Results: Phylogenetic analyses: MP and MCMC resolved the same major clades. Xerus erythropus is basal to the remain

1. Evolution of the African ground squirrel genus Xerus:Phylogenetic and phylogeographic patterns reflect the influence of climate changeMatthew D. Herron, Christopher L. Parkinson and Jane M. WatermanDepartment of Biology, University of Central Florida • Results: • Phylogenetic analyses: • MP and MCMC resolved the same major clades. • Xerus erythropus is basal to the remaining Xerus. • The sister group to (X. inauris + X. princeps) is X. rutilus. • Xerus inauris is sister to X. princeps, and each is monophyletic. • Within X. princeps, a northern (Clade I) and a southern (Clade II) clade were resolved. • Three major clades were resolved within X. inauris: clade I, restricted to west-central Namibia; clade II, restricted to South Africa east and south of the Kalahari, and clade III, overlapping the distributions of clades I and II. • Nested clade analysis of X. inauris: • All haplotypes were included within three four-step clades (4-1 through 4-3 in Fig. 3), which correspond to the three major clades recovered in the MP and MCMC analyses (Fig. 2). • Restricted gene flow (RGF) with isolation by distance (IBD) was evident in two high level clades, including the total cladogram. • Two third-level clades (3-1 and 3-3) exhibited IBD or long distance dispersal (LDD), and one third level clade (3-2) showed past habitat fragmentation. Abstract:We used Bayesian and maximum parsimony phylogenetic methods and nested clade analysis to infer relationships among African ground squirrels (Xerus) using mitochondrial cytochrome b sequences. We inferred relationships among the four species of Xerus, evaluated the specific distinctness of Cape (X. inauris)and mountain (X. princeps) ground squirrels, and tested hypotheses for historical patterns of gene flow within X. inauris.The inferred phylogeny supports the existence of an “arid corridor” from the Horn of Africa to the Cape region until at least the Early Pleistocene. Our analyses suggest that X. inauris and X. princeps represent distinct monophyletic lineages, and thus appear to be valid taxa. Phylogenetic and phylogeographic patterns suggest distinct effects of climate change on population differentiation and speciation within Xerus. Introduction:Africa has undergone several major episodes of climate change since the Pliocene, which likely caused shifts in vegetation. Several authors have hypothesized that an “arid corridor” stretched from the Horn of Africa to the Cape of Good Hope during drier periods, providing a route by which xeric plants and animalsdispersed from central to southern Africa. This hypothesis is supported by disjunct distributions of several arid-adapted plant taxa found in both southern and eastern Africa. The ground squirrel tribe Xerini (Sciuridae) is an example of an arid-adapted animal taxon with a disjunct distribution in Africa. Two of the three extant genera in this tribe are endemic to Africa. The Barbary ground squirrel, Atlantoxerus getulus,is restricted to the extreme northwest of Africa, while the genus Xerus is broadly distributed across the arid and semiarid zones of sub-Saharan and southwest Africa. The third member, the long-clawed ground squirrel (Spermophilopsis leptodactylus),occupies a range east of the Caspian Sea in central Asia. The four species making up the genus Xerus occupy two disjunct ranges separated by nearly 2000 km (Fig. 1). With the exception of X. erythropus, which occupies a variety of habitats, Xerus is restricted to arid semi-desert and savannah habitats. The striped ground squirrel (X. erythropus) and the unstriped ground squirrel (X. rutilus) occupy overlapping ranges in sub-Saharan Africa, while the Cape ground squirrel (X. inauris) and the mountain ground squirrel (X. princeps) occupy overlapping ranges in the southern Africa. Pleistocene changes in precipitation patterns are likely to have fragmented arid habitats in southern Africa (Matthee & Flemming 2002), thus present patterns of genetic structure in arid-adapted taxa may have been significantly impacted by past isolation of populations in xeric refugia. Xerus inauris and X. princeps are almost morphologically indistinguishable and occur sympatrically in portions of their respective ranges. Both are terrestrial, semi-fossorial and diurnal (Dorst & Dandelot 1970). In spite of their morphological and ecological similarities, X. inauris and X. princeps have strikingly different social behaviors; X. inauris forms large bands, whereas X. princeps is mainly solitary (Herzig-Straschil & Herzig 1989; Waterman 1995). Female X. inauris live in highly social kin clusters of up to three adult females and subadult offspring (Waterman 1995; Waterman 1997). Males form social groups composed of unrelated individuals; their range overlaps the ranges of several female social groups (Waterman 1995; Waterman 1997). Both males and females forage as groups and share sleeping burrows (Waterman 1995). Little is known regarding the phylogeny of the Xerini, and several studies have presented conflicting hypotheses for their phylogenetic position within the Sciuridae. Here, we (1) infer phylogenetic hypotheses for the evolutionary relationships among the four currently recognized species of the genus Xerus, (2) investigate the historical biogeographic patterns leading to the current distributions of Xerus species, and (3) investigate the roles of gene flow and climate change in the phylogeographic history of X.inauris. Figure 1.Map of Africa showing ranges of Atlantoxerus getulus (vertical stripes), Xerus erythropus (dots), X. inauris (light gray), X. princeps (diagonal stripes) and X. rutilus (dark gray). Inset: southern Africa with ranges and sampling localities for Xerus inauris (clade I: gray circles; clade II: white circles; clade III: black circles; see text) and X. princeps (squares). • Conclusions: • Xerus phylogeny is consistent with the former existence of a hypothesized “arid corridor” running from the Horn of Africa to the Cape of Good Hope. Fossils identified as Xerus cf. inauris, dated from1.75 MYA, in the Olduvai Gorge of Tanzania (Fernandez-Jalvo et al. 1998) suggest that X. inauris may have originated in eastern Africa near the current range of X. rutilus. • Recognition of X. princeps and X. inauris as distinct species is warranted. Restriction of X. princeps to the western escarpment underscores the importance of adding all or part of this region to Namibia’s network of protected lands. • Gene flow among populations of X. inauris was restricted due to isolation by distance early in the species’ evolutionary history. Suitable habitat for X. inauris was widespread and essentially uninterrupted throughout the SWA early in the species’ evolution, as it is today. • A more recent restriction of gene flow resulted from habitat fragmentation in southern Africa, probably due to intermittently wetter conditions. The return of drier conditions allowed isolated populations of X. inauris to expand, eventually coming into secondary contact. This inferred secondary contact of X. inauris clades is observed in western Namibia and northern South Africa. Figure 1.Map of Africa showing ranges of Atlantoxerus getulus (vertical stripes), Xerus erythropus (dots), X. inauris (light gray), X. princeps (diagonal stripes) and X. rutilus (dark gray). Inset: southern Africa with ranges and sampling localities for Xerus inauris (clade I: gray circles; clade II: white circles; clade III: black circles; see text) and X. princeps (squares). Methods:We extracted total DNA using proteinase K digestion followed by phenol/chloroform/isoamyl alcohol organic separation methods. We PCR-amplified the cyt b gene using 45 cycles of denaturation at 94°C for 1 min, annealing at 45-60°C for 1 min, and extension at 72°C for 2 min. PCR products were excised from agarose gels and purified using the MinElute™ Gel Extraction kit (Qiagen). All sequencing reactions were run using the CEQ Dye Terminator Cycle Sequencing Quick Start Kit (Beckman Coulter), electrophoresed on a Beckman CEQ8000 automated sequencer, according to the manufacturer’s protocols. We edited raw sequence chromatographs in Sequencher 4.1.2 (Gene Codes Corp.) and aligned sequences manually in GeneDoc 2.6.002 (Nicholas & Nicholas 1997). We estimated phylogenetic relationships using both Bayesian Metropolis-coupled Markov Chain Monte Carlo (MCMC) and maximum parsimony (MP) methods. A cyt-b sequence from Spermophilopsis leptodactylus was the outgroup for all analyses. All MCMC phylogenetic reconstructions were conducted in MrBayes Version 3.0b4 (Ronquist & Huelsenbeck 2003), with nodal support evaluated by posterior probabilities. We conducted equally-weighted MP searches in PAUP* v4.0b10 (Swofford 2002) and assessed nodal support with nonparametric bootstrapping. We constructed statistical parsimony networks for X. inauris using a 95% confidence limit in TCS Version 1.13 (Clement et al. 2000). We nested the resulting haplotype network following the procedures outlined in Templeton et al. (1987), Templeton & Sing (1993) and Crandall (1996). We performed nested contingency analysis and nested geographical distance analysis using GeoDis Version 2.0 (Posada et al. 2000) and interpreted the results using the revised inference key (http://inbio.byu.edu/Faculty/kac/crandall_lab/geodis.htm). Figure 1.Map of Africa showing ranges of Atlantoxerus getulus (vertical stripes), Xerus erythropus (dots), X. inauris (light gray), X. princeps (diagonal stripes) and X. rutilus (dark gray). Inset: southern Africa with ranges and sampling localities for Xerus inauris (clade I: gray circles; clade II: white circles; clade III: black circles; see text) and X. princeps (squares). Acknowledgements:We thank Suzanne B. McLaren of the Division of Mammals, Carnegie Museum of Natural History, for providing two samples of Xerus rutilus; Dr. Mary Denver of the Baltimore Zoo for providing a sample of X. erythropus; and Dr. Eileen Lacey for providing samples of X. erythropus and X. inauris. Additional thanks go to Beryl Wilson of the McGregor Museum, Kimberly, South Africa and Alexandra Jo Newman for help collecting samples. We also thank Todd Castoe, John Fauth, James Roth and Franklin Snelson for their suggestions and comments on various drafts of the manuscript. This work was supported, in part, by a Graduate Thesis Grant from the University of Central Florida Department of Biology and by NSF grant #0130600.