African Clawed Frog &#X2018;Xenopus Laevis&#X201

Molecular Ecology (2015) 24, 909–925 doi: 10.1111/mec.13076

Pan-African phylogeography of a model organism, the African clawed frog ‘Xenopus laevis’

BENJAMIN L. S. FURMAN,* ADAM J. BEWICK,*1 TIA L. HARRISON,*2 ELI GREENBAUM,† VACLAV GVOZDIK,‡ § CHIFUNDERA KUSAMBA¶ and BEN J. EVANS* *Biology Department, McMaster University, Life Sciences Building, 1280 Main Street West, Hamilton, Ontario, Canada L8S 4K1, †Department of Biological Sciences, University of Texas at El Paso, 500 West University Avenue, El Paso, TX 79968, USA, ‡Institute of Vertebrate Biology, Academy of Sciences of the Czech Republic, Kvetna 8, 603 65 Brno, Czech Republic, §Department of Zoology, National Museum, 193 00 Prague, Czech Republic, ¶Laboratoire d’Herpetologie, Departement de Biologie, Centre de Recherche en Sciences Naturelles, Lwiro, Republique Democratique du Congo

Abstract The African clawed frog Xenopus laevis has a large native distribution over much of sub-Saharan Africa and is a model organism for research, a proposed disease vector, and an invasive species. Despite its prominent role in research and abundance in nature, surprisingly little is known about the phylogeography and evolutionary history of this group. Here, we report an analysis of molecular variation of this clade based on 17 loci (one mitochondrial, 16 nuclear) in up to 159 individuals sampled throughout its native distribution. Phylogenetic relationships among mitochondrial DNA haplotypes were incongruent with those among alleles of the putatively female-speciﬁc sex-determining gene DM-W, in contrast to the expectation of strict matrilineal inheritance of both loci. Population structure and evolutionarily diverged lineages were evidenced by analyses of molecular variation in these data. These results further contextualize the chronology, and evolutionary relationships within this group, support the recognition of X. laevis sensu stricto, X. petersii, X. victorianus and herein revalidated X. poweri as separate species. We also propose that portions of the currently recognized distributions of X. laevis (north of the Congo Basin) and X. petersii (south of the Congo Basin) be reassigned to X. poweri.

Keywords: gene ﬂow, phylogeography, population genetics, species limits Received 20 June 2014; revision received 6 January 2015; accepted 8 January 2015

native range of this species spans much of sub-Saharan Introduction Africa (Tinsley et al. 1996). Established invasive colonies The African clawed frog Xenopus laevis, sensu lato (Kobel exist in portions of Europe, North America and South et al. 1996), has an unusual connection with humans, hav- America (McCoid & Fritts 1980, 1989; Tinsley & McCoid ing been used in the early 20th century as a pregnancy 1996; Measey & Tinsley 1998; Lobos & KJaksic 2005), and assay (Shapiro & Zwarenstein 1934; Weisman & Coates nonpersistent populations have been reported in other 1941) and more recently as a model organism for research localities, including parts of Asia (Measey et al. 2012). (Cannatella & de Sa 1993; Gurdon 1996; Gurdon & Hop- Xenopus laevis has been identiﬁed as a potential vector for wood 2000). Also called the Common Platanna, the the amphibian pathogens Batrachochytrium dendrobatidis (Weldon et al. 2004) and ranavirus (Robert et al. 2007), although a causal role between X. laevis and the dispersal Correspondence: Ben J. Evans, Fax: 905-522-6066; of these pathogens has not been demonstrated (Measey E-mail: [email protected] et al. 2012). Xenopus laevis has potentially harmful conse- 1Current address: Department of Plant Biology, University of Georgia, Miller Plant Sciences Bldg., Athens, GA 30602, USA quences for Xenopus gilli (Evans et al. 2004, 2005; Evans 2Current address: Department of Ecology and Evolutionary Biol- 2007), which is endangered (South African Frog Re- ogy, University of Toronto, Toronto, Ontario Canada M5S3B2 assessment Group (SA-FRoG) ISASG (2013), through

© 2015 John Wiley & Sons Ltd 910 B. L. S. FURMAN ET AL. ecological competition and hybridization (Tinsley 1981; (Daudin 1802), X. petersii from southern Central Africa Simmonds 1985; Picker 1993; Picker et al. 1996; Evans (du Bocage 1895) and X. victorianus from Eastern Africa et al. 1997, 1998; Fogell et al. 2013). (Ahl 1924). A fourth species, X. poweri, was described Xenopus laevis is generally found in slow moving or based on specimens from the area of the Victoria Falls stagnant water, and occasionally disperses over land (Zambia–Zimbabwe border) by Hewitt (1927) but con- (Measey & Tinsley 1998; Eggert & Fouquet 2006). It is a sidered a subspecies of X. laevis by some authors generalist species that does well in disturbed habitat (Schmidt et al. 1959; Poynton 1964), or a synonym with and has a high capacity to tolerate drought conditions, X. (laevis) petersii (e.g. Parker 1936; Poynton & Broadley salinity, starvation, anoxia and temperature fluctuations 1985). Xenopus laevis sensu lato also includes two pro- (reviewed in Measey et al. 2012). This species (and other posed subspecies: X. l. bunyoniensis (Loveridge 1932) frogs in the Family Pipidae) has unusual adaptations in and X. l. sudanensis (Perret 1966). Additional informa- adults for a mostly aquatic lifestyle, including lateral tion on the taxonomic history of this clade is provided line sensory organs and a complex communication sys- in Supplementary Information. tem involving a unique mechanism of sound produc- Diversity within X. laevis sensu lato has been explored tion, context- and sex-specific vocalizations and female in terms of molecular and morphological variation (Carr phonotaxis (Tobias et al. 1998, 2004, 2011; Kelley et al. 1987; Grohovaz et al. 1996; Evans et al. 1997, 2004; & Tobias 1999). Kobel et al. 1998; Measey & Channing 2003; Du Preez et al. 2009) and variation in vocalization (Tobias et al. 2011). In general, these studies consistently found that Tetraploidization, sex determination populations in different parts of Africa, including popu- An ancestor of Xenopus laevis experienced genome lations from different portions of South Africa, are dif- duplication during its evolution, probably by allopoly- ferentiated. The distribution of variation within ploidization between two diploid ancestors with 18 mitochondrial DNA is perhaps best relayed in terms of chromosomes, to create a genome with 36 chromosomes four geographical zones of sub-Saharan Africa, which (Tymowska 1991). However, its genome is now func- we will refer to as ‘southern Africa’ (including South tionally diploidized in that, during cell division, each Africa and Malawi), East Africa (including Tanzania, chromosome aligns with only one other homologous Kenya, Uganda, Burundi, Rwanda and the eastern por- chromosome (Tymowska 1991). Tetraploidization in the tion of the Democratic Republic of the Congo, ‘Central ancestor of X. laevis duplicated essentially all genes in Africa’ (including Nigeria, Cameroon, western Zambia its genome, although many of these duplicates were and northern Botswana), and ‘West Central Africa’ reduced to a single copy as a result of pseudogenization (including the southern Republic of Congo, the western (Chain & Evans 2006; Morin et al. 2006; Hellsten et al. portion of the Democratic Republic of the Congo and 2007; Semon & Wolfe 2008; Chain et al. 2011). Angola) (Fig. 1). Evans et al. (2004) analysed mitochon- Of particular interest is the chromosome W-linked drial DNA sequences from X. laevis sensu lato from each DM-domain containing gene (DM-W), which is present of these zones. Their analysis recovered paraphyly of as a single allele in female X. laevis, but absent in males, the group of mitochondrial DNA sequences from south- and is the sex-determining gene in this species (Yoshim- ern Africa with relatively weak support, but recovered oto et al. 2008). DM-W originated by partial segmental strong support for monophyly of the group of mito- duplication of one of the two copies (paralogs) of the chondrial DNA sequences from East and Central Africa double sex- and mab-3-related transcription factor 1 (Evans et al. 2004). Mitochondrial DNA from one sam- gene (DMRT1) that arose when an ancestor of X. laevis ple from the Republic of Congo (West Central Africa) experienced tetraploidization (Bewick et al. 2011). If was closely related to a clade containing mitochondrial DM-W was strictly maternally inherited over evolution- DNA from Central and East Africa (Evans et al. 2004). ary time, its evolutionary history is expected to match Within the country of South Africa, Grohovaz et al. that of mitochondrial DNA, which is also thought to be (1996) and Measey & Channing (2003) found a popula- maternally inherited and nonrecombining in most spe- tion of X. laevis sensu lato sampled near the town of cies. Niewoudtville to be distinct from populations in other parts of the country. Measey & Channing (2003) also identified a zone of admixture of mitochondrial haplo- Taxonomy and phylogeography of X. laevis sensu lato types from Niewoudtville and haplotypes from the Xenopus laevis sensu lato (Kobel et al. 1996) comprises south-western Cape Region in the vicinity of the town three currently recognized species (AmphibiaWeb 2014; of Vredendal (not sampled in the current study), which Frost 2014): X. laevis sensu stricto from southern Africa is ~100 km south-west of Niewoudtville. Du Preez et al. and a disjunct population north of the Congo Basin (2009) further identified a second zone of admixture

a species as a ‘separately evolving metapopulation lineage’ (de Queiroz 2007). The term ‘metapopulation’ refers 18 10°N 20 to a set of subpopulations that are interconnected by Nigeria 19 gene ﬂow, and ‘lineage’ refers to the ancestor–descen- 17 dant relationship between metapopulations of the same Cameroon 23 Uganda species through time (de Queiroz 2007). Lendu Plateau 25 24 C t 22 R. Congo f 0° e i n R 26 D. R. 27 t X r Congo e Rwanda a n 28 Kenya . l i p t o A r 30 31 29 Methods e East

16 w f

b 33 r l e i BurundiAfrica

r c A 32 21 15 i a X. victorianus 34 West Samples and molecular data Central Tanzania 10°S Africa X. petersii 13 A total of 183 samples of X. laevis sensu lato from 14 14 Malawi Zambia countries were used in this study, including 104 sam- 11 10 Angola ples obtained from South Africa, 37 from Democratic Republic of the Congo (hereafter DRC), 12 from Bur- 12 20°S undi, 8 from Zambia, 7 from Cameroon, 3 from Nigeria, Namibia SouthernAfrica Current Revised . laevis Botswana X 3 from Uganda, 2 from Kenya, 2 from Botswana and 1 X. laevis X. petersii each from Rwanda, the Republic of Congo, Angola, 9 X. victorianus 500 km South Malawi and Tanzania (see Table S1, Supporting infor- X. poweri NA Africa 8 <1,500 m mation for specific locality information). These tissue 30°S 6 7 1,500–2,000 m 5 samples were obtained from field collections, tissue 2 3 >2,000 m 4 donations from institutional archives (California Acad- 1 emy of Sciences, the Museum of Comparative Zoology 0° 10°E 20°E 30°E 40°E at Harvard University, the Natural History Museum of Geneva and the Zoological Research Museum – Alexan- Fig. 1 Xenopus laevis sensu lato sampling localities, and currently recognized and revised species ranges. Numbers inside der Koenig), a collection of live Xenopus that was at the circles indicate locality numbers that correspond with samples University of Geneva, and colleagues (T. Hayes, L. Ka- listed in Table S1. Unfilled polygons with different lines indi- lous, R. Tinsley and P. Wagner). cate the currently recognized distributions of X. laevis, X. pet- Sequences from a portion of the mitochondrial 12S ersii (including X. poweri as a synonym) and X. victorianus, and 16S rDNA genes and the intervening tRNAVal gene respectively (AmphibiaWeb, 2014; Frost, 2014). Filled polygons were obtained using primers from Evans et al. (2004) for indicate four geographical regions that are referred to in the 159 X. laevis sensu lato individuals (87% of the samples text, each of which corresponds to the distribution of a species (named below the geographical region) that is supported by in this study), with an average of 907 base pairs (bp) per this study. No locus in this study has data from every sample individual (range: 623–2374 bp). Sequences from the depicted; mitochondrial DNA has the least missing data (see female-specific W chromosome gene DM-W and flank- Table S1 for details). Additional shading refers to meters above ing regions were obtained using primers detailed in sea level as indicated. Bewick et al. (2011) for 96 female X. laevis sensu lato individuals, with an average of 1734 bp per individual (range: 1036–2049 bp). Autosomal DNA sequences were near the town of Laingsburg (sampled in the current obtained from portions of the protein coding region of study), South Africa, between South African popula- 15 loci ranging in length from 341 to 618 bp (Table 1) tions to the north-east and south-west of this locality for 113–136 individuals per locus, using paralog-specific based on variation in mitochondrial DNA and two primers detailed in Bewick et al. (2011). Sequence data autosomal genes. were aligned by eye, and homologies of the aligned The main goal of this study is to further characterize characters were unambiguous. Sequences of individual the evolutionary history of X. laevis sensu lato in terms of autosomal alleles were inferred using the ‘best guess’ the phylogenetic relationships, divergence times and estimates of allelic states from PHASE v.2.1.1 using default geographic distributions of diverged evolutionary lin- parameters (Stephens et al. 2001; Stephens & Donnelly eages. We additionally evaluate support for previously 2003), and both alleles were analysed for the population proposed species designations within X. laevis sensu lato assignment tests detailed below. DNASP v.5.10.01 (Librado (X. laevis sensu stricto, X. victorianus, X. petersi and & Rozas 2009) was used to quantify descriptive statistics X. poweri). For evaluating support for the previously pro- of the sequence data, and formula 5 of Kimura & Ohta posed species, we adopt the ‘General Lineage Concept’ (1972) to calculate 95% confidence intervals (95% CI) for (GLC; de Queiroz 1998, 2007) of a species, which defines pairwise nucleotide diversity at synonymous sites. All

Table 1 Polymorphism statistics for autosomal loci for Xenopus laevis, X. petersii, X. powerii and X. victorianus, including the gene acronym (gene), number of base pairs sequenced (bp), number of alleles sequenced (No. of allelles), number of unique haplotypes (No. of haplotypes), number of synonymous sites (SSites), the number of nonsynonymous sites (NSites), Jukes Cantor corrected pairwise nucleotide diversity for synonymous (pS) and nonsynonmous (pN) sites, the number of segregating synonymous (SS) and non- * synonymous (SN) sites, and Tajima’s D based on synonymous sites (DS), with indicating signiﬁcant departure from zero. For some loci, NA indicates that Tajima’s D could not be calculated due to insufﬁcient molecular diversity or data

p p Gene bp No. of alleles No. of haplotypes SSites NSites S N SSSNDS

Southern Africa (X. laevis) AR 339 160 9 83 256 0.0015 0.0035 3 6 À1.28 prmt6 612 170 61 146 466 0.0237 0.0033 17 10 0.36 mogA 619 176 26 155 463 0.0031 0.0037 3 17 À0.13 c7orf25 531 184 16 119 412 0.0127 0.0005 11 4 À0.53 nfil3 534 188 21 120 415 0.0172 0.0016 12 7 À0.06 pigo 494 184 28 127 365 0.0230 0.0029 15 12 0.25 Sugp2 438 182 19 108 330 0.0103 0.0027 11 9 À1.16 mastl 537 184 46 119 418 0.0126 0.0073 11 32 À0.53 zbed4 471 170 17 110 361 0.0100 0.0029 9 9 À0.72 Rassf10 486 186 36 98 388 0.0307 0.0070 19 11 À0.29 p7e4 522 186 30 121 401 0.0274 0.0009 16 5 0.45 fem1c 474 146 24 107 367 0.0192 0.0010 17 21 À0.94 znf238.2 531 186 24 117 410 0.0030 0.0073 4 17 0.17 bcl9 489 188 18 114 372 0.0123 0.0010 10 10 À0.45 nufip2 473 156 24 105 363 0.0148 0.0019 27 29 1.97* West Central Africa (X. petersii ) AR 339 8 6 83 256 0.0065 0.0058 1 4 1.17 prmt6 612 2 2 145 467 0.0210 0.0021 3 1 NA mogA 619 10 7 155 463 0.0000 0.0139 0 19 NA c7orf25 531 10 9 118 413 0.0549 0.0099 18 11 À0.10 nfil3 534 10 6 120 414 0.0211 0.0013 6 1 0.72 pigo 494 10 5 127 365 0.0167 0.0006 6 1 À0.06 Sugp2 438 10 4 108 330 0.0052 0.0026 2 3 À0.69 mastl 537 10 4 119 418 0.0056 0.0018 2 3 À0.18 zbed4 471 8 2 109 357 0.0023 0.0000 1 0 À1.05 Rassf10 486 8 4 98 388 0.0118 0.0018 2 2 1.80 p7e4 522 10 3 122 400 0.0089 0.0000 3 0 0.02 fem1c 474 8 3 107 367 0.0175 0.0000 4 0 0.79 znf238.2 531 10 4 117 414 0.0052 0.0027 3 3 À1.56 bcl9 489 10 3 114 372 0.0000 0.0022 0 2 NA nufip2 473 10 4 106 365 0.0070 0.0018 2 2 0.12 Central Africa (X. poweri) AR 339 26 5 83 256 0.0071 0.0020 4 2 À1.20 prmt6 612 22 12 145 467 0.0155 0.0017 8 6 0.03 mogA 619 24 8 155 463 0.0000 0.0036 0 10 NA c7orf25 531 26 7 118 413 0.0043 0.0023 3 3 À0.89 nfil3 534 26 16 120 414 0.0130 0.0023 5 8 0.49 pigo 494 26 7 127 365 0.0087 0.0014 5 4 À0.48 Sugp2 438 24 5 109 329 0.0000 0.0017 0 4 NA mastl 537 18 14 119 418 0.1052 0.0194 33 27 0.90 zbed4 471 28 7 109 355 0.0095 0.0042 3 6 0.80 Rassf10 486 18 7 95 376 0.0126 0.0021 3 3 0.99 p7e4 522 26 4 121 401 0.0028 0.0005 3 2 À1.29 fem1c 474 26 16 107 367 0.0312 0.0004 13 2 À0.10 znf238.2 531 26 9 118 413 0.0102 0.0033 4 7 0.36 bcl9 489 26 4 114 372 0.0059 0.0016 3 2 À0.36 nufip2 473 22 10 106 365 0.0135 0.0036 7 5 À0.86

Table 1 Continued

p p Gene bp No. of alleles No. of haplotypes SSites NSites S N SSSNDS

East Africa (X. victorianus) AR 339 40 7 84 255 0.0007 0.0034 2 4 0.38 prmt6 612 38 10 145 467 0.0104 0.0001 8 1 À0.63 mogA 619 42 10 155 463 0.0009 0.0054 2 9 À1.30 c7orf25 531 40 5 117 413 0.0053 0.0009 3 2 À0.28 nﬁl3 534 40 9 120 414 0.0052 0.0016 3 5 À0.27 pigo 494 34 7 127 365 0.0052 0.0006 5 2 À1.20 Sugp2 438 42 4 108 330 0.0009 0.0004 2 1 À1.50 mastl 537 34 16 119 418 0.0173 0.0043 11 12 À0.78 zbed4 471 32 10 109 353 0.0039 0.0041 4 8 À1.50 Rassf10 486 42 15 99 387 0.0100 0.0058 4 13 0.12 p7e4 522 40 6 121 401 0.0090 0.0000 6 0 À0.62 fem1c 474 40 11 107 367 0.0137 0.0009 6 4 0.05 znf238.2 531 38 10 117 414 0.0057 0.0034 4 7 À0.73 bcl9 489 40 11 113 370 0.0157 0.0026 7 6 0.18 nuﬁp2 473 38 16 106 365 0.0192 0.0021 11 10 À0.67

new sequence data are deposited in GenBank (Accession was the general time reversible model (Tavare 1986), nos. KP343951–KP345838), and Accession nos. of other with a proportion of invariant sites, a gamma-distrib- data in these analyses are listed in previous studies uted heterogeneity in the rate of evolution and esti- (Evans et al. 2004, 2005, 2008, 2011; Evans 2007; Bewick mated base frequencies (GTR+I+Γ+bf). For the DM-W et al. 2011). data set, the preferred model was the Hasegawa, Ki- sino and Yano model (Hasegawa et al. 1985), with a proportion of invariant sites and estimated base fre- Phylogenetic analyses quencies (HKY+I+bf). We used BEAST v.1.6 (Drummond & Rambaut 2007) to To examine evolutionary relationships among the generate time calibrated trees for the mitochondrial autosomal genes, three approaches were taken. First, and for the DM-W data. For each locus, we performed phylogenetic networks were generated among phased four independent runs, 50 million generations each, autosomal alleles from each locus using SPLITSTREE using a strict clock. Previously published orthologous v.4.13.1 (Huson & Bryant 2006). We used Jukes–Can- sequences from X. gilli were used as outgroups. The tor corrected distances between alleles and the Neigh- timing of divergence of X. laevis and X. gilli was set bor-Net algorithm (Bryant & Moulton 2004). Support to 16.7 million years ago (Ma) with a standard devia- for the splits in the networks was determined with a tion of 3.62 Ma (Evans et al. 2004), to calibrate these bootstrap analysis with 1000 replicates. Second, we analyses. This divergence time is based on the performed a phylogenetic analysis on concatenated assumption that the separation of the South Atlantic autosomal data using BEAST version 1.7.4 (Drummond Ocean triggered the diversiﬁcation of South American et al. 2012) including individuals with less than 50% from African pipid frogs ~100 Ma (Pitman et al. 1993; missing data (i.e. the same individuals that were Maisey 2000; McLoughlin 2001; Sereno et al. 2004; Ali included in the population assignment analyses & Aitchison 2008) and was based on analysis of data described below). And third, we estimated a species from mitochondrial DNA (Evans et al. 2004). We tree using *BEAST with a reduced data set of 70 indi- tested for convergence of the MCMC chains on the viduals and 10 genes that minimized missing data posterior distribution by calculating effective sample across individuals and loci. For the phylogenetic sizes (ESSs) of post-burn-in likelihoods using TRACER analyses of concatenated autosomal data, we used a v.1.5 (Rambaut & Drummond 2007), and inspecting model of evolution selected by MrModeltest, with the traces of parameter estimates. This led us to discard a same calibration procedure as detailed above for mito- burn-in of 25% of the generations from each analysis. chondrial DNA and DM-W, and we performed four For each analysis, the model of evolution was selected independent runs for 20 million generations each. For by the program MrModeltest version 2 (Nylander the *BEAST analysis, we assumed a strict molecular 2004) based on the Akaike information criterion. The clock with an exponentially distributed mutation rate À preferred model for the mitochondrial DNA analysis with a mean of 4.7 9 10 10 substitutions/site/

© 2015 John Wiley & Sons Ltd 914 B. L. S. FURMAN ET AL. generation following Bewick et al. (2012). This muta- collapsed during the Markov Chain (Yang & Rannala tion rate estimate is based on a multilocus analysis of 2010). We ran two independent chains for each algo- data from >100 genes from pipid frogs and relied on rithm using a gamma prior G (2, 1000) for both the the same assumption about the geological trigger for population size and tree root age priors, with auto- diversification of pipid frogs as the mitochondrial matic adjustments of step lengths in the MCMC algo- DNA analysis above. To achieve convergence, it was rithm made by the program. In addition, we explored necessary to use a simpler model of evolution than an alternative prior for both of these parameters in that recommended by MrModeltest for the concate- which we calculated a scale parameter (b) for the nated data set (we used HKY+Γ+bf instead of gamma distribution, by dividing 1 (a diffuse value for GTR+I+Γ+bf). For the *BEAST analysis, we linked the the shape parameter (a) of the gamma distribution) by model of evolution across all data partitions, and the mutation rate used in the *BEAST analysis (Bewick unlinked the phylogeny of each partition. A priori spe- et al. 2012) multiplied by an estimated divergence time cies designations were based on eight clades that were from the outgroup taxon X. gilli of 16.7 Ma (Evans observed in the concatenated analysis of autosomal et al. 2004), which resulted in a value of 126. We then DNA, including: (i) Nigeria and Cameroon, (ii) Bots- ran two independent chains for both algorithms with wana and Zambia, (iii) Angola and western DRC, (iv) this new gamma prior distribution G (1, 126) for both eastern DRC, Uganda and Burundi, (v) Malawi, and the ancestral population size and the tree root age. For the South African localities Kimberly and Victoria each prior setting, the MCMC was run for 100 000 West, (vi) the South African locality Niewoudtville, generations, and 20 000 generations were discarded as (vii) the South African localities Betty’s Bay, Garden burn-in, based on visual inspection of the posterior Route National Park and some individuals from La- distribution of likelihoods. ingsburg and (viii) the South African localities Beau- fort West, and other individuals from Laingsburg. The Population assignment *BEAST analysis was performed with four independent runs, each for 100 million generations. For BEAST and The phylogenetic analyses detailed above evaluate evo- *BEAST analyses, orthologs from X. gilli were used as lutionary relationships in the context of a bifurcating the outgroup, and independent runs were combined phylogeny. However, autosomal DNA relationships using LOGCOMBINER version 1.7.4 (Drummond et al. may reticulate or be inconsistent among loci as a result 2012). Similar to the other phylogenetic analyses, 25% of gene flow, lineage sorting and recombination, and of the run was discarded as burn-in, and convergence this is a particular concern when analysing intraspe- was assessed based on ESSs of the parameters as cal- cific samples. More specifically, use of a phylogeny culated by TRACER version 1.6 (Drummond & Rambaut estimated from autosomal DNA to guide the *BEAST 2007). A maximum clade credibility tree with median and BP&P analyses comes with the caveat that we did node heights was constructed using TREEANNOTATOR not explore all possible groupings or (for BP&P) all pos- version 1.7.4 (Drummond et al. 2012). sible relationships among these groups, and therefore the results are contingent on the a priori groups and guide tree that we used for *BEAST and BP&P, respec- BP&P analysis tively. Population assignment tests (and also the Splits- We used BP&P version 2.2 (Yang & Rannala 2010) to Tree analysis discussed above) therefore offer a test for evidence of species limits. This analysis uses complementary perspective on the nature of multilocus molecular data and a ‘guide’ phylogeny, which is a molecular variation among taxa because they do not hypothesized relationship among populations or spe- interpret evolutionary relationships in the context of a cies, to evaluate the posterior probability of a species bifurcating tree. To assess the degree of population tree. The species tree is assumed to either be the same structure and assign individual genotypes to putative as the guide tree, or alternatively to be a simplified populations, we used the programs TESS v.2.3 (Chen version of the guide tree that can be obtained by col- et al. 2007) and STRUCTURE v.2.3 (Pritchard et al. 2000). lapsing one or more nodes. We used a guide tree Both approaches estimate the probability that each based on clusters obtained from the phylogenetic individual is assigned to K populations, with an aim analysis of concatenated autosomal DNA, and of minimizing Hardy–Weinberg and linkage disequilib- included only the autosomal DNA sequences that were ria within the populations (Pritchard et al. 2000; Chen analysed in the *BEAST analyses detailed above. BP&P et al. 2007; Francßois & Durand 2010). Unlike Structure, has two different reversible jump proposal algorithms Tess incorporates spatial information on geographic for species delimitation that influence the probability distances between sampling points (based on GPS that nodes within the guide tree are expanded or coordinates) into the prior distribution when

© 2015 John Wiley & Sons Ltd EVOLUTION OF XENOPUS LAEVIS 915 calculating individual assignment probabilities (Chen Results et al. 2007; Francßois & Durand 2010). For Tess and Structure analyses, we excluded data Phylogenetic incongruence between maternally from individuals with missing data from more than half inherited loci of the loci. Data from mitochondrial DNA and DM-W were also excluded so that these analyses would pro- Estimated phylogenetic relationships from mitochon- vide a perspective on diversification independent from drial DNA and from DM-W each resolve sequences into the analyses of the maternally inherited loci. The 135 geographically clustered clades that correspond with individuals in the analysis had an average of 6.3% of one another, and both analyses recover strong and the loci with missing data. We ran Tess for 100 000 gen- congruent support for paraphyly of the group of erations with a burn-in of 10 000, using the conditional haplotypes from individuals in one pond in Laingsburg, auto-regressive admixture model (Durand et al. 2009), South Africa. However, there are strongly supported starting from a neighbour-joining tree and using 10 iter- inconsistencies in the estimated relationships among ations for each value of K ranging from 2 to 10. Because these clades (Fig. 2; see Fig 1 and insert in Fig. 3 for some individuals were sampled from the same location, sampling locations). The mitochondrial DNA phylogeny we used the ‘generate spatial coordinates for individu- supports monophyly of the group of sequences from als’ option in Tess, with a standard deviation equal to the following South African localities: Niewoudtville, 1.0. Convergence was based on inspection of post-run Beaufort West, Laingsburg, De Doorns, Betty’s Bay, log-likelihood plots, and support for alternative K val- GRNP, Hoekwil and Cape Town, whereas the DM-W ues was assessed by inspection of the deviance informa- phylogeny supports paraphyly of this group of tion criterion (DIC) (Spiegelhalter et al. 2002); models sequences (Fig. 2). Another difference with strong sta- with lower DIC values are preferred. tistical support is seen in relationships among samples For Structure analysis, we ran 20 million generations from East Africa. In the mitochondrial DNA phylogeny, with a burn-in of 2.5 million generations for values of K all East Africa sequences that are not from or near the equal to 2–10, with five iterations for each value of K. Lendu Plateau form a strongly supported clade. But in We specified the ‘admixture model’ (Falush et al. 2003) the DM-W phylogeny, this group of sequences is and assumed no correlation between alleles. The inferred to be paraphyletic. Strongly supported incon- post-run likelihood values were stable and support for sistent relationships were also inferred when we alternative K values was evaluated using the DK statis- restricted both analyses to include only those individu- tic (Evanno et al. 2005), as calculated with STRUCTURE als for whom data were collected from both loci (data HARVESTER WEB v.0.6.93 (Earl & von Holdt 2012), and the not shown). ad hoc method outlined in Pritchard et al. (2000). The samples used in the population assignment Molecular variation, evolutionary relationships and = analyses comprise more from South Africa (n 107) species delimitation using autosomal loci than from other portions of the distribution of X. laevis sensu lato that are not from South Africa (n = 41). To Table 1 presents polymorphism statistics for four examine whether this uneven geographic sampling diverged lineages of X. laevis sensu lato that correspond affected our results, we reran the Tess analysis with a to previously proposed species within this group as random subsample of only five individuals from each redefined below. All of the loci were polymorphic sampling locality in South Africa, plus all other sam- within X. laevis sensu lato. One locus exhibited a Taj- ples from other countries. This reduced data set had a ima’s (1989) D value that was significantly greater than total of 66 individuals, of which 25 were from South zero within a geographical region depicted in Fig. 1, an Africa. We ran this analysis for 100 000 generations observation that could reflect a signature of balancing with 10 000 discarded as burn-in, for K values ranging selection. After weighting individual locus values by from 2 to 10, with 10 iterations for each value. Other the number of synonymous sites at each locus, the larg- analytical details were identical to those discussed est average pairwise diversity of synonymous sites was above. similar for individuals from southern Africa (0.0148; – – Tess and Structure runs were post-processed using 95% CI: 0.0091 0.0206), Central Africa (0.0159; 0.0100 – CLUMPP v.1.1.2 (Jakobsson & Rosenberg 2007), which 0.0218) and West Central Africa (0.0124; 0.0072 0.0176), averages assignment probabilities across iterations. but about half as large for individuals from East Africa – Clumpp offers three separate algorithms that maximize (0.0081; 0.0039 0.0124). similarity across all of the iterations of a given K;we Phylogenetic analysis of concatenated autosomal data selected an algorithm as recommended in the program from 135 X. laevis sensu lato individuals provided strong documentation (Jakobsson & Rosenberg 2007). support for multiple diverged evolutionary lineages

A Mitochondrial DNA X. gilli

>95% East Africa, Central Africa, 47% 90-95% West Central Africa 85-90% AMNH17301 Malawi PF BJE3580 Potchefstroom BJE3577 BJE3579 BJE3609 BJE3574 Kimberley BJE3576 BJE3582 BJE3578 BJE3608 BJE3573 BJE3575 BJE3581 BJE3628 X. laevis BJE3641 BJE3642 BJE3631 BJE3645 BJE3651 BJE3637 BJE3644 BJE3648 South BJE3640 BJE3627 Africa AMNH17259 BJE3639 VG09_128 Nigeria BJE3635 AMNH17260 BJE3630 AMNH17262 BJE3649 XEN058 BJE3633 Niewoudtville VG09_100 BJE3646 BJE3255 Cameroon BJE3647 BJE3253 BJE3629 BJE3254 BJE3643 BJE3252 BJE3632 X. poweri ZFMK3TN02 BJE3650 ZFMK1TN10 BJE3537 PRIVATE33 BJE3540 AMNH17263 BJE3527 ZFMK3TN8 Zambia BJE3539 PRIVATE31 BJE3541 ZFMK1TN01 BJE3531 RT5 BJE3530 RT4 Botswana BJE3542 ZFMK2TN13 BJE3533 Zambia BJE3538 ELI502 BJE3532 ELI1369 BJE3529 Beaufort West ELI527 BJE3534 66% ELI1370 BJE3535 Laingsburg ELI526 BJE3525 ELI503 BJE3543 ELI1462 BJE3545 ELI1461 BJE3546 CK003 Eastern DRC, BJE3536 EBG2872 ELI1298 Burundi CFS1091 BJE3558 BJE266 BJE3550 CFS1090 BJE3555 BJE267 BJE3506 BJE263 BJE3553 ELI1064 BJE3547 ELI1142 BJE3511 EBG2147 BJE3509 ELI1065 BJE3552 ELI1141 BJE3512 ELI1139 BJE3559 49% ELI1066 BJE3557 BJE3507 Laingsburg XEN234 BJE3554 ELI965 BJE3513 De Doorns XEN232 Tanzania, BJE3514 ZFMK86160 BJE3556 Betty’s Bay ZFMK63120 Kenya, BJE3528 GRNP ZFMK63119 Burundi, BJE3551 X. victorianus ZFMK86159 BJE3526 Hoekwil CAS168711 Rwanda, BJE3572 ELI1013 BJE3548 Cape Town ELI1012 Uganda BJE3549 ELI1011 BJE3505 ELI1010 BJE3508 ELI938 LG12-3 BJE2898 BJE3510 EBG2463 Lendu Plateau, KML7 EBG2464 KML5 EBG2329 Eastern DRC KML8 EA4 VG08_81 Angola EA7 X. petersii PM086 RGL1 PM118 XSL5 PM109 Western DRC KML6 PM085 XSL4 PM108 AMNH17324R. Congo 15.0 12.5 10.0 7.5 5.0 2.5 0.0 5.0 2.5 0.0

Fig. 2 Chronogram among (A) mitochondrial DNA and (B) DM–W haplotypes in X. laevis sensu lato. Shaded dots over nodes indicate posterior probabilities, expressed as percentages as indicated; some terminal support values were omitted for clarity, and the posterior probabilities of various poorly supported nodes are indicated. The scale bar indicates divergence time from the present in millions of years. With the exception of the sample from Malawi, shaded branches in southern Africa correspond to sampling localities depicted in Fig. 3. Small arrows indicate relationships that are well supported in each phylogeny but discordant between them.

B DM-W X. gilli X. poweri AMNH17262 AMNH17260 Nigeria BJE2900 >95% BJE2898 EBG2329 Lendu Plateau, 90-95% BJE2902 Eastern DRC 85-90% EBG2464 EBG2463 ELI503 ELI527 X. victorianus ELI1370 ELI1369 ELI1462 ELI1461 LWIRO162 BJE0262 Southern Albertine Rift LWIRO161 Eastern DRC LWIRO174 BJE0267 BJE0260 LWIRO172 LWIRO171 BJE0263 BJE0266 1 LWIRO173 CFS1091 X. petersii PM108 1 PM109 PM118 Western DRC AMNH17301Malawi BJE3574 PF BJE3580 BJE3577 BJE3579 Potchefstroom BJE3582 Kimberley BJE3581 BJE3573 BJE3609 BJE3575 KML6 XSL3 EA7 KML8 0 BJE3549 KML7 RGL1 BJE3547 0 BJE3557 BJE3510 Betty’s Bay BJE3548 RGL2 GRNP 0 BJE3513 De Doorns 0 EA4 Hoekwil EA6 Laingsburg BJE3509 BJE3526 KML5 BJE3506 XSL18 0 BJE3572 XSL5 EA3 XSL4 X. laevis RGL3 BJE3558 BJE3505 BJE3550 BJE3538 BJE3540 BJE3536 BJE3537 BJE3535 Beaufort West BJE3543 Laingsburg BJE3545 BJE3539 1 BJE3534 BJE3525 BJE3527 BJE3639 BJE3650 BJE3629 BJE3630 BJE3632 BJE3636 BJE3643 Niewoudtville BJE3637 1 BJE3628 BJE3633 BJE3646 BJE3640 BJE3631 BJE3634 BJE3648 BJE3649

15.0 12.5 10.0 7.5 5.0 2.5 0.0

Fig. 2 Continued.

AMNH17301 Malawi Fig. 3 Phylogenetic analysis of concate- X. gilli PF Potchefstroom BJE3609 nated data from up to 15 autosomal loci AMNH17259 Nigeria BJE3608 BJE3255 BJE3576 VG09_100 BJE3575 per individual. A map shows sampling VG09_128 BJE3577 BJE3253 Cameroon BJE3582 localities with dots, a plus sign indicates X. poweri BJE3254 BJE3581 Kimberley BJE3252 BJE3580 a zone of admixture in Vredendal AMNH17263 BJE3579 ZFMK1TN01 BJE3578 Zambia BJE3573 between mitochondrial DNA lineages >95% RT5 BJE3574 Victoria West 90-95% RT4 Botswana 52% BJE3542 ZFMK3TN02 from Niewoudtville and the south-wes- 85-90% Zambia BJE3545 ELI1369 BJE3540 ELI1462 tern Western Cape Province identiﬁed by BJE3535 ELI1298 BJE3546 ELI1370 Beaufort West BJE3543 Measey & Channing (2003), and short EBG2147 BJE3541 ELI503 BJE3537 ELI502 dotted lines indicate the approximate ELI536 Eastern DRC, BJE3536 ELI527 southern BJE3544 locations of conﬁrmed or hypothesized EBG2872 BJE3534 Albertine BJE3538 ELI1461 BJE3539 contact zones between X. laevis popula- CFS1090 Rift BJE3530 CFS1091 BJE3525 tions. Long dotted lines indicate major X. victorianus ZFMK63120 BJE3532 Laingsburg ZFMK63119 Uganda BJE3531 ELI1142 geological formations. The scale bars Burundi BJE3529 BJE2897 BJE3533 BJE2900 indicates divergence time from the pres- BJE3631 EBG2464 Eastern DRC, BJE3642 EBG2463 BJE3635 ent in millions of years. BJE2902 Lendu Plateau BJE3628 EBG2329 BJE3641 BJE2898 BJE3640 VG08_81 BJE3644 South X. petersii PM108 Angola BJE3627 Africa PM118 BJE3633 PM085 Western DRC BJE3630 PM109 BJE3645 BJE3643 BJE3632 BJE3648 Niewoudtville BJE3651 BJE3649 BJE3647 X. laevis BJE3637 BJE3639 BJE3636 BJE3650 17.5 15.0 12.5 10.0 7.5 5.0 2.5 0.0 BJE3646 BJE3629 BJE3634 BJE3506 BJE3514 De Doorns BJE3509 RGL1 KML8 wana EA7 Betty’s Bay Bots EA6 Namibia BJE3552 Potchefstroom BJE3549 BJE3551 De Doorns, BJE3505 GRNP, BJE3557 Hoekwil BJE3548 BJE3558 BJE3555 G BJE3572 re BJE3528 at Kimberley Laingsburg

South Africa E RGL2

s XSL15 c BJE3507 ar Lesotho 30°S pm BJE3508 ent XSL4 KML7 De Doorns, BJE3512 Betty’s Bay, BJE3510 Hoekwil, Niewoudtville Victoria West BJE3513 GRNP + KML6 Vredendal EA5 BJE3511 Beaufort West BJE3553 RGL3 Laingsburg BJE3547 De Doorns Swartberg Range BJE3559 BJE3556 Lan Hoekwil BJE3554 Lewis ge GRNP b erg Range BJE3550 Gay Dam 200 km BJE3526 Betty’s BJE3527 Laingsburg Bay 20°E 25°E 5.0 2.5 0.0

(Fig. 3), many of which correspond to those identiﬁed in that the group of samples from East Africa that were in the analyses of mitochondrial DNA and DM-W. not from or near the Lendu Plateau were inferred to be Diverged lineages in the analysis of concatenated auto- monophyletic. The analysis of concatenated autosomal somal data include (i) individuals from southern Africa data differs from the analyses of mitochondrial DNA (South Africa and Malawi), (ii) individuals from East and DM-W in that the former supports monophyly of Africa (Uganda, Burundi, eastern DRC), (iii) individuals the group of samples from East Africa (Figs 2 and 3). from Central Africa (Nigeria, Cameroon, Zambia, Bots- Both maternally inherited loci supported a close rela- wana) and (iv) individuals from West Central Africa tionship between haplotypes from Niewoudtville, South (Angola and western DRC). The topology of relation- Africa and those from Beaufort West, South Africa and ships among the geographically clustered clades was a few from Laingsburg, South Africa, but this relation- more similar to that inferred from DM-W than ship was not observed in the analysis of concatenated mitochondrial DNA in the sense that the group of autosomal loci. samples from Malawi and South Africa are inferred to Geographical clustering of variation was observed in be monophyletic. However, it was more similar to the the concatenated analysis. Within Central Africa, for mitochondrial DNA phylogeny than the DM-W phylogeny example, samples from Nigeria and Cameroon form a

X. gilli the guide tree, with each cluster corresponding to the Posterior * Probability Nigeria, terminal taxa presented in Fig. 4 for the BEAST analysis. >95% X. poweri Cameroon 90-95% Zambia, Results were consistent for both species delimitation 85-90% Botswana algorithms, and for both prior settings that we tried. X. victorianus Eastern DRC, Burundi, Kenya X. petersii Angola, Western DRC Population assignment Malawi, Kimberley South Africa{Victoria West Tess and Structure population assignment analyses South Africa Niewoudtville X. laevis recovered similar results and support the existence of South Africa Beaufort West substantial population structure in X. laevis sensu lato {Laingsburg Betty’s Bay Hoekwil, GRNP (Fig. 5, Figs S2 and S3, Supporting information). The South Africa De Doorns {Laingsburg – 15.0 12.5 10.0 7.5 5.0 2.5 0.0 DIC plots suggest that 6 7 populations are preferred by Millions of years ago the Tess analysis and the method of Evanno et al. (2005) Fig. 4 Species tree analysis by *BEAST. In southern Africa, supports 5 populations in the Structure analysis. As a shaded branches correspond to the sampling localities depicted consequence of isolation by distance (identified using in Fig. 3. Partial Mantel tests, data not shown), we expected the ad hoc method of Pritchard et al. (2000) to deliver an clade that is most closely related to a clade comprising overestimation of the number of clusters due to depar- samples from Botswana and Zambia. Within East ture of the observed data from a model of multiple pan- Africa, samples from or near the Lendu Plateau form a mictic populations (Pritchard et al. 2000). As expected, clade that is most closely related to a clade containing this method supported the maximum number of clus- other samples from the rest of the Albertine Rift and ters we tested (K = 10, P < 0.001). samples from Uganda and Burundi. Within South Individuals assigned to each cluster were nearly Africa, geographically structured clades were recovered identical in both analyses at most values of K. Clusters from multiple regions, including (i) samples from identified by Tess and Structure at higher values of K Malawi and northern South Africa (Potchefstroom, corresponded to clades identified in the phylogenetic Kimberley, Victoria West) and (ii) samples from Nie- analyses, and to the species identified by BP&P analy- woudtville, South Africa, (iii) samples from Beaufort sis. Similar to the phylogenetic analyses, these assign- West and some samples from Laingsburg and (iv) other ment tests also highlight genetic uniqueness of the samples from Laingsburg plus samples from south- X. laevis sensu lato population from or near the Lendu western Western Cape Province. Plateau and that from Niewoudtville, and also distin- Species tree analyses with *BEAST (Heled & Drum- guish populations in the northern and southern por- mond 2010) supported the same relationships among tions of West Central Africa, the former of which most clusters of sequences as the concatenated analysis corresponds to a proposed subspecies X. l. sudanensis (Fig. 4). The exception to this is that the species tree (Perret 1966). analysis infers a monophyletic relationship between the two populations that included individuals from the Discussion admixed population in Laingsburg, whereas the concatenated analysis supported a paraphyletic relationship Phylogenetic incongruence among maternally inherited between these two populations with respect to other loci populations (Figs 3 and 4). Similar to the phylogenetic analyses discussed above, We observed well supported, discordant relationships the network analysis of phased X. laevis biparentally among lineages of two putatively maternally inherited inherited alleles reveals strong geographic association genomic regions in the frog X. laevis sensu lato: of molecular variation. Twelve of 15 networks placed mitochondrial DNA and the female-specific gene DM- molecular variation from southern Africa and the rest W (Yoshimoto et al. 2008). This observation could of sub-Saharan Africa on distinct portions of the net- reflect error in phylogenetic inference (that is, an incor- work (Fig. S1, Supporting information). Variation in rect phylogeny may have been inferred in one or both East Africa also tended to cluster in portions of these loci) or it could be a ‘real’ (biological) difference. Miss- networks that were distinct from variation in other ing data, long-branch attraction and model misspecifi- parts of Central Africa or West Central Africa. cation, for example, may affect phylogenetic inference Using the species delimitation program BP&P, we (Lemmon & Moriarty 2004; Kuck€ et al. 2012; Roure recovered strong support (a posterior probably of ~1) et al. 2013). A biological difference in these phylogenies for separate species statuses for all of the clusters in could arise if either of these markers was not strictly

A B 1.0 K = 7 0.0 20 500 1.0 K = 6 0.0 1.0 K = 5 18 250

0.0 DIC 1.0 K = 4 0.0 1.0 K = 3 16 000 0.0 Probability of assignment Probability 1.0 23456107 8 9 K = 2 Number of clusters (K) 0.0 Niewoudtville BW Lain GRNP Betty’s Ba y West Lendu Kimberly VW

Angola Malawi BotswanaDRC BurundiUganda South Zambia Africa Nigeria

Cameroon

Fig. 5 (A) Results of TESS analysis with population clusters (K) ranging from 2–7. (B) The deviance information criterion (DIC) for each value of K, with bars (most not visible) indicating the standard deviation of this estimate across iterations. In (A), two localities are labelled within the Democratic Republic of the Congo (DRC) including the western DRC (West DRC) and a region including or near the Lendu Plateau (Lendu). Localities in South Africa include Kimberley, Victoria West (VW), Niewoudtville, Beaufort West (BW), Laingsburg (Lain), Garden Route National Park (GRNP) and Betty’s Bay. maternally inherited, or if either experienced recombi- onomic revision from the standpoint of the GLC (de nation. If individuals carrying DM-W occasionally Queiroz 1998, 2007). Within X. laevis sensu lato, almost developed as phenotypic males, for instance, this could all of our analyses recovered support for at least four lead to a mode of inheritance that is not strictly mater- evolutionarily diverged lineages in the following geo- nal. Periodic phenotypic sex reversal coupled with sex- graphical regions: (i) southern Africa, including Malawi specific rates of recombination (specifically, a lower and South Africa, (ii) Central Africa, including Nigeria, recombination in the heterogametic sex) has been Cameroon, Zambia and Botswana and (iii) West Central proposed as a mechanism for maintaining nondiverged Africa, including the Republic of Congo, western DRC (homomorphic) sex chromosomes in other frogs (Perrin and Angola, and (iv) East Africa, including Kenya, 2009; Stock€ et al. 2011), and indeed, X. laevis has homo- Uganda, Rwanda, Burundi, eastern DRC and Tanzania. morphic sex chromosomes (Tymowska 1991). Further Each of these groups was identified as a differentiated information on rates of recombination in male and cluster in the population assignment tests, and lineages female X. laevis would be useful to evaluate the appli- i–iii were recovered in each of the phylogenetic cability of this hypothesis to X. laevis. It is also possi- analyses we performed (mitochondrial DNA, DM-W ble that DM-W or its flanking region exist in duplicate and concatenated and species tree analysis of autoso- copies in some females, and that these copies could mal DNA). Lineage iv formed a clade in the phyloge- occasionally undergo ectopic recombination events, netic analyses of autosomal DNA and DM-W, but not even if this locus were strictly maternally inherited. in the analysis of mitochondrial DNA. These four lin- While evidence for recombination in mitochondrial eages, respectively, correspond to four currently or pre- DNA has been reported in various taxa (Piganeau et al. viously recognized species: X. laevis, X. poweri, 2004; Tsaousis et al. 2005), many statistical approaches X. petersii and X. victorianus, but we argue for a revised to detect recombination are prone to false positives (In- distribution for two of them (X. laevis and X. poweri). A nan & Nordborg 2002; Galtier et al. 2006; Sun et al. revision of the distributions of X. laevis and X. poweri is 2011), and we view this as an unlikely explanation for warranted because individuals from the north of the our observations. Congo Basin (Cameroon, Nigeria) are more closely related to individuals from the south of the Congo Basin (Zambia, Botswana) than they are to individuals Statuses of previously proposed species from other parts of Africa, including southern Africa, Our results provide novel perspectives on the evolu- which is where X. laevis occurs. Thus, we reassign the tionary history of X. laevis sensu lato, and argue for tax- population of X. laevis sensu lato from Nigeria and

Cameroon to X. poweri instead of X. laevis. We note that are consistent with the findings from these and other a subspecies of X. laevis, X. l. sudanensis, from the Ad- studies (Grohovaz et al. 1996; Kobel et al. 1998; Measey amawa Region in Cameroon was described by Perret & Channing 2003). Similar to Du Preez et al. (2009), we (1966). Our data potentially support the transfer of found evidence for extensive introgression between X. l. sudanensis to the synonymy of X. poweri instead of populations south-west and north-east of the locality of X. laevis, although additional data from the type locali- Laingsburg. This was evinced by (i) individuals from ties or examination of the type specimens is needed. this locality having a diversity of evolutionary affinities Similarly, another subspecies of X. laevis, X. l. bunyoni- in the mitochondrial DNA, DM-W and concatenated ensis (Loveridge 1932), should be tentatively considered analysis of autosomal DNA and (ii) admixed population a synonym of X. victorianus, as evidenced by the affinities that were identified by population assignment inferred phylogeography of X. laevis sensu lato and by tests. We did not recover qualitative evidence for exten- phylogenetic position of our sample from south-wes- sive gene flow between other populations of X. laevis tern Uganda. Although again we note that this study sensu lato based on the population assignment tests. lacks samples directly from the type locality of One possibility is that this could be an artefact of miss- X. l. bunyoniensis, which should be investigated in the ing genetic information from animals in the contact future. Under our proposed taxonomy, relationships zones between these lineages, for example in the Congo among mitochondrial DNA variants of X. laevis, and Basin and south of the Congo Basin, or between differ- X. victorianus may be paraphyletic within each species; entiated populations in South Africa (Fig. 1). Reciprocal we note also that monophyly is not a requirement of crosses between X. laevis sensu lato individuals that the GLC (de Queiroz 2007). were probably from South Africa, and individuals from Although the question of whether further taxonomic Uganda or Botswana both produced fertile offspring of division is warranted is beyond the scope of this both sexes (Blackler et al. 1965; Blackler & Fischberg study, we do note that genetic variation within X. la- 1968). Thus, gene flow between these species is possi- evis, X. victorianus, and X. poweri is substantial. Within ble. Analysis of additional material from poorly sam- X. laevis, differentiated populations were identified in pled regions therefore could provide novel insights into the following regions: (i) south-western Western Cape the nature of gene flow among species and populations Province, (ii) Niewoudtville, (iii) Kimberley, Victoria identified here. West and Malawi. The south-western Western Cape Province lineage is comprised of two geographically Phylogeographic implications clustered demes with admixture detected at the location of Laingsburg. Individuals from south-western Vegetation in sub-Saharan Africa can be broadly classi- and north-eastern South Africa also differ in body size fied into ‘savanna’ habitat, which is open habitat where - and in the frequency of naturally occurring testicular aC4 carbon fixation grass layer exists, and ‘non savanna’ oocytes (Du Preez et al. 2009). Within X. poweri,a (i.e. tropical forest) habitat, which is closed and lacks a population from Cameroon and Nigeria is differenti- C4 carbon fixation grass layer, with the distribution of ated from a population from Botswana and Zambia. each habitat type being largely dependent on the extent Clades within X. laevis and within X. poweri were and seasonality of rainfall (Jacobs 2004; Lehmann et al. delimited from one another by the species delimita- 2011). The distributions of these habitat types cycled tion program BP&P. However, significant evidence during climatic oscillations, with savanna habitat was recovered for isolation by distance using a partial becoming more extensive or shifting to lower latitudes Mantel test (data not shown), and these data therefore during glacial periods (Dupont 2011). Within these hab- violate an assumption (panmixia) of the BP&P analy- itat types, there is also variation in the seasonality of sis. Within X. victorianus, the population from or near rain, a factor that may have played a role in the differ- the Lendu Plateau is differentiated from other popula- entiation of Xenopus laevis in South Africa (Grohovaz tions. The finding of substantial genetic differentiation et al. 1996). Thus, over the last 15 Myr or so, the evolu- in these species supports the point made by Du Preez tion of X. laevis sensu lato took place on a varied and et al. (2009) that the geographic provenance of experi- dynamic ecological and climatic landscape. It is also mental animals is an important experimental variable likely that geological features had an impact on popula- because of among-population variation in genetic tion structure within X. laevis. In particular, the Great backgrounds. Escarpment (Fig. 3) lies between the population that Our results, which include some of the individuals ranges from Victoria West to Malawi and another popu- from Central Africa studied by Evans et al. (2004) and lation that ranges from Beaufort West to Laignsburg Du Preez et al. (2009), but a different suite of individu- (Fig. 3). To the south-west of the Great Escarpment, the als sampled in South Africa from Du Preez et al. (2009), Cape Fold Belt, including the Swartberge Range and

© 2015 John Wiley & Sons Ltd 922 B. L. S. FURMAN ET AL. the Langeberg Range (Fig. 3), lie between the Beaufort of populations of X. laevis sensu lato from Nigeria and West/Laignsburg population and the coastal population Cameroon to X. poweri instead of X. laevis and with the in the south-western Western Cape Province, South assignment of populations of X. laevis sensu lato from Africa. The Niewoudtville population is also on top of Botswana and Zambia to X. poweri instead of X. peters- the Great Escarpment, and has a zone of contact with ii. In doing so, this study clarifies the evolutionary his- the south-western Western Cape Province population tory of one of the most intensively studied amphibian nearby in Vredendal (Measey & Channing 2003), which species in the context of its closely related relatives, is at the bottom of the Great Escarpment. and identifies additional differentiated populations that We present four molecular clock analyses (mitochon- may themselves be meritorious of species status. drial DNA, DM-W, concatenated autosomal DNA and species tree analysis of autosomal DNA) that assumed a Acknowledgements strict molecular clock that was calibrated in two different ways (Methods). Despite these different calibration We thank four external reviewers, each of whom provided extensive and helpful comments on earlier versions of this approaches, divergence times were quite similar across manuscript. We thank P. Staab, D. Metzler, J. Measey, J. McGu- these analyses, although this does not necessarily indi- ire, and W. Conradie for helpful discussion or comments on cate that these estimates are accurate. We resorted to this manuscript, Brian Golding for access to computer relatively crude models of evolution in these analyses resources, and T. Hayes, L. Kalous, R. Tinsley, and P. Wagner in order to achieve convergence on the posterior distri- for providing genetic samples. We thank M. D. Picker, Z. T. bution of the parameters. Clearly, error in divergence Nagy, M. M. Aristote, W. M. Moninga, M. Zigabe, A. M. Mar- times and evolutionary relationships could arise due to cel, M. Luhumyo, J. F. Akuku, F. I. Alonda, A. M’Mema, F. B. Murutsi, B. Bajope, M. Manunu, M. Collet, and the Institut model misspecification, and other model violations. For * Congolais pour la Conservation de la Nature for hospitality, example, because BEAST does not account for migration, assistance with fieldwork, logistical support and permits. Major divergence times may be underestimated in the pres- support for this study was provided by the National Science ence of migration (Leache et al. 2014). and Engineering Research Council of Canada (RGPIN/283102- Although our divergence estimates for species within 2012 to B.J.E.). Additional support was from IVB institutional X. laevis sensu lato predate the Pleistocene, the geo- support (RVO: 68081766), the Ministry of Culture of the Czech graphic locations of proposed Pleistocene savanna refu- Republic (DKRVO 2015/15, National Museum, 00023272), the Percy Sladen Memorial Fund, an IUCN/SSC Amphibian Spe- gia (see Fig. 2 in Lorenzen et al. 2012) coincide with the cialist Group Seed Grant, K. Reed, M.D., research funds from distributions of diverged evolutionary lineages in X. la- the Department of Biology at Villanova University, a National evis sensu lato. A possible mechanism for these congru- Geographic Research and Exploration Grant (no. 8556-08), the ent areas of endemism is that diversification of many of University of Texas at El Paso and the National Science Foun- these evolutionary lineages, including X. laevis sensu dation (DEB-1145459). lato, was sculpted by the distributions and connectivity of suitable habitat, which waned and waxed over time. Being mostly aquatic, Xenopus presumably are particu- References € larly sensitive to ecological factors such as the abun- Ahl E (1924) Uber eine Froschsammlung aus Nordost-Afrika dance and seasonality of rainfall that affect und Arabien. Mitteilungen aus dem Zoologischen Museum in opportunities for dispersal over land and time to Berlin, 11,1–12. complete metamorphosis. Regions with consistently Ali JR, Aitchison JC (2008) Gondwana to Asia: plate tectonics, paleogeography and the biological connectivity of the Indian habitable habitats potentially acted as ‘lifeboats’ that sub-continent from the Middle Jurrasic through latest Eocene sustained divergent lineages that evolved before habitat (166–35 Ma). Earth-Science Reviews, 88, 145–166. contraction (Evans et al. 2004). AmphibiaWeb (2014) Information on Amphibian Biology and Conservation [Web Application]. AmphibiaWeb, Berkeley, Califor- Conclusions nia. Bewick AJ, Anderson DW, Evans BJ (2011) Evolution of the This study reports for the first time, evidence of phy- closely related, sex-related genes DM-W and DMRT1 in Afri- – logenetic discordance between two putatively mater- can clawed frogs (Xenopus). Evolution, 65, 698 712. Bewick AJ, Chain FJJ, Heled J, Evans BJ (2012) The pipid root. nally inherited genomic regions, mitochondrial DNA Systematic Biology, 61, 913–926. and DM-W, in the frog lineage X. laevis sensu lato.We Blackler AW, Fischberg M (1968) Hybridization of Xenopus la- do not know of a methodological explanation for this evis petersi (poweri) and X. l. laevis. Revue Suisse de Zoologie, discordance, opening the possibility that there is a bio- 75, 1023–1032. logical cause. Results also support the recognition of Blackler AW, Fischberg M, Newth DR (1965) Hybridisation of X. laevis sensu stricto, X. victorianus, X. petersii and two subspecies of Xenopus laevis (Daudin). Revue Suisse de Zo- – newly revalidated X. poweri, but with the assignment ologie, 72, 841 857.

© 2015 John Wiley & Sons Ltd EVOLUTION OF XENOPUS LAEVIS 923 du Bocage JVB (1895) Herpetologie d’Angola et du Congo : ouvrage Evans BJ, Kelley DB, Melnick DJ, Cannatella DC (2005) Evolu- publie sous les auspices du Ministere de la marine et des colonies. tion of RAG-1 in polyploid clawed frogs. Molecular Biology Impr. Nationale, Lisbonne. and Evolution, 22, 1193–1207. Bryant D, Moulton V (2004) Neighbor-Net: an agglomerative Evans BJ, Carter TF, Hanner R et al. (2008) A new species of method for the construction of phylogenetic network. Molec- clawed frog (genus Xenopus), from the Itombwe Plateau, ular Biology and Evolution, 21, 255–265. Democratic Republic of the Congo: implications for DNA Cannatella DC, de Sa RO (1993) Xenopus laevis as a model barcodes and biodiversity conservation. Zootaxa, 1780,55–68. organism. Systematic Biology, 42, 476–507. Evans BJ, Greenbaum E, Kusamba C et al. (2011) Description of Carr S, Brothers A, Wilson A (1987) Evolutionary inferences a new octoploid frog species (Anura: Pipidae: Xenopus) from from restriction maps of mitochondrial DNA from nine taxa the Democratic Republic of the Congo, with a discussion of of Xenopus frogs. Evolution, 41, 176–188. the biogeography of African clawed frogs in the Albertine Chain FJJ, Evans BJ (2006) Multiple mechanisms promote the Rift. Journal of Zoology, London, 283, 276–290. retained expression of gene duplicates in the tetraploid frog Falush D, Stephens M, Pritchard JK (2003) Inference of pop- Xenopus laevis. PLoS Genetics, 2, e56. ulation structure using multilocus genotype data: linked Chain FJJ, Dushoff J, Evans BJ (2011) The odds of duplicate gene loci and correlated allele frequencies. Genetics, 164, 1567– persistence after polyploidization. BMC Genomics, 12, 599. 1587. Chen C, Durand E, Forbes F, Francßois O (2007) Bayesian clus- Fogell DJ, Tolley KA, Measey GJ (2013) Mind the gaps: investi- tering algorithms ascertaining spatial population structure: a gating the cause of the current range disjunction in the Cape new computer program and a comparison study. Molecular Platanna, Xenopus gilli (Anura: Pipidae). PeerJ, 1, e166. Ecology Notes, 7, 747–756. Francßois O, Durand E (2010) Spatially explicit Bayesian cluster- Daudin FM (1802) Histoire Naturelle des Rainettes, des Grenouilles ing models in population genetics. Molecular Ecology et des Crapauds. Levrault, Paris. Resources, 10, 773–784. Drummond AJ, Rambaut A (2007) BEAST: Bayesian evolutionary Frost DR (2014) Amphibian Species of the World: an Online Refer- analysis by sampling trees. BMC Evolutionary Biology, 7, 214. ence. Version 6 (27 August 2014). American Museum of Nat- Drummond AJ, Suchard MA, Xie D, Rambaut A (2012) Bayes- ural History, New York. ian phylogenetics with BEAUti and the BEAST 1.7. Molecular Galtier N, Enard D, Radondy Y, Basin E, Belkhir K (2006) Biology and Evolution, 29, 1969–1973. Mutation hotspots in mammalian mitochondrial DNA. Gen- Du Preez LH, Kunene N, Hanner R et al. (2009) Population- ome Research, 16, 215–222. specific occurrence of testicular ovarian follicles in Xenopus Grohovaz GS, Harley B, Fabran B (1996) Significant mitochon- laevis from South Africa. Aquatic Toxicology, 95,10–16. drial DNA sequence divergence in natural populations of Dupont L (2011) Orbital scale vegetation change in Africa. Qua- Xenopus laevis (Pipidae) from South Africa. Herpetologica, 52, ternary Science Reviews, 30, 3589–3602. 247–253. Durand E, Jay F, Gaggiotti OE, Francßois O (2009) Spatial infer- Gurdon JB (1996) Introductory comments: Xenopus as a labora- ence of admixture proportions and secondary contact zones. tory animal. In: The Biology of Xenopus (eds Tinsley RC, Kobel Molecular Biology and Evolution, 26, 1963–1973. HR), pp. 3–6. Oxford University Press, Oxford. Earl DA, von Holdt BM (2012) Structure Harvester: a website Gurdon JB, Hopwood N (2000) The introduction of Xenopus la- and program for visualizing Structure output and imple- evis into developmental biology: of empire, pregnancy testing menting the Evanno method. Conservation Genetics Resources, and ribosomal genes. International Journal of Developmental 4, 359–361. Biology, 44,43–50. Eggert C, Fouquet A (2006) A preliminary biotelemetric study Hasegawa M, Kishino H, Yano T (1985) Dating of the human- of a feral invasive Xenopus laevis population in France. Alytes, ape splitting by a molecular clock of mitochondrial DNA. 23, 144–149. Journal of Molecular Evolution, 21, 160–174. Evanno G, Regnaut S, Goudet J (2005) Detecting the number of Heled J, Drummond AJ (2010) Bayesian inference of species clusters of individuals using the software structure: a simula- trees from multilocus data. Molecular Biology and Evolution, tion study. Molecular Ecology, 14, 2611–2620. 27, 570–580. Evans BJ (2007) Ancestry influences the fate of duplicated Hellsten U, Khokha MK, Grammer TC et al. (2007) Accelerated genes millions of years after duplication in allopolyploid gene evolution and subfunctionalization in the pseudotetrap- clawed frogs (Xenopus). Genetics, 176, 1119–1130. loid frog Xenopus laevis. BMC Evolutionary Biology, 5, 31. Evans BJ, Morales JC, Picker MD, Kelley DB, Melnick DJ (1997) Hewitt J (1927) Further descriptions of reptiles and batrachians Comparative molecular phylogeography of two Xenopus spe- from South Africa. Record of the Albany Museum. Grahams- cies, X. gilli and X. laevis, in the southwestern Cape Prov- town, 3, 371–415. ince, South Africa. Molecular Ecology, 6, 333–343. Huson DH, Bryant D (2006) Application of phylogenetic net- Evans BJ, Morales JC, Picker MD, Kelley DB, Melnick DJ (1998) works in evolutionary studies. Molecular Biology and Evolu- Absence of extensive introgression between Xenopus gilli and tion, 23, 254–267. Xenopus laevis laevis (Anura: Pipidae), in southwestern Cape Innan H, Nordborg M (2002) Recombination or mutational hot Province, South Africa. Copeia, 1998, 504–509. spots in human mtDNA? Molecular Biology and Evolution, 19, Evans BJ, Kelley DB, Tinsley RC, Melnick DJ, Cannatella DC 1122–1127. (2004) A mitochondrial DNA phylogeny of clawed frogs: Jacobs BF (2004) Palaeobotanical studies from tropical Africa: phylogeography on sub-Saharan Africa and implications for relevance to the evolution of forest, woodland and savanna polyploid evolution. Molecular Phylogenetics and Evolution, 33, biomes. Philosophical Transactions of the Royal Society of London 197–213. B, 359, 1573–1583.

Jakobsson M, Rosenberg NA (2007) CLUMPP: a cluster match- Xenopus laevis and Xenopus tropicalis reveals post-tetraploi- ing and permutation program for dealing with label switch- dization transcriptome remodeling. Genome Research, 16, ing and multimodality in analysis of population structure. 796–803. Bioinformatics, 23, 1801–1806. Nylander JAA (2004) MrModeltest v2. Evolutionary Biology Kelley DB, Tobias ML (1999) The vocal repertoire of Xenopus Centre, Uppsala University, Uppsala. laevis. In: The Design of Animal Communication (eds Hauser M, Parker HW (1936) Reptiles and amphibians collected by the Konishi M), pp. 9–35. MIT Press, Cambridge. Lake Rudoph Rift Valley Expedition, 1934. Annals and Maga- Kimura M, Ohta T (1972) On the stochastic model for estima- zine of Natural History, Series 10, 18, 594–609. tion of mutational distance between homologous proteins. Perret JL (1966) Les amphibiens du Cameroun. Zoologische Jah- € Journal of Molecular Evolution, 2,87–90. rbucher.€ Abteilung fur€ Systematik, Okologie und Geographie, 93, Kobel HR, Loumont C, Tinsley RC (1996) The extant species. 289–464. In: The Biology of Xenopus (eds Tinsley RC, Kobel HR), pp. 9– Perrin N (2009) Sex reversal: a fountain of youth for sex chro- 33. Clarendon Press, Oxford. mosomes? Evolution, 63, 3043–3049. Kobel H, Barandun B, Thiebaud CH (1998) Mitochondrial Picker MD (1993) Embryonic tolerance of Xenopus (Anura) to rDNA phylogeny in Xenopus. Herpetological Journal, 8,13–17. acidic blackwater. Copeia, 3, 574–580. Kuck€ P, Mayer C, Wagele€ J-W, Misof B (2012) Long branch Picker MD, Harrison JA, Wallace D (1996) Natural hybridiza- effects distort maximum likelihood phylogenies in simulations tion between Xenopus laevis laevis and X. gilli in the south- despite selection of the correct model. PLoS ONE, 7, e36593. western Cape Province, South Africa. In: The Biology of Leache AD, Harris RB, Rannala B, Yang Z (2014) The influence Xenopus (eds Tinsley RC, Kobel HR), pp. 61–71. Clarendon of gene flow on species tree estimation: a simulation study. Press, Oxford. Systematic Biology, 63,17–30. Piganeau G, Gardner M, Eyre-Walker A (2004) A broad survey Lehmann CER, Archibald SA, Hoffman WA, Bond WJ (2011) of recombination in animal mitochondria. Molecular Biology Deciphering the distribution of the savanna biome. New Phy- and Evolution, 21, 2319–2325. tologist, 191, 197–209. Pitman WC III, Cande S, LaBrecque J, Pindell J (1993) Frag- Lemmon AR, Moriarty EC (2004) The importance of proper mentation of Gondwana: the separation of Africa from South model assumptions in Bayesian phylogenetics. Systematic America. In: Biological Relationships Between Africa and South Biology, 53, 265–277. America (ed. Goldblatt P), pp. 15–34. Yale University Press, Librado R, Rozas J (2009) DnaSP v5: a software for comprehen- New Haven. sive analysis of DNA polymorphism data. Bioinformatics, 25, Poynton JC (1964) The Amphibia of southern Africa: a faunal 1451–1452. study. Annals of the Natal Museum, 17,1–334. Lobos G, KJaksic FM (2005) The ongoing invasion of African Poynton JC, Broadley DG (1985) Amphibia Zambesiaca 1. Scol- clawed frogs (Xenopus laevis) in Chile: causes of concern. Bio- ecomorphidae, Pipidae, Microhylidae, Hemisidae, Arthrolep- diversity and Conservation, 14, 429–439. tidae. Annals of the Natal Museum, 26, 503–553. Lorenzen ED, Heller R, Siegismund HR (2012) Comparative Pritchard JK, Stephens M, Donnelley JM (2000) Inference of phylogeography of African savannah ungulates. Molecular population structure using multilocus genotype data. Genet- Ecology, 21, 3656–3670. ics, 155, 945–959. Loveridge A (1932) New races of a skink (Siaphos) and frog de Queiroz K (1998) The general lineage concept of species, (Xenopus) from the Uganda Protectorate. Proceedings of the species criteria, and the process of speciation. In: Endless Biological Society of Washington, 45, 113–116. Forms: Species and Speciation (eds Howard D, Berlocher S), Maisey JG (2000) Continental break up and the distribution of pp. 57–75. Oxford Press, New York. fishes of Western Gondwana during the Early Cretaceous. de Queiroz K (2007) Species concepts and species delimitation. Cretaceous Research, 21, 281–314. Systematic Biology, 56, 879–886. McCoid MJ, Fritts TH (1980) Observations of feral populations Rambaut A, Drummond AJ (2007) Tracer v1.5, Available from of Xenopus laevis (Pipidae) in southern California. Bulletin of http://beast.bio.ed.ac.uk/Tracer. the Southern California Academy of Sciences, 79,82–86. Robert J, Abramowitz L, Gantress J, Morales HD (2007) Xeno- McCoid MJ, Fritts TH (1989) Growth and fatbody cycles in pus laevis: a possible vector of ranavirus infection? Journal of feral populations of the African clawed frog: Xenopus laevis Wildlife Diseases, 43, 645–652. (Pipidae), in California with comments on reproduction. The Roure B, Baurain D, Phillippe H (2013) Impact of missing data Southwestern Naturalist, 34, 499–505. on phylogenies inferred from empirical phylogenomic data McLoughlin S (2001) The breakup history of Gondwana and its sets. Molecular Biology and Evolution, 30, 197–214. impact on pre-Cenozoic floristic provincialism. Australian Schmidt KP, Inger RF (1959) Amphibians exclusive of the gen- Journal of Botany, 49, 271–300. era Afrixalus and Hyperolius. Exploration du Parc National de Measey GJ, Channing A (2003) Phylogeography of the genus l’Upemba, Mission G.F. de Witte, en Collaboration avec W. Adam, Xenopus in southern Africa. Amphibia-Reptilia, 24, 321–330. A. Janssens, L. Van Meel, et R. Verheyen (1946–1949), 56,1–264. Measey GJ, Tinsley RC (1998) Feral Xenopus laevis in South Semon M, Wolfe KH (2008) Preferential subfunctionalization of Wales. Herpetological Journal, 8,23–27. slow-evolving genes after allopolyploidization in Xenopus la- Measey GJ, Rodder D, Green SL et al. (2012) Ongoing invasions evis. Proceedings of the National Academy of Sciences of the Uni- of the African clawed frog, Xenopus laevis: a global review. ted States of America, 105, 8333–8338. Biological Invasions, 14, 2255–2270. Sereno PC, Wilson JA, Conrad JL (2004) New dinosaurs link Morin RD, Chang E, Petrescu A et al. (2006) Sequencing southern landmasses in the Mid-Cretaceous. Proceedings of and analysis of 10,967 full-length cDNA clones from the Royal Society of London Series B, 271, 1325–1330.

Shapiro HA, Zwarenstein H (1934) A rapid test for pregnancy Yang Z, Rannala B (2010) Bayesian species delimitations using of Xenopus laevis. Nature, 133, 762. multilocus sequence data. Proceedings of the National Academy Simmonds MP (1985) Interactions between Xenopus species in of Sciences of the United States of America, 107, 9264–9269. the southwestern Cape Province, South Africa. South African Yoshimoto S, Okada E, Umemoto H et al. (2008) A W-linked Journal of Science, 81, 200. DM-domain gene, DM-W, participates in primary ovary South African Frog Re-assessment Group (SA-FRoG) ISASG development in Xenopus laevis. Proceedings of the National (2013) Xenopus gilli. In: IUCN Red List of Threatened Species. Academy of Sciences of the United States of America, 105, 2469– Version 2013.1. IUCN 2013. 2474. Spiegelhalter SD, Best NG, Carlin BP, Lingde AVD (2002) Bayes- ian measures of model complexity and fit. Journal of the Royal Statistical Society, Series B (Statistical Methodology), 64, 583–639. Stephens M, Donnelly P (2003) A comparison of Bayesian The goals of this study were developed by B.L.S.F., methods for haplotype reconstruction. American Journal of A.J.B. and B.J.E. Fieldwork for this study was per- Human Genetics, 73, 1162–1169. formed by B.L.S.F., A.J.B., E.G., V.G., C.K. and B.J.E. Stephens M, Smith NJ, Donnelly P (2001) A new statistical Molecular data were collected by B.L.S.F., A.J.B., T.L.H. method for haplotype reconstruction from population data. and B.J.E. Analysis was performed by B.L.S.F., A.J.B., 86 – American Journal of Human Genetics, , 978 989. T.L.H. and B.J.E. The paper was initially written by Stock€ M, Horn A, Grossen C et al. (2011) Ever-young sex chro- B.L.S.F., E.G., V.G. and B.J.E., and all authors then edi- mosomes in European tree frogs. PLoS Biology, 9, e1001062. Sun S, Evans BJ, Golding GB (2011) Heterotachy and false posi- ted the paper. tives for recombination in animal mtDNA. Molecular Biology and Evolution, 28, 2549–2559. Tajima F (1989) Statistical method for testing the neutral mutation hypothesisbyDNApolymorphism.Genetics,123,585–595. Data accessibility Tavare S (1986) Some probabilistic and statistical problems on the analysis of DNA sequences. Lectures on Mathematics in the Sequence data in this study have been deposited in Life Sciences, 17, 277–290. GenBank (new sequences not previously in GenBank Tinsley RC (1981) Interactions between Xenopus species (Anura have Accession nos. KP343951– KP345838), input files – Pipidae). Monitore Zoologico Italiano, 15, 133 150. and tree files have been deposited in Dryad Tinsley RC, McCoid MJ (1996) Feral populations of Xenopus (doi:10.5061/dryad.4n2c4) and sampling localities are outside Africa. In: The Biology of Xenopus (eds Tinsley RC, Kobel HR), pp. 81–94. Clarendon Press, Oxford. available in the Table S1. Tinsley RC, Loumont C, Kobel HR (1996) Geographical distribution and ecology. In: The Biology of Xenopus (eds Tinsley RC, Kobel HR), pp. 35–59. Clarendon Press, Oxford. Tobias ML, Viswanathan SS, Kelley DB (1998) Rapping, a Supporting information female receptive call, initiates male-female duets in the Additional supporting information may be found in the online ver- South African clawed frog. Proceedings of the National Acad- sion of this article. emy of Sciences of the United States of America, 95, 1870–1875. Tobias M, O’Hagan R, Horng S, Kelley D (2004) Vocal commu- Table S1 Locality information for genetic samples used in this nication between male Xenopus laevis; behavioral context and study and details on sequence data collected for each sample. sexual state. Animal Behavior, 67, 353–365. Tobias ML, Evans BJ, Kelley DB (2011) Evolution of advertise- Fig. S1 Gene networks of the 15 autosomal loci for X. laevis. ment call in African clawed frogs. Behaviour, 148, 519–549. Fig. S2 Tess analysis with all individuals from North and Tsaousis AD, Martin DP, Ladoukakis D, Posada D, Zouros E Central Africa (n = 41), with a reduced representation of indi- (2005) Widespread recombination in published animal mtDNA viduals from South Africa (n = 25), including (a) individual sequences. Molecular Biology and Evolution, 22, 925–933. assignments and (b) the deviance information criterion (DIC) Tymowska J (1991) Polyploidy and cytogenetic variation in begins to level off at values of K greater than 5, supporting 6–7 frogs of the genus Xenopus. In: Amphibian Cytogenetics and clusters. Labeling follows Fig. 5. Evolution (eds Green DS, Sessions SK), pp. 259–297. Aca- demic Press, San Diego. Fig. S3 Structure analysis of (A) 135 X. laevis individuals for 15 Weisman AI, Coates CW (1941) The “frog” test (Xenopus laevis), loci with labeling following Fig. 5. (b) The DK statistic (Evanno as a rapid diagnostic test for early pregnancy. Endocrinology, et al. 2005) indicates support for 5 clusters, which corresponds 28, 141–142. with the value of K where the log-likelihood (lnL) begins to Weldon C, du Preez L, Hyatt AD, Muller R, Speare R (2004) level off. Origin of the amphibian chytrid fungus. Emerging Infectious Diseases, 19, 2100–2105.

Taxonomy

The Amphibian Species of the World database (Frost, 2014) and AmphibiaWeb

(AmphibiaWeb, 2014) list Xenopus laevis sensu lato as comprising three species: X. laevis (the African clawed frog), X. petersii (Peters’ clawed frog; including X. poweri as a junior synonym), and X. victorianus (the Lake Victoria clawed frog). As currently recognized by these two databases, Xenopus laevis includes populations from southern Africa (e.g. South Africa, Namibia, Botswana, Zimbabwe, southern

Zambia, Malawi) and the northwestern portion of Central Africa (e.g. Central African

Republic, Cameroon, eastern Nigeria, Fig. 1), and would thus include at least two named subspecies: X. l. laevis and X. l. sudanensis (Perret, 1966) in northern Central

Africa.

Xenopus petersii, originally described by du Bocage (1895) based on specimens from

Angola, was later classified as a subspecies of X. laevis (Parker, 1936). The taxon was again recognized as a distinct species by Channing (2001), and support for this classification was also argued on the basis of divergence in mitochondrial DNA

(Measey & Channing, 2003). Xenopus poweri (Hewitt, 1927) was considered a subspecies of X. laevis by some authors (Schmidt & Inger, 1959; Poynton, 1964), or a synonym with X. (laevis) petersii (e.g. Parker, 1936; Poynton & Broadley, 1985). If X. poweri is a synonym of X. petersii, this species occurs in northern Namibia, northern

1 Botswana, northern Zimbabwe, Zambia, Angola, the Democratic Republic of the

Congo, Republic of the Congo, and southern Gabon (Fig. 1; AmphibiaWeb, 2014;

Frost, 2014).

Xenopus victorianus was described by Ahl (1924), subsequently synonymized with

X. laevis (Loveridge, 1925; Loveridge, 1933), and was again recognized as a species by Channing and Howell (2006) and Pickersgill (2007). Xenopus victorianus occurs in Tanzania, Burundi, Rwanda, eastern Democratic Republic of the Congo, Uganda, southern South Sudan, and Kenya (Fig. 1; AmphibiaWeb, 2014; Frost, 2014). The type locality of the subspecies X. laevis bunyoniensis (the western shore of Lake

Bunyonyi in southwest Uganda), described by Loveridge (1932), is also situated in this region.

There are multiple lines of evidence to support the recognition of X. laevis, X. victorianus, and X. petersii as separate species. For instance, X. laevis is larger than X. petersii and X. victorianus, with females averaging 110 mm snout–vent length (Kobel et al., 1996), although X. victorianus is of comparable size to X. petersii (females are

~62 or 65 mm respectively). Xenopus laevis has flattened black claws; dorsal and ventral patterning and coloration is highly variable, and the venter is often white or grey. Xenopus victorianus has more slender black claws, and highly variable coloration of the venter (Loveridge, 1933), whereas Xenopus petersii has narrow rounded black claws with a mottled black venter (Loveridge, 1933). Additionally, X. laevis and X. petersii differ in the number of sensory organs surrounding the eye (the

2 latter has 14 instead of 17; Channing, 2001). All three of these species have distinct mating calls (Vigny, 1979; Tobias et al., 2011); X. laevis has pronounced intraspecific variation in mating calls between animals from Malawi and the Western Cape

Province, South Africa (Tobias et al., 2011).

The recognition of X. laevis as including populations from southern Africa and also populations from Central Africa (Fig. 1; AmphibiaWeb, 2014; Frost, 2014) is inconsistent with the current understanding of evolutionary relationships within X. laevis sensu lato, which suggests that populations from West Central Africa are more closely related to populations from East Africa than to those from southern Africa

(Evans et al., 2004; Evans et al., 2011a; Evans et al., 2011b). If taxonomy is to reflect evolutionary relationships among diverged lineages, this could be reconciled either by (i) splitting X. laevis sensu lato into more than one species – but not in the way that X. laevis, X. victorianus, and X. petersii are currently recognized by at least two taxonomic databases, or by (ii) recognizing only X. laevis, which would include the currently recognized X. victorianus and X. petersii.

References

Ahl E (1924) Über eine Froschsammlung aus Nordost-Afrika und Arabien. Mitteilungen aus dem Zoologischen Museum in Berlin 11, 1–12. AmphibiaWeb (2014) Information on amphibian biology and conservation [web application]. Berkeley, California: AmphibiaWeb.

3 Channing A (2001) Amphibians of Central and Southern Africa. Cornell University Press, Ithaca. Channing A, Howell KM (2006) Amphibians of East Africa. Cornell University Press, Ithaca. Cuvier GLCFD (1830) Le Règne Animal Distribué d'Après son Organisation, pour Servir de Base à l'Histoire Naturelle des Animaux et d'Introduction à l'Anatomie Comparée. Nouvelle Edition, Revue et Augmentée par P.A. Latreille., Paris: Deterville. Daudin FM (1802) Histoire Naturelle des Rainettes, des Grenouilles et des Crapauds. Levrault, Paris. du Bocage JVB (1895) Herpétologie d'Angola et du Congo : ouvrage publié sous les auspices du Ministère de la marine et des colonies. Impr. Nationale, Lisbonne. Evans BJ, Bliss SM, Mendel SA, Tinsley RC (2011a) The Rift Valley is a major barrier to dispersal of African clawed frogs (Xenopus) in Ethiopia. Molecular Ecology 20, 4216–4230. Evans BJ, Greenbaum E, Kusamba C, et al. (2011b) Description of a new octoploid frog species (Anura: Pipidae: Xenopus) from the Democratic Republic of the Congo, with a discussion of the biogeography of African clawed frogs in the Albertine Rift. Journal of Zoology, London 283, 276–290. Evans BJ, Kelley DB, Tinsley RC, Melnick DJ, Cannatella DC (2004) A mitochondrial DNA phylogeny of clawed frogs: phylogeography on sub-Saharan Africa and implications for polyploid evolution. Molecular Phylogenetics and Evolution 33, 197–213. Frost DR (2014) Amphibian Species of the World: an Online Reference. Version 6 (27 August 2014). American Museum of Natural History, New York, USA. Hewitt J (1927) Further descriptions of reptiles and batrachians from South Africa. Record of the Albany Museum. Grahamstown 3, 371–415. Kobel HR, Loumont C, Tinsley RC (1996) The extant species. In: The Biology of Xenopus (eds. Tinsley RC, Kobel HR), pp. 9–33. Clarendon Press, Oxford.

4 Loveridge A (1925) Notes on East African batrachians, collected 1920–1923, with the description of four new species. Proceedings of the Zoological Society of London 1925, 763–791. Loveridge A (1932) New races of a skink (Siaphos) and frog (Xenopus) from the Uganda Protectorate. . Proceedings of the Biological Society of Washington 45, 113–116. Loveridge A (1933) Reports on the scientific results of an expedition to the southwestern highlands of Tanganyika Territory. VII. Herpetology. Bulletin of the Museum of Comparative Zoology 74, 197–416. Measey GJ, Channing A (2003) Phylogeography of the genus Xenopus in southern Africa. Amphibia-Reptilia 24, 321–330. Parker HW (1936) Reptiles and amphibians collected by the Lake Rudoph Rift Valley Expedition, 1934. Annals and Magazine of Natural History, Series 10 18, 594–609. Perret JL (1966) Les amphibiens du Cameroun. Zoologische Jahrbücher. Abteilung für Systematik, Ökologie und Geographie. 93, 289–464. Pickersgill M (2007) Frog Search. Results of Expeditions to Southern and Eastern Africa from 1993-1999. Chimaira, Frankfurt. Poynton JC (1964) The Amphibia of Southern Africa: a faunal study. . Annals of the Natal Museum 17, 1–334. Poynton JC, Broadley DG (1985) Amphibia Zambesiaca 1. Scolecomorphidae, Pipidae, Microhylidae, Hemisidae, Arthroleptidae. . Annals of the Natal Museum 26, 503–553. Schmidt KP, Inger RF (1959) Amphibians exclusive of the genera Afrixalus and Hyperolius. . Exploration du Parc National de l'Upemba. Mission G.F. de Witte, en Collaboration avec W. Adam ... [et al.] (1946–1949) 56, 1–264. Smith A (1831) Contributions to the natural history of South Africa, & c. South African Quarterly Journal 5, 9–24. Tobias ML, Evans BJ, Kelley DB (2011) Evolution of advertisement call in African clawed frogs. Behaviour 148, 519–549.

5 Vigny C (1979) The mating calls of 12 species and sub-species of the genus Xenopus (Amphibia: Anura). J. Zool., London 188, 103–122. Wagler J (1827) Untitled footnote. Isis von Oken 20, 726.

Supplemental Table 1. Locality information for genetic samples used in this study and details on sequence data collected for each sample (Please see excel file).

6 Supplemental Figure 1. Gene networks of the 15 autosomal loci for X. laevis. Scale bars indicate substitutions/site, and values are bootstrap support for selected major lineages, which usually connect sequences from southern and Central Africa.

Lineages from East Africa, Central Africa, and West Central Africa are shaded in green, orange, and pink, respectively.

7 Supplemental Figure 1 continued.

8 Supplemental Figure 2. Tess analysis with all individuals from North and Central Africa (n = 41), with a reduced representation of individuals from South Africa (n = 25), including (A) individual assignments and (B) the deviance information criterion (DIC) begins to level off at values of K greater than 5, supporting 6–7 clusters. Labeling follows Fig. 5.

9 Supplemental Figure 3. Structure analysis of (A) 135 X. laevis individuals for 15 loci with labeling following Fig. 5. (B) The ∆K statistic (Evanno et al. 2005) indicates support for 5 clusters, which corresponds with the value of K where the log-likelihood

(lnL) begins to level off.

10 Supplementary Table 1. Locality information for genetic samples used in this study and details on sequence data collected for each sample. Gene acronyms follow Bewick et al. 2011. The following country names are abbreviated: Democratic Republic of the Congo (DRC), the Republic of the Congo (R. Congo). Museum identification numbers (Museum IDs) refer to the Muséum d'histoire naturelle in Geneva (MHNG), the American Museum of Natural History (AMNH), the Museum of Comparative Zoology at Harvard (MCZ), the California Academy of Sciences (CAS), the Zoologisches Forschungsmuseum Alexander Koenig (ZFMK), the National Museum in Prague (NMP6V), and the University of Texas at El Paso Biodiversity Collection (UTEP). Locality numbers (Locality #) correspond to those in Fig. 1. For each locus, a "1"or a "0" indicates if data were or were not collected for each sample respectively. Museum ID Sample ID Country Locality Locality # latitude long prmt6 mogA c7orf25 nfil3 pigo sugp2 mastl zbed4 Rassf10 p7e4 fem1c znf238.2 bcl9 nufip2 AR DM-W 16A

Southern Africa (X. laevis ) – EA12-1 South Africa Western Cape Province, near 1 -34.3500 18.8167 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Betty's Bay – EA3 South Africa Western Cape Province, near 1 -34.3500 18.8167 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 Betty's Bay – EA4 South Africa Western Cape Province, near 1 -34.3500 18.8167 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 Betty's Bay – EA5 South Africa Western Cape Province, near 1 -34.3500 18.8167 0 0 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 Betty's Bay – EA6 South Africa Western Cape Province, near 1 -34.3500 18.8167 1 1 1 1 1 1 1 0 1 1 1 1 1 0 0 1 0 Betty's Bay – EA7 South Africa Western Cape Province, near 1 -34.3500 18.8167 1 1 1 1 1 1 1 1 1 1 0 1 1 0 0 1 0 Betty's Bay – KML5 South Africa Western Cape Province, near 1 -34.3470 18.9833 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 Betty's Bay – KML6 South Africa Western Cape Province, near 1 -34.3470 18.9833 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 1 0 Betty's Bay – KML7 South Africa Western Cape Province, near 1 -34.3470 18.9833 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 1 0 Betty's Bay – KML8 South Africa Western Cape Province, near 1 -34.3470 18.9833 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 1 0 Betty's Bay – LG12-3 South Africa Lewis Gay Dam, Cape Point 1 -34.2680 18.4137 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 Peninsula – RGL1 South Africa Western Cape Province, near 1 -34.3370 18.9982 0 0 1 1 1 1 1 1 1 1 1 1 1 1 0 1 0 Betty's Bay – RGL2 South Africa Western Cape Province, near 1 -34.3370 18.9982 0 0 1 1 0 1 1 1 1 1 1 1 1 1 0 1 0 Betty's Bay – RGL3 South Africa Western Cape Province, near 1 -34.3370 18.9982 0 0 0 1 1 1 1 1 1 1 1 1 1 1 0 1 0 Betty's Bay – XSL15 South Africa Western Cape Province, near 1 -34.3567 18.9333 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 Betty's Bay – XSL18 South Africa Western Cape Province, near 1 -34.3567 18.9333 0 0 0 1 0 1 0 1 0 1 0 0 0 0 0 1 0 Betty's Bay – XSL3 South Africa Western Cape Province, near 1 -34.3567 18.9333 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 Betty's Bay – XSL4 South Africa Western Cape Province, near 1 -34.3567 18.9333 0 0 1 1 1 1 1 1 1 1 0 1 1 0 0 1 0 Betty's Bay – XSL5 South Africa Western Cape Province, near 1 -34.3567 18.9333 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 1 0 Betty's Bay – BJE3505 South Africa Western Cape Province, near 2 -33.4109 19.7326 1 1 0 0 0 0 0 0 0 1 1 1 1 1 0 1 1 De Doorns, midway between Cape Town and Laignsburg, Dion Polo Vineyard – BJE3506 South Africa Western Cape Province, near 2 -33.4109 19.7326 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 De Doorns, midway between Cape Town and Laignsburg, Dion Polo Vineyard – BJE3507 South Africa Western Cape Province, near 2 -33.4109 19.7326 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 De Doorns, midway between Cape Town and Laignsburg, Dion Polo Vineyard – BJE3508 South Africa Western Cape Province, near 2 -33.4109 19.7326 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 0 1 De Doorns, midway between Cape Town and Laignsburg, Dion Polo Vineyard – BJE3509 South Africa Western Cape Province, near 2 -33.4109 19.7326 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 De Doorns, midway between Cape Town and Laignsburg, Dion Polo Vineyard – BJE3510 South Africa Western Cape Province, near 2 -33.4109 19.7326 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 De Doorns, midway between Cape Town and Laignsburg, Dion Polo Vineyard – BJE3511 South Africa Western Cape Province, near 2 -33.4109 19.7326 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 De Doorns, midway between Cape Town and Laignsburg, Dion Polo Vineyard – BJE3512 South Africa Western Cape Province, near 2 -33.4109 19.7326 1 1 1 1 1 1 1 0 1 1 1 1 1 0 1 0 1 De Doorns, midway between Cape Town and Laignsburg, Dion Polo Vineyard – BJE3513 South Africa Western Cape Province, near 2 -33.4109 19.7326 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 De Doorns, midway between Cape Town and Laignsburg, Dion Polo Vineyard – BJE3514 South Africa Western Cape Province, near 2 -33.4109 19.7326 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 De Doorns, midway between Cape Town and Laignsburg, Dion Polo Vineyard MCZ A-149234 BJE3525 South Africa Western Cape Province, 3 -33.2230 20.8519 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Laignsburg, Jacques Boubon Leftley farm MCZ A-149235 BJE3526 South Africa Western Cape Province, 3 -33.2230 20.8519 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Laignsburg, Jacques Boubon Leftley farm MCZ A-149236 BJE3527 South Africa Western Cape Province, 3 -33.2230 20.8519 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 Laignsburg, Jacques Boubon Leftley farm MCZ A-149237 BJE3528 South Africa Western Cape Province, 3 -33.2230 20.8519 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 Laignsburg, Jacques Boubon Leftley farm MCZ A-149238 BJE3529 South Africa Western Cape Province, 3 -33.2230 20.8519 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 Laignsburg, Jacques Boubon Leftley farm MCZ A-149239 BJE3530 South Africa Western Cape Province, 3 -33.2230 20.8519 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 0 1 Laignsburg, Jacques Boubon Leftley farm MCZ A-149240 BJE3531 South Africa Western Cape Province, 3 -33.2230 20.8519 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 Laignsburg, Jacques Boubon Leftley farm MCZ A-149241 BJE3532 South Africa Western Cape Province, 3 -33.2230 20.8519 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 Laignsburg, Jacques Boubon Leftley farm MCZ A-149242 BJE3533 South Africa Western Cape Province, 3 -33.2230 20.8519 1 1 1 1 1 1 1 1 1 0 1 1 1 0 1 0 1 Laignsburg, Jacques Boubon Leftley farm MCZ A-149256 BJE3547 South Africa Western Cape Province, 4 -33.9665 22.6021 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Hoekwil, Terry Smith's pond MCZ A-149257 BJE3548 South Africa Western Cape Province, 4 -33.9665 22.6021 1 1 1 1 1 1 1 1 1 0 0 1 1 1 1 1 1 Hoekwil, Terry Smith's pond MCZ A-149258 BJE3549 South Africa Western Cape Province, 4 -33.9665 22.6021 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Hoekwil, Terry Smith's pond – BJE3550 South Africa Western Cape Province, 4 -33.8924 22.8780 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 Garden Route National Park, next to Forestry Office – BJE3551 South Africa Western Cape Province, 4 -33.8924 22.8780 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 Garden Route National Park, next to Forestry Office MCZ A-140159 BJE3552 South Africa Western Cape Province, 4 -33.8924 22.8780 1 1 1 1 1 1 1 0 1 1 1 1 1 0 1 0 1 Garden Route National Park, next to Forestry Office MCZ A-140160 BJE3553 South Africa Western Cape Province, 4 -33.8924 22.8780 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 0 1 Garden Route National Park, next to Forestry Office MCZ A-140161 BJE3554 South Africa Western Cape Province, 4 -33.8924 22.8780 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 Garden Route National Park, next to Forestry Office – BJE3555 South Africa Western Cape Province, 4 -33.8924 22.8780 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 Garden Route National Park, next to Forestry Office – BJE3556 South Africa Western Cape Province, 4 -33.8924 22.8780 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 Garden Route National Park, next to Forestry Office – BJE3557 South Africa Western Cape Province, 4 -33.8924 22.8780 1 1 1 1 0 1 0 1 1 1 0 1 1 1 1 1 1 Garden Route National Park, next to Forestry Office – BJE3558 South Africa Western Cape Province, 4 -33.8924 22.8780 1 1 1 1 1 1 0 1 1 1 0 1 1 1 0 1 1 Garden Route National Park, next to Forestry Office – BJE3559 South Africa Western Cape Province, 4 -33.8924 22.8780 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 Garden Route National Park, next to Forestry Office MCZ A-149259 BJE3572 South Africa Western Cape Province, 4 -33.9009 22.9124 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 Garden Route National Park, pond in forest on right side of road traveling away from forestry office MCZ A-149243 BJE3534 South Africa Western Cape Province, 5 -32.2404 22.5858 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Beaufort West, KoTara Dam; Lemoenfontein Game Lodge, MCZ A-149244 BJE3535 South Africa Western Cape Province, 5 -32.2404 22.5858 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 Beaufort West, KoTara Dam; Lemoenfontein Game Lodge, MCZ A-149245 BJE3536 South Africa Western Cape Province, 5 -32.2404 22.5858 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 Beaufort West, KoTara Dam; Lemoenfontein Game Lodge, MCZ A-149246 BJE3537 South Africa Western Cape Province, 5 -32.2404 22.5858 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Beaufort West, KoTara Dam; Lemoenfontein Game Lodge, MCZ A-149247 BJE3538 South Africa Western Cape Province, 5 -32.2404 22.5858 0 1 1 1 1 1 1 0 1 1 1 1 1 0 1 1 1 Beaufort West, KoTara Dam; Lemoenfontein Game Lodge, MCZ A-149248 BJE3539 South Africa Western Cape Province, 5 -32.2404 22.5858 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 Beaufort West, KoTara Dam; Lemoenfontein Game Lodge, MCZ A-149249 BJE3540 South Africa Western Cape Province, 5 -32.2404 22.5858 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 Beaufort West, KoTara Dam; Lemoenfontein Game Lodge, MCZ A-149250 BJE3541 South Africa Western Cape Province, 5 -32.2404 22.5858 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 Beaufort West, KoTara Dam; Lemoenfontein Game Lodge, MCZ A-149251 BJE3542 South Africa Western Cape Province, 5 -32.2404 22.5858 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 0 1 Beaufort West, KoTara Dam; Lemoenfontein Game Lodge, MCZ A-149252 BJE3543 South Africa Western Cape Province, 5 -32.2404 22.5858 1 1 1 1 1 1 1 1 1 1 0 1 1 0 0 1 1 Beaufort West, KoTara Dam; Lemoenfontein Game Lodge, MCZ A-149253 BJE3544 South Africa Western Cape Province, 5 -32.2404 22.5858 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 Beaufort West, KoTara Dam; Lemoenfontein Game Lodge, MCZ A-149254 BJE3545 South Africa Western Cape Province, 5 -32.2404 22.5858 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Beaufort West, KoTara Dam; Lemoenfontein Game Lodge, MCZ A-149255 BJE3546 South Africa Western Cape Province, 5 -32.2404 22.5858 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 0 1 Beaufort West, KoTara Dam; Lemoenfontein Game Lodge, MCZ A-149260 BJE3573 South Africa Northern Cape Province, 6 -31.5621 23.0615 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Victoria West, Jas Fontein Farm MCZ A-149261 BJE3574 South Africa Northern Cape Province, 6 -31.5621 23.0615 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Victoria West, Jas Fontein Farm – BJE3628 South Africa Northern Cape Province, 7 -31.4747 19.0610 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 Oorlogskloof Nature Reserve, Niewoudtville – BJE3629 South Africa Northern Cape Province, 7 -31.4747 19.0610 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 Oorlogskloof Nature Reserve, Niewoudtville – BJE3630 South Africa Northern Cape Province, 7 -31.4747 19.0610 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 Oorlogskloof Nature Reserve, Niewoudtville – BJE3631 South Africa Northern Cape Province, 7 -31.4747 19.0610 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Oorlogskloof Nature Reserve, Niewoudtville – BJE3632 South Africa Northern Cape Province, 7 -31.4747 19.0610 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Oorlogskloof Nature Reserve, Niewoudtville – BJE3633 South Africa Northern Cape Province, 7 -31.4747 19.0610 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 Oorlogskloof Nature Reserve, Niewoudtville – BJE3634 South Africa Northern Cape Province, 7 -31.4747 19.0610 1 1 1 1 1 1 1 1 1 1 0 1 0 1 1 1 0 Oorlogskloof Nature Reserve, Niewoudtville – BJE3635 South Africa Northern Cape Province, 7 -31.4747 19.0610 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 Oorlogskloof Nature Reserve, Niewoudtville – BJE3636 South Africa Northern Cape Province, 7 -31.4747 19.0610 0 1 0 1 1 1 1 1 1 1 0 1 1 1 1 1 0 Oorlogskloof Nature Reserve, Niewoudtville – BJE3636 South Africa Northern Cape Province, 7 -31.4747 19.0610 0 1 0 1 1 1 1 1 1 1 0 1 1 1 1 0 0 Oorlogskloof Nature Reserve, Niewoudtville – BJE3637 South Africa Northern Cape Province, Dam 7 -31.3788 19.1035 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 near caravan park, Niewoudtville – BJE3638 South Africa Northern Cape Province, Dam 7 -31.3788 19.1035 0 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 near caravan park, Niewoudtville MCZ A-149265 BJE3639 South Africa Northern Cape Province, 7 -31.4720 19.0475 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Waternear jeep track on way to Pramkoppie, Niewoudtville MCZ A-149266 BJE3640 South Africa Northern Cape Province, 7 -31.4720 19.0475 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 Waternear jeep track on way to Pramkoppie, Niewoudtville MCZ A-149267 BJE3641 South Africa Northern Cape Province, 7 -31.4706 19.0748 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 0 1 Kareebos, Niewoutdville MCZ A-149268 BJE3642 South Africa Northern Cape Province, 7 -31.4698 19.0779 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 Joining of Rietvei and Oorlogskloof River, Niewoudtville MCZ A-149269 BJE3643 South Africa Northern Cape Province, 7 -31.4698 19.0779 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Joining of Rietvei and Oorlogskloof River, Niewoudtville MCZ A-149270 BJE3644 South Africa Northern Cape Province; 7 -31.3281 19.0837 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 William's Farm, Hollow River, Niewoudtville MCZ A-149271 BJE3645 South Africa Northern Cape Province; 7 -31.3281 19.0837 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 William's Farm, Hollow River, Niewoudtville MCZ A-149272 BJE3646 South Africa Northern Cape Province; 7 -31.3281 19.0837 1 1 1 1 1 1 1 1 1 1 0 1 1 0 1 1 1 William's Farm, Hollow River, Niewoudtville MCZ A-149273 BJE3647 South Africa Northern Cape Province, 7 -31.4023 19.0501 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 Farmer's property bordering front entrance of Oologskloof N.P., Niewoudtville MCZ A-149274 BJE3648 South Africa Northern Cape Province, 7 -31.4023 19.0501 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Farmer's property bordering front entrance of Oologskloof N.P., Niewoudtville MCZ A-149275 BJE3649 South Africa Northern Cape Province, 7 -31.4023 19.0501 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Farmer's property bordering front entrance of Oologskloof N.P., Niewoudtville MCZ A-149276 BJE3650 South Africa Northern Cape Province, 7 -31.4023 19.0501 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 Farmer's property bordering front entrance of Oologskloof N.P., Niewoudtville MCZ A-149277 BJE3651 South Africa Northern Cape Province, 7 -31.4023 19.0501 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 Farmer's property bordering front entrance of Oologskloof N.P., Niewoudtville MCZ A-149262 BJE3575 South Africa Northern Cape Province,30 km 8 -29.1093 24.5894 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 south of Kimberley, Debonaire Farm – BJE3576 South Africa Northern Cape Province,30 km 8 -29.1093 24.5894 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 0 1 south of Kimberley, Debonaire Farm – BJE3577 South Africa Northern Cape Province,30 km 8 -29.1093 24.5894 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 south of Kimberley, Debonaire Farm – BJE3578 South Africa Northern Cape Province,30 km 8 -29.1093 24.5894 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 south of Kimberley, Debonaire Farm – BJE3579 South Africa Northern Cape Province,30 km 8 -29.1093 24.5894 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 south of Kimberley, Debonaire Farm – BJE3580 South Africa Northern Cape Province,30 km 8 -29.1093 24.5894 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 south of Kimberley, Debonaire Farm – BJE3581 South Africa Northern Cape Province,30 km 8 -29.1093 24.5894 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 south of Kimberley, Debonaire Farm – BJE3582 South Africa Northern Cape Province,30 km 8 -29.1093 24.5894 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 south of Kimberley, Debonaire Farm MCZ A-149263 BJE3608 South Africa Northern Cape Province,30 km 8 -29.1093 24.5894 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 south of Kimberley, Debonaire Farm MCZ A-149264 BJE3609 South Africa Northern Cape Province,30 km 8 -29.1093 24.5894 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 south of Kimberley, Debonaire Farm – BJE3627 South Africa Northern Cape Province,30 km 8 -31.4747 19.0610 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 south of Kimberley, Debonaire Farm – PF South Africa Potchefstrum 9 -26.7150 27.1033 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 MHNG 2644.61 AMNH17301 Malawi Near Blantyre 10 -15.7800 35.0000 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

Central Africa (X. poweri ) – RT4 Botswana Okavanga 12 -18.9700 22.5700 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 – RT5 Botswana Okavanga 12 -18.9700 22.5700 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 0 1 MCZ A-148106 BJE3252 Cameroon Lower Bafut 17 6.1753 10.0746 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 0 1 MCZ A-148107 BJE3253 Cameroon Lower Bafut 17 6.1753 10.0746 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 0 1 MCZ A-148108 BJE3254 Cameroon Lower Bafut 17 6.1753 10.0746 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 MCZ A-148109 BJE3255 Cameroon Lower Bafut 17 6.1753 10.0746 0 1 1 0 1 1 0 1 1 1 1 1 1 1 1 0 1 NMP6V 74582/2 VG09_100 Cameroon Tchabal Gangdaba, loc.3 19 7.7436 12.7162 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 0 1 NMP6V 74586/2 VG09_128 Cameroon Mabor 19 7.7015 12.6767 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 0 1 – Xen058 Cameroon Ngaoundere 20 7.3200 13.5800 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 MHNG 2644.52 AMNH17259 Nigeria near Jos 18 9.9000 8.9000 1 1 1 1 1 1 1 1 0 1 0 0 0 0 1 0 1 AMNH17260 Nigeria near Jos 18 9.0000 8.9000 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 AMNH17262 Nigeria near Jos 18 9.0000 8.9000 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 MHNG 2644.53 AMNH17263 Zambia near Lusaka 11 -15.5000 28.2000 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 private 31 Zambia Ikelenge 13 -11.2379 24.2682 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 1 private 33 Zambia Ikelenge 13 -11.2379 24.2682 1 1 1 0 0 1 1 1 0 0 0 0 0 0 1 0 1 ZFMK#1 TN01 Zambia Ikelenge 13 -11.2379 24.2682 1 0 1 1 1 1 0 1 0 1 1 1 1 1 1 1 1 ZFMK#1 TN10 Zambia Ikelenge 13 -11.2379 24.2682 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 ZFMK#2 TN13 Zambia Ikelenge 13 -11.2379 24.2682 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 1 ZFMK#3 TN02 Zambia Ikelenge 13 -11.2379 24.2682 0 0 0 1 0 0 0 1 0 1 1 1 1 1 0 0 1 ZFMK#3 TN08 Zambia Ikelenge 13 -11.2379 24.2682 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 1

West central Africa (X. petersii ) NMP6V 74142 VG08_81 Angola Nequilo, Bié Province 14 -12.6000 17.0700 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 0 1 PM085 DRC Luki Reserve, forest - camp 15 -5.6167 13.1593 0 1 1 1 1 1 1 1 1 1 0 1 1 1 0 0 1 PM086 DRC Luki Reserve, forest - camp 15 -5.6167 13.1593 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 PM108 DRC Luki Reserve, forest - camp 15 -5.6167 13.1593 0 1 1 1 1 1 1 0 1 1 1 1 1 1 1 0 1 PM109 DRC Luki Reserve, forest - camp 15 -5.6167 13.1593 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 PM118 DRC Tsumbakituti, agricult. 15 -5.6581 13.1995 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 MHNG 2644.67 AMNH17324 R. Congo between Loubomo and Pointe- 16 -4.3000 12.4000 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1

East Africa (X. victorianus) UTEP 21041 ELI1010 Burundi Gihofi 29 -4.0277 30.1507 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 UTEP 21042 ELI1011 Burundi Gihofi 29 -4.0277 30.1507 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 UTEP 21043 ELI1012 Burundi Gihofi 29 -4.0277 30.1507 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 UTEP 21044 ELI1013 Burundi Gihofi 29 -4.0277 30.1507 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 UTEP 21036 ELI1064 Burundi Reserve Naturelle Rusizi, near 31 -3.3534 29.2712 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 Bujumbura UTEP 21037 ELI1065 Burundi Reserve Naturelle Rusizi, near 31 -3.3534 29.2712 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 Bujumbura UTEP 21038 ELI1066 Burundi Reserve Naturelle Rusizi, near 31 -3.3534 29.2712 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 Bujumbura UTEP 21033 ELI1139 Burundi Bubanza, near Terra Nova 31 -3.0819 29.4047 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 Hotel UTEP 21034 ELI1141 Burundi Mpishi 31 -3.0697 29.4845 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 UTEP 21035 ELI1142 Burundi Mpishi, Kibira Forest 31 -3.0697 29.4845 1 1 1 1 0 1 0 1 1 1 1 1 1 1 1 0 1 UTEP 21040 ELI938 Burundi Nyamugwaga Swamp near 31 -3.9433 29.6283 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 Bururi UTEP 21039 ELI965 Burundi Mivgaro Swamp, Bururi 31 -3.9483 29.6191 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 MCZ A-140118 BJE2897 DRC Lendu Plateau 23 1.9856 30.8613 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 0 0 MCZ A-140119 BJE2898 DRC Lendu Plateau 23 1.9856 30.8613 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 0 0 MCZ A-140121 BJE2900 DRC Lendu Plateau 23 1.9856 30.8613 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 0 0 MCZ A-140123 BJE2902 DRC Lendu Plateau 23 1.9856 30.8613 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 UTEP 20170 EBG2329 DRC Orientale Province: Banywani 25 1.5505 30.2529 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 0 1 UTEP 20174 EBG2463 DRC Orientale Province: Bunia 25 1.3387 30.1517 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 UTEP 20175 EBG2464 DRC Orientale Province: Bunia 25 1.3387 30.1517 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 UTEP 21046 CK003 DRC North Kivu Province: Virunga 26 -1.2511 29.0599 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 MCZ A-138177 BJE260 DRC Bukavu 28 -2.5028 28.8755 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 MCZ A-138179 BJE262 DRC Lwiro 28 -2.2455 28.8124 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 MCZ A-138180 BJE263 DRC Lwiro 28 -2.2455 28.8124 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 MCZ A-138183 BJE266 DRC Lwiro 28 -2.2455 28.8124 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 MCZ A-138184 BJE267 DRC Lwiro 28 -2.2455 28.8124 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 UTEP 21050 CFS1090 DRC South Kivu, Idjwi island 28 -2.0550 29.0548 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 UTEP 21051 CFS1091 DRC South Kivu, Idjwi island 28 -2.0550 29.0548 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 – LWIRO 16-1 DRC Lwiro 28 -2.2455 28.8124 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 – LWIRO 16-2 DRC Lwiro 28 -2.2455 28.8124 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 – LWIRO 17-1 DRC Lwiro 28 -2.2455 28.8124 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 – LWIRO 17-2 DRC Lwiro 28 -2.2455 28.8124 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 – LWIRO 17-3 DRC Lwiro 28 -2.2455 28.8124 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 – LWIRO 17-4 DRC Lwiro 28 -2.2455 28.8124 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 UTEP 20165 EBG2147 DRC South Kivu Province: Sangya 30 -3.8212 29.0961 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 River near Nundu UTEP 21047 ELI1298 DRC South Kivu: Baraka 30 -4.1078 29.0972 1 1 1 1 0 1 1 0 1 1 0 0 0 0 1 0 1 UTEP 21048 ELI526 DRC Byonga, South Kivu 30 -3.3403 28.1309 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 UTEP 21049 ELI527 DRC Byonga, South Kivu 30 -3.3403 28.1309 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 UTEP 21054 ELI502 DRC South Kivu: Nyamibungu 32 -3.2573 28.1180 0 1 0 0 0 1 1 1 1 1 1 1 1 1 1 0 1 UTEP 21055 ELI503 DRC South Kivu: Nyamibungu 32 -3.2573 28.1180 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 UTEP 21052 ELI1369 DRC South Kivu: Kihungwe 33 -4.0843 28.1537 1 1 1 1 1 1 0 0 1 1 1 1 1 1 1 1 1 UTEP 21053 ELI1370 DRC South Kivu: Kihungwe 33 -4.0843 28.1537 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 UTEP 21056 ELI1461 DRC South Kivu: Fizi 33 -4.3000 28.9413 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 UTEP 21057 ELI1462 DRC South Kivu: Fizi 33 -4.3000 28.9413 1 1 1 1 1 1 0 1 1 0 1 1 1 1 1 1 1 UTEP 21045 EBG2872 DRC Katanga Province: road S of 34 -6.6820 29.0811 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 0 1 Kalemie ZFMK86159 Kenya Kakamega Forest 22 0.2667 34.8833 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 ZFMK86160 Kenya Kakamega Forest 22 0.2667 34.8833 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 – xen232 Rwanda Shama 29 -1.7000 30.5500 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 CAS168711 CAS168711 Tanzania Tanga Region, Muheza District, 21 -5.0700 38.7200 1 0 1 0 0 1 0 0 1 0 0 0 0 0 1 0 1 ZFMK63119 Uganda Semliki NP 24 0.8333 30.0500 1 1 1 1 1 0 1 1 1 1 1 1 1 0 1 0 1 ZFMK63120 Uganda Semliki NP 24 0.8333 30.0500 0 1 1 1 1 1 1 1 1 0 1 1 1 1 1 0 1 – xen234 Uganda Kitanga 27 -1.1167 30.0333 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1