Molecular phylogeography of Dolichopoda cave crickets (Orthoptera, Rhaphidophoridae): A scenario suggested by mitochondrial DNA

Molecular Phylogenetics and Evolution 37 (2005) 153–164 www.elsevier.com/locate/ympev

Molecular phylogeography of Dolichopoda cave crickets (Orthoptera, Rhaphidophoridae): A scenario suggested by mitochondrial DNA Giuliana Allegrucci ¤, Valentina Todisco, Valerio Sbordoni Department of Biology, University of Rome Tor Vergata, Via della Ricerca ScientiWca, 00133 Rome, Italy Received 8 November 2004; revised 29 April 2005 Available online 16 June 2005

Abstract This study focuses on the phylogenetic relationships among a number of West-Mediterranean cave crickets species belonging to Dolichopoda; primarily a Mediterranean genus, distributed from eastern Pyrenees to Caucasus. In this paper, 11 Dolichopoda species from the French Pyrenees (D. linderi), the island of Corsica (D. bormansi and D. cyrnensis), and northern, central, and southern Italy (D. ligustica, D. schiavazzii, D. aegilion, D. baccettii, D. laetitiae, D. geniculata, D. capreensis, and D. palpata) were studied. Two more species, one from the Caucasus, D. euxina, and one from Greece, D. remyi, were also included in the analyses, together with more distant species within the same family to be used as outgroups. Fifteen hundred base pairs of mitochondrial DNA, corresponding to the small subunit of the ribosomal RNA (16S rRNA) and to the subunit I of the cytochrome oxidase I (COI), were sequenced in order to clarify the phylogenetic relationships and biogeography of this group of Mediterranean cave crickets. The molecular data are congruent with a phylogeographic pattern; with the geographically close species also the most related ones. Based on mtDNA divergence, the present-day distribution of genetic diversity seems to have been impacted by climatic events due to glacial and interglacial cycles that have characterized the Pleistocene era.  2005 Elsevier Inc. All rights reserved. Keywords: Phylogeography; Biogeography; Cytochrome oxidase I; 16S rRNA; Mitochondrial DNA; Divergence times; Cave crickets; Dolichopoda

1. Introduction The family Rhaphidophoridae (Orthoptera, Gryllacridoidea) includes a large number of cave-adapted genera and species with a worldwide distribution. Morphologically, the cave species are quite similar: all species have long legs and antennae, small eyes and poor pigmentation. Several species also occur in epigean habitats, being mostly forest dwellers in tropical areas. Caveadapted species occur in temperate Holarctic (included in the subfamilies Dolichopodinae, Rhaphidophorinae, and Ceuthophilinae), Afrotropical, Neotropical, and Australian regions (Macropathinae) and they are thought to derive from sylvicolous ancestors (Di Russo *

Corresponding author. Fax: +39 06 72595965. E-mail address: [email protected] (G. Allegrucci).

1055-7903/$ - see front matter  2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ympev.2005.04.022

and Sbordoni, 1998; Hubbel and Norton, 1978; Leroy, 1967). Fossil representatives of such fauna have been recorded in Europe from Baltic ambers dating back to the Oligocene. Species within the Mediterranean genera, Troglophilus and Dolichopoda, are thought to originate from such ancestors (Chopard, 1936). However, phyletic relations within Rhaphidophoridae are controversial (Gorochov, 2001). Object of this study are the West-Mediterranean species of genus Dolichopoda belonging to the subfamily Dolichopodinae. This subfamily includes also the Nearctic genera Hadenoecus and Euhadenoecus (Hubbel and Norton, 1978). Dolichopoda is a circum-Mediterranean genus, consisting of around 30 species distributed throughout the North Mediterranean regions from the Pyrenees to Turkish Armenia and the Caucasus. Some Dolichopoda species are strictly

154

G. Allegrucci et al. / Molecular Phylogenetics and Evolution 37 (2005) 153–164

allopatric and are well diVerentiated both morphologically and karyologically. On the other hand, low morphological diVerentiation is associated with identical or variable chromosome numbers (Baccetti, 1982; Saltet, 1967) in species that show quasi-parapatric ranges. The highest species diversity is present in insular and peninsular Greece and Italy. In particular, nine species of Dolichopoda occur in Italy, where they range from the Maritime Alps to the southern tip of the Italian peninsula (Fig. 1). Most species of this genus are strictly dependent upon caves. However, especially in the northern part of the range, several populations inhabit cave-like habitat such as rock-crevices and ravines and individuals are often observed outside in moist or mesic woods. In peninsular Italy, Dolichopoda populations often live in cellars, catacombs, aqueducts, Etruscan tombs, and other similar man-made hypogean environments. Population sizes can be small and constant over long periods, at least in natural caves (Carchini et al., 1983; Sbordoni et al., 1987). Based on morphology three subgenera have been described in insular and peninsular Italy: Dolichopoda (Dolichopoda), D. (Chopardina), and D. (Capraiacris) (Baccetti, 1975; Baccetti and Capra, 1959, 1970). One of the main morphological diVerences between these subgenera is presence or absence of spinulation on the appendages. The subgenus Dolichopoda includes the

highest number of species distributed throughout the range of the genus, except for some coastal areas and it is characterized by the presence of spines on the anterior tibiae. Species belonging to Chopardina subgenus are mostly found in insular and peninsular Tuscany in Italy and in Corsica, with a single species found also in Greece (Macedonia). Chopardina species are characterized by the presence of several spines also on the hind femurs. The subgenus Capraiacris includes only two species, restricted to the Giglio Island and Monte Argentario in the Tuscan archipelago. Its two species (D. aegilion and D. baccettii) are distinct from the species in the other two subgenera because of the lack of spines on the anterior tibiae and on the hind femur. For the present study we sequenced 1500 bp of a mitochondrial DNA, corresponding to two fragments of the small subunit of the ribosomal RNA (16S rRNA) and to the subunit I of cytochrome oxidase I (COI) in order to clarify the phylogenetic relationships and the historical processes that shaped the current geographic distributions of the 11 Dolichopoda species that occur in the Western Mediterranean area from the French Pyrenees to the southernmost of Italian peninsula (Fig. 1). To attempt to encompass a larger geographic window we also included two species from the easternMediterranean region, one from Greece and one from Caucasus (D. remyi and D. euxina, respectively). These

Fig. 1. Sampling sites and ranges of the West-Mediterranean Dolichopoda species studied. The symbol shape denotes species; shading of a given symbol denotes subspecies.

G. Allegrucci et al. / Molecular Phylogenetics and Evolution 37 (2005) 153–164

two species are formally included in the Chopardina and Dolichopoda subgenera, respectively. To understand the relation of the genus Dolichopoda within the family Rhaphidophoridae we included in the analyses four additional taxa, belonging to four diVerent genera within the family. These species are considered the closest living relative to the genus Dolichopoda (Hubbel and Norton, 1978). Based on genetic divergence, we attempted to reconstruct possible ancient scenarios explaining the present-day distribution of the West-Mediterranean Dolichopoda species. The Mediterranean region is well suited for testing biogeographic hypotheses. Its fauna and Xora has evolved through a complex interplay of geological and paleoclimatic vicariance events (Blondel and Aronson, 1999). The geological and paleontological scenarios are here compared with genetic divergence data in order to test alternative hypotheses such as dispersal and/or vicariance. These two mechanisms have been identiWed as the principal ones responsible for the formation of biogeographic patterns. Cave crickets as Dolichopoda are favorable organisms to test both dispersal and vicariance hypotheses, since genetic diVerentiation of populations and speciation could have been inXuenced by active dispersal during the moist and warm interglacials, and by vicariance during the glacial ones. During these periods refugial populations experienced extremely low dispersal opportunities, because of the widespread steppic environment, caused by the dry and cold climate.

2. Materials and methods 2.1. Taxon sampling A total of 52 individuals representing 25 populations of 13 species of the genus Dolichopoda and eight individuals of four outgroup species were sampled in this study (Table 1). The species within the subgenus Dolichopoda included in this study are: D. linderi from the Pyrenees, D. ligustica from the Maritime Alps, D. laetitiae from the northern-central Apennines, D. geniculata from the central-southern Apennines, D. capreensis endemic from the island of Capri, D. palpata (syn. D. calabra Galvagni) from the southernmost part of the Apennines, and D. euxina from the Caucasus (Russia, not included in Fig. 1). We analyzed all the species within the subgenus Chopardina: D. bormansi and D. cyrnensis from the island of Corsica, D. schiavazzii from the coastal Tuscany (Fig. 1), and D. remyi from Macedonia (Greece, not included in Fig. 1). Similarly, we studied the two species included in the subgenus Capraiacris: D. aegilion from the island of Giglio and D. baccettii from Monte Argentario, both in Tuscany (Fig. 1). As outgroups we

155

used Ceuthophilus gracilipes, Troglophilus cavicola, Hadenoecus cumberlandicus, and Euhadenoecus insolitus. These taxa were chosen as representatives of diVerent subfamilies of Rhaphidophoridae. The Nearctic H. cumberlandicus and E. insolitus belong to the same subfamily Dolichopodinae as the species here studied, while the Palaearctic T. cavicola and the Nearctic C. gracilipes belong to the Troglophilinae and Ceuthophilinae subfamilies, respectively. Two individuals from each population were assayed to provide an approximate idea of intra-speciWc variation, but mostly verify the accuracy of the sequencing results. 2.2. Laboratory procedures DNA was isolated from leg muscle using a C-TAB protocol (Doyle and Doyle, 1987). Liquid nitrogen was used during the homogenization phase. Fresh, frozen or 95% ethanol-preserved specimens usually gave the same quality and quantity of DNA. Two overlapping fragments of cytochrome oxidase I gene (COI, total of 971 bp) and a 527-bp fragment of the 16S rRNA gene were ampliWed through the polymerase chain reaction (PCR) from each individual DNA sample. Primers used were J1751, N2191 (Simon et al., 1994), UEA1, UEA5, UEA8, and UEA10 (Lunt et al., 1996) for the COI gene, and 16Sa (Kocher et al., 1989) and 16Sb (Palumbi, 1996) for the 16S rRNA gene. Double-stranded ampliWcations were performed with a Perkin-Elmer-Cetus thermal cycler in 50
l of a solution containing genomic DNA (10–100 ng), 1.5 mM MgCl2, 2.5 mM of each dNTP, 0.5
M primer, 1 U of Amplitaq (Perkin-Elmer-Cetus) and the buVer supplied by the manufacturer. Optimal cycling parameters varied for each primer pair used. To obtain two overlapping fragments of COI (total of 971 bp) we performed a nested-PCR. In the Wrst PCR we used UEA1, UEA10 primers to obtain an invisible fragment of about 1500 bp. This fragment was then used as template to obtain two partially overlapped DNA fragments, 480 bp (using J1751 and N2198 primers) and 650 bp (using UEA5 and UEA8 primers) long. The Wrst PCR included a 10 min denaturation at 94 °C followed by 5 cycles at 98 °C for 15 s, 48 °C for 45 s, and 72 °C for 2 min 30 s and 30 cycles at 92 °C for 15 s, 52 °C for 45 s, and 62 °C for 2 min 30 s with a Wnal extension at 62 °C for 7 min. The nested-PCRs included a 10 min denaturation at 94 °C followed by 30 cycles at 95 °C for 30 s, 50 °C for 30 s, and 72 °C for 1 min 30 s with Wnal extension at 72 °C for 10 min. 16S PCR was performed as follows: after a 5 min denaturation step at 94 °C, each cycle of PCR consisted of denaturation for 30 s at 95 °C, annealing for 30 s at 48 °C, and extension for 45 s at 72 °C. After 35 cycles, a 5 min extension step at 72 °C followed. Double-stranded ampliWed products were

156

Table 1 Dolichopoda Species and outgroup taxa included in this study Collection locality

GenBank Accession No.

Outgroups Ceuthophilus gracilipes Troglophilus cavicola Hadenoecus cumberlandicus Euhadenoecus insolitus

Hamden, CT, USA Covoli di Veroli Cave, Veneto, Northern Italy Bat Cave, Carter Cave State Park, Carter Co., KY, USA Indian Grave Point Cave, De Kalb Co., TN, USA

16S: AY793561 COI: AY793593 16S: AY793564 COI: AY793624 16S: AY793562 COI: AY793592 16S: AY793563 COI: AY793591

Ingroups Dolichopoda Dolichopoda euxina linderi ligustica ligustica septentrionalis laetitiae

Vorontzovskaya Cave (VOR), Caucasus, Russia; Golova Otapa Cave (GOL), Caucasus, Russia Sirach Cave (SIR) Eastern Pyrenees, Western-South France Corno Cave (CON), Piemonte, Western-North Italy Pugnetto Cave (PUG), Piemonte, Western-North Italy Poscola Cave (PSC), Veneto, Northern-East Italy; Piane Cave (GDP), Umbria, Central Italy Diavolo Cave (DIA), Tuscany, Central Italy Valmarino Cave (VAL), Latium, Central-South Italy; Fontanelle Cave (FON), Campania, Southern-West Italy; Ischia cellars (ISC), Ischia Island, Campania, Southern-West Italy Roman Aqueduct(PNZ), Ponza Island, Latium, Southern-West Italy San Michele Cave (CPR), Capri Island, Campania, Southern-West Italy Tremusa cave (TRE), Calabria, Southern Italy

16S: AY793566/AY793565 COI: AY793622/AY793623 16S: AY793567 COI: AY793598/AY793599 16S: AY793568 COI: AY793604/AY793605 16S: AY793569 COI: AY793601/AY793602/AY793603 16S: AY793581/AY793582 COI: AY793611/AY793613/AY793610/AY793612 16S: AY793580 COI: AY793614/AY793615 16S: AY793583/AY793584/AY793585 COI: AY793616/AY793617 AY793594/AY793595 16S: AY793586 COI: AY793596/AY793597 16S: AY793587 COI: AY793606/AY793607 16S: AY793588 COI: AY793608/AY793609

Punta degli Stretti Cave (PST), Tuscany, Central-West Italy Campese Mine (CAM), Giglio Island, Tuscany, Central-West Italy

16S: AY793571 COI: AY793639/AY793640 16S: AY793570 COI: AY793600 16S: AY793589/AY793590 COI: AY793637/AY793638

schiavazzii caprai bormansi

Edessa Cave (EDE), Macedonia, Northern-East Greece; Pozarska Cave (POZ), Macedonia, Northern-East Greece Pipistrelli Cave (ORS), Tuscany, Central-West Italy; Marciana Cave (MRC), Elba Island, Tuscany, Central-West Italy Fichino Cave (FIC), Tuscany, Central-West Italy Brando Cave (BRA), Corsica Island, France; Sisco Cave (SIS), Corsica Island, France

cyrnensis

Valletto Cave (VLT), Corsica Island, France; Sabara Cave (SAB), Corsica Island, France

laetitiae etrusca geniculata geniculata pontiana capreensis palpata Dolichopoda Capraiacris baccettii aegilion Dolichopoda Chopardina remyi schiavazzii

16S: AY793573/AY793572 COI: AY793633/AY793635 16S: AY793574 COI: AY793634/AY793636 16S: AY793578/AY793579 COI: AY793631/AY793632/ AY793627/AY793625/AY793626/AY793628 16S: AY793577/AY793576 COI: AY793620/AY793621/ AY793618/AY793619

G. Allegrucci et al. / Molecular Phylogenetics and Evolution 37 (2005) 153–164

Species

G. Allegrucci et al. / Molecular Phylogenetics and Evolution 37 (2005) 153–164

checked for the expected size by electrophoresis of 1/10 of the product through a 1% agarose gel. PCR fragments were puriWed by using the GFXTM DNA and Gel Band puriWcation kit (Amersham Pharmacia Biotech), directly sequenced (in both directions) using the BigDye terminator ready-reaction kit, and resolved on either ABI 310 or 3100 Genetic Analyzer (PE Applied Biosystems), following the manufacturer’s protocols. Sequence data were edited and compiled using Sequencher version 4.1 (Gene Codes). All sequences were submitted to GenBank (Accession Nos. AY793561–AY793640). 2.3. Phylogenetic analysis For the 16S rRNA fragment, DNA sequences were aligned using CLUSTAL X 1.81 (Thompson et al., 1997) with opening gap D 10 and extending gap D 0.10. The alignment, also checked by eye, did not require further improvement by considering a secondary structure model. Cytochrome oxidase I nucleotide sequences were assembled, aligned, and translated with Sequencher 4.1 (Gene Codes). Phylogenetic structure in the data resulting from both COI and 16S genes was assessed using the permutation tail probability (PTP test; Faith, 1991) as implemented in PAUP* version 4.0b10 (SwoVord, 2000), with 1000 random matrices and randomizing of ingroup taxa only. Phylogenetic analyses were performed using maximum-parsimony (MP; Farris, 1970), and maximum-likelihood (ML; Felsenstein, 1981) methods, as implemented in PAUP*. A Bayesian analysis was performed using MRBAYES version 3.0 (Ronquist and Huelsenbeck, 2003). MP trees were inferred with a heuristic search using stepwise addition of taxa with 10 random replications and ACCTRAN character-state optimization. Gaps were treated as missing data. The consistency index (CI; Kluge and Farris, 1969), calculated after the exclusion of uninformative characters (Sanderson and Donoghue, 1989), was used to examine overall homoplasy levels. MP searches were run using both all substitutions un-weighted and using only transversions (Tv) in third position for the COI data set. The latter was used because saturation analyses revealed a small amount of saturation of third codon transitions (analysis available from the authors). The appropriate model of DNA substitution for ML analysis was determined using MODELTEST version 3.06 (Posada and Crandall, 1998) for each of the two genes considered, separately. Bayesian analysis was performed using four search chains for 1,000,000 generations, sampling trees every 100 generation and considering the same model as in ML analysis. The Wrst 1000 trees were discarded as burn-in. Data were partitioned by gene and, for COI, by codon position. Site-speciWc rate variation was also

157

calculated. Parameter stability was estimated by plotting log-likelihood values against generation time, and a consensus tree with posterior probabilities was then generated in PAUP*. To check if the data from the two mtDNA fragments could be combined we used the partition-homogeneity test (Farris et al., 1994, 1995) as implemented in PAUP*, with 10,000 iterations including and excluding uninformative sites. Bootstrap supports for the resulting topologies were calculated using 1000 replicates for MP and 100 replicates for the ML tree, as implemented in PAUP*. Twelve competing phylogenetic hypotheses were tested using the approximately unbiased (AU; Shimodaira, 2002) tree selection test in the software package CONSEL (Shimodaira and Hasegawa, 2001). For comparison we also performed the Shimodaira–Hasegawa test (SH; Shimodaira and Hasegawa, 1999). A Mantel test (1967), considering all ingroup taxa, was carried out to assess possible correlation between genetic and geographic distances. To compare the mtDNA data with previously studied allozymes data (Allegrucci et al., 1992, 1997; Di Russo, 1993 and unpublished data) we carried out a neighbor joining (NJ; Saitou and Nei, 1987) analysis, using ML distances based on the parameters inferred by MODELTEST. Allele frequencies were combined from the diVerent studies and a NJ tree was constructed on the basis of Nei’s (1978) genetic distance index. 2.4. Dating of the cladogenetic events To test whether signiWcant rate diVerence occurs among the whole group of considered species, we carried out for each gene, 240 Tajima’s relative rate tests (1993), using both the 1D (all substitutions combined) and 2D (Ti and Tv considered separately) methods, as implemented in MEGA version 2.1 (Kumar et al., 2001). To approximate absolute ages of divergence among haplotypes, we applied substitution rates previously reported for insect mitochondrial COI and 16S genes and calibrated using geological evidence. Two COI substitution rates (i.e., 1.2 and 2.3% per lineage, per million years) were used. The Wrst one was calculated for cave beetles (Coleoptera, Bathyiscinae, Caccone and Sbordoni, 2001) and the second one was obtained by combining multiple rates from a variety of insects (Brower, 1994). Only one substitution rate was used for 16S gene (i.e., 1.1% per lineage, per million years). For this rate, reported in Brower (1994), no molecular evolutionary model was available. These rates were applied to nucleotide divergences among haplotypes estimated using Tamura and Nei (1993) distances to follow the same model of molecular evolution as in Caccone and Sbordoni (2001).

158

G. Allegrucci et al. / Molecular Phylogenetics and Evolution 37 (2005) 153–164

3. Results 3.1. Levels of sequence variation in the 16S and COI genes A total of 971 bp from COI gene was sequenced for each individual. Percent of sequence divergence ranged from 19.6 to 3% between species and from 0.1 to 1% within species, with one to three haplotypes found in each species. The highest inter-speciWc values were usually associated to comparisons including D. euxina and/ or D. remyi, the two Eastern Mediterranean species. Among all considered species, 332 sites are variable and 267 are parsimony-informative. Out of 267 parsimonyinformative sites, 215 are found in the third positions. The transitions/transversions (Ti/Tv) ratio ranged from 0.6 to 6. Ti’s accounted for about 52 or 67% of all substitutions, if outgroups are included or excluded, respectively. This suggests some Ti’s saturation due to multiple substitutions, especially when the outgroup comparisons were included. This is also revealed by the plot of the number of Ti’s and/or Tv’s vs. genetic distances (data available from the Wrst author). However, signiWcant phylogenetic structure was detected with PTP test (P D 0.001). Total alignment length for the 16S rRNA sequences was 552 bp (obtained using CLUSTAL X). Percent of sequence divergence ranged from 0.2 to 5% between spe-

cies and from 0 to 0.6% within species, with one haplotype found in each species. When all species were considered 157 sites were variable and 97 parsimonyinformative, the Ti/Tv ratio ranged from 0.39 to 7. The partition-homogeneity test did not reject the null hypothesis that COI and 16S data sets are any diVerent from random partitions of the pooled data (P D 0.998), this was true also when we tested codon positions within COI (P D 0.996). Thus, we combined both data for all subsequent analyses. Results from separate data set were congruent to the results of the combined data set and available from the authors upon request. The combined data set included 1523 sites, 504 of which were variable and 360 were parsimony-informative. SigniWcant phylogenetic structure was detected by the PTP test (P D 0.001). MODELTEST indicated the general time reversible model with among-site rate heterogeneity and proportion of invariant sites (GTR + I + ; Gu et al., 1995; Lanave et al., 1984; Yang, 1994) as the best Wt for both genes and the combined data set. Rate matrix parameters for the combined data set were: A–C D 2.646, A–G D 15.640, A–T D 10.715, C–G D 2.103, C–T D 50.026, G–T D 1.000. Base frequencies were: A D 0.325, C D 0.141, G D 0.156 and T D 0.379. Among-site rate variation was approximated with the gamma distribution shape parameter
D 0.812. Proportion of invariant sites was 0.539. These

Fig. 2. Relationships among species of Dolichopoda, inferred by maximum likelihood (ML) from mtDNA sequences. Values above and below branches indicate bootstrap percentage for MP and ML methods and posterior probabilities derived from Bayesian analysis (Wrst, second, and third value, respectively). Only bootstrap values higher than 50% are reported.

G. Allegrucci et al. / Molecular Phylogenetics and Evolution 37 (2005) 153–164

parameters were used for subsequent phylogenetic analyses. 3.2. Phylogenetic analysis of the combined mtDNA data The MP search using equal weights for all substitution produced two most parsimonious trees 1226 steps long (CI D 0.582). Fig. 2 shows the tree obtained under ML analysis. On each node bootstrap values (BP), relative to MP or ML, and posterior probability (PP) values are reported. Generally, this topology is strongly supported in all performed analyses. Within the Italian species, three main groups could be distinguished (Fig. 2). The Wrst group included the continental species from Liguria to Southern Italy (D. ligustica, D. laetitiae, and D. geniculata). The second group was represented by the southernmost species D. palpata and its sister taxon D. capreensis, endemic of the island of Capri. All these species belong to subgenus Dolichopoda. The third group consisted of the species belonging to the subgenera Capraiacris and Chopardina, from continental and insular Tuscany and Corsica (D. aegilion, D. baccettii, D. schiavazzii, D. cyrnensis, and D. bormansi). Each of these three clusters is strongly supported with BP values ranging from 65 to 100% and PP values ranging from 95 to 100%. D. euxina (subgenus Dolichopoda) from Caucasus was the most diVerentiated species. The Pyrenean species, D. linderi, clustered as the sister taxon of the rest of Italian-Corsican species, even if it belongs to subgenus Dolichopoda. The Greek species, D. remyi (subgenus Chopardina), linked outside the Italian-Corsican group of species. D. euxina (subgenus Dolichopoda) from the Caucasus clustered as basal to all the other Dolichopoda species included in this study. All these nodes were highly supported both from BP and PP values and suggested that both Dolichopoda and Chopardina subgenera are polyphyletic. This is also conWrmed by results from AU and SH likelihood ratio tests carried out for a set of alternative trees, suggested by traditional taxonomy, and compared to our overall optimal tree. The AU and SH tests are employed to determine the conWdence set of trees (Shimodaira, 2002; Shimodaira and Hasegawa, 1999). The AU test is a powerful test for general tree selection problems involving a priori and a posteriori hypotheses. The SH test is biased against rejecting trees when comparing many trees and it is provided here for comparison. Eight of the 12 suboptimal trees can be rejected in favor of optimal tree using AU test, while Wve can be rejected using the more conservative SH test (data available from the Wrst author). In particular, topologies considering D. euxina as sister relative to the other species belonging to subgenus Dolichopoda or D. remyi as sister of the other species belonging to subgenus Chopardina are rejected by both AU and SH tests in favor of our optimal tree (P < 0.001, in all cases). Within subgenus Dolichopoda, three alternative topolo-

159

gies were considered. The Wrst one considers D. capreensis/D. palpata clade as sister relative to the D. geniculata populations and it is rejected by both AU and SH tests (P < 0.05), while in the second one this clade is sister to D. geniculata/D. laetitiae/D. ligustica cluster and it is accepted by both AU and SH tests (P D 0.279 and 0.923, respectively). Placement of D. linderi as sister relative to D. ligustica or to the Tuscan/Corsican species is not supported by AU test (P < 0.001). Finally, topology considering D. ligustica as sister of Chopardina–Capraiacris species is not accepted by both AU and SH tests (P < 0.05). Results from Mantel (1967) test, carried out to explore a possible correlation between geographic and genetic distance in all studied taxa, suggested a strong phylogeographic pattern with correlation coeYcient, r D 0.785 (
< 0.05). 3.3. Comparison of diVerent genetic markers In our laboratory, Dolichopoda cave crickets have been the object of extensive genetic research. Allozyme variation was thoroughly investigated in several populations of D. laetitiae, D. geniculata, D. baccettii, D. aegilion (Sbordoni et al., 1985), D. linderi (Di Russo, 1993; Sbordoni et al., 2000), D. schiavazzii (Allegrucci et al., 1997), D. ligustica, D. capreensis, and D. palpata (unpublished data). Allozyme data were not available only for D. euxina and D. remyi. Combining allele frequencies for 22 loci from diVerent studies we obtained the NJ tree illustrated in Fig. 3A. This tree, based on Nei’s (1978) genetic distance index, is compared to the NJ tree from present data set including exactly the same populations and species (Fig. 3B). Generally, the two topologies are almost identical but, as shown by branch lengths there are striking diVerences at the inter- and intra-speciWc level, with allozymes 15 times running faster than mtDNA. 3.4. Dating of cladogenetic events Results from the Tajima’s (1993) relative rate tests indicated that the molecular clock hypothesis could not be rejected in the whole group of considered species. None of the 480 comparisons performed returned a statistically signiWcant value (P ranging from 0.100 to 1.00, data available from Wrst author). To date the main lineage-splitting events in Dolichopoda group of species, we used substitution rates previously reported in other insect species for mitochondrial COI and 16S genes (Brower, 1994; Caccone and Sbordoni, 2001; see also Section 2). Given the large error associated with the use of this approach, resulting divergence times must be treated with caution as tentative estimates (Grauer and Martin, 2004). In particular, according to our estimates based on a range of available

160

G. Allegrucci et al. / Molecular Phylogenetics and Evolution 37 (2005) 153–164

Fig. 3. Genetic relationships in diVerent populations and species of Dolichopoda based (A) on allozymes and (B) on mtDNA data. Values at nodes are bootstrap percentage of 1000 replications for neighbor joining method. Only bootstrap values higher than 50% are reported.

rates (from 1.1 to 2.3%, per lineage, per million years) divergence times between the coastal group (subgenera Capraiacris and Chopardina) and the inland group of species (Dolichopoda spp.) range from 2.4 to 1.2 million years (Myr) ago. The same temporal window is outlined when divergence time is estimated between D. Capraiacris–D. Chopardina and D. linderi species. Narrower temporal windows are delineated when comparison are made within the coastal group or the inland group of species, the divergence time estimates ranging from 2.1 and 0.9 and from 1.6 to 0.8 Myr ago, respectively.

4. Discussion 4.1. Phylogenetic reconstruction As outlined in Section 3, three main groups can be distinguished within the West-Mediterranean species (Fig. 2). The Wrst one includes the inland species distributed from Liguria to southern Italy, the second one comprises the coastal species occurring in Corsica and in coastal Tuscany, and the third one consists of the southernmost species occurring in Calabria and on the Island of Capri. Within the Wrst group, the relationships among populations of D. geniculata are not completely resolved. At least some populations of this species are closely related to D. laetitiae, and results from hybridization studies in laboratory and in nature, indicated the lack of pre-mating barriers (Allegrucci et al., 1982; Sbordoni et al., 1987). Allozyme studies indicated that peripheral popu-

lations of D. geniculata Xanking the Tyrrhenian and the Adriatic coasts were genetically more diVerentiated from each other, unlike the inland, montane group of populations. This group, consisting of both D. laetitiae and D. geniculata populations, is genetically more homogeneous, regardless of their taxonomic assignment and/or their geographic distance (Allegrucci et al., 1987; Cesaroni et al., 1997; Sbordoni et al., 1985). Our choice to sample coastal populations of D. geniculata for this study was based on these hints. However, results from COI and 16S data sets, while emphasizing the close link between D. laetitiae and D. geniculata, do not clearly resolve relationships among D. geniculata populations. Comparison of the results from the diVerent data sets (allozymes, morphology, and mitochondrial DNA) suggests that D. geniculata is a complex of sibling species gradually divergent from D. laetitiae. Dolichopoda ligustica, a species distributed in northwest Italy, appears to be more closely related to the laetitiae–geniculata group than to the other species included in this study. This is supported by high BP and PP values in the ML and Bayesian analyses (75 and 96%, respectively, Fig. 2). These results along with the AU test falsify the hypothesis of D. ligustica as a sister taxon of the Pyrenean species D. linderi. Actually, D. ligustica is the westernmost species among the Italian ones and it could be considered as putative sister taxon to D. linderi, based on geographical considerations. A controversial point concerns the relationships between the Tuscany insular endemics D. aegilion and D. baccettii, for which the subgenus Capraiacris has been erected, and the Tuscan-Corsican species included in the

G. Allegrucci et al. / Molecular Phylogenetics and Evolution 37 (2005) 153–164

subgenus Chopardina (D. schiavazzii, D. bormansi, and D. cyrnensis). These two subgenera were erected by Baccetti (1975) on the basis of the absence/presence of series of spines on the femurs and the dorsal margin of the protibiae. However, despite these diVerences, previous studies based on allozymes and scn DNA–DNA hybridization suggested a close aYnity between these two groups (Allegrucci et al., 1992), a result strongly supported by the present study. Within the Tuscan and Corsican species group, two main clusters can be distinguished; one constituted by the Tuscan species (D. aegilion, D. baccettii, and D. schiavazzii) and the other by the Corsican species (D. cyrnensis and D. bormansi). D. remyi, another putative Chopardina species occurring in Greece, branches outside the Italian-Corsican group of species, suggesting that the subgenus Chopardina is polyphyletic. This result is also conWrmed by AU and SH tests; suboptimal topologies considering D. remyi as sister of the other species belonging to subgenus Chopardina are rejected in favor of our optimal tree. Since the main morphological character deWning the subgenus Chopardina is the occurrence of a series of spines on the inferior margin of metafemura, it could be concluded that the presence/absence of spines on femurs is subjected to homoplasy. We argue that lack of spines is an adaptive character state, tracing the evolution to cave environment, as we can also observe in other species of Gryllidae where spines on femur become lost in cave obligate species (Desutter-Grandcolas, 1993; Leroy, 1967). It could be speculated that, in epigean crickets, occurrence of spines on legs represents a defence against vertebrate predators, which are absent in caves. The third clearly distinct clade in our phylogenetic reconstruction (Fig. 2) includes two other species of the subgenus Dolichopoda, D. capreensis and D. palpata. Their relatedness as sister taxa was already highlighted at morphological level, from male genitalia morphology (Capra, 1968); however, relationships of this clade with the other species are not resolved, as also indicated by the AU and SH tests. The Pyrenean species, D. linderi, is the sister taxon to the three clades described above, suggesting an older diVerentiation of this species from the Italian ones (Fig. 2). DiVerent mitochondrial haplotypes clustered following a geographic pattern, strongly supported by BP, PP analyses and Mantel test. This phylogeographical pattern typically results from isolation due to environmental or geographical barriers. The results shown in this study are in broad agreement with previous studies on Wve Italian Dolichopoda species based on diVerent genetic markers (allozymes, DNA–DNA hybridization, and RFLP mtDNA; Allegrucci et al., 1992; Venanzetti et al., 1993), since clustering trees produced with all these markers produced topologies reXecting the same topological relationships amongst the Wve species as obtained

161

in the present work. For a larger population data set, which included multiple populations of all the ingroup taxa in this study but for D. euxina and D. remyi, we could directly compare divergence levels between allozyme and mtDNA data (Fig. 3). Also for this data set the pattern of divergence between the two genetic markers is quite concordant, the main diVerence between the two trees being on the relative branch lengths (the mtDNA tree branches are 15 times shorter than the allozymic tree ones) rather than on topology. This result is not unexpected given our knowledge on the evolutionary rates in the two compartments of the genome tracked by these two markers. 4.2. Timing of the cladogenetic events Estimation of molecular rates is, in principle, a powerful tool to trace the tempo and mode of cladogenetic events; however, the reliability of such estimates strongly depends on the accuracy by which genetic distances are estimated and on the appropriateness of the calibration method (Arbogast et al., 2002; Bromham et al., 1999; Bromham and Penny, 2003; Grauer and Martin, 2004). The choice of calibration events is therefore crucial to the accuracy of molecular dating. In the present case, diVerent scenarios would be suitable to calibrate our molecular clock, based on diVerent interpretations of the complex geological history of the Mediterranean basin. The current distribution of Dolichopoda in Corsica and on the islands oV the coast of Tuscany could be the outcome of vicariance events dating back to the Messinian (5.5 Myr ago), during the Mediterranean salinity crisis when landmasses around the Mediterranean basin were reconnected as a result of the desiccation of the Mediterranean Sea (Boccaletti et al., 1990). Alternatively, the presence of Dolichopoda in Corsica and in Tuscany could be due to active dispersal during the Plio-Pleistocene (2– 0.5 Myr ago), when recurrent marine regressions led to the formation of land bridges, or a series of stepping stone islands, between Corsica and Tuscany (Burgassi et al., 1983; La Greca, 1990; Lipparini, 1976). More ancient paleogeographic events, such as the separation of the Corsican-Sardinian plate from the Pyrenees, making the Wrst move during the Oligocene–Miocene transition (between 29 and 24 Myr ago; Alvarez, 1972; Bellon et al., 1977; Carmignani et al., 1995), could, at least speculatively, explain the occurrence of Dolichopoda in Corsica and the Pyrenees. The present distribution of the Italian inland Dolichopoda species could be the outcome of vicariance and/or dispersal events caused by the alternation of glacials and interglacials during the whole Quaternary. Even if in our taxon samples several insular Dolichopoda are represented, the recurrent occasions for isolation and dispersal occurred from Miocene to Pleistocene make diYcult tracing a particular calibration event. Therefore, rather than exploiting paleogeographic

162

G. Allegrucci et al. / Molecular Phylogenetics and Evolution 37 (2005) 153–164

information to calibrate the molecular clock, we decided to use molecular rates previously calculated for insects in both COI and 16S genes (Brower, 1994; Caccone and Sbordoni, 2001). By this procedure, we attempt to consider geological and paleoclimatic scenarios for speciation as dependent variables, rather than predetermined bases for calibration. Based on these molecular rates, ranging from 1.1 to 2.3% (per lineage, per million years), the approximate timing of divergences among major haplotype clusters would correspond to the paleoclimatic events that inXuenced the West Mediterranean region during the Pleistocene. Whatever COI or 16S rate we consider, the same broad temporal window is outlined. The most important speciation events, driving the subdivision of species belonging to diVerent subgenera would have occurred between 2.4 and 1.2 Myr ago. These events date back to the Plio-Pleistocene era, when repeated changes between marine regressions and transgressions might have favored dispersal and allopatric separation of the Chopardina–Capraiacris species, respectively. Interestingly, a comparable time of divergence was hypothesized for evolutionary splitting of the isopod Stenasellus racovitzai species group, which is distributed in Corsica and Sardinia islands and in coastal Tuscany. According to Ketmaier et al. (2003) the presence of Stenasellus in coastal Tuscany might be due to dispersal through a continuous hydrographic system on a land bridge that connected these regions during the Quaternary. Also the presence of sister taxa of Dolichopoda in Corsica and in coastal Tuscany might be due to active dispersal through habitat suited corridors connecting the surfaced lands, during the Pleistocene. The subsequent isolation and speciation of insular taxa could have occurred during one of the recurring marine transgressions. Most of these speciation episodes should have started between 1.2 and 0.9 Myr ago. At this time, around the Calabrian–Ionian transition an important marine transgression took place, lasting at least 500,000 years. This transgression represents a likely event driving genetic diVerentiation in isolation of the coastal and the insular Dolichopoda in the Capraiacris– Chopardina group. Moreover, allozyme data indicated a divergence time between the two Capraiacris species, D. aegilion and D. baccettii, corresponding to 800,000 years ago (Allegrucci et al., 1992) and between D. baccettii and D. schiavazzii dating to 940,000 years ago (Sbordoni et al., 1985), in excellent agreement with present mitochondrial DNA dating, if 2.3%, per lineage, per million years, (for COI) and 1.1%, per lineage, per million years, (for 16S) molecular rates are considered. Divergence time between D. linderi and the Chopardina–Capraiacris species would date back to 1.4–1.1 Myr, in disagreement with the hypothesis that the current distribution of Dolichopoda in Corsica and in Pyrenees derived from the much older connection of the Corsican-Sardinian plate to Pyrenees, and their subsequent separation in early Miocene.

Within the subgenus Dolichopoda, assuming 2.3% (per lineage, per million years) for COI and 1.1% (per lineage, per million years) for 16S molecular rates, the split between D. ligustica and D. laetitiae–D. geniculata clade would have occurred between 1.3 and 0.9 Myr ago, while divergence time estimates between D. laetitiae and D. geniculata are 0.8 Myr in all cases. These estimates date again back to the Pleistocene, when continental populations experienced repeated instances of active dispersal, during interglacial periods in mesic woods, alternating to episodes of population fragmentation and reduction of gene Xow during the dry cold climatic phases. During these periods ancestral epigean forest populations of Dolichopoda were forced to use forest remains and caves as refugia. In conclusion, we can observe that the temporal sequence of cladogenetic events is in good agreement with the geographic distribution of the studied taxa, supporting the issue that speciation events have been strictly allopatric and mostly determined by isolation of diVerent populations in isolated cave systems, following the refugium model (Barr and Holsinger, 1985; Sbordoni, 1982; Sbordoni et al., 2000). This model assumes fragmentation of a widespread epigean species and isolation of small populations in cave refugia, where divergence is mainly driven by genetic drift. Actually, the vast majority of Dolichopoda species is allopatric and comprises genetically well-diVerentiated populations, as indicated by low levels of current and historical gene Xow among geographically close conspeciWc populations (Cesaroni et al., 1997, and unpublished data). Acknowledgments The authors express their gratitude to Claudio Di Russo, Mauro Rampini, and Nikolai Mugue for collecting samples of D. linderi, D. remyi, and D. euxina, respectively. Gisella Caccone kindly provided the DNA from the North American cave crickets. Stefano De Felici assisted us in the use of UNIX system. We thank Saverio Vicario for valuable comments on the manuscript. References Allegrucci, G., Caccone, A., Cesaroni, D., Cobolli Sbordoni, M., De Matthaeis, E., Sbordoni, V., 1982. Natural and experimental interspeciWc hybridization between populations of Dolichopoda cave crickets. Experientia 38, 96–98. Allegrucci, G., Cesaroni, D., Sbordoni, V., 1987. Adaptation and speciation of Dolichopoda cave crickets (Orth. Rhaph.): geographic variation of morphometric indices and allozyme frequencies. Biol. J. Linn. Soc. 31, 151–160. Allegrucci, G., Caccone, A., Cesaroni, D., Sbordoni, V., 1992. Evolutionary divergence in Dolichopoda cave crickets: a comparison of single copy DNA hybridization data with allozymes and morphometric distances. J. Evol. Biol. 5, 121–148. Allegrucci, G., Minasi, M.G., Sbordoni, V., 1997. Patterns of gene Xow and genetic structure in cave-dwelling crickets of the Tuscan

G. Allegrucci et al. / Molecular Phylogenetics and Evolution 37 (2005) 153–164 endemic, Dolichopoda schiavazzii (Orthoptera, Rhaphidophoridae). Heredity 78, 665–673. Alvarez, W., 1972. Rotation of the Corsica–Sardinia microplate. Nature 235, 103–105. Arbogast, B.S., Edwards, S.V., Wakeley, J., Beerli, P., Slowinski, J.B., 2002. Estimating divergence times from molecular data on phylogenetic and population genetics timescales. Annu. Rev. Ecol. Syst. 33, 707–740. Baccetti, B., 1975. Notulae orthopterologicae. XXXII. Una nuova specie di Dolichopoda raccolta sull’Isola del Giglio. Boll. Ist. Entom. Univ. Bologna 31, 205–214. Baccetti, B., 1982. Ortotteri cavernicoli italiani (Notulae orthopterologicae XXXVI). Lav. Soc. Ital. Biogeogr. Verona 1978 (n.s.) 6, 195–200. Baccetti, B., Capra, F., 1959. Notulae Orthopterologicae. XII. Revisione delle specie italiane del genere Dolichopoda Bol. (Orth. Rhaph.). Redia 44, 165–217. Baccetti, B., Capra, F., 1970. Notulae Orthopterologicae. XXVII. Nuove osservazioni sistematiche su alcune Dolichopoda italiane esaminate anche al microscopio elettronico a scansione. Mem. Soc. Entomol. Ital. 48, 351–367. Barr Jr., T.C., Holsinger, J.R., 1985. Speciation in cave faunas. Annu. Rev. Ecol. Syst. 16, 313–337. Bellon, H., Coulon, C., Edel, J.B., 1977. Le déplacement de la Sardaigne synthèse des données géochronologiques, magmatiques et paléomagnétiques. Bull. Soc. Geol. France 7, 825–831. Blondel, J., Aronson, J., 1999. Biology and Wildlife of the Mediterranean Region. Oxford University Press, New York. Boccaletti, M., CiaranW, N., Cosentino, D., et al., 1990. Palinspatic restoration and paleogeographic reconstruction of the peri-Thyrrhenian area during the Neogene. Palaeogeogr. Palaeoclimatol. Palaeoecol. 77, 41–50. Bromham, L.D., Phillips, M.J., Penny, D., 1999. Growing up with dinosaurs: molecular dates and the mammalian radiation. Trends Ecol. Evol. 14, 113–118. Bromham, L.D., Penny, D., 2003. The modern molecular clock. Nature 4, 216–224. Brower, A.V.Z., 1994. Rapid morphological radiation and convergence among races of the butterXy Heliconius hereto inferred from patterns of mitochondrial DNA evolution. Proc. Natl. Acad. Sci. USA 91, 6491–6495. Burgassi, P.D., Decandia, F.A., Lazzaretto, A., 1983. Elementi di stratigraWa e paleogeograWa nelle colline metallifere (Toscana) dal Trias al Quaternario. Memorie della Società Geologica Italiana 25, 27–50. Caccone, A., Sbordoni, V., 2001. Molecular biogeography of cave life: a study using mitochondrial DNA from Bathysciinae beetles. Evolution 55, 122–130. Capra, F., 1968. Una nuova specie dell’isola di Capri. Fragm. Entomol. 6, 39–44. Carchini, G., Rampini, M., Severini, C., Sbordoni, V., 1983. Population size estimates of four species of Dolichopoda in natural and artiWcial caves of Central Italy (Orth. Raph.). Mém. Biospéol. 10, 341– 347. Carmignani, L., Decandia, F.A., Disperati, L., Fantozzi, P.L., Lazzarotto, A., Liotta, D., Oggiano, G., 1995. Relationships between the tertiary structural evolution of the Sardinia-Corsica-Provencal domain and the northern Apennines. Terra Nova 7, 128–137. Cesaroni, D., Matarazzo, P., Allegrucci, G., Sbordoni, V., 1997. Comparing patterns of geographic variation in cave crickets by combining geostatistic methods and Mantel tests. J. Biogeogr. 24, 419–431. Chopard, L., 1936. Biospeleologica N. LXIII. Orthopteres et Dermaptéres (Premiere serie). Arch. Zool. Exp. Gen. Paris 78, 195–214. Desutter-Grandcolas, L., 1993. The cricket fauna of Chiapanecan caves (Mexico): systematics, phylogeny and the evolution of troglobitic life (Orthoptera, Grylloidea, Phalangopsidae, Luzarinae). Int. J. Speleol. 22, 1–82. Di Russo, C., 1993. Adaptations et divergence évolutive des populations des Dolichopodes des Pyrénées orientales, These de Doctorat

163

en Ecologie et Biologie des Populations. Université Paul Sabatier de Toulose, France, 70pp. Di Russo, C., Sbordoni, V., 1998. Gryllacridoidea. Encyclopaedia Biospeologica 2, 979–988. Doyle, J.J., Doyle, J.L., 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochem. Bull. 19, 11–15. Faith, D.P., 1991. Cladistic permutation tests for monophyly and nonmonophyly. Syst. Zool. 40, 366–375. Farris, J.S., 1970. Methods for computing Wagner trees. Syst. Zool. 18, 374–385. Farris, J.S., Källersjö, M., Kluge, A.G., Bult, C., 1994. Testing signiWcance of incongruence. Cladistics 10, 315–319. Farris, J.S., Källersjö, M., Kluge, A.G., Bult, C., 1995. Constructing a signiWcance test for incongruence. Syst. Biol. 44, 570–572. Felsenstein, J., 1981. Evolutionary trees from DNA sequences, a maximum likelihood approach. J. Mol. Evol. 17, 368–376. Gorochov, A.V., 2001. The higher classiWcation, phylogeny and evolution of the superfamily Stenopelmatoidea. In: The Biology of Wetas, King Crickets and their Allies, Field ed., pp. 3–33. Grauer, D., Martin, W., 2004. Reading the entrails of chickens: molecular timescales of evolution and the illusion of precision. Trends Genet. 20 (2), 80–86. Gu, X., Fu, Y.X., Li, W.H., 1995. Maximum likelihood estimation of heterogeneity of substitution rate among nucleotide sites. Mol. Biol. Evol. 12, 546–557. Hubbel, T.H., Norton, R.M., 1978. The systematics and the biology of the cave-crickets of the North American tribe Hadenoecinni (Orthoptera Saltatoria: Rhaphidophoridae: Dolichopodinae). Misc. Publ. Mus. Zool. Univ. Mich. 156, 1–125. Ketmaier, V., Argano, R., Caccone, A., 2003. Phylogeography and molecular rates of subterranean aquatic Stenasellid Isopods with a peri-Thyrrenian distribution. Mol. Ecol. 12, 547–555. Kluge, A.G., Farris, J.S., 1969. Quantitative phyletics and the evolution of anurans. Syst. Zool. 18, 1–32. Kocher, T.D., Thomas, W.K., Meyer, A., Edwards, S.V., Paabo, S., Villablanca, F.X., Wilson, A.C., 1989. Dynamics of mitochondrial DNA evolution in animals: AmpliWcation and sequencing with conserved primers. Proc. Natl. Acad. Sci. USA 86, 6196–6200. Kumar, S., Tamura, K., Jakobsen, I.B., Nei, M., 2001. MEGA2: Molecular Evolutionary Genetics Analysis software, Arizona State University, Tempe, Arizona, USA. La Greca, M., 1990. The insect biogeography of west Mediterranean islands. Atti dei Convegni Lincei 85, 459–468. Lanave, C., Preparata, C., Saccone, C., Serio, G., 1984. A new method for calculating evolutionary substitution rates. J. Mol. Evol. 20, 86– 93. Leroy, Y., 1967. Gryllides et grillacrides cavernicoles. Ann. Speol. 22, 659–722. Lipparini, T., 1976. Per la storia del popolamento delle isole dell’Arcipelago Toscano (contributo geo-paleontologico). Lavori della Società Italiana di BiogeograWa 5, 13–25. Lunt, D.H., Zhang, D.X., Szymura, J.M., Hewitt, G.M., 1996. The insect cytochrome oxidase I gene: evolutionary patterns and conserved primers for phylogenetic studies. Insect Mol. Biol. 5, 153–165. Mantel, N., 1967. The detection of disease clustering and a generalized regression approach. Cancer Res. 27, 209–220. Nei, M., 1978. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 89, 583–590. Palumbi, S.R., 1996. Nucleic acids II: the polymerase chain reaction. In: Hillis, D.H., Moritz, C., Mable, B.K. (Eds.), Molecular Systematics. Sinauer, Sunderland, MA, pp. 205–248. Posada, D., Crandall, K., 1998. Modeltest: testing the model of DNA substitutions. Bioinformatics 14, 817–818. Ronquist, F., Huelsenbeck, J.P., 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19, 1572–1574.

164

G. Allegrucci et al. / Molecular Phylogenetics and Evolution 37 (2005) 153–164

Saitou, M., Nei, M., 1987. The neighbourhood-Joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4, 406–425. Saltet, P., 1967. Les Dolichopodes de Corse. I. Etude cytologique préliminaire. Bull. Soc. Hist. Nat. Touluse 103, 265–268. Sanderson, M.J., Donoghue, M.J., 1989. Patterns of variation in levels of homoplasy. Evolution 43, 1781–1795. Sbordoni, V., 1982. Advances in speciation of cave animals. In: Bigozzi, C. (Ed.), Mechanism of Speciation. A.R. Liss, New York, pp. 219– 240. Sbordoni, V., Allegrucci, G., Cesaroni, D., Cobolli Sbordoni, M., De Matthaeis, E.,1985. Genetic structure of populations and species of Dolichopoda cave crickets: evidence of peripatric divergence. In: Sbordoni, V. (Ed.), Genetics and Ecology in Contact Zones of Populations. Boll. Zool. 52, 95–114. Sbordoni, V., Allegrucci, G., Caccone, A., Carchini, G., Cesaroni, D., 1987. Microevolutionary studies in Dolichopodinae cave crickets. In: Baccetti, B. (Ed.), Evolutionary Biology of Orthopteroid Insects. Horwood Ltd. Publ., Chichester, UK, pp. 514–540. Sbordoni, V., Allegrucci, G., Cesaroni, D., 2000. Population genetic structure, speciation and evolutionary rates in cave dwelling organisms. In: Wilkens, H., Culver, D., Humphrey, B. (Eds.), Ecosystems of the World, Subterranean Ecosystems. Elsevier, Amsterdam, pp. 450–483 (Chapter 24). Shimodaira, H., 2002. An approximately unbiased test of phylogenetic tree selection. Syst. Biol. 51, 492–508.

Shimodaira, H., Hasegawa, M., 1999. Multiple comparisons of log-likelihoods with applications to phylogenetic inference. Mol. Biol. Evol. 16, 1114–1116. Shimodaira, H., Hasegawa, M., 2001. CONSEL: for assessing the conWdence of phylogenetic tree selection. Bioinformatics 17, 1246–1247. Simon, C., Frati, F., Beckenbach, A., Crespi, B., Liu, H., Flook, P., 1994. Evolution, weighting and phylogenetic utility of mitochondrial gene sequences and a compilation of conserved polymerase chain reaction primers. Ann. Entomol. Soc. Am. 87, 651–701. SwoVord, D., 2000. PAUP* B-Version 4.0. Sinauer, Sunderland. Tajima, F., 1993. Simple methods for testing the molecular evolutionary clock hypothesis. Genetics 135, 599–607. Tamura, K., Nei, M., 1993. Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Mol. Biol. Evol. 10, 512–526. Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., Higgins, D.G., 1997. The ClustalX windows interface: Xexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. 24, 4876–4882. Venanzetti, F., Cesaroni, D., Mariottini, P., Sbordoni, V., 1993. Molecular phylogenies in Dolichopoda cave crickets and mtDNA rate calibration. Mol. Phylogenet. Evol. 2, 275–280. Yang, Z., 1994. Maximum likelihood phylogenetic estimation from DNA sequences with variable rates over sites: approximate methods. J. Mol. Evol. 39, 306–314.