cratogeomys castanops species group

Journal of Mammalogy, 89(1):190–208, 2008
EVOLUTIONARY RELATIONSHIPS OF POCKET GOPHERS
(CRATOGEOMYS CASTANOPS SPECIES GROUP) OF THE
MEXICAN ALTIPLANO
DAVID J. HAFNER,* MARK S. HAFNER, GERALD L. HASTY, THERESA A. SPRADLING,
AND JAMES
W. DEMASTES
New Mexico Museum of Natural History, Albuquerque, NM 87104, USA (DJH)
Department of Biological Sciences and Museum of Natural Science, Louisiana State University, Baton Rouge,
LA 70803, USA (MSH, GLH)
Department of Biology, University of Northern Iowa, Cedar Falls, IA 50614, USA (TAS, JWD)
The Southern Coahuila Filter-Barrier (SCFB) effectively subdivides the mammalian fauna of the Mesa del Norte,
the northern and most extensive section of the Mexican Altiplano. Pocket gophers of the genus Cratogeomys
north and south of this filter-barrier have been informally recognized as 2 distinct species, C. castanops and
C. goldmani, respectively. Support for species recognition derives from early morphological comparisons and
recent chromosomal and ectoparasite studies. Contradictory conclusions based on the only comprehensive morphometric study have prevented formal recognition of C. goldmani. A morphometric reevaluation based on ratiotransformed data reveals that the previous analysis was unduly biased by size, an ecophenotypically plastic
character. When this factor is removed, morphometric variation is fully concordant with chromosomal diploid
number and sequence data of mitochondrial and nuclear DNA. We provide synonymies and descriptions for
C. goldmani apart from C. castanops, and revise the number of subspecies from a total of 26 to 2 subspecies in each
species. The SCFB is most effective in its central portion (Desierto Mayrán), and least effective in its western
portion (Rı́o Nazas), which should be geographically broadened to include the neighboring Rı́o Aguanaval.
Key words:
systematics
chromosomes, Cratogeomys, filter-barrier, mitochondrial DNA, morphology, nuclear DNA, pocket gophers,
the castanops species group. Species of the C. fumosus species
group occupy deep soils of pine–oak woodlands of the western
Trans-Mexican Volcanic Belt and the periphery of the Mesa
Central, whereas species of the castanops species group, with
the exception of C. castanops itself, occur in and around the
arid Oriental Basin (Cuenca Oriental) of Puebla and surrounding states, extending up into the high volcanoes of the region.
Reflecting its probable origin in the aridlands of central Mexico
(Hafner et al. 2005), C. castanops is distributed in the highelevation, desert grasslands of the Chihuahuan Desert and the
high plains and plateaus near the western edge of the Great
Plains.
The yellow-faced pocket gopher (Cratogeomys castanops)
has the largest and northernmost distribution of all species in
the genus, extending northward from the southern end of the
Mesa del Norte (¼ Central Plateau or Mexican Plateau) on the
Mexican Altiplano (228N latitude, 2,100 m elevation, 4,454 m
equator-equivalent—Harris 1985) into southeastern Colorado
and southwestern Kansas (388N latitude, 750 m elevation,
4,816 m equator-equivalent; Fig. 1). The other 5 species in the
genus are arrayed along the Trans-Mexican Volcanic Belt of
central Mexico, from Colima to Veracruz (Hafner et al. 2004,
2005). Recent revisions of Cratogeomys of the Trans-Mexican
Volcanic Belt have recognized 2 species (fumosus and
planiceps) of the fumosus species group (formerly divided into
5 species and recognized as the gymnurus species group
[Hafner et al. 2004]) and 3 cryptic species within C. merriami
(fulvescens, merriami, and perotensis [Hafner et al. 2005]) of
TAXONOMIC HISTORY OF CRATOGEOMYS CASTANOPS
Although C. castanops is evolutionarily rooted in the south
(Hafner et al. 2005), description of its distribution and
taxonomy proceeded from its initial description in the north
(Pseudostoma castanops Baird, 1852), extended south across
the Rı́o Grande (Geomys clarkii Baird, 1855) and the Southern
Coahuila Filter-Barrier (SCFB—Baker 1956 [C. c. goldmani
Merriam, 1895]), and finally into the southern Mesa del Norte
(C. c. peridoneus Nelson and Goldman, 1934; Fig. 1). Baker
* Correspondent: [email protected]
Ó 2008 American Society of Mammalogists
www.mammalogy.org
190
February 2008
HAFNER ET AL.—SYSTEMATICS OF CRATOGEOMYS
FIG. 1.—Geographic distribution of Cratogeomys castanops indicating subspecies groups defined by Russell (1968): subnubilus (light
gray) and excelsus (darker gray). Darkest gray indicates 3 areas where
Russell (1968) believed the 2 subspecies groups were in sympatry
and acting as full species, as opposed to other sites of purported
introgression. Dashed lines indicate state boundaries, and stitched
lines indicate subspecies boundaries; 2 broadly overlapping subspecies
(surculus and goldmani) are identified. Localities mentioned in text: 1)
Pine Springs Canyon, Texas; 2) Ojinaga Valley, Chihuahua; 3) Rodeo,
Durango; 4) Torreón, Coahuila; 5) La Flor de Jimulco, Coahuila; 6)
Parras, Coahuila; 7) Cañon Santo Domingo, Coahuila; 8) Montemorelos,
Nuevo León; 9) Hacienda Atotonilco, Durango; 10) Cañitas,
Zacatecas; 11) Villa de Cos, Zacatecas.
(1956) defined the SCFB as formed by the Rı́o Nazas on the
west and the Sierra de Parras and Sierra de Guadalupe (which
includes the Sierra de Patagalana and Sierra La Concordia) on
the east (Baker 1956; Baker and Greer 1962; Peterson 1976;
Schmidly 1977). The central region of the SCFB is augmented
on the north by an extensive dry lake (Laguna Mayrán) in the
Mayrán Basin, a terminal basin fed by the Rı́o Nazas and Rı́o
Aguanaval. The Laguna Mayrán has been a playa throughout
the Holocene, although catastrophic floods induced by
hurricane-related precipitation events inundate much of the
Rı́o Nazas alluvial plain, and open water may persist in regions
of the Mayrán Basin for several years. It is likely that water
levels were higher and more constant during pluvial periods of
the Pleistocene (K. Butzer, pers. comm.).
191
Merriam (1895) recognized the 1st specimens of C.
castanops from south of the SCFB (a series of 5 females from
Cañitas, Zacatecas; Fig. 1, locality 10) as a distinct subspecies,
C. c. goldmani. Interestingly, Merriam (1895:151) mistakenly
included goldmani in his ‘‘key to species of Cratogeomys,’’
perhaps emphasizing the distinctiveness of the subspecies.
Based on new material gathered over the ensuing 40 years,
Nelson and Goldman (1934) added 13 new subspecies of C.
castanops, including 4 subspecies (all from south of the SCFB)
that they allied with C. c. goldmani (still restricted to the type
locality) as opposed to the 2 other existing and 9 new subspecies from north of the SCFB, which they allied with C. c.
castanops. Although not explicitly stating a dichotomy between subspecies north and south of the SCFB, subspecies
descriptions indirectly but clearly conveyed the existence of 2
morphologically distinct geographic groups: castanops (north
of the SCFB) and goldmani (south).
Russell and Baker (1955) examined geographic variation
in C. castanops from Coahuila based on additional material,
named 4 new subspecies in addition to the previously recognized 16 subspecies, and extended the distribution of C. c.
goldmani to include new records from near La Flor de Jimulco
on the Rı́o Aguanaval (south of the SCFB; Fig. 1, locality 5) as
well as a new record from 1.5 miles north of Parras (north of the
SCFB; Fig. 1, locality 6). Subspecies comparisons of Russell
and Baker (1955) appeared to distinguish subspecies generally
south of the SCFB (goldmani, planifrons, and subnubilus) from
those to the immediate north (excelsus and subsimus). They
noted the morphological distinctiveness of C. c. subnubilus
and C. c. planifrons from northern neighbors (Russell and
Baker 1955:606) and the close geographic proximity between
C. c. subnubilus and C. c. subsimus (2 miles—Russell and
Baker 1955:607), but they explained the morphological distinctiveness of C. c. subnubilus and C. c. planifrons as resulting
from ‘‘their isolation in an elevated habitat’’ (Russell and
Baker 1955:595), rather than due to any broad geographic or
evolutionary subdivision.
Hall and Kelson (1959) considered C. c. goldmani to occur
in eastern Durango, southern Coahuila, and across Zacatecas
to western San Luis Potosı́. Baker and Greer (1962) reported
the 1st specimens of Cratogeomys from Durango south of
the Rı́o Nazas, from Hacienda Atotonilco (Fig. 1, locality 9).
They assigned these 2 specimens to C. c. goldmani, and emphasized the role of the Rı́o Nazas in isolating C. c. goldmani
(Hacienda Atotonilco) from C. c. excelsus (Bolsón de Mapimı́
of northeastern Durango).
In his revision of the genus Pappogeomys (including
Cratogeomys), Russell (1968) revised subspecies boundaries,
named 7 new subspecies, subsumed 2 others (C. c. lacrimalis
and C. c. convexus), and recognized 2 distinct morphological
groups within these 25 subspecies: the excelsus group in the
northern and eastern part of the species’ range and the
subnubilus group in the southern and western part of the range
(Fig. 1). Russell (1968:625) demonstrated a strong, positive
correlation between length of palate and condylobasal length
among female C. castanops, with specimens of the subnubilus
group distinctly smaller in both measurements compared to
192
Vol. 89, No. 1
JOURNAL OF MAMMALOGY
specimens of the excelsus group. Russell (1968:623–624)
indicated that the 2 groups formed a ‘‘rassenkreis,’’ acting as
full species in most places where populations met but joined by
interbreeding populations in other places. Russell (1968)
believed that the 2 groups acted as full species at Pine Springs
Canyon (Guadalupe Mountains, Texas; Fig. 1, locality 1), in
the Ojinaga Valley (northeastern Chihuahua; Fig. 1, locality 2),
and at 2 sites along the SCFB: near Rodeo on the Rı́o Nazas
(west end of the SCFB; Fig. 1, locality 3) and near the mouth of
Cañon Santo Domingo in the Sierra de Parras (east end; Fig. 1,
locality 7). At the same time, Russell (1968) believed that
intergradation was demonstrated between the 2 groups in
a narrow zone of contact near Villa de Cos, Zacatecas (Fig. 1,
locality 11). He emphasized size differences between the
larger, northeastern excelsus group (in which he included
goldmani) and the smaller, generally southwestern subnubilus
group. Russell (1968) indicated that 2 of the new subspecies
(parviceps in southern New Mexico and Texas and perexiguus
along the Chihuahuan–Coahuilan border) along with consitus
(in Chihuahua) represented broad contiguity of the 2 groups
throughout the species’ range. He extended the range of
goldmani to include most of eastern Durango, where he
indicated broad sympatry between goldmani (of the excelsus
group) and surculus (of the subnubilus group; Fig. 1). Álvarez
and Álvarez-Castañeda (1996) subsequently named 1 additional subspecies (maculatus) from northern San Luis Potosı́
and eastern Zacatecas based largely on the presence of white
spots on the pelage (a marking frequently observed in this
and other species of pocket gophers—D. J. Hafner and M. S.
Hafner, in litt.). Dalquest (1953:101) previously noted the
presence of white spots in pocket gophers from San Luis
Potosı́, and Russell (1968:685) described in detail the existence
of this pattern in 3 adjacent subspecies.
Berry and Baker (1972) surveyed non–preferentially stained
chromosomal complements throughout the distribution of
C. castanops. They indicated 2 clearly defined geographic
groups that occurred north (2n ¼ 46, FN ¼ 86) and south (2n ¼
42, FN ¼ 78) of the SCFB. Specimens from near Cañitas,
Zacatecas (the type locality for C. c. goldmani), possessed the
southern karyotype, as did 2 specimens from Villa de Cos,
Zacatecas, where Russell (1968) had indicated interbreeding
between his 2 morphological groups. Berry and Baker
(1972:308) chose not to recognize these 2 chromosomal
groups as species because of the ‘‘striking differences between
the interpretation of evolutionary affinities based on gross
morphology [Russell 1968] and karyotypes.’’ They specifically
noted the confusion surrounding C. c. goldmani, stating that
specimens that they karyotyped from Rodeo, Durango, and
from near the type locality of Cañitas, Zacatecas, were smaller
than those assigned to goldmani by Russell (Rodeo) or the
initial type series from Cañitas. It is not clear how comparisons
were made between the 2 sets of specimens from Cañitas;
Berry and Baker (1972:308) stated only that the new specimens
‘‘do not resemble the holotype and topotypes of goldmani.’’
Lee and Baker (1987:3) compared preferentially stained
(G-band and C-band) karyotypes representing both chromosomal forms, and noted that the distribution of parasitic lice
(genus Geomydoecus—Hellenthal and Price 1976) on the
C. castanops complex was ‘‘completely compatible with
patterns of chromosomal races [Berry and Baker 1972], but
not with variation in cranial and body size’’ (Russell 1968).
Lee and Baker (1987:13) concluded that ‘‘it probably is best
to recognize 2 species’’ (C. castanops and C. goldmani or C.
subnubilus), but did not formally recognize C. goldmani as
a distinct species because ‘‘individuals collected in the 1970s
from near the type locality (Cañitas, Zacatecas) . . . differ
morphologically from the type specimen and paratypes of
goldmani.’’ Despite the lack of formal elevation of goldmani
to species status, and inclusion by Russell (1968) of C. c.
goldmani along with C. c. castanops in the excelsus species
group, Davidow-Henry et al. (1989) restricted their description
of C. castanops to vaguely north of 258N latitude, and DeWalt
et al. (1993), Demastes et al. (2002), Álvarez and ÁlvarezCastañeda (1996), Hafner et al. (2004), Hafner et al. (2005),
and Patton (2005) all have recognized C. goldmani as
specifically distinct from C. castanops.
There are 2 contrasting views of relationships within
C. castanops: 2 distinct groups north (castanops) and south
(goldmani) of the SCFB (Berry and Baker 1972; Lee and Baker
1987; Merriam 1895; Nelson and Goldman 1934), and a mosaic
of 2 groups (excelsus and subnubilus) that are sympatric at
multiple sites, interbreeding at some and not at others (Russell
1968). We reexamined relationships among populations of
C. castanops based on cranial morphology, nonpreferentially
stained karyotypes, and mitochondrial (mtDNA) and nuclear
(nuDNA) sequences, with particular attention to relationships
among populations in the vicinity of the SCFB and the 2
subspecies, C. c. goldmani and C. c. surculus, which supposedly
occur sympatrically across the SCFB (Russell 1968; Fig. 1).
MATERIALS AND METHODS
Specimens examined.— We examined a total of 284 specimens of C. castanops from 100 localities (Fig. 2): 199
specimens in the morphometric analysis (Appendix I), 79 in
the chromosomal analyses (Appendix II), 55 in the mtDNA
analyses, and 54 in the nuDNA analyses (Appendix III).
Cytochrome-c oxidase subunit I (CoI) sequence data for 2
specimens in Appendix I were taken from Hafner et al. (2005
[AY331076 and AY331077]), and b-fibrinogen gene (b-fib)
sequence data for 2 specimens in Appendix III were taken from
Spradling et al. (2004 [AY331241 and AY331242]). The
remaining specimens are new to this study and were captured
in the wild using standard trapping methods approved by the
American Society of Mammalogists (Gannon et al. 2007).
Outgroup taxa consisted of specimens of C. fumosus,
C. merriami, and C. planiceps. Sequence data for these species
were obtained from Demastes et al. (2002 [GenBank accession
nos. AF302179 and AF302183]), Spradling et al. (2004
[AY331075, AY331238–AY331240, and AY331243]), Hafner
et al. (2004 [AY545541]), or Hafner et al. (2005 [AF302158
and AY331078]).
Chromosomal analysis.— Non–preferentially stained chromosome preparations were made from 42 individuals from 11
February 2008
HAFNER ET AL.—SYSTEMATICS OF CRATOGEOMYS
FIG. 2.—Collecting localities listed in Appendices I–III. Shading
indicates geographic distribution of Cratogeomys castanops (sensu
Hall [1981], with recent updates); dashed line indicates location of
Southern Coahuila Filter-Barrier (Baker 1956).
localities following the postmortem field protocol described by
Hafner and Sandquist (1989). This sample was augmented by
37 individuals from 16 localities reported by Berry and Baker
(1972) and Lee and Baker (1987), for a total of 79 individuals
from 27 localities (Appendix II; Fig. 3B). Most localities were
represented by 1 or 2 karyotypes. We sampled the purported
zone of contact in Cañon Domingo, Sierra de Guadalupe
(Russell 1968) more extensively (n ¼ 25 karyotypes). Diploid
number (2n) and fundamental number (FN) were determined
for each individual.
Mitochondrial DNA analysis.— Sequence data were obtained
for 55 specimens of C. castanops from 28 collecting localities
(Fig. 3C; Appendix III). Whole DNA extractions (DNeasy
Tissue Kit; Qiagen Inc., Valencia, California) used 20–25 mg
of tissue. Extractions followed the manufacturer’s protocol,
with the 2 final elutions of 100 ll combined for each sample.
Extractions were assayed and sequenced by polymerase chain
reaction (PCR) for 2 mitochondrial loci, CoI (1,551 base pairs
[bp]; n ¼ 33) and cytochrome b (Cytb, 1,140 bp; n ¼ 55).
Typically, PCR protocols consisted of 1 ll of DNA extraction,
2.5 ll of 10 PCR buffer (with 15 mM MgCl2; Applied
Biosystems Group, Foster City, California), 2.5 ll of 8 mM
deoxynucleoside triphosphates, 1 ll each of 10 lM PCR
primers, 0.1 ll of 5 U/ll Taq (Applied Biosystems Group), and
16.9 ll of water. Thermal cycling profiles consisted of an initial
denaturation at 948C for 2–3 min, followed by 40 cycles of
948C for 30–45 s, 45–558C for 45 s, and 728C for 45–60 s.
Positive PCR products (CoI primers [Spradling et al. 2004]:
CoI-5285F, CoI-6929R, CoI-1039F [59-AGC YCT AGG CTT
TAT YTT CC-39], and Gco1R1; Cytb primers: L14724 [Irwin
et al. 1991], H15906 [Spradling et al. 2001], H15154 [primer
193
MVZ04 in Smith and Patton 1993], and L15171 [Spradling
et al. 2001]) were cleaned before sequencing (QIAquick PCR
Purification Kit; Qiagen, Inc.). Cycle sequencing reactions used
BigDye version 3.1 dye-terminator cycle sequencing reagents
(Applied Biosystems Group; Foster City, California) and
typically consisted of 1 ll of cleaned PCR product, 1 ll of
3.3 lM primer (CoI [Spradling et al. 2004]: PCR primers plus
CoI-570 and its reverse complement, Gco1F1, the reverse
complement of Gco1R1, CoI-30; Cytb: PCR primers plus
H15579 [Spradling et al. 2001], the reverse complement of
H15408), 2 ll of 5 reaction buffer, 0.5 ll of BigDye reaction
mixture, and 5.5 ll water. Sequencing products were cleaned
using Sephadex G-50 (Sigma-Aldrich, St. Louis, Missouri) in
Centri-Sep columns (Princeton Separations, Princeton, New
Jersey) or Performa DTR 96 well standard plates (Edge
Biosystems, Gaithersburg, Maryland), and sequence data were
obtained using an ABI 3100 automated sequencing machine
(Applied Biosystems Group) following the manufacturer’s
protocols. Sequences were aligned using Aligner version 1.4.6
(CodonCode Corporation, Dedham, Massachusetts) and reference to sequences available from GenBank for Cratogeomys
CoI (C. castanops AY331076 and AY331077, C. merriami
AY331078, and C. fumosus AY331075) and Cytb (C. fumosus
AF302179, C. merriami AF302158, and C. planiceps
AF302183 and AY545541). All sequenced regions received
at least 2 times coverage, by sequencing in both strand directions or by repeated coverage using different sequencing
primers and reactions. Sequences were submitted to GenBank
(GenBank accession nos. EF607212–EF607277).
Phylogenetic analyses were conducted using PAUP* version
4.0b10 (Swofford 2002) for maximum parsimony and maximum likelihood and MrBayes version 3.1.1 (Ronquist and
Huelsenbeck 2003) for Bayesian analyses. Phylogenetic
congruence of the locus-specific data was tested by the
incongruence length difference test (‘‘homogeneity partition
test’’ of PAUP*) as a prelude to combining the separate locusspecific data into 1 data set. ModelTest version 3.7 (Posada and
Crandall 1998) was used as an aid in selecting nucleotide
substitution models used in phylogenetic analyses. All 26
specimens from the purported zone of contact in Cañon
Domingo, Sierra de Guadalupe, were sequenced for Cytb; 5
representative specimens were sequenced for CoI. Outgroups
for Cytb analyses included 2 specimens of C. planiceps
(AY545541 and AF302183) and 1 specimen each of C.
fumosus (AF302179) and C. merriami (AF302158). Outgroups
for CoI analyses included 1 specimen each of C. merriami
(AY331078) and C. fumosus (AY331075). Species designations follow the taxonomy of Hafner et al. (2004; 2005).
Maximum-parsimony analyses weighted all nucleotide
changes equally and used 5,000 replicates of random, stepwise
addition of taxa in heuristic searches with tree-bisectionreconnection branch swapping. Equally parsimonious trees
were combined into 50% majority-rule consensus trees. Branch
support was estimated as nonparametric bootstrap support
from 5,000 replicates of random taxon addition in a heuristic
search using tree-bisection-reconnection branch swapping. For
maximum-likelihood analyses, model parameters were estimated
194
JOURNAL OF MAMMALOGY
Vol. 89, No. 1
FIG. 3.—A) Major geographic features of the study area and components of the Southern Coahuila Filter-Barrier (stitched line); shading
indicates geographic distribution of Cratogeomys castanops (sensu Hall [1981], with recent updates). B) Collecting localities for chromosomal
analyses (listed in Appendix II) and distribution of 2 major karyotypes: 2n ¼ 46 (open circles) and 2n ¼ 42 (closed circles). C) Collecting
localities for sequence analyses (mitochondrial and nuclear; listed in Appendix III). D) Collecting localities for morphometric analyses (listed
in Appendix I) and identification of reference samples used in discriminant function analyses (black-filled circles ¼ castanops, gray-filled
circles ¼ goldmani).
by using an initial neighbor-joining tree of uncorrected
p-distances subjected to successive iterations of heuristic
searches (Swofford et al. 1996). All Bayesian analyses consisted of paired runs of 4 Markov chain Monte Carlo analyses
each, using default settings and iterated for 6 106 generations sampled every 500 generations. The stationarity of lnlikelihood (ln L) scores of reconstructed trees was evaluated
using the methods of Geweke (1992), Heidelberger and Welch
(1983), and Raftery and Lewis (1992a, 1992b) as implemented
with default settings in the R package BOA version 1.1.5-2
(Smith 2005). All trees sampled before stationarity were
discarded from subsequent analyses, including the construction of 50% majority-rule (MrBayes ‘‘half-compatible’’) consensus trees.
Nuclear DNA sequence analysis.— Sequence data were
obtained for 54 specimens of C. castanops from 28 collecting
localities (Fig. 3C; Appendix III). Amplification of a portion of
the 7th intron (464 bp) and the 8th exon (32 bp) of b-fib was
accomplished using the FIB-B17U and FIB-B17L primers of
Prychitko and Moore (1997) as outlined by Spradling et al.
(2004). b-fib PCR products were prepared for sequencing using
the QIAquick PCR Purification Kit (Qiagen, Inc.). Sequencing
reactions were performed at Iowa State University’s DNA
Facility using their ABI 3730 DNA Analyzer (Applied
Biosystems, Inc.). Sequences were aligned and heterozygosity
was evaluated by eye using Sequencher 4.1.2 software (Gene
Codes Corporation, Ann Arbor, Michigan). Sequences were
submitted to GenBank (GenBank accession nos. DQ518821–
DQ518872). Previously published sequences from C. castanops (GenBank AY331241) and C. goldmani (AY331242)
were used. Sequences from 4 Cratogeomys were used in outgroup analysis: C. merriami (AY331243) and 3 genetically
different C. fumosus (AY331238–AY331240). PAUP* version
4.0b10 was used for parsimony and neighbor-joining analyses
(Swofford 2002).
Morphological analysis.— Nine of the 11 mensural characters described and used by Russell (1968) were recorded from
199 adult female specimens of C. castanops from 82 localities
February 2008
HAFNER ET AL.—SYSTEMATICS OF CRATOGEOMYS
(Fig. 3D; Appendix I). These characters included condylobasal
length, zygomatic breadth, length of palate, length of nasals,
breadth of braincase, squamosal breadth, breadth of rostrum,
length of rostrum, and alveolar length of maxillary toothrow.
Although these characters differ slightly from a set that has
proven useful in previous morphometric analyses of pocket
gophers (Hafner et al. 2004; Patton and Smith 1990; Smith and
Patton 1988), they were chosen to allow direct comparison with
subspecies means and ranges reported in Russell (1968). Two
characters employed by Russell (1968) were excluded because
they could not be measured consistently: palatofrontal depth
and breadth across angular processes. Only females were
measured because past work has demonstrated strong secondary sexual dimorphism in pocket gophers (Hafner et al. 2004;
Patton and Smith 1990; Smith and Patton 1988). Specimens
initially were assigned to age categories following criteria
described in Russell (1968), but it immediately became
apparent that these criteria were selecting only older adults
and that many females of small size but clearly of adult age
(based on tooth wear and suture closure) would be excluded
using these criteria. We instead employed criteria of Daly
and Patton (1986 [for Thomomys bottae]), who demonstrated
that individuals with fused exoccipital–supraoccipital and
basioccipital–basisphenoid sutures were breeding members of
the adult population.
Measurements were transformed to ratios (divided by
condylobasal length) and standardized to reduce the effects
of overall size and individual size variation, respectively.
Lilliefors test (SYSTAT 7.0—Wilkinson 1997) was used to
test for normality of the original and transformed data. Discriminant function analyses and principal component analyses (both implemented in SYSTAT 7.0 [Wilkinson 1997])
were performed on the transformed, standardized characters to
determine if pocket gophers could be separated based on an
a priori hypothesis of group membership in clades identified
by genetic (mtDNA, nuDNA, and chromosomal) analyses.
Principal component analyses and box plots of Russell’s (1968)
subspecies means, our untransformed morphometric data, and
transformed, standardized variables were employed to allow
visual inspection of qualitative differences among cranial
dimensions of taxa.
Geographic information system predictions of species
distributions.— The geographic distributions of the northern
(2n ¼ 46, C. castanops) and southern (2n ¼ 42, C.
‘‘goldmani’’) chromosomal groups were predicted independently using only localities from which we had chromosomal
(Appendix II) or genetic (Appendix III) data. The geocoordinates for each locality were determined in the field
using a Garmin 12XL global positioning system instrument
(Garmin, Olathe, Kansas). We used the DOMAIN algorithm
(DIVA-GIS package, version 5.2—Hijmans et al. 2001) to
classify unsampled areas as castanops or goldmani environments (Carpenter et al. 1993) based on similarities for 19 basic
and composite climatic variables (Hijmans et al. 2005) when
compared with the localities that we sampled. These variables
were not significantly correlated based on random samples of
10,000 points taken from the total extent of the modeled area,
195
and included: 1) annual mean temperature, 2) mean monthly
temperature range, 3) isothermality (variable 2 divided by
variable 7), 4) temperature seasonality (coefficient of variation
[CV] of monthly means), 5) maximum temperature of warmest
month, 6) minimum temperature of coldest month, 7) annual
temperature range, 8) mean temperature of wettest quarter (3
consecutive months), 9) mean temperature of driest quarter, 10)
mean temperature of warmest quarter, 11) mean temperature of
coldest quarter, 12) annual precipitation, 13) precipitation of
wettest month, 14) precipitation of driest month, 15) precipitation seasonality (CV of monthly means), 16) precipitation
of wettest quarter, 17) precipitation of driest quarter, 18)
precipitation of warmest quarter, and 19) precipitation of
coldest quarter values. Spatial resolution for the basic climatic
variables (Hijmans et al. 2005) and the modeled distribution
was 30 arc seconds per cell (approximately 800-m resolution at
298N latitude). Multiple localities per cell were consolidated
into a single locality at the center of the cell. The predicted
distribution was compared visually with the known distribution
of the combined northern (2n ¼ 46) and southern (2n ¼ 42)
chromosomal groups (Hall 1981; MaNIS 2007).
RESULTS
Chromosomal variation.— Our results were consistent with
those of Berry and Baker (1972) and Lee and Baker (1987)
in showing a clear geographic pattern of 2 karyotypes, 1 north
(2n ¼ 46) and 1 south (2n ¼ 42) of the SCFB, except that
specimens from south of the Rı́o Nazas in Durango exhibited
the northern karyotype (2n ¼ 46; Fig. 2, localities 46 and 60;
Fig. 3B). No zone of sympatry was identified in Cañon
Domingo; instead, specimens living at low elevation (1,710 m)
and possessing the northern karyotype (2n ¼ 46, n ¼ 2) were
separated from specimens at higher elevation (2,064–2,115 m)
with the southern karyotype (2n ¼ 42, n ¼ 23) by .10 km of
rocky, steep canyon that did not appear to harbor hospitable
habitat for pocket gophers. The most promising potential zone
of sympatry between the 2 karyotypic forms is along the Rı́o
Aguanaval, between La Unión (2n ¼ 46; Fig. 2, locality 46)
and La Flor de Jimulco (2n ¼ 42; Fig. 2, locality 61).
Fieldwork in late 2006 along the intervening 25 km (permanent
water bordered by hospitable habitat of sandy soil and agricultural fields) documented the 2 forms within 4 km of each
other along the Rı́o Aguanaval approximately midway between
La Unión and La Flor de Jimulco, with no apparent hybridization (based on karyotype and gross external morphology).
The distribution of the 2 karyotypic forms is concordant with
the 2 morphological groups identified by Merriam (1895)
and Nelson and Goldman (1934): castanops north of the
SCFB and goldmani south of the SCFB, with the exception of
the populations south of the Rı́o Nazas in Durango (Fig. 2,
localities 46 and 60).
Mitochondrial DNA analysis.— The amount of locus-specific
sequence obtained varied across samples. As a result, 1,410 bp
(positions 142–1,551) of CoI data and 1,029 bp (positions
22–426 and 454–1,077) of Cytb were retained for analysis.
Alignments resulted in nucleotide variability across samples
196
JOURNAL OF MAMMALOGY
Vol. 89, No. 1
FIG. 4.—Left) Maximum-parsimony majority-rule consensus tree based on 1,410 base pairs of the mitochondrial cytochrome-c oxidase subunit
I gene, for 29 populations of Cratogeomys castanops (Fig. 2; Appendix III) and 2 outgroup species (C. merriami and C. fumosus). Numbers above
branches indicate percentage of bootstrap replicates containing the node (.50% of 5,000 replicates); Bayesian posterior probabilities are shown
below nodes. b-fibrinogen genotypes are as defined in text and Fig. 5. Right) Geographic depiction of groups indicated by maximum-parsimony
tree. Two major clades are indicated (C. castanops, solid lines; C. goldmani, dashed lines), with 2 subclades within each major clade.
and loci consisting entirely of substitutions. Four nucleotide
substitution models were chosen for use across maximumlikelihood and Bayesian analyses. In ModelTest terminology,
these were (with number of free parameters): K81ufþG (6),
HKYþIþG (6), TVMþIþG (9), and GTRþIþG (10). These
were drawn from the 95% AIC (Akaike’s Information
Criterion) and BIC (Bayesian Information Criterion) credibility
sets constructed by ModelTest for CoI and Cytb data and
included the best supported models as indicated by ModelTest
AIC, BIC, and hLRT (hierarchical likelihood–ratio test)
criteria.
Forty-four and 60 most-parsimonious trees were recovered,
respectively, in maximum-parsimony analyses of CoI and Cytb
data. One maximum-likelihood tree was recovered for each
analysis, and all Bayesian analyses converged on stationarity of
tree ln L values within 600,000 generations.
The CoI and Cytb data were not combinable under the
incongruence length difference test (P ¼ 0.001) because of
heterogeneity within the Cytb data, which contained notably
fewer variable and apomorphic nucleotide sites in the 59 than in
the 39 end of the locus. Even so, all phylogenetic analyses of
the CoI and Cytb data returned very similar topological results
(Fig. 4). In all analyses of both genes, ingroup taxa were
consistently separated into 2 broad groups, a northern group
(castanops) of 20 samples and a southern group (goldmani)
of 9 samples, which were consistent with karyotypic designations (2n ¼ 46 and 2n ¼ 42, respectively). These 2 groups
had .99% bootstrap support in maximum-parsimony and
maximum-likelihood analyses and .94% posterior probability
support in Bayesian analyses of both genes.
Substructure within the broad groups of castanops and
goldmani also was consistent across analyses of mtDNA. The
northern castanops group was consistently divided into
a northern clade (localities 1–8, and 11; Fig. 4) and a southern
clade (localities 9, 10, 13, 14, 17, 21, 36, 38, 42, and 60). The
Cytb analyses differed from the CoI analyses only by inclusion
of locality 4 (from New Mexico) in the southern, rather than
northern, castanops clade and placement of localities 9, 10,
and 13 as outgroups to all other castanops specimens in the
Cytb trees (not shown). These differences received only weak
bootstrap support in the Cytb analyses, as did many internal
branches in the Cytb trees.
Maximum-likelihood analysis constraining the specimen
from locality 11 (Primero de Mayo, Coahuila) to group with the
southern clade resulted in a tree with a significantly lower
likelihood score than the tree depicted in Fig. 4. This confirms
that the specimen from locality 11 belongs with the northern
castanops clade, which otherwise includes only taxa from New
Mexico, Oklahoma, and Texas in this study (Fig. 4).
One of 3 specimens sequenced from locality 10 (Ocampo,
Coahuila) grouped with specimens from the nearby (50 km
distant) locality 13 (Cuatrociénegas, Coahuila; 100% bootstrap
support for both genes). The other 2 specimens from locality 10
grouped with specimens from locality 9 (Gallego, Chihuahua),
February 2008
HAFNER ET AL.—SYSTEMATICS OF CRATOGEOMYS
FIG. 5.—Network of 8 b-fibrinogen alleles found in Cratogeomys
castanops (locality numbers from Fig. 2 and Appendix III are given in
parentheses). Neighbor-joining and parsimony analyses each indicate
the same 2 distinct groups of alleles: c-alleles north of the Southern
Coahuila Filter-Barrier (SCFB; castanops), and g-alleles south of the
SCFB (goldmani).
which is nearly 500 km northwest of the Ocampo locality
(bootstrap support 93% in the CoI analysis and 64% in the
Cytb analysis).
The southern group (goldmani) was divided consistently into
an eastern clade (Fig. 4; localities 53, 54, 57, and 64) and
a western clade (localities 61, 78, 81, and 86). In the analyses
of CoI data, the southernmost specimen included in this study
(locality 100; Fig. 2) was depicted as the outgroup to the
eastern clade (bootstrap support ¼ 74%), whereas the Cytb
analyses showed this specimen as outgroup to the western
clade (bootstrap support ¼ 70%). In general, reconstructions
based on mtDNA data had short branch lengths within clades
relative to those among clades and between ingroup and
outgroup taxa (Fig. 4).
Nuclear DNA sequence analysis.— Of the 496 bp of the b-fib
intron-7 and exon-8 regions sequenced, there were 7 nucleotide
positions that varied in the 54 specimens of C. castanops
sampled. There were 8 unique b-fib alleles (Figs. 4 and 5), and
all variable nucleotide positions were located in the intron.
Whether midpoint rooted or rooted using outgroup taxa,
neighbor-joining and parsimony analyses indicate 2 distinct
groups of pocket gophers. Two nucleotide differences (1
transition and 1 transversion) distinguished these 2 groups
and sorted specimens as either northern (castanops) or southern
(goldmani) in a manner consistent with karyotypic and
mtDNA-sequence designations.
Within castanops, there were 4 unique alleles. Allele c1 was
found in gophers from New Mexico, Oklahoma, Texas, and
localities 11 and 13 in northern Mexico; allele c2 was concentrated in the southern portion of the castanops range; and
alleles c3 and c4 were rare (Fig. 4). The geographic distribution
of the c1 and c2 alleles was concordant with the 2 clades within
castanops identified by mtDNA haplotypes, with the exception
of 2 localities closest to the meeting point of the 2 clades.
The single individual from locality 13 (in the southern clade
of castanops) possessed the c1 allele, and the single individual
from locality 11 (in the northern clade) was heterozygous for
both alleles.
Within goldmani, there also were 4 b-fib alleles. However,
allele g2 was much more common and widespread than were
197
the other 3 alleles (Fig. 5). It is possible that the g2 allele may
be a primitive allele for both the castanops and goldmani
groups, because for each of the 7 variable nucleotide positions,
this allele matches the base composition of the outgroup
species, C. merriami and C. fumosus.
Morphological variation.— Although mensural characters
were selected to permit comparison with the subspecies means
reported by Russell (1968), Russell’s mean values were
consistently higher than comparable values from this analysis,
so there was no attempt to combine values from both studies. A
principal component analysis of standardized mean measurements for females of each subspecies (as reported by Russell
[1968]) resulted in only 1 informative component (eigenvalue
¼
. 1), which had high, positive component loadings (X
0.914) for all 9 characters. Taken together, these results
indicate strong influence of overall body size. Subspecies
of the subnubilus group (Russell 1968) are all located along
1 end of a clinal array of component scores (Fig. 6A). When
these subspecies means are transformed to ratios (divided
by condylobasal length) to remove the influence of size, 3
informative components result and mean component loading
is near 0 (0.080). A plot of principal component I scores
(Fig. 6B) from this principal component analysis separates 5
subspecies (all of which are located south of the SCFB) from
the clinal array. All 5 of these southern subspecies possess
a diploid number of 42. Two of the remaining subspecies
(surculus and goldmani) include both 2n ¼ 42 and 2n ¼ 46
karyotypes, but only the mean for goldmani includes specimens from localities with 2n ¼ 42. It is not clear which
localities were included in this mean value; Russell (1968:644)
states only that it includes ‘‘six females . . . from the drainage of
the Rı́o Aguanaval and Rı́o Nazas.’’
Standardized raw measurements for all specimens combined
displayed significant departures from a normal distribution
(P , 0.05) for 8 of 9 variables. After assignment to northern
(castanops) or southern (goldmani) groups based on the subset
of reference specimens, only 2 standardized raw measurements
differed significantly from normality (a different variable in
each group). Whether combined or separated into assigned
groups, standardized ratio data were all normally distributed.
A principal component analysis of standardized ratio
variables revealed no clearly separated groups within the 82
localities in the vicinity of the SCFB. However, discriminant
function analyses of genetically identified reference populations (n ¼ 19 castanops, n ¼ 44 goldmani; Fig. 3D) easily
discriminated between the 2 groups (P ¼ 0.0000), correctly
identifying 98% of the reference specimens to the appropriate
group (Fig. 7). The single misidentified specimen (P ¼ 0.73)
was from locality 88 (near the type locality of Cañitas; Fig. 2),
1 of 5 karyotyped specimens collected from the vicinity of
Cañitas in the 1970s (Berry and Baker 1972). Of 10 total
specimens (6 reference and 4 unassigned) from near the type
locality (localities 82 and 86–89; Fig. 2), 9 were correctly
identified as goldmani. These include the holotype and 2 other
female specimens in the type series. The skull of 1 of the latter
was broken, and a separate discriminant function analysis run
without rostral length identified it as goldmani (P ¼ 1.000).
198
JOURNAL OF MAMMALOGY
Vol. 89, No. 1
FIG. 6.—Results of principal components analyses of A) raw and B) ratio-corrected and standardized subspecies means for 9 mensural
characters of the cranium reported by Russell (1968). Values on principal component I for subspecies included in the subnubilus (open circles) and
excelsus (closed circles) subspecies groups by Russell (1968) are listed in increasing magnitude. The dashed line in B represents the location of the
Southern Coahuila Filter-Barrier (SCFB) relative to the subspecies’ location. Note that when mensural data are ratio-transformed and standardized,
values for the subspecies consitus, surculus, perexiguus, and parviceps (all located north of the SCFB) dramatically shift position relative to the
value for rubellus (south of the SCFB).
Assignment of nonreference populations was concordant with
genetic patterns (i.e., north and south of the SCFB; 92% of the
specimens ‘‘correctly’’ identified) with the exception of locality
88 (above), 3 localities represented by only a single specimen
(localities 20, 80, and 85), and 3 localities in the vicinity of
Tepeyac, San Luis Potosı́ (localities 95–97; Fig. 2).
Specimens of C. goldmani from which mtDNA data are
available (Figs. 3D and 4) served as reference samples in
a discriminant function analysis (n ¼ 14 from the eastern
FIG. 7.—Distribution of individual scores along discriminant
function I of a discriminant function analysis for reference specimens
(lower graph; Fig. 3D) and total combined individuals (upper graph)
from north (castanops) and south (goldmani) of the Southern Coahuila
Filter-Barrier. Dots between the 2 graphs indicate scores for 1
specimen from La Unión (black) and 3 specimens from La Flor de
Jimulco (gray), separated by 25 km along the Rı́o Aguanaval at the
Durango–Coahuila border (see Fig. 1 and text).
portion of the range; n ¼ 10 from the western portion) that
distinguished the 2 groups morphologically (P ¼ 0.0111) and
correctly assigned 96% of the reference specimens (Fig. 8).
Localities were then assigned to the 2 groups based on average
posterior probability. Specimens from 3 localities along the
extreme southwestern edge of the species’ range (localities 87,
92, and 93; Fig. 2) and 1 locality at the extreme eastern edge
(locality 75) departed from the general east–west pattern
indicated by mtDNA analysis. Locality 87, Cañitas, Zacatecas,
is the type locality for C. c. goldmani. Of 3 females measured
from this locality, 2 (including the holotype) were assigned to
the eastern group (posterior probabilities ¼ 0.92 and 0.93). A
separate discriminant function analysis run without rostral
length assigned the 3rd (damaged) specimen to the western
group (posterior probability ¼ 0.85).
FIG. 8.—Distribution of individual scores along discriminant
function I of a discriminant function analysis for reference specimens
(lower graph) of 2 subclades within Cratogeomys goldmani indicated
by mitochondrial DNA (mtDNA; Figs. 3D and 4), and total combined
individuals of C. goldmani (upper graph). Ninety-six percent of the
reference specimens were correctly identified to their proper mtDNA
clade.
February 2008
HAFNER ET AL.—SYSTEMATICS OF CRATOGEOMYS
FIG. 9.—Geographic information system prediction using the
DOMAIN algorithm of the geographic distribution of Cratogeomys
castanops (85% classification confidence; shading) based on similarity
of values for 19 climatic variables among 46 localities of C. castanops
(larger black dots with white borders). The predicted distribution of
C. castanops closely matches marginal localities reported in Hall (1981;
smaller black dots) and also predicts the distribution of C. goldmani
(large open dots).
The 2 geographic groups identified in the mtDNA and
morphometric analyses correspond generally to sets of eastern
(elibatus, peridoneus, planifrons, and subnubilus) and western
(rubellus and surculus) subspecies of Russell’s (1968:662)
subnubilus group, except for slight differences in the vicinities
of Concepción del Oro, Zacatecas, and Matehuala, San Luis
Potosı́ (Fig. 2, localities 70 and 84–85, respectively). Samples
from 3 localities in the vicinity of Concepción del Oro exhibit mixtures of eastern and western forms, with individuals
variously assigned to either group with high posterior prob ¼ 0.97).
abilities (0.81–1.00, X
Geographic information system prediction of species
distributions.— The predicted distribution of the northern chromosomal form (2n ¼ 46) of C. castanops (Fig. 9) is divided
into 3 classification confidence intervals: .85–90%, .90–
95%, and .95–100%. This prediction, which was based on
only 46 localities and 19 climatic variables, conforms well to
the known distribution of C. castanops (Hall 1981:520; MaNIS
2007; augmented by additional distributional data known to the
authors and data from R. C. Dowler [pers. comm.]). The model
199
FIG. 10.—Summary distribution of subspecies of Cratogeomys
castanops (dark gray) and C. goldmani (light gray) recognized in this
paper (heavy dashed lines ¼ subspecies boundaries). The distributions
have been modified based on the results of the geographic information
system prediction (Fig. 9), and reallocation of the Montemorelos,
Nuevo León, population to C. castanops.
overpredicts the known distribution of castanops (but mostly at
the low confidence level) in northern and eastern portions of
its range, and underpredicts the known distribution in western
Texas (Fig. 9). Two large areas demarcated by the high
confidence level (.95%) include a broad area in eastern New
Mexico and western Texas (roughly corresponding to the Llano
Estacado), and a broad area extending north of the SCFB into
northern Durango and central Coahuila (Fig. 9). The predicted distribution of C. castanops (Fig. 9) also includes most
of the known localities of the southern chromosomal form
(goldmani), even though goldmani localities were not used as
input data in the geographic information system model.
DISCUSSION
Systematic relationships.— Genetic groups defined by karyotype, mtDNA sequence analysis, and nuDNA sequence
analysis are fully concordant with morphological groupings
defined by Merriam (1895) and Nelson and Goldman (1934),
but not with those recognized by Russell (1968). Analysis of
200
JOURNAL OF MAMMALOGY
ratio-transformed morphometric variables resulted in patterns
fully concordant with the original castanops and goldmani
morphological groups. Contrary to Russell (1968), we found
no evidence for sympatry at any localities thus far examined
in any analysis. The 2 genetic units approach contact along
the SCFB as defined by Baker (1956), except that the elevated region between the drainages of the Rı́os Nazas and
Aguanaval, rather than the Rı́o Nazas, separates the 2 genetic
units along the western portion of the SCFB. The forms are
likely sympatric (current or intermittent) along the lower
reaches of the Rı́o Aguanaval, between La Unión, Durango
(Fig. 2, locality 46), and La Flor de Jimulco, Coahuila (Fig. 2,
locality 61). Based on this clear and well-supported distributional pattern and the absence of any evidence of introgressive
hybridization between the genetic forms, we recognize 2
separate species, each of which is characterized by reciprocally
monophyletic nuDNA sequences and mtDNA haplotypes,
distinct morphology, and different diploid numbers (Fig. 10).
As noted by Hafner et al. (2005), diploid number differences in
pocket gophers usually indicate reproductive barriers between
species (Hafner et al. 1987; Patton 1985; but see Patton et al.
[1984] and Thaeler [1968] for contrary examples).
Additional support for dividing C. castanops sensu lato into
northern (C. castanops sensu stricto) and southern (C. goldmani) species derives from the distribution of ectoparasitic
lice (genus Geomydoecus—Hellenthal and Price 1976). Two
species of Geomydoecus, G. expansus and G. texanus, are
represented by different subspecies generally north and south
of the SCFB. However, the detailed distributions of the lice are
not ‘‘completely compatible’’ with genetic patterns in their
hosts, as stated by Lee and Baker (1987:3). For example, in
G. expansus, broad mixing of the 2 subspecies is evident
among pocket gophers along the eastern edge of the SCFB,
with the northern subspecies, G. e. expansus, extending south
into San Luis Potosı́ and the southern subspecies, G. e. martini,
extending slightly north of the SCFB into Coahuila. Along the
western margin of the SCFB, the population of C. castanops at
Hacienda Atotonilco, Durango (Fig. 2, locality 60), hosts the
southern subspecies, G. e. martini, whereas C. goldmani from
La Flor de Jimulco (Fig. 2, locality 61) is parasitized by the
northern subspecies, G. e. expansus. The Hacienda Atotonilco
population of C. castanops also possesses the southern subspecies of G. texanus (G. t. subnubili), and this subspecies
extends into populations of C. castanops north into southeastern Coahuila at the eastern end of the SCFB. These discrepancies between gopher and louse distributions are further
evidence that C. castanops and C. goldmani are (or have been)
in contact, where they have exchanged lice, but not genes.
Morphometric discordances.—It is evident from comparison
of analyses of untransformed and ratio-transformed morphometric variables (Fig. 6) that Russell (1968) was unduly biased
by size in his evaluation of geographic variation in C.
castanops (sensu lato). The ecophenotypic plasticity of body
size in pocket gophers has been amply demonstrated (Patton
and Brylski 1987): populations living in habitat that supports
increased food resources are of larger body size. Populations
north of the SCFB that Russell (1968) included in the smaller-
Vol. 89, No. 1
sized subnubilus group are all from arid regions that support
lower levels of food resources (e.g., lower-elevation deserts of
Chihuahua and the arid Tularosa Basin of New Mexico). We
consider the few specimens of pocket gopher assigned to the
‘‘wrong’’ genetic group based on our morphometric analyses to
represent geographic variation and homoplastic convergence
rather than genetic introgression between the 2 forms. Finally,
we refer the population from Montemorelos, Nuevo León,
along the eastern edge of the Sierra Madre Oriental (Fig. 1,
locality 8; Fig. 2, locality 58), to C. castanops (as originally
assigned by Nelson and Goldman [1934], contra Russell
[1968]). Efforts to locate living populations of pocket gophers
in the vicinity of Montemorelos in 2006 were fruitless.
Geographic variation.— The most comprehensive evaluation
of geographic variation within C. castanops (sensu lato—
Russell 1968) resulted in recognition of 25 subspecies, but was
heavily influenced by an ecophenotypically plastic character,
body size. Our analysis of geographic variation within C.
goldmani, based on adult female specimens from most known
localities, revealed morphological groups concordant with
those identified in the mtDNA analysis. Below, we formally
recognize these 2 groups as subspecies (C. g. goldmani and
C. g. subnubilus), with a narrow transition zone in the vicinity
of Concepción del Oro, Zacatecas.
We have chosen to recognize subspecies within C. castanops
based primarily on genetic groups, rather than retain morphologically based subspecific determinations that we know to be
flawed. Our recognition of subspecies is in agreement with
genetic definitions outlined by Lidicker (1960, 1962) and
expanded by Endler (1977). Our sampling of C. castanops
(sensu stricto) was not designed to evaluate geographic
variation in morphology throughout its distribution. However,
we expect that geographic variation in morphology would be
concordant with the distribution of C. castanops mitochondrial
haplotypes, as it was for C. goldmani. Furthermore, mtDNA
and nuDNA analyses were concordant in identifying 2 genetic
subunits within C. castanops (Figs. 4 and 5), which we
recognize herein as subspecies. The oldest available names are
C. c. castanops (northeastern subunit) and C. c. consitus
(southwestern subunit). The single individual examined from
1.7 km north of Primero de Mayo, Coahuila (where the 2
subspecies meet; Fig. 2, locality 11), was heterozygous for the
diagnostic c1 and c2 alleles, whereas a single individual from
a nearby locality (2 km northwest of Cuatro Ciénegas; Fig. 2,
locality 13) possessed the southern mtDNA haplotype and the
northern nuDNA allele. This region marked the boundary
between the former subspecies C. c. jucundus and C. c.
bullatus (Russell 1968). The 2 (revised) subspecies may come
into secondary contact in the vicinity of Ojinaga, Coahuila,
near the confluence of the Rı́o Conchos and Rı́o Grande.
Russell (1968:671) noted disparity in size and habitat in this
area between populations of his excelsus group (larger pocket
gophers inhabiting deep soils along the rivers) and his
subnubilus group (smaller pocket gophers restricted to thinner,
upland soils). This disparity in size may represent simply
ecophenotypic plasticity, genetic-based differences between the
2 subspecies, or a combination thereof.
February 2008
HAFNER ET AL.—SYSTEMATICS OF CRATOGEOMYS
We recognize that our evaluation of geographic variation
within each species and our subsequent recognition of
subspecies is a marked departure from traditional methods
based on morphological variation alone. It is evident to us that
homoplasy in morphological characters among geographically
isolated populations, compounded by a misleading emphasis
on size in the interpretation of variation, has led to oversplitting
of these pocket gophers at the subspecific level. Patton and
Smith (1990) used similar logic to reduce the number of
Thomomys taxa in California from 46 to 15 subspecies.
Geographic information system prediction of species
distributions.— We used a geographic information system
model based on climatic variables at a limited number of localities of C. castanops to predict with considerable accuracy the
entire known distribution of the species (Fig. 9). The ability to
predict pocket gopher distribution based on simple climatic
variables runs counter to the common perception that pocket
gophers (and other subterranean organisms, in general) live in
‘‘sealed [eco]systems’’ that are effectively buffered from climatic variables extrinsic to the system (Nevo 1979:272). The
fact that the model could predict the distribution of C. castanops
without inclusion of soil-related variables suggests that specific
characteristics of the soil may be less important determinants
of pocket gopher distribution than are precipitation and temperature, the latter acting directly on the food resources (plants) in
these ‘‘preferred’’ climates. Our trapping experience suggests
that gophers will live in almost any type of soil as long as it does
not flood and is sufficiently deep to buffer the nest from surface
temperatures and provide shelter from predators.
The 2 large areas with high (.95%) classification confidence for C. castanops (Fig. 9) correspond reasonably well
with the 2 major genetic clades within C. castanops (Fig. 4).
These clades, which we recognize at the subspecies level,
are separated by physiographic discontinuities, including the
arid lowlands of eastern Chihuahua (,200 mm annual
precipitation—Anderson 1972) and the Sierra del Carmen–
Sierra Madre Oriental Filter-Barrier (Baker 1956).
The castanops-only geographic information system model
predicted the present distribution of C. goldmani quite accurately (Fig. 9), and the goldmani-only geographic information
system prediction (not shown) included most localities of the
southern (but not the northern) genetic clade of C. castanops.
Together, these results suggest that the climatic niche of
goldmani is a subset of the castanops climatic niche. If so, the
present-day separation of the 2 species in southern Coahuila
may be more a consequence of historical dispersal and vicariance events associated with the filter barrier, than it is the
result of ecological adaptation to different habitats north and
south of the barrier.
Southern Coahuila Filter-Barrier.— Baker (1956), Baker
and Greer (1962), Peterson (1976), and Schmidly (1977) all
have emphasized the importance of the SCFB in influencing
mammalian distribution in the Chihuahuan Desert. More
broadly, Arriaga et al. (1997) and Morrone (2005) have
recognized 2 major subregions of the Chihuahuan Desert (¼
Provincia del Altiplano Mexicano) that meet at the SCFB
(Altiplano Norte and Altiplano Sur), based on topography,
201
climate, potential vegetation, and faunal analyses. Only
Morafka (1977) de-emphasized the effect of the SCFB, maintaining that the herpetofauna of the southern subregion (Saladan
Desert) was merely a depauperate subset of the northern
subregion (Mapimian division). In contrast, Smith (1941)
defined a southern province (Austro-Central) separate from
the ‘‘Chihuahuense’’ province, based mostly on species of the
lizard genus Sceloporus.
The SCFB has 3 distinct segments: a western portion in which
2 rivers (Rı́o Nazas and Rı́o Aguanaval) approach one another
from the west and south, respectively; the central Mayrán Basin,
which alternates temporally between alkali flats and a shallow
laguna fed with periodic floodwaters from the 2 rivers; and the
eastern mountain ranges (western extensions of the Sierra
Madre Oriental), which are broken intermittently by low passes.
These 3 segments would be expected to exert different influences on species variously adapted to rocky versus sandy
substrates, or to arid versus riparian-edge situations.
Sixteen species of Chihuahuan Desert rodents occur in the
vicinity of the SCFB (Hall 1981). Six species occur generally
(or are restricted) to the north (Ammospermophilus interpres,
Dipodomys nelsoni, and C. castanops), or the south (D.
spectabilis, D. phillipsii, and C. goldmani). One other species
(Neotoma goldmani) has a continuous distribution across the
eastern edge of the SCFB, with no recognized subspecific
differentiation. The remaining 9 species appear to occur continuously across the SCFB. All except Chaetodipus eremicus
have named subspecies whose distributional boundaries coincide with the SCFB (Ch. hispidus, Ch. nelsoni, D. merriami,
D. ordii, Perognathus flavus, N. leucodon, Onychomys
arenicola, and Peromyscus eremicus). The impact of the
central segment of the SCFB is the most intense of the
3 segments, marking the distribution limit of 6 species and
delineating subspecies within another 8 species (i.e., affecting
93% of the species). Only Ch. eremicus appears to be continuously distributed along the margins of the Mayrán Basin,
although there are no records from the basin floor. The eastern
segment is the distributional limit of 4 species, and it delineates
subspecies within another 6 species (affecting 67% of the
species). The southern species of Cratogeomys, C. goldmani,
extends just 7 km north of a 2,000-m-elevation pass on the
eastern edge of the Sierra Guadalupe (near Agua Nueva). The
western segment appears to have the weakest impact, marking
the distributional limit of only 3 species and delineating
subspecies in another 3 species (affecting only 43% of the
species). Distributional shifts across the Rı́o Nazas are
primarily to the south (2 species and 4 subspecies). Although
the Rı́o Nazas has been viewed as the single component of the
western segment of the SCFB (Baker and Greer 1962; Peterson
1976), detailed genetic analysis of the relationships among
conspecific populations north and south of the Rı́o Nazas and
in the valley of the Rı́o Aguanaval may reveal additional
components to this segment: the elevated divide between the
valleys of the 2 rivers, or the Rı́o Aguanaval itself. Thus, the
western segment of the SCFB may extend in a fan shape
(anchored in the Mayrán Basin) from the Sierra Jimulco, west
across the valley of the Rı́o Aguanaval, northwest across the
202
JOURNAL OF MAMMALOGY
divide into the valley of the Rı́o Nazas, and north into the
Bolson Mapimı́. Based on current taxonomy, this would
include the distribution limits of 5 species and delineate subspecies in another 7 species, thus affecting 86% of the species
and excepting only the sand-dwelling Ch. eremicus and the
rock-dwelling Ch. nelsoni.
TAXONOMIC CONCLUSIONS
Results of karyotypic, mtDNA and nuDNA sequence, and
morphometric analyses are concordant in identifying 2 species
within C. castanops that meet along the SCFB of Durango and
Coahuila, Mexico. Populations of Cratogeomys north of the
SCFB and throughout Durango are considered C. castanops,
whereas those south of the SCFB and east of Durango are
C. goldmani, in agreement with publications subsequent to Lee
and Baker (1987) that employed these names. There is no
evidence of introgression between the species from any of the
analyses, and the only known potential zone of contact between
the species is along the Rı́o Aguanaval between La Unión,
Durango, and La Flor de Jimilco, Coahuila. Merriam (1895)
and Nelson and Goldman (1934) correctly envisioned this
pattern of species distribution, including assignment of specimens on the eastern versant of the Sierra Madre Oriental
near Montemorelos to castanops rather than goldmani
(Russell [1968] considered it to be a member of his southern
subnubilus group).
Our recognition of 2 subspecies within C. goldmani is
based on concordance between patterns of morphometric and
mtDNA sequence variation. Recognition of subspecies within
C. castanops is based on concordant patterns of mtDNA and
nuDNA sequence variation. The 2 points of contact between
these subspecies match subspecific boundaries identified in the
more comprehensive (albeit size-biased) morphologic analysis
of Russell (1968). Primero de Mayo, Coahuila, is between the
subspecies bullatus and jucundus, and a specimen from this
locality was heterozygous for the subspecies’ diagnostic alleles.
The other point of contact, along the Rı́o Grande at Ojinaga,
Chihuahua, was considered by Russell (1968) to be a zone of
sympatry between clarkii (of his excelsus-group) and consitus
(of his subnubilus-group).
Cratogeomys Merriam, 1895
Cratogeomys Merriam, 1895:150. Type species Geomys
merriami Thomas. Cratogeomys was regarded as a subgenus
of Pappogeomys by Russell (1968:592), but was returned to
generic status by Honeycutt and Williams (1982:212).
Platygeomys Merriam, 1895:162. Type species Geomys
gymnurus Merriam. Regarded as inseparable from Cratogeomys by Hooper (1946:397).
In the most recent comprehensive revision of the genus (as
part of the genus Pappogeomys), Russell (1968) recognized 2
species groups: the castanops species group, with 2 species
(castanops, 2 subspecies groups and 25 subspecies, and
merriami, with 7 subspecies) and the gymnurus species group,
with 5 species (3 monotypic species, C. neglectus, C. zinseri,
and C. fumosus; and 2 polytypic species, C. tylorhinus [6 sub-
Vol. 89, No. 1
species] and C. gymnurus [3 subspecies]). Álvarez and ÁlvarezCastañeda (1996) subsequently added 1 additional subspecies of
C. castanops. Hafner et al. (2004) revised the gymnurus species
group to include 2 species, C. fumosus (4 subspecies) and the
monotypic C. planiceps, and renamed it as the fumosus species
group. Hafner et al. (2005) revised C. merriami to include 2
additional species, C. fulvescens and C. perotensis. Berry and
Baker (1972) and Lee and Baker (1987) suggested that
C. goldmani might be specifically distinct from C. castanops,
but did not formally elevate goldmani to species status.
Our recognition of C. goldmani as a distinct species results
in 5 species in the castanops species group: C. castanops,
C. goldmani, C. fulvescens, C. merriami, and C. perotensis.
Cratogeomys castanops (Baird, 1852)
Yellow-faced Pocket Gopher
(Synonymy under subspecies)
Geographic range.— Patchily distributed from the grammagrass–dominated high plains and plateaus of the western edge
of the Great Plains in southeastern Colorado and southwestern
Kansas (388N latitude, 750 m elevation), throughout the Llano
Estacado of New Mexico and Texas, along the Rı́o Grande (to
near its mouth at the Gulf of Mexico) and south into the highelevation, desert grasslands of the Chihuahuan Desert to the
Sierra Guadalupe, Sierra Parras, and Mayrán Basin of southern
Coahuila, and across the Rı́o Nazas and west of the Rı́o
Aguanaval in Durango to the northeastern slopes of the Sierra
de Yerbanı́s (258N latitude, 2,000 m elevation). Elevational
range approximately 10 m (near the mouth of the Rı́o Grande)
to 2,150 m.
Description.—Medium body size for genus; generally yellowish brown in dorsal coloration; skull long (relative to
C. goldmani), particularly in rostrum and braincase. Compared
to C. goldmani, cranium larger in all measurements, but with
relatively longer palate and nasals, and relatively narrower bullar
breadth. Body size highly influenced by local climate: smaller
in drier regions, larger in wetter regions with more dense vegetation. Within C. c. castanops, body size largest in the southern
distribution, smaller in intermediate latitudes of Trans-Pecos
Texas, and growing larger in higher latitudes. Within C. c.
consitus, body size largest in south, and decreasing significantly
with the shift to open Chihuahuan Desert of more northern
latitudes (R2 ¼ 0.857, P ¼ 0.029; comparison of principal component I of a principal component analysis based on Russell’s
[1968] subspecies means with maximum northern latitude).
Cratogeomys castanops castanops (Baird, 1852)
Pseudostoma castanops Baird, 1852:313. Type locality
‘‘prairie road to Bents Fort’’ near the present town of Las
Animas, Colorado. Type specimen age and sex unknown,
skin and skull, United States National Museum number
4007/3861, collected in 1845 by Lieutenant Abert. ‘‘The
type specimen, formerly in the Patent Office, is now in the
National Museum, but is in very poor condition, having been
exposed to the light for nearly forty years, as a result of
which it is so faded that no trace of the original color
February 2008
HAFNER ET AL.—SYSTEMATICS OF CRATOGEOMYS
remains’’ (Merriam 1895:160), and ‘‘The specimen was at
first mounted; later it was made into a study skin’’ (Russell
1968:635).
Cratogeomys castanops: Merriam, 1895:159; name combination.
C. c. castanops: Miller, 1912:247. First use of current name
combination.
C. c. angusticeps Nelson and Goldman, 1934:139. Type
locality ‘‘Eagle Pass, [Maverick County,] Texas.’’
C. c. bullatus Russell and Baker, 1955:597. Type locality ‘‘2
mi. S and 6 1/2 mi. E Nava, 810 ft., Coahuila.’’
Geomys clarkii Baird, 1855:332. Type locality ‘‘Presidio del
Norte, [at or near the present town of Ojinaga], on the Rı́o
Grande, Chihuahua.’’
C. c. convexus Nelson and Goldman, 1934:142. Type locality
‘‘seven miles east of Las Vacas, Rio Grande Valley,
Coahuila, Mexico (opposite Del Rio, Texas).’’
C. c. dalquesti Hollander, 1990:45. Type locality ‘‘1 mi. N and
4 mi. W Sterling City, Sterling Co., Texas.’’
C. c. hirtus Nelson and Goldman, 1934:138. Type locality
‘‘Albuquerque, [Bernalillo Co.,] New Mexico (altitude 5,000
feet).’’
C. c. lacrimalis Nelson and Goldman, 1934:137. Type locality
‘‘Roswell, Chaves County, New Mexico (altitude 3,500
feet).’’
Pappogeomys castanops parviceps Russell, 1968:673. Type
locality ‘‘18 mi. SW Alamogordo, 4,400 ft., Otero Co., New
Mexico.’’
C. c. perplanus Nelson and Goldman, 1934:136. Type locality
‘‘Tascosa, Oldham County, Texas (altitude 3,000 feet).’’
Pappogeomys castanops pratensis Russell, 1968:653. Type
locality ‘‘8 mi. W and 3 mi. S Alpine, 5,100 ft., Brewster
Co., Texas.’’
Pappogeomys castanops simulans Russell, 1968:656. Type
locality ‘‘17 mi. SE Washburn, Armstrong County, Texas.’’
C. c. tamaulipensis Nelson and Goldman, 1934:141. Type
locality ‘‘Matamoros, Tamaulipas, Mexico.’’
Pappogeomys castanops torridus Russell, 1968:665. Type
locality ‘‘3 mi. E Sierra Blanca, about 4,000 ft., Hudspeth
Co., Texas.’’
C. c. ustulatus Russell and Baker, 1955:598. Type locality
‘‘Don Martı́n, 800 ft., Coahuila.’’
Geographic range.— Patchily distributed from the grammagrass–dominated high plains and plateaus of the western edge
of the Great Plains in southeastern Colorado and southwestern
Kansas, throughout the Llano Estacado of New Mexico and
Texas, along and between the Pecos River and Rı́o Grande (to
near its mouth at the Gulf of Mexico) and south onto the
eastern flanks of the northern Sierra Madre Oriental near Primer
de Mayo, Coahuila, and up the Rı́o San Juan to near
Montemorelos, Nuevo León. Elevational range approximately
10 m (near mouth of the Rı́o Grande) to 2,150 m.
Cratogeomys castanops consitus Nelson and
Goldman, 1934
C. c. consitus Nelson and Goldman, 1934:140. Type locality
‘‘Gallego, Chihuahua, Mexico (altitude 5,500 feet).’’
203
C. c. excelsus Nelson and Goldman, 1934:143. Type locality
‘‘San Pedro [¼ San Pedro de las Colonias], 10 miles west of
Laguna de Mayrán, Coahuila, Mexico.’’
C. c. goldmani Merriam, 1895:160. Type locality ‘‘Cañitas,
Zacatecas, Mexico.’’ Part, not specimens from the Rı́o
Aguanaval drainage and eastward.
C. c. jucundus Russell and Baker, 1955:599. Type locality
‘‘Hermanas, 1,205 ft., Coahuila.’’
Pappogeomys castanops perexiguus Russell, 1968:676. Type
locality ‘‘6 mi. E Jaco, Chihuahua, 4,500 ft., in Coahuila.’’
C. c. sordidulus Russell and Baker, 1955:600. Type locality
‘‘1.5 mi. NW Ocampo, 3,300 ft., Coahuila.’’
C. c. subsimus Nelson and Goldman, 1934:144. Type locality
‘‘Jaral [¼ San Antonio de Jaral], southeastern Coahuila,
Mexico.’’
Pappogeomys castanops surculus Russell, 1968:688. Type
locality ‘‘La Zarca, Durango.’’ Part, not specimens from the
Rı́o Aguanaval drainage and eastward.
Geographic range.— Patchily distributed throughout the
high-elevation, desert grasslands of the Chihuahuan Desert
from Samalayuca, Chihuahua, across the Rı́o Conchos and
Mapimı́ Basin (Bolsón de Mapimı́) to the Sierra Guadalupe,
Sierra Parras, and Mayrán Basin of southern Coahuila, and
across the Rı́o Nazas and west of the Rı́o Aguanaval in
Durango to the northeastern slopes of the Sierra de Yerbanı́s at
Hacienda Atotonilco, Durango. Elevational range approximately 850–2,150 m.
Cratogeomys goldmani Merriam, 1895
Goldman’s Yellow-faced Pocket Gopher
(Synonymy under subspecies)
Geographic range.— Patchily distributed throughout the
arid, high-elevation Mesa del Norte south of the Mayrán
Basin, Sierra Parras, and Sierra Guadalupe of southern
Coahuila from the drainage of the Rı́o Aguanaval (Coahuila,
Durango, and Zacatecas) to the western flanks of the Sierra
Madre Oriental, and south to the Rı́o Verde in San Luis Potosı́.
Elevational range approximately 750–2,650 m.
Description.— Body size small for genus; generally more
ochraceous in dorsal coloration than C. castanops; skull short
(relative to C. castanops), particularly in rostrum and braincase.
Compared to C. castanops, cranium smaller in all measurements, but with relatively shorter palate and nasals, and
relatively broader bullar breadth. C. g. goldmani is larger than
C. g. subnubilus in all cranial characters, but with relatively
narrower bullar breadth and squamosal breadth. The smaller
subspecies (C. g. subnubilus) occurs in higher elevations (along
the western flanks of the Sierra Madre Oriental) and lower
elevations (of San Luis Potosı́) relative to the intermediate
elevations of the Mesa del Norte occupied by C. g. goldmani.
As noted by Dalquest (1953:101) and Russell (1968:685),
populations from eastern Zacatecas and western San Luis
Potosı́ assigned by Russell (1968) to the subspecies rubellus,
surculus, and subnubilus, and assigned herein to goldmani and
subnubilus, display a high incidence (89%) of white spots on
the belly, sides, or rump.
204
Vol. 89, No. 1
JOURNAL OF MAMMALOGY
Cratogeomys goldmani goldmani Merriam, 1895
Cratogeomys castanops goldmani Merriam, 1895:160. Type
locality ‘‘Cañitas, Zacatecas, Mexico.’’ Type specimen
young adult female, skin and skull, United States National
Museum number 57965, collected 24 December 1893 by
E. A. Goldman, original number 286. Part, not specimens
from the Rı́o Nazas drainage and northward.
C[ratogeomys] goldmani: Lee and Baker, 1987:13. Name
combination.
C. g. goldmani: DeWalt et al., 1993:200. First use of current
name combination.
Cratogeomys castanops rubellus Nelson and Goldman,
1934:147. Type locality ‘‘Soledad, near San Luis Potosı́,
San Luis Potosı́, Mexico (altitude 6,400 feet).’’
C. g. rubellus: DeWalt et al., 1993:200. First use of current
name combination.
Pappogeomys castanops surculus Russell, 1968:688. Type
locality ‘‘La Zarca, Durango.’’ Part, not specimens from the
Rı́o Nazas drainage and northward.
Geographic range.— Patchily distributed in the more arid
regions of the Mesa del Norte south of the Mayrán Basin from
the drainage of the Rı́o Aguanaval (Coahuila, Durango, and
Zacatecas) to the higher elevations of western San Luis Potosı́.
Elevational range approximately 1,300–2,350 m.
Cratogeomys goldmani subnubilus Nelson and
Goldman, 1934
Cratogeomys castanops subnubilus Nelson and Goldman,
1934:145. Type locality ‘‘Carneros, Coahuila, Mexico
(altitude 6,800 feet).’’
Pappogeomys castanops elibatus Russell, 1968:672. Type
locality ‘‘12 mi. W San Antonio de Alazanas, about 7,500 ft.,
Coahuila.’’
C. g. maculatus Álvarez and Álvarez-Castañeda, 1996:38.
Type locality ‘‘1.5 km SE Cedral, 1,600 m, San Luis Potosı́,
México.’’
Cratogeomys castanops peridoneus Nelson and Goldman,
1934:148. Type locality ‘‘Rı́o Verde, 3,000 ft., San Luis Potosı́.’’
Cratogeomys castanops planifrons Nelson and Goldman,
1934:146. Type locality ‘‘Miquihuana, Tamaulipas.’’
Geographic range.—Patchily distributed in the relatively
wetter regions of the Mesa del Norte south of the Sierra Parras
and Sierra Guadalupe of southeastern Coahuila, along the
western flanks of the Sierra Madre Oriental and south to the
Rı́o Verde in San Luis Potosı́. Elevational range approximately
750–2,650 m.
RESUMEN
La barrera-filtro del sur de Coahuila subdivide efectivamente
la fauna de mamı́feros de la Mesa del Norte, la sección norteña
y más extensa del Altiplano Mexicano. Las tuzas del género
Cratogeomys al norte y sur de ésta barrera-filtro han sido
reconocidas informalmente como dos especies distintas, C.
castanops y C. goldmani (respectivamente). El apoyo para el
reconocimiento a nivel especı́fico proviene de comparaciones
morfológicas antiguas y estudios recientes de cromosomas y
ectoparásitos. Conclusiones contradictorias basadas en el único
estudio morfométrico completo han impedido el reconocimiento formal de C. goldmani. Una reevaluación morfométrica
basada en datos proporcionales transformados revela que los
análisis previos estuvieron influenciados indebidamente por la
talla, un carácter ecofenotı́picamente plástico. Cuando este
factor es eliminado, la variación morfométrica es completamente coherente con el número cromosómico diploide y con
secuencias de ADN mitocondrial y nuclear. En esta contribución proveemos sinonimias y descripciones para C. goldmani
separándola de C. castanops, y revisamos el número de
subespecies de un total de 26 a 2 subespecies en cada especie.
La barrera-filtro del sur de Coahuila es más efectiva en su parte
central (Desierto Mayrán), y menos efectiva en su porción
occidental (Rı́o Nazas), que deberı́a ser ampliada geográficamente para incluir el vecino Rı́o Aguanaval.
ACKNOWLEDGMENTS
We thank our Mexican collaborator, F. A. Cervantes, and his
students C. Ballesteros, J. A. Fernández, X. Isidro, M. E. Mancera,
L. Mondragón, A. Montiel, and J. P. Ramı́rez for their hospitality and
helpful field assistance in Mexico. A. Waychoff did much of the work
to generate the b-fib sequences. J. C. Hafner provided helpful advice
on statistics, K. Butzer provided useful information on Laguna
Mayrán, R. C. Dowler reported a new extralimital locality for
C. castanops in Texas, and J. A. Fernández provided the Spanish
translation of the abstract. A. Gardner, J. L. Patton, and T. Holmes
generously and carefully measured selected specimens (including type
specimens) in their care. We thank the following institutions and
curators for providing or examining museum specimens in their care:
Universidad Nacional Autónoma de México Colección Nacional de
Mamı́feros (F. A. Cervantes and Y. Hortelano), University of Kansas
Natural History Museum (R. M. Timm, T. Holmes, and H. Campbell);
University of California Museum of Vertebrate Zoology (J. L. Patton);
The Museum, Texas Tech University (R. Bradley, H. Garner, and M.
Renick); University of Michigan Museum of Zoology (P. Myers and
S. Hinshaw); Michigan State University Museum (B. Lundrigan and
L. Abraczinskas), and the United States National Museum of Natural
History (A. Gardner). Animals collected for this project were treated
in a humane manner as approved by the University of New Mexico
and Louisiana State University Institutional Animal Care and Use
Committees following guidelines of the American Society of
Mammalogists (Gannon et al. 2007). This research was supported
by National Science Foundation grants 0236957 (to D. J. Hafner) and
0343869 (to M. S. Hafner).
LITERATURE CITED
ÁLVAREZ, T., AND S. T. ÁLVAREZ-CASTAÑEDA. 1996. Descripción de
una nueva subespecie de tuza, Cratogeomys goldmani (Rodentia:
Geomyidae), de San Luis Potosı́, México. Acta Zoologica Mexicana
(Nueva Serie) 68:37–43.
ANDERSON, S. 1972. Mammals of Chihuahua: taxonomy and distribution. Bulletin of the American Museum of Natural History 148:
149–410.
ARRIAGA, L., C. AGUILARA, D. ESPINOSA-ORGANISTA, AND R. JIMÉNEZ.
1997. Regionalización ecológica y biogeográfica de México. Taller
de la Comisión Nacional para el Conocimiento y Uso de la
Biodiversidad (CONABIO), México, Distrito Federal, México.
February 2008
HAFNER ET AL.—SYSTEMATICS OF CRATOGEOMYS
BAIRD, S. F. 1852. Mammals. Pp. 309–313 (appendix C) in
Exploration and survey of the valley of the Great Salt Lake of
Utah, including a reconnoissance of a new route through the Rocky
Mountains (H. Stansbury, J. W. Gunnison, and C. Preuss, eds.).
Lippincott, Grambo & Company, Philadelphia, Pennsylvania.
BAIRD, S. F. 1855. Characteristics of some new species of North
American Mammalia, collected by the U.S. and Mexican Boundary
Survey, Major W. H. Emory, U.S.A., Commissioner. Proceedings
of the Academy of Natural Sciences of Philadelphia 7:331–333.
BAKER, R. H. 1956. Mammals of Coahuila, México. University of
Kansas Publications, Museum of Natural History 9:125–335.
BAKER, R. H., AND J. K. GREER. 1962. Mammals of the Mexican state
of Durango. Publications of the Museum, Michican State University
Biological Series 2:25–154.
BERRY, D. L., AND R. J. BAKER. 1972. Chromosomes of pocket
gophers of the genus Pappogeomys, subgenus Cratogeomys.
Journal of Mammalogy 53:303–309.
CARPENTER, G., A. N. GILLISON, AND J. WINTER. 1993. DOMAIN:
a flexible modelling procedure for mapping potential distributions
of plants and animals. Biodiversity and Conservation 2:667–680.
DALQUEST, W. W. 1953. Mammals of the Mexican state of San Luis
Potosı́. Louisiana State University Studies, Biological Science
Series 1:1–229.
DALY, J. C., AND J. L. PATTON. 1986. Growth, reproduction and sexual
dimorphism in Thomomys bottae pocket gophers. Journal of
Mammalogy 67:256–265.
DAVIDOW-HENRY, B. R., J. K. JONES, JR., AND R. R. HOLLANDER. 1989.
Cratogeomys castanops. Mammalian Species 338:1–6.
DEMASTES, J. W., T. A. SPRADLING, M. S. HAFNER, D. J. HAFNER, AND
D. L. REED. 2002. Systematics and phylogeography of pocket
gophers in the genera Cratogeomys and Pappogeomys. Molecular
Phylogenetics and Evolution 22:144–154.
DEWALT, T. S., P. D. SUDMAN, M. S. HAFNER, AND S. K. DAVIS. 1993.
Phylogenetic relationships of pocket gophers (Cratogeomys and
Pappogeomys) based on mitochondrial DNA cytochrome b sequences. Molecular Phylogenetics and Evolution 2:193–204.
ENDLER, J. A. 1977. Geographic variation, speciation, and clines.
Monographs in Population Biology 10:1–246.
GANNON, W. L., AND R. S. SIKES, AND THE ANIMAL CARE AND USE
COMMITTEE OF THE AMERICAN SOCIETY OF MAMMALOGISTS. 2007.
Guidelines of the American Society of Mammalogists for the
use of wild mammals in research. Journal of Mammalogy 88:
809–823.
GEWEKE, J. 1992. Evaluating the accuracy of sampling-based
approaches to calculating posterior moments. Pp. 169–194 in
Bayesian statistics 4 (J. M. Bernardo, J. O. Berger, J. O. Dawid, and
A. F. M. Smith, eds.). Clarendon Press, Oxford, United Kingdom.
HAFNER, J. C., AND D. R. SANDQUIST. 1989. Postmortem field
preparation of bird and mammal chromosomes: an evaluation
involving the pocket gopher Thomomys bottae. Southwestern
Naturalist 34:330–337.
HAFNER, M. S., J. C. HAFNER, J. L. PATTON, AND M. F. SMITH. 1987.
Macrogeographic patterns of genetic differentiation in the pocket
gopher Thomomys umbrinus. Systematic Zoology 36:18–34.
HAFNER, M. S., J. E. LIGHT, D. J. HAFNER, S. V. BRANT, T. A.
SPRADLING, AND J. W. DEMASTES. 2005. Cryptic species in the
Mexican pocket gopher Cratogeomys merriami. Journal of
Mammalogy 86:1095–1108.
HAFNER, M. S., T. A. SPRADLING, J. E. LIGHT, D. J. HAFNER, AND J. R.
DEMBOSKI. 2004. Systematic revision of pocket gophers of the
Cratogeomys gymnurus species group. Journal of Mammalogy
85:1170–1183.
205
HALL, E. R. 1981. The mammals of North America. Vol. 1. John
Wiley & Sons, Inc., New York.
HALL, E. R., AND K. R. KELSON. 1959. The mammals of North
America. Vol. 1. Ronald Press Company, New York.
HARRIS, A. H. 1985. Late Pleistocene vertebrate paleoecology of the
West. University of Texas Press, Austin.
HEIDELBERGER, P., AND P. WELCH. 1983. Simulation run length control in the presence of an initial transient. Operations Research 31:
1109–1144.
HELLENTHAL, R. A., AND R. D. PRICE. 1976. Louse–host associations of
Geomydoecus (Mallophaga: Trichodectidae) with the yellow-faced
pocket gopher, Pappogeomys castanops (Rodentia: Geomyidae).
Journal of Medical Entomology 13:331–386.
HIJMANS, R. J., S. E. CAMERON, J. L. PARRA, P. G. JONES, AND A.
JARVIS. 2005. Very high resolution interpolated climate surfaces
for global land areas. International Journal of Climatology 25:
1965–1978.
HIJMANS, R. J., L. GUARINO, M. CRUZ, AND E. ROJAS. 2001. Computer
tools for spatial analysis of plant genetic resources data: 1. DIVAGIS. Plant Genetic Resources Newsletter 127:15–19.
HOLLANDER, R. R. 1990. Biosystematics of the yellow-faced pocket
gopher, Cratogeomys castanops (Rodentia: Geomyidae) in the
United States. Special Publications, The Museum, Texas Tech
University 33:1–62.
HONEYCUTT, R. L., AND S. L. WILLIAMS. 1982. Genic differentiation in
pocket gophers of the genus Pappogeomys, with comments on
intergeneric relationships in the subfamily Geomyinae. Journal of
Mammalogy 63:208–217.
HOOPER, E. T. 1946. Two genera of pocket gophers should be
congeneric. Journal of Mammalogy 27:397–399.
IRWIN, D. M., T. D. KOCKER, AND A. C. WILSON. 1991. Evolution of
the cytochrome-b gene of mammals. Journal of Molecular
Evolution 32:128–144.
LEE, H. K., AND R. J. BAKER. 1987. Cladistical analysis of
chromosomal evolution in pocket gophers of the Cratogeomys
castanops complex (Rodentia: Geomyidae). Occasional Papers, The
Museum, Texas Tech University 114:1–15.
LIDICKER, W. Z., JR. 1960. An analysis of intraspecfic variation in the
kangaroo rat Dipodomys merriami. University of California
Publications in Zoology 67:125–218.
LIDICKER, W. Z., JR. 1962. The nature of subspecies boundaries in
a desert rodent and its implications for subspecies taxonomy.
Systematic Zoology 11:160–171.
MANIS. 2007. Mammal Networked Information System. http://
manisnet.org. Accessed October 2007.
MERRIAM, C. H. 1895. Monographic revision of the pocket gophers,
family Geomyidae (exclusive of the species of Thomomys). North
American Fauna 8:1–220.
MILLER, G. S., JR. 1912. List of North American land mammals in the
United States National Museum, 1911. Bulletin of the United States
National Museum 79:1–455.
MORAFKA, D. J. 1977. Is there a Chihuahuan Desert? A quantitative
evaluation through a herpetofaunal perspective. Pp. 437–454 in
Transactions of the Symposium on the Biological Resources of the
Chihuahuan Desert Region, United States and Mexico (R. H. Wauer
and D. H. Riskind, eds.). National Park Service Transactions and
Proceedings Series 3:1–658.
MORRONE, J. J. 2005. Hacia una sı́ntesis biogeográfica de México.
Revista Mexicana de Biodiversidad 76:207–252.
NELSON, E. W., AND E. A. GOLDMAN. 1934. Revision of the pocket
gophers of the genus Cratogeomys. Proceedings of the Biological
Society of Washington 47:135–154.
206
Vol. 89, No. 1
JOURNAL OF MAMMALOGY
NEVO, E. 1979. Adaptive convergence and divergence of subterranean
mammals. Annual Review of Ecology and Systematics 10:269–308.
PATTON, J. L. 1985. Population structure and the genetics of speciation
in pocket gophers, genus Thomomys. Acta Zoologica Fennica
170:109–114.
PATTON, J. L. 2005. Family Geomyidae. Pp. 859–870 in Mammal
species of the world: a taxonomic and geographic reference (D. E.
Wilson and D. M. Reeder, eds.). 3rd ed. Johns Hopkins University
Press, Baltimore, Maryland.
PATTON, J. L., AND P. V. BRYLSKI. 1987. Pocket gophers in alfalfa
fields: causes and consequences of habitat-related body size
variation. American Naturalist 130:493–506.
PATTON, J. L., AND M. F. SMITH 1990. The evolutionary dynamics of
the pocket gopher Thomomys bottae, with emphasis on California
populations. University of California Publications in Zoology
123:i–viiþ 1–161.
PATTON, J. L., M. F. SMITH, R. D. PRICE, AND R. A. HELLENTHAL. 1984.
Genetics of hybridization between the pocket gophers Thomomys
bottae and Thomomys townsendii in northeastern California. Great
Basin Naturalist 44:431–440.
PETERSON, M. K. 1976. The Rio Nazas as a factor in mammalian
distribution in Durango, Mexico. Southwestern Naturalist 20:
495–502.
POSADA, D., AND K. A. CRANDALL. 1998. MODELTEST: testing the
model of DNA substitution. Bioinformatics 14:817–818.
PRYCHITKO, T. M., AND W. S. MOORE. 1997. The utility of DNA
sequences of an intron from the b-fibrinogen gene in phylogenetic
analysis of woodpeckers (Aves: Picidae). Molecular Phylogenetics
and Evolution 8:193–204.
RAFTERY, A. E., AND S. M. LEWIS. 1992a. Comment: one long run with
diagnostics: implementation strategies for Markov chain Monte
Carlo. Statistical Science 7:493–497.
RAFTERY, A. E., AND S. M. LEWIS. 1992b. How many iterations in the
Gibbs sampler? Pp. 763–764 in Bayesian statistics 4 (J. M.
Bernardo, J. O. Berger, J. O. Dawid, and A. F. M. Smith, eds.).
Clarendon Press, Oxford, United Kingdom.
RONQUIST, F., AND J. P. HUELSENBECK. 2003. MrBayes 3: Bayesian
phylogenetic inference under mixed models. Bioinformatics
19:1572–1574.
RUSSELL, R. J. 1968. Revision of pocket gophers of the genus
Pappogeomys. University of Kansas Publications, Museum of
Natural History 16:581–776.
RUSSELL, R. J., AND R. H. BAKER. 1955. Geographic variation in the
pocket gopher, Cratogeomys castanops, in Coahuila, México.
University of Kansas Publications, Museum of Natural History
7:591–608.
SCHMIDLY, D. J. 1977. Factors governing the distribution of mammals
in the Chihuahuan Desert region. Pp. 163–192 in Transactions of
the Symposium on the Biological Resources of the Chihuahuan
Desert Region, United States and Mexico (R. H. Wauer and D. H.
Riskind, eds.). National Park Service Transactions and Proceedings
Series 3:1–658.
SMITH, B. J. 2005. The BOA package: Bayesian Output Analysis
Program (BOA) for MCMC, version 1.1.3-1. Available for
download via the The Comprehensive R Archive Network. http://
cran.r-project.org/. Accessed October 2007.
SMITH, H. M. 1941. Las provincias bióticas de México, según la
distribución geográfica de las lagartijas del género Sceloporus.
Anales de la Escuela Nacional de Ciencias Biológicas 2:103–110.
SMITH, M. F., AND J. L. PATTON. 1988. Subspecies of pocket gophers:
causal bases for geographic differentiation in Thomomys bottae.
Systematic Zoology 37:163–178.
SMITH, M. F., AND J. L. PATTON. 1993. The diversification of SouthAmerican murid rodents—evidence from mitochondrial-DNA
sequence data for the akodontine tribe. Biological Journal of the
Linnean Society 50:149–177.
SPRADLING, T. A., S. V. BRANT, M. S. HAFNER, AND C. J. DICKERSON.
2004. DNA data support a rapid radiation of pocket gopher
genera (Rodentia: Geomyidae). Journal of Mammalian Evolution
11:105–125.
SPRADLING, T. A., M. S. HAFNER, AND J. W. DEMASTES. 2001.
Differences in rate of cytochrome-b evolution among species of
rodents. Journal of Mammalogy 82:65–80.
SWOFFORD, D. L. 2002. PAUP*: phylogenetic analysis using
parsimony (* and other methods). Version 4.0b10. Sinauer
Associates, Inc., Publishers, Sunderland, Massachusetts.
SWOFFORD, D. L., G. J. OLSEN, P. J. WADDELL, AND D. M. HILLIS. 1996.
Phylogenetic inference. Pp. 407–514 in Molecular systematics
(D. M. Hillis, C. Moritz, and B. K. Mable, eds.). Sinauer
Associates, Inc., Publishers, Sunderland, Massachusetts.
THAELER, C. S. 1968. Karyotypes of sixteen populations of the
Thomomys talpoides complex of pocket gophers (Rodentia: Geomyidae). Chromosoma 25:172–183.
WILKINSON, L. 1997. Systat 7.0 for Windows. SPSS, Inc., Chicago,
Illinois.
Submitted 2 January 2007. Accepted 2 May 2007.
Associate Editor was Jesús Maldonado.
APPENDIX I
The 67 specimens of Cratogeomys castanops from 35 localities and
132 specimens of C. goldmani from 47 localities (total of 199
specimens from 82 localities plotted in Figs. 2 and 3D) examined in
morphometric analyses are listed below by subspecies, locality, and
museum acronym. Locality numbers are given in parentheses.
Collection acronyms are as follows: Colección Nacional de Mamı́feros, Universidad Nacional Autónoma de México (CNMA); Museum
of Natural History, University of Kansas (KU); Museum of Natural
Science, Louisiana State University (LSUMZ); Michigan State
University Museum (MSU); Museum of Vertebrate Zoology,
University of California (MVZ); New Mexico Museum of Natural
History (NMMNH); Museum of Texas Tech University (TTU);
Museum of Zoology, University of Michigan (UMMZ); and United
States National Museum of Natural History (USNM).
Cratogeomys castanops castanops (n ¼ 3).— (11) Coahuila: 1.7 km
N Primero de Mayo, 390 m (LSUMZ 36449–36451).
Cratogeomys castanops consitus (n ¼ 64).— (10) Coahuila: 1.5
miles NW Ocampo, 3,300 feet (NMMNH 3626, 3628); (13) Coahuila:
2 km (by road) NW Cuatro Ciénegas, 776 m (LSUMZ 36456, 36457);
(14) Coahuila: Santa Teresa de Sofı́a, 2,500 feet (NMMNH 3615);
(15) Coahuila: Hisachalo [¼ Huisachalo] (KU 58078, 58079); (16)
Coahuila: 3 miles S, 3 miles E La Muralla (KU 48513); (17) Coahuila:
Plan de Guadalupe, 1,040 m (LSUMZ 36448); (18) Coahuila: 46 miles
NE San Pedro de las Colonias (TTU 10504); (19) Durango: 4 miles
NNE Boquilla (MSU 3341); (20) Durango: Tlahualilo (USNM
246530); (21) Durango: 7 miles NNW La Zarca, 5,700 feet (NMMNH
2467); (22) Durango: 7 miles NW La Zarca (MSU 3344, 3346–3348);
(24) Durango: La Zarca (KU 62469, 62470); (25) Durango: 12 miles
E La Zarca (KU 62467, 62468); (26) Coahuila: 3 miles N Santa Cruz,
32 miles (by road) N Saltillo (KU 48516); (27) Coahuila: San Pedro
(¼ San Pedro de las Colonias) (USNM 246534, 246535); (28) Coahuila:
11 miles N, 10 miles W San Lorenzo (KU 48520); (29) Coahuila: 10
February 2008
HAFNER ET AL.—SYSTEMATICS OF CRATOGEOMYS
miles N, 11 miles W San Lorenzo (KU 48518); (30) Coahuila:
Hacienda El Tulillo, 5 km S Hipólito (KU 35772); (31) Coahuila: 12
miles N, 10 miles E Parras (KU 34937); (32) Coahuila: Jaral, 3,860
feet (¼ San Antonio de Jaral) (USNM 51046, 51047); (33) Coahuila:
0.25 miles SE San Antonio de Jaral (MVZ 76355–76358); (34)
Coahuila: 3 miles N, 5 miles W La Rosa (KU 48519); (35) Coahuila: 2
miles E Torreón (KU 40223, 40225, 40226, 40228–40231, MSU
165); (36) Durango: 4 miles WSW Lerdo (KU 40234, MSU 166); (38)
Coahuila: San Lorenzo, 1,380 m (NMMNH 5105); (40) Durango: 10
miles WSW Lerdo, San Juan, 3,800 feet (UMMZ 90004); (41)
Coahuila: 21 km W Saltillo on Hwy. 40 (TTU 10503); (42) Coahuila:
7 km S, 14 km W General Cepeda, 1,710 m (LSUMZ 36446, 36447);
(43) Coahuila: 10 miles S, 5 miles W General Cepeda, N foot of Sierra
Guadalupe (KU 55586); (44) Durango: 11 miles N Rodeo, Rı́o Nazas
(TTU 12065); (45) Durango: 6 miles NW Rodeo, Rı́o Nazas (KU
62472, 62475, 62477, 62479, 62480); (46) Durango: 2 km SW La
Unión (NMMNH 5240); (58) Nuevo León: Montemorelos, 2,500 feet
(USNM 116845, 116846, 116849, 116850); (60) Durango: Hacienda
Atotonilco, 1,976 m (KU 67622, NMMNH 4482).
Cratogeomys goldmani goldmani (n ¼ 46).— (61) Coahuila: 1 km
NW La Flor de Jimulco, 1,295 m (NMMNH 5070, 5071, 5073); (65)
Zacatecas: 10 km ESE Charcos (¼ San Juan de los Charcos) (MSU
25132); (76) Zacatecas: 10.5 miles S Concepción del Oro (Hwy. 54)
(TTU 45088, 45089, 45091, 45093, 45094); (77) Zacatecas: 15 miles
S Concepción del Oro (KU 58129, 58130); (78) Zacatecas: 25 km S
Concepción del Oro, 1,864 m (LSUMZ 36437, 36439–36441); (80)
Zacatecas: 4 km N Nieves (MSU 26441, 26443); (81) Zacatecas: 30
km NW Rı́o Grande, 2,095 m (NMMNH 5076, 5077); (82) Zacatecas:
6 km SE Tetillas (MSU 30003); (83) Zacatecas: 8 miles S Majoma
(KU 58133, 58135); (86) Zacatecas: 0.5 km N Cañitas, 2,018 m
(NMMNH 5074); (87) Zacatecas: Cañitas de Felipe Pescador (USNM
57965 [holotype], 57966, 57968); (88) Zacatecas: 1 mile S Cañitas
(TTU 11994); (89) Zacatecas: 11 miles SW Cañitas (TTU 9701–9703,
9705); (91) Zacatecas: Villa de Cos (KU 58139, 58146, 58148–
58150); (92) Zacatecas: 1 mile (by road) SW Villa de Cos (TTU 9707,
9708, 10145); (93) Zacatecas: 45 km (by road) NE Morelos Junction,
Rancho El Amarillo (TTU 8720, 8722, 9268); (95) San Luis Potosı́:
Presa de Guadalupe (LSUMZ 5089).
Cratogeomys goldmani subnubilus (n ¼ 86).— (47) Coahuila: 4
miles S, 6 miles E Saltillo, Sierra Guadalupe (KU 35777, 35778); (48)
Coahuila: 7 miles S, 4 miles E Bella Unión (KU 48526, 48527,
48529); (49) Coahuila: 12 miles S, 2 miles E Arteaga (KU 33117,
33119–33121, 33124, 33125); (50) Coahuila: 1 mile N, 14 miles W
San Antonio de las Alazanas (KU 58086); (51) Coahuila: 12 miles W
San Antonio de las Alazanas (KU 58081, 58084, 58089–58091,
58092); (52) Coahuila: 1 mile N Agua Nueva, 1,922 m (KU 33127);
(53) Coahuila: 2 miles E Agua Nueva, 2,017 m (LSUMZ 36444,
36445); (54) Coahuila: 17 km S, 16 km W General Cepeda, 2,064 m
(LSUMZ 36434, NMMNH 5081, 5083, 5090); (55) Coahuila: 11
miles S, 4 miles W General Cepeda, Sierra Guadalupe (KU 55587);
(56) Coahuila: 19 km S, 17 km W General Cepeda, 2,115 m
(NMMNH 4641); (57) Coahuila: 19 km S, 18 km W General Cepeda,
2,091 m (NMMNH 5087, 5089, 5091, 5095, 5097–5099); (59) Nuevo
León: 7 miles NW La Providencia (KU 100449, 100451); (62)
Coahuila: Carneros, 6,800 feet (USNM 79485); (63) Coahuila: 1 mile
S Carneros (KU 33128); (66) Coahuila: 2 miles W San Miguel (KU
33132); (67) Nuevo León: Laguna (KU 58094, 58096, 58097, 58099);
(68) Coahuila: 8 miles N La Ventura (KU 33136, 34590, 34930,
34933–34936); (70) Zacatecas: Concepción del Oro (KU 58121–
58124, 58126); (71) Coahuila: La Ventura, 5,600 feet and 5,650 feet
(USNM 79477, 79479, 79481, 79489, 79491, 79494); (72) Nuevo
León: 8 miles (by road) S San Roberto Junction (TTU 8309); (73)
207
Zacatecas: 7 km W San Felipe de Nuevo Mercurio (Rancho San
Marcos) (MSU 25294); (74) Zacatecas: 3 miles N Lulú (Estación
Lulú in San Luis Potosı́) (MVZ 91265, 91266, 91271, 91274); (75)
Nuevo León: 16 km S, 2 km E Pablillo (MVZ 158036); (79) Nuevo
León: 5 miles W Ascensión (KU 58100–58105, 58107–58109,
58111–58113); (84) San Luis Potosı́: 11.5 miles N Matehuala
(TTU 57067); (85) San Luis Potosı́: 6 km S Matehuala (LSUMZ
5084); (90) Nuevo León: 1 mile W Doctor Arroyo (KU 58117); (96)
San Luis Potosı́: 10 km N Tepeyac (TTU 15607–15609); (97) San
Luis Potosı́: 5.7 miles E junction Hwy. 80 and 101, near Tepeyac
(TTU 15611, 15612); (100) San Luis Potosı́: 6 km E Rı́o Verde, 3350
feet (LSUMZ 36087).
APPENDIX II
The 17 specimens of Cratogeomys castanops from 11 localities and
62 specimens of C. goldmani from 16 localities (total of 79 specimens
from 27 localities plotted in Figs. 2 and 3B) examined in the
chromosomal analysis are listed below by subspecies, locality, and
museum acronym (listed in Appendix I). Locality numbers are given
in parentheses. TTU specimens are reported in Berry and Baker (1972)
or Lee and Baker (1987).
Cratogeomys castanops consitus (n ¼ 17).— (12) Coahuila: 12
miles N Monclova (TTU, 1); (18) Coahuila: 46 miles NE San Pedro de
las Colonias (TTU, 1); (23) Durango: 8 km N La Zarca Junction
(TTU, 1); (37) Durango: 5 km SW Lerdo, 1,158 m (NMMNH 4472);
(38) Coahuila: San Lorenzo, 1,380 m (NMMNH 5104–5105); (39)
Coahuila: 60 km W Saltillo on Hwy. 40 (TTU, 1); (41) Coahuila: 21
km W Saltillo on Hwy. 40 (TTU, 2); (42) Coahuila: 7 km S, 14 km W
General Cepeda, 1,710 m (LSUMZ 36446, 36447); (44) Durango: 11
miles N Rodeo, Rı́o Nazas (TTU, 3); (46) Durango: 2 km SW La
Unión (NMMNH 5240, 5241); (60) Durango: Hacienda Atotonilco,
1,976 m (NMMNH 4482).
Cratogeomys goldmani goldmani (n ¼ 34).— (61) Coahuila: 1 km
NW La Flor de Jimulco, 1,295 m (NMMNH 5069–5073); (69)
Zacatecas: 3.5 miles (by road) E Mazapil (TTU, 1); (76) Zacatecas:
10.5 miles S Concepción del Oro (Hwy. 54) (TTU, 11); (81)
Zacatecas: 30 km NW Rı́o Grande, 2,095 m (NMMNH 5076, 5077);
(86) Zacatecas: 0.5 km N Cañitas, 2,018 m (NMMNH 5074, 5075);
(88) Zacatecas: 1 mile S Cañitas (TTU, 1); (89) Zacatecas: 11 miles
SW Cañitas (TTU, 3); (92) Zacatecas: 1 mile (by road) SW Villa de
Cos (TTU, 2); (93) Zacatecas: 45 km (by road) NE Morelos Junction,
Rancho El Amarillo (TTU, 4); (94) Zacatecas: 20 miles (by road) NE
Morelos Junction (TTU, 3).
Cratogeomys goldmani subnubilus (n ¼ 28).— (54) Coahuila: 17
km S, 16 km W General Cepeda, 2,064 m (LSUMZ 36434, 36436,
NMMNH 5081–5085, 5090); (56) Coahuila: 19 km S, 17 km W
General Cepeda, 2,115 m (NMMNH 4641); (57) Coahuila: 19 km S,
18 km W General Cepeda, 2,091 m (NMMNH 5086–5089, 5091–
5098, 5100–5103); (72) Nuevo León: 8 miles (by road) S San Roberto
Junction (TTU, 1); (98) San Luis Potosı́: 35 km W Ciudad de Maı́z
(TTU, 1); (99) San Luis Potosı́: 12 miles W Ciudad de Maı́z (TTU, 1).
APPENDIX III
The 25 specimens of Cratogeomys castanops from 19 localities
and 35 specimens of C. goldmani from 10 localities (total of 60
specimens from 29 localities plotted in Figs. 2 and 3C) and 7 outgroups examined in mitochondrial DNA (mt), and nuclear DNA (nu)
analyses are listed below by subspecies, locality, and museum acronym (listed in Appendix I). Locality numbers are given in parentheses.
Each individual is represented by mt (cytochrome-b [Cytb] and
208
JOURNAL OF MAMMALOGY
cytochrome-c oxidase subunit I [CoI]) and nu sequences unless
otherwise noted.
Cratogeomys castanops castanops (n ¼ 9 [mt], 10 [nu]).— (1)
Oklahoma: Cimarron County; 1.5 miles S, 3 miles E Kenton (TTU
43257); (2) Texas: Moore County; 3 miles S Dumas (TTU 42767); (3)
New Mexico: Roosevelt County; 2.5 miles E Tolar, 1,306 m
(NMMNH 4341); (4) New Mexico: Lincoln County; 2.5 miles W
Ancho, 6,000 feet (LSUMZ 29324); (5) Texas: Cochran County; 0.5
miles W Morton, 1,172 m (NMMNH 4340); (6) New Mexico: Chaves
County; 6.5 miles W Caprock, 4,300 feet (NMMNH 4317); (7) New
Mexico: Otero County; 25 miles SW Alamogordo, 3,800 feet
(LSUMZ 31454, 31455 [nu only]); (8) Texas: Presidio County; Big
Bend Ranch State Natural Area (TTU 68426); (11) Coahuila: 1.7 km
N Primero de Mayo, 390 m (LSUMZ 36453).
Cratogeomys castanops consitus (n ¼ 13 [mt], 11 [nu]).— (9)
Chihuahua: Gallego, 1,627 m (NMMNH 5106); (10) Coahuila: 1.5
miles NW Ocampo, 3,300 feet (NMMNH 3626, 3628 [mt only],
CNMA 42291 [mt only]); (13) Coahuila: 2 km (by road) NW Cuatro
Ciénegas, 776 m (LSUMZ 36456); (14) Coahuila: Santa Teresa de
Sofı́a, 2,500 feet (NMMNH 3615); (17) Coahuila: Plan de Guadalupe,
1,040 m (LSUMZ 36448); (21) Durango: 7 miles NNW La Zarca,
5,700 feet (NMMNH 2467); (37) Durango: 5 km SW Lerdo, 1,158 m
(NMMNH 2488 [nu only], NMMNH 4472 [mt only]); (38) Coahuila:
San Lorenzo, 1,380 m (NMMNH 5104, 5105 [nu only]); (42)
Coahuila: 7 km S, 14 km W General Cepeda, 1,710 m (LSUMZ
36446, 36447 [mt only]); (60) Durango: Hacienda Atotonilco, 1,976
m (NMMNH 4482).
Cratogeomys goldmani goldmani (n ¼ 4 [mt], 4 [nu]).— (61)
Coahuila: 1 km NW La Flor de Jimulco, 1,295 m (NMMNH 5072);
Vol. 89, No. 1
(78) Zacatecas: 25 km S Concepción del Oro, 1,864 m (LSUMZ
36442); (81) Zacatecas: 30 km NW Rı́o Grande, 2,095 m (NMMNH
5078); (86) Zacatecas: 0.5 km N Cañitas, 2,018 m (NMMNH 5075).
Cratogeomys goldmani subnubilus (n ¼ 29 [mt], 29 [nu]).— (52)
Coahuila: 1 mile N Agua Nueva, 1,922 m (LSUMZ 36443 [nu only]);
(53) Coahuila: 2 miles E Agua Nueva, 2,017 m (LSUMZ 36444 [mt
only]); (54) Coahuila: 17 km S, 16 km W General Cepeda, 2,064 m
(LSUMZ 36434–36436, NMMNH 5081–5085, 5090); (57) Coahuila:
19 km S, 18 km W General Cepeda, 2,091 m (NMMNH 5086–5089,
5091–5103); (64) Coahuila: 44 km SSW Saltillo, 6,500 feet (CNMA
42292); (100) San Luis Potosı́: 6 km E Rı́o Verde, 3,350 feet (LSUMZ
36086 [mt only], CNMA 39924 [nu only]).
Cratogeomys fumosus angustirostris.— México: Jalisco; 3 km NE
Lagos de Moreno, 6,150 feet (LSUMZ 36084, GenBank AY331239
[nu]).
Cratogeomys fumosus fumosus.— México: Colima; 5 km S Colima,
1,000 feet (CNMA 39925, GenBank AY331075 [mt CoI], GenBank
AY331240 [nu]).
Cratogeomys fumosus imparilis.— México: Michoacán; 1 km S
Tacambaro, 5,100 feet (LSUMZ 36129, GenBank AF302179 [mt Cytb]).
Cratogeomys fumosus tylorhinus.— México: Queretero; La Cañada,
9 km by road SW Pinal de Amoles, 9,000 feet (CNMA 39926,
GenBank AY331238 [nu]).
Cratogeomys merriami.— México: México; 5 km SSW Texcoco,
7,000 feet (LSUMZ 36065, GenBank AY331078 [mt CoI and Cytb],
GenBank AY331243 [nu]).
Cratogeomys planiceps.— México: México; 10 km S, 16 km W
Toluca, 3,000 m (LSUMZ 34901, GenBank AY545541 [mt Cytb],
LSUMZ 36121, AF302183 [mt Cytb]).

Download Report

cratogeomys castanops species group

Paperzz.com

Your Paperzz