Urban Habitat Fragmentation and Genetic

Population Structure of Bobcats in Coastal Southern

California
Author(s): E.W. Ruell, S.P.D. Riley, M.R. Douglas, M.F. Antolin,
J.R. Pollinger, J.A. Tracey, L.M. Lyren, E.E. Boydston, R.N. Fisher,
and K.R. Crooks
Source: The American Midland Naturalist, 168(2):265-280. 2012.
Published By: University of Notre Dame
URL: http://www.bioone.org/doi/full/10.1674/0003-0031-168.2.265

BioOne (www.bioone.org) is a nonprofit, online aggregation of core

research in the biological, ecological, and environmental sciences. BioOne
provides a sustainable online platform for over 170 journals and books
published by nonprofit societies, associations, museums, institutions, and
presses.
Your use of this PDF, the BioOne Web site, and all posted and associated
content indicates your acceptance of BioOne’s Terms of Use, available at
www.bioone.org/page/terms_of_use.
Usage of BioOne content is strictly limited to personal, educational, and
non-commercial use. Commercial inquiries or rights and permissions
requests should be directed to the individual publisher as copyright holder.
BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors,
nonprofit publishers, academic institutions, research libraries, and research funders in the common goal of
maximizing access to critical research.
Am. Midl. Nat. (2012) 168:265–280

Urban Habitat Fragmentation and Genetic Population

Structure of Bobcats in Coastal Southern California
E. W. RUELL1
Department of Fish, Wildlife, and Conservation Biology, Colorado State University, Fort Collins 80523

S. P. D. RILEY
Santa Monica Mountain National Recreation Area, National Park Service, Thousand Oaks, California 91360

M. R. DOUGLAS
Department of Fish, Wildlife, and Conservation Biology, Colorado State University, Fort Collins 80523

M. F. ANTOLIN
Department of Biology, Colorado State University, Fort Collins 80523

J. R. POLLINGER
Department of Ecology and Evolutionary Biology, University of California, Los Angeles 90095

J. A. TRACEY
Department of Fish, Wildlife, and Conservation Biology, Colorado State University, Fort Collins 80523

L. M. LYREN
U.S. Geological Survey, Western Ecological Research Center, Carlsbad, California 92011

E. E. BOYDSTON
U.S. Geological Survey, Western Ecological Research Center, Thousand Oaks, California 91360

R. N. FISHER
U.S. Geological Survey, Western Ecological Research Center, San Diego, California 92101

AND

K. R. CROOKS
Department of Fish, Wildlife, and Conservation Biology, Colorado State University, Fort Collins 80523

ABSTRACT.—Although habitat fragmentation is recognized as a primary threat to biodiversity,

the effects of urban development on genetic population structure vary among species and
landscapes and are not yet well understood. Here we use non-invasive genetic sampling to
compare the effects of fragmentation by major roads and urban development on levels of
dispersal, genetic diversity, and relatedness between paired bobcat populations in replicate
landscapes in coastal southern California. We hypothesized that bobcat populations in sites
surrounded by urbanization would experience reduced functional connectivity relative to less
isolated nearby populations. Our results show that bobcat genetic population structure is
affected by roads and development but not always as predicted by the degree that these
landscape features surround fragments. Instead, we suggest that urban development may affect
functional connectivity between bobcat populations more by limiting the number and genetic
diversity of source populations of migrants than by creating impermeable barriers to dispersal.

1
Corresponding author: Department of Biology, Colorado State University, Fort Collins 80523;
e-mail: eruell@gmail.com

265
266 THE AMERICAN MIDLAND NATURALIST 168(2)

INTRODUCTION
Habitat loss and fragmentation are recognized as primary threats to biodiversity (Wilcove
et al., 1998), and urbanization is a leading agent of habitat destruction and cause of species
endangerment (Czech et al., 2000). Yet, how habitat fragmentation, the subdivision of
natural habitats into isolated and small patches, impacts genetic population structure is not
well understood, because fragmentation effects are heterogeneous across populations,
species, and habitats (Thompson et al., 2001; Jellinek et al., 2004; Culley et al., 2007;
Keyghobadi, 2007; Santos et al., 2008; Delaney et al., 2010). The majority of studies published
between 1996 and 2006 that investigated the effects of fragmentation on genetic population
structure across replicate landscapes found repeated patterns of lower genetic diversity
(,58% of studies) and greater population differentiation (,69%) in fragmented habitats,
as would be predicted for small and isolated populations (Keyghobadi, 2007); the remaining
studies, however, reported reverse relationships or showed that fragmentation impacts
varied among landscapes. Fragmentation effects vary due to a number of factors, such as the
size and age of the fragment, the permeability of the intervening matrix, the spatial
proximity of fragments to other populations, and the genetic variability of any potential
source populations (Jellinek et al., 2004; Culley et al., 2007; Santos et al., 2008; Delaney et al.,
2010). As a result, the genetic effects of fragmentation across species, or even within a
species across different landscapes, may prove difficult to predict, and examination of such
impacts should include replication to evaluate if observed effects are consistent (Culley et al.,
2007; Keyghobadi, 2007).
Fragmentation can reduce functional connectivity, or the flow of organisms between
populations (Tischendorf and Fahrig, 2000). Preserving connectivity is an important means
of preventing detrimental genetic effects that can occur in small isolated populations due to
inbreeding and genetic drift, processes that can increase extinction risk (Keller and Waller,
2002; Frankham, 2005; Wright et al., 2008). Genetic diversity is considered essential for the
ability of species to avoid inbreeding depression and to adapt to environmental change
(Girman et al., 1997; Frankham, 2005). The risk of inbreeding depression is greatest in
species without histories of close inbreeding and whose decline is primarily caused by
gradual habitat loss (Jamieson, 2007). Gene flow, a measure of functional connectivity, may
vary among species because it is largely determined by the extent to which landscapes are
permeable for dispersal between fragments (Johansson et al., 2005; Uezu et al., 2005; Crooks
and Sanjayan, 2006). Studies on species with limited mobility have shown that populations
that are fragmented by roads and other urban development have substantially reduced gene
flow and genetic diversity, even for populations that have been isolated only recently
(Vandergast et al., 2009; Delaney et al., 2010). Yet urban fragmentation may not affect
functional connectivity for species with greater mobility in the same manner (Culley et al.,
2007).
Restricted connectivity and subsequent genetic effects are of particular concern for
mammalian carnivores because of their relatively low population densities, large ranges, and
sensitivity to habitat fragmentation (Paetkau et al., 1998). Carnivores avoid inbreeding and
loss of genetic diversity primarily through dispersal (Creel, 1998; Girman et al., 1997), but
relatively little is known about carnivore dispersal because it is infrequent and survival rates
of dispersers are relatively low (Van Vuren, 1998; Kenward et al., 2002). Thus, genetic
monitoring may be able to detect landscape features that restrict connectivity sooner than
do traditional monitoring methodologies, such as telemetry (Vos et al., 2001; Vandergast
et al., 2009). Only a few studies have examined the effects of habitat fragmentation on
genetic population structure of mammalian carnivores within two or more replicate
2012 RUELL ET AL.: URBAN FRAGMENTATION AND GENETIC STRUCTURE OF BOBCATS 267

landscapes (Kyle and Strobeck, 2001; Dallas et al., 2002; Banks et al., 2005), and these did not
address the effects of isolation resulting from urban development.
In coastal southern California, urban fragmentation in conjunction with high levels of
species endemism has created a global hotspot of endangerment and extinction (Wilson,
1992; Dobson et al., 1997). Large carnivores can serve as valuable indicator species of
landscape connectivity and isolation in coastal southern California (Crooks, 2002; Hunter
et al., 2003). Specifically, bobcats (Lynx rufus) are sensitive to fragmentation in southern
California and have a low probability of persisting as habitat fragments become increasingly
isolated and small (Crooks, 2002). Dispersal of bobcats across major roads and urban
development results in lower than expected levels of gene flow (Riley et al., 2006). Thus,
urbanization likely affects the genetic population structure of bobcats in the south coast
ecoregion.
The objectives of this study were to use non-invasive sampling to examine whether the
degree to which major roads and urban development surrounding habitat fragments can be
used to predict the relative genetic isolation of bobcat populations in coastal southern
California. We compared genetic diversity, relatedness between individuals, genetic
structure, and dispersal between populations that differed in their level of isolation by
surrounding urban development within paired sites in two replicate landscapes. We
hypothesized that bobcat populations in sites that were more completely surrounded by
roads and development would experience less functional connectivity relative to less
isolated nearby populations. Populations in fragments with less functional connectivity
should then exhibit characteristics of greater genetic isolation, such as reduced genetic
diversity, greater relatedness between individuals, distinct genetic structure from nearby
populations, and reduced dispersal.

MATERIALS AND METHODS

STUDY SITES

We conducted our surveys on populations of bobcats within paired sites of natural habitat
in two regions of coastal southern California: the Coast and Central sites in the Orange
County (OC) region, south of Los Angeles, and the Simi Hills and Topanga sites in the
Santa Monica Mountains (SMM) region, north of Los Angeles (Fig. 1). Within each pair,
sites differed in the degree to which they were surrounded by adjacent major roads and
urban development. This paired design allowed us to better control for factors that may be
affecting bobcat populations other than urbanization and proximity to other areas of
natural habitat.
We classified land use within and surrounding the study sites using the Southern
California Association of Governments (SCAG, 2005) land-use dataset and vegetation
cover data downloaded in 2007 (LANDFIRE, http://www.landfire.gov/) with ArcView 3.3
(Environmental Systems Research Institute, Redlands, CA). We then defined patches of
natural habitat versus urban areas using a majority rule, which allowed us to view the
landscape at a coarser grain than the original land use data (Chrisman, 1997; Johnston,
1998; Figs. 1A and 1B). Under the majority rule, each 30 m by 30 m raster cell was part of a
habitat patch if the majority ($50%) of a 1-km diameter circle around the center of each
raster cell was coastal sage scrub, chaparral, oak woodlands, riparian woodlands, or
grasslands; this broad set of natural vegetation types was selected because bobcats are habitat
generalists (Crooks, 2002; Riley et al., 2003, 2010; Ordeñana et al., 2010). Otherwise, the
raster cell was not included as part of a habitat patch. Study site boundaries were
approximated from the boundaries of the sampled habitat patches as defined by the
268 THE AMERICAN MIDLAND NATURALIST 168(2)

FIG. 1.—Study sites in Orange County (OC), California, south of Los Angeles, and in the Santa
Monica Mountains (SMM), California, north of Los Angeles, are shown with areas of natural habitat and
urban development. Inset (A) shows the Simi Hills and Topanga sites in the SMM region, sampled in
2004. Inset (B) shows the Coast site sampled in 2003–2004 and 2006–2007, and the Central site sampled
in 2003–2004 in the OC region. Insets depict study sites and other habitat patches as defined by the
majority rule and the three criteria used to determine fragment isolation: one kilometer buffers
surrounding each site, proportion of each site’s boundary adjacent to other habitat patches, and the
major roads and ocean that encircle each site (see text)

majority rule, except for portions of the boundaries of the Simi Hills and Topanga study
sites, which were delineated by roads where sampling transects ended; and the southern and
eastern boundaries of the Central site, which were defined using a convex hull around
trapping locations. The entire boundary of the Coast site corresponded with the boundary
of the sampled habitat patch as defined by the majority rule.

ISOLATION BY MAJOR ROADS AND URBAN DEVELOPMENT

We measured isolation by major roads and urban development for each site using three
criteria we developed based on features of urban development that are known to affect
bobcat movement and habitat use (Fig. 1). The first criterion was an approximation of the
relative permeability of the urban development immediately surrounding each study site,
which we defined as the percentage of all natural habitat vegetation types in a 1-km buffer
zone surrounding each site’s boundary. For this analysis, land cover within the buffer was
not reduced to a coarser grain with a majority rule. The rationale for the first criterion is
that bobcats rely on natural areas and tend to avoid urban development (Crooks, 2002;
2012 RUELL ET AL.: URBAN FRAGMENTATION AND GENETIC STRUCTURE OF BOBCATS 269

Lyren et al., 2008; Ordeñana et al., 2010; Riley et al., 2003, 2006, 2010). Telemetry locations
of radio-collared bobcat within urban development are not commonly recorded farther
than 1-km from areas of natural habitat in this region (Riley et al., 2003, 2006; Lyren et al.,
2006).
The second criterion was the percentage of each study site’s boundary that was directly
adjacent to another habitat patch (see Fig. 1). For this analysis, habitat patches adjacent to
each study site boundary were defined by the majority rule, as described above. The
rationale for this criterion is that habitat patches directly adjacent to the study site
boundaries are more likely to facilitate functional connectivity than adjacent urban
development (Tigas et al., 2002; Crooks and Sanjayan, 2006; Riley et al., 2010).
The final criterion was whether or not major roads (i.e., freeways or major secondary roads
that had speed limits .50 mph, $four lanes of traffic, and divided by direction) or ocean
completely encircled the site, separating it from other natural habitat. Bobcats generally
avoid crossing busy, paved roads, particularly freeways, and territories will tend to build up
alongside major roads rather than straddle over them (Riley et al., 2003, 2006, 2010).
Thus, within each region, the more isolated site of the pair had a smaller proportion of
natural habitat in the 1-km buffer zone, a smaller proportion of its boundary connected to
natural habitat, and was completely encircled by major roads or ocean.

SAMPLE COLLECTION AND GENOTYPING

To ensure that every individual in each population had a reasonable chance of being
sampled, fresh (#4 d old) scat and hair samples of mammalian carnivores were sampled
using scat transects and hair snares along common paths of movement (roads, trails, and dry
creek beds; MacDonald, 1980) that thoroughly covered each site (Ruell and Crooks, 2007;
Ruell et al., 2009). In the OC region, the Coast site was sampled in 2003–2004 and then
again 2006–2007. We analyzed samples from each period separately, because they likely did
not represent the same set of individuals (i.e., many bobcats sampled in the first period were
likely absent from the second period and vice versa). During the first sampling period in
2003–2004, hair and fresh scat of mammalian carnivores were collected at 3 to 4-d intervals
over a 14-d period in spring 2003 (Ruell and Crooks, 2007) and then re-sampled for fresh
scat at 4-d intervals over a 16-d period in summer 2004. During the second sampling period
in 2006–2007, blood and tissue samples were obtained from bobcats trapped along paths of
movement throughout the Coast site as part of an ongoing GPS telemetry study and from
bobcats killed by vehicle collisions within or directly bordering the site during this period
(Lyren et al., 2008). In the Central site, fresh scat and hair samples were collected at 3 to
4-d intervals over a 14-d period in spring 2003 (Ruell and Crooks, 2007). In addition,
we obtained blood samples from bobcats trapped along common paths of movement
throughout the Central site in 2003 and 2004 as part of an ongoing GPS telemetry study
(Lyren et al., 2006; Fig. 2). In the SMM region, fresh scat samples from mammalian
carnivores were collected throughout the Simi Hills and Topanga sites at 4-d intervals over a
16-d period in summer 2004 (Ruell et al., 2009).
We extracted genomic DNA from blood samples using the QIAampH DNA Mini Kit blood
spin protocol, from hair and tissue samples using QIAampH DNA Mini Kit standard tissue
protocol, and from scat samples using the QIAampH DNA Stool Mini Kit stool protocol
(Qiagen, Inc., Valencia, CA). We identified hair and scat samples from bobcat using the 16S
mtDNA restriction digestion method (Mills et al., 2000). Samples from other species were
not analyzed further. All confirmed bobcat samples were genotyped using four
microsatellite loci (FCA026, FCA045, FCA077, and FCA132; Menotti-Raymond et al.,
1999), using M13-tailed primers (Boutin-Ganache et al., 2001) and a modified multiple
270 THE AMERICAN MIDLAND NATURALIST 168(2)

FIG. 2.—A female bobcat that was trapped within the Central site in 2004 as part of a concurrent GPS
telemetry study

tubes approach (Ruell and Crooks, 2007). The use of only four loci and the multiple tube
approach for hair and scat samples minimized the risk of adding ghost individuals to the
sample due to higher genotyping error rates for noninvasive samples, which can bias
estimates of allele frequencies and heterozygosities (Taberlet et al., 1996, 1999; Ruell and
Crooks, 2007). Ruell and Crooks (2007) found that these four polymorphic microsatellite
loci were sufficient to differentiate individuals sampled at nearby sites with a high level of
confidence; the observed probability of identity (P(ID)obs) was zero with just three loci, and
the conservative probability of identity between siblings (P(ID)sib) was estimated to be 0.02
with four loci (calculated as P(ID)sib 5 0.25 + (0.5gpi2) + (0.5g(pi2)2) 2 (0.25gpi4), where pi
is the frequency of the ith allele; Waits et al., 2001; Ruell and Crooks, 2007).
Genotypes were matched using the Excel Microsatellite Toolkit (Park, 2001) to identify
samples that originated from the same individual. Closely matching genotypes were
rechecked for scoring errors. Final lists of individual genotypes were compiled for each site.
In the OC region, there were 15 bobcats from the Coast site in 2003–2004, 22 from the Coast
site in 2006–2007 (four individuals were sampled in the Coast site in both 2003–2004 and
2006–2007) and 38 from the Central site. In the SMM region, there were 19 bobcats from the
Simi Hills site and 19 from the Topanga site. Despite our relatively small sample sizes, we are
confident that a majority of individuals at each site were sampled. Based on population size
estimates using a closed-population heterogeneity estimator, our samples represented 68% of
the population in the Simi Hills site (19/28) and 73% of the population in the Topanga site
(19/26; Ruell et al., 2009) when using only scat sampling; trapping and collecting road-killed
bobcats provided even larger sample sizes than scat sampling of the same sites. Thus, we
assumed that our samples were representative of the populations in subsequent analyses.
2012 RUELL ET AL.: URBAN FRAGMENTATION AND GENETIC STRUCTURE OF BOBCATS 271

Each population was tested for Hardy-Weinberg proportions of genotypes at each locus
and for linkage equilibrium between pairs of loci using program GENEPOP (Raymond and
Rousset, 1995) and a Bonferroni correction for a family-wide a 5 0.05 (Rice, 1989).

GENETIC DIVERSITY AND RELATEDNESS

We compared levels of genetic diversity between populations within each region using the
average number of alleles per locus, the observed heterozygosity (HO), and the expected
heterozygosity (HE) for each population, estimated using program GDA (Lewis and Zaykin,
2001), and the average allelic richness (A) per locus, corrected for differences in sample
size, using program FSTAT version 2.9.3 (Goudet, 2001).
We compared the proportion of close relatives between populations within each region to
evaluate the effects of isolation on the genetic relatedness (R) between pairs of individuals.
We identified close relatives as pairs of individuals with R $ 0.25 with statistical significance
(a 5 0.05) using log-likelihood calculations in program KINSHIP version 1.3.1 (Goodnight,
2006). We compared the proportion of close relatives between populations within each
region using Pearson’s Chi-square tests using Epi Info version 3.3.2 (CDCP, 2005).

POPULATION STRUCTURE

We tested for population genetic structure within each region using pair-wise FST values
with 95% confidence intervals obtained by bootstrapping across loci using program GDA.
Although the use of FST to measure population differentiation has recently been questioned
(e.g., Jost, 2008), when mutation rates are relatively high (as with microsatellites), but low
relative to migration rates, FST likely remains the best tool for assessing demographic
processes that lead to population structure (Whitlock, 2011). We also used program
STRUCTURE 2.3.3 to test for population substructure across all populations and across
populations within each region without using a priori population assignments (Pritchard
et al., 2000), and then using collection location (i.e., the study site from which the individual
was sampled) as a Bayesian prior (Waples and Gaggiotti, 2006). For the analysis including all
four putative populations (two samples from the Coast population), we ran K values from
one to eight to find likelihood values associated with the different estimates of the number
of populations, because the roads bisecting the Coast and Central sites could have caused
population substructure in each site. We ran each K value 10 times to test for consistency
when using a burn-in of 50,000 iterations, 500,000 Markov chain Monte Carlo (MCMC)
repetitions, and allowing for population admixture. The most likely number of clusters was
determined using the methods of Evanno et al. (2005).

DISPERSAL

A dispersal estimate was obtained for each population by identifying the number of first
generation migrants with $95% confidence using the L-home likelihood computation, the
Bayesian method of classification (Rannala and Mountain, 1997), and the Monte-Carlo
simulation algorithm (Paetkau, 2004) in program GENECLASS2, which does not assume that
the population of origin was always sampled (Piry et al., 2004). We compared the
proportions of first generation migrants between populations within each region using two-
tailed Fisher’s exact tests (Langsrud, 2006). For each population, we also compared the
proportion of individuals that were excluded from either site within a region with
assignment probabilities ,0.5, using the Monte-Carlo re-sampling procedure of Paetkau
et al. (2004) in program GENECLASS2. These individuals could have originated from
unsampled populations. We compared the proportion of excluded individuals between
the populations within each region using two-tailed Fisher’s exact tests (Langsrud, 2006).
272 THE AMERICAN MIDLAND NATURALIST 168(2)

TABLE 1.—Indices of isolation for the Coast site in 2003–2004 and 2006–2007 and the Central site in
2003–2004 in the Orange County region in California, and the Simi Hills and Topanga sites in the Santa
Monica Mountains region in California in 2004. The first, labeled ‘‘habitat in buffer,’’ was the
proportion of natural habitat in a one-kilometer buffer zone surrounding each study site boundary. The
second, labeled ‘‘habitat adjacent to boundary,’’ was the proportion of each site’s boundary that was
directly adjacent to another habitat patch, defined using a majority rule. The third, labeled ‘‘encircling
roads/ocean,’’ was whether or not major roads or ocean completely encircled each site

Indices of isolation
Region and study site Habitat in buffer Habitat adjacent to boundary Encircling roads/oceans

Orange County
Coast 29% 0% Yes
Central 57% 47% No
Santa Monica Mountains
Simi Hills 32% 10% Yes
Topanga 44% 44% No

RESULTS
ISOLATION BY MAJOR ROADS AND URBAN DEVELOPMENT

In the OC region, we designated the Coast site as more isolated based on all three criteria
(Table 1). The Coast site was bordered to the west by the Pacific Ocean, and to the north,
east, and south by major (U.S. Interstate 405, U.S. Interstate 5, and State Route 55) and
secondary roads, as well as residential, military, and agricultural development (Fig. 1). The
Central site was designated as the less isolated site of the pair. Although it was partially
bordered to the north, south, and west by major roads (U.S. Interstate 5, State Route 91,
State Route 55, and State Route 261) and residential, military, and agricultural development, a
large proportion of its eastern border was contiguous with large areas of natural habitat in the
adjacent Cleveland Natural Forest. The Coast and Central sites were each bisected by a two-
lane secondary road and a major road that had several wildlife underpasses with documented
bobcat crossings (Crooks and Jones, 1998; Lyren et al., 2006; Lyren et al., 2008).
In the SMM region, we designated the Simi Hills site as more isolated based on all three
criteria, although the difference between sites was not as marked as in the OC region
(Table 1). The Simi Hills site was bordered to the west by residential development and a
major road (State Route 23), separated from the Santa Susana Mountains to the north by
residential development and a major road (State Route 118), bordered to the east by
.20 km of residential development and major roads, and separated from the Santa Monica
Mountains by residential development and a major road (U.S. Highway 101; Fig. 1). The
Topanga site was designated as less isolated primarily because it was located within the large
natural areas of the Santa Monica Mountains and thus, had a large area of natural habitat
along its western border. However, the Topanga site was bordered to the south by
residential development, roads, and the Pacific Ocean, and to the north and east by major,
multi-lane roads (U.S. Highway 101 and U.S. Interstate 405), secondary roads and
residential development.

GENOTYPING

Across all individuals (n 5 109), the mean number of alleles per locus was 8.75 (range: 6–
13 alleles per locus). There was no deviation from Hardy-Weinberg in the proportions of
homozygotes and heterozygote expected in any population. Only one of 30 tests for linkage
2012 RUELL ET AL.: URBAN FRAGMENTATION AND GENETIC STRUCTURE OF BOBCATS 273

TABLE 2.—Measures of genetic diversity (mean number of alleles per locus, mean allelic richness per
locus corrected for sample size (A), observed heterozygosity (HO), expected heterozygosity (HE)), the
proportion of closely related pairs of individuals (R $ 0.25 with 95% confidence), the proportion of
individuals sampled that were first generation migrants, and the proportion of individuals that did not
assign to either sampled population in the region for bobcats in southern California. In the Orange
County region, data are presented for the Coast population in 2003–2004 and 2006–2007, and the
Central population in 2003–2004. In the Santa Monica Mountains region, data are presented for the
Simi Hills and Topanga populations in 2004

Mean no. Closely related 1st generation Excluded from both

alleles per indiv. migrants populations (assign.
Region and Population n locus A HO HE (R $ 0.25) (a 5 0.05) prob. , 0.05)

Orange County
Coast in 2003–2004 15 6.3 6.1 0.64 0.73 34% 13% 13%
Coast in 2006–2007 22 5.5 6.1 0.68 0.69 8% 9% 32%
Central 38 7.3 6.4 0.78 0.80 11% 5% 47%
Santa Monica Mountains
Simi Hills 19 6.5 5.0 0.79 0.78 23% 11% 32%
Topanga 19 5.5 5.0 0.61 0.66 36% 16% 37%

disequilibrium was significant (FCA045 and FCA026 in the Central population). However, this
result was likely to have arisen by chance since there was no evidence for linkage disequilibrium
between these loci in the other populations, so they were likely not physically linked.

GENETIC DIVERSITY AND RELATEDNESS

Following our expectation, the more isolated Coast population had lower genetic diversity
for all measures in both 2003–2004 and 2006–2007 than the less isolated population,
Central, in the OC region (Table 2). Contrary to our expectation for the SMM region, the
population initially designated as more isolated, Simi Hills, had slightly higher genetic
diversity for all measures except A (mean allelic richness) than the population designated as
less isolated, Topanga (Table 2).
As expected, the proportion of closely related individuals (R $ 0.25 at the 95%
confidence level) was greater in the Coast population in 2003–2004 than in the Central
population in the OC region (x2 5 39.79, d.f. 5 1, P , 0.001; Table 2). However, when
comparing the Coast population in 2006–2007 to the Central population, there was no
longer a difference in the proportion of closely related individuals (x2 5 2.40, d.f. 5 1, P 5
0.121; Table 2). Also, contrary to expectation, the proportion of closely related individuals
was significantly lower in the Simi Hills population than in the Topanga population in the
SMM region (x2 5 6.84, d.f. 5 1, P 5 0.009; Table 2).

POPULATION STRUCTURE

Populations exhibited moderate levels of genetic population structure that were

significantly different from zero within both regions. Pairwise FST values for the OC region
were 0.072 (95% CI 5 0.016–0.120) between the Coast population in 2003–2004 and the
Central population, and 0.092 (95% CI 5 0.022–0.149) between the Coast population in
2006–2007 and the Central population. Similarly, FST was 0.089 (95% CI 5 0.037–0.134)
between the Simi Hills and Topanga populations in the SMM region.
The STRUCTURE analysis indicated that K 5 2 population clusters resulted in the largest
change in log likelihood values, and based upon the delta-K method (Evanno et al., 2005)
274 THE AMERICAN MIDLAND NATURALIST 168(2)

FIG. 3.—(A) The STRUCTURE analysis of the most likely number of clusters based upon the delta-K
method (Evanno et al., 2005) without using a priori population assignments showed that the most
probable number of clusters was K 5 2, where the largest change in likelihood occured. Genetic
assignment of individuals from all populations is shown for two inferred clusters (K 5 2). Individuals are
represented by columns and grouped by the region and study site from which they were sampled. The
proportion of each shade of gray in each column represents that individual’s probabilities of assignment
to the corresponding cluster. Individuals from Central, Topanga, and Simi Hills primarily assigned to
one cluster and individuals from the Coast (in both the 2003–2004 and 2006–2007 samples) primarily
assigned to the other. (B) The STRUCTURE analysis with collection location as a Bayesian prior detected
population structure among all four populations (K 5 4 clusters)

was the true number of clusters (Fig. 3A). The majority of individuals in the Coast
populations in 2003–2004 and 2006–2007 assigned primarily to one cluster and the majority
of individuals in the Central, Simi Hills, and Topanga populations assigned primarily to the
other (Fig. 3A). However, the STRUCTURE analysis using sample location as a Bayesian prior
showed population structure among all four populations (K 5 4; Fig. 3B).

DISPERSAL

Contrary to our expectation, the proportion of first generation migrants was similar between
populations in each region (two-tailed Fisher’s exact tests: P 5 0.586 for the Coast population in
2003–2004 versus the Central population, P 5 0.619 for the Coast population in 2006–2007
versus the Central population, and P 5 1.000 for the Simi Hills population versus the Topanga
population; Table 2). Consistent with our prediction, assignment tests indicated a significantly
smaller proportion of individuals in the Coast population in 2003–2004 than in the Central
population were either migrants from unsampled populations or descendants of migrants (two-
tailed Fisher’s exact test, P 5 0.028; Table 2). The proportion of individuals from outside
populations, however, was not significantly different between the Coast population in 2006–
2007 and the Central population (two-tailed Fisher’s exact test, P 5 0.286) or between
populations in the SMM region (two-tailed Fisher’s exact test, P 5 1.000; Table 2).
2012 RUELL ET AL.: URBAN FRAGMENTATION AND GENETIC STRUCTURE OF BOBCATS 275

DISCUSSION
Fragmentation by urban development likely affected the functional connectivity between
bobcat populations in both study regions in southern California but not always as predicted
by the degree to which major roads and urban development surrounded populations. As
predicted, greater isolation of bobcat habitat in the Orange County region was associated
with greater genetic isolation, particularly during our first sampling period. Bobcats in the
more isolated Coast population in 2003–2004 had lower genetic diversity and higher
relatedness relative to the less isolated Central population. Although smaller sample sizes in
the Coast population could have lowered some of our estimates of genetic diversity, allelic
richness values, which were corrected for sample size, were also lower. In the Central
population, we detected a greater proportion of bobcats that did not assign to either
population in the OC region and thus, likely originated from unsampled populations in the
extensive core habitat in the adjacent Cleveland National Forest. Pairwise FST values for the
Coast and Central populations (0. 072 for 2003–2004 and 0.092 for 2006–2007) were larger
than those reported for other bobcat (0.067; Millions and Swanson, 2006; 0.060; Millions
and Swanson, 2007) and Canada lynx (Lynx canadensis; #0.051; Schwartz et al., 2002)
populations separated by water or much larger geographic distances. Additionally, the
STRUCTURE analysis supports limited gene flow between these populations (Fig. 3B). This was
not surprising, given that the Coast site was separated from other areas of natural habitat by
the 26-lane interchange between U.S. Interstate 405 and U.S. Interstate 5.
By our second sampling period in 2006–2007, however, the relationship between isolation
and functional connectivity was not as pronounced in the OC region, with partially
conflicting results as to the genetic isolation of the Coast population. As predicted, genetic
diversity in the Coast population in 2006–2007 was still significantly lower than in the
Central population and the STRUCTURE analysis indicated that the two populations were still
genetically distinct (Fig. 3B). However, in contrast to predictions, by 2006–2007 relatedness
was no longer higher in the Coast population compared to the Central population. Given
that the Coast population was relatively small, gene flow between the sampling periods, even
if limited, may have been sufficient to counteract some of the effects of isolation that we
detected in 2003–2004. Likewise, by 2006–2007, the proportion of individuals excluded
from both populations in their region no longer significantly differed between the Coast
and the Central populations, potentially indicating that more individuals had been able to
disperse into the Coast site from populations other than the Central population. However,
given the small number of bobcats and loci available in this study, we lacked the statistical
power to determine whether the observed fluctuations in genetic isolation of the Coast site
were simply an artifact or represented real differences between sampling periods due to
gene flow. Nonetheless, the detection of first generation migrants in both sampling periods
suggests that isolation by major roads and urban development has not completely
eliminated gene flow into the Coast population and that the matrix surrounding the Coast
site was still somewhat permeable to bobcat movement. The exact locations of movement
routes into the Coast population are unknown, but immigrating bobcats might be
originating from natural areas to the southeast (Lyren et al., 2008; Fig. 1).
Similar to the OC populations, the pairwise FST value for the Simi Hills and Topanga
populations (0.089) was relatively large compared to those reported elsewhere. This FST
values was also larger than that reported in Riley et al., (2006; #0.064) for nearby bobcat
populations in the SMM region that were closer in proximity than the Simi Hills and
Topanga populations but were still separated by U.S. Interstate 101. Despite this, the
STRUCTURE analysis could not detect structure between the Simi Hills, Topanga, or Central
276 THE AMERICAN MIDLAND NATURALIST 168(2)

study sites without the a priori use of collection location. There are several potential
explanations for the lack of differentiation. First, the Simi Hills, Topanga, and Central sites
may still be connected enough that there is sufficient gene flow among them to maintain
similar genetic structure. Second, these three populations may not have been isolated long
enough to diverge from each other, and given enough time may be as isolated as the Coast
population is from the rest. However, these northern populations are located on the
opposite side of the sprawling metropolis of greater Los Angeles from Central (Fig. 1).
Thus, the third and most likely explanation is that we simply lacked the statistical power to
differentiate between populations due to the fairly small samples for Simi Hills and
Topanga, and the small number of loci used in the analyses (Waples and Gaggiotti, 2006).
Once collection locations were used as a Bayesian prior in the STRUCTURE analysis, the four
populations were more clearly differentiated (Fig. 3B).
However, the direction of the genetic differences in the SMM region were largely opposite
of what we expected a priori based on levels of urban development surrounding both sites.
Although similar across many measures of genetic isolation, the population with less
surrounding development, Topanga, appeared more genetically isolated than the site we
initially designated as more structurally isolated, Simi Hills, because Topanga had lower
genetic diversity and a significantly larger proportion of close relatives. Because bobcats can
navigate through urban development, particularly via movement corridors or small habitat
patches within the urban matrix (Riley et al., 2010), the genetic variability of populations in
close proximity to the sites in the SMM region may have contributed to the unexpected
patterns of genetic isolation. The only large nearby source of genetic variation for the
Topanga population was the rest of the Santa Monica Mountains along its western border,
which itself still constitutes a habitat fragment, although a larger one than Topanga
(Fig. 1A; Riley et al., 2006). In contrast, the Simi Hills population was adjacent to two likely
source populations, potentially receiving migrants from both the larger Santa Susana
Mountains populations to the north and from the Santa Monica Mountains to the south.
These results reinforce the finding that measuring functional connectivity across more
than one landscape and at multiple geographic and temporal scales is necessary to test the
relationship between measures of habitat fragmentation and genetic population structure
(Keyghobadi, 2007). The regional differences in the relationship between development
intensity surrounding sites and genetic population structure suggest that, at least for
relatively mobile species, close proximity to a genetically diverse source population or
multiple source populations of genetic variation, even if some urban development
intervenes, may be as or more important for preserving genetic diversity than more direct
habitat connectivity to another population with reduced genetic diversity. Population
genetic models demonstrate that the viability of fragmented populations depends not only
on the number of migrants entering per generation but also on the amount of gene flow
into the source populations from where the fragmented populations receive migrants
(Couvet, 2002). Thus, even when there are seemingly adequate levels of dispersal into
fragmented populations, their viability may still be threatened if source populations do not
have adequate connectivity to other sources of genetic variation.

Acknowledgments.—Thanks to C. E. Lee at UW-Madison, R. K. Wayne at UCLA, and M. E. Douglas

at CSU for laboratory facilities, laboratory expertise, and valuable guidance on this project and
manuscript. R. Alonso, E. Ambat, M. Cegelski, M. Ervin, G. Geye, C. Haas, R. Mowry, J. Naegele, T.
Smith, G. Turschak, and E. York provided field support. C. Talbert and M. Ackerman assisted in the
laboratory. Thanks to J. Lee and S. Vandewoude for their help. The Nature Conservancy, The Irvine
Company, The Nature Reserve of Orange County, Transportation Corridor Agencies, Oscar and Isabel
2012 RUELL ET AL.: URBAN FRAGMENTATION AND GENETIC STRUCTURE OF BOBCATS 277

Anderson Graduate Scholarship, Douglas L. Gilbert Memorial Scholarship, Colorado State University
College of Natural Resources Need-Based Scholarship, Theodore Roosevelt Memorial Fund Grant,
Rocky Mountain Goats Foundation’s Bill Burtness Fellowship, and the National Science Foundation
Ecology of Infectious Disease research program (NSF EF-0723676) supported this work. Any use of
trade, product, or firm names is for descriptive purposes only and does not imply an endorsement by
the U.S. Government.

LITERATURE CITED
BANKS, S. C., G. R. FINLAYSON, S. J. LAWSON, D. B. LINDENMAYER, D. PAETKAU, S. J. WARD, AND A. C. TAYLOR.
2005. The effects of habitat fragmentation due to forestry plantation establishment on the
demography and genetic variation of a marsupial carnivore, Antechinus agilis. Biol. Conserv.,
122:581–597.
BOUTIN-GANACHE, I., M. RAPOSO, M. RAYMOND, AND C. F. DESCHEPPER. 2001. M13-tailed primers improve the
readability and usability of microsatellite analyses performed with two different allele-sizing
methods. Biotechniques, 31:24–26.
CENTERS FOR DISEASE CONTROL AND PREVENTION (CDCP). 2005. Epi Info (version 3.3.2). http://www.cdc.
gov/epiinfo/.
CHRISMAN, N. 1997. Exploring Geographic Information Systems. John Wiley & Sons, New York. 298 p.
COUVET, D. 2002. Deleterious effects of restricted gene flow in fragmented populations. Conserv. Biol.,
16:369–376.
CREEL, S. 1998. Social organization and effective population size in carnivores, p. 246–265. In: T. Caro
(ed.). Behavioral Ecology and Conservation Biology. Oxford University Press, Oxford. 582 p.
CROOKS, K. AND D. JONES. 1998. Monitoring program for carnivore corridor use in the Nature Reserve
of Orange County. Annual Report to Nature Reserve of Orange County, Santa Ana,
California(unpublished report).
CROOKS, K. R. 2002. Relative sensitivities of mammalian carnivores to habitat fragmentation. Conserv.
Biol., 16:488–502.
——— AND M. SANJAYAN. 2006. Connectivity Conservation. Cambridge University Press, Cambridge.
693 p.
CULLEY, T. M., S. J. SBITA, AND A. WICK. 2007. Population genetic effects of urban habitat fragmentation in
the perennial herb Viola pubescens (Violaceae) using ISSR markers. Ann. Bot., 100:91–100.
CZECH, B., P. R. KRAUSMAN, AND P. K. DEVERS. 2000. Economic associations among causes of species
endangerment in the United States. Bioscience, 50:593–601.
DALLAS, J. F., F. MARSHALL, S. B. PIERTNEY, P. J. BACON, AND P. A. RACEY. 2002. Spatially restricted gene flow
and reduced microsatellite polymorphism in the Eurasian otter Lutra lutra in Britain. Conserv.
Genet., 3:15–29.
DELANEY, K. S., S. P. D. RILEY, AND R. N. FISHER. 2010. A rapid, strong, and convergent genetic response to
urban habitat fragmentation in four divergent and widespread vertebrates. PLoS ONE,
5:e12767.
DOBSON, A. P., J. P. RODRIGUEZ, W. M. ROBERTS, AND D. S. WILCOVE. 1997. Geographic distribution of
endangered species in the United States. Science, 275:550–553.
EVANNO, G., S. REGNAUT, AND J. GOUDET. 2005. Detecting the number of clusters of individuals using the
software STRUCTURE: a simulation study. Mol. Ecol., 14:2611–2620.
FRANKHAM, R. 2005. Genetics and extinction. Biol. Conserv., 126:131–140.
GIRMAN, D., G. MILES, E. GEFFEN, AND R. WAYNE. 1997. A molecular genetic analysis of social structure,
dispersal, and pack interactions in the African wild dog (Lycaon pictus). Behav. Ecol. Sociobiol.,
40:187–198.
GOODNIGHT, K. F. 2006. KINSHIP (version 1.3.1). http://www.gsoftnet.us/GSoft.html
GOUDET, J. 2001. FSTAT (version 2.9.3): a program to estimate and test gene diversities and fixation
indices. http://www.unil.ch/izea/softwares/fstat.html.
HUNTER, R., R. FISHER, AND K. R. CROOKS. 2003. Landscape-level connectivity in coastal southern
California as assessed by carnivore habitat suitability. Nat. Areas J., 23:302–314.
278 THE AMERICAN MIDLAND NATURALIST 168(2)

JAMIESON, I. G. 2007. Has the debate over genetics and extinction of island endemics truly been resolved?
Anim. Conserv., 10:139–144.
JELLINEK, S., D. A. DRISCOLL, AND J. B. KIRKPATRICK. 2004. Environmental and vegetation variables have a
greater influence than habitat fragmentation in structuring lizard communities in remnant
urban bushland. Austral Ecol., 29:294–304.
JOHANSSON, M., C. R. PRIMMER, J. SAHLSTEN, AND J. MERILÄ. 2005. The influence of landscape structure on
occurrence, abundance and genetic diversity of the common frog, Rana temporaria. Glob. Change
Biol., 11:1664–1679.
JOHNSTON, C. A. 1998. Geographic Information Systems in Ecology. Blackwell Science, Oxford, U.K. 239 p.
JOST, L. 2008. GST and its relatives do not measure differentiation. Mol. Ecol., 17:4015–4026.
KELLER, L. F. AND D. M. WALLER. 2002. Inbreeding effects in wild populations. Trends Ecol. Evol.,
17:230–241.
KENWARD, R. E., S. P. RUSHTON, C. M. PERRINS, D. W. MACDONALD, AND A. B. SOUTH. 2002. From marking to
modeling: dispersal study techniques for land vertebrates, p. 50–71. In: J. M. Bullock, R. E.
Kenward, and R. S. Hails (eds.). Dispersal Ecology. Blackwell Publishing, Oxford, U.K. 458 p.
KEYGHOBADI, N. 2007. The genetic implications of habitat fragmentation for animals. Canadian J. Zool.,
85:1049–1064.
KYLE, C. J. AND C. STROBECK. 2001. Genetic structure of North American wolverine (Gulo gulo) popu-
lations. Mol. Ecol., 10:337–347.
LANGSRUD, Ø. 2006. Fisher’s Exact Test. http://www.langsrud.com/fisher.htm
LEWIS, P. O. AND D. ZAYKIN. 2001. GDA (version 1.0). http://hydrodictyon.eeb.uconn.edu/people/
plewis/software.php.
LYREN, L. M., R. S. ALONSO, K. R. CROOKS, AND E. E. BOYDSTON. 2008. GPS telemetry, camera trap, and
mortality surveys of bobcats in the San Joaquin Hills, Orange County, California. Department
of the Interior, U.S. Geological Survey, Western Ecological Research Center, Sacramento-
Technical Report.
———, G. M. TURSCHAK, E. S. AMBAT, C. D. HAAS, J. A. TRACEY, J. A., E. E. BOYDSTON, S. A. HATHAWAY, R. N.
FISHER, AND K. R. CROOKS. 2006. Carnivore activity and movement in a southern California
protected area, the North/Central Irvine Ranch. Orange County, California. Department of
the Interior, U.S. Geological Survey, Western Ecological Research Center, Sacramento,
California. Technical Report.
MACDONALD, D. W. 1980. Patterns of scent marking with urine and faeces amongst carnivore
communities. Symp. Zool. Soc. Lond., 45:107–139.
MENOTTI-RAYMOND, M. A., V. A. DAVID, L. A. LYONS, A. A. SCHÄFFER, J. F. TOMLIN, AND S. J. O’BRIEN. 1999. A
genetic linkage map of microsatellites in the domestic cat (Felis catus). Genetics, 57:9–23.
MILLIONS, D. G. AND B. J. SWANSON. 2006. An application of Manel’s Model: detecting bobcat poaching in
Michigan. Wildl. Soc. Bull., 34:150–155.
——— AND ———. 2007. Impact of natural and artificial barriers to dispersal on the population
structure of bobcats. J. Wildl. Mgmt., 71:96–102.
MILLS, S. L., K. L. PILGRIM, M. K. SCHWARTZ, AND K. MCKELVEY. 2000. Identifying lynx and other North
American felids based on MtDNA analysis. Conserv. Genet., 1:285–288.
ORDENAÑA, M. A., K. R. CROOKS, E. E. BOYDSTON, R. N. FISHER, L. M. LYREN, S. SIUDYLA, C. D. HASS, S. HARRIS,
S. A. HATHAWAY, G. M. TURSCHAK, A. K. MILES, AND D. H. VAN VUREN. 2010. Effects of urbanization
on carnivore species distribution and richness. J. Mammal., 91:1322–1331.
PAETKAU, D. 2004. Genetic assignment methods for the direct, real-time estimation of migration rate: a
simulation-based exploration of accuracy and power. Mol. Ecol., 13:55–65.
———, R. SLADE, M. BURDEN, AND A. ESTOUP. 2004. Genetic assignment methods for the direct, real-time
estimation of migration rate: a simulation-based exploration of accuracy and power. Mol. Ecol.,
13:55–65.
———, L. P. WAITS, P. L. CLARKSON, L. CRAIGHEAD, E. VYSE, R. WARD, AND C. STROBECK. 1998. Variation in
genetic diversity across the range of North American brown bears. Conserv. Biol., 12:418–429.
PARK, S. D. E. 2001. The Excel Microsatellite Toolkit. http://animalgenomics.ucd.ie/sdepark/
ms-toolkit/.
2012 RUELL ET AL.: URBAN FRAGMENTATION AND GENETIC STRUCTURE OF BOBCATS 279

PIRY, S., A. ALAPETITE, J.–M. CORNUET, D. PAETKAU, L. BAUDOUIN, AND A. ESTOUP. 2004. GENECLASS2:
a software for genetic assignment and first-generation migrant detection. J. Hered., 95:536–
539.
PRITCHARD, J. K., M. STEPHENS, AND P. DONNELLY. 2000. Inference of population structure using multilocus
genotype data. Genetics, 155:945–959.
RANNALA, B. AND J. L. MOUNTAIN. 1997. Detecting immigration by using multilocus genotypes. Proc. Natl.
Acad. Sci. USA, 94:9197–9201.
RAYMOND, M. AND F. ROUSSET. 1995. GENEPOP (version 1.2): population genetics software for exact tests
and ecumenicism. J. Hered., 86:248–249.
RICE, W. R. 1989. Analyzing tables of statistical tests. Evolution, 43:223–225.
RILEY, S. P. D. 2006. Spatial ecology of bobcats and gray foxes in urban and rural zones of a national
park. J. Wildl. Mgmt., 70:1425–1435.
———, E. E. BOYDSTON, K. R. CROOKS, AND L. M. LYREN. 2010. Bobcats (Lynx rufus), p. 120–138. In: S. D.
Gehrt, S. P. D. Riley, and B. Cypher (eds.). Urban Carnivores: Ecology, Conflict, and
Conservation. Johns Hopkins University Press, Baltimore. 285 p.
———, J. P. POLLINGER, R. M. SAUVAJOT, E. C. YORK, C. BROMLEY, T. K. FULLER, AND R. K. WAYNE. 2006. A
southern California freeway is a physical and social barrier to gene flow in carnivores. Mol. Ecol.,
10:1–9.
———, R. M. SAUVAJOT, T. K. FULLER, E. C. YORK, D. A. KAMRADT, C. BROMLEY, AND R. K. WAYNE. 2003.
Effects of urbanization and habitat fragmentation on bobcats and coyotes in Southern
California. Conserv. Biol., 17:566–576.
RUELL, E. W. AND K. R. CROOKS. 2007. Evaluation of noninvasive genetic sampling methods for felid and
canid populations. J. Wildl. Manage., 71:1690–1694.
———, S. P. D. RILEY, M. R. DOUGLAS, J. P. POLLINGER, AND K. R. CROOKS. 2009. Estimating bobcat
population sizes and densities in a fragmented urban landscape using noninvasive capture-
recapture sampling. J. Mammal., 90:129–135.
SANTOS, T., J. A. DÍAZ, J. PÉREZ-TRIS, R. CARBONELL, AND J. L. TELLERÍA. 2008. Habitat quality predicts the
distribution of a lizard in fragmented woodlands better than habitat fragmentation. Anim.
Conserv., 11:46–56.
SCHWARTZ, M. K., L. S. MILLS, K. S. MCKELVEY, L. F. RUGGIERO, AND F. W. ALLENDORF. 2002. DNA reveals high
dispersal synchronizing the population dynamics of Canada lynx. Nature, 415:520–522.
SOUTHERN CALIFORNIA ASSOCIATION OF GOVERNMENTS (SCAG). 2005. Land-use data for Los Angeles and
Ventura counties. SCAG, Los Angeles.
TABERLET, P., S. GRIFFIN, B. GOOSSENS, S. QUESTIAU, V. MANCEAU, N. ESCARAVAGE, L. P. WAITS, AND J. BOUVET.
1996. Reliable genotyping of samples with very low DNA quantities using PCR. Nucleic Acids Res.,
26:3189–3194.
———, L. P. WAITS, AND G. LUIKART. 1999. Noninvasive genetic sampling: look before you leap. Trends
Ecol. Evol., 14:323–327.
THOMPSON, J. N., O. J. REICHMAN, P. J. MORIN, G. A. POLIS, M. E. POWER, R. W. STERNER, C. A. COUCH, L.
GOUGH, R. HOLT, D. U. HOOPER, F. KEESING, C. R. LOVELL, B. T. MILNE, M. C. MOLLES, D. W.
ROBERTS, AND S. Y. STRAUSS. 2001. Frontiers of ecology. BioScience, 51:15–24.
TIGAS, L. A., D. H. VAN VUREN, AND R. M. SAUVAJOT. 2002. Behavioral responses of bobcats and coyotes to
habitat fragmentation and corridors in an urban environment. Biol. Cons., 108:299–306.
TISCHENDORF, L. AND L. FAHRIG. 2000. On the usage and measurement of landscape connectivity. Oikos,
90:7–19.
UEZU, A., J. P. METZGER, AND J. M. E. VIELLIARD. 2005. Effects of structural and functional connectivity and
patch size on the abundance of seven Atlantic Forest bird species. Biol. Conserv., 123:507–519.
VAN VUREN, D. 1998. Mammalian dispersal and reserve design, p. 369–393. In: T. Caro (ed.). Behavioral
Ecology and Conservation Biology. Oxford University Press, Oxford. 582 p.
VANDERGAST, A. G., E. A. LEWALLEN, J. DEAS, A. J. BOHONAK, D. B. WEISSMAN, AND R. N. FISHER. 2009. Loss of
genetic connectivity and diversity in urban microreserves in a southern California endemic
Jerusalem cricket (Orthoptera: Stenopelmatidae: Stenopelmatus n. sp. ‘‘santa monica’’). J. Insect
Conserv., 13:329–345.
280 THE AMERICAN MIDLAND NATURALIST 168(2)

VOS, C. C., A. G. ANTONISSE-DE JONG, P. W. GOEDHARD, AND M. J. M. SMULDERS. 2001. Genetic similarity as a
measure for connectivity between fragmented populations of the moor frog (Rana arvalis).
Heredity, 86:598–608.
WAPLES, R. S. AND O. GAGGIOTTI. 2006. What is a population? An empirical evaluation of some genetic
methods for identifying the number of gene pools and their degree of connectivity. Mol. Ecol.,
15:1419–1439.
WAITS, L. P., L. GORDON, AND P. TABERLET. 2001. Estimating the probability of identity among genotypes in
natural populations: cautions and guidelines. Mol. Ecol., 10:249–256.
WHITLOCK, M. C. 2011. G’ST and D do not replace FST. Mol. Ecol., 20:1083–1091.
WILCOVE, D. S., D. ROTHSTEIN, J. DUBOW, A. PHILLIPS, AND E. LOSOS. 1998. Quantifying threats to imperiled
species in the United States. BioScience, 48:607–615.
WILSON, E. O. 1992. The Diversity of Life. W. W. Norton & Company, New York. 424 p.
WRIGHT, L. I., T. TREGENZA, AND D. J. HOSKEN. 2008. Inbreeding, inbreeding depression, and extinction.
Conserv. Genet., 9:833–843.

SUBMITTED 14 JULY 2011 ACCEPTED 3 APRIL 2012