Amphibia-Reptilia 34 (2013): 337-351

Gene flow among deeply divergent mtDNA lineages of Testudo

graeca (Linnaeus, 1758) in Transcaucasia

Viola Mashkaryan1,2,* , Melita Vamberger1,* , Marine Arakelyan2 , Nasim Hezaveh2 ,

Miguel A. Carretero3 , Claudia Corti4 , D. James Harris3 , Uwe Fritz1,**

Abstract. Using 10 polymorphic microsatellite loci and sequences of the mitochondrial cytochrome b gene, we examine
gene flow in more than 100 spur-thighed tortoises from Transcaucasia and compare our findings with previously published
AFLP and mtDNA data. While mtDNA sequences correspond to three deeply divergent clades and AFLP data suggest
two distinct groups, microsatellite data indicate weak differentiation and extensive gene flow. We conclude that each
marker system reflects a distinct episode in the evolutionary history of the Testudo graeca complex, corresponding to
phases of vicariance and extensive gene flow. We hypothesize that the reciprocally monophyletic and deeply divergent
mtDNA lineages reflect old vicariance events, while the conflicting nuclear markers are the legacy of younger episodes
of extensive gene flow. The differentiation pattern found in the AFLP markers, with AFLP groups matching allopatrically
or parapatrically distributed mitochondrial lineages, is likely to be older than the microsatellite differentiation, which could
correspond to Holocene range expansions into Caucasian valleys and lower altitudes. Owing to the largely mutually exclusive
distribution ranges of the deeply divergent mtDNA lineages, their identification with distinct subspecies is a reasonable and
straightforward classification that facilitates communication and acknowledges the conspecifity of the involved evolutionary
units. Consequently, Transcaucasia constitutes an intergradation zone of T. g. armeniaca, T. g. buxtoni and T. g. ibera.

Keywords: Armenia, Caucasus, Georgia, Iran, Nagorno Karabakh, phylogeography, population genetics, tortoise.

Introduction based on morphological evidence (Chkhikvadze

The taxonomy of spur-thighed tortoises (Tes-
tudo graeca complex) has been in a state
of flux during the past 25 years. While this
group had traditionally been regarded as a
widely distributed polytypic species (Wermuth
and Mertens, 1961, 1977; Anderson, 1979;
Pritchard, 1979), many subspecies were later el-
evated to species level and additional species
were described or resurrected from synonymy
1 - Museum of Zoology (Museum fr Tierkunde), Senck-
enberg Dresden, A.B. Meyer Building, 01109 Dresden,
Germany
2 - Faculty of Biology, Yerevan State University, Alek
Manoogian 1, Yerevan, 0025, Armenia
3 - CIBIO, Centro de Investigao em Biodiversidade e
Recursos Genticos, Universidade do Porto, Campus
Agrrio de Vairo, 4485-661 Vairo, Portugal
4 - Museo di Storia Naturale dellUniversit di Firenze,
Sezione di Zoologia La Specola, Via Romana, 17,
50125 Firenze, Italy
* These authors contributed equally to this publication.
** Corresponding author;
e-mail: uwe.fritz@senckenberg.de
and Tuniyev, 1986; Weissinger, 1987; High-
field and Martin, 1989a, b, c; Highfield, 1990;
Chkhikvadze and Bakradze, 1991, 2002; Perl,
1996, 2002; Pieh, 2001; Pieh and Perl, 2002,
2004; Chkhikvadze, Mazanaeva and Sham-
makov, 2011). As a result, some authors claimed
that the T. graeca complex consists of ap-
proximately 20 distinct species with mutu-
ally exclusive allopatric or parapatric ranges in
the western Mediterranean region, the Balkan
peninsula and the Near and Middle East (for a
review, see Guyot-Jackson, 2004; Fritz et al.,
2007). However, molecular genetic investiga-
tions found much less differentiation (van der
Kuyl et al., 2002; Harris et al., 2003; van der
Kuyl, Ballasina and Zorgdrager, 2005; Parham
et al., 2006; Fritz et al., 2007, 2009), with
only six deeply divergent mitochondrial lin-
eages identified. These results suggested that
many of the morphologically defined taxa are
invalid and that external morphology of spur-
thighed tortoises is heavily impacted by envi-
ronmental factors (Carretero et al., 2005; Fritz et

Koninklijke Brill NV, Leiden, 2013. DOI:10.1163/15685381-00002895

338 V. Mashkaryan et al.

Figure 1. Mitochondrial phylogeny of the Testudo graeca complex as inferred from ML analyses of a 1216-bp-long alignment
(cyt b gene plus adjacent DNA coding for tRNAs) of 140 ingroup sequences (data set of Fritz et al., 2007 combined with
46 sequences generated for the present study; two sequences of North African tortoises corrected according to Vamberger
et al., 2011). Terminal clades collapsed to cartoons; grey cartoons indicate that haplotypes of the respective clade were
represented among the samples sequenced for the present study. Numbers at nodes are ML bootstrap values and posterior
probabilities from Bayesian inference. Root length shortened by 80%. On the right, the correlation of mtDNA clades with
AFLP groups (Mikulicek et al., 2013) and microsatellite clusters (K = 3, this study) is shown. Note that mtDNA clades
A and E represent one and the same AFLP cluster, although they are not sister groups. Merging colours for microsatellite
data symbolize admixed individuals. The letter C in the Caucasian AFLP group indicates the occurrence of a mitochondrial
haplotype of clade C in a tortoise from Katekh, Azerbaijan, belonging to the Caucasian AFLP group (Mikulicek et al., 2013).

al., 2007). In the following, we use the terminol-
ogy of Fritz et al. (2007) and label the six major
mitochondrial clades by the upper-case letters
A-F (fig. 1).
Using nuclear-genomic ISSR fingerprinting,
a method useful in discriminating the tradi-
tionally recognized Testudo species T. graeca,
T. hermanni, T. horsfieldii, T. kleinmanni and
T. marginata, Fritz et al. (2007) found only
negligible differentiation among tortoises rep-
resenting these six mitochondrial clades and
concluded that they are best regarded as con-
specific. Acknowledging the deep mitochon-
drial divergences, Fritz et al. (2007) proposed
to identify each clade with a distinct sub-
species, while other authors preferred to recog-
nize no subspecies at all (Parham et al., 2006).
In a follow-up study, Mikulicek et al. (2013)
re-examined the differentiation within the T.
graeca complex using more sensitive AFLP
fingerprints and found four well-differentiated
AFLP groups. Two of these groups, the west-
ern Mediterranean AFLP group and the central-
eastern Iranian AFLP group (Mikulicek et al.,
2013), correspond to the completely or largely
allopatric mitochondrial lineages from the west-
ern Mediterranean (mtDNA clade B) and east-
ern Iran (mtDNA clade F; fig. 1). Each of the
remaining two AFLP clusters includes two ge-
ographically neighbouring mtDNA clades, sug-
gestive of extensive gene flow among tortoises
harbouring the respective mitochondrial lin-
eages. The Caucasian AFLP group comprises
the mtDNA clades A and E. These two clades
Gene flow in Testudo graeca
are not sister groups in phylogenetic analy-
ses (fig. 1). The Balkans-Middle Eastern AFLP
group embraces the mtDNA clades C and D,
which are sister groups. The ranges of these two
distinct AFLP groups abut in the central Cauca-
sus, and in one site (Katekh, Azerbaijan) two
tortoises were found that belonged to each of
the two distinct AFLP groups, without any in-
dication of nuclear genomic admixture. On the
other hand, both tortoises yielded haplotypes of
mtDNA clade C, indicating recent or past gene
flow between the Caucasian and the Balkans-
Middle Eastern AFLP groups. Since mtDNA
is known to be transferred relatively easy from
one species to another as the result of inter-
specific hybridization (Mallet, 2005; Currat et
al., 2008), this complicated situation raises the
question of whether distinct reproductively iso-
lated species might be involved (Mikulicek et
al., 2013), which hybridize only rarely.
In the present study we focus exactly on this
question: Could some of the distinct genetic lin-
eages of Caucasian spur-thighed tortoises rep-
resent reproductively isolated units qualifying
as Biological Species sensu Mayr (1942, 1963)
or do they alternatively represent one and the
same Biological Species? For testing these hy-
potheses, we use rapidly evolving microsatel-
lite markers, which are an ideal tool for as-
sessing gene flow. We genotyped more than
100 tortoises from Armenia, Georgia, Iran and
Nagorno Karabakh using 10 polymorphic mi-
crosatellite loci. The sampling region is known
to harbour three distinct mtDNA clades corre-
Table 1. Microsatellite loci, multiplex sets, allele size ranges and number of alleles of the individual loci.
Locus Original reference Multiplex set Fluorescent label
GmuB08 King and Julian (2004) 1 ATTO 565
Goag6 Edwards et al. (2003) 1 6-FAM
Test76 Forlani et al. (2005) 1 HEX
GmuD51 King and Julian (2004) 2 ATTO 550
Test56 Forlani et al. (2005) 2 HEX
Test71 Forlani et al. (2005) 2 6-FAM
Gp61 Schwartz et al. (2003) 3 6-FAM
Gp81 Schwartz et al. (2003) 3 HEX
Test10 Forlani et al. (2005) 3 HEX
Test21 Forlani et al. (2005) 3 ATTO 565
339
sponding to two AFLP clusters (Fritz et al.,
2007; Mikulicek et al., 2013). We analyze our
data set with an unsupervised Bayesian clus-
tering approach as implemented in STRUCTURE
2.3.3 (Pritchard, Stephens and Donnelly, 2000;
Hubisz et al., 2009) and compare the results
with mtDNA sequence variation and the AFLP
groups of Mikulicek et al. (2013). If distinct re-
productively isolated species were represented
among our samples, they should correspond
to clearly differentiated microsatellite clusters
with little or no evidence of gene flow. By con-
trast, if only one more-or-less panmictic species
were concerned, the expectation is little nuclear
genomic differentiation and, in secondary con-
tact zones of distinct microsatellite clusters, ex-
tensive gene flow among clusters.
Materials and methods
Sampling and laboratory procedures
DNA of 121 ethanol-preserved samples of tortoises from
26 sites in Armenia, Georgia, Iran and Nagorno Karabakh
(Appendix) was isolated using the InnuPREP DNA Mini
Kit (Analytik Jena AG, Jena, Germany) or the NucleoSpin
Tissue Kit (Macherey-Nagel, Dren, Germany). Following
Salinas et al. (2011), the samples were genotyped using
10 microsatellite loci; three or four loci each were com-
bined in multiplex PCRs (table 1). The final volume of each
multiplex PCR was 10 l containing 0.5 units Taq poly-
merase (Bioron, Ludwigshafen, Germany) with the buffer
recommended by the supplier and a final concentration of
1.5 mM MgCl2 (Bioron), 0.2 mM of each dNTP (Thermo-
Scientific, St. Leon-Rot, Germany), 2 g Bovine Serum Al-
bumin (Thermo-Scientific), approximately 10-20 ng of to-
tal DNA and primer concentrations ranging from 0.2 M
to 0.8 M. Forward primers were fluorescent-labelled. The
Allele size range [bp] Number of alleles
195-249 17
282-404 19
112-138 4
132-224 17
192-274 31
119-141 10
179-193 6
361-363 2
179-241 26
193-237 13
340 V. Mashkaryan et al.

Table 2. Testing for Hardy-Weinberg equilibrium (HWE)

PCR cycling conditions were as follows: 43 cycles with de-
using tortoises from three populations or groups of neigh-
naturation at 94C for 60 s but for 11 min in the first cy-
bouring populations with sufficient sample sizes (collecting
cle, annealing at 58C for 45 s and extension at 72C for
sites 8, 9 and sites 19-24 lumped together; see Appendix).
45 s but for 30 min in the final cycle. Fragment lengths
Yes = in HWE, no = not in HWE. Asterisks indicate pres-
were determined on an ABI 3130xl Genetic Analyzer using
ence of null alleles.
the GeneScanTM-600 LIZ Size Standard and the software
GENEMAPPER 4.0 (Applied Biosystems, Foster City, USA). Locus Population/collecting site
For 14 samples no microsatellite data could be produced,
because the respective sample was used up or PCR failed 8 9 (19-24)
consistently. GmuB08 yes yes yes
In addition, the mitochondrial cyt b gene was se- Goag6 no yes no
quenced for 46 samples representing all collecting sites (Ap- Test76 no no no
pendix). Two mtDNA fragments overlapping by approxi- GmuD51 monomorphic yes no
mately 300 bp were amplified using the primer pairs CytbG Test56 no no yes
plus mt-E-Rev2 (Spinks et al., 2004; Fritz et al., 2006) and Test71 yes yes no
mt-c-For2 plus mt-f-na (Fritz et al., 2006). PCR was per- Gp61 no no no
formed in a final volume of 20 l using 1 unit Taq poly- Gp81 monomorphic monomorphic monomorphic
merase (Bioron) with the buffer recommended by the sup- Test10 no yes yes
plier and a final concentration of 0.25 mM of each dNTP Test21 no yes no
(Thermo-Scientific), 0.5 M of each primer and approxi-
mately 10-40 ng of total DNA. The PCR cycling conditions
were as follows: 35-40 cycles with denaturation at 94C for
30 and an admixture scenario with allele frequencies cor-
45 s but for 5 min in the first cycle, annealing at 50C for
related, allowing the individuals to have mixed ancestries.
30 s, and extension at 72C for 60 s but for 10 min in the fi-
The burn-in was set to 25 104 and the number of fur-
nal cycle. PCR products were purified using the ExoSAP-IT
ther MCMC runs to 75 104 . Calculations were repeated
enzymatic cleanup (USB Europe GmbH, Staufen, Germany;
1:20 dilution; modified protocol: 30 min at 37C, 15 min 10 times for each K; convergence of likelihood values was
at 80C) and sequenced on an ABI 3130xl Genetic Ana- reached after the burn-in. Clustering results and individual
lyzer using the BigDye Terminator v3.1 Cycle Sequencing admixture were visualized using bar plots with the soft-
Kit (Life Technologies, Darmstadt, Germany) and the PCR ware DISTRUCT 1.1 (Rosenberg, 2004). In a conservative
primers. DNA sequences were aligned in BIOEDIT 7.1.11 approach following Randi (2008), individuals with propor-
(Hall, 1999) with previously published data (Fritz et al., tions for cluster membership below 80% were treated as
2007), yielding a 1216-bp-long alignment of 140 sequences having mixed ancestries.
comprising the complete cyt b gene plus adjacent DNA cod- Genetic differentiation among different subsets of the
ing for tRNAs. Two erroneous North African sequences of microsatellite data was inferred by FST values and analy-
Fritz et al. (2007) were corrected according to Vamberger et ses of molecular variance (AMOVA) using ARLEQUIN 3.11
al. (2011). GenBank accession numbers for sequences gen- (10 000 permutations). For comparing number and size of
erated for the present study are HF954117-HF954162. microsatellite alleles, a frequency table was produced us-
ing CONVERT 1.31 (Glaubitz, 2004) and locus-specific ob-
served (HO ) and expected heterozygosities (HE ) were es-
Data analysis
timated in ARLEQUIN. Locus-specific allelic richness (AR)
Pairwise linkage disequilibrium between microsatellite loci values were obtained in FSTAT 2.9.3.2 (Goudet, 1995).
and Hardy-Weinberg equilibrium were tested in ARLEQUIN Phylogenetic relationships of mtDNA sequences were
3.11 (Excoffier, Laval and Schneider, 2005) using tortoises inferred by Maximum Likelihood (ML) analyses using
from three populations or groups of neighbouring popula- RAxML 7.2.6 (Stamatakis, 2006) and the implemented evo-
tions with sufficient sample sizes (collecting sites 8, 9 and lutionary model GTR + G. Five independent ML searches
sites 19-24 lumped together; see Appendix). There was no were performed using different starting conditions and the
evidence for linkage disequilibrium, and microsatellite data fast bootstrap algorithm to explore the robustness of the
of all tortoises were then subjected to unsupervised cluster phylogenetic trees by comparing the best trees. Subse-
analysis using STRUCTURE 2.3.3 (Pritchard, Stephens and quently, 1000 non-parametric thorough bootstrap replicates
Donnelly, 2000; Hubisz et al., 2009). However, many loci were calculated and the values plotted against the best
were found to be not in Hardy-Weinberg equilibrium, and tree. In addition, analyses using MRBAYES 3.2.1 (Ronquist
MICRO - CHECKER 2.2.3 (van Oosterhout et al., 2004) sug- et al., 2012) were run. The best evolutionary model was
gested the presence of null alleles (table 2). Therefore, the established in JMODELTEST 0.1.1 (Posada, 2008) by the
data set was corrected for null alleles according to Falush, Bayesian Information Criterion, resulting in the TPM1uf +
Stephens and Pritchard (2007). The optimal number of clus- G model. Phylogenetic analyses were performed using two
ters was determined by the K method of Evanno, Regnaut parallel runs (each with four chains) and default parameters.
and Goudet (2005). Considering the number of collecting Both chains ran for 10 million generations with every 100th
sites for which microsatellite data were available (n = 26), generation sampled. Using a burn-in of 2.5 million genera-
population structure was modelled using an upper bound of tions, only the plateau of the most likely trees was sampled
Gene flow in Testudo graeca 341

for generating a 50% majority rule consensus. The posterior When genetic diversity indices for the dis-
probability of any individual clade in this consensus tree tinct STRUCTURE clusters (K = 3 and K =
corresponds to the percentage of all trees containing that
clade, and is a measure of clade frequency and credibility. 4) are compared and tortoises with mixed an-
In both approaches, a sequence of Testudo hermanni (Gen- cestries are excluded, the red cluster with the
Bank accession number AJ888362) served for tree-rooting. least amount of admixture has the lowest val-
Additionally, uncorrected p distances of mtDNA clades
were calculated using MEGA 5.05 (Tamura et al., 2011) and ues, with just two private alleles, despite more
the pairwise deletion option. or less even sample sizes (table 4). Fixation in-
dices (FST values) for K = 3 range from 0.12
to 0.21 (table 5), and according to an AMOVA,
Results 17% of the molecular variance occurs among
clusters and 83% within clusters. For K = 4,
The 10 studied microsatellite loci are in part
FST values range from 0.15 to 0.26 (table 5),
highly polymorphic, with allele numbers rang-
with 22% of the molecular variance occurring
ing from 2 to 31 per locus (table 1) and a to-
among clusters and 78% within clusters.
tal allele number of 145. The K method of
From a morphological point of view, our tor-
Evanno, Regnaut and Goudet (2005) suggested
toise samples represent two principal pheno-
for the STRUCTURE results four as the optimal
types, individuals with depressed shells, as orig-
number of clusters. When K = 4 is used as
inally described as typical for Testudo graeca
a framework, there is a high degree of admix-
armeniaca (Chkhikvadze and Bakradze, 1991;
ture evident. However, in some collecting sites
Pieh, Fritz and Berglas, 2002), and tortoises
more or less pure genotypes are detected be-
with domed shells, the normal character state in
sides admixed individuals (fig. 2). The phyloge-
the T. graeca complex. When morphologically
netic analyses of mtDNA sequences of tortoises
similar tortoises (M. Arakelyan, V. Mashkaryan,
representing all sites found three deeply diver-
M.A. Carretero, C. Corti, unpubl.) are lumped
gent mitochondrial lineages among the studied
together, three geographically coherent groups
samples (figs 1 and 2), corresponding to the
emerge (fig. 2: top). However, in the two north-
mtDNA clades A, C and E of Fritz et al. (2007).
ern groups (Araxes valley) there are also some
Mean uncorrected p distances of the cyt b gene
individuals whose morphological assignment to
between these and the remaining three other
the domed vs. flat phenotype is ambiguous. In
mtDNA clades of the Testudo graeca complex
any case, using microsatellite data these three
amount to 2.50-5.70%. In particular, the diver-
groups differ by FST values ranging from 0.10
gences between clades A, C and E range from
to 0.15 and they do very roughly correspond to
4.15% to 4.90% (table 3). Haplotypes of clade A
the three microsatellite clusters under K = 3,
are widely distributed in the Araxes valley and
including admixed individuals. According to an
Nagorno Karabakh, while haplotypes of clade C
AMOVA, 11% of the observed global molec-
are largely confined to the northernmost Arme-
ular variance occurs among the three morpho-
nian sites. However, in two sites in the Araxes
logical groups and 89% within the groups (tor-
valley haplotypes of clades A and C occur syn-
toises with mixed genetic ancestry included).
topically. Haplotypes of clade E are only found
in northwestern Iran (fig. 2; Appendix). None
of the three mtDNA clades corresponds to any
Discussion
of the four clusters revealed by the microsatel-
lite analyses. Also when barplots for K = 3 are In agreement with a previous paper (Fritz et
compared to mtDNA clades, the general picture al., 2007), our present investigation confirmed
does not change and the incongruity between deep mitochondrial divergences in Transcau-
mitochondrial haplotypes and nuclear-genomic casian tortoises of the Testudo graeca complex
genotypes remains (fig. 2). (table 3). In our study area occur three of the
342 V. Mashkaryan et al.
Gene flow in Testudo graeca 343

six mitochondrial clades of the T. graeca com- groups. Two of these groups perfectly corre-
plex. Compared to mtDNA variation, Mikulicek spond to mitochondrial clades in the western
et al. (2013) found less differentiation using Mediterranean and in central and eastern Iran
AFLP data (152 markers), with only four AFLP (mtDNA clade B and F, respectively). Due to
their completely allopatric or parapatric distri-
Table 3. Mean uncorrected p distances (percentages) be-
bution ranges, and hence the impossibility of
tween and within the six major mtDNA clades A-F of the
Testudo graeca complex based on the data set of Fritz et al. any recent gene flow, such a pattern is not too
(2007), merged with the sequences generated in the present surprising. Some other findings by Mikulicek
study (1144 bp cyt b). Two erroneous sequences of western et al. (2013) require more attention: (1) One of
Mediterranean tortoises of Fritz et al. (2007) were replaced
for the corrected sequences (see Vamberger et al., 2011). the two remaining AFLP groups, the so-called
Below the diagonal, divergences between clades are given Balkans-Middle Eastern AFLP group, embraces
and on the diagonal in boldface, divergences within the re- tortoises harbouring mtDNA clades C and D,
spective clade.
which are sister groups. (2) The fourth AFLP
A B C D E F group, the so-called Caucasian AFLP group, in-
A (n = 50) 0.35 cludes mtDNA clades A and E, which are not
B (n = 7) 4.48 1.02 sister groups (Fritz et al., 2007). This suggests
C (n = 27) 4.81 5.70 0.26
D (n = 37) 4.42 5.40 2.50 0.78
in both cases extensive gene flow within each
E (n = 12) 4.15 5.11 4.90 4.37 1.06 AFLP group. (3) The distribution ranges of the
F (n = 7) 3.84 3.90 4.01 3.48 3.02 0.14 Caucasian and Balkans-Middle Eastern AFLP
groups abut in Transcaucasia, and Mikulicek
et al. (2013) found two tortoises representing
Table 4. Genetic diversity of Testudo graeca, based on 10
microsatellite loci (individuals with mixed ancestries ex- each of these AFLP groups syntopically at one
cluded). Cluster names (colours) refer to fig. 2. Abbrevia-
tions: n, number of individuals; nA , number of alleles; nA , Table 5. Fixation indices (FST values) for microsatellite
average number of alleles; nP , number of private alleles; data under K = 3 and K = 4 (individuals with mixed an-
AR, allelic richness; HO , average observed heterozygosity; cestries excluded). All FST values are significantly different
HE , average expected heterozygosity. from zero.

Cluster n nA nA nP AR HO HE Cluster Red Blue Yellow Green

K=3 K=3
Red 27 43 1.59 2 4.300 0.207 0.411 Red
Blue 27 65 2.41 14 6.500 0.374 0.563 Blue 0.21
Yellow 33 120 3.64 62 11.290 0.412 0.707 Yellow 0.19 0.12

K=4 K=4
Red 25 42 1.68 2 3.838 0.252 0.438 Red
Blue 18 63 3.50 15 6.120 0.416 0.566 Blue 0.25
Yellow 16 48 3.00 4 4.800 0.306 0.459 Yellow 0.26 0.24
Green 19 106 5.58 47 9.918 0.447 0.748 Green 0.24 0.15 0.16

Figure 2. Cluster assignment of 107 spur-thighed tortoises from 26 collecting sites in Transcaucasia using STRUCTURE 2.3.3,
based on 10 polymorphic microsatellite loci. Shown are the STRUCTURE runs with the highest probability values for each K.
In the barplots, the genotype of an individual tortoise is represented by a vertical column. Below a threshold of 80% for cluster
membership, distinct colours within one column indicate mixed ancestry. The white lines in the barplots represent 20% and
the numbers below the barplots refer to collecting sites. Different collecting sites are separated by black vertical lines. Colours
of sampling sites in the maps correspond to STRUCTURE clusters; slices indicate tortoises with mixed ancestries or conflicting
cluster assignment. Top: Results for K = 3. Encircled collecting sites in the map denote morphological groups (for other
localities, no morphological data were available); the black bar approximately indicates the Lesser Caucasus. Bottom: Results
for K = 4. For encircled collecting sites, mtDNA data were available as indicated. At locality 4, four tortoises harboured
haplotypes of clade A and two of clade C. At locality 13, five tortoises yielded haplotypes of clade A and one of clade C.
Inset shows location of map section.
344
site (Katekh, Azerbaijan), without any admix-
ture. This could indicate that the two involved
AFLP groups represent reproductively isolated
units, and hence, that each of these units could
qualify as a full species under the Biological
Species Concept (Mayr, 1942, 1963). However,
both tortoises from Katekh harboured mtDNA
haplotypes representing the same clade (clade
C), providing evidence either for intraspecific
gene flow or interspecific introgression.
If two reproductively largely isolated species
were involved, and the shared mitochondrial
haplotypes of the tortoises from Katekh were
the result of interspecific hybridization or in-
trogression, it should be expected that our
microsatellite markers would reveal largely
distinct nuclear gene pools for the involved
species as is the case, for instance, even in
biological species which lost their original mi-
tochondrial genome by introgressive hybridiza-
tion (e.g. Lissotriton montandoni, Zielinski et
al., 2013). Such distinct gene pools should cor-
respond to highly distinct clusters in unsuper-
vised STRUCTURE analyses and should be char-
acterized by pronounced genetic differences, as
reflected by many private alleles and highly dis-
tinct fixation indices of microsatellite loci. On
the other hand, if the observed mismatches be-
tween nuclear genomic and mitochondrial dif-
ferentiation result rather from recent secondary
contact of two reproductively fully compatible
evolutionary lineages, the differentiation of dis-
tinct clusters should be rather weak and gene
flow should be extensive as indicated by many
individuals with mixed ancestries. Moreover,
private alleles should be rare, and fixation in-
dices should be low.
In agreement with the latter, the STRUCTURE
analyses of our microsatellite data provide ev-
idence for a generally high degree of admix-
ture (fig. 2), with relatively few individuals
assigned to more or less pure clusters under
K = 4. This high degree of admixture reflects
the difficulties to find data partitions in Hardy-
Weinberg equilibrium, one of the two main
criteria of STRUCTURE for cluster delineation
V. Mashkaryan et al.
(Pritchard, Stephens and Donnelly, 2000). How-
ever, Hardy-Weinberg equilibrium is not ex-
pected in recently admixed populations.
STRUCTURE revealed the most pronounced
differentiation for the red cluster, which corre-
sponds with some exceptions to flat-shelled
tortoises matching the original description of T.
g. armeniaca Chkhikvadze and Bakradze, 1991.
Such tortoises occur in a geographically quite
isolated situation in the valley of the Araxes
River, embedded between the eastern Anato-
lian mountains and the Lesser Caucasus. Pos-
sibly their flat-shelled morphotype is associated
with burrow-digging, while the domed tortoises
corresponding to the other microsatellite clus-
ters generally do not dig deep burrows. How-
ever, neither fixation and diversity indices of
microsatellites (tables 4 and 5) nor AFLP data
or mtDNA sequences (figs 1 and 2) support
a genetic isolation of the flat-shelled tortoises.
The flat-shelled tortoises harbour mitochondrial
haplotypes of clades A and C, which do also oc-
cur in tortoises with domed shells, and the Cau-
casian AFLP group embraces flat-shelled and
domed tortoises. Even if tortoises with mixed
ancestry are completely disregarded, FST val-
ues between the red cluster and the other clus-
ters are not extraordinarily high for microsatel-
lites, ranging from 0.19 to 0.21 for K = 3
and from 0.24 to 0.26 for K = 4 (table 5).
When topotypic individuals of mixed ancestry
are included, FST values are for obvious rea-
sons lower and amount only to 0.10-0.15. Also
within other tortoise species, similar or even
distinctly higher FST values have been reported
(Chelonoidis chilensis, based on 10 microsatel-
lite loci: 0.10-0.21, Fritz et al., 2012a; Gopherus
agassizii, based on 20 microsatellite loci: 0.01-
0.13, Hagerty and Tracy, 2010; G. polyphe-
mus, based on 9 microsatellite loci: 0.06-0.51,
Schwartz and Karl, 2005; Testudo marginata,
based on 11 microsatellite loci: 0.05-0.16, Perez
et al., 2012). Most notably, based on seven mi-
crosatellite loci FST values of up to 0.24 were
found within one and the same subspecies of T.
Gene flow in Testudo graeca
graeca, the western Mediterranean T. g. graeca
(Graci et al., 2013).
In addition, the red cluster is characterized
by the lowest number of private alleles of all
(table 4). This, together with the evidence for
a high degree of admixture and FST values
matching the differentiation within other tor-
toise species, supports the view that the T.
graeca complex represents only one Biologi-
cal Species sensu Mayr (1942, 1963), despite
deeply divergent mitochondrial lineages (ta-
ble 3). The three mtDNA clades A, C and E
occurring in our study region differ by 4.15%
to 4.90% in uncorrected p distances of the
cyt b gene. These values exceed uncorrected p
distances as observed among some other con-
generic tortoise species (3.7% to 12.7%; Fritz
et al., 2012a; Kindler et al., 2012), underlining
that species delineation in chelonians should not
rely on genetic distances alone (Vargas-Ramrez
et al., 2010; Praschag et al., 2011; Stuckas and
Fritz, 2011; Fritz et al., 2012a, b; Kindler et al.,
2012).
In phylogenetic analyses of mtDNA se-
quences, one of the Caucasian clades, clade
A, constitutes the sister group of the west-
ern Mediterranean clade B (Fritz et al., 2007,
2009; fig. 1), and another clade occurring in
the Caucasus region, clade C, is distributed
over a highly disjunct range including parts of
the Balkan peninsula, Turkey, the Russian and
Georgian Black Sea coast and the Caucasus.
The fossil record suggests that this patchy range
is a consequence of Pleistocene extinction (Fritz
et al., 2007). These observations provide evi-
dence that the mitochondrial clades within T.
graeca are very old. Using a fossil-calibrated
molecular clock, Fritz et al. (2009) estimated
that the six major clades of T. graeca have
evolved approximately 4.2-1.8 million years
ago. Despite these old and deep divergences,
microsatellite differentiation is rather weak and
differentiation patterns of AFLP and microsatel-
lite markers are not congruent. This indicates
repeated phases of extensive gene flow and vi-
cariance, which is not surprising when the Pleis-
345
tocene history is considered. We hypothesize
that the deeply divergent mtDNA lineages re-
flect old vicariance events, while the conflict-
ing nuclear markers are the legacy of younger
episodes of extensive gene flow. The differenti-
ation pattern found in the AFLP markers, with
AFLP groups matching allopatrically or para-
patrically distributed mitochondrial lineages, is
likely to be older than the microsatellite differ-
entiation, which could correspond to Holocene
range expansions into Caucasian valleys and
lower altitudes. It is expected that genealogies
of some nuclear genes will also deviate, id-
iosyncratically reflecting individual gene trees,
gene flow and recombination. This suggests that
coalescent-based species delineation attempts
(e.g. Yang and Rannala, 2010) will be fraught
with difficulties in sexually reproducing organ-
isms which experienced repeated switches be-
tween vicariance and extensive gene flow, such
as the T. graeca complex. Any resulting classi-
fication based on genetic coalescence along the
time axis necessarily has therefore to conflict
with population genetic differentiation and gene
flow in the here and now.
Owing to the largely mutually exclusive dis-
tribution ranges of the mtDNA lineages (fig. 3),
the approach of Fritz et al. (2007, 2009), who
have identified distinct mtDNA lineages with
subspecies, is a reasonable and straightforward
classification that facilitates communication and
acknowledges on the one hand the deep mi-
tochondrial divergences and on the other the
conspecifity of the involved evolutionary units.
Consequently, Transcaucasia constitutes an in-
tergradation zone of Testudo graeca armeni-
aca sensu Fritz et al. (2007), corresponding to
mtDNA clade A, T. g. ibera sensu Fritz et al.
(2007), corresponding to mtDNA clade C, and
T. g. buxtoni sensu Fritz et al. (2007), corre-
sponding to mtDNA clade E.
The observed differences among the three
marker systems (fig. 1) do not only provide
insights in biogeography and taxonomy. They
also underscore that the different modes of in-
346
Figure 3. Distribution ranges of eastern Testudo graeca subspecies according to Fritz et al. (2007, 2009) and the present study.
The distribution of T. graeca is shown in grey. Subspecies ranges are indicated, as far as known, by hatching; cross-hatching,
secondary intergradation zones; question marks, unknown subspecies allocation.
heritance of mtDNA, AFLP and microsatellite
markers may contribute to distinct differentia-
tion patterns. Mitochondrial DNA is generally
inherited only through the maternal line (Bal-
lard and Whitlock, 2004; Currat et al., 2008),
while AFLP and microsatellite markers are bi-
parentally inherited with dominant or codom-
inant modes of inheritance, respectively (Bee-
bee and Rowe, 2008). Mitochondrial DNA typ-
ically experiences less gene flow, but intro-
gresses more easily than nuclear DNA (Currat
et al., 2008), and sex-specific differences in dis-
persal may further contribute to mismatches be-
tween the marker systems. This is in line with
the observation that males of T. graeca have dis-
tinctly larger home ranges than females (Daz-
Paniagua, Keller and Andreu, 1995), suggesting
that gene flow is mainly male-mediated and that
the persistence of deeply divergent mtDNA lin-
eages is fostered by smaller home ranges of fe-
males.
V. Mashkaryan et al.
Acknowledgements. Christian Kehlmaier, Anke Mller
and Anja Rauh helped in the laboratory and Christian
Kehlmaier genotyped and sequenced some of the sam-
ples. Field work in Armenia and Georgia and part of the
laboratory work was funded by the Gulbenkian Founda-
tion, Portugal (program Preserving Armenian biodiver-
sity: Joint Portuguese-Armenian program for training in
modern conservation biology) and by the Fundao para
a Cincia e a Tecnologia, FCT, Portugal (PTDC/BIA-
BEC/101256/2008). Melita Vamberger was funded by a
PhD fellowship of the German Academic Exchange Service
(DAAD; A/09/91179).
References
Anderson, S.C. (1979): Synopsis of the turtles, crocodiles,
and amphisbaenians of Iran. Proc. California Acad. Sci.,
4th Ser. 41: 501-528.
Ballard, J.W.O., Whitlock, M.C. (2004): The incomplete
natural history of mitochondria. Mol. Ecol. 13: 729-744.
Beebee, T.J.C., Rowe, G. (2008): An Introduction to Molec-
ular Ecology, Second Edition. Oxford University Press,
New York.
Carretero, M.A., Znari, M., Harris, D.J., Mac, J.C. (2005):
Morphological divergence among populations of Tes-
tudo graeca from west-central Morocco. Animal Biol.
55: 259-279.
Gene flow in Testudo graeca
Chkhikvadze, V.M., Bakradze, M.A. (1991): O sistematich-
eskom polozhenii sovremennoi sukhoputnoi cherepakhi
iz doliny reki Araks. Trudy Tbilissk. gosudarstven. Univ.
305: 59-63.
Chkhikvadze, V.M., Bakradze, M.A. (2002): Novyi podvid
sukhoputnoi cherepakhi iz Dagestana. Trudy Inst. Zool.,
Akad. Nauk Gruzii 21: 276-279.
Chkhikvadze, V.M., Tuniyev, B.S. (1986): O sistem-
aticheskom polozhenii sovremennoi sukhoputnoi
cherepakhi zapadnogo Zakavkazya. Soobshch. Akad.
Nauk Gruzinsk. SSR 124: 617-620.
Chkhikvadze, V.M., Mazanaeva, L.F., Shammakov, S.M.
(2011): Kratkie svedeniya o novom vide sukhoputnoi
cherepakh iz Dagestana. In: Proceedings of the Interna-
tional Conference Biological Diversity and Conserva-
tion Problems of the Fauna of the Caucasus, p. 336-340.
Yerevan, National Academy of Sciences, Republic of
Armenia, Division of Natural Sciences, Scientific Center
of Zoology and Hydroecology and American University
of Armenia Foundation (Acopian Center for the Envi-
ronment).
Currat, M., Ruedi, M., Petit, R.J., Excoffier, L. (2008): The
hidden side of invasions: massive introgression by local
genes. Evolution 62: 1908-1920.
Daz-Paniagua, C., Keller, C., Andreu, A.C. (1995): Annual
variation of activity and daily distances moved in adult
spur-thighed tortoises, Testudo graeca, in southwestern
Spain. Herpetologica 51: 225-233.
Edwards, T., Goldberg, C.S., Kaplan, M.E., Schwalbe, C.R.,
Swann, D.E. (2003): PCR primers for microsatellite loci
in the desert tortoise (Gopherus agassizii, Testudinidae).
Mol. Ecol. Notes 3: 589-591.
Evanno, G., Regnaut, S., Goudet, J. (2005): Detecting the
number of clusters of individuals using the software
STRUCTURE: a simulation study. Mol. Ecol. 14: 2611-
2620.
Excoffier, L., Laval, G., Schneider, S. (2005): ARLEQUIN
(version 3.0): an integrated software package for popu-
lation genetics data analysis. Evol. Bioinform. Online 1:
47-50.
Falush, D., Stephens, M., Pritchard, J.K. (2007): Inference
of population structure using multilocus genotype data:
dominant markers and null alleles. Mol. Ecol. Notes 7:
574-578.
Forlani, A., Crestanello, B., Mantovani, S., Livoreil, B.,
Zane, L., Bertorelle, G., Congiu, L. (2005): Identifica-
tion and characterization of microsatellite markers in
Hermanns tortoise (Testudo hermanni, Testudinidae).
Mol. Ecol. Notes 5: 228-230.
Fritz, U., Auer, M., Bertolero, A., Cheylan, M., Fattizzo, T.,
Hundsdrfer, A.K., Martn Sampayo, M., Pretus, J.L.,
irok, P., Wink, M. (2006): A rangewide phylogeog-
raphy of Hermanns tortoise, Testudo hermanni (Rep-
tilia: Testudines: Testudinidae): implications for taxon-
omy. Zool. Scr. 35: 531-543.
Fritz, U., Hundsdrfer, A.K., irok, P., Auer, M., Kami,
H., Lehmann, J., Mazanaeva, L.F., Trkozan, O., Wink,
M. (2007): Phenotypic plasticity leads to incongruence
between morphology-based taxonomy and genetic dif-
ferentiation in western Palaearctic tortoises (Testudo
graeca complex; Testudines, Testudinidae). Amphibia-
Reptilia 28: 97-121.
347
Fritz, U., Harris, D.J., Fahd, S., Rouag, R., Graci Martnez,
E., Gimnez Casalduero, A., irok, P., Kalboussi,
M., Jdeidi, T.B., Hundsdrfer, A.K. (2009): Mitochon-
drial phylogeography of Testudo graeca in the Western
Mediterranean: Old complex divergence in North Africa
and recent arrival in Europe. Amphibia-Reptilia 30: 63-
80.
Fritz, U., Alcalde, L., Vargas-Ramrez, M., Goode, E.V.,
Fabius-Turoblin, D.U., Praschag, P. (2012a): Northern
genetic richness and southern purity, but just one species
in the Chelonoidis chilensis complex. Zool. Scr. 41: 220-
232.
Fritz, U., Stuckas, H., Vargas-Ramrez, M., Hundsdrfer,
A.K., Maran, J., Pckert, M. (2012b): Molecular phy-
logeny of Central and South American slider turtles:
implications for biogeography and systematics (Tes-
tudines: Emydidae: Trachemys). J. Zool. Syst. Evol. Res.
50: 125-136.
Glaubitz, J.C. (2004): CONVERT: a user friendly program
to reformat diploid genotypic data for commonly used
population genetic software packages. Mol. Ecol. Notes
4: 309-310.
Goudet, J. (1995): FSTAT (version 1.2): a computer program
to calculate F -statistics. J. Hered. 86: 485-486.
Graci, E., Gimnez, A., Anadn, J.D., Harris, D.J., Fritz,
U., Botella, F. (2013): The uncertainty of Late Pleis-
tocene range expansions in the western Mediterranean:
a case study of the colonization of south-eastern Spain
by the spur-thighed tortoise, Testudo graeca. J. Bio-
geogr. 40: 323-334.
Guyot-Jackson, G., Ed. (2004): Numro spcial Testudo.
Manouria 7(22): 1-52.
Hagerty, B.E., Tracy, C.R. (2010): Defining population
structure for the Mojave desert tortoise. Conserv. Genet.
11: 1795-1807.
Hall, T.A. (1999): BIOEDIT: a user friendly biological se-
quence alignment editor and analysis program for Win-
dows 95/98/NT. Nucl. Acids Symp. Ser. 41: 95-98.
Harris, D.J., Znari, M., Mac, 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.
Highfield, A.C. (1990): Tortoises of north Africa; taxonomy,
nomenclature, phylogeny and evolution with notes on
field studies in Tunisia. J. Chelon. Herpetol. 1(2): 1-56.
Highfield, A.C., Martin, J. (1989a): A revision of the Tes-
tudines of North Africa, Asia and Europe. Genus: Tes-
tudo. J. Chelon. Herpetol. 1(1): 1-12.
Highfield, A.C., Martin, J. (1989b): Testudo whitei Bennett
1836. New light on an old carapace Gilbert Whites
Selborne tortoise re-discovered. J. Chelon. Herpetol.
1(1): 13-22.
Highfield, A.C., Martin, J. (1989c): Description of a minia-
ture tortoise Testudo flavominimaralis n. species from
North Africa. London, The Tortoise Trust/Tortoise Sur-
vival Project.
Hubisz, M.J., Falush, D., Stephens, M., Pritchard, J.K.
(2009): Inferring weak population structure with the
assistance of sample group information. Mol. Ecol. Res.
9: 1322-1332.
348
Kindler, C., Branch, W.R., Hofmeyr, M.D., Maran, J.,
irok, P., Vences, M., Harvey, J., Hauswaldt, J.S.,
Schleicher, A., Stuckas, H., Fritz, U. (2012): Molecu-
lar phylogeny of African hinge-back tortoises (Kinixys):
implications for phylogeography and taxonomy (Tes-
tudines: Testudinidae). J. Zool. Syst. Evol. Res. 50: 192-
201.
King, T.L., Julian, S.E. (2004): Conservation of microsatel-
lite DNA flanking sequence across 13 emydid genera as-
sayed with novel bog turtle (Glyptemys muhlenbergii)
loci. Conserv. Genet. 5: 719-725.
Mallet, J. (2005): Hybridization as an invasion of the
genome. Trends Ecol. Evol. 20: 229-237.
Mayr, E. (1942): Systematics and the Origin of Species
from the Viewpoint of a Zoologist. Columbia University
Press, New York.
Mayr, E. (1963): Animal Species and Evolution. Belknap
Press, Cambridge, Massachusetts.
Mikulcek, P., Jandzik, D., Fritz, U., Schneider, C., irok,
P. (2013): AFLP analysis shows high incongruence
between genetic differentiation and morphology-based
taxonomy in a widely distributed tortoise. Biol. J. Linn.
Soc. 108: 151-160.
Parham, J.F., Trkozan, O., Stuart, B.L., Arakelyan, M.,
Shafei, S., Macey, J.R., Papenfuss, T.J. (2006): Genetic
evidence for premature taxonomic inflation in Middle
Eastern tortoises. Proc. California Acad. Sci., 4th Ser.
57: 955-964.
Perl, J. (1996): Etel-Turkin maakilpikonnista. Morfolo-
gisia ja ekologisia eroja (Testudo ibera Pallas 1814 &
Testudo ibera anamurensis Weissinger 1987) sek uuden
maakilpikonnalajin kuvaus + 15 kuvaa. In: Herpetokon-
gressi I, Virallinen kongressijulkaisu, p. 14-26. Perl,
J., Vikberg, J., Kanza, M., Eds, Suomen herpetologinen
yhdistys ry, Helsinki.
Perl, J. (2002): Morphological variation among Middle
Eastern Testudo graeca L., 1758 (sensu lato), with a
focus on taxonomy. Chelonii 3: 78-108.
Perez, M., Leblois, R., Livoreil, B., Bour, R., Lambourdire,
J., Samadi, S., Boisselier, M.-C. (2012): Effects of land-
scape features and demographic history on the genetic
structure of Testudo marginata populations in the south-
ern Peloponnese and Sardinia. Biol. J. Linn. Soc. 105:
591-606.
Pieh, A. (2001): Testudo graeca soussensis, eine neue Un-
terart der Maurischen Landschildkrte aus dem Sousstal
(Nordwest-Marokko). Salamandra 36(2000): 209-222.
Pieh, A., Perl, J. (2002): Variabilitt von Testudo graeca
Linnaeus, 1758 im stlichen Nordafrika mit Beschrei-
bung eines neuen Taxons von der Cyrenaika (Nord-
ostlibyen). Herpetozoa 15: 3-28.
Pieh, A., Perl, J. (2004): Variabilitt der Maurischen
Landschildkrten (Testudo graeca Linnaeus, 1758
Komplex) im zentralen und nordwestlichen Marokko
mit Beschreibung zweier neuer Taxa. Herpetozoa 17:
19-47.
Pieh, A., Fritz, U., Berglas, R. (2002): New data on
morphology, distribution and nomenclature of Testudo
graeca armeniaca Chkhikvadze & Bakradze, 1991.
Faun. Abh. Mus. Tierkd. Dresden 22: 329-345.
V. Mashkaryan et al.
Posada, D. (2008): JMODELTEST: phylogenetic model aver-
aging. Mol. Biol. Evol. 25: 1253-1256.
Praschag, P., Stuckas, H., Pckert, M., Maran, J., Fritz, U.
(2011): Mitochondrial DNA sequences suggest a revised
taxonomy of Asian flapshell turtles (Lissemys Smith,
1931) and the validity of previously unrecognized taxa
(Testudines: Trionychidae). Vertebr. Zool. 61: 147-160.
Pritchard, J.K., Stephens, M., Donnelly, P. (2000): Inference
of population structure using multilocus genotype data.
Genetics 155: 945-959.
Pritchard, P.C.H. (1979): Encyclopedia of Turtles. T.F.H.
Publications, Hong Kong.
Randi, E. (2008): Detecting hybridization between wild
species and their domesticated relatives. Mol. Ecol. 17:
285-293.
Ronquist, F., Teslenko, M., van der Mark, P., Ayres, D.L.,
Darling, A., Hhna, S., Larget, B., Liu, L., Suchard,
M.A., Huelsenbeck, J.P. (2012): MRBAYES 3.2: effi-
cient Bayesian phylogenetic inference and model choice
across a large model space. Syst. Biol. 61: 539-542.
Rosenberg, N.A. (2004): DISTRUCT: a program for the
graphical display of population structure. Mol. Ecol.
Notes 4: 137-138.
Salinas, M., Altet, L., Clavel, C., Almela, R.M., Bayn, A.,
Burguete, I., Snchez, A. (2011): Genetic assessment,
illegal trafficking and management of the Mediterranean
spur-thighed tortoise in Southern Spain and Northern
Africa. Conserv. Genet. 12: 1-13.
Schwartz, T.S., Karl, S.A. (2005): Population and conserva-
tion genetics of the gopher tortoise (Gopherus polyphe-
mus). Conserv. Genet. 6: 917-928.
Schwartz, T.S., Osentoski, M., Lamb, T., Karl, S.A. (2003):
Microsatellite loci for the North American tortoises
(genus Gopherus) and their applicability to other turtle
species. Mol. Ecol. Notes 3: 283-286.
Spinks, P.Q., Shaffer, H.B., Iverson, J.B., McCord, W.P.
(2004): Phylogenetic hypotheses for the turtle family
Geoemydidae. Mol. Phylogenet. Evol. 32: 164-182.
Stamatakis, A. (2006): RAxML-VI-HPC: maximum like-
lihood-based phylogenetic analyses with thousands of
taxa and mixed models. Bioinformatics 22: 2688-2690.
Stuckas, H., Fritz, U. (2011): Identity of Pelodiscus sinensis
revealed by DNA sequences of an approximately 180-
year-old type specimen and a taxonomic reappraisal of
Pelodiscus species (Testudines: Trionychidae). J. Zool.
Syst. Evol. Res. 49: 335-339.
Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M.,
Kumar, S. (2011): MEGA 5: Molecular Evolutionary Ge-
netics Analysis using maximum likelihood, evolutionary
distance, and maximum parsimony methods. Mol. Biol.
Evol. 28: 2731-2739.
Vamberger, M., Corti, C., Stuckas, H., Fritz, U. (2011): Is
the imperilled spur-thighed tortoise (Testudo graeca) na-
tive in Sardinia? Implications from population genetics
and for conservation. Amphibia-Reptilia 32: 9-25.
van der Kuyl, A.C., Ballasina, D.L.P., Dekker, J.T., Maas,
H., Willemsen, R.E., Goudsmit, J. (2002): Phylogenetic
relationships among the species of the genus Testudo
(Testudines: Testudinidae) inferred from mitochondrial
12S rRNA gene sequences. Mol. Phylogenet. Evol. 22:
174-183.
Gene flow in Testudo graeca 349

van der Kuyl, A.C., Ballasina, D.L.P., Zorgdrager, F. (2005):
Mitochondrial haplotype diversity in the tortoise species
Testudo graeca from North Africa and the Middle East.
BMC Evol. Biol. 5: 29.
van Oosterhout, C., Hutchinson, W.F., Wills, D.P.M., Ship-
ley, P. (2004): MICRO - CHECKER: software for identify-
ing and correcting genotyping errors in microsatellite
data. Mol. Ecol. Notes 4: 535-538.
Vargas-Ramrez, M., Vences, M., Branch, W.R., Daniels,
S.R., Glaw, F., Hofmeyr, M.D., Kuchling, G., Maran,
J., Papenfuss, T.J., irok, P., Vieites, D.R., Fritz, U.
(2010): Deep genealogical lineages in the widely dis-
tributed African helmeted terrapin: evidence from mi-
tochondrial and nuclear DNA (Testudines: Pelomedusi-
dae: Pelomedusa subrufa). Mol. Phylogenet. Evol. 56:
428-440.
Weissinger, H. (1987): Testudo graeca anamurensis ssp.
nov. aus Kleinasien. GH-Nachrichten 10/11: 14-18.
Wermuth, H., Mertens, R. (1961): Schildkrten, Krokodile,
Brckenechsen. VEB Gustav Fischer Verlag, Jena.
Wermuth, H., Mertens, R. (1977): Testudines, Crocodylia,
Rhynchocephalia. Das Tierreich 100: I-XXVII, 1-174.
Yang, Z., Rannala, B. (2010): Bayesian species delimitation
using multilocus sequence data. Proc. Nat. Acad. Sci.
USA 107: 9264-9269.
Zielinski, P., Nadachowska-Brzyska, K., Wielstra, B., Szko-
tak, R., Covaciu-Marcov, D., Cogalniceanu, D., Babik,
W. (2013): No evidence for nuclear introgression despite
complete mtDNA replacement in the Carpathian newt
(Lissotriton montandoni). Mol. Ecol. 22: 1884-1903.
Submitted: March 15, 2013. Final revision received: April
30, 2013. Accepted: May 24, 2013.
Associated Editor: Sylvain Ursenbacher.

Appendix. Studied samples of the Testudo graeca complex.

Population Lab code Collection site N E K=3 K=4 mtDNA clade

1 AC558 Georgia: Bolnisi 41.44 44.53 yellow admixed n/a

2 AC073 Armenia: Tavush: Kokhb 40.90 45.39 yellow green C
2 AC519 Armenia: Tavush: Kokhb 40.90 45.39 admixed admixed n/a
2 AC520 Armenia: Tavush: Kokhb 40.90 45.39 blue blue n/a
3 R-3 Armenia: Noyemberyan 41.18 45.01 n/a n/a C
3 T1975 Armenia: Noyemberyan 41.18 45.01 n/a n/a C
3 AC038 Armenia: Noyemberyan 41.18 45.01 yellow green C
4 AC041 Armenia: Armavir 40.15 43.84 n/a n/a A
4 AC042 Armenia: Armavir 40.15 43.84 n/a n/a A
4 AC044 Armenia: Armavir 40.15 43.84 red red n/a
4 AC045 Armenia: Armavir 40.15 43.84 yellow admixed n/a
4 AC213 Armenia: Armavir 40.15 43.84 red red n/a
4 AC064 Armenia: Armavir 40.15 43.84 yellow admixed n/a
4 AC200 Armenia: Armavir 40.15 43.84 red red A
4 AC201 Armenia: Armavir 40.15 43.84 yellow green n/a
4 AC202 Armenia: Armavir 40.15 43.84 admixed blue A
4 AC203 Armenia: Armavir 40.15 43.84 yellow green C
4 AC204 Armenia: Armavir 40.15 43.84 yellow green C
5 AC074 Armenia: Garni 40.12 44.74 red red A
5 AC076 Armenia: Garni 40.12 44.74 red red n/a
5 AC075 Armenia: Garni 40.12 44.74 n/a n/a A
6 AC001 Armenia: Armavir: Vanand 40.11 43.82 red red A
6 AC002 Armenia: Armavir: Vanand 40.11 43.82 red red n/a
6 AC003 Armenia: Armavir: Vanand 40.11 43.82 red red n/a
6 AC005 Armenia: Armavir: Vanand 40.11 43.82 red red n/a
6 AC006 Armenia: Armavir: Vanand 40.11 43.82 n/a n/a A
6 AC007 Armenia: Armavir: Vanand 40.11 43.82 red red n/a
6 AC008 Armenia: Armavir: Vanand 40.11 43.82 red red n/a
6 AC160 Armenia: Armavir: Vanand 40.11 43.82 red red n/a
7 AC039 Armenia: Gorovan 39.89 44.71 red red A
8 AC061 Armenia: Ararat: Urtsadzor 39.91 44.81 n/a n/a A
8 AC063 Armenia: Ararat: Urtsadzor 39.91 44.81 n/a n/a A
8 Ac065 Armenia: Ararat: Urtsadzor 39.91 44.81 n/a n/a A
8 AC066 Armenia: Ararat: Urtsadzor 39.91 44.81 red red n/a
8 AC067 Armenia: Ararat: Urtsadzor 39.91 44.81 red red n/a
8 AC068 Armenia: Ararat: Urtsadzor 39.91 44.81 red red n/a
8 AC069 Armenia: Ararat: Urtsadzor 39.91 44.81 red red n/a
350 V. Mashkaryan et al.

Appendix. (Continued.)

Population Lab code Collection site N E K=3 K=4 mtDNA clade

8 AC071 Armenia: Ararat: Urtsadzor 39.91 44.81 red red n/a

8 AC072 Armenia: Ararat: Urtsadzor 39.91 44.81 red red n/a
8 AC142 Armenia: Ararat: Urtsadzor 39.91 44.81 red admixed n/a
8 AC143 Armenia: Ararat: Urtsadzor 39.91 44.81 red admixed n/a
8 AC146 Armenia: Ararat: Urtsadzor 39.91 44.81 red red n/a
8 AC147 Armenia: Ararat: Urtsadzor 39.91 44.81 red red n/a
8 AC149 Armenia: Ararat: Urtsadzor 39.91 44.81 red red n/a
8 AC151 Armenia: Ararat: Urtsadzor 39.91 44.81 admixed admixed n/a
8 AC153 Armenia: Ararat: Urtsadzor 39.91 44.81 blue blue n/a
8 AC156 Armenia: Ararat: Urtsadzor 39.91 44.81 red red n/a
8 AC157 Armenia: Ararat: Urtsadzor 39.91 44.81 red red n/a
8 AC158 Armenia: Ararat: Urtsadzor 39.91 44.81 admixed admixed n/a
8 AC205 Armenia: Ararat: Urtsadzor 39.91 44.81 red red A
8 AC206 Armenia: Ararat: Urtsadzor 39.91 44.81 n/a n/a A
9 T-2 Nagorno Karabakh: Kashatagh: Tcobi 39.02 46.65 n/a n/a A
9 AC009 Nagorno Karabakh: Kashatagh: Tcobi 39.02 46.65 admixed admixed n/a
9 AC010 Nagorno Karabakh: Kashatagh: Tcobi 39.02 46.65 blue admixed A
9 AC011 Nagorno Karabakh: Kashatagh: Tcobi 39.02 46.65 blue admixed n/a
9 AC012 Nagorno Karabakh: Kashatagh: Tcobi 39.02 46.65 blue blue n/a
9 AC013 Nagorno Karabakh: Kashatagh: Tcobi 39.02 46.65 blue yellow n/a
9 AC014 Nagorno Karabakh: Kashatagh: Tcobi 39.02 46.65 admixed admixed n/a
9 AC015 Nagorno Karabakh: Kashatagh: Tcobi 39.02 46.65 blue admixed n/a
9 AC017 Nagorno Karabakh: Kashatagh: Tcobi 39.02 46.65 blue admixed A
9 AC018 Nagorno Karabakh: Kashatagh: Tcobi 39.02 46.65 blue yellow n/a
9 AC019 Nagorno Karabakh: Kashatagh: Tcobi 39.02 46.65 blue blue A
9 AC034 Nagorno Karabakh: Kashatagh: Tcobi 39.02 46.65 blue admixed n/a
9 AC207 Nagorno Karabakh: Kashatagh: Tcobi 39.02 46.65 blue blue A
9 AC208 Nagorno Karabakh: Kashatagh: Tcobi 39.02 46.65 blue blue A
9 AC209 Nagorno Karabakh: Kashatagh: Tcobi 39.02 46.65 blue blue A
9 AC210 Nagorno Karabakh: Kashatagh: Tcobi 39.02 46.65 blue blue n/a
9 DB7821 Nagorno Karabakh: Kashatagh: Tcobi 39.02 46.65 blue blue A
9 AC211 Nagorno Karabakh: Kashatagh: Tcobi 39.02 46.65 admixed admixed A
9 AC212 Nagorno Karabakh: Kashatagh: Tcobi 39.02 46.65 blue blue n/a
10 AC022 Nagorno Karabakh: Kashatagh: Lusavan 39.19 47.02 admixed yellow A
10 AC023 Nagorno Karabakh: Kashatagh: Lusavan 39.19 47.02 blue blue A
10 AC024 Nagorno Karabakh: Kashatagh: Lusavan 39.19 47.02 admixed admixed n/a
10 AC025 Nagorno Karabakh: Kashatagh: Lusavan 39.19 47.02 blue yellow n/a
10 AC026 Nagorno Karabakh: Kashatagh: Lusavan 39.19 47.02 blue blue n/a
10 AC029 Nagorno Karabakh: Kashatagh: Lusavan 39.19 47.02 blue admixed n/a
10 AC030 Nagorno Karabakh: Kashatagh: Lusavan 39.19 47.02 blue blue A
10 AC032 Nagorno Karabakh: Kashatagh: Lusavan 39.19 47.02 n/a n/a A
10 AC033 Nagorno Karabakh: Kashatagh: Lusavan 39.19 47.02 admixed yellow n/a
10 AC035 Nagorno Karabakh: Kashatagh: Lusavan 39.19 47.02 admixed yellow n/a
10 AC036 Nagorno Karabakh: Kashatagh: Lusavan 39.19 47.02 admixed admixed A
11 AC021 Nagorno Karabakh: Tigranavan 39.38 46.67 blue yellow A
12 DB4773 Nagorno Karabakh: Hadrut: Aknyakbur 39.51 47.02 n/a n/a A
12 DB7811 Nagorno Karabakh: Hadrut: Aknyakbur 39.51 47.02 yellow green A
13 AC037 Armenia: Nrnadzor 38.92 46.43 admixed admixed A
13 AC154 Armenia: Nrnadzor 38.92 46.43 blue blue n/a
13 AC155 Armenia: Nrnadzor 38.92 46.43 admixed admixed n/a
13 DB12516 Armenia: Nrnadzor 38.92 46.43 blue blue A
13 DB12394 Armenia: Nrnadzor 38.92 46.43 yellow green A
13 5614 Armenia: Nrnadzor 38.92 46.43 yellow green C
13 DB12069 Armenia: Nrnadzor 38.92 46.43 yellow green n/a
13 DB12453 Armenia: Nrnadzor 38.92 46.43 yellow green A
13 DB12518 Armenia: Nrnadzor 38.92 46.43 blue blue A
Gene flow in Testudo graeca 351

Appendix. (Continued.)

Population Lab code Collection site N E K=3 K=4 mtDNA clade

14 AC106 Iran: West Azerbaijan Province 37.27 44.59 blue blue n/a
14 AC107 Iran: West Azerbaijan Province 37.27 44.59 admixed yellow n/a
15 AC108 Iran: East Azerbaijan Province 37.54 46.16 n/a n/a E
15 AC109 Iran: East Azerbaijan Province 37.54 46.16 yellow yellow n/a
15 AC110 Iran: East Azerbaijan Province 37.54 46.16 admixed yellow n/a
15 AC111 Iran: East Azerbaijan Province 37.54 46.16 yellow admixed n/a
16 AC098 Iran: Maragheh 37.37 46.42 yellow admixed n/a
17 AC089 Iran: Varzeqan 38.50 46.65 yellow green n/a
17 AC090 Iran: Varzeqan 38.50 46.65 admixed yellow n/a
17 AC091 Iran: Varzeqan 38.50 46.65 yellow admixed n/a
18 AC097 Iran: Ahar 38.46 47.06 admixed yellow n/a
19 AC092 Iran: Miyaneh 37.43 47.70 yellow green n/a
19 AC093 Iran: Miyaneh 37.43 47.70 yellow green n/a
20 AC082 Iran: Angoran 36.44 47.28 yellow admixed n/a
21 AC094 Iran: Sohrein 36.51 48.24 yellow admixed n/a
21 AC095 Iran: Sohrein 36.51 48.24 yellow green n/a
22 AC085 Iran: Molalar 36.42 48.13 yellow green n/a
22 AC086 Iran: Molalar 36.42 48.13 yellow green n/a
22 AC087 Iran: Molalar 36.42 48.13 yellow green n/a
22 AC088 Iran: Molalar 36.42 48.13 yellow green n/a
23 AC096 Iran: between Buqda Kandi and Ahmad Kandi 36.34 48.16 yellow admixed n/a
24 AC077 Iran: Zanjan 36.70 48.29 yellow yellow n/a
24 AC078 Iran: Zanjan 36.70 48.29 admixed yellow E
24 AC079 Iran: Zanjan 36.70 48.29 yellow admixed E
24 AC081 Iran: Zanjan 36.70 48.29 admixed yellow n/a
25 AC083 Iran: Tarom 36.59 48.68 yellow green n/a
25 AC084 Iran: Tarom 36.60 49.68 yellow yellow n/a
26 AC099 Iran: Khoda Bande 35.40 51.25 yellow admixed n/a
T 2265* Iran: S Resht: between Saravan and Rostamabad 36.57 49.33 n/a n/a E
T 2267* Iran: S Resht: between Saravan and Rostamabad 36.57 49.33 n/a n/a E
T 2268* Iran: S Resht: between Saravan and Rostamabad 36.57 49.33 n/a n/a E
T 1999* Iran: Sefid Rud 37.22 48.08 n/a n/a E
T 1427* Iran: 6 km NE Meshgn Shahr 38.26 47.42 n/a n/a E
T 1428* Iran: 6 km NE Meshgn Shahr 38.26 47.42 n/a n/a E

Data from Fritz et al. (2007).