Landscape Ecology **: *** ***, ****.
**** ****** ******** **********. ******* in the Netherlands.
Comparing the landscape level perceptual abilities of forest sciurids in
fragmented agricultural landscapes
Patrick A. Zollner
Department of Life Sciences, Indiana State University, Terre Haute, IN 47809, USA; Present address: USDA
Forest Service, North Central Research Station, 5985 Highway K, Rhinelander, WI 54501-9128, U.S.A. (e-mail:
abo95v@r.postjobfree.com)
Received 26 January 1999; Revised 22 September 1999; Accepted 14 October 1999
Key words: chipmunks (Tamias striatus), connectivity, dispersal, fox squirrel (Sciurus niger), gray squirrel (Sciurus
carolinensis), habitat isolation, inter-patch movements, perceptual range
Abstract
Perceptual range is the maximum distance from which an animal can perceive the presence of remote landscape
elements such as patches of habitat. Such perceptual abilities are of interest because they in uence the probability
that an animal will successfully disperse to a new patch in a landscape. Furthermore, understanding how perceptual
range differs between species may help to explain differential species sensitivity to patch isolation. The objective
of this research was to assess the perceptual range of eastern chipmunks (Tamias striatus), gray squirrels (Sciurus
carolinensis), and fox squirrels (Sciurus niger) in fragmented agricultural landscapes. Animals were captured in
remote woodlots and translocated to unfamiliar agricultural elds. There they were released at different distances
from a woodlot and their movements towards or away from the woodlot were used to assess their ability to perceive
forested habitat. Observed perceptual ranges of approximately 120 m for chipmunks, 300 m for gray squirrels, and
400 m for fox squirrels, suggest that differences in landscape-level perceptual abilities may in uence the occurrence
of these species in isolated habitat patches.
Introduction landscape by different species (With, 1994; Crist and
Wiens 1995; Diffendorffer et al. 1995; With and Crist
Interspeci c differences in patterns of landscape use 1995, 1996; Wiens et al. 1997; Firle et al. 1998; Had-
often can be attributed to behavioral phenomena (Ims dad 1999). Such differences can alter the dynamics
1995). For example, habitat specialists may be reluc- and structure of populations (Crist and Wiens 1995),
tant to traverse large areas of matrix habitat (Laurance often increasing the susceptibility of isolated popula-
1990; Rail et al. 1997; Heinen et al. 1998), and tions of poor dispersers to local extinction (Petterson
such behavior can affect the persistence of a frag- 1985; Andren 1994). Thus, by re ning our knowl-
mented population (Andren 1994; Laurance 1995; edge of what animals know about their surroundings
With and Crist 1995). Mortality risks during disper- and how they make decisions as they move through
sal, as well as life history traits such as vagility and landscapes (Crist and Wiens 1995; Roitberg and Man-
body size, may also in uence the distribution of an- gel 1997; Pither and Taylor 1998; Turchin 1998), we
imals in fragmented landscapes (Taylor et al. 1993; should increase our understanding of how the spatial
Lidicker and Koenig 1996; Zollner and Lima 1999a). con guration of habitat affects different species (Ims
Indeed, several experiments have documented dif- 1995; Gustafson and Gardner 1996; Zollner and Lima
ferential use of and movement through a common 1999a; Haddad 1999).
The ability of animals to perceive habitat at a
The U.S. Government s right to retain a non-exclusive,
distance is a behavioral mechanism that may be an im-
royalty-free licence in and to any copyright s acknowledged.
portant component of dispersal success in fragmented
524
landscapes (Lima and Zollner 1996; Zollner and Lima success in fragmented landscapes, as well as provide
1997). An animal s perceptual range will determine information on a mechanism that may contribute to
the ease with which it can locate habitat patches and observed difference in patterns of patch occupancy.
hence the time spent searching in a hostile matrix for Thus, the objective of this work was to examine
such habitat (Zollner and Lima 1999a). Consequently, the perceptual ranges of these three species, eastern
a species sensitivity to habitat fragmentation may be chipmunks, eastern gray squirrels, and fox squirrels
to a great extent a function of its perceptual range. Un- in fragmented agricultural landscapes of east central
fortunately, empirical information on the perceptual Illinois and west-central Indiana.
abilities of vertebrates is rare and based on a few sin-
gle species studies performed in different landscapes
(Yoemans 1995; Zollner and Lima 1997; Andreassen Methods
et al. 1998; Gillis and Nams 1998; Zollner and Lima
1999c). Thus, a study comparing the perceptual abil- General methods
ities of several species within a common landscape
The ability of chipmunks and squirrels to orient to-
and relating these abilities to each species occurance
wards forested habitat from a distance was used as a
in isolated habitat patches should clarify the in uence
behavioral assay of their ability to perceive forested
of perceptual range on dispersal success.
habitat at a distance. These abilities were assessed by
Forest-dwelling-sciurids have received consider-
capturing chipmunks and squirrels at distant woodlots
able attention in studies of habitat fragmentation, and
and moving them to an unfamiliar, bare, fallow eld,
they appear to be sensitive to the effects of patch
which was devoid of fence rows. At these novel elds
isolation (Henderson et al. 1985; Verboom and Van
animals were released at several distances (determined
Apeldoorn 1990; Fitzgibbon 1993; Van Appeldoorn
by pilot work; Zollner unpublished data) from the edge
et al. 1994; Wauters et al. 1994; Sheperd and Swi-
of a mature woodlot. The orientation of the movement
hart 1995; Rushton et al. 1997; Heinen et al. 1998).
path at each release distance was measured to assess
However, species may differ in their sensitivity to the
perceptual range of each species (Zollner and Lima
effects of isolation, and these differences are likely to
1997; Zollner and Lima 1999c). Critical to this work
re ect a species mobility in matrix habitat or some
is the assumption that the perception of a woodlot is
correlate such as body size (Swihart and Nupp 1998).
equivalent to movement towards the woodlot. Few en-
Among the forest-dwelling sciurids of eastern North
vironments would appear as hostile to a chipmunk or
America, it is clear that gray squirrels (Sciurus car-
squirrel as a barren eld, and survival in such an en-
olinensis) are more sensitive to patch isolation than
vironment requires locating forested habitat as soon
fox squirrels (Sciurus niger; Table 1). A third for-
as possible. Thus, it is reasonable to assume that if
est dwelling sciurid common to this area the eastern
these animals perceived forested habitat they should
chipmunk (Tamias striatus) is also negatively effected
have attempted to reach it.
by isolation of habitat (Table 1). However, the sensi-
I captured chipmunks with Sherman live traps and
tivity of chipmunks to patch isolation relative to the
squirrels with Tomahawk live traps in mature woodlots
other two species is unclear (Table 1). Working in
5 29 km from the release site. This 5 km mini-
east-central Illinois Rosenblatt et al. (1999) only found
mum distance and movement barriers (roads, streams,
chipmunks in the three largest and best connected
etc.) between capture and release sites minimized the
patches of ten which they surveyed, while gray and
chance that animals had prior experience at the release
fox squirrels were present in six and nine patches re-
site. I used adult males and females, and all pregnant
spectively. However, Nupp and Swihart (1998) found
or lactating females were excluded from the experi-
chipmunks in all four patches that they surveyed in
ments. Traps were checked twice each day, once in
northwestern Indiana, including one patch that was
the morning and once in the evening. Prior to their re-
870 m from the nearest woodlot. More extensive sur-
lease, animals were provisioned with seeds and housed
veys in northwestern Indiana indicate that chipmunks
overnight in their traps in a small, unheated shed.
are less sensitive to patch isolation than either gray
Releases were accomplished using a standard re-
squirrels or fox squirrels (Nupp and Swihart in re-
lease mechanism . The mechanism was constructed of
view). An assessment of the perceptual abilities of
a 40 cm long piece of PVC pipe that was either 6.5 cm
each of these three species should help clarify the role
(chipmunks) or 13 cm (squirrels) in diameter. A metal
that perceptual range plays in in uencing dispersal
525
Table 1. Reported sensitivities of three woodland sciurids to patch isolation.
Effect of patch isolation Reference
Fox squirrels
Fox squirrels were insensitive to levels of fragmentation in this study. Nupp and Swihart (in review)
Fox squirrels were present in 31 of 37 patches surveyed and are apparently
capable of crossing agricultural matrix.
Wide distribution of fox squirrels (present in 9 out of 10 woodlots) indicates Rosenblatt et al. (1999)
high mobility or prolonged persistence in isolated patches.
Fox squirrels are ubiquitous in non-urban woodlots of row-crop dominated Rosenblatt (in review)
east central Illinois; attributed to their superior dispersal ability.
Relative to gray squirrels, fox squirrels move greater distances, and visit
more patches when translocated.
None of 49 radio-collared fox squirrels moved between isolated patches Sheperd and Swihart (1995)
across agricultural matrix.
Some fox squirrels did move 200 500 m away from woodlots through
fencerows and one patch was colonized by a non-collared squirrel from
at least 800 m away.
Spatially explicit simulation found interpatch movements by fox squirrels Swihart and Nupp (1998)
were not constrained by patch isolation.
Empirical surveys found fox squirrels in all 18 patches examined even the
most isolated ones.
Gray squirrels
Gray squirrels are less likely to be found in patches > 500 m from other Fitzgibbon (1993)
woodlots and not connected by hedgerows.
Sites that historically contained gray squirrels lost them as landscapes were Nixon et al. (1978)
reduced to less than 20% forested coverage.
Among 54 habitat variables analyzed, the best predictor of gray squirrel
presence was the amount of forested habitat within 23.31 km2 of a site.
Gray squirrels were restricted to continuous forests and large sites (only Nupp and Swihart (in review)
present in 7 of 37 patches surveyed).
The best- t logistic regression model of gray squirrel presence was posi-
tively associated with patch area and negatively associated with isolation.
Gray squirrels were present in only 6 of 10 patches surveyed and were Rosenblatt et al. (1999)
absent from isolated rural woodlots.
Gray squirrels are restricted to towns and riparian forests in row-crop dom- Rosenblatt (in review)
inated east central Illinois, which was attributed to their hesitancy to move
across open elds.
Relative to fox squirrels gray squirrels move shorter distances, and visit
fewer patches when translocated.
Successful introduction of gray squirrels to woodlots where they were ab- Rosenblatt (1999)
sent for 20+ years supports the hypothesis that absence was maintained by
isolation.
Spatially explicit simulation found interpatch movements by gray squirrels Swihart and Nupp (1998)
were constrained by patch isolation.
Empirical surveys found gray squirrels in 4 of 18 patches examined but not
in the isolated ones.
526
Table 1. Continued.
Effect of patch isolation Reference
Eastern Chipmunks
Eastern chipmunks reside in fencerows and use them as movement path- Bennett et al. (1994)
ways between patches.
Two chipmunks were observed moving > 500 m between patches pre-
sumably through fencerows and numerous other long movements were
documented within fencerows.
One individual regularly crossed 70 m of open eld. Forsyth and Smith (1973)
Spatially explicit simulation found that patch connectivity was the most Henein et al. (1998)
important factor determining persistence of eastern chipmunk populations
in fragmented agricultural landscapes.
Chipmunks were observed to move as far as 1560 m between woodlots and
this travel was presumed to be largely through fencerows.
Henderson et al. (1985)
Chipmunks probably never crossed areas of matrix larger than 20 60 m
although one individual may have moved as far as 460 m across elds.
Eastern chipmunks were present in 4 out of 4 patches surveyed including Nupp and Swihart 1998
one isolated patch.
Chipmunks appear to be negatively in uenced by forest fragmentation as
survival rates for individuals living in patches were signi cantly lower than
for those in continuous forest.
Eastern chipmunks were present in 32 out of 37 patches and no signi cant Nupp and Swihart (in review)
model of presence/absence could be developed based on landscape metrics.
Chipmunks never crossed roads with clearances > 30 m and roadways > Oxley et al. (1974)
90 m act as barriers.
Eastern chipmunks were only detected in large forest tracts that were well Rosenblatt et al. (1999)
connected to other forested areas (3 out of 10 patches surveyed).
Fencerows provide corridors for movements by chipmunks and reduce Wegner and Merriam (1979)
isolation of woodlots.
Chipmunks were never captured in agricultural elds.
spike (30 cm long) was placed through two holes in Baumgartner 1940). Next, a unique tag was attached to
the pipe at one end, and driven into the ground to se- an animal s ear. A tracking spool (1.7 g, 180 m, denier
cure the mechanism. This secured end of the pipe was two-ply nylon No. 2 quilting bobbin; Barbour Threads
covered with an opaque cap that prevented the animal Inc., Anniston, AL, USA) was then glued to the an-
from exiting the release mechanism. The other end of imal s back, and the loose end of this spool was tied
the pipe was left open until the animal was placed in- to the release mechanism (Boonstra and Crane 1986;
side it, at which time the pipe was sealed with a plastic Key and Woods, 1996). I followed the spool-and-line
cap. technique described by Key and Woods (1996) with
At the time of release, I transported chipmunks and a few modi cations: no animals used in this study
squirrels to the study in opaque boxes that prevented were anaesthetized, and rather than wrapping tracking
them from visually assessing their surroundings. The spools in adhesive tape, I placed them in small dark
actual release locations were placed in straight lines brown unin ated rubber ballons. After the spool was
parallel to the edge of the woods at different distances securely attached (20 30 s) each animal was lowered
from the woods for each species. Along these parallel into the PVC pipe facing the back, and a plastic cap
lines the release sites were spaced such that no animals was placed over the open end of the pipe.
were released within 70 m of each other on any given The release itself was done remotely from a dis-
day. At the release site, I removed animals from their tance of 60 m, so that my presence did not in uence
traps and restrained them in either a heavy black cotton animal movements. This remote release was accom-
bag (chipmunks) or a wire handling cone (squirrels; plished by pulling on a string, thereby removing the
527
plastic cap and opening one end of the release mecha- tions with Mardia-Watson-Wheeler (MWW) pairwise
nism. After removing the caps, I immediately left the tests (Batschelet 1981). No test demonstrated any sex-
study site and did not return until the following day. related differences in orientation towards the woods
(MWW test, 2 0.1), hence data
While releasing animals and leaving the study site, I
was at the same distance from the woods as the release from both sexes were combined before performing
mechanism, so my presence should not have biased statistical analyses.
the animals to move towards or away from the woods.
Additionally, remote observations through a spotting Species-speci c methods
scope during pilot work indicated that chipmunks and
Eastern fox squirrels
squirrels remained inside the release mechanism for
During April May, 1997, I captured 47 fox squirrels
30 45 min after it was opened (P. A. Zollner pers.
in a mature oak hickory woodlot in east-central Clark
obs.).
County, Illinois. These animals were released at a site
The tracking spool left a trail of thread record-
26 km away in north-central Edgar County, Illinois.
ing an animal s movements after it exited the release
This release site was a large 132-ha eld bordered to
mechanism. The day following release, stick ags
the south by a second growth oak hickory forest, to
were placed in the ground along each thread trail at
the north by a road, and to the north, east and west
approximately 3 m intervals and at all points where the
by additional, large agricultural elds. Other than the
animal turned sharply. Each trail was followed until
forest along the southern edge of the release site, the
the thread ended or it reached the woods (chipmunks
nearest trees were 1.8 km away. All fox squirrels were
only). Trails that did not reach the woods were on aver-
age ( SE) 147.1 m ( 3.6) long. Animals occasionally released in this eld between 10:00 and 14:00. Fox
squirrels were released 300 m (15 squirrels), 500 m
broke the thread before travelling the full 180 m al-
(16 squirrels), and 800 m (16 squirrels) from the
though all trails included in these results exceeded
forested southern edge of the eld. During April May,
100 m in length. After tracking, I used a sighting com-
1998, while studying gray squirrels (see below), I
pass (Brunton Sight Master 80NL) and eld tape to
captured an additional 24 fox squirrels, which were
measure the bearing and distance from the point of
released 300 m (8 squirrels), 400 m (8 squirrels) and
release to (i) each ag in the trail and (ii) the nearest
500 m (8 squirrels) from the forested southern edge
point along the woodlot edge.
of the eld. These additional releases of fox squirrels
I assessed perceptual abilities by determining
allowed for the de nition of fox squirrel perceptual
whether the animals locations after traveling a pre-
range at the same scale used for gray squirrels (see
scribed distance (chipmunks 50 m; squirrels 100 m)
below).
were oriented towards the woods. These prescribed
distances were shorter than the minimum distance to
Gray squirrels
the woods (chipmunks 60 m, squirrels 300 m) but long
During April May, 1998, I captured 28 gray squir-
enough to allow an animal to orient after dashing out
rels in a mature oak hickory woodlot in south-central
of the release mechanism. The use of this minimum
Clark County, Illinois. These squirrels were all of
distance ensured that animals did not reach the woods
the gray color morph, although there is no reason to
as a result of random wandering (Goodwin et al. 1999;
expect melanistic animals would have behaved differ-
Zollner and Lima 1999b). The angle to each animals
ently (Gustafson and Van Druff 1990). These animals
location after travelling the prescribed distance was
were released at the site used for the fox squirrel re-
calculated from the recorded movement pathways us-
leases (see above), which was approximately 29 km
ing trigonometry. V-tests were used to assess whether
away in Edgar County, Illinois. All gray squirrels were
these angels were signi cantly oriented towards the
released between 10:00 and 14:00. Squirrels were re-
woods for each species. The V-test is a modi cation
leased at 300 m (8 individuals), 400 m (10 individuals)
of a Rayleigh test which examines whether observed
and 500 m (10 individuals) from the forested southern
angles are statistically clustered around a hypothesized
edge of the eld. Only 8 gray squirrels were released
angle (Batschelet 1981). I also used V-tests to deter-
at 300 m because this species was dif cult to capture
mine whether the locations of the last points to which
locally, and it was apparent that they were orienting
animals were tracked were oriented towards the home
towards the woods from 300 m.
woodlot (site of capture). Furthermore, I examined the
possibility of sex-related effects on angular orienta-
528
Eastern chipmunks
The chipmunk releases took place at two sites over the
course of two different seasons. This was necessitated
by dif culty in capturing enough animals in a single
season and changes in the crop rotations between the
years at the release sites. Tests indicated no difference
between releases at these two sites, hence data were
combined for all analyses (see below). Overall, 21
chipmunks were released 60 m from the woods, while Figure 1. Angular orientations of fox squirrels released during
1997. Fox squirrels were released 300, 500, and 800 m from the
20 chipmunks were released at both 120 and 180 m
woods; angular orientations were assessed after 100 m of travel.
from the woods. The solid square in the center of each panel represents the site where
During May 1996, I captured 22 chipmunks in animals were released and the trees show the direction to the woods.
The angular orientation of each fox squirrel is depicted as an open
a mature oak hickory woodlot in northwestern Vigo
circle on the unit circle. Vectors indicate average angle and degree
County, Indiana. These animals were released at a
of orientation and are displayed only for cases with statistically
site 6.4 km away in west-central Vigo County. This signi cant orientation towards the woods.
release site was a 10-ha eld bordered to the east by
a mature oak hickory forest, to the north and south
by additional agricultural elds and to the west by a
road beyond which there were more agricultural elds.
Other than the eastern edge of the release site, the
nearest trees were 250 m away. As with the squirrels,
all chipmunks were released between 10:00 and 14:00.
Fourteen chipmunks were released 60 m and 8 chip-
munks were released 120 m from the forested eastern
edge of the eld. Figure 2. Angular orientations of fox squirrels released during
1998. Fox squirrels were released 300, 400, and 500 m from the
During October November 1997, I captured 39
woods; and angular orientations were assessed after 100 m of travel.
chipmunks in a mature oak hickory woodlot in eastern All symbols are as in Figure 1.
Clark County, Illinois. These animals were released
at a site approximately 5 km away in central Clark
County, Illinois. This release site was 7 km northwest P 0.1; 800 m releases, V-test: u =
a mature oak hickory forest, to the east and west by
0.03, P > 0.1).
additional elds and to the south by a road beyond
which more elds were located. Other than the north- Releases conducted during 1998 suggest that the
ern edge of the release site, the nearest other trees were perceptual range of fox squirrels was between 400 and
360 m away. All chipmunks were released in the eld 500 m (Figure 2). The angular orientation of fox squir-
between 10:00 and 14:00. Chipmunks were released rels released 300 m from the woods was signi cantly
oriented towards the woods (V-test: u = 2.23, P
Releases conducted during 1997 demonstrated that the
0.1).
perceptual range of fox squirrels was between 300
Fox squirrels were released 300 and 500 m from
and 500 m (Figure 1). The angular orientation of fox
the woods during both 1997 and 1998. Mardia
squirrels released 300 m from the woods was signi -
cantly oriented towards the woods (V-test: u = 4.93, Watson Wheeler pairwise comparisons of fox squirrel
529
Figure 4. Angular orientations of chipmunks released during 1996
Figure 3. Angular orientations of gray squirrels released during
and 1997 (see text). Chipmunks were released 60, 120, and 180 m
1998. Gray squirrels were released 300, 400, and 500 m from the
from the woods; and angular orientations were assessed after 50 m
woods; and angular orientations were assessed after 100 m of travel.
of travel. All symbols are as in Figure 1.
All symbols are as in Figure 1.
P > 0.1) or 120 m (MWW test, 2 = 5.06, d.f. =
data found no year-speci c difference in the angular
2, P > 0.1) from the forest edge. Thus, data from
orientations of fox squirrels released at 300 m (MWW
test, 2 = 3.01, d.f. = 2, P > 0.1) or 500 m (MWW each site were pooled for chipmunks released at 60
test, 2 = 3.71, d.f. = 2, P > 0.1) from the forest and 120 m during all subsequent analyses; all 180 m
releases were done at the second site.
edge. After combining the data from the two years,
The perceptual range of eastern chipmunks was
fox squirrels released 300 m from the woods were
between 120 m and 180 m (Figure 4). The locations
still signi cantly oriented towards the woods (V-test:
u = 5.29, P 0.1). Finally, fox squirrels
0.0001; 120 releases, V-test: u = 2.64, P
they were captured (500 m releases, V-tests, u = 0.51,
P > 0.1; 800 m releases, V-tests, u = 0.15, P > 0.1). 0.1); these chipmunks also failed to show a signi -
cant orientation towards the site at which they were
captured (V-test, u = 0.81, P > 0.1).
Gray squirrels
The perceptual range of gray squirrels was between
300 and 400 m (Figure 3). Gray squirrels released Discussion
300 m from the woods were signi cantly oriented to-
wards the woods (V-test: u = 2.85, P 0.1; was less than that of fox squirrels. This ordering of
500 m releases, V-test: u = 0.46, P > 0.1). Fi- perceptual ranges is inversely related to the reported
nally, gray squirrels not signi cantly oriented towards sensitivities of these two species to habitat isolation
the woods also failed to show a signi cant orientation (Table 1), which suggests that differences in percep-
towards the site at which they were captured (400 m tual abilities contribute to the occurrences of these
releases, V-test, u = 0.06, P > 0.1; 500 m releases, species in fragmented agricultural landscapes. Prox-
V-test, u = 0.32, P > 0.1). imately, the differences in the perceptual abilities of
these species may also be related to differences in
Eastern chipmunks body size (Gillis and Nams 1998). The ultimate ori-
gin of these differences in perceptual abilities may be
Recall that circumstances dictated the use of two sites
related to historical differences in habitat occupied by
over two eld seasons for the chipmunk releases (see
these species. Historically, fox squirrels were common
above). Mardia Watson Wheeler pairwise compar-
at the interface of the eastern deciduous forests and
isons of chipmunk locations found no site-speci c
the prairie, while gray squirrels were found in interior
differences in the angular orientations of chipmunks
forest habitat (Allen 1943; Smith and Follmer 1972;
released at 60 m (MWW test, 2 = 1.49, d.f. = 2,
530
Swihart and Nupp 1998). Fox squirrels presumably be good approximations of the maximum perceptual
have a longer evolutionary history with large areas of range for each of these species.
open habitat, while gray squirrels have been exposed Perceptual range in uences dispersal success most
to open habitat only since the recent clearing of forests when species face an intermediate probability of suc-
for agriculture. This is consistent with the observation cessful dispersal. Fahrig (1988) demonstrated this pat-
that fox squirrels forage as patch transients while tern for simulated animals facing different proportions
gray squirrels forage as patch residents (Steele and of suitable habitat in a landscape. This same principle
Weigl 1992). is likely to apply to other factors that affect dispersal
The observed perceptual range of chipmunks sug- success. For example, species using highly effective
gests they should be more sensitive to the effect of search strategies will nd patches quickly no matter
fragmentation than either squirrel species, but disper- what their perceptual range (Zollner and Lima 1999a).
sal success is likely to be in uenced by a variety of Alternatively, species facing very high mortality risks
factors (see below). Chipmunks clearly are sensitive may never successfully disperse because they will die
to the effects of habitat isolation (Wegner and Merriam before reaching new habitat even when endowed with
1979; Henderson et al. 1985; Heinen et al. 1998; Nupp vast perceptual abilities (Swihart and Nupp 1998).
and Swihart 1998; Rosenblatt et al. 1999), however Such differences may explain why some simulations
their sensitivity relative to that of the squirrels remains have found dispersal success to be sensitive to percep-
unresolved (Rosenblatt et al. 1999; Nupp and Swihart tual range (Fahrig 1988; Pulliam et al. 1992; Turner
in review). This ambiguity may in part be attributable et al. 1993) while others have indicated that perceptual
to differences between study sites such as the pro- range is inconsequential (Liu et al. 1995; Swihart and
portion of the landscape containing forested habitat, Nupp 1988).
the quality of the habitat for chipmunks, or the range A key assumption in this experiment was that if
of isolation values investigated. Additional factors the animals could perceive forested habitat they would
such as the occurrence of fence rows in these land- move towards it. This is a reasonable assumption since
scapes might confound comparisons because fence all of these species are woodland resident animals
rows containing resident populations of chipmunks (Snyder 1982; Koprowski 1994a, 1994b) that face an
(Bennett et al. 1994) may not be included in calcu- increased risk of predation in open habitat (Bowers
lations of patch isolation. Finally, differences might and Ellis 1993; Bowers et al. 1993; Lima 1998). This
also be attributable to geographic variation in either increase in risk is demonstrated by the observation
historical habitat or land use patterns. Note that the that all of these species will forage in open elds, but
perceptual ranges reported here are consistent with the only when they are close to forested habitat or other
sensitivities to patch isolation observed by Rosenblatt cover (Lima and Valone 1986; Sheperd and Swihart
et al. (1999) who s work occurred in close geographic 1995; McAdam and Kramer 1998). Further support
proximity to these study sites. for this assumption was provided by the observation
The perceptual range which each of these species that several chipmunks released at 180 m actually dug
effectively experience during dispersal may not always shallow tunnels, presumably to reduce their risk of
be as great as the values reported here. All three of predation while lost. In contrast, chipmunks released
these species are known to disperse during times of at 60 and 120 m moved directly towards the woods and
the year when crops are present in the elds, and never dug such tunnels. This protocol also assumed
visually obstructive crops may reduce the ability to that the release sites were unfamiliar to the subjects.
perceive distant habitat (Zollner and Lima 1997). Ani- This assumption is supported by several lines of evi-
mals might minimize the perceptual constraints which dence: (i) all translocation distances exceeded reported
crops impose by climbing trees prior to dispersal, but homing abilities of all species (Hungerford and Wilder
any such gains would only apply to the immediate 1941; Seidel 1961; Bendel and Therres 1994); (ii) no
vicinity of their point of origin and not their entire homeward orientation was detected for any of these
search path (Zollner and Lima 1999c). Furthermore, releases; and (iii) no marked animals were recaptured
dispersal is typically done by juveniles which may at trapping sites. Finally, pilot observations indicate
have more limited perceptual ranges than the adults that animals remain in the release mechanisms for at
used in these experiments (Zollner unpublished data). least 0.5 hours after the cap is removed. Thus distress
Nonetheless, the estimates presented here are likely to caused by handling the animals (Goodwin et al. 1999)
should have subsided prior to the recorded move-
531
ments. Furthermore, such stress should only increase Acknowledgements
the desire of these woodland resident animals to move
towards the forest if they perceive it (Zollner and Lima J. Deloughry, M. Evrard, T. Fitzpatrick, N. Hogan, J.
1999b). Irwin, S. Lima, D. Smith, F. Stepp, M. Young and
Perceptual range is certainly not the only factor Indiana State University kindly provided permission
affecting dispersal success. Landscape characteristics to use their properties in my experiments. K. Crane
such as the presence of fence rows or the composi- and E. Sprague provide assistance with the eld work.
tion of the matrix may be important for dispersing E. Gustafson, S. Lima, D. Rosenblatt, R. Swihart, and
chipmunks and squirrels (Wegner and Merriam 1979; V. Quinn commented on early versions of this man-
Fitzgibbon 1993; Bennett et al. 1994; Sheperd and uscript. Funding for this work was provided by the
Swihart 1995). Mortality risks are also known to in- Theodore Roosevelt Fund of the American Museum
crease during dispersal (Larsen and Boutin 1994; Van of Natural History, Indiana State University, and NSF
Vuren 1998), and chipmunks and squirrels may face grant IBN-9221925.
different mortality risks while moving through agri-
cultural elds (Smith and Follmer 1972; Swihart and
Nupp 1998). Such differences in mortality would de- References
termine the best dispersal search strategy (Zollner and
Allen, D.L. 1943. Michigan fox squirrel management. Dept. of
Lima 1999a), and even the willingness of animals to
Cons., Game Div. Publ. 100: 1 404.
cross matrix habitat (Lidicker and Koenig 1996). Ad- Andreassen, H.P., Bjornbom, E., Carlson, R.G., Gunderson, G. and
ditionally, factors such as energetic reserves (Nunes Gunderson, H. 1998. Visual cues as determinants of perceptual
and Holekamp 1996) and social conditions (Nixon range in root voles Microtus oeconmus. Acta Theriologica 43:
371 378.
et al. 1986) are likely to in uence dispersal decisions
Andren, H. 1994. Effects of habitat fragmentation on birds and
and success. Clearly, dispersal success is affected by mammals in landscapes with different proportions of suitable
a wide variety of behavioral attributes and landscape habitat: a review. Oikos 71: 355 363.
components (Wiens et al. 1997), and we have only a Batschelet, E. 1981. Circular Statistics in Biology. Academic Press,
New York, NY, USA.
minimal understanding of the relative contribution of
Baumgartner, L.L. 1940. Trapping, handling, and marking fox
these different elements to dispersal success. squirrels. J. Wildl. Manage. 4: 479 484.
In summary, the perceptual abilities of eastern Bennett, A.F., Henein, K. and Merriam, G. 1994. Corridor use and
the elements of corridor quality: chipmunks and fencerows in a
chipmunks, gray squirrels, and fox squirrels suggest
farmland mosaic. Biol. Cons. 68: 155 165.
that perceptual range may contribute to dispersal suc-
Bendel, P.R. and Therres, G.D. 1994. Movements, site delity and
cess for these species in agriculturally fragmented survival of Delmarva fox squirrels following translocation. Am.
landscapes. Certainly, other behaviors are also likely Midl. Nat. 132: 227 233.
Boonstra, R. and Craine, I.T.M. 1986. Natal nest location and small
to in uence dispersal success and thus warrant fur-
mammal tracking with a spool and line technique. Can. J. Zool.
ther investigation with the comparative approach used
64: 1034 1036.
in this paper. A series of such investigations could Bowers, M.A. and Ellis, A. 1993. Load size variation in the east-
be used to parameterize spatially explicit simulation ern chipmunk, Tamias striatus: the importance of distance from
burrow and canopy cover. Ethology 94: 72 82.
models and perform sensitivity analysis, greatly in-
Bowers, M.A., Jefferson, J.L. and Kuebler, M.G. 1993. Variation in
creasing our understanding of which behavioral char- giving up densities of foraging chipmunks (Tamiais striatus) and
acteristics in uence dispersal success to what extent. squirrels (Sciurus carolinensis). Oikos 66: 229 236.
Thus, the present study is an important step towards Crist, T.O. and Wiens, J.A. 1995. Individual movements and es-
timation of population size in darkling beetles (Coleoptera:
the development of a more general understanding of
Tenebrionidae). J. Anim. Ecol. 64: 733 746.
how behavioral processes in uence patterns of distri- Diffendorfer, J.E., Gaines, M.S. and Holt, R.D. 1995. Habitat frag-
bution across landscapes. Ultimately, this understand- mentation and movements of three small mammals (Sigmodon,
ing is of interest because of the role that dispersal Mocrotus, and Peromyscus). Ecology. 76: 827 839.
Fahrig, L. 1988. A general model of populations in patchy habitats.
success plays in maintaining viable populations in
Appl. Math. C. 27: 53 66.
fragmented landscapes (Andren 1994; With and Crist Firle, S., Bommarco, R., Ekbom, R. and Natiello, M. 1998. The
1995). in uence of movement and resting behavior on the range of three
carabid beetles. Ecology 79: 2113 2122.
Fitzgibbon, C.D. 1993. The distribution of grey squirrel dreys
in farm woodland: the in uence of wood area, isolation and
management. J. Appl. Ecol. 30: 736 742.
532
Nixon, C.M., Havera, S.P. and Greenberg, R.E. 1978. Distribution
Forsyth, D.J., and Smith, D.A. 1973. Temporal variability in home
and abundance of the gray squirrel in Illinois. Illinois Natural
ranges of eastern chipmunks (Tamias striatus) in a South-eastern
History Survey Biological Notes, Vol. 105, 55p.
Ontario woodlot. Am. Midl. Nat. 90: 107 117.
Nixon, C.M., Hansen, L.P. and Havera, S.P. 1986. Demographic
Goodwin, B.J., Bender, D.J., Contreras, T.A., Fahrig L. and Weg-
characteristics of an unexploited population of fox squirrels
ner, J.F. 1999. Testing for habitat detection distances using
(Sciurus niger). Can. J. Zool. 64: 512 521.
orientation data. Oikos, 84: 160 163.
Nunes, S. and Holekamp, K.E. 1996. Mass and fat in uence timing
Gillis, E.A. and Nams, V.O. 1998. How red backed voles nd habitat
of natal dispersal in Belding s ground squirrels. J. Mammal. 77:
patches. Can. J. Zool. 76: 791 794.
807 817.
Gustafson, E.J. and Gardner, R.H. 1996. The effect of landscape
Nupp. T.E. and Swihart, R.K. 1998. Effects of forest fragmenta-
heterogeneity on the probability of patch colonization. Ecology
tion on population dynamics of white-footed mice and eastern
77: 94 107.
chipmunks. J. Mammal. 79: 1234 1243.
Gustafson, E.J. and Van Druff, L.W. 1990. Behavior of black and
Nupp. T.E. and Swihart, R.K. in review. Landscape-level correlates
gray morphs of Sciuris carolinensis in an urban environment.
of small mammal assemblages in forest fragments of farmlands.
Am. Midl. Nat. 123: 186 192.
J. Mammal.
Haddad, N.M. 1999. Corridor use predicted from behaviors at
Pettersson, B. 1985. Extinction of an isolated population of the mid-
habitat boundaries. Amer. Nat. 153: 215 227.
dle spotted woodpecker Dendrocopos medius (L.) in Sweden and
Henderson, M.T., Merriam G. and Wegner, J. 1985. Patchy environ-
its relation to general theories on extinction. Biol. Conserv. 32:
ments and species survival: chipmunks in an agricultural mosaic.
335 353.
Biol. Conserv. 31: 95 105.
Pither, J. and Taylor, P.D. 1998. An experimental assessment of
Henein, K., Wegner, J. and Merriam, G. 1998. Population effects
landscape connectivity. Oikos 83: 166 174.
of landscape model manipulation on two behaviourally different
Pulliam, H.R., Dunning, J.B. Jr. and Liu, J. 1992. Population dy-
woodland small mammals. Oikos 81: 168 186.
namics in complex landscapes: a case study. Ecol. Appl. 2:
Hungerford, K.E. and Wilder, N.G. 1941. Observations on the hom-
165 177.
ing behavior of the gray squirrel (Sciurus carolinensis). J. Wildl.
Oxley, D.J., Fenton, M.B. and Carmody, G.R. 1974. The effects
Manage. 5: 458 460.
of roads on populations of small mammals. J. Appl. Ecol. 18:
Ims, R.A. 1995. Movement patterns related to spatial structures. In
51 59.
Mosaic landscapes and cological rocesses. pp. 85 109. Edited
Rail, J-F., M. Darveau, A. Desrochers, and J. Huot. 1997. Territorial
by L. Hansson, Fahrig, L. and Merriam, G. Chapman and Hall,
responses of boreal forest birds to habitat gaps. Condor 99: 976
London.
980.
Key, G.E. and Woods, R.D. 1996. Spool-and-line studies on the be-
Roitberg, B.D. and Mangel, M. 1997. Individuals on the landscape:
havioural ecology of rats (Rattus spp.) in the Galapagos Islands.
behavior can mitigate landscape differences among habitats.
Can. J. Zool. 74: 733 737.
Oikos 80: 234 240.
Koprowski, J.L. 1994a. Sciurus niger. Mammalian Species Account
Rosenblatt, D.L. in review. An experimental examination of tree
479: 1 9.
squirrel vagility in an agricultural landscape. Landscape Ecol.
Koprowski, J.L. 1994b. Sciurus carolinensis. Mammalian Species
Rosenblatt, D.L. 1999. The effect of habitat fragmentation on forest
Account 480: 1 9.
mammals: An experiemntal analysis of tree squirrel distributions
Larsen, K.W. and Boutin, S. 1994. Movements, survival, and
in the agricultural landscape of east central Illinois. Unpublished
settlement of red squirrel (Tamiasciurus hudsonicus) offspring.
Ph.D. Dissertation, University of Illinois, Urbana, Illinois.
Ecology 75: 214 223.
Rosenblatt, D.L., Heske, E.J., Nelson, S.L., Barber, D.M., Miller,
Laurance, W.F. 1995. Extinction and survival of rainforest mammals
M.A. and MacAllister, B. 1999. Forest fragments in east-central
in a fragmented tropical landscape. In Landscape approaches
Illinois: island or habitat patches for mammals? Am. Midl. Nat.
in mammalian ecology and conser
scapes*