Phylogeographic diversiﬁcation of antelope squirrels ... · Phylogeographic diversiﬁcation of...

Phylogeographic diversification of antelope squirrels(Ammospermophilus) across North American deserts

STACY J. MANTOOTH1,2, DAVID J. HAFNER3,*, ROBERT W. BRYSON, JR4 andBRETT R. RIDDLE1

1School of Life Sciences, University of Nevada, Las Vegas, 4505 S Maryland Parkway, Las Vegas, NV89154-4004, USA2Department of Physical and Life Sciences, Nevada State College, 1125 Nevada State Dr., Henderson,NV 89002, USA3Museum of Southwestern Biology, University of New Mexico, Albuquerque, NM 87131, USA4Department of Biology & Burke Museum of Natural History and Culture, University of Washington,Box 351800, Seattle, WA 98195-1800, USA

Received 13 November 2012; revised 13 February 2013; accepted for publication 14 February 2013

We investigated the biogeographic history of antelope squirrels, genus Ammospermophilus, which are widelydistributed across the deserts and other arid lands of western North America. We combined range-wide samplingof all currently recognized species of Ammospermophilus with a multilocus data set to infer phylogeneticrelationships. We then estimated divergence times within identified clades of Ammospermophilus using fossil-calibrated and rate-calibrated molecular clocks. Lastly, we explored generalized distributional changes of Ammo-spermophilus since the last glacial maximum using species distribution models, and assessed responses toQuaternary climate change by generating demographic parameter estimates for the three wide-ranging clades ofA. leucurus. From our phylogenetic estimates we inferred strong phylogeographic structure within Ammosper-mophilus and the presence of three well-supported major clades. Initial patterns of historical divergence werecoincident with dynamic alterations in the landscape of western North America, and the formation of regionaldeserts during the Late Miocene and Pliocene. Species distribution models and demographic parameter estimatesrevealed patterns of recent population expansion in response to glacial retreat. When combined with evidence fromco-distributed taxa, the historical biogeography of Ammospermophilus provides additional insight into the mecha-nisms that impacted diversification of arid-adapted taxa across the arid lands of western North America. Wepropose species recognition of populations of the southern Baja California peninsula to best represent our currentunderstanding of evolutionary relationships among genetic units of Ammospermophilus. © 2013 The LinneanSociety of London, Biological Journal of the Linnean Society, 2013, 109, 949–967.

ADDITIONAL KEYWORDS: biogeography – divergence dating – Neogene – North America – Pleistocene –species distribution models – vicariance.

INTRODUCTION

Genetic differentiation within and between speciesoften coincides with significant geological or climaticchanges that have shaped species ranges and alteredthe connectivity between populations over time.Within the North American deserts, many endemictaxa experienced high levels of initial divergence

associated with the geological transformations of theNeogene, with subsequent diversification and geo-graphical structuring of populations associated withclimatic changes during the Quaternary (Riddle,1995; Hafner & Riddle, 1997, 2005, 2011). Climaticoscillations throughout the Pleistocene led torepeated cycles of glacial expansion and retreat, withexpansion reaching its maximum extent duringthe last glacial maximum (LGM), approximately21 000 BP. The distributions of many arid-land (ageneral term we use to describe deserts as well as*Corresponding author. E-mail: [email protected]

bs_bs_banner

Biological Journal of the Linnean Society, 2013, 109, 949–967. With 6 figures

© 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 109, 949–967 949

semi-arid shrublands and grasslands) species showgenetic signatures of a history of population isolationand reconnection, as well as distribution changesresulting from glacial cycles, with the most recentexpansion of xeric habitats following the LGM. Com-parative analyses of similarly distributed specieshave identified complex patterns of genetic relation-ships within and between areas of endemism acrossthe major warm deserts (Riddle & Hafner, 2006), butwith an underlying signature of congruent geneticdiscontinuities between evolutionarily distinct line-ages across taxa at several pronounced biogeographicbarriers. These complex patterns are the results ofboth shared and unique responses to various eventsisolating and reconnecting populations.

The biogeographic history of Ammospermophilus(antelope squirrels) represents an opportunity to addto the growing literature that addresses the develop-ment and assembly of the arid-lands biota of NorthAmerica. Ammospermophilus represents a distinctgenus of four (following Álvarez-Castañeda, 2007) orfive extant species within the rodent family Sciuridaethat are distributed across the deserts and arid lands

of western North America (Hall, 1981; Wilson &Reeder, 2005). Ammospermophilus first appears inthe fossil record during the mid-Miocene, 11.5 Mya, inthe Cuyama Valley of southern California (James,1963), prior to the expansion of the semi-desert eco-systems during the Pliocene. Black (1963) provision-ally assigned material of a similar age (ClarendonianNorth American Land Mammal Age, NALMA, 13.6–10.3 Mya) from eastern Oregon to Ammosper-mophilus, while noting the contrary opinion ofShotwell & Russell (1963). Miller (1980) reported theearly Pliocene (Blancan NALMA, 4.75–1.8 Mya)Ammospermophilus jeffriesi from the Cape Region,southernmost Baja California peninsula, andGustafson (1978) reported the early Pliocene(Blancan) Ammospermophilus hanfordi from south-central Washington. The depth of this fossil historysuggests a causal association between the formationof North American regional deserts and the origin anddiversification of Ammospermophilus, which is foundthroughout, and is restricted to, most of the desertand semi-desert regions in North America (except thesouthern portion of the Chihuahuan Desert; Fig. 1).

Figure 1. (A) Distribution of North American regional deserts (shaded area; following Shreve, 1942; modified accordingto Hafner & Riddle, 1997, 2011): Great Basin, San Joaquin, Mojave, Peninsular, Sonoran, and Chihuahuan. Abbrevia-tions: AZ, Arizona; BC, Baja California; BCS, Baja California Sur; CA, California; CFB, Cochise filter-barrier (Morafka,1977); CHI, Chihuahua; ID, Idaho; ISM, Isla San Marcos; MH, Mesa Huatamote (Hafner & Riddle, 2011); NM, NewMexico; OR, Oregon; SCFB, Southern Coahuila filter-barrier (Baker, 1956); SON, Sonora; ST, Salton Trough; TX, Texas;UT, Utah; VS, Vizcaíno Seaway (Upton & Murphy, 1997). (B) Distribution of Ammospermophilus species (light and darkshaded areas; Hall, 1981; modified according to Hall, 1946; Armstrong, 1972; Findley et al., 1975; Schmidly, 1977; Hafner,1981; Hoffmeister, 1986; Verts & Carraway, 1998; Yensen & Valdés-Alarcón, 1999; and Geluso, 2009). White circlesindicate the localities sampled in this study (numbered in Figure 3 and listed in Appendix S1).

950 S. J. MANTOOTH ET AL.

© 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 109, 949–967

The white-tailed antelope squirrel, Ammosper-mophilus leucurus, is the most widespread member ofthis genus, occurring from the northern Great Basinto the southern tip of the Baja California peninsula,and into central New Mexico. This distributionencompasses an ecologically broad area throughoutboth warm desert regions in the south and cold shrub-steppe regions in the north.

Recent molecular evidence suggests that popula-tions of A. leucurus from the northern Baja Californiapeninsula expanded northwards into the continentaldeserts, and following an episode of isolation, formeda lineage distinct from a southern peninsular lineagethat includes both A. leucurus and Ammosper-mophilus insularis (Riddle et al., 2000a; Whorley,Álvarez-Castañeda & Kenagy, 2004; Álvarez-Castañeda, 2007). Based on a preliminary mitochon-drial DNA (mtDNA) analysis with small samplesizes (Riddle et al., 2000a), these northern popula-tions probably share a more recent common evolu-tionary history with Ammospermophilus harrisii, amorphologically distinct species that is geographicallyseparated by the Colorado River, indicating thatA. leucurus may represent a paraphyletic assemblagewith regard to A. harrisii and A. insularis. SeparatemtDNA sequence and morphological analyses of A. in-sularis, on Isla Espíritu Santo in the Gulf of Califor-nia, and a population of A. leucurus, on Isla SanMarcos in the Gulf of California (Álvarez-Castañeda,2007; Fig. 1B), indicated that these insular forms mayboth represent recently isolated populations of A. leu-curus, and Álvarez-Castañeda (2007) recommendedthat A. insularis be considered a subspecies of A.leucurus (Ammospermophilus leucurus insularis),whereas the population on Isla San Marcos repre-sented the mainland subspecies Ammospermophilusleucurus extimus. Álvarez-Castañeda (2007) identifiedone locality (Volcán Tres Vírgenes) where northernand southern haplotypes were found in sympatry (oneindividual each), and suggested that the two lineagesmay be morphologically as well as genetically diag-nosable, and may represent distinct species. Collec-tively, these studies suggest that the biogeographichistory of Ammospermophilus is more complex thanthat suggested by current taxonomy.

Although previous studies have examined the phy-logeography of some of the currently recognizedspecies of Ammospermophilus (Riddle et al., 2000a;Whorley et al., 2004; Álvarez-Castañeda, 2007), todate none have included individuals from all speciesand from across the broad geographical ranges of themore widespread species. Here, we combine range-wide sampling of all Ammospermophilus species witha broad multilocus data set and a tiered methodologyto present a comprehensive examination of the bio-geographic and evolutionary history of this genus,

and revise the species-level taxonomy to best reflectcurrently understood relationships among identifiedclades.

MATERIAL AND METHODSTAXONOMIC SAMPLING

We collected tissue from 86 Ammospermophilus speci-mens, including representatives of all five nominalspecies (Appendix S1). We included geographicallywidespread and representative samples for A. leucu-rus (62 individuals from 28 localities), A. harrisii(nine individuals from four localities), and Ammosper-mophilus interpres (ten individuals from five locali-ties). The two species with restricted distributions,Ammospermophilus nelsoni and A. insularis, wererepresented by two and three individuals, respec-tively. The localities sampled are mapped inFigures 1B and 3, and are listed in Appendix S1.Based on previous molecular research into the higher-level systematic relationships within Sciuridae(Harrison et al., 2003; Mercer & Roth, 2003; Herron,Castoe & Parkinson, 2004), we used Cynomys gun-nisoni and Xerospermophilus tereticaudus as out-groups (Appendix S1).

LABORATORY PROTOCOLS

We extracted total genomic DNA from liver or kidneytissues following either a lysis buffer protocol(Longmire, Maltbie & Baker, 1997) or a QiagenDNeasy Tissue Extraction Kit (Qiagen Inc.). Toexplore range-wide geographical variation in Ammo-spermophilus, we sequenced the mitochondrial genescytochrome oxidase 3 (CO3) and a section of D-loopwithin the control region (CR) for all 86 samples ofAmmospermophilus and the two out-group taxa. Wethen sequenced six additional genes for 22 individualsof Ammospermophilus, representing each major cladeinferred from the analyses of the complete CO3 andCR data. These additional six genes included twonuclear markers, exon 1 of the interphotoreceptorretinoid-binding protein (IRBP) and the recombina-tion activating gene 2 (RAG2), and four mitochondrialgenes, including cytochrome oxidase 1 (CO1, theputative animal DNA barcoding gene), cytochrome b(Cytb), the small subunit 12S ribosomal RNA (12S),and the large subunit 16S ribosomal RNA (16S). Weamplified all genes using the polymerase chain reac-tion (PCR) with gene-specific primers and tempera-ture profiles (Appendix S2). Double-stranded PCRproducts were qualitatively examined using a 0.8%agarose gel. The amplified PCR fragments were puri-fied using the GeneClean II Kit (BIO 101 Inc.),Qiaquick PCR Purification Kit (Qiagen Inc.), or Exo-SAP IT (USB Corp.), following the manufacturers’

DIVERSIFICATION OF AMMOSPERMOPHILUS 951


protocols. The purified PCR fragments (including boththe light and heavy DNA strands) were sequencedusing the ABI PRISM BigDye v3.1 Cycle Sequencingchemistry (Applied Biosystems Inc.), using thesequencing primers identified in Appendix S2. Unin-corporated dye terminators were removed usingSephedex spin columns (Centri-Sep Inc.), andsequence data were generated on either an ABI 310 or3130 Genetic Analyzer (Applied Biosystems Inc). Weunambiguously aligned complementary strands ofeach gene using SEQUENCHER 4.9 (Gene CodesCorp.), followed by manual proofreading. The protein-coding sequences were translated into amino acidsusing MACCLADE 4 (Maddison & Maddison, 2005),and were compared with Rattus and Mus to confirmthe correct reading frame and to check for the pres-ence of stop codons. All aligned sequences were depos-ited in the Dryad repository (http://www.datadryad.org/; Appendix S1).

MATERNAL PHYLOGENY

We assessed range-wide geographical variation inAmmospermophilus using maximum likelihood (ML)and Bayesian inference (BI) phylogenetic methods.For these analyses, variation within a portion of theprotein-coding CO3 (691 bp) and non-coding CR(503 bp) from the mitochondrial genome was exam-ined. We used JMODELTEST 0.1.1 (Guindon &Gascuel, 2003; Posada, 2008) to identify the mostappropriate model of nucleotide evolution chosenunder the Akaike’s information criterion (AIC) for theconcatenated data set. We conducted ML phylogeneticanalysis in TREEFINDER 2008 (Jobb, von Haeseler& Strimmer, 2004) with non-parametric bootstrap-ping (100 replicates; Felsenstein, 1985). The BI analy-ses were implemented in MRBAYES 3.1.2 (Ronquist& Huelsenbeck, 2003). We ran analyses for 10 ¥ 106

generations with an initial burn-in of 2 ¥ 106 genera-tions (25 000 trees), with four Monte Carlo Markovchains, and a temperature value of 0.05. Convergenceof runs was estimated by examining the posteriorprobabilities of clades for non-overlapping samples oftrees using AWTY (Wilgenbusch, Warren & Swofford,2004).

MULTILOCUS PHYLOGENY

We examined higher-level relationships of all nominalspecies and the major genetic lineages of Ammosper-mophilus using ML and BI analyses of our multilocusdata set. We used JMODELTEST to select the appro-priate models of sequence evolution for each of theseven genes. TREEFINDER was used to perform theML analyses and calculate bootstrap values after1000 replicates. BI analyses were run for 4 ¥ 106

generations using the default parameters of fourMarkov chains per generation, with random startingtrees, and subsequent trees sampled every 100 gen-erations. We assessed the stationarity of the analysesby examining the stabilization of cold-chain likelihoodscores and parameter estimates using TRACER 1.4.1(Rambaut & Drummond, 2007). The convergence ofruns was assessed by examining the posterior prob-abilities of clades for non-overlapping samples of treesusing AWTY. After excluding the trees generatedduring the ‘burn-in’ period prior to stable equilibrium(10 000 trees), we generated a 50% majority-ruleconsensus tree.

DIVERGENCE TIMES

We estimated divergence times within Ammosper-mophilus from our multilocus data set using anuncorrelated lognormal relaxed clock model imple-mented in BEAST 1.4.8 (Drummond et al., 2006;Drummond & Rambaut, 2007). The data set waspartitioned into three separate alignments: one align-ment contained the concatenated mtDNA data set(four protein-coding and two ribosomal genes, alllinked within the mitochondrial genome); a secondalignment with the IRBP data; and a third alignmentwith the RAG2 sequences. JMODELTEST was usedto select a model of sequence evolution for each sepa-rate alignment, and a Yule process speciation modelwas used to set the prior on the tree.

We used fossil data to calibrate our phylogeny,setting Ammospermophilus fossilis from the Claren-donian NALMA of the mid-Miocene (James, 1963) asthe Ammospermophilus stem-node calibration point.By rooting at the stem of Ammospermophilus, theminimum constraint on the out-group node wasestablished, yielding a conservative estimate forminimum divergence time within Ammosper-mophilus. This stem placement was further sup-ported by a mid-Miocene estimate of the divergence ofAmmospermophilus from the out-group taxa(Cynomys and Xerospermophilus) generated fromseveral independent and external fossil calibrationsthroughout the sciurid phylogeny (Mercer & Roth,2003). Analyses in BEAST were run for 4 ¥ 107 gen-erations, sampling every 1000 generations. The first4 ¥ 106 (10%) generations were discarded as burn-in.TRACER was used to verify proper mixing of thechains and to ensure that the analyses reached sta-tionarity. To increase the effective sample size (ESS)values, the analyses were repeated, and the data fromthe two separate runs were combined.

Two additional calibration methods were tested tocompare divergence estimates within the phylogeny.Using the same BEAST methods, the fossil calibra-tion was placed at the basal node (crown group) of



Ammospermophilus divergence. We additionally useda mutation rate-calibrated estimation of divergencetimes performed on just the Cytb data set. We usedthe standard mutation rate of 2% per Myr (Arbogast& Slowinski, 1998). JMODELTEST was used tochoose the appropriate models of sequence evolutionfor the Cytb data set. We employed an uncorrelatedexponential relaxed clock with a coalescent model ofexponential growth, and ran analyses for 4 ¥ 107 gen-erations, sampling every 1000 generations, and dis-carded the first 10% as burn-in.

DISTRIBUTION AND DEMOGRAPHIC CHANGES

To explore the generalized distributional changes ofAmmospermophilus since the LGM, we constructedspecies distribution models (SDMs) for each of thethree major lineages inferred from our phylogeneticanalyses. We built models for present (0 kya) and past(LGM, c. 21 kya) climatic conditions based on occur-rence data. This data set included occurrence recordsof individuals examined in this study (Appendix S1)as well as a subset of available records accurate towithin 5 km listed in MaNIS (http://manisnet.org/). Intotal we compiled 44 records for the interpres clade,255 records for the leucurus north clade, and 50records for the leucurus south clade. These occurrencerecords spanned the geographical distribution of eachof the three lineages. The maximum entropy methodimplemented in MAXENT 3.2.1 was used to generatethese models (Phillips et al., 2006a; Phillips,Anderson & Schapire, 2006b). MAXENT has beenshown to outperform similar habitat estimators(Phillips & Dudik, 2008; Elith & Graham, 2009), andits utility in phylogeographic studies was summarizedby Carstens & Knowles (2007) and Waltari &Guralnick (2009). The predictions for this analysiswere based on elevation plus a suite of 19 bioclimaticparameters previously compiled from the WorldClimclimate layers (Hijmans et al., 2005; Waltari et al.,2007), with a 2.5-arcminute/pixel resolution.

Model calibrations were performed using 75% of thedata as a training group, and the predicted distribu-tion models were then tested with the remaining 25%(Evans et al., 2009). Default parameters were used(500 maximum iterations; convergence threshold of0.00001; regularization multiplier of 1; 10 000 back-ground points) with random seeding, removal of mul-tiple presence records from individual cells resultingfrom sampling localities within 5 km2 (i.e. one pixel),and logistic probabilities for the output (Phillips &Dudik, 2008). A split-sample approach to separate thegeographically closest sample pairs between thetraining and test groups reduced the effects of spatialautocorrelation (Fielding & Bell, 1997; Parolo, Rossi& Ferrarini, 2008).

An initial model (including all 20 variables) wasrun to produce ‘area under the receiver operationcharacteristic curve’ (AUC) values for each bioclimaticparameter. A minimum AUC of 0.75 for the test groupwas considered the threshold for good model perform-ance (Elith et al., 2006; Suárez-Seoane et al., 2008;Elith & Graham, 2009). Models were then run usingtemporal transfer modelling from the current distri-bution to the LGM, incorporating information in theCommunity Climate System Model (CCSM 3; Collinset al., 2006) and the Model for InterdisciplinaryResearch on Climate (MIROC; Hasumi & Emori,2004). MAXENT analyses were performed three sepa-rate times using both the CCSM and MIROC climatereconstructions, and the habitat model results fromboth were averaged, accepting only those areas thatboth methods agreed were suitable (Waltari &Guralnick, 2009). Averaging the three independentMAXENT runs using the Spatial Analyst feature inArcGIS produced presence/absence binary habitatmodels using ArcGIS 9.2 (ESRI Corp., Redlands, CA,USA). The models were evaluated across four logisticthresholds: fixed cumulative value of 10.0; equaltraining sensitivity and specificity; equal test sensi-tivity and specificity; and equal entropy of thresh-olded and non-thresholded distributions. Thesethreshold values were used to assess a range of sen-sitivities and specificities to ensure that model inter-pretations were robust. Ultimately, the chosen cut-offof suitable habitat had a fixed cumulative probabilityof 10, a level that rejected the lowest 10% of predictedlogistic values. This value, although conservative,maintained a low omission rate (Pearson et al., 2007),consistent with the expectation that the occurrencerecords contain some georeferencing errors.

We additionally assessed responses to Pleistoceneclimatic oscillations by exploring the demographichistory of three widespread geographical clades ofA. leucurus inferred from our maternal phylogeny(see below). We created mismatch distributions ofpairwise distances among haplotypes and estimatedseveral demographic parameters for each clade,including nucleotide diversity (p), haplotype diversity(h), and Tajima’s D (Tajima, 1989), using DNASP 5.0(Rozas et al., 2003). If SDMs generally predicted anexpansion of favourable habitat since the LGM formost lineages of Ammospermophilus, we hypoth-esized that demographic parameter estimates for thethree wide-ranging clades of A. leucurus would simi-larly suggest population expansion.

RESULTSMATERNAL PHYLOGENY

The HKY + I + G model of nucleotide evolution wasselected for the combined CO3 and CR data set.



Analyses of the CO3 and CR sequence data for allsamples indicated strong phylogeographic structurewithin Ammospermophilus (Fig. 2). We inferred threewell-supported major lineages: A. interpres (hereafterreferred to as the interpres clade), A. insularis +A. leucurus from the southern Baja California penin-sula (leucurus south clade), and A. harrisii +A. nelsoni + A. leucurus from the northern Baja Cali-fornia peninsula and the USA (leucurus north clade).Samples of A. insularis formed a distinct clade withinthe leucurus south clade. Ammospermophilus harrisiiappeared to be polyphyletic, with two lineages fallingout in different places within the leucurus northclade. Samples of A. harrisii collected from Sonora,Mexico, and one sample from western Arizona, werebasal to a clade that contained all other A. harrisii, aswell as A. nelsoni and two geographical clades ofA. leucurus within the leucurus north clade (one inthe northern Baja California peninsula, the otherincluding all US samples of A. leucurus). Relation-ships among samples within each of the three cladesof A. leucurus (USA, northern peninsula, and south-ern peninsula; Fig. 3) lacked resolution (based onposterior probabilities and bootstrap values), consist-ent with a relatively recent diversification withineach lineage. Individuals with alternate mtDNA com-plements representative of either the northern penin-sula or southern peninsula clades (representing twoof the three major clades within the genus) werefound together at two localities in northern Baja

California Sur: 32 km east of San Ignacio (one north-ern and one southern individual) and 32 km south-east of Santa Rosalía (two northern and one southernindividual; Fig. 3B). These localities are in the vicin-ity of Volcán Tres Vírgenes, a sympatric site previ-ously identified by Álvarez-Castañeda (2007).

MULTILOCUS PHYLOGENY

The aligned multilocus data set contained 5962 bpand variable numbers of informative sites per gene:CO1, 76 informative sites/691 bp; CO3, 115/690 bp;Cytb, 173/1143 bp; 12S, 54/832 bp; 16S, 36/550 bp;IRBP, 43/1087 bp; Rag2, 17/969 bp. Models of nucle-otide evolution selected for each gene were GTR + G(Cytb, CO1, CO3), GTR + I (12S, 16S), HKY + I(Rag2), and HKY + I + G (IRPB). We were unable toobtain complete sequence data for western samplesof A. harrisii (Sonora and western Arizona), so thislineage was not included in the multigene data set.

Both ML and BI phylogenetic analyses resulted ina phylogenetic tree with three well-supported clades(Fig. 4). One clade was composed of A. harrisii, A. leu-curus samples from the northern Baja Californiapeninsula and USA, and A. nelsoni. Samples of A. leu-curus from the southern Baja California peninsulaformed a well-supported clade that included A. insu-laris, consistent with previous studies (Riddle et al.,2000a; Álvarez-Castañeda, 2007). A third, well-supported major clade was composed of samples ofA. interpres. These three clades corresponded to thesame clades inferred in our mtDNA range-wideassessments, referred to as our leucurus north, leu-curus south, and interpres clades, respectively.

Although the monophyly of Ammospermophiluswas strongly supported, the relationship among thethree major clades within this genus was unclear.Analysis of the nuclear data set alone (results notshown) confirmed the monophyly of Ammosper-mophilus, but did not show support for any structurewithin the genus. Importantly, however, the mtDNAand combined nuclear data were not in conflict,although western samples of A. harrisii were lackingfrom the nuclear data set.

DIVERGENCE TIME ESTIMATES

Based on our Ammospermophilus stem-calibratedphylogeny, the three major clades of Ammosper-mophilus may have diverged at around 4 Mya(event B in Fig. 5, Table 1). Divergences withinA. interpres (Fig. 5, event D) probably started nearthe end of the Pliocene, around 3 Mya. Westernsamples of A. harrisii were lacking from the nucleardata set, and were not included in divergence timeestimates, but the estimated date of divergence

Figure 2. Bayesian inference tree depicting relationshipsamong all haplotypes sampled across the range of allspecies of Ammospermophilus based on mitochondrialDNA (CO3 + CR). Asterisks represent nodes with posteriorprobabilities � 0.95.



Figure 3. (A) Relationships among species of Ammospermophilus (exclusive of A. interpres) based on Bayesian andmaximum likelihood analyses of 1194 base pairs of mitochondrial sequences (CO3 + CR), depicting three geographicallywidespread clades of Ammospermophilus leucurus in the USA, northern Baja California (BC) peninsula, and southernBaja California peninsula. Strongly supported nodes (� 0.95 Bayesian posterior probability, � 70% ML bootstrap) aredenoted by asterisks. AZ, Arizona; NM, New Mexico. (B) Distribution of Ammospermophilus (light and dark shaded area)and location of sample localities (numbered as in Figure 3A). Inset: detail of localities (numbers from this study,unnumbered circles from Álvarez-Castañeda, 2007) in relation to the purported former location of a Vizcaíno Seaway:northern peninsular clade (dark shading and black circles), southern peninsular clade (light shading and white circles),and three localities (half white, half black; boldface in Figure 3A) in which mtDNA haplotypes of both northern andsouthern Baja California peninsular clades are found.

Table 1. Estimated divergence dates within Ammospermophilus for each node depicted in Figure 5 based on threecalibration methods: fossil calibration placement at node A; fossil calibration placement at node B, and a rate calibrationusing cytochrome b (Cytb) data

A B Cytb

A 11.14 (9.99–13.25)* 25.21 (11.19–41.24) 12.45 (6.21–23.68)B 4.13 (2.11–6.6) 11.77 (10.1–13.37)* 5.05 (1.19–5.18)C 3.58 (1.34–4.66) 5.08 (2.22–8.21) 1.05 (0.48–1.95)D 2.91 (0.47–3.75) 2.37 (0.5–5.02) 0.53 (0.03–0.76)E 2.52 (0.99–4.25) 4.17 (1.6–7.34) 0.63 (0.13–1.16)F 2.01 (0.95–3.69) 3.42 (1.33–5.83) 1.05 (0.48–1.95)G 0.64 (0.02–1.13) 0.85 (0.05–2.31) 0.27 (0.0–0.38)H 0.15 (0.03–0.92) 1.05 (0.09–2.78) 0.54 (0.0–0.51)I 0.17 (0.01–0.98) 0.64 (0.04–1.68) 0.06 (0.0–0.19)

Divergence times in Mya followed by 95% posterior credibility intervals in parentheses. Asterisks denote fossil calibrationpoints.



between A. nelsoni and northern peninsular–USAA. leucurus + A. harrisii (event C) was near 3.6 Mya,and the divergence between A. harrisii, northernpeninsular A. leucurus, and USA leucurus (event F)was estimated at about 2 Mya. The time to themost recent common ancestor (tmrca) of southernA. leucurus + A. insularis (event E) was estimated at2.5 Mya, whereas within-population diversification inA. insularis (event G), A. nelsoni (event H), andA. harrisii (event I) was more recent, and well withinthe Pleistocene.

Our alternative calibrations resulted in differentestimates of divergence times. Based on an Ammos-permophilus crown-calibrated phylogeny, the diver-gence of Ammospermophilus from the out-group taxawas estimated to have occurred at 25.2 Mya, andestimates of clade and lineage divergence withinAmmospermophilus were generally much older thanestimates generated with the placement of fossil cali-bration on the stem node (Table 1). Dates from arate-calibrated Cytb data set were more congruentwith the placement of the fossil calibration at thestem node of the phylogeny, and suggested thatAmmospermophilus diverged from the out-group taxaat approximately 12.5 Mya. This rate calibrationmethod also indicated a basal divergence among themajor Ammospermophilus lineages at around 5 Mya,

Figure 4. Relationships among the major lineages ofAmmospermophilus inferred from our multilocus analyses.Asterisks represent strongly supported nodes (� 0.95Bayesian posterior probability, � 70% ML bootstrap).Horizontal dashed lines separate the three major clades.Sample localities are numbered as in Figures 3A and 3Band Appendix S1; BC, Baja California.

Figure 5. Ultrametric chronogram depicting divergence dates within Ammospermophilus. Arrows represent alternativeplacements of our fossil calibration. Numbers in bold below each node represent divergence time estimates (in millionsof years) based on the fossil calibration placed at node A. Grey bars represent the 95% posterior credibility intervalssurrounding these estimates. Numbers above each node represent divergence time estimates based on the fossilcalibration placed at node B (no credibility intervals are shown for these values). Geological time scale and branch lengthscorrespond to fossil calibration at node A. Node letters correspond to divergence time estimates listed in Table 1. Samplelocalities are numbered as in Figures 3 and 4; BC, Baja California.



consistent with the placement of a fossil calibration atthe stem node of the phylogeny.

DISTRIBUTION AND DEMOGRAPHIC CHANGES

The results of all SDMs were significantly better thanrandom samples (AUC = 0.5) in receiver operatingcharacteristic analyses (interpres clade, trainingAUC = 0.977, test AUC = 0.962; leucurus north clade,training AUC = 0.960, test AUC = 0.950; leucurussouth clade, training AUC = 0.996, test AUC = 0.995).The AUC values for all variables were greater than0.75. The predicted high-quality habitat representedin the present-day SDMs for each of the three clades(Fig. 6A, C, and E) largely captured the current dis-tribution of the species in each clade (Fig. 1). For theleucurus north clade, the present-day SDM (Fig. 6A)indicated continuous, high-quality habitat from thecentral Baja California peninsula north throughoutthe Mojave and Great Basin Deserts, and into the SanJoaquin Valley of California. Disjunct or lower-qualityhabitat extended into the eastern reaches of theGreat Basin (see Fig. 1), into the Sonoran Desert, andinto the westernmost portion of the Chihuahuan sub-region of the Chihuahuan Desert (considered to betransitional between Sonoran and Chihuahuan ele-ments by Hafner & Riddle, 2011). The present-daymodel for the leucurus south clade (Fig. 6C) indicatedthat continuous high-quality habitat was restricted tothe southern half of the Baja California peninsulaand to the adjacent, coastal continental SonoranDesert. The present-day SDM for the interpres clade(Fig. 6E) indicated that appropriate habitat was gen-erally concentrated throughout the two northern sub-regions of the Chihuahuan Desert (Chihuahuan andTrans-Pecos; Hafner & Riddle, 2011).

The reconstructions of LGM models for each of thethree clades of Ammospermophilus predicted anoverall loss of suitable habitat for each clade,although these models do not portray significantexposures of potential habitat that resulted fromlowered full-glacial sea levels (perhaps 200 km of dryland at the head of the Gulf of California and exten-sive continental shelf along the west coast of the BajaCalifornia peninsula; Hafner & Riddle, 2011: fig. 4.3).The high-quality habitat for northern leucurus(Fig. 6B) was compressed to the central Baja Califor-nia peninsula and the western Mojave Desert. Thepredicted habitat for the southern leucurus clade(Fig. 6D) was similarly compressed, confined withinthe southern extent of the Baja California peninsula.The LGM model for the interpres clade (Fig. 6F) indi-cated that habitat for this clade was dramaticallycompressed to a much smaller pocket of suitablehabitat in the Coahuilan and Zacatecan subregions

(Hafner & Riddle, 2011) of the southern ChihuahuanDesert.

Demographic parameter estimates calculated foreach of the three widespread clades of A. leucurus(USA, southern peninsula, and northern peninsula)yielded similar values. Values indicated that eachclade had low nucleotide diversity (USA, 0.0019;southern peninsula, 0.00319; northern peninsula,0.00315) and relatively high haplotype diversity(0.616; 0.889; 0.805). Tajima’s D-values for each cladewas significantly negative (–2.16215; -1.87675;-2.16208), consistent with probable population expan-sion. Mismatch distributions of haplotypes from eachlineage (Appendix S3) indicated a unimodal distribu-tion of haplotypes for all three clades, congruent withthe demographic parameters and with a model ofrecent demographic expansion.

DISCUSSION

Genetic patterns in a suite of arid-land taxa havebeen used to investigate the biogeographic history ofNorth American deserts. These include mammals(Riddle et al., 2000a; Riddle, Hafner & Alexander,2000b, c; Álvarez-Castañeda & Patton, 2004; Jezkovaet al., 2009; Bell et al., 2010), birds (Zink et al., 2001;Zink, 2002), reptiles (Upton & Murphy, 1997; Lindell,Mendez-de la Cruz & Murphy, 2005; Douglas et al.,2006; Leaché, Crews & Hickerson, 2007; Leaché &Mulcahy, 2007), amphibians (Jaeger, Riddle &Bradford, 2005; Bryson et al., 2012), invertebrates(Ayoub & Reichert, 2004; Crews & Hedin, 2006;Wilson & Pitts, 2010; Graham et al., 2013), and plants(Nason, Hamrick & Fleming, 2002; Garrick et al.,2009), as well as fish species bordering warm desertregions (Bernardi & Lape, 2005; Reginos, 2005). Thisbroad set of exemplar taxa has demonstrated acomplex history of vicariance and dispersal inresponse to both geological forces and climatic cycles.Our results suggest that Ammospermophilus hasresponded in similar fashion to this shared history,and provide additional insight into the mechanismsthat impacted diversification of arid-adapted taxaacross the arid lands of western North America.

PRE-QUATERNARY DIVERSIFICATION

Initial patterns of historical divergences withinAmmospermophilus appear coincident with dynamicalterations in the landscape of western NorthAmerica and the formation of the regional desertsduring the late Neogene (Riddle et al., 2000a). Fol-lowing a basal divergence of Ammospermophilus intothree major lineages at the Miocene–Pliocene bound-ary (Fig. 5), diversification within lineages occurredin rapid succession. Many of these divergence events



appear attributable to a number of vicariant eventsthat occurred prior to the Quaternary period.

Ammospermophilus interpres in the ChihuahuanDesert, the easternmost component of the NorthAmerican deserts, represents a divergent lineage thatmay have been isolated during the Late Miocene–early Pliocene. During this time, secondary upliftingof the Sierra Madre Occidental occurred simultane-ously with global climate change and the developmentof regional deserts (Axelrod, 1979; Webb, 1983; Zinket al., 2000; Riddle et al., 2000a). The synergisticeffects of these Neogene events probably subdividedmany taxa spanning the Sonoran and Chihuahuandeserts (Riddle & Hafner, 2006; Bryson, García-Vázquez & Riddle, 2011), with the more stronglydesert-adapted taxa likely to be isolated in aridregions subdivided by the Sierra Madre Occidental(Pyron & Burbrink, 2010).

Ammospermophilus leucurus, as currently recog-nized, represents a geographically structuredpolyphyletic species forming at least three distinctlineages: the USA lineage and northern peninsularlineage (aligned with A. nelsoni and A. harrisii in theleucurus north clade), and the southern peninsularlineage of the leucurus south clade (Fig. 3). The north-ern and southern peninsular lineages meet at theeastern edge of the Vizcaíno Desert of the centralBaja California peninsula. This peninsular pattern,inferred previously in Ammospermophilus (Riddleet al., 2000a), is similar to patterns of mid-peninsulardivergences detected in a suite of other taxa (e.g.Riddle et al., 2000b, c; Lindell et al., 2005; Crews &Hedin, 2006; Lindell, Méndez-de la Cruz & Murphy,2008; Bryson et al., 2012). Spatial and temporal diver-gences in these taxa support a mid-peninsular vicari-ance event potentially caused by the Vizcaíno Seaway(Figs 1A, 3B) during the late Miocene or earlyPliocene (Upton & Murphy, 1997; Riddle et al., 2000a;Whorley et al., 2004; Hafner & Riddle, 2011). Contactlocalities between the two clades of Ammosper-mophilus occur along the eastern edge of the hypoth-esized seaway, which may have been bridged byregional uplift and associated eruption of the TresVírgenes volcanic field, initiated 1.2 Mya and continu-ing today (Hafner & Riddle, 2011). Although directgeological evidence is seemingly absent for the Viz-caíno Seaway (Crews & Hedin, 2006; Lindell, Ngo &

Murphy, 2006), there may be a strong set of environ-mental factors driving divergences across this area(Grismer, 2000, 2002). Regardless of the exact causalmechanism, diversification across the mid-peninsularregion during the Late Miocene and Pliocene appearsto be a robust pattern.

Although the distinction between the northern andsouthern peninsular lineages of A. leucurus is basedprimarily on mtDNA sequence differences, studiesbased on other data also support this division. Hafner(1981) recognized that the southern peninsular formwas basal to all other recognized species in the genus(except A. interpres) based on limited allozyme dataand morphometric analyses, and tentatively sug-gested the recognition of a new species, Ammosper-mophilus canfieldae. Cladistic analysis of 23presumptive loci encoding for 13 proteins, withA. interpres designated as the out-group, placed thesouthern peninsular form (represented by a singlepopulation from the southernmost Cape Region)outside all remaining samples of (in turn) A. nelsoni,A. harrisii, and A. leucurus (including two popula-tions from the northern peninsula; Hafner, 1981).Similarly, phenetic analysis of bacular morphology(Hafner, 1981) indicated that those of A. interpreswere the most divergent, followed (in order) by thosefrom the southern peninsula, A. nelsoni, A. harrisii,and other A. leucurus. Multivariate analysis of 26cranial and dentary measurements from a totalsample including 253 specimens from the Baja Cali-fornia peninsula (Hafner, 1981) indicated an abruptshift in morphology at approximately 29°30′N lati-tude, the northernmost margin of the peninsular fogdesert (Mesa Huatamote filter barrier; Hafner &Riddle, 2011; Fig. 1A). This morphological shift wasmore abrupt than that across the Colorado Riverbetween the recognized species A. leucurus andA. harrisii, but was substantially north of the Viz-caíno Desert contact between northern and southernpeninsula clades identified herein by mtDNAsequence analysis. Hafner (1981) indicated a subtleshift in morphology in the Vizcaíno region (based oncondylobasal length), and also demonstrated a geo-graphically separate shift in pelage coloration(banding pattern of tail undersurface, a characterused to distinguish species of Ammospermophilus inNew Mexico; Findley et al., 1975). This shift was even

Figure 6. Species distribution models of predicted distributions for Ammospermophilus in our leucurus north, leucurussouth, and interpres clades (Figure 4). Models predicted under current climatic conditions (present day) in the left panels(A, C, and E), and models of predicted distribution at the last glacial maximum in the right panels (B, D, and F). Thegreyscale shading represents the probability of occurrence, with the darkest colour (black) indicating the most suitablepredicted habitat and the lightest shading (light grey) indicating the least suitable predicted habitat. A probability ofoccurrence less than 10% is indicated by the white areas, and represents unsuitable habitat.�



farther north, coinciding with the versants of thePeninsular Range (Sierras Juárez and San PedroMártir): Ammospermophilus of the more mesicwestern slopes and south on the peninsula have thepattern characteristic of A. harrisii and A. interpres,whereas those of the drier eastern slopes have thetypical A. leucurus banding pattern.

The evolutionary histories of A. harrisii, A. nelsoni,and northern populations of A. leucurus appear to beclosely linked. Together, these taxa probably divergedin the early Pliocene, around 3.6 Mya, into western(A. nelsoni) and eastern (A. harrisii + northern A. leu-curus) groups, coincident with geological events thatisolated the Mojave Desert and San Joaquin Valleyfrom the Peninsular and Sonoran Deserts (Bell et al.,2010). These included the intensified uplift of theSierra Nevada and Transverse ranges (Norris &Webb, 1990), marine inundation of the Salton Trough(Boehm, 1984), and the formation of a through-flowing Colorado River (the freshwater interpretationof the Bouse Embayment of Luchitta, 1979; Spencer& Pearthree, 2001, 2005; Fig. 1A).

Interestingly, a distinct second lineage of A. harrisiifrom Sonora and western Arizona may have alsodiverged from other lineages within our leucurus northclade during the late Miocene–Pliocene. Although welacked nuclear sequence data for A. harrisii fromSonora and western Arizona, and so did not includethis lineage in multilocus estimates of divergencetimes, our phylogenetic analyses of mtDNA data(Figs 2, 3) suggest a basal split of this lineage fromother lineages within our leucurus north clade. Assum-ing multilocus data are in accord with this estimate,this divergence may date to sequential development ofthe through-flowing Colorado River (Spencer &Pearthree, 2001, 2005) along the Bouse Embayment(Buising, 1990). This geological event is thought tohave driven divergences in a number of co-distributedtaxa (Lamb, Avise & Gibbons, 1989; Jones, 1995;Riddle et al., 2000b, 2000c; Pellmyr & Segraves, 2003;Murphy, Trépanier & Morafka, 2006; Castoe, Spencer& Parkinson, 2007). The region surrounding the headof the Gulf of California, at the meeting point of theMojave, Peninsular, and Sonoran regional deserts, hashad a complex geological history, resulting in complexpatterns of species diversity (e.g. Hafner & Riddle,2011) and within-species groups (e.g. Neotoma lepidaspecies group; Patton, Huckaby & Álvarez-Castañeda,2007). Genetic relationships within A. harrisii in thisregion should be evaluated in future studies employingmultilocus data and larger sample sizes of A. harrisiifrom the western part of the Sonoran Desert, andinvestigating possible introgression between the twolineages of A. harrisii: our sample from locality 32(Fig. 3B) contained haplotypes from eastern andwestern A. harrisii (Fig. 3A).

All of the species of Ammospermophilus possesslarge, conservatively retained blocks of constitutiveheterochromatin (accounting for a 75% increase intotal DNA versus other ground squirrels), despiteextensive non-Robertsonian chromosomal rearrange-ments among the species (Mascarello & Mazrimas,1977). Mascarello & Mazrimas (1977: 215) concludedthat these unusual blocks of heterochromatin servean important function, perhaps of ‘a regulator ofrecombination and, hence, of variation in popula-tions.’ If these blocks of heterochromatin limit nuclearDNA variation, it may partially explain the absence ofgeographic structuring found in the two nuclear DNAmarkers compared with the extensive differentiationin mtDNA genes.

QUATERNARY CLIMATE CHANGE AND

GENETIC CONSEQUENCES

Population contractions during the LGM and subse-quent expansions northward are characteristic of anumber of North American warm desert taxa (e.g.Riddle et al., 2000a; Jaeger et al., 2005; Cicero & Koo,2012). Based on our species distribution models, therewas an overall restriction of suitable habitat for eachof the three major clades of Ammospermophilus at theLGM. Suitable habitat subsequently expanded alongwith the desert regions of western North Americafollowing glacial retreat.

Genetic patterns evident in the three geographi-cally widespread clades of A. leucurus suggest thatthese clades tracked this expanding habitat. Ourpopulation demographic analyses indicate that eachclade of A. leucurus has undergone recent expansionsfrom smaller ancestral populations, leading to anexcess of low-frequency polymorphisms. These resultsare in general agreement with the data reported byWhorley et al. (2004), who found evidence of popula-tion expansion in A. leucurus. In particular, the USAleucurus lineage is widespread across the MojaveDesert and Great Basin, displaying generally smooth,clinal variation in cranial morphology, with moreabrupt change occurring along the Peninsular Rangesof the northern Baja California peninsula (Ammos-permophilus leucurus peninsulae), and along theupper Colorado River and lower San Juan River ofsouth-eastern Utah (Ammospermophilus leucuruspennipes; Hafner, 1981). While noting often markedvariation in pelage coloration associated with moremesic climate (darker coloration) or substrate(reddish), Hafner (1981) recommended combining thefive recognized subspecies (Hall, 1981) spanning theMojave Desert and Great Basin into a single subspe-cies (Ammospermophilus leucurus leucurus).

Soon after the marine inundation of the SaltonTrough, sediment from the fully formed Colorado



River had filled in much of the Salton Trough andisolated it from the Gulf (Spencer & Pearthree, 2001,2005); however, portions of the Salton Trough wereperiodically submerged during the Late Pleistocene–Holocene (Stokes et al., 1997). Collectively, these his-torical inundations of the Salton Trough appear tohave driven and maintained the phylogenetically dis-tinct lineages of USA and northern peninsular Ammo-spermophilus. Similarly, abrupt ecological andclimatic shifts in the Vizcaíno Desert region of thecentral Baja California peninsula (Grismer, 2002),coupled with climatic oscillations of the later Pleis-tocene, may have produced alternate pulses of contactand isolation between the northern and southernpeninsular forms of Ammospermophilus subsequentto their divergence in isolation.

Rivers appear to represent partial or complete bar-riers to dispersal in Ammospermophilus. Forexample: the Colorado River between A. leucurus andA. harrisii downstream from the central GrandCanyon of Arizona; the Río Grande between A. leucu-rus and A. interpres in New Mexico; the western (RíoGrande) and eastern (Pecos River) limits of A. inter-pres in Texas; and the few records of A. leucurus northof the Snake River in Idaho are believed to haveresulted from crossings of man-made bridges (Davis,1939), as is a single record on the south side of thebridge across the Grand Canyon of the Colorado Riverin Arizona (Hoffmeister, 1986). It is not clear whetherthese limits are imposed by the rivers themselves orby the absence of suitable habitat in the vicinity ofthe rivers. The lower reaches of the Colorado Riverand the Río Grande have both been historicallysubject to extensive meanderings, effecting a river‘crossing’ in static rodent populations. The Río Grandehas historically been reduced seasonally to a shallowstream and mudflats, whereas the Colorado Riverwas fully diverted for 2 years in 1905 (forming theSalton Sea; Sykes, 1937), and flow to the Gulf wasreduced substantially by completion of the HooverDam (in 1935), and ceased entirely for 20 years fol-lowing construction of the Glen Canyon Dam in 1962(Brusca et al., 2005). However, two of three specimens(MVZ) assigned to A. harrisii from the vicinity ofYuma, Arizona, and collected prior to these modernperturbations (1892 and 1899), were morphologicallyidentified as A. leucurus with a posterior probabilityof > 0.99 (Hafner, 1981).

The isolation of A. harrisii and A. interpres appearsto have been maintained by the Sierra Madre Occi-dental and the ranges of eastern Arizona and westernNew Mexico. The Cochise filter-barrier (Morafka,1977; Fig. 1A), a lower-elevation gap in these elevatedhighlands along the southern Arizona–New Mexicoborder, probably posed a complete barrier to Ammo-spermophilus during full-glacial periods, when pluvial

lakes filled the adjacent closed basins (pluvial lakeCochise in Arizona, Animas in New Mexico, andPalomas in Chihuahua; Smith & Street-Perrott,1983). Although the Cochise filter-barrier has main-tained isolation in a large number of co-distributedtaxa (Pyron & Burbrink, 2010), A. harrisii has sub-sequently spread east (Geluso, 2009), yet remainsisolated from A. interpres by the intervening playasand the Río Grande in southern New Mexico(Fig. 1B). In central New Mexico, A. leucurus occurson bajadas on the west bank of the Río Grande,whereas A. interpres is restricted to piñon–juniperwoodland habitat 10–15 km east of the river (Fig. 1B;D. J. Hafner, pers. observ.). Although we did notanalyse the demographic history of A. interpresbecause of the low sample size, this species is likely tohave contracted and expanded in response to glacial–interglacial cycles throughout the Pleistocene. OurSDM at the LGM (Fig. 6F) indicates a severe com-pression of the distribution of A. interpres into thesouthernmost Chihuahuan Desert, very similar to arecent SDM constructed for the desert pocket mouse,Chaetodipus eremicus (Jezkova et al., 2009). In starkcontrast, our present-day models (Fig. 6E) suggestlarge swathes of suitable habitat extending from thesouthern Chihuahuan Desert, across all of Trans-Pecos Texas, and most of New Mexico, in broad agree-ment with the current distribution of A. interpres.Presumably future studies will uncover a markedgenetic response in A. interpres as this speciestracked habitat expansion northwards.

CONCLUSION

The apparent lack of zones of sympatry betweenneighboring species of Ammospermophilus has frus-trated attempts to evaluate the degree of geneticisolation between recognized species (Findley et al.,1975; Hafner, 1981). Clear morphological and ecologi-cal differences exist among A. leucurus and theperipheral allopatric species A. nelsoni (larger bodysize; more yellowish pelage coloration; more opendesert habitat) and A. interpres (longer relative tailand hindlimbs; closer association with piñon–juniperwoodland and rocky habitat); however, such distinc-tions are far less pronounced between A. leucurus andA. harrisii, separated only (and perhaps incompletely)by the lower Colorado River (Hafner, 1981). There areno obvious morphological or ecological differencesbetween the northern peninsular and southern penin-sular lineages of A. leucurus described herein(Hafner, 1981), although Álvarez-Castañeda (2007)indicated possible morphological differences.Although the two lineages are defined primarily onmtDNA molecular sequence differences, these arenonetheless greater than those among A. nelsoni,



A. leucurus, and A. harrisii. In order to reconcilethese patterns pending more intensive genetic analy-sis of populations along purported or possible areas ofcontact (Vizcaíno Desert and lower Colorado River),we propose the recognition of the southern peninsularlineage of A. leucurus as a distinct species. We believethat this alternative best represents our currentunderstanding of evolutionary relationships amonggenetic forms of Ammospermophilus, rather thanretaining obvious paraphyly of A. leucurus relative toA. nelsoni and A. harrisii (based on molecularsequence data and supported by preliminary allozymeevidence; Hafner, 1981), or to resolve the paraphylyby combining A. leucurus (Merriam, 1889) and A. nel-soni (Merriam, 1893) under A. harrisii (Audubon &Bachman, 1854).

AMMOSPERMOPHILUS INSULARIS NELSON

& GOLDMAN, 1909PENINSULAR ANTELOPE SQUIRREL

(SYNONYMY UNDER SUBSPECIES)

Geographic range: Inhabiting generally rocky deserthabitat in Baja California Sur, Mexico (including IslaEspíritu Santo and Isla San Marcos) from the CapeRegion (excluding higher elevation coniferous forestsof the Sierra Laguna) north to the eastern edge of theVizcaíno Desert, in the vicinity of the Tres Vírgenesvolcanic field. Elevational range from approximatelysea level to approximately 600 m a.s.l. (possiblyhigher in the Tres Vírgenes, Sierra Giganta, andother arid ranges of Baja California Sur).

Description: Mid-dorsal summer pelage colorationvariable, but generally among the darker (highereumelanin) and more rufous (higher phaeomelanin)for the genus; specimens from the eastern edge of theVizcaíno Desert and the Llano de Magdalena notablylighter and less rufous; two clearly defined blackbands alternating with three white (including distal)bands on underside of tail (Hafner, 1981). Small toaverage sized for the genus, with a relatively longtail; specimens from eastern edge of Vizcaíno Desertnotably smaller than others; short, robust rostrumand short tooth row; small auditory bullae; baculumwith long distal recurved wings, short proximallateral expansions, deep shaft angle, and deep distalcup (Hafner, 1981).

AMMOSPERMOPHILUS INSULARIS INSULARIS

NELSON & GOLDMAN, 1909

Ammospermophilus leucurus insularis Nelson &Goldman, 1909: 24. Type locality ‘Espíritu SantoIsland, Gulf of California, Baja California [Sur],Mexico.’ Type specimen adult female, skin and skull,

United States National Museum (USNM) 146783collected 7 February 1906 by E. W. Nelson andE. A. Goldman, original number 19072.

Citellus leucurus insularis Elliot, 1917: 28. Namecombination.

Citellus insularis Howell, 1938: 181. Namecombination.

Ammospermophilus insularis Hall & Kelson,1959: 334. Name combination.

Ammospermophilus leucurus insularis Álvarez-Castañeda, 2007: 1167. Name combination.

Geographic range: Known only from Isla EspírituSanto and Isla Partida Sur (the two islands in Bahíade la Paz are connected by a narrow isthmus).

Description: Similar to Ammospermophilus insularisextimus of the peninsular mainland except slightlylarger and darker, with vestigial or absent upperpremolar (Pm3). Skull larger, particularly zygomaticarches, nasals, and auditory bullae (Álvarez-Castañeda, 2007). The Pm3 is extremely variable inthe family Sciuridae, and is occasionally absent, uni-laterally or bilaterally, in all members of the genus.Whereas the frequency of an anomalous Pm3 is below15% in other species of Ammospermophilus, it is over50% in A. i. insularis. In addition to the absenceof one or both permanent Pm3, anomalies includethe absence of one or both deciduous Pm3 (dPm3) orretention in adults of the dPm3 (Hafner, 1981;Álvarez-Castañeda, 2007).

AMMOSPERMOPHILUS INSULARIS EXTIMUS

NELSON & GOLDMAN, 1929

Ammospermophilus leucurus extimus Nelson &Goldman, 1929: 281. Type locality ‘Saccaton, 15 mi. Nof Cape San Lucas, Lower California [Baja CaliforniaSur], Mexico.’ Type specimen adult female, skin andskull, USNM 146587, collected 29 December 1905 byE.W. Nelson and E.A. Goldman, original number18805.

Geographic range: Eastern edge of the VizcaínoDesert (vicinity of Tres Vírgenes volcanic field) in thecentral Baja California peninsula, northern Baja Cali-fornia Sur, south throughout the peninsula to theCape Region, excluding the higher elevation, conifer-ous forests of the Sierra Laguna.

Description: Same as for species.

ACKNOWLEDGEMENTS

This research was funded by NSF grants to B.R.R.(DEB-9629787 and DEB-0237166) and D.J.H. (DEB-



9629840 and DEB-0236957), and a Major Instrumen-tation grant (DBI-0421519) to UNLV. Additionalfunding was provided to S.J.M. by NSF EPSCoR, theAmerican Museum of Natural History Theodore Roo-sevelt Memorial Fund, the American Society of Mam-malogists Grant-In-Aid of Research Program, and theUNLV Graduate & Professional Student Association.The following institutions provided samples used inthis study: New Mexico Museum of Natural History(P. Gegick), Museum of Natural Science, LouisianaState University (M. Hafner), Burke Museum ofNatural History and Culture, University of Washing-ton (G. Kenagy), Monte L. Bean Life Science Museum,Brigham Young University (D. Rogers), ColecciónNacional de Mamíferos, Instituto de Biología, Univer-sidad Nacional Autónoma de México (F. Cervantes andY. Hortelano), and The Museum, Texas Tech Univer-sity (H. Garner). We thank M. Chmiel, L. Alexander,J. Berger, M. Krainock, and D. Bangle for laboratoryassistance; D. Houston, C. Lowrey, K. Kleyboecker,and Z. Marshall for assistance with fieldwork; D. Hou-ston, M. Mika, C. Herr, S. Neiswenter, and A. Lelandfor analytical assistance; and J. A. Allen and threeanonymous reviewers for their helpful comments. Allfield and laboratory protocols followed the recommen-dations of the American Society of MammalogistsAnimal Care and Use Committee (Gannon et al.,2007), and were approved by the UNLV IACUC (pro-tocol no. R701-0703-180). The paper fulfilled partialrequirements for the PhD degree to S.J.M., withspecial thanks to committee members G. Voelker,J. Klicka, and A. Kirk.

REFERENCES

Álvarez-Castañeda ST. 2007. Systematics of the antelopeground squirrel (Ammospermophilus) from islands adjacentto the Baja California Peninsula. Journal of Mammalogy88: 1160–1169.

Álvarez-Castañeda ST, Patton JL. 2004. Geographicgenetic architecture of pocket gopher (Thomomys bottae)populations in Baja California, Mexico. Molecular Ecology13: 2287–2301.

Arbogast BS, Slowinski JB. 1998. Pleistocene speciationand the mitochondrial clock. Science 282: 1955a.

Armstrong DM. 1972. Distribution of mammals in Colorado.University of Kansas Museum of Natural History Mono-graph 3: 1–415.

Axelrod DI. 1979. Age and origin of Sonoran Desert vegeta-tion. Occasional Papers of the California Academy of Sci-ences 132: 1–74.

Ayoub NA, Reichert SE. 2004. Molecular evidence for Pleis-tocene glacial cycles driving diversification of a NorthAmerican desert spider, Agelenopsis aperta. MolecularEcology 13: 3453–3465.

Baker RH. 1956. Mammals of Coahuila, México. University of

Kansas Museum of Natural History Special Publication 9:125–335.

Bell KC, Hafner DJ, Leitner P, Matocq MD. 2010. Phy-logeography of the ground squirrel subgenus Xerosper-mophilus and assembly of the Mojave desert biota. Journalof Biogeography 37: 363–378.

Bernardi G, Lape J. 2005. Tempo and mode of speciation inthe Baja California disjunct fish species Anisotremus dav-idsonii. Molecular Ecology 14: 4085–4096.

Black CC. 1963. A review of the North American TertiarySciuridae. Bulletin of the Museum of Comparative Zoology,Harvard University 130: 113–248.

Boehm MC. 1984. An overview of the lithostratigraphy, bios-tratigraphy, and paleoenvironments of the Late NeogeneSan Felipe marine sequence, Baja California, Mexico. In:Frizzell VA, ed. Geology of the Baja California peninsula.Los Angeles: Society of Economic Paleontologists and Min-eralogists, Pacific Section, 253–265.

Brusca RC, Findley LT, Hastings PA, Hendrickx ME,Cosio JT, Van Der Heiden AM. 2005. Macrofaunal diver-sity in the Gulf of California. In: Cartron J-LE, Ceballos G,Felger RS, eds. Biodiversity, ecosystems, and conservation inNorthern Mexico. New York: Oxford University Press, 179–202.

Bryson RW, García-Vázquez UO, Riddle BR. 2011. Phy-logeography of Middle American gophersnakes: mixedresponses to biogeographical barriers across the MexicanTransition Zone. Journal of Biogeography 38: 1570–1584.

Bryson RW, Jaeger JR, Lemos-Espinal JA, Lazcano D.2012. A multilocus perspective on the speciation history ofa North American aridland toad (Anaxyrus punctatus).Molecular Phylogenetics and Evolution 64: 393–400.

Buising AV. 1990. The Bouse Formation and bracketingunits, southeastern California and western Arizona: impli-cations for the evolution of the proto-Gulf of California andthe lower Colorado River. Journal of Geophysical Research95: 111–120.

Carstens BC, Knowles LL. 2007. Shifting distributions andspeciation: species divergence during rapid climate change.Molecular Ecology 16: 619–627.

Castoe TA, Spencer CL, Parkinson CL. 2007. Phylogeo-graphic structure and historical demography of the westerndiamondback rattlesnake (Crotalus atrox): a perspective onNorth American desert biogeography. Molecular Phylogenet-ics and Evolution 42: 193–212.

Cicero C, Koo MS. 2012. The role of niche divergence andphenotypic adaptation in promoting lineage diversification inthe Sage Sparrow (Artemisiospiza belli, Aves: Emberizidae).Biological Journal of the Linnean Society 107: 332–354.

Collins WD, Bitz CM, Blackmon ML, Bonan GB, Brether-ton CS, Carton JA, Chang P, Doney SC, Hack JJ,Henderson TB, Kiehl JT, Large WG, McKenna DS,Santer BD, Smith RD. 2006. The Community ClimateSystem Model version 3 (CCSM3). Journal of Climate 19:2122–2143.

Crews SC, Hedin M. 2006. Studies of morphological andmolecular phylogenetic divergence in spiders (Araneae:Homalonychus) from the American southwest, including



divergence along the Baja California Peninsula. MolecularPhylogenetics and Evolution 38: 470–487.

Davis WB. 1939. The Recent mammals of Idaho. Caldwell:Caxton Printers, Ltd.

Douglas ME, Douglas MR, Schuett GW, Porras LW. 2006.Evolution of rattlesnakes (Viperidae; Crotalus) in the warmdeserts of western North America shaped by Neogene vicari-ance and Quaternary climate change. Molecular Ecology 15:3353–3374.

Drummond A, Rambaut A. 2007. BEAST: Bayesian evolu-tionary analysis by sampling trees. BMC EvolutionaryBiology 7: 214.

Drummond AJ, Ho SYW, Phillips MJ, Rambaut A. 2006.Relaxed phylogenetics and dating with confidence. PLoSBiology 4: e88.

Elith J, Graham CH. 2009. Do they? How do they? WHY dothey differ? On finding reasons for differing performances ofspecies distribution models. Ecography 32: 66–77.

Elith J, Graham CH, Anderson RP, Dudík M, Ferrier S,Guisan A, Hijmans RJ, Huettmann F, Leathwick JR,Lehmann A, Li J, Lohmann LG, Loiselle BA, ManionG, Moritz C, Nakamura M, Nakazawa Y, OvertonJMCM, Peterson AT, Phillips SJ, Richardson K,Scachetti-Pereira R, Schapire RE, Soberón J, Wil-liams S, Wisz MS, Zimmermann NE. 2006. Novelmethods improve prediction of species’ distributions fromoccurrence data. Ecography 29: 129–151.

Evans MEK, Smith SA, Flynn RS, Donoghue MJ. 2009.Climate, niche evolution, and diversification of the ‘Bird-cage’ evening primroses (Oenothera, Section Anogra andKleinia). The American Naturalist 173: 225–240.

Felsenstein J. 1985. Confidence limits on phylogenies: anapproach using the bootstrap. Evolution 39: 783–791.

Fielding AH, Bell JF. 1997. A review of methods for theassessment of prediction errors in conservation presence/absence models. Environmental Conservation 24: 38–49.

Findley JS, Harris AH, Wilson DE, Jones C. 1975.Mammals of New Mexico. Albuquerque: University of NewMexico Press.

Gannon WL, Sikes RS, The Animal Care and Use Com-mittee of the American Society of Mammalogists.2007. Guidelines of the American Society of Mammalogistsfor the use of wild mammals in research. Journal of Mam-malogy 88: 809–823.

Garrick RC, Nason JD, Meadows CA, Dyer RJ. 2009. Notjust vicariance: phylogeography of a Sonoran Deserteuphorb indicates a major role of range expansion along theBaja peninsula. Molecular Ecology 18: 1916–1931.

Geluso K. 2009. Distributional records for seven species ofmammals in southern New Mexico. Museum of Texas TechUniversity, Occasional Papers 287: 1–7.

Graham MR, Jaeger JR, Prendini L, Riddle BR. 2013.Phylogeography of the Arizona hairy scorpion (Hadrurusarizonensis) supports a model of biotic assembly in theMojave Desert and adds a new Pleistocene refugium.Journal of Biogeography Online Early View doi: 10.1111/jbi.12079.

Grismer LL. 2000. Evolutionary biogeography on Mexico’s

Baja California peninsula: a synthesis of molecules andhistorical geology. Proceedings of the National Academy ofSciences of the United States of America 97: 14017–14018.

Grismer LL. 2002. A re-evaluation of the evidence for amid-Pleistocene mid-peninsular seaway in Baja California:a reply. Herpetological Review 33: 15–16.

Guindon S, Gascuel O. 2003. A simple, fast, and accuratealgorithm to estimate large phylogenies by maximum like-lihood. Systematic Biology 52: 696–704.

Gustafson EP. 1978. The vertebrate faunas of the PlioceneRingold formation, south-central Washington. Bulletin of theMuseum of Natural History, University of Oregon, Eugene 23:1–62.

Hafner DJ. 1981. Evolution and historical zoogeography ofantelope ground squirrels, genus Ammospermophilus(Rodentia: Sciuridae). Unpublished Ph.D. thesis, Universityof New Mexico.

Hafner DJ, Riddle BR. 1997. Biogeography of Baja Califor-nia peninsular desert mammals. In: Yates TL, Gannon WL,Wilson DE, eds. Life among the muses: papers in honor ofJames S. Findley. Albuquerque: The Museum of Southwest-ern Biology, University of New Mexico, 39–68.

Hafner DJ, Riddle BR. 2005. Mammalian phylogeographyand evolutionary history of northern Mexico’s deserts. In:Cartron J-LE, Ceballos G, Felger RS, eds. Biodiversity,ecosystems, and conservation in Northern Mexico. New York:Oxford University Press, 225–245.

Hafner DJ, Riddle BR. 2011. Boundaries and barriers ofNorth American warm deserts: an evolutionary perspective.In: Upchurch P, McGowan A, Slater C, eds. Palaeogeogra-phy and palaeobiogeography: biodiversity in space and time.Boca Raton: CRC Press, The Systematics AssociationSpecial Volume Series, 75–113.

Hall ER. 1946. Mammals of Nevada. Berkeley and LosAngeles: University of California Press.

Hall ER. 1981. The mammals of North America. New York:John Wiley & Sons.

Harrison RG, Bogdanowicz SM, Hoffman RS, Yensen E,Sherman PW. 2003. Phylogeny and evolutionary history ofthe ground squirrels (Rodentia: Marmotinae). Journal ofMammalian Evolution 10: 249–276.

Hasumi H, Emori S. 2004. K-1 coupled GCM (MIROC)description. Technical Report No. 1. Center for ClimateSystem Research, University of Tokyo, Tokyo.

Herron MD, Castoe TA, Parkinson CL. 2004. Sciurid phy-logeny and the paraphyly of Holarctic ground squirrels(Spermophilus). Molecular Phylogenetics and Evolution 31:1015–1030.

Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A.2005. Very high resolution interpolated climate surfaces forglobal land areas. International Journal of Climatology 25:1965–1978.

Hoffmeister DF. 1986. Mammals of Arizona. Tucson: Uni-versity of Arizona Press and Arizona Game and FishDepartment.

Jaeger JR, Riddle BR, Bradford DF. 2005. CrypticNeogene vicariance and Quaternary dispersal of the red-spotted toad (Bufo punctatus): insights on the evolution of



North American warm deserts biota. Molecular Ecology 14:3033–3048.

James GT. 1963. Paleontology and nonmarine stratigraphyof the Cuyama Valley Badlands, California. Part 1. Geology,faunal interpretations, and systematic descriptions of Chi-roptera. University of California Publications in GeologicalSciences 45: 1–170.

Jezkova T, Jaeger JR, Marshall ZL, Riddle BR. 2009.Pleistocene impacts on the phylogeography of the desertpocket mouse (Chaetodipus penicillatus). Journal of Mam-malogy 90: 306–320.

Jobb G, von Haeseler A, Strimmer K. 2004. TREEF-INDER: a powerful graphical analysis environment formolecular phylogenetics. BMC Evolutionary Biology 4: 18.

Jones KB. 1995. Phylogeography of the desert horned lizard(Phrynosoma platyrhinos) and the short-horned lizard(Phrynosoma douglassi): patterns of divergence and diver-sity. Unpublished PhD thesis, University of Nevada, LasVegas, NV.

Lamb T, Avise JC, Gibbons JW. 1989. Phylogeographicpatterns in mitochondrial DNA of the desert tortoise (Xero-bates agassizi), and evolutionary relationships among theNorth American desert tortoises. Evolution 43: 76–87.

Leaché AD, Crews SC, Hickerson MJ. 2007. Two waves ofdiversification in mammals and reptiles of Baja Californiarevealed by hierarchical Bayesian analysis. Biology Letters3: 646–650.

Leaché AD, Mulcahy DG. 2007. Phylogeny, divergencetimes and species limits of spiny lizards (Sceloporus mag-ister species group) in western North American deserts andBaja California. Molecular Ecology 16: 5216–5233.

Lindell J, Mendez-de la Cruz FR, Murphy RW. 2005.Deep genealogical history without population differentia-tion: discordance between mtDNA and allozyme divergencein the zebra-tailed lizard (Callisaurus draconoides). Molecu-lar Phylogenetics and Evolution 36: 682–694.

Lindell J, Méndez-de la Cruz FR, Murphy RW. 2008.Deep biogeographical history and cytonuclear discordancein the black-tailed brush lizard (Urosaurus nigricaudus) ofBaja California. Biological Journal of the Linnean Society94: 89–104.

Lindell J, Ngo A, Murphy RW. 2006. Deep genealogies andthe mid peninsular seaway of Baja California. Journal ofBiogeography 33: 1327–1331.

Longmire JL, Maltbie M, Baker RJ. 1997. Use of ‘lysisbuffer’ in DNA isolation and its implication for museumcollections. Occasional Papers of the Museum of Texas TechUniversity 163: 1–4.

Luchitta I. 1979. Late Cenozoic uplift of the southwesternColorado Plateau and adjacent lower Colorado River region.Tectonophysics 61: 63–95.

Maddison DR, Maddison WP. 2005. MacClade 4: analysisof phylogeny and character evolution. Version 4.08. Sunder-land, MA: Sinauer Associates.

Mascarello JT, Mazrimas JA. 1977. Chromosomes of ante-lope squirrels (genus Ammospermophilus): a systematicbanding analysis of four species with unusual constitutiveheterochromatin. Chromosoma 64: 207–217.

Mercer JM, Roth VL. 2003. The effects of Cenozoic globalchange on squirrel phylogeny. Science 299: 1568–1572.

Miller WE. 1980. The late Pliocene Las Tunas local faunafrom southernmost Baja California, Mexico. Journal of Pale-ontology 54: 762–805.

Morafka DJ. 1977. A biogeographical analysis of the Chihua-huan Desert through its herpetofauna. Biogeographica 9:1–313.

Murphy RW, Trépanier TL, Morafka DJ. 2006. Conserva-tion genetics, evolution and distinct population segments ofthe Mojave fringe-toed lizard, Uma scoparia. Journal ofArid Environments 67: 226–247.

Nason JD, Hamrick JL, Fleming TH. 2002. Gene migra-tion, range expansion, and phylogeography of the columnarcactus genus Lophocereus. Evolution 56: 2214–2226.

Norris RM, Webb RW. 1990. Geology of California, 2nd edn.New York: John Wiley and Sons.

Parolo G, Rossi G, Ferrarini A. 2008. Toward improvedspecies niche modelling: Arnica montana in the Alps as acase study. Journal of Applied Ecology 45: 1410–1418.

Patton JL, Huckaby DG, Álvarez-Castañeda ST. 2007.The evolutionary history and a systematic revision ofwoodrats of the Neotoma lepida group. University of Cali-fornia Publications in Zoology 135: 1–411.

Pearson RG, Raxworthy CJ, Nakamura M, Peterson AT.2007. Predicting species distributions from small numbersof occurrence records: a test case using cryptic geckos inMadagascar. Journal of Biogeography 34: 102–117.

Pellmyr O, Segraves KA. 2003. Pollinator divergence withinan obligate mutualism: two yucca moth species (Lepidop-tera; Prodoxidae: Tegeticula) on the Joshua tree (Yuccabrevifolia; Agavaceae). Annals of the Entomological Societyof America 96: 716–722.

Phillips M, McLenachan P, Down C, Gibb G, Penny D.2006a. Combined mitochondrial and nuclear DNAsequences resolve the interrelations of the major Australa-sian marsupial radiations. Systematic Biology 55: 122–137.

Phillips SJ, Anderson RP, Schapire RE. 2006b. Maximumentropy modeling of species geographic distributions. Eco-logical Modelling 190: 231–259.

Phillips SJ, Dudik M. 2008. Modeling of species distribu-tions with Maxent: new extensions and a comprehensiveevaluation. Ecography 31: 161–175.

Posada D. 2008. jModelTest: phylogenetic model averaging.Molecular Biology and Evolution 25: 1253–1256.

Pyron RA, Burbrink FT. 2010. Hard and soft allopatry:physically and ecologically mediated modes of geographicspeciation. Journal of Biogeography 37: 2005–2015.

Rambaut A, Drummond AJ. 2007. Tracer v.1.4. Availablefrom http://beast.bio.ed.ac.uk/Tracer.

Reginos C. 2005. Cryptic vicariance in Gulf of Californiafishes parallels vicariant patterns found in Baja Californiamammals and reptiles. Evolution 59: 2678–2690.

Riddle BR. 1995. Molecular biogeography in the pocket mice(Perognathus and Chaetodipus) and grasshopper mice (Ony-chomys): the late Cenozoic development of a North Ameri-can aridlands rodent guild. Journal of Mammalogy 76:283–301.



Riddle BR, Hafner DJ. 2006. A step-wise approach to inte-grating phylogeographic and phylogenetic biogeographicperspectives on the history of a core North American warmdeserts biota. Journal of Arid Environments 66: 435–461.

Riddle BR, Hafner DJ, Alexander LF. 2000b. Comparativephylogeography of Bailey’s pocket mouse (Chaetodipusbaileyi) and the Peromyscus eremicus species group: histori-cal vicariance of the Baja California Peninsular Desert.Molecular Phylogenetics and Evolution 17: 161–172.

Riddle BR, Hafner DJ, Alexander LF. 2000c. Phylogeog-raphy and systematics of the Peromyscus eremicus speciesgroup and the historical biogeography of North Americanwarm regional deserts. Molecular Phylogenetics and Evolu-tion 17: 145–160.

Riddle BR, Hafner DJ, Alexander LF, Jaeger JR. 2000a.Cryptic vicariance in the historical assembly of a BajaCalifornia peninsular desert biota. Proceedings of theNational Academy of Sciences of the United States ofAmerica 97: 14438–14443.

Ronquist F, Huelsenbeck JP. 2003. MrBayes 3: bayesianphylogenetic inference under mixed models. Bioinformatics19: 1572–1574.

Rozas J, Sanchez-DelBarrio JC, Messeguer X, Rozas R.2003. DnaSP, DNA polymorphism analyses by the coales-cent and other methods. Bioinformatics 19: 2496–2497.

Schmidly DJ. 1977. Factors governing the distribution ofmammals in the Chihuahuan Desert Region. In: Wauer RH,Riskind DH, eds. Transactions of the Symposium on theBiological Resources of the Chihuahuan desert region,United States and Mexico. National Park Service Transac-tions and Proceedings Series 3. Washington: National ParkService, 163–192.

Shotwell JA, Russell DE. 1963. Mammalian fauna of theUpper Juntura Formation, the Black Butte local fauna.Transactions of the American Philosophical Society, newseries 53: 42–69.

Shreve F. 1942. The desert vegetation of North America.Botanical Review 8: 195–246.

Smith GI, Street-Perrott FA. 1983. Pluvial lakes of thewestern United States. In: Wright Jr. HE, ed. Late-Quaternary environments of the United States, Vol. 1.London: Longman Group Limited, 190–212.

Spencer JE, Pearthree PA. 2001. Headward erosion versusclosed-basin spillover as alternative causes of Neogenecapture of the ancestral Colorado River by the Gulf ofCalifornia. In: Young RA, Spamer EE, eds. The ColoradoRiver: origin and evolution. Grand Canyon AssociationMonograph 12. Arizona: Grand Canyon Association, 215–219.

Spencer JE, Pearthree PA. 2005. Abrupt initiation of theColorado River and initial incision of the Grand Canyon.Arizona Geology 35: 1–4.

Stokes S, Kocurek G, Pye K, Winspear NR. 1997.New evidence for the timing of aeolian sand supply tothe Algodones dunefield and East Mesa area, southeasternCalifornia, USA. Palaeogeography, Palaeoclimatology,Palaeoecology 128: 63–75.

Suárez-Seoane S, García de la Morena EL, Morales

Prieto MB, Osborne PE, de Juana E. 2008. Maximumentropy niche-based modelling of seasonal changes in littlebustard (Tetrax tetrax) distribution. Ecological Modelling219: 17–29.

Sykes G. 1937. The Colorado delta. New York: CarnegieInstitute of Washington, American Geographical Society.

Tajima F. 1989. Statistical method for testing the neutralmutation hypothesis by DNA polymorphism. Genetics 123:585–595.

Upton DE, Murphy RW. 1997. Phylogeny of the side-blotchedlizards (Phrynosomatidae: Uta) based on mtDNA sequences:support for a midpeninsular seaway in Baja California.Molecular Phylogenetics and Evolution 8: 104–113.

Verts BJ, Carraway LN. 1998. Land mammals of Oregon.Berkeley and Los Angeles: University of California Press.

Waltari E, Guralnick RP. 2009. Ecological niche modellingof montane mammals in the Great Basin, North America:examining past and present connectivity of species acrossbasins and ranges. Journal of Biogeography 36: 148–161.

Waltari E, Hijmans RJ, Peterson AT, Nyari AS, PerkinsSL, Guralnick RP. 2007. Locating Pleistocene refugia:comparing phylogeographic and ecological niche model pre-dictions. PLoS ONE 2: e563.

Webb SD. 1983. The rise and fall of the late Miocene ungulatefauna in North America. In: Nitecki MH, ed. Coevolution.Chicago: University of Chicago Press, 267–306.

Whorley JR, Álvarez-Castañeda ST, Kenagy GJ.2004. Genetic structure of desert ground squirrels overthe 20–degree–latitude transect from Oregon through theBaja California peninsula. Molecular Ecology 13: 2709–2720.

Wilgenbusch JC, Warren DL, Swofford DL. 2004. AWTY:a system for graphical exploration of MCMC convergence inBayesian phylogenetic inference. Available at http://ceb.csit.fsu.edu/awty

Wilson DE, Reeder DM, eds. 2005. Mammal species of theworld. Baltimore: The Johns Hopkins University Press.

Wilson JS, Pitts JP. 2010. Phylogeographic analysis of thenocturnal velvet ant genus Dilophotopsis (Hymenoptera:Mutillidae) provides insights into diversification in theNearctic deserts. Biological Journal of the Linnean Society101: 360–375.

Yensen E, Valdés-Alarcón M. 1999. Family Sciuridae. In:Álvarez-Castañeda ST, Patton JL, eds. Mamíferos delNoroeste de México. La Paz: Centro de InvestigacionesBiológicas del Noroeste, S.C., 239–320.

Zink RM. 2002. Methods in comparative phylogeography, andtheir application to studying evolution in the North Ameri-can aridlands. Integrative and Comparative Biology 42:953–959.

Zink RM, Barrowclough GF, Atwood JL, Blackwell-Rago RC. 2000. Genetics, taxonomy, and conservation ofthe threatened California gnatcatcher. Conservation Biology14: 1394–1405.

Zink RM, Kessen AE, Line TV, Blackwell–Rago RC. 2001.Comparative phylogeography of some aridland bird species.Condor 103: 1–10.



SUPPORTING INFORMATION

Additional supporting information may be found in the online version of this article at the publisher’s web-site:

Appendix S1. Voucher specimens for tissue used in this study are housed in the New Mexico Museum ofNatural History (NMMNH), Museum of Southwestern Biology, University of New Mexico (MSB), BurkeMuseum of Natural History and Culture, University of Washington (UWBM), Monte L. Bean Life ScienceMuseum, Brigham Young University (BYU), Colección Nacional de Mamíferos, Instituto de Biología, Univer-sidad Nacional Autónoma de México (CNMA), or The Museum, Texas Tech University (TTU). Tissue samplesare stored in NMMNH, UWBM, BYU, CNMA, TTU, or Louisiana State University Museum of Natural Science(M-numbers); corresponding LVT numbers refer to sequences deposited in Dryad (http://www.datadryad.org/;doi:10.5061/dryad.gs94r). Boldface numbers in parentheses before locality names refer to mapped localities inFigs. 1 and 3. Localities for A. interpres are listed north to south and mapped in Fig. 1.Appendix S2. Oligonucleotide primers used to amplify and sequence mitochondrial and nuclear DNA inAmmospermophilus and the two outgroup taxa Cynomys and Xerospermophilus. IRBP (nuclear Interphotore-ceptor Retinoid-Binding Protein), Rag2 (nuclear Recombining Activating Gene 2), 12S (small subunitmitochondrially-encoded ribosomal RNA), 16S (large subunit mitochondrially-encoded ribosomal RNA), Cytb(mitochondrial cytochrome b), CO1 (mitochondrial cytochrome oxidase 1), CO3 (mitochondrial cytochromeoxidase 3), and CR (mitochondrial Control Region). Primer names and sequences follow the cited sources, exceptAmmo LS1, which was generated specifically for this study.Appendix S3. Mismatch distributions (expected = solid line; observed = dashed line) representing the haplotypefrequency distribution of three geographically widespread clades of Ammospermophilus leucurus: (A) USA; (B)southern Baja California peninsula; and (C) northern Baja California peninsula.



Phylogeographic diversiﬁcation of antelope squirrels ... · Phylogeographic diversiﬁcation of...

Documents

Transcript of Phylogeographic diversiﬁcation of antelope squirrels ... · Phylogeographic diversiﬁcation of...