Species boundaries, geographic distribution and evolutionary history of the Western Palaearctic freshwater mussels Unio (Bivalvia: Unionidae)

Species boundaries, geographic distribution and evolutionary history of the Western Palaearctic... Abstract Delimiting species boundaries is a fundamental yet challenging task that has been especially problematic in morphologically conserved lineages. The use of genetic markers and phylogenetic approaches has proven of paramount importance to resolve these cases. Using the genus Unio as a case study, we have analysed two mitochondrial markers, cytochrome c oxidase subunit I (COI) and 16S rRNA (16S), phylogenetically and phylogeographically to explore species limits and evolutionary lineages within the group. We followed two different approaches to define species boundaries: the generalized mixed Yule coalescence model in a Bayesian framework and the Poisson tree process model. Phylogenetic relationships among species and divergence times were also assessed using relaxed molecular clock analyses. Taken together, we provide a sound systematic framework for Western Palaearctic Unio species, addressing phylogenetic relationships, and the taxonomy, biogeographic patterns and evolutionary history of this group in this region. The Unio clade from the studied area had a clear phylogenetic structure with four robust lineages that include the following species: (1) Unio durieui + Unio gibbus; (2) Unio tumidus; (3) Unio pictorum, Unio delphinus + Unio foucauldianus, Unio elongatulus, Unio mancus, Unio ravoisieri, Unio tigridis; and (4) Unio tumidiformis + (Unio bruguierianus + (Unio crassus + Unio ionicus)). Deeper analyses, including filling gaps in distributions, should be conducted to disentangle the U. crassus, U. pictorum and U. tigridis species complexes. INTRODUCTION Delimiting species boundaries is a fundamental yet challenging task. The challenge derives from species being a dynamic spatial and temporal continuum (e.g. de Queiroz, 2005, 2007; Wake, 2009). It also derives from the complexity and the variety of possible mechanisms from which new species originate (Turelli, Barton & Coyne, 2001). Much debate exists about these topics, which is reflected in the literature on epistemological/philosophical issues related to the nature of species [i.e. what is actually a species? (e.g. Coyne, Orr & Futuyma, 1988; Wilson, 1995; Baum, 1998; Wheeler & Meier, 2000; Lee & Wolsan, 2002; de Queiroz, 2007; Ereshefsky, 2014)] and on the methodological problems associated with defining species [i.e. what criteria should be used to define and recognize species? (e.g. Mayden, 1997; Sites & Marshall, 2004; de Queiroz, 2005, 2007; Carstens et al., 2013)]. Several factors may contribute to the difficulty of resolving species boundaries observed in nature. Time from speciation is a key factor for detecting new species as recently divergent lineages would go through polyphyletic and paraphyletic stages, the so-called ‘grey zone of speciation’, before all their traits or characters eventually appear as reciprocally monophyletic (de Queiroz, 2005; Roux et al., 2016). Secondary contact and hybridization between differentiated lineages is also an obvious confounding factor. Moreover, given that taxonomists traditionally rely on morphology (in a broad sense) to describe and classify diversity, delimiting species boundaries in morphologically conserved lineages is especially problematic. Discordant within-clade diversity and disparity (e.g. Rowe et al., 2011), morphological stasis among species (e.g. Van Bocxlaer & Hunt, 2013), rampant homoplasy and constraints (e.g. Wake, Wake & Specht, 2011), and the presence of very simplified or specialized morphologies (e.g. parasitic and miniaturized species; Hanken & Wake, 1993) may result in taxonomic uncertainty within groups. The use of genetic markers has proven to be of paramount importance to resolve these cases. Furthermore, there has also been recent renewed interest in using phylogenetic approaches for species delimitations, particularly with the development of DNA barcodes and new species delimitation algorithms based on genotype data (e.g. Choi, 2016). However, results from these single-locus or multilocus DNA species delimitation algorithms, although claimed to be ‘objective’, should not be directly translated into new taxonomies. Instead, they should be considered as a powerful tool to hypothesize potential historical lineages/species to be tested and validated with other sources of information. In this context, we selected as a case study the genus Unio, a Palaearctic freshwater mussel group distributed throughout North Africa (Morocco, Algeria, Tunisia), Europe and parts of Asia, with the exception of the South African U. caffer and the Ethiopian Unio abyssinicus. The fossil record of the group extends back to the Triassic (Moore, 1969). In a seminal work, Haas (1969) considered 42 taxonomic units within Unio that included 12 ‘fundamental Unio species’ comprising different ‘races’ or incipient species; all the descriptions were based on external shell morphology. However, based on preliminary mitochondrial DNA analyses and geographic distribution, the following Western Palaearctic species are currently accepted (Araujo, Gómez & Machordom, 2005; Araujo, Toledo & Machordom, 2009a; Araujo et al., 2009b; Khalloufi et al., 2011; Prié, Puillandre & Bouchet, 2012; Reis, Machordom & Araujo, 2013; Prié & Puillandre, 2014; Lopes-Lima et al., 2017): Unio delphinus Spengler, 1793 (Atlantic Iberian rivers); U. tumidiformis Castro, 1885 (South Atlantic Iberian rivers); Unio mancus Lamarck, 1819 (Atlantic and Mediterranean French rivers and Mediterranean Iberian rivers); U. gibbus Spengler, 1793 (North Africa and the Spanish Barbate River); Unio ravoisieri Deshayes, 1847 (Algeria, Tunisia and two North East Iberian basins); and Unio durieui Deshayes, 1847 (Algeria and Tunisia). Recently, Froufe et al. (2016a) divided the species U. delphinus into the Iberian U. delphinus and the Atlantic Moroccan species U. foucauldianus Pallary, 1936. Other Western Palaearctic species with a wide distribution that were previously cited by Haas (1969), such as Unio crassus Retzius, 1788, Unio pictorum Linnaeus, 1758, Unio tumidus Retzius, 1788, Unio elongatulus C. Pfeiffer, 1825, and Unio tigridis Bourguignat, 1852, still require validation and an update on their distribution (Araujo & de Jong, 2015; Lopes-Lima et al., 2017). Previous taxonomy based on shell characters does not reflect the evolutionary history of the genus, nor does it adequately account for current species. The results obtained in previous works indicate that morphological characters alone are insufficient to establish reliable taxonomic and phylogenetic classification of Unionoida. Furthermore, many authors have concluded that the range of variation of shell shape within Unio species is very wide (Ball, 1922; Haas, 1969; Eagar, 1978; Zieritz et al., 2010), suggesting it may be a survival advantage, provided that such plasticity has functional significance. Given these uncertainties and the poor conservation status of freshwater mussels (Vaughn and Hakenkamp, 2001; Strayer et al., 2004; Lopes-Lima et al., 2017), there is an urgent need for a phylogeographic and taxonomic study of the genus Unio to determine the current Western Palaearctic lineages, their distribution and implications for their conservation. Understanding phylogenetic diversity, to maximize current and future levels of biodiversity, is crucial for conservation strategies and prioritization of freshwater mussels that are among the most threatened freshwater taxa worldwide (IUCN Red List of Threatened Species; Lydeard et al., 2004). However, conservation efforts in West Palaearctic Unionidae are hindered by taxonomic uncertainties (Araujo & de Jong, 2015; Lopes-Lima et al., 2017), and many species have been assigned a Data Deficient status by the IUCN (Cuttelod, Seddon & Neubert, 2011). Here, we sampled individuals across the distribution ranges of the currently recognized Unio species and used coalescence methods to delimit species boundaries in Western Palaearctic lineages. We then validated these results with phylogenetic (well-supported phylogenetically independent lineages), phylogeographic (values of genetic divergence and their intraspecific and interspecific gaps) and geographic (biogeographically distinct distributions) evidence. The resulting delimited species were contrasted with the previous taxonomy based on shell characters. Taken together, we provide a sound systematic framework for Western Palaearctic Unio species, addressing the phylogenetic relationships within the group, and its taxonomy, biogeographic patterns and evolutionary history in the region. MATERIAL AND METHODS We sampled 518 specimens along the distribution range of most of the recognized Unio morphospecies (Supporting Information, Table S1). Sequences from GenBank were also used to complete the ingroup data set (Supporting Information, Table S1) and to test genus monophyly using representatives of Margaritiferidae [Margaritifera auricularia (Spengler, 1793)] and Unionidae [Anodonta anatina (Linnaeus, 1758), Alasminodonta marginata Say, 1818, Alasminodonta triangulata (Lea, 1858) and Pyganodon grandis (Say, 1829)]. We sequenced two mitochondrial markers, cytochrome c oxidase subunit I (COI) and 16S rRNA (16S), for the sampled specimens. Genomic DNA was extracted with the DNeasy kit (Qiagen) following manufacturer’s instructions, except for a longer Proteinase K homogenization (overnight) and the use of RNAse. Small tissue samples were taken from the mantle or foot prior to returning most live specimens back to their habitat. Voucher specimens from each locality were included in the Malacological Collection of the Museo Nacional de Ciencias Naturales (MNCN-CSIC) in Madrid. PCR amplifications were performed with 2 µL of a 1:20 dilution of gDNA, 1× buffer, 2 mM MgCl2, 0.2 mM of each dNTPs, 0.16 µM of each primer (Folmer et al., 1994 and Machordom et al., 2003 for COI and Lydeard, Mulvey & Davis, 1996 for 16S), 1.5 U of Taq polymerase (Biotools) and ddH2O for a final volume of 50 µL. The amplification cycles were as follows: 94 °C for 4 min and 40 cycles of 94 °C for 45 s, 45 °C (for COI amplification) or 50 °C (for 16S amplification) for 1 min, 72 °C for 1 min and a final extension at 72 °C for 10 min. After analysis on 0.8% agarose gels stained with SYBRsafe, the positively amplified fragments were purified and sequenced on an ABI 3730 XL sequencer (Life Technologies) using the Big Dye Terminator Kit (Life Technologies). Electropherograms were trimmed, edited and assembled in Sequencher (Gene Codes Corporation, Ann Arbor, MI, USA). Raw matrices were prepared for the two markers separately. The COI matrix did not need gaps to be aligned, and the 16S required the addition of short gaps, mainly for alignment with the outgroup. The maximum number of gaps was 10 at 6 different regions of the 16S sequences (for U. elongatulus). Unique haplotypes were selected using Collapse (Posada, 2004) for each gene separately; new matrices considering only those were then prepared. Finally, other matrices were prepared containing information for both genes. Delimiting species boundaries We followed two different approaches to define species boundaries in our data set. First, we used the generalized mixed Yule coalescence model in a Bayesian framework (bGMYC) to delimit the number of lineages/species (Pons et al., 2006; Reid & Carstens, 2012). Basically, this model identifies a threshold in branching rates in time-calibrated trees that distinguishes population-level branching processes (i.e. coalescence processes) from species-level cladogenesis (Yule model of diversification). Without an a priori assignment of individuals to putative species, the model returns the probability of groups of individuals to conform to natural species. This analysis was carried out with the bGMYC 1.0.2 package (Reid & Carstens, 2012) in R 3.3.2 (R core Team, 2016). To obtain the calibrated trees necessary to perform the bGMYC analysis, we ran a BEAST 2.4.5 analysis (Bouckaert et al., 2014), partitioning the haplotypic data set by genes. We used the ‘Beast Model test’ option to both estimate the substitution model for each partition during the analysis and calibrate independent relaxed molecular clocks for the two genes with substitution rates retrieved from the literature (see Araujo et al., 2017; Froufe et al., 2016b). We applied lognormal distributions for the COI (mean = 0.003; SD = 0.1) and the 16S (mean = 0.002; SD = 0.1) ucld. mean priors, to specify lineages substitution rates of 0.3 and 0.2% per Myr, respectively. The analysis was run for 100 × 106 generations, saving one in every 10000 trees and associated parameters. The posterior distributions of the topologies and all parameters were synthesized in a maximum clade credibility (MCC) tree in TreeAnnotator 2.4.5 (Drummond & Rambaut, 2007). To run the bGMYC analysis, we randomly selected 100 trees from the post-burnin posterior distribution of ultrametric trees. After checking for convergence of the parameters in a single tree, we ran the bGMYC analysis using the 100 topologies with the following uniform priors: Yule process rate change (0–5), coalescence process rate change (0–2) and threshold parameter (1–50). The latter defines the number of potential species. We let this number vary from 1 to 50, which is over three times the number of currently recognized morphospecies. The GMYC approach is especially sensitive to misspecification of calibrations and branch smoothing, which could bias the accuracy of the resulting number of species (Zhang et al., 2013; Tang et al., 2014). We, therefore, ran a second analysis using the Poisson tree process (PTP) model (Zhang et al., 2013) to compare the number of species inferred. Like the GMYC, the PTP model runs without a priori assigning individuals to potential species, and models the branching rates to distinguish between intra- and interspecific processes. Unlike the GMYC, however, the PTP models speciation in terms of number of substitutions in branches, not time, and it can be run with non-ultrametric phylogenetic trees. To account for potential differences in branch lengths that could be obtained through different phylogenetic inference methods, we ran two separate analyses with trees derived from maximum likelihood (ML) and Bayesian inference (BI). The ML tree was generated on the RAxML webserver (Stamatakis, Hoover & Rougemont, 2008) and the BI tree in the parallel version of MrBayes 3.2.5 (Huelsenbeck & Ronquist, 2001; Ronquist & Huelsenbeck, 2003). The Bayesian analysis was run for 100 × 106 generations, discarding the first 25% as burnin and allowing the program to estimate the substitution model during the Metropolis-coupled Markov chain Monte Carlo (MC3) search. For both ML and BI reconstructions, a sequence from the Chinese Unio douglasiae, the only Eastern Palaearctic species included in this study, was specified as the outgroup. The M-PTP model for the ML and BI trees was run in m-PTP v.0.2.1 (Zhang et al., 2013; Kapli et al., 2016) with a Markov chain Monte Carlo search for 100 × 106 generations, sampling every 10000 generations and discarding the first 10 × 106 as burnin. We ran 100 independent analyses for each tree to confirm that the analyses were properly converging. All bGMYC and M-PTP analyses were run on a haplotypic data set to avoid zero branch lengths in the trees derived from identical sequences, which could severely mislead boundaries delimitation. Species validation: phylogenetic, phylogeographic, morphological and geographic evidence The validity of previously defined species boundaries was tested against historical (phylogenetic), morphological and geographic evidence. To determine the phylogenetic relationships for the species, we conducted maximum parsimony (MP), ML and BI analyses, using PAUP* 4.0a147 (Swofford, 2002), PHYML v.3.0 (Guindon et al., 2010) and MrBayes v3.1.2 (Huelsenbeck & Ronquist, 2001; Ronquist & Huelsenbeck, 2003), respectively. MP analyses were carried out through heuristic searches, using the tree bisection and reconnection algorithm, including ten random stepwise additions. For the ML analyses, we estimated best-fit models in jModelTest (Posada, 2008). Supports for both MP and ML analyses were determined by bootstrapping (Felsenstein, 1985) 1000 pseudoreplicates. BI was performed through 10 × 106 generations in two parallel runs, with invgamma and nst = 6 options. A consensus tree was computed after eliminating the first 25% of sampled generations as burnin, evaluating node supports by the posterior probabilities. Tracer (Rambaut et al., 2014) was used to verify that the two runs were long enough to converge and reach stability. These phylogenetic analyses were run on the haplotypic data set with the full set of outgroups. Phylogenetic relationships among species and their divergence times were also assessed using relaxed molecular clock analyses in a Bayesian-coalescence approach with the software starBEAST2 in BEAST 2.4.5 (Drummond & Rambaut, 2007; Drummond et al., 2012), using the UncorrelatedLogNormal template. We ran different analyses grouping the individuals according to the results of the bGMYC and M-PTP analyses. To calibrate the relaxed molecular clocks, we used a lognormal distribution analysis with a mean substitution rate of 0.003 (0.3% substitutions/lineage/Myr) and 0.002 (0.2% substitutions/lineage/Myr) for COI and 16S, respectively, and a SD of 0.1 for both genes (see Araujo et al., 2016; Froufe et al., 2016b). All analyses were run for 100 × 106 generations, saving one in every 10000 trees and associated parameters. The posterior distributions of the topologies and all parameters generated during the analyses were synthesized in an MCC tree in TreeAnnotator 2.4.5 (Drummond & Rambaut, 2007). The population geographic structure for each species or lineage was assessed through haplotype networks in Haploviewer (http://www.cibiv.at/~greg/haploviewer), based on the compound haplotypes (COI + 16S) of all specimens analysed (i.e. using the non-collapsed matrices), their frequencies and sample locations. To analyse the data with Haploviewer, positions with ambiguous nucleotides (like ‘N’) were eliminated because the program treats them as differences, and in most cases, these Ns were added to GenBank sequences as they were shorter than those obtained here. The mean genetic distances between groups, previously determined by the phylogenetic analyses, were calculated with DnaSP (Rozas & Rozas, 1995; Librado & Rozas, 2009), on the COI data, since this gene is the most studied and referenced for the analysis of divergence among related species, and is commonly used as a barcode (Hebert et al., 2003; Prié & Puillandre, 2014). The studied specimens were photographed to illustrate each species variability of shell phenotypes. Pictures of U. tumidiformis, U. gibbus, U. ravoisieri and U. durieui were not included because their shell variation has been illustrated previously in Reis & Araujo (2009), Araujo et al. (2009a) and Khalloufi et al. (2011). RESULTS The sequence data obtained from COI [657 base pairs (bp)] and 16S (498 bp, after alignment) (GenBank accession numbers from KX399859 to KX400569, with the exception of KX400076, KX400078 and KX400319; Supporting Information, Table S1) were concatenated into a unique matrix after verifying that no major incongruences exist between the two genes. For the phylogenetic analyses, we used the concatenated matrix that included only unique haplotypes (144 for COI and 118 for 16S) and the specimens that had information for both genes (resulting in a matrix with 164 individuals and 1155 characters). For the phylogeographic analyses, all specimens in common were included (368). The models that best fit the data were 012232 + I + G + F for COI, TIM1 + I + G for 16S and 010012 + I + G + F for the concatenated matrix (under Bayesian information criterion). Species boundaries To delimit the number of species in the bGMYC analysis, we first defined a probability threshold above which the haplotypes would represent individuals from different species (Yule process). To avoid over-splitting or over-lumping samples, we adopted the posterior mean of the analysis (P = 0.5) as the threshold. Following this, 24 species were supported (Fig. 1 and see below). The M-PTP analyses, either using ML or Bayesian trees, resulted in a lower number of inferred species (ML approach: 19 species; Bayesian approach: 18 species) (Fig. 1). Figure 1. View largeDownload slide Bayesian tree reconstruction based on the two mitochondrial genes analysed. Values on the branches indicate Bayesian posterior probabilities, ML and MP bootstrap values. Results from the species delimitation analyses are also shown in this figure (green: M-PTP with a Bayesian tree; red: M-PTP with an ML tree; blue: bGMYC analysis). Names for the currently recognized morphospecies are also indicated in the phylogenetic tree. Figure 1. View largeDownload slide Bayesian tree reconstruction based on the two mitochondrial genes analysed. Values on the branches indicate Bayesian posterior probabilities, ML and MP bootstrap values. Results from the species delimitation analyses are also shown in this figure (green: M-PTP with a Bayesian tree; red: M-PTP with an ML tree; blue: bGMYC analysis). Names for the currently recognized morphospecies are also indicated in the phylogenetic tree. Eight of the fourteen currently recognized morphospecies analysed were consistently recovered in both the bGMYC and M-PTP analyses (U. delphinus, U. douglasiae, U. durieui, U. elongatulus, U. foucauldianus, U. mancus, U. tumidiformis and U. tumidus). Disagreement resulted in the remaining six species: (i) Unio bruguierianus and Unio ionicus: previously considered as subspecies of U. crassus by Haas (1969), they were consistently recovered as independent units in all analyses. (ii) Unio crassus: M-PTP analyses split this taxon in two, one including samples from Greece, Turkey and one haplotype from central Europe (sequence from GenBank), and the other grouping samples from Slovakia, France and Germany. The bGMYC analysis further divided the ‘crassus’ group in up to six distinct lineages. (iii) Unio gibbus: bGMYC and M-PTP analyses split this taxon into two, one including the samples from Morocco and Spain and the other including the samples from Tunisia. (iv) Unio pictorum: all analyses consistently split U. pictorum into two lineages, one including only the haplotypes from Thrichomida Lake and the Axios River in Greece and the other containing the remaining European samples and the Greek haplotypes from Volvi Lake. (v) Unio ravoisieri: the M-PTP analysis with the ML tree and the bGMYC model split this taxon into two groups with no distinctive geographic distribution. (vi) Unio tigridis: the bGMYC model identified the two haplotypes analysed (Israel and Turkey) as different species. Phylogenetic analyses, geographic distribution and morphological characterisation Unio was consistently recovered as a monophyletic group in all analyses (bootstrap values were 83% for ML, 90% for MP and 1.00 posterior probability for BI), with U. douglasiae as the sister group of all other species (Fig. 1). The Unio clade from the Western Palaearctic had a clear structure with four robust and well-supported lineages that included the following morphospecies (Fig. 1): (1) U. gibbus + U. durieui; (2) U. tumidus; (3) (U. foucauldianus + U. delphinus) + U. tigridis, U. ravoisieri, U. mancus, U. elongatulus, U. pictorum; and (4) U. tumidiformis + [(U. bruguierianus + (U. ionicus + U. crassus)]. Clades 3 and 4 formed a well-supported monophyletic group, but the relationships between the other clades [1, 2 and (3 + 4)] were not clearly resolved. The U. durieui + U. gibbus lineage This clade grouped the samples from the morphospecies U. durieui and U. gibbus as sister lineages (Fig. 1). The U. durieui lineage only included samples from Tunisia and consisted of five different haplotypes among the specimens from the three rivers sampled. Distinct haplotypes were present in each river, but the most frequent haplotype was shared among specimens from the Ziatine and El Maaden rivers (Supporting Information, Fig. S1). Morphologically, U. durieui specimens were elongated with an intermediate shape between U. mancus and U. pictorum, and probably impossible to distinguish in the absence of geographical information. There were also rounded morphotypes similar to U. gibbus, as in the case of El Maaden (Khalloufi et al., 2011: fig. 5). This species was not found in Morocco but is present in Tunisia and Algeria to the east of the Moulouya River (Fig. 2). Figure 2. View largeDownload slide Schematic distribution of the Unio species in the Western Palaearctic. Points indicate the general vicinity of sampled localities. See Supporting Information, Table S1 for details. Figure 2. View largeDownload slide Schematic distribution of the Unio species in the Western Palaearctic. Points indicate the general vicinity of sampled localities. See Supporting Information, Table S1 for details. For the morphospecies U. gibbus, we obtained two clearly separated groups (genetic distance of 2.03%, Table 1): one representing haplotypes from the Iberian Peninsula (Barbate River) and Morocco, and another consisting of Tunisian haplotypes (Fig. 2; Supporting Information, Fig. S2). These two subclades were consistently recovered as distinct species in the bGMYC and M-PTP analyses. From 16 to 22 substitutions separated the Spanish-Moroccan haplotypes from the Tunisian ones. Although the eastern haplotypes were more similar, none of the Spanish, Moroccan and Tunisian specimens shared haplotypes among the 10 detected. Table 1. COI genetic distances between different lineages   1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  1. Outgroup  15.20                                                  2. U. douglasiae  15.30  –                                                3. U. crassus 1  15.44  14.56  0.20                                              4. U. crassus 2  15.42  13.85  2.13  0                                            5. U. crassus 3  15.12  13.85  3.10  2.34  0.09                                          6. U. crassus 4  15.50  14.46  0.96  2.18  2.85  0.27                                        7. U. crassus 5  15.85  14.92  2.60  3.28  4.10  2.22  0.51                                      8. U. crassus 6  15.65  14.80  2.69  2.60  3.64  2.33  2.89  0                                    9. U. ionicus  15.15  14.31  4.24  4.28  4.30  3.99  4.07  3.93  0.04                                  10. U. bruguierianus  15.9  14.12  3.87  4.22  4.21  3.82  4.45  3.86  4.28  0.08                                11. U. mancus  16.22  14.18  10.33  9.72  9.82  10.10  10.16  10.54  9.44  9.47  0.49                              12. U. pictorum 1  17.19  14.45  11.17  10.52  10.59  10.68  10.15  10.46  10.49  10.16  3.85  0.38                            13. U. pictorum 2  16.88  14.61  11.26  10.35  10.77  11.01  10.12  10.57  10.63  10.62  4.50  2.38  1.42                          14. U. elongatulus  15.94  15.16  9.80  8.98  9.21  9.30  9.01  9.35  9.43  9.25  3.89  3.26  4.46  0.28                        15. U. delphinus  13.88  12.90  11.32  9.80  9.70  10.81  9.75  9.79  9.04  8.53  3.81  3.66  4.49  4.29  0.50                      16. U. foucauldianus  15.92  14.30  10.84  10.21  9.91  10.38  9.75  9.63  9.90  9.87  4.10  3.94  4.52  3.71  3.32  0.20                    17. U. ravoisieri 1  16.26  14.79  10.65  9.73  10.45  10.50  9.50  9.86  10.02  10.30  5.01  4.36  4.75  4.46  4.72  4.30  0.20                  18. U. ravoisieri 2  16.25  14.70  10.40  9.79  10.03  10.24  8.97  9.28  10.10  10.07  5.82  4.38  4.96  4.81  4.21  5.20  2.55  0.26                19. U. tigridis 1  16.8  14.76  11.11  10.81  11.20  10.96  10.72  10.73  10.33  11.07  5.24  4.86  6.09  5.35  4.79  5.98  6.84  7.56  xx              20. U. tigridis 2  16.65  14.76  10.25  9.59  9.98  10.05  10.05  10.08  9.72  10.46  4.86  4.55  5.43  4.99  5.36  5.22  5.77  6.19  2.89  0            21. U. tumidiformis  15.74  13.26  8.31  7.99  7.85  8.86  8.51  7.95  9.30  8.22  11.54  11.63  11.90  11.95  10.72  10.98  11.18  10.92  12.83  12.86  0.50          22. U. tumidus  13.91  12.18  12.16  11.87  11.58  12.10  12.41  11.52  11.87  11.15  10.80  11.14  12.02  10.80  9.50  10.76  11.46  11.95  12.04  12.04  12.77  0.18        23. U. gibbus 1  14.24  11.13  11.89  11.65  11.25  11.82  11.80  11.40  11.13  10.93  12.35  12.86  13.33  13.22  12.12  13.12  13.77  14.14  13.69  13.90  13.02  10.19  0.37      24. U. gibbus 2  14.4  11.24  11.60  11.24  10.85  11.24  11.11  10.87  10.63  10.59  12.51  13.28  13.73  13.26  11.71  12.90  13.65  14.03  13.83  13.79  12.48  10.51  2.03  0.46    25. U. durieui  14.98  11.70  11.60  11.45  11.06  11.25  11.06  10.92  10.99  10.65  12.69  13.74  14.13  13.47  11.70  12.79  13.84  13.91  14.74  14.58  13.29  9.99  5.46  4.87  0.07    1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  1. Outgroup  15.20                                                  2. U. douglasiae  15.30  –                                                3. U. crassus 1  15.44  14.56  0.20                                              4. U. crassus 2  15.42  13.85  2.13  0                                            5. U. crassus 3  15.12  13.85  3.10  2.34  0.09                                          6. U. crassus 4  15.50  14.46  0.96  2.18  2.85  0.27                                        7. U. crassus 5  15.85  14.92  2.60  3.28  4.10  2.22  0.51                                      8. U. crassus 6  15.65  14.80  2.69  2.60  3.64  2.33  2.89  0                                    9. U. ionicus  15.15  14.31  4.24  4.28  4.30  3.99  4.07  3.93  0.04                                  10. U. bruguierianus  15.9  14.12  3.87  4.22  4.21  3.82  4.45  3.86  4.28  0.08                                11. U. mancus  16.22  14.18  10.33  9.72  9.82  10.10  10.16  10.54  9.44  9.47  0.49                              12. U. pictorum 1  17.19  14.45  11.17  10.52  10.59  10.68  10.15  10.46  10.49  10.16  3.85  0.38                            13. U. pictorum 2  16.88  14.61  11.26  10.35  10.77  11.01  10.12  10.57  10.63  10.62  4.50  2.38  1.42                          14. U. elongatulus  15.94  15.16  9.80  8.98  9.21  9.30  9.01  9.35  9.43  9.25  3.89  3.26  4.46  0.28                        15. U. delphinus  13.88  12.90  11.32  9.80  9.70  10.81  9.75  9.79  9.04  8.53  3.81  3.66  4.49  4.29  0.50                      16. U. foucauldianus  15.92  14.30  10.84  10.21  9.91  10.38  9.75  9.63  9.90  9.87  4.10  3.94  4.52  3.71  3.32  0.20                    17. U. ravoisieri 1  16.26  14.79  10.65  9.73  10.45  10.50  9.50  9.86  10.02  10.30  5.01  4.36  4.75  4.46  4.72  4.30  0.20                  18. U. ravoisieri 2  16.25  14.70  10.40  9.79  10.03  10.24  8.97  9.28  10.10  10.07  5.82  4.38  4.96  4.81  4.21  5.20  2.55  0.26                19. U. tigridis 1  16.8  14.76  11.11  10.81  11.20  10.96  10.72  10.73  10.33  11.07  5.24  4.86  6.09  5.35  4.79  5.98  6.84  7.56  xx              20. U. tigridis 2  16.65  14.76  10.25  9.59  9.98  10.05  10.05  10.08  9.72  10.46  4.86  4.55  5.43  4.99  5.36  5.22  5.77  6.19  2.89  0            21. U. tumidiformis  15.74  13.26  8.31  7.99  7.85  8.86  8.51  7.95  9.30  8.22  11.54  11.63  11.90  11.95  10.72  10.98  11.18  10.92  12.83  12.86  0.50          22. U. tumidus  13.91  12.18  12.16  11.87  11.58  12.10  12.41  11.52  11.87  11.15  10.80  11.14  12.02  10.80  9.50  10.76  11.46  11.95  12.04  12.04  12.77  0.18        23. U. gibbus 1  14.24  11.13  11.89  11.65  11.25  11.82  11.80  11.40  11.13  10.93  12.35  12.86  13.33  13.22  12.12  13.12  13.77  14.14  13.69  13.90  13.02  10.19  0.37      24. U. gibbus 2  14.4  11.24  11.60  11.24  10.85  11.24  11.11  10.87  10.63  10.59  12.51  13.28  13.73  13.26  11.71  12.90  13.65  14.03  13.83  13.79  12.48  10.51  2.03  0.46    25. U. durieui  14.98  11.70  11.60  11.45  11.06  11.25  11.06  10.92  10.99  10.65  12.69  13.74  14.13  13.47  11.70  12.79  13.84  13.91  14.74  14.58  13.29  9.99  5.46  4.87  0.07  The intraspecific divergences within species divided into different units were Unio crassus: 2.10, Unio pictorum: 0.72, Unio ravoisieri: 1.36, Unio tigridis: 0.96 and Unio gibbus: 1.18%. View Large This species had a rounded shape similar to Potomida, but clear diagnostic characters such as the lateral teeth, the umbonal sculpture and shell thickness differentiate the two taxa (Araujo et al., 2009a: fig. 2). The U. tumidus lineage The U. tumidus lineage included samples from European rivers (namely, the Seine, Danube, Rhine, Thames and Weser), a Swedish lake, the Izorka and Kovash rivers in Russia and the Dnieper River in Ukraine (Fig. 2). This species is absent from the Iberian Peninsula and North Africa. Samples from the Rhine and Thames rivers and Bjornsjon Lake (Sweden) shared the same haplotype (Supporting Information, Fig. S3). The species had an elongated shell with an anterior umbo, but a characteristic cuneiform posterior end with a descending dorsal posterior margin, which clearly distinguishes this taxon from the other species. Nevertheless, there were specimens with short shells that were very difficult to identify (Fig. 3). We did not have access to the Ukrainian and Russian shells. Figure 3. View largeDownload slide Differing shell shapes of Unio tumidus. A, Franconian Saale, a tributary of the Main River (Rhine), Germany. B, Fulda River (Weser), Germany. C, Okna River (Danube), Slovakia. D, Danube River, Slovakia. E, Ferma Lake (Rhine), Germany. F, Thames River, UK. G, Fulda River (Weser), Germany. H, Rhine River, Germany. I, Horloff River (Rhine), Germany. Scale bar 2 cm. Figure 3. View largeDownload slide Differing shell shapes of Unio tumidus. A, Franconian Saale, a tributary of the Main River (Rhine), Germany. B, Fulda River (Weser), Germany. C, Okna River (Danube), Slovakia. D, Danube River, Slovakia. E, Ferma Lake (Rhine), Germany. F, Thames River, UK. G, Fulda River (Weser), Germany. H, Rhine River, Germany. I, Horloff River (Rhine), Germany. Scale bar 2 cm. The (U. foucauldianus + U. delphinus) + U. tigridis, U. ravoisieri, U. mancus, U. elongatulus, U. pictorum lineage This lineage was the most complex within the genus (Fig. 1). All morphospecies were recovered as monophyletic groups, although some had high levels of within-clade diversity and phylogenetic structuring. The relationships among lineages, however, were not well supported. The first well-supported clade of this lineage included two closely related morphospecies with a genetic distance of 3.32% (Fig. 1; Table 1): U. delphinus from Atlantic Iberia and U. foucauldianus from Atlantic and Mediterranean Morocco (Fig. 2). The network analysis showed a minimum of 24 steps separating the two species (Supporting Information, Fig. S4). There was greater Atlantic/South differentiation in U. delphinus (no shared haplotypes between the 2 regions and up to 24 substitutions among the most differentiated haplotypes) than in U. foucauldianus (in this case, the main haplotype was found in both the Atlantic and Mediterranean basins). Many of the U. delphinus and all of the U. foucauldianus specimens from Mediterranean Morocco had an elongated shell with a pictorum shape, while specimens from the Atlantic rivers of Morocco and southern Spain were more rounded, resembling the mancus shape (Figs 4, 5). Figure 4. View largeDownload slide Differing shell shapes of Unio foucauldianus. A, Loukos River. B, Oum Er Rbia River. C, Molouya River. D, Mda River. E, Martil River. F, Beth River (Sebou). G, Loukos River. Scale bar 2 cm. Figure 4. View largeDownload slide Differing shell shapes of Unio foucauldianus. A, Loukos River. B, Oum Er Rbia River. C, Molouya River. D, Mda River. E, Martil River. F, Beth River (Sebou). G, Loukos River. Scale bar 2 cm. Figure 5. View largeDownload slide Differing shell shapes of Unio delphinus. A, Guadalmez River (Guadiana). B, Landrinos River (Tagus). C, Ulla River. D, Hozgarganta River. E, Barbate River. F, Deza River (Ulla). G, Guadalporcún River (Guadalete). Scale bar 2 cm. Figure 5. View largeDownload slide Differing shell shapes of Unio delphinus. A, Guadalmez River (Guadiana). B, Landrinos River (Tagus). C, Ulla River. D, Hozgarganta River. E, Barbate River. F, Deza River (Ulla). G, Guadalporcún River (Guadalete). Scale bar 2 cm. The second clade comprised U. tigridis samples from Israel (Lake Kinneret in the Jordan River basin) and Turkey (the Tersakan stream) (Fig. 2). The M-PTP analysis considered these samples as two different species (Fig. 1), with a genetic distance of 2.89% (Table 1). Only six specimens were analysed from this area, and the two haplotypes detected were separated by 32 substitutions (Supporting Information, Fig. S5). This species is typified by a short shell with anteriorly shifted umbos, which was observed in the Israeli specimens. The Turkish specimens resembled the U. mancus morph (Fig. 6). Figure 6. View largeDownload slide Differing shell shapes of Unio tigridis. A, Lake Kinneret, Israel. B, Tersakan River, Southwest Turkey. Figure 6. View largeDownload slide Differing shell shapes of Unio tigridis. A, Lake Kinneret, Israel. B, Tersakan River, Southwest Turkey. The next lineage combined the U. ravoisieri specimens into two subclades with a genetic distance of 2.55% (Table 1; Fig. 2). These subclades were also consistently recovered as distinct species in the bGMYC and M-PTP analyses. The two most frequent haplotypes (Supporting Information, Fig. S6) were separated by 18 steps. One of these haplotypes represented the Spanish samples, and was very close to a haplotype present in the Tunisian Kebir River (separated by only one step), and to one of the two divergent El Maaden haplotypes (separated by four steps). The shell of this species typically has a modified U. mancus shape, but the morphs we obtained were highly variable, especially those from the Spanish Ser River (Khalloufi et al., 2011: fig. 3). The clade consisting of U. mancus (Fig. 1) included specimens from Mediterranean basins in Spain (the Ebro, Fluviá, Ter, Sonella, Llobregat and Júcar rivers), France (the Hérault and Argens rivers, the Tet River in the Basse Basin, Lake Bourget and the Ognon River, both within the Rhône Basin, many rivers in Corsica and the Atlantic rivers Charente, Seine and Drée, a tributary of the Loire) and Italy (many rivers in Sardinia) (Fig. 2). The network analysis showed haplotypes with some biogeographic incongruences (Supporting Information, Fig. S7). For instance, the Júcar and Sonella rivers (eastern Spain) shared a haplotype with some French populations, but not with any of the northeastern Spanish populations. This high-frequency haplotype appeared in both Atlantic and Mediterranean French and Spanish rivers. Of the 15 different haplotypes found, five were exclusive to Sardinian-Corsican populations, one to the French Lake Bourget and one appeared with a high frequency in the Catalonian (eastern Spain) rivers. This species had a more rounded shell than the other Mediterranean species U. elongatulus, with the exception of specimens from northeastern Spain, Corsica and Sardinia, which were elongated (Figs 7, 8). Figure 7. View largeDownload slide Differing shell shapes of Unio mancus. A, Stabiacciu River, Corsica. B, Liscia River, Sardinia. C, Cedrino River, Corsica. D, River at Banyoles Lake, Spain. E, F, Araxisi River, Sardinia. Scale bar 2 cm. Figure 7. View largeDownload slide Differing shell shapes of Unio mancus. A, Stabiacciu River, Corsica. B, Liscia River, Sardinia. C, Cedrino River, Corsica. D, River at Banyoles Lake, Spain. E, F, Araxisi River, Sardinia. Scale bar 2 cm. Figure 8. View largeDownload slide Differing shell shapes of Unio mancus. A, Brugent River (Ter), Spain. B, Bourget Lake (Rhône), France. C, Drée River (Loire), France. D, Golo River, Corsica. E, Ebro River, Spain. F, Stabiacciu River, Corsica. G, Orbu River, Corsica. Scale bar 2 cm. Figure 8. View largeDownload slide Differing shell shapes of Unio mancus. A, Brugent River (Ter), Spain. B, Bourget Lake (Rhône), France. C, Drée River (Loire), France. D, Golo River, Corsica. E, Ebro River, Spain. F, Stabiacciu River, Corsica. G, Orbu River, Corsica. Scale bar 2 cm. The next clade included the U. elongatulus (Fig. 1) specimens from Mediterranean freshwaters in Italy (Po and Isonzo rivers and Alpine lakes), Croatia (Mirna and Zrmanja rivers, Bačinska Lake) and Albania (Scutari Lake) (Fig. 2). The Croatian haplotypes appeared at both ends of the network (Supporting Information, Fig. S8), therefore, shared haplotypes between Croatia and Italy would be expected. However, no haplotypes were shared among any of the specimens from these regions. The specimens had more elongated shells than other Mediterranean species (U. mancus) and a posterior umbo, although some populations had more rounded shells (Fig. 9). Figure 9. View largeDownload slide Differing shell shapes of Unio elongatulus. A, Po di Tolle River, Italy. B, Lake Candia, Italy. C, Venice, Italy. D, Lake Cestella, Italy. E, Lake Bačinska, Croatia. F, G, Mirna River, Croatia. H, Zrmanja River, Croatia. I, Lake Scutari, Albania. Scale bar 2 cm. Figure 9. View largeDownload slide Differing shell shapes of Unio elongatulus. A, Po di Tolle River, Italy. B, Lake Candia, Italy. C, Venice, Italy. D, Lake Cestella, Italy. E, Lake Bačinska, Croatia. F, G, Mirna River, Croatia. H, Zrmanja River, Croatia. I, Lake Scutari, Albania. Scale bar 2 cm. The last clade represented the species U. pictorum, which included two groups having a genetic distance of 2.38% (Table 1). The first group (1) consisted of specimens from Western and Central Europe (Lake Bourget in the Rhône Basin and the Adour, Seine, Oder, Danube and Thames basins), eastern Greece (Strymonas and Axios rivers and Lake Volvi), Ukraine (Teteriv at the Black Sea), Iran and Russia (Dzhankhot River at the Black Sea, Vyborg River in Karelia). The second group (2) comprised specimens from both western (Trichonida Lake at the Acheloos basin) and eastern (Axios River) Greece (Fig. 2). These two subclades were consistently recovered as distinct species in the bGMYC and M-PTP analyses. Up to 16 different haplotypes (Supporting Information, Fig. S9), with from one to a maximum of 23 substitutions, were found among the specimens of this species. There was a maximum of 8 steps of differentiation within group ‘1’ and 17 between two of the three ‘2’ haplotypes. The network shape of group ‘1’ represented polymorphic populations with two relatively frequent haplotypes, one of which was only found in specimens from the Thames River and the other shared by samples as far away as Iran and Greece. Haplotypes found in the Danube and Dzhankhot rivers and in Lake Volvi seemed to derive from the latter one. This highly polymorphic species had a typical elongated shell. However, some specimens were very different, for example those from Strymonas River, Lake Volvi and the Axios River (within both clades). Some populations resembled U. mancus (Lake Volvi, Thichonida and Strymonas rivers) (Figs 10, 11). Figure 10. View largeDownload slide Differing shell shapes of Unio pictorum. A, B, Lake Trichonida, Greece. C, Thames River, UK. D, Axios River, Greece. E, Okna River (Danube), Slovakia. F, Axios River, Greece. Scale bar 2 cm. Figure 10. View largeDownload slide Differing shell shapes of Unio pictorum. A, B, Lake Trichonida, Greece. C, Thames River, UK. D, Axios River, Greece. E, Okna River (Danube), Slovakia. F, Axios River, Greece. Scale bar 2 cm. Figure 11. View largeDownload slide Differing shell shapes of Unio pictorum. A, B, Lake Volvi, Greece. C–F, Strymonas River, Greece. Scale bar 2 cm. Figure 11. View largeDownload slide Differing shell shapes of Unio pictorum. A, B, Lake Volvi, Greece. C–F, Strymonas River, Greece. Scale bar 2 cm. The U. tumidiformis + [U. bruguierianus + (U. ionicus + U. crassus)] lineage This lineage included two monophyletic groups with a COI genetic distance greater than 7.85%: a western group that included the endemic Iberian species U. tumidiformis (Fig. 2) and a second group (supported by a posterior probability of 0.99 and bootstrap values of 79 for ML and 94 for MP) comprising three morphospecies with a complex geographic structure (Fig. 1; Table 1). Up to nine haplotypes were found among the samples analysed for U. tumidiformis (Supporting Information, Fig. S10). Two were exclusive to the Portuguese Sado River but were close to some Spanish haplotypes. The most frequent haplotype was found in the Mira and Guadiana rivers. Within the second group, U. bruguierianus from eastern Greece (the Axios and Pinios rivers) (Fig. 2) was the sister species of U. ionicus + U. crassus (genetic distance of 4.28–4.45%), although this relationship was not highly supported. Unio ionicus lives in Albania and the Acheloos River (including Lake Lysimacheia) in western Greece (Fig. 2) and was the sister group of U. crassus (genetic distance of 4.30%). Unio crassus presented five subclades with a minimum divergence of 2.16% (Table 1) and haplotypes separated by at least 19 steps (Fig. 12). These five subclades included (1) U. crassus courtillierii Hattemann, 1859 from France and one sample from Sweden (both sequences obtained from GenBank); (2) samples from the Sofaditikos (Pinios catchment), Aliakmon/Aliakmonas and Sperchios rivers in eastern Greece; (3) Central European samples from the Rhine, Danube, and Rhône catchments; (4) samples from the Lissos River (eastern Greece); and (5) samples from western Turkey (Fig. 2). This group had the most complexity, which is also reflected in the bGMYC and M-PTP analyses in which up to six species were revealed. Figure 12. View largeDownload slide Network of Unio crassus haplotypes. A, U. crassus courtillierii and Sweden. B, eastern Greece (Sofaditikos, Aliakmon and Sperchios). C, Central European (Rhine, Danube and Rhône). D, eastern Greece (Lissos River). E, western Turkey. Figure 12. View largeDownload slide Network of Unio crassus haplotypes. A, U. crassus courtillierii and Sweden. B, eastern Greece (Sofaditikos, Aliakmon and Sperchios). C, Central European (Rhine, Danube and Rhône). D, eastern Greece (Lissos River). E, western Turkey. All U. crassus haplotypes were represented in a unique network (Fig. 12). No clear biogeographic structure was observed, and some haplotypes belonging to specimens from the same rivers had a high number of substitutions (up to 33). From the right end of Figure 12, three haplotypes from eastern Greece were assigned to U. bruguierianus. Three distinct haplotypes from western Greece and Albania were considered to represent U. ionicus. Three haplotypes were found from the Turkish locations analysed, but with a smaller genetic distance compared with the other haplotypes. A polymorphic shell shape was observed for U. bruguierianus: some specimens from the Axios and Pinios rivers had a shell shape resembling some pictorum specimens from the Axios River, while other specimens from the Pinios River were identical to some U. crassus specimens from the Lissos River or Turkey (Fig. 13). This polymorphic shell shape was also observed for U. ionicus and U. crassus (Figs 13, 14). Unio ionicus had an almost identical shell to some U. crassus specimens from the Sofaditikos (Pinios Basin), Sperchios and Lissos rivers in Greece. In contrast, other U. crassus specimens from the Lissos River and from Turkey presented an elongated shell shape. There were also specimens with and without sculpture in the umbo area from the same locality with the same haplotype, such as those from the Lissos River (Fig. 14). Figure 13. View largeDownload slide A–C, Unio bruguierianus. A, Pinios River, Greece. B, Axios River, Greece. C, Pinios River, Greece. D–E, Unio ionicus. D, River at Lake Lysimacheia, Greece. E, Perroi i Bistrices, Albania. F–H, Unio crassus. F, Sperchios River, Greece. G, H, Aliakmonas River, Greece. Scale bar 2 cm. Figure 13. View largeDownload slide A–C, Unio bruguierianus. A, Pinios River, Greece. B, Axios River, Greece. C, Pinios River, Greece. D–E, Unio ionicus. D, River at Lake Lysimacheia, Greece. E, Perroi i Bistrices, Albania. F–H, Unio crassus. F, Sperchios River, Greece. G, H, Aliakmonas River, Greece. Scale bar 2 cm. Figure 14. View largeDownload slide Differing shell shapes of Unio crassus. A, B, Sofaditikos River (Pinios), Greece. C, Matzenheim, France. D, Çine Çayi, Mugla, Turkey. E, F, Lissos River, Greece. G, Limagne, France. Scale bar 2 cm. Figure 14. View largeDownload slide Differing shell shapes of Unio crassus. A, B, Sofaditikos River (Pinios), Greece. C, Matzenheim, France. D, Çine Çayi, Mugla, Turkey. E, F, Lissos River, Greece. G, Limagne, France. Scale bar 2 cm. The coalescence-based molecular clock analyses for the species tree resulted in very broad temporal intervals for most of the clades. Although the results have to be interpreted with caution, they provide a broad temporal framework for the diversification of the group. The coalescence reconstruction (Fig. 15) placed the common ancestor of the Unio clade and the origin of the Western Palaearctic Unio in the Eocene. In fact, the Western Palaearctic species underwent two main cladogenetic events. The first event occurred in the Early Eocene (U. gibbus, U. durieui split), and the second during the Oligocene, involving the divergence of the most speciose clade. Most of the modern species appeared during the Miocene. The most recent cladogenetic event involved the U. delphinus and U. foucauldianus clade at the end of the Miocene (Messinian). Notably, the reconstructed species tree agrees with the general topology of the concatenated matrix analyses except in the phylogenetic position of U. tumidus. In the species tree, U. tumidus appeared with strong support as sister group of the U. tigridis + [U. pictorum, U. ravoisieri (U. mancus + U. elongatulus, U. delphinus + U. foucauldianus)] lineage (Fig. 15). Figure 15. View largeDownload slide Coalescence-based species tree generated in BEAST. The x-axis scale is in millions of years. Bars indicate 95% high probability density intervals. Asterisks (*) in the tree indicate posterior probabilities pp > 0.9. Figure 15. View largeDownload slide Coalescence-based species tree generated in BEAST. The x-axis scale is in millions of years. Bars indicate 95% high probability density intervals. Asterisks (*) in the tree indicate posterior probabilities pp > 0.9. DISCUSSION In his classic study, Haas (1969) divided the Western Palaearctic lineages of Unio into 42 different taxa (Table 2), including 12 ‘fundamental species’ with many subspecies or races. In recent years, molecular markers have been instrumental in unravelling both the taxonomy and the phylogeny of this group (Araujo et al., 2005; Araujo et al., 2009a; Khalloufi et al., 2011; Prié et al., 2012; Reis et al., 2013; Prié & Puillandre, 2014; Froufe et al., 2016a). To further advance our understanding of species limits within this group and their evolutionary history and relationships, we have analysed the patterns of genetic variability in two mitochondrial genes in 518 samples of Unio from North Africa (Morocco, Tunisia), Europe (from Portugal to Greece) and the Middle East (Turkey and Israel). Table 2. Information on the Western Palaearctic Unio taxonomy Haas (1969)   Author  Accepted species  References  Unio tumidus tumidus  Retzius, 1788  Unio tumidus  This paper  Unio tumidus borysthenensis  Kobelt, 1879  Unio tumidus  This paper  Unio pictorum pictorum  Linnaeus, 1758  Unio pictorum  This paper  Unio pictorum praeposterus  Küster, 1854  Unio pictorum  This paper  Unio pictorum latirostris  Küster, 1854  Unio pictorum  This paper  Unio pictorum platyrhynchus  Rossmässler, 1835  Unio pictorum  This paper  Unio pictorum schrenkianus  Clessin, 1880  Unio pictorum  This paper  Unio pictorum ascanius  Kobelt, 1913  Unio pictorum  This paper  Unio pictorum proechistus  Bourguignat, 1870  Unio pictorum  This paper  Unio pictorum gaudioni  Drouët, 1881  Unio pictorum  This paper  Unio pictorum rostratus  Lamarck, 1819  Unio pictorum  This paper  Unio pictorum platyrhynchoideus  Dupuy, 1849  Unio pictorum  This paper  Unio pictorum mucidus  Morelet, 1845  Unio delphinus  Araujo et al. (2009); Reis et al. (2013); this paper  Unio pictorum delphinus  Spengler, 1793  Unio delphinus  Araujo et al. (2009); Reis et al. (2013); this paper  Unio pictorum ravoisieri  Deshayes, 1848  Unio ravoisieri  Araujo et al. (2009); Khalloufi et al. (2011); this paper  Unio tigridis tigridis  Bourguignat, 1852  Unio tigridis  This paper  Unio tigridis terminalis  Bourguignat, 1852  Unio tigridis  This paper  Unio abyssinicus  Martens, 1866  ¿?    Unio elongatulus elongatulus  C. Pfeiffer, 1825  Unio elongatulus  Prié et al. (2012); this paper  Unio elongatulus pallens  Rossmässler, 1842  Unio elongatulus + Unio pictorum  This paper  Unio elongatulus bourgueticus  Bourguignat, 1882  Unio pictorum  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus turtoni  Payraudeau, 1826  Unio mancus  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus moquinianus  Dupuy, 1843  Unio mancus  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus aleroni  Companyo & Massot, 1845  Unio mancus  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus rousii  Dupuy, 1849  Unio pictorum  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus mancus  Lamarck, 1819  Unio pictorum  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus requienii  Michaud, 1831  Unio mancus + Unio pictorum  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus penchinatianus  Bourguignat, 1865  Unio ravoisieri  Khalloufi et al. (2011)   Unio elongatulus valentinus  Rossmässler, 1854  Unio mancus  Araujo et al. (2005); Araujo et al. (2009)   Unio elongatulus gargottae  Philippi, 1836  Unio mancus?  This paper  Unio elongatulus lawleyianus  Gentiluomo, 1868  Unio mancus + Unio pictorum?  This paper  Unio elongatulus glaucinus  Porro, 1838  Unio elongatulus + Unio mancus + Unio pictorum  This paper  Unio elongatulus eucirrus  Bourguignat, 1860  Unio tigridis + Unio pictorum  This paper  Unio elongatulus dembeae  Sowerby, 1865  ¿?    Unio elongatulus durieui  Deshayes, 1847  Unio durieui + Unio foucauldianus  Araujo et al. (2009); Khalloufi et al. (2011); Froufe et al. (2016a)   Unio crassus crassus  Retzius, 1788  Unio crassus  This paper  Unio crassus cytherea  Küster, 1833  Unio crassus  This paper  Unio crassus batavus  Maton & Racket, 1807  Unio crassus  This paper  Unio crassus carneus  Küster, 1854  Unio crassus  This paper  Unio crassus ionicus  Drouet, 1879  Unio ionicus  This paper  Unio crassus gontieri  Bourguignat, 1856  Unio crassus  This paper  Unio crassus bruguierianus  Bourguignat, 1853  Unio bruguierianus  This paper  Unio crassus mongolicus  Middendorff, 1851  Unio crassus  This paper  Haas (1969)   Author  Accepted species  References  Unio tumidus tumidus  Retzius, 1788  Unio tumidus  This paper  Unio tumidus borysthenensis  Kobelt, 1879  Unio tumidus  This paper  Unio pictorum pictorum  Linnaeus, 1758  Unio pictorum  This paper  Unio pictorum praeposterus  Küster, 1854  Unio pictorum  This paper  Unio pictorum latirostris  Küster, 1854  Unio pictorum  This paper  Unio pictorum platyrhynchus  Rossmässler, 1835  Unio pictorum  This paper  Unio pictorum schrenkianus  Clessin, 1880  Unio pictorum  This paper  Unio pictorum ascanius  Kobelt, 1913  Unio pictorum  This paper  Unio pictorum proechistus  Bourguignat, 1870  Unio pictorum  This paper  Unio pictorum gaudioni  Drouët, 1881  Unio pictorum  This paper  Unio pictorum rostratus  Lamarck, 1819  Unio pictorum  This paper  Unio pictorum platyrhynchoideus  Dupuy, 1849  Unio pictorum  This paper  Unio pictorum mucidus  Morelet, 1845  Unio delphinus  Araujo et al. (2009); Reis et al. (2013); this paper  Unio pictorum delphinus  Spengler, 1793  Unio delphinus  Araujo et al. (2009); Reis et al. (2013); this paper  Unio pictorum ravoisieri  Deshayes, 1848  Unio ravoisieri  Araujo et al. (2009); Khalloufi et al. (2011); this paper  Unio tigridis tigridis  Bourguignat, 1852  Unio tigridis  This paper  Unio tigridis terminalis  Bourguignat, 1852  Unio tigridis  This paper  Unio abyssinicus  Martens, 1866  ¿?    Unio elongatulus elongatulus  C. Pfeiffer, 1825  Unio elongatulus  Prié et al. (2012); this paper  Unio elongatulus pallens  Rossmässler, 1842  Unio elongatulus + Unio pictorum  This paper  Unio elongatulus bourgueticus  Bourguignat, 1882  Unio pictorum  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus turtoni  Payraudeau, 1826  Unio mancus  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus moquinianus  Dupuy, 1843  Unio mancus  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus aleroni  Companyo & Massot, 1845  Unio mancus  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus rousii  Dupuy, 1849  Unio pictorum  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus mancus  Lamarck, 1819  Unio pictorum  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus requienii  Michaud, 1831  Unio mancus + Unio pictorum  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus penchinatianus  Bourguignat, 1865  Unio ravoisieri  Khalloufi et al. (2011)   Unio elongatulus valentinus  Rossmässler, 1854  Unio mancus  Araujo et al. (2005); Araujo et al. (2009)   Unio elongatulus gargottae  Philippi, 1836  Unio mancus?  This paper  Unio elongatulus lawleyianus  Gentiluomo, 1868  Unio mancus + Unio pictorum?  This paper  Unio elongatulus glaucinus  Porro, 1838  Unio elongatulus + Unio mancus + Unio pictorum  This paper  Unio elongatulus eucirrus  Bourguignat, 1860  Unio tigridis + Unio pictorum  This paper  Unio elongatulus dembeae  Sowerby, 1865  ¿?    Unio elongatulus durieui  Deshayes, 1847  Unio durieui + Unio foucauldianus  Araujo et al. (2009); Khalloufi et al. (2011); Froufe et al. (2016a)   Unio crassus crassus  Retzius, 1788  Unio crassus  This paper  Unio crassus cytherea  Küster, 1833  Unio crassus  This paper  Unio crassus batavus  Maton & Racket, 1807  Unio crassus  This paper  Unio crassus carneus  Küster, 1854  Unio crassus  This paper  Unio crassus ionicus  Drouet, 1879  Unio ionicus  This paper  Unio crassus gontieri  Bourguignat, 1856  Unio crassus  This paper  Unio crassus bruguierianus  Bourguignat, 1853  Unio bruguierianus  This paper  Unio crassus mongolicus  Middendorff, 1851  Unio crassus  This paper  View Large Using a coalescence-based approach to delimit species boundaries, we identified eight lineages in congruence with the current taxonomy of the group. Some of these species were formerly considered at the subspecific level (e.g. Haas, 1969); however, their taxonomic statuses were recently revised following recent molecular analyses (e.g. Araujo & de Jong, 2015; Lopes-Lima et al., 2017). In this study, we show that U. delphinus (Araujo et al., 2009b) and U. ravoisieri (Khalloufi et al., 2011), named as subspecies of U. pictorum (Haas, 1969), actually belong to different phylogenetic groups. We also propose the validity of the Mediterranean U. mancus from France, Sardinia (Italy) and Spain, considered by Haas (1969) as several U. elongatulus subspecies (Table 2) (Araujo et al., 2005; Araujo et al., 2009b; Prié, 2011; Prié et al., 2012; Prié & Puillandre, 2014). The presence of Atlantic populations of this species in France (Drée at the Loire, Garonne and Seine rivers) (Prié et al., 2012; Prié & Puillandre, 2014) may be explained by palaeohydrological events of stream capture or by river connections via the construction of old channels (during the 18th century) in France that joined Mediterranean and Atlantic rivers (Prié et al., 2012; Prié & Puillandre, 2014). The western species U. delphinus and U. foucauldianus were probably misidentified as U. pictorum delphinus and U. elongatulus durieui by Haas (1969) (Khalloufi et al., 2011; Table 2). Although these two species are separated by the smallest genetic distance, they are independent lineages, a finding also confirmed by microsatellite marker analyses (Froufe et al., 2016a), and are distributed on both sides of the Gibraltar Strait (Araujo et al., 2009b; Froufe et al., 2016a). In addition, we found U. foucauldianus in some Mediterranean rivers of Morocco (Martil, Laou and Moulouya), where it had not been cited previously. The GenBank numbers used in Froufe et al. (2016a) were not available when we analysed our data, but we sampled many of the same Atlantic Moroccan rivers. We also restored the species U. ionicus and U. bruguierianus, which were both included by Haas (1969) in the U. crassus clade. The populations from the Axios (= Vardar River) and Pinios rivers in eastern Greece belong to U. bruguierianus, although other populations of the Sofaditikos River in the Pinios Basin belong to U. crassus. The presence of these two species in Greek rivers in the same basin can be explained by river connections and the complex palaeogeography of this peninsula (Steininger & Rogl, 1984; Economidis & Miller, 1990). These processes were responsible for the great species radiation in Barbus (Karakousis et al., 1995; Gante, 2011), one of the main host fishes of Unio. Recently, Froufe et al. (2016b) and Araujo et al. (2016) demonstrated a similar case of endemism for Potomida acarnanica Kobelt, 1879 in the Balkans. Of particular interest, however, are the six main discordances found in our species delimitation analysis and the established taxonomy: Unio tigridis was divided into two subspecies by Haas (1969). The studied samples from Israel and western Turkey were also retrieved as two separate species in our analyses. However, until more samples are studied from the eastern Turkish rivers, Syria and Iraq, we propose maintaining the name U. tigridis. The lack of comprehensive geographic sampling emphasizes the need for in-depth study. This was noted by Falkner (1994) who referred to the ‘puzzling form chaos’ of U. tigridis and Unio terminalis in the Middle East. Interestingly, Froufe et al. (2016b) observed a similar distribution in Potomida semirugata Lamarck, 1819. Haas (1969) described 12 subspecies within U. pictorum (Table 2). With the exception of U. delphinus (Araujo et al., 2009b) and U. ravoisieri (Khalloufi et al., 2011), both considered as species in this study, the other 10 subspecies grouped into two main clades/species, including one that consists of only specimens from northeastern Greece (2.38% genetic distance from the other specimens of this clade). Unio pictorum is distributed from France to Ukraine, Russia and Iran, but not in North Africa, the Iberian Peninsula or, with the exception of Greece and some localities in the Rhône catchment, the Mediterranean rivers of Europe. The presence of this species in Greece can be explained by dispersion from the Danube, as in some Greek Barbus species (Karakousis et al., 1995). Nagel (2000) concluded that the present-day population structure of U. pictorum in central Europe was best explained by river connections during the Pliocene and Pleistocene. Whether the populations of northeastern Greece constitute a different species, as suggested in our study, deserves further studies. Similarly, the specific status of the two main lineages/species of U. gibbus revealed in our study (and in a previous study; Khalloufi et al., 2011), corresponding to the populations of South Spain and Morocco, and Tunisia, respectively, should be further explored. The close relationship between the southern Iberian and northwestern Maghreb species, compared with that among North African and Moroccan populations, or the Iberian with other European populations, provides further evidence of the presence of the Messinian Betic-Riffian Massif, which disappeared some 5.3 Mya (Krijgsman et al., 1999). This biogeographic pattern is similar to that shown by cyprinid fish, and other fish lineages, that were considered different species (Machordom & Doadrio, 2001). The other North African species, U. ravoisieri, included in the pictorum group by Haas (1969), had been redescribed and discussed by Khalloufi et al. (2011). Regarding their Spanish populations in Catalonia, Khalloufi et al. (2011) proposed two hypotheses to explain their current distribution patterns: some historical human transport by the Phoenicians, Romans or Almohads from Tunisian rivers, as has been hypothesized for other animal groups (see Recuero et al., 2007; Araujo et al., 2017), or, more doubtfully, a connection among freshwater courses between these areas during the Messinian crisis, which dried up the Mediterranean (5.5 Mya), leading to the subsequent extinction of connecting populations. These hypotheses suggest a possible foreign origin for the Spanish U. ravoisieri populations; however, no clear geographic pattern is found among the samples included in the two clades obtained in our study. Finally, the species with a larger distribution and more taxonomic problems was U. crassus. The species delimitation analyses revealed from two to six main lineages/species with no clear geographic structuring. However, the absence of morphological features and clear biogeographic patterns require a deeper survey including a greater number of locations and specimens. The phylogenetic and the calibrated species tree revealed a strong congruence between analyses. In general, we obtained very broad time intervals for all nodes of the phylogeny, which is expected given the lack of reliable fossils or well-characterized palaeobiogeographic events to optimally calibrate the molecular clocks and the phylogenetic uncertainties. Temporal estimates, therefore, should be considered with caution and only used to obtain a broad picture of the diversification patterns within the group. Some relevant aspects deserve further study: (1) In general, diversification is old among and between groups. Most of the specific diversity is of Miocene origin. Species are long lived and characterized by some morphological stasis. Furthermore, recent palaeogeographic and climatic events, such as Holocene glaciations, seem to have affected the population and demographic dynamics of the species but not diversification per se. This new temporal and systematic framework for the group may help to reinterpret the idea of some species of Unio forming a polytypic species (ring species or Rassenkreis), with continuous gene flow and thus, very recent or as yet incomplete separation of lineages (e.g. Nagel, Badino & Celebrano, 1998; Nagel & Badino, 2001). (2) The three species with broad European distribution ranges (U. crassus, U. pictorum and U. tumidus) are included in three well-differentiated clades in the phylogenetic analyses. Species with more restricted distribution, such as the ones present in the Iberian Peninsula and Maghreb (U. delphinus + U. foucauldianus, U. tumidiformis, U. ravoisieri, U. gibbus + U. durieui) were also well differentiated. They all colonized or differentiated in the Iberian Peninsula at very different times, from different origins and following very different palaeogeographic events. Unio gibbus, for instance, is probably of North African origin, while U. tumidiformis is related to European U. crassus, U ionicus and U. bruguierianus. Differences between U. crassus and the Iberian endemic U. tumidiformis were previously described by Reis & Araujo (2009), who suggested that U. tumidiformis may have evolved from a common ancestor in the Betic-Riffian Massif, which eventually contributed to the formation of the Guadiana and Guadalquivir basins (Vargas, Real & Guerrero, 1998). Using a relaxed molecular clock based on a mean COI rate substitution of 0.3% per million years, we estimate that the common ancestor of U. tumidiformis and the U. crassus complex existed about 15 Mya, during the Mid-Miocene. A formal comparative biogeographic study, however, would be necessary to fully understand the evolutionary history of the species in the region. (3) The phylogenetic position of U. tumidus is controversial in our study (Figs 1, 15). This clearly deserves further sampling and study as the evolutionary and biogeographic interpretations could drastically change depending on these results. Freshwater mussels are among the most endangered freshwater species worldwide. Providing robust phylogenetic and systematic hypotheses for the group is of paramount importance to design effective conservation and management plans, either at local or regional scales. The different approaches addressed in this study have helped establishing a new and robust systematic framework for Western Palaearctic Unio, which we hope translates into new conservation and management plans. While most of the unionid populations are in regression in the studied areas, new taxa identified here, and those with limited distribution, should be urgently protected and conserved, particularly in the case of U. bruguierianus, U. ionicus or U. elongatulus. Moreover, the potential existence of new species within currently recognized morphospecies (e.g. U. gibbus, U. tigridis or U. crassus) should prompt further detailed studies. Some of the new species could represent very localized endemics with very limited distribution areas and, hence, be highly endangered. SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article at the publisher’s web-site: Table S1. Information on the specimens analysed and GenBank accession numbers. *Misnamed in GenBank. Figure S1. Network of U. durieui haplotypes. Figure S2. Network of U. gibbus haplotypes. Figure S3. Network of U. tumidus haplotypes. Figure S4. Network of U. delphinus (Iberia) and U. foucauldianus (Morocco) haplotypes. Figure S5. Network of U. tigridis haplotypes. Figure S6. Network of U. ravoisieri haplotypes. Figure S7. Network of U. mancus haplotypes. Figure S8. Network of U. elongatulus haplotypes. Figure S9. Network of U. pictorum haplotypes. Figure S10. Network of U. tumidiformis haplotypes. ACKNOWLEDGEMENTS We thank Carlos Toledo for his work on some of the analyses. Joaquim Reis, José Miguel Barea, María José Madeira, Bülent Yorulmaz, Günter Hartz, Peter Reischütz, Joseph Heller, Mohamed Ghamizi and Noureddine Khalloufi helped us collect the specimens. Arthur Bogan sent us the Ukrainian and Russian specimens. The plates were compiled by Jesús Muñoz from the photography facility of the MNCN. This study was partially funded by the Spanish Ministry of Economy and Competitiveness (Ref. CTM2014-57949-R). Dimitrios Germanos provided us with the collection permits to sample in Greece, Pëllumb Abeshi in Albania and the Ministry of Culture, Nature Protection Directorate in Croatia. The help of Eladio Fernández Galiano (Council of Europe) was invaluable. Sarah Young and Melinda Modrell reviewed the English version of the manuscript. REFERENCES Araujo R, Buckley D, Nagel K-O, Machordom A. 2017. Potomida littoralis (Bivalvia, Unionidae) evolutionary history: slow evolution or recent speciation? Zoological Journal of the Linnean Society  179: 277– 290. doi: 10.1111/zoj.12470. Araujo R, de Jong Y. 2015. Fauna Europaea: Mollusca-Bivalvia. Biodiversity Data Journal  3: e5211. doi: 10.3897/BDJ.3.e5211. Google Scholar CrossRef Search ADS   Araujo R, Gómez I, Machordom A. 2005. The identity and biology of Unio mancus (= U. elongatulus) (Bivalvia: Unionidae) in the Iberian Peninsula. Journal of Molluscan Studies  71: 25– 31. Google Scholar CrossRef Search ADS   Araujo R, Reis J, Machordom A, Toledo C, Madeira MJ, Gómez I, Velasco JC, Morales J, Barea JM, Ondina P, Ayala I. 2009b. Las náyades de la península Ibérica. Iberus  27: 7– 72. Araujo R, Schneider S, Erpenbeck D, Roe KJ, Machordom A. 2017. The origin and phylogeny of Margaritiferidae (Bivalvia: Unionoida): a synthesis of fossil and molecular data. Zoologica Scripta  46: 289– 307. Google Scholar CrossRef Search ADS   Araujo R, Toledo C, Machordom A. 2009a. Redescription of Unio gibbus Spengler, 1793, a west Palaearctic freshwater mussel with hookless glochidia. Malacologia  51: 131– 141. Google Scholar CrossRef Search ADS   Ball GH. 1922. Variation in freshwater mussels. Ecology  3: 93– 121. Google Scholar CrossRef Search ADS   Baum DA. 1998. Individuality and the existence of species through time. Systematic Biology  47: 641– 653. Google Scholar CrossRef Search ADS PubMed  Bouckaert R, Heled J, Kühnert D, Vaughan T, Wu CH, Xie D, Suchard MA, Rambaut A, Drummond AJ. 2014. BEAST 2: a software platform for Bayesian evolutionary analyses. PLoS Biology  10: e1003537. Google Scholar CrossRef Search ADS   Carstens BC, Pelletier TA, Reid NM, Satdler JD. 2013. How to fail at species delimitation. Molecular Ecology  22: 4369– 4383. Google Scholar CrossRef Search ADS PubMed  Choi SC. 2016. Methods for delimiting species via population genetics and phylogenetics using genotype data. Genes and Genomics  38: 905– 915. Google Scholar CrossRef Search ADS   Coyne JA, Orr HA, Futuyma DJ. 1988. Do we need a new species concept? Systematic Zoology  37: 190– 200. Google Scholar CrossRef Search ADS   Cuttelod A, Seddon M, Neubert E. 2011. European red list of non-marine Molluscs . Luxembourg: Publications Office of the European Union. Drummond AJ, Rambaut A. 2007. BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evolutionary Biology  7: 214. Google Scholar CrossRef Search ADS PubMed  Drummond AJ, Suchard MA, Xie D, Rambaut A. 2012. Bayesian phylogenetics with BEAUti and the BEAST 1.7. Molecular Biology and Evolution  29: 1969– 1973. Google Scholar CrossRef Search ADS PubMed  Eagar RMC. 1978. Shape and function of the shell: a comparison on some living and fossil bivalve molluscs. Biological Reviews  53: 169– 210. Google Scholar CrossRef Search ADS   Economidis P, Miller P. 1990. Systematics of freshwater gobies from Greece (Teleostei; Gobiidae). Journal of Zoology  221: 125– 170. Google Scholar CrossRef Search ADS   Ereshefsky M. 2014. Species, historicity, and path dependency. Philosophy of Science  81: 714– 726. Google Scholar CrossRef Search ADS   Falkner G. 1994. Systematik vorderorientalischer Najaden als Vorstudie zur Bearbeitung archäologischer Funde. Forschungen und Berichte zur Vor- und Frühgeschichte in Baden-Württemberg  53: 135– 162. Felsenstein J. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution  39: 783– 791. Google Scholar CrossRef Search ADS PubMed  Folmer O, Black M, Hoeh W, Lutz R, Vrijenhoek R. 1994. DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Molecular Marine Biology and Biotechnology  3: 294– 299. Google Scholar PubMed  Froufe E, Gonçalves DV, Teixeira A, Sousa R, Varandas S, Ghamizi M, Zieritz A, Lopes-Lima M. 2016a. Who lives where? Molecular and morphometric analyses clarify which Unio species (Unionida, Mollusca) inhabit the southwestern Palearctic. Organisms Diversity & Evolution  16: 597– 611. doi: 10.1007/s13127-016-0262-x. Google Scholar CrossRef Search ADS   Froufe E, Prié V, Faria J, Ghamizi M, Gonçalves DV, Gürlek ME, Karaouzas I, Kebapçi Ü, Sereflisan H, Sobral C, Sousa R, Teixeira A, Varandas S, Zogaris S, Lopes-Lima M. 2016b. Phylogeny, phylogeography, and evolution in the Mediterranean region: news from a freshwater mussel (Potomida, Unionida). Molecular Phylogenetics and Evolution  100: 322– 332. Google Scholar CrossRef Search ADS   Gante HF. 2011. Diversification of Circum-Mediterranean barbels. In: Grillo O, Venora G, eds. Changing diversity in changing environment . Rijeka: InTech, 283– 298. Guindon S, Dufayard JF, Lefort V, Anisimova M, Hordijk W, Gascuel O. 2010. New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the performance of PhyML 3.0. Systematic Biology  59: 307– 321. Google Scholar CrossRef Search ADS PubMed  Haas F. 1969. Superfamilia Unionacea. Das Tierreich  88: 1– 663. Hanken J, Wake DB. 1993. Miniaturization of body size: organismal consequences and evolutionary significance. Annual Review of Ecology and Systematics  24: 501– 519. Google Scholar CrossRef Search ADS   Hebert PDN, Cywinska A, Ball SL, De Waard JR. 2003. Biological identifications through DNA barcodes. Proceedings of the Royal Society of London B: Biological Sciences  270: 313– 321. Google Scholar CrossRef Search ADS   Huelsenbeck JP, Ronquist F. 2001. MrBayes: Bayesian inference of phylogenetic trees. Bioinformatics  17: 754– 755. Google Scholar CrossRef Search ADS PubMed  Kapli P, Lutteropp S, Zhang J, Kobert K, Pavlidis P, Stamatakis A, Flouri T. 2016. Multi-rate tree processes for single-locus species delimitation under maximum likelihood and Markov chain Monte Carlo. bioRxiv. Available at: http://biorxiv.org/content/early/2016/07/14/063875 Karakousis Y, Machordom A, Doadrio I, Economidis PS. 1995. Phylogenetic relationships of Barbus peloponnesius Valenciennes, 1842 (Osteichthyes: Cyprinidae) from Greece and other species of Barbus as revealed by allozyme electrophoresis. Biochemical Systematics and Ecology  23: 365– 375. Google Scholar CrossRef Search ADS   Khalloufi N, Toledo C. Machordom A, Boumaïza M, Araujo R. 2011. The unionids of Tunisia: taxonomy and phylogenetic relationships, with redescription of Unio ravoisieri Deshayes, 1847 and U. durieui Deshayes, 1847. Journal of Molluscan Studies  77: 1– 13. Google Scholar CrossRef Search ADS   Krijgsman W, Hilgen FJ, Raffi I, Sierro FJ, Wilson DS. 1999. Chronology, causes and progression of the Messinian salinity crisis. Nature  400: 652– 655. Google Scholar CrossRef Search ADS   Lee M, Wolsan M. 2002. Integration, individuality, and species concepts. Biology and Philosophy  17: 651– 660. Google Scholar CrossRef Search ADS   Librado P, Rozas J. 2009. DnaSP v5: a software for comprehensive analysis of DNA polymorphism data. Bioinformatics  25: 1451– 1452. Google Scholar CrossRef Search ADS PubMed  Lopes-Lima M, Sousa R, Geist J, Aldridge DC, Araujo R, Bergengren J, Bespalaja Y, Bódis E, Burlakova L, Van Damme D, Douda K, Froufe E, Georgiev D, Gumpinger C, Karatayev A, Kebapçi Ü, Killeen I, Lajtner J, Larsen BM, Lauceri R, Legakis A, Lois S, Lundberg S, Moorkens E, Motte G, Nagel KO, Ondina P, Outeiro A, Paunovic M, Prié V, von Proschwitz T, Riccardi N, Rudzīte M, Rudzītis M, Scheder C, Seddon M. Şereflişan H, Simić V, Sokolova S, Stoeckl K, Taskinen J, Teixeira A, Thielen F, Trichkova T, Varandas S, Vicentini H, Zajac K, Zajac T, Zogaris S. 2017. Conservation status of freshwater mussels in Europe: state of the art and future challenges. Biological Reviews  92: 572– 607. Google Scholar CrossRef Search ADS PubMed  Lydeard C, Cowie RH, Ponder WF, Bogan AE, Bouchet P, Clarck SA, Cummings KS, Frest TJ, Gargominy O, Herbert DG, Hershler R, Perez KE, Roth B, Seddon M, Strong EE, Thompson FG. 2004. The global decline of non-marine molluscs. BioScience  54: 321– 330. Google Scholar CrossRef Search ADS   Lydeard C, Mulvey M, Davis GM. 1996. Molecular systematics and evolution of reproductive traits of North American freshwater unionacean mussels (Mollusca: Bivalvia) as inferred from 16S rRNA gene sequences. Philosophical Transactions of the Royal Society of London B: Biological Sciences  351: 1593– 1603. Google Scholar CrossRef Search ADS   Machordom A, Araujo R, Erpenbeck D, Ramos MA. 2003. Phylogeography and conservation genetics of endangered European Margaritiferidae (Bivalvia: Unionoidea). Biological Journal of the Linnean Society  78: 235– 252. Google Scholar CrossRef Search ADS   Machordom A, Doadrio I. 2001. Evidence of a cenozoic Betic-Kabilian connection based on freshwater fish phylogeography (Luciobarbus, Cyprinidae). Molecular Phylogenetics and Evolution  18: 252– 263. Google Scholar CrossRef Search ADS PubMed  Mayden RL. 1997. A hierarchy of species concepts: the denouement in the saga of the species problem. In: Claridge MF, Dawah HA, Wilson MR, eds. Species: the units of biodiversity . New York: Chapman & Hall, 381– 424. Moore RC. 1969. Treatise on invertebrate paleontology: Mollusca 6 Bivalvia, Part N, Vol. 1 . Boulder: The University of Kansas and The Geological Society of America Inc., 1– 489. Nagel KO. 2000. Testing hypotheses on the dispersal and evolutionary history of freshwater mussels (Mollusca: Bivalvia: Unionidae). Journal of Evolutionary Biology  13: 854– 865. Google Scholar CrossRef Search ADS   Nagel KO, Badino G. 2001. Population genetics and systematics of European Unionoidea. In: Bauer G, Wächtler K, eds. Ecological studies: ecology and evolution of the freshwater mussels Unionoida, Vol. 145 . Berlin: Springer-Verlag, 51– 80. Google Scholar CrossRef Search ADS   Nagel KO, Badino G, Celebrano G. 1998. Systematics of European naiades (Bivalvia: Margaritiferidae and Unionidae): a review and some new aspects. Malacological Review (Bivalvia I)  ( suppl. 7): 83– 104. Pons J, Barraclough TG, Gomez-Zurita J, Cardoso A, Duran DP, Hazell S, Kamoun S, Sumlin WD, Vogler AP. 2006. Sequence-based species delimitation for the DNA taxonomy of undescribed insects. Systematic Biology  55: 595– 609. Google Scholar CrossRef Search ADS PubMed  Posada D. 2004. Collapse ver. 1.2: a tool for collapsing sequences to haplotypes . Available at: http://darwin.uvigo.es Posada D. 2008. jModelTest: phylogenetic model averaging. Molecular Biology and Evolution  25: 1253– 1256. Google Scholar CrossRef Search ADS PubMed  Prié V. 2011. Les sous-espèces de la Mulette méridionale Unio mancus Lamarck 1819 (Bivalvia, Unionidea) en France: descriptions originales et matériel topotypique. MalaCo  8: 1– 19. Prié V, Puillandre N. 2014. Molecular phylogeny, taxonomy, and distribution of French Unio species (Bivalvia, Unionidae). Hydrobiologia  735: 95– 110. Google Scholar CrossRef Search ADS   Prié V, Puillandre N, Bouchet P. 2012. Bad taxonomy can kill: molecular reevaluation of Unio mancus Lamarck, 1819 (Bivalvia, Unionodae) and its accepted subspecies. Knowledge and Management of Aquatic Ecosystems  405: 1– 18. de Queiroz K. 2005. Ernst Mayr and the modern concept of species. Proceedings of the National Academy of Sciences of the United States of America  102: 6600– 6607. Google Scholar CrossRef Search ADS PubMed  de Queiroz K. 2007. Species concepts and species delimitation. Systematic Biology  56: 879– 886. Google Scholar CrossRef Search ADS PubMed  R Core Team. 2016. R: a language and environment for statistical computing . Vienna: R Foundation for Statistical Computing. Available at: http://www.R-project.org/ Rambaut A, Suchard MA, Xie D, Drummond AJ. 2014. Tracer v1.6 . Available at: http://beast.bio.ed.ac.uk/Tracer Recuero E, Iraola A, Machordom A, García-París M. 2007. Mitochondrial differentiation and biogeography of Hyla meridionalis (Anura: Hylidae): an unusual phylogeographical pattern. Journal of Biogeography  34: 1207– 1219. Google Scholar CrossRef Search ADS   Reid NM, Carstens BC. 2012. Phylogenetic estimation error can decrease the accuracy of species delimitation: a Bayesian implementation of the general mixed Yule-coalescent model. BMC Evolutionary Biology  12: 196. Google Scholar CrossRef Search ADS PubMed  Reis J, Araujo R. 2009. Redescription of Unio tumidiformis Castro, 1899 (Bivalvia: Unionoidea), an endemism from Southwestern Iberian Peninsula. Journal of Natural History  43: 1929– 1945. Google Scholar CrossRef Search ADS   Reis J, Machordom A, Araujo R. 2013. Morphological and molecular diversity of Unionidae (Mollusca: Bivalvia) from Portugal. Graellsia  69: 17– 36. Ronquist F, Huelsenbeck JP. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics  19: 1572– 1574. Google Scholar CrossRef Search ADS PubMed  Roux C, Fraïsse C, Romiguier J, Anciaux Y, Galtier N, Bierne N. 2016. Shedding light on the grey zone of speciation along a continuum of genomic divergence. PLoS Biology  14: e2000234. Google Scholar CrossRef Search ADS PubMed  Rowe KC, Aplin KP, Baverstock PR, Moritz C. 2011. Recent and rapid speciation with limited morphological disparity in the genus Rattus. Systematic Biology  60: 188– 203. Google Scholar CrossRef Search ADS PubMed  Rozas J, Rozas R. 1995. DnaSP, DNA sequence polymorphism: an interactive program for estimating population genetics parameters from DNA sequence data. Computer Applications in the Biosciences  11: 621– 625. Google Scholar PubMed  Sites JWJr, Marshall JC. 2004. Operational criteria for delimiting species. Annual Review of Ecology and Systematics  35: 199– 227. Google Scholar CrossRef Search ADS   Stamatakis A, Hoover P, Rougemont J. 2008. A rapid bootstrap algorithm for the RAxML Web servers. Systematic Biology  57: 758– 771. Google Scholar CrossRef Search ADS PubMed  Steininger F, Rogl F. 1984. Paleogeography and palinspastic reconstruction of the Neogene of the Mediterranean. In: Dixon J, Robertson A, eds. The geological evolution of the Eastern Mediterranean. Oxford: Blackwell Scientific Publications, 17: 659– 668. Google Scholar CrossRef Search ADS   Strayer DL, Downing JA, Haag WR, King TL, Layzer JB, Newton TJ, Nichols SJ. 2004. Changing perspectives on pearly mussels, North America’s most imperilled animals. BioScience  54: 429– 439. Google Scholar CrossRef Search ADS   Swofford DL. 2002. PAUP*: Phylogeny Analysis Using Parsimony (*and other methods), Version 4.0a147 . Sunderland: Sinauer Associates Inc. Tang CQ, Humphreys AM, Fontaneto D, Barraclough TG. 2014. Effects of phylogenetic reconstruction method on the robustness of species delimitation using single-locus data. Methods in Ecology and Evolution  5: 1086– 1094. Google Scholar CrossRef Search ADS PubMed  Turelli M, Barton NH, Coyne JA. 2001. Theory and speciation. Trends in Ecology & Evolution  16: 330– 343. Google Scholar CrossRef Search ADS PubMed  Van Bocxlaer B, Hunt G. 2013. Morphological stasis in an ongoing gastropod radiation from Lake Malawi. Proceedings of the National Academy of Sciences of the United States of America  110: 13892– 13897. Google Scholar CrossRef Search ADS PubMed  Vargas JM, Real R, Guerrero JC. 1998. Biogeographical regions of the Iberian Peninsula based on freshwater fish and amphibian distributions. Ecography  21: 371– 382. Google Scholar CrossRef Search ADS   Vaughn CC, Hakenkamp CC. 2001. The functional role of burrowing bivalves in freshwater ecosystems. Freshwater Biology  46: 1431– 146. Google Scholar CrossRef Search ADS   Wake DB. 2009. What salamanders have taught us about evolution. Annual Review of Ecology, Evolution and Systematics  40: 333– 352. Google Scholar CrossRef Search ADS   Wake DB, Wake MH, Specht CD. 2011. Homoplasy: from detecting pattern to determining process and mechanism of evolution. Science  331: 1032– 1035. Google Scholar CrossRef Search ADS PubMed  Wheeler QD, Meier R, eds. 2000. Species concepts and phylogenetic theory . New York: Columbia University Press. Wilson BE. 1995. A (not-so) radical solution to the species problem. Biology and Philosophy  10: 339– 356. Google Scholar CrossRef Search ADS   Zhang J, Kapli P, Pavlidis P, Stamatakis A. 2013. A general species delimitation method with applications to phylogenetic placements. Bioinformatics  29: 2869– 2876. Google Scholar CrossRef Search ADS PubMed  Zieritz A, Hoffman JI, Amos W, Aldridge DC. 2010. Phenotipic plasticity and genetic isolation-by-distance in the freshwater mussel Unio pictorum (Mollusca: Unionoida). Evolutionary Ecology  24: 923– 938. Google Scholar CrossRef Search ADS   © 2017 The Linnean Society of London, Zoological Journal of the Linnean Society http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Zoological Journal of the Linnean Society Oxford University Press

Species boundaries, geographic distribution and evolutionary history of the Western Palaearctic freshwater mussels Unio (Bivalvia: Unionidae)

Loading next page...
 
/lp/ou_press/species-boundaries-geographic-distribution-and-evolutionary-history-of-QyQ7bhn0dn
Publisher
The Linnean Society of London
Copyright
© 2017 The Linnean Society of London, Zoological Journal of the Linnean Society
ISSN
0024-4082
eISSN
1096-3642
D.O.I.
10.1093/zoolinnean/zlx039
Publisher site
See Article on Publisher Site

Abstract

Abstract Delimiting species boundaries is a fundamental yet challenging task that has been especially problematic in morphologically conserved lineages. The use of genetic markers and phylogenetic approaches has proven of paramount importance to resolve these cases. Using the genus Unio as a case study, we have analysed two mitochondrial markers, cytochrome c oxidase subunit I (COI) and 16S rRNA (16S), phylogenetically and phylogeographically to explore species limits and evolutionary lineages within the group. We followed two different approaches to define species boundaries: the generalized mixed Yule coalescence model in a Bayesian framework and the Poisson tree process model. Phylogenetic relationships among species and divergence times were also assessed using relaxed molecular clock analyses. Taken together, we provide a sound systematic framework for Western Palaearctic Unio species, addressing phylogenetic relationships, and the taxonomy, biogeographic patterns and evolutionary history of this group in this region. The Unio clade from the studied area had a clear phylogenetic structure with four robust lineages that include the following species: (1) Unio durieui + Unio gibbus; (2) Unio tumidus; (3) Unio pictorum, Unio delphinus + Unio foucauldianus, Unio elongatulus, Unio mancus, Unio ravoisieri, Unio tigridis; and (4) Unio tumidiformis + (Unio bruguierianus + (Unio crassus + Unio ionicus)). Deeper analyses, including filling gaps in distributions, should be conducted to disentangle the U. crassus, U. pictorum and U. tigridis species complexes. INTRODUCTION Delimiting species boundaries is a fundamental yet challenging task. The challenge derives from species being a dynamic spatial and temporal continuum (e.g. de Queiroz, 2005, 2007; Wake, 2009). It also derives from the complexity and the variety of possible mechanisms from which new species originate (Turelli, Barton & Coyne, 2001). Much debate exists about these topics, which is reflected in the literature on epistemological/philosophical issues related to the nature of species [i.e. what is actually a species? (e.g. Coyne, Orr & Futuyma, 1988; Wilson, 1995; Baum, 1998; Wheeler & Meier, 2000; Lee & Wolsan, 2002; de Queiroz, 2007; Ereshefsky, 2014)] and on the methodological problems associated with defining species [i.e. what criteria should be used to define and recognize species? (e.g. Mayden, 1997; Sites & Marshall, 2004; de Queiroz, 2005, 2007; Carstens et al., 2013)]. Several factors may contribute to the difficulty of resolving species boundaries observed in nature. Time from speciation is a key factor for detecting new species as recently divergent lineages would go through polyphyletic and paraphyletic stages, the so-called ‘grey zone of speciation’, before all their traits or characters eventually appear as reciprocally monophyletic (de Queiroz, 2005; Roux et al., 2016). Secondary contact and hybridization between differentiated lineages is also an obvious confounding factor. Moreover, given that taxonomists traditionally rely on morphology (in a broad sense) to describe and classify diversity, delimiting species boundaries in morphologically conserved lineages is especially problematic. Discordant within-clade diversity and disparity (e.g. Rowe et al., 2011), morphological stasis among species (e.g. Van Bocxlaer & Hunt, 2013), rampant homoplasy and constraints (e.g. Wake, Wake & Specht, 2011), and the presence of very simplified or specialized morphologies (e.g. parasitic and miniaturized species; Hanken & Wake, 1993) may result in taxonomic uncertainty within groups. The use of genetic markers has proven to be of paramount importance to resolve these cases. Furthermore, there has also been recent renewed interest in using phylogenetic approaches for species delimitations, particularly with the development of DNA barcodes and new species delimitation algorithms based on genotype data (e.g. Choi, 2016). However, results from these single-locus or multilocus DNA species delimitation algorithms, although claimed to be ‘objective’, should not be directly translated into new taxonomies. Instead, they should be considered as a powerful tool to hypothesize potential historical lineages/species to be tested and validated with other sources of information. In this context, we selected as a case study the genus Unio, a Palaearctic freshwater mussel group distributed throughout North Africa (Morocco, Algeria, Tunisia), Europe and parts of Asia, with the exception of the South African U. caffer and the Ethiopian Unio abyssinicus. The fossil record of the group extends back to the Triassic (Moore, 1969). In a seminal work, Haas (1969) considered 42 taxonomic units within Unio that included 12 ‘fundamental Unio species’ comprising different ‘races’ or incipient species; all the descriptions were based on external shell morphology. However, based on preliminary mitochondrial DNA analyses and geographic distribution, the following Western Palaearctic species are currently accepted (Araujo, Gómez & Machordom, 2005; Araujo, Toledo & Machordom, 2009a; Araujo et al., 2009b; Khalloufi et al., 2011; Prié, Puillandre & Bouchet, 2012; Reis, Machordom & Araujo, 2013; Prié & Puillandre, 2014; Lopes-Lima et al., 2017): Unio delphinus Spengler, 1793 (Atlantic Iberian rivers); U. tumidiformis Castro, 1885 (South Atlantic Iberian rivers); Unio mancus Lamarck, 1819 (Atlantic and Mediterranean French rivers and Mediterranean Iberian rivers); U. gibbus Spengler, 1793 (North Africa and the Spanish Barbate River); Unio ravoisieri Deshayes, 1847 (Algeria, Tunisia and two North East Iberian basins); and Unio durieui Deshayes, 1847 (Algeria and Tunisia). Recently, Froufe et al. (2016a) divided the species U. delphinus into the Iberian U. delphinus and the Atlantic Moroccan species U. foucauldianus Pallary, 1936. Other Western Palaearctic species with a wide distribution that were previously cited by Haas (1969), such as Unio crassus Retzius, 1788, Unio pictorum Linnaeus, 1758, Unio tumidus Retzius, 1788, Unio elongatulus C. Pfeiffer, 1825, and Unio tigridis Bourguignat, 1852, still require validation and an update on their distribution (Araujo & de Jong, 2015; Lopes-Lima et al., 2017). Previous taxonomy based on shell characters does not reflect the evolutionary history of the genus, nor does it adequately account for current species. The results obtained in previous works indicate that morphological characters alone are insufficient to establish reliable taxonomic and phylogenetic classification of Unionoida. Furthermore, many authors have concluded that the range of variation of shell shape within Unio species is very wide (Ball, 1922; Haas, 1969; Eagar, 1978; Zieritz et al., 2010), suggesting it may be a survival advantage, provided that such plasticity has functional significance. Given these uncertainties and the poor conservation status of freshwater mussels (Vaughn and Hakenkamp, 2001; Strayer et al., 2004; Lopes-Lima et al., 2017), there is an urgent need for a phylogeographic and taxonomic study of the genus Unio to determine the current Western Palaearctic lineages, their distribution and implications for their conservation. Understanding phylogenetic diversity, to maximize current and future levels of biodiversity, is crucial for conservation strategies and prioritization of freshwater mussels that are among the most threatened freshwater taxa worldwide (IUCN Red List of Threatened Species; Lydeard et al., 2004). However, conservation efforts in West Palaearctic Unionidae are hindered by taxonomic uncertainties (Araujo & de Jong, 2015; Lopes-Lima et al., 2017), and many species have been assigned a Data Deficient status by the IUCN (Cuttelod, Seddon & Neubert, 2011). Here, we sampled individuals across the distribution ranges of the currently recognized Unio species and used coalescence methods to delimit species boundaries in Western Palaearctic lineages. We then validated these results with phylogenetic (well-supported phylogenetically independent lineages), phylogeographic (values of genetic divergence and their intraspecific and interspecific gaps) and geographic (biogeographically distinct distributions) evidence. The resulting delimited species were contrasted with the previous taxonomy based on shell characters. Taken together, we provide a sound systematic framework for Western Palaearctic Unio species, addressing the phylogenetic relationships within the group, and its taxonomy, biogeographic patterns and evolutionary history in the region. MATERIAL AND METHODS We sampled 518 specimens along the distribution range of most of the recognized Unio morphospecies (Supporting Information, Table S1). Sequences from GenBank were also used to complete the ingroup data set (Supporting Information, Table S1) and to test genus monophyly using representatives of Margaritiferidae [Margaritifera auricularia (Spengler, 1793)] and Unionidae [Anodonta anatina (Linnaeus, 1758), Alasminodonta marginata Say, 1818, Alasminodonta triangulata (Lea, 1858) and Pyganodon grandis (Say, 1829)]. We sequenced two mitochondrial markers, cytochrome c oxidase subunit I (COI) and 16S rRNA (16S), for the sampled specimens. Genomic DNA was extracted with the DNeasy kit (Qiagen) following manufacturer’s instructions, except for a longer Proteinase K homogenization (overnight) and the use of RNAse. Small tissue samples were taken from the mantle or foot prior to returning most live specimens back to their habitat. Voucher specimens from each locality were included in the Malacological Collection of the Museo Nacional de Ciencias Naturales (MNCN-CSIC) in Madrid. PCR amplifications were performed with 2 µL of a 1:20 dilution of gDNA, 1× buffer, 2 mM MgCl2, 0.2 mM of each dNTPs, 0.16 µM of each primer (Folmer et al., 1994 and Machordom et al., 2003 for COI and Lydeard, Mulvey & Davis, 1996 for 16S), 1.5 U of Taq polymerase (Biotools) and ddH2O for a final volume of 50 µL. The amplification cycles were as follows: 94 °C for 4 min and 40 cycles of 94 °C for 45 s, 45 °C (for COI amplification) or 50 °C (for 16S amplification) for 1 min, 72 °C for 1 min and a final extension at 72 °C for 10 min. After analysis on 0.8% agarose gels stained with SYBRsafe, the positively amplified fragments were purified and sequenced on an ABI 3730 XL sequencer (Life Technologies) using the Big Dye Terminator Kit (Life Technologies). Electropherograms were trimmed, edited and assembled in Sequencher (Gene Codes Corporation, Ann Arbor, MI, USA). Raw matrices were prepared for the two markers separately. The COI matrix did not need gaps to be aligned, and the 16S required the addition of short gaps, mainly for alignment with the outgroup. The maximum number of gaps was 10 at 6 different regions of the 16S sequences (for U. elongatulus). Unique haplotypes were selected using Collapse (Posada, 2004) for each gene separately; new matrices considering only those were then prepared. Finally, other matrices were prepared containing information for both genes. Delimiting species boundaries We followed two different approaches to define species boundaries in our data set. First, we used the generalized mixed Yule coalescence model in a Bayesian framework (bGMYC) to delimit the number of lineages/species (Pons et al., 2006; Reid & Carstens, 2012). Basically, this model identifies a threshold in branching rates in time-calibrated trees that distinguishes population-level branching processes (i.e. coalescence processes) from species-level cladogenesis (Yule model of diversification). Without an a priori assignment of individuals to putative species, the model returns the probability of groups of individuals to conform to natural species. This analysis was carried out with the bGMYC 1.0.2 package (Reid & Carstens, 2012) in R 3.3.2 (R core Team, 2016). To obtain the calibrated trees necessary to perform the bGMYC analysis, we ran a BEAST 2.4.5 analysis (Bouckaert et al., 2014), partitioning the haplotypic data set by genes. We used the ‘Beast Model test’ option to both estimate the substitution model for each partition during the analysis and calibrate independent relaxed molecular clocks for the two genes with substitution rates retrieved from the literature (see Araujo et al., 2017; Froufe et al., 2016b). We applied lognormal distributions for the COI (mean = 0.003; SD = 0.1) and the 16S (mean = 0.002; SD = 0.1) ucld. mean priors, to specify lineages substitution rates of 0.3 and 0.2% per Myr, respectively. The analysis was run for 100 × 106 generations, saving one in every 10000 trees and associated parameters. The posterior distributions of the topologies and all parameters were synthesized in a maximum clade credibility (MCC) tree in TreeAnnotator 2.4.5 (Drummond & Rambaut, 2007). To run the bGMYC analysis, we randomly selected 100 trees from the post-burnin posterior distribution of ultrametric trees. After checking for convergence of the parameters in a single tree, we ran the bGMYC analysis using the 100 topologies with the following uniform priors: Yule process rate change (0–5), coalescence process rate change (0–2) and threshold parameter (1–50). The latter defines the number of potential species. We let this number vary from 1 to 50, which is over three times the number of currently recognized morphospecies. The GMYC approach is especially sensitive to misspecification of calibrations and branch smoothing, which could bias the accuracy of the resulting number of species (Zhang et al., 2013; Tang et al., 2014). We, therefore, ran a second analysis using the Poisson tree process (PTP) model (Zhang et al., 2013) to compare the number of species inferred. Like the GMYC, the PTP model runs without a priori assigning individuals to potential species, and models the branching rates to distinguish between intra- and interspecific processes. Unlike the GMYC, however, the PTP models speciation in terms of number of substitutions in branches, not time, and it can be run with non-ultrametric phylogenetic trees. To account for potential differences in branch lengths that could be obtained through different phylogenetic inference methods, we ran two separate analyses with trees derived from maximum likelihood (ML) and Bayesian inference (BI). The ML tree was generated on the RAxML webserver (Stamatakis, Hoover & Rougemont, 2008) and the BI tree in the parallel version of MrBayes 3.2.5 (Huelsenbeck & Ronquist, 2001; Ronquist & Huelsenbeck, 2003). The Bayesian analysis was run for 100 × 106 generations, discarding the first 25% as burnin and allowing the program to estimate the substitution model during the Metropolis-coupled Markov chain Monte Carlo (MC3) search. For both ML and BI reconstructions, a sequence from the Chinese Unio douglasiae, the only Eastern Palaearctic species included in this study, was specified as the outgroup. The M-PTP model for the ML and BI trees was run in m-PTP v.0.2.1 (Zhang et al., 2013; Kapli et al., 2016) with a Markov chain Monte Carlo search for 100 × 106 generations, sampling every 10000 generations and discarding the first 10 × 106 as burnin. We ran 100 independent analyses for each tree to confirm that the analyses were properly converging. All bGMYC and M-PTP analyses were run on a haplotypic data set to avoid zero branch lengths in the trees derived from identical sequences, which could severely mislead boundaries delimitation. Species validation: phylogenetic, phylogeographic, morphological and geographic evidence The validity of previously defined species boundaries was tested against historical (phylogenetic), morphological and geographic evidence. To determine the phylogenetic relationships for the species, we conducted maximum parsimony (MP), ML and BI analyses, using PAUP* 4.0a147 (Swofford, 2002), PHYML v.3.0 (Guindon et al., 2010) and MrBayes v3.1.2 (Huelsenbeck & Ronquist, 2001; Ronquist & Huelsenbeck, 2003), respectively. MP analyses were carried out through heuristic searches, using the tree bisection and reconnection algorithm, including ten random stepwise additions. For the ML analyses, we estimated best-fit models in jModelTest (Posada, 2008). Supports for both MP and ML analyses were determined by bootstrapping (Felsenstein, 1985) 1000 pseudoreplicates. BI was performed through 10 × 106 generations in two parallel runs, with invgamma and nst = 6 options. A consensus tree was computed after eliminating the first 25% of sampled generations as burnin, evaluating node supports by the posterior probabilities. Tracer (Rambaut et al., 2014) was used to verify that the two runs were long enough to converge and reach stability. These phylogenetic analyses were run on the haplotypic data set with the full set of outgroups. Phylogenetic relationships among species and their divergence times were also assessed using relaxed molecular clock analyses in a Bayesian-coalescence approach with the software starBEAST2 in BEAST 2.4.5 (Drummond & Rambaut, 2007; Drummond et al., 2012), using the UncorrelatedLogNormal template. We ran different analyses grouping the individuals according to the results of the bGMYC and M-PTP analyses. To calibrate the relaxed molecular clocks, we used a lognormal distribution analysis with a mean substitution rate of 0.003 (0.3% substitutions/lineage/Myr) and 0.002 (0.2% substitutions/lineage/Myr) for COI and 16S, respectively, and a SD of 0.1 for both genes (see Araujo et al., 2016; Froufe et al., 2016b). All analyses were run for 100 × 106 generations, saving one in every 10000 trees and associated parameters. The posterior distributions of the topologies and all parameters generated during the analyses were synthesized in an MCC tree in TreeAnnotator 2.4.5 (Drummond & Rambaut, 2007). The population geographic structure for each species or lineage was assessed through haplotype networks in Haploviewer (http://www.cibiv.at/~greg/haploviewer), based on the compound haplotypes (COI + 16S) of all specimens analysed (i.e. using the non-collapsed matrices), their frequencies and sample locations. To analyse the data with Haploviewer, positions with ambiguous nucleotides (like ‘N’) were eliminated because the program treats them as differences, and in most cases, these Ns were added to GenBank sequences as they were shorter than those obtained here. The mean genetic distances between groups, previously determined by the phylogenetic analyses, were calculated with DnaSP (Rozas & Rozas, 1995; Librado & Rozas, 2009), on the COI data, since this gene is the most studied and referenced for the analysis of divergence among related species, and is commonly used as a barcode (Hebert et al., 2003; Prié & Puillandre, 2014). The studied specimens were photographed to illustrate each species variability of shell phenotypes. Pictures of U. tumidiformis, U. gibbus, U. ravoisieri and U. durieui were not included because their shell variation has been illustrated previously in Reis & Araujo (2009), Araujo et al. (2009a) and Khalloufi et al. (2011). RESULTS The sequence data obtained from COI [657 base pairs (bp)] and 16S (498 bp, after alignment) (GenBank accession numbers from KX399859 to KX400569, with the exception of KX400076, KX400078 and KX400319; Supporting Information, Table S1) were concatenated into a unique matrix after verifying that no major incongruences exist between the two genes. For the phylogenetic analyses, we used the concatenated matrix that included only unique haplotypes (144 for COI and 118 for 16S) and the specimens that had information for both genes (resulting in a matrix with 164 individuals and 1155 characters). For the phylogeographic analyses, all specimens in common were included (368). The models that best fit the data were 012232 + I + G + F for COI, TIM1 + I + G for 16S and 010012 + I + G + F for the concatenated matrix (under Bayesian information criterion). Species boundaries To delimit the number of species in the bGMYC analysis, we first defined a probability threshold above which the haplotypes would represent individuals from different species (Yule process). To avoid over-splitting or over-lumping samples, we adopted the posterior mean of the analysis (P = 0.5) as the threshold. Following this, 24 species were supported (Fig. 1 and see below). The M-PTP analyses, either using ML or Bayesian trees, resulted in a lower number of inferred species (ML approach: 19 species; Bayesian approach: 18 species) (Fig. 1). Figure 1. View largeDownload slide Bayesian tree reconstruction based on the two mitochondrial genes analysed. Values on the branches indicate Bayesian posterior probabilities, ML and MP bootstrap values. Results from the species delimitation analyses are also shown in this figure (green: M-PTP with a Bayesian tree; red: M-PTP with an ML tree; blue: bGMYC analysis). Names for the currently recognized morphospecies are also indicated in the phylogenetic tree. Figure 1. View largeDownload slide Bayesian tree reconstruction based on the two mitochondrial genes analysed. Values on the branches indicate Bayesian posterior probabilities, ML and MP bootstrap values. Results from the species delimitation analyses are also shown in this figure (green: M-PTP with a Bayesian tree; red: M-PTP with an ML tree; blue: bGMYC analysis). Names for the currently recognized morphospecies are also indicated in the phylogenetic tree. Eight of the fourteen currently recognized morphospecies analysed were consistently recovered in both the bGMYC and M-PTP analyses (U. delphinus, U. douglasiae, U. durieui, U. elongatulus, U. foucauldianus, U. mancus, U. tumidiformis and U. tumidus). Disagreement resulted in the remaining six species: (i) Unio bruguierianus and Unio ionicus: previously considered as subspecies of U. crassus by Haas (1969), they were consistently recovered as independent units in all analyses. (ii) Unio crassus: M-PTP analyses split this taxon in two, one including samples from Greece, Turkey and one haplotype from central Europe (sequence from GenBank), and the other grouping samples from Slovakia, France and Germany. The bGMYC analysis further divided the ‘crassus’ group in up to six distinct lineages. (iii) Unio gibbus: bGMYC and M-PTP analyses split this taxon into two, one including the samples from Morocco and Spain and the other including the samples from Tunisia. (iv) Unio pictorum: all analyses consistently split U. pictorum into two lineages, one including only the haplotypes from Thrichomida Lake and the Axios River in Greece and the other containing the remaining European samples and the Greek haplotypes from Volvi Lake. (v) Unio ravoisieri: the M-PTP analysis with the ML tree and the bGMYC model split this taxon into two groups with no distinctive geographic distribution. (vi) Unio tigridis: the bGMYC model identified the two haplotypes analysed (Israel and Turkey) as different species. Phylogenetic analyses, geographic distribution and morphological characterisation Unio was consistently recovered as a monophyletic group in all analyses (bootstrap values were 83% for ML, 90% for MP and 1.00 posterior probability for BI), with U. douglasiae as the sister group of all other species (Fig. 1). The Unio clade from the Western Palaearctic had a clear structure with four robust and well-supported lineages that included the following morphospecies (Fig. 1): (1) U. gibbus + U. durieui; (2) U. tumidus; (3) (U. foucauldianus + U. delphinus) + U. tigridis, U. ravoisieri, U. mancus, U. elongatulus, U. pictorum; and (4) U. tumidiformis + [(U. bruguierianus + (U. ionicus + U. crassus)]. Clades 3 and 4 formed a well-supported monophyletic group, but the relationships between the other clades [1, 2 and (3 + 4)] were not clearly resolved. The U. durieui + U. gibbus lineage This clade grouped the samples from the morphospecies U. durieui and U. gibbus as sister lineages (Fig. 1). The U. durieui lineage only included samples from Tunisia and consisted of five different haplotypes among the specimens from the three rivers sampled. Distinct haplotypes were present in each river, but the most frequent haplotype was shared among specimens from the Ziatine and El Maaden rivers (Supporting Information, Fig. S1). Morphologically, U. durieui specimens were elongated with an intermediate shape between U. mancus and U. pictorum, and probably impossible to distinguish in the absence of geographical information. There were also rounded morphotypes similar to U. gibbus, as in the case of El Maaden (Khalloufi et al., 2011: fig. 5). This species was not found in Morocco but is present in Tunisia and Algeria to the east of the Moulouya River (Fig. 2). Figure 2. View largeDownload slide Schematic distribution of the Unio species in the Western Palaearctic. Points indicate the general vicinity of sampled localities. See Supporting Information, Table S1 for details. Figure 2. View largeDownload slide Schematic distribution of the Unio species in the Western Palaearctic. Points indicate the general vicinity of sampled localities. See Supporting Information, Table S1 for details. For the morphospecies U. gibbus, we obtained two clearly separated groups (genetic distance of 2.03%, Table 1): one representing haplotypes from the Iberian Peninsula (Barbate River) and Morocco, and another consisting of Tunisian haplotypes (Fig. 2; Supporting Information, Fig. S2). These two subclades were consistently recovered as distinct species in the bGMYC and M-PTP analyses. From 16 to 22 substitutions separated the Spanish-Moroccan haplotypes from the Tunisian ones. Although the eastern haplotypes were more similar, none of the Spanish, Moroccan and Tunisian specimens shared haplotypes among the 10 detected. Table 1. COI genetic distances between different lineages   1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  1. Outgroup  15.20                                                  2. U. douglasiae  15.30  –                                                3. U. crassus 1  15.44  14.56  0.20                                              4. U. crassus 2  15.42  13.85  2.13  0                                            5. U. crassus 3  15.12  13.85  3.10  2.34  0.09                                          6. U. crassus 4  15.50  14.46  0.96  2.18  2.85  0.27                                        7. U. crassus 5  15.85  14.92  2.60  3.28  4.10  2.22  0.51                                      8. U. crassus 6  15.65  14.80  2.69  2.60  3.64  2.33  2.89  0                                    9. U. ionicus  15.15  14.31  4.24  4.28  4.30  3.99  4.07  3.93  0.04                                  10. U. bruguierianus  15.9  14.12  3.87  4.22  4.21  3.82  4.45  3.86  4.28  0.08                                11. U. mancus  16.22  14.18  10.33  9.72  9.82  10.10  10.16  10.54  9.44  9.47  0.49                              12. U. pictorum 1  17.19  14.45  11.17  10.52  10.59  10.68  10.15  10.46  10.49  10.16  3.85  0.38                            13. U. pictorum 2  16.88  14.61  11.26  10.35  10.77  11.01  10.12  10.57  10.63  10.62  4.50  2.38  1.42                          14. U. elongatulus  15.94  15.16  9.80  8.98  9.21  9.30  9.01  9.35  9.43  9.25  3.89  3.26  4.46  0.28                        15. U. delphinus  13.88  12.90  11.32  9.80  9.70  10.81  9.75  9.79  9.04  8.53  3.81  3.66  4.49  4.29  0.50                      16. U. foucauldianus  15.92  14.30  10.84  10.21  9.91  10.38  9.75  9.63  9.90  9.87  4.10  3.94  4.52  3.71  3.32  0.20                    17. U. ravoisieri 1  16.26  14.79  10.65  9.73  10.45  10.50  9.50  9.86  10.02  10.30  5.01  4.36  4.75  4.46  4.72  4.30  0.20                  18. U. ravoisieri 2  16.25  14.70  10.40  9.79  10.03  10.24  8.97  9.28  10.10  10.07  5.82  4.38  4.96  4.81  4.21  5.20  2.55  0.26                19. U. tigridis 1  16.8  14.76  11.11  10.81  11.20  10.96  10.72  10.73  10.33  11.07  5.24  4.86  6.09  5.35  4.79  5.98  6.84  7.56  xx              20. U. tigridis 2  16.65  14.76  10.25  9.59  9.98  10.05  10.05  10.08  9.72  10.46  4.86  4.55  5.43  4.99  5.36  5.22  5.77  6.19  2.89  0            21. U. tumidiformis  15.74  13.26  8.31  7.99  7.85  8.86  8.51  7.95  9.30  8.22  11.54  11.63  11.90  11.95  10.72  10.98  11.18  10.92  12.83  12.86  0.50          22. U. tumidus  13.91  12.18  12.16  11.87  11.58  12.10  12.41  11.52  11.87  11.15  10.80  11.14  12.02  10.80  9.50  10.76  11.46  11.95  12.04  12.04  12.77  0.18        23. U. gibbus 1  14.24  11.13  11.89  11.65  11.25  11.82  11.80  11.40  11.13  10.93  12.35  12.86  13.33  13.22  12.12  13.12  13.77  14.14  13.69  13.90  13.02  10.19  0.37      24. U. gibbus 2  14.4  11.24  11.60  11.24  10.85  11.24  11.11  10.87  10.63  10.59  12.51  13.28  13.73  13.26  11.71  12.90  13.65  14.03  13.83  13.79  12.48  10.51  2.03  0.46    25. U. durieui  14.98  11.70  11.60  11.45  11.06  11.25  11.06  10.92  10.99  10.65  12.69  13.74  14.13  13.47  11.70  12.79  13.84  13.91  14.74  14.58  13.29  9.99  5.46  4.87  0.07    1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  1. Outgroup  15.20                                                  2. U. douglasiae  15.30  –                                                3. U. crassus 1  15.44  14.56  0.20                                              4. U. crassus 2  15.42  13.85  2.13  0                                            5. U. crassus 3  15.12  13.85  3.10  2.34  0.09                                          6. U. crassus 4  15.50  14.46  0.96  2.18  2.85  0.27                                        7. U. crassus 5  15.85  14.92  2.60  3.28  4.10  2.22  0.51                                      8. U. crassus 6  15.65  14.80  2.69  2.60  3.64  2.33  2.89  0                                    9. U. ionicus  15.15  14.31  4.24  4.28  4.30  3.99  4.07  3.93  0.04                                  10. U. bruguierianus  15.9  14.12  3.87  4.22  4.21  3.82  4.45  3.86  4.28  0.08                                11. U. mancus  16.22  14.18  10.33  9.72  9.82  10.10  10.16  10.54  9.44  9.47  0.49                              12. U. pictorum 1  17.19  14.45  11.17  10.52  10.59  10.68  10.15  10.46  10.49  10.16  3.85  0.38                            13. U. pictorum 2  16.88  14.61  11.26  10.35  10.77  11.01  10.12  10.57  10.63  10.62  4.50  2.38  1.42                          14. U. elongatulus  15.94  15.16  9.80  8.98  9.21  9.30  9.01  9.35  9.43  9.25  3.89  3.26  4.46  0.28                        15. U. delphinus  13.88  12.90  11.32  9.80  9.70  10.81  9.75  9.79  9.04  8.53  3.81  3.66  4.49  4.29  0.50                      16. U. foucauldianus  15.92  14.30  10.84  10.21  9.91  10.38  9.75  9.63  9.90  9.87  4.10  3.94  4.52  3.71  3.32  0.20                    17. U. ravoisieri 1  16.26  14.79  10.65  9.73  10.45  10.50  9.50  9.86  10.02  10.30  5.01  4.36  4.75  4.46  4.72  4.30  0.20                  18. U. ravoisieri 2  16.25  14.70  10.40  9.79  10.03  10.24  8.97  9.28  10.10  10.07  5.82  4.38  4.96  4.81  4.21  5.20  2.55  0.26                19. U. tigridis 1  16.8  14.76  11.11  10.81  11.20  10.96  10.72  10.73  10.33  11.07  5.24  4.86  6.09  5.35  4.79  5.98  6.84  7.56  xx              20. U. tigridis 2  16.65  14.76  10.25  9.59  9.98  10.05  10.05  10.08  9.72  10.46  4.86  4.55  5.43  4.99  5.36  5.22  5.77  6.19  2.89  0            21. U. tumidiformis  15.74  13.26  8.31  7.99  7.85  8.86  8.51  7.95  9.30  8.22  11.54  11.63  11.90  11.95  10.72  10.98  11.18  10.92  12.83  12.86  0.50          22. U. tumidus  13.91  12.18  12.16  11.87  11.58  12.10  12.41  11.52  11.87  11.15  10.80  11.14  12.02  10.80  9.50  10.76  11.46  11.95  12.04  12.04  12.77  0.18        23. U. gibbus 1  14.24  11.13  11.89  11.65  11.25  11.82  11.80  11.40  11.13  10.93  12.35  12.86  13.33  13.22  12.12  13.12  13.77  14.14  13.69  13.90  13.02  10.19  0.37      24. U. gibbus 2  14.4  11.24  11.60  11.24  10.85  11.24  11.11  10.87  10.63  10.59  12.51  13.28  13.73  13.26  11.71  12.90  13.65  14.03  13.83  13.79  12.48  10.51  2.03  0.46    25. U. durieui  14.98  11.70  11.60  11.45  11.06  11.25  11.06  10.92  10.99  10.65  12.69  13.74  14.13  13.47  11.70  12.79  13.84  13.91  14.74  14.58  13.29  9.99  5.46  4.87  0.07  The intraspecific divergences within species divided into different units were Unio crassus: 2.10, Unio pictorum: 0.72, Unio ravoisieri: 1.36, Unio tigridis: 0.96 and Unio gibbus: 1.18%. View Large This species had a rounded shape similar to Potomida, but clear diagnostic characters such as the lateral teeth, the umbonal sculpture and shell thickness differentiate the two taxa (Araujo et al., 2009a: fig. 2). The U. tumidus lineage The U. tumidus lineage included samples from European rivers (namely, the Seine, Danube, Rhine, Thames and Weser), a Swedish lake, the Izorka and Kovash rivers in Russia and the Dnieper River in Ukraine (Fig. 2). This species is absent from the Iberian Peninsula and North Africa. Samples from the Rhine and Thames rivers and Bjornsjon Lake (Sweden) shared the same haplotype (Supporting Information, Fig. S3). The species had an elongated shell with an anterior umbo, but a characteristic cuneiform posterior end with a descending dorsal posterior margin, which clearly distinguishes this taxon from the other species. Nevertheless, there were specimens with short shells that were very difficult to identify (Fig. 3). We did not have access to the Ukrainian and Russian shells. Figure 3. View largeDownload slide Differing shell shapes of Unio tumidus. A, Franconian Saale, a tributary of the Main River (Rhine), Germany. B, Fulda River (Weser), Germany. C, Okna River (Danube), Slovakia. D, Danube River, Slovakia. E, Ferma Lake (Rhine), Germany. F, Thames River, UK. G, Fulda River (Weser), Germany. H, Rhine River, Germany. I, Horloff River (Rhine), Germany. Scale bar 2 cm. Figure 3. View largeDownload slide Differing shell shapes of Unio tumidus. A, Franconian Saale, a tributary of the Main River (Rhine), Germany. B, Fulda River (Weser), Germany. C, Okna River (Danube), Slovakia. D, Danube River, Slovakia. E, Ferma Lake (Rhine), Germany. F, Thames River, UK. G, Fulda River (Weser), Germany. H, Rhine River, Germany. I, Horloff River (Rhine), Germany. Scale bar 2 cm. The (U. foucauldianus + U. delphinus) + U. tigridis, U. ravoisieri, U. mancus, U. elongatulus, U. pictorum lineage This lineage was the most complex within the genus (Fig. 1). All morphospecies were recovered as monophyletic groups, although some had high levels of within-clade diversity and phylogenetic structuring. The relationships among lineages, however, were not well supported. The first well-supported clade of this lineage included two closely related morphospecies with a genetic distance of 3.32% (Fig. 1; Table 1): U. delphinus from Atlantic Iberia and U. foucauldianus from Atlantic and Mediterranean Morocco (Fig. 2). The network analysis showed a minimum of 24 steps separating the two species (Supporting Information, Fig. S4). There was greater Atlantic/South differentiation in U. delphinus (no shared haplotypes between the 2 regions and up to 24 substitutions among the most differentiated haplotypes) than in U. foucauldianus (in this case, the main haplotype was found in both the Atlantic and Mediterranean basins). Many of the U. delphinus and all of the U. foucauldianus specimens from Mediterranean Morocco had an elongated shell with a pictorum shape, while specimens from the Atlantic rivers of Morocco and southern Spain were more rounded, resembling the mancus shape (Figs 4, 5). Figure 4. View largeDownload slide Differing shell shapes of Unio foucauldianus. A, Loukos River. B, Oum Er Rbia River. C, Molouya River. D, Mda River. E, Martil River. F, Beth River (Sebou). G, Loukos River. Scale bar 2 cm. Figure 4. View largeDownload slide Differing shell shapes of Unio foucauldianus. A, Loukos River. B, Oum Er Rbia River. C, Molouya River. D, Mda River. E, Martil River. F, Beth River (Sebou). G, Loukos River. Scale bar 2 cm. Figure 5. View largeDownload slide Differing shell shapes of Unio delphinus. A, Guadalmez River (Guadiana). B, Landrinos River (Tagus). C, Ulla River. D, Hozgarganta River. E, Barbate River. F, Deza River (Ulla). G, Guadalporcún River (Guadalete). Scale bar 2 cm. Figure 5. View largeDownload slide Differing shell shapes of Unio delphinus. A, Guadalmez River (Guadiana). B, Landrinos River (Tagus). C, Ulla River. D, Hozgarganta River. E, Barbate River. F, Deza River (Ulla). G, Guadalporcún River (Guadalete). Scale bar 2 cm. The second clade comprised U. tigridis samples from Israel (Lake Kinneret in the Jordan River basin) and Turkey (the Tersakan stream) (Fig. 2). The M-PTP analysis considered these samples as two different species (Fig. 1), with a genetic distance of 2.89% (Table 1). Only six specimens were analysed from this area, and the two haplotypes detected were separated by 32 substitutions (Supporting Information, Fig. S5). This species is typified by a short shell with anteriorly shifted umbos, which was observed in the Israeli specimens. The Turkish specimens resembled the U. mancus morph (Fig. 6). Figure 6. View largeDownload slide Differing shell shapes of Unio tigridis. A, Lake Kinneret, Israel. B, Tersakan River, Southwest Turkey. Figure 6. View largeDownload slide Differing shell shapes of Unio tigridis. A, Lake Kinneret, Israel. B, Tersakan River, Southwest Turkey. The next lineage combined the U. ravoisieri specimens into two subclades with a genetic distance of 2.55% (Table 1; Fig. 2). These subclades were also consistently recovered as distinct species in the bGMYC and M-PTP analyses. The two most frequent haplotypes (Supporting Information, Fig. S6) were separated by 18 steps. One of these haplotypes represented the Spanish samples, and was very close to a haplotype present in the Tunisian Kebir River (separated by only one step), and to one of the two divergent El Maaden haplotypes (separated by four steps). The shell of this species typically has a modified U. mancus shape, but the morphs we obtained were highly variable, especially those from the Spanish Ser River (Khalloufi et al., 2011: fig. 3). The clade consisting of U. mancus (Fig. 1) included specimens from Mediterranean basins in Spain (the Ebro, Fluviá, Ter, Sonella, Llobregat and Júcar rivers), France (the Hérault and Argens rivers, the Tet River in the Basse Basin, Lake Bourget and the Ognon River, both within the Rhône Basin, many rivers in Corsica and the Atlantic rivers Charente, Seine and Drée, a tributary of the Loire) and Italy (many rivers in Sardinia) (Fig. 2). The network analysis showed haplotypes with some biogeographic incongruences (Supporting Information, Fig. S7). For instance, the Júcar and Sonella rivers (eastern Spain) shared a haplotype with some French populations, but not with any of the northeastern Spanish populations. This high-frequency haplotype appeared in both Atlantic and Mediterranean French and Spanish rivers. Of the 15 different haplotypes found, five were exclusive to Sardinian-Corsican populations, one to the French Lake Bourget and one appeared with a high frequency in the Catalonian (eastern Spain) rivers. This species had a more rounded shell than the other Mediterranean species U. elongatulus, with the exception of specimens from northeastern Spain, Corsica and Sardinia, which were elongated (Figs 7, 8). Figure 7. View largeDownload slide Differing shell shapes of Unio mancus. A, Stabiacciu River, Corsica. B, Liscia River, Sardinia. C, Cedrino River, Corsica. D, River at Banyoles Lake, Spain. E, F, Araxisi River, Sardinia. Scale bar 2 cm. Figure 7. View largeDownload slide Differing shell shapes of Unio mancus. A, Stabiacciu River, Corsica. B, Liscia River, Sardinia. C, Cedrino River, Corsica. D, River at Banyoles Lake, Spain. E, F, Araxisi River, Sardinia. Scale bar 2 cm. Figure 8. View largeDownload slide Differing shell shapes of Unio mancus. A, Brugent River (Ter), Spain. B, Bourget Lake (Rhône), France. C, Drée River (Loire), France. D, Golo River, Corsica. E, Ebro River, Spain. F, Stabiacciu River, Corsica. G, Orbu River, Corsica. Scale bar 2 cm. Figure 8. View largeDownload slide Differing shell shapes of Unio mancus. A, Brugent River (Ter), Spain. B, Bourget Lake (Rhône), France. C, Drée River (Loire), France. D, Golo River, Corsica. E, Ebro River, Spain. F, Stabiacciu River, Corsica. G, Orbu River, Corsica. Scale bar 2 cm. The next clade included the U. elongatulus (Fig. 1) specimens from Mediterranean freshwaters in Italy (Po and Isonzo rivers and Alpine lakes), Croatia (Mirna and Zrmanja rivers, Bačinska Lake) and Albania (Scutari Lake) (Fig. 2). The Croatian haplotypes appeared at both ends of the network (Supporting Information, Fig. S8), therefore, shared haplotypes between Croatia and Italy would be expected. However, no haplotypes were shared among any of the specimens from these regions. The specimens had more elongated shells than other Mediterranean species (U. mancus) and a posterior umbo, although some populations had more rounded shells (Fig. 9). Figure 9. View largeDownload slide Differing shell shapes of Unio elongatulus. A, Po di Tolle River, Italy. B, Lake Candia, Italy. C, Venice, Italy. D, Lake Cestella, Italy. E, Lake Bačinska, Croatia. F, G, Mirna River, Croatia. H, Zrmanja River, Croatia. I, Lake Scutari, Albania. Scale bar 2 cm. Figure 9. View largeDownload slide Differing shell shapes of Unio elongatulus. A, Po di Tolle River, Italy. B, Lake Candia, Italy. C, Venice, Italy. D, Lake Cestella, Italy. E, Lake Bačinska, Croatia. F, G, Mirna River, Croatia. H, Zrmanja River, Croatia. I, Lake Scutari, Albania. Scale bar 2 cm. The last clade represented the species U. pictorum, which included two groups having a genetic distance of 2.38% (Table 1). The first group (1) consisted of specimens from Western and Central Europe (Lake Bourget in the Rhône Basin and the Adour, Seine, Oder, Danube and Thames basins), eastern Greece (Strymonas and Axios rivers and Lake Volvi), Ukraine (Teteriv at the Black Sea), Iran and Russia (Dzhankhot River at the Black Sea, Vyborg River in Karelia). The second group (2) comprised specimens from both western (Trichonida Lake at the Acheloos basin) and eastern (Axios River) Greece (Fig. 2). These two subclades were consistently recovered as distinct species in the bGMYC and M-PTP analyses. Up to 16 different haplotypes (Supporting Information, Fig. S9), with from one to a maximum of 23 substitutions, were found among the specimens of this species. There was a maximum of 8 steps of differentiation within group ‘1’ and 17 between two of the three ‘2’ haplotypes. The network shape of group ‘1’ represented polymorphic populations with two relatively frequent haplotypes, one of which was only found in specimens from the Thames River and the other shared by samples as far away as Iran and Greece. Haplotypes found in the Danube and Dzhankhot rivers and in Lake Volvi seemed to derive from the latter one. This highly polymorphic species had a typical elongated shell. However, some specimens were very different, for example those from Strymonas River, Lake Volvi and the Axios River (within both clades). Some populations resembled U. mancus (Lake Volvi, Thichonida and Strymonas rivers) (Figs 10, 11). Figure 10. View largeDownload slide Differing shell shapes of Unio pictorum. A, B, Lake Trichonida, Greece. C, Thames River, UK. D, Axios River, Greece. E, Okna River (Danube), Slovakia. F, Axios River, Greece. Scale bar 2 cm. Figure 10. View largeDownload slide Differing shell shapes of Unio pictorum. A, B, Lake Trichonida, Greece. C, Thames River, UK. D, Axios River, Greece. E, Okna River (Danube), Slovakia. F, Axios River, Greece. Scale bar 2 cm. Figure 11. View largeDownload slide Differing shell shapes of Unio pictorum. A, B, Lake Volvi, Greece. C–F, Strymonas River, Greece. Scale bar 2 cm. Figure 11. View largeDownload slide Differing shell shapes of Unio pictorum. A, B, Lake Volvi, Greece. C–F, Strymonas River, Greece. Scale bar 2 cm. The U. tumidiformis + [U. bruguierianus + (U. ionicus + U. crassus)] lineage This lineage included two monophyletic groups with a COI genetic distance greater than 7.85%: a western group that included the endemic Iberian species U. tumidiformis (Fig. 2) and a second group (supported by a posterior probability of 0.99 and bootstrap values of 79 for ML and 94 for MP) comprising three morphospecies with a complex geographic structure (Fig. 1; Table 1). Up to nine haplotypes were found among the samples analysed for U. tumidiformis (Supporting Information, Fig. S10). Two were exclusive to the Portuguese Sado River but were close to some Spanish haplotypes. The most frequent haplotype was found in the Mira and Guadiana rivers. Within the second group, U. bruguierianus from eastern Greece (the Axios and Pinios rivers) (Fig. 2) was the sister species of U. ionicus + U. crassus (genetic distance of 4.28–4.45%), although this relationship was not highly supported. Unio ionicus lives in Albania and the Acheloos River (including Lake Lysimacheia) in western Greece (Fig. 2) and was the sister group of U. crassus (genetic distance of 4.30%). Unio crassus presented five subclades with a minimum divergence of 2.16% (Table 1) and haplotypes separated by at least 19 steps (Fig. 12). These five subclades included (1) U. crassus courtillierii Hattemann, 1859 from France and one sample from Sweden (both sequences obtained from GenBank); (2) samples from the Sofaditikos (Pinios catchment), Aliakmon/Aliakmonas and Sperchios rivers in eastern Greece; (3) Central European samples from the Rhine, Danube, and Rhône catchments; (4) samples from the Lissos River (eastern Greece); and (5) samples from western Turkey (Fig. 2). This group had the most complexity, which is also reflected in the bGMYC and M-PTP analyses in which up to six species were revealed. Figure 12. View largeDownload slide Network of Unio crassus haplotypes. A, U. crassus courtillierii and Sweden. B, eastern Greece (Sofaditikos, Aliakmon and Sperchios). C, Central European (Rhine, Danube and Rhône). D, eastern Greece (Lissos River). E, western Turkey. Figure 12. View largeDownload slide Network of Unio crassus haplotypes. A, U. crassus courtillierii and Sweden. B, eastern Greece (Sofaditikos, Aliakmon and Sperchios). C, Central European (Rhine, Danube and Rhône). D, eastern Greece (Lissos River). E, western Turkey. All U. crassus haplotypes were represented in a unique network (Fig. 12). No clear biogeographic structure was observed, and some haplotypes belonging to specimens from the same rivers had a high number of substitutions (up to 33). From the right end of Figure 12, three haplotypes from eastern Greece were assigned to U. bruguierianus. Three distinct haplotypes from western Greece and Albania were considered to represent U. ionicus. Three haplotypes were found from the Turkish locations analysed, but with a smaller genetic distance compared with the other haplotypes. A polymorphic shell shape was observed for U. bruguierianus: some specimens from the Axios and Pinios rivers had a shell shape resembling some pictorum specimens from the Axios River, while other specimens from the Pinios River were identical to some U. crassus specimens from the Lissos River or Turkey (Fig. 13). This polymorphic shell shape was also observed for U. ionicus and U. crassus (Figs 13, 14). Unio ionicus had an almost identical shell to some U. crassus specimens from the Sofaditikos (Pinios Basin), Sperchios and Lissos rivers in Greece. In contrast, other U. crassus specimens from the Lissos River and from Turkey presented an elongated shell shape. There were also specimens with and without sculpture in the umbo area from the same locality with the same haplotype, such as those from the Lissos River (Fig. 14). Figure 13. View largeDownload slide A–C, Unio bruguierianus. A, Pinios River, Greece. B, Axios River, Greece. C, Pinios River, Greece. D–E, Unio ionicus. D, River at Lake Lysimacheia, Greece. E, Perroi i Bistrices, Albania. F–H, Unio crassus. F, Sperchios River, Greece. G, H, Aliakmonas River, Greece. Scale bar 2 cm. Figure 13. View largeDownload slide A–C, Unio bruguierianus. A, Pinios River, Greece. B, Axios River, Greece. C, Pinios River, Greece. D–E, Unio ionicus. D, River at Lake Lysimacheia, Greece. E, Perroi i Bistrices, Albania. F–H, Unio crassus. F, Sperchios River, Greece. G, H, Aliakmonas River, Greece. Scale bar 2 cm. Figure 14. View largeDownload slide Differing shell shapes of Unio crassus. A, B, Sofaditikos River (Pinios), Greece. C, Matzenheim, France. D, Çine Çayi, Mugla, Turkey. E, F, Lissos River, Greece. G, Limagne, France. Scale bar 2 cm. Figure 14. View largeDownload slide Differing shell shapes of Unio crassus. A, B, Sofaditikos River (Pinios), Greece. C, Matzenheim, France. D, Çine Çayi, Mugla, Turkey. E, F, Lissos River, Greece. G, Limagne, France. Scale bar 2 cm. The coalescence-based molecular clock analyses for the species tree resulted in very broad temporal intervals for most of the clades. Although the results have to be interpreted with caution, they provide a broad temporal framework for the diversification of the group. The coalescence reconstruction (Fig. 15) placed the common ancestor of the Unio clade and the origin of the Western Palaearctic Unio in the Eocene. In fact, the Western Palaearctic species underwent two main cladogenetic events. The first event occurred in the Early Eocene (U. gibbus, U. durieui split), and the second during the Oligocene, involving the divergence of the most speciose clade. Most of the modern species appeared during the Miocene. The most recent cladogenetic event involved the U. delphinus and U. foucauldianus clade at the end of the Miocene (Messinian). Notably, the reconstructed species tree agrees with the general topology of the concatenated matrix analyses except in the phylogenetic position of U. tumidus. In the species tree, U. tumidus appeared with strong support as sister group of the U. tigridis + [U. pictorum, U. ravoisieri (U. mancus + U. elongatulus, U. delphinus + U. foucauldianus)] lineage (Fig. 15). Figure 15. View largeDownload slide Coalescence-based species tree generated in BEAST. The x-axis scale is in millions of years. Bars indicate 95% high probability density intervals. Asterisks (*) in the tree indicate posterior probabilities pp > 0.9. Figure 15. View largeDownload slide Coalescence-based species tree generated in BEAST. The x-axis scale is in millions of years. Bars indicate 95% high probability density intervals. Asterisks (*) in the tree indicate posterior probabilities pp > 0.9. DISCUSSION In his classic study, Haas (1969) divided the Western Palaearctic lineages of Unio into 42 different taxa (Table 2), including 12 ‘fundamental species’ with many subspecies or races. In recent years, molecular markers have been instrumental in unravelling both the taxonomy and the phylogeny of this group (Araujo et al., 2005; Araujo et al., 2009a; Khalloufi et al., 2011; Prié et al., 2012; Reis et al., 2013; Prié & Puillandre, 2014; Froufe et al., 2016a). To further advance our understanding of species limits within this group and their evolutionary history and relationships, we have analysed the patterns of genetic variability in two mitochondrial genes in 518 samples of Unio from North Africa (Morocco, Tunisia), Europe (from Portugal to Greece) and the Middle East (Turkey and Israel). Table 2. Information on the Western Palaearctic Unio taxonomy Haas (1969)   Author  Accepted species  References  Unio tumidus tumidus  Retzius, 1788  Unio tumidus  This paper  Unio tumidus borysthenensis  Kobelt, 1879  Unio tumidus  This paper  Unio pictorum pictorum  Linnaeus, 1758  Unio pictorum  This paper  Unio pictorum praeposterus  Küster, 1854  Unio pictorum  This paper  Unio pictorum latirostris  Küster, 1854  Unio pictorum  This paper  Unio pictorum platyrhynchus  Rossmässler, 1835  Unio pictorum  This paper  Unio pictorum schrenkianus  Clessin, 1880  Unio pictorum  This paper  Unio pictorum ascanius  Kobelt, 1913  Unio pictorum  This paper  Unio pictorum proechistus  Bourguignat, 1870  Unio pictorum  This paper  Unio pictorum gaudioni  Drouët, 1881  Unio pictorum  This paper  Unio pictorum rostratus  Lamarck, 1819  Unio pictorum  This paper  Unio pictorum platyrhynchoideus  Dupuy, 1849  Unio pictorum  This paper  Unio pictorum mucidus  Morelet, 1845  Unio delphinus  Araujo et al. (2009); Reis et al. (2013); this paper  Unio pictorum delphinus  Spengler, 1793  Unio delphinus  Araujo et al. (2009); Reis et al. (2013); this paper  Unio pictorum ravoisieri  Deshayes, 1848  Unio ravoisieri  Araujo et al. (2009); Khalloufi et al. (2011); this paper  Unio tigridis tigridis  Bourguignat, 1852  Unio tigridis  This paper  Unio tigridis terminalis  Bourguignat, 1852  Unio tigridis  This paper  Unio abyssinicus  Martens, 1866  ¿?    Unio elongatulus elongatulus  C. Pfeiffer, 1825  Unio elongatulus  Prié et al. (2012); this paper  Unio elongatulus pallens  Rossmässler, 1842  Unio elongatulus + Unio pictorum  This paper  Unio elongatulus bourgueticus  Bourguignat, 1882  Unio pictorum  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus turtoni  Payraudeau, 1826  Unio mancus  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus moquinianus  Dupuy, 1843  Unio mancus  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus aleroni  Companyo & Massot, 1845  Unio mancus  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus rousii  Dupuy, 1849  Unio pictorum  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus mancus  Lamarck, 1819  Unio pictorum  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus requienii  Michaud, 1831  Unio mancus + Unio pictorum  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus penchinatianus  Bourguignat, 1865  Unio ravoisieri  Khalloufi et al. (2011)   Unio elongatulus valentinus  Rossmässler, 1854  Unio mancus  Araujo et al. (2005); Araujo et al. (2009)   Unio elongatulus gargottae  Philippi, 1836  Unio mancus?  This paper  Unio elongatulus lawleyianus  Gentiluomo, 1868  Unio mancus + Unio pictorum?  This paper  Unio elongatulus glaucinus  Porro, 1838  Unio elongatulus + Unio mancus + Unio pictorum  This paper  Unio elongatulus eucirrus  Bourguignat, 1860  Unio tigridis + Unio pictorum  This paper  Unio elongatulus dembeae  Sowerby, 1865  ¿?    Unio elongatulus durieui  Deshayes, 1847  Unio durieui + Unio foucauldianus  Araujo et al. (2009); Khalloufi et al. (2011); Froufe et al. (2016a)   Unio crassus crassus  Retzius, 1788  Unio crassus  This paper  Unio crassus cytherea  Küster, 1833  Unio crassus  This paper  Unio crassus batavus  Maton & Racket, 1807  Unio crassus  This paper  Unio crassus carneus  Küster, 1854  Unio crassus  This paper  Unio crassus ionicus  Drouet, 1879  Unio ionicus  This paper  Unio crassus gontieri  Bourguignat, 1856  Unio crassus  This paper  Unio crassus bruguierianus  Bourguignat, 1853  Unio bruguierianus  This paper  Unio crassus mongolicus  Middendorff, 1851  Unio crassus  This paper  Haas (1969)   Author  Accepted species  References  Unio tumidus tumidus  Retzius, 1788  Unio tumidus  This paper  Unio tumidus borysthenensis  Kobelt, 1879  Unio tumidus  This paper  Unio pictorum pictorum  Linnaeus, 1758  Unio pictorum  This paper  Unio pictorum praeposterus  Küster, 1854  Unio pictorum  This paper  Unio pictorum latirostris  Küster, 1854  Unio pictorum  This paper  Unio pictorum platyrhynchus  Rossmässler, 1835  Unio pictorum  This paper  Unio pictorum schrenkianus  Clessin, 1880  Unio pictorum  This paper  Unio pictorum ascanius  Kobelt, 1913  Unio pictorum  This paper  Unio pictorum proechistus  Bourguignat, 1870  Unio pictorum  This paper  Unio pictorum gaudioni  Drouët, 1881  Unio pictorum  This paper  Unio pictorum rostratus  Lamarck, 1819  Unio pictorum  This paper  Unio pictorum platyrhynchoideus  Dupuy, 1849  Unio pictorum  This paper  Unio pictorum mucidus  Morelet, 1845  Unio delphinus  Araujo et al. (2009); Reis et al. (2013); this paper  Unio pictorum delphinus  Spengler, 1793  Unio delphinus  Araujo et al. (2009); Reis et al. (2013); this paper  Unio pictorum ravoisieri  Deshayes, 1848  Unio ravoisieri  Araujo et al. (2009); Khalloufi et al. (2011); this paper  Unio tigridis tigridis  Bourguignat, 1852  Unio tigridis  This paper  Unio tigridis terminalis  Bourguignat, 1852  Unio tigridis  This paper  Unio abyssinicus  Martens, 1866  ¿?    Unio elongatulus elongatulus  C. Pfeiffer, 1825  Unio elongatulus  Prié et al. (2012); this paper  Unio elongatulus pallens  Rossmässler, 1842  Unio elongatulus + Unio pictorum  This paper  Unio elongatulus bourgueticus  Bourguignat, 1882  Unio pictorum  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus turtoni  Payraudeau, 1826  Unio mancus  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus moquinianus  Dupuy, 1843  Unio mancus  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus aleroni  Companyo & Massot, 1845  Unio mancus  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus rousii  Dupuy, 1849  Unio pictorum  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus mancus  Lamarck, 1819  Unio pictorum  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus requienii  Michaud, 1831  Unio mancus + Unio pictorum  Prié et al. (2012); Prié & Puillandre (2014); this paper  Unio elongatulus penchinatianus  Bourguignat, 1865  Unio ravoisieri  Khalloufi et al. (2011)   Unio elongatulus valentinus  Rossmässler, 1854  Unio mancus  Araujo et al. (2005); Araujo et al. (2009)   Unio elongatulus gargottae  Philippi, 1836  Unio mancus?  This paper  Unio elongatulus lawleyianus  Gentiluomo, 1868  Unio mancus + Unio pictorum?  This paper  Unio elongatulus glaucinus  Porro, 1838  Unio elongatulus + Unio mancus + Unio pictorum  This paper  Unio elongatulus eucirrus  Bourguignat, 1860  Unio tigridis + Unio pictorum  This paper  Unio elongatulus dembeae  Sowerby, 1865  ¿?    Unio elongatulus durieui  Deshayes, 1847  Unio durieui + Unio foucauldianus  Araujo et al. (2009); Khalloufi et al. (2011); Froufe et al. (2016a)   Unio crassus crassus  Retzius, 1788  Unio crassus  This paper  Unio crassus cytherea  Küster, 1833  Unio crassus  This paper  Unio crassus batavus  Maton & Racket, 1807  Unio crassus  This paper  Unio crassus carneus  Küster, 1854  Unio crassus  This paper  Unio crassus ionicus  Drouet, 1879  Unio ionicus  This paper  Unio crassus gontieri  Bourguignat, 1856  Unio crassus  This paper  Unio crassus bruguierianus  Bourguignat, 1853  Unio bruguierianus  This paper  Unio crassus mongolicus  Middendorff, 1851  Unio crassus  This paper  View Large Using a coalescence-based approach to delimit species boundaries, we identified eight lineages in congruence with the current taxonomy of the group. Some of these species were formerly considered at the subspecific level (e.g. Haas, 1969); however, their taxonomic statuses were recently revised following recent molecular analyses (e.g. Araujo & de Jong, 2015; Lopes-Lima et al., 2017). In this study, we show that U. delphinus (Araujo et al., 2009b) and U. ravoisieri (Khalloufi et al., 2011), named as subspecies of U. pictorum (Haas, 1969), actually belong to different phylogenetic groups. We also propose the validity of the Mediterranean U. mancus from France, Sardinia (Italy) and Spain, considered by Haas (1969) as several U. elongatulus subspecies (Table 2) (Araujo et al., 2005; Araujo et al., 2009b; Prié, 2011; Prié et al., 2012; Prié & Puillandre, 2014). The presence of Atlantic populations of this species in France (Drée at the Loire, Garonne and Seine rivers) (Prié et al., 2012; Prié & Puillandre, 2014) may be explained by palaeohydrological events of stream capture or by river connections via the construction of old channels (during the 18th century) in France that joined Mediterranean and Atlantic rivers (Prié et al., 2012; Prié & Puillandre, 2014). The western species U. delphinus and U. foucauldianus were probably misidentified as U. pictorum delphinus and U. elongatulus durieui by Haas (1969) (Khalloufi et al., 2011; Table 2). Although these two species are separated by the smallest genetic distance, they are independent lineages, a finding also confirmed by microsatellite marker analyses (Froufe et al., 2016a), and are distributed on both sides of the Gibraltar Strait (Araujo et al., 2009b; Froufe et al., 2016a). In addition, we found U. foucauldianus in some Mediterranean rivers of Morocco (Martil, Laou and Moulouya), where it had not been cited previously. The GenBank numbers used in Froufe et al. (2016a) were not available when we analysed our data, but we sampled many of the same Atlantic Moroccan rivers. We also restored the species U. ionicus and U. bruguierianus, which were both included by Haas (1969) in the U. crassus clade. The populations from the Axios (= Vardar River) and Pinios rivers in eastern Greece belong to U. bruguierianus, although other populations of the Sofaditikos River in the Pinios Basin belong to U. crassus. The presence of these two species in Greek rivers in the same basin can be explained by river connections and the complex palaeogeography of this peninsula (Steininger & Rogl, 1984; Economidis & Miller, 1990). These processes were responsible for the great species radiation in Barbus (Karakousis et al., 1995; Gante, 2011), one of the main host fishes of Unio. Recently, Froufe et al. (2016b) and Araujo et al. (2016) demonstrated a similar case of endemism for Potomida acarnanica Kobelt, 1879 in the Balkans. Of particular interest, however, are the six main discordances found in our species delimitation analysis and the established taxonomy: Unio tigridis was divided into two subspecies by Haas (1969). The studied samples from Israel and western Turkey were also retrieved as two separate species in our analyses. However, until more samples are studied from the eastern Turkish rivers, Syria and Iraq, we propose maintaining the name U. tigridis. The lack of comprehensive geographic sampling emphasizes the need for in-depth study. This was noted by Falkner (1994) who referred to the ‘puzzling form chaos’ of U. tigridis and Unio terminalis in the Middle East. Interestingly, Froufe et al. (2016b) observed a similar distribution in Potomida semirugata Lamarck, 1819. Haas (1969) described 12 subspecies within U. pictorum (Table 2). With the exception of U. delphinus (Araujo et al., 2009b) and U. ravoisieri (Khalloufi et al., 2011), both considered as species in this study, the other 10 subspecies grouped into two main clades/species, including one that consists of only specimens from northeastern Greece (2.38% genetic distance from the other specimens of this clade). Unio pictorum is distributed from France to Ukraine, Russia and Iran, but not in North Africa, the Iberian Peninsula or, with the exception of Greece and some localities in the Rhône catchment, the Mediterranean rivers of Europe. The presence of this species in Greece can be explained by dispersion from the Danube, as in some Greek Barbus species (Karakousis et al., 1995). Nagel (2000) concluded that the present-day population structure of U. pictorum in central Europe was best explained by river connections during the Pliocene and Pleistocene. Whether the populations of northeastern Greece constitute a different species, as suggested in our study, deserves further studies. Similarly, the specific status of the two main lineages/species of U. gibbus revealed in our study (and in a previous study; Khalloufi et al., 2011), corresponding to the populations of South Spain and Morocco, and Tunisia, respectively, should be further explored. The close relationship between the southern Iberian and northwestern Maghreb species, compared with that among North African and Moroccan populations, or the Iberian with other European populations, provides further evidence of the presence of the Messinian Betic-Riffian Massif, which disappeared some 5.3 Mya (Krijgsman et al., 1999). This biogeographic pattern is similar to that shown by cyprinid fish, and other fish lineages, that were considered different species (Machordom & Doadrio, 2001). The other North African species, U. ravoisieri, included in the pictorum group by Haas (1969), had been redescribed and discussed by Khalloufi et al. (2011). Regarding their Spanish populations in Catalonia, Khalloufi et al. (2011) proposed two hypotheses to explain their current distribution patterns: some historical human transport by the Phoenicians, Romans or Almohads from Tunisian rivers, as has been hypothesized for other animal groups (see Recuero et al., 2007; Araujo et al., 2017), or, more doubtfully, a connection among freshwater courses between these areas during the Messinian crisis, which dried up the Mediterranean (5.5 Mya), leading to the subsequent extinction of connecting populations. These hypotheses suggest a possible foreign origin for the Spanish U. ravoisieri populations; however, no clear geographic pattern is found among the samples included in the two clades obtained in our study. Finally, the species with a larger distribution and more taxonomic problems was U. crassus. The species delimitation analyses revealed from two to six main lineages/species with no clear geographic structuring. However, the absence of morphological features and clear biogeographic patterns require a deeper survey including a greater number of locations and specimens. The phylogenetic and the calibrated species tree revealed a strong congruence between analyses. In general, we obtained very broad time intervals for all nodes of the phylogeny, which is expected given the lack of reliable fossils or well-characterized palaeobiogeographic events to optimally calibrate the molecular clocks and the phylogenetic uncertainties. Temporal estimates, therefore, should be considered with caution and only used to obtain a broad picture of the diversification patterns within the group. Some relevant aspects deserve further study: (1) In general, diversification is old among and between groups. Most of the specific diversity is of Miocene origin. Species are long lived and characterized by some morphological stasis. Furthermore, recent palaeogeographic and climatic events, such as Holocene glaciations, seem to have affected the population and demographic dynamics of the species but not diversification per se. This new temporal and systematic framework for the group may help to reinterpret the idea of some species of Unio forming a polytypic species (ring species or Rassenkreis), with continuous gene flow and thus, very recent or as yet incomplete separation of lineages (e.g. Nagel, Badino & Celebrano, 1998; Nagel & Badino, 2001). (2) The three species with broad European distribution ranges (U. crassus, U. pictorum and U. tumidus) are included in three well-differentiated clades in the phylogenetic analyses. Species with more restricted distribution, such as the ones present in the Iberian Peninsula and Maghreb (U. delphinus + U. foucauldianus, U. tumidiformis, U. ravoisieri, U. gibbus + U. durieui) were also well differentiated. They all colonized or differentiated in the Iberian Peninsula at very different times, from different origins and following very different palaeogeographic events. Unio gibbus, for instance, is probably of North African origin, while U. tumidiformis is related to European U. crassus, U ionicus and U. bruguierianus. Differences between U. crassus and the Iberian endemic U. tumidiformis were previously described by Reis & Araujo (2009), who suggested that U. tumidiformis may have evolved from a common ancestor in the Betic-Riffian Massif, which eventually contributed to the formation of the Guadiana and Guadalquivir basins (Vargas, Real & Guerrero, 1998). Using a relaxed molecular clock based on a mean COI rate substitution of 0.3% per million years, we estimate that the common ancestor of U. tumidiformis and the U. crassus complex existed about 15 Mya, during the Mid-Miocene. A formal comparative biogeographic study, however, would be necessary to fully understand the evolutionary history of the species in the region. (3) The phylogenetic position of U. tumidus is controversial in our study (Figs 1, 15). This clearly deserves further sampling and study as the evolutionary and biogeographic interpretations could drastically change depending on these results. Freshwater mussels are among the most endangered freshwater species worldwide. Providing robust phylogenetic and systematic hypotheses for the group is of paramount importance to design effective conservation and management plans, either at local or regional scales. The different approaches addressed in this study have helped establishing a new and robust systematic framework for Western Palaearctic Unio, which we hope translates into new conservation and management plans. While most of the unionid populations are in regression in the studied areas, new taxa identified here, and those with limited distribution, should be urgently protected and conserved, particularly in the case of U. bruguierianus, U. ionicus or U. elongatulus. Moreover, the potential existence of new species within currently recognized morphospecies (e.g. U. gibbus, U. tigridis or U. crassus) should prompt further detailed studies. Some of the new species could represent very localized endemics with very limited distribution areas and, hence, be highly endangered. SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article at the publisher’s web-site: Table S1. Information on the specimens analysed and GenBank accession numbers. *Misnamed in GenBank. Figure S1. Network of U. durieui haplotypes. Figure S2. Network of U. gibbus haplotypes. Figure S3. Network of U. tumidus haplotypes. Figure S4. Network of U. delphinus (Iberia) and U. foucauldianus (Morocco) haplotypes. Figure S5. Network of U. tigridis haplotypes. Figure S6. Network of U. ravoisieri haplotypes. Figure S7. Network of U. mancus haplotypes. Figure S8. Network of U. elongatulus haplotypes. Figure S9. Network of U. pictorum haplotypes. Figure S10. Network of U. tumidiformis haplotypes. ACKNOWLEDGEMENTS We thank Carlos Toledo for his work on some of the analyses. Joaquim Reis, José Miguel Barea, María José Madeira, Bülent Yorulmaz, Günter Hartz, Peter Reischütz, Joseph Heller, Mohamed Ghamizi and Noureddine Khalloufi helped us collect the specimens. Arthur Bogan sent us the Ukrainian and Russian specimens. The plates were compiled by Jesús Muñoz from the photography facility of the MNCN. This study was partially funded by the Spanish Ministry of Economy and Competitiveness (Ref. CTM2014-57949-R). Dimitrios Germanos provided us with the collection permits to sample in Greece, Pëllumb Abeshi in Albania and the Ministry of Culture, Nature Protection Directorate in Croatia. The help of Eladio Fernández Galiano (Council of Europe) was invaluable. Sarah Young and Melinda Modrell reviewed the English version of the manuscript. REFERENCES Araujo R, Buckley D, Nagel K-O, Machordom A. 2017. Potomida littoralis (Bivalvia, Unionidae) evolutionary history: slow evolution or recent speciation? Zoological Journal of the Linnean Society  179: 277– 290. doi: 10.1111/zoj.12470. Araujo R, de Jong Y. 2015. Fauna Europaea: Mollusca-Bivalvia. Biodiversity Data Journal  3: e5211. doi: 10.3897/BDJ.3.e5211. Google Scholar CrossRef Search ADS   Araujo R, Gómez I, Machordom A. 2005. The identity and biology of Unio mancus (= U. elongatulus) (Bivalvia: Unionidae) in the Iberian Peninsula. Journal of Molluscan Studies  71: 25– 31. Google Scholar CrossRef Search ADS   Araujo R, Reis J, Machordom A, Toledo C, Madeira MJ, Gómez I, Velasco JC, Morales J, Barea JM, Ondina P, Ayala I. 2009b. Las náyades de la península Ibérica. Iberus  27: 7– 72. Araujo R, Schneider S, Erpenbeck D, Roe KJ, Machordom A. 2017. The origin and phylogeny of Margaritiferidae (Bivalvia: Unionoida): a synthesis of fossil and molecular data. Zoologica Scripta  46: 289– 307. Google Scholar CrossRef Search ADS   Araujo R, Toledo C, Machordom A. 2009a. Redescription of Unio gibbus Spengler, 1793, a west Palaearctic freshwater mussel with hookless glochidia. Malacologia  51: 131– 141. Google Scholar CrossRef Search ADS   Ball GH. 1922. Variation in freshwater mussels. Ecology  3: 93– 121. Google Scholar CrossRef Search ADS   Baum DA. 1998. Individuality and the existence of species through time. Systematic Biology  47: 641– 653. Google Scholar CrossRef Search ADS PubMed  Bouckaert R, Heled J, Kühnert D, Vaughan T, Wu CH, Xie D, Suchard MA, Rambaut A, Drummond AJ. 2014. BEAST 2: a software platform for Bayesian evolutionary analyses. PLoS Biology  10: e1003537. Google Scholar CrossRef Search ADS   Carstens BC, Pelletier TA, Reid NM, Satdler JD. 2013. How to fail at species delimitation. Molecular Ecology  22: 4369– 4383. Google Scholar CrossRef Search ADS PubMed  Choi SC. 2016. Methods for delimiting species via population genetics and phylogenetics using genotype data. Genes and Genomics  38: 905– 915. Google Scholar CrossRef Search ADS   Coyne JA, Orr HA, Futuyma DJ. 1988. Do we need a new species concept? Systematic Zoology  37: 190– 200. Google Scholar CrossRef Search ADS   Cuttelod A, Seddon M, Neubert E. 2011. European red list of non-marine Molluscs . Luxembourg: Publications Office of the European Union. Drummond AJ, Rambaut A. 2007. BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evolutionary Biology  7: 214. Google Scholar CrossRef Search ADS PubMed  Drummond AJ, Suchard MA, Xie D, Rambaut A. 2012. Bayesian phylogenetics with BEAUti and the BEAST 1.7. Molecular Biology and Evolution  29: 1969– 1973. Google Scholar CrossRef Search ADS PubMed  Eagar RMC. 1978. Shape and function of the shell: a comparison on some living and fossil bivalve molluscs. Biological Reviews  53: 169– 210. Google Scholar CrossRef Search ADS   Economidis P, Miller P. 1990. Systematics of freshwater gobies from Greece (Teleostei; Gobiidae). Journal of Zoology  221: 125– 170. Google Scholar CrossRef Search ADS   Ereshefsky M. 2014. Species, historicity, and path dependency. Philosophy of Science  81: 714– 726. Google Scholar CrossRef Search ADS   Falkner G. 1994. Systematik vorderorientalischer Najaden als Vorstudie zur Bearbeitung archäologischer Funde. Forschungen und Berichte zur Vor- und Frühgeschichte in Baden-Württemberg  53: 135– 162. Felsenstein J. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution  39: 783– 791. Google Scholar CrossRef Search ADS PubMed  Folmer O, Black M, Hoeh W, Lutz R, Vrijenhoek R. 1994. DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Molecular Marine Biology and Biotechnology  3: 294– 299. Google Scholar PubMed  Froufe E, Gonçalves DV, Teixeira A, Sousa R, Varandas S, Ghamizi M, Zieritz A, Lopes-Lima M. 2016a. Who lives where? Molecular and morphometric analyses clarify which Unio species (Unionida, Mollusca) inhabit the southwestern Palearctic. Organisms Diversity & Evolution  16: 597– 611. doi: 10.1007/s13127-016-0262-x. Google Scholar CrossRef Search ADS   Froufe E, Prié V, Faria J, Ghamizi M, Gonçalves DV, Gürlek ME, Karaouzas I, Kebapçi Ü, Sereflisan H, Sobral C, Sousa R, Teixeira A, Varandas S, Zogaris S, Lopes-Lima M. 2016b. Phylogeny, phylogeography, and evolution in the Mediterranean region: news from a freshwater mussel (Potomida, Unionida). Molecular Phylogenetics and Evolution  100: 322– 332. Google Scholar CrossRef Search ADS   Gante HF. 2011. Diversification of Circum-Mediterranean barbels. In: Grillo O, Venora G, eds. Changing diversity in changing environment . Rijeka: InTech, 283– 298. Guindon S, Dufayard JF, Lefort V, Anisimova M, Hordijk W, Gascuel O. 2010. New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the performance of PhyML 3.0. Systematic Biology  59: 307– 321. Google Scholar CrossRef Search ADS PubMed  Haas F. 1969. Superfamilia Unionacea. Das Tierreich  88: 1– 663. Hanken J, Wake DB. 1993. Miniaturization of body size: organismal consequences and evolutionary significance. Annual Review of Ecology and Systematics  24: 501– 519. Google Scholar CrossRef Search ADS   Hebert PDN, Cywinska A, Ball SL, De Waard JR. 2003. Biological identifications through DNA barcodes. Proceedings of the Royal Society of London B: Biological Sciences  270: 313– 321. Google Scholar CrossRef Search ADS   Huelsenbeck JP, Ronquist F. 2001. MrBayes: Bayesian inference of phylogenetic trees. Bioinformatics  17: 754– 755. Google Scholar CrossRef Search ADS PubMed  Kapli P, Lutteropp S, Zhang J, Kobert K, Pavlidis P, Stamatakis A, Flouri T. 2016. Multi-rate tree processes for single-locus species delimitation under maximum likelihood and Markov chain Monte Carlo. bioRxiv. Available at: http://biorxiv.org/content/early/2016/07/14/063875 Karakousis Y, Machordom A, Doadrio I, Economidis PS. 1995. Phylogenetic relationships of Barbus peloponnesius Valenciennes, 1842 (Osteichthyes: Cyprinidae) from Greece and other species of Barbus as revealed by allozyme electrophoresis. Biochemical Systematics and Ecology  23: 365– 375. Google Scholar CrossRef Search ADS   Khalloufi N, Toledo C. Machordom A, Boumaïza M, Araujo R. 2011. The unionids of Tunisia: taxonomy and phylogenetic relationships, with redescription of Unio ravoisieri Deshayes, 1847 and U. durieui Deshayes, 1847. Journal of Molluscan Studies  77: 1– 13. Google Scholar CrossRef Search ADS   Krijgsman W, Hilgen FJ, Raffi I, Sierro FJ, Wilson DS. 1999. Chronology, causes and progression of the Messinian salinity crisis. Nature  400: 652– 655. Google Scholar CrossRef Search ADS   Lee M, Wolsan M. 2002. Integration, individuality, and species concepts. Biology and Philosophy  17: 651– 660. Google Scholar CrossRef Search ADS   Librado P, Rozas J. 2009. DnaSP v5: a software for comprehensive analysis of DNA polymorphism data. Bioinformatics  25: 1451– 1452. Google Scholar CrossRef Search ADS PubMed  Lopes-Lima M, Sousa R, Geist J, Aldridge DC, Araujo R, Bergengren J, Bespalaja Y, Bódis E, Burlakova L, Van Damme D, Douda K, Froufe E, Georgiev D, Gumpinger C, Karatayev A, Kebapçi Ü, Killeen I, Lajtner J, Larsen BM, Lauceri R, Legakis A, Lois S, Lundberg S, Moorkens E, Motte G, Nagel KO, Ondina P, Outeiro A, Paunovic M, Prié V, von Proschwitz T, Riccardi N, Rudzīte M, Rudzītis M, Scheder C, Seddon M. Şereflişan H, Simić V, Sokolova S, Stoeckl K, Taskinen J, Teixeira A, Thielen F, Trichkova T, Varandas S, Vicentini H, Zajac K, Zajac T, Zogaris S. 2017. Conservation status of freshwater mussels in Europe: state of the art and future challenges. Biological Reviews  92: 572– 607. Google Scholar CrossRef Search ADS PubMed  Lydeard C, Cowie RH, Ponder WF, Bogan AE, Bouchet P, Clarck SA, Cummings KS, Frest TJ, Gargominy O, Herbert DG, Hershler R, Perez KE, Roth B, Seddon M, Strong EE, Thompson FG. 2004. The global decline of non-marine molluscs. BioScience  54: 321– 330. Google Scholar CrossRef Search ADS   Lydeard C, Mulvey M, Davis GM. 1996. Molecular systematics and evolution of reproductive traits of North American freshwater unionacean mussels (Mollusca: Bivalvia) as inferred from 16S rRNA gene sequences. Philosophical Transactions of the Royal Society of London B: Biological Sciences  351: 1593– 1603. Google Scholar CrossRef Search ADS   Machordom A, Araujo R, Erpenbeck D, Ramos MA. 2003. Phylogeography and conservation genetics of endangered European Margaritiferidae (Bivalvia: Unionoidea). Biological Journal of the Linnean Society  78: 235– 252. Google Scholar CrossRef Search ADS   Machordom A, Doadrio I. 2001. Evidence of a cenozoic Betic-Kabilian connection based on freshwater fish phylogeography (Luciobarbus, Cyprinidae). Molecular Phylogenetics and Evolution  18: 252– 263. Google Scholar CrossRef Search ADS PubMed  Mayden RL. 1997. A hierarchy of species concepts: the denouement in the saga of the species problem. In: Claridge MF, Dawah HA, Wilson MR, eds. Species: the units of biodiversity . New York: Chapman & Hall, 381– 424. Moore RC. 1969. Treatise on invertebrate paleontology: Mollusca 6 Bivalvia, Part N, Vol. 1 . Boulder: The University of Kansas and The Geological Society of America Inc., 1– 489. Nagel KO. 2000. Testing hypotheses on the dispersal and evolutionary history of freshwater mussels (Mollusca: Bivalvia: Unionidae). Journal of Evolutionary Biology  13: 854– 865. Google Scholar CrossRef Search ADS   Nagel KO, Badino G. 2001. Population genetics and systematics of European Unionoidea. In: Bauer G, Wächtler K, eds. Ecological studies: ecology and evolution of the freshwater mussels Unionoida, Vol. 145 . Berlin: Springer-Verlag, 51– 80. Google Scholar CrossRef Search ADS   Nagel KO, Badino G, Celebrano G. 1998. Systematics of European naiades (Bivalvia: Margaritiferidae and Unionidae): a review and some new aspects. Malacological Review (Bivalvia I)  ( suppl. 7): 83– 104. Pons J, Barraclough TG, Gomez-Zurita J, Cardoso A, Duran DP, Hazell S, Kamoun S, Sumlin WD, Vogler AP. 2006. Sequence-based species delimitation for the DNA taxonomy of undescribed insects. Systematic Biology  55: 595– 609. Google Scholar CrossRef Search ADS PubMed  Posada D. 2004. Collapse ver. 1.2: a tool for collapsing sequences to haplotypes . Available at: http://darwin.uvigo.es Posada D. 2008. jModelTest: phylogenetic model averaging. Molecular Biology and Evolution  25: 1253– 1256. Google Scholar CrossRef Search ADS PubMed  Prié V. 2011. Les sous-espèces de la Mulette méridionale Unio mancus Lamarck 1819 (Bivalvia, Unionidea) en France: descriptions originales et matériel topotypique. MalaCo  8: 1– 19. Prié V, Puillandre N. 2014. Molecular phylogeny, taxonomy, and distribution of French Unio species (Bivalvia, Unionidae). Hydrobiologia  735: 95– 110. Google Scholar CrossRef Search ADS   Prié V, Puillandre N, Bouchet P. 2012. Bad taxonomy can kill: molecular reevaluation of Unio mancus Lamarck, 1819 (Bivalvia, Unionodae) and its accepted subspecies. Knowledge and Management of Aquatic Ecosystems  405: 1– 18. de Queiroz K. 2005. Ernst Mayr and the modern concept of species. Proceedings of the National Academy of Sciences of the United States of America  102: 6600– 6607. Google Scholar CrossRef Search ADS PubMed  de Queiroz K. 2007. Species concepts and species delimitation. Systematic Biology  56: 879– 886. Google Scholar CrossRef Search ADS PubMed  R Core Team. 2016. R: a language and environment for statistical computing . Vienna: R Foundation for Statistical Computing. Available at: http://www.R-project.org/ Rambaut A, Suchard MA, Xie D, Drummond AJ. 2014. Tracer v1.6 . Available at: http://beast.bio.ed.ac.uk/Tracer Recuero E, Iraola A, Machordom A, García-París M. 2007. Mitochondrial differentiation and biogeography of Hyla meridionalis (Anura: Hylidae): an unusual phylogeographical pattern. Journal of Biogeography  34: 1207– 1219. Google Scholar CrossRef Search ADS   Reid NM, Carstens BC. 2012. Phylogenetic estimation error can decrease the accuracy of species delimitation: a Bayesian implementation of the general mixed Yule-coalescent model. BMC Evolutionary Biology  12: 196. Google Scholar CrossRef Search ADS PubMed  Reis J, Araujo R. 2009. Redescription of Unio tumidiformis Castro, 1899 (Bivalvia: Unionoidea), an endemism from Southwestern Iberian Peninsula. Journal of Natural History  43: 1929– 1945. Google Scholar CrossRef Search ADS   Reis J, Machordom A, Araujo R. 2013. Morphological and molecular diversity of Unionidae (Mollusca: Bivalvia) from Portugal. Graellsia  69: 17– 36. Ronquist F, Huelsenbeck JP. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics  19: 1572– 1574. Google Scholar CrossRef Search ADS PubMed  Roux C, Fraïsse C, Romiguier J, Anciaux Y, Galtier N, Bierne N. 2016. Shedding light on the grey zone of speciation along a continuum of genomic divergence. PLoS Biology  14: e2000234. Google Scholar CrossRef Search ADS PubMed  Rowe KC, Aplin KP, Baverstock PR, Moritz C. 2011. Recent and rapid speciation with limited morphological disparity in the genus Rattus. Systematic Biology  60: 188– 203. Google Scholar CrossRef Search ADS PubMed  Rozas J, Rozas R. 1995. DnaSP, DNA sequence polymorphism: an interactive program for estimating population genetics parameters from DNA sequence data. Computer Applications in the Biosciences  11: 621– 625. Google Scholar PubMed  Sites JWJr, Marshall JC. 2004. Operational criteria for delimiting species. Annual Review of Ecology and Systematics  35: 199– 227. Google Scholar CrossRef Search ADS   Stamatakis A, Hoover P, Rougemont J. 2008. A rapid bootstrap algorithm for the RAxML Web servers. Systematic Biology  57: 758– 771. Google Scholar CrossRef Search ADS PubMed  Steininger F, Rogl F. 1984. Paleogeography and palinspastic reconstruction of the Neogene of the Mediterranean. In: Dixon J, Robertson A, eds. The geological evolution of the Eastern Mediterranean. Oxford: Blackwell Scientific Publications, 17: 659– 668. Google Scholar CrossRef Search ADS   Strayer DL, Downing JA, Haag WR, King TL, Layzer JB, Newton TJ, Nichols SJ. 2004. Changing perspectives on pearly mussels, North America’s most imperilled animals. BioScience  54: 429– 439. Google Scholar CrossRef Search ADS   Swofford DL. 2002. PAUP*: Phylogeny Analysis Using Parsimony (*and other methods), Version 4.0a147 . Sunderland: Sinauer Associates Inc. Tang CQ, Humphreys AM, Fontaneto D, Barraclough TG. 2014. Effects of phylogenetic reconstruction method on the robustness of species delimitation using single-locus data. Methods in Ecology and Evolution  5: 1086– 1094. Google Scholar CrossRef Search ADS PubMed  Turelli M, Barton NH, Coyne JA. 2001. Theory and speciation. Trends in Ecology & Evolution  16: 330– 343. Google Scholar CrossRef Search ADS PubMed  Van Bocxlaer B, Hunt G. 2013. Morphological stasis in an ongoing gastropod radiation from Lake Malawi. Proceedings of the National Academy of Sciences of the United States of America  110: 13892– 13897. Google Scholar CrossRef Search ADS PubMed  Vargas JM, Real R, Guerrero JC. 1998. Biogeographical regions of the Iberian Peninsula based on freshwater fish and amphibian distributions. Ecography  21: 371– 382. Google Scholar CrossRef Search ADS   Vaughn CC, Hakenkamp CC. 2001. The functional role of burrowing bivalves in freshwater ecosystems. Freshwater Biology  46: 1431– 146. Google Scholar CrossRef Search ADS   Wake DB. 2009. What salamanders have taught us about evolution. Annual Review of Ecology, Evolution and Systematics  40: 333– 352. Google Scholar CrossRef Search ADS   Wake DB, Wake MH, Specht CD. 2011. Homoplasy: from detecting pattern to determining process and mechanism of evolution. Science  331: 1032– 1035. Google Scholar CrossRef Search ADS PubMed  Wheeler QD, Meier R, eds. 2000. Species concepts and phylogenetic theory . New York: Columbia University Press. Wilson BE. 1995. A (not-so) radical solution to the species problem. Biology and Philosophy  10: 339– 356. Google Scholar CrossRef Search ADS   Zhang J, Kapli P, Pavlidis P, Stamatakis A. 2013. A general species delimitation method with applications to phylogenetic placements. Bioinformatics  29: 2869– 2876. Google Scholar CrossRef Search ADS PubMed  Zieritz A, Hoffman JI, Amos W, Aldridge DC. 2010. Phenotipic plasticity and genetic isolation-by-distance in the freshwater mussel Unio pictorum (Mollusca: Unionoida). Evolutionary Ecology  24: 923– 938. Google Scholar CrossRef Search ADS   © 2017 The Linnean Society of London, Zoological Journal of the Linnean Society

Journal

Zoological Journal of the Linnean SocietyOxford University Press

Published: Feb 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 12 million articles from more than
10,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Unlimited reading

Read as many articles as you need. Full articles with original layout, charts and figures. Read online, from anywhere.

Stay up to date

Keep up with your field with Personalized Recommendations and Follow Journals to get automatic updates.

Organize your research

It’s easy to organize your research with our built-in tools.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

Monthly Plan

  • Read unlimited articles
  • Personalized recommendations
  • No expiration
  • Print 20 pages per month
  • 20% off on PDF purchases
  • Organize your research
  • Get updates on your journals and topic searches

$49/month

Start Free Trial

14-day Free Trial

Best Deal — 39% off

Annual Plan

  • All the features of the Professional Plan, but for 39% off!
  • Billed annually
  • No expiration
  • For the normal price of 10 articles elsewhere, you get one full year of unlimited access to articles.

$588

$360/year

billed annually
Start Free Trial

14-day Free Trial