A molecular phylogeny of the cosmopolitan hyperdiverse ...molevol.cmima.csic.es/ribera/pdfs/Trizzino_etal2013_Hydraena.pdf · Systematic Entomology (2013), 38, 192–208 DOI: 10.1111/j.1365-3113.2012.00654.x

Systematic Entomology (2013), 38, 192–208 DOI: 10.1111/j.1365-3113.2012.00654.x

A molecular phylogeny of the cosmopolitanhyperdiverse genus Hydraena Kugelann (Coleoptera,Hydraenidae)

M A R C O T R I Z Z I N O1, M A N F R E D A . J A C H2, P A O L O A U D I S I O3,R O C I O A L O N S O4 and I G N A C I O R I B E R A4

1Department of Theoretical and Applied Sciences, Insubria University, Varese, Italy, 2Naturhistorisches Museum Wien, Wien,Austria, 3Department of Biology and Biotechnologies ‘C. Darwin’, Sapienza University, Rome, Italy and 4Animal Biodiversityand Evolution, Institute of Evolutionary Biology (CSIC-UPF), Barcelona, Spain

Abstract. With almost 900 described species, Hydraena Kugelann (Hydraenidae)is one of the largest genera among Coleoptera. The subgeneric classification ofHydraena has been controversial, with 11 subgeneric names having so far beenattributed to it. Some of these, Haenydra Rey and Spanglerina Perkins, havebeen treated as valid genera, as subgenera or as species groups. The most recentcomplete treatment of the genus, based on a cladistic analysis of morphologicalcharacters, recognized two major lineages, and only these were classified assubgenera: Hydraenopsis (mainly Gondwanan distribution), and Hydraena s.str.(mainly Laurasian). Here, we reconstruct the phylogeny of Hydraena using 212species plus several outgroups and approximately 4 kb of sequence data from twonuclear (SSU and LSU ) and four mitochondrial genes (cox1, rrnL, trnL and nad1 ).Data were aligned with two different strategies of multiple alignment (implementedin mafft and prank), and the phylogenies reconstructed using maximum likelihoodand Bayesian methods. We estimated approximate ages of the main nodes usinga relaxed molecular clock with Bayesian methods, and an a priori evolutionaryrate of 0.01 substitutions/site/million years (Ma) plus a calibration point based ona biogeographical split. We found strong support for the monophyly of Hydraenaand many of the clades recognized with morphological data. The following clades areconsidered as subgenera: Phothydraena Kuwert, Spanglerina Perkins, HolcohydraenaKuwert, Hydraenopsis Janssens and Hydraena s.str. The placement of three speciesgroups, two Neotropical (H. multispina group, H. paeminosa group) and one SouthAfrican/Madagascan (H. monikae group), is uncertain, and they are consideredincertae sedis within Hydraena. The origin of the genus was estimated to be inthe Lower Eocene, with many species complexes diversifying in the Pleistocene.Dispersal events seem to have played a key role in order to determine thecurrent distribution of the species groups in the southern hemisphere (mainly inHydraenopsis).

Correspondence: Ignacio Ribera, Institute of Evolutionary Biology (CSIC-UPF), Passeig Maritim de la Barceloneta 37, 08003 Barcelona,Spain. E-mail: [email protected]

192 © 2012 The Royal Entomological Society

Phylogeny of Hydraena 193

Introduction

The genus Hydraena Kugelann, with almost 900 describedspecies and hundreds of undescribed species, is the largestwater beetle genus (see Appendix S1 for a complete checklistof Hydraena spp. with geographical distribution). Adultsof Hydraena are small (approximately 1–3 mm long), andusually inconspicuously coloured. Most species are aquatic,living in the benthos of many different types of freshwaterhabitats, particularly in small streams and other bodies ofrunning water. A few species are terrestrial, dwelling in thewet leaf litter and soil of tropical rainforests. The genus occurson all continents except Antarctica. Many species have veryrestricted distributions (Jach et al., 2005; Jach & Balke, 2008).

The concept of the genus is well defined and its monophylyis well supported by a number of morphological apomorphies(e.g. presence of a labral-mandibular interlocking device, andmentum with acute projection anteriorly; see Jach et al., 2000)and by molecular data (Ribera et al., 2011). Like all hyperdi-verse groups, the generic/subgeneric classification of Hydraenahas been subject to a complex history. Various authors havesplit Hydraena into several, often paraphyletic genera andsubgenera (Rey, 1886; Kuwert, 1888; Seidlitz, 1888; Perkins,1980, 1997; Berthelemy, 1986; Hansen, 1991, 1998), whereasother workers synonymized putatively monophyletic subgen-era with Hydraena s.str. (Hansen, 1991, 1998; Perkins, 1997).Berthelemy (1986) discussed the evolution of some morpho-logical characters of the genus, mainly the number of ely-tral striae and the morphology of the aedeagus. He recog-nized Hydraenopsis, Phothydraena, Haenydra and Hadrenyaas valid subgenera. Perkins (1980, 1997) analysed in detailthe exocrine secretion delivery system, a series of glandularpores and associated structures on the base of head and pro-thorax. Based on this system and the number of elytral striae,Perkins (1997) recognized Haenydra, Hadrenya and Phothy-draena as valid subgenera, but synonymized Hydraenopsiswith Hydraena s.str. Similarly, Spanglerina, described as agenus by Perkins (1980) and reduced to subgenus by Perkins(1989), was finally synonymized with Hydraena s.str. byPerkins (1997) due to similarities in the exocrine delivery sys-tem.

In the only global phylogeny of the genus publishedso far, Jach et al. (2000) carried out a cladistic analysisusing morphological characters of 24 representative taxa,and recognized two major lineages, which they regarded assubgenera: Hydraenopsis, with a hypothesized Gondwananorigin, and Hydraena s.str., with a Laurasian origin. Speciesof Hydraenopsis were found to share some characters of theadult head (incomplete or absent medial genal suture, shortor absent pregula) and pronotum (structure of the hypomeralantennal cleaner) (Jach et al., 2000). Six clades in additionto Hydraena s.str. (Holcohydraena, H . monikae, H . circulatagroup, Phothydraena, Spanglerina, H . palustris group) werenot granted subgeneric status, because their relationshipswere not resolved satisfactorily in that analysis. Within themain lineage of Hydraena s.str., most of the commonlyrecognized species groups (or subgenera) were represented

by a single species, and thus their monophyly could not betested in the cladistic analysis, although a number of potentialautapomorphies were discussed based on the examination ofhundreds of additional species.

Ribera et al. (2011) included some representative speciesof Hydraena as outgroups for the focal clade (‘Haenydra’lineage), but once again incomplete sampling prevented anygeneral conclusions being drawn about the phylogeny of theentire genus.

In this complex scenario, we aimed to reconstruct themolecular phylogeny of Hydraena using a comprehensivesample of more than 200 species of all main lineages anda combination of nuclear and mitochondrial markers. Wespecifically tested the phylogenetic scenario proposed byJach et al. (2000) using morphological data, and providehere a new subgeneric classification of the genus basedon the results of the sequencing and on the congruencewith morphological characters. Due to the comprehensivesampling, we were also able to assess the monophylyof some of the commonly recognized species groups andto provide an approximate temporal framework for thediversification of the genus, including discussions about themain trends in the geographical distribution of the species ofHydraena.

Materials and methods

Taxon sampling

The complete ingroup dataset comprises 213 specimensbelonging to 212 species and one subspecies of Hydraena(H. gracilis balcanica), representing most of the generally rec-ognized species groups of the genus. Sequences of ‘Haenydra’(70 specimens belonging to 69 species and one subspecies), aswell as sequences of 41 species of Hydraena previously usedas outgroups for ‘Haenydra’ (see Appendix S2), were selectedfrom Ribera et al. (2011) and Trizzino et al. (2011). Someadditional sequences were obtained from Ribera et al. (2010a)and Abellan & Ribera (2011). In total, sequences of 102 speciesof Hydraena were newly obtained for this work (Appendix S2).As outgroups we used a selection of seven species of differentgenera of Hydraenidae and Ptiliidae, the latter being generallyrecognized as the sister group of Hydraenidae (Caterino et al.,2005; Hunt et al., 2007) and used to root the tree (AppendixS2).

For the Palaearctic species of Hydraena, the taxonomy andnomenclature of Jach (2004) have been followed, except forH. subintegra aroensis, which was elevated to species rank byJach & Díaz (2012). For non-Palaearctic species we followedthe taxonomy and nomenclature of Jach et al. (2000) andPerkins (2007). Seven of the 102 newly analysed specimensrepresent new species to be formally described elsewhere (seeAppendix S2).

© 2012 The Royal Entomological Society, Systematic Entomology, 38, 192–208

194 M. Trizzino et al.

DNA extraction and sequencing

Specimens were collected in the field and directly pre-served in 96% ethanol or in pure acetone (some specimensof the ‘Haenydra’ lineage). DNA was extracted from wholespecimens by a standard phenol-chloroform extraction or byDNeasy Tissue Kit columns (Qiagen GmbH, Hilden, Ger-many), following the manufacturer’s instructions. Vouchersand DNA samples of the newly analysed specimens are keptin the collections of the Museo Nacional de Ciencias Natu-rales (MNCN, Madrid), and in the Institute of EvolutionaryBiology (CSIC-UPF, Barcelona). DNA extraction was nonde-structive, to preserve voucher specimens for subsequent mor-phometric and morphological analyses. Usually only maleswere sequenced, and the male genitalia (used for the identi-fication of the species) were dissected and mounted prior tothe extraction to ensure correct identification.

For all the 109 newly analysed specimens (including theseven outgroups) we amplified and sequenced four fragments,two mitochondrial [3′-end of cytochrome c oxidase subunit 1(cox1 ), and 3′-end of large ribosomal unit plus the leucinetransfer plus the 5′-end of NADH dehydrogenase subunit1 (rrnL + trnL + nad1 )], and two nuclear [small ribosomalunit (SSU ) and large ribosomal unit (LSU ); see AppendixS3 for primers and Appendix S2 for the new sequences].For 59 species the 5′-end of cox1 (the ‘barcode fragment’,Hebert et al., 2003) was also sequenced (Appendix S2). Forsome specimens the 3′-end cox1 fragment was amplifiedusing internal primers to obtain two smaller fragments ofapproximately 400 bp each (Appendix S3).

Amplifications were performed with the following generalcycle conditions: initial denaturation at 95◦C for 5 min,followed by 33–38 cycles of denaturation at 94◦C for 1 min,annealing at 47–58◦C for 30 s, 1 min extension at 72◦Cand a last 7 min elongation step at 72◦C. Reactions wereperformed in a 25 μL volume containing 16 mm (NH4)2SO4,67 mm Tris–HCl (pH 8.8 at 25◦C), 3 mM MgCl2, 1 mm ofeach dNTP, 0.8 pmol of each primer and 1.25 units of TaqDNA polymerase. Sequences were assembled and edited withgeneious 5.4 (Drummond et al., 2011) or by sequencher 4.7(Gene Codes, Inc., Ann Arbor, MI, U.S.A.). A total of 327 newsequences have been deposited in GenBank (EMBL accessionnumbers HE970771-HE971097; see Appendix S2).

Phylogenetic analyses

For the length-variable regions, we used two alignmentstrategies: multiple pairwise comparisons using the online ver-sion of mafft v.6.8 and the g–ins–i algorithm (MF; Katoh &Toh, 2008), and multiple progressive alignment modelling theevolution of indels with prank (PR; Loytynoja & Goldman,2005). Phylogenetic analyses were conducted with Bayesiananalysis and an approximate maximum likelihood algorithm(ML), as these methods allow the use of a molecular-clockapproach to estimate divergence times. Bayesian analyses wereconducted on a combined data matrix with mrbayes 3.1.2

(Huelsenbeck & Ronquist, 2001), using six partitions cor-responding to the six genes (the rrnL + trnL fragment wasconsidered a single partition, whereas the two cox1 fragmentsas two different partitions due to the large amount of missingdata in the barcode fragment). The use of alternative parti-tions (e.g. by codon position) was not explored due to thescarcity of data for some protein-coding genes. Also, previousresults with similar data and related groups show no relevanttopological differences between the gene and codon partitions(Hidalgo-Galiana & Ribera, 2011; Trizzino et al., 2011). Thebest-fitting model to analyse each partition was selected byjmodeltest (Posada, 2008) using the Akaike information cri-terion (AIC). The mrbayes runs used default values and savedtrees every 1000. ‘Burn-in’ values were established after visualexamination of plots of standard deviation of the split frequen-cies between the two simultaneous runs. Convergence betweenruns was assessed with tracer v1.5 (Drummond & Ram-baut, 2007).

A ML phylogenetic reconstruction was performed withraxml v7.0.4, using the same partition as in mrbayes and witha GTR + G + I model. We ran 100 replicates and selected thetree with the highest likelihood. Node supports were estimatedusing a fast bootstrapping algorithm (1000 replicates with theCAT approximation; Stamatakis et al., 2008), using the samepartition by genes.

We ran some additional analyses with the nuclear data only(102 species, including representatives of all main clades;Appendix S2), using raxml and the same alignments (MFand PR) and analytical procedure as for the combined data.

Estimation of divergence times

To estimate the relative age of divergence of the lineages, weused the Bayesian relaxed phylogenetic approach implementedin beast v1.6.1 (Drummond & Rambaut, 2007), which allowsvariation in substitution rates among branches. We usedan uncorrelated lognormal relaxed molecular-clock model toestimate substitution rates and the Yule process of speciationas the tree prior. We ran two independent analyses samplingeach 1000 generations, and used tracer v1.5 to determineconvergence, measure the effective sample size of eachparameter and calculate the mean and 95% highest posteriordensity interval for divergence times. The initial 10% ofeach run was discarded as burn-in. Results of the two runswere combined with logcombiner v1.6.2 and the consensustree compiled with treeannotator v1.5.4 (Drummond &Rambaut, 2007).

Because of the absence of fossil records to calibrate thetrees, we used as a prior a rate of 2.0% of pairwise divergenceper Ma, established for a closely related family (Leiodidae) fora combination of mitochondrial markers, including those usedhere (Ribera et al., 2010b). This rate was estimated using thecombined mitochondrial protein coding and ribosomal genes,so we applied a GTR + I + G model of DNA substitutionwith four rate categories for the combined mitochondrial dataset only, omitting the nuclear markers and pruning all the



outgroups with the exception of the closer Adelphydraena.Main well-supported nodes were constrained to obtain thesame topology as in the phylogenetic analyses. We set asa prior rate a normal distribution with average rate of 0.01substitutions/site/Ma and a standard deviation of 0.001. Wealso included a calibration point based on the separationbetween the Peloponnesus and mainland Greece, estimated tohave occurred approximately 2.5 Ma (Simaiakis & Mylonas,2008). This age was set as the upper limit for the split betweenthe Greek Hydraena vedrasi and its sister based on moleculardata, the Peloponnesus endemic H. jaechiana (Trizzino et al.,2011). Therefore, for this clade of sister species we set a timingof most recent common ancestor (tmrca) with a truncatednormal distribution, with an upper value equal to 2.5, a meanof 1.5 and a standard deviation of 0.5 (Trizzino et al., 2011).

Results

Molecular phylogeny

The final data matrix included 220 terminals (213 ingroupand seven outgroup species; see Appendix S2) and 3558aligned characters with MF and 3823 with PR. The differencein length corresponds mostly to the expansion of hypervariableregions of the nuclear ribosomal genes. In both cases, theselected evolutionary model was GTR + I + G for each of themitochondrial gene partitions, whereas HKY + G was selectedfor the gene SSU and K80 + G for LSU. A HKY + G modelwas implemented in mrbayes separately for the nuclear genes,as K80 + G cannot be implemented in mrbayes.

The two runs of mrbayes converged to split frequencieslower than 0.05 after 12 Ma generations in the run with MF,and 0.01 after 38 Ma generations in the run with PR. Weset the burn-in fraction to 2.5 and 4.0 Ma generations in theruns with MF and PR, respectively, by estimating an optimaleffective sample size of trees with tracer v1.5 (Drummond& Rambaut, 2007).

The four topologies obtained with the two alignmentmethods and with Bayesian probabilities and ML were verysimilar, especially in the nodes with good support (Figs 1, 2;Appendix S4). Differences affected nodes with low support,basically the resolution among some of the main clades (seebelow).

All analyses strongly supported the monophyly of Hydraena(Fig. 1), and the sister relationship between H. paeminosa andall other Hydraena. The ribosomal sequences of H . paeminosahad several long unique insertions, resulting in a verylong branch that may have introduced some artefact intheir phylogenetic placement (see Discussion). The nextcladogenetic event splits a group including Phothydraena andSpanglerina, and then two main clades with, on the onehand, the rest of the groups considered to be plesiomorphicand, on the other, the derived species of Hydraena s.str.according to Jach et al. (2000). The first clade includesthe H. rugosa and H. circulata groups as sisters, being inturn sisters of the South African/Madagascan species of the

analyses (= H. monikae group) plus Hydraenopsis. Only inthe Bayesian analysis with the two alignments were the SouthAfrican/Madagascan species monophyletic (although with lowsupport). In the ML analyses, they form a paraphyletic group,with H. monikae and an undescribed South African speciesas sister of Hydraenopsis, again with low support (Figs 1, 2;Appendix S4a,c).

The topology of the large clade of the remaining Hydraenas.str. was also very similar across methods. In all cases, theH. palustris group was sister to the rest, followed by the H.bisulcata group, and then two sister clades including, on oneside, the ‘Haenydra’ lineage and, on the other side, a poorlyresolved clade including most of the currently recognizedPalaearctic species groups (Fig. 1), which were, in general,monophyletic and well supported.

Within Hydraenopsis, two main clusters were detected(Fig. 1): (i) a small clade including the Chinese H. cordiformisplus a mixed group of species from India, China and thePhilippines; and (ii) a large, poorly resolved clade, groupingan assemblage of species from all biogeographical regions.Within this latter clade, some of the species groups, previouslyidentified by morphology, were recognizable: H. quadricollis,H. scabra, H. jojoorculloi, H. miyatakei, H. castanea and H.leechi groups (see, e.g., Jach & Díaz, 1998; Freitag & Jach,2007; Fig. 2).

The nuclear markers had, in general, lower variability,and the trees resulting from the raxml analyses had poorresolution and support (Appendix S5). Most of the main cladeswere, however, recovered with both alignments, includingthe monophyly of Hydraena, the monophyly of Spanglerinaplus Phothydraena (with a paraphyletic Phothydraena), themonophyly of the H. rugosa plus H. circulata groups (witha paraphyletic H. rugosa group) and of Hydraenopsis plusthe H. monikae group, and the monophyly of Hydraenopsis(Appendix S5). The clade of Hydraena s.str. was monophyleticwith the MF alignment, but included Phothydraena plusSpanglerina with the PR alignment (Appendix S5). Withinthis large group of Hydraena s.str., the resolution was verypoor due to the low variation of the sequences, but somespecies groups could be recognized, such as, for example,the H. bisulcata, H. holdhausi or H. minutissima groups (seeDiscussion).

Molecular clocks and diversification

For the analysis with beast (Fig. 3) we excluded H. paemi-nosa because of its very long branch and the possibility thatits position was due to some analytical artefact. Differencesbetween the length of the mitochondrial sequence with MFand PR were smaller than for the nuclear sequence (2297 and2449 bp, respectively), and mostly reduced to some variableregions in the gene rrnL. We constrained nine basal nodes toensure that the topologies of the main clades were congru-ent with those obtained in the phylogenetic analyses, most ofthem with high support (Fig. 1). The estimated ages were verysimilar for the two alignments (Table 1).



H. cirrata gr+H. eichleri cplx

Hydraenopsis

Ptiliolum AI649

Spanglerina

H. bisulcata gr

H. grandis gr

Limnebius zaerensis AC14

Asiobates minimus AI447

H. pulchella gr

Adelphydraena orchymonti AI356

H. riparia gr

Ptenidium AI515

H. holdhausi gr

Enicocerus melanescens AI344

[H. monikae gr]

Ochthebius subpictus AI452

[Hydraenopsis paeminosa AI557]

H. circulata gr

H. palustris gr

Phothydraena

H. rugosa gr

H. nigrita gr

H. minutissima gr

‘Haenydra’

H. rufipes gr

Holcohydraena

Hydraena s.str.

Hydraena

H. notsui gr

1/760.85/x

0.97/84 -/x

0.5/ -0.67/-

1/991/97

1/731/67

1/921/91

0.99/630.99/59

1/521/54

1/921/93

1/551/56

1/961/94

1/930.99/77

0.61/-0.74/56

1/991/99

1/x1/98

0.99/730.94/74 0.93/62

0.97/61

1/87 1/90

1/1001/100

1/1001/100

1/951/76

1/981/95

1/1001/100

1/881/91

1/1001/100

1/1001/100

1/981/99

1/971/100

1/82 1/84

0.98/720.99/79

1/951/96

1/1001/100

0.55/x0.55/x

1/1001/100

0.76/740.83/75

Fig. 1. Summarized cladogram of the phylogenetic relationships of the main lineages within Hydraena, according to the results of the differentanalyses. See Fig. 2 and Appendix S2 for the species included in the terminal groups and their geographical origin. Numbers in nodes, Bayesianposterior probabilities in mrbayes (Bpp), when > 0.5/bootstrap values in raxml (bt), when > 50%. Upper row, PR alignment, lower row, MFalignment; ‘–’, compatible nodes (i.e. present with Bpp < 0.5; bt < 50%); ‘x’, incongruent nodes (see Appendix S4 for alternative topologies).In square brackets, species ‘incertae sedis’ (see Appendix S1). Dotted line, species of the H. monikae group, not monophyletic in some analyses(Fig. 2a). Habitus photographs, from top to bottom: H. berthelemyiana, H. grandis, H. janeceki, H. sanagergelyae, H. paganettii (not at the samescale).

Using a calibration of 0.01 substitutions/site/Ma in combi-nation with the constraint of the age of separation betweenH. vedrasi and H. jaechiana (see Materials and methods),the origin of Hydraena was estimated to be at approxi-mately 43 Ma (Lower Eocene), with a wide confidence inter-val (as is usual in this type of analysis; Table 1). Thesplit between Phothydraena + Spanglerina and the remain-ing Hydraena occurred approximately 34–35 Ma, whereas theseparation between Hydraena s.str. and the remaining groups(incl. Hydraenopsis) occurred approximately 32 Ma. The sep-aration between the H. rugosa group + H. circulata group andHydraenopsis occurred 31–32 Ma, while the major derived

Hydraena clades likely originated about 8–25 Ma (Miocene;Fig. 3). Finally, the majority of species groups were lessthan 15 Ma old, whereas most of the terminal clades of theallopatric sibling species were estimated to be less than 3 Maold (i.e. of Plio-/Pleistocene origin).

Subgeneric classification

Our phylogenetic results support the classification of thespecies of Hydraena s.l. in five monophyletic groups, wellsupported and largely concordant with previously recognized



H. isabelae AI170

H. vandykei AI1276

Hydraenopsis scabra AH138

Ptiliolum AI649

H. circulata RA330

Hydraenopsis abyssinica AI1228

Hydraenopsis sp TAZ RA321

Hydraenopsis accurata AH97

Hydraenopsis cristatigena AI1096

Hydraenopsis cordiformis AI416

Hydraenopsis cooperi AI456

H. brevis AI122

Hydraenopsis tricantha AF166

Hydraenopsis sp NCA AI429Hydraenopsis sp AUS AF90

Hydraenopsis sp MAY AF202

Hydraenopsis castanea AF89

Hydraenopsis quadricollis AI1312

Ochthebius subpictus AI452


Hydraenopsis claudia AH134

H. monikae AI455

H. hernandoi AI435

Hydraenopsis sp PHI AH139

Hydraenopsis cf cyclops AI418

Hydraenopsis fontana AI708

Hydraenopsis sp MAD AI547

H. pacifica AI1151

H. sp SAF RA298

Ptenidium AI515

H. serricollis AI1093

H. arenicola AI504

H. testacea AI566

Hydraenopsis sp FIJ AI419

Asiobates minimus AI447

Limnebius zaerensis AC14

Hydraenopsis jojoorculloi AH132

Hydraenopsis hosiwergi AH136


H. sp MAD AI542

H. sp CAN AI610

H. cf pacifica RA302

H. exarata AI169

Hydraenopsis sp PER AI687

Hydraenopsis quadricollis gr RA98

Hydraenopsis castanescens AH137

H. sp MAD AI541

Hydraenopsis kodadai AH131Hydraenopsis sp TAZ RA323

H. rugosa AI392

H. petila AI465

H. pallidula AF211

Hydraenopsis victoriae AC29

Hydraenopsis cf concinna AF194

Hydraenopsis manguao AH135

Hydraenopsis paeminosa AI557

Hydraenopsis palawanensis AH133

Hydraenopsis williamsensis AF87

Hydraenopsis cf fundata AF88

H. putearius RA95

Hydraenopsis cf rhinoceros RA319

H. cf tuolumne RA329

H. marinae AI173

Hydraenopsis cf luminicollis AI423

Hydraenopsis scintillabella AI688

Hydraenopsis miyatakei AI828

H. atrata AI314

Enicocerus melanescens AI344

Hydraenopsis scintillutea gr AI709

0.65/- x/x

0.91/- x/x

0.67/-0.68/-

0.99/63

(a)

0.99/59

1/521/54

1/551/56

1/871/90

0.94/660.58/-

0.93/95

0.85/90

0.54/- - / -

0.57/x x/x

0.88/- 1 /-

0.9/841/84

1/971/88

0.59/- x /x

0.54/- x/ x

0.61/- x/x

0.82/-0.93/55

1/630.99/91

0.95/- x /x

1/881/91

0.66/940.58/94

0.87/630.99/1

1/951/76

0.99/- x/x

0.97/84 1/x

0.60/-0.88/-

0.5/-0.67/-

0.56/x0.75/-

0.88/- - / -

0.84/-0.90/-

0.64/x0.80/x

0.88/611/65

1/731/67

1/86 x/x

0.94/610.96/63

0.79/-0.68/x

0.98/790.99/84

0.89/- - / -

subg. Phothydraena

subg. Spanglerina

subg. Holcohydraena

H. rugosa gr

H. circulata gr

subg. Hydraenopsis

‘incertae sedis’

Fig. 2b

1/760.85/x

H. monikae gr

Fig. 2. Phylogram obtained by mrbayes with the combined data and the alignment with prank. Numbers in nodes, Bayesian posterior probabilitiesin mrbayes (Bpp), when > 0.5/bootstrap values in raxml (bt), when > 50%. Upper row, PR alignment, lower row, MF alignment; ‘−’, compatiblenodes (i.e. Bpp < 0.5; bt < 50%); ‘x’, incongruent nodes (see Appendix S4 for alternative topologies). For simplicity, nodes with support values ofBpp = 1 and bt = 100 for the two alignments are indicated by a black circle, and values of Bpp = 1 and bt = 100 for one alignment are indicatedby ‘1’.



H. watanabei AC31

H. schubertorum AI820

H. palustris AI309

H. egoni RA112

H. alia H006_1

H. helena AI784

H. nigrita AI345

H. dochula AI518

H. riparia AI312

H. similis AH177

H. marcosae AI904

H. holdhausi AI1025

H. subjuncta AI1230

H. pachyptera RA82

H. riberai AI568

H. gavarrensis AI1288

H. reyi AI320

H. britteni AI1124

H. quilisi AI1085

H. minutissima AI604

H. sicula AC40

H. pygmaea AI346

H. brachymera AI323

H. allomorpha AI316

H. barrosi AI954

H. curta AI383

H. subacuminata AI305

H. inapicipalpis AI1043

H. sardoa AF206

H. subirregularis AH183H. delia AI1061

H. melas AI370

H. andalusa AI168

H. balearica AI175

H. albai AI989

H. lucasi AI864

H. bolivari AI171

H. rufipes AI308

H. subsequens AI306

H. croatica RA52

H. andreinii AI873

H. canakcioglui AI790

H. angulosa AI1283

H. graphica AI1267

H. rigua AI644

H. sp OMA RA97

H. unca AI276

H. alcantarana RA38

H. stussineri AI196

H. intermedia AI436

H. coryleti AI767

H. servilia AI274

H. speciosa AI1297

H. platycnemis RA44

H. chiesai RA346

H. aethaliensis RA143H. hayashii AI691

H. assimilis AI1050

H. cordata AI388

H. armipes RA41

H. tsushimaensis AC32

H. corrugis AI1016

H. bisulcata AI172

H. dentipalpis AI492

H. antiatlantica AI567

H. pamphylia AI495

H. corinna AI284

H. sharpi AI448

H. persica RA305

H. grandis AI516

H. kasyi AI1026

H. biltoni RA111

H. carbonaria AI947

H. mecai PA280

H. attaleiae AI496

H. pulchella RA348

H. affusa AI1042

H. capta AI167

H. subimpressa AI289

H. morio AI1203

0.76/1000.83/75

0.55/x0.55/x

0.96/670.92/71

0.93/620.97/61

0.97/85 x/x

0.54/-0.55/-

0.63/10.67/1

10.96/100

1/851/82

1/x1/x

1/861/90

0.66/750.68/77

0.96/820.97/87

0.7/-0.7/-

0.7/-0.7/-

0.83/x0.55/x

1/871/87

0.85/960.99/96

1/841/92

0.59/-0.62/-

1/x1/x

0.99/800.99/85

0.56/x x/x

1/561/58

1/861/87

0.97/990.56/84

0.74/940.78/97

0.79/71(b)0.74/67

0.88/x0.89/x

0.61/-0.74/56

0.67/x0.77/x 1/93

0.99/77

0.70/-0.74/50

0.85/550.79/56

10.9/100

1/831/84

0.73/760.77/72

0.99/-0.94/74

0.89/750.86/77

0.97/710.96/79

0.96/1000.96/86

0.84/x0.81/x

1/81x/x

0.8/0.8/-

0.77/520.68/54

1/821/84

H. palustris gr

H. bisulcata gr

H. nigrita gr

H. holdhausi gr

H. riparia gr

H. pulchella gr

H. grandis gr

H. rufipes gr

H. minutissima gr

H. cirrata gr

Fig. 2a

Fig. 2c

Hydraena s.str.

H. eichleri cplx

H. notsui gr

Fig. 2. Continued



H. gracilidelphis AI510

H. vedrasi RA94

H. belgica AI426

H. caucasica AI493

H. gracilis balcanica AI338

H. lazica 138_1

H. saga AI485

H. sinope AI788

H. akbesiana 49_1

H. solarii 62_1

H. tarvisina AI1129

H. iberica AI181

H. diazi AI479

H. excisa AI391

H. madronensis AI424

H. graciloides 139_1

H. occitana 24_1

H. gaditana AI166

H. epeirosi 89_1

H. devincta AI966

H. alpicola AI347

H. crepidoptera 137_1

H. bitruncata AI354

H. pangaei 80_1

H. tatii AI164H. manfredjaechi AI313

H. gracilis AI332

H. samnitica 78_1

H. aroensis 82_1

H. fontiscarsavii 134_1

H. leonhardi AI339

H. altamirensis AI425

H. catalonica AI1060

H. rosannae 16_1

H. polita AI165

H. tyrrhena 21_1

H. larissae AI303

H. devillei AI288

H. bicuspidata 163_1

H. fosterorum AI481

H. solodovnikovi 167_1

H. plumipes 34_1

H. muelleri 67_1H. truncata AI357

H. dentipes AI361

H. bononiensis 22_1

H. czernohorskyi 65_1

H. christinae 106_2

H. subintegra AI340

H. scitula 142_1

H. septemlacuum AI795

H. emarginata AI325

H. monstruosipes AI439

H. carniolica AI1049

H. dalmatina RA80

H. lapidicola AI292

H. heterogyna AI294

H. lusitana AI385

H. decolor 28_1

H. evanescens AI286

H. producta AI402

H. anatolica AI802

H. zezerensis AI182

H. sanfilippoi 23_1

H. schuleri AI400

H. discreta 127_1

H. integra AI783

H. hispanica AI329

H. bensae AI293

H. jaechiana 99_2

H. exasperata AI523

0.67/-0.60/-

0.99/750.93/68

0.69/x0.67/x

0.63/-0.52/x

0.5/x0.5/-

0.64/970.58/92

0.9/600.96/59

0.85/60(c)0.86/56

0.86/580.79/62

0.94/-0.62/-

0.51/x0.51/58

0.9/-0.9/x

0.98/670.99/66

10.8/99

0.96/600.98/62

1/891/89

1/820.97/81

1/891/88

0.56/-0.58/-

1/851/84

0.7/-0.7/-

0.93/-0.78/-

0.67/600.62/65

0.8/560.8/58

0.87/620.92/70

1/871/84

0.96/600.98/62

H. caucasica clade

H. gracilis clade

H. iberica clade

H. dentipes clade

Hydraena s.str.”Haenydra” lineage

Fig. 2b

Fig. 2. Continued



0 Ma1020304050



Hydraenopsis sp PHI AH139

H. brevis AF122

Hydraenopsis castanea AF89

Hydraenopsis scabra AH138

Hydraenopsis sp FIJ AI419

H. cf pacifica RA302


Hydraenopsis cf concinna AF194

H. rugosa AI392

Hydraenopsis tricantha AF166

H. atrata AI314

H. marinae AI173

Hydraenopsis cf cyclops AI418

H. serricollis AI1093

H. circulata RA330

Hydraenopsis sp PER AI687

Hydraenopsis palawanensis AH133

H. sp MAD AI542


Hydraenopsis cristatigena AI1096

H. vandykei AI1276

Hydraenopsis miyatakei AI828

H. cf tuolumne RA329

Hydraenopsis kodadai AH131

H. exarata AI169

Hydraenopsis jojoorculloi AH132

H. sp MAD AI541

Hydraenopsis cf rhinoceros RA319

Hydraenopsis accurata AH97

Hydraenopsis hosiwergi AH136

Hydraenopsis sp MAD AI547

H. monikae AI455

Hydraenopsis scintillabella AI688

H. pacifica AI1151

Hydraenopsis sp NCA AI429

Hydraenopsis cooperi AI456

Hydraenopsis sp MAY AF202

H. sp CAN AI610

H. sp SAF RA298

Hydraenopsis castanescens AH137

Hydraenopsis fontana AI708

H. testacea AI566

Hydraenopsis victoriae AC29

Hydraenopsis manguao AH135

Hydraenopsis cf luminicollis AI423Hydraenopsis sp AUS AF90

Hydraenopsis cf fundata AF88

H. arenicola AI504

Hydraenopsis williamsensis AF87

H. petila AI465

H. hernandoi AI435

Hydraenopsis quadricollis gr RA98Hydraenopsis quadricollis AI1312

Hydraenopsis claudia AH134

H. putearius RA95

H. isabelae AI170

Hydraenopsis abyssinica AI1228

H. pallidula AF211

Hydraenopsis cordiformis AI416

Hydraenopsis scintillutea gr AI709

8.4

4.9

20.9

25.9

20.5

19.3

19.4

43.4

0.8

21.8

30.4

15.5

34.3

0.4

8.83.7

19.2

30.6

17.4

10.5

11.2

22.7

21.5

31.1

7.6

5.7

12.1

17.7

21.1

0.8

26.5

8.3

14.6

10.3

3.8

19.0

24.4

4.317.5

23.0

12.0

18.1

28.0

12.8

10.0

4.4

13.4

0.15

12.0

17.4

31.6

(a)

0.7

13.0

24.2

22.4

0.5

5.09.9

16.4

6.5

9.7

17.1

Eocene Oligocene Miocene Pliocene Pleistocene

Spanglerina

Phothydraena

Holcohydraena

Hydraenopsis

inc. sed.

Fig. 3b

Hydraena

Fig. 3. Ultrametric tree obtained by beast with the mitochondrial sequences of the PR alignment, with a molecular rate of 2% divergence/ Ma andthe constraint of the split between H. jaechiana and H. vedrasi (red circle, see Materials and methods) at 1.5 Ma. Some nodes were constrained tobe congruent with those obtained with the combined dataset (black circles).



H. dentipalpis AI492

H. antiatlantica AI567

H. subirregularis AH183

H. croatica RA52

H. barrosi AI954

H. aethaliensis RA143

H. palustris AI309

H. britteni AI1124

H. pygmaea AI346

H. kasyi AI1026

H. riparia AI312

H. chiesai RA346

H. pamphylia AI495

H. attaleiae AI496

H. pachyptera RA82

H. affusa AI1042

H. nigrita AI345

H. riberai AI568

H. reyi AI320

H. schubertorum AI820

H. melas AI370

H. canakcioglui AI790

H. carbonaria AI947

H. graphica AI1267

H. persica RA305

H. egoni RA112

H. hayashii AI691

H. alcantarana RA38

H. similis AH177

H. albai AI989

H. rufipes AI308

H. grandis AI516

(b)

H. sp OMA RA97

H. minutissima AI604

H. brachymera AI323

H. allomorpha AI316

H. alia H006_1

H. intermedia AI436

H. inapicipalpis AI1043

H. subimpressa AI289

H. assimilis AI1050

H. cordata AI388

H. corrugis AI1016

H. andalusa AI168

H. sardoa AF206

H. marcosae AI904

H. angulosa AI1283

H. armipes RA41

H. sicula AC40

H. tsushimaensis AC32

H. capta AI167

H. gavarrensis AI1288

H. corinna AI284

H. balearica AI175

H. watanabei AC31

H. andreinii AI873

H. holdhausi AI1025

H. pulchella RA348

H. platycnemis RA44

H. quilisi AI1085

H. bisulcata AI172

H. unca AI276

H. subacuminata AI305

H. subsequens AI306

H. curta AI383

H. helena AI784

H. rigua AI644

H. sharpi AI448

H. coryleti AI767

H. biltoni RA111

H. dochula AI518

H. mecai PA280

H. stussineri AI196

H. subjuncta AI1230

H. bolivari AI171

H. speciosa AI1297

H. servilia AI274

H. lucasi AI864

H. delia AI1061

H. morio AI1203

9.5

10.6

1.0

3.3

3.2

0.18

8.3

13.4

3.2

1.3

9.9

10.7

3.7

6.6

6.3

4.2

0.7

6.5

2.2

0.6

10.5

4.8

8.5

12.1

4.2

0.23

13.9

1.8

0.6

2.8

5.3

0.8

0.5

9.0

7.5

22.6

0.36

17.7

5.2

10.5

4.8

8.6

3.1

7.3

8.5

1.2

0.47

1.29.0

2.6

3.4

3.0

5.1

0.17

2.6

14.6

24.7

3.5

16.3

1.7

2.1

11.2

9.7

4.9

7.0

10.2

7.5

0.9

4.7

13.2

5.9

13.1

28.0

0.32

1.0

0.7

0.6

5.2

1.7

5.4

Fig. 3c

Hydraena s.str.

H. palustr is gr

H. bisulcata gr

H. minutissima gr

H. holdhausi gr

H. nigrita gr

H. riparia gr

H. pulchella gr

H. grandis gr

H. cirrata gr

H. rufipes gr

H. rufipes gr

0 Ma102030405 0

Eocene Oligocene Miocene Pliocene Pleistocene

Fig. 3a

H. eichleri cplx

H. notsui gr

Fig. 3. Continued



H. czernohorskyi 65_1

H. lusitana AI385

H. dentipes AI361

H. anatolica AI802H. gracilidelphis AI510

H. lazica 138_1

H. altamirensis AI425

H. rosannae 16_1

H. alpicola AI347

H. akbesiana 49_1

H. gracilis AI332

H. monstruosipes AI439

H. integra AI783

H. dalmatina RA80H. christinae 106_2

H. fontiscarsavii 134_1

H. bicuspidata 163_1H. polita AI165

H. lapidicola AI292

H. pangaei 80_1

H. sinope AI788

H. iberica AI181

H. exasperata AI523

H. aroensis 82_1

H. fosterorum AI481

H. plumipes 34_1

H. bononiensis 22_1

H. catalonica AI1060

H. leonhardi AI339

H. saga AI485

H. gracilis balcanica AI338

H. zezerensis AI182

H. emarginata AI325

H. subintegra AI340

H. producta AI402

H. carniolica AI1049

H. muelleri 67_1

H. bensae AI293

H. manfredjaechi AI313

H. occitana 24_1

H. truncata AI357

H. tarvisina AI1129

H. graciloides 139_1

H. vedrasi RA94

H. scitula 142_1

H. schuleri AI400

H. septemlacuum AI795

H. devillei AI288

H. samnitica 78_1

H. sanfilippoi 23_1

H. devincta AI966

H. gaditana AI166

H. epeirosi 89_1

H. tatii AI164

H. solodovnikovi 167_1

H. discreta 127_1

H. diazi AI479

H. solarii 62_1

H. tyrrhena 21_1

H. madronensis AI424

H. jaechiana 99_2

H. crepidoptera 137_1

H. excisa AI391

H. hispanica AI329

H. belgica AI426

H. evanescens AI286

H. decolor 28_1

H. bitruncata AI354

H. larissae AI303

H. heterogyna AI294

H. caucasica AI493

0.37

2.0

1.4

1.3

6.3

2.0

2.4

3.4

0.18

3.1

8.0

10.4

1.4

1.7

4.3

9.1

0.10

0.17

1.2

1.0

0.22

6.2

4.2

6.0

5.8

1.8

1.55

4.7

0.5

5.6

4.6

1.7

0.6

2.70.11

5.5

1.4

1.8

0.4

0.6

0.9

0.5

6.7

2.1

5.5

1.2

8.7

2.4

1.9

0.25

0.5

4.2

0.8

0.17

1.5

6.0

3.8

0.05

3.2

3.9

1.3

0.38

0.5

0.39

2.2

0.33

3.8

0.9

2.9

2.4

‘Haenydra’ l ineage

0 Ma1020304050

Eocen

(c)

e Oligocene Miocene Pliocene Pleistocene

Fig. 3b

Fig. 3. Continued



Table 1. Estimated ages [Ma (95% confidence interval)] obtained inbeast for some nodes in the phylogeny of Hydraena with the twoalignments.

Clade prank mafft

Hydraena, stem group 43.4 (31–58) 43.3 (30–58)Hydraena* 34.3 (27–42) 34.5 (27–42)Phothydraena + Spanglerina* 34.2 (22–39) 29.4 (21–38)Phothydraena* 17.4 (11–24) 21.5 (14–29)Hydraena excl.* (Phothy-

draena + Spanglerina)31.1 (25–38) 32.1 (26–40)

Holcohydraena* 22.4 (15–30) 22.3 (15–30)H. rugosa group 10.5 (6–16) 10.9 (6–16)H. circulata group 9.7 (6–13) 9.7 (6–13)Hydraenopsis* + H. monikae

group30.4 (24–37) 30.8 (24–38)

Hydraenopsis* 28.0 (22–34) 28.2 (22–34)Hydraena s.str.* 28.0 (22–35) 27.9 (22–35)H. palustris group 13.9 (9–19) 14.2 (9–19)Hydraena s.str. excl. palustris

group*24.7 (19–31) 25.3 (20–32)

H. bisulcata group 10.5 (7–13) 10.8 (7–14)Hydraena s.str. excl. (palustris

and bisulcata grs)*22.6 (17–28) 23.6 (18–30)

H. minutissima + notsui + 17.7 (14–22) 18.5 (14–23)holdhausi + nigrita + riparia +pulchella + cirrata + eichleri+grandis + rufipes groupsH. minutissima group 10.2 (6–15) 10.7 (6–15)H. holdhausi group 11.2 (8–15) 11.4 (8–15)H. nigrita group 10.7 (8–14) 10.4 (8–13)H. riparia group 9.9 (7–12) 10.1 (7–13)H. pulchella group 8.5 (6–12) 9.1 (6–13)H. cirrata group 7.0 (5–10) 7.0 (5–10)H. grandis group 4.2 (3–6) 4.1 (3–6)H. rufipes group – 9.1 (7–12)Haenydra lineage 10.4 (8–13) 10.6 (8–14)

Unless specified, all ages refer to crown groups (i.e. the age of thelast common ancestor of all included species in the clade). Asterisksindicate nodes which monophyly was constrained in the beast analyses(see Fig. 3).

taxa that we reconsider here as subgenera (see Appendix S1for a complete list of subgenera and species):

Genus Hydraena Kugelann, 1794. Type species: H. ripariaKugelann, 1794, by monotypy.

1. Subgenus Phothydraena Kuwert, 1888. Type speciesHydraena testacea Curtis, 1830, by monotypy. Originallydescribed as subgenus, in the more recent literature treatedas subgenus (e.g. Berthelemy, 1986; Perkins, 1997) or asspecies group of Hydraena s.str. (Jach et al., 2000). Itcurrently includes nine species of western Palaearctic dis-tribution.

2. Subgenus Spanglerina Perkins, 1980. Type species Span-glerina ingens Perkins, 1980, by original designation.Originally described as genus, later downgraded to sub-genus (Perkins, 1989) and subsequently synonymized withHydraena s.str. (Perkins, 1997). Currently it includes fourspecies distributed in Central America (Perkins, 1980).

3. Subgenus Holcohydraena Kuwert, 1888 (synonym: Taenhy-draena Kuwert, 1888). Type species Hydraena rugosa Mul-sant, 1844, designated by Berthelemy (1986). Originallydescribed as subgenus, it was synonymized with Hydraenas.str. by Berthelemy (1986). The 24 species are distributedin the western Palaearctic (H. rugosa group, three species)and in the Nearctic (H. circulata group, 21 species).

4. Hydraenopsis Janssens, 1972. Type species Hydraenopsisvietnamensis Janssens, 1972, by original designation.Originally described as a genus, it was considered as asubgenus by several authors (e.g. Jach, 1986; Perkins,1989; Jach et al., 2000). Perkins (1997) synonymized itwith Hydraena s.str. However, our phylogenetic conceptof Hydraenopsis agrees with Jach et al. (2000), butexcludes H. paeminosa. Currently, this subgenus includes458 species, mostly distributed in the southern hemisphere,but occurring also in the Nearctic and in the southern andeastern Palaearctic (Appendix S1).

5. Hydraena s.str. Kugelann, 1794 (synonyms: Hadrenya Rey,1886; Haenydra Rey, 1886; Hoplydraena Kuwert, 1888).According to our analyses, this subgenus corresponds tothe concept of Hydraena s.str. published by Jach et al.(2000) with the exclusion of the species of Holcohydraena,Phothydraena and the H. monikae group. In accordancewith Jach et al. (2000), Haenydra and Hadrenya remainsynonyms, as their consideration as subgenera would renderHydraena s.str. paraphyletic. Since numerous species(especially from the eastern Palaearctic Region) were notavailable for sequencing, we have not been able to furtherdivide the subgenus or to draw final conclusions aboutspecies groups. Currently, Hydraena s.str. includes 385species living in the Palaearctic Region and the northernmargin of the Oriental Realm.

Three species groups, all of them plesiomorphic, have notbeen included in any of these subgenera: the NeotropicalH. multispina and H. paeminosa groups (Perkins, 2011b) andthe South African/Madagascan H. monikae group (includesone described species and several undescribed ones). Unfortu-nately, no species of the H. multispina group could be obtainedfor sequencing. Due to the uncertainty of the phylogenetic posi-tion of these three species groups, plus the lack of availablesubgeneric names, we prefer to provisionally consider them as‘incertae sedis’ within the genus Hydraena.

Discussion

Phylogeny and biogeography of Hydraena s.l.

We obtained a very robust phylogeny of Hydraena s.l.despite minor differences in the alignment and reconstructionmethod. Differences were restricted to poorly supported nodesand generally affected the internal phylogeny of the mainclades, not the backbone of the tree, which was remarkablystable. According to our results, Hydraena is monophyleticand originated in the lower Eocene. The incomplete sampling



of some of the lineages of the genus, and in particularHydraenopsis, does not allow a formal reconstruction of itsgeographical origin. It is interesting to note, however, thattwo of the main clades have a sister relationship betweenwestern Palaearctic and Nearctic or Neotropical species(Phothydraena plus Spanglerina, and the two species groupsof Holcohydraena: H. rugosa and H. circulata; see Fig. 1).These disjunct distribution patterns cannot be explained by thetectonic opening of the Atlantic (which was fully completedby the Early Eocene; Hallam, 1994). Holcohydraena probablydispersed across the Bering Strait during the Miocene.

The position of H. paeminosa as sister of the remainingspecies of Hydraena is surprising but very consistent acrossmethods. Morphologically, it combines plesiomorphic (gono-coxite divided) and highly derived (aedeagal morphology)characters. The lack of data of other species of the H. paemi-nosa group and the probably related H. multispina group(Perkins, 2011b) does not allow final conclusions about itsphylogenetic position. As noted earlier, the long branch of thespecies, due to some long insertions in the nuclear ribosomalgenes, may have introduced some artefacts in the analyses.

The subgenus Phothydraena is characterized by severalunambiguous autapomorphies in the structure of the elytralpunctures and in the presence of two pairs of glabrousmetaventral plaques (Berthelemy, 1986; Jach et al., 2000).

The only species of Spanglerina included in the study,Hydraena brevis, was in all analyses sister of Phothydraena.The species of Spanglerina share with Phothydraena thepresence of a hyaline membrane on the pronotum (presentalso in H. paeminosa; see Jach et al., 2000) and the very widepseudepipleura. Synapomorphies of the species of Spanglerinaare the strongly arched fronto-clypeal suture, the anteriorlyand posteriorly strongly attenuate pronotum, and a derivedaedeagus (Perkins, 1980).

The species of Holcohydraena share some plesiomorphicgenital characters, and the absence of conspicuous malesecondary sexual characters. It should be noted that nounequivocal morphological synapomorphies between the twospecies groups of this subgenus (H. rugosa group andH. circulata group) have so far been identified (Jach et al.,2000). Further studies are thus required to better characterizethis subgenus. On the other hand, a separate subgeneric statusfor the H. circulata group could be taken into consideration.However, since there is no published name for this groupavailable, we refrain from a formal description.

The position of the South African/Madagascan species ofthis analysis (H. monikae group) is not well established in ourphylogenies. They seem to be related to Hydraenopsis, butwhether they are monophyletic or paraphyletic still remainsquestionable. We have studied the only formally describedspecies of this group (H. monikae) plus three undescribedspecies, but the inclusion of more of the many undescribedspecies of the area (P. Perkins, personal communication)may contribute to the clarification of their phylogeneticposition. As noted by Jach et al. (2000), H. monikae sharessome plesiomorphic characters with species belonging toPhothydraena, Holcohydraena and some Hydraenopsis (e.g.

presence of an intercoxal cavity on abdominal sternites II andIII).

Hydraenopsis is, together with Hydraena s.str., the mostdiverse clade of the genus. The analysed dataset comprises justa small representation of the 459 described species (Janssens,1972; Jach et al., 2000; Perkins, 1980, 2007, 2011a,b; seeAppendix S1 for an updated checklist). Notwithstandingthis, our sampling included species from a wide geographicrange (Australia & Pacific, South America, Asia, Africa),representing many of the morphologically identified speciesgroups. All the analyses grouped the sampled Hydraenopsisspp. into a well supported monophyletic group, with severalmorphological synapomorphies in the structure of the gena andthe gula (Jach et al., 2000).

The relatively recent origin for the diversification within thegenus Hydraena estimated by beast excludes the hypothesisthat the current distribution of Hydraenopsis was drivenby vicariance as a consequence of the fragmentation ofGondwana (which was completed by the Late Cretaceous,as noted earlier). Our data suggest instead that Hydraenopsisprobably started to diversify ∼ 28 Ma in the Oriental Regionand that the current geographical distribution is due todispersal events to the Australian, African and Americancontinents. The Neotropical species have a rather derivedposition within Hydraenopsis (Fig. 2A), suggesting a relativelyrecent colonization from either Africa or eastern Asia. Thepresence of Hydraenopsis in some remote Pacific Islands (e.g.Fiji; Appendix S1) demonstrates their ability for long-distancetransoceanic dispersal. The colonization of South Americafrom Africa by dispersal has been hypothesized for reptilesand plants, thus indicating that dispersal may have been moreimportant than traditionally assumed (de Queiroz, 2005; Roweet al., 2010; Oaks, 2011; Townsend et al., 2011).

Within Hydraenopsis, the phylogeny remains largely unre-solved, probably because of the limited number of analysedspecies. The sister relationships of a clade of Oriental species,including H. cordiformis, with the rest of the subgenus agreeswith the morphological analyses of Jach et al. (2000) but isnot generally well supported. Within the subgenus, some of thepreviously defined species groups could be recognized (e.g. H.quadricollis, H. scabra, H. jojoorculloi, H. miyatakei, H. cas-tanea and H. leechi groups), but there are no clear geographicalpatterns and at this stage it seems impossible to define thecomposition of these lineages with any confidence.

Hydraena s.str. includes the majority of the Palaearcticspecies of the genus (except for Phothydraena, Holcohydraenaand several species of Hydraenopsis). Within this subgenus,several monophyletic species groups were recognized.

The Hydraena palustris group includes a small numberof European/North African species (Figs 1, 2B). This cladewas, in all analyses, sister to the rest of Hydraena s.str.Species of the H. palustris group share a single, presumablyplesiomorphic character (i.e. the structure of the intercoxalcavity) with H. monikae, Holcohydraena and Phothydraena(Jach et al., 2000). The species of the H. palustris groupare characterized by a similar habitus, by the anteriorlystrongly emarginate pronotum, by the absence of conspicuous



secondary sexual characters and by rather simple malegenitalia, resembling those of other plesiomorphic groups ofHydraena.

The rest of the large and poorly resolved Hydraena cladeincludes three main lineages: H. bisulcata group, H. ripariaand related groups, and the ‘Haenydra’ lineage.

The H. bisulcata group includes eight species with south-western European/northeastern African distribution. In previ-ous analyses it was found to be sister to ‘Haenydra’ (Riberaet al., 2011; Trizzino et al., 2011), although in general with lowsupport. It must also be noted that our dataset did not includeany species of the Chinese H. armipalpis group, which maybe closely related to ‘Haenydra’ (Jach et al., 2000). There-fore, the phylogenetic placement of this large clade within thesubgenus Hydraena could not be settled satisfactorily.

Within the lineage including H. riparia (plus related speciesgroups), species with close morphological affinities (e.g.aedeagal similarities) were usually placed together with strongsupport. Some of these clades correspond to traditionallyrecognized species groups, such as the H. riparia, H. nigrita, H.minutissima, H. grandis, H. cirrata, H. holdhausi or H. rufipesgroups (Figs 1, 2A; e.g. Jach, 1988; Berthelemy et al., 1991;Jach & Skale, 2009). However, the composition of these groupsand the relationships between them are still uncertain, with lowsupport for some nodes and alternative topologies.

The H. riparia and H. nigrita groups were recovered in asingle monophyletic clade by some analyses, but paraphyleticin others. The H. riparia group includes species usually largerin size and with more marked secondary sexual characters(especially in the maxillary palps), while the species of theH. nigrita group are usually smaller, with less conspicuoussecondary sexual characters.

Species of the H. grandis group share a very large body size(sometimes being more than 3 mm long), marked secondarysexual characters and very complex male genitalia. Accordingto molecular data, this clade is monophyletic, well supportedand sister to the few analysed species of the H. cirratagroup + H. eichleri complex (sensu Jach & Kasapoglu, 2006)(Fig. 2B). The split between the H. grandis group andthe H. cirrata group + H. eichleri complex occurred around8–9 Ma, whereas the majority of species arose more recently,during the Pleistocene glacial cycles (Fig. 3).

The H. rufipes group was not monophyletic in some analyses(Fig. 2B; Appendix S4). Further morphological studies arerequired to clarify the relationships within this group. Inall analyses, H. pulchella and its allies were placed in thesame clade as the H. grandis and H. rufipes groups, althoughgenerally with low support (Fig. 2B).

The H. minutissima group was considered as a valid sub-genus or even as a genus (Hadrenya) by several authors(Berthelemy, 1986; Hansen, 1991; Perkins, 1997). In agree-ment with Jach et al. (2000), the analyses of the molecular dataplaced the species of this group within the clade of H. ripariaand related groups. The H. minutissima group was hypothe-sized to be related to ‘Haenydra’ by Perkins (1997), whichcould not be confirmed in our analyses.

Finally, the H. holdhausi group (Berthelemy et al., 1991) ismonophyletic and well supported, with the Thyrrhenian species(H. subacuminata from Corsica, H. sardoa from Sardinia, andH. aethaliensis from Elba) being sisters of the rest (Fig. 2B).The molecular-clock estimations suggested that the currentdistribution of the species of this group could not be due tovicariance as a consequence of the fragmentation and rotationof the Sardinian micro-plate from the Iberian Peninsula, whichoccurred some 33–25 Ma ago (Schettino & Turco, 2006). Ourestimations suggest that the H. holdhausi group originated atthe end of the Miocene, with subsequent dispersals during theTortonian and Messinian (Fig. 3), a biogeographical scenariosimilar to that recently proposed for the origin and radiation ofCorso-Sardinian members of the ‘Haenydra’ lineage (Trizzinoet al., 2011).

Some aspects of the phylogeny and diversification of the‘Haenydra’ lineage were discussed in detail by Ribera et al.(2011) and Trizzino et al. (2011). This large derived grouporiginated approx. 9–10 Ma in the Tortonian (Fig. 3). Thereare four main clades – the H. iberica, H. dentipes, H. caucasicaand H. gracilis clades – which, based on our calibration, splitat about 7–8 Ma (Fig. 2C).

The H. iberica clade comprises a group of four species, allendemic to the Iberian Peninsula. According to DNA data thisgroup originated in the late Miocene, although it diversifiedmore recently, in the Plio-/Pleistocene.

Within the H. gracilis clade, which includes more than25 species with an articulated aedeagal distal lobe (Trizzino,2011; Trizzino et al., 2011), our analyses recognized threemonophyletic main groups (Fig. 2C), corresponding to theH. gracilis, H. excisa and H. emarginata complexes. With theexception of H. gracilis and H. excisa, both widely distributedin Europe, the species of this clade are usually highly endemic.Some of the Greek species of the H. gracilis lineage wereplaced in an isolated position, as sister to the rest of the clade:H. jaechiana (Peloponnesus), H. vedrasi (Balkan Peninsula)and H. epeirosi (Greece, including Peloponnesus) (Audisioet al., 1996).

The H. caucasica clade, corresponding to the H. caucasicagroup (sensu Trizzino, 2011), includes the easternmost speciesof ‘Haenydra’. These species are characterized by a relativelyhomogeneous habitus and by male genitalic features. Theirrelationships with the rest of ‘Haenydra’ are not wellestablished (Fig. 2C; Trizzino et al., 2011).

The H. dentipes clade includes a series of well supportedsubclades, corresponding to species complexes, which canbe recognized based on their morphology and distribution:H. dentipes, H. lapidicola, H. truncata, and H. evanescenscomplexes (Ribera et al., 2011; Trizzino et al., 2011).

The H. dentipes complex includes ten species with amostly western Palaearctic distribution. With the exceptionof two relatively isolated species (H. dentipes and H. pro-ducta), the H. dentipes complex is split into two mono-phyletic and allopatric clades including H. heterogyna andallies and H. polita and allies (Ribera et al., 2011; Trizzinoet al., 2011). According to our calibration, these two cladeshad a common origin at the end of the Messinian, although



the relatives of H. heterogyna diversified more recently, around1.4 Ma, whereas for H. polita and allies, the origin was esti-mated in the cold Plio-/Pleistocene transition, approximately2.2 Ma.

The Hydraena lapidicola complex includes a dozen species(Fig. 2C), almost all exhibiting restricted and allopatric geo-graphic ranges. All species are characterized by distinctmale secondary sexual characters (e.g. concerning maxillarypalps and/or legs) and relatively large body sizes. Thereare five monophyletic groups, which, in general, correspondto previously morphologically identified species complexes,including (i) H. hungarica, (ii) H. lapidicola, (iii) H. dev-illei, (iv) H. tatii, and (v) H. monstruosipes and relatives(Trizzino, 2011).

The Hydraena truncata clade includes three species:H. truncata, widespread in the western Palaearctic from Por-tugal to Ukraine, as well as H. muelleri and H. czer-nohorskyi, both endemic to an area comprising southern Aus-tria, northeastern Italy, northwestern Slovenia and northernCroatia. These three species, seemingly well differentiated inexternal morphology (Trizzino, 2011), are of pre-Pleistoceneorigin.

The Hydraena evanescens clade includes seven species char-acterized by a circum-Tyrrhenian distribution. Molecular datasuggest a split of this clade into two subclades correspondingto H. decolor and relatives (Audisio & De Biase, 1995), onone side, and the Corso-Sardinian H. evanescens and its twosister species on the other (Audisio et al., 2009).

Finally, H. solodovnikovi, from southwestern Russia, hasan isolated and uncertain position within the H. dentipesclade.

Supporting Information

Additional Supporting Information may be found in the onlineversion of this article under the DOI reference:10.1111/j.1365-3113.2012.00654.x

Appendix S1. Checklist of the known species of thegenus Hydraena Kugelann, 1794 (as of July 2012),including data on geographical distribution.

Appendix S2. List of studied material, with vouchernumber, locality, collector and accession numbers. In bold,sequences newly included in this study.

Appendix S3. List of primers used for sequencing.

Appendix S4. Phylogenetic trees obtained with (a) raxmlwith prank alignment; (b) mrbayes and mafft alignment;(c) raxml with mafft alignment. Numbers at nodes,Bayesian posterior probability values (b) or bootstrapvalues (a, c).

Appendix S5. Phylogenetic trees obtained with thenuclear markers only and (a) raxml with prank alignment;(b) raxml with mafft alignment.

Acknowledgements

We thank all collectors mentioned in Appendix S2 forspecimens they provided for this study. Visits of MT andIR to the Vienna Natural History Museum were supportedby Synthesys (Applications AT–TAF–53 and AT–TAF–217,respectively). MT is indebted to ‘Unita Analisi & Gestionedelle Risorse Ambientali’ (Insubria University, Varese) andto ZooPlantLab (Bicocca University, Milan) for their kindhospitality. Some analyses were run using the Universityof Oslo Bioportal webserver (www.bioportal.uio.no). Wethank Ana Izquierdo, Amparo Hidalgo, Pedro Abellan andArnaud Faille for laboratory work. Part of the work wasdone at the Museo Nacional de Ciencias Naturales (MNCN,CSIC, Madrid). Funds were provided by Spanish ‘PlanNacional’ projects CGL2007-61665 and CGL2010-15755 toIR.

References

Abellan, P. & Ribera, I. (2011) Geographic location and phylogenyare the main determinants of the size of the geographical range inaquatic beetles. BMC Evolutionary Biology, 11, 344.

Audisio, P. & De Biase, A. (1995) Materiali per un’analisi bio-geografica delle specie italiane dei generi Hydraena Kuge-lann ed Haenydra Rey (Coleoptera, Hydraenidae) (Coleoptera,Hydraenidae). Fragmenta Entomologica, 27, 163–189.

Audisio, P., De Biase, A. & Jach, M.A. (1996) Greek speciesof the genus Hydraena (subgenus Haenydra Rey) (Coleoptera,Hydraenidae). Aquatic Insects, 18, 65–90.

Audisio, P., Trizzino, M., De Biase, A., Mancini, E. & Antonini, G.(2009) A new species of Hydraena (Coleoptera: Hydraenidae) ofthe H. evanescens complex from Sardinia. Zootaxa, 2318, 281–289.

Berthelemy, C. (1986) Remarks on the genus Hydraena and revisionof the subgenus Phothydraena (Coleoptera: Hydraenidae). Annalesde Limnologie, 22, 181–193.

Berthelemy, C., Kaddouri, A.T. & Richoux, P. (1991) Revision of thegenus Hydraena Kugelann, 1794 from North Africa (Coleoptera:Hydraenidae). Elytron, 5, 181–213.

Caterino, M.S., Hunt, T. & Vogler, A.P. (2005) On the constitutionand phylogeny of Staphyliniformia (Insecta: Coleoptera). MolecularPhylogenetics and Evolution, 34, 655–672.

Curtis, J. (1830) British Entomology, Vol. VII. (ed. by J. Curtis).London.

Drummond, A. & Rambaut, A. (2007) Beast: Bayesian evolution-ary analysis by sampling trees. BMC Evolutionary Biology,7, 214.

Drummond, A.J., Ashton, B., Buxton, S. et al. (2011) Geneious v5.4[WWW document]. URL http://www.geneious.com [accessed on 20July 2011].

Freitag, H. & Jach, M.A. (2007) Revision of the species of HydraenaKugelann (Coleoptera: Hydraenidae) from Palawan and Busuanga,with descriptions of eleven new species, and redescription ofHydraena (Hydraenopsis) scabra d’Orchymont, 1925. Zootaxa,1431, 1–44.

Hallam, A. (1994) An Outline of Phanerozoic Biogeography. OxfordUniversity Press, Oxford.

Hansen, M. (1991) A review of the genera of the beetle familyHydraenidae (Coleoptera). Steenstrupia, 17, 1–52.

Hansen, M. (1998) World Catalogue of Insects, Vol. 1: Hydraenidae(Coleoptera). Apollo Books, Stenstrup.



Hebert, P.D.N., Cywinska, A., Ball, S.L. & deWaard, J.R. (2003)Biological identifications through DNA barcodes. Proceedings ofthe Royal Society, B, 270, 313–321.

Hidalgo-Galiana, A. & Ribera, I. (2011) Late Miocene diversificationof the genus Hydrochus (Coleoptera, Hydrochidae) in the westMediterranean area. Molecular Phylogenetics and Evolution, 59,377–385.

Huelsenbeck, J.P. & Ronquist, F. (2001) MrBayes: Bayesian inferenceof phylogenetic trees. Bioinformatics, 17, 754–755.

Hunt, T., Bergsten, J., Levkanicova, Z. et al. (2007) A comprehensivephylogeny of beetles reveals the evolutionary origins of a superra-diation. Science, 318, 1913–1916.

Jach, M.A. (1986) A synopsis of the genus Hydraena (Coleoptera:Hydraenidae) of the Middle East. Israel Journal of Entomology, 20,13–23.

Jach, M.A. (1988) Updating the Hydraena fauna of Turkey(Coleoptera, Hydraenidae). Entomologica Basiliensia, 12, 241–258.

Jach, M.A. (2004) Hydraenidae. Catalogue of Palaearctic Coleoptera,Volume 2: Hydrophiloidea – Histeroidea – Staphylinoidea (ed. byI. Lobl and A. Smetana), pp. 102–122. Apollo Books, Stenstrup.

Jach, M.A. & Balke, M. (2008) Global diversity of water beetles(Coleoptera) in freshwater. Hydrobiologia, 595, 419–442.

Jach, M.A. & Díaz, J.A. (1998) Hydraenidae: I. The Taiwanesespecies of the genus Hydraena Kugelann (Coleoptera). Water Bee-tles of China, Vol. II (ed. by M.A. Jach and L. Ji), pp. 147–171.Zoologisch-Botanische Gesellschaft in Osterreich and WienerColeopterologenverein, Wien.

Jach, M.A. & Díaz, J.A. (2012) New and little known Palearcticspecies of the genus Hydraena (s.l.) Kugelann X. Descriptionsof four new species from southern Europe and taxonomic reviewof the H. subintegra species complex (Coleoptera: Hydraenidae).Koleopterologische Rundschau, in press.

Jach, M.A. & Kasapoglu, A. (2006) Hydraena (s. str.) emineae sp. n.from Antalya, southern Turkey (Coleoptera: Hydraenidae). Zootaxa,1133, 39–43.

Jach, M.A. & Skale, A. (2009) Revision of the Hydraena (s.str.) cir-rata species group (Insecta: Coleoptera: Hydraenidae). Biodiversitatund Naturausstattung im Himalaya, Vol. III (ed. by M. Hartmannand J. Weipert), pp. 181–236. Verein der Freunde & Forderer desNaturkundemuseums Erfurt e.V., Erfurt.

Jach, M.A., Beutel, R.G., Díaz, J.A. & Kodada, J. (2000) Subgenericclassification, description of head structures, and world checklist of Hydraena Kugelann (Insecta: Coleoptera: Hydraenidae).Annalen des Naturhistorischen Museums in Wien, Series B, 102,177–258.

Jach, M.A., Beutel, R.G., Delgado, J.A. & Díaz, A. (2005)Hydraenidae. Coleoptera, Beetles, Vol. 1 (ed. by R.G. Beutel andR.A.B. Leschen), pp. 224–250. W. de Gruyter, Berlin.

Janssens, E. (1972) Essai sur la systematique des Hydraena desregions intertropicales. Bulletin et Annales de la Societe RoyaleBelge de l’Entomologie, 108, 253–261.

Katoh, K. & Toh, H. (2008) Recent developments in the MAFFTmultiple sequence alignment program. Briefings in Bioinformatics,9, 286–298.

Kugelann, J.G. (1794) Verzeichniss der in einigen Gegenden Preussensbis jetzt entdeckten Kafer-Arten, nebst kurzen Nachrichten vondenselben. Neustes Magazin fur die Liebhaber der Entomologie,Herausgegeben von D.H. Schneider, 1, 513–582.

Kuwert, A. (1888) Generalubersicht der Hydraenen der europaischenFauna. Deutsche Entomologische Zeitschrift, 32, 113–123.

Loytynoja, A. & Goldman, N. (2005) An algorithm for progressivemultiple alignment of sequences with insertions. Proceedings of theNational Academy of Sciences of the United States of America, 102,10557–10562.

Mulsant, E. (1844) Histoire naturelle des coleopteres de France.Palpicornies. L. Maison, Paris.

Oaks, J.R. (2011) A time-calibrated species tree of Crocodylia revealsa recent radiation of the true Crocodiles. Evolution, 65, 3285–3297.

Perkins, P.D. (1980) Aquatic beetles of the family Hydraenidae inthe Western Hemisphere: classification, biogeography and inferredphylogeny (Insecta, Coleoptera). Quaestiones Entomologicae, 16,3–554.

Perkins, P.D. (1989) Adelphydraena, new genus, and two new speciesfrom Venezuela, and remarks on phylogenetic relationships withinthe subtribe Hydraenina (Coleoptera: Hydraenidae). Proceedings ofthe Biological Society of Washington, 102, 447–457.

Perkins, P.D. (1997) Life on the effective bubble: exocrine secretiondelivery systems (ESDS) and the evolution and classification ofbeetles in the family Hydraenidae (Insecta: Coleoptera). Annals ofthe Carnegie Museum, 66, 89–207.

Perkins, P.D. (2007) A revision of the Australian species of thewater beetle genus Hydraena Kugelann (Coleoptera: Hydraenidae).Zootaxa, 1489, 1–207.

Perkins, P.D. (2011a) New species (130) of the hyperdiverse aquaticbeetle genus Hydraena Kugelann from Papua New Guinea,and a preliminary analysis of areas of endemism (Coleoptera:Hydraenidae). Zootaxa, 2944, 1–417.

Perkins, P.D. (2011b) New records and description of fifty-four newspecies of aquatic beetles in the genus Hydraena Kugelann fromSouth America (Coleoptera: Hydraenidae). Zootaxa, 3074, 1–198.

Posada, D. (2008) JModelTest phylogenetic model averaging. Molec-ular Biology and Evolution, 25, 1253–1256.

de Queiroz, A. (2005) The resurrection of oceanic dispersal inhistorical Biogeography. Trends in Ecology & Evolution, 20, 68–73.

Rey, C. (1886) Histoire naturelle des Coleopteres de France (suite).Annales de la Societe Linneenne de Lyon, 32, 1–187, pl. 1–2.

Ribera, I., Castro, A. & Hernando, C. (2010a) Ochthebius (Enico-cerus) aguilerai sp.n. from central Spain, with a molecular phy-logeny of the Western Palaearctic species of Enicocerus (Coleoptera,Hydraenidae). Zootaxa, 2351, 1–13.

Ribera, I., Fresneda, J., Bucur, R., Izquierdo, A., Vogler, A.P., Sal-gado, J.M. & Cieslak, A. (2010b) Ancient origin of a westernMediterranean radiation of subterranean beetles. BMC EvolutionaryBiology, 10, 29.

Ribera, I., Castro, A., Díaz, J.A., Garrido, J., Izquierdo, A., Jach,M.A. & Valladares, L.F. (2011) The geography of speciation innarrow range endemics of the “Haenydra” lineage (Coleoptera,Hydraenidae, Hydraena). Journal of Biogeography, 38,502–516.

Rowe, D.L., Dunn, K.A., Adkins, R.M. & Honeycutt, R.L. (2010)Molecular clocks keep dispersal hypotheses afloat: evidence fortrans-Atlantic rafting by rodents. Journal of Biogeography, 37,305–324.

Schettino, A. & Turco, E. (2006) Plate kinematics of the westernMediterranean Region during the Oligocene and Early Miocene.Geophysical Journal International, 166, 1398–1423.

Seidlitz, G. (1888) Fauna Baltica. Die Kaefer (Coleoptera) derdeutschen Ostseeprovinzen Russlands, 2nd edn. Hartungsche Ver-lagsdruckerei, Konigsberg.

Simaiakis, S. & Mylonas, M. (2008) The Scolopendra species(Chilopoda: Scolopendromorpha: Scolopendridae) of Greece (E-Mediterranean): a theoretical approach on the effect of geogra-phy and palaeogeography on their distribution. Zootaxa, 1792,39–53.

Stamatakis, A., Hoover, J. & Rougemont, J. (2008) A rapid bootstrapalgorithm for the RAxML web–servers. Systematic Biology, 75,758–771.



Townsend, T.M., Leavitt, D.H. & Reeder, T.W. (2011) Intercontinen-tal dispersal by a microendemic burrowing reptile (Dibamidae).Proceedings of the Royal Society, B, 278, 2568–2574.

Trizzino, M. (2011) Molecular phylogeny, taxonomy and biogeogra-phy of the “Haenydra” lineage (Coleoptera, Hydraenidae, genusHydraena). PhD Dissertation, University of Rome “Sapienza”,Department of Biology and Biotechnologies “C. Darwin”,Rome.

Trizzino, M., Audisio, P., Antonini, G., Mancini, E. & Ribera, I.(2011) Molecular phylogeny and diversification of the “Haenydra”lineage (Hydraenidae, genus Hydraena), a north-Mediterraneanendemic-rich group of rheophilic Coleoptera. Molecular Phyloge-netics and Evolution, 61, 772–783.

Accepted 15 August 2012First published online 22 October 2012


Documents

A molecular phylogeny of the cosmopolitan hyperdiverse ...molevol.cmima.csic.es/ribera/pdfs/Trizzino_etal2013_Hydraena.pdf · Systematic Entomology (2013), 38, 192–208 DOI: 10.1111/j.1365-3113.2012.00654.x