Fossorial Morphotype Does Not Make a Species in Water Voles

Home , Arvicola, Arvicolini, Chionomys, Common vole, Neodon, Proedromys

Mammalia 2015; 79(3): 293–303

Boris Kryštufek, Toni Koren, Simon Engelberger, Győző F. Horváth, Jenő J. Purger, Atilla Arslan, Gabriel Chişamera* and Dumitru Murariu Fossorial morphotype does not make a species in water voles

Abstract: Phenetic and ecological plasticity in Arvicola has caused a long-standing dispute over the number of species Introduction within the genus, which is currently thought to consist of Water voles (genus Arvicola Lacépède, 1799) are extraor- two aquatic (sapidus, amphibius) and one fossorial species dinary among the Palaearctic rodents for their outstand- (scherman). We used mitochondrial cytochrome b (cytb) ing phenetic and ecological plasticity. Two morphological gene sequences to reconstruct phylogenetic relationships types can be distinguished, the aquatic and the fossorial, between the fossorial and the aquatic water voles from each with different living habits (Meylan 1977). This varia- the various regions of their European and Asiatic range. bility has caused a long-standing dispute over the number These two types differed morphologically and exhibited of species within the genus. Whereas Ellerman and Mor- allopatric ranges. Our study provided 50 new haplotypes, rison-Scott (1951) clumped a wide array of morphological generating a total dataset of 70 different water vole cytb forms under one nominal species, A. terrestris (Linnaeus, haplotypes. Phylogenetic reconstructions retrieved two 1758), Hinton (1926) recognized four species and Miller major lineages that were in a sister position to A. sapidus: (1912) listed seven species for Western Europe only. Since a fossorial Swiss lineage and a widespread cluster, which 1970, the majority of authors have accepted a division into contained aquatic and fossorial water voles from Europe two species: A. sapidus Miller, 1908 with a small range in and western Siberia. The phylogeographic architecture in Western Europe, and a widespread and polytypic A. ter- water voles is explained by Quaternary climatic dynamics. restris (Corbet 1978, Niethammer and Krapp 1982 Gromov Our results show that A. scherman in its present scope is and Polyakov 1992, Shenbrot and Krasnov 2005). Within not a monophyletic taxon. the latter, Ognev (1964) distinguished two species, the fossorial A. scherman (Shaw, 1801) and the aquatic A. terrestris. Keywords: Arvicola amphibius; Arvicola scherman; cyto- Ognev’s taxonomic arrangement was largely ignored, but chrome b; molecular phylogeny; Quaternary refugia. gained credit after being adopted in the influential compi- lation by Musser and Carleton (2005). The genus Arvicola DOI 10.1515/mammalia-2014-0059 is currently regarded as consisting of two aquatic (sapidus, Received April 15, 2014; accepted July 3, 2014; previously published amphibius) and one fossorial species (scherman). Musser online August 5, 2014 and Carleton (2005) even anticipated “that future revision- ary research will […] converge towards Miller’s (1912) rec- *Corresponding author: Gabriel Chişamera, Department of ognition of biodiversity”, i.e., in an increase in the species Terrestrial Fauna, “Grigore Antipa” National Museum of Natural History, Sos. Kiseleff no. 1, Bucharest 1, RO-011341, Romania, of water voles. Evidently, our understanding of species e-mail: [email protected] limits in the genus Arvicola is still far from final. Boris Kryštufek: Slovenian Museum of Natural History, Prešernova Arvicola sapidus, which is endemic to the Pyrenean 20, SI-1000 Ljubljana, Slovenia Peninsula and France (Shenbrot and Krasnov 2005), pos- Toni Koren: Science and Research Centre, University of Primorska, sesses a unique set of chromosomes (Matthey 1956) and Garibaldijeva 1, SI-6000 Koper, Slovenia is therefore well defined taxonomically. We shall subse- Simon Engelberger: Museum of Natural History Vienna, Burgring 7, 1010 Wien, Austria quently focus on two species that result from splitting A. Győző F. Horváth and Jenő J. Purger: Institute of Biology, Faculty of terrestris (sensu Corbet 1978) into A. amphibius (Linnaeus, Sciences, Department of Ecology, University of Pécs, H-7624 Pécs, 1758) and A. scherman (Musser and Carleton 2005). The Hungary names amphibius and terrestris are synonymous and the Atilla Arslan: Faculty of Science, Department of Biology, Selçuk former has priority based on the principle of first reviser University, 42031 Konya, Turkey Dumitru Murariu: Department of Terrestrial Fauna, “Grigore Antipa” (Corbet 1978). Both names were used interchangeably in National Museum of Natural History, Sos. Kiseleff no. 1, Bucharest 1, the past. To avoid confusion, we shall subsequently use RO-011341, Romania amphibius sensu stricto (s.s.) in the sense of Musser and 294 B. Kryštufek et al.: Fossorial morphotype does not make a species in water voles

Carleton (2005), and amphibius sensu lato (s.l.) to contain Musser and Carleton’s amphibius s.s. and scherman. The range of Arvicola amphibius s.s. stretches from Western Europe to the River Lena in eastern Siberia (Bat- saikhan et al. 2008). The geographic scope of A. scherman is more restricted, but it is also poorly resolved. Hinton (1926) believed that its range extends over “Central Europe from the Baltic southwards to the Pyrenees”, also encompassing the Alps. Panteleyev (2001) and Amori et al. (2008a) mapped A. scherman for the mountainous regions in northern Spain and the Pyrenees, the Alps, the mountains of Central Europe, and the Carpathians. The two species are presumably allopatric in the south, but show overlapping ranges in the north (e.g., Meylan 1977). Arvicola voles, which typically possess rootless molars, presumably evolved from the extinct rhizodont Mimomys Forsyth Major, 1902, and have been known in the Figure 1 Distribution of water vole samples sequenced for cytochrome b gene. Approximate ranges of Arvicola amphibius s.s. fossil record since the beginning of the Early Pleistocene and A. scherman follow Batsaikhan et al. (2008) and Amori et al. (Gromov and Polyakov 1992). Therefore, the taxonomic (2008a), respectively. Symbols for morphotypes (Δ – fossorial; □ richness in Arvicola results from a relatively recent radia- – aquatic) correspond to those in Figure 4 and Table 1. Samples of tion and the species may still be in an ongoing speciation unknown morphotype are denoted by circles. process. In such cases, which are common in arvicolines (e.g., Jaarola et al. 2004), molecular data provide vital insight into past cladogenetic events and current species of Biology, Selçuk University (Konya, Turkey). Tissue limits. In the present study, we used mitochondrial cyto- samples are deposited in the PMS and NMW. chrome b (cytb) gene sequences to reconstruct phylogenetic relationships between the fossorial and the aquatic water voles from various regions of their European and Morphotypes Asiatic range. Our aim was to test the taxonomic arrangement of A. amphibius s.l. Our null hypothesis predicted We examined museum specimens of water voles for their the reciprocal monophyly of the aquatic and the fossorial morphology. Included in the study were individuals morphotypes. If our hypothesis would not be rejected, sequenced for cytb but also museum vouchers of unknown the results will support the current tripartite taxonomic genetic makeup from the same site as the genotyped mate- arrangement of the genus Arvicola. rial. Swiss and German sequences were downloaded from GenBank and their morphotype was not reported (Pfunder et al. 2004, Schlegel et al. 2012b). Swiss samples were from the range of a subspecies scherman (Meylan and Saucy Materials and methods 1995) and were assigned on this ground to the fossorial morphotype. German water voles were not classified into Samples the morphotype. Identification was based on character states and five We studied the mitochondrial genetic makeup of water linear measurements (three external and two cranial) that voles from Switzerland, Austria, Hungary, Slovenia, are used in determination keys (see references in the next Bosnia and Herzegovina, Serbia, Romania, Turkey, and paragraph). Measurements were scored only in full grown Russia (Figure 1, Table 1). Voucher specimens (skins and voles (age groups V and VI of Hinton 1926) with complete skulls, or wet specimens) were preserved for the majority skulls: H&B – length of head and body; TL – length of of sequenced individuals. They are deposited in the Slo- tail; HF – length of hind foot (without claws); condyloba- venian Museum of Natural History (PMS; Ljubljana, Slo- sal length of skull (CbL); and length of molar tooth-row venia), Naturhistorisches Museum Wien (NMW; Vienna, (MxT). External measurements were obtained from speci- Austria), “Grigore Antipa” National Museum of Natural men tags and skull dimensions were measured using a History (Bucharest, Romania), and the Department dial calliper to the nearest 0.1 mm. B. Kryštufek et al.: Fossorial morphotype does not make a species in water voles 295

Table 1 Water voles included in the phylogenetic analysis, arranged according to morphotypes (Δ – fossorial; □ – aquatic). Names of countries are abbreviated as follows: A – Austria; BH – Bosnia and Herzegovina; CH – Switzerland; D – Germany; H – Hungary; RO – Romania; RU – Russia; SL – Slovenia; SR – Serbia; TR – Turkey.

Location Sample locality Longitude/latitude N Morphotype Haplotype GenBank access. no.

1 Hagen, D 51°11′/07°58′ D1 JF318969a 2 Eckhardtshausen, D 50°52′/10°16′ D2 HQ728473b D3 HQ728474b 3 Dürrweitzschen, D 51°10′/12°43′ D4 HQ728469b D5 HQ728470b 4 Scharfenberg, D 51°09′/13°29′ D6 HQ728467b D7 HQ728468b 5 Winnenden, D 48°52′/09°24′ D8 HQ728477b 6 Romerstein, D 48°30′/09°31′ D9 HQ728476b 7 Wiesensteig, D 48°31′/09°37′ D10 HQ728475b D11 HQ728478b 8 Noflen, CH 46°48′/07°33′ Δ CH1 HQ728471b Δ CH2 HQ728472b 9 Schötz, CH 47°08′/07°59′ Δ CH3 AY332708c 10 Kriens, CH 47°03′/08°17′ Δ CH4 AY332709c 11 Hohenems, A 47°22′/09°41′ 2 Δ A1 KM004997 1 Δ A2 KM004998 1 Δ A3 KM004999 1 Δ A4 KM005000 1 Δ A5 KM005001 12 Brandenberg, A 47°32′/11°54′ 5 Δ A6 KM005002 1 Δ A7 KM005003 13 Leogang, A 47°27′/12°40′ 2 Δ A8 KM005004 1 Δ A9 KM005005 1 Δ A10 KM005006 14 Landl, A 47°39′/14°42′ 1 Δ A11 KM005007 1 Δ A12 KM005008 15 Maribor, SL 46°33′/15°38′ 1 Δ SL1 KM005036 16 Vransko, SL 46°15′/14°57′ 1 Δ SL2 KM005037 17a Blagovica, SL 46°10′/14°48′ 1 Δ SL2 KM005037 1 Δ SL3 KM005038 1 Δ SL4 KM005039 17b Trebeljevo, SL 46°01′/14°44′ 1 Δ SL5 KM005040 18 Mt. Snežnik, SL 45°36′/14°23′ 1 Δ SL6 KM005041 1 Δ SL7 KM005042 1 Δ SL8 KM005043 19 Kis-Balaton, H 46°37′/17°08′ 1 □ H1 KM005011 1 □ H2 KM005012 1 □ H3 KM005013 1 □ H4 KM005014 1 □ H5 KM005015 1 □ H6 KM005016 1 □ H7 KM005017 1 □ H8 KM005018 3 □ H9 KM005019 1 □ H10 KM005020 1 □ H11 KM005021 1 □ H12 KM005022 2 □ H13 KM005023 1 □ H14 KM005024 1 □ H15 KM005025 20 Lake Matty, H 45°47′/18°16′ 1 □ H16 KM005026 1 □ H17 KM005027 21 Dubovac, SR 44°47′/21°11′ 1 □ SB1 KM005044 296 B. Kryštufek et al.: Fossorial morphotype does not make a species in water voles

(Table 1 Continued)

Location Sample locality Longitude/latitude N Morphotype Haplotype GenBank access. no.

22 Kupres, BH 43°56′/17°14′ 1 □ BH1 KM005009 5 □ BH2 KM005010 23 Argeş, RO 45°30′/25°10′ 1 Δ RO1 KM005028 24 Tulcea, RO 45°08′/28°35′ 1 □ RO2 KM005029 1 □ RO3 KM005030 25 Ankara, Ayas, TR 40°01′/32°19′ 2 □ TU1 KM005045 26 Aksaray, TR 38°15′/34°17′ 1 □ TU2 KM005046 1 □ TU2 KM005047 27 Novosibirsk, RU 54°42′/83°14′ 1 □ RUS1 KM005031 1 □ RUS2 KM005032 28 Aktash, RU 50°17′/87°43′ 1 □ RUS3 KM005033 29 Kuray steppe, RU 50°10′/87°54′ 2 □ RUS4 KM005034 3 □ RUS5 KM005035 aSchlegel et al., 2012a; bSchlegel et al., 2012b; cPfunder et al., 2004.

Morphological differences between the morpho- sequences were checked for the absence of stop-codons types were compiled from Miller (1910), Hinton (1926), and chimeric sequences. Ognev (1964), and Panteleyev (2001). In line with these Nucleotide, amino acid composition, and genetic authors, the aquatic morphotype attains larger dimen- distances were analyzed assuming a Kimura 2 parameter sions (H&B > 165 mm, HF ≥ 28 mm, CbL > 36 mm, MxT ≥ 9 (K2P) sequence evolution with 104 bootstraps in the MEGA mm), has a longer tail ( > 98 mm), less reduced palmar v. 6 program (Tamura et al. 2013). The most appropriate and plantar pads, orthodont upper incisors, and nearly models of DNA substitution for the data were identified vertically truncated occiput. The fossorial morphotype using MRMODELTEST 2.3 (Nylander 2004). Both the Akaike is smaller (H&B ≤ 160 mm, HF ≤ 27 mm, CbL ≤ 36 mm, information criterion (AIC) and the hierarchical likelihood MxT ≤ 9 mm), with a shorter tail (TL ≤ 98 mm), more ratio test (hLRT) were used. Phylogenetic analysis was con- reduced palmar and plantar pads, protruding (proodont) ducted with the Bayesian inference (BI), using the program incisors that are less concealed by the lips, and the occiput MRBAYES 3.1.2 (Huelsenbeck and Ronquist 2001, Ronquist is inclined. and Huelsenbeck 2003), and maximum likelihood (ML) as implemented in the program PhyML 2.4.5 (Guindon and Gascuel 2003, Anisimova and Gascuel 2006). Molecular protocols The phylogenetic inferences were performed with a general time-reversible model (GTR) + gamma distri- This study consisted of 69 Arvicola samples from 29 locali- bution (G) + proportion of invariable sites (I) (G = 0.6125 ties in Austria, Hungary, Slovenia, Bosnia and Herzegovina, and I = 0.5557). Four Monte Carlo Markov chains were run Serbia, Romania, Turkey, and Russia. A further 15 haplo- simultaneously for 6.5 × 106 generations, with the resulting types from Germany and Switzerland and five haplotypes trees sampled every 500 generations. Bayesian posterior of A. sapidus (FJ539341-5; Centeno-Cuadros and Godoy 2010) probabilities (BPP) were used to assess branch support were downloaded from the GenBank database (Table 1). of the BI tree. Convergence for posterior probabilities was DNA was extracted from muscle samples preserved checked by examining the generation plot visualized with in 96% ethanol using a QIAamp DNA Mini kit (Qiagen, TRACER v1.4 (Rambaud and Drummond 2007). Valencia, CA, USA). A 1117-bp mitochondrial cytb frag- The GTRGAMMA model was used for ML analysis. ment was amplified using a polymerase chain reaction Branch support (BP) in the ML tree was estimated by combining the following primers: L14725 (Steppan et al. 103 bootstrap replicates. The topologies resulting from 1999); L14931 (Harrison et al. 2003); H15915-SP (Jaarola these two methods were compared using a Shimodaira- and Searle 2002); L15162Marv, L15408Marv (Haynes et al. Hasegawa test (Shimodaira and Hasegawa 1999) imple- 2003); and H15497-SP (Jaarola et al. 2004). Forward and mented in PAUP* 4.010b (Swofford 2002) with 103 reverse sequencing was performed on an ABI PRISM 3130 bootstrap replicates. In line with other authors, we Genetic Analyzer (Applied Biosystems, Foster City, CA, accepted BPP > 0.95 as “good”, and BPP = 0.90–0.95 as USA) using BigDye chemistry (Applied Biosystems). The “moderate” support. For branch support in the ML tree we B. Kryštufek et al.: Fossorial morphotype does not make a species in water voles 297 considered BP > 90% as “good” support, and BP = 80–90% as “moderate” support. Trees were rooted with 11 species of Palaearctic arvicolines from three genera, which are probably closely related to Arvicola (Martínková and Moravec 2012): Chionomys gud (Satunin, 1909) (GQ352457; Bannikova et al. 2010), C. roberti (Thomas, 1906) (GQ352459; Bannikova et al. 2010), C. nivalis (Martins, 1842) (AY513848; Jaarola et al. 2004), Microtus arvalis (Pallas, 1778) (GU197810; Martínk- ová et al. 2013), M. agrestis (Linnaeus, 1761) (DQ662099; Schmidt-Chanasit et al. 2010), M. cabrerae (Thomas, 1906) (JX457758; Barbosa et al. 2013), M. duodecimcostatus (de Selys-Longchamps, 1839) (JX457806; Barbosa et al., 2013), M. lusitanicus (Gerbe, 1879) (JX457796 JX457796; Barbosa et al. 2013), M. subterraneus (de Selys-Longchamps, 1839) (JN019755; Rovatsos and Giagia-Athanasopoulou 2012), Neodon irene (Thomas, 1911) (JF906127; Chen et al. 2012), and N. leucurus (Blyth, 1863) (AM392371; Galewski et al. 2006). Figure 2 Skulls (in lateral view) of fossorial (A) and aquatic (B) We further analyzed the data by SAMOVA 1.0 imple- water voles: (A) Slovenia, Ljubljana, Blagovica, CbL = 35.5 mm (PMS mented in Dupanloup et al. (2002). The objective was to 18874); (B) Bosnia and Herzegovina, Kupres, CbL = 39.6 mm (PMS find groups of sampling localities that were maximally 2541). Skulls are deposited in the Slovenian Museum of Natural differentiated from each other. The number of popula- History (PMS). Scale bar = 5 mm. tions that maximized variation among groups (FCT) was the number of populations assumed to be correct. The comparison of all pairwise K2P values and the geographic Molecular analyses distances was performed using 1000 permutations in the Mantel test implemented in Rohlf (1994). Altogether, 50 new haplotypes were found in our material from Europe and Asia, generating a total dataset of 70 different water vole cytb haplotypes (Pfunder et al. 2004, Centeno-Cuadros and Godoy 2010, Schlegel et al. Results 2012a,b). Within the 1117-bp long sequences considered here, 106 polymorphic sites were found with a total of 140 Morphotypes mutations, 105 of which were parsimony informative. No stop codons, insertions, or deletions were observed in the Of the 67 adult water voles examined morphologi- alignment. The mean transition/transversion ratio was cally, 44 individuals were fossorial and 23 were aquatic. 5.90. The nucleotide composition was characterized by a There was a good match between the orientation of the deficit of guanines (13.8%), similar to that described in upper incisors (Figure 2) and size. In line with published other mammals (Irwin et al. 1991). data (see references in Material and methods), voles Phylogenetic relationships among Arvicola hap- with proodont upper incisors (fossorial type) were also lotypes that were reconstructed by the two different smaller (CbL = 32.42 ± 0.407), whereas the orthodont voles methods (ML and BI) yielded highly similar results, and of the aquatic morphotype attained larger dimensions the Shimodaira–Hasegawa test did not reveal significant (CbL = 36.84 ± 0.622). In addition, the fossorial water voles differences between them (SH = 0.23). Consequently, only showed a shorter tail relative to HbL (TL = 40–59% of HbL) the BI tree is shown (Figure 3). In line with published than aquatic voles (TL = 48–82% of HbL). The two morpho- results (Taberlet et al. 1998), both trees retrieved strongly types showed strong geographic associations (Figure 1). (BP = 100%, PP = 1.00) supported basal dichotomy into A. In accordance with published evidence (e.g., Panteleyev sapidus and A. amphibius s.l. Within the latter, a branch- 2001), our fossorial voles originated from the Alps and the ing pattern yielded congruent results at terminal nodes Carpathians, and aquatic voles were collected in south- but presented inconsistency at deeper nodes. Fossorial eastern Europe, Turkey in Asia, and western Siberia. water voles from Switzerland hold a sister position against 298 B. Kryštufek et al.: Fossorial morphotype does not make a species in water voles

Figure 3 Bayesian inference tree reconstructed from cytochrome b sequences of water vole Arvicola. The tree is rooted with 11 species of Arvicolinae: Microtus agrestis, M. cabrerae, M. subterraneus, M. lusitanicus, M. duodecimcostatus, M. arvalis, Neodon irene, N. leucurus, Chionomys nivalis, C. roberti, and C. gud. The branching pattern and branch lengths follow the Bayesian analysis, whereas the first and second numbers on the branches correspond to posterior probability values and bootstrap support in the maximum likelihood tree analyses, respectively. Symbols for morphotypes (Δ – fossorial; □ – aquatic) correspond to those in Figure 1 and Table 1. the remaining haplotypes, which demonstrated strong remained unresolved due to poor branch supports. The support (PP = 0.99) in the BI analysis. The basal position next major branching was a supported (PP = 0.95) diver- of the Turkish haplotypes was evident in both trees, but gence into the Siberian and European lineages. The B. Kryštufek et al.: Fossorial morphotype does not make a species in water voles 299 prevailing pattern within the European lineage was a sup- 40 chromosomes) and the remaining water voles with 36 ported (PP = 0.99) polytomy. Geographic association was chromosomes (e.g., Meylan 1977) is undoubtedly indica- particularly obvious in a cluster of Austrian and German tive of a speciation event. The two cytotypes are broadly haplotypes that radiated further into a number of sub- sympatric in Spain and France (Shenbrot and Krasnov lineages. The Turkish and Siberian lineages contained 2005), and the genetic variation from the mitochondrial only aquatic water voles, whereas the only monophyl- cytb gene (7.6% in Taberlet et al. 1998, 6.2% in our results) etic lineage of purely fossorial water voles was the Swiss is within the range for interspecific differences in rodents lineage. The major European lineage contained both large (Baker and Bradley 2006). Furthermore, there is no evi- aquatic and small fossorial water voles. These morpho- dence of mitochondrial gene flow between A. sapidus and types were strictly allopatric in our samples. A. amphibius s.l. in their zone of contact (Centeno-Cuadros The highest percentage of variance among groups et al. 2009). K2P divergence between A. terrestris italicus (67.2%) was obtained when samples of Arvicola amphibius (sensu Taberlet et al. 1998; hereafter the Italian lineage) s.l. were classified into three groups, which corresponded and the remaining lineages of A. amphibius s.l. is tenta- to the three lineages retrieved in the above phylogenetic tively estimated from Taberlet et al. (1998) to be about 5%, analyses: Swiss, Turkish, and European + Siberian com- and the divergence of the Swiss fossorial lineage in our bined. The results were therefore in congruence with the study accounts for 4.1%. In rodents, genetic divergences haplotype geographic distribution and supported strong of ≤ 4.1% are tentatively indicative of intraspecific varia- association of haplotypes with geographic location. tion and those of ≥ 4.9% putatively suggest interspecific The highest K2P genetic distance (in percentages) variation (Baker and Bradley 2006). Therefore, major was between Arvicola sapidus and the remaining lineages divergences in A. amphibius s.l. can be equally well inter- (6.2 ± 0.9). Within A. amphibius s.l., K2P genetic distances preted as high intraspecific or low interspecific variation. were not correlated with geographical distances among Graf (1982) reached similar conclusions in a study based samples (matrix correlation r = 0.136, t = 0.479, p = 0.68). on the electrophoresis of allozymes. Captive crossbreeding The K2P value was the highest (4.1 ± 0.72) between the fos- between two morphotypes from Switzerland (putatively sorial Swiss water voles and the remaining samples, the representing the fossorial Swiss lineage in our results, lowest (1.6 ± 0.3) between the Siberian and the European and the Italian lineage) retrieved partial sterility of the F1 lineages, and intermediate (2.3 ± 0.5) between the Turkish progeny, therefore suggesting some reproductive isolation and the combined Siberian + European lineages. Within- between major phylogeographic lineages of A. amphibius group divergence was the highest in the European lineage (Saucy et al. 1994). Conversely, Kleist (1996) successfully (1.2 ± 0.2), lowest in the Siberian (0.5 ± 0.2) and Turkish lin- crossbred aquatic and fossorial water voles originating eages (0.6 ± 0.3), and intermediate in the Swiss fossorial from Germany. lineage (0.9 ± 0.2). Clearly, further sampling is essential to define geographic limits of major phylogroups and explore zones of contact or overlap between them. Molecular results that have been obtained thus far and discussed above suggest Discussion that the taxonomy of A. amphibius s.l. is a complex issue that cannot be resolved by a simple division into the Species of water voles aquatic and fossorial species. The data presented herein make it very unlikely that A. scherman forms a single Molecular reconstructions of extant water voles retrieved monophyletic taxon. However, a very recent split between four major evolutionary lineages (Taberlet et al. 1998, fossorial and aquatic morphotypes in the European and our results) instead of three clusters as predicted by lineage may still be detected after analyzing additional the current taxonomic arrangement (Musser and Carle- genes and obtaining a larger dataset of sequences. ton 2005). Topology of phylogenetic trees provided no support for basal aquatic-fossorial dichotomy in Arvicola amphibius s.l., and instead suggests multiple origins of Survival in refugia small proodont water voles. Similarly, the pattern of variation in the distribution and amount of constitutive (C-) het- The four lineages of water voles bear witness to the main erochromatin variation is not related to a morphotype of cladogenetic events during the evolutionary history of individual populations of A. amphibius (Arslan et al. 2011). the genus Arvicola. Considering the time of split of the Basal dichotomy in Arvicola between A. sapidus (having common ancestor of A. sapidus and A. amphibius s.l., 300 B. Kryštufek et al.: Fossorial morphotype does not make a species in water voles which is estimated at about 252 kya (Centeno-Cuadros adaptation and respond morphologically even to sea- et al. 2009), the lineages of the latter must have diverged sonal dynamics in resources, e.g., those occupying the during the last two major glacial-interglacial cycles, the banks of large rivers change morphologically in response Saalian (Riss) and the Vistulian (Würm). Clearly, the phy- to floods (Panteleyev et al. 1978). It is intriguing to specu- logeographic architecture in water voles is a legacy of late that fossorial and aquatic water voles might be at the Quaternary climatic dynamics. Just as it was the case with extremes of a putative phenotypic continuum, rather than numerous temperate taxa (Taberlet et al. 1998, Hewitt representing two discrete morphotypes. Well-controlled 2000), the range of Arvicola was repeatedly fragmented morphometric analyses must be an integral part of any by expanding glaciers and isolated populations began to comprehensive taxonomic assessment of this interesting diverge in relatively small refugial areas. rodent genus. Deeply divergent lineages of water voles were retrieved only at the extreme western edge of the exten- Acknowledgments: We would like to thank Mrs. Karolyn sive range of the genus, suggesting that Western and Close for the English editing and Mr. Peter Glasnović for southwestern Europe have acted as a major centre for the his help with Figure 1. diversification of Arvicola. Namely, the more ancient the lineage, the more westward it tends to be located. There- fore, A. sapidus occupies the extreme western border of Appendix Arvicola, being followed further east by two ancient lineages, the Swiss and the Italian. Phylogeographic studies The Appendix is aimed to assist the reader with the of widespread Palaearctic arvicolines repeatedly retrieved nomenclatural component of our study by listing main high genetic diversity in Western and Central Europe, synonyms and those taxonomic names that are related which contrasts the genetic uniformity in Eastern Europe to our sampling localities. Names are given chronologi- and Asia. Although each species has its own geographic cally and follow the taxonomic arrangement of Musser pattern of genetic architecture, a west-to-east decay in dis- and Carleton (2005). Junior synonyms of Arvicola tribution of genetic diversity is clearly evident in various amphibius and A. scherman (sensu Musser and Carleton arvicolines, e.g., Clethrionomys glareolus (Wójcik et al. 2005) are listed as trinomials but with no intention of 2010), Microtus oeconomus (Brunhoff et al. 2003), Micro- implying their actual taxonomic status. Therefore, tri- tus agrestis (Jaarola and Searle 2002, Herman and Searle nomials are the available species group names although 2012), and Microtus arvalis (Haynes et al. 2003, Tougard they do not necessarily merit a subspecific rank. If not et al. 2008). indicated otherwise, type localities are from Ellerman and Morrison-Scott (1951). Chromosomal data follow Zima and Král (1984). Morphotypes Arvicola amphibius (Linnaeus, 1758) Many authors report fossorial water voles for higher alti- Type locality: England. tudes and aquatic populations for lower altitudes (see Pan- Notes: The oldest available species group name in the teleyev 2001), which suggests the competitive advantage genus, which is based on large, orthodont and strictly of a particular morphotype under certain environmental aquatic water voles (Miller 1912) with 36 chromosomes. conditions. However, the reality is more puzzling and fos- Precedence of amphibius over terrestris is legitimized by sorial water voles are also frequently encountered in the the Principle of the First Reviser (Article 24.2. of the Inter- lowlands (Meylan 1977), whereas aquatic animals occupy national Code of Zoological Nomenclature, 4th ed.) and mountains in some parts of Europe (e.g., the Balkans) and follows Blasius (1857) (cf. Corbet 1978). Asia (Asia Minor, Caucasus, Altai; authors’ field observa- tions). The issue is further complicated by morphological Arvicola terrestris (Linnaeus, 1758) forms “which are [not] solely fossorial” (Meylan 1977). Type locality: Upsala, Sweden. The fossorial morphotype presumably originated Notes: Moderately large aquatic voles with a skull showing from the aquatic one through a heterochronic process of “a decided tendency to assume a fossorial structure.” “[H] accelerated dwarfism (Cubo et al. 2006). Although some abits both aquatic and mole-like” (Miller 1912). Priority of populations remain morphologically stable across years, terrestris over amphibius was accepted by Ellerman and others are more plastic in this respect (Panteleyev 2001). Morrison-Scott (1951), Ognev (1964), and numerous sub- Namely, water voles can show considerable ecological sequent authors. B. Kryštufek et al.: Fossorial morphotype does not make a species in water voles 301

Arvicola scherman (Shaw, 1801) Arvicola amphibius variabilis Ognev, 1933 Type locality: Strasbourg, Bas Rhin, Eastern France. Type locality: Barabinsk steppes Govt. Tomsk, Siberia. Notes: Small and proodont water voles adapted to fosso- Notes: Large, orthodont, and aquatic water voles. Arvicola rial mode of life. Recognized as species on its own right terrestris variabilis Ognev, 1933, is preoccupied by Arvicola already by Miller (1912) and Ognev (1964). arvalis form variabilis Rörig et Börner, 1905 (now a junior synonym of Microtus arvalis) and was renamed Microtus Arvicola amphibius italicus Savi, 1838 terrestris barabensis Heptner, 1948. This name (sensu Type locality: Vicinity of Pisa, Italy. Gromov and Erbajeva 1995) is applicable to our samples Notes: Well-defined form of large, orthodont, and from Russia (pts. 27–29 on Figure 1). strictly aquatic water voles with 36 chromosomes. For 1838 as the year of publication (instead of 1839; cf. Arvicola amphibius martinoi Petrov, 1949 Miller 1912) see Gippoliti (2012). Range is restricted to Type locality: Vicinity of Belgrade, Serbia (Petrov 1949). peninsular Italy (Amori et al. 2008b) and Canton Tessin Notes: Includes large and aquatic water voles (Petrov 1949). in Switzerland (Meylan and Saucy 1995); borders are Our specimen from Dubovac, Serbia (pt. 21 in Figure 1) origi- not resolved. nates from the vicinity of the type locality for this name.

Arvicola scherman exitus Miller, 1910 Type locality: St. Gallen, Switzerland. Notes: Small, proodont, and strictly fossorial water voles References (Miller 1912). The name exitus is probably applicable to Amori, G., R. Hutterer, B. Kryštufek, N. Yigit, G. Mitsain and L.J.P. sequences from Switzerland (pts. 8–10 on Figure 1), which Muñoz. 2008a. Arvicola scherman. In: IUCN 2012. IUCN Red we downloaded from the GenBank. List of Threatened Species. Version 2012.2. < www.iucnredlist. org > . Accessed 17 May 2013. Arvicola amphibius persicus De Filippi, 1865 Amori, G., L. Contoli and A. Nappi. 2008b. Fauna d’Italia. Mammalia Type locality: Sultanieh, south of Elbruz Mountains, II: Erinaceomorpha, Soricomorpha, Lagomorpha, Rodentia. Calderini, Milano. pp. 736. Persia ( = Iran). Anisimova, M. and O. Gascuel. 2006. Approximate likelihood ratio Notes: Contains large, orthodont, and strictly aquatic test for branches: a fast, accurate and powerful alternative. water voles (Kryštufek and Vohralík 2005) with 36 chro- Syst. Biol. 55: 539–552. mosomes (Arslan et al. 2013). This is the oldest name for Arslan, A., T. Yorulmaz, K. Toyran, S. Gözütok and J. Zima. 2011. water voles from the Middle East and is applicable to our C-heterochromatin variation and NOR distribution in the karyo- samples from Anatolia (pts. 25 and 26 on Figure 1). Taxo- type of water vole, Arvicola terrestris (Mammalia, Rodentia). Caryologia 64: 215–222. nomic scope is not resolved unambiguously but persicus Baker, R.J. and R.D. Bradley. 2006. Speciation in mammals and the probably contains as junior synonyms Microtus terrestris genetic species concept. J. Mammalogy 87: 643–662. armenius Thomas, 1907 (Type locality: Van, 5000 ft., Bannikova, A., V. Lebedev, A. Lissovsky, V. Matrosova, N. Abramson, Eastern Asia Minor) and Arvicola terrestris hintoni Aharoni, E. Obolenskaya and A. Tesakov. 2010. Molecular phylogeny 1932 (Type locality: Island of Tel el Sultan, Antioch Lake, and evolution of the Asian lineage of vole genus Microtus (Arvi- colinae, Rodentia) inferred from mitochondrial cytb sequence. Northern Syria) (Ellerman and Morrison-Scott 1951, Ognev Biol. J. Linn. Soc. 99: 595–613. 1964, Kryštufek and Vohralík 2005). Barbosa, S., J. Pauperio, J.B. Searle and P.C. Alves. 2013. Genetic identification of Iberian rodent species using both mitochon- Arvicola amphibius illyricus (Barrett-Hamilton, 1899) drial and nuclear loci: application to noninvasive sampling. Type locality: Bosnia (no exact locality). Mol. Ecol. Resour. 13: 43–56. Notes: Contains large, orthodont, and strictly aquatic Batsaikhan, N., H. Henttonen, H. Meinig, G. Shenbrot, A. Bukhni- kashvili, G. Amori, R. Hutterer, B. Kryštufek, N. Yigit, G. Mitsain water voles (Petrov 1949). Our sample from Kupres (pt. 22 and L.J. Palomo. 2008. Arvicola amphibius. In: IUCN 2012. in Figure 1) is topotypical with this name. IUCN Red List of Threatened Species. Version 2012.2. < www. iucnredlist.org > . Accessed 17 May 2013. Arvicola sapidus Miller, 1908 Blasius, J.H. 1857. Naturgeschichte der Säugetiere Deutschlands Type locality: Santo Domingo de Silos, Burgos, Spain. und der angenzenden Länder von Mitteleuropa. Brauschweig. pp. 549. Notes: A large, orthodont, and strictly aquatic species Brunhoff, C., K.E. Galbreath, V.B. Fedorov, J.A. Cook and M. Jaarola. with 40 chromosomes. Range restricted to Spain, Portu- 2003. Holarctic phylogeography of the root vole (Microtus gal, and western France; mapped in Ventura (2002) and oeconomus): implications for late Quaternary biogeography of Shenbrot and Krasnov (2005). high latitudes. Mol. Ecol. 12: 957–968. 302 B. Kryštufek et al.: Fossorial morphotype does not make a species in water voles

Centeno-Cuadros, A. and J.A. Godoy. 2010. Structure, organization Irwin, D.M., T.D. Kocher and A.C. Wilson. 1991. Evolution of the and nucleotide diversity of the mitochondrial control region cytochrome b gene of mammals. J. Mol. Evol. 32: 128–144. and cytochrome b of southern water vole (Arvicola sapidus). Jaarola, M. and J.B. Searle. 2002. Phylogeography of field voles Mitochondrial DNA 21: 48–53. (Microtus agrestis) in Eurasia inferred from mitochondrial DNA Centeno-Cuadros, A., M. Delibes and J.A. Godoy. 2009. Dating the sequences. Mol. Ecol. 11: 2613–2621. divergence between Southern and European water voles using Jaarola, M., N. Martínková, I. Gündüz, C. Brunhoff, J. Zima, A. molecular coalescent-based methods. J. Zool. (Lond.) 279: Nadachowski, G. Amori, N.S. Bulatova, B. Chondropoulos, S. 404–409. Fraguedakis-Tsolis, J. González-Esteban, M.J. López-Fuster, Chen, W.-C., H..B. Hao, Z.-Y. Sun, Y. Liu, S.-Y. Liu and B.-S. Yue. 2012. A.S. Kandaurov, H. Kefelioglu, M.D. Mathias, I. Villate and J.B. Phylogenetic position of the genus Proedromys (Arvicolinae, Searle. 2004. Molecular phylogeny of the speciose vole genus Rodentia): Evidence from nuclear and mitochondrial DNA. Microtus (Arvicolinae, Rodentia) inferred from mitochondrial Biochem. Syst. Ecol. 42: 59–68. DNA sequences. Mol. Phyl. Evol. 33: 647–663. Corbet, G.B. 1978. The mammals of the Palaearctic region: a taxo- Kleist, D. 1996. Stellung der oberen Incisivi und Proodontie nomic review. British Museum (Natural History), London. pp. 314. bei terrestrischen und aquatischen Schermäusen Arvicola Cubo, J., J. Ventura and A. Casinos. 2006. A heterochronic inter- terrestris (Linnaeus, 1758). Bonn. zool. Beit. 46: 275–282. pretation of the origin of digging adaptations in the northern Kryštufek, B. and V. Vohralík. 2005. Mammals of Turkey and Cyprus. water vole, Arvicola terrestris (Rodentia: Arvicolidae). Biol. J. Rodentia I: Sciuridae, Dipodidae, Gliridae, Arvicolinae. Založba Linn. Soc. 87: 381–391. Annales, Koper. pp. 292. Dupanloup, I., S. Schneider and L. Excoffier. 2002. A simulated Martínková, N. and J. Moravec. 2012. Multilocus phylogeny of annealing approach to define the genetic structure of popula- arvicoline voles (Arvicolini, Rodentia) shows small tree terrace tions. Mol. Ecol. 11: 2571–2581. size. Folia Zool. 61: 254–267. Ellerman, J.R. and T.C.S. Morrison-Scott. 1951. Checklist of Palaearc- Martínková, N., R. Barnett, T. Cucchi, R. Struchen, M. Pascal, tic and Indian mammals 1758 to 1946. British Museum (Natural M.C. Fischer, T. Higham, S. Brace, S.Y.W. Ho, J. Quéré, P. History), London. pp. 810. O’Higgins, L. Excoffier, G. Heckel, A.R. Hoelzel, K.M. Dobney Galewski, T., M.K. Tilak, S. Sanchez, P. Chevret, E. Paradis and and J.B. Searle. 2013. Divergent evolutionary processes E.J. Douzery. 2006. The evolutionary radiation of Arvicolinae associated with colonization of offshore islands. Mol. Ecol. rodents (voles and lemmings): relative contribution of nuclear 22: 5205–5220. and mitochondrial DNA phylogenies. BMC Evol. Biol. 6. Matthey, R. 1956. Cytologie chromosomique comparée et systéma- doi: 10.1186/1471-2148-6-80. tique des Muridae. Mammalia 20: 29–123. Gippoliti, S. 2012. The name of the Italian water vole Arvicola cf. Meylan, A. 1977. Fossorial forms of the water vole, Arvicola terrestris amphibius (Linnaeus, 1758). Hystrix, It. J. Mamm. 23: 87–89. (L.), in Europe. EPPO Bul. 7: 209–221. Graf, J.-D. 1982. Génétique biochimique, zoogégraphie et taxonomie Meylan, A. and F. Saucy. 1995. Arvicola terrestris (L., 1758). In: (J. des Arvicolidae (Mammalia, Rodentia). Rev. Suisse Zool. 89: Hausser, ed.) Mammifères de la Suisse. Répartition, Biologie, 749–787. Ecologie. Bikhäuser Verlag, Basel. pp. 303–313. Gromov, I.M. and M.A. Erbajeva. 1995. The mammals of Russia Miller, G.S. 1910. Brief synopsis of the water rats of Europe. Proc. and adjacent territories: Lagomorphs and rodents. Zoological Biol Soc. Wash. 23: 19–22. Institute, Russian Academy of Sciences, St. Petersburg. pp. 521 Miller, G.S. 1912. Catalogue of the mammals of Western Europe (in Russian). (Europe exclusive of Russia) in the collection of the British Gromov, I.M. and I.Ya. Polyakov. 1992. Voles (Microtinae). Smithso- Museum. British Museum, London. pp. 1019. nian Institution Libraries, Washington, DC. pp. 725. Musser, G.G. and M.D. Carleton. 2005. Superfamily Muroidae. Guindon, S. and O. Gascuel. 2003. A simple, fast, and accurate In: (D.E. Wilson and D.A.M. Reeder, eds.) Mammal species of algorithm to estimate phylogenies by maximum likelihood. the World. A taxonomic and geographic reference. 3rd ed., Syst. Biol. 52: 696–704. Vol. 2. John Hopkins University Press, Baltimore, MD. Harrison, R.G., S.M. Bogdanowicz, R.S. Hoffmann, E. Yensen and pp. 894–1531. P.W. Sherman. 2003. Phylogeny and evolutionary history of the Niethammer, J. and F. Krapp. 1982. Handbuch der Säugetiere ground squirrels (Rodentia: Marmotinae). J. Mamm. Evol. 10: Europas: Band 2/I, Rodentia II. Akademische Verlagsgesells- 249–276. chaft, Wiesbaden. pp. 649. Haynes, S., M. Jaarola and J.B. Searle. 2003. Phylogeography of the Nylander, J.A.A. 2004. Mrmodeltest, Version 2.2. Uppsala: Uppsala common vole (Microtus arvalis) with particular emphasis on the University, Department of Systematic Zoology. colonization of the Orkney archipelago. Mol. Ecol. 12: 951–956. Ognev, S.I. 1964. Mammals of the U.S.S.R. and adjacent countries. Herman, J.S. and J.B. Searle. 2012. Post-glacial partitioning of mito- Mammals of Eastern Europe and northern Asia. Vol. VII, Rodents. chondrial genetic variation in the field vole. Proc. R. Soc. Lond. Israel Program for Scientific Translations, Jerusalem. pp. 626. B. 278: 3601–3607. Panteleyev, P.A. 2001. The water vole: mode of the species. Nauka, Hewitt, G.M. 2000. The genetic legacy of the Quaternary ice ages. Moscow. pp. 528. Nature 405: 907–913. Panteleyev, P.A., A.N. Terekhina and A.A. Varshavsky. 1978. Effect of Hinton, M.A.C. 1926. Monograph of voles & lemmings (Microtinae) floods on morphological characters of Arvicola terrestris. Zool. living and extinct. The British Museum (Natural History), Lon- Zh. 57: 759–767. don. pp. 488. Petrov, B. 1949. Materials for classification and geographical distri- Huelsenbeck, J.P. and F. Ronquist. 2001. MrBayes: Bayesian infer- bution of watervoles (Arvicola terrestris L.) in Serbia. Glasnik ence of phylogeny. Bioinformatics 17: 754–755. prirodnjačkog muzeja Beograd B 1–2: 171–200. B. Kryštufek et al.: Fossorial morphotype does not make a species in water voles 303

Pfunder, M., O. Holzgang and J.E. Frey. 2004. Development of Shenbrot, G.I. and B.R. Krasnov. 2005. An atlas of the geographic microarray-based diagnostics of voles and shrews for use in distribution of the arvicoline rodents of the world (Rodentia, biodiversity monitoring studies, and evaluation of mitochon- Muridae: Arvicolinae). Pensoft, Sofia. pp. 336. drial cytochrome oxidase I vs. cytochrome b as genetic mark- Shimodaira, H. and M. Hasegawa. 1999. Multiple comparisons of ers. Mol. Ecol. 13: 1277–1286. log-likelihoods with applications to phylogenetic inference. Rambaud, A. and A.J. Drummond. 2007. Tracer, version 1.4. http:// Mol. Biol. Evol. 16: 1114–1116. beast.bio.ed.ac.uk/Tracer. Accessed on June 11, 2013. Steppan, S.J., M.R. Akhverdyan, E.A. Lyapunova, D.G. Fraser, N.N. Rohlf, F.J. 1994. NTSYS-pc. Numerical taxonomy and multivariate Vorontsov, R.S. Hoffmann and M.J. Braun. 1999. Molecular phy- analysis system. Version 1.80. Exeter Publishing, Setauket, NY. logeny of the marmots (Rodentia: Sciuridae): test of evolutionary Ronquist, F. and J.P. Huelsenbeck. 2003. MRBAYES 3: Bayesian and biogeographic hypotheses. Syst. Biol. 48: 715–734. phylogenetic inference under mixed models. Bioinformatics 19: Swofford, D.L. 2002. PAUP*. Phylogenetic analysis using parsimony 1572–1574. (*and other methods), Version 4.0b10. Sinauer Associates, Rovatsos, M.T. and E.B. Giagia-Athanasopoulou. 2012. Taxonomical Sunderland, MA. status and phylogenetic relations between the ‘thomasi’and Taberlet, P., L. Fumagalli, A.-G. Wust-Saucy and J.-F. Cosson. 1998. ‘atticus’ chromosomal races of the underground vole Microtus Comparative phylogeography and postglacial colonization thomasi (Rodentia, Arvicolinae). Mamm. Biol. 77: 6–12. routes in Europe. Mol. Ecol. 7: 453–464. Saucy, F., A.-G. Wust Saucy and H.-J. Pelz. 1994. Biochemical Tamura, K., G. Stecher, D. Peterson, A. Filipski and S. Kumar. 2013. polymorphism and genetic variability in aquatic and fossorial MEGA6: Molecular evolution genetics analysis version 6.0. populations of the water vole, Arvicola terrestris, in Western Mol. Biol. Evol. 30: 2725–2729. Europe. Pol. ecol. Stud. 20: 559–573. Tougard, C., E. Renvoisé, A. Petitjean and J.-P. Quéré. 2008. New Schlegel, M., H.S. Ali, N. Stieger, M.H. Groschup, R. Wolf and R.G. insight into the colonization processes of common voles: Ulrich. 2012a. Molecular identification of small mammal inferences from molecular and fossil evidence. PloS ONE 3: species using novel cytochrome B gene-derived degenerated e3532. primers. Biochem. Genet. 50: 440–447. Ventura, J. 2002. Arvicola sapidus Miller, 1908, Rata de aqua. In: Schlegel, M., E. Kindler, S.S. Essbauer, R. Wolf, J. Thiel, M.H. Gros- (L.J. Palomo and J. Gisbert, eds.) Atlas de los Mamíferos Ter- chup, G. Heckel, R.M. Oehme and R.G. Ulrich. 2012b. Tula virus restres de España. Dirección General de Conservación de la infections in the Eurasian water vole, Central Europe. Vector- Naturaleza-SECEM-SECEMU, Madrid. pp. 362–365. Borne Zoon. Dis. 12: 503–513. Wójcik, J.M., A. Kawałko, S. Marková, J.B. Searle and P. Kotlík. 2010. Schmidt-Chanasit, J., S. Essbauer, R. Petraityte, K. Yoshimatsu, Phylogeographic signatures of northward post-glacial coloni- K. Tackmann, F.J. Conraths, K. Sasnauskas, J. Arikawa, A. zation from high-latitude refugia: a case study of bank voles Thomas, M. Pfeffer, J.J. Scharninghausen, W. Splettstoesser, using museum specimens. J. Zool. (Lond.) 281: 249–262. M. Wenk, G. Heckel and R.G. Ulrich. 2010. Extensive host Zima, J. and B. Král. 1984. Karyotypes of European mammals II. Acta sharing of central European Tula virus. J. Virol. 84: 459–474. Sc. Nat. Brno 18: 1–62.