Content uploaded by Anna Sztencel-Jabłonka
Author content
All content in this area was uploaded by Anna Sztencel-Jabłonka on Dec 13, 2017
Content may be subject to copyright.
Phylogeography of the smooth snake Coronella
austriaca (Serpentes: Colubridae): evidence for a
reduced gene pool and a genetic discontinuity in
Central Europe
ANNA SZTENCEL-JABŁONKA1*, TOMASZ D. MAZGAJSKI1, STANISŁAW BURY2,
BARTŁOMIEJ NAJBAR3, MARIUSZ RYBACKI4, WIESŁAW BOGDANOWICZ1and
JOANNA MAZGAJSKA1
1Museum and Institute of Zoology, Polish Academy of Sciences, ul. Wilcza 64, 00-679 Warszawa,
Poland
2Institute of Environmental Sciences, Jagiellonian University, ul. Gronostajowa 7, 30-387 Kraków,
Poland
3Faculty of Biological Sciences, University of Zielona Góra, ul. Prof. Z. Szafrana 1, 65-516 Zielona
Góra, Poland
4Department of Zoology, Kazimierz Wielki University, Al. Ossolin´ skich 12, 85-067 Bydgoszcz, Poland
Received 14 November 2014; revised 28 December 2014; accepted for publication 30 December 2014
The present study considers the genetic structure and phylogeography of the smooth snake (Coronella austriaca)
in Central Europe, as analyzed on the basis of 14 microsatellite markers and a 284-bp fragment of cytochrome b.
We found deep divergence between western and south-eastern Poland, suggesting at least two different colonization
routes for Central Europe, originating in at least two different refugia. The west/south-east divide was reflected in
the haplotype distribution and topology of phylogenetic trees as defined by mitochondrial DNA, and in population
structuring seen in the admixture analysis of microsatellite data. The well supported western European clade
suggests that another refugium might have existed. We also note the isolation-by-distance and moderate-to-
pronounced structuring in the examined geographical demes. Our data fit the assumption of the recently suggested
sex-biased dispersal, in that we found a strong divide in the maternal line, as well as evidence for a small but
existent gene flow based on biparentally inherited microsatellite markers. All studied populations were very similar
in respect of allelic richness, observed and expected heterozygosities, and inbreeding coefficients. However, some
genetic characteristics were different from those expected compared to a similar fine-scale study of C. austriaca
from Great Britain. In the present study, we observed heterozygosity deficit, high inbreeding, and low Garza–
Williamson indices, suggesting a reduction in population size. © 2015 The Linnean Society of London, Biological
Journal of the Linnean Society, 2015, ••, ••–••.
ADDITIONAL KEYWORDS: biogeography – genetic diveristy – microsatellites – mtDNA – refugium.
INTRODUCTION
Habitat loss and fragmentation that result in increas-
ing habitat-discontinuities in species with restricted
vagility may severely affect gene flow and the genetic
structure of populations (Tilman et al., 1994, 2001;
Fahrig, 2003; Frankham, 2005; Hamer & McDonnell,
2010; Keyghobadi, 2007; Krauss et al., 2010). Limited
gene flow can cause an increase in mating among kin,
leading to an increased individual homozygosity and
the loss of genetic variability. Additionally, when a
species has specific habitat demands, it often becomes
limited to small patchy habitats in which small iso-
lated populations with no or limited gene flow are
present (Saccheri et al., 1998; Keller & Waller, 2002).
*Corresponding author. E-mail: brysio@miiz.waw.pl
bs_bs_banner
Biological Journal of the Linnean Society, 2015, ••, ••–••. With 6 figures
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2015, ••, ••–•• 1
In small isolated populations, the impact of genetic
drift increases as well (Masel, 2011).
In such cases, often only well-planned conservation
management can prevent a species from declining.
However, for successful management, clear designa-
tion of the management unit becomes essential
(Palsbøll, Bérubé & Allendorf, 2007). Accordingly, a
multifaceted approach is required, including informa-
tion on the population density, species’ biology and
ecology (e.g. habitat demands), as well as on species’
phylogeography and population genetics (Avise,
2004).
The smooth snake, Coronella austriaca Laurenti,
1768,is a small nonvenomous colubrid occurring in
most of the European continent and part of Asia
(Arnold & Ovenden, 2002). Smooth snakes feed
mainly on lacertid lizards and, to a smaller extent, on
small mammals (the diet depending on sex and age;
Brown, Ebenezer & Symondson, 2013; Reading &
Jofré, 2013), and are associated with heathlands,
hedgerows, wood-edges, open woods and bushy and
rocky slopes (Arnold & Ovenden, 2002).
Smooth snakes have limited dispersal ability and
small home ranges (Reading, 2012). The narrow spec-
trum of habitats used, trophic specialization, and
conservative use of territory (Drobenkov, 2000) all
make C. austriaca especially prone to fragmentation
and disruption of suitable habitats, as well as
decreasing food availability (a phenomena affecting
many European species; Tilman et al., 1994, 2001;
Krauss et al., 2010). In these circumstances,
C. austriaca may suffer from severe population
decline and, at the genetic level, from heterozygosity
loss and even inbreeding depressions.
Reading et al. (2010) documented widespread
declines in populations of snakes, including
C. austriaca, in recent years, also indicating that six
of the eight declining species are characterized by
small home ranges, sedentary habits, and ambush-
based foraging strategies. As noted above,
C. austriaca is characterized by the first two features,
whereas ambush also plays an important (if not the
most important) role in its feeding strategy (Amo,
López & Martín, 2004).
Santos et al. (2009) link the vulnerability to local
extinction, which is typical for the smooth snake in
the southern Iberian mountains, to dietary speciali-
zation and a low rate of reproduction. Coronella
austriaca in the Iberian Peninsula is documented as
a lizard-eating specialist (82.1% frequency; Galán,
1988), and such dietary specialization has been iden-
tified as a factor increasing vulnerability to extinction
in snakes (Dodd, 1993).
Indeed, C. austriaca is listed under Appendix II of
the Bern Convention and Annex IVa of the EC Habi-
tats Directive, as well as being red-listed in 11 Euro-
pean countries. However, the species status was
assessed on the basis of observational data, with
neither the genetic pool of the species, nor its man-
agement units having been evaluated. This is not
attributed to a lack of interest in the species but,
instead, to the fact that C. austriaca is not easy to
study as a result of its secretive mode of life and low
densities in a number of areas within its geographical
distribution (Sewell et al., 2012). To date, only two
studies based on molecular markers have been con-
ducted for the species: one based on microsatellite
markers that evaluated the fine-scale population
differentiation in one locality in Great Britain
(Pernetta et al., 2011) and the other using
mitochondrial cytochrome bsequences to elucidate
the phylogeography of the species, mostly focusing on
the south-western edge of its range (Santos et al.,
2008), where two subspecies have been distinguished:
Coronella austriaca acutirostris in the north of
Iberian Peninsula, and Coronella austriaca fitzingeri
in southern Italy and Sicily (Arnold & Ovenden,
2002).
Based on biparentally inherited microsatellite loci,
Pernetta et al. (2011) suggested sex-biased dispersal
in the species. Taking the results of their study into
consideration, we chose to determine the genetic
structure of the species on the basis of both biparen-
tally and uniparentally (maternally) inherited
markers. The phylogeographical study (Santos et al.,
2008) revealed a complex and highly diversified
pattern of haplotypes in the Iberian Peninsula, with
three distinct clades. However, because that study
mostly examined individuals at the south-western
edge of the species’ range and considering its distri-
butional gap between the Peninsula and the rest of
Europe (Arnold & Ovenden, 2002), as well as possibly
different routes of postglacial colonization, we were
interested in evaluating how phylogeographical rela-
tionships are shaped in Central Europe. This was all
the more relevant given the findings from other
phylogenetic studies of reptiles that reveal this region
to be a meeting point of distant lineages from differ-
ent refugia as, for example, in the case of the Euro-
pean adder (Ursenbacher et al., 2006), the European
grass snake (Kindler et al., 2013), and the slow worm
(Gvoždík et al., 2010). Moreover, two previous molecu-
lar studies involving C. austriaca were located at the
limits of the species’ range and on an island, allowing
us to evaluate whether high population structuring
and high diversity of cytochrome bhaplotypes are
also present in the core area of the smooth snake’s
contiguous range (i.e. Central Europe).
Therefore, the present study aimed: (1) to deter-
mine the genetic structure patterns of smooth snake
populations, as well as differences in genetic struc-
ture and other population traits between an insular
2A. SZTENCEL-JABŁONKA ET AL.
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2015, ••, ••–••
population (Great Britain) and one from within the
contiguous range (Poland); (2) to infer the genetic
structure of the species (e.g. whether the sex-biased
dispersal observed in the course of small-scale studies
influences the structure over a longer evolutionary
and larger spatial scale) by contrasting patterns seen
in biparentally inherited (nuclear microsatellites) and
maternally inherited mitochondrial markers; and (3)
to determine the phylogenetic relationships of popu-
lations from Central Europe with other parts of the
species’ distribution, and to infer where possible the
colonization pathways of Central Europe.
Finally, we discuss how genetic and phylogenetic
data we obtained could be implemented in the con-
servation management plans for this species.
MATERIAL AND METHODS
STUDY AREA AND SAMPLING STRATEGY
From 2010 to 2013, we caught snakes and collected
genetic samples, mainly in Poland, at historically
known localities. Coronella austriaca has a patchy
distribution in Poland (Głowacin´ ski & Rafin´ ski,
2003), which is reflected in the distribution of
obtained samples (Fig. 1), and has not been noted in
north of the country near the Baltic sea. All snakes
were captured by hand and sexed by reference to
their tail length and shape (Reading, 2004). We also
weighed them and measured their total body length,
photographed the scale pattern on their heads, and
recorded the GPS (Global Positioning System) loca-
tion of capture. We obtained samples for genetic
analyses using cotton-tipped sterile swabs and
clipped one ventral scale for marking purpose (Ethic
Committee agreement number 63/2010). We also col-
lected shed skins, tissue (road kills) and remains of
eggs. The buccal swab samples and shed skin samples
were stored dry at −20 °C; all other samples were
kept in 95% ethanol. We obtained 148 samples in
total, including 136 from Poland and 12 from other
countries: Slovakia (N= 1), the Czech Republic
(N= 3), Croatia (N= 3), and Armenia (N= 5) (Fig. 1).
Samples donated to us were all from road-killed
animals.
Figure 1 A, distribution of cytochrome bhaplotypes of Coronella austriaca austriaca in Europe (those original to the
present study are represented by small-sized squares supplemented with H1–H10, those from GenBank are shown by
different symbols). The range of clades designated by Santos et al. (2008) for the Iberian Peninsula is also provided. B,
the map of Poland showing geographical demes designated for analyses of microsatellite markers.
SMOOTH SNAKE PHYLOGEOGRAPHY AND POPULATION STRUCTURE 3
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2015, ••, ••–••
DNA ISOLATION,MICROSATELLITE GENOTYPING,AND
MTDNA SEQUENCING
Genomic DNA was isolated using the QuickGene
DNA tissue kit S (Kurabo Industries Lrd) in accord-
ance with the manufacturer’s instructions for swab or
tissue samples on an automated nucleic acid extrac-
tion Quick Gene-810 system (FujiFilm). Using the
polymerase chain reaction (PCR), we genotyped a
total of 148 snake samples with 16 microsatellite
markers (Bond et al.,2005), in three sets of multiplex
reactions, with the forward primers labelled with
WellRed Dyes D2, D3 and D4 (set 1, D3: Ca43,Ca66;
D2: Ca16,Ca19; D4: Ca27,Ca63; set 2, D3: Ca20,
Ca45; D4: Ca30,Ca612; D2: Ca40; and set 3, D3:
Ca47; D4: Ca26,Ca78; D2: Ca79). We performed the
PCRs in a total volume of 10 μl, consisting of 10 ng
of DNA template, 0.2 μM of each primer, 5 μlof
Multiplex PCR Master Mix (Qiagen), and water. For
PCR amplification, we used a thermal cycler (Applied
Biosystems) with the PCR profile: initial denaturation
at 95 °C for 15 min (hot start), 40 cycles of 30 s at
94 °C, 90 s at 56 °C, 90 s at 72 °C, followed by a final
elongation step at 72 °C for 10 min. PCR products
were genotyped on an CEQ 8000 DNA fragment
analyzer (Beckman Coulter) and the genotypes were
scored using FRAGMENT ANALYSIS software
(Beckman Coulter).
Primers used for the fragment of cytochrome b,in
both amplification and sequencing, were L14846 and
H15149 (Kocher et al., 1989). Amplification was
carried out in a total volume of 25 μl, consisting of
20 ng of DNA template, 0.5 μM of each primer,
12.5 μl of REDTaq ReadyMix PCR Reaction Mix
(Sigma-Aldrich), and water; with the PCR profile:
initial denaturation at 94 °C for 2 min, 35 cycles of
15 s at 94 °C, 20 s at 45 °C, 60 s at 72 °C, followed
by a final elongation step at 72 °C for 10 min.
Sequencing PCR was carried out with a BigDye Ter-
minator v3.1 Cycle Sequencing Kit, with sequencing
on an ABI 3500 XL Genetic Analyzer, in accordance
with the manufacturer’s instructions (Applied
Biosystems).
PHYLOGEOGRAPHICAL ANALYSIS
The 146 sequences obtained (284 bp of cytochrome b;
10 haplotypes) and 37 homologous sequences from
GenBank (popset 158819496 from Santos et al., 2008;
sample from Ilgaz, Turkey: AY486930.1 from Nagy
et al., 2004; samples from Serra da Estrela, Portugal:
JQ904296 and Madrid, Spain: JQ904299 from Santos
et al.,2012) were aligned using MEGA, version 6
(Tamura et al., 2013). The popset included two
sequences of Coronella girondica used as outgroups
for tree rooting. The best fitting evolutionary model
for the sequences was inferred using the Akaike
criterion corrected for the finite sample size in
JMODELTEST, version 2.1.3 (Posada, 2008). The
aligned DNA data were further analyzed using the
maximum likelihood (ML) and Bayesian inference
(BI) methods. For ML, PHYML, version 3.0 (Guindon
et al., 2010) was applied with 1000 bootstraps to
assess the reliability of the ML tree. The bootstrap
value (BP) is presented as a percentage of the number
of times a particular clade was found. The BI was
carried out in MrBayes, version 3.2.1. (Ronquist et al.,
2012) and the implemented Metropolis-coupled
Markov chain Monte Carlo (MCMC) algorithm. Two
parallel runs with one cold and three heated runs
were conducted, with every 500th generation sampled
for 1 × 107generations. To generate the final 50%
majority-rule consensus tree, a burn-in of 25%, was
used to sample only the most likely trees. Conver-
gence was checked by examining the generation plot
visualized with TRACER, version 1.6 (Rambaud,
Suchard & Drummond, 2013). Bayesian posterior
probabilities (PP) were used to assess branch support
of the BI tree. We considered PP ≥0.95 as ‘strong’ and
PP = 0.90–0.95 as ‘good’ support, in accordance with
other studies. For BP, we considered good support to
be no less than 700 per 1000 (i.e. 70%). According to
the simulation-based study of Hillis & Bull (1993),
bootstrap proportions in majority-rule consensus
trees provide biased but highly conservative esti-
mates of the probability of correctly inferring corre-
sponding clades. Under conditions of equal rates of
change, symmetric phylogenies, and internodal
change of maximally 20% of the characters, bootstrap
proportions of ≥70% usually correspond to a probabil-
ity of ≥95%, thereby indicating a high level of support
for the corresponding clade. To further examine
phylogeographical relationships, a network for the
same data set was constructed in NETWORK, version
4.6.1.2 (Bandelt, Forster & Röhl, 1999), using the
median joining option. The uncorrected genetic dis-
tances (pi) with standard error within and among
sequence groups were calculated in MEGA, version 6
(Tamura et al., 2013) and are given as percentage
values.
Names of the sequences are given as: SPA55, where
letters designate the country of origin and the
number comprises the two last digits of the GenBank
accession number, which are as listed by Santos et al.
(2008), except for three samples for which the acces-
sion numbers are given above. For C. a. acutirostris
and C. a. fitzingeri, letters A and F are given respec-
tively at the end of the sequence name (i.e. ITL70F)
(ARM, Armenia; CES, Czech Republic; CRO, Croatia;
FRA, France; GRB, Great Britain; GRE, Greece; ITL,
Italy; AUS, Austria; POL, Poland; POR, Portugal;
RUS, Russia; SLO, Slovakia; SPA, Spain; TUR,
Turkey). Instead of accession number last digits,
4A. SZTENCEL-JABŁONKA ET AL.
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2015, ••, ••–••
haplotypes original to the present study have symbols
H1, H2, H3, etc.
STATISTICAL AND POPULATION GENETIC ANALYSIS OF
MICROSATELLITE MARKERS
Genotypic data were checked for amplification errors
using MICRO-CHECKER, version 2.2.3 (van
Oosterhout et al., 2003). As a result of poor amplifi-
cation success, loci Ca19 and Ca45 were excluded
from further analyses. All analyses were based on 14
microsatellite markers. Because we used shed skins
as a DNA source, all genotypes were checked for their
uniqueness. Two genotypes were found to be repeated
(both with 100% identity); thus, 146 genotypes were
used in the analyses. As a result of uneven sampling
success in various geographical regions, in most
microsatellite analyses, samples from seven geo-
graphical demes were used, encompassing 122 indi-
viduals: Bieszczady Mountains (BIE), Pieniny
National Park (PPN), Roztocze and Puszcza Solska
Landscape Park (ROZ), Lubuskie Voivodeship upriver
of the Odra (LUB), Lubuskie Voivodeship downriver
of the Odra (ODR), Western Pomerania Voivodeship
(POM), and Gostynin-Włocławek Landscape Park
(CEN) (Fig. 1). The linkage disequilibrium and con-
formance with Hardy–Weinberg equilibrium (HWE)
were calculated using GENEPOP, version 4.2
(Raymond & Rousset, 1995; Rousset, 2008), using an
exact probability test (Markov chain parameters:
10 000 dememorizations, 100 batches, 1000 iterations
per batch), with Bonferroni correction (Rice, 1989). As
a result of deviation from HWE being detected,
hypotheses regarding heterozygosity deficit and
excess were also tested. Because groups from geo-
graphical demes/loci deviated from HWE after the
Bonferroni correction, we did not check for the pres-
ence of null alleles because of difficulty in distinguish-
ing this phenomenon from heterozygotes deficit
(David et al., 2007; Da˛ browski et al., 2013).
Allele frequency, mean number of alleles per locus
(NA), inbreeding coefficient (FIS) with a randomization
(N=1960) test for its significance, and allelic richness
(AR) as a standardized measure of the number of
alleles corrected by sample size were all calculated
using FSTAT, version 2.9.3 (Goudet, 2001). Observed
(HO) and expected (HE) heterozygosities were calcu-
lated in ARLEQUIN, version 3.5.1.2 (Excoffier &
Lischer, 2010). Allelic richness and measures of
genetic diversity were compared between geographi-
cal demes, using Kruskal–Wallis tests in PAST,
version 2.17c (Hammer, Harper & Ryan, 2001).
The hypothesis of a bottleneck effect was tested
in BOTTLENECK (Cornuet & Luikart, 1996). We
assumed that the two-phase model (Di Rienzo et al.,
1994) was the one best suiting our markers; and
performed the Wilcoxon test and allele frequency
curve shape evaluation (Luikart & Cornuet, 1998).
Additionally, the Garza–Williamson index (M; i.e. the
number of alleles divided by the allelic range),
expected to be low in populations that have under-
gone size reduction, was measured using ARLEQUIN,
version 3.5.1.2 (Excoffier & Lischer, 2010).
We tested for the presence of isolation-by-distance
using Mantel’s test (Mantel, 1967) between the
matrix of linearized pairwise FST(FST/(1 – FST); values
of FST from FSTAT) and a matrix of linear geographi-
cal distance between populations in GENALEX,
version 6.5 (Peakall & Smouse, 2006, 2012).
STRUCTURE, version 2.3.4 software (Pritchard,
Stephens & Donnelly, 2000) was used to first deter-
mine the number of genetic clusters (K) and then to
assign probabilities of an individual belonging to each
cluster. We tested the number of clusters from 1 to 14
with 10 iterations for each K(50 000 burn-ins,
100 000 MCMC replicates in each run) using the
admixture model and assuming correlated allele fre-
quencies. We determined the final number of clusters
from ΔK, the rate of change in the log probability
over all 10 iterations (Evanno, Regnaut & Goudet,
2005) using STRUCTURE HARVESTER (Earl & von
Holdt, 2012). Because different solutions may result
from replicated cluster analyses, we used CLUMPP
(Jakobsson & Rosenberg, 2007) with the ‘greedy’
option amd with 10 000 random input orders to iden-
tify the optimal individual alignments of replicated
cluster analyses. The population mixture analyses
(clustering of groups and spatial clustering of groups)
was carried out in BAPS6 (Corander, Sirén & Arjas,
2008), with Kranging from 1 to 20, each repeated 10
times. As noted in Corander et al. (2008), the ration-
ale of using spatial information is to assign a biologi-
cally relevant non-uniform prior distribution over the
space of clustering solutions, which expects that
underlying clusters are spatially smooth at least to
some extent. This increases the power to correctly
detect the underlying population structure.
RESULTS
PHYLOGEOGRAPHICAL ANALYSIS
Of the 284 bp of mtDNA analyzed, 81 positions were
variable and 62 were parsimony informative. The best
model of DNA evolution for the examined cytochrome
bfragment was HKY + G (Hasegawa, Kishino and
Yano + gamma distribution). Both the ML and BI
trees resulted in a similar general picture. There were
some differences in higher branches; however, topo-
logical incongruences occurred only in places where
both methods showed weak nodal support. Therefore,
trees are presented as one figure (Fig. 2), on which
SMOOTH SNAKE PHYLOGEOGRAPHY AND POPULATION STRUCTURE 5
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2015, ••, ••–••
Figure 2. Phylogenetic tree based on the cytochrome bsequences: Bayesian inference tree with posterior probability
values (under the line: bootstrap values shown as a percentage of the number of times a particular node was found in
the maximum likelihood analysis).
6A. SZTENCEL-JABŁONKA ET AL.
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2015, ••, ••–••
the topology of the BI tree is given and values of BP
are given under the PP values separated by a line.
Polish samples had five unique haplotypes (H2, H4,
H6, H7, H8) and two shared with other samples: H1,
as described earlier from north-east Italy (Treviso)
and shared with all our samples from Croatia, and
H3, also found in Slovakia (Fig. 1). Unique haplotypes
were also observed in samples from the Czech Repub-
lic (H5) and Armenia (H9, H10). Haplotype H1 in
Poland had both the highest frequency (56.7%) and
the largest geographical range. It was discovered in
the Pieniny National Park (PPN) and near Cracow
and Opole (KRA) in the south, in central Poland
(CEN), and in all western regions (POM, LUB, ODR).
The frequency of haplotypes H3 from Roztocze (ROZ)
and H7 from Lubuskie Voivodeship (LUB) amounted
to 24.6% and 9.7%, respectively. Haplotype H2,
unique to south-eastern Poland (the Bieszczady
Mountains), accounted for 6.0%. Three haplotypes
were very rare: H8 was noted in two individuals from
PPN and one from LUB (2.2%), H4 in one individual
from the vicinity of Opole, KRA (0.7%), and H6 in one
individual from Pomerania, POM (0.7%).
Haplotype H1 (occurring in south, central and
eastern Poland, Croatia, and Italy) plus haplotypes
H4, H6, H7, and H8 from south-western Poland, and
H5 from the Czech Republic, grouped in one well-
supported (PP =1,BP = 97) North Balkan Clade with
a within-group genetic distance pi = 0.59 ± 0.26,
whereas haplotypes H2 and H3 from south-eastern
Poland (BIE and ROZ, respectively) and Slovakia
(H3) grouped with the haplotypes from Turkey and
Greece, respectively. Nevertheless, details of tree
topology differ depending on the method used and are
not always strongly supported. Haplotype H3 and one
from Greece are in one well-supported clade (pi =
1.06 ± 0.59; PP = 0.9, BP = 73), whereas haplotypes
H2 and haplotype from Turkey form separate clades
(Fig. 2).
The network of haplotypes designates haplotype
H1 in the centre from which haplotypes H4, H5, H6,
H7, and H8 diverge, differing by one substitution
step (Fig. 3). Haplotypes H2 and H3 are more than
14 substitution steps away from the above mentioned
group, and haplotype H3 groups with the Greek one
and haplotype H2 groups with the Turkish one
(the number of substitution steps between H2
and Turkey is considerably high, reaching two and
eight steps from common node, respectively). This
topology is in congruence with both trees topologies.
According to the topologies of the phylogenetic trees
and the network, the Russian samples group with
samples from Armenia (PP = 0.8, BP = 71) and they
are two (Armenia) and five steps (Russia) from the
common node in the network (Eastern Clade;
pi = 1.43 ± 0.47).
In both trees, the Eastern Clade is grouped
together with the North Balkan Clade, although
support is very weak (PP = 0.5, BP = 51); indeed, on
the network, these groups are not closely connected
and pi distance between them is also high
(pi = 5.8 ± 1.2). Regarding samples from the Iberian
Peninsula, the topologies of both trees and the
network point to a well-supported clade of C. a.
acutirostris (occurring in the Iberian Peninsula only;
PP =1, BP = 100; Clade I in Santos et al., 2008;
pi = 0.8 ± 0.32), separated from its sister group
formed by all other samples composed of C. a.
austriaca. All remaining Iberian samples are divided
into two well-supported clades (named Clade II and
Clade III in Santos et al., 2008). In our analysis, the
samples of C. a. fitzingeri (from Sicily) constitute a
separate well-supported taxon (PP =1, BP = 100).
Another main clade (Western Europe Clade,
pi = 0.18 ± 0.17) is formed by samples from northern
France, Great Britain, and Austria (PP =1,
BP = 100), and it is also clearly designated in the
network topology.
The North Balkan Clade has the highest values of
uncorrected pi distance from other clades, ranging
from 5.75 ± 1.2 (Eastern Clade) to 9.1 ± 1.5 (Clade II),
whereas, for the rest of the clades, pi is within the
range of 4.46–6.81.
GENETIC VARIABILITY AND SPATIAL STRUCTURE
In all examined geographical demes apart from ODR
(P<0.029), the observed allele frequencies showed
significant deviations from the HWE after the
Bonferroni correction. Statistically significant depar-
ture from HWE was also observed in the case of
11 loci other than Ca61,Ca79, and Ca612.Inall
demes and all loci, apart from Ca66 and Ca612,
heterozygosity deficiency was also highly significant
(P≤0.001). No systematic linkage disequilibrium
between loci was detected (data not shown). Genetic
parameters such as observed heterozygosity
(Kruskal–Wallis H= 0.71, P= 0.99), expected heter-
ozygosity (H= 6.73, P= 0.35), allelic richness (H=
9.08, P= 0.17), Garza–Williamson index (H= 3.93,
P= 0.69), and inbreeding coefficient (H= 3.66,
P= 0.72) did not vary significantly between geo-
graphical demes (Table 1). Values of FST between
demes varied from moderate to high (Table 2), in
some cases accounting for more than 20% of the total
genetic variation among geographical demes.
Assuming that loci fit the two-phase model, no
bottleneck was confirmed in any of the geographical
demes studied because the Wilcoxon tests for
heterozygosity excess gave nonsignificant values and
allele frequency curve shapes were not shifted.
However, in all demes, the Garza–Williamson index
SMOOTH SNAKE PHYLOGEOGRAPHY AND POPULATION STRUCTURE 7
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2015, ••, ••–••
was very low, ranging from 0.23 (BIE) to 0.31 (ODR)
(i.e. suggestive of a reduction in population size)
(Table 1).
Isolation-by-distance examined for the seven most-
numerous groups was found to be significant at
P<0.01, with approximately 8.8% of genetic variation
explained by geographical distance (Fig. 4).
In STRUCTURE analyses, with the Evanno method
applied, two genetic clusters were identified (see Sup-
porting information, Fig. S1). Nevertheless, taking
into account the posterior probability Pr(X|K) distri-
bution (see Supporting information, Fig. S2) and the
fact that isolation-by-distance was confirmed in the
examined dataset, which may bias an outcome of
the Evanno method, a scenario for seven clusters is
also presented.
In a two-cluster scenario (Fig. 5A), a divide between
south-eastern Poland (in black: BIE, PPN, and ROZ)
and western Poland (in grey: LUB, POM, and ODR) is
visible. The group from central Poland (CEN) showed
an admixed character. In a seven-cluster scenario, the
population structuring was in concordance with the
geographical distribution of sampled groups, although
a considerable amount of admixture is visible
(Fig. 5B). The western groups POM, LUB, and ODR
were well differentiated from the rest of the popula-
tion, with the two geographically closest (POM, LUB)
forming one genetic cluster, whereas the third was
not only dominated by a separate cluster, but also has
a cluster specific for POM and LUB admixed. The
largest group ROZ was heterogenic and comprised
two dominating and regionally specific clusters, with
Figure 3. Network of cytochrome bhaplotypes. Numeric values represent the number of substitution steps between
haplotypes, with the size of the circle being proportional to the number of individuals with a given haplotype from
different localities.
8A. SZTENCEL-JABŁONKA ET AL.
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2015, ••, ••–••
some signs of admixture of clusters from the well
differentiated BIE group, and the second largest
group PIE was dominated by one cluster, with some
admixture from the CEN group (Fig. 5B). As in the
two-cluster scenario, a divide between western and
south-eastern demes is visible, with the CEN group
having an intermixed character.
The mixture analyses in BAPS6 of both the spatial
group clustering (with information about geographi-
cal distribution of samples included) and the ‘stand-
ard’ group clustering divided the data into six clusters
with P= 1. Sampling sites POM and LUB were
grouped in one cluster and all other geographical
demes assigned to separate genetic clusters (Fig. 6A).
The spatial group clustering with the samples that
were not included in previous analyses as a result of
either small numbers or too extensive geographical
spread to be regarded as a single deme (i.e. Croatia,
CRO, N=3; Czech Republic, CZE, N=3; Armenia,
ARM, N=5; south-western Poland, Kłodzko Valley,
KLO, N=2); southern Poland, near Cracow and
Opole, KRA (N=8) also resulted in the designation of
six clusters (P= 0.99): BIE, PPN, ROZ (all from
south-eastern Poland), CRO, and ARM constituting
five clusters, and the rest of the regions POM, LUB,
ODR, KLO, KRA, and CEN (all from either central or
western part of Poland), as well as CZE, forming one
genetic cluster (Fig. 6B).
DISCUSSION
PHYLOGEOGRAPHY
The ML and BI trees confirmed the topology of the
tree in Santos et al. (2008) (i.e. the same three clades
being designated for samples from the Iberian Pen-
insula, with two additional samples from Santos
et al., 2012). Samples from northern France, Great
Britain, and Austria constitute one group (the
Western Europe Clade), and this monophyly is
strongly supported on both the BI and ML trees and
in Santos et al. (2008). The number of haplotypes in
Central Europe is comparable with that found in the
Iberian Peninsula. The group of haplotypes from
Poland belonging to the North Balkan Clade probably
diverged quite rapidly and recently, as seen in the
network results, within which a star-like topology
leading from haplotype H1 is visible. In addition, the
high values for pi distances between this clade and
other clades, coupled with a low within-group pi dis-
tance, indicate a scenario of rapid recent radiation for
this group (post-glacial or occurring in the last stages
of the glaciation period).
It remains impossible to draw general conclusions
concerning the phylogeography of the C. austriaca
Table 1. Genetic population structure of C. austriaca in Poland
Deme nH
EHONAARMF
IS
CEN 10 0.64 (0.15) 0.45 (0.29) 4.71 (2.02) 3.69 (1.28) 0.30 (0.17) 0.31
LUB 15 0.55 (0.20) 0.43 (0.34) 4.64 (1.69) 3.31 (1.10) 0.27 (0.15) 0.22
POM 6 0.61 (0.22) 0.51 (0.28) 3.50 (1.40) 3.35 (1.30) 0.28 (0.14) 0.19
ODR 12 0.58 (0.20) 0.45 (0.32) 4.00 (1.79) 3.21 (1.12) 0.31 (0.17) 0.24
PPN 39 0.68 (0.15) 0.46 (0.17) 7.29 (2.73) 4.06 (1.10) 0.29 (0.08) 0.33
ROZ 33 0.68 (0.22) 0.49 (0.24) 8.93 (4.12) 4.51 (1.59) 0.30 (0.08) 0.28
BIE 7 0.69 (0.17) 0.48 (0.22) 4.50 (1.79) 4.03 (1.52) 0.23 (0.10) 0.30
Total 122 0.70 (0.17) 0.46 (0.21) 12.14 (4.44) 4.55 (1.34) 0.32 (0.07) 0.34
CEN, Gostynin-Włocławek Landscape Park; LUB, Lubuskie Voivodeship upriver of the Odra; POM, Western Pomerania
Voivodeship; ODR, Lubuskie Voivodeship downriver of the Odra; PPN, Pieniny National Park; ROZ, Roztocze and Puszcza
Solska Landscape Park; BIE, Bieszczady Mountains. n, sample size; HE, expected heterozygosity; HO, observed
heterozygosity; NA, mean number of alleles; AR, allelic richness; M, Garza–Williamson index; FIS, inbreeding coefficient. SD
in parentheses, significant FIS values indicated in bold.
Table 2. Pairwise estimates of FST between smooth snake
geographical demes in Poland
Deme CEN LUB POM ODR PPN ROZ
CEN –
LUB 0.08 –
POM 0.14 0.14 –
ODR 0.12 0.13 0.16 –
PPN 0.06 0.09 0.13 0.11 –
ROZ 0.08 0.07 0.13 0.10 0.04 –
BIE 0.12 0.20 0.24 0.19 0.13 0.13
CEN, Gostynin-Włocławek Landscape Park; LUB,
Lubuskie Voivodeship upriver of the Odra; POM, Western
Pomerania Voivodeship; ODR, Lubuskie Voivodeship
downriver of the Odra; PPN, Pieniny National Park; ROZ,
Roztocze and Puszcza Solska Landscape Park; BIE,
Bieszczady Mountains. Pairwise FST values that are sig-
nificantly different from zero following the Bonferroni cor-
rection are indicated in bold.
SMOOTH SNAKE PHYLOGEOGRAPHY AND POPULATION STRUCTURE 9
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2015, ••, ••–••
group as a whole because of unresolved relationships
of the main clades of the ML and BI trees. However,
certain conclusions on a smaller scale may be arrived
at. All three methods (trees and the network) indicate
a considerable difference between haplotypes H2 and
H3 and the remaining Polish haplotypes grouped
within the North Balkan Clade. H3 from Poland and
Slovakia constitutes one clade with the sample from
Greece and is well supported, whereas H2 forms a
separate clade (BI) or with the Turkish sample (in
line with the network topology). Both haplotypes were
found in south-eastern Poland, which, taking into
account their grouping along with samples from
Greece and possibly Turkey, may indicate a south-
Figure 4. Relationship between pairwise FST values and geographical distance between smooth snake geographical
demes in Poland.
Figure 5. Plot of posterior probabilities (Q) of individual’s assignment to a given genetic cluster: a two-cluster (A) and
a seven-cluster scenarios (B) are shown. Geographical demes are presented in order from the most western to the most
eastern one.
10 A. SZTENCEL-JABŁONKA ET AL.
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2015, ••, ••–••
eastern colonization route for this part of Poland. It
may be presumed that the route in question skirted
the Carpathian Mountains on the eastern side, or
even originated from the Carpathians, because they
are known to have constituted a refugium during the
Younger Dryas stadial; for example, in the case of the
European adder Vipera berus (Ursenbacher et al.,
2006) and the natterjack toad Bufo calamita (Rowe,
Harris & Beebee, 2006). To fully resolve this issue, it
will be necessary to analyze samples from Greece,
Bulgaria, Romania, Ukraine, and northern Turkey.
On the other hand, haplotype H1 occurring in
western and central Poland was also found in Croatia
and north Italy (the North Balkan Clade) (Fig. 1),
with this pointing to a south-western route by which
both Poland and Central Europe in general could
have been colonized from the Balkans. Not much is
known about the history of Europe’s colonization by
C. austriaca. Strijbosch (1997) mentions that coloni-
zation after the glaciations took place from the
Adriato-Mediterranean region. Our results point to
the Adriatic or even a Carpathian scenario in the case
of south-eastern Poland but, as far as western Poland
is concerned, both the Mediterranean and the Adri-
atic scenarios are possible because H1 was found in
both Italy and Croatia. Again, if C. austriaca followed
the pattern of V. berus (Ursenbacher et al., 2006),
which is quite probable considering the similarities in
the present ranges of the species (both having a very
wide range covering almost all of Europe and parts of
Asia; with V. berus distributed further north-east
being the most widely-distributed snake species),
then the colonization may have started from the
Balkans, spreading from there to the Mediterranean
region, as well as west-central Europe. Although we
opt for this scenario of colonization, there is no doubt
that much more detailed research is needed to fully
resolve this issue, with regard to south-eastern
Europe and most especially the region of the
Carpathian Mountains.
The grouping of haplotypes from Western Europe
(France, Great Britain, Austria) in a separate
Western Europe Clade, as confirmed by ML, BI, and
the network, suggests that this area might have been
Figure 6. Group clustering of the Polish populations with Voronoi tessellations (A) and of all genotyped samples (B).
SMOOTH SNAKE PHYLOGEOGRAPHY AND POPULATION STRUCTURE 11
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2015, ••, ••–••
colonized from yet another refugium, presumably
from the Apennine Peninsula.
Because the European temperate fauna and flora
have experienced multiple range shifts driven by
severe palaeoclimatic oscillations (Taberlet et al.,
1998; Hewitt, 2000, 2004), it is not only the spatial
scale, but also different time scales, that need to be
considered. It is very probable that colonization of
western and south-eastern Poland occurred in differ-
ent periods. However, this will again need confirming
in the course of further research.
GENETIC STRUCTURE
Compared to other snake species studied (Reading
et al., 2010) and considering that C. austriaca is a
species of restricted vagility and high specialization,
we expected to find a highly-structured population
with high inbreeding coefficients and some genetic
signature of a bottleneck. However, in comparison
with the population from Great Britain with a patchy
distribution and a very limited range (Pernetta et al.,
2011), we expected to find slightly higher levels of
heterozygosity, and lower inbreeding levels with
greater pairwise genetic distances between geo-
graphical demes, as a result of the broader scale of
the present study and its localization in the core area
of the species’ range, where we expected to find a
larger population less prone to genetic drift.
We did observe a significant isolation-by-distance
effect, and both mixture and admixture analyses
revealed separate genetic clusters much in concord-
ance with geographical demes, although this was not
as pronounced as expected. The correlation between
genetic and geographical distance, although signifi-
cant, was not very high, and not as pronounced as in
C. austriaca in Great Britain (Pernetta et al., 2011).
This could be explained by the fact that sampling at
large distances compared to dispersal distance may
result in an absence of isolation-by-distance (Rousset,
1997). However, admixture analysis also showed some
sign of gene flow, and pairwise genetic distances (FST)
between examined groups were rather moderate,
especially when compared with the Pernetta et al.
(2011) study, in which very similar values for genetic
distances were observed, although the distance
between analyzed demes were less than 6 km as
opposed to up to 600 km in the present study (the
eight markers used in the present study were the
same as in the British research, and so a quite direct
comparison of FST is justified).
A probable explanation for structuring not being as
pronounced as expected may lay in the long-term
effect of sex-biased dispersal, which was found
recently in the smooth snake (Pernetta et al., 2011).
Unfortunately, for our dataset, the tests for sex-biased
dispersal, as used by Pernetta et al. (2011), did not
prove conclusive as a result of strong female-biased
sampling. Another factor that could strengthen the
effect of male-induced gene flow is multipaternity,
which has been observed in many snake species
(Ursenbacher, Erny & Fumagalli, 2009). This scenario
is also in concordance with the much clearer separa-
tion stated in maternally-inherited cytochrome b
sequences, in which a strong divide between different
geographical regions of Poland is visible because most
geographical demes are monohaplotypic and only two
haplotypes (H1 and H8) are not specific for one
region. Of course, variability in mitochondrial
markers by nature, as a result of differences in muta-
tion rate and inheritance mode, is expected to be more
limited than in microsatellites but, considering the
topology of the phylogenetic trees and network, the
divide is deep, reflecting the history of Poland’s colo-
nizations and indicating limited subsequent gene flow
in the maternal line.
Similar FST values in Poland and Great Britain (i.e.
on completely different spatial scales), may also be
explained to some extent by the fact that fine-scale
genetic structure is likely to arise as a result of
founder effects during colonization of suitable habitat
patches (with the population in Great Britain also
subject to range shifts), and may be obscured over
time and by scale-dependent modes of dispersal. Such
a mechanism has been observed even over a short
time scale; for example, in a population of the African
gecko Hemidactylus mabouia during range expansion
(Short & Petren, 2011). Considering the smooth
snake’s distribution in Great Britain, where it is very
rare, patchily distributed and dependent on the pres-
ence of heathland, this appears to be a reasonable
explanation: a founder effect and genetic drift may be
much more pronounced there than in the larger and
more continuous population in Poland.
We did not detect a significant bottleneck in any of
the geographical demes. However, in all demes, the
Garza–Williamson index was very low, indicating that
in all examined regions the populations have been
reduced in size. This contradicting result can be
explained by the fact that M-ratio is expected to have
a long recovery time, whereas heterozygosity excess
and allele frequency distributions will recover rela-
tively quickly (Garza & Williamson, 2001). As such,
populations that have primarily been influenced by
historic decline(s) will show low M-ratios but nonsig-
nificant heterozygosity excess and allele shifts,
whereas populations recently or currently in decline
will not have time to recover from the genetic signa-
tures associated with any of the bottleneck detection
methods (Spear et al., 2006). The method imple-
mented in BOTTLENECK is more likely to detect a
major, rapid decline, whereas, in the examined popu-
12 A. SZTENCEL-JABŁONKA ET AL.
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2015, ••, ••–••
lation, the more probable scenario is that it suffered
from a long-term, slow decline, buffered in some way
by the life strategy of the species but still not
negligible.
In line with this finding, the heterozygosity deficit
was considerable, and inbreeding values were high in
all examined groups, and also more pronounced than
in the British populations. This is an unexpected
result because a population within a continuous
range is expected to be less influenced by phenomena
such as genetic drift and heterozygosity loss com-
pared to an insular population, even more so that
many of our sampling sites were located in protected
areas with well preserved, nonfragmented habitats.
Regarding all population parameters, apart from the
relatedness coefficient, we found all geographical
demes to be very similar; all had low heterogeneity
(much lower than expected), similarly low allelic rich-
ness, and high values for inbreeding coefficients.
Moreover, all these parameters were very similar to
those in the study by Pernetta et al. (2011). It appears
that the answer to such demographic stability of
C. austriaca despite such different study scales must
be sought in the life history of the species. Even in a
much more continuous range of suitable habitat, the
distribution of C. austriaca would not be fully con-
tinuous because these animals are bound up with
habitats that are likely to occur in patches, such as
heathlands, meadows, forest clearings, and edges
(Strijbosch, 1997). The best model fitting the
C. austriaca population would presumably be that of
a metapopulation (a system of small, patchy, regional
populations with small but existent gene flow). This is
important in terms of the conservation management
of this species, which is considered as endangered or
vulnerable across most of its range (Joint National
Conservation Committee, 2010). Hence, it would
appear that special care needs to be taken in moni-
toring the dynamics of such patchy subpopulations,
with more care perhaps being needed in sustaining
numerous small populations, as opposed to concen-
trating on several larger ones that will not be carry-
ing a large gene pool.
In their study of populations of rattlesnakes
Sistrurus catenatus catenatus, Gibbs et al. (1997)
stated that low levels of gene flow and genetic isola-
tion on very fine spatial scales may be the evolution-
ary norm for this species, and this may be a reason
why these snakes persist even as isolated populations
in small, recently fragmented landscapes. It is possi-
ble that this applies to C. austriaca as well, although
a high level of persistence of the species in the face
of unfavourable circumstances should not, in any
way, lessen the effort undertaken for the sake of
species conservation. Low allelic richness and low
heterozygosity, and high values for coefficients of
inbreeding and low values of the Garza–Williamson
index, all suggest a reduced gene pool and size reduc-
tion at the population level, and thus a danger that
the species will not be able to adapt when sudden
changes in the environment occur.
Knowledge of the spatial genetic structure of wild
populations is central to their management and con-
servation, and necessary if management units are to
be identified (Palsbøll et al., 2007). We therefore took
special care to make inferences about the number of
genetic clusters in the examined population. The
issue of inferring the most likely cluster number in a
population has been the subject of broad discussion
(Frantz et al., 2009; Guillot et al., 2009; Jombart,
Devillard & Balloux, 2010; Aurelle & Ledoux, 2013).
The Evanno method appears to represent the most
popular approach taken, although the observed
isolation-by-distance to which the method is prone
(Frantz et al., 2009) leads us to give consideration to
the approach proposed by Pritchard et al. (2000). It
had been shown that unbalanced sample numbers
(unlikely to be avoided when a rare and secretive
species is being dealt with) may influence the results
of an admixture analysis implemented in STRUC-
TURE (Kalinowski, 2011). Accordingly, sensu Aurelle
& Ledoux (2013) with respect to simulative data, we
employed the mixture analysis implemented in BAPS
because, in some cases, this outperformed other
methods when it came to identifying the real number
of clusters. The result for six/seven clusters (BAPS/
STRUCTURE with the method from Pritchard et al.,
2000, respectively) is very much in concordance with
geographical data, and appears very reasonable,
whereas the scenario involving two genetic clusters
(with the Evanno method) is in accordance with the
division of Poland visible in mitochondrial data. It
may be that both scenarios are true because investi-
gation of the proposed divide indicates that they do
not preclude each other. This may represent a case of
historically hierarchical structuring; the older one, in
evolutionary terms, being reflected in a two-cluster
scenario and connected to the history of colonization
of Poland, in particular, and Central Europe in
general, by the species, and the six/seven cluster
scenario with further differentiation arising in iso-
lated populations. Indeed, it has been suggested that
Evanno’s delta Kmethod identifies the highest level
of population structure (Waples & Gaggiotti, 2006)
and Kmay therefore be underestimated when there is
a hierarchical structure.
CONCLUSIONS
The basic conclusion to be drawn from the present
study is that there were at least two different coloni-
zation routes for Poland, and probably Central
SMOOTH SNAKE PHYLOGEOGRAPHY AND POPULATION STRUCTURE 13
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2015, ••, ••–••
Europe, originating in at least two different refugia,
as reflected in a west/south-east divide of Poland
visible in the haplotype distribution and topology of
the network, and also in the detected population-
structuring in the admixture analysis of
microsatellite data. The well supported designation of
the Western Europe Clade suggests that, apart from
these two refugia, another refugium might have
occurred, from which Western Europe has been colo-
nized. Clarification of this issue will, however, require
further research. Low allelic richness and levels of
heterozygosity, and high inbreeding, as well as low
Garza–Williamson indices, point to a reduced gene
pool, and suggest size reduction at the population
level.
ACKNOWLEDGEMENTS
We thank Grzegorz Hebda, Stanisław Pagacz,
Przemysław Stachyra, Piotr Zielin´ ski, Michał
Z
˙mihorski, and Professor Eduard G. Yavruyan and
his group for sharing with us their samples of
C. austriaca; Jan Pomorski for valuable advice in the
field of phylogeny; and Łukasz Kajtoch and Ronald A.
Van Den Bussche for their valuable comments on the
manuscript. We thank three anonymous reviewers for
their helpful comments. We also thank Włocławek
Forest Inspectorate and Józefów Forest Inspectorate
for their logistic support. The study was founded by
grant number N N303 570539 from the Ministry of
Science and Higher Education/National Science
Centre.
REFERENCES
Amo L, López P, Martín J. 2004. Chemosensory recognition
of its lizard prey by the ambush smooth snake, Coronella
austriaca.Journal of Herpetology 38: 451–454.
Arnold N, Ovenden D. 2002. Reptiles and amphibians of
Europe. Princeton, NJ: Princeton University Press.
Aurelle D, Ledoux J-B. 2013. Interplay between isolation by
distance and genetic clusters in the red coral Corallium
rubrum: insights from simulated and empirical data. Con-
servation Genetics 14: 705–716.
Avise JC. 2004. Molecular markers, natural history, and
evolution, 2nd edn. Sunderland, MA: Sinauer.
Bandelt H-J, Forster P, Röhl A. 1999. Median-joining net-
works for inferring intraspecific phylogenies. Molecular
Biology and Evolution 16: 37–48.
Bond M, Porteous R, Hughes S, Mogg RJ, Gardner MG,
Reading CJ. 2005. Polymorphic microsatellite markers,
isolated using a simple enrichment procedure, in the threat-
ened smooth snake (Coronella austriaca). Molecular Ecology
Notes 5: 42–44.
Brown DS, Ebenezer KL, Symondson WOC. 2013.
Molecular analysis of the diets of snakes: changes in prey
exploitation during development of the rare smooth snake
Coronella austriaca.Molecular Ecology 23: 3734–3743.
Corander J, Sirén J, Arjas E. 2008. Bayesian spatial mod-
elling of genetic population structure. Computational Sta-
tistics 23: 111–129.
Cornuet JM, Luikart G. 1996. Description and power analy-
sis of two tests for detecting recent population bottlenecks
from allele frequency data. Genetics 144: 2001–2014.
Da˛ browski M, Pilot M, Kruczyk M, Z
˙mihorski M, Umer
HM, Gliwicz J. 2013. Reliability assessment of null allele
detection: inconsistencies between and within different
methods. Molecular Ecology Resources 14: 361–373.
David P, Pujol B, Viard F, Castella V, Goudet J. 2007.
Reliable selfing rate estimates from imperfect population
genetic data. Molecular Ecology 16: 2474–2487.
Di Rienzo A, Peterson AC, Garza JC, Valdes AM, Slatkin
M, Freimer NB. 1994. Mutational processes of simple
sequence repeat loci in human populations. Proceedings of
the National Academy of Sciences of the United States of
America 91: 3166–3170.
Dodd CK Jr. 1993. Strategies for snake conservation. In:
Seigel RA, Collins JT, eds. Snakes – ecology and behavior.
New York, NY: McGraw Hill, 363–394.
Drobenkov SV. 2000. Ecology of the smooth snake
(Coronella austriaca Laur.) in Belarus. Russian Journal of
Herpetology 7: 135–138.
Earl DA, von Holdt BM. 2012. STRUCTURE HARVESTER:
a website and program for visualizing STRUCTURE output
and implementing the Evanno method. Conservation Genet-
ics and Resources 4: 359–361.
Evanno G, Regnaut S, Goudet J. 2005. Detecting the
number of clusters of individuals using the software
STRUCTURE: a simulation study. MolecularEcology 14:
2611–2620.
Excoffier L, Lischer HEL. 2010. Arlequin suite ver 3.5: a
new series of programs to perform population genetics
analyses under Linux and Windows. Molecular Ecology and
Resources 10: 564–567.
Fahrig L. 2003. Effects of habitat fragmentation on biodiver-
sity. Annual Review of Ecology, Evolution and Systematics
34: 487–515.
Frankham R. 2005. Genetics and extinction. Biological Con-
servation 126: 131–140.
Frantz AC, Cellina S, Krier A, Schley L, Burke T. 2009.
Using spatial Bayesian methods to determine the genetic
structure of a continuously distributed population: clusters
or isolation by distance? Journal of Applied Ecology 46:
493–505.
Galán P. 1988. Segregación ecológica en una comunidad de
ofidios. Doñana, Acta Vertebrata 15: 59–78.
Garza JC, Williamson EG. 2001. Detection of reduction in
population size using data from microsatellite loci. Molecu-
lar Ecology 10: 305–331.
Gibbs HL, Prior K, Weatherhead PJ, Johnson G. 1997.
Genetic structure of populations of the threatened eastern
massasauga rattlesnake, Sistrurus c. catenatus: evidence
from microsatellite DNA markers. Molecular Ecology 6:
1123–1132.
14 A. SZTENCEL-JABŁONKA ET AL.
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2015, ••, ••–••
Głowacin´ ski J, Rafin´ ski Z, eds. 2003. Atlas of the
amphibians and reptiles of Poland. Status–distribution–
conservation [with abstracts in English], Warszawa-
Kraków: Inspekcja Ochrony S
´odowiska.
Goudet J. 2001. FSTAT, a program to estimate and test gene
diversities and fixation indices, Version 2.9.3. Available at:
http://www2.unil.ch/popgen/softwares/fstat.htm.
Guillot G, Leblois R, Coulon A, Frantz AC. 2009. Statis-
tical methods in spatial genetics. Molecular Ecology 18:
4734–4756.
Guindon S, Dufayard JF, Lefort V, Anisimova M, Hordijk
W, Gascuel O. 2010. New algorithms and methods to
estimate maximum-likelihood phylogenies: assessing the
performance of PhyML 3.0. Systematic Biology 59: 307–321.
Gvoždík V, Jandzik D, Lymberakis P, Jablonski D,
Moravec J. 2010. Slow worm, Anguis fragilis (Reptilia:
Anguidae) as a species complex: genetic structure reveals
deep divergences. Molecular Phylogenetics and Evolution
55: 460–472.
Hamer AJ, McDonnell MJ. 2010. The response of
herpetofauna to urbanization: inferring patterns of persis-
tence from wildlife databases. Austral Ecology 35: 568–580.
Hammer Ø, Harper DAT, Ryan PD. 2001. PAST:
paleontological statistics software package for education
and data analysis. Palaeontologia Electronica 4: 1–9.
Hewitt GM. 2000. The genetic legacy of the Quaternary ice
ages. Nature 405: 907–913.
Hewitt GM. 2004. Genetic consequences of climatic oscilla-
tions in the Quaternary. Philosophical Transactions of the
Royal Society of London Series B, Biological Sciences 359:
183–195.
Hillis DM, Bull JJ. 1993. An empirical test of bootstrapping
as method for assessing confidence in phylogenetic analysis.
Systematic Biology 42: 182–192.
Jakobsson M, Rosenberg NA. 2007. CLUMPP: a cluster
matching and permutation program for dealing with label
switching and multimodality in analysis of population struc-
ture. Bioinformatics 23: 1801–1806.
Joint National Conservation Committee. 2010. UK Pri-
ority Species data collation Coronella austriaca version 2
updated on 15/12/2010. Available at: http://jncc.defra.gov.uk/
_speciespages/2194.pdf
Jombart T, Devillard S, Balloux F. 2010. Discriminant
analysis of principal components: a new method for the
analysis of genetically structured populations. BMC Genet-
ics 11: 94.
Kalinowski ST. 2011. The computer program STRUCTURE
does not reliably identify the main genetic clusters within
species: simulations and implications for human population
structure. Heredity 106: 625–632.
Keller LF, Waller DM. 2002. Inbreeding effects in wild
populations. Trends in Ecology and Evolution 17: 230–241.
Keyghobadi N. 2007. The genetic implications of habitat
fragmentation for animals. Canadian Journal of Zoology 85:
1049–1064.
Kindler C, Böhme W, Corti C, Gvoždík V, Jablonski D,
Jandzik D, Metallinou M, Široký P, Fritz U. 2013.
Mitochondrial phylogeography, contact zones and taxonomy
of grass snakes (Natrix natrix, N. megalocephala).Zoologica
Scripta 42: 458–472.
Kocher TD, Thomas WK, Meyer A, Edwards SV, Pääbo
S, Villablanca F, Wilson AC. 1989. Dynamics of
mitochondrial DNA evolution in animals: amplification and
sequencing with conserved primers. Proceedings of the
National Academy of Sciences of the United States of
America 86: 6196–6200.
Krauss J, Bommarco R, Guardiola M, Heikkinen RK,
Helm A, Kuussaari M, Lindborg R, Ockinger E, Pärtel
M, Pino J, Pöyry J, Raatikainen K, Sang A, Stefanescu
C, Teder T, Zobel M, Steffan-Dewenter I. 2010. Habitat
fragmentation causes immediate and time-delayed biodiver-
sity loss at different trophic levels. Ecology Letters 13:
597–605.
Luikart G, Cornuet JM. 1998. Empirical evaluation of a test
for identifying recently bottlenecked populations from allele
frequency data. Conservation Biology 12: 228–237.
Mantel N. 1967. The detection of disease clustering and a gen-
eralized regression approach. Cancer Research 27: 209–220.
Masel J. 2011. Genetic drift. Current Biology 21: R837–R838.
Nagy ZT, Lawson R, Joger U, Wink M. 2004. Molecular
systematics of racers, whipsnakes and relatives (Reptilia:
Colubridae) using mitochondrial and nuclear markers.
Philosophical Transactions of the Royal Society of London
Series B, Biological Sciences 42: 223–233.
van Oosterhout C, Hutchinson WF, Wills DPM, Shipley
P. 2003. Micro-checker, Version 2.2.3. Department of Bio-
logical Sciences and Department of Computer Science, Uni-
versity of Hull, Hull. Available at: http://www.microchecker
.hull.ac.uk
Palsbøll PJ, Bérubé M, Allendorf FW. 2007. Identification
of management units using population genetic data. Trends
in Ecology and Evolution 22: 11–16.
Peakall R, Smouse PE. 2006. GENALEX 6: genetic analysis
in Excel. Population genetic software for teaching and
research. Molecular Ecology Notes 6: 288–295.
Peakall R, Smouse PE. 2012. GenAlEx 6.5 genetic analysis
in Excel. Population genetic software for teaching and
research – an update. Bioinformatics 28: 2537–2539.
Pernetta AP, Allen JA, Beebee TJC, Reading CJ. 2011.
Fine-scale population genetic structure and sex-biased dis-
persal in the smooth snake (Coronella austriaca) in south-
ern England. Heredity 107: 231–238.
Posada D. 2008. jModelTest: phylogenetic model averaging.
Molecular Biology and Evolution 25: 1253–1256.
Pritchard JK, Stephens M, Donnelly P. 2000. Inference of
population structure using multilocus genotype data. Genet-
ics 155: 945–959.
Rambaud A, Suchard M, Drummond AJ. 2013. Tracer,
Version l.6. Available at: http://beast.bio.ed.ac.uk/Tracer
Raymond M, Rousset F. 1995. GENEPOP (version 1.2):
population genetics software for exact tests and
ecumenicism. Journal of Heredity 86: 248–249.
Reading C, Jofré G. 2013. Diet composition changes corre-
lated with body size in the smooth snake, Coronella
austriaca, inhabiting lowland heath in southern England.
Amphibia-Reptilia 34: 463–470.
SMOOTH SNAKE PHYLOGEOGRAPHY AND POPULATION STRUCTURE 15
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2015, ••, ••–••
Reading CJ. 2004. Age, growth and sex determination in a
population of smooth snakes, Coronella austriaca in south-
ern England. Amphibia-Reptilia 25: 137–150.
Reading CJ. 2012. Ranging behaviour and home range size
of smooth snakes inhabiting lowland heath in southern
England. Herpetological Journal 22: 241–247.
Reading CJ, Luiselli LM, Akani GC, Bonnet X, Amori G,
Ballouard JM et al. 2010. Are snake populations in wide-
spread decline? Biological Letters 6: 77–778.
Rice WR. 1989. Analyzing tables of statistical tests. Evolu-
tion 43: 223–225.
Ronquist F, Teslenko M, van der Mark P, Ayres D,
Darling A, Höhna S et al. 2012. MrBayes 3.2: efficient
Bayesian phylogenetic inference and model choice across a
large model space. Systematic Biology 61: 539–542.
Rousset F. 1997. Genetic differentiation and estimation of
gene flow from F-statistics under isolation by distance.
Genetics 145: 1219–1228.
Rousset F. 2008. Genepop’007: a complete reimplementation
of the Genepop software for Windows and Linux. Molecular
Ecology Resources 8: 103–106.
Rowe G, Harris DJ, Beebee TJC. 2006. Lusitania revisited:
a phylogeographic analysis of the natterjack toad Bufo
calamita across its entire biogeographical range. Molecular
Phylogenetics and Evolution 39: 335–346.
Saccheri I, Kuussaari M, Kankare M, Vikman P,
Wilhelm F, Hanski I. 1998. Inbreeding and extinction in a
butterfly metapopulation. Nature 392: 491–494.
Santos X, Brito JC, Caro J, Abril AJ, Lorenzo M, Sillero
N, Pleguezuelos JM. 2009. Habitat suitability, threats and
conservation of isolated populations of the smooth snake
(Coronella austriaca) in the southern Iberian Peninsula.
Biological Conservation 142: 344–352.
Santos X, Rato C, Carranza S, Carretero MA,
Pleguezuelos JM. 2012. Complex phylogeography in the
southern smooth snake (Coronella girondica) supported by
mtDNA sequences. Journal of Zoological Systematics and
Evolutionary Research 50: 210–219.
Santos X, Roca J, Pleguezuelos JM, Donaire D,
Carranza S. 2008. Biogeography and evolution of the
smooth snake Coronella austriaca (Serpentes: Colubridae)
in the Iberian Peninsula: evidence for Messinian refuges
and Pleistocenic range expansions. Amphibia-Reptilia 29:
35–47.
Sewell D, Guillera-Arroita G, Griffiths RA, Beebee TJC.
2012. When is a species declining? Optimizing survey effort
to detect population changes in reptiles. PLoS ONE 7:
e43387.
Short KH, Petren K. 2011. Fine-scale genetic structure
arises during range expansion of an invasive gecko. PLoS
ONE 6: e26258.
Spear SF, Peterson CR, Matocq MD, Storfer A. 2006.
Molecular evidence for historical and recent population size
reductions of tiger salamanders (Ambystoma tigrinum)in
Yellowstone National Park. Conservation Genetics 7: 605–
611.
Strijbosch H. 1997. Coronella austriaca Laurenti, 1768. In:
Gasc JP, Cabela A, Crnobrja-Isailovic J, Dolmen D,
Grossenbachner K, Haffner P et al, eds. 344–345. Atlas of
amphibians and reptiles in Europe. Paris: Societas Europea
Herpetologica & Muséum National d’Historie Naturelle
(IEGB/SPN).
Taberlet P, Fumagalli L, Wust-Saucy AG, Cosson J. 1998.
Comparative phylogeography and post glacial colonization
routes in Europe. Molecular Ecology 7: 453–464.
Tamura K, Stecher G, Peterson D, Filipski A, Kumar S.
2013. MEGA6: molecular Evolutionary Genetics Analysis
version 6.0. Molecular Biology and Evolution 30: 2725–
2729.
Tilman D, Fargione J, Wolff B, D’Antonio C, Dobson A,
Howarth R. 2001. Forecasting agriculturally driven global
environmental change. Science 292: 281–284.
Tilman D, May RM, Lehman CL, Nowak MA. 1994.
Habitat destruction and the extinction debt. Nature 371:
65–66.
Ursenbacher S, Carlsson M, Heifer V, Tegelstrom H,
Fumagalli L. 2006. Phylogeography and Pleistocene
refugia of the adder (Vipera berus) as inferred from
mitochondrial DNA sequence data. Molecular Ecology 15:
3425–3437.
Ursenbacher S, Erny C, Fumagalli L. 2009. Male repro-
ductive success and multiple paternity in wild, low-density
populations of the adder (Vipera berus). Journal of Heredity
100: 365–370.
Waples RS, Gaggiotti O. 2006. What is a population? An
empirical evaluation of some genetic methods for identifying
the number of gene pools and their degree of connectivity.
Molecular Ecology 15: 1419–1439.
SHARED DATA
All sequences obtained were submitted to GenBank (accession numbers: KP756615 to KP756624). The genotypic
data with sampling localities are available at the Research Gate (https://www.researchgate.net/profile/
Anna_Sztencel-Jabtonka).
SUPPORTING INFORMATION
Additional Supporting Information may be found in the online version of this article at the publisher’s web-site:
Figure S1. Delta Kvalues in relation to the number of clusters (K).
Figure S2. Distribution of posterior probability Pr(X|K) for the subsequent number of clusters.
16 A. SZTENCEL-JABŁONKA ET AL.
© 2015 The Linnean Society of London, Biological Journal of the Linnean Society, 2015, ••, ••–••
