Molecular Phylogenetics and Evolution 54 (2010) 97–106

Contents lists available at ScienceDirect

Molecular Phylogenetics and Evolution

journal homepage: www.elsevier.com/locate/ympev

A molecular phylogeny shows the single origin of the Pyrenean subterranean ground (Coleoptera: Carabidae)

A. Faille a,b,*, I. Ribera b,c, L. Deharveng a, C. Bourdeau d, L. Garnery e, E. Quéinnec f, T. Deuve a a Département Systématique et Evolution, ‘‘Origine, Structure et Evolution de la Biodiversité” (C.P.50, UMR 7202 du CNRS/USM 601), Muséum National d’Histoire Naturelle, Bât. Entomologie, 45 rue Buffon, F-75005 Paris, b Institut de Biologia Evolutiva (CSIC-UPF), Passeig Maritim de la Barceloneta 37-49, 08003 Barcelona, Spain c Museo Nacional de Ciencias Naturales (CSIC), José Gutiérrez Abascal 2, 08006 Madrid, Spain d 5 chemin Fournier-Haut, F-31320 , France e Laboratoire Evolution, Génomes, Spéciation, CNRS UPR9034, Gif-sur-Yvette, France f Unité ‘‘Evolution & Développement”, UMR 7138 ‘‘Systématique, Adaptation, Evolution”, Université P. & M. Curie, 9 quai St–Bernard, F-75005 Paris, France article info abstract

Article history: Trechini ground beetles include some of the most spectacular radiations of cave and endogean Coleoptera, Received 16 March 2009 but the origin of the subterranean taxa and their typical morphological adaptations (loss of eyes and Revised 1 October 2009 wings, depigmentation, elongation of body and appendages) have never been studied in a formal phylo- Accepted 5 October 2009 genetic framework. We provide here a molecular phylogeny of the Pyrenean subterranean Trechini based Available online 21 October 2009 on a combination of mitochondrial (cox1, cyb, rrnL, tRNA-Leu, nad1) and nuclear (SSU, LSU) markers of 102 specimens of 90 species. We found all Pyrenean highly modified subterranean taxa to be monophyletic, to Keywords: the exclusion of all epigean and all subterranean species from other geographical areas (Cantabrian and Subterranean environment Iberian mountains, Alps). Within the Pyrenean subterranean clade the three genera (Geotrechus, Convergence Endogean Aphaenops and Hydraphaenops) were polyphyletic, indicating multiple origins of their special adaptations Troglobitic to different ways of life (endogean, troglobitic or living in deep fissures). Diversification followed a geographical pattern, with two main clades in the western and central-eastern Pyrenees respectively, Aphaenops and several smaller lineages of more restricted range. Based on a Bayesian relaxed-clock approach, and using as an approximation a standard mitochondrial mutation rate of 2.3% MY, we estimate the origin of the subterranean clade at ca. 10 MY. Cladogenetic events in the Pliocene and Pleistocene were almost exclusively within the same geographical area and involving species of the same morphological type. Ó 2009 Elsevier Inc. All rights reserved.

1. Introduction (Caccone, 1985). Amongst , many groups of Coleoptera have repeatedly colonised subterranean habitats, but two of them are The origin and evolution of cave organisms has fascinated evo- particularly diverse: Leiodidae (especially subfamily Cholevinae) lutionists and biologists for more than two hundred years, since in the suborder Polyphaga, and Carabidae of the subfamily Trechi- the discovery of the first troglobitic species (Proteus anguinus, de- nae in the suborder (Casale et al., 1998). Subterranean scribed by Laurenti, 1768). Organisms living in a subterranean species of both groups share morphological modifications consid- environment tend to show a highly modified morphology and biol- ered to be adaptations to a subterranean lifestyle: loss of metatho- ogy, and a mixture of losses (eye degeneration, depigmentation) racic wings, eyes and pigment, similar changes in body shape and and adaptations (development of sensory organs, changes in the size (Jeannel, 1926a,b; Vandel, 1964; Barr and Holsinger, 1985), life cycle and metabolism, body shape modifications) (Racovitza, and modifications in their way of life (Deleurance, 1958). The 1907; Vandel, 1964; Culver et al., 1990). Troglobitic invertebrates extensive convergence in morphological characters obscures the isolated in karstic areas are also very good models to study speci- phylogenetic relationships among species (Marquès and Gnaspini, ation and diversification, because of the isolation of populations in 2001; Desutter-Grandcolas et al., 2003), which has resulted in a well-defined karstic units with highly restricted gene flow high number of taxonomic arrangements with non-monophyletic taxa (see e.g. Fresneda et al., 2007 for an example with a lineage of Leiodidae cave beetles). * Corresponding author. The Pyrenean Chain is known to be one of the main world hot- E-mail addresses: [email protected], [email protected] (A. Faille), igna spots for subterranean invertebrate fauna (Culver et al., 2006). The [email protected] (I. Ribera), [email protected] (L. Deharveng), Lionel.Gar [email protected] (L. Garnery), [email protected] (E. Quéinnec), deuve phylogenetic relationships among the subterranean species of @mnhn.fr (T. Deuve). Pyrenean Trechini, one of the groups which have experienced

1055-7903/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.ympev.2009.10.008 98 A. Faille et al. / Molecular Phylogenetics and Evolution 54 (2010) 97–106 extensive diversification in the area (Jeannel, 1941), are poorly species (A. schmidtii Sturm). Putzeys (1870) transferred these known, and studies have so far been based on morphological char- species to the genus Clairville, which also includes epigean acters only (Jeannel, 1941; Casale et al., 1998; see below). species. Bonvouloir (1862) erected the genus Aphoenops for the The subterranean Trechini of the Pyrenees include ca. 80 species subterranean species A. leschenaulti Bonvouloir on the basis of the in three genera, Geotrechus, Aphaenops and Hydraphaenops (Mora- non-dilated protarsi of the male, a character currently considered vec et al., 2003; see Appendix for nomenclatorial remarks). All of reduced phylogenetic relevance (Bedel and Simon, 1875). See the species are Pyrenean endemics (most of them with very nar- the appendix for the use of Aphaenops in place of Aphoenops. row distributions), with different adaptations for life in subterra- The current concept of the genus Aphaenops includes 41 species nean habitats. They are all completely blind and apterous, with a on both sides of the Pyrenees, all of them highly modified and slender body form, and (in some species) an extreme elongation exclusive to karst areas, living either in deep cavities or, in some of the head, pronotum and appendages (Jeannel, 1941; Casale cases, in the Superficial Hypogean Compartment (‘‘Milieu souter- et al., 1998), resulting in a very characteristic appearance, the rain superficiel”, MSS, Juberthie and Bouillon, 1983). Diagnostic ‘‘aphaenopsian” morphological type (Jeannel, 1941; Vandel, characters are the presence of incomplete frontal furrows (vs. com- 1964)(Fig. 1). Many subterranean insects around the world have plete in Geotrechus), very elongated legs and antennae, body pale, independently developed similar characteristics, and ‘‘aphaenop- completely depigmented, and a pronounced narrowing (a ‘‘neck”) sian”, ‘‘aphaenopsoid” or ‘‘Aphaenops-like” is commonly used to re- at the base of the head (Coiffait, 1962)(Fig. 1). It is subdivided in fer to this syndrome in other groups of Carabidae (Barr, 1979; six subgenera (for the taxonomic ordination of the group we follow Deuve, 2001; Ortuño et al., 2004; Uéno and Clarke, 2007), and even the recent catalogue of Moravec et al., 2003, although we do not other insects (e.g. Hymenoptera, Roncin and Deharveng, 2003). consider subspecies unless otherwise stated): In this study we provide for the first time a phylogenetic frame- work obtained with numerical algorithms to study the origin and (1) Aphaenops Bonvouloir, 1862: 10 species, mainly found in the diversification of the subterranean species of Pyrenean Trechini, western Pyrenees. based on a combination of nuclear and mitochondrial genes. We (2) Geaphaenops Cabidoche, 1965: 7 species, also in the western include a broad sample of the three subterranean genera (51 spe- Pyrenees. All the species of this group seem to be endogean, cies, some with repeated examples), plus a representation of other and their external morphology is very homogeneous. troglobitic species and potential relatives living on the surface in (3) Cerbaphaenops Coiffait, 1962: 16 species, mainly found in the the Pyrenees and other west Mediterranean areas. Our specific central and eastern Pyrenees, between Bagnères-de-Bigorre aims were to (1) determine the origin of the subterranean genera and the Ariège River. This is also a group with a very homo- and their relationships with epigean species, (2) investigate the geneous morphology, although no clear diagnostic charac- monophyly of traditional taxa (genera and subgenera), established ters were given by Coiffait (1962) (pubescent head, short on external morphological characters, and (3) investigate the rela- mandibles). tionship between endogean and cave species. (4) Pubaphaenops Genest, 1983: a single species from a cave in Ariège, A. laurenti Genest, fully pubescent. 2. Materials and methods (5) Arachnaphaenops Jeanne, 1967: three species, one in the western Pyrenees, two in Ariège and Haute-Garonne respec- 2.1. Historical and taxonomic background of Pyrenean subterranean tively, all with very long legs and antennae, which give them Trechini the appearance of an arachnid. (6) Cephalaphaenops Coiffait, 1962: two species, one in the wes- The first known Pyrenean cave ground-beetles were included in tern Pyrenees, the other in Ariège and Haute-Garonne, with the genus Anophthalmus, created for an eastern Alpine hypogean a large and pubescent head and long mandibles.

Fig. 1. Habitus of (1) Aphaenops alberti Jeannel (troglobitic), (2) Aphaenops pluto Dieck (troglobitic), (3) Hydraphaenops navaricus Coiffait & Gaudin (troglobitic), (4) Geotrechus seijasi Español (endogean), and (5) (Schrank) (epigean). Scale bars, 1 mm. Photos 1–3 P. Déliot, 4 A. Faille, 5 U. Schmidt. A. Faille et al. / Molecular Phylogenetics and Evolution 54 (2010) 97–106 99

The genus Geotrechus was created by Jeannel (1919) for some Table 1 Checklist of genera and subgenera of subterranean species of Pyrenean Trechini, with blind species with an ‘‘Anophthalmus-like” habitus, in opposition total number of species and species included in the study. follows Moravec to the species of Aphaenops. Diagnostic characters of Geotrechus et al. (2003) updated. are the presence of frontal furrows and their more robust appear- Genus Subgenus N. spp. Sampled spp. ance, with short legs and antennae (Coiffait, 1962; Fig. 1). Jeannel (1919) considered the genera Aphaenops and Geotrechus as two clo- Aphaenops Aphaenops 10 8 Aphaenops Geaphaenops 72 sely related but distinct lineages. Most of the 22 known species of Aphaenops Cerbaphaenops 16 14 Geotrechus are endogean, and although some populations can be Aphaenops Cephalaphaenops 21 locally abundant in caves, most of them seem to be more common Aphaenops Arachnaphaenops 33 in the ground at the entrance of the cavities (Jeannel, 1926b, 1941). Aphaenops Pubaphaenops 11 Some species can also be found under large stones in forests, when Hydraphaenops Hydraphaenops 18 8 Geotrechus Geotrechus 84 hydric conditions are favourable. Geotrechus Geotrechidius 15 6 Hydraphaenops was first described as a subgenus of Aphaenops Trechus Trechus 17a 7 by Jeannel (1926a), and subsequently upgraded by Coiffait a Species occurring in the Pyrenees, 11 of them endemic. (1962). Currently it includes 18 species, characterised by an elon- gated and parallel-sided, almost cylindrical head, sharp, sickle- shaped mandibles, short appendages and the body at least partially covered with pubescence (Jeannel, 1941; Coiffait, 1962)(Fig. 1). We sequenced three mitochondrial (50 end of cytochrome c oxi- Most species are exceedingly rare, some of them being known from dase subunit 1, cox1; cytochrome b, cyb,50 end of large ribosomal only one or two specimens, and their biology is virtually unknown unit plus the Leucine transfer plus the 30 end of NADH dehydroge- (Cabidoche, 1966). They do seem to be highly hygrophilous, requir- nase subunit 1, rrnl+tRNA-Leu+nad1) and two nuclear (small ribo- ing a water-saturated atmosphere to colonise karstic areas (Deleu- somal unit, SSU, large ribosomal unit, LSU) gene fragments (see rance-Glaçon, 1963). Some species are known at low altitude (e.g. Table 2 for the primers used). Sequences were assembled and edi- H. galani Español, found at sea level), while others have only been ted with Bioedit v. 7.00 (Hall, 1999) or Sequencher 4.6 (Gene found in high altitude shafts in direct contact with ice (e.g. H. pena- Codes, Inc., Ann Arbor, MI). New sequences have been deposited collaradensis Dupré, H. mouriesi Genest) (Español, 1968; Dupré, in GenBank with Acc. Nos. GQ293502–GQ293896 (395 sequences) 1991; Genest, 1983). As happens with Aphaenops, Hydraphaen- (Suppl. Table 1). For some species, the final sequence is a chimera ops-like species are known in other lineages of subterranean Tre- of sequences obtained from different specimens (labelled with the chinae (Deuve, 2000; Casale, 2004). two voucher numbers in all Figures, see Suppl. Table 1). Protein coding genes were not length variable, and the ribosomal genes were aligned with the online version of MAFFT v.6 using the G- 2.2. Taxon sampling INS-i algorithm and default parameters (Katoh et al., 2002; Katoh and Toh, 2008). Trechini species were collected in caves, shafts and MSS from the Pyrenean chain, in France and Spain as listed in Suppl. Table 1. Single individuals were used for amplification and sequencing. 2.4. Phylogenetic analyses We included as outgroups several examples of Trechus from the Pyrenees (mainly epigean, some of them hypogean), plus some Bayesian analyses were conducted on a combined data matrix other genera from different geographical areas, including both epi- with MrBayes 3.1.2 (Huelsenbeck and Ronquist, 2001), using five gean and subterranean species (Suppl. Table 1). To root the tree we partitions corresponding to the five sequenced fragments. Evolu- used one species of Anillini (Typhlocharis Dieck) and two of Bem- tionary models were estimated prior to the analysis with Model- bidiini ( Netolitzky and Philochthus Stephens), which Test 3.7 (Posada and Crandall, 1998). MrBayes ran for 6 106 are clearly outside Trechini (Grebennikov and Maddison, 2005; generations using default values, saving trees each 500. ‘‘Burn-in” Grebennikov, 2008). In total, we sampled 50 specimens of 32 spe- values were established after visual examination of a plot of the cies of Aphaenops, 11 specimens of 9 species of Hydraphaenops and standard deviation of the split frequencies between two simulta- 12 specimens of 10 species of Geotrechus (Table 1; Suppl. Table 1). neous runs. We used two additional phylogenetic approaches for compara- 2.3. DNA extraction, PCR amplification and sequencing tive purposes, maximum likelihood with a genetic algorithm implemented in Garli v0.9 (Zwickl, 2006), using an estimated Specimens were collected alive in the field and directly killed GTR+I+G model for the combined sequence and the default set- and preserved in 96% ethanol. DNA was extracted from whole spec- tings, and parsimony in PAUP v4.b10 (Swofford, 2002), with imens by a standard phenol–chloroform extraction (Blin and Staf- 10,000 random replicates, swapping on best trees only and not sav- ford, 1976). DNA extraction was usually non-destructive, to ing multiple trees. Node support was measured with the posterior preserve voucher specimens for subsequent morphometric and probabilities in MrBayes, and 1000 bootstrap replicates (Felsen- morphological study (Pons, 2006; Gilbert et al., 2007; Rowley stein, 1985) in Garli and PAUP. To reduce computation time in Gar- et al., 2007). Specimens were incubated overnight in a mix of li, the number of generations without improving the topology 500 ll of buffer (10 mM Tris, pH 8.0; 0.5% SDS; 0.1 M EDTA, pH necessary to complete each replica was reduced to 5000 instead 8.0) and 25 ll of proteinase K (20 mg/ml) at 55 °C, with the abdom- of the default 10,000. In PAUP we performed heuristic searches inal ventrites slightly opened to facilitate the action of the digestion with random addition of taxa with 10 repetitions for each of enzyme. The use of non-destructive methods allowed the molecular 1000 replications. Differences between alternative topologies were study of very rare species, as even fragile structures of taxonomic evaluated using the tests of Templeton (1983) for parsimony and importance, like the chaetotaxy or the internal structures of the Shimodaira and Hasegawa (1999) for maximum likelihood. aedeagus, were perfectly preserved after extraction (Pons, 2006; To check for possible topological incongruences we did maxi- Gilbert et al., 2007; Rowley et al., 2007). Voucher specimens are mum likelihood analyses in Garli with the nuclear sequence alone, kept in the MNHN (Paris), DNA aliquots are kept in the tissue collec- using the GTR+I+G evolutionary model and estimating node sup- tions of the MNHN and IBE (CSIC-UPF, Barcelona). port with 1000 bootstrap replicas as above. 100 A. Faille et al. / Molecular Phylogenetics and Evolution 54 (2010) 97–106

Table 2 Primers used in the study.

Marker Primer Sequence Ref. cox1 RON 50GGATCACCTGATATAGCATTCCC30 Simon et al. (1994) HOBBES 50AAATGTTNGGRAAAAATGTTA30 Monteiro and Pierce (2001) TONYA 50GAAGTTTATATTTTAATTTTACCGG30 Monteiro and Pierce (2001) NANCY 50CCCGGTAAAATTAAAATATAAACTTC30 Simon et al. (1994) JERRY 50CAACATTTATTTTGATTTTTTGG30 Simon et al. (1994) PAT 50TCCAATGCACTAATCTGCCATATTA30 Simon et al. (1994) cyb CB1 50TATGTACTACCATGAGGACAAATATC30 Simon et al. (1994) CP1 50GATGATGAAATTTTGGATC30 Kergoat (pers. comm., 2004) TSERco 50TATTTCTTTATTATGTTTTCAAAAC30 Simon et al. (1994) rrnl+tRNA-Leu+nad1 NDIA 50GGTCCCTTACGAATTTGAATATATCCT30 Simon et al. (1994) 16SaR 50CGCCTGTTTATCAAAAACAT30 Simon et al. (1994) LSU D3 50GCATAGTTCACCATCTTTC30 Ober (2002) D1 50GGGAGGAAAAGAAACTAAC30 Ober (2002) SSU 18S-50 50GACAACCTGGTTGATCCTGCCAGT30 Shull et al. (2001) 18S-b5.0 50TAACCGCAACAACTTTAAT30 Shull et al. (2001)

2.5. Estimation of divergence times 14 bp between Geotrechus vandeli and some species of Aphaenops (A. alberti, A. cabidochei, A. ochsi). To estimate the relative age of divergence of the lineages we The optimal evolutionary model for the mitochondrial genes, as used the Bayesian relaxed phylogenetic approach implemented measured with Modeltest under the Akaike information criterion, in BEAST v1.4.7 (Drummond and Rambaut, 2007), which allows was GTR+I+G. For the SSU the optimal model was TVMef+I, and variation in substitution rates among branches (Drummond et al., for the LSU TVM+I+G. The runs of MrBayes converged at ca. 2006). We implemented a GTR+I+G model of DNA substitution 2 106 generations, with a standard deviation of the split frequen- with four rate categories using the mitochondrial data set only cies between the two runs of ca. 0.015. The two runs were inter- and pruning species with more than one missing gene fragment, rupted at 5 106 generations (see the estimated parameters in with an uncorrelated lognormal relaxed molecular clock model Suppl. Table 2). A heuristic search using PAUP and assuming an to estimate substitution rates and the Yule process of speciation equal weight for all characters resulted in 2151 trees of 4537 steps as the tree prior. The main nodes of the topology were constrained (consistency index, CI = 0.41, retention index, RI = 0.60). to match that of the tree obtained with the whole dataset (mito- The topology of the tree, and the support for the main nodes, chondrial plus nuclear) in MrBayes. We ran two independent anal- were very similar for the three reconstruction methods (Bayesian yses for each group, sampling each 500 generations, and used probabilities, maximum likelihood and parsimony) (Fig. 2, Suppl. TRACER version 1.4 to determine convergence, measure the effec- Fig. 1). In all cases the three subterranean genera of the Pyrenees tive sample size of each parameter, and calculate the mean and (Aphaenops, Hydraphaenops and Geotrechus) formed a clade with 95% highest posterior density interval (HPD) for divergence times. exclusion of all epigean species, with strong support (Fig. 2, Suppl. Results of the two runs were combined with LogCombiner v1.4.7 Fig. 1). The Pyrenean subterranean lineage was sister to a poorly and the consensus tree compiled with TreeAnnotator v1.4.7 supported clade including all species of Trechus of different areas (Drummond and Rambaut, 2007). (including the Pyrenees), plus some other subterranean taxa out- The analyses were run for 25 106 generations, with the initial side the Pyrenees (Apoduvalius, Cantabrian mountains; Duvalius, 10% discarded as burn-in. Because of the absence of fossil record Alps; Paraphaenops, Iberian System) (Suppl. Table 1; Fig. 2, Suppl. for both groups, to calibrate the trees we used as a prior a normal Fig. 1). Basal relationships within this clade were not supported. distribution with average equal to the standard rate of 2.3% MY, Within the Pyrenean subterranean clade, the three genera were equivalent to a per-branch rate of 0.0115 substitutions/site/MY polyphyletic under all reconstruction methods, with at least one (Brower, 1994), and a standard deviation of 0.0001. This rate is well-supported node determining the polyphyly in each case lower to that obtained by Contreras-Díaz et al. (2007) for the genus (Fig. 2). We constrained the monophyly of the three genera and Trechus, using calibration points based on the colonisation of the searched the best topology compatible with this constrain both Canary islands (0.015 substitutions/site/MY), although the later in PAUP using parsimony and in Garli with maximum likelihood. was based on cox1 and cox2 only, which have faster evolutionary The search in PAUP with the constraint of the monophyly of the rates than the ribosomal rrnL (Ribera et al., 2001). three subterranean genera resulted in 227 trees of 4691 steps (CI = 0.39; RI = 0.57). The resulting topologies were significantly 3. Results worse, as tested both for parsimony (Templeton test, p < 0.0001) and maximum likelihood (Shimodaira–Hasegawa test, p < 0.0005). 3.1. Phylogenetic analysis The basal nodes of the subterranean clade were not well-sup- ported, but the best topologies in Bayesian analyses and maximum The aligned data matrix had 3653 characters, of which 932 were likelihood placed a paraphyletic series of species of Geotrechus parsimony informative. There was no length variation in the pro- from the Eastern Pyrenees at the base (Fig. 2), included in the sub- tein coding genes, and variation in the ribosomal genes was mostly genus Geotrechidius (the ‘‘vulcanus group” sensu Coiffait, 1962). The concentrated in the LSU, ranging from 862 (Typhlocharis) to 909 bp rest of the species were included in two main well-supported (Perileptus), both among the outgroups. Length variation in the in- clades (pp = 1, bootstrap >70% in all analyses) plus some western group LSU was reduced to between 867 (G. saulcyi, G. seijasi) and lineages of Hydraphaenops and Geotrechus (Figs. 2 and 3). The 898 bp (Hydraphaenops galani, H. delicatulus). For the SSU fragment two well-supported main lineages were (1) species of Aphaenops there was only three bp maximum length difference, and for the distributed in the western Pyrenees (clade W), and (2) a clade of rrnL+tRNA-LEU fragment the maximum length difference was species of Aphaenops and Hydraphaenops from the eastern Pyrenees A. Faille et al. / Molecular Phylogenetics and Evolution 54 (2010) 97–106 101

69/0.97/78 80/1/84 A. vandeli MNHN-AF43 x/0.77/63 A. vandeli MNHN-AF44 A. vandeli MNHN-AF45 59/0.96/67 A. bouiganensis MNHN-AF46 54/0.74/x A. crypticola MNHN_AF52 84/0.99/79 A. parallelus MNHN_AF53 97/1/95 61/0.72/70 A. crypticola MNHN_AF49 67/0.96/80 A. crypticola MNHN_AF47 76/1/69 A. crypticola MNHN_AF48 72/1/70 A. crypticola MNHN_AF51 A. bouilloni MNHN_AF56 75/0.98/87 A. mariarosae MNHN_AF57 x/1/59 85/-/71 A. crypticola MNHN_AF50 68/0.62/67 A. sp MNHN_AF133 A. sioberae MNHN_AF54 x/0.99/- A. pluto MNHN_AF58 x/0.57/x A. carrerei MNHN_AF34 x/0.9/- A. laurenti MNHN_AF63 x/0.94/- A. michaeli MNHN_AF35 97/1/95 A. bonneti MNHN_AF38 A. delbreili MNHN_AF37 100/1/99 A. cerberus MNHN_AF20_AF30 56/0.77/x A. jauzioni MNHN_AF33 57/0.79/x A. crypticola MNHN_AF135 x/0.64/61 52/1/53 A. hustachei MNHN_AF39 86/1/x A. aeacus MNHN_AF40 95/1/97 A. crypticola MNHN_AF134 x/0.88/69 A. sp MNHN_AF42 83/1/73 A. bucephalus MNHN_AF62 100/1/100 . A. chappuisi MNHN_AF61 Eastern clade 98/1/89 A. tiresias MNHN_AF59_AF60 x/1/- H. bourgoini MNHN_AF68 H. bourgoini MNHN_AF69 100/1/100 H. ehlersi MNHN_AF64 x/0.55/- H. pecoudi MNHN_AF72 H. elegans MNHN_AF120 86/0.99/69 H. penacollaradensis MNHN_AF121 100/1/99 A. abodiensis MNHN_AF4 A. bessoni MNHN_AF122 100/1/100 A. loubensi MNHN_AF3 59/0.95/- A. ludovici MNHN_AF15 100/1/100 A. rhadamanthus MNHN_AF13_AF14 98/1/96 A. jeanneli MNHN_AF11 66/0.99/63 A. orionis MNHN_AF9_AF10 97/1/88 A. alberti MNHN_AF12 58/0.95/- . A. cabidochei MNHN_AF5_AF6 Western clade 74/0.99/76 100/1/100 A. ochsi MNHN_AF7_AF8 x/0.63/- A. catalonicus MNHN_AF2 63/1/57 A. leschenaulti MNHN_AF1 H. galani MNHN_AF67 77/1/78 H. vasconicus MNHN_AF65 99/1/93 71/0.92/x G. jeanneli MNHN_AF77 G. gallicus MNHN_AF76 x/0.63/x H. pandellei MNHN_AF70 100/1/99 100/1/100 H. pandellei MNHN_AF71 x/1/- 99/1/97 G. discontignyi MNHN_AF92 G. orcinus MNHN_AF85 59/0.81/x G. trophonius MNHN_AF83 H. delicatulus MNHN_AF66 G. orpheus MNHN_AF79_AF81 Pyrenean hypogean 99/0.78/100 G. saulcyi MNHN_AF86 .100/1/x clade G. saulcyi MNHN_AF87 G. vandeli MNHN_AF88 92/1/80 G. vulcanus MNHN_AF91 G. seijasi MNHN_AF89 64/0.94/x T. escalerae MNHN_AF104 x/0.53/x T. saxicola MNHN_AF100 68/1/58 A. alberichae MNHN_AF105 72/0.99/70 T. navaricus MNHN_AF103 x/0.53/- T. uhagoni MNHN_AF102 53/1/92 x/0.84/x Apoduvalius sp MNHN_AF106 x/0.61/x T. fulvus MNHN_AF98 94/1/97 T. barnevillei MNHN_AF97 86/1/71 T. obtusus MNHN_AF126 87/1/71 T. ceballosi MNHN_AF128 100/1/100 T. distigma MNHN_AF94 x/1/51 T. quadristriatus MNHN_AF96 S. mayeti MNHN_AF107 73/0.98/97 100/1/99 T. comasi MNHN_AF127 T. schaufussi MNHN_AF101 x/0.91/x P. breuilianus MNHN_AF108 x/0.68/x Agostinia gaudini MNHN_AF116 90/1/x x/0.6/x D. berthae MNHN_AF114_AF115 100/1/100 D. roberti MNHN_AF129 A. robini MNHN_AF112 x/0.86/x I. bolivari MNHN_AF111 L. deharvengi MNHN_AF117 P. areolatus MNHN_AF113 x/0.7/x P. bisulcatus MNHN_AF131 P. lunulatus MNHN_AF118 Typhlocharis MNHN_AF119 x/1/x

0.09

Fig. 2. Phylogram of subterranean Trechini of the Pyrenees obtained with maximum likelihood in Garli, using the combined data matrix. Number in nodes, ML bootstrap/ Bayesian posterior probability, obtained in MrBayes/parsimony bootstrap (see Section 2 for details). ‘‘Western” and ‘‘Eastern” clades marked with ‘‘W” and ‘‘E” respectively (see text). In red, species of Aphaenops; in green, species of Hydraphaenops; in blue, species of Geotrechus. Habitus, from top to bottom: Aphaenops pluto, A. bessoni, A. alberti, Hydraphaenops galani, Geotrechus gallicus, G. seijasi, Trechus sp., Paraphaenops breuilianus, Duvalius berthae (see Suppl. Table 1). (For interpretation of colour mentioned in this figure the reader is referred to the web version of the article.) 102 A. Faille et al. / Molecular Phylogenetics and Evolution 54 (2010) 97–106

Fig. 3. Distribution of the main clades of subterranean Trechini of the Pyrenees, according to the phylogeny in Fig. 2. ‘‘Western” and ‘‘Eastern” clades marked with ‘‘W” and ‘‘E” respectively (see text).

(clade E). The Eastern group of Aphaenops species corresponds site MY. The estimated age of the origin of the subterranean clade mostly to the subgenus Cerbaphaenops sensu Coiffait (1962), plus was 9.7MY, with a 95% interval of confidence between 7.6 and some morphologically characteristic species so far placed in the 12.2MY (Fig. 4). The origin of the main clades (Eastern and Wes- subgenera Arachnaphaenops, Cephalaphaenops and Pubaphaenops tern), and that of the different lineages within each genus, was esti- (A. bucephalus, A. laurenti, A. chappuisi, A. pluto and A. tiresias; Suppl. mated to be in the Upper Miocene, before the end of the Messinian Table 1). (Fig. 4). Cladogenetic events within the Pliocene and Pleistocene The Western clade of Aphaenops included all species of the sub- were almost exclusively within the same geographical area and genus Geaphaenops (forming a monophyletic lineage) plus species involving species of the same morphological type (i.e. within lin- of Aphaenops s.str. and A.(Arachnaphaenops) alberti. The eastern- eages of each of the traditional genera) (Fig. 4). most species of this clade is Aphaenops catalonicus, which is also the southern-most species of Aphaenops, with records from the 4. Discussion Pre-Pyrenees in the Ribagorza valley (Fig. 3). It has morphological affinities to the northern species and in particular to its sister A. 4.1. Origin of the subterranean Pyrenean clade leschenaulti (specially the male genitalia, Faille et al., 2006). Within the Eastern clade (Cerbaphaenops sensu lato), what is The most remarkable result of our work was the finding that all currently known as A. crypticola is polyphyletic, with some lineages the highly modified species of subterranean Trechini from the associated to other Aphaenops species according to their geo- Pyrenees share a common origin, to the exclusion of all sampled graphic distribution. These affinities are also supported by mor- epigean species and all highly modified subterranean species con- phological characters (see Section 4). Similarly, the only species sidered by some authors to belong to the phyletic lineage of of Pubaphaenops (Genest, 1983), A. laurenti, with a peculiar mor- Aphaenops from other geographical areas (Apoduvalius, Speotrechus, phology, is grouped in a clade (albeit with low support) with the Paraphaenops, Suppl. Table 1). Jeannel (1928) hypothesised a com- species in the same geographical area, between the Lez and the mon origin for Aphaenops (plus Hydraphaenops) and Geotrechus, Vicdessos valleys, at the eastern limit of the distribution of Aphaen- well separated from the epigean Trechus, but included in this sub- ops (Figs. 2 and 3). terranean ‘‘phyletic series” other genera from outside the Pyrenees. In all trees the genera Hydraphaenops and Geotrechus (and the According to our results, these highly modified subterranean spe- subgenera Geotrechus and Geotrechidius of the later) were polyphy- cies from nearby areas, or less modified troglobitic species from letic, with strong support (Fig. 2, Suppl. Fig. 1). The two Aphaenops the Pyrenees, were nested within Trechus sensu lato, and not di- lineages were sister to some species of Hydraphaenops, while Geo- rectly related with the subterranean clade. This was the case of trechus was split between a paraphyletic basal series and some Speotrechus from the Cevennes (Jeannel, 1922), Apoduvalius from species in a lineage with Hydraphaenops (Fig. 2). the Cantabrian chain (Vives, 1980; although see Faille, 2006 for a In the analyses of the nuclear sequence we excluded six speci- different view), Paraphaenops from the Iberian system, or the micr- mens because of missing data (see Suppl. Table 1). The tree ob- ophthalmous (but not blind) Trechus navaricus from the Pyrenees. tained with Garli with the combined SSU + LSU had the same Other subterranean genera, such as Duvalius and Agostinia, have basic topology as the combined tree (Suppl. Fig. 2), with a well- traditionally being considered as part of a distinct lineage (the supported monophyletic lineage for all subterranean species from ‘‘Duvalius phyletic lineage”), not directly related to Aphaenops,in the Pyrenees, and the polyphyly of all three genera. The main sub- agreement with our results (e.g. Jeannel, 1928; Casale et al., 1998). terranean clades found in the combined tree (including the basal There are several obvious possible caveats to this conclusion: paraphyly of species of Geotrechus) were also present with boot- (1) there could be some un-sampled epigean species which could strap values above 70%, although, due to the lower variability of belong to this clade, (2) there could be some un-sampled Pyrenean the nuclear genes, relationships within the two main clades (W subterranean species outside this clade (i.e. sharing a most recent and E in Figs. 2 and 3) were not recovered. common ancestor with other epigean species, not with the subter- ranean clade), or (3) there could be some un-sampled non-Pyre- 3.2. Divergence time estimates nean subterranean species inside this clade. Based on previous morphological analyses there are no obvious candidate species We combined the results of the two independent runs of Beast, for the first two cases, but for the third only the study of potential with a final estimation of the rate at 0.0115 ± 0.0002 substitutions/ candidates (e.g. Sardaphaenops, Italaphaenops, Allegrettia; Casale A. Faille et al. / Molecular Phylogenetics and Evolution 54 (2010) 97–106 103

0.2 [0.05,0.38] A_parallelus_MNHN_AF53 [0.12,0.57] A_crypticola_MNHN_AF49 0.33 [0.39,1.11] A_crypticola_MNHN_AF47 0.73 [0.72,1.75] 0.27 A_vandeli_MNHN_AF44 [0.07,0.51] [1.17,2.51] 1.2 A_vandeli_MNHN_AF45 A_mariarosae_MNHN_AF57 2.4 1.81 A_pluto_MNHN_AF58 [1.66,3.2] 3.26 A_sp_MNHN_AF133 [2.4,4.16] 4.01 2.1 A_delbreili_MNHN_AF37 [3.06,4.95] [0.84,3.32] A_laurenti_MNHN_AF63 1.55 A_cerberus_MNHN_AF20_AF30 3.35 [0.72,2.51] 4.24 A_jauzioni_MNHN_AF33 [2.3,4.38] [3.26,5.22] A_bucephalus_MNHN_AF62 0.67 5.19 A_hustachei_MNHN_AF39 2.6 [0.14,1.37] [4.02,6.4] A_crypticola_MNHN_AF135 E [1.25,4.07] A_sp_MNHN_AF42 6.73 A_tiresias_MNHN_AF59_AF60 [5.47,8.13] 1.05 H_bourgoini_MNHN_AF68

[6.28,9.16] 4.76 [0.29,1.98] H_bourgoini_MNHN_AF69 [2.99,6.47] 2.02 H_ehlersi_MNHN_AF64 7.69 [0.81,3.39] H_pecoudi_MNHN_AF72 3.54 H_penacollaradensis_MNHN_AF121 [0.84,6.96] H_elegans_MNHN_AF120 0.86 A_abodiensis_MNHN_AF4 [0.25,1.64] A_loubensi_MNHN_AF3 3.71 A_alberti_MNHN_AF12 8.41 [4.59,7.07] [2.27,5.23] A_jeanneli_MNHN_AF11 [7.11,9.83] W 5.82 1.16 A_ludovici_MNHN_AF15 [0.35,2.11] A_rhadamanthus_MNHN_AF13_AF14 6.29 1.45 A_cabidochei_MNHN_AF5_AF6 [5.07,7.52] [0.52,2.53] A_ochsi_MNHN_AF7_AF8 [5.97,8.6] 4.33 A_leschenaulti_MNHN_AF1 7.26 [2.27,6.12] A_catalonicus_MNHN_AF2 9.69 6.28 H_vasconicus_MNHN_AF65 [7.62,12.25] 7.93 [4.3,8.12] H_galani_MNHN_AF67 [6.65,9.25] H_delicatulus_MNHN_AF66 6.01 G_gallicus_MNHN_AF76 [4.1,7.78] G_trophonius_MNHN_AF83 H_pandellei_MNHN_AF70 3.05 G_vulcanus_MNHN_AF91 7.7 [0,7.35] G_saulcyi_MNHN_AF86 [5.12,10.12] G_orpheus_MNHN_AF79_AF81

1.0

Fig. 4. Ultrametric tree of the Phylogeny of subterranean Trechini of the Pyrenees obtained with Beast, using a standard mitochondrial rate (0.0115 substitutions/site/MY). Number above nodes, estimated age (in MY); numbers below nodes, 95% confidence intervals. and Laneyrie, 1982) can establish their phylogenetic relationships character states: loss of eyes, apterism, depigmented body, a sub- with certain confidence. terranean life, and a requirement for high levels of humidity (e.g. The sister lineage of the subterranean clade was not well-de- Jeannel, 1926a; Vannier and Thibaud, 1971). These are also traits fined in our analyses, as support for the basal nodes of the lineage that have been linked with a reduced dispersal ability (Kane including Trechus and related (mostly subterranean) genera was et al., 1992; Barr and Holsinger, 1985; Caccone, 1985), and thus low. What seems clear from our analyses is that, under its current run against the interpretation of a single origin of subterranean concept, Trechus, with more than 440 species distributed in the adaptations with subsequent diversification over a relatively large northern Hemisphere and the mountains of sub-Saharan Africa geographical area (ca. 360 km, from the Puigmal massif, Geotrechus (Casale and Laneyrie, 1982), includes epigean or weakly modified puigmalensis Lagar, 1981, to Guipuzcoa, Hydraphaenops galani subterranean species with a plesiomorphic morphology, forming Español, 1968). The traditional solution to this dilemma was the a largely paraphyletic series with numerous genera of highly mod- assumption that there have been multiple independent active col- ified species nested within. A thorough taxonomic revision of Tre- onisations of the subterranean environment restricted to a very chus sensu lato (including the Pyrenean subterranean taxa) would limited geographical area, each derived from different epigean be highly desirable, but impossible until a more comprehensive ancestors and with a secondary reduction of gene flow (the ‘‘adap- phylogeny is available. tive shift hypothesis”, Howarth, 1982; Peck and Finston, 1993; The monophyly of all the highly modified subterranean species Chapman, 1993; Desutter-Grandcolas and Grandcolas, 1996; Rive- of the Pyrenees strongly suggests a single origin of their shared ra et al., 2002). 104 A. Faille et al. / Molecular Phylogenetics and Evolution 54 (2010) 97–106

Under this scenario, one would expect to find multiple in- The main lineages within the subterranean clade seem to be stances of species with ‘‘intermediate” morphologies (e.g. partly geographically well differentiated, with successive splits between depigmented bodies, reduced eyes) interspersed among the highly the eastern and western Pyrenees resulting in several geographi- modified or epigean ones. Although still based on a limited number cally well-defined clades (Fig. 3). Relationships among closely re- of species, this seems to be the case for the Cantabrian chain, with lated species reflect geographical proximity more than general species of Apoduvalius intermixed with epigean Trechus (Fig. 2), morphological similarities, with morphologically highly divergent and also of the Trechus radiation in the Canary islands, where the species found in close proximity, as found for other subterranean subterranean species of the archipelago are closely related to the organisms (Fišer et al., 2008). This is for example the case of epigean species from the same geographical area (Contreras-Díaz Hydraphaenops pandellei and Geotrechus gallicus, of very different et al., 2007; Borges et al., 2007). Recent molecular work on other morphology and ecology, or Aphaenops jeanneli and A. alberti. The cave-dwelling species also suggests frequent multiple colonisa- latter (Fig. 1) is a very distinct and scarce species endemic to the tions of the subterranean environment in the same geographic area Arbailles massif, in the western Pyrenees, previously assumed to (Crustacean isopods, Rivera et al., 2002; Amphipods, Fišer et al., be related to some species from the Eastern clade (Cerbaphaenops 2008). sensu lato), such as A. bucephalus (Jeannel, 1939; Coiffait, 1962) In our case, this interpretation would require the complete or A. pluto (Jeanne, 1967). It occurs in the same caves with A. jean- extinction of all species with intermediate morphologies which neli, to which it is closely related according to our molecular data could be included in the Pyrenean subterranean clade. This has despite its very different body shape, suggesting an ecological been hypothesised for areas subjected to strong fluctuating cli- differentiation. mate, in particular in areas subjected to dry periods which could On the other hand, species that were previously considered to result in the extinction of epigean hygrophilous species: the ‘‘cli- be closely related based on their general appearance, but occurring matic relict hypothesis” of Jeannel (1943) and Peck and Finston in different geographical areas, were found to be included in their (1993). According to our estimations based on a standard mito- local clades. Thus, the species of the subgenera Arachnaphaenops chondrial rate, the origin of the Pyrenean subterranean clade (A. pluto, A. tiresias and A. alberti), with a very similar appearance, would be mid-late Miocene, a time in when the general climate were included in three different clades with other Aphaenops spe- in the area seems to have been warmer and wetter than today, cies according to their distributions. Similarly, according to our re- with extensive forested areas (Bruch et al., 2007; Jiménez-Moreno sults what is currently considered as Aphaenops crypticola, and Suc, 2007). A possible dry period producing this generalised distributed from caves between Haute-Garonne and Hautes-Pyré- extinction, and the separation between the main Eastern and Wes- nées, would be polyphyletic. The populations of the western part tern clades, could have been the Messinian salinity crisis at the of the range (Aure valley, Mont Né) are very close to A. crypticola Miocene–Pliocene boundary, although recent data suggest that aeacus and A. hustachei from the same area, while populations from the vegetation of the north Mediterranean area may not have been the eastern part are subdivided in two groups delimited by the deeply affected (e.g. Bertini, 2006). In any case, it must be stressed Garonne valley: a western (A. crypticola MNHN-51 and 48) and that these dates are based on a fixed rate estimated from a combi- an eastern group (A. crypticola MNHN-47, 49, 52, 136). Due to nation of genes in several groups (0.0115 substitutions/ the lack of resolution of the nuclear data (Suppl. Fig. 2) it is not pos- site/MY, Brower, 1994), and thus have to be considered as merely sible to discard the possibility of local introgression among some of orientative. The only estimate of mitochondrial evolutionary rate these closely related species, but there is no evidence of incongru- of a closely related group (0.015 substitutions/site/MY for the ence between the nuclear and mitochondrial genomes in any of the genus Trechus) is based on the colonisation of the Canary islands, lineages for which there is enough resolution, contrary to what (Contreras-Díaz et al., 2007). As already noted, this was based on happens in other groups of Carabidae, in which introgression a combination of cox1 and cox2, known to have faster rates than through hybridisation is common (e.g. Sota and Vogler, 2001; Deu- ribosomal genes, and thus not directly applicable to our dataset. ve, 2004; Streiff et al., 2005; Zhang and Sota, 2007). A re-examina- tion of the morphology of the different populations of A. crypticola in the light of our results revealed differences in the shape of the 4.2. Diversification of the subterranean Pyrenean clade aedeagus and some male secondary sexual characters consistent with this geographical split (Faille, 2006). Within the subterranean Pyrenean clade, the three currently We found clear differences in the pattern of diversification be- recognised genera (Aphaenops, Hydraphaenops and Geotrechus) tween the Western and Eastern clades. The Western clade, be- were found to be polyphyletic. These genera were originally de- tween Bagnères-de-Bigorre and the Arbailles massif (clade W, fined according to their general body shape, especially the head Aphaenops s.str. plus A. (Arachnaphaenops) alberti and Geaphaen- and elytra (Jeannel, 1926b; Coiffait, 1962; see Section 2 above). ops), seems to be the oldest lineage of troglobitic species, with an These are likely to be characters reflecting different adaptations estimated Late Miocene basal diversification (Fig. 4). Some of the to the subterranean environment: even if most species are only species within this group have secondarily developed endogean known from caves, species of Geotrechus are mostly endogean, liv- habits, with a reverse to a more stout (i.e. less ‘‘aphaenopsian”) ing in deep humid soil, while species of Aphaenops live in more body shape (A. ludovici, A. rhadamanthus). They were included in open subterranean spaces, such as caves or the interstices of the the subgenus Geaphaenops by Cabidoche (1965), together with MSS (Jeannel, 1926b; Juberthie and Bouillon, 1983). A particularly other endogean species of more uncertain relationships not in- interesting case is the apparently highly specialised habit of most cluded in our study (A. linderi Jeannel, 1938, A. rebereti Gaudin, of the species of the genus Hydraphaenops, which seem to live in 1947, and also A. cissauguensis Faille and Bourdeau, 2008). the cracks of karstic massifs and are only occasionally found in The main clade of the Eastern Pyrenees, between Bagneres-de- caves. They have a cylindrical head and long and sickled mandibles, Bigorre and the Ariege River (Cerbaphaenops plus the morphologi- likely to be adapted to an unknown prey (Jeannel, 1926b; Deleu- cally distinct species A. laurenti, A. bucephalus, A. chappuisi, A. pluto rance-Glaçon, 1963). According to our results, it seems that the and A. tiresias), seems to be of more recent origin, with a Pliocene– general body shape is associated with the particular ecological Pleistocene diversification (Fig. 4) and species with a more homo- and physical conditions of the subterranean environment colon- geneous morphology (Coiffait, 1962). The sampling of this clade ised by these species, with a high degree of homoplasy and conver- was complete, with two exceptions: (1) A. bourdeaui Coiffait, gence (Marquès and Gnaspini, 2001; Fišer et al., 2008). 1976, considered as part of Cerbaphaenops despite being found in A. Faille et al. / Molecular Phylogenetics and Evolution 54 (2010) 97–106 105 the area of distribution of the W clade. It is only known from two Appendix B. Supplementary data females collected the same day (Coiffait, 1976), but never found again despite numerous visits to the cave. The lack of males and Supplementary data associated with this article can be found, in its geographical distribution cast doubts about its affinities, which the online version, at doi:10.1016/j.ympev.2009.10.008. could only be solved with molecular data. (2) A. hidalgoi Español and Comas, 1985, also from the W Pyrenees. It was described as References Cerbaphaenops (Español and Comas, 1985), but it is a Hydraphaen- ops-like species, apparently close to H. penacolladarensis—which is Abeille de Perrin, E., 1872. Etudes sur les Coléoptères cavernicoles suivies de la found in the same geographical area (Faille, unpublished observa- description de 27 Coléoptères nouveaux français. Marius Olive, Marseille, 41 p. tions). The Western and Eastern clades overlap in the Bigorre area, Barr, T.C., 1979. The taxonomy, distribution, and affinities of , with notes on associated species of (Coleoptera, Carabidae). Am. where one species of each group occur sympatrically in a few Mus. Novit. 2682, 1–20. caves: Aphaenops leschenaulti (Eastern clade) and Aphaenops crypti- Barr, T.C., Holsinger, J.R., 1985. Speciation in cave faunas. Ann. Rev. Ecol. Syst. 16, cola aeacus (Western clade) (Fresneda et al., 2009). 313–337. Bedel, L., Simon, E., 1875. Liste générale des Articulés Cavernicoles de l’Europe. P. A potential explanation for the differences between the eastern Gervais, Paris, IV, pp. 1–69. and western lineages of Pyrenean subterranean Trechini could be Bertini, A., 2006. The Northern Apennines palynological record as a contribute for the different pattern of the limestone areas in which they are the reconstruction of the Messinian palaeoenvironments. Sedim. Geol. 188– 189, 235–258. found. In the west, areas of suitable karstified habitat tend to be Blin, N., Stafford, D.W., 1976. A general method for isolation of high molecular larger and more homogeneous, frequently with continuous patches weight DNA from Eucaryots. Nucleic Acids Res. 3, 2303. of ca. 150 km2 (e.g. the Arbailles massif, Vanara, 2000). On the con- Bonvouloir, H. de, 1862. Description d’un genre nouveau et de deux espèces nouvelles de coléoptères de France. Ann. Soc. Entomol. Fr. 4 (1), 567–571 (1861). trary, in the Eastern Pyrenees karstified limestone is highly frag- Borges, P.A.V., Oromí, P., Serrano, A.R.M., Amorim, I.R., Pereira, F., 2007. Biodiversity mented, opening opportunities for the development of multiple patterns of cavernicolous ground-beetles and their conservation status in the isolated local populations leading to allopatric speciation (Culver, Azores, with the description of a new species: Trechus isabelae n. sp. (Coleoptera: Carabidae: Trechinae). Zootaxa 1478, 21–31. 1970; Crouau-Roy, 1986; Faille and Déliot, 2007). Brower, A.V.Z., 1994. Rapid morphological radiation and convergence among races of the butterfly Heliconius erato inferred from patterns of mitochondrial DNA evolution. Proc. Natl. Acad. Sci. USA 91, 6491–6495. Acknowledgments Bruch, A.A., Uhl, D., Mosbrugger, V., 2007. Miocene climate in Europe—Patterns and evolution. A first synthesis of NECLIME. Palaeogeog. Palaeoclim. Palaeoecol. 253, 1–7. We wish to thank J.P. Besson, F. Brehier, E. Dupré, F. Fadrique, J. Cabidoche, M., 1965. Sur les Aphaenops du groupe rhadamanthus (Col. Carab.). Ann. Fresneda, G. Jauzion, J.M. Salgado, E. Ollivier, J. Raingeard, Spéléol. 20 (4), 523–528. Cabidoche, M., 1966. Contribution à la connaissance de l’écologie des Trechinae C. Vanderbergh and all the collectors mentioned in the Suppl. Table cavernicoles pyrénéens. Unpublished thesis. Paris, 228 pp. 1 for their help during field work, the members of the Groupe Caccone, A., 1985. Gene flow in cave : a qualitative and quantitative Spéléologique du Couserans and Groupe Spéléologique Minos, approach. Evolution 39, 1223–1234. Casale, A., 2004. Two new hypogean beetles from Sardinia, Sardaphaenops adelphus without whom visiting some particularly difficult cavities would n. sp. (Coleoptera Carabidae) and Patriziella muceddai n. sp. (Coleoptera have been impossible, G. Kergoat for the primer cp1, U. Schmidt Cholevidae) and their biogeographical significance. Boll. Soc. Entomol. Ital. (http://www.kaefer-der-welt.de/) for the photo of T. quadristriatus, 136 (1), 3–31. A. Hassanin for help during the PhD of A.F., A. Cieslak, J. Fresneda, Casale, A., Laneyrie, R., 1982. Trechodinae et Trechinae du monde. Tableau des sous- familles, tribus, séries phylétiques, genres, et catalogue général des espèces. A. Casale for multiple discussions on the evolution of the subterra- Mém. Biospéol. 9, 1–226. nean beetles, D.T. Bilton for reading the manuscript, and three Casale, A., Vigna Taglianti, A., Juberthie, C., 1998. Coleoptera Carabidae. In: anonymous referees for useful comments on previous versions of Juberthie, C., Decu, V. (Eds.), Encyclopaedia Biospeologica, tome II. Société Internationale de Biospéologie. Moulis, France, pp. 1047–1081. our work. A.F. and C.B. thank the Subterranean laboratory of Moulis Chapman, P., 1993. Caves and Cave Life. Harper Collins, London. pp. 1–224. (currently Station d’écologie expérimentale du CNRS) for support Coiffait, H., 1962. Monographie des Trechinae cavernicoles des Pyrénées. Ann. during field work. This study was funded in part by the project Spéléol. 17 (1), 119–170. Coiffait, H., 1976. Aphaenops (Cerbaphaenops) bourdeaui, nouvelle espèce des Synthesys (ES-TAF-1540) to A.F. and I.R., the Société Entomologi- Pyrénées Atlantiques. Nouv. Rev. Entomol. 6 (3), 247–248. que de France (grants ‘‘Germaine Cousin”) to A.F., and the Spanish Contreras-Díaz, H.G., Moya, O., Oromí, P., Juan, C., 2007. Evolution and MICINN project CGL2007-61665 to I.R. We dedicate this work to diversification of the forest and hypogean ground- genus Trechus in the Canary Islands. Mol. Phylogenet. Evol. 42, 687–699. Philippe Déliot, recently disappeared, without whom collecting Crouau-Roy, B., 1986. Population studies on pyrenean troglobitic beetles: local all the species necessary to this study could not have been genetic differentiation and microgeographic variations in natural populations. possible. Biochem. Syst. Ecol. 14 (5), 521–526. Culver, D.C., 1970. Analysis of simple cave communities. I. Caves as islands. Evolution 24, 463–474. Culver, D.C., Kane, T.C., Fong, D.W., Jones, R., Taylor, M.A., Sauereisen, S.C., 1990. Appendix A Morphology of cave organisms—is it adaptative? Mém. Biospéol. 17, 13–26. Culver, D.C., Deharveng, L., Bedos, A., Lewis, J.J., Madden, M., Reddell, J.R., Sket, B., Trontelj, P., White, D., 2006. The mid-latitude biodiversity ridge in terrestrial Bonvouloir (1862) described the genus Aphoenops for a species cave fauna. Ecography 29, 120–128. collected in a cave of the central French Pyrenees, A. leschenaulti Deleurance, S., 1958. La contraction du cycle évolutif des Coléoptères Bathysciinae et Trechinae en milieu souterrain. C. R. Séances Acad. Sci. 247, 752–753. Bonvouloir, 1862. Subsequently, the genus name was written as Deleurance-Glaçon, S., 1963. Contribution à l’étude des coléoptères cavernicoles de Aphaenops by Grenier (1864), and this has been the spelling used la sous-famille des Trechinae. Ann. Spéléol. 18 (2), 227–265. afterwards by all authors with the only exception of Abeille de Per- Desutter-Grandcolas, L., Grandcolas, P., 1996. The evolution toward troglobitic life: rin (1872), Bedel and Simon (1875) and Peyerimhoff (1915). Re- a phylogenetic reappraisal of climatic relict and local habitat shift hypotheses. Mém. Biospéol. 23, 57–63. cently Moravec et al. (2003) resurrected the original graphy Desutter-Grandcolas, L., D’Haese, C., Robillard, T., 2003. The problem of characters Aphoenops, which has been subsequently used by Lorenz (2005) susceptible to parallel evolution in phylogenetic reconstructions: a reply to and some other authors. An opinion to the ICZN is in preparation Marquès & Gnaspini (2001) with emphasis on cave life phenotypic evolution. Cladistics 19, 131–137. to conserve Aphaenops, based on the acceptance by the ICZN of Deuve, T., 2000. Nouveaux Trechidae cavernicoles chinois, découverts dans les the equivalence of oe and ae for species-level names (not specified confins karstiques du Sichuan, du Hubei et du Yunnan (Coleoptera, Adephaga). for genus-level names), and the prevalence of use for near 150 Rev. Fr. Entomol. (n.s.) 21, 151–161. Deuve, T., 2001. Le genre Eustra Schmidt-Goebel, 1846, insectes (Coleoptera, years, with hundreds of examples of the use of Aphaenops and vir- Paussidae, Ozaeninae) à genitalia femelles orthotopiques. Zoosystema 23 (3), tually no use of Aphoenops. 547–578. 106 A. Faille et al. / Molecular Phylogenetics and Evolution 54 (2010) 97–106

Deuve, T., 2004. Illustrated catalogue of the genus Carabus of the world (Coleoptera: Lorenz, W., 2005. A Systematic List of Extant Ground Beetles of the World. Carabidae). Pensoft Publ., Sofia, Moscow, x + 461 pp. (Coleoptera ‘‘Geadephaga”: Trachypachidae and Carabidae incl. Paussinae, Drummond, A.J., Rambaut, A., 2007. BEAST: Bayesian evolutionary analysis by Cicindelinae, Rhysodinae), second ed. Tutzing, Germany. 530p.. sampling trees. BMC Evol. Biol. 7, 214. Marquès, A.C., Gnaspini, P., 2001. The problem of characters susceptible to parallel Drummond, A.J., Ho, S.Y.W., Phillips, M.J., Rambaut, A., 2006. Relaxed phylogenetics evolution in phylogenetic reconstructions: suggestion of a practical method and and dating with confidence. PLoS Biol. 4 (5), e88. doi:10.1371/ its application to cave . Cladistics 17, 371–381. journal.pbio.0040088. Monteiro, A., Pierce, N.E., 2001. Phylogeny of Bicyclus (Lepidoptera: Nymphalidae) Dupré, E., 1991. Trechini nouveaux ou peu connus de France et d’Espagne inferred from COI, COII, and Ef-1a gene sequences. Mol. Phylogenet. Evol. 18, (Coléoptères Carabidae). Mém. Biospéol. 18, 287–299. 264–281. Español, F., 1968. Un nuevo Hydraphaenops de la provincia de Guipúzcoa (Col. Moravec, P., Uéno, S.-I., Belousov, I.A., 2003. Tribe Trechini. In: Löbl, Smetana (Eds.), Trechidae). Misc. Zool. 2 (3), 55–58. Catalogue of Palaearctic Coleoptera, Archostemata, Myxophaga, Adephaga, vol. Español, F., Comas, J., 1985. Un nuevo Aphaenops Bonv. de la vertiente española de 1. Apollo Books, Stenstrup, pp. 288–346 los Pirineos (Col., Carabidae, Trechinae). Misc. Zool. 9, 219–221. Ober, K.A., 2002. Phylogenetic relationships of the carabid subfamily Faille, A., 2006. Endémisme et adaptation à la vie cavernicole chez les Trechinae (Coleoptera) based on molecular sequence data. Mol. Phylogenet. Evol. 24, 228– Pyrénéens (Coleoptera: Carabidae). Approches moléculaire et morphométrique. 248. Unpublished thesis. Muséum National d’Histoire Naturelle, Paris, 319 pp. Ortuño, V.M., Sendra, A., Montagud, S., Teruel, S., 2004. Systématique et biologie Faille, A., Déliot, P., Quéinnec, E., 2007. A new cryptic species of Aphaenops d’une espèce paléoendémique hypogée de la péninsule Ibérique: Ildobates (Coleoptera: Carabidae: Trechinae) from a French Pyrenean cave: congruence neboti Español 1966 (Coleoptera: Carabidae: Dryptinae). Ann. Soc. Entomol. Fr. between morphometrical and geographical data confirm species isolation. Ann. (n.s.) 40 (3–4), 459–475. Soc. Entomol. Fr. (n.s.) 43 (3), 363–370. Peck, S.B., Finston, T.L., 1993. Galapagos islands troglobites: the question of tropical Faille, A., Fresneda, J., Déliot, P., Bourdeau, C., 2006. Description du mâle troglobites, parapatric distributions with eyed sister-species, and their origin by d’Aphaenops catalonicus Escolà & Canció (Coleoptera, Trechinae). Bull. Soc. parapatric speciation. Mém. Biospéol. 20, 19–37. Entomol. Fr. 111 (2), 247–250. Peyerimhoff, P. de, 1915. Variations de contours et de la chétotaxie chez Trechus Felsenstein, J., 1985. Confidence limits on phylogenies: an approach using the (Trechopsis) lapiei Peyer.; démonstration de sa parente phylogénique avec bootstrap. Evolution 39, 783–791. Aphoenops iblis Peyer. Bull. Soc. Entomol. Fr., 128–133. Fišer, C., Sket, B., Trontelj, P., 2008. A phylogenetic perspective on 160 years of Pons, J., 2006. DNA-based identification of preys from non-destructive, total DNA troubled taxonomy of Niphargus (Crustacea: Amphipoda). Zool. Script. 37 (6), extractions of predators using arthropod universal primers. Mol. Ecol. Notes 6, 665–680. 623–626. Fresneda, J., Salgado, J.M., Ribera, I., 2007. Phylogeny of western Mediterranean Posada, D., Crandall, K.A., 1998. Modeltest: testing the model of DNA substitution. Leptodirini, with an emphasis on genital characters (Coleoptera: Leiodidae: Bioinformatics 14 (9), 817–818. Cholevinae). Syst. Entomol. 32, 332–358. Putzeys, J., 1870. Trechorum oculatorum Monographia. Stett. Entomol. Zeit. 31 (1– Fresneda, J., Bourdeau, C., Faille, A., 2009. Baronniesia delioti gen. n. sp. n., a new 3), 7–48. 145–201. subterranean Leptodirini from the French Pyrenees (Coleoptera: Leiodidae: Racovitza, E.G., 1907. Essai sur les problèmes biospéologiques. Arch. Zool. Exp. Gén. Cholevinae). Zootaxa 1993, 1–16. 6 (4ème sér.), 371–488. Genest, L.C., 1983. Nouvelles espèces d’Aphaenops et d’Hydraphaenops des Pyrénées Ribera, I., Hernando, C., Aguilera, P., 2001. Agabus alexandrae n.sp. from Morocco, Centrales (Coléoptères Trechinae). Mém. Biospéol. 10, 305–310. with a molecular phylogeny of the western Mediterranean species of the A. Gilbert, M.T.P., Moore, W., Melchior, L., Worobey, M., 2007. DNA extraction from dry guttatus group (Coleoptera: Dytiscidae). Ins. Syst. Evol. 32, 253–262. museum beetles without conferring external morphological damage. PLoS ONE Rivera, M.A.J., Howarth, F.G., Taiti, S., Roderick, G.K., 2002. Evolution in Hawaiian 2 (3), e272. doi:10.1371/journal.pone.0000272. cave-adapted isopods (Oniscidea: Philosciidae): vicariant speciation or adaptive Grebennikov, V.V., 2008. Tasmanitachoides belongs to Trechini (Coleoptera: shifts? Mol. Phylogenet. Evol. 25, 1–9. Carabidae): discovery of the larva, its phylogenetic implications and revised Roncin, E., Deharveng, L., 2003. Leptogenys khammouanensis sp. nov. (Hymenoptera: key to Trechitae genera. Invertebr. Syst. 22, 479–488. Formicidae). A possible troglobitic species of Laos, with a discussion on cave Grebennikov, V.V., Maddison, D.R., 2005. Phylogenetic analysis of Trechitae ants. Zool. Sci. 20, 919–924. (Coleoptera: Carabidae) based on larval morphology, with description of first- Rowley, D.L., Coddington, J.A., Gates, M.W., Norrbom, A.L., Ochoa, R.A., Vandenberg, instar Phrypeus and a key to genera. Syst. Entomol. 30, 38–59. N.J., Greenstone, M.H., 2007. Vouchering DNA-barcoded specimens: test of a Grenier, A., 1864. Description de nouvelles espèces de Coléoptères Français des nondestructive extraction protocol for terrestrial arthropods. Mol. Ecol. Notes 7, genres Cionus, Raymondia et Anophthalmus et quelques réflexions sur les yeux 915–925. de certaines espèces réputées aveugles. Ann. Soc. Entomol. Fr. 33, 133–140. Shimodaira, H., Hasegawa, M., 1999. Multiple comparisons of log-likelihoods with Hall, T.A., 1999. BioEdit: a user-friendly biological sequence alignment editor and applications to phylogenetic inference. Mol. Biol. Evol. 16, 1114–1116. analysis program for Windows 95/98/NT. Nucleic Acids Symp. Ser. 41, 95–98. Shull, V.L., Vogler, A.P., Baker, M.D., Maddison, D.R., Hammond, P.M., 2001. Sequence Howarth, F.G., 1982. Bioclimatic and geologic factors governing the evolution and alignment of 18S Ribosomal RNA and the basal relationships of Adephagan distribution of Hawaiian cave insects. Entomol. Gen. 8 (1), 17–26. beetles: evidence for monophyly of aquatic families and the placement of Huelsenbeck, J.P., Ronquist, F., 2001. MRBAYES: bayesian inference of phylogenetic Trachypachidae. Syst. Biol. 50, 945–969. trees. Bioinformatics 17, 754–755. Simon, C., Frati, F., Beckenbach, A., Crespi, B., Liu, H., Flook, P., 1994. Evolution, Jeanne, C., 1967. Carabiques de la péninsule ibérique (5ème note). Actes Soc. Linn. weighting and phylogenetic utility of mitochondrial gene sequences and a Bordeaux, Sér. A 10 (104), 1–22. compilation of conserved polymerase chain reaction primers. Ann. Entomol. Jeannel, R., 1919. Diagnoses préliminaires de Trechinae [Col. Carabidae] Soc. Am. 87, 651–701. cavernicoles nouveaux de France. Bull. Soc. Entomol. Fr., 253–255. Sota, T., Vogler, A.P., 2001. Incongruence of mitochondrial and nuclear gene trees in Jeannel, R., 1922. Les Trechinae de France (Deuxième partie). Ann. Soc. Entomol. Fr. the Carabid beetles Ohomopterus. Syst. Biol. 50, 39–59. 90, 295–345 (1921). Streiff, R., Veyrier, R., Audiot, P., Meusnier, S., Brouat, C., 2005. Introgression in Jeannel, R., 1926a. Faune cavernicole de la France avec une étude des conditions natural populations of bioindicators: a case study of Carabus splendens and d’existence dans le domaine souterrain. Lechevalier, Paris. 334 pp. Carabus punctatoauratus. Mol. Ecol. 14, 3775–3786. Jeannel, R., 1926b. Monographie des Trechinae. Morphologie comparée et Swofford, D.L., 2002. PAUP*: Phylogenetic Analysis Using Parsimony ( and Other distribution d’un groupe de Coléoptères. L’Abeille 32, 221–550. Methods) Version 4. Sinauer Associates, Sunderland, MA. Jeannel, R., 1928. Monographie des Trechinae. Morphologie comparée et Templeton, A.R., 1983. Phylogenetic inference from restriction endonuclease distribution d’un groupe de Coléoptères. Troisième Livraison: les Trechini cleavage site maps with particular reference to the evolution of man and the cavernicoles. L’Abeille 35, 1–808. apes. Evolution 37, 221–224. Jeannel, R., 1939. Un nouvel Aphaenops de la grotte d’Oxibar. Rev. Fr. Entomol. 6 (3– Uéno, S.I., Clarke, A.K., 2007. Discovery of a new aphaenopsoid Trechine Beetle 4), 83–85. (Coleoptera, Trechinae) in northeastern Jiangxi, East China. Elytra, Tokyo 35 (1), Jeannel, R., 1941. Faune de France. Coléoptères Carabiques I, vol. 39. Lechevalier, 267–278. Paris. 571 pp. Vanara, N., 2000. Le karst des Arbailles (Pyrénées-Atlantiques, France). Karstologia Jeannel, R., 1943. Les fossiles vivants des cavernes. Gallimard, Paris. 321 pp. 36, 23–42. Jiménez-Moreno, G., Suc, J.P., 2007. Middle Miocene latitudinal climatic gradient in Vandel, A., 1964. Biospéologie. La Biologie des Animaux Cavernicoles. Gauthier- Western Europe: evidence from pollen records. Palaeogeog. Palaeoclim. Villars, Paris. 619 pp. Palaeoecol. 253, 208–225. Vannier, G., Thibaud, J.M., 1971. Relation entre l’activité motrice d’une espèce de Juberthie, C., Bouillon, M., 1983. Présence des Aphaenops (Coléoptères Trechinae) dans Collemboles cavernicoles et les variations de température dans son biotope. le milieu souterrain superficiel des Pyrénées françaises. Mém. Biospéol. 10, 91–98. Rev. Ecol. Biol. Sol. 8 (2), 261–286. Kane, T.C., Barr Jr., T.C., Badaracca, W.J., 1992. Cave beetle genetics: geology and Vives, E., 1980. Revision del género Apoduvalius Jeannel (Col. Trechinae). Speleon 25, gene flow. Heredity 68, 277–286. 15–21. Katoh, K., Toh, H., 2008. Recent developments in the MAFFT multiple sequence Zhang, A.-B., Sota, T., 2007. Nuclear gene sequences resolve species phylogeny and alignment program. Brief. Bioinform. 9 (4), 286–298. mitchondrial introgression in Leptocarabus beetles showing trans-species Katoh, K., Misawa, K., Kuma, K., Miyata, T., 2002. MAFFT: a novel method for rapid polymorphisms. Mol. Phylogenet. Evol. 45, 534–546. multiple sequence alignment based on fast Fourier transform. Nucleic Acids Zwickl, D.J., 2006. Genetic algorithm approaches for the phylogenetic analysis of Res. 30 (14), 3059–3066. large biological sequence datasets under the maximum likelihood criterion. Laurenti, J.N., 1768. Specimen Medicum, Exhibens Synopsin Reptilium Emendatam Ph.D. dissertation. The University of Texas at Austin. Available from: Cum Experimentis Circa Venena et Antidota Reptilium Austriacorum. Wien. 214 pp. .

Supplementary Fig. 1. Phylogeny of subterranean Trechini of the Pyrenees obtained with parsimony in PAUP, using the combined data matrix. Number in nodes, bootstrap values (if above 50%). “Western” and “Eastern” clades marked with “W” and “E”, respectively (see text).

Supplementary Fig. 2: Phylogeny of subterranean Trechini of the Pyrenees obtained with maximum likelihood in Garli, using only the nuclear sequences (LSU+SSU). Number in nodes, bootstrap values (if above 50%). “Western” clade marked with “W”; “E+” Eastern clade plus some additional species (see text).

Suppl. Table 1. Sequenced specimens, with locality, collectors, sequence accession numbers and ecology (T: troglobitic, E: endogean, Ep: Epigean). Code of specimens used to build composite sequences marked with stars.

No sp locality collector biology code SSU LSU cox1 rrnL tRNA-Leu nad1 cyb 1 Trechini 2 Aphaenops Bonvouloir, 1862 3 Aphaenops Bonvouloir, 1862 (sensu stricto) 4 Aphaenops leschenaulti Bonvouloir, 1861 Grotte de Castelmouly - Bagnères-de-Bigorre (France-65) C. Bourdeau, P. Déliot, A. Faille T MNHN-AF1 GQ293593 GQ293629 GQ293739 GQ293757 GQ293822 GQ293886 5 Aphaenops catalonicus Escolà & Canció, 1983 Cova des Toscllosses - Bonansa (Spain-Huesca) C. Bourdeau, P. Déliot, J. Fresneda T MNHN-AF2 GQ293508 GQ293674 GQ293699 GQ293756 GQ293821 6 Aphaenops loubensi Jeannel, 1953 Salle de la Verna - Sainte-Engrâce (France-64) C. Bourdeau, P. Déliot, A. Faille T MNHN-AF3 GQ293660 GQ293863 7 Aphaenops abodiensis Dupré, 1988 Villanueva de Aezkoa - Sierra de Abodi - P70 (Spain-Navarra) C. Bourdeau, A. Faille, E. Quéinnec T MNHN-AF4 GQ293555 GQ293627 GQ293862 8 Aphaenops bessoni Cabidoche, 1961 Gouffre du Col d’Aran 3 - (France-64) C. Bourdeau, E. Ollivier T MNHN-AF122 GQ293554 9 Aphaenops cabidochei Coiffait, 1959 Salle de la Verna - Sainte-Engrâce (France-64) C. Bourdeau, P. Déliot, A. Faille T MNHN-AF5* GQ293556 GQ293667 GQ293890 10 Villanueva de Aezkoa - Sierra de Abodi - P70 (Spain-Navarra) C. Bourdeau, A. Faille, E. Quéinnec T MNHN-AF6* GQ293520 GQ293741 GQ293778 GQ293831 11 Aphaenops ochsi Gaudin, 1925 Grotte d’Ayssaguer - (France-64) C. Bourdeau, P. Déliot, A. Faille T MNHN-AF7* GQ293666 GQ293892 12 Sima de Garralda - P10 (Spain-Navarra) C. Bourdeau, P. Déliot, A. Faille T MNHN-AF8* GQ293521 GQ293601 GQ293740 GQ293777 GQ293830 13 Aphaenops jeanneli (Abeille de Perrin, 1905) Aven d’Istaurdy - (France-64) C. Bourdeau, P. Déliot, A. Faille T MNHN-AF11 GQ293594 GQ293661 GQ293891 14 Aphaenops orionis Fagniez, 1913 Mine de Larrey - (France-64) C. Bourdeau, P. Déliot, A. Faille T MNHN-AF10* GQ293664 GQ293885 15 Gouffre EL71 - Château-Pignon (France-64) C. Bourdeau, P. Déliot, A. Faille T MNHN-AF9* GQ293507 GQ293592 16 Geaphaenops Cabidoche, 1966 17 Aphaenops rhadamanthus (Linder, 1860) Doline de la Sablère - (France-64) C. Bourdeau E MNHN-AF14* GQ293677 GQ293717 GQ293776 GQ293827 GQ293895 18 Aven de Nabails - Arthez d' (France-64) C. Bourdeau, P. Déliot, A. Faille E MNHN-AF13* GQ293506 19 Aphaenops ludovici Colas & Gaudin, 1935 Grotte d’Ambielle - (France-64) C. Bourdeau, P. Déliot, A. Faille E MNHN-AF15* GQ293550 GQ293676 GQ293716 GQ293775 GQ293828 GQ293896 20 Cerbaphaenops Coiffait, 1962 21 Aphaenops cerberus (Dieck, 1869) Grotte du Sendé - Moulis (France-09) P. Déliot, A. Faille T MNHN-AF30* GQ293589 GQ293646 GQ293718 GQ293779 GQ293835 GQ293871 22 Grotte de L'Estelas - (France-09) P. Déliot, A. Faille T MNHN-AF20* GQ293526 23 Aphaenops jauzioni Faille, Déliot & Quéinnec, 2007 Grotte d’Artigouli - (France-31) P. Déliot, A. Faille T MNHN-AF33 GQ293581 GQ293640 GQ293877 24 Aphaenops carrerei Coiffait, 1953 Gouffre du Trapech d’en Haut - Bordes-sur-Lez (France-09) C. Bourdeau, P. Déliot, A. Faille T MNHN-AF34 GQ293512 GQ293572 GQ293641 25 Aphaenops michaeli Fourès, 1954 Grotte de Noël - (France-09) A. Faille T MNHN-AF35 GQ293515 GQ293585 GQ293722 GQ293768 GQ293815 26 Aphaenops delbreili Genest, 1983 Gouffre du Petit Mirabat - Ercé (France-09) C. Bourdeau, P. Déliot, A. Faille T MNHN-AF37 GQ293513 GQ293570 GQ293638 GQ293720 GQ293773 GQ293804 27 Aphaenops bonneti Fourès, 1948 Trou du Rantou - Suc-et-Sentenac (France-09) P. Déliot, A. Faille T MNHN-AF38 GQ293571 GQ293721 GQ293774 GQ293805 28 Aphaenops hustachei Jeannel, 1916 Grotte de l'Eglise - Nistos (France-65) C. Bourdeau, P. Déliot, A. Faille T MNHN-AF39 GQ293514 GQ293573 GQ293636 GQ293711 GQ293751 GQ293844 29 Aphaenops sp. Grotte de Frechet-Aure - Frechet-Aure (France-65) C. Bourdeau, P. Déliot, A. Faille T MNHN-AF134 GQ293643 30 Aphaenops sp. Grotte de la Cascade - Sarrancolin (France-65) C. Bourdeau, P. Déliot, A. Faille T MNHN-AF135 GQ293576 GQ293635 GQ293710 GQ293749 GQ293842 GQ293883 31 Aphaenops crypticola aeacus (Saulcy, 1864) Grotte de Castelmouly - Bagnères-de-Bigorre (France-65) C. Bourdeau, P. Déliot, A. Faille T MNHN-AF40 GQ293574 GQ293712 GQ293750 GQ293845 32 Aphaenops sp. Tuto de la Cigalero - Ferrère (France-65) P. Déliot, A. Faille T MNHN-AF42 GQ293583 GQ293637 GQ293696 GQ293752 GQ293843 GQ293884 33 Aphaenops vandeli Fourès, 1954 Grotte de Payssa - Salsein (France-09) P. Déliot, A. Faille T MNHN-AF44 GQ293584 GQ293657 GQ293706 GQ293762 GQ293808 GQ293869 34 MSS S100 - (France-09) P. Déliot, A. Faille MSS MNHN-AF43 GQ293648 GQ293856 35 Grotte SL1 - Saint-Lary (France-09) P. Déliot, A. Faille T MNHN-AF45 GQ293656 GQ293705 GQ293761 GQ293807 GQ293867 36 Aphaenops vandeli bouiganensis Fourès, 1954 Grotte de L'Ournas - Saint-Lary (France-09) P. Déliot, A. Faille T MNHN-AF46 GQ293510 GQ293590 GQ293654 37 Aphaenops crypticola (Linder, 1859) Gouffre de Peyreigne - Tibiran-Jaunac (France-65) P. Déliot, A. Faille T MNHN-AF51 GQ293580 GQ293642 38 Grotte d’Aron - d’Aspet (France-31) P. Déliot, A. Faille T MNHN-AF52 GQ293591 GQ293651 GQ293855 39 Grotte de Gouillou - Aspet (France-31) P. Déliot, A. Faille T MNHN-AF47 GQ293577 GQ293671 GQ293709 GQ293765 GQ293812 GQ293866 40 Grotte de Terreblanque - Aspet (France-31) P. Déliot, A. Faille T MNHN-AF50 GQ293579 GQ293650 GQ293858 41 Grotte de l’Haiouat de Pelou - Nistos (France-65) C. Bourdeau, P. Déliot, A. Faille T MNHN-AF48 GQ293578 GQ293655 GQ293859 42 Grotte de la Maouro - Izaut-de-l'Hôtel (France-31) P. Déliot, A. Faille T MNHN-AF49 GQ293653 GQ293708 GQ293764 GQ293811 GQ293868 43 Aphaenops parallelus Coiffait, 1954 Grotte de la Buhadère - Coulédoux (France-31) P. Déliot, A. Faille MSS/T MNHN-AF53 GQ293582 GQ293652 GQ293707 GQ293763 GQ293810 GQ293865 44 Aphaenops sioberae Fourès, 1954 Grotte de Payssa - Salsein (France-09) P. Déliot, A. Faille T MNHN-AF54 GQ293516 GQ293587 GQ293639 45 Aphaenops bouilloni Coiffait, 1955 Grotte de Pétillac - Bordes-sur-Lez (France-09) P. Déliot, A. Faille T MNHN-AF56 GQ293575 GQ293644 GQ293870 46 Aphaenops sp. Grotte d'Aulignac - Bordes-sur-Lez (France-09) A. Faille T MNHN-AF133 GQ293509 GQ293586 GQ293645 GQ293694 GQ293758 GQ293806 GQ293888 47 Aphaenops mariaerosae Genest, 1983 Gouffre du Trapech d’en Haut - Bordes-sur-Lez (France-09) C. Bourdeau, P. Déliot, A. Faille T MNHN-AF57 GQ293568 GQ293649 GQ293704 GQ293760 GQ293809 GQ293860 48 Aphaenops chappuisi Coiffait, 1955 Grotte de la Maouro - Izaut-de-l'Hôtel (France-31) P. Déliot, A. Faille T MNHN-AF61 GQ293525 GQ293563 GQ293684 49 Arachnaphaenops Jeanne, 1967 50 Aphaenops pluto (Dieck, 1869) Grotte du Sendé - Moulis (France-09) P. Déliot, A. Faille T MNHN-AF58 GQ293567 GQ293647 GQ293864 51 Aphaenops tiresias (Piochard de La Brûlerie, 1872) Gouffre de la Peyrère - Balaguères (France-09) A. Faille T MNHN-AF59* GQ293527 GQ293596 GQ293658 GQ293713 GQ293748 GQ293800 52 Grotte du Goueil-di-Her - (France-31) C. Bourdeau, P. Déliot, A. Faille T MNHN-AF60* GQ293889 53 Aphaenops alberti Jeannel, 1939 Aven prox. Istaurdy - Aussurucq (France-64) C. Bourdeau T MNHN-AF12 GQ293595 GQ293662 GQ293700 GQ293782 GQ293829 GQ293853 54 Cephalaphaenops Coiffait, 1962 55 Aphaenops bucephalus (Dieck, 1869) Gouffre de la Peyrère - Balaguères (France-09) A. Faille T MNHN-AF62 GQ293511 GQ293588 GQ293675 GQ293693 GQ293747 GQ293814 GQ293876 56 Pubaphaenops Genest, 1983 57 Aphaenops laurenti Genest, 1983 Grotte de Bordes de Crues - Seix (France-09) A. Faille T MNHN-AF63 GQ293569 GQ293634 GQ293719 GQ293767 GQ293813 GQ293873 58 Hydraphaenops Jeannel, 1926 59 Hydraphaenops ehlersi (Abeille de Perrin, 1872) Goueil-di-Her - Arbas (France-31) C. Bourdeau, P. Déliot, A. Faille T MNHN-AF64 GQ293565 GQ293683 GQ293875 60 Hydraphaenops vasconicus (Jeannel, 1913) Aven d’Istaurdy - Aussurucq (France-64) C. Bourdeau, P. Déliot, A. Faille T MNHN-AF65 GQ293530 GQ293622 GQ293628 GQ293698 GQ293759 GQ293803 61 Hydraphaenops vasconicus delicatulus Coiffait, 1962 Salle de la Verna - Sainte-Engrâce (France-64) C. Bourdeau, P. Déliot, A. Faille T MNHN-AF66 GQ293600 GQ293663 GQ293695 GQ293753 GQ293818 GQ293872 62 Hydraphaenops galani Español, 1968 Guardetxe Koleccia - Usurbil (Spain-Guipuzcoa) C. Bourdeau T MNHN-AF67 GQ293524 GQ293602 GQ293697 GQ293746 GQ293817 63 Hydraphaenops bourgoini (Jeannel, 1945) Grotte de la Maouro - Izaut-de-l'Hôtel (France-31) P. Déliot, A. Faille T MNHN-AF69 GQ293553 GQ293672 GQ293734 GQ293755 GQ293824 GQ293894 64 Grotte de l'Eglise - Nistos (France-65) C. Bourdeau, P. Déliot, A. Faille T MNHN-AF68 GQ293552 GQ293659 GQ293733 GQ293772 GQ293826 GQ293851 65 Hydraphaenops pandellei (Linder, 1859) Grotte d’Arréglade - Rébénacq (France-64) C. Bourdeau T MNHN-AF70 GQ293545 GQ293681 GQ293880 66 Grotte d' Ambielle - Arette (France-64) C. Bourdeau, A. Faille T MNHN-AF71 GQ293546 67 Hydraphaenops pecoudi (Gaudin, 1938) Gouffre du Barroti - (France-09) A. Faille T MNHN-AF72 GQ293566 GQ293673 GQ293738 GQ293878 68 Hydraphaenops elegans Gaudin, 1945 Subterranean river of Artigaléou-Arodets - Esparros (France-65) C. Bourdeau, E. Ollivier, E. Quéinnec T MNHN-AF120 GQ293562 GQ293703 GQ293754 GQ293816 69 Hydraphaenops penacollaradensis Dupré, 1991 Aven El Sinistro, Villanúa (Spain-Huesca) C. Bourdeau, E. Ollivier T MNHN-AF121 GQ293564 GQ293702 GQ293771 GQ293823 70 Geotrechus Jeannel, 1919 71 Geotrechus Jeannel, 1919 (sensu stricto) 72 Geotrechus discontignyi (Fairmaire, 1863) Grotte du Tuco - Bagnères-de-Bigorre (France-65) C. Bourdeau, P. Déliot, A. Faille E/T MNHN-AF92 GQ293560 GQ293893 73 Geotrechus orcinus (Linder, 1859) Gouffre de Peyreigne - Tibiran (France-65) C. Bourdeau, P. Déliot, A. Faille E/T MNHN-AF85 GQ293519 GQ293559 GQ293744 GQ293789 GQ293802 74 Geotrechus orpheus (Dieck, 1869) Grotte de la Quère - Mérigon (France-09) P. Déliot, A. Faille E/T MNHN-AF81* GQ293528 GQ293597 GQ293665 GQ293874 75 Grotte de - Ganties (France-31) P. Déliot, A. Faille E/T MNHN-AF79* GQ293723 GQ293780 GQ293834 76 Geotrechus trophonius (Abeille de Perrin, 1872) Grotte de Tuto Heredo - Merigon (France-09) A. Faille E/T MNHN-AF83 GQ293561 GQ293631 GQ293715 GQ293766 GQ293825 GQ293852 77 Geotrechidius Jeannel, 1947 78 Geotrechus gallicus (Delarouzee, 1857) Aven de Nabails - Arthez d'Asson (France-64) C. Bourdeau, P. Déliot, A. Faille E MNHN-AF76 GQ293518 GQ293557 GQ293670 GQ293724 GQ293769 GQ293798 79 Geotrechus jeanneli Gaudin, 1938 Grotte de la Bouhadère - Saint-Pé-de-Bigorre (France-65) C. Bourdeau, P. Déliot, A. Faille E MNHN-AF77 GQ293517 GQ293558 GQ293725 GQ293770 GQ293799 80 Geotrechus saulcyi (Argod-Vallon, 1913) Grotte du Ker - Rivérenert (France-09) P. Déliot, A. Faille E/T MNHN-AF87 GQ293522 GQ293548 GQ293669 81 Gouffre du Barroti - Lacourt (France-09) A. Faille E/T MNHN-AF86 GQ293668 GQ293887 82 Geotrechus seijasi Español, 1969 Cova d'en Manent - Isòvol (Spain-Girona) P. Déliot, A. Faille E/T MNHN-AF89 GQ293529 GQ293598 GQ293854 83 Geotrechus vandeli Coiffait, 1959 Aven d'Anglade - Couflens (France-09) P. Déliot, A. Faille E MNHN-AF88 GQ293523 GQ293549 GQ293714 GQ293784 GQ293833 84 Geotrechus vulcanus (Abeille de Perrin, 1904) Perte du Fustié - Saint-Martin-de-Caralp (France-09) C. Bourdeau, A. Faille E/T MNHN-AF91 GQ293599 GQ293701 GQ293786 GQ293832 85 Trechus Clairville, 1806 86 Trechus distigma Kiesenwetter, 1851 Aven de Nabails - Arthez d'Asson (France-64) C. Bourdeau, P. Déliot, A. Faille Ep MNHN-AF94 GQ293611 GQ293678 GQ293879 87 Trechus quadristriatus (Schrank, 1781) Collau de la Plana del Turbón - Egea (Spain-Huesca) P. Déliot, A. Faille, J. Fresneda Ep MNHN-AF96 GQ293534 GQ293619 GQ293743 GQ293745 GQ293841 88 Trechus barnevillei Pandellé, 1867 Cueva del Pis - Penilla, Santiurde de Toranzo (Spain-Cantabria) C. Bourdeau, P. Déliot, A. Faille Ep/T MNHN-AF97 GQ293533 GQ293607 GQ293680 GQ293727 GQ293783 GQ293848 89 Dejean, 1831 Cueva del Pis - Penilla, Santiurde de Toranzo (Spain-Cantabria) C. Bourdeau, P. Déliot, A. Faille Ep/T MNHN-AF98 GQ293613 GQ293729 90 Trechus saxicola Putzeys, 1870 Braña Caballo - Piedrafita (Spain-León) C. Bourdeau, P. Déliot, A. Faille Ep MNHN-AF100 GQ293614 GQ293682 GQ293882 91 Trechus schaufussi Putzeys, 1870 Ciudad Real-Navas de Estena-"El Boqueron" (Spain-Toledo) A. Faille E/MSS MNHN-AF101 GQ293532 GQ293620 GQ293737 GQ293788 GQ293820 92 Trechus grenieri uhagoni Crotch, 1869 Cueva de Orobe - Alsasúa (Spain-Navarra) C. Bourdeau T MNHN-AF102 GQ293540 GQ293616 GQ293730 93 Trechus navaricus (Vuillefroy, 1869) Grotte de Sare - Sare (France-64) C. Bourdeau T MNHN-AF103 GQ293539 GQ293603 GQ293687 94 Trechus escalerai Abeille de Perrin, 1903 Cueva de Porro Covañona - Covadonga (Spain-Asturias) J.M. Salgado T MNHN-AF104 GQ293538 GQ293612 GQ293731 GQ293793 GQ293839 95 Erichson, 1837 Saint-Pé-de-Bigorre (France-65) C. Bourdeau, A. Faille Ep MNHN-AF126 GQ293608 GQ293726 GQ293795 GQ293847 96 Trechus comasi Hernando, 2001 Cueva Basaura - Barindano (Spain-Navarra) J. Fresneda T MNHN-AF127 GQ293617 97 Trechus ceballosi Mateu, 1953 Aven de Licie , Lanne-en Barétous (France-64) C. Bourdeau, A. Faille Ep MNHN-AF128 GQ293610 GQ293728 GQ293791 GQ293850 98 Apoduvalius Jeannel, 1953 99 Apoduvalius alberichae Español, 1971 Cova de Agudir - Cardano de abajo - Palencia (Spain-Asturias) J.M. Salgado T MNHN-AF105 GQ293536 GQ293618 GQ293632 GQ293732 GQ293794 GQ293840 100 Apoduvalius sp. Cueva Requexada - Piloñeta (Spain-Asturias) J.M. Salgado T MNHN-AF106 GQ293537 GQ293609 GQ293736 GQ293796 GQ293846 101 Speotrechus Jeannel, 1922 102 Speotrechus mayeti (Abeille de Perrin, 1875) Perte du Rimouren - Saint-Montant (France-07) J-Y. Bigot T MNHN-AF107 GQ293535 GQ293547 GQ293881 103 Paraphaenops Jeannel, 1916 104 Paraphaenops breuilianus (Jeannel, 1916) Cova Cambra - Tortosa (Spain-Tarragona) C. Bourdeau, P. Déliot, A. Faille T MNHN-AF108 GQ293541 GQ293551 GQ293685 105 Iberotrechus Jeannel, 1920 106 Iberotrechus bolivari (Jeannel, 1913) Cueva del Pis - Penilla, Santiurde de Toranzo (Spain-Cantabria) C. Bourdeau, P. Déliot, A. Faille Ep/T MNHN-AF111 GQ293615 GQ293679 GQ293735 GQ293781 GQ293819 GQ293861 107 Duvalius Delarouzée, 1859 108 Duvalius berthae (Jeannel, 1910) Cova d’en Xoles - Pratdip (Spain-Tarragona) C. Bourdeau, P. Déliot, F. Fadrique, A. Faille T MNHN-AF115* GQ293606 GQ293626 GQ293857 109 Cova Massega - Llaberia (Spain-Tarragona) C. Bourdeau, P. Déliot, F. Fadrique, A. Faille T MNHN-AF114* GQ293531 110 Duvalius roberti (Abeille de Perrin, 1903) Grotte de Peïra Cava - Peïra Cava (France-06) A. Coache, J. Raingeard T MNHN-AF129 GQ293605 GQ293691 GQ293785 GQ293837 111 Agostinia Jeannel, 1928 112 Agostinia gaudini (Jeannel, 1952) Puits des Bauges - Dévoluy (France-05) J-Y. Bigot T MNHN-AF116 GQ293543 GQ293604 GQ293692 GQ293787 GQ293838 113 Laosaphaenops Deuve, 2000 114 Laosaphaenops deharvengi Deuve, 2000 Vang Vieng-Nam Xang Tai (Laos) A. Bedos, L. Deharveng T MNHN-AF117 GQ293542 GQ293621 GQ293630 115 Aepopsis Jeannel, 1922 116 Aepopsis robini (Laboulbène, 1849) Plage du Toëno - Trébeurden (France-22) A. Faille Ep MNHN-AF112 GQ293504 GQ293623 GQ293689 GQ293792 GQ293836 117 Perileptus Schaum, 1860 118 Perileptus areolatus (Creutzer, 1799) Immouzer des Ida Outanane (Maroc) P. Aguilera, C. Hernando, I. Ribera Ep MNHN-AF113 GQ293503 GQ293625 GQ293688 119 120 Philochthus Stephens, 1828 121 Philochthus lunulatus (Fourcroy, 1795) Guadalajara - El Pobo de Dueñas (Spain-Guadalajara) A. Cieslak, I. Ribera Ep MNHN-AF118 GQ293505 GQ293686 GQ293690 GQ293797 GQ293801 122 Typhlocharis Dieck, 1869 123 Typhlocharis sp. Santa Almagrera (Spain-Almería) C. Andujar E MNHN-AF119 GQ293502 GQ293624 GQ293633 124 Porotachys Netolitzky, 1914 125 (Nicolaï, 1822) Grotte des Fées - Saint-Cricq-du-Gave (France-40) C. Bourdeau, A. Faille Ep/T MNHN-AF131 GQ293544 GQ293742 GQ293790 GQ293849 Suppl. Table 2. Estimated parameters in the MrBayes run.

Partition Gene 1 LSU 2 cox1 3 cyb 4 SSU 5 rrnL-tRNA-Leu-nad1

Parameter Mean Variance Lower Upper Median PSRF * TL{all} 4.559693 0.063014 4.07 5.043 4.559 1.018 r(A<->C){1} 0.049424 0.000068 0.034689 0.066694 0.049003 1 r(A<->G){1} 0.315247 0.000824 0.261275 0.37451 0.314323 1 r(A<->T){1} 0.230424 0.00033 0.196747 0.267916 0.230173 1 r(C<->G){1} 0.021475 0.000042 0.010314 0.035718 0.020852 1 r(C<->T){1} 0.31653 0.000831 0.262069 0.374831 0.315589 1 r(G<->T){1} 0.066901 0.000136 0.045651 0.091642 0.066263 1 r(A<->C){2} 0.040399 0.00011 0.022005 0.063336 0.039562 1.002 r(A<->G){2} 0.311296 0.001581 0.234833 0.392138 0.310238 1.005 r(A<->T){2} 0.048516 0.000052 0.03531 0.063409 0.048136 1.01 r(C<->G){2} 0.118613 0.000945 0.065049 0.184674 0.116375 1.003 r(C<->T){2} 0.46844 0.002121 0.377717 0.555796 0.468631 1.005 r(G<->T){2} 0.012735 0.00003 0.004212 0.025192 0.012044 1 r(A<->C){3} 0.047638 0.000106 0.030334 0.070141 0.046778 1.003 r(A<->G){3} 0.411278 0.003426 0.297987 0.521927 0.413283 1.01 r(A<->T){3} 0.037981 0.000054 0.025427 0.053689 0.037463 1.015 r(C<->G){3} 0.060626 0.000519 0.023866 0.112859 0.058092 1 r(C<->T){3} 0.4055 0.00329 0.302696 0.520306 0.402505 1.008 r(G<->T){3} 0.036978 0.000132 0.017882 0.06153 0.035806 1.002 r(A<->C){4} 0.01421 0.00008 0.001436 0.035942 0.012656 1.003 r(A<->G){4} 0.05063 0.000178 0.030033 0.081783 0.04883 1.001 r(A<->T){4} 0.861162 0.000686 0.800711 0.903696 0.864183 1.003 r(C<->G){4} 0.003883 0.000004 0.000898 0.00901 0.003525 1 r(C<->T){4} 0.064736 0.000298 0.037439 0.103357 0.062386 1 r(G<->T){4} 0.005379 0.000018 0.000194 0.015923 0.004471 1 r(A<->C){5} 0.021898 0.000096 0.006474 0.044816 0.020428 1 r(A<->G){5} 0.62562 0.001727 0.538516 0.703213 0.627552 1 r(A<->T){5} 0.115637 0.000267 0.086505 0.150775 0.114925 1 r(C<->G){5} 0.024756 0.000488 0.000957 0.082658 0.018881 1 r(C<->T){5} 0.144934 0.000931 0.092562 0.212862 0.142116 1 r(G<->T){5} 0.067155 0.000222 0.040165 0.098204 0.066398 1 pi(A){1} 0.318408 0.000161 0.293197 0.343815 0.318377 1.001 pi(C){1} 0.229642 0.000138 0.207182 0.25343 0.229434 1 pi(G){1} 0.1729 0.000108 0.152807 0.193682 0.172902 1 pi(T){1} 0.27905 0.000141 0.256189 0.302581 0.279116 1 pi(A){2} 0.358209 0.000261 0.326175 0.390413 0.358168 1.001 pi(C){2} 0.099804 0.000049 0.086864 0.113943 0.09955 1 pi(G){2} 0.086904 0.000125 0.067076 0.110516 0.086183 1.01 pi(T){2} 0.455084 0.000241 0.424815 0.485691 0.454603 1.002 pi(A){3} 0.393645 0.000368 0.35688 0.43187 0.393724 1.003 pi(C){3} 0.123165 0.000088 0.105748 0.142669 0.122806 1.006 pi(G){3} 0.05234 0.000123 0.034735 0.076711 0.050999 1.013 pi(T){3} 0.430851 0.000311 0.396442 0.465501 0.430869 1.002 pi(A){5} 0.388421 0.000215 0.359728 0.417285 0.388326 1 pi(C){5} 0.058252 0.000063 0.044143 0.074907 0.057884 1.001 pi(G){5} 0.08781 0.000081 0.071917 0.107191 0.087314 1 pi(T){5} 0.465517 0.000239 0.435488 0.496042 0.465426 1 alpha{1} 0.384221 0.00168 0.312122 0.473951 0.381721 1.002 alpha{2} 0.666526 0.005281 0.531185 0.81887 0.66541 1 alpha{3} 0.887375 0.027705 0.595748 1.240414 0.876048 1.002 alpha{5} 0.65994 0.011901 0.47105 0.89661 0.652117 1 pinvar{1} 0.221579 0.001691 0.134414 0.297168 0.223342 1 pinvar{2} 0.693532 0.000355 0.654392 0.728343 0.694088 1 pinvar{3} 0.546503 0.001226 0.470943 0.606305 0.549672 1.001 pinvar{4} 0.592707 0.000356 0.554866 0.628952 0.592955 1.001 pinvar{5} 0.496361 0.0011 0.42531 0.555367 0.498523 1.002