982. Refuge de-
1039
sign and island biogeographic theory: effects of
fragmentation. The American Naturalist, 120:41-50.
SOKOL, R. R., AND F. J. ROHLF. 1981. Biometry: the
principles and practice of statistics in biological research. W. H. Freeman Co. New York, 859 pp.
SONGER, M. A 1996. Habitat selection and niche dynamics of two species of Peromyscus in Olympic
National Forest. M.S. thesis. University of Oklahoma, Norman, 64 pp.
SULLIVAN, T. P. 1979. Demography of populations of
desert mice in coastal forest and clear-cut (logged)
habitats. Canadian Journal Zoology, 57:1636-1648.
SYSTAT, INC. 1992. SYSTAT for windows: statistics,
version 5 edition. SYSTAT, Inc., Evanston, Illinois,
750 pp.
TEVIS, L., JR. 1956. Responses of small mammal populations to logging of Douglas-fir. Journal of Mammalogy, 57:189-196.
VAN DROP, D., AND P. F. M. OPDAM. 1987. Effects of
patch size, isolation, and regional abundance on forest bird communities. Landscape Ecology, 1:59-73.
VERBOOM, B., AND R. VAN APELDOORN. 1990. Effects
of habitat fragmentation on the red squirrel, Sciurus
vulgaris. Landscape Ecology, 4: 171-176.
WECKER, S. C. 1963. The role of early experience in
habitat selection by the prairie deerrnouse, Peromyscus maniculatus bairdi. Ecological Monographs, 33:
307-325.
- - - . 1964. Habitat selection. Scientific American,
211:109-116.
WILCOX, B. A, AND D. D. MURPHY. 1985. Conservation strategy: the effects of fragmentation on extinction. The American Naturalist, 125:879-887.