Phylogeny, evolutionary history and taxonomy of the Mustelidae based on sequences of the cytochrome b gene and a complex repetitive flanking region

Phylogeny, evolutionary history and taxonomy of the Mustelidae based on sequences of the cytochrome b gene and a complex repetitive flanking region Blackwell Publishing, Ltd. JOSEP MARMI, JUAN FRANCISCO LÓPEZ-GIRÁLDEZ & XAVIER DOMINGO-ROURA Accepted: 30 December 2003 Marmi, J., López-Giráldez, J. F. & Domingo-Roura, X. (2004). Phylogeny, evolutionary history and taxonomy of the Mustelidae based on sequences of the cytochrome b gene and a complex repetitive flanking region. — Zoologica Scripta, 33, 481– 499. The Mustelidae is a diverse family of carnivores which includes weasels, polecats, mink, tayra, martens, otters, badgers and, according to some authors, skunks. Evolutionary relationships within the family are under debate at a number of different taxonomic levels, and incongruencies between molecular and morphological results are important. We analysed a total of 241 cytochrome b (cyt b) gene sequences and 33 sequences of a complex repetitive flanking region from 33 different species to compile an extensive molecular phylogeny for the Mustelidae. We analysed these sequences and constructed phylogenetic trees using Bayesian and neighborjoining methods that are evaluated to propose changes to the taxonomy of the family. The peripheral position of skunks in phylogenetic trees based on both loci suggests that they should be considered a separate family, Mephitidae. The subfamily Melinae is the basal group within the Mustelidae and trees based on the cyt b gene suggest that the American badger, Taxidea taxus, should be considered a separate monotypic subfamily, Taxidiinae. Otters classified within the genera Lutra, Amblonyx and Aonyx are grouped within the same clade in cyt b and combined partial cyt b and flanking region trees and show reduced levels of inter specific divergence, suggesting that they could be classified together under a single genus, Lutra. The Bayesian tree based on combined data from both loci supports the idea that subfamily Mustelinae is paraphyletic, as otters (subfamily Lutrinae) are included in this subfamily. Low levels of genetic divergence among European polecat, Mustela putorius, steppe polecat, Mustela eversmannii, and European mink, Mustela lutreola, suggest that these species could be considered subspecies within a single species, Mustela putorius. Our results are consistent with a rapid diversification of mustelid lineages in six different radiation episodes identified since the Early Eocene, the oldest events being the separation of subfamilies and the split of marten (Martes, Gulo) and weasel (Mustela) lineages in the Early Middle Miocene. The separation of New World from Old World lineages and the split of the remaining genera are estimated to have occurred in Late Miocene. The most recent events have been the differentiation of species within genera and this probably occurred in four radiation episodes at the end of Late Miocene, Early Pliocene, Late Pliocene and Pleistocene epochs. Josep Marmi, Juan Francisco López-Giráldez & Xavier Domingo-Roura, Departament de Ciències Experimentals i de la Salut, Universitat Pompeu Fabra, Dr Aiguader 80, 08003 Barcelona, Spain. E-mail: josep.marmi@upf.edu Introduction The Mustelidae is a diverse and widespread family of carnivores consisting of approximately 65 species which are distributed across the majority of the New and Old Worlds (Nowak 1991). Extant members of the family Mustelidae are considered to be a monophyletic group on the basis of the loss of the carnassial notch on the upper fourth molar, the loss of the upper second molar and the enlargement of scent glands (Martin 1989; Bryant et al. 1993). However, there remains much debate regarding the phylogenetic relationships within and among subfamilies within this family. In the past, reports by Simpson (1945) and Stains (1984) were widely cited in textbooks on this subject (e.g. Nowak 1991). Stains (1984) assigned the 23 extant genera of mustelids to the five subfamilies defined previously by Simpson (1945): (i) Mustelinae (weasels, mink, martens and wolverine): Eira, © The Norwegian Academy of Science and Letters • Zoologica Scripta, 33, 6, November 2004, pp481– 499 481 Phylogeny, evolution and taxonomy of the Mustelidae • J. Marmi et al. Galictis, Ictonyx, Lyncodon, Mustela, Poecilictis, Poecilogale, Vormela, Martes and Gulo; (ii) Lutrinae (otters): Aonyx, Enhydra, Lutra and Pteronura; (iii) Mellivorinae (honey-badger): Mellivora; (iv) Melinae (true badgers): Arctonyx, Meles, Melogale, Mydaus and Taxidea; and (v) Mephitinae (skunks): Conepatus, Mephitis and Spilogale. The phylogenetic relationship between skunks and other mustelids is the most controversial issue arising from this classification. Based on morphological characteristics, skunks have been described as a sister group to the badgers (Simpson 1945) and otters (Hunt 1974; Wozencraft 1989). However, these relationships may have been described on the basis of plesiomorphic character states, such as the unexpanded auditory bulla shared between skunks and otters (Dragoo & Honeycutt 1997). Chromosomal data (Wurster & Benirschke 1968) and serum protein analyses (Ledoux & Kenyon 1975) have, however, indicated a remarkable difference between skunks and other species of the Mustelidae family. Differences were confirmed with the arrival of molecular data when analyses of the cytochrome b (cyt b) gene indicated that skunks diverged before the rest of the mustelids (Ledge & Arnason 1996). It is becoming accepted that the Mustelidae, as defined by Simpson, represents a paraphyletic group, with skunks and the oriental stink badger (Mydaus spp.) forming a monophyletic clade. This clade has been referred to as the Mephitidae family, separated from the rest of the Mustelidae and the monophyletic Procyonidae (Dragoo & Honeycutt 1997). Relationships involving the remaining subfamilies have also been debated. Willemsen (1992) concluded from fossil evidence that badgers and otters share a common ancestor, the genus Mionictis, an otter-like animal from the Miocene. However, more recent molecular evidence suggests that otters may be closer to mustelines than to badgers (Masuda & Yoshida 1994; Koepfli & Wayne 1998; Hosoda et al. 2000). The monophyly of the subfamilies is also a matter of debate. Based on analyses of two mitochondrial genes (12S and 16S ribosomal RNAs), only the Lutrinae represent a monophyletic group (Dragoo & Honeycutt 1997). At the opposite extreme, Bininda-Emonds et al. (1999) combined phylogenetic information from morphological and molecular data to demonstrate that only badgers were polyphyletic. The monophyly of the Mustelinae is not supported by several other studies (e.g. Koepfli & Wayne 1998; Hosoda et al. 2000). The Melinae clearly seem to be a polyphyletic group, the badger ecomorph having evolved at least three times within the Mustelidae (Bryant et al. 1993). Most of the disagreements concerning phylogenetic positions and taxonomic status within the Mustelidae may be a consequence of a basal diversification at the origin of this family in which the major lineages split rapidly from one another. This argument was used to explain the lack of resolution among clades found in the subfamilies Lutrinae and 482 Mustelinae (Koepfli & Wayne 1998). Fast radiation events occurring during the Miocene have also been reported in other mammalian families, making it difficult to resolve the phylogenetic relationships and the taxonomy of taxa involved (e.g. Ursidae, Waits et al. 1999; Bovidae, Gatesy et al. 1997; Muridae, Martin et al. 2000). The evolution of the cyt b gene is well characterized (Irwin et al. 1991) and this gene provides relevant phylogenetic information in moderately diverged species. Public databases contain a large number of mustelid cyt b sequences published in a number of reports. Our aim in this study was to resolve the main disagreements about phylogenetic relationships and taxonomy among the Mustelidae. In order to do this, we pooled all mustelid cyt b sequences published before December 2002 and complemented them with sequences produced in our own laboratory from the same region of relevant species to obtain an extensive molecular phylogeny for the mustelids. We included several individuals for certain species to increase the resolution and reliability of the data set (e.g. Gagneux et al. 1999). We explored phylogenetic relationships with neighborjoining and Bayesian methods (Huelsenbeck & Ronquist 2001), and proposed taxonomic changes accordingly. Many authors have warned that it may be desirable to use a combination of different genomic markers to guarantee the accuracy of phylogenetic inference (e.g. Kluge 1989). We combined cyt b mitochondrial and nuclear data from a highly variable and informative complex repetitive flanking region in 17 relevant mustelid species to clarify the deepest evolutionary roots of the mustelid group. Materials and methods Sequence acquisition We collected a total of 122 entire (1140 bp) and 105 partial (337 bp) sequences of the cyt b gene from GenBank/ EMBL. These samples originated from 31 and 33 species, respectively, these representing four different mustelid subfamilies (Mustelinae, Lutrinae, Melinae and Taxidiinae) and the skunk family (Mephitidae). Three species of procyonids (racoon, Procyon lotor; olingo, Bassaricyon gabbii; and red panda, Ailurus fulgens), and one otarid (Antarctic fur seal, Arctocephalus gazella) were used as outgroups. In addition to data obtained from public databases, we sequenced 14 partial cyt b sequences and 33 Mel08 complex repetitive flanking regions (Mel08fr). Mel08 is a complex repetitive region which includes up to four different repetitive motives showing size variability in mustelids, skunks, procyonids, otarids and phocids (DomingoRoura et al. unpublished data). This region was first described in the Eurasian badger, Meles meles, where it shows a GTG2T2C2TG(CA)6C4AC5AC(CA)2G2 repetitive motive (Domingo-Roura 2002). The species analysed are given in Appendices 1 and 2. Zoologica Scripta, 33, 6, November 2004, pp481– 499 • © The Norwegian Academy of Science and Letters J. Marmi et al. • Phylogeny, evolution and taxonomy of the Mustelidae A fragment of 402 bp of the cyt b gene was amplified using primers L14724 (5′-GATATGAAAAACCATCGTTG-3′ ) and H15149 (5′-CTCAGAATGATATTTGTCCTCA-3′ ) (Masuda & Yoshida 1994). Twenty-five microlitre polymerase chain reactions (PCRs) consisted of 100 ng of target DNA, 16.6 mM (NH4 )2SO4, 67.0 mM Tris–HCl (pH = 8,8), 0.01% Tween-20, 2.5 mM MgCl2, 2.5 mM of each nucleotide, 1.7 pmol of each primer and 0.85 units of Taq DNA polymerase (Ecogen, Madrid, Spain). The program used for thermal cycling was: an initial cycle of denaturation at 94 °C for 5 min, 30 cycles divided into three steps of 1 min each at 94 °C (denaturing), 49 °C (annealing) and 72 °C (extension), and a final extension at 72 °C for 5 min. The Mel08 complex repetitive region was amplified using primers Mel08 F (5′-CTGCCCTTAACTGTAGTC-3′) and Mel08 R (5′-CCTTACCATCCATCAGCTTC-3′) (DomingoRoura 2002). The PCR conditions were as described for cyt b gene amplification and the program used for thermal cycling was: an initial cycle of denaturation at 94 °C for 3 min, 32 cycles divided into three steps of 45 s each at 94 °C (denaturing), 52 °C (annealing) and 72 °C (extension), and a final extension at 72 °C for 5 min. The amplification products were purified with Geneclean (Qbiogen, Carlsbad, CA, USA) and sequenced using the dRhodamine Terminator Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA) following the manufacturer’s instructions. Each reaction tube contained 4 µL of PCR product, 1.9 µL of H2O, 1.6 pmol of primer and 2.5 µL of Sequencing Kit in a total volume of 10 µL. Sequencing reactions were precipitated and run in an ABI Prism 377TM automated DNA sequencer (Applied Biosystems). Data analyses Sequences were aligned using CLUSTALW (Thompson et al. 1994) and double-checked visually. Sequences from both sides of the Mel08 complex repetitive region were combined to obtain a total of 154 bp of nuclear sequence. Mel08fr sequences contained nucleotide deletions ranging from 1 to 16 bp in the skunks as well as in the outgroups, and these deletions were treated as a fifth mutational state. We computed the number of polymorphic sites and the number of synonymous and non-synonymous sites for the entire cyt b gene data set using DNASP v.3.51 (Rozas & Rozas 1999). We estimated the shape parameter (α) of the gamma distribution to correct for substitution rate heterogeneity across the sequence using MRBAYES v.2.01 (Huelsenbeck & Ronquist 2001). The average number of substitutions, transitions and transversions between each pair of species, as well as the average number of substitutions within each species, were computed for each locus assuming the gamma-corrected Tamura–Nei model and using MEGA v.2.1 (Kumar et al. 2001). This model, as with the other substitution models used in this work, assumes substitution rate differences between nucleotides and inequality of nucleotide frequencies ( Tamura & Nei 1993). A recent report based on tree distances (also known as ‘patristic distances’) of complete cyt b sequences, suggested a reference mean value of 0.2 substitutions per site (or 0.1 substitutions per site per lineage) separating different genera (Castresana 2001). Even if the sole use of molecular data for taxonomic classification has been debated (Seberg et al. 2003) we also obtained a matrix of patristic distances, between pairs of species, from complete sequences of cyt b with the aim of discussing the application of this measure at the taxonomy of genera within the Mustelidae. In this case we assumed the HKY model (Hasegawa et al. 1985), using a discrete gamma distribution with six rate categories, and a neighbor-joining tree with branch lengths further optimized by maximum likelihood. See Castresana (2001) for further details. The number of transitions was plotted against the number of transversions, to see the level of transitional saturation of sequences for both loci. In addition, Mantel’s test was performed between matrices of transitions and transversions using ARLEQUIN v.2.000 (Schneider et al. 2000). We included and excluded skunks to obtain a correlation coefficient. Phylogenetic analyses were performed for the entire cyt b data set. A neighbor-joining tree was constructed using DNADIST — assuming the F84 model (Felsenstein & Churchill 1996), gamma-distributed rates across sites, and a transition/ transversion ratio equal to 7.8 as estimated in Koepfli & Wayne (1998) — and NEIGHBOR programs. Bootstrap values (B) were obtained from 1000 replications using SEQBOOT, and the consensus tree was obtained with CONSENSE. All these programs are included in PHYLIP v.3.6a3 (Felsenstein 2002). A Bayesian tree was constructed with MRBAYES using the following parameters: ‘lset nst = 6’ (the general time reversible model); ‘sitepartition = bycodon’ (partition is defined for each base position within codons); ‘rates = ssgamma’ (specific gammadistributed rate variation for each partition); and ‘basefreq = estimate’ (estimated proportion of base types from the data). In the Monte Carlo process, four chains ran simultaneously for 1 200 000 generations. Trees were sampled every 100 generations for a total of 12 000 trees in the initial sample. ‘Stationarity’ was determined to have occurred by the 2000th tree and therefore ‘burnin’ was completed by this stage (i.e. the first 2000 trees were discarded). The whole procedure was repeated three times and the tree topologies obtained were the same. To support the deepest evolutionary roots of the mustelids we estimated a Bayesian tree for Mel08fr sequences, assuming the same parameters as used for the cyt b gene, but without partition and using ‘rates = gamma’ (gamma-distributed rate variation). In addition, we repeated the analysis using combined data from these two independent loci. We obtained © The Norwegian Academy of Science and Letters • Zoologica Scripta, 33, 6, November 2004, pp481– 499 483 Phylogeny, evolution and taxonomy of the Mustelidae • J. Marmi et al. partial consensus cyt b sequences with BIOEDIT Sequence Alignment Editor v.5.0.9 (Hall 1999). We tested these cyt b sequences and the Mel08fr sequences for congruence using the partition homogeneity test (Farris et al. 1995) in PAUP v.4.0b10 (Swofford 2002). This involved 1000 replicates and the heuristic search option with a simple addition sequence. The phylogenetic relationships of taxa included in the analysis of combined loci were also estimated by Bayesian inference as in the cyt b analysis, except that in this case ‘sitepartition = bygene’ (partition is defined by locus). In the cyt b gene, third-position transversions accumulate approximately linearly with time for up to 80 million years or more (Irwin et al. 1991). Therefore, we used the mean pairwise divergence of third-position transversions among mustelid taxa, and calibrated the rate of substitution with the divergence time between mustelids and procyonids. We estimated divergence time among mustelid lineages following Koepfli & Wayne (1998) except that we did not calibrate the molecular clock with the genus Mustelavus as it has been proposed recently that this genus is neither a mustelid nor a member of the mustelid–procyonid clade (Wolsan 1999). Instead we based divergence times on the oldest mustelid and procyonid fossils known, Plesictis dated at 24.3 million years ago and Pseudobassaris dated at 28.5 million years ago, respectively ( Wolsan 1999), which suggested a rate of 0.67% third-position transversions per million years. Fig. 1 A, B. Plots of the number of transitions per site against the Results number of transversions per site to detect transitional saturation. —A. The complete cytochrome b gene. —B. Mel08fr. Sequence variability The entire cyt b gene (1140 bp) had 552 (48.4%) variable sites, of which 495 (43.4%) were maximum parsimony informative within the Mustelidae including skunks. If skunks were removed, the number of variable sites decreased to 534 (46.8%). The mean ratio of synonymous and non-synonymous substitutions for pairwise comparisons was 7.9. Plots of the number of transitions against the number of transversions showed that transitions reached saturation quickly in pairwise comparisons among mustelid species ( Fig. 1A). The plateau of transitional saturation was around 0.12 transitions per site. The upper limit of transversions among mustelids was near 0.05, whereas in comparisons among mustelids and skunks the number of transversions ranged between 0.06 and 0.08. Mantel’s test gave a higher correlation coefficient between transitions and transversions excluding skunks (r = 0.63) rather than including them (r = 0.41), although in both cases this correlation was highly significant (P < 0.001). The shape parameter (α) was equal to 0.23. Mel08fr (154 bp) showed 43 (27.9%) variable sites when skunks were included and 23 (14.9%) when they were not. Of the 43 variable sites, 33 (21.4%) were maximum parsimony informative. The numbers of transitional and transversional differences in comparisons among mustelids and skunks were higher than in comparisons within mustelids (Fig. 1B). This locus did not show transitional saturation. The correlation coefficient between transitions and transversions when skunks were excluded was equal to 0.4 (P < 0.001), whereas when skunks were included it rose to 0.92 (P < 0.001). The shape parameter for this locus was equal to 2.03. The combined cyt b–Mel08fr alignment consisted of 491 bp with 132 maximum parsimony informative polymorphic sites within the Mustelidae including skunks. The partial cyt b gene and the Mel08fr region had 29 and 21% maximum parsimony informative polymorphic sites, respectively. The shape parameter for the combined loci was equal to 0.32. Ranges and values of sequence divergence were obtained for the two loci for comparisons at different taxonomic levels (Table 1) and for comparisons between pairs of species ( Tables 2, 3). The range of intraspecific sequence divergence was very large. Yellow-throated marten, Martes flavigula, showed the highest level of intraspecific divergence due to the high number of polymorphic sites found between sequences obtained by Kurose et al. (1999) and Hosoda et al. (2000). For Mel08fr, only the Eurasian badger, Meles meles, and the eastern spotted skunk, Spilogale putorius, showed intraspecific 484 Zoologica Scripta, 33, 6, November 2004, pp481– 499 • © The Norwegian Academy of Science and Letters J. Marmi et al. • Phylogeny, evolution and taxonomy of the Mustelidae Table 1 Percentage of sequence divergence based on pairwise comparisons within the Mustelidae. Comparisons were made at different taxonomic levels using gamma-corrected Tamura–Nei distances for the complete cytochrome b gene (cyt b) and for Mel08fr. As we found that Martes pennanti, Gulo gulo, Amblonyx cinereus, Aonyx capensis, Meles meles and Arctonyx collaris each had an unclear taxonomic status, these species were excluded from intergeneric and intrageneric comparisons within subfamilies. cyt b Mel 08fr Taxonomic level Mean Range Mean Range Overall Mustelidae Between subfamilies Between genera Between genera within Mustelinae Between genera within Lutrinae Between species within genus Mustela Between species within genus Martes Between species within genus Lutra Between species within genus Lontra 35.6 44.4 41.9 27.7 40.8 15.2 12.3 27.0 14.7 0.2–63.8 34.0–53.0 25.6–59.0 NA 25.6–55.8 0.2–37.0 2.8–25.0 NA 7.7–19 3.7 5.2 4.3 4.9 2.3 1.2 0.9 NA NA 0.0–7.5 4.0–6.0 0.7–7.5 NA 0.7–3.4 0.0–2.7 0.7–1.4 NA NA NA, not available. variation (0.7% of sequence divergence), but few individuals were analysed per species. We evaluated the values of the patristic distances shown in Table 2 and how far these values were from the mean value of 0.2 substitutions per site (20% of sequence divergence), that is 0.1 substitutions per site per lineage, obtained for different genera to explore the consistency of these distances with the current mustelid generic taxonomy and tree topologies obtained in this study. In general, the patristic distances agreed with the current classification of genera within Mustelidae with most mustelid species classified within the same genus showing patristic divergences below 20% ( Table 2). However, among the values which deviated more from this 20%, the wolverine, Gulo gulo, had an average sequence divergence of 18.7% in relation to marten species. Within the Lutrinae, the short-clawed otter, Amblonyx cinereus, and the Cape clawless otter, Aonyx capensis, showed average sequence divergences from species belonging to the genus Lutra, which ranged between 14.8 and 20.6%, and sequence divergence between the genera Amblonyx and Aonyx was 13.4%. Within the Melinae, sequence divergence between the genera Meles and Arctonyx was 14.7%. We estimated divergence times of mustelid lineages based on cyt b third-codon transversions and described six radiation episodes after the separation of the Mustelidae from the Procyonidae and Mephitidae ( Table 4). We suggest that in a first episode, between the Early and the Middle Miocene, subfamilies and the oldest genera may have separated. The second radiation episode might have occurred during the Late Miocene when we estimate that Old and New World lineages, all genera, and the oldest lineages leading to the present species separated. In the remaining four episodes proposed to have occurred between the Late Miocene and the Pleistocene, the remaining lineages, which lead to the present species, may have separated. Phylogenetic relationships Entire cyt b gene data set. Terminal branches were longer than internal branches in both trees, although the resolution of internal branches was better in the Bayesian (Fig. 2) than in the neighbor-joining tree (Fig. 3). The Bayesian tree also had deeper internodes supported by high clade credibility values (CCV). The two methods recovered the same clades within the Mustelidae, and skunks appeared either as a sister group to the remaining clades including the Procyon– Bassaricyon clade (Fig. 2) or within the paraphyletic Procyonidae (Fig. 3). The monophyly of the Mustelidae, excluding skunks, was strongly supported by Bayesian inference (CCV = 100%) and moderately supported by the neighbor-joining tree method (B = 75%). The clade comprising Melinae and Taxidiinae subfamilies was the sister group of the rest of mustelid subfamilies. Both trees suggested that the Mustelinae was paraphyletic as it included the subfamily Lutrinae, although statistical support for this grouping was low (CCV = 81%, B < 50%). In the Bayesian tree, the Lutrinae formed a clade with the genus Mustela (CCV = 91%), but this did not include the genera Martes or Gulo (Fig. 2). Within the Lutrinae, both methods clearly supported the monophyly of the genus Lontra (CCV = 100%, B = 100%) and the clade formed by Lutra lutra, Amblonyx cinereus and Aonyx capensis, as well as the separation of these three species from other otters (CCV = 100%, B = 86%). The nodes of the spot-necked otter, Lutra maculicollis, and the sea otter, Enhydra lutris, were more strongly supported by Bayesian inference, in which they were grouped together with Lutra lutra, Amblonyx cinereus and Aonyx capensis (CCV = 96%, Fig. 2). Two clades were recognized within the Mustelinae, one for the genus Mustela and another for the genera Martes and Gulo together. The Bayesian tree showed that each one of these two clades was monophyletic (CCV = 100% in both cases). The nodes within the genus Mustela were more strongly supported by the Bayesian than by the neighbor-joining tree. The American mink, Mustela vison, was the first species to diverge from the other members of this genus, followed by the ermine, Mustela erminea. The least weasel, Mustela nivalis, and the mountain weasel, Mustela altaica, were identified as sister taxa (CCV = 91%, B = 55%), while the Japanese weasel, Mustela itatsi, and the kolinsky, Mustela sibirica, appeared as sister groups of the European mink, Mustela lutreola, the steppe polecat, Mustela eversmanii, and the European polecat, Mustela putorius clade (CCV = 100%, B = 98%). These last three species were particularly closely related. Within the © The Norwegian Academy of Science and Letters • Zoologica Scripta, 33, 6, November 2004, pp481– 499 485 Zoologica Scripta, 33, 6, November 2004, pp481– 499 • © The Norwegian Academy of Science and Letters Mama Mazi Mame Maam Mafo Mafl Mape Gugu Muev Mulu Mupu Mung* Musi Muit Mual Muni Muer Muvi Mufr * Lulu Luma Amci Aoca Enlu Lofe Lolo Loca Ptbr Meme Arco Tata Mpmp Sppu Mama Mazi Mame Maam Mafo Mafl Mape Gugu Muev Mulu Mupu Mung* Musi Muit Mual Muni Muer 1.1(0.7) 2.8 3.6 6.9 12.0 18.1 29.7 28.8 26.7 28.5 26.4 22.8 30.0 31.4 26.1 30.1 25.2 47.6 35.9 31.5 35.1 42.4 34.6 36.9 44.2 30.6 37.0 48.1 47.6 45.2 31.8 74.0 50.9 2.9 0.7(1.1) 4.9 7.8 13.8 20.2 32.6 29.9 27.7 29.4 27.3 24.2 32.2 32.5 26.6 29.5 26.1 44.2 39.2 32.6 37.4 47.1 32.2 39.6 46.5 33.5 41.4 50.8 48.2 45.1 35.1 77.6 54.0 3.0 4.3 0.7(0.3) 6.1 11.9 15.9 27.9 28.3 24.7 26.3 24.4 17.2 27.0 30.8 23.7 24.8 23.0 45.0 30.2 29.2 32.6 37.2 33.4 32.8 41.3 29.5 35.6 47.5 48.1 45.6 28.5 73.9 45.6 5.7 7.0 5.1 1.8(1.2) 14.1 21.4 32.2 32.3 29.2 31.2 29.0 26.1 30.9 36.9 27.2 31.3 26.9 43.4 32.0 34.3 39.5 41.8 35.1 38.3 42.0 34.4 40.0 54.5 46.2 49.7 33.2 69.6 51.3 9.2 10.5 8.7 10.9 2.2(1.4) 25.0 31.1 31.0 36.7 36.7 36.4 19.8 35.7 35.7 27.2 30.3 29.5 44.9 30.6 34.6 31.4 47.0 38.3 32.2 45.9 33.3 37.3 48.0 48.9 36.1 34.8 72.5 47.0 10.5 11.9 10.0 12.2 13.8 5.6(22.6) 36.0 28.3 33.7 34.5 33.7 27.6 37.5 35.8 33.6 39.1 29.1 52.4 48.1 28.8 40.7 38.6 38.6 40.8 43.1 37.5 43.7 50.7 54.9 58.5 37.9 71.0 51.6 19.4 20.7 18.8 21.1 22.7 19.1 0.06(0.0) 34.7 34.0 36.4 33.7 27.8 35.0 31.0 41.8 36.2 27.9 41.1 34.3 36.8 40.5 49.6 36.7 31.9 44.4 41.8 42.4 56.3 47.9 52.7 47.2 79.1 59.8 17.8 19.1 17.2 19.5 21.1 17.5 20.9 0.8(2.2) 37.9 37.5 37.5 27.7 42.1 35.0 34.3 34.1 30.3 46.9 45.6 34.7 42.5 38.9 28.5 36.6 49.6 42.5 42.0 54.0 55.1 52.7 33.8 59.9 62.1 22.3 23.6 21.8 24.0 25.6 22.1 24.9 23.8 NA(0.7) 0.8 0.2 2.7 4.5 8.3 12.4 12.9 14.1 30.6 27.2 32.8 35.7 38.5 33.5 30.0 40.9 38.2 38.0 39.9 42.7 52.2 34.0 60.4 50.6 22.1 23.4 21.6 23.8 25.4 21.9 24.7 23.6 0.8 NA(0.3) 1.1 2.7 4.3 7.3 13.9 12.7 13.9 31.3 29.7 31.1 35.1 36.2 32.9 29.1 41.4 37.0 37.2 37.1 45.6 50.1 36.6 58.7 47.6 22.7 24.0 22.1 24.3 26.0 22.4 25.3 24.2 0.3 1.1 0.4(0.7) 3.2 4.7 8.2 12.5 12.9 14.2 30.9 25.8 33.0 35.9 38.6 33.8 29.9 40.9 37.9 38.0 40.1 42.9 52.1 33.8 60.3 50.4 NA NA NA NA NA NA NA NA NA NA NA NA(0.0) 4.5 6.2 12.2 7.4 10.2 31.3 28.0 29.7 23.8 30.3 25.2 19.1 30.1 24.7 32.9 64.6 51.0 46.4 34.8 55.0 46.8 22.4 23.7 21.8 24.0 25.7 22.1 25.0 23.9 4.0 3.8 4.3 NA 0.6(1.1) 7.3 13.7 13.2 13.5 31.3 32.8 33.3 36.8 35.8 30.6 33.4 38.6 34.9 36.8 43.6 44.2 50.2 36.3 61.9 46.2 22.3 23.6 21.8 24.0 25.6 22.1 24.9 23.8 6.3 6.1 6.7 NA 6.4 NA(0.8) 20.9 13.9 13.3 30.5 27.8 32.6 33.3 37.5 30.4 29.8 41.3 36.0 32.9 41.8 39.8 47.7 35.8 48.8 42.3 22.7 24.0 22.2 24.4 26.0 22.5 25.3 24.2 11.9 11.6 12.2 NA 11.9 11.9 0.7(0.0) 11.0 12.0 33.0 21.4 32.2 30.9 38.4 40.5 34.8 46.1 36.0 39.9 52.4 49.6 45.6 36.1 56.7 52.8 23.1 24.4 22.5 24.7 26.4 22.8 25.7 24.6 12.2 12.0 12.6 NA 12.3 12.2 8.5 0.6(0.7) 13.2 37.0 25.6 41.2 30.5 44.8 39.2 32.2 44.7 39.4 43.6 50.2 47.0 46.9 36.7 69.3 53.2 21.4 22.7 20.8 23.0 24.7 21.1 24.0 22.9 10.5 10.3 10.8 NA 10.6 10.5 9.5 9.9 2.4(2.1) 32.2 20.0 29.2 29.3 35.1 29.5 24.5 34.4 31.1 33.5 46.0 42.0 44.2 33.5 72.8 58.9 Phylogeny, evolution and taxonomy of the Mustelidae • J. Marmi et al. 486 Table 2 Percentage of mean cytochrome b gene (cyt b) sequence divergence, based on gamma-corrected Tamura–Nei distances for 1140 bp, among species (below diagonal) and within species (diagonal). Species marked with an asterisk and values within parentheses are based on 337 bp. Cyt b patristic distances based on one sequence of 1140 bp per species are above the diagonal. Mama Mazi Mame Maam Mafo Mafl Mape Gugu Muev Mulu Mupu Mung* Musi Muit Mual Muni Muer Muvi Mufr * Lulu Luma Amci Aoca Enlu Lofe Lolo Loca Ptbr Meme Arco Tata Mpmp Sppu Muvi Mufr* Lulu Luma Amci Aoca Enlu Lofe Lolo Loca Ptbr Meme Arco Tata Mpmp Sppu 24.7 26.0 24.1 26.3 28.0 24.4 27.3 26.2 18.6 18.4 19.0 NA 18.7 18.6 19.0 19.4 17.7 0.6(0.4) 19.2 46.8 48.8 46.6 46.0 37.6 49.3 46.1 42.2 51.3 52.3 57.6 50.9 63.8 68.8 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA(NA) 58.8 45.8 26.9 41.8 37.9 35.9 32.1 37.9 70.8 36.9 35.1 33.5 41.7 64.7 23.0 24.3 22.4 24.6 26.3 22.7 25.6 24.5 22.1 21.9 22.4 NA 22.1 22.1 22.5 22.8 21.1 24.4 NA 0.09(0.1) 27.0 23.3 22.9 25.6 31.6 33.0 34.8 49.8 54.0 54.4 36.3 66.2 49.4 23.5 24.8 23.0 25.2 26.8 23.3 26.1 25.0 22.6 22.4 23.0 NA 22.7 22.6 23.0 23.4 21.7 25.0 NA 18.3 NA(NA) 33.9 34.6 30.0 44.4 41.7 44.1 61.7 56.7 58.0 38.9 81.1 53.2 25.3 26.6 24.7 26.9 27.2 25.0 27.9 26.8 24.4 24.2 24.7 NA 24.5 24.4 24.8 25.1 23.4 26.7 NA 16.2 20.6 NA(0.3) 20.2 32.1 35.8 34.9 40.9 58.1 53.9 63.8 44.6 65.9 50.5 23.9 25.2 23.3 25.5 27.2 23.6 26.5 25.4 23.0 22.8 23.3 NA 23.0 23.0 23.4 23.7 22.0 25.3 NA 14.8 19.2 13.4 NA(NA) 25.3 37.1 38.9 43.0 49.4 47.4 48.8 35.4 77.8 56.1 22.9 24.3 22.4 24.6 26.2 22.7 25.5 24.4 22.0 21.8 22.4 NA 22.1 22.0 22.4 22.8 21.1 24.4 NA 18.4 19.0 20.7 19.3 0.06(0.1) 42.9 41.0 40.8 39.3 51.2 47.4 43.6 65.0 45.3 26.3 27.6 25.7 27.9 29.6 26.0 28.9 27.7 25.4 25.1 25.7 NA 25.4 25.3 25.7 26.1 24.4 27.7 NA 24.3 24.8 26.6 25.2 24.2 NA(NA) 7.7 19.0 60.0 48.7 51.1 45.9 77.3 70.1 25.1 26.4 24.5 26.7 28.4 24.8 27.7 26.6 24.2 24.0 24.5 NA 24.3 24.2 24.6 24.9 23.2 26.5 NA 23.1 23.7 25.4 24.0 23.1 6.1 NA(NA) 17.5 52.4 51.1 48.5 46.9 78.5 67.4 25.2 26.6 24.7 26.9 28.6 25.0 27.9 26.7 24.3 24.1 24.7 NA 24.4 24.3 24.7 25.1 23.4 26.7 NA 23.3 23.8 25.6 24.2 23.2 12.4 11.2 NA(NA) 55.0 56.9 49.6 47.8 71.7 71.1 26.9 28.2 26.4 28.6 30.2 26.7 29.5 28.4 26.0 25.8 26.4 NA 26.1 26.0 26.4 26.8 25.1 28.4 NA 26.7 27.2 29.0 27.6 26.6 29.9 28.8 28.9 NA(NA) 59.0 57.4 46.4 78.6 73.4 28.3 29.6 27.7 29.9 31.6 28.0 30.9 29.7 30.1 29.9 30.5 NA 30.2 30.1 30.5 30.9 29.2 32.5 NA 30.8 31.3 33.1 31.7 30.7 34.1 32.9 33.1 34.7 4.4(4.4) 16.1 53.0 85.8 72.2 28.6 29.9 28.1 30.3 31.9 28.4 31.2 30.1 30.5 30.3 30.9 NA 30.6 30.5 30.9 31.3 29.6 32.9 NA 31.1 31.7 33.5 32.1 31.1 34.4 33.3 33.4 35.1 14.7 NA(NA) 53.4 86.4 62.2 23.3 24.6 22.7 24.9 26.6 23.0 25.9 24.8 25.2 24.9 25.5 NA 25.2 25.2 25.6 25.9 24.2 27.5 NA 25.8 26.4 28.1 26.7 25.8 29.1 27.9 28.1 29.7 28.3 28.7 NA(0.6) 66.1 56.0 42.7 44.0 42.2 44.4 46.0 42.5 45.3 44.2 44.6 44.4 45.0 NA 44.7 44.6 45.0 45.3 43.6 46.9 NA 45.2 45.8 47.6 46.1 45.2 48.5 47.4 47.5 49.2 44.4 44.8 42.8 NA(0.0) 36.6 39.5 40.8 38.9 41.1 42.8 39.2 42.1 41.0 41.4 41.1 41.7 NA 41.4 41.4 41.7 42.1 40.4 43.7 NA 42.0 42.6 44.3 42.9 42.0 45.3 44.1 44.3 45.9 41.2 41.6 39.5 20.2 NA(0.3) See Appendix 1 for full species names. NA, not available. 487 J. Marmi et al. • Phylogeny, evolution and taxonomy of the Mustelidae © The Norwegian Academy of Science and Letters • Zoologica Scripta, 33, 6, November 2004, pp481– 499 Table 2 Continued. Phylogeny, evolution and taxonomy of the Mustelidae • J. Marmi et al. Table 3 Percentage of mean Mel08fr sequence divergence based on gamma-corrected Tamura–Nei distances for 154 bp among species (below diagonal) and within species (diagonal). Mama Mame Mafo Gugu Mulu Mupu Mung Mual Muni Muer Lulu Amci Enlu Loca Meme Mpmp Sppu Mama Mame Mafo Gugu Mulu Mupu Mung Mual Muni Muer Lulu Amci Enlu Loca Meme Mpmp Sppu 0.0 0.7 1.4 1.3 4.9 4.9 4.9 4.8 6.3 4.8 4.1 6.3 4.8 5.5 6.0 22.4 22.2 NA 0.7 0.7 4.2 4.2 4.2 4.1 5.6 4.1 3.4 5.5 4.1 4.8 5.3 21.0 20.8 0.0 1.3 4.9 4.9 4.9 4.8 6.3 4.8 4.1 6.3 4.8 5.6 6.0 22.5 22.3 NA 3.4 3.4 3.4 3.4 4.8 3.4 2.7 4.8 3.4 4.1 4.5 20.2 19.9 NA 0.0 0.0 1.3 1.3 1.3 2.0 4.1 2.7 3.4 5.3 21.1 20.8 0.0 0.0 1.3 1.3 1.3 2.0 4.1 2.7 3.4 5.3 21.1 20.8 NA 1.3 1.3 1.3 2.0 4.1 2.7 3.4 5.3 21.1 20.8 NA 1.3 2.7 2.0 4.1 2.8 4.9 5.2 18.9 19.9 NA 2.7 3.4 5.6 4.1 4.8 6.7 20.7 21.8 0.0 3.4 5.6 4.1 3.4 6.8 23.0 22.8 0.0 2.0 0.7 2.7 4.5 20.8 20.6 NA 2.7 4.8 6.7 21.2 20.9 NA 3.4 5.3 19.4 19.2 NA 7.5 22.1 21.9 0.7 19.5 19.2 NA 3.8 0.7 See Appendix 1 for full species names. NA, not available. Divergence event Million years ago Late Eocene –Late Oligocene: origin of Mustelidae Mustelidae + (Procyonidae, Mephitidae) 37–23 Early–Middle Miocene: split of Mustelidae subfamilies, differentiation of first genera (Melinae, Taxidiinae) + (Lutrinae, Mustelinae) 20–12 Lutrinae + Mustelinae 15–8 Taxidiinae + Melinae 17–16 (Gulo, Martes) + Mustela 14–11 Late Miocene: split of Old and New World lineages, differentiation of genera and most divergent species within genera Lontra + (other otter species) 10.6 –7.5 Pteronura + (other otter species) 10.6 –7.0 Enhydra lutris + (Lutra, Aonyx, Amblonyx) 10.1–7.5 Mustela vison + (other Mustela species) 9.5–6.6 Gulo + Martes 8.6–5.8 Martes pennanti + (remaining Martes species) 7.6–6.6 Late Miocene–Early Pliocene: differentiation within genera Lutra maculicollis + (Lutra lutra, Aonyx, Amblonyx) 6.6–5.2 Lutra lutra + (Aonyx, Amblonyx) 5.7–4.5 Martes flavigula + (remaining Martes species) 6.1–4.6 Early Pliocene: differentiation within genera Mustela erminea + (remaining Mustela species) 4.8–3.0 (Mustela nivalis, Mustela altaica) + (remaining Mustela species) 4.0–3.4 Aonyx + Amblonyx 3.6 Late Pliocene: differentiation within genera Meles + Arctonyx 3.3 Martes foina + (Martes americana, Martes zibellina, Martes martes, Martes melampus ) 3.1–2.2 Lontra canadensis + (Lontra felina, Lontra longicaudis ) 2.4–1.9 Pleistocene: differentiation within genera Martes americana + (Martes zibellina, Martes martes, Martes melampus ) 1.8–0.9 Lontra felina + Lontra longicaudis 1.2 Martes melampus + (Martes martes, Martes zibellina) 1.0 (Mustela itatsi, Mustela sibirica) + (Mustela lutreola, Mustela eversmannii, Mustela putorius ) 0.7–0.4 Mustela lutreola + (Mustela eversmannii, Mustela putorius ) 0.15 488 Table 4 Estimates of divergence times of radiations and lineage splits within the Mustelidae. These were based on thirdcodon tranversions from complete cytochrome b gene sequences. Taxonomic changes proposed in this work were assumed to define divergence events, although in this table we use current nomenclature for taxa. Zoologica Scripta, 33, 6, November 2004, pp481– 499 • © The Norwegian Academy of Science and Letters J. Marmi et al. • Phylogeny, evolution and taxonomy of the Mustelidae Fig. 2 The Bayesian tree obtained from complete cytochrome b sequences (1140 bp). The numbers indicate clade credibility values which exceeded 50%, as observed in 10 000 trees, and the ln likelihood was −14 055.33. martens, four groups were distinguished: (i) Gulo gulo; (ii) the fisher, Martes pennanti; (iii) Martes flavigula and (iv) the beech marten, Martes foina, the American marten, Martes americana, the pine marten, Martes martes, the sable, Martes zibellina; and the Japanese marten, Martes melampus. Bayesian analysis was not able to resolve the phylogenetic relationships within this last group of marten species. Mel08fr and combined partial cyt b gene and Mel08fr data sets. The P-value resulting from the partition homogeneity test conducted with PAUP was 0.41, P = 0.05, indicating congruence between cyt b and Mel08fr data sets (Farris et al. 1995). Combined (Fig. 4A) and Mel08fr (Fig. 4B) trees were consistent with the exclusion of skunks from the mustelid clade, and in the monophyly of the remaining taxa included within the Mustelidae (CCV = 100% in both cases). The Mel08fr tree supported the monophyly of martens and the wolverine (genera Martes and Gulo, CCV = 100%) but was not able to resolve the remaining phylogenetic relationships within the Mustelidae. However, the combined data also provided strong support for a sister group relationship between the Lutrinae and the genus Mustela, and for the monophyly of the Lutrinae, martens including the wolverine and the genus Mustela (CCV = 100% in all cases). Discussion Usefulness of the cyt b gene and the Mel08fr region for clarifying phylogenetic relationships of the Mustelidae Analysis of the cyt b gene is a commonly used tool for phylogenetic inference within carnivore families (e.g. Ursidae, © The Norwegian Academy of Science and Letters • Zoologica Scripta, 33, 6, November 2004, pp481– 499 489 Phylogeny, evolution and taxonomy of the Mustelidae • J. Marmi et al. Fig. 3 The neighbor-joining tree obtained from complete cytochrome b sequences (1140 bp). The numbers indicate bootstrap values which exceeded 50%, as obtained from 1000 iterations. Waits et al. 1999; Otariidae, Wynen et al. 2001). Within the Mustelidae, excluding skunks, the cyt b gene strongly supported the monophyly of the basal and terminal clades (Figs 2, 3). However, the relationships among subfamilies or among some genera were not clearly resolved, possibly due to the effects of transitional saturation and /or fast radiation. In addition, cyt b sequence divergences including patristic distances should be used with caution in genera diagnosis because it is known that the cyt b gene evolves at different rates, even within mammals, and taxonomic changes proposed according to these distances should be effective only if they agree with tree topologies and morphological synapomorphies. 490 In this study, the levels of sequence divergence showed wide variations across the mustelid species (Tables 1, 2). This could be due to some lineages leading to modern species being much older than others. However, other factors, such as differences in evolutionary rates, generation time, and selection could also result in wide variations in sequence divergence. In addition, the family includes from highly endangered to widely distributed species and differences in effective population sizes are likely to be involved. Repetitive flanking region sequences might show an adequate degree of conservation for PCR primers to amplify across species which diverged millions of years ago, and yet Zoologica Scripta, 33, 6, November 2004, pp481– 499 • © The Norwegian Academy of Science and Letters J. Marmi et al. • Phylogeny, evolution and taxonomy of the Mustelidae Fig. 4 A, B. Bayesian trees for combined loci. —A. The partial cytochrome b gene (337 bp) and Mel08fr (154 bp). —B. Mel08fr alone. The numbers indicate those clade credibility values which exceeded 50%, as observed in 10 000 trees. Values of ln likelihood were −2951.99 and − 624.56, respectively. still be sufficiently variable to contain relevant phylogenetic information (e.g. Zardoya et al. 1996). In spite of its short length, Mel08fr has a high mutation rate. Alone, and when combined with cyt b data, Mel08fr showed good resolution at the basal level of the phylogeny. It supported the monophyly of the Mustelidae (Melinae, Lutrinae and Mustelinae) and the exclusion of skunks (Mephitidae) from the rest of the mustelids, but it failed to resolve some phylogenetic relationships within Mustelidae (Fig. 4). Phylogenetic relationships and taxonomy among Simpsonian subfamilies The classical division of the family Mustelidae into five subfamilies, as proposed by Simpson (1945), has not been supported by molecular data, especially regarding skunks. Our results, based on cyt b transversional and Mel08fr transitional and transversional differences (Fig. 1), and on phylogenetic trees (Figs 2– 4), support the assignment of skunks to the Mephitidae family, as has been proposed by other authors (e.g. Ledge & Arnason 1996; Dragoo & Honeycutt 1997). Once skunks are excluded, badgers remain external to the other two remaining subfamilies analysed in this study, the Lutrinae and the Mustelinae, which are sister groups. The phylogenetic trees suggest that the genus Mustela and the Lutrinae subfamily might share an ancestor, which is not common to the martens (Figs 2, 4), but further research combining different markers will be needed to confirm this. Phylogenetic relationships and taxonomy within the Melinae The monophyly of the Melinae has not been supported by either morphological (Bryant et al. 1993), or molecular techniques (Dragoo & Honeycutt 1997), nor has it been upheld by combining these two sources of data (Bininda-Emonds et al. 1999). Classically, the American badger, Taxidea taxus, was classified within the subfamily Melinae. We found that Taxidea was not closely related to Meles or Arctonyx (Figs 2, 3), an observation which is consistent with the findings of authors who proposed the separation of Taxidea into a monotypic subfamily, Taxidiinae (see Wozencraft 1993; Macdonald 2001). On the other hand, because the genera Meles and Arctonyx formed a monophyletic group in the phylogenetic trees (Figs 2, 3) and both cyt b gene sequence divergence and © The Norwegian Academy of Science and Letters • Zoologica Scripta, 33, 6, November 2004, pp481– 499 491 Phylogeny, evolution and taxonomy of the Mustelidae • J. Marmi et al. patristic distances were low (Table 2), we suggest that these two genera may be classified within the same genus, Meles. We found significant levels of intraspecific sequence divergence in the cyt b gene within the Eurasian badger, Meles meles (Table 2). These sequences belonged to individuals from different geographical origins (see Appendix 1) and formed three clusters which were strongly supported by high bootstrap values: (Europe) + (Central Russia) + ( Japan) ( Figs 2, 3). Based on cyt b sequences, Kurose et al. (2001) suggested that the Japanese badger (Meles meles anakuma) is a different subspecies from the continental badgers (Meles meles meles). Our study predicts the existence of three different subspecies, whereas several reports based on morphology propose the existence of two or three species of Meles (Baryshnikov & Potapova 1990; Abramov 2002). Phylogenetic relationships within the Lutrinae Phylogenetic relationships within the Lutrinae were explored by Koepfli & Wayne (1998). They distinguished three monophyletic groups of otters containing the following genera: (Lontra) + (Lutra, Aonyx, Amblonyx and Enhydra) + (Pteronura). Our work includes sequences already analysed by these authors and nine additional otter cyt b sequences which confirm these three groups. These authors also suggested that the genus Lutra was paraphyletic, because the Eurasian otter, Lutra lutra, was more phylogenetically related to the genera Amblonyx and Aonyx than to its congeneric species, the spotnecked otter, Lutra maculicollis. Our results based on cyt b gene trees (Figs 2, 3) and distances (Table 2) agree with the inclusion of Lutra, Amblonyx and Aonyx, and even Enhydra, within a single genus, Lutra. Phylogenetic relationships and taxonomy within the Mustelinae Two clades were found within the subfamily Mustelinae, one containing the genera Martes and Gulo, and the other containing the genus Mustela. The genus Martes is divided into three subgenera (Anderson 1970). Phylogenetic trees based on cyt b sequences (Figs 2, 3) were consistent with the subdivision of the genus Martes into subgenera Charronia (Martes flavigula) and Martes (Ma. foina, Ma. martes, Ma. zibellina, Ma. melampus and Ma. americana). Neither of the trees based on cyt b sequences (Figs 2, 3, 4A) or values of sequence divergences or patristic distances (Table 2) justify the placement of Gulo gulo as a separate genus. Within the Martes subgenus, all members are closely related, with the possible exception of Martes foina. Stone & Cook (2002) found polytomies within the subgenus Martes and were not able to resolve the phylogenetic relationships of their members. Similar results are reported in our present work (Figs 2, 3). These species maintain allopatric and parapatric distributions, and it has been suggested that they form a superspecies (Anderson 1970). 492 However, none of the marten species is paraphyletic, except Martes martes and Martes zibellina. These two species are not reproductively isolated and successful hybridization between them is possible (Grakov 1994). Recently, the genus Mustela has been divided into nine subgenera according to skull structure, dentition, bacular structure and external characteristics (Abramov 2000). As a result, Mustela erminea and the long-tailed weasel, Mustela frenata, were included within the subgenus Mustela. However, we propose that Mustela frenata should be excluded from this subgenus as this species and Mustela erminea are highly divergent compared with other pairs of Mustela species (Table 2). Abramov (2000) placed Mustela sibirica and Mustela itatsi in the subgenus Kolonokus. However, our phylogenetic trees based on the cyt b gene (Figs 2, 3) suggest that these two species, together with species in the subgenera Putorius (Mustela putorius, Mustela eversmannii and the black-footed ferret, Mustela nigripes) and Lutreola (Mustela lutreola), should be included in the same subgenus. Our results are consistent with Abramov’s (2000) assignment of Mustela nivalis and Mustela altaica to the subgenus Gale. The genetic divergence among Mustela putorius, Mustela eversmannii and Mustela lutreola is similar to, or less than, that found within other Mustela species (Table 2). Moreover, hybridization between pairs of these three species has been reported (Heptner et al. 1967; Ternovsky 1977; Davison et al. 2000). Thus, we think that these three species could be considered as subspecies of a single species, Mustela putorius. On the other hand, in our study, Mustela vison was the most divergent species within Mustela, but we found no support to classify this species as a separate genus named Neovison, as suggested by Abramov (2000). Temporal mode of diversification of mustelid lineages The fossil record suggests that between the Eocene and Early Oligocene epochs, diversification among the Miacids resulted in the production of modern carnivore families (Anderson 1970). Plesictis plesictis, from the Late Oligocene of Cournon (France), is the oldest mustelid fossil known (Wolsan 1999). The first stage of Mustelid radiation occurred in the Old World, and at some point between the Early and Middle Miocene epochs all of the Mustelidae subfamilies known today became recognizable (Anderson 1970; Bryant et al. 1993). Our estimates of divergence times ( Table 4) are either consistent with, or older than, predictions based on fossil evidence. This might be explained by the fact that, in general, gene evolution predates species evolution (Graur & Li 2000), and, moreover, the fossil record for the Mustelidae is incomplete for most lineages (Anderson 1989). We suggest at least six radiation episodes within the Mustelidae ( Table 4), adding a further radiation (during the Pliocene) to those proposed by Hosoda et al. (2000). Zoologica Scripta, 33, 6, November 2004, pp481– 499 • © The Norwegian Academy of Science and Letters J. Marmi et al. • Phylogeny, evolution and taxonomy of the Mustelidae We suggest that the mustelid subfamilies differentiated and that martens and weasels separated within the Mustelinae in the first radiation that we estimate that took place between the Early and the Middle Miocene. On the basis of fossil evidence, the first member of the Lutrinae may have been Paralutra, from the Early Miocene of Europe (Savage & Russell 1983). Plesiogale and Paragale, which have been placed in the same epoch and continent, have been considered early members of the Mustelinae (Wolsan 1993), whereas the earliest known marten, Martes laevidens, was found in the Early Miocene of Germany (Anderson 1994). Dehmictis vorax and Trochictis artenensis from the Early Miocene have been considered the earliest melines (Ginsburg & Morales 2000). We propose that most New World lineages differentiated from Old World lines during the second radiation, which may have occurred in the Late Miocene epoch. This was also the time at which we estimate that martens divided into the Martes and Gulo lineages. Fossil data agree with our estimates of the separation age of Gulo and Enhydra (Anderson 1989; Willemsen 1992). Martes palaeosinensis is the earliest known ancestor of Martes pennanti, and this was discovered in Early Pliocene deposits in China (Anderson 1994). In the USA, fossil records exist for Mustela vison from as far back as the Early Pleistocene (Anderson 1989). However, our molecular estimates for the appearance of both Martes palaeosinensis and Mustela vison are earlier than fossil records suggest. In the third radiation, which we suggest took place between the end of the Late Miocene and the beginning of the Early Pliocene, differentiation within genera increased. Again, however, our estimates of the dates involved are earlier than those suggested by the fossil record. The genus Lutra was first identified in deposits of the Late Pliocene in Asia (Savage & Russell 1983). Martes lydekkeri, the oldest species within Charronia, was found in deposits from the Early Pliocene (Anderson 1994), also in Asia. We propose that lineages which resulted in the present-day otter genera Amblonyx and Aonyx, musteline subgenus Gale, and Mustela erminea separated in the fourth radiation, which may have occurred during the Early Pliocene epoch. Our estimates of the time at which the Mustela erminea lineage appeared agree with the fossil record (Anderson 1989). However, the fossil record suggests an older origin for Aonyx and Amblonyx genera, in the Late Miocene (Radinsky 1968). Finally, we propose that lineages leading to the genera Meles and Arctonyx and our present-day species, differentiated in the most recent two radiations, which we estimate took place between the Late Pliocene and the Pleistocene epochs. The fossil record suggests that the genus Arctonyx separated from the Meles line during the Plaisancien age within the Late Pliocene (Petter 1971). Martes vetus is considered to be the common ancestor of Martes foina and Martes martes, and this species has been found in deposits from the beginning of the Middle Pleistocene of Germany (Anderson 1994). Martes americana and Martes melampus have also occurred in the fossil record since the Middle Pleistocene. Mustela lutreola is, however, only known in the Holocene (Anderson 1989), supporting its unclear specific status proposed in this research. Acknowledgements Researchers and institutions mentioned in the appendices kindly provided samples for this study. We thank A. Ferrando, O. Andrés, A. Pérez-Lezaun and M. Vallès for their help and advice in the laboratory. J. Castresana (Centre de Regulació Genòmica) and A. Susanna (Institut Botànic de Barcelona, CSIC) assisted with data analyses. Sandra Baker provided editorial assistance. The project was financed by the Ministerio de Educación y Cultura, Spain (reference PB981064) and the Departament d’Universitats, Recerca i Societat de la Informació, Generalitat de Catalunya (reference 2001SGR 00285). JM and JF L-G were supported by scholarships from the latter institution (references 2000FI-00698 and 2001FI-00625, respectively). We thank A. Abramov and two anonymous reviewers for providing useful comments to improve the manuscript. References Abramov, A. V. (2000). A taxonomic review of the genus Mustela (Mammalia, Carnivora). Zoosystematica Rossica, 8, 357–364. Abramov, A. V. (2002). Variation of the baculum structure of the Paleartic badger (Carnivora, Mustelidae, Meles). Russian Journal of Theriology, 1, 57–60. Anderson, E. (1970). Quaternary evolution of the genus Martes (Carnivora, Mustelidae). Acta Zoologica Fennica, 130, 1–132. Anderson, E. (1989). The phylogeny of mustelids and the systematics of ferrets. In U. S. Seal, E. T. Thorne, M. A. Bogan & S. H. Anderson (Eds) Conservation Biology and the Black-footed Ferret (pp. 10–20). Connecticut: Yale University Press. Anderson, E. (1994). Evolution, prehistoric distribution and systematics of Martes. In S. W. Buskirk, A. S. Harestad, M. G. Raphael & R. A. Powell (Eds) Martens, Sables and Fishers: Biology and Conservation (pp. 13–25). Ithaca: Cornell University Press. Baryshnikov, G. F. & Potapova, O. R. (1990). Variability of the dental system in badgers (Meles, Carnivora) of the USSR fauna. Zoologicheskii Zhurnal, 69, 84–97. Bininda-Emonds, O. R. P., Gittleman, J. L. & Purvis, A. (1999). Building large trees by combining phylogenetic information: a complete phylogeny of the extant Carnivora (Mammalia). Biological Reviews, 74, 143–175. Bryant, H. N., Russell, A. P. & Fitch, W. B. (1993). Phylogenetic relationships within the extant Mustelidae (Carnivora): appraisal of the cladistic status of the Simpsonian subfamilies. Zoological Journal of the Linnaean Society, 108, 301–334. Castresana, J. (2001). Cytochrome b phylogeny and the taxonomy of great apes and mammals. Molecular Biology and Evolution, 18, 465– 471. Davison, A., Griffiths, H. I., Brookes, R. C., Maran, T., Macdonald, D. W., Sidorovich, V. E., Kitchener, A. C., Irizar, I., Villate, I., GonzálezEsteban, J., Ceña, J. C., Ceña, A., Moya, I. & Palazón, S. (2000). © The Norwegian Academy of Science and Letters • Zoologica Scripta, 33, 6, November 2004, pp481– 499 493 Phylogeny, evolution and taxonomy of the Mustelidae • J. Marmi et al. Mitochondrial DNA and palaeontological evidence for the origins of endangered European mink, Mustela lutreola. Animal Conservation, 4, 345–355. Domingo-Roura, X. (2002). Genetic distinction of marten species by fixation of a microsatellite region. Journal of Mammalogy, 83, 907– 912. Dragoo, J. W. & Honeycutt, R. L. (1997). Systematics of mustelidlike carnivores. Journal of Mammalogy, 78, 426–443. Farris, J. S., Källersjö, M., Kluge, A. G. & Bult, C. (1995). Testing significance of incongruence. Cladistics, 10, 315–319. Felsenstein, J. (2002). PHYLIP (Phylogeny Inference Package), version 3.6a3. Seattle: Department of Genetics, University of Washington. Felsenstein, J. & Churchill, G. A. (1996). A hidden Markov model approach to variation among sites in rate of evolution. Molecular Biology and Evolution, 13, 93–104. Gagneux, P., Wills, C., Gerloff, U., Tautz, D., Morin, P. A., Boesch, C., Fruth, B., Hohmann, G., Ryder, O. A. & Woodruff, D. S. (1999). Mitochondrial sequences show diverse evolutionary histories of African hominoids. Proceedings of the National Academy of Sciences USA, 96, 5077–5082. Gatesy, J., Amato, G., Vrba, E., Schaller, G. & DeSalle, R. (1997). A cladistic analysis of mitochondrial ribosomal DNA from the Bovidae. Molecular Phylogenetics and Evolution, 7, 303–319. Ginsburg, L. & Morales, J. (2000). Origine et évolution des Melinae (Mustelidae, Carnivora, Mammalia). Compte Rendu de l’Académie des Sciences de Paris (2A), 330, 221–225. Grakov, N. N. (1994). Kidus — a hybrid of the sable and the pine marten. Lutreola, 3, 1– 4. Graur, D. & Li, W.-H. (2000). Fundamentals of Molecular Evolution, 2nd edn. Sunderland, MA: Sinauer Associates. Hall, T. A. (1999). Bioedit, version 5.0.9 ( http://www.mbio.ncsu.edu/ BioEdit/page2.html). Hasegawa, M., Kishino, H. & Yano, T. (1985). Dating of the human– ape splitting by molecular clock of mitochondrial DNA. Journal of Molecular Evolution, 22, 160–174. Heptner, V. G., Naumov, N. P., Yurgenson, P. B., Sludskiy, A. A., Chirkova, A. F. & Bannikov, A. G. (1967). Mammals of the Soviet Union, Vol. 2. Sea Cows and Carnivora. Moscow: Vyshaya Shkola [in Russian]. Hosoda, T., Suzuki, H., Harada, M., Tsuchiya, K., Han, S.-H., Zhang, Y., Kryukov, A. P. & Lin, L.-K. (2000). Evolutionary trends of the mitochondrial lineage differentiation in species of genera Martes and Mustela. Genes, Genetics and Systematics, 75, 259– 267. Huelsenbeck, J. P. & Ronquist, F. (2001). MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics Application Notes, 17, 754–755. Hunt, R. M. Jr (1974). The auditory bulla in Carnivora: an anatomical basis for reappraisal of carnivore evolution. Journal of Morphology, 143, 21–76. Irwin, D. M., Kocher, T. D. & Wilson, A. C. (1991). Evolution of the cytochrome b gene of mammals. Journal of Molecular Evolution, 32, 128–144. Kluge, A. G. (1989). A concern for evidence and a phylogenetic hypothesis of relationships among Epicrates (Boidae, Serpentes). Systematic Zoology, 38, 7–25. Koepfli, K. P. & Wayne, R. K. (1998). Phylogenetic relationships of otters (Carnivora: Mustelidae) based on mitochondrial cytochrome b sequences. Journal of Zoology, London, 246, 401–416. 494 Kumar, S., Tamura, K., Jakobsen, I. B. & Nei, M. (2001). MEGA2: molecular evolutionary genetics analysis software. Bioinformatics, 17, 1244–1245. Kurose, N., Kaneko, Y., Abramov, A. V., Siriaroonrat, B. & Masuda, R. (2001). Low genetic diversity in Japanese populations of the Eurasian badger Meles meles (Mustelidae, Carnivora) revealed by mitochondrial cytochrome b gene sequences. Zoological Science, 18, 1145–1151. Kurose, N., Masuda, R., Siriaroonrat, B. & Yoshida, M. C. (1999). Intraspecific variation of mitochondrial cytochrome b gene sequences of the Japanese marten Martes melampus and the sable Martes zibellina (Mustelidae, Carnivora, Mammalia) in Japan. Zoological Science, 16, 693–700. Ledge, C. & Arnason, U. (1996). Phylogenetic analyses of complete cytochrome b genes of the order Carnivora with particular emphasis on the Caniformia. Journal of Molecular Evolution, 42, 135–144. Ledoux, R. G. & Kenyon, A. J. (1975). Protides of the Mustelidae — II. Immunologic relatedness. Comparative Biochemistry and Physiology. Comparative Physiology, 51, 213–217. Macdonald, D. W. (2001). The New Encyclopaedia of Mammals. Oxford: Oxford University Press. Martin, L. D. (1989). Fossil history of the terrestrial Carnivora. In J. L. Gittleman (Ed.) Carnivore Behaviour, Ecology and Evolution (pp. 536–568). New York: Comstock Publishing. Martin, Y., Gerlach, G., Schlötterer, C. & Meyer, A. (2000). Molecular phylogeny of European muroid rodents based on complete cytochrome b sequences. Molecular Phylogenetics and Evolution, 16, 37–47. Masuda, R. & Yoshida, M. C. (1994). A molecular phylogeny of the family Mustelidae (Mammalia, Carnivora), based on comparison of mitochondrial cytochrome b nucleotide sequences. Zoological Science, 11, 605–612. Nowak, R. M. (1991). Walker’s Mammals of the World, 5th edn. Baltimore: Johns Hopkins University Press. Petter, G. (1971). Origine, phylogenie et systematique des blaireaux. Mammalia, 35, 567–597. Radinsky, L. B. (1968). Evolution of somatic sensory specialization in otter brains. Journal of Comparative Neurobiology, 134, 495–506. Rozas, J. & Rozas, R. (1999). DnaSP, version 3: an integrated program for molecular population genetics and molecular evolution analysis. Bioinformatics, 15, 174–175. Savage, D. E. & Russell, D. E. (1983). Mammalian Paleofaunas of the World. London: Addison-Wesley. Schneider, S., Roessli, D. & Excoffier, L. (2000). Arlequin, Version 2.000: A Software for Population Genetics Data Analysis. Geneva: Genetics and Biometry Laboratory, University of Geneva. Seberg, O., Humphries, C. J., Knapp, S., Stevenson, D. W., Petersen, G., Scharff, N. & Andersen, N. M. (2003). Shortcuts in systematics? A commentary on DNA-based taxonomy. Trends in Ecology and Evolution, 18, 63–65. Simpson, G. G. (1945). The principles of classification and a classification of mammals. Bulletin of the American Museum of Natural History, 85, 1–350. Stains, H. J. (1984). Carnivores. In S. Anderson & J. K. Jones, Jr (Eds) Orders and Families of Recent Mammals of the World (pp. 491– 522). New York: John Wiley and Sons. Stone, K. D. & Cook, J. A. (2002). Molecular evolution of Holarctic martens (genus Martes, Mammalia: Carnivora: Mustelidae). Molecular Phylogenetics and Evolution, 24, 169–179. Zoologica Scripta, 33, 6, November 2004, pp481– 499 • © The Norwegian Academy of Science and Letters J. Marmi et al. • Phylogeny, evolution and taxonomy of the Mustelidae Swofford, D. L. (2002). PAUP*, Phylogenetic Analysis Using Parsimony (*and Other Methods), version 4.0b10. Sunderland, MA: Sinauer Associates. Tamura, K. & Nei, M. (1993). Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Molecular Biology and Evolution, 10, 512–526. Ternovsky, D. V. (1977). Biology of Mustelidae. Novosibirsk: Nauka [in Russian]. Thompson, J. D., Higgins, D. G. & Gibson, T. J. (1994). CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting position-specific gap penalties and weight matrix choice. Nucleic Acids Research, 22, 4673–4680. Waits, L. P., Sullivan, J., O’Brien, S. J. & Ward, R. H. (1999). Rapid radiation events in the family Ursidae indicated by likelihood phylogenetic estimation from multiple fragments of mtDNA. Molecular Phylogenetics and Evolution, 13, 82–92. Willemsen, G. F. (1992). A revision of the Pliocene and Quaternary Lutrinae from Europe. Scripta Geologica, 101, 1–115. Wolsan, M. (1993). Phylogeny and classification of early European Mustelida (Mammalia: Carnivora). Acta Theriologica, 38, 345– 384. Wolsan, M. (1999). Oldest mephitine cranium and its implications for the origin of skunks. Acta Palaeontologica Polonica, 44, 223– 230. Wozencraft, W. C. (1989). The phylogeny of the recent Carnivora. In J. L. Gittleman (Ed.) Carnivore Behaviour, Ecology and Evolution (pp. 495–453). New York: Cornell University Press. Wozencraft, W. C. (1993). Carnivora. In D. E. Wilson & D. M. Reeder (Eds) Mammal Species of the World: a Taxonomic Geographic Reference (pp. 279–348). Washington: Smithsonian Institution Press. Wurster, D. H. & Benirschke, K. (1968). Comparative cytogenetics studies in the order Carnivora. Chromosoma (Berlin), 24, 336 –382. Wynen, L. P., Goldsworthy, S. D., Insley, S. J., Adams, M., Bickham, J. W., Francis, J., Gallo, J. P., Hoelzel, A. R., Majluf, P., White, R. W. G. & Slade, R. (2001). Phylogenetic relationships within the eared seals (Otariidae: Carnivora): implications for the historical biogeography of the family. Molecular Phylogenetics and Evolution, 21, 270 – 284. Zardoya, R., Vollmer, D. M., Craddock, C., Streelman, J. T., Karl, S. & Meyer, A. (1996). Evolutionary conservation of microsatellite flanking regions and their use in resolving the phylogeny of cichlid fishes (Pisces: Perciformes). Proceedings of the Royal Society of London B Biological Sciences, 263, 1589–1598. © The Norwegian Academy of Science and Letters • Zoologica Scripta, 33, 6, November 2004, pp481– 499 495 Phylogeny, evolution and taxonomy of the Mustelidae • J. Marmi et al. Appendix 1 List of samples used in cytochrome b analyses. Species MUSTELIDAE Mustelinae Martes martes, Mama zibellina, Mazi melampus, Mame americana, Maam foina, Mafo flavigula, Mafl pennanti, Mape Gulo gulo, Gugu Mustela eversmannii, Muev lutreola, Mulu putorius, Mupu nigripes, Mung sibirica, Musi itatsi, Muit altaica, Mual 496 Origin n Accession numbers Spain Germany Sweden Belarus Finland Russia Unknown Russia Japan Unknown Japan North America 1 4 2 1 1 2 4 5 8 2 20 25 Spain Germany Belarus China Unknown Thailand Russia China USA Canada Unknown 1 6 1 1 1 2 1 1 1 2 1 AJ536007* AB051251–52, AF154975, AF448239 AF448240 –41 AJ536008* AJ536009* AB051237, AB051253 L39275, L77958, AF336974–75 AB029420–21, AF448242–44 D26519, AB012356–61, AB029423 L39276, L77957 D26518, AB012341–55, AB029424–26, AB051238 L39270–74, L77952–53, AB051234, AF057130, AF154964–74, AF268272–74, AF448237–38 AJ536005* AB051247–50, AF448245–46 AJ536006* AB051236 AF336976 AB012362–63 AB051235 AB051246 AF057131 AF448247–48 L77959 Sweden Russia Unknown 1 1 1 X94921 AB051245 L77960 Serbia Russia Mongolia Eastern Europe Russia Spain UK Slovenia Germany Eastern Europe Russia Unknown USA Captive Russia Korea Japan Taiwan Japan Russia Mongolia 1 2 2 3 3 1 1 4 3 1 2 3 1 1 5 3 8 2 5 3 1 AF068540 AB026102, AB051261 AF068541–42 AF207712–14 AB026105, AB051263, AF068544 AF207716 AF068538 AF068534–37 AB051273–75 AF207715 AB026107, AB051276 U12845, X94925, AF057128 AF068543 AJ489325* AB029428, AB051277–80 AB051281–83 D26132, AB026108, AB051242, AB051284–88 AB051243, AB051289 D26130–31, AB026104, AB029427, AB051262 AB051239, AB051254–55 AB026100 Zoologica Scripta, 33, 6, November 2004, pp481– 499 • © The Norwegian Academy of Science and Letters J. Marmi et al. • Phylogeny, evolution and taxonomy of the Mustelidae Appendix 1 Continued n Species Origin nivalis, Muni Slovenia Germany Russia Korea Japan Taiwan UK Germany Russia Japan North America Spain UK USA/UK Japan Unknown Colombia 1 3 3 1 6 1 1 1 2 5 8 1 1 1 1 3 1 AF068545 AB051264–66 AB051267–69 AB051270 D26133, D26516, AB026106, AB051241, AB051271–72 AB046612 AF068546 AB051256 AB051257–58 D26515, AB026101, AB051240, AB051259–60 AF057127, AF271060–66 AJ536003* AJ536004* AF068548 D26517 L39278, AB026109, AF057129 AF068547 Spain UK Norway Belarus Russia Unknown Captive 1 1 1 1 1 1 1 AJ536012* AJ536010* AF057124 AJ536011* D26521 X94923 AF057125 Thailand Unknown 1 1 AJ536013* AF057119 South Africa 1 AF057118 USA North Pacific Unknown 2 1 2 AB051244, AF057120 D26522 U12835, X94924 Chile South America USA 1 1 1 AF057122 AF057123 AF057121 Unknown 1 AF057126 erminea, Muer vison, Muvi frenata, Mufr Lutrinae Lutra lutra, Lulu maculicollis, Luma Amblonyx cinereus, Amci Aonyx capensis, Aoca Enhydra lutris, Enlu Lontra felina, Lofe longicaudis, Lolo canadensis, Loca Pteronura brasiliensis, Ptbr Melinae Meles meles, Meme Arctonyx collaris, Arco Taxidiinae Taxidea taxus, Tata Accession numbers UK Sweden Russia Japan 1 1 3 18 Thailand 1 AB049810 USA Unknown 1 2 AF057132 L39277, L77961 © The Norwegian Academy of Science and Letters • Zoologica Scripta, 33, 6, November 2004, pp481– 499 AF068549 X94922 AB049807–09 D26520, AB049790–806 497 Phylogeny, evolution and taxonomy of the Mustelidae • J. Marmi et al. Appendix 1 Continued n Species Origin MEPHITIDAE Mephitis mephitis, Mpmp Spilogale putorius, Sppu USA 2 X94927, AJ536014* USA 2 X94928, AJ536015* Unknown 1 X94931 Unknown 1 X94930 Unknown 1 X94919 Unknown 1 X82292 PROCYONIDAE Bassaricyon gabbii Procyon lotor Ailurus fulgens OTARIIDAE Arctocephalus gazella Accession numbers n, number of sequences. Samples sequenced in our laboratory are marked with an asterisk. 498 Zoologica Scripta, 33, 6, November 2004, pp481– 499 • © The Norwegian Academy of Science and Letters J. Marmi et al. • Phylogeny, evolution and taxonomy of the Mustelidae Appendix 2 List of samples sequenced for the Mel08 complex repetitive flanking region. Origin n Accession number Supplier and institution Spain France Finland Belarus Japan Spain Belarus 1 1 1 1 1 1 1 AJ489568 AJ489567 AJ489569 AJ489570 S. Lavín, Universitat Autónoma de Barcelona X. Domingo-Roura, Universitat Pompeu Fabra K. Kauhala, Finnish Game and Fisheries Research Institute V. Sidorovich, National Academy of Sciences of Belarus Y. Fukue, Tokio University of Agriculture and Technology S. Lavín, Universitat Autónoma de Barcelona V. Sidorovich, National Academy of Sciences of Belarus Canada 1 AJ489571 M. A. Ramsay, University of Saskatchewan C. M. Pond, The Open University Belarus Spain UK Captive Belarus UK 1 1 1 1 1 2 AJ489560 AJ489562 V. Sidorovich, National Academy of Sciences of Belarus C. Rosell, Minuartia, Sant Celoni A. Grogan, WildCRU, University of Oxford A. Kitchener, National Museums of Scotland V. Sidorovich, National Academy of Sciences of Belarus R. A. Macdonald, University of Bristol UK Belarus 1 1 AJ489574 A. Bradshaw, University of Cardiff V. Sidorovich, National Academy of Sciences of Belarus Thailand 1 AJ489575 S. O’Brien, Laboratory of Genomic Diversity, National Cancer Institute USA 1 AJ489576 S. O’Brien, Laboratory of Genomic Diversity, National Cancer Institute USA 1 AJ489573 S. O’Brien, Laboratory of Genomic Diversity, National Cancer Institute Spain UK Austria Greece Japan 1 1 1 1 2 AJ309847 AJ489572 M. Miralles, Rectoria Vella, Sant Celoni D. W. Macdonald and C. Newman, WildCRU, University of Oxford J. Brabec, University of Innsbruck D. W. Macdonald and R. Woodroffe, WildCRU, University of Oxford Y. Fukue, Tokio University of Agriculture and Technology M. Saeki, WildCRU, University of Oxford MEPHITIDAE Mephitis mephitis Spilogale putorius USA 1 AJ489577 S. O’Brien, Laboratory of Genomic Diversity, National Cancer Institute USA 2 AJ489578–79 S. O’Brien, Laboratory of Genomic Diversity, National Cancer Institute PROCYONIDAE Procyon lotor Ailurus fulgens USA 1 AJ489580 J. F. López-Giráldez, Universitat Pompeu Fabra captive 1 AJ489583 J. Fernández, Parc Zoològic de Barcelona S.A. South Georgia 1 AJ489584 J. P. Arnould, Macquarie University C. M. Pond, The Open University Species MUSTELIDAE Mustelinae Martes martes melampus foina Gulo gulo Mustela lutreola putorius nigripes nivalis erminea Lutrinae Lutra lutra Amblonyx cinereus Enhydra lutris Lontra canadensis Melinae Meles meles OTARIIDAE Arctocephalus gazella AJ489561 AJ489566 AJ489564–65 n, number of samples. © The Norwegian Academy of Science and Letters • Zoologica Scripta, 33, 6, November 2004, pp481– 499 499