Amphibia-Reptilia 30 (2009): 245-257

Mitochondrial diversity of the widespread Central Asian steppetortoise (Testudo horsfieldii Gray, 1844): implications for taxonomy

and relocation of confiscated tortoises

Uwe Fritz1,*, Markus Auer1, Marina A. Chirikova2, Tatyana N. Duysebayeva2,

Valery K. Eremchenko3, Haji Gholi Kami4, Roman D. Kashkarov5, Rafaqat Masroor6,

Yoshan Moodley7, Attaullah Pindrani8, Pavel Široký9, Anna K. Hundsdörfer1

Abstract. Using a nearly range-wide sampling, we investigated phylogeographic differentiation and mitochondrial diversityof Testudo horsfieldii, the only tortoise species confined to Central Asia. We identified three major haplotype clades withmainly parapatric distribution that do not correspond well to the currently recognized three subspecies. One clade is restrictedto the Fergana Valley and seems to represent a previously overlooked evolutionarily significant unit. Another clade, consistingof several largely parapatrically distributed haplotypes, occurs in the north and the central southern part of the species’ range.The third clade, likewise comprising several largely parapatrically distributed haplotypes, was identified from the south-eastern corner of the Caspian Sea in the west, from Afghanistan and Pakistan in the east and from two more northerlysites in western and south-eastern Uzbekistan. It is possible that this clade also occurs in eastern Turkmenistan and adjacentAfghanistan, regions not sampled for the present study. The generally parapatric distribution of individual haplotypes, evenwithin each of the three major clades, suggests advanced lineage sorting, either due to limited dispersal abilities, glacialisolation in distinct local microrefuges or both acting in accord. The localized distribution of endemic haplotypes in thenorthern and central plains as well as in the mountainous eastern and southern parts of the distribution range supports theexistence of multiple microrefuges there. Records of haplotypes of distinct clades in sympatry or close geographic proximityare likely the result of Holocene range expansions. In recent years, thousands of confiscated steppe tortoises were released intothe wild. The detected mitochondrial differentiation offers a powerful tool for nature conservation, as a means of determiningthe geographic origin of confiscated tortoises and selecting suitable reintroduction regions.

Keywords: Central Asia, phylogeography, subspecies, Testudo horsfieldii horsfieldii, Testudo horsfieldii kazachstanica,Testudo horsfieldii rustamovi.

1 - Museum of Zoology (Museum für Tierkunde), NaturalHistory State Collections Dresden, A.B. Meyer Build-ing, D-01109 Dresden, Germany

2 - Institute of Zoology, Ministry of Education and Science,Al-Faraby av. 93, Almaty, 050060, Kazakhstan

3 - NABU Central Asia, Tabachnaya Street 24, Bishkek,720011, Kyrgyz Republic

4 - Department of Biology, Faculty of Sciences, GorganUniversity of Agricultural Sciences and Natural Re-sources, Gorgan, Golestan Province, Iran

5 - Department of Zoology, Biological Faculty, Na-tional University of Uzbekistan, Vuzgorodok, Tashkent,100174, Uzbekistan

6 - Pakistan Museum of Natural History, Shakarparian,Garden Avenue, Islamabad, Pakistan

7 - Max Planck Institute for Infection Biology, Departmentof Molecular Biology, Charitéplatz 1, D-10117 Berlin,Germany

8 - Sustainable Use Specialist Group Central Asia (SUSG-CAsia), Habitat and Species Conservation Project,BRSP House, 5-A Saryab Road, PO Box 216, Quetta,Pakistan

Introduction

The small to medium-sized Testudo horsfieldiiGray, 1844 (maximum shell length approx. 30cm) is the only tortoise species confined toCentral Asia. Its range encompasses easternIran, Afghanistan, Turkmenistan, Uzbekistan,Tajikistan, Kyrgyzstan, southern Kazakhstan,the westernmost part of Xinjiang (China),and in Pakistan part of the region borderingAfghanistan (Bannikov et al., 1977; Iverson,1992; Kuzmin, 2002; Fritz and Havaš, 2007).This is an area of about 2500 km in an east-west

9 - Department of Biology and Wildlife Diseases, Facultyof Veterinary Hygiene and Ecology, University of Vet-erinary and Pharmaceutical Sciences, Palackého 1-3,CZ-612 42 Brno, Czech Republic∗Corresponding author; e-mail:uwe.fritz@snsd.smwk.sachsen.de

© Koninklijke Brill NV, Leiden, 2009. Also available online - www.brill.nl/amre

246 U. Fritz et al.

direction and 1500 km in a north-south direc-tion. The natural history of the steppe tortoisediffers significantly from that of its Mediter-ranean congeners. Testudo horsfieldii is a spe-cialized burrow-dwelling steppicolous and de-serticolous species with a very short activity pe-riod to cope with extreme summer heat and win-ter cold (Kami, 1999; Kuzmin, 2002; Lagardeet al., 2002). For instance, steppe tortoises areactive for less than three months in southernUzbekistan (Lagarde et al., 2002). As requiredby their burrow-digging mode of life, steppetortoises have a distinctive morphology. Theirstrong fore-feet are shovel-like with reduceddigits and phalanges; the shell is almost as longas broad and conspicuously depressed (Ban-nikov et al., 1977; Ernst et al., 2000; Kuzmin,2002; Hitschfeld et al., 2008). Due to the mor-phological distinctiveness of T. horsfieldii, itsgeneric allocation to Testudo has repeatedlybeen challenged. However, removal from thisgenus seems unwarranted since all phylogeneticanalyses using a five-gene data set (mtDNA:12S rRNA, 16S rRNA, cyt b; nDNA: C-mos,Rag2) of approximately two-thirds of all testu-dinid species consistently found the five Testudospecies monophyletic, albeit with moderate sup-port values (Fritz and Bininda-Emonds, 2007).

In contrast to many Mediterranean species(e.g., Weiss and Ferrand, 2006; Schmitt, 2007),little is known about the phylogeography ofCentral Asian species (Fritz et al., 2008), mak-ing T. horsfieldii an attractive object for suchinvestigations. In the Mediterranean, phylogeo-graphic structures of widespread reptiles gen-erally correlate with their glacial refugia (e.g.,Joger et al., 2007; Schmitt 2007; Weiss andFerrand, 2007). Therefore, we expect that thephylogeography of a widely distributed CentralAsian species like T. horsfieldii could be help-ful in locating its refuges as well. Moreover, ob-served phylogeographic structures should cor-respond to subspecific differentiation, whereasconflicts between subspecies delineations andphylogeographic patterns often provide evi-dence for the need of taxonomic revisions

(Avise, 2000; Zink and Barrowclough, 2008).Currently, three subspecies of T. horsfieldii arerecognized (Fritz and Havaš, 2007), and thesediffer mainly in shell shape, size and col-oration (Chkhikvadze, 1988; Chkhikvadze etal., 1999; Kuzmin, 2002). However, such char-acters proved to be of little help for reveal-ing evolutionarily significant units in other tes-tudinids (Testudo: Carretero et al., 2005; Fritzet al., 2005, 2006, 2007; Attum et al., 2007;Široký and Fritz, 2007; Indotestudo: Ives etal., 2008). Although the exact distribution ofthe three putative T. horsfieldii subspecies isnot entirely clear, T. h. kazachstanica (Chkhik-vadze, 1988) is thought to occur in the north-ern part of the range; T. h. rustamovi (Chkhik-vadze, Amiranashvili and Ataev, 1990) in south-western Turkmenistan, adjacent Iran and south-western Kazakhstan; and T. h. horsfieldii Gray,1844 in the rest of the species’ range (Fritzand Havaš, 2007; Vasilyev et al., 2008). A fewyears ago, it was proposed that these three sub-species, as well as Testudo baluchiorum An-nandale, 1906, described from what is now theAfghan-Pakistani border region and previouslyconsidered a junior synonym of T. h. horsfieldii,represent distinct species (Perälä, 2002; Vet-ter, 2002). But a recent investigation using 12SrRNA sequences of 59 samples from Iran, east-ern Uzbekistan and eastern Kazakhstan foundlittle variation and concluded that the observedvariation supports conspecificity of the studiedpopulations (Vasilyev et al., 2008). In this pa-per, only four haplotypes differing by 1-3 siteswere identified. One haplotype occurred in allnorthern sites of Uzbekistan and Kazakhstan;the other three were found in four localities inIran and south-easternmost Uzbekistan. How-ever, the 12S rRNA gene is slowly evolving andonly moderately informative in revealing phy-logeographic differentiation of chelonians. Forinstance, low variation was found among 12SrRNA sequences of T. graeca (van der Kuyl etal., 2002; Harris et al., 2003), while the morevariable mitochondrial cytochrome b gene (cytb) yielded a much more detailed and clearly

Diversity of Testudo horsfieldii 247

structured phylogeographic pattern (Fritz et al.,2007, 2009).

In the present paper, we use cyt b sequencevariation in a nearly range-wide sampling of thesteppe tortoise to address the following ques-tions: (i) is T. horsfieldii a phylogeographi-cally structured species, and if so, (ii) does thephylogeographic pattern correlate with putativeglacial refuges, and (iii) does it match the rangesof the currently recognized subspecies?

Beyond science, a better knowledge of thephylogeography of T. horsfieldii could alsodirectly improve future conservation strate-gies. In the former Central Asian Soviet Re-publics, the species is massively harvested forthe pet-trade. As corollary of the implementa-tion of CITES, thousands of confiscated tor-toises have been released into the wild, of-ten without knowledge of their geographic ori-gin (M.A. Chirikova, T.N. Duysebayeva, V.K.Eremchenko, R.D. Kashkarov, own observ.; T.Harder, pers. comm.) and risking ‘genetic pol-lution’ of native populations. Given that a clearphylogeographic structure exists, mitochondrialhaplotyping of confiscated tortoises would helpto determine their geographic origin and to se-lect suitable reintroduction regions.

Materials and methods

Sampling

Eighty blood or saliva samples of Testudo horsfieldii werefield-collected, representing populations of most of thespecies’ range and all three currently recognized subspecies(samples were assigned to subspecies according to their ge-ographic origin). In addition, nine cyt b sequences fromGenBank and 11 tissue samples of pet-trade tortoises fromthe collection of the Museum of Zoology Dresden wereused. This total sample corresponds to 58 sampling sites(Appendix).

DNA extraction, PCR and sequencing

Total genomic DNA was extracted by overnight incubationat 55◦C in lysis buffer (6% DTAB, 1.125 M NaCl, 75 mMTris-HCl, 37.5 mM EDTA, pH 8.0) including 0.5 mg of pro-teinase K (Merck, Whitehouse Station, NJ) and subsequentpurification following the DTAB method (Gustincich et al.,1991). DNA was precipitated from the supernatant with 0.2

volumes of 4 M LiCl and 0.8 volumes of isopropanol, cen-trifuged, washed, dried and resuspended in TE buffer.

Our target sequence was an mtDNA fragment contain-ing the complete cyt b gene and approximately 20 bp ofthe adjacent tRNA-Thr gene. For polymerase chain reaction(PCR) and sequencing, the primers CytbG (Spinks et al.,2004), mt-c-For2, mt-f-na3, and mt-E-Rev2 (Praschag et al.,2007) were used. PCR amplification, PCR product purifi-cation and sequencing followed Praschag et al. (2007). Se-quencing was performed on an ABI 3130 (Applied Biosys-tems, Foster City, CA). None of the sequences contained in-ternal stop codons, and nucleotide frequencies correspondedto those of coding mtDNA; therefore we conclude to haveamplified and sequenced mtDNA and not nuclear copies ofmitochondrial genes.

Sequence analyses

The 100 Testudo horsfieldii sequences were manually col-lapsed into haplotypes, resulting in 26 distinct haplotypes.Six of these haplotypes are GenBank sequences and 15 hap-lotypes were found only once (Appendix). Five of the Gen-Bank haplotypes were represented by 429-bp-long to 1063-bp-long sequences only, while the sixth GenBank haplo-type and our own sequences were 1167 bp long. Phyloge-netic analyses were run for two data sets, one including all26 haplotypes and another with the shorter GenBank hap-lotypes removed. Data were analysed under the optimalitycriteria Maximum Parsimony (MP; equal weighting, com-mand: hs add = cl) and Maximum Likelihood (ML) usingPAUP*4.0b10 (Swofford, 2002) and under Bayesian infer-ence of phylogeny (BA) using MrBayes 3.1 (Ronquist andHuelsenbeck, 2003; settings: ngen = 10 000 000 nchains =4 nrun = 1 sample = 500 temp = 0.2 savebrlens = yes start-ingtree = random; burn-in set to sample only the plateauof most likely trees). Two Genbank sequences of a closelyand a distantly related testudinid species served as out-groups (AM230496, Testudo h. hermanni; DQ497301, Stig-mochelys p. pardalis). Testudo hermanni is the sister speciesof T. horsfieldii (Fritz and Bininda-Emonds, 2007). For thedata set of 26 haplotypes, 1103 of 1167 aligned characterswere constant in the ingroup sequences; 31 characters werevariable and parsimony-informative; 33 variable characterswere singletons. In the smaller data set, 1122 sites wereconstant; 25 were variable and parsimony-informative and20 variable characters were parsimony-uninformative. Thebest evolutionary model was established using Modeltest3.06 (Posada and Crandall, 1998; AIC best-fit model: TrN +I). Bootstrap support values were obtained in PAUP*4.0b10(MP: nreps = 1000, maxtree = 100; ML: nreps = 100, max-tree = 1). Uncorrected p distances were calculated usingMEGA 3.1 (Kumar et al., 2004).

Within the same species or between closely relatedspecies, relationships of haplotypes are likely to be retic-ulate and ancestral haplotypes may persist, which is why in-traspecific gene evolution may be only imperfectly reflectedby dichotomous trees (Posada and Crandall, 2001). There-fore, we also calculated a parsimony network using TCS1.21 (Clement et al., 2000). This software is based on sta-tistical parsimony and connects haplotypes via a minimal

248 U. Fritz et al.

number of mutational steps and allows for alternative path-ways. A further advantage of network analyses is that infor-mation about the age of haplotypes may be obtained. Interi-orly located haplotypes, having more than one mutationalconnection, are thought to be ancestral to and older thantip haplotypes (Posada and Crandall, 2001). To avoid am-biguities caused by much shorter GenBank sequences, weincluded in network analysis only the 91 sequences of 1167bp length produced in this investigation, the correspondingfragment of the complete mitochondrial genome from Gen-Bank (DQ080045), and in the last position of the data set,the 1140-bp-long GenBank sequence DQ497322.

Further, population-based analyses were performed forthe 1167-bp-long sequences to better elucidate past demo-graphic events as well as population structure. The softwareDnaSP 4.1 (Rozas et al., 2003) was used to carry out mis-match distribution analyses (Rogers and Harpending, 1992)as well as Fu and Li’s (1993) tests for neutrality. A Man-tel regression of a pairwise geographic distance matrix (thatis, great-arc distances obtained from Cartesian co-ordinates)with a pairwise genetic matrix of uncorrected p distances,

obtained in MEGA 3.1, was performed to ascertain if thedistribution of genetic diversity across the species was spa-tially correlated.

Results

For both data sets, all phylogenetic analyses re-vealed the same three major clades (indicatedin red, blue and black in fig. 1). The basalbranching pattern of these three major cladesvaried between the tree-building methods, how-ever. While ML suggested, with weak support,that the ‘red clade’ forms the sister group of aclade comprising the ‘blue clade’ and the ‘blackclade’, Bayesian analysis placed all of the threemajor clades in a basal polytomy. Under MP,

Figure 1. One of nine ML trees for 26 mtDNA haplotypes of Testudo horsfieldii (1167 bp, cyt b and partial tRNA-Thrgenes; − ln L = 2990.5257). The other eight ML trees differ only in the branching pattern within the ‘red clade’. Majorhaplotype clades indicated on the right. Individual known-locality haplotypes are indicated by letters; haplotypes identifiedin more than one individual in upper case and bold; haplotypes identified only once, lower case. For haplotypes of unknowngeographic origin, GenBank accession numbers or voucher numbers are given (MTD T = Museum of Zoology Dresden,tissue collection). AJ888365, AY678410, AY678438, ‘r’, and ‘s’ are haplotypes represented by shorter GenBank sequences(429-1063 bp). Numbers above nodes are ML bootstrap values; below nodes, Bayesian posterior probabilities and MPbootstraps. The large asterisk indicates this branch was not found under BA and MP (see text). For the data set of 21haplotypes, one ML tree with the same major topology was obtained (− ln L = 2815.0805). Support values were for this dataset (ML/BA/MP): red clade, 72/0.85/70; blue clade, 92/0.93/92, black clade, 95/1.0/95; under ML: (blue clade + blackclade), 52; under MP: (red clade + black clade), 53.

Diversity of Testudo horsfieldii 249

Table 1. Uncorrected p distances (percentages) of the three Testudo hors-fieldii clades (all 26 haplotypes). Below diagonal, distances between clades;on the diagonal, within-clade distances in bold (average and ranges).

Red Blue Black

Red 0.50 (0.09-1.37)Blue 1.01 (0.60-1.65) 0.09 (–)Black 1.57 (0.72-2.79) 1.31 (0.94-2.23) 1.02 (0.09-2.03)

24 equally parsimonious trees were obtained forthe 26 haplotype data set (tree length = 294;CI = 0.9252, RI = 0.8866); the 21 haplotypedata set resulted in two equally parsimonioustrees (tree length = 272; CI = 0.8105, RI =0.8896). In all MP trees, a basal position ofthe ‘blue clade’ and a sister group relationshipof the ‘red clade’ and the ‘black clade’ werefavoured, although with weak bootstrap support(

250 U. Fritz et al.

Figure 2. Geographic distribution of mtDNA haplotypes of Testudo horsfieldii. Colours correspond to other figures. Noteoccurrence of haplotypes of the ‘red clade’ and the ‘black clade’ in the same sites or in close geographic proximity. Simplifiedrange borders of the three currently recognized subspecies indicated. Stars represent type localities: 1 – T. h. rustamovi(Chkhikvadze, Amiranashvili and Ataev, 1990); 2 – T. h. kazachstanica (Chkhikvadze, 1988); 3 – T. h. horsfieldii Gray, 1844;4 – ‘T. baluchiorum Annandale, 1906’.

population is expected to result in a peak inthe distribution of pairwise differences (plottedred lines) between sequences (Rogers and Harp-ending, 1992), whereas a population maintain-ing a constant size through time is expected tofollow a trend similar to that plotted by bluelines. Analyses of the entire data set, as wellas for each clade, showed a large proportion ofclosely related haplotypes. However, a past sud-den population expansion is most likely to haveoccurred in the ‘black clade’, while the ‘redclade’ shows moderate signs of demographicexpansion. The ‘blue clade’ perfectly fits themodel of constant population size. Fu and Li’s(1993) D* and F* indices of neutrality were allslightly negative for all tested groups (fig. 4),and in the case of a neutral marker such as mi-tochondrial DNA, this indicates a greater num-ber of pairwise differences relative to segregat-ing sites and a possibly expanding population.

However, none of these indices were significant.A Mantel regression found that only 24% of thevariation in the genetic data could be attributedto geographic variation, precluding the assump-tion of a clinal population structure of isolation-by-distance.

Discussion

Although we found only moderate variation inthe studied mtDNA fragment, our results pro-vide evidence that Testudo horsfieldii is a phylo-geographically structured species. We identifiedthree major clades of haplotypes with mainlyparapatric ranges. Even within each of the threemajor clades, the more frequent haplotypes havelargely parapatric distributions, suggestive ofadvanced lineage sorting, either due to lim-ited dispersal abilities, glacial isolation in dis-

Diversity of Testudo horsfieldii 251

Figure 3. Parsimony network of 93 mtDNA sequencesof Testudo horsfieldii (1167 bp, cyt b and partial tRNA-Thr genes; spring tree). Colours correspond to other fig-ures; symbol size, approximate haplotype frequency; miss-ing node haplotypes, open circles. Each line joining hap-lotypes indicates one substitution except when hatchmarksacross lines are present; then each mark indicates one step.Only known-locality haplotypes bear letters. Haplotype fre-quencies are B: n = 4; D: n = 14; E: n = 3; F: n = 27; I:n = 12; K: n = 7; M: n = 2; O: n = 6; P: n = 6; T: n = 2.Other haplotypes were identified only once.

tinct local microrefuges or both acting in con-cert. The average uncorrected p distances of thethree clades (table 1) fall within the range ob-served between North African subspecies of T.graeca, arguing for Early to Middle Pleistocenedifferentiation in both cases and in favour ofrepeated glacial fragmentation (cf. Fritz et al.,2009). Furthermore, the localized distributionof endemic haplotypes in the northern and cen-tral plains as well as in the mountainous easternand southern parts of the distribution range ofT. horsfieldii supports the existence of multiplemicrorefuges. This is in line with our results of

the Mantel test, indicating discontinuous diver-gence in geographic distribution.

The three major clades of T. horsfieldii do notagree well with its currently recognized threesubspecies and also do not support the validityof T. baluchiorum Annandale, 1906, describedfrom the Afghan-Pakistani border region (fig.2). One of our three major clades, the ‘blueclade’ from the Fergana Valley within the rangeof T. h. kazachstanica, seems to represent apreviously overlooked evolutionarily significantunit. Its occurrence in the Fergana Valley sug-gests that microrefuges were located not onlywithin the extant southern range, but also fairlyfar north. The ‘blue clade’, isolated in the Fer-gana Valley, has maintained a constant popu-lation size throughout its demographic history,while the other clades show at least some evi-dence of recent population expansion that seemsto be related to Holocene warming and range ex-tension.

Except the Fergana Valley, the distribution ofour ‘red clade’ in the north matches the rangeof T. h. kazachstanica and the distribution ofthe northern 12S rRNA haplotype of Vasilyev etal. (2008). However, in the south the same ‘redclade’ penetrates deep into the putative rangesof T. h. horsfieldii and T. h. rustamovi, propos-ing a distinctly larger range of T. h. kazachstan-ica than thought before. The other haplotypesidentified from the ranges of T. h. horsfieldii andT. h. rustamovi represent another major clade(‘black clade’), suggesting weak differentiationand synonymy of both subspecies. In some sitesin Iran and western and south-eastern Uzbek-istan, haplotypes of the ‘black clade’ were re-corded together with or in close proximity tohaplotypes of the ‘red clade’. This, as well asthe record of a southern 12S rRNA haplotypein south-eastern Uzbekistan (Vasilyev et al.,2008), may indicate the existence of secondarycontact zones due to Holocene range expansionsand a wider distribution of the ‘black clade’in Turkmenistan and northern Afghanistan, re-gions not sampled for the present study.

252 U. Fritz et al.

Figure 4. Mismatch distributions and demographic indices for the entire data set (panel A) and for each of the threephylogenetic clades (panels B-D). Observed mismatch distributions are graphed in grey columns, whereas line distributionscorrespond to those expected under a model of constant population size (blue) and a model of population expansion (red).NS, not significant.

In the former Central Asian Soviet Republics,steppe tortoises have been heavily exploited forthe pet-trade for decades. From 1976 to 1993,more than one million steppe tortoises werecollected in Kazakhstan alone (United NationsEnvironment Programme, World ConservationMonitoring Centre, 2004). With the implemen-tation of CITES, collection and export quo-tas were introduced (Rhodin, 2003; Bykova etal., 2007). For instance, in 2002 Kazakhstan,Tajikistan and Uzbekistan had export quotasof 40 000, 20 000 and 30 000 live tortoises, re-spectively (United Nations Environment Pro-gramme, World Conservation Monitoring Cen-tre, 2004). Because of such harvest quotas andpoaching, the number of confiscated tortoisesincreased considerably and thousands of confis-cated tortoises were released into the wild, of-ten without knowledge of their geographic ori-

gin (M.A. Chirikova, T.N. Duysebayeva, V.K.Eremchenko, R.D. Kashkarov, own observ.; T.Harder, pers. comm.). While we excluded fromsampling any sites that could harbour such re-leased tortoises, it is obvious that the currentrepatriation practice bears the risk of admixtureof distinct genetic lineages and should not becontinued.

Conclusions

Testudo horsfieldii is a phylogeographicallystructured species. Its differentiation patternstrongly suggests that the current diversity wasshaped by glacial range interruptions. Refugesand microrefuges were most probably not onlylocated in the south of the extant range, butalso fairly far north. The weak agreement be-tween subspecies and phylogroups of T. hors-

Diversity of Testudo horsfieldii 253

fieldii suggests that the morphological char-acters used for subspecies delineation are oflimited taxonomic value, as in other Testudospecies (Carretero et al., 2005; Fritz et al., 2005,2006, 2007, 2009; Attum et al., 2007; Širokýand Fritz, 2007). This is also underlined bythe fact that some allegedly diagnostic osteo-logical characters of T. horsfieldii subspecies(Perälä, 2002) were found to be ontogeneticvariation (Hitschfeld et al., 2008). One of thethree mtDNA clades identified in the present pa-per seems to represent a previously overlookedevolutionarily significant unit. Future investiga-tions should re-analyze morphological variationin the light of phylogeographic differentiation tounravel the impact of ecological constraints andancestry.

The phylogeographic pattern of T. horsfieldiiargues for great caution with further releasesof confiscated steppe tortoises without knowl-edge of their origin. However, our study of-fers a highly conservation-relevant applicationof phylogeography. Mitochondrial haplotypingof confiscated tortoises would allow for a muchbetter selection of suitable releasing sites, basedon the largely allopatric distribution of individ-ual haplotypes.

Acknowledgements. Thorsten Harder (NABU CentralAsia) helped with field work in Kyrgyzstan and providedvaluable information. Thanks go to the local nature conser-vation authorities and the German Agency for Nature Con-servation (BfN) for permits. Lab work was done by AnkeMüller (Museum of Zoology Dresden).

References

Attum, O., Baha el Din, S., Carranza, S., Earley, R., Arnold,E.N., Kingsbury, B. (2007): An evaluation of the tax-onomic validity of Testudo werneri. Amphibia-Reptilia28: 393-401.

Avise, J.C. (2000): Phylogeography. The History and For-mation of Species. Cambridge, MA and London, Har-vard University Press.

Bannikov, A.G., Darevskii, I.S., Ishchenko, V.G., Rustamov,A.K., Shcherbak, N.N. (1977): Opredelitel’ zemnovod-nykh i presmykayushchikhsya fauny SSSR. Moscow,Prosveshcheniye.

Bykova, E.V., Sorochinsky, V.G., Golenkevich, A.V., Pere-gontsev, E.A., Sorochinskaya, I.N., Sorochinsky, G.Y.(2007): Ranching as a method of conservation of thewild population of Horsfield’s tortoise, Agrionemyshorsfieldii. Russ. J. Herpetol. 14: 232-236.

Carretero, M.A., Znari, M., Harris, D.J., Macé, J.C. (2005):Morphological divergence among populations of Tes-tudo graeca from west-central Morocco. Animal Biol.55: 259-279.

Chkhikvadze, V.M. (1988): O sistematicheskom polozheniisovremennykh sukhoputnykh cherepakh Srednei Azii iKazakhstana. Izv. Akad. Nauk GSSR, Ser. biol. 14: 110-113.

Chkhikvadze, V.M., Amiranashvili, N.G., Ataev, Ch.(1990): Novyi podvid sukhoputnoi cherepakhi iz yugo-zapadnogo Turkmenistana. Izv. Akad. Nauk Turkmensk.SSR, Ser. biol. 1: 72-75.

Clement, M., Posada, D., Crandall, K.A. (2000): TCS: acomputer program to estimate gene genealogies. Mol.Ecol. 9: 1657-1660.

Ernst, C.H., Altenburg, R.G.M., Barbour, R.W. (2000): Tur-tles of the World. World Biodiversity Database, CD-ROM Series, Windows Version 1.2. Amsterdam, Biodi-versity Center of ETI.

Fritz, U., Bininda-Emonds, O.R.P. (2007): When genesmeet nomenclature: Tortoise phylogeny and the shiftinggeneric concepts of Testudo and Geochelone. Zoology110: 298-307.

Fritz, U., Havaš, P. (2007): Checklist of chelonians of theworld. Vertebr. Zool. 57: 149-368.

Fritz, U., Široký, P., Kami, H., Wink, M. (2005): Environ-mentally caused dwarfism or a valid species – Is Testudoweissingeri Bour, 1996 a distinct evolutionary lineage?New evidence from mitochondrial and nuclear genomicmarkers. Mol. Phylogenet. Evol. 37: 389-401.

Fritz, U., Auer, M., Bertolero, A., Cheylan, M., Fattizzo, T.,Hundsdörfer, A.K., Martín Sampayo, M., Pretus, J.L.,Široký, P., Wink, M. (2006): A rangewide phylogeog-raphy of Hermann’s tortoise, Testudo hermanni (Rep-tilia: Testudines: Testudinidae): implications for taxon-omy. Zool. Scr. 35: 531-543.

Fritz, U., Hundsdörfer, A.K., Široký, P., Auer, M., Kami,H., Lehmann, J., Mazanaeva, L.F., Türkozan, O., Wink,M. (2007): Phenotypic plasticity leads to incongruencebetween morphology-based taxonomy and genetic dif-ferentiation in western Palaearctic tortoises (Testudograeca complex; Testudines, Testudinidae). Amphibia-Reptilia 28: 97-121.

Fritz, U., Ayaz, D., Buschbom, J., Kami, H.G., Mazanaeva,L.F., Aloufi, A.A., Auer, M., Rifai, L., Šilić, T., Hunds-dörfer, A.K. (2008): Go east: Phylogeographies ofMauremys caspica and M. rivulata – Discordance ofmorphology, mitochondrial and nuclear genomic mark-ers and rare hybridization. J. Evolution. Biol. 21: 527-540.

Fritz, U., Harris, D.J., Fahd, S., Rouag, R., Graciá Martínez,E., Giménez Casalduero, A., Široký, P., Kalboussi, M.,Jdeidi, T.B., Hundsdörfer, A.K. (2009): Mitochondr-ial phylogeography of Testudo graeca in the WesternMediterranean: Old complex divergence in North Africaand recent arrival in Europe. Amphibia-Reptilia 30: 63-80.

254 U. Fritz et al.

Fu, Y.-X., Li, W.-H. (1993): Statistical tests of neutrality ofmutations. Genetics 133: 693-709.

Gustincich, S., Manfioletti, G., del Sal, G., Schneider, C.,Carninci, C. (1991): A fast method for high-qualitygenomic DNA extraction from whole human blood.BioTechniques 11: 298-302.

Harris, D.J., Znari, M., Macé, J.-C., Carretero, M.A. (2003):Genetic variation in Testudo graeca from Morocco esti-mated using 12S rRNA sequencing. Rev. Esp. Herpetol.17: 5-9.

Hitschfeld, E., Auer, M., Fritz, U. (2008): Phalangeal for-mulae and ontogenetic variation of carpal morphology inTestudo horsfieldii and T. hermanni. Amphibia-Reptilia29: 93-99.

Iverson, J.B. (1992): A Revised Checklist with DistributionMaps of the Turtles of the World. Richmond, IN, Pri-vately Printed.

Ives, I., Spinks, P.Q., Shaffer, H.B. (2008): Morphologicaland genetic variation in the endangered Sulawesi tortoiseIndotestudo forstenii. Conserv. Genet. 9: 709-713.

Joger, U., Fritz, U., Guicking, D., Kalyabina-Hauf, S.,Nagy, Z.T., Wink, M. (2007): Phylogeography of west-ern Palaearctic reptiles – Spatial and temporal speciationpatterns. Zool. Anz. 246: 293-313.

Kami, H.G. (1999): On the biology of the Afghan Tortoise,Testudo horsfieldi, in north-eastern Iran. Zool. MiddleEast 19: 43-54.

Kumar, S., Tamura, K., Nei, M. (2004): MEGA3: Integratedsoftware for Molecular Evolutionary Genetics Analysisand sequence alignment. Brief. Bioinform. 5: 150-163.

van der Kuyl, A.C., Ballasina, D.L.P., Dekker, J.T., Maas,H., Willemsen, R.E., Goudsmit, J. (2002): Phylogeneticrelationships among the species of the genus Testudo(Testudines: Testudinidae) inferred from mitochondrial12S rRNA gene sequences. Mol. Phylogenet. Evol. 22:174-183.

Kuzmin, S.L. (2002): The Turtles of Russia and Other Ex-Soviet Republics. Frankfurt am Main, Chimaira.

Lagarde, F., Bonnet, X., Nagy, K., Henen, B., Corbin, J.,Naulleau, G. (2002): A short spring before a long jump:the ecological challenge in the steppe tortoise (Testudohorsfieldi). Can. J. Zool. 80: 493-502.

Perälä, J. (2002): Biodiversity in relatively neglected taxa ofTestudo L., 1758 s. l. Chelonii 3: 40-53.

Posada, D., Crandall, K.A. (1998): Modeltest: testing themodel of DNA substitution. Bioinformatics 14: 817-818.

Posada, D., Crandall, K.A. (2001): Intraspecific gene ge-nealogies: trees grafting into networks. Trends Ecol.Evol. 16: 37-45.

Praschag, P., Hundsdörfer, A.K., Fritz, U. (2007): Phy-logeny and taxonomy of endangered South and South-east Asian freshwater turtles elucidated by mtDNA se-quence variation (Testudines: Geoemydidae: Batagur,Callagur, Hardella, Kachuga, Pangshura). Zool. Scr.36: 429-442.

Rhodin, A.G.J. (2003): Conservation and Trade of AsianFreshwater Turtles and Tortoises: Updated Reviewof Status and Threats with Recommendations forCITES Listings. Geneva, CITES. Available at http://www.cites.org/eng/com/ac/19/E19-15-1.doc.

Rogers, A.R., Harpending, H. (1992): Population growthmakes waves in the distribution of pairwise geneticdifferences. Mol. Biol. Evol. 9: 552-569.

Ronquist, F., Huelsenbeck, J.P. (2003): MrBayes 3: Bayes-ian phylogenetic inference under mixed models. Bioin-formatics 19: 1572-1574.

Rozas, J., Sanchez-DelBarrio, J.C., Messeguer, X., Rozas,R. (2003): DnaSP, DNA polymorphism analyses by thecoalescent and other methods. Bioinformatics 19: 2496-2497.

Schmitt, T. (2007): Molecular biogeography of Europe:Pleistocene cycles and postglacial trends. Front. Zool.4 (11): 1-12.

Široký, P., Fritz, U. (2007): Is Testudo werneri a distinctspecies? Biologia 62: 228-231.

Spinks, P.Q., Shaffer, H.B., Iverson, J.B., McCord, W.P.(2004): Phylogenetic hypotheses for the turtle familyGeoemydidae. Mol. Phylogenet. Evol. 32: 164-182.

Swofford, D.L. (2002): PAUP*. Phylogenetic Analysis Us-ing Parsimony (*and Other Methods), Version 4. Sun-derland, MA, Sinauer Associates.

United Nations Environment Programme, World Conser-vation Monitoring Centre (2004): Review of Signifi-cant Trade. Analysis of Trade Trends with Notes onthe Conservation Status of Selected Species. AnnexC: Reptiles and Amphibians. Geneva, CITES. Avail-able at http://www.cites.org/common/com/ac/20/E20-inf-12.pdf.

Vasilyev, V.A., Bondarenko, D.A., Peregontsev, E.A.,Voronov, A.S., Ryskov, A.P., Semenova, S.K. (2008):Polimorfism gena 12S rRNK i filogeografiya sredneazi-atskoy cherepakhi Agrionemys horsfieldii Gray, 1844.Genetika 55: 784-788.

Vetter, H. (2002): Terralog: Turtles of the World, Vol. 1.Africa, Europe and Western Asia. Frankfurt am Main,Chimaira.

Weiss, N., Ferrand, N. (2007): Phylogeography of South-ern European Refugia. Berlin, Heidelberg, New York,Springer.

Zink, R.M., Barrowclough, G.F. (2008): MitochondrialDNA under siege in avian phylogeography. Mol. Ecol.17: 2107-2121.

Received: September 10, 2008. Accepted: December 22,2008.

Diversity of Testudo horsfieldii 255A

ppen

dix.

Sam

ples

and

Gen

Ban

kse

quen

ces

ofTe

stud

oho

rsfie

ldii

and

thei

rha

plot

ypes

.Sam

ple

num

bers

are

eith

erM

TD

Tnu

mbe

rs(M

useu

mof

Zoo

logy

Dre

sden

,tis

sue

colle

ctio

n)or

Gen

Ban

kac

cess

ion

num

bers

.Hap

loty

pes

iden

tified

inm

ore

than

one

case

inbo

ld.A

cces

sion

num

bers

ofha

plot

ypes

iden

tified

inth

isst

udy

are

FM88

3673

-FM

8836

92.

Sam

ple

Subs

peci

esL

ocal

ityL

atitu

deL

ongi

tude

Hap

loty

pe

5034

hors

field

iiA

fgha

nist

an:G

hazn

i33

◦ 33′

N68

◦ 26′

En

3049

hors

field

iiIr

an:K

hora

san:

Sarb

ishe

h32

◦ 35′

N59

◦ 48′

EK

3048

hors

field

iiIr

an:K

hora

san:

vici

nity

ofB

irja

nd32

◦ 53′

N59

◦ 12′

EK

5362

hors

field

iiPa

kist

an:H

azar

ganj

i30

◦ 02.

820′

N66

◦ 52.

945′

El

5359

hors

field

iiPa

kist

an:H

azar

ganj

i30

◦ 02.

994′

N66

◦ 52.

872′

EO

5361

hors

field

iiPa

kist

an:H

azar

ganj

i30

◦ 03.

033′

N66

◦ 53.

983′

EO

5355

hors

field

iiPa

kist

an:K

arkh

asa

30◦ 1

0.94

2′N

66◦ 5

7.25

1′E

O53

56ho

rsfie

ldii

Paki

stan

:Kuc

hlac

k30

◦ 22.

489′

N66

◦ 58.

416′

EM

5358

hors

field

iiPa

kist

an:Q

uetta

30◦ 0

8.95

7′N

66◦ 5

6.93

1′E

O53

60ho

rsfie

ldii

Paki

stan

:Que

tta30

◦ 08.

957′

N66

◦ 56.

931′

EO

5357

hors

field

iiPa

kist

an:T

orgh

ar31

◦ 11.

888′

N68

◦ 27.

359′

EM

AJ8

8836

5ka

zach

stan

ica

Kaz

akhs

tan

––

AJ8

8836

554

16ka

zach

stan

ica

Kaz

akhs

tan:

Kyz

ylku

msa

nds,

WN

WC

hard

ara

villa

ge41

◦ 14′

N67

◦ 42′

EF

5418

kaza

chst

anic

aK

azak

hsta

n:K

yzyl

kum

sand

s,W

NW

Cha

rdar

avi

llage

41◦ 1

5′N

67◦ 4

2′E

F54

15ka

zach

stan

ica

Kaz

akhs

tan:

SA

rys

sand

s42

◦ 09′

N68

◦ 25′

EF

5414

kaza

chst

anic

aK

azak

hsta

n:A

rys

sand

s42

◦ 13′

N68

◦ 19′

ED

5417

kaza

chst

anic

aK

azak

hsta

n:A

rys

sand

s42

◦ 13′

N68

◦ 19′

ED

5413

kaza

chst

anic

aK

azak

hsta

n:A

rys

sand

s42

◦ 13′

N68

◦ 19′

Eh

5412

kaza

chst

anic

aK

azak

hsta

n:E

Kyz

ylku

msa

nds;

197

mel

evat

ion

42◦ 1

6′N

67◦ 5

6′E

F54

09ka

zach

stan

ica

Kaz

akhs

tan:

Mt.

Kar

atau

43◦ 4

0′N

69◦ 0

1′E

F30

64ka

zach

stan

ica

Kaz

akhs

tan:

Jam

bylr

egio

n:M

erke

Dis

tric

t:vi

cini

tyof

Ken

esh

villa

ge43

◦ 57.

666′

N73

◦ 34.

217′

ED

3062

kaza

chst

anic

aK

azak

hsta

n:Ja

mby

lreg

ion:

Mer

keD

istr

ict:

vici

nity

ofK

enes

hvi

llage

43◦ 5

7.66

6′N

73◦ 3

4.21

7′E

F54

07ka

zach

stan

ica

Kaz

akhs

tan:

Zhu

sand

ala

valle

y,S

Tauk

umsa

nds

(sou

ther

nL

ake

Bal

khas

hre

gion

);37

7m

elev

atio

n44

◦ 30.

266′

N75

◦ 03.

697′

EF

5408

kaza

chst

anic

aK

azak

hsta

n:Z

husa

ndal

ava

lley,

STa

ukum

sand

s(s

outh

ern

Lak

eB

alkh

ash

regi

on);

377

mel

evat

ion

44◦ 3

0.26

6′N

75◦ 0

3.69

7′E

F54

05ka

zach

stan

ica

Kaz

akhs

tan:

coas

tof

Ala

kull

ake

(sou

ther

nm

argi

nof

Lak

eB

alkh

ash)

;340

mel

evat

ion

44◦ 5

0.24

9′N

74◦ 1

2.77

8′E

g54

06ka

zach

stan

ica

Kaz

akhs

tan:

10km

SEK

ashk

aten

izvi

llage

(wes

tern

coas

tof

Lak

eB

alkh

ash)

;359

mel

evat

ion

45◦ 4

2.88

3′N

73◦ 2

4.30

5′E

F54

19ka

zach

stan

ica

Kaz

akhs

tan:

SWco

astL

ake

Bal

khas

h,5

kmto

Min

aral

villa

ge45

◦ 24.

067′

N73

◦ 38.

280′

EF

5411

kaza

chst

anic

aK

azak

hsta

n:A

ralS

ea:B

arsa

kelm

esis

land

45◦ 4

0′N

59◦ 5

5′E

F54

10ka

zach

stan

ica

Kaz

akhs

tan:

Eco

astA

ralS

ea,3

0km

SEK

arat

eren

villa

ge45

◦ 47′

N61

◦ 19′

EF

3060

kaza

chst

anic

aK

yrgy

zsta

n:Fe

rgan

aV

alle

y:Ja

lalA

bad:

Kyz

yk-K

iyya

Dis

tric

t,40

◦ 23′

N72

◦ 15′

EI

18km

NE

Kyz

yl-K

iyya

tow

n,vi

cini

tyof

Kha

uzvi

llage

3061

kaza

chst

anic

aK

yrgy

zsta

n:Fe

rgan

aV

alle

y:Ja

lalA

bad:

Kyz

yk-K

iyya

Dis

tric

t,40

◦ 23′

N72

◦ 15′

EI

18km

NE

Kyz

yl-K

iyya

tow

n,vi

cini

tyof

Kha

uzvi

llage

3056

kaza

chst

anic

aK

yrgy

zsta

n:Fe

rgan

aV

alle

y:Ja

lalA

bad:

May

li-Su

uca

nyon

,7km

from

Kok

-Tas

hvi

llage

41◦ 1

4′N

72◦ 2

7′E

I

256 U. Fritz et al.A

ppen

dix.

(Con

tinue

d).

Sam

ple

Subs

peci

esL

ocal

ityL

atitu

deL

ongi

tude

Hap

loty

pe

3057

kaza

chst

anic

aK

yrgy

zsta

n:Fe

rgan

aV

alle

y:Ja

lalA

bad:

May

li-Su

uca

nyon

,7km

from

Kok

-Tas

hvi

llage

41◦ 1

4′N

72◦ 2

7′E

I30

58ka

zach

stan

ica

Kyr

gyzs

tan:

Ferg

ana

Val

ley:

Jala

lAba

d:M

ayli-

Suu

cany

on,7

kmfr

omK

ok-T

ash

villa

ge41

◦ 14′

N72

◦ 27′

EI

3059

kaza

chst

anic

aK

yrgy

zsta

n:Fe

rgan

aV

alle

y:Ja

lalA

bad:

May

li-Su

uca

nyon

,7km

from

Kok

-Tas

hvi

llage

41◦ 1

4′N

72◦ 2

7′E

j30

54ka

zach

stan

ica

Kyr

gyzs

tan:

Ferg

ana

Val

ley:

Jala

lAba

d:12

kmSW

Tash

-Kom

urto

wn

41◦ 1

8′N

72◦ 0

8′E

I30

55ka

zach

stan

ica

Kyr

gyzs

tan:

Ferg

ana

Val

ley:

Jala

lAba

d:12

kmSW

Tash

-Kom

urto

wn

41◦ 1

8′N

72◦ 0

8′E

I30

50ka

zach

stan

ica

Kyr

gyzs

tan:

Ferg

ana

Val

ley:

Jala

lAba

d:W

Ala

-Buk

avi

llage

41◦ 2

4′N

71◦ 2

8′E

I30

51ka

zach

stan

ica

Kyr

gyzs

tan:

Ferg

ana

Val

ley:

Jala

lAba

d:W

Ala

-Buk

avi

llage

41◦ 2

4′N

71◦ 2

8′E

I30

63ka

zach

stan

ica

Kyr

gyzs

tan:

Ferg

ana

Val

ley:

Jala

lAba

d:W

Ala

-Buk

avi

llage

41◦ 2

4′N

71◦ 2

8′E

I30

52ka

zach

stan

ica

Kyr

gyzs

tan:

Ferg

ana

Val

ley:

Jala

lAba

d:vi

cini

tyof

Ker

ben

(Kar

avan

)vi

llage

41◦ 2

9′N

71◦ 4

5′E

I30

53ka

zach

stan

ica

Kyr

gyzs

tan:

Ferg

ana

Val

ley:

Jala

lAba

d:vi

cini

tyof

Ker

ben

(Kar

avan

)vi

llage

41◦ 2

9′N

71◦ 4

5′E

I30

65ka

zach

stan

ica

Kyr

gyzs

tan:

Chu

yre

gion

:7.5

kmN

Tele

kvi

llage

,lef

tban

kof

Ak-

Suu

rive

r43

◦ 14′

N74

◦ 03′

NF

3066

kaza

chst

anic

aK

yrgy

zsta

n:C

huy

regi

on:7

.5km

NTe

lek

villa

ge,l

eftb

ank

ofA

k-Su

uri

ver

43◦ 1

4′N

74◦ 0

3′N

F35

32ka

zach

stan

ica

Uzb

ekis

tan:

Eco

cent

er“D

jeir

an”;

223

mel

evat

ion

39◦ 3

4.73

7′N

64◦ 4

1.29

6′E

E35

28ka

zach

stan

ica

Uzb

ekis

tan:

Zar

avsh

anN

atur

eR

eser

ve;7

20m

elev

atio

n39

◦ 40.

292′

N67

◦ 05.

492′

ED

3536

kaza

chst

anic

aU

zbek

ista

n:10

kmS

Tur

kmen

villa

ge;8

00m

elev

atio

n39

◦ 53.

520′

N68

◦ 30.

000′

EO

3533

kaza

chst

anic

aU

zbek

ista

n:20

kmW

Ora

zjan

villa

ge;3

73m

elev

atio

n40

◦ 31.

773′

N65

◦ 10.

315′

ED

3537

kaza

chst

anic

aU

zbek

ista

n:25

kmE

cros

sroa

dsJa

ngig

azga

n-Z

afar

abad

(sou

ther

nK

yzyl

kum

);35

7m

elev

atio

n40

◦ 36.

414′

N65

◦ 16.

705′

ED

3525

kaza

chst

anic

aU

zbek

ista

n:w

este

rnpa

rtof

Aja

kagi

tma

depr

essi

on;1

00m

elev

atio

n40

◦ 39.

672′

N64

◦ 26.

847′

ED

3529

kaza

chst

anic

aU

zbek

ista

n:20

kmE

Jang

igaz

gan

villa

ge;1

95m

elev

atio

n40

◦ 43.

344′

N64

◦ 52.

290′

ED

3531

kaza

chst

anic

aU

zbek

ista

n:K

arak

ata

depr

essi

on:v

icin

ityof

Azn

ekw

ell;

193

mel

evat

ion

40◦ 4

3.38

4′N

64◦ 5

2.21

9′E

D35

34ka

zach

stan

ica

Uzb

ekis

tan:

2km

WK

yzyl

cha

villa

ge(n

orth

ern

foot

hills

ofN

urat

auri

dge)

;410

mel

evat

ion

40◦ 4

4.27

1′N

66◦ 0

4.03

4′E

D35

35ka

zach

stan

ica

Uzb

ekis

tan:

10km

NA

jakk

uduk

villa

ge(n

orth

ern

foot

hills

ofSa

ngru

ntau

ridg

e);4

50m

elev

atio

n41

◦ 19.

170′

N65

◦ 05.

416′

ED

3538

kaza

chst

anic

aU

zbek

ista

n:40

kmS

Uzu

nkud

ukvi

llage

;357

mel

evat

ion

41◦ 2

3.69

4′N

63◦ 1

0.80

9′E

F35

23ka

zach

stan

ica

Uzb

ekis

tan:

Buc

hara

regi

on:K

ara-

kyr

lake

;98

mel

evat

ion

41◦ 2

4.10

2′N

63◦ 1

4.42

5′E

F35

24ka

zach

stan

ica

Uzb

ekis

tan:

wes

tern

part

ofN

urat

auri

dge:

20km

NN

urat

aci

ty;4

01m

elev

atio

n41

◦ 24.

102′

N65

◦ 47.

695′

ED

3526

kaza

chst

anic

aU

zbek

ista

n:15

kmN

Tam

dy-t

ruba

villa

ge;9

9m

elev

atio

n41

◦ 31.

885′

N64

◦ 01.

906′

EF

3527

kaza

chst

anic

aU

zbek

ista

n:D

jam

anku

msa

nds:

vici

nity

ofA

ktak

yrvi

llage

;398

mel

evat

ion

41◦ 4

4.20

4′N

63◦ 1

9.24

4′E

F35

30ka

zach

stan

ica

Uzb

ekis

tan:

Buk

anta

um

ount

ain

ridg

e:C

hing

ildy

gorg

e;41

7m

elev

atio

n42

◦ 41.

514′

N63

◦ 25.

550′

EF

3521

kaza

chst

anic

aU

zbek

ista

n:U

styu

rtpl

atea

u;24

0m

elev

atio

n43

◦ 19.

201′

N58

◦ 09.

710′

EF

3522

kaza

chst

anic

aU

zbek

ista

n:U

styu

rtpl

atea

u:K

arak

alka

n,10

0km

SEK

ungr

ad;2

46m

elev

atio

n43

◦ 30.

017′

N58

◦ 10.

006′

EF

3520

kaza

chst

anic

aU

zbek

ista

n:U

styu

rtpl

atea

u:vi

cini

tyof

Chu

ruk

wel

l;21

4m

elev

atio

n44

◦ 59.

853′

N56

◦ 54.

800′

EP

1429

rust

amov

iIr

an:S

emna

n:M

arko

uh,6

0km

SD

amgh

an35

◦ 38′

N54

◦ 23′

EB

2671

rust

amov

iIr

an:K

hora

san:

Nis

habo

or(N

eysh

abou

r)36

◦ 03′

N59

◦ 07′

EB

3027

rust

amov

iIr

an:K

hora

san:

Anj

irB

elag

,35

kmSE

Bes

anga

n,S

Shar

akhs

;834

mel

evat

ion

36◦ 0

5.59

2′N

60◦ 3

9.24

0′E

a30

26ru

stam

ovi

Iran

:Kho

rasa

n:A

njir

Bel

ag,3

5km

SEB

esan

gan,

SSh

arak

hs;8

34m

elev

atio

n36

◦ 05.

592′

N60

◦ 39.

240′

Eq

Diversity of Testudo horsfieldii 257A

ppen

dix.

(Con

tinue

d).

Sam

ple

Subs

peci

esL

ocal

ityL

atitu

deL

ongi

tude

Hap

loty

pe

2274

rust

amov

iIr

an:K

hora

san:

Bes

anga

nla

ke,5

0km

EM

assh

ad36

◦ 15′

N59

◦ 00′

EK

2275

rust

amov

iIr

an:K

hora

san:

Bes

anga

nla

ke,5

0km

EM

assh

ad36

◦ 15′

N59

◦ 00′

EK

2276

rust

amov

iIr

an:K

hora

san:

Bes

anga

nla

ke,5

0km

EM

assh

ad36

◦ 15′

N59

◦ 00′

EK

5241

rust

amov

iIr

an:K

hora

san:

Bes

anga

nla

ke,5

0km

EM

assh

ad36

◦ 15′

N59

◦ 00′

EK

3025

rust

amov

iIr

an:K

hora

san:

Ab

Bel

utch

,17

kmSE

Bes

anga

n,S

Shar

akhs

36◦ 1

6′N

60◦ 3

9′E

P30

30ru

stam

ovi

Iran

:Kho

rasa

n:K

hang

iran

,Cha

hA

rtes

ian,

near

Tur

kmen

ian

bord

er,W

Shar

akhs

;152

mel

evat

ion

36◦ 3

4.85

5′N

60◦ 4

5.63

0′E

B26

70ru

stam

ovi

Iran

:Gol

esta

n:H

ezar

pich

villa

gene

arG

orga

n36

◦ 46′

N54

◦ 28′

EB

1419

rust

amov

iIr

an:G

oles

tan:

vici

nity

ofG

orga

nci

ty36

◦ 51′

N54

◦ 28′

EP

AJ8

8836

6ru

stam

ovi

Iran

:Gol

esta

n:vi

cini

tyof

Gor

gan

city

36◦ 5

1′N

54◦ 2

8′E

r22

79ru

stam

ovi

Iran

:Gol

esta

n:M

iank

aleh

peni

nsul

a36

◦ 54.

337′

N54

◦ 00.

413′

EP

AJ8

8836

7ru

stam

ovi

Iran

:Gol

esta

n:M

iank

aleh

peni

nsul

a36

◦ 54.

337′

N54

◦ 00.

413′

Es

1422

rust

amov

iIr

an:G

oles

tan:

Mia

nkal

ehpe

nins

ula

36◦ 5

4.33

7′N

54◦ 0

0.41

3′E

T14

23ru

stam

ovi

Iran

:Gol

esta

n:M

iank

aleh

peni

nsul

a36

◦ 54.

337′

N54

◦ 00.

413′

ET

2674

rust

amov

iIr

an:K

hora

san:

Baj

gira

n37

◦ 36′

N58

◦ 24′

EK

1424

rust

amov

iIr

an:G

oles

tan:

Mor

aveh

Tapp

eh37

◦ 55′

N55

◦ 57′

EP

1425

rust

amov

iIr

an:G

oles

tan:

Mor

aveh

Tapp

eh37

◦ 55′

N55

◦ 57′

EP

DQ

0800

45ru

stam

ovi

Tur

kmen

ista

n:K

opet

Dag

37◦ 5

0′N

58◦ 1

5′E

c

AY

6784

10?

unkn

own

––

AY

6784

10A

Y67

8434

?un

know

n–

–A

Y67

8410

AY

6784

38?

unkn

own

––

AY

6784

38D

Q44

5847

?un

know

n–

–E

DQ

4973

22?

unkn

own

––

F66

?un

know

n–

–M

TD

T66

83?

unkn

own

––

F84

?un

know

n–

–F

941

?un

know

n–

–D

1296

?un

know

n–

–M

TD

T12

9616

95?

unkn

own

––

F22

56?

unkn

own

––

F34

35?

unkn

own

––

E34

36?

unkn

own

––

D44

24?

unkn

own

––

F44

25?

unkn

own

––

MT

DT

4425