Mountainous genus Anterastes (Orthoptera, Tettigoniidae): autochthonous survival across several glacial ages via vertical range shifts

Zoologica Scripta Mountainous genus Anterastes (Orthoptera, Tettigoniidae): autochthonous survival across several glacial ages via vertical range shifts € € UZ BATTAL C ß IPLAK, SARP KAYA, ZEHRA BOZTEPE & I_ SLAM GUND

Submitted: 3 December 2014 Accepted: 8 April 2015 doi:10.1111/zsc.12118

_ (2015). Mountainous genus Anterastes C und€ uz, I.. ß ıplak, B., Kaya, S., Boztepe, Z., G€ (Orthoptera, Tettigoniidae): autochthonous survival across several glacial ages via vertical range shifts. —Zoologica Scripta, 00, 000–000. Although the high-latitude range margins in Europe and North America are intensively studied, attention is gradually turned towards the taxa/populations inhabiting glacial refugia. Here, we evaluate the genealogical history of the cold-adapted Anatolio-Balkan genus Anterastes especially to test the possible effects of intrarefugial vertical range shifts during climatic oscillations of the Quaternary. Using concatenated data from sequences of COI+16S and ITS1–5.8S–ITS2, intrageneric relationships and the time of speciation events were estimated. Thirteen different demographic analyses were performed using a data set produced from sequences of 16S. Different phylogenetic analyses recovered similar lineages with high resolution. The molecular chronogram estimated speciation events in a period ranging from 5.60 to 1.22 Myr. Demographic analyses applied to 13 populations and five lineages suggested constant population size. Genetic diversity is significantly reduced in a few populations, while not in others. Fixation indices suggested extremely diverged populations. In the light of these data, the following main conclusions were raised: (i) although glacial refugia are the biodiversity hotspots, species level radiation of the cold-adapted lineages is mainly prior to the Mid-Pleistocene transition; (ii) heterogeneous topography provides refugial habitats and allows populations to survive through vertical range shifts during climatic fluctuations; (iii) prolonged isolation of refugial populations do not always result in reduced intrapopulation diversity, but in high level of genetic differentiation; (iv) the cold-adapted lineages with low dispersal ability might have not colonised the area out of Anatolian refugium during interglacial periods; and (v) populations of invertebrates may have restricted ranges, but this does not mean that they have small effective population size. Corresponding author: Battal C ß ıplak, Department of Biology, Faculty of Science, Akdeniz University, 07058 Antalya, Turkey. E-mail: [email protected] Battal C ß ıplak, Sarp Kaya, and Zehra Boztepe, Department of Biology, Faculty of Science, Akdeniz University, Antalya, Turkey. E-mail:[email protected] and [email protected] _ Islam G€und€ uz, Department of Biology, Faculty of Art & Science, Ondokuz Mayıs University, Samsun, Turkey. E-mail:[email protected]

Introduction Molecular and analytical tools have greatly improved in recent years. Gene-based phylogenies allow characterisation of the evolutionary history of lineages in combination with geographic/climatologic evolution of their range areas (Avise 2009). The vast majority of phylogeographic studies mostly focus on the European and North American lineages (Hewitt 1996, 2000; Schmitt 2007). Although Anatolia is considered the origin for several Mediterranean taxa/ lineages and an important biodiversity hotspot in the West Palaearctic (Myers et al. 2000), the phylogeography of the Anatolian lineages remains understudied and its potential

ª 2015 Royal Swedish Academy of Sciences

for being a glacial refugium has only been recently acknowledged (Ansell et al. 2011; Sa glam et al. 2013; Korkmaz et al. 2014; Kaya et al. 2013; Kaya et al. 2015). The simplest generalisation developed in parallel with the phylogeographic theory is that the range shifts of species/populations caused by climatic fluctuations. Two main patterns of range shifts have been proposed: (i) horizontal or latitudinal and (ii) vertical or altitudinal. The horizontal pattern assumes that a range of species/populations shift from glaciated areas towards ice-/frost-free refugia during glacial periods and in a reverse direction during interglacials (Hewitt 1996; Schmitt 2007). The same is also 1

Phylogeography of Anterastes




B. C ß ıplak et al.

expected to occur within a refugium (C glam ß ıplak 2004; Sa et al. 2013; Kaya et al. 2015). The vertical pattern proposes that species/populations inhabiting intrarefugial highlands extended their range towards lower elevations at each cooling period, but receding again to the summits at each interglacial. The horizontal pattern mainly concerns range changes in non-refugial areas and possibly trivial for the intrarefugial range shifts. But the vertical pattern is expected to be experienced by almost all populations inhabiting a refugium, thus more important for the evolution of refugial biodiversity. Well-documented examples on the larger scale are provided by European mountain ranges (Schmitt 2009). It has already been shown that many contemporary coldadapted species with disjunct distribution patterns are postice age relicts (Habel & Assmann 2011). However, there is still need for studies on disjunct species and their phylogeography on a smaller scale in other areas. There is a strong relation between range shifts and variable topography. In Anatolia, the mean altitude is around 1100 m, with thousands of summits higher than 2000 m and some even above 3000 m (http://peakery.com/region/ turkey-mountains/). Several summits were glaciated during glacial periods (Sarikaya et al. 2011), and such summits may be considered as intrarefugial arctic poles from the perspective of Quaternary biogeography (C ß ıplak 2008). Thus, Anatolia serves as a natural laboratory to study altitudinal range shifts during climatic oscillations (C ß ıplak 2004, 2008; S ß ekercio glu et al. 2011). Although there are phylogeographic studies on Anatolian or East Mediterranean lineages/populations (e.g. Veith et al. 2003; Mutun 2010; Ansell et al. 2011; Kaya et al. 2015; Korkmaz et al. 2014), persistence of refugial populations across several glacial periods has been rarely studied (Sa glam et al. 2013) and needs elaboration. The genus Anterastes Brunner von Wattenwyl (Orthoptera, Tettigoniidae) is almost strictly Anatolian, with 13 species present in Anatolia and one species (A. serbicus) both in Anatolia and the Balkans (Eades et al. 2014). All species have disjunct distributions in highland meadows indicating their habitat preference and sensitivity to climatic parameters. Several of the species are known from a single summit in a very restricted area and those occurring on more than one summit also have isolated patchy populations (C ß ıplak 2004; C ß ıplak et al. 2010; Kaya & C ß ıplak 2011; Kaya et al. 2012a). Because of being flightless (short wings in male function as a stridulatory organ only) and univoltine occurring in a short summer period, they have limited dispersal ability. Based on these features, C ß ıplak (2004) suggested that speciation events might have been directed by glacial periods of the Pleistocene. A phylogeographic study of the A. serbicus species group suggested an earlier radiation period, but still a pattern strictly correlated with mountain belts in Anatolia and the Balkans (C ß ıplak

2

et al. 2010). These features make the genus a good candidate of a model lineage to test altitudinal range shifts. In this study, we examine the inter- and intraspecies disjunct distribution pattern of Anterastes to provide insights into how species and populations have previously responded to episodes of climatic change and the way in which these have shaped the distribution of genetic diversity across the distribution ranges. We will construct a molecular phylogeny using two different mitochondrial and two nuclear markers. We will then benefit from a molecular clock estimation to correlate speciation steps with historical geographic events. If the time estimation indicates that species/populations of Anterastes survived across two or more glacial periods, our null hypothesis will be that the survival during glacial periods of Quaternary is achieved by vertical range shift and each species/population evolved autochthonous in their highlands. As independence of lineages should increase with divergence time and independent range shifts limit gene flow (Avise 2000; Knowles & Carstens 2007), the following results will be considered as support to the null hypothesis: (i) lack of long-distance dispersals, (ii) absence of hybridisation events between species or population of a species, thus phylogenetic independence of each species/populations, (iii) correlation of genetic similarity or phylogenetic relationship pattern of the species/population with geographic proximity, (iv) species/highland populations having unique gene pools and (v) historically or relatively stable population sizes. The contrary results will reject the null hypothesis. Finally, the results regarding range extent and genetic diversity indices will be discussed in regard to conservation of Anterastes.

Material and methods Sampling, DNA extraction, PCR amplification and sequencing Samples of Anterastes were collected from Anatolia and Balkans during 2007–2011. Twenty-six locations were sampled, each from an isolated mountain population (Table 1, Fig. 1). Specimens were identified according to C ß ıplak (2004), Kaya & C ß ıplak (2011) and Kaya et al. (2012a), and the cryptic species were denoted by ‘cf. the closest morphospecies name’ (for taxonomical remarks, see Appendix S1). The sequences of two mitochondrial (16S rDNA and cytochrome c oxidase subunit 1 – COI) and two nuclear (internal transcribed spacers and 5.8S rDNA in between – ITS1–5.8S–ITS2) markers were used in phylogenetic, time estimation, population genetics and demographic analyses. MtDNA genes were amplified following the reaction procedure given by C ß ıplak et al. (2010) and the nuclear genes described in Kaya et al. (2013) (see Appendix S1 for details). Nucleotide sequences of each unique haplotype identified were deposited in the GenBank database (for accession numbers, see Table S1).

ª 2015 Royal Swedish Academy of Sciences

B. C ß ıplak et al.




Phylogeography of Anterastes

Table 1 Sampling details (see Fig. 1) of specimens used for DNA studies (except three locations from Bulgaria, all others are from Anatolian Turkey) Locality/Population

Latitude 0

Longitude 0

Elevation (m)

Abb.

Species

Bolu Antalya Isparta Bursa

Kartalkaya Mt. Akdag Mt. Davraz Mt. Uludag Mt.

40°35 47.0″N 36°340 56.4″N 37°460 18.7″N 40°070 18.2″N

31°46 59.3″E 29°340 49.9″E 30°440 34.2″E 29°080 57.0″E

2001 2247 1950 1903

BKdis AAant IDdav BUulu

A. A. A. A.

disparalatus antecessor davrazensis uludaghensis

Antalya Tokat Sivas Denizli Antalya Antalya Konya Isparta Antalya Antalya

Tahtalıdag Mt. Cß amlıbel pass K€osedag Babadag Mt. Begisß Erentepe Mt. Kızkayası Davraz Mt. _Imecik yayla Cß amkuyusu

26°320 12.4″N 39°570 48.6″N 40°090 39.3″N 37°470 45.8″N 36°560 10.6″N 36°440 22.1″N 38°050 19.7″N 37°460 18.5″N 36°500 16.3″N 36°360 11.0″N

30°260 21.7″E 36°300 27.1″E 37°510 02.5″E 28°460 03.7″E 30°050 38.2″E 29°380 27.4″E 32°160 56.0″E 30°440 34.9″E 30°180 43.2″E 29°370 49.4″E

1850 1698 1809 1450 1309 1982 1644 1630 1790–2140 2177

ATtur TCnig SKnig DBbab ABbab AEbab KKcuca IDuca AIuca ACuca

A. A. A. A. A. A. A. A. A. A.

turcicus niger niger babadaghi babadaghi babadaghi cf. ucari ucari ucari ucari

Bulgaria

Rila Mt. Stara Mt. Osogovska Mt. Ilgaz Mt. ßSebinkarahisar K€osedag Mt. Cß erkesß, Isßık Mt. Uludag Mt. Murat Mt. Bozdag Mt. Kartalkaya Mt. Seydisßehir Sandıklı, ßSuhut Tavsßanlı

41°520 11.5″N 42°290 34.2″N 41°570 56.4″N 41°030 24.3″N 40°270 46.2″N 40°090 29.7″N 40°410 27.8″N 40°070 19.2″N 38°560 54.7″N 38°210 15.1″N 40°350 47.4″N 41°030 24.0″N 38°280 08.9″N 37°010 30.7″N

23°500 36.6″E 25°190 07.8″E 23°200 35.3″E 33°430 05.8″E 38°420 48.3″E 37°510 02.6″E 32°440 06.4″E 29°080 57.3″E 29°370 29.4″E 28°060 14.8″E 31°460 52.6″E 33°430 05.7″E 30°220 39.1″E 33°160 45.4″E

2400–2850 1800 2100–2200 2045 2294 1802 1595 1734 1785 1478 2001 1551 1802 1386

BRser

A. serbicus

KIcser GScser SKcser CCcser BUbur UMcbur IBtol BKant KSant ASant KTant

A. A. A. A. A. A. A. A. A. A. A.

Kastamonu Giresun Sivas Cß ankırı Bursa Usßak _Izmir Bolu Konya Afyon Karaman

DNA sequence evaluation and descriptive genetic patterns Sequences were checked, edited and visually optimised using SEQUENCHER v.4.1.4 (GeneCodes Corp., Ann Arbor, MI, USA). Multiple alignment of the each segment was accomplished using ClustalW as implemented in MEGA v.5 (Tamura et al. 2011) with default parameter settings and checked by eye. Each of the COI sequences was checked by DNASP v.5 (Librado & Rozas 2009) to determine unique haplotypes and to calculate their frequency. To exclude possible numts (nuclear mitochondrial pseudogenes), the COI nucleotide sequences were translated to amino acid sequences to determine the presence of any stop codons or non-sense mutations using DNASP v.5. The obtained nucleotide and amino acid sequences were checked for homology and pseudogenes using the NCBI BLAST program (http:// www.ncbi.nlm.nih.gov/BLAST/). Three different data matrices were generated: the first from sequences of 16S, the second from concatenated sequences of COI and 16S and the third from ITS1–5.8S– ITS2. The first matrix was used for descriptive genetic and demographic analyses, while the second and third were used for phylogenetic and time estimation analyses. The

ª 2015 Royal Swedish Academy of Sciences

A. babadaghi species group

A. serbicus species group

cf. serbicus cf. serbicus cf. serbicus cf. serbicus burri cf. burri tolunayi antitauricus antitauricus antitauricus antitauricus

matrix of 16S sequences was used to calculate the number of haplotypes, either with gaps (K) or without gaps (k), haplotype diversity (h), the number of polymorphic sites (S), nucleotide diversity (p) and the mean number of pairwise differences between K haplotypes for each population under ARLEQUIN v.3.5.2 (Excoffier et al. 2005). F statistics (FST) were calculated and tested by 1000 permutations in ARLEQUIN v.3.5.2 to describe population genetic structure. Phylogenetic analyses Phylogenetic analyses were performed separately using data matrices of sequences COI+16S and ITS1–5.8S–ITS2. To establish a concatenated data set from 16S and COI sequences of the same specimen, we performed incongruence length difference (ILD) test [also named the partition homogeneity test (Farris et al. 1994)] using PAUP v.4.10b (Swofford 2000). The nucleotide composition, the number of variable and parsimony informative sites, and the transition/transversion ratio were calculated by MEGA v.5. The matrices of COI+16S and ITS1–5.8S–ITS2 were analysed under maximum parsimony (MP), maximum likelihood (ML) and Bayesian inference (BI) approaches (for

3

Phylogeography of Anterastes




B. C ß ıplak et al.

Fig. 1 Sampling localities of the Anterastes (for details, see Table 1; Anatolian diagonal showed by dashed lines).

details, see Appendix S1). Anadolua schwarzi Ramme, Parap€ holidoptera distincta Karaba g and Bolua turkiyae Unal were included as out-groups, the first two for ITS1–5.8S–ITS2 and the last two for COI+16S (see Appendix S1 for details). Estimation of divergence times The time estimations of the speciation/divergence events within Anterastes lineage are crucial for the study. We established a data set of two sequence blocks as the mitochondrial 16S+COI and the nuclear ITS1–5.8S–ITS2 and analysed as ‘unlink’ in a simultaneous run of BEAST v.1.7.2 (Drummond & Rambaut 2007). We first tested diversification rates and rate shifts for COI+16S sequences using the topological program SYMMETREE v.1.1 (Chan & Moore 2005) (for application of the analysis, see Appendix S1). We also applied a ML scores to nuclear ITS1–ITS2 sequences using MEGA v. 5 (Tamura & Nei, 1993; Tamura et al. 2011) to determine whether they behave clock-like or not. The nodes of the main mitochondrial and nuclear clades suggested by the each data set were dated using a Bayesian coalescence approach, as implemented in BEAST. The time to the most recent common ancestor (TMRCA) for each clade was estimated under the model parameters highlighted in MODELTEST applied to each of the mitochondrial and nuclear data block separately (for application of BEAST, see Appendix S1).

4

Sequences two mitochondrial loci (COI and 16S) were concatenated in the matrix. Substitution rate estimations for COI in insects range from 1.68% to 3.15% per site/ million years (Percy et al. 2004; Papadopoulou et al. 2010; Allegrucci et al. 2011). Even so, there are extreme rate estimations for COI running in to 10 % per site/million years (Shapiro et al. 2006). The rates for 16S are around 0.54-0.7 % per site/million years (Brower 1994; Papadopoulou et al. 2010; Allegrucci et al. 2011). We calibrated BEAST with %2.3 divergence rate per site per million years for concatenated matrix to allow for a faster average sequence divergence rate. This conclusion was made because of two reasons. First, the rates for COI, which constitute the main amount of base pairs of the matrix, are higher than that of 16S. Second, for our hypotheses it is important to determine if speciation events are prior or during glacial cycles and a faster rate will allow us a more robust time estimation in this respect. There are controversies for divergence rate of ITS1–ITS2, and these estimations vary from 0.00172 to 0.0258 per site/million years in eukaryotes (Kasuga et al. 2002; Bargues et al. 2006) and from 0.0025 to 0.0258 in insects (Percy et al. 2004). As the lower limit for insects produced consistent estimations with COI in other Orthoptera (Kaya et al. 2013), molecular clock analysis of ITS1–ITS2 sequences was calibrated by this rate (for application of BEAST, see Appendix S1).

ª 2015 Royal Swedish Academy of Sciences

B. C ß ıplak et al.

Demographic analyses Several different approaches were applied to examine the past dynamics of the populations and internal phylogroups of Anterastes using the 16S rDNA sequences, the marker for which we have sufficient samples. Firstly, Tajima’s D (Tajima 1989), Fu’s FS (Fu 1997), Fu & Li’s F* (Fu & Li 1993) and Ramos-Onsins & Rozas’ (2002) R2 tests for mutation/drift equilibrium were conducted to estimate past population expansions and bottlenecks. We also used the coalescent-based approach to estimate exponential growth rate (g) for the populations/internal phylogroups. Secondly, demographic history was explored by mismatch distribution analyses of observed haplotype pairwise differences following a sudden population expansion model (Rogers & Harpending 1992). The parameters of the demographic expansion s, h0 and h1 were estimated by a generalised nonlinear least-square approach. A parametric bootstrapping approach (Schneider & Excoffier 1999) with 1000 replicates was used to estimate the significance of the Harpending’s raggedness index (Hri; Harpending 1994) and the sum of squared deviations (SSD) between the observed and expected mismatch distribution as a test statistic. Thirdly, historical dynamics of populations and internal phylogroups were investigated using a coalescence approach of Bayesian skyline plots (BSPs; Drummond et al. 2005). BSP model generates a posterior distribution of effective population size through time using a Markov chain Monte Carlo (MCMC) sampling. This approach is a nonparametric estimate of the changes in population size that makes no a priori assumptions on the demographic model (for further details of demographic analyses, see Appendix S1).

Results Descriptive genetics Final length of sequences after alignment and trimming, their characteristics as numbers of variable, invariable, indels and parsimony informative position and the number of the haplotypes for each of 16S, COI, COI+16S and ITS1– 5.8S–ITS2 region are given in Table 2. Genetic diversity Table 2 Description of the sequences used in analyses (*two are from out-group and others from in-group)

*N of the sequences Total length as base pair after alignment and trimming N of the constant sites N of the variable sites N of the indel sites N of the parsimony informative sites *N of the haplotype with indels *N of the haplotype without indels

ª 2015 Royal Swedish Academy of Sciences

COI

16S

COI+16S

ITS1–5.8S–ITS2

128 1126

341 521

– 1647

136 937

707 419 – 370 – 67

402 119 9 94 91 83

1116 531 – 461 – 67

725 97 115 62 77 42




Phylogeography of Anterastes

indices calculated for 22 populations with a sample size of N ≥ 6 using 16S marker were presented in Table 3. Seven of 22 populations contain a single unique haplotype, and their indices are zero. There are no haplotypes of 16S and COI shared by two or more populations. As there are uncertainties for species status of three phylogroups, FST values were calculated for the population pairs belonging to A. serbicus group (12 populations) and for those to A. babadaghi group (10 populations) separately. Group level pairwise FST values are rarely below 0.5 and mostly run into 1.00 both for the population pairs belonging to A. serbicus (Table 4) and A. babadaghi groups (Table 5). In the ITS1–5.8S–ITS2, intra-individual sequence variation has been found in one specimen of AAant, ASant, AIuca and BKdis, two specimens of BRser and three specimens of BUbur populations. Of the 40 in-group haplotypes of ITS1–5.8S–ITS2, only AIuca-1 is shared by seven populations (Table 1) of A. babadaghi and A. ucari, while all others are unique to the populations sampled. A medianjoining network of haplotypes put this shared haplotype as ancestral for other haplotypes of the A. babadaghi group (Fig. S1). Phylogenetic analyses As the homogeneity test suggested a homogenous evolution (P = 0.5) for both 16S and COI sequences, they were concatenated in a single matrix (Table 2). MODELTEST applied to this matrix suggested GTR+G +I according to Akaike’s information criterion (AIC) with gamma correction (G) of 1.0299 and invariable sites (I) of 0.6222. These parameters were set in PAUP v.4.10b and MRBAYES v.3.1 for ML and BI, respectively. The MP analysis resulted in 16 equally parsimonious trees (tree length = 1820, CI = 0.427; RI = 0.817; RC = 0.349; HI = 0.573). The MP, ML and BI analyses of 67 haplotypes produced trees with similar topologies (Fig. 2). The differences between trees occur in two nodes: (i) the MP and ML suggest A. antecessor + A. disparalatus clade as the most basal node of the in-group and the A. uludaghensis + A. davrazensis clade as the following, while BI suggests the reverse order, each with significant statistical support, and (ii) the ML and BI suggest a high bootstrap/posterior priority values for the monophyly of A. antecessor + A. disparalatus, while MP bootstrap is below 50. All of MP, ML and BI trees suggest a sister phylogroup relationship both for A. antecessor + A. disparalatus and for A. uludaghensis + A. davrazensis occupying the two most basal branches in the in-group (Fig. 2). Descending from these basal nodes, there is a clade including the remaining 54 haplotypes from other populations of Anterastes. This crown clade constitutes two subphylogroups with high bootstrap (MP and ML) and posterior probability (BI) val-

5

Phylogeography of Anterastes




B. C ß ıplak et al.

Table 3 Genetic diversity indices of 26 populations of Anterastes (shown from left to right are sample sizes (N), number, number of haplotypes without indels (k), number of haplotypes including indels (K), haplotype diversity (h), nucleotide diversity (p), mean number of pairwise differences between K haplotypes (p). Indices were calculated using the model TrN+Γ (0.1320). The last two columns show the sample size (N) and number of unique haplotypes (K or k) per populations for COI and ITS1–5.8S–ITS2 (ITS), respectively. NA indicates to populations not analysed. N for ITS1–5.8S–ITS2 haplotypes indicates to the 136 sequenced produced from 126 individuals 16S rDNA

COI p

Population

Species

N

K

k

h

AAant BKdis IDdav BUulu ATtur TCnig SKnig DBbab ABbab AEbab KKcuca IDuca AIuca ACuca BRser KIcser GScser SKcser CCcser BUbur UMcbur IBtol BKant KSant ASant KTant

A. A. A. A. A. A. A. A. A. A. A. A. A. A. A. A. A. A. A. A. A. A. A. A. A. A.

3 8 4 8 14 14 13 23 22 20 18 9 13 9 13 4 6 7 15 19 16 21 13 29 17 3

2 2 2 1 3 3 1 3 9 3 2 3 4 5 8 3 1 1 5 11 1 1 6 4 4 1

2 1 2 1 3 2 1 3 7 3 2 3 4 3 8 3 1 1 5 11 1 1 4 4 4 1

NA 0.2500 NA 0.0000 0.3619 0.4835 0.0000 0.4348 0.9048 0.3526 0.2092 0.6389 0.4231 0.8056 0.8077 NA 0.0000 0.0000 0.6952 0.9240 0.0000 0.0000 0.7179 0.6921 0.5515 NA

antecessor disparalatus davrazensis uludaghensis turcicus niger niger babadaghi babadaghi babadaghi cf. ucari ucari ucari ucari serbicus cf. serbicus cf. serbicus cf. serbicus cf. serbicus burri cf. burri tolunayi antitauricus antitauricus antitauricus antitauricus

NA 0.00048 NA 0.00000 0.00073 0.00264 0.0000 0.00090 0.00811 0.00125 0.00041 0.00236 0.00218 0.00258 0.01358 NA 0.00000 0.00000 0.00378 0.00522 0.00000 0.00000 0.00302 0.00552 0.00194 NA

 0.1802            

0.0000 0.1448 0.1420 0.0000 0.1111 0.0311 0.1227 0.1163 0.1258 0.1645 0.1196 0.1131

        

0.0000 0.0000 0.1092 0.0375 0.0000 0.0000 0.1279 0.0401 0.1163

P

 0.00068            

0.00000 0.00082 0.00193 0.00000 0.00091 0.00466 0.00113 0.00058 0.00187 0.00170 0.00200 0.00766

        

0.00000 0.00000 0.00253 0.00324 0.00000 0.00000 0.00215 0.00332 0.00153

NA 0.25318 NA 0.00000 0.38466 1.38487 0.00000 0.47306 4.26828 0.65905 0.21282 1.23942 1.14516 1.35722 7.11786 NA 0.00000 0.00000 1.98144 2.73523 0.00000 0.00000 1.59051 2.90054 1.01458 NA

 0.31356            

0.00000 0.38318 0.90474 0.00000 0.42681 2.19884 0.52970 0.26921 0.86247 0.79118 0.92206 3.57178

        

0.00000 0.00000 1.18469 1.51724 0.00000 0.00000 1.00853 1.56798 0.71503

ITS

N

k

N

K

k

6 3 3 1 3 1 5 3 22 7 3 4 3 2 8 – 2 8 4 5 6 6 9 5 6 3

6 2 2 1 1 1 3 2 7 4 2 3 2 2 3 – 1 2 2 5 3 2 3 3 2 1

7 6 2 1 1 5 10 2 13 5 4 4 3 11 6 – 5 5 5 10 8 6 5 5 6 1

6 2 1 1 1 1 3 1 7 5 1 4 2 4 4 – 2 2 5 3 5 5 3 2 4 1

4 2 1 1 1 1 2 1 4 3 1 3 1 3 3 – 2 1 2 1 3 1 3 2 4 1

Table 4 Pairwise FST values (below diagonal) and their P-values (above diagonal) for 12 populations in A. serbicus species group [model used = Tamura + G = (0.0110)]

serbicus BRser BRser KIcser GScser SKcser CCcser BUbur UMcbur IBtol BKant KSant ASant KTant

0.72902 0.80296 0.79391 0.79154 0.82686 0.89911 0.91186 0.79224 0.79878 0.82777 0.70715

cf. serbicus KIcser

GScser

SKcser

CCcser