This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Amphibia-Reptilia 28 (2007): 97-121
Phenotypic plasticity leads to incongruence betweenmorphology-based taxonomy and genetic differentiation in
western Palaearctic tortoises (Testudo graeca complex;Testudines, Testudinidae)
Uwe Fritz1, Anna K. Hundsdörfer1, Pavel Široký2, Markus Auer1, Hajigholi Kami3, Jan Lehmann4,
Lyudmila F. Mazanaeva5, Oguz Türkozan6, Michael Wink7
Abstract. Tortoises of the Testudo graeca complex inhabit a patchy range that covers part of three continents (Africa, Europe,Asia). It extends approximately 6500 km in an east-west direction from eastern Iran to the Moroccan Atlantic coast andabout 1600 km in a north-south direction from the Danube Delta to the Libyan Cyrenaica Peninsula. Recent years haveseen a rapid increase of recognized taxa. Based on morphological investigations, it was suggested that this group consistsof as many as 20 distinct species and is paraphyletic with respect to T. kleinmanni sensu lato and T. marginata. Based onsamples from representative localities of the entire range, we sequenced the mitochondrial cytochrome b gene and conductednuclear genomic fingerprinting with ISSR PCR. The T. graeca complex is monophyletic and sister to a taxon consistingof T. kleinmanni sensu lato and T. marginata. The T. graeca complex comprises six well-supported mtDNA clades (A-F).Highest diversity is found in the Caucasian Region, where four clades occur in close neighbourhood. This suggests, inagreement with the fossil record, the Caucasian Region as a radiation centre. Clade A corresponds to haplotypes from theEast Caucasus. It is the sister group of another clade (B) from North Africa and western Mediterranean islands. Clade Cincludes haplotypes from western Asia Minor, the southeastern Balkans and the western and central Caucasus Region. Itssister group is a fourth, widely distributed clade (D) from southern and eastern Asia Minor and the Levantine Region (NearEast). Two further clades are distributed in Iran (E, northwestern and central Iran; F, eastern Iran). Distinctness of these sixclades and sister group relationships of (A + B) and (C + D) are well-supported; however, the phylogeny of the resultingfour clades (A + B), (C + D), E and F is poorly resolved. While in a previous study (Fritz et al., 2005a) all traditionallyrecognized Testudo species were highly distinct using mtDNA sequences and ISSR fingerprints, we detected within the T.graeca complex no nuclear genomic differentiation paralleling mtDNA clades. We conclude that all studied populations ofthe T. graeca complex are conspecific under the Biological Species Concept. There is major incongruence between mtDNAclades and morphologically defined taxa. Morphologically well-defined taxa, like T. g. armeniaca or T. g. floweri, nest withinclades comprising also geographically neighbouring, but morphologically distinctive populations of other taxa (clade A:T. g. armeniaca, T. g. ibera, T. g. pallasi; clade D: T. g. anamurensis, T. g. antakyensis, T. g. floweri, T. g. ibera, T. g.terrestris), while sequences of morphologically similar tortoises of the same subspecies (T. g. ibera sensu stricto or T. g. iberasensu lato) scatter over two or three genetically distinct clades (A, C or A, C, D, respectively). This implies that pronouncedmorphological plasticity, resulting in phenotypes shaped by environmental pressure, masks genetic differentiation. To achievea more realistic taxonomic arrangement reflecting mtDNA clades, we propose reducing the number of T. graeca subspeciesconsiderably and regard in the eastern part of the range five subspecies as valid (T. g. armeniaca, T. g. buxtoni, T. g. ibera, T.g. terrestris, T. g. zarudnyi). As not all North African taxa were included in the present study, we refrain from synonymizingNorth African taxa with T. g. graeca (mtDNA clade B) that represents a further valid subspecies.
1 - Museum of Zoology (Museum für Tierkunde), Nat-ural History State Collections Dresden, KönigsbrückerLandstr. 159, D-01109 Dresden, Germanye-mail: [email protected]
2 - 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
3 - Department of Biology, Faculty of Sciences, GorganUniversity of Agricultural Sciences and Natural Re-sources, Golestan Province, Iran
4 - Haspekrogen 10, DK-2880 Bagsværd, Denmark
Introduction
Tortoises of the genus Testudo Linnaeus, 1758are distributed over most of the southwestern
5 - Department of Zoology, Dagestan State University, 37aM. Gadyeva st., 367025 Makhachkala, Russia
6 - Department of Biology, Faculty of Science and Arts,Adnan Menderes University, 09010 Aydın, Turkey
7 - Institute of Pharmacy and Molecular Biotechnology(IPMB), Heidelberg University, INF 364, D-69120 Hei-delberg, Germany
Palaearctic Region (Ernst et al., 2000; Fritz andCheylan, 2001). The greatest portion of thisarea is inhabited by representatives of the Tes-tudo graeca complex. These chelonians are alsoknown as spur-thighed tortoises, a name refer-ring to two prominent horny thigh tuberclespresent in most individuals. The patchy rangeof spur-thighed tortoises covers part of threecontinents (Africa, Europe, Asia) and extendsapproximately 6500 km in an east-west direc-tion from eastern Iran to the Moroccan Atlanticcoast and about 1600 km in a north-south direc-tion from the Danube Delta to the Libyan Cyre-naica Peninsula (fig. 1). Spur-thighed tortoisesoccur under very different climatic and envi-ronmental conditions, ranging from a Mediter-ranean climate with mild, frost-free winters toan extreme continental steppe climate with se-vere winter frost.
While T. graeca Linnaeus, 1758 was longaccepted as single polytypic species with fourto seven subspecies (Mertens, 1946; Wermuth,1958; Wermuth and Mertens, 1961, 1977;Anderson, 1979; Ernst and Barbour, 1989;Gasperetti et al., 1993; Fritz et al., 1996; Ernst etal., 2000; Buskirk et al., 2001; Fritz and Chey-lan, 2001), its taxonomy has fluctuated greatlyin recent years. Based on morphology, manyspecies and subspecies were described or resur-rected, in part in grey literature (Chkhikvadzeand Tuniyev, 1986; Weissinger, 1987; High-field and Martin, 1989a, b, c; Highfield, 1990;Chkhikvadze and Bakradze, 1991, 2002; Perälä,1996, 2002a; Pieh, 2001; Pieh and Perälä, 2002,2004; van der Kuyl et al., 2002; table 1), andsome authors suggested that as many as 20 dis-tinct species were traditionally lumped underT. graeca (Bour, 1989; Highfield and Martin,1989a, b, c; Highfield, 1990; Gmira, 1993a, b,1995; David, 1994; Pieh, 2001; Perälä, 2002a,b; Pieh and Perälä, 2002, 2004; see also Guyot-Jackson, 2004). In contrast, Fritz et al. (1996)argued that most taxa within the T. graeca com-plex are badly defined and that two distinct lin-eages exist, one including the populations ofsmall to moderately sized tortoises with light
coloration occurring in Spain, North Africa andalong the Levantine coast, and a second lin-eage consisting of the populations with largerand darker coloured tortoises from the rest ofthe range. Gmira (1993a, b, 1995) and Perälä(2002b) proposed that species of the T. graecacomplex are paraphyletic with respect to twoother traditional Testudo species that were neverconsidered part of the T. graeca complex (T.kleinmanni Lortet, 1883 sensu lato, T. mar-ginata Schoepff, 1792). As a consequence, sys-tematics and taxonomy in these tortoises be-came a much debated field (Gasperetti et al.,1993; Ernst et al., 2000; Buskirk et al., 2001;Fritz and Cheylan, 2001; van der Kuyl et al.,2002, 2005; Harris et al., 2003; Guyot-Jackson,2004; Perälä, 2004a; Semyenova et al., 2004;Carretero et al., 2005; Korsunenko et al., 2005).In this study we treat all taxa within the T.graeca complex provisionally as subspecies ofT. graeca.
Some recent studies on mtDNA sequencevariation of T. graeca, mainly in the westernMediterranean (Álvarez et al., 2000; van derKuyl et al., 2002, 2005; Harris et al., 2003), usedthe slowly evolving mitochondrial 12S rRNAgene and in part 426 bp of the cytochrome b
gene (Álvarez et al., 2000) and 411 bp of the D-loop (van der Kuyl et al., 2005). These studiessuggested distinctly less taxonomic differentia-tion than morphological investigations. In con-trast, Semyenova et al. (2004) and Korsunenkoet al. (2005) found four eastern subspecies (T.g. ibera, T. g. nikolskii, T. g. pallasi, T. g.terrestris) distinct using nuclear genomic fin-gerprinting (Randomly Amplified PolymorphicDNA = RAPDs). If it is considered that RAPDsmay reflect population-specific, and not taxon-specific, differentiation (e.g. Haig et al., 1994,1997; Prior et al., 1997; Tassanakajon et al.,1997; Gomes et al., 1998; Vanlerberghe-Masuttiand Chavigny, 1998; Palkovacs et al., 2003),these contradictory findings are not surprising.
Here we present for the first time a range-wide phylogeography for the T. graeca com-plex that includes most nominal taxa (table 1).
Testudo graeca complex 99
Figure 1. Range of the Testudo graeca complex (shaded) and collection sites of samples. Neighbouring localities combined;for exact localities see table 2. Symbols correspond to distinct mtDNA clades. Range according to Gasperetti et al. (1993)and Buskirk et al. (2001); some records of introduced tortoises omitted.
We used the mitochondrial cytochrome b genethat in testudinids is phylogenetically more in-formative than the 12S rRNA gene and a power-ful tool for revealing their matrilineal differen-tiation (Caccone et al., 1999; Palkovacs et al.,2002; Austin et al., 2003; Fritz et al., 2005a,2006). To find out whether gene flow takes placebetween populations harbouring different ma-trilineages, we applied ISSR (Inter-Simple Se-quence Repeat) nuclear genomic fingerprinting,a technique that has been shown to be useful forTestudo species (Fritz et al., 2005a) and otherchelonians as well (Wink et al., 2001; Guick-ing et al., 2002; Schilde et al., 2004; Fritz et al.,2005b). ISSR PCR produces species-specificnuclear fingerprints in a wide range of organ-isms (e.g. Gupta et al., 1994; Zietkiewicz et al.,1994; Wink et al., 1998; Bornet and Branchard,2001; Fritz et al., 2005a, b; Hundsdörfer andWink, 2005), allowing identification of inter-specific hybrids (Wink et al., 2001; Schilde etal., 2004), and of gene flow and introgression(Wolfe et al., 1998; Nagy et al., 2003; Fritz etal., 2005b).
Materials and methods
Sampling
We studied 94 samples from localities covering most of thedistribution range of the Testudo graeca complex (fig. 1;table 2). Blood samples were obtained by coccygeal veinpuncture of wild or captive tortoises. Alternatively, muscletissue was extracted from thighs of frozen carcasses priorto alcohol preservation. Samples were either preserved inan EDTA buffer (0.1 M Tris, pH 7.4, 10% EDTA, 1%NaF, 0.1% thymol) or in ethanol, and stored at −20◦C untilprocessing. Remaining blood, tissue and DNA samples arepermanently kept at −80◦C in the blood and tissue samplecollection of the Museum of Zoology, Dresden.
DNA extraction, marker gene amplification and sequencing
Total genomic DNA was isolated by an overnight incubationat 37◦C in lysis buffer (10 mM Tris, pH 7.5, 25 mMEDTA, 75 mM NaCl, 1% SDS) including 1 mg of proteinaseK (Roth or Merck), followed by purification with eitherthe DTAB method (Gustincich et al., 1991) or a phenol-chloroform protocol (Sambrook et al., 1989). The primersmt-A1 5′-CCC CCT ACC AAC ATC TCA GCA TGA TGAAAC TTC G-3′ or mt-a-neu2 5′-CTC CCA GCC CCA TCCAAC ATC TCA GCA TGA TGA AAC-3′ and mtf-na 5′-AGG GTG GAG TCT TCA GTT TTT GGT TTA CAAGAC CAA TG-3′ or mt-Fr 5′-CTA AGA AGG GTG GAGTCT TCA GTT TTT GGT TTA CAA-3′ were used foramplification of an approximately 1150 bp long mtDNA
100 U. Fritz et al.Ta
ble
1.Ta
xaof
the
Test
udo
grae
caco
mpl
ex.S
ever
alau
thor
sbe
lieve
that
furt
her
taxa
exis
tbut
assi
gnno
nam
es(t
wo
addi
tiona
lLev
antin
eta
xaw
ithin
T.g.
anta
kyen
sis:
Perä
lä,2
002a
,b;
furt
her
taxa
with
inT.
g.ib
era:
Perä
lä,2
002b
;Pie
het
al.,
2002
;afu
rthe
rta
xon
inT
urkm
enia
:Pie
han
dPe
rälä
,200
1).I
tis
deba
ted
whe
ther
the
nam
eTe
stud
ow
hite
iBen
nett
inW
hite
,183
6is
appl
icab
leto
any
Nor
thA
fric
anpo
pula
tion
orno
t(Pi
ehan
dPe
rälä
,200
2,20
04;v
ande
rK
uyle
tal.,
2002
,200
5;Pe
rälä
,200
4a).
Pieh
and
Perä
lä(2
002,
2004
)tr
eatT
.whi
teia
sdo
ubtf
ulna
me
that
cann
otbe
allo
cate
dto
ace
rtai
npo
pula
tion,
whi
leva
nde
rK
uyl
etal
.(20
02,
2005
)us
eT.
g.w
hite
ifo
rse
vera
lN
orth
Afr
ican
popu
latio
nsfr
omw
ithin
the
rang
eof
othe
rta
xa.
Mor
eove
r,va
nde
rK
uyle
tal.
(200
2)pr
opos
edth
atSa
rdin
ian
T.gr
aeca
mig
htre
pres
enta
nad
ditio
nals
ubsp
ecie
sal
thou
ghth
eydi
dno
trep
ortm
orph
olog
ical
orge
netic
dist
inct
ion.
Val
idity
ofTe
stud
ofla
vom
inim
aral
isH
ighfi
eld
and
Mar
tin,1
989
isge
nera
llyno
tac
cept
ed(E
rnst
etal
.,20
00;
Bus
kirk
etal
.,20
01;
Pieh
and
Perä
lä,2
002,
2004
;va
nde
rK
uyl
etal
.,20
02,2
005)
.(+
)Sa
mpl
esof
resp
ectiv
eta
xon
stud
ied
for
pres
ent
pape
r,(−
)no
tst
udie
d,(?
)un
clea
r,se
eab
ove.
MSL
=m
axim
umsh
ell
leng
th.M
SLfo
rTe
stud
ogr
aeca
grae
cafr
omPi
ehan
dPe
rälä
(200
4),f
orT.
g.ib
era
sens
ula
tofr
omB
uski
rket
al.(
2001
),fo
rT.
g.la
mbe
rti
and
T.g.
mar
okke
nsis
from
Pieh
and
Perä
lä(2
004)
,and
for
T.g.
whi
tei
from
Ern
stet
al.(
2000
);ot
her
data
from
Guy
ot-J
acks
on(2
004)
.Dis
trib
utio
nra
nges
asde
fined
inor
igin
alde
scri
ptio
ns,i
fno
toth
erw
ise
stat
ed.
Taxo
nM
SL(m
m)
MtD
NA
clad
eTy
pelo
calit
yA
ppro
xim
ate
rang
e
+Te
stud
ogr
aeca
anam
uren
sis
Wei
ssin
ger,
1987
267
DA
nam
urum
,sou
ther
nT
urke
yC
entr
alM
edite
rran
ean
coas
tof
Tur
key
+Te
stud
ogr
aeca
anta
kyen
sis
Perä
lä,1
996
250
DA
ntak
ya,s
outh
east
ern
Tur
key
Lev
ant
with
out
cent
ral
and
sout
hern
coas
tal
part
(Per
älä,
2002
a)+
Test
udo
grae
caar
men
iaca
Chk
hikv
adze
and
Bak
radz
e,19
9125
8A
Meg
ri,s
outh
east
ern
Arm
enia
Ara
xes
Val
ley
and
adja
cent
regi
ons
inT
urke
y,A
rmen
ia,
Aze
rbai
jan,
and
Iran
(Pie
het
al.,
2002
)+
Test
udo
grae
cabu
xton
iBou
leng
er,1
921
267
EM
anjil
,bet
wee
nR
esht
and
Kas
win
,Ira
nSo
uthw
este
rnco
rner
ofC
aspi
anSe
a,Ir
an(P
eräl
ä,20
04b)
+Te
stud
ogr
aeca
cyre
naic
aPi
ehan
dPe
rälä
,200
220
5B
Dar
nah,
east
ern
Lib
yaC
yren
aica
Peni
nsul
a,L
ibya
+Te
stud
ogr
aeca
flow
eriB
oden
heim
er,1
935
154
DN
egev
,pro
babl
yvi
cini
tyof
Gaz
a,Pa
lest
ine
(res
tric
ted
byB
our,
1989
,but
see
Perä
läan
dSh
acha
m,2
004)
Coa
stal
plai
nsof
Gaz
a,Is
rael
and
Leb
anon
(Per
älä,
2002
a;Pe
rälä
and
Shac
ham
,200
4)
+Te
stud
ogr
aeca
grae
caL
inna
eus,
1758
belo
w21
4B
Sant
aC
ruz
near
Ora
n,A
lger
ia(r
estr
icte
dby
Mer
tens
and
Mül
ler,
1928
)N
orth
east
ern
Mor
occo
(eas
tern
part
ofM
edite
rran
ean
coas
tan
dad
jace
ntin
land
regi
ons)
,A
lger
ia(P
ieh
and
Perä
lä,
2004
)+
Test
udo
grae
caib
era
Palla
s,18
14se
nsu
stri
cto:
260
sens
ula
to:
358
sens
ust
rict
o:A
,Cse
nsu
lato
:A
,C,D
Tbi
lisi,
Geo
rgia
(des
igna
ted
byB
our,
1987
)Se
nsu
stri
cto:
Kur
aR
iver
Bas
in,
Cau
casu
s(P
eräl
ä,20
02a;
Dan
ilov
and
Milt
o,20
04);
iber
ase
nsu
lato
incl
udes
also
sout
heas
tern
Eur
ope
and
vast
part
sof
Asi
aM
inor
(Bus
kirk
etal
.,20
01)
−Te
stud
ogr
aeca
lam
bert
iPie
han
dPe
rälä
,200
421
4?
22km
nort
hof
Tét
ouan
,Mor
occo
Nor
ther
nM
oroc
co(w
este
rnpa
rtof
Med
iterr
anea
nco
ast)
−Te
stud
ogr
aeca
mar
okke
nsis
Pieh
and
Perä
lä,
2004
237
?Ta
rmile
te,M
oroc
coN
orth
wes
tern
Mor
occo
(Atla
ntic
coas
tan
dad
jace
ntin
land
regi
on)
Testudo graeca complex 101
Tabl
e1.
(Con
tinue
d).
Taxo
nM
SL(m
m)
MtD
NA
clad
eTy
pelo
calit
yA
ppro
xim
ate
rang
e
+Te
stud
ogr
aeca
nabe
ulen
sis
(Hig
hfiel
d,19
90)
belo
w18
0B
Fore
st7-
8km
nort
hwes
tof
Nab
eul,
indi
-re
ctio
nof
Gro
mba
lia,T
unis
ia(r
estr
icte
dby
Pieh
and
Perä
lä,2
004)
Tun
isia
,adj
acen
tLib
ya(P
ieh
and
Perä
lä,2
002)
+Te
stud
ogr
aeca
niko
lski
iC
hkhi
kvad
zean
dT
u-ni
yev,
1986
297
CN
ebug
Settl
emen
t,T
uaps
eC
ount
y,K
rasn
odar
Dis
tric
t,R
ussi
aN
orth
ern
Bla
ckSe
aco
ast
ofR
ussi
aan
dG
eorg
ia(B
uski
rket
al.,
2001
)+
Test
udo
grae
capa
llas
iC
hkhi
kvad
zean
dB
akra
dze,
2002
247
AG
ilyar
y-D
agSe
ttlem
ent,
Dag
esta
n,R
ussi
aC
aspi
anSe
aco
asta
lare
ain
Dag
esta
nan
dad
jace
ntA
zerb
ai-
jan
(Dan
ilov
etal
.,20
04)
+Te
stud
ogr
aeca
pers
esPe
rälä
,200
224
0E
,F3
mile
sW
Lal
abad
Vill
age,
25m
iles
NW
Ker
man
shah
,Ker
man
shah
anPr
ovin
ce,
wes
tern
cent
ralI
ran
Zag
ros
Mou
ntai
ns(I
ran)
and
adja
cent
mou
ntai
nch
ains
−Te
stud
ogr
aeca
sous
sens
isPi
eh,2
001
249
?A
gadi
r,M
oroc
coSo
uthw
este
rnM
oroc
co,
regi
onof
Mar
rake
chan
dre
gion
sout
hwes
tof
Hig
hA
tlas
Mou
ntai
ns(P
ieh
and
Perä
lä,2
004)
+Te
stud
ogr
aeca
terr
estr
isFo
rssk
ål,1
775
254
DA
lepp
o,Sy
ria
(by
neot
ype
desi
gnat
ion;
Perä
läan
dB
our,
2004
)N
orth
ern
Mes
opot
amia
(Bou
ran
dPe
rälä
,200
4;Pe
rälä
and
Bou
r,20
04)
?Te
stud
ogr
aeca
whi
teiB
enne
ttin
Whi
te,1
836
292
?A
lgie
rsan
dits
envi
rons
,A
lger
ia(d
esig
-na
ted
byH
ighfi
eld
and
Mar
tin,
1989
b,bu
tse
ere
mar
ks)
See
rem
arks
+Te
stud
ogr
aeca
zaru
dnyi
Nik
olsk
y,18
9627
5F
Bır
jand
,K
hora
san
Prov
ince
,no
rthe
aste
rnIr
an(r
estr
icte
dby
Perä
lä,2
002a
)E
aste
rnIr
an(P
eräl
ä,20
02a)
102 U. Fritz et al.Ta
ble
2.St
udie
dsa
mpl
esof
the
Test
udo
grae
caco
mpl
exan
dou
tgro
ups.
Que
stio
nm
arks
deno
teun
cert
ain
taxo
nom
ical
loca
tions
due
toco
llect
ion
site
sal
ong
rang
ebo
rder
s(f
orA
diya
man
,T
urke
y,se
eT
ürko
zan
etal
.,20
03).
MT
DD
num
bers
refe
rto
com
plet
evo
uche
rsp
ecim
ens,
MT
DT
num
bers
tobl
ood,
tissu
eor
DN
Asa
mpl
esin
the
colle
ctio
nof
the
Mus
eum
ofZ
oolo
gy,
Dre
sden
.Pro
visi
onal
HD
num
bers
ofsa
mpl
esst
udie
dby
Fritz
etal
.(20
05a)
repl
aced
bype
rman
entM
TD
num
bers
.Acc
essi
onnu
mbe
rsre
fer
tom
tDN
Ase
quen
ces.
Sam
ple
Taxo
nL
ocal
ityM
TD
Acc
essi
onnu
mbe
rIS
SR(G
AA
) 5
ISSR
(GA
CA
) 4
ISSR
L18
AN
AM
UR
EN
SIS
1Te
stud
ogr
aeca
anam
uren
sis
Tur
key:
Ana
mur
um;3
6◦03
′ N32
◦ 51′
ET
257
AJ8
8834
7A
NA
MU
RE
NSI
S2
Test
udo
grae
caan
amur
ensi
sT
urke
y:G
azip
asa;
36◦ 1
7′N
32◦ 1
8′E
T25
9A
J888
348
AN
AM
UR
EN
SIS
3Te
stud
ogr
aeca
anam
uren
sis
Tur
key:
Gaz
ipas
a;36
◦ 17′
N32
◦ 18′
ET
258
AM
2309
46A
NA
MU
RE
NSI
S4
Test
udo
grae
caan
amur
ensi
sT
urke
y:Sa
hayi
Sitig
eS
Seri
kan
dA
spen
dos;
36◦ 5
0′02
′′ N31
◦ 09′
01′′ E
T15
1A
J888
356
++
+A
NA
MU
RE
NSI
S5
Test
udo
grae
caan
amur
ensi
sT
urke
y:Si
de;3
6◦46
′ 52′′
N31
◦ 23′
56′′ E
T61
0A
M23
1002
++
+A
NA
MU
RE
NSI
S6
Test
udo
grae
caan
amur
ensi
sT
urke
y:Si
de;3
6◦47
′ 18′′
N31
◦ 24′
25′′ E
T60
9A
M23
0948
++
+A
NA
MU
RE
NSI
S7
Test
udo
grae
caan
amur
ensi
sT
urke
y:W
Side
;36◦
49′ N
31◦ 2
1′E
T25
3A
M23
1003
++
AN
AM
UR
EN
SIS
8Te
stud
ogr
aeca
anam
uren
sis
Tur
key:
WSi
de;3
6◦49
′ N31
◦ 21′
ET
254
AM
2310
04+
++
?A
NA
MU
RE
NSI
S9
Test
udo
grae
caan
amur
ensi
s?T
urke
y:M
ersi
n;36
◦ 49′
N34
◦ 39′
ET
260
AM
2310
05A
NTA
KY
EN
SIS
1Te
stud
ogr
aeca
anta
kyen
sis
Isra
el:T
iber
ias;
32◦ 4
8′N
35◦ 3
1′E
T16
44A
M23
0947
AN
TAK
YE
NSI
S2
Test
udo
grae
caan
taky
ensi
sIs
rael
:Tib
eria
s;32
◦ 48′
N35
◦ 31′
ET
1376
AJ8
8834
4A
NTA
KY
EN
SIS
3Te
stud
ogr
aeca
anta
kyen
sis
Isra
el:T
iber
ias;
32◦ 4
8′N
35◦ 3
1′E
T13
77A
J888
345
AN
TAK
YE
NSI
S4
Test
udo
grae
caan
taky
ensi
sJo
rdan
:Jar
ash;
32◦ 1
1′N
35◦ 5
1′E
T13
80A
J888
346
AN
TAK
YE
NSI
S5
Test
udo
grae
caan
taky
ensi
sSy
ria:
Ant
iLeb
anon
Mts
.(Ja
bale
shSh
arqi
):M
a’lu
la;3
3◦51
′ N36
◦ 33′
ET
2179
AM
2309
49A
NTA
KY
EN
SIS
6Te
stud
ogr
aeca
anta
kyen
sis
Syri
a:A
ntiL
eban
onM
ts.(
Jaba
lesh
Shar
qi):
Sayd
naya
;33◦
42′ N
36◦ 2
2′E
T21
74A
M23
0950
AN
TAK
YE
NSI
S7
Test
udo
grae
caan
taky
ensi
sSy
ria:
Jaba
lalN
usay
rıya
h:A
ynal
Bay
dah;
34◦ 5
9′N
36◦ 2
0′E
T21
83A
M23
0951
++
+A
NTA
KY
EN
SIS
8Te
stud
ogr
aeca
anta
kyen
sis
Syri
a:Ja
bala
lNus
ayrı
yah:
Jour
ine
(atQ
alat
Mer
za);
35◦ 3
9′N
36◦ 1
6′E
T21
88A
M23
0952
AN
TAK
YE
NSI
S9
Test
udo
grae
caan
taky
ensi
sSy
ria:
Jaba
lalN
usay
rıya
h:M
asya
f;35
◦ 04′
N36
◦ 21′
ET
2181
AM
2309
53+
++
AN
TAK
YE
NSI
S10
Test
udo
grae
caan
taky
ensi
sSy
ria:
Jaba
lDur
uz:A
lKaf
r;32
◦ 39′
N36
◦ 38′
ET
2169
AM
2309
54A
NTA
KY
EN
SIS
11Te
stud
ogr
aeca
anta
kyen
sis
Syri
a:Ja
balD
uruz
:As
Suw
ayda
’;32
◦ 43′
N36
◦ 35′
ET
2162
AM
2309
55A
NTA
KY
EN
SIS
12Te
stud
ogr
aeca
anta
kyen
sis
Syri
a:Ja
balD
uruz
:hal
f-w
ayA
sSu
way
da’
toSa
leh;
32◦ 4
1′N
36◦ 4
1′E
T21
63A
M23
0956
AR
ME
NIA
CA
Test
udo
grae
caar
men
iaca
Tur
key:
Ara
xes
Riv
erV
alle
y:V
ilaye
tIgd
ır:M
elek
li;39
◦ 57N
’44
◦ 06′
ET
2095
AM
2309
57+
++
BU
XTO
NI
1Te
stud
ogr
aeca
buxt
oni
Iran
:Now
shar
near
Man
jil;3
6◦44
′ 03′′
N49
◦ 25′
12′′ E
T22
60A
M23
0958
+B
UX
TON
I2
Test
udo
grae
cabu
xton
iIr
an:S
Res
ht:b
etw
een
Sara
van
and
Ros
tam
abad
;36◦
57′ N
49◦ 3
3′E
T22
65A
M23
0959
++
+B
UX
TON
I3
Test
udo
grae
cabu
xton
iIr
an:S
Res
ht:b
etw
een
Sara
van
and
Ros
tam
abad
;36◦
57′ N
49◦ 3
3′E
T22
67A
M23
0960
BU
XTO
NI
4Te
stud
ogr
aeca
buxt
oni
Iran
:SR
esht
:bet
wee
nSa
rava
nan
dR
osta
mab
ad;3
6◦57
′ N49
◦ 33′
ET
2268
AM
2309
61+
++
BU
XTO
NI
5Te
stud
ogr
aeca
buxt
oni
Iran
:Sefi
dR
ud;3
7◦22
′ 20′′
N48
◦ 08′
10′′ E
T19
99A
M23
0962
++
+C
YR
EN
AIC
ATe
stud
ogr
aeca
cyre
naic
aL
ibya
:Cyr
enai
ca,a
ppro
x.32
◦ 25′
N21
◦ 20′
ED
4281
9A
J888
341
++
+
Testudo graeca complex 103Ta
ble
2.(C
ontin
ued)
.
Sam
ple
Taxo
nL
ocal
ityM
TD
Acc
essi
onnu
mbe
rIS
SR(G
AA
) 5
ISSR
(GA
CA
) 4
ISSR
L18
FL
OW
ER
I1
Test
udo
grae
caflo
wer
iIs
rael
:Bet
Zev
i,K
arm
elM
ts.;
32◦ 4
3′N
34◦ 5
8′E
T13
74A
M23
0963
FL
OW
ER
I2
Test
udo
grae
caflo
wer
iIs
rael
:Bet
Zev
i,K
arm
elM
ts.;
32◦ 4
3′N
34◦ 5
8′E
T13
75A
M23
1006
FL
OW
ER
I3
Test
udo
grae
caflo
wer
iIs
rael
:Ros
hH
a’A
yin;
32◦ 0
7′N
34◦ 5
8′E
T13
78A
M23
0964
++
+F
LO
WE
RI
4Te
stud
ogr
aeca
flow
eri
Isra
el:R
osh
Ha’
Ayi
n;32
◦ 07′
N34
◦ 58′
ET
1379
AM
2310
07+
+G
RA
EC
ATe
stud
ogr
aeca
grae
caM
oroc
co:N
EO
tatO
ulad
elH
ajj;
33◦ 3
7′20
′′ N03
◦ 06′
39′′ W
T22
94A
M23
0965
++
+M
AL
LO
RC
ATe
stud
ogr
aeca
grae
cas.
l.Sp
ain:
Mal
lorc
a:N
Cal
vià;
39◦ 3
6′N
02◦ 3
1′E
D42
822
AJ8
8834
2SA
RD
INIA
Test
udo
grae
cagr
aeca
s.l.
Ital
y:Sa
rdin
ia:S
inis
Peni
nsul
a;ap
prox
.40◦
00′ N
08◦ 2
5′E
T11
13A
J888
343
SIC
ILY
1Te
stud
ogr
aeca
grae
cas.
l.It
aly:
Sici
ly:M
arsa
la;3
7◦48
′ N12
◦ 27′
ET
2230
AM
2309
66IB
ER
A1
Test
udo
grae
caib
era
Aze
rbai
jan:
Kat
ekh;
41◦ 3
9′N
46◦ 3
4′E
T14
61A
M23
1008
++
IBE
RA
2Te
stud
ogr
aeca
iber
aA
zerb
aija
n:K
atek
h;41
◦ 39′
N46
◦ 34′
ET
1462
AM
2310
09+
++
IBE
RA
3Te
stud
ogr
aeca
iber
aA
zerb
aija
n:SW
Bey
laqa
n,ne
arD
ashb
urun
;39◦
43′ N
47◦ 3
4′E
T14
46A
M23
0969
++
+IB
ER
A4
Test
udo
grae
caib
era
Aze
rbai
jan:
SWB
eyla
qan,
near
Das
hbur
un;3
9◦43
′ N47
◦ 34′
ET
1447
AM
2309
67+
+IB
ER
A5
Test
udo
grae
caib
era
Aze
rbai
jan:
SWB
eyla
qan,
near
Das
hbur
un;3
9◦43
′ N47
◦ 34′
ET
1448
AM
2309
68+
++
IBE
RA
6Te
stud
ogr
aeca
iber
aB
ulga
ria:
Alb
ena;
43◦ 2
2′27
′′ N28
◦ 05′
00′′ E
T72
4A
M23
1010
IBE
RA
7Te
stud
ogr
aeca
iber
aB
ulga
ria:
Alb
ena;
43◦ 2
2′53
′′ N28
◦ 04′
97′′ E
T72
5A
J888
349
IBE
RA
8Te
stud
ogr
aeca
iber
aB
ulga
ria:
Alb
ena;
43◦ 2
2′37
′′ N28
◦ 05′
14′′ E
T72
7A
J888
350
IBE
RA
9Te
stud
ogr
aeca
iber
aB
ulga
ria:
Piri
n;41
◦ 33′
N23
◦ 34′
ET
1365
AM
2310
11IB
ER
A10
Test
udo
grae
caib
era
Bul
gari
a:Ž
elez
ino;
41◦ 2
8′N
25◦ 5
6′E
T23
32A
M23
1012
IBE
RA
11Te
stud
ogr
aeca
iber
aG
eorg
ia:M
tskh
eta;
41◦ 5
0′N
44◦ 4
3′E
D40
654
AM
2310
13+
+IB
ER
A12
Test
udo
grae
caib
era
Gre
ece:
15km
EA
lexa
ndro
úpol
is;4
0◦50
′ N26
◦ 00′
ET
1615
AM
2310
14+
++
IBE
RA
13Te
stud
ogr
aeca
iber
aG
reec
e:be
twee
nD
rám
aan
dX
anth
i;41
◦ 11′
N24
◦ 37′
ET
1614
AM
2310
15+
++
IBE
RA
14Te
stud
ogr
aeca
iber
aG
reec
e:E
vros
Del
ta,a
ppro
x.25
kmE
Ale
xand
roúp
olis
;40◦
46′ N
26◦ 0
5′E
T16
16A
M23
1016
IBE
RA
15Te
stud
ogr
aeca
iber
aG
reec
e:K
osIs
land
;36◦
52′ 3
7′′N
27◦ 1
7′76
′′ ET
821
AJ8
8835
1IB
ER
A16
Test
udo
grae
caib
era
Gre
ece:
Kos
Isla
nd;3
6◦53
′ 36′′
N27
◦ 17′
49′′ E
T82
3A
J888
352
IBE
RA
17Te
stud
ogr
aeca
iber
aR
epub
licof
Mac
edon
ia:S
kopj
e;42
◦ 00′
N21
◦ 28′
ET
1975
AM
2310
17IB
ER
A18
Test
udo
grae
caib
era
Rom
ania
:His
tria
;44◦
35′ N
28◦ 4
2′E
T13
62A
M23
1018
++
+IB
ER
A19
Test
udo
grae
caib
era
Rom
ania
:His
tria
;44◦
35′ N
28◦ 4
2′E
T13
99A
M23
1019
++
+IB
ER
A20
Test
udo
grae
caib
era
Rom
ania
:His
tria
;44◦
35′ N
28◦ 4
2′E
T16
34A
M23
1020
IBE
RA
21Te
stud
ogr
aeca
iber
aT
urke
y:M
enem
en;3
8◦37
′ N27
◦ 04′
ET
268
AM
2310
21IB
ER
A22
Test
udo
grae
caib
era
Tur
key:
Seyd
iseh
ir;3
7◦25
′ 19′′
N31
◦ 50′
05′′ E
T15
0A
J888
355
++
+
104 U. Fritz et al.
Tabl
e2.
(Con
tinue
d).
Sam
ple
Taxo
nL
ocal
ityM
TD
Acc
essi
onnu
mbe
rIS
SR(G
AA
) 5
ISSR
(GA
CA
) 4
ISSR
L18
IBE
RA
23Te
stud
ogr
aeca
iber
aT
urke
y:L
ake
Van
:10
kmN
Van
;38◦
34′ N
43◦ 2
4′E
T38
2A
J888
354
++
IBE
RA
24Te
stud
ogr
aeca
iber
aT
urke
y:L
ake
Van
:Bud
akli;
38◦ 2
3′56
′′ N42
◦ 37′
56′′ E
T38
1A
M23
1022
+IB
ER
A25
Test
udo
grae
caib
era
Tur
key:
Lak
eV
an:K
üçük
su;3
8◦27
′ N42
◦ 20′
ET
380
AM
2310
23+
+IB
ER
A26
Test
udo
grae
caib
era
Tur
key:
Lak
eV
an:E
rcis
;39◦
02′ N
43◦ 2
3′E
T27
21A
J888
353
++
+SI
CIL
Y2
Test
udo
grae
caib
era
Ital
y:Si
cily
:Cam
pobe
llodi
Maz
ara;
37◦ 3
8′N
12◦ 4
5′E
T21
42A
M23
0970
NA
BE
UL
EN
SIS
1Te
stud
ogr
aeca
nabe
ulen
sis
Tun
isia
:Sou
sse
(Sus
ah);
35◦ 5
0′N
10◦ 3
9′E
T14
9A
M23
0971
NA
BE
UL
EN
SIS
2Te
stud
ogr
aeca
nabe
ulen
sis
Tun
isia
D42
893
AM
2309
72N
IKO
LSK
II1
Test
udo
grae
cani
kols
kii
Rus
sia:
NT
uaps
e;44
◦ 11′
N39
◦ 06′
ET
1808
AM
2309
73+
+N
IKO
LSK
II2
Test
udo
grae
cani
kols
kii
Rus
sia:
Sukk
o,SE
Ana
pa;4
4◦47
′ N37
◦ 23′
ET
1806
AM
2309
74+
PAL
LA
SI1
Test
udo
grae
capa
llas
iA
zerb
aija
n:K
olan
i;41
◦ 11′
N49
◦ 08′
ET
1450
AM
2309
75+
++
PAL
LA
SI2
Test
udo
grae
capa
llas
iA
zerb
aija
n:K
olan
i;41
◦ 11′
N49
◦ 08′
ET
1451
AM
2309
76+
+PA
LL
ASI
3Te
stud
ogr
aeca
pall
asi
Aze
rbai
jan:
Kol
ani;
41◦ 1
1′N
49◦ 0
8′E
T14
52A
M23
0977
++
PAL
LA
SI4
Test
udo
grae
capa
llas
iR
ussi
a:D
ages
tan:
Dag
esta
nski
eO
gni;
42◦ 0
7′N
48◦ 1
1′E
T23
88A
M23
0978
++
PAL
LA
SI5
Test
udo
grae
capa
llas
iR
ussi
a:D
ages
tan:
Dag
esta
nski
eO
gni;
42◦ 0
7′N
48◦ 1
1′E
T23
89A
M23
0979
++
+PA
LL
ASI
6Te
stud
ogr
aeca
pall
asi
Rus
sia:
Dag
esta
n:D
ages
tans
kie
Ogn
i;42
◦ 07′
N48
◦ 11′
ET
2390
AM
2309
80+
+PA
LL
ASI
7Te
stud
ogr
aeca
pall
asi
Rus
sia:
Dag
esta
n:Z
elen
omor
sk;4
2◦45
′ N47
◦ 42′
ET
873
AM
2310
00+
++
PAL
LA
SI8
Test
udo
grae
capa
llas
iR
ussi
a:D
ages
tan:
Zel
enom
orsk
;42◦
45′ N
47◦ 4
2′E
T87
4A
M23
1001
++
+?
PAL
LA
SI9
Test
udo
grae
capa
llas
i?A
zerb
aija
n:A
bshe
ron
Peni
nsul
a:N
ovkh
ani;
40◦ 3
2′N
49◦ 4
7′E
T14
53A
M23
0981
++
+?
PAL
LA
SI10
Test
udo
grae
capa
llas
i?A
zerb
aija
n:A
bshe
ron
Peni
nsul
a:N
ovkh
ani;
40◦ 3
2′N
49◦ 4
7′E
T14
54A
M23
0982
++
+P
ER
SES
1Te
stud
ogr
aeca
pers
esIr
an:A
njır
Ava
nd;3
2◦30
′ N54
◦ 26′
ET
2283
AM
2309
83P
ER
SES
2Te
stud
ogr
aeca
pers
esIr
an:6
kmN
EM
eshg
ınSh
ahr;
38◦ 2
6′N
47◦ 4
2′E
T14
27A
M23
0984
++
PE
RSE
S3
Test
udo
grae
cape
rses
Iran
:6km
NE
Mes
hgın
Shah
r;38
◦ 26′
N47
◦ 42′
ET
1428
AM
2309
85+
+P
ER
SES
4Te
stud
ogr
aeca
pers
esIr
an:N
eyrı
z;29
◦ 12′
N54
◦ 19′
ET
2282
AM
2309
86+
++
PE
RSE
S5
Test
udo
grae
cape
rses
Iran
:Sha
hr-e
Bab
ak,M
aym
and;
appr
ox.3
0◦07
′ N55
◦ 07′
ET
2284
AM
2309
87+
++
PE
RSE
S6
Test
udo
grae
cape
rses
Iran
:Sha
hr-e
Bab
ak,M
aym
and;
appr
ox.3
0◦07
′ N55
◦ 07′
ET
2285
AM
2309
88+
++
PE
RSE
S7
Test
udo
grae
cape
rses
Iran
:EE
sfah
an:K
uhpa
yeh;
32◦ 4
3′N
52◦ 2
6′E
T22
69A
M23
0989
++
+P
ER
SES
8Te
stud
ogr
aeca
pers
esIr
an:N
ır;3
1◦30
′ N54
◦ 08′
ET
2272
AM
2309
90+
+?
PE
RSE
S9
Test
udo
grae
cape
rses
?Ir
an:b
etw
een
Ger
mıa
ndR
azay
-Am
ırA
bad
(Am
ırK
andı
);38
◦ 53′
N48
◦ 00′
ET
2286
AM
2309
91+
++
Testudo graeca complex 105
Tabl
e2.
(Con
tinue
d).
Sam
ple
Taxo
nL
ocal
ityM
TD
Acc
essi
onnu
mbe
rIS
SR(G
AA
) 5
ISSR
(GA
CA
) 4
ISSR
L18
TE
RR
EST
RIS
1Te
stud
ogr
aeca
terr
estr
isSy
ria:
hills
NW
Ale
ppo:
Dar
Ta’i
zzah
;36◦
17′ N
36◦ 5
1′E
T21
93A
M23
0992
++
+T
ER
RE
STR
IS2
Test
udo
grae
cate
rres
tris
Syri
a:hi
llsN
WA
lepp
o:Q
alat
Sam
an;3
6◦22
′ N36
◦ 51′
ET
2194
AM
2309
93T
ER
RE
STR
IS3
Test
udo
grae
cate
rres
tris
Tur
key:
Mar
din;
37◦ 1
9′N
40◦ 4
7′N
T26
7A
M23
0994
TE
RR
EST
RIS
4Te
stud
ogr
aeca
terr
estr
isT
urke
y:St
reet
E-9
9be
twee
nH
ilvan
and
Çay
larb
asi;
37◦ 3
9′N
T36
4A
M23
0995
39◦ 0
6′E
?T
ER
RE
STR
IS5
Test
udo
grae
cate
rres
tris
?T
urke
y:A
diya
man
;37◦
46′ 0
9′′N
38◦ 2
1′35
′′ ET
378
AM
2309
96?
TE
RR
EST
RIS
6Te
stud
ogr
aeca
terr
estr
is?
Tur
key:
Adi
yam
an;3
7◦46
′ N38
◦ 16′
NT
264
AM
2309
97Z
AR
UD
NY
I1
Test
udo
grae
caza
rudn
yiIr
an:S
agha
nd,s
outh
ern
bord
erof
Kav
ırD
eser
t;32
◦ 33′
N55
◦ 13′
ET
2280
AM
2309
98+
++
ZA
RU
DN
YI
2Te
stud
ogr
aeca
zaru
dnyi
Iran
:Tab
as,s
outh
ern
bord
erof
Kav
ırD
eser
t;33
◦ 35′
N56
◦ 56′
ET
2273
AM
2309
99+
++
OU
TG
RO
UPS
:H
ER
MA
NN
IH
ER
MA
NN
I1
Test
udo
herm
anni
herm
anni
Ital
y:R
occa
tede
righ
i;43
◦ 02′
N11
◦ 05′
ED
4159
0A
J888
362
HE
RM
AN
NI
HE
RM
AN
NI
2Te
stud
ohe
rman
nihe
rman
niSo
uth
Ital
yD
4297
2A
J888
364
HO
RSF
IEL
DII
KA
ZA
CH
STA
NIC
ATe
stud
oho
rsfie
ldii
kaza
chst
anic
aK
azak
hsta
nD
4420
1A
J888
365
HO
RSF
IEL
DII
RU
STA
MO
VI
Test
udo
hors
field
iiru
stam
ovi
Iran
:Gor
gan;
36◦ 5
0′N
54◦ 2
6′E
T14
19A
J888
366
KL
EIN
MA
NN
I1
Test
udo
klei
nman
niU
nkno
wn
D44
284
AJ8
8837
0K
LE
INM
AN
NI
2Te
stud
okl
einm
anni
Unk
now
nD
4428
5A
J888
371
MA
RG
INA
TA1
Test
udo
mar
gina
taG
reec
e:Pe
lopo
nnes
e:D
ídim
i;37
◦ 28′
N23
◦ 07′
ET
388
AJ8
8831
8M
AR
GIN
ATA
2Te
stud
om
argi
nata
Ital
y:Sa
rdin
ia:P
ittul
ongo
;40◦
59′ N
09◦ 3
5′E
T11
17A
J888
332
106 U. Fritz et al.
fragment (cytochrome b gene and adjacent portion of tRNA-Thr gene). PCR was performed in a 50 µl volume (BioronPCR buffer or 50 mM KCl, 1.5 mM MgCl2, and 10 mMTris-HCl, 0.5% Triton X-100, pH 8.5) containing 1 unit ofTaq DNA polymerase (Bioron), 10 pmol dNTPs and 5 or10 pmol of each primer. The PCR program consisted of a5 min initial denaturation at 94◦C, followed by 35 cyclesof 45 s at 94◦C, 52 s at 55-60◦C and 80 s at 72◦C witha final elongation step of 10 min at 72◦C, or alternatively,a 4 min initial denaturation at 94◦C, then 30 cycles of45 s at 94◦C, 60 s at 52◦C and 120 s at 72◦C with a10 min final elongation step at 72◦C. PCR products weresequenced directly on both strands on ABI or MegaBace1000 (Amersham Biosciences) sequencers using the internalprimers mt-c2 5′-TGA GGA CAA ATA TCA TTC TGAGG 3′ or mt-c-For2 5′-TGA GGV CAR ATA TCA TTYTGA G-3′ and mt-E-Rev 5′-GCA AAT AGG AAG TATCAT TCT GG-3′ or mt-E-Rev2 5′-GCR AAT ARR AAGTAT CAT TCT GG-3′.
Sequence alignment, phylogenetic analyses and geneticdistances
The corresponding sequences from both strands were com-bined with the software DNAsis 7.00 (copyright HitachiSoftware Engineering Company, 1991) for every samplewith the resulting consensus sequences being aligned man-ually using BioEdit 7.0.5.2 (Hall, 1999). Sixteen sequences(AJ888341-56) were already published in a previous study(Fritz et al., 2005a); sequences of all other Testudo speciesfrom this paper served as outgroups for our phylogeneticanalyses (table 2). Data were analyzed using Bayesian in-ference of phylogeny as implemented in MrBayes 3.1 (Ron-quist and Huelsenbeck, 2003), maximum parsimony (MP;equal weighting) and neighbour joining (NJ; with model-corrected maximum likelihood distances) as implementedin PAUP* 4.0b10 (Swofford, 2002). Bayesian analysis wasperformed using four chains of 1 000 000 generations sam-pling every 100 generations and with the first 1000 gen-erations discarded as burn-in (with which only the plateauof the most likely trees was sampled). The best evolution-ary model for the data (also included in the analysis) wasestablished by hierarchical likelihood testing using Mod-eltest 3.06 (Posada and Crandall, 1998). For the ingroupspecies, 934 of 1134 aligned sites were constant, 74 char-acters were variable but parsimony-uninformative, and 126variable characters were parsimony-informative. Due to theprotracted computation times for MP analyses, the num-ber of equally most parsimonious solutions that were savedneeded to be limited using the maxtrees option. However,to determine whether or not this constraint was limitingthe search unduly, maxtrees was set to 10 000, 50 000 and100 000 in three separate analyses and the resolution andtopologies of the resulting strict consensus trees were com-pared. All calculations resulted in MP trees of 603 steps(CI = 0.6385, RI = 0.9135). To test the robustness of ob-tained branching patterns, bootstrap permutations (Felsen-stein, 1985) were run under both MP (setting maxtrees =1000, nreps = 1000) and NJ (ML-distances; nreps = 1000).Genetic distances (uncorrected p distances) were calculated
with PAUP* 4.0b10; previously published sequences (Fritzet al., 2005a) were used for comparison with other Testudospecies.
Nuclear genomic fingerprinting and analysis
ISSR employs a single PCR primer, binding to di- or trin-ucleotide repeat motifs (microsatellites), which are abun-dant in eukaryotic genomes (Tautz and Renz, 1984; Con-dit and Hubbell, 1991). Since sequences of microsatellitesare conserved over a wide range of organisms, universalprimers can be applied. Amplified regions correspond to thenucleotide sequence between two inverted simple sequencerepeat (SSR) priming sites (Wolfe et al., 1998; Bornet andBranchard, 2001). SSR regions appear to be scattered evenlythroughout the genome (Tautz and Renz, 1984; Condit andHubbell, 1991), resulting in a large number of polymorphicbands. ISSR markers are inherited in a dominant or codom-inant Mendelian fashion (Gupta et al., 1994; Tsumura et al.,1996). They are interpreted as dominant markers scored asdiallelic with ‘band present’ or ‘band absent’. The absenceof a band is interpreted as primer divergence or loss of alocus through the deletion of the SSR site or chromosomalrearrangement (Wolfe and Liston, 1998; Wolfe et al., 1998).ISSR fingerprints are usually diagnostic for species-leveltaxa (e.g. Gupta et al., 1994; Zietkiewicz et al., 1994; Winket al., 1998; Bornet and Branchard, 2001; Fritz et al., 2005a,b; Hundsdörfer and Wink, 2005). Individuals from contactzones of distinct subspecies as well as interspecific hybridsshare with both parental taxa diagnostic bands. Thus, lim-ited or non-existing gene flow is reflected by distinct band-ing patterns for reproductively isolated taxa, while a highpercentage of shared bands is indicative of gene flow orincomplete differentiation (Wolfe et al., 1998; Nagy et al.,2003; Fritz et al., 2005b).
We used two non-anchored primers that yielded species-diagnostic banding patterns for Testudo in a previous study(Fritz et al., 2005a), (GAA)5, annealing temperature 40◦C,and (GACA)4, annealing temperature 55◦C. In addition, weapplied the short anchored primer L18 (CTC GGG AAGGGA), annealing temperature 45◦C, that was useful in an-other chelonian genus, Emys (Fritz et al., 2005b). Each PCRwas performed with approximately 60 ng template DNA ina 25 µl volume [10 pmol of the primer and 0.625 nmol ofeach dNTP, except dATP: 0.28 nmol cold dATP plus 0.1µl radioactive α-33P-dATP solution (370 MBq/ml, Amer-sham Biosciences), 0.75 units of Taq polymerase (SIGMA)and water, buffered with 10 mM Tris-HCl, 50 mM KCl,0.5% Triton X-100, 1.5 mM MgCl2] covered by two dropsof mineral oil. Thermo-cycling was performed with a TrioThermo block TB1 (Biometra, Göttingen). Following an ini-tial 5 min denaturation at 94◦C, the program consisted of 30cycles of 45 s at 94◦C, 60 s at the respective annealing tem-perature, 120 s at 72◦C and 5 min at 72◦C for final elonga-tion. DNA fragments were separated by PAGE in a verticalapparatus (Base Acer Sequencer, Stratagene) for 2.5-4 h at65 W. The denaturing gel [6 M Urea, 100 ml Long RangerSolution, Biozym (PA), 100 ml TBE buffer (10x: 1 M Tris,0.83 M Boric Acid, 10 mM EDTA, pH 8.6)] had a size of45 × 30 cm and a thickness of 0.25 mm. After drying, an
Testudo graeca complex 107
X-ray film (Hyperfilm-MP, Amersham) was exposed to thegel for at least 12 h and developed.
Fragment patterns were analyzed manually. Unmis-takably identifiable bands were transferred into a pres-ence/absence matrix scoring each particular fragment.(GAA)5 yielded 33 fragments for 50 samples, (GACA)4 29fragments for 45 samples, and L18 resulted in 89 fragmentsfor 46 samples. To enable comparison of variation withinT. graeca and other organisms, we report average Dice andJaccard distance values (Jaccard, 1901, 1908; Dice, 1945;Nei and Li, 1979), calculated with RAPDistance 1.04 (Arm-strong et al., 1996), following the appeal for standardizationin ISSR PCR analyses by Hundsdörfer and Wink (2005).Based on the presence/absence matrix, cluster analyses (NJtrees) as implemented in PAUP* 4.0b10 (Swofford, 2002)were calculated for banding patterns of each primer and fora combined dataset for all three primers (151 characters for36 samples). Robustness of trees was tested by bootstrap-ping (2000 replicates) under the 50% consensus criterion.
Results
MtDNA sequence variation and phylogeny
Monophyly of the Testudo graeca complex wasconfirmed by high posterior probability andbootstrap values under all tree building meth-ods; its sister taxa are T. kleinmanni and T. mar-ginata (fig. 2). All phylogenetic analyses re-sulted in six well-supported major clades (A-F)within the T. graeca complex, and the MP topol-ogy obtained was corroborated by all maxtreessettings.
Average uncorrected p distances between theT. graeca complex and other Testudo speciesrange between 8.870% and 12.662%. Withinthe T. graeca complex, a mean of 3.346% wasobserved. This value clearly exceeds sequencedivergences within other Testudo species, andwithin-divergences of clades B and E corre-spond approximately to that within the poly-typic T. hermanni (table 3).
Within the T. graeca complex, clade A com-prises haplotypes distributed in the East Cauca-sus. It is sister to another clade (B) that corre-sponds to haplotypes from the western Mediter-ranean islands of Mallorca, Sardinia and Sicily(which are thought to be inhabited by allochtho-nous tortoise populations; Buskirk et al., 2001)and from North Africa. Clade C includes haplo-
types from western Asia Minor, the southeast-ern Balkans and the western and central Cauca-sus Region. Its sister group is a fourth, widelydistributed clade (D) from southern and east-ern Asia Minor and the Levantine Region (NearEast). Two further clades (E, F) are distributedin Iran (one clade in northwestern and centralIran; the other in eastern Iran).
While the distinctness of each of these sixclades and the sister group relationships be-tween (A + B) and between (C + D) were un-equivocally supported, the phylogeny of the re-sulting four clades (A + B), (C + D), E and Fwas poorly resolved (small cladograms in fig.2). Under Bayesian analysis, there was weaksupport for a sister group relationship of the twomajor western clades, i.e. ((A + B) + (C + D)),whereas MP and NJ resulted in a sister group re-lation between a weakly supported clade ((A +B) + E, F) and (C + D).
Nuclear genomic fingerprinting
In contrast to obvious geographical sequencedifferentiation of the mitochondrial genome, wefound a quite homogenous pattern in nuclear ge-
Figure 2. Bayesian phylogram of mtDNA haplotypes forspur-thighed tortoises (Testudo graeca complex), rootedwith all other Testudo species. Taxon names within T.graeca complex are assigned according to morphologyand geographic provenance and follow Chkhikvadze andTuniyev (1986), Weissinger (1987), Chkhikvadze andBakradze (1991, 2002), Perälä (2002a, b, 2004b), Pieh etal. (2002) and Pieh and Perälä (2002, 2004). Questionmarks denote uncertain taxonomic allocations due to local-ities along range borders. Locality names (Mallorca, Sar-dinia, Sicily) instead of taxon names indicate introducedor allochthonous tortoises from the western Mediterranean.Numbers following taxon or locality names refer to table 2,where exact localities are given. Symbols correspond to dis-tinct mtDNA lineages and to figure 1. Branch lengths areBayesian estimates and proportional to the scale bar withthe unit being a mean number of nucleotide changes persite. Numbers at crucial nodes represent posterior proba-bilities and MP and NJ bootstrap values (1000 replicates)greater than 50; NJ using ML distances. Dashes indicatethat the respective branch was not supported. Small clado-grams show alternative topologies for major groups withinT. graeca complex. For clade D from southern and easternAsia Minor and the Levant, see also figure 7.
108 U. Fritz et al.
Testudo graeca complex 109
Tabl
e3.
Unc
orre
cted
pdi
stan
ces
(per
cent
ages
)w
ithin
and
betw
een
clad
esof
the
Test
udo
grae
caco
mpl
exan
dot
her
Test
udo
spec
ies,
base
don
ada
tase
tof
1000
bp(m
tDN
A,c
ytoc
hrom
eb
gene
).M
ean
dist
ance
sbe
twee
nsp
ecie
sor
clad
esar
egi
ven
belo
w,r
ange
s(i
nita
lics)
abov
eth
edi
agon
al.T
hew
ithin
-tax
onse
quen
cedi
verg
ence
isgi
ven
inbo
ldon
the
diag
onal
(mea
nan
dra
nge)
.Pol
ytyp
icsp
ecie
sas
teri
sked
.Seq
uenc
esfr
omFr
itzet
al.(
2005
a):T
.her
man
nibo
ettg
eriA
J888
357-
60,T
.her
man
nihe
rman
niA
J888
361-
64,T
.hor
sfiel
diik
azac
hsta
nica
AJ8
8836
5,T.
hors
field
iiru
stam
oviA
J888
366,
T.kl
einm
anni
AJ8
8837
0-71
,T.m
argi
nata
AJ8
8830
8-40
.
nTe
stud
ogr
aeca
*–
all
Cla
deA
Cla
deB
Cla
deC
Cla
deD
Cla
deE
Cla
deF
Test
udo
herm
anni
*Te
stud
oho
rsfie
ldii
*Te
stud
okl
einm
anni
Test
udo
mar
gina
ta
Test
udo
grae
ca*
–al
l
943.
346
(0.0
00-7
.953
)–
––
––
–10
.562
-15.
087
9.84
5-14
.609
8.05
7-10
.928
7.25
6-10
.665
Cla
deA
14–
0.30
4(0
.000
-1.1
48)
3.90
8-6.
149
4.20
8-5.
423
4.00
8-5.
834
3.70
7-5.
732
3.30
7-4.
020
12.0
94-1
4.05
011
.186
-13.
060
10.0
20-1
0.55
18.
868-
10.2
56
Cla
deB
7–
4.66
41.
515
(0.0
00-2
.440
)5.
206-
7.95
34.
868-
7.42
94.
709-
7.44
73.
707-
4.99
011
.208
-13.
314
11.0
25-1
3.60
88.
994-
10.4
548.
870-
10.3
82
Cla
deC
20–
4.67
75.
990
0.24
3(0
.000
-0.9
21)
2.09
2-4.
081
3.97
4-6.
212
3.70
7-4.
509
12.0
20-1
5.08
710
.587
-14.
609
9.19
7-10
.928
7.65
8-9.
844
Cla
deD
37–
4.48
65.
591
2.61
20.
789
(0.0
00-2
.509
)3.
607-
6.93
93.
009-
4.88
311
.049
-13.
692
9.84
5-12
.544
8.05
7-9.
703
7.28
8-9.
514
Cla
deE
9–
4.27
45.
419
4.85
84.
409
1.34
3(0
.000
-3.0
46)
2.50
5-4.
309
11.2
28-1
4.02
810
.374
-13.
916
8.91
8-10
.922
7.89
4-10
.665
Cla
deF
7–
3.54
54.
061
3.95
43.
510
3.08
10.
150
(0.0
00-0
.401
)10
.562
-12.
224
10.0
90-1
2.31
48.
460-
9.21
87.
256-
8.71
7
Test
udo
herm
anni
*8
12.6
6213
.203
12.4
0413
.154
12.4
5312
.605
11.6
151.
183
(0.0
00-2
.752
)9.
931-
13.0
5710
.673
-11.
907
9.87
1-11
.864
Test
udo
hors
field
ii*
211
.737
11.9
9412
.328
12.3
0811
.243
11.9
3011
.420
11.6
222.
461
(–)
9.99
0-11
.333
10.0
72-1
1.78
8
Test
udo
klei
nman
ni2
9.41
010
.184
9.60
19.
872
8.82
89.
746
8.99
511
.454
10.6
620.
000
(–)
6.45
2-7.
301
Test
udo
mar
gina
ta33
8.87
09.
433
9.70
39.
207
8.48
18.
973
7.88
411
.086
10.9
496.
847
0.22
0(0
.000
-1.1
01)
110 U. Fritz et al.
nomic fingerprints. All studied samples shared ahigh percentage of bands. The primer (GACA)4
resulted in average distances of 0.30 (Dice; Neiand Li) and 0.44 (Jaccard), (GAA)5 in 0.35(Dice; Nei and Li) and 0.50 (Jaccard), and L18,which produced the most variable fingerprints,in average distances of 0.23 (Dice; Nei and Li)and 0.37 (Jaccard).
Using NJ cluster analyses, the datasets forthe three single primers resulted in unresolvedpolytomies for most samples under the 50%bootstrap consensus criterion, and those sam-ples that clustered with weak to high bootstrapsupport often belong to different mtDNA lin-eages and different taxa. The NJ tree of the(GAA)5 dataset serves as example (fig. 3, left).The combined dataset of 151 characters fromthe three primers resulted in fewer polytomies(fig. 3, right); however, also in this tree none ofthe major clusters corresponds with taxon limitsor mtDNA lineages.
Systematics, zoogeography and discussion
Incongruence of morphology, taxa and geneticdata
Populations of spur-thighed tortoises comprisein some regions (Spain, North Africa, Levant)small to medium-sized, mainly yellow-colouredtortoises, while in other parts of the range largersized, dark individuals occur. However, onto-genetic variation of shell coloration and pro-portions is considerable. Different age classesof Bulgarian Testudo graeca display nearly theentire range of shell coloration, with young,small adults being light-coloured and aged,large adults being dark-coloured. Also shellproportions are known to change much duringgrowth (Fritz et al., 1996).
Nevertheless, such characters (shell shapeand proportions, colour pattern, in part alsoscutellation characters) were repeatedly usedfor taxon delineation, leading to recent de-scriptions or resurrections of several morpho-logically diagnosable taxa (Pieh, 2001; Perälä,
2002a; Pieh and Perälä, 2002, 2004). For Mo-roccan T. graeca it was demonstrated, how-ever, that such variation corresponds rather topopulation-specific differentiation than to tax-onomic distinctness (Harris et al., 2003; Car-retero et al., 2005).
Using ISSR fingerprinting we found no sup-port for any taxonomic differentiation and noneof our six mtDNA clades within the T. graecacomplex is congruent with any morphologi-cally defined taxon. Also a distinction betweena North African-Levantine lineage, comprisingmainly yellow-coloured tortoises of small tomoderate size and a second lineage of larger anddarker tortoises, occurring in Southeast Europeand the Near and Middle East, as suggested byFritz et al. (1996), is not corroborated by ourdata (figs 1-2). While all studied North Africantortoises fall into one clade (B), the sister groupof this clade is not formed by Levantine haplo-types but by haplotypes from the East Caucasus(clade A), representing populations with largeand dark tortoises.
According to our mtDNA data, this East Cau-casian clade A is composed of haplotypes fromtortoises that belong to T. g. armeniaca, T. g. ib-era and T. g. pallasi. The North African cladeB contains haplotypes of the nominal taxa T.g. graeca, T. g. cyrenaica and T. g. nabeulen-sis as well as most specimens from the westernMediterranean islands. Clade C, found in theBalkans, in western Asia Minor, and in the west-ern and central Caucasus, comprises sequencesrepresenting T. g. ibera and T. g. nikolskii. CladeD from southern and eastern Asia Minor and theLevant is composed of haplotypes of T. g. ana-murensis, T. g. antakyensis, T. g. floweri, T. g.ibera, T. g. terrestris and one (allochthonous)tortoise from Sicily. Clade E from northwest-ern and central Iran consists of haplotypes fromT. g. buxtoni and T. g. perses, and the easternIranian clade F contains haplotypes assigned toT. g. perses and T. g. zarudnyi.
From a morphological point of view, cladesA, C, E and F contain populations with dark-coloured tortoises of medium to large size.
Testudo graeca complex 111
Figure 3. Unrooted NJ bootstrap 50% majority-rule consensus trees of ISSR fingerprints for samples of the Testudograeca complex. Left, primer (GAA)5, based on 50 samples and 33 fragments. Right, combined dataset of primers(GAA)5, (GACA)4 and L18, based on 36 samples and 151 fragments. Numbers along branches represent bootstrap values(2000 replicates) greater than 50. For further explanations see figure 2.
Taxa in these clades differ mainly in shellshape and in part in colour pattern. Tortoisesof clade B are generally lighter coloured andsmall- to moderately-sized (table 1). Most vari-ation is found in clade D that comprises small-to medium-sized, mainly yellow-coloured tor-toises (T. g. antakyensis, T. g. floweri, T. g.terrestris) as well as huge, dark brownishindividuals, either with flat, elongated shells (T.g. anamurensis) or distinctly domed shells (T. g.ibera from Lake Van Region, Turkey).
However, even within populations attributedto T. g. antakyensis by Perälä (2002a) consider-able morphological differences are known to oc-cur, reflecting the same general pattern. WithinSyria, morphological variation was shown tofollow Gloger’s Rule in that tortoises fromarid regions are distinctly lighter coloured (andsmaller) than tortoises from more humid re-
gions. Further, tortoises living on dark basalticsoil, as in the Jabal Duruz Region, are signifi-cantly darker coloured than tortoises from otherregions with light substrate (figs 4-6; Fritz etal., 1996). We observed the same pattern also inother regions of the Near East during field work.This suggests stabilizing selection by environ-mental pressure as a source of morphologicalvariation. Similarity of general body and sub-strate coloration is a well-known phenomenonalso in other animal species (‘substrate races’;Mayr, 1963).
Haplotypes within clade D, that containsamong others all T. g. antakyensis sequences,group into three weakly to well-supported sub-clades. These subclades rather correspond tosequences from neighbouring regions than totaxonomic segregation or morphological sim-ilarity (fig. 7). This is most obvious in the
112 U. Fritz et al.
well-supported subclade from Israel, Jordan and
southern Syria, which consists of sequences
from the nominal taxa T. g. antakyensis and T.
g. floweri. While the sequences of the four T. g.
floweri and a T. g. antakyensis from Jordan (an-
takyensis 4) represent small-sized, mainly yel-
low tortoises from arid regions, other sequences
from T. g. antakyensis originate from larger and
darker tortoises that live in moister environ-
ments (antakyensis 1-3: Tiberias, Israel). The
three Syrian T. g. antakyensis sequences in this
subclade (antakyensis 10-12) are from dark-
coloured tortoises from the basaltic Jabal Duruz
Region.
Testudo graeca armeniaca is a further promi-
nent example for incongruence of morpholo-
gical and genetic differentiation. This taxon is
characterized by a depressed shell and flat-
tened forearms. In general appearance it resem-
bles T. horsfieldii Gray, 1844 (Chkhikvadze and
Bakradze, 1991; Pieh et al., 2002), a species that
is not closely related to the T. graeca complex
and allocated to the genus Agrionemys by some
authors (Fritz and Cheylan, 2001; Parham et al.,
2006). The mtDNA haplotype of a T. g. arme-
niaca is nested within the East Caucasian clade
A together with other, morphologically distinc-
tive tortoises with domed shell from the same
region (fig. 2; table 2). Both T. g. armeniaca
and T. horsfieldii occur in regions with strictly
continental climate with severe, cold winters.
Farther, both are steppicolous tortoises that dig
and inhabit deep burrows, suggesting that envi-
ronmental pressure and the similar mode of life
results via selection in similar phenotypes and
masks phylogenetic relationships.
Also in another Testudo species, T. mar-
ginata, environmental pressure was shown to re-
sult in considerable phenotypic variation (size
reduction, persisting juvenile characteristics in
a poor region of the range; Fritz et al., 2005a),
suggestive of considerable phenotypic plasticity
in Testudo species.
How many species?
Based on morphology and zoogeographic con-siderations, it was repeatedly suggested thatTestudo graeca is composed of several distinctspecies (Bour, 1989; Gmira, 1993a, b, 1995;David, 1994; Pieh, 2001; Perälä, 2002a, b; Piehand Perälä, 2002, 2004). In contrast to Gmira(1993a, b, 1995) and Perälä (2002b), whosemorphological datasets argued for the para-phyly of the T. graeca complex with respectto other Testudo species (T. kleinmanni sensulato, T. marginata), our mtDNA data confirmeda monophyletic T. graeca complex and its sis-ter group relationship to T. kleinmanni and T.marginata. This agrees with other recent stud-ies employing genetic markers (mtDNA data:van der Kuyl et al., 2002; Fritz et al., 2005a;Parham et al., 2006; nuclear genomic finger-printing: Fritz et al., 2005a).
Obviously, species delineation depends onthe adopted species concept, and there is an on-going debate about the benefits and shortcom-ings of different competing species concepts(e.g. Ereshefsky, 1992; Wheeler and Meier,2000; Agapow et al., 2004; Coyne and Orr,2004). There is no doubt that the six mtDNAlineages within the T. graeca complex couldbe recognized as representing full species un-der the Evolutionary (e.g. Wiley and May-den, 2000) or Phylogenetic Species Concepts(e.g. Cracraft, 1983, 1987; Mishler and The-riot, 2000; Wheeler and Platnick, 2000; seealso the review in Coyne and Orr, 2004), al-though morphological characterization of thesespecies would be difficult (see below). How-ever, the most restrictive species concept forbisexual animals, the Biological Species Con-cept of Mayr (e.g. 1942, 1963, 2000), de-fines species as reproductively (and thus ge-netically) isolated groups of populations. Un-der this species concept, mitochondrial genomicmarkers alone are not conclusive, mainly dueto their matrilineal inheritance and the differ-ent effective population sizes of mitochondrialand nuclear genomes, resulting in different pat-terns for incomplete lineage sorting and intro-
Testudo graeca complex 113
gression (Funk and Omland, 2003; Ballard andWhitlock, 2004).
ISSR PCR provides an ideal tool for obtain-ing information about nuclear genomic differ-entiation and gene flow. In distinct ‘biologicalspecies’, differentiation of nuclear and mito-chondrial genomes is typically congruent andISSR profiles differ clearly (e.g. for cheloni-ans: Wink et al., 2001; Guicking et al., 2002;Schilde et al., 2004; Fritz et al., 2005a, b). Insuch cases, introgression and hybridization areindicated by shared bands in otherwise distinctbanding patterns. In contrast, if populations stillform a genetic continuum, ISSR profiles do notdiffer markedly. ISSR also allows identificationof taxa in which introgression of mtDNA mightmask nuclear genomic differentiation (Gupta etal., 1994; Zietkiewicz et al., 1994; Wink et al.,1998, 2001; Wolfe and Liston, 1998; Wolfe etal., 1998; Guicking et al., 2002; Nagy et al.,2003; Schilde et al., 2004; Fritz et al., 2005a, b).
In a previous study, ISSR profiles of all fivetraditionally recognized Testudo species wereshown to be significantly distinct (in NJ clus-ter analysis bootstrap support values of 100 foreach species; Fritz et al., 2005a). Using the sameprimers, (GAA)5 and (GACA)4, plus one ad-ditional primer (L18) shown to be useful inother chelonians (Emys; Fritz et al., 2005b), wefound little differentiation within the T. graecacomplex, despite extensive mtDNA differenti-ation. As we can exclude homoplasy due tothe excellent discrimination of the other Testudospecies, the similarity of fingerprints within theT. graeca complex indicates either a high levelof gene flow or a low level of differentiation(including incomplete lineage sorting). We con-clude that T. graeca represents a single poly-typic species under the Biological Species Con-cept.
How many subspecies?
Based on a combined data set of 393-394 bp ofthe 12S rRNA gene and 411 bp of the D-loop,van der Kuyl et al. (2005) found among 22 spur-thighed tortoises only two well-supported hap-
lotype clades, corresponding to tortoises from(i) North Africa and (ii) Turkey (one sample)and the Near East and presumed that theserepresent only two subspecies, Testudo graecagraeca and T. g. ibera, while genetic evi-dence for the existence of other subspecies wasthought to be weak. According to locality data,the ‘ibera clade’ of van der Kuyl et al. (2005)could be identical with our clade D. However,a problem of the papers by van der Kuyl etal. (2002, 2005) was that many tortoises of un-known provenance were studied. Such speci-mens cannot be used when assessing geographicvariation (Harris et al., 2003). Moreover, sub-species names assigned by van der Kuyl et al.(2002, 2005) often did not fulfil nomenclaturalrequirements and locality data in part did notmatch the ranges of the respective taxa (Perälä,2004a). For example, van der Kuyl et al. (2005)assign tortoises from Bulgaria in part to T. g.terrestris; this taxon does not occur in Bulgaria(Bour and Perälä, 2004; Perälä and Bour, 2004;see also table 1).
In accordance with the Biological SpeciesConcept, we understand subspecies as geneti-cally distinct, geographically vicariant groupsof populations between which gene flow occurs.Distinct mitochondrial matrilineages often, butnot always, mirror such subspecific differenti-ation (Avise, 2000; but see Funk and Omland,2003).
Using the mitochondrial cytochrome b gene,we discovered distinctly more variation thanprevious studies based on the 12S rRNA gene(van der Kuyl et al., 2002, 2005; Harris et al.,2003). This is in line with other investigationsthat found in testudinids the cytochrome b geneto be more variable than the 12S rRNA gene(Caccone et al., 1999; Palkovacs et al., 2002;Fritz et al., 2005a, 2006). Neither the four sub-species that were originally thought to composeT. graeca (T. g. graeca from Spain and north-western Africa; T. g. terrestris from northeast-ern Africa, southeastern Turkey and the Lev-antine Region; T. g. ibera from Southeast Eu-rope, Asia Minor, the Caucasus Region, west-
114 U. Fritz et al.
Figure 4. Light coloured spur-thighed tortoise from arid region in Syria (Saydnaya, Anti Lebanon Mts.; maximum weight990 g, n = 17). Note healed serious injury on top of carapace.
Figure 5. Dark coloured spur-thighed tortoise from more humid environment in Syria (Qalat Saman, hills NW Aleppo,maximum weight 1900 g, n = 15).
Testudo graeca complex 115
Figure 6. Dark coloured spur-thighed tortoise from basaltic region in Syria (Jabal Duruz, maximum weight 1320 g, n = 28).
ern and central Iran; T. g. zarudnyi from easternIran; Mertens, 1946; Wermuth, 1958; Wermuthand Mertens, 1961, 1977), nor the many taxathat were later described or resurrected (table1) correspond perfectly to any of our six ma-jor mtDNA clades. This could be indicative ofinaccurate taxon limits.
Several authors agree that T. g. ibera, as cur-rently understood, might be a conglomerate ofdistinct taxa, and that the name ibera shouldbe used only for tortoises from the CaucasianKura River Basin, the area around the type lo-cality (Perälä, 2002a; Pieh et al., 2002; Danilovand Milto, 2004). However, the disagreementof subspecies delineation and mtDNA cladeswould not be resolved by using a restricted sub-species concept for T. g. ibera (table 1). Evenour sequences of T. g. ibera sensu stricto fromthe Kura River Basin are scattered over twoclades (A, C). The sequences from central Cau-casian ibera localities, one from Mtskheta only15 km away from the type locality of ibera(Tbilisi, Georgia), belong to clade C, which is
otherwise found along the northeastern BlackSea coast (=T. g. nikolskii), in the Balkans andwestern Asia Minor (=T. g. ibera sensu lato),while the ibera sequences from the more east-ern Caucasian locality Dashburun (Azerbaijan)appear in the same clade (A) as T. g. pallasiand the morphologically very distinctive T. g.armeniaca, so that a taxonomic break must beassumed within the Kura River Basin.
Until now, taxa within the T. graeca complexwere defined exclusively by morphology. Ourdata imply that genetic differentiation is maskedby pronounced morphological plasticity. Thisresults in the unexpected nesting of morpholog-ically well-defined taxa, like T. g. armeniaca orT. g. floweri, within clades comprising also ge-ographically neighbouring, but morphologicallydistinctive populations of other taxa (clade A: T.g. armeniaca, T. g. ibera, T. g. pallasi; clade D:T. g. anamurensis, T. g. antakyensis, T. g. flow-eri, T. g. ibera, T. g. terrestris), or the scatteringof sequences of tortoises of the same subspecies(T. g. ibera sensu stricto or T. g. ibera sensu lato)
116 U. Fritz et al.
Figure 7. Detail of Bayesian tree of mtDNA haplotypes for spur-thighed tortoises from southern and eastern Asia Minor andthe Levant. For further explanations see figure 2.
over two or three genetically distinct clades (A,C or A, C, D, respectively; fig. 2). Future re-search is needed to find out whether any mor-phological differences exist paralleling the sixmtDNA clades.
In any case, it is obvious that the sug-gested existence of up to 20 distinct taxa withinthe T. graeca complex (Highfield and Mar-tin, 1989a, b, c; Highfield, 1990; Pieh, 2001;Perälä, 2002a, b; Pieh and Perälä, 2002, 2004)is exaggerated and does not match genetic dif-ferentiation. To achieve a more realistic taxo-nomic arrangement reflecting mtDNA clades,we propose usage of the taxonomic arrange-ment in table 4, which is also most parsimo-nious with respect to nomenclatural changesof the formerly widely accepted subspecies de-lineation of Mertens (1946), Wermuth (1958)and Wermuth and Mertens (1961, 1977). Asnot all North African taxa were included inthe present study, we refrain from synonymiz-
ing North African taxa with T. g. graeca Lin-naeus, 1758 (mtDNA clade B) that representsa further valid subspecies. We admit that oursubspecies delineation may be imprecise alongrange borders due to introgression of mitochon-drial genomes.
Historic zoogeography
The distribution of the six major mtDNA cladeswithin the Testudo graeca complex allows somezoogeographical insights. Like in many otherorganisms (Myers et al., 2000), a remarkablyhigh diversity is found in the Caucasus Region,where four of the six mtDNA clades occur (A,C, D, E; fig. 1). The high diversity of the Cau-casus Region is often explained by an admixtureof old endemics and taxa of different geographicorigin (e.g. Satunin, 1910; Darevskii, 1967; Tu-niyev, 1990, 1995), highlighting the role of thismountain range as centre of endemism, refugearea and crossroads between the Mediterranean,
Testudo graeca complex 117
Table 4. Proposed new subspecies delineation for the eastern part of the range of Testudo graeca acknowledging mtDNAclades. For most type localities see table 1. Type localities of taxa that have not been regarded as valid in past decadesare: Testudo ibera racovitzai Calinescu, 1931 – Tutrakan, Bulgaria; Testudo ibera var. bicaudalis Venzmer, 1920 – TaurusMts., Cilicia, Turkey (Buskirk et al., 2001). We refrain from assigning the doubtful name Testudo ecaudata Pallas, 1814(see Buskirk et al., 2001; Pieh et al., 2002) to clade E although its type locality (forested parts of Persia at the Caspian Sea)suggests identity with T. g. buxtoni.
MtDNAclade
Subspecies Junior synonyms Range
A Testudo graeca armeniacaChkhikvadze andBakradze, 1991
Testudo graeca pallasi Chkhikvadzeand Bakradze, 2002
Central west coast of Caspian Sea; easternCaucasus Region and parts of central Cau-casus Region, including Araks River Valley(Armenia, Turkey)
C Testudo graeca iberaPallas, 1814
Testudo ibera racovitzai Calinescu,1931Testudo graeca nikolskiiChkhikvadze and Tuniyev, 1986
Southeast Europe, western Asia Minor,Russian and Georgian Black Sea coast,central Caucasus Region (Georgia, adjacentAzerbaijan)
D Testudo graeca terrestrisForsskål, 1775
Testudo ibera var. bicaudalisVenzmer, 1920
Southern and eastern Asia Minor, Levant
Testudo floweri Bodenheimer, 1935Testudo graeca anamurensisWeissinger, 1987Testudo antakyensis Perälä, 1996
E Testudo graeca buxtoniBoulenger, 1921
Testudo perses Perälä, 2002 Northwestern and central Iran, in the east-ern Caucasus Region probably also adjacentparts of other countries
F Testudo graeca zarudnyiNikolsky, 1896
– East Iran
Europe, and Central Asia. The complicated geo-logical history of the Caucasus Region and adja-cent East Anatolia includes repeated Oligoceneand Miocene episodes of isolation during ma-rine transgressions and reconnections with Eu-ropean and Asiatic landmasses, in part open-ing corridors to Africa via Arabia (Rögl, 1998,1999), resulting in several vicariance events forbiota.
Not only the current high diversity in the Cau-casus Region, but also the fossil record suggestsa Caucasian origin for the Testudo graeca com-plex. Fossils referred to the extinct species T.burtschaki and T. eldarica from the Medial toUpper Miocene of Azerbaijan and eastern Geor-gia (Chkhikvadze, 1983, 1989; Danilov, 2005)are the oldest representatives of this group. Inthe Pliocene, tortoises resembling T. graecawere already widely distributed, with recordsin the central and northern Caucasus Region,Ukraine, Moldavia, Romania (T. cernovi, T. ku-curganica; Chkhikvadze, 1983, 1989; Danilov,2005) and Morocco (T . aff. kenitrensis; de Lap-
parent de Broin, 2000), indicating rapid disper-sal.
The fossil record also provides evidence thatthe current patchy distribution of clade C (fig.1) is a consequence of secondary range inter-ruption. Besides the above-mentioned Pliocenefossils, there are many Pleistocene tortoise find-ings from the northern Caucasus and Black SeaRegions that are in part referred to T. graecaalready (Chkhikvadze, 1983, 1989). Thus, thecurrent gaps in the range of clade C tortoises aremost probably the result of glacial extinction.
All tree building methods favour a sistergroup relationship of the East Caucasian andNorth African clades A and B, although both arefully allopatric today and separated by a largedistribution gap (figs 1-2). From Central andWest Europe there are no tortoise fossils knownresembling unambiguously the T. graeca com-plex, despite an excellent fossil record (Bailón,2001; Buskirk et al., 2001; de Lapparent deBroin, 2001). This suggests that the founder in-dividuals of the current Spanish (and Italian)
118 U. Fritz et al.
populations were introduced. If the East Cau-casian and the North African clades A and B aresister groups indeed and an invasion of NorthAfrica via West Europe is to be excluded, a col-onization route via the Levantine Region mustbe hypothesized. However, along the Levant an-other clade (D) is distributed today, so that atleast two colonization waves must have reachedthis region.
A first dispersal event from the Caucasus Re-gion to North Africa could have occurred inthe Middle to Late Miocene, when intermittentland bridges connected North Africa with theNear and Middle East due to the northeastwardmovement of the Arabian Plate, leading laterto complete closure of the Tethys Sea (Rögl,1998, 1999). An invasion of North Africa viasuch a temporary land bridge is an attractivescenario, explaining later extermination and re-placement of geographically connecting popu-lations through clade D, now distributed in theLevantine Region (fig. 1).
Our investigation confirms that most of theintroduced spur-thighed tortoise populations onwestern Mediterranean islands originated inNorth Africa, as suggested by the previous allo-cation of these populations to the subspecies T.g. graeca (Wermuth and Mertens, 1961, 1977;Buskirk et al., 2001). However, the record of aSicilian tortoise belonging to the eastern cladeD instead of the North African clade B providesevidence for multiple introductions from differ-ent regions, and that some individuals definitelydid not come from North Africa.
Acknowledgements. F. Ahmadzadeh, H. Bringsøe, M.Cheylan, S. d’Angelo, D. Frynta, E. Geffen, M. Jirku, M.Kamler, P. Mikulícek, D. Modrý, O. Paknia, E. Roitberg,W. Wegehaupt and B.H. Zarei assisted in obtaining bloodsamples or provided samples. Most lab work was done byA. Müller (Museum of Zoology, Dresden) and H. Sauer-Gürth (IPMB, Heidelberg). The work of P. Široký profitedfrom the IGA VFU grant (2/2004/FVHE) and O. Türkozanwas supported by TÜBITAK project TBAG-2206 102T104.Thanks for comments on this paper go to O. Bininda-Emonds, S.D. Busack, D.J. Harris, M. Vences and oneanonymous reviewer.
References
Agapow, P.-M., Bininda-Emonds, O.R.P., Crandall, K.A.,Gittleman, J.L., Mace, G.M., Marshall, J.C., Purvis, A.(2004): The impact of species concept on biodiversitystudies. Quart. Rev. Biol. 79: 161-179.
Álvarez, Y., Mateo, J.A., Andreu, A.C., Díaz-Paniagua, C.,Diez, A., Bautista, J.M. (2000): Mitochondrial DNAhaplotyping of Testudo graeca on both continental sidesof the Straits of Gibraltar. J. Heredity 91: 39-41.
Anderson, S.C. (1979): Synopsis of the turtles, crocodiles,and amphisbaenians of Iran. Proc. California Acad. Sci.,4th Ser. 41 (22): 501-528.
Armstrong, J., Gibbs, A., Peakall, R., Weiller, G. (1996):RAPDistance 1.04. Canberra, Australian National Uni-versity.
Austin, J.J., Arnold, E.N., Bour, R. (2003): Was there a sec-ond adaptive radiation of giant tortoises in the IndianOcean? Using mitochondrial DNA to investigate specia-tion and biogeography of Aldabrachelys. Mol. Ecol. 12:1415-1424.
Avise, J.C. (2000): Phylogeography. The History and For-mation of Species. Cambridge, Massachusetts, and Lon-don, Harvard University Press.
Bailón, S. (2001): Revisión de la asignación a Testudo cf.graeca del yacimiento del Pleistoceno superior de CuevaHorá (Darro, España). Rev. Esp. Herpetol. 15: 61-65.
Ballard, J.W.O., Whitlock, M.C. (2004): The incompletenatural history of mitochondria. Mol. Ecol. 13: 729-744.
Bornet, B., Branchard, M. (2001): Nonanchored inter sim-ple sequence repeat (ISSR) markers: reproducible andspecific tools for genome fingerprinting. Plant Mol. Biol.Rep. 19: 209-215.
Bour, R. (1987): L’identité des tortues terrestres eu-ropéennes: spécimens-types et localités-types. Rev. Fr.Aquariol. 13 (1986): 111-122.
Bour, R. (1989): Caractères diagnostiques offerts par lecrâne des tortues terrestres du genre Testudo. Mésogée48 (1988): 13-19.
Bour, R., Perälä, J. (2004): Testudo [graeca] terrestrisForsskål, 1775. In: Numéro spécial Testudo. Guyot-Jackson, G., Ed., Manouria 7 (22): 40-41.
Buskirk, J.R., Keller, C., Andreu, A.C. (2001): Testudograeca Linnaeus, 1758 – Maurische Landschildkröte. In:Handbuch der Reptilien und Amphibien Europas. Band3/IIIA: Schildkröten (Testudines) I, p. 125-178. Fritz,U., Ed., Wiebelsheim, Aula-Verlag.
Caccone, A., Amato, G., Gratry, O.C., Behler, J., Powell,J.R. (1999): A molecular phylogeny of four endangeredMadagascar tortoises based on mtDNA sequences. Mol.Phylogenet. Evol. 12: 1-9.
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., Bakradze, M.A. (1991): O sistematich-eskom polozhenii sovremennoi sukhoputnoi cherepakhiiz doliny reki Araks. Trudy Tbilissk. gosudarstven. Univ.305: 59-63.
Chkhikvadze, V.M., Bakradze, M.A. (2002): Novyi podvidsukhoputnoi cherepakhi iz Dagestana. Trudy Inst. Zool.,Akad. Nauk Gruzii 21: 276-279.
Chkhikvadze, V.M., Tuniyev, B.S. (1986): O sistemati-cheskom polozhenii sovremennoi sukhoputnoi chere-pakhi zapadnogo Zakavkazya. Soobshch. Akad. NaukGruzinsk. SSR 124: 617-620.
Condit, R., Hubbell, S.P. (1991): Abundance and DNAsequence of two-base repeat regions in tropical treegenomes. Genome 34: 66-71.
Cracraft, J. (1983): Species concepts and speciation analy-sis. Curr. Ornithol. 1: 159-187.
Cracraft, J. (1987): Species concepts and the ontology ofevolution. Biol. Phil. 2: 329-346.
Danilov, I.G. (2005): Die fossilen Schildkröten Europas. In:Handbuch der Reptilien und Amphibien Europas. Band3/IIIB: Schildkröten (Testudines) II, p. 329-441. Fritz,U., Ed., Wiebelsheim, Aula-Verlag.
David, P. (1994): Liste des reptiles actuels du monde. I.Chelonii. Dumerilia 1: 7-127.
de Lapparent de Broin, F. (2000): African chelonians fromthe Jurassic to the present: phases of development andpreliminary catalogue of the fossil record. Palaeontolo-gia Africana 36: 43-82.
de Lapparent de Broin, F. (2001): The European turtle faunafrom the Triassic to the present. Dumerilia 4: 155-217.
Dice, L.R. (1945): Measures of the amount of ecologicassociation between species. Ecology 26: 297-302.
Ereshefsky, M. (1992): The Units of Evolution. Essays onthe Nature of Species. Cambridge, Massachusetts, andLondon, Bradford Book, MIT Press.
Ernst, C.H., Barbour, R.W. (1989): Turtles of the World.Washington, DC, Smithsonian Institution Press.
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.
Felsenstein, J. (1985): Confidence limits on phylogenies: anapproach using the bootstrap. Evolution 39: 783-791.
Fritz, U., Bischoff, W., Martens, H., Schmidtler, J.F. (1996):Variabilität syrischer Landschildkröten (Testudo graeca)sowie zur Systematik und Zoogeographie im NahenOsten und in Nordafrika. Herpetofauna 18 (104): 5-14.
Fritz, U., Cheylan, M. (2001): Testudo Linnaeus, 1758 –Eigentliche Landschildkröten. In: Handbuch der Rep-tilien und Amphibien Europas. Band 3/IIIA: Schild-kröten (Testudines) I, p. 113-124. Fritz, U., Ed.,Wiebelsheim, Aula-Verlag.
Fritz, U., Široký, P., Kami, H., Wink, M. (2005a): 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., Fattizzo, T., Guicking, D., Tripepi, S., Pennisi,M.G., Lenk, P., Joger, U., Wink, M. (2005b): A newcryptic species of pond turtle from southern Italy, thehottest spot in the range of the genus Emys. Zool. Scr.34: 351-371.
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.
Gmira, S. (1993a): Une nouvelle espèce de tortue Testu-dininei (Testudo kenitrensis n. sp.) de l’Inter Amirien-Tensiftien de Kénitra (Maroc). C.R. Acad. Sci. Paris 316(II): 701-707.
Gmira, S. (1993b): Nouvelles données sur les espècesactuelles de Testudo. Bull. Soc. Herpétol. Fr. 65/66: 49-56.
Gmira, S. (1995): Étude des chéloniens fossiles du Maroc.Paris, Cahiers de Paléontologie, CNRS Éditions.
Gomes, C., Dales, R.B.G., Oxenford, H.A. (1998): Theapplication of RAPD markers in stock discriminationof the four-wing flyingfish, Hirundichthys affinis in thecentral western Atlantic. Mol. Ecol. 7: 1029-1039.
Guicking, D., Fritz, U., Wink, M., Lehr, E. (2002): Newdata on the diversity of the Southeast Asian leaf turtlegenus Cyclemys Bell, 1834. Molecular results. Faun.Abh. Mus. Tierkd. Dresden 23: 75-86.
Gupta, M., Chyi, Y.-S., Romero-Severson, J., Owen, J.L.(1994): Amplification of DNA markers from evolution-ary diverse genomes using single primers of simple-sequence repeats. Theor. Appl. Genet. 89: 998-1006.
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.
Haig, S.M., Gratto-Trevor, C.L., Mullins, T.D., Colwell,M.A. (1997): Population identification of western hemi-sphere shorebirds throughout the annual cycle. Mol.Ecol. 6: 413-427.
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., 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.
Highfield, A.C. (1990): Tortoises of north Africa; taxonomy,nomenclature, phylogeny and evolution with notes onfield 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 Bennett1836. New light on an old carapace – Gilbert White’sSelborne 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 fromNorth Africa. London, The Tortoise Trust/Tortoise Sur-vival Project.
Hundsdörfer, A.K., Wink, M. (2005): New source of geneticpolymorphisms in Lepidoptera? Z. Naturforsch. 60c:618-624.
Jaccard, P. (1901): Étude comparative de la distributionflorale dans une portion des Alpes et des Jura. Bull. Soc.Vaud. Sci. Nat. 37: 547-579.
Jaccard, P. (1908): Nouvelles recherches sur la distributionflorale. Bull. Soc. Vaud. Sci. Nat. 44: 223-270.
Korsunenko, A., Vasilyev, V., Pereshkolnik, S., Mazanaeva,L., Lapid, R., Bannikova, A., Semyenova, S. (2005):DNA polymorphism and genetic differentiation of Tes-tudo graeca L. In: Herpetologica Petropolitana. Pro-ceedings of the 12th Ordinary General Meeting of theSocietas Europaea Herpetologica, August 12-16, 2003.Ananjeva, N., Tsinenko, O., Eds, Russ. J. Herpetol. 12(Suppl.): 40-42.
Mayr, E. (1942): Systematics and the Origin of Speciesfrom the Viewpoint of a Zoologist. New York, ColumbiaUniversity Press.
Mayr, E. (1963): Animal Species and Evolution. Cam-bridge, Massachusetts, Belknap Press.
Mayr, E. (2000): The Biological Species Concept. In:Species Concepts and Phylogenetic Theory. A Debate,p. 17-29. Wheeler, Q.D., Meier, R., Eds, New York,Columbia University Press.
Mertens, R. (1946): Über einige mediterrane Schildkröten-Rassen. Senckenbergiana 27: 111-118.
Mertens, R., Müller, L. (1928): Liste der Amphibien undReptilien Europas. Abh. Senckenberg. Naturforsch. Ges.41: 1-62.
Mishler, B.D., Theriot, E.C. (2000): The PhylogeneticSpecies Concept (sensu Mishler and Theriot): mono-phyly, apomorphy, and phylogenetic species concepts.
In: Species Concepts and Phylogenetic Theory. A De-bate, p. 44-54. Wheeler, Q.D., Meier, R., Eds, New York,Columbia University Press.
Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fon-seca, G.A.B., Kent, J. (2000): Biodiversity hotspots forconservation priorities. Nature 403: 853-858.
Nagy, Z.T., Joger, U., Guicking, D., Wink, M. (2003):Phylogeography of the European whip snake Coluber(Hierophis) viridiflavus as inferred from nucleotide se-quences of the mitochondrial cytochrome b gene andISSR genomic fingerprinting. Biota 3: 109-118.
Nei, M., Li, W.-H. (1979): Mathematical model for studyinggenetic variation in terms of restriction endonucleases.PNAS 76: 5269-5273.
Palkovacs, E.P., Gerlach, J., Caccone, A. (2002): The evo-lutionary origin of Indian Ocean tortoises (Dipsochelys).Mol. Phylogenet. Evol. 24: 216-227.
Palkovacs, E.P., Marschner, M., Ciofi, C., Gerlach, J., Cac-cone, A. (2003): Are the native giant tortoises from theSeychelles really extinct? A genetic perspective basedon mtDNA and microsatellite data. Mol. Ecol. 12: 1403-1413.
Parham, J.F., Macey, J.R., Papenfuss, T.J., Feldman, C.R.,Türkozan, O., Polymeni, R., Boore, J. (2006): The phy-logeny of Mediterranean tortoises and their close rel-atives based on complete mitochondrial genome se-quences from museum specimens. Mol. Phylogenet.Evol. 38: 50-64.
Perälä, J. (1996): Etelä-Turkin maakilpikonnista. Morfolo-gisia ja ekologisia eroja (Testudo ibera Pallas 1814 &Testudo ibera anamurensis Weissinger 1987) sekä uudenmaakilpikonnalajin kuvaus + 15 kuvaa. In: Virallinenkongressijulkaisu, p. 14-26. Herpetokongressi I. Perälä,J., Vikberg, J., Kanza, M., Eds, Helsinki, Suomen her-petologinen yhdistys ry.
Perälä, J. (2002a): Morphological variation among MiddleEastern Testudo graeca L., 1758 (sensu lato), with afocus on taxonomy. Chelonii 3: 78-108.
Perälä, J. (2002b): The genus Testudo (Testudines: Testu-dinidae): phylogenetic inferences. Chelonii 3: 32-39.
Perälä, J. (2004a): Tortoise systematics: a critique of arecent paper by van der Kuyl et al. (2002). Herpetol. J.14: 51-53.
Pieh, A. (2001): Testudo graeca soussensis, eine neue Un-terart der Maurischen Landschildkröte aus dem Sousstal(Nordwest-Marokko). Salamandra 36 (2000): 209-222.
Pieh, A., Fritz, U., Berglas, R. (2002): New data onmorphology, distribution and nomenclature of Testudograeca armeniaca Chkhikvadze & Bakradze, 1991.Faun. Abh. Mus. Tierkd. Dresden 22: 329-345.
Pieh, A., Perälä, J. (2001): Eine ungewöhnliche Landschild-kröte des Testudo graeca-Komplexes aus Krasnowodsk(Turkmenien). Herpetozoa 14: 65-73.
Testudo graeca complex 121
Pieh, A., Perälä, J. (2002): Variabilität von Testudo graecaLinnaeus, 1758 im östlichen Nordafrika mit Beschrei-bung eines neuen Taxons von der Cyrenaika (Nord-ostlibyen). Herpetozoa 15: 3-28.
Pieh, A., Perälä, J. (2004): Variabilität der MaurischenLandschildkröten (Testudo graeca Linnaeus, 1758 –Komplex) im zentralen und nordwestlichen Marokkomit Beschreibung zweier neuer Taxa. Herpetozoa 17:19-47.
Posada, D., Crandall, K.A. (1998): Modeltest: testing themodel of DNA substitution. Bioinformatics 14: 817-818.
Prior, K.A., Gibbs, H.L., Weatherhead, P.J. (1997): Popula-tion genetic structure in the black rat snake: implicationsfor management. Conserv. Biol. 11: 1147-1158.
Rögl, F. (1998): Paleogeographic considerations for Medi-terranean and Paratethys seaways (Oligocene toMiocene). Ann. Naturhist. Mus. Wien 99A: 279-310.
Rögl, F. (1999): Mediterranean and Paratethys. Facts andhypotheses of an Oligocene to Miocene palaeogeogra-phy (short overview). Geologica Carpathica 50: 339-349.
Schilde, M., Barth, D., Fritz, U. (2004): An Ocadia sinensis× Cyclemys shanensis hybrid. Asiatic Herpetol. Res. 10:120-125.
Semyenova, S.K., Korsunenko, A.V., Vasilyev, V.A.,Pereschkolnik, S.L., Mazanaeva, L.F., Bannikova, A.A.,Ryskov, A.P. (2004): RAPD variation in Mediterraneanturtle Testudo graeca L. Russ. J. Genetics 40: 1348-1355.
Swofford, D.L. (2002): PAUP*. Phylogenetic Analysis Us-ing Parsimony (*and Other Methods), Version 4.0b10.Sunderland, Massachusetts, Sinauer Associates.
Tassanakajon, A., Pongsomboon, S., Rimphanitchayakit, V.,Jarayabhand, P., Boonsaeng, V. (1997): Random ampli-fied polymorphic DNA (RAPD) markers for determina-tion of genetic variation in wild populations of the blacktiger prawn (Penaeus monodon) in Thailand. Mol. Mar.Biol. Biotechnol. 6: 110-115.
Tautz, D., Renz, M. (1984): Simple sequences are ubiq-uitous repetitive components of eukaryontic genomes.Nucl. Acids Res. 12: 4127-4138.
Tsumura, Y., Ohba, K., Strauss, S.H. (1996): Diversity andinheritance of inter-simple sequence repeat polymor-phisms in Douglas fir (Pseudotsuga menziesii) and sugi(Cryptomeria japonica). Theor. Appl. Genet. 92: 40-45.
Tuniyev, B.S. (1990): On the independence of the ColchisCenter of amphibian and reptile speciation. Asiatic Her-petol. Res. 3: 67-84.
Tuniyev, B.S. (1995): On the Mediterranean influence onthe formation of herpetofauna of the Caucasian isthmusand its main xerophylous refugia. Russ. J. Herpetol. 2:95-119.
Türkozan, O., Kumlutas, Y., Arikan, H., Ilgaz, Ç., Avcı,A. (2003): Morphological and serological comparisonof Mediterranean spur-thighed tortoises, Testudo graeca,from the Aegean region and southeastern Turkey. Zool.Middle East 29: 41-50.
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.
van der Kuyl, A.C., Ballasina, D.L.P., Zorgdrager, F. (2005):Mitochondrial haplotype diversity in the tortoise speciesTestudo graeca from North Africa and the Middle East.BMC Evol. Biol. 5: 29.
Vanlerberghe-Masutti, F., Chavigny, P. (1998): Host-basedgenetic differentiation in the aphid Aphis gossypiiGlover, evidenced from RAPD fingerprints. Mol. Ecol.7: 905-914.
Weissinger, H. (1987): Testudo graeca anamurensis ssp.nov. aus Kleinasien. ÖGH-Nachrichten 10/11: 14-18.
Wermuth, H. (1958): Status und Nomenklatur der Mau-rischen Landschildkröte, Testudo graeca, in SW-Asienund NO-Afrika. Senckenbergiana Biol. 39: 149-153.
Wermuth, H., Mertens, R. (1961): Schildkröten, Krokodile,Brückenechsen. Jena, Fischer.
Wermuth, H., Mertens, R. (1977): Testudines, Crocodylia,Rhynchocephalia. Das Tierreich 100: I-XXVII, 1-174.
Wheeler, Q.D., Meier, R., Eds, (2000): Species Con-cepts and Phylogenetic Theory. A Debate. New York,Columbia University Press.
Wheeler, Q.D., Platnick, N.I. (2000): The PhylogeneticSpecies Concept (sensu Wheeler and Platnick). In:Species Concepts and Phylogenetic Theory. A Debate,p. 55-69. Wheeler, Q.D., Meier, R., Eds, New York,Columbia University Press.
Wiley, E.O., Mayden, R.L. (2000): The EvolutionarySpecies Concept. In: Species Concepts and PhylogeneticTheory. A Debate, p. 70-89. Wheeler, Q.D., Meier, R.,Eds, New York, Columbia University Press.
Wink, M., Sauer-Gürth, H., Martinez, F., Doval, G., Blanco,G., Hatzofe, O. (1998): Use of GACA-PCR for molec-ular sexing of Old World vultures (Aves: Accipitridae).Mol. Ecol. 7: 779-782.
Wink, M., Guicking, D., Fritz, U. (2001): Molecular evi-dence for hybrid origin of Mauremys iversoni Pritchardet McCord, 1991, and Mauremys pritchardi McCord,1997. Zool. Abh. Mus. Tierkd. Dresden 51: 41-49.
Wolfe, A.D., Liston, A. (1998): Contributions of PCR-basedmethods to plant systematics and evolutionary biology,p. 43-86. In: Molecular Systematics of Plants II. DNASequencing. Soltis, D.E., Soltis, P.S., Doyle, J.J., Eds,Boston, Massachusetts, Kluwer Academic Publishers.
Wolfe, A.D., Xiang, Q.-Y., Kephart, S.R. (1998): Assessinghybridization in natural populations of Penstemon (Scro-phulariaceae) using hypervariable intersimple sequencerepeat (ISSR) bands. Mol. Ecol. 7: 1107-1125.
Zietkiewicz, E., Rafalski, A., Labuda, D. (1994): Genomefingerprinting by simple sequence repeat (ISSR)-anchored polymerase chain reaction amplification. Ge-nomics 20: 176-183.