TEXAS TECH UNIVERSITY

Natural Science Research Laboratory

Occasional Papers

Museum of Texas Tech University

Number 289 13 October 2009

Genetic Variation Within Populations of a Dietary Specialist, Neotoma
stephensi (Stephen’s Woodrat), in Arizona

Michelle L. Haynie, Charles F. Fulhorst, and Robert D. Bradley

Abstract

Seven microsatellite loci were used to develop multilocus genotypes for 49 individuals
of Neotoma stephensi (Stephen’s woodrat) collected from nine sites in central and northern
Arizona. Several statistical analyses were used to determine genetic structure, levels of genetic
variability, and degree of relatedness. Structure analyses estimated two distinct groups within
localities. The F ST value indicated moderate genetic differentiation among groups and sites. All
populations displayed low to moderate levels of genetic diversity in terms of mean expected
heterozygosity, and low to moderate levels of genetic diversity in terms of mean polymorphic
information content. Mean relatedness values were low within groups and sites. Comparison
of genetic diversity in N. stephensi to other species of Neotoma indicated that levels of genetic
variation were comparable to other species of woodrats, including habitat specialists such as N.
magister and wide ranging species such as N micropus.

Key words: genetic variation, microsatellites, Neotoma stephensi , population structure

Introduction

Neotoma stephensi (Stephen’s woodrat) is re¬
stricted to northern Arizona, northwestern New Mexico,
and southwestern Utah (Hall 1981; Hoffmeister 1986;
Jones and Hildreth 1989). It lives in rock outcroppings
in association with juniper and pinyon (Jones and Hil¬
dreth 1989). Vaughan (1982) described N. stephensi as
“a relict species, mostly restricted to a series of isolated
or semi-isolated remnants of xeric juniper woodland... ”
This species is unique among Neotoma in that it is a
dietary specialist, feeding almost exclusively on juni¬
pers (Vaughan 1982; Dial 1988). Because this species
has a restricted range and habitat, little is known about
its biology; although, it has been widely studied from
a physiological standpoint (Dearing et al. 2000, 2001,

2002; Boyle and Dearing 2003; Sorensen and Dearing
2003, 2004; Sorensen et al. 2004) and moderately ex¬
amined from a systematic standpoint (Goldman 1910;
Hoffmeister and de la Torre 1960; Hooper 1960; Planz
et al. 1996; Edwards and Bradley 2002).

Only a few studies have examined genetic diver¬
sity at the population level within this genus (Castle¬
berry et al. 2002; Matocq 2002, 2004; Monty et al.
2003; Mendez-Harclerode et al. 2005, 2007; Haynie
et al. 2007). Castleberry et al. (2002) examined 357
N. magister (Allegheny woodrat) from nine popula¬
tions throughout its range. Although the range of this
species is moderate in size, occurring in the eastern

2

Occasional Papers, Museum of Texas Tech University

United States in the Appalachian and Interior Highland
regions (Hall 1981; Castleberry et al. 2002), it is con¬
sidered to be endangered, threatened, or a species of
concern throughout its range (Castleberry et al. 2002).
Like A. stephensi, A. magister is a habitat specialist
found in isolated rocky outcroppings in forested areas
(Castleberry et al. 2002). Similarly, Haynie et al. (2007)
and Matocq (2002, 2004) examined population level
questions in A. macrotis (large-eared woodrat) and A.
fuscipes (dusky-footed woodrat). Both of these species
have moderately-sized ranges, with the distribution of
A. macrotis encompassing southern portions of the Si¬
erra Nevada Range, South Coast Ranges, southern Cali¬
fornia, and Baja California and the range of A. fuscipes
extending from western Oregon to northern California,
inner Coast Ranges, and northern portions of the Sierra
Nevada Range (Matocq 2002). These species occur in
a variety of habitats including chaparral, coastal sage-
scrub, and densely wooded areas (Murray and Barnes
1969; Carraway and Verts 1991; Tietje and Vreeland
1997). Mendez-Harclerode et al. (2007) examined a
single population (n = 549) of A. micropus (southern
plains woodrat) from south Texas. This species is dis¬
tributed in Kansas, Oklahoma, Texas, New Mexico, and
northeastern Mexico (Hall 1981). Finally, Monty et al.
(2003) examined four populations from a small portion
of the range of A. floridana (eastern woodrat). This
species has a wide range extending throughout large
portions of the central and southeastern United States
(Hall 1981). To date, no population genetic studies
have been performed on A. stephensi. Of the five spe¬

cies examined from a population genetics context, none
have a range that is as restricted as A. stephensi , nor are
any dietary specialists. Based on ranges of the studied
species, genetic diversity within A. stephensi would be
expected to be lower than the more abundant species
with a wider geographic range (i.e. A. floridana or A.
micropus ), but greater than that found in a potentially
endangered species such as A. magister.

Abbott et al. (2004) examined 1,610 Neotoma
from 51 localities in Arizona; 114 (7.1%) were A. ste¬
phensi compared to 1,250 (77.6%) A. albigula (white-
throated woodrat). Eighty-five samples of A. stephensi
were from a single locality and the 29 remaining
samples were distributed among eight other localities.
In another study (Kosoy et al. 1996), 756 Neotoma were
collected; 26 (3.4%) were A. stephensi compared to 395
(52.3%) A. albigula. Lack of A. stephensi samples in
these studies compared to other species of Neotoma
(especially A. albigula ) may be due to several factors
including the fact that A. stephensi are habitat special¬
ists with a restricted range; thus, little is known about
the population biology of this species.

The objectives of this study were to: 1) examine
genetic structure; 2) examine levels of genetic diversity
within populations; and 3) determine degree of genetic
relatedness within populations. To achieve these ob¬
jectives, multilocus microsatellite genotypes were
developed for individuals collected from throughout
this species range in Arizona.

Methods

Collecting localities andDNA extraction. —One
hundred fourteen individuals of A. stephensi were col¬
lected from nine sites throughout central and northern
Arizona (Table 1; Fig. 1). These animals were collected
to determine arenavirus prevalence in multiple woodrat
species (Abbott et al. 2004). Eighty-five samples were
from a single locality (Site 7; Fig. 1), and a random
subsample of 20 was used in this study. The 29 other
samples were distributed throughout the remaining
eight localities. Genomic DNA was extracted from
approximately 25 mg of liver using a DNeasy tissue
extraction kit (Qiagen, Valencia, California).

Microsatellite analysis. —Castleberry et al.
(2000) tested cross-species amplification of 13 loci de¬
veloped for A. magister. Twelve of these primer pairs,
known to amplify up to seven different species of Neo¬
toma (A. magister , A cinerea - bushy-tailed woodrat,
A. floridana, A. fuscipes, A. lepida- desert woodrat, A.
mexicana - Mexican woodrat, and A. micropus), were
used in this study (Table 2) and were amplified via the
polymerase chain reaction (PCR). PCR amplifications
were conducted in 25 pi volumes containing 1 to 1.5
pi 50 ng genomic DNA, 0.6 pi 10 pM each primer, 2.5
pi 10 X PCR buffer, 0.75 to 2 pi 25 mM MgCl 2 , 0.75

Haynie et al.—Genetic Variation in Neotoma stephensi

3

Table 1. Locality data for the 49 individuals o/Neotoma stephensi collected from nine sites in central and northern
Arizona. For each locality, site number (Site; corresponding to Fig. 1), specific locality name (Name), latitude/lon¬
gitude (Lat/Long), number of individuals (N), and ID number (ID#) of individuals at each site (corresponding to Fig.
2) are provided.

Site

Name

Lat/Long

N

ID#

1

AZ: Apache Co.; Three Turkey

36° 1 '447-109 o 24'46"

1

26

2

AZ: Apache Co.; Saint Johns

34°28 , 287-109°19T8"

2

27, 29

3

AZ: Apache Co.; Little Colorado River North

34 o llT67-109°18T7"

3

31-33

4

AZ: Navajo Co.; Lone Pine Reservoir

34°20'42"/-110°4'53"

7

24, 28, 30, 34,41-43

5

AZ: Navajo Co.; Trick Tank Draw

34°33'437-110°46T3"

13

25, 35-40, 44-49

6

AZ: Yavapai Co.; Granite Dells Ranch

34°36'557-l 12°23'44"

1

1

7

AZ: Yavapai Co.; Pine Flat

35°0 , 67-112°50'43 M

20

3-22

8

AZ: Yavapai Co.; Sycamore Station

34°23 , 287-112°3T"

1

23

9

AZ: Yavapai Co.; Sayer Spring

34°F07-112°39 , 4"

1

2

Figure 1. Map of Arizona showing nine collecting localities for Neotoma
stephensi. Site numbers correspond to data in Table 1.

4

Occasional Papers, Museum of Texas Tech University

Table 2. Microsatellite loci examined for Neotoma stephensi. Castleberry et al. (2000) tested cross-species
amplification in seven woodrat species for these 12 loci; however, no tests were performed on N. stephensi.
Product length (PL), number of alleles (A), and sample size (N) for each locus for N. magister (Castleberry
et al. 2000) and N. stephensi (this study) are shown. Loci NmaOl, Nma02, Nma03, Nma08, and Nmal2
were removedfrom this study due to amplification difficulties.

Locus

N. magister

N. stephensi

PL

A

N

PL

A

N

NmaOl

314—322

6

28

NA

NA

NA

Nma02

197-205

4

33

NA

NA

NA

Nma03

180

1

12

NA

NA

NA

Nma04

145—163

7

33

130—168

12

49

Nma05

227—232

4

38

204—231

8

45

Nma06

215—223

5

39

202—281

24

47

Nma08

125—125

7

38

NA

NA

NA

NmalO

186—224

14

39

248—284

17

44

Nmall

150—160

8

8

144—207

27

49

Nmal2

115—127

3

3

NA

NA

NA

Nmal4

144—160

7

7

135—203

16

49

Nmal5

120—136

10

10

105—139

16

48

pi 10 mM dNTPs, and 0.25 pi 5U/pl Taq. The thermal
profile was modified from Castleberry et al. (2000) and
consisted of a denaturation and enzyme activation cycle
at 94°C (2 min); 35 cycles of 94°C (30 s) denaturation,
55 to 58°C (30 s) annealing, 72°C (1 min) elongation;
followed by a final elongation cycle at 72°C (10 min).

Variation at individual loci was examined using
a 3100 -Avant Genetic Analyzer (Applied Biosystems
Inc., Foster City, California). Reactions included 13.5
to 14 pi Hi-Di Formamide (Applied Biosystems Inc.),
0.5 pi 400HD ROX size standard (Applied Biosystems
Inc.), and 0.5 to 1 pi PCR product. Genotypes were
scored using GeneMapper version 3.0 software (Ap¬
plied Biosystems Inc.). Alleles that did not amplify
above a predetermined peak height (signal strength),
were difficult to score, or appeared aberrant were
reamplified and rescored.

Statistical analyses. —The program Cervus 3.0.3
(Marshall et al. 1998) was used to compare alleles
to bin files generated from GeneMapper software to
identify typing errors that may have occurred during
data entry. Micro-Checker version 2.2.1 software (Van
Oosterhout et al. 2004) was used to test for presence
of null alleles, large allele drop out, and error due to

stutter. A random sample of at least four individuals
per locus was genotyped twice without knowledge of
previous scores. Using these samples, an error rate
was calculated by dividing the number of erroneous
allele scores at each locus by the total number of al¬
lele scores for all individuals for which at least two
genotypes existed.

Structure version 2.2 software (Pritchard et al.
2000) was used to estimate the number of populations
represented by collection localities. This was done in
two ways. First, all individuals were grouped without
prior information of collection site. This analysis was
referred to as “no priors.” Parameters used to determine
number of populations were: burn-in length = 90,000;
Monte Carlo Markov Chains (MCMC) repetitions after
the burn-in = 900,000; ancestry model = admixture;
allele frequency model = allele frequencies correlated;
k = 9 (k is the number of potential populations tested
which, in this study, represented the number of col¬
lection localities); and iterations for each population
test = 5. An individual was assigned to a population
if it had at least an 80% posterior probability of being
included in that population. The second approach used
“prior knowledge” from the above analysis. Individu¬
als were sorted into groups using results from the first

Haynie et al.—Genetic Variation in Neotoma stephensi

5

analysis and population assignment was analyzed.
Parameters for the population assignment test were:
bum-in length = 90,000; MCMC repetitions after the
bum-in = 900,000; ancestry model = prior population
information; allele frequency model = allele frequen¬
cies correlated; and G = 2. The G value estimates the
probability of each individual having an ancestor that
immigrated from another population. An individual
was considered to be assigned correctly if it had at least
an 80% posterior probability of being included in the
population to which it originally was grouped based
on geographic locality.

The program Cervus 3.0.3 (Marshall et al. 1998)
was used to estimate allele frequencies, observed and
expected heterozygosities, null allele frequencies, and
polymorphic information content (PIC; index of vari¬
ability associated with expected heterozygosity). Prob¬
ability of identity (PI) was estimated with IDENTITY
1.0 software (Wagner and Sefc 1999), using equations
reported by Paetkau et al. (1995). This program also
was used to identify identical genotypes among samples
and indicate parent-offspring combinations. Pairwise
and mean relatedness values were estimated with Relat¬
edness 5.0 software (Queller and Goodnight 1989).

The program Fstat 2.9.3.2 (Goudet 2001) was
used to estimate deviations from Hardy-Weinberg
equilibrium (HWE), linkage disequilibrium, F-statistics
(Weir and Cockerham 1984), R ST (Slatkin 1995; Rous-
set 1996; Goodman 1997), pairwise tests of differentia¬
tion, and relatedness values (Hamilton 1971; Queller
and Goodnight 1989). Deviations from HWE were
estimated by evaluating whether F IS values within each
sample were significantly different from zero. Tests of
disequilibrium were performed between all pairs of loci
over all samples and between all pairs of loci within
each sample. The G-statistic, used for pairwise tests
of differentiation, was corrected for comparisons over
multiple loci, unlike traditional F-statistics. Hamilton’s
(1971) relatedness was estimated using an equation that
is comparable to Queller and Goodnight (1989). This
estimator compares average relatedness within groups
compared to all other samples. Sequential Bonferroni
corrections (Holm 1979; Rice 1989) were performed on
all analyses. The indicative adjusted nominal level was
set at 5%, following traditional tests for significance at
the 95% level. For all tests, 1,000 permutations were
performed.

Results

Five loci were removed from this study due to an
inability to amplify samples (Table 2). The remaining
seven loci (Table 2) were used for all further analyses.
No data entry errors or genotype scoring errors were
detected for any locus. Additionally, no evidence for
scoring error due to stutter or large allele drop was
detected at any locus using Micro-Checker software.
The putative presence of null alleles was found at all
loci except NmalO, Nmall, and Nmal4.

Using the program Structure and assuming no
knowledge of geographic localities, two groups were
detected (Fig. 2). Group I consisted of 20 individu¬
als (4, 7, 8, 19, 22, 26-32, 36, 37, 39, 40, 44, 45, 47,
and 49; Fig. 2). Six of these individuals had posterior
probabilities <0.800 (range = 0.725-0.790), but were
assigned to group I as they consistently clustered with
this group. Group II consisted of 19 individuals (2,3,5,
9-12,15,17,18,20,21,23,24, 33,41-43, and 46; Fig.

2). Five of these individuals had posterior probabilities
<0.800 (range = 0.742-0.793), but were assigned to
group II as they consistently clustered with this group.
Ten individuals (1,6,13,14,16,25,34,35,38, and 48;
Fig. 2) could not be assigned to either group with any
confidence. When “prior knowledge” was considered
(i.e. samples were assigned as a member of group I or
group II, or were left unassigned, based on results for
the “no priors” analysis) the same results were obtained
(data not shown).

Mean number of alleles was 8.29 within group
I, 13.43 within group II, and 17.1 over all samples
(Table 3). Allele frequencies and null allele frequen¬
cies are reported by Haynie (2006) or are available
from the senior author upon request. Mean observed
heterozygosity was 0.687 in group 1,0.753 in group II,
and 0.709 over all samples (Table 3). Mean expected
heterozygosity was 0.769 in group 1,0.887 in group II,

6

Occasional Papers, Museum of Texas Tech University

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49

2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48

Figure 2. Bar graph of Structure analysis showing placement of all 49 samples collected in this study in either group
I or group II. Dark gray bars represent group I and light gray bars represent group II. Ten individuals (1, 6, 13, 14,
16, 25, 34, 35, 38, 48) could not be placed in either group with confidence. A list of corresponding sites for each
individual can be found in Table 1.

Table 3. Summary statistics for Neotoma stephensi for each group, site, and samples combined. Allele frequencies
are reported by Haynie (2006) or are available from the senior author upon request. Site numbers correspond to
localities in Fig. 1 and Table 1. Four localities (Sites 1, 6, 8, 9) contained a single individual and were not included
in any population level analyses. Number of individuals (N), mean number of alleles (A), mean observed (H ( j and
expected (HJ heterozygosity, F fS , HWE P-values over all loci (P), mean polymorphic information content (PIC), and
mean relatedness (R) values are shown below. Bonferroni corrected a for tests of HWE was 0.004for comparisons
among groups and 0.001 for comparisons among sites.

Group/Site

N

A

H 0

H F

F, s

P

PIC

R

Group I

20

8.3

0.687

0.769

0.174

0.004

0.713

0.004

Group II

19

13.4

0.753

0.887

0.110

0.011

0.849

0.108

2

2

2.3

0.571

0.619

0.143

0.405

0.366

NA

3

3

4.9

0.714

0.924

0.268

0.005

0.732

NA

4

7

7.4

0.735

0.811

0.101

0.041

0.728

NA

5

13

8.1

0.723

0.765

0.058

0.135

0.706

NA

7

20

12.0

0.695

0.840

0.177

0.001

0.796

NA

Total

49

17.1

0.709

0.854

NA

NA

0.827

-0.022

and 0.854 over all samples (Table 3). PIC was 0.713 in
group I, 0.849 in group II, and 0.827 over all samples
(Table 3). No identical genotypes were detected at
any site. Additionally, no parent-offspring groupings
were found. PI was 1.3e' 9 (1 chance in 800 million
of randomly selecting two individuals with the same
genotype). Mean relatedness values, estimated using
Relatedness software, were 0.004 in group I, 0.108 in
group II, and 0.067 over all samples (Table 3). Related¬
ness, estimated using FSTAT software, within groups
compared to all other samples, with 95% confidence
intervals in parentheses, was 0.127 (0.069, 0.194).

F IS was 0.174 within group I and 0.110 within
group II. Both group I (Bonferroni corrected a = 0.004,
P = 0.004) and group II (Bonferroni corrected a = 0.004,
P = 0.011) were in HWE (Table 3). When tests of geno¬
typic disequilibrium were performed over all samples,

no evidence of genotypic disequilibrium was detected
for any locus among regions (Bonferroni corrected a =
0.002, P > 0.107 for all pairwise comparisons). When
comparisons were made between loci within groups,
no evidence for genotypic disequilibrium was detected
within either group (Bonferroni corrected a = 0.001,
P > 0.169 for all pairwise comparisons within group I,
P > 0.104 for all pairwise comparisons within group
II, and P > 0.105 for all pairwise comparisons among
loci over all groups).

F it , F st , and F IS between the two groups, with 95%
confidence intervals in parentheses, were as follows:
F it = 0.210 (0.096, 0.357), F ST = 0.077 (0.039, 0.129),
and Fjg = 0.144 (0.050,0.268). Pairwise differentiation
values (i.e. G-statistics), estimated using the program
FSTAT, indicated that the two groups were significantly
different from one another (Bonferroni corrected a =

Haynie et al.—Genetic Variation in Neotoma stephensi

7

0.050; P = 0.050). Three estimators of R ST were esti¬
mated among regions: weighted = 0.193, Goodman =
0.149, and unweighted = 0.139.

Population genetic variables also were estimated
for each site. Sites 1, 6, 8, and 9 contained a single
individual (Table 1) and were not considered in any
site-level analyses. Mean number of alleles within
sites ranged from 2.3 (Site 2) to 12.0 (Site 7) (Table 3).
Allele frequencies and null allele frequencies for indi¬
vidual localities are reported by Haynie (2006) or are
available from the senior author upon request. Mean
observed heterozygosity ranged from 0.571 (Site 2) to
0.735 (Site 4), mean expected heterozygosity ranged
from 0.619 (Site 2) to 0.924 (Site 3), and mean PIC val¬
ues ranged from 0.366 (Site 2) to 0.796 (Site 7) (Table
3). Mean relatedness and pairwise relatedness values
were not estimated for sites. However, relatedness,
estimated using FSTAT software, within sites compared
to all other samples, with 95% confidence intervals in
parentheses, was 0.096 (0.057, 0.136).

F IS ranged from 0.058 (Site 5) to 0.268 (Site
3) within sites (Table 3). Across all loci, sites were

in HWE (Table 3), although Site 7 showed marginal
significance (Bonferroni corrected a = 0.001, P =
0.001). When tests of genotypic disequilibrium were
performed over all samples, no evidence of genotypic
disequilibrium was detected for any locus among sites
(Bonferroni corrected a = 0.002, P > 0.038 for all
pairwise comparisons). When comparisons were made
between loci within sites, no evidence for genotypic
disequilibrium was detected within any site (Bonferroni
corrected a = 0.0003, P > 0.026 for all pairwise com¬
parisons among loci over all sites). F-statistic values
among sites, with 95% confidence intervals in paren¬
theses, were as follows: F IT = 0.184 (0.061,0.349), F ST
= 0.057 (0.031, 0.091), and F [S = 0.134 (0.029,0.283).
Pairwise differentiation values (i.e. G-statistics), esti¬
mated using the program FSTAT, indicated that 2 sets
of sites were significantly different, Site 4 and Site 7,
and Site 5 and Site 7 (Bonferroni corrected a = 0.001;
P = 0.001 for both comparisons). Three estimators of
R st were calculated among sites: weighted = 0.012,
Goodman = 0.024, and unweighted = 0.021.

Discussion

Population assignment .—Structure analyses
estimated two distinct groups within the N. stephensi
samples. There appeared to be a general north-south
break among groups; however, individuals collected at
the same sample site often were assigned to different
groups (see Fig. 2 and corresponding information in
Table 1). Additionally, there appeared to be no recog¬
nizable geographic or ecological barrier between the
two groups. Currently, there are two described subspe¬
cies within N. stephensi (Hoffmeister 1986). The split
between these two subspecies is located in the north¬
eastern part of the state. Based on information provided
by Hoffmeister (1986), Site 1 (Fig. 1) should contain
a different subspecies compared to the rest of the col¬
lection sites. Results of this study indicate that the line
between the two subspecies as currently drawn may not
be accurate. The two groups detected using Structure
software may correspond to two different subspecies,
although there are not sufficient samples to address this
possibility. Additional studies need to be performed to
reassess the demarcation of the subspecies.

All sites and groups were in HWE, although Site
7 did show marginal significance (Table 3). Several
localities (Sites 1, 6, 8, and 9) contained a single in¬
dividual (Table 1) and were not useful for population
level analyses. Collection of additional samples may
aid in defining population boundaries.

Genetic structure .—The positive F value esti¬
mated over all sites suggested heterozygote deficiency
within sites. Comparison of the F gT value among groups
(0.077) to guidelines provided by Wright (1978) indi¬
cated moderate genetic differentiation between groups.
The same was true for differentiation among sites (F §T =
0.057). Based on examination of pairwise differentia¬
tion values that were corrected for comparisons over
multiple loci, Site 4-Site 7 and Site 5-Site 7 showed
significant levels of differentiation (Bonferroni cor¬
rected a = 0.001;P = 0.001). Overall, there was little
genetic structure among the samples. Lack of structure
within the samples could be the result of small sample
sizes available for most of the sites. Only sites 5 and

8

Occasional Papers, Museum of Texas Tech University

7 contained >10 individuals. All other sites contained
<7 individuals.

Genetic variation. —Estimators of variability
(H 0 , H e , and PIC) suggested moderate variation within
groups and collection sites (Table 3), with the exception
of site 2. Site 2 contained two samples, both of which
were homozygous at three of seven loci, thus reducing
heterozygosity and PIC values. However, each site
had a unique allele for at least one locus. Addition¬
ally, no identical genotypes existed among samples.
PI, which was estimated over all samples, suggested
that these loci were highly variable and that individual
genotypes were unique. Small samples sizes for most
of the populations were reflected in the moderate levels
of variation.

Relatedness. —-No family groups were detected
within any site, using Identity software. Relatedness
within groups compared to all other samples was 0.127,
indicative of a third-order (e.g. cousin) relationship
among individuals. Relatedness within sites compared
to all other samples was 0.096, indicative of a fourth-
order (e.g. distant cousin) relationship. Relatedness
over all samples, as estimated using Relatedness soft¬
ware, was 0.067, again indicative of a fourth-order
relationship among all samples. Relatedness within
group I was 0.004, indicative of a very low level of
relatedness among samples, and relatedness within

group II was 0.108, indicative of a third-order rela¬
tionship among samples. These results suggest that
individuals within groups and sites are more closely
related to other individuals within the group or site
than to individuals outside the group or site. Pairwise
comparisons ranged from negative to highly positive,
indicating that some individuals were closely related
to others. Interestingly, the pairwise relatedness value
between the sample from site 8 and the sample from
site 9 was 0.562 (indicative of a parent-offspring or full¬
sibling relationship), despite the geographic distance
between the two sites (-100 km). This was indicative
of the potential for dispersal and gene flow between
these localities. The relatedness values estimated for
N. stephensi were similar to what has been recorded
for other species of Neotoma (Matocq and Lacey 2004;
Haynie et al. 2007). Relatedness values in this study
may reflect small sample sizes and sampling strategy.

Comparison to other Neotoma species. —Despite
low to moderate levels of genetic variation, based on
loci characterized by Castleberry et al. (2000), within
these sites, number of alleles, expected heterozygosity
values, and PIC values present in this study fell within
the range reported for other species of Neotoma (Table
4). For example, number of alleles ranged from eight
(Nma05) to 27 (Nmall) with a mean of 17.1 over all
sites for this study (Table 2). Number of alleles for
N. magister (Castleberry et al. 2000) ranged from one

Table 4. Comparison of genetic variation in Neotoma stephensi to five other species o/Neotoma. N = number of
specimens examined, NP = number ofpopulations examined, NL = number of loci examined, AR = range in number
of alleles, MA = mean number of alleles, H E = expected heterozygosity, PIC = polymorphic information content, and
F st represents the value among populations. NA indicates that no data was available for that study. N. macrotis-7
andN. fuscipes-7 data are from Haynie et al. (2007), N. magister data are from Castleberry et al. (2002), N. floridana
data are from Monty et al. (2003), N. micropus data are from Mendez-Harclerode et al. (2007), N. macrotis-2 data are
from Matocq (2004), and N. fuscipes-2 data are from Matocq (2002). All studies used loci developed by Castleberry
et al. (2000) except Matocq (2002, 2004).

Species

N

NP

NL

AR

MA

H b

PIC

F S t

N. stephensi

49

7

7

8-27

17.1

0.854

0.827

0.052

N. macrotis-X

127

4

5

21-36 25.6

0.934

0.930

0.028

N. fuscipes-X

29

5

5

9-21

13.4

0.804

0.761

0.249

N. magister

357

9

11

5-19

10.4

0.618

NA

0.170

N. floridana

84

5

6

1-9

2.8

0.365

0.666

0.522

N. micropus

549

1

5

16-47

26.0

0.845

0.829

NA

N. macrotis-2

195

1

5

11-19

15.0

0.850

0.840

NA

N. fuscipes-2

81

8

3

16-21

18.3

NA

NA

NA

Haynie et al.—Genetic Variation in Neotom a stephensi

9

(Nma03) to 14 (NmalO) with a mean of 6.3 over 12 loci,
and from four (Nma05) to 14 (NmalO) with a mean of
7.9 over the same seven loci used in this study (Table
2). In a more comprehensive study, Castleberry et al.
(2002) found number of alleles to range from five to
19 with a mean of 10.4 over 11 loci from a sample of
357 TV. magister (Table 4). In a study which examined
five loci in a single population of TV. micropus , number
of alleles ranged from 16 to 47 with a mean of 26.0
(Mendez-Harclerode et al. 2007). The large number
of alleles in this population could be attributed to the
large sample size (n = 549) and indicated a high degree
of genetic variability within this population.

Mean heterozygosity values ranged from 0.365
in five populations of TV. floridana (Monty et al. 2003)
to 0.934 in four populations of TV. macrotis (Haynie et
al. 2007) among the species of Neotoma (Table 4). The
mean heterozygosity value for TV. stephensi (0.854) fell
well within this range. The same was true for mean
PIC values which ranged from 0.666 in TV. floridana
(Monty et al. 2003) to 0.930 in N. macrotis (Haynie
et al. 2007; Table 4), with the value for TV. stephensi
(0.827) falling within this range.

Comparing levels of genetic differentiation (F ST )
among populations among the species of Neotoma ,
the value reported for TV. stephensi (0.057) was lowest
among the species (Table 4). This F ST value indicated
moderate levels of genetic differentiation among sites
in this study, even though sites were widely distrib¬
uted. Lack of structure may have been an artifact of
small sample sizes. F ST values suggested moderate
levels of genetic differentiation among populations of
TV. macrotis and high levels among populations of TV.
fuscipes (Table 4; Haynie et al. 2007). High levels of
genetic differentiation were found among populations
of TV. magister (Castleberry et al. 2002) collected from
a large portion of this species range (Table 4). Finally,
Monty et al. (2003) reported great degrees of genetic
differentiation among five populations of TV. floridana ,
despite studying a limited portion of this species range
(Table 4). Even though TV. stephensi is a dietary and
habitat specialist with a restricted range, levels of
genetic diversity within sites examined in this study
were comparable to levels found in other species of
Neotoma.

Acknowledgments

This project was financially supported by Na¬
tional Institutes of Health grant AI-41435 (“Ecology
of emerging arenaviruses in the southwestern United
States”). The authors thank Kenneth D. Abbott and
Josh Keith (Yavapai College, Prescott, Arizona) for

sample collection and producing the map; Ronald A.
Van Den Bussche and Stephanie Davis (Oklahoma
State University, Stillwater, Oklahoma) for insights
into data analysis; and Richard E. Strauss (Texas Tech
University) for his help and insight.

Literature Cited

Abbott, K. D., M. L. Milazzo, J. Keith, R. D. Bradley, and
C. F. Fulhorst. 2004. Epizootiology of arenaviral
infections in the white-throated woodrat (Muridae:
Sigmondontinae) and other woodrats in Arizona.
Journal of Vector Ecology 29:355-364.

Boyle, R., and M. D. Dearing. 2003. Ingestion of juniper fo¬
liage reduces metabolic rates in woodrat ( Neotoma )
herbivores. Zoology 106:151-158.

Carraway, L. N., and B. J. Verts. 1991. Neotoma fuscipes.
Mammalian Species 386:1-10.

Castleberry, S. B., T. L. King, P. B. Wood, and W. M. Ford.
2000. Microsatellite DNA markers for the study
of Allegheny woodrat (Neotoma magister) popula¬
tions and cross-species amplification in the genus
Neotoma. Molecular Ecology 9:817-829.

Castleberry, S. B., T. L. King, P. B. Wood, and W. M. Ford.
2002. Microsatellite DNA analysis of population
structure in Allegheny woodrats {Neotoma magis¬
ter). Journal of Mammalogy 83:1058-1070.

10

Occasional Papers, Museum of Texas Tech University

Dearing, M. D., A. M. Mangione, andW. H. Karasov. 2000.
Diet breadth of mammalian herbivores: nutri¬
ent versus detoxification constraints. Oecologia
123:397-405.

Dearing, M. D., A. M. Mangione, and W. H. Karasov.

2001. Plant secondary compounds as diuretics:
an overlooked consequence. American Zoology
41:890-901.

Dearing, M. D., A. M. Mangione, and W. H. Karasov.

2002. Ingestion of plant secondary compounds
causes diuresis in desert herbivores. Oecologia
130:576-584.

Dial, K. P. 1988. Three sympatric species of Neotoma.
dietary specialization and co-existence. Oecologia
76:531-537.

Edwards, C. W., and R. D. Bradley. 2002. Molecular
systematics of the genus Neotoma. Molecular
Phylogenetics and Evolution 25:489-500.

Goldman, E. A. 1910. Revision of the wood rats of the genus
Neotoma. North American Fauna 31:1 -124.

Goodman, S. J. 1997. Rst Calc: a collection of computer
programs for calculating estimates of genetic dif-
ferentition from microsatellite data and determining
their significance. Molecular Ecology 6:881-885.

Goudet, J. 2001. FSTAT, a program to estimate and test
gene diversities and fixation indices (version 2.9.3).
Available from http://www.unil.ch/izea/softwares/
fstat.html.

Hall, E. R. 1981. The mammals of North America, second
edition. Wiley, New York.

Hamilton, W. D. 1971. Selection of selfish and altruistic be¬
havior in some extreme models. In Man and Beast:
Comparative Social Behavior. (J. F. Eisenberg, and
W. S. Dillon, eds.) Smithsonian Institution Press,
Washington, D.C.

Haynie, M. L. 2006. Population genetics of four species
of Neotoma from the southwestern United States.
Ph.D. dissertation, Texas Tech University, Lub¬
bock.

Haynie, M. L., S. G. Bennett, M. Rood, B. Hess, C. F.
Fulhorst, and R. D. Bradley. 2007. Genetic varia¬
tion in multilocus microsatellite genotypes in two
species of woodrats {Neotoma macrotis and N.
fuscipes) from California. Journal of Mammalogy
88:745-758.

Hoffmeister, D. F. 1986. Mammals of Arizona. The Uni¬
versity of Arizona Press and Arizona Game and
Fish Department, Arizona.

Hoffmeister, D. F., and L. de la Torre. 1960. A revision of
the wood rat Neotoma stephensi. Journal of Mam¬
malogy 41:476-491.

Holm, S. 1979. A simple sequentially rejective multiple
test procedure. Scandinavian Journal of Statistics
6:56-70.

Hooper, E. T. 1960. The glans penis in Neotoma (Roden-
tia) and allied genera. Occasional Papers of the
Museum of Zoology, University of Michigan
618:1-20.

Jones, C., and N. J. Hildreth. 1989. Neotoma stephensi.
Mammalian Species 328:1-3.

Kosoy, M. Y„ L. H. Elliott, T. G. Ksiazek, C. F. Fulhorst, P. E.
Rollin, J. E. Childs, J. N. Mills, G. O. Maupin, and
C. J. Peters. 1996. Prevalence of antibodies to are¬
naviruses in rodents from the southern and western
United States: evidence for an arenavirus associated
with the genus Neotoma. American Journal of
Tropical Medicine and Hygiene 54:570-576.

Matocq, M. D. 2002. Morphological and molecular analysis
of a contact zone in the Neotoma fuscipes species
complex. Journal of Mammalogy 83:866-883.

Matocq, M. D. 2004. Reproductive success and effective
population size in woodrats {Neotoma macrotis ).
Molecular Ecology 13:1635-1642.

Matocq, M. D., and E. A. Lacey. 2004. Philopatry, kin
cluster, and genetic relatedness in a population of
woodrats {Neotoma macrotis ). Behavioral Ecology
15:647-653.

Marshall, T. C., J. Slate, L. E. B. Kruuk, and J. M. Pemberton.
1998. Statistical confidence for likelihood-based
paternity inference in natural populations. Molecu¬
lar Ecology 7:639-655.

Mendez-Harclerode, F. M., J. D. Hanson, C. F. Fulhorst, M.
L. Milazzo, D C. Ruthven III, and R. D. Bradley.
2005. Genetic diversity within the Southern Plains
woodrat {Neotoma micropus ) in south Texas. Jour¬
nal of Mammalogy 86:180-190.

Mendez-Harclerode, F. M., R. E. Strauss, C. F. Fulhorst, M.
L. Milazzo, D. C. Ruthven III, and R. D. Brad¬
ley. 2007. Molecular evidence for high levels
of intrapopulation genetic diversity in woodrats
{Neotoma micropus). Journal of Mammalogy
88:360-370.

Monty, A-M., E. J. Heist, E. R. Wagle, R. E. Emerson, E. H.
Nicholson, and G. A. Feldhamer. 2003. Genetic
variation and population assessment of eastern
woodrats in southern Illinois. Southeastern Natu¬
ralist 2:243-260.

Haynie et al.—Genetic Variation in Neotoma stephensi

11

Murray, K. F., and A. M. Barnes. 1969. Distribution and
habitat of the woodrat, Neotoma fuscipes, in
northeastern California. Journal of Mammalogy
50:43-48.

Paetkau, D., W. Calvert, I. Stirling, and C. Strobeck. 1995.
Microsatellite analysis of population structure in
Canadian polar bears. Molecular Ecology 4:347-
354.

Planz, J. V., E. G. Zimmerman, T. A. Spradling, and D. R.
Akins. 1996. Molecular phylogeny of the Neotoma
floridana species group. Journal of Mammalogy
77:519-535.

Pritchard, J. K., M. Stephens, and P. Donnelly. 2000. In¬
ference of population structure using multilocus
genotype data. Genetics 155:945-959.

Queller, D. C., and K. F. Goodnight. 1989. Estimating
relatedness using genetic markers. Evolution
43:258-275.

Rice,W. R. 1989. Analysing tables of statistical tests. Evo¬
lution 43:223-225.

Rousset, F. 1996. Equilibrium values of measures of popula¬
tion subdivision for stepwise mutation processes.
Genetics 142:1357-1362.

Slatkin, M. 1995. A measure of population subdivision
based on microsatellite allele frequency. Genetics
139:457-462.

Sorensen, J. S., and M. D. Dearing. 2003. Elimination of
plant toxins by herbivorous woodrats: revisiting
an explanation for dietary specialization in mam¬
malian herbivores. Oecologia 134:88-94.

Addresses of authors:

Michelle L. Haynie

University of Central Oklahoma
Department of Biology
100 N University Dr
Edmond, OK 73034 USA
mhaynie@uco. edu

Robert D. Bradley

Department of Biological Sciences and
Natural Science Research Laboratory, The Museum
Texas Tech University
Lubbock, TX 79409-3131 USA
robert. bradley@ttu. edu

Sorensen, J. S., and M. D. Dearing. 2004. Physiological
limitations of dietary specialization in herbivorous
woodrats (Neotoma spp.). International Congress
Series 1275:313-320.

Sorensen, J. S., C. A. Turnbull, and M. D. Dearing. 2004. A
specialist herbivore (Neotoma stephensi ) absorbs
fewer plant toxins than does a generalist (Neotoma
albigula). Physiological and Biochemical Zoology
77:139-148.

Tietje, W. D., and J. K. Vreeland. 1997. Vertebrates diverse
and abundant in well-structured oak woodland.
California Agriculture 51:8-14.

Van Oosterhout, C., W. F. Hutchinson, D. P. M. Willis, and
P. Shipley. 2004. Micro-Checker: software for
identifying and correcting genotyping errors in
microsatellite data. Molecular Ecology Notes
4:535-538.

Vaughan, T. A. 1982. Stephen’s woodrat, a dietary specialist.
Journal of Mammalogy 63:53-62.

Wagner, H.W., and K.M.Sefc. 1999. IDENTITY 1.0. Cen¬
tre for Applied Genetics, University of Agricultural
Sciences, Vienna.

Weir, B. S., and C. C. Cockerham. 1984. Estimating F-
statistics for the analysis of population structure.
Evolution 38:1358-1370.

Wright, S. 1978. Evolution and the genetics of populations.
Volume 4. Variability within and among natural
populations. University of Chicago Press, Chicago.

Charles F. Fulhorst

Department of Pathology
University of Texas Medical Branch
Galveston, TX 77555-0609 USA
cfulhors@utmb. edu

Publications of the Museum of Texas Tech University

Institutional subscriptions are available through the Museum of Texas Tech University, attn: NSRL
Publications Secretary, Box 43191, Lubbock, TX 79409-3191. Individuals may also purchase separate numbers
of the Occasional Papers directly from the Museum of Texas Tech University.

ISSN 0149-175X

Museum of Texas Tech University, Lubbock, TX 79409-3191